Cost effective production of pullulan from agri-industrial residues using response surface...

transcript

cTnripawibeTapwtauapal

International Journal of Biological Macromolecules 64 (2014) 252– 256

Contents lists available at ScienceDirect

International Journal of Biological Macromolecules

jo ur nal home p age: www. elsev ier .com/ locate / i jb iomac

ost effective production of pullulan from agri-industrial residuessing response surface methodology

nanya Mehta, G.S. Prasad, Anirban Roy Choudhury ∗

SIR-Institute of Microbial Technology (IMTECH), Council of Scientific and Industrial Research (CSIR), Sector – 39A, Chandigarh 160036, India

r t i c l e i n f o

rticle history:eceived 23 October 2013eceived in revised form 6 December 2013ccepted 7 December 2013

a b s t r a c t

Response surface methodology was used to develop an economically feasible process for the fermentativeproduction of pullulan using agri-industrial residues, jaggery, de-oiled jatropha seed cake (DOJSC) andcorn steep liquor (CSL), as sole media components. A second order polynomial model was obtained using

vailable online 14 December 2013

eywords:ullulanesponse surface methodologygri-industrial waste

central composite design to understand the effects of interactions among these substrates on pullulanbiosynthesis. Results indicated that, lower concentrations of CSL and DOJSC and higher concentrationsof jaggery favoured pullulan production. The optimal nutrient composition (18% jaggery, 3% DOJSC and0.97% CSL) as suggested by the model resulted in production of 66.25 g/L pullulan with a productivity of0.92 g/L h. Analysis of raw material cost component for pullulan production suggested that sole utilizationof agri-residues may lead to development of cost effective process for pullulan production.

. Introduction

Pullulan is a water soluble neutral extracellular homopolysac-haride synthesized by yeast like fungus Aureobasidium pullulans.his biomacromolecule is composed of maltotriose subunits con-ected by �-(1→4) and �-(1→6) glycosidic linkages in a unique 2:1atio. This unique linkage pattern confers pullulan higher solubil-ty and structural flexibility and renders it special physicochemicalroperties like fibre and film forming capability, low oxygen perme-bility, higher mechanical strength, etc. These properties coupledith non-toxic and non-immunogenic nature makes pullulan an

deal choice for applications in food, pharmaceuticals, cosmetic,iomedical industries, etc. Despite these advantages pullulan is lessxploited due to its high cost compared to other biopolymers [1].he high cost of pullulan is associated with cost of raw materi-ls, low productivity and yield during fermentative production andurification of the final product. One of the problems associatedith productivity and yield is inability of the A. pullulans strains to

olerate higher concentrations of glucose or other carbon sourcesnd this may be overcome by using osmotolerant strains of A. pull-lans [2]. Attempts were made to increase the fermentative yieldnd productivity by examining the effect of media composition,

∗ Corresponding author. Tel.: +91 172 6665312; fax: +91 1722695215.E-mail addresses: anirban@imtech.res.in, wb.anirban@gmail.com

A.R. Choudhury).

parameters [10–17]. However, in all these cases, expensive con-ventional media components like dextrose, yeast extract, peptone,etc. were used and replacing these costly nutrients by low costsubstrates may be economically advantageous. Media containinga combination of conventional nutrients along with agri-industrialresidues were also used as substrates for fermentative productionof pullulan [18–23]. However, there are no reports of using onlyagri-industrial residues as substrates for production of pullulan.In the present study, a medium solely composed of agri-industrialresidues viz. jaggery, de-oiled jatropha seed cake (DOJSC) and cornsteep liquor (CSL) was used for production of pullulan using anosmotolerant strain of A. pullulans. RSM was applied to develop amodel using central composite design to understand the interac-tions among these nutrients and their effect on pullulan production.The results indicated that pullulan production was favoured athigher concentrations of jaggery and lower concentrations of DOJSCand CSL. Process economic analysis was carried out about thefeasibility of using these low cost raw materials for commercialproduction of pullulan. To the best of our knowledge, this is thefirst report of pullulan production using a media containing onlyagri-industrial residues as sole media components.

2. Materials and methods

2.1. Material

DOJSC was obtained from Biodiesel Production Facility, Cen-tre for Rural Development, Indian Institute of Technology, Delhi,India. A proximate analysis data indicated that DOJSC mostly con-sists of protein and carbohydrate [24]. CSL was obtained from

Biological Macromolecules 64 (2014) 252– 256 253

Bsbfps

Ccvtwcpa

CmmmlAw

ac1pbpudCwu

ilaocav3(ioctfircw

Table 1Experimental range of the variables studied using CCD in terms of coded and actualfactors.

Factors Symbols Coded levels

Min.(%)

Low (−1)(%)

Mid (o)(%)

High(+1) (%)

Max.(%)

Std.Dev.

Jaggery A 9.95 12 15 18 20.05 2.48DOJSC B 1.98 3 4.5 6 7.02 1.24CSL C 0.33 0.50 0.75 1 1.17 0.21

Table 2Experimental design used in RSM studies to understand interaction amongnutrients.

Run no. Factor1 Factor2 Factor3 Predictedresponse

Observedresponse

Jaggery (%) DOJSC (%) CSL (%) Pullulan(g/L)

Pullulan(g/L)

1 15 4.50 0.75 53.70 54.022 15 4.50 0.33 56.07 54.983 18 6.00 1.00 55.82 54.504 15 4.50 1.17 48.40 48.435 15 4.50 0.75 53.73 54.026 18 3.00 1.00 57.61 58.597 12 3.00 1.00 47.28 47.498 12 3.00 0.50 51.73 53.789 18 3.00 0.50 55.82 55.89

10 15 4.50 0.75 54.53 54.0211 15 4.50 0.75 53.04 54.0212 15 4.50 0.75 55.80 54.0213 15 1.98 0.75 60.61 58.9814 12 6.00 1.00 40.21 40.8715 15 7.02 0.75 52.93 53.5016 15 4.50 0.75 53.16 54.0217 18 6.00 0.50 55.47 55.9918 9.95 4.50 0.75 43.90 42.65

used to predict optimum media composition for pullulan produc-

A. Mehta et al. / International Journal of

harath Starch Industries Limited, Yamunanagar, India. Literatureuggested that protein is the major component of this agri-residualy product [25]. Jaggery was procured from the local market andound to be mainly composed of sucrose [26]. The proximate com-osition of all the media components used are summarized inupplementary Table 1.

.2. Microorganism and inoculum development

Aureobasidium pullulans RBF 4A3, isolated from inflorescence ofaseulia axillaries [2] was used for pullulan production. The stockulture was stored in 20% glycerol at −80 ◦C for long term preser-ation. The organism was sub-cultured in YPD agar plates prioro each experimental run. Freshly grown cultures from YPD platesere transferred to a 250 ml conical flask containing 50 ml of media

omprised of 2% (w/v) dextrose, 1% (w/v) yeast extract and 2% (w/v)eptone for inoculum development [27]. The culture was incubatedt 28 ◦C at 200 rpm for 24 h.

.3. Shake flask fermentation

Pullulan was produced in shake flasks using jaggery, DOJSC andSL as media components. The concentrations of the nutrients inedia were varied as suggested by the statistical design of experi-ents as described earlier [27]. In each case, 20 ml of the productionedia in a 100 ml conical flask was inoculated with 5%(v/v) inocu-

um and incubated at 28 ◦C in a rotary shaker with 250 rpm for 72 h.ll the experiments were carried out in triplicates and average dataere presented.

.4. Purification, analysis and characterization of pullulan

Pullulan was harvested and purified by following the methods described earlier [2]. Briefly, the fermentation broth wasentrifuged at 16,000 × g for 20 min at 4 ◦C using Sigma 6K-5 centrifuge, followed by filtration through 0.22 � filters. Theolysaccharide was precipitated from this cell free fermentationroth by adding cold absolute ethanol in the ratio of 1:2 (v/v). Therecipitate obtained was dried at 80 ◦C overnight to remove resid-al ethanol. This crude polysaccharide was further purified usingialysis and re-precipitation techniques as described earlier byhoudhury et al. [2]. Pullulan content in the purified polysaccharideas measured by enzymatic method [22] and was characterizedsing FT-IR spectroscopy [2].

.5. Design of experiments and statistical analysis

Response surface methodology was used to understand thenteractions among nutrients and their effect on production of pul-ulan. A rotatable central composite design (CCD) was developeds a full factorial matrix to identify the key factors for higher yieldf pullulan. In full factorial design �-value gives the distance fromentre point to the axial point and also determines the location ofxial point in a design space. In case of a full factorial design, thealue of ̨ is equal to (2k)1/4. In present case the value of k is equal to

and hence the value of ̨ is 1.68179. The three media componentsjaggery, DOJSC and CSL) were used as numeric factors and variedn 5 different levels (−1 − �, −1, 0, +1, +1 + �) to generate secondrder response surface (Table 1). The upper and lower limits of con-entrations for all media components were decided on the basis ofhe results obtained from preliminary experiments. Thus using thisull factorial matrix a set of 20 experiments were designed which

19 12 6.00 0.50 51.60 51.3520 20.05 4.50 0.75 55.69 55.88

by performing these experiments and their effect on the selectedresponse (production of pullulan). A second order polynomial equa-tion was developed to define the predicted responses in terms ofindependent numeric variables and the equation is as follows:

Y = xo + x1A + x2B + x3C + x11A2 + x22B2 + x33C2 + x12AB

+ x13AC + x23BC.

where Y is response, xo is intercept coefficient, x1, x2, x3 are linearcoefficients, x11, x22, x33 are squared coefficients and x12, x13, x23are interaction coefficients.

The model developed by regression analysis is a rotatable designand hence may be used to measure and analyze the responsesobtained from any combination of the variables within the entirerange. The regression model developed was further analyzed sta-tistically after obtaining the responses from the experimental runsperformed to determine analysis of variants (ANOVA). Suitabil-ity of the model from statistical point of view was evaluatedusing Fishers test value (F value) and proportion of variance (R2

value) [27]. Design Expert software (ver 8.0) was used to gener-ate contour plots and response surface graphs for each variable.The three dimensional response surface plots were used to under-stand the individual and combinatorial effect of the interactionsamong variables on the desired response. These plots were also

tion. Further the model was validated performing experimentalruns suggested during statistical optimization of the model. Exper-iments were carried out in triplicates and average data werereported.

254 A. Mehta et al. / International Journal of Biolog

Table 3Analysis of variance (ANOVA) for all terms of model.

Source ANOVA for pullulan production

Sum of squares df F value Probability > F

Model 408.22 9 32.32 <0.0001A-Jaggery 211.38 1 133.39 <0.0001B-DOJSC 36.27 1 22.97 0.0006C-CSL 51.81 1 32.82 0.0001AC 40.41 1 25.60 0.0004BC 8.78 1 5.56 0.0479A2 37.91 1 22.37 0.0270B2 10.27 1 5.16 0.0425C2 8.31 1 6.5 0.0472Residual 17.37 11Lack of fit 13.49 5 2.79 0.1146

sDosmdhodhtRwmatAdew

atnssfcegmtRwaprp

pullulan production. The overall data obtained in the present study,

Pure error 5.31 5Cor total 425.59 19

. Results and discussion

.1. Development of rotatable central composite design

A rotatable central composite design was developed to under-tand the interactions among different media components (Jaggery,OJSC, CSL) and their effect on pullulan production. The rangesf variables, i.e. media components are listed in Table 1. Analy-is of ‘design of experiment’ indicated that quadratic model will beore suitable for this set of experiments. The overall analysis of the

esign of experiments in present study illustrated that this modelas 9 degrees of freedom and it has 5 degree of freedom for both lackf fit and pure error ensuring a valid lack of fit test. The analysis ofesign of experiments also demonstrated that the designed modelas a VIF value very close to 1.0, which is ideal and clearly showedhe ability of the model to estimate coefficients properly [27]. Thei-squared values in case of all the terms in the model equationere very close to 0, which suggested development of a suitableodel for all the variables considered. The Fisher statistical test for

nalysis of variance (ANOVA) was employed to understand statis-ical significance and reliability of the regression model (Table 3).NOVA suggested that, all the model terms, except AB, used forevelopment of quadratic equation were valid and hence the finalquation in terms of coded factor for pullulan production may beritten as follows:

ullulan = 53.95 + 3.93A − 1.63B − 1.95C + 2.25AC − 1.05BC

− 1.62A2 + 0.84B2 − 0.76C2

here A, B and C are jaggery, DOJSC and CSL respectively.A high F-value (32.32) implied that the model was significant

nd there was negligible chance of obtaining such high F value dueo noise. The lack of fit value for the response (2.79) indicated aon-significant lack of fit in comparison to pure error which clearlyhowed that this model might be used for navigating the designpace. The degree of precision with which experiments were per-ormed is indicated by the coefficient of variation (CV). In presentase a low CV value (2.48%) indicated greater reproducibility ofxperimental runs. A high adequate precision value (21.021) sug-ested an adequate signal and thus confirmed suitability of thisodel for using navigation of the design space. Whereas, aptness of

he model was clearly demonstrated by the R2 (0.96) and adjusted2 (0.93) values [15,27]. Three dimensional response surface plotsere developed to understand the interaction between substrates

nd also to obtain optimum level of substrates required for pullulan

ical Macromolecules 64 (2014) 252– 256

Fig. 1A represents the interaction between jaggery and DOJSCat specific concentration of CSL and demonstrates its effect onpullulan production. The response graph indicates that pullulanproduction was favoured by higher jaggery concentration andlower DOJSC concentration. However, it also suggests that a con-centration of jaggery beyond 18% (w/v) may have negative impacton pullulan production. These results confirm the earlier observa-tions, where higher concentrations of dextrose was found to havenegative effect on pullulan production [27]. Fig. 1B represented theeffect of interactions among jaggery and CSL at a defined concen-tration of DOJSC. The figure indicated that higher concentrations ofjaggery and lower concentrations of CSL is most favourable for pul-lulan production. Whereas, the interaction between CSL and DOJSCat a specific jaggery concentration is illustrated in Fig. 1C, whichindicated that, lower concentrations of DOJSC and moderate CSLconcentration is suitable for pullulan production. A media contain-ing 15% jaggery, 1.98% DOJSC and 0.75% CSL resulted in productionof 60.61 g/L pullulan. This is significantly higher compared to ear-lier report of when 5% jaggery was used as one of the substratesalong with yeast extract and inorganic salts to produce 23.19 g/Lpullulan [21].

3.2. Understanding the effect of interactions amongagri-industrial wastes on pullulan production

The observed and predicted responses obtained from the designof experiments are found to be well in agreement (Table 2). The datashows that pullulan production increased with increase in jaggeryconcentrations. As the jaggery concentrations were increased from12% to 18% without changing the DOJSC and CSL concentrations,the pullulan elaboration was found to increase from 47.49 g/L to58.59 g/L (Run nos. 7 and 6). Similar observations were also noted inRun nos. 8 and 9 and Run nos. 3 and 14 where, pullulan productionenhanced with increase in jaggery concentrations. These observa-tions are in line with our earlier observations, where A. pullulansRBF 4A3 was found to be osmotolerant and capable of producingmore than 60 g/L pullulan with 15% dextrose as carbon source [2].

Concentration of CSL has significant influence on the pullulanyield (Table 2), as can be seen from Run nos. 14 and 19, reducing theCSL concentration from 1% (w/v) to 0.5% (w/v) resulted around 20%increase in the pullulan yield. Similarly decreasing CSL concentra-tion from 1.17% (w/v) to 0.75% (w/v) increased pullulan productionfrom 48.43 g/L to 54.02 g/L (Run nos. 4 and 5), further decrease inCSL concentration to 0.33% (w/v) had marginally increased the pul-lulan production to 54.98 g/L (Run no. 2). This observation is in linewith earlier published report [23] where, lower concentrations ofCSL favoured higher elaboration of pullulan. This may be due tothe fact that CSL has a high lactic acid content which may haveinhibitory effects on pullulan production.

The observed responses indicated that DOJSC also has similareffect as compared to CSL. In case of Run nos. 15, 1 and 13, the pullu-lan elaboration increased from 52.93 g/L to 60.61 g/L with decreasein DOJSC concentration from 7.02% (w/v) to 1.98% (w/v) in themedia. Similar results were also obtained in case of Run nos. 3 and6 where pullulan production increased from 55.82 g/L to 57.61 g/Lwhen DOJSC concentration was reduced from 6% (w/v) to 3% (w/v).

Earlier reports showed that swollen cells and chlamydosporesare the morphological types of A. pullulans majorly responsible forpullulan production [28] and indicated that high carbon to nitro-gen ratio favoured conversion of blastospores to swollen cells andchlamydospores [29]. Catley et al. [30] also reported similar obser-vations indicating higher carbon to nitrogen ratio is better for

indicated that lower concentrations of CSL and DOJSC, which pre-dominantly acted as nitrogen sources and higher concentrations ofjaggery, major carbon source, favoured pullulan production. Thus,

A. Mehta et al. / International Journal of Biological Macromolecules 64 (2014) 252– 256 255

F ion ofv ween

cojcwoI

ig. 1. (A) Effect of interaction between DOJSC and Jaggery at a varying concentratarying concentration of DOJSC on pullulan production. (C) Effect of interaction bet

t may be commented that in this study also, nitrogen limitation hasesulted in formation of swollen cells and chlamydospores leadingo enhanced production of pullulan. However, being osmotoler-nt, A. pullulans RBF 4A3 was able to synthesize higher amount ofullulan compared to the other mentioned reports.

. Model validation and process optimization

Model was validated by performing shake flask fermentationsarried out with the predicted optimum media composition. Theptimum media composition for production of pullulan was 18%aggery, 3% DOJSC and 0.97% CSL. The actual experimental data indi-

bservations clearly ensured validity of the model. Further, the FT-R spectra of the polysaccharide produced was found to be almost

CSL on pullulan production. (B) Effect of interaction between CSL and jaggery at a CSL and DOJSC at a varying concentration of jaggery on pullulan production.

identical with standard pullulan obtained from sigma (Supplemen-tary Fig. 1).

5. Process economic feasibility analysis

One of the major bottlenecks for commercialization of pullulanis high cost of the final product. This challenge may be overcomeby enhancing productivity during fermentation through media andprocess parameter optimization along with exploitation of low costsubstrates. Recent reports on fermentative production of pullulanindicated that mostly conventional substrates like glucose, yeastextract, peptone were used for pullulan production sometimescoupled with low cost substrates like soyabean pomace, hydrol-ysed sweet potato, hydrolysed potato starch waste, etc. (Table 4).

256 A. Mehta et al. / International Journal of Biological Macromolecules 64 (2014) 252– 256

Table 4Comparative of pullulan production using different substrates.

Major substrates Pullulan produced (g/L) Productivity (g/L h) Reference

Carbon source Nitrogen sourceGlucose Yeast extract and peptone 66.79 0.69 [23]Glucose Yeast extract, ammonium sulphate 25.65 0.43 [31]Hydrolysed potato starch waste Yeast extract, ammonium sulphate 19.2 0.17 [32]Sucrose Yeast extract 30.28 0.25 [16]Soybean pomace Yeast extract, ammonium sulphate 7.5 0.1 [33]Hydrolysed sweet potato Yeast extract 29.43 0.31 [34]

drmawHcscmt[nT

ifriStccDrap

1330–1337.[33] H. Seo, C. Son, C. Chung, D. Jung, S. Kim, R.A. Gross, D.L. Kaplan, J. Lee, Bioresour.

Technol. 95 (2004) 293–299.

Sucrose Yeast extract, ammonium sulphateSucorse Yeast extract, ammonium sulphate

Jaggery CSL, DOJSC

e-oiled jatropha seed cake, corn steep liquor, jaggery, etc. foreducing the cost of pullulan production. For example, cost of rawaterial for pullulan production was reduced by more than 90%

s compared to pullulan produced using conventional substrateshen DOJSC [22] and CSL [23] were used as one of the nutrients.owever, in both cases, the low cost raw material was used alongonventional nutrients, whereas in present case the media wasolely composed with low cost raw agri-residues. Further, a pro-ess economic calculation demonstrated that in present study rawaterial cost for production is only $0.58/kg pullulan which is more

han 50% less as compared to earlier published reports where DOJSC22] and CSL [23] were used as nutrient in place of conventionalitrogen sources like yeast extract and peptone (Supplementaryable 2).

. Conclusion

The present study shows effective utilization low cost agri-ndustrial residues (jaggery, DOJSC and CSL) as media componentsor pullulan production. A quadratic model was developed usingotatable central composite design to understand the effects ofnteractions among these components on production of pullulan.ignificantly high yield of pullulan (66.25 g/L) was obtained usinghe optimized media as predicted by the model. The model alsolearly indicated that pullulan production was supported at higheroncentrations of jaggery and moderate concentrations of CSL andOJSC. Process economic analysis suggested that agri-industrial

esidues such as jaggery, DOJSC and CSL may be used as potentiallternatives to conventional media components for cost effectiveroduction of pullulan.

cknowledgements

The authors are thankful to the Council of Scientific and Indus-rial Research (CSIR), Government of India for financial support.he authors also wish to thank Dr. Sunil Khare, IIT-Delhi, Indiand Bharath Starch Industries Limited for providing DOJSC and CSLespectively.

ppendix A. Supplementary data

Supplementary data associated with this article can be found,n the online version, at http://dx.doi.org/10.1016/j.ijbiomac.013.12.011.

26.13 0.54 [35]44.4 0.26 [15]61.17 0.92 Present study

References

[1] T.D. Leathers, Appl. Microbiol. Biot. 62 (2003) 468–473.[2] A.R. Choudhury, P. Saluja, G.S. Prasad, Carbohyd. Polym. 83 (2011)

1547–1552.[3] M.S. Deshpande, V.B. Rale, J.M. Lynch, Enzyme Microb. Tech. 14 (1992)

514–527.[4] A. Leduy, J. Yarmoff, A. Chagraoui, Biotechnol. Lett. 5 (1983) 49–54.[5] A. Leduy, J.M. Boa, Can. J. Microbiol. 29 (1983) 143–146.[6] C. Lacroix, A. Leduy, G. Noel, L. Choplin, Biotechnol. Bioeng. 27 (1985)

202–207.[7] K. Ono, Y. Kawahara, S. Ueda, Agric. Biol. Chem. 41 (1977) 2313–2317.[8] B. McNeil, B. Kristiansen, Enzyme Microb. Tech. 12 (1990) 521–526.[9] M. Reeslev, B. Jensen, Appl. Microbiol. Biotechnol. 42 (1995) 910–915.10] Z. Chi, S. Zhao, Enzyme Microb. Tech. 33 (2003) 206–211.11] J.H. Lee, J.H. Kim, I.H. Zhu, X.B. Zhan, J.W. Lee, D.H. Shin, S.K. Kim, Biotechnol.

Lett. 23 (2001) 817–820.12] S.F.G. Oskouei, F. Tabandeh, B. Yakhchali, F. Eftekhar, Biochem. Eng. J. 39 (2008)

37–42.13] G. Prakash, A.K. Srivastava, Process Biochem. 40 (12) (2005) 3795–3800.14] R.K. Sen, T. Swaminathan, Appl. Microbiol. Biotechnol. 47 (1997) 358–363.15] A.M.L. Teruel, E. Gontier, C. Bienaime, J.E. Nava, B.J. Saucedo, Enzyme Microb.

Tech. 21 (1997) 314–320.16] R.S. Singh, G.K. Saini, Appl. Biochem. Biotech. 152 (2008) 42–53.17] L. Jiang, Carbohyd. Polym. 79 (2010) 414–417.18] C. Barnett, A. Smith, B. Scanlon, C.J. Israilides, Carbohyd. Polym. 38 (1999)

203–209.19] T. Roukas, M. Liakopoulou-Kriakides, J. Food Eng. 40 (1999) 89–94.20] K. Thirumavalavan, T.R. Manikkandan, R. Dhanesekar, Biotechnology 7 (2008)

317–322.21] S.V.N. Vijayendra, D. Bansal, M.S. Prasad, K. Nand, Process Biochem. 37 (2001)

359–364.22] A.R. Choudhury, N. Sharma, G.S. Prasad, Microb. Cell Fact. 11 (2012) 39.23] N. Sharma, G.S. Prasad, A.R. Choudhury, Carbohyd. Polym. 93 (2013) 95–101.24] U.V. Inekwe, M.O. Gauje, A.M. Dakare, O. Nkeonye, S. Bala, G. Mohammed,

Trends Adv. Sci. Eng. 2 (2012) 151–154.25] T.A. Nasir, M. Sarwar, F. Ahmad, M.A. Tipu, I. Hussain, J. Anim. Plant Sci. 22

(2012) 296–300.26] P.V.K. Jaganndha Rao, M. Das, S.K. Das, Indian J. Trad. Knowledge 6 (2007)

95–102.27] A.R. Choudhury, M.S. Bhattacharya, G.S. Prasad, Biocatal. Agric. BioTechnol. 1

(2012) 232–237.28] L. Simon, C. Caye-vaugen, M. Bouchonneau, J. Gen. Microbiol. 139 (1993)

979–985.29] A. Kockova-Kratochilova, A. Gernakova, E. Slavikova, Folia Microbiol. 25 (1980)

56–67.30] B.J. Catley, Appl. Microbiol. 22 (1971) 641–649.31] X. Yu, Y. Wang, G. Wei, Y. Dong, Carbohyd. Polym. 89 (2012) 928–934.32] Y. Goksungur, P. Uzunogullari, S. Dagbagli, Carbohyd. Polym. 83 (2011)

34] S. Wu, Z. Jin, Q. Tong, H. Chen, Carbohyd. Polym. 76 (2009) 645–649.35] J. Chen, S. Wu, S. Pan, Carbohyd. Polym. 87 (2012) 771–774.

Cost effective production of pullulan from agri-industrial residues using response surface...

Documents