G Model

ARTICLE IN PRESS

BIOTEC-6837; No. of Pages 9

Journal of Biotechnology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Journal of Biotechnology journal homepage: www.elsevier.com/locate/jbiotec

Synergistic modular promoter and gene optimization to push cellulase secretion by Pichia pastoris beyond existing benchmarks Andrea Mellitzer a , Claudia Ruth a , Claes Gustafsson b , Mark Welch b , Ruth Birner-Grünberger c , Roland Weis d , Thomas Purkarthofer d , Anton Glieder a,∗ a

ACIB GmbH Graz University of Technology, Institute of Molecular Biotechnology, NAWI Graz, Petersgasse 14, A8010 Graz, Austria DNA2.0 Inc., Menlo Park, CA 94025, USA c Center for Medical Research, Graz, Austria d VTU Technology GmbH, Grambach, Austria b

a r t i c l e

i n f o

Article history: Received 24 March 2014 Received in revised form 20 August 2014 Accepted 25 August 2014 Available online xxx Keywords: Pichia pastoris Synonymous codon optimization Synthetic promoters Gene dosage Cellulase secretion

a b s t r a c t Although successfully used for heterologous gene expression for more than twenty years, general knowledge about all factors influencing protein expression by Pichia pastoris is still lacking. For high titers of protein clones are optimized individually for each target protein. Optimization efforts in this study were focused on the DNA level, evaluating a set of 48 different individual synthetic genes (TrCBH2) coding for the same protein sequence of a Trichoderma reesei cellulase in combination with three different promoter sequences: PGAP (constitutive) and the synthetic AOX1 promoter variants PDeS (derepressed) and PEn (enhanced, inducible). Expression of active secreted enzyme varied from undetectable to ∼300% of the best known gene, as determined by secreted enzyme activity analyses of supernatants from 96 well plate and bioreactor cultivations. Finally, the best optimized gene and new promoters were combined to engineer highly productive P. pastoris CBH2 expression strains. Although no methanol was used for induction a final titer of more than 18 g/l of secreted protein was produced under controlled conditions in small scale bioreactor cultivations after 60–70 h of growth limiting glycerol feed. This is the highest concentration of secreted enzyme in P. pastoris published so far and single parts of the expression cassette could be independently optimized showing additive effects for improvements in protein production by P. pastoris. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The methylotrophic yeast Pichia pastoris is a unicellular eukaryotic micro-organism with many advantages compared to other microbial expression systems (Cregg et al., 2009; Macauley-Patrick et al., 2005; Zhang et al., 2009). Nevertheless, to further increase its value as host system for heterologous protein production new tools and techniques are required to push expression efficiencies and product quality to new limits and to simplify production processes. There are several well-known factors influencing protein production by P. pastoris, such as synonymous codon substitution, promoter choice and gene dosage (Abad et al., 2010a,b; Hartner et al., 2008; Macauley-Patrick et al., 2005; Mattanovich et al., 2004; Mellitzer et al., 2012b). However, most of these factors were not systematically studied in detail for P. pastoris so far. Synonymous

∗ Corresponding author. Tel.: +43 3168739300. E-mail addresses: [email protected], [email protected] (A. Glieder).

codon substitution studies performed in Escherichia coli obtained partly contradictory results. Welch et al. (2009) demonstrated that large variations of expression can be caused by synonymous codon substitution independent of local predicted mRNA structure near the translation initiation region and speculated that an imbalance of tRNA consumption may limit expression levels. Other researchers have shown that strong mRNA structure near the translation initiation site can impair expression (Kosuri et al., 2013; Kudla et al., 2009). For P. pastoris different results were expected, since post-translational processes (e.g. folding and secretion) and protein degradation of misfolded or unfolded proteins are common limiting factors for protein production (Mattanovich et al., 2004). The stress-induced extra energy requirement needed for recombinant protein production was summarized as metabolic burden (Glick, 1995). If the metabolic burden is too high this can result in inhibition of growth, low level of product accumulation and plasmid or genetic instabilities in yeasts. In case of P. pastoris genetic instabilities were observed in multi-copy strains by Zhu et al. (2009).

http://dx.doi.org/10.1016/j.jbiotec.2014.08.035 0168-1656/© 2014 Elsevier B.V. All rights reserved.

Please cite this article in press as: Mellitzer, A., et al., Synergistic modular promoter and gene optimization to push cellulase secretion by Pichia pastoris beyond existing benchmarks. J. Biotechnol. (2014), http://dx.doi.org/10.1016/j.jbiotec.2014.08.035

G Model BIOTEC-6837; No. of Pages 9

ARTICLE IN PRESS A. Mellitzer et al. / Journal of Biotechnology xxx (2014) xxx–xxx

2

The most popular promoters in P. pastoris are the wild-type promoters of the glyceraldehyde-3-phosphate dehydrogenase (GAP) (Waterham et al., 1997) and alcohol oxidase 1 (AOX1) (Tschopp et al., 1987) genes since they are strong promoters, easy to handle and available by commercial suppliers. However, due to the fact that PGAP is a constitutive promoter, it is not preferable for the expression of physiologically problematic or cytotoxic proteins (Macauley-Patrick et al., 2005). On the other hand PAOX1 , which can be tightly regulated by the available C-source (Hartner and Glieder, 2006; Tschopp et al., 1987) has the disadvantage that the use of methanol in industrial applications is not desired by some industries due to potential safety issues and potential methanolinduced cell lysis and proteolysis (Macauley-Patrick et al., 2005; Mattanovich et al., 2004; Menéndez et al., 2004; Zhang et al., 2009). Meanwhile, there are many alternatives. For example a set of synthetic reengineered PAOX1 promoter variants (Hartner et al., 2008) which were generated by selected short sequence deletions and enabled methanol-inducible and methanol-independent gene expression at varying strength or PGAP promoter variants generated by error-prone PCR (Qin et al., 2011). Efficient natural carbon source regulated promoters were identified by a transcriptomics approach and applied by Gasser et al. (2012) and Stadlmayr et al. (2010). The aim of the present work was to evaluate independently the potential of synonymous codon substitution, promoter choice and gene dosage for a well expressed gene and secreted enzyme in P. pastoris. Cellobiohydrolase 2 from Trichoderma reesei (TrCBH2,) was chosen as model enzyme since we previously showed that it is well expressed and secreted by P. pastoris (Mellitzer et al., 2012b). Moreover, T. reesei is well known for its cellulolytic activity and the main cellulases produced by this fungus are CBH1 and CBH2 (Miettinen-Oinonen et al., 2005). A gene set of 48 differently optimized TrCBH2 variants was expressed under the control of distinctly regulated promoters with different sequences requiring different growth conditions and cell physiology for protein expression. Well-defined strains were generated to further characterize the synthetic promoter performances on protein level in fed-batch bioreactor cultivations. 2. Materials and methods 2.1. Chemicals and materials Oligonucleotide primers were obtained from Integrated DNA Technologies (Leuven, Belgium). For plasmid isolation the GeneJETTM Plasmid Miniprep Kit of Fermentas (Burlington, Ontario, Canada) was used. All DNA-modifying enzymes were obtained from Fermentas GmbH (Burlington, Ontario, Canada). Chemicals were purchased if not stated otherwise from Becton, Dickinson and Company (Franklin Lakes, NJ, USA), Fresenius Kabi Austria (Graz, Austria) and Carl Roth (Karlsruhe, Germany). p-Hydroxybenzoic acid hydrazide (order no. 54600) and d-(+)-cellobiose were obtained from Fluka (Hamburg, Deutschland). Cellobiase from Aspergillus niger (C6105) and cellulase from T. reesei ATCC 26921 (C2730) were obtained from Sigma–Aldrich (St. Louis, MO, USA). 2.2. Media For E. coli standard LB-medium containing 25 ␮g/ml zeocin was used. YPD for P. pastoris contained 10 g/l yeast extract, 20 g/l peptone and 20 g/l glucose. For antibiotic selection 100 ␮g/ml zeocin were used. 15 g/l agar was added for plate media. Buffered minimal media BMD (1%), BMM2 and BMM10 consisted per liter of 200 ml 1 M potassium phosphate buffer (pH 6), 13.4 g yeast nitrogen base without amino acids, 0.0004 g/l biotin and 11 g/l glucose or

1 or 5% (v/v) methanol, respectively. All pre-cultures were prepared using YPhyD medium containing 20 g/l phytone–peptone, 10 g/l bacto-yeast extract and 20 g/l glucose. BSM medium contained per liter CaSO4 2H2 O 0.47 g, K2 SO4 9.1 g, KOH 2.07 g, MgSO4 7H2 O 7.5 g, EDTA 0.6 g, H3 PO4 (85%) 13.4 ml, glycerol 40.0 g, NaCl 0.22 g and 4.35 ml PTM1. PTM1 trace elements solution contained per liter 0.2 g biotin, 6.0 g CuSO4 5H2 O, 0.09 g KI, 3.0 g MnSO4 H2 O, 0.2 g Na2 MoO4 2H2 O, 0.02 g H3 BO3 , 0.5 g CoCl2 , 42.2 g ZnSO4 7H2 O, 65 g Fe(II)SO4 7H2 O and 5 ml H2 SO4 . The fed-batch feed media were either 60% (w/w) Glycerol or concentrated MeOH and were supplemented with 12 ml/l PTM1 mineral salts solution. 2.3. Design of initial gene variant sets Gene sequences encoding cellobiohydrolase 2 (TrCBH2) from T. reesei (Mellitzer et al., 2012b) including the native secretion leader were designed and synthesized by DNA2.0 using the process described in detail in Welch et al. (2009). A total of 48 gene variants were designed by back-translating the TrCBH2 protein sequence using a Monte Carlo repeated random sampling algorithm to design a set of maximally systematically varied TrCBH2-encoding gene variants (Welch et al., 2009). This algorithm selects a codon for each position at a probability defined in a codon frequency lookup table. A variety of different lookup tables and constraints were applied to create variant designs that differed in the parameters that have been associated with expression effects in the literature (Gustafsson et al., 2012). Codon usage was varied in bias on a codon-by-codon basis relative to P. pastoris genomic bias and in the inclusion or exclusion of seven naturally rare codons (defined as codons used at a frequency lower than 0.25× that of the highest frequency codon for the cognate amino acid). Codon usage frequency ranges sampled are tabulated in Table S1 and a correlation between usage frequencies of pairs of codons within the 48 TrCBH2 gene set can be seen in Fig. S1. We also specifically varied the first 15 codons toward higher or lower GC content. Across the set, GC content in these codons ranged from 33 to 53%. Overall gene GC content ranged from 46 to 53%. A 48-run Plackett–Burman Design of Experiments (DoE) methodology (Hair et al., 1998) was used to minimize co-variation of the constraints and maximize systematic diversity among the variants. GC and purine/pyrimidine biases in codon usage for each amino acid were independently varied. This experimental design approach allows for the identification and quantification of independently contributing factors. Due to the use of Monte-Carlo sampling in the gene design, all of these variants were highly divergent in sequence identity from each other (Fig. S2). Average pair-wise nucleotide sequence similarities were 77 ± 1%. Based on our results, the most important gene sequences are shown in Figs. S3, S4 and S5. 2.4. Construction of P. pastoris strains To further optimize translation all genes were cloned after a defined Kozak consensus sequence (gaaacg). The synthetic genes were cloned into the multiple cloning site of the E. coli/P. pastoris shuttle vector pPpB1 (Abad et al., 2010a,b) via EcoRI/NotI sites. This vector is based on the plasmid pPpT4 (Näätsaari et al., 2012), but with a weaker promoter for the selection marker. The TrCBH2 gene variants were cloned downstream of the wild type promoters PGAP and PAOX1 and synthetic promoter variants with distinctly different regulation patterns were also included, namely PEn and PDeS . PEn can be induced by methanol and showed more than 60% higher GFP expression compared to the wild-type promoter PAOX1 . In Hartner et al. (2008) this promoter is referred to as P(d1). The PDeS promoter is based on the sequence of P(d6*) variant of the P. pastoris AOX1 promoter (Hartner et al., 2008) and enables strong transcription based on simple derepression without necessity to induce by

Please cite this article in press as: Mellitzer, A., et al., Synergistic modular promoter and gene optimization to push cellulase secretion by Pichia pastoris beyond existing benchmarks. J. Biotechnol. (2014), http://dx.doi.org/10.1016/j.jbiotec.2014.08.035

G Model BIOTEC-6837; No. of Pages 9

ARTICLE IN PRESS A. Mellitzer et al. / Journal of Biotechnology xxx (2014) xxx–xxx

methanol. Plasmids were linearized with either BglII or BamHI, subsequently purified and concentrated using the Wizard SV Gel and PCR Cleanup System (Promega Corp.). Electro-competent P. pastoris CBS 7435 mutS cells (Näätsaari et al., 2012) were prepared and transformed with 0.1- to 2 ␮g of the BglII-linearized pPpB1 or pPpT4 vector construct according to Lin-Cereghino (Lin-Cereghino et al., 2005). Transformants were plated on YPD-Zeocin (100 ␮g/ml Zeocin) agar plates and grown at 28 ◦ C for 48 h. 2.5. Micro-scale cultivation and high-throughput screening P. pastoris strains expressing TrCBH2 were cultivated in 96-deep well plates as described by Weis et al. (2004). Incubation was done in shakers (INFORS Multitron, Bottmingen, Switzerland) at 28 ◦ C, 320 rpm, and 80% relative humidity. After an initial batch phase for 60 h on 1% glucose the cultures were induced with 0.5% of methanol in the media for a total of 72 h for expression under the control of PAOX1 . Glucose was used instead of methanol for PGAP and for PDeS sorbitol was used instead of methanol, similar to the methanol induction procedure. After induction the cells were pelleted at 4000 rpm and enzymatic activities were determined in the supernatants using the pHBAH-assay as previously described by Mellitzer et al. (2012a). Substrate conversions were performed in 50 mM citrate buffer containing 1% Avicel® at 50 ◦ C for 2 h. For the subsequent reducing sugar assay 50 ␮L of the substrate reaction (or, in the case of the standard sugars, appropriate dilutions of the reducing sugars) were pipetted into 150 ␮l of the pHBAH working solution in a 96-well PCR plate. The plate was sealed and incubated at 95 ◦ C for 5 min and then cooled to 4 ◦ C. 150 ␮l of the assay samples were transferred to a new micro-titer-plate and the absorption measured at 410 nm in a SPECTRA MAX Plus384 plate reader (Molecular Devices Corp., Sunnyvale, CA, USA). For exact quantification of reducing sugars a standard curve of the reducing sugar (0–1 mg/ml) was included on each plate. Activity units refer to the amount of released reducing sugar over time and correspond to the standard IUPAC definition ␮M/min. 2.6. Copy number determination by quantitative real-time PCR Copy numbers of integrated expression cassettes in the Pichia genome were determined using quantitative real-time PCR (qRTPCR) as described by Abad et al. (2010a,b). 2.7. Fed-batch small scale bioreactor cultivations Pre-cultures of individual strains were grown in 50 and 200 ml YPhyD medium containing 20 g/l Bacto-Yeast Extract and 20 g/l glucose in wide-necked, baffled shake flasks at 120 rpm at 28 ◦ C. BSM medium contained per liter CaSO4 2H2 O 0.47 g, K2 SO4 9.1 g, KOH 2.07 g, MgSO4 7H2 O 7.5 g, EDTA 0.6 g, H3 PO4 (85%) 13.4 ml, Glycerol 40.0 g, NaCl 0.22 g and 4.35 ml PTM1. PTM1 Trace elements solution contained per liter 0.2 g Biotin, 6.0 g CuS04 5H2 O, 0.09 g KI, 3.0 g MnSO4 H2 O, 0.2 g Na2 MoO4 2H2 O, 0.02 g H3 BO3 , 0.5 g CoCl2 , 42.2 g ZnSO4 7H2 O, 65 g Fe(II) SO4 7H2 O and 5 ml H2 SO4 . Each bioreactor (Infors Multifors system (Infors AG, Bottmingen, Switzerland)) containing 450 ml BSM-media (pH 5.0) was inoculated from the pre-culture to an OD600 of 2.0. During the batch phase P. pastoris was grown on glycerol (4%) at 28 ◦ C. At the beginning of the glycerol feeding phase the temperature was decreased to 24 ◦ C. For methanol-fed cultures, the fed-batch phase was started upon depletion of initial batch glycerol with 16 g/(l h) glycerol feed solution followed by methanol induction. In the early stages, the methanol-feed was set to 2 g/(l h) and was gradually increased within the next 70 h to 6 g/(l h). Likewise, the glycerol-feed was phased down during the first hour of methanol induction to 0 g/(l h). Dissolved oxygen was set to 30% throughout the whole process.

3

After 91.5 h of methanol induction the bioreactor cultivations were stopped. For glycerol-fed strains, the batch phase was directly followed by a constant glycerol-feed with 5 g/(l h). 2.8. Determination of protein concentration Protein concentrations were either determined by the Bradford method or by microfluidic capillary electrophoresis (CE). For the Bradford method the Bio-Rad Protein Assay kit (cat no. 500-0201) was carried out according to the manufactures instructions based on the microassay protocol. For protein concentrations determined by CE a fluorescence detection system was used (LabChip® GX II Gel Electrophoresis, PerkinElmer, USA). Standard deviations of this robust system are usually below 10%, even at high protein loads. Therefore, just single measurements of every sample were performed if not stated differently. More specifically, proteins were quantified by calibrating the integrated areas of the protein-specific peaks in the electropherograms to an external reference protein standard (BSA) of known concentration. For glycosylated proteins, peak areas of diluted deglycosylated samples were compared to those of untreated samples to compensate for glycosylation-related differences in quantification. Samples were treated with EndoH for deglycosylation according to the manufactures instructions (Biolabs, catalog# P0702L). The dilutions of samples were in a range to give peak areas of the samples that were comparable to those of the reference protein standard. 3. Results 3.1. Design of the expression constructs The effects of synonymous codon substitution and promoter choice were studied in more detail to improve heterologous protein expression in P. pastoris. An initial set of 48 TrCBH2 genes was made by DNA2.0 employing an in-silico design setup, which applied distinct biases to various gene features, including codon usage frequencies. This design allowed covering a broad range of variation while minimizing co-variation between the features. This was important to be able to distinguish independent effects of the features. For all genes, known splicing and polyA motifs as well as large GC, AT and homonucleotide runs of 6 or more were avoided. This was to avoid local extremes, which have been associated with various local effects, such as translational frame shifts and transcriptional termination. No special mRNA structures near the translational start or elsewhere were included in the design scheme and CAI (codon adaptation index) or the tAI (tRNA adaptation index) were not specifically included as variables in the DoE, although CAI/tAI and related global variables are indirectly varied as a consequence of the codon specific gene variations and such sequence–function relationship can be reconstructed from the experimental data. Possible correlations between synonymous codon substitution and expression effects under different physiological production conditions were also examined by employing three different promoters, namely PGAP , PEn and PDeS . PGAP constitutively expressed the 48 different TrCBH2 genes whereas PEn was induced by MeOH addition similar to PAOX1 (Hartner et al., 2008). PDeS is an enhanced synthetic promoter which is repressed at high concentrations of e.g. glycerol or glucose and can be induced by derepression. Induction of this promoter occurs either by low amounts of repressing C-sources or similar to PAOX1 , also by MeOH addition. The genes were sequence-verified and cloned into the P. pastoris shuttle vector system (pPpT4). Due to a strong promoter for the selection marker single copy integration of the expression cassette is favored especially if low amounts (