TUGAS

TEKNOLOGI FERMENTASI

MEDIUM FERMENTASI

Disusun Oleh:

ADITYA PUJASAKTI YUSWI

H0912002

ILMU DAN TEKNOLOGI PANGAN

FAKULTAS PERTANIAN

UNIVERSITAS SEBELAS MARET

SURAKARTA

2015

Formulation of fermentation media from flour-rich waste streams for microbial

lipid production by Lipomyces starkeyi

Sofia Tsakona, Nikolaos Kopsahelis, Afroditi Chatzifragkou, Seraphim

Papanikolaou, Ioannis K. Kookos, Apostolis A. Koutinas

Pada fermentasi ini digunakan metode fermentasi (Solid State Fermentation)

SSF dilakukan dengan jamur strain A. awamori 2B.361 U 2/1 yang tersedia dari

Profesor Colin Webb (University of Manchester, Inggris).

Selama penelitian, diamati bahwa produksi minyak mikroba efisien dalam

kultur bioreaktor dengan L. starkeyi sangat tergantung pada produksi inokulum padat.

Untuk mencapai hal tersebut dalam fed-batch fermentasi bioreaktor, pra-kultur yang

diproduksi di hidrolisat (Flour-rich waste) FRW yang menyebabkan peningkatan

efisiensi. Fermentasi dilakukan baik menggunakan hidrolisat FRW atau media sintetis.

Komposisi Medium

Dalam kasus terakhir komposisi media (dalam g/L): glukosa, 105; ekstrak ragi,

2,0; (NH4)2SO4, 1,0; KH2PO4, 7,0; Na2HPO4, 2,5; MgSO4.7H2O, 1.5; FeCl3.6H2O,

0,15; ZnSO4.7H2O, 0,02; MnSO4.H2O, 0,06; CaCl2.2H2O, 0,15.

Tetapi pada penelitian ini dilakukan dengan cara sebagai berikut:

Spora jamur disimpan pada 4C dikemiringan yang mengandung 5% (b/v)

dedak gandum (WB) dan 2% (b/v) agar. Strain ragi oleaginous L. starkeyi DSM 70296

ditambahkan dalam fermentasi untuk mikroba produksi minyak. Dipertahankan pada

4oC diagar miring yang mengandung glukosa (10 g/L), ekstrak ragi (10 g/L), pepton

(10 g/L) dan agar (2%, b/v). Sebuah medium cair dari komposisi yang sama pada

glukosa, ekstrak ragi dan pepton tersebut digunakan untuk persiapan inokulum

fermentasi.

*) Medium yang digunakan pada penelitian ini hanya glukosa, ekstrak ragi, pepton dan

agar.

Fungsi dalam Proses Fermentasi

Gandum giling digunakan dalam Solid State Fermentation (SSF) dari

Aspergillus awamori untuk produksi enzim kasar, terutama glukoamilase dan protease.

Stream (Flour-rich waste) FRW yang disuplai oleh Jotis SA, sebuah industri gula

Yunani yang menghasilkan berbagai macam produk permen dan makanan untuk bayi.

FRW diuji sebagai substrat padat tambahan di SSF untuk mengevaluasi potensi untuk

meningkatkan produksi enzim amilolitik dan proteolitik. FRW utamanya digunakan

dalam percobaan hidrolisis bertujuan untuk menghasilkan media fermentasi generik.

Stream FRW digunakan untuk optimasi hidrolisis enzimatik dan mikroba produksi

minyak yang dihasilkan selama proses pembuatan makanan bagi bayi. Enzyme-rich

SSF padatan yang kemudian digunakan untuk hidrolisis FRW stream menjadi kaya gizi

media fermentasi.

Hasil yang terbaik

Batch percobaan hidrolitik dengan 2 metode yaitu Batch hidrolisis dan Fed-

batch hidrolisis. Menggunakan konsentrasi (Flour-rich waste) FRW hingga 205 g/L

menghasilkan lebih tinggi dari 90% (b/b) pati menjadi glukosa hasil konversi dan 40%

(b/b) Total Kjeldahl Nitrogen (TKN) untuk hasil konversi amino nitrogen bebas

(FAN). Pati menjadi glukosa hasil konversi dari 98,2; 86,1 dan 73,4% (w/w) yang

diperoleh ketika konsentrasi FRW awal 235, 300 dan 350 g/L ditambahkan pada fed-

batch percobaan hidrolitik, masing-masing. Hidrolisat minyak mentah digunakan

sebagai media fermentasi dalam kultur labu kocok dengan ragi oleaginous Lipomyces

starkeyi DSM 70296 mencapai total berat kering 30,5 g/L dengan jumlah minyak yang

dihasilkan mikroba dari 40,4% (w/w), lebih tinggi dari yang dicapai media sintetis.

Kultur bioreaktor Fed-batch menyebabkan total berat kering 109,8 g/L dengan mikroba

menghasilkan kandungan minyak 57,8% (b/b) dan produktivitas sebesar 0,4 g/L/jam.

Sehingga dapat disimpulkan, dengan penambahan medium Flour-rich waste (FRW)

yang akan menghasilkan hidrolisat minyak mentah dan dengan menggunakan metode

hidrolisis Fed-batch diperoleh hasil total berat kering, jumlah minyak dan produktifitas

yang lebih baik dan efisien.

Journal of Biotechnology 189 (2014) 3645

Contents lists available at ScienceDirect

Journal of Biotechnology

j ourna l ho me pa ge: www.elsev ier .com/ locate / jb io tec

Formulation of fermentation media from our-rimicrobial lipid production by Lipomyces starkeyi

Soa Tsa ub,Ioannis Ka Department o 1855 Ab Department o , UKc Department o

a r t i c l

Article history:Received 21 MReceived in reAccepted 12 AAvailable onlin

Keywords:Food waste valorisationSolid state fermentationEnzymatic hydrolysisLipomyces starMicrobial oil

ct straw milling

of cruSSF solids were subsequently employed for hydrolysis of FRW streams into nutrient-rich fermentationmedia. Batch hydrolytic experiments using FRW concentrations up to 205 g/L resulted in higher than 90%(w/w) starch to glucose conversion yields and 40% (w/w) total Kjeldahl nitrogen to free amino nitro-gen conversion yields. Starch to glucose conversion yields of 98.2, 86.1 and 73.4% (w/w) were achievedwhen initial FRW concentrations of 235, 300 and 350 g/L were employed in fed-batch hydrolytic experi-ments, respectively. Crude hydrolysates were used as fermentation media in shake ask cultures with the

1. Introdu

The devthrough vaby current iply chain wsupply chaimillion t of food manuvices (14%)2010). Repoof annual flost or wasto the cons

Corresponand PreservatiNutrition, Agrfax: +30 210 5

E-mail add

http://dx.doi.o0168-1656/ keyioleaginous yeast Lipomyces starkeyi DSM 70296 reaching a total dry weight of 30.5 g/L with a microbial oilcontent of 40.4% (w/w), higher than that achieved in synthetic media. Fed-batch bioreactor cultures ledto a total dry weight of 109.8 g/L with a microbial oil content of 57.8% (w/w) and productivity of 0.4 g/L/h.

2014 Elsevier B.V. All rights reserved.

ction

elopment of sustainable processes could be achievedlorisation of waste and by-product streams generatedndustrial processes (Koutinas et al., 2014a,b). Food sup-aste and by-product streams are generated along anyn of food industrial sectors and sub-sectors. Around 89food waste is generated annually in the EU-27 throughfacturing (39%), households (42%), food catering ser-

and retail/wholesale of food products (Monier et al.,rts published by FAO estimate more than 1.3 billion t

ood supply chain waste corresponding to 50% of foodted along the supply chain from the agricultural eldumer (Partt et al., 2010). Flour-rich waste (FRW) and

ding author at: Laboratory of Food Process Engineering, Processingon of Agricultural Products, Department of Food Science and Humanicultural University of Athens, Greece. Tel.: +30 210 5294729;294729.ress: akoutinas@aua.gr (A.A. Koutinas).

by-product streams are generated by many industrial food sectorsbelonging mainly to the following categories as have been classiedby the PRODCOM List 2013 (Anonymous, 2014a):

manufacture of grain mill products (PRODCOM code 10.61) thatconstitute the 4th (56.9 106 t in EU-27 in 2012) most importantfood sector in terms of production capacity;

manufacture of bread, fresh pastry goods and cakes (PRODCOM code10.71) that constituted the 7th (26.4 106 t in EU-27 in 2012)most important food sector in terms of production capacity;

manufacture of rusks, biscuits and preserved pastry goods and cakes(PRODCOM code 10.72);

various types of confectionery products and food for infants.

Flour-rich waste streams are mainly generated during the man-ufacturing process, disposed by consumers and catering services orare returned from the market as end-of-date products. To providean estimate of the capacity of such waste streams, it could be men-tioned that around 800,000 t of bakery waste is produced annuallyin the UK (Anonymous, 2011). FRW contain signicant quantitiesof starch and protein as well as various micro-nutrients that could

rg/10.1016/j.jbiotec.2014.08.0112014 Elsevier B.V. All rights reserved.konaa, Nikolaos Kopsahelisa, Afroditi Chatzifragko. Kookosc, Apostolis A. Koutinasa,

f Food Science and Human Nutrition, Agricultural University of Athens, 75 Iera Odos, 1f Food and Nutritional Sciences, University of Reading, Whiteknights, Reading RG6 6ADf Chemical Engineering, University of Patras, 26504 Patras, Rio, Greece

e i n f o

ay 2014vised form 6 August 2014ugust 2014e 20 August 2014

a b s t r a c t

Flour-rich waste (FRW) and by-produplants could be employed as the sole for microbial oil synthesis. Wheat mAspergillus awamori for the productionch waste streams for

Seraphim Papanikolaoua,

thens, Greece

reams generated by bakery, confectionery and wheat millingaterials for generic fermentation media production, suitable

by-products were used in solid state fermentations (SSF) ofde enzymes, mainly glucoamylase and protease. Enzyme-rich

S. Tsakona et al. / Journal of Biotechnology 189 (2014) 3645 37

be used for the production of generic fermentation feedstocks.The main by-product stream generated by wheat our millers isproduced via milling and sifting of wheat and contains mainly branand varying quantities of endosperm depending on the type ofwheat ouranimal feedmilling by-via solid staThus, FRW renewable chemical pr

Microbiaresearch dualternative renewable heterotrophcapable of bioconversilipids are mwith quantand polar frthe oleaginolipids may pvarying proof microbiachain saturageneral accduction (Viother microhydrogenathols constitproducts su(Naik et al.products wdevelopmeof oleochemnature (BtMetzger anfor natural by-productural fats andcurrent manthrough 201of key vegeutilisation othesis can bbiorenery

Recent rbial oil proLeiva-Cand2013) and e2014b). Utmercial andare of parammicrobial oed microbbioreactor pbial oil conwhen the cunitary costditure for that bioreacto2014b). Incto a microbhigher bioreate a cost-c

recently demonstrated the potential to satisfy the required criteriato achieve industrial implementation. Lin et al. (2011) reported theproduction of a total dry weight (TDW) of 104.6 g/L with a microbialoil content of 64.9% (w/w) at a productivity of 1.2 g/L/h via fed-batch

ctor l meturesture

micL/h.s stuedstoial of Aspd tominsphoixed

feedsduct. There rwithydroted hction

The Fry lin

pro wasr hydmogtione

lacto usestrateous

teria

icroo

werthat

of Mrageas e

in slo

oleaenta

on ), pee cr th

l woreacton of atatiod to imentheticn g/L4, 7.7H2O produced. This by-product stream is mainly used as, but surplus quantities are treated as wastes. Wheatproducts (WMB) can be used for enzyme productionte fermentations (SSF) using appropriate fungal strains.and by-product streams could be employed as the soleresources for bioprocess development employed foroduction (Lin et al., 2013; Koutinas et al., 2014a).l lipid production represents a signicant eld ofring the last decades. The increasing interest towardspathways for cost-effective production of biofuels fromresources has set the ground for intensive search onic (yeast and fungi) or phototrophic (algae) organisms,

accumulating high oil contents, accompanied by highon yields (Papanikolaou and Aggelis, 2010). Microbialainly composed of triacylglycerols (TAGs), together

ities of free-fatty acids and to a lesser extent sterolsactions (Papanikolaou and Aggelis, 2011). Depending onus strain and the applied cultivation method, microbialresent a diversied fatty acid composition attributingperties to the produced lipid bodies. The compositionl lipids is similar to vegetable oils, comprised of longted and unsaturated fatty acids, a fact that justies theireptance as suitable starting material for biodiesel pro-ncente et al., 2009). Besides fatty acids methyl esters,bial oil derivatives such as fatty acids, glycerol anded products of fatty acid methyl esters and fatty alco-ute base materials for a broad spectrum of oleochemicalch as surfactants, lubricants, polymers and plastics

, 2010). The need to replace petrochemically derivedith bio-based ones has paved the way for research andnt towards vegetable oil utilisation for the productionicals due to their non-toxic and readily biodegradabletcher et al., 2009; Buchholz and Bornscheuer, 2005;d Bornshcheuer, 2006). Moreover, worldwide demandfatty acids and glycerol, nowadays available as a crude

stream from biodiesel production processes using nat- oils, is expected to grow about 9.8% annually, from theufacturing value of $7.7 billion in 2011 to $13.5 billion7, based on the expected continuous increase of pricestable oils and animal fats (Anonymous, 2014a,b). Thef microbial lipids as precursors for oleochemical syn-e considered as a promising path, especially within aconcept.esearch has focused on the improvement of micro-duction efciency (Koutinas and Papanikolaou, 2011;ia et al., 2014; Papanikolaou et al., 2013; Shen et al.,valuation of its cost-competitiveness (Koutinas et al.,

ilising crude renewable resources rather than com- puried carbon sources and nutrient supplementsount importance towards the development of viable

il production processes. The production cost of puri-ial oil at an annual production capacity of 10,000 t, aroductivity of 0.54 g/L/h, a TDW of 106.5 g/L and micro-tent of 67.5% (w/w) has been estimated at $3.4 per kgost of glucose is negligible (Koutinas et al., 2014b). The

of puried microbial oil production at negligible expen-e carbon source could be closer to that of vegetable oilsr productivities higher than 2.5 g/L/h (Koutinas et al.,

reasing the price of glucose to $400 per kg correspondsial oil production cost of $5.5 per kg and thereforeactor productivities should be achieved in order to cre-ompetitive process. The strain Lipomyces starkeyi has

bioreaminerator culon mixwith a0.29 g/

Thition femicrobtions orequireilable aof phowere mtation oil pro70296literatumedia FRW hcentraliquefaours.industtioneryin thisof outhis hoconfecstarch,will bedemonerogen

2. Ma

2.1. M

SSFU 2/1 versityand sto(Koutinat 4 C agar.

Thein fermat 4 C(10 g/Lthe samused fomentain biorductiofermenthat le

Feror syntwas (iKH2POZnSO4cultures of L. starkeyi AS 2.1560 using glucose-baseddium. Tapia et al. (2012) carried out fed-batch bioreac-

of a mutant strain of L. starkeyi DSM 70296 cultivateds of glucose and xylose to achieve a TDW of 88.7 g/Lrobial oil content of 55.2% (w/w) at a productivity of

dy focuses on the production of a generic fermenta-ck from WMB and FRW streams that could be used foril production. WMB were used in solid state fermenta-ergillus awamori for the production of all the enzymes

convert starch into glucose, protein into directly assim-o acids and peptides, and generate an inorganic sourcerus via phytic acid hydrolysis. Enzyme-rich SSF solids

with FRW suspensions for the production of fermen-tocks that were subsequently evaluated for microbialion using the oleaginous yeast strain L. starkeyi DSM

obtained results are among the highest reported in theegarding microbial oil production from glucose-based

the strain L. starkeyi. Multi-enzyme production andlysis were optimised in order to achieve highly con-ydrolysates similar to those produced in traditional

and saccharication processes of starch and cerealRW used in this study were produced by a confectionerye producing food for infants and our-based confec-

ducts. Therefore, starch was the carbohydrate containedte stream. This study was focused on the optimisationrolysis and evaluation of microbial oil production usingeneous FRW stream. In a forthcoming publication mixedry waste streams containing various carbohydrates (i.e.se, sucrose and fructose) as well as proteins and lipids

d for fermentative microbial oil production in order toe the efciency of microbial oil production using het-

FRW streams.

ls and methods

rganisms and media

e carried out with the fungal strain A. awamori 2B.361was kindly provided by Professor Colin Webb (Uni-anchester, UK). The origin, purication, sporulation

methods have been described in previous publicationst al., 2005; Wang et al., 2009). Fungal spores were storedpes containing 5% (w/v) wheat bran (WB) and 2% (w/v)

ginous yeast strain L. starkeyi DSM 70296 was employedtions for microbial oil production. It was maintainedagar slopes containing glucose (10 g/L), yeast extractptone (10 g/L) and agar (2%, w/v). A liquid medium ofomposition in glucose, yeast extract and peptone wase preparation of fermentation inocula. During experi-k, it was observed that efcient microbial oil productionr cultures by L. starkeyi is highly dependent on the pro-

dense inoculum. To achieve this in fed-batch bioreactorns, pre-cultures were produced in FRW hydrolysatesmproved fermentation efciency.ations were carried out using either FRW hydrolysates

media. In the latter case the composition of the medium): glucose, 105.0; yeast extract, 2.0; (NH4)2SO4, 1.0;0; Na2HPO4, 2.5; MgSO47H2O, 1.5; FeCl36H2O, 0.15;, 0.02; MnSO4H2O, 0.06; CaCl22H2O, 0.15.

38 S. Tsakona et al. / Journal of Biotechnology 189 (2014) 3645

2.2. Raw materials

WMB were used as substrate for solid state fermentation of A.awamori. The starch, protein, phosphorus and moisture contentof WMB usrespectivelyfectionery iproducts ansubstrate induction of aused in hydmentation of enzymatproduced dstarch, protprocesses woptimisatiowaste streaalso evaluathe develop

2.3. Solid st

SSF was (100:0, 90:1the total drsubstrates wErlenmeyermL was useinoculum, tat 30 C on Erlenmeyerlum for SSFVarying moilised tap wevaluated a60, 65, 70 a

2.4. Produc

Remainiafter suspenwere produthe enzymepension wa4, 4.5, 5, 540, 45, 50, of experimeried out at tconditions. enzyme act175, 205, 23(0.24, 0.49, 16.13 U/mLried out in bhydrolysis ssion, while approach eFRW concen

Samplesseparated vwas lteredwas subseqamino nitrotation medi(Polycap TM

adjusted to the optimum pH value of 6.0 for yeast strains growthusing 5 M NaOH.

2.5. Shake ask fermentations

ke aeyertic m

the thatEachl pren thee. Flaeries

rpmlue gas p. Distive

oreac

-batcew BL usi

wa A 10-culof 15

of svm ad ind to as ad200 hermihrouin)

wasorus, enere te. D

.

alyt

coamtrati

phas ususptibiliounay coteaseysis oC w

d as the rples

qual e en

g foed in this study were 12%, 20%, 1.1% and 9.7% (w/w),. FRW streams were supplied by Jotis S.A., a Greek con-ndustry that produces a wide range of confectioneryd food for infants. FRW were tested as additional solid

SSF to evaluate the potential to enhance the pro-mylolytic and proteolytic enzymes. FRW were mainlyrolysis experiments aiming to produce a generic fer-medium. The FRW streams used for the optimisationic hydrolysis and the production of microbial oil wereuring the manufacturing process of food for infants. Theein and moisture content of FRW used in fermentationere 84.8%, 7.3% and 5% (w/w), respectively. After then of starch and protein hydrolysis, various our-basedms that contained only starch as carbohydrate wereted in order to demonstrate the wide applicability ofed process.

ate fermentations

carried out at 30 C using various ratios of WMB to FRW0, 80:20, 70:30 and 60:40, on a weight basis). In all casesy weight of the substrate used for SSF was 5 g. Solidere initially sterilised at 121 C for 20 min in 250 mL

asks. A fungal spore suspension of 2 106 spores perd as inoculum. To achieve this spore concentration in thehe fungal strain A. awamori was sporulated for 5 daysa solid medium (5% WMB and 2% agar, w/v) in 250 mL

asks. The procedure followed to produce the inocu- has been presented by Kachrimanidou et al. (2013).isture contents were achieved through addition of ster-ater. The production of glucoamylase and protease wast varying WMB to FRW ratios, moisture content (50, 55,nd 75%, w/w) and fermentation duration.

tion of FRW hydrolysate

ng SSF solids were macerated using a kitchen blendersion in 500 mL sterilised tap water. FRW hydrolysatesced by mixing varying quantities of FRW streams with-rich suspension in 1 L Duran bottles. Mixing of the sus-s achieved using magnetic stirrers. The effect of pH (3.5,.5, 6, 6.5 and uncontrolled pH) and temperature (30,55, 60 and 65 C) was evaluated in two different setsnts. All subsequent hydrolytic experiments were car-he optimum temperature of 55 C and uncontrolled pHFRW hydrolysis was optimised using different initialivities and seven initial FRW concentrations (98, 140,5 and 350 g/L, on a wet basis). Four initial glucoamylase0.73 and 0.97 U/mL) and protease (4.03, 8.06, 12.9 and) activities were used. Hydrolytic experiments were car-oth batch and fed-batch mode. In the latter case, FRWtarted by adding only 80% of the enzyme-rich suspen-the remaining 20% was added at approximately 6 h. Thisnhanced the nal starch to glucose conversion yield attrations higher than 205 g/L.

were collected at random intervals and the solids wereia centrifugation (3000 g for 10 min). The supernatant

(Whatman No 1) to remove insoluble materials anduently used for the determination of glucose and freegen (FAN) content. FRW hydrolysates used as fermen-a were initially lter-sterilised using a 0.2 m lter unit

AS, Whatman Ltd.) and the pH of the hydrolysate was

ShaErlenmsynthecation,meansasks. nentiaused ipepton211C S180 5in a va(5 M) wneededa selec2004).

2.6. Bi

Fedtor (Nof 1.5 The pHNaOH.tial prerange at 20%at 1.0 vinitiatereducew/v) wup to were tactor tfor 10 mnatantphosphof TDWtions wtriplicaments

2.7. An

Gluconcenin 0.2 M(5%) wstarch suscepthe amthe ass

Prohydroland 55deneunder

Samwith evate th(9000 sk cultivations of L. starkeyi were carried out in 250 mL asks using 50 mL of either FRW hydrolysate or aedium. To ensure the reliability of microbial oil quanti-content of a whole ask was used for each sample. This

each fermentation required the preparation of up to 10 shake ask was inoculated with 1 mL of a 24 h expo--culture (around 13 107 cells per mL). The mediumse pre-cultures contained glucose, yeast extract andsks were incubated at 30 C in an orbital shaker (ZHWY-

Floor Model Incubator, PR China) at an agitation rate of. In all experiments it was desirable to maintain the pH

reater than 5.2, therefore an appropriate volume of KOHeriodically and aseptically added into the asks whensolved oxygen (DO) in shake asks was measured withelectrode as previously described (Papanikolaou et al.,

tor fermentations

h bioreactor cultures were carried out in a 3 L bioreac-runswick Scientic Co., USA) with a working volumeng FRW hydrolysates as the sole source of nutrients.s maintained at 6.0 0.1 by automatic addition of 5 M% (v/v) inoculum was employed using a 24 h exponen-ture. The agitation rate in the bioreactor was set in the0500 rpm in order to maintain the DO concentrationaturation. Aeration and temperature were maintainednd 30 C, respectively. The fed-batch fermentation was

batch mode and when the glucose concentration wasless than 20 g/L a concentrated glucose solution (60%,ded in the bioreactor. Similar additions were performed

fermentation when both TDW and lipid productionnated. Samples were taken periodically from the biore-ghout fermentation. They were centrifuged (9000 g

to separate yeast cells from the supernatant. The super- used for the analysis of glucose, FAN and inorganic

(IP), while the yeast biomass was used for the analysisdo-polysaccharides and intracellular lipids. Fermenta-carried out in duplicate and the respective analyses inata presented are the mean values of those measure-

ical methods

ylase activity was assayed by measuring the glucoseon produced during hydrolysis of 20 g/L (db) pure starchosphate buffer at pH 6.0 and 55 C within 15 min. TCAsed to stop the reaction (ratio 1:1). Gelatinisation ofensions at 75 C for 20 min was used to increase enzymety. One unit (U) of glucoamylase activity was dened ast of enzyme that releases 1 mg glucose in 1 min undernditions.

activity was measured by the production of FAN duringf 15 g/L (db) casein in 0.2 M phosphate buffer at pH 6.0ithin 30 min. One unit (U) of proteolytic activity was

the amount of enzyme that releases 1 g FAN in 1 mineaction conditions.

(2 mL) from enzymatic hydrolysis of FRW were mixedvolume of trichloroacetic acid (TCA, 5% w/v) to deacti-zymes and the mixture was subsequently centrifugedr 5 min). The supernatant was used for the analysis of

S. Tsakona et al. / Journal of Biotechnology 189 (2014) 3645 39

glucose, FAN and IP. The analytical methods for FAN, total Kjel-dahl Nitrogen (TKN) and IP were reported by Kachrimanidou et al.(2013). Glucose was analysed using a High Performance LiquidChromatography unit (Waters 600E) equipped with an AminexHPX-87H coferential refas follows: ow rate 0.

TDW washake ask bial oil analcells and eFolch (1957latter case, 500 mg of d1 h. Cellular10 min) andmixture at aoration, the

The analried out thfollowing a methoxide methanol. Tfatty acids ibial oil hydrGas Chromacolumn (60carrier gas with the injidentied b

Quantialso carriedmass was itioned prevremoved bypH 7.0 withThe concenof 3,5-dinitrglucose equ

3. Results

One of thbioprocess relying entiwheat ourFlour contaysed into gas amino acmentation malso used fophosphorusand by-proding the requmicrobial bvia SSF usinganism for (Koutinas ewere subsecomponentgal biomasshydrolysateprocess as cthe fact tha

0

50

100

150

200

250

14

16

18

75%70%65%60%55%50%

Pro

teo

lytic

activ

ity (U

/g)

(U

/g)

Moisture content (%)

lycoamylase () and proteolytic () activities at 48 h SSF carried out by A.i on wheat milling by-products using different initial moisture contents.

l treatment and for this reason the gelatinisation step couldtted.

timisation of solid state fermentation

imismylat thend inn usi/g) aial SSpresemoisratioois

d in attrties or (Lo

2 shc (23s usit of

0

100

200

150100500

Pro

teo

lytic

activ

ity (U

/g)

Time (h)

lycoamylase () and proteolytic () activities produced during SSF of A.i cultivated on wheat milling by-products with an initial moisture contentlumn (300 mm 7.8 mm, Bio-Rad, CA, USA) and a dif-ractometer (RI Waters 410). Operating conditions weresample volume 20 L; mobile phase 0.005 M H2SO4;6 mL min1; column temperature 65 C.s determined by drying the yeast biomass produced inand bioreactor fermentations at 105 C for 24 h. Micro-ysis was carried out by disrupting the dry mass of yeastxtraction of oil using either the method proposed by) or heat treatment with HCl (Tapia et al., 2012). In thehydrochloric acid (6 mL of 4 M HCl) was added for eachry yeast mass and the mixture was heated at 80 C for

debris was removed by centrifugation (9000 g, 4 C, the lipids were extracted with a chloroform/methanol

ratio of 2:1 (v/v). Prior to solvent separation via evap- extracted oil was washed with 0.88% KCl (w/v).ysis of fatty acid composition of microbial oil was car-rough production of fatty acid methyl esters (FAME)two-step reaction using transesterication with sodium(MeONa) followed by HCl catalyzed esterication withhe latter stage was followed in order to esterify freento FAME that may have been produced due to micro-olysis by intracellular lipase. FAME were analysed by atography Fisons 8060 unit equipped with a chrompack

m 0.32 mm) and a FID detector. Helium was used as(2 mL min1). The analysis was carried out at 200 Cection at 240 C and the detector at 250 C. FAME werey reference to standards.cation of total intra-cellular polysaccharides (IPS) was

our during fed-batch bioreactor cultures. The dry yeastnitially disrupted by heat treatment with HCl as men-iously in the case of lipid analysis. Cellular debris was

centrifugation and the supernatant was neutralised to 2 M NaOH. The nal volume was adjusted to 20 mL.tration of total sugars was determined with the assayosalicylic acid (Miller, 1959) and IPS were expressed asivalents.

and discussion

e targets of this study is the development of a two stagefor the production of a generic fermentation feedstockrely on FRW and by-product streams generated from

mills, bakeries and confectionery production processes.ins mainly starch and protein that should be hydrol-lucose and directly assimilable nitrogen sources suchids and peptides in order to create highly efcient fer-edia. WMB is a suitable substrate for SSF. It could ber the generation of various micronutrients including

and other minerals. The combination of these wasteuct streams results in a self-sustained process generat-ired hydrolytic enzymes and nutrient supplements forioconversions. The hydrolytic enzymes were producedg the fungal strain A. awamori, a well-known microor-the production of glucoamylase, protease and phytaset al., 2005, 2007; Wang et al., 2009). These enzymesquently employed to hydrolyse FRW and remainings (e.g. phytic acid) from WMB. Simultaneous lysis of fun-

contributed to enhanced nutrient release in the nal (Koutinas et al., 2005). One major advantage of thisompared to conventional starch or our hydrolysis ist the majority of FRW streams are already subjected to

0

2

4

6

8

10

12

Glu

co

am

yla

se a

ctivity

Fig. 1. Gawamor

thermabe omi

3.1. Op

Optglucoapresen75%) aductio(16.6 Uan initiment initial SSF duinitial mresultemainlypropertransfe

Fig.teolytiprocesconten

0

5

10

15

20

25

Glu

co

am

yla

se a

ctivity (

U/g

)

Fig. 2. Gawamorof 65%.ation of SSF by A. awamori aimed to maximise primarilyse and subsequently proteolytic activity. Figs. 1 and 2

effect of initial moisture content (50, 55, 60, 65, 70 andcubation duration on glucoamylase and protease pro-ng WMB as the sole substrate. Maximum glucoamylasend proteolytic (208.6 U/g) activities were achieved atF moisture content of 65%. The duration of each exper-nted in Fig. 1 was 48 h. Enzyme activities at different

ture contents were also measured in shorter and longern and the results observed showed the same optimumture content. Further increase in initial moisture contenta signicant decrease in both enzyme activities, a factibuted to the interference of moisture on the physicalf the solid particles which could cause limited oxygennsane et al., 1985).ows that maximum glucoamylase (17.4 U/g) and pro-3.6 U/g) activities were achieved at 60 h during an SSFng only WMB as solid substrate with an initial moisture65%. At prolonged SSF duration, glucoamylase activity

300

40 S. Tsakona et al. / Journal of Biotechnology 189 (2014) 3645

was gradually reduced, whereas proteolytic activity remained con-stant. A set of experiments was also carried out using various ratiosof WMB to FRW streams generated from a process line producingfood for infants. Enzyme production was reduced with increasingcontent of Flow particlecited studiewaste as suawamori (W2014). Wanand waste glucoamylausing the sato approximlower than duction, thewaste breastudy (Wanduction of protease dusolid substrprocess usi(38.8%) of thmilling by-253.7 U/g oof enzyme tafter 10 dasolid substrbakery wasduction of gA. awamori

In subseysis was cawith initial

3.2. Optimi

FRW hydwith unprorated fungacomplex hyof FRW masis (causedautolysis ofvaluable nu2005). The sis was opticoncentrati

Fig. 3 pruncontrolleproduced dtion. All exwas carriedthe potentiin Fig. 3, thuncontrolleuncontrolle5.0. For thisout pH contbe develope

Fig. 4 pat seven tusing uncotion of 14activities otemperatur

0

50

100

150

200

250

300

350

400

0

0

0

0

0

100

120F

AN

(mg

/L)

ion (

g/L

)

pH

ffect of pH on glucose () and FAN () production during hydrolysis ofinitial FRW concentration using crude enzymes produced via SSF of A.i.

of A. awamori strain are 60 and 55 C, respectively (Wang2009). Hydrolysis at 60 C resulted in maximum glucosetion, whereas similar FAN production was achieved at atempigniespevalue cont 55

C a sed (1t expaneot-ricnt nustritionzym

costourysis ctionnow

0

50

100

150

200

250

300

350

65605550454030

FA

N (m

g/L

)

Temperature (oC)

ffect of temperature on glucose () and FAN () production during hydrol-40 g/L initial FRW concentration with crude enzymes produced via SSF of A.i.RW (data not shown). This could be attributed to the size of this particular FRW stream. Previous literature-s have indicated wheat pieces, pastries and mixed fooditable substrates for multi-enzyme production by A.ang et al., 2007, 2009; Lam et al., 2013; Pleissner et al.,g et al. (2009) reported the utilisation of wheat piecesbread for the production of 81.3 U/g and 78.4 U/g ofse activities, respectively, after more than 130 h of SSFme strain of A. awamori. It should be stressed that upately 50 h of SSF the production of glucoamylase was

5 U/g (Wang et al., 2009). Contrary to glucoamylase pro- proteolytic activity produced from wheat pieces and

d was signicantly lower (less than 80 U/g) than thisg et al., 2009). Wang et al. (2007) reported the pro-almost 40 U/g of glucoamylase activity and 20 U/g ofring SSF of A. awamori when wheat bran was used asate which was produced by a simplied wheat millingng a laboratory mill. However, the high starch contentis bran-rich stream is not common in industrial wheat

products. Lam et al. (2013) reported the production off glucoamylase (the activity was dened as the amounthat releases 1 mol of glucose per minute of reaction)ys of SSF of A. awamori using pastry waste as the soleate. Pleissner et al. (2014) reported the valorisation ofte from canteens (cakes, pastry and bun) for the pro-lucoamylase and proteases using the fungal strains ofand Aspergillus oryzae.quent experiments, enzyme production for FRW hydrol-rried out at 60 h SSF using WMB as the sole substrate

moisture content of 65%.

sation of FRW hydrolysis

rolysis was carried out by mixing macerated SSF solidscessed FRW. The addition of crude enzymes and mace-l mass into the hydrolysis mixture denoted the onset ofdrolytic reactions that involved not only the hydrolysiscromolecules, but also the hydrolysis or even autoly-

by oxygen depletion) of macerated fungal cells. The the fungus is reported to encourage the regeneration oftrient components from fungal biomass (Koutinas et al.,production of glucose and FAN during FRW hydroly-mised regarding reaction temperature, pH, initial FRWon and initial activities of glucoamylase and protease.esents the effect of pH (3.5, 4, 4.5, 5, 5.5, 6, 6.5 andd pH) on the maximum glucose and FAN concentrationuring hydrolysis of 140 g/L (wb) initial FRW concentra-periments were carried out at 55 C. One experiment

out under uncontrolled pH conditions to evaluateal of avoiding pH control during hydrolysis. As showne maximum glucose production was achieved underd pH conditions. The FAN concentration achieved underd pH conditions was similar to the one produced at pH

reason, subsequent experiments were carried out with-rol as in this way a more cost-competitive process couldd.resents the glucose and FAN production achievedemperatures (30, 40, 45, 50, 55, 60 and 65 C)ntrolled pH conditions, an initial FRW concentra-0 g/L (wb) and initial glucoamylase and proteasef 0.24 U/mL and 4.3 U/mL, respectively. The optimumes for glucoamylase and protease produced by this

2

4

6

8

Glu

co

se c

on

ce

ntr

at

Fig. 3. E140 g/L awamor

strain et al., producwider were s60 C, rimum glucosthan aat 60

achievsequensimultnutriensufcie

Indproduccial enenergycereal hydrolliquefation is

0

20

40

60

80

100

120

140

Glu

co

se

co

nce

ntr

atio

n (

g/L

)

Fig. 4. Eysis of 1awamorerature range (4555 C). Glucose and FAN productioncantly reduced at temperatures higher than 55 C andctively. Both glucose and FAN production reached max-es at approximately 24 h hydrolysis. Higher starch toversion yield can be achieved at 60 C (90.1%) ratherC (86.6%). However, if hydrolysis of FRW is carried outignicantly lower TKN to FAN conversion yield will be3.8% at 55 C and 7.3% at 60 C). For this reason, all sub-eriments were carried out at 55 C in order to achieve

usly sufcient starch and protein hydrolysis. Producingh fermentation feedstocks requires the production ofitrogen sources besides glucose.al hydrolysis of starch or whole cereal our for the

of fermentation media is carried out using commer-es at high initial concentrations in order to minimises. Up to 300 g/L and 350 g/L are usually used for whole

and starch hydrolysis, respectively. Industrial starchis carried following a two stage process based on/gelatinisation and saccharication. Starch liquefac-adays carried out simultaneously with gelatinisation at

400

S. Tsakona et al. / Journal of Biotechnology 189 (2014) 3645 41

Table 1Starch and protein (expressed as TKN) conversion yields achieved during batch and fed-batch hydrolysis experiments using various initial FRW concentrations and fourinitial glucoamylase and protease activities produced via SSF of A. awamori cultivated on WMB.

FRW concentration (g/L) Starch to glucose conversion yield (%) TKN to FAN conversion yield (%)

Glucoamylase activity (U/mL)

0.24 0.49 0.73 0.97

Batch hydrolysis98 89.4 0.3 93.2 0.1 93.1 0.2 92.8 0.3140 86.6 0.6 90.1 0.3 91.9 0.3 93.1 0.1 175 85.1 0.2 88.6 0.6 89.7 0.5 92.9 0.4 205 82.4 0.5 86.0 0.3 89.0 0.3 93.0 0.2 235 68.0 0.4 70.2 0.6 73.5 0.2 84.1 0.6

Fed-batch hydrolysis235 92.1 0.2 98.2 0.4 300 86.1 0.1 350 73.4 0.2

temperaturwhereas stglucoamylaefcient hyavoiding thas well as tgelatinisati

Table 1 (98, 140, 1lase (0.24, 0and 16.13 Uconversion adding theture and thwere carriebe easily obhydrolytic etein (expresthe case of in crude enstarch to glsation of FRhydrolysis ythe high vishydrolysis starch to gTable 1 thaconversion compared t

The stardicted by aexpressing yield versustial concentpresented iof the linea

Table 2Linear equatioversus glucoamFRW concentr

Initial FRW c

234 205 175 140 98

pectr thuld gluco

e(%)

ch opn inie andg/L,

correr FA

pretionsns. Ts esse ml appty ofd thbatcto glu

hyded. T

exp). Sated atningt seeher Fig. 5bes up to 105 C using thermostable bacterial -amylase,arch saccharication to glucose is carried out usingse and pullulanase at 60 C. This study aimed to achievedrolysis of similar FRW concentrations in a single stagee use of high temperatures and commercial enzymeshe application of a separate stage for liquefaction andon.presents the effect of ve initial FRW concentrations75, 205 and 235 g/L, wb) and four initial glucoamy-.49, 0.73 and 0.97 U/mL) and protease (4.03, 8.06, 12.9/mL) activities on starch to glucose and TKN to FANyields. These hydrolytic reactions were carried out by

SSF solids containing the crude multi-enzyme mix-e FRW at the beginning of the process. All experimentsd out at 55 C and uncontrolled pH conditions. It canserved that increased initial enzyme activities in batchxperiments led to increased starch to glucose and pro-sed as TKN) to FAN conversion yields. For instance, inFRW initial concentration of 235 g/L, a 4-fold increasezyme activities resulted in substantial improvement ofucose conversion yield from 68% to 84% (w/w). Utili-W concentrations higher than 235 g/L resulted in lowields with simultaneous operational problems due tocosity of the suspension. It should be stressed that FRWat 60 C would result in approximately 3.55% higherlucose conversion yields than the values reported int were achieved at 55 C. However, the TKN to FANyields achieved at 60 C were more than 50% lowero the respective yields achieved at 55 C.ch to glucose conversion yields (YGlucose) could be pre-

simple empirical Eq. (1). Initially, linear equationsthe relationship between starch to glucose conversion

the initial glucoamylase activity (GA) at different ini-rations of FRW were plotted (Table 2) using the resultsn Table 1. Subsequently, the set of coefcients and

the resused fothat coinitial

YGlucos

Batwhen aglucos937.2 mtrationA highproteinmentasolutiosible iproducationaquantiysis anas fed-starch at 24 hwas usin this(Fig. 5achievmentiodid nono higysis (Fr equations presented in Table 2 were associated with

ns (YGlucose = GA + ) expressing starch to glucose conversion yieldylase activity (0.24, 0.49, 0.73 and 0.97 U/mL) at ve different initial

ations.

oncentration (g/L) Coefcients of linear equation

22.234 62.99314.817 7810.304 82.998.24 85.494.978 88.97

sufcient torespondingto FAN con803 mg/L). Tand WMB htion (228.9 300 g/L (corsimilar to cohydrolysis omaximum cing on the wcanteens wcessively adfrom an initProteolytic activity (U/mL)

4.03 8.06 12.9 16.13

18.2 0.8 27.2 0.8 42.1 1.3 42.0 1.113.8 1.0 25.9 1.2 38.7 1.2 41.9 0.911.7 1.2 22.7 0.9 35.6 0.9 42.1 1.210.1 0.5 19.9 1.1 30.4 0.6 41.2 0.68.4 1.1 17.3 0.9 29.8 1.2 38.5 0.3

17.3 1.1 30.8 1.0 24.9 0.6 22.7 0.7

ive initial FRW concentration (CFRW). Eq. (1) could bee calculation of the starch to glucose conversion yieldbe achieved at various initial FRW concentrations andamylase activities.

= (0.1148 CFRW 7.874) GA+ (0.0019 C2FRW + 0.456 CFRW + 61.97) (1)

eration led to satisfactory starch and protein hydrolysistial FRW concentration up to 205 g/L (wb) was used. The

FAN concentrations produced were up to 168.9 g/L and respectively. It should be stressed that this FAN concen-esponds only to the FAN produced from FRW hydrolysis.N concentration was produced due to hydrolysis of

sent in WMB. However, the operation of fed-batch fer- requires the utilisation of highly concentrated feedinghus, generating as concentrated hydrolysates as pos-ential in order to minimise energy costs. In order toore concentrated FRW hydrolysates, a different oper-roach was followed in which half of the crude enzyme

known activity was added at the beginning of hydrol-e rest after 6 h (this mode of operation is designatedh hydrolysis in this study). Through this method, thecose conversion yield was increased up to 98.2% (w/w)rolysis when an initial FRW concentration of 235 g/Lhe total glucoamylase and protease activities employederiment were 0.73 U/mL and 12.9 U/mL, respectivelyisfactory starch and protein hydrolysis could be also

initial FRW concentrations of 300 g/L (Table 1). Worth is the fact that this gradual addition of crude enzymesm to positively affect protein conversion yield, sinceAN production was noted compared to batch hydrol-). As shown in Fig. 5, hydrolysis duration of 24 h was achieve 98.2% starch to glucose conversion yield (cor- to a glucose concentration of 204.5 g/L) and 30.8% TKNversion yield (corresponding to a FAN concentration ofhe nal FAN concentration corresponding to both FRWydrolysis was 1142 mg/L. A higher glucose concentra-g/L) was achieved when an initial FRW concentration ofresponding to 86.1% conversion yield) was used that isncentrations used in traditional industrial processes forf whole cereal ours. Pleissner et al. (2014) reported aonversion yield of starch to glucose of 8090% (depend-aste) during hydrolysis of bakery and food waste from

hen solid mashes of A. awamori and A. oryzae were suc-ded. A glucose concentration of 143 g/L was producedial food waste concentration of 432 g/L containing 0.36 g

42 S. Tsakona et al. / Journal of Biotechnology 189 (2014) 3645

0%

20%

40%

60%

80%

100%

Sta

rch

to

glu

co

se

co

nve

rsio

n y

ield

(%

)

0%

10%

20%

30%

0

TK

N to

FA

N c

on

ve

rsio

n y

ield

(%

)

Fig. 5. Compafed-batch () hlase and prote

starch per gcose conver

3.3. Shake production

Major oFRW hydroof lipids bytions were cor glucose-sole sourceto-FAN ratithe ratio of (311 g/g). Tto the carbexpressed a

Fig. 6 prconsumptiosynthetic mglucose conThe initial The initial CIn both casinoculationsubstantiall

30

35

100

120(a)

(b)

0

20

40

60

80

Glu

co

se

(g

/L)

20

40

60

80

100

Glu

co

se

(g

/L)30252015105

Hydrolysis time (h)

rison of starch (a) and TKN (b) conversion yields during batch () andydrolysis of 235 g/L initial FRW concentration using a total glucoamy-ase activities of 0.73 U/mL and 12.9 U/mL, respectively.

solids corresponding to an approximate starch to glu-sion yield of 0.84 g/g (Pleissner et al., 2014).

ask cultures of L. starkeyi for microbial lipid

bjective of the present study was the evaluation oflysates as suitable feedstock for microbial production

the oleaginous yeast L. starkeyi. Preliminary fermenta-arried out in shake asks using either FRW hydrolysatesbased synthetic media. FRW hydrolysates served as the

of carbon, nitrogen and micronutrients. The carbon-o of the synthetic medium was adjusted according tothe FRW hydrolysate used in shake ask fermentationshe C/FAN ratio of the FRW hydrolysate correspondedon content of the glucose and the nitrogen contents FAN concentration in the hydrolysate.esents the production of TDW and lipids as well as then of glucose during cultivation of L. starkeyi on eitheredium (Fig. 6a) or FRW hydrolysate (Fig. 6b). The initialcentration was around 106 and 109 g/L, respectively.FAN concentration of the hydrolysate was 140 mg/L./N or C/FAN ratio for both fermentations was 311 g/g.

es, fermentation was completed in around 200 h after. However, TDW synthesis and lipid accumulation werey enhanced when FRW hydrolysate was employed as

0

0

Fig. 6. Consumbial lipids () (b) lter-steril

the sole su30.5 g/L witwas used. were 21.6 gonset of lipimedium, apThese resulnutrient-costarkeyi grocontain a bathat favour

The esseof cellular lnitrogen-lim2011; Ratlegen depleti(ICDH) enzlipid biosynence of orgaof oleaginouand Ratledinuenced contained result of dRatledge, 19tivation on 0

5

10

15

20

25

TD

W a

nd

Lip

ids (g

/L)

(a)

10

15

20

25

30

TD

W a

nd

Lip

ids (g

/L)

(b)0

5

2001601208040

Fermentation time (h)

ption of glucose () and production of total dry weight () and micro-during shake ask cultures of L. starkeyi on (a) synthetic medium andised FBW hydrolysate.

bstrate (Fig. 6). Specically, maximum TDW reachedh a lipid content of 40.4% (w/w) when FRW hydrolysateThe respective values achieved in synthetic medium/L and 34.8% (w/w), respectively. In both cases, thed production occurred after FAN was depleted from theproximately after 48 h of inoculation (data not shown).ts indicate the suitability of the produced feedstock as amplete fermentation feedstock that can sustain both L.wth and lipid accumulation. FRW hydrolysates seem tolanced composition of glucose, FAN and other nutrients

microbial lipid production.ntial condition that triggers the de novo accumulationipids in oleaginous microorganisms is the presence ofited cultivation conditions (Papanikolaou and Aggelis,

dge and Wynn, 2002). In such case, extracellular nitro-on results in the inhibition of isocitrate dehydrogenaseyme that regulates carbon excess channelling towardsthesis (Ratledge and Wynn, 2002). However, the pres-nic nitrogen can favour lipid accumulation in a numbers microorganisms, including strains of L. starkeyi (Evans

ge, 1984). The rate of microbial lipid production isby the products of catabolism of the nitrogen sourcein the fermentation medium, rather than being theirect stimulation by the nitrogen source (Evans and84). Thus, the enhanced lipid accumulation during cul-

FRW hydrolysates could be attributed to the presence of

S. Tsakona et al. / Journal of Biotechnology 189 (2014) 3645 43

250

300

0

20

40

60

80

100

Glu

co

se

(g

/L)

(a)

0

20

40

60

80

100

0

TD

W a

nd

Lip

ids (

g/L

)

Fig. 7. Consum(), endo-polysolution () oL. starkeyi usin

complex orby the micr

3.4. Fed-bahydrolysate

Althoughpromising, hydrolysatecultures. Foin a lab-scalipid accumused was lFAN concenture was inglucose solubelow 20 gand IP conpolysacchar

The inocwas produceffect on yfermentatiocommerciaefciency wis a critical f

producing the inoculum in FRW hydrolysate, the cells were suf-ciently adjusted to the culture environment and in the same timea signicantly high concentration of yeast cells was achieved. Itwas observed that cultivating the inoculum in FRW hydrolysates a

yeasl nuantlte wd celmend acc

depainedtatiod lipfcieion erme

h. Liols. F

undeatmresul

This dur

of fous lytic

xtracd pro0

50

100

150

200

FA

Nand IP

(mg/L

)

4

8

12

16 En

do

-po

lysa

cch

arid

es (g

/L)

(b)

highermerciasigniction raage anand fer

LipialmostIP remfermeninducewas suutilisatactor fat 175protoc(1957)heat tr2012) tation.causedreleaseoleaginthe analipid eassiste0

2001601208040

Fermentation time (h)

ption of glucose (), FAN () and IP () and production of TDWsaccharides () and lipids extracted either with chloroform:methanolr with HCl () during lter-sterilised fed-batch bioreactor culture ofg FRW hydrolysate as the sole fermentation medium.

ganic nitrogen sources, which was directly assimilableoorganism.

tch bioreactor culture of L. starkeyi on FRW

the results of shake ask fermentations were veryit was essential to verify the suitability of FRWs as medium for microbial oil production in bioreactorr this reason, a fed-batch fermentation was carried outle bioreactor aiming to enhance microbial growth andulation by L. starkeyi. The initial fermentation mediumter-sterilised FRW hydrolysate with initial glucose andtrations of 90 g/L and 180 mg/L, respectively. The cul-itiated in batch mode and feeding of a concentratedtion started when the glucose concentration dropped

/L. Fig. 7 presents the prole change of glucose, FANsumption as well as TDW, microbial oil and endo-ide production during fermentation.ulum employed in fed-batch cultivations with L. starkeyied in FRW hydrolysates. This approach had a profoundeast growth and fermentation efciency. Fed-batchns were also carried out with inocula cultivated inl nutrient supplements and the obtained fermentationas rather low. The physiological state of the inoculumactor in determining the course of the fermentation. By

low lipid cothe presentcell densitythe sufcieof microbiadevelopmeliterature-cinvolving hsurements 2012; Xu etby Tapia etwas 64.5 g/of 58.7% (wcontent of 6

During treached a plation of eare regardemushroomproteoglycaammatoryOn a biocheformation atration of v(i.e. nitrogeoccurs, a rathe functioaccumulatiexcreted toCL) in oleagenerate ce2002). In thaccumulateculture mefructokinasof 6-phospt cell concentration than the one produced with com-trient supplements could be achieved. In this way, ay lower lag phase and higher growth and lipid produc-ere achieved. Lin et al. (2011) reported that inoculuml density play a signicant role on lipid accumulationtation efciency by L. starkeyi AS 2.1560.umulation started at approximately 40 h when FAN wasleted from the fermentation broth. The concentration of

constant at approximately 30 mg/L until the end of then indicating that IP was not the limiting nutrient that

id accumulation. The initial IP concentration of 161 mg/Lnt for microbial growth. This also indicates that the

of WMB provides adequate quantities of IP for biore-ntations. The maximum TDW of 113 g/L was observedpid accumulation was determined with two analyticalig. 7 demonstrates that the analytical protocol of Folcherestimates lipid accumulation. The protocol involvingent with HCl prior to solvent extraction (Tapia et al.,ted in higher lipid concentrations throughout fermen-

could be attributed to the sufcient cellular breakageing HCl heat treatment, which facilitated an improvedatty acids. Due to the presence of thick cell wall, someyeasts are often resistant to solvent treatments. Besides,al protocol of Folch (1957) was originally described fortion from animal tissue and, similarly to other solvent-tocols, its efciency is best validated in the presence ofntents in the biological sample. Based on the results of

study, the applied protocol for lipid extraction in high cultures of oleaginous microorganisms should ensurent disruption of the cell wall. The successful recoveryl lipid constitutes an issue of major importance for thent of viable microbial oil production processes. Recentited publications have also reported that the methodeat treatment with HCl result in more realistic mea-of lipid accumulation (Tapia et al., 2012; Gong et al.,

al., 2012). According to the analytical method reported al. (2012), the maximum lipid concentration achievedL at 160 h corresponding to an intracellular oil content/w) and a productivity of 0.4 g/L/h. The maximum oil1.5% (w/w) was observed at 148 h.he nal 40 h of the fed-batch culture, lipid productionlateau, while TDW increase was due to the accumu-ndo-polysaccharides (Fig. 7). Endo-polysaccharidesd as bioactive compounds and numerous types of

-derived polysaccharides (glucans, galactomannans,ns) have been reported to possess antitumor, antin-

and immunomodulating properties (Wasser, 2002).mical level, the events that lead to endo-polysaccharidere related with a transition in the intracellular concen-arious metabolites, triggered after nutrient limitationn) in the culture medium. When nitrogen depletionpid decrease of intracellular AMP causes alterations inn of the Krebs cycle and results in intra-mitochondrialon of citric acid, which after exceeding a critical value, is

the cytoplasm. The presence of ATP-citrate lyase (ATP-ginous microorganisms will cleavage citric acid andllular fatty acids via -oxidation (Ratledge and Wynn,e absence of ATP-CL enzymatic complex, citric acid isd in the cytoplasm and will be either excreted into thedium, or will provoke the inhibition of 6-phosphoro-e (PFK). The latter, in combination with the activityhoglucose isomerase (PGI) will result in intracellular

44 S. Tsakona et al. / Journal of Biotechnology 189 (2014) 3645

Table 3Fatty acid content (%, w/w) of total cellular lipids produced by L. starkeyi during fed-batch cultivation on FRW hydrolysates.

Cultivation time (h) C14:0 C16:0 9C16:1 C18:0 9C18:1 9,12C18:2 9,12,15C18:3 C22:0 Other

48 0.8 24.4 8.4 52.9 11.3 2.2 66 1.0 25.8 7.4 53.1 10.6 2.1 86 1.4 27.6 0.5 7.6 52.4 9.1 1.4

112 0.3 33.1 1.2 6.4 50.6 4.4 0.6 2.3 1.1136 1.2 27.1 0.6 8.7 52.9 8.5 0.3 0.3 0.4

accumulation of polysaccharides (Papanikolaou and Aggelis, 2011).In the present study, the accumulation of endo-polysaccharidesat the late stages of the fermentation could be attributed to thesaturation of ATP-CL enzyme, an event which in turn channelledglucose catabolism towards fructose 6-phosphate pathway.

Table 3 presents the fatty acid composition of the produced oilat different cultivation times. The microbial lipids produced by L.starkeyi mainly contained oleic acid (9C18:1) followed by palmiticacid (C16:0), linoleic acid (9,12C18:2) and stearic acid (C18:0). Thedistribution of oleic acid, palmitic acid and linoleic acid remainedat similar levels throughout the process. The content of saturatedfatty acids was higher than 35% (w/w) of total fatty acid contentafter 86 h fermentation. The oleic acid content was always higherthan 50% (w/w). Considering the content of saturated fatty acids,the microbial oil produced by L. starkeyi is closer to palm oil. How-ever, an oleic acid content higher than 50% is usually observed inrapeseed oil and in some cases in palm oil. Similar distribution infatty acid composition has been reported for other yeast strains ofL. starkeyi and Rhodosporidium toruloides (Li et al., 2007; Zhao et al.,2008; Lin et al., 2011).

Table 4 presents literature-cited results regarding microbial oilproductionsources. Momineral-baof glucose, for microbi2.1560, wittation of msame strain

containing a mixture of glucose and xylose (Zhao et al., 2008).Furthermore, mixtures of xylose and glucose were employed infed-batch cultures for the production of 88.7 g/L of TDW with amicrobial oil content of 55.2% (w/w) when a mutant strain of L.starkeyi DSM 70296 was used (Tapia et al., 2012). Lin et al. (2011)has reported the highest TDW of 104.6 g/L and lipid content of64.9% (w/w) at an overall productivity of 1.2 g/L/h when glucoseand mineral media were used in fed-batch bioreactor cultures witha L. starkeyi strain obtained from the China Microbiological Cul-ture Collection Center. The fermentation results reported in thisstudy are also promising and were achieved without optimisa-tion. Liu et al. (2012) have also reported the utilisation of a wastestream, monosodium glutamate wastewater, for the production ofmicrobial lipids using the wild-type strain L. starkeyi GIM2.142.Moreover, Huang et al. (2011) demonstrated the utilisation of sh-meal wastewater supplemented with glucose for microbial lipidproduction by L. starkeyi HL, leading to the production of 17.6 g/Lof TDW with an oil content of 15.3% (w/w).

Fed-batch cultivations offer the advantage of overcoming sub-strate inhibition obstacles and achieving high density microbialcultures and are often preferred in microbial lipid production stud-

articulatin of ttch cultuer sting n th

Table 4Literature-cite rious

L. starkeyi st TDW

AS 2.1560 20.5

DSM 7029513.3 9.3

AS 2.1560

27.9 23.8 22.3 31.5

25.5

04.6

DSM 70296(wild-type)

12.3 75.3

DSM 70296 mutant)

13.3 88.7

GIM2.142

HL

DSM 70296

a F: ask an by various L. starkeyi strains using different carbonst of the studies employed pure carbon sources and

sed media. Gong et al. (2012) investigated the suitabilityxylose, cellobiose and their mixtures as carbon sourcesal lipid production using the yeast strain L. starkeyi ASh the most promising results achieved during fermen-edia containing a mixture of cellobiose and xylose. The

accumulated 61.5% (w/w) of oil in fermentation media

ies. In paccumtivatiofed-bathese cin othpromisfocus o

d results on microbial lipid production by various L. starkeyi strains cultivated on va

rain Carbon source Nitrogen source Culture mode a

Glucose and xylose Variable F

Glucose Basal mediumFGlucose Waste water

Cellobiose

Mineral medium FGlucose Xylose Cellobiose andxyloseCellobiose, xyloseand glucoseGlucose Minimal medium FBB 1

XyloseMineral medium

F Glucose and xylose FBB

(A1 XyloseMineral medium

F Glucose and xylose FBB Glucose Monosodiumglutamate wastewater

F 4.6

Glucose Fishmeal wastewater F 17.6

Glucose Mineral medium F 21.6 Flour based industrialwaste streams

F 30.5 FBB 109.8

d FBB: fed-batch bioreactor.ular, Li et al. (2007) reported sufcient growth and lipidon (106.5 g/L and 67.5%, w/w, respectively) during cul-he yeast strain R. toruloides Y4 on glucose in pilot-scaleultures. The overall lipid productivity reported underre conditions was 0.54 g/L/h. From the results presentedudies, it becomes obvious that L. starkeyi is a highlystrain for microbial oil production. Future work wille optimisation of microbial oil production by L. starkeyi

carbon and nitrogen sources in ask or fed-batch bioreactor cultures.

(g/L) Lipid content(%)

Productivity(g/L/h)

Reference

61.5 0.1 Zhao et al. (2008)

56.3 0.03 Angerbauer et al.(2008)72.3 0.03

50 0.13

Gong et al. (2012)53 0.157 0.0855 0.12

52 0.12

64.9 1.2 Lin et al. (2011)

35 0.03

Tapia et al. (2012)45.4 0.2439.6 0.0355.2 0.2924.7 0.01 Liu et al. (2012)

15.3 0.01 Huang et al. (2011)

34.8 0.04 Present study40.4 0.0658.7 0.4

S. Tsakona et al. / Journal of Biotechnology 189 (2014) 3645 45

cultivated in various industrial waste streams. Moreover, the dif-ferent stages of primary and secondary metabolism of the strainshould be studied, in order to achieve high-cell density culture inthe rst stage of the fed-batch culture that will offer high oil yieldsduring lipid accumulation phase.

4. Conclusions

A novel tionery indwere successerve as a gThe integratries could streams, thcould be uoleochemic

Acknowled

This wo(09SYN-32-erence Fram(Greek Minand Commis also incluvalorisation

References

Anonymous, 2RBC820-00

Anonymous, nomenclatStrLangua2014).

Anonymous, 2prnewswi18599981

Angerbauer, Csewage slusour. Tech

Bttcher, D., Scand stereoProtein En

Buchholz, K., Bmarks and

Evans, C.T., Ration by Rh

Folch, J., Lees, puricatio

Gong, Z., Wanlobiose an117, 2024

Huang, L., Zhaa potentiaEnviron. T

KachrimanidoKookos, I., biodiesel pWaste Bio

Koutinas, A.A.Papanikolaand by-prbiopolyme

Koutinas, A.A.,Design andable resou

Koutinas, A.A.Luque, R., cesses and177198.

Koutinas, A.A., Arifeen, N., Wang, R., Webb, C., 2007. Cereal-based biorenery devel-opment: integrated enzyme production for cereal our hydrolysis. Biotechnol.Bioeng. 97, 6172.

Koutinas, A.A., Wang, R.H., Webb, C., 2005. Development of a process for the pro-duction of nutrient supplements for fermentations based on fungal autolysis.Enzyme Microb. Technol. 36, 629638.

Lam, W.C., Pleissner, D., Lin, C.S.K., 2013. Production of fungal glucoamylase forglucose production from food waste. Biomolecules 3, 651661.

Leiva-Candia, D.E., Pinzi, S., Redel-Macas, M.D., Koutinas, A., Webb, C., Dorado, M.P.,. The he prohao, Zoridiu317.., PfaKoutinamedhe propectivhen, HipomyechnoYue, Qosodi. Biore, B.K., cts of, J.O., Bical a

echno.L., 19rs. AnV., M

toux, study://ec.e/2014., Gou

ion bioolaou,lis, G.e yeaolaou,ngle colaou,

for t654.

olaou,tierella872

., Barthtica308

r, D., tationesour., C., Wation i., Gonvation9.E., Ans

of lipagene, G., B

., Torreleagin., Codrovingentat., Ji, Ye ethmE Pa

S.P., 2ion poao, Xoducts and

J. 65, , Konguctiont Lipobiorenery and bioprocess concept, based on confec-ustry wastes was developed in the present study. FRWsfully converted into a rich nutrient medium that couldeneric feedstock for the production of microbial lipids.tion of such a biorenery concept into food indus-promote the efcient valorisation of generated wasterough their bioconversion into microbial lipids thatsed as feedstock for the production of biodiesel andals.

gments

rk is funded by the research project NUTRI-FUEL621), implemented within the National Strategic Ref-ework (NSRF) 20072013 and co-nanced by National

istry General Secretariat of Research and Technology)unity Funds (E.U. European Social Fund). This workded in the Cost Action TD1203 entitled Food waste

for sustainable chemicals, materials & fuels (EUBis).

011. Reducing household bakery waste. Wrap report, Project code3.

2014a. PRODCOM List 2013. http://ec.europa.eu/eurostat/ramon/ures/index.cfm?TargetUrl=LST NOM DTL&StrNom=PRD 2013&geCode=EN&IntPcKey=&StrLayoutCode=HIERARCHIC (visited: 03/

014b. Global Markets for Oleochemical Fatty Acids. http://www.re.com/news-releases/global-markets-for-oleochemical-fatty-acids-2.html (visited: 03/2014).., Siebenhofer, M., Mittelbach, M., Guebitz, G.M., 2008. Conversion ofdge into lipids by Lipomyces starkeyi for biodiesel production. Biore-nol. 99, 30513056.hmidt, M., Bornscheuer, U.T., 2009. Alteration of substrate specicityselectivity of lipases and esterases. In: Lutz, S., Bornscheuer, U.T. (Eds.),gineering Handbook. Wiley-VCH, Weinheim, Germany, pp. 753775.ornscheuer, U.T., 2005. Highlights in biocatalysis historical land-

current trends. Eng. Life Sci. 5, 309323.tledge, C., 1984. Inuence of nitrogen metabolism on lipid accumula-odosporidium toruloides CBS 14. J. Gen. Microbiol. 130, 17051710.M., Sloane-Stanley, G.H., 1957. A simple method for the isolation andn of total lipids from animal tissues. J. Biol. Chem. 226, 497509.g, Q., Shen, H., Hu, C., Jin, G., Zhao, Z.K., 2012. Co-fermentation of cel-d xylose by Lipomyces starkeyi for lipid production. Bioresour. Technol..ng, B., Gao, B., Sun, G., 2011. Application of shmeal wastewater asl low-cost medium for lipid production by Lipomyces starkeyi HL.echnol. 32, 19751981.u, V., Kopsahelis, N., Chatzifragkou, A., Papanikolaou, S., Yanniotis, S.,Koutinas, A.A., 2013. Utilisation of by-products from sunower-basedroduction processes for the production of fermentation feedstock.

mass Valor. 4, 529537., Vlysidis, A., Pleissner, D., Kopsahelis, N., Garcia, I.L., Kookos, I.K.,ou, S., Kwan, T.H., Lin, C.S.K., 2014a. Valorization of industrial waste

oduct streams via fermentation for the production of chemicals andrs. Chem. Soc. Rev. 43, 25872627.

Chatzifragkou, A., Kopsahelis, N., Papanikolaou, S., Kookos, I.K., 2014b. techno-economic evaluation of microbial oil production as a renew-rce for biodiesel and oleochemical production. Fuel 116, 566577., Papanikolaou, S., 2011. Biodiesel production from microbial oil. In:Campelo, J., Clark, J.H. (Eds.), Handbook of Biofuels Production Pro-

Technologies. Woodhead Publishing Limited, Cambridge, UK, pp.

2014for t

Li, Y., Zdosp312

Lin, C.S.KJ.H., Mohfor tpers

Lin, J., Sby LBiot

Liu, J.X., monlipid

Lonsaneaspe

MetzgerchemBiot

Miller, Gsuga

Monier, Montoryhttp16/3

Naik, S.Nerat

PapanikAggein th

Papanikof si

Papanikused639

PapanikMor95, 2

Partt, Jquan3065

PleissnemenBior

Ratledgemul

Shen, Wculti858

Tapia, V.tionmut

VincenteR.Man o

Wang, RImpferm

Wang, RvativIChe

Wasser,ulat

Xu, J., ZhbyprloideEng.

Zhao, X.prodyeaspotential for agro-industrial waste utilization using oleaginous yeastduction of biodiesel. Fuel 123, 3342.., Bai, F., 2007. High-density cultivation of oleaginous yeast Rho-m toruloides Y4 in fed-batch culture. Enzyme Microb. Technol. 41,

ltzagraff, L.A., Herrero-Davila, L., Mubofu, E.B., Abderrahim, S., Clark,as, A.A., Kopsahelis, N., Stamatelatou, K., Dickson, F., Thankappan, S.,

, Z., Brocklesby, R., Luque, R., 2013. Food waste as a valuable resourceduction of chemicals, materials and fuels. Current situation and globale. Energy Environ. Sci. 6, 426464.., Tan, H., Zhao, X., Wu, S., Hu, C., Zhao, Z.K., 2011. Lipid productionces starkeyi cells in glucose solution without auxiliary nutrients. J.l. 152, 184188..-Y., Gao, B.-Y., Ma, Z.-H., Zhang, P.-D., 2012. Microbial treatment of theum glutamate wastewater by Lipomyces starkeyi to produce microbialsour. Technol. 106, 6973.Ghildyal, N.P., Budiateman, S., Ramakrishna, S.V., 1985. Engineering

solid state fermentation. Enzyme Microb. Technol. 7, 258265.ornshcheuer, U., 2006. Lipids as renewable resources: current state ofnd biotechnological conversion and diversication. Appl. Microbiol.l. 71, 1322.59. Use of dinitrosalicylic acid reagent for determination of reducingal. Chem. 31, 426428.udgal, S., Escalon, V., OConnor, C., Gibon, T., Anderson, G.,

H., Reisinger, H., Dolley, P., Ogilvie, S., Morton, G., 2010. Prepara- on food waste across EU-27. Technical Report 2010 054.uropa.eu/environment/eussd/pdf/bio foodwaste report.pdf (visited:).d, V.V., Rout, P.K., Dalai, A.K., 2010. Production of rst and second gen-fuels: a comprehensive review. Renew. Sust. Energ. Rev. 14, 578597.

S., Beopoulos, A., Koletti, A., Thevenieau, F., Koutinas, A.A., Nicaud, J.M.,, 2013. Importance of the methyl-citrate cycle on glycerol metabolismst Yarrowia lipolytica. J. Biotechnol. 168, 303314.

S., Aggelis, G., 2011. Lipids of oleaginous yeasts. Part I: biochemistryell oil production. Eur. J. Lipid Sci. Technol. 113, 10311051.

S., Aggelis, G., 2010. Yarrowia lipolytica: a model microorganismhe production of tailor-made lipids. Eur. J. Lipid Sci. Technol. 112,

S., Komaitis, M., Aggelis, G., 2004. Single cell oil (SCO) production by isabellina grown on high-sugar content media. Bioresour. Technol.91.el, M., Macnaughton, S., 2010. Food waste within food supply chains:

tion and potential for change to 2050. Philos. Trans. Roy. Soc. B 365,1.Kwan, T.H., Lin, C.S.K., 2014. Fungal hydrolysis in submerged fer-

for food waste treatment and fermentation feedstock preparation. Technol. 158, 4854.ynn, J.P., 2002. The biochemistry and molecular biology of lipid accu-n oleaginous microorganisms. Adv. Appl. Microbiol. 51, 151.g, Z., Yang, X., Jin, G., Bai, F., Zhao, Z.K., 2013. Kinetics of continuous

of the oleaginous yeast Rhodosporidium toruloides. J. Biotechnol. 168,

chau, A., Coradini, L.V.A., Franco, T.T., Deckmann, A.C., 2012. Optimiza-id production by the oleaginous yeast Lipomyces starkeyi by randomsis coupled to cerulenin screening. AMB Express 2, 6472.autista, L.F., Rodrguez, R., Gutirrez, F.J., Sdaba, I., Ruiz-Vzquez,s-Martnez, S., Garre, V., 2009. Biodiesel production from biomass ofous fungus. Biochem. Eng. J. 48, 2227.oy, L.C., Shaarani, S.M., Melikoglu, M., Koutinas, A., Webb, C., 2009.

wheat our hydrolysis by an enzyme mixture from solid state fungalion. Enzyme Microb. Technol. 44, 223228.., Melikoglu, M., Koutinas, A., Webb, C., 2007. Optimization of inno-anol production from wheat by response surface methodology. Trans.rt B 85, 19.002. Medicinal mushrooms a source of antitumor and immunomod-lysaccharides. Appl. Microbiol. Biotechnol. 10, 1332.., Wang, W., Du, W., Liu, D., 2012. Microbial conversion of biodiesel

glycerol to triacylglycerols by oleaginous yeast Rhodosporidum toru-the individual effect of some impurities on lipid production. Biochem.3036., X., Hua, Y., Feng, B., Zhao, Z.K., 2008. Medium optimization for lipid

through co-fermentation of glucose and xylose by the oleaginousmyces starkeyi. Eur. J. Lipid Sci. Technol. 110, 405412.

Formulation of fermentation media from flour-rich waste streams for microbial lipid production by Lipomyces starkeyi1 Introduction2 Materials and methods2.1 Microorganisms and media2.2 Raw materials2.3 Solid state fermentations2.4 Production of FRW hydrolysate2.5 Shake flask fermentations2.6 Bioreactor fermentations2.7 Analytical methods

3 Results and discussion3.1 Optimisation of solid state fermentation3.2 Optimisation of FRW hydrolysis3.3 Shake flask cultures of L. starkeyi for microbial lipid production3.4 Fed-batch bioreactor culture of L. starkeyi on FRW hydrolysate

4 ConclusionsAcknowledgmentsReferences