quinoafacts others

25

41

ReviewReceived: 10 March 2010 Revised: 6 August 2010 Accepted: 6 August 2010 Published online in Wiley Online Library: 2 September 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4158

Nutrition facts and functional potentialof quinoa (Chenopodium quinoa willd.),an ancient Andean grain: a reviewAntonio Vega-Galvez,a∗ Margarita Miranda,a Judith Vergara,a Elsa Uribe,a

Luis Puenteb and Enrique A Martınezc

Abstract

Quinoa, Chenopodium quinoa Willd., is an Amaranthacean, stress-tolerant plant cultivated along the Andes for the last7000 years, challenging highly different environmental conditions ranging from Bolivia, up to 4.500 m of altitude, to sea level,in Chile. Its grains have higher nutritive value than traditional cereals and it is a promising worldwide cultivar for humanconsumption and nutrition. The quinoa has been called a pseudo-cereal for botanical reasons but also because of its unusualcomposition and exceptional balance between oil, protein and fat. The quinoa is an excellent example of ‘functional food’that aims at lowering the risk of various diseases. Functional properties are given also by minerals, vitamins, fatty acids andantioxidants that can make a strong contribution to human nutrition, particularly to protect cell membranes, with provengood results in brain neuronal functions. Its minerals work as cofactors in antioxidant enzymes, adding higher value to its richproteins. Quinoa also contains phytohormones, which offer an advantage over other plant foods for human nutrition.c© 2010 Society of Chemical Industry

Keywords: Andean crops; functional foods; human nutrition; physiologically active compounds; quinoa; stress tolerance

INTRODUCTIONA strong earthquake and tsunami recently impacted our Chileanterritory.1 This strike left two million people without shelter andfood supplies.2 This situation is imposing a large-scale challengefor good-quality food to be readily available. One alternativesource of staple food is Chenopodium quinoa Willd., a crop presentin Chile, although poorly known, and it emerges as a good foodcandidate due to its exceptional nutritive value but also due tothe strong tolerance to stressing abiotic conditions. The genusChenopodium is distributed worldwide and includes 250 species.This review focuses on the nutritional and functional properties ofChenopodium quinoa, which is a tetraploid species, a close relativeof beets and amaranth that originated in the Andean region ofBolivia and Peru.3,4 It has been cultivated in this area for the last5000–7000 years5,6 and from there it was transmitted by livestockmigrations and traded to other ancient cultures to the northern(Venezuela) and southern extremes of South America, namelyArgentina and Chile.7 This is why it is known with different localnames, according to voices of different cultures such as ‘tupapasupha’ in Aymara, ‘suba’ in Chibcha, ‘ayara’ in Quechua, ‘dawe’in Mapudungun (southern Chile) or just quinoa or quinua. Thisplant was called by the Incas ‘the mother grain’ and it was given asacred status, a gift from their gods. After the Spanish conquest,it remained only where Europeans could not arrive and introducegrains such as wheat, rye and oat (Altiplano in the High Andesabove 3500 m above sea level (a.s.l.)) or in isolated regions whereroads are cut off in winter or where the ancient cultures still remainstrong and attached to their agricultural practices and to theirtraditional food consumption habits (aymaras in the northernChilean Altiplano, isolated farmers of the coast of central Chile

and within the mapuches people in southern Chile). Concerningthe more accessible arable lands, the European-introduced cropsreplaced quinoa.5,7,8 This separation and subsequent isolationdetermined a strong pattern of genetic differentiation: the highAndes ecotypes are genetically different from the southern ones, asdetected by microsatellites, a kind of highly polymorphic molecularmarker.9 About 3000 varieties are conserved in South Americangermplasm banks assuring conservation and characterization, andopening possibilities for informed interchange of seed materials.Quinoa has been authorized to be sown in Europe, North America,Asia and Africa.10 Particularly in Europe a project was approvedin 1993 entitled ‘Quinoa: a multipurpose crop for the EuropeanCommunity’, for agriculture diversification.11,12

While quinoa is an ancient crop, available technical informationregarding the properties of chemical composition and functionalproperties is limited. Therefore, this work is an updated reviewof the chemical composition, physiologically active compoundsand some functional properties of Chenopodium quinoa that givesthis grain outstanding potential in human nutrition. Its varied

∗ Correspondence to: Antonio Vega-Galvez, Department of Food Engineering,Universidad de La Serena, Av. Raul Bitran s/n, Box 599, La Serena, Chile.E-mail: [email protected]

a Department of Food Engineering, Universidad de La Serena, La Serena, Chile

b Department of Food Science and Chemical Technology, Universidad de Chile,7500906, Santiago, Chile.

c Center for Advanced Studies in Arid Zones, CEAZA, Universidad de La Serena,La Serena, Chile

J Sci Food Agric 2010; 90: 2541–2547 www.soci.org c© 2010 Society of Chemical Industry

25

42

www.soci.org A Vega-Galvez et al.

Figure 1. Quinoa panicules (Chenopodium quinoa Willd).

nutritional properties are better understood when the botanicaland environmental diversity of quinoa is also known. We start ourreview with a brief description on its great adaptation to abioticstress, a particularly relevant aspect in a world with strong trendstowards increased soil degradation, desertification and criticalclimatic change.

BOTANICAL AND AGRONOMICAL DIVERSITYOF QUINOA IN VARIED AND STRESSINGENVIRONMENTAL CONDITIONSQuinoa is a plant that produces grains even if cultivated upto 4500 m a.s.l. and with higher nutritive value than traditionalcereals,8 as for instance Amarilla de Marangani and Blanca de Junın,two commercial varieties grown in greater proportion in the Andesof southern Peru.13 Most varieties of quinoa commonly differ inthe morphology, phenology and the chemical composition of thetissues.14

The quinoa (a dicot plant) is not a true grain, like typical cereal(monocot) grains, it is rather a fruit, so that it has been calleda pseudo-cereal and even a pseudo-oilseed. This is also becauseof its unusual composition and exceptional balance between oil,protein and fat.15

Quinoa, according to sowing density, can grow from 1 to 3 mhigh. The seeds can germinate very fast, i.e. in a few hours afterhaving been exposed to moisture. The roots can reach a depth ofup to 30 cm if sown deep in the soil. The stem is cylindrical, 3.5 cmin diameter; it can be either as a straight stem or branched and itscolor is variable. Depending on the variety, it changes from white,yellow or light brown to red. Leaves are shaped like a goose foot.The flowers are incomplete and do not have petals. Quinoa hasboth hermaphrodite flowers, located at the distal end of a group,and female flowers, located at the proximal end(Fig. 1).16

Quinoa seeds are round and flattened, and they may measurefrom about 1.5 mm in diameter to 4 mm; about 350 seeds weigh1 g.17 Seed size and color are variable.18 Seed colors go from whiteto grey and black, potentially having tones of yellow, rose, redand purple and violet, often with very colorful mixes in the samepanicule, where black is dominant over red and yellow, whichin turn are dominant to white seed color (as seen in Fig. 1).19

The classification of quinoa was first made from the color of theplant and fruits. Subsequently, it was based on the morphological

Table 1. Main uses of quinoa

Main uses Component implied Plant organ

Foods and drinks Vitamins Seeds and leaves

Proteins Seeds and leaves

Animal food Vitamins Whole plant

Proteins Whole plant

Medicine Immune system Leaves and seeds

Skin applications Leaves and seeds

Circulatory applications Leaves and seeds

Repellent Insects Leaves and seed coat

FAO.26

types of the plant. Despite the wide variation observed, quinoa isconsidered to be one single species.16

The cultivation of quinoa is related to crop rotation withpotatoes, also a crop of Andean origin. This is a common practice,which improves quinoa yield and preserves soil fertility. Moreover,the biological cycle of several pathogenic microorganisms isbroken down.16

The cultivation cycle lasts 8 months in the high Andes but itcan be as short as 4 months in arid central Chile.20 It is sown inNovember in the Altiplano, close to the Equator (close to 12 hdaylight) and from September to August in the lowlands of moresouthern latitudes (longer spring and summer days). Maturationand harvest, according to daylength, is done in May in the Altiplanoand from February to March in the center–south of Chile. Here,some ecotypes could attain maturity and seed production underirrigation equivalent to only 50 mm of rainfall per season, whichis an extremely low irrigation for any crop species.21 It also seemsto have exceptional physiological adaptations for high wateruse efficiency under stomatal closure21 besides efficient roots forwater capture, as earlier pointed out by Wood.22 In arid regions theaddition of organic matter also increases water use efficiency andgrain yields.21 Strong tolerance has been also demonstrated forother stressing conditions such as salty soils and cold climate.12,23

Quinoa can be grown on various types of soils, including marginalsoils,8 under a wide range of acid/alkaline conditions (from pH6.0 to 8.5). The plant is not affected from around −1 ◦C. However,it tolerates high temperatures up to 35 ◦C. Quinoa is resistantto freezing temperatures if the frost occurs before flowering.However, if the frost occurs after flowering, significant damagemay affect the plant.

As mentioned above, quinoa is a drought-tolerant crop havinglow water requirements, although yield is significantly affectedby irrigation.24 It is able to develop even in regions where theannual rainfall is in the range of 200–400 mm,16 but it has beenproven that it can be grown in southern Chile with an annualprecipitation as high as 3000 mm.20 Although having good aresponse in poor soils, quinoa does respond well to nitrogenfertilization. Thus nitrogen significantly increases seed productionand protein content.25

BIOCHEMICAL AND NUTRITIONALCOMPOSITION OF QUINOAConsumption of seeds is the most common use of quinoa (Table 1)and the review will be focused on its composition (Table 2).However, consumption of sprouts is becoming increasingly

wileyonlinelibrary.com/jsfa c© 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 2541–2547

25

43

Nutrition and functional potential of quinoa www.soci.org

Table 2. Proximate analysis of quinoa (g 100 g−1 fresh weight)

References

Component Kozioł27 Wright et al.28 De Bruin29 Dini et al.30

Protein 16.5 16.7 15.6 12.5

Fat 6.3 5.5 7.4 8.5

Ash 3.8 3.2 3.0 3.7

Carbohydrate 69.0 74.7 69.7 60.0

Crude fiber 3.8 10.5 2.9 1.92

popular among people interested in improving and maintainingtheir health status by changing dietary habits. The seeds andsprouts are both excellent examples of ‘functional food’, definedas lowering the risk of various diseases and/or exerting health-promoting effects, in addition to its nutritive value.31 Most ofthe recently published papers are focused mainly on studies oftypical sprouts such as buckwheat, broccoli, mung bean, andsoybean, which are already readily available on the market. Thesprouts of amaranth and quinoa are ‘new’ vegetables, which canbe used in vegetarian nutrition and as a common diet too.31

Today’s health-conscious consumers are illustrating a preferencetowards value-added products. The opportunity to supplement orcompletely replace common cereal grains (corn, rice and wheat)with a cereal of higher nutritional value (such as quinoa) isinherently beneficial to the public interest.32 We will review theknown composition and nutritional facts reported for quinoa,before describing the potential for functional properties and forhuman health, particularly for certain consumers (the elderly,children, high-performance athletes, diabetics, celiacs, people whoare gluten or lactose intolerant).

ProteinsProtein nutritional quality is determined by the proportions ofessential amino acids, which cannot be synthesized by animalsand hence must be provided in the diet. If only one of these aminoacids is limiting, the others will be broken down and excreted,resulting in poor growth of livestock and humans and loss ofnitrogen in the diet. Ten amino acids are strictly essential: lysine,isoleucine, leucine, phenylalanine, tyrosine, threonine, tryptophan,valine, histidine and methionine, all of which are present in quinoa(Table 3), providing it with a similar value to casein, the proteinof milk.33,35 Koziol27 showed that protein content in quinoa grainranges from 13.8% to 16.5%, with an average 15%. Wright et al.28

reported a protein content of 14.8% and 15.7% for sweet andbitter quinoa, respectively, from Bolivia. De Bruin29 studied theprotein content of four genotypes of quinoa, reporting a range of12.9–15.1%.

According to values indicated by FAO/WHO/UNU,36,37 quinoaprotein can supply around 180% of the histidine, 274% of theisoleucine, 338% of the lysine, 212% of the methionine + cysteine,320% of the phenylalanine+ tyrosine, 331% of the threonine, 228%of the tryptophan and 323% of the valine recommended in proteinsources for adult nutrition. In addition, the sulfur-containing aminoacids cystine, and methionine are found in concentrations thatare unusually high compared to other plants,38 probably dueto the type of land (volcanic) where this plant originated. Thecontent of essential amino acids in quinoa (Table 3) is higher thanin common cereals.28,39 Mahoney et al.,40 working with Bolivianquinoa, concluded that protein contained high amounts of lysine

Table 3. Essential amino acid profile (g 100 g−1 protein)

References

Aminoacid Kozioł27

Diniet al.30

Repo-Carrascoet al.33

Wrightet al.28

Gonzalezet al.34

His 3.2 2.0 2.7 3.1 ND

Ile 4.4 7.4 3.4 3.3 ND

Leu 6.6 7.5 6.1 5.8 ND

Met + Cys 4.8 4.5 4.8 2.0a 2.4a

Phe + Tyr 7.3 7.5 6.2 6.2 ND

Thr 3.8 3.5 3.4 2.5 ND

Val 4.5 6.0 4.2 4.0 ND

Lys 6.1 4.6 5.6 6.1 6.6

Trp 1.2 ND 1.1 ND 1.1

His, histidine; Ile, isoleucine; Leu, leucine; Met + Cys, methionine +cystine; Phe + Tyr, phenylalanine + tyrosine; Thr, threonine; Val, valine;Lys, lysine; Trp, tryptophan.ND, not detected.a Only methionine reported.

and methionine even though there is considerable variationbetween these varieties in the contents of such amino acids. Diniet al.,30 using decorticated quinoa, found that the composition ofquinoa is nutritionally comparable or superior to other commonlyconsumed cereals. When extracted, quinoa proteins solubilitycould be improved by enzymatic hydrolisis.41 Quinoa is alsoconsidered as one of the best leaf protein concentrate sources andso has potential as a protein substitute for food and fodder and inthe pharmaceutical industry.42

CarbohydratesStarch, the major biopolymeric constituent of plants (grains, seedsand tubers) occurs typically as granular forms of various shapes andsizes.43 Starch provides the major source of physiological energyin the human diet and accordingly it is classified, in general, asavailable carbohydrate.44 In quinoa, starch is the most importantcarbohydrate in all grains, making up approximately 58.1–64.2%of the dry matter, according to studies of Repo-Carrasco et al.,33

of which 11% is amylose.45,46 Granules of quinoa starch have apolygonal form with a diameter of 2 µm, being smaller than starchof the common grains. The extremely small size of the starchgranule can be beneficially exploited by using it as a biodegradablefiller in polymer packaging. Its excellent freeze–thaw stabilitymakes it an ideal thickener in frozen foods and other applicationswhere resistance to retrogradation is desired.47 In addition, quinoaflour contains high percentages of D-xylose and maltose, and lowcontents of glucose and fructose, which allows its use in malteddrink formulations.48 Also, its content of D-ribose and D-galactoseand maltose would result in a low fructose glycemic index. Repo-Carrasco et al.33 reported for quinoa 1.70 mg 100 g−1 of glucose,0.20 mg 100 g−1 of fructose, 2.90 mg 100 g−1 of saccharose and1.40 mg 100 g−1 of maltose.

MineralsQuinoa has a high content of calcium, magnesium, iron, copper andzinc.27,29,33,49 Many minerals in quinoa are found at concentrationsgreater than that reported for most grain crops. Providing they arefound in bioavailable forms, calcium, magnesium and potassiumare found in sufficient quantities for a balanced human diet38

J Sci Food Agric 2010; 90: 2541–2547 c© 2010 Society of Chemical Industry wileyonlinelibrary.com/jsfa

25

44


Table 4. Mineral composition (mg kg−1 dry weight)

Minerals

References Ca P Mg Fe Zn K Cu

Kozioł27 1487 3837 2496 132 44 9267 51

Repo-Carrasco et al.33 940 1400 2700 168 48 ND 37

Ruales and Nair17 874 5300 260 81 36 12000 10

Bhargava et al.10 1274 3869 ND 20 48 6967 ND

Konishi et al.50 863 4110 5020 150 40 7320 ND

Dini et al.30 275 4244 ND 26 27.5 75 ND

Sanders52 565 4689 1760 14 28 11930 2

Gonzalez et al.34 1020 1400 ND 105 ND 8225 ND

ND, not detected.

Table 5. Vitamin concentration (mg 100 g−1 dry weight)

References

Vitamin Kozioł27 Ruales and Nair17

Ascorbic acid (C) 4.0 16.4

α-Tocoferol (E) 5.37 2.6

Thiamin (B1) 0.38 0.4

Riboflavin (B2) 0.39 ND

Niacin (B3) 1.06 ND

ND,not detected.

(Table 4). For instance, Ruales and Nair39 and Ahamed et al.47

reported that iron, calcium and phosphorus levels are higher thanthose of maize and barley: iron was 81 mg kg−1 and calcium was874 mg kg−1. It also has about 0.26% of magnesium in comparisonto 0.16% of wheat and 0.14% of corn. Schlick and Bubenheim,38

comparing quinoa from different sources (USA, Peru, Bolivia andChile), reported that the mineral concentrations for quinoa seemto vary dramatically. This may occur due to the soil type andmineral composition of the region and/or fertilizer application.

With respect to the distribution of minerals within the grain,Konishi et al.50 used scanning electron microscopy with energy-dispersive X-ray detection on seed with and without epicarp,finding that minerals like P, K and Mg were located in theembryo, while Ca and P in the pericarp were associated withpectic compounds of the cell wall. Thus abrasive procedures toremove saponins might cause losses of 40% and 10%, respectively.Sulfur is found uniformly distributed within the embryo. The ironhas been reported as highly soluble and thus could be easilyavailable to anemic populations.51

VitaminsVitamins are compounds essential for the health of humans andanimals; according to their solubility they are divided into twogroups: hydro- and lipo-soluble. Traditionally, vitamins have beenamong the most widely applied chemical agents to enhance thenutritional values of food products. Some vitamins may also helpto lower the levels of toxic compounds formed in chemical re-actions such as the Maillard reaction.53 Table 5 shows the mainvitamins found in quinoa. The quinoa is found to be rich inα-carotene and niacin. Ruales and Nair39 have reported

Table 6. Unsaturated fatty acid (g 100 g−1 of oil extract)

Fatty acid

Reference Oleic Linoleic Linolenic

Kozioł27 23.3 53.1 6.2

Repo-Carrasco et al.33 26.0 50.2 4.8

Ruales and Nair17 24.8 52.3 3.9

appreciable amounts of thiamin (0.4 mg 100 g−1), folic acid(78.1 mg 100 g−1) and vitamin C (16.4 mg 100 g−1). Kozioł27 com-pared the vitamin contents of quinoa with some cereals (rice,barley and wheat) and reported that quinoa contains substantiallymore riboflavin (B2), α-tocopherol (vitamin E) and carotene thanthose cereals. In terms of a 100 g edible portion, quinoa supplies0.20 mg vitamin B6, 0.61 mg pantothenic acid, 23.5 g folic acid and7.1 g biotin.10 Repo-Carrasco et al.33 also reported that quinoa isrich in vitamin A, B2 and E. The content of vitamin E in quinoais important since this vitamin acts as an antioxidant at the cellmembrane level, protecting the fatty acids of the cell membranesagainst damage caused by free radicals.33

LipidsOil content in quinoa ranges from 1.8% to 9.5%, with an average of5.0–7.2%, which is higher than that of maize (3–4%).27 Numerousfatty acids are synthesized by the human body, and these areknown as ‘non-essential fatty acids’ because they are not essentiallyneeded in the diet.54 However, because the body cannot produceall the types of fatty acids it requires, some must come from thediet; these fatty acids are called ‘essential fatty acids’ or EFAs(Table 6). The EFAs are metabolized to longer-chain fatty acids of20 and 22 carbon atoms.37 There are two known families of EFAs:omega-3 (ω-3) and omega-6 (ω-6).55 Linoleic acid is metabolizedto arachidonic acid and linolenic acid to eicosapentaenoic acid(EPA) and docosahexaenoic acid (DHA). Linoleic acid is one ofthe most abundant polyunsaturated fatty acids identified inquinoa; polyunsaturated fatty acids have several positive effectson cardiovascular disease and improved insulin sensitivity.37 Thereported total lipid content in quinoa is 14.5% with an unsaturatedlevel of about 70%, having linoleic and oleic acids in percentagesof 38.9% and 27.7% respectively,30 while Ahamed et al.47 reportedin another study that quinoa fat had a high content of oleic acid(24%) and linoleic acid (52%). All fatty acids present in quinoaare well protected by the presence of vitamin E, which actsas a natural antioxidant.56 Repo-Carrasco et al.33 reported fromPeruvian quinoa the highest percentage of fatty acids being 50.2%for linoleic acid (ω-6), 4.8% of linolenic acid (ω-3).

ANTIOXIDANT ACTIVITYRecently, much attention has been given to naturally occurringantioxidants, which may play an important role in inhibiting bothfree radicals and oxidative chain reactions within tissues andmembranes.57

Therefore, the evaluation of antioxidant activities of extractsand fractions is considered an important step prior to the isolationof antioxidant phytochemicals they contain. Antioxidants arecompounds that can delay or inhibit the oxidation of lipidsor other molecules by inhibiting the initiation or propagation


25

45


of oxidizing chain reactions. When added to foods, antioxidantsminimize rancidity, delay the formation of toxic oxidation products,maintain nutritional quality and increase shelf life.52

Pasko et al.31 showed that pseudocereal seeds and sproutsshow relatively high antioxidant activity. Nsimba et al.57 evaluatedthe antioxidant activity of various extracts from quinoa (Japan)and from its relative Amaranthus, finding different values amongthe samples. In addition, Pasko et al.31 reported that quinoapresents higher antioxidant activity than amaranth using differentmethodologies: ferric reducing/antioxidant potential (FRAP),2,2′-azinobis(3-ethylbenzothiazoline 6-sulfonate) (ABTS) and 2,2-diphenyl-2-picryl-hydrazyl (DPPH). Antioxidant activity of quinoamight be of particular interest to medical researchers and needsfurther attention regarding its utilization as a natural potentantioxidant.10

ANTINUTRITIONAL FACTORSSeveral antinutritional substances have been found in quinoa,such as, saponins, phytic acid, tannins and protease inhibitors,34,58

which can have a negative effect on performance and survival ofmonogastric animals when it is used as the primary dietary energysource.58

Saponins were found to be the primary anti-quality factorsassociated with quinoa,58 but they have also some interesting bio-logical properties.59 Saponins are natural detergents made of gly-cosylated secondary metabolites, distributed throughout the plantkingdom; they include a diverse group of compounds character-ized by their structure containing a steroidal or triterpenoid agly-cone and one or more sugar chains.60 The quinoa is surrounded byan epicarp that contains saponins showing a characteristic bitter orastringent taste.61 Quinoa can be classified in accordance with thesaponin concentration and their content depends on the quinoavariety: ‘sweet’ (free from or containing <0.11% of free saponins)or ‘bitter’ (containing >0.11% of free saponins).21,27,62 – 64 Stuardoand San Martin63 reported that the content of saponins varies inquinoa between 0.1% and 5%. From the nutritional or pharmaco-logical point of view saponins could also have some value. They canincrease membrane permeability, thus enabling use for increasingfood intake at the intestinal level or even for drug assimilation.64,65

Other applications include raw materials for production ofhormones66 and immunological adjuvants,67 and there are alsoreported to be active ingredients in various natural health prod-ucts, such as herbal extracts.68 Stuardo and San Martın,63 Keukenset al.69 and Armah et al.70 reported antifungal activity of quinoasaponins due to its capacity to associate with steroids of fungalmembranes, causing damage to its integrity and pore formation,probably the basis of the novel molluscicide derived from thehusks of quinoa, discovered and developed by San Martın et al.71

Phytic acid is not only present in the outer layers of quinoa, asin the case of rye and wheat,47 but is also evenly distributed in theendosperm. Phytic acid binds minerals, thereby rendering themunavailable for metabolism.47,72 Ranges of 10.5–13.5 mg g−1 ofphytic acid for five different varieties of quinoa were reported byKozioł.27

Protease inhibitors, broadly distributed in nature, are proteinsthat form very stable complexes with proteolytic enzymes.73

The concentration of protease inhibitors in quinoa seeds is<50 p.p.m.47 Ahamed et al.47 and Improta and Kellems58 reportedthat quinoa contains small amounts of trypsin inhibitors which aremuch lower than those in commonly consumed grains and hencedo not pose any serious concern.

SUMMARY OF QUINOA FUNCTIONALPOTENTIAL FOR HUMAN DIETSome Leguminosae in combination with some cereals mightimprove proteic profiles of high-quality foods due to amino acidcompensation, a good strategy also used with quinoa food forchildren in the Andean region. Cerezal et al.74 designed a food for3- to 5-year-old children, with high amino acid content (35–40%of daily requirements). Nsimba et al.57 used quinoa and amaranthin products such as bread, pastas and baby foods. Lorenz andCoulter45 evaluated quinoa flour extrusion mixed with maize gritsto develop snacks with moderate acceptance. Moreover, thereis evidence concerning other physiologically active compoundspresent in quinoa seeds such as tannins17 and betaines.75 Tannins,which are polyphenolic compounds, form complexes with dietaryproteins and also with digestive enzymes.76 In addition to proteins,humans require minerals for their normal life processes, particularlyessential minerals, those necessary to support adequate growth,reproduction and health throughout the life cycle. Because theycannot be synthesized, minerals are necessarily obtained from thediet, and thus animals require a mineral intake for a long-termmaintenance of body mineral reserves.77 Minerals are involvedin many important functions in the body, e.g. cofactors ofhundreds of enzymatic reactions, bone mineralization, as well asprotection of cells and lipids in biological membranes (antioxidantproperties). Low intake or reduced bioavailability of minerals maylead to deficiencies, which causes serious impairment of bodyfunctions.78 Quinoa content is rich in minerals such as calcium,iron, zinc, magnesium and manganese, which give the grainshigh value for different target populations: for instance, adultsand children benefit from calcium for bones and from iron forblood functions.27,33 Antioxidant properties conferred by vitaminE and ω-3 fatty acids, plus the neuronal activity of tryptophanamino acid and vitamin B complex, can be powerful aids in brainfunction. Strong effects on protection of stressed neurons given byquinoa consumption in lab rats has recently been demonstrated,with evident effects on neuronal gene expression under stressingconditions, and also on improving spatial memory and promotionof low anxiety in the same animals.79 All these effects should beimportant in adults as well as in child populations. Besides, zinchelps the immunological system and magnesium is also importantduring the formation of neuromessengers and neuron modulators.Quinoa also improves some insulin-like forms which are active asgrowth hormones.80 The low glycemic index makes quinoa goodfor diabetic patients (low fructose and glucose), as mentioned byOshodi et al.81 Celiac and lactose-intolerant subjects should alsobe quinoa consumers because of its gluten-free nature and its richprotein levels, similar to milk casein quality.82

ISOFLAVONESFinally, a recent yet unpublished thesis83 showed that quinoa seedsfrom different origins, including long-distance regions of Chile,show different isoflavone concentrations, particularly daidzeinand genistein. These hormones are implicated in plant physiology(protection from pathogens, from UV light and nitrogen-limitedsoils) and could be recognized by alpha and beta receptorsof estrogens in humans. These endoplasmic reticulum-linkedreceptors are implicated as inhibitors of tyrosine kinase enzymes,and as antagonists of vessel contraction. They also reduce arterialresistance, benefit bone density and stimulate osteoprogerinsecretion by osteoblasts, in addition to its antioxidant properties.83


25

46


CONCLUSIONSThe outstanding physicochemical, nutritional and functionalproperties of quinoa have been reviewed. From ancient andhistorical data to current laboratory scientific evidence, quinoawas cultivated for its nutritional value, and after being abandonedin favor of old-world crops it is now starting to be rediscoveredby modern scientific approaches. Bitter seed coat saponins,while probably giving slower speed to quinoa recovery, mightnow be helpful for its ‘take-off’ among farmers for a broaderrange of consumers, as such saponins also have important agro-pharmacological and cosmetic industrial uses. From the point ofview of vegetarian consumers, quinoa in combination with othercereals might easily replace meat, with a great future in modern,conscient and more ecological food habits. Functional propertiesgiven by strongly active compounds like minerals, vitamins, fattyacids and antioxidants make of this small and noble grain astrong contribution to human nutrition, particularly for all cellprocesses requiring antioxidant protection of membranes, likeneuronal activity, with minerals and amino acid contents withpotential implications for aiding memory and lowering anxietyunder stressful conditions.

ACKNOWLEDGEMENTSThe authors gratefully acknowledge the Research Department atUniversidad de La Serena (DIULS), Chile, for providing financialsupport to the project (DIULS PI07302). In addition, we wishto thank Project Fondecyt 1060281, 1100638, and funding byANR-IMAS, TWAS-ICGEB and IRSES agencies.

REFERENCES1 USGS, United States Geological Survey. [Online]. Available: http://

earthquake.usgs.gov/earthquakes/eqinthenews/2010/us2010tfan/[2 March 2010].

2 Legrand C, [Online]. Available: <http://www.lemonde.fr/ameriques/article/2010/03/01/des-millions-de-refugies-au-chili-apres-un-fort-seisme 1312672 3222.html#xtor=EPR-32280229-[NL Titresdujour]-20100301-[zoneb]&ens id=1312017> [2 March2010].

3 Maughan PJ, Kolano BA, Maluszynska ND, Coles ND, Bonifacio A,Rojas J, et al, Molecular and cytological characterization ofribosomal RNA genes in Chenopodium quinoa and Chenopodiumberlandieri. Genome 49:825–839 (2006).

4 Christensen SA, Pratt DB, Nelson PT, Stevens MR, Jellen EN,Coleman CE, et al, Assessment of genetic diversity in the USDA andCIP-FAO international nursery collections of quinoa (Chenopodiumquinoa Willd.) using microsatellite markers. Plant Genet Resour5:82–95 (2007).

5 National Research Council (NRC), Lost Crops of the Incas: Little KnownPlants of the Andes with Promise for Worldwide Cultivation. NationalAcademy Press, Washington, DC, pp. 148–161 (1989).

6 Mujica A, Izquierdo J and Marathee JP, Origen y descripcion de laquinua, in Quinua (Chenopodium quinoa Willd.): Ancestral cultivoandino, alimento del presente y futuro, ed. by Mujica A, JacobsenS-E, Izquierdo J, Marathee J. FAO, UNA-Puno, CIP, Santiago, Chile,pp. 9–29 (2001).

7 Tagle MB and Planella MT, La Quinoa en la zona central de Chile:Superviviencia de una tradicion prehispana. Editorial IKU, Santiago,Chile (2002).

8 Tapia M, Cultivos andinos subexplotados y su aporte a la alimentacion(2nd edn). FAO, Oficina Regional para America Latina y el Caribe,Santiago, Chile (1997).

9 Fuentes FF, Martınez EA, Hinrichsen PV, Jellen EN and Maughan PJ,Assessment of genetic diversity patterns in Chilean quinoa(Chenopodium quinoa Willd.) germplasm using multiplexfluorescente microsatellite markers. Conserv Genet 10:369–377(2008).

10 Bhargava A, Shukla S and Ohri D, Chenopodium quinoa: an Indianperspective. Ind Crops Prod 23:73–87 (2006).

11 Jacobsen SE, Adaptation of quinoa (Chenopodium quinoa) to northernEuropean agriculture: studies on developmental pattern. Euphytica96:41–48 (1997).

12 Jacobsen SE, Developmental stability of quinoa under Europeanconditions. Ind Crops Prod 7:169–174 (1998).

13 Ormachea E and Quispe D, Evaluacion de parasitoides de la ‘polilla dela quinua’ Eurysacca melanocampta, en el Cusco, in XXXVConvencionNacional de Entomologıa, Arequipa, Peru, 11–14 November (1993).

14 Bertero HD, De La Vega AJ, Correa G, Jacobsen SE and Mujica A,Genotype and genotype-by-environment interaction effects forgrain yield and grain size of quinoa (Chenopodium quinoa Willd.)as revealed by pattern analysis of international multi-environmenttrials. Field Crops Res 89:299–318 (2004).

15 Cusack DF, Quinua: grain of the Incas. Ecologist 14:21–31 (1984).16 Valencia-Chamorro SA, Quinoa, in Encyclopedia of Food Science

and Nutrition, ed. by Caballero B. Academic Press, Amsterdam,pp. 4895–4902 (2003).

17 Ruales J and Nair BM, Saponins, phytic acid, tannins and proteaseinhibitors in quinoa (Chenopodium quinoa Willd.). J Sci Food Agric54:211–219 (1993).

18 Mujica A, Andean grains and legumes, in Neglected Crops: 1492 froma Different Perspective, ed. by Hernando Bermujo JE, Leon J. FAO,Rome, pp. 131–148 (1994).

19 Risic JC and Galwey NW, The Chenopodium grains of the Andes: Incacrops for modern agriculture, in Advances in Applied Microbiology,ed. by Coaker TH. Academic Press, London, pp. 145–217 (1984).

20 Martınez EA, Delatorre J and Von Baer I, Quınoa: las potencialidadesde un cultivo sub-utilizado en Chile. Tierra Adentro (INIA) 75:24–27(2007).

21 Martınez EA, Veas E, Jorquera C, San Martın R and Jara P, Re-introduction of Chenopodium quinoa Willd. into arid Chile:cultivation of two lowland races under extremely low irrigation.J Agron Crop Sci 195:1–10 (2009).

22 Wood TR, Tale of a food survivor: Quinoa. East West J 4:63–68 (1985).23 Mujica A, Jacobsen SE and Izquierdo J, Resistencia a factores adversos

de la quinua, in Quinua (Chenopodium quinoa Willd.): AncestralCultivo Andino, Alimento del Presente y Futuro, ed. by Mujica A,Jacobsen S-E, Izquierdo J, Marathee JP. FAO, UNA-Puno, CIP,Santiago, pp. 162–183 (2001).

24 Oelke EA, Putnam DH, Teynor TM and Oplinger ES, Alternative fieldcrops manual. University of Wisconsin Cooperative ExtensionService, University of Minnesota Extension Service, Centre forAlternative Plant and Animal Products (1992).

25 Thanapornpoonpong SN, Vearaslip S, Pawelzik E and Gorinstein S,Influence of various nitrogen applications on protein and amino acidprofiles of amaranth and quinoa. JAgricFoodChem 56:11464–11470(2008).

26 FAO, Food and Agriculture Organization, Ecocrop Database(2007). [Online]. Available: http://ecocrop.fao.org/ecocrop/srv/en/dataSheet?id=2509 [29 May 2009].

27 Kozioł M, Chemical composition and nutritional evaluation of Quinoa.(Chenopodium quinoa Willd). J Food Comp Anal 5:35–68 (1992).

28 Wright KH, Pike OA, Fairbanks DJ and Huber SC, Composition ofAtriplex hortensis, sweet and bitter Chenopodium quinoa seeds.Food Chem Toxicol 67:1383–1385 (2002).

29 De Bruin A, Investigation of the food value of quinoa and canihuaseed. J Food Sci 29:872–876 (1963).

30 Dini A, Rastrelli L, Saturnino P and Schettino O, A compositional studyof Chenopodium quinoa seeds. Nahrung 36:400–404 (1992).

31 Pasko P, Barton H, Zagrodzki P, Gorinstein S, Folta M and Zachwieja Z,Anthocyanins, total polyphenols and antioxidant activity inamaranth and quinoa seeds and sprouts during their growth.Food Chem 115:994–998 (2009).

32 Brady K, Ho C, Rosen R, Sang S and Karwe M, Effects of processingon the nutraceutical profile of quinoa. Food Chem 100:1209–1216(2007).

33 Repo-Carrasco R, Espinoza C and Jacobsen SE, Nutritional value anduse of the Andean crops quinoa (Chenopodium quinoa) and kaniwa(Chenopodium pallidicaule). Food Rev Int 19:179–189 (2003).

34 Gonzalez JA, Roldan A, Gallardo M, Escudero T and Prado FE,Quantitative determinations of chemical compounds withnutritional value from inca crops: Chenopodium quinoa (‘quinoa’).Plant Foods Hum Nutr 39:331–337 (1989).


25

47


35 Ridout CL, Price KP, Dupont MS, Parker ML and Fenwiclk GR, Quinoasaponins: analysis and preliminary investigation into the effects ofreduction by processing. J Sci Food Agric 54:165–176 (1991).

36 FAO/WHO/UNU, Food and Agriculture Organization of the UnitedStates/World Health Organization/United Nations University,Energy and protein requirements. Report of a joint FAO/WHO/UNUmeeting. World Health Organization, Geneva (1985).

37 Abugoch LE, Quinoa (Chenopodium quinoa Willd.): composition,chemistry, nutritional, and functional properties. Adv Food NutrRes 58:1–31 (2009).

38 Schlick G and Bubenheim DL, Quinoa: candidate crop for NASA’scontrolled ecological life support systems, in Progress in New Crops,ed. by Janick J. ASHS Press, Arlington, VA, pp. 632–640 (1996).

39 Ruales J and Nair BM, Properties of starch and dietary fibre in raw andprocessed quinoa (Chenopodium quinoa, Willd.) seeds. Plant FoodsHum Nutr 45:223–246 (1994).

40 Mahoney AW, Lopez JG and Hendricks DG, An evaluation of theprotein quality of quinoa. J Agric Food Chem 23:190–193 (1975).

41 Aluko RE and Monu E, Functional and bioactive properties of quinoaseed protein hydrolysates. J Food Sci 68:1254–1258 (2003).

42 Bhargava A, Rana TS, Shukla S and Ohri D, Seed proteinelectrophoresis of some cultivated and wild species ofChenopodium. Biol Plant 49:505–511 (2005).

43 Tharanathan RN, Starch: the polysaccharide of high abundance andusefulness. J Sci Ind Res 54:452–458 (1995).

44 Tharanathan RN and Mahadevamma S, Grain legumes: a boom tohuman nutrition. Trends Food Sci Technol 14:507–518 (2003).

45 Lorenz K and Coulter L, Quinoa flour in baked products. Plant FoodsHum Nutr 41:213–223 (1991).

46 Jian YQ and Kuhn M, Characterization of Amaranthus cruentus andChenopodium quinoa starch. Starch–Starke 51:116–120 (1999).

47 Ahamed NT, Singhal RS, Kulkarni PR and Mohinder P, A lesser-knowngrain, Chenopodium quinoa: review of the chemical composition ofits edible parts. Food Nutr Bull 19:61–70 (1998).

48 Ogunbengle HN, Nutritional evaluation and functional properties ofquinoa (Chenopodium quinoa) flour. Int J Food Sci Nutr 54:153–158(2003).

49 Dini I, Tenore GD and Dini A, Nutritional and antinutritionalcomposition of Kancolla seeds: an interesting and underexploitedandine food plant. Food Chem 92:125–132 (2005).

50 Konishi Y, Hirano S, Tsuboi H and Wada M, Distribution of mineralsin quinoa (Chenopodium quinoa Willd.) seeds. Biosci BiotechnolBiochem 68:231–234 (2004).

51 Valencia S, Svanberg U, Sandberg AS and Ruales J, Processing ofquinoa (Chenopodium quinoa Willd.): effects on in vitro ironavailability and phytate hydrolysis. Int J Food Sci Nutr 50:203–211(1999).

52 Sanders M, Estudio del secado industrial de la quinoa (Chenopodiumquinoa Willd.) cultivada en Chile: efecto de la temperatura sobre sucomposicion. Tesis de Pregrado. Departament of Food Engineering,Universidad de La Serena, Chile (2009).

53 Zeng X, Cheng K-W, Jiang Y, Lin Z-X, Shi J-J, Ou S-Y, et al, Inhibitionof acrylamide formation by vitamins in model reactions and friedpotato strips. Food Chem 116:34–39 (2009).

54 Insel P, Turner RE and Ross D, Discovering Nutrition. American DieteticAssociation, Jones & Bartlett, Boston, MA (2003).

55 Moyad MA, An introduction to dietary/supplemental omega-3 fattyacids for general health and prevention. Part I. Urol Oncol 23:28–35(2005).

56 Ng SC, Anderson A, Coker J and Ondrus M, Characterization of lipidoxidation products in quinoa (Chenopodium quinoa). Food Chem101:185–192 (2007).

57 Nsimba RY, Kikuzaki H and Konishi Y, Antioxidant activity of variousextracts fractions of Chenopodium quinoa and Amaranthus spp.seeds. Food Chem 106:760–766 (2008).

58 Improta F and Kellems RO, Comparison of raw, washed andpolished quinoa (Chenopodium quinoa Willd.) to wheat, sorghumor maize based diets on growth and survival of broilerchicks. [Online]. Livest Res Rural Dev 13 (2001). Available:http://www.cipav.org.co/lrrd/lrrd13/1/impr131.htm [29 May 2009].

59 Sparg SG, Light ME and Van Staden J, Biological activities anddistribution of plant saponins. J Ethnopharmacol 94:219–243(2004).

60 Guclu-Ustundag O and Mazza G, Saponins: properties, applicationsand processing. Crit Rev Food Sci Nutr 47:231–258 (2007).

61 Tarade KM, Singhal RS, Jayram RV and Pandit AB, Kinetics ofdegradation of saponins in soybean flour (Glycine max) duringfood processing. J Food Eng 76:440–445 (2006).

62 Soliz-Guerrero JB, Jasso de Rodrıguez D, Rodrıguez-Garcıa R, Angulo-Sanchez JL and Mendez-Padilla G, Quinoa saponins: concentrationand composition analysis, in Trends in New Crops and New Uses, ed.by Janick J and Whipkey A. ASHS Press, Alexandria, VA, pp. 110–114(2002).

63 Stuardo M and San Martın R, Antifungal properties of quinoa(Chenopodium quinoa Willd.) alkali treated saponins against Botrytiscinerea. Ind Crops Prod 27:296–302 (2008).

64 Gee JM, Price KR, Ridout CL, Wortley GM, Hurrel RF and Johnson IT,Saponins of quinoa (Chenopodium quinoa): effect of processing ontheir abundance in quinoa products and their biological effects onintestinal mucosal tissue. J Sci Food Agric 63:201–209 (1993).

65 Oakenfull D and Sidhu G, Could saponins be a useful treatment forhypercholesterolaemia? Eur J Clin Nutr 44:79–88 (1990).

66 Blunden G, Culling C and Jewers K, Steroidal sapogenins: a review ofactual and potential plant sources. Trop Sci 17:139–154 (1975).

67 Kensil CR, Mo AX and Truneh A, Current vaccine adjuvants: anoverview of a diverse class. Frontiers Biosci 9:2972–2988 (2004).

68 Balandrin MF, Commercial utilization of plant-derived saponins: anoverview of medicinal, pharmaceutical, and industrial applications,in Saponins Used in Traditional Medicine, ed. by Waller GR andYamasaki K. Plenum Press, New York, pp. 1–14 (1996).

69 Keukens AJ, De Vrije T, Van den Boom C, De Waard P, Plasman HH,Thiel F, et al, Molecular basis of glycoalkaloid induced membranedisruption. Biochem Biophys Acta 1240:216–228 (1995).

70 Armah CN, Mackie AR, Roy C, Price K, Osbourn AE, Bowyer P, et al,The membrane permeabilizing effect of avenacin A-1 involves thereorganization of bilayer cholesterol. Biophys J 76:281–290 (1999).

71 San Martın R, Ndjoko K and Hostettmann K, Novel molluscicide againstPomacea canaliculata based on quinoa (Chenopodium quinoa)saponins. Crop Prot 27:310–319 (2008).

72 Khattak AB, Zeb A, Bibi N, Khalil SA and Khattak MS, Influence ofgermination techniques on phytic acid and polyphenols contentof chickpea (Cicer arietinum L.) sprouts. Food Chem 104:1074–1079(2007).

73 Aguirre C, Valdez-Rodrıguez S, Mendoza-Hernandez G, Rojo-Domınguez A and Blanco-Labra A, A novel 8.7 kDA proteaseinhibitor from chan seeds (Hyptis suaveolens L.) inhibits proteasesfrom the larger grain borer Prostephanus truncatus (Coleoptera:Bostrichidae). Comp Biochem Physiol Part B 138:81–89 (2004).

74 Cerezal P, Carrasco A, Pinto K, Romero N and Arcos R, Suplementoalimenticio de alto contenido proteico para ninos de 2–5anos, desarrollo de la formulacion y aceptabilidad. Interciencia32:857–864 (2007).

75 Dini I, Tenore GC, Trimarco E and Dini A, Two novel betaine derivativesfrom Kancolla seeds (Chenopodiaceae). Food Chem 98:209–213(2006).

76 Singh U and Eggum BO, Factors, affecting the quality of pigeonpea(Cajanus cajan L.). Plant Foods Hum Nutr 34:273–283 (1984).

77 McDowell LR, Minerals in Animal and Human Nutrition (2nd edn).Elsevier, Amsterdam (2003).

78 Schlenker ED and Williams SR, Essentials of Nutrition and Diet Therapy.Elsevier, Amsterdam (2003).

79 Munoz-Llancao P, Martınez EA, Wyneken U and Dagnino-Subiabre A,Pseudo cereal quinoa improves memory and decreases anxiety ofstressed rats: behavioral and molecular approaches, in I IBRO/LARCCongress of Neurosciences of Latin America, Caribbean and IberianPeninsula, Buzios, Brazil, 1–4 September (2008).

80 Ruales J, de Grijalva Y, Lopez-Jaramillo P and Nair BM, The nutritionalquality of an infant food from quinoa and its effect on the plasmalevel of insulin-like growth factor-1 (IGF-1) in undernourishedchildren. Int J Food Sci Nutr 53:143–154 (2002).

81 Oshodi AA, Ogungbenle HN and Oladimeji MO, Chemicalcomposition, nutritionally valuable minerals and functionalproperties of beniseed (Sesamum radiatum), pearl millet(Pennisetum typhoides) and quinoa (Chenopodium quinoa) flours.Int J Food Sci Nutr 50:325–331 (1999).

82 Herencia LI, Alıa MJ, Gonzalez A and Urbano P, Cultivo de la quinoa(Chenopodium Quinoa Willd.) en la region Centro. [The culture ofquinoa in the central region]. Vida Rural VI:28–33 (1999).

83 Martınez AM, Contenidodedaidzeinay genisteinaenecotipos desemillasde quınoa (Chenopodium quinoa Willd.). Thesis, Chemistry andPharmacy, Universidad de Valparaıso, Chile (2008).


25

48

Research ArticleReceived: 26 October 2009 Revised: 1 July 2010 Accepted: 2 July 2010 Published online in Wiley Online Library: 17 August 2010


Cultivar choice provides options for localproduction of organic and conventionallyproduced tomatoes with higher qualityand antioxidant contentHeather Troxell Aldrich,a Karen Salandanan,a Patricia Kendall,b

Marisa Bunning,b Frank Stonaker,a Oktay Kulena and Cecil Stushnoffa∗

Abstract

BACKGROUND: Tomatoes (Solanum lycopersicum L.) are widely consumed and well known for their health benefits, many ofwhich have been associated with the high levels of antioxidants present in tomatoes. With a growing interest in local andorganic foods, it would be helpful to determine whether farmers could naturally improve the quality and antioxidant contentof tomatoes for sale in local markets. This study evaluated antioxidant properties, quality attributes, and yield for 10 tomatocultivars grown for 2 years using certified organic and conventional practices.

RESULTS: Cultivar and year effects impacted (P < 0.05) all tests conducted, while growing method influenced (P < 0.05)yield, soluble solids content, ascorbic acid, and antioxidant radical scavenging capacity. Even when accounting for year-to-yearvariability, cultivars in the highest groups had 1.35- to 1.67-fold higher antioxidant levels than cultivars in the lowest groups.‘New Girl’, ‘Jet Star’, ‘Fantastic’, and ‘First Lady’ were always in the highest groups, while ‘Roma’ and ‘Early Girl’ consistentlyhad the lowest antioxidant content.

CONCLUSION: Compared to production practices and environmental effects of years that are generally beyond the control ofsmall-scale producers, choice of cultivar provides the simplest and most effective means of increasing antioxidant properties.Knowledge of tomato cultivars with naturally higher antioxidant levels could assist smaller-scale producers to grow fruitthat may provide a competitive advantage and the opportunity to capitalize on the increasing popularity of locally grown,high-quality fresh produce.c© 2010 Society of Chemical Industry

Keywords: antioxidants; total phenolics; ascorbic acid; cultivar; organic production

INTRODUCTIONTomatoes (Solanum lycopersicum L.) are widely consumed, rankingsecond to potatoes among vegetable and melon per capita usein the USA.1 Tomatoes are good sources of vitamin C, folate, andpotassium, as well as many phytochemicals.2 – 5

Awareness for the important role fruits and vegetables have ina healthy diet is increasing,6,7 as is interest in organic foods.

Organic foods have become one of the fastest-growing foodcategories, with sales increasing nearly 20% each year since 1990.8

Consumer studies have shown that organic produce is perceivedto be safer, more nutritious, and better tasting than conventionallygrown produce.9 – 12 However, research comparing such attributeshas produced inconsistent or inconclusive results, most likelydue to uncontrolled variables such as differences in growingconditions and cultivars studied.13,14 Additional well-controlledstudies are needed to better understand the impact of organicand conventional growing methods on produce quality andnutritional attributes.15

Farmers’ markets are becoming a popular place to purchaseproduce in the USA. According to the United States Department

of Agriculture (USDA), the number of US farmers’ markets morethan doubled from 1755 in 1994 to 4685 in 2008.16 Also, based onthe results of a national survey conducted in 2006, three out offour respondents had shopped at a farmers’ market within the lastyear.17 Consumers indicated greater willingness to pay a premiumfor produce that had a higher nutritional value (i.e. melons with25% more vitamin C), was grown locally, and/or was grown usingorganic production methods. Support for local farmers was foundto be more important than organic production and documentednutritional attributes also seemed to be favored over productioncertifications.18

∗ Correspondence to: Cecil Stushnoff, Department of Horticulture and LandscapeArchitecture, Colorado State University, Fort Collins, CO 80523-1173, USA.E-mail: [email protected]

a Department of Horticulture and Landscape Architecture, Colorado StateUniversity, Fort Collins, CO 80523, USA

b Department of Food Science and Human Nutrition, Colorado State University,Fort Collins, CO 80523, USA


25

49

Cultivar choice provides nutritious tomato options www.soci.org

This increased popularity and demand for high-quality, freshproduce, along with the growing interest in locally produced foodand organic production,17,18 provides small-scale growers witha unique opportunity. The opportunity, however, can best berealized if local small-scale growers can differentiate their producefrom that available from other sources. The objectives of thisstudy were to compare the antioxidant levels, quality attributes,and yield among 10 tomato cultivars grown using organic andconventional production methods, and to determine whetherproduction method and/or cultivar choice would assist small- andmedium-scale farmers to market nutritionally superior tomatoes.

MATERIALS AND METHODSPlant materialTomatoes were grown at Colorado State University’s (CSU)Horticulture Field Research Center (HFRC) in Fort Collins, CO,during the summers of 2005 and 2006. Ten cultivars of tomatoeswere grown simultaneously on organic and conventional plots.The cultivars grown included ‘Big Beef’, ‘Early Girl’, ‘Celebrity’,‘Fantastic’, ‘First Lady’, ‘Husky Red’, ‘Jet Star’, ‘Red Sun’, ‘NewGirl’, and ‘Roma’. Seeds were obtained from Johnny’s SelectedSeeds, Winslow, ME, USA, and Harris Seeds, Rochester, NY, USA. Allcultivars are beefsteak type except ‘Roma’ (a plum type).

Plants were started in the CSU Plant Environmental ResearchCenter greenhouses in Landmark T1204 cell packs (LandmarkPlastic Corp., Akron, OH, USA) using Sunshine Organic Basicplanting media (Sun Gro Horticulture, Bellvue, WA, USA) with20% vermicompost (local source). After 6 weeks, the tomatoeswere transplanted to the field, spaced evenly in black plasticmulched beds (plants approximately 45 cm apart within a row,and rows 150 cm apart).

Soil at the HFRC is classified as Nunn clay with a pH of 7.8.The organic plots have been USDA certified organic since 2001.The tomatoes were planted in a split plot design with productionsystem as the whole plot factor. Tomato cultivars were randomizedas sub-plots within each of three replicated plots (15 plants perplot) in 2005 and within four replicated plots (eight plants per plot)in 2006. The two production systems have identical soil texturesand were located approximately 50 m from each other, separatedby an appropriate buffer zone with vegetation.19

Prior to planting, soil tests were conducted on the organic andconventional plots to determine pH, electroconductivity (EC), lime,organic matter, N, P, K, Zn, Fe, Mn, and Cu levels. Soil fertility ofthe two plots was equilibrated at the beginning of the growingseason using either organic or conventional fertilizers. Based onthe soil tests, 22 679 kg ha−1 of ‘Evergreen’ poultry compost (A1Organics, Eaton, CO, USA) was applied pre-planting to the organicplots. The compost was applied with a spreader (Millcreek Mfg,Leola, PA, USA) and disked into the soil immediately followingthe application. To match the amount of nutrients in the organicblock, 389 kg ha−1 of urea (45N-0P-0K) and 882 kg ha−1 of triplesuperphosphate (0N-20.1P-0K) were applied to the conventionalblock using a broadcast spreader.

Crops were irrigated using drip irrigation with low-salt-content,high-quality municipal water. Soil moisture levels were determinedusing ‘Watermark’ granular matrix sensors (Irrometer Co., Riverside,CA, USA). Soil moisture was monitored to ensure the tomatoeswere watered adequately in order to prevent water stress by notallowing water tension to drop below 100 kPa throughout thegrowing season.

Pest management was needed in 2005, as there was acombination of pressure from potato psyllid (Paratrioza cockerelli)and beet leafhopper (Circulifer tenellus). Potato psyllids werecontrolled in the organic plots using an approved botanicallyderived pyrethrum (’Pyganic,’ MGK Co., Golden Valley, MN, USA)insecticide. ‘Provado’ (Imidicloprid, Bayer CropScience, ResearchTriangle Park, NC, USA) was applied to control psyllids andleafhoppers in the conventional plots. There was no psyllidinfestation during the 2006 growing season.

Temperature and solar radiation data collectionData on temperature and solar radiation during the 2005 and 2006growing seasons were obtained from a Northern Colorado WaterConservancy District weather station, located within 100 m of theresearch plots. To examine possible effects of temperature, dailygrowing degree days (GDD) was computed by subtracting thebase temperature (10 ◦C) from the average of the daily maximumand minimum temperatures (daily GDD = [(Tmax + Tmin)/2] − basetemperature, where Tmax and Tmin are maximum and minimumdaily air temperatures). Each daily GDD was summed over thegrowing season and 30 days prior to harvest. Solar radiation datawere recorded with an Eppley pyranometer (Newport, RI, USA)and expressed as langleys.

Post-harvest analysesOnce the tomatoes reached the vine ripe stage, indicated by fullred color and the onset of softening, they were harvested manuallyearly each morning. Note that while the latest harvest date in 2005was 20 days longer than in 2006 due to the difference in heat unitaccumulation, fruit was harvested when it was judged to be atessentially the same stage of ripeness for each cultivar.

All tomatoes were harvested and weighed from each replicate,and yield was expressed in kg per plant. On one day during theheight of the growing season, three uniformly sized and coloredtomato fruits per replicate plot were randomly selected from thebulked harvest. The tomato samples (n = 180 in 2005 and 240 in2006) were transported at ambient temperature to the laboratorywithin 30 min for evaluation of soluble solids content and pH, aswell as preparation for antioxidant analyses.

Soluble solids content and pHPercent soluble solids of each tomato sample was measured usinga handheld Reichert temperature-compensated refractometer(Reichert Analytical Instruments, Depew, NY, USA). The pH ofthe tomatoes was tested using a Beckman pH meter (Fullerton,CA, USA).

Antioxidant analysesSample preparationTomatoes were washed well to remove any debris on their outersurface, and then cut in half vertically. Thin radial slices were cutfrom the tomato halves. Thin slices from each tomato (35–40 g)were freeze-dried using a Genesis freeze drier (Virtis, Inc., Gardiner,NY, USA). Lyophilized samples were weighed to determine percentdry matter content and ground in preparation for extraction. Thedried samples were ground into a fine powder using a mortar andpestle and sieved with a No. 20 Tyler sieve (WS Tyler Inc., Mentor,OH, USA).

Samples were prepared by extracting 200 mg lyophilizedpowder in 5 mL of 80% acetone (Fisher Scientific, Fair Lawn, NJ,


25

50

www.soci.org HT Aldrich et al.

USA) with vortexing and 15 min rotary mixing in 15 mL centrifugetubes in the dark at 4 ◦C. The samples were centrifuged (4 ◦C;1771 g relative centrifugal force) for 15 min. One-milliliter aliquotsof supernatant were vacufuged at 45 ◦C to dryness (approximately2–3 h). Samples were stored at −20 ◦C until analytical tests werecompleted.

Total phenolic contentTotal phenolic content was measured using a microplate-basedFolin–Ciocalteu assay adapted from Singleton and Rossi,20 Spanosand Wrolstad,21 and Rivera et al.22 Briefly, vacufuged extractionswere reconstituted with 1.0 mL of 80% acetone, diluted 1/9with nanopure water and reacted with 150 µL of 0.2 mol L−1

Folin–Ciocalteu reagent (Sigma-Aldrich, Inc., St Louis, MO, USA).After 5 min at room temperature, microplate samples were reactedwith 115 µL of 7.5% (w/v) Na2CO3 (Fisher Scientific), incubated at45 ◦C for 30 min and cooled to room temperature for 1 h, beforereading absorbance at 765 nm in a Spectra Max Plus (MolecularDevices, Sunnyvale, CA, USA) spectrophotometer using SoftmaxPro software (Molecular Devices). Total phenolic content wascalculated by regression from a gallic acid (Sigma Chemical Co.,St Louis, MO, USA) standard curve and expressed as milligrams ofgallic acid equivalents per kilogram of tomato fresh weight (mgGAE kg−1 FW).

Ascorbic acid contentAscorbic acid (vitamin C) content was determined using high-performance liquid chromatography (HPLC) as described by Riveraet al.22 and modified from Dale et al.23 Freeze-dried samples wereextracted for 15 min at 4 ◦C in the dark with a 5% w/v aqueoussolution of metaphosphoric acid containing 1% w/v dithiothreitol(DTT) (Promega Corp., Madison, WI, USA). The samples werecentrifuged for 5 min at 1771 × g at 4 ◦C and the supernatantwas filtered through a 0.45 µm nylon syringe filter. The extractionprocess was repeated and the supernatant from both extractionswas placed in an amber HPLC vial. Ascorbic acid standards wereprepared with 100 mg DTT (Promega Corp.), 10 mg ascorbic acid(Sigma-Aldrich), and 10 mL of 100% methanol before diluting tofive concentrations for the standard curve. All analyses were runin duplicate and were analyzed by HPLC (Hewlett Packard Model1050 Series, Palo Alto, CA, USA) using Chem Station for LC Rev A09.01 software (Agilent Technologies, Palo Alto, CA, USA). Sampleswere injected into an Inertsil C4 column (Agilent Technologies)run with a phosphoric acid/methanol gradient and absorbanceread at 254 nm.

ABTS Trolox equivalent antioxidant capacityThe 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) di-ammonium salt (ABTS) assay was used to estimate antioxidanthydroxyl radical scavenging capacity. The protocol used was basedon the microplate method described by Rivera et al.,22 as modifiedfrom Miller and Rice-Evans.24 The ABTS solution was prepared bymixing 40 mg ABTS (Calbiochem, EMD Biosciences, La Jolla, CA,USA), 15 mL distilled water, and 2.0 ± 0.5 g MnO2 (Sigma-Aldrich).After 20 min, the MnO2 was removed using double filtration, firstwith a vacuum filtration and second with a 0.2 µm syringe filter.The absorbance value of the ABTS solution was adjusted to 0.70 AUat 734 nm in the Spectra Max Plus spectrophotometer using Soft-max Pro software with 5.0 mmol L−1 phosphate buffer solution.The ABTS solution was held at 30 ◦C and used within 4 h.

Vacufuged samples were reconstituted with 1 mL of 80%acetone (Fisher Scientific). Twenty-five microliters of sample werereacted with 250 µL of the ABTS solution, and the absorbance valuewas read after exactly 60 s at 30 ◦C in a temperature-controlledmicroplate reader. ABTS antioxidant capacity was reported asTrolox equivalent antioxidant capacity (TEAC) per gram of freshsample (TEAC g−1 FW) and was calculated by regression of a Trolox(Calbiochem) standard curve. Analyses were run in triplicate atthree dilutions for a total of nine assays per sample.

DPPH Trolox equivalent antioxidant capacityThe 2,2-diphenyl-1-picrylhydrazyl (DPPH) assay was also used toestimate antioxidant capacity according to the method of Luand Foo25 with some modifications. Briefly, vacufuged sampleswere reconstituted with 1.0 mL of 5.0 mmol L−1 phosphatebuffer solution and reacted with 0.1 mmol L−1 DPPH solutionin 100% methanol. Absorbance was read in the Spectra MaxPlus spectrophotometer using Softmax Pro software at 515 nmand adjusted to 0.95 AU. Fifteen microliters of the reconstitutedsamples were mixed with 285 µL of the DPPH solution, held forexactly 3 min at 25 ◦C, and read at 515 nm. The results weredetermined by regression from a Trolox (Calbiochem) standardcurve and expressed as µmol TEAC kg−1 FW.

Data and statistical analysisResults were analyzed using SAS statistical software (Version 9.1,Cary, NC, USA). A factorial analysis of variance was performed(SAS Proc Mixed) with differences between means assessed usinga significance of P < 0.05 with the Tukey–Kramer adjustmentfor multiple comparisons. Fixed effects included cultivar, growingmethod, and year; replication was included as a random effect.Correlation analysis was done using Pearson distribution (SAS ProcCorr).

RESULTSTemperature and solar radiationFrom field planting to harvest, GDD was higher in 2006 thanin 2005; thus tomatoes grown in 2006 were exposed to highercumulative heat units (Fig. 1(A)). During the 2006 growing season,there were also more days with maximum temperatures above30 ◦C (Fig. 1(B)). Solar radiation received by tomato plants fromplanting to harvest was nearly the same for both 2005 and 2006(Fig. 1(C)).

YieldYear, growing method, and cultivar all impacted (P < 0.0001)tomato yields (Table 1). Overall, 2006 yields per plant were higherthan 2005 yields (average of 4.17 kg per plant versus 2.99 kg perplant, respectively, data not shown); and conventional yields werehigher than organic yields (average of 4.03 kg per plant versus3.14 kg per plant, respectively; Table 2). While average cultivaryields varied by growing method, ‘Early Girl’ consistently producedthe highest yields and ‘Roma’ the lowest yields.

Percent dry matterThe percent dry matter (% DM) was significantly affected by year(Table 1; average of 5.3% in 2005 versus 7.4% in 2006), whilegrowing method did not impact % DM, with both methodsaveraging 6.3%. ‘Early Girl’ had the lowest % DM (P < 0.05), whilecultivars ‘First Lady,’ ‘New Girl,’ and ‘Jet Star’ had the highest(P < 0.05) over both years (Table 2).


25

51


0

200

400

600

800

1,000

1,200A

B

C

1 11 21 31 41 51 61 71 81 91 101

1 11 21 31 41 51 61 71 81 91 101

Days After Transplanting

Hea

t A

ccu

mu

lati

on

2005 2006

0

5

10

15

20

25

30

35

40


1 2111 31 41 51 61 71 81 91 101 111


Max

imu

m t

emp

erat

ure

s(d

egre

es C

)

2005 2006

0

100

200

300

400

500

600

700

800

So

lar

rad

iati

on

(L

ang

leys

)

2005 2006

Figure 1. (A) Heat accumulation (in growing degree-days or GDD) fromplanting to harvest, GDD = [(Tmax + Tmin)/2] −10 ◦C, where Tmax and Tminare maximum and minimum daily air temperatures; 10 ◦C is the basetemperature for warm season crops. (B) Number of days with temperaturegreater than 30 ◦C as indicator of heat stress. (C) Daily net solar radiation(in langleys) from planting to harvest.

pH and soluble solids contentCooler conditions in 2005 significantly impacted pH, with 2005tomatoes being more acidic than in 2006 (overall average of3.84 versus 4.14, respectively), while growing method did notsignificantly affect pH (Tables 1 and 2). The pH also varied bycultivar, with ‘Celebrity’ and ‘Early Girl’ ranking among the lowestin pH and ‘Jet Star’ the highest across growing method and year.Cultivar pH tended to rank in a similar order, regardless of growingmethod and year (Table 2).

Soluble solids content (SSC) was impacted by year (P < 0.0001;Table 1), with 2006 tomatoes having higher SSC than in 2005(average of 4.9 versus 4.3 ◦Brix, respectively). Overall, conventionaltomatoes had a slight, though significantly higher SSC thanorganic (means 4.6 versus 4.4 ◦Brix, respectively). SSC also varied

(P < 0.001) by cultivar, with ‘Early Girl’ having the lowest SSC(Table 2).

Antioxidant analysesOverall, year had a significant impact on all antioxidant tests(Table 1). Total phenolic content was higher in 2005, the cooleryear, than in 2006 (average of 0.78 versus 0.74 g GAE kg−1 FW,respectively), while the 2006 fruit had higher ABTS values (averageof 1533 versus 1400 µmol TEAC kg−1 FW), higher DPPH (averageof 2573 versus 1705 µmol TEAC kg−1 FW), and higher ascorbic acid(average of 0.23 versus 0.15 g kg−1 FW; data not shown).

Organically grown tomatoes had higher overall (P < 0.0001)ABTS (average of 1546 versus 1387 µmol TEAC kg−1 FW), DPPH(average of 2194 versus 2084 µmol TEAC kg−1 FW), and ascorbicacid (average of 0.198 versus 0.185 g kg−1 FW) compared to theconventionally grown fruit (Table 3). Cultivar differences werevery significant (P < 0.0001) for all antioxidant analyses (Table 1),though they varied by test (Table 3).

To compare cultivars by growing method, we placed cultivarsinto lowest, midrange, and highest groups, with values for cultivarsin lowest and highest groups significantly different (P < 0.05)based on Tukey–Kramer mean separation tests, using the methoddeveloped by Salandanan et al.26 This was done for both organicand conventional results averaged for the 2 years, since yeardifferences are impossible to control (Table 4). Based on ratioscomparing highest and lowest cultivar groups, cultivars falling inthe highest category were found to have 1.35- to 1.67-fold highervalues for all the antioxidant tests compared to cultivars in thelowest groups. The spread between lowest and highest groupswas similar for organic and conventionally grown tomatoes. Ratiosused to compare organic versus conventional results for the lowest,midrange, and highest cultivar groups tended to hover around1.00, while organic results were slightly higher for all but twogroups (Table 4).

A simple additive antioxidant index was used to comparecultivars over both years regardless of growing method. This indexwas calculated by determining the average of the total phenoliccontent, ascorbic acid content, and antioxidant capacity (using anaverage of the TEAC values obtained from the ABTS and DPPHassays; Fig. 2). The antioxidant index rank of the cultivars variedslightly between the 2 years; however, despite the environmentaleffects of each growing season, the same cultivars appear in the topand bottom tiers. ‘Roma’ and ‘Early Girl’ consistently had the lowestantioxidant index scores, while ‘New Girl’, ‘Jet Star’, ‘Fantastic’, and‘First Lady’ consistently had the highest index scores (Fig. 2 andTable 3).

Correlation analysisWhen all 10 cultivars were considered over both years andproduction methods, fruit with the highest % DM also had thehighest pH and SSC, as well as TEAC radical scavenging capacityand ascorbic acid content (Table 5). Total phenolic levels were notstrongly related to fruit quality parameters nor to ascorbic acid orTEAC values. While yield was significantly correlated with a fewparameters, none, except possibly DPPH, had high r values andwere therefore considered not to contribute substantially to thesecorrelations (Table 5).

DISCUSSIONIn this study, year-to-year variability was significant, with tomatoesgrown in the warmest year, 2006, having higher (P < 0.05) ascorbic


25

52


Table 1. Analysis of variance for the main effects year (Y), cultivar (C), growing method (M), and their interactions on yield, dry matter (% DM), pH,soluble solids content (SSC), total phenolic content (TPC), ascorbic acid (AA), and antioxidant activity (ABTS and DPPH)

Yield % DM pH SSC TPC AA ABTS DPPH

Year <0.0001 <0.0001 <0.0001 <0.0001 0.0034 <0.0001 0.0008 <0.0001

Cultivar <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001

Y × C 0.0437 0.3034 0.0210 0.0079 <0.0001 0.0356 <0.0001 <0.0001

Method <0.0001 0.7891 0.1126 0.0034 0.4944 0.0001 <0.0001 <0.0001

Y × M 0.8109 <0.0001 0.0022 0.0473 <0.0001 0.5746 0.4440 0.0027

C × M 0.5094 0.1612 0.1269 0.0020 <0.0001 0.0083 <0.0001 <0.0001

Y × C × M 0.3712 0.4037 0.2641 0.0907 <0.0001 0.0119 <0.0001 <0.0001

Table 2. Yield, % DM, pH, and soluble solids content resultsa for 10 tomato cultivars grown organically and conventionally in 2005 and 2006

Yield(kg per plant) % DM pH

Soluble solidscontent (◦Brix)

Cultivar Organic Conv. Organic Conv. Organic Conv. Organic Conv.

Big Beef 2.74bc 3.32bc 6.5bc 6.4bcde 3.97abc 4.02abc 4.62c 4.96cde

Celebrity 2.89bc 3.89abc 6.2bc 5.6ab 3.83a 3.94ab 4.33bc 4.01ab

Early Girl 4.51a 5.21a 5.1a 5.2a 3.85ab 3.89a 3.52a 3.46a

Fantastic 2.96abc 4.44ab 6.5bc 6.6cde 4.01cd 3.97ab 4.60c 5.22de

First Lady 3.66ab 4.33ab 7.1c 7.2e 3.91ab 3.95ab 4.82c 5.22de

Husky Red 3.40ab 4.18ab 6.3bc 6.3bcde 4.02cde 4.03abc 4.50bc 4.28bc

Jet Star 2.66bc 4.20ab 6.8c 6.8de 4.18e 4.14c 4.66c 4.58bcde

New Girl 3.57ab 3.74abc 6.7c 7.2e 3.95abc 3.97abc 4.39bc 5.31e

Red Sun 3.35ab 4.59ab 6.2bc 5.9abc 3.98bcd 4.04bc 4.26abc 4.48bcd

Roma 1.62c 2.39c 5.7ab 6.0abcd 4.13de 4.08bc 3.79ab 4.03ab

Mean 3.14 4.03 6.3 6.3 3.98 4.00 4.35 4.56

SEb ±0.32 ±0.18 ±0.03 ±0.15

a Results represent the mean (±SE) of seven replications (three replications in 2005 and four replications in 2006).b Standard error of means listed in column.Cultivars for the same year and growing method with different letters are significantly different (P < 0.05).

Table 3. Total phenolic content (TPC), ascorbic acid, and antioxidant activity (ABTS and DPPH) resultsa for 10 tomato cultivars grown organicallyand conventionally in 2005 and 2006

TPC(g GAE kg−1 FW)

Ascorbic Acid(g kg−1 FW)

ABTS(µmol TEAC kg−1 FW)

DPPH(µmol TEAC kg−1 FW)

Cultivar Organic Conv. Organic Conv. Organic Conv. Organic Conv.

Big Beef 0.78abc 0.66ab 0.17a 0.18abc 1318a 1474ab 1859ab 1939ab

Celebrity 0.80abc 0.70ab 0.20abc 0.21bc 1641abcA 1106aB 1874ab 1877a

Early Girl 0.69ab 0.56a 0.17a 0.17ab 1222a 1278ab 1956abc 1751a

Fantastic 0.87c 0.94d 0.22bc 0.19abc 1967cd 1568b 2379d 2436c

First Lady 0.77abc 0.67ab 0.22bc 0.20bc 2117dA 1534bB 2711fA 2292cB

Husky Red 0.64a 0.75bc 0.17a 0.18abc 1241a 1284ab 2083bc 2217bc

Jet Star 0.89c 0.90cd 0.22bcA 0.18abcB 1501ab 1527b 2585defA 2219bcB

New Girl 0.81bc 0.93d 0.23c 0.21c 1775bcA 1340abB 2605ef 2466c

Red Sun 0.69ab 0.67ab 0.19ab 0.18abc 1405ab 1403ab 2166cd 1958ab

Roma 0.68ab 0.73b 0.19ab 0.15a 1274a 1354ab 1722a 1688a

Mean 0.76 0.74 0.198 0.185 1546 1387 2194 2084

SEb ±0.032 ±0.007 ±91 ±55

a Results represent the mean (±SE) of seven replications for TPC and ABTS•+ (three replications in 2005 and four replications in 2006) and sixreplications for DPPH+ and ascorbic acid (three replications both years).b Standard error of means listed in column.Cultivars for the same year and growing method with different lower-case letters are significantly different (P < 0.05).Growing methods for the same year and cultivar with different upper-case letters are significantly different (P < 0.05).


25

53


Table 4. Comparison of means for total phenolics, antioxidantactivity (ABTS and DPPH), and ascorbic acid for high, midrange, andlow cultivar groups in organic and conventional production systemsaveraged over 2 years

Group Organic z Conventional z

Ratio(organic/

conventional)

Total phenolic content (g GAE kg−1 FW)

Lowest 0.64 1 0.56 1 1.14

Midrange 0.75 7 0.73 7 1.02

Highest 0.88 2 0.94 2 0.94

Ratio (highest/lowest) 1.38 1.67

Ascorbic acid (g kg−1 FW)

Lowest 0.17 3 0.15 1 1.13

Midrange 0.21 6 0.19 8 1.11

Highest 0.23 1 0.21 1 1.10

Ratio (Highest/lowest) 1.35 1.38

ABTS antioxidant activity (µmol TEAC kg−1 FW)

Lowest 1264 4 1106 1 1.14

Midrange 1658 5 1356 6 1.22

Highest 2117 1 1544 3 1.37


DPPH antioxidant activity (µmol TEAC kg−1 FW)

Lowest 1722 1 1772 3 0.97

Midrange 2188 8 2083 4 1.05

Highest 2711 1 2398 3 1.13


z, number of cultivars in lowest, midrange, and highest groups withvalues for cultivars in lowest and highest groups significantly different(P < 0.05) based on Tukey–Kramer mean separation tests.

acid, antioxidant activity (ABTS and DPPH), soluble solids content,pH, yield per plant and percent dry matter, while tomatoes grownin 2005 had higher (P < 0.05) total phenolic content. Higher lightintensity during production has been associated with higher ascor-bic acid content in fresh produce,27 which may explain why theascorbic acid levels were higher in 2006 than in 2005. Also, environ-mental temperatures during tomato growth have been shown toaffect antioxidant levels, though mechanisms are not well under-stood and optimal temperatures may differ for different groups ofphytochemicals.28,29 Although precipitation amounts were quitedifferent in the 3 months preceding harvest in 2005 (10.5 cm)and 2006 (3.5 cm), soil moisture was monitored and plots were

0

20

40

60

80

100

120

140

New G

irl

Jet S

tar

Fanta

stic

First L

ady

Red S

un

Husky

Red

Celebr

ity

Big Bee

f

Roma

Early

Girl

Cultivar

An

tio

xid

ant

Ind

ex

2005 2006

Figure 2. Antioxidant Index (±SE) of 10 tomato cultivars for 2005 and2006 growing seasons (n = 6 for each bar, three replications of twogrowing methods). Antioxidant index was calculated as [(� vit. C + TPC+ antioxidant capacity)/3], with antioxidant capacity being an average ofthe ABTS and DPPH antioxidant capacity assay results. Antioxidant indiceswere calculated per 100 g to allocate equal weight to each factor.

irrigated to prevent water stress. Clearly, environmental factorssuch as temperature, precipitation, and light intensity have im-portant implications in the development of quality and nutritionalcharacteristics of crops, but they are difficult to control and moreresearch is needed to better understand their specific impacts.30

Some evidence indicates there may be a trade-off betweenyield and nutritional quality of produce.31 – 34 However, suchstudies have primarily focused on comparing historical variationof fruits and vegetables, which entail averaging many foodsover time and include a great deal of confounding variablessuch as production methods, sampling procedures, and testmethodologies.33 As a result, Davis recommended the use ofside-by-side comparisons as a more accurate, controlled approachin which specific foods and nutrients can be evaluated.33 Few suchside-by-side comparisons have been conducted, though studieslooking at potato,35 wheat,36,37 and broccoli38 have found thatcultivars with higher yields (or larger heads for broccoli) oftenexperience low to moderate declines in many nutrients comparedto lower-yielding (or smaller-head) cultivars.

Our side-by-side comparative study does not support theargument that high-yielding cultivars compensate by reducingnutritional quality. Low Pearson r values and non-significance ofsome correlations suggest only a small proportion of the variation

Table 5. Correlation analysis (Pearson r) between percent dry matter (% DM), pH, soluble solids content (SSC), total phenolics content (TPC),antioxidant activity (ABTS and DPPH tests), ascorbic acid (AA), and yield. Correlation results include means from 10 cultivars, 2 years, two growingmethods, and three replications (n = 120)

% DM pH SSC TPC ABTS DPPH AA

% DM

pH 0.65∗∗∗

SSC 0.79∗∗∗ 0.46∗∗∗

TPC 0.13 n.s. −0.05 n.s. 0.25∗∗

ABTS•+ 0.40∗∗∗ 0.16 n.s. 0.31∗∗ 0.33∗∗∗

DPPH+ 0.86∗∗∗ 0.61∗∗∗ 0.69∗∗∗ 0.23∗ 0.50∗∗∗

AA 0.81∗∗∗ 0.66∗∗∗ 0.57∗∗∗ 0.10 n.s. 0.42∗∗∗ 0.83∗∗∗

Yield 0.29∗∗ 0.23∗∗ 0.18∗ −0.06 n.s. 0.09 n.s. 0.48∗∗∗ 0.35∗∗∗

n.s., not significant; asterisks indicate significance at ∗ P < 0.05; ∗∗ P < 0.01; ∗∗∗ P < 0.0001.


25

54


in quality and antioxidant parameters may be directly related toyield on a per plant basis. We chose to normalize yield on a perplant basis to avoid confounding effects that may occur fromdensity and planting configuration.

Organic versus conventional growing method differences havebeen a topic of much research and debate. A literature reviewof nutritional differences by Magkos and colleagues39 found aslight trend toward higher ascorbic acid content in organic leafyvegetables and potatoes and a trend toward slightly lower (buthigher-quality) protein levels in some organic crops compared toconventional counterparts. In another review article, organicallygrown produce was found to have significantly more vitaminC, iron, magnesium, and phosphorus, as well as significantlylower nitrate levels compared to conventionally grown produce,40

while others have indicated organic produce may be higher inantioxidants than conventional.41,42 However, these authors andothers acknowledge that most organic and conventional researchstudies did not control cultivars, maturity at harvest, soil, and otherenvironmental conditions that may impact results as much as, ifnot more than, growing method.13 – 15,26

Recently, more controlled studies have been conducted com-paring various attributes of tomatoes grown on paired organic andconventional farms. In two studies done in California, researchersfound the soluble solids content to be higher in the organic toma-toes than in the conventionally grown tomatoes.43,44 However,another project with a similar design conducted in Taiwan foundno consistent differences between soluble solids content fororganic and conventional tomatoes.45 All three of these studiesalso found no significant yield differences or clear trends that onemethod produces higher antioxidant levels in tomatoes.43 – 45

Another study comparing tomatoes grown on organic andconventional plots on the same farm for 3 years found organictomatoes to be higher in ascorbic acid and soluble solids content,but year and cultivar variability made it difficult to draw conclusionsabout other nutrient and quality characteristics studied.46 In astudy by Caris-Veyrat et al.,47 organically grown tomatoes hadsignificantly higher levels of antioxidants compared to thosegrown conventionally when reporting on a fresh weight basis,but the results were not significant on a dry weight basis.

Our results indicate there may be a difference in yield potential,with conventional production yielding more tomatoes thanorganic production. This trend was also seen in a study donein Tunisia, which found organically grown tomato yields to be 63%of conventionally grown tomato yields.48 Riahi and colleaguesfound no consistent difference in quality and antioxidant resultsbetween the two growing systems.

The results of our project found method overall to impact(P < 0.05) antioxidant capacity (ABTS and DPPH), ascorbic acidcontent, and soluble solids content, with organic tomatoes beinghigher in antioxidant capacity and ascorbic acid content andconventional tomatoes being higher in soluble solids content.With some results favoring each method, it is difficult to argue thatone method produces higher quality tomatoes than the other;continued research in well-controlled farm settings is encouragedto further explore this question.

Cultivar differences greatly influenced (P < 0.05) all of theparameters studied in this project, and this is an area wherefarmers can easily make choices that will increase the qualityand nutritional factors of the crops they grow. Chassy et al.46

also found cultivar to have the greatest impact on phytochemicalresults in a comparison of two cultivars grown organically andconventionally. High genetic diversity with regard to antioxidant

potential was found in a screening of 50 tomato cultivars.49 Eventhough the fruit of many tomato cultivars look visually the same,slight changes in size and pigmentation may influence antioxidantlevels.50 The importance of genotype on antioxidant compositionhas also been documented in other fruit, such as strawberries,51

stone fruit,52 – 54 and blueberries.55 Growers and seed companiesmay wish to consider incorporating additional antioxidant andquality parameters into their cultivar trials in addition to yield, pestresistance, appearance, and other standard criteria. With researchshowing consumers may be willing to pay more for produce withhigher nutritional levels, as well as produce that is produced locallyor organically,12,18,56 this appears to be a promising area wherefarmers, especially those selling at farmers’ markets and other localvenues, could differentiate their produce from others.

CONCLUSIONSA better understanding of how the complex interaction ofenvironmental effects, growing method, and cultivar choicesinfluence the quality and antioxidant properties of tomatoesand other fresh produce is fundamental to producing high-qualityfood. In this study, year-to-year variability and the productionmethod used to grow tomatoes affected quality and nutritionalcharacteristics, though the trends were not consistent. Genotypedifferences were shown to dramatically impact quality andantioxidant characteristics, and based on the cultivars used inthis study, by choosing cultivars in the highest groups, a 1.35- to1.67-fold gain in antioxidant levels could be achieved.

Further research on the importance of cultivar choices thatsupport production of nutritiously superior tomatoes could benefitlocal producers who target consumers seeking nutritious produce.

ACKNOWLEDGEMENTSThis project was supported by the National Research Initiative ofthe USDA Cooperative State Research, Education and ExtensionService, grant number 2005-55618-15634, and Colorado StateUniversity Agricultural Experiment Station project number 690.

REFERENCES1 Lucier G and Dettman R, Vegetables and Melons Situation and Outlook

Yearbook. [Online]. Economic Research Service, United StatesDepartment of Agriculture (2008). Available: http://www.ers.usda.gov/Publications/VGS/2008/05May/VGS2008.pdf [12 August 2008].

2 Beecher GR, Nutrient content of tomatoes and tomato products. ProcSoc Exp Biol Med 218:98–100 (1998).

3 Leonardi C, Ambrosino P, Esposito F and Fogliano V, Antioxidativeactivity and carotenoid and tomatine contents in differenttypologies of fresh consumption tomatoes. J Agric Food Chem48:4723–4727 (2000).

4 Djuric Z and Powell LC, Antioxidant capacity of lycopene-containingfoods. Int J Food Sci Technol 52:143–149 (2001).

5 Willcox JK, Catignani GL and Lazarus S, Tomatoes and cardiovascularhealth. Crit Rev Food Sci 43:1–18 (2003).

6 Goldman IL, Recognition of fruit and vegetables as healthful: vitaminsand phytonutrients. HortTechnology 13:252–258 (2003).

7 Bazzano LA, The high cost of not consuming fruits and vegetables. JAm Diet Assoc 106:1364–1368 (2006).

8 Winter CK and Davis SF, Organic foods. J Food Sci 71:R117–R124(2006).

9 Shepherd R, Magnusson M and Sjoden PO, Determinants of consumerbehavior related to organic foods. Ambio 34:352–359 (2005).

10 Torjusen H, Lieblein G, Wandel M and Francis CA, Food systemorientation and quality perception among consumers andproducers of organic food in Hedmark County, Norway. Food QualPref 12:207–216 (2001).


25

55


11 Magnusson MK, Arvola A, Hursti UKK, Aberg L and Sjoden PO, Choiceof organic foods is related to perceived consequences forhuman health and to environmentally friendly behaviour. Appetite40:109–117 (2003).

12 Yiridoe EK, Bonti-Ankomah S and Martin RC, Comparison of consumerperceptions and preference toward organic versus conventionallyproduced foods: a review and update of the literature. Renew AgricFood Syst 20:193–205 (2005).

13 Harker FR, Organic food claims cannot be substantiated throughtesting of samples intercepted in the marketplace: ahorticulturalist’s opinion. Food Qual Pref 15:91–95 (2004).

14 Lester GE, Organic versus conventionally grown produce: qualitydifferences, and guidelines for comparison studies. HortScience41:296–300 (2006).

15 Bourn D and Prescott J, A comparison of the nutritional value,sensory qualities, and food safety of organically and conventionallyproduced foods. Crit Rev Food Sci Nutr 42:1–34 (2002).

16 USDA-AMS, Farmers Market Growth. [Online]. United States Depart-ment of Agriculture Agricultural Marketing Service (2008). Available:http://www.ams.usda.gov/farmersmarkets/FarmersMarketGrowth.htm [25 May 2009].

17 Keeling-Bond J, Thimany D and Bond C, Direct marketing of freshproduce: understanding consumer purchasing decisions. Choices21:229–235 (2006).

18 Bond CA, Thilmany D and Keeling Bond J, Understanding consumerinterest in product and process-based attributes for fresh produce.Agribusiness 24:231–252 (2008).

19 USDA-NOP, Code of Federal Regulations Title 7, Part 205:National Organic Program. (2010). [Online]. Available:http://www.access.gpo.gov/nara/cfr/waisidx 10/7cfr205 10.html[6 May 2010].

20 Singleton VL and Rossi JA, Colorimetry of total phenolics withphosphomolybdic–phosphotungstic acid reagents. Am J EnolViticult 16:144–158 (1965).

21 Spanos GA and Wrolstad RE, Influence of processing and storage onthe phenolic composition of Thompson seedless grape juice. J AgricFood Chem 38:1565–1571 (1990).

22 Rivera JRE, Stone MB, Stushnoff C, Pilon-Smits E and Kendall PA, Effectsof ascorbic acid applied by two hydrocooling methods on physicaland chemical properties of green leaf lettuce stored at 5 ◦C. J FoodSci 71:S270–S276 (2006).

23 Dale MFB, Griffiths DW and Todd DT, Effects of genotype,environment, and postharvest storage on the total ascorbatecontent of potato (Solanum tuberosum) tubers. J Agric Food Chem51:244–248 (2003).

24 Miller NJ and Rice-Evans CA, Factors influencing the antioxidantactivity determined by the ABTS radical cation assay. Free RadicRes 26:195–199 (1997).

25 Lu YR and Foo LY, Antioxidant and radical scavenging activities ofpolyphenols from apple pomace. Food Chem 68:81–85 (2000).

26 Salandanan K, Bunning M, Stonaker F, Kulen O, Kendall P andStushnoff C, Comparative analysis of antioxidant properties andfruit quality attributes of organically and conventionally grownmelons (Cucumis melo L.). HortScience 44:1–8 (2009).

27 Lee SK and Kader AA, Preharvest and postharvest factors influencingvitamin C content of horticultural crops. Postharvest Biol Technol20:207–220 (2000).

28 Dumas Y, Dadomo M, Di Lucca G and Grolier P, Effects ofenvironmental factors and agricultural techniques on antioxidantcontent of tomatoes. J Sci Food Agric 83:369–382 (2003).

29 Gautier H, Diakou-Verdin V, Benard C, Reich M, Buret M, Bourgaud F,et al, How does tomato quality (sugar, acid, and nutritional quality)vary with ripening stage, temperature, and irradiance? J Agric FoodChem 56:1241–1250 (2008).

30 Weston LA and Barth MM, Preharvest factors affecting postharvestquality of vegetables. HortScience 32:812–816 (1997).

31 Mayer A-M, Historical changes in the mineral content of fruits andvegetables. Br Food J 99:207–211 (1997).

32 Davis DR, Epp MD and Riordan HD, Changes in USDA foodcomposition data for 43 garden crops, 1950 to 1999. J Am CollNutr 23:669–682 (2004).

33 Davis DR, Declining fruit and vegetable nutrient composition: what isthe evidence? HortScience 44:15–19 (2009).

34 White PJ and Broadley MR, Historical variation in the mineralcomposition of edible horticultural products. J Hortic Sci Biotechnol80:660–667 (2005).

35 White PJ, Bradshaw JE, Dale MFB, Ramsay G, Hammond JP andBroadley MR, Relationships between yield and mineralconcentrations in potato tubers. HortScience 44:6–11 (2009).

36 Fan MS, Zhao FJ, Fairweather-Tait SJ, Poulton PR, Dunham SJ andMcGrath SP, Evidence of decreasing mineral density in wheat grainover the last 160 years. J Trace Elem Med Biol 22:315–324 (2008).

37 Garvin DF, Welch RM and Finley JW, Historical shifts in the seedmineral micronutrient concentration of US hard red winter wheatgermplasm. J Sci Food Agric 86:2213–2220 (2006).

38 Farnham MW, Grusak MA and Wang M, Calcium and magnesiumconcentration of inbred and hybrid broccoli heads. J Am Soc HorticSci 125:344–349 (2000).

39 Magkos F, Arvaniti F and Zampelas A, Organic food: nutritious food orfood for thought? A review of the evidence. Int J Food Sci Technol54:357–371 (2003).

40 Worthington V, Nutritional quality of organic versus conventionalfruits, vegetables, and grains. J Altern Complement Med 7:161–173(2001).

41 Zhao X, Carey EE, Wang WQ and Rajashekar CB, Does organicproduction enhance phytochemical content of fruit andvegetables? Current knowledge and prospects for research.HortTechnology 16:449–456 (2006).

42 Brandt S, Pek Z, Barna E, Lugasi A and Helyes L, Lycopene contentand colour of ripening tomatoes as affected by environmentalconditions. J Sci Food Agric 86:568–572 (2006).

43 Barrett DM, Weakley C, Diaz JV and Watnik M, Qualitative andnutritional differences in processing tomatoes grown undercommercial organic and conventional production systems. J FoodSci 72:C441–C451 (2007).

44 Pieper JR and Barrett DM, Effects of organic and conventionalproduction systems on quality and nutritional parameters ofprocessing tomatoes. J Sci Food Agric 89:177–194 (2009).

45 Juroszek P, Lumkin HM, Yang RY, Ledesma DR and Ma CH, Fruitquality and bioactive compounds with antioxidant activity oftomatoes grown on-farm: comparison of organic and conventionalmanagement systems. J Agric Food Chem 57:1188–1194 (2009).

46 Chassy AW, Bui L, Renaud ENC, Van Horn M and Mitchell AE, Three-year comparison of the content of antioxidant microconstituentsand several quality characteristics in organic and conventionallymanaged tomatoes and bell peppers. J Agric Food Chem54:8244–8252 (2006).

47 Caris-Veyrat C, Amiot MJ, Tyssandier V, Grasselly D, Buret M,Mikolajczak M, et al, Influence of organic versus conventionalagricultural practice on the antioxidant microconstituent contentof tomatoes and derived purees: consequences on antioxidantplasma status in humans. J Agric Food Chem 52:6503–6509 (2004).

48 Riahi A, Hdider C, Sanaa M, Tarchoun N, Ben Kheder M and Guezal I,Effect of conventional and organic production systems on the yieldand quality of field tomato cultivars grown in Tunisia. J Sci FoodAgric 89:2275–2282 (2009).

49 Hanson PM, Yang RY, Wu J, Chen JT, Ledesma D, Tsou SCS, et al,Variation for antioxidant activity and antioxidants in tomato. JAm Soc Hortic Sci 129:704–711 (2004).

50 Toor RK, Lister CE and Savage GP, Antioxidant activities of NewZealand-grown tomatoes. Int J Food Sci Nutr 56:597–605 (2005).

51 Scalzo J, Politi A, Pellegrini N, Mezzetti B and Battino M, Plant genotypeaffects total antioxidant capacity and phenolic contents in fruit.Nutrition 21:207–213 (2005).

52 Gil MI, Tomas-Barberan FA, Hess-Pierce B and Kader AA, Antioxidantcapacities, phenolic compounds, carotenoids, and vitamin Ccontents of nectarine, peach, and plum cultivars from California. JAgric Food Chem 50:4976–4982 (2002).

53 Drogoudi PD, Vemmos S, Pantelidis G, Petri E, Tzoutzoukou C andKarayiannis I, Physical characters and antioxidant, sugar, andmineral nutrient contents in fruit from 29 apricot (Prunus armeniacaL.) cultivars and hybrids. J Agric Food Chem 56:10754–10760 (2008).

54 Tavarini S, Degl’Innocenti E, Remorini D, Massai R and Guidi L,Preliminary characterisation of peach cultivars for their antioxidantcapacity. Int J Food Sci Technol 43:810–815 (2008).

55 Howard LR, Clark JR and Brownmiller C, Antioxidant capacity andphenolic content in blueberries as affected by genotype andgrowing season. J Sci Food Agric 83:1238–1247 (2003).

56 Lin BH, Smith TA and Huang CL, Organic premiums of US freshproduce. Renew Agric Food Syst 23:208–216 (2008).


25

56

Research ArticleReceived: 5 April 2010 Revised: 15 June 2010 Accepted: 5 July 2010 Published online in Wiley Online Library: 5 August 2010


Inhibition of enzymatic browning in actualfood systems by the Maillard reaction productsBurce Atac Mogol, Aslı Yıldırım and Vural Gokmen∗

Abstract

BACKGROUND: The Maillard reaction occurring between amino acids and sugars produces neo-formed compounds havingcertain levels of antioxidant activity depending on the reaction conditions and the type of reactants. The objective of this studywas to investigate enzymatic browning inhibition capacity of Maillard reaction products (MRPs) formed from different aminoacids including arginine (Arg), histidine (His), lysine (Lys) and proline (Pro).

RESULTS: The inhibitory effects of the MRPs on polyphenol oxidase (PPO) were determined. The total antioxidant capacity (TAC)of MRPs derived from different amino acids were in the order Arg > His > Lys > Pro. The TAC and PPO inhibition of MRPs wereevaluated as a function of temperature (80–120 ◦C), time (1–6 h) and pH (2–12). Arg-Glc and His-Glc MRPs exhibited strong TACand PPO inhibition. Increasing temperature (up to 100 ◦C) and time also increased TAC and PPO inhibition. Kinetics analysisindicated a mixed type inhibition of PPO by MRPs.

CONCLUSION: The results indicate that the MRPs derived from Arg and His under certain reaction conditions significantlyprevent enzymatic browning in actual food systems. The intermediate compounds capable of preventing enzymatic browningare reductones and dehydroreductones, as confirmed by liquid chromatographic–mass spectrometric analyses.c© 2010 Society of Chemical Industry

Keywords: Maillard reaction products; enzymatic browning inhibition; polyphenol oxidase; reductones; dehydroreductones

INTRODUCTIONEnzymatic browning is catalyzed by polyphenol oxidase (PPO; EC1.10.3.1), which oxidizes o-quinones to polymerized dark-coloredpigments. This enzymatic reaction is considered to be the maincontributor to browning in fruits and vegetables and greatlyaffects nutritional, functional and organoleptic properties of theproduct.1 There are many chemical compounds with differentmechanisms avoiding or minimizing this discoloration. Sulfitingagents are the most effective and the cheapest antibrowningagents. They act both on o-quinones and enzymes. However, theyhave adverse health effects on asthmatic people and their usetends to be limited or even prohibited by food regulations.2,3

Thus the search for alternative inhibitors in the prevention of suchoxidative reactions has great importance.

The Maillard reaction is a chemical reaction between aminogroups and reducing sugars and forms colored and/or colorlessproducts. The functional properties of the Maillard reactionproducts depend on the reaction conditions,4 pH,5 and typeof reactants.6 The MRPs have been found to have antioxidantproperties.7 MRPs formed from glycine and glucose, cysteine-related compounds and various monosaccharides have beenreported as natural antibrowning agents.2,8 – 10

The objective of this study was to investigate the antioxidantcapacity of MRPs formed from different amino acids, and theirinhibitory effect on enzymatic browning. The Maillard reactionintermediates responsible for the inhibition of antioxidant activitywere identified by liquid chromatography coupled to massspectrometry (LC-MS). The inhibition of apple PPO was kineticallycharacterized in the presence of MRPs. The antibrowning potential

of MRPs was also tested for apple puree homogenized with MRP,and for potato cubes dipped in MRP solution.

EXPERIMENTALChemicals and consumablesApples of the variety Golden Delicious and potato tubers of thevariety Agria were obtained from a local supermarket in Ankara,Turkey. Glucose, arginine, histidine, proline, disodium hydrogenphosphate anhydrous and potassium dihydrogen phosphate werepurchased from Merck (Darmstadt, Germany). Lysine, catechol,citric acid, polyvinylpolypyrrolidone (PVPP), methanol, ABTS (2,2′-azino-bis(3-ethylbenz-thioline-6-sulfonic acid)) and DPPH (1,1-diphenyl-2-picrylhydrazyl) were purchased from Sigma (St Louis,MO, USA). Trolox (6-hydroxy-2,5,7,8-tetra-methylchromone-2-carboxylic acid) and potassium persulfate (di-potassium perox-disulfate) were purchased from Fluka (Buchs, Switzerland).

Preparation of Maillard reaction productsEquimolar amounts of different amino acids (Arg, His, Lys andPro) and glucose (Glc) were mixed in 1.5 mL glass vials to a finalconcentration of 0.125 mol L−1. The mixtures were heated in anoven (Memmert UNE 400, Braunschweig, Germany) at different

∗ Correspondenceto:Vural Gokmen,DepartmentofFoodEngineering,HacettepeUniversity, 06800 Beytepe, Ankara, Turkey. E-mail: [email protected]

Department of Food Engineering, Hacettepe University, 06800 Beytepe, Ankara,Turkey


25

57

Maillard reaction products inhibit enzymatic browning www.soci.org

temperatures (80–120 ◦C) for different times (1–6 h). In orderto determine the effect of pH, initial pH values of the reactionmediums were adjusted between 2.0 and 12.0 with H3PO4 (85%)and NaOH (4 mol L−1). The mixtures were heated at 100 ◦C for 3 h.The color of MRPs was measured as absorbance at 420 nm usinga variable-wavelength UV-visible spectrometer (model 2101 PC,Shimadzu, Kyoto, Japan).

Measurement of the total antioxidant capacity (TAC)The TAC of MRPs was measured by using DPPH and ABTS radicalscavenging methods. The results were expressed as mmol L−1 ofTrolox equivalent antioxidant capacity (TEAC).

DPPH radical scavenging methodA working solution of DPPH radicals was prepared in methanolto a final absorbance of ∼1.00 at 517 nm.11 The reaction mixturecontaining 1.9 mL DPPH working solution and 0.1 mL MRPs waskept at room temperature. Absorbance was measured at 517 nmexactly after 30 min using a Shimadzu 2101 PC model UV-visiblespectrophotometer. The control was prepared in the same manner,except that distilled water was added instead of MRPs.

ABTS radical scavenging methodThe blue-green ABTS stock solution was produced by reacting7 mmol L−1 ABTS with 2.45 mmol L−1 potassium persulfate andallowing the mixture to stand in the dark at room temperature for12–16 h before use.12 The ABTS working solution was obtainedby diluting the stock solution with water to an absorbance of 0.70(±0.02) at 734 nm. The reaction mixture containing 1.95 mL ABTSand 50 µL MRPs was kept at room temperature. The absorbancewas measured at 734 nm exactly after 6 min using a Shimadzu2101 PC model UV-visible spectrophotometer. The control wasprepared with distilled water as described previously.

Inhibition of PPOPPO extractionWashed, cored and peeled apples were used to extract PPO. Apple(100 g) was blended at 4 ◦C with 160 mL of pH 6.5 phosphatebuffer. PVPP (2 g) was added to prevent excessive browningduring blending. The homogenate was squeezed through layersof cloth and centrifuged at 10 000 × g for 15 min at 4 ◦C. Thesupernatant was used as crude PPO extract.

Measurement of PPO activityThe PPO activity was assayed spectrophotometrically.13 Inhibitionstudies were carried out under buffered conditions. The samplecuvette contained 1.6 mL of pH 4.5 citrate–phosphate buffer (McIlvaine’s), 200 µL MRPs, 200 µL of 1 mol L−1 catechol as substrateand 1 mL PPO extract (3 mL total volume). PPO was added lastto initiate the enzymatic reaction. The control was prepared withdistilled water instead of MRPs. The increase in absorbance at400 nm was recorded up to 3 min. The reaction rate was calculatedfrom the initial slope of the progress curve and percentage PPOinhibition values were calculated as follows:

PPO Inhibition (PPO) =(

1 − Initial rate (sample)

Initial rate (control)

)· 100 (1)

PPO inhibition kineticsTo determine the kinetics of PPO inhibition by MRPs, catecholconcentration was varied from 3.33 to 166.67 mmol L−1 in thereaction volume of 3 mL. MRPs (100 µL) prepared from Arg-Glc orHis-Glc was added as inhibitors for each substrate concentration.The reaction rates were determined from the initial slopes ofabsorbance versus time plots (Abs. min−1). Inhibition constants (Ki

and Ki′) were calculated according to Lineweaver–Burk plots.14

The velocity equation for mixed type inhibition is

V = Vmax[S]

Kmα + [S]α′ (2)

and the Lineweaver–Burk equation for this type inhibition is

1

V= α

Km

Vmax

1

[S]+ α′ 1

Vmax(3)

where

α = 1 + [I]

Ki(4)

and

α′ = 1 + [I]

Ki(5)

Antibrowning activity of MRPsAntibrowning activity of MRPs was tested in apple puree andpotato cubes as the actual food systems. Apple (2 g) washomogenized with 2 mL Arg-Glc or His-Glc MRPs and the volumewas made up to 12 mL with distilled water. The control wasprepared with 2 mL distilled water instead of MRPs. The potatotuber was washed, peeled and sliced into 1 cm cubes. Three cubesof potato were dipped into 20 mL MRPs for 1 min. Change of colorwas monitored using a color spectrometer (CM-3600d, Minolta,)up to 2 h and 6 h for apple puree and potato cubes, respectively.The total color difference �E was calculated for each time intervalusing the following equation:

�E =√

(L∗t − L∗

0)2 + (a∗t − a∗

0)2 + (b∗0 − b∗

0)2 (6)

The browning inhibition was calculated as follows:

Browning Inhibition % =(

1 − AUC (inhibitor)

AUC (control)

)· 100 (7)

where AUC stands for the area under the curve. The trapezoid rulewas used to calculate AUC from the plots of �E versus time up to2 h and 6 h for apple puree and potato cubes, respectively.

Analysis of MRPs by LC-MSLC-MS analyses of MRPs were performed using an Agilent 1200HPLC system (Waldbronn, Germany) consisting of a binary pump,autosampler and temperature-controlled column oven, coupledto an Agilent 6130 MS detector equipped with an electrosprayionization (ESI) interface operated using the following interfaceparameters: drying gas (N2, 30 psig) flow 13 L min−1, nebulizerpressure 30 psig, drying gas temperature 300 ◦C, capillary voltage4 kV, and fragmenter voltage 80 eV.

Analytical separation was performed on an Atlantis T3 column(250×4.6 mm, i.d. 5 µm) using the isocratic mixture of 10 mmol L−1

formic acid and methanol (90 : 10, v/v) at a flow rate of 0.5 mL min−1


25

58

www.soci.org BA Mogol, A Yıldırım, V Gokmen

at 25 ◦C. The LC eluent was directed to the MS system forthe detection of Maillard reaction intermediates formed duringheating in Arg-Glc and His-Glc mixtures. Data acquisition wasperformed in scan mode with positive ionization in an m/z rangeof 50–1000. The presence of intermediate compounds such asN-glycosylamine, reductone, dehydroreductone, Schiff base anddecarboxylated Schiff base of Arg and His were investigated inMRPs. Total ion chromatograms were extracted to confirm thepresence of corresponding intermediate structures.

Statistical analysisThe least significance difference test was used to determinesignificant difference between means at a significance level ofP < 0.05.

RESULTS AND DISCUSSIONEffect of heating time and temperature on TAC and PPOinhibition of MRPsSome MRPs, especially melanoidins, have been known to haveantioxidant activity through scavenging oxygen radicals orchelating metals.15,16 In this study, effect of heating time andtemperature on TAC and PPO inhibition capability of the MRPsobtained from different amino acids (Arg, His, Lys and Pro)and sugars (Glc and Fru) were investigated. Preliminary studiesrevealed that MRPs prepared by glucose showed greater TAC andPPO inhibition than those obtained from fructose (P < 0.05) (datanot shown). Other researchers have previously reported similarresults.2,10

Figure 1 shows the TAC of MRPs prepared from different aminoacids as influenced by temperature and heating time. The reactionmixtures were found to have TAC levels less than 0.05 mmol L−1

TEAC initially. In general, increasing the temperature also increased

the TAC of resulting MRPs for all amino acids. The TAC of MRPsderived from different amino acids were in the order Arg > His> Lys > Pro. It should be noted that Lys-Glc and Pro-Glc MRPshad very low levels of TAC compared to Arg-Glc and His-Glc MRPs.For example, the MRPs derived from Arg at 100 ◦C for 3 h hadapproximately 50 times greater TAC than the MRPs prepared fromLys and Pro under the same conditions (P < 0.05). It has beenpreviously reported that His-Glc MRPs possess peroxyl radicalscavenging activity assayed by the ORAC method.15 The TACvalues determined by ABTS and DPPH methods were found tobe comparable (data not shown). There was a positive correlation(r = 0.93) between the TAC measures obtained by ABTS and DPPHmethods for His-Glc MRPs.

In this study, inhibition potency of MRPs derived from differentamino acids was also characterized on apple PPO. Figure 2 showsthe PPO inhibition percentages of MRPs together with their TACvalues. The reaction mixtures prepared from Arg-Glc and His-Glchad a PPO inhibition potency of 15.01% and 13.68%, respectively. Ingeneral, MRPs having higher TAC also showed higher percentageinhibition of PPO regardless of the type of amino acid. Therewas a positive correlation between the TAC and PPO inhibitionpercentages of MRPs. The correlation coefficient was determinedto be 0.81 and 0.83 for Arg-Glc and His-Glc MRPs, respectively.

Similar to TAC values, MRPs derived from Arg were found tobe the best inhibitor of PPO, followed by MRPs derived from His,Lys and Pro (Fig. 2(a, b)). At 100 ◦C, increasing the reaction timealso increased PPO inhibition potency of MRPs derived from Argand His, while it was relatively stable for MRPs derived from Lys(Fig. 2(c)). MRPs formed at temperatures above 100 ◦C showeddifferent PPO inhibition potency. The PPO inhibition capacity ofMRPs first increased to an apparent maximum and then decreasedgradually regardless of the type of amino acid (data not shown).This fact has been explained as the progressive formation of

(a) (b)

(c) (d)

Figure 1. The TAC of MRPs determined using DPPH radical scavenging method: (a) Arg-Glc; (b) His-Glc; (c) Lys-Glc; (d) Pro-Glc (vertical bars indicatestandard deviations).


25

59


(a) (b)

(c)

Figure 2. TAC and PPO inhibition properties of MRPs synthesized from different amino acids under different conditions: (a) 100 ◦C for 3 h; (b) 120 ◦C for3 h; (c) 100 ◦C for 1, 2, 3 and 6 h (vertical bars indicate standard deviations).

insoluble melanoidins during heating.17 However, our resultsdid not confirm this explanation, because the MRPs obtainedwere optically clear. No insoluble melanoidins formed under thereaction conditions applied in this study. Therefore, intermediatesformed during the Maillard reaction rather than melanoidins arethought to be responsible for the inhibition of PPO.

Color was measured as absorbance at 420 nm in MRPs. Ingeneral, Arg generated darker MRPs during heating, followedby His, Lys and Pro (data not shown). This order was valid forthe TAC, which meant there was a positive correlation betweenbrowning intensity and TAC. Other researchers have also foundpositive correlation between color and antioxidant properties inprocessed foods.18

Even though Lys was known to be a very reactive amino acidin the Maillard reaction,19 it was surprising that the MRPs derivedfrom Lys showed very low levels of TAC and PPO inhibitioncapability. Therefore, this study was continued with Arg-Glc andHis-Glc MRPs for further characterization.

Effect of initial pH of MRPs on enzyme inhibition andantioxidant activityInitial pH of the reaction mixture (Arg-Glc or His-Glc) was adjustedbetween 2.0 and 12.0 in order to see the effect of pH on TACand PPO inhibition of MRPs formed at 100 ◦C for 3 h. Initial pH ofthe reaction mixture had a significant effect on the characteristicsof MRPs. As shown in Table 1, the MRPs derived from Arg underacidic conditions had very low levels of TAC and PPO inhibitioncapacity. For Arg, reaction conditions having pH 10.02 and 11.97were found more favorable to form MRPs having both higher TACand PPO inhibition. The PPO inhibition percentage was highest ata pH of 11.97 for Arg-Glc MRPs. For His, MRPs formed at pH valuesof 6.00 and 8.02 showed remarkable PPO inhibition capacity. Theinitial pH values of unheated Arg-Glc and His-Glc solutions are

Table 1. Influence of initial pH on TAC and PPO inhibition capacity ofMRPs prepared from Arg-Glc and His-Glc mixtures by heating at 100 ◦Cfor 3 h

pHTAC PPO inhibition

MRP Initial Final (mmol L−1 TEAC) (%)

Arg-Glc 2.02 2.28 0.39 ± 0.02f 12.12 ± 0.61o

2.98 2.86 0.35 ± 0.02f 7.48 ± 0.37p

3.98 4.00 0.22 ± 0.01f 12.52 ± 0.36o

5.08 4.79 1.32 ± 0.07e 16.94 ± 0.85n

6.06 5.47 6.46 ± 0.32d 24.08 ± 1.20m

8.06 6.59 11.50 ± 0.57a 10.07 ± 0.50of

10.02 8.95 8.36 ± 0.42c 29.86 ± 1.49l

11.97 9.32 9.89 ± 0.49b 83.11 ± 4.16k

His-Glc 2.02 2.05 0.68 ± 0.03f′ 0.00n′

2.99 3.00 0.73 ± 0.04f′ 0.00n′

4.04 3.99 0.75 ± 0.04f′ 0.00n′

5.05 4.87 1.57 ± 0.08e′ 19.46 ± 0.97l′

6.00 5.72 3.21 ± 0.16c′ 54.50 ± 2.72k′

8.02 7.14 2.13 ± 0.11d′ 55.14 ± 2.76k′

10.00 8.74 8.32 ± 0.42b′ 12.54 ± 0.63m′

11.98 9.27 12.74 ± 0.64a′ 10.63 ± 0.53m′

Different letters and letters with prime symbols indicate statisticallydifferent groups for Arg-Glc and His-Glc MRPs, respectively.

10.76 and 7.59, respectively. These pH values make MRPs fromboth amino acids practically applicable without changing theirinitial pH values.

The absorbance of MRPs measured at 420 nm changed as theinitial pH of reaction medium changed. Absorbance increased


25

60


gradually with pH for both Arg and His (data not shown). The MRPsderived from His-Glc had lighter color intensity, whereas Arg-GlcMRPs were darker.

Inhibition kineticsTo characterize the inhibition kinetics of MRPs on PPO, Arg-Glcand His-Glc MRPs were chosen from among others because theyshowed higher TAC and PPO inhibition percentages. A mixedtype inhibition activity of Arg-Glc and His-Glc MRPs on PPO wasdetermined from the Lineweaver–Burk double reciprocal plots(Fig. 3). The Km and Vmax values of enzyme without inhibitor werecalculated as 7.10 mmol L−1 and 0.24 Abs. min−1, respectively. TheKm increased while Vmax reduced in the presence of inhibitors. Inmixed type inhibition, inhibitor can interact not only with the freeenzyme but also with enzyme–substrate complex at a site otherthan the active site.

The Ki and Ki′ values were calculated as 22.79 µL and 82.90 µL for

Arg-Glc MRPs by means of the Lineweaver–Burk plots and relatedequations (Eqns (2)–(5)). The Ki and Ki

′ values were significantlyhigher for His-Glc MRPs than Arg-Glc MRPs, and determinedas 49.98 µL and 159.27 µL, respectively. The higher constantscalculated for the MRPs derived from His confirmed that they werestronger PPO inhibitors (P < 0.05). For both Arg-Glc and His-GlcMRPs, Ki values were lower than Ki

′ values, which indicated thatthe affinity of MRPs towards the free enzyme was higher than theenzyme–substrate complex.

Brun-Merimee et al. have reported that MRPs prepared using anaqueous model system containing equimolar glucose or fructosewith glutathione (GSH) were mixed-type inhibitors, glucose/GSHbeing the most efficient model system.20

Potential use of MRPs as antibrowning agents in foodsTo test the potential use of Arg-Glc and His-Glc MRPs for theinhibition of PPO in actual food systems, apple puree and potatocubes were used as the model systems. Apple puree was preparedby homogenizing 10 g apple slices with 1 mL MRPs or 1 mL waterfor control. Measuring the color for up to 2 h monitored browningtendency of the puree. Total color difference (�E) was calculatedto evaluate the antibrowning activity of MRPs in apple puree.Figure 4(a) shows the change of �E with time in apple puree inthe presence of His-Glc MRPs. It is clear from the results that �Erapidly increased, reaching a plateau value within 1 h at roomtemperature. Compared to the control, addition of His-Glc MRPssignificantly decreased the plateau value of �E attained after 1 h

Figure 3. Lineweaver–Burk double-reciprocal plots showing inhibition ofPPO by MRPs synthesized at 100 ◦C for 3 h.

in the apple puree model system (P < 0.05). From a practical pointof view, a �E value that exceeded 3.0 was undesirable because thecolor of the puree became apparently brownish after this point.The color turned to brownish (�E > 3.0) within 10 min in control,while it was relatively stable for 2 h in puree treated with His-GlcMRPs. Change in the surface color of potato cubes dipped in MRPsor water as control was monitored for 6 h. Similar to the resultsobtained in the apple puree system, enzymatic browning could besignificantly prevented in potato cubes dipped in MRPs (Fig. 4(b)).

Using Eqn (7), percentage inhibition of browning was calculatedby comparing the area under the curves of �E versus time plotsfor apple puree and potato cubes. In apple puree kept at roomtemperature for 2 h, browning inhibition rates were 75%±4% and52% ± 3% for Arg-Glc and His-Glc MRPs, respectively. Browningrates could be effectively decreased in potato cubes dipped inMRPs derived from Arg and His. In potato cubes, browning ratesdecreased by 51% ± 3% and 53% ± 3% for Arg-Glc and His-GlcMRPs, respectively, within 6 h of storage at room temperature.These results revealed that the MRPs derived from Arg and Hiscan act as potential antibrowning agents in actual food systems.It has been reported that Arg-Glc MRPs are limited in practical usebecause of their intense color.21 Since they have a lighter color,His-Glc MRPs are considered more desirable for food applications.

(a)

(b)

Figure 4. Change of total color difference (�E) with time in actual foodsystems as influenced by MRPs: (a) apple puree homogenized with water(control) and His-Glc MRPs prepared at 100 ◦C for 3 and 6 h; (b) potatocubes dipped in water (control), His-Glc MRPs prepared at 100 ◦C for 6 h,and Arg-Glc MRPs prepared at 100 ◦C for 3 h.


25

61


Analysis of MRPs by LC-MSThe MRPs derived from Arg and His were analyzed by LC-MS in scan mode with positive ionization in an m/z rangeof 50–1000 to confirm the formation of reductones anddehydroreductones as potential antioxidants and antibrowningagents. Total ion chromatograms were extracted to confirm thepresence of corresponding intermediate structures using theparent and compound-specific ions having certain m/z valuesfor Arg-Glc and His-Glc. In positive electrospray ionization mode,the parent ions used to confirm N-glycosylamines, reductones,dehydroreductones and Schiff bases were 337, 301, 299, and283, respectively, for Arg-Glc MRPs. They were 318, 282, 280,and 264 for His-Glc MRPs. The extracted ion chromatogramsconfirmed the presence of reductones and dehydroreductones, aswell as N-glycosylamines and Schiff bases in the MRPs. Among theintermediates, N-glycosylamines were the most abundant formsin the MRPs prepared by heating mixtures of Arg-Glc and His-Glcat 100 ◦C. This indicated that the Maillard reaction was at early andintermediate stages under the reaction conditions applied in thisstudy. The rate of change of reductones in the MRPs of Arg-Glcand His-Glc was comparable during heating at 100 ◦C for up to6 h (Fig. 5(a)), but levels of dehydroreductones were significantlyhigher in the Arg-Glc mixture than in the His-Glc mixture (Fig. 5(b)).It is thought that the initial pH of 10.47 accelerated the formationof dehydroreductones in the Arg-Glc mixture during heating.

Since the reaction products were mainly characterized bythe intermediate compounds formed during the Maillard reac-tion at 100 ◦C, high TAC and PPO inhibition capacity shouldbe due to these intermediates rather than high-molecular-weight melanoidins. Among these intermediates, reductonesand dehydroreductones are capable of active involvement in

(a)

(b)

Figure 5. Changes in peak areas for reductones and dehydroreductonesin MRPs prepared at 100 ◦C: (a) Arg-Glc; (b) His-Glc.

oxidation–reduction reactions, possibly by a hydrogen transfermechanism.22 Due to the fact that we determined significant lev-els of reductones and dehydroreductones in the MRPs derivedfrom Arg and His, these reductones formed during the Mail-lard reaction might be responsible for reversing the oxidation ofpolyphenols to quinones catalyzed by PPO.

CONCLUSIONFunctional characteristics of MRPs were investigated in terms ofTAC, PPO inhibition and antibrowning capacity in this study. Thetype of amino acids, initial pH of the reaction medium, temperatureand time were found to influence these characteristics in theresulting MRPs. Among others, the MRPs derived from Argand His under certain reaction conditions had higher TACand significantly prevented enzymatic browning in actual foodsystems. Owing to its lighter color intensity, the MRPs from His areconsidered more desirable for food applications. In-depth massspectrometric analyses confirmed the presence of reductones anddehydroreductones in MRPs derived from Arg and His. Thesecompounds, especially reductones, are thought to be potentiallyactive inhibitors of enzymatic browning. Since similar MRPs areproduced under food-processing conditions such as baking androasting, they are considered to be usual constituents of foods asPPO inhibitors that may facilitate their approval as antibrowningagents. However, the absence of any harmful compound inthe mixture of MRPs that may be formed during the Maillardreaction should be confirmed by detailed investigation. If desired,removal of aroma compounds in the mixture of MRPs shouldalso be considered to prevent any adverse effect on the sensorialproperties of foods.

ACKNOWLEDGEMENTSWe thank the Scientific and Technical Research Council of Turkeyfor financial support (Project TOVAG COST 928).

REFERENCES1 Mayer AM and Harel E, Polyphenol oxidases in plants. Phytochemistry

18:193–215 (1979).2 Billaud C, Maraschin C and Nicolas J, Inhibition of polyphenoloxidase

from apple by Maillard reaction products prepared from glucoseor fructose with L-cysteine under various conditions of pH andtemperature. LWT Food Sci Technol 37:69–78 (2004).

3 Nicolas J, Richard-Forget F, Goupy PM, Amiot M-J and Aubert SY,Enzymatic browning reactions in apple and apple products. CritRev Food Sci Nutr 34:109–157 (1994).

4 Renn PT and Sathe SK, Effects of pH, temperature, and reactant molarratio on L-leucine and D-glucose Maillard browning reaction in anaqueous system. J Agric Food Chem 45:3782–3787 (1997).

5 Cioroi M, Leonte M, Munari M, Mastracola D and Lerici CR,Chemical–physical changes in sugar–amino acid model systemsdue to Maillard reaction. Note 1. The evolution of the pH and thecolour of the glucose–lysine model system heat treated. Rev RoumChimie 45:153–157 (2000).

6 Ajandouz EH and Puigserver A, Nonenzymatic browning reaction ofessential amino acids: effect of pH on caramelization and Maillardreaction kinetics. J Agric Food Chem 47:1786–1793 (1999).

7 Franzke C and Iwansky H, Zur antioxydativen Wirksamveit derMelanoidine. Dtsch Lebens Rundsch 50:251–254 (1954).

8 Lee M-K and Park I, Inhibition of potato polyphenol oxidase by Maillardreaction products. Food Chem 91:57–61 (2005).

9 Nicoli MC, Elizalde BE, Pitotti A and Lerici CR, Effects of sugars andMaillard reaction products on polyphenol oxidase and peroxidaseactivity in food. J Food Biochem 15:169–184 (1991).


25

62


10 Billaud C, Maraschin C, Chow Y-N, Cheriot S, Peyrat-Maillard M-N andNicolas J, Maillard reaction products as ‘natural antibrowning’agents in fruit and vegetable technology. Mol Nutr Food Res49:656–662 (2005).

11 Brand-Williams W, Cuvelier ME and Berset C, Use of a free radicalmethod to evaluate antioxidant activity. LWT Food Sci Technol28:25–30 (1995).

12 Re R, Pellegrini N, Proteggente A, Pannala A, Yang M and Rice-Evans C,Antioxidant activity applying an improved ABTS radical cationdecolorization assay. Free Radic Biol Med 26:1231–1237 (1999).

13 Billaud C, Brun-Merimee S, Louarme L and Nicolas J, Effect ofglutathione and Maillard reaction products prepared from glucoseor fructose with glutathione on polyphenoloxidase from apple. I.Enzymatic browning and enzyme activity inhibition. Food Chem84:223–233 (2004).

14 Lineweaver H and Burk D, The determination of enzyme dissociationconstants. J Am Chem Soc 56:658–666 (1934).

15 Yilmaz Y and Toledo R, Antioxidant activity of water-soluble Maillardreaction products. Food Chem 93:273–278 (2005).

16 Rendleman JA, Complexation of calcium by melanoidin and its role indetermining bioavailability. J Food Sci 52:1699–1705 (1987).

17 Maillard M-N, Billaud C, Chow C-N, Ordonaud C and Nicolas J, Freeradical scavenging, inhibition of polyphenoloxidase activity andcopper chelating properties of model Maillard systems. LWT FoodSci Technol 40:1434–1444 (2007).

18 Manzocco L, Calligaris S, Mastracola D, Nicoli MC and Lerici CR, Reviewof nonenzymatic browning and antioxidant capacity in processedfoods. Trends Food Sci Technol 11:340–346 (2001).

19 Morales FJ and Jimenez-Perez S, Free radical scavenging capacity ofMaillard reaction products as related to colour and fluorescence.Food Chem 72:119–125 (2001).

20 Brun-Merimee S, Billaud C, Louarme L and Nicolas J, Effect ofglutathione and Maillard reaction products prepared from glucoseor fructose with glutathione on polyphenoloxidase from apple. II.Kinetic study and mechanism of inhibition. Food Chem 84:235–241(2004).

21 Tan BK and Harris ND, Maillard reaction products inhibit applepolyphenoloxidase. Food Chem 53:267–273 (1995).

22 Belitz H-D, Grosch W and Schieberle P, Food Chemistry (4th edn).Springer, Berlin, pp. 248–337 (2009).


25

63



Development of efficient enzymatic productionof theanine by γ -glutamyltranspeptidase froma newly isolated strain of Bacillus subtilis,SK11.004Yuying Shuai,a,b Tao Zhang,a Bo Jianga,∗ and Wanmeng Mua

Abstract

BACKGROUND: Theanine, a unique amino acid found almost exclusively in tea plants, has various favourable physiologicaland pharmacological functions in humans. Gamma-glutamyltranspeptidase (GGT, EC 2.3.2.2) is considered to be the mosteffective enzyme for the production of theanine. In fact, GGT can catalyse the transfer of γ -glutamyl moieties from γ -glutamylcompounds to water (hydrolysis) or to amino acids and peptides (transpeptidation).

RESULTS: A novel strain, SK11.004, which produces GGT with high theanine-forming ability was isolated from fermentedshrimp paste and identified as Bacillus subtilis through its physiological and biochemical properties as well as its 16S rDNAsequence analysis. Theanine (18.9 mmol L−1) was synthesised by GGT (0.06 U mL−1) through transfer reaction in the presence ofglutamine (20 mmol L−1) as a donor and ethylamine HCl (50 mmol L−1) as an acceptor at pH 10 and 37 ◦C for 4 h, the conversionrate being up to 94%.

CONCLUSION: The enzymatic synthesis of theanine using GGT from a newly isolated strain Bacillus subtilis SK11.004 was foundto be an efficient method. Moreover, compared with others, the GGT from B. subtilis SK11.004 exhibited the highest ratioof transferring activity to hydrolytic activity using glutamine, suggesting a high potential application in the production oftheanine and other functional γ -glutamyl compounds.c© 2010 Society of Chemical Industry

Keywords: theanine; gamma-glutamyltranspeptidase; enzymatic synthesis; Bacillus subtilis

INTRODUCTIONTheanine (γ -glutamylethylamide), a unique amino acid foundalmost exclusively in tea plants, elicits umami (flavour potential)and a sweet taste, which determine the quality of green tea.Moreover, theanine offers various favourable physiological andpharmacological functions: it promotes relaxation,1 helps toreduce blood pressure,2 inhibits the negative effects of caffeine,3

enhances anti-tumour activity,4 provides neuroprotection,5 actsas an anti-obesity compound,6 and improves learning ability.7

The US Food and Drug Administration (FDA) admitted andaccepted theanine early in 1985, and confirmed that theanineis Generally Recognised As Safe (GRAS) with no limitation onthe quantity during application. For instance, Sun-theanine

(Taiyo Kagaku, Co. Ltd, Yokkaichi, Mie, Japan) has been used tofacilitate relaxation and in the treatment of anxiety and depression,which are prominent psychiatric illnesses in today’s society.Therefore, many attempts have been made to produce theaninecommercially, including tea callus cultivation, chemical synthesis,extraction from fresh leaves of the tea plant and enzymaticsynthesis. The latter has been intensively investigated, involvingtheanine synthetase (EC 6.3.1.6), glutamine synthetase (EC6.3.1.2),8 glutaminase (EC 3.5.1.2) and γ -glutamyltranspeptidase(gamma-glutamyltransferase, GGT, EC 2.3.2.2).9 However, most of

these methods appear to use complicated processes with a lowyield and are expensive.

GGT catalyses the transfer of the γ -glutamyl moieties from γ -glutamyl compounds to water (hydrolysis) or to amino acids andpeptides (transpeptidation). Its transpeptidation reaction couldbe exploited for the production of theanine, and seemed to beone of the most effective methods due to the availability ofenzyme source, short reaction time and non-requirement of ATP.However, theanine was produced at a relatively low yield with ahigh quantity of the by-product γ -glutamylglutamine using GGTfrom Escherichia coli K-12.9 In fact, various microorganisms suchas Aspergillus oryzae,10 E. coli K-12,11 Helicobacter pylori,12 Proteusmirabilis,13 Saccharomyces cerevisiae,14 Treponema denticola15 andmany Bacillus species, including Bacillus sp. KUN-17,16 Bacillus

∗ Correspondence to: Bo Jiang, State Key Laboratory of Food Science and Tech-nology, Jiangnan University, 1800 Lihu Avenue, 214122 Wuxi, Jiangsu, China.E-mail: [email protected]

a State Key Laboratory of Food Science and Technology, Jiangnan University,1800 Lihu Avenue, 214122 Wuxi, Jiangsu, China

b School of Food Science and Technology, Jiangnan University, 1800 Lihu Avenue,214122 Wuxi, Jiangsu, China


25

64

www.soci.org Y Shuai et al.

pumilus,17 Bacillus subtilis 168,18 Bacillus subtilis NAFM5,19 Bacillussubtilis natto20 and Bacillus subtilis NX-2,21 have been found toproduce GGT. Fermented shrimp paste, a traditional and popularseasoning in Asian countries, was shown to contain many strainsof Bacillus species.22

The main objective of this study was to isolate strains producingGGT with high theanine-forming activity from a traditionalfermented shrimp paste and develop a more efficient methodto produce theanine.

MATERIALS AND METHODSStrain isolation and identificationShrimp paste (1 g; Lee Kum Kee) from a local market wastransferred into a 250 mL Erlenmeyer flask containing 30 mLmedium A containing 20 g glucose, 15 g yeast extract, 10 g cornsteep liquor, 0.5 g MgSO4·7H2O and 1 g K2HPO4·3H2O L−1 at pH7.2. The inoculated medium was incubated at 37 ◦C in a rotaryshaker at 200 rpm. After 30 h, a loop of the culture was spread overseparate plates of solid medium B containing 10 g glucose, 5 gyeast extract, 5 g glutamate and 0.5 g KH2PO4 L−1 at pH 7.2, andincubated at 37 ◦C. After 48 h, fast-growing colonies were pickedand incubated in medium A again. Then the culture broth wascentrifuged at 10 000 × g and 4 ◦C for 30 min to remove insolublematerials. The supernatant was tested for GGT activity. The strainsthat produced GGT with high activity were taken for a secondscreening to determine the theanine-forming capability of theirGGTs according to the enzymatic synthesis described below.

The strain producing GGT with high theanine-forming capabilitywas selected for further research and then identified. The morpho-logical and biochemical tests of the isolated strain were conductedusing Biolog Microstation System (Biolog Inc., Hayward, CA, USA).Genomic DNA was extracted with a Bacteria Genomic DNA Iso-lation Kit (Shanghai Chaoshi Bio Technologies Co. Ltd, Shanghai,China). Liquor (1–3 µL) of the supernatant was used as the tem-plate DNA for PCR amplification. The 16S rDNA gene was amplifiedwith bacterial universal primers 27F (5′-GAGTTTGATCCTGGCTCAG-3′) and 1527R (5′-AGAAAGGAGGTGATCCAGCC-3′), which wereused for sequencing (China Center for Type Culture Collection,Wuhan, China). Homology searches were performed against thesequences of database using the BLAST program (NCBI, Bethesda,MD, USA).

Enzyme productionFor the production of GGT with a high activity, a loop of theisolated strain was inoculated into 30 mL optimised mediumcontaining 25 g sucrose, 5 g tryptone, 15 g corn steep liquor, 0.5 gMgSO4·7H2O and 1 g K2HPO4·3H2O L−1 at pH 7.2 in 250 mL flaskand incubated at 37 ◦C in a rotary shaker at 200 rpm for 20 h. Theculture broth was centrifuged at 10 000 × g and 4 ◦C for 30 minto remove insoluble materials. The supernatant was used as theenzyme.

Assay of γ -glutamyltranspeptidase activityThe enzyme assay was performed using the method of Orlowskiand Meister with slight modifications.23 The assay mixture(1 mL) was composed of 50 mmol L−1 borate–NaOH (pH 10),5 mmol L−1 L-γ -glutamyl-p-nitroanilide (γ -GpNA; Sigma, St Louis,MO, USA), 20 mmol L−1 glycylglycine and the enzyme solution.After incubation at 37 ◦C for 30 min, the reaction was terminatedby the addition of 0.1 mol L−1 HCl. The transferase activity

was calculated from the difference in absorbance at 410 nmbetween reaction mixtures with and without glycylglycine. Oneunit of GGT was defined as the amount of enzyme that released1 µmol of p-nitroaniline per minute from γ -GpNA through thetranspeptidation reaction.

Enzymatic synthesis of theanine with γ -glutamyl-transpeptidaseThe reaction system (10 mL in a flask) constituted of0.04 U mL−1 GGT, 20 mmol L−1 glutamine (Gln) and 100 mmol L−1

ethylamine HCl. The reaction was carried out in a shaking waterbath at 37 ◦C and 150 rpm for 5 h. Reaction mixtures withoutthe addition of GGT served as controls and were incubated sim-ilarly. The reaction was terminated by the addition of 100 g L−1

trichloroacetic acid.

Analysis of theanine, glutamine and glutamic acidTheanine, Gln and glutamic acid (Glu) were determined us-ing HPLC equipped with a C18 ODS HYPERSIL column (Agilent1200; Agilent Technologies, Palo Alto, CA, USA), with a gra-dient elution at 40 ◦C and a flow rate of 1 mL min−1. Thegradient of mobile phase was formed with buffer A (8 g L−1

CH3COONa·3H2O, containing 0.16 mg L−1 triethylamine, and4.4 mg L−1 tetrahydrofuran, pH 7.2), and buffer B (20 g L−1

CH3COONa·3H2O, pH 7.2/acetonitrile/methanol (1 : 2 : 2, by vol-ume). A/B ratios were 92 : 8, 62 : 38, 0 : 100, 0 : 100 and 92 : 8,at run times of 0, 20, 24, 25.5 and 28.5 min, respectively.o-Phthaldialdehyde (Sigma) was used as the pre-column deriva-tion reagent and detection was performed by a UV detector(model LC-9A; Shimadzu, Kyoto, Japan) at 338 nm, with excitationat 262 nm. Typically, Glu, Gln and theanine standards (Sigma) wereeluted after 3.2, 7.5 and 13.6 min, respectively.

Analysis of purified theanineThe purified theanine was identified by a Waters Synapt Q-TOF mass spectrometer (Waters Corporation, Manchester, UK).The instrument was operated in positive ion mode with theelectro spray ionisation (ESI) capillary set to 3.5 kV. The nitrogendesolvation gas was set to 600 L h−1, the source block temperatureto 100 ◦C, the desolvation temperature to 250 ◦C, and the conevoltage to 30 V.

RESULTS AND DISCUSSIONScreening and identification of strains that produceγ -glutamyltranspeptidaseAbout 20 GGT-producing strains were isolated from fermentedshrimp paste, among which four strains produced GGT withtransferring activity (above 2.0 U mL−1) relatively close to thehighest GGT activity (3.2 U mL−1) exhibited by B. subtilis NX-2.21

Some characteristics of GGT produced by the four strains weredetermined and compared (Table 1). Results showed that differentGGTs have dissimilar specificity toward the same substrate Gln.Obviously, using GGT from strain SK11.004 as a catalyst, the highestproduction of theanine was 16 mmol L−1 with a conversion yieldof 80%. The amount of theanine formed was much greater thanthat produced by GGT from E. coli K-12,9 which has been reportedto be efficient for theanine production. It is worth noting that theratio of transferring activity to hydrolytic activity of Gln was foundto be 16.8 and this ratio is much higher than those observed for γ -glutamyl transferring enzymes from other microorganisms.11,20,24


25

65

Enzymatic production of theanine by B. subtilis GGT www.soci.org

Table 1. Characteristics of the GGTs produced by four isolated strainsof B. subtilis

StrainGGT activity

(U mL−1)Conversionratioa (%) Theanine:Glub

SK11.001 2.33 ± 0.05 56.3 ± 0.14 5.29 ± 0.16

SK11.002 2.01 ± 0.04 62.5 ± 0.16 8.00 ± 0.21

SK11.003 2.15 ± 0.03 64.2 ± 0.1 6.42 ± 0.33

SK11.004 2.30 ± 0.04 80.0 ± 0.2 16.8 ± 0.26

a The conversion rate was calculated taking into account the amountof substrate (glutamine) converted into product (theanine).b Ratio of theanine to glutamic acid (Glu). Data were expressed as mean± SD from three independent experiments.

0.01 0.02 0.03 0.05 0.070

5

10

15

20

The

anin

e (m

M)

Ethylamine HCI concentration (mol L-1)

0

1

2

3

4

5

Glu

(m

M)

0.060.04

Figure 1. Effect of ethylamine HCl concentration on theanine synthesis.The reaction was carried out with 20 mmol L−1 Gln and 0.04 U mL−1 GGT atpH 10 and 37 ◦C for 5 h, the concentration of ethylamine HCl being variedas indicated in the figure. Theanine (shaded columns) and Glu (unfilledcolumns) were analysed by HPLC. Data were expressed as mean ± SD fromthree independent experiments.

These findings suggested a potential application of strain SK11.004in theanine production.

Strain SK11.004 was a rod-shaped, endospore-forming andGram-positive bacterium, which grows well in aerobic culture.Its 16S rDNA sequence (1432 bp) showed 99% of similarity withBacillus subtilis 79-9 16S rDNA gene (EU624 322.1) and has beendeposited in GenBank under the accession number FJ437210.It was identified as B. subtilis based on its physiological andbiochemical properties as well as its 16S rDNA sequence analysis.

Enzymatic synthesis of theanine with γ -glutamyl-transpeptidaseTo increase the yield of theanine, the reaction conditions for itssynthesis were investigated. When the concentration of donor(Gln) was fixed at 20 mmol L−1, the optimal concentration ofacceptor (ethylamine HCl) was found to be 50 mmol L−1 (Fig. 1).Moreover, Gln was not detected in the reaction product, indicatingthat Gln has been used up as a γ -glutamyl donor. Similarly, theoptimum concentration of GGT was 0.06 U mL−1 (Fig. 2). Whenhigher concentrations of GGT were used, the amounts of theanineand Glu decreased after 5 h. This may be due to the fact thatGGT catalyses the auto-transpeptidation reaction in which Glnitself serves as an acceptor of the γ -glutamyl group when the

0.02 0.04 0.06 0.08 0.100

5

10

15

20

The

anin

e (m

mol

L-1

)

GGT concentration (U mL-1)

0

1

2

3

4

5

Glu

(m

mol

L-1

)

Figure 2. Effect of enzyme concentration on theanine synthesis. Thereaction was carried out with 20 mmol L−1 Gln and 50 mmol L−1

ethylamine HCl, at pH 10 and 37 ◦C for 5 h, the concentration of GGTbeing varied as indicated in the figure. Theanine (shaded columns) and Glu(unfilled columns) were analysed by HPLC. Data were expressed as mean± SD from three independent experiments.

9.0 9.5 10.0 10.5 11.00

5

10

15

20T

hean

ine

(mm

ol L

-1)

pH values

0

1

2

3

4

5

Glu

(m

mol

L-1

)

Figure 3. Effect of pH on theanine synthesis. The reaction was carriedout with 20 mmol L−1 Gln, 50 mmol L−1 ethylamine HCl, and 0.06 U mL−1

GGT at 37 ◦C for 5 h, the pH of reaction mixture being varied as indicatedin the figure. Theanine (shaded columns) and Glu (unfilled columns)were analysed by HPLC. Data were expressed as mean ± SD from threeindependent experiments.

concentration of GGT is increased. The effect of pH was also studied(Fig. 3). After 5 h of incubation, the best conversion yield wasobtained at pH 10. The content of Glu increased with decreasingpH from 10 to 7. Results revealed that at pH values lower than10, the hydrolytic reaction is preferred. However, when the pH isabove 10, the auto-transpeptidation reaction occurs. These factsindicated that if the pH of reaction mixture is well adjusted, GGT canselectively catalyse the transpeptidation reaction. The optimumincubation time was 4 h (Fig. 4). The optimum conditions forthe synthesis of theanine were determined to be 20 mmol L−1

Gln, 50 mmol L−1 ethylamine HCl, 0.06 U mL−1 GGT, pH 10, andincubation at 37 ◦C for 4 h. The yield was 18.9 mmol L−1 and theconversion rate of Gln to theanine reached 94% (Fig. 5). Theanineproduction was achieved in a shorter time with a small amountof enzyme and by-product. However, the yield was higher thanthose obtained using GGT from Pseudomonas taetrolens Y-30,8

Methylovorus mays No. 925 and Pseudomonas citronellosis GEA.26


25

66

www.soci.org Y Shuai et al.

10

5

10

15

20T

hean

ine

(mm

ol L

-1)

Incubation time (h)

0

1

2

3

4

5

Glu

(m

mol

L-1

)

2 3 4 5 6

Figure 4. Effect of incubation time on theanine synthesis. The reactionwas carried out with 20 mmol L−1 Gln, 50 mmol L−1 ethylamine HCl, and0.06 U mL−1 GGT at 37 ◦C and pH 10. Theanine (shaded columns) and Glu(unfilled columns) were analysed by HPLC. Data were expressed as mean± SD from three independent experiments.

Purification and identification of theanineAfter theanine had been synthesised, the enzymatic reactionmixture (20 mL) was applied to a column (20 mL) of 201 × 7 strongbasic resin (Cl− form) to absorb Glu, theanine being eluted withMilli Q water. The fractions containing theanine were then appliedto a column (20 mL) of 001 × 7 strong acidic resin (H+ form).Theanine was eluted with ammonia solution, and the fractionscontaining only theanine were collected and lyophilised afterremoving ammonia under vacuum at 60 ◦C.

The purified theanine was subjected to positive-mode ESI-MSanalysis. The molecular mass of theanine was determined to be174 Da (Fig. 6). As shown in Fig. 6, a strong signal with an m/z valueof 175, which corresponds to the ionised theanine in positivemode, was observed. The protonated molecule at m/z 175 losesthe ammonia residue to yield a fragment at m/z 158, which in turn

0 5 10 15 200

100

200

300

400

0 5 10 15 200

20

40

60

80

100 Glu

Gln

A

B

Res

pons

e va

lue

(mA

u)

Time (min)

Theanine

Figure 5. HPLC of the enzymatic synthesis of theanine by GGT underoptimised conditions: (A) product of enzymatic synthesis; (B) Glu, Gln andtheanine standards.

loses the ethylamine group to generate an ion at m/z 130. Theseresults indicated that theanine had been successfully synthesisedand purified.

CONCLUSIONSIn this study, Bacillus subtilis SK11.004, which produces GGTwith high theanine-forming activity, was isolated from fermentedshrimp paste. The goal of developing an effective method forthe production of theanine was achieved using the GGT fromB. subtilis SK11.004. These findings are of considerable impor-tance in the feasibility of industrial production of this func-tional food additive and pharmaceutical intermediate to meet

Figure 6. Q-TOF mass spectra of the [M+H]+ ion of the purified theanine.


25

67

Enzymatic production of theanine by B. subtilis GGT www.soci.org

the increasing needs of humans. Moreover, compared withothers, the GGT from B. subtilis SK11.004 exhibited the high-est ratio of transferring activity to hydrolytic activity of Gln(the donor of γ -glutamyl moiety), suggesting a potential ap-plication in the production of other functional γ -glutamylcompounds. However, further investigation is required to ex-tensively characterise the GGT produced by the newly isolatedstrain.

ACKNOWLEDGEMENTSThis research was financially supported by the National NaturalScience Foundation of China (No. 20376029), and the ResearchProgram of State Key Laboratory of Food Science and Technology,Jiangnan University (No. SKLF-MB-200804 and SKLF-TS-200805).

REFERENCES1 Lu K, Gray MA, Oliver C, Liley DT, Harrison BJ, Bartholomeusz CF, et al.,

The acute effects of L-theanine in comparison with alprazolamon anticipatory anxiety in humans. Hum Psychopharmacol Clin Exp19:457–465 (2004).

2 Yokogoshi H, Kato Y, Sagesaka YM, Takihara-Matsuura T, Kakuda T andTakeuchi N, Reduction effect of theanine on blood pressure andbrain 5-hydroxyindoles in spontaneously hypertensive rats. BiosciBiotechnol Biochem 59:615–618 (1995).

3 Kakuda T, Nozawa A, Unno T, Okamura N and Okai O, Inhibiting effectsof theanine on caffeine stimulation evaluated by EEG in the rat.Biosci Biotechnol Biochem 64:287–293 (2000).

4 Sadzuka Y, Sugiyama T and Sonobe T, Improvement of idarubicininduced antitumor activity and bone marrow suppression bytheanine, a component of tea. Cancer Lett 158:119–124 (2000).

5 Egashira N, Hayakawa K, Mishima K, Kimura H, Iwasaki K andFujiwara M, Neuroprotective effect of γ -glutamylethylamide(theanine) on cerebral infarction in mice. Neurosci Lett 363:58–61(2004).

6 Zheng G, Sayama K, Okubo T, Juneja LR and Oguni I, Anti-obesityeffects of three major components of green tea, catechins, caffeineand theanine, in mice. In Vivo 18:55–62 (2004).

7 Juneja LR, Chu DC, Okubo T, Nagato Y and Yokogoshi H, L-theanine – a unique amino acid of green tea and its relaxationeffect in humans. Trends Food Sci Technol 10:199–204 (1999).

8 Yamamoto S, Wakayama M and Tachiki T, Cloning and expression ofPseudomonas taetrolens Y-30 gene encoding glutamine synthetase:an enzyme available for theanine production by coupledfermentation with energy transfer. Biosci Biotechnol Biochem70:500–507 (2006).

9 Suzuki H, Izuka S, Miyakawa N and Kumagai H, Enzymatic productionof theanine, an ‘‘umami’’ component of tea, from glutamineand ethylamine with bacterial γ -glutamyltranspeptidase. EnzymeMicrob Technol 31:884–889 (2002).

10 Tomita K, Ito M, Yano T, Kumagai H and Tochikura T, γ -Glutamyltranspeptidase activity and the properties of theextracellular glutaminase from Aspergillus oryzae. Agric Biol Chem52:1159–1163 (1988).

11 Suzuki H, Kumagai H and Tochikura T, γ -Glutamyltranspeptidasefrom Escherichia coli K-12: purification and properties. J Bacteriol168:1325–1331 (1986).

12 Boanca G, Sand A and Barycki JJ, Uncoupling the enzymaticand autoprocessing activities of Helicobacter pylori γ -glutamyltranspeptidase. J Biol Chem 281:19029–19037 (2006).

13 Nakayama R, Kumagai H and Tochikura T, Purification and propertiesof γ -glutamyltranspeptidase from Proteus mirabilis. J Bacteriol160:341–346 (1984).

14 Mehdi K, Thierie J and Penninckx MJ, γ -Glutamyltranspeptidase in theyeast Saccharomyces cerevisiae and its role in the vacuolar transportand metabolism of glutathione. Biochem J 359:631–637 (2001).

15 Chu L, Xu X, Dong Z, Cappelli D and Ebersole JL, Role for recombinantγ -glutamyltransferase from Treponema denticola in glutathionemetabolism. Infect Immun 71:335–342 (2003).

16 Hwang SY, Ryang JH, Lim WJ, Yoo ID and Oishi K, Purification andproperties of γ -glutamyl transpeptidase from Bacillus sp. KUN-17.J Microb Biotechnol 6:238–244 (1996).

17 Moallic C, Dabonne S, Colas B and Sine JP, Identification andcharacterization of a γ -glutamyltranspeptidase from a thermo-alcalophile strain of Bacillus pumilus. Protein J 25:391–397 (2006).

18 Minami H, Suzuki H and Kumagai H, Salt-tolerant γ -glutamyltranspeptidase from Bacillus subtilis 168 with glutaminaseactivity. Enzyme Microb Technol 32:431–438 (2003).

19 Kimura K, Tran LS, Uchida I and Itoh Y, Characterization of Bacillussubtilis γ -glutamyltransferase and its involvement in thedegradation of capsule poly-gamma-glutamate. Microbiology150:4115–4123 (2004).

20 Ogawa Y, Hosoyama H, Hamano M and Motai H, Purification andproperties ofγ -glutamyltranspeptidase from Bacillus subtilis (natto).Agric Biol Chem 55:2971–2977 (1991).

21 Wu Q, Xu H, Zhang L, Yao J and Ouyang P, Production, purificationand properties of γ -glutamyltranspeptidase from a newly isolatedBacillus subtilis NX-2. J Mol Catal B: Enzym 43:113–117 (2006).

22 Phromraksa P, Nagano H, Kanamaru Y, Izumi H, Yamada C andKhamboonruang C, Characterization of Bacillus subtilis isolated fromAsian fermented foods, Food Sci Technol Res 15:659–666 (2009).

23 Orlowski M and Meister A, γ -Glutamyl-p-nitroanilide: a newconvenient substrate for determination and study of L-and D-γ -glutamyltranspeptidase activities. Biochim Biophys Acta73:679–681 (1963).

24 Nakayama R, Kumagai H and Tochikura T, Purification and propertiesof γ -glutamyltranspeptidase from Proteus mirabilis. J Bacteriol160:341–346 (1984).

25 Sachiko Y, Mamoru W and Takashi T, Characterization of theanine-forming enzyme from Methylovorus mays No. 9 in respect toutilization of theanine production. Biosci Biotechnol Biochem71:545–552 (2007).

26 Tachiki T, Okada Y, Ozeki M, Okubo T, Juneja L and Yamazaki N,Process for producing theanine. JP Patent 2002–229026 (2002).


25

68

Research ArticleReceived: 4 February 2010 Revised: 7 July 2010 Accepted: 9 July 2010 Published online in Wiley Online Library: 17 August 2010


Freshness characterisation of whiting(Merlangius merlangus) using an SPME/GC/MSmethod and a statistical multivariate approachGuillaume Duflos,a∗ Francois Leduc,a Assi N’Guessan,b Frederic Krzewinski,c

Ossarath Kolc and Pierre Mallea

Abstract

BACKGROUND: The freshness of whiting was studied at five stages of ice storage by comparing the analysis of volatilecompounds obtained through solid phase microextraction/gas chromatography/mass spectrometry (SPME/GC/MS) with twosensory methods.

RESULTS: Of the volatile compounds identified, 38 were analysed using a statistical multivariate approach and classifiedaccording to their role in the estimation of freshness during storage as markers of freshness or spoilage. Regarding the evolutionof the presence or absence of individual compounds, three categories were defined. For example, the volatile compoundspropanal, hexanal, 1-penten-3-ol, pentanal, 2,3-pentanedione, 1-penten-3-one, heptanal, (E)-2-pentenal, 2,3-octanedione,(Z)-2-penten-1-ol, 1-pentanol, butanal, octanal, 3,5,5-trimethyl-2-hexene, 1-hexanol and 4,4-dimethyl-1,3-dioxane appearedhighly relevant, because they are found throughout storage and can be divided into several categories that are directly relatedto the quality of fish.

CONCLUSION: SPME/GC/MS combined with a statistical multivariate approach may be a useful method to identify volatilecompounds and characterise fish freshness during storage.c© 2010 Society of Chemical Industry

Keywords: fish freshness; SPME; GC/MS; principal component analysis; volatile compounds

INTRODUCTIONFish is a fragile product. Its quality is a health and commercialissue that concerns industrialists, official inspection authoritiesand consumers. Fish freshness is therefore a key attribute of itsquality.

Odour is one of the main indicators that consumers use to assessthe freshness of fish. The smell of fish changes rapidly according tothe product’s degree of freshness, and this is why sensory analysesare used by consumers and industrialists to assess fish quality.Since odour is one of the main parameters used to determinethe sensory quality of products, the key volatile compounds thatcontribute to this characteristic odour can be measured and usedas quality indicators.1 – 3 These volatile aromatic compounds thatcharacterise smell are generated by the action of bacterial andtissue enzymes and lipid autoxidation.

The microbial production of volatile compounds is caused byspecific spoilage bacteria.4 These include mainly Gram-negativepsychrotrophs such as Pseudomonas spp. and Alteromonas spp.5

and more specifically Shewanella putrefaciens and Photobacteriumphosphoreum.6,7 These bacteria generate odours by producingalcohols, carbonyls, esters and sulfur compounds8 – 10 correspond-ing to different kinds of odour such as flat and neutral odours, andthen rancid, putrid odours or odours of ammonia or sulfur chang-ing the flavour and taste of fish during storage.11 – 13 On the otherhand, odours of fresh fish are mainly associated with compounds

that contain carbonyls and alcohols and are characterised by light,green, delicate, melon, marine plant and iodised notes.11,13

In view of the role that volatile compounds play in theassessment of fish freshness, studies have been undertaken toidentify and quantify these compounds in various fish (mackerel,herring, cod, sardine and sea bream) using either static headspaceanalysis (SHA), dynamic headspace analysis (DHA) or an electronicnose in various storage phases.3,4,14 – 18

Another technique is solid phase microextraction (SPME). SPMEis a relatively new method that can be used to extract volatilesfrom foods. It is a rapid and simple extraction technique that doesnot require the use of solvents. SPME was first used to assess fish

∗ Correspondence to: Guillaume Duflos, ANSES, Laboratoire des Produits de laPeche, Bld Bassin Napoleon, F-62200 Boulogne-sur-Mer, France.E-mail: [email protected]

a ANSES, Laboratoire des Produits de la Peche, Bld Bassin Napoleon, F-62200Boulogne-sur-Mer, France

b Laboratoire Paul Painleve, UMR CNRS 8524 & Ecole Polytechnique Universitairede Lille, Universite de Lille 1, Avenue Paul Langevin, F-59655 Villeneuve d’AscqCedex, France

c Unite de Glycobiologie Structurale et Fonctionnelle, UMR CNRS 8576, Universitede Lille 1, F-59655 Villeneuve d’Ascq Cedex, France


25

69

Freshness characterisation of whiting www.soci.org

quality by Bene et al.19,20 when studying the volatile substances insalmon and whiting. Other researchers have used the techniqueon crayfish, rainbow trout, salmon, mackerel, sardine, tuna, seabream, cod and shrimp to study their volatile compounds duringstorage.16,21 – 26

SHA, DHA and SPME are often combined with gas chromatog-raphy/mass spectrometry (GC/MS) to determine the quality andquantity of various volatile aromatic compounds that can berelevant quality indicators.

The aims of this study were to examine the various volatilecompounds in whiting during storage using the SPME techniquecombined with GC/MS, to ascertain if there is a relationshipbetween these compounds and fish quality as assessed througha sensory analysis method and to determine which of thesecompounds may be markers of fish freshness or spoilage using astatistical multivariate approach.

EXPERIMENTALChemicalsCarboxen (CAR)/polydimethylsiloxane (PDMS) StableFlex fibre(65 µm) came from Supelco (Bellefonte, PA, USA). Before firstuse, each SPME fibre was conditioned as recommended by themanufacturer. NaCl came from Oxoid Ltd (Basingstoke, UK), Milli-Q water (high-performance liquid chromatographic water) fromFisher Scientific Labosi (Elancourt, France) and 3-methyl-3-buten-1-ol from Sigma-Aldrich (Saint Quentin Fallavier, France).

Sample preparationWhiting (Merlangius merlangus), caught the night before the startof the study, was acquired from Cooperative Maritime Etaploise(Boulogne-sur-Mer, France). Two different catches (20 and 15 fishrespectively) were analysed. The fish were stored in crushed ice at4 ◦C in self-draining polystyrene boxes for 7 days. Fresh crushedice was added daily. Sensory evaluation and volatile analysis wereperformed on days 1, 2, 3, 4 and 7 (on seven different fish eachday). Sensory evaluation was performed by two panellists familiarwith the sensory evaluation of fish. Following sensory evaluation,each fish was filleted. The next step was according to previouswork.27 The fillets were cut into 1 cm cubes, then 50 g of fleshwas introduced into a stomacher bag with 100 mL of ultrapurewater saturated with NaCl. The contents were mixed for 2 min in aStomacher Lab-Blender 400 (Seward, Thetford, UK). The aqueousphase was removed and centrifuged at 12 000 × g for 10 minat 4 ◦C (Multifuge 3 S-R Heraeus, Kendro Laboratory Products,Courtaboeuf, France).

Sensory evaluationTwo methods were used for the sensory evaluation of fish.

The first method is the European Union’s grading system pre-sented in European Directive 2406/96. This system distinguishesbetween three freshness categories, E, A and B, corresponding tovarious levels of spoilage. Category E corresponds to the highestquality level, followed by categories A and B, while fish gradedbelow B is considered to be non-edible. Another category cor-responds to the product’s discard level. In order to rate thisevaluation, these letters have been replaced with numbers: 0 = E,1 = A, 2 = B and 3 = unacceptable. The lower the number, thefresher the fish; conversely, the higher the number, the greater thespoilage (below 18 the fish is acceptable).

The second method is the quality index method (QIM) evaluationsystem adapted to whiting. It is based on changes in the sensorycharacteristics of raw fish during spoilage. Scores of 0–1, 0–2 or0–3 demerit (or index) points are attributed according to changesobserved in the smell, texture, eye appearance, skin and gills. Thepoints are added up to obtain an overall sensory score or qualityindex (QI).

SPME procedureAccording to previous studies,27 11 mL of aqueous phase fromthe supernatant of sample preparation was introduced into ahermetically sealed 20 mL vial. The vial was placed in the sampletray of a Combi PAL (CTC Analytics, Zwingen, Switzerland) and thentransferred to the mixer, where it was heated at 50 ◦C and mixedat 500 rpm for 10 min. After this equilibrium time the CAR/PDMSSPME fibre was inserted into the headspace of the sample andheld there for 40 min at 50 ◦C. The fibre was then removed fromthe headspace and inserted into the Merlin Microseal injector(250 ◦C) of a GC-17A gas chromatograph (GC) equipped with anMS-QP5000 mass spectrometer (MS) (Shimadzu, Kyoto, Japan) fordesorption. The fibre was maintained 10 seconds.

GC/MS procedureThe GC was equipped with a BPX5 capillary column (60 m ×0.25 mm×0.25 µm) (SGE, Courtaboeuf, France). The GC conditionswere as follows: oven temperature set initially at 35 ◦C (5 min hold),increased to 100 ◦C at 10 ◦C min−1, then increased to 280 ◦C at20 ◦C min−1 and maintained at 280 ◦C for 5 min; the splitless modewas used for injection, with a purge time of 2 min. The fibre wasmaintained in the injection port for 10 s. The electron impactMS conditions were as follows: temperature of interface, 260 ◦C;ionisation voltage, 70 eV; mass range, m/z 33–200; scan speed,250 per 0.5 s. After each injection the fibre was heated to 300 ◦Cfor 10 min in the SPME fibre conditioner. Volatile compoundswere identified by matching their mass spectra with Mass SpectralLibraries 21 and 107 of the National Institute of Standards andTechnology (NIST) (developed for Shimadzu by NIST, July 2002).Semi-quantification of the components was based on arbitraryunits of total current ion peak area counts.

Statistical analysisThe statistical analysis of the 35 samples was divided into twoparts. First we described and analysed the freshness and spoilageindices on the various sampling days using PASW Statistics 18(SPSS, Paris, France). Two graphic approaches were used to thatend. One approach, using box plots, shows a statistical summaryof the indices by sampling day. Another graphic approach basedon control charts of the individual index values was used for thisfirst part. These control charts were used to analyse the dailyprogression of freshness and/or spoilage indices. The second partof our statistical analysis proposes a grouping of analysed fishsamples and identifies the volatile compounds that characterisetheir categories. To do so, we used SPAD 7 (SPAD, Paris, France) toselect a set of relevant volatile compounds via the first two trendsin a principal component analysis (PCA). We then performed anascending hierarchical classification (AHC) of the first two factors.The categories were then analysed in relation to the selectedvolatile compounds and the freshness and/or spoilage indices.


25

70

www.soci.org G Duflos et al.

Table 1. Volatile compounds identified during spoilage analysis of whiting

Variablename Compound name

Rentiontime (min)

Variablename Compound name

Rentiontime (min)

a Acetaldehyde 3.099 ag 3-Methyl-1-butanol 8.790

b Methanethiol 3.400 ah 2-Methyl-1-butanol 8.801

c Trimethylamine 3.426 aj (E)-2-Pentenal 9.331

d Ethanol 3.527 ak 1-Pentanol 9.565

e Pentane 3.699 al (Z)-2-Penten-1-ol 9.640

f Propanal 3.795 am 2-Hexanone 10.154

g Dimethyl sulfide 4.019 an Hexanal 10.484

h Methylene chloride 4.195 ao 4,4-Dimethyl-1,3-dioxane 11.484

i Carbon disulfide 4.253 ap (E)-2-Hexenal 11.882

j 2,3-Butanedione 5.009 aq 1-Hexanol 12.122

k Butanal 5.075 ar 3-Heptanone 12.497

l 2-Butanone 5.110 as (Z)-4-Heptenal 12.817

m 2-Butanol 5.370 at Heptanal 12.858

o Ethyl acetate 5.442 au Cyclohexanone 12.968

p Acetic acid, ethyl ester 5.639 av 1-Heptanol 14.018

q 2-Methyl-1-propanol 5.793 aw 3,5,5-Trimethyl-2-hexene 14.106

r 3-Methyl butanal 6.525 ax Benzaldehyde 14.142

s 1-Butanol 6.695 ay 1-Octen-3-ol 14.192

t 2-Methyl butanal 6.725 az 2,3-Octanedione 14.248

v 1-Penten-3-ol 7.197 ba 2,4-Heptadienal 14.534

w 1-Penten-3-one 7.298 bb Octanal 14.567

x Heptane 7.453 bc (E,E)-2,4-Heptadienal 14.769

y 2-Ethyl furan 7.577 bd 2-Ethyl-1-hexanol 14.818

z 3-Pentanone 7.598 be Limonene 14.856

aa 2,3-Pentanedione 7.610 bf 1-Octanol 15.052

ab Pentanal 7.643 bg 3,5-Octadien-2-one 15.459

ac 3-Pentanol 7.892 bh Nonanal 15.852

ae 3-Hydroxy-2-butanone 8.503

Figure 1. Comparison of freshness indices according to sampling day.

RESULTS AND DISCUSSIONAnalysis of freshness or spoilage indices by sampling dayThe study consists of 35 measurements taken during five periods,D1 (day 1), D2 (day 2), D3 (day 3), D4 (day 4) and D7 (day 7), ata rate of seven measurements per period and some 60 volatilecompounds observed per sample (Table 1). Figures 1 and 2 showthe respective freshness indices and QIM scores associated withfreshness and spoilage of the analysed whiting catches.

Figure 2. Comparison of QIM scores according to sampling day.

They illustrate both the average trend and the variability of thetwo indices (freshness and QIM) by sampling day. We note thatthese two freshness and/or spoilage indicators tend to rise oversampling time.

The daily variability of the two indicators is illustrated by theheight of each box plot. This variation is greatest on D3 and D4for the freshness index and on D3 and D7 for the QIM score.Values outside the boxes (outliers) testify to the atypical characterof some indicator values (abnormally high or low). Two fish have


25

71


Figure 3. Chart of individual freshness indices: , mean; - - - - ,upper and lower control limits.

freshness indices that are relatively far from the average on D1and D3, while one fish has a QIM score that is relatively far fromthe average on D1.

We note for example that the mean value of the freshness indexincreases from 1.393 on D1 to 14.000 on D7 with high variabilityon D3 (standard deviation of 1.1603). As for the QIM score, it varieson average from 2.976 on D1 to 18.333 on D7 with fairly highvariability on D3 (standard deviation of 2.0189).

Concerning the freshness index, the estimate of means andstandard deviations clearly shows a relatively significant differencebetween D2 and D3 on the one hand and between D4 and D7on the other hand. It thus appears that D3 represent a breakingpoint between high and low quality of the analysed fish samples’freshness.

Freshness therefore decreases between D2 and D3. Then thereis a transition phase from D3 to D4, leading to a confirmed spoilagezone starting on D4.

This grouped analysis of freshness indices is combined witha descriptive analysis of each of the two indicators’ individualvalues in Figs 3 and 4. These two graphs explain the upward trendobserved above and specify the quite linear nature of this trend inrelation to the sampling days. They also highlight a transition zonedelimited by the line located between the two broken lines (upperand lower control limits) and marked by a first change in valuesof the two indicators between D2 and D3 and a second changebetween D4 and D7. These jumps appear clearly evident for theQIM indicator. Statistical tests for small samples were applied andgave a significance of around 10−3 (Fig. 4): the mean value on D2 issignificantly different from that on D3 with an error level of 5% forfreshness index and QIM score. D4 presents the same differenceas D7 for the two indicators.

Classification of fish using PCAThe results of the PCA show that the first two components, alinear combination of volatile compounds, account for about 65%(Tables 2 and 3) of the total information contained in the data

Figure 4. Chart of individual QIM scores: , mean; - - - - , upper andlower control limits.

table. The first principal component with an eigenvalue of 28.477(Table 2, column VP) accounts for 54.75% (projected inertia) of the65%, while the second principal component with an eigenvalue of5.535 (Table 3, column VP) accounts for only 10.64%.

The circle of correlation (Fig. 5: projection of variables on theplane formed by the first two principal components) clearlyillustrates two major volatile compound trends. The first/secondtrend involves a group of volatile compounds that are positivelyprojected onto the first/second principal component. Fromthese two trends we extracted the most characteristic volatilecompounds (Tables 2 and 3) in order of their importance in eachof the two principal components. The COOR column correspondsto the coordinate of the projection of the volatile compound oneach factor, while the columns QLT and CRT respectively representthe quality of the representation and the relative contribution ofthis compound to the formation of the principal trend.

We note therefore that 30 volatile compounds are largelyresponsible for the formation of axis 1 (Fig. 5). These 30 volatilecompounds account for more than 80% of the 54.75% of inertiaprojected on the first component and range from 1-pentanol (ak)to (E)-2-hexenal (ap) and methylene chloride (h) and from 2-ethyl-1-hexanol (bd) to octanal (bb), heptanal (at) and 1-octen-3-ol(ay). As for the second principal component, it largely compriseseight volatile compounds, ethanol (d), 3-methyl-1-butanol (ag),2,3-butanedione (j), 2-methyl-1-butanol (ah), 3-methyl butanal(r), 2-methyl-1-propanol (q), limonene (be) and ethyl acetate (o),which make up nearly 63% of the 10.64% of inertia projected onthis second component.

Classification description and interpretationAfter the PCA an AHC was conducted using SPAD 7 and the firsttwo principal components to define categories.

Table 4 presents a distribution into three categories and givesthe composition of each of them as well as the fish with profilesthat are closest to the category’s centre of gravity (distance to thecentre of the category’s column).


25

72


Table 2. Variables of interpretation from first principal component

Variable name COOR QLT VP CRT

ak 0.981 0.963 28.477 0.034

ap 0.976 0.952 28.477 0.033

ba 0.969 0.939 28.477 0.033

bg 0.969 0.938 28.477 0.033

av 0.964 0.930 28.477 0.033

aj 0.964 0.929 28.477 0.033

al 0.963 0.928 28.477 0.033

as 0.963 0.928 28.477 0.033

w 0.956 0.914 28.477 0.032

aw 0.949 0.900 28.477 0.032

bb 0.946 0.894 28.477 0.031

at 0.945 0.893 28.477 0.031

an 0.940 0.883 28.477 0.031

bc 0.937 0.878 28.477 0.031

ab 0.935 0.874 28.477 0.031

ao 0.933 0.870 28.477 0.031

az 0.931 0.866 28.477 0.030

ay 0.928 0.861 28.477 0.030

v 0.927 0.859 28.477 0.030

am 0.927 0.859 28.477 0.030

ax 0.924 0.854 28.477 0.030

aa 0.905 0.819 28.477 0.029

k 0.901 0.811 28.477 0.028

bh 0.895 0.802 28.477 0.028

aq 0.891 0.794 28.477 0.028

b 0.879 0.772 28.477 0.027

f 0.870 0.757 28.477 0.027

x 0.854 0.729 28.477 0.026

bd 0.839 0.703 28.477 0.025

h 0.827 0.684 28.477 0.024

% total contribution 0.905

Table 3. Variables of interpretation from second principalcomponent

Variable name COOR QLT VP CRT

d 0.779 0.607 5.535 0.110

ag 0.769 0.591 5.535 0.107

j 0.747 0.559 5.535 0.101

ah 0.741 0.549 5.535 0.099

r 0.695 0.483 5.535 0.087

q 0.644 0.415 5.535 0.075

be 0.635 0.403 5.535 0.073

o 0.634 0.402 5.535 0.073

% total contribution 0.724

• Category 1/3 consists of 17 fish from the two catches of whitingwith samples taken on D1, D2 and D3.

• Category 2/3 consists of 7 fish from the two catches withsamples taken on D3 and D4.

• Category 3/3 consists of 11 fish from the two catches. It includessome of the samples taken on D4 and all samples taken on D7.

Figure 6 illustrates the simultaneous projection of the centres ofthe three categories and the fish. The plan of projection comprising

Figure 5. Circle of correlations in principal plane.

the first two principal components represents more than 75% ofthe total information. The diameter of each category (each class)allows us to assess the quality of representation and importanceof the category in the projection plan. The large the diameter, themore the category is well represented: better quality for category1/3 than for categories 3/3 and 2/3. The analysis shows thatthe centre of category 1/3 projects negatively onto the first twoprincipal components. The squared cosine of the angle formed bythe category 1/3 centre of gravity and the first principal componentis 0.794 with a significance of about 10−4. As a result, if we use anerror level of 5% (five out of 100 chances that the conclusion will befalse), the correlation between this category and the first principalcomponent is statistically significant. The results also show thatcategory 1/3 is statistically significantly correlated with the secondprincipal component with a correlation of 0.191.

Similarly, category 2/3 is significantly positively correlated withthe second principal component with a correlation of 0.517 anda virtually null significance. Category 3/3 is significantly positivelycorrelated with the first principal component with a correlation of0.991 and a virtually null significance.

The diagrams in Figs 7 and 8 make it possible to quantify thedegree of freshness and/or spoilage of each of the three fishcategories. Overall, category 1/3 has the lowest values of freshnessindex and QIM score. The freshness quality of the fish in thiscategory is excellent. In category 2/3 the freshness indices andQIM scores are higher than in category 1/3 and overlap category3/3, with significant dispersion within this category only for thefreshness index. It appears to be an intermediate category betweenfreshness and spoilage. Category 3/3 shows higher freshness indexvalues with less dispersion. This category is synonymous withspoilage (poor freshness quality of fish).

To characterise each of the three fish categories with volatilecompounds, we compared the compound’s mean category valuewith its mean value calculated from all analysed fish. The selectedcompounds were then compared with those from the PCA.We therefore propose three typologies of volatile compounds:compounds present in the category on average and highlysignificant, those present and not highly significant and thosecompletely and significantly absent on average (Table 4).

In category 3/3, which consists of fish at an advanced stageof spoilage (Figs 7 and 8), the volatile compounds whose averagepresence is significant are those that contribute strongly to the


25

73


Table 4. Characterisation of categories by volatile compounds

Category

Distance tothe centre ofthe category Name

Highly significantaverage presence

Slightly significantaverage presence

Highly significantaverage absence

3/3 5.932 L01P03D07 f, an, v, ab, aa, w, h, at, aj, az,al, ak, ay, aq, b, aw, as, bb,k, av, ap, ao, bc, ax, bd, bh,ba, x, bg, am

r

6.780 L01P02D07

10.452 L02P01D04

13.765 L01P01D07

18.874 L02P01D07

22.325 L02P02D04

23.574 L01P04D04

39.567 L02P02D07

39.972 L02P03D04

43.094 L02P03D07

56.905 L01P04D07

2/3 9.989 L02P03D03 d, ag, r, ah, q, j, o, be w, ao am

12.390 L01P03D04

20.353 L01P02D04

38.469 L01P02D03

49.812 L01P01D03

52.347 L02P02D03

57.513 L01P01D04

1/3 0.680 L02P01D01 f, an, v, ab, aa, w, at, aj, ay, ak,az, al, k, bb, aw, aq, ao, b,av, ap, bd, bh, bc, ax, bg d,q, o, j

h, as, x, ba, am,

1.532 L01P03D02

2.169 L01P04D01

2.366 L02P01D02

3.336 L01P02D01

3.376 L01P04D02

3.427 L01P02D02

3.547 L02P01D03

4.516 L01P03D01

4.550 L02P03D01

6.350 L02P02D01

7.754 L01P04D03

10.033 L02P02D02

10.781 L01P01D02

12.540 L01P01D01

14.050 L01P01D03

20.665 L02P03D02

formation of the first trend in the PCA (Table 2). The majorityof these compounds have a category mean that is significantlyhigher than their general mean (mean calculated using all samplesof analysed fish). The most characteristic compounds on averageare propanal (f), hexanal (an), 1-penten-3-ol (v), pentanal (ab),2,3-pentanedione (aa), 1-penten-3-one (w), methylene chloride(h), heptanal (at), (E)-2-pentenal (aj), 2,3-octanedione (az), (Z)-2-penten-1-ol (al), 1-pentanol (ak), 1-octen-3-ol (ay), 1-hexanol (aq),methanethiol (b), 3,5,5-trimethyl-2-hexene (aw), (Z)-4-heptenal(as), octanal (bb), butanal (k), 1-heptanol (av), (E)-2-hexenal(ap), 4,4-dimethyl-1,3-dioxane (ao), (E,E)-2,4-heptadienal (bc),benzaldehyde (ax), 2-ethyl-1-hexanol (bd), nonanal (bh), 2,4-heptadienal (ba), heptane (x), 3,5-octadien-2-one (bg) and 2-hexanone (am). We also noted the average presence of a volatile

compound that contributes to the second principal component:3-methyl butanal (r). This compound has a mean in the categorythat is lower than its general mean, but the difference betweenthe two means is not statistically significant.

Category 2/3, comprising seven fish, can be qualified as anintermediate category because it contains neither fish with anexcellent state of freshness nor fish with significant spoilage(fish with a poor state of freshness). This transitional categorybetween freshness and spoilage is mainly characterised by thecompounds ethanol (d), 3-methyl-1-butanol (ag), 2,3-butanedione(j), 2-methyl-1-butanol (ah), 3-methyl butanal (r), 2-methyl-1-propanol (q), limonene (be) and ethyl acetate (o) whose meanin the category is significantly higher than their general mean.These volatile compounds are also those involved in the second


25

74


Figure 6. Projections of categories and fish on first two factors.

Figure 7. Characterisation of categories by freshness indices.

Figure 8. Characterisation of categories by QIM scores.

main trend determined by the PCA. This category 2/3 alsocontains volatile compounds related to the formation of the PCA’sfirst principal component: 1-penten-3-one (w), 4,4-dimethyl-1,3-dioxane (ao) and 2-hexanone (am). These latter three volatilecompounds contribute to the formation of the PCA’s first principalcomponent and the characterisation of category 3/3. Unlike the

compounds that characterise this second category, these threevolatile compounds have a mean value in category 2/3 that isslightly lower or even null compared with their general mean.

Category 1/3 consists of fish with an excellent state of freshness.The volatile compounds that characterise it generally have acategory mean that is significantly smaller than their general mean.These compounds (d, q, o, j) are linked to both the first and secondcomponents in the PCA. The most prominent volatile compoundsin category 1/3 can be divided into two groups: a group ofcompounds (f, an, v, ab, aa, w, at, aj, ay, ak, az, al, k, bb, aw, aq, ao,d, q, o, j, b, av, ap, bd, bh, bc, ax, bg) whose mean in the categoryis non-null but significantly smaller than the mean in the sample(‘very significant average presence’) and a group of compounds(h, as, x, ba, am) whose mean in the category is totally null (‘verysignificant average absence’). This latter group of compounds canalso be found in category 3/3 but with the completely oppositebehaviour. The fish L02P01D01 has the profile that is closest to thecentre of the category. The compounds that characterise thesetwo axes (Fig. 6) may be compounds that represent fish freshnessand spoilage.

To sum up the three categories, the first category 1/3corresponds to a category with an excellent freshness index,which characterises excellent product quality. For category 2/3,this quality decreases and the whiting starts to deteriorate. Forcategory 3/3, which is the worst category from a freshness/qualityviewpoint, spoilage of whiting continues but the product remainsacceptable for consumption (freshness index <18, Fig. 3). We mayconclude that axis 1 represents the freshness of whiting and axis2 characterises spoilage.

CONCLUSIONSThe volatile compounds propanal (f), hexanal (an), 1-penten-3-ol (v), pentanal (ab), 2,3-pentanedione (aa), 1-penten-3-one(w), heptanal (at), (E)-2-pentenal (aj), 2,3-octanedione (az), (Z)-2-penten-1-ol (al), 1-octen-3-ol (ay), 1-pentanol (ak), butanal(k), octanal (bb), 3,5,5-trimethyl-2-hexene (aw), 1-hexanol (aq),4,4-dimethyl-1,3-dioxane (ao), methanethiol (b), 1-heptanol (av),(E)-2-hexenal (ap), 2-ethyl-1-hexanol (bd), nonanal (bh), (E,E)-2,4-heptadienal (bc), benzaldehyde (ax) and 3,5-octadien-2-one (bg)are found in both category 1/3 (category with excellent freshnessquality; negative test values) and category 3/3 (considered a


25

75


spoilage category; positive test values). Thus they are relevantcompounds because they are found throughout the spoilageprocess but with a difference in quantity, which shows analteration in the whiting’s quality. These volatile compounds maybe considered representative markers of whiting and markers offuture freshness.

The compounds methylene chloride (h), (Z)-4-heptenal (as),heptane (x), 2,4-heptadienal (ba) and 2-hexanone (am) aresignificantly absent from category 1/3 and highly present incategory 3/3. They are therefore relevant in the sense that theycan be considered spoilage markers. A very fresh product willnot contain these compounds; conversely, they will be found in adeteriorated product.

Other compounds present in category 2/3 could be repre-sentative of spoilage in whiting and can be considered to bespoilage markers: ethanol (d), 3-methyl-1-butanol (ag), 3-methylbutanal (r), 2-methyl-1-butanol (ah), 2-methyl-1-propanol (q), 2,3-butanedione (j), ethyl acetate (o) and limonene (be). 2-Ethyl furan(y) is worth exploring for use as a freshness marker, which does notcharacterise axes 1 and 2 but is only significant for category 1/3.

This method is well adapted for recovery of the volatile com-pounds, because our results present a large number of compoundsdescribed in previous studies:3,4,15,17,23,24,26 compounds character-ising freshness, i.e. hexanal (an), heptanal (at), 1-penten-3-ol (v),(Z)-2-penten-1-ol (al), 1-octen-3-ol (ay), octanal (bb), (E)-2-pentenal(aj) and (E)-2-hexenal, or spoilage, i.e. 3-methyl-1-butanol (ag),ethanol (d), 2-methyl-1-propanol (q), 3-methyl butanal (r), ethylacetate (o), 2,3-butanedione (j) and 3-hydroxy-2-butanone (ae).SPME/GC/MS may be a useful method at the same time to identifythe volatile compounds then to characterise freshness. Comple-mentary studies could allow applications to other species.

ACKNOWLEDGEMENTSWe thank the Nord-Pas de Calais region (France) and FranceAgrimer for financial assistance.

REFERENCES1 Olafsdottir G and Fleurence J, Evaluation of fish freshness using

volatile compounds – classification of volatile compounds in fish, inMethods to Determine the Freshness of Fish in Research and Industry,Proceedings of the Final Meeting of the Concerted Action ‘Evaluationof Fish Freshness’ AIR3CT94 2283, ed by Olafsdottir G. Internationalinstitute of refrigeration, Paris, France, pp. 51–69 (1997).

2 Alasalvar C, Aishima T and Quantick PC, Dynamic headspace analysisof volatile aroma products in fresh and deteriorated mackerel(Scomber scombrus). Food Sci Technol Int 1:125–127 (1995).

3 Aro T, Tahovonen R, Koskinen L and Kallio H, Volatile compoundsof Baltic herring analysed by dynamic headspace sampling–gaschromatography–mass spectrometry. Eur Food Res Technol216:483–488 (2003).

4 Olafsdottir G, Jonsdottir R, Lauzon HL, Luten J and Kristbergsson K,Characterization of volatile compounds in chilled cod (Gadusmorhua) fillets by gas chromatography and detection of qualityindicators by an electronic nose. J Agric Food Chem 53:10140–10147(2005).

5 Shewan JM, The bacteriology of fresh and spoiling fish and somerelated chemical changes. Recent Adv Food Sci 1:167–193 (1962).

6 Van Spreekens KJA, Characterization of some fish and shrimp spoilingbacteria. Antonie van Leeuwenhoek: Int J Gen Mol Microbiol43:283–303 (1977).

7 Olafsdottir G, Li X, Lauzon HL and Jonsdottir R, Precision andapplication of electronic nose for freshness monitoring of whole

redfish (Sebastes marinus) stored in ice and modified atmospherebulk storage. J Aquat Food Prod Technol 11:229–249 (2002).

8 Miller III A, Scanlan RA, Lee JS and Libbey LM, Identification of thevolatile compounds produced in sterile fish muscle (Sebastesmelanops) by Pseudomonas fragi. J Appl Microbiol 25:952–955(1973).

9 Miller III A, Scanlan RA, Lee JS, Libbey LM and Morgan ME, Volatilecompounds produced in sterile fish muscle (Sebastes melanops) byPseudomonas perolens. J Appl Microbiol 25:257–261 (1973).

10 Lindsay RC, Josephson DB and Olafsdottir G, Chemical andbiochemical indices for assessing the quality of fish packagedin controlled atmospheres, in Seafood Quality Determination,Proceedings of an International Symposium (University of Alaska SeaGrant Program, Anchorage, Alaska), ed. by Kramer DE and Liston J.Elsevier Science, Amsterdam, pp. 221–234 (1986).

11 Josephson DB and Lindsay RC, Enzymic generation of volatile aromacompounds from fresh fish. ACS Symp Ser 317:201–219 (1986).

12 Durnford E and Shahidi F, Flavour of fish meat, in Flavor of Meat, MeatProducts and Seafoods, ed. by Shahidi F. Springer - Verlag, Berlin,pp. 131–158 (1998).

13 Kawai T, Fish flavor. Crit Rev Food Sci 36:257–298 (1996).14 Alasalvar C, Quantick PC and Grigor JM, Aroma compounds of fresh

and stored mackerel (Scomber scombrus). ACS Symp Ser 674:39–54(1997).

15 Alasalvar C, Taylor KDA and Shahidi F, Comparison of volatiles ofcultured and wild sea bream (Sparus aurata) during storage inice by dynamic headspace analysis/gas chromatography–massspectrometry. J Agric Food Chem 53:2616–2622 (2005).

16 Jonsdottir R, Bragadottir M and Olafsdottir G, The role of volatilecompounds in odor development during hemoglobin-mediatedoxidation of cod muscle membrane lipids. J Aquat Food Prod Technol16:67–86 (2007).

17 Prost C, Hallier A, Cardinal M, Serot T and Courcoux P, Effect of storagetime on raw sardine (Sardina pilchardus) flavor and aroma quality.J Food Sci 69:198–204 (2004).

18 Duflos G, Cornu M, Coin VM, Antinelli JF and Malle P, Determinationof volatile compounds to characterize fish spoilage using HS/MSand SPME–GC/MS analysis. J Agric Food Chem 86:600–611 (2006).

19 Bene A, Fornage A, Luisier J, Pichler P and Villettaz J, A new methodfor the rapid determination of volatile substances: the SPME-directmethod. Part I: Apparatus and working conditions. Sensors ActuatorsB 72:184–187 (2001).

20 Bene A, Hayman A, Reynard E, Luisier J and Villettaz J, A new methodfor the rapid determination of volatile substances: the SPME-directmethod. Part II: Determination of the freshness of fish. SensorsActuators B 72:204–207 (2001).

21 Baek HH and Cadwallader KR, Volatile compounds in flavorconcentrates produced from crayfish-processing by-products withand without protease treatement. J Agric Food Chem 44:3262–3267(1996).

22 Guillen MD and Errecalde MC, Volatile components of raw andsmoked black bream (Brama raii) and rainbow trout (Oncorhynchusmykiss) studied by means of solid phase microextraction and gaschromatography/mass spectrometry. J Sci Food Agric 82:945–952(2002).

23 Mansur MA, Bhadra A, Takamura H and Matoba T, Volatile flavorcompounds of some sea fish and prawn species. Fish Sci 69:864–866(2003).

24 Wierda RL, Fletcher G, Xu L and Dufour J-P, Analysis of volatilecompounds as spoilage indicators in fresh king salmon(Oncorhynchus tshawytscha) during storage using SPME-GC–MS.J Agric Food Chem 54:8480–8490 (2006).

25 Edirisinghe RKB, Graffham AJ and Taylor SJ, Characterisation of thevolatiles of yellowfin tuna (Thunnus albacares) during storage bysolid phase microextraction and GC–MS and their relationship tofish quality parameters. Int J Food Sci Technol 42:1139–1147 (2007).

26 Iglesias J and Medina I, Solid-phase microextraction method for thedetermination of volatile compounds associated to oxidation offish muscle. J Chromatogr A 1192(1):9–16 (2008).

27 Duflos G, Coin VM, Moine F and Malle P, Determination of volatilecompounds in whiting using SPME GC–MS. J Chromatogr Sci43:304–312 (2005).


25

76

Research ArticleReceived: 23 July 2009 Revised: 8 July 2010 Accepted: 12 July 2010 Published online in Wiley Online Library: 5 August 2010


Release of phenolic acids from defatted ricebran by subcritical water treatment†

Cynthia Fabian, Ngoc Yen Tran- Thi, Novy S Kasim and Yi-Hsu Ju∗

Abstract

BACKGROUND: Oil production from rice bran, an undervalued by-product of rice milling, produces defatted rice bran (DRB)as a waste material. Although it is considered a less valuable product, DRB still contains useful substances such as phenoliccompounds with antioxidant, UV-B-protecting and anti-tumour activities. In this study the phenolic acids in DRB were extractedwith subcritical water at temperatures of 125, 150, 175 and 200 ◦C.

RESULTS: Analysis of total phenolics using Folin–Ciocalteu reagent showed about 2–20 g gallic acid equivalent kg−1 bran inthe extracts. High-performance liquid chromatography analysis showed low contents of phenolic acids (about 0.4–2 g kg−1

bran). Ferulic, p-coumaric, gallic and caffeic acids were the major phenolic acids identified in the extracts. Thermal analysis ofthe phenolic acids was also done. The thermogravimetric curves showed that p-coumaric, caffeic and ferulic acids started todecompose at about 170 ◦C, while gallic acid did not start to decompose until about 200 ◦C.

CONCLUSION: Subcritical water can be used to hydrolyse rice bran and release phenolic compounds, but the high temperaturesused in the extraction can also cause the decomposition of phenolic acids.c© 2010 Society of Chemical Industry

Supporting information may be found in the online version of this article.

Keywords: antioxidant; phenolic acids; rice bran; subcritical water; thermogravimetry

INTRODUCTIONLipid oxidation decreases the quality and nutritional value offoods.1,2 To remedy this, antioxidants that can prevent or retardthe oxidation of fats and oils are introduced. Increasing evidencehas shown that intake of antioxidants can also lower the riskof degenerative diseases such as cancer, cardiovascular diseases,diabetes, arthritis, cataract formation and aging.3 – 5

Interest in antioxidants is being directed towards the identifica-tion and extraction of natural compounds because of the growingconcern about the safety of synthetic ingredients contained invarious foods.6 Cereals, including rice (Oryza sativa L.), contain awide range of phenolic compounds and are claimed to be a goodsource of natural antioxidants.4,7 – 10 Some of these compounds arepredominantly found in grains and are not present significantlyin fruits and vegetables.11 These phenolic acids are known tocontribute to the antioxidant potential of cereal grains.12 – 15

Zhou et al.16 found high levels of phenolic acids (ferulic andp-coumaric acids) in brown rice but lower levels in milled rice.This indicates that phenolic acids are concentrated in rice bran.However, there are few studies reporting the profiles of phenolicacids in rice bran and defatted rice bran (DRB).

Rice bran is a by-product of the rice-milling industry and isgenerally used as an animal feed.7,17 – 19 In some countries, oilis extracted from rice bran for food application, which producesDRB as a waste material. Renuka Devi and Arumughan4 studiedthe antioxidant efficacy of DRB extract and its phytochemicalconstituents, which include ferulic acid. Since phenolic acids arepolar compounds, most are retained in the bran after extraction

of lipids. They can be utilised if extracted and separated fromthe DRB.

Extraction of phenolic compounds from DRB would facilitatethe production of value-added products. Also, the extraction ofbioactive components from biodiesel residue provides options tolower the production cost of biodiesel.20 The utilisation of DRB,a by-product of biodiesel production if rice bran is used as theraw material, has the potential to reduce the cost of biodieselproduction.

In this study, subcritical water was employed to extractphenolic compounds from DRB. Subcritical water extraction, anew, environmentally friendly technique, has been considered asan alternative for the isolation of antioxidant constituents.2,21,22

Since only water is employed for extraction, the process avoids theuse and disposal of large volumes of flammable or toxic solvents.Subcritical water is achieved at temperatures between 100 and374 ◦C (the critical point of water is at 374 ◦C and 22 MPa) andat a pressure high enough to keep the water in a liquid state.23

At subcritical condition the ionic product of water increases suchthat the water becomes a rich source of H+ and OH− ions suitable

∗ Correspondence to: Yi-Hsu Ju, National Taiwan University of Science andTechnology, 43 Keelung Road Sec. 4, Taipei 106-07, Taiwan.E-mail: [email protected]

† Part of this paper was presented at the Conference of Food Engineering,Columbus, Ohio, USA, 5–8 April 2009.

National Taiwan University of Science and Technology, 43 Keelung Road Sec.4, Taipei 106-07, Taiwan


25

77

Phenolic acids from defatted rice bran www.soci.org

for hydrolysis reactions.24 Also, the dielectric constant of waterdecreases as it approaches the critical point, so a wide range oforganic compounds can be solubilised and extracted.25

Previous studies by Wiboonsirikul et al.,18,19 Hata et al.26 andSereewatthanawut et al.27 on subcritical extraction from rice branfocused on proteins, carbohydrates and antioxidant capacity ofthe subcritical extracts. The identification of antioxidants suchas phenolic acids in the rice bran extracts was not performed.In the present study, high-performance liquid chromatography(HPLC) was utilised to identify phenolic acids in the extracts.Since high temperatures were employed during extraction,thermogravimetric analyses of the identified phenolic acids werealso carried out to determine their decomposition temperatures.

MATERIALS AND METHODSMaterialsRice bran fresh from milling was purchased from a local rice millin Taoyuan County (Taiwan). The bran is not specifically fromone variety of rice but is a mixture of rice harvested in northernTaiwan. Bran collected from the mill was stored at −60 ◦C beforeuse. Defatting of rice bran was done using hexane in a Soxhletextractor at 60 ◦C for 4 h.

Phenolic acid standards (gallic (monohydrate), caffeic, p-coumaric and ferulic acids) and the Folin–Ciocalteu (FC) reagentused for total phenolic analysis were purchased from Sigma Aldrich(St Louis, MO, USA). HPLC-grade methanol and acetonitrile wereused in the analyses. Other chemicals used were of analyticalgrade.

Extraction with subcritical waterDRB was mixed with distilled water to a content of about 400 g L−1

solids. About 70 mL of this mixture was charged into a 90 mLstainless steel batch reactor (Jia Chan Company, Tainan, Taiwan)equipped with a thermocouple and a pressure gauge. The pressurewas increased to 20 bar using nitrogen gas. The reactor was thenheated to the desired temperature (125, 150, 175 or 200 ◦C), whichtook about 30 min, and held at that temperature for 5 min. Thereactor was then cooled to room temperature and its contentwas centrifuged. The supernatant was collected, freeze-dried andanalysed for phenolic acid content. All extraction experiments andsubsequent analyses were carried out at least twice.

Analysis of phenolic acidsThe supernatant was first analysed for total phenolic compoundsextracted using the methods of Wiboonsirikul et al.18 About100 µL of the diluted subcritical water extract (10× dilution) wascombined with 400 µL of FC reagent and 1 mL of 75 g L−1 sodiumcarbonate solution. Distilled water was added to the mixture toa total volume of 5 mL and the solution was kept in the darkat ambient temperature for 2 h to complete the reaction. Totalphenolic content was determined by measuring the absorbanceat 765 nm using a Jasco UV-V 550 UV–visible spectrophotometer(Jasco, Easton, MD, USA). Gallic acid was used as a standard, andresults were calculated as g gallic acid equivalent kg−1 bran.

Individual phenolic acids in the DRB extract were identifiedusing an HPLC UV–visible detector. Since many compounds werepresent in the subcritical extract, phenolic acids were separatedby solvent extraction from the freeze-dried supernatant. About0.1 g of the freeze-dried supernatant was mixed with 4 mL of800 mL L−1 methanol at 30 ◦C for 2 h. The mixture was centrifuged

and separated to obtain the methanolic extract. The extract wasthen passed through a 0.45 µm filter and analysed by HPLC.

The HPLC analysis of phenolic acids employed in this study wasbased on the method of Zhou et al.16 with modification. A 20 µLaliquot of the methanolic extract was separated using a JascoHPLC system. Peaks were detected with a Jasco multiwavelengthUV–visible detector operated at 280 and 325 nm. The absorbanceat 325 nm is optimal for all hydroxycinnamic acids, while 280 nmis the most accepted compromise for all phenols, includinghydroxybenzoic and hydroxycinnamic acids. Separations wereachieved on a 5 µm Phenomenex Luna C18 column (240 mm× 4.6 mm; Phenomenex, Torrance, CA, USA). Gradient elutionwas performed using solvent A (water/acetic acid, 100 : 1 v/v)and solvent B (methanol/acetonitrile/acetic acid, 95 : 5:1 v/v/v)in the following sequence: 0–2 min, 5% B; 2–10 min, 5–25%B; 10–20 min, 25–40% B; 20–30 min, 40–50% B; 30–40 min,50–100% B; 40–45 min, 100% B; 45–55 min, 100–5% B. Thesolvent flow rate was fixed at 1 mL min−1. Peak identities wereconfirmed from retention data and by spiking extracts withauthentic standards.

HPLC data were processed using ChromPass Version 1.8.6.1(Jasco, Tokyo, Japan). Quantification was accomplished viaexternal standard calibration curves plotted using pure standardsof the four phenolic acids (Supporting information). Regressionanalysis of each calibration curve was also carried out toobtain a correlation coefficient between peak area and standardconcentration. These correlation coefficients were later used inthe quantification of phenolic acid content in samples.

Thermogravimetric analysisA Perkin Elmer Diamond TG/DTA instrument (PerkinElmer, Shelton,CT, USA) was used for thermal stability studies. Approximately6 mg of phenolic acid was placed on a platinum pan. The samplewas then heated from 30 to 950 ◦C at 10 ◦C min−1 to determinethe temperature at which decomposition occurred. During theentire run, air at atmospheric pressure was allowed to flow at20 mL min−1 through the system containing the sample.

RESULTS AND DISCUSSIONTotal phenolics in subcritical extracts from DRBThe phenolic acids in DRB were extracted by subcritical water andanalysed using two different methods, i.e. FC reagent and HPLC.As can be seen in Table 1, the FC method showed an increasein the amount of phenolic compounds in the subcritical extractwith increasing temperature. At 125 ◦C the amount of phenolicswas about 2 g kg−1 bran, but this increased significantly to about20 g kg−1 bran at 200 ◦C. This trend is similar to the findings ofWiboonsirikul et al.,18 with differences in specific values beingattributable to the different sources of rice bran. The increase inphenolic content of the extract at high temperatures is probablydue to the release of phenolic substances, interfering substancesor both by subcritical water treatment. At subcritical condition,hydrolysis of rice bran is favoured, because the ionic product ofwater increases such that there are more H+ and OH− ions availableto catalyse the hydrolysis.25 Hydrolysis of the bran resulted inan increase in released phenolic acids, because hydroxycinnamicacids such as ferulic, sinapic, caffeic and p-coumaric acids are foundnot only as soluble forms in the cytoplasm but also covalentlyattached to the plant cell wall.28

Quantification of the four major phenolic acids found in therice bran extract (gallic, caffeic, ferulic and p-coumaric acids) by


25

78

www.soci.org C Fabian et al.

Table 1. Phenolic content of subcritical water extracts of defattedrice bran

Temperature(◦C)

Folin–Ciocalteu method(g gallic acid equivalent

kg−1 bran)

HPLC analysis(g total phenolic acids

kg−1 bran)

125 2.011 ± 0.056 0.528 ± 0.006

150 3.552 ± 0.046 1.794 ± 0.045

175 7.833 ± 0.178 2.125 ± 0.077

200 19.480 ± 0.148 0.380 ± 0.029

the HPLC method (based on the study by Zhou et al.16) showed adifferent trend of phenolic acid content in rice bran extract fromthat determined by the FC method. The total amount of phenolicacids found was about 0.5 g kg−1 bran at 125 ◦C, increased to about2 g kg−1 bran at 175 ◦C and then fell sharply to about 0.4 g kg−1

bran at 200 ◦C. Zhou et al.16 extracted about 0.5 and 0.105 g kg−1

phenolic acids from brown and milled rice respectively. Comparingthese three sets of data, there is a noticeable tendency that thephenolic acid content (about 0.4–2 g kg−1 bran) in rice branmatrix is higher than those in brown and milled rice matrices. Thusfurther study needs to be carried out to investigate the likelihoodof the phenolic acid distribution being in the following order ofmagnitude: phenolic acid content in rice bran > phenolic acidcontent in brown rice > phenolic acid content in milled rice.

The comparative analyses of all major phenolic compoundsusing FC reagent and HPLC were also studied, since the twomethods gave different trends of phenolic acid content in rice branextract. Emphasis was placed on the applicability and reliabilityof FC reagent to analyse the phenolic acid content in rice bran.Data on rice phenolic content reported by Zhou et al.,16 Bunzelet al.,29 Hudson et al.30 and Kuroda et al.31 suggest that most of thephenolic compounds in rice bran are phenolic acids. A significantdifference in the quantification results of the two methods wasobserved, which may be attributed to interfering compoundsduring the analysis using FC reagent. According to Yu,32 the useof FC reagent for the spectrophotometric analysis of phenoliccompounds has the disadvantages of low specificity for phenolsand interference of other reducing agents that can react with thereagent. Wiboonsirikul et al.18 also mentioned that the subcriticalextract contains carbohydrates, proteins and amino acids, whichare considered as interfering substances for FC analysis. Thesecompounds can reduce the reagent. FC analysis will give poorreliability in terms of accuracy and data reproduction in analysingthe phenolic acid content in rice bran; therefore this analysis maybe inappropriate for the determination of phenolic acid contentin rice bran matrix.

It was also noted that at 175 and 200 ◦C there was a pronouncedbrown discolouration of the extract, and phenolic acids were notthe major components in the methanolic extract. A high-intensitypeak with a retention time of about 10 min was observed at thesetemperatures (Fig. 1).

The brown pigment in the extract is a polymeric materialresulting from oxidative polymerisation of phenols as well aspossible formation of Maillard products. According to Somoza,33

the Maillard reaction is a series of reactions between proteins andcarbohydrates during heating. The results of early-stage Maillardreactions are ‘Amadori products’. When higher temperatures areapplied for longer times, advanced brown-pigmented Maillardreaction products termed ‘melanoidins’ are formed. Adams et al.34

stated that other food constituents were incorporated in the

13

2

45

a

b

c

d

24 26 28 30

Retention time (min)

13

2

45

a

b

c

d

222018161412108642

Figure 1. HPLC chromatograms of methanolic extract from freeze-driedsubcritical water extract of defatted rice bran at (a) 125, (b) 150, (c) 175 and(d) 200 ◦C recorded by UV detection at 280 nm. Peak numbers: 1, gallicacid; 2, unknown compound; 3, caffeic acid; 4, p-coumaric acid; 5, ferulicacid.

melanoidin structure as well, such as lipid oxidation productsin tomato melanoidins and phenolic compounds in coffeemelanoidins. Wiboonsirikul et al.18 confirmed the occurrence ofan advanced Maillard reaction when rice bran was subjected tosubcritical water at a temperature above 150 ◦C.

From Fig. 1 it can be seen that, when the temperature wasraised to 175 ◦C, the intensities of peak 4 (p-coumaric acid) andpeak 5 (ferulic acid) decreased while the intensity of peak 2increased sharply. It is highly unlikely that the latter peak is dueto a melanoidin, since melanoidins are heterogeneous polymersand will appear in an HPLC chromatogram as a broad regionof intensity. From the fact that the intensities of peaks 4 and 5decreased while the intensity of peak 2 increased, it is likely thatpeak 2 represents a product derived from p-coumaric and ferulicacids. Further investigations are needed to identify peak 2. Thechromatogram of the subcritical extract at 175 ◦C was comparedwith that of the subcritical extract at 175 ◦C treated with 2 mol L−1

NaOH for 4 h at 25 ◦C.35 It was observed that the area of peak 2decreased by 94% while the area of the ferulic acid peak increasedby about 700% (data not shown). This may be a result of thebreakdown of ester linkages of ferulic acid by alkaline hydrolysis.35

Phenolic acid profilesFrom the HPLC method based on that of Zhou et al.,16 fourphenolic acids, i.e. gallic, caffeic, p-coumaric and ferulic acids, weredetected in all extracts obtained at different subcritical extractiontemperatures. Gallic acid was observed in the rice bran extracts.During milling, some rice hull ends up in the rice bran fraction. Ricehull is abundant in gallic acid because of the presence of tannin,


25

79


which upon hydrolysis yields gallic acid.36 Hydroxycinnamic acidssuch as ferulic, p-coumaric and caffeic acids were also found inthe extracts, since these phenolic acids are constituents of the cellwall and cytoplasm of plant cells.28

Figure 2 presents the relation between the amount of eachphenolic acid extracted and the temperature of extraction.Figure 2(a) shows that, when the temperature was increasedfrom 150 to 175 ◦C, gallic acid was the rice bran phenolic acidto increase most rapidly in the extract. When the temperaturewas further increased to 200 ◦C, gallic acid was the one thatdecreased most rapidly (see also Table 1). This is because gallicacid is the most abundant of all phenolic acids in rice bran extract.Moreover, thermogravimetric analysis of the four phenolic acids(Figs 3–5) indicated that ferulic acid was the most heat-labilephenolic acid in rice bran, and it lost its weight because of itsthermal decomposition at a lower temperature.

During subcritical extraction, phenolic acids are released fromthe bran. This is probably due to several reasons, such as hydrolysisof the rice bran cell wall, a change in solvent viscosity allowingpermeability through the bran, and the behaviour of water atsubcritical condition being similar to that of alcohol in terms ofpolarity. Since a mixture of phenolic acids is present in the bran,the decreasing polarity of water allows the extraction of morephenolic acids as the temperature is increased. At 175 ◦C theextracted caffeic and p-coumaric acids have already decreased,while the extracted ferulic and gallic acids decrease only at 200 ◦C.

The decrease in extracted phenolic acids with increasingtemperature of subcritical treatment is attributable to two possiblephenomena, i.e. thermal degradation of these phenolic acidsand their involvement in the Maillard reaction. Thermal stabilitystudies of the four phenolic acids were carried out. For thethermogravimetric analysis the weight loss of each phenolic acid as

a function of temperature is presented in Figs 3–5. The derivativeof the weight loss curve is also shown in the figures to indicate thetemperature at which weight loss is apparent.

Figure 3 presents the thermogravimetric curve for gallic acidmonohydrate. The first abrupt decrease in weight of the samplewas observed at 91.7 ◦C, which corresponds to the release of waterattached to the acid. The second (207–308 ◦C), third (308–365 ◦C)and fourth (376–576 ◦C) events in the thermogravimetric curvecorrespond to the thermal degradation of gallic acid. Major weightloss was observed at 207–308 ◦C. In the thermal analysis ofgallic acid, Oridorga et al.37 noted that the peak at about 260 ◦Ccorresponds to the formation of decarboxylation products fromgallic acid. Heating of gallic acid at this temperature caused theevolution of gaseous products and the formation of pyrogallol. Atabout 308 ◦C, which corresponds to the boiling point of pyrogallol,further weight loss of the sample occurred. Lastly, heating above308 ◦C resulted in the subsequent decomposition of pyrogallol.

From Fig. 3 it can be seen that the decomposition of gallic acidstarted at about 207 ◦C. The subcritical extraction at 200 ◦C alsoshowed a decrease in the amount of gallic acid in the extract(Fig. 2). The decrease in gallic acid in the extract at 200 ◦C may beattributed to the conversion of gallic acid into its decarboxylationproducts such as pyrogallol and other gaseous products.

Similar thermogravimetric curves for caffeic and p-coumaricacids were obtained (Fig. 4). Four events of thermal decompositionoccurred, with peaks at 221, 300, 332 and 539 ◦C for caffeic acidand at 229, 360, 394 and 547 ◦C for p-coumaric acid. Thermaldecomposition started at about 172 ◦C for these two phenolicacids, and the amount extracted decreased at 175 ◦C (Fig. 2).

In the thermal stability analysis of ferulic acid it was observedthat two decomposition steps occurred (Fig. 5). The majordecomposition of ferulic acid started at 173 ◦C and peaked at

Figure 2. Concentrations of phenolic acids in subcritical water extract of defatted rice bran at different temperatures: (a) gallic acid; (b) caffeic acid;(c) p-coumaric acid; (d) ferulic acid.


25

80

www.soci.org C Fabian et al.

Figure 3. Thermogravimetric curves of gallic acid: , weight loss;- - - - , derivative of weight loss.

Figure 4. Thermogravimetric curves of (a) caffeic acid and (b) p-coumaricacid: , weight loss; - - - - , derivative of weight loss.

245 ◦C. This result agrees with those of Taniguchi et al.,38 whostated that ferulic acid starts to decompose at about 176 ◦C,and Fiddler et al.,39 who reported that the peak decompositionof ferulic acid occurs at 245 ◦C. Fiddler et al.39 stated that thefirst step in the decomposition of ferulic acid corresponds toits decarboxylation, while the second step comprises differentreactions occurring simultaneously. It is impossible to propose asingle mechanism.

Although the decomposition of ferulic acid started at 173 ◦C,in the subcritical extract, ferulic acid kept increasing up to 175 ◦C(Fig. 2). This can be expected, since ferulic acid is the major phenolic

Figure 5. Thermogravimetric curves of ferulic acid: , weight loss;- - - - , derivative of weight loss.

acid identified in rice bran cell wall, where it may be bound toarabinoxylans or crosslinked to cell wall components.28 At 175 ◦C,ferulic acid was extracted by subcritical water, but the free ferulicacid started to decompose, resulting in a lower yield. The amountof ferulic acid extracted decreased at 200 ◦C.

CONCLUSIONSSubcritical water can be used to hydrolyse rice bran and releasephenolic compounds. Phenolic acids at a level of about 2 g kg−1

rice bran can be obtained at 175 ◦C. At higher temperatures,phenolic acids may be damaged and/or Maillard reaction productsmay be formed that may incorporate the phenolic acids. Thetemperature used in the extraction must be carefully chosen so asto release the maximum amount of phenolics while simultaneouslyminimising their destruction.

ACKNOWLEDGEMENTSFinancial support from Taiwan’s Ministry of Education, AmiaEnterprise Co., Ltd and National Taiwan University of Scienceand Technology through project 94DI062 is greatly appreciated.The authors thank Dr Fred Quarnstrom for his assistance duringthe preparation of this manuscript.

Supporting informationSupporting information may be found in the online version of thisarticle.

REFERENCES1 Cosgrove J, Church D and Pryor W, The kinetics of the autoxidation of

polyunsaturated fatty acids. Lipids 22:299–304 (1987).2 Ibanez E, Kubatova A, Senorans FJ, Cavero S, Reglero G and

Hawthorne SB, Subcritical water extraction of antioxidantcompounds from rosemary plants. J Agric Food Chem 51:375–382(2002).

3 Landrault N, Poucheret P, Ravel P, Gasc F, Cros G and Teissedre P-L,Antioxidant capacities and phenolics levels of French wines fromdifferent varieties and vintages. J Agric Food Chem 49:3341–3348(2001).

4 Renuka Devi R and Arumughan C, Phytochemical characterization ofdefatted rice bran and optimization of a process for their extractionand enrichment. Bioresour Technol 98:3037–3043 (2007).


25

81


5 Wu X, Gu L, Holden J, Haytowitz DB, Gebhardt SE, Beecher G, et al,Development of a database for total antioxidant capacity in foods:a preliminary study. J Food Compos Anal 17:407–422 (2004).

6 Prior RL, Phytochemicals: mechanism of action, in Absorption andMetabolism of Anthocyanins: Potential Health Effects, ed. byMeskin M, Bidlack WR, Devies AJ, Lewis DS and Randolph RK. CRCPress, Boca Raton, FL, pp. 1–9 (2004).

7 Iqbal S, Bhanger MI and Anwar F, Antioxidant properties andcomponents of some commercially available varieties of rice branin Pakistan. Food Chem 93:265–272 (2005).

8 Krings U, El-Saharty YS, El-Zeany BA, Pabel B and Berger RG,Antioxidant activity of extracts from roasted wheat germ. FoodChem 71:91–95 (2000).

9 Renuka Devi R, Jayalekshmy A and Arumughan C, Antioxidant efficacyof phytochemical extracts from defatted rice bran in in-vitro modelemulsions. Int J Food Sci Technol 43:878–885 (2008).

10 Shin ZI, Chang YS, Kang WS and Jung SU, Antioxidant extracted fromthe defatted rice bran. US Patent 5175 (1992).

11 Wojdyło A and Oszmainski J, Comparison of the content of phenolicacid, α-tocopherol and their antioxidant activity in oat and nakedand weeded. ElectronJEnvironAgricFood Chem 6:1980–1988 (2007).

12 Adom KK, Sorrells ME and Liu RH, Phytochemicals and antioxidantactivity of milled fractions of different wheat varieties. J Agric FoodChem 53:2297–2306 (2005).

13 Maillard M-N and Berset C, Evolution of antioxidant activity duringkilning: role of insoluble bound phenolic acids of barley and malt.J Agric Food Chem 43:1789–1793 (2002).

14 Goupy P, Hugues M, Boivin P and Amiot MJ, Antioxidant compositionand activity of barley (Hordeum vulgare) and malt extracts andof isolated phenolic compounds. J Sci Food Agric 79:1625–1634(1999).

15 Pussayanawin V, Wetzel DL and Fulcher RG, Fluorescence detectionand measurement of ferulic acid in wheat milling fractions bymicroscopy and HPLC. J Agric Food Chem 36:515–520 (2002).

16 Zhou Z, Robards K, Helliwell S and Blanchard C, The distribution ofphenolic acids in rice. Food Chem 87:401–406 (2004).

17 Juliano BO, Rice: Chemistry and Technology. American Association ofCereal Chemists, St Paul, MN (1985).

18 Wiboonsirikul J, Kimura Y, Kadota M, Morita H, Tsuno T and Adachi S,Properties of extracts from defatted rice bran by its subcritical watertreatment. J Agric Food Chem 55:8759–8765 (2007).

19 Wiboonsirikul J, Kimura Y, Kanaya Y, Tsuno T and Adachi S, Productionand characterization of functional substances from a by-product ofrice bran oil and protein production by a compressed hot watertreatment. Biosci Biotechnol Biochem 72:384–392 (2008).

20 Zullaikah S, Lai C-C, Vali SR and Ju Y-H, A two-step acid-catalyzedprocess for the production of biodiesel from rice bran oil. BioresourTechnol 96:1889–1896 (2005).

21 Rodrıguez-Meizoso I, Marin FR, Herrero M, Senorans FJ, Reglero G,Cifuentes A, et al, Subcritical water extraction of nutraceuticals

with antioxidant activity from oregano. Chemical and functionalcharacterization. J Pharmaceut Biomed Anal 41:1560–1565 (2006).

22 Ju Z and Howard LR, Subcritical water and sulfured water extractionof anthocyanins and other phenolics from dried red grape skin. JFood Sci 70:S270–S276 (2005).

23 Ramos L, Kristenson EM and Brinkman UAT, Current use of pressurisedliquid extraction and subcritical water extraction in environmentalanalysis. J Chromatogr 975:3–29 (2002).

24 Patrick HR, Griffith K, Liotta CL, Eckert CA and Glaser R, Near-criticalwater: a benign medium for catalytic reactions. Ind Eng Chem Res40:6063–6067 (2001).

25 Galkin AA and Lunin VV, Subcritical and supercritical water: a universalmedium for chemical reactions. Russ Chem Rev 74:21–35 (2005).

26 Hata S, Wiboonsirikul J, Maeda A, Kimura Y and Adachi S, Extractionof defatted rice bran by subcritical water treatment. Biochem Eng J40:44–53 (2008).

27 Sereewatthanawut I, Prapintip S, Watchiraruji K, Goto M, Sasaki M andShotipruk A, Extraction of protein and amino acids from deoiled ricebran by subcritical water hydrolysis. Bioresour Technol 99:555–561(2008).

28 Faulds CB and Williamson G, The role of hydroxycinnamates in theplant cell wall. J Sci Food Agric 79:393–395 (1999).

29 Bunzel M, Allerdings E, Sinwell V, Ralph J and Steinhart H, Cell wallhydroxycinnamates in wild rice (Zizania aquatica L.) insolubledietary fibre. Eur Food Res Technol 214:482–488 (2002).

30 Hudson EA, Dinh PA, Kokubun T, Simmonds MSJ and Gescher A,Characterization of potentially chemopreventive phenols inextracts of brown rice that inhibit the growth of human breast andcolon cancer cells. Cancer Epidemiol Biomarkers Prev 9:1163–1170(2000).

31 Kuroda K-I, Suzuki A, Kato M and Imai K, Analysis of rice (Oryza sativa L.)lignin by pyrolysis–gas chromatography. J Anal Appl Pyrol 34:1–12(1995).

32 Yu L, Wheat Antioxidants. Wiley-Interscience, Hoboken, NJ (2008).33 Somoza V, Five years of research on health risks and benefits of Maillard

reaction products: an update. Mol Nutr Food Res 49:663–672 (2005).34 Adams A, Borrelli RC, Fogliano V and De Kimpe N, Thermal

degradation studies of food melanoidins. J Agric Food Chem53:4136–4142 (2005).

35 Ribereau-Gayon P, Plant Phenolics. Oliver and Boyd, Edinburgh (1972).36 Whiting D, Natural phenolic compounds 1900–2000: a bird’s eye view

of a century’s chemistry. Nat Prod Rep 18:583–606 (2001).37 Oridorga VA, Korobeinikova II and Yasnitski BG, Thermographic study

of gallic acid. Khimiya Prirodnykh Soedinenii 4:527–528 (1978).38 Taniguchi H, Nomura E, Hosoda A, Tsuno T and Maruta Y, Thermally

stable ferulic acid derivatives. US Patent 20040152912 (2005).39 Fiddler W, Parker WE, Wasserman AE and Doerr RC, Thermal

decomposition of ferulic acid. J Agric Food Chem 15:757–761 (1967).


25

82

Research ArticleReceived: 12 May 2009 Revised: 15 February 2010 Accepted: 12 July 2010 Published online in Wiley Online Library: 17 August 2010


Nutritional composition and condensed tanninconcentration changes as browse leavesbecome litterAmanda Acero,a James P Muirb∗ and Richard M Wolfeb

Abstract

BACKGROUND: The role of condensed tannins (CT) in ruminant nutrition and health makes changes in leaf litter (LL) afterabscission of interest. This study compared the effect of different drying methods of green leaves (GL) with that of naturaldrying of LL on CT, fibre, crude protein (CP) and phosphorus (P) concentrations in nine Texas browse species. Leaves harvestedbefore autumn shedding were oven-dried (OD) or freeze-dried (FD).

RESULTS: Where different (P < 0.05), extractable CT concentrations were higher while protein- and fibre-bound CTconcentrations were lower in GL-FD than in LL. Drying method changed total CT concentration in three species. Wheredifferent, fibre fraction concentrations were greater in LL than in GL, regardless of drying method. In some species, CP andP concentrations were lower in LL than in GL, but in five species they did not change (P > 0.05) from GL to LL, with CPconcentrations ranging from 63 to 151 g kg−1 in the latter.

CONCLUSION: Browse LL had high nutritive value and CT concentrations, explaining why browsing ruminants utilise this feedresource. However, changes in nutrient and CT concentrations as leaves become litter in some species mean that informationon one is not necessarily applicable to the other.c© 2010 Society of Chemical Industry

Keywords: browse; leaf litter; nutritive value; tree forage quality; drying method

INTRODUCTIONGoats use tree, brush and vine leaf litter (LL) as a feed sourcewhen green browse is not available. This has been documentedin warm climates with prolonged annual droughts1,2 as well asin cooler climates where leaf loss is associated with shorter daylength and cool temperatures.3 Although studies have indicatedthat the nutritive value of LL is generally lower than that ofgreen leaves (GL),3,4 the dynamics of condensed tannin (CT)concentrations and CT fractions as leaves change to litter arenot known. Greater knowledge of CT and other nutritive valuessuch as crude protein (CP), fibre fractions and phosphorus (P) in LLwill help to understand the role they play in environments wherethis feed resource is important to browsing ruminants.

Tannins are polyphenolic compounds with the ability toprecipitate proteins and play a role in ecological processes suchas LL decomposition, nutrient cycling, nitrogen sequestration andmicrobial activity.5,6 They can be divided into two major classes,namely condensed and hydrolysable.7 CT have both positive andnegative effects in ruminants, depending on their concentration inforages. According to Barry and McNabb,8 the ideal concentrationof CT in forage legumes fed to ruminants is between 20 and40 g kg−1.

In higher concentrations, CT can reduce feed intake, protein andcarbohydrate digestibility and animal performance.9 Moderatelevels of CT increase the efficiency of CP utilisation throughgreater flow of protein to the duodenum10 by inhibiting ruminaldegradation of plant proteins and enhancing ruminal protein

escape (bypass).11 Research has also shown that plants containingCT may have anthelmintic properties.12 – 14 Feeding ruminantsthese plants can suppress gastrointestinal parasite faecal eggcounts.15

Previous research on herbaceous species indicates that con-centrations of CT, neutral detergent fibre (NDF), acid detergentfibre (ADF) and acid detergent lignin (ADL)16,17 change with dry-ing method, i.e. oven-drying (OD) or freeze-drying (FD), owingto the formation of complexes between tannins and cell wallcompounds, increasing CT bound to protein and fibre.18 However,the effect of natural drying of LL on CT fractions present beforeabscission has not been thoroughly studied.

To better understand the CT and nutritive value changes thatoccur as leaves become LL, browse species known to have CTin greater concentrations than 20 g kg−1 on a dry matter (DM)basis were selected for this study. These included two vine species,Smilax rotundifolia (SR) and Smilax bona-nox (SB), and seven tree orshrub species, Celtis occidentalis (CO), Quercus stellata (QS), Ulmus

∗ Correspondence to: James P Muir, Texas AgriLife Research, 1229 North U.S.Highway 281, Stephenville, TX 76401, USA. E-mail: [email protected]

a Universidad de Cundinamarca, Facultad de Ciencias Agropecuarias, Diagonal18 No. 20-29, Fusagasuga, Colombia

b Texas AgriLife Research, 1229 North U.S. Highway 281, Stephenville, TX76401, USA


25

83

Compositional changes as browse leaves become litter www.soci.org

crassifolia (UC), Bumelia lanuginosa (BL), Quercus virginiana (QV),Carya illinoinensis (CI) and Gleditsia triacanthos (GT). The objectiveof our study was to determine the composition of key compoundsand nutrients in pre-abscission leaves and LL in order to betterpredict their potential nutritive value and health contributions toruminants consuming LL during winter or dry seasons. We alsowanted to compare the effect of OD or FD of leaves with that ofthe natural drying that occurs when leaves become LL.

MATERIALS AND METHODSLeaves from plants of each species were harvested at threelocations in October 2007 at the Texas AgriLife Research Center,Stephenville, TX, USA (32◦ 15′ N, 98◦ 12′ W, altitude 395 m) andimmediately frozen with dry ice. Subsequently, half were oven-dried (GL-OD) at 55 ◦C for 48 h and the other half were freeze-dried(GL-FD) at −40 ◦C for 72 h (FTS System Dur-Top MP, SP Industries,Warminster, PA, USA). Shade cloth nets were placed under theplants at the same time to collect LL throughout January 2008.

Nutrient and CT fraction compositions were determined for GL-OD, GL-FD and LL. Samples were digested by the Dumas method19

using an Elementar Vario Macro C : N analyser (Elementar Americas,Inc., Mt Laurel, NJ, USA) to determine nitrogen (N) concentration.CP was estimated as N × 6.25.20 An ANKOM200 fibre analyser(Ankom Technologies, Macedon, NY, USA) was used to determineneutral detergent fibre (NDFom), acid detergent fibre (ADFom)and acid detergent lignin (ADLom) in a sequential analysis withthe addition of sodium sulfite and heat-stable α-amylase. Thesevalues were corrected to an organic matter (om) basis. Samplesof 0.5 g were digested in sulfuric acid and extracts were used todetermine P concentration by a modified ascorbic acid method.21

Extractable (ECT), protein-bound (PBCT), fibre-bound (FBCT) andtotal (TCT) CT in GL-OD, GL-FD and LL were determined by thebutanol/HCl procedure.22 CT were extracted from each species todevelop a specific standard curve for each species.17

Plant harvested (three for each species) was considered asreplicate. Data were analysed for variance by species using a leastsignificant difference test (P ≤ 0.05) to determine mean treatmentdifferences. Differences were considered significant at P ≤ 0.05unless noted otherwise in the text. The general lineal modelprocedure of SAS/STAT 9.123 was used based on the statisticalmodel Yij = µ + αj + Eij , where αj is the effect of drying method.

The dependent variables were ECT, PBCT, FBCT, TCT, CP, P, NDFom,ADFom and ADLom.

RESULTSNutritional compositionThere were no differences in CP concentration among GL-OD,GL-FD and LL in all species except QS, UC and CI, where CPconcentration was lower in LL (Table 1). P concentration decreasedin LL relative to GL-FD of SB, QS, UC and BL as well as relative toGL-OD of SR, SB, CO, QS, UC and BL.

NDFom concentration was affected by drying method in SR, CIand GT, where higher values were found in LL (Table 2). The vinesSR and SB had higher ADFom concentrations in LL than in GL-FD,but there were no differences with GL-OD. ADFom concentrationsin QS, CI, UC and GT were higher in LL than in GL-OD and GL-FD.ADLom concentration was likewise higher in LL of QS and CI, whileSR, SB and CO had higher ADLom concentrations in LL and GL-ODthan in GL-FD.

Condensed tanninsECT concentration was higher in GL-FD than in GL-OD and LL ofSB, CO and QS, while CI, UC and GT had higher ECT concentrationin GL-FD than in LL, but no different from GL-OD (Table 3). PBCTconcentration was greater in GL-OD than in GL-FD and LL of UCand CI. In contrast, there were no differences in PBCT concentrationbetween GL-OD and LL of SB and BL, but PBCT concentration inGL-FD was lower in these species. FBCT concentration was higherin GL-OD than in GL-FD and LL of SR, SB, QS and UC, while QV hadhigher FBCT concentration in LL than in GL-OD and GL-FD. Dryingmethod changed TCT only in GT, UC and CI.

DISCUSSIONCP concentrations measured in the LL samples indicate that somespecies may be good sources of CP for concentrate-selectiveruminants such as goats. The lowest CP concentration in LL was63 g kg−1 for CI, while the highest was 151 g kg−1 for SB. Theseconcentrations are appropriate for satisfying protein requirementsin goats, whose reported requirements vary between 2.03 and3.07 g kg−1 (body weight)0.75 for maintenance.24,25 LL from most

Table 1. Crude protein (CP) and phosphorus (P) concentrations in green leaves oven-dried (GL-OD) or freeze-dried (GL-FD) and leaf litter (LL) ofTexas trees and vines

CP (g kg−1) P (g kg−1)

Species GL-FD GL-OD LL P ≤ GL-FD GL-OD LL P ≤Bumelia lanuginosa 112a 112a 95a 0.34 1.1a 1.2a 0.6b 0.04

Carya illinoinensis 120a 125a 63b 0.01 1.2a 1.4a 0.8a 0.06

Celtis occidentalis 123a 135a 87a 0.05 1.2b 1.7a 1.1b 0.01

Gleditsia triacanthos 134a 143a 107a 0.10 2.3a 2.7a 1.8a 0.66

Quercus stellata 121a 124a 67b 0.01 1.1b 1.6a 0.2c 0.01

Quercus virginiana 91a 91a 102a 0.42 2.0a 1.7a 0.2a 0.34

Smilax rotundifolia 139a 155a 137a 0.10 1.4ab 2.0a 0.9b 0.02

Smilax bona-nox 159a 164a 151a 0.33 1.7b 2.2a 1.0c 0.01

Ulmus crassifolia 121a 120a 87b 0.01 1.2b 1.5a 0.4c 0.01

Means within rows under the same subheading followed by different letters differ according to a least significant difference multiple mean separation(P ≤ 0.05).


25

84

www.soci.org A Acero, JP Muir, RM Wolfe

Table 2. Neutral detergent fibre (NDFom), acid detergent fibre (ADFom) and acid detergent lignin (ADLom) concentrations in green leavesoven-dried (GL-OD) or freeze-dried (GL-FD) and leaf litter (LL) of Texas trees and vines

NDFom (g kg−1) ADFom (g kg−1) ADLom (g kg−1)

Species GL-FD GL-OD LL P ≤ GL-FD GL-OD LL P ≤ GL-FD GL-OD LL P ≤Bumelia lanuginosa 250a 352a 341a 0.57 166a 173a 233a 0.46 74a 87a 117a 0.29

Carya illinoinensis 248b 282b 349a 0.01 148b 160b 211a 0.01 63b 73b 100a 0.01

Celtis occidentalis 344a 323a 330a 0.87 166a 153a 209a 0.09 44b 59a 80a 0.01

Gleditsia triacanthos 322b 325a 420a 0.01 238b 244b 319a 0.01 144b 168ab 211a 0.03

Quercus stellata 396a 399a 398a 0.98 225b 225b 263a 0.01 102b 99b 125a 0.02

Quercus virginiana 460a 450a 482a 0.32 316a 310a 333a 0.44 121a 127a 136a 0.37

Smilax bona-nox 344a 388a 409a 0.13 215b 250a 258a 0.01 60b 95a 86a 0.01

Smilax rotundifolia 336c 378b 411a 0.01 220b 242ab 261a 0.04 60b 87a 95a 0.01

Ulmus crassifolia 501a 482a 421a 0.22 168b 193b 265a 0.01 58b 85ab 123a 0.01


Table 3. Extractable (ECT), protein-bound (PBCT), fibre-bound (FBCT) and total (TCT) condensed tannin concentrations in green leaves oven-dried(GL-OD) or freeze dried (GL-FD) and leaf litter (LL) of Texas trees and vines

ECT (g kg−1) PBCT (g kg−1) FBCT (g kg−1) TCT (g kg−1)

SpeciesGL-FD

GL-OD LL

P≤

GL-FD

GL-OD LL

P≤

GL-FD

GL-OD LL

P≤

GL-FD

GL-OD LL

P≤

Bumelia lanuginosa 62.8a 32.1a 37.0a 0.10 21.7b 35.7a 35.2a 0.04 2.7a 7.0a 4.1a 0.05 76.2a 87.2a 74.9a 0.63

Carya illinoinensis 25.3a 15.3ab 5.8b 0.02 20.6b 40.5a 24.0b 0.01 7.7a 8.6a 5.6a 0.27 53.6a 64.4a 35.5b 0.05

Celtis occidentalis 19.0a 1.1b 0.0b 0.01 3.7a 2.8a 4.7a 0.09 5.0a 12.4a 5.6a 0.06 27.6a 16.3a 10.2a 0.08

Gleditsia triacanthos 36.9a 24.6ab 8.2b 0.03 27.6a 27.7a 25.7a 0.95 8.5a 6.9a 5.6a 0.27 73.0a 45.6b 39.6b 0.01

Quercus stellata 22.2a 6.1b 10.4b 0.01 7.8a 12.5a 6.9a 0.07 2.5b 7.4a 2.9b 0.01 32.5a 26.0a 30.2a 0.49

Quercus virginiana 40.1a 34.3a 30.3a 0.27 9.4a 16.4a 13.2a 0.08 2.2b 2.8b 5.8a 0.01 51.7a 53.5a 49.3a 0.85

Smilax bona-nox 44.0a 17.3c 25.0b 0.01 13.8b 27.5a 25.5a 0.01 3.5b 14.8a 2.6b 0.01 61.3a 59.6a 53.1a 0.15

Smilax rotundifolia 16.9a 8.1a 23.2a 0.08 16.4a 21.2a 22.1a 0.46 3.9b 23.3a 2.6b 0.003 37.2a 52.6a 48.0a 0.23

Ulmus crassifolia 37.2a 36.4a 5.3b 0.01 9.8b 41.6a 15.6b 0.03 4.3c 23.7a 10.2b 0.01 51.3b 91.7a 31.1c 0.01


of these species could contribute to meeting these requirementswhen other feed alternatives are limited. However, digestibilityand intake of this material, with resulting nutrient bioavailabilityto ruminants, must be determined before usefulness of this feedcan be defined.

The lower P concentration in LL compared with GL may bedue to the normal process of translocating nutrients in the plantbefore LL fall.26 This phenomenon may not be as developed for Nin most of the species studied, since only QS, UC and CI had lowerN concentration in LL compared with GL.

The effect of drying method on the fibre fraction becameprogressively more evident with each assay step. The NDFomfraction was greater in the LL of only three species, ADFomwas greater in six species, whereas ADLom was more concen-trated in seven species compared with GL. This indicates thatADLom is more stable and degrades less in LL compared withhanging leaves than ADFom, which in turn is more stable thanNDFom.

Schofield et al.,27 in a study of willow tree LL, found thatextractable phenolics and tannins were rapidly lost fromthe leaves and had half-lives of approximately 2.4 weeks.Lower-molecular-weight tannins were lost more rapidly than

higher-molecular-weight tannins, suggesting that the primaryroute for loss of tannins is leaching. These results parallel thoseof three species in our study, but not the six for which there wasno decline in TCT from GL to LL. Our results would indicate thatmolecular weights in the latter six species may be greater thanthose in the first three species.

TCT concentrations in LL ranged between 10 and 75 g kg−1

DM. Four species in our study had LL TCT concentrations greaterthan 45 g kg−1, levels that might adversely affect the palatabilityto or performance of ruminants.28 Further research is needed todetermine if the dynamics of the CT–ruminant interactions in LLfollow the same general trends as those of fresh browse or forage,especially since PBCT and FBCT increase in LL compared with GLas a proportion of TCT.

Further research is also recommended on those species withhigh variability in CT concentrations among collection sites (repli-cations) as indicated by lack of statistical differences among treat-ments with large numerical differences, for example S. rotundifolia(Table 3). Hattenschwiler et al.29 reported differences in tree pop-ulation CT concentrations as influenced by soil properties intropical forests, so this or other factors may be in play and shouldbe identified.


25

85

Compositional changes as browse leaves become litter www.soci.org

Effect of drying method on PBCT and FBCT fractions differedamong species. Three had greater PBCT concentrations in GL-OD,with no differences between GL-FD and LL. The same pattern wasobserved in FBCT concentrations of three species, where GL-ODhad greater concentrations compared with GL-FD and LL. Thiscould be partly a result of OD causing the formation of proteinand fibre tannin complexes as reported in previous studies,16,17

but why this pattern was not uniform across all species needs tobe further studied.

CONCLUSIONSThe transformation from GL to LL reduces nutritive value in somespecies but not others, and these characteristics may explainruminant selectivity and performance where LL is an importantruminant feed resource. LL of some species may be a usefulsource of CP for browsers in situations, such as winter, wheregreen material is not available. Fibre and CT concentrationsare not excessive in the species studied, so these should notcompromise animal performance. However, the implications ofdecreased proportions of ECT and increased PBCT and FBCTin some species as leaves abscise need to be further studied,especially in regard to ruminal protein bypass (escape) or othersecondary effects on ruminant health.

The effect of drying method on the variables evaluated didnot show a uniform pattern among the species studied; however,OD of GL tended to closer emulate LL compared with FD ofGL, although patterns were not consistent across species. Morestudies are necessary to better understand the interaction betweendrying, CT and cell wall fibre components. For example, a betterunderstanding of the CT chemical characteristics in these speciescould help to explain the variation in response to leaf drying andabscission.

ACKNOWLEDGEMENTThis research was funded in part by grant 2005-51300-02392 fromthe USDA-CSREES Integrated Organic Program.

REFERENCES1 Pfister JA and Malechek JC, Dietary selection by goats and sheep

in a deciduous woodland of northeastern Brazil. J Range Manag39:24–28 (1984).

2 Scholte PT, Leaf litter and Acacia pods as feed for livestock during thedry season in Acacia-Commiphora bushland, Kenya. J Arid Environ22:271–276 (1992).

3 Packard CE, Muir JP and Wittie RD, Effects of groundnut stoveror bermudagrass hay supplementation to doe kids on winterhardwood range. Small Rumin Res 67:1–6 (2007).

4 Kronberg SL and Malechek JC, Relationships between nutrition andforaging behavior of free-ranging sheep and goats. J Anim Sci75:1756–1763 (1997).

5 Kraus TEC, Dahlgren RA and Zasoski RJ, Tannins in nutrient dynamicsof forest ecosystems. Plant Soil 256:41–66 (2003).

6 Haase K and Wantzen KM, Analysis and decomposition of condensedtannins in tree leaves. Environ Chem Lett 6:71–75 (2008).

7 Reed JD, Nutritional toxicology of tannins and related polyphenols inforage legumes. J Anim Sci 73:1516–1528 (1995).

8 Barry TN and McNabb WC, The implications of condensed tannins onthe nutritive value of temperate forages fed to ruminants. Br J Nutr81:263–272 (1999).

9 Reed JD, Soller H and Woodward A, Fodder tree and straw diets forsheep: intake, growth, digestibility, and the effects of phenolics onnitrogen utilization. Anim Feed Sci Technol 30:39–50 (1990).

10 McNabb WC, Waghorn GC, Barry TN and Shelton ID, The effect ofcondensed tannins of Lotus pedunculatus on the digestion andmetabolism of methionine, cystine and inorganic sulphur in sheep.Br J Nutr 70:647–661 (1993).

11 Makkar HPS, Effects and fate of tannins in ruminant animals,adaptation to tannins, and strategies to overcome detrimentaleffects of feeding tannin-rich feeds. Small Rumin Res 49:241–256(2003).

12 Molan AL, Waghorn GC, Min BR and McNabb WC, The effect ofcondensed tannins from seven herbages on Trichostrongyluscolubriformis larval migration in vitro. Folia Parasitol 47:39–44(2000).

13 Paolini V, Bergeaud JP, Grized C, Prevot F, Dorchies P and Hoste H,Effects of condensed tannins on goats experimentally infected withHaemonchus contortus. Vet Parasitol 113:253–261 (2003).

14 Lange KC, Olcott DD, Miller JE, Mosjidis JA, Terrill TH, Burke JM, et al,Effect of Sericea lespedeza (Lespedeza cuneata) fed as hay, on naturaland experimental Haemonchus contortus infections in lambs. VetParasitol 141:273–278 (2006).

15 Min BR and Hart SP, Tannins for suppression of internal parasites.J Anim Sci 81:E102–E109 (2003).

16 Steward JL, Mould F and Mueller-Harvey I, The effect of dryingtreatment on the fodder quality and tannin content of twoprovenances of Calliandra calothyrsus Meissner. J Sci Food Agric80:1461–1468 (2000).

17 Wolfe RM, Terrill TH and Muir JP, Drying method and origin ofstandard affect condensed tannin (CT) concentrations in perennialherbaceous legumes using simplified butanol-HCl CT analysis. J SciFood Agric 88:1060–1067 (2008).

18 Terrill HT, Windham WR, Evans JJ and Hoveland C, Effect of dryingmethod and condensed tannin on detergent fiber analysis ofSericea lespedeza. J Sci Food Agric 66:337–343 (1994).

19 AOAC, Official Methods of Analysis (15th edn), Vol. 1. Association ofOfficial Analytical Chemists. Arlington, VA (1990).

20 Van Soest PJ, Nutritional Ecology of the Ruminant (2nd edn). CornellUniversity Press, Ithaca, NY (1994).

21 APHA, Phosphorus, in Standard Methods for the Examination of Waterand Wastewater, ed. by Eaton AD, Clesceri LS and Greenberg AE.American Public Health Association, Bethesda, MD, pp. 106–115(1995).

22 Terrill TH, Rowan AM, Douglas GD and Barry TN, Determination ofextractable and bound condensed tannin concentrations in forageplants, protein concentrate meals and cereal grains. J Sci Food Agric58:321–329 (1992).

23 SAS, SAS/STAT 9.1 User’s Guide. SAS Institute, Cary, NC (2004).24 NRC, Nutrient Requirements of Small Ruminants: Sheep, Goats, Cervids,

and New World Camelids. National Research Council, Washington,DC (2007).

25 Luo J, Goetsch AL, Nsahlai IV, Sahlu T, Ferrell CL, Owens FN, et al,Metabolizable protein requirements for maintenance and gainof growing goats. Small Rumin Res 53:309–326 (2004).

26 Mafongoya PL, Guiller KE and Palm CA, Decomposition and nitrogenrelease patterns of tree prunings and litter. Agroforest Syst 38:77–97(1998).

27 Schofield JA, Hagerman AE and Harold A, Loss of tannins and otherphenolics from willow leaf litter. J Chem Ecol 24:1409–1421 (1998).

28 Min BR, Barry TN, Attwood GT and McNabb WC, The effect ofcondensed tannins on the nutrition and health of ruminants fedfresh temperate forages: a review. Anim Feed Sci Technol 106:3–19(2003).

29 Hattenschwiler S, Hagermand AE and Vitousek PM, Polyphenols inlitter from tropical montane forests across a wide range in soilfertility. Biogeochemistry 64:129–148 (2003).


25

86



Fruit dry matter concentration: a new qualitymetric for applesJohn W Palmer,a∗ F Roger Harker,b D Stuart Tustinc and Jason Johnstonb

Abstract

BACKGROUND: In the fresh apple market fruit must be crisp and juicy to attract buyers to purchase again. However, recentstudies have shown that consumer acceptability could be further enhanced by improving taste. This study evaluates the use offruit dry matter concentration (DMC) as a new fruit quality metric for apple.

RESULTS: Fruit samples collected at harvest, in the two main fruit growing regions of New Zealand, showed a variation in meanfruit DMC from 130 to 156 g kg−1 with ‘Royal Gala’ and with ‘Scifresh’ from 152 to 176 g kg−1. Individual fruit DMC showed alarger range, from 108 to 189 g kg−1 with ‘Royal Gala’ and from 125 to 201 g kg−1 with ‘Scifresh’. Fruit DMC proved a morereliable predictor of total soluble solids after 12 weeks of air storage at 0.5 ◦C than TSS at harvest for both ‘Royal Gala’ and‘Scifresh’. Fruit DMC was also positively related to flesh firmness, although this relationship was not as strong as that seen withsoluble solids and was more dependent on cultivar. Consumer studies showed that consumer preference was positively relatedto fruit DMC of ‘Royal Gala’ apples.

CONCLUSION: Fruit DMC can therefore be measured before or at harvest, and be used to predict the sensory potential for thefruit after storage.c© 2010 Society of Chemical Industry

Keywords: Malus domestica Borkh.; flesh firmness; fruit maturity; soluble solids; consumer panel; dry matter concentration

INTRODUCTIONEating quality in the broadest sense (texture, taste, odour) is oneof the primary reasons consumers purchase fruit1 – 3 and conse-quently the development of knowledge and indices to predicteating quality has become a significant focus for postharvest bi-ologists and technologists.4 – 6 Research has generally sought toestablish key objective measurements that can be used to predictconsumer liking and willingness to purchase fruit4,7 or specificsensory attributes as measured by trained panellists.8 – 10 Thisneed for objective measurements to predict consumer responseshas evolved in two ways: (1) a move from destructive to non-destructive instruments so that each individual fruit can be sortedinto the appropriate quality categories;11,12 and (2) the desire toapply quality indices based on the fundamental biology of thecrop, which can be monitored through fruit growth and develop-ment and applied at harvest to represent the quality of an entirecrop.13 The advantage of the latter approach is that it providesearly insights for logistics and marketing and can help to identifyphysiological processes and crop management practices that canbe modified in order to provide solutions that improve qualityacross an entire fruit sector. Traditionally, the best examples ofthe biologically based quality standards involve the monitoring offruit maturity to support decisions on the timing of harvest.14 – 16

In the fresh apple market, fruit with good textural properties areparamount; the fruit must be crisp and juicy to attract buyers topurchase again. A recent consumer study of fruit from five applecultivars in the USA showed that flesh firmness is the primaryedible quality factor determining consumer acceptance.4 Much ofpostharvest science over the last 30 years has been devoted to

the control and maintenance of flesh firmness in apples. However,once fruit are above a minimum threshold for flesh firmness,consumer acceptability could be further enhanced by improvingtaste (total soluble solids (TSS) and titratable acidity).4 In otherwords, although texture is paramount, the customer is also lookingfor high taste intensity in the fresh product. While volatiles providethe signature orthonasal and retronasal stimuli responsible forcharacteristic flavour of apple,17 our ability to establish whichindividual or combinations of aroma compounds contribute toconsumers’ preferences for whole apples has proven difficult untilrecently,18 probably due to biological variability associated withboth apples and human ability to perceive odour. Thus there maybe opportunities for a more holistic predictor of eating qualitybased on knowledge of physiological and metabolic processes.

In recent times there has been an increasing focus on fruit drymatter concentration (DMC), either as it relates to maturity (e.g.avocado) or to consumer preference in itself.13,19 For example,fruit DMC of kiwifruit at harvest has been shown to be a goodpredictor of ripe soluble solids content after storage.13 Consumer

∗ Correspondence to: John W Palmer, New Zealand Institute for Plant and FoodResearch Limited (PFR), 55 Old Mill Road, RD 3, Motueka 7198, New Zealand.E-mail: [email protected]

a New Zealand Institute for Plant and Food Research Limited (PFR), Motueka7198, New Zealand

b PFR, Mt Albert Research Centre, Auckland 1142, New Zealand

c PFR, Hawke’s Bay Research Centre, Havelock North, Hastings 4157, New Zealand


25

87

Apple fruit dry matter concentration www.soci.org

studies have also shown that kiwifruit harvested with high DMC arepreferred, after ripening, to fruit harvested with low DMC levels.13

As these authors point out, the improvement in consumer likingis not just associated with the higher TSS, because, for example,titratable acidity was also positively correlated with fruit DMC. Insimilar studies with avocado, consumers showed a progressiveincrease in both liking and intent to buy as DMC increasedfrom ∼200 g kg−1 (minimally mature) to nearly 400 g kg−1 (verymature).19 It is surprising that there has not been an earlier attemptto use fruit DMC as a quality index for apples, given its successwith other fruits.

It has been shown that fruit DMC of ‘Royal Gala’ apple at harvestcan be used to predict TSS after cool storage.20 Few people seemto have followed up on this linkage between apple fruit DMC andTSS after storage. This current work set out firstly to quantify thevariation in apple fruit DMC, within and between fruit samples, andsecondly to extend this earlier work across several apple cultivarsbut, in addition, to examine possible links between fruit DMC andother key quality attributes such as flesh firmness, in the light ofthe importance of fruit texture for apple. However, in attemptingto develop a new quality index, one of the critical early steps is toestablish that the measurement can predict consumer liking, andso the third part of this work addresses this issue. We believe thisis the first extensive study of its kind with apple.

MATERIALS AND METHODSFifty-fruit samples of ‘Royal Gala’ and ‘Scifresh’ fruit were collectedjust before the start of commercial harvest in 2005 from 12orchards in the Nelson (latitude 41 ◦S, longitude 173 ◦E) and 17 inthe Hawke’s Bay (latitude 40 ◦S, longitude 177 ◦E) regions of NewZealand. Fruit DMC was measured by drying two wedges per fruitat 65 ◦C in a forced draught oven for at least 48 h to a constantweight. The fruit wedges were taken from opposite sides of eachfruit, using two longitudinal cuts. This sampling procedure wasadopted to allow for the variation in fruit DMC within individualfruit.21 The sampled tissue fresh weight was ∼15–25 g beforedrying.

Further fruit samples were taken from the first and the thirdselect picks of ‘Royal Gala’ and ‘Scifresh’ from four orchards in eachof Nelson and Hawke’s Bay. From each pick, 10 fruit were used torecord fruit DMC and a further 10 fruit used to assess at-harvestquality: TSS using an Atago Pocket PAL-1 digital refractometer,flesh firmness using a Guss Fruit Texture Analyser fitted with an11.1 mm probe, starch pattern index (SPI) using a scale from 0 to 6,no clearance to full starch clearance, and skin background colourusing ENZA swatches from 1 to 10, green to yellow. Additional40-fruit samples were air-stored for either 6 or 12 weeks at 0.5 ◦C.After storage, fruit DMC, fruit firmness and TSS were recorded. Thiswas repeated for ‘Royal Gala’ in the 2005/6 growing season.

A further dataset was obtained in the 2005/6 growing seasonfrom eight commercial apple cultivars (’Cox’s Orange Pippin’ clone‘Greenmeadows’, ‘Fuji’, ‘Granny Smith’, ‘Scifresh’, ‘Sciearly’, ‘Sciros’,Pink Lady and ‘Royal Gala’). The fruit were sourced from Hawke’sBay during the commercial harvesting period, couriered to PFR,Auckland, and stored at 0.5 ◦C, or 3 ◦C for ‘Cox’. After storage, twoadjacent wedges of fruit were removed from each of 22 fruit ofeach cultivar and one used for measurement of TSS and one forfruit DMC. In this case, TSS was determined on juice obtained bypassing the fruit wedge through a domestic juicer, rather thanusing the juice on the tip of the penetrometer probe used in theother trials.

For sensory assessment experiments, ‘Royal Gala’ apples(100–110 count) were harvested from nine orchard blocks inHawke’s Bay and nine orchard blocks in Nelson during February2008. Within each region there were three DMC categories (low,moderate and high), and three orchard blocks per DMC. Fruitwere representative of those harvested during the first maincommercial pick (excluding an initial skim pick), with fruit eithersampled directly from the picked bins or directly from the tree1–2 days before commercial harvest. All fruit met commercialrequirements for blush coverage and background colour. A30-fruit sample from each orchard block was then assessedfor at-harvest characteristics (within 24 h of harvest), and theremainder (170–200 fruit) were placed in air storage at 0.5 ◦C for10–12 weeks. At-harvest assessments included skin backgroundcolour, internal ethylene concentration, fruit DMC, firmness, TSSand SPI (using the methodology already described). Internalethylene concentration was determined by injecting a 1 mL coregas sample from each fruit into a gas chromatograph equippedwith a flame ionisation detector. Post storage, each apple wasassessed for firmness, DMC, TSS and titratable acidity before usingthe remaining tissue for sensory assessment. A wedge adjacentto the puncture test was used for titratable acidity determination.A subsample of 5 g of tissue was homogenised in 25 mL waterusing a Polytron homogeniser (Kinematica, Luzern, Switzerland),and titrated to an endpoint of pH 8.1 with 0.1 mol L−1 sodiumhydroxide (NaOH) using an automatic titrator (Model 702 SMTitrino, Metrohm, Herisau, Switzerland). Results were expressed asmalic acid equivalent (g kg−1).

One hundred and four consumers from the general populationof New Zealand living in Auckland were recruited on the basisof being between the ages of 18 and 65 years and being regularapple eaters, i.e. eating apples more than once a week. Fifty-twopercent were male, 29% aged 18–30, 38% aged 31–45 and 33%aged 46–60 years. The vast majority (92%) consumed ‘Royal Gala’apples.

Each consumer tasted four sets of three fruit. Two sets werefrom Nelson and two from Hawke’s Bay, and each set includeda fruit from each DMC category. The order in which fruit weretasted was according to a randomised complete block statisticaldesign that accounted for order and carryover effects. Care wastaken to ensure all orchards were equally represented withineach DMC category, although each consumer did not taste fruitfrom all orchards. Consumers’ liking of apples was assessed on a9-point category scale using a modified rank-rating approach,in which consumers were asked to consider the relative aswell as absolute scores.13,19 Following this they indicated tasteacceptability and their willingness to purchase fruit if availableat an average price ($NZ 2.60 kg−1) using a 6-category scale.19

Tastings were undertaken under red-coloured lighting in sensoryfacilities described previously.3

Statistical analysis of variation in fruit DMC across orchards andrelationships between fruit DMC and other quality variates wascompleted using linear regression with Genstat (version 10.1.0.71,VSN International Ltd, Hemel Hempstead, UK) or Origin (version7.5, OriginLab Corp., Northampton, MA, USA).

Mixed-effects models relating instrumental measurements toconsumer responses were fitted using PROC Mixed in the SAS 9.1statistical package. The region and DMC categories were treatedas fixed effects, with the effect related to orchard being modelledas random. This allowed the estimates to take into account thestructure of the data, in that fruit from within the same orchard maybe more similar than fruit from different orchards. Additionally the


25

88

www.soci.org JW Palmer et al.

Table 1. Fruit DMC (g kg−1) of orchard samples (50 fruit) of ‘RoyalGala’ and ‘Scifresh’ apples harvested just prior to or on the day of thefirst commercial harvest. Samples were collected from 12 orchards inthe Nelson and 17 orchards in the Hawke’s Bay (HB) regions of NewZealand

Cultivar: ’Royal Gala’ ’Scifresh’

Region: Nelson HB Nelson HB

Mean 138 147 159 164

Range of samples 130–150 137–156 152–165 156–176

Range of individual fruit 108–175 113–189 132–198 125–201

panellist may have a similar effect on liking scores and purchaseintent. For these analyses, the panellist effect was also treated asrandom. Since the acceptability data were binary, with ‘acceptable’or ‘not acceptable’ the only possible answers, a generalised linearmixed effects model was used to account for the orchard andpanellist effects. This was estimated in Genstat, and used theapproach of Breslow and Clayton.22 The fitted curves in the plots ofphysical measurements against liking scores were estimated usinga LOESS smoother routine using R version 2.90 (R DevelopmentCore Team http://www.R-project.org). This had the advantage ofnot having to constrain the data to a preconceived theoreticalmodel.

RESULTSVariation across orchard blocksConsiderable variation in fruit DMC was evident among orchardsamples of apples (Table 1), with a 13–20 g kg−1 difference amongorchard samples and a ∼70 g kg−1 difference between individualfruit for any region/cultivar combination. Within one region,‘Scifresh’ orchard samples had a consistently higher fruit DMCthan samples of ‘Royal Gala’.

Relationship between fruit DMC and total soluble solidsAlthough the relationship between fruit DMC at harvest and TSS atharvest of ‘Royal Gala’ orchard samples was statistically significant(r2 = 0.32), fruit DMC at harvest was more closely related toTSS after storage, with an improvement in the relationship thelonger the fruit remained in storage. After 12 weeks of storage therelationship accounted for 82% of the variation (Fig. 1). The resultsfrom the 2005/6 season were similar to those of the 2004/5 seasonand the combined relationship after 12 weeks of storage is givenin Fig. 2. The effect of year was not statistically significant.

When comparisons were made using TSS at harvest or fruitDMC at harvest as the predictive variate for TSS of ‘Royal Gala’after 6 weeks of storage, the relationships were similar; bothrelationships accounted for 49% of the variation. However, after12 weeks of storage, fruit DMC accounted for 81% of the variationin TSS compared with 37% of the variation where TSS at harvestwas used as the predictive variate.

It was obviously not possible to use fruit DMC at harvest as apredictor of TSS after storage on the same fruit, as fruit DMC is adestructive test. However, TSS could be related to fruit DMC whenthe fruit DMC was determined. On an individual fruit basis, TSS of‘Royal Gala’ at harvest was poorly related to fruit DMC at harvest(r2 = 0.10). This relationship improved with storage after 6 weeks(r2 = 0.52) and after 12 weeks (r2 = 0.63).

Similar trends were evident with ‘Scifresh’ orchard samples tothose observed with ‘Royal Gala’ orchard samples. The relationshipbetween fruit DMC at harvest and TSS improved with storage time(Fig. 3). In this case, the relationship was strongest after 6 weeksin storage. Fruit DMC at harvest again proved a better predictorof TSS after storage than TSS at harvest. After 6 weeks of storage,variance accounted for by the regression between harvest fruitDMC and TSS was 78%, compared with 57% when harvest TSSwas used. After 12 weeks, the two values were 54% and 32%respectively.

On an individual fruit basis at harvest, TSS of ‘Scifresh’ was poorlyrelated to fruit DMC (r2 = 0.17). This relationship improved withstorage after 6 weeks (r2 = 0.61) and after 12 weeks (r2 = 0.68).

Considering orchard samples of both cultivars together, therelationship between TSS and fruit DMC, where both variateswere measured after 12 weeks, was highly statistically significant(r2 = 0.97). In the absence of a significant effect of cultivar, bothcultivars could be represented by the same linear relationship(Fig. 4).

A similar result was found in the study of individual fruit fromeight cultivars (Fig. 5) where TSS after storage was highly correlatedwith fruit DMC recorded after storage (r2 = 0.77).

Relationship between fruit DMC and flesh firmnessThere was no statistically significant relationship between fruitDMC at harvest and firmness for ‘Royal Gala’ orchard samples atharvest or firmness after 6 or 12 weeks of storage. Flesh firmnessafter storage was primarily related to flesh firmness at harvest,with 52% and 55% of the variation at 6 and 12 weeks explained byflesh firmness at harvest. However, when the data for individualfruit were analysed, there was a significant positive relationshipbetween fruit DMC and firmness, although the proportion of thevariation accounted for by the regression was 6%, 10% and 12%for samples at harvest, 6 weeks and 12 weeks respectively.

However, with ‘Scifresh’ orchard samples, fruit DMC at harvestwas significantly related to flesh firmness both at harvest andafter storage, with the regressions accounting for 22%, 52% and44% of the variation in flesh firmness at harvest and after 6 weeksand 12 weeks of storage, respectively. Flesh firmness at harvest,however, was a better predictor of flesh firmness after storage thanfruit DMC at harvest, with regressions accounting for 70% and 73%of the variation in flesh firmness after 6 weeks and 12 weeks ofstorage respectively. Post storage, flesh firmness in the orchardsamples was significantly related to fruit DMC, where fruit DMCwas recorded after storage, with the regressions accounting for70% and 67% of the variance in flesh firmness at 6 weeks and12 weeks respectively.

Relationship between fruit DMC and consumer responseAll apples were allocated into treatments on the basis of whetherthey were harvested from low, medium or high DM orchards,which is fundamentally different from previous studies, whichhave allocated fruit to treatments on the basis of instrumentalmeasurement made just before consumer assessments.4,13,19

Given the potential high risk that individual apples allocatedto treatments on the basis of orchard might not have met requiredDMC criteria, it was important to evaluate the effectiveness of thesorting process. The process of sorting the apples into differentDMC categories was successful (Table 2), with differences betweencategories pronounced. All DMC categories were significantlydifferent from each other (P < 0.001 and P < 0.001 for all three


25

89


120 125 130 135 140 145 150 155 16010

11

12

13

14 Hawke's Bay

A B

C

Nelson

Sol

uble

sol

ids

at h

arve

st (

°Brix

)

Dry matter concentration at harvest (g kg-1)

r2 = 0.32

120 125 130 135 140 145 150 155 16010

11

12

13

14Hawke's BayNelson

Sol

uble

sol

ids

afte

r 6

wee

ks s

tora

ge (

°Brix

)


r2 = 0.53

120 125 130 135 140 145 150 155 16010

11

12

13

14Hawke's BayNelson

Sol

uble

sol

ids

afte

r 12

wee

ks s

tora

ge (

°Brix

)


r2 = 0.82

Figure 1. Relationship between harvest dry matter concentration and (A) total soluble solids at harvest, (B) after six weeks of cool storage and (C) after12 weeks of cool storage of ‘Royal Gala’ apple fruit from the first and third select picks of four orchards in Hawke’s Bay and four orchards in Nelson, NewZealand, 2005.

120 125 130 135 140 145 150 155 16010

11

12

13

14

Sol

uble

sol

ids

afte

r 12

wee

ks s

tora

ge (

°Brix

)

Fruit dry matter concentration at harvest (g kg-1)

2004/5

2005/6

r2 = 0.78

Figure 2. Relationship between harvest DMC and TSS after 12 weeks ofcool storage of ‘Royal Gala’ apple fruit over two seasons (2004–06) fromthe first and third select picks of four orchards in Hawke’s Bay and fourorchards in Nelson, New Zealand.

pairwise comparisons after Tukey’s HSD adjustment for multipletesting differences). The process of allocating apples to individualconsumers ensured minimal overlap in DMC categories for any setof three apples tasted by an individual consumer. Most pairs offruit had positive differences in DMC values (i.e. very few fruit froma higher DMC category that were actually lower in DMC) and fewfruit that had no (zero) difference between DMC categories. Rather,the average DMC differences going from ‘low’ to ‘moderate’, from‘moderate’ to ‘high’ and from ‘low’ to ‘high’ were 11, 14 and25 g kg−1 respectively.

Consumers were asked to taste apples of differing DMC toindicate their absolute level of liking of each fruit, and topay attention to the relative pattern of liking between thesedifferent apples. There was a significant increase in consumerliking for apples from the high DMC category (P = 0.001; pairwisecomparisons after Tukey’s HSD adjustment for multiple testingdifferences, high versus moderate P = 0.004, high versus lowP < 0.001, moderate versus low P = 0.318). Average scores forconsumer liking for the apples increased from 5.2 for the low DMCcategory to 6.3 for the highest DMC category (Fig. 6(A)).

For the supplementary question (’Was the taste of the fruitacceptable or unacceptable?’), the acceptability of the DMCcategories increased from 69% to 83% of apples as the DMCincreased from low to high (Fig. 6(B)). DMC categories had asignificant influence on acceptability of apples (P = 0.028).There was a significant increase in consumer acceptance for


25

90


140 145 150 155 160 165 170 175 18012

13

14

15

16 Hawke's BayNelson

Sol

uble

sol

idsa

t har

vest

(°B

rix)


r2 = 0.42

140 145 150 155 160 165 170 175 18012

13

14

15

16Hawke's BayNelson

Sol

uble

sol

ids

afte

r 6

wee

ks s

tora

ge (

°Brix

)

Drymatter concentration at harvest (g kg-1)

r2 = 0.79

140 145 150 155 160 165 170 175 18012

13

14

15

16Hawke's BayNelson

Sol

uble

sol

ids

afte

r 12

wee

ks s

tora

ge (

°Brix

)


r2 = 0.57

A B

C

Figure 3. Relationship between harvest DMC and (A) TSS at harvest, (B) after 6 weeks of cool storage and (C) after 12 weeks of cool storage of ‘Scifresh’apple fruit from the first and third select picks of four orchards in Hawke’s Bay and five orchards in Nelson, New Zealand, 2005.

130 140 150 160 170 18011

12

13

14

15

16

Sol

uble

sol

ids

(°B

rix)

Fruit dry matter concentration (g kg-1)

Royal GalaScifresh

r2 = 0.97

Figure 4. Relationship between DMC and TSS after 12 weeks of coolstorage of ‘Royal Gala’ and ‘Scifresh’ apple fruit from the first and thirdselect picks of four orchards in Hawke’s Bay and four orchards in Nelson,New Zealand, 2005.

pairwise comparisons between low and high DMC (P < 0.011),while comparisons between high and moderate (P = 0.109) andmoderate and low (P = 0.221) were not significant.

There were significant increases in consumers’ intention topurchase apples at the stated price of $2.60 kg−1 for eachstepwise increase in DMC band (P = 0.001). Tukey’s pairwisecomparisons showed significant differences between high tolow DMC (P = 0.001) and high to moderate DMC (P = 0.022),but no significant differences between moderate and low DMC(P = 0.209) (Fig. 6(C)). An alternative approach to investigatingthe relation between DMC and consumer responses is to groupthe results by actual DMC rather by an arbitrary DMC category(e.g. low, moderate and high) allocated at harvest. This type ofanalysis is only possible in retrospect, since actual DMC were notavailable until some days after the consumer tasting sessions.Analysis of these more detailed categories shown in Fig. 7(A)suggests a curvilinear relationship between increasing DMC ofapples and increasing consumer liking for fruit, although thereare limited sample sizes at very high and very low fruit DMC.The curve (LOESS smoother) was generated using all individualconsumers’ responses, and the points represent consumer likingas was calculated for discrete bands of DMC.

It was also possible to examine the influence of otherinstrumental attributes (TSS, fruit firmness and titratable acidity)besides that of fruit DMC. Fruit from categories with higher DMC


25

91


100 110 120 130 140 150 160 170 180 1909

10

11

12

13

14

15

16

17

Sol

uble

sol

ids

conc

entr

atio

n (°

Brix

)


CoxRoyal GalaPink Lady®

Pacific Rose™Pacific Beauty™ScifreshGranny SmithFuji

r2 = 0.77

Figure 5. Relationship between DMC and TSS after 12 weeks of cool storage of 22 individual fruit from eight apple cultivars sourced from Hawke’s Bay,New Zealand, 2006.

Table 2. DMC, TSS, fruit firmness and titratable acidity (malic acidequivalent g kg−1) of ‘Royal Gala’ apple fruit after 10–12 weeks at0.5 ◦C and 1 day at 20 ◦C. Means are based on six orchard blocks (tworegions, three orchard blocks per region, 86–91 fruit per block), analysisof variance P-values (12 d.f.e.), and 5% least significant differences (LSD)are displayed

Dry mattercategory

DMC(g kg−1)

Soluble solids(◦Brix)

Firmness(kgf)

Titratableacidity

(g kg−1)

High 157 13.6 6.62 2.9

Moderate 143 12.7 6.21 2.6

Low 132 11.8 6.19 2.4

P-value <0.001 <0.001 0.1133 0.02

5% LSD 4.5 0.40 0.45 0.30

also had higher TSS, firmness and titratable acidity (Table 2).Plots of consumer liking against each of these physical attributesindicate that all these traits were also positively associated withincreasing consumer preference (Fig. 7(B)–(D)).

There were no significant differences in consumer liking forNelson compared with Hawke’s Bay apples (data not shown;P = 0.6), or for apples harvested in different seasons (data notshown). This is in keeping with the instrumental measurements,where no regional differences were detected in firmness, TSS andtitratable acidity. The relationship of increased consumer likingwith increasing DMC was similar for both regions. The lack of anyregional differences suggests that DMC is a robust predictor ofconsumer preference.

An important component of this study was to evaluate whichat-harvest characteristic of ‘Royal Gala’ best predicted consumerresponses after storage. For the orchard blocks in this study, at-harvest values ranged from 0.1 to 3.0 µL L−1 for internal ethylene,7.3–8.8 kgf for firmness, 10.7–12.6% for TSS, 126–164 g kg−1

for fruit DMC and 1.6–4.7 for SPI. Background colour did notvary greatly, but this is to be expected given this index is used

commercially to pick fruit selectively. For DMC to be a successfulquality index in a commercial context, it has to be sufficientlyrobust to contend with this background variation.

It is therefore possible to compare the effectiveness of DMC as anindex for predicting eating quality before storage with other moretraditional indices such as firmness, TSS and SPI. However, thisshould be done with care, since the experiment was not designedwith this in mind. Correlations of at-harvest characteristics andsensory responses after storage showed that for mean sampleliking scores DMC (r2 = 0.58) and SPI (r2 = 0.53) were the twobest predictors, with firmness having a lower predictive association(r2 = 0.45). The percentage of consumers that rated the apples asacceptable was best predicted by DMC and SPI, although both r2

values were low to moderate (r2 = 0.41). The willingness to buywas best predicted by DMC (r2 = 0.55), followed by SPI (r2 = 0.46).The co-correlation amongst the different at-harvest variables alsomeans that there is limited value in performing multivariateprocedures, such as multiple linear regressions, to determineif combinations of two or more variables explain a higher degreeof variation in consumer responses. Overall these results suggestthat DMC was the best at-harvest predictor of sensory responses,followed closely by SPI, and that the relationships were linear.

DISCUSSIONVariation in fruit DMCThe variation in apple fruit DMC observed in this study was similarto that reported elsewhere. Belgian work showed that fruit DMCof individual ‘Jonagold’ apples varied within a single year from134 to 196 g kg−1.23 The range of just over 60 g kg−1 is verysimilar to the ranges recorded for both ‘Royal Gala’ and ‘Scifresh’.Although the mean fruit DMC of ‘Scifresh’ was higher than that of‘Royal Gala’, the range of fruit DMC across individual apples wassimilar. This variation in DMC in fruit within commercially pickedsamples represents a considerable variation in possible consumeracceptability. It must be noted that these fruit samples were fromthe first commercial pick. This spread in fruit DMC might be wider


25

92


Low Moderate High

0

2

4

6

8

A B C

Liki

ng S

core

DMC Category

bb

a

Low Moderate High

0

20

40

60

80

100

Acc

epta

bilit

y %

DMC Category

b

ab

a

Low Moderate High

0

20

40

60

80

100

Like

lihoo

d of

Pur

chas

e %

DMC Category

bb

a

Figure 6. Consumers’ scores for (A) liking, (B) acceptability and (C) purchase intent for ‘Royal Gala’ apples from different DMC categories after 10–12 weeksof cool storage. Letters show pairwise comparisons after Tukey’s HSD adjustment for multiple testing differences.

100 110 120 130 140 150 160 170 180 190

3

4

5

6

7

8

Liki

ng s

core


4 6 8

3

4

5

6

7

8

Liki

ng s

core

Flesh firmness (kgf)

9 10 11 12 13 14 15 16

3

4

5

6

7

8

Liki

ng s

core

Soluble solids (°Brix)

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

3

4

5

6

7

8

Liki

ng s

core

Titratable acidity (malic acid equivalents g kg-1)

5 7 9

AB

C D

Figure 7. Relationship between consumer liking scores (scale 1 – dislike extremely; to 9 – like extremely) and fruit DMC (A), fruit firmness (B), TSS (C) andtitratable acidity (D) for ‘Royal Gala’ apples after 10–12 weeks of cool storage. The range of each of the four physical measures was divided into 20 equallyspaced bins. The mean liking scores of each of these bins are shown plotted against the respective bin mid-points. The points denoted by the star symbolindicate mean liking scores based on five or fewer fruit. The curves were calculated from a LOESS smoother routine with a span of 0.5 and a degree of 1.


25

93


if later-picked fruit included those from the more shaded regionsof older, larger trees, which have a reduced fruit DMC.24

The variation in fruit DMC between samples from differentregions appeared to be similar. The tendency for higher fruit DMCto be recorded in Hawke’s Bay than in Nelson, particularly for ‘RoyalGala’, may suggest a regional difference related to differences inseasonal conditions, or to the small number of sampled orchardblocks.

In general, fruit DMC was higher for ‘Scifresh’ than for ‘RoyalGala’, suggesting cultivar differences. In the comparison of eightcultivars, Pink Lady apples were notable for their high fruit DMC(mean of 172 g kg−1) and ‘Royal Gala’ apples for their low fruitDMC (mean of 130 g kg−1). Similar consistent differences in fruitDMC between cultivars have been reported by several authors25,26

and recently a survey of 106 apple cultivars showed a range inflesh DMC from 80 to 186 g kg−1.27

Fruit growth modelling of apples normally involves using carbonas the currency, with an unwritten assumption that fruit DMC isconstant. The results presented here clearly show that fruit DMCcan vary from fruit to fruit and from orchard sample to orchardsample. Additionally, mean fruit DMC of ‘Cox’s Orange Pippin’apples from the same trees varied from 159 to 206 g kg−1 overa 9-year period.28 Similarly fruit DMC of ‘Golden Delicious’ variedfrom 139 to 176 g kg−1 in two consecutive seasons.29

Relationship between fruit DMC and TSSThe results presented here show a clear relationship betweenfruit DMC recorded at harvest and TSS after storage demonstratedacross several cultivars. This therefore confirms and extends earlierwork,20 which was done solely with the cultivar ‘Royal Gala’. Thestrong relationship between fruit DMC and TSS, particularly whenmeasured on the same fruit, follows from the high proportion offruit dry matter contributed by soluble compounds, particularlysugars, in apple fruit. The proportion of dry matter contributed bysugars, starch and organic acids in fresh apples of seven cultivarswas reported to range from 71% to 82%, with a mean of 78%.30 Anine-apple cultivar study showed a similar range.31 Most apples areharvested before starch solubilisation is complete, and thereforeTSS at harvest may not be a good predictor of TSS after storage, ashas been clearly shown by this work.

Relationship between fruit DMC and flesh firmnessThe linkage between flesh firmness and fruit DMC on these dataappears be very cultivar specific. We might hypothesise thatfruit with DMC would have more cell wall material and/or ahigher osmotic potential and therefore greater flesh firmness.Fruit maturity, however, is likely to be a major determinant of fleshfirmness, so that with both cultivars the relationship between fleshfirmness at harvest and flesh firmness after storage accounted fora greater proportion of the variation in flesh firmness after storagethan fruit DMC at harvest. Nevertheless, with ‘Scifresh’, fruit withhigh DMC at harvest were associated with fruit with greater fleshfirmness. This connection between flesh firmness and fruit DMCin ‘Royal Gala’ needs further attention, particularly consideringthe stronger relationship between these two variates seen in theorchard samples taken for the sensory study in 2008 than in theorchard samples taken in 2005.

Consumer studiesThis study provides evidence of a good relationship betweenDMC and consumer preference for ‘Royal Gala’ apples, with the

response being similar for fruit grown in two regions and intwo seasons. Given the robust relationship between DMC andconsumer preference, it is perhaps useful to speculate what DMCrepresents in terms of sensory quality. Accumulation of DMC isa fundamental measure of complex physiological and metabolicprocesses that contribute to the flow of carbon, nitrogen andminerals into fruit.13 This flux of material into the fruit willinclude (1) compounds that may contribute directly to flavour,(2) compounds that contribute to flavour as a result of metabolicprocesses that occur within the fruit and (3) compounds that will beincorporated in the structural elements of the cell associated withtexture. Thus a holistic measure such as DMC offers a complex arrayof potential physical–chemical interactions that individually and incombination contribute to sensory outcomes. For apples, firmnessand TSS are sometimes thought of as proxies for texture attributessuch as ‘crispness’ and taste attributes such as ‘sweetness’ and‘flavour’, respectively. However, caution is needed before makingthese assumptions here, as the flavour or texture characteristics inthese DMC samples can only be determined by a trained sensorypanel, which was not used in the current study. Trained panels areused to describe the individual textural and flavour characteristicsof fruit, whereas (untrained) consumer panels (as employed inthis study) are used to determine preference. At this stage, wesuspect that the consumer responses to DMC are not specificallydriven by TSS, as industry might be tempted to conclude. Rather,we speculate that DMC is a more holistic measure of quality.Dry matter concentration specifically represents the biologicalprocesses responsible for setting up the textural characteristics,carbohydrate status and flavour potential of the fruit.

Application of DMC as a new quality predictorIt must be emphasised that we are proposing fruit DMC as anew quality metric for apples. It is not to be confused with orthought of as a maturity metric. For both avocado and kiwifruit,fruit DMC has been proposed as both a quality metric and a harvestmaturity metric. Just before and during harvest of apple, maturityvariates such as background colour, SPI and internal ethyleneconcentration can change rapidly, but fruit DMC shows only minorchange over this period of time.32 Fruit DMC can therefore bemeasured before or at harvest, and used to predict the sensorypotential for the fruit after many months of storage. In contrast,measurements such as firmness and titratable acidity are moredynamic and change appreciably during maturation and storage.The dynamic nature of these latter indices makes it difficult topredict storage responses from their measurement at harvest.However, this does not mean that the traditional harvest indicesare redundant, as they are indicators of harvest maturity; instead,DMC can be viewed as a complementary quality index. Fruit DMCcan be used to compare and contrast the sensory potential fordifferent orchard blocks before or at harvest, while firmness andother quality indices can be used to monitor the progression ofthe fruit during the harvest window and in storage, to determine ifthe sensory potential is realised in the marketplace. For example,a high DMC fruit will only achieve its high sensory potential if itis harvested at the correct stage of maturity and then stored in amanner in which firmness and acidity are optimally conserved. Itis very unlikely that high DMC will compensate for poor texture.

CONCLUSIONSThis work has highlighted the variation in fruit DMC that canoccur within individual cultivars and also between cultivars,


25

94


but we believe that fruit DMC will find its prime usefulness indistinguishing between fruit from the same cultivar, rather thandistinguishing between different cultivars. The innate geneticcontrol within individual cultivars of characters such as size, shape,colour, sugar–acid balance, flavour and crispness will determinethe potential for that cultivar to appeal to individual customers,but we believe that high dry matter fruit of any cultivar will proveto be more acceptable to the customer than low dry matter fruit.

ACKNOWLEDGEMENTSThe authors would like to gratefully acknowledge the major inputfrom our colleagues Murray Oliver, Robert Diack, Daya Dayatilake,Shona Seymour, Robert Henriod, Ken Breen, Amy Paisley, MarkWohlers, Michelle Beresford and Judith Bowen; help from NewZealand apple growers, particularly Heartland Fruit in Nelson andFruitpackers Cooperative (Hawke’s Bay) Ltd; and funding fromPipfruit New Zealand Inc. and the New Zealand Foundation forScience Research and Technology (Contract No. CO6X0705).

REFERENCES1 Anon, State of the plate 2005: Study on America’s consumption of fruits

and vegetables. Produce for Better Heath Foundation, Wilmington,DE (2005).

2 Harker FR, Gunson FA and Jaeger SR, The case for fruit quality: aninterpretive review of consumer attitudes, and preferences forapples. Postharvest Biol Technol 28:333–347 (2003).

3 Jaeger SR, Rossiter KL, Wismer WV and Harker FR, Consumer-drivenproduct development in the kiwifruit industry. Food Qual Pref14:187–198 (2003).

4 Harker FR, Kupferman EM, Marin AB, Gunson FA and Triggs CM, Eatingquality standards for apples based on consumer preferences.Postharvest Biol Techmol 50:70–78 (2008).

5 Herregods M, Postharvest market quality preferences for fruit andvegetables. Acta Hortic 518:207–212 (2000).

6 Crisosto CH, Garner D, Crisosto GM and Bowerman E, Increasing‘Blackamber’ plum (Prunus salicina Lindell) consumer acceptance.Postharvest Biol Technol 34:237–244 (2004).

7 McCluskey JJ, Mittelhammer RC, Marin AB and Wright KS, Effect ofquality characteristics on consumers’ willingness to pay for Galaapples. Can J Agric Econ 55:217–231 (2007).

8 Harker FR, Maindonald J, Murray SH, Gunson FA, Hallett IC andWalker SB, Sensory interpretation of instrumental measurements.1. Texture of apple fruit. Postharvest Biol Technol 24:225–239 (2002).

9 Harker FR, Marsh KB, Young H, Murray SH, Gunson FA and Walker SB,Sensory interpretation of instrumental measurements. 2. Sweet andacid taste of apple fruit. Postharvest Biol Technol 24:241–250 (2002).

10 Mehinagic E, Royer G, Symoneaux R, Bertrand D and Jourjon F,Prediction of the sensory quality of apples by physicalmeasurements. Postharvest Biol Technol 34:257–269 (2004).

11 Molina-Delgado D, Alegrec S, Barreirod P, Valerod C, Ruiz-Altisent Mand Recasensa I, Addressing potential sources of variation in severalnon-destructive techniques for measuring firmness in apples.Biosyst Eng 104:33–46 (2009).

12 Hertog MLATM, Nicolai BM, De Ketelaere B, Lammertyn J and DeBaerdemaeker J, Non-destructive techniques and quality modelsfor the supply chain: a review. Acta Hortic 768:375–384 (2008).

13 Harker FR, Carr BT, Lenjo M, MacRae EA, Wismer WV, Marsh KB, et al,Consumer liking for kiwifruit flavour: a meta-analysis of five studieson fruit quality. Food Qual Pref 20:30–41 (2009).

14 Blankenship SM, Parker M and Unrath CR, Use of maturity indicesfor predicting poststorage firmness of ‘Fuji’ apples. HortScience32:909–910 (1997).

15 Knee M, Hatfield SGS and Smith SM, Evaluation of various indicatorsof maturity for harvest of apple fruit intended for. J Hortic Sci64:403–411 (1989).

16 Kingston CM, Maturity indices for apple and pear. Hortic Rev13:407–432 (1992).

17 Rowan DD, Hunt MB, Alspach PA, Whitworth CJ and Oraguzie NC,Heritability and genetic and phenotypic correlations of apple(Malus domestica) fruit volatiles in a genetically diverse breedingpopulation. J Agric Food Chem 57:7944–7952 (2009).

18 Altisent R, Echeverria G, Lara I, Lopez ML and Graell J, Shelf-life of‘Golden Reinders’ apples after ultra low oxygen storage: effect onaroma volatile compounds, standard quality parameters, sensoryattributes and acceptability. Food Sci Technol Int 15:481–493 (2009).

19 Gamble J, Harker FR, Jaeger SR, White A, Bava C, Beresford M, et al, Theimpact of dry matter, ripeness and internal defects on consumerperceptions of avocado quality and intentions to purchase.Postharvest Biol Technol 57:35–43 (2010).

20 McGlone VA, Jordan RB, Seelye R and Clark CJ, Dry matter – a betterpredictor of the post-storage soluble solids in apples? PostharvestBiol Technol 28:431–435 (2003).

21 Perring MA, Changes in dry matter concentration gradients in storedapples. J Sci Food Agric 46:439–449 (1989).

22 Breslow NE and Clayton DG, Approximate inference in generalizedlinear mixed models. J Am Statist Assoc 88:9–25 (1993).

23 Moons E, Sinnaeve G and Dardenne P, Non-destructive visible and NIRspectroscopy measurement for the determination of apple internalquality. Acta Hortic 517:441–448 (2000).

24 Haynes RJ and Goh KM, Variation in the nutrient content of leaves andfruit with season and crown position for two apple varieties. Aust JAgric Res 31:739–748 (1980).

25 Planchon V, Lateur M, Dupont P and Lognay G, Ascorbic acid levels ofBelgian apple genetic resources. Sci Hortic 100:51–61 (2004).

26 Lata B, Relationship between apple peel and the whole fruitantioxidant content: year and cultivar variation. J Agric Food Chem55:663–671 (2007).

27 Roen D, Ekholm A and Rumpunen K, Estimating useful diversity inthe Norwegian core collection of apples. Acta Hortic 814:131–136(2009).

28 Perring MA, The technique and application of fruit analysis, inProceedings of the 8th International Colloquium on Plant Analysisand Fertilizer Problems, Auckland, New Zealand NZ DSIR InformationSeries No 134, ed. by Ferguson AR, Bieleski RL and Ferguson IB.Government Printer, Wellington, pp. 375–382 (1978).

29 Marguery P and Sangwan BS, Sources of variation between applefruits within a season, and between seasons. J Hortic Sci 68:309–315(1993).

30 Suni M, Nyman M, Eriksson NA, Bjork L and Bjorck I, Carbohydratecomposition and content of organic acids in fresh and storedapples. J Sci Food Agric 80:1538–1544 (2000).

31 Salo ML and Korhonen I, Carbohydrate and acid composition of someapple varieties. J Sci Agric Soc Finl 44:63–67 (1972).

32 Bowen JH and Watkins CB, Fruit maturity, carbohydrate and mineralcontent relationships with watercore in ‘Fuji’ apples. PostharvestBiol Technol 11:31–38 (1997).


25

95

Research ArticleReceived: 13 April 2010 Revised: 13 July 2010 Accepted: 15 July 2010 Published online in Wiley Online Library: 17 August 2010


Characterization and discriminationof premium-quality, waxy,and black-pigmented rice based on odor-activecompoundsDong Sik Yang,a,b Kyu-Seong Leec and Stanley J. Kaysa∗

Abstract

BACKGROUND: Odor-active compounds have been studied in cooked aromatic rice, but not in specialty rice types that havedistinctly different flavors. We analyzed the odor-active compounds emanating from three different types of specialty rice(premium-quality, waxy and black-pigmented) and identified the differences in odor-active compounds among them.

RESULTS: Twenty-one, 21 and 23 odorants were detected using GC-O for cooked samples of premium-quality, waxy and black-pigmented rice cultivars, respectively. Hexanal was the main odorant in premium-quality and waxy cultivars; however, waxycultivars had 16 times higher hexanal odor activity values (OAVs) than premium-quality cultivars, indicating premium-qualityrice had a less pronounced overall aroma. 2-Acetyl-1-pyrroline was the main contributor to overall aroma in black-pigmentedrice, followed by guaiacol. The three types of specialty rice were clearly discriminated based on the OAVs of their odor-activecompounds using multivariate analyses. Six odor-active compounds [2-acetyl-1-pyrroline, guaiacol, hexanal, (E)-2-nonenal,octanal and heptanal] contributed the most in discriminating the three types of specialty rice. Six very similar superior cultivarsof premium rice could likewise be readily separated using aroma chemistry.

CONCLUSION: The ability to discriminate the aroma among rice types using the OAVs of the principal odor-active compoundsfacilitates our understanding of the aroma chemistry of specialty rice.c© 2010 Society of Chemical Industry

Keywords: specialty rice; hierarchical cluster analysis (HCA); principal component analysis (PCA); GC-O; plant breeding; selection criteria

INTRODUCTIONSpecialty rice is a term used to distinguish cultivars of rice thathave unique properties (e.g. flavor, color, nutrition, chemicalcomposition). The demand has been increasing in recent yearsfor various types of specialty rice, which are sold for as much as50% more than traditional rice cultivars.1 They are widely grownin India, Pakistan and Thailand, and are popular in Asia, the MiddleEast, Europe, and the United States.2 Examples of types of specialtyrice are: aromatic (unique aromas); premium (glossiness, stickiness,and smooth texture); black (unique color and flavor); and waxy(very low amylose content and superior processing quality).1,3,4

New rice cultivars are developed in breeding programs wherethe progeny are screened for a wide range of quantitativeand qualitative traits. The criteria used in evaluating ricevaries among programs, reflecting differences in preferencesamong consumer populations. Many traits can be fairly readilyand accurately quantified. For example, traits such as millingquality, grain size, shape, color and appearance, and cookingcharacteristics (e.g. gelatinization temperature, amylose content,grain elongation, aroma), are commonly used in making progenyselection decisions.5,6 Assessment of other traits (e.g. flavor, insectresistance) tend to be much more difficult and are often moresubjective. Flavor, for example, requires the use of sensory panels

that typically require 14–15 individuals and generally only fiveto six samples can be accurately tested at one setting. As aconsequence, the time, cost, and subjectivity of traditional sensoryanalysis tremendously limit the number of progeny that can beassessed. The net effect is that, out of necessity, flavor is generallyrelegated to one of the last traits in the selection hierarchy, greatlydecreasing the rate at which improvement can be made. If sensoryanalysis could be replaced in initial progeny screening with a rapid,accurate analytical technique for the measurement of flavor, largenumbers of progeny could be screened, greatly facilitating thedevelopment of superior new cultivars.7

∗ Correspondence to: Stanley J. Kays, 1111 Plant Science Building, Departmentof Horticulture, University of Georgia, Athens, GA 30602, USA.E-mail: [email protected]

a 1111 Plant Science Building, Department of Horticulture, University of Georgia,Athens, GA 30602, USA

b Plant Biology Division, Samuel Roberts Noble Foundation, Ardmore, OK 73401,USA

c Green Growth & Future Strategy Team, RDA, Suwon 441-707, Republic of Korea


25

96

www.soci.org DS Yang, K-S Lee, SJ Kays

In rice, aroma is the primary contributor to the overall flavor.As a consequence, considerable interest has been focused onthe identification and quantification of the volatile compoundsemanating from cooked rice.8 Over 300 volatiles have beenreported using a cross-section of isolation techniques.9 – 13

Dynamic headspace trapping onto Tenax TA (2,6-diphenyleneoxide-based porous polymer), in particular, is widely used forstudying organic volatiles in air, food and plants due to its abilityto adsorb and desorb a wide range of compounds.14 Tenax TA,which has a low affinity for water, is used for collecting volatilesfrom high moisture samples such as cooked foods. A dynamicheadspace system using a Tenax trap in conjunction with shortpath thermal desorption can detect volatile organic compoundsin the ppb range.15

Among the volatiles identified in cooked rice, approximately20 odor-active compounds (e.g. 2-acetyl-1-pyrroline, hexanal, (E)-2-nonenal, 2-amino acetophenone) are known to contribute tothe unique flavor of a cross-section of rice cultivars.10,13,16 Forexample, the popcorn-like odor in aromatic rice is conferred by2-acetyl-1-pyrroline, while hexanal is thought to be associatedwith the development of a rancid odor in aged rice.9 2-Amino acetophenone, conferring medicinal and phenolic odors,contributes to the overall aroma of cooked brown rice.16

Recently the odor-active compounds in black-pigmented rice weredetermined and 2-acetyl-1-pyrroline and guaiacol were found tobe the primary components of its unique aroma.4 In addition,differences in the unique aromas of basmati, jasmine, black-pigmented, and other rice flavor types were explained using 13critical odor-active compounds.17 Characterization of odor-activecompounds, however, has been mainly limited to aromatic rice,even though other specialty rices have distinctly different flavors.A better understanding of the aroma chemistry of speciality ricewill facilitate the development of superior new cultivars in ricebreeding programs focusing on flavor.

As an initial step in developing an analytical method forscreening rice progeny for flavor, we characterized the odor-activecompounds emanating from three different types of specialtyrice (premium-quality, waxy and black-pigmented). Differencesin odor-active compounds affecting flavor quality among themain specialty rice types were assessed using hierarchical clusteranalysis (HCA) and principal component analysis (PCA).

MATERIALS AND METHODSMaterialsTen cultivars including six premium-quality [Oryza sativa L. japon-ica cv. Hwaseongbyeo (P1), Ilpumbyeo (P2), Gopumbyeo (P3),Taebongbyeo (P4), Chucheongbyeo (P5), and Samkwangbyeo(P6)], two waxy cultivars [O. sativa L. japonica cv. Hwaseonchal-byeo (W1), and O. sativa L. indica cv. Hangangchalbyeo (W2)], andtwo black-pigmented [O. sativa L. japonica cv. Heugjinjubyeo (B1),and Heugkwangbyeo (B2)] were grown at the National Instituteof Crop Science, Suwon, South Korea in 2006. Rice transplants ofeach cultivar were planted on 24 May. Heugjinjubyeo and Tae-bongbyeo were harvested on 22 September; Heugkwangbyeo,Hangangchalbyeo, Hwaseonchalbyeo and Hwaseongbyeo on 3October; Gopumbyeo, Ilpumbyeo, Chucheongbyeo and Samk-wangbyeo on 22 October. N, P, and K fertilizers were applied atthe rate of 9.0×105, 4.5×105, and 5.7×105 kg m−2, respectively.During growth, the average temperature was 21.7 ± 3.6 ◦C andthere was 1150 mm of precipitation. After harvest and milling,

the rice was sealed in glass containers and held at −20 ◦C untilanalysis.

ChemicalsAnalytical standards utilized for identification and quantificationwere: benzaldehyde, decane, decanal, (E)-2-decenal, guaiacol,heptanal, heptane, 2-heptanone, nonane, nonanal, octane, pen-tadecane, 1-pentanol, tridecane, tetradecane, undecane (Sigma-Aldrich Inc., St Louis, MO, USA); (E,E)-2,4-decadienal, hexanal,(E)-2-hexenal, 2-nonanone, (E)-2-octenal, 1-octen-3-ol (AldrichChem. Co., Milwaukee, USA); p-menthan-3-one, naphthalene, oc-tanal, p-xylene (Fluka Chem. Co., St. Louis, MO, USA); dodecane,(E,E)-2,4-nonadienal, (E)-2-nonenal and 2-pentylfuran (TCI Amer-ica, Portland, OR, USA); 3-octen-2-one (Alfa Aesar, Ward Hill, MA,USA). 2,4,6-Trimethylpyridine (TMP) was purchased from AldrichChemical Co. and is routinely used as a internal standard to quantifythe concentration of 2-acetyl-1-pyrroline.

Preparation of cooked riceRice samples (100 g) were cooked in distilled water (150 mL) for30 min at 100 ◦C in a specially constructed 1 L glass beaker withentry and exit ports. The entry and exit ports were tightly sealedwith aluminum foil during cooking. Due to differences in waterabsorption, 100 mL of distilled water was added to black ricesamples (100 g) and similarly cooked for 30 min as describedpreviously by Yang et al.18 Three replications of each cultivar wereanalyzed.

Dynamic headspace using Tenax trapIsolation of volatile compounds emanating from cooked ricewas performed using an all-glass dynamic headspace systemdescribed previously.4,17 The glass beaker with freshly cookedrice was immediately placed in a hot water bath and maintainedat 70 ◦C during sampling. Headspace volatiles emanating fromcooked rice samples were collected using a Tenax trap and avacuum sampling pump (Aircheck Sampler, Model 224-44XR; SKCInc., Eighty Four, PA, USA) that swept volatiles from the sampleheadspace. Incoming air was purified by passing it through acharcoal filter (Alltech Assoc. Inc., Deerfield, IL, USA; 1 cm i.d., 10 cmlong Pyrex glass tube with a 7 cm bed of charcoal) connected tothe entry port of the glass beaker at a rate of 150 mL min−1 for60 min. A 50 mL Erlenmeyer flask was placed between the exitport and the trap to collect any condensation. One milliliter of18.34 mg L−1 TMP solution in 0.1 mol L−1 HCl was injected intothe Erlenmeyer flask at the beginning of volatile collection. TMPwas chosen as an internal standard for quantification of 2-APdue to similar properties with 2-AP (e.g. basic, water solubility,volatility, stability, retention time).19

After sampling, the Tenax trap was connected to an automatedshort path thermal desorption system (Model TD-5; ScientificInstrument Services, Ringoes, NJ, USA) on the injector portof the gas chromatograph/mass spectrometer (GC/MS, Model6890N/5973; Agilent, Wilmington, DE, USA). The volatiles weredesorbed at 250 ◦C for 5 min with He at a flow rate of 10 mL min−1

and the analytes collected on the first 4 cm of the GC column usinga CO2 cooled cryofocus trap (−40 ◦C) (SIS 2 Cryo-Trap; ScientificInstrument Services). After desorption, the cryofocus trap wasrapidly heated to 200 ◦C and condensed analytes were separatedutilizing GC temperature programming.


25

97

Aroma of premium-quality, waxy and black-pigmented rice www.soci.org

GC/MS and GC-OlfactometryDesorbed odorants were split between the mass spectrometer foridentification and quantification and an Olfactory Detector Outlet(ODO II; SGE Intl., Austin, TX, USA) used to describe the individualodorants and to assess their intensity. The GC was equipped witha 30 m length, 0.25 mm i.d., 0.25 µm film thickness (5% phenylmethyl silicone) fused silica capillary column (HP-5 MS; Agilent).The injection port temperature was 225 ◦C with a split ratio of0.5 : 1. Helium was used as the carrier gas with a flow rate of2.0 mL min−1. The column temperature was programmed to holdat 40 ◦C for 1 min and then increase 1.5 ◦C min−1 to 65 ◦C for1 min, then 2 ◦C min−1 to 120 ◦C for 1 min, and finally 15 ◦C min−1

to 280 ◦C for 5 min. MS conditions were: ion source 230 ◦C; electronenergy 70 eV; multiplier voltage 1247 V; a transfer line 280 ◦C; anda scan range of 35–350 atomic mass units.

GC-O analysis was made by three assessors who had consid-erable previous experience in sensory tests. Acceptance of anodorant required substantiation by at least two of the three as-sessors with each individual assessing replicated samples of thecultivars.17 The assessors were trained by describing 15 materialswith different odors: popcorn-like (popcorn), starchy (rice starch),woody (toothpicks), cooked grain (cream of wheat), corn (creamstyle corn), nutty (roasted peanut), floral (jasmine scent), dairy(2% milk), hay (hay), barn (white pepper), buttery (butter), green(alfalfa), rancid (vegetable oil), waxy (candle) and earthy (mush-room) odors. An aroma extract of the sample was characterized bydescribing the aroma of the individual components and assessingtheir odor intensity on a scale of 1 to 5, where 1 = very weak;2 = weak; 3 = intermediate; 4 = strong; 5 = very strong. Odor-ants perceived by all three assessors were considered odor-activecompounds (odor description and intensity data are not shown).

Identification and quantification of odorantsOdorants were identified based on their mass spectra usingNIST 02 and Wiley 7 libraries. Identification was confirmedusing Kovats retention indices (RI) and odor descriptorswith authentic standards. RIs were determined using anon-polar HP-5 MS column and a series of n-hydrocarbons(C7 –C15) and compared with those reported previously andat http://webbook.nist.gov/chemistry/name-ser.html/. In the ab-sence of an authentic standard, the identification of 2-acetyl-1-pyrroline (2-AP) was confirmed by mass spectra, RI and itsdistinctive popcorn odor. The concentration of each odorant wasquantified using standard curves for each compound in hexane (5,10, 20, 50, 100, 200, and 500 µ L−1) and three replications of 1 µLof each standard solution was injected directly into the GC-MSusing a microsyringe. Linearity, sensitivity, and precision of eachcompound were shown in Yang et al.4 2-AP was expressed as TMPequivalents.

Odor thresholds in air and odor activity valuesOdor thresholds in air were determined using a modified Ulrichand Grosch GC-O method by the assessors to ascertain the impactof individual odorants.20 A solution of each odorant (10 mg) wasprepared in 10 mL of hexane and 0.5 µL of the solution wasinjected into the GC and sniffed using the olfactory detectoroutlet. The solution was diluted stepwise (1 : 1, v/v) until theodorant was not detected (minimum detectable concentration).The odor threshold in air for each odorant was calculatedrelative to minimum detectable concentration of (E)-2-decenalin air (2.7 ng L−1). The odor threshold for 2-AP was reported

previously.21 Odor activity values (OAVs) were calculated bydividing the emission concentration of each odorant in a sampleover a constant collection period by its odor threshold in air.

Statistical data analysisHierarchical cluster analysis (HCA) and principal componentanalysis (PCA) were carried out using the SAS system for Windowsv8 with three replications of each cultivar (six premium quality,two waxy and two black-pigmented cultivars). HCA was performedusing Ward minimum variance method to determine if the OAVsof the odor-active compounds in each specialty rice cultivars wereable to segregate the rice samples. PCA was performed using thecovariance matrix obtained from the data matrix (i.e. OAVs of theprimary odor-active compounds) to determine which odor-activecompounds had the highest loadings and contributed the mostto differences in aroma among the samples.

RESULTS AND DISCUSSIONOdor-active compounds and their odor activity valuesAmong the premium-quality cultivars (Hwaseongbyeo, Ilpum-byeo, Gopumbyeo, Taebongbyeo, Chucheongbyeo and Samk-wangbyeo) there were 21 odorants identified that included 12aldehydes, two alcohols, three ketones, and four aromatic com-pounds (Table 1). Based on the OAV, hexanal was the dominantodorant, followed by (E)-2-nonenal, octanal, heptanal, (E,E)-2,4-nonadienal and nonanal in all six cultivars. Collectively, thesecompounds accounted for over 95% of the total OAV and werekey aroma compounds in premium rice. However, these cultivarshad lower OAVs than waxy and black-pigmented rice cultivars,indicating that the aroma of the premium cultivars was lesspronounced.

The six premium-quality cultivars were selected in breedingprograms based on six primary quality traits: (1) grain length/widthratio prior to milling (1.7–2.0); (2) appearance (size 172 g per1000 milled grains), translucency, gloss and very low chalkiness;(3) physicochemical properties (amylose content below 20%,protein content of brown rice 7–9%, high alkali digestion value);(4) appearance and palatability when cooked (glossy appearance,intact grains, weakly aromatic fragrance, sticky, good taste andtexture); (5) slow retrogradation of cooked rice; and (6) millingrecovery from rough to polished rice of above 75% and a headrice ratio of above 90%.3 With regard to the palatability of cookedKorean premium-quality rice, our results indicated that the sixpremium-qualtiy cultivars have a less pronounced overall aromathat corresponds to the preference of Korean consumers.3

Twenty-one odorants were identified in the waxy rice culti-var Hwaseonchalbyeo (W1) and 20 in Hangangchalbyeo (W2)that included 12 aldehydes, one alcohol, three ketones, and fouraromatic compounds. 1-Pentanol was present only in Hwaseon-chalbyeo (Table 1). Hexanal in Hwaseonchalbyeo (316), a Japonicatype and Hangangchalbyeo (448), an Indica type, had the high-est OAVs, 8.8 and 12.4 times higher than that in Samkwangbyeo(P6) which had the highest OAV in the premium-quality cultivars(Table 1). The very high hexanal concentration appears to be dueto the higher lipid content in waxy rice and its subsequent degra-dation resulting in the formation of hexanal.22 Hexanal, a linoleicacid oxidation product, had a higher concentration in broken ricethan in head rice due to greater surface lipids and free fatty acids.The concentration of hexanal in broken rice increased signifi-cantly during storage, resulting in the development of off-flavor.23

The odor of hexanal is described as green or green tomato


25

98


Tab

le1

.O

do

ract

ivit

yva

lue

(OA

V)a

ofo

do

r-ac

tive

com

po

un

ds

insi

xp

rem

ium

-qu

alit

y,tw

ow

axy,

and

two

bla

ck-p

igm

ente

dri

cecu

ltiv

ars

Prem

ium

-qu

alit

yri

ceb

Wax

yri

cec

Bla

ck-p

igm

ente

dri

ced

No

.O

do

ran

tRI

eID

fP1

P2P3

P4P5

P6W

1W

2B

1B

2O

Tg

11-

Pen

tan

ol

766

A0.

01±

0.00

trh

0.01

±0.

000.

01±

0.00

0.01

±0.

00tr

0.02

±0.

00N

D0.

01±

0.00

0.01

±0.

0015

3i

2H

exan

al80

3A

15.5

3±

4.79

13.4

3±

5.55

22.4

9±

2.08

31.7

0±

8.30

26.6

7±

0.30

35.2

6±

12.1

531

5.52

±4.

3744

7.53

±9.

3988

.52

±3.

8836

.61

±8.

421.

1i

3(E

)-2-

Hex

enal

857

A0.

05±

0.01

0.03

±0.

020.

05±

0.01

0.05

±0.

010.

06±

0.01

0.04

±0.

030.

15±

0.08

0.30

±0.

000.

10±

0.01

0.11

±0.

033.

1i

4p-

Xyl

ene

859

A0.

01±

0.00

0.01

±0.

000.

01±

0.00

0.01

±0.

000.

01±

0.00

0.01

±0.

000.

01±

0.00

0.03

±0.

000.

02±

0.00

0.06

±0.

0146

.9

52-

Hep

tan

on

e89

5A

0.16

±0.

030.

10±

0.10

0.20

±0.

030.

25±

0.05

0.23

±0.

030.

24±

0.05

1.14

±0.

262.

29±

0.01

0.68

±0.

000.

93±

0.00

3.5i

6H

epta

nal

903

A1.

08±

0.23

1.37

±0.

491.

50±

0.20

1.82

±0.

441.

43±

0.02

2.14

±0.

485.

22±

1.22

7.35

±0.

032.

73±

0.07

4.00

±0.

780.

9i

72-

Ace

tyl-

1-p

yrro

line

918

BN

DN

DN

DN

DN

DN

DN

DN

D56

9.24

±94

.96

285.

60±

114.

740.

02j

8(E

)-2-

Hep

t-an

al95

2A

0.09

±0.

050.

07±

0.06

0.15

±0.

040.

17±

0.04

0.14

±0.

010.

19±

0.03

0.52

±0.

100.

74±

0.03

ND

ND

13.4

9B

enza

ldeh

yde

954

AN

DN

DN

DN

DN

DN

DN

DN

D0.

01±

0.01

0.05

±0.

0285

i

101-

Oct

en-3

-ol

984

A0.

14±

0.05

0.13

±0.

070.

21±

0.05

0.27

±0.

070.

20±

0.00

0.23

±0.

031.

19±

0.28

2.73

±0.

380.

51±

0.06

0.24

±0.

192.

7i

112-

Pen

tylfu

ran

992

A0.

07±

0.01

0.05

±0.

010.

08±

0.01

0.09

±0.

030.

08±

0.00

0.10

±0.

020.

30±

0.08

0.53

±0.

050.

15±

0.00

0.10

±0.

0219

i

12O

ctan

al10

05A

1.52

±0.

232.

22±

0.81

2.16

±0.

433.

30±

0.82

2.18

±0.

033.

13±

0.97

11.2

4±

2.67

6.80

±0.

133.

87±

0.12

5.02

±1.

180.

4i

133-

Oct

en-2

-o

ne

1036

A0.

02±

0.01

0.01

±0.

000.

04±

0.02

0.05

±0.

010.

01±

0.00

0.02

±0.

010.

14±

0.04

0.27

±0.

030.

05±

0.03

0.08

±0.

046.

7i

14(E

)-2-

Oct

enal

1058

A0.

31±

0.10

0.24

±0.

090.

39±

0.05

0.42

±0.

080.

33±

0.01

0.42

±0.

081.

55±

0.34

1.24

±0.

710.

60±

0.34

0.56

±0.

122.

7i

15G

uai

aco

l10

86A

ND

ND

ND

ND

ND

ND

ND

ND

1.12

±0.

631.

07±

0.20

1.5i

162-

No

nan

on

e10

93A

trtr

trtr

trtr

0.01

±0.

000.

01±

0.00

0.01

±0.

00N

D31

i

17N

on

anal

1106

A0.

65±

0.18

1.14

±0.

411.

04±

0.35

1.68

±0.

401.

02±

0.10

1.47

±0.

503.

48±

0.91

2.26

±0.

052.

37±

0.26

2.94

±0.

682.

6i

18p-

Men

than

-3-

on

e11

52A

0.02

±0.

010.

02±

0.02

0.04

±0.

030.

01±

0.01

0.04

±0.

030.

04±

0.03

0.03

±0.

010.

33±

0.04

0.03

±0.

000.

10±

0.09

4.7i

19(E

)-2-

No

nen

al11

60A

3.32

±0.

583.

58±

1.24

4.51

±1.

475.

77±

0.83

4.66

±0.

536.

37±

1.25

10.4

2±

1.72

13.3

8±

1.04

5.46

±1.

087.

29±

1.82

0.09

i

20N

aph

thal

ene

1172

A0.

02±

0.01

0.03

±0.

010.

03±

0.01

0.03

±0.

010.

03±

0.00

0.03

±0.

010.

02±

0.01

0.03

±0.

000.

04±

0.01

0.04

±0.

0133

21D

od

ecan

e12

00A

ND

ND

ND

ND

ND

ND

trtr

ND

ND

202

22D

ecan

al12

06A

0.12

±0.

040.

15±

0.07

0.16

±0.

060.

23±

0.08

0.16

±0.

020.

26±

0.06

0.33

±0.

060.

22±

0.04

0.27

±0.

030.

38±

0.07

2.6i

23(E

,E)-

2,4-

No

nad

ien

al12

12A

0.89

±0.

120.

80±

0.10

1.27

±0.

191.

23±

0.11

1.04

±0.

061.

22±

0.05

2.38

±0.

841.

86±

0.50

1.64

±0.

071.

79±

0.04

0.2i

24(E

)-2-

Dec

enal

1262

A0.

07±

0.02

0.07

±0.

010.

09±

0.03

0.10

±0.

010.

09±

0.02

0.13

±0.

020.

22±

0.06

0.11

±0.

020.

07±

0.01

0.09

±0.

042.

7i

25(E

,E)-

2,4-

Dec

adie

nal

1315

A0.

24±

0.07

0.16

±0.

060.

25±

0.10

0.36

±0.

040.

30±

0.02

0.40

±0.

010.

34±

0.11

0.35

±0.

020.

14±

0.03

0.23

±0.

082.

3i

aO

AV

sar

eg

iven

asth

em

ean

±st

and

ard

dev

iati

on

oft

hre

ere

plic

atio

ns

ofe

ach

cult

ivar

.b

P1,H

was

eon

gb

yeo

;P2,

Ilpu

mb

yeo

;P3,

Go

pu

mb

yeo

;P4,

Taeb

on

gb

yeo

;P5,

Ch

uch

eon

gb

yeo

;P6,

Sam

kwan

gb

yeo

.c

B1,

Heu

gjin

jub

yeo

;B2,

Heu

gkw

ang

bye

o.

dW

1,H

was

eon

chal

bye

o;W

2,H

ang

ang

chal

bye

o.

eRe

ten

tio

nin

dex

bas

edo

nH

P-5

MS

colu

mn

usi

ng

ase

ries

ofn

-hyd

roca

rbo

ns.

fM

eth

od

ofi

den

tific

atio

n:A

,co

mp

ou

nd

sw

ere

iden

tifie

db

yco

mp

aris

on

ofm

ass

spec

tru

man

dRI

ofa

np

erce

ived

com

po

un

dw

ith

tho

seo

fan

auth

enti

cco

mp

ou

nd

.;B

,co

mp

ou

nd

sw

ere

ten

tati

vely

iden

tifie

db

yco

mp

aris

on

ofm

ass

spec

tru

m,o

do

rdes

crip

tio

n,a

nd

RIw

ith

tho

sefr

om

the

liter

atu

rew

hen

refe

ren

ceco

mp

ou

nd

sw

ere

no

tav

aila

ble

.g

OT,

od

ort

hre

sho

ldin

air

(ng

L−1).

htr

,tra

ceva

lue

(<0.

01);

ND

,no

td

etec

ted

.iD

ata

fro

mY

ang

etal

.17

jD

ata

fro

mSc

hie

ber

le.21


25

99


and has been associated with rancidity and consumer rejectionof rice.9 Monsoor and Proctor23 suggested that water washingof rice grains can reduce the off-flavor products includinghexanal. In addition to hexanal, (E)-2-nonenal, octanal, heptanal,nonanal, (E,E)-2,4-nonadienal, (E)-2-octenal, 1-octen-3-ol, and 2-heptanone had OAVs greater than 1 (Table 1). OAVs of octanal inHwaseonchalbyeo and heptanal in Hanganchalbyeo were 4.7 and4.6 times higher than the average OAVs of octanal and heptanalin the six premium-quality cultivars, respectively.

Twenty-three odorants were identified in the black-pigmentedcultivars Heugjinjubyeo (B1) and 22 in Heugkwangbyeo (B2)(Table 1). Eleven aldehydes, two alcohols, two ketones, sixaromatics and one nitrogen-containing compound were foundin all cultivars. 2-Nonanone was present in Heugjinjubyeo butnot Heugkwangbyeo. In Heugjinjubyeo and Heugkwangbyeo, 2-acetyl-1-pyrroline (2-AP) had the highest OAV (e.g. 569 and 286,respectively) followed by hexanal, (E)-2-nonenal, octanal, heptanal,nonanal, (E,E)-2,4-nonadienal and guaiacol. 2-AP, which confersa popcorn-like odor, was a significant component in the aromaof the cultivar. The concentration of 2-AP, used as an indicatorof aromatic rice during selection in some breeding programs,varied substantially between the two black rice cultivars from alow of 5.71 ng g−1 in Heugnambyeo to a high of 11.38 ng g−1 inHeugjinjubyeo (data not shown). According to a previous report,there was a wide variation in 2-AP concentration among aromaticcultivars.24 Guaiacol was detected only in black rice where it is akey aroma compound that confers a ‘smoked odor’ and has a lowodor threshold.4 It had an OAV of 1.12 and 1.07 in Heugjinjubyeo,and Heugkwangbyeo, respectively. Even though hexanal had highOAVs, guaiacol and 2-AP contributed to the unique aroma of blackrice due to their unique odors and low odor threshold values.

Rice aroma is affected by genetic, preharvest and postharvest(e.g. drainage and harvest dates, harvest moisture content, degreeof milling, storage conditions, washing, soaking, cooking method)factors.25 The different aroma patterns of cultivars among eachrice type (premium-quality, waxy, and black-pigmented rice types)appear to be predominantly due to genetic differences sinceall cultivars were cultivated, harvested, and stored the same.Genetic differences in existing compounds (e.g. 2-AP) and thosesynthesized during cooking from the basic chemical componentsof rice (e.g. lipids, proteins, starch) largely dictated the differencesamong cultivars in the volatiles released from cooked rice.18

HCA and PCA of the odor-active compounds in specialty riceThe dendogram obtained by HCA demonstrated distinct separa-tion of three different classifications in flavor among samples thatcoincided with the three types of specialty rice (Fig. 1): cluster 1(premium-quality rice; P1 to P6), cluster 2 (waxy rice; W1 and W2),and cluster 3 (black-pigmneted rice; B1 and B2). Premium-quality(P1 to P6), waxy (W1 and W2), and black-pigmented (B1 and B3)types were clearly separated with the first (PC 1) and second (PC2) principal components accounting for 62.8% and 37.2% of thetotal variance, respectively (Fig. 2), confirming the clustering ob-tained from HCA. The black-pigmented cultivars positioned on thepositive side of PC 1, were clearly separated from the premium-quality and waxy cultivars positioned on the negative side ofPC 1. 2-Acetyl-1-pyrroline (7) was the most important contribu-tor influencing the separation of black-pigmented cultivars frompremium-quality and waxy cultivars, followed by guaiacol (15).The premium-quality cultivars positioned on the negative side ofPC 2 were clearly separated from the waxy cultivars positionedon the positive side of PC 2. The main compounds influencing

the differences on PC 2 were, in decreasing order of contribution,hexanal (2), (E)-2-nonenal (19), octanal (12) and heptanal ( 6).

Potential for using aroma chemistry to make progeny selectiondecisions in breeding programsThe potential for using aroma chemistry for making progenyselection decisions in rice breeding programs is predicated onbeing able to screen large numbers of progeny and accuratelyidentify those with desirable characteristics. The six cultivars ofpremium rice in the test that were very uniform in their overallflavor as indicated by their very tight clustering in Fig. 2, werereadily separated using HCA (Fig. 3). HCA indicated three differentflavor groups within the cultivars: cluster 1 (P1 and P2), cluster 2 (P3and P5), and cluster 3 (P4 and P6) (Fig. 3) with cluster 1 being moredistant than clusters 2 and 3. Figure 4 illustrates the pattern ofodor-active compounds for the six cultivars with PC 1, accountingfor 99.6% of the total variance. The separations were matched toclusters obtained from HCA. The main compound influencing PC1 was hexanal (2). Based on PC 2, the groups were separated byheptanal (6), octanal (12), nonanal (17), and (E)-2-nonenal (19);however, PC 2 explained only 0.3% of total variance, indicatingtheir contribution is very weak.

The method allowed precise separation of the premium-qualityrice cultivars into three groups, which was confirmed by HCA andPCA. It should be noted that in a breeding program, one of theprimary advantages of using an analytical method for makingselection decisions is to allow screening large numbers of progenyfor flavor, in contrast to the very restricted number that can becurrently assessed using sensory analysis. The cultivars testedrepresent superior lines with very desirable flavors. In a breedingprogram, the range in flavor chemistry among progeny will befar greater than the cultivars tested herein, and therefore, lessfavorable lines will be separated much more distinctly from theexisting cultivars. Screening a far greater number of progeny toidentify desirable lines facilitates the development of new cultivarswith superior flavor. Selection decisions are also made using otheranalytical assessments methods (e.g. texture, stickiness, amyloseto amylopectin ratio) that are currently widely used. An analyticalmethod for assessing flavor does not totally eliminate the use ofsensory analysis; however, it allows testing only a small number ofadvanced lines prior to final acceptance of potential new cultivars.

CONCLUSIONSThe OAV of odor-active compounds was useful in classifyingthe ten cultivars into three groups (premium-quality, waxy, andblack-pigmented) (Figs. 1 and 2). Six odor-active compounds [2-acetyl-1-pyrroline, guaiacol, hexanal, (E)-2-nonenal, octanal, andheptanal] contributed the most in distinguishing differences inaroma among the rice types. The ability to discriminate the aromaamong rice types using the OAVs of odor-active compounds willfacilitate defining the flavor chemistry contributing to differencesin consumer flavor preference. The methods employed alsodemonstrated distinct differences within rice types (i.e. premium).If sufficiently precise separation can be made among progenywithin a type of specialty rice, the analytical method describedfor assessing the aroma component of flavor could potentiallybe used for making progeny selection decisions in rice breedingprograms, characterizing the chemistry of preference of targetedconsumer populations and greatly accelerating the developmentof superior new cultivars.


26

00


Figure 1. HCA dendrogram of ten cooked specialty rice cultivars based on their average odor active values (OAVs) (n = 3): premium-quality rice (P1-P6),waxy rice (W1 and W2), and black-pigmented rice (B1 and B2).

Figure 2. PCA biplot of odor-active compounds in cooked specialty rices based on their average odor active values (OAVs) (n = 3). Numbers correspondto compounds listed in Table 1.

Figure 3. HCA dendrogram of six cooked premium-quality rice cultivars based on their average odor active values (OAVs) (n = 3): Hwaseongbyeo (P1),Ilpumbyeo (P2), Gopumbyeo (P3), Taebongbyeo (P4), Chucheongbyeo (P5), Samkwangbyeo (P6).


26

01


Figure 4. PCA biplot of odor-active compounds in cooked premium-quality rices based on their average odor active values (OAVs) (n = 3). Numberscorrespond to compounds listed in Table 1.

ACKNOWLEDGEMENTSThe authors wish to express their appreciation to Drs H.P. Moonand S.K. Hahn for their contributions to our rice flavor chemistryresearch program.

REFERENCES1 Chaudhary RC, Speciality rices of the world: effect of WTO and IPR on

its production trend and marketing. J Food Agric Environ 1:34–41(2003).

2 Singh RK, Singh US and Khush GS, Prologue, in Aromatic Rices, ed. bySingh RK, Singh US and Khush GS. Science Publishers, Enfield, NH,pp. 1–3 (2000).

3 Choi HC, Current status and perspectives in varietal improvement ofrice cultivars for high-quality and value-added products. Kor J CropSci 47:15–32 (2000).

4 Yang DS, Lee KS, Jeong OY, Kim KJ and Kays SJ, Characterization ofvolatile aroma compounds in cooked black rice. J Agric Food Chem56:235–240 (2008).

5 Abdel-Aal ESM, Young JC and Rabalski I, Anthocyanin composition inblack, blue, pink, purple, and red cereal grains. J Agric Food Chem54:4696–4704 (2006).

6 Dela Cruz N and Khush GS, Rice grain quality evaluation procedures,in Aromatic Rices, ed. by Singh RK, Singh US and Khush GS. SciencePublishers, Enfield, NH, pp. 15–28 (2000).

7 Yang DS, Lee KS, Jeong OY, Kim KJ and Kays SJ, Fingerprinting riceflavor, in Improving Human Health Through Biofortified Rice.International Symposium on Rice Fortification. National Instituteof Crop Science, Suwon, S. Korea, pp. 71–95 (2006).

8 Champagne E and Bett-Garber K, Challenges of measuring ricearoma and flavor, in Proceedings of the United States –Japan UJNRCooperativePrograminNaturalResources,Food and AgriculturePanel,36th Annual Meeting, National Food Research Institute, Tsukuba,Japan, pp. 217–222 (2007).

9 Bergman CJ, Delgado JT, Bryant R, Grimm C, Cadwallader KR andWebb BD, Rapid gas chromatographic technique for quantifying2-acetyl-1-pyrroline and hexanal in rice (Oryza sativa L.). CerealChem 77:454–458 (2000).

10 Widjaja R, Craske JD and Wootton M, Comparative studies on volatilecomponents of non-fragrant and fragrant rices. J Sci Food Agric70:151–161 (1996).

11 Tava A and Bocchi S, Aroma of cooked rice (Oryza sativa): comparisonbetween commercial Basmati and Italian line B5-3. Cereal Chem76:526–529 (1999).

12 Grimm CC, Bergman C, Delgado JT and Bryant R, Screening for 2-acetyl-1-pyrroline in the headspace of rice using SPME/GC-MS.J Agric Food Chem 49:245–249 (2001).

13 Buttery RG, Turnbaugh JG and Ling LC, Contribution of volatiles to ricearoma. J Agric Food Chem 36:1006–1009 (1988).

14 Reineccius G, Flavor Chemistry and Technology, 2nd edition. Taylor andFrancis, Boca Raton, FL (2006).

15 Peng CY and Batterman S, Performance evaluation of a sorbent tubesampling method using short path thermal desorption for volatileorganic compounds. J Environ Monit 2:313–324 (2000).

16 Jezussek M, Juliano BO and Schieberle P, Comparison of key aromacompounds in cooked brown rice varieties based on aroma extractdilution analysis. J Agric Food Chem 50:1101–1105 (2002).

17 Yang DS, Shewfelt RL, Lee KS and Kays SJ, Comparison of odor-activecompounds from six distinctly different rice flavor types. J AgricFood Chem 56:2780–2787 (2008).

18 Yang DS, Lee KS, Kim KJ and Kays SJ, Site of origin of volatilecompounds in cooked rice. Cereal Chem 85:591–598 (2008).

19 Buttery RG, Ling LC and Thomas RM, Quantitative analysis of 2-acetyl-1-pyrroline in rice. J Agric Food Chem 34:112–114 (1986).

20 Ullrich F and Grosch W, Identification of the most intense volatileflavour compounds formed during autoxidation of linoleicacid. Zeitschrift fur Lebensmitteluntersuchung und -Forschung A184:277–282 (1987).

21 Schieberle P, Primary odorants in popcorn. J Agric Food Chem39:1141–1144 (1991).

22 Shin HS and Rhee JY, Comparative studies on the lipid content andneutral lipid composition in nonglutinous and glutinous rice. Kor JFood Sci Technol 18:1337–142 (1986).

23 Monsoor MA and Proctor A, Volatile component analysis ofcommercially milled head and broken rice. J Food Sci 69:C632–C636(2004).

24 Buttery RG, Juliano BO and Ling LC, Cooked rice aroma and 2-acetyl-1-pyrroline. J Agric Food Chem 31:823–826 (1983).

25 Champagne E, Rice aroma and flavor: a literature review. Cereal Chem85:445–454 (2008).


26

02

Research ArticleReceived: 11 May 2010 Revised: 14 July 2010 Accepted: 19 July 2010 Published online in Wiley Online Library: 5 August 2010


Relationships between disease control, greenleaf duration, grain quality and the productionof alcohol from winter wheatAndrew M Watson,a∗ Martin C Hare,a Peter S Kettlewell,a James M Brosnanb

and Reginald C Agub

Abstract

BACKGROUND: Since demand for distilling wheat is expected to increase rapidly as a result of the development of thebioethanol industry, efficient production will become of increasing importance. Achieving this will require an understandingof the agronomic factors that influence both grain yield and alcohol yield. Therefore five field experiments using the winterdistilling wheat variety Glasgow were conducted over three seasons (2006–2007, 2007–2008 and 2008–2009) to study therelationships between foliar disease and alcohol yield.

RESULTS: There was a significant relationship between alcohol yield and the severity of the disease septoria leaf blotch (Septoriatritici), which was present in the experiments from natural infection. Retention of green flag leaf area as affected by diseasecontrol following fungicide application was also shown to be important for achieving high alcohol yields. Measurements of grainquality showed that high thousand-grain weight and low grain protein concentration were significantly related to increasedalcohol yield.

CONCLUSION: The experiments showed the importance of disease management to protect alcohol yields in the distilling wheatcrop. Fungicides that provide greater disease control and improved green leaf retention are likely to be beneficial to alcoholyield.c© 2010 Society of Chemical Industry

Keywords: grain quality; alcohol; bioethanol; wheat; disease; green leaf duration; fungicide

INTRODUCTIONTraditionally, distilling wheat in the UK has been grown for theScotch whisky industry. This market requires an annual productionof 7×105 t of wheat to produce approximately 3×108 L of alcohol(LA) each year.1 There is growing interest in the productionof distilling wheat for the emerging fuel alcohol market. Fuelalcohol, commonly known as bioethanol, is produced using similartechniques to those used by the potable alcohol industry. Growinginterest in biofuels has led to concerns over the production oflarge areas of non-food crops threatening the availability andaffordability of food.2 There have also been concerns over theamount of CO2 emitted during production and processing of thefuel.3 If biofuels are to be produced, it is important to address theseconcerns by maximising efficiency both in the field and duringprocessing. Achieving a high level of alcohol production perhectare will help minimise the cultivated area of non-food crops,easing pressures on both food production and the environment.Maximising the efficiency of crop production can also help limit theamount of CO2 emitted during production. Achieving a high grainalcohol yield (LA t−1) will also help maximise processing efficiencyand reduce energy requirements. To achieve these objectives, itis necessary to have a thorough understanding of the agronomicfactors that influence alcohol production. Previous investigationshave studied relationships between disease control, green flag

leaf area duration and grain quality,4,5 but none has looked at theinfluence of these factors on alcohol yield. This study is thereforedesigned to expand on what is known to determine how thesefactors influence the production of alcohol.

MATERIALS AND METHODSExperimental designFive field experiments were conducted in Shropshire, UK at HarperAdams University College. During the study, experiment 1 wasestablished for the 2006–2007 growing season, experiments2 and 3 were conducted in the 2007–2008 growing seasonand experiments 4 and 5 were conducted in the 2008–2009growing season. Each experiment used the soft winter wheatvariety Glasgow recommended for distilling. All experiments wereconducted using a randomised block design with seven replicates.

∗ Correspondence to: Andrew M Watson, Crop and Environment Research Centre,Harper Adams University College, Newport, Shropshire TF10 8NB, UK.E-mail: [email protected]

a Crop and Environment Research Centre, Harper Adams University College,Newport, Shropshire TF10 8NB, UK

b Scotch Whisky Research Institute (SWRI), Riccarton, Edinburgh EH14 4AP, UK


26

03

Relationships affecting alcohol production from winter wheat www.soci.org

Table 1. Fungicide treatments applied at GS 59

Treatment Fungicide treatment at GS 59

1 Untreateda

2 Prothioconazole + tebuconazole 62.5 : 62.5 g ha−1

3 Prothioconazole + tebuconazole 62.5 : 62.5 g ha−1 +azoxystrobin 62.5 g ha−1b

4 Prothioconazole + tebuconazole 62.5 : 62.5 g ha−1 +azoxystrobin 125 g ha−1

5 Prothioconazole + tebuconazole 62.5 : 62.5 g ha−1 +azoxystrobin 187.5 g ha−1

6 Prothioconazole + tebuconazole 62.5 : 62.5 g ha−1 +azoxystrobin 250 g ha−1

a Received no fungicide at GS 59 (i.e. only received fungicides at GS 30,32 and 39).b Azoxystrobin 62.5 g ha−1 treatment was not present in experiment 1.

Each plot measured 12 m × 1.8 m with a 1.5 m buffer zonebetween replicates and a 0.2 m buffer zone between plots withineach replicate.

All plots in all five experiments received a fungicide applicationof 300 g L−1 metrafenone (Attenzo, BASF, Cheadle Hulme, UK)at 0.2 L ha−1 applied with 375 g L−1 chlorothalonil + 62.5 g L−1

propiconazole + 50 g L−1 cyproconazole (Cherokee, SyngentaCrop Protection UK Ltd, Fulbourn, UK) at 0.75 L ha−1 at growthstage (GS) 30 (pseudostem erect).6 A further 125 g L−1 epoxicona-zole (Opus, BASF) at 0.6 L ha−1 + 500 g L−1 chlorothalonil (Bravo500, Syngenta Crop Protection UK Ltd) at 1 L ha−1 was applied atboth GS 32 (second node detectable) and GS 39 (flag leaf blade allvisible). Experimental treatments were 125 g L−1 prothioconazole+ 125 g L−1 tebuconazole (Prosaro, Bayer CropScience, Milton,UK) at 0.5 L ha−1 with varying rates of 250 g L−1 azoxystrobin(Amistar, Syngenta Crop Protection UK Ltd) applied at GS 59 (earfully emerged) (Table 1). The base fungicide programme used forthis study was designed to provide a high level of disease controlto enable any potential physiological effects of the strobilurinfungicide azoxystrobin on green leaf duration, grain quality andalcohol to be detected.

Crop establishment and managementAll five experiments were established as the first wheat cropafter oilseed rape into a plough-based seedbed consisting of afree-draining sandy loam soil. The experiments were all drilledin early October using a seed rate of 300 seeds m−2. In theautumn of each year a broad-spectrum herbicide along with aninsecticide was applied to control weeds and aphids in each ofthe experiments. Soil nitrogen was measured prior to fertiliserapplications, enabling nitrogen inputs to be adjusted accordingly.This ensured that crops grown in each experiment would haveapproximately 180 kg available nitrogen ha−1. In experiment 1(2007), nitrogen was applied on 8 March (GS 25) at 40 kg ha−1,on 14 April (GS 32) at 60 kg ha−1 and on 27 April (GS 37) at 60 kgha−1. In experiments 2 and 3 (2008), applications were made on 6March (GS 25) at 42.5 kg ha−1, on 15 April (GS 32) at 42.5 kg ha−1

and on 3 May (GS 37) at 40 kg ha−1. In experiments 4 and 5 (2009),nitrogen was applied on 13 March (GS 29) at 40.5 kg ha−1, on 9April (GS 31) at 45 kg ha−1 and on 15 May (GS 39) at 41.5 kg ha−1.A growth regulator of chlormequat (1.29 kg ha−1) was applied atGS 31. Fungicide treatments at GS 30, 32 and 39 were appliedto all plots using a 12 m tractor-mounted sprayer. Experimental

treatments at GS 59 were applied using a hand-held field trialsprayer propelled by CO2. Spraying was performed using a spraypressure of 2 bar to provide medium spray quality, with LurmarkF110 02 nozzles. Treatments were applied using a water volumeof 200 L ha−1 and a forward speed of 1 m s−1.

Weather dataWeather data were collected from a meteorological station within1 km of the field experiments. Data for monthly rainfall and averageminimum and maximum temperatures were compared with datacollected over a period of 30 years (1961–1990) from the samemeteorological station.

Preharvest assessmentsSeptoria leaf blotch (Septoria tritici) was the only foliar disease thatcaused significant symptoms in these experiments throughoutthe growing season. This disease was assessed on the flag leaf60 days after leaf emergence on ten leaves per plot using visualassessment keys,7 which provided a pictorial comparison of %disease symptoms at given levels (1, 2 and 5%). Despite being afirst wheat, parts of experiment 2 became infected with take-all(Gaeumannomyces graminis). This was visually assessed using themethod of Tilston et al.,8 whereby the number of white headswithin 1 m2 were counted at three points in each plot using asystematic sampling method. These data were then used as acovariate when analysing the results for this experiment.

Towards the end of the growing season the decline in green flagleaf area was monitored by performing visual assessments of %green leaf area every 3 days throughout the period of senescenceon ten randomly selected shoots per plot.

Harvest and postharvest assessmentsYield and grain weightExperiments were harvested using a plot combine, a 1 kg sampleof grain being taken from each plot; all further measurementsconducted derived from this sample. Thousand-grain weight wasmeasured using an automated seed counter with a known weightof grain. From this the average weight of a grain was determinedand then multiplied by 1000.

Grain protein concentrationGrain nitrogen concentration was measured by near-infrared (NIR)using an Infratec 1241 grain analyser (Foss, Warrington, UK). Grainprotein concentration was measured by multiplying the nitrogenconcentration by 5.7.9

Predicted alcohol yieldThe Scotch Whisky Research Institute (SWRI) has developed acalibration for the NIR grain analyser to predict alcohol yields,using grain samples from across the UK collected over 7 years.This technique is now commercially used within Scottish graindistilleries and was employed in this study to predict the alcoholyields on the plot grain samples. In each of the three years, tensamples were selected for their high and low alcohol yields. Theseselected samples were subjected to a wheat cook method10 toconfirm the accuracy of the predicted alcohol yield data. Thismethod closely simulates the distillation process conducted ingrain distilleries.


26

04

www.soci.org AM Watson et al.

Table 2. Monthly rainfall, average minimum temperature and aver-age maximum temperature for May, June, July and August 2007, 2008and 2009 in comparison with 30 year mean (1961–1990) of recordingstaken at Harper Adams University College, Shropshire, UK

Year May June July August

Rainfall (mm)

2007 107.2 236.8 125.8 20.4

2008 47.5 35.2 94.4 83.6

2009 50.2 92.2 110.6 37.8

1961–1990 57.2 54.2 49.1 60.4

Average minimum temperature (◦C)

2007 7.2 11.4 11.4 11.1

2008 9.1 9.0 12.0 12.8

2009 7.2 9.7 11.5 11.8

1961–1990 6.0 8.8 10.6 10.5

Average maximum temperature (◦C)

2007 16.8 19.7 19.5 20.8

2008 18.2 19.2 21.4 20.6

2009 17.5 20.4 21.0 21.7

1961–1990 15.6 18.7 20.5 20.2

Statistical analysisData analysis was performed using the statistical program GenstatVersion 12.0. Green flag leaf area duration was assessed by usingregression analysis with standard curves and fitting the Gompertzmodel.11 This enabled data to be combined from assessmentsmade throughout the period of green leaf decline to give the time(days) to 37% green flag leaf area (Gompertz M) for each plot.A large proportion of the plots had low levels of disease, whichskewed the data. Loge transformation was performed. The resultsfrom experiment 2 were analysed using take-all as a covariate.

All data from the five experiments were combined in aspreadsheet, and one-way analysis of variance (ANOVA) inrandomised blocks was performed using experiment as a factor.The residual data from the ANOVA table were then used for furtheranalysis. The residuals were used to remove the systematic effectof experiment, enabling overall regressions to be performed onthe data. The general mean was then added to the residuals forpresentation of the results. Correlation and linear regression wereused to study the relationships between the various parameters.

RESULTSThe weather data collected showed that there was a high level ofrainfall in May, June and July 2007 (experiment 1), July 2008 andJune and July 2009 (Table 2). Temperature measurements showedno major variations from the 30 year mean. Therefore water doesnot appear to be a limiting factor to grain development withinthese experiments.

The experiments were designed to study how the alcohol yieldof the grain was influenced by factors occurring in the field, suchas disease infection, green leaf retention and the relationshipsthese had with grain quality. The overall relationships from thefive experiments are presented in Table 3. Increasing levels ofseptoria leaf blotch disease symptoms resulted in a significant(P < 0.001) reduction in green flag leaf area duration, whichshowed a significant (P < 0.001) relationship with increasinggrain weight. Increases in thousand-grain weight were related

Table 3. Correlation matrix showing relationships between septorialeaf blotch on flag leaf 45 days after leaf emergence (SLB, %), greenflag leaf area duration (GFLAD, days to Gompertz M), grain yield at 85%dry matter (GY, t ha−1) and grain quality parameters thousand-grainweight (TGW, g), grain protein concentration (GPC, %) and predictedalcohol yield (PAY, LA t−1)

SLB GFLAD GY TGW GPC

GFLAD −0.4660∗∗∗

GY −0.3047∗∗∗ 0.8117∗∗∗

TGW −0.1307NS 0.3313∗∗∗ 0.3108∗∗∗

GPC 0.1807∗ 0.2085∗∗ 0.2660∗∗∗ 0.1318NS

PAY −0.4778∗∗∗ 0.3953∗∗∗ 0.4188∗∗∗ 0.2382∗∗ −0.3349∗∗∗

Data presented are from all five experiments. Significance: ∗ P < 0.05;∗∗ P < 0.01; ∗∗∗ P < 0.001; NS, not significant (P > 0.05).

(P = 0.003) to an increase in the predicted alcohol yield observedin the experiments (Table 3 and Fig. 1).

The data can be linked together through the physiologicalmechanisms involved. This shows how disease infection in thefield results in a decline in green leaf area, which in turn isimportant for grain quality, which ultimately influences predictedalcohol yield (Fig. 1).

Where levels of septoria leaf blotch symptoms were less, greenflag leaf area duration was significantly (P < 0.001) greater.Increased green flag leaf area duration showed a relationshipwith increased grain filling, resulting in a higher thousand-grainweight, which in turn resulted in an increased predicted alcoholyield.

Septoria leaf blotch and green flag leaf area duration can havea large impact on grain quality and were therefore directly linkedto the predicted alcohol yield of the grain (Figs 2 and 3).

Septoria leaf blotch and green flag leaf area durationSeptoria leaf blotch was the only foliar disease present at highenough levels for assessments to be conducted. The averageseverity of symptoms in plots that received no fungicide at GS59 across the five experiments was 17.0% on the flag leaf inmid-July (60 days after flag leaf emergence). Greater symptomsof the disease inevitably resulted in a decline in green leaf area(P < 0.001).

Septoria leaf blotch and grain qualityThe disease appeared to decrease thousand-grain weight, butthe relationship was not significant (P = 0.157). This is likelyto be the result of the relatively low disease levels present inthese experiments. A significant (P = 0.050) relationship betweenseptoria leaf blotch and grain protein concentration was detected,however. This showed that, where there was a 1% increase inthe level of septoria leaf blotch on the flag leaf, grain proteinconcentrations increased by 0.04%. In addition, it was shown thata 1% increase in the disease reduced alcohol yields by 1.12 LA t−1

(P < 0.001).

Green flag leaf area duration, grain quality and yieldIncreased green flag leaf area duration had a significant (P < 0.001)relationship with thousand-grain weight, which showed anincrease of 0.18 g for each day taken to reach Gompertz M. Asignificant (P < 0.001) increase in grain yield of 0.20 t ha−1 was


26

05


SLB

y = −1.3907x − 0.0006R2= 0.22

(P < 0.001)

y = 0.1816x − 0.0406R2 = 0.11

(P < 0.001)

PAY

y = 0.3546x + 0.0233R2 = 0.05

(P = 0.006)

GFLAD

GFLAD

TGW

TGW

Figure 1. Effect of septoria leaf blotch (SLB) on green flag leaf area duration (GFLAD) and how this influences thousand-grain weight (TGW) to ultimatelyinfluence predicted alcohol yield (PAY) of grain.

4475

4500

4525

4550

4575

10

4600

8

4625

6

4650

43 95 7

4450

SLB (%)

PA

Y (

LA t−

1 )

y = −1.12253x + 464.15

R2 = 0.23 (P < 0.001)

Figure 2. Overall relationship between septoria leaf blotch (SLB) on flagleaf (60 days after leaf emergence) and predicted alcohol yield (PAY) fromall five experiments (data presented are residuals added to general mean).

also observed. Grain protein concentration increased (P = 0.006)by 0.02% for each day taken to reach Gompertz M. Predictedalcohol yield increased by 0.34 LA t−1 for each day taken to reachGompertz M (P < 0.001).

Grain quality and predicted alcohol yieldThousand-grain weightIncreasing thousand-grain weight appeared to result in ahigher grain protein concentration in these experiments, butthe relationship was not significant (P = 0.096). A significant(P = 0.003) relationship was seen with predicted alcohol yield,which increased by 0.36 LA t−1 for each 1 g increase in thousand-grain weight.

655550 60 70

GFLAD (days to Gompertz M)

4475

4500

4525

4550

4575

4600

4625

4650

4450

PA

Y (

LA t−1

)

y = 0.3359x + 437.26

R2 = 0.15 (P < 0.001)

Figure 3. Overall relationship between green flag leaf area duration(GFLAD) and predicted alcohol yield (PAY) from all five experiments(data presented are residuals added to general mean).

Protein and predicted alcohol yieldAn inverse relationship (P < 0.001) between grain proteinconcentration and predicted alcohol yield was shown in theseexperiments. For every 1% increase in the concentration of proteinwithin the grain, the predicted alcohol yield fell by 3.23 LA t−1.

Predicted alcohol yield and actual alcohol yieldDistillations were conducted to determine the accuracy of thepredicted alcohol yields. The results showed that there was asignificant (P < 0.001) relationship between the two methods.For every 1 LA t−1 increase in predicted alcohol yield the actualalcohol yield increased by 0.71 LA t−1.


26

06

www.soci.org AM Watson et al.

DISCUSSIONThe results show that factors affecting the crop later in the season,such as disease and green flag leaf area duration, can havesubstantial effects on alcohol yield. Maintenance of green flagleaf area is of key importance in achieving high crop yields (tha−1) as well as helping to maximise the alcohol yield of the grain(LA t−1). Delaying senescence helps maximise the crop’s abilityto photosynthesise. In these experiments, green flag leaf areaduration was manipulated by fungicide applications. Prolongedgreen flag leaf area duration is largely the result of disease controlwhen nutrition and water are not limiting factors. In addition,fungicides have been shown to prolong green flag leaf areaduration through physiological effects such as promoting thegrowth hormone cytokinin and delaying the inhibitor ethylene.12

Other studies have shown increases in green flag leaf areaduration through the control of phylloplane microflora such asCladosporium herbarum.13,14

Prolonged green flag leaf area duration increases the potentialgrain-filling period and hence increases grain weight. In this studyan average thousand-grain weight increase of 0.18 g occurred foreach additional day the flag leaf retained above 37% green leafarea. However, studies by Dimmock and Gooding15 showed thatthe ability of the crop to prolong grain filling with increased greenflag leaf area duration was dependent on the variety grown. Theyfound that early-maturing varieties benefited less from extendinggreen flag leaf area duration, as these varieties had a limited grain-filling period compared with later-maturing varieties. The varietyGlasgow has an average maturity date which is identical (0 days ±)to the standard solstice.16 Therefore it is likely that the applicationof fungicides that promote green flag leaf area duration wouldbe more beneficial in increasing alcohol yield in late-maturingvarieties such as Invicta (+3 days). However, it should be notedthat green flag leaf area duration and grain filling will also bestrongly influenced by the growing conditions of the crop, suchas temperature and the availability of water and nitrogen.17 – 19

Therefore this must be taken into account when making decisionson whether to apply additional fungicides to distilling wheat crops.

Where increases in thousand-grain weight occur, it could beassumed that grain protein concentration would be diluted asa result of greater accumulation of carbohydrates later in theseason. However, despite increases in thousand-grain weight inthis study, no dilution of grain protein was observed, suggesting alate season accumulation of nitrogen within the grain. Thereforegrain protein increased at a comparable rate to starch. As a result,there was a relatively high level of protein at a given starchconcentration. Therefore the relationship between protein andalcohol yield was less (−3.23 LA t−1 per 1% increase in grainprotein) than in previous studies (−7.50 LA t−1).1 Weather datacollected over the experiments showed a high level of rainfall in thelater period of the growing season in each of the years. AccordingMonaghan et al.,20 this can result in an increased proportion ofnitrogen accumulated post-anthesis. In addition, Gooding et al.21

and Ford et al.22 observed that late season fungicide applicationshad the potential to delay senescence in the root system, thusproviding crops with the potential to translocate a greater amountof nitrogen to the grain. This may therefore offer a potentialexplanation for the results in the current study.

Studies of the growing crop also showed that the control ofseptoria leaf blotch had a significant relationship with alcoholyield. For every 1% increase in septoria leaf blotch severity on theflag leaf (60 days after flag leaf emergence) there was a reductionin alcohol yield of 1.12 LA t−1. Septoria leaf blotch is a facultative

pathogen. Facultative pathogens are highly destructive, secretingenzymes that destroy the plant cell wall to enable them to feed onthe glucose within the cell.23 This reduces the plant’s grain-fillingpotential, thus reducing the thousand-grain weight. Accordingto Gooding et al.,24 facultative pathogens are more detrimentalto carbohydrate accumulation than to nitrogen. This was seen inthe present study with an increase in grain protein concentrationof 0.04% for every 1% increase in septoria leaf blotch severity.Previous studies25 – 27 have also observed increases in grain proteinconcentration as a result of septoria leaf blotch. The abilityof this disease to reduce carbohydrate accumulation, therebyincreasing grain protein concentration, can be directly related tothe reductions in predicted alcohol yield observed in the presentstudy. These findings suggest that other facultative pathogenssuch as eyespot (Oculimacula spp.) and take-all (G. graminis var.tritici) may also have a similar relationship, resulting in reducedalcohol yields. The occurrence of obligate pathogens such asrust (Puccinia spp.) and mildew (Blumeria graminis), however, islikely to have a lesser effect. These pathogens are less destructive,developing a haustorial feeding complex to continuously absorbnutrients from the living plant cell,28 which results in less ofan effect upon starch translocation. Studies by Dimmock andGooding29 suggest that mildew retains nitrogen in the leaves,which is likely to reduce the grain protein concentration further.This is supported in studies by Johnson et al.30 and Puppalaet al.,26 whose results appear to show a reduction in grain proteinconcentration with increased levels of mildew. Therefore infectionwith these diseases is likely to have little effect on grain alcoholyield per tonne but will reduce the overall crop yield and thereforealcohol yield per hectare.

In conclusion, this study has shown that fungicide applicationshave the potential to substantially increase the alcohol yield ofwheat grain as a result of disease control and extended green flagleaf area duration when the crop is not sink-limited. In addition toincreasing the alcohol yield per tonne of grain, the concomitanteffect of increased crop yield will result in a substantial increasein alcohol yield per hectare. In the present study, disease levelsremained relatively low in the later stages of crop development.This was the result of the robust fungicide programme applied;in situations where disease levels are greater, the influence ofapplying a late season fungicide is likely to be of increasedimportance. It should be noted that at the time of writing thispaper there are no known price incentives in place for growers toproduce crops with high alcohol yields in the UK. However, thecost of fungicide inputs in this study (£12.50/ha + £10.80/ha forapplication) was more than covered by the additional yield benefitobserved (+0.85 t ha−1 at £100/t = £85/ha).31

ACKNOWLEDGEMENTSThe authors thank Syngenta Crop Protection UK Ltd for sponsor-ship of the project, David Ranner of Syngenta Crop ProtectionUK Ltd for advice, Tom Bringhurst of the Scotch Whisky ResearchInstitute for advice and the technical staff of the Crop and Envi-ronment Research Centre, Harper Adams University College forestablishment and support of the field experiments.

REFERENCES1 Smith TC, Kindred DR, Brosnan JM, Weightman RM, Shepherd M and

Sylvester-Bradley R, Wheat as a Feedstock for Alcohol Production(HGCA Research Review No. 61). HGCA, London (2006).


26

07


2 Gurgel A, Reilly JM and Paltsev S, Potential land use implications forglobal biofuels industry. J Agric Food Ind Organisat 5:1–34 (2007).

3 Reijnders L and Huijbregts MAJ, Life cycle greenhouse gas emissions,fossil fuel demand and solar energy conversion efficiency inEuropean bioethanol production for automotive purposes. J CleanerProd 15:1806–1812 (2007).

4 Ruske RE, Gooding MJ and Jones SA, The effects of addingpicoxystrobin, azoxystrobin and nitrogen to a triazole programmeon disease control, flag leaf senescence, yield and grain quality ofwinter wheat. Crop Protect 22:975–987 (2003).

5 Paveley ND, Sylvester-Bradley R, Scott RK, Craigon J and Day W,Steps in predicting the relationship of yield on fungicide dose.Phytopathology 91:708–716 (2001).

6 Zadocks JC, Chang TT and Konzak CF, A decimal code for the growthstages of cereals. Weed Res 114:415–421 (1974).

7 Cereal Disease Guide. AgrEvo UK Ltd, King’s Lynn (1997).8 Tilston EL, Pitt D, Fuller MP and Groenhof AC, Compost increases yield

and decreases take-all severity in winter wheat. Field Crops Res94:176–188 (2005).

9 Teller GL, Non-protein nitrogen compounds in cereals and theirrelation to the nitrogen factor for protein in cereals and bread.Cereal Chem 9:261–274 (1932).

10 Agu RC, Bringhurst TA and Brosnan JM, Production of grain whiskyand ethanol from wheat, maize and other cereals. J Inst Brew112:314–332 (2006).

11 Gooding MJ, Dimmock JPRE, France J and Jones SA, Green leaf areadecline of wheat flag leaves: the influence of fungicides andrelationships with mean grain weight and grain quality. Ann ApplBiol 136:77–84 (2000).

12 Grossmann K and Retzlaff G, Bioregulatory effects of the fungicidalstrobulurin kresoxim-methyl in wheat (Triticum aestivum). Pestic Sci50:11–20 (1997).

13 Bertelsen JR, Neergaard E and Smedegaard-Petersen V, Fungicidaleffects of azoxystrobin and epoxiconazole on phyllosphere fungi,senescence and yield of winter wheat. Plant Pathol 50:190–205(2001).

14 Hussain Z and Leith MH, The effect of applied sulphur on the growth,grain yield and control of powdery mildew in spring wheat. AnnAppl Biol 147:49–56 (2005).

15 Dimmock JPRE and Gooding MJ, The effects of fungicides on rate andduration of grain filling in winter wheat in relation to maintenanceof flag leaf green area. J Agric Sci 138:1–16 (2002).

16 Home Grown Cereals Authority, HGCARecommended List,WinterWheat2010/11. HGCA, London (2010).

17 Pan J, Yan J and WeiXing C, Modelling plant carbon flow and grainstarch accumulation in wheat. Field Crops Res 101:276–284 (2007).

18 Gooding MJ, Ellis RH, Shewry PR and Schofield JD, Effects of restrictedwater availability and increased temperature on the grain filling,drying and quality of winter wheat. J Cereal Sci 37:295–309 (2003).

19 Kindred DR, Verhoeven TMO, Weightman RM, Swanston JS, Agu RC,Brosnan JM, et al., Effects of variety and fertiliser nitrogen on alcoholyield, grain yield, starch and protein composition of winter wheat.J Cereal Sci 48:46–57 (2008).

20 Monaghan JM, Snape JW, Chojecki JS and Kettlewell PS, The use ofgrain protein deviation for identifying wheat cultivars with highprotein concentrations and yield. Euphytica 122:309–317 (2001).

21 Gooding MJ, Gregory PJ, Ford KE and Pepler S, Fungicide and cultivaraffect post-anthesis patterns of nitrogen uptake, remobilizationand utilization efficiency in wheat. J Agric Sci 143:503–518 (2005).

22 Ford KE, Gregory PJ, Gooding MJ and Pepler S, Genoptype andfungicide effects on late-season root growth of winter wheat.Plant Soil 284:33–44 (2006).

23 Farrar JF and Lewis DH, Nutrient relations in biotrophic interactions, inFungal Infection of Plants, ed. by Pegg GF and Ayres AG. CambridgeUniversity Press, Cambridge, pp. 92–132 (1987).

24 Gooding MJ, Smith SP, Davies WP and Kettlewell PS, Effects of late-season applications of propiconazole and tridemorph on disease,senescence, grain development and the breadmaking quality ofwinter wheat. Crop Protect 13:362–369 (1994).

25 Pepler S, Gooding MJ, Ford KE, Ellis RH and Jones SA, Delayingsenescence of wheat with fungicides has interacting effects withcultivar on grain sulphur concentration but not with sulphur yieldor nitrogen : sulphur ratios. Eur J Agron 22:405–416 (2005).

26 Puppala V, Herrman TJ, Bockus WW and Loughin T, Quality responseof twelve hard red winter wheat cultivars to foliar disease acrossfour locations in central Kansas. Cereal Chem 75:94–99 (1998).

27 Ruske RE, Gooding MJ and Dobraszczyk BJ, Effects of triazole andstrobulurin fungicide programmes with and without late seasonnitrogen fertiliser, on baking quality of Malacca winter wheat.J Cereal Sci 40:1–8 (2004).

28 Sutton PN, Gilbert MJ, Williams LE and Hall JL, Powdery mildewinfection of wheat leaves changes host solute transport andinvertase activity. Physiol Plant 129:787–795 (2007).

29 Dimmock JPRE and Gooding MJ, The effects of fungicides on wheatgrain protein concentration. Aspects Appl Biol 64:217–218 (2001).

30 Johnson JW, Baenziger PS, Yamazaki WT and Smith RT, Effects ofpowdery mildew on yield and quality of isogenic lines of Chancellorwheat. Crop Sci 19:349–352 (1979).

31 Nix J, The John Nix Farm Management Pocketbook (40th edn).Andersons Centre, Melton Mowbray (2009).


26

08



Modulation of salt (NaCl)-induced effects on oilcomposition and fatty acid profile of sunflower(Helianthus annuus L.) by exogenousapplication of salicylic acidSibgha Noreena and Muhammad Ashrafa,b∗

Abstract

BACKGROUND: Salicylic acid (SA) is a potential endogenous plant hormone that plays an important role in plant growth anddevelopment. Since sunflower yield and its seed oil yield are adversely affected by salinity, in this study the role of SA inmodulating salt (NaCl)-induced effects on various yield and oil characteristics of sunflower was investigated. For this purposea greenhouse experiment comprising two sunflower hybrid lines (Hysun-33 and SF-187), two NaCl levels (0 and 120 mmol L−1)and four SA levels (0, 100, 200 and 300 mg L−1) was conducted.

RESULTS: Salt stress markedly reduced yield, oil content, linoleic acid and δ-tocopherol in both sunflower lines, while it increasedlinolenic acid, palmitic acid, stearic acid and α- and γ -tocopherols. However, increasing levels of foliar-applied SA resulted inimproved achene yield and hundred-achene weight in both lines. Foliar-applied SA caused a significant decrease in oil stearicacid and α- and γ -tocopherols in both lines under non-saline and saline conditions.

CONCLUSION: Salt-induced harmful effects on achene yield and oil characteristics of sunflower could be alleviated by exogenousapplication of SA. High doses of SA caused a marked increase in sunflower achene oil content as well as some key fatty acids.c© 2010 Society of Chemical Industry

Keywords: sunflower; salt stress; salicylic acid; oil composition; tocopherols; antioxidant activity

INTRODUCTIONPlant growth and development are directly regulated by planthormones. These influence plants in multifarious ways, affectinga number of physiological/biochemical processes in plantssubjected to abiotic stresses.1 – 4 Salicylic acid (SA) and relatedcompounds belong to a diverse group of plant phenolics. Thisacid acts as an endogenous signal molecule and is responsible foralleviating adverse effects of abiotic stresses on plants.5,6 SA hasalso shown promise in mitigating oxidative and osmotic stressesdue to salt stress by enhancing proteins, total free amino acids andsoluble sugars in different plant parts of several crops.3,7 – 9 Apartfrom enhancement of growth and development in various cropplants, foliar-applied SA has been reported to improve grain yieldof mung bean (Vigna radiata),10 Phaseolus vulgaris11 and wheat.3,12

A significant increase in number of pods per plant and grain yieldof soybean by exogenously applied SA was also observed.10

Sunflower (Helianthus annuus L.) is one of a number of potentialoilseed crops and is gaining considerable popularity because itis a short-duration crop. The oil content of sunflower varies from40 to 45% depending on the quality of seed. It contains 85–95%polyunsaturated fatty acids. Oleic and linoleic acids are abundant,while it is low in cholesterol.13 Sunflower is rated as a moderatelysalt-tolerant crop.14 – 16 However, as the salinity of irrigation waterincreases, plant height, stem diameter and head diameter ofsunflower decrease significantly.17

Significant decreases in achene oil concentration and oil yieldmay occur when the salt content of irrigation water exceeds l gL−1.18 Foliar injury from saline sprinkling water containing highNa+ or Cl− concentration may occur when the salt level is greaterthan 20 mmol L−1. Francois13 and Maas16 reported that soil salinityup to ECe 4.8 dS m−1 did not have an adverse effect on relativeseed yield of four hybrids of sunflower, but there was a reductionin yield by 5% for each unit increase in salinity above 4.8 dSm−1. However, it has been observed that fatty acid compositionis greatly affected by salinity: under saline conditions, oleic acidincreased and linoleic acid decreased progressively with increasingsalinity level owing to salt-induced inhibition of oleate desaturaseenzyme.19 Various researchers have reported the effects of salinityon oil content and fatty acid composition in different oilseed cropssuch as safflower (Carthamus tinctorius L.),20,21 olive,22 rapeseed(Brassica napus L.), stock (Matthiola tricuspidata) and eveningprimose (Oenothera biennis).23 Oil yield decreased significantlyfrom 38.3 to 3.8 g per head and unsaturated fatty acids such as

∗ Correspondence to: Muhammad Ashraf, Department of Botany, University ofAgriculture, Faisalabad 38040, Pakistan. E-mail: [email protected]

a Department of Botany, University of Agriculture, Faisalabad, Pakistan

b King Saud University, Riyadh, Saudi Arabia


26

09

Modulation of salt-induced effects on sunflower by salicylic acid www.soci.org

oleic and linoleic acids, which account for 85–95% of the totalfatty acids in sunflower seed, were greatly affected by salt stress.24

The oil content of sunflower seed decreased from 524 to 508 mgoil g−1 seed with increasing salt level, while oleic acid increasedfrom 82.8 to 86.0%; conversely, linoleic acid decreased from 6.9to 2.8% owing to inhibition of oleate desaturase occurring undersalt stress.25 Francois and Kleiman25 also reported an increase inoleic acid with a concurrent decrease in linoleic acid in safflower(C. tinctorius L.) under saline conditions (3.7–7.9 dS m−1). Theincrease in oleic/linoleic acid ratio could be due to water stress,lipid peroxidation and/or accumulation of Na+ and Cl− under saltstress.26

In view of the great economic importance of sunflower asa potential oilseed crop, it was hypothesised that exogenouslyapplied SA could improve salt tolerance in sunflower for profitableyield production as well as improved quality of oil. The objectiveof this study was to examine whether or not exogenously appliedSA could alter the quantity and quality of achene oil from salt-stressed sunflower plants. In addition, relationships between yieldattributes and various oil components of sunflower supplied withSA were established.

MATERIALS AND METHODSThe present investigation was conducted in a wire-house at theUniversity of Agriculture, Faisalabad, Pakistan under natural light.During the experimental period, mean maximum temperaturesranged from 27.6 to 40.3 ◦C, mean minimum temperatures from13.0 to 27.3 ◦C and mean relative humidity from 26.2 to 52.4%. Theexperimental design was a completely randomised three-factorial(2 × 2 × 4) design with four replications. Two sunflower (H. annuusL.) hybrid lines (Hysun-33 and SF-187), two salinity levels (0 and120 mmol L−1 NaCl) and four SA levels (0, 100, 200 and 300 mg L−1)were assessed. Before the start of the experiment the seeds of bothsunflower hybrids were treated with 5 g L−1 sodium hypochloritesolution for 5 min to prevent the occurrence of seedling diseases.The seeds were then washed several times with distilled waterto remove excess sodium hypochlorite solution. The sterilisedseeds were air dried at room temperature. Plastic pots of 24.5 cmdiameter and 28 cm depth were each filled with 10 kg of air-dried river sand. The sunflower achenes were selected on thebasis of uniform size and maturity and freedom from physicaldamage or infection. Five achenes were dibbled at 5 mm depthin each pot, and irrigation with modified half-strength Hoagland’snutrient solution was carried out every 2 days after sowing.27

After complete emergence, i.e. on day 4 after sowing, thinning tothree seedlings per pot was carried out. Neutral NaCl was usedto impose salt stress. Two levels of salinity (0 and 120 mmol L−1

NaCl) were prepared in full-strength Hoagland’s nutrient solution.The salt stress was imposed on day 15 after sowing. The plantswere salinised progressively at 1 day intervals at an increment of40 mmol L−1 NaCl. SA (2-hydroxybenzoic acid (C7H6O3), molecularweight 138.1; Sigma Aldrich, Tokyo, Japan) was initially dissolvedin a few drops of dimethyl sulfoxide and then in distilled waterand the pH was adjusted to 5.5 ± 0.2 with 0.5 mol L−1 KOH.Four levels of SA (0, 100, 200 and 300 mg L−1) were prepared.The SA was applied exogenously in combination with 1 mL L−1

Tween-20 (polyoxyethylene sorbitan laurate) as a surfactant toensure spreading of the applied solution on the leaf surface andto maximise penetration of the exogenously applied SA into theleaf tissues. The first foliar application of SA was done on day 20after sowing and the second on day 3 after the first application. A

volume of 9 mL of solution per three plants was sprayed in all caseswith a manually operated sprayer. The nozzle of the hand-heldsprayer was adjusted to deliver 1 mL of solution per spray, witheach plant receiving 3 mL of the SA solution.

Yield and yield componentsCapitulum weight per plant, hundred-achene weight and acheneyield per plant were measured at maturity on day 113 after sowing.Harvested heads were sun dried and thrashed manually to cleanthe grains.

Oil extractionDried seeds (100 g) from each treatment were crushed. The oil wasextracted with 0.5 L of n-hexane using a Soxhlet apparatus. Oilcontent was determined by evaporating the extractant in a rotaryevaporator.

Refractive indexThe refractive index of sunflower oil was determined using an Abberefractometer. The temperature of the prism was maintained at40 ◦C. The refractive index (n) was read after putting a clear dropof oil on the prism:

n = sin(i)/ sin(r) = velocity of light in air/velocity of light in oil

Specific gravityThe specific gravity of sunflower oil was determined as follows:

specific gravity = density of oil at

20 ◦C/density of water at 20 ◦C

Fatty acid compositionFirst, fatty acid methyl esters (FAMEs) were prepared by thestandard IUPAC method. Samples were derivatised into FAMEs oftriglycerides by saponification of glycosides and esterification offatty acids using methanol. The oil sample (0.2 g) was placed ina 100 mL round-necked and round-bottomed flask, then 30 mLof methanol and one pellet of KOH were added. The contentsof the flask were refluxed for 25 min until the droplets of fatdissolved. On cooling, the reaction mixture was gently transferredto a separating funnel and a small amount of n-hexane was added.The funnel was shaken gently by rotating it several times and theupper layer of n-hexane was separated. The solution was driedover anhydrous sodium sulfate, filtered and stored in a sealedsample tube in a freezer until analysis. The fatty acid composition(stearic, palmatic, oleic and linoleic acids) was determined by gaschromatography using a Shimadzu 17-A gas chromatograph fittedwith an SP-2330 methyl lignoserate-coated polar capillary column(30 m × 0.32 mm) and a flame ionisation detector (Shimadzu,Kyoto, Japan).

Determination of tocopherolsTocopherols were analysed by the method of Lee et al.28 A high-performance liquid chromatograph (HPLC) fitted with an S-3210UV detector (Sykam GmbH, Kleinsthein, Germany) was used. Theoil sample (1 g) was placed in a 10 mL volumetric flask and thevolume was made up to the mark with acetonitrile. The flaskwas wrapped in foil to retard the oxidation process. The materialwas subsequently filtered and a 20 µL aliquot was injected into


26

10

www.soci.org S Noreen, M Ashraf

Table 1. Mean squares from analysis of variance of data for yield attributes, oil physical properties and oil composition of sunflower (Helianthusannuus L.) when varying levels of salicylic acid were applied as a foliar spray to 24-day-old plants subjected to normal or saline conditions

Source ofvariation DF

Hundred-acheneweight

Achene yieldper plant

Capitulum weightper plant

Seed oilcontent

Refractiveindex

Salt (S) 1 2.080∗∗∗ 12.60∗∗ 1.32ns 142072.4∗∗∗ 0.0010ns

Hybrid line (HBL) 1 0.151ns 64.20∗∗∗ 138.40∗∗∗ 15221.4∗∗∗ 0.0070∗∗∗

Salicylic acid (SA) 3 0.235∗∗ 0.85ns 0.90ns 23167.5∗∗∗ 0.0010ns

S × HBL 1 3.850∗∗ 39.20∗∗ 61.60∗∗∗ 858.4∗ 0.0001ns

S × SA 3 0.077ns 4.25∗ 12.80∗∗ 1500.5∗∗∗ 0.0004ns

HBL × SA 3 0.257ns 1.40ns 5.25ns 4531.7∗∗∗ 0.0020∗∗

S × HBL × SA 3 0.489∗ 2.72ns 5.85ns 5996.2∗∗∗ 0.0001ns

Error 48 0.163 1.12 2.65 164.1 0.004

Source of variation DF Specific gravity α-Tocopherol δ-Tocopherol γ -Tocopherol Oleic acid

Salt (S) 1 0.443∗∗∗ 508.50∗∗∗ 3609.00∗∗∗ 1160.00∗∗∗ 183.40∗

Hybrid line (HBL) 1 0.382∗∗ 20652.00∗∗∗ 44.40ns 1559.00∗∗∗ 5114.00∗∗∗

Salicylic acid (SA) 3 0.023ns 18166.00∗∗∗ 603.10∗∗∗ 727.10∗∗∗ 14.53ns

S × HBL 1 0.163∗ 2458.00∗∗∗ 53.30∗ 1.40∗ 79.81ns

S × SA 3 0.024ns 2477.00∗∗∗ 76.60∗∗ 4.45∗ 3.95ns

HBL × SA 3 0.011ns 3981.00∗∗∗ 21.30ns 252.30∗∗∗ 44.10ns

S × HBL × SA 3 0.004ns 111.20∗∗∗ 29.60ns 19.40∗ 64.40ns

Error 48 0.036 31.6 12.90 14.70 34.40

Source of variation DF Linoleic acid Linolenic acid Palmitic acid Stearic acid

Salt (S) 1 371.60∗∗∗ 2.080∗∗∗ 107.40∗∗∗ 508.50∗∗∗

Hybrid line (HBL) 1 1863.00∗∗∗ 0.102∗ 0.47ns 20652.00∗∗∗

Salicylic acid (SA) 3 120.90∗∗ 0.072∗∗∗ 0.12ns 18166.00∗∗∗

S × HBL 1 11.50ns 0.00001ns 3.60ns 2458.00∗∗∗

S × SA 3 3.39ns 0.007ns 1.57ns 2477.00∗∗∗

HBL × SA 3 4.46ns 0.012ns 0.66ns 3981.00∗∗∗

S × HBL × SA 3 11.20ns 0.003ns 0.37ns 111.20∗

Error 48 20.6 0.017 2.60 31.60

Significance: ∗ P < 0.05; ∗∗ P < 0.01: ∗∗∗ P < 0.001; ns, not significant.

a Hypersil ODS reverse phase (C-18) guard column for elutionwith a solvent comprising HPLC-grade methanol and acetonitrile(65 : 35 v/v). A flow rate of 1.3 mL min−1 at 30 ◦C was used toachieve separation. The absorbance of the separated solutionwas detected at 292 nm. The different forms of tocopherol wereidentified by comparison of retention times. Quantification wascarried out by comparing the peak areas of the unknown sampleswith those of pure blank standards of α-, γ - and δ-tocopherols(Sigma Chemical Co., St Louis, MO, USA). Data were calculated fromthe peak areas using PeakSimple chromatography data acquisitionand integration software (SRI Instruments, Torrance, CA, USA).

Statistical analysisAnalysis of variance of all data was carried out using CoStatsoftware (CoHort, Berkeley, CA, USA). Significant differencesbetween mean values were assessed using the least significantdifference test at P < 0.05.29

RESULTSSalt stress (120 mmol L−1 NaCl) was found to be very effective inreducing hundred-achene weight in both sunflower hybrid lines.There were no significant differences between the two hybridlines in this parameter (Table 1, Fig. 1). However, exogenous

application of varying levels of SA caused a significant changein grain weight of both sunflower hybrid lines under salt-stressedand non-stressed conditions. Increasing levels of foliar-applied SAresulted in improved achene weight in both hybrid lines. Acheneyield in both hybrid lines was significantly reduced by increasingsalt concentration in the substrate. However, the responses ofthe two lines to salt stress with respect to this yield attributewere significantly different. Exogenous application of 100 and300 mg L−1 SA caused a substantial increase in grain yield of lineSF-187 under salt-stressed and non-stressed conditions (Table 1,Fig. 1). Another potential yield component, capitulum weight, wasnot significantly affected by salinity. However, the two hybrid linesdiffered significantly in this attribute (Table 1, Fig. 1). Under non-stressed conditions, SF-187 attained more capitulum weight afterapplication of 100 mg L−1 SA, but no gain in capitulum weight ofHysun-33 was observed until application of 200 mg L−1 SA (Fig. 1).Exogenous application of 300 mg L−1 SA proved most effectivein increasing capitulum weight in both hybrid lines under saltstressed and non-stressed conditions.

Addition of NaCl to the growth medium caused a substantialreduction in oil content of sunflower. The two hybrid lines differedsignificantly in seed oil content. Hysun-33 produced more seedoil than SF-187 under both saline and normal growth conditions.Seed oil content was significantly improved by foliar applicationof SA in both hybrid lines under salt-stressed and non-stressed


26

11


0

0.5

1

1.5

2

2.5

3

3.5

Hun

dred

-ach

ene

wei

ght (

g)

0

1

2

3

4

5

6

7

8

9

Ach

ene

yiel

d (g

per

pla

nt)

0 mg L−1 100 mg L−1 200 mg L−1 300 mg L−1 Salicylic acid

02468

10121416

ControlSF-187

Saline ControlHYSUN-33

Saline

Cap

itulu

m w

eigh

t (g

per

plan

t)

Figure 1. Yield attributes of two sunflower (Helianthus annuus L.) hybrid lines when varying levels of salicylic acid were applied as a foliar spray to24-day-old plants subjected to normal or saline conditions (mean ± standard error, n = 4).

conditions (Table 1, Fig. 2). Thus exogenously applied SA had apromotive effect on seed oil content and was more effective when300 mg L−1 SA was applied as a foliar spray. Furthermore, plants ofHysun-33 grown on saline medium and sprayed with SA producedmore seed oil than those of SF-187, while the opposite was trueunder non-saline conditions (Fig. 2).

The refractive index of seed oil of both lines was little affectedby salinity. Foliar application of SA did not significantly affect therefractive index of either line (Table 1, Fig. 2). The specific gravityof achene oil of both sunflower lines was significantly reducedby addition of salt to the growth medium. The two lines differedsignificantly under non-saline conditions. Seed oil of SF-187 hadhigher specific gravity than that of Hysun-33 under non-salineconditions. The specific gravity of seed oil of both hybrid lines waslittle influenced by foliar spraying of SA on plants subjected tosaline conditions.

Salt stress significantly increased seed oil linolenic, palmitic andstearic acid contents but decreased seed oil linoleic acid contentin both lines of sunflower (Table 1). The two lines also differedsignificantly in all these fatty acids except palmitic acid. Hysun-33was higher in oil palmitic and linolenic acid contents but lower inoil stearic acid content than SF-187 under saline conditions. Foliar

application of 300 mg L−1 SA resulted in increased oil stearic acidcontent in both lines under saline conditions. Similarly, foliar appli-cation of various levels of SA caused a marked improvement in oillinolenic acid content in both hybrid lines under saline and normalconditions (Fig. 3). Saline conditions caused a significant reductionin oil oleic acid content, particularly in SF-187 (Table 1, Fig. 3). How-ever, Hysun-33 had higher oil oleic acid content than SF-187 underboth salt-stressed and non-stressed conditions. Exogenously ap-plied SA improved oil oleic acid content only in salt-stressed plantsof SF-187 (Fig. 3). Linoleic acid content in both sunflower lines wassignificantly reduced (Table 1). SF-187 was higher in oil linoleic acidcontent than Hysun-33 under both salt-stressed and non-stressedconditions (Fig. 3). Exogenously applied SA resulted in a significantreduction in oil linoleic acid content in plants of both hybrid linesgrown under saline and normal conditions.

Linolenic acid content in both lines was higher under salt stressthan under normal conditions. It generally increased owing toimposition of salt on the rooting medium (Fig. 3). Hysun-33 hadhigher oil linolenic acid content than SF-187 under both non-salineand saline conditions.

Oil palmitic acid content in both sunflower hybrids was higherunder salt stress than under non-stressed conditions. Exogenous


26

12


0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Spe

cific

gra

vity

0

50

100

150

200

250

300

350

400

450

See

d oi

l con

tent

s (g

oil

kg-1

see

d)


1.34

1.36

1.38

1.4

1.42

1.44

1.46

1.48

1.5

ControlSF-187


Saline

Ref

ract

ive

inde

x

Figure 2. Physical properties of seed oil of two sunflower (Helianthus annuus L.) hybrid lines when varying levels of salicylic acid were applied as a foliarspray to 24-day-old plants subjected to normal or saline conditions (mean ± standard error, n = 4).

application of SA had a positive effect in increasing palmitic acidcontent in plants under saline conditions. Hysun-33 maintainedhigher oil palmitic acid content than SF-187 under salt stress. Incontrast, SF-187 maintained higher oil palmitic acid content thanHysun-33 under non-saline conditions (Fig. 3).

Oil stearic acid content decreased with increasing concentrationof foliar-applied SA under non-stressed conditions. On the otherhand, it increased with exogenous SA application under salt stress.SF-187 sprayed with 300 mg L−1 SA had maximum oil stearic acidcontent under salt stress. Foliar application of 200 mg L−1 SAresulted in a reduction in oil stearic acid content compared withfoliar application of 100 mg L−1 SA in plants of both hybrid linesunder salinity stress.

Analysis of tocopherol data showed that salt stress signifi-cantly increased α- and γ -tocopherol contents but decreasedδ-tocopherol content in the seed oil of both sunflower lines(Table 1, Fig. 4). SF-187 was higher in oil α- and γ -tocopherol con-tents than Hysun-33 under both salt-stressed and non-stressedconditions. However, the two lines showed little difference inoil δ-tocopherol content under salt-stressed and non-stressedconditions. Foliar-applied SA caused a significant decrease in allthree tocopherols in both lines under non-saline and saline condi-tions. Maximum reduction in all seed oil tocopherols in both lines

was observed at 300 mg L−1 SA under salt-stressed and normalconditions.

The analysis of correlation coefficients (r) among various yieldand achene oil characteristics presented in Table 2 shows thatonly yield attributes such as hundred-achene weight/acheneyield (r = 0.94∗∗), seed oil content/achene yield (r = 0.76∗∗)and specific gravity/achene yield (r = 0.64∗) are significantlycorrelated; among the oil composition characteristics, α-, γ - and δ-tocopherols are significantly correlated with oleic acid (r = 0.75∗∗,0.76∗∗ and 0.76∗∗ respectively).

DISCUSSIONCrop yield is a vital determinant for assessing crop productivity ofa plant under stress conditions. In sunflower, size of capitulum andnumber and size of achenes per capitulum determine the actualyield, since various physiological and biochemical processes thatgovern yield are adversely affected by salinity.30 – 34 Therefore itis not difficult to suggest that decline in achene yield due tosalt stress with concurrent enhancement in achene yield due toSA application was principally reponsible for change in acheneyield (achene yield vs hundred-achene weight, r = 0.94∗∗∗).These results can be discussed by reviewing earlier studies where


26

13


01020304050607080

Ole

ic a

cid

(%)

0

10

20

30

40

50

60

Lino

leic

aci

d (%

)

0

2

4

6

8

10

12

Pal

miti

c ac

id (

%)


0

0.2

0.4

0.6

0.8

1

ControlSF-187


Saline

Lino

leni

c ac

id (

%)

Figure 3. Fatty acid profile of seed oil of two sunflower (Helianthus annuus L.) hybrid lines when varying levels of salicylic acid were applied as a foliarspray to 24-day-old plants subjected to normal or saline conditions (mean ± standard error, n = 4).

plants treated with salicylates had higher yield when cultivatedin either greenhouse or open conditions. For example, it wasreported that SA caused production of larger-sized tubers of carrot(60%), beet (16%) and radish (200%).35 Increase in capitulumsize of sunflower lines through foliar-applied SA could havebeen due to either development of additional flowers or rapidflower initiation with subsequent flower development.36 Forexample, in a study of Xanthium strumarium, Cleland and Ajami37

reported that exogenously applied SA stimulated flowering.Similarly, flowering in Pistia stratiotes, a member of the familyArecaceae, was also accelerated by incorporating SA in the culturemedium.38 In another study, Kumar et al.39 compared the flower-inducing effect of SA with that of gibberellic acid (GA), kinetin,1-naphthaleneacetic acid, ethral and chlorocholine chloride. Theyfound that SA and/or GA were more effective than any othercombination of these hormones. However, the specific flower-inducing mechanism that involves SA is yet to be elucidated. Basedon information available in the literature, it is evident that floweringis induced by chelating agents in Laminaceae40,41 through theactivity of benzoic acid42,43 and other non-chelating phenolic

compounds.44 It was hypothesised that the free o-hydroxyl groupof benzoic acid induces a metal-chelating characteristic thatfavours flower induction.45 – 47 Moreover, it was concluded thatadditional flower-inducing mechanisms may be heavily involvedin this process.46,47 Thus the results from the present study alsosupport this notion that SA-induced increase in achene yield in salt-stressed plants of both sunflower lines might have been due to thedevelopment of additional flowers, thereby increasing capitulumsize. In the present study, SA-induced increase in achene sizeof salt-stressed plants of both sunflower lines might have beendue to transport of an additional amount of photoassimilatesfrom vegetative organs (source) to achenes (sink) during theirdevelopment, thereby increasing achene size. Zhou et al.48 alsoreported a 9% higher grain weight of maize by injecting stemswith SA compared with sucrose and/or distilled water treatments.Similarly, in a study of wheat, Arfan et al.12 observed that SA-induced improvement in grain yield occurred because of SA-induced enhancement in grain size.

Achene oil content in both sunflower lines was decreasedby salinity. This result is in line with previous studies in which


26

14


0

10

20

30

40

50

60

70

80

d-T

ocop

hero

l (m

g kg

−1 o

il)

0

50

100

150

200

250

a-T

ocop

hero

l (m

g kg

−1 o

il)


0

10

20

30

40

50

60

70

ControlSF-187


Saline

g-T

ocop

hero

l (m

g kg

−1 o

il)

Figure 4. Tocopherol contents of seed oil of two sunflower (Helianthus annuus L.) hybrid lines when varying levels of salicylic acid were applied as a foliarspray to 24-day-old plants subjected to normal or saline conditions (mean ± standard error, n = 4).

it was shown that salt stress reduced seed oil content in flaxand safflower49 and sunflower.19 However, exogenously appliedSA improved achene oil content in both sunflower lines. Thisimprovement was greatest in salt-stressed plants of Hisun-33treated with 300 mg L−1 SA foliar spray. It is widely acceptedthat the quality of achene oil is associated with its fatty acidcomposition, mainly the proportion of oleic, linoleic and linolenicacids. In the present study, salinity stress increased palmitic, stearicand linolenic acid contents in both sunflower lines, while linoleicacid content decreased. Thus salt stress reduced achene oil qualityin both sunflower lines, which is similar to what was observedpreviously in stock (M. tricuspidata), chia (Salvia hispanica) andevening primrose (O. biennis)23 and Matthiola incana.50 Of all fattyacids, the quantity of oleic and linoleic acids is most importantfrom the viewpoint of oil quality. Exogenous application of SA didnot change palmitic, stearic and oleic acid contents, while linolenicacid content increased considerably in both sunflower lines. Theseresults suggest that exogenous SA application counteracted thesalt-induced harmful effects on achene oil quality in both sunflowerlines.

Tocopherols (vitamin E) are non-enzymatic antioxidant com-pounds found to a varying extent in seed oils. They counteract

oxidative stress, thereby maintaining oil quality.51 Althoughdifferent types of tocopherol occur in oils, α-tocopherol is themost prominent, as it has considerable antioxidant activity.52

Salinity stress can generate reactive oxygen species (ROS) in plantsto a great extent,31,53 – 55 which damage seed lipids.56 To combatROS, plants generate various types of antioxidant, including toco-pherols (vitamin E).57 In the present study, salt stress significantlyincreased α- and γ -tocopherol contents in the seed oil of bothsunflower hybrids. These results are in agreement with those ofAnwar et al.,58 who reported that salt stress increased tocopherolcontent in the seed oil of Moringa oleifera plants. A considerablevariation was observed in both sunflower cultivars with regard tothis attribute. However, when parallels were drawn between toco-pherol content and yield attributes, a non-significant associationwas observed.

In conclusion, salt-induced harmful effects on achene yield andoil characteristics of sunflower could be allayed by exogenousapplication of SA, but the effectiveness of SA in mitigating thenegative effects of salt stress was dose-dependent. However, highdoses of SA caused a marked increase in sunflower achene oilcontent as well as some key fatty acids such as linolenic, oleic andpalmitic acids.


26

15


Table 2. Correlation coefficients (r) among yield attributes, oil physical properties and oil composition of sunflower (Helianthus annuus L) whenvarying levels of salicylic acid were applied as a foliar spray to 24-day-old plants subjected to normal or saline conditions

Acheneyield


Capitulumweight

Seed oilcontent

Refractiveindex

Specificgravity

Palmiticacid

Stearicacid

Oleicacid

Linoleicacid

Linolenicacid α-Toc γ -Toc

Achene yield


0.94∗∗∗

Capitulumweight

0.24ns 0.32ns

Seed oilcontent

0.76∗∗ 0.8∗∗ 0.38ns

Refractiveindex

0.28ns 0.13ns −0.25ns −0.04ns

Specificgravity

0.64∗ 0.6∗ −0.10ns 0.13ns 0.61∗

Palmitic acid −0.91∗∗∗ −0.79∗∗ −0.35ns −0.69∗ −0.2ns −0.52ns

Stearic acid −0.81∗∗ −0.81∗∗ −0.45ns −0.79∗∗ −0.05ns −0.37ns 0.87∗∗

Oleic acid 0.06ns 0.12ns 0.27ns 0.4ns −0.57ns −0.4ns −0.18ns −0.4ns

Linoleic acid 0.42ns 0.29ns −0.26ns −0.09ns 0.74∗∗ 0.79∗∗ −0.29ns −0.12ns −0.75∗∗

Linolenic acid −0.85∗∗ −0.73∗∗ −0.17ns −0.44ns −0.39ns −0.79∗∗ 0.87∗∗ 0.76∗∗ 0.05ns −0.6∗

α-Tocopherol −0.23ns −0.25ns 0.16ns −0.07ns −0.53ns −0.44ns −0.05ns −0.16ns 0.75∗∗ −0.6∗ 0.08ns

γ -Tocopherol −0.41ns −0.37ns 0.07ns −0.18ns −0.6∗ −0.54ns 0.17ns 0.03ns 0.76∗∗ −0.7∗∗ 0.95∗∗∗ 0.95∗∗∗

δ-Tocopherol −0.41ns −0.37ns 0.07ns −0.18ns −0.6∗ −0.55ns 0.17ns 0.03ns 0.76∗∗ −0.7∗∗ 0.95∗∗∗ 0.99∗∗∗ 0.99∗∗∗

Significance: ∗ P < 0.05; ∗∗ P < 0.01: ∗∗∗ P < 0.001; ns, not significant.

REFERENCES1 Reymond P and Farmer EE, Jasmonate and salicylate as global signals

for defense gene expression. Curr Opin Plant Biol 1:404–411 (1998).2 Steduto P, Albrizio R, Giorio P and Sorrentino G, Gas exchange

response and stomatal and non-stomatal limitations to carbonassimilation of sunflower under salinity. Environ Exp Bot 44:243–255(2000).

3 Shakirova FM, Sakhabutdinova AR, Bezrukova MV, Fathutdinova RAand Fathutdinova DR, Changes in hormonal status of wheatseedlings induced by salicylic acid and salinity. Plant Sci164:317–322 (2003).

4 Ashraf M, Inducing drought tolerance in plants: recent advances.Biotechnol Adv 28:169–183 (2010).

5 Senaratna T, Touchell D, Bunn T and Dixon K, Acetyl salicylic (aspirin)and salicylic acid induce multiple stress tolerance in bean andtomato plants. Plant Growth Regul 30:157–161 (2000).

6 Khodary SEA, Effect of salicylic acid on growth, photosynthesis andcarbohydrate metabolism in salt stressed maize plants. Int J AgricBiol 6:5–8 (2004).

7 Borsani O, Valpuesta V and Botella MA, Evidence for a role of salicylicacid in the oxidative damage generated by NaCl and osmotic stressin Arabidopsis seedlings. Plant Physiol 126:1024–1030 (2001).

8 El-Tayeb MA, Response of barley grains to the interactive effect ofsalinity and salicylic acid. Plant Growth Regul 45:215–224 (2005).

9 Gunes A, Inal A, Alpaslan M, Eraslan F, Bagci EG and Cicek N,Salicylic acid induced changes on some physiological parameterssymptomatic for oxidative stress and mineral nutrition in maize (Zeamays L.) grown under salinity. J Plant Physiol 164:728–736 (2006).

10 Singh G and Kaur M, Effect of growth regulators on podding andyield of mung bean (Vigna radiata L.) Wilezck. Indian J Plant Physiol23:366–370 (1980).

11 Rendon SLA, Hormonal control of reproductive organs in Phaseolusvulgaris L. cv. Cacahuate-72. Master of Science Thesis, C.P. Chapingo,Mexico, (1983).

12 Arfan M, Athar HR and Ashraf M, Does exogenous application ofsalicylic acid through the rooting medium modulate growthand photosynthetic capacity in differently adapted spring wheatcultivated under salt stress? J Plant Physiol 6:685–694 (2007).

13 Francois LE, Salinity effects on four sunflower hybrids. Agron J88:215–219 (1996).

14 Richards RA, Increasing salinity tolerance of grain crops. Is itworthwhile? Plant Soil 146:89–98 (1992).

15 Flowers TJ and Lauchli A, Sodium versus potassium: substitution andcompartmentation, in Inorganic Plant Nutrition, Vol. 15B, ed. byLauchli A and Bieleski RL. Springer, Berlin, pp. 651–681 (1983).

16 Maas EV, Salt tolerance in plants. Appl Agric Res 1:12–26 (1986).17 Farah MA, Daoud AM, Barakat MA and Anta IM, The effect of the depth

to water table and salinity of irrigation water on growth and variousparameters of sunflower. Agric Res Rev 59:199–209 (1984).

18 Raju KV and Ranganayakulu C, Effect of saline water irrigation on saltdistribution in profile and performance of sunflower crop. Indian JAgric Res 12:183–186 (1978).

19 Flagella Z, Giuliani MM, Rotunno T, Caterina RD and De Caro A, Effectof saline water on oil yield and quality of a high oleic sunflower(Helianthus annuus L.) hybrid. Eur J Agron 21:267–272 (2004).

20 Irving DW, Shannon MC, Breda VA and Mackey BE, Salinity effects onyield and oil quality of high-linoleate and high-oleate cultivars ofsafflower (Carthamus tinctorius L.). J Agric Food Chem 36:37–42(1988).

21 Bassil ES and Kaffka SR, Response of safflower (Carthamus tinctorius L.)to saline soils and irrigation: II. Crop response to salinity. Agric WaterManag 54:81–92 (2002).

22 Zarrouk M, Marzouk B, Ben-Miled D and Cherif A, Oil accumulationin olives and effect of salt on their composition. Olivae 61:41–45(1996).

23 Heuer B, Yaniv Z and Ravina I, Effect of late salinization of chia (Salviahispanica), stock (Matthiola tricuspidata) and evening primrose(Oenothera biennis) on their oil content and quality. Indian CropProd 15:163–167 (2002).

24 Lagravere T, Kleiber D and Dayde J, Perfomance agronomique etconduits culturales du tournesol oleique. Realites et perspectives.Oleagineux, Corps Gras, Lipides 5:477–485 (1998).

25 Francois LE and Kleiman R, Salinity effects on vegetative growth, seedyield, and fatty acid composition of crambe. Agron J 82:1110–1114(1990).

26 Baldini M, Giovanardi R, Tahmasebi-Enferadi S and Vannozzi GP,Effects of water regime on fatty acid accumulation and final fattyacid composition in the oil of standard and high oleic sunflowerhybrids. Ital J Agron 6:119–126 (2002).


26

16


27 Hoagland DR and Arnon DI, The water culture method for growingplants without soil. Calif Agric Exp Sta Circ. University of California,Berkeley, 347 p. 32 (1950).

28 Lee BL, New AL and Ong CN, Simultaneous determination oftocotrienols, tocopherols, retinol, and major carotenoids in humanplasma. Clin Chem 49:2056–2066 (2003).

29 Snedecor GW and Cochran WG, Statistical Methods (7th edn). The IowaState University Press, Ames, IA (1980).

30 Ashraf M, Some important physiological selection criteria for salttolerance in plants. Flora 199:361–376 (2004).

31 Munns R and Tester D, Mechanisms of salinity tolerance. Annu RevPlant Biol 59:651–681 (2008).

32 Ashraf M and Akram NA, Improving salinity tolerance of plants throughconventional breeding and genetic engineering: an analyticalcomparison. Biotechnol Adv 27:744–752 (2009).

33 Noreen S and Ashraf M, Alleviation of adverse effects of salt-stresson sunflower (Helianthus annuus L.) by exogenous application ofsalicylic acid: growth and photosynthesis. Pak J Bot 40:1657–1663(2008).

34 Noreen S, Ashraf M, Hussain M and Jamil A, Exogenous applicationof salicylic acid enhances antioxidative capacity in salt-stressedsunflower (Helianthus annuus L.) plants. Pak J Bot 41:473–479(2009).

35 Aristeo-Cortes P, Reguladores de crecimiento. XIV: Efectos del acidosalicılico ydimetilsulfoxido en el crecimiento de zanahoria, betabely rabano. Tesis de Licenciatura, Facultad de Ciencias UNAM, Mexico(1998).

36 Hayat S, Ali B and Ahmad A, Salicylic acid biosynthesis, metabolismand physiological role in plants, in Salicylic Acid. A Plant Hormone,ed. by Hayat S and Ahmad A. Springer, Berlin, pp. 1–14 (2007).

37 Cleland CF and Ajami A, Identification of the flower-inducing factorisolated from aphid honeydew as being salicylic acid. Plant Physiol54:904–906 (1974).

38 Piterse AH, A review of chemically induced flowering in Lemma gibbaG3 and Pistia stratiotes. Aquat Bot 13:21–28 (1982).

39 Kumar P, Lakshmi NJ and Mani VP, Interactive effects of salicylic acidand phytohormones on photosynthesis and grain yield of soybean(Glycine max L. Merrill). Physiol Mol Biol Plants 6:179–186 (2000).

40 Seth PN, Venkatarman R and Maheshwari SC, Studies on growth andflowering of a short-day plant, Wolffia microscopica. III. Role of metalions and chelates. Planta 90:349–359 (1970).

41 Oota Y, The response of Lemma gibba G3 to a single long day in thepresence of EDTA. Plant Cell Physiol 13:575–580 (1972).

42 Watanabe K and Takimoto A, Flower inducing effect of benzoic acidand some related compounds in Lamma paucicostata. Plant CellPhysiol 22:1469–1479 (1979).

43 Fujioka S, Yamaguchi I, Murofushi N, Takahashi N, Kaihara S andTakimoto A, The role of plant hormones and benzoic acid in

flowering of Lamma paucicostata. Plant Cell Physiol 24:241–246(1983).

44 Watanabe K, Fujita T and Takimoto A, Relationship between structureand flower inducing activity of benzoic acid derivatives in Lemmapaucicostata. Plant Cell Physiol 20:847–850 (1981).

45 Oota Y, Short-day flowering of Lamma gibba G3 induced by salicylicacid. Plant Cell Physiol 16:1131–1135 (1975).

46 Raskin I, Role of salicylic acid in plants. Annu Rev Plant Physiol Plant MolBiol 43:439–463 (1992).

47 Raskin I, Salicylate, a new plant hormone. Plant Physiol 99:799–803(1992).

48 Zhou XM, Mackeuzie AF, Madramootoo CA and Smith DLJ, Effect ofsome injected plant growth regulators, with or without sucrose,on grain production, biomass and photosynthetic activity of field-grown corn plants. J Agron Crop Sci 183:103–110 (1999).

49 Beke GJ and Volkmar KM, Mineral composition of flax (Linumusitatissimum L.) and safflower (Carthamus tinctorius L.) on a salinesoil high in sulphate salts. Can J Plant Sci 75:399–404 (1995).

50 Heuer B, Ravina I and Davidov S, Seed yield, oil content, and fatty acidcomposition of stock (Matthiola incana) under saline irrigation. AustJ Agric Res 56:45–47 (2005).

51 Kamal-Eldin A and Appelqvist LA, The chemistry and antioxidantproperties of tocopherols and tocotrienols. Lipids 31:631–701(1996).

52 Pongracz G, Weiser H and Matzinger D, Tocopherol. Antioxidant derNat. Fat Sci Technol 97:90–104 (1995).

53 Wise RR and Naylor AW, Chilling-enhanced photo-oxidation: evidencefor the role of singlet oxygen and endogenous antioxidants. PlantPhysiol 83:278–282 (1987).

54 Mittova V, Guy M, Tal M and Volokita M, Response of the cultivatedtomato and its wild salt-tolerant relative Lycopersicon pennellii tosalt-dependent oxidative stress: increased activities of antioxidantenzymes in root plastids. Free Radic Res 36:195–202 (2002).

55 Mittova V, Volokita M, Guy M and Tal M, Activities of SOD and theascorbate–gluthione cycle enzymes in subcellular compartmentsin leaves and roots of the cultivated tomato and its wild salt-tolerantrelative Lycopersicon pennellii. Physiol Plant 110:42–51 (2000).

56 Dat JS, Vandenabeele E, Vranova M, Montagu V, Inze D andBreusegem FV, Dual action of the active oxygen species duringplant stress responses. Cell Mol Life Sci 57:779–795 (2000).

57 Foyer CH and Noctor G, Redox sensing and signalling associated withreactive oxygen in chloroplasts, peroxisomes and mitochondria.Physiol Plant 119:355–364 (2003).

58 Anwar F, Hussain AI, Ashraf M, Jamil A and Iqbal S, Effect of salinityon yield and quality of Moringa oleifera seed oil. Grasas Aceites57:394–401 (2006).


26

17

Research ArticleReceived: 15 April 2010 Revised: 15 July 2010 Accepted: 16 July 2010 Published online in Wiley Online Library: 17 August 2010


Use of microfungi in the treatment of oak chips:possible effects on wineLeonardo Petruzzi,a Antonio Bevilacqua,a,b∗ Claudio Ciccarone,c

Giuseppe Gambacorta,a,b† Giuseppina Irlante,a Sandra Patia,b

and Milena Sinigagliaa,b

Abstract

BACKGROUND: Oak barrels are commonly used in the aging of wines and spirits because of their positive effects on the product.In recent years the addition of oak chips has been used to introduce desirable wood aromas and flavours into wines. In this study,oak chips in saline solution or laboratory medium were inoculated with Penicillium purpurogenum, Aureobasidium pullulans,Phialemonium obovatum, Phanerochaete chrysosporium and a combination of Ph. chrysosporium and A. pullulans. After 12 weeksof incubation, oak chips (2 g L−1) were macerated in a red wine for 17 days. Gas chromatography/mass spectrometry and high-performance liquid chromatography were used to evaluate 14 compounds, namely furfural, furfuryl alcohol, guaiacol, syringol,cis-β-methyl-γ -octalactone, 2-phenylethanol, 4-vinylguaiacol, benzyl alcohol, 2,3-butanediol, γ -butyrolactone, benzaldehyde,4-ethylguaiacol, gallic acid and ellagic acid.

RESULTS: The microfungal treatments increased the concentration of some components. In particular, P. purpurogenum resultedin a significant improvement in the levels of guaiacol, furfural, syringol, furfuryl alcohol and 2-phenylethanol.

CONCLUSION: Penicillium purpurogenum and Ph. chrysosporium showed a constant trend (enrichment of furfural andbenzaldehyde) independent to some extent of the medium used for chip treatment.c© 2010 Society of Chemical Industry

Keywords: wood; seasoning; oak chips; fungi; wine aging

INTRODUCTIONOak wood has traditionally been used in barrel making becauseof both its mechanical properties and extractable compounds,which can induce positive changes in the composition and flavourof aged wine.1 The use of oak barrels for wine aging involves ahigh cost outlay for wineries; therefore some suitable alternativesto this traditional method have been sought.2 In recent yearsthe addition of oak chips has been used to introduce desirableoak aromas and flavours into wines and accelerate the agingprocess.2 – 4 Oak chips, obtained from wood scrap wastes producedduring barrel manufacturing, are prepared following traditionalcooperage methods;5 their use is considered a valid alternative totraditional barrel aging or barrel fermentation.4

Perez-Coello et al.6 reported that the amount of compoundsreleased into the wine depends on the type of wood employed,the seasoning and charring treatments applied, as well as otherfactors, such as how long the wood is in contact with the wine,the temperature and humidity, etc. The production of staves forbarrels or chips includes an outdoor seasoning step that generallytakes between 24 and 36 months; this is not a simple processof drying but a refining stage, since the wood undergoes slowchemical and biochemical transformations of biopolymers andextractable compounds by fungi and bacteria that contribute toits maturation.7 The importance of mycotic wood infection is amatter discussed as a main theme in several studies.8 – 11

Focusing on the effect of fungi, basidiomycetes are the mostimportant organisms able to degrade lignocelluloses, mainly thewhite rot and related litter-decomposing fungi.12 In particular,Phanerochaete chrysosporium is the most studied among thewood-rotting fungi owing to its production of two peroxidases,lignin peroxidase (LP) and manganese peroxidase (MnP), knownto play an important role in the degradation of lignin.13

Vivas et al.10 studied the evolution of the microflora of woodstaves throughout seasoning and found that 83% of the totalfungal community was represented by Aureobasidium pullulans,followed by 15% of Trichoderma harzianum and Trichodermakonigii. These three fungi have an enzymatic activity able to

∗ Correspondence to: Antonio Bevilacqua, Department of Food Science, Facultyof Agricultural Science, University of Foggia, Via Napoli 25, I-71122 Foggia,Italy. E-mail: [email protected]; [email protected]

† Current address: Dipartimento di Biologia e Chimica Agro-forestalee Ambientale (DiBCA), University of Bari, Via Amendola 165, Bari, Italy.

a Department of Food Science, Faculty of Agricultural Science, University ofFoggia, Via Napoli 25, I-71122 Foggia, Italy

b Food Quality and Health Research Center (BIOAGROMED), University of Foggia,Via Napoli 52, I-71122 Foggia, Italy

c Department of Agro-Environmental, Chemistry and Crop-Protection, Faculty ofAgricultural Science, University of Foggia, Via Napoli 25, I-71122 Foggia, Italy


26

18

www.soci.org L Petruzzi et al.

hydrolyse many wood heterosides (ellagitannins, coumarins andpolysaccharides), thus resulting in a decrease in astringency andbitterness.9,11

In particular, A. pullulans is a ubiquitous species producingcellulosolytic, pectinolytic and ligninolytic enzymes and is able toadapt to different environmental conditions.

After 24–36 months of stave seasoning, it is possible to note anevolution of the fungal community, with an increase in species ofthe genera Penicillium, Geomyces and Geotrichum; throughout thisperiod, Penicillium purpurogenum is the secondary species mostrepresented.10 This fungus is an active producer of a variety ofxylan-degrading enzymes and several esterases.14

An ecological study carried out by Roulland et al.8 revealed thatCandida sp., Paecilomyces variotii, Phialemonium sp. and the strainE were the moulds that colonised the internal layers of staves. Theauthors measured the ability of these species to degrade cellulose,xylan and aesculin: in a synthetic liquid medium the four fungalspecies degraded aesculin and xylan, but only Phialemonium sp.showed a measurable growth rate in a cellulose medium.8

Some authors hypothesised the possibility to inoculate staveswith selected mould strains to control metabolic reactions of fungi;in particular, the use of A. pullulans was suggested.11 – 13 The use ofselected fungal inocula on a large scale could lead to the creationof new seasoning places.

Jourez et al.15 reported that a speeding-up of the aging ofstaves is possible by treatment with a solution of natural enzymesobtained from fungal species usually colonising wood staves;this enzymatic treatment could be considered a very interestingeconomical alternative, permitting a reduction in storage time ofthe shooks from 12–36 months to only 1 month before assembly.

Therefore, as an alternative to artificial drying, this paperproposes the treatment of chips with four fungi (P. purpurogenum,A. pullulans, Phialemonium obovatum and Ph. chrysosporium)and a combination of two of them (Ph. chrysosporium + A.pullulans), representative of both the primary and secondarymicroflora recovered in the natural seasoning of staves. Thistreatment was proposed to improve the impact of oak chips inthe maceration of red wine and attain an effect comparable tothat observed in barrel aging. The chips were inoculated with theaforementioned fungi both in saline solution and in laboratorymedium and used after 12 weeks for the accelerated aging of ared wine in order to study the effect of the fungal inocula onthe concentration of volatile compounds (furfural, furfuryl alcohol,guaiacol, syringol, cis-β-methyl-γ -octalactone, 2-phenylethanol,4-vinylguaiacol, benzyl alcohol, 2,3-butanediol, γ -butyrolactone,benzaldehyde and 4-ethylguaiacol) and gallic and ellagic acids inthe wine. Gas chromatography/mass spectrometry (GC/MS) andhigh-performance liquid chromatography (HPLC) were used toevaluate these 14 compounds.

EXPERIMENTALOak chip treatmentMicro-organismsThe following fungi purchased from the culture collection ofMycotheca Universitatis Taurinensis (MUT), University of Torino,Italy were used throughout this study: Ph. chrysosporium Burds.(MUT 2660), P. purpurogenum Stoll (MUT 3316), A. pullulans (deBary) G. Arnaud (MUT 3237) and Phi. obovatum W. Gams & McGinnis(MUT 2702). All fungi were non-mycotoxinogenic.

Fungi were maintained on potato dextrose agar (DifcoLaboratories, Detroit, MI, USA) plates at 4 ◦C, then revived on malt

extract agar (Difco Laboratories) plates and kept at 70% relativehumidity in complete darkness as follows: Ph. chrysosporium wasincubated for 7 days at 37 ◦C; P. purpurogenum was incubated for7 days at 28 ◦C; A. pullulans and Phi. obovatum were incubated for14 days at 24 ◦C. Afterwards, an agar plug (8 mm in diameter, cutalong the edge of an actively growing colony) of each fungus wasused for preparation of the inoculum for oak chips (reported in thefollowing subsections and divided into different steps: preculture,fungal inoculum preparation and oak chip inoculation).

PrecultureFungi were grown on Petri plates (90 mm in diameter) with dif-ferent lignocellulosic supplements: wheat bran (Ph. chrysosporiumand A. pullulans), wheat straw (P. purpurogenum) and oat bran(Phi. obovatum). The medium was prepared according to Moredoet al.16 and comprised 10 g L−1 glucose (J.T. Baker, Milan, Italy),15 g L−1 agar technical No. 3 (Oxoid, Milan, Italy), 3.5 g L−1 maltextract (Oxoid) and 7.5 g L−1 triturated lignocellulosic material.

Fungi were incubated at 30 ◦C for 7 days (Ph. chrysosporium andP. purpurogenum), 10 days (A. pullulans) or 14 days (Phi. obovatum).

Fungal inoculum preparationEach inoculum was prepared in several steps. First, the centralpart (5 cm in diameter) of the colony cultivated on a precultureplate was blended for 30 s with 50 mL of sterile distilled water in alaboratory blender (Stomacher 400, PBI International, Milan, Italy).Then 5 mL of homogenate was added to a 250 mL Erlenmeyer flaskcontaining 50 mL of growth medium (sterilised in an autoclave at121 ◦C for 15 min) (see Appendix). All experiments were done induplicate.

Cultures were incubated statically (Ph. chrysosporium, 37 ◦C for7 days; P. purpurogenum, 37 ◦C for 7 days; Phi. obovatum, 28 ◦C for6 days) or with agitation on a rotary shaker (A. pullulans, 28 ◦C for3–5 days, 200 rpm). The flasks, capped with cellulose stoppers topermit passive aeration, were maintained in complete darkness at90% relative humidity.

Afterwards, fungus and medium were aseptically homogenisedin a blender (3 × 20 s cycles with 1 min interval). The resultinghomogenate (5 mL) was added to a 250 mL Erlenmeyer flaskcontaining 50 mL of growth medium (sterilised in an autoclave at121 ◦C for 15 min).

Cultures were incubated statically (Ph. chrysosporium, 30 ◦C for5 days) or with agitation (P. purpurogenum, 30 ◦C for 5–7 days,135 rpm; Phi. obovatum, 28 ◦C for 7 days, 150 rpm; A. pullulans,28 ◦C for 7 days, 200 rpm).

Afterwards, fungus and medium were aseptically homogenisedin a blender (Ph. chrysosporium and P. purpurogenum, 3 × 20 scycles with 1 min interval; A. pullulans and Phi. obovatum, 20 s).

The final homogenate (fungus+medium) of each fungal specieswas used to inoculate the oak chips.

Oak chip inoculationFrench chips toasted at low degree and of medium size (F4LBoiselevage Lavoisier, AEB Group SpA, Bologna, Italy) were used.A 3 mL aliquot of homogenate was added to a 250 mL Erlenmeyerflask containing 4 g of oak chips and 12 mL of production medium(laboratory medium) (see Appendix) or 12 mL of saline solution;chips were sterilised with laboratory medium or saline solution at121 ◦C for 15 min. The saline solution contained 3 g L−1 NH4NO3,3 g L−1 KH2PO and 0.5 g L−1 MgSO4· 7H2O;17 all components werepurchased from J.T. Baker.


26

19

Oak chips inoculated with moulds www.soci.org

Cultures were incubated for 12 weeks in darkness under staticconditions at different temperatures: 24–26 ◦C (P. purpurogenum,A. pullulans and Phi. obovatum) or 30 ◦C (Ph. chrysosporium). Eachexperiment was performed twice (on two independent batches).Aliquots of laboratory medium or saline solution containing oakchips but not inoculum were used as controls.

In addition to the single inocula, a combination of A. pullulansand Ph. chrysosporium was investigated. Oak chips were firstinoculated with A. pullulans for 6 weeks, then sterilised, chargedwith laboratory medium or saline solution and inoculated with Ph.chrysosporium for 6 weeks.

Wine agingAfter the 12 week incubation period, oak chips were removedfrom the flasks, cleaned with a small brush to remove visiblemycelia and used for aging red wine elaborated by the traditionalred winemaking process from two grape varieties, Montepulcianod’Abruzzo (70%) and Merlot (30%) (2006 vintage), from vineyardsin southern Italy (San Severo, Apulia).

Artificial aging was performed by placing 1 g (dry weight)portions of oak chips in 500 mL glass bottles containing red wineand storing the bottles in a cool dark room (temperature of about20 ◦C at 82% relative humidity) for 17 days.

Chemical analysesIdentification and quantification of volatile compounds by GCand GC/MSThe method of Di Stefano,18 appropriately modified, was used forthe extraction and concentration of volatile compounds (furfural,furfuryl alcohol, guaiacol, syringol, cis-β-methyl-γ -octalactone,2-phenylethanol, 4-vinylguaiacol, benzyl alcohol, 2,3-butanediol,γ -butyrolactone, benzaldehyde and 4-ethylguaiacol).

Liquid/liquid extraction was carried out in a glass flask atroom temperature under magnetic stirring, using 7.5 mL ofdichloromethane (J.T. Baker) as extracting solvent, 0.25 mL of200 mg L−1 2-octanol solution (Fluka, Buchs, Switzerland) asinternal standard and 50 mL of wine. The emulsion was recoveredafter 1 h of extraction and frozen for 2 h at −20 ◦C. After freezing,the organic phase was passed through two separation funnels(100 and 25 mL respectively), dried over Na2SO4 (J.T. Baker) andconcentrated to 1.5 mL under a stream of pure N2 (0.5 L min−1) forGC and GC/MS analyses.

A 1 µL aliquot of concentrated extract was injected in splitlessmode into an Agilent 6890N gas chromatograph (AgilentTechnologies, Palo Alto, CA, USA) equipped with a split/splitlessinjector, a DB-Wax fused silica capillary column (60 m length,0.25 mm i.d., 0.25 µm film thickness; J&W Scientific, Folsom, CA,USA) and a flame ionisation detector.

Helium was used as the carrier gas at a linear velocity of37 cm s−1. The oven temperature was programmed as follows:40 ◦C for 3 min; from 40 to 200 ◦C at a rate of 4 ◦C min−1; 200 ◦Cfor 20 min. The detector and injector temperatures were 240 and230 ◦C respectively.

Identification of compounds was performed using an Agilent5975C quadrupole mass spectrometer coupled with an Agilent6890N gas chromatograph. The same column and same tem-perature programme as for GC analysis were employed. Electronimpact mass spectra were recorded with an ion source energy of70 eV.

Chromatographic peaks were identified by comparing theirmass spectra with those of standards (Sigma-Aldrich, St Louis, MO,USA) and those reported in the NIST 2.0 commercial library.

Identification and quantification of gallic and ellagic acids by HPLCAcid hydrolysis was used to determine the content of gallic andellagic acids,19 modified as follows: 7.5 mL of HCl was added to10 mL of wine and heat-shocked for 3 h in a water bath held at100 ◦C.

After filtration through a 13 mm polytetrafluoroethylene filter(0.20 µm pore size), a 10 µL sample was directly injected into anAgilent 1200 HPLC system equipped with an autosampler, a binarypump, a diode array detector and a 4 µm Synergi MAX-RP C12column (150 mm × 2 mm; Phenomenex, Milan, Italy).

The eluent was fed at 0.2 mL min−1 and consisted of twosolvents: A, water; B, acetonitrile. Both solvents were acidified with10 mL L−1 formic acid for the determination of ellagic acid andwith 20 mL L−1 formic acid for the determination of gallic acid.Acetonitrile and water were of HPLC ultra gradient grade and werepurchased from J.T. Baker.

The elution gradient for the determination of ellagic acid wasas follows: from 0 to 4 min, 18% B; from 4 to 14 min, 20% B; thenan increase to 90% B (time 14.2 min). The elution gradient for thedetermination of gallic acid was as follows: from 0 to 10 min, 1%B; then an increase to 95% B (time 10.2 min).

The spectra were obtained by scanning from 225 to 390 nm.Ellagic acid was detected at 250 nm and gallic acid at 280 nm.

Peaks were identified via spectra and retention times ofstandards (Sigma-Aldrich). Chemstation Version 03.01 (AgilentTechnologies) was used for data acquisition and analysis.

Quantification was achieved by means of a calibration curveobtained by injecting solutions with variable and known amountsof standards in order to cover the desired concentration range.The main calibration characteristics are reported in Table 1.

Statistical analysisAll analyses were performed in duplicate over two differentbatches.

Data were submitted to one-way analysis of variance (ANOVA)and Tukey’s test using Statistica for Windows (StatSoft, Tulsa, OK,USA).

In addition, differences among samples were determined byprincipal component analysis (PCA) using the Excel componentXLSTAT 7 (Addinsoft, Paris, France).

RESULTS AND DISCUSSIONIt is known that some compounds, namely furfural, furfuryl alcohol,guaiacol, syringol and cis-β-methyl-γ -octalactone, are present inwine largely as a result of their extraction from oak wood. Othercompounds such as 2-phenylethanol and 4-vinylguaiacol are tosome extent fermentation products, but oak aging has been shownto increase levels of these chemicals to a certain degree.

Benzyl alcohol, 2,3-butanediol, γ -butyrolactone and benzalde-hyde are grape or fermentation products and are not knownto be affected by oak aging.20 4-Ethylguaiacol, produced by thecontaminant yeasts Brettanomyces and Dekkera, is related to the‘Brett character’ of spoiled wines and has been associated withdescriptive expressions such as ‘bacon’ or ‘smoked’.21

Gallic and ellagic acids were studied in this research becausethey contribute to wine astringency and bitterness.22

It is interesting to note that cis-β-methyl-γ -octalactone, whichis associated with oaky, coconut and vanilla notes, appeared intrace concentrations in all samples.4,23 The concentration of cis-β-methyl-γ -octalactone is strongly related to the origin of the


26

20


Table 1. Main calibration( characteristics related to ellagic and gallic acid determination

CompoundRetentiontime (min) Linear regression equation

Detection limit(S/N = 3) mg L−1)

Concentration rangeof calibration (mg L−1)

Ellagic acid 8.15 y = 268.06x − 136.58 (R2 = 0.9987) 0.6 1–10

Gallic acid 6.12 y = 177.62x − 2.91 (R2 = 0.9999) 0.07 2–10

S/N, signal/noise ratio; R2, coefficient of determination.

Table 2. Chemicals recovered in wine aged with oak chips treated in (a) saline solution (S) or (b) laboratory medium (M) and inoculated with fungi

(a)

Compound TS PCS PPS APS POS A+PS

Furfural (µg L−1) 22.45 ± 0.85 113.5 ± 11.5 130 ± 10 25.5 ± 1.5 16.5 ± 1.5 20 ± 7

Guaiacol (µg L−1) 13.5 ± 3.53 11.5 ± 3.53 44.5 ± 3.53 62.5 ± 6.36 50.5 ± 9.19 19 ± 7.07

Syringol (µg L−1) 227.5 ± 45.96 232.5 ± 38.89 615 ± 91.92 362 ± 59.39 502.5 ± 51.61 269.5 ± 71.41

Furfuryl alcohol (µg L−1) 100 ± 14.14 43.5 ± 10.6 57.5 ± 17.67 116 ± 2.82 72 ± 1.41 115 ± 7.07

2-Phenylethanol (mg L−1) 9.57 ± 0.26 9.51 ± 1.52 10.47 ± 1.73 9.34 ± 0.19 11.67 ± 0.65 8.2 ± 0.65

4-Vinylguaiacol (µg L−1) 66 ± 31.11 213 ± 5.65 162 ± 26.87 342.5 ± 45.96 401 ± 24.04 376.5 ± 38.98

Benzyl alcohol (mg L−1) 3.58 ± 0.16 1.59 ± 0.02 1.62 ± 0.13 2.62 ± 0.09 2.62 ± 0.28 2.14 ± 0.42

2,3-Butanediol (mg L−1) 11.01 ± 0.97 3.8 ± 0.39 3.26 ± 1.49 9.99 ± 0.02 3.66 ± 1.04 4.92 ± 1.01

γ -Butyrolactone (mg L−1) 0.4 ± 0.17 6.65 ± 0.78 5.6 ± 0.17 6.98 ± 0.01 7.19 ± 0.26 5.27 ± 0.07

Benzaldehyde (µg L−1) 42.5 ± 9.19 74.5 ± 9.19 32.5 ± 9.19 6 ± 2.82 4 ± 1.41 7.1 ± 0.14

4-Ethylguaiacol (µg L−1) 57.5 ± 10.6 75 ± 7.07 55 ± 2.82 56.5 ± 6.36 65.5 ± 9.19 76 ± 7.07

Ellagic acid (mg L−1) 7.18 ± 0.86 2.81 ± 0.8 3.22 ± 0.78 1.32 ± 0.25 1.14 ± 0.5 2.14 ± 0.44

Gallic acid (mg L−1) 8.72 ± 0.36 7.39 ± 0.64 7.81 ± 1.44 4.75 ± 0.35 3.49 ± 0.58 5.04 ± 1.08

cis-β-Methyl-γ -octalactone (µg L−1) Tr Tr Tr Tr Tr Tr

(b)

Compound TM PCM PPM APM POM A+PM

Furfural (µg L−1) 16.5 ± 2.5 168.5 ± 5.5 185 ± 5 14 ± 3 23.5 ± 4.5 21.5 ± 2.5

Guaiacol (µg L−1) 2.85 ± 0.77 8.5 ± 3.53 50 ± 11.31 119 ± 15.55 17.5 ± 3.53 15 ± 7.07

Syringol (µg L−1) 35.5 ± 3.53 123.5 ± 0.7 580 ± 84.85 725.5 ± 92.63 139.5 ± 57.27 225 ± 63.63

Furfuryl alcohol (µg L−1) 11.5 ± 2.12 255 ± 35.35 230 ± 28.28 120 ± 28.28 106.5 ± 9.19 105 ± 21.21

2-Phenylethanol (mg L−1) 0.88 ± 0.11 10.49 ± 0.57 47.99 ± 0.14 10.06 ± 2.04 11.28 ± 0.21 8.58 ± 1.52

4-Vinylguaiacol (µg L−1) 26 ± 2.82 123.5 ± 0.7 230 ± 55.15 228.5 ± 45.96 210.5 ± 38.89 145.5 ± 44.54

Benzyl alcohol (mg L−1) 0.16 ± 0.007 1.88 ± 0.02 1.93 ± 0.07 2.26 ± 0.35 4.21 ± 0.65 3.5 ± 0.68

2,3-Butanediol (mg L−1) 10.39 ± 0.74 1.5 ± 0.62 2.68 ± 0.72 3.31 ± 1.15 5.34 ± 1.22 7.44 ± 1.04

γ -Butyrolactone (mg L−1) 0.49 ± 0.02 6.19 ± 0.14 6.26 ± 1.15 5.36 ± 0.007 7.96 ± 0.29 6.11 ± 1.07

Benzaldehyde (µg L−1) 4 ± 1.41 88 ± 12.72 45 ± 8.48 55.5 ± 12.02 9.5 ± 2.12 79.5 ± 16.26

4-Ethylguaiacol (µg L−1) 48.5 ± 9.19 53 ± 1.41 63.5 ± 7.77 56.5 ± 6.36 65 ± 7.07 83.5 ± 2.12

Ellagic acid (mg L−1) 8.21 ± 1 1.37 ± 0.22 6.65 ± 1.83 1.59 ± 0.02 1.36 ± 0.43 1.97 ± 0.38

Gallic acid (mg L−1) 9.34 ± 0.25 8.31 ± 0.89 9.25 ± 1.76 9.23 ± 0.14 3.95 ± 1.24 3.96 ± 0.1

cis-β-Methyl-γ -octalactone (µg L−1) Tr Tr Tr Tr Tr Tr

Data are the mean of two values ± standard deviation. TS/TM, control; PCS/PCM, Ph. chrysosporium; PPS/PPM, P. purpurogenum; APSS/APM, A.pullulans; POS/POM, Phi. obovatum; A+PS/A+PM, A. pullulans + Ph. chrysosporium; Tr, trace amount.

wood and the country of seasoning.23 Moreover, Guchu et al.4

reported that non-toasted oak chips released more oak lactonesin wine than toasted oak chips did, probably owing to the thermaldegradation of these heat-sensitive compounds or their loss byvolatilisation when oak wood is treated at very high temperatures.

Concerning volatile compounds and phenols, their concentra-tion was influenced by the kind of medium and the mould usedfor the chip treatment. Table 2 reports the concentrations of winecompounds. The chemicals for which an influence of the fungalinoculum was found are discussed in the following sections.

FurfuralThe concentration of furfural was significantly affected by Ph.chrysosporium (PC) and P. purpurogenum (PP) and increased from22.45 to 113.5–130 µg L−1 and from 16.5 to 168.5–185 µg L−1

for saline solution- and laboratory medium-treated oak chipsrespectively, without attaining the perception threshold level(14 100 µg L−1)24 (Fig. 1). In the other samples, no significantchanges were recorded. Furfural contributes to the characterof ‘dried fruits’, particularly to that of ‘burned almonds’.3 Owingto the high threshold, Garde-Cerdan et al.25 reported that this


26

21


a

bb

aa

aa

b

b

aa a

0

25

50

75

100

125

150

175

200

225

250

TS PCS PPS APS POS A+PS TM PCM PPM APM POM A+PM

Am

ou

nt

(µg

L-1

)

Perception threshold (1/60)

Figure 1. Amounts (± standard deviation) of furfural in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) and inoculatedwith fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A.pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are significantly different (one-way ANOVA and Tukey’s test,P < 0.05). Perception threshold: the line represents a fraction of the real breakpoint (1/60).

compound does not play an important role in the aroma of wine,although it might strengthen the aroma of lactones.

Furfural is formed by the degradation of hemicelluloses duringwood toasting and is an indicator of the relative toast levelof wood.3,24 Therefore Spillman et al.26 reported that furfuralappears to be the most susceptible of the various oak woodvolatile compounds to microbial transformations, being almosttotally transformed during alcoholic and malolactic fermentationand during red wine maturation.

GuaiacolThe treatment of oak chips with fungi caused a significant increasein the content of guaiacol. In particular in the case of salinesolution (Fig. 2), the increase was observed for chips inoculatedwith P. purpurogenum (PPS), A. pullulans (APS) and Phi. obovatum(POS) (the average value of this compound was 44.5–62.5 µg L−1,compared with 13.5 µg L−1 for non-treated chips). On the otherhand, in the case of laboratory medium the differences weresignificant only for P. purpurogenum (PPM) and A. pullulans (APM)(50 and 119 µg L−1 respectively). It is interesting to note that inthe case of A. pullulans the content of guaiacol was above itsperception threshold (75 µg L−1).27

Guaiacol is produced by the breakdown of lignin during woodtoasting and is responsible for the burnt overtones of wine aroma.3

SyringolFor saline solution-treated chips the increase in syringol was sig-nificant only in the case of P. purpurogenum (PPS) (615 µg L−1)in comparison with the control (227.5 µg L−1) (Fig. 3). In lab-oratory medium the control showed a syringol content of35.5 µg L−1, while higher levels (above the perception thresholdof 570 µg L−1)28 were found in wine treated with oak chips inoc-ulated with P. purpurogenum (PPM) (580 µg L−1) and A. pullulans(APM) (725.5 µg L−1).

Syringol is synthesised during wood toasting and is an indicatorof the relative toast level of wood. In comparison with guaiacol,syringol has a weak odour and little impact on wine flavour.29

BenzaldehydeIn saline solution the fungal treatment decreased the concentra-tion of benzaldehyde (ranging between 4 and 32.5 µg L−1, whilein the control its content was 42.5 µg L−1), with the sole exceptionof Ph. chrysosprorium (PCS) (74.5 µg L−1) (Fig. 4). In the case of oakchips aged in laboratory medium, a significant increase in ben-zaldehyde was observed (45–88 µg L−1, compared with 4 µg L−1

in the control).A characteristic bitter almond odour is generally attributed to

the presence of benzaldehyde, which has a perception thresholdof 20 mg L−1.30 Benzaldehyde in wine is probably formed by theoxidation of benzyl alcohol,31 which is occasionally used as aplasticiser in epoxy resins or found as a contaminant in liquidgelatin.32 Benzaldehyde in wine is also formed by the action ofmicro-organisms on aromatic amino acids (e.g. phenylalanine),phenol compounds of the grape or some secondary compoundssuch as phenyl acetic acid and p-hydroxybenzoic acid.31 Severalmicro-organisms can oxidise benzyl alcohol to benzaldehyde,including Botrytis cinerea and the yeasts Schizosaccharomycespombe and Zygosaccharomyces bailii.32

2,3-ButanediolFigure 5 shows the content of 2,3-butanediol. Saline solutionand laboratory medium showed similar trends. In saline solutionthe level of this chemical was lower in wine samples aged withoak chips inoculated with fungi (3.26–4.92 mg L−1, comparedwith 11.01 mg L−1 in the control), with the sole exception of A.pullulans (APS). Similar results were obtained for oak chips treatedin laboratory medium.


26

22


a a

b

b

b

a

aa

b

c

a a

Perception threshold

0

30

60

90

120

150

180


Am

ou

nt (

µg L

-1)

Figure 2. Amounts (± standard deviation) of guaiacol in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) and inoculatedwith fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A.pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are significantly different (one-way ANOVA and Tukey’s test,P < 0.05).

a a

c

a, b

b, c

a, b

a

a

b

b

a

a

Perception threshold

0

100

200

400

300

600

500

700

800

900


Am

ou

nt

(µg

L-1

)

Figure 3. Amounts (± standard deviation) of syringol in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) and inoculatedwith fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A.pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are significantly different (one-way ANOVA and Tukey’s test,P < 0.05).

Although 2,3-butanediol is odourless, as its perception thresh-old is very high (600 mg L−1), it contributes to the sweet taste inwine.33 Bartowsky and Henschke34 reported that 2,3-butanediol isa stable compound derived from the reduction of acetoin.

Influence of chip fungal treatment on ellagic and gallic acidsFigure 6 presents the results for ellagic acid. In saline solutionthis compound was found at mean levels ranging from 1.14 to3.22 mg L−1 (compared with 7.18 mg L−1 for the control), whilein laboratory medium it varied from from 1.36 to 1.97 mg L−1

(compared with 8.21 mg L−1 for the control), thus suggesting

that ellagic acid was probably metabolised by fungi, the onlyexception being represented by P. purpurogenum in laboratorymedium (PPM) (6.65 mg L−1). Similar results were found in thecase of gallic acid (data not shown).

A possible explanation can be found in the paper of Roullandet al.,8 who reported that some fungi (in particular Phialemoniumsp.) were able to metabolise these two acids.

Principal component analysisAs a final step in this paper we propose a multivariate approach totry to give a simple insight into a complex phenomenon, such as the


26

23


c

d

b, c

a a a a

c

b

b, c

a

b, c


0

25

75

50

100

125

150


Am

ou

nt

(µg

L-1

)

Figure 4. Amounts (± standard deviation) of benzaldehyde in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) andinoculated with fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum;A+P, A. pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are significantly different (one-way ANOVA and Tukey’stest, P < 0.05). Perception threshold: the line represents a fraction of the real breakpoint (1/15).

0

3

6

9

12

15

18

21

24

a

b b

a

bb

d

aa, b

a, b

b, c

c, d

1/60



Am

ou

nt

(µg

L-1

)

Figure 5. Amounts (± standard deviation) of 2,3-butanediol in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) andinoculated with fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum;A+P, A. pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are significantly different (one-way ANOVA and Tukey’stest, P < 0.05). Perception threshold: the line represents a fraction of the real breakpoint (1/30).

pretreatment with some selected fungi and the chip effect on thechemical profile of phenols and volatile compounds of wine. Onlythose compounds showing significant differences among samples,i.e. furfural, benzaldehyde, 2,3-butanediol, guaiacol, syringol, gallicacid and ellagic acid, were used as input data. The results are shownin Fig. 7. The different samples are represented as a function ofcomponents 1 and 2, which together accounted for 79% (salinesolution, Fig. 7(a)) and 68% (laboratory medium, Fig. 7(b)) of thetotal variance. Table 3 lists the correlation coefficients of eachcompound with axis components.

In the case of saline solution, benzaldehyde, guaicol, ellagicacid and gallic acid were associated with principal component

1 (PC1); on the other hand, principal component 2 showed agood correlation with 2,3-butanediol. Finally, syringol showeda partial correlation with both PC1 and PC2 (−0.571 and 0.599respectively).

Figure 7(a) shows the distribution of the different samples in a2D plot. As evidenced by this score plot, the fungal treatment ofchips significantly affected the chemical profile of wine, resultingin a distribution into two groups, i.e. group A represented byPhi. obovatum (POS), A. pullulans + Ph. chrysosporium (A+PS) andA. pullulans (APS) and group B represented by P. purpurogenum(PPS) and Ph. chrysosporium (PCS). Samples treated with oak chipsof group A showed an increase in guaiacol and syringol levels,


26

24


b

a a

a a

a

b

a

b

a aa

0

2

4

6

8

10

12


Am

ou

nt

(mg

L-1

)

Figure 6. Amounts (± standard deviation) of ellagic acid in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) and inoculatedwith fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A.pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are significantly different (one-way ANOVA and Tukey’s test,P < 0.05).

whereas samples treated with oak chips of group B showed higherconcentrations of furfural and benzaldehyde.

The effect of the fungi appeared to be quite variable and wasinfluenced by the kind of medium used for the treatment of chips,as evidenced by the score plot in Fig. 7(b). In particular, in the caseof oak chips treated with laboratory medium, component 1 wasassociated with 2,3-butanediol, guaiacol and syringol, while gallicacid and ellagic acid showed high correlation coefficients withcomponent 2.

Although the variance explained for laboratory medium-treatedchips was lower (68%), the PCA showed some qualitative trend,confirming the data obtained in saline solution, but with somedifferences. The samples were again distributed into two groups,i.e. group C represented by Phi. obovatum (POM) and A. pullulans+ P. purpurogenum (A+PM) and group D represented by P.purpurogenum (PPM), A. pullulans (APM) and Ph. chrysosporium(PCM). Group D was characterised by higher concentrations offurfural, benzaldehyde, guaiacol and syringol, thus confirming theeffects of P. purpurogenum and Ph. chrysosporium.

CONCLUSIONSOak chips are a suitable alternative to barrels to obtain, in amuch shorter period of time (17 days), wine with the peculiarcharacteristics given by oak wood.

Based on the results reported in this paper, we could suggestthat microfungal treatment of chips increased the concentrationof some components during the aging period of wine. Inparticular, some interesting data were obtained in the case of P.purpurogenum and Ph. chrysosporium, as they showed a constanttrend (enrichment of furfural and benzaldehyde) independent tosome extent of the medium used for chip treatment.

Objections to the proposed method could arise owing to thepossible production of toxins by fungi; however, the fungi usedthroughout this study are non-mycotoxigenic. Moreover, this

aspect should be considered as a fundamental prerequisite forthe selection of suitable strains.

Further investigations are required to provide a more completeunderstanding of the biology and enzymatic profile of fungi.The present promising results suggest the application of thismethodology on the one hand to determine the extent of therelationship between P. purpurogenum and oak chips as a functionof their different sizes and degrees of toasting and on the otherhand to study other fungi (e.g. Trichoderma sp.) that colonise oakwood more or less deeply during seasoning or other white rotfungi such as Bjerkandera adusta and Trametes versicolor.

APPENDIX: MEDIA FOR FUNGIGrowth mediaA liquid medium adapted from Keller et al.35 was used forPh. chrysosporium. It consisted of 3 g L−1 NaNO3, 0.5 g L−1 KCl,0.5 g L−1 MgSO4· 7H2O, 0.5 g L−1 FeSO4· 7H2O, 1 g L−1 KH2PO4,20 g L−1 glucose and 1 g L−1 wheat bran. All components werepurchased from J.T. Baker.

The liquid medium described by Kheng and Omar36 was usedfor P. purpurogenum, with some modification. It was preparedwith the following composition: 0.3 g L−1 urea (Sigma-Aldrich),0.7 g L−1 bacteriological peptone (Oxoid), 0.2 g L−1 yeast extract(Oxoid), 1 g L−1 KH2PO4, 0.3 g L−1 CaCl2 (J.T. Baker), 1.4 g L−1

(NH4)2SO4 (J.T. Baker), 0.3 g L−1 MgSO4· 7H2O, 10 g L−1 glucose,1 g L−1 wheat straw, 5 mg L−1 FeSO4· 7H2O, 1.6 mg L−1 MnSO4·4H2O, 1.4 mg L−1 ZnSO4· 7H2O (J.T. Baker) and 20 mg L−1 CoCl2·6H2O (J.T. Baker).

Aureobasidium pullulans was grown on the liquid mediumreported by Lee et al.,37 modified as follows: 2.5 g L−1 yeast extract,0.6 g L−1 (NH4)2SO4, 5 g L−1 KH2PO4, 1 g L−1 NaCl, 0.2 g L−1

MgSO4· 7H2O, 0.2 g L−1 chloramphenicol (C. Erba, Milan, Italy),50 g L−1 glucose and 1 g L−1 wheat bran.

A liquid medium adapted from De Jong et al.38 was used for Phi.obovatum. It contained 5 g L−1 bacteriological peptone, 3 g L−1


26

25


A+PS

POS

APS

PPS

PCS

TS

1

2

3

4

5

6

7

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

3

-3 -2 -1 0 1 2 3

-- Component 1 (49%) -->

-- C

om

po

nen

t 2

(30%

) --

>

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

(b)

(a)

2.5

-- C

om

po

nen

t 2

(27%

) --

>

A

B

A+PM

POM

APM

PPM

PCM

TM

1

2

34

5

6 7

-3 -2 -1 0 1 2 3

-- Component 1 (41%) -->

C

D

Figure 7. PCA of chemical profile of wine aged with oak chips treated in (a) saline solution (S) or (b) laboratory medium (M) and inoculated with fungi:T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A. pullulans +Ph. chrysosporium. Compounds: 1, furfural; 2, benzaldehyde; 3, 2,3-butanediol; 4, guaiacol; 5, syringol; 6, gallic acid; 7, ellagic acid.

Table 3. Compounds associated with principal components 1 (PC1)and 2 (PC2) in PCA: correlation coefficients

S M

Compound PC1 PC2 PC1 PC2

Furfural (1) 0.372 0.890 −0.528 0.082

Benzaldehyde (2) 0.855 0.381 −0.601 −0.535

2,3-Butanediol (3) 0.230 −0.826 0.885 0.225

Guaiacol (4) −0.834 0.071 −0.726 0.271

Syringol (5) −0.571 0.599 −0.808 0.306

Gallic acid (6) 0.905 0.185 −0.414 0.792

Ellagic acid (7) 0.828 −0.281 0.344 0.858

S, oak chips treated in saline solution; M, oak chips treated in laboratorymedium. The numbers in parentheses correspond to compoundnumbers in Fig. 7.

yeast extract, 1 g L−1 K2HPO4, 0.5 g L−1 MgSO4· 7H2O, 0.2 g L−1

chloramphenicol, 20 g L−1 glucose and 1 g L−1 oat bran.

Production mediaThe composition of the production media was almost the sameas reported for the growth media; however, they did not containlignocellulosic materials and there were some slight differences.

• The liquid medium for P. chrysosporium was supplementedwith 0.5 g L−1 asparagine (Sigma-Aldrich), 1 mg L−1 vitamin B1(Sigma-Aldrich) and 1 mL L−1 Tween 80 (Biolife, Milan, Italy);the content of glucose was only 10 g L−1.

• The liquid medium for P. purpurogenum was supplementedwith 0.2 mL L−1 Tween 80.

• The liquid medium for A. pullulans was supplemented with5 mL L−1 Tween 80; the content of glucose was only 30 g L−1.

• In the liquid medium for Phi. obovatum the content of MgSO4·7H2O was 4 g L−1 and the content of KH2PO4 was 5 g L−1; thecontent of glucose was only 15 g L−1.


26

26


REFERENCES1 Citron G, Uso del legno in enologia: specie botaniche utilizzate,

anatomia e classificazione. Inform Agrar 50:69–72 (2003).2 Garde-Cerdan T and Ancın-Azpilicueta C, Review of quality factors on

wine ageing in oak barrels. Trends Food Sci Technol 17:438–447(2006).

3 Arapitsas P, Antonopoulos A, Stefanou E and Dourtoglou VG, Artificialageing of wines using oak chips. Food Chem 86:563–570 (2004).

4 Guchu E, Dıaz-Maroto MC, Perez-Coello MS, Gonzales-Vinas MAand Cabezudo Ibanez MD, Volatile composition and sensorycharacteristics of Chardonnay wines treated with American andHungarian oak chips. Food Chem 99:350–359 (2006).

5 Bozalongo R, Carrillo JD, Torroba MAF and Tena MT, Analysis ofFrench and American oak chips with different toastingdegrees by headspace solid-phase microextraction–gaschromatography–mass spectrometry. J Chromatogr A 1173:10–17(2007).

6 Perez-Coello MS, Sanchez MA, Garcıa E, Gonzales-Vinas MA, Sanz Jand Cabezudo MD, Fermentation of white wines in the presenceof wood chips of America and French oak. J Agric Food Chem48:885–889 (2000).

7 Ribereau-Gayon P, Glories Y, Maujean A and Dubourdieu D, Handbookof Enology. The Chemistry of Wine, Stabilization and Treatments (2ndedn), Vol. 2. Wiley, Chichester (2006).

8 Roulland C, Snakkers G and Cantagrel R, Approche experimentale durole des micro-organismes dans le processus de maturation desbois de tonnellerie. J Int Sci Vigne Vin 33:67–78 (1999).

9 Vivas N, Glories Y, Doneche B and Gueho E, Observation sur lamicroflore du bois de chene (Quercus sp.) au cours de son sechagenaturel. Ann Sci Nat Bot 11:149–153 (1991).

10 Vivas N, Saint-Cricq De Gaulejac N, Doneche B and Glories Y, Incidencede la duree du sechage naturel de Quercus petraea Liebl. et Quercusrobur L. sur la diversite de la flore fongique en place et sur quelquesaspects de son ecologie. J Sci Technol Tonnel 3:17–25 (1997).

11 Vivas N and Glories Y, Etude de la flore fongique du chene (Quercus sp.)caracteristique du sechage naturel des bois destines a la tonnellerie.Cryptogr Mycol 14:127–148 (1993).

12 Dhouib A, Hamza M, Zouari H, Mechichi T, H’midi R, Labat M, et al,Autochthonous fungal strains with high ligninolytic activities fromTunisian biotopes. Afr J Biotechnol 4:431–436 (2005).

13 Orth AB, Royse DJ and Tien M, Ubiquity of lignin-degradingperoxidases among various wood-degrading fungi. Appl EnvironMicrobiol 59:4017–4023 (1993).

14 Chavez R, Bull P and Eyzaguirre J, The xylanolytic enzyme system fromthe genus Penicillium. J Biotechnol 123:413–433 (2006).

15 Jourez B, Charron S and Quin GP, Proprietes des merrains affines dansune solution d’enzymes naturels et destines a la tonnellerie. AnnForest Sci 60:123–130 (2003).

16 Moredo N, Lorenzo M, Domınguez A, Moldes D, Cameselle C andSanroman A, Enhanced ligninolytic enzyme production anddegrading capability of Phanerochaete chrysosporium and Trametesversicolor. World J Microbiol Biotechnol 19:665–669 (2003).

17 Nilsson T, Studies on wood degradation and cellulolytic activity ofmicrofungi. Stud Forest Suecica 104:1–40 (1973).

18 Di Stefano R, Metodi chimici nella caratterizzazione varietale. Ann IstSper Enol 27:33–53 (1996).

19 Puech JL, Rabier P, Bories-Azeau J, Sarni F and Moutounet M,Determination of ellagitannins in extracts of oak wood and indistilled beverages matured in oak barrels. J Assoc Off Anal Chem73:498–501 (1990).

20 Towey JP and Waterhouse AL, The extraction of volatile compoundsfrom French and American oak barrels in Chardonnay during threesuccessive vintages. Am J Enol Vitic 47:163–172 (1996).

21 Suarez R, Suarez-Lepe JA, Morata A and Calderon F, The production ofethylphenols in wine by yeasts of the genera Brettanomyces andDekkera: a review. Food Chem 102:10–21 (2006).

22 Parodi G, L’affinamento del vino in legno. Vignevini 10:49–59 (1997).23 Spillman PJ, Sefton MA and Gawel R, The contribution of volatile

compounds derived during oak barrel maturation to the aroma ofa Chardonnay and Cabernet Sauvignon wine. Aust J Grape Wine Res10:227–235 (2004).

24 Ferreira V, Lopez R and Cacho JF, Quantitative determination of theodorants of young red wines from different grape varieties. J SciFood Agric 80:1659–1667 (2000).

25 Garde-Cerdan T, Rodrıguez-Mozaz A and Ancın-Azpilicueta C, Volatilecomposition of aged wine in used barrels of French oak and ofAmerican oak. Food Res Int 35:603–610 (2002).

26 Spillman PJ, Iland PG and Sefton MA, Accumulation of volatile oakcompounds in a model wine stored in American and Limousin oakbarrels. Aust J Grape Wine Res 4:67–73 (1998).

27 Boidro JN, Chatonnet P and Pons M, Influence du bois sur certainessubstances odorantes des vins. Conn Vigne Vin 22:275–294 (1988).

28 Lopez R, Aznar M, Cacho JF and Ferreira V, Determination of minor andtrace volatile compounds in wine by solid-phase extraction and gaschromatography with mass spectrometric detection. J ChromatogrA 966:167–177 (2002).

29 Sefton MA, How does oak barrel maturation contribute to wineflavour? Aust NZ Wine Ind J 6:17–20 (1991).

30 Peinado RA, Moreno J, Bueno JE, Moreno JA and Mauricio JC,Comparative study of aromatic compounds in two young whitewines subjected to pre-fermentative cryomaceration. Food Chem84:585–590 (2004).

31 Genovese A, Gambuti A, Piombino P and Moio L, Sensory propertiesand aroma compounds of sweet Fiano wine. Food Chem103:1228–1236 (2007).

32 Jackson RS, Wine Science: Principles and Applications (3rd edn).Academic Press, London (2008).

33 Gonzalez-Marco A, Jimenez-Moreno N and Ancın-Azpilicueta C,Concentration of volatile compounds in Chardonnay winefermented in stainless steel tanks and oak barrels. Food Chem108:213–219 (2008).

34 Bartowsky EJ and Henschke PA, The ‘buttery’ attribute ofwine – diacetyl – desirability, spoilage and beyond. Int J FoodMicrobiol 96:235–252 (2004).

35 Keller FA, Hamilton JE and Nguyen QA, Microbial pretreatment ofbiomass: potential for reducing severity of thermochemical biomasspretreatment. Appl Biochem Biotechnol 105/108:27–41 (2003).

36 Kheng PP and Omar IC, Xylanase production by a local fungal isolate,Aspergillus niger USM AI 1 via solid state fermentation usingpalm kernel cake (PKC) as substrate. Songklanakarin J Sci Technol27:325–336 (2005).

37 Lee JH, Kim JH, Zhu IH, Zhan XB, Lee JW, Shin DH, et al, Optimization ofconditions for the production of pullulan and high molecular weightpullulan by Aureobasidium pullulans. Biotechnol Lett 23:817–820(2001).

38 De Jong E, Chandra RP and Saddler JN, Effects of a fungal treatmenton the brightness and strength properties of a mechanical pulpfrom Douglas-fir. Bioresour Technol 61:61–68 (1997).


26

27

Research ArticleReceived: 31 March 2010 Revised: 20 June 2010 Accepted: 19 July 2010 Published online in Wiley Online Library: 17 August 2010


Temperature effects on type Ipepsin-solubilised collagen extractionfrom silver-line grunt skin and its in vitro fibrilself-assemblyNuntaporn Aukkanit and Wunwiboon Garnjanagoonchorn∗

Abstract

BACKGROUND: Fish skin is a potential source of collagen. Increasing the extraction temperature increases the yield of collagen.However, it may also result in degradation of the peptide chains, thus damaging the 3D structure of collagen that is vital forits application as a biomaterial. This work investigated the effects of extraction temperature on the yield and characteristics,including fibril self-assembly, of type I pepsin-solubilised fish skin collagen.

RESULTS: Pepsin-solubilised collagens were extracted from fresh skin of silver-line grunt at 4, 10, 20 and 28◦C for 6 h. Extractionat 10 ◦C gave the highest yield of collagens (439.32 ± 96.43 mg g−1 fresh skin, dry basis), which were identified as type I andcomprised β, α1 and α2 subunits. Extraction at higher temperatures (20 and 28 ◦C) resulted in the formation of low-molecular-weight peptide fragments, thus reducing the yield of the resultant type I collagen. The denaturation temperatures of collagensextracted at 4 and 10 ◦C, as determined by thermal analysis using differential scanning calorimetry, were 39.5 and 37.5 ◦Crespectively. In vitro fibril self-assembly of 1 mg mL−1 collagen solution (pH 6) incubated at 25 ◦C was only observed withcollagens extracted at 4 and 10 ◦C. The 10 ◦C collagen not only showed a higher rate of self-assembly, but its matrix also had alarger fibril diameter of 0.50 ± 0.07 µm (compared with 0.41 ± 0.07 µm for the 4 ◦C collagen) after 4 h of incubation.

CONCLUSION: The results indicated strong effects of extraction temperature on the yield and characteristics of the collagenobtained. Extraction of pepsin-solubilised collagen from silver-line grunt skin at 4–10 ◦C gave a high yield of type I collagenwith molecular integrity suitable for tissue-engineering applications.c© 2010 Society of Chemical Industry

Keywords: type I collagen; extraction; temperature; fibril self-assembly

INTRODUCTIONCollagen is widely used as a food and cosmetic ingredient andas a biomaterial. The collagen molecule must maintain its native3D structure to function properly when used as a biomaterial, e.g.to form films or porous sheets.1,2 It has also been shown thatfish skins from many species are good sources of collagen.3 – 6 Toobtain collagen from animal skin, the cleaned skin is hydrolysedwith an acid such as acetic acid7 or treated with an acid and thendigested with enzymes such as pepsin to remove telopeptides.8,9

After digestion, type I collagen is separated by salting out with0.9 mol L−1 NaCl in 0.5 mol L−1 acetic acid. Most extraction studieshave been conducted at low temperatures (usually 4 ◦C) to avoiddamaging collagen integrity. However, Lin and Liu10 reportedthe optimum temperature for collagen extraction from pepsin-treated bird feet as 12 ◦C, while the optimum temperature fordigestion with commercial porcine pepsin was specified as 37 ◦C.Moreover, collagen has been shown to denature at temperaturesclose to animal body temperature.11 For example, the acid-solublecollagen and pepsin-solubilised collagen extracted from the skinof silver-line grunt, a warm water fish, denature at 34 and 33.8 ◦Crespectively.8 It is therefore of interest to investigate the effects ofdifferent temperatures (4, 10, 20 and 28 ◦C) on pepsin-solubilised

collagen extraction from silver-line grunt skin, one of the majorby-products from the frozen fish industry, and to determinethe characteristics of the extracted collagen, including its fibrilself-assembly.

MATERIALS AND METHODSCollagen preparation from fish skinSilver-line grunt (Pomadasys kaakan) skins taken from fish ofaverage length 37 cm were obtained as frozen blocks from UnionFrozen Product Co., Ltd (Samutsakorn, Thailand). Fish skins werecleaned by scraping off scales and remaining meat and washedin distilled water at room temperature. The skins were thensuspended in 0.05 mol L−1 NaOH (pH 12) for 4 h at 4 ◦C to removenon-collagenous proteins. Finally, the treated skins were cut into

∗ Correspondence to: Wunwiboon Garnjanagoonchorn, Department of FoodScienceandTechnology,FacultyofAgro-Industry,KasetsartUniversity,Bangkok10900, Thailand. E-mail: [email protected]

Department of Food Science and Technology, Faculty of Agro-Industry,Kasetsart University, Bangkok 10900, Thailand


26

28

www.soci.org N Aukkanit, W Garnjanagoonchorn

2–3 cm pieces and collected as a pooled sample for collagenextraction. To extract collagen, 20 g portions of the collectedskins were blended with 200 mL of extracting solution (specifiedlater) at a medium speed for 3 min. The collagen extraction wascarried out at different temperatures (4, 10, 20 and 28 ◦C) with1 mg mL−1 pepsin (EC 3.4.23.1, activity 936 units mg−1 protein;Sigma Chemicals, St Louis, MO, USA) in 0.5 mol L−1 acetic acid (ata skin/acetic acid ratio of 1 g/10 mL) by continuous shaking for 6 hin a shaking water bath (BS-11, Jeio Tech Co., Ltd, Daejeon, Korea)fitted with a temperature controller. The extract was centrifuged at20 000×g for 30 min at 4 ◦C in a refrigerated centrifuge (Sorvall RC5C, Dupont, Newtown, CT, USA) to remove undigested materials.The type I pepsin-solubilised collagen was then precipitatedaccording to Noitup et al.5 by adjusting the concentration of NaClin the supernatant to 0.9 mol L−1 and allowed to stand for 30 minat 4 ◦C. The resultant precipitate was collected by centrifugationat 20 000 × g for 30 min, dissolved in four volumes of 0.5 mol L−1

acetic acid and dialysed (Cellu Sep dialysis membrane, molecularweight cut-off 12 000–14 000; Membrane Filtration Products Co.,Ltd, Seguin, Texas, USA) at 4 ◦C against distilled and deionisedwater until the medium maintained neutral pH. The dialysate wasfinally lyophilised to obtain the collagen sample. The yield ofcollagen (mg g−1 fresh skin, dry basis) was calculated as follows:

yield = [weight of lyophilised collagen (g)/

weight of fresh skin (g, dry basis)] × 1000

Viscosity of extracting solutionViscosity was determined with a Brookfield viscometer (RVT, Brook-field Engineering Laboratories, Inc., Middleboro, Massachusetts,USA) using spindle No. 5 at a speed of 5 rpm for the solutions ex-tracted at 4, 10 and 20 ◦C and a speed of 100 rpm for the solutionextracted at 28 ◦C. The viscosity was read hourly over a 6 h period.

Collagen subunitsCollagen subunit patterns were determined by sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) (modifiedmethod of Laemmli12) on 75 g L−1 separating gel with 40 g L−1

stacking gel, using Tris-glycine containing 1 g L−1 SDS (pH 8.3) aselectrophoresis buffer. Lyophilised samples were solubilised with50 g L−1 SDS solution (containing 50 g L−1 2-mercaptoethanol asreducing agent) to a concentration of 5 mg mL−1, then 5 µL sam-ple solutions were applied onto the gel. The molecular weights ofthe protein bands were determined by comparing their mobilitieswith those of high-molecular-weight markers ranging from 36to 205 kDa (Sigma Chemicals), namely rabbit myosin (205 kDa),Escherichia coli β-galactosidase (116 kDa), rabbit phosphorylase(97 kDa), rabbit fructose-6-phosphate kinase (84 kDa), bovine albu-min (66 kDa), bovine glutamic dehydrogenase (55 kDa), ovalbumin(45 kDa) and rabbit glyceraldehyde-3-phosphate dehydrogenase(36 kDa). After electrophoresis was completed, the protein bandswere stained with Coomassie brilliant blue R-250. The subunit pat-terns were compared with that of the standard acid-soluble type Icollagen from calf skin (ICN 9007-34-5, ICN Biomedical, Livermore,CA, USA). The proportions of protein bands on three gel slabs(from triplicate experiments) were determined by densitometry.

Thermal analysisCollagen samples were prepared by the modified method ofKomsa-Penkova et al.13 and Rochdi et al.,14 whereby lyophilised

collagens were dissolved in deionised water (1 : 40 w/v) andallowed to stand at 4 ◦C for 24 h. A 15 mg aliquot of samplesolution was sealed in a differential scanning calorimetry (DSC)cell and kept at 4 ◦C for 24 h. Thermal analysis was carried out byDSC (Peris 1, Perkin Elmer Inc., Waltham, MA, USA). Samples werescanned at 1 ◦C min−1 over the temperature range 20–50 ◦C. Thecharacteristic onset (To), peak (Tp) and recovery (Tr) temperatureswere recorded from each DSC curve.

Collagen self-assemblyIn vitro self-assembly was performed according to the method ofNoitup et al.5. Lyophilised collagen was dissolved in 0.5 mol L−1

acetic acid at 4 ◦C to give a concentration of 1 mg mL−1 anddialysed against 67 mmol L−1 phosphate buffer (pH 6). Aftercentrifugation at 20 000 × g for 30 min the supernatant wasincubated at 25 ◦C to reconstruct the matrix. The rate of collagenself-assembly was monitored every 15 min by measuring theturbidity changes as observed by the increase in absorbance at310 nm using a UV spectrophotometer (Spectro 22, Labomed Inc.,Culver city, CA, USA).

Structure of collagen matrix reconstructThe collagen solution was prepared according to the methoddescribed in the previous section. The supernatant was incubatedat 25 ◦C for 4 or 5 h to reconstruct the matrix. At the end of theincubation period the collagen matrix reconstruct was separatedby centrifugation at 5000 × g for 10 min and the structure wasresolved according to the method of Wong et al.15 The matrixwas stained with fluorescein isothiocyanate (FITC) (Invitrogen,Carlsbad, CA, USA) for 1 h and rinsed for 30 s with distilled waterto remove excess FITC. The sample was then covered with acoverglass and its structure was examined using a confocal laserscanning microscope (Carl Zeiss LSM 5 PASCAL, Axio Imager MI,Carl Zeiss Pte Ltd, Gottingen, Germany) and an He/Ne laser with anexcitation wavelength of 488 nm. Fibril diameters of the collagenmatrix were measured in two areas of 0.6 cm2 and the mean of100 fibrils was reported.

Statistical analysisData were subjected to analysis of variance and further analysedusing Duncan’s multiple range test to determine differencesbetween treatment means at a significance level of 95%. Unlessspecified otherwise, the mean of at least two replicates was usedin each study.

RESULTS AND DISCUSSIONViscosity of extracting solutionDifferent temperatures and times of pepsin digestion affect boththe yield and the properties of the collagen obtained. Temperaturehas been shown to affect the protein digestion rate of pepsin.10

At the early stage of skin collagen extraction by acid containing1 mg mL−1 pepsin the collagen fibres swelled, causing an increasein solution viscosity as shown in Fig. 1(a). Increasing extractiontime resulted in loss of skin integrity, and the solution appearedmore viscous. After 6 h the extraction at 10 ◦C gave the highestviscosity, followed by those at 4, 20 and 28 ◦C respectively. Athigh extraction temperature (28 ◦C) the viscosity of the extractingsolution decreased abruptly after 3 h owing to the rapid rate ofacid and pepsin digestion of collagen molecules (Fig. 1(b)). Theincreased viscosity of the extracting solution was related to the


26

29

Temperature effects on collagen extraction from silver-line grunt skin www.soci.org

Figure 1. Viscosity of fish skin collagen solutions extracted at different temperatures: (a) 4, 10 and 20 ◦C; (b) 28 ◦C.

Figure 2. Yield of lyophilised fish skin type I collagen at 4, 10, 20 and 28 ◦C after 6 h extraction time. Values with different letters are significantly different(P < 0.05). Bars indicate standard deviation of the means.

increased yield of pepsin-solubilised collagen, as can be seen inthe following section. However, the severe digestion conditionsat 28 ◦C resulted in the formation of low-molecular-weightfragments, which lowered the viscosity of the extracting solution(Fig. 1) and the yield of the extracted collagen as shown in Fig. 2.

Yield of collagenThe yield of type I collagen extract, isolated by precipitation with0.9 mol L−1 NaCl in 0.5 mol L−1 acetic acid, was calculated from

the final lyophilised product. Extraction temperature also had asignificant effect on the yield of lyophilised collagen obtained(P < 0.05) owing to the reduced activity of pepsin at lowertemperature. The extraction at 10 ◦C for 6 h gave the highest yieldof 439.32 ± 96.43 mg g−1 fresh skin (dry basis) (Fig. 2), leavingskin residues of 544.60 mg g−1 fresh skin (dry basis) that could gothrough the extraction process again. According to the viscositymeasurement at low temperature (Fig. 1), the yield of collagenincreased with the viscosity of the extracting solution. However,


26

30


Figure 3. SDS-PAGE patterns of fish skin collagen extracted at 4, 10, 20 and 28 ◦C on 75 g L−1 gel: M, standard marker proteins; C, calf skin collagen.

Table 1. Composition of collagen subunits on SDS-PAGE gel slab of fish skin collagen extracted at 4, 10, 20 and 28 ◦C and calf skin collagen

Fish skin collagen

Collagen subunita Calf skin collagen 4 ◦C 10 ◦C 20 ◦C 28 ◦C

>250 kDa 0.096 ± 0.015 0.045 ± 0.017 0.039 ± 0.016 – –

β 0.512 ± 0.031 0.566 ± 0.018 0.561 ± 0.020 0.132 ± 0.013 0.033 ± 0.010

α1 + α2 0.392 ± 0.045 0.389 ± 0.023 0.400 ± 0.034 0.254 ± 0.065 0.177 ± 0.055

<50 kDa – – – 0.614 ± 0.072 0.790 ± 0.059

Total 1.000 1.000 1.000 1.000 1.000

a Collagen subunits as shown in Fig. 3.

the low-molecular-weight peptides that resulted from the higherdigestion temperatures (20 and 28 ◦C) could be responsible forthe lower yield, since small peptides do not precipitate in the0.5 mol L−1 acetic acid solution containing 0.9 mol L−1 NaCl thatwas used to fractionate type I collagen.10

SDS-PAGE patternsThe SDS-PAGE patterns of fish skin collagen obtained byprecipitation with 0.9 mol L−1 NaCl in 0.5 mol L−1 acetic acid areshown in Fig. 3. The collagen samples, extracted by acid pepsindigestion at 4, 10, 20 and 28 ◦C for 6 h, all contained type I collagensimilar to that in calf skin. However, different proportions of eachsubunit were seen at different extraction temperatures (Table 1).The 4 and 10 ◦C collagens gave very similar results. The 20 and 28 ◦Cextractions produced collagen with much reduced proportions ofbothα andβ subunits but with a high proportion of low-molecular-weight peptides (<50 kDa) owing to the excessive digestion athigh temperatures. The temperature of the extraction processshould therefore be kept between 4 and 10 ◦C to ensure thatthe extraction product contains type I collagen with a minimumamount of low-molecular-weight peptide fragments.

Table 2. Thermal analysis of fish skin collagen extracted at 4 and10 ◦C and calf skin collagen

Denaturation temperaturea (◦C)

Sample To Tp Tr

Fish skin, 4 ◦C 38.5 ± 1.48b 39.5 ± 0.63b 40.1 ± 0.68b

Fish skin, 10 ◦C 35.9 ± 0.21c 37.5 ± 0.07c 39.6 ± 0.65b

Calf skinb 45.0 ± 0.03a 45.6 ± 0.00a 46.9 ± 0.02a

Values with different letters within a column are significantly different(P < 0.05).a To, onset temperature; Tp, peak temperature; Tr, recovery tempera-ture.b Calf skin acid-soluble collagen type I.

Thermal analysisThe thermal analysis of silver-line grunt skin collagen extractedat 4 and 10 ◦C and calf skin acid-soluble type I collagen isshown in Table 2. The denaturation temperature (Td) or peaktemperature (Tp) of pepsin-solubilised fish skin collagen extractedat 4 ◦C (Tp 39.5 ◦C) was significantly higher (P < 0.05) than that


26

31

Temperature effects on collagen extraction from silver-line grunt skin www.soci.org

Figure 4. Rate of self-assembly of fish skin collagen extracted at 4, 10, 20 and 28 ◦C.

of collagen extracted at 10 ◦C (Tp 37.5 ◦C), which could be due tothe reason suggested by Noitup et al.8 and Wong,16 namely thatcollagen that retains a higher degree of inter- and intramolecularcrosslinks is more heat-stable. In the case of calf skin collagenthe denaturation temperature reported was higher than thatof fish skin collagen owing to the higher imino acid content,which results in more stable helices.8,16 The higher Td of pepsin-solubilised collagen from silver-line grunt skin in this study ascompared with that reported by Noitup et al.8 was mainly due todifferences in the sample preparation methods used for thermalanalysis. Denaturation is thought to involve the disruption of inter-and intramolecular crosslinking of collagen molecules.17 In thepresent study, lyophilised collagens were dissolved in deionisedwater, whereas Noitup et al.8 used acetate buffer (pH 3.5) in whichcollagen solubility is higher. The characteristic temperatures To,Tp and Tr of the collagens extracted at 20 and 28 ◦C could notbe detected under the present experimental conditions used forthermal analysis.

Rate of collagen self-assemblyThe rate of collagen fibril reconstruction is influenced by severalexperimental variables, including the concentration of collagenand the pH and temperature of the solution.18 In this study,collagen self-assembly was performed according to Noitup et al.,5

who reported that 1 mg mL−1 collagen solution (pH 6) incubated at25 ◦C gave the highest rate of in vitro collagen self-assembly. Theprogressive aggregation of collagen molecules was monitoredby the increase in absorbance at 310 nm of the solution, andthe changes in absorbance at 310 nm against incubation timeindicated the rate of collagen matrix reconstruction. The resultsobtained (Fig. 4) show typical sigmoidal curves similar to thosereported by Engel19 and Veis and George.20 The fish skin collagensextracted at 20 and 28 ◦C did not show self-assembly, while thatextracted at 10 ◦C again exhibited the highest rate of collagenfibril reconstruction (Fig. 4). Similarly, Lin and Liu10 reported thatbird feet collagen extracted at 12 ◦C had a higher rate of self-assembly than those extracted at 4, 18 and 24 ◦C. The results frompeptide SDS-PAGE patterns indicated that the fish skin collagensextracted at 4 and 10 ◦C contained fewer low-molecular-weight

peptides with higher α1, α2 and β subunits when compared withthe samples treated at 20 and 28 ◦C. These data support previousreports1,2 that molecular integrity of the collagen molecule isnecessary for its application as a biomaterial for tissue-engineeringpurposes.

Structure of collagen matrix reconstructThe matrices reconstructed from fish skin type I collagens extractedat 4 and 10 ◦C with 4 and 5 h of assembly incubation time wereexamined under a confocal laser scanning microscope. Fibrilswere observed as green fluorescence against a dark background;however, all micrographs in Fig. 5 are formatted and shown ingrey scale. The resultant micrographs revealed that after 4 h ofincubation the diameter of the fibrils formed from acid/pepsin-digested collagen extracted at 10 ◦C (0.50 ± 0.07 µm, mean ±standard deviation) (Fig. 5(c)) was significantly larger (P < 0.05)than that of the fibrils formed from collagen extracted at 4 ◦C(0.41 ± 0.07 µm) (Fig. 5(a)). Incubation time also plays a role:a longer incubation time leads to an increase in fibril diameter(Figs 5(a) and 5(b)). Likewise, the absorbance at 310 nm is related tothe diameter of fibrils: the larger the fibril diameter, the higher is theabsorbance (Fig. 4). Different rates of collagen monomer assemblyhave also been shown to produce fibrils with different diameters.21

In this study we found that under the same assembly environmentthe fibril diameter increases as the collagen self-assembly rateincreases.

CONCLUSIONSDifferent temperatures and times of pepsin digestion affect boththe yield and the properties of the collagen obtained. Collagenextraction from fresh silver-line grunt skin with 0.5 mol L−1

acetic acid containing 1 mg mL−1 pepsin (at a skin/acid ratioof 1 g/10 mL) incubated at 10 ◦C for 6 h gave the highestyield (439.32 ± 96.43 mg g−1 fresh skin, dry basis). Extractionat higher temperatures (20 and 28 ◦C) gave lower yields ofcollagen, resulting from the formation of low-molecular-weightpeptide fragments not precipitated by 0.9 mol L−1 NaCl. The Td

of collagen extracted at 4 ◦C was higher than that of collagen


26

32


Figure 5. Confocal laser scanning micrographs of reconstructed fish skincollagen fibrils obtained using a 100× oil immersion objective lens: (a) fishskin collagen extracted at 4 ◦C incubated for 4 h; (b) fish skin collagenextracted at 4 ◦C incubated for 5 h; (c) fish skin collagen extracted at 10 ◦Cincubated for 4 h. Fibril diameters (mean± standard deviation) determinedfrom 100 fibrils are shown at the bottom of each micrograph; means withdifferent superscripts are significantly different (P < 0.05). The scale bar in(a) represents 5 µm.

extracted at 10 ◦C, but the collagen extracted at 10 ◦C had ahigher rate of collagen self-assembly with a larger diameterof reconstructed fibrils. In summary, the extraction of collagenfrom silver-line grunt skin by acid/pepsin digestion at 10 ◦C(or lower) produced collagen in high yield and with promisingmolecular integrity for its application in tissue-engineeringwork.

ACKNOWLEDGEMENTSThe authors thank The Graduate School, Kasetsart University forfinancial support and Union Frozen Product Co., Ltd, Samutsakorn,Thailand for the generous supply of fish skins.

REFERENCES1 Lee CH, Singla A and Lee Y, Biomedical applications of collagen. Int J

Pharmaceut 221:1–22 (2001).2 Ikada Y, Biological materials, in Integrated Biomaterials Science, ed.

by Barbucci R. Kluwer Academic/Plenum, New York, NY, pp. 1–7(2001).

3 Kimura S, Zhu X, Matsui R, Shijoh M and Takamizawa S,Characterization of fish muscle type I collagen, J Food Sci53:1315–1318 (1988).

4 Sato K, Yoshinaka R, Itoh Y and Sato M, Molecular species of collagenin the intra-muscular connective tissue of fish. Comp Biochem PhysiolB 92:87–92 (1989).

5 Noitup P, Morrissey MT and Garnjanagoonchorn W, In vitro self-assembly of silver line grunt type I collagen: effect of collagenconcentrations, pH and temperatures on collagen self-assembly.J Food Biochem 30:547–555 (2006).

6 Zhang JJ, Duan R, Tian Y and Konno K, Characterisation of acid-solublecollagen from skin of silver carp (Hypophthalmichthys molitrix). FoodChem 116:318–322 (2009).

7 Kittipattanabowon P, Benjakul S, Visessanguan W, Nagai T andTanaka M, Characterisation of acid-soluble collagen from skinand bone of bigeye snapper (Priacanthus tayenus). Food Chem89:363–372 (2005).

8 Noitup P, Garnjanagoonchorn W and Morrissey MT, Fish skin typeI collagen: characteristic comparison of albacore tuna (Thunnusalalunga) and silver-line grunt (Pomadasys kaaka). J Aquat FoodProd Technol 14:17–28 (2005).

9 Senaratne LS, Park PJ and Kim SK, Isolation and characterization ofcollagen from brown backed toadfish (Lagocephalus gloveri) skin.Bioresour Technol 97:191–197 (2006).

10 Lin YK and Liu DC, Effects of pepsin digestion at different temperaturesand times on properties of telopeptide-poor collagen from bird feet.Food Chem 94:621–625 (2006).

11 Bailey AJ and Light ND, Molecular and fiber structure of collagen, inConnective Tissue in Meat and Meat Products, ed. by Bailey AJ andLight ND. Elsevier Science, London, pp. 33–34 (1989).

12 Laemmli UK, Cleavage of structural proteins during assembly of thehead of bacteriophage T4. Nature 227:680–685 (1970).

13 Komsa-Penkova R, Koyonava R, Kostov G and Tenchov B, Discretereduction of type I collagen thermal stability upon oxidation.Biophys Chem 83:185–195 (1999).

14 Rochdi A, Foucat L and Renou J, NMR and DSC studies during thermaldenaturation of collagen. Food Chem 69:295–299 (2000).

15 Wong JW, Frank JF and Bailey S, Visualization of eggshell membranesand their interaction with Salmonella enteritidis using confocal laserscanning microscopy. J Food Protect 60:1022–1028 (1997).

16 Wong DWS, Proteins, in Mechanism and Theory in Food Chemistry, ed.by Wong DWS. Van Nostrand Reinhold, New York, NY, pp. 48–97(1989).

17 Purslow PP, The fracture properties and thermal analysis ofcollagenous tissues, in Advances in Meat Research, ed. byPearson AM, Dutson TR and Bailey AJ. AVI Books, New York, NY,pp. 187–208 (1987).

18 Hulmes DJS, The collagen superfamily – diverse structures andassemblies. Essays Biochem 27:49–67 (1992).

19 Engel J, Concepts of self-assembly in biological systems, in ExtracellularMatrix Assembly and Structure, ed. by Yurchenco PD, Birk DE andMecham RP. Academic Press, San Diego, CA, pp. 1–14 (1994).

20 Veis A and George A, Fundamentals of interstitial collagen self-assembly, in Extracellular Matrix Assembly and Structure, ed. byYurchenco PD, Birk DE and Mecham RP. Academic Press, San Diego,CA, pp. 15–45 (1994).

21 Miyahara M, Njieha FK and Prockop DJ, Formation of collagen fibrilsin vitro by cleavage of procollagen with procollagen proteinase.J Biol Chem 257:8442–8448 (1982).


26

33



Iron supply to soybean plants throughthe foliar application of IDHA/Fe3+: effectof plant nutritional status and adjuvantsPatricia Rodrıguez-Lucena, Edgar Ropero, Lourdes Hernandez-Apaolazaand Juan J Lucena∗

Abstract

BACKGROUND: Synthetic Fe chelates are commonly used to overcome Fe deficiencies in crops, but most of them are scarcelybiodegradable. Iminodisuccinic acid (IDHA) is a biodegradable chelating agent that is currently being evaluated as an alternativeto EDTA. In this work, the efficacy of the foliar application of IDHA/Fe3+ to soybean chlorotic plants under controlled conditionswas studied, testing the influence of the adjuvant used and of the plant nutritional status.

RESULTS: When IDHA/Fe3+ was applied to soybean plants with severe Fe chlorosis and the foliar sprays were the sole source ofFe, this chelate behaved similarly to the EDTA/Fe3+ and the recovery of the plants was slight in both cases. The same chelateswere tested when foliar sprays were an additional source of Fe for mildly chlorotic plants, which were also being supplied withlow concentrations of Fe applied to the nutrient solution. Then, plant recovery was appreciable in all cases, and the IDHA/Fe3+was as effective as EDTA/Fe3+. Among the adjuvants studied, a urea-based product was the only one that did not damage theleaf surface and that could improve the efficiency of IDHA/Fe3+ up tp the level of EDTA/Fe3+.

CONCLUSIONS: Thus, it was concluded the foliar application of IDHA/Fe3+ can be an environmentally friendly alternative to thenon-biodegradable chelate EDTA/Fe3+ when the appropriate adjuvant is used.c© 2010 Society of Chemical Industry

Keywords: biodegradable; chelate; chlorosis; foliar spray; IDHA; soybean

INTRODUCTIONIron chlorosis is a widespread agricultural problem, especiallyfor crops grown in calcareous soils, where calcium carbonatebuffers soil solution pH in the 7.5–8.5 range1 and there is a highbicarbonate concentration. The solubility of Fe in soil is controlledby Fe oxides2 and the most soluble Fe oxide limits total solubleFe concentration to around 10−10 mol L−1 in calcareous soils,much lower than that required (10−8 mol L−1) for optimal plantgrowth.3 Caused by a reduction in leaf photosynthetic pigmentconcentration, intervenial leaf yellowing is the most characteristicvisual symptom in chlorotic plants, whose fruit quality, size andyield are severely reduced.

Fe chelates applied to soils are the most efficient remedy tocontrol Fe chlorosis. Most of these chelates degrade very slowlyin the environment, and concern about the environmental riskof their application4 has grown in recent decades. Moreover, therisk of leaching out of the root zone in regimes of high wateravailability, which can be very high for the highly efficient chelateEDDHA/Fe3+,5 constitutes an important constraint when syntheticchelates are applied to soils.6 The biodegradable chelating agentN-(1,2-dicarboxyethyl)-D,L-aspartic acid, commonly known as imin-odisuccinic acid or IDHA (Fig. 1), has recently been proposed for itsuse in agriculture.7,8 IDHA shares structural similarities with EDTA(Fig. 1), but contains only five functional groups able to complexFe. Due to this, IDHA/Fe3+ records a lower stability than EDTA/Fe3+

and a higher reactivity in agronomic conditions. However, theefficiency of IDHA/Fe3+ has been comparable to EDTA/Fe3+ atproviding Fe through nutrient solution to cucumber and soybeanplants, cultivated in a growth chamber in hydroponics undercalcareous soil conditions.8 When applied to tomato or green beanplants grown under field conditions or in commercial hydroponics,IDHA/Fe3+ behaved similarly to EDTA/Fe3+ solving Fe chlorosis.9

As an alternative to a conventional supply of Fe to the soilor nutrient solution, the foliar application of Fe compounds isnow under consideration, but controversial results have beenobtained due to the numerous uncertainties involved in this type offertilisation. The aerial parts of plants are covered by a cuticle that isthe major barrier to be overcome when chemicals are sprayed ontoleaves. Cuticles have both hydrophilic and lipophilic properties,and the routes and factors controlling penetration of ionic speciesare still poorly understood. The use of adjuvants enhances theretention of foliar sprays and can increase cuticular penetration,diffusion into the apoplast and uptake by leaf cells.10 Nevertheless,

∗ Correspondence to: Juan J Lucena, Agricultural Chemistry Department,Autonoma University of Madrid, E-28049 Madrid, Spain.E-mail: [email protected]

Agricultural Chemistry Department, Universidad Autonoma de Madrid,Francisco Tomas y Valiente N◦ 7, 28049 Madrid, Spain


26

34

www.soci.org P Rodrıguez-Lucena et al.

HO

OH

O

OH

NH

HOO

HN

o,o-EDDHA

HOOC

H N

HOOC

COOH

COOH

IDHA

EDTA

HOOC

HOOC

COOH

NHNH

COOH

Figure 1. Fe chelates described in the text.

as many adjuvants can traverse the cuticle and damage cellsand cell membranes, those applied with foliar fertilisers shouldeither not penetrate or do so very slowly. Moreover, they shouldbe biodegradable and non-phytotoxic products and performefficiently when applied at low concentration.

The foliar application of IDHA/Fe3+ has been assayed previouslyunder field conditions with variable results. Although some studiesshowed that FeSO4 could re-green chorotic leaves more efficientlythan other Fe compounds when supplied through foliar sprays,11,12

the beneficial effect seems related to the low pH used in theapplication of Fe sulfate (pH 4) in order to avoid natural oxidation ofFe. In addition, the Fe supplied by this salt may not remain availablein leaves after a short period of time and leaf surface is damageddue to the low pH of the solution.13,14 On the other hand, recentstudies have corroborated that synthetic chelates behave moreefficiently that sulfate when the appropriate adjuvant is includedin the formulation.13,14 With regard to the type of chelate used,working with Fe-deficient peach trees Fernandez et al.13 observedthat EDTA/Fe3+ was more efficient than IDHA/Fe3+ when appliedwith an alkyl-glucoside adjuvant. In a different experiment testingthe influence of the type of adjuvant on the performance of thesechelates when applied to chlorotic peach trees, the same authorsconcluded that their effectiveness was highly dependent on thetype of adjuvant used. In general, EDTA/Fe3+ promoted higherleaf re-greening than IDHA/Fe3+, while leaf Fe concentration wasgreater in the trees sprayed with IDHA/Fe3+.14

Given the variability of the results obtained in field experiments,the aim of this work was to asses the ability of IDHA/Fe3+ foliarsprays to overcome Fe chlorosis under controlled conditions whenapplied to soybean (Glycine max. cv Stine 0480), a non-efficientmodel plant for Fe nutrition. First, the effectiveness of the foliarapplication of IDHA/Fe3+ to soybean plants with severe chlorosis

was studied when the foliar sprays were the sole source usedto provide Fe to plants, and different adjuvants (non-ionic orglycine-based) were applied in combination with the chelate. In asecond experiment, the efficiency of the IDHA/Fe3+ applied withurea or with a non-ionic adjuvant to provide foliar Fe to plants wastested when applied to mildly chlorotic plants that were also beingsupplied with Fe at low concentrations (as EDTA/Fe3+) through thenutrient solution to simulate the low availability of Fe in calcareoussoils. For comparative purposes, EDTA/Fe3+, frequently used infoliar applications in the field, was evaluated in both experiments.

EXPERIMENTALFe-containing compoundsIDHA/Fe3+ (8.7% (w/w) Fe as IDHA/Fe3+) was provided by ADOBPPC (Poznan, Poland) and prepared by dissolution of the solidand filtration, while EDTA/Fe3+ was prepared in the laboratory;to do so, the ligand (Na2EDTA; Merck, Darmstadt, Germany) wasfirst dissolved and then an amount of Fe(NO3)3·9H2O (Merck),calculated to be 5% in excess of the molar amount of ligand,was added slowly. During chelation, the pH was maintainedbetween 6.0 and 8.0 and adjusted to 5.5 at the end. The solutionwas left to stand overnight to allow any excess element toprecipitate as oxides. Finally, the solution was filtered througha 0.45 µm Millipore membrane and made up to volume withwater. Exposure of all the chelate solutions to light was avoidedduring their preparation and storage because of the potentialphotodecomposition of chelates.15 Both chelates were preparedat a concentration of Fe 5 mmol L−1.

AdjuvantsDifferent adjuvants were tested: two non-ionic products withethylene oxide groups as hydrophilic component (an alkylpolyglu-coside (APG) and polyoxyethylene (20) sorbitan monooleate(PS)), an anionic glycine-based solution (GLY) and a urea-basedadjuvant (U).

Plant material, experimental design and treatmentsSoybean seeds (Glycine max L. cv Stine 0480, kindly providedby Professor R. Goos, North Dakota State University, Fargo, USA)were germinated using a standard seed-growing procedure insterilised trays. The seeds were washed with water for 30 minand then placed in trays between two sheets of cellulosepaper moistened with distilled water. The trays were kept indarkness at 28 ◦C for 3 days in a thermostatted oven. Aftergermination, the seedlings were transferred to a growth chamber,where they were grown under controlled climatic conditions:day/night photoperiod, 16/8 h; temperature (day/night) 30/25 ◦C;relative humidity (RH) (day/night) 50/70%. Seedlings of similardevelopment were placed on a holed plate, floating in containerswith a continuously aerated 1/5 diluted nutrient solution for6 days. The diluted nutrient solution was then replaced by a full-strength solution with the following composition: (macronutrientsin mmol L−1) 1.0 Ca(NO3)2, 0.9 KNO3, 0.3 MgSO4, 0.1 KH2PO4;(cationic micronutrients in µmol L−1, as buffered micronutrientsolution) 2.5 MnSO4, 1.0 CuSO4, 10.0 ZnSO4, 1.0 NiCl2, 1.0 CoCl2,5.0 Fe(NO3)3, 120.5 Na2EDTA, 50.0 KOH; (anionic micronutrientsin µmol L−1) 35.0 NaCl, 10.0 H3BO3, 0.05 Na2MoO4. The pH wasbuffered with HEPES 1.0 × 10−4 mol L−1 and adjusted to 7.5 withKOH 1.0 mol L−1. The seedlings were kept in this solution for6 days, until chlorotic symptoms were observed. After this time,


26

35

Foliar application of IDHA/Fe3+ to soybean www.soci.org

Table 1. Treatments, adjuvants and doses tested in experiment I

Foliar

Fetreatments Treatment Adjuvant

Nutrientsolution

FeIDHA +APG

5 mmol L−1

IDHA/Fe3+Oligomeric

alkylpolygluco-side (APG) 0.1%(v/v)

No added Fe

FeIDHA + GLY Glycine based(GLY) 0.3% (v/v)

FeIDHA + PS Polyoxyethylene(20) sorbitanmonooleate(PS) 0.1% (v/v)

FeEDTA +APG

5 mmol L−1

EDTA/Fe3+Oligomeric


Control

Control – Fe No added Fe No added Fe

Table 2. Treatments, adjuvants and doses tested in Experiment II

Foliar

Fetreatments Treatment Adjuvant

Nutrientsolution

FeIDHA +APG I

5 mmol L−1

IDHA/Fe3+Oligomeric


No added Fe

FeIDHA +APG II

5 µmol L−1

EDTA/Fe3+

FeIDHA + U Urea (U) 0.15%(v/v)

FeEDTA +APG

5 mmol L−1

EDTA/Fe3+Oligomeric


Control

Control – Fe No added Fe No added Fe

ControlFeEDTA

No added Fe 5 µmol L−1

EDTA/Fe3+

the stems of two plants were wrapped together with foam, andplaced in 2 L polyethylene vessels (three holes in the lid, sixplants per pot) containing 2 L of a full strength nutrient solutionwith the same composition as the initial one, except in thecontent of micronutrients (unbuffered micronutrient solution, inµmol L−1): 1.0 MnSO4, 0.5 M CuSO4, 0.5 ZnSO4, 0.1 NiCl2, 0.1 CoCl2.In Experiment I, Fe was not added to this nutrient solution, whilein Experiment II, 5.0 µmol L−1 EDTA/Fe3+ was added to avoidsevere chlorosis. The pH was adjusted to 7.5 with KOH 1.0 mol L−1

and buffered with HEPES 1.0 × 10−4 mol L−1, and 0.4 g−1 of solidCaCO3 per pot. Water was added every 2 days, and the solutionwas renewed every week. Three replicates were prepared for eachtreatment and for each control.

At this point, Fe treatments were applied (Table 1 for ExperimentI and Table 2 for Experiment II), and applications were repeatedafter 8 and 15 days.

Experiment I: Foliar application of IDHA/Fe3+ and EDTA/Fe3+ tosoybean plants affected by severe chlorosis without an additionalsource of FeEach pair of plants was sprayed with 2 mL of the 5 mmol L−1

IDHA/Fe3+ or EDTA/Fe3+ solutions at pH 5.5 (Table 1). Leaf sprayswere applied both on the adaxial and the abaxial surface with anebuliser system. The corresponding adjuvant (Table 1) was addedto each solution just before leaf spraying at the rate indicated inTable 1. A control was included in which no Fe was supplied(control –Fe).

Experiment II: Foliar application of IDHA/Fe3+ and EDTA/Fe3+ tomildly chlorotic soybean plants, as a complement to Fe supplythrough the rootsApplication was performed as described in Experiment I, but onlythe first four levels of leaves were treated in order to evaluatethe redistribution of the Fe applied through foliar sprays. Thecorresponding adjuvant was added to each solution just beforeleaf spraying (Table 2), with the rate varying depending on thesurface active agent. Two control treatments (control –Fe andcontrol FeEDTA) were established. For control –Fe, no Fe wasadded to the nutrient solution or applied through foliar sprays. Inthe case of control FeEDTA (used to evaluate the influence on plantdevelopment of the Fe supplied through the nutrient solution)5.0 µmol L−1 EDTA/Fe3+ was applied to the nutrient solution. Forcontrol treatments, foliar applications were not performed.

MeasurementsDuring the experiments, regular Soil and Plant Analyzer Devel-opment (SPAD) readings were taken with a chlorophyll meter(Minolta SPAD-502; Minolta, Osaka, Japan) for all the leaf stages(average of three readings per leaf). Whole plants were sampled8 days (two pairs of plants) and 22 days (one pair of plants) af-ter the first application of the treatment. Sampled roots, stems,treated leaves and untreated leaves (only in Experiment II) wereseparated, weighed, washed with 0.1% HCl and 0.01% non-ionicdetergent solution, and rinsed twice with ultrapure water. Thesamples were then dried in a forced air oven at 65 ◦C for 3 days.Micronutrients were determined in the leaves and roots after drydigestion procedure by Atomic absorption spectroscopy (AAS,PerkinElmer AAnalyst 800; Perkin Elmer, Waltham, MA, USA).

Statistical analysisData were statistically evaluated using analysis of variance(ANOVA) with the program SPSS 15.0 to asses the significanceof the main factors and interactions. Means were also comparedusing Duncan’s test at P ≤ 0.05 in order to find significantdifferences between treatments.

RESULTSExperiment I: foliar application of IDHA/Fe3+ and EDTA/Fe3+to soybean plants affected by severe chlorosis without anadditional source of FeMicronutrient contents and biometric dataIn this experiment, the foliar sprays described in Table 1 werethe sole source of Fe for plants, which showed severe chlorosissymptoms. The effect of the foliar application of the treatmentson plant dry weight, root Fe and Mn concentrations, and Fe/Mnand Fe/(Mn + Zn + Cu) ratios was studied using one-wayANOVA. Statistical differences were recorded only for root Fe


26

36


Table 3. SPAD values, dry weight, root Fe and Mn concentrations, and root Fe/Mn and Fe/(Mn + Zn + Cu) ratios in soybean plants sprayed with theFe chelates tested, at the end of Experiment I

SPADRoot DW (g

plant−1 DW)[Fe] (µg g−1

DW)[Mn] (µgg−1 DW) Fe/Mn

Fe/(Mn + Zn +Cu)

FeIDHA + APG 11.3 ± 0.7a 0.51 ± 0.05ns 61.1 ± 1.4ab 166 ± 21b 0.38 ± 0.04a 0.20 ± 0.02a

FeIDHA + GLY 8.9 ± 0.6ab 0.57 ± 0.02 74.2 ± 7.9a 186 ± 28ab 0.36 ± 0.07a 0.20 ± 0.02a

FeIDHA + PS 9.6 ± 0.7ab 0.62 ± 0.13 50.6 ± 10.7b 187 ± 7ab 0.28 ± 0.07ab 0.15 ± 0.04ab

FeEDTA + APG 11.6 ± 1.3a 0.69 ± 0.12 50.2 ± 0.5b 236 ± 21ab 0.21 ± 0.01ab 0.13 ± 0.01ab

Control – Fe 7.3 ± 1.1b 0.43 ± 0.12 41.5 ± 3.3b 306 ± 73a 0.15 ± 0.03b 0.09 ± 0.01b

Data are means ± standard error (SE) of three independent replicates, except for SPAD values (average of 18 measurements). Different letters in thesame column denote significant differences between the treatments (P ≤ 0.05).DW: Dry weight basis; ns, not significant.

concentration (at P ≤ 0.05), indicating that the combination ofFeIDHA + GLY was more effective than the other Fe-containingcompounds and adjuvants tested.

In order to evaluate the redistribution of the foliarly appliedFe from leaves to other plant organs, Fe and other micronutrientconcentrations in roots were measured. Fe and Mn concentrations,as well as the Fe/Mn and Fe/(Mn + Zn + Cu) ratios in the roots,are presented in Table 3. The same measurements were also per-formed on treated leaves, but as it was not possible to ensure thatthe fraction of Fe applied through foliar sprays and not absorbedby leaves was completely removed after washing, these data werenot taken into consideration. The highest root Fe concentrationscorresponded to FeIDHA + GLY, which was the only treatmentstatistically different to control –Fe, and to FeIDHA + APG. Mnconcentrations also indicate that the uptake of Mn from thenutrient solution was lower in plants treated with Fe sprays, eventhough most of these treatments were not statistically differentto control –Fe. The low Fe/Mn and Fe/(Mn + Cu + Zn) ratios forcontrol –Fe with regard to treated plants indicate the increasedcapacity of roots to take up other cations under Fe deficiencyconditions. This favoured uptake is related to the involvement ofthe same transporter implied in the uptake of Fe by roots, the Ironregulated transporter 1 (IRT1), in the absorption of other cationssuch as Mn or Zn.16 In spite of the differences in micronutrientconcentrations and ratios, none of the treatments presentedstatistical differences in plant dry weight with control –Fe.

SPAD measurementsSPAD was measured periodically in all leaf levels. Two-way ANOVAanalysis showed that the SPAD values measured for each treatmentin the third level of leaves were strongly influenced (at P ≤ 0.001)by the treatment applied and by the age of the plants (dayssince the first application of the treatments), but there was nointeraction between the two factors. Average SPAD values arereported in Table 3, and indicate that for all the treatments theaverage SPAD was higher than for control –Fe. The best resultscorresponded to plants treated with Fe compounds combinedwith APG as adjuvant (FeIDHA + APG and FeEDTA + APG), whilethe other treatments were not statistically different to control –Fe.

Experiment II: foliar application of IDHA/Fe3+ and EDTA/Fe3+to mildly chlorotic soybean plants, as a complement to Fesupply through rootsIn the second experiment, foliar sprays were a complement toFe nutrition through the roots (Table 2), as usually occurs whenFe foliar sprays are applied under field conditions. Consequently,

plants exhibited mild chlorosis. In order to test the redistributionfrom treated leaves to untreated leaves only the first four levelsof leaves were sprayed. Regarding the adjuvants, a urea-basedsolution (U) was compared with the APG used in ExperimentI when applied in combination with IDHA/Fe3+ through foliarsprays.

Micronutrient contents and biometric dataThe influence of the treatments on plant dry weight, Fe andMn concentrations, and Fe/Mn and Fe/(Mn + Zn + Cu) ratios inuntreated leaves and roots was tested using one-way ANOVA. Thetreatments strongly affected Fe and Mn concentrations and theFe/Mn and Fe/(Mn + Cu + Zn) ratios in roots (at P ≤ 0.001 in allthe cases).

The data corresponding to dry weight, Fe and Mn concentrationsand the Fe/Mn and Fe/(Mn + Zn + Cu) ratios in untreated leavesand roots at the end of the experiment are reported in Table 4. Inuntreated leaves, no statistical differences due to the treatmentswere observed for any of the parameters studied. However, whenmicronutrient concentrations and ratios were analyzed in roots,IDHA treatments including the supply of Fe to the roots (FeIDHA+ APG II and FeIDHA + U) were the most effective and presentedstatistical differences with control –Fe. Nevertheless, in most casesthese treatments were not statistically different to control FeEDTA(where foliar applications were not performed). The only treatmentin which Fe was not added to the nutrient solution (FeIDHA + APGI) recorded no statistical differences with control –Fe. With regardto root dry weight, the highest values were found for FeIDHA + Uand FeEDTA + APG, while no statistical differences were observedin root dry weight between FeIDHA + APG I (with no Fe supplythrough the nutrient solution) and control –Fe.

SPAD measurementsSPAD was measured periodically in all leaf levels. Two-way ANOVAanalysis showed that SPAD values for each treatment in thethird level of leaves were strongly influenced (at P ≤ 0.001)by the treatment applied and by the age of the plants (dayssince the first application of the treatments), but there was nointeraction between the two factors. Average SPAD values arereported in Table 4, and indicate that all the treatments exceptFeIDHA + APG I (with no Fe added to the nutrient solution) werestatistically different to control –Fe. On the other hand, very lowSPAD measurements were recorded when no foliar sprays wereperformed (control FeEDTA). The best treatments for improvingSPAD values were FeIDHA + U and FeEDTA + APG.


26

37


Tab

le4

.SP

AD

valu

es,u

ntr

eate

dle

aves

and

roo

tdry

wei

gh

t,u

ntr

eate

dle

aves

and

roo

tFe

and

Mn

con

cen

trat

ion

s,an

du

ntr

eate

dle

aves

and

roo

tsFe

/Mn

and

Fe/(

Mn

+Z

n+

Cu

)rat

ios

inso

ybea

np

lan

tssp

raye

dw

ith

the

Fech

elat

este

sted

,at

the

end

ofE

xper

imen

tII

Bio

mas

s(g

pla

nt−1

DW

)Fe

con

cen

trat

ion

(µg

g−1

DW

)M

nco

nce

ntr

atio

n(µ

gg

−1D

W)

Fe/M

nFe

/(M

n+

Cu

+Z

n)

SPA

DU

ntr

eate

dle

afRo

ot

Un

trea

ted

leaf

Roo

tU

ntr

eate

dle

afRo

ot

Un

trea

ted

leaf

Roo

tU

ntr

eate

dle

afRo

ot

FeID

HA

+A

PGI

4.4

±0.

7d0.

56±

0.10

ab47

.9±

1.9c

215.

9±

36.6

a0.

2±

0.1c

0.20

±0.

02b

FeID

HA

+A

PGII

10.1

±1.

4bc

1.20

±0.

11ab

0.93

±0.

24ab

46.5

±0.

5ns

135.

3±

21.7

a70

.8±

1.8n

s23

.3±

2.1b

0.7

±0.

01n

s5.

9±

1.1b

0.4

±0.

1ns

2.3

±0.

3a

FeID

HA

+U

14.8

±1.

4a1.

50±

0.37

a1.

64±

0.47

a45

.1±

8.9

105.

3±

9.1ab

63.8

±12

.112

.5±

1.9b

0.8

±0.

38.

7±

0.9a

0.5

±0.

22.

3±

0.1a

FeED

TA+

APG

14.1

±1.

6ab0.

49±

0.15

b1.

55±

0.59

ab40

.7±

3.3

89.4

±10

.4b

70.4

±20

.68.

6±

0.2b

0.7

±0.

210

.4±

1.2a

0.4

±0.

12.

0±

0.1a

Co

ntr

ol

–Fe

3.2

±0.

4d–

0.46

±0.

04b

–33

.1±

1.8c

–19

4.8

±26

.9a

–0.

2±

0.1c

–0.

10±

0.04

b

Co

ntr

olF

eED

TA6.

5±

0.7cd

0.58

±0.

10ab

1.07

±0.

22ab

31.4

±4.

214

0.4

±17

.9a

97.6

±25

.515

.8±

0.4b

0.4

±0.

18.

9±

1.1a

0.2

±0.

12.

5±

0.2a

Dat

aar

em

ean

s±

stan

dar

der

ror

(SE)

oft

hre

ein

dep

end

entr

eplic

ates

,exc

eptf

or

SPA

Dva

lues

(ave

rag

eo

f24

mea

sure

men

ts).

Diff

eren

tle

tter

sin

the

sam

eco

lum

nd

eno

tesi

gn

ifica

nt

diff

eren

ces

bet

wee

nth

etr

eatm

ents

(P<

0.05

,n=

3).

DW

:Dry

wei

gh

tb

asis

.Fe

IDH

A+

APG

Idid

no

tdev

elo

pu

ntr

eate

dle

aves

.


26

38


DISCUSSIONExperiment I: foliar application of IDHA/Fe3+ and EDTA/Fe3+to soybean plants affected by severe chlorosis without anadditional source of FeComparison of Fe compoundsIDHA is a biodegradable chelating agent recently proposed foragricultural use.7,8 Consequently, there are currently few studiesavailable that evaluate the effectiveness of IDHA/Fe3+ to overcomeFe chlorosis. The ability of this chelate to solve Fe deficiencieswhen applied to roots under controlled8,9 or field conditions9

has been tested successfully, but due to its relatively low stabilityits application through foliar sprays is also under consideration.Accordingly, several field trials to assess its performance have beencarried out in recent years,13 although to our knowledge the foliarapplication of IDHA/Fe3+ chelates under controlled conditions tocorrect Fe deficiencies in plants has yet to be studied. One ofthe objectives of this work was to evaluate the behaviour of thischelate under controlled conditions. All possible sources of Fewould therefore be considered in the experimental design andany improvement in plant Fe nutrition could be attributed to theFe supplied through the treatments.

In the studies assessing the effectiveness of IDHA/Fe3+ foliarsprays when applied in the field to peach trees,13,14 IDHA/Fe3+

performed as efficiently as EDTA/Fe3+ when the appropriateadjuvant was employed, although the results were always veryvariable. Foliar sprays are not the only source of Fe, since in soilexperiments a fraction of Fe is available for plants, even undercalcareous soil conditions. Thus, the results will be influenced bysoil conditions and Fe availability in it.

In the first experiment described in this work, foliar sprays werethe sole source of Fe for plants. In general, IDHA/Fe3+ behaved asefficiently as EDTA/Fe3+. Root Fe concentration and the other nu-tritional parameters reported in Table 3 suggest that Fe was redis-tributed from leaves to roots. Evidence of translocation from leaveshas been observed in other experiments working with cucumberand tomato and using labelled Fe.17 For our results, differences inFe and Mn concentrations and ratios in roots with regard to con-trol –Fe were observed, even though these differences were notsignificant in all cases. This suggests that the foliarly applied Fe wasredistributed, although at low rates due to the limited absorptionof Fe through leaf surfaces18 and to the scarce mobility of this mi-cronutrient through the phloem.19 Data corresponding to Fe/Mnand Fe/(Mn + Zn + Cu) ratios indicate that the affinity of roots forMn, Cu and Zn decreased when Fe sprays were performed, sup-porting the idea of Fe redistribution to roots. Taken together, thesedata would to some extent reflect the relative normalisation of themetabolic processes of the plants, as already observed in previousworks with IDHA/Fe3+ applied through foliar sprays.13 Moreover,these results indicate that the foliar application of biodegradableIDHA/Fe3+ is a good alternative to EDTA/Fe3+ when applied un-der appropriate conditions, as expected due to the similaritiesbetween both chelates. The better performance of EDTA/Fe3+with regard to other chelates when applied through foliar sprayshas been attributed, at least partially, to the photoreduction of thischelate20 that would favour Fe uptake by leaf cells. However, ourresults suggest that this phenomenon cannot explain the differentbehaviour observed for EDTA/Fe3+ and IDHA/Fe3+ in previous as-says, which can be attributed to different experimental conditions.The biodegradation of IDHA, which may be favoured more in cropsgrown in the field than under the controlled conditions of a growthchamber, could explain the lower effectiveness of IDHA/Fe3+ withregard to EDTA/Fe3+ when used under field conditions.

Comparison of adjuvant solutionsThree biodegradable adjuvants were tested in this first experi-ment: (1) two non-ionic products (polyoxyethylene (20) sorbitanmonooleate (PS) and an oligomeric alkylpolyglucoside (APG) prod-uct); and (2) an anionic glycine-based (GLY) solution (Table 1).Non-ionic adjuvants are the most commonly used in foliar appli-cations, since interaction with the active ingredient is minimised,albeit not completely avoided,18 in this type of compounds. Atthe same time, they increase the cuticular penetration of the ac-tive ingredient through complex interactions between the activeingredient, the adjuvants and leaf.21 PS adjuvants have not beencommonly applied in foliar formulations. In the case of APG, theseadjuvants favour cuticular penetration of solutions in general22

and Fe solutions in particular.13 For anionic adjuvants, it is impor-tant to bear in mind that these molecules may interact with thecations of the Fe compound solutions, forming large moleculesthat may block cuticular pores and interfere with the process of leafFe penetration.23 However, when applied through foliar sprays,GLY could penetrate leaf surfaces and translocate to other plantorgans,24 so solutions based on this amino acid may improve Fepenetration.

Previous studies13,14 have tested the foliar application ofIDHA/Fe3+ and EDTA/Fe3+ in combination with an APG and aGLY-based adjuvant, concluding that both chelates could increasegreen surface and chlorophyll content in peach leaves. However,as far as we know, the foliar application of solutions containingonly Fe compounds and GLY-based or PS adjuvants has notbeen assayed previously. Our results indicate that the applicationof IDHA/Fe3+ in combination with GLY (FeIDHA + GLY) or PSadjuvants (FeIDHA + PS) was slightly less efficient than the useof an APG adjuvant (FeIDHA + APG), in spite of the interactionbetween IDHA/Fe3+ and APG adjuvants described elsewhere.13,14

Regardless of the Fe compound or adjuvant used, leaf burnwas always observed in treated leaves, even when the non-phytotoxic APG surface active agent was applied. This observationis consistent with the results obtained by Fernandez et al.,18 whodetected the ionisation of a non-ionic adjuvant in the presence ofEDTA. However, these symptoms were similar to the necrotic spotsobserved for control –Fe, so it was not possible to confirm whetherthey were due to the foliar sprays (the damage observed on leavesmight be related to the Fe compound applied, to the dissolutionof the cuticle as a negative consequence of the application ofadjuvants, to a detrimental interaction between the adjuvant andthe Fe compound) or to the severe chlorosis of the plants.

Experiment II: foliar application of IDHA/Fe3+ and EDTA/Fe3+to mildly chlorotic soybean plants, as a complement to Fesupply through rootsComparison of Fe compoundsThe results of Experiment I indicated that foliarly applied Fe canre-green leaves and be distributed to other plant organs, althoughdue to the limited absorption of Fe through leaf surfaces18 and tothe low mobility of Fe in the phloem19,25 this type of applicationcannot fully satisfy plant Fe demand. Since Fe fertilisation throughfoliar sprays can be used only as a complementary techniqueto the direct application into soil or to a nutrient solution of Fechelates,11,26 in the second experiment presented in this workfoliar sprays were applied as a complement to the application ofFe chelates to the nutrient solution of soybean plants grown inhydroponics under calcareous soil conditions.

The results of this assay confirmed that even though Fe foliarsprays cannot fully substitute the conventional supply of Fe


26

39


chelates to the roots, they help to overcome Fe chlorosis whenapplied as a complementary source of Fe. Under these conditions,the re-greening of leaves and the redistribution of Fe from treatedleaves to other organs occurred. The foliar application of Fechelates when EDTA/Fe3+ was added to the nutrient solution(FeIDHA + APG II, FeIDHA + U and FeEDTA + APG) increasedroot Fe concentration. This suggests that Fe was translocated tothe roots, as already observed when labeled Fe was used,17,27,28

and/or that Fe uptake from the nutrient solution was enhanceddue to the transmission of signals from the shoots29 that activatethe root Fe uptake mechanism. None of the treatments tested wasclearly more efficient than the rest at favouring Fe redistributionto untreated organs, but the highest SPAD values correspondedto FeIDHA + U.

Comparison of adjuvant solutionsAlthough the APG adjuvant was the most efficient product usedin combination with IDHA/Fe3+, the interaction between bothcompounds has been described in previous studies. Thus, theeffectiveness of the APG was compared with urea when added inIDHA/Fe3+ formulations.

The use of urea in foliar sprays favours Fe assimilation byleaves,30 given both its capacity to increase the permeability ofleaf membranes and its surfactant properties. At the same time,these applications can play an additive role. It has already beendemonstrated that under Fe deficiency conditions the capacity ofroots to acquire nitrate is limited and this anion accumulates atlow rates in leaf tissues.31 Thus, the simultaneous application ofFe and N (as urea) through foliar sprays may have deal with thisdeficiency and favoured leaf re-greening. This beneficial effect ofthe application would explain the high SPAD values for the FeIDHA+ U treatment. In any case, urea must be supplied at low rates toavoid the inhibition of photosynthesis.32

Regarding burn signs on leaf surfaces, they were less abundantin the mildly chlorotic plants used in this experiment. However,except for FeIDHA + U, treated leaves always presented dark spotsthat could be associated with leaf damage. This confirms that theuse of APG negatively affected the leaf surface even though thistype of adjuvant has been described as non-phytotoxic. Negativeinteractions between spray components and leaf surfaces did notseem to occur for urea.

CONCLUSIONSIDHA/Fe3+ performed similarly to EDTA/Fe3+ and revealed a po-tential ability to provide Fe to soybean when applied throughfoliar sprays under controlled conditions and with the appropri-ate adjuvant. Moreover, the application of IDHA/Fe3+ was moreeffective when urea was also used. Thus, foliar application withIDHA/Fe3+ and urea formulations, both degradable compounds,may constitute an environmentally friendly alternative to the tradi-tional use of EDTA/Fe3+ in foliar Fe fertiliser formulations, althoughfurther research is recommended to optimise formulations andapplication procedures.

ACKNOWLEDGEMENTSThis work was partly supported by ADOB PPC, and by theSpanish Ministry of Science and Education Project AGL2007-63756, co-financed with ERDF and the European Commission.P. Rodrıguez-Lucena was on a Spanish Ministry of Science and

Education ‘FPI’ pre-doctoral contract co-financed by the EuropeanSocial Fund.

REFERENCES1 Lindsay WL and Schwab AP, The chemistry of iron soils and its

availability to plants. J Plant Nutr 5:821–840 (1982).2 Lindsay WL, Iron oxide solubilization by organic matter and its effect

on iron availability, in Iron Nutrition and Interaction in Plants, ed.by Chen Y and Hadar Y. Kluwer Academic Publishers, Dordrecht,pp. 29–36 (1991).

3 Romheld V and Marschner H, Mobilization of iron in the rhizosphereof different plant species, in Advances in Plant Nutrition, vol. 2, ed.by Tinker B and Lauchli A. Praeger, New York, pp. 155–204 (1986).

4 Hyvonen H, Orama M, Saarinen H and Aksela R, Studies onbiodegradable chelating agents: complexation of iminodisuccinicacid (ISA) with Cu(II), Zn(II), Mn(II) and Fe(III) ions in aqueous solution.Green Chem 5:410–414 (2003).

5 Cesco S, Romheld V, Varanini Z and Pinton R, Solubilization of ironby water-extractable humic-substances. J Plant Nutr 163:285–290(2000).

6 Rombola AD and Tagliavini M, Iron nutrition of fruit tree crops, in IronNutrition in Plants and Rhizospheric Microorganisms, ed. by Barton LLand Abadıa J. Springer-Verlag Academic Publishers, Dordrecht,pp. 61–83 (2006).

7 Mitschker A, Moritz RJ and Nawrocki A, Chelated plant micro-nutrients. Bayer Chemicals A.-G., Germany; PrzedsiebiorstwoProdukcyjno Consultingowe Adob. European Patent EP1411037A120040421Appl (2004).

8 Villen M, Garcıa-Arsuaga A and Lucena JJ, Potential useof biodegradable chelate N-(1,2-dicarboxyethyl)-D,L-asparticacid/Fe3+ as an Fe fertilizer. J Agric Food Chem 55:402–407 (2007).

9 Lucena JJ, Sentıs JA, Villen M, Lao T and Perez-Saez M, IDHA chelatesas a micronutrient source for green bean and tomato in fertigationand hydroponics. Agron J 100:813–818 (2008).

10 Schonherr J and Baur P, Effects of temperature, surfactants and otheradjuvants on rates of uptake of organic compounds, in PlantCuticles – An Integrated Functional Approach, ed. by Kerstiens G.Bios Science Publisher Ltd, Oxford, pp. 135–156 (1996).

11 Alvarez-Fernandez A, Garcıa-Lavina P, Fidalgo C, Abadıa J andAbadıa A, Foliar fertilization to control iron chlorosis in pear (Pyruscommunis L.) trees. Plant Soil 263:5–15 (2004).

12 Rombola AD, Bruggemann W, Tagliavini M, Marangoni B andMoog PR, Iron source affects iron reduction and re-greening ofkiwifruit (Actinidia deliciosa) leaves. J Plant Nutr 23:1751–1765(2000).

13 Fernandez V, Rıo V, Abadıa J and Abadıa A, Foliar iron fertilizationof peach (Prunus persica (L.) Batsch): effects of iron compounds,surfactants and other adjuvants. Plant Soil 289:239–252 (2006).

14 Fernandez V, Del Rıo V, Pumarino L, Igartua E, Abadıa J and Abadıa A,Foliar fertilization of peach (Prunus persica (L.) Batsch) with differentiron formulations: Effects on re-greening, iron concentration andmineral composition in treated and untreated leaf surfaces. SciHortic 117:241–248 (2008).

15 Hill-Cottingham DG, Photosensitivity of iron chelates. Nature175:347–348 (1955).

16 Vert G, Grotz N, Dedaldechamp F, Gaymard F, Guerinot ML, Briat JF,et al, IRT1, an arabidopsis transporter essential for iron uptake fromthe soil and for plant growth. Plant Cell 14:1223–1233 (2002).

17 Rodrıguez-Lucena P, Tomasi T, Pinton R, Hernandez-Apaolaza L,Lucena JJ and Cesco S, Evaluation of 59Fe-lignosulfonatescomplexes as Fe-sources for plants. Plant Soil 325:53–63 (2009).

18 Fernandez V, Orera I, Abadıa J and Abadıa A, Foliar iron-fertilisation offruit trees: present knowledge and future perspectives – a review.J Hort Sci Biotechnol 84:1–6 (2009).

19 Kim AS and Guerinot ML, Mining iron: iron uptake and transport inplants. FEBS Lett 581:2273–2280 (2007).

20 Bityustkii NP, Effects of carboxylic and phosphonic Fe-chelates on rootand foliar plant nutrition. Russ J Plant Physiol 42:444–453 (1995).

21 Stock D and Holloway PJ, Possible mechanisms for surfactant-inducedfoliar uptake of agrochemicals. Pestic Sci 38:165–177 (1993).

22 Schonherr J, Cuticular penetration of calcium salts: effects of humidity,anions, and adjuvants. J Plant Nutr Soil Sci 164:225–231 (2001).

23 Fernandez V and Ebert G, Foliar iron fertilization: a critical review.J Plant Nutr 28:2113–2124 (2005).


26

40


24 Espasa-Manresa R, La fertilizacion foliar con aminoacidos. Horticultura2:33–35 (1983).

25 Briat JF, Curie C and Gaymard F, Iron utilization and metabolism inplants. Plant Biol 10:276–282 (2007).

26 Pestana M, Correia PJ, De Varennes A, Abadıa J and Araujo-Faria E,Effectiveness of different foliar iron applications to control ironchlorosis in orange trees grown on a calcareous soil. J Plant Nutr24:613–622 (2001).

27 Huve K, Remus R, Luttschwager D and Merbach W, Transport of foliarapplied iron (59Fe) in Vicia faba. J Plant Nutr 26:2231–2242 (2003).

28 Nikolic M, Cesco S, Romheld V, Varanini Z and Pinton R, Uptake ofiron (59Fe) complexed to water-extractable humic substances bysunflower leaves. J Plant Nutr 26:2243–2252 (2003).

29 Romera FJ, Lucena C and Alcantara E, Plant hormones influencingiron uptake in plants, in Iron Nutrition in Plants and RhizosphericMicroorganisms, ed. by Barton LL and Abadıa J. Springer-VerlagAcademic Publishers, Dordrecht, pp. 255–278 (2006).

30 Swietlik D and Faust M, Foliar nutrition of fruit crops. Hort Rev6:287–356 (1984).

31 Nikolic M, Cesco S, Romheld V, Varanini Z and Pinton R, Short-terminteractions between nitrate and iron nutrition in cucumber. FunctPlant Biol 34:402–408 (2007).

32 Orbovic V, Jifon JL and Syvertsen JP, Foliar-applied surfactants andurea temporally reduce carbon assimilation of grapefruit leaves.J Am Soc Hort Sci 126:486–490 (2001).


26

41



Influence of genetic matrix and crop yearon chemical and sensory profiles of Italianmonovarietal extra-virgin olive oilsAnnalisa Rotondi,a∗ Barbara Alfei,b Massimiliano Maglia

and Giorgio Pannellic

Abstract

BACKGROUND: Commercial virgin olive oils belonging to the cultivars (Bosana, Carolea, Coratina, Frantoio, Itrana, Leccino,Moraiolo, Peranzana, Piantone di Mogliano and Ravece) most represented at the Italian National Review of Monovarietal oliveoils (Rassegna Nazionale Italiana degli oli Monovarietali) were considered. The evaluation of the influence of the cultivar andof the crop year as well as their interaction on oil composition were statistically analysed by a complete factorial design byprincipal components analysis and by linear discriminant analysis.

RESULTS: In fatty acids composition, the effect of the cultivar and crop year and their interaction were highly significant. Thestatistical analysis showed that the sensory attributes (olive fruity, grassy, fresh almond, artichoke, tomato, aromatic herbs,bitter and pungent) were strongly influenced by the cultivar. The prevalent effect of the cultivar on the sensory profile was alsodemonstrated by the low or absent level of significance observed in the crop year.

CONCLUSION: The construction of a databank based on a large number of samples, which is available at URL http://www.olimonovarietali.it, has contributed to the reduction of the variable effects involved in the oil production process. Knowledgeof the chemical and sensory profiles of the Italian monovarietal olive oils could start a certification process of these oils, thusgiving greater guarantees about their origin.c© 2010 Society of Chemical Industry

Keywords: monovarietal in extra virgin olive oils; Olea europaea; cultivar; chemical and sensory identity; statistical analyses

INTRODUCTIONThe safeguarding of Italian autochtonous olive cultivars representsan important aim which each region is pursuing. This isdemonstrated by the continuous increase of monovarietal oliveoils participating in the Italian National Review of Monovarietalolive oils organised in Marche Region. Between 2005 and 2008, 986olive oil samples originating from 117 varieties and 18 differentItalian regions, have reached the Review organised by AgenziaServizi Settore Agroalimentare delle Marche (ASSAM), IstitutoCentro di Ricerca per l’Olivicoltura e l’Industria Olearia, sededistaccata di Spoleto (CRA- OLI) and Il Sole24ore-Business Media.

The richness of Italian autochthonous olive germplasm, repre-sented by 822 cultivars,1 which are still increasing, guarantees theproduction of high quality extra-virgin olive oils, contributing insuch a way as to maintain a large part of the olive genetic ancientbiodiversity.

Since the chemical and sensory characters of extra-virgin oliveoil (EVOO) are strongly affected by the genotype of origin, thesafeguarding and characterisation of cultivars and clones play akey role in the marketing of high quality standard olive oils.

Several typical Italian EVOOs have received a EuropeanProtected Origin Denomination (POD) trademark or a EuropeanProtected Geographical Indication (PGI) trademark. These typicaloils are generally blends of different varieties or monovarietal

EVOOs. The studies of the quality of monovarietal oils increase thevalue of the product and, at the same time, inform the consumerof their nutritional and organoleptic value.

In Italy the market of monovarietal and organic oils is increasingthanks to the presence of consumers who are paying greaterattention to the pleasurable aspect as well as the health aspect.2,3

The knowledge and the increase in the characteristics of Italianmonovarietal EVOOs will also improve the knowledge of areaswhere these oils are produced, with a consequent increasein tourism, a very important sector for the Italian economy.In Italy, a new regulation was recently introduced compellingvirgin and extra-virgin olive oil producers to indicate the

∗ Correspondence to: Annalisa Rotondi, Istituto di Biometeorologia,Consiglio Nazionale delle Ricerche, Via Gobetti 101, I-40129 Bologna, Italy.E-mail: [email protected]

a Istituto di Biometeorologia, Consiglio Nazionale delle Ricerche, Via Gobetti 101,I-40129 Bologna, Italy

b Agenzia Servizi Settore Agroalimentare delle Marche, Via Alpi 21 I-60131Ancona, Italy

c CRA, Istituto Centro di Ricerca per l’Olivicoltura e l’Industria Olearia, sededistaccata di Spoleto, Via Nursina 2, I-06049 Spoleto, Perugia, Italy


26

42

www.soci.org A Rotondi et al.

Table 1. Cultivar, geographical region and crop year of 489 monovarietal olive oils

Crop year (no. of samples)

Cultivar Geographical region 2005/2006 2006/2007 2007/2008 2008/2009Total no.

of samples

Bosana Sardegna 11 20 28 23 82

Carolea Calabria 8 5 5 3 21

Coratina Puglia 8 11 16 14 49

Frantoio Marche, Lazio, Toscana, Umbria 14 19 21 13 67

Itrana Lazio 13 19 9 14 55

Leccino Abruzzo, Marche, Toscana, Umbria 8 13 21 17 59

Moraiolo Toscana, Umbria 7 14 17 15 53

Peranzana Puglia 7 7 7 6 27

Piantone Di Mogliano Marche 10 6 7 11 34

Ravece Campania 4 10 14 14 42

Total no. of samples 90 124 145 130 489

0

100

200

300

400

500

600

Bosana Carolea Coratina Frantoio Itrana Leccino Moraiolo Peranzana Piantone M. Ravece

Varietes

To

tal p

hen

ols

(m

g g

allic

aci

d k

g-1

)

Figure 1. Mean values and standard error of the mean of the total phenols relative to the 489 monovarietal olive oil samples. Contents were expressedas mg kg−1 acid gallic of oil.

geographical location of olive harvesting and oil productionon the label.4 More recently, the European Community (EC)Council of Regulation established compulsory standards on oliveoil production regarding labelling giving the origin for extra-virginand virgin olive oils.5 Quality and typicality of EVOO are primarilydetermined by genetic, agronomical and environmental factors,even if the technological parameters of oil processing also play animportant role.6,7

Nutritional and sensory properties of EVOOs are stronglyaffected by several agronomical factors such as cultivar, fruitripeness, crop yield, and growing area.8 – 10 With regard to fattyacid composition, Uceda and Hermoso,11 in a preliminary olivegermplasm bank evaluation, concluded that the cultivar was themain source of variability for the major fatty acids. The strongrelationship between the variety and the composition of fattyacids is also shown by several studies.12,13

EVOO is also well known for its high content of pheno-lic substances, which are thought to have health-promoting

properties.14,15 Phenolic compounds are the main respon-sible agents for the resistance against autoxidation andphotoxidation.16 They also influence some sensorial propertiesof olive fruits and virgin olive oils.17,18 The large increase of thedemand of EVOO is due not only to its health virtues, but also toits organoleptic properties. The richness of Italian olive cultivarsallows the production of different monovarietal EVOOs charac-terised by a wide range of pleasant sensory flavours. Also, thephenols, as well as the sensory properties of EVOO, are influencedby the cultivar19 and environmental factors20 or their interaction.

Some authors investigated the effect of cultivar–environmentinteraction on the composition of EVOO.21,22 Environments whereolive trees grow may influence plant physiology as well as the oliveripening process. During the ripening process, the fruit changesits chemical composition, with the activation and inhibition ofdifferent enzymatic activities, also affecting oil composition. Asmaller number of studies has considered the seasonal effect onthe chemical and sensory profile of EVOOs. The seasonality which


26

43

Genetic matrix and crop year effects on Italian monovarietal olive oil quality www.soci.org

Tab

le2

.M

ean

valu

esan

dst

and

ard

erro

ro

fth

em

ean

oft

he

fatt

yac

ids

rela

tive

toth

e48

9m

on

ova

riet

alo

live

oil

sam

ple

s

Fatt

yac

idB

osa

na

Car

ole

aC

ora

tin

aFr

anto

ioIt

ran

aLe

ccin

oM

ora

iolo

Pera

nza

na

Pian

ton

ed

iM

og

lian

oRa

vece

Eico

san

oic

0.41

±0.

010.

39±

0.02

0.40

±0.

010.

36±

0.01

0.31

±0.

010.

32±

0.01

0.36

±0.

010.

39±

0.01

0.42

±0.

010.

57±

0.14

Eico

sen

oic

0.26

±0.

010.

28±

0.02

0.33

±0.

030.

29±

0.01

0.22

±0.

020.

22±

0.01

0.27

±0.

020.

35±

0.04

0.32

±0.

030.

01±

0.06

Hep

tad

ecan

oic

0.07

±0.

010.

14±

0.02

0.07

±0.

010.

06±

0.01

0.06

±0.

000.

06±

0.01

0.05

±0.

000.

05±

0.00

0.12

±0.

010.

06±

0.01

Hep

tad

ecen

oic

0.10

±0.

010.

24±

0.04

0.08

±0.

000.

10±

0.00

0.09

±0.

000.

10±

0.00

0.09

±0.

000.

08±

0.00

0.23

±0.

010.

09±

0.01

Lin

ole

ic10

.24

±0.

147.

13±

0.26

7.10

±0.

167.

01±

0.12

6.38

±0.

146.

69±

0.17

7.14

±0.

129.

50±

0.26

6.67

±0.

199.

52±

0.20

Lin

ole

nic

0.72

±0.

020.

61±

0.02

0.78

±0.

020.

69±

0.01

0.77

±0.

020.

74±

0.02

0.76

±0.

020.

78±

0.02

0.79

±0.

020.

78±

0.02

Ole

ic73

.05

±0.

2074

.81

±0.

6078

.37

±0.

2476

.90

±0.

2378

.12

±0.

2575

.56

±0.

2676

.23

±0.

2373

.88

±0.

4077

.44

±0.

3573

.43

±0.

28

Palm

itic

12.4

8±

0.11

13.3

1±

0.34

10.7

7±

0.16

12.1

3±

0.14

11.6

2±

0.15

13.6

5±

0.13

12.5

4±

0.17

12.4

8±

0.25

11.4

3±

0.22

12.3

1±

0.13

Palm

ito

leic

0.76

±0.

021.

16±

0.11

0.45

±0.

030.

82±

0.03

0.83

±0.

031.

28±

0.17

0.76

±0.

020.

75±

0.02

0.69

±0.

030.

66±

0.03

Stea

ric

1.93

±0.

051.

92±

0.08

1.65

±0.

061.

63±

0.04

1.59

±0.

051.

54±

0.04

1.53

±0.

051.

75±

0.06

1.86

±0.

052.

56±

0.08


26

44


is deeply related to the different climate events of the crop year,may influence the ripening process of olives, affecting in this waythe oil composition.23,24

In order to create a database of chemical and sensory profiles ofItalian Monovarietal EVOOs, an elevated number of observations,available every year, could allow the continuous update of thedatabase and give a more accurate data set.

The aim of the study is to describe the nutritional properties,expressed as fatty acid and total phenols contents, and the sensoryprofiles of Italian monovarietal EVOOs. In addition, the evaluationof the influence of the cultivar and seasonality, as well as theirinteraction on oil composition, were analysed.

MATERIALS AND METHODSCommercial virgin olive oil samples (n = 489) belonging to the tencultivars most represented (n ≥ 20) at the Review and declaredmonovarietal by the producer, were obtained from the ItalianNational Review of Monovarietal Olive Oils during four successiveeditions (crops 2005/2006, 2006/2007, 2007/2008 and 2008/2009).Table 1 shows geographical region, cultivar, crop year and numberof samples.

Sensory and chemical analyses of oils were evaluated each yearduring the Review 2–4 months after their processing.

Sensory analysis was determined by the ‘ASSAM – Marche Panel’recognised by the IOOC (International Olive Oil Council) and theItalian Ministry for Agriculture, Food and Forestry Policy. Theevaluation of the samples was performed under the conditionsdescribed in the EC Regulation 640/2008.25 Each taster smelledand tasted the oil under consideration, in order to analysethe olfactory, gustatory, tactile and kinaesthetic characteristics.Thirteen attributes were evaluated: nine during the olfactoryphase (olive fruity, olive fresh leaf, grass, fresh almond, artichoke,tomato, apple, berries and aromatic herbs) and four during thegustatory phase (olive fruity, bitter, pungent and the fluidity).Attributes were assessed on an oriented 10 cm line scale andquantified measuring the location of the mark from the origin.The data obtained for the 13 descriptors were used to define thesensory profile of each sample using the median values.

Fatty acids composition was determined by gas chromatogra-phy (GC) according to Regulation EC Reg.796/200226 methodologyby Centro Agrochimico ASSAM, Jesi (AN), Italy. Total phenolic con-tent, expressed in mg kg−1 acid gallic of oil, was determined bya colorimetric method using a spectrophotometer (Perkin Elmermodel Lambda 3b; Perkin Elmer, Waltham, MA, USA). Chemicaland sensory data were processed using SAS software 9.1.3 (SASInstitute Inc., Cary, NC, USA).

Explorative analysis and descriptive statistics were performedfor each set of data in order to identify outliers, extremeobservations and to obtain distributional properties of the data.Descriptive measures (moments, basic measures of location andvariability, confidence intervals for the mean, standard deviation,and variance) of chemical and sensory variables were calculatedfor each monovarietal oil. Subsequently the statistical procedurewas based on the analysis of variance (ANOVA) by a completefactorial design in order to examine treatment interdependencies(variety and crop year). A principal components analysis (PCA)and a linear discriminant analysis (LDA) were also performed onchemical and sensory data separately, using mean values of eachcrop year of each cultivar collected.

The fatty acid content of monovarietal olive oil samples wassubmitted to the ANOVA procedure by a complete factorial design.

Table 3. Variability expressed as % of the total sum of the squares forfatty acid composition and total phenols

Parameter Cultivar Crop year Cultivar × crop year

Eicosanoic 19.67∗∗∗ 10.56∗∗∗ 69.77∗∗∗

Eicosenoic 12.75∗∗∗ 78.53∗∗∗ 8.72∗∗∗

Heptadecanoic 53.08∗∗∗ 20.69∗∗∗ 26.23∗∗

Heptadecenoic 67.04∗∗∗ 7.92∗∗∗ 25.04∗∗∗

Linoleic 86.17∗∗∗ 6.07∗∗∗ 7.76∗∗∗

Linolenic 12.14∗∗∗ 81.20∗∗∗ 6.66∗∗∗

Oleic 82.28∗∗∗ 9.06∗∗∗ 8.66∗∗∗

Palmitic 63.90∗∗∗ 28.27∗∗∗ 7.83ns

Palmitoleic 62.85∗∗∗ 12.10∗∗∗ 25.05ns

Stearic 53.39∗∗∗ 36.45∗∗∗ 10.16∗∗∗

Total phenols 37.40∗∗∗ 49.02∗∗∗ 13.58∗

∗ , ∗∗ , ∗∗∗ Significant F-values: the ∗ 0.05, ∗∗ 0.01 or ∗∗∗ 0.001 levels,respectively; ns = not significant.

Fatty acids with the highest index of variability (heptadecenoic,linoleic, oleic, stearic and palmitic) were selected according totheir P level and F values and submitted to PCA and LDA.

RESULTS AND DISCUSSIONTable 2 reports mean values and standard errors of the mean of thefatty acids belonging to the 489 monovarietal olive oil samples.Regarding the oleic acid, whose range contents reported to ECOfficial Reg. EC Regulation 702/200727 varies from 55% to 83%,the monovarietal EVOOs considered in this study are generallycharacterised by high levels (superior to 73%), moreover Coratinaand Itrana monovarietal EVOOs showed the highest amount ofacid oleic (above 78%). The percentage of the other individualfatty acids (markers of genuineness), such as myristic, linolenic,arachidonic, eicosenoic, and behenic acid were within the limitsset by the official normal standard.27

As shown in Fig. 1, oil of the Coratina cultivar presented thehighest value of total phenols (543.06 mg kg−1), while Itrana oilsshowed the lowest level of 305.02 mg kg−1. However, the phenoliccontent of the other monovarietal EVOOs, having a range between376.37 and 498.90, showed their high levels of nutritional quality.

In order to evaluate the treatment interdependencies (cultivarand crop year) on the fatty acid and phenolic contents, an ANOVAprocedure by complete factorial design was made. In Table 3 it isreported that in all fatty acids analysed, the effect of the cultivarand crop year is highly significant. The effect of the interactionbetween the two factors is also highly significant on the contentof fatty acid, except palmitic and palmitoleic acids. Also, the totalphenolic contents are deeply influenced both by the cultivar andyear, while the interaction between the two factors has showeda minor level of significance. It is interesting to underline thatboth factors (cultivar and crop year) influence at the same levelof significance the contents of the most important fatty acidsas linoleic, linolenic, oleic, palmitic and palmitoleic; however foroleic and linoleic acid the main factor was the cultivar, in factANOVA procedure explains the 82.28% and 86.17% of its variationrespectively. The cultivar did not represent a great source ofvariability for linoleic acid: only 12.14%, while the crop year showsa variation of 81.20%.

Figure 2 shows the PCA, applied on the mean values of eachmonovarietal olive oil and fatty acids content. The figure describes


26

45


Figure 2. Principal component analysis applied to the fatty acids (heptadecenoic, linoleic, oleic, stearic and palmitic) expressing the highest index ofvariability. The figure describes each monovarietal EVOO considering the four following crop years individually. B, Bosana; P, PeranzanA; R, Ravece; CA,Carolea; L, Leccino; PM, Piantone di Mogliano; M, Moraiolo; C, Coratina; F, Frantoio; I, Itrana.

each monovarietal EVOO considering the four following cropyears individually. This analysis explains the 71.16% of variabilityrepresenting the discrimination among some cultivars. At the topof the figure are localised Bosana (B), Peranzana (P) and Ravece(R) oils. These are characterised by a higher level of linoleic acid,while cv. Itrana (I) and Coratina (C) by a higher level of oleic acid.Cv. Carolea (CA) showed a higher level of heptadecenoic acid. Dataplotted on the principal components are, however, spread insidesome cultivars, confirming the significant effect of the cultivar andcrop year on the fatty acid composition, as previously shown by theANOVA procedure. Linear discriminant analysis (LDA) on fatty acidsconfirmed the result of the PCA. The analysis of the misclassifiedobservations showed that the LDA model presented a relativelylow specificity (77.5%), thus, while being able to identify fourvarieties (Itrana, Peranzana, Piantone di Mogliano and Ravece),its ability to classify the rest of observations is fairly limited; themajority of these misclassified observations were Bosana (50%)and Frantoio (75%) (Table 4).

Considering organoleptic quality for each of the 489 olive oilsamples collected, the sensory profile was defined. Table 5 showsmean values and standard error of the mean of the sensoryattributes of the monovarietal EVOOs. The monovarietal EVOOsproduced by cultivars Bosana, Frantoio, Itrana, Peranzana andRavece exhibited a higher intensity (mean, about 5) of olivefruity odour. The intensity of olive fruity perceived in the abovementioned oils, was also confirmed by the gustative route. Allmonovarietal EVOOs have presented significant intensities of grassattribute with the highest levels of 3.01 and 2.90, respectively,perceived in Itrana and Ravece oils. Considering the peculiarattributes which are deeply cultivar-dependent, like fresh almond,artichoke, tomato, aromatic herbs and berries,28 – 30 oils producedby Coratina, Frantoio, Leccino, Moraiolo and Piantone di Mogliano

Table 4. Linear discriminant analysis (error count estimates forvarieties) applied on fatty acids

Cultivar

Bosana 0.50

Carolea 0.25

Coratina 0.25

Frantoio 0.75

Itrana 0.00

Leccino 0.25

Moraiolo 0.25

Peranzana 0.00

Piantone di Mogliano 0.00

Ravece 0.00

are distinguished for their high intensity of fresh almond, atypical pleasant flavour which characterised these cultivars. Allmonovarietal oils considered in this study presented a significantintensity of artichoke flavours. Bosana, Peranzana and Ravece andoils exhibited the highest intensity, while in Piantone di Moglianoand Leccino oils, artichoke and tomato attributes were slightlyperceived. With regard to this last attribute, oils of Ravece andItrana are distinguished also for their high intensity of tomatoflavour. Itrana oil, previously described for its high intensity ofolive fruity odour, is also characterised by a high intensity oftomato flavours. Aromatic herbs flavour is significantly presentonly in Itrana and Ravece oils.

All monovarietal EVOOs considered in this study were charac-terised by a significant level of bitterness showing a range from3.65 to 5.14. In particular, Piantone di Mogliano, Leccino and Itrana


26

46


Table 5. Mean values and standard error of the median of the sensory attributes relative to the 489 monovarietal olive oil samples

Sensoryattribute Bosana Carolea Coratina Frantoio Itrana Leccino Moraiolo Peranzana

PiantoneDi

Mogliano Ravece

Olive fruity(olf.)

5.08 ± 0.07 4.81 ± 0.24 4.80 ± 0.11 5.05 ± 0.09 5.39 ± 0.15 4.63 ± 0.13 4.91 ± 0.09 5.06 ± 0.09 4.31 ± 0.13 5.69 ± 0.10

Olive fruity(gus)

5.00 ± 0.06 4.63 ± 0.19 4.83 ± 0.08 4.92 ± 0.08 5.08 ± 0.12 4.38 ± 0.11 4.81 ± 0.10 4.88 ± 0.12 4.28 ± 0.14 5.39 ± 0.09

Grass 2.61 ± 0.13 2.14 ± 0.33 1.56 ± 0.19 2.39 ± 0.17 3.01 ± 0.20 1.61 ± 0.19 1.84 ± 0.19 2.72 ± 0.28 1.50 ± 0.20 2.90 ± 0.20

Freshalmond

1.64 ± 0.12 1.61 ± 0.24 2.20 ± 0.18 2.61 ± 0.13 1.22 ± 0.16 2.34 ± 0.13 2.15 ± 0.15 1.51 ± 0.20 2.09 ± 0.15 1.08 ± 0.15

Artichoke 2.16 ± 0.12 1.73 ± 0.26 1.69 ± 0.17 1.51 ± 0.16 1.83 ± 0.17 0.99 ± 0.14 1.80 ± 0.16 2.28 ± 0.23 0.98 ± 0.17 2.15 ± 0.15

Tomato 0.76 ± 0.11 0.41 ± 0.14 0.43 ± 0.11 0.39 ± 0.10 2.14 ± 0.18 0.19 ± 0.06 0.53 ± 0.12 1.11 ± 0.23 0.41 ± 0.14 2.47 ± 0.22

Aromaticherbs

0.09 ± 0.04 0.06 ± 0.06 0.06 ± 0.03 0.02 ± 0.02 0.29 ± 0.09 0.03 ± 0.02 0.02 ± 0.02 0.14 ± 0.08 0.09 ± 0.06 0.30 ± 0.12

Bitter 4.84 ± 0.08 4.30 ± 0.33 5.14 ± 0.14 4.61 ± 0.13 4.01 ± 0.13 3.99 ± 0.16 4.58 ± 0.15 4.13 ± 0.18 3.65 ± 0.15 4.70 ± 0.13

Pungent 4.52 ± 0.07 3.78 ± 0.24 4.79 ± 0.11 4.39 ± 0.10 3.85 ± 0.13 3.99 ± 0.13 4.39 ± 0.12 4.08 ± 0.12 4.03 ± 0.16 4.75 ± 0.12

Table 6. Variability expressed as % of the total sum of the squares forsensory attributes

Parameter Cultivar Crop year Cultivar × crop year

Olive fruity (olf.) 72.65∗∗∗ 8.84∗ 18.51ns

Olive fruity (gus.) 68.07∗∗∗ 5.01ns 26.92ns

Grass 65.34∗∗∗ 5.29ns 29.37ns

Fresh almond 71.53∗∗∗ 7.48∗ 20.99ns

Artichoke 56.63∗∗∗ 4.87ns 38.50∗

Tomato 83.58∗∗∗ 5.16ns 11.26∗

Aromatic herbs 42.22∗∗∗ 10.49ns 47.29ns

Bitter 73.20∗∗∗ 8.21∗ 18.59ns

Pungent 57.34∗∗∗ 10.69∗∗ 31.97∗

∗ , ∗∗ , ∗∗∗ Significant F-values the ∗ 0.05, ∗∗ 0.01 or ∗∗∗ 0.001 levels,respectively; ns = not significant.

oils presented the lowest intensity of bitterness. It is interestingto underline that the same oils were also characterised by thelowest contents of total phenols (see Fig. 1). On the contrary, theCoratina oil which exhibited the highest intensity of bitterness,also presents the highest phenolic content. These results showa clear positive correlation between bitterness and total phenols(r2 = 0.65) according to other authors.31

In the present study the monovarietal oils show values ofpungency quite similar among them. In fact the range varied from3.78 to 4.75, but correlation between bitterness and pungency ishigh (r2 = 0.75).

The sensory profiles of the 489 monovarietal olive oil sampleswere submitted to the ANOVA procedure by a complete factorialdesign. Table 6 shows that all sensory attributes: olive fruity(odour), olive fruity (taste), grass, fresh almond, artichoke, tomato,aromatic herbs, bitter and pungent were strongly influencedby the cultivar. The prevalent effect of the cultivar on thesensory profile of monovarietal oils was demonstrated by thelow or absent level of significance observed in the cropyear, relative interaction and by the variability, expressed aspercentage of the total sum of the squares, of the factorcultivar characterised by a range from 42.22% to 83.58%. Theprevalent effect of the cultivar with respect to the crop year isalso evidenced by the PCA, where the data plotted showed a

clear grouping inside each cultivar (Fig. 3), explaining the 79%of variability. LDA analysis on sensory attributes confirmed theresult of the PCA. The analysis of the misclassified observationsshowed that the LDA model presented a high specificity(85.0%). The model is able to identify correctly five varieties(Coratina, Itrana, Leccino, Peranzana and Ravece); the majority ofthese misclassified observations were Bosana observations (50%)(Table 7).

CONCLUSIONThe decision to carry out a study of an increased number oflabelled commercial extra-virgin olive oils has had the precise aimto provide the consumer with information about the chemicaland sensory properties of extra-virgin olive oils which are actuallyavailable on the Italian market. The authors are aware of thenumerous variables: mills typology, olive ripening index andagronomic practices, which influence the overall olive oil quality.These variables are not usually known for commercial oils.

The construction of a databank based on a large number ofsamples which is available at URL http://www.olimonovarietali.it,has contributed to the reduction of the variable effects involvedin the oil production process. Moreover the continuous increaseof oil samples participating in the Italian National Review ofmonovarietal olive oils will allow the improvement of the chemicaland sensory profiles of each oil.

The exact knowledge of sensory profiles of the Italianmonovarietal olive oils represents an important step, consideringthat the production of monovarietal EVOOs has greatly increasedduring the past few years. As reported by Chiavaro et al.,2 thecharacterisation of monovarietal oils is needed to identify specificproperties that distinguish them and increase their value withrespect to other niches of EVOOs such as PDO.

In this study only ten of the 117 cultivars, available in thedatabank, which represent the Italian olive genetic resourcesare reported. Further studies will be carried out involving othercultivars. Considering the cultivars Frantoio and Leccino, whichare widely diffused along the Italian peninsula, studies are inprogress to estimate the effect of the environment (climate,altitude and latitude) on the chemical and sensory profiles ofolive oils participating in the Italian National Review.


26

47


Figure 3. Principal component analysis applied to the sensory attributes (olfactive and gustative olive fruity, grass, fresh almond, artichoke, tomato,aromatic herbs, bitter and pungent). The figure describes each monovarietal EVOO considering the four following crop years individually. B, Bosana; P,PeranzanA; R, Ravece; CA, Carolea; L, Leccino; PM, Piantone di Mogliano; M, Moraiolo; C, Coratina; F, Frantoio; I, Itrana.

Table 7. Linear discriminant analysis (error count estimates forvarieties) applied on sensory attributes

Cultivar

Bosana 0.50

Carolea 0.25

Coratina 0.00

Frantoio 0.25

Itrana 0.00

Leccino 0.00

Moraiolo 0.25

Peranzana 0.00

Piantone di Mogliano 0.25

Ravece 0.00

The knowledge and the information of the chemical andsensory profiles of the Italian monovarietal olive oils couldstart a certification process of these oils, giving in such a waygreater guarantees about their origin and consequently greaterguarantees of quality for the consumer. The importance ofmonovarietal oils is becoming more significant in restaurantsand eating establishment, as a result of associating local oils withtypical dishes.

ACKNOWLEDGEMENTSWe gratefully thank ‘Il Sole24ore’-Business Media, Srl, Via Patecchio2, Milano, Italy.

REFERENCES1 Bartolini G, Olive Germplasm (Olea europaea L.): cultivars,

synonyms, cultivation area, collections (2008). Available:http://www.oleadb.eu/. [30 April 2010].

2 Chiavaro E, Vittadini E, Rodriguez-Estrada MT, Cerretani L andBendini A, Monovarietal extra virgin olive oils. Correlationbetween thermal properties and chemical composition: heatingthermograms. J Agric Food Chem 56:496–501 (2008).

3 Consolandi C, Palmieri L, Severgnini M, Maestri E, Marmiroli N,Agrimonti C, et al, A procedure for olive oil traceability andauthenticity: DNA extraction, multiplex PCR and LDR-universal arrayanalysis. Eur Food Res Technol 227:1429–1438 (2008).

4 Italian Government Legislative Decree of 18 October 2007 Regulationon compulsory indications which appear on the labelling of virginand extra virgin olive oil. Official Gazette Italian Republic 243:8–9(2007).

5 European Commission Regulation 182/2009. Official Journal ofEuropean Community, 6 March 2009, amending Regulation (EC)N. 1019/2002, L63:6–8 (2009).

6 Cerretani L, Bendini A, Rotondi A, Lercker G and Gallina Toschi T,Analytical comparison of monovarietal virgin olive oils obtainedby both a continuous industrial plant and a low-scale mill. Eur J LipidSci Technol 107:93–100 (2005).

7 Angerosa F, Mostallino R, Basti C and Vito R, Influence of malaxationtemperature and time on the quality of virgin olive oils. Food Chem72:19–28 (2001).

8 Lazzez A, Perri E, Caravita M, Khalifa M and Cossentini M, Influenceof olive maturity stage and geographical origin on some minorcomponents in virgin olive oil of the Chemlali variety. J Agric FoodChem 56:982–988 (2008).

9 Rotondi A, Bendini A, Cerretani L, Mari M, Lercker G and GallinaToschi T, Effect of olive ripening degree on the oxidative stabilityand organoleptic properties of cv. Nostrana di Brisighella extravirgin olive oil. J Agric Food Chem 52:3649–3654 (2004).

10 Morello JR, Romero MP and Motilva MJ, Effect of the maturationprocess of the olive fruit on the phenolic fraction of drupes and


26

48


oils from Arbequina, Farga, and Morrut cultivars. J Agric Food Chem52:6002–6009 (2004).

11 Uceda M and Hermoso M, La Calidad del Aceite de Oliva, in El Cultivodel Olivo, ed. by Junta de A, Sevilla and Mundi-Prensa, Madrid,pp. 589–614 (2001).

12 D’Imperio M, Dugo G, Alfa M, Mannina L and Segre AL, Statisticalanalysis on Sicilian olive oils. Food Chem 102:956–965 (2007).

13 Allalout A, Krichnene D, Methenni K, Taamalli A, Oueslati I, Daoud D,et al, Characterization of virgin olive oil from super intensive Spanishand Greek varieties grown in northern Tunisia. Sci Hort 129:77–83(2009).

14 Menendez JA, Vasquez-Martın A, Colomer R, Carrasco-Pancorbo A,Garcıa-Villalba R, Fernandez-Gutierrez A, et al, Olive oil’s bitterprinciple reverses acquired autoresistance to trastuzumab(Herceptin) in HER2-overexpressing breast cancer cells. BMC Cancer7:7–80 (2007).

15 Bendini A, Cerretani L, Carrasco-Pancorbo A, Gomez-Caravaca AM,Segura-Carretero A, Fernandez-Gutierrez A, et al, Phenolicmolecules in virgin olive oils. Survey of their sensory properties,health effects, antioxidant activity and analytical methods. AnOverview of the last decade. Molecules 12:1678–1719 (2007).

16 Baccouri O, Cerretani L, Bendini A, Lercker G, Zarrouk M and BenMiled D, Phenol contents correlated to antioxidant activity andgustative characteristics of Tunisian monovarietal virgin olive oils.Riv Ital Sost Grasse 85:189–195 (2008).

17 Botia J, Ortuno A, Benavente-Garcia O and Baidez AG, Modulation ofthe biosynthesis of some phenolic compounds in Olea europaea L.fruits: their influence on olive quality. J Agric Food Chem 49:355–358(2001).

18 Carrasco-Pancorbo A, Gomez-Caravaca AM, Segura-Carretero A,Cerretani L, Bendini A and Fernandez-Gutierrez A, Use of capillaryelectrophoresis with UV detection to compare the phenolic profilesof extra virgin olive oils belonging to Spanish and Italian PDOs andtheir relation to sensorial properties. J Sci Food Agric 89:2144–2155(2009).

19 Oliveras-Lopez MJ, Innocenti M, Giaccherini C, Ieri F, Romani A andMulinacci N, Study of the phenolic composition of Spanish andItalian monocultivar extra virgin olive oils: distribution of lignans,secoiridoidic, simple phenols and flavonoids. Talanta 73:726–732(2007).

20 Paz Aguilera M, Beltran G, Ortega D, Fernandez A, Jimenez A andUceda M, Characterization of virgin olive oil of Italian olivecultivars: ‘‘Frantoio’’ and ‘‘Leccino’’, grown in Andalusia. Food Chem89:387–391 (2005).

21 Kotti F, Chiavaro E, Cerretani L, Barnaba C, Gargouri M and Bendini A,Chemical and thermal characterization of Tunisian extra virgin oliveoil from Chetoui and Chemlali cultivars and different geographicalorigin. Eur Food Res Technol 228:735–742 (2009).

22 Rotondi A and Lapucci C. Nutritional properties of extra virgin olive oilsfrom the Emilia-Romagna region: profiles of phenols, vitamins andfatty acids, in Olives and Olive Oil in Health and Disease Prevention, ed.by Preedy VR and Watson RR. Academic Press, Oxford, pp. 725–733(2010).

23 Rotondi A and Magli M, Ripening of olives var. Correggiolo:modification of oxidative stability of oils during fruit ripening andoil storage. J Food Agric Environ 2:193–199 (2004).

24 Beltran G, Del Rio C, Sanchez S and Martınez L, Influence of harvestdate and crop yield on the fatty acid composition of virgin olive oilsfrom cv. Picual. J Agric Food Chem 52:3434–3440 (2004).

25 European Commission Regulation No. 640/2008. Off J Eur Commun L178:11–16 (2008).



28 Angerosa F, Influence of volatile compounds on virgin olive oil qualityevaluate by analytical approaches and sensory panels. Eur J Lipid SciTechnol 104:639–660 (2002).

29 Aparicio R, Morales MT and Alonso V, Authentication of Europeanvirgin olive oils by their chemical compounds, sensory attributesand consumer attitudes. J Agric Food Chem 45:1076–1083 (1997).

30 Tura D, Prenzler PD, Bedgood D R, Antolovich M and Robards K,Varietal and processing effects on the volatile profile of Australianolive oils. Food Chem 84:341–349 (2004).

31 Inarejos-Garcıa AM, Androulaky A, Salvador D, Fregapane G andTsimidou, MZ, Discussion on the objective evaluation of virginolive oil bitterness. Food Res Int 42:279–284 (2009).


26

49

Research ArticleReceived: 27 February 2010 Revised: 23 July 2010 Accepted: 23 July 2010 Published online in Wiley Online Library: 25 August 2010


Physiological optimality, allocation trade-offsand antioxidant protection linked to better leafyield performance in drought exposedmulberryAnirban Guha, Debashree Sengupta and Attipalli Ramachandra Reddy∗

Abstract

BACKGROUND: Mulberry (Morus spp. L.), usually linked to silkworm rearing, is now considered as a potential forage for livestockfeeding and has great potential in world agriculture. Trait-based investigations for leaf yield stability in mulberry under waterstress have not been studied extensively. The present study aims to identify candidate traits conferring leaf yield stability inmulberry under drought.

RESULTS: Four popular, indigenous mulberry cultivars (Morus indica L. cvs AR-12, K-2, M. Local and V-1) were investigated.Low leaf temperature (Tl), higher internal/ambient CO2 ratios (Ci/Ca), greater stomatal conductance to CO2 (gs) and stability inphotosystem II efficiency were associated with better net photosynthetic rates (Pn) in V-1, generating maximum leaf yield whencompared to other drought-exposed cultivars. Increased accumulation of foliar α-tocopherol and ascorbic acid–glutathionepool, associated with higher carotenoids, proline and glycine betaine, facilitated lower lipid peroxidation and better leaf yieldin V-1 under drought.

CONCLUSION: Minimal plasticity in photosynthetic gas exchange traits and better quantitative growth characteristics wereattributed to leaf yield stability under drought. Lower photoinhibition, stabilized photochemistry, effective osmoregulationand enhanced activity of foliar antioxidants extensively contributed to drought tolerance and higher leaf yield in mulberry.c© 2010 Society of Chemical Industry

Keywords: antioxidants; chlorophyll a fluorescence; drought; leaf yield stability; mulberry; photosynthesis

INTRODUCTIONMulberry (Morus spp L.), a pioneer tree of secondary succession,was one of the first commercialized foliage crops in theworld and has been cultivated transcontinentally, covering 50countries across the globe.1 In the past, mulberry cultivationwas predominantly linked to the silkworm (Bombyx mori L.)and sericulture industry. However, research studies over thelast two decades have revealed several other potential uses ofmulberry. Recently, mulberry has been gaining popularity asforage for livestock feeding as the leaves are highly rich in protein,antioxidants and minerals, without any toxic elements. Severalreports have emphasized the potential of mulberry foliage asa feed for ruminant and non-ruminant animals.2,3 Hence thepurpose of mulberry research is largely aimed at producingmore foliage with palatable, succulent leaves of high nutritivevalue. Recently evolved high-yielding mulberry genotypes andimproved crop husbandry techniques have caused a massiveincrease in harvestable leaf yield in the moriculture sectors ofIndia.4 However, mulberry is also cultivated under resource-limited agro-ecosystems and many regions of the countryare currently facing the prospect of producing yields withlow water availability. Nearly 48% of the cultivated mulberryareas in India have been clustered under rainfed water-stressconditions.

Mulberry requires 500–700 L water to produce 1 kg of freshleaf.5 High-yielding mulberry cultivars (cvs) consume largequantities of water due to their faster growth rate, high biomass-producing foliage, large leaf area and canopy size; hence waterdeprivation can severely arrest mulberry growth, leading toyield loss. The relationship between yield loss and water stressseverity can differ largely among crop genotypes. This hasgiven rise to the concept of drought tolerance (DT), with somegenotypes performing better under a given severity of waterstress than others. Even though such variations have beenrecognized for many years, progress towards understandingand exploiting the mechanisms that confer tolerance has beenslow in many crop species due to inconsistent use of the termDT and the difficulties encountered while quantifying it. Froman agro-economical viewpoint, the functional definition of DT

∗ Correspondence to: Attipalli Ramachandra Reddy, Photosynthesis and PlantStress Biology Laboratory, Department of Plant Sciences, School of Life Sciences,University of Hyderabad, Hyderabad 500046, India.E-mail: [email protected]

Photosynthesis and Plant Stress Biology Laboratory, Department of PlantSciences, School of Life Sciences, University of Hyderabad, Hyderabad 500046,India


26

50

www.soci.org A Guha, D Sengupta, AR Reddy

should be based on yield stability, which precisely indicatesless fluctuation in yield components in a drought-tolerantgenotype compared to the susceptible when exposed to waterdeficit. Stabilizing yield performance requires optimization ofthe physiological processes involved in the critical stages ofplant response to soil dehydration. Several morphophysiologicaland biochemical mechanisms including water use efficiency,photosynthetic capacity, radiation use efficiency, rooting vigour,antioxidative protection, osmotic adjustment and staying green,are linked to enhanced performance and yield stability inagreement with stress tolerance syndrome.6,7 A priori knowledgeof these candidate crop traits contributing to DT is essential beforedesigning crop improvement programmes for any crop model.

In spite of being an age-old crop, such trait-based investigationsfor yield stability in drought-exposed mulberry cvs have notbeen carried out extensively, except for a few reports.8 – 11 Hencethe present investigation was focused on identifying candidatecrop traits conferring DT in mulberry. Our major objectives were:(a) to identify the morphophysiological traits that would maximizebiomass production as well as the use of available soil moistureunder drought; (b) to investigate the drought-induced allocationtrade-offs, partitioning to leaf biomass for enhanced crop waterproductivity (yield achieved per amount of water used); and (c) tolink the oxidative stress and mechanisms of antioxidant protectionto understand the overall physiological and growth performanceof mulberry cvs under water deficit conditions.

MATERIALS AND METHODSField experimentsPlant materials, study site and climatic conditionsThe tested plant materials consisted of four 2-year-old indigenousMorus indica cvs – AR-12, K-2, M. Local and V-1 – grown at theexperimental plots of the University of Hyderabad, India (17.3◦ 10′

N, 78◦ 23′ E; 542.6 m above sea level). The origin and someof the important genotypic characters of the tested cvs arelisted in Table 1. Hyderabad, the region of the present study,is in the hottest part of the state of Andhra Pradesh in Indiaand has a hot steppe climate characterized by high irradianceand air temperature during summer (February–July), followedby southwest monsoon (August–October). All samplings andestimations were performed during the dry summer months(February–July) of 2009 over two consecutive growth seasons:

season I (February–April 2009) and season II (May–July 2009).Mean precipitation during the experimental period was only49 mm and was considered insignificant to influence the study;mean air temperature recorded during daytime ranged from30.7 to 40.8 ◦C; mean photosynthetically active radiation (PAR)measured between 09.00 and 11.00 solar hours ranged from 1230to 2500 µmol m−2 s−1. The soil of the experimental site was sandyloam with a pH of 7.5.

Experimental design, stress treatments and samplingThe experimental design was a randomized block with a pitsystem of plantation and split plot arrangement of treatments,4,12

including four replications. The mulberry cvs were evaluated undertwo water regimes: well-watered (control) and water-stressed(drought). The control plot received two irrigations per week,whereas the stressed plot was exposed to natural drought andonly received survival irrigation (two to three irrigations in anentire growing season) to avoid depletion of soil moisture contentbelow 20% (at a soil depth of 30–40 cm). To verify the degreeof drought intensity, soil moisture content at 30–40 cm depthof the control and stressed plots was determined gravimetricallyon every sampling day. The soil moisture content of the well-watered and water-stressed plots ranged from 55% to 65% and22% to 28%, respectively, throughout the experimental period.The mulberry plantations were maintained as medium bushby pruning techniques and other farm practices were followedas recommended.4 All the samplings and measurements wereperformed in fully developed young leaves of the third or fourthposition from the apex of top branches. The first sampling andmeasurements were conducted in the first week of March 2009.Thereafter, all estimations were periodically conducted at intervalsof 20–25 days in each growing season and the results reported arethe mean of all periodic data obtained over two summer growingseasons (seasons I and II).

Measurement of leaf gas exchange parametersThe rate of leaf gas exchange was measured using a portableinfrared CO2/H2O gas analyser (IRGA) (LCpro-32 070, ADC Bio-scientific Ltd, Great Amwell, UK) equipped with a broad leafchamber. The gas analyser was used to measure instantaneousnet photosynthetic rate (Pn), stomatal conductance to CO2 (gs) andtranspiration rate (E), periodically during each growing season be-tween 10.00 and 11.00 a.m. on clear sunny days. The instant water

Table 1. Genotypic characteristics of four mulberry cvs (AR-12, K-2, M. Local and V-1) selected for drought tolerance experiment in this study

Characteristics AR-12 K-2 M. Local V-1

Genetic origin Cross-breed OPHa selection Land race OPH selection

Branching nature Erect Erect Erect Erect

Phyllotaxy 1/3 Mixed 1/3 Mixed

Internodal distance (cm) 4.6 4.5 3.7 4.5

Leaf nature Homophyllous Homophyllous Heterophyllous Homophyllous

Leaf lobation Unlobed Unlobed Medium lobed Unlobed

Leaf shape Cordate Ovate Ovate Narrow ovate

Leaf colour Dark green Light green Dark green Dark green

Sex expression Unisexual male Unisexual female Unisexual male or female Unisexual male

Ploidy level 3n = 42 2n = 30 2n = 28 2n = 28

Rooting (%) >80 >80 >90 >90

a OPH, open pollinated hybrid.


26

51

Characteristics conferring leaf yield stability in drought-stressed mulberry www.soci.org

use efficiency (WUEi) was calculated (WUEi = Pn/E). The plantswere also analysed for internal CO2 concentration (Ci) and internalto ambient CO2 ratios (Ci/Ca). Leaf temperature (Tl) was measuredsimultaneously by an integral leaf thermistor probe (M.PLC-011,LICOR, ADC) attached to the leaf chamber. All photosyntheticmeasurements were performed in situ on young, well-expandedand light-exposed leaves, randomly chosen from the upper half ofthe plant canopy of uniform plants in each replicate.

Measurement of leaf yield (LY) and determination of droughtsusceptibility index (DSI)At the end of each growing season, the leaves were harvestedand the weight recorded as fresh weight (FW) of leaves ingrams. Drought susceptibility index (DSI) was calculated:13 DSI= (1 − Yd/Yc)/DI where Yd and Yc are the leaf yields of the water-stressed and well-watered plants of one particular cv., respectivelyand DI is the drought intensity for the conditions of the givenexperimental time span. DI is defined as DI = 1 − Xd/Xc, where,Xd and Xc are the mean leaf yields of all compared cvs underwater-stressed and well-watered conditions, respectively Thus,DSI = (1 − Yd/Yc)/(1 − Xd/Xc); if DSI <1 the cv. is drought tolerant,whereas if DSI >1 the cv. is drought susceptible (in terms of leafyield harvest).

Glasshouse experimentsExperimental design, glasshouse conditions and samplingThe second phase of study was executed under a glasshousechamber to avoid difficulties in field observations due to thesouthwest monsoon, which usually arrives in our study zoneimmediately after the dry spell of 6 months (February–July). Three-month-old, healthy potted (pot size 35 L) saplings of all the testedmulberry cvs were arranged in a completely randomized blockdesign (CRBD) with four replications. The plants were randomlysubmitted to two watering treatments: water-stressed pots weremaintained at 25–30% pot capacity (PC), whereas the well-wateredplants were maintained at 100% PC. The study was undertakenfor a time period of 75 days (20 August 2009 to 30 October2009), calculated from the day after initiation of treatments.Chamber walls and ceiling were transparent to sunlight. Meanair temperature inside the glasshouse ranged from 22 ± 1 ◦C (earlymorning) to 34 ± 4 ◦C (early afternoon) and relative humidity from20% ± 5% to 41% ± 2%. Samplings were conducted periodicallyat an interval of 15–18 days and the results represented are themean values of all periodic data.

Measurement of leaf water statusTo monitor the effectiveness of the dry-down treatment, leafrelative water content (RWC) was measured and calculated:14 RWC(%) = [(FW − DW)/(SW − DW)] × 100, where FW is fresh weight,SW is the turgid mass after rehydration obtained by storing leafsamples for 24 h in distilled water and DW is oven-dried weight(105 ◦C for 24 h) of leaves. Leaf moisture content (LMC) wasestimated by using the formula LMC (%) = [(FW − DW)/FW] × 100.

Measurement of light-saturated net photosynthetic rates (LSPn),photosynthetic radiation use efficiency (PRUE) and chlorophyll afluorescenceNon-detached, young and fully expanded leaves from each cv.under each treatment were used to measure LSPn during 09.00to 11.00 a.m. by using a portable IRGA at a saturating PAR of

1600 µmol m−2 s−1. The PAR was provided by an LED light sourceattached to a broad light unit (LCpro Lamp 32 070 – Broad, ADC)of the leaf chamber. The air humidity in the leaf chamber wasabout 60% with a CO2 concentration of 350–360 µmol mol−1 andambient air temperature was 25 ± 2 ◦C. PRUE was determined:15

PRUE = light-saturated net photosynthetic rate (LSPn)/incidentPAR. Chlorophyll a fluorescence was measured using a portableplant efficiency analyser–fluorometer (Handy PEA-2126, Hansat-ech, King’s Lynn, UK). Measurements were made on dark-adapted(30 min) leaves at pre-dawn (5.00–5.30 a.m.). The ground fluo-rescence yield (Fo) and maximal fluorescence induction (Fm) wereobtained by illuminating the leaves with a beam of saturating light(3000 µmol m−2 s−1) of 650 nm peak wavelength obtained fromthree light-emitting diodes, focused on a circle of 5 mm diameterof the leaf sample. The first reliably measured point of the fluo-rescence transient is at 20 µs, which was taken as Fo. The variablefluorescence (Fv = Fm − Fo) and the maximum quantum yield ofPSII (Fv/Fm) was then estimated. Measurements were done in 10replications. A single leaf per plant constituted each replicate.

Investigation on growth, biomass allocation and yield parametersPeriodically, some of the potted mulberry plants were completelyharvested (both shoots and roots), while remaining plants wereharvested at the end of the experiment (75th day) to obtainperiodic as well as final growth and yield records. The numbersof branches were calculated and the length of all the branchesof a plant was added to obtain the total shoot length (TSL). Leafweight and stem weight of each plant were recorded as freshweight in grams and added to obtain above-ground biomass. Leafarea (cm2) was recorded using a LICOR leaf area meter (LI 1600)and instant leaf area duration (LAD, cm2 d−1) was computed.Root fresh weight (g) was recorded and root volume (cm3) wasmeasured following the water displacement method.16 Total plantbiomass, leaf mass ratio (LMR), stem mass ratio (SMR), root massratio (RMR) and root : shoot ratio were calculated using the dataof leaf, stem and root weights. The periodically as well as finallyharvested plant tissues were oven-dried at 70 ◦C for 72 h and,based on these data, relative growth rate (RGR), net assimilationrate (NAR) and biomass duration (BMD) were calculated.17

Analysing antioxidative protection, osmoprotectants and lipidperoxidationIn the third phase of study we narrowed down to only twomulberry cvs (one drought tolerant and one drought susceptible,screened out from the first two phases of our experiment).Carotenoids were extracted from fresh leaves (0.5 g) with 10 mLof 80% (v/v) acetone followed by centrifugation at 10 000 × g for5 min. The absorbance of cleared extract was read at 663.2, 646.8and 470 nm in a UV-visible 160A spectrophotometer (Shimadzu,Tokyo, Japan) and total carotenoid content was calculated.18 Forfoliar ascorbic acid(AA) determination, fresh leaf tissue (0.5 g)was homogenized with 5 mL of 10% trichloroacetic acid (TCA),centrifuged at 10 000 × g for 20 min at room temperature,re-extracted twice, supernatants were pooled and 0.5 mL ofthe extract was used to estimate AA.19 For measuring totalglutathione, the reaction mixture (1 mL) contained 100 µL sample,100 µL distilled water, 700 µL of 0.3 mmol L−1 NADPH in potassiumphosphate buffer (20 mmol L−1, pH 7.5) and 6 mmol L−1 5′-dithio-bis(2-nitrobenzoic acid) (DNTB). The rate of DNTB reduction wasrecorded by adding 10 µL glutathione reductase (GR) at 412 nmfor 3 min.20 For quantifying α-tocopherol,21 fresh leaf tissue


26

52


Figure 1. Effect of water deficit on photosynthetic gas exchange characteristics and leaf yield performance: (A) net photosynthetic rate (Pn); (B) stomatalconductance to CO2 (gs); (C) transpiration rate (E); (D) instant water use efficiency (WUEi); (E) internal CO2 concentration (Ci); (F) minimal value of internalto ambient CO2 ratio (Ci/Ca)min; (G) leaf temperature (Tl); (H) leaf yield/plant; (I) drought susceptibility index (DSI, based on leaf yield) in four mulberry cvsgrown under two water regimes (well-watered and water-stressed). All measurements and leaf harvests were conducted during two summer seasons ofthe study zone: season I (February–April 2009) and season II (May–July 2009). Data represented are the average over two summer seasons. Values aremean ± SD.

(1 g) was homogenized in liquid nitrogen with 10 mL of cold80% methanol, centrifuged at 3000 × g for 15 min at 4 ◦C. Thesupernatant was filtered through a 25 mm Millipore filter and thefiltrate was stored in ice under dark conditions. The α-tocopherolcontent of the filtrate was determined by high-performanceliquid chromatography (HPLC) with a C18 reverse-phase column(250 × 10.00 mm 5 µm, Phenomenex, Macclesfield, UK) at a flowrate of 1 mL min−1, using an isocratic solvent system – methanoland ethyl acetate (1 : 4, v/v) – as an eluant. α-Tocopherol levelwas quantified by a UV-visible detector system at 295 nm (LC-10AT VP, Shimadzu, Japan). Peak identification was performed bycomparing the retention times with pure α-tocopherol standard(Sigma, St Louis, MO, USA).

To determine the content of free proline, fresh leaf tissue(0.5 g) was homogenized in 10 mL of 3% sulfosalicyclic acid. Thehomogenate was centrifuged at 9000 × g for 15 min at roomtemperature and the supernatant was used to determine freeproline content.22 For glycine betaine (GB) quantification, freshleaf tissue was finely ground, mechanically shaken with 20 mLdeionized water for a period for 24 h at 25 ◦C. The samples werethen filtered and the collected filtrate was used for estimationof GB.23 The extent of lipid peroxidation in leaf tissues wasdetermined by measuring the malondialdehyde (MDA) content.Tissue samples (0.5 g) were homogenized in liquid nitrogen with5 mL of 0.1% TCA followed by centrifugation at 5000×g for 10 minat 4 ◦C. Thereafter, 500 µL of the extract was used for determinationof MDA equivalents.24

Statistical analysisResults were represented as mean ± standard deviation (SD).The significance of the differences between the treatments wasdetermined using paired t-tests. Correlation coefficient (r) andcoefficient of determinations (r2) of linear relationships betweenthe investigated parameters were established by using linearregression. The linear regression slopes were analysed usingbivariate correlation significance tests. Data were also analysed byanalysis of variance (ANOVA) to determine significant differences

between individual cv. and water stress treatments. All statisticalanalyses were performed and graphs drawn using the statisticalpackages Excel and Sigma Plot 11.0.

RESULTSField experimentsPhotosynthetic leaf gas exchange and leaf yield performanceEffect of drought stress on Pn, gs and E depicted contrastingvariability among the cvs and between the treatments. Significantgenotypic variation was recorded for Pn among the fourmulberry cvs in both water regimes (Fig. 1(A)). Under well-wateredconditions, cv. AR-12 exhibited the highest Pn (13.2 µmol m−2 s−1),followed by V-1 (12.58 µmol m−2 s−1), K-2 (11.45 µmol m−2 s−1)and M. Local (9.03 µmol m−2 s−1). Water deficit significantly(P < 0.05) reduced Pn in all the mulberry cvs. Reduction in Pn wasstrong in cv. AR-12 (70.7%), followed by K-2 (64%) and M. Local(61.4%) under drought stress. However, higher Pn under waterdeficit was recorded in V-1 (7.9 µmol m−2 s−1) with a minimumreduction of 37.2% compared to well-watered counterparts. Thegs ranged from 0.46 to 0.26 mol m−2 s−1 and from 0.19 to0.05 mol m−2 s−1 in control and stress treatments, respectively(Fig. 1(B)). Under well-watered conditions, V-1 (0.46 mol m−2 s−1)exhibited maximum gs, followed by K-2 (0.40 mol m−2 s−1) andAR-12 (0.38 mol m−2 s−1), whereas gs was lower in M. Local(0.26 mol m−2 s−1). Soil water deprivation caused a dramaticdownregulation in the rates of gs for all mulberry cvs. Thelowest value for gs under low water regime was recorded inAR-12 (0.05 mol m−2 s−1), followed by K-2 (0.06 mol m−2 s−1) andM. Local (0.07 mol m−2 s−1). However, cv. V-1 exhibited maximumgs (0.19 mol m−2 s−1) under low soil moisture. Drought causedsimilar drastic reduction of E (∼70%) associated with lack ofirrigation in all the cvs (Fig. 1(C)). V-1 maintained relatively higherrates for E under low water regimes. Relative to unstressed plants,drought led to a significant (P < 0.01) increase in WUEi in mulberrycvs V-1 (26.6%) and M. Local (36.6%) (Fig. 1(D)). In K-2 there was amarginal increment of 9.7% in WUEi under water stress, whereasno significant change in WUEi of drought-exposed AR-12 was


26

53


Figure 2. Relationships between: (A) stomatal conductance to CO2 (gs) versus net photosynthetic rate (Pn); (B) transpiration rate (E) versus instant wateruse efficiency (WUEi); (C) internal CO2 concentration (Ci) versus net photosynthetic rate (Pn); (D) stomatal conductance to CO2 (gs) versus internal CO2concentration (Ci). Linear regressions were fitted to data for four mulberry cvs grown under well-watered and water-stressed conditions during February2009 to July 2009. Each point represents the mean of five independent measurements from an individual plant. The correlations were significant atP < 0.001.

witnessed when compared to control plants. Ci of well-watered cvsranged from 257 to 267 µmol mol−1; however, drought resultedin an apparent reduction of Ci in all the mulberry cvs, although toa different extent depending upon the genotype (Fig. 1(E)). Thereduction in Ci was significantly (P < 0.05) sharper in K-2 (39.4%)and M. Local (36.3%) compared to V-1 (11.4%) during the droughtstress period. The lowest values recorded for Ci/Ca[(Ci/Ca)min]under dry-land conditions were 0.23, 0.29, 0.33 and 0.47 for K-2,AR-12, M. Local and V-1, respectively (Fig. 1(F)). Water stress causedsignificant (P < 0.01) elevation in Tl for the drought exposedmulberry cvs (Fig. 1(G)). The highest Tl of stressed mulberry leaveswas observed in AR-12 (36.5 ◦C) followed by M. Local (35.9 ◦C)and K-2 (35.4 ◦C), while V-1 maintained the lowest value for Tl

(∼32.6 ◦C) throughout the periods of water deprivation.The assessment of leaf yield production depicted wide genetic

variability between both the cv. and treatment levels (Fig. 1(H)).Under well-irrigated conditions, the highest leaf yield per plantwas recorded in V-1 (1190 g per plant) followed by AR-12 (698 gper plant) and M. Local (624 g per plant). A significant (P < 0.01)decline in leaf yield was recorded in all the cvs under water-stressedtreatments. However, maximum leaf yield per plant was obtainedin V-1 (450 g per plant) under drought conditions exhibiting a DSIof 0.85, whereas minimum leaf yield was recorded in AR-12 (166 gper plant, DSI 1.12) followed by K-2 (168 g per plant, DSI 1.02) andM. Local (180 g per plant, DSI 1.05) (Fig. 1(I)).

Relationships among photosynthetic gas exchange variablesThe drought-induced regulation of Pn by gs followed a linearfunction, resulting in a significant positive correlation between Pn

and gs in all the mulberry cvs (Fig. 2(A)). The regression slopeswere steeper for the data from K-2 (r2 = 0.93, P < 0.001) andM. Local (r2 = 0.81, P < 0.001). For V-1, the correlation betweenPn and gs was comparatively weak though significant (r2 = 0.64,P < 0.001). Significant negative correlations were evident when

Table 2. Leaf moisture content and leaf relative water content infour mulberry cvs grown in a glasshouse under two water regimes(well-watered and water-stressed). Values are mean ± SD

Leaf moisturecontent (%)

Leaf relativewater content (%)

Cultivars Well-watered Water-stressedWell-

watered Water-stressed

AR-12 74.3 ± 1.2 66.3 ± 0.7 82.5 ± 2.0 67.4 ± 1.7

K-2 72.5 ± 1.1 67.5 ± 0.9 82.1 ± 3.8 68.5 ± 0.9

M. Local 70.2 ± 1.4 66.5 ± 0.7 80.2 ± 1.9 68.6 ± 2.6

V-1 76.1 ± 0.8 72.6 ± 0.8 86.8 ± 2.7 74.4 ± 1.9

data for E and WUEi were plotted together for all the mulberrycvs (Fig. 2(B)). The correlations were stronger for V-1 (r2 = −0.97,P < 0.001) and M. Local (r2 = −0.82, P < 0.001); however, theywere weak for K-2 (r2 = −0.26 P < 0.001) and AR-12 (r2 = −0.41,P < 0.001). The relationships between Pn versus Ci and Ci versusgs were linear and positively correlated (Fig. 2(C, D)). However,the stomatal closure induced a prompt and higher inhibition ofPn compared to Ci. In fact, for a reduction of 0.28 mol m−2 s−1 ings, Pn decreased 58.3%, while Ci decreased 30%. The decline inCi seemed to depend linearly on gs and was strongly correlatedfor all four mulberry cvs. The slope of regression for Ci versus gs

curve was steeper for AR-12 (r2 = 0.94, P < 0.001), followed byK-2 (r2 = 0.89, P < 0.001), M. Local (r2 = 0.86, P < 0.001) and V-1(r2 = 0.84, P < 0.001).

Glasshouse experimentsLeaf water statusEffect of drought stress on LMC and leaf RWC of mulberry cvs arepresented in Table 2. For all the cvs LMC decreased under water-


26

54


Table 3. Changes in growth and yield characteristics in four mulberry cvs subjected to drought stress under glasshouse conditions

AR-12 K-2 M. Local V-1

VariablesWell-

watered Water-stressedWell-



watered Water-stressed

Plant height (cm) 76.6 36.2∗∗∗ 84.2 31.1∗∗∗ 95.3 37.3∗∗∗ 97.4 46.3∗∗∗

TSL (cm) 173.1 77.2∗∗∗ 211.3 76.2∗∗∗ 188.4 71.3∗∗∗ 196.5 113.2∗∗

No. branches per plant 6.2 3.7∗∗ 7.2 3.2∗∗∗ 7.0 3.2∗∗∗ 7.5 4.5∗

Total leaf area (cm2) 2894.2 908.6∗∗∗ 2782.5 924.4∗∗∗ 2112.3 921.7∗∗∗ 3041.6 1274.6∗∗∗

Leaf weight (g) 156.2 18.4∗∗∗ 125.3 21.6∗∗∗ 148.4 24.3∗∗∗ 191.7 46.5∗∗∗

LAD (cm2 d−1) 41.5 13.2∗∗∗ 40.9 13.4∗∗∗ 31.5 13.5∗∗∗ 44.6 20.7∗∗

Stem weight (g) 118.1 24.3∗∗∗ 128.2 34.9∗∗∗ 136.6 46.5∗∗∗ 241.7 42.7∗∗∗

Root weight (g) 142.7 58.8∗∗∗ 144.8 66.2∗∗∗ 166.4 76.2∗∗∗ 171.1 108.3∗∗∗

Root length (cm) 87.4 67.8∗∗ 86.5 65.6∗∗∗ 108.1 78.4∗∗∗ 108.6 88.9∗∗

Root volume (cm3) 55.5 30.5∗∗ 58.7 35.5∗∗ 67.2 38.8∗∗∗ 98.4 65.3∗∗∗

Root : shoot (g g−1) 0.52 1.38∗∗∗ 0.57 1.17∗∗∗ 0.58 1.07∗∗∗ 0.40 1.21∗∗∗

Above ground biomass (g) 274.3 42.7∗∗∗ 253.5 56.5∗∗∗ 285.3 70.8∗∗∗ 433.4 89.2∗∗∗

Total plant biomass (g) 417.4 101.5∗∗∗ 398.3 122.5∗∗∗ 451.4 147.1∗∗∗ 604.5 197.5∗∗∗

RGR (g g−1 d−1) 0.31 0.24∗ 0.32 0.22∗∗ 0.30 0.26∗ 0.36 0.31 n

NAR (g m−2 d−1) 1.6 0.28∗∗∗ 1.5 0.33∗∗∗ 1.7 0.35∗∗∗ 1.8 0.77∗∗∗

BMD (g d−1) 6.1 1.4∗∗∗ 5.9 1.8∗∗∗ 6.6 2.2∗∗ 8.9 3.1∗∗

Effects of drought were tested by paired t-test. ∗∗∗ P < 0.001; ∗∗ P < 0.01; ∗ P < 0.05; n.s., not significant.

stressed conditions, but for V-1 this decrease was minimal andthe cv. maintained a comparatively higher LMC of 72.6%. Underlow water regimes, the RWC was reduced significantly (P < 0.05)in all four mulberry cvs compared to control (Table 2). However,V-1 showed tolerance to drought with the highest RWC of 74.4%,whereas the decrease in RWC was higher in AR-12 and dropped to67% during periods of water deprivation.

PRUE and chlorophyll a fluorescenceSignificant genotypic variation was encountered for LSPn andPRUE among the four mulberry cvs in both water regimes(Fig. 3(A)). LSPn showed essentially the same variations as PRUE

(hence the LSPn values were not shown), since the PAR wasfixed and due to the high correlation between LSPn and PRUE(r2 = 0.96, P < 0.001). Under well-irrigated conditions, PRUEranged from 8.8 to 6.2 (mmol (CO2) mol−1 (photons)). AR-12 hadthe highest PRUE, followed by V-1 (8.8 and 8.3 mmol (CO2) mol−1

(photons), respectively) under normal physiological conditions,while PRUE values were comparatively low in the control plantsof M. Local (6.2 mmol (CO2) mol−1 (photons)). Drought imposedon the mulberry cvs decreased PRUE significantly (P < 0.05) andthe values ranged from 5.2 to 2.3 (mmol (CO2) mol−1 (photons))under low water regimes. The maximum drop in PRUE was evidentin stressed AR-12 plants (71.5%), followed by K-2 (64.4%) and

Figure 3. Effect of water stress on photosynthetic radiation use efficiency (PRUE) and chlorophyll a fluorescence characteristics: (A) PRUE; (B) percentagereduction in PRUE; (C) ground fluorescence (Fo); (D) maximal fluorescence (Fm); (E) maximal quantum yield of PSII (Fv/Fm) in four mulberry cvs grown in aglasshouse under two water regimes (well-watered and water-stressed). Values are mean ± SD.


26

55


Figure 4. Biomass allocation patterns in leaf (leaf mass ratio, LMR) stem(stem mass ratio, SMR) and root (root mass ratio, RMR) fractions of fourmulberry cvs grown in a glasshouse under well-watered (left half of piediagrams) and water-stressed (right half of the pie diagrams) conditions.Data on fresh weights of leaf, stem and root fractions from Table 3 havebeen taken to depict the pie diagrams.

M. Local (61.6%) (Fig. 3(B)). However, the highest PRUE value wasobserved in V-1 (5.2 mmol (CO2) mol−1 (photons)) during waterdeficit conditions. V-1 also executed a minimum reduction of37.3% in PRUE compared to the well-watered counterparts.

Chlorophyll a fluorescence parameters were significantly af-fected by drought stress. In all the tested mulberry cvs, droughtstress induced an increase in Fo which was significantly more inK-2, AR-12 and M. Local when compared to V-1 (Fig. 3(C)). Unlike Fo,a reduction in Fm was encountered in the stressed leaves of mul-berry cvs compared to the well-watered counterparts (Fig. 3(D)).Depression in Fm was more discrete in AR-12 (12%), whereas thereduction was minimal in the stressed leaves of V-1 (0.8%). TheFv/Fm ratio in control leaves of four mulberry cvs remained above0.8 throughout the experiment. The ratio was also practically notaltered in V-1 (0.81) under low water regimes. However, in K-2 theratio dropped significantly (0.70, P < 0.05) in response to droughtwhen compared to well-watered plants (Fig. 3(E)).

Growth, allocation trade-offs and yield performanceSubstantial alterations in growth, allocation and productivity traitswere detected in response to drought in four mulberry cvs(Table 3). The main effect of water stress was a significant reductionin all shoot criteria (plant height, TSL, number of branches,stem weight, leaf weight, cumulative leaf area and above-groundbiomass) as assessed at the end of experimental period. AR-12, K-2and M. Local cvs suffered greater reductions in all shoot growthcriteria than V-1. The interaction treatment × cv. was significantfor all these parameters (Table 4). The four mulberry cvs also variedsignificantly in respect to different root characteristics (Table 3).Discrete reduction in root FW was observed in mulberry cvs(∼15.5%) under water stress treatments; however, V-1 exhibitedhighest root FW compared to other cvs. Vertical proliferation ofroots was retarded, in all the mulberry cvs under low water regimes,

Table 4. Significance of the influence of treatment, cultivar andproduct of treatment × cultivar on the variance of 10 growthparameters measured under glasshouse conditions according tomultivariate analyses of variance

Growth parameter Treatment Cultivar Treatment × cultivar

TSL ∗∗∗ ∗∗ ∗∗∗

Total leaf area ∗∗∗ ∗∗∗ ∗∗∗

Leaf weight ∗∗∗ ∗∗∗ ∗∗∗

Stem weight ∗∗∗ ∗∗ ∗∗∗

Above-ground biomass ∗∗∗ ∗∗ ∗∗∗

Root length ∗∗ ∗∗ ∗∗

Root weight ∗∗∗ ∗∗ ∗∗∗

Root volume ∗∗∗ ∗∗ ∗∗∗

Root : shoot mass ratio ∗∗∗ ∗ ∗∗∗

Total plant biomass ∗∗∗ ∗∗ ∗∗∗

∗∗∗ P < 0.001; ∗∗ P < 0.01; ∗ P < 0.05.

which was reflected in reduction of root length in the mulberry cvsof either group. However, the root length in V-1 was relatively largerirrespective of the watering regimes (108.6 cm and 88.9 cm incontrol and water stress treatments, respectively). The root volumewas reduced to 45% and 42.2% in AR-12 and M. Local, respectively,in response to water stress; however, the reduction was minimumin drought-exposed V-1 (31%). The interaction treatment × cvwas significant for the majority of root characteristics (Table 4).Drought stress also led to an apparent reduction in total plantbiomass accumulation in all the mulberry cvs (Table 3). However,V-1 maintained comparatively higher plant biomass under water-limited conditions when compared to others.

The drought-treated plants showed a larger reduction in shootgrowth compared to root, indicating a shift in biomass allocationtowards below-ground organs. However, the four mulberry cvsdiffered in their response to biomass allocation. As indicated inFig. 4, under well-watered conditions the biomass allocations toleaf and root tissues in the mulberry cvs were approximately33.5%, and 33.7%, respectively. A substantial shift to root biomassallocation (RMR ∼0.56 g g−1) was recorded in the water-stressedcvs compared to control counterparts (RMR ∼0.33 g g−1), whichconcurrently increased the root : shoot mass ratios in all testedmulberry cvs (Table 3). A discrete depression in leaf biomassallocation was recorded in AR-12 (LMR, 0.18 g g−1), K-2 (LMR,0.17 g g−1) and M. Local (LMR, 0.16 g g−1) cvs, while drought-exposed V-1 exhibited significantly more allocation to leaf biomass(LMR, 0.24 g g−1) than the remaining cvs. V-1 had a greaterallocation to stem (SMR, 0.39 g g−1) compared to AR-12 (SMR,0.28 g g−1), M. Local (SMR, 0.31 g g−1) and K-2 (SMR, 0.32 g g−1)under well-irrigated conditions (Fig. 4(D)). However, allocation tostem was maximal in M. Local (SMR, 0.32 g g−1) and minimal in V-1(SMR, 0.22 g g−1) under water-limited conditions (Fig. 4(B,D)).

LAD ranged from 31.5 to 44.6 cm2 d−1 and from 13.2 to 20.7 cm2

d−1 under control and drought treatments, respectively (Table 3).Deprivation in soil moisture led to a severe reduction in LAD inall the mulberry cvs; however, V-1 exhibited the highest value forLAD during water stress periods. Under well-watered conditions,the cvs did not differ much in respect to RGR and NAR; however,drought stress led to a significant drawdown in RGR and NAR in allthe tested cvs (Table 3). V-1 exhibited relatively higher values forRGR (0.31 g g−1 d−1) and NAR (0.77 g m−2 d−1) under low waterregimes, whereas RGR and NAR were found to be lowest in K-2 and


26

56


AR-12, respectively (Table 3). Regardless of the watering regimes,V-1 maintained higher values for BMD, while drought caused adramatic reduction of BMD in other mulberry cvs with an averagedecline of 69.2% (Table 3).

Significant positive correlations were achieved when data forPRUE and leaf weight were plotted together for all the cvs(Fig. 5(A)). The regression slopes were steeper and the degreeof significance was high for all the mulberry cvs (AR-12, r2 = 0.96,P < 0.001; K-2, r2 = 0.98 P < 0.001; M. Local, r2 = 0.97, P < 0.001;V-1, r2 = 0.98, P < 0.001). The response of PRUE to total plantbiomass accumulation was also strongly significant for all mulberrycvs (M. Local (r2 = 0.98, P < 0.001), V-1 (r2 = 0.95, P < 0.001),AR-12 (r2 = 0.93, P < 0.001) and K-2 (r2 = 0.92, P < 0.001)](Fig. 5(B)).

Antioxidants and osmoprotectantsTotal carotenoid content was not significantly different in theleaves of well-watered cvs. However, an increasing trend in totalcarotenoid accumulation was recorded in both the cvs underdrought stress (Fig. 6(A)). V-1 showed a maximum level of totalcarotenoids (0.69 mg g−1 FW) compared to K-2 (0.55 mg g−1 FW)under drought conditions. The foliar AA concentration of well-irrigated plants was not significantly different between the twotested cvs (AA ∼1.05 mg g−1 FW). However, an increase in AAlevel was witnessed in the leaves of stressed V-1 plants exhibitingsignificant elevation (about 36% increase, P < 0.05) comparedto the minor increment observed in K-2 (only 7.5% increase)(Fig. 6(B)). α-Tocopherol content showed contrasting variation

Figure 5. Relationships between photosynthetic radiation use efficiency(PRUE) versus (A) leaf weight and (B) total plant biomass. Linear regressionswere fitted to data for four mulberry cvs grown in a glasshouse underwell-watered and water-stressed conditions. The solid lines are the linearregressions fitted to all points for each cv. and are significant for P ≤ 0.001.

Figure 6. Effect of water stress on the contents of (A) total carotenoids, (B) ascorbic acid (AA), (C) α-tocopherol, (D) glutathione, (E) proline, (F) glycinebetaine and (G) foliar malondialdehyde (MDA) in two mulberry cvs (V-1 and K-2) grown under glasshouse conditions. Values are mean ± SD.


26

57


among the two mulberry cvs (Fig. 6(C)). V-1 exhibited a significantly(P < 0.05) elevated level of α-tocopherol under drought stresswith respect to control, whereas the endogenous α-tocopherollevel of K-2 was affected severely and a net α-tocopherol losswas visible in stressed leaves of K-2. V-1 had endogenouslyhigher concentration of glutathione compared to K-2 (5.9 and3.7 µmol g−1 FW, respectively). Under water deficit conditions,V-1 exhibited more glutathione accumulation (7 µmol g−1 FW)than K-2 (4 µmol g−1 FW) (Fig. 6(D)).

Differential levels of endogenous proline content were moni-tored in the two mulberry cvs in both well-watered and water-stressed plants (Fig. 6(E)). V-1 showed inherently higher prolinecontent compared to K-2 during unstressed conditions and alsoaccumulated significantly (threefold increase, P < 0.05) under lowwater regimes. Similar to proline, concentration of GB increasedin the water-stressed leaves of both mulberry cvs (Fig. 6(F)). How-ever, increment in the level of GB was substantial in V-1 (52%,P < 0.05) when compared to K-2 (31%, P < 0.05). Water defi-ciency augmented the MDA content in leaf tissues of the twotested mulberry cvs. An apparent increase in leaf MDA contentoccurred in stressed K-2 (46.2% increase, P < 0.05) comparedto control, while MDA accumulation was considerably less in thestressed leaf tissues of V-1. An increase of 29% (P < 0.05) inMDA content was evident in drought-exposed V-1 leaves whencompared to control (Fig. 6(G)).

DISCUSSIONIn the present study, the most important plastic response ofdrought stress in mulberry was the reduction in leaf yield perplant and associated traits, which included cumulative leaf area,leaf biomass, number of branches and LAD. However, suchhypersensitive, phenotypic plasticity causing diminution of foliagecannot be linked to enhanced performance under drought as leafyield is equally important besides crop sustenance in mulberry.Anomalous to drought-susceptible mulberry cvs (AR-12, M. Localand K-2) that underwent severe loss in leaf yield (DSI >1),V-1 showed drought tolerance (DSI <1), maintaining relativelyhigher and stable leaf yield during drought-exposed conditions.In agreement with the stress tolerator syndrome, not only higherbiomass allocation to yield organ (LMR) but also a minimal plasticityin foliar gas exchange physiology conferred enhanced genotypicperformance to V-1 under water-limited conditions. Focusing on‘leaf yield harvest per plant’, we analysed the evidence of ‘yieldlosses’ in the susceptible mulberry cvs compared to tolerant V-1.Limitation in leaf yield due to drought stress can be explained as aconsequence of significant decline in the rates of photosynthesis(Pn), stomatal conductance to CO2 (gs) and transpiration rates (E)in the tested mulberry cvs. As Pn decreased in parallel with gs,stomatal inhibition to CO2 conductance seem to be the principalcause of photosynthetic downregulation in mulberry leaves underwater stress. The reductions in Pn and gs were accompaniedby acute depressions in Ci during drought stress, supportingstrong stomatal inhibition. Stomatal closure is a drought avoidancestrategy that allows leaf moisture restoration, but a concomitantdrawdown in Ci can dramatically downregulate Pn.25 – 27 There maybe uncertainties in calculating Ci in water-stressed, heteroboricmulberry leaves due to stomatal patchiness.28,29 As we did notmeasure patchy stomatal conductance,30 – 32 we cannot be surethat our Ci values were not subjected to the patchiness problems.However, it should be noted that: (i) patchiness should notbe an important problem under progressively drought-exposed

conditions in the field; and (ii) gs values measured during thewater-stressed periods for two consecutive summer monthswere always higher than 0.03 mol m−2 s−1, below which thepatchiness phenomenon can be important.32,33 The explicit co-relationships between the photosynthetic gas exchange traitsprovided a vivid understanding of the functional interrelationsamongst those traits. The positive linear relationships betweenPn –gs, Pn –Ci and Ci –gs, as evident within the tested mulberrycvs, prominently indicate the regulatory functions of gs and Ci

on Pn in the dehydrated leaves of mulberry. Elevated Tl (>35 ◦C),as recorded in control and stressed leaves of drought-susceptiblemulberry cvs might also adversely affect mesophyll conductanceto CO2 and cause damage to the photosynthetic machinery.34

Interestingly, the drought-tolerant cv. V-1 was able to maintainlow Tl and therefore heat-induced damage on the photosyntheticapparatus was also presumably less critical for the same cv. Thenegative relationship between E and WUEi of the mulberry cvswas in agreement with findings in other tree species such asEuropean chestnut35 and western red cedar.36 However, in spiteof sufficient reduction in E, WUEi was not found to be significantlyimproved in the susceptible mulberry cvs, owing to their inabilityto maintain higher Pn under water stress conditions, whereas inthe tolerant V-1 Pn was found to be higher and less sensitiveto E; hence WUEi was consequently much improved in thesame cv.

The leaf-level efficiency of carboxylation, commonly termed‘photosynthetic efficiency’, is the central process for plant perfor-mance and can be determined in terms of instantaneous PRUE,which characterizes net CO2 fixation per incoming PAR on the leafsurface.37 Moreover, the efficiency with which absorbed photonsare finally used for photosynthetic electron transport and carbonfixation can be determined using chlorophyll a fluorescence tech-niques and denoted as maximum PSII photochemical efficiency(Fv/Fm).38 All three processes (photochemical efficiency of PSII,leaf-level PRUE and Pn) are ultimately light-driven and determinethe carbon gain and biomass acquisition in plants. In the presentstudy, water constraint directly affected Fv/Fm, which presum-ably contributed to jeopardizing Pn and PRUE in the susceptiblemulberry cvs compared to tolerant V-1. The light-harvesting com-plexes reduced their efficiency due to drought-induced damageto the antennae, as evidenced by an increase in Fo in drought-exposed leaves of susceptible cvs; however, in the stressed leavesof drought-tolerant V-1 the increase in Fo was relatively less andwas balanced by the maintenance of Fm, which contributed tokeep similar Fv/Fm. The decrease in Fm in the susceptible mulberrycvs may be related to the decrease in activity of the water-splittingenzyme complex and perhaps a concomitant cyclic electron trans-port with or around PSII.39 The decrease in Fv/Fm in the susceptiblecvs suggests a chronic photoinhibition due to photoinactivationof PSII centres, possibly attributable to D1 protein damage.40

The strong, positive influence of PRUE on leaf weight and totalplant biomass (Fig. 5) suggests the importance of unimpairedphotochemical efficiency for photosynthetic yield acquisition inmulberry cvs under well-watered as well as water deficit situations.

The susceptible mulberry cvs revealed significant differencesin their responses to drought with respect to leaf and rootmorphology, whole-plant productivity and allocation whencompared to tolerant V-1. Mulberry cv. V-1 not only showedthe maximum LMR and cumulative leaf area in both the droughtand control treatments, but also excelled in growth rates (NAR,RGR) and BMD (Table 3). A comparatively well-developed rootsystem as evident in V-1 might be helpful in better hydraulic


26

58


conductance of the root system during drought and subsequentlycould maintain higher rates of E and gs, displaying highly efficientwater use strategy.41 It is a widespread opinion that plantsadapted to dry habitats reduce their foliage and thus theirwater loss as a conservative water use strategy. However, stableoptimization strategies can confer the ability to consume largeamounts of resources due to high explorative ability and thusallow stabilization in leaf morphophysiological characteristics, aswitnessed in drought-tolerant cv. V-1.42

Better physiological functioning and growth performance incrops under water stress conditions have been connected tothe enhanced activity of antioxidants for scavenging reactiveoxygen species (ROS).43 Our results indicate a paramount increasein the concentrations of non-enzymatic antioxidants in V-1compared to the susceptible K-2, exhibiting elevated antioxidativeprotection under water-deprived situations. AA was the majorwater-soluble antioxidant in V-1 leaves and the enhancedlevels of AA, glutathione and α-tocopherol in water-stressedV-1 elucidate the well-maintained redox buffering capacity ofthe acorbate–glutathione cycle.44 An apparent increase in totalcarotenoids and α-tocopherol content in V-1 indicates efficientscavenging of highly destructive singlet oxygen (1O2) and lipidperoxidation products, thereby stabilizing the photosyntheticmembranes of chloroplasts.45 It is interesting to note that thedrought-susceptible K-2 showed endogenous loss in α-tocopherolwhen exposed to drought which might occur due to irreversibledegradation of α-tocopherol as a limited AA availability inchloroplast of the same cv.46 The efficient antioxidative protectionin stressed V-1 facilitated low MDA accumulation, whereas in K-2increased levels of MDA accumulated as an aftermath of poorantioxidative defence response, making the cv succumb more tooxidative damage. A positive relationship between accumulationof compatible solutes (proline, GB) and drought tolerance was alsoobserved in V-1, which could provide better osmotic equilibriumand cell membrane stability during water stress regime.

CONCLUSIONFrom our present investigation, the following conclusions canbe ascertained: (i) in agreement with stress tolerance syndrome,not only an increased biomass allocation to yield organ butalso a minimal plasticity in foliar gas exchange characteristicsconferred enhanced leaf yield performance to cv. V-1 underwater-limited conditions; (ii) less photoinhibition and stabilizedphotochemistry facilitated higher PRUE in tolerant cv. V-1, showingenhanced ‘photosynthetic yield’ under drought; (iii) the effectiveosmoregulation and increased activities of antioxidants also com-plemented drought tolerance in cv. V-1. Net photoassimilationrates and carbon allocation are the culmination of all thesebiological responses and chemical processes, which were rel-atively less affected in the drought-tolerant mulberry cv. V-1,resulting in better yield performance under a water stress envi-ronment.

ACKNOWLEDGEMENTSWe are grateful to Regional Sericultural Research Station, Salem,India, for providing the mulberry germplasm. We acknowledgethe Department of Science and Technology (DST), Government ofIndia, for financial assistance (Grant SR/SO/PS-27/05). AG gratefullyacknowledges DST for a research fellowship.

REFERENCES1 Vijayan K and Chatterjee SN, ISSR profiling of Indian cultivars of

mulberry (Morus spp.) and its relevance to breeding programs.Euphytica 131:53–63 (2003).

2 Papanastasis VP, Yiakoulaki MD, Decandia M and Dini-Papanastasi O,Integrating woody species into livestock feeding in theMediterranean areas of Europe. Animal Feed Sci Technol 140:1–17(2008).

3 Machii H, Koyama A and Yamanouchi H, Mulberry breeding,cultivation and utilization in Japan, in MulberryforAnimalProduction:FAO Electronic Conference, Feed Resources Group (AGAP), FAO,Rome, 1 May–31 June (2000).

4 Dandin SB, Jayaswal J and Giridhar K, Mulberry cultivation, inHandbook of Sericulture Technologies, ed. by Dandin SB, Jayaswal Jand Giridhar K. Central Silk Board, Bangalore, pp. 31–45 (2003).

5 Karaba NN, Sheshshayee MS and Udaykumar M, Biotech News: Optionsfor Improvement – Mulberry, Vol. III(5) (2008).

6 Chaves MM, Flexas J and Pinheiro C, Photosynthesis under droughtand salt stress: regulation mechanisms from whole plant to cell.Ann Bot 103:551–560 (2009).

7 Murchie EH, Pinto M and Horton P, Agriculture and the new challengesfor photosynthesis research. New Phytol 181:532–552 (2009).

8 Susheelamma BN, Jolly MS, Giridhar K and Sengupta K, Evaluation ofgermplasm genotypes for the drought resistance in mulberry.Sericologia 30:327–340 (1990).

9 Ramanjulu S, Giridara NKS and Sudhakar C, Photosyntheticcharacteristics in mulberry during water stress and rewatering.Photosynthetica 35:259–263 (1998).

10 Chaitanya KV, Jutur PP, Sundar D and Reddy AR, Water stress effectson photosynthesis in different mulberry cultivars. Plant GrowthRegul 40:75–80 (2003).

11 Karatassiou M, Parissi ZM, Abraham EM and Kyriazopoulos AP, Growthof Morus alba L. under water deficit conditions. OptionsMediterraneennes Series A 70:315–318 (2008).

12 Chaturvedi HK and Sarkar A, Optimum size and shape of the plot formulberry experiments. Indian J Seric 39:66–69 (2000).

13 Fisher RA and Maurer R, Drought resistance in spring wheat cultivars.I. Grain yield response. Aust J Agric Res 29:897–912 (1978).

14 Castillo FJ, Antioxidative protection in the inducible CAM plant Sedumalbum L. following the imposition of severe water stress andrecovery. Oecologia 107:469–477 (1996).

15 Ma CC, Gao YB, Guo HY and Wang JL, Photosynthesis, transpiration,and water use efficiency of Caragana microphylla, C. intermedia, andC. korshinskii. Photosynthetica 42:65–70 (2004).

16 Burdett AN, A nondestructive method for measuring the volume ofintact plant parts. Can J For Res 9:120–122 (1979).

17 Lamers John PA, Khamzina A and Worbes M, The analyses ofphysiological and morphological attributes of 10 tree speciesfor early determination of their suitability to afforest degradedlandscapes in the Aral Sea Basin of Uzbekistan. For Ecol Manage221:249–259 (2006).

18 Lichtenthaler HK, Chlorophylls and carotenoids: pigments ofphotosynthetic biomembranes. Methods Enzymol 148:350–382(1987).

19 Omaye ST, Turnbull JD and Sauberilich HE, Selected methods for thedetermination of ascorbic acid in animal cells, tissues and fluids.Methods Enzymol 62:3–11 (1979).

20 Griffith OW and Meister A, Potent and specific inhibition of glutathionesynthesis by buthionine sulfoximine (s-n-butylhomocysteinesulfoximine). J Biol Chem 254:7558–7560 (1979).

21 Yen GC, Wu SC and Duh PD, Extraction and identification ofantioxidant components from the leaves of mulberry (Morus albaL.). J Agric Food Chem 44:1687–1690 (1996).

22 Bates LS, Walderen RP and Teare ID, Rapid determination of freeproline for water stress studies. Plant Soil 39:205–207 (1973).

23 Grieve CM and Grattam SR, Rapid assay for determination of watersoluble quaternary ammonium compounds. Plant Soil 70:303–307(1983).

24 Fu J and Huang B, Involvement of antioxidants and lipid peroxidationin the adaptation of two cool-season grasses to localized droughtstress. Environ Exp Bot 45:105–114 (2001).

25 Cornic G, Drought stress inhibits photosynthesis by decreasingstomatal aperture – not by affecting ATP synthesis. Trends PlantSci 5:187–188 (2000).


26

59


26 Flexas J and Medrano H, Drought inhibition of photosynthesis in C3plants: stomatal and non stomatal limitations revisited. Ann Bot89:183–189 (2002).

27 Medrano H, Escalona JM, Bota J, Gulıas J and Flexas J, Regulationof photosynthesis of C3 plants in response to progressivedrought: stomatal conductance as a reference parameter. AnnBot 89:895–905 (2002).

28 Nikolopoulos D, Liakopoulos G, Drossopoulos I and Karabourniotis G,The relationship between anatomy and photosyntheticperformance of heteroboric leaves. Plant Physiol 129:235–243(2002).

29 Gomes FP, Oliva MA, Mielke MS, Almeida A-AF de, Leite HG andAquino LA, Photosynthetic limitations in leaves of young BrazilianGreen Dwarf coconut (Cocos nucifera L. ‘nana’) palm under well-watered conditions or recovering from drought stress. Environ ExpBot 62:195–204 (2008).

30 Terashima I, Wong SC, Osmond CB and Farquhar GD, Characterisationof non-uniform photosynthesis induced by abscisic acid in leaveshaving different mesophyll anatomies. Plant CellPhysiol 29:385–394(1988).

31 Mott KA and Buckley TN, Patchy stomatal conductance: emergentcollective behaviour of stomata. Trends Plant Sci 5:258–262 (2000).

32 Flexas J, Bota J, Escalona JM, Sampol B and Medrano H, Effects ofdrought on photosynthesis in grapevines under field conditions: anevaluation of stomatal and mesophyll limitations. Funct Plant Biol29:461–471 (2002).

33 Grassi G and Magnani F, Stomatal, mesophyll conductance andbiochemical limitations to photosynthesis as affected by droughtand leaf ontogeny in ash and oak trees. Plant Cell Environ28:834–849 (2005).

34 Bernacchi CJ, Portis AR, Caemmerer HS von and Long SP, Temperatureresponse of mesophyll conductance: implications for thedetermination of Rubisco enzyme kinetics and for limitations tophotosynthesis in vivo. Plant Physiol 130:1992–1998 (2002).

35 Lauteri M, Pliura A, Monteverdi MC, Brugnoli E, Villani F and Eriksson G,Genetic variation in carbon isotope discrimination in six European

populations of Castanea sativa Mill. originating from contrastinglocalities. J Evol Biol 17:1286–1296 (2004).

36 Fan S, Grossnickle SC and Russell JH, Morphological and physiologicalvariation in western redcedar (Thuja plicata) populations undercontrasting soil water conditions. Trees 22:671–683 (2008).

37 Winkel T, Methy M and Thenot F, Radiation use efficiency,chlorophyll fluorescence, and reflectance indices associated withontogenic changes in water-limited Chenopodium quinoa leaves.Photosynthetica 40:227–232 (2002).

38 Baker NR, Chlorophyll fluorescence: a probe of photosynthesis in vivo.Annu Rev Plant Biol 59:89–113 (2008).

39 Zlatev ZS and Yordanov IT, Effects of soil drought on photosynthesisand chlorophyll fluorescence in bean plants. Bul J Plant Physiol30:3–18 (2004).

40 Ohnishi N, Allakhverdiev SI, Takahashi S, Higashi S, Watanabe M,Nishiyama Y, et al, Two-step mechanism of photo damage tophotosystem II: step 1 occurs at the oxygen-evolving complex andstep 2 occurs at the photochemical reaction centre. Biochemistry44:8494–8499 (2005).

41 Susiluoto S and Berninger F, Interactions between morphological andphysiological drought responses in Eucalyptus microtheca. SilvaFenn 41:221–233 (2007).

42 Aspelmeier S and Leuschner C, Genotypic variation in droughtresponse of silver birch (Betula pendula Roth): leaf and rootmorphology and carbon partitioning. Trees 20:42–52 (2006).

43 Foyer CH and Noctor G, Oxygen processing in photosynthesis:regulation and signalling. New Phytol 146:359–388 (2000).

44 Reddy AR, Chaitanya KV and Vivekanandan M, Drought-inducedresponses of photosynthesis and antioxidant metabolism in higherplants. J Plant Physiol 161:1189–1202 (2004).

45 Munne-Bosch S, The role of α-tocopherol in plant stress tolerance.J Plant Physiol 162:743–748 (2005).

46 Munne-Bosch S and Jon F, New insights into the function oftocopherols in plants. Planta 218:323–326 (2004).


26

60

Research ArticleReceived: 12 November 2009 Revised: 24 June 2010 Accepted: 23 July 2010 Published online in Wiley Online Library: 18 August 2010


Effects of dietary taurine on egg production,egg quality and cholesterol levels in JapanesequailFu-Rong Wang,a,b Xiao-Fang Dong,b Xiao-Ming Zhang,a Jian-Ming Tong,a,b∗Zhong-Guo Xiea and Qi Zhangb

Abstract

BACKGROUND: Taurine is a semi-essential amino acid and has many biological properties. The objective of this study was todetermine the effect of dietary supplementation with taurine on egg production, egg quality, and cholesterol level in serumand egg yolk of quails. A total of 108 quails aged 6 weeks were randomly allocated to three dietary treatments. Each treatmentconsisted of four replicates of nine quails. The diets were supplemented with 0, 100, and 500 mg kg−1 of taurine for 8 weeks.

RESULTS: Dietary 500 mg kg−1 taurine significantly affected egg production rate and feed conversion ratio, but had nosignificant effects on body weight gain, feed consumption, or egg weight. Dietary taurine had no significant effect on eggquality parameters studied. Serum triglyceride concentration was reduced significantly with supplementation of taurine at100 and 500 mg kg−1. Egg yolk cholesterol content was reduced significantly, and the contents of serum taurine and egg yolktaurine were increased significantly with taurine supplementation at 500 mg kg−1.

CONCLUSION: Results of the present study indicated that adding 500 mg kg−1 taurine reduced yolk cholesterol concentrationand increased yolk taurine content without adverse effects on performance and egg quality of laying quails.c© 2010 Society of Chemical Industry

Keywords: taurine; quail; performance; egg quality; cholesterol

INTRODUCTIONTaurine (2-aminoethanesulfonic acid) is a free, semi-essential,sulfur-containing β-amino acid.1 It is not incorporated intoproteins and is in fact the most abundant free amino acid inmany tissues.2 Taurine has been reported to have many biologicalproperties, including osmoregulation,3 membrane stabilization,4

antioxidant,5 calcium homeostasis,6 modulation of immunity,7

and growth modulation.8

Taurine plays an important role in conjugation of bile acids thatare formed from cholesterol in the liver, suggesting that there isa close relationship between taurine and cholesterol metabolism.Taurine also plays an important role in lipid metabolism. Recentstudies have shown the hypolipidemic effect of taurine in variousspecies, including rats,9 hamsters,10 mice,11 and rabbits,12 withhypocholesterolemia induced by feeding a high-fat diet.

A previous study found that the serum cholesterol concentrationwas significantly reduced in quails receiving taurine when com-pared to a control treatment without taurine supplementation.13

However, the literature concerned with the effects of taurine onegg production, egg quality, egg yolk cholesterol and egg taurinecontent is limited. The specific objectives of this experiment wereto determine the effect of dietary taurine supplementation onperformance, egg quality, cholesterol in serum and egg yolk, andtaurine content in serum and egg yolk in Japanese quails (Coturnixcoturnix japonica).

MATERIALS AND METHODSBirds, diets, and managementA total of 108 quails (Coturnix coturnix japonica) aged 6 weeks wererandomly allocated to three dietary treatments. The birds were fedeither a basal diet or the basal diet with taurine supplementationat a rate of 100 or 500 mg kg−1 feed. Each treatment consistedof four replicates of nine quails. They were housed in cages in awindowed poultry house with a light regime of 16L:8D. Feed andwater were provided for ad libitum consumption. The experimentwas conducted for 8 weeks.

The ingredient and chemical composition of the diets are givenin Table 1. The basal diet contained 2810 kcal kg−1 metabolizableenergy (ME) and 201.3 g kg−1 crude protein, and was calculated to

∗ Correspondence to: Jian-Ming Tong, State Key Laboratory of Food Science andTechnology, and School of Food Science and Technology, Jiangnan University,Jiangsu Wuxi 214122, China. E-mail: [email protected]

a State Key Laboratory of Food Science and Technology, and School of FoodScience and Technology, Jiangnan University, Jiangsu Wuxi 214122, China

b State Key Laboratory of Animal Nutrition, Institute of Animal Science, ChineseAcademy of Agricultural Science, Beijing 100193, Chinasupported by theNational Scientific and Technological Supporting Project of People’s Republic ofChina (2006BAD12B05), Agricultural S&T Achievement Transfer Fund Program(2007 GB23260400), and Basic Science Research Program


26

61

Effects of taurine on laying quails www.soci.org

Table 1. Composition of the basal level(%) diet and its nutrient levels

Ingredient Composition (g kg−1)

Corn 511.7

Soybean meal 32.50

Corn protein 3.50

Soybean oil 30.0

CaHPO4 14.0

Limestone 55.0

Methionine 1.3

Vitamin–mineral premixa 10.0

Salt 3.0

Nutrient composition Level

MEb (kcal kg−1) 2810

Crude protein 201.3

Calcium 25.1

Phosphorus 4.6

Lysine 10.3

Methionine 4.5

a Supplied per kg diet: 3300 IU vitamin A, 900 IU cholecalciferol, 25IU vitamin E, 1.04 mg vitamin K3, 5 mg riboflavin, 2.04 mg thiaminemononitrate, 7.5 mg D-biotin, 3450 mg choline chloride, 1.03 mg folicacid, 20.10 mg niacin, 16.64 mg calcium pantothenate, 3.06 mg pyri-doxine hydrochloride, 0.3 mg vitamin B12 (cyanocobalamin), 142.83 mgZnSO4.H2O, 189.67 mg FeSO4.H2O, 13.08 mg CuSO4, 0.46 mg Na2SeO3,192.53 mg MnSO4.H2O, 0.41 mg KI. 2ME: Metabolizable energy.

meet or slightly exceed the nutrient requirements recommendedby the National Research Council (1994).14

Laying performanceQuails were weighed individually at the beginning and at the endof the experiment. Eggs were collected daily and egg productionwas calculated on a bird–day basis. Egg weight was recorded dailyfor each replicate. Feed consumption was recorded weekly. Thefeed conversion rate (FCR) was calculated as kilograms of feed perkilogram of egg. Mortality was recorded when it occurred and wasexpressed as a percentage.

Egg quality measurementsEight eggs from each treatment (two eggs from each replicate)were randomly collected on the 2nd, 4th, 6th and 8th week of theexperiment to study the egg quality traits: egg weight (g), egglength and width (cm), eggshell strength (kg cm−2) and thickness(mm), albumen height (mm), width and height (mm) of yolk.Eggs were weighed by electronic scale; the eggshell strength wasmeasured in kilograms per centimeters squared using an eggshellforce gauge (model-II, Robotmation Co. Ltd, Tokyo, Japan). Eggalbumen height and yolk height were measured using an eggquality gauge, and yolk width was measured by vernier caliper.The eggshell thickness was measured using the vernier caliper atthree points: the two narrow ends and in the middle of the egg.Egg shape index was measured by vernier caliper that considerswidth : length ratio as a percentage. The yolk index was calculatedas height of yolk (mm)/average width of yolk (mm). Interior eggquality was measured by Haugh unit. Haugh unit was calculatedas 100 × log (H − 1.7 W0.37 + 7.57), where H is albumen height(mm) and W is egg weight in (g).

Table 2. The effects of taurine on performance of laying quails

Taurine (mg kg−1)

0 100 500

Body weight gain (g) 29.46 ± 6.99a 28.64 ± 4.42a 27.44 ± 2.99a

Egg weight (g) 10.86 ± 0.26a 10.88 ± 0.15a 10.80 ± 0.22a

Laying rate (%) 82.72 ± 0.93a 83.42 ± 1.46a 86.19 ± 1.26b

Daily feed intake (g) 23.00 ± 0.68a 22.54 ± 0.78a 21.89 ± 0.60a

Feed conversion rate 2.59 ± 0.12b 2.53 ± 0.07b 2.36 ± 0.10a

Means in the same row with no common letters differ significantly(P < 0.05).

Cholesterol analysesFeed was removed 12 h before collecting blood. Blood sampleswere collected by cardiac puncture from eight randomly selectedquails per treatment group (two from each replicate) at theend of the experiment and centrifuged at 1500 × g for 10 min.The serum was then collected from the blood and stored at−20 ◦C for determination of serum parameters. The levels oftotal cholesterol and triglyceride in the serum were determinedenzymatically with commercially available reagent kits accordingto the manufacturer’s instructions (Bio Sino Bio-technology andScience Inc., Beijing, China).

Yolk cholesterol was determined by using a spectrophotometerduring the last week of the trial by the method of Rudel andMorris.15

Taurine analysesPlasma serum was deproteinized by 6% sulfosalicylic acid andcentrifugation at 3000 × g for 15 min. The extraction of taurinefrom egg yolk was carried out by homogenization in 6%sulfosalicylic acid and centrifugation at 3000 × g for 15 min. Thesupernatants of serum and egg yolk were collected separately,and taurine content was then determined by high-performanceliquid chromatography (HPLC) as described by Chen et al.16

Statistical analysesAll the data were analyzed by SPSS 13.0 software for Windows(SPSS Inc., Chicago, IL, USA). The differences among treatmentsin each group were determined by one-way analysis of variance(ANOVA). The significance of mean differences between groupswas tested by Duncan’s test. The means with P-values ≤0.05 wereconsidered significantly different. Data were expressed as mean± SD.

RESULTSLaying performanceThe effects of taurine supplementation on performance are shownin Table 2. There were no significant differences (P > 0.05) inbody weight gain, egg weight or daily feed intake among thetreatments. Furthermore, 100 mg kg−1 taurine had no significanteffect on egg production rate or FCR (P > 0.05). However, theaddition of 500 mg kg−1 taurine in the diet significantly increasedegg production rate and decreased FCR (P < 0.05).


26

62

www.soci.org F-R Wang et al.

Table 3. The effects of taurine on egg quality of laying quails

Taurine (mg kg−1)

0 100 500

Eggshell thickness (mm) 0.185 ± 0.017a 0.193 ± 0.016a 0.193 ± 0.013a

Eggshell strength (kg cm−2) 1.02 ± 0.26a 1.09 ± 0.16a 1.18 ± 0.22a

Egg shape index (%) 80.48 ± 3.48a 80.16 ± 3.46a 80.25 ± 2.48a

Egg yolk index (%) 41.16 ± 2.67a 41.92 ± 3.43a 41.49 ± 2.16a

Egg Haugh unit 78.28 ± 6.97a 78.76 ± 4.07a 78.23 ± 4.77a

Means in the same row with no common letters differ significantly (P < 0.05).

Table 4. The effects of taurine on serum biochemical, egg yolk cholesterol and taurine concentrations of laying quails

Taurine (mg kg−1)

0 100 500

Blood serum cholesterol (mg dL−1) 144.51 ± 10.58a 142.81 ± 10.61a 140.41 ± 22.58a

Blood serum triglyceride (mmol L−1) 15.29 ± 0.45b 14.81 ± 1.51a 13.15 ± 0.72a

Blood serum taurine (µg 100 mL−1) 38.32 ± 4.33a 42.67 ± 7.94ab 49.12 ± 4.07b

Egg yolk cholesterol (g kg−1 yolk) 17.94 ± 0.44b 17.28 ± 0.66ab 16.81 ± 0.97a

Egg yolk taurine (mg kg−1) 44.68 ± 3.58a 48.57 ± 5.04ab 50.91 ± 2.39b

Means in the same row with no common letters differ significantly (P < 0.05).

Egg qualityThe effect of dietary supplementation of taurine on egg quality ispresented in Table 3. There were no significant differences in theeggshell thickness, egg shape index, eggshell strength, egg yolkindex or egg Haugh unit among the three groups (P > 0.05).

Serum cholesterol and triglyceride concentration and eggcholesterolThere was no significant reduction in serum cholesterol con-centration when laying quails were fed the diets with 100 or500 mg kg−1 taurine (Table 4). The serum triglyceride concentra-tion decreased significantly due to supplementation of taurineat 100 and 500 mg kg−1. The addition of 500 mg kg−1 taurinesignificantly decreased yolk cholesterol content (P < 0.05).

Serum and egg yolk taurineThe serum concentration of taurine and egg yolk taurine contentfrom the quails fed 500 mg kg−1 taurine were significantly higherthan those of the control group (Table 4).

DISCUSSIONThe results indicated that the supplementation of taurine at500 mg kg−1 improved the egg production and feed conversionratio in laying quails, while egg weight and body weight gainwere not influenced. Yamazaki and Takemasa17 found that dietary2.5 g kg−1 taurine in laying hens enhanced egg production andfeed conversion but the egg weight decreased. They also observedthat egg weight was decreased without affecting egg production,feed conversion, or body weight by dietary 5 g kg−1 taurine inlaying hens.

No information is available with regard to the effect ofdietary taurine on serum triglyceride concentration of the

laying quails. Our present study demonstrated that taurinesupplementation decreased serum triglyceride concentration. Thisresult is consistent with similar findings in other animals andhumans.18 – 20

Though the addition of taurine to the diet did not reduceserum level of cholesterol in normal rats,21,22 many studiesshowed that it can diminish the degree of increase in the serumcholesterol concentration induced by feeding a diet containing alarge amount of cholesterol.9 – 13 Our present study demonstratedthat the taurine supplementation at 500 mg kg−1 level decreasedthe cholesterol level significantly in the egg yolk. The results ofthese studies tend to vary depending on the species of animalused, and the composition of taurine supplementation and of theexperimental diet.

Taurine is the most abundant free amino acid in severaltissues and in the cellular components of blood.23 Dietary taurinesupplementation at 500 mg kg−1 increased serum taurine levelsignificantly. Yokogoshi et al.22 found that the serum taurineconcentration in rat was decreased by the intake of the high-cholesterol diet, but the decrease was gradually restored by theadministration of taurine in a dose-dependent manner. Similarly,serum taurine concentration in cats increased with increaseddietary supplementation of taurine.24 These results imply thatdietary taurine could increase serum taurine in a dose-dependentmanner. Moreover, earlier observations in rats,25,26 cats27 andhumans28 obtained similar results.

Taurine mostly existed only in the yolk, but not in the albumen.The present study showed that the egg yolk taurine level wassignificantly increased with the dietary taurine supplementationat 500 mg kg−1. Hu et al.29 found that adding taurine to thediet similarly increased milk taurine content. Taurine has manybiological functions; taurine supplements have been shown tobe beneficial for infants and some groups of elderly subjects,and it is especially essential to the fetus and newborn for their


26

63

Effects of taurine on laying quails www.soci.org

development.24,30 Thus the results from the present study, alongwith findings from the previous studies mentioned above, suggestthat increased levels of taurine in egg yolk through dietarysupplementation may aid in improving the nutritional value ofeggs.

CONCLUSIONSOn the basis of the above experiment, it is concluded that dietarysupplementation of taurine at 500 mg kg−1 in laying quails’ dietscould enhance egg production, improve the egg yolk taurinecontent, and reduce the egg yolk cholesterol concentration andserum triglyceride level without adversely affecting on egg quality.

ACKNOWLEDGEMENTSThis research was supported, in part, by the National Scientific andTechnological Supporting Project of the People’s Republic of China(2006BAD12B05), Agricultural Science and Technology Achieve-ments Transformation Fund Programs (2007 GB23260400), and theBasic Research Operating Expenses of the Central-level Non-profitResearch Institutes (ywf-td-4). The authors thank all the peoplewho offered help in this study.

REFERENCES1 Gaull GE, Taurine in human milk: growth modulator or conditionally

essential amino acid. J Pediatr Gastroenterol Nutr 21:5266–5271(1983).

2 Huxtable RJ, Physiological actions of taurine. Physiol Rev 72:101–163(1992).

3 Pasantes-Morales H and Schousboe A, Role of taurine inosmoregulation in brain cells: mechanisms and functionalimplications. Amino Acids 12:281–292 (1997).

4 Pasantes-Morales H, Wright CE and Gaull GE, Taurine protectionof lymphoblastoid cells from iron-ascorbate induced damage.Biochem Pharmacol 34:2205–2207 (1985).

5 Gurer H, Ozgunes H, Saygin E and Ercal N, Antioxidant effect of taurineagainst lead-induced oxidative stress. Arch Environ Contam Toxicol41:397–402 (2001).

6 Foos TM and Wu JY, The role of taurine in the central nervoussystem and the modulation of intracellular calcium homeostasis.Neurochem Res 27:21–26 (2002).

7 Redmond HP, Stapleton PP, Neary P and Bouchier-Hayes D,Immunonutrition: the role of taurine. Nutrition 14:599–604 (1998).

8 Hayes KC, Stephen ZF and Sturman JA, Growth depression in taurine-depleted infant monkeys. J Nutr 110:2058–2064 (1980).

9 Park T and Lee K, Dietary taurine supplementation reduces plasma andliver cholesterol and triglyceride levels in rats fed a high-cholesterolor a cholesterol-free diet. Adv Exp Med Biol 442:319–325 (1998).

10 Murakami S, Kondo Y, Toda Y, Kitajima H, Kameo K, Sakono M, et al,Effect of taurine on cholesterol metabolism in hamsters: up-regulation of low density lipoprotein (LDL) receptor by taurine.Life Sci 70:2355–2366 (2002).

11 Murakami S, Kondo-Ohta Y and Tomisawa K, Improvement incholesterol metabolism in mice given chronic treatment of taurineand fed a high-fat diet. Life Sci 64:83–91 (1999).

12 Petty MA, Kintz J and DiFrancesco GF, The effects of taurine onatherosclerosis development in cholesterol-fed rabbits. Eur JPharmacol 180:119–127 (1990).

13 Jackson JA and Burns MJ, Effects of cystine, niacin and taurine oncholesterol concentration in the Japanese quail with comments onbile acid metabolism. Comp Biochem Physiol 48:61–68 (1974).

14 NRC, Nutrient Requirements of Poultry (9th rev. edn). NationalAcademies Press, Washington, DC (1994).

15 Rudel LL and Morris MD, Determination of cholesterol usingo-phthaladehyde. J Lipid Res 14:364–366 (1973).

16 Chen ZL, Xu G, Specht K, Yang RJ and She SW, Determination of taurinein biological samples by reversed-phase liquid chromatographywith precolumn derivatization with dinitrofluorobenzene. AnalChim Acta 296:249–253 (1994).

17 Yamazaki M and Takemasa M, Effects of dietary taurine on egg weight.Poult Sci 77:1024–1026 (1998).

18 Nakaya Y, Minami A, Harada N, Sakamoto S, Niwa Y and Ohnaka M,Taurine improves insulin sensitivity in the Otsuka Long-EvansTokushima Fatty rat, a model of spontaneous type 2 diabetes.Am J Clin Nutr 71:54–58 (2000).

19 Balkan J, Kanbagli O, Hatipoglu A, Kucuk M, Cevikbas U, Aykac-Toker G, et al, Improving effect of dietary taurine supplementationon the oxidative stress and lipid levels in the plasma, liver and aortaof rabbits fed on a high-cholesterol diet. Biosci Biotechnol Biochem66:1755–1758 (2002).

20 Zhang M, Bi LF, Fang JH, Su XL, Da GL, Kuwamori T, et al, Beneficialeffects of taurine on serum lipids in overweight or obese non-diabetic subjects. Amino Acids 26:267–271 (2004).

21 Mochizuki H, Oda H and Yokogoshi H, Increasing effect of dietarytaurine on the serum HDL-cholesterol concentration in rats. BiosciBiotechnol Biochem 62:578–579 (1998).

22 Yokogoshi H, Mochizuki H, Nanami K, Hida Y, Miyachi F and Oda H,Dietary taurine enhances cholesterol degradation and reducesserum and liver cholesterol concentrations in rats fed a high-cholesterol diet. J Nutr 129:1705–1711 (1999).

23 Schuller-Levis G and Park E, Taurine: new implications for an old aminoacid. Fems Microbiol Lett 226:195–202 (2003).

24 Sturman JA, Taurine in development. Physiol Rev 73:119–147 (1993).25 Lombardini JB and Medina EV, Effects of dietary inorganic sulfate,

taurine, and methionine on tissue levels of taurine in the growingrat. J Nutr 108:428–433 (1978).

26 Dawson R, Liu S, Eppler B and Patterson T, Effects of dietary taurinesupplementation or deprivation in aged male Fischer 344 rats. MechAgeing Dev 107:73–91 (1999).

27 Laidlaw SA, Sturman JA and Kopple JD, Effect of dietary taurine onplasma and blood cell taurine concentrations in cats. J Nutr117:1945–1949 (1987).

28 Vinton NE, Laidlaw SA, Ament ME and Kopple JD, Taurine concen-trations in plasma and blood cells of patients undergoing long-termparenteral nutrition. Am J Clin Nutr 44:398–404 (1986).

29 Hu JM, Rho JY, Suzuki M, Nishihara M and Takahashi M, Effect oftaurine in rat milk on the growth of offspring. J Vet Med Sci62:693–698 (2000).

30 Cho KH, Kim ES, Chen JD, Zhang S, Kim H, Kim SY, et al, Serum andurine taurine levels in elderly patients undergoing long-term enteralnutrition are reduced overtime. Nutr Res 22:1017–1025 (2002).


26

64

Research ArticleReceived: 13 October 2009 Revised: 24 June 2010 Accepted: 23 July 2010 Published online in Wiley Online Library: 24 August 2010


Postmortem degradation of desminand calpain in breast and leg and thigh musclesfrom Taiwan black-feathered country chickensYa-Shiou Chang and Rong-Ghi R. Chou∗

Abstract

BACKGROUND: Several studies have reported that the postmortem changes are more rapid in breast muscles (BM) than in legand thigh muscles (LM) of chickens. However, the reasons for the differences in postmortem proteolysis of BM and LM are stilluncertain. The purpose of this study was therefore to compare the postmortem degradation of desmin and calpains in BM andLM from Taiwan black-feathered country chickens at 5 ◦C.

RESULTS: The pH was lower (P < 0.05) in BM than in LM. Western blot indicated that postmortem desmin degradation was morerapid in BM than in LM. Casein zymograms showed that at-death µ-calpain activity was higher in BM than in LM. As postmortemtime proceeded, µ-calpain was activated and autolyzed more extensively in BM than in LM. However, the µ/m-calpain activityremained stable during postmortem storage in both BM and LM.

CONCLUSION: Our results suggest that the more rapid postmortem proteolysis found in BM than in LM at 5 ◦C similar with theprevious study could be mainly explained by both greater amounts and faster activation and autolysis of µ-calpain in BM.c© 2010 Society of Chemical Industry

Keywords: chicken muscle; postmortem changes; calpain; desmin

INTRODUCTIONIt is generally accepted that postmortem degradation of cytoskele-tal proteins such as desmin by calcium-dependent proteases(calpains) may improve meat tenderness.1 – 3 There are two ubiq-uitous calpains (µ- and m-calpain) and one tissue-specific calpain,p94, that are present in mammalian muscle cells.4 Instead ofm-calpain, µ/m-calpain, which contains sequence homology andCa2+ sensitivity between µ- and m-calpain, is present in chickenmuscle.5 Free calcium concentration required for chicken µ- andµ/m-calpain activation is lower than for mammalian µ- and m-calpain.5 Among those calpains, µ-calpain is thought to play anessential role in the postmortem proteolysis of bovine muscleat 5 ◦C,6 although m-calpain may also be involved.7,8 Severalstudies9,10 have indicated that postmortem proteolysis is morerapid in avian breast muscles (BM) than in leg and thigh muscles(LM). Early studies9,11 reported that the difference in fiber typecomposition between BM (white) and LM (red) may explain thedifference in the rate of postmortem proteolysis in these twomuscle types. However, studies show that because red muscles(m. vastus intermedius and m. soleus) exhibit the same rate ofdesmin and troponin-T postmortem degradation as white muscle(m. semitendinosus) in pigs, the difference in rate of postmortemproteolysis cannot be solely explained by the difference in fibertype composition.12 Therefore, more studies are needed to ex-amine the differences in postmortem proteolysis of BM and LM.Taiwan black-feathered country chickens were very popular inlocal markets, and its production value in 2008 was close to$300 million in US currency.13 Little information, however, wasavailable regarding postmortem changes in the muscles of this

strain. Additionally, because previous studies10 found that desmindegraded very rapidly in 1-day postmortem chicken BM, it wouldbe interesting to examine the relationship between calpain autol-ysis and desmin degradation within 24 h post mortem. Thus thepurpose of this study was to compare the degradation of desminand calpains within 24 h post mortem in BM and LM from Taiwanblack-feathered country chickens at 5 ◦C.

MATERIALS AND METHODSSample preparationTaiwan black-feathered country chickens (female, ∼100 days oldwith an average live weight of ∼1.5 kg) were slaughtered in a localabattoir by using standard commercial practices. The carcasses(30–40 min post mortem) were vacuum-packed and stored at 5 ◦C.Breast and leg and thigh muscles were sampled at 3, 6, 12 and 24 hof storage. The 0 h samples were taken at the time of killing (within5–10 min post mortem). Twenty chickens were randomly assignedto each of the five sampling times. Four chickens were sampled ateach time in each of three replications. Muscle samples (70–80 g)from the four chickens at each sampling time were combinedand ground through a 3 mm plate. The ground samples were

∗ Correspondence to: Rong-Ghi R. Chou, Department of Animal Science,National Chiayi University, 300 University Road, Chiayi City, 60083 Taiwan.E-mail: [email protected]

Department of Animal Science, National Chiayi University, Chiayi City, Taiwan


26

65

Postmortem changes in breast and leg muscles in chickens www.soci.org

immediately frozen in liquid nitrogen and stored at −80 ◦C beforepH measurement, western blot analysis, and casein zymography.The pH was determined by the method of Farouk and Swan.14

Western blot analysisBreast and leg and thigh myofibrils from the pooled sam-ples were isolated via the method of Huff-Lonergan et al.15

The protein concentration of myofibrils was determined byusing a modified biuret method.16 Myofibril samples forsodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) were prepared by the method of Wang et al.17 TheSDS-PAGE was done in a 120 g kg−1 tris-glycine slab gel(acrylamide : methylenebisacrylamide = 37.5 : 1, w/w) by themethod of Laemmli.18 The same amount of protein (150 µg)from each sample was loaded on each well of the gels. Proteinswere transferred from a 120 g kg−1 slab gel to a nitrocellulosemembrane by the method of Towbin et al.19 After the transfer,the membrane was incubated in a 50 g kg−1 bovine serum albu-min–phosphate buffer solution (BSA-PBS) for 30 min at 37 ◦C andthen washed three times (5 min each) in a 1 g kg−1 BSA-PBS solu-tion at room temperature. A desmin monoclonal antibody (cloneDE-U-10; D-1033, Sigma-Aldrich Co., St Louis, MO, USA) was usedas a primary antibody. The membrane was incubated with theprimary antibody for 2 h at room temperature, washed three times(5 min each) in 1 g kg−1 BSA-PBS, incubated with a secondary an-tibody – goat anti-mouse-HRP (A-9044, Sigma-Aldrich) – for 2 h atroom temperature, and washed twice (5 min each) in 1 g kg−1 BSA-PBS solution and twice (1 min each) in deionized water. The colorwas developed by SIGMAFAST 3,3′-diaminobenzidine tablets(D-4418, Sigma-Aldrich).

Casein zymographyThe procedure used for protein extraction was based on themethod of Veiseth et al.20 Briefly, a 5 g sample was homogenizedin three volumes of extraction buffer (100 mmol L−1 Tris base,10 mmol L−1 ethylenediaminetetraacetic acid (EDTA), 0.5 g kg−1

2-mercaptoethanol, pH 8.3) at 5 ◦C. Homogenates were cen-trifuged at 22 000×g for 25 min at 5 ◦C. The protein concentrationof the supernatant was determined by using a modified biuretmethod.16 Zymograms were routinely run in 120 g kg−1 gels(acrylamide : methylenebisacrylamide = 37.5 : 1, w/w) containing2.1 g kg−1 casein (w/v) by the method of Raser et al.21 The sam-ple buffer (150 mmol L−1 Tris-HCl, pH 6.8, 200 g kg−1 glycerol,7.5 g kg−1 2-mercaptoethanol (MCE), 0.2 g kg−1 (w/v) bromophe-nol blue) was added to the protein extract at a ratio of two partsof the buffer to three parts of protein extract (v/v). The sameamount of protein (250 µg) from each sample was loaded onto each well of the casein gels. The casein minigels (0.75 mm,Bio-Rad Laboratories, Hercules, CA, USA) were pre-run at 100 Vfor 15 min (5 ◦C), with a running buffer containing 25 mmol L−1

Tris-HCl, 0.5 g kg−1 MCE, 192 mmol L−1 glycine, and 1 mmol L−1

EDTA (pH 8.3) before samples were loaded on to the wells. Thegels were run at 100 V for 2 h (5 ◦C) and then incubated at roomtemperature in a 50 mmol L−1 Tris-HCl (pH 7.5) buffer containing0.5 g kg−1 MCE and CaCl2 (0.01, 0.03, 0.1 or 4 mmol L−1) with slowshaking for 1 h (three changes of the buffer). This was followed bya 16 h incubation in the same buffer at 37 ◦C. The gels were thenstained for 2 h with Coomassie blue (R-250) and destained with200 g kg−1 methanol and 70 g kg−1 acetic acid.

Image analysisTwo to three representative gels or blots from each replicationwere used for image analysis. The bands in blots and caseingels incubated with 4 mmol L−1 Ca2+ to activate all calpainisoforms were digitized using a scanner (model J131B, EpsonTaiwan Technology and Trading LTD, Taiwan) using Photoshopsoftware. The resulting signals were quantified using Image Gauge(version 3.46, Science Lab 99 for Windows, Fuji Film, Tokyo, Japan).

Statistical analysisSplit-plot design was used in this study. Whole unit was the typeof muscle, and subunit was the muscle samples taken at eachsampling time. This experiment was done with three replications,and triplicate samples were taken at each sampling time in eachof three replications. All data were analyzed by the procedureof the SAS-GLM program (SAS Institute Inc., Cary NC, USA), andmean separation was determined using the least squares meansprocedure.22

RESULTS AND DISCUSSIONThe average at-death pH was 6.11± 0.05 and 6.45± 0.08 in breast(BM) and leg and thigh (LM) muscles, respectively, and the pHdecreased significantly (P < 0.05) to 5.70 ± 0.04 and 6.28 ± 0.06,respectively, by 6 h post mortem (Fig. 1). However, the pH changedinsignificantly (P > 0.05) from 6 to 24 h post mortem in the BM(5.68 ± 0.07) and LM (6.27 ± 0.06) samples (Fig. 1). These resultswere consistent with the findings of Lee et al.,23 who showed thatthe ultimate pH in chicken muscle was reached by 6 h post mortem.Figure 1 also shows that the pH was significantly (P < 0.05) lowerin the BM than in the LM samples after 24 h postmortem storage at5 ◦C, consistent with our previous report.24 This difference mightbe attributed to the higher content25 and to the faster utilization23

of muscle glycogen in chicken breast muscle.Desmin, an intermediate filament protein, is a calpain-sensitive

protein.26 Western blots labeled with a desmin monoclonalantibody indicate that BM desmin (Fig. 2) decreased more rapidlythan LM desmin (Fig. 2), confirming previous studies.24 In BMsamples, most of the desmin remained intact in 3 h postmortemsamples, which contained ∼85% of the desmin present inBM sampled at death. However, the most extensive desmindegradation (P < 0.05) was observed in 6 h postmortem samples,which lost ∼75% of the at-death desmin. Between 6 and 12 hpost mortem, there was a small decrease (<5%) in BM desmin andlittle additional desmin decrease was seen in 24 h BM samples. Onthe other hand, ∼87% of the at-death LM desmin was left in 3 hpostmortem samples and remained after 6–24 h post mortem.

Casein gel zymography is a sensitive method to detect calpainproteases.20 In agreement with previous studies on chickenmuscle,27 casein zymograms showed that the bands of µ- andµ/m-calpain in the at-death BM and LM samples began to appearfaintly in the presence of 10 (Fig. 3(A)) and 30 µmol L−1 Ca2+(Fig. 3(B)), respectively. The bands on casein zymograms becamemore apparent as Ca2+ concentration increased. However, noextra bands were found in the presence of 100 µmol L−1 (Fig. 3(C))or 4 mmol L−1 Ca2+ (Fig. 3(D)) in both BM and LM at-deathsamples. Results also indicated that µ-calpain, instead of µ/m-calpain, migrated as a doublet on gels, similar to the studies ofVeiseth et al.20 in postmortem ovine muscle, which showed thatnative µ-calpain (lower band) and its autolyzed form (upper band)migrated very closely on casein gels as seen on ours. Therefore it


26

66

www.soci.org Y-S Chang, R-GR Chou

5.0

5.5

6.0

6.5

7.0

0 3 6 9 12 15 18 21 24

Time post mortem, h

pH

BMLM

Figure 1. Changes in pH of breast (BM) and leg (LM) muscles from Taiwanblack-feathered country chickens during postmortem storage at 5 ◦C. •,BM samples; ◦, LM samples.

Figure 2. Western blotting showing changes in desmin of BM (A) andLM (B) samples from the Taiwan black-feathered country chickens duringpostmortem storage at 5 ◦C. These blots are representative of threereplications of combined samples. D, Desmin; lane 1, at-death; lane 2, 3 h;lane 3, 6 h; lane 4, 12 h; lane 5, 24 h.

Figure 3. Zymograms showing µ- and µ/m-calpains in the presenceof 10 µmol L−1 (A), 30 µmol L−1 (B), 100 µmol L−1 (C) and 4 mmol L−1

(D) calcium in at-death BM and LM samples from Taiwan black-featheredcountry chickens. These gels are representative of three replications ofcombined samples. Lane 1, BM; lane 2, LM. µ- = µ-calpain; µ/m- = µ/m-calpain.

was reasonable to predict that the doublet found in BM samplesmight consist of µ-calpain and its autolyzed form. Because Jiand Takahashi28 reported that the free calcium concentration inchicken muscle is 70 µmol L−1 20 min post mortem, it is possiblethat µ-calpain in the at-death samples, which were collectedwithin 5–10 min post mortem, had been activated and autolyzed.Image analysis (Fig. 4) results show that the at-death activity ofµ-calpain was approximately two times higher (P < 0.05) inBM than in LM samples. The at-death activity of µ/m-calpain

0

25

50

75

100

125

150

µ-calpain µ/m-calpain

Rel

ativ

e ac

tivity

, %

a

b

a a

Figure 4. Activity of µ- and µ/m-calpain in at-death BM and LM samplesfrom Taiwan black-feathered country chickens. Activities measured in theat-death BM samples are taken as 100%. Vertical bars show the standarddeviation of the means. Different letters within each calpain isoformindicate significant differences. �, BM; �, LM.

was approximately 10% lower in BM than in LM samples (Fig. 3),although the difference was insignificant (P > 0.05).

Figure 5 shows casein zymograms ofµ- andµ/m-calpain activityin BM and LM samples at different times of postmortem storageat 5 ◦C. The results indicated that µ-calpain activity decreasedas postmortem time proceeded (Fig. 5). Figure 6 shows that thegreatest reduction (P < 0.05) of µ-calpain activity was found inthe 6 h postmortem BM samples, which lost ∼60% of the at-deathBM µ-calpain activity. This implies that µ-calpain autolysis wasvery extensive by 6 h post mortem and explains why BM desmindegraded extensively simultaneously (Fig. 2). By 6 h post mortem,the level of µ-calpain activity was similar in BM and LM samplesbecause the reduction of µ-calpain activity was much slower inLM, if the at-death BM µ-calpain activity (approximately two timeshigher than the at-death LM samples) was taken as 100% (Fig. 6).After 24 h post mortem, µ-calpain activity in both BM and LMsamples further decreased to ∼20% of the at-death BM activity(Fig. 6).

Recent studies28 in postmortem bovine muscle showed thatthe 80 kDa subunit of µ-calpain was autolyzed first to a 78 kDaintermediate product, from which NH2-terminal 14 amino acids ofthe 80 kDa subunit were removed,4,29 and completely autolyzedto produce a 76 kDa form, from which an additional NH2-terminal12 amino acids were removed.4,29 However, the 28 kDa smallsubunit of µ-calpain was not autolyzed in postmortem bovinemuscle.28 Camou et al.29 concluded that the 76/28 kDa µ-calpainwas proteolytically inactive and that this accounted for the loss ofµ-calpain activity during postmortem storage.

It has been reported that postmortem proteolysis is more rapidin BM than in LM.9 – 11 As shown in Fig. 1, the LM samples hada higher ultimate pH (∼6.3) than the BM samples (∼5.7). It wasalso reported that little calpain activity could be found at pH 5.8in bovine muscle.7 Presumably µ-calpain, which is essential forpostmortem proteolysis at 5 ◦C,6,30 was more active in the LMsamples than in the BM samples and might favor more extensivepostmortem desmin degradation. However, results (Fig. 4) not onlyshowed that BM contained approximately two times greater µ-calpain activity than LM in the at-death samples, but also indicatedthat most of the µ-calpain (∼60% of the at-death activity) wasautolyzed in BM samples (Fig. 6) before the pH reached ∼5.7 by6 h postmortem (Fig. 1). It was recently proposed that calpainactivation might explain the variation in postmortem desmindegradation in pig muscle.3 Therefore the differences in the extent


26

67

Postmortem changes in breast and leg muscles in chickens www.soci.org

Figure 5. Zymograms showing postmortem changes in µ- and µ/m-calpains in the presence of 4 mmol L−1 calcium in BM (A) and LM (B) samples fromTaiwan black-feathered country chickens at 5 ◦C. These gels are representative of three replications of combined samples. µ, µ-calpain; µ/m, µ/m-calpain.Lane 1, at-death; lane 2, 3 h; lane 3, 6 h; lane 4, 12 h; lane 5, 24 h.

0

20

40

60

80

100

120

0 3 6 9 12 15 18 21 24

Postmortem time, h

Rel

ativ

e ac

tivity

, %

Figure 6. Postmortem changes in µ-calpain activity of BM and LM samplesfrom Taiwan black-feathered country chickens at 5 ◦C. Activity is expressedas a percentage of the at-death BM µ-calpain activity, which is taken as100%. ◦, BM; •, LM.

0

50

100

150

200

0 3 6 9 12 15 18 21 24

Postmortem time, h

Rel

ativ

e ac

tivity

, %

Figure 7. Postmortem changes in µ/m-calpain activity of BM and LMsamples from Taiwan black-feathered country chickens at 5 ◦C. Activity isexpressed as percentage of the at-death BM µ/m-calpain activity, which istaken as 100%. ◦, BM; •, LM.

of postmortem proteolysis between BM and LM probably dependmore on µ-calpain level and the rate of its activation and autolysisthan on the difference in final pH between the two muscle types.

On the other hand, µ/m-calpain activity was very stable duringpostmortem storage in both BM and LM samples (Figs 5 and 7).The bands below the µ/m-calpain band in the 6, 12 and 24 hpostmortem BM and LM samples were visible in the presenceof 30 µmol L−1 Ca2+ (results not shown) and clearly seen inthe presence of 4 mmol L−1 Ca2+ (Fig. 5). It was suggested thatthese bands might be generated from µ/m-calpain23 because thefree calcium concentration in 6 h postmortem chicken muscleexceeded 120 µmol L−1,28 which is enough for µ/m-calpainactivation and autolysis. Additionally, these bands were moreapparent in the LM than in BM samples (Fig. 5) because the higherultimate pH would favor continued proteolytic activity (Fig. 1).Results indicated that the µ/m-calpain in BM and LM was activeafter 6 h post mortem. However, our western blot (Fig. 2) showed

that little BM or LM desmin degraded after 6 h post mortem. Thisimplies that µ/m-calpain may play a relatively minor role in desmindegradation after 6 h post mortem. Therefore more research wouldbe needed to understand the precise role of µ/m-calpain in thepostmortem proteolysis of chicken muscle.

In summary, we compared postmortem degradation of desminand calpain in BM and LM from Taiwan black-feathered countrychicken at 5 ◦C. The results showed that the pH was lower(P < 0.05) in BM than in LM at death and at 3, 6, 12 and 24 hpost mortem. Western blot indicated that postmortem desmindegradation was more rapid in BM than in LM. Casein zymogramsshowed that at-death µ-calpain activity was higher in BM thanin LM. As postmortem time proceeded, µ-calpain was activatedand autolyzed more extensively in BM than in LM. However, theµ/m-calpain activity remained stable during postmortem storagein both BM and LM. Therefore our results suggest that the morerapid postmortem proteolysis found in BM than in LM at 5 ◦C couldbe mainly explained by both greater amounts and faster activationand autolysis of µ-calpain in BM.

REFERENCES1 Hopkins DL and Thompson JM, Factors contributing to proteolysis and

disruption of myofibrillar proteins and the impact on tenderizationin beef and sheep meat. Aust J Agric Res 53:149–166 (2002).

2 Robson RM, Huff-Lonergan E, Parrish FC Jr, Ho C-Y, Stromer MH,Huiatt TW, et al, Postmortem changes in the myofibrillar and othercytoskeletal proteins in muscle. Proc Recip Meat Confer 50:43–52(1997).

3 Huff-Lonergan E and Lonergan SM, Mechanism of water-holdingcapacity of meat: the role of postmortem biochemical and structuralchanges. Meat Sci 71:194–204 (2005).

4 Goll DE, Thompson VF, Li H, Wei W and Cong J, Calpain system. PhysiolRev 83:731–801 (2003).

5 Sorimachi H, Tsukahara T, Okada-Ban M, Sugita H, Ishiura S andSuzuki K, Identification of a third ubiquitous calpain species:chicken muscle expresses four distinct calpains. Biochim BIophysActa 1261:381–393 (1995).

6 Geesink GH, Kuchay S, Chishti AH and Koohmaraie M, µ-Calpain isessential for postmortem proteolysis of muscle proteins. J Anim Sci84:2834–2840 (2006).

7 Camou JP, Marchello JA, Thompson VF, Mares SW and Goll DE, Effectof postmortem storage on activity ofµ- and m-calpain in five bovinemuscles. J Anim Sci 85:2670–2681 (2007).

8 Pomponio L, Lametsch R, Karlsson AH, Costa LN, Grossi A andErtbjerg P, Evidence for post-mortem m-calpain autolysis in porcinemuscle. Meat Sci 80:761–764 (2008).

9 Paxhia JM and Parrish Jr FC, Effect of postmortem storage on titinand nebulin in pork and poultry light and dark muscles. J Food Sci53:1599–1601 (1988).

10 Chou R-GR, Tseng T-F, Lin K-J and Yang J-H, Post-mortem changes inmyofibrillar proteins of breast and leg muscles from broilers, spenthens and Taiwanese country chickens. J Sci Food Agric 65:297–302(1994).

11 Hay JD, Currie RW and Wolf FH, Effect of postmortem aging on chickenmuscle fibrils. J Food Sci 38:981–987 (1973).


26

68

www.soci.org Y-S Chang, R-GR Chou

12 Christensen M, Henckel P and Purslow PP, Effect of muscle type onthe rate of post-mortem proteolysis in pigs. Meat Sci 66:595–601(2004).

13 Council of Agriculture, Executive Yuan, ROC. Agricul-ture Statistics Yearbook (2008). [Online]. Available:http://www.coa.gov.tw/htmlarea file/web articles/coa/11499/010.xls [26 April 2010].

14 Farouk MM and Swan JE, Acceptability and functional properties ofrestructured roast from frozen pre-rigor injected beef. Meat Sci46:57–66 (1997).

15 Huff-Lonergan E, Parrish FC Jr and Robson RM, Effects of postmortemaging time, animal age, and sex on degradation of titin and nebulinin bovine longissimus muscle. J Anim Sci 73:1064–1073 (1995).

16 Robson RM Goll DE and Temple MJ, Determination of protein in ‘Tris’buffer by the biuret reaction. Anal Biochem 24:339–341 (1968).

17 Wang S-M, Greaser ML, Schultz E, Bulinski JC, Lin JJ-C and Lessard JL,Studies on cardiac myofibrillogenesis with antibodies to titin, actin,tropomyosin, and myosin. J Cell Biol 107:1075–1083 (1988).

18 Laemmli UK, Cleavage of structural proteins during the assembly ofthe head of bacteriophage T4. Nature 227:680–685 (1970).

19 Towbin H, Staehelin T and Gordon J, Electrophoretic transfer ofproteins from polyacrylamide gels to nitrocellulose sheet:procedure and some application. Proc Natl Acad Sci USA76:4350–4354 (1979).

20 Veiseth E, Shackelford SD, Wheeler TL and Koohmaraie M, Effect ofpostmortem storage on µ-calpain and m-calpain in ovine skeletalmuscle. J Anim Sci 79:1502–1508 (2001).

21 Raser KJ, Posner A and Wang KW, Casein zymography: a method tostudyµ-calpain, m-calpain and their inhibitory agents. Arch BiochemBiophys 319:211–216 (1995).

22 SAS Institute Inc., User Guide: Statistics, version 8.01. SAS Institute, Cary,NC (1986).

23 Lee HL, Sante-Lhoutellier V, Vigouroux S, Briand Y and Briand M, Roleof calpains of postmortem proteolysis in chicken muscle. Poult Sci87:2126–2132 (2008).

24 Tsai S-F, Lin C-Y, Lu J-J and Chou R-GR, Postmortem proteolysis ofbreast and leg muscles from Taiwan colored chickens and silkiebantam. Asia Aus J Anim Sci 19:739–743 (2006).

25 Schreurs FJG, Post-mortem changes in chicken muscle. World’s PoultSci J 56:319–346 (2000).

26 O’Shea JM, Robson RM, Hartzer MK, Huiatt TW, Rathbun WE andStromer MH, Purification of desmin from adult mammalian skeletalmuscle. Biochem J 195:345–356 (1981).

27 Lee HL, Sante-Lhoutellier V, Vigouroux S, Briand Y and Briand M,Calpain specificity and expression in chicken tissue. Comp BiochemPhysiol B 146:88–93 (2007).

28 Ji JR and Takahashi K, Changes in concentration of sarcoplasmic freecalcium during post mortem ageing of meat. Meat Sci 73:395–403(2006).

29 Camou JP, Marchello JA, Thompson VF, Mares SW and Goll DE, Effectof postmortem storage on activity of µ- and m-calpain in five bovinemuscles. J Anim Sci 85:2670–2681 (2007).

30 Zimmerman UJP and Schlaepfer WW, Two stage autolysis of thecatalytic subunit initiates activation of calpain. Biochim BiophysActa 1078:192–198 (1991).


26

69

Research ArticleReceived: 9 October 2009 Revised: 27 June 2010 Accepted: 23 July 2010 Published online in Wiley Online Library: 25 August 2010


Biogenic amine formation in Turkish dryfermented sausage (sucuk) as affected by nisinand nitriteSukru Kurta∗ and Omer Zorbab

Abstract

BACKGROUND: The effects of nitrite (0, 100, and 200 mg kg−1) and nisin (0, 250, and 500 mg kg−1) on biogenic amine formationin sucuk were investigated by utilising a central composite design of response surface methodology.

RESULTS: The addition of nitrite led to decreased levels of tryptamine, 2-phenylethylamine, putrescine, cadaverine, tyramine,and histamine, whereas nisin decreased the tryptamine level and counts of lactic acid bacteria. However, nisin increasedputrescine, cadaverine, and spermidine levels. Their interactive effect was also found to be significant (P < 0.05) for putrescinevalues.

CONCLUSION: The additional nitrite levels can be decreased by the addition of nisin, which will hinder biogenic amine formation.c© 2010 Society of Chemical Industry

Keywords: biogenic amines; dry fermented sausage; nitrite; nisin; sucuk

INTRODUCTIONSucuk is a dry fermented sausage produced from beef orsheep meat. A variety of microorganisms are involved in sucukfermentation. The production conditions of this meat product mayresult in the formation of biogenic amines, such as putrescine,cadaverine, tyramine, and histamine. Biogenic amines are organicbases with aliphatic, aromatic, or heterocyclic structures that canbe found in several food products. They are mainly generated viadecarboxylation of the corresponding precursors (i.e. amino acids),through substrate-specific microbial enzymes. Microbial growth,acidification, and proteolysis provide favourable conditions for theformation of biogenic amines during meat fermentation. Some ofthese microbial activities are of concern in relation to human healthdue to the potential for toxicological effects.1 Some biogenicamines, such as putrescine and cadaverine, form carcinogenicnitrosamines by reacting with nitrite.

Nitrite is an indispensable additive that plays an important rolein the generation of high-quality sucuk. However, nitrite may alsocause haemoglobinaemia.2 In addition to the toxicological effectof biogenic amines, these metabolic products are also of concernin relation to food hygiene. High amounts of certain amines maybe found in food as a consequence of the using poor-quality rawmaterials, contamination, and inappropriate conditions duringfood processing and storage.3,4 Therefore, biogenic amines andnitrite levels are factors that need to be considered in theproduction of dry fermented sausages. Decreasing nitrite levelsand preventing the accumulation of biogenic amines are critical inthe production of dry fermented sausages.3 To prevent biogenicamine accumulation, raw materials must be of higher microbialquality, through controlled production and ripening conditions.

New approaches, such as using bacteriocinogenic lactic acidbacteria (LAB) cultures and/or the bacteriocins of these cultures,to control pathogenic and spoilage microorganisms have beendeveloped in order to increase food safety.5,6 Among bacteriocins,nisin is mainly used in food, and particularly in dairy products.However, less is known regarding the effects of nisin in meatproducts.5

Nisin exhibits activity against a range of Gram-positive bacteria.The spectrum of activity of this compound may also be extended toGram-negative bacteria through utilising nisin in combination withother agents. Such a combinatory application may also enhancethe activity of nisin against Gram-positive bacteria.7 Nisin is still theonly bacteriocin approved as a food preservative (E234) in morethan 50 countries worldwide, including the USA, the EuropeanUnion, Brazil, and China.5

The effects of nitrite and nisin levels on biogenic amineformation were evaluated using response surface methodology inthis study.

MATERIALS AND METHODSReagents and standardsDansyl chloride from Acros (Acros Organics, Geel, Belgium), andammonia (25%), acetone, acetonitrile (high-performance liquid

∗ Correspondence to: Sukru Kurt, Vocational School, University of Adıyaman,TR-02040 Adiyaman, Turkey. E-mail: [email protected]

a Department of Food Technology, Vocational School, University of Adıyaman,TR-02040 Adiyaman, Turkey

b Department of Food Engineering, University of Yuzuncu Yıl, 65080 Van, Turkey


26

70

www.soci.org S Kurt, O Zorba

Table 1. Central composite design of two independent variables

Codified level Actual level

Runorder

X1(nitrite)

X2(nisin)

Nitrite(mg kg−1)

Nisin(mg kg−1)

1 −1 −1 0 0

2 −1 0 0 250

3 −1 1 0 500

4 0 −1 100 0

5 0 0 100 250

6 0 0 100 250

7 0 1 100 500

8 1 −1 200 0

9 1 0 200 250

10 1 1 200 500

Table 2. Results of biogenic amine analysis of fresh meat

Biogenic amine (mg kg−1)

Tryptamine 0.00

2-Phenylethylamine 0.00

Histamine 0.00

Spermidine 3.24

Cadaverine 8.41

Tyramine 15.40

Putrescine 17.28

Spermine 27.34

chromatography (HPLC) grade), ammonium acetate, and per-chloric acid from Merck (Darmstadt, Germany) were used in HPLCanalysis. Biogenic amine standards used were 1,7-diaminoheptane(internal standard), spermidine trihydrochloride, spermine tetrahy-drochloride from Sigma (St Louis, MO, USA), and cadaverine dihy-drochloride, histamine dihydrochloride, putrescine dihydrochlo-ride, tryptamine hydrochloride, 2-phenylethylamine hydrochlo-ride, and tyramine hydrochloride from Acros (Acros Organics).Nisin (Nisaplin MS-50) was obtained from Danisco (Niebull, Ger-many). Nitrite was used in the form of sodium nitrite (Merck).

Sucuk preparationSucuk was prepared according to the following recipe:8 84.6%meat (beef), 9.4% lamb tail fat, 1.9% salt, 0.94% garlic, 0.66%red pepper, 0.47% black pepper, 0.85% cumin, 0.24% allspice,0.47% sugar and 0.47% phosphate (K2HPO4; Merck). The meatand fat pieces (∼4 cm3 in size), spices, garlic, salt, sugar, andphosphate were mixed and minced in a grinder (Cem, Turkey).Starter cultures (Lactobacillus sake, Pediococcus pentosaceus,Staphylococcus carnosus, and Staphylococcus xylosus; Bactoferm;Chr. Hansen, Hørsholm, Denmark) were added to the sucuk batterand mixed in. Sucuk batter was divided into 10 equal parts andvarying amounts of nisin and nitrite, which were dissolved in20 mL distilled water, were added to each part as shown in Table 1.Each of the resulting batches of batter was rested for 12 h at4 ◦C and stuffed into collagen casings (Naturin Darm, Weinheim,Germany) of 35 mm diameter using a stuffing machine (Cem,Istanbul, Turkey). Each sample was washed under running water,and then a 10% potassium sorbate solution was sprayed on to it.

Figure 1. Three-dimensional plots of the effects of nitrite and nisin on LAB.

Figure 2. Three-dimensional plots of the effects of nitrite and nisin ontryptamine.

Samples were ripened at 20 ± 1 ◦C for 13 days. For equilibration,relative humidity was adjusted to 60 ± 3% during the first 6 h ofthe ripening period and then increased to 87 ± 3% and decreasedevery day by 1 unit. After the ripening period, samples were storedat −20 ± 1 ◦C until analysis.

Microbiology analysisA 25 g sample was aseptically collected from each sucuk andhomogenised in 225 mL of a sterile salt solution (0.85% NaCl;Merck) using a blender (Waring 80 011S, Torrington, CT, USA). TheLAB number was determined on De Man Rogosa Sharpe Agar(MRS; Oxoid, Basingstoke, UK), which was incubated at 30 ◦C for72 h.

Determination of biogenic aminesThe chromatographic method of Eerola et al.9 was used for thedetermination of tyramine, putrescine, cadaverine, histamine,


26

71

Biogenic amine formation in Turkish dry fermented sausage www.soci.org

Table 3. Analysis of variance of the effects of nitrite and nisin on biogenic amine values of sucuk

Source ofvariation d.f. LAB Tryptamine 2-Phenylethylamine Tyramine Putrescine Cadaverine Histamine Spermidine Spermine

R2 = 0.82 R2 = 0.95 R2 = 0.83 R2 = 0.97 R2 = 0.95 R2 = 0.95 R2 = 0.87 R2 = 0.86 R2 = 0.67

F-value F-value F-value F-value F-value F-value F-value F-value F-value

X1 (nitrite) 1 0.211 51.468∗∗ 11.378∗ 117.031∗∗ 42.639∗∗ 16.331∗ 22.003∗∗ 2.721 5.753

X2 (nisin) 1 7.895∗ 19.410∗ 5.676 1.756 19.607∗ 44.081∗∗ 3.463 12.623∗ 1.292

X1 × X2 1 0.316 3.866 1.156 0.235 10.439∗ 4.222 0.603 0.002 0.185

X1 × X1 1 4.538 1.706 0.352 15.140∗ 0.013 2.354 0.266 0.133 0.021

X2 × X2 1 3.271 5.450 0.748 0.121 4.565 10.352∗ 0.866 8.755∗ 0.943

Lack of fit 3 0.267 33.799 153.558 8.817 10.173 22.013 6.580 11.935 11.765

Corrected total 9

∗∗ P < 0.01 significance level; ∗ P < 0.05 significance level; X1, nitrite (mg kg−1); X3, nisin (mg kg−1); d.f., degrees of freedom.

Table 4. Predicted model equations for the effects of nitrite (X1) andnisin (X2) on biogenic amine values of sucuk

Parameter Predicted model equation

LAB Y = 8.512 + 0.013X1 − 0.082X2 −0.099X1

2 − 0.084X22 + 0.020X1X2

Tryptamine Y = 7.754 − 12.050X1 − 7.400X2 +3.518X1

2 + 6.288X22 + 4.045X1X2

2-Phenylethylamine Y = 5.820 − 3.138X1 − 2.217X2 +0.885X1

2 + 1.290X22 + 1.225X1X2

Putrescine Y = 163.764 − 102.653X1 +69.610X2 − 2.902X1

2 −53.862X2

2 − 62.208X1X2

Tyramine Y = 283.139−65.777X1 −8.057X2 −37.937X1

2 −3.387X22 +3.610X1X2

Cadaverine Y = 343.075 − 86.443X1 +142.020X2 − 52.625X1

2 −110.365X2

2 − 53.828X1X2

Histamine Y = 2.397 − 1.727X1 + 0.685X2 −0.304X1

2 − 0.549X22 − 0.350X1X2

Spermidine Y = 4.332 − 0.087X1 + 0.187X2 +0.031X1

2 − 0.249X22 − 0.003X1X2

Spermine Y = 37.650 + 1.502X1 + 0.712X2 +0.145X1

2 + 0.975X22 − 0.330X1X2

2-phenylethylamine, tryptamine, spermidine, and spermine levels.Amines were separated using HPLC (Agilent 1100, Agilent Tech-nologies, Boeblingen, Germany). The separation was carried out bygradient elution with 0.1 mol L−1 ammonium acetate/acetonitrileon a reverse-phase column (Spherisorb ODS-2; 5 µm, 125 × 4 mm;Waters Corporation, Milford, MA, USA) at a flow rate of 1 mL min−1

using a diode array detector (G1315B DAD, Agilent Technologies)at 254 nm with 550 nm as a reference.

A 4 g sample was weighed into test tube and added 250 µLinternal standard (1.7 diaminoheptane) and 10 mL of 0.4 mol L−1

perchloric acid solution and homogenised with a homogeniser(Pro260, Pro, Oxford, CT, USA). The homogenised sample wascentrifuged (Universal 32R, Hettich International, Tuttlingen,Germany) for 10 min at 2400 × g and rinsed with supernatantinto a 25 mL bottle through filter paper. The extraction wasrepeated with 10 mL of 0.4 mol L−1 perchloric acid solution, mixedthoroughly with a Vortex mixer (Reax Top, Heidolph, Schwabach,Germany) and centrifuged as above. Supernatants were combinedand adjusted to 25 mL with 0.4 mol L−1 perchloric acid solution.

Figure 3. Three-dimensional plots of the effects of nitrite and nisin on2-phenylethylamine.

The alkalinity of a 500 µL sample extract was adjusted using200 µL of 2 mol L−1 sodium hydroxide solution. A 300 µL saturatedsodium bicarbonate was added as a buffer. A 1 mL dansyl chloridesolution (10 mg mL−1 acetone) was added and incubated at 40 ◦Cfor 45 min. Residual dansyl chloride was removed by adding 100 µLammonia (25%). After 30 min, dansylated extract was adjusted to5 mL with 0.1 mol L−1 ammonium acetate/acetonitrile (1 : 1), andfiltered through a 0.45 µm syringe filter (Sartorius, Goettingen,Germany).

Quantities of biogenic amines in the sample were calculated as

Cu = 250 × RF × (HA/Hi) × Ci/WS

where Cu is the unknown (mg kg−1 sample), 250 is the dilutionfactor, RF is the response factor, HA is the peak height of unknown,Hi is the peak height of internal standard, Ci is the concentrationof the internal standard, and WS is the weight of sample.

Experimental design and statistical analysisThe experimental design and statistical analysis were performedusing JMP Software (SAS Institute Inc., Cary, NC, USA). Theexperiments, based on a central composite design with a total


26

72


Figure 4. Three-dimensional plots of the effects of nitrite and nisin ontyramine.

of 10 combinations, including two replicates of the central point,were carried out in random order. The codified and actual levelsare given in Table 1. The variables were coded according to thefollowing equation:

Xi = (xi − xi)/�xi

where Xi is the coded value of an independent variable, xi is thereal value of an independent variable, xi is the real value of anindependent variable at the centre point, and �xi is the stepchange.

The variance for each factor assessed was partitioned into linear,quadratic, and interactive components and were representedusing a second-order polynomial equation. The equation is

Y = β0 +k∑

i=1

βixi +k∑

i=1

βiix2ii +

k∑i=1i<j

k∑j=1

βijxixj

where Y is the estimated response, β0, βi , βii , and βij are constantcoefficients, k is the number of factor variables, and Xi, Xii , andXij represent the linear, quadratic, and interactive effects of theindependent variables, respectively. The analysis was performedusing uncoded units.

RESULTS AND DISCUSSIONThe results of a biogenic amine analysis of fresh meat aresummarised in Table 2. The addition of nitrite and/or nisin affectedsome biogenic amine values, such as values for tryptamine, 2-phenylethylamine, putrescine, cadaverine, histamine, tyramine,and spermidine. The linear effects of nisin were found to besignificant (P < 0.05) on LAB (Table 3). As shown in Fig. 1,the increasing level of nisin decreased counts of LAB, whichmay be the result of the inhibitory effect of nisin on Gram-positive microorganisms.6,10 Dykes et al.11 reported that nisinhad significant effects on Lactobacillus plantarum, Lactobacillusbrevis, and Lactococcus lactis. This result is important becausethese bacteria play an important role in dry fermented sausageproduction.

Figure 5. Three-dimensional plots of the effects of nitrite and nisin onputrescine.

Figure 6. Three-dimensional plots of the effects of nitrite and nisin oncadaverine.

The linear effects of nitrite and nisin on tryptamine valueswere found to be significant (P < 0.01, P < 0.05 respectively;Table 3). Tryptamine values decreased with increasing nitrite ornisin levels (Fig. 2). These results are important, since tryptamineis a vasoactive amine.

The 2-phenylethylamine level decreased in sucuk with theaddition of nitrite, which was statistically significant (P < 0.05)(Table 3; Fig. 3). 2-Phenylethylamine generally occurs when a highlevel of tyramine is present.12 Nitrite was also found to have asimilar effect on tyramine values (Fig. 4). The linear effect of nitriteon tyramine was found to be significant (P < 0.01) (Table 3), andthe quadratic effect was also found to be significant (P < 0.05) fortyramine levels. The increasing rate of nitrite decreased tyraminevalues (Fig. 4), where this may result from the antimicrobialeffects of nitrite on proteolytic organisms, which causes tyramineformation. Higher nitrite levels were more effective than lowerconcentrations at decreasing the rate of tyramine formation


26

73

Biogenic amine formation in Turkish dry fermented sausage www.soci.org

Figure 7. Three-dimensional plots of the effects of nitrite and nisin onhistamine.

(Fig. 4). The findings of previous investigations8,13,14 support thisresult for sucuk production. The effect of nisin was not found toresult in significant effects (P > 0.05) on tyramine levels (Table 3).Espinosa et al.15 similarly found that the addition of nisin to Cluacheese did not reduce the amount of tyramine.

The linear effect of nitrite was found to be significant (P < 0.01)for putrescine (Table 3). The increasing rate of nitrite decreasedputrescine values (Fig. 5). However, nisin increased putrescinevalues (P < 0.05) (Table 3 and Fig. 5). Their interactive effect wasalso found to be significant (P < 0.05) for putrescine values. Someprevious studies reported that the formation of putrescine isclosely linked to the total aerobic viable count.16 – 18 In addition toaerobic bacteria, some LAB and Enterobacter can cause putrescineformation.4,12,19,20 Kurt,8 as well as Bozkurt and Erkmen,21,22

reported that nitrite significantly decreased the counts of totalaerobic mesophilic bacteria in sucuk production. Raju et al.23

reported that nisin at 50 p.p.m. was effective in reducing the totalplate count and spore count in the fish sausage after 2 days ofstorage.

The linear effect of nisin was found to be significant (P < 0.01)on cadaverine levels (Table 3), and the quadratic effect was alsofound to be significant (P < 0.05). The increased rate of nisinresulted in an augmentation of increased cadaverine values, andthis increased rate was higher in the presence of low levelsof nisin rather than in higher levels (Fig. 6). Espinosa et al.15

reported that the addition of nisin increased the amounts ofcadaverine in Clua cheese. However, nitrite decreased cadaverinevalues significantly (P < 0.05) (Fig. 6). The effects of nitriteon cadaverine were supported by previous investigations ofsucuk with similar results.8,13,14 Moreover, nitrite was reportedto decrease cadaverine concentrations in dry fermented Italiansausage.12 Members of Enterobacteriaceae and enterococci playan important role in the formation of cadaverine.12,19,24 Ercoskunet al.16 reported that cadaverine concentrations were directlyrelated to Enterobacteriaceae counts.

Histamine and spermidine were found at lower concentrationsin sucuk (Figs 7 and 8). Nitrite significantly (P < 0.01) decreasedhistamine levels. However, histamine levels were not significantlyaffected by the addition of nisin (Table 3). The linear and quadratic

Figure 8. Three-dimensional plots of the effects of nitrite and nisin onspermidine.

effects of nisin were found to have significant effects (P < 0.05)on spermidine levels (Table 3). In particular, lower levels of nisinincreased spermidine values (Fig. 8). The increasing amount ofspermidine was less than 1 ppm. Spermidine and spermine werenot considered an indicator of spoilage at certain levels, sincethese compounds can be natural components of muscles.25

The effects of nitrite and nisin on LAB count and biogenic aminesare also expressed mathematically in Table 4. These predictedmodel equations are useful for understanding the significance ofamine levels and the interactions between studied factors.

CONCLUSIONDecreases in tryptamine, 2-phenylethylamine, putrescine, cadav-erine, tyramine, and histamine levels demonstrated that nitrite isan indispensable component of sucuk, since the physical, chem-ical, and microbial quality of sucuk is partially related to nitritepresence. However, the additional nitrite levels can be decreasedwith the addition of nisin, which hinders biogenic amine forma-tion. Some antimicrobial agents, which were effective on generaof Gram-negative bacteria, may be considered for utilisation incombination with nisin in dry fermented sausage to prevent theformation of biogenic amines. Thus nisin can play a more ef-fective role in decreasing additional nitrite levels. However, thephenomenon of increasing levels of the some biogenic aminesdue to the addition of nisin needs further study to determine thenisin–biogenic amine interactions.

REFERENCES1 Bover-Cid S, Izquierdo-Pulido M and Vidal-Carou MC, Influence of

hygienic quality of raw materials on biogenic amine productionduring ripening and storage of dry fermented sausages. J Food Prot63:1544–1550 (2000).

2 Gokalp HY, N-Nitroso bile ikleri, kanserojenik etkileri, ce itli gıdalarınN-nitrosamin icerikleri ve ce itli kaynaklardan bunyeye alınan N-nitrosamin miktarları (in Turkish). Gıda 6:317–324 (1984).

3 Halasz A, Barath A, Simon-Sarkadi L and Holzapfel W, Biogenic aminesand their production by microorganisms in food. Trends Food SciTechnol 5:42–49 (1994).

4 Bover-Cid S, Izquierdo-Pulido M and Vidal-Carou MC, Changes inbiogenic amine and polyamine contents in slightly fermented


26

74


sausages manufactured with and without sugar. Meat Sci57:215–221 (2001).

5 Reunanen J and Saris PEJ, Bioassay for nisin in sausage; a shelf lifestudy of nisin in cooked sausage. Meat Sci 66:515–518 (2004).

6 de Martinis ECP, Alves VF and Franco BDGM, Fundamentals andperspectives for the use of bacteriocins produced by lactic acidbacteria in meat products. Food Rev Int 18:191–208 (2002).

7 Singh B, Falahee MB and Adams MR, Synergistic inhibition of Listeriamonocytogenes by nisin and garlic extract. Food Microbiol18:133–139 (2001).

8 Kurt S, The effects of fermentation time, nitrite level and heat treatmenton biogenic amine formation and some properties of sucuk. PhDthesis. Yuzuncu Yıl University, Van, Turkey (2006).

9 Eerola S, Hinkkanene R, Lindfors E and Hirvi T, Liquid chromatographicdetermination of biogenic amines in dry sausages. J AOAC Int76:575–578 (1993).

10 Chen H and Hoover DG, Bacteriosins and their food applications. CompRev Food Sci Food Saf 2:82–100 (2003).

11 Dykes GA, Amarowicz R and Pegg RB, Enhancement of nisinantibacterial activity by a bearberry (Arctostaphylos uva-ursi) leafextract. Food Microbiol 20:211–216 (2003).

12 Suzzi G and Gardini F, Biogenic amines in dry fermented sausages: areview. Int J Food Microbiol 2731:1–14 (2003).

13 Kurt S and Zorba O, The effects of ripening period, nitrite level andheat treatment on biogenic amine formation of ‘sucuk; – a Turkishdry fermented sausage. Meat Sci 82:179–184 (2009).

14 Genccelep H, Kaban G and Kaya M, Effects of starter cultures andnitrite levels on formation of biogenic amines in sucuk. Meat Sci77:424–430 (2007).

15 Espinosa D, de Lamo S, Hernandez M, Lopez T and Roig AX, Formationof biogenic amines in Clua cheese treated by high hydrostatic

pressure, in Health Implications of Dietary Amines: A Joint COSTAction 922 and Biochemical Society Focused Meeting, University ofAberdeen, 19–21 October (2006).

16 Ercoskun H, Con AH and Gokalp HY, Biyojenik aminler ve gıdalardamikroorganizmalarca uretimi (in Turkish). Standard 56–61 (2000).

17 Ruiz-Capillas C and Jimenez-Colmenero F, Biogenic amines in meatand meat products. Crit Rev Food Sci Nutr 44:489–499 (2004).

18 Bozkurt H, Utilization of natural antioxidants: green tea extract andThymbra spicata oil in Turkish dry-fermented sausage. Meat Sci73:442–450 (2006).

19 Shalaby AR, Significance of biogenic amines to food safety and humanhealth. Food Res Int 29:675–690 (1996).

20 Bover-Cid S, Miguelez-Arrizado J and Vidal-Carou MC, Biogenic amineaccumulation in ripened sausages affected by the addition ofsodium sulphite. Meat Sci 59:391–396 (2001).

21 Bozkurt H and Erkmen O, Effect of nitrate/nitrite on the quality ofsausage (sucuk) during ripening and storage. J Sci Food Agric84:279–286 (2004).

22 Bozkurt H and Erkmen O, Effects of some commercial additives onthe quality of sucuk (Turkish dry-fermented sausage). Food Chem101:1482–1490 (2007).

23 Raju CV, Shamasundar BA and Udupa KS, The use of nisin as apreservative in fish sausage stored at ambient (28 ± 2 ◦C) andrefrigerated (6 ± 2 ◦C) temperatures. Int J Food Sci Technol38:171–185 (2003).

24 Montel MC, Masson F and Talon R, Bacterial role in flavourdevelopment. Meat Sci 49:111–113 (1998).

25 Hernandez-Jover T, Izquierdo-Pulido M, Veciana-Nogues MT, Marine-Font A and Vidal-Carou MC, Biogenic amine and polyaminecontents in meat and meat products. J Agric Food Chem45:2098–2102 (1997).


26

75

Research ArticleReceived: 31 May 2010 Revised: 5 July 2010 Accepted: 26 July 2010 Published online in Wiley Online Library: 2 September 2010


Relationship between changes in the totalconcentration of acetic acid bacteria and majorvolatile compounds during the acetic acidfermentation of white wineSilvia Baena-Ruano,a Ines M Santos-Duenas,a Juan C Mauriciob

and Isidoro Garcıa-Garcıaa∗

Abstract

BACKGROUND: In the scope of the wine vinegar production, this paper provides comprehensive information about theevolution of some volatile compounds during the biological acetification cycle. These data were compared with the acidity,cell concentration and ethanol concentration. Such information may allow a better understanding of the complex biologicalprocesses involved.

RESULTS: The volatile compounds 2-phenylethanol, diethyl succinate (diethyl butanedioate), meso-2,3-butanediol (meso-butane-2,3-diol), levo-2,3-butanediol (levo-butane-2,3-diol), methanol and ethyl acetate exhibited no significant changesbetween the starting wine and produced vinegar, whereas the rest [acetoin (3-hydroxybutan-2-one) excepted] ethyl lactate(ethyl 2-hydroxypropanoate), isoamyl alcohols (3-methylbutan-1-ol and 2-methylbutan-1-ol), isobutanol (2-methylpropan-1-ol), 1-propanol (propan-1-ol), and acetaldehyde were consumed in substantial amounts during the process. Additionally, theirspecific evolution patterns alongside bacterial cell concentrations, acidity and ethanol concentration are shown.

CONCLUSION: Concentrations of acetic acid bacteria at the end of the acetification cycle were found to vary because of celllysis, a result of the high acidity and low ethanol concentration of the medium. Variations were similar to those in some volatilecompounds, which suggests their involvement in the metabolism of acetic bacteria. The results testify to the usefulness of thispioneering study and suggest that there should be interest in similar, more detailed studies for a better knowledge of thepresence of certain volatile compounds and metabolic activity in cells effecting the acetification of wine.c© 2010 Society of Chemical Industry

Keywords: wine; acetification; vinegar; volatile compounds; acetic acid bacteria

INTRODUCTIONWine vinegar is an increasingly appreciated product by virtue ofits sensory quality and richness. In fact, today, as much importanceis attached to vinegar as to wine in winemaking areas which havetraditionally obtained the vinegar as a by-product;1 this had ledto the application of strict analytical and control methods to theacetification process with a view to improving the quality of the endproduct, and facilitating its characterisation and discrimination.2,3

Vinegar owes much of its sensory character to its aroma, whichis a combination of the individual contributions of many volatileproducts.4 – 7 The final composition of vinegar in such productsdepends on the particular raw material used, the way the alcoholicfermentation and subsequent biological oxidation are conducted,and the ageing procedure employed, if any. Many vinegars, somewith a designation of origin included, are biologically oxidised inmodern industrial fermentation tanks where the culture mediumis subjected to substantial aeration. Despite the high efficiency ofthe aeration process, there is always the risk of some volatilecompounds being lost and the quality of the end productdiminished as a result. However, using volatile condensers at

the gas output and optimising oxygen input to the reactor duringthe aeration process can help to substantially avoid volatile losses.

Although, as noted earlier, the volatile composition of vinegaris widely variable, it usually includes higher alcohols, esters andsome aldehydes and ketones such as acetoin, acetaldehyde, ethyllactate, 2,3-butanediol, isoamyl alcohols, ethyl acetate, methanoland 2-phenylethanol as major components.8

∗ Correspondence to: Isidoro Garcıa-Garc ıa, Departamento de Ingenier ıaQuımica, Edificio Marie Curie, Facultad de Ciencias, Universidad de Cordoba,Campus Universitario de Rabanales, 14071 Cordoba, Spain.E-mail: [email protected]

a Departamento de Ingenier ıa Quımica, Edificio Marie Curie, Facultad deCiencias, Universidad de Cordoba, Campus Universitario de Rabanales,14071 Cordoba, Spain

b Departamento de Microbiolog ıa, Edificio Severo Ochoa, Facultad de Ciencias,Universidad de Cordoba, Campus Universitario de Rabanales, 14071 Cordoba,Spain


26

76

www.soci.org S Baena-Ruano et al.

Acetaldehyde present in vinegar is an intermediate in thecatabolism of carbon-containing substrates used by aceticacid bacteria.9 This aldehyde is formed by oxidation of bothethanol and pyruvate, which in turn can be produced by theEntner–Doudoroff reaction of sugars, if available, or from lactate,especially at low ethanol levels.10

Acetoin is produced in substantial amounts during theacetification process. In fact, it is often used as a marker for thebiological origin of vinegar since its synthesis is related to cellmetabolism in acetic bacteria.11

Alcohols are also important to vinegar. Thus, isoamyl alcoholsare usually consumed during the acetification of wine.12 On theother hand, levo- and meso-2,3-butanediol in wine are reportedlyoxidised to acetoin,5 but were found to change little in contentwith respect to the starting wine here.

Esters are also important as regards concentration and theirpotential influence on vinegar aroma. Wine vinegar usuallycontains some, such as ethyl lactate, diethyl succinate, ethylacetate, methyl acetate and isoamyl acetate.13 Also, isoamylacetate and ethyl acetate are among the compounds with thehighest odour activity value (OAV) in vinegar.14

Although the volatile composition of vinegar is widelydocumented, its changes during biological acetification of winehave seemingly been studied only once.13 Also, the study inquestion was quite comprehensive and interesting, the authorsonly analysed samples at the beginning, middle and end of theacetification cycle, which was inadequate to assess changes involatile compounds throughout the acetification cycle with a viewto identifying potential alterations in their assumed evolutionpatterns. Recently,15 changes in volatiles in balsamic and red winevinegars stored in wooden casks and bottles were studied, butonly between the start and end of the process.

No study of the potential relationship between volatiles andmicrobial concentration changes during the acetification cycleappears to have been conducted to date. Elucidating such arelationship, if it does exist, would be helpful with a view torelating the synthesis and evolution of these compounds withchemical and biological activity in the acetification system.

In this work, we studied the variations in major volatiles in winevinegar and examined their potential relationship to the totalconcentration of cells throughout the acetification cycle and tovarious other important variables.

EXPERIMENTALMicroorganismThe inoculum used was a mixed culture of acetic bacteriafrom an industrial fermentation tank in full operation. The total

concentration of cells as measured with a method describedelsewhere16 ranged from about 1 × 108 to 3.5 × 108 cells mL−1

during the acetification cycle.

Culture medium and fermentation conditionsThe raw material used was white wine from the Montilla–Morilesregion (Spain) (which is similar to sherry wine) containing89.7 ± 3.9 g ethanol L−1 and having an initial acidity of 4.0 gacetic acid L−1.

The bioreactor employed was operated in a semi-continuousmode (Fig. 1). Thus, once an ethanol concentration of 3.9 g ethanolL−1 was reached, a portion of 75% of the total volume of culturemedium was unloaded and the reactor replenished at a constantflow-rate of 0.01 L wine min−1. This cycle can be repeated anindefinite number of times.

The reactor was a Frings 8 L fermenter and operated at31 ◦C, using an aeration regime of 7.5 L air h−1 L−1 medium.The fermenter was loaded, unloaded and monitored in anautomatic manner according to a programmed sequence. Properoperation in this situation entailed careful on-line measurementof the total ethanol concentration and volume of medium. Anonline probe Alcosens (Heinrich Frings GmbH & Co. KG, Bonn,Germany) and a differential pressure sensor (Yokogawa Iberia S.A.,Madrid, Spain) were used for ethanol and volume determinationrespectively.17

Analysis of volatilesMajor volatile compounds and polyols were quantified on aModel 6890 gas chromatograph from Agilent Technologies (PaloAlto, CA, USA), using the method described by Peinado et al.18

A CP-WAX 57 CB capillary column (60 m long × 0.25 mm i.d.,0.4 µm film thickness) from Varian (Palo Alto, CA) was used,and 0.5 µL aliquots from 10 mL samples previously suppliedwith 1 mL of 1 g L−1 4-methyl-2-pentanol as internal standardwere injected into the instrument. Tartaric acid in the winewas removed by precipitation with 0.2 g of calcium carbonateand centrifugation at 1380 × g. Quantification was based onthe response factors obtained for standard solutions of eachcompound. A split ratio of 30 : 1, an FID, and a temperatureprogram involving an initial temperature of 50 ◦C (15 min), a4 ◦C min−1 ramp and a final temperature of 190 ◦C (35 min)were used. The injector and detector temperatures were 270and 300 ◦C, respectively. The flow rate of carrier gas (helium)was initially set at 0.7 mL min−1 (16 min) and followed by a0.2 mL min−1 ramp to the final value (1.1 mL min−1), which washeld for 52 min.

Figure 1. Stages of a semi-continuous acetification process.


26

77

Relationship between acetic acid bacteria and volatile compounds in wine vinegar www.soci.org

Figure 2. Variation of the concentrations of ethanol and oxygen, as well as the acidity and volume of the medium, in relation to the total number of cellsduring the acetification cycle. Bars represent standard deviations. The corresponding mean standard deviations for ethanol, oxygen and volume wereabout 3%, 2% and 2%, respectively.

Figure 3. Contents in volatiles of the starting wine and the resulting vinegar. Bars represent standard deviations.

RESULTS AND DISCUSSIONFigures 2 to 8 show the experimental data. The results shown inthis work are the means for four cycles. Bars represent standarddeviations.

Figure 2 shows the variation of the main system variablesduring the acetification cycle. As can be seen, the volume andconcentration of ethanol in the medium increased with timeduring the first 10 h (loading stage); by contrast, the concentrationof bacterial cells and the acidity of the medium decreased markedlyby effect of dilution over the same period. The ethanol and acidityvalues at the end of the loading phase are the result of boththe addition of wine as well as the bacteria activity. In fact, aglobal mass balance can shows that, approximately, 10 g L−1 ofacetic acid has been produced during this stage. Acetic bacteria

were subjected to a high stress as a result of abrupt changes intheir environment during the loading stage, which was followedby an adaptation lag stage that lasted 8 h. Then, bacterial cellsentered an exponential growth stage which lasted about 4 h.The last stage of the vinegar production cycle was especiallyinteresting on account of the marked changes undergone by thebacterial cells. In a previous study19 on the variation of aminoacids concentrations during the vinegar production cycle, suchchanges were found to be due to cell lysis phenomena. In fact,each anecdotal decrease in cell concentration was accompaniedby a simultaneous anecdotal increase in the contents of manyamino acids present that was ascribed to cellular autolysis. Therewas also evidence of the opposite changes (i.e. an increase in cellconcentration concomitant with a decrease in amino acid levels).


26

78


Figure 4. Variation of the concentrations of acetoin and ethyl acetate in relation to the total number of cells during the acetification cycle. Bars representstandard deviations.

Figure 5. Variation of the concentrations of isoamyl alcohols, isobutanol and 1-propanol in relation to the total number of cells during the acetificationcycle. Bars represent standard deviations.

These facts may be indicative or even provide evidence for acomplex response of bacterial cells to stressing conditions in orderto ensure survival of their population via ‘programmed cell death’,which allows cells to live on lysis products from dead cells.

The previous results led us to examine changes in variouscompounds with a strong impact on the sensory properties ofvinegar.5 Fig. 3 shows their concentrations in the starting wineand end product (vinegar), as well as the intervening changes.The compounds inside the box exhibited no significant changes,whereas those outside it, acetoin excepted, were consumed insubstantial amounts during the process. Acetoin is known to beinvolved in the biological oxidation of ethanol by acetic bacteria.Therefore, it is present in virtually all types of vinegar, albeit atwidely variable concentrations. Thus, its content in pineapple

vinegar is typically in the region of 2 mg L−1,20 whereas thatin vinegar from sherry wines21 or cider22,23 can be as high as1000 mg L−1 or even higher. Acetoin can form in various ways5,11

including the condensation of two acetaldehyde molecules, thereaction between pyruvate and acetaldehyde, and the oxidationof 2,3-butanediol.

Figure 4 shows the variation of the acetoin concentrationalongside that of the bacterial cell concentration. As can be seen,both evolved virtually in parallel throughout the process. Becauseof its low level in the starting wine, the acetoin concentrationdecreased by effect of dilution during the reactor loading stage.Worthy of special note were the oscillations observed at theend of the process, consistent with changes in the total cellconcentration which, as noted earlier, may be a result of cell lysis


26

79


Figure 6. Variation of the concentrations of ethyl lactate and acetaldehyde in relation to the total number of cells during the acetification cycle. Barsrepresent standard deviations.

Figure 7. Variation of the concentrations of methanol and 2-phenylethanol in relation to the total number of cells during the acetification cycle. Barsrepresent standard deviations.

caused by an increased acidity and a reduced nutrient availabilityin the medium. The concomitance of these oscillations in theacetoin and cell concentrations provides further evidence for arelationship of the synthesis and changes in acetoin to biologicalactivity in the system.

Figure 4 also shows the variation of the ethyl acetate concentra-tions. Although the final content of the vinegar in this compoundwas roughly the same as in the starting wine, this does not excludepotential changes in its concentration during the process. In fact,as can clearly be seen from the figure, this compound exhibitedstrong changes and an early evolution pattern differing markedlyfrom that for acetoin. Judging by the high concentrations of aceticacid and ethanol present in the medium, the ester was most likely

formed by esterification outside bacterial cells. As fresh wine wasadded to the fermenter during the loading stage, the mediumwas supplied with additional ethanol that reacted with acetic acidaccumulating in it to give the ester. Once loading was finished andethanol started to be consumed by acetic bacteria, the reversereaction (hydrolysis of ethyl acetate) gradually prevailed and led toa decrease in the ester concentration. Although the esterificationreaction must occur outside cells, oscillations in cell concentra-tions had a marked effect on the ethyl acetate concentration (forexample, the release of acetate to the medium by effect of celldisruption can displace equilibria to the ester formation). Despitethe subsequent oscillations in cell concentrations, the increasingscarcity of ethanol in the medium led to a more limited esterifica-


26

80


Figure 8. Variation of the concentration of 2,3-butanediol in relation to the total number of cells during the acetification cycle. Bars represent standarddeviations.

tion reaction and prevented further ester concentration rises byeffect of cell lysis.

The four alcohols of Fig. 5 exhibited an increase in concentrationat the start of the cycle as a result of their being supplied during theloading process. Subsequently, the alcohols started to decrease inparallel with the increase in acidity of the medium, which suggeststhat esterification with acetic acid was the main factor governingtheir evolution. However, the small, transient increase at the startof the oscillations in cell concentrations may also indicate that thethree alcohols are involved in the metabolism of acetic bacteria.

As can be seen from Fig. 6, there was a close relationshipbetween cell, lactate and acetaldehyde concentrations. This isunsurprising if one considers the significance of acetaldehydeas an intermediate in the oxidative metabolism of ethanol andvarious other compounds in acetic bacteria.9 Also, there is thewell-known ability of acetic bacteria in using lactate to produceacetate via pyruvate first and acetaldehyde then.

The other volatile compounds (Figs 7 and 8) exhibited nosignificant differences in concentration between the startingwine and the vinegar. Again, there were anecdotal changes inconcentration at the end of the production cycle, which suggeststhe involvement of these compounds in the metabolism of aceticbacteria.

CONCLUSIONIn summary, this paper provides for the first time comprehensiveinformation about the evolution of some volatile compoundsduring the biological acetification cycle. Such information mayallow a better understanding of the complex biological processesinvolved. Thus, the oscillations in cell concentrations observed atthe end of the cycle, which can be ascribed to cell adaptationand survival mechanisms, coincided with similar oscillations inthe concentrations of the target compounds. Their apparentrelationship may be of use to identify the specific compoundsinvolved in the metabolism of acetic bacteria. To our knowledge,no study of this type has been conducted to date.

ACKNOWLEDGEMENTSThe authors are grateful to Spain’s Ministry of Education andScience, and to Grupo SOS, S.A. for funding this research in theframework of Projects AGL2002-01712, PET2006-0827, AGL2005-2494-E-ALI and AGL2009-08117-E-ALI. Co-funding by FEDER is alsogratefully acknowledged.

REFERENCES1 Duran-Guerrero E, Control de los procesos de elaboracion, calidad y

trazabilidad del Vinagre de Jerez. PhD thesis. Puerto Real, Cadiz(2008).

2 Callejon RM, Morales ML, Troncoso AM and Ferreira ACS, Targetingkey aromatic substances on the typical aroma of sherry vinegar.J Agric Food Chem 56:6631–6639 (2008).

3 Tesfaye W, Morales ML, Garcia-Parrilla MC and Troncoso AM,Improvement of wine vinegar elaboration and quality analysis:instrumental and human sensory evaluation. Food Rev Int25:142–156 (2009).

4 Blanch GP, Tabera J, Sanz J, Herraiz M and Reglero G, Volatilecomposition of vinegars – simultaneous distillation extraction andgas-chromatographic mass-spectrometric analysis. J Agric FoodChem 40:1046–1049 (1992).

5 Tesfaye W, Morales ML, Garcıa-Parrilla MC and Troncoso AM, JerezVinegar, in Vinegars of the World, ed. by Solieri L and Giudici P.Springer-Verlag Italia, Milan, pp. 180–195 (2009).

6 Troncoso AM, Evaluacion fisicoquımica y sensorial de vinagres, inSecond Symposium on R+D+I for Vinegar Production, ed. by Garcıa-Garcıa I. Servicio Publicaciones de la Universidad de Cordoba,Cordoba, pp. 195–199 (2006).

7 Duran-Guerrero E, Marin RN, Mejias RC and Barroso CG, Stir barsorptive extraction of volatile compounds in vinegar: validationstudy and comparison with solid phase microextraction.J Chromatogr A 1167:18–26 (2007).

8 Callejon RM, Torija MJ, Mas A, Morales ML and Troncoso AM, Changesof volatile compounds in wine vinegars during their elaborationin barrels made from different woods. Food Chem 120:561–571(2010).

9 Bidan P, Divies C and Cachon R, Microorganismos de alteracion de losvinos, in Enologıa: Fundamentos Cient ıficos y Tecnologicos, ed. byFlanzy C. A Madrid Vicente Ediciones and Ediciones Mundi-Prensa,Madrid, pp. 344–356 (2000).

10 Morales ML, Gonzalez GA, Casas JA and Troncoso AM, Multivariateanalysis of commercial and laboratory produced sherry wine


26

81


vinegars: influence of acetification and aging. Eur Food Res Technol212:676–682 (2001).

11 Polo MC and Sanchez-Luengo AA, Las bacterias aceticas, in El Vinagrede Vino, ed. by Llaguno C and Polo MC. Consejo Superior deInvestigaciones Cientıficas, Madrid, pp. 25–47 (1991).

12 Nieto J, Gonzalez-Vinas MA, Barba P, Martın-Alvarez PJ, Aldave L,Garcıa-Romero E and Cabezudo MD, Recent progress in winevinegar R&D and some indicators for the future, in Food Flavour,Ingredients and Composition, ed. by Charalombous G. ElsevierScience, New York, pp. 469–499 (1993).

13 Morales M, Tesfaye W, Garcia-Parrilla MC, Casas JA and Troncoso AM,Sherry wine vinegar: physicochemical changes during theacetification process. J Sci Food Agric 81:611–619 (2001).

14 Callejon RM, Morales ML, Ferreira ACS and Troncoso AM, Defining thetypical aroma of sherry vinegar: sensory and chemical approach.J Agric Food Chem 56:8086–8095 (2008).

15 Callejon RM, Tesfaye W, Torija MJ, Mas A, Troncoso AM andMorales ML, Volatile compounds in red wine vinegars obtainedby submerged and surface acetification in different woods. FoodChem 113:1252–1259 (2009).

16 Baena-Ruano S, Jimenez-Ot C, Santos-Duenas I, Cantero-Moreno D,Barja F and Garcia-Garcia I, Rapid method for total, viable andnon-viable acetic acid bacteria determination during acetificationprocess. Process Biochem 41:1160–1164 (2006).

17 Garcıa-Garcıa I, Santos-Duenas I, Jimenez-Ot C, Jimenez-Hornero Jand Bonilla-Venceslada J, Vinegar Engineering, in Vinegars of the

World, ed. by Solieri L and Giudici P. Springer-Verlag Italia, Milan,pp. 97–120 (2009).

18 Peinado R, Moreno J, Munoz D, Medina M and Moreno J, Gaschromatographic quantification of major volatile compounds andpolyols in wine by direct injection. J Agric Food Chem 52:6389–6393(2004).

19 Maestre O, Santos-Duenas I, Peinado R, Jimenez-Ot C, Garcia-Garcia Iand Mauricio JC, Changes in amino acid composition during winevinegar production in a fully automatic pilot acetator. ProcessBiochem 43:803–807 (2008).

20 Ou ASM and Chang RC, Taiwan fruit vinegar, in Vinegars of theWorld, ed. by Solieri L and Giudici P. Springer-Verlag Italia, Milan,pp. 223–242 (2009).

21 Morales ML, Tesfaye W, Garcia-Parrilla MC, Casas JA and Troncoso AM,Evolution of the aroma profile of sherry wine vinegars during anexperimental aging in wood. J Agric Food Chem 50:3173–3178(2002).

22 Llaguno C, Definicion y tipos de vinagre, in El Vinagre de Vino, ed.by Llaguno C and Polo MC. Consejo Superior de InvestigacionesCientıficas, Madrid, pp. 133–145 (1991).

23 Valero E, Berlanga T, Roldan P, Jimenez C, Garcia I and Mauricio JC,Free amino acids and volatile compounds in vinegars obtainedfrom different types of substrate. J Sci Food Agric 85:603–608(2005).


26

82

Research ArticleReceived: 2 November 2009 Revised: 28 April 2010 Accepted: 26 July 2010 Published online in Wiley Online Library: 14 October 2010


Chemical composition, antioxidantand antibacterial properties of the essential oilsof Etlingera elatior and Cinnamomumpubescens KochummenSiddig Ibrahim Abdelwahab,a∗,† Faridah Qamaruz Zaman,b,c†

Abdalbasit Adam Mariod,d Muhammad Yaacob,c Adil Hassan AhmedAbdelmageedc and Shamsul Khamisc

Abstract

BACKGROUND: Plant essential oils are widely used as fragrances and flavours. Therefore, the essential oils from the leaves ofCinnamomum pubescens Kochummen (CP) and the whole plant of Etlingera elatior (EE) were investigated for their antioxidant,antibacterial and phytochemical properties.

RESULTS: CP and EE were found to contain appreciable levels of total phenolic contents (50.6 and 33.41 g kg−1 as gallicacid equivalent) and total flavonoid contents (205.6 and 244.8 g kg−1 as rutin equivalent), respectively. DPPH free radicalscavenging activity of CP is superior to EE (P < 0.05) showing IC50 of 77.2 and 995.1 µg mL−1, respectively. Methicillin-resistantStaphylococcus aureus (MRSA), Bacillus subtilis, Pseudomonas aeruginosa and Salmonella choleraesuis were tested against CP andEE. Only MRSA was the most susceptible bacteria to CP. GC/MS studies resulted in the identification of 79 and 73 compounds inCP and EE, respectively. The most abundant components of EE included β-pinene (24.92%) and 1-dodecene (24.31%). While themajor compound in CP were 1,6-octadien-3-ol,3,7-dimethyl (11.55%), cinnamaldehyde (56.15%) and 1-phenyl-propane-2,2-dioldiethanoate (11.38%).

CONCLUSION: This study suggests that the essential oils from Cinnamomum pubescens Kochummen and Etlingera elatior couldbe potentially used as a new source of natural antioxidant and antibacterial in the food and pharmaceutical industries.c© 2010 Society of Chemical Industry

Keywords: antibacterial activities; antioxidant; chemical composition; Cinnamomum pubescens Kochummen; Etlingera elatior

INTRODUCTIONEssential oils from herbal sources are used in food flavours,perfumes and pharmaceutical preparations for their functionalproperties.1 The commercial use of essential oils in aromatherapyconstitutes little more than 2.0% of the total market.2 Moreover,the antibacterial properties of these herbal essential oils andtheir components are exploited in such diverse commercialproducts as dental root canal sealers, antiseptics and animalfeed supplements.3 Besides the antibacterial properties, some ofessential oils were proven to possess antioxidant properties.4,5

Williams and Harborne screened 39 species of ginger (Zingib-eraceae) for their phytochemical constituents. Leaves of Alpiniaand Zingiber were found to contain kaempferol and quercetin gly-cosides, and myricetin and quercetin glycosides, respectively.6

Flavonoids in the leaves of Etlingera elatior (Zingiberaceae)have been identified as kaempferol 3-glucuronide, quercetin 3-glucuronide, quercetin 3-glucoside, and quercetin 3-rhamnoside.Members of the Etlingera genus have various remedial and com-mercial uses. Young shoots, flowers and fruits are eaten either raw,cooked as a vegetable, or used as a condiment.7 Inflorescences

of E. elatior are widely cultivated as spice for curry.8 Fruits areused to treat ear itching, while leaves are applied for healingwounds.7 E. elatior was also found to have anti-tumour promotingand cytotoxic activities.9,10

∗ Correspondence to: Siddig Ibrahim Abdelwahab, Department of Pharmacy,Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia.E-mail: [email protected]

† S.I. Abdelwahab and F.Q. Zaman contributed equally to this paper.

a Department of Pharmacy, Faculty of Medicine, University of Malaya, KualaLumpur, Malaysia

b Department of Biology, Faculty of Science, Universiti Putra Malaysia, Serdang,Malaysia

c Biodiversity Unit, Laboratory of Natural Products, Institute of Bioscience,Universiti Putra Malaysia, Serdang, Malaysia

d Food Science & Technology Department, College of Agricultural Studies, SudanUniversity of Science & Technology, Khartoum north, Sudan


26

83

Composition and properties of oils from E. elatior and C. pubescens www.soci.org

Cinnamomum pubescens Kochummen (family: Lauraceae) is asmall tree up to 7 m tall. It is indigenous to Peninsular Malaysia;the leaves are opposite to alternate, with a hairy stalk 1–2 cm inlength. The blade is leathery, drying greenish yellow, lanceolate,apex pointed, base cuneate, finely pubescent on the under surface,midrib and secondary nerves flattened to sunken on the uppersurface, raised on the lower surface.11,12 As with all cinnamomumspecies, the whole parts of the tree are aromatic and are used intraditional medicine as medang, but this species has not been usedcommercially.13 The leaf oil of C. pubescens showed significantlarvacidal and platelet-activating factor (PAF) receptor-bindingantagonist activities.14

In this present study, the essential oils obtained from thewhole plant of E. elatior and leaves of C. pubescens Kochummen(referred to as C. pubescens for the remainder of this paper)were analysed by GC/MS and their antioxidant activities usingdiphenylpicrylhydrazyl (DPPH) free-radical scavenging activity,antibacterial properties and their chemical composition werecompared. To the best of our knowledge, this study is the first ofits kind to report the antioxidant activities and antibacterial effectof essential oils from E. elatior and C. pubescens.

MATERIALS AND METHODSAll solvents used were of analytical grade. Methanol, ethylacetate, hexane, chloroform, butylated hydroxytoluene (BHT) andFolin–Ciocalteu reagent were obtained from Merck (Darmstadt,Germany).

Plant materialsFresh leaves of C. pubescens and whole plant of E. elatior werecollected from Pahang and Selangor states, Malaysia, respectively,in 2009. Plants were identified by Assistant Professor ShamsulKhamis at the Unit of Biodiversity, Institute of Bioscience, UniversitiPutra Malaysia, Malaysia. The voucher specimens under the plants’names were deposited in the unit herbarium.

Isolation procedure for the essential oilsThe fresh whole plant of E. elatior and leaves of C. pubescens,were steam distilled, separately, in a hydrodistillation apparatus(Clevenger-type) for 8 h. The essential oils were dried overanhydrous sodium sulfate and stored at 4–6 ◦C before analysis.

Antioxidant activityDetermination of total phenolic contentTotal phenolic content (TPC) in essential oils of E. elatior andC. pubescens was determined with Folin–Ciocalteu reagentfollowing the method of Kaur et al.15 Stock solutions of oilswere prepared in a concentration of 10 mg mL−1, and a 50 µLfrom this solution was transferred to a test tube (n = 3). Tothis tube, 0.4 mL of Folin–Ciocalteu reagent (1 : 10) was addedand the tube was shaken thoroughly. After 1 min, 0.8 mL ofsodium bicarbonate solution (0.9 mol L−1) was added and themixture allowed standing in dark room for 30 min with intermittentshaking. Absorbance was measured at 765 nm using a ShimadzuUV–visible spectrophotometer (Mini 1240; Shimadzu, Columbia,MD, USA). The total phenolic content (TPC) was expressed asgallic acid equivalent (GAE) in mg per g oil from the calibrationcurve of gallic acid standard solution. For the gallic acid, thecurve was established by plotting concentration (mg mL−1) versusabsorbance (nm) (y = 5.145x + 0.014; R2 = 0.9975), where y isthe absorbance and x is the concentration.

Determination of total flavonoid contentTotal flavonoid content (TFC) was determined by the AlCl3method, using rutin as a standard.16 The test samples weredissolved in dimethyl sulfoxide (DMSO). The sample solution(1.0 mL) was mixed with 1.0 mL of AlCl3 (0.15 mol L−1). After10 min of incubation at ambient temperature, the absorbanceof the supernatant was measured at 435 nm using a ShimadzuUV–visible spectrophotometer (Mini 1240). Three replicates weremade for each test sample. The total flavonoid content wasexpressed as rutin equivalents (RE, mg g−1). For the rutin, thecurve was established by plotting concentration (mg mL−1) versusabsorbance (nm) (y = 5.6752x − 0.0312; R2 = 0.994), where y isthe absorbance and x is the concentration.

DPPH radical scavenging antioxidant assayRadical scavenging activity of plant essential oils against sta-ble DPPH (2, 2-diphenyl-2-picrylhydrazyl hydrate) (Sigma–AldrichChemie, Steinheim, Germany) was determined spectrophotomet-rically. When DPPH reacts with an antioxidant compound, whichcan donate hydrogen and is reduced, the changes in colour(from deep-violet to light-yellow) were measured at 517 nmwavelength.17 Radical scavenging activity of essential oils wasmeasured by a slight modification of the method by Ao et al.18

Stock solutions were prepared in 10 mg mL−1 in methanol. Theworking solution was prepared using methanol in a concen-tration of 2.0 mg mL−1 (Labsystems iEMS Reader MF; Eichen-weg, Aumuhle, Germany). The solution of DPPH in methanol(1 mmol L−1) was freshly prepared, before UV measurements.Five microlitres of this solution were mixed with 100 µL of serialdilutions of samples (15.625–2000 µg mL−1) in a 96-well plate.The samples were kept in the dark for 30 min at ambient tem-perature and then the decrease in absorption was measuredeach 30 min for 2 h. Absorption of a blank sample containingthe same amount of methanol and DPPH solution was preparedand measured daily. The experiment was carried out in tripli-cate. Radical scavenging activity was calculated by the following% inhibition = [(AB − AA)/AB] × 100, where AB is the absorp-tion of blank sample (t = 0 min); and AA is the absorption oftested samples (t = 30 min). The inhibitory concentration 50%was determined as well as the kinetics of DPPH scavenging reac-tion. Commercial standard antioxidant butylated hydroxytoluene(BHT) was also tested against DPPH and used as a reference.

Antibacterial assayMicrobial strainsThe antibacterial activity of essential oil samples was evaluated us-ing two Gram-positive bacteria, methicillin resistant Staphylococ-cus aureus (MRSA) and Bacillus subtilis B29, and two Gram-negativebacteria, Pseudomonas aeruginosa 60 690 and Salmonella choler-aesuis. All bacterial strains were obtained from the Laboratoryof Molecular Biomedicine, Institute of Bioscience, Universiti PutraMalaysia, Serdang, Malaysia.

Disc diffusion methodScreening for the antibacterial effect of the essential oils was carriedout by determining the zone of inhibition using paper disc (6 mm indiameter, Whatman No. 1) diffusion method.19 The microorganismstrains obtained were inoculated in a Petri dish containing nutrientbroth at 37 ◦C for 24 h and were referred as seeded broth. Thedensity of the bacterial suspension was standardised and theconcentrations of the cultures were adjusted turbidometrically at


26

84

www.soci.org SI Abdulwahab et al.

wavelength of 600 nm to 5 × 105 to 106 colony forming unitsper mL. The essential oils were dissolved in DMSO which waspreviously tested for antibacterial activity against all test bacteriaand found to have no activity. Essential oils were diluted to aconcentration of 100 mg mL−1 and finally sterilised by filtrationusing 0.45 µm Millipore filters. The sterile discs were impregnatedwith oil solution (0.05 mL from 100 mg mL−1) to achieve desiredconcentration and placed in inoculated agar. Streptomycin(10 µg mL−1) susceptibility discs and methanol-impregnated discswere used as positive and negative controls, respectively. Afterincubation overnight at 37 ◦C, inhibition zones were measuredand recorded as mean diameter (mm). Antibacterial activity wasalso expressed as inhibition percentage of streptomycin.

Minimum inhibitory concentrationThe least possible inhibitory concentrations of essential oilsagainst MRSA were estimated using the agar disc method (ADM).Inoculation of 1.0 mL of MRSA was poured into each Petri dishand the agar was later dispensed and permitted to set. Wells werebored using a sterile 3.0 mm cork borer. Serial dilutions of theessential oil were added into the wells. The plates were incubatedat 37 ◦C for 24 h. The growth was observed to determine thesensitivity of MRSA using clear zones of no microbial growth. Theleast concentration of the essential oil that had inhibitory effectwas taken as the minimum inhibitory concentration (MIC).

Gas chromatography mass spectrometryThe essential oils of E. elatior and C. pubescens were analysedby Shimadzu GC-MS (Model GC-17A). A FT-DB-5 capillary column(30 m×0.25 mm× i.d. 0.25 µm) was used for gas chromatographicseparation of the analytes.20 The injection volume was 1.0 µL witha split ratio of 13 : 1; the injector temperature was held constant at230 ◦C. Helium was used as the carrier gas with an inlet pressureof 21.0 kPa, corresponding to a flow rate of 1.0 mL min−1. Thecolumn oven temperature was set at 30 ◦C (held for 3 min), raisedat 8 ◦C min−1 to 230 ◦C (held for 5 min), and finally held at 245 ◦Cfor 10 min. The mass spectrometer was operated in the electronimpact (EI) mode with ionisation energy of 70 eV. The transferline was set at 290 ◦C. The chemical constituents of the analyteswere identified by comparing the MS fragmentation patterns withthose of NIST/EPA/NIH mass special database library of the GC/MSsystem.

Statistical analysesIn order to determine whether there is a statistically significantdifference between the obtained results for the different assays theindependent t-test was carried out using the SPSS 17.0 softwarepackage.

RESULTS AND DISCUSSIONAntioxidant capacityHydro-distillation of fresh leaves of C. pubescens and the wholeplant of E. elatior, afforded colourless pleasant-smelling essentialoils. The total phenolic content (TPC) of these essential oils wasdetermined using the Folin–Ciocalteu method and expressed inmg GAE g−1. Results presented in Table 1 showed that C. pubescens(50.6 ± 0.58 g GAE kg−1) had higher TPC when compared to E.elatior (33.41 ± 0.92 g GAE kg−1). Independent t-test statisticalanalysis showed the mean TPC of C. pubescens is significantlydifferent (n = 3) at the 0.05 level of significance. Total flavonoid

Table 1. Total phenolic content, total flavonoid content and DPPHIC50 (µg mL−1) of Etlingera elatior and Cinnamomum pubescensKochummen

Sample

Total flavonoidcontent

(RE g kg−1)

Total phenoliccontent

(GAE g kg−1)DPPH IC50(µg mL−1)

E. elatior 244.83 ± 15.5 3341.2 ± 92.1 995.1 ± 123

C. pubescens 205.65 ± 30.4 5060.5 ± 58.6 77.2 ± 8.5

Results are expressed as average ± SD (n = 3).RE, rutin equivalent; GAE, gallic acid equivalent.

content (TFC) was determined by the AlCl3 method. Results wereexpressed as milligrams of rutin equivalent (RE) in one gram of oil(Table 1). The TFC for E. elatior was observed to be 244.83 ± 15.5 gRE kg−1, which is not statistically different from the TFC for C.pubescens (205.65 ± 30.4 g RE kg−1) (Table 1). The current studyshows that the essential oil from E. elatior has high TPC and TFC.However, it was reported earlier that a methanolic extract (not theessential oil) from fresh leaves of E. elatior has also shown highTPC.7 Antioxidant activity of essential oils from E. elatior and C.pubescens was also evaluated using the DPPH assay, which wasconducted for 240 min (Fig. 1). Samples were able to reduce violetDPPH to the yellow DPPH-H, with an IC50 of 995.1 ± 123 and77.2 ± 8.5 µg mL−1 for C. pubescens and E. elatior, respectively(Table 1). No earlier reports are available regarding the DPPHradical scavenging activity of the essential oils of E. elatior andC. pubescens with which to compare the results of our presentanalysis. However, Chan et al., reported the IC50 for a methanolicextract of fresh leaves of E. elatior to be 37.5 mg kg−1.7

The higher TPC of C. pubescens may explain its superioritycompared to E. elatior. In addition, DPPH protection by C. pubescensmay be due to the antioxidative action of eugenol (7.27%),which has been detected by GC-MS. This phenolic compound hasbeen detected earlier in the essential oils of eight Cinnamomumspecies.21

Despite the fact that E. elatior essential oil has a higherflavonoid content (Table 1), significant contributors to the highantiradical effect, it did not show a scavenging effect comparedto C. pubescens. These surprising results might be explained bythe existence of β-pinene (monoterpene hydrocarbons). Someisolated terpenes have been previously tested individually in orderto determine the antioxidant nature of the oils, such as β-pinene,but none has exhibited antioxidant activity.22 To the best of ourknowledge, no data have been published on antioxidant activity,using the DPPH method, on E. elatior and C. pubescens essentialoils.

Antibacterial activitiesAmong the tested bacteria, only MRSA was the most sensitiveorganism to the essential oils of C. pubescens. The mean diameterof the zone of inhibition of C. pubescens was 15 mm for MRSAwhich represents 75% inhibition compared to streptomycinMRSA (Table 2). The essential oil of C. pubescens was reportedearlier to be antifungal.21 Cinnamaldehyde (the major constituentof C. pubescens, 56.15%) is known to be antibacterial againstEscherichia coli and Salmonella typhimurium; it did not disintegratethe outer membrane or deplete the intracellular ATP pool.23

The carbonyl group of some kind of potential antibacterialagents is thought to bind to proteins, preventing the action


26

85


Figure 1. Antioxidant activity of Etlingera elatior (upper photo), Cinnamomum pubescens (lower photo).

Table 2. Antibacterial activity of essential oils of Etlingera elatior and Cinnamomum pubescens Kochummen against bacteria using the disc-diffusionmethod and minimum inhibitory concentration

Methicillin resistantStaphylococcus aureus

Pseudomonasaeruginosa

Salmonellacholeraesuis Bacillus subtilis

Sample

Inhibitionzone(mm)

MIC(mg

mL−1)

Inhibitionzone(mm)

MIC(mg

mL−1)

Inhibitionzone(mm)

MIC(mg

mL−1)

Inhibitionzone(mm)

MIC(mg

mL−1)

Etlingera elatior 15 (75) 10 – – – – – –

Cinnamomum pubescens Kochummen – – – – – – – –

Control (streptomycin) 20 – 20 – 23 – 23 –

Methanol – – – – – – – –

a The screening of the essential oils antibacterial effect was carried out by determining the zone of inhibition using paper disc (6 mm in diameter,Whatman No. 1) diffusion method (n = 2). Figures in parentheses are inhibition percentages compared to streptomycin.MIC: Minimum inhibitory concentration.


26

86


of amino acid decarboxylases in Enterobacter aerogenes.24 Gram-positive Bacillus subtilis was observed to be resistant for both E.elatior and C. pubescens. Previous studies on the antimicrobialactivity of essential oils obtained from ginger species alsoshowed weak inhibition of bacteria.25 The mean diameter ofthe zone of inhibition of streptomycin was 20 mm for MRSA andPseudomonas aeruginosa and 23 mm for Salmonella choleraesuisand B. subtilis (Table 2). The solvent used to prepare the referenceand test samples showed no inhibitory effect on the bacteriaused.

Essential oils of E. elatior and C. pubescens failed to inhibitGram-negative S. choleraesuis and P. aeruginosa (Table 2). Selectivepermeability of the outer membrane of Gram-negative bacteriamakes it generally less susceptible to volatile oils than theGram-positive bacteria. Gram-negative P. aeruginosa is known tohave a high level of intrinsic resistance to virtually almost all knownantimicrobials and antibiotics, due to a very restricted outer mem-brane barrier, highly resistant even to synthetic drugs.26 To thebest of our knowledge, this is the first study reporting antibacterialactivities of the essential oils of E. elatior and C. pubescens.

Table 3. Compounds tentatively identified in the essential oil of Cinnamomum pubescens Kochummen

Compound no. RTa RCb (%) Compoundc Molecular weight Similarity (%)

1 14.132 0.94 Bicyclo[3.1.1]hept-2-ene, 2,6,6-trimethyl 136 95

2 15.019 0.71 Camphene 136 94

3 16.377 0.77 Benzaldehyde 106 94

4 16.475 0.51 β-Pinene 136 88

5 19.408 0.72 Eucalyptol 154 85

6 23.132 11.55 1,6-Octadien-3-ol,3,7-dimethyl 154 95

7 26.547 1.40 Benzene propanal 134 95

8 26.731 1.13 Borneol 154 94

9 27.870 0.82 (−)-α-Terpineol (p-menth-1-en-8-ol) 154 94

10 32.470 56.15 Cinnamaldehyde 132 98

11 35.476 7.27 Phenol, 2-methoxy-4-(2-propenyl)-, acetate (eugenol) 206 76

12 39.822 11.38 1-Phenyl-propane-2,2-diol diethanoate 236 77

12 out of 79 Total 93.35

a RT, Retention time (min).b Relative area percentage (peak area relative to the total peak area percentage).c Compounds are listed in order of their relative area percentage.

Table 4. Compounds tentatively identified in the essential oil of Etlingera elatior

Compound no. RTa RCb (%) Compoundc Molecular weight Similarity (%)

1 14.151 11.59 Bicyclo[3.1.1]hept-2-ene, 2,6,6-trimethyl 136 97

2 16.528 24.92 β-Pinene 136 96

3 17.206 0.60 β-Myrcene 136 91

4 26.919 0.86 Bicyclo[3.1.1]heptan-3-one, 2,6,6-trimethyl 152 92

5 27.908 0.66 (−)-α-Terpineol (p-menth-1-en-8-ol) 154 86

6 27.917 0.74 7-Methylene-9-oxabicyclo[6.1.0]non-2-ene 136 78

7 28.325 0.69 Decanal 156 94

8 32.448 0.74 2-Undecanone 170 95

9 32.639 1.38 3-Bromo-7-methyl-1-adamantane carboxylic acid 272 78

10 37.937 8.15 Dodecanal 184 97

11 38.074 2.49 β-Farnesene 204 87

12 39.554 2.41 1,6,10-Dodecatriene, 7,11-dimethyl-3-methylene 204 91

13 39.744 1.99 α-Caryophyllene 204 94

14 40.935 24.31 1-Dodecene 168 96

15 41.675 0.90 2-Tridecanone 198 96

16 45.447 2.56 trans-(Z)-α-Bisabolene epoxide 220 86

17 46.405 3.49 Acetic acid 228 87

18 46.663 1.22 2-Pentadecyn-1-ol 224 83

19 48.837 0.63 (E)-10-Pentadecenol 226 92

20 49.334 2.27 1,3-Propanediol, 2-dodecyl 244 93

20 out of 73 Total 92.6%

a RT, Retention time (min).b Relative area percentage (peak area relative to the total peak area percentage).c Compounds are listed in order of their relative area percentage.


26

87


Figure 2. Total ionic chromatogram (GC-MS) of essential oils of Etlingera elatior (A) and Cinnamomum pubescens Kochummen (B) obtained with 70 eVusing a A FT-DB-5 capillary column (30 m × 0.25 mm × i.d. 0.25 µm). Helium was used as the carrier gas with an inlet pressure of 21.0 kPa, correspondingto a flow rate of 1.0 mL min−1.

Chemical compositionThe chemical composition of E. elatior and C. pubescens es-sential oils studied by GC-MS is presented in Tables 3 and 4.Total ionic chromatograms of both E. elatior and C. pubescensare presented in Fig. 2. A total of 79 compounds were char-acterised in C. pubescens while the essential oil of E. elatiorconsists of 73 compounds. The most abundant components inthe leaf essential oil of E. elatior included bicyclo[3.1.1] hept-2-ene,2,6,6-trimethyl (11.59%), β-pinene (24.92%), 3-bromo-7-methyl-1-adamantanecarboxylic acid (1.38%), dodecanal (8.15%), β-farnesene (2.49%), 1,6,10,-dodecatriene, 7,11-dimethyl-3-meth-ylene (2.41%), α-caryophyllene (1.99%), 1-dodecene (24.31%),trans-(Z)-α-bisabolene epoxide (2.56%), acetic acid (3.49%)and 1,3-propanediol, 2-dodecyl (2.27%) (Table 3). 1,6-Octadien-3-ol,3,7-dimethyl (11.55%), borneol (1.13%), cinnamaldehyde(56.15%), phenol, 2-methoxy-4-(2-propenyl)-, acetate (7.27%) and1-phenyl-propane-2,2-diol diethanoate (11.38%) were the majorcompounds in C. pubescens (Table 3).

Bicyclo [3.1.1] hept-2-ene, 2,6,6-trimethyl was found to be moreabundant in the essential oil of E. elatior (11.59%) compared toC. pubescens (0.94%), while β-pinene is abundant in the latter

(24.92%) compared to C. pubescens (0.51%) (Tables 3 and 4).Caryophyllene was reported previously in E. elatior and C.pubescens; however, the current study revealed the absence ofthis sesquiterpene hydrocarbon in C. pubescens.25

Previously, Jaafar et al. analysed the essential oils isolated fromdifferent parts (leaves, stems, flowers and rhizomes) of MalaysianE. elatior using GC-MS.9 The leaf essential oil was found to containβ-pinene (19.7%), β-caryophyllene (15.4%) and trans-β-farnesene(27.1%) as the major compounds whereas the stem essentialoil was largely dominated by 1,1-dodecanediol diacetate (34.3%)and trans-5-dodecene (27.0%). The essential oils of the flowers andrhizomes contained 1,1-dodecanediol diacetate (24.4% and 40.4%,respectively) and cyclododecane (47.3% and 34.5%, respectively)as the major compounds. The current study demonstrated ahigher percentage of β-pinene and lower for α-caryophyllene andtrans-β-farnesene.

CONCLUSIONSEssential oils of C. pubescens and E. elatior have significantdifferences in their chemical composition and antibacterial


26

88


activities. C. pubesecens essential oil showed an interestingantibacterial effect against methicillin resistant Staphylococcusaureus, a bacterium responsible for difficult-to-treat infectionsin humans. In view of the antioxidant properties of these twoessential oils they might be considered for inclusion as naturalantioxidants in nutraceutical and pharmaceutical preparations.

REFERENCES1 Surburg H, Panten J and Bauer K, Common Fragrance and Flavor

Materials: Preparation, Properties and Uses. Vch VerlagsgesellschaftMbh, Weinheim, West Germany (2006).

2 Burt S, Essential oils: their antibacterial properties and potentialapplications in foods – a review. Int J Food Microbiol 94:223–253(2004).

3 Doel MA and Segrott J, Materializing complementary and alternativemedicine: aromatherapy, chiropractic, and Chinese herbal medicinein the UK. Geoforum 35:727–738 (2004).

4 Kelen M and Tepe B, Chemical composition, antioxidant and anti-microbial properties of the essential oils of three Salvia species fromTurkish flora. Bioresour Technol 99:4096–4104 (2008).

5 Kivrak , Duru ME, Ozturk M, Mercan N, Harmandar M and Topcu G,Antioxidant, anticholinesterase and antimicrobial constituents fromthe essential oil and ethanol extract of Salvia potentillifolia. FoodChem 116:470–479 (2009).

6 Williams CA and Harborne JB, The leaf flavonoids of the Zingiberales.Biochem Syst Ecol 5:221–229 (1977).

7 Chan EWC, Lim YY and Omar M, Antioxidant and antibacterial activityof leaves of Etlingera species (Zingiberaceae) in Peninsular Malaysia.Food Chem 104:1586–1593 (2007).

8 Lemmens R, Soerianegara I and Wong WC, Plant Resources of South-east Asia No. 5 (2). Timber Trees: Minor Commercial Timbers. BackhuysPublishers, Leiden (1995).

9 Jaafar FM, Osman CP, Ismail NH and Awang K, Analysis of essential oilsof leaves, stems, flowers and rhizomes of Etlingera elatior (Jack) RMSmith. Malaysian J Anal Sci 11:269–273 (2007).

10 Habsah M, Ali AM, Lajis NH, Sukari MA, Yap YH and Kikuzaki H,Antitumor promoting and cytotoxic constituents of Etlingera elatior.Malaysian J Med Sci 12:6–12 (2005).

11 Verheij EWM and Coronel RE, Plant Resources of South-East Asia. Pudoc,Jakarta, Kuala Lumpur (1991).

12 Kochummen KM, Family: Lauraceae, Tree Flora of Malaya, Vol. 4.Longmans, Kuala Lumpur (1989).

13 Burkill IH and Birtwistle W, A Dictionary of the Economic Products ofthe Malay Peninsula. Published on behalf of the Governments of

the Straits Settlements and Federated Malay States by the CrownAgents for the Colonies, London (1935).

14 Jantan I, Rafi IAA and Jalil J, Platelet-activating factor (PAF) receptor-binding antagonist activity of Malaysian medicinal plants. Phytomed12:88–92 (2005).

15 Kaur R, Arora S and Singh B, Antioxidant activity of the phenol richfractions of leaves of Chukrasia tabularis A. Juss. Bioresour Technol99:7692–7698 (2008).

16 Quettier-Deleu C, Gressier B, Vasseur J, Dine T, Brunet C and Luyckx M,Phenolic compounds and antioxidant activities of buckwheat(Fagopyrum esculentum Moench) hulls and flour. J Ethnopharmacol72:35–42 (2000).

17 Brand-Williams W, Cuvelier ME and Berset C, Use of a free radicalmethod to evaluate antioxidant activity. LWT-Food Sci Technol28:25–30 (1995).

18 Ao C, Li A, Elzaawely AA, Xuan TD and Tawata S, Evaluation ofantioxidant and antibacterial activities of Ficus microcarpa L. fil.extract. Food Cont 19:940–948 (2008).

19 Sahoo S, Kar DM, Mohapatra S, Rout SP and Dash SK, Antibacterialactivity of Hybanthus enneaspermus against selected urinary tractpathogens. Indian J Pharm Sci 68:653 (2006).

20 Ibrahim H, Aziz AN, Syamsir DR, Ali NAM, Mohtar M and Ali RM,Essential oils of Alpinia conchigera Griff. and their antimicrobialactivities. Food Chem 113:575–577 (2009).

21 Jantan I, Karim Moharam BA, Santhanam J and Jamal JA, Correlationbetween chemical composition and antifungal activity of theessential oils of eight Cinnamomum species. Pharma Biol46:406–412 (2008).

22 Marin R, Apel MA, Limberger RP, Raseira MCB, Pereira JFM andZuanazzi JAS, Volatile components and antioxidant activity fromsome Myrtaceous fruits cultivated in Southern Brazil. Lat Am J Pharm27:172–177 (2008).

23 Friedman M, Kozukue N and Harden LA, Cinnamaldehyde contentin foods determined by gas chromatography-mass spectrometry.J Agric Food Chem 48:5702–5709 (2000).

24 Wendakoon CN and Sakaguchi M, Inhibition of amino acid de-carboxylase activity of Enterobacter aerogenes by active com-ponents in spices. J Food Protect 58:280–283 (1995).

25 Mackeen MM, Ali AM, El-Sharkawy SH, Manap MY, Salleh KM andLajis NH, Antimicrobial and cytotoxic properties of some Malaysiantraditional vegetables (ulam). Pharma Biol 35:174–178 (1997).

26 Mann CM, Cox SD and Markham JL, The outer membrane ofPseudomonas aeruginosa NCTC 6749 contributes to its toleranceto the essential oil of Melaleuca alternifolia (tea tree oil). Lett ApplMicrobiol 30:294–297 (2000).


26

89

Research ArticleReceived: 9 June 2010 Revised: 21 July 2010 Accepted: 28 July 2010 Published online in Wiley Online Library: 17 September 2010


Morphological and qualitative characterisationof globe artichoke head from newseed-propagated cultivarsAnna Bonasia,a∗ Giulia Conversa,a Corrado Lazzizera,a

Giuseppe Gambacortab and Antonio Eliaa

Abstract

BACKGROUND: Three new artichoke seed-propagated hybrids (Tempo, Opal and Madrigal) were compared with two standardcultivated varietal types [Catanese and Violet du Provence (VP)] in terms of head morphology, processing performance,nutritional or technological qualitative traits, in order to define their best use.

RESULTS: Compared to the other genotypes, Opal and Madrigal had more rounded, heavier, larger heads, higher processingyield (>400 g of heart kg−1 raw head) and lower total phenol (TP) content (2.4 g of gallic acid equivalents kg−1 FW). VP gave ahigher processing yield than Catanese and showed the highest TP content (6.5 g kg−1 FW). Tempo hearts were more similar tothose of VP in biometrical and chemical terms (P, Na, K, Ca); they had the highest dry matter content (163 g kg−1 FW) and thewaste left after peeling had the highest TP content.

CONCLUSIONS: Hybrid artichokes, especially Opal and Madrigal, appear more suitable for the processing industry and also forfresh-cut production due to their highest processing yield and lowest total phenol content. Because of its high total phenolcontent, Tempo waste represents a possible source of natural antioxidant in the pharmaceutical field and in the food industry(as a food additive).c© 2010 Society of Chemical Industry

Keywords: processing yield; polyphenols; antioxidant activity

INTRODUCTIONGlobe artichoke [Cynara cardunculus L. subsp. scolymus (L.) Hayek]is a perennial rosette plant grown for its ‘capitula’, commonlyreferred to as ‘heads’ or ‘buds’. The edible part (‘heart’) consists ofa receptacle and innermost tender bracts.

Artichoke has a marked antioxidative and health protectivepotential due to its high levels of phenolic compounds, which arealso important from a technological point of view. In fact, in storedartichoke enzymatic or chemical reactions are responsible for theappearance of browning phenomena,1 representing a significantproblem both for industrial processing and, above all, for fresh-cutpreparation.

Globe artichoke production is traditionally based on vegeta-tively propagated cultivated varietal types. Those widely grownin Apulia are Violet du Provence (VP) and Catanese, both with re-flowering characteristic, whose heads are harvested both duringautumn–winter (early production) and spring (late production).Normally, early production is for the fresh market, whereas thelast part of the late production is industrially processed (canned orfrozen).

In the last three decades artichoke breeding programmeshave produced some seed (achens) propagated hybrids, whichare starting to be considered in the main artichoke producingareas (even in Apulia) as an alternative to the traditional ones.Being propagated from ‘seeds’, these hybrids are free of the

main endemic diseases (e.g. viruses and fungi, like Verticilliumdaliae), infecting propagation material of cultivated. They havehigh agronomic performance2 – 4 and they are also suitable bothfor the fresh market and processing industry,5,6 even if they havea late or medium–late harvest.

In previous studies agronomic,2,6,7 morphological orqualitative8,9 traits were evaluated on different hybrids. However,there is little or no information on the head/heart biometrical andqualitative traits of the new seed-propagated hybrids Madrigal,Opal and Tempo.

With the aim of defining the nutritional and technologicalquality of the heads of these latter hybrids, in the present workthe morphological and the chemical characteristics of heads inMadrigal, Opal and Tempo were evaluated and compared withthose of the standard cultivated varietal types (Violet du Provence

∗ Correspondence to: Anna Bonasia, Department of Agro-Environmental Science,Chemistry and Plants Protection – DiSACD – University of Foggia, via Napoli25, 71100 Foggia, Italy. E-mail: [email protected]

a Department of Agro-Environmental Science, Chemistry and Plants Protec-tion – DiSACD – University of Foggia, via Napoli 25, 71100 Foggia, Italy

b Department of Engineering and Management of the Agricultural, Livestockand Forest Systems – PROGESA – University of Bari, Via Amendola 165/A, 70126Bari, Italy


26

90

www.soci.org A Bonasia et al.

and Catanese) for the Apulia region, one of the most importantareas for artichoke production in Italy.

MATERIALS AND METHODSChemicalsβ-Carotene, linoleic acid and Tween 40 were purchased fromSigma Chemical Co. (St Louis, MO, USA). Sodium bicarbonate,gallic acid, Folin–Ciocalteu reagent, hydrochloric acid, chloroformand methanol analytical grade from Carlo Erba (Rodano, Milan,Italy). Ultrapure water (Millipore, Billerica, MA, USA) was used.

Collection of raw materialOn 18 April 2008 artichoke heads of the cultivated types (Cataneseand Violet du Provence (VP)) and of the new hybrids (Tempo, Opaland Madrigal (Nunhems Netherlands BV, Haelen, The Netherlands)were simultaneously harvested from commercial fields in theApulia region. The heads were picked at the optimal stage for freshconsumption and, from the whole mass of heads, four replicationsof 20 heads were randomly selected for each genotype.

Biometric measurements and head peelingAfter completely removing the floral stem, the heads wereprocessed to obtain the edible part (hearts) as follows: the 30outermost tough bracts were first removed and the apex of thehead was cut across at 70% of head height, then starting from the31st outermost bract, the cutting force of each single bract wasmeasured and it was eliminated if the cutting force was higherthan a threshold value. A digital pressure tester (model 53 205,TR; Turoni & C. s.n.c., Forlı, Italy), equipped with a prismatic bodywas used (width 3 mm and length 10 mm). The threshold valuewas defined as the level of tenderness acceptable for processingand was fixed at 40 N by preliminary tests. The cutting force wasmeasured 5 mm below the transversal point of the upper cut ofthe bract, parallel to this and in the middle position with respectto the breadth of the bract. The inedible apex and the eliminatedbracts were collected as waste.

The fresh weight (FW), diameter and length of raw head, FWand diameter of heart, number and FW of inner bracts, number ofouter bracts and the FW of waste were all measured. In addition,the ratio between length and diameter of heads was calculatedas a representative shape index (SI). Moreover, the processedyield was expressed as grams of produce obtained from 1 kg ofraw head, and the number of heads needed to obtain 1 kg ofproduce was determined. Samples of hearts were dried at 65 ◦Cin a thermo-ventilated oven till constant weight to determine dryweight (DW).

Preparation of plant material for chemical analysesSamples of hearts and waste were separately sliced into smallpieces, treated by liquid nitrogen, lyophilised (model Lio5P;CinquePascal s.r.l., Trezzano, Milano, Italy), and subsequently finelyground using a mortar.

The modified method described by Gil-Izquierdo et al.10 wasused for preparation of plant extracts: lyophilised samples (1 g)were extracted with 2 × 20 mL of water/methanol (20 : 80; v/v)solution using a refrigerated centrifuge Beckman Coulter Allegra

25 (Fullerton, CA, USA) at 21 000 × g for 15 min at 4 ◦C. Thesupernatant was extracted twice, recovered and filtered throughWhatman no. 1 filter paper. The filtrate was considered as theartichoke extract and kept in the dark at −20 ◦C until use in thefollowing assays. All samples were analysed in triplicate.

Content of inorganic cations and phosphorusInorganic cations were extracted from 1 g of lyophilised heartsamples, previously ashed in a muffle furnace at 550 ◦C for 6 h, anddigested with 20 mL of 1 mol L−1 HCl in boiling water for 30 min.They were then determined by ion chromatography (Dionex ICS3000; Dionex, Sunnivale, CA, USA) with a conductivity detector,using the pre-column IonPack CG12A and the column of separationIonPack CS12A (4 × 250 mm, 5 µm), according to the methodreported by Serio et al.11

Phosphorus was determined on lyophilised heart samplesby spectrophotometry (Shimadzu UV-1800; Shimadzu ScientificInstruments, Columbia, MD, USA), following the method proposedby Miller.12

Total phenol content as determined by the Folin–CiocalteuassayTotal phenol (TP) content was determined on methanolic heartand waste extracts according to the method of Singleton andRossi.13 The methanolic extracts (100 µL) were mixed with 0.5 mLof Folin–Ciocalteu reagent and left to stand at room temperaturefor 5 min before 1.0 mL of sodium bicarbonate solution (Na2CO3,20%) was added to the mixtures. After 45 min at 30 ◦C, absorbancewas read at 750 nm (Shimadzu UV-1800). Results were expressedas grams of gallic acid equivalents kg−1 FW, using a calibrationcurve.

Antioxidant activity by β-carotene/linoleic acid assayThe antioxidant activity (AA) of heart and waste methanolicextracts was assayed based on the β-carotene bleaching methoddeveloped by Ismail et al.14 In this assay, the solution ofβ-carotene/linoleic acid mixture was prepared as follows: in3 mL β-carotene solution (5 mg β-carotene in 50 mL chloroform)were dissolved in Tween 40 (0.400 g) and linoleic acid (0.040 g).Chloroform was removed under vacuum at reduced pressure usinga rotary evaporator Buchi R-200 (Buchi Labortechnik AG, Flawil,Switzerland). Following evaporation, 100 mL of distilled watersaturated with oxygen was added to the mixture, with vigorousshaking to form an emulsion. The mixture was then added toartichoke extracts or methanol (as control). Then, the test tubeswere incubated in a water bath at 50 ◦C for 2 h. The absorbancewas measured before and after the incubation phase at 470 nm(Shimadzu UV-1800). AA was expressed as a percentage, using theformula

AA = Ai,s − Af,s

Ai,c − Af,c

where Ai,s is the initial absorbance of the sample; Af,s is the finalabsorbance of the sample; Ai,c is the initial absorbance of thecontrol; and Af,c is the final absorbance of the control.

Statistical analysisData were analysed by ANOVA using the GLM procedure of the SASSoftware (SAS, Cary, NC, USA) and mean separation was performedusing the LSD test.

RESULTSMorphological characteristics and processing featuresof headsRaw heads of Madrigal and Opal had the highest valuesin morphological parameters (on average, FW 277.8 g, length


26

91

Quality in new seed-propagated artichokes www.soci.org

Table 1. Morphological characteristics of raw artichoke heads and related processing performance in the different artichoke genotypes

Raw head Heart Waste

GenotypeDiameter

(mm)Length(mm)

FW(g)

Diameter(mm)

FW(g)

Bracts(n)

Bracts(g)

Bracts(n)

FW(g)‡

Processed yield(g kg−1 raw head)

Catanese 65.4c 88.3abc 117.8c 36.4c 35.4c 40c 19.8d 33 80.1c 300d

VP 69.4c 81.7c 141.9bc 46.8b 52.5b 56ab 30.8c 34 89.2c 384bc

Madrigal 94.9a 94.5a 292.9a 58.3a 119.5a 50b 74.7a 38 172.7a 408ab

Opal 93.8a 94.1ab 262.7a 59.2a 112.3a 63a 65.6b 37 149.5b 427a

Tempo 76.6b 87.3bc 164.9b 47.8b 61.5b 49b 36.8c 37 102.8c 374c

Significance† ∗∗∗ ∗∗ ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ NS ∗∗∗ ∗∗∗

a – d Means in columns not sharing the same letter are significantly different according to LSD test (P < 0.05).† NS, ∗∗ , ∗∗∗ not significant or significant at P ≤ 0.01 or P ≤ 0.001, respectively.‡ It is included also the weight of head apex.

94.3 mm, diameter 94.3 mm), while raw heads of standardcultivated types showed the lowest values (on average, FW129.8 g, length 85.0 mm, diameter 67.4 mm). The morphologicalparameters in Tempo raw head were intermediate between thoseof cultivated types and hybrids (diameter) or similar to VP (lengthand FW). The results for Tempo were similar to standard type VPeven in shape (shape index (SI) = 1.1; data not shown). For thestandard types, the raw head of Catanese was slightly more oblong(SI = 1.3) compared to VP (data not shown), even if the lengthand diameter of the raw heads were not significantly differentbetween the two cultivated varietal types (on average, 85.0 and67.4 mm, respectively) (Table 1). Moreover, Madrigal and Opal rawheads resulted were completely round in shape, with SI = 1.0, onaverage (data not shown).

Once processed, Madrigal and Opal also showed the highestheart FW and diameter (on average, 115.9 g FW, 58.7 mmdiameter). Tempo showed hearts similar to those of VP (onaverage, 57.0 g FW, 47.3 mm diameter), while Catanese had thesmallest hearts (35.4 g FW, 36.4 mm diameter). After peeling,Opal hearts presented the highest number of remaining innerbracts (63), followed by those of Madrigal (50) and Tempo (49).Catanese hearts showed the lowest number of inner bracts (40),while in VP it was similar to the hybrids. The FW of innerbracts was the highest in Madrigal hearts (74.7 g), followed byOpal hearts (65.6 g); in Tempo it was similar to VP (on average,33.8 g), while the lowest value was observed in Catanese hearts(19.8 g).

The number of (outer) bracts eliminated during the peelingprocess was not affected by the genotype, being 36, on average.

Waste FW (outer bracts plus head apex) was the highest inMadrigal (172.7 g), followed by Opal (149.5 g), while Tempo andthe two cultivated varietal types showed the lowest value (onaverage, 90.7 g).

The highest processing yield was in Opal and Madrigal (onaverage, 417 g kg−1 raw head), while it was lowest in Catanese(300 g kg−1 raw head). Tempo recorded a intermediate value toVP (on average, 379 g kg−1 raw head).

Quantitative characterisation of the hearts and the wasteResults for the quantitative characterisation of hearts and wasteare given in Table 2. The hearts of Tempo had the highest drymatter content (163 g kg−1 FW), followed by Catanese (133 g kg−1

FW), while Madrigal showed the lowest value (107 g kg−1 FW).Intermediate values, not dissimilar from those of Madrigal andCatanese, were found in VP and Opal (on average, 122 g kg−1 FW).

The highest contents of mineral elements in the hearts wereobserved in Madrigal for P and K (6.7 and 29.9 g kg−1 DW,respectively), in Madrigal and Catanese for Ca (on average,3.5 g kg−1 DW), and in Opal and Catanese for Na (on average,4.9 g kg−1 DW). The lowest values were observed in Tempo andVP for P and K content (on average, 4.2 and 14.5 g kg−1 DW,respectively), in Tempo and Opal for Ca content (on average,2.0 g kg−1 DW) and in Tempo, VP and Madrigal for Na level(on average, 1.6 g kg−1 DW). Intermediate values were found in

Table 2. Quantitative characteristics of the heart, and total phenol and antioxidant activity of the waste in the different artichoke genotypes

Minerals in heart (g kg−1 DW)Total phenol (g gallic

acid equivalent kg−1 FW)Antioxidantactivity (%)

Dry matter

Genotype (g kg−1 FW) P Na K Mg Ca Heart Waste Heart Waste

Catanese 133b 5.7ab 5.6a 24.2ab 3.2 3.5a 3.8b 6.9a 72a 63

VP 117bc 4.0b 1.8b 16.4bc 2.6 2.5ab 6.5a 7.2a 78a 67

Madrigal 107c 6.7a 1.5b 29.9a 2.3 3.5a 2.3c 2.0b 74a 64

Opal 127bc 6.1ab 4.2a 20.3abc 2.5 2.8b 2.6c 3.5b 77a 67

Tempo 163a 4.5b 1.6b 12.7c 1.3 1.3b 2.6c 8.0a 53b 61

Significance† ∗∗ ∗∗∗ ∗∗ ∗ NS ∗ ∗∗∗ ∗∗∗ ∗∗∗ NS

† NS, ∗ , ∗∗ , and ∗∗∗ , not significant or significant at P ≤ 0.05, P ≤ 0.01 or P ≤ 0.001, respectively.a – c Means in columns not sharing the same letter are significantly different according to LSD test (P < 0.05).


26

92

www.soci.org A Bonasia et al.

Catanese and Opal for P and K (on average, 5.9 and 22.0 g kg−1 DW,respectively) and in VP for Ca (2.5 g kg−1 DW). Mg heart contentwas not affected by the genotype (on average 2.4 g kg−1 DW).

VP hearts had the highest TP content (6.5 g of gallic acidequivalent kg−1 FW), followed by Catanese (3.8 g kg−1 FW), whilethe hybrids had the lowest content (on average, 2.5 g kg−1 FW).Tempo, VP and Catanese showed the highest TP content in waste(on average, 7.4 g kg−1 FW), Madrigal and Opal had the lowest (onaverage, 2.7 g kg−1 FW).

Tempo hearts had lower AA than the other genotypes (53% vs.75%). The AA in waste parts was not affected by the genotype,being 64%, on average.

DISCUSSIONThough in the last decade a number of papers have confirmedthe superiority of seed-propagated artichoke hybrids in terms ofyield and quality compared to traditional cultivar types,2,4,6,7 thesize of their heads when used for processing is still questionablefrom a technical point of view. Peeling machines in artichokecanning factories require correctly sized vegetables, and they areset to work with the small-sized heads of cultivated varietal types,especially at the end of the harvesting season. These new cultivarsmay be profitably used for heart production, but only by settingthe peeling machines to work with larger-sized artichokes. Thenew peeling machines are able to work with a head size rangingfrom 30 to 100 mm in diameter.

Among the genotypes tested in this research, the seed-propagated hybrids, in particular Madrigal and Opal, stand outcompared to the vegetatively propagated types; in fact, they hadthe largest heads in terms of both weight and size, especially dueto their larger diameter.

The number of bracts eliminated during the peeling processto obtain the edible hearts did not differ between the testedgenotypes, but the weight of waste (outer bracts + apex) wasconsiderably higher in Madrigal and Opal (Table 1). Since thisweight consists almost exclusively of the external bracts, the latterobservation confirms that in Madrigal and Opal external bractsare thicker than in the other genotypes, as also revealed by visualanalysis.

Despite the higher waste weight recorded in Madrigal andOpal, its incidence on the total head weight was lower than in thestandard types, especially Catanese. Indeed, the lowest processingyield was found in this latter cultivated type (300 g kg−1 raw head),while in Madrigal and Opal it was more than 400 g kg−1 raw head(Table 1). Tempo and VP processing yield values were more similarto the latter (Table 1).

Confirming the results of previous work on Madrigal,15,16 nineheads were necessary for 1 kg of hearts with Madrigal and Opal,while a considerably higher number was necessary in Tempo (17),VP (20), and especially in Catanese (29) (data not shown).

Catanese showed a processing yield in agreement with thatobserved in other research15 and with the average processingyield recorded for the most cultivated European varieties;17 thesevalues are largely lower compared to those gathered from the besttwo hybrids considered in this research.

As the inner bract FW in Madrigal and Opal was almost doublethat of Tempo and VP and almost four-fold higher than in Catanese,and considering that their number was on average slightly higheronly if compared to Catanese (Table 1), their greater thickness andfleshiness is confirmed.

After head peeling, the incidence of the remaining bracts onthe heart weight was about 59% for all the genotypes, underlininga substantial equilibrium between the two parts of the heart(receptacle and inner bracts).

Tempo was similar to VP both from a biometric point of view(FW, diameter and shape of head, number and FW of inner bract,processed yield) (Table 1) and in chemical composition (P, Na, K,Ca) (Table 2).

The TP content in the hearts of standard types, and particularlythose of VP (more than 6.0 g of gallic acid equivalent kg−1 FW),was higher than that of the hybrids (Table 2). This is in agreementwith Lombardo et al.,9 who have recently found a lower TP contentin the inner bract and receptacle of Madrigal and Tempo than inVioletto di Sicilia (a cultivated varietal type very close to Catanese)and VP.

The high abundance of these compounds in the standard typesgives them a greater nutritional value. However, the lower phenolcontent in seed-propagated hybrids brings them the advantageof lower browning processes during storage, thus minimising theneed for antioxidant treatments in processing steps.18 This factordemonstrates the greater suitability of the hybrids compared tostandard types for standard industrial processing and especially infresh-cut produce.

A recent study evaluating different artichoke genotypes forfresh-cut preparation has underlined that those with the lowestpolyphenol content are more suitable for this transformation.19

Therefore, the polyphenol content of an artichoke cultivar plays akey role in defining how it can best be used.1

Although TP content was far lower in the hybrid hearts thanin those of standard types, the antioxidant capacity did not differbetween hybrid, especially in Madrigal and Opal, and cultivatedvarietal type hearts (Table 2). This indicates that the AA capacityin Madrigal and Opal hearts, as well as the phenol fraction, isalso formed by other concurrent antioxidative compounds. Whenscreening different artichoke cultivars, Cabezas-Serrano et al.19

found the lowest TP content in Catanese, but it had the highestAA and vitamin C, and these authors concluded that the latter isthe concurrent antioxidant compound.

The average values of AA indicate there is good control capacityof oxidative damage in the tested artichoke genotypes (Table 2),in agreement with Lattanzio et al.20

In addition to processing aspects, Madrigal and Opal also standout for their high nutritional value in terms of mineral contentand particularly that of P, K and Ca (Table 2). The cation profilehas been determined for vegetatively propagated types such asVioletto di Toscana,21 but no reference data has been found forthe new hybrid cultivars considered in this research. By comparingour findings in mineral composition with the only available values,found on the INRAN website (National Research Institute for Foodand Nutrition – Italian Ministry of Agriculture), P content overlapswith INRAN data,22 while the content of the main cations is, onaverage, lower.

CONCLUSIONSTempo and Catanese stand out for their higher phenol contentin the waste than in the heart. The chemical characterisation ofthe waste contributes to indicating artichoke by-products as apotential source of natural (and non-toxic) antioxidant phenols.23

The results of this work contribute to the characterisation of newcultivars of artichokes and define the best way they can be used.Opal and Madrigal have high potential suitability in the processing


26

93

Quality in new seed-propagated artichokes www.soci.org

industry, especially in fresh-cut production, due to their very lowTP content, a key factor for limiting browning phenomena inprocessed artichokes. Moreover, their high AA in hearts could bestrategic in prolonging shelf life.

In contrast, vegetatively propagated types can be assessed asbest suited to fresh consumption, because they are characterisedby a lower processing yield compared to seed-propagatedtypes, and by a high nutritional value in terms of antioxidantcompounds.

Among the hybrids studied, Tempo was the most similar to thestandard types, especially to VP, from a biometric point of view, inchemical composition and in processing performance, but standsout for the high dry-matter content of its hearts, which couldbe useful in some processing preparations (dehydratated or friedproduce), and for the high TP content of the waste, which couldrepresent a source of natural antioxidants.

ACKNOWLEDGEMENTSWe are grateful to Mrs M.A. Previtali for technical assistance withthe laboratory analyses.

REFERENCES1 Lattanzio V, Kroon PA, Linsalata V and Cardinali A, Globe artichoke: A

functional food and source of nutraceutical ingredients. J FunctionalFoods 1:131–144 (2009).

2 Miguel A, Baixauli C, Aguilar JM, Giner A, Maroto JV, Loperz S, et al,Cultivar trials of seed propagated artichoke. Acta Hortic660:111–114 (2004).

3 Calabrese N, De Palma E and Bianco VV, Yield and quality of newcommercial seed grown artichoke hybrids. Acta Hortic 660:77–82(2004).

4 Calabrese N, De Palma E and Bianco VV, Yield and quality of seedpropagated artichoke hybrid cultivars grown for four years. ActaHortic 681:135–142 (2005).

5 Calabrese N, Cardinali A, Di Venere D, Linsalata V, Pieralice M, Sergio L,et al, Technological parameters and suitability to freezing of seedgrown artichoke hybrids. Acta Hortic 681:495–501 (2005).

6 Baixauli C, Giner A, Miguel S, Lopez B, Pascual B and Maroto JV,Agronomic behaviour of seed propagated artichoke cultivars inthe spanish mediterranean area. Acta Hortic 730:143–147 (2007).

7 Mauromicale G and Ierna A, Characteristics of head of seed-grownglobe artichoke [(Cynara cardunculus L. var. scolymus (L.) Fiori]as affected by harvest period, sowing date and gibberellic acid.Agronomie 20:197–204 (2000).

8 Di Venere D, Linsalata V, Calabrese N, Cardinali A, Sergio L andPieralice M, Biochemical characterization of new seed propagatedartichoke cultivars. Acta Hortic 681:517–522 (2005).

9 Lombardo S, Pandino G, Mauromicale G, Knodler M, Carle R andSchieber A, Influence of genotype, harvest time and plant part onpolyphenolic composition of globe artichoke [Cynara cardunculusL. var. scolymus (L.) Fiori]. Food Chem 119:1175–1181 (2009).

10 Gil-Izquierdo A, Gil MI, Conesa MA and Ferreres F, The effect of storagetemperatures on vitamin C and phenolics content of artichoke(Cynara scolymus L.) heads. Innov Food Sci Emerg Technol 2:199–202(2001).

11 Serio F, De Gara L, Caretto S, Leo L and Santamaria P, Influence of anincreased NaCl concentration on yield and quality of cherry tomatogrown in posidonia (Posidonia oceanica (L) Delile). J Sci Food Agric84:1885–1890 (2004).

12 Miller RO, Extractable chloride, nitrate, orthophosphate, potassium,and sulphate-sulfur in plant tissue: 2% acetic acid extraction, inHandbook of Reference Methods for Plant Analysis, ed. by Klara YP.CRC Press, Boca Raton, FL, pp. 115–118 (1998).

13 Singleton VL and Rossi Jr JA, Colorimetry of total phenolics withphosphomolybdic–phosphotungstic acid reagents. Am J Enol Vitic16:144–158 (1965).

14 Ismail A, Marjan ZM and Foong CW, Total antioxidant activity andphenolic content in selected vegetables. Food Chem 87:581–586(2004).

15 Conversa G, Bonasia A, Lazzizera C and Elia A, Head processingsuitability in ‘Madrigal’ and ‘Brindisino’ artichoke cultivar. ActaHortic (2010), in press.

16 Del Nobile MA, Conte A, Scrocco C, Laverse J, Brescai I, Conversa G,et al, New packaging strategies to preserve fresh-cut artichokequality during refrigerated storage. Innov Food Sci Emerg Technol10:128–133 (2009).

17 Lahoz L, Malumbres A, Macua JI, Bozal JM, Urmentea I and Arrondo MA,Agricultural and industrial study of the principal European varietiesof artichoke in Navarra. Acta Hortic 660:531–537 (2004).

18 Garcia EL and Barrett DM, Preservative treatments for fresh-cutfruits and vegetables, in Fresh-Cut Fruits and Vegetables. Science,Technology and Market, ed. by Lamikanra O. CRC Press, Boca Raton,FL, pp. 267–303 (2002).

19 Cabezas-Serrano AB, Amodio ML, Cornacchia R, Rinaldi R and Colelli G,Screening quality and browning susceptibility off five artichokecultivars for fresh-cut processing. J Sci Food Agric 89:2588–2594(2009).

20 Lattanzio V, Cicco N and Linsalata V, Antioxidant activities of artichokephenolics. Acta Hortic 681:421–427 (2005).

21 Romani A, Pinelli P, Cantini C, Cimato A and Heimler D, Char-acterization of Violetto di Toscana, a typical Italian variety ofartichoke (Cynara scolymus L.). Food Chem 95:221–225 (2006).

22 INRAN. Istituto Nazionale di ricerca per gli Alimenti e la Nutrizione.Ministero delle Politiche Agricole, Alimentari e Forestali, Tabelledi composizione degli alimenti. Available: http://www.inran.it/646/tabelle di composizione degli alimenti.html?idalimento=005120&quant=100 [10 May 2010].

23 Llorach R, Espın JC, Tomas-Barberan FA and Ferreres F, Artichoke(Cynara scolymus L.) by-products as a potential source of health-promoting antioxidant phenolics. J Agric Food Chem 50:3458–3464(2002).


26

94

Research ArticleReceived: 26 May 2010 Revised: 27 July 2010 Accepted: 28 July 2010 Published online in Wiley Online Library: 2 September 2010


Effect of storage on chemical and sensoryprofiles of peanut pastes preparedwith high-oleic and normal peanutsCecilia G Riveros,a Marta G Mestrallet,a Maria F Gayol,b Patricia R Quiroga,a

Valeria Nepoteb and Nelson R Grossoa∗

Abstract

BACKGROUND: Peanut paste and peanut butter have high oil contents and are thus susceptible to developing rancidity andoff-flavours through lipid oxidation. Preservation of the chemical and sensory quality of these products is one of the mainproblems in the peanut industry. The purpose of this study was to compare the chemical and sensory stability of peanut pasteprepared with high-oleic peanuts (cv. Granoleico, GO-P) with that of peanut paste prepared with normal peanuts (cv. Tegua,T-P) from Argentina.

RESULTS: Chemical (peroxide and p-anisidine values and conjugated dienes) and sensory (roasted peanutty, oxidised andcardboard flavours) indicators of lipid oxidation were measured in peanut pastes stored at 4, 23 and 40 ◦C. Chemical indicatorvalues and oxidised and cardboard flavours showed lower increments in GO-P than in T-P during storage. T-P had significantlyhigher peroxide value than GO-P. Roasted peanutty flavour showed a lower decrease in GO-P. Peanut paste prepared withhigh-oleic peanuts had four (at 4 ◦C), two (at 23 ◦C) and three (at 40 ◦C) times longer shelf-life than peanut paste prepared withnormal peanuts.

CONCLUSION: These results indicate that high-oleic Granoleico kernels provide peanut paste with higher protection againstlipid oxidation.c© 2010 Society of Chemical Industry

Keywords: peanuts; high oleic; sensory; stability

INTRODUCTIONA large proportion of peanut production worldwide is des-tined for domestic foods.1 Peanuts contain high levels of oil(450–540 g kg−1) and protein (250–310 g kg−1). Owing to theirhigh oil content and elevated unsaturated fatty acid concentration(300–350 g kg−1 linoleic acid, 450–500 g kg−1 oleic acid), peanutsare susceptible to lipid oxidation.2 Peanut butter and peanut pasteare important peanut products used for direct consumption or asingredients in the preparation of other foods. Preservation of thechemical and sensory quality of these products is an ongoingconcern in the peanut industry.

Many factors influence the shelf-life of peanut products, suchas variety, kernel ripeness at harvest, seed size, processingand storage conditions (temperature, time, light and oxygen).Researchers have shown increased interest in high-oleic peanutcultivars owing to their low degree of lipid oxidation duringstorage, which significantly improves the preservation of sensoryand chemical quality parameters. This high-oleic trait was initiallyreported by Norden et al.3 Other researchers have reportedpeanut lines from the USA with about 800 g kg−1 oleic acidand 20–30 g kg−1 linoleic acid.4,5 Peanut products made fromhigh-oleic varieties are expected to have higher stability. Thiseffect was detected in roasted peanuts6 and fried-salted peanuts.7

Furthermore, studies have shown that the consumption of

high-oleic oils has potential health benefits, such as loweringblood cholesterol levels in hypercholesterolaemic women.2 Theconcentration of oleic acid in high-oleic peanut oil is higherthan that in olive oil,8 which is known for its heart-healthycharacteristics.

Dry-roasted peanuts6 and fried-salted peanuts7 preparedwith high-oleic peanuts showed longer stability during storage.However, the chemical and sensory stability of peanut pasteprepared with high-oleic peanuts has not been studied in depth,especially that prepared with Argentinean peanut varieties. Theuse of high-oleic peanuts rather than normal peanuts wouldincrease shelf-life and improve the oxidative stability of peanutpaste, thus preventing loss of sensory and nutritional quality.

The purpose of this study was to compare the chemical andsensory stability of peanut paste prepared with high-oleic peanuts

∗ Correspondence to: Nelson R Grosso, Quımica Biologica, Facultad de CienciasAgropecuarias (UNC), IMBIV-CONICET, CC 509, 5000 Cordoba, Argentina.E-mail: [email protected]

a Quımica Biologica, Facultad de Ciencias Agropecuarias (UNC), IMBIV-CONICET,CC 509, 5000 Cordoba, Argentina

b ICTA, Facultad de Ciencias Exactas, F ısicas y Naturales, IMBIV-CONICET,Cordoba, Argentina


26

95

Peanut pastes prepared with high-oleic peanuts www.soci.org

with that of peanut paste prepared with normal peanuts fromArgentina.

MATERIALS AND METHODSPeanut samplesHigh-oleic peanuts (cv. Granoleico) and normal peanuts (cv. Tegua)were provided by Lorenzati, Ruetsch & Cia (Ticino, Cordoba,Argentina). Sound and mature seeds of blanched peanuts, size40–50 kernels oz−1, were selected.

To prepare ‘Granoleico’ (GO-P) and ‘Tegua’ (T-P) peanut pastes,the blanched peanuts were heated in an oven (Model 600, Memert,Schwabach, Germany) at 140 ◦C for 30 min to a medium roast,measured as an average Hunter colour lightness (L) value of50 ± 1.9 The roasted peanuts were then ground in a colloid mill(COLMIL Mod. AD 50 VR, Munro, Buenos Aires, Argentina).

Storage conditions and samplingAfter preparation, GO-P and T-P samples were packaged in 350 gplastic jars and stored at 4 ◦C (refrigeration chamber), 23 ◦C (con-ditioner room) and 40 ◦C (oven) to reproduce refrigeration, roomtemperature and accelerated storage conditions respectively.10

Samples were removed from storage after 0, 35, 70, 105, 140 and175 days for chemical and descriptive analyses.

Chemical analysisOil was obtained from the peanut paste by cold pressing using a20 ton press (HE-DU, Hermes I. Dupraz SRL, Cordoba, Argentina).The peanut oil was used for chemical analyses: fatty acid, peroxide,p-anisidine and conjugate diene determinations.

Fatty acid methyl esters were prepared from peanut paste (GO-Pand T-P) oils by transmethylation with a 30 g L−1 solution of sulfuricacid in methanol as described by Grosso et al.11 The fatty acidmethyl esters of total lipids were analysed in a Clarus 500 gas/liquidchromatograph (Perkin Elmer, Waltham, MA, USA) equipped witha flame ionisation detector. A CP-Wax 52 CB capillary column(30 m × 0.25 mm × 0.25 µm; Varian, Lake Forest, CA, USA) wasused. The column temperature was increased from 180 ◦C (heldfor 1 min) to 230 ◦C at 4 ◦C min−1. The carrier gas was nitrogen ata flow rate of 1 mL min−1. The separated fatty acid methyl esterswere identified by comparing their retention times with those ofauthentic samples purchased from Sigma Chemical Co. (St Louis,MO, USA). Quantitative fatty acid analysis was performed usingheptadecanoic acid methyl ester (Sigma Chemical Co.) as internalstandard. Iodine value (IV) was calculated using the formula11

IV = (%C18 : 1 × 0.8601) + (%C18 : 2 × 1.7321)

+ (%C20 : 1 × 0.7854)

Peroxide value (PV) was evaluated by the AOAC12 standard methodand expressed as milliequivalents active oxygen (meqO2) kg−1 oil.

Conjugated dienes (CD) and p-anisidine value (AV) were evalu-ated in an HP 8452A UV–visible diode array spectrophotometer(Hewlett Packard, Palo Alto, CA, USA) according to the IUPAC13

and COI14 standard methods respectively.

Descriptive analysisA total of ten trained panellists (six female and four male),each with at least 6 years of experience in evaluating peanutproducts, participated in the descriptive analysis of peanut paste

samples. All panellists were selected according to the followingcriteria: (a) people with natural dentition; (b) people without foodallergies; (c) non-smokers; (d) people between the ages of 18 and64; (e) people who consumed roasted peanuts and/or peanutproducts at least once a month; (f) people available for all sessions;(g) people interested in participating; (h) people able to verballycommunicate their observations regarding the product.15 Beforebeing qualified, all panellists showed a perfect score in a tastesensitivity test and the ability to identify five of seven commonlyfound food flavours.

The panellists were trained and calibrated in six training sessionsfor evaluating peanut pastes. Each training session lasted 3 h. Thedescriptive analysis test procedures described by Meilgaard et al.,16

Grosso and Resurreccion10 and Nepote et al.17 were used to trainthe panellists. A ‘hybrid’ descriptive analysis method combiningthe Quantitative Descriptive Analysis (Tragon Corp., RedwoodCity, CA, USA) and Spectrum Analysis (Sensory Spectrum, Inc.,Chatham, NJ, USA) methods was used by the panellists forevaluating samples. A 150 mm unstructured linear scale wasemployed.15 A list of definitions and a sheet with warm-up andreference intensity ratings (Table 1) were developed during thetraining sessions.10 The attribute definitions were based on thepeanut lexicon.9

All samples were evaluated in partitioned booths underfluorescent light at room temperature. Product samples (10 g)were placed in plastic cups with lids coded with three-digitrandom numbers. The final lists of warm-up and reference intensityratings and definitions (Table 1) were posted in the booths for alltest sessions.10 The panellists were instructed to first familiarisethemselves with the reference standard intensities (Table 1) andthen evaluate the sensory attributes of the peanut paste samples. Acompletely randomised block design was used for testing samples.The data were registered on paper ballots.

Statistical analysisThe experiment was replicated three times. The data were analysedusing InfoStat Version 1.1 (Facultad de Ciencias Agropecuarias,Universidad Nacional de Cordoba, Argentina). Means and standarddeviations were calculated. Analysis of variance and Duncantests (α = 0.05) were used to detect significant differences insensory attributes and chemical analyses between sampling days.Pearson coefficients were used to calculate correlations betweendependent variables. Second-order polynomial equations wereused in the regression analyses.

RESULTS AND DISCUSSIONChemical analysesGO-P had higher oleic acid content but lower linoleic acid content(785 and 46 g kg−1 respectively) than T-P (458 and 333 g kg−1

respectively). Similar results for these fatty acid contents were alsofound in others peanut products6,7,17 prepared with high-oleic andnormal peanut lines from Argentina and other countries.18,19 Theoleic/linoleic ratio was 17.06 in GO-P and 1.38 in T-P. IV was lower inGO-P (77) than in T-P (98). In addition, GO-P had higher eicosenoicacid content but lower palmitic acid content (25.1 and 57.7 g kg−1

respectively) than T-P (17.2 and 99.3 g kg−1 respectively). Otherfatty acids did not differ significantly between GO-P and T-P.

The changes in PV, CD and AV during storage at 4, 23 and 40 ◦Care shown in Fig. 1. Except for GO-P samples stored at 4 ◦C, PV andCD increased significantly with storage time. In each treatment,


26

96

www.soci.org CG Riveros et al.

Table 1. Definitions of attributes, standard references and warm-up intensity ratings used in descriptive analysis of peanut pastes

Attributea Definition Reference Reference intensityb Warm-up intensityb,c

Appearance

1. Brown colour Intensity or strength of brown colourfrom light to dark brown

Hazelnut and cocoa spreadbutterd

120 44

2. Uneven colour Amount of speckles in peanut pastecoming from residual particles ofpeanut skins


75 15

3. Glossiness Appearance associated with amount oflight reflected by product surface


100 115

Aroma

4. Roasted peanutty Aroma associated withmedium-roasted peanuts

Dry-roasted peanutse 44 55

5. Oxidised Aroma associated with rancid fats andoils

Rancid peanuts 75 1

6. Cardboard Aroma associated with wet cardboard Moist cardboard 45 5

7. Burnt Aroma associated with over-roastedpeanuts

8 g of coffeef in 250 mL ofdistilled water

80 20

8. Raw/beany Aroma associated with uncooked orraw peanuts

Raw peanutsg 65 30

Taste

9. Sweetness Taste on tongue associated withsucrose solutions

20 g kg−1 sucrose solution 20 25

50 g kg−1 sucrose solution 50

100 g kg−1 sucrose solution 100

10. Saltiness Taste on tongue associated withsodium chloride solutions

2 g kg−1 NaCl solution 25 10

3.5 g kg−1 NaCl solution 50

5 g kg−1 NaCl solution 85

11. Sourness Taste on tongue associated with acidagents such as citric acid solutions

0.5 g kg−1 citric acid solution 20 5

0.8 g kg−1 citric acid solution 50

1.5 g kg−1 citric acid solution 100

12. Bitterness Taste on tongue associated with bittersolutions such as caffeine

0.5 g kg−1 caffeine solution 20 15

0.8 g kg−1 caffeine solution 50

1.5 g kg−1 caffeine solution 100

Mouthfeel

13. Astringency Puckering or drying sensation onmouth or tongue surface

8 g of coffeef in 250 mL ofdistilled water

60 35

Textureh

14. Oiliness Degree to which free oil is perceived inmouth


40 55

15. Adhesiveness Force required for removing materialthat adheres to palate during normaleating process


45 80

16. Graininess Degree to which grains or granules areperceived in mouth


25 35

a Attributes listed in same order as perceived by panellists.b Intensity ratings based on 150 mm unstructured linear scale.c Standard peanut paste (L = 50 ± 1) prepared with roasted peanuts (cv. Runner, blanched).d Hazelnut and cocoa spread butter: Nutella (Ferrero SpA, Alba, Italy).e Dry-roasted peanuts: cv. Runner (JL SA, Ticino, Cordoba, Argentina).f Coffee: Nescafe Clasico (Nestle Argentina SA, Buenos Aires, Argentina).g Raw peanuts: size 40–50 kernels oz−1 (Lorenzati, Ruetsch & Cia, Ticino, Cordoba, Argentina).h The texture descriptors and their corresponding definitions were adapted from Muego-Ganasekharan and Resurreccion.22

AV did not show any marked increase during the storage period.However, GO-P samples stored at 4 and 23 ◦C had significantlylower AV than the other treatments. T-P had higher PV and showedsignificant differences during storage after day 0 at 23 and 40 ◦Cand after day 35 at 4 ◦C with respect to GO-P. T-P had slightlyhigher AV than GO-P.

Abegaz et al.20 reported that storage time was a significantfactor influencing PV in model peanut butter confections storedat 21 ◦C. Their results indicated that PV increased markedlywithin 4 weeks. In roasted peanuts in shell,21 roasted peanuts6

and fried-salted peanuts,7 PV and CD obtained at the end ofthe storage period were higher than those found in peanut


26

97


0 35 70 105 140 175

Days

0

2

5

7

9

11

14

16

Per

oxid

e va

lue

(meq

O2

Kg−1

)

0 35 70 105 140 175

Days

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Con

juga

ted

Die

nes

(E1;

%23

2nm

)

0 35 70 105 140 175

Days

0

1

2

p-A

nisi

dine

val

ue

(a)

(c)

(b)

Figure 1. Changes in (a) peroxide value, (b) conjugated dienes and (c) p-anisidine value in Granoleico (GO-P) and Tegua (T-P) peanut pastes duringstorage: , GO-P at 4 ◦C; , GO-P at 23 ◦C; , GO-P at 40 ◦C; ♦, T-P at 4 ◦C;�, T-P at 23 ◦C; �, T-P at 40 ◦C.

pastes (T-P and GO-P) in the present study. These results indicatethat peanut paste is more resistant to lipid oxidation processesin comparison with other peanut products. After the roastingprocess, roasted peanuts release moisture, forming small cellsthat trap air inside. It is likely that the oxygen in these cellsaccelerates the lipid oxidation process. When roasted peanutsare finely ground to produce peanut paste, these air cells aredestroyed, so peanut paste traps less air than roasted peanutkernels.

0 35 70 105 140 175

Days

0

2

5

7

9

11

14

16

Oxi

dize

d in

tens

ity

rati

ng

0 35 70 105 140 175

Days

0

2

5

7

9

12

14

Car

dboa

rd in

tens

ity

rati

ng

0 35 70 105 140 175

Days

47

49

51

53

54

56

58

60

Roa

sted

pea

nutt

y in

tens

ity

rati

ng

(a)

(b)

(c)

Figure 2. Changes in intensity rating of (a) oxidised, (b) cardboard and(c) roasted peanutty flavours in Granoleico (GO-P) and Tegua (T-P) peanutpastes during storage: , GO-P at 4 ◦C; , GO-P at 23 ◦C; , GO-P at 40 ◦C;♦, T-P at 4 ◦C; �, T-P at 23 ◦C; �, T-P at 40 ◦C.

Descriptive analysesSixteen sensory attributes were evaluated, but only three (roastedpeanutty, oxidised and cardboard flavours) showed significantchanges in their intensity ratings during storage (Fig. 2). Theintensity of oxidised flavour was higher in T-P except for samplesstored at 4 ◦C, in which there were no significant differences


26

98

www.soci.org CG Riveros et al.

Table 2. Coefficients and adjusted R2 values from regression equations of peroxide value (PV), conjugated dienes (CD), p-anisidine value (AV) andsensory attributes (oxidised, cardboard and roasted peanutty flavours) in Granoleico (GO-P) and Tegua (T-P) peanut pastes at 4, 23 and 40 ◦C

Regression coefficientsa

Sample Dependent variable Temperature (◦C) β0 β1 β11 R2

GO-P PV 4 −0.006779 −0.000620 0.000017 0.955723

23 0.112643 −0.013597 0.000155 0.984468

40 0.155357 −0.015676 0.000198 0.968748

CD 4 0.912810 0.002166 −0.000011 0.700428

23 0.926633 −0.000322 0.000015 0.933126

40 0.928497 0.00019 0.000017 0.991511

AV 4 0.030389 0.000437 −0.000001 0.949463

23 0.042354 0.002560 −0.000009 0.946595

40 0.109461 0.008736 −0.000035 0.852145

Oxidised 4 2.618214 0.025170 0.000028 0.924024

23 2.726071 0.024197 0.000033 0.979966

40 2.848929 0.050405 −0.000101 0.994754

Cardboard 4 4.473929 0.021881 −0.000032 0.943499

23 4.132857 0.039245 −0.000097 0.992439

40 4.553214 0.043870 −0.000121 0.951792

Roasted peanutty 4 57.351071 −0.045866 0.000152 0.930223

23 57.217857 −0.037104 0.000047 0.967110

40 56.917857 −0.060618 0.000182 0.984936

T-P PV 4 −0.493571 0.064076 −0.000042 0.979634

23 −0.182893 0.114488 −0.000273 0.996934

40 0.597857 0.142239 −0.000390 0.981749

CD 4 1.134650 0.006293 −0.000003 0.913256

23 1.193871 0.014912 −0.000036 0.982638

40 1.253916 0.021565 −0.000072 0.980057

AV 4 0.454742 0.000771 −0.000005 0.069723

23 0.449254 0.002440 −0.000005 0.642913

40 0.437830 0.000962 0.000004 0.905300

Oxidised 4 2.990357 0.005997 0.000164 0.984725

23 2.771429 0.040857 0.000023 0.957019

40 2.761071 0.057497 0.000043 0.974840

Cardboard 4 4.850000 0.019253 0.000055 0.993361

23 4.793214 0.025691 0.000055 0.989509

40 4.280000 0.049404 −0.000019 0.953164

Roasted peanutty 4 57.209643 −0.090605 0.000302 0.992060

23 57.210714 −0.094533 0.000265 0.996901

40 57.143929 −0.107528 0.000352 0.994401

a Regression coefficients for the general regression equation Y = β0 +β1X +β11X2, where Y is the dependent variable (PV, CD, AV, oxidised, cardboardor roasted peanutty) and X is the independent variable (days of storage).

between GO-P and T-P. Significant differences (α = 0.05) inoxidised flavour intensity between GO-P and T-P were observedfrom day 70 at 23 and 40 ◦C. The intensity ratings of cardboardflavour were higher in T-P. Significant differences (α = 0.05) incardboard flavour intensity between GO-P and T-P were detectedfrom day 70 at 4, 23 and 40 ◦C. Other authors reported similarresults for normal peanut (cv. Florunner) paste stored at 40 ◦C for161 days22 and model (ideal) peanut butter confection stored at21 ◦C20,23 with respect to the results obtained for T-P in the presentstudy.

Roasted peanutty flavour is the attribute used to characterisetypical roasted peanut flavour in peanut products.9 Bett andBoylston24 and Nepote et al.7 reported that roasted peanuttyflavour intensity decreased in roasted peanuts during storage.In the present study the intensity ratings of roasted peanutty

flavour in GO-P and T-P decreased during storage. The intensitiesof this sensory attribute were lower in T-P. Significant differences(α = 0.05) in roasted peanutty intensity between GO-P and T-Pwere detected from day 35 at all temperatures tested. Similarresults were observed for peanut butter prepared with normalpeanuts and stored at 25 ◦C for 29 days25 in comparison with theresults observed for T-P in the present study.

Correlation and regression analysesThe variables that changed during storage (PV, CD, AV and oxi-dised, cardboard and roasted peanutty flavours) were correlatedusing Pearson coefficients. Positive correlations (higher than 0.60)were detected between PV, CD, AV and oxidised and cardboardflavours, the highest being between PV and CD (0.98) and betweenoxidised and cardboard flavours (0.92). Negative correlations of


26

99


Table 3. Shelf-life (days to reach a peroxide value of 10 meqO2 kg−1)of Granoleico (GO-P) and Tegua (T-P) peanut pastes at 4, 23 and 40 ◦Cestimated using regression equation

Shelf-life (days)

Sample 4 ◦C 23 ◦C 40 ◦C

T-P 187 128 87

GO-P 786 300 266

roasted peanutty flavour with PV, CD, AV and oxidised and card-board flavours were found, the highest being with PV (−0.87),oxidised flavour (−0.87) and CD (−0.83). These negative correla-tions indicated that roasted peanutty flavour decreased as lipidoxidation indicators increased during storage time.

The regression equations of dependent variables (PV, CV, AVand oxidised, cardboard and roasted peanutty flavours) for GO-Pand T-P are presented in Table 2. The dependent variables showedR2 > 0.60 except for AV in T-P stored at 4 ◦C. These regressionsequations could be used to estimate the effect of storage timeat 4, 23 and 40 ◦C on peanut paste prepared with normal andhigh-oleic peanuts.

In the Food Code of Argentina, 10 meqO2 kg−1 is the maximumPV allowed for peanut products.26 A PV of 10 meqO2 kg−1 may beuseful as a reference value for the endpoint of peanut paste shelf-life. Using the prediction equation, the shelf-lives of peanut pastesprepared with high-oleic (GO-P) and normal (T-P) peanuts arepresented in Table 3. PV levels above 10 meqO2 kg−1 were reachedin T-P before GO-P during storage at different temperatures. Usingthe prediction equation, it was estimated that 10 meqO2 kg−1 isreached on day 128 in T-P and on day 300 in GO-P when the peanutpaste is stored at 23 ◦C. The shelf-life of GO-P is approximately four(at 4 ◦C), two (at 23 ◦C) and three (at 40 ◦C) times longer than thatof T-P. These results indicate that high-oleic Granoleico kernelsprovide peanut paste (or peanut butter) with higher protectionagainst lipid oxidation and improve the shelf-life of this productconsiderably.

CONCLUSIONSThis study provides an equation to estimate the shelf-life ofpeanut paste using chemical measurements of PV, AV and CD andintensity ratings of sensory attributes (oxidised, cardboard androasted peanutty flavours) from descriptive analysis. The resultsobtained show that peanut paste prepared with high-oleic peanuts(GO-P) has higher storage stability than peanut paste preparedwith normal peanuts (T-P). This is due to the increased resistanceof the former to lipid oxidation.

ACKNOWLEDGEMENTSThis work was supported by CONICET and SECYT-UNC.

REFERENCES1 Riveros CG, Mestrallet MG, Nepote V and Grosso NR, Chemical

composition and sensory analysis of peanut pastes elaboratedwith high-oleic and regular peanut from Argentina. Grasas Aceites60:388–395 (2009).

2 Frankel EN, Lipid Oxidation. The Oily Press, Bridgewater (2005).3 Norden AAJ, Gorbet DW, Knauff DDA and Young CT, Variability in oil

quality among peanut genotypes in the Florida Breeding Program.Peanut Sci 14:7–11 (1987).

4 Braddock JC, Sims CA and O’Keefe SF, Flavour and oxidative stabilityof roasted high oleic acid peanuts. J Food Sci 60:489–493 (1995).

5 Mugendi JB, Sims CA, Gorbet DW and O’Keefe SF, Flavor stability ofhigh oleic peanuts stored at low humidity. J Am Oil Chem Soc75:21–25 (1998).

6 Nepote V, Mestrallet MG, Accietto RH, Galizzi M and Grosso NR,Chemical and sensory stability of roasted high-oleic peanuts fromArgentina. J Sci Food Agric 86:944–952 (2006).

7 Nepote V, Mestrallet MG and Grosso NR, Oxidative stability in fried-salted peanuts elaborated with high-oleic and regular peanut fromArgentina. Int J Food Sci Technol 41:900–909 (2006).

8 Cabrini L, Barzanti V, Cipollone M, Florenini D, Grossa G, Tolomelli B,et al, Antioxidants and total perosyl radical-trapping ability of oliveand seed oils. J Agric Food Chem 49:6026–6032 (2001).

9 Johnsen PB, Civille GV, Vercellotti JR, Sanders TH and Dus CA, De-velopment of a lexicon for the description of peanut flavor. J SensoryStud 3:9–17 (1988).

10 Grosso NR and Resurreccion AVA, Predicting consumer acceptanceratings of cracker-coated and roasted peanuts from descriptiveanalysis and hexanal measurements. J Food Sci 67:1530–1537(2002).

11 Grosso NR, Nepote V and Guzman CA, Chemical composition of somewild peanut species (Arachis L.) seeds. JAgric Food Chem 48:806–809(2000).

12 AOAC, Official Methods of Analysis (16th edn). Association of OfficialAnalytical Chemists, Washington, DC (1995).

13 IUPAC, Method number 2.504. Determination of the p-anisidine value(p-A.V.), in Standard Methods for the Analysis of Oils, Fats andDerivatives (7th edn), ed. by Paquot C and Hautfenne A. Blackwell,Oxford, pp. 1–347 (1987).

14 COI, Metodo de analisis, prueba espectrofotometrica en el ultravioleta.Document COI/T, 20/Doc. No. 19/Rev. 1, International Olive OilCouncil, Madrid (2001).

15 Plemmons LE and Resurreccion AVA, A warm-up sample improvesreliability of responses in descriptive analysis. J Sensory Stud13:359–376 (1998).

16 Meilgaard M, Civille GV and Carr BT, Sensory Evaluation Techniques(2nd edn). CRC Press, Boca Raton, FL, pp. 135–235 (1991).

17 Nepote V, Olmedo RH, Mestrallet MG and Grosso NR, A study of therelationships among consumer acceptance, oxidation chemicalindicators and sensory attributes in high-oleic and normal peanuts.J Food Sci 74:1–8 (2009).

18 Ozcan M and Seven S, Physical and chemical analysis and fatty acidcomposition of peanut, peanut oil and peanut butter from COMand NC-7 cultivars. Grasas Aceites 54:12–18 (2003).

19 Isleib TG, Pattee HE, Sanders TH, Hendrix KW and Dean LO, Com-positional and sensory comparisons between normal and high-oleic peanuts. J Agric Food Chem 54:1759–1763 (2006).

20 Abegaz EG, Kerr WL and Koehler PE, The role of moisture in flavorchanges of model peanut confections during storage. LebensmWiss Technol 37:215–225 (2004).

21 Mozingo RW, O’Keefe SF, Sanders TH and Hendrix KW, Improving shelflife of roasted and salted in shell peanuts using high oleic fatty acidchemistry. Peanut Sci 31:40–45 (2004).

22 Muego-Ganasekharan KF and Resurreccion AVA, Physicochemicaland sensory characteristics of peanut paste stored at differenttemperatures. J Food Sci 57:1385–1389 (1992).

23 Abegaz EG, Kerr WL and Koehler PE, Descriptive sensory analysis ofstored model peanut confections with different sugar, moistureand antioxidant levels. Peanut Sci 33:53–59 (2006).

24 Bett KL and Boylston TD, Effect of storage on roasted peanut quality.ACS Symp Ser 500:322–343 (1992).

25 Felland SL and Koehler PE, Sensory, chemical and physical changesin increased water activity peanut butter products. J Food Qual20:145–156 (1997).

26 CAA, Capıtulo VII, Alimentos Grasos, Artıculo No 531 (Res. 2012,19.10.84). Codigo Alimentario Argentino, Ley 18.248 (18.7.1969), De laCanal y Asociados, Buenos Aires, pp. 164–165 (1996).


27

00



Meta-analyses of effects of phytochemicalson digestibility and rumen fermentationcharacteristics associated withmethanogenesisAmlan K Patra∗

Abstract

BACKGROUND: A meta-analysis study was conducted to investigate the changes in rumen fermentation characteristics whenmethane inhibition by phytochemicals is employed. The whole database containing 185 treatment means from 36 publishedstudies was divided into four subsets according to the major phytochemicals used in the studies, i.e. saponins, tannins, essentialoils (EO) and organosulfur compounds (OS).

RESULTS: Changes in protozoal numbers showed linear relationships with changes in methane production by saponins(R2 = 0.48), tannins (R2 = 0.30) and EO (R2 = 0.20) but not OS. Concentrations of total volatile fatty acids (VFA) and acetatedid not show any relationship (P > 0.1) with changes in methane due to saponins. However, propionate production increasedlinearly with increasing inhibition of methane (R2 = 0.31), which resulted in a linear (R2 = 0.26) decrease in acetate/propionateratio (A/P) with decreasing methane production. Concentrations of total VFA, acetate and propionate did not change withchanges in methane production by tannins. However, A/P showed a significant linear relationship (R2 = 0.27) with decreasingmethane formation. Concentrations of total VFA (R2 = 0.44) and propionate (R2 = 0.15) changed linearly and positively withchanges in methane production by EO. However, acetate production (R2 = 0.22) and A/P (R2 = 0.17) increased linearly withincreasing inhibition of methane by EO. Changes in concentrations of total VFA (R2 = 0.60) and acetate (R2 = 0.35) decreasedlinearly while those of propionate increased linearly (R2 = 0.23) with increasing inhibition of methane by OS. Consequently,A/P decreased linearly (R2 = 0.30) with decreasing methane production by OS. Digestibilities of organic matter (OM) andneutral detergent fibre were not affected by inhibition of methane production by saponins, EO and OS, but digestibility of OMdecreased with decreasing methane production by tannins.

CONCLUSION: The inhibition of methane production by phytochemicals results in changes in rumen fermentation that differdepending on the types of phytochemicals.c© 2010 Society of Chemical Industry

Keywords: phytochemicals; methane production; rumen fermentation; meta-analysis

INTRODUCTIONMethane is the second most anthropogenic greenhouse gas aftercarbon dioxide and thus contributes to global warming andclimate change.1 Agriculture is responsible for about 47% of totalanthropogenic methane emissions, of which 32% comes fromenteric fermentation in livestock.1 Enteric methane is producedduring the fermentation of feeds by methanogenic archaea, mostlyin the rumen. Hence, in addition to greenhouse effects, methaneproduction in ruminants represents a loss of about 2–15% offeed energy,2 which decreases the metabolisable energy contentof feeds. From the calculation of energy balances3,4 as cited byBeauchemin et al.,5 it has been suggested that a 25% decrease inmethane production in ruminants might result in an increase inmilk production of 1 L day−1 in high-yielding dairy cows or growthof 75 g day−1 in growing cattle. Therefore the development offeeding strategies to decrease methane emissions in ruminantsmerits research attention for long-term mitigation of greenhouse

gas emissions into the atmosphere as well as short-term economicbenefits.6

The use of phytochemicals to inhibit methane production inthe rumen could provide benefits over chemical feed additivesin relation to the presence of chemical residues in animal-derived foods. Many phytochemicals, broadly saponins, essentialoils (EO), tannins and organosulfur compounds (OS), have beenshown to lower methane production in vitro and in vivo.7 – 10 Thesephytochemicals also affect rumen metabolism, e.g. volatile fatty

∗ Correspondence to: Amlan K Patra, Department of Animal Nutrition, Faculty ofVeterinary and Animal Sciences, West Bengal University of Animal and FisherySciences, 37 K. B. Sarani, Belgachia, Kolkata 700037, India.E-mail: patra [email protected]

Department of Animal Nutrition, Faculty of Veterinary and Animal Sciences,West Bengal University of Animal and Fishery Sciences, 37 K. B. Sarani, Belgachia,Kolkata 700037, India


27

01

Effects of phytochemicals on rumen fermentation associated with methanogenesis www.soci.org

acid (VFA) production pattern and digestibility of nutrients,7,9 andultimately animal performance. Hence it is important to considerrumen fermentation characteristics when methane mitigationstrategies involving these phytochemicals are employed. A meta-analysis of data obtained from previously published studiescan explain the responses of rumen fermentation pattern whenmethane inhibition is targeted by different phytochemicals undera variety of experimental conditions.11,12 Therefore a meta-analysiswas conducted to study rumen fermentation and digestibility inrelation to methane inhibition by phytochemicals.

MATERIALS AND METHODSDescription of databaseA meta-analysis of the effects of phytochemicals on rumenfermentation characteristics and digestibilities in associationwith methanogenesis was conducted by pooling and analysingdata from previously published studies. The studies that wereincluded in the database reported the effects of phytochemicalsor plants/plant extracts rich in phytochemicals on methanogenesisand rumen fermentation and/or digestibilities of nutrients.These studies were published in peer-reviewed English languagejournals. Although several other studies have reported the effectsof phytochemicals on methanogenesis, they were not included inthe database, since the main active components were generally notidentified. Overall, 36 published studies of 185 different treatmentgroups were considered (Appendix). The database included dataon composition of feeds, digestibilities of organic matter (OM),neutral detergent fibre (NDF), acid detergent fibre (ADF) andcrude protein (CP), protozoal numbers and rumen fermentationcharacteristics. Changes in rumen fermentation, digestibilities andprotozoal numbers compared with control values were calculated(i.e. change = (value in phytochemical-supplemented treatment- control value)/control value)) to investigate how changes inmethanogenesis by phytochemicals alter digestibilities, VFA andprotozoal counts. Changes in methane production were calculatedafter expressing methane per unit of dry matter (DM) or OMdigestibility. All variables were not available across all observationsin the database. Hence the numbers of observations used forregression analyses varied between dietary and response variablesdepending on the regressor variables available. Since many ofthe studies did not report concentrations of phytochemicals, itwas not possible to consider this factor in the database. Manystudies also reported additional outcomes, but only the outcomesof interest were used here. Data reported in different units ofmeasurement were transformed to the same units. Some recordswere incomplete or not reported uniformly, which necessitatedcalculations from the reported data. When a study did not reportall possible outcomes and it was not possible to calculate from thereported data, missing variables were considered as missing data.

The whole database was divided into four subsets accordingto the major phytochemicals used in the studies, i.e. saponins(n = 53), tannins (n = 48), EO (n = 44) and OS (n = 40). Most ofthe data were from in vitro measurements. Saponins and saponin-containing plants or extracts included Yucca saponaria, Sapindussaponaria, tea saponins, alfalfa saponins, Acacia concinna, Quillajasaponaria, Enterolobium cyclocarpum, Trigonella foenum-graecum,Pithecellobium saman and Sesbania sesban. Tannins included que-bracho, Castanea sativa, Acacia mearnsii, myrabolam, chestnut,Lespedeza striata, Lespedeza cuneata, Biophytum petersianum, Aca-cia mangium, Psidium guajava, Jatropa curcas, Phaleria papuana,Quercus incana and Persea americana. EO included peppermint

oil, ropadiar, cinnamon oil, juniper berry oil, eucalyptus oils, Syzy-gium aromaticum, Foeniculum vulgare, anethol and cymene. OSincluded allicin, diallyl disulfide, garlic oil, allyl isothiocyanate andhorseradish oil.

Statistical analysisAlthough the statistical analysis procedure used for meta-analysisof the database has been described elsewhere,13 a brief account ispresented here. All statistical computations were carried out usingthe PROC MIXED, PROC REG and PROC CORR procedures of SASVersion 8.2.14 Data were analysed according to St-Pierre,15 takinginto account the random effect of study, using PROC MIXED14 withthe following model:

Yij = B0 + B1Xij + B2X2ij + si + biXij + eij

where Yij is the expected outcome for the dependent variable Yobserved at level j of the continuous variable X in study i; B0 isthe overall intercept across all studies (fixed effect); B1 and B2 arethe overall linear and quadratic regression coefficients of Y onX respectively across all studies (fixed effect); Xij is the syntheticdatum value j of the continuous variable X in study i; si is therandom effect of study i; bi is the random effect of study i on theregression coefficient of Y on X in study i; and eij is the unexplainedresidual error.

The variable study was declared in the CLASS statement.The slope and intercept by study were included as randomeffects, as suggested by St-Pierre.15 In addition, an unstructuredvariance–covariance matrix (type = un) was performed in therandom part of the model, as suggested by St-Pierre,15 toavoid the positive correlation between intercept and slope.When random covariance of the slope and intercept or randomslope and intercept was not significant (P > 0.1), a variancecomponent (type = vc) of variance–covariance structure wasperformed or slope and intercept were not included in the modelrespectively.15 When the square term of slope was not significant(P > 0.1), it was not included in the model. Because the dataon changes in rumen fermentation parameters, protozoal countsand digestibility involve relative rates, they were subjected tologarithmic transformation before analysis to meet the criterionof homogeneity of variance. Since the values of some of theseparameters were negative, data were coded by adding 1 prior tologarithmic transformation (i.e. log(x + 1), where x is the variable).

For proper graphical representation of statistical results fromthe multidimensional space of studies in two-dimensional space,the Y observation was adjusted to take into account the randomeffect of study.15 The regression coefficient (R2) calculations alsoused the adjusted Y observation.15

RESULTSDescription of data setMeans and standard deviations for selected variables in the foursubsets are reported in Table 1. In the saponin subset, NDFconcentrations of feeds ranged from 344 to 682 g kg−1 DM witha mean value of 468 g kg−1 DM, while CP contents ranged from73 to 287 g kg−1 DM with a mean value of 148.5 g kg−1 DM. Theeffect of saponins on methane production varied from stimulatory(19%) to inhibitory (42%) with an average of 11% inhibition,while the growth of protozoa ranged from an increase by 60%to a decrease by 79% with an average of 28% lower in the


27

02

www.soci.org AK Patra

Table 1. Descriptive statistics of dietary composition and percentage changes in rumen fermentation and digestibility variables relative to controlvalues (i.e. % change versus control) as affected by different phytochemicals

Saponins Tannins Essential oils Organosulfurs

Parameter Mean SD Mean SD Mean SD Mean SD

OM content (g kg−1) 917.6 21.03 927.9 33.44 907.2 8.74 914.4 8.44

NDF content (g kg−1) 467.9 97.87 490.8 112.65 609.9 85.07 434.3 110.19

ADF content (g kg−1) 273.4 58.56 254.5 50.27 335.6 29.56 292.3 71.50

CP content (g kg−1) 148.5 40.75 132.7 23.39 119.5 11.21 154.9 27.39

Methane −10.96 12.96 −19.13 18.57 −35.35 32.41 −23.09 30.82

Protozoa −28.27 34.52 −14.61 28.14 −34.92 29.86 14.58 31.59

TVFA 1.24 8.11 −5.75 5.70 1.06 10.91 3.00 10.48

Propionate 8.03 11.03 −4.47 6.46 −3.29 9.07 6.64 9.94

Acetate −0.60 5.02 −4.01 5.24 −1.30 5.66 −1.84 5.82

A/P −8.61 11.62 −0.13 8.77 3.38 14.10 −7.67 12.53

OM digestibility −4.34 6.52 −2.63 5.10 −8.46 9.19 −2.01 4.95

NDF digestibility −10.56 17.57 −2.95 9.63 −23.16 9.56 −3.77 9.22

ADF digestibility −5.27 9.15 0.62 9.32 ND ND ND ND

CP digestibility −1.32 4.23 −3.69 6.79 ND ND ND ND

OM, organic matter; NDF, neutral detergent fibre; ADF, acid detergent fibre; CP, crude protein; TVFA, total volatile fatty acids; A/P, acetate/propionateratio; SD, standard deviation; ND, not determined owing to insufficient data.

saponin database. Acetate concentrations were lower by 12% tohigher by 18% compared with controls. Changes in propionateconcentration ranged from −8 to 39% compared with controls,which resulted in a mean decrease in acetate/propionate ratio(A/P) by 8.6%. Overall, digestibility of OM (4.3%), especially NDF(10.6%), was reduced by saponins, whereas digestibility of CP(−1.3%) was hardly affected. In the tannin subset, NDF andCP contents ranged from 401 to 711 g kg−1 DM and from 97to 229 g kg−1 DM respectively, suggesting that roughage-basedmedium-quality feeds were used in the studies. Like saponins,tannins affected protozoal numbers (−69 to 53%) and methaneproduction (−50 to 32%) both positively and negatively, withoverall decreases in protozoal counts and methane production. Inthis data set, total VFA (16 to 5%), acetate (−14 to 6%), propionate(−21 to 7%) and OM (−16 to 11%) and NDF (−20 to 15%)digestibilities were little affected in general. In the EO data set,there was substantial inhibition of methane production (−90 to21%) and protozoal growth (−88 to 18%) compared with controls.Digestibilities of OM (−28 to 5%) and NDF (33 to 10%) werealso reduced considerably by EO. Although the effects of EO onrumen VFA varied noticeably, overall effects were minimal. Inthe OS data set, methane production decreased by −90 to 9%.Overall, protozoal growth was stimulated (15%) by OS. While totalVFA (−21 to 19%) and acetate (−12 to 6%) concentrations wereonly slightly affected, propionate (−8 to 32%) concentrationsincreased considerably, resulting in significant reductions in A/P(−33 to 10%). The minimum and maximum values of OM andNDF digestibility changes compared with control values were −17and 3% and −27 and 5% respectively. The results presented hereshould be used with caution when input variables are outside therange of variables in the data set.

CorrelationsPearson correlation matrices for the data sets are shown inTable 2. In the saponin data set, changes in methane productioncorrelated positively (P < 0.1) with changes in protozoal numbersand OM digestibility. Other parameters did not show significant

Table 2. Simple correlation coefficients (P < 0.1) between changesin methane production and changes in rumen fermentation character-istics and digestibility as affected by different phytochemicals

Parameter Saponins Tannins Essential oils Organosulfurs

Protozoa 0.64 0.65 0.48 NS

TVFA NS 0.28 0.38 NS

Propionate NS NS 0.43 −0.32

Acetate NS 0.55 −0.55 0.61

OM digestibility 0.29 0.39 NS 0.58

NDF digestibility NS NS NS 0.67

ADF digestibility NS NS ND ND

CP digestibility NS NS ND ND

TVFA, total volatile fatty acids; OM, organic matter; NDF, neutraldetergent fibre; ADF, acid detergent fibre; CP, crude protein; NS,not significant; ND, not determined owing to insufficient data.

correlations with methane production. In the tannin data set,there were also positive correlations (P < 0.1) between changes inmethane production and changes in protozoal counts, total VFAconcentration, acetate concentration and OM digestibility. In theEO data set, changes in protozoal populations and concentrationsof total VFA and propionate showed significant (P < 0.1) positivecorrelations with changes in methane production. However,changes in acetate concentration correlated (P < 0.1) negativelywith changes in methane production. In the OS data set, protozoalnumbers and total VFA concentration did not show any correlation(P > 0.1) with changes in methane production. Changes inmethane production correlated negatively with changes inpropionate concentration but positively with changes in acetateconcentration and digestibilities of OM and NDF.

Effects of saponins on protozoa, VFA and digestibilityChanges in protozoal numbers showed a linear relationship(R2 = 0.48) with changes in methane production caused by


27

03


-0.15

-0.1

-0.05

0

0.05

0.1

0.15

-0.4 -0.2 0 0.2 0.4 0.6

Changes in protozoal numbers (log transformed)

Cha

nges

in m

etha

ne (

log

tran

sfor

med

)

Figure 1. Effect of changes in protozoal numbers on changes in methaneproduction by saponins. The relationship was as follows: log(change inmethane +1) = −0.0253 (SE = 0.00875, P < 0.01) + log(change inprotozoal numbers +1) × 0.0.184 (SE = 0.0802, P = 0.04), R2 = 0.48.

saponins (Fig. 1):

log(change in methane + 1) = −0.0253

(SE = 0.00875, P < 0.01)

+ log(change in protozoal numbers + 1)

× 0.184 (SE = 0.0802, P = 0.04)

Concentrations of total VFA and acetate did not show any relation-ship with methane due to saponins (Table 3). However, propionateproduction increased linearly (P = 0.10) with increasing inhibi-tion of methane, though the relationship was weak (R2 = 0.31).This probably resulted in the linear decrease in A/P (P = 0.07,R2 = 0.26) with decreasing methane production. Digestibilities ofall nutrients did not change with decreasing methane productionby saponins.

Effects of tannins on protozoa, VFA and digestibilityInhibition of methane production by tannins, like that by saponins,changed linearly (R2 = 0.30) with decreasing numbers of protozoa(Fig. 2):


(SE = 0.0113, P < 0.01)


× 0.327 (SE = 0.155, P = 0.08)

Concentrations of total VFA, acetate and propionate did notchange with changes in methane production by tannins (Table 3).However, A/P showed a significant linear relationship (P = 0.09,R2 = 0.27) with increasing inhibition of methane production.Digestibilities of OM (P = 0.08, R2 = 0.19) and NDF (P = 0.12, R2 =0.23) decreased linearly with decreasing methane production bytannins, but digestibility of CP did not show any relationship withmethane inhibition by tannins.

Effects of essential oils on protozoa, VFA and digestibilityFor EO, changes in methane production were linearly associatedwith changes in protozoal numbers, though the relationship was

weak (R2 = 0.20) (Fig. 3):


(SE = 0.0374, P = 0.02)


× 0.513 (SE = 0.179, P = 0.08)

Concentrations of total VFA (P = 0.06, R2 = 0.44) and propionate(P = 0.09, R2 = 0.15) changed linearly and positively withpositive changes in methane production (Table 3). However,acetate concentrations decreased linearly (P = 0.01, R2 = 0.22)with positive in methane production. A/P increased linearly(P = 0.02) with increasing inhibition of methane by EO, though therelationship was very weak (R2 = 0.17). Changes in digestibilitiesof OM and NDF were not affected by inhibition of methane by EO.

Effects of organosulfur compounds on protozoa, VFAand digestibilityMethane inhibition by OS, unlike that by other phytochemicals,was not associated with protozoal populations. Changes inconcentrations of total VFA (P = 0.07, R2 = 0.60) and acetate(P = 0.02, R2 = 0.35) decreased linearly while those of propionateincreased linearly (P = 0.06, R2 = 0.23) with increasing inhibitionof methane by OS (Table 3). Hence changes in A/P decreasedlinearly (P = 0.03, R2 = 0.30) with decreasing methane productionby OS. Digestibilities of OM and NDF were not affected by inhibitionof methane production by OS.

Effects of dietary composition on methane inhibitionby phytochemicalsIn the whole database, CP content did not affect methaneinhibition by phytochemicals. However, suppression of methaneby saponins decreased with increasing CP content in diets, but thedegree of determination was low (R2 = 0.23):

log(change in methane + 1) = −0.0875(SE = 0.0208, P < 0.01)

+ CP content (%) × 0.00392 (SE = 0.00194, P = 0.06)

For other types of phytochemicals, no significant relationshipwas evident. The above relationship signifies that methaneinhibition by saponins could be higher for low-protein dietsthan for high-protein diets. In the whole database, methaneproduction was inhibited by phytochemicals increasingly withgreater concentrations of NDF in diets (R2 = 0.15):

log(change in methane + 1) = 0.0173 (SE = 0.0431, P = 0.65)

−NDF content (%) × 0.00143 (SE = 0.00301, P = 0.09)

Data sets for individual types of phytochemicals did not revealany relationship between NDF content in diets and methanedepression by phytochemicals. This, in general, suggests that theeffect of phytochemicals on methane inhibition could be greaterfor roughage-based diets than for concentrate-based diets.

DISCUSSIONIn order to take advantage of the effects of phytochemicalson methane inhibition as well as to design future methanemitigation technologies, it is necessary to understand the effectsof phytochemicals on rumen metabolism. It appears that different


27

04


Table 3. Effects (equations for linear regressionsa) of changes in methane production on changes in rumen fermentation parameters and digestibilityas impacted by different phytochemicals

Parameter Intercept SEIntercept P Slope SESlope P RMSEAdj.R2

Saponins

TVFA 0.0068 0.0104 0.52 0.141 0.101 0.17 0.0132 0.14

Acetate −0.00285 0.00422 0.51 0.0340 0.0843 0.69 0.0125 0

Propionate 0.0175 0.00935 0.08 −0.294 0.164 0.10 0.0311 0.31

A/P −0.0178 0.00761 0.03 0.262 0.127 0.07 0.0222 0.26

OM digestibility −0.00636 0.00581 0.29 0.0857 0.0121 0.50 0.0128 0.08

NDF digestibility −0.0321 0.0141 0.04 0.0251 0.211 0.90 0.0443 0

Tannins

TVFA −0.0143 0.00606 0.04 0.135 0.133 0.36 0.0166 0.24

Acetate −0.00592 0.00461 0.23 0.0629 0.0857 0.50 0.0122 0.08

Propionate −0.0178 0.00620 0.02 0.00946 0.0678 0.89 0.0258 0

A/P −0.00337 0.00976 0.73 0.179 0.0909 0.09 0.0211 0.27


NDF digestibility −0.0157 0.0167 0.38 0.233 0.152 0.12 0.0305 0.23

Essential oils

TVFA 0.00602 0.0164 0.72 0.219 0.119 0.06 0.0164 0.44

Acetate −0.00656 0.00258 0.04 −0.0526 0.0193 0.01 0.00940 0.22

Propionate −0.00059 0.00956 0.95 0.125 0.0714 0.09 0.0348 0.15

A/P −0.0740 0.0135 0.58 −0.265 0.104 0.02 0.0501 0.17


NDF digestibility −0.0529 0.0186 0.04 0.211 0.146 0.22 0.0424 0.10

Organosulfurs

TVFA −0.0127 0.0151 0.43 0.368 0.175 0.07 0.00792 0.60

Acetate −0.00499 0.00835 0.57 0.0771 0.0319 0.02 0.0103 0.35

Propionate 0.00587 0.0210 0.72 −0.193 0.0981 0.06 0.0333 0.23

A/P −0.0149 0.0176 0.42 0.168 0.0742 0.03 0.0247 0.30


NDF digestibility 0.00319 0.0151 0.75 0.273 0.197 0.28 0.00828 0.32

TVFA, total volatile fatty acids; A/P, acetate/propionate ratio; OM, organic matter; NDF, neutral detergent fibre; SE, standard error; RMSE, root meansquare error.a No significant quadratic relationship was noted between changes in methane and changes in volatile fatty acid concentrations and digestibility.Data on changes in methane production, rumen fermentation and digestibility are logarithmically transformed (log(x + 1), where x is the variable).

-0.2

-0.15

-0.05

-0.1

0.05

0

0.1

0.15

-0.4 -0.2 0 0.2 0.4


Cha

nges

in m

etha

ne (

log

tran

sfor

med

)

Figure 2. Effect of changes in protozoal numbers on changes in methaneproduction by tannins. The relationship was as follows: log(change inmethane +1) = −0.0609 (SE = 0.0113, P < 0.01) + log(change inprotozoal numbers +1) × 0.327 (SE = 0.155, P = 0.08), R2 = 0.30.

-0.3

-0.25

-0.15

-0.2

-0.05

0.05

-0.1

0

0.1

-0.4 -0.2 0-0.1 0.1 0.2


Cha

nges

in m

etha

ne (

log

tran

sfor

med

)

Figure 3. Effect of changes in protozoal numbers on changes in methaneproduction by essential oils. The relationship was as follows: log(changein methane +1) = −0.0628 (SE = 0.0374, P = 0.02) + log(change inprotozoal numbers +1) × 0.513 (SE = 0.179, P = 0.08), R2 = 0.20.


27

05


types of phytochemicals such as saponins, tannins, EO andOS have different effects on rumen fermentation and micro-organisms. Inhibition of methane production by saponins wasassociated with an increase in propionate concentration andthus resulted in a decrease in A/P. For tannins, a decrease inmethane emission did not relate to changes in acetate andpropionate concentrations, although A/P decreased quadraticallywith inhibition of methane. For EO, acetate and propionateconcentrations relative to controls decreased quadratically andlinearly respectively with increasing inhibition of methane. Unlikesaponins, tannins and OS, A/P increased with decreasing methaneproduction by EO. OS-mediated methane inhibition resulted ina linear decrease in acetate concentration, a linear increase inpropionate concentration and a linear decrease in A/P, whichwas very similar to decreases in methane by direct inhibition ofmethanogens in the rumen.

According to the results of this meta-analysis, it appears thatdifferent types of phytochemicals inhibit methane by differentmodes, resulting in different patterns of rumen fermentation.A decrease in methane production by phytochemicals may beconfounded with several factors such as suppression of proto-zoa, archaea and hydrogen-producing microbial populations anddecreased fibre digestion in the rumen.6,16 From the relation-ship between changes in methane production and protozoalcounts for different phytochemicals, it is apparent that each 1%suppression of protozoal numbers accounted for about 0.17,0.29 and 0.45% inhibition of methane by saponins, tannins andEO respectively. Thus methane inhibition by saponins perhapsresults predominantly from decreased protozoal populations.Methanogens associated with protozoa account for decreasedmethane production of about 9–25%17 or as much as 37%.18

Sterol-binding capability of saponins19 has been implicated inthe destruction of protozoal cell membranes.9,16 A decrease inthe activities of methane-producing genes or rate of methaneproduction per methanogenic cell by saponins has also beensuggested,20,21 which might increase propionate production as aresult of rechannelling of hydrogen from methane to propionate.22

Hence a decrease in protozoal numbers by saponins may result inan increase in propionate concentration and a decrease in A/P. Be-sides, saponins sometimes stimulate the growth of Selenomonasruminantium,23 which predominantly produces propionate fromsuccinate metabolism.

Suppression of methane by tannins could entail reductionsin numbers of protozoa and methanogens depending on thechemical structure of tannins and methanogenic species.6,24

Besides their direct effect on methanogens, tannins exertantimicrobial action on microbial growth, including cellulolyticbacteria and fungi,16 which could decrease hydrogen availabilityand thus lower methanogenesis25 to some extent. Thereforepropionate and acetate production was not significantly affectedby tannin-induced suppression of methane, but A/P decreasedwith increasing methane inhibition.

In contrast to saponins, EO-mediated inhibition of methanecould primarily involve suppression of methanogens7,9,16 andhydrogen-producing micro-organisms such as Lachnospira multi-parus,Ruminococcusalbus and Ruminococcusflavifaciens, includinginhibition of protozoal growth.7,16 Because of less formation ofhydrogen, propionate concentrations decrease, resulting in anincrease in A/P. Besides, EO such as thymol may decrease pro-pionate formation by inhibiting the growth of S. ruminantium.26

These characteristic changes in rumen fermentation by EO are

not nutritionally favourable for ruminant animals in the con-text of energy utilisation. In an in vitro study with mixed rumenmicrobial populations, Evans and Martin26 also reported that thy-mol (a component of essential oils) decreased concentrationsof acetate, propionate, methane and hydrogen and increasedA/P.

OS might specifically inhibit methanogenic archaea.7,27 Viaanalysis by real-time polymerase chain reaction, McAllisterand Newbold22 observed that allicin at 20 mg L−1 decreasedmethanogen numbers in Rusitec, but not at 2 mg L−1. It hasbeen suggested that OS found in garlic oil perhaps directlyinhibit rumen methanogenic archaea through inhibition of theenzyme 3-hydroxy-3-methyl-glutaryl coenzyme A (HMG-CoA)reductase.27 Methanogenic archaea have unique membrane lipidsthat contain glycerol joined by ether linkages to long-chainisoprenoid alcohols, which is not present in other rumen micro-organisms.28 The synthesis of isoprenoid units in methanogenicarchaea is catalysed by HMG-CoA reductase.27 Inhibition ofmethane production usually increases the concentration ofhydrogen, which is rechannelled to other hydrogen sinkssuch as propionate, resulting in increased concentration ofpropionate.22 Conversely, production of acetate in the rumenresults in large quantities of hydrogen and depends on theavailability of reducing equivalents27 such as NAD+. The highpartial pressure of hydrogen and high NADH/NAD+ ratio inthe rumen due to inhibition of methanogenesis may resultin a decrease in acetate production.27,29 The characteristicsof VFA patterns noted due to suppression of methane byOS suggest a direct effect on methanogens without mucheffect on other rumen microbial populations. A strategy forsuppressing methane inhibition from ruminants would onlybe practical if it has no hostile effects on ruminal dynamics.Among the phytochemicals, OS hold promise to decrease entericmethane emissions, as they do not exert much adverse effecton ruminal fermentation and digestibility compared with otherphytochemicals.

CONCLUSIONSFrom the patterns of changes in protozoal numbers and VFA profileby different groups of phytochemicals, it appears that methaneinhibition by saponins may predominantly involve suppressionof protozoal growth, which is responsible for greater propionateproduction and lower A/P. However, the decrease in methaneproduction by tannins may be due to inhibition of protozoa,methanogens and, to a lesser extent, hydrogen-producingmicrobial population. For EO, depression of methane productionmay primarily entail inhibition of methanogens, hydrogen-producing rumen micro-organisms and, to a lesser extent,protozoa, resulting in decreased concentrations of propionate andacetate and increased A/P. OS may have a direct antimethanogeniceffect, which results in decreased acetate, increased propionateand decreased A/P. The different phytochemicals exert theircharacteristic effects on rumen fermentation in association withthe inhibition of methanogenesis.

ACKNOWLEDGEMENTThe research grant provided by the Indian Council of AgriculturalResearch, New Delhi is gratefully acknowledged.


27

06


APPENDIX

Table A1. List of previously published studies included in database for analysing changes in rumen fermentation characteristics associated withmethanogenesis

Phytochemicals Study Diet METH PROT VFA DIG

Saponins Goel et al.1 Hay; hay/concentrate (32 : 68) Y Y N N

Hess et al.2 Forage/concentrate (2 : 1) Y Y Y Y

Hess et al.3 Meadow grass/Arachis pintoi hay/barley straw (56 : 22:11) Y Y Y Y

Hess et al.4 Brachiaria grass (66–100%) Y Y Y Y

Holtshausen et al.5 Barley silage/concentrate (51 : 49) Y N Y Y

Holtshausen et al.5 Barley silage/concentrate (51 : 49) Y Y Y Y

Hu et al.6 Grass hay/corn (1 : 1) Y Y Y Y

Hu et al.7 Grass hay/corn (1 : 1) Y Y Y Y

Klita et al.8 Grass hay Y N Y Y

Lila et al.9 Corn starch; potato starch; hay/concentrate (3 : 2) Y Y Y Y

Patra et al.10 Wheat straw/concentrate (1 : 1) Y Y Y Y

Pen et al.11 Ryegrass hay/concentrate (3 : 2) Y Y Y Y

Santoso et al.12 Orchard grass silage/concentrate (7 : 3) Y N Y Y

Sliwinski et al.13 Hay/barley-based concentrate (1 : 1) Y Y Y Y

Sliwinski et al.14 Grass silage and hay/barley (77 : 23) Y Y Y Y

Wang et al.15 Mixed hay/concentrate (3 : 1) Y N Y Y

Tannins Animut et al.16 Lespedeza forage/sorghum-sudan grass (1 : 2, 2 : 1 and 3 : 0) Y Y Y Y

Beauchemin et al.17 Barley silage 69.5%, barley grain 21.5% Y N Y Y

Bhatta et al.18 Timothy hay/concentrate (65 : 35) Y Y Y Y

Bodas et al.19 Lucerne/grass hay/barley grain (5 : 4:1) Y N Y Y

Carrula et al.20 Ryegrass/lucerne (1 : 1) Y Y Y Y

Hariadi and Santoso21 Elephant grass Y Y Y Y

Hess et al.4 Brachiaria grass (66–100%) Y Y Y Y




Puchala et al.24 Lespedeza cuneata pasture Y N Y Y

Sliwinski et al.13 Hay/barley-based concentrate (1 : 1) Y Y Y Y

Sliwinski et al.14 Grass silage and hay/barley (77 : 23) Y Y Y Y

Essential oils Agarwal et al.25 Wheat straw/concentrate (1 : 1) Y Y Y Y

Beauchemin and McGinn26 Barley silage/concentrate (3 : 1) Y N Y Y

Chaves et al.27 Soluble starch Y Y Y Y

Kumar et al.28 Wheat straw/concentrate (1 : 1) Y Y Y Y


Sallam et al.30 Roughage/concentrate (1 : 1) Y Y N Y

Tatsuoka et al.31 Glucose/cellobiose (1 : 1) Y N Y Y

Wang et al.15 Mixed hay/concentrate (3 : 1) Y N Y Y

Organosulfurs Busquet et al.32 Lucerne hay/concentrate (1 : 1) Y N Y Y

Chaves et al.27 Soluble starch Y Y Y Y

Garcia-Gonzalez et al.33 Lucerne/grass hay/barley (5 : 2:3) Y N Y Y

Kamel et al.34 Lucerne hay/concentrate (1 : 1) Y N Y Y

Lila et al.35 Corn starch; potato starch; hay/concentrate (3 : 2) Y Y Y Y

Mohammed et al.36 Sudan grass/concentrate (3 : 2) Y Y Y Y



Tatsuoka et al.31 Soluble sugar Y N Y Y

METH, methane; PROT, protozoa; VFA, volatile fatty acids; DIG, digestibility; Y, reported; N, not reported.1 Goel G, Makkar HPS and Becker K, Effect of Sesbania sesban and Carduus pycnocephalus leaves and fenugreek (Trigonella foenum-graecum L.) seedsand their extracts on partitioning of nutrient from roughage and concentrate based feeds to methane. Anim Feed Sci Technol 147:72–89 (2008).2 Hess HD, Beuret RA, Lotscher M, Hindrichsen IK, Machmuller A, Carulla JE, et al., Ruminal fermentation, methanogenesis and nitrogen utilization ofsheep receiving tropical grass hay–concentrate diets offered with Sapindus saponaria fruits and Cratylia argentea foliage. Anim Sci 79:177–189 (2004).3 Hess HD, Kreuzer M, Diaz TE, Lascano CE, Carulla JE and Solvia CR, Saponin rich tropical fruits affect fermentation and methanogenesis in faunatedand defaunated fluid. Anim Feed Sci Technol 109:79–94 (2003).4 Hess HD, Monsalve LM, Lascano CE, Carulla JE, Diaz TE and Kreuzer M, Supplementation of a tropical grass diet with forage legumes and Sapindussaponaria fruits: effects on in vitro ruminal nitrogen turnover and methanogenesis. Aust J Agric Res 54:703–713 (2003).


27

07


Table A1. (Continued)

5 Holtshausen L, Chaves AV, Beauchemin KA, McGinn SM, McAllister TA, Odongo NE, et al., Feeding saponin-containing Yucca schidigera and Quillajasaponaria to decrease enteric methane production in dairy cows. J Dairy Sci 92:2809–2821 (2009).6 Hu W, Wu Y, Liu J, Guo Y and Ye J, Tea saponins affect in vitro fermentation and methanogenesis in faunated and defaunated rumen fluid. J ZhejiangUniv Sci B 6:787–792 (2005).7 Hu W, Liu J, Wu Y, Guo Y and Ye J, Effects of tea saponins on in vitro ruminal fermentation and growth performance in growing Boer goat. Arch AnimNutr 60:89–97 (2006).8 Klita PT, Mathison GW, Fenton TW and Hardin RT, Effects of alfalfa root saponins on digestive function in sheep. J Anim Sci 74:1144–1156 (1996).9 Lila ZA, Mohammed N, Kanda S, Kamada T and Itabashi H, Effect of sarsaponin on rumen fermentation with particular reference to methaneproduction in vitro. J Dairy Sci 86:3330–3336 (2003).10 Patra AK, Kamra DN and Agarwal N, Effect of plant extracts on in vitro methanogenesis, enzyme activities and fermentation of feed in rumen liquorof buffalo. Anim Feed Sci Technol 128:276–291 (2006).11 Pen B, Takaura K, Yamaguchia S, Asa R and Takahashi J, Effects of Yucca schidigera and Quillaja saponaria with or without β-1,4 galacto-oligosaccharides on ruminal fermentation, methane production and nitrogen utilization in sheep. Anim Feed Sci Technol 138:75–88 (2007).12 Santoso B, Mwenya B, Sar C, Gamo Y, Kobayashi T, Morikawa R, et al., Effects of supplementing galacto-oligosaccharides, Yucca schidigera and nisinon rumen methanogenesis, nitrogen and energy metabolism in sheep. Livest Prod Sci 91:209–217 (2004).13 Sliwinski BJ, Kreuzer M, Wettstein HR and Machmuller A, Rumen fermentation and nitrogen balance of lambs fed diets containing plant extractsrich in tannins and saponins and associated emissions of nitrogen and methane. Arch Anim Nutr 56:379–392 (2002).14 Sliwinski BJ, Solvia CR, Machmuller A and Kreuzer M, Efficacy of plant extracts rich in secondary constituents to modify rumen fermentation. AnimFeed Sci Technol 101:101–114 (2002).15 Wang CJ, Wang SP and Zhou H, Influences of flavomycin, ropadiar, and saponin on nutrient digestibility, rumen fermentation, and methaneemission from sheep. Anim Feed Sci Technol 148:157–166 (2009).16 Animut G, Goetsch AL, Puchala R, Patra AK, Sahlu T, Varel VH, et al., Methane emission by goats consuming diets with different levels of condensedtannins from lespedeza. Anim Feed Sci Technol 144:212–227 (2008).17 Beauchemin KA, McGinn SM, Martinez TF and McAllister TA, Use of condensed tannin extract from quebracho trees to reduce methane emissions.J Anim Sci 85:1990–1996 (2007).18 Bhatta R, Uyeno Y, Tajima K, Takenaka A, Yabumoto Y, Nonaka I, et al., Difference in the nature of tannins on in vitro ruminal methane and volatilefatty acid production and on methanogenic archaea and protozoal populations. J Dairy Sci 92:5512–5522 (2009).19 Bodas R, Lopez S, Fernandez M, Garcia-Gonzalez R, Rodrıguez AB, Wallace RJ, et al., In vitro screening of the potential of numerous plant species asantimethanogenic feed additives for ruminants. Anim Feed Sci Technol 145:245–258 (2008).20 Carulla JE, Kreuzer M, Machmuller A and Hess HD, Supplementation of Acacia mearnsii tannins decreases methanogenesis and urinary nitrogen inforage-fed sheep. Aust J Agric Res 56:961–970 (2005).21 Hariadi BT and Santoso B, Evaluation of tropical plants containing tannin on in vitro methanogenesis and fermentation parameters using rumenfluid. J Sci Food Agric 90:456–461 (2010).22 Patra AK, Kamra DN and Agarwal N, Effect of extracts of leaves on rumen methanogenesis, enzyme activities and fermentation in in vitro gasproduction test. Indian J Anim Sci 78:91–96 (2008).23 Patra AK, Kamra DN and Agarwal N, Effect of plants containing secondary metabolites on in vitro methanogenesis, enzyme profile and fermentationof feed with rumen liquor of buffalo. Anim Nutr Feed Technol 6:203–213 (2006).24 Puchala R, Min BR, Goetsch AL and Sahlu T, The effect of a condensed tannin-containing forage on methane emission by goats. J Anim Sci83:182–186 (2005).25 Agarwal N, Shekhar C, Kumar R, Chaudhary LC and Kamra DN, Effect of peppermint (Mentha piperita) oil on in vitro methanogenesis and fermentationof feed with buffalo rumen liquor. Anim Feed Sci Technol 148:321–327 (2009).26 Beauchemin KA and McGinn SM, Methane emissions from beef cattle: effects of fumaric acid, essential oil, and canola oil. J Anim Sci 84:1489–1496(2006).27 Chaves AV, He ML, Yang WZ, Hristov AN, McAllister TA and Benchaar C, Effects of essential oils on proteolytic, deaminative and methanogenicactivities of mixed ruminal bacteria. Can J Anim Sci 89:97–104 (2008).28 Kumar R, Kamra DN, Agrawal N and Chaudhary LC, Effect of eucalyptus (Eucalyptus globulus) oil on in vitro methanogenesis and fermentation offeed with buffalo rumen liquor. Anim Nutr Feed Technol 9:237–243 (2009).29 Patra AK, Kamra DN and Agarwal N, Effects of extracts of spices on rumen methanogenesis, enzyme activities and fermentation of feeds in vitro. JSci Food Agric 90:511–520 (2010).30 Sallam SMA, Bueno ICS, Brigide P, Godoy PB, Vitti DMSS and Abdalla AL, Efficacy of eucalyptus oil on in vitro rumen fermentation and methaneproduction. Options Mediterraneennes 85:267–272 (2009).31 Tatsuoka N, Hara K, Mikuni K, Hara K, Hashimoto H and Itabashi H, Effects of the essential oil cyclodextrin complexes on ruminal methane productionin vitro. Anim Sci J 79:68–75 (2008).32 Busquet M, Calsamiglia S, Ferret A, Carro MD and Kamel C, Effect of garlic oil and four of its compounds on rumen microbial fermentation. J DairySci 88:4393–4404 (2005).33 Garcia-Gonzalez R, Lopez S, Fernandez M, Bodas R and Gonzalez JS, Screening the activity of plants and spices for decreasing ruminal methaneproduction in vitro. Anim Feed Sci Technol 147:36–52 (2008).34 Kamel C, Greathead HMR, Tejido ML, Ranilla MJ and Carro MD, Effects of allicin and diallyl disulfide on in vitro rumen fermentation of a mixed diet.Anim Feed Sci Technol 145:351–363 (2008).35 Lila ZA, Mohammed N, Kanda S, Kamada T and Itabashi H, Effect of α-cyclodextrin allyl isothiocyanate on ruminal microbial methane productionin vitro. Anim Sci J 74:321–326 (2003).36 Mohammed N, Ajisaka N, Lila ZA, Hara K, Mikuni K, Hara K, et al., Effect of Japanese horseradish oil on methane production and ruminal fermentationin vitro and in steers. J Anim Sci 82:1839–1846 (2004).


27

08


REFERENCES1 IPCC, Summary for policymakers, in Climate Change 2007: the Physical

ScienceBasis. ContributionofWorkingGroupItotheFourthAssessmentReport of the Intergovernmental Panel on Climate Change, ed. bySolomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, et al.Cambridge University Press, Cambridge, pp. 1–18 (2007).

2 Johnson KA and Johnson DE, Methane emission from cattle. J Anim Sci73:2483–2492 (1995).

3 Bruinenberg MH, van der Honing Y, Agnew RE, Yan T, van Vuuren AMand Valk H, Energy metabolism of dairy cows fed on grass. LivestProd Sci 75:117–128 (2002).

4 Nkrumah JD, Okine EK, Mathison GW, Schmid K, Li C, Basarab JA, et al,Relationships of feedlot feed efficiency, performance, and feedingbehavior with metabolic rate, methane production, and energypartitioning in beef cattle. J Anim Sci 84:145–153 (2006).

5 Beauchemin KA, McGinn SM, Martinez TF and McAllister TA, Use ofcondensed tannin extract from quebracho trees to reduce methaneemissions from cattle. J Anim Sci 85:1990–1996 (2007).

6 Animut G, Goetsch AL, Puchala R, Patra AK, Sahlu T, Varel VH, et al,Methane emission by goats consuming diets with different levelsof condensed tannins from lespedeza. Anim Feed Sci Technol144:212–227 (2008).

7 Hart KJ, Yanez-Ruiz DR, Duval SM, McEwan NR and Newbold CJ, Plantextracts to manipulate rumen fermentation. Anim Feed Sci Technol147:8–35 (2008).

8 Kamra DN, Patra AK, Chatterjee PN, Kumar R, Agarwal N andChaudhary LC, Effect of plant extract on methanogenesis andmicrobial profile of the rumen of buffalo: a brief overview. AustJ Exp Agric 48:175–178 (2008).

9 Patra AK and Saxena J, A review of the effect and mode of action ofsaponins on microbial population and fermentation in the rumenand ruminant production. Nutr Res Rev 22:204–219 (2009).

10 Patra AK, Kamra DN and Agarwal N, Effects of extracts of spices onrumen methanogenesis, enzyme activities and fermentation offeeds in vitro J Sci Food Agric 90:511–520 (2010).

11 Patra AK, Responses of intake, digestibility and nitrogen utilization ingoats fed low-quality roughages supplementing with tree foliages.J Sci Food Agric 89:1462–1472 (2009).

12 Sauvant D, Schmidely P, Daudin JJ and St-Pierre NR, Meta-analyses ofexperimental data in animal nutrition. Animal 2:1203–1214 (2008).

13 Patra AK, Aspects of nitrogen utilization in sheep fed mixeddiets containing foliages from trees and browses. Br J Nutr103:1319–1330 (2010).

14 SAS, SAS/STAT User’s Guide, Version 8.2. SAS Institute, Cary, NC (2001).15 St-Pierre NR, Integrating quantitative findings from multiple studies

using mixed model methodology. J Dairy Sci 84:741–755 (2001).

16 Patra AK and Saxena J, Dietary phytochemicals as rumen modifiers:a review of the effects on microbial populations. Antonie vanLeeuwenhoek 96:363–375 (2009).

17 Newbold CJ, Lassalas B and Jouany JP, The importance ofmethanogenesis associated with ciliate protozoa in ruminalmethane production in vitro. Lett Appl Microbiol 21:230–234 (1995).

18 Finlay BJ, Esteban G, Clarke KJ, Williams AG, Embley TM and Hirt RR,Some rumen ciliates have endosymbiotic methanogens. FEMSMicrobiol Lett 117:157–162 (1994).

19 Hostettmann K and Marston A, Saponins. Cambridge University Press,Cambridge (1995).

20 Guo YQ, Liu J-X, Lu Y, Zhu WY, Denman SE and McSweeney CS, Effectof tea saponin on methanogenesis, microbial community structureand expression of mcrA gene, in cultures of rumen micro-organisms.Lett Appl Microbiol 47:421–426 (2008).

21 Hess HD, Monsalve LM, Lascano CE, Carulla JE, Diaz TE and Kreuzer M,Supplementation of a tropical grass diet with forage legumesand Sapindus saponaria fruits: effects on in vitro ruminal nitrogenturnover and methanogenesis. Aust J Agric Res 54:703–713 (2003).

22 McAllister TA and Newbold CJ, Redirecting rumen fermentation toreduce methanogenesis. Aust J Exp Agric 48:7–13 (2008).

23 Wang Y, McAllister TA, Yanke LJ and Cheeke PR, Effect of steroidalsaponin from Yucca schidigera extract on ruminal microbes. J ApplMicrobiol 88:887–896 (2000).

24 Tavendale MH, Meagher LP, Pacheco D, Walker N, Attwood GTand Sivakumaran S, Methane production from in vitro rumenincubations with Lotus pedunculatus and Medicago sativa,and effects of extractable condensed tannin fractions onmethanogenesis. Anim Feed Sci Technol 123/124:403–419 (2005).

25 Carulla JE, Kreuzer M, Machmuller A and Hess HD, Supplementationof Acacia mearnsii tannins decreases methanogenesis and urinarynitrogen in forage-fed sheep. Aust J Agric Res 56:961–970 (2005).

26 Evans JD and Martin SA, Effects of thymol on ruminal micro-organisms.Curr Microbiol 41:336–340 (2000).

27 Busquet M, Calsamiglia S, Ferret A, Carro MD and Kamel C, Effectof garlic oil and four of its compounds on rumen microbialfermentation. J Dairy Sci 88:4393–4404 (2005).

28 De Rosa M, Gambacorta A and Gliozzi A, Structure, biosynthesis, andphysicochemical properties of archaebacterial lipids. Microbiol Rev50:70–80 (1986).

29 Miller TL, Ecology of methane production and hydrogen sinks inthe rumen, in Ruminant Physiology: digestion, metabolism, growthand reproduction. ed. by Engelhardt W, Leonhard-Marek S, Breves Gand Giesecke D. Ferdinand Enke Verlag, Stuttgart, Germany,pp. 317–331 (1995).


27

09

Research ArticleReceived: 15 April 2010 Revised: 22 June 2010 Accepted: 30 July 2010 Published online in Wiley Online Library: 2 September 2010


Red mold dioscorea-induced G2/M arrestand apoptosis in human oral cancer cellsWei-Hsuan Hsu, Bao-Hong Lee and Tzu-Ming Pan∗

Abstract

BACKGROUND: Monascus-fermented products are among the most commonly used traditional food supplements. Dioscorea isknown to exhibit anticancer properties. In this study the effects of the ethanol extract of red mold dioscorea (RMDE) on cellproliferation, cell cycle and apoptosis in human oral cancer cells were investigated.

RESULTS: RMDE exercised growth inhibition on squamous cell carcinoma-25 (SCC-25) cells. RMDE-mediated G2/M phase arrestwas associated with the down-regulation of NF-κB, resulting in the inhibition of cyclin B1 and CDK1 expression; this may be themechanism by which RMDE inhibits cancer cells. Furthermore, the proapoptotic activity of RMDE was revealed by the AnnexinV-FITC/PI double-staining assay. In addition, the proapoptotic effect of RMDE was evident by the inhibition of Bax expressionin the mitochondria, resulting in the activation of caspase-9 and caspase-3 and subsequent triggering of the mitochondrialapoptotic pathway. RMDE also enhanced caspase-8 activity, indicating the involvement of the death receptor pathway inRMDE-mediated SCC-25 cell apoptosis.

CONCLUSION: RMDE treatment inhibited the growth of SCC-25 cells by arresting cell cycle at the G2/M phase and inducedapoptosis in a time- and dose-dependent manner. Therefore RMDE may be a good candidate for development as a dietarysupplement against oral cancer.c© 2010 Society of Chemical Industry

Supporting information may be found in the online version of this article.

Keywords: apoptosis; cell cycle; red mold dioscorea; Monascus-fermented products; oral cancer

INTRODUCTIONOral cancer is the fifth most common neoplasm worldwide,accounting for more than 500 000 cases annually.1 In Taiwanit has the fastest-rising incidence and mortality rate of any cancerand is the sixth most common cause of cancer death, beingmore prevalent in males than in females. Tobacco and alcoholconsumption have been reported to be the major factors in thedevelopment of oral cancer.2 Diets low in carotenoids and vitaminA, poor oral hygiene and indoor air pollution are also recognisedas factors in oral cancer.3,4 However, betel quid chewing is oneof the most important causes of oral cancer in Taiwan, with highmortality and poor prognosis. Therefore, in an effort to improvepatient survival and quality of life, new therapeutic approachesfocusing on the molecular target and mechanism that mediatetumour cell growth or cell death have gained much attention.

In recent years, natural food products have received increasedattention because of their potential role in the preventionand/or intervention of carcinogenesis and neoplastic progression.Monascus species have been used as traditional food fungi ineastern Asia for several centuries. In previous studies, Monascus-fermented rice, known as red mold rice (RMR), was found toshow antioxidative ability5 as well as anti-Alzheimer’s diseaseand anticancer effects.6 RMR has many functional secondarymetabolites. One of these secondary metabolites, monacolinK, acts by competitively inhibiting 3-hydroxy-3-methylglutarylcoenzyme A (HMG-CoA) reductase in the cholesterol biosyntheticpathway7 and also plays a role in antitumorigenic activity.8 In

addition, several yellow pigments such as monascin and ankaflavinfrom Monascus have been reported to have anti-inflammatorypotential and cytotoxic/cytostatic activities.9,10

Monascus-fermented dioscorea, known as red mold dioscorea(RMD), comprises a dioscorea root substance as well as severalMonascus metabolites. Dioscorea is regarded as a functional foodor a valuable herb because of its content of many beneficialingredients for the prevention of various diseases.11 Dioscorin,polysaccharides, flavones, vitamin C and polyphenols of dioscoreahave been proven to exhibit high antioxidative ability.12 Inaddition, dioscorea shows antitumour activity.13,14 Thus Monascusfermentation of dioscorea should strengthen the anticancer effectsof RMD. In a previous study the ethanol extract of RMR (RMRE)was found to effectively inhibit oral cancer carcinogenesis in ahamster buccal pouch model.15 Hence Monascus fermentation ofdioscorea may lead to stronger anticancer effects. In the presentstudy, to understand the effect of the ethanol extract of RMD(RMDE) fermented by Monascus purpureus NTU 568 on humanoral cancer cells, we selected human tongue cancer squamouscell carcinoma-25 (SCC-25) cells for examining the effects on cell

∗ Correspondence to: Tzu-Ming Pan, Institute of Microbiology and Biochemistry,College of Life Science, National Taiwan University, No. 1, Sec. 4, RooseveltRoad, Taipei 10617, Taiwan. E-mail: [email protected]

Institute of Microbiology and Biochemistry, College of Life Science, NationalTaiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei 10617, Taiwan


27

10

www.soci.org W-H Hsu, B-H Lee, T-M Pan

viability, cell proliferation and the mechanisms underlying theinduction of cell cycle arrest and apoptosis.

MATERIALS AND METHODSChemicals and reagentsCrystal violet, propidium iodide (PI), sodium dodecyl sulfate (SDS),Triton X-100, trypsin, trypan blue, Folin–Ciocalteu’s reagent, gallicacid and quercetin were purchased from Sigma Chemical Co. (StLouis, MO, USA). Foetal bovine serum (FBS) was purchased fromLife Technologies (Auckland, New Zealand). Dimethyl sulfoxide(DMSO) was purchased from Wako Pure Chemical Industries(Saitama, Japan). Ethanol (950 mL L−1) was purchased from TaiwanTobacco and Liquor Corp. (Taipei, Taiwan). Dulbecco’s modifiedEagle’s medium (DMEM), Ham’s F-12, sodium bicarbonate,hydrocortisone, penicillin and streptomycin were purchasedfrom HyClone Laboratories (Logan, UT, USA). Aluminium nitrate(Al(NO3)3) and potassium acetate (CH3COOK) were purchasedfrom J.T. Baker Co. (Phillipsburg, NJ, USA).

Preparation of RMDE and RMREThe M. purpureus NTU 568 culture strain was maintained onpotato dextrose agar (Difco Co., Detroit, MI, USA) slants at 10 ◦Cand transferred monthly. Dioscorea root (Dioscorea batatas Dence)purchased from a local supermarket in Taiwan was used to produceRMD by the method of solid state culture.14 The preparation ofRMR was carried out with a substrate of long-grain rice (Oryzasativa) using the method of solid state culture. Briefly, 500 g ofsubstrate was soaked in deionised water for 8 h and excess waterwas removed with a sieve. The substrate was autoclaved (HL-341,Gemmy Corp., Taipei, Taiwan) at 121 ◦C for 20 min in a wooden ‘kojidish’ with dimensions of 30 cm × 20 cm × 5 cm. After the substratehad cooled, it was inoculated with 50 g L−1 spore suspension andcultivated at 30 ◦C for 10 days. During the culturing stage, 100 mLof water was added daily to the substrate from the second dayto the fifth day. At the end of cultivation the crushed and driedproduct with the mold was extracted with 950 mL L−1 ethanolor water at 50 ◦C for 3 days. Extracts were further freeze-dried topowder and stored at −20 ◦C until use. Samples were dissolvedin DMSO and the concentration was kept below 1 mL L−1 in theexperimental design.

Determination of total phenolic compounds and flavonoidsFor the determination of total phenolic compounds, freeze-dried dioscorea or RMD powder was dissolved in deionisedwater and the concentration of total phenolic compoundswas measured spectrophotometrically using Folin–Ciocalteu’sreagent. Sample solution (100 µL), Folin–Ciocalteu’s reagent(500 µL), sodium carbonate (400 µL, 75 g L−1) and deionised water(5 mL) were mixed thoroughly and kept at room temperature for30 min before the absorbance at 760 nm was measured. Totalphenolic content was determined using gallic acid as standard.For the determination of flavonoids, sample solution (500 µL),ethanol (1.5 mL), Al(NO3)3 (100 µL, 100 g L−1), CH3COOK (100 µL,1 mol L−1) and water (2.8 mL) were mixed thoroughly and kept atambient temperature for 40 min before the absorbance at 415 nmwas measured with a spectrophotometer (U-2000, Hitachi, Tokyo,Japan). Total flavonoid content was calculated according to astandard curve established with quercetin. The results indicatedthat the levels of total phenols and flavonoids were 119 and191 mg kg−1 dioscorea respectively and 179 and 249 mg kg−1

RMD respectively.

Cell cultureHuman SCC-4, SCC-9 and SCC-25 tongue, FaDu and HEp-2 pharynxcell lines were obtained from the Bioresource Collection andResearch Center (Hsinchu, Taiwan). SCC cells were maintainedin DMEM/Ham’s F-12 (1 : 1 v/v) medium supplemented with100 mL L−1 FBS, 1.5 g L−1 sodium bicarbonate, 400 ng mL−1

hydrocortisone and 10 mL L−1 antibiotic/antimitotic solution.FaDu and HEp-2 cells were maintained in DMEM mediumcontaining 100 mL L−1 FBS and 10 mL L−1 antibiotic/antimitoticsolution. Cells were incubated in 5% CO2/95% humidifiedatmosphere at 37 ◦C.

Determination of cell viabilityThe cell-killing effect of RMDE and RMRE against oral cancer cellswas measured using the crystal violet staining assay. Cells wereseeded on 24-well plates (3 × 104 cells per well) and treatedwith various concentrations of RMDE and RMRE for 24 and48 h. The medium was then removed, washed with phosphate-buffered saline (PBS) and stained with 2 g L−1 crystal violet in100 mL L−1 phosphate-buffered formaldehyde for 20 min beforebeing washed with water. The crystal violet bound to the cellswas dissolved in 20 g L−1 SDS solution and its absorbance at600 nm was measured. The 50% inhibitory concentration (IC50)of RMDE/RMRE was calculated from a sigmoidal dose–responsecurve.

Clonogenic survival assayCells were seeded on six-well plates at 250 cells per well in afinal volume of 2 mL of medium containing either vehicle or anappropriate RMDE concentration. All cultures were incubated foran additional 14 days until colonies were large enough to beclearly discerned. At this point the medium was aspirated and thedishes were washed once with PBS and stained with a filteredsolution of 5 g L−1 crystal violet for 10 min. Colonies containing50 cells were scored. Clonogenic survival was expressed as thepercentage of colonies formed in RMDE-treated cells with respectto vehicle-treated cells.

Cell cycle distributionAfter 12 and 24 h of exposure to RMDE the medium was aspiratedand adherent cells were harvested and centrifuged at 300 × g for5 min. Cells were washed with PBS, fixed with 700 mL L−1 ice-coldethanol at −20 ◦C overnight and then stained with PI at roomtemperature for 30 min. The cell cycle distribution was analysedby flow cytometry using a FACScan-LSR flow cytometer equippedwith CellQuest software (BD Biosciences, San Jose, CA, USA).

Apoptosis analysisFor apoptosis detection, floating cells in the medium and adherentcells were collected after 12 and 24 h of RMDE treatment. Cellswere harvested, washed in ice-cold PBS and resuspended in 200 µLof binding buffer before being incubated in 5 µL of Annexin V-FITC(BD Biosciences) solution and 5 µL of PI at room temperature for15 min in the dark. Then 300 µL of binding buffer was added. Cellswere analysed by flow cytometry. Untreated cells were used as thecontrol for double staining.

Reverse transcription polymerase chain reaction (RT-PCR)Total RNA was isolated using Trizol (Life Technologies) accord-ing to the manufacturer’s instructions. A 5 µg aliquot of purified


27

11

Effects of red mold dioscorea on SCC-25 cells www.soci.org

total RNA was employed for reverse transcription using Super-Script III (Life Technologies). The reaction mixture was incubatedat 42 ◦C for 1 h and the reaction was terminated by heatingat 70 ◦C for 10 min. Amplification of the RT product by PCRwas performed using Promega Taq DNA Polymerase (PromegaCo., Madison, WI, USA). All reactions were carried out in a ther-mal cycler (Model 2400, Perkin-Elmer, Norwalk, CT, USA) withthe following primers: GAPDH sense, 5′-CATCAAGAAGGTGGT-GAAGCAG-3′, and antisense, 5′-CCACCACCCTGTTGCTGTAGCCA-3′, GST-P sense, 5′-TCATCTACACCAACTATGAG-3′ , and antisense,5′-GCCACATAGGCAGAGAGCAG-3′, cyclin B1 sense, 5′-CTTATAC-TAAGCACCAAATC-3′ , and antisense, 5′-CTTGGCTAAATCTTGAACT-3′, and CDK1 sense, 5′-CTTATGCAGGATTCCAGGTT-3′, and an-tisense, 5′-GGTGCCTATACTCCAAATGTC-3′ (PREMIER Biosoft Int.,Palo Alto, CA, USA). The conditions for standard amplificationwere 95 ◦C for 5 min, then 30 cycles of 95 ◦C for 30 s (to amplifyGAPDH, cyclin B1 and CDK1) or 95 ◦C for 1 min (to amplify GST-P),65 ◦C for 30 s, 74 ◦C for 30 s and 74 ◦C for 10 min. Products ofthe reaction were separated on 20 g L−1 agarose gel, stained with1 µg mL−1 ethidium bromide and visualised using a UVPGDS-7900digital imaging system (UVP AutoChemi System, Cambridge, UK).All amplifications were conducted within the linear range of theassay, normalised to respective GAPDH levels using SPSS Version17.0 (SPSS Inc., Chicago, IL, USA).

Colorimetric estimation of caspase-3, caspase-8 and caspase-9activitiesCaspase-3, caspase-8 and caspase-9 activities were determinedusing kits from Biovision (Mountain View, CA, USA). Cells (1 × 106)were treated with RMDE for 6, 12 and 24 h, washed with PBS,suspended in 50 µL of cell lysis buffer and incubated on icefor 10 min. Following centrifugation at 10 000 × g for 1 min, thesupernatant (cytosolic extract) was transferred to a fresh tubeand put on ice for immediate assay. Protein concentration wasassayed using 50–200 µg of standard protein in 50 µL of cell lysisbuffer for each assay. The cytosolic extract was mixed with 50 µLof 2× reaction buffer (containing 10 mmol L−1 dithiothreitol) and5 µL of 4 mmol L−1 substrate (200 µmol L−1 final concentration),incubated at 37 ◦C for 1 h and its absorbance at 405 nm wasmeasured.

Immunoblot analysisProteins separated by SDS polyacrylamide gel electrophoresiswere electrophoretically transferred to polyvinylidene difluoridemembranes. Blots were first incubated in PBS containing 50 g L−1

non-fat dry milk for 2 h to block non-specific binding sites, then in a1 : 1000 dilution of primary antibodies at 4 ◦C overnight and finally,after washing, in a 1 : 20 000 dilution of horseradish peroxidase-conjugated secondary antibodies (GeneTex, Inc., San Antonio, TX,USA) at room temperature for 1 h. After washing, immunoreactiveproteins were visualised using enhanced chemiluminescencedetection reagents (Sigma Chemical Co.). Densitometry wasquantitated with SPSS Version 17.0 (SPSS Institute, Inc.).

Statistical analysisData were expressed as mean ± standard deviation (SD). Statisticalsignificance was determined by one-way analysis of variance(ANOVA) using the general linear model procedure of SPSS Version17.0 (SPSS Institute, Inc.), followed by ANOVA with Duncan’s test.All comparisons were made relative to controls, and significantdifferences are indicated as ∗P < 0.05, ∗∗P < 0.01 or ∗∗∗P < 0.001.

Figure 1. Representative cell viability of different oral cancer cells:(A) effects of 24 h treatment with ethanol and water extracts of RMRand RMD on viability of SCC-25 cells; (B) clonogenic survival assay forRMDE on oral cancer cells. Results are expressed as mean ± SD (n = 3).

Table 1. Growth inhibition (IC50, µg mL−1) of different oral cancercell lines by RMRE and RMDE

Cell line

Extract SCC-4 SCC-9 SCC-25 FaDu HEp-2

RMRE

24 h >250 >250 >250 >250 >250

48 h >250 237.3 188.6 >250 >250

RMDE

24 h >250 >250 78.1 244.8 162.6

48 h 142.6 151.9 36.4 117.8 155.5

RESULTSEffect of RMD/RMR ethanol and water extracts on cell viabilityand proliferationFigure 1(A) shows the dose-dependent effects of ethanol andwater extracts of RMD and RMR on the viability of oral cancerSCC-25 cells after a 24 h treatment. Water extracts of RMD (RMDW)and RMR (RMRW) exerted only weak cell toxicity, whereas RMDEand RMRE reduced cell viability significantly. Thus we investigatedthe growth-inhibitory effects of RMDE and RMRE on differentoral cancer cell lines. The results showed that IC50 of RMDE wasless than that of RMRE, with a time-dependent decrease in growthinhibition. Among the five cell lines tested, the maximum cytotoxiceffect was observed on SCC-25 cells (Table 1).

RMDE affected the long-term survival of oral cancer cellsmore severely (Fig. 1(B)) than their short-term viability (Table 1),


27

12


Table 2. Cell cycle distribution (%) in SCC-25 cells treated with RMDEfor 24 h

Phase

Treatment G0/G1 S G2/M

Untreated 74.0 ± 5.02 6.1 ± 0.15 18.9 ± 2.04

100 µg mL−1 55.1 ± 2.20 10.0 ± 0.19 34.0 ± 1.94

200 µg mL−1 50.6 ± 3.04 9.8 ± 0.12 38.3 ± 3.06

Results are expressed as mean ± SD (n = 3).

indicating that the treatment with RMDE resulted in long-term celldamage and subsequent inhibition of cell proliferation.

Effect of RMDE on cell cycle distributionTo elucidate the cytotoxic mechanism of RMDE, further experi-ments were performed using the SCC-25 cell line. To examine themechanism underlying the inhibitory effects of RMDE on SCC-25cell proliferation, the cell cycle distribution was evaluated usingflow cytometry. A 24 h treatment with RMDE caused cell cyclearrest at the G2/M phase, and this effect was dose-dependent(Table 2). Thus the cytotoxic effect of RMDE on SCC-25 cells mightbe attributed to the induction of G2/M arrest.

Effect of RMDE on GST-P, cyclin B1 and CDK1 expressionWe next examined the effect of RMDE on cell cycle-regulatorymolecules. Figure 2 shows the results of the RT-PCR analysis of

Figure 2. Changes in level of mRNA associated with cell proliferation byRT-PCR: (A) expression of GST-P, cyclin B1 and CDK1 after exposure to100 µg mL−1 RMDE for 3 and 6 h; (B) densitometric analysis of RT-PCRimages.

cell survival markers and G2/M phase arrest-associated factors ofSCC-25 cells. In comparison with the control group, the expression

Figure 3. Effects of RMDE-induced apoptosis on SCC-25 cells after 24 h (cells in lower right quadrant represent early apoptosis; cells in upper rightquadrant represent late apoptosis): (A) untreated group; (B) 100 µg mL−1 RMDE treatment; (C) 200 µg mL−1 RMDE treatment; (D) analysis of cells in earlyand late stage apoptosis.


27

13


of GST-P, cyclin B1 and CDK1 was decreased in cells treated with100 µg mL−1 RMDE for 3 and 6 h. Thus the inhibitory effect of RMDEon SCC-25 cell proliferation and cell cycle progression might beattributable to the down-regulation of GST-P, cyclin B1 and CDK1.

Effect of RMDE on SCC-25 cell apoptotic inductionTo further elucidate the mechanism of action of RMDE on SCC-25cells, we examined the effect of RMDE on apoptosis by performingAnnexin V-FITC/PI double staining. This staining method alongwith flow cytometry enables the quantitative assessment of living(Annexin V-FITC negative/PI negative), early apoptotic (AnnexinV-FITC positive/PI negative), late apoptotic/necrotic (AnnexinV-FITC positive/PI positive) and dead (Annexin V-FITC negative/PIpositive) cells. The effects of a 24 h RMDE treatment on SCC-25cell apoptosis are shown in Fig. 3. The cells in the lower rightquadrant represent early apoptosis, while those in the upper rightquadrant represent late apoptosis. The results indicated that, aftertreatment with 100 µg mL−1 RMDE, most cells were in the earlyapoptotic stage (69.88%) and a few were in the late apoptotic stage(7.87%), i.e. 77.75% of cells underwent apoptosis. On the otherhand, 95.93% of cells underwent apoptosis in the group treatedwith 200 µg mL−1 RMDE, 52.3% being in the early apoptotic stageand 43.63% in the late apoptotic stage.

Effect of RMDE on caspase-3, caspase-8 and caspase-9activitiesApoptosis has been found to be regulated by two main pathways,the mitochondrial pathway and the death receptor pathway,both of which activate caspase-3. Caspase-8 and caspase-9are activators/initiators of the death receptor pathway and themitochondrial pathway respectively.16 We examined the effect ofRMDE on caspase activities in order to understand the mechanismof RMDE cell proliferation inhibition.

Figure 4 indicates that RMDE promoted the expression ofcaspase-3, caspase-8 and caspase-9 in a dose- and time-dependentmanner. These results suggested that the RMDE-induced apoptosiswas mediated by the induction of both the death receptor pathwayand the mitochondrial pathway.

Effect of RMDE on NF-κB, I-κB and Bax expressionThe NF-κB transcription factor is reported to be overexpressedin highly proliferating tumours.17 Besides aiding tumour cellproliferation, NF-κB has also been reported to enhance tumourdevelopment by inhibiting apoptosis.18 NF-κB is sequesteredin the cytoplasm by inhibitory proteins such as I-κB, whichmask the nuclear localisation signal of NF-κB. An increasein Bax expression can block mitochondria-mediated apoptosisby preventing cytochrome c release from the mitochondria,thereby inhibiting the activation of caspase-3. Figure 5 showsa representative Western blot analysis of SCC-25 cells treatedwith RMDE. The expression of NF-κB, I-κB and Bax was detectedas bands of molecular weight 65, 43 and 21 kDa respectively. Incomparison with the control group, RMDE treatment decreasedthe expression of NF-κB and increased the expression of I-κBand Bax.

DISCUSSIONDioscorea is proven to have antioxidative as well as antitumourproperties.13,14,19 The antiproliferative activity of polyphenols suchas delphinidin, cyanidin, peonidin, petunidin and malvidin has also

Figure 4. Effects of RMDE on (A) caspase-3, (B) caspase-8 and (C) caspase-9activities: �, 50 µg mL−1 RMDE; �, 100 µg mL−1 RMDE; , 200 µg mL−1

RMDE. Results are expressed as mean ± SD (n = 3). ∗Significantly differentfrom control group at P < 0.05. ∗∗Significantly different from control groupat P < 0.01.

been reported.20 Additionally, a polyphenol in black tea has beenshown to effectively inhibit hamster buccal pouch carcinogenesisand induce apoptosis,21,22 and its antioxidative property preventsoral carcinogenesis.23 In a previous study we found that ethanol ex-tracts of Monascus-fermented products had antioxidative proper-ties, including reducing power and 1,1-diphenyl-2-pichrylhydrazylradical-scavenging activity.24 Furthermore, Monascus-fermentedproducts contain various antioxidants such as dimerumic acid,tannins and phenols.14 Thus Monascus-fermented products ex-hibit tumour-inhibitory properties, and Monascus fermentation ofdioscorea may result in stronger anticancer effects. The levels oftotal phenols and flavonoids were 119 and 191 mg kg−1 dioscorearespectively and 179 and 249 mg kg−1 RMD respectively (data notshown). Thus the total phenol and flavonoid levels increased as aresult of Monascus fermentation, and this may confer RMD withanti-oral cancer properties. On the other hand, Monascus speciesproduce yellow pigments such as monascin and ankaflavin, whichhave been reported to have cancer cell-cytotoxic activities.9,10 Asa result, apart from analysing phenol and flavonoid contents, we


27

14


Figure 5. (A) Western blot analysis of protein extracts obtained from SCC-25 cells treated with different concentrations of RMDE for different times.Densitometric analysis of (B) NF-κB, (C) I-κB and (D) Bax.

also used high-performance liquid chromatography to identify thebioactive chemical constituents in RMDE. Monascin and ankaflavinwere present at concentrations of 137.2 and 34.14 g kg−1 respec-tively in Monascus-fermented dioscorea,25 higher than the levelsfound in RMRE.15 Dioscorea has proven antitumour ability,13 andRMDE has stronger anticancer activity than RMRE, RMDW andRMRW used in this study. We suggest that the greater anticanceractivity of RMDE compared with RMRE might be attributableto RMDE containing higher levels of anticancer substances suchmonascin and ankaflavin.

Monascus-fermented products have been reported to inhibittumour progression and tumour metastasis-associated factors andto reduce angiogenesis.8 RMRE is known to exert inhibitory effectsagainst oral carcinogenesis.15 However, in the present study, RMDEexerted stronger cell cytotoxicity than RMRE (Table 1) and was amore potent cell proliferation inhibitor than RMDW and RMRW(Fig. 1(A)). These findings suggested that RMDE had the potentialto repress oral cancer cell growth. Therefore we examined theeffect of RMDE on the induction of apoptosis.

Phytochemicals have been shown to induce cell cycle arrest,cause apoptosis and affect the differentiation and proliferation ofcells mediated by the effect of intracellular reactive oxygen specieson the signal transduction pathway.26 Further, two apoptosissignalling pathways converge during the activation of caspasesin the inhibition of cell proliferation. Capsase-8 is activated bythe death pathway, which is initiated by the death receptorthat contains the intracellular death domain and is propagatedthrough the Fas-associated protein with death domain (FADD)adaptor protein. The ligand bound to the death receptor forms a

signal complex, FADD, which increases caspase-8 and caspase-3activities to induce apoptosis.16,27 The mitochondrial pathway isregulated by mitochondria-released cytochrome c, which leads tocaspase-9 activation via apoptotic protease-activating factor-1.

Treatment of cells with RMDE for 24 h resulted in cell cycle arrestat the G2/M phase (Table 2). This effect was also associated withthe repression of CDK1 and cyclin B1 mRNA levels, resulting incell proliferation inhibition (Fig. 2). The overexpression of GST-Pand NF-κB enhances cell proliferation and prevents apoptosisreaction. RMDE treatment decreased the mRNA levels of GST-P(Fig. 2), indicating that RMDE inhibited cell proliferation, possiblyvia lowering CDK1 and cyclin B1 expression and then arresting cellsat the G2/M phase. In addition, RMDE effectively induced SCC-25cell apoptosis in a dose-dependent manner (Figs 3 and 5(D)). Ourresults suggested that the 6, 12 and 24 h RMDE (200 µg mL−1)treatments induced the activity of caspase-3, caspase-8 andcaspase-9 in SCC-25 cells (Fig. 4). The activity of caspase-8 wassignificantly increased compared with that of caspase-9, indicatingthat RMDE induced SCC-25 cell apoptosis mainly via activatingcaspase-8. RMDE treatment also resulted in a decrease in NF-κBlevels and an increase in I-κB (Fig. 5), thus indicating that RMDEinhibited cell proliferation by repressing NF-κB transcription.

In conclusion, RMDE treatment resulted in the followingsignificant changes in SCC-25 cells: (a) selective inhibition of SCC-25 cells, (b) dose-dependent cell cycle arrest at the G2/M phaseand (c) time- and dose-dependent induction of apoptosis (Fig. 6).Therefore RMDE has potential to be used as a functional food inadjuvant chemotherapy for treating human oral cancer.


27

15


Figure 6. Proposed signal pathway of RMDE-induced G2/M arrest andapoptosis in human oral cancer SCC-25 cells.

ACKNOWLEDGEMENTThis research work and subsidiary spending were supported byPaolyta Co., Ltd (Taipei, Taiwan).

Supporting informationSupporting information may be found in the online version of thisarticle.

REFERENCES1 Parkin DM, Bray F, Ferlay J and Pisani P, Estimating the world cancer

burden: Globocan 2000. Int J Cancer 94:153–156 (2001).2 Blot WJ, McLaughlin JK, Winn DM, Austin DF, Greenberg RS, Preston-

Martin S, et al., Smoking and drinking in relation to oral andpharyngeal cancer. Cancer Res 48:3282–3287 (1988).

3 McLaughlin JK, Gridley G, Block G, Winn DM, Preston-Martin S,Schoenberg JB, et al., Dietary factors in oral and pharyngeal cancer.J Natl Cancer Inst 80:1237–1243 (1988).

4 Pintos J, Franco EL, Kowalski LP, Oliveira BV and Curado MP, Use ofwood stoves and risk of cancers of the upper aero-degestive tract:a case-control study. Int J Epidemiol 27:936–940 (1998).

5 Lee BH, Ho BY, Wang CT and Pan TM, Red mold rice promotedantioxidase activity against oxidative injury and improvedthe memory ability of zinc-deficient rats. J Agric Food Chem57:10600–10607 (2009).

6 Lee CL, Kuo TF, Wang JJ and Pan TM, Red mold riceameliorates impairment of memory and learning ability inintracerebroventricular amyloid beta-infused rat by repressingamyloid beta accumulation. J Neurosci Res 85:3171–3182 (2007).

7 Endo A and Monacolin K, a new hypocholesterolemic agent producedby a Monascus species. J Antibiot 32:852–854 (1979).

8 Ho BY and Pan TM, The Monascus metabolite monacolin K reducestumor progression and metastasis of Lewis lung carcinoma cells.J Agric Food Chem 57:8258–8265 (2009).

9 Akihisa T, Tokuda H, Ukiya M, Kiyota A, Yasukawa K, Sakamoto N, et al.,Anti-tumor-initiating effects of monascin, an azaphilonoid pigmentfrom the extract of Monascus pilosus fermented rice (red-mold rice).Chem Biodiversity 2:1305–1309 (2005).

10 Su NW, Lin YL, Lee MH and Ho CY, Ankaflavin from Monascus-fermented red rice exhibits selective cytotoxic effect and inducescell death on Hep G2 cells. J Agric Food Chem 53:1949–1954 (2005).

11 Chang WC, Yu YM, Wu CH, Tseng YH and Wu KY, Reduction ofoxidative stress and atherosclerosis in hyperlipidemic rabbits byDioscorea rhizome. Can J Physiol Pharmacol 83:423–430 (2005).

12 Wang G, Chen H, Huang M, Wang N, Zhang J, Zhang Y, et al., Methylprotodioscin induces G2/M cell cycle arrest and apoptosis in HepG2liver cancer cells. Cancer Lett 241:102–109 (2006).

13 Park MK, Kwon HY, Ahn WS, Bae S, Rhyu MR and Lee Y, Estrogenactivities and the cellular effects of natural progesterone fromwild yam extract in mcf-7 human breast cancer cells. Am J Chin Med37:159–167 (2009).

14 Lee CL, Wang JJ, Kuo SL and Pan TM, Monascus fermentation ofdioscorea for increasing the production of cholesterol-loweringagent-monacolin K and anti-inflammation agent-monascin. ApplMicrobiol Biotechnol 72:1254–1262 (2006).

15 Tsai RL, Ho BY and Pan TM, Red mold rice mitigates oralcarcinogenesis in 7,12-dimethyl-1,2-benz[a]anthracene-inducedoral carcinogenesis in hamster. Evid Based Compl Altern Med DOI:10.1093/ecam/nep215 (2009).

16 Nagata S, Apoptosis by death factor. Cell 88:355–365 (1997).17 Dolcet X, Llobet D, Pallares J and Matias-Guiu X, NF-κB in development

and progression of human cancer. Virchows Arch 446:475–482(2005).

18 Tamatani M, Che YH, Matsuzaki H, Ogawa S, Okado H, Miyake S, et al.,Tumor necrosis factor induces Bcl-2 and Bcl-x expression throughNFkappaB activation in primary hippocampal neurons. J Biol Chem274:8531–8538 (1999).

19 Lin YM and Lin KW, Antioxidative ability, dioscorin stability, and thequality of yam chips from various yam species as affected byprocessing method. J Food Sci 74:118–125 (2009).

20 Jacob JK, Hakimuddin F, Paliyath G and Fisher H, Antioxidant andantiproliferative activity of polyphenols in novel high-polyphenolgrape lines. Food Res Int 41:419–428 (2008).

21 Chandra-Mohan KVP, Devaraj H, Prathiba D, Hara Y and Nagini S,Antiproliferative and apoptosis inducing effect of lactoferrin andblack tea polyphenol combination on hamster buccal pouchcarcinogenesis. Biochim Biophys Acta 1760:1536–1544 (2006).

22 Chandra-Mohan KVP, Vidjaya-Letchoumy P, Hara Y and Nagini S,Combination chemoprevention of hamster buccal pouchcarcinogenesis by bovine milk lactoferrin and black tea polyphenols.Cancer Invest 26:193–201 (2008).

23 Vidjaya-Letchoumy P, Chandra-Mohan KVP, Stegeman JJ, Gelboin HV,Hara Y and Nagini S, In vitro antioxidative potential of lactoferrin andblack tea polyphenols and protective effects in vivo on carcinogenactivation, DNA damage, proliferation, invasion, and angiogenesisduring experimental oral carcinogenesis. Oncol Res 17:193–203(2008).

24 Lee CL, Hung HK, Wang JJ and Pan TM, Red mold dioscorea has greaterhypolipidemic and antiatherosclerotic effect than traditional redmold rice and unfermented dioscorea in hamsters. J Agric FoodChem 55:7162–7169 (2007).

25 Hsu WH, Lee BH and Pan TM, Protection of Monascus-fermenteddioscorea against DMBA-induced oral injury in hamster by anti-inflammatory and antioxidative potentials. J Agric Food Chem58:6715–6720 (2010).

26 Hu R and Kong AHT, Activation of MAP kinases, apoptosis andnutrigenomics of gene expression elicited by dietary cancer-prevention compounds. Nutrition 20:83–88 (2004).

27 Pietenpol JA and Stewart ZA, Cell cycle checkpoint signaling: cell cyclearrest versus apoptosis. Toxicology 181/182:475–481 (2002).


27

16

Research ArticleReceived: 6 May 2010 Revised: 3 August 2010 Accepted: 5 August 2010 Published online in Wiley Online Library: 2 September 2010


Effects of postharvest treatments on fruitquality of sweet pepper at low temperaturePaula Cuadra-Crespo and Francisco M del Amor∗

Abstract

BACKGROUND: Postharvest storage of sweet pepper fruits (Capsicum annuum L.) at low temperatures could impair their physicaland chemical composition. Therefore, maintenance of essential nutrition support or altered gas exchange could preserve fruitquality, minimizing chilling injury. Thus our aim was to determine the response to postharvest application of a low concentrationof nitrogen (urea) or antitranspirant (pinolene) during a period of 21 days at 5 ◦C.

RESULTS: The results indicate that storage at 5 ◦C was effective with respect to maintaining firmness of sweet pepper fruitsfor 21 days, while application of antitranspirant increased firmness compared with non-sprayed fruits. Additionally, ureamaintained color while increasing total phenolics and the activity of catalase and ascorbate peroxidase, lowering lipidperoxidation. Composition of free amino acids was affected to a minor extent.

CONCLUSION: Maintaining quality is of paramount importance in the postharvest period. This study shows the effect of bothtemperature and spraying treatments with regard to maintaining fruit quality during this period, and provides new insightsinto the physiological role of enzymes of the antioxidant system during pepper storage at low temperature.c© 2010 Society of Chemical Industry

Keywords: Capsicum annuum L.; urea; antitranspirant; color; antioxidant enzymes; chilling injury

INTRODUCTIONWith the increasing demand for fresh fruit and vegetables,postharvest technology for extending shelf-life of these perishablecommodities has gained significant importance in recent years.1

The principal physiological factors that negatively impact pepperfruit during shipment and storage and subsequent marketing arewater loss and chilling injury.2 The skin of fruit and vegetablesplays an important role in gas exchange between the product andthe surrounding environment,3 and for this reason the protectionof the pericarp against dehydration is particularly important afterharvest, when fruits do not receive water or nutrients from theplant. Therefore, implementation of techniques to preserve thephysicochemical properties of the pericarp could help to preservefruit quality during the storage period.

Storage temperature has a great influence on the physiologicalresponse, since temperature and humidity are the environmentalfactors that have the strongest influence on fruit quality.3 Manystudies have pointed out an influence of storage temperatureon water loss rate, texture and overall quality of fresh pepperfruits. Thus temperature regulation is the most effective tool forextending the storage life of fresh commodities, including pepper.4

However, storage at low temperatures could produce unusualripening, water loss, an increase of CO2, and higher permeabilityin cellular membranes, inducing ion leakage; the damage isprogressively more severe during long-term storage at lowtemperatures.5 Gonzalez-Aguilar et al.6 indicated that symptomsare accompanied by biochemical and physiological changes,which are generated by the direct effect of low temperatureon cellular constituents.

If free radicals are not neutralized in humans and plants, theydamage cells and organs, causing many degenerative diseasesand susceptibility to biotic and abiotic stresses.7 Reactive oxygenspecies (ROS) are strongly associated with lipid peroxidationand consequent deterioration of food materials, but have alsobeen involved in the development of several diseases.8 Lowtemperatures can induce free radical production and ROSmay contribute to the loss of cellular functions through lipidperoxidation.9 Oxidation and peroxidation of membrane lipidsand proteins could be caused by ROS.10 Therefore, antioxidantenzymes are the most important components in the scavengingsystem of ROS11 and peppers contain many active substances thatare important for protection against oxidative damage by freeradicals.12

Plants are able to use several forms of N, nitrate and ammoniumbeing the most important ones. Urea is one of the most widely usedfoliar-N fertilizers, characterized by high leaf penetration rate andlow cost. Additionally, urea has been considered the most suitableform of foliar N because of its rapid absorption, low phytotoxicityand high solubility in both oil and water.13 The nutrient with thesingle greatest effect on fruit quality is N.14 Moreover, pepper fruityield and quality are affected by different agricultural practices and

∗ Correspondence to: Francisco M del Amor, Departamento de Citricultura yCalidad Alimentaria, Instituto Murciano de Investigacion y Desarrollo Agrarioy Alimentario (IMIDA), C/Mayor s/n, 30150 Murcia, Spain.E-mail: [email protected]

Departamento de Citricultura y Calidad Alimentaria, Instituto Murciano deInvestigacion y Desarrollo Agrario y Alimentario (IMIDA), C/Mayor s/n, 30150Murcia, Spain


27

17

Postharvest quality of sweet pepper www.soci.org

by the availability of N to the plant.15,16 No N source is providedto the fruit once detached from the plant, but many physiologicalprocess could still require N to avoid metabolic unbalance inthe pericarp. Therefore, rapid absorption of N applied directlyto the fruit could help to maintain its characteristics. Thus, ureasprays enhanced the green pigmentation in apple after a storageperiod,17 increased chlorophyll content in broccoli heads18 andhad a beneficial effect in pepper leaves, maintaining membranepermeability.19

Antitranspirants are compounds applied to plant leaves toreduce transpiration, since they reduce the stomatal opening andincrease the leaf resistance to water vapor diffusion, and theyhave often been applied in an attempt to prevent water stress.20

Reducing rates of transpiration directly by antitranspirant spraysavoided the need for drastically altering environmental conditionsin experiments designed to evaluate the effects of transpiration.21

Antitranspirants have been used successfully in agriculture tocontrol leaf transpiration and improve quality of sweet pepper,22

tomatoes,23 onions24 and potatoes.25

The aim of this study was to determine the effect of foliarurea and antitranspirant applications on pepper fruits, regardingboth physical and chemical postharvest characters (oxidativemetabolism and amino acids). Therefore, we studied severalquality parameters such as fruit firmness, color, lipid peroxidation,antioxidants such as catalase (CAT), ascorbate peroxidase (APOX)and phenolic compounds, and the free amino acids profile. Ourresult could help to define an efficient postharvest strategy (froma physiological and agronomical perspective) to maintain fruitquality of sweet pepper fruits under low temperature.

MATERIALS AND METHODSPlant material and storage conditionsUniform sweet pepper plants (Capsicum annuum L.) cv. Herminiowere obtained from a commercial nursery. Plants were grownin 1.2 m long coconut fiber-filled bags in a greenhouse equippedwith a computer-regulated drip irrigation system, under controlledenvironmental conditions. Each bag had three plants with three4 L h−1 drippers. Irrigation management was according to localcommercial soilless cultivation and the drainage percentagewas maintained at 30%.26 The pH of the nutrient solution wasmaintained between 5.6 and 6.0. Plants were irrigated with nutrientsolution of the following composition (meq L−1): NO3

−, 12.5;H2PO4

−, 1.5; SO42−, 7.5; K+, 7.5; Ca2+, 9.5; Mg2+, 4.5. One hundred

and fifty days after transplanting (DAT), fruits were harvested,weighed and stored at 5 ◦C and 90–95% relative humidity in adark chamber for 21 days. Fruit harvesting was performed at thegreen stage of ripening. Treatments consisted of the applicationof antitranspirant (AT, pinolene 5%, as commercial preparation‘Vapor Gard’ = 96% pinolene in 4% inert ingredients), foliar urea(UR, 15 g L−1) and non-sprayed fruits (control (CN)). At 0 (beforestorage), 7, 14 and 21 days after storage, 15 fruits per treatmentand per day of storage were processed by measuring color,firmness, total phenolic compounds, lipid peroxidation (TBARS),amino acids, and the antioxidant enzymes catalase (CAT) andascorbate peroxidase (APOX). Each fruit was considered a sample.

Skin color and firmnessFirmness was determined from the surface in the equatorial area,using a Bertuzzi FT011 penetrometer (Fruit tester, Alfonsine, Italy)fitted with an 8 mm diameter probe. Color was determined with

a Konica-Minolta CR-300 colorimeter (Minolta, Osaka, Japan), lightsource D65, making three measurements for each pepper. Afterfirmness and color determinations, destructive measurementswere carried out for measuring the enzymatic response; thereforethese parameters were measured in different fruits of the sametreatment at each harvest time. Color data are provided asCIELAB (L∗ a∗ b∗) coordinates, which define the color in athree-dimensional space: L∗ indicates lightness, and a∗ and b∗are the chromaticity coordinates green–red and blue–yellow,respectively. L∗ is an approximate measurement of luminosity,which is the property according to which each color can beconsidered as equivalent to a member of the gray scale, betweenblack and white, taking values within the range 0–100; a∗ takespositive values for reddish colors and negative values for greenishones, whereas b∗ takes positive values for yellowish colors andnegative values for bluish ones.16

Total phenolic compoundsTotal phenolic compounds were extracted from 0.4 g of frozenpepper fruits (−80 ◦C) with 4 mL methanol and 0.1 mol L−1 HCl.The homogenate was centrifuged at 15 000 × g for 20 min at4 ◦C. For the determination, Folin–Ciocalteu reagent was useddiluted with distilled water (1 : 10). The diluted reagent (2 mL) wasmixed with 400 µL supernatant and 1600 µL sodium carbonate(7.5%) was added. The mixture was kept for 30 min in the darkand then centrifuged at 5000 × g for 5 min. The supernatant wasseparated and absorbance was measured at 765 nm according tomethodology of Kahkonen et al.27 The total phenolic content wasexpressed as gallic acid equivalents in mg mL−1 fresh weight.

Antioxidant enzymesAntioxidant enzymes were determined according to Del Amoret al.28 Briefly, extracts for the determination of catalase (CAT)and ascorbate peroxidase (APOX) activities were homogenizedand centrifuged and the supernatant was used for the assays.CAT activity was determined by measuring the decrease inabsorption at 240 nm. The H2O2-dependent oxidation of ascorbatewas followed as a decrease in absorbance at 290 nm (e,2.8 mmol L−1 cm−1).

Lipid peroxidationLipid peroxidation was extracted using the method described byour group in 2009.28 Briefly, fresh fruit were homogenized andcentrifuged. The supernatant was separated and a mixture oftrichloroacetic acid (TCA), thiobarbituric acid (TBA) and butylatedhydroxytoluene (BHT) was added. The mixture was heated andthen quickly cooled on ice. The contents were centrifuged andthe absorbance was measured at 532 nm. The value for non-specific absorption at 600 nm was subtracted. The concentrationof TBARS was calculated using an extinction coefficient of155 mmol L−1 cm−1.29

Free amino acidsFree amino acids were extracted from fruits frozen at −80 ◦C:sap was extracted after vortexing at 5000 rpm (10 min, 4 ◦C)and determined following the AccQ·Tag-ultra ultra-performanceliquid chromatography (UPLC) method.30 For derivatization, 70 µLborate buffer was added to the hydrolyzed sample or to 10 µLof the fruit sap. Next, 20 µL reagent solution was added. Thereaction mixture was mixed immediately and heated at 55 ◦C


27

18

www.soci.org P Cuadra-Crespo, FM del Amor

for 10 min. After cooling, an aliquot of the reaction mixture wasused for UPLC injection. UPLC was performed on an Acquitysystem (Waters, Milford, MA, USA), equipped with a fluorescencedetection (FLR) system. The column used was a BEH C18 100 mm× 2.1 mm, 1.7 µm (Waters). The flow rate was 0.7 mL min−1 andthe column temperature was kept at 55 ◦C. The injection volumewas 1 µL. Wavelength excitation (λex) and emission (λem) wereset at 266 and 473 nm, respectively. The solvent system consistedof two eluents: (A) AccQ·Tag-ultra eluent A concentrate (5%, v/v)and water (95%, v/v); (B) AccQ·Tag ultra eluent B. The followinggradient elution was used: 0–0.54 min, 99.9% A–0.1% B; 5.74 min,90.9% A–9.1% B; 7.74 min, 78.8% A–21.2% B; 8.04 min, 40.4%A–59.6% B; 8.05–8.64 min. 10% A–90% B; 8.73–10 min, 99.9%A–0.1% B. Empower 2 (Waters) software was used for systemcontrol and data acquisition. External standards (Thermo scientific)were used for quantification of (NH3) ammonia; (ala) alanine; (arg)arginine; (asp) aspartic acid; (cys) cysteine; (glu) glutamic acid;(gly) glycine; (his) histidine; (ile) isoleucine; (leu) leucine; (lys)lysine; (met) methionine; (phe) phenylalanine; (pro) proline; (ser)serine; (thr) threonine; (tyr) tyrosine; (val) valine.

Statistical analysisData were tested first for homogeneity of variance and normalityof distribution, and were analyzed by analysis of variance usingthe Duncan multiple range test to determine differences betweenmeans (P ≤ 0.05) for treatments at different days of storage.The statistical analyses were done using SPSS 12.0 (SPSS Science,Chicago, IL, USA).

RESULTS AND DISCUSSIONFirmness and colorFruit firmness measurement is a good way to monitor fruitsoftening and to predict bruising damage during harvest andpostharvest handling;31 thus accelerated loss of texture isconsidered one of the main factors that limit the shelf-life of freshtissue.32 In our study, storage at 5 ◦C was effective for maintenanceof fruit firmness of control fruits (non-sprayed) during 21 days(Fig. 1). Already after 7 days, a significant increase in firmness wasobserved for those fruits sprayed with urea or AT compared withcontrol fruits. After 14 days, firmness was reduced in those fruitssprayed with urea to values close to the control fruits, but ATincreased fruit firmness by 29.1% compared with the non-sprayedfruits after 14 days of storage, and by 29.5% after 21 days.

Lowering the temperature of non-climacteric fruits such assweet pepper lowers their rate of ripening and deterioration.33

However, fresh peppers are highly sensitive to freezing injury andsusceptible to chilling injury, and the characteristic symptomsof chilling injury in sweet pepper are softening, pitting and apredisposition to decay.34 Our study with cv. Herminio showed novariations in firmness at 5 ◦C for control fruits, while a transientincrease in firmness was observed for urea. However, a morestable and long-lasting difference compared with the controlwas maintained when AT was applied. Softening of fruits duringripening is characterized by the solubilization of pectins.35 Thusthe observed effect might be due to an increase of the pectinviscosity of cell walls, implicating a rapid and short-lasting effectof urea due to a quick absorption, and a longer-lasting effect forAT, due to an effective covering of the skin of the fruit until 21 dayswith no absorption or alteration.

Color change in the pepper surface takes place as a result ofchlorophyll degradation and a considerable increase in carotenoid

Figure 1. Effects of the application of antitranspirant (AT) or foliar urea(UR) on fruit firmness. CN indicates control non-sprayed fruit. Vertical barsindicate standard errors of means, only shown when larger than symbolsize. Values with the same letter are not significantly different at P < 0.05(Duncan’s multiple range test).

content.16 Meir36 reported that pepper fruit harvested at themature-green stage is sensitive to temperatures below 6 ◦C anddeveloped chilling injury. Color parameter L∗ represents lightness,ranging between 0 (black) to 100 (white), and color change isobserved as a decrease in L∗.37 Thus already after 7 days of storageat 5 ◦C, both sprayed and non-sprayed fruits become darker butafter 21 days no significant differences were observed betweenthe treatments (Fig. 2(A)). The parameter a∗ showed an inversepattern compared with the L∗, but after 21 days fruits treated withurea had significantly reduced values of this parameter comparedwith the control fruits (Fig. 2(B)). This increase in a∗ in controlor antitranspirant treatments will result in reddish fruits andtherefore an increase in ripening (these pepper fruits turn fromgreen to red). Thus urea could delay senescence as it was able tomaintain lower a∗ values after 21 days. Additionally, the higher a∗values of control and antitranspirant fruits compared with ureacould indicate a significant degradation of chlorophyll pigments inthese treatments. In a previous study38 we demonstrated that thechlorophyll pigment concentration was related to the supply of Nto the roots, and urea could provide a source of N for maintainingthese pigments. Similar results were found by del Amor et al.:28

applications of urea during pepper cultivation increased L∗ and a∗

and reduced b∗, compared with the limited N supply treatment.Some differences in the responses to urea between that study,under greenhouse conditions, and this one (postharvest) could bealso due to differences in the intensity of the N stress, as our fruitswere harvested from well-nourished plants and the additionaleffect in L∗ could be minimized.

Total phenolic compoundsPhenolic compounds are secondary metabolites in plants. Theirfunctions are not always known, but some are structural polymers,UV screens, antioxidants or attractants, and others are involved innon-specific defense mechanisms.39 The phenolic composition offruits and hence their antioxidant properties may be modifiedby environmental and postharvest factors, including storageand processing.40 Also, phenolics are of great importance indetermining some quality attributes and properties in fresh fruitsand vegetables, like color, texture, taste and flavour. Additionally,


27

19


Figure 2. Effects of the application of antitranspirant (AT) or foliar urea(UR) on fruit color. CN indicates control non-sprayed fruit. Vertical barsindicate standard errors of means, only shown when larger than symbolsize. Values with the same letter are not significantly different at P < 0.05(Duncan’s multiple range test).

phenolic compounds may act as antioxidants in different plantreactions. Thus numerous studies have demonstrated thatthe accumulation of phenolic compounds such as flavonoidsand anthocyanins provides a defensive mechanism, and theconcentration of flavonoids in plants could be affected by NO3

−

supply.28,41 In our study, total phenolics of control fruits were notaffected at 21 days of storage at 5 ◦C (Fig. 3(A)), while a significantincrease was observed when urea was applied. Some researchersfound that NO3

− application was favorable to phenolic compoundaccumulation.42 Del Amor et al.28 showed the effect of foliarN fertilization on some phenolic compounds like anthocyanins,which increased with the application of urea. In the case of ATtreatment, phenolic compounds did not change and they followed

a pattern similar to that of the control fruits. Shin et al.43 observedthat phenolic compounds increased during storage but in ourstudy this effect was only as a result of urea application. Paddaand Picha44 showed that phenolic contents are influenced bycultivar and other pre- and/or postharvest conditions and long-term exposure to low temperature. Cordenunsi et al.45 found thattotal phenolics contents remained constant or even decreased.In contrast, Robards et al.40 found that total phenol contentincreased significantly in apples stored at 0 ◦C, while Ayala-Zavala et al.46 found that total phenolics increased during coldstorage of strawberry fruits. In our control fruits, the total phenoliccompounds showed a slight tendency to increase after 7 days ofstorage.

Antioxidant enzymesPlants have an antioxidant defense system that can preventthe accumulation of ROS and repair oxidative damage. Chillinginjury in plants results in elevated levels of ROS and antioxidantenzymes protect against these potentially dangerous molecules.47

Hydrogen peroxide (H2O2) is a potentially toxic compound whichis reduced to water by CAT and APOX.48 Thus catalase protectscells against the H2O2 that is generated in these cells, catalyzingits conversion to H2O and molecular O2, and destroying toxicsubstances, which could enter the cells.49 However the responseof CAT and APOX was not exactly the same when urea was appliedto growing, N-deficient fruits28 or to well-nourished fruits understorage at low temperatures. In this study, catalase was affectedalready after 7 days by urea or AT, while it was reduced in controlfruits to a minor extent (Fig. 3(B)). Thus the catalase activity inAT fruits first increased transiently during the first 7 days andthen decreased permanently over the next period. However, theeffect of urea was the inverse of that observed for AT (whileAT decreased, urea increased it) but at 21 days both treatmentsshowed the same activity, significantly higher than for controlfruits. A similar pattern was observed for APOX activity (Fig. 3(C)).However, differences between AT and urea were still maintainedat 21 days although with the opposite tendency. Imahori et al.50

found that fruit stored at 6 ◦C showed a gradual decline in CATand APOX activity at 15 days of storage and reported that changesin CAT activity during cold storage are also related to both thechilling resistance and the development of oxidative stress. ThusBaker51 found that CAT activity in pepper leaves declined beforevisible symptoms of senescence were observed, while Zilkahet al.52 also reported a beneficial effect of foliar urea: increasedfreezing tolerance in avocado and peach. Consequently, our resultsindicate that the increased enzyme activity triggered by urea couldconfer increased protection during storage at low temperaturesas, compared with the control, lower activity of defensive enzymesat chilling temperatures can impair the plant’s ability to breakproducts of oxygen photoreduction.51 Additionally, Lim et al.53

found that chilling increases the level of active oxygen species(AOS) in chilling-sensitive plants and reduction of the chilling inresistant plants may be related to their ability to reduce and/orscavenge free radicals through increased enzyme activity. Thuslower activities of CAT were observed in chilling-sensitive fruitthan in tolerant fruit after storage at 0 ◦C.54 However, as previouslypointed out, CAT and APOX tended to decrease with AT butincrease for urea after 7 days. A different pattern during the first7 days of storage compared with the rest of the storage periodwas observed.


27

20


Figure 3. Effects of the application of antitranspirant (AT) or foliar urea (UR) on total phenolics compounds (A) catalase (B), ascorbate peroxidase (C) andlipid peroxidation (D). CN indicates control non-sprayed fruit. Vertical bars indicate standard errors of means, only shown when larger than symbol size.Values with the same letter are not significantly different at P < 0.05 (Duncan’s multiple range test).

Lipid peroxidationEnvironmental stress induces active oxygen species, which couldlead to oxidation of membrane lipids and disrupted membranes.55

Lipid peroxidation contributes to the development of chillinginjury56 and chilling has been found to induce lipid degradationin cucumber fruit and tomato pericarp.57 The thiobarbituric acid(TBA) assay is widely used to measure thiobarbituric acid-reactivesubstances (TBARS) resulting from lipid oxidation, the TBARSreaction being an indicator of lipid peroxidation.58 Changes inlipid peroxidation levels in a tissue can be a good indicator ofthe structural integrity of the membranes of plants subjected tolow temperature.59 Our data show that lipid peroxidation wasreduced in all treatments during the first 7 days of storage atlow temperature. Afterwards, control fruits showed a tendencyto increase TBARS, while a significant increase was observedin fruits treated with AT after 14 days of storage (Fig. 3(D)).Fruits treated with urea showed significantly lower values ofTBARS than control or AT fruits. Oxidative stress may be definedas an increment of oxidant species and/or a depletion ofantioxidant defenses. Thus this differential effect of AT couldbe attributed to a sharply decreasing activity of CAT andAPOX observed after 7 days of storage. Thus it is envisagedthat both activities in AT fruits will continue decreasing belowthe control levels. However, the dramatic increase in TBARSwas produced before a more evident decrease of antioxidantenzymes in the AT treatment was observed. This could indicatea delayed induction of the defense system that counteractsthe increase of TBARS at low temperature in this treatment.Wismer et al.60 reported that low temperatures can modify thebiophysical properties of membranes via the composition ofmembrane lipids, which contributes to the visible symptomsof damage and induces oxidative stress in the cell. Therefore,maintenance of the membrane integrity at low temperature hasbeen considered important in the resistance to low temperature.Thus the observed increase in TBARS could be associated withchanges in fatty acid unsaturation and in the phospholipidcomposition of mitochondrial membranes that, in turn, causechanges in membrane fluidity and the activity of respiratory

complexes.61 However, it is at least as likely that the TBARSincrease resulted from peroxidation of plastid galactolipids,which are rich in linoleic and linolenics acids (source ofmalondialdehyde).

Amino acid compositionAmino acid metabolism is one of the most important biochemicalprocesses in plants. Free amino acids are involved in secondaryplant metabolism and the biosynthesis of compounds, such asglucosinolates and phenolics, which directly or indirectly playan important role in plant–environment interaction and humanhealth; thus free amino acid profile determination is important.62

Additionally, amino acids are important for human nutrition andaffect food quality, including taste, aroma and color.63 In thisstudy arginine was the main free amino acid in sweet pepper andaccounted for half of the total amino acid content, while valueswere moderate for Asp, Thr, Ala, Lys and Try and very low for Pro,Ile, Leu and Phe (Table 1). His, Gly, Glu, Cys were not detected in thesweet pepper extracts. In general, concentrations of amino acidswere not affected when pepper was stored at 5 ◦C for 21 days, andwere maintained close to the initial concentrations before storage.Arginine, the main amino acid in sweet pepper, is considered asemi-essential amino acid for humans, being required to ensurethat the liver, joints, muscles (including the heart muscle) and skinare kept healthy. Additionally, Oliveira et al.64 found that arginineis also of great importance as an intermediary product in ureasynthesis, being (theoretically) the most efficient form of storageN because of its low C/N ratio.65 The preservation of the amino acidprofile at the studied temperature is important for the nutritionalcharacteristics of sweet pepper. Therefore, our results demonstratethat although the enzymatic metabolism related with antioxidantenzymes or color was altered by low temperature, only minorchanges were observed for the amino acid composition of sweetpepper. Additionally, a positive differential effect was found inthe urea treatment, which improved color and the mitigationof lipid peroxidation through increased catalase and ascorbateperoxidase.


27

21


Table 1. Effect of foliar urea (UR) and antitranspirant (AT) on the free amino acid concentration of sweet pepper at 7, 14 and 21 days of storage at5 ◦C

Free amino acids (µmol L−1)

Days Treatment Ser Arg Asp Thr Ala Pro Lys Tyr Val Ile Leu Phe

0 5.08 16.13ab 4.11a 2.54 3.01 0.00 2.15 0.42a 2.02a 0.00a 0.98 0.22

7 CN 4.31 15.89ab 5.22bc 2.35 3.30 0.00 1.97 3.67d 2.55b 0.00a 0.24 0.00

UR 4.25 16.42ab 5.87cd 2.52 3.34 0.00 1.74 1.97abcd 1.87a 0.00a 0.00 0.00

AT 5.59 18.86bc 6.18d 2.92 3.35 0.25 2.29 2.53bcd 2.03a 0.23ab 0.70 0.24

14 CN 4.60 16.00ab 4.11a 2.31 3.19 0.00 1.92 1.41abc 2.52b 0.00a 0.54 0.00

UR 4.33 16.44ab 5.66bcd 1.97 2.56 0.00 1.90 1.65abc 1.86a 0.00a 0.24 0.00

AT 5.63 22.67c 6.25d 2.54 3.31 0.00 2.71 3.19cd 1.74a 0.00a 0.74 0.24

21 CN 5.29 16.02ab 4.91ab 2.38 3.45 0.25 2.50 0.95ab 1.79a 0.00a 0.30 0.24

UR 5.98 13.00a 5.12bc 2.68 3.58 0.27 2.34 0.96ab 1.78a 0.48b 1.04 0.25

AT 5.70 15.37ab 5.92cd 2.58 3.18 0.00 1.59 1.45abc 1.68a 0.00a 1.12 0.00

Values with the same letter within the same column are not significantly different at P < 0.05 (Duncan’s multiple range test).

ACKNOWLEDGEMENTSPaula Cuadra-Crespo is the recipient of a pre-doctoral fellowshipfrom the IMIDA. The authors thank MC Pinero for his technicalassistance. This work has been supported by the Instituto Nacionalde Investigaciones Agrarias (INIA), through project RTA2008-00089and POI 07-021. Part of this work was also funded by the EuropeanSocial Fund.

REFERENCES1 Banaras M, Bosland PW and Lownds NK, Effects of harvest time and

growth conditions on storage and post-storage quality of freshpeppers (Capsicum annuum L.). Pak J Bot 37:337–344 (2005).

2 Smith DL, Stommel JR, Fung RWM, Wang CY and Whitaker BD,Influence of cultivar and harvest method on postharvest storagequality of pepper (Capsicumannuum L.) fruit. PostharvestBiolTechnol42:243–247 (2006).

3 Dıaz-Perez JC, Muy-Rangel MD and Mascorro AG, Fruit size and stageof ripeness affect postharvest water loss in bell pepper fruit(Capsicum annuum L.). J Sci Food Agric 87:68–73 (2007).

4 Paull RE, Chilling injury of crops of tropical and subtropical origin, inChilling Injury of Horticultural Crops, ed. by Wang CY. CRC Press, BocaRaton, FL, pp. 17–36 (1990).

5 Kehr E, Susceptibilidad a dano por enfriamiento en poscosechade pimiento y tratamientos para disminuir su efecto. Agric Tec62:509–518 (2002).

6 Gonzalez-Aguilar GA, Gayoso L, Cruz R, Fortiz J, Baez R and Wang CY,Polyamines induced by hot water treatments reduce chilling injuryand decay in pepper fruit. Postharvest Biol Technol 18:19–26 (2000).

7 Singh BK, Sharma SR and Singh B, Combining ability for superoxidedismutase, peroxidase and catalase enzymes in cabbage head(Brassica oleracea var. capitata L.). Sci Hortic 122:195–199 (2009).

8 Murcia MA, Jimenez AM and Martınez-Tome M, Vegetablesantioxidant losses during industrial processing and refrigeratedstorage. Food Res Int 42:1046–1052 (2009).

9 Boonsiri K, Ketsa S and Van Doorn WG, Seed browning of hotpeppers during low temperature storage. Postharvest Biol Technol45:358–365 (2007).

10 Koc A, Gasch AP, Rutherford JC, Kim HY and Gladyshev VN, Methioninesulfoxide reductase regulation of yeast lifespan reveals reactiveoxygen species-dependent and -independent components ofaging. Proc Natl Acad Sci USA 101:7999–8004 (2004).

11 Nunez M, Mazzafera P, Mazorra LM, Siquira WJ and Zullo MAT,Influence of a brassinosteroid analogue on antioxidant enzymes inrice grown in culture medium with NaCl. Biol Plant 47:67–70 (2003).

12 Ogiso Y, Hosoda-Yabe R, Kawamoto Y, Kawamoto T, Kato K andYabe T, An antioxidant of dried chili pepper maintained its activity

through postharvest ripening for 18 months. Biosci BiotechnolBiochem 72:3297–3300 (2008).

13 Knoche M, Petracek PD and Bukovac MJ, Urea penetration of isolatedtomato fruit cuticles. J Am Soc Hortic Sci 119:761–764 (1994).

14 Crisosto CH and Mitchell JP, Preharvest factors affecting fruit andvegetable quality, in Postharvest Technology of Horticultural Crops,ed. by Kader AA. University of California, Oakland, CA, p. 519 (2002).

15 del Amor FM, Yield and fruit quality response of sweet pepperto organic and mineral fertilization. Renewable Agric Food Syst22:233–238 (2006).

16 Perez-Lopez AJ, Lopez-Nicolas JM, Nunez-Delicado E, Del Amor FMand Carbonell-Barrachina AA, Effects of agricultural practices oncolor, carotenoids composition, and minerals contents of sweetpeppers, cv. Almuden. J Agric Food Chem 55:8158–8164 (2007).

17 Meheriuk M, McKenzie DL, Neilsen GH and Hall JW, Fruit pigmentationof four green apple cultivars responds to urea sprays but not tonitrogen fertilization. HortScience 31:992–993 (1996).

18 Yildirim E, Guvenc I, Turan M and Karatas A, Effect of foliar ureaapplication on quality, growth, mineral uptake and yield of broccoli(Brassica oleracea L., var. italica). Plant Soil Environ 53:120–128(2007).

19 Kaya C and Higgs D, Relationship between water use and ureaapplication in salt-stressed pepper plants. J Plant Nutr 26:19–30(2003).

20 Harris JR and Bassuk NL, Effects of defoliation and antitranspiranttreatments on transplant response of scarlet oak, green ash andTurkish hazelnut. J Arboric 21:33–36 (1995).

21 Gale J and Poljakoff-Mayber A, Effect of antitranspirant spray on theuptake and transport of rubidium by plants. Plant Cell Physiol4:285–288 (1963).

22 del Amor FM and Rubio JS, Effects of antitranspirant spray andpotassium : calcium:magnesium ratio on photosynthesis, nutrientand water uptake, growth, and yield of sweet pepper. J Plant Nutr32:97–111 (2009).

23 Phelps E and Morelock T, Antitranspirants as a selection techniquein breeding for tomato fruit cracking resistance. HortScience22:728–733 (1987).

24 Lipe WN, Hondnet K, Gerst M and Wendt CW, Effects ofantitranspirants on water use and yield of greenhouse and fieldgrown onion. HortScience 17:242–244 (1982).

25 Byari SH and Okeefe RB, Effect of antitranspirants on heat and droughtresponse of potato varieties. HortScience 17:513–523 (1982).

26 del Amor FM and Gomez-Lopez MD, Agronomical response and wateruse efficiency of sweet pepper plants grown in different greenhousesubstrates. Hortscience 44:810–814 (2009).

27 Kahkonen MP, Hopia AI, Vuorela HJ, Rauha JP, Pihlaja K, Kujala TS,et al, Antioxidant activity of plant extracts containing phenoliccompounds. J Agric Food Chem 47:3954–3962 (1999).


27

22


28 del Amor FM, Cuadra-Crespo P, Varo P and Gomez MC, Influence offoliar urea on the antioxidant response and fruit color of sweetpepper under limited N supply. J Sci Food Agric 89:504–510 (2009).

29 Heath RL and Packer L, Photoperoxidation in isolated chloroplasts. I.Kinetics and stoichiometry of fatty acid peroxidation. Arch BiochemBiophys 125:189–198 (1968).

30 Waters, UPLC Amino Acid Analysis Solution. Waters Corporation, Milford,MA (2006).

31 ElsevieValero C, Crisosto CH and Slaughter D, Relationship betweennondestructive firmness measurements and commerciallyimportant ripening fruit stages for peaches, nectarines and plums.Postharvest Biol Technol 44:248–253 (2007).

32 King AD and Bolin HR, Physiological and microbiological storagestability of minimally processed fruits and vegetables. Food Technol43:132–135 (1989).

33 Kays SJ, Post-harvest Physiology and Handling of Perishable PlantProducts. Van Nostrand-Reinhold, New York, p. 733 (1991).

34 Encyclopedia of Fruits and Nuts, ed. by Janick J and Paull RE. PurdueUniversity, West Lafayette. IN; University of Hawaii at Manoa, HI,p. 976 (2008).

35 Von Mollendorff LJ, De Villiers OT, Jacobs G and Westraad I, Molecularcharacteristics of pectic constituents in relation to firmness,extractable juice, and woolliness in nectarines. J Am Soc HorticSci 118:77–80 (1993).

36 Meir S, Rosenberger I, Aharon Z, Grinberg S and Fallik E, Improvementof the postharvest keeping quality and colour development ofbell pepper (cv. ‘Maor’) by packaging with polyethylene bags at areduced temperature. Postharvest Biol Technol 5:303–309 (1995).

37 Cox KA, McGhie TK, White A and Woolf AB, Skin colour and pigmentchanges during ripening of ‘Hass’ avocado fruit. Postharvest BiolTechnol 31:287–294 (2004).

38 del Amor FM, Growth, photosynthesis and chlorophyll fluorescenceof sweet pepper plants as affected by the cultivation method. AnnAppl Biol 148:133–139 (2006).

39 Papoulias E, Siomos AS, Koukounaras A, Gerasopoulos D andKazakis E, Effects of genetic, pre- and post-harvest factors onphenolic content and antioxidant capacity of white asparagusspears. Int J Mol Sci 10:5370–5380 (2009).

40 Robards K, Prenzler PD, Tucker G, Swatsitang P and Glover W, Phenoliccompounds and their role in oxidative processes in fruits. Food Chem66:401–436 (1999).

41 Chimphango SBM, Musil CF and Dakora FD, Response of purelysymbiotic and NO3

− fed nodulated plants of Lupinus luteus andVicia atropurpurea to ultraviolet-B radiation. J Exp Bot 54:1771–1784(2003).

42 Yang SH, Tao J, Liu XF, Guo DA and Zheng JH, Effects of carbon sourceand nitrogen source on callus growth and flavonoid content inGlycyrrhiza uralensis. J Chin Mater Med 31:1857–1859 (2006).

43 Shin Y, Liu RH, Nock JF, Holliday D and Watkins CB, Temperatureand relative humidity effects on quality, total ascorbic acid,phenolics and flavonoid concentrations, and antioxidant activityof strawberry. Postharvest Biol Technol 45:349–357 (2007).

44 Padda MS and Picha DH, Effect of low temperature storage on phenoliccomposition and antioxidant activity of sweet potatoes. PostharvestBiol Technol 47:176–180 (2008).

45 Cordenunsi BR, Genovese MI, do Nascimento JR, Hassimotto NM, dosSantos RJ and Lajolo FM, Effects of temperature on the chemicalcomposition and antioxidant activity of three strawberry cultivars.Food Chem 91:113–121 (2005).

46 Ayala-Zavala JF, Wang SY, Wang CY and Gonzalez-Aguilar GA, Effectof storage temperatures on antioxidant capacity and aromacompounds in strawberry fruit. Lebensm Wiss Technol 37:687–695(2004).

47 Wongsheree T, Ketsa S and van Doorn WG, The relationship betweenchilling injury and membrane damage in lemon basil (Ocimum ×citriodourum) leaves. Postharvest Biol Technol 51:91–96 (2009).

48 Imahori Y, Kanetsune Y, Ueda Y and Chachin K, Changes in hydrogenperoxide content and antioxidative enzyme activities during thematuration of sweet pepper (Capsicum annuum L.) fruit. J Japan SocHortic Sci 69:690–695 (2000).

49 Cadenas E and Davies JA, Mitochondrial free radical generation:oxidative stress and aging. Free Radic Biol Med 29:220–230 (2000).

50 Imahori Y, Takemura M and Bai J, Chilling-induced oxidative stressand antioxidant responses in mume (Prunus mume) fruit during lowtemperature storage. Postharvest Biol Technol 49:54–60 (2008).

51 Baker NR, Chilling stress and photosynthesis, in Causes ofPhotooxidative Stress and Amelioration of Defense Systems in Plants,ed. by Foyer CH and Mullineaux PM. CRC Press, Boca Raton, FL,pp. 127–154 (1994).

52 Zilkah S, Wiesmann Z, Klein I and David I, Foliar applied urea improvesfreezing protection to avocado and peach. Sci Hortic 66:85–92(1996).

53 Lim CS, Kang SM and Cho JL, Antioxidizing enzyme activities inchilling-sensitive and chilling-tolerant pepper fruit as affected bystage of ripeness and storage temperature. J Am Soc Hortic Sci134:156–163 (2009).

54 Ju Z, Yuan Y, Liou C, Zhan S and Xiu S, Effects of low temperature onH2O2 and heart browning of Chili and Yali pear (Pyrus bretscheideriR.). Sci Agric Sinica 27:77–81 (1994).

55 Blokhina O, Virolainen E and Fagerstedt KV, Antioxidants, oxidativedamage and oxygen deprivation stress: a review. Ann Bot91:179–194 (2003).

56 Wang CY, Kramer GF, Whitaher BD and Lusby WR, Temperaturepreconditioning increases tolerance to chilling injury and alterslipid composition in zucchini squash. J Plant Physiol 140:229–235(1992).

57 Wang CY, Effects of temperature preconditioning on catalase,peroxidase, and superoxide dismutase in chilled zucchini squash.Postharvest Biol Technol 5:67–76 (1995).

58 Davey MW, Stals E, Panis B, Keulemans J and Swennen RL, High-throughput determination of malondialdehyde in plant tissues.Anal Biochem 347:201–207 (2005).

59 Posmyk MM, Bailly C, Szafranska K, Janas KM and Corbineau F,Antioxidant enzymes and isoflavonoids in chilled soybean (Glycinemax. (L.) Merr.) seedlings. J Plant Physiol 162:403–412 (2005).

60 Wismer WV, Worthing WM, Yada RY and Marangoni AG, Membranelipid dynamics and lipid peroxidation in the early stages of low-temperature sweetening in tubers of Solanum tuberosum. PhysiolPlant 102:396–410 (1998).

61 Gualanduzzi S, Baraldi E, Braschi I, Carnevali F, Gessa CE and DeSantis A, Respiration, hydrogen peroxide levels and antioxidantenzyme activities during cold storage of zucchini squash fruit.Postharvest Biol Technol 52:16–23 (2009).

62 Gomes MH and Rosa E, Free amino acid composition in primaryand secondary inflorescences of 11 broccoli (Brassica oleracea var.italica) cultivars and its variation between seasons. J Sci Food Agric81:295–299 (2000).

63 Belitz HD and Grosch W, Amino acids, peptides, and proteins, FoodChemistry. Springer, Berlin, pp. 8–34 (1999).

64 Oliveira AP, Pereira DM, Andrade PB, Valenta P, Sousa C, Pereira JA,et al, Free amino acids of tronchuda cabbage (Brassica oleracea L.Var. costata DC): influence of leaf position (internal or external) andcollection time. J Agric Food Chem 56:5216–5221 (2008).

65 Titus JS and Kang SM, Nitrogen metabolism, translocation, andrecycling in apple trees. Hortic Rev 4:204–246 (1982).


27

23

Short CommunicationReceived: 21 May 2010 Revised: 5 July 2010 Accepted: 9 July 2010 Published online in Wiley Online Library: 5 August 2010


Factors affecting ANKOM fiber analysisof forage and browse varying in condensedtannin concentrationThomas H Terrill,a Richard M Wolfeb and James P Muirb∗

Abstract

BACKGROUND: Browse species containing condensed tannins (CTs) are an important source of nutrition for grazing/browsinglivestock and wildlife in many parts of the world, but information on fiber concentration and CT–fiber interactions for theseplants is lacking.

RESULTS: Ten forage or browse species with a range of CT concentrations were oven dried and freeze dried and then analyzedfor ash-corrected neutral detergent fiber (NDFom) and corrected acid detergent fiber (ADFom) using separate samples (ADFSEP)and sequential NDF-ADF analysis (ADFSEQ) with the ANKOM fiber analysis system. The ADFSEP and ADFSEQ residues werethen analyzed for nitrogen (N) concentration. Oven drying increased (P < 0.05) fiber concentrations with some species, but notwith others. For high-CT forage and browse species, ADFSEP concentrations were greater (P < 0.05) than NDFom values andapproximately double the ADFSEQ values. Nitrogen concentration was greater (P < 0.05) in ADFSEP than ADFSEQ residues,likely due to precipitation with CTs.

CONCLUSION: Sequential NDF-ADF analysis gave more realistic values and appeared to remove most of the fiber residuecontaminants in CT forage samples. Freeze drying samples with sequential NDF-ADF analysis is recommended in the ANKOM

fiber analysis system with CT-containing forage and browse species.c© 2010 Society of Chemical Industry

Keywords: acid detergent fiber; ANKOM ; freeze drying; neutral detergent fiber; condensed tannins; oven drying

INTRODUCTIONBrowse species containing condensed tannins (CTs), includingherbaceous legumes and woody species, are an importantcomponent of livestock and wildlife diets throughout the world1,2

but there is little information available on the nutritional adequacyof CT-containing browse species. There have been a number ofreports on the potential benefits of including low to medium levelsof CT-containing forage and browse species in ruminant diets,including reducing bloat,3 production of greenhouse gases,4 andinfection by gastrointestinal nematodes,5,6 as well as improvingprotein utilization efficiency and increasing weight gains, milkproduction, and reproductive performance of livestock.7 – 9 Intakeof high CT-containing forages has been improved by sun-drying10

or treating with polyethylene glycol to bind the tannins.11 Asinterest in CT forage and browse species as components of thediet for both domestic livestock and ruminant wildlife continuesto grow, accurate laboratory analysis of quality indices for theseforages is critical. Muir et al.12 and Wolfe et al.13 reported a widerange in CT and N concentrations of several native, herbaceouslegumes from Texas, USA, but information on fiber concentrationand possible CT–fiber interactions is limited for these and otherbrowse species found throughout the world.

The ANKOM detergent fiber analysis system with filterbag technology14,15 has replaced the crucible method16 inmany herbage nutritive value laboratories throughout theworld, although both are based on the same general chemical

principles.17 Terrill and Koivisto18 reported higher ADFSEP thanADFSEQ values for oven-dried cool-season and warm-seasonpasture legumes high in CT using the ANKOM method, butlittle information is available on the application of this systemto CT-containing browse species. The purpose of the currentinvestigation was to measure the effect of herbage sampledrying method and ANKOM detergent fiber analysis method onresulting fiber constituents from herbaceous and woody speciesvarying in CT concentration.

MATERIALS AND METHODSLeaf material from 10 native and introduced plant species varyingin CT concentration and consumed by goats and deer in centralTexas (Table 1) were collected from three replicates of 1.5 × 3.0 mplots at the Texas AgriLife Research Center, Stephenville, Texas.

∗ Correspondence to: James P Muir, Texas AgriLife Research, Texas A&M System,1229 North U.S. Highway 281, Stephenville, TX 76401, USA.E-mail: [email protected]

a Agricultural Research Station, Fort Valley State University, Fort Valley, GA 31030,USA

b Texas AgriLife Research, Texas A&M System, 1229 North U.S. Highway 281,Stephenville, TX 76401, USA


27

24

www.soci.org TH Terrill, RM Wolfe and JP Muir

Table 1. Names, growth habits and condensed tannin (CT) concentrations of herbage species used in the trial

SpeciesCT concentration

Scientific name Common name Type (g kg−1)

Acacia angustissima var. hirta Eastern prairie acacia Herbaceous perennial legume 5∗

Desmodium paniculatum Panicled tick-clover Herbaceous perennial legume 121

Lespedeza procumbens Creeping bush-clover Herbaceous perennial legume 66

Lespedeza stuevei Tall bush-clover Herbaceous perennial legume 123

Smilax rotundifolia Common smilax Perennial vine 34

Quercus sinuata var.breviloba Shin oak Woody perennial 125

Leucaena retusa Golden-ball lead-tree Woody perennial 13

Gleditsia tricanthos Common Honey-locust Woody perennial 92

Cynodon dactylon Common Bermudagrass Perennial grass Trace

Panicum virgatum Switchgrass cv. Alamo Perennial grass Trace

∗ Total CT determined according to Terrill et al.21 using a self-standard except for species without measurable CT.

For woody perennials and Smilax, three different plants in a 10 kmradius of Stephenville were used as replications, with leaf materialcollected below the approximate browse height for goats (1.5 m).For non-CT controls, leaf material of two species of grasses wascollected (Table 1). Half of each sample was oven dried at 55 ◦C for48 h, and the other half was freeze dried for 72 h. The dry sampleswere then ground with a Wylie mill to pass a 1 mm mesh screenand stored in air-tight plastic bottles at room temperature.

Samples were analyzed for dry matter using AOAC procedures,19

and neutral detergent fiber (NDF), corrected acid detergentfiber (ADFom) using separate samples (ADFSEP), and sequentialNDFom-ADFom (ADFSEQ) concentrations using the ANKOMModel 200 and the fiber bag technique developed by ANKOM

(ANKOM Technology, Macedon, NY, USA). Heat-stable α-amylaseand sodium sulfite were added to the neutral detergent solutionas recommended with the ANKOM system.20 After ADFSEP andADFSEQ extractions were completed, the fiber mat was removedfor each sample and analyzed for total N using an Elementar VarioMacro combustion analyzer (Elementar Americas, Inc., Mt. Laurel,NJ, USA). All NDF and ADF residues were analyzed and correctedfor ash content.19 The ash-corrected values were expressed asNDFom and ADFom. The freeze-dried samples were analyzed fortotal condensed tannin (TCT) concentration using the methodof Terrill et al.,21 with purified CTs from each species used asstandards.13

The data were analyzed as a completely randomized blockdesign with three replicates (plots or cuttings) using the GLMprocedure of SAS.22 Differences were considered significant atP < 0.05 and means were separated using the LSMeans procedurewhere appropriate. For NDFom data, species and drying methodwere included in the model, while species, drying method, andADF analysis method were included in the model for ADFom andADFN data.

RESULTSTotal condensed tannin concentration ranged from trace amountsin grass samples to 125 g kg−1 in shin oak (Table 1). There wasa wide range of TCT levels in both the herbaceous legumes(5–123 g kg−1) and woody species (13–125 g kg−1) tested.

There was a cultivar × drying method interaction for NDFomdata, while a cultivar × drying method × ADF analysis methodinteraction was also a factor for ADFom and ADFN (Table 2).

Drying method had no effect on NDFom concentration ofherbaceous legumes and grasses, but oven drying increasedNDFom concentration compared with freeze drying for woodyspecies. For herbaceous native legumes, differences in ADFom dueto drying method increased as CT levels increased, with greaterADFSEP in oven-dried than freeze-dried bush-clovers, which havegreater CTs than eastern prairie acacia.13 These differences wereremoved with sequentially analyzed ADFom for seven of the tenspecies tested.

For woody species, ADFSEP and ADFSEQ residues were greaterin oven-dried than freeze-dried material, with greater differencesin the ADFSEP residues. Drying method and ADFom analysisprotocol had no effect on ADFom residues of grasses. Dryingmethod also had little effect on ADFSEP-N and ADFSEQ-N valuesfor the species tested, but N concentrations were approximatelydouble in ADFSEP compared with ADFSEQ residues.

DISCUSSIONThe TCT levels in these plants were generally lower thanTCT values reported for similar browse species analyzed us-ing purified Quebracho CT as the standard.13 Although it ismore time consuming to purify CTs from individual species touse as standards, this provides a more accurate and mean-ingful analysis of TCT using the butanol HCl colorimetricmethod.13

Forage drying method (oven dried versus freeze dried) hada greater effect on ANKOM fiber analysis of woody speciesthan herbaceous forages, but the major effect of CT in thisstudy appears to be on ADFSEP and ADFSEQ values, withmuch greater ADFSEP concentrations for most of the browsespecies tested, regardless of drying method (Table 2). In fact,ADFSEP was greater than NDFom in six of eight herbaceousnative legumes and/or woody species tested. As hemicelluloseconcentration of forages is estimated by subtracting ADF fromNDF, these results would give negative values for this fiberfraction, which is not possible. Similar results have been reportedpreviously for herbaceous and brushy CT-containing speciesusing the crucible method of detergent fiber analysis20 but thedifferences were greater using the ANKOM system in the currentinvestigation.

The contaminants in ADFSEP residues were N and othermaterials, most likely CTs. Terrill et al.23 reported higher N and


27

25

Condensed tannins and ANKOM fiber analysis www.soci.org

Table 2. Ankom detergent fiber analysis (g kg−1 dry matter) of oven-dried (OD) and freeze-dried (FD) forage and browse species with a range ofcondensed tannin levels∗

Constituents of the Ankom Fiber Analysis SystemDrying

Species method NDFom ADFSEP ADFSEQ ADFSEP N ADFSEQ N

Eastern prairie acacia OD 250a 150aA 93aB 32.9aA 25.6aB

FD 243a 197bA 96aB 44.7bA 20.9bB

Panicled tick-clover OD 276a 451aA 181aB 25.2aA 11.9aB

FD 291a 455aA 194aB 29.6bA 11.6aB

Creeping bush-clover OD 322a 458aA 206aB 25.9aA 11.8aB

FD 290a 412bA 196aB 25.2aA 11.3aB

Tall bush-clover OD 285a 398aA 201aB 22.1aA 9.8aB

FD 284a 434bA 207aB 23.9aA 9.0aB

Common smilax OD 391a 551aA 304aB 25.0aA 19.5aB

FD 292b 443bA 227bB 22.2aA 7.8bB

Shin Oak OD 406a 351aA 240aB 15.5aA 9.8aB

FD 353b 262bA 216bB 9.4bA 8.7aB

Golden-ball lead-tree OD 336a 375aA 186aB 39.6aA 14.4aB

FD 250b 352aA 159aB 41.8aA 11.9bB

Honey locust OD 440a 506aA 285aB 25.6aA 18.5aB

FD 333b 386bA 231bB 21.3bA 10.5bB

Common bermudagrass OD 611a 307aA 280aA 4.7aA 4.5aA

FD 595a 299aA 274aA 4.8aA 4.0aA

Alamo switchgrass OD 581a 294aA 262aA 3.8aA 3.6aA

FD 560a 297aA 264aA 3.7aA 3.9aA

∗ Standard error for cultivar × drying method interaction (P < 0.001) for NDFom data = 13.2 g kg−1 dry matter. Standard error for cultivar × dryingmethod × ADF analysis method interactions (P < 0.05) for ADFom and ADFN data = 12.9 and 1.3 g kg−1 dry matter, respectively.Means within columns for each species comparing OD and FD followed by different lower-case letters differ at P < 0.05.Means within lines under columns ADFSEP and ADFSEQ or under columns ADFSEP N and ADFSEQ N followed by different upper-case letters differ atP < 0.05.NDFom, neutral detergent fiber; ADFSEP, acid detergent fiber (ADFom), separate sample analysis; ADFSEQ, ADFom sequential analysis; N, nitrogen.

CT concentrations in NDF and ADFSEP residues from the highCT plant sericea lespedeza [Lespedeza cuneata (Dum-Cours.) G.Don.] after detergent fiber analysis using the crucible method.In their study, the CTs were recovered in the lignin fraction. In acrucible method study with a range of CT-containing plants, Pagan-Riestra et al.20 reported that adding sodium sulfite to NDF analysisfollowed by sequential ADF analysis removed most of the fiberresidue contaminants and gave more accurate fiber estimates. Thecurrent investigation confirmed that a similar protocol should beused for the ANKOM analysis system. With this system, sodiumsulfite and α-amylase are routinely added to the neutral detergentextraction step, and after rinsing and drying the filter bags withNDF residues, these residues are then extracted with ADF solution.This protocol greatly reduced N in ADF residues and lowered theADF values by over 50% for both oven-dried and freeze-driedsamples for several of the forage/browse species containing CT inthe current investigation.

As the ANKOM detergent fiber analysis system continues togrow in popularity in animal nutrition laboratories around theworld, as does the interest and benefits of including CT-containingplants in ruminant diets, the following analytical protocols thatprovide realistic fiber values for these forage and browse speciesis critical. This is particularly true for ADF, as it is often used toestimate digestibility of forages. With separate NDF-ADF analysisusing the ANKOM system, inflated ADF values would lead toan underestimation of TDN for CT-containing forage and browsespecies.

CONCLUSIONSSeparate sample ADFom analysis using the ANKOM systemcan increase fiber estimates in CT-containing herbaceous legumeand browse-type plant species. Use of freeze drying and sequentialNDF-ADF analysis minimizes ADFom contaminants and gives morerealistic fiber estimates when using the ANKOM fiber analyzerwith plant material containing CTs.

REFERENCES1 van Rooyen N, Grunow JO and Theron GK, Veld management, in Game

Ranch Management, ed. by Bothma J du P. J.L. van Shaik, Pretoria,pp. 567–607 (1989).

2 Berg WA, Native forb persistence under grazing of a southern GreatPlains planting, in Proceedings of the Vth Int. Rangeland Congress,Society for Range Management, Denver, Colorado. Society for RangeManagement, Wheat Ridge CO., USA, pp. 46–47 (1996).

3 Min BR, Pinchak WE, Fulford JD and Puchala R, Wheat pasture bloatdynamics, in vitro ruminal gas production, and potential bloatmitigation with condensed tannins. JAnimSci 83:1322–1331 (2005).

4 Puchala R, Min BR, Goetsch AL and Sahlu T, The effect of a condensedtannin-containing forage on methane emission by goats. J Anim Sci83:182–186 (2005).

5 Shaik SA, Terrill TH, Miller JE, Kouakou B, Kannan G, Kaplan RM,et al, Sericea lespedeza hay as a natural deworming agentagainst gastrointestinal nematode infection in goats. Vet Parasitol139:150–157 (2006).

6 Lange K, Olcott DD, Miller JE, Mosjidis JA, Terrill TH, Burke JM, et al,Effect of sericea lespedeza (Lespedeza cuneata) fed as hay, on naturaland experimental Haemonchus contortus infections in lambs. VetParasitol 141:273–278 (2006).


27

26

www.soci.org TH Terrill, RM Wolfe and JP Muir

7 Waghorn GC, Ulyatt MJ, John A and Fisher MT, The effect of condensedtannins on the site of digestion of amino acids and other nutrientsin sheep fed on Lotus corniculatus L. Br J Nutr 57:115–126 (1987).

8 Min BR, McNabb WC, Barry TN, Kemp PD, Waghorn GC andMcDonald MF, The effect of condensed tannins in Lotus corniculatusupon reproductive efficiency and wool production in sheep duringlate summer and autumn. J Agric Sci 132:323–334 (1999).

9 Barry TN and McNabb WC, The implications of condensed tannins onthe nutritive value of temperate forages fed to ruminants. Br J Nutr81:263–272 (1999).

10 Terrill TH, Windham WR, Evans JJ and Hoveland CS, Condensedtannins in sericea lespedeza: Effect of preservation method ontannin concentration. Crop Sci 30:219–224 (1990).

11 Jones WT and Mangan JL, Complexes of the condensed tannins ofsainfoin (Onobrychis viciifolia Scop.) with fraction 1 leaf protein andwith submaxillary mucoprotein, and their reversal by polyethyleneglycol and pH. J Sci Food Agric 28:126–136 (1977).

12 Muir JP, Taylor J and Interrante SM, Herbage and seed from nativeperennial herbaceous legumes of Texas. Range Ecol Manage58:643–651 (2005).

13 Wolfe RM, Terrill TH and Muir JP, Season and drying method effectson condensed tannin levels in perennial herbaceous legumes. J SciFood Agric 88:1060–1067 (2008).

14 Komarek AR, Robertson JB and Van Soest PJ, A comparison of methodsfor determining ADF using the filter bag technique versusconventional filtration. J Dairy Sci 77:24–26 (1993).

15 Vogel KP, Pederson JF, Masterson SD and Toy JJ. Evaluation of a filterbag system for NDF, ADF, and IVDMD forage analysis. Crop Sci39:276–279 (1999).

16 Van Soest PJ. Use of detergents in the analysis of fibrous feeds. 1.Preparation of fiber residues of low nitrogen content. J Assoc OffAnal Chem 46:825–829 (1963).

17 Van Soest PJ, Robertson JB and Lewis BA, Methods for dietary fiber,neutral detergent fiber, and nonstarch polysaccharides in relationto animal nutrition. J Dairy Sci 74:3583–3597 (1991).

18 Terrill TH, and Koivisto JM, Comparison of the ANKOM , fibrecaps,and fibrebags systems for detergent fiber analysis of condensedtannin (CT)-containing and non-CT forages. Proc Am For Grassl Coun12:263–267 (2003).

19 Association of Official Analytical Chemists, Official Methods of Analysis,14th edition. AOAC, Washington, DC, pp. 129–130 (1984).

20 Pagan-Riestra S, Wolfe RM, Terrill TH and Muir JP, Effect of dryingmethod and assay methodology on detergent fiber analysisin plants containing condensed tannins. Anim Feed Sci Technol154:119–124 (2009).

21 Terrill TH, Rowan AW, Douglas GB and Barry TN, Determination ofextractable and bound condensed tannin concentration in forageplants, protein concentrate meals, and cereal grains. J Sci Food Agric58:321–329 (1992).

22 SAS Institute, SAS/STAT Software: changes and enhancements.Release 6.07, SAS Technical Report, SAS Institute, Cary, NC (1992).

23 Terrill TH, Windham WR, Evans JJ and Hoveland CS, Effect of dryingmethod and condensed tannin on detergent fiber analysis of sericealespedeza. J Sci Food Agric 66:337–343 (1994).


quinoafacts others

Documents

clover creeping bush

clover tall bush

wiley online library

neutral detergent ber

van doorn wg

performance liquid chromatography

acid detergent ber

detergent ber analysis