Protein Expression and Purification 92 (2013) 235–244

Contents lists available at ScienceDirect

Protein Expression and Purification

journal homepage: www.elsevier .com/ locate /yprep

Cloning, expression and optimized production in a bioreactor of bovinechymosin B in Pichia (Komagataella) pastoris under AOX1 promoter

1046-5928/$ - see front matter � 2013 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.pep.2013.08.018

⇑ Corresponding author at: Instituto de Investigaciones Biotecnológicas-InstitutoTecnológico de Chascomús (IIB-INTECH), Universidad Nacional de San Martín(UNSAM) – Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET),San Martín, Buenos Aires, Argentina. Tel.: +54 11 4006 1500.

E-mail address: mag@di.fcen.uba.ar (M.A. Galvagno).1 These authors contributed equally to this work.

Diego Gabriel Noseda a,1, Matías Nicolás Recúpero a,1, Martín Blasco a, Gastón Ezequiel Ortiz a,Miguel Angel Galvagno a,b,⇑a Instituto de Investigaciones Biotecnológicas-Instituto Tecnológico de Chascomús (IIB-INTECH), Universidad Nacional de San Martín (UNSAM) – Consejo Nacional de InvestigacionesCientíficas y Técnicas (CONICET), San Martín, Buenos Aires, Argentinab Departamento de Ingeniería Química, Facultad de Ingeniería, Universidad de Buenos Aires, Pabellón de Industrias, Ciudad Universitaria, (1428) Buenos Aires, Argentina

a r t i c l e i n f o

Article history:Received 26 April 2013and in revised form 24 July 2013Available online 16 October 2013

Keywords:Pichia (Komagataella) pastorisMethanol-inducible AOX1 promoterBovine chymosinOptimizationFed-batch fermentation processPurification

a b s t r a c t

The codon sequence optimized bovine prochymosin B gene was cloned under the control of the alcoholoxidase 1 promoter (AOX1) in the vector pPIC9K and integrated into the genome of the methylotrophicyeast Pichia (Komagataella) pastoris (P. pastoris) strain GS115. A transformant clone that showed resis-tance to over 4 mg G418/ml and displayed the highest milk-clotting activity was selected. Cell growthand recombinant bovine chymosin production were optimized in flask cultures during methanol induc-tion phase achieving the highest coagulant activity with low pH values, a temperature of 25 �C and withthe addition of sorbitol and ascorbic acid at the beginning of this period. The scaling up of the fermenta-tion process to lab-scale stirred bioreactor using optimized conditions, allowed to reach 240 g DCW/L ofbiomass level and 96 IMCU/ml of milk-clotting activity. The enzyme activity corresponded to 53 mg/L ofrecombinant bovine chymosin production after 120 h of methanol induction. Western blot analysis of theculture supernatant showed that recombinant chymosin did not suffer degradation during the proteinproduction phase. By a procedure that included high performance gel filtration chromatography and3 kDa fast ultrafiltration, the recombinant bovine chymosin was purified and concentrated from fermen-tation cultures, generating a specific activity of 800 IMCU/Total Abs280 nm and a total activity recovery of56%. This study indicated that P. pastoris is a suitable expression system for bioreactor based fed-batchfermentation process for the efficient production of recombinant bovine chymosin under methanol-inducible AOX1 promoter.

� 2013 Elsevier Inc. All rights reserved.

Introduction

Bovine chymosin is a special member of the aspartic proteasegroup of enzymes synthesized in the fourth stomach (abomasum)of neonatal calves. This enzyme (323 amino acids, 35.6 kDa) is se-creted by the cells of the gastric mucosa as a zymogene, known asprochymosin (365 amino acids, 40.8 kDa). Under the acidic condi-tions of the stomach lumen, the enzyme precursor is convertedinto the active form through the autocatalytic breakdown of the42-amino acid N-terminal prosequence [1]. Chymosin containstwo aspartic acid residues at the active site, Asp32 and Asp215,which catalyze the selective cleavage of Phe105-Met106 peptidebond in j-casein molecules that stabilize milk micelles. This

cleavage causes the destabilization of the micelles and conse-quently induces milk clotting. Chymosin displays an extremelylow non-specific proteolytic activity, avoiding the furtherdegradation of milk proteins [2]. These characteristics make bovinechymosin appropriate for the coagulation of milk for cheesemanufacturing.

The production of recombinant bovine chymosin has beenachieved in diverse microorganism expression systems, such asEscherichia coli [3,4], Saccharomyces cerevisiae [5,2], and Kluyver-omyces lactis [6]. However the intracellular expression of chymosinin E. coli and S. cerevisiae generates an insoluble protein, which re-quires refolding so as to obtain an active enzyme [2,3]. In addition,the active extracellular chymosin produced in S. cerevisiae exhib-ited low secretion efficiency of prochymosin [7]. Moreover, chymo-sin has been expressed in filamentous fungi, such as Aspergillusspp. [8,9] and Trichoderma reesei [10].

The methylotrophic yeast Pichia (Komagataella) pastoris(P. pastoris) has become an important expression system for theproduction of active heterologous proteins [11,12]. In recent years,many proteins of different origin have been successfully cloned

236 D.G. Noseda et al. / Protein Expression and Purification 92 (2013) 235–244

and expressed in this organism [11,13]. The advantages of this sys-tem comprise simple molecular manipulation [14], high expressionlevel of recombinant proteins [15,16], post-translational modifica-tions, such as folding and glycosylation [13,14], the efficient secre-tion of extracellular proteins [17] and growth to high biomasslevels on defined minimal medium [11,18]. There are two reasonsfor the great success of the P. pastoris expression system. The first isa strong and efficient methanol inducible promoter from the alco-hol oxidase I gene (AOX1)2 that is frequently utilize to express for-eign genes [19]. Secondly, the preference of P. pastoris forrespiratory rather than fermentative metabolism, avoiding the gen-eration of ethanol and acetic acid during high-cell density processes[13]. Furthermore, P. pastoris does not contain the potentially onco-genic or viral nucleic acids present in mammalian cells nor the cellwall pyrogens found in E. coli, which in part supports the GRAS sta-tus of this yeast [20,21]. Due to its advantageous features, P. pastorisis also a useful expression system for scaling up recombinant proteinproduction [22].

Milk-clotting enzymes from different sources have been effi-ciently produced using P. pastoris system, such as Mucor rennin[23], buffalo chymosin [24], and goat chymosin [25]. Most of theseenzymes were expressed under the control of the AOX1 promoter.A distinctive feature of this promoter is that it is strongly repressedin the presence of certain carbon sources, such as glucose and glyc-erol, but induced over 1000-fold with methanol as the sole carbonsource. These characteristics ensure high cell densities before therecombinant proteins are overexpressed during methanol induc-tion [12,13].

Moreover, P. pastoris has been used for the constitutive expres-sion of bovine chymosin under the control of the glyceraldehyde-3-phosphate dehydrogenase gene (GAP) promoter in shaker flaskcultures [26]. However, some studies have reported that the con-stitutive expression of heterologous proteins might cause cytotoxiceffects in P. pastoris [11,12].

In the present study, the optimized codon sequence for the bo-vine prochymosin B gene was cloned into the pPIC9K vector andexpressed in a P. pastoris system under the control of the AOX1 pro-moter. In addition, we performed the optimization of cell growthand recombinant bovine chymosin production in P. pastoris usingshake flask cultures. Furthermore, we scaled up the production ofrecombinant chymosin in a stirred bioreactor using fed-batchmethanol feeding strategies under optimal physical and nutritionalconditions. Therefore, we obtained increased levels of active bo-vine chymosin in the fermentation supernatants. Additionally, weconducted the purification and concentration of recombinant bo-vine chymosin using high performance gel filtration chromatogra-phy and fast ultrafiltration obtaining a remarkable increase in thespecific milk-clotting activity.

Materials and methods

Reagents, plasmid and strains

Taq DNA polymerase and restriction enzymes were obtainedfrom Promega (Madison, WI). E. coli DH5a, P. pastoris GS115 andpPIC9K vector were acquired from Invitrogen (Carlsbad, CA). PCRcloning Kit was purchased from Qiagen (Hilden, Germany). Com-mercial recombinant bovine chymosin (Maxiren-DSM; Heerlen,Netherlands) was used as the chymosin standard for milk clotting,SDS–PAGE and Western blot assays.

2 Abbreviations used: AOX1, alcohol oxidase 1 promoter; GPA, glyceraldehyde-3-phosphate; BSM, basal salts medium; HSA, human serum albumin; DO, dissolvedoxygen; MWCO, molecular weight cutoff; IMCU, international milk clotting units.

Dry cell weight determination

The optical density of the samples was measured at 600 nmusing an UV–Vis spectrophotometer and converted to dry cellweights (DCW, in g/L) with a previously obtained calibration curveaccording to the formula:

OD600nm ¼ 2:3375� DCW ; R2 ¼ 0:9911 ð1Þ

Metabolites determination

Glucose was measured in the cell-free supernatant samplesfrom the different culture times using an enzymatic method (Wie-ner Lab S.A.I.C., Buenos Aires, Argentina). Glycerol levels were as-sessed by the microplate-adapted periodate method according toBondioli and Della Bella [27].

Milk-clotting determination

Milk-clotting activity was determined according to two tech-niques: the standard protocol of the International Dairy Federation(IDF) (ISO 11815 IDF 157) and the end-point dilution assay basedon previously described methods [3,28]. For both methodologies,powdered skimmed milk used as substrate was reconstituted at26% (w/v) in 0.5 g/L CaCl2 (pH 6.5), mixed at 25 ± 1 �C for 30 minand preincubated at 37 ± 1 �C for 20 min. The IDF protocol deter-mines the time of milk flocculation through the action of culturesupernatant samples. Hence, 100 ll of each sample was added to5 ml of milk solution in test tubes. After agitation, the tubes wereincubated in a water bath at 37 ± 1 �C with swirling on a rotatingspindle to avoid foam formation. The clotting time (s) was deter-mined when the first flocculation was observed in the substratefilm on the wall of the test tube. The procedure was repeated witheach sample to obtain duplicate values. Commercial recombinantbovine chymosin was utilized as a standard to compare the clottingtime values.

For the end-point dilution assay, culture supernatants wereserial diluted to half using CH3COOH/CH3COONa�3H2O buffer(pH 5.5) in a 96-well plastic plate. Milk solution (50 ll) wasadded to each well containing the supernatant dilutions (50 ll).After stirring the mixtures, the plate was incubated at 37 ± 1 �Cfor 10 min and centrifuged at 2000g for 5 min. The milk-clottingactivity was determined using the highest dilution that inducedmilk coagulation, visualized as the formation of clots in thebottom of the wells. The highest dilution values were comparedwith those acquired using a solution of commercial recombinantbovine chymosin (600 IMCU/ml) to obtain the international milkclotting units (IMCU). As a negative control, blank culturemedium was used instead of the culture supernatant. Themilk-clotting activity of each sample was determined induplicate.

SDS–PAGE and Western blot

The cell-free culture supernatants were mixed with crackingbuffer and subjected to 12% SDS–PAGE using a standard protocol[29]. The gels were either stained with Coomassie brilliant blueG250 to visualize the protein bands or transferred to a PVDF mem-brane (Thermo Scientific; Waltham, MA). After transfer, the mem-brane was incubated in PBSTM (10 mM phosphate-buffered salinewith 0.05% Tween 20 and 3% powdered milk) for 60 min. Afterwashing with PBST (10 mM phosphate-buffered saline with 0.05%Tween 20), the membrane was incubated with chicken (IgY)anti-chymosin antibodies (National Institute on Agricultural Tech-nology, Argentina) (1:1000 in PBSTM) for 90 min, followed by

D.G. Noseda et al. / Protein Expression and Purification 92 (2013) 235–244 237

washing 3 times with PBST. Subsequently, the membrane wasincubated with secondary rabbit anti-chicken IgG peroxidase-la-beled antibodies (1:1000 in PBSTM) and finally washed 3 timeswith PBST. The Super Signal Western blotting kit (Thermo Scien-tific) was used for visualization. Standard protein markers wereutilized to estimate the molecular weight of the proteins fromthe culture supernatants. Recombinant bovine chymosin concen-tration in the supernatants was determined using a BSA standardcalibration curve analyzed through SDS–PAGE and band estimatescanning using ImageJ image densitometry software (http://www.rsb.info.nih.gov/ij).

Media composition

Luria Bertani (LB) medium containing 100 lg/ml of ampicillinwas utilized for the selection of the E. coli clones transformed withpPIC9K vector grown at 37 ± 1 �C. P. pastoris growth on solid med-ium was performed at 30 ± 1 �C using YPD medium containing (ingrams per liter): peptone, 20; yeast extract, 10; glucose, 20 andagar, 20. Yeast Nitrogen Base (YNB, Difco 239210) agar mediumwas used for the selection of P. pastoris transformants based on his-tidine auxotrophy.

Basal salts medium (BSM) supplemented with trace metal solu-tion (PTM1) and biotin was utilized as the liquid culture media forthe P. pastoris experiments according to a previously describedcomposition [30]. Glycerol and methanol (alone or combined withsorbitol and/or ascorbic acid) were used as carbon sourcesthroughout the growth and induction phases, respectively.

Construction of expression vector

The pPIC9K vector was used to achieve the secreted expressionof bovine prochymosin B. This plasmid contains the tightly regu-lated AOX1 promoter and the S. cerevisiae a-factor secretion signallocated immediately upstream of the multiple cloning site [31].pPIC9K plasmid was digested with XhoI and EcoRI restriction en-zymes. The sequence of bovine prochymosin B gene (GenBank:J00003.1) was optimized in the codon sequences using Gene De-signer from DNA 2.0 [32]. After removing the signal peptide se-quence, the resulting prochymosin B sequence was synthesizedand subsequently cloned into pPIC9K at the XhoI and EcoRI

Fig. 1. Schematic representation of the pPIC9K-prochyB plasmid. The sequenceencoding bovine prochymosin B was inserted between the XhoI and EcoRI sites andexpressed under the control of the AOX1 promoter, fused at the N-terminus to thecleavable S. cerevisiae a-factor secretion signal. 50 AOX1: AOX1 promoter region, a-factor SP: a-factor secretion signal, 30 AOX1-TT: AOX1 transcriptional terminationregion, Kanamycin and Ampicillin Resistance: selectable markers, XhoI and EcoRI:sites for ligation of the insert, SacI: site for plasmid linearization, HIS 4: auxotrophicselectable marker.

restriction sites (Fig. 1). The recombinant plasmid pPIC9K-prochyBwas amplified through transformation into E. coli DH5a. The se-quence of the cloned prochymosin B gene was confirmed by DNAsequencing.

Transformation and selection of recombinant yeast

For P. pastoris GS115 transformation, 3 lg of pPIC9K-prochyBplasmid were linearized with SacI restriction enzyme and trans-formed into cells through electroporation using 2 mm gap cuvetteswith a BTX 630 electroporator at 1500 V, 125 X, 50 lF. The trans-formed cells were selected by a two-step procedure: first, the inte-grants were seeded onto plates containing YNB medium withoutamino acids for histidine prototrophy selection, and second, theintegrants were seeded on YPD medium containing geneticin(G418) at 0.5, 2 and 4 mg/ml for resistance selection. Coloniesexhibiting enhanced growth in medium supplemented with genet-icin were further grown in glass test tubes containing YPD brothand after 24 h, cultures were induced by adding 1% (v/v) puremethanol every 24 h. Supernatants were sampled after 72 h ofinduction and used to evaluate the milk-clotting activity accordingthe standard protocol of the IDF, and the relative amount of recom-binant chymosin produced was analyzed through Western blot-ting. The positive control corresponded to the commercialrecombinant bovine chymosin diluted 1/10, and the negative con-trol corresponded to the culture supernatant of P. pastoris trans-formed with the human serum albumin (HSA) gene (LifeTechnologies Inc., Carlsbad, CA).

Optimization of cell growth and chymosin production in P. pastoris

Cell growth and recombinant bovine chymosin production dur-ing the induction phase was optimized in shake flasks, analyzingthe effect of pH, temperature and the addition of sorbitol andascorbic acid at the beginning of the stage. For this purpose, P. pas-toris cells previously grown on YPD agar plate were utilized toinoculate BSM (supplemented with PTM1 and biotin) containingglycerol (10 g/L) in Erlenmeyer flasks incubated on a rotary shakerat 250 rpm and 30 ± 1 �C. After 24 h of cultivation, the cells wereharvested through centrifugation at 3500g for 15 min at room tem-perature and resuspended in shake flasks using BSM containingmethanol (1%, v/v) so as to obtain an OD600 of 10. The incubationwas continued on a shaker, with the addition of 1% (v/v) puremethanol every 24 h for a period of 120 h to sustain the inductionconditions.

To analyze the effect of pH, citrate–phosphate buffers wereadded every 24 h to maintain the cultures with different pH values(3, 4, 5 and 6), such that the final concentration in the cultures was0.1 M. The temperature effect was studied by incubating the cul-tures at different temperatures (20, 25 and 30 ± 1 �C) using the pre-determined optimal pH value. Finally, the effect of sorbitol andascorbic acid was evaluated under four conditions: (1) sorbitolwas added to the culture at a concentration of 50 g/L at the begin-ning of the induction phase, according to Celik et al. [33]; (2) theculture was supplemented with ascorbic acid such that the con-centration was 10 mmol/L at the beginning of the induction,according to Xiao et al. [30]; (3) both, sorbitol and ascorbic acidwere incorporated to the culture at the beginning of the phase atthe concentrations described above; and (4) neither sorbitol norascorbic acid were added during induction, as the control condi-tion. All flasks were incubated at the predetermined optimal tem-perature, adding methanol and citrate–phosphate buffer at every24 h to maintain the expression conditions.

The cultures were sampled every 24 h of methanol inductionand its DCW was determined. Then, culture samples were centri-fuged at 3500g for 15 min, and the supernatants were used for

238 D.G. Noseda et al. / Protein Expression and Purification 92 (2013) 235–244

the milk-clotting end-point dilution assay and glycerol levelmeasurement.

Bioreactor fermentation cultures of chymosin transformed P. pastoris

Fermentation was performed in four phases based on Celik et al.[33], with minor modifications. The first phase consisted in a batchculture with glycerol (40 g/L) as unique carbon source, whereyeasts achieved high cell densities whilst repressing foreign geneexpression, as the AOX1 promoter is repressed by excess amountsof glycerol. In the second phase, glycerol (600 g/L solution supple-mented with 12 ml/L PTM1) was fed to the culture at a growth-limiting rate to continue with the increase of the biomass leveland allow the gradual derepression of AOX1 promoter [34]. Glyc-erol feeding was conducted using a predetermined constant feed-ing profile, calculated according to the constant flow rateequation: F ¼ l0V0CX0=YX=SCS, where lo is the initial specificgrowth rate (h�1); Vo, the initial volume (l); CXo, the initial cell con-centration (g/L); Yx/s, the cellular yield coefficient based on sub-strate consumption (g cell/g substrate); and CS, the feed substrateconcentration (g/L). In this stage the pre-fixed parameters were ta-ken as: lo = 0.17 h�1, Vo = 3.2 L, Yx/s = 0.60 g/g and CS = 600 g/L.Next, a short transition stage was carried out by feeding with aglycerol:methanol (3:1) mixture, allowing the adaptation of cellsfor the growth on methanol. Finally, the induction phase was con-ducted by adding 100% methanol (supplemented with 12 ml/LPTM1) in a fed-batch mode with a growth-limiting rate. Methanolfeeding was achieved with a continuous feeding profile given bythe constant flow rate F, where the pre-fixed parameters in thisstage were taken as: lo = 0.06 h�1, Vo = 4 L, Yx/s = 0.30 g/g andCS = 800 g/L. Recombinant bovine prochymosin was synthesizedand secreted during this last phase.

In order to obtain the inoculum for fermentation, cells grown onYPD agar plates were inoculated into a 100-ml flask containing20 ml YPD medium and cultured overnight at 30 ± 1 �C. A volumeof 200 ml of BSM with 40 g/L glycerol supplemented with 4 ml/LPTM1 in a 1-L flask was inoculated with the overnight cultureand incubated at 30 ± 1 �C until the culture reached an OD600 of�25. This culture was utilized to inoculate 3 L of the aforemen-tioned culture media. Fermentation was conducted in a 7-L BioFlo110 bioreactor (New Brunswick Scientific; Edison, NJ) interfacedwith Biocommand Bioprocessing software (New Brunswick Scien-tific) for information acquisition and parameter control. Thetemperature was maintained at 30 ± 1 �C during batch, glycerolfed-batch and transition phases and at 25 ± 1 �C throughout induc-tion period (as determined below). A pH value of 5 was maintainedthrough the first three stages, and a pH of 4 was maintained in thelast phase (as established below); these values were controlled bythe addition of H3PO4 (85%) and 25% (v/v) NH4OH which alsoserved as nitrogen source. The dissolved oxygen (DO) was keptabove 30–40% of saturation and controlled through the agitationcascade (maximum of 1200 rpm) and the supply of filter-sterilized(0.22 lm) air. The pH was measured using a pH electrode (Mettler-Toledo GmbH, Germany) and the oxygen concentration was deter-mined with a polarographic probe (InPro6110/320, Mettler-ToledoGmbH). Foam formation was avoided through the addition of 0.3%(v/v) antifoam 289 (Sigma–Aldrich; St. Louis, MO). The incorpora-tion of both sorbitol and ascorbic acid to the fermentation cultureat the beginning of the methanol induction phase was comparedwith a strategy where these compounds were not added. The cul-tures were sampled from the beginning of the fermentationthroughout all bioprocess phases to measure the biomass leveland recombinant bovine chymosin concentration. Samples werealso utilized to determinate the milk-clotting activity using theend-point dilution assay.

Purification and concentration process of recombinant bovinechymosin

A bioreactor fermentation culture corresponding to a 96 h fed-batch methanol feeding performed with the supplementation ofsorbitol and ascorbic acid was centrifuged at 3500g for 15 min,and the supernatant was used as an enzyme source. The cell-freesupernatant (5 ml) was filtered through a 0.22-lm cellulose ace-tate filter (Thermo Scientific) to remove any large protein aggre-gate. Then, a 3.5 ml filtered supernatant volume was subjected tofast ultrafiltration in order to concentrate the recombinant bovinechymosin. For this, an Amicon Ultra-4 device (Millipore Corp.,Billerica, MA) with a molecular weight cutoff (MWCO) of 3 kDawas used at 7500g for 40 min and 5 �C. Later, 250 ll of the concen-trated supernatant was applied to a Superdex 75 HR 10/30 pre-packed column (Pharmacia Biotech Inc., Piscataway, NJ)connected to FPLC system to conduct high performance gel filtra-tion chromatography at an analytical scale. The column was equil-ibrated with two column volumes of a buffer consisting of 0.05 MNaH2PO4 and 0.15 M NaCl at pH 7.0 and the elution was carried outwith the same buffer at a flow rate of 0.25 ml/min. All the chroma-tography fractions were analyzed to determine the milk-clottingactivity. The fraction that exhibited clotting activity (1.5 ml) wassubjected to 3 kDa fast ultrafiltration by the above mentionedmethod to concentrate the purified recombinant chymosin. Theprotein profile of each stage of the whole procedure was analyzedby 12% SDS–PAGE and the gel was stained with Coomassie brilliantblue G-250. Total protein contents were estimated by measuringthe absorbance at 280 nm using a Beckman spectrophotometer.Volumetric milk-clotting activities (IMCU/ml) were determinedby the end-point dilution assay comparing with the activity valueof the chymosin standard.

Results and discussion

Selection of multi copy transformants

Ten transformant clones that exhibited enhanced growth in thepresence of a high concentration of geneticin (bigger colonies)(Fig. 2a) were grown in YPD broth and induced with methanol. Re-combinant chymosin production was evaluated through the deter-mination of the milk-clotting time (Fig. 2b). Clones 1, 2, 5 and 9exhibited higher clotting activity levels than the other clones ana-lyzed. These clones were subsequently analyzed with Westernblot, showing that clone 1 produced the highest amount of recom-binant bovine chymosin (Fig. 2c). Therefore, clone 1 was chosen foroptimizing cell growth and recombinant chymosin production aswell as for scaling-up the process.

Effect of pH on cell growth and chymosin production

It has been shown that low pH values increase the production ofrecombinant proteins through the reduction of protease activity incultures [21,35,36]. Thus, we analyzed the effect of medium pH oncell growth and recombinant chymosin production using P. pastorisclone 1. As shown in Fig. 3, the highest milk-clotting activity(IMCU/ml) was obtained at pH 3 and 4 from 96 h of methanolinduction. This result may be attributable to a diminution in theexpression or inactivation of native proteases in P. pastoris cells,as was previously reported by Shi et al. [36]. In addition, biomasslevel achieved at pH 5 and 6 was 1.2-fold higher than that at pH3 and 4 after 120 h of methanol induction. Although biomass accu-mulation at pH 3 and 4 was slightly lower than at pH 5 and 6, re-combinant chymosin production (IMCU/ml) was 2-fold higher(Fig. 3). Moreover, the experiment demonstrated that cell density

Fig. 2. Selection of multi-copy transformants using auxotrophy and antibiotic resistance. (A) Transformed colonies with different sizes in YPD medium containing 4 mg/ml ofgeneticin. Numbers (1–10) indicate the colonies that were selected for culture in YPD broth with methanol induction. (B) Milk-clotting activity, expressed in clotting time(sec), of the 10 transformed colonies (clones) selected. (C) Western blot analysis of the clones that presented higher clotting activity. A 15-ll volume of the culturesupernatant was loaded for each sample. Positive control (+) correspond to the commercial recombinant bovine chymosin diluted 1/10. Negative control (�) correspond to aculture supernatant of P. pastoris expressing HSA. Protein molecular weight markers expressed in kDa are shown in the right margin.

D.G. Noseda et al. / Protein Expression and Purification 92 (2013) 235–244 239

increased until 96 h of methanol induction, with similar growthrates observed at both pH 3 or 4 and at pH 5 or 6. This results showthat P. pastoris clone 1 is more efficient in expressing bovine chy-mosin at low pH values, therefore subsequent experiments wereconducted at pH 4 throughout the induction period.

Effect of temperature on cell growth and chymosin production

The P. pastoris clone 1 grew almost equally at temperatures of30 and 25 �C during methanol induction, reaching biomass levelsof 1.3-fold higher than those observed at 20 �C (Fig. 4). We also ob-served that biomass concentration increased until 96 h of induc-tion, with similar rate when growing at 30 or 25 �C. Furthermore,this experiment demonstrated that induction at 25 �C significantlyincreased recombinant chymosin production compared with thatat 30 and 20 �C (Fig. 4). Although the cell concentration was similarto that obtained at 25 �C, the low chymosin production observed at30 �C might reflect higher proteolytic activity. In addition, low chy-mosin expression at 20 �C may be due to the low growth rateachieved during methanol induction. Hence, subsequent assayswere performed at 25 �C during the protein production period.The results obtained agreed with previous reports where lowerculture temperatures produce an enhancement in the recombinantproteins expression yields in P. pastoris without detrimental effect

on the cell growth [36,37]. Shi et al. [36] have proposed that thiseffect could be explained by a decrease in extracellular proteolysis.

Effect of sorbitol and ascorbic acid on cell growth and chymosinproduction

It has been reported that adding sorbitol to the medium en-hances biomass production, achieving higher recombinant protein[33]. Moreover, the presence of sorbitol in the broth culture re-duces both the oxygen consumption and heat production rates[38]. Furthermore, Xiao et al. [30] reported that the incorporationof antioxidant ascorbic acid to the broth may decrease damagestress caused by reactive oxygen species, increasing the viabilityof P. pastoris cells and reducing the proteolytic degradation of het-erologous proteins in cultures containing methanol. Therefore, theeffect of adding sorbitol and ascorbic acid on cell growth and re-combinant chymosin production was investigated during metha-nol induction phase at the predetermined optimal values of pHand temperature. As sorbitol is a non-repressing carbon sourcefor AOX1 promoter [39,40], this substance, alone or in combinationwith ascorbic acid, was added to the cultures at the beginning ofthe induction period. Fig. 5 shows that when cultures were sup-plied with sorbitol or both sorbitol and ascorbic acid, the biomasswas 1.6-fold higher than with ascorbic acid or methanol alone at

Fig. 3. Effect of pH on cell growth and recombinant chymosin production during methanol induction. P. pastoris cells were grown in shake flasks containing BSM withmethanol (1% v/v) at different pH values (3, 4, 5 and 6). Cultures were incubated on a shaker with the addition of 1% (v/v) methanol every 24 h. Citrate–phosphate bufferswere added to maintain the pH values. Cultures were sampled every 24 h for biomass level and milk-clotting activity determination.

Fig. 4. Effect of temperature on cell growth and recombinant chymosin production during methanol induction. P. pastoris cells were grown in shake flasks containing BSMand methanol (1% v/v) at the optimal pH value and different temperatures (20, 25 and 30 �C). Methanol and citrate–phosphate buffer were added every 24 h to sustain theinduction conditions. Aliquots were withdrawn every 24 h to measure cell concentration and milk-clotting activity.

240 D.G. Noseda et al. / Protein Expression and Purification 92 (2013) 235–244

120 h post induction. Notably, biomass accumulation increaseduntil 96 h of induction with a similar growth rate under thesetwo conditions. Furthermore, the maximum production of recom-binant chymosin was obtained from 96 h after the start of metha-nol induction in cultures supplemented with both sorbitol andascorbic acid, with values 2-fold higher than those observed afteradding sorbitol or ascorbic acid alone. When inducing with meth-anol alone, recombinant chymosin production was 4-fold lowerthan when supplying both sorbitol and ascorbic acid at the begin-ning of the induction stage (Fig. 5). This last condition was selectedfor scaling up the production of recombinant bovine chymosin in astirred bioreactor and compared with bioprocessing without theaddition of such both compounds.

Production of recombinant bovine chymosin in bioreactor

P. pastoris clone 1 was grown in a 7-L stirred bioreactor using aprocess that included the four distinct stages described above. Inthe batch phase, cell concentration reached a maximum level of29 g DCW/L after 25 h of cultivation (Fig. 6A). At this stage, P. pas-toris exhibited a maximum specific growth rate (lmax) of 0.17 h�1

and a biomass yield coefficient (Yx/s) of 0.72 g DCW/g of glycerol.After a spike of dissolved oxygen, glycerol fed-batch phase wasstarted with a constant flow rate of 42 ml/L consisting of 600 g/Lglycerol. This feeding rate was maintained for 20 h, when biomassaccumulation reached a value of 108 g DCW/L. During this period,the yeast showed an average specific growth rate (l) of 0.085 h�1

Fig. 5. Effect of sorbitol and ascorbic acid on cell growth and recombinant chymosin production during methanol induction. P. pastoris cells were grown in shake flaskscontaining BSM with methanol (1% v/v) at the optimal pH and temperature under four conditions: sorbitol addition, ascorbic acid incorporation, sorbitol and ascorbic acidaddition and methanol alone. Methanol and citrate–phosphate buffer were added every 24 h. Culture samples were taken every 24 h to determine cell density and clottingactivity.

Fig. 6. Cell growth and recombinant chymosin production during bioreactor fermentation. Fermentation was performed in a 7-L bioreactor with or without the addition ofsorbitol and ascorbic acid at the beginning of the methanol induction period. (A) Variations in the biomass level with fermentation time. P1: glycerol batch phase; P2: glycerolfed-batch phase; P3: transition phase; P4: methanol fed-batch phase. (B) Production of recombinant bovine chymosin during methanol induction stage.

D.G. Noseda et al. / Protein Expression and Purification 92 (2013) 235–244 241

Fig. 7. Time course of recombinant chymosin expression. Production of recombi-nant bovine chymosin by bioreactor fermentation with the addition of sorbitol andascorbic acid was evaluated as a function of methanol induction times using SDS–PAGE and Western blotting. (A) SDS–PAGE showing protein profiles from super-natants of different induction times. Lane M: protein molecular weight marker(kDa); Lane C: commercial recombinant bovine chymosin; Lane 1: start ofinduction; Lane 2–6: 24, 48, 72, 96 and 120 h after induction. (B) Western blotanalysis showing production of recombinant chymosin at diverse times ofinduction. Lane C: commercial recombinant bovine chymosin; Lane 1: start ofinduction; Lane 2–4: 48, 96, 120 h after induction. Protein molecular weightmarkers (kDa) are shown in the right margin. A 20-ll volume of supernatant wasloaded for each sample. Arrows indicate the bands of recombinant bovinechymosin.

242 D.G. Noseda et al. / Protein Expression and Purification 92 (2013) 235–244

and an Yx/s of 0.68 g/g. Then, glycerol feeding was stopped and thetransition stage was performed by feeding with a glycerol:metha-nol (3:1) mixture for 8 h, to allow cell adaptation for the efficientutilization of methanol. At the end of this stage, the biomass levelreached 130 g DCW/L. Next, the methanol fed-batch phase was ini-tiated using a constant flow rate of 70 ml/L to induce recombinantbovine prochymosin expression. At this time point, two strategieswere conducted: (1) continuous feeding with methanol alone and(2) the incorporation of 50 g/L sorbitol and 10 mmol/L ascorbicacid to the fermentation culture at the beginning of the feeding.Methanol induction conducted with strategy (2) facilitated bio-mass accumulation to 240 g DCW/L, which was 1.5-fold higherthan the value obtained with methanol alone (160 g DCW/L)(Fig. 6A). In addition, during the induction period with strategy(2), cell concentration increased with a higher average specificgrowth rate (l = 0.042 h�1) compared with that obtained usingstrategy (1) (l = 0.015 h�1). Furthermore, strategy (2) achievedan Yx/s of 0.32 g/g, while with strategy (1) a value of 0.20 g/g wasobtained throughout the product formation phase. Noteworthy,the growth-limiting feeding rate utilized during the inductionphase permitted the maintenance of a low residual methanol

Table 1Purification and concentration process of recombinant bovine chymosin from Pichia pasto

Process stage Volume(ml)

Abs280 TotalAbs280

aMilk-clotting activity

Volumetric(IMCU/ml)b

Total(IMC

Cell free supernatant 5 13.5 67.5 96 480Filtration by 0.22 lm

filter5 13.2 66 96 480

3 kDa fastultrafiltration

1.43 26 37.1 384 549

High performance gelfiltration

8.40 0.08 0.67 24 202

3 kDa fastultrafiltration

1.40 0.24 0.34 192 269

a Absorbance at 280 nm multiplied by the volume.b Milk-clotting activity in international milk clotting units per millimeter (IMCU/ml).c Milk-clotting activity in IMCU/ml multiplied by the total volume.d Specific activity expressed as the total milk-clotting activity (IMCU) divided by thee The specific activity of a stage divided by the specific activity of the cell-free supernf The total activity of a stage divided by the total activity of the cell-free supernatant

concentration in the fermentation cultures, ensuring a slower pro-tein synthesis rate and preventing damage to the folding andexportation machinery. Moreover, the observed reduction of thespecific growth rate throughout fermentation phases reflectedchanges in the cultivation mode (batch or fed-batch) and the car-bon and energy sources (glycerol, methanol or sorbitol).

Furthermore, Fig. 6B shows that the recombinant chymosin pro-duction during fed-batch methanol feeding was higher for strategy(2), reaching a maximum milk-clotting activity of 96 IMCU/mlfrom 96 h of induction compared with that observed in the strat-egy (1), which exhibited a coagulant activity of 48 IMCU/ml. After125 h of methanol induction, no further increase in cell density andchymosin production was observed for both strategies. These re-sults confirmed that the incorporation of both sorbitol and ascorbicacid to the fermentation culture at the beginning of the inductionphase achieved a more effective production of recombinant bovinechymosin, due to the higher biomass levels obtained with sorbitoladdition and the enhanced viability generated through the actionof ascorbic acid. Notably, the milk-clotting activity attained afterthe incorporation of sorbitol and ascorbic acid during bioreactorfermentation was 8-fold higher than that obtained with shake flaskcultures.

Post induction samples corresponding to the fermentation pro-cess performed using strategy (2) were also analyzed throughSDS–PAGE and Western blotting (Fig. 7). In both cases, the recom-binant chymosin band migrated in the range of 30 and 45 kDa,which was expected, as the molecular weight of bovine chymosinis 35.6 kDa [1]. The concentration of the expressed recombinantchymosin, calculated by densitometric analysis, was 53 mg/L after120 h of induction with strategy (2). Moreover, the pronouncedincrease in the intensity of the recombinant chymosin bands ob-served with both SDS–PAGE and Western blot analysis confirmedthat efficient induction was achieved during methanol fed-batchphase. In addition, Western blotting revealed that recombinantbovine chymosin did not undergo detectable degradation duringprotein production phase (Fig. 7B).

Purification and concentration of recombinant chymosin

Recombinant bovine chymosin was purified and concentratedfrom the fermentation supernatant by a procedure that includedhigh performance gel filtration and fast ultrafiltration. The firststep of the process, 0.22-lm filtration, did not change the valuesof total protein content and milk-clotting activity (Table 1), whichsuggests that large protein aggregates were not generated. Before

ris fermentation cultures.

Increase in specificactivity (fold)e

Total activityrecovery (%)f

U)cSpecific (IMCU/totalAbs280)d

7 1 1007 1 100

15 2 114

300 43 42

800 114 56

total Abs280.atant.and multiplied by 100.

Fig. 8. SDS–PAGE analysis from the purification and concentration process.Recombinant bovine chymosin was subjected to purification and concentrationsteps by a procedure that included analytical-scale gel filtration chromatographyand fast ultrafiltration. Lane M: protein molecular weight marker; Lane C:commercial recombinant bovine chymosin; Lane 1: Cell-free supernatant, Lane 2:Filtration by 0.22 lm filter, Lane 3: 3 kDa fast ultrafiltration, Lane 4: highperformance gel filtration, Lane 5: 3 kDa fast ultrafiltration. A 20-ll sample volumewas loaded for each process stage. Arrow indicates the bands of recombinant bovinechymosin.

D.G. Noseda et al. / Protein Expression and Purification 92 (2013) 235–244 243

gel filtration chromatography, the 0.22-lm filtered supernatantwas desalted and concentrated by fast ultrafiltration. By this tech-nique the fermentation supernatant was 4-fold concentrated, elim-inating salts from the medium, as well as peptides and smallproteins with less than 3 kDa. Residual methanol was also removedin this stage of the procedure. In addition, by means of this step thevolumetric clotting activity (IMCU/ml) increased 4 times. The in-crease of the total activity recovery after this step may be due tothe error of the method for clotting activity determination(Table 1). Analytical-scale high performance gel filtration chroma-tography, using a Superdex 75 HR 10/30 column, allowed a properseparation of the supernatant proteins obtaining a unique fractionwith milk-clotting activity. This chromatography fraction showedthe presence of chymosin as major protein (96% of total proteins)(Fig. 8), indicating that the recombinant bovine chymosin was suc-cessfully purified from the proteins of high molecular weight. Thisprocess step presented a specific milk-clotting activity of300 IMCU/Total Abs280, which corresponds to a 43-fold increaseof such activity (Table 1). Then, this fraction was 6-fold concen-trated by 3 kDa fast ultrafiltration resulting in a specific clottingactivity of 800 IMCU/Total Abs280 nm and a final increase in thisactivity of 114 times. In this last step of the process, the totalmilk-clotting activity recovery reached a value of 56%. The recov-ery increase after this stage could be explained by the removal ofa putative inhibitor or a protease of bovine chymosin.

Conclusion

In this study, bovine prochymosin B was produced from P. pas-toris under the control of the methanol-inducible AOX1 promoter.The results demonstrated that the a-factor signal peptide from S.cerevisiae directed the secretion of the expressed prochymosin intothe culture broth. The optimization experiments showed that thehighest recombinant bovine chymosin production was achievedat low pH values (3 and 4) and with a temperature of 25 �C. More-over, adding sorbitol and ascorbic acid to the culture at the begin-ning of the methanol induction phase increased both the biomasslevel and milk-clotting activity, as sorbitol acts as a co-substrateand ascorbic acid increases cell viability, reducing the enzymaticdegradation of heterologous proteins. Furthermore, the scalingup of the process using a laboratory-scale bioreactor facilitatedthe production of 240 g/L dry biomass with a remarkable

milk-clotting activity of 96 IMCU/ml when both sorbitol and ascor-bic acid were added to the fermentation culture; these values wererespectively 1.5- and 2-fold higher than those obtained withoutthe addition of these compounds. The amount of recombinant bo-vine chymosin produced after 120 h of methanol induction in thebioreactor under optimal conditions was 53 mg/L. Additionally,the recombinant bovine chymosin was purified and concentratedfrom the fermentation supernatant by a process that included highperformance gel filtration chromatography and fast ultrafiltrationwith a molecular weight cutoff of 3 kDa. This process allowedobtaining a specific milk-clotting activity of 800 IMCU/TotalAbs280 nm, which represented a final increase of 114 times and a to-tal milk-clotting activity recovery of 56%. Interestingly, secretedbovine prochymosin is activated at low pH values during methanolinduction stage (pH 4), without the need for further in vitro pro-cessing. In conclusion, the results of the present study show thatthe utilization of P. pastoris as an expression system under the con-trol of the AOX1 promoter, using a fed-batch fermentation processin a lab-scale bioreactor with the predetermined optimal condi-tions, allow the efficient production of recombinant bovine chymo-sin. The expressed recombinant bovine chymosin with a high levelof purity could be used to perform suitable milk clotting in the pro-cedure of cheese manufacture.

Acknowledgments

This work was supported by the PICT Start Up 2010 1662 grantfrom the National Agency for Science and Technology Promotionfrom the National Ministry of Science and Technology of Argentina(issued to M.A. Galvagno) and the grant 2010 SJ10/31 from the Uni-versidad Nacional de San Martin (issued to M. Blasco). We thankDr. Andres Wigdorowitz and Dr. Gustavo Asenzo from INTA(National Institute on Agricultural Technology, Buenos Aires,Argentina) for obtaining the chicken anti-chymosin antibodiesand providing the anti-chicken secondary antibody.

References

[1] A.G. Attallah, Characters of chymosin gene isolated from different animalsources at molecular level, J. Appl. Sci. Res. 3 (2007) 904–907.

[2] J. Mellor, M.J. Dobson, N.A. Roberts, M.F. Tuite, J.S. Emtage, S. White, P.A. Lowe,T. Patel, A.J. Kingsman, S.M. Kingsman, Efficient synthesis of enzymaticallyactive calf chymosin in Saccharomyces cerevisiae, Gene 24 (1983) 1–14.

[3] J.S. Emtage, S. Angal, M.T. Doel, T.J. Harris, B. Jenkins, G. Lilley, P.A. Lowe,Synthesis of calf prochymosin (prorennin) in Escherichia coli, Proc. Natl. Acad.Sci. USA 80 (1983) 3671–3675.

[4] A.K. Mohanty, U.K. Mukhopadhyay, S. Grover, V.K. Batish, Bovine chymosin:production by rDNA technology and application in cheese manufacture,Biotechnol. Adv. 17 (1999) 205–217.

[5] C.G. Goff, D.T. Moir, T. Kohno, T.C. Gravius, R.A. Smith, E. Yamasaki, A. Taunton-Rigby, Expression of calf prochymosin in Saccharomyces cerevisiae, Gene 27(1984) 35–46.

[6] J.A. Van den Berg, K.J. Van der Laken, A.J. Van Ooyen, T.C. Renniers, K. Rietveld,A. Schaap, A.J. Brake, R.J. Bishop, K. Schultz, D. Moyer, Kluyveromyces as a hostfor heterologous gene expression: expression and secretion of prochymosin,Biotechnology (NY) 8 (1990) 135–139.

[7] R.A. Smith, M.J. Duncan, D.T. Moir, Heterologous protein secretion from yeast,Science 229 (1985) 1219–1224.

[8] D. Cullen, G.L. Gray, L.J. Wilson, K.J. Hayenga, M.H. Lamsa, M.W. Rey, S. Norton,R.M. Berka, Controlled expression and secretion of bovine chymosin inAspergillus Nidulans, BioTechnology (NY) 5 (1987) 369–376.

[9] K. Tsuchiya, K. Gomi, K. Kitamoto, C. Kumagai, G. Tamura, Secretion of calfchymosin from the filamentous fungus Aspergillus oryzare, Appl. Microbiol.Biotechnol. 40 (1993) 327–332.

[10] A. Harkki, J. Uusitalo, M. Bailey, M. Penttilä, J.K.C. Knowles, A novel fungalexpression system: secretion of active calf chymosin from the filamentousfungus Trichoderma reesei, Biotechnology (NY) 7 (1989) 596–603.

[11] J.L. Cereghino, J.M. Cregg, Heterologous protein expression in themethylotrophic yeast Pichia pastoris, FEMS Microbiol. Rev. 24 (2000) 45–66.

[12] S. Macauley-Patrick, M.L. Fazenda, B. McNeil, L.M. Harvey, Heterologousprotein production using the Pichia pastoris expression system, Yeast 22(2005) 249–270.

244 D.G. Noseda et al. / Protein Expression and Purification 92 (2013) 235–244

[13] G.P.L. Cereghino, J.L. Cereghino, C. Ilgen, J.M. Cregg, Production of recombinantproteins in fermenter cultures of the yeast Pichia pastoris, Curr. Opin.Biotechnol. 13 (2002) 329–332.

[14] J.M. Cregg, J.L. Cereghino, J. Shi, D.R. Higgins, Recombinant protein expressionin Pichia pastoris, Mol. Biotechnol. 16 (2000) 23–52.

[15] S. Richter, J. Nieveler, H. Schulze, T.T. Bachmann, R.D. Schmid, High yieldproduction of a mutant Nippostrongylus brasiliensis acetylcholinesterase inPichia pastoris and its purification, Biotechnol. Bioeng. 93 (2006)1017–1022.

[16] M. Rodríguez, R. Rubiera, M. Penichet, R. Montesinos, J. Cremata, V. Falcón, G.Sánchez, R. Bringas, C. Cordovés, M. Valdés, High level expression of the B.microplus Bm86 antigen in the yeast Pichia pastoris forming highlyimmunogenic particles for cattle, J. Biotechnol. 33 (1994) 135–146.

[17] L. Peng, X. Zhong, J. Ou, S. Zheng, J. Liao, L. Wang, A. Xu, High-level secretoryproduction of recombinant bovine enterokinase light chain by Pichia pastoris, J.Biotechnol. 108 (2004) 185–192.

[18] F. Hong, N.Q. Meinander, L.J. Jönsson, Fermentation strategies for improvedheterologous expression of laccase in Pichia pastoris, Biotechnol. Bioeng. 79(2002) 438–449.

[19] J.M. Cregg, K.R. Madden, Development of the methylotrophic yeast, Pichiapastoris, as a host system for the production of foreign proteins, J. Ind.Microbiol. 29 (1988) 33–41.

[20] V. Ciofalo, N. Barton, J. Kreps, I. Coats, D. Shanahan, Safety evaluation of a lipaseenzyme preparation, expressed in Pichia pastoris, intended for use in thedegumming of edible vegetable oil, Regul. Toxicol. Pharmacol. 45 (2006) 1–8.

[21] J.M. Cregg, T.S. Vedvick, W.C. Raschke, Recent advances in the expression offoreign genes in Pichia pastoris, Biotechnology (NY) 11 (1993) 905–910.

[22] P. Baumgartner, R.J.M. Raemaekers, A. Durieux, A. Gatehouse, H. Davies, M.Taylor, Large-scale production, purification, and characterisation ofrecombinant Phaseolus vulgaris phytohemagglutinin E-form expressed in themethylotrophic yeast Pichia pastoris, Protein Expr. Purif. 26 (2002)394–405.

[23] A. Beldarrain, N. Acosta, R. Montesinos, M. Mata, J. Cremata, Characterizationof Mucor pusillus rennin expressed in Pichia pastoris : enzymic, spectroscopicand calorimetric studies, Biotechnol. Appl. Biochem. 84 (2000) 77–84.

[24] J.A. Vallejo, J.M. Ageitos, M. Poza, T.G. Villa, A comparative analysis ofrecombinant chymosins, J. Dairy Sci. 95 (2012) 609–613.

[25] J.A. Vallejo, J.M. Ageitos, M. Poza, T.G. Villa, Cloning and expression of buffaloactive chymosin in Pichia pastoris, J. Agric. Food Chem. 56 (2008) 10606–10610.

[26] X.P. Jiang, M.L. Yin, P. Chen, Q. Yang, Constitutive expression, purification andcharacterization of bovine prochymosin in Pichia pastoris GS115, World J.Microbiol. Biotechnol. 28 (2012) 2087–2093.

[27] P. Bondioli, L. Della Bella, An alternative spectrophotometric method for thedetermination of free glycerol in biodiesel, Eur. J. Lipid Sci. Technol. 107 (2005)153–157.

[28] N.S. Dunn-Coleman, P. Bloebaum, R.M. Berka, E. Bodie, N. Robinson, G.Armstrong, M. Ward, M. Przetak, G.L. Carter, R. LaCost, Commercial levels ofchymosin production by Aspergillus, BioTechnology (NY) 9 (1991) 976–981.

[29] U.K. Laemmli, Cleavage of structural proteins during the assembly of the headof bacteriophage T4, Nature 227 (1970) 680–685.

[30] A. Xiao, X. Zhou, L. Zhou, Y. Zhang, Improvement of cell viability and hirudinproduction by ascorbic acid in Pichia pastoris fermentation, Appl. Microbiol.Biotechnol. 72 (2006) 837–844.

[31] D.R. Higgins, K. Busser, J. Comiskey, P.S. Whittier, T.J. Purcell, J.P. Hoeffler,Small vectors for expression based on dominant drug resistance with directmulticopy selection, in: D.R. Higgins, J.M. Cregg (Eds.), Methods in MolecularBiology: Pichia Protocols, Humana Press, New Jersey, 1998, pp. 28–41.

[32] A. Villalobos, J.E. Ness, C. Gustafsson, J. Minshull, S. Govindarajan, Genedesigner: a synthetic biology tool for constructing artificial DNA segments,BMC Bioinformatics 7 (2006) 285–293.

[33] E. Celik, P. Calık, S. Oliver, Fed-batch methanol feeding strategy forrecombinant protein production by Pichia pastoris in the presence of co-substrate sorbitol, Yeast 26 (2009) 473–484.

[34] P. Bhatacharya, G. Pandey, K.J. Mukherjee, Production and purification ofrecombinant human granulocyte-macrophage colony stimulating factor (GM-CSF) from high cell density cultures of Pichia pastoris, Bioprocess Biosyst. Eng.30 (2007) 305–312.

[35] L.M. Damasceno, I. Pla, H.J. Chang, L. Cohen, G. Ritter, L.J. Old, C.A. Batt, Anoptimized fermentation process for high-level production of a single-chain Fvantibody fragment in Pichia pastoris, Protein Expr. Purif. 37 (2004) 18–26.

[36] X. Shi, T. Karkut, M. Chamankhah, M. Alting-Mees, S.M. Hemmingsen, D.Hegedus, Optimal conditions for the expression of a single-chain antibody(scFv) gene in Pichia pastoris, Protein Expr. Purif. 28 (2003) 321–330.

[37] Z. Li, F. Xiong, Q. Lin, M. d’Anjou, A.J. Daugulis, D.S. Yang, C.L. Hew, Low-temperature increases the yield of biologically active herring antifreezeprotein in Pichia pastoris, Protein Expr. Purif. 21 (2001) 438–445.

[38] C. Jungo, J. Schenk, M. Pasquier, I.W. Marison, U. Von Stockar, A quantitativeanalysis of the benefits of mixed feeds of sorbitol and methanol for theproduction of recombinant avidin with Pichia pastoris, J. Biotechnology 131(2007) 57–66.

[39] M. Inan, M.M. Meagher, Non-repressing carbon sources for alcohol oxidase(AOX1) promoter of Pichia pastoris, J. Biosci. Bioeng. 92 (2001) 585–589.

[40] J. Xie, Q. Zhou, P. Du, R. Gan, Q. Ye, Use of different carbon sources incultivation of recombinant Pichia pastoris for angiostatin production, EnzymeMicrob. Technol. 36 (2005) 210–216.