Optimization of production of the anti-keratin 8 single-chain Fv TS1-218 in Pichia pastoris using design of experiments
© Jafari et al; licensee BioMed Central Ltd. 2011
Received: 22 December 2010
Accepted: 16 May 2011
Published: 16 May 2011
Optimization of conditions during recombinant protein production for improved yield is a major goal for protein scientists. Typically this is achieved by changing single crucial factor settings one at a time while other factors are kept fixed through trial-and-error experimentation. This approach may introduce larger bias and fail to identify interactions between the factors resulting in failure of finding the true optimal conditions.
In this study we have utilized design of experiments in order to identify optimal culture conditions with the aim to improve the final yield of the anti-keratin 8 scFv TS1-218, during expression in P. pastoris in shake flasks. The effect of: pH, temperature and methanol concentration on the yield of TS1-218 using buffered minimal medium was investigated and a predictive model established. The results demonstrated that higher starting pH and lower temperatures during induction significantly increased the yield of TS1-218. Furthermore, the result demonstrated increased biomass accumulation and cell viability at lower temperatures which suggested that the higher yield of TS1-218 could be attributed to lower protease activity in the culture medium. The optimal conditions (pH 7.1, temperature 11°C and methanol concentration 1.2%) suggested by the predictive model yielded 21.4 mg TS1-218 which is a 21-fold improvement compared to the yield prior to optimization.
The results demonstrated that design of experiments can be utilized for a rapid optimization of initial culture conditions and that P. pastoris is highly capable of producing and secreting functional single-chain antibody fragments at temperatures as low as 11°C.
Single-chain variable fragments (scFv) are small recombinant antibodies that consist of the variable binding domains of the light and heavy-chain (VL, VH) joined together with a short peptide linker [1, 2]. ScFvs retain the binding specificity of their parent immunoglobulins but are easier to manipulate and their expression is facilitated and can be readily expressed in different expression systems. During the last decades, the methylotropic yeast, Pichia pastoris (P. pastoris) has been proven to be a powerful candidate for high level expression of functional antibody fragments with reports of yields ranging from 10 mg up to 4.88 gram per liter of culture [3–9]. P. pastoris is easier to manipulate and culture than other eukaryotic cells and is also capable of performing many of the post-translational modifications seen in higher eukaryotes such as disulfide bond formation, glycosylation and proteolytic processing. In addition, P. pastoris provides the possibility of extracellular secretion of recombinant proteins and the low level of secreted endogenous P. pastoris proteins enables easier purification of recombinant proteins . Recently, we have utilized the P. pastoris KM71H strain for overexpression of the anti-keratin 8 scFv (TS1-218) in shake flasks where a twenty-fold increase in yield of soluble TS1-218 was obtained compared to overexpression of the same scFv in E. coli. However, the yield of TS1-218 from routine expression in P. pastoris using shake flasks could still become a substantial bottleneck for further progress due to the nature of shake flask cultures (i.e poor control).
Optimization of production conditions for overproduction of recombinant proteins are routinely achieved by varying single factors at a time until an apparent optimum is reached [7, 8, 12]. This approach could be labor intensive and assumes that all single parameters are mutually independent of one another and fails to identify interactions between the different factors involved . The consequences could be failure in identifying the true optimal conditions for protein production. Recently, several groups have adopted the statistical design of experiments (DoE) methodology in order to address these limitations during optimization of the conditions in protein expression [14–18]. The major advantages of DoE are that interactions between multiple factors can be identified, and that a more reliable prediction of the true optimum can be achieved. In addition, a more structured approach towards experimental setup can be undertaken which can aid in reducing the number of experiments and facilitate data analysis. Today there are numerous software packages available which facilitate the application of DoE for non-statisticians.
In this study we have applied DoE for optimization of the initial culture conditions in order to improve the final yield of the TS1-218 during expression in shake flask cultures by the P. pastoris KM71H MutS strain. We have investigated whether yield of the TS1-218 can be improved without addition of supplementary compounds to the culture medium in order to facilitate downstream processing. Three factors; temperature, pH and methanol (MeOH) concentration and their effects, alone and in combination, were investigated using response surface methodology (RSM) . Enzyme-Linked Immunosorbent Assay was used to assess the level of TS1-218 production.
Cell strains and media
P. pastoris strain KM71H MutS; arg4 aox1::ARG4 (Invitrogen) expressing anti-keratin 8 scFv, TS1-218. The P. pastoris clone, was maintained on Yeast extract Peptone Dextrose medium (YPD) agar plates (1% (w/v) yeast extract, 2% (w/v) peptone, 2% (w/v) dextrose, 1% (w/v) agar). For biomass production, Buffered Minimal Glycerol medium (BMG; 100 mM potassium phosphate pH 6.0, 1.34% (w/v) yeast nitrogen base, 4 × 10-5% (w/v) biotin, 1% (v/v) glycerol) was used. Buffered Minimal Methanol medium (BMM; 100 mM potassium phosphate pH 4-8, 1.34% (w/v) yeast nitrogen base, 4 × 10-5% (w/v) biotin, 0.1-3% (v/v) MeOH was used for induction of protein expression.
Experimental setup and analysis
Temperature and pH are two of the most common crucial factors that can influence protein production in various expression systems as well as production of scFvs in P. pastoris[7–9, 19]. MeOH concentration is another important factor in protein expression in P. pastoris under the regulation of the alcohol oxidase promoter (AOX) since MeOH can be used as the inducer and sole carbon source [20, 21]. Consequently these three factors; temperature, pH and MeOH concentration were chosen as quantitative factors in our investigation. The absorbance values from ELISA assays were chosen as a response for the level of TS1-218 production. In the final experiments the viability of the P. pastoris cells and change in wet cell weight (WCW) were also used as responses.
Box-Benhken design matrix and experimental results for the different factors and their effect on the responses.
Expression of TS1-218
Single P. pastoris colonies were grown overnight in BMG medium at 30°C. The overnight culture was used to inoculate a larger volume of fresh BMG in baffled shake flasks (Nalgene). The cultures were grown until the cells reached log phase (OD600 2-6). The cultures were then divided equally between conical sterile 50 mL tubes and centrifuged at 1500 × g for 5 minutes at room temperature and the supernatant was discarded and the wet cell weights (WCW) were measured (mean WCW: 564.4 ± SD 0.006 mg). To induce protein expression the pellets were resuspended with 5 mL BMM to a final OD600 of ~50 and incubated for 72 h with vigorous shaking. Methanol was added (v/v) every 24 h to the cultures in order to maintain induction. The MeOH concentration and pH of the BMM medium and incubation temperatures varied depending on the experimental runs (Table 1). After 72 h of induction small samples were taken from the cultures for further analysis and the cultures were centrifuged at 3000 × g for 10 minutes at 4°C. The culture supernatants were stored at 4°C and the WCWs were measured.
Cell viability assay
The viability of the P. pastoris cells were evaluated by propidium iodide (PI) staining . Samples taken from the BMG cultures prior to induction and from each experimental run at the end point (72 h) of each experiment were diluted in phosphate buffered saline (PBS, 137 mM NaCl, 3 mM KCl, 8 mM Na2HPO4 and 1 mM KH2PO4, pH 7.4). The cells were stained with PI at a final concentration of 5 μg/mL for 10 minutes at room temperature. The samples were briefly centrifuged at 1500 × g and the supernatants were discarded. The cells were washed twice with PBS and finally analyzed in a BD FACScan™ flow cytometer (Becton Dickinson). A total count of 50000 cells was collected for each sample. PI negative cells were considered as viable cells and PI positive cells were regarded as dead. As a positive control, heat-killed P. pastoris cells were stained with PI as described above. As negative control, P. pastoris cells without PI staining were used.
Enzyme Linked Immunosorbent Assay
Microtiter 96 wells plates (Nunc) were coated over night at 4°C with 100 μL per well of 5 μg/mL keratin 8/18/19 complex in 0.1 mM NaOH . The culture medium from different runs where diluted 60 times with (50 mM sodium phosphate buffer pH 6.5, 350 mM NaCl) and added in triplicates with 100 μL sample per well. The plates were incubated on a shaker for 2 h at room temperature. After washing with 50 mM Tris-HCl pH 7.4, 150 mM NaCl and 0.1% (v/v) Tween-20 (TBST) the plates were incubated for 2 h on a shaker at room temperature with 100 μL per well (0.85 μg/mL) monoclonal mouse anti-polyhistidine-alkaline phosphatase conjugated antibody clone his-1 (Sigma-Aldrich) diluted with TBST. After washing with TBST the plates were developed with 3 mM para-Nitro Phenyl Phosphate (pNPP) in 50 mM 2-amino-2-metyl-1-propanol, 1 mM MgCl2 pH10. The absorbance was measured at 405 nm and the absorbance at 650 nm was subtracted. The mean absorbance for triplicate samples were calculated for the different runs before analysis in MODDE.
Purification of TS1-218
The hexahistidine-tagged TS1-218 from the scaled up one liter cultures were purified from the culture medium with immobilized-metal affinity chromatography (IMAC) using Ni-NTA agarose (Invitrogen). The TS1-218 in the culture medium were allowed to bind to a 10 mL Ni-NTA agarose gel bed in a K26/60 column (Pharmacia) followed by a wash of the Ni-NTA agarose gel bed with wash buffer (50 mM sodium phosphate buffer pH 6.5, 350 mM NaCl) and another wash with a wash buffer containing 25 mM imidazole. The TS1-218 was eluted with 50 mM sodium phosphate buffer pH 6.5, 350 mM NaCl, 300 mM imidazole. The samples were concentrated and the buffer was exchanged against wash buffer using Amicon Ultra-15 centrifugal filter units. BCA Protein Assay Kit (Pierce) was used to determine the concentration of TS1-218.
The purified TS1-218 samples from the validations experiments were adjusted to equal volumes and separated on NuPage® Novex Bis-Tris 4-12% polyacrylamide gels with NuPAGE® MES SDS Running Buffer according to the manufacturer's recommendations (Invitrogen) and stained with Coomassie Blue R-250.
Results and Discussion
The aim of this study was to identify optimal culture conditions for TS1-218 production in BMM medium in shake flasks by the methylotrophic yeast P. pastoris using design of experiments. There are many different factors that can influence protein expression in P. pastoris, such as pH, temperature, MeOH concentration, cell density, medium composition or additives (casamino acids, sorbitol, EDTA) to name a few. In a previous study, P. pastoris KM71H/MutS clones expressing hexahistidine-tagged (His-tagged) TS1-218 under the control of the AOX promoter were cultured in complex medium containing yeast extract and peptone at pH 6, temperature 20°C and 0.5% (v/v) methanol during expression . Some components of the complex medium were retained in the metal affinity purification matrix and co-eluted with the TS1-218. These components interfered with the downstream processing of the purified proteins and needed to be removed prior to further analysis. Additives such as casamino acids and EDTA have been shown to increase the yield of secreted proteins in many cases [7, 25, 26], probably due to their inhibitory effect on extracellular proteases present in the culture medium. However, these medium additives can also interfere with the purification matrix during affinity purification of His-tagged proteins as mentioned earlier. For these reasons, buffered minimal medium without any supplementary additives which provides less background was used for both biomass production and protein induction in this study.
Generally, when using P. pastoris MutS strains for protein production, high cell densities are used since the yield of secreted proteins are proportional to cell density . In addition, the MutS strains grow at a slower rate when methanol is used as the sole carbon source. However, very high cell densities can result in rapid depletion of the nutrients and lead to increased accumulation of extracellular protease activity due to increased cell death [20, 27]. In contrast, in low cell density cultures more resources are employed for biomass production than protein production [7, 8]. Therefore we decided to use a cell density (OD600) of ~50 during the induction phase throughout the entire study.
Initially the effect of pH, temperature and MeOH concentration were investigated using a wide range between the factor levels; pH 4-8, temperature 15-30°C and MeOH concentration 0.1-3%. However, at the highest temperature (30°C) and lowest pH (pH 4) investigated, no signal were detected in the ELISA assays for any of the experimental runs, nor were there any TS1-218 detected after Ni-NTA purification and SDS-PAGE analysis of the culture medium from these runs (results not shown). The results from the initial experiment indicated that lower temperatures and a more narrow range of the factors should be investigated in order to achieve a more comprehensive model.
Validation of the predicted optimal culture conditions for optimal TS1-218 yield in one liter cultures.
1.0 ± 0.14
15.3 ± 1.29
21.4 ± 1.25
P. pastoris has several endogenous proteases although very little is known about these different proteases. Shi et al.  have described an increased production of a scFv at higher pH (7.5-8) in shake flask cultures, although the scFv production did not correlate with protease activity. Other studies have however shown that lower pH in the bioreactor cultures resulted in increased protease activity, which suggest that the majority of the proteases released by P. pastoris in these studies have an optimal activity at acidic conditions [28, 29]. When measuring the pH in the culture medium after 72 h of induction it was found that the pH had decreased markedly in the different experimental runs even though a buffered medium was used (Table 1). It is likely that the TS1-218 was also subjected to increased proteolytic degradation in culture medium with lower pH resulting in weaker absorbance signal in the ELISA assays for the experimental runs with a lower starting pH. This could probably be avoided if the pH in the cultures could be maintained at strict intervals by improving the buffer capacity in the culture medium or using bioreactors.
The aim of this study was to utilize DoE in order to investigate the effect of pH, temperature and MeOH concentration on production of the anti-keratin 8 single-chain Fv, TS1-218, by P. pastoris strain KM71H/MutS with the purpose to optimize production. The application of DoE facilitated the experimental setup and data analysis and identified several interactions between the investigated factors. The model suggested several factor settings for optimal protein production that deviate considerably from the generally recommended conditions for recombinant protein expression in P. pastoris, specially the temperature. The yield of the TS1-218 was improved 21-fold compared to the pre-optimization conditions (pH 6, temperature 20°C, MeOH concentration 0.5%) by lowering the temperature to 11°C and increasing the starting pH and MeOH concentration of the cultures to 7.1 and 1.2% (v/v) respectively during induction. This study shows that P. pastoris is highly capable of producing recombinant proteins at temperatures as low as 11°C.
The authors would like to thank Lars Renman at Karlstad University for valuable input.
- Bird RE, Hardman KD, Jacobson JW, Johnson S, Kaufman BM, Lee SM, Lee T, Pope SH, Riordan GS, Whitlow M: Single-chain antigen-binding proteins. Science. 1988, 242: 423-426. 10.1126/science.3140379.View Article
- Huston JS, Levinson D, Mudgett-Hunter M, Tai MS, Novotny J, Margolies MN, Ridge RJ, Bruccoleri RE, Haber E, Crea R, et al: Protein engineering of antibody binding sites: recovery of specific activity in an anti-digoxin single-chain Fv analogue produced in Escherichia coli. Proc Natl Acad Sci USA. 1988, 85: 5879-5883. 10.1073/pnas.85.16.5879.View Article
- Chang HJ, Choi SW, Chun HS: Expression of functional single-chain variable domain fragment antibody (scFv) against mycotoxin zearalenone in Pichia pastoris. Biotechnol Lett. 2008, 30: 1801-1806. 10.1007/s10529-008-9770-x.View Article
- Eldin P, Pauza ME, Hieda Y, Lin G, Murtaugh MP, Pentel PR, Pennell CA: High-level secretion of two antibody single chain Fv fragments by Pichia pastoris. J Immunol Methods. 1997, 201: 67-75. 10.1016/S0022-1759(96)00213-X.View Article
- Freyre FM, Vazquez JE, Ayala M, Canaan-Haden L, Bell H, Rodriguez I, Gonzalez A, Cintado A, Gavilondo JV: Very high expression of an anti-carcinoembryonic antigen single chain Fv antibody fragment in the yeast Pichia pastoris. J Biotechnol. 2000, 76: 157-163. 10.1016/S0168-1656(99)00183-2.View Article
- Rahbarizadeh F, Rasaee MJ, Forouzandeh M, Allameh AA: Over expression of anti-MUC1 single-domain antibody fragments in the yeast Pichia pastoris. Mol Immunol. 2006, 43: 426-435. 10.1016/j.molimm.2005.03.003.View Article
- Shi X, Karkut T, Chamankhah M, Alting-Mees M, Hemmingsen SM, Hegedus D: Optimal conditions for the expression of a single-chain antibody (scFv) gene in Pichia pastoris. Protein Expr Purif. 2003, 28: 321-330. 10.1016/S1046-5928(02)00706-4.View Article
- Wan L, Cai H, Yang H, Lu Y, Li Y, Li X, Li S, Zhang J, Cheng J, Lu X: High-level expression of a functional humanized single-chain variable fragment antibody against CD25 in Pichia pastoris. Appl Microbiol Biotechnol. 2008, 81: 33-41. 10.1007/s00253-008-1568-1.View Article
- Damasceno LM, Pla I, Chang HJ, Cohen L, Ritter G, Old LJ, Batt CA: An optimized fermentation process for high-level production of a single-chain Fv antibody fragment in Pichia pastoris. Protein Expr Purif. 2004, 37: 18-26. 10.1016/j.pep.2004.03.019.View Article
- Cregg JM, Vedvick TS, Raschke WC: Recent advances in the expression of foreign genes in Pichia pastoris. Biotechnology (N Y). 1993, 11: 905-910. 10.1038/nbt0893-905. 10.1038/nbt0893-905.View Article
- Jafari R, Holm P, Piercecchi M, Sundstrom BE: Construction of divalent anti-keratin 8 single-chain antibodies (sc(Fv)(2)), expression in Pichia pastoris and their reactivity with multicellular tumor spheroids. J Immunol Methods. 2011, 364: 65-76. 10.1016/j.jim.2010.11.003.View Article
- Cai H, Chen L, Wan L, Zeng L, Yang H, Li S, Li Y, Cheng J, Lu X: High-level expression of a functional humanized anti-CTLA4 single-chain variable fragment antibody in Pichia pastoris. Appl Microbiol Biotechnol. 2009, 82: 41-48. 10.1007/s00253-008-1744-3.View Article
- Eriksson L: Design of experiments : principles and applications. 2000, Umea: Umetrics,
- Chambers SP, Swalley SE: Designing experiments for high-throughput protein expression. Methods Mol Biol. 2009, 498: 19-29. 10.1007/978-1-59745-196-3_2.View Article
- Holmes WJ, Darby RA, Wilks MD, Smith R, Bill RM: Developing a scalable model of recombinant protein yield from Pichia pastoris: the influence of culture conditions, biomass and induction regime. Microb Cell Fact. 2009, 8: 35- 10.1186/1475-2859-8-35.View Article
- Islam RS, Tisi D, Levy MS, Lye GJ: Framework for the rapid optimization of soluble protein expression in Escherichia coli combining microscale experiments and statistical experimental design. Biotechnol Prog. 2007, 23: 785-793.View Article
- Lin H, Kim T, Xiong F, Yang X: Enhancing the production of Fc fusion protein in fed-batch fermentation of Pichia pastoris by design of experiments. Biotechnol Prog. 2007, 23: 621-625.View Article
- Swalley SE, Fulghum JR, Chambers SP: Screening factors effecting a response in soluble protein expression: formalized approach using design of experiments. Anal Biochem. 2006, 351: 122-127. 10.1016/j.ab.2005.11.046.View Article
- Tanfous NGB, Kallel H, Jarbouia MA, Fathallah DM: Expression in Pichia pastoris of a recombinant scFv form of MAb 107, an anti human CD11b integrin antibody. Enzyme and microbal technology. 2005, 38: 636-642.View Article
- Daly R, Hearn MT: Expression of heterologous proteins in Pichia pastoris: a useful experimental tool in protein engineering and production. J Mol Recognit. 2005, 18: 119-138. 10.1002/jmr.687.View Article
- Cereghino JL, Cregg JM: Heterologous protein expression in the methylotrophic yeast Pichia pastoris. FEMS Microbiol Rev. 2000, 24: 45-66. 10.1111/j.1574-6976.2000.tb00532.x.View Article
- Box G, Behnken D: Some new three level designs for the study of quantitative variables. Technometrics. 1960, 2: 455-475. 10.2307/1266454. Richmond, Va : American Society for Quality Control ; Richmond, Va : American Statistical Association, 1959,
- Zhao HL, Xue C, Wang Y, Yao XQ, Liu ZM: Increasing the cell viability and heterologous protein expression of Pichia pastoris mutant deficient in PMR1 gene by culture condition optimization. Appl Microbiol Biotechnol. 2008, 81: 235-241. 10.1007/s00253-008-1666-0.View Article
- Sundstrom BE, Nathrath WB, Stigbrand TI: Diversity in immunoreactivity of tumor-derived cytokeratin monoclonal antibodies. J Histochem Cytochem. 1989, 37: 1845-1854. 10.1177/37.12.2479674.View Article
- Sreekrishna K, Brankamp RG, Kropp KE, Blankenship DT, Tsay JT, Smith PL, Wierschke JD, Subramaniam A, Birkenberger LA: Strategies for optimal synthesis and secretion of heterologous proteins in the methylotrophic yeast Pichia pastoris. Gene. 1997, 190: 55-62. 10.1016/S0378-1119(96)00672-5.View Article
- Clare JJ, Romanos MA, Rayment FB, Rowedder JE, Smith MA, Payne MM, Sreekrishna K, Henwood CA: Production of mouse epidermal growth factor in yeast: high-level secretion using Pichia pastoris strains containing multiple gene copies. Gene. 1991, 105: 205-212. 10.1016/0378-1119(91)90152-2.View Article
- Cregg JM, Cereghino JL, Shi J, Higgins DR: Recombinant protein expression in Pichia pastoris. Mol Biotechnol. 2000, 16: 23-52. 10.1385/MB:16:1:23.View Article
- Hong F, Meinander NQ, Jonsson LJ: Fermentation strategies for improved heterologous expression of laccase in Pichia pastoris. Biotechnol Bioeng. 2002, 79: 438-449. 10.1002/bit.10297.View Article
- Kobayashi K, Kuwae S, Ohya T, Ohda T, Ohyama M, Ohi H, Tomomitsu K, Ohmura T: High-level expression of recombinant human serum albumin from the methylotrophic yeast Pichia pastoris with minimal protease production and activation. J Biosci Bioeng. 2000, 89: 55-61. 10.1016/S1389-1723(00)88050-0.View Article
- Dragosits M, Stadlmann J, Albiol J, Baumann K, Maurer M, Gasser B, Sauer M, Altmann F, Ferrer P, Mattanovich D: The effect of temperature on the proteome of recombinant Pichia pastoris. J Proteome Res. 2009, 8: 1380-1392. 10.1021/pr8007623.View Article
- Li Z, Xiong F, Lin Q, d'Anjou M, Daugulis AJ, Yang DS, Hew CL: Low-temperature increases the yield of biologically active herring antifreeze protein in Pichia pastoris. Protein Expr Purif. 2001, 21: 438-445. 10.1006/prep.2001.1395.View Article
- Mattanovich D, Gasser B, Hohenblum H, Sauer M: Stress in recombinant protein producing yeasts. J Biotechnol. 2004, 113: 121-135. 10.1016/j.jbiotec.2004.04.035.View Article
- Gasser B, Maurer M, Rautio J, Sauer M, Bhattacharyya A, Saloheimo M, Penttila M, Mattanovich D: Monitoring of transcriptional regulation in Pichia pastoris under protein production conditions. BMC Genomics. 2007, 8: 179- 10.1186/1471-2164-8-179.View Article
- Jahic M, Wallberg F, Bollok M, Garcia P, Enfors SO: Temperature limited fed-batch technique for control of proteolysis in Pichia pastoris bioreactor cultures. Microb Cell Fact. 2003, 2: 6- 10.1186/1475-2859-2-6.View Article
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.