INTRODUCTION

Color is important to many foods, boththose that are unprocessed and those that aremanufactured. Together with flavor and tex-ture, color plays an important role in foodacceptability. In addition, color may providean indication of chemical changes in a food,such as browning and caramelization. For afew clear liquid foods, such as oils and bev-erages, color is mainly a matter of transmis-sion of light. Other foods are opaque—theyderive their color mostly from reflection.

Color is the general name for all sensationsarising from the activity of the retina of theeye. When light reaches the retina, the eye'sneural mechanism responds, signaling coloramong other things. Light is the radiantenergy in the wavelength range of about 400to 800 nm. According to this definition, color(like flavor and texture) cannot be studiedwithout considering the human sensory sys-tem. The color perceived when the eye viewsan illuminated object is related to the follow-ing three factors: the spectral composition ofthe light source, the chemical and physicalcharacteristics of the object, and the spectralsensitivity properties of the eye. To evaluatethe properties of the object, we must stan-dardize the other two factors. Fortunately,the characteristics of different people's eyes

for viewing colors are fairly uniform; it is nottoo difficult to replace the eye by someinstrumental sensor or photocell that can pro-vide consistent results. There are several sys-tems of color classification; the mostimportant is the CIE system (CommissionInternational de 1'Eclairage—InternationalCommission on Illumination). Other systemsused to describe food color are the Munsell,Hunter, and Lovibond systems.

When the reflectance of different coloredobjects is determined by means of spectro-photometry, curves of the type shown in Fig-ure 6-1 are obtained. White materials reflectequally over the whole visible wavelengthrange, at a high level. Gray and black materi-als also reflect equally over this range but toa lower degree. Red materials reflect in thehigher wavelength range and absorb theother wavelengths. Blue materials reflect inthe low-wavelength range and absorb thehigh-wavelength light.

CIE SYSTEM

The spectral energy distribution of CIElight sources A and C is shown in Figure 6-2.CIE illuminant A is an incandescent lightoperated at 28540K, and illuminant C is thesame light modified by filters to result in a

Color

CHAPTER 6

Figure 6-1 Spectrophotometric Curves of Col-ored Objects. Source: From Hunter AssociatesLab., Inc.

spectral composition that approximates thatof normal daylight. Figure 6-2 also showsthe luminosity curve of the standard observeras specified by CIE. This curve indicates

how the eyes of normal observers respond tothe various spectral light types in the visibleportion of the spectrum. By breaking downthe spectrum, complex light types are re-duced to their component spectral lighttypes. Each spectral light type is completelydetermined by its wavelength. In some lightsources, a great deal of radiant energy is con-centrated in a single spectral light type. Anexample of this is the sodium lamp shown inFigure 6-3, which produces monochromaticlight. Other light sources, such as incandes-cent lamps, give off a continuous spectrum.A fluorescent lamp gives off a continuousspectrum on which is superimposed a linespectrum of the primary radiation producedby the gas discharge (Figure 6-3).

In the description of light sources, refer-ence is sometimes made to the black body.This is a radiating surface inside a hollowspace, and the light source's radiation comesout through a small opening. The radiation isindependent of the type of material the lightsource is made of. When the temperature isvery high, about 600O0K the maximum ofthe energy distribution will fall about in themiddle of the visible spectrum. Such energydistribution corresponds with that of daylighton a cloudy day. At lower temperatures, themaximum of the energy distribution shifts tolonger wavelengths. At 3000° K, the spectralenergy distribution is similar to that of anincandescent lamp; at this temperature theenergy at 380 nm is only one-sixteenth ofthat at 780 nm, and most of the energy isconcentrated at higher wavelengths (Figure6-3). The uneven spectral distribution ofincandescent light makes red objects lookattractive and blue ones unattractive. This iscalled color rendition. The human eye hasthe ability to adjust for this effect.

The CIE system is a trichromatic system;its basis is the fact that any color can be

REFLE

CTAN

CE (%

)

WAVELENGTH

WAVE LENGTH nmFigure 6-2 Spectral Energy Distribution ofLight Sources A and C, the CIE, and RelativeLuminosity Function y for the CIE StandardObserver

RELA

TIVE

LUM

INOS

ITY

(y)

RELA

TIVE

ENER

GY (

A AN

D C)

C.I.E. STANDARDOBSERVER

matched by a suitable mixture of three pri-mary colors. The three primary colors, or pri-maries, are red, green, and blue. Any possiblecolor can be represented as a point in a trian-gle. The triangle in Figure 6-4 shows howcolors can be designated as a ratio of the threeprimaries. If the red, green, and blue values ofa given light type are represented by a, b, andc, then the ratios of each to the total light aregiven by a/(a + b + c), bl(a + b + c), and cl(a+ b + c), respectively. Since the sum of theseis one, then only two have to be known toknow all three. Color, therefore, is deter-mined by two, not three, of these mutuallydependent quantities. In Figure 6-4, a colorpoint is represented by P. By determining thedistance of P from the right angle, the quanti-ties al(a + b + c) and bl(a + b + c) are found.The quantity cl(a + b + c) is then found, byfirst extending the horizontal dotted linethrough P until it crosses the hypotenuse at Qand by then constructing another right angletriangle with Q at the top. All combinations

of a, b, and c will be points inside the trian-gle.

The relative amounts of the three primariesrequired to match a given color are called the

WAVELENGTH NM

Figure 6-3 Spectral Energy Distribution of Sunlight (S), CIE Illuminant (A), Cool White FluorescentLamp (B), and Sodium Light (N)

REL

ATI

VE E

NER

GY

Figure 6-4 Representation of a Color as a Pointin a Color Triangle

COLOR

tristimulus values of the color. The CIE pri-maries are imaginary, because there are noreal primaries that can be combined to matchthe highly saturated hues of the spectrum.

In the CIE system the red, green, and blueprimaries are indicated by X, Y9 and Z. Theamount of each primary at any particularwavelength is given by the values J, y, and z.These are called the distribution coefficientsor the red, green, and blue factors. They rep-resent the tristimulus values for each chosenwavelength. The distribution coefficients forthe visible spectrum are presented in Figure6-5. The values of y correspond with theluminosity curve of the standard observer(Figure 6-2). The distribution coefficientsare dimensionless because they are the num-bers by which radiation energy at each wave-length must be multiplied to arrive at the X,

y, and Z content. The amounts of X, Y, and Zprimaries required to produce a given colorare calculated as follows:

780

X=Ix IRdh

380

780

XY = J y IRdh

380

780

XZ = J z IRdh

380where

/ = spectral energy distribution of illu-minant

R = spectral reflectance of sampledh = small wavelength intervaljc, y, ~z = red, green, and blue factors

The ratios of the primaries can beexpressed as

_ X

*"x+y+z

_ yy"x+y+z

_ zZ~X+Y+Z

The quantities x and y are called the chroma-ticity coordinates and can be calculated foreach wavelength from

Figure 6-5 Distribution Coefficients JC, y, and zfor the Visible Spectrum. Source: From HunterAssociates Lab., Inc.

WAVELENGTH (NANOMETERS)

RELA

TIVE A

MOUN

T

jc = xf(x + y + z)

y= y/(x + y + z)

z=l-(x + y)

A plot of jc versus y results in the CIE chro-maticity diagram (Figure 6-6). When thechromaticities of all of the spectral colors areplaced in this graph, they form a line calledthe locus. Within this locus and the line con-necting the ends, represented by 400 and 700nm, every point represents a color that can bemade by mixing the three primaries. Thepoint at which exactly equal amounts of each

of the primaries are present is called theequal point and is white. This white pointrepresents the chromaticity coordinates ofilluminant C. The red primary is located at jc= 1 and y = O; the green primary at x = O andy = 1; and the blue primary at x = O and y = O.The line connecting the ends of the locusrepresents purples, which are nonspectralcolors resulting from mixing various amountsof red and blue. All points within the locusrepresent real colors. All points outside thelocus are unreal, including the imaginary pri-maries X, Y, and Z. At the red end of thelocus, there is only one point to represent thewavelength interval of 700 to 780 nm. This

Figure 6-6 CIE Chromaticity Diagram

means that all colors in this range can besimply matched by adjustment of luminosity.In the range of 540 to 700 nm, the spectrumlocus is almost straight; mixtures of twospectral light types along this line segmentwill closely match intervening colors withlittle loss of purity. In contrast, the spectrumlocus below 540 nm is curved, indicating thata combination of two spectral lights alongthis portion of the locus results in colors ofdecreased purity.

A pure spectral color is gradually dilutedwith white when moving from a point on thespectrum locus to the white point P. Such astraight line with purity decreasing from 100to O percent is known as a line of constantdominant wavelength. Each color, except thepurples, has a dominant wavelength. Theposition of a color on the line connecting thelocus and P is called excitation purity (pe)and is calculated as follows:

_ *-xw _ y-y*f — —— -̂— — _———.

XP ~~xw yp~~ y\vwhere

jc and y are the chromaticity coordinates ofa color

xw and yw are the chromaticity coordinatesof the achromatic source

xp and yp are the chromaticity coordinatesof the pure spectral color

Achromatic colors are white, black, andgray. Black and gray differ from white onlyin their relative reflection of incident light.The purples are nonspectral chromatic col-ors. All other colors are chromatic; for exam-ple, brown is a yellow of low lightness andlow saturation. It has a dominant wavelengthin the yellow or orange range.

A color can be specified in terms of the tri-stimulus value Y and the chromaticity coor-

dinates x and y. The Y value is a measure ofluminous reflectance or transmittance and isexpressed in percent simply as 7/1000.

Another method of expressing color is interms of luminance, dominant wavelength,and excitation purity. These latter are roughlyequivalent to the three recognizable attrib-utes of color: lightness, hue, and saturation.Lightness is associated with the relativeluminous flux, reflected or transmitted. Hueis associated with the sense of redness, yel-lowness, blueness, and so forth. Saturation isassociated with the strength of hue or the rel-ative admixture with white. The combinationof hue and saturation can be described aschromaticity.

Complementary colors (Table 6-1) areobtained when a straight line is drawnthrough the equal energy point P. When thisis done for the ends of the spectrum locus,the wavelength complementary to the 700 to780 point is at 492.5 nm, and for the 380 to410 point is at 567 nm. All of the wave-lengths between 492.5 and 567 nm are com-plementary to purple. The purples can bedescribed in terms of dominant wavelengthby using the wavelength complementary toeach purple, and purity can be expressed in amanner similar to that of spectral colors.

Table 6-1 Complementary Colors

Wavelength Complementary(nm) Color Color

~400 Violet ~450 Blue Y e " ° W

^ Orange500 Green *550 Yellow 6

600 Orange650 Red ^ UG

700 G r e e n

An example of the application of the CIEsystem for color description is shown in Fig-ure 6-7. The curved, dotted line originatingfrom C represents the locus of the chromatic-ity coordinates of caramel and glycerol solu-tions. The chromaticity coordinates of maplesyrup and honey follow the same locus. Threetriangles on this curve represent the chroma-ticity coordinates of U.S. Department of Agri-culture (USDA) glass color standards for

maple syrup. These are described as lightamber, medium amber, and dark amber. Thesix squares are chromaticity coordinates ofhoney, designated by USDA as water white,extra white, white, extra light amber, lightamber, and amber. Such specifications areuseful in describing color standards for a vari-ety of products. In the case of the light amberstandard for maple syrup, the following valuesapply: x = 0.486, y = 0.447, and T = 38.9 per-

Figure 6-7 CIE Chromaticity Diagram with Color Points for Maple Syrup and Honey Glass ColorStandards

X

y

GLASS COLOR STANDARDSA FOR MAPLE SYRUP• FOR HONEY

cent. In this way, x and y provide a specifica-tion for chromaticity and T for luminoustransmittance or lightness. This is easilyexpressed as the mixture of primaries underilluminant C as follows: 48.6 percent of redprimary, 44.7 percent of green primary, and6.7 percent of blue primary. The light trans-mittance is 38.9 percent.

The importance of the light source andother conditions that affect viewing of sam-ples cannot be overemphasized. Many sub-stances are metameric; that is, they may haveequal transmittance or reflectance at a certainwavelength but possess noticeably differentcolors when viewed under illuminant C.

MUNSELL SYSTEM

In the Munsell system of color classifica-tion, all colors are described by the threeattributes of hue, value, and chroma. Thiscan be envisaged as a three-dimensional sys-tem (Figure 6-8). The hue scale is based onten hues which are distributed on the circum-ference of the hue circle. There are five hues:red, yellow, green, blue, and purple; they arewritten as R, Y, G, B, and P. There are alsofive intermediate hues, YR, GY, BG, PB, andRP. Each of the ten hues is at the midpoint ofa scale from 1 to 10. The value scale is alightness scale ranging from O (black) to 10(white). This scale is distributed on a lineperpendicular to the plane of the hue circleand intersecting its center. Chroma is a mea-sure of the difference of a color from a grayof same lightness. It is a measure of purity.The chroma scale is of irregular length, andbegins with O for the central gray. The scaleextends outward in steps to the limit of purityobtainable by available pigments. The shapeof the complete Munsell color space is indi-cated in Figure 6-9. The description of acolor in the Munsell system is given as //,VIC. For example, a color indicated as 5R

Figure 6-8 The Munsell System of Color Clas-sification

2.8/3.7 means a color with a red hue of 5R, avalue of 2.8, and a chroma of 3.7. All colorsthat can be made with available pigments arelaid down as color chips in the Munsell bookof color.

Figure 6-9 The Munsell Color Space

Black

Whte

Purple

Black

Yellow

Green

White

Blue

Red

SaturationChroma

Lightn

ess

Value

HUNTER SYSTEM

The CIE system of color measurement isbased on the principle of color sensing by thehuman eye. This accepts that the eyes containthree light-sensitive receptors—the red, green,and blue receptors. One problem with thissystem is that the X, Y, and Z values have norelationship to color as perceived, though acolor is completely defined. To overcome thisproblem, other color systems have been sug-gested. One of these, widely used for foodcolorimetry, is the Hunter L, a, fo, system. Theso-called uniform-color, opponent-colors colorscales are based on the opponent-colorstheory of color vision. In this theory, it isassumed that there is an intermediate signal-switching stage between the light receptors inthe retina and the optic nerve, which trans-mits color signals to the brain. In this switch-ing mechanism, red responses are comparedwith green and result in a red-to-green colordimension. The green response is comparedwith blue to give a yellow-to-blue colordimension. These two color dimensions are

represented by the symbols a and b. The thirdcolor dimension is lightness L, which is non-linear and usually indicated as the square orcube root of K This system can be repre-sented by the color space shown in Figure6-10. The L, a, b, color solid is similar to theMunsell color space. The lightness scale iscommon to both. The chromatic spacing isdifferent. In the Munsell system, there are thepolar hue and chroma coordinates, whereas inthe L, a, b, color space, chromaticity isdefined by rectangular a and b coordinates.CIE values can be converted to color valuesby the equations shown in Table 6-2 into L, a,b, values and vice versa (MacKinney and Lit-tle 1962; Clydesdale and Francis 1970). Thisis not the case with Munsell values. These areobtained from visual comparison with colorchips (called Munsell renotations) or frominstrumental measurements (called Munsellrenotations), and conversion is difficult andtedious.

The Hunter tristimulus data, L (value), a(redness or greenness), and b (yellowness orblueness), can be converted to a single color

Figure 6-10 The Hunter L, a, b Color Space. Source: From Hunter Associates Lab., Inc.

L=O

BLACK

BLUE

RED

YELLOW

GRAYGREEN

L--100

WHITE

function called color difference (AE) byusing the following relationship:

AE = (AL)2 + (Afl)2 + (Ab)2

The color difference is a measure of the dis-tance in color space between two colors. Itdoes not indicate the direction in which thecolors differ.

LOVIBOND SYSTEM

The Lovibond system is widely used forthe determination of the color of vegetableoils. The method involves the visual compar-ison of light transmitted through a glass

cuvette filled with oil at one side of aninspection field; at the other side, coloredglass filters are placed between the lightsource and the observer. When the colors oneach side of the field are matched, the nomi-nal value of the filters is used to define thecolor of the oil. Four series of filters areused—red, yellow, blue, and gray filters. Thegray filters are used to compensate for inten-sity when measuring samples with intensechroma (color purity) and are used in thelight path going through the sample. The red,yellow, and blue filters of increasing inten-sity are placed in the light path until a matchwith the sample is obtained. Vegetable oilcolors are usually expressed in terms of red

Source: From Hunter Associates Lab., Inc.

Table 6-2 Mathematical Relationship Between Color Scales

To Convert To L, a, b To X%, Y, Z% ToY,x,y

From

From

From

and yellow; a typical example of the Lovi-bond color of an oil would be Rl.7 Y17. Thevisual determination of oil color by the Lovi-bond method is widely used in industry andis an official method of the American OilChemists' Society. Visual methods of thistype are subject to a number of errors, andthe results obtained are highly variable. Astudy has been reported (Maes et al., 1997)to calculate CIE and Lovibond color valuesof oils based on their visible light transmis-sion spectra as measured by a spectropho-tometer. A computer software has beendeveloped that can easily convert light trans-mission spectra into CIE and Lovibond colorindexes.

GLOSS

In addition to color, there is another impor-tant aspect of appearance, namely gloss.Gloss can be characterized as the reflectingproperty of a material. Reflection of light canbe diffused or undiffused (specular). In spec-ular reflection, the surface of the object actsas a mirror, and the light is reflected in ahighly directional manner. Surfaces can rangefrom a perfect mirror with completely specu-lar reflection to a surface reflecting in a com-pletely diffuse manner. In the latter, the lightfrom an incident beam is scattered in alldirections and the surface is called matte.

FOOD COLORANTS

The colors of foods are the result of naturalpigments or of added colorants. The naturalpigments are a group of substances present inanimal and vegetable products. The addedcolorants are regulated as food additives, butsome of the synthetic colors, especially ca-rotenoids, are considered "nature identical"

and therefore are not subject to stringent tox-icological evaluation as are other additives(Dziezak 1987).

The naturally occurring pigments embracethose already present in foods as well asthose that are formed on heating, storage, orprocessing. With few exceptions, these pig-ments can be divided into the following fourgroups:

1. tetrapyrrole compounds: chlorophylls,hemes, and bilins

2. isoprenoid derivatives: carotenoids3. benzopyran derivatives: anthocyanins

and flavonoids4. artefacts: melanoidins, caramels

The chlorophylls are characteristic of greenvegetables and leaves. The heme pigmentsare found in meat and fish. The carotenoidsare a large group of compounds that arewidely distributed in animal and vegetableproducts; they are found in fish and crusta-ceans, vegetables and fruits, eggs, dairyproducts, and cereals. Anthocyanins and fla-vonoids are found in root vegetables andfruits such as berries and grapes. Caramelsand melanoidins are found in syrups andcereal products, especially if these productshave been subjected to heat treatment.

Tetrapyrrole Pigments

The basic unit from which the tetrapyrrolepigments are derived is pyrrole.

The basic structure of the heme pigmentsconsists of four pyrrole units joined togetherinto a porphyrin ring as shown in Figure 6-11.

Figure 6-11 Schematic Representation of theHeme Complex of Myoglobin. M = methyl, P =propyl, V = vinyl. Source: From C.E. Bodwelland RE. McClain, Proteins, in The Sciences ofMeat Products, 2nd ed., I.E. Price and B.S. Sch-weigert, eds., 1971, W.H. Freeman & Co.

In the heme pigments, the nitrogen atoms arelinked to a central iron atom. The color ofmeat is the result of the presence of two pig-ments, myoglobin and hemoglobin. Bothpigments have globin as the protein portion,and the heme group is composed of the por-phyrin ring system and the central iron atom.In myoglobin, the protein portion has amolecular weight of about 17,000. In hemo-globin, this is about 67,000—equivalent tofour times the size of the myoglobin protein.The central iron in Figure 6-11 has six coor-dination bonds; each bond represents anelectron pair accepted by the iron from fivenitrogen atoms, four from the porphyrin ringand one from a histidyl residue of the globin.The sixth bond is available for joining withany atom that has an electron pair to donate.The ease with which an electron pair isdonated determines the nature of the bondformed and the color of the complex. Other

factors playing a role in color formation arethe oxidation state of the iron atom and thephysical state of the globin.

In fresh meat and in the presence of oxy-gen, there is a dynamic system of three pig-ments, oxymyoglobin, myoglobin, and met-myoglobin. The reversible reaction with oxy-gen is

Mb + O2 ̂ MbO2

In both pigments, the iron is in the ferrousform; upon oxidation to the ferric state, thecompound becomes metmyoglobin. Thebright red color of fresh meat is due to thepresence of oxymyoglobin; discoloration tobrown occurs in two stages, as follows:

MbO2 ^ Mb ^ MetMbRed Purplish red Brownish

Oxymyoglobin represents a ferrous covalentcomplex of myoglobin and oxygen. Theabsorption spectra of the three pigments areshown in Figure 6-12 (Bodwell and McClain1971). Myoglobin forms an ionic complexwith water in the absence of strong electronpair donors that can form covalent com-plexes. It shows a diffuse absorption band inthe green area of the spectrum at about 555nm and has a purple color. In metmyoglobin,the major absorption peak is shifted towardthe blue portion of the spectrum at about 505nm with a smaller peak at 627 nm. The com-pound appears brown.

As indicated above, oxymyoglobin andmyoglobin exist in a state of equilibriumwith oxygen; therefore, the ratio of the pig-ments is dependent on oxygen pressure. Theoxidized form of myoglobin, the metmyo-globin, cannot bind oxygen. In meat, there isa slow and continuous oxidation of the heme

Globin

pigments to the metmyoglobin state. Reduc-ing substances in the tissue reduce the met-myoglobin to the ferrous form. The oxygenpressure, which is so important for the stateof the equilibrium, is greatly affected bypackaging materials used for meats. Themaximum rate of conversion to metmyoglo-bin occurs at partial pressures of 1 to 20 nmof mercury, depending on pigment, pH, andtemperature (Fox 1966). When a packagingfilm with low oxygen permeability is used,the oxygen pressure drops to the point whereoxidation is favored. To prevent this, Lan-drock and Wallace (1955) established thatoxygen permeability of the packaging filmmust be at least 5 liters of oxygen/squaremeter/day/atm.

Fresh meat open to the air displays thebright red color of oxymyoglobin on the sur-face. In the interior, the myoglobin is in thereduced state and the meat has a dark purplecolor. As long as reducing substances arepresent in the meat, the myoglobin willremain in the reduced form; when they areused up, the brown color of metmyoglobinwill predominate. According to Solberg(1970), there is a thin layer a few nanome-ters below the bright red surface and justbefore the myoglobin region, where a defi-nite brown color is visible. This is the areawhere the oxygen partial pressure is about1.4 nm and the brown pigment dominates.The growth of bacteria at the meat surfacemay reduce the partial oxygen pressure to

Oxymyoglobin

Metmyoglolmi

Myoglobin

Wavelength (in/x)

Ext

inct

ion

coef

fici

ent

(cn

v/m

g^

Figure 6-12 Absorption Spectra of Myoglobin, Oxymyoglobin, and Metmyoglobin. Source: FromC.E. Bodwell and RE. McClain, Proteins, in The Sciences of Meat Products, 2nd ed., I.E. Price andB.S. Schweigert, eds., 1971, W.H. Freeman & Co.

below the critical level of 4 nm. Microor-ganisms entering the logarithmic growthphase may change the surface color to thatof the purplish-red myoglobin (Solberg1968).

In the presence of sulfhydryl as a reducingagent, myoglobin may form a green pigment,called sulfmyoglobin. The pigment is greenbecause of a strong absorption band in thered region of the spectrum at 616 nm. In thepresence of other reducing agents, such asascorbate, cholemyoglobin is formed. In thispigment, the porphyrin ring is oxidized. Theconversion into sulfmyoglobin is reversible;cholemyoglobin formation is irreversible,and this compound is rapidly oxidized toyield globin, iron, and tetrapyrrole. Accord-ing to Fox (1966), this reaction may happenin the pH range of 5 to 7.

Heating of meat results in the formation ofa number of pigments. The globin is dena-tured. In addition, the iron is oxidized to theferric state. The pigment of cooked meat isbrown and called hemichrome. In the pres-ence of reducing substances such as thosethat occur in the interior of cooked meat, theiron may be reduced to the ferrous form; theresulting pigment is pink hemochrome.

In the curing of meat, the heme reacts withnitrite of the curing mixture. The nitrite-heme complex is called nitrosomyoglobin,which has a red color but is not particularlystable. On heating the more stable nitrosohe-mochrome, the major cured meat pigment isformed, and the globin portion of the mole-cule is denatured. This requires a tempera-ture of 650C. This molecule has been callednitrosomyoglobin and nitrosylmyoglobin,but Mohler (1974) has pointed out that theonly correct name is nitric oxide myoglobin.The first reaction of nitrite with myoglobin isoxidation of the ferrous iron to the ferricform and formation of MetMb. At the same

time, nitrate is formed according to the fol-lowing reaction (Mohler 1974):

4MbO2 + 4NO2- + 2H2O -»4MetMbOH + 4NO3" + O2

During the formation of the curing pigment,the nitrite content is gradually lowered; thereare no definite theories to account for thisloss.

The reactions of the heme pigments inmeat and meat products have been summa-rized in the scheme presented in Figure 6-13(Fox 1966). Bilin-type structures are formedwhen the porphyrin ring system is broken.

Chlorophylls

The chlorophylls are green pigmentsresponsible for the color of leafy vegetablesand some fruits. In green leaves, the chloro-phyll is broken down during senescence andthe green color tends to disappear. In manyfruits, chlorophyll is present in the unripestate and gradually disappears as the yellowand red carotenoids take over during ripen-ing. In plants, chlorophyll is isolated in thechloroplastids. These are microscopic parti-cles consisting of even smaller units, calledgrana, which are usually less than one micro-meter in size and at the limit of resolution ofthe light microscope. The grana are highlystructured and contain laminae betweenwhich the chlorophyll molecules are posi-tioned.

The chlorophylls are tetrapyrrole pigmentsin which the porphyrin ring is in the dihydroform and the central metal atom is magne-sium. There are two chlorophylls, a and b,which occur together in a ratio of about 1:25.Chlorophyll b differs from chlorophyll a inthat the methyl group on carbon 3 is replacedwith an aldehyde group. The structural for-

mula of chlorophyll a is given in Figure 6-14. Chlorophyll is a diester of a dicarboxylicacid (chlorophyllin); one group is esterifiedwith methanol, the other with phytyl alcohol.The magnesium is removed very easily byacids, giving pheophytins a and b. The actionof acid is especially important for fruits thatare naturally high in acid. However, itappears that the chlorophyll in plant tissuesis bound to lipoproteins and is protectedfrom the effect of acid. Heating coagulatesthe protein and lowers the protective effect.The color of the pheophytins is olive-brown.Chlorophyll is stable in alkaline medium.The phytol chain confers insolubility inwater on the chlorophyll molecule. Uponhydrolysis of the phytol group, the water-sol-

uble methyl chlorophyllides are formed. Thisreaction can be catalyzed by the enzymechlorophyllase. In the presence of copper orzinc ions, it is possible to replace the magne-sium, and the resulting zinc or copper com-plexes are very stable. Removal of the phytolgroup and the magnesium results inpheophorbides. All of these reactions aresummarized in the scheme presented in Fig-ure 6-15.

In addition to those reactions describedabove, it appears that chlorophyll can bedegraded by yet another pathway. Chichesterand McFeeters (1971) reported on chloro-phyll degradation in frozen beans, whichthey related to fat peroxidation. In this reac-tion, lipoxidase may play a role, and no

Figure 6-13 Heme Pigment Reactions in Meat and Meat Products. ChMb, cholemyoglobin (oxidizedporphyrin ring); O2Mb, oxymyoglobin (Fe+2); MMb metmyoglobin (Fe+3); Mb, myoglobin (Fe+2);MMb-NO2, metmyoglobin nitrate; NOMMb, nitrosylmetmyoglobin; NOMb, nitrosylmyoglobin;NMMb, nitrimetmyoglobin; NMb, nitrimyoglobin, the latter two being reaction products of nitrous acidand the heme portion of the molecule; R, reductants; O, strong oxidizing conditions. Source: FromJ.B. Fox, The Chemistry of Meat Pigments, J. Agr. Food Chem., Vol. 14, no. 3, pp. 207-210, 1966,American Chemical Society.

Nitrosyl-hemochrome

Denatured GlobinHemichrome

HeminTetrapyrroles Nitrihemin

heat acid heat acidacid

BilePigment

FRESH CURED

pheophytins, chlorophyllides, or pheophor-bides are detected. The reaction requiresoxygen and is inhibited by antioxidants.

Carotenoids

The naturally occurring carotenoids, withthe exception of crocetin and bixin, are tet-raterpenoids. They have a basic structure ofeight isoprenoid residues arranged as if two20-carbon units, formed by head-to-tail con-densation of four isoprenoid units, hadjoined tail to tail. There are two possibleways of classifying the carotenoids. The firstsystem recognizes two main classes, the car-

otenes, which are hydrocarbons, and thexanthophylls, which contain oxygen in theform of hydroxyl, methoxyl, carboxyl, keto,or epoxy groups. The second system dividesthe carotenoids into three types (Figure 6-16),acyclic, monocyclic, and bicyclic. Examplesare lycopene (I)—acyclic, y-carotene (II)—monocyclic, and a-carotene (III) and p-car-otene (IV)—bicyclic.

The carotenoids take their name from themajor pigments of carrot (Daucus car old).The color is the result of the presence of asystem of conjugated double bonds. Thegreater the number of conjugated doublebonds present in the molecule, the further themajor absorption bands will be shifted to the

Figure 6-14 Structure of Chlorophyll a. (Chlorophyll b differs in having a formyl group at carbon 3).Source: Reprinted with permission from J.R. Whitaker, Principles of Enzymology for the Food Sciences,1972, by courtesy of Marcel Dekker, Inc.

region of longer wavelength; as a result, thehue will become more red. A minimum ofseven conjugated double bonds are requiredbefore a perceptible yellow color appears.Each double bond may occur in either cis ortrans configuration. The carotenoids in foodsare usually of the all-trans type and onlyoccasionally a mono-cis or di-cis compoundoccurs. The prefix neo- is used for stereoiso-mex.1 with at least one cis double bond. Theprefix pro- is for poly-a's carotenoids. Theeffect of the presence of cis double bonds onthe absorption spectrum of p-carotene isshown in Figure 6-17. The configuration hasan effect on color. The all-trans compoundshave the deepest color; increasing numbersof cis bonds result in gradual lightening ofthe color. Factors that cause change of bondsfrom trans to cis are light, heat, and acid.

In the narrower sense, the carotenoids arethe four compounds shown in Figure 6-16—a-, p-, and y-carotene and lycopene—poly-ene hydrocarbons of overall compositionC40H56. The relation between these and caro-tenoids with fewer than 40 carbon atoms isshown in Figure 6-18. The prefix apo- isused to designate a carotenoid that is derivedfrom another one by loss of a structural ele-ment through degradation. It has been sug-gested that some of these smaller carotenoidmolecules are formed in nature by oxidativedegradation of C40 carotenoids (Grob 1963).

Several examples of this possible relation-ship are found in nature. One of the bestknown is the formation of retinin and vita-min A from p-carotene (Figure 6-19).Another obvious relationship is that of lyco-pene and bixin (Figure 6-20). Bixin is a food

Figure 6-15 Reactions of Chlorophylls

chlorinpurpurins

alkaliO2

alkaliO2

acidalkali°2

pheophytinphytol

acidpheophorbide acid methyl

chlorophyllide

phytol

chlorophyllasechlorophyll

strongacid

acid

phytol

Figure 6-16 The Carotenoids: (I) Lycopene, (II)y-carotene, (III) a-Carotene, and (IV) p-Caro-tene. Source: From E.G. Grob, The Biogenesis ofCarotenes and Carotenoids, in Carotenes andCarotenoids, K. Lang, ed., 1963, Steinkopff Ver-lag.

color additive obtained from the seed coat ofthe fruit of a tropical brush, Bixa orellana.The pigment bixin is a dicarboxylic acidesterified with one methanol molecule. Apigment named crocin has been isolatedfrom saffron. Crocin is a glycoside contain-ing two molecules of gentiobiose. Whenthese are removed, the dicarboxylic acid cro-cetin is formed (Figure 6-21). It has thesame general structure as the aliphatic chainof the carotenes. Also obtained from saffronis the bitter compound picrocrocin. It is aglycoside and, after removal of the glucose,yields saffronal. It is possible to imagine acombination of two molecules of picrocrocinand one of crocin; this would yield protocro-cin. Protocrocin, which is directly related tozeaxanthin, has been found in saffron (Grob1963).

The structure of a number of importantxanthophylls as they relate to the structure ofP-carotene is given in Figure 6-22. Carot-enoids may occur in foods as relatively sim-ple mixtures of only a few compounds or asvery complex mixtures of large numbers ofcarotenoids. The simplest mixtures usuallyexist in animal products because the animalorganism has a limited ability to absorb anddeposit carotenoids. Some of the most com-plex mixtures are found in citrus fruits.

Beta-carotene as determined in fruits andvegetables is used as a measure of the provi-tamin A content of foods. The column chro-matographic procedure, which determinesthis content, does not separate cc-carotene, p-carotene, and cryptoxanthin. Provitamin Avalues of some foods are given in Table 6-3.Carotenoids are not synthesized by animals,but they may change ingested carotenoidsinto animal carotenoids—as in, for example,salmon, eggs, and crustaceans. Usually carot-enoid content of foods does not exceed 0.1percent on a dry weight basis.

In ripening fruit, carotenoids increase atthe same time chlorophylls decrease. Theratio of carotenes to xanthophylls also in-creases. Common carotenoids in fruits are oc-and y-carotene and lycopene. Fruit xantho-phylls are usually present in esterified form.Oxygen, but not light, is required for caro-tenoid synthesis and the temperature range iscritical. The relative amounts of differentcarotenoids are related to the characteristiccolor of some fruits. In the sequence ofpeach, apricot, and tomato, there is an in-creasing proportion of lycopene and increas-ing redness. Many peach varieties are devoidof lycopene. Apricots may have about 10percent and tomatoes up to 90 percent. Thelycopene content of tomatoes increases dur-ing ripening. As the chlorophyll breaks downduring ripening, large amounts of carot-

enoids are formed (Table 6-4). Color is animportant attribute of citrus juice and isaffected by variety, maturity, and processingmethods. The carotenoid content of orangesis used as a measure of total color. Curl andBailey (1956) showed that the 5,6-epoxidesof fresh orange juice isomerize completely to5,8-epoxides during storage of canned juice.This change amounts to the loss of one dou-ble bond from the conjugated double bondsystem and causes a shift in the wavelengthof maximum absorption as well as a decrease

in molar absorbance. In one year's storage at7O0F, an apparent carotenoid loss of 20 to 30percent occurs.

Peaches contain violaxanthin, cryptoxan-thin, p-carotene, and persicaxanthin as wellas 25 other carotenoids, including neoxan-thin. Apricots contain mainly p- and ycaro-tene, lycopene, and little if any xanthophyll.Carrots have been found to have an averageof 54 ppm of total carotene (Borenstein andBunnell 1967), consisting mainly of a-, p,and ^-carotene and some lycopene and xan-

Figure 6-17 Absorption Spectra of the Three Stereoisomers of Beta Carotene. B = neo-p-carotene; U =neo-p-carotene-U; T = all-trans-p-carotene. a, b, c, and d indicate the location of the mercury arc lines334.1, 404.7, 435.8 and 491.6 nm, respectively. Source: From F Stitt et al., Spectrophotometric Deter-mination of Beta Carotene Stereoisomers in Alfalfa, /. Assoc. Off. Agric. Chem. Vol. 34, pp. 460-471,1951.

Wavelength ( n m)

Abs

orpt

ivity

(l

/g-c

m)

Figure 6-18 Relationship Between the Caroteneand Carotenoids with Fewer than 40 Carbons

thophyll. Canning of carrots resulted in a 7 to12 percent loss of provitamin A activitybecause of cis-trans isomerization of a- andp-carotene (Weckel et al. 1962). In dehy-

drated carrots, carotene oxidation and off-flavor development have been correlated(Falconer et al. 1964). Corn contains aboutone-third of the total carotenoids as car-otenes and two-thirds xanthophylls. Com-pounds found in corn include zeaxanthin,cryptoxanthin, p-carotene, and lutein.

One of the highest known concentrationsof carotenoids occurs in crude palm oil. Itcontains about 15 to 300 times more retinolequivalent than carrots, green leafy vegeta-bles, and tomatoes. All of the carotenoids incrude palm oil are destroyed by the normalprocessing and refining operations. Recently,improved gentler processes have been devel-oped that result in a "red palm oil" thatretains most of the carotenoids. The compo-sition of the carotenes in crude palm oil witha total carotene concentration of 673 mg/kgis shown in Table 6-5.

Milkfat contains carotenoids with sea-sonal variation (related to feed conditions)ranging from 2 to 13 ppm.

Figure 6-19 Formation of Retinin and Vitamin A from p-Carotene. Source: From B.C. Grob, The Bio-genesis of Carotenes and Carotenoids, in Carotenes and Carotenoids, K. Lang, ed., 1963, SteinkopffVerlag.

Vilomin A

Retinin

ft-Corottn*

AZAFRIN IONONE

13C27C

PlCROCROClN CROCIN PICROCROCIN

1OC2OC1OC

BIXINMETHYLHEPTENONE8C24C

METHYLHEPTENONE8C

VITAMIN A

2OC

VITAMINA

2OC

CAROTENES

4OC

lycopene

Bixin

Figure 6-20 Relationship Between Lycopene and Bixin. Source: From E.G. Grob, The Biogenesis ofCarotenes and Carotenoids, in Carotenes and Carotenoids, K. Lang, ed., 1963, Steinkopff Verlag.

Zcaxanthin

Protocrocin

PicrocrocinCf OC in Picrocrocin

Glucose

Gentiobiote Genttobiose

Gluco»e

Crocetin

S of fronal SaHronol

Figure 6-21 Relationship Between Crocin and Picrocrocin and the Carotenoids. Source: From B.C.Grob, The Biogenesis of Carotenes and Carotenoids, in Carotenes and Carotenoids, K. Lang, ed., 1963,Steinkopff Verlag.

Figure 6-22 Structure of Some of the Important Carotenoids. Source: From B. Borenstein and R.H.Bunnell, Carotenoids: Properties, Occurrence, and Utilization in Foods, in Advances in Food Research,Vol. 15, C.O. Chichester et al., eds., 1967, Academic Press.

Astaxanthin

Torularhodin

Canthaxanthin

Physalien

Zeaxanthin

Isozeaxanthin

Lutein

Cryptoxanthin

p-Apo-8'-carotenal

p-Carotene

Capsorubin

Capsanthin

Table 6-3 Provitamin A Value of Some Fruits andVegetables

Product IU/100g

Carrots, mature 20,000Carrots, young 10,000Spinach 13,000Sweet potato 6,000Broccoli 3,500Apricots 2,000Lettuce 2,000Tomato 1,200Asparagus 1,000Bean, trench 1,000Cabbage 500Peach 800Brussels sprouts 700Watermelon 550Banana 400Orange juice 200

Source: From B. Borenstein and R.H. Bunnell, Caro-tenoids: Properties, Occurrence, and Utilization inFoods, in Advances in Food Research, Vol. 15, C.O.Chichester et al., eds., 1967, Academic Press.

Egg yolk contains lutein, zeaxanthin, andcryptoxanthin. The total carotenoid contentranges from 3 to 89 ppm.

Crustaceans contain carotenoids bound toprotein resulting in a blue or blue-gray color.When the animal is immersed in boilingwater, the carotenoid-protein bond is brokenand the orange-red color of the free car-

otenoid appears. Widely distributed in crus-taceans is astaxanthin. Red fish containastaxanthin, lutein, and taraxanthin.

Common unit operations of food process-ing are reported to have only minor effectson the carotenoids (Borenstein and Bunnell1967). The carotenoid-protein complexes aregenerally more stable than the free car-otenoids. Because carotenoids are highly un-saturated, oxygen and light are major factorsin their breakdown. Blanching destroysenzymes that cause carotenoid destruction.Carotenoids in frozen or heat-sterilized foodsare quite stable. The stability of carotenoidsin dehydrated foods is poor, unless the foodis packaged in inert gas. A notable exceptionis dried apricots, which keep their color well.Dehydrated carrots fade rapidly.

Several of the carotenoids are now com-mercially synthesized and used as food col-ors. A possible method of synthesis isdescribed by Borenstein and Bunnell (1967).Beta-ionone is obtained from lemon grass oiland converted into a C14 aldehyde. The C14aldehyde is changed to a C16 aldehyde, thento a C19 aldehyde. Two moles of the C19aldehyde are condensed with acetylene di-magnesium bromide and, after a series ofreactions, yield p-carotene.

Three synthetically produced carotenoidsare used as food colorants, p-carotene, p-apo-8'-carotenal (apocarotenal), and can-thaxanthin. Because of their high tinctorialpower, they are used at levels of 1 to 25 ppm

Pigment

LycopeneCaroteneXanthophyllXanthophyll ester

Green(mg/100g)

0.110.160.02O

Half-ripe(mg/100g)

0.840.430.030.02

Ripe(mg/100g)

7.850.730.060.10

Table 6-4 Development of Pigments in the Ripening Tomato

in foods (Dziezak 1987). They are unstablein light but otherwise exhibit good stabilityin food applications. Although they are fatsoluble, water-dispersible forms have beendeveloped for use in a variety of foods. Beta-carotene imparts a light yellow to orangecolor, apocarotenal a light orange to reddish-orange, and canthaxanthin, orange-red tored. The application of these compounds in avariety of foods has been described by Coun-sell (1985). Natural carotenoid food colorsare annatto, oleoresin of paprika, and unre-fined palm oil.

Anthocyanins and Flavonoids

The anthocyanin pigments are present in

the sap of plant cells; they take the form of

glycosides and are responsible for the red,

blue, and violet colors of many fruits and

vegetables. When the sugar moiety is re-

Table 6-5 Composition of the Carotenes in

Crude Palm Oil

% of TotalCarotene Carotenes

Phytoene 1.27Cis-p-carotene 0.68Phytofluene 0.06p-carotene 56.02cc-carotene 35.06^-carotene 0.69y-carotene 0.336-carotene 0.83Neurosporene 0.29p-zeacarotene 0.74oc-zeacarotene 0.23Lycopene 1.30

Source: Reprinted with permission from Choo YuenMay, Carotenoids from Palm Oil, Palm Oil Develop-ments, Vol. 22, pp. 1-6, Palm Oil Research Institute ofMalaysia.

moved by hydrolysis, the aglucone remainsand is called anthocyanidin. The sugar partusually consists of one or two molecules ofglucose, galactose, and rhamnose. The basicstructure consists of 2-phenyl-benzopyry-lium or flavylium with a number of hydroxyand methoxy substituents. Most of the antho-cyanidins are derived from 3,5,7-trihydroxy-flavylium chloride (Figure 6-23} and thesugar moiety is usually attached to thehydroxyl group on carbon 3. The anthocya-nins are highly colored, and their names arederived from those of flowers. The structureof some of the more important anthocyani-dins is shown in Figure 6-24, and the occur-rence of anthocyanidins in some fruits andvegetables is listed in Table 6-6. Recentstudies have indicated that some anthocya-nins contain additional components such asorganic acids and metals (Fe, Al, Mg).

Substitution of hydroxyl and methoxylgroups influences the color of the anthocya-nins. This effect has been shown by Braver-man (1963) (Figure 6-25). Increase in thenumber of hydroxyl groups tends to deepenthe color to a more bluish shade. Increase inthe number of methoxyl groups increasesredness. The anthocyanins can occur in dif-ferent forms. In solution, there is an equilib-rium between the colored cation R+ oroxonium salt and the colorless pseudobaseROH, which is dependent on pH.

R+ + H2O ̂ ROH + H+

As the pH is raised, more pseudobase isformed and the color becomes weaker. How-ever, in addition to pH, other factors influ-ence the color of anthocyanins, includingmetal chelation and combination with otherflavonoids and tannins.

Anthocyanidins are highly colored instrongly acid medium. They have two ab-sorption maxima—one in the visible spec-

trum at 500-550 nm, which is responsiblefor the color, and a second in the ultraviolet(UV) spectrum at 280 nm. The absorptionmaxima relate to color. For example, therelationship in 0.01 percent HCl in methanolis as follows: at 520 nm pelargonidin is scar-let, at 535 nm cyanidin is crimson, and at 546nm delphinidin is blue-mauve (Macheix etal. 1990).

About 16 anthocyanidins have been identi-fied in natural products, but only the followingsix of these occur frequently and in many dif-ferent products: pelargonidin, cyanidin, del-phinidin, peonidin, malvidin, and petunidin.The anthocyanin pigments of Red Deliciousapples were found to contain mostly cyanidin-3-galactoside, cyanidin-3-arabinoside, and cya-nidin-7-arabinoside (Sun and Francis 1968).Bing cherries contain primarily cyanidin-3-rutinoside, cyanidin-3-glucoside, and smallamounts of the pigments cyanidin, peonidin,peonidin-3-glucoside, and peonidin-3-ruti-

noside (Lynn and Luh 1964). Cranberryanthocyanins were identified as cyanidin-3-monogalactoside, peonidin-3-monogalacto-side, cyanidin monoarabinoside, and peonidin-3-monoarabinoside (Zapsalis and Francis1965). Cabernet Sauvignon grapes containfour major anthocyanins: delphinidin-3-monoglucoside, petunidin-3-monoglucoside,malvidin-3-monoglucoside, and malvidin-3-monoglucoside acetylated with chlorogenicacid. One of the major pigments is petunidin(Somaatmadja and Powers 1963).

Anthocyanin pigments can easily bedestroyed when fruits and vegetables are pro-cessed. High temperature, increased sugarlevel, pH, and ascorbic acid can affect therate of destruction (Daravingas and Cain1965). These authors studied the change inanthocyanin pigments during the processingand storage of raspberries. During storage,the absorption maximum of the pigmentsshifted, indicating a change in color. The

Figure 6-23 Chemical Structure of Fruit Anthocyanidins

R, = H

R1 =OH

R1 = OH

R, = OCH3

R1 = OCH,

R1 = OCH3

R2 = H

R2 = H

R2 = OH

R2=H

R2 = OH

R2 = OCH3

PELARGONIDIN

CYANIDIN

DELPHINIDIN

PEONIDIN

PETUNIDIN

MALVIDIN

Table 6-6 Anthocyanidins Occurring in SomeFruits and Vegetables

Fruit or Vegetable Anthocyanidin

Apple CyanidinBlack currant Cyanidin and delphinidinBlueberry Cyanidin, delphinidin, mal-

vidin, petunidin, andpeonidin

Cabbage (red) CyanidinCherry Cyanidin and peonidinGrape Malvidin, peonidin, delphini-

din, cyanidin, petunidin,and pelargonidin

Orange Cyanidin and delphinidinPeach CyanidinPlum Cyanidin and peonidinRadish PelargonidinRaspberry CyanidinStrawberry Pelargonidin and a little

cyanidin

Source: From P. Markakis, Anthocyanins, in Encyclo-pedia of Food Technology, A.M. Johnson and M.S.Peterson, eds., 1974, AVI Publishing Co.

level of pigments was lowered by prolongedtimes and higher temperatures of storage.Higher concentration of the ingoing sugarsyrup and the presence of oxygen resulted ingreater pigment destruction.

The stability of anthocyanins is increasedby acylation (Dougall et al. 1997). Theseacylated anthocyanins may occur naturallyas in the case of an anthocyanin from thepurple yam (Yoshida et al. 1991). This antho-cyanin has one sinapic residue attachedthrough a disaccharide and was found to bestable at pH 6.0 compared to other anthocya-nins without acylation. Dougall et al. (1977)were able to produce stable anthocyanins byacylation of carrot anthocyanins in cell cul-tures. They found that a wide range of aro-matic acids could be incorporated into theanthocyanin.

Anthocyanins can form purplish or slate-gray pigments with metals, which are calledlakes. This can happen when canned foodstake up tin from the container. Anthocyanins

Figure 6-24 Structure of Some Important Anthocyanidins

Delphinidin Peonidin

Cyanidin

can be bleached by sulfur dioxide. Accord-ing to Jurd (1964), this is a reversible processthat does not involve hydrolysis of the glyco-sidic linkage, reduction of the pigment, oraddition of bisulfite to a ketonic, chalconederivative. The reactive species was found tobe the anthocyanin carbonium ion (R+),which reacts with a bisulfite ion to form acolorless chromen-2(or 4)-sulfonic acid(R-SO3H), similar in structure and proper-ties to an anthocyanin carbinol base (R-OH).This reaction is shown in Figure 6-26.

The colors of the anthocyanins at acid pHvalues correspond to those of the oxoniumsalts. In slightly alkaline solutions (pH 8 to10), highly colored ionized anhydro basesare formed. At pH 12, these hydrolyze rap-idly to fully ionized chalcones (Figure 6-27).Leuco bases are the reduced form of theanthocyanins. They are usually withoutmuch color but are widely distributed infruits and vegetables. Under the influence ofoxygen and acid hydrolysis, they maydevelop the characteristic color of the car-

Figure 6-25 Effect of Substituents on the Color of Anthocyanidins. Source: Reprinted with permissionfrom J.B.S. Braverman, Introduction to the Biochemistry of Foods, © 1963, Elsevier Publishing Co.

pelargonidin cyanidin delphinidin

petunidin

malvidin

shade of blue

shade of red

Figure 6-26 Reaction of Bisulfite with the Anthocyanin Carbonium Ion

bonium ion. Canned pears, for example, mayshow "pinking"—a change from the leucobase to the anthocyanin.

The flavonoids or anthoxanthins are glyco-sides with a benzopyrone nucleus. The fla-vones have a double bond between carbons 2and 3. The flavonols have an additionalhydroyxl group at carbon 3, and the fla-vanones are saturated at carbons 2 and 3(Figure 6-28). The flavonoids have low col-

oring power but may be involved in discolor-ations; for example, they can impart blue andgreen colors when combined with iron.Some of these compounds are also potentialsubstrates for enzymic browning and cancause undesirable discoloration through thismechanism. The most ubiquitous flavonoidis quercetin, a 3,5,7,3',4'-pentahydroxy fla-vone (Figure 6-29). Many flavonoids containthe sugar rutinose, a disaccharide of glucose

Figure 6-27 Structure of Anhydro Base (I) and Chalcone (II)

I I I

(1) flavones (positions 2:3 unsaturated)

(2) flavonols (an additional OH at position 3)

(3) flavanones (saturated at positions 2:3)

(4) flavanonols (position 3 saturated and extrahydroxyl group)

(5) isoflavones (phenol ring B at position 3)

Figure 6-28 Structure of Flavones, Flavonals, Flavanones, Flavanonols, and Isoflavones

Figure 6-29 Structure of Quercetin

and rhamnose. Hesperidin is a flavanoneoccurring in citrus fruits and, at pH 12, theinner ring opens to form a chalcone in a sim-ilar way as shown for the anthocyanins. Thechalcones are yellow to brown in color.

Tannins

Tannins are polyphenolic compounds pres-ent in many fruits. They are important ascolor compounds and also for their effect ontaste as a factor in astringency (see Chapter7). Tannins can be divided into two classes—hydrolyzable tannins and nonhydrolyzable orcondensed tannins. The tannins are charac-terized by the presence of a large number ofhydroxyl groups, which provide the ability toform reversible bonds with other macromol-ecules, polysaccharides, and proteins, as wellas other substances such as alkaloids. Thisbond formation may occur during the devel-opment of the fruit or during the mechanicaldamage that takes place during processing.

Hydrolyzable tannins are composed ofphenolic acids and sugars that can be brokendown by acid, alkaline, or enzymic hydroly-

sis. They are polyesters based on gallic acidand/or hexahydroxydiphenic acid (Figure6-30). The usual sugar is D-glucose andmolecular weights are in the range of 500 to2,800. Gallotannins release gallic acid onhydrolysis, and ellagitannins produce ellagicacid. Ellagic acid is the lactone form ofhexahydroxydiphenic acid, which is thecompound originally present in the tannin(Figure 6-30).

Nonhydrolyzable or condensed tanninsare also named proanthocyanidins. These arepolymers of flavan-3-ols, with the flavanbonds most commonly between C4 and C8or C6 (Figure 6-23) (Macheix et al. 1990).Many plants contain tannins that are poly-mers of (-f)-catechin or (-)-epicatechin.These are hydrogenated forms of flavonoidsor anthocyanidins. Other monomers occupy-ing places in condensed fruit tannins havetrihydroxylation in the B-ring: (+)-gallocat-echin and (-)-epigallocatechin. Oligomericand polymeric procyanidins are formed byaddition of more flavan-3-ol units and resultin the formation of helical structures. Thesestructures can form bonds with proteins.

Tannins are present in the skins of redgrapes and play an important part in the fla-vor profile of red wine. Tannins in grapes areusually estimated in terms of the content ofgallic acid (Amerine and Joslyn 1970).

Oxidation and polymerization of phenoliccompounds as a result of enzymic activity ofphenoloxidases or peroxidases may result in

ELLAGIC ACIDHEXAHYDROXYDIPHENIC ACIDGALLIC ACID

Figure 6-30 Structure of Components of Hydrolyzable Tannins

the formation of brown pigments. This cantake place during the growth of fruits (e.g., indates) or during mechanical damage in pro-cessing.

Betalains

Table beets are a good source of red pig-ments; these have been increasingly used forfood coloring. The red and yellow pigments

obtained from beets are known as betalainsand consist of the red betacyanins and theyellow betaxanthins (Von Elbe and Maing1973). The structures of the betacyanins areshown in Figure 6-31. The major betacyaninis betanin, which accounts for 75 to 95 per-cent of the total pigments of beets. Theremaining pigments contain isobetanin, pre-betanin, and isoprebetanin. The latter two aresulfate monoesters of betanin and isobetanin,

III BETANINIV ISOBETANIN, C-15

EPIMER OF BETANIN

I BETANIDINII ISOBETANIDIN, C-15

EPIMER OF BETANIDIN

V VULGAXANTHIN-I Vl VULGAXANTHIN-II

Figure 6-31 Structure of Naturally Occurring Betalains in Red Beets. Source: From J.H. Von Elbe andL-Y. Maing, Betalains as Possible Food Colorants of Meat Substitutes, Cereal ScL Today, Vol. 18, pp.263-264,316-317, 1973.

respectively. The major yellow pigments arevulgaxanthin I and vulgaxanthin II. Betaninis the glucoside of betanidin, and isobetaninis the C-15 epimer of betanin.

Betanidin has three carboxyl groups (pka =3.4), two phenol groups (pHa = 8.5), andasymmetric carbons at positions 2 and 15.The 15-position is easily isomerized underacid or basic conditions in the absence ofoxygen to yield isobetanidin. Under alkalineconditions and in the presence of glutamineor glutamic acid, betanin can be converted tovulgaxanthin (Mabry 1970).

The color of betanin solutions is influ-enced by pH. In the range of 3.5 to 7.0, thespectrum shows a maximum of 537 nm (Fig-ure 6-32). Below pH 3, the intensity of thismaximum decreases and a slight increase inthe region of 570 to 640 nm occurs and thecolor shifts toward violet. At pH values over

7, a shift of the maximum occurs to longerwavelength. At pH 9, the maximum is about544 nm and the color shifts toward blue. VonElbe et al. (1974) found that the color of bet-anin is most stable between pH 4.0 and 6.0.The thermostability is greatest between pH4.0 and 5.0. Light and air have a degradingeffect on betanin, and the effect is cumula-tive.

Caramel

Caramel color can be produced from a vari-ety of carbohydrate sources, but usually cornsugar syrup is used. Corn starch is first hydro-lyzed with acid to a DE of 8 to 9, followed byhydrolysis with bacterial oc-amylase to a DEof 12 to 14, then with fungal amyloglucosi-dase up to a DE of 90 to 95. Several types ofcaramel are produced. The largest amount is

Figure 6-32 Visible Spectra of Betanin at pH Values of 2.0, 5.0, and 9.0. Source: From J.H. Von Elbe,L-Y. Maing, and C.H. Amundson, Color Stability of Betanin, Journal of Food Science, Vol. 39, pp. 334-337, 1974, Institute of Food Technologists.

A ( nm )

electropositive or positive caramel, which ismade with ammonia. Electronegative or neg-ative caramel is made with ammonium salts.A slightly electronegative caramel is solublein alcohol and is used for coloring beverages(Greenshields 1973). The composition andcoloring power of caramel depends on thetype of raw materials and the process used.Both Maillard-type reactions and pure cara-melizing reactions are thought to be involved,and the commercial product is extremelycomplex in composition. Caramels containhigh and low molecular weight colored com-pounds, as well as a variety of volatile com-ponents.

Other Colorants

Synthetic colorants, used commercially,are also known as certified color additives.There are two types, FD&C dyes and FD&Clakes. FD&C indicates substances approvedfor use in food, drug, and cosmetic use byU.S. federal regulations. Dyes are water-solu-ble compounds that produce color in solu-tion. They are manufactured in the form ofpowders, granules, pastes, and dispersions.They are used in foods at concentrations ofless than 300 ppm (Institute of Food Technol-ogists 1986). Lakes are made by combiningdyes with alumina to form insoluble colo-rants, which have dye contents in the range of20 to 25 percent (Pearce 1985). The lakesproduce color in dispersion and can be usedin oil-based foods when insufficient water ispresent for the solubilization of the dye. Thelist of approved water-soluble colorants haschanged frequently; the current list is givenin Chapter 11.

The uncertified color additives (Institute ofFood Technologists 1986) include a numberof natural extracts as well as inorganic sub-stances such as titanium dioxide. Some ofthese can be used only with certain restric-tions (Table 6-7). The consumer demand formore natural colorants has provided an impe-tus for examining many natural coloring sub-stances. These have been described in detailby Francis (1987). The possibility of usingplant tissue culture for the production of nat-ural pigments has also been considered (Ilker1987).

Table 6-7 Color Additives Not RequiringCertification

Colorant Restriction

Annatto extract —Beta-apo-8'-carotenal 33 mg/kgBeta-carotene —Beet powder —Canthaxanthin 66 mg/kgCaramel —Carrot oil —Cochineal extract —

(carmine)Ferrous gluconate Ripe olives onlyFruit juice —Grape color extract Nonbeverage

foods onlyGrape skin extract Beverages

(enocianina)Paprika and its oleoresin —Riboflavin —Saffron —Titanium dioxide 1 %Turmeric and its oleo- —

resinVegetable juice —

REFERENCES

Amerine, M.A., and M.A. Joslyn. 1970. Table wines.The technology of their production. Berkeley, CA:University of California Press.

Bodwell, C.E., and RE. McClain. 1971. Proteins. InThe sciences of meat products, ed. J. Price and B. S.Schweigert. San Francisco: W.H. Freeman and Co.

Borenstein, B., and R.H. Bunnell. 1967. Carotenoids:Properties, occurrence and utilization in foods. InAdvances in food research, Vol. 15, ed. C.O. Chich-ester, E.M. Mrak, and G.F. Stewart. New York: Aca-demic Press.

Braverman, J.B.S. 1963. Introduction to the biochemis-try of foods. New York: Elsevier Publishing Co.

Chichester, C.O., and R. McFeeters. 1971. Pigmentdegeneration during processing and storage. In Thebiochemistry of fruits and their products, ed. A.C.Hulme. New York: Academic Press.

Clydesdale, F.M., and FJ. Francis. 1970. Color scales.Food Prod. Dev. 3: 117-125.

Counsell, J.N. 1985. Uses of carotenoids in foods.IFST Proceedings 18: 156-162.

Curl, A.L., and G.F. Bailey. 1956. Carotenoids of agedcanned Valencia orange juice. J. Agr. Food Chem. 4:159-162.

Daravingas, G., and R.F. Cain. 1965. Changes in theanthocyanin pigments of raspberries during process-ing and storage. /. Food ScL 30: 400-405.

Dougall, D.K., et al. 1997. Biosynthesis and stability ofmonoacylated anthocyanins. Food TechnoL 51, no.11:69-71.

Dziezak, J.D. 1987. Applications of food colorants.Food TechnoL 41, no. 4: 78-88.

Falconer, M.E., et al. 1964. Carotene oxidation and off-flavor development in dehydrated carrot. J. ScL FoodAgr. 15: 897-901.

Fox, J.B. 1966. The chemistry of meat pigments. /.Agr. Food Chem. 14: 207-210.

Francis, FJ. 1987. Lesser known food colorants. FoodTechnoL 41, no. 4: 62-68.

Greenshields, R.N. 1973. Caramel—Part 2. Manufac-ture, composition and properties. Process Biochem.8, no. 4: 17-20.

Grob, E.C. 1963. The biogenesis of carotenes and caro-tenoids. In Carotenes and carotenoids, ed. K. Lang.Darmstadt, Germany: Steinkopff Verlag.

Ilker, R. 1987. In-vitro pigment production: An alter-native to color synthesis. Food TechnoL 41, no. 4:70-72.

Institute of Food Technologists. 1986. Food colors:Scientific status summary. Food TechnoL 40, no. 7:49-56.

Jurd, L. 1964. Reactions involved in sulfite bleachingof anthocyanins. J. Food ScL 29: 16-19.

Landrock, A.H., and G.A. Wallace. 1955. Discolora-tion of fresh red meat and its relationship to filmoxygen permeability. Food TechnoL 9: 194-196.

Lynn, D.Y.C., and B.S. Luh. 1964. Anthocyanin pig-ments in Bing cherries. J. Food ScL 29: 735-743.

Mabry, TJ. 1970. Betalains, red-violet and yellowalkaloids of Centrospermae. In Chemistry of theAlkaloids, ed. S.W. Pelletier. New York: Van Nos-trand Reinhold Co.

Macheix, JJ., et al. 1990. Fruit phenolics. Boca Raton,FL: CRC Press.

MacKinney, G., and A.C. Little. 1962. Color of foods.Westport, CT: AVI Publishing Co.

Maes, PJ.A., et al. 1997. Converting spectra into colorindices. Inform 8: 1245-1252.

Mohler, K. 1974. Formation of curing pigments bychemical, biochemical or enzymatic reactions. InProceedings of the International Symposium onNitrite in Meat Products. Wageningen, The Nether-lands: Center for Agricultural Publishing and Docu-mentation.

Pearce, A. 1985. Current synthetic food colors. IFSTProceedings 18: 147-155.

Solberg, M. 1968. Factors affecting fresh meat color.Proc. Meat Ind. Research Conference, Chicago.March 21,22.

Solberg, M. 1970. The chemistry of color stability inmeat: A review. Can. Inst. Food TechnoL J. 3: 55-62.

Somaatmadja, D., and JJ. Powers. 1963. Anthocya-nins, IV: Anthocyanin pigments of Cabernet Sauvi-gnon grapes. / Food ScL 28: 617-622.

Sun, B.H., and FJ. Francis. 1968. Apple anthocyanins:Identification of cyanidin-7-arabinoside. J. Food ScL32: 647-649.

Von Elbe, J.H., and L-Y. Maing. 1973. Betalains aspossible food colorants of meat substitutes. CerealScL Today 18: 263-264, 316-317.

Von Elbe, J.H., L-Y. Maing, and C.H. Amundson.1974. Color stability of betanin. J. Food ScL 39:334-337.

Weckel, K.G., et al. 1962. Carotene components of fro-zen and processed carrots. Food Technol. 16, no. 8:91-94.

Yoshida, K., et al. 1991. Unusually stable monoacy-lated anthocyanin from purple yam Dioscorea alata.Tetrahedron Lett. 32: 5579-5580.

Zapsalis, C., and FJ. Francis. 1965. Cranberry antho-cyanins. J. Food ScL 30: 396-399.

INTRODUCTION

Flavor has been defined by Hall (1968) asfollows: "Flavor is the sensation produced bya material taken in the mouth, perceivedprincipally by the senses of taste and smell,and also by the general pain, tactile and tem-perature receptors in the mouth. Flavor alsodenotes the sum of the characteristics of thematerial which produce that sensation."

This definition makes clear that flavor is aproperty of a material (a food) as well as ofthe receptor mechanism of the personingesting the food. The study of flavorincludes the composition of food com-pounds having taste or smell, as well as theinteraction of these compounds with thereceptors in the taste and smell sensoryorgans. Following an interaction, the organsproduce signals that are carried to the cen-tral nervous system, thus creating what weunderstand as flavor. This process is proba-bly less well understood than the processesoccurring in other organs (O'Mahony1984). Beidler (1957) has represented thetaste process schematically (Figure 7-1).

Although flavor is composed mainly oftaste and odor, other qualities contribute tothe overall sensation. Texture has a verydefinite effect. Smoothness, roughness,granularity, and viscosity can all influence

Figure 7-1 Schematic Representation of theTaste Process. Source: From LM. Beidler, Factsand Theory on the Mechanism of Taste and OdorPerception, in Chemistry of Natural Food Fla-vors, 1957, Quartermaster Food and ContainerInstitute for the Armed Forces.

flavor, as can hotness of spices, coolness ofmenthol, brothiness or fullness of certainamino acids, and the tastes described asmetallic and alkaline.

TASTE

It is generally agreed that there are onlyfour basic, or true, tastes: sweet, bitter, sour,

Flavor

CHAPTER 7

TASTE SENSATIONS

BRAIN

NEURAL PATTERNS OF ACTIVITY

TACTILE TONGUE PAIN

WARM

TASTE

COLD

and salty. The sensitivity to taste is located intaste buds of the tongue. The taste buds aregrouped in papillae, which appear to be sen-sitive to more than one taste. There isundoubtedly a regional distribution of thefour kinds of receptors at the tongue, creat-ing areas of sensitivity—the sweet taste atthe tip of the tongue, bitter at the back, sourat the edges, and salty at both edges and tip(Figure 7-2). The question of how the fourtypes of receptors are able to respond thisspecifically has not been resolved. Accord-ing to Teranishi et al. (1971), perception ofthe basic taste qualities results from a patternof nerve activity coming from many tastecells; specific receptors for sweet, sour, bit-ter, and salty do not exist. It may be envi-sioned that a single taste cell possessesmultiple receptor sites, each of which mayhave specificity.

The mechanism of the interaction betweenthe taste substance and the taste receptor isnot well understood. It has been suggested

that the taste compounds interact with spe-cific proteins in the receptor cells. Sweet-and bitter-sensitive proteins have beenreported. Dastoli and Price (1966) isolated aprotein from bovine tongue epithelium thatshowed the properties of a sweet taste recep-tor molecule. Dastoli et al. (1968) reportedisolating a protein that had the properties of abitter receptor.

We know that binding between stimulusand receptor is a weak one because no irre-versible effects have been observed. A mech-anism of taste stimulation with electrolyteshas been proposed by Beidler (1957); it isshown in Figure 7-3. The time required fortaste response to take place is in the order of25 milliseconds. The taste molecule isweakly adsorbed, thereby creating a distur-bance in the molecular geography of the sur-face and allowing an interchange of ionsacross the surface. This reaction is followedby an electrical depolarization that initiates anerve impulse.

The taste receptor mechanism has beenmore fully described by Kurihara (1987). Theprocess from chemical stimulation to trans-mitter release is schematically presented inFigure 7-4. The receptor membranes containvoltage-dependent calcium channels. Tastecompounds contact the taste cells and depo-larize the receptor membrane; this depolar-ization spreads to the synaptic area, activatingthe voltage-dependent calcium channels.Influx of calcium triggers the release of thetransmitter norepinephrine.

The relationship between stimulus concen-tration and neural response is not a simpleone. As the stimulus concentration increases,the response increases at a decreasing rateuntil a point is reached where further in-crease in stimulus concentration does notproduce a further increase in response.Beidler (1954) proposed the following equa-

B I T T E R

Figure 7-2 Areas of Taste Sensitivity of theTongue

S W E E T

Figure 7-4 Diagram of a Taste Cell and the Mechanism of Chemical Stimulation and TransmitterRelease. Source: Reprinted with permission from Y. Kawamura and M.R. Kare, Umami: A Basic Tale,© 1987, Marcel Dekker, Inc.

Release of transmitter(norepinephrine)

Tastenerve

Ca influx

Activation ofvoltage-dependent Ca channel

Receptor potential

Adsorption

Receptormembrane

Electriccurrent

Synapse

Figure 7-3 Mechanism of Taste Stimulation as Proposed by Beidler. Source: From L.M. Beidler, Factsand Theory on the Mechanism of Taste and Odor Perception, in Chemistry of Natural Food Flavors,1957, Quartermaster Food and Container Institute for the Armed Forces.

NERVE ACTION POTENTIALS

SENSE CELL DEPOLARIZATION

CELLULAR CHANGESSTRUCTURAL

CHEMICAL

PHYSICOCHEMICAL CHANGESSPATIAL ARRANGEMENTS

CHARGE DENSITIES

BINDING SITESPROTEINS

LIPIDS

HYDRATEDIONS

tion relating magnitude of response and stim-ulus concentration:

£ - — —R " Ws

+ KR s

whereC = stimulus concentrationR = response magnitudeRs = maximum responseK = equilibrium constant for the stimulus-

receptor reaction

K values reported by Beidler for many sub-stances are in the range of 5 to 15.

It appears that the initial step in the stimu-lus-receptor reaction is the formation of aweak complex, as evidenced by the smallvalues of K. The complex formation resultsin the initiation of the nerve impulse. Tasteresponses are relatively insensitive to changesin pH and temperature. Because of thedecreasing rate of response, we know that thenumber of receptor sites is finite. The tasteresponse is a function of the proportion ofsites occupied by the stimulus compound.

According to Beidler (1957), the thresholdvalue of a substance depends on the equilib-rium constant and the maximum response.Since K and Rx both vary from one substanceto another and from one species to another,the threshold also varies between substancesand species. The concentration of the stimu-lus can be increased in steps just large enoughto elicit an increase in response. This amountis called the just noticeable difference (JND).

There appear to be no significant age- orsex-related differences in taste sensitivity(Fisher 1971), but heavy smoking (more than20 cigarettes per day) results in a deteriora-tion in taste responsiveness with age.

Differences in taste perception betweenindividuals seem to be common. Peryam

(1963) found that sweet and salt are usuallywell recognized. However, with sour and bit-ter taste some difficulty is experienced.Some tasters ascribe a bitter quality to citricacid and a sour quality to caffeine.

Chemical Structure and Taste

A first requirement for a substance to pro-duce a taste is that it be water soluble. Therelationship between the chemical structureof a compound and its taste is more easilyestablished than that between structure andsmell. In general, all acid substances aresour. Sodium chloride and other salts aresalty, but as constituent atoms get bigger, abitter taste develops. Potassium bromide isboth salty and bitter, and potassium iodide ispredominantly bitter. Sweetness is a propertyof sugars and related compounds but also oflead acetate, beryllium salts, and many othersubstances such as the artificial sweetenerssaccharin and cyclamate. Bitterness is exhib-ited by alkaloids such as quinine, picric acid,and heavy metal salts.

Minor changes in chemical structure maychange the taste of a compound from sweetto bitter or tasteless. For example, Beidler(1966) has examined saccharin and its sub-stitution compounds. Saccharin is 500 timessweeter than sugar (Figure 7-5). Introduc-tion of a methyl group or of chloride in thepara position reduces the sweetness by half.Placing a nitro group in the meta positionmakes the compound very bitter. Introduc-tion of an amino group in the para positionretains the sweetness. Substitutions at theimino group by methyl, ethyl, or bromoethylgroups all result in tasteless compounds.However, introduction of sodium at this loca-tion yields sodium saccharin, which is verysweet.

The compound 5-nitro-otoluidine is sweet.The positional isomers 3-nitro-o-toluidineand 3-nitro-p-toluidine are both tasteless(Figure 7-6). Teranishi et al. (1971) pro-vided another example of change in tasteresulting from the position of substituentgroup: 2-amino-4-nitro-propoxybenzene is4,000 times sweeter than sugar, 2-nitro-4-amino-propoxybenzene is tasteless, and 2,4-dinitro-propoxybenzene is bitter (Figure 7-7).Dulcin (p-ethoxyphenylurea) is extremelysweet, the thiourea analog is bitter, and the0-ethoxyphenylurea is tasteless (Figure 7-8).

Just as positional isomers affect taste, sodo different stereoisomers. There are eightamino acids that are practically tasteless. Agroup of three has varying tastes; except forglutamic acid, these are probably derivedfrom sulfur-containing decomposition prod-

ucts. Seven amino acids have a bitter taste inthe L form or a sweet taste in the D form,except for L-alanine, which has a sweet taste(Table 7-1). Solms et al. (1965) reported onthe taste intensity, especially of aromaticamino acids. L-tryptophan is about half asbitter as caffeine; D-tryptophan is 35 timessweeter than sucrose and 1.7 times sweeterthan calcium cyclamate. L-phenylalanine isabout one-fourth as bitter as caffeine; the Dform is about seven times sweeter thansucrose. L-tyrosine is about one-twentieth asbitter as caffeine, but D-tyrosine is still 5.5times sweeter than sucrose.

Some researchers claim that differencesexist between the L and D forms of some sug-ars. They propose that L-glucose is slightlysalty and not sweet, whereas D-glucose issweet. There is even a difference in taste

Figure 7-5 The Effect of Substitutions in Saccharin on Sweetness. Source: From L.M. Beidler, Chem-ical Excitation of Taste and Odor Receptors, in Flavor Chemistry, I. Hornstein, ed., 1966, AmericanChemistry Society.

Sweet Tasteless Tasteless Tasteless Sweet

Sweet Sweet Sweet Bitter Sweet

Figure 7-6 Taste of Nitrotoluidine Isomers

SWEET TASTELESS TASTELESS

between the two anomers of D-mannose. Thea form is sweet as sugar, and the (3 form is bit-ter as quinine.

Optical isomers of carvone have totallydifferent flavors. The D+ form is characteris-tic of caraway; the L- form is characteristicof spearmint.

The ability to taste certain substances isgenetically determined and has been studiedwith phenylthiourea. At low concentrations,about 25 percent of subjects tested do nottaste this compound; for the other 75 percent,the taste is bitter. The inability to taste phen-ylthiourea is probably due to a recessivegene. The compounds by which tasters andnontasters can be differentiated all containthe following isothiocyanate group:

SIi

- C - N -

These compounds—phenylthiourea, thio-urea, and thiouracil—are illustrated in Figure

7-9. The corresponding compounds thatcontain the group,

OIl

- C - N -

phenylurea, urea, and uracil, do not showthis phenomenon. Another compound con-taining the isothiocyanate group has beenfound in many species of the Cruciferae fam-ily; this family includes cabbage, turnips,and rapeseed and is well known for itsgoitrogenic effect. The compound is goitrin,5-vinyloxazolidine-2-thione (Figure 7-10).

Sweet Taste

Many investigators have attempted to relatethe chemical structure of sweet tasting com-pounds to the taste effect, and a series of theo-ries have been proposed (Shallenberger 1971).Shallenberger and Acree (1967, 1969) pro-

TASTELESSBITTERSWEET

Figure 7-8 Taste of Substituted Ethoxybenzenes

BITTERTASTELESSSWEETFigure 7-7 Taste of Substituted Propoxybenzenes

posed a theory that can be considered a refine-ment of some of the ideas incorporated inprevious theories. According to this theory,called the AH,B theory, all compounds thatbring about a sweet taste response possess anelectronegative atom A, such as oxygen ornitrogen. This atom also possesses a protonattached to it by a single covalent bond; there-fore, AH can represent a hydroxyl group, animine or amine group, or a methine group.

Within a distance of about 0.3 nm from theAH proton, there must be a second electrone-gative atom B, which again can be oxygen ornitrogen (Figure 7-11). Investigators haverecognized that sugars that occur in a favoredchair conformation yield a glycol unit confor-mation with the proton of one hydroxyl groupat a distance of about 0.3 nm from the oxygenof the next hydroxyl group; this unit can beconsidered as an AH,B system. It was alsofound that the K bonding cloud of the benzenering could serve as a B moiety. This explainsthe sweetness of benzyl alcohol and thesweetness of the anti isomer of anisaldehydeoxime, as well as the lack of sweetness of thesyn isomer. The structure of these compoundsis given in Figure 7-12. The AH,B systempresent in sweet compounds is, according toShallenberger, able to react with a similarAH,B unit that exists at the taste bud receptorsite through the formation of simultaneoushydrogen bonds. The relatively strong natureof such bonds could explain why the sense ofsweetness is a lingering sensation. Accordingto the AH,B theory, there should not be a dif-ference in sweetness between the L and D iso-mers of sugars. Experiments by Shallenberger(1971) indicated that a panel could not distin-guish among the sweet taste of the enantio-morphic forms of glucose, galactose, man-nose, arabinose, xylose, rhamnose, and gluco-heptulose. This suggests that the notion that Lsugars are tasteless is a myth.

Phenylthiourea Thiourea Thiouracil

Figure 7-9 Compounds Containing theDifferentiated

Group by Which Tasters and Nontasters Can Be

SIl

- C - N -

Table 7-1 Difference in Taste Between the L-and D-Forms of Amino Acids

Amino Acid

AsparagineGlutamic

acidPhenylala-

nineLeucine

Valine

Serine

Histidine

lsoleucineMethionineTryptophane

Taste ofL lsomer

InsipidUnique

Faintly bitter

Flat, faintlybitter

Slightlysweet,bitter

Faintly sweet,stale after-taste

Tasteless tobitter

BitterFlatBitter

Taste ofD lsomer

SweetAlmost taste-

lessSweet, bitter

aftertasteStrikingly

sweetStrikingly

sweet

Strikinglysweet

Sweet

SweetSweetVery sweet

Figure 7-10 5-Vinyloxazolidine-2-thione

Spillane (1996) has pointed out that theAH,B theory appears to work quite well,although spatial, hydrophobic/hydrophilic,and electronic effects are also important.Shallenberger (1998) describes the initiationof sweetness as being due to a concertedintermolecular, antiparallel hydrogen-bond-ing interaction between the glycophore(Greek glyks, sweet; phoros, to carry) andreceptor dipoles. The difficulty in explainingthe sweetness of compounds with differentchemical structures is also covered by Shal-lenberger (1998) and how this has resulted inalternative taste theories. The application ofsweetness theory is shown to have importantapplications in the food industry.

Extensive experiments with a large num-ber of sugars by Birch and Lee (1971) sup-port Shallenberger's theory of sweetnessand indicate that the fourth hydroxyl groupof glucopyranosides is of unique impor-tance in determining sweetness, possibly bydonating the proton as the AH group. Ap-

parently the primary alcohol group is of lit-tle importance for sweetness. Substitutionof acetyl or azide groups confers intensebitterness to sugars, whereas substitution ofbenzoyl groups causes tastelessness.

As the molecular weight of saccharidesincreases, their sweetness decreases. This isbest explained by the decrease in solubilityand increase in size of the molecule. Appar-ently, only one sugar residue in each oli-gosaccharide is involved in the interaction atthe taste bud receptor site.

The relative sweetness of a number of sug-ars and other sweeteners has been reportedby Solms (1971) and is given in Table 7-2.These figures apply to compounds tasted sin-gly and do not necessarily apply to sugars infoods, except in a general sense. The relativesweetness of mixtures of sugars changeswith the concentration of the components.Synergistic effects may increase the sweet-ness by as much as 20 to 30 percent in suchmixtures (Stone and Oliver 1969).

Sour Taste

Although it is generally recognized thatsour taste is a property of the hydrogen ion,there is no simple relationship between sour-ness and acid concentration. Acids have dif-ferent tastes; the sourness as experienced inthe mouth may depend on the nature of theacid group, pH, titratable acidity, buffering

S W E E T

COMPOUND

RECEPTOR

S I T E

Figure 7-11 The AH,B Theory of Sweet Taste Perception

effects and the presence of other compounds,especially sugars. Organic acids have agreater taste effect than inorganic acids (suchas hydrochloric acid) at the same pH. Infor-mation on a number of the most commonacids found in foods and phosphoric acid(which is also used in soft drinks) has beencollected by Solms (1971) and comparedwith hydrochloric acid. This information ispresented in Table 7-3.

According to Beatty and Cragg (1935), rel-ative sourness in unbuffered solutions ofacids is not a function of molarity but is pro-portional to the amount of phosphate bufferrequired to bring the pH to 4.4. Ough (1963)determined relative sourness of four organicacids added to wine and also preference forthese acids. Citric acid was judged the mostsour, fumaric and tartaric about equal, andadipic least sour. The tastes of citric and tar-taric acids were preferred over those offumaric and adipic acids.

Pangborn (1963) determined the relativesourness of lactic, tartaric, acetic, and citricacid and found no relation between pH, totalacidity, and relative sourness. It was alsofound that there may be considerable differ-ences in taste effects between sugars andacids when they are tested in aqueous solu-tions and in actual food products.

Table 7-2 Relative Sweetness of Sugars andOther Sweeteners

Compound Relative Sweetness

Sucrose 1

Lactose 0.27

Maltose 0.5

Sorbitol 0.5

Galactose 0.6

Glucose 0.5-0.7

Mannitol 0.7

Glycerol 0.8

Fructose 1.1-1.5

Cyclamate 30-80

Glycyrrhizin 50

Aspartyl-phenylalanine 100-200

methylester

Stevioside 300

Naringin dihydrochal- 300

cone

Saccharin 500-700

Neohesperidin 1000-1500dihydrochalcone

Source: From J. Solms, Nonvolatile Compounds andthe Flavor of Foods, in Gustation and Olfaction, G.Ohloff and A.F. Thomas, eds., 1971, Academic Press.

SWEET TASTELESS

Figure 7-12 Anfr'-Anisaldehyde Oxime, Sweet; and Syrc-Anisaldehyde Oxime, Tasteless

Buffering action appears to help determinethe sourness of various acids; this mayexplain why weak organic acids taste moresour than mineral acids of the same pH. It issuggested that the buffering capacity of salivamay play a role, and foods contain many sub-stances that could have a buffering capacity.

Wucherpfennig (1969) examined the sourtaste in wine and found that alcohol maydecrease the sourness of organic acids. Heexamined the relative sourness of 17 organicacids and found that the acids tasted at thesame level of undissociated acid have greatlydifferent intensities of sourness. Partiallyneutralized acids taste more sour than pureacids containing the same amount of undis-sociated acids. The change of malic into lac-tic acid during the malolactic fermentation ofwines leads to a decrease in sourness, thusmaking the flavor of the wine milder.

Salty Taste

The salty taste is best exhibited by sodiumchloride. It is sometimes claimed that thetaste of salt by itself is unpleasant and thatthe main purpose of salt as a food componentis to act as a flavor enhancer or flavor poten-tiator. The taste of salts depends on thenature of both cation and anion. As themolecular weight of either cation or anion—or both—increases, salts are likely to tastebitter. The lead and beryllium salts of aceticacid have a sweet taste. The taste of a num-ber of salts is presented in Table 7-4.

The current trend of reducing sodiumintake in the diet has resulted in the formula-tion of low-sodium or reduced-sodium foods.It has been shown (Gillette 1985) thatsodium chloride enhances mouthfeel, sweet-ness, balance, and saltiness, and also masks

Table 7-3 Properties of Some Acids, Arranged in Order of Decreasing Acid Taste and with Tartaric Acidas Reference

Properties ofO.OSN Solutions

Acid

HydrochloricTartaricMalic

PhosphoricAceticLacticCitric

Propionic

Taste

+1.43O-0.43

-1.14-1.14-1.14-1.28

-1.85

TotalAcidg/L

1.853.753.35

1.653.004.503.50

3.70

pH

1.702.452.65

2.252.952.602.60

2.90

lonizationConstant

1.04 x 10~3

3.9X10"4

7.52 x 1Q-3

1.75 x 10~5

1.26 x 1Q-4

8.4 x 1Q-4

1.34 x 10~5

TasteSensation

HardGreen

IntenseVinegarSour, tartFresh

Sour,cheesy

Found In

GrapeApple, pear, prune,

grape, cherry,apricot

Orange, grapefruit

Berries, citrus,pineapple

Source: From J. Solms, Nonvolatile Compounds and the Flavor of Foods, in Gustation and Olfaction, G. Ohloffand A.F. Thomas, eds., 1971, Academic Press.

or decreases off-notes. Salt substitutes basedon potassium chloride do not enhancemouthfeel or balance and increase bitter ormetallic off-notes.

Bitter Taste

Bitter taste is characteristic of many foodsand can be attributed to a great variety ofinorganic and organic compounds. Manysubstances of plant origin are bitter. Al-though bitter taste by itself is usually consid-ered to be unpleasant, it is a component ofthe taste of many foods, usually those foodsthat are sweet or sour. Inorganic salts canhave a bitter taste (Table 7-4). Some aminoacids may be bitter (Table 7-1). Bitter pep-tides may be formed during the partial enzy-mic hydrolysis of proteins—for example,during the ripening of cheese. Solms (1969)has given a list of peptides with differenttaste sensations (Table 7-5).

The compounds best known for their bit-ter taste belong to the alkaloids and glyco-sides. Alkaloids are basic nitrogen-containingorganic compounds that are derived frompyridine, pyrrolidine, quinoline, isoquino-line, or purine. Quinine is often used as astandard for testing bitterness (Figure 7-13).

The bitterness of quinine hydrochloride isdetectable in a solution as dilute as 0.00004molar, or 0.0016 percent. If 5 mL of thissolution is tasted, the amount of substance aperson detects would be 0.08 mg (Moncri-eff 1951). Our sensitivity to bitterness ismore extreme than our sensitivity to othertastes; the order of sensitivity is from bitterto sour to salty and our least sensitivity is tosweet taste. Threshold values reported byMoncrieff are as follows: sour—0.007 per-cent HCl; salt—0.25 percent NaCl; andsweet—0.5 percent sucrose. If the artificialsweeteners such as saccharine are consid-ered, the sweet sensitivity is second to bit-ter. Quinine is used as a component of somesoft drinks to produce bitterness. Otheralkaloids occurring as natural bitter constit-uents of foods are caffeine and theobromine(Figure 7-14), which are derivatives ofpurine. Another naturally occurring bittersubstance is the glycoside naringin, whichoccurs in grapefruit and some other citrusfruits. Naringin in pure form is more bitterthan quinine and can be detected in concen-

Table 7-4 Taste Sensations of Salts

Taste

Salty

Salty and bitterBitterSweet

1 Extremely toxic

Salts

LiCI, LiBr, LiI, NaNO3, NaCI,NaBr, NaI, KNO3, KCIKBr, NH4ICsCI, CsBr, Kl, MgSO4

Lead acetate,1 berylliumacetate1

Table 7-5 Taste of Some Selected Peptides

Taste

Flat

Sour

Bitter

Sweet

Biting

Composition of Peptides

L-Lys-L-Glu, L-PhE-L-Phe, GIy-GIy-GIy-GIyL-Ala-L-Asp, y-L-Glu-L-Glu, GIy-L-Asp-L-Ser-GlyL-Leu-L-Leu, L-Arg-L-Pro, L-VaI-L-VaI-L-VaIL-Asp-L-Phe-OMe, L-Asp-L-Met-OMey-L-Glutamyl-S-(prop-1 -enyl)-L-cystein

Source: From J. Solms, Nonvolatile Compounds andthe Flavor of Foods, in Gustation and Ol faction, G.Ohloff and A.F. Thomas, eds., 1971, Academic Press.

trations of less than 0.002 percent. Naringin(Figure 7-15) contains the sugar moietyrutinose (L-rhamnose-D-glucose), whichcan be removed by hydrolysis with boilingmineral acid. The aglucose is called narin-genin, and it lacks the bitterness of narin-gin. Since naringin is only slightly solublein water (0.05 percent at 2O0C), it may crys-tallize out when grapefruit is subjected tobelow-freezing temperatures. Hesperidin(Figure 7-15) occurs widely in citrus fruitsand is also a rutinose glycoside. It occurs inoranges and lemons. Dried orange peel maycontain as much as 8 percent hesperidin.The aglycone of hesperidin is called hes-peretin. The sugar moiety is attached to car-bon 7. Horowitz and Gentili (1969) havestudied the relationship between bitternessand the structure of 7-rhamnoglycosides ofcitrus fruits; they found that the structure ofthe disaccharide moiety plays an importantrole in bitterness. The point of attachmentof rhamnose to glucose determines whetherthe substance will be bitter or tasteless.Thus, neohesperidin contains the disaccha-ride neohesporidose, which contains rham-

nose linked l->2 to glucose; therefore, thesugar moiety is 2-0-oc-L-rhamnopyranosyl-D-glucose. Glycosides containing this sugar,including neohesperidin, have a bitter taste.When the linkage between rhamnose andglucose is 1—>6, the compound is tastelessas in hesperidin, where the sugar part, ruti-nose, is 6-O-a-L-rhamnopyranosyl-D-glu-cose.

Bitterness occurs as a defect in dairy prod-ucts as a result of casein proteolysis byenzymes that produce bitter peptides. Bitterpeptides are produced in cheese because of anundesirable pattern of hydrolysis of milkcasein (Habibi-Najafi and Lee 1996). Ac-cording to Ney (1979), bitterness in aminoacids and peptides is related to hydrophobic-ity. Each amino acid has a hydrophobicityvalue (Af), which is defined as the free energyof transfer of the side chains and is based onsolubility properties (Table 7-6). The averagehydrophobicity of a peptide, Q, is obtained asthe sum of the Af of component amino acidsdivided by the number of amino acid resi-dues. Ney (1979) reported that bitterness isfound only in peptides with molecular weights

A B

Figure 7-14 (A) Caffeine and (B) Theobromine

Figure 7-13 Structure of Quinine. This has an intensely bitter taste.

below 6,000 Da when their Q value is greaterthan 1,400. These findings indicate the im-portance of molecular weight and hydropho-bicity. In a more detailed study of the compo-sition of bitter peptides, Kanehisa (1984)reported that at least six amino acids arerequired for strong bitterness. A bitter peptiderequires the presence of a basic amino acid atthe N-terminal position and a hydrophobicone at the C-terminal position. It appears thatat least two hydrophobic amino acids arerequired in the C-terminal area of the peptideto produce intense bitterness. The high hydro-phobicity of leucine and the number of leu-

Table 7-6 Hydrophoblcity Values (Af) of the SideChains of Amino Acids

Abbrevia- Af (cal/Amino Acid tion mol)

Glycine GIy OSerine Ser 40Threonine Thr 440Histidine His 500Aspartic acid Asp 540Glutamic acid GIu 550Arginine Arg 730Alanine Ala 730Methionine Met 1,300Lysine Lys 1,500Valine VaI 1,690Leucine Leu 2,420Proline Pro 2,620Phenylalanine Phe 2,650Tyrosine Tyr 2,870lsoleucine lie 2,970Tryptophan Trp 3,000

Source: Reprinted with permission from K.H. Ney,Bitterness of Peptides: Amino Acid Composition andChain Length, in Food Taste Chemistry, J.C. Boudreau,ed., ACS Symp. Ser. 115, © 1979, American ChemicalSociety.

cine and possibly proline residues in the pep-tide probably play a role in the bitterness.

Other Aspects of Taste

The basic sensations—sweet, sour, salty,and bitter—account for the major part of thetaste response. However, it is generallyagreed that these basic tastes alone cannotcompletely describe taste. In addition to thefour individual tastes, there are importantinterrelationships among them. One of themost important in foods is the interrelation-ship between sweet and sour. The sugar-acid

Figure 7-15 (A) Naringin; (B) Hesperidin; (C)Rutinose, 6-0-oc-L-Rhamnopyranosyl-D-Glu-copyranose

C

Br ut i nose

Arutinose

ratio plays an important part in many foods,especially fruits. Kushman and Ballinger(1968) have demonstrated the change insugar-acid ratio in ripening blueberries (Table7-7). Sugar-acid ratios play an importantrole in the flavor quality of fruit juices andwines (Ough 1963). Alkaline taste has beenattributed to the hydroxyl ion. Caustic com-pounds can be detected in solutions contain-ing only 0.01 percent of the alkali. Probablythe major effect of alkali is irritation of thegeneral nerve endings in the mouth. Anothereffect that is difficult to describe is astrin-gency. Borax is known for its ability to pro-duce this effect, as are the tannins present infoods, especially those that occur in tea.Even if astringency is not considered a partof the taste sense, it must still be considereda feature of food flavor.

Another important taste sensation is cool-ness, which is a characteristic of menthol.The cooling effect of menthol is part of themint flavor complex and is exhibited by onlysome of the possible isomeric forms. Only (-)and (+) menthol show the cooling effect, theformer to a higher degree than the latter, but

Table 7-7 Change in Sugar-Acid Ratio DuringRipening of Blueberries*

Unripe Ripe Overripe

Total sugar (%) 5̂ 8 7J9 12̂ 4pH 2.83 3.91 3.76Titr acidity 23.9 12.9 7.5(mEq/100g)Sugar-acid ratio 3.8 9.5 25.8

*The sugars are mainly glucose and fructose, andthe acidity is expressed as citric acid.

Source: From LJ. Kushman and W.E. Ballinger,Acid and Sugar Changes During Ripening in WolcottBlueberries, Proc. Amer. Soc. Hort. Soc., Vol. 92, pp.290-295,1968.

the isomers isomenthol, neomenthol, andneoisomenthol do not give a cooling effect(Figure 7-16) (Kulka 1967). Hotness is aproperty associated with spices and is alsoreferred to as pungency. The compound pri-marily responsible for the hotness of blackpepper is pipeline (Figure 7-17). In red pep-per or capsicum, nonvolatile amides areresponsible for the heat effect. The heat effectof spices and their constituents can be mea-sured by an organoleptic threshold method(Rogers 1966) and expressed in heat units.The pungent principle of capsicum is capsai-cin. The structure of capsaicin is given in Fig-ure 7-18. Capsaicin shows similarity to thecompound zingerone, the pungent principleof ginger (Figure 7-19).

Govindarajan (1979) has described therelationship between pungency and chemicalstructure of pungent compounds. There arethree groups of natural pungent compounds—the capsaicinoids, piperine, and the gin-gerols. These have some common structuralaspects, including an aromatic ring and analkyl side chain with a carbonyl function(Figures 7-18 and 7-19). Structural varia-tions in these compounds affect the intensityof the pungent response. These structuralvariations include the length of the alkyl sidechain, the position of the amide group nearthe polar aromatic end, the nature of thegroupings at the alkyl end, and the unsatura-tion of the alkyl chain.

The metallic taste has been described byMoncrieff (1964). There are no receptor sitesfor this taste or for the alkaline and meatytastes. However, according to Moncrieff,there is no doubt that the metallic taste is areal one. It is observable over a wide area ofthe surface of the tongue and mouth and, likeirritation and pain, appears to be a modalityof the common chemical sense. The metallictaste can be generated by salts of metals such

Figure 7-16 Isomeric Forms of Menthol

as mercury and silver (which are mostpotent) but normally by salts of iron, copper,and tin. The threshold concentration is in theorder of 20 to 30 ppm of the metal ion. Incanned foods, considerable metal uptakemay occur and the threshold could be

exceeded in such cases. Moncrieff (1964)also mentions the possibility of metallic ionexchange between the food and the con-tainer. The threshold concentration of copperis increased by salt, sugar, citric acid, andalcohol. Tannin, on the other hand, lowersthe threshold value and makes the coppertaste more noticeable. The metallic taste isfrequently observed as an aftertaste. The leadsalt of saccharin gives an impression ofintense sweetness, followed by a metallicaftertaste. Interestingly, the metallic taste isfrequently associated with oxidized prod-ucts. Tressler and Joslyn (1954) indicate that20 ppm of copper is detectable by taste inorange juice. Copper is well known for itsability to catalyze oxidation reactions. Starkand Forss (1962) have isolated and identifiedoct-l-en-3-one as the compound responsiblefor the metallic flavor in dairy products.

Taste Inhibition and Modification

Some substances have the ability to modifyour perception of taste qualities. Two suchcompounds are gymnemagenin, which isable to suppress the ability to taste sweet-ness, and the protein from miracle fruit,which changes the perception of sour tosweet. Both compounds are obtained fromtropical plants.

The leaves of the tropical plant Gymnemasylvestre, when chewed, suppress the abilityto taste sweetness. The effect lasts for hours,and sugar seems like sand in the mouth. Theability to taste other sweeteners such as sac-charin is equally suppressed. There is also adecrease in the ability to taste bitterness. Theactive principle of leaves has been namedgymnemic acid and has been found (Stocklinet al. 1967) to consist of four components,designated as gymnemic acids, A1, A2, A3,and A4. These are D-glucuronides of acety-

(db)-Neomenthol (± )-Neoisomenthol

(rt)-Menthol (dr)-Isomenthol

Figure 7-17 Piperine, Responsible for the Hot-ness of Pepper

lated gymnemagenins. The unacetylatedgymnemagenin is a hexahydroxy pentacy-clic triterpene; its structure is given in Figure7-20.

The berries of a West African shrub (Syn-sepalum dulcificum) contain a substance thathas the ability to make sour substances tastesweet. The berry, also known as miraclefruit, has been shown to contain a taste-mod-ifying protein (Kurihara and Beidler 1968;1969). The protein is a basic glycoproteinwith a molecular weight of 44,000. It is sug-gested that the protein binds to the receptormembrane near the sweet receptor site. Thelow pH changes the conformation of themembrane so that the sugar part of the pro-tein fits into the sweet receptor site. Thetaste-modifying protein was found to con-tain 6.7 percent of arabinose and xylose.

These taste-modifying substances providean insight into the mechanism of the produc-tion of taste sensations and, therefore, are avaluable tool in the study of the interrelation-ship between taste and chemical structure.

Flavor Enhancement—Umami

A number of compounds have the abilityto enhance or improve the flavor of foods. Ithas often been suggested that these com-pounds do not have a particular taste of theirown. Evidence now suggests that there is abasic taste response to amino acids, espe-cially glutamic acid. This taste is sometimesdescribed by the word umami, derived fromthe Japanese for deliciousness (Kawamuraand Kare 1987). It is suggested that a pri-mary taste has the following characteristics:

• The receptor site for a primary tastechemical is different from those of otherprimary tastes.

• The taste quality is different from others.• The taste cannot be reproduced by a

mixture of chemicals of different pri-mary tastes.

From these criteria, we can deduce that theglutamic acid taste is a primary taste for thefollowing reasons:

• The receptor for glutamic acid is differ-ent from the receptors for sweet, sour,salty, and bitter.

• Glutamic acid does not affect the taste ofthe four primary tastes.

• The taste quality of glutamic acid is dif-ferent from that of the four primarytastes.

Figure 7-18 Capsaicin, the Pungent Principle of Red Pepper

Figure 7-19 Zingerone, the Pungent Principle ofGinger

• Umami cannot be reproduced by mixingany of the four primary tastes.

Monosodium glutamate has long been rec-ognized as a flavor enhancer and is nowbeing considered a primary taste, umami.The flavor potentiation capacity of monoso-dium glutamate in foods is not the result ofan intensifying effect of the four primarytastes. Glutamate may exist in the L and Dforms and as a racemic mixture. The L formis the naturally occurring isomer that has aflavor-enhancing property. The D form isinert. Although glutamic acid was first iso-lated in 1866, the flavor-enhancing proper-ties of the sodium salt were not discovereduntil 1909 by the Japanese chemist Ikeda.Almost immediately,' commercial produc-tion of the compound started and total pro-duction for the year 1954 was estimated at13,000,000 pounds. The product as firstdescribed by Ikeda was made by neutralizinga hydrolysate of the seaweed Laminariajaponica with soda. Monosodium glutamateis now produced from wheat gluten, beetsugar waste, and soy protein and is used inthe form of the pure crystallized compound.It can also be used in the form of protein

hydrolysates derived from proteins that con-tain 16 percent or more of glutamic acid.Wheat gluten, casein, and soy flour are goodsources of glutamic acid and are used to pro-duce protein hydrolysates. The glutamic acidcontent of some proteins is listed in Table7-8 (Hall 1948). The protein is hydrolyzedwith hydrochloric acid, and the neutralizedhydrolysate is used in liquid form or as a drypowder. Soy sauce, which is similar to thesehydrolysates, is produced wholly or partiallyby enzymic hydrolysis. This results in theformation of ammonia from acid amides; soysauce contains ammonium complexes ofamino acids, including ammonium gluta-mate.

The flavor of glutamate is difficult todescribe. It has sometimes been suggestedthat glutamate has a meaty or chickeny taste,but it is now generally agreed that glutamateflavor is unique and has no similarity tomeat. Pure sodium glutamate is detectable inconcentrations as low as 0.03 percent; at 0.05percent the taste is very strong and does notincrease at higher concentrations. The tastehas been described (Crocker 1948) as a mix-ture of the four tastes. At about 2 thresholdvalues of glutamate concentration, it could

Figure 7-20 Structure of Gymnemagenin

Table 7-8 Glutamic Acid Content of SomeProteins

Glutamic AcidProtein Source (%)

Wheat gluten 36.0Corn gluten 24.5Zein 36.0Peanut flour 19.5Cottonseed flour 17.6Soybean flour 21.0Casein 22.0Rice 24.1Egg albumin 16.0Yeast 18.5

Source: From L.A. Hall, Protein Hydrolysates as aSource of Glutamate Flavors, in Monosodium Glut-amate—A Symposium, 1948, Quartermaster Food andContainer Institute for the Armed Forces.

be well matched by a solution containing 0.6threshold of sweet, 0.7 of salty, 0.3 of sour,and 0.9 of bitter. In addition, glutamate issaid to cause a tingling feeling and a markedpersistency of taste sensation. This feeling ispresent in the whole of the mouth and pro-vides a feeling of satisfaction or fullness.Apparently glutamate stimulates our tactilesense as well as our taste receptors. The pres-ence of salt is required to produce theglutamate effect. Glutamate taste is mosteffective in the pH range of 6 to 8 anddecreases at lower pH values. Sugar contentalso affects glutamate taste. The taste in acomplex food, therefore, depends on a com-plex interaction of sweet, sour, and salty, aswell as the added glutamate.

Monosodium glutamate improves the fla-vor of many food products and is thereforewidely used in processed foods. Productsbenefiting from the addition of glutamate

include meat and poultry, soups, vegetables,and seafood.

For many years glutamate was the onlyknown flavor enhancer, but recently a num-ber of compounds that act similarly havebeen discovered. The 5'-nucleotides, espe-cially 5'-inosinate and 5'-guanylate, haveenhancement properties and also show a syn-ergistic effect in the presence of glutamate.This synergistic effect has been demonstratedby determining the threshold levels of thecompounds alone and in mixtures. The datain Table 7-9 are quoted from Kuninaka(1966). The 5'-nucleotides were discoveredmany years ago in Japan as components ofdried bonito (a kind of fish). However, theywere not produced commercially and used asflavor enhancers until recently, when techni-cal problems in their production were solved.The general structure of the nucleotides withflavor activity is presented in Figure 7-21.There are three types of inosinic acid, 2'-, 3'-,and 5'-isomers; only the 5'-isomer has flavoractivity. Both riboside and S'-phosphomon-oester linkages are required for flavor activ-ity, which is also the case for the OH group atthe 6-position of the ring. Replacing the OHgroup with other groups, such as an aminogroup, sharply reduces flavor activity but thisis not true for the group at the 2-position.Hydrogen at the 2-position corresponds withinosinate and an amino group with guany-late; both have comparable flavor activity,and the effect of the two compounds is addi-tive.

The synergistic effect of umami substancesis exceptional. The subjective taste intensityof a blend of monosodium glutamate and di-sodium 5'-inosinate was found to be 16 timesstronger than that of the glutamate by itself atthe same total concentration (Yamaguchi1979).

S'-nucleotides can be produced by degra-dation of ribonucleic acid. The problem isthat most enzymes split the molecule at the3'-phosphodiester linkages, resulting innucleotides without flavor acitivity. Suitableenzymes were found in strains of Penicilliumand Streptomyces. With the aid of theseenzymes, the 5'-nucleotides can be manufac-tured industrially from yeast ribonucleic

acid. Another process produces the nucleo-side inosine by fermentation, followed bychemical phosphorylation to 5'-inosinic acid(Kuninaka 1966).

The search for other flavor enhancers hasbrought to light two new amino acids, tri-cholomic acid and ibotenic acid, obtainedfrom fungi (Figure 7-22). These aminoacids have flavor activities similar to that ofmonosodium glutamate. Apparently, the fla-vor enhancers can be divided into twogroups; the first consists of 5'-inosinate and5'-guanylate with the same kind of activityand an additive relationship. The othergroup consists of glutamate, tricholomic,and ibotenic acid, which are additive inaction. Between the members of the twogroups, the activity is synergistic.

A different type of flavor enhancer is mal-tol, which has the ability to enhance sweet-ness produced by sugars. Maltol is formedduring roasting of malt, coffee, cacao, andgrains. During the baking process, maltol isformed in the crust of bread. It is also foundin many dairy products that have beenheated, as a product of decomposition of thecasein-lactose system. Maltol (Figure 7-23)is formed from di-, tri-, and tetrasaccharidesincluding isomaltose, maltotretraose, and

Table 7-9 Threshold Levels of Flavor Enhancers Alone and in Mixtures in Aqueous Solution

Threshold Level (%)

Solvent

Water0.1%glutamate0.01%inosinate

Disodium 5'-lnosinate

0.0120.0001

Disodium5'-Guanylate

0.00350.00003

MonosodiumL-Glutamate

0.03

0.002

Source: From A. Kuninaka, Recent Studies of S'-Nucleotides as New Flavor Enhancers, In Flavor Chemistry, I.Hornstein, ed., 1966, American Chemical Society.

Figure 7-21 Structure of Nucleotides with Fla-vor Activity

panose but not from maltotriose. Formationof maltol is brought about by high tempera-tures and is catalyzed by metals such as iron,nickel, and calcium.

Maltol has antioxidant properties. It hasbeen found to prolong storage life of coffeeand roasted cereal products. Maltol is used asa flavor enhancer in chocolate and candies,ice cream, baked products, instant coffee andtea, liqueurs, and flavorings. It is used inconcentrations of 50 to 250 ppm and is com-mercially produced by a fermentation pro-cess.

ODOR

The olfactory mechanism is both morecomplex and more sensitive than the processof gustation. There are thousands of odors,and the sensitivity of the smell organ is about10,000 times greater than that of the tasteorgan. Our understanding of the odor recep-tor's mechanism is very limited, and there isno single, generally accepted theory account-ing for the relationship between molecularstructure and odor. The odorous substancearrives at the olfactory tissue in the nasalcavity, contained in a stream of air. Thismethod of sensing requires that the odorouscompound be volatile. Most odorous com-pounds are soluble in a variety of solvents,but it appears that solubility is less importantthan type of molecular arrangement, whichconfers both solubility and chemical reactiv-

ity (Moncrieff 1951). The number of volatilecompounds occurring in foods is very high.Maarse (1991) has given the following num-bers for some foods: beef (boiled, cooked)—486; beer—562; butter—257; coffee—790;grape—466; orange—203; tea—541; toma-to—387; and wine (white)—644. Not all ofthese substances may be essential in deter-mining the odor of a product. Usually, therelative amounts of a limited number of thesevolatile compounds are important in estab-lishing the characteristic odor and flavor of afood product.

The sensitivity of the human olfactoryorgan is inferior to that of many animals.Dogs and rats can detect odorous compoundsat threshold concentrations 100 times lowerthan man. When air is breathed in, only asmall part of it is likely to flow over theolfactory epithelium in the upper nasal cav-ity. When a smell is perceived, sniffing mayincrease the amount reaching the olfactorytissue. When foods are eaten, the passage ofbreath during exhalation reaches the nasalcavity from the back. Doving (1967) hasquoted the threshold concentrations of odor-ous substances listed in Table 7-10. Appar-ently, it is possible to change odor thresholdsby a factor of 100 or more by stimulating thesympathetic nervous system so that moreodor can reach the olfactory tissue. What isremarkable about the olfactory mechanism isnot only that thousands of odors can be rec-ognized, but that it is possible to store the

A B

Figure 7-22 (A) Tricholomic and (B) Ibotenic Acid

information in the brain for retrieval afterlong periods of time. The ability to smell isaffected by several conditions, such as colds,menstrual cycle, and drugs such as penicillin.Odors are usually the result of the presenceof mixtures of several, sometimes many, dif-ferent odorous compounds. The combinedeffect creates an impression that may be very

Table 7-10 Odor Threshold Concentrations ofOdorous Substances Perceived During NormalInspiration

Threshold ConcentrationCompound (Molecules/cc)

AIIyI mercaptan 6 x 107

Sec. butyl mercaptan 1 x 108

lsopropyl mercaptan 1 x 108

lsobutyl mercaptan 4 x 108

Tert. butyl mercaptan 6 x 108

Thiophenol 8x108

Ethyl mercaptan 1 x 109

1,3-Xylen-4-ol 2x1012

ji-Xylene 2x1012

Acetone 6x1013

Source: From K.B. Doving, Problems in the Physiol-ogy of Olfaction, in Symposium on Foods: The Chem-istry and Physiology of Flavors, H.W. Schultz et al.,eds., 1967, AVI Publishing.

different from that of the individual compo-nents. Many food flavors, natural as well asartificial, are of this compound nature.

Odor and Molecular Structure

M. Stoll wrote in 1957: "The whole sub-ject of the relation between molecular struc-ture and odor is very perplexing, as there isno doubt that there exist as many relation-ships of structure and odor as there are struc-tures of odorous substances." In 1971(referring to Stoll 1957), Teranishi wrote:"The relation between molecular structureand odor was perplexing then. It is now." Wecan observe a number of similarities betweenthe chemical structure of compounds andtheir odors. However, the field of food fla-vors, as is the field of perfumery, is still verymuch an art, albeit one greatly supported byscientists' advancing ability to classify struc-tures and identify the effect of certain molec-ular configurations. The odor potency of va-rious compounds ranges widely. Table 7-11indicates a range of about eight orders ofmagnitude (Teranishi 1971). This indicatesthat volatile flavor compounds may be presentin greatly differing quantities, from traces torelatively large amounts.

The musks are a common illustration ofcompounds with different structures that all

Figure 7-23 Some Furanones (1,2,3), Isomaltol (4), and Maltol (5)

i 2 3

4 5

Table 7-11 Odor Thresholds of CompoundsCovering a Wide Range of Intensity

Odorant Threshold (\ig/L of Water)

Ethanol 100,000Butyric acid 240Nootkatone 170Humulene 160Myrcene 15n-Amyl acetate 5A7-Decanal 0.0a- and p-Sinensal 0.05Methyl mercaptan 0.02p-lonone 0.0072-methoxy-3- 0.002

isobutylpyra-zine

Source: From R. Teranishi, Odor and MolecularStructure, in Gustation and Olfaction, G. Ohloff and A.F.Thomas, eds., 1971, Academic Press.

give similar odors. These may include tricy-clic compounds, macrocyclic ketones andlactones, steroids, nitrocyclohexanes, indanes,tetrahydronaphthalenes, and acetophenoses.Small changes in the structure of these mol-ecules may significantly change in potencybut will not affect quality, since all aremusky. There are also some compounds thathave similar structures and very differentodors, such as nootkatone and related com-pounds (Teranishi 1971). Nootkatone is aflavor compound from grapefruit oil. Thiscompound and 1,10-dihydronootkatone havea grapefruity flavor (Figure 7-24). Severalother related compounds have a woody fla-vor. The odor character of stereoisomersmay be quite different. The case of mentholhas already been described. Only mentholisomers have peppermint aroma. The iso-,neo-, and neoisomenthols have an unpleas-ant musty flavor. Naves (1957) describes the

difference between the cis- and trans- forms of3-hexenol (CH2OH-CH2-CH=CH-CH2CH3).The ds-isomer has a fresh green odor,whereas the frans'-isomer has a scent remi-niscent of chrysanthemum. The 2-trans-6-cisnonadienal smells of cucumber and is quitedifferent from the smell of the 2-trans-6-trans isomer (nonadienal, CHO-CH=CH-(CH2)2-CH=CH-CH2-CH3). Lengtheningof the carbon chain may affect odorousproperties. The odor of saturated acidschanges remarkably as chain length in-creases. The lower fatty acids, especiallybutyric, have very intense and unpleasantflavors, because an increased chain lengthchanges flavor character (Table 7-12) andlessens intensity. The fatty acids with 16 or18 carbon atoms have only a faint flavor.

Another example is given by Kulka (1967).Gamma-nonalactone has a strong coconut-like flavor; y-undecalactone has a peacharoma. As the chain length is increased byone more carbon atom, the flavor characterbecomes peach-musk. The lactones are com-pounds of widely differing structure andodor quality and are found as components ofmany food flavors. Gamma- and 5-lactoneswith 10 to 16 carbon atoms have beenreported (Juriens and OeIe 1965) as flavorcomponents of butter, contributing to the but-ter flavor in concentrations of only parts permillion. The flavor character and chemicalstructure of some y-lactones as reported byTeranishi (1971) are shown in Figure 7-25.One of these, the y-lactone with a total chainlength of 10 carbons, has peach flavor. Thea-hydroxy-p-methyl-y-carboxy-A^-y-hex-eno-lactone occurs in protein hydrolysateand has very strong odor and flavor of beefbouillon. Gold and Wilson (1963) found thatthe volatile flavor compounds of celery con-tain a number of phthalides (phthalides arelactones of phthalic acid, lactones are inter-

nal esters of hydroxy acids). These includethe following:

• 3-isobutyliden-3a,4-dihydrophthalide(Figure 7-26)

• 3-isovalidene-3a,4-dihydrophthalide• 3-isobutylidene phthalide• 3-isovalidene phthalide

These compounds exhibit celery-like odorsat levels of 0.1 ppm in water. Pyrazines havebeen identified as the compounds giving thecharacteristic intense odor of green peppers(Seifert et al. 1970). A number of pyrazinederivatives were tested and, within this singleclass of compounds, odor potencies showed arange of eight orders of magnitude equal tothat of the widely varying compounds listed inTable 7-11. The compounds examined bySeifert et al. (1970) are listed in Table 7-13.2-methoxy-3-isobutylpyrazine appears to bethe compound responsible for the green pep-per odor. Removal of the methoxy- or alkyl-

groups reduces the odor potency by 105 to 106

times, as is the case with 2-methoxypyrazine,2-iosbutylpyrazine, and 2,5-dimethylpyrazine.Thus, small changes in molecular structuremay greatly affect flavor potency. The odorsof isobutyl, propyl, and hexyl methoxypyra-zines are similar to that of green peppers. Theisopropyl compound is moderately similar topeppers and its odor is somewhat similar toraw potato. The ethyl compound is even moresimilar to raw potato and less to pepper. Infact, this compound can be isolated from pota-toes. The methyl compound has an odor likeroasted peanuts. The structure of some of thepyrazines is shown in Figure 7-27. Pyrazineshave been identified as flavor components in anumber of foods that are normally heated dur-ing processing. Rizzi (1967) demonstrated thepresence of seven alkyl-substituted pyrazinesin chocolate aroma. These were isolated bysteam distillation, separated by gas-liquidchromatography, and identified by mass spec-trometry. The components are methyl pyra-

Figure 7-24 Odor Character of Nootkatone and Related Compounds

Uonootkotone Eremophilone

4—Epinootkotone Tetrohydronootk atone 1U2-Dihydronootkotone

WOODY1,10—DihvdronootlcafoneNootkatone

GRAPEFRUITY

Table 7-12 Flavor Character of Some N-Carboxylic Acids

Acid Flavor Character

Formic Acid, pungentAcetic Acid, vinegary, pungentPropionic Acid, pungent, rancid,

cheesyButyric Acid, rancidHexanoic Sweaty, goatyOctanoic RancidDecanoic WaxyLaurie TallowyMyristic Soapy, cardboardPalmitic Soapy

zine; 2,3-dimethylpyrazine; 2-ethyl-5-methyl-pyrazine; trimethylpyrazine; 2,5-dimethyl-3-ethylpyrazine; 2,6-dimethyl-3-ethylpyrazine;and tetramethylpyrazine. Other researchers(Flament et al. 1967; Marion et al. 1967) haveisolated these and other pyrazines from thearoma components of cocoa. Pyrazines arealso aroma constituents of coffee. Goldman etal. (1967) isolated and identified 24 pyra-zines and pyridines and revealed the pres-ence of possibly 10 more. Bondarovich et al.(1967) isolated and identified a large numberof pyrazines from coffee aroma and drew

attention to the importance of pyrazines anddihydropyrazines to the flavor of roasted orotherwise cooked foods. These authors alsodrew attention to the instability of the dihy-dropyrazines. This instability not only makestheir detection and isolation difficult, but mayhelp explain why flavors such as that ofroasted coffee rapidly change with time.Another roasted product from which pyra-zines have been isolated is peanuts. Mason etal. (1966) found methylpyrazine; 2,5-dimeth-ylpyrazine; trimethylpyrazine; methylethyl-pyrazine; and dimethylethylpyrazine in theflavor of roasted peanuts. The pyrazinesappear to be present in unprocessed as well asin heated foods.

Another group of compounds that havebeen related to the aroma of heated foods isthe furanones. Teranishi (1971) summarizedthe findings on several of the furanones (seeFigure 7-23). The 4-hydroxy-2,5-dimethyl-3-dihydrofuranone (1) has a caramel or burntpineapple odor. The 4-hydroxy-5-methyl-3-dihydrofuranone (2) has a roasted chicoryroot odor. Both compounds may contributeto beef broth flavor. The 2,5-dimethyl-3-dihydrofuranone (3) has the odor of freshlybaked bread. Isomaltol (4) and maltol (5) areproducts of the caramelization and pyrolysisof carbohydrates.

Beef bouillon

Figure 7-25 Flavor Character of Some Lactones. Source: From R. Teranishi, Odor and MolecularStructure, in Gustation and Olfaction, G. Ohloff and A.F. Thomas, eds., 1971, Academic Press.

R = 05!-! J1 (coconut)R = CgH13 (peach)R = C7H15 (peach)R = C8H17 (peach-musk)

Figure 7-26 Phthalides of Celery Volatiles

Theories of Olfaction

When an odoriferous compound, or odor-

ivector, arrives at the olfactory organ, a reac-

tion takes place between the odor molecules

and the chemoreceptors; this reaction pro-

Table 7-13 Odor Threshold of Pyrazine and

Derivatives

duces a neural pulse, which eventuallyreaches the brain. The exact nature of theinteraction between odorivector and chemo-receptor is not well known. The number ofolfactory receptors in the smell organs is inthe order of 100 million, and Moncrieff(1951) has calculated that the number ofmolecules at the threshold concentration ofone of the powerful mercaptans in a sniff(about 20 mL) of air would be 1 x 1010 mole-cules. Obviously, only a fraction of thesewould interact with the receptors, butundoubtedly numerous interactions are re-quired to produce a neural response.Dravnieks (1966) has indicated that accord-ing to information theory, 13 types of sensorsare needed to distinguish 10,000 odors on ayes-or-no basis, but more than 20 might berequired to respond rapidly and withouterror. Many attempts have been made to clas-sify odors into a relatively small number ofgroups of related odors. These so-called pri-mary odors have been used in olfaction theo-ries to explain odor quality. One theory, thestereochemical site theory (Amoore et al.1964; Amoore 1967), is based on molecularsize and shape. Amoore compared the vari-ous odor qualities that have been used tocharacterize odors and concluded that sevenprimary odors would suffice to cover themall: camphoraceous, pungent, ethereal, floral,

Compound

2-methoxy-3-hexylpyrazine2-methoxy-3-isobutylpyra-

zine2-methoxy-3-propylpyra-

zine2-methoxy-3-isopropylpyra-

zine2-methoxy-3-ethylpyrazine2-methoxy-3-methylpyra-

zine2-methoxypyrazine2-isobutylpyrazine2-5-dimethylpyrazinepyrazine

Odor Threshold(Parts perl O12

Parts of Water)

12

6

2

4004000

700,000400,000

1,800,000175,000,000

Source: From R.M. Seifert et al., Synthesis of Some2-Methoxy-3-Alkylpyrazines with Strong Bell Pepper-Like Odors, J. Agr. Food Chem., Vol. 18, pp. 246-249,1 970, American Chemical Society.

pepperminty, musky, and putrid. Table 7-14lists some of the chemical compounds thatcan be used to demonstrate these primaryodors. The theory is based on the assumptionthat all odorous compounds have a distinc-tive molecular shape and size that fit into asocket on the receptor site. This would besimilar to the "lock-and-key" concept ofenzyme action. Five of the receptor siteswould accept the flavor compound accordingto shape and size, and two (pungent andputrid) on the basis of electronic status (Fig-ure 7-28). The site-fitting concept as initiallyproposed was inadequate because it assessedonly one-half of the molecule; subsequentrefinements considered all aspects of molec-ular surface in a "shadow-matching" tech-nique (Amoore 1967). It was also suggestedthat there may be more than seven primaries.The primary odors may have to be split intosubgroups and others added as new prima-ries. Molecular model silhouettes as devel-oped for five primary odors are reproducedin Figure 7-29.

A membrane-puncturing theory has beenproposed by Davis (Dravnieks 1967). Ac-cording to this theory, the odorous substancemolecules are adsorbed across the interfaceof the thin lipid membrane, which forms partof the cylindrical wall of the neuron in thechemoreceptor and the aqueous phase thatsurrounds the neuron. Adsorbed moleculesorient themselves with the hydrophilic endtoward the aqueous phase. When theadsorbed molecules are desorbed, they moveinto the aqueous phase, leaving a defect. Ions

may adsorb into this puncture and cause aneural response. This theory could be con-sidered a thermodynamic form of the profilefunctional group concept, since the freeenergy of adsorption of the odor substance atthe interface is related to shape, size, func-tional groups and their distribution, and posi-tion. The adsorption is a dynamic processwith a free energy of adsorption of about 1 to8 kcal/mole for different substances. Daviesprepared a plot of molecular cross-sectionalarea versus free energy of adsorption andobtained a diagram (Figure 7-30) in whichgroups of related odors occupy distinct areas.

The suggestion that odorous character isrelated to vibrational specificity of odor mol-ecules has led to the vibrational theory ofolfaction (Wright 1957). Vibrational energylevels can be derived from the infrared orRaman spectra. The spectral area of greatestinterest is that below 700 cm'1, which isrelated to vibrations of chains and flexing ortwisting of bonds between groups of atomsin the molecule. Wright and others havedemonstrated that correlations exist betweenspectral properties and odor quality in anumber of cases, but inconsistencies in othercases have yet to be explained.

Obviously, none of the many theories ofolfaction proposed so far have been entirelysatisfactory. It might be better to speak ofhypotheses rather than of theories. Most ofthese theories deal with the explanation ofodor quality and do not account for the quan-titative aspects of the mechanism of olfac-tion. The classification of odor and the

A B C

Figure 7-27 (A) Pyrazine, (B) 2-Methoxypyrazine, and (C) 2-Methoxy-3-Hexylpyrazine

correlation of chemical structure and odorremain difficult to resolve.

Odor Description

An odor can be described by the combina-tion of threshold value and odor quality. Thethreshold value, the lowest concentration thatcreates an odor impression, can be consid-ered the intensity factor, whereas the odorquality describes the character of the aroma.As has been mentioned under olfactory theo-ries, attempts at reducing the number ofcharacteristic odor qualities to a small num-ber have not been successful. In many cases,the aroma and flavor of a food can be relatedto the presence of one or a few compoundsthat create an impression of a particular foodwhen smelled alone. Such compounds havebeen named contributory flavor compoundsby Jennings and Sevenants (1964). Somesuch compounds are the pyrazines, whichgive the odor quality of green bell peppers;nootkatone for grapefruit; esters for fruits;

and nona-2-frans-6-a's-dienal for cucum-bers (Forss et al. 1962). In a great number ofother cases, there are no easily recognizablecontributory flavor compounds, but the fla-vor seems to be the integrated impression ofa large number of compounds.

Determining the threshold value is difficultbecause subthreshold levels of one com-pound may affect the threshold levels ofanother. Also, the flavor quality of a com-pound may be different at threshold level andat suprathreshold levels. The total range ofperception can be divided into units that rep-resent the smallest additional amount thatcan be perceived. This amount is called justnoticeable difference (JND). The wholeintensity scale of odor perception coversabout 25 JNDs; this is similar to the numberof JNDs that comprise the scale of tasteintensity. Flavor thresholds for some com-pounds depend on the medium in which thecompound is dispersed or dissolved. Patton(1964) found large differences in the thresh-old values of saturated fatty acids dissolvedin water and in oil.

Table 7-14 Primary Odors for Humans and Compounds Eliciting These Odors

Primary Odor

CamphoraceousPungentEthereal

FloralPepperminty

Musky

Putrid

Odor Compounds

Borneol, terf-butyl alcohol d-camphor, cineol, pentamethyl ethyl alcoholAIIyI alcohol, cyanogen, formaldehyde, formic acid, methylisothiocyanateAcetylene, carbon tetrachloride, chloroform, ethylene dichloride, propyl

alcoholBenzyl acetate, geraniol, a-ionone, phenylethyl alcohol, terpineolte/t-butylcarbinol, cyclohexanone, menthone, piperitol, 1,1,3-trimethyl-

cyclo-5-hexanoneAndrostan-3oc-ol (strong), cyclohexadecanone, ethylene cebacate, 17-

methylandrostan-3a-ol, pentadecanolactoneAmylmercaptan, cadaverine, hydrogen sulfide, indole (when concentrated,

floral when dilute), skatole

Source: From J.E. Amoore et al., The Stereochemical Theory of Odor, Sc/. Am., Vol. 210, No. 2, pp. 42-49, 1964.

PEPPERMINTY ETHEREAL PUTRID

PUNGENT

FLORALMUSKYCAMPHORACEOUS

Figure 7-28 Olfactory Receptor Sites According to the Stereochemical Theory of Odor

Top silhouette Front silhouette Right silhouette

1,2-Dichloroethane(ethereal)

1,8-Cineole(camphoraceous)

15-Hydroxypentadecanoicacid lactone

(musky)

d,/-p-Phenylethyl methylethylcarbinol

(floral)

d,l-Menthone(minty)

Figure 7-29 Molecular Model Silhouettes of Five Standard Odorants. Source: From J. Amoore, Stere-ochemical Theory of Olfaction, in Symposium on Foods: The Chemistry and Physiology of Flavors,H.W. Schultz et al., eds., 1967, AVI Publishing Co.

DESCRIPTION OF FOOD FLAVORS

The flavor impression of a food is influ-enced by compounds that affect both tasteand odor. The analysis and identification ofmany volatile flavor compounds in a largevariety of food products have been assistedby the development of powerful analyticaltechniques. Gas-liquid chromatography waswidely used in the early 1950s when com-mercial instruments became available. Intro-duction of the flame ionization detectorincreased sensitivity by a factor of 100 and,together with mass spectrometers, gave amethod for rapid identification of many com-ponents in complex mixtures. These methodshave been described by Teranishi et al.(1971). As a result, a great deal of informa-tion on volatile flavor components has beenobtained in recent years for a variety of foodproducts. The combination of gas chroma-tography and mass spectrometry can provideidentification and quantitation of flavor com-pounds. However, when the flavor consistsof many compounds, sometimes several hun-

dred, it is impossible to evaluate a flavorfrom this information alone. It is then possi-ble to use pattern recognition techniques tofurther describe the flavor. The pattern rec-ognition method involves the application ofcomputer analysis of complex mixtures ofcompounds. Computer multivariate analysishas been used for the detection of adultera-tion of orange juice (Page 1986) and Spanishsherries (Maarse et al. 1987).

Flavors are often described by using thehuman senses on the basis of widely recog-nized taste and smell sensations. A proposedwine aroma description system is shown inFigure 7-31 (Noble et al. 1987). Such sys-tems attempt to provide an orderly and reli-able basis for comparison of flavor descrip-tions by different tasters.

The aroma is divided into first-, second-,and third-tier terms, with the first-tier termsin the center. Examination of the descriptorsin the aroma wheel shows that they can bedivided into two types, flavors and off-fla-vors. Thus, it would be more useful to dividethe flavor wheel into two tables—one for fla-

Figure 7-30 Plot of Molecular Cross-Sectional Area Versus Free Energy of Adsorption for Davies'Theory of Olfaction

-AG0/w(CALORIES MOLE"')

(J. T. OAVlES )

MO

LECU

LAR

2C

RO

SS-S

ECTI

ON

AL A

REA

(A

.)

Next page