The methylotrophic yeast Pichia pastoris is an established host system for producing heterologous proteins for both academic and commercial purposes. P. pastoris combines benefits from bacterial and eukaryotic expression systems (Cregg et al., 1985, 2009; Lin-Cereghino et al., 2002; Lin-Cereghino and Cregg, 2000; Macauley-Patrick et al., 2005). P. pastoris is also a model system for studying peroxisome biogenesis and autophagy (Faber et al., 1998) and the organization and biogenesis of secretory pathway organelles (Esaki et al., 2006; Soderholm et al., 2004). Research into the metabolic features of P. pastoris (Caspeta et al., 2012; Chung et al., 2010; Jorda et al., 2012; Sohn et al., 2010; Solà et al., 2004, 2007) suggests that it is also a promising platform for primary and secondary metabolic engineering (Araya-Garay et al., 2012; Gasser et al., 2010; Marx et al., 2008).
When P. pastoris is used for recombinant protein production, heterologous protein expression levels are commonly increased by introducing multiple copies of an expression vector (Sunga and Cregg, 2004; Sunga et al., 2008). This is achieved through multiple transformations, using vectors with the same expression cassette but different selectable markers. Increasing the number of integrated mini-proinsulin cassettes by retransformation, for example, increases mini-proinsulin yield by > 10-fold (Mansur et al., 2005). The retransformation method is also useful for introducing multiple genes for metabolic engineering. In a recent study, two haploid P. pastoris strains were mated: one strain carried an expression cassette for the light chain of anti-HER2 and the other carried a cassette for the heavy chain. Markers were for Zeocin or nourseothricin resistance. The resulting diploid secreted fully assembled anti-HER2 into the medium (Chen et al., 2012).
Several selectable marker genes are currently available, including biosynthetic genes for arginine, adenine, histidine, uracil or methionine (Lin-Cereghino and Cregg, 2000; Nett and Gerngross, 2003; Nett et al., 2005; Thor et al., 2005); and dominant antibiotic-resistance genes. The advantage of antibiotic-resistance markers is that they allow straightforward selection of strains with a high copy number of the integrated plasmid. Alternatively, plasmid copy number can be amplified by post-transformational vector amplication (PTVA), which is cultivating transformants in a medium containing increasingly higher concentrations of drug (Sunga et al., 2008). Strains originally transformed with one or a few vector copies can be subjected to further selection with high levels of drug, even long after initial transformation. Selection typically results in clones containing multiple head-to-tail copies of the entire vector integrated at a single genomic locus. A similar method, chemically inducible chromosomal evolution (CIChE), is reported for bacteria (Tyo et al., 2009), with recA homologous recombination used to evolve a chromosome with approximately 40 consecutive copies of the target genes.
Currently, the dominant antibiotic makers that are available for P. pastoris are limited to the Sh ble gene from Streptoalloteichus hindustanus (Zeocin resistance) (Drocourt et al., 1990), the blasticidin S-deaminase gene from Aspergillus tererus (blasticidin resistance) (Kimura et al., 1994), the nourseothricin acetyltransferase gene from Streptomyces noursei (nourseothricin resistance) (Chen et al., 2012) and the kanamycin-resistance gene of transposon Tn903 (G418 resistance) (Lin-Cereghino et al., 2008; Papakonstantinou et al., 2009; Scorer et al., 1994).
In this study, we describe new P. pastoris expression vectors that confer hygromycin resistance from the Klebsiella pneumoniae hph gene under the control of the Saccharomyces cerevisiae transcription elongation factor TEF1 promoter. We demonstrated the application of the new vectors by expressing green fluorescent protein (GFP) and human serum albumin (HSA) in P. pastoris. We also demonstrated the application of the vectors for introducing multiple genes into P. pastoris by transforming P. pastoris with the S. cerevisiae genes GSH1, GSH2 and SAM2, on vectors that confer Zeocin, G418 and hygromycin resistance as dominant selectable markers.
Materials and methods
Strains, plasmids and growth conditions
Strains and plasmids used or constructed in this study are in Table 1. P. pastoris strains X33, GS115 (Mut+) and KM71 (MutS) were used (Invitrogen). Recombinant DNA manipulations were in Escherichia coli strain DH5α. Yeast strains were grown in YPD [1% yeast extract (Oxoid or Angel Yeast), 2% peptone (Oxoid or Angel Yeast), 2% glucose] or minimal medium [1.4% yeast nitrogen base, 4 × 10–5% biotin, carbon source] with 50 mg/l amino acids, 100 mg/l Zeocin (Invitrogen), 200 mg/l G418 (Invitrogen) or 200 mg/l hygromycin (Sinoreagent or Hisun) as required. For minimal medium, carbon sources were 2% glucose (MD medium), 2% glycerol (MGY medium) or 0.5% methanol (MM medium). E. coli strains were cultured in low-salt Luria–Bertani (LB) medium (1% tryptone, 0.5% yeast extract, 0.5% NaCl, pH 7.5) supplemented with 100 mg/l ampicillin, 50 mg/l kanamycin, 25 mg/l Zeocin or 100 mg/l hygromycin, as required.
Table 1. Strains and plasmids used in this study
Strain or plasmid
AmpR, ampicillin resistance; KmR, kanamycin resistance; ZeoR, Zeocin resistance; HygR, hygromycin resistance; GFP, green fluorescent protein; HSA, human serum albumin.
E. coli DH5α
General cloning host strain
S. cerevisiae BY4742
General S. cerevisiae strain
his4, host strain
Wild-type, host strain
his4,arg4,aox1::ARG4, host strain
GFP expression strain
GFP expression strain
HSA expression strain
HSA expression strain, derived from H30 via PTVA
S. cerevisiae SAM2, GSH1 and GSH2 co-expression strain
Oligonucleotides used are shown in Table 2. PCR reactions were performed using Taq (Fermentas) or KOD (Toyobo) polymerase.
Table 2. Oligonucleotide primers used in this study
Enzyme restriction sites are underlined.
Construction of the expression vectors
A 1.7 kb DNA fragment containing the hygromycin-resistance gene ORF, the TEF1 promoter and the TEF1 transcription terminator sequences was amplified from pAG32 (Goldstein and McCusker, 1999), using the primer pair M13R/Ttef_dn and digested with BglII. The resulting fragment was used to replace the 1.4 kb BamHI/EcoRV region containing the Zeocin-resistance gene in the expression plasmids pZAB (Yang et al., 2009) and pGAPZB. The resulting hygromycin plasmids were named pAHYB and pGHYB, with the the inducible AOX1 in pAHYB and the constitutive GAP promoter in pGHYB for target gene expression.
We also constructed the hygromycin-resistance plasmid series pPICHY(A–C), pPICHYα(A–C), pGHY(A–C) and pGHYα(A–C), by replacing the Zeocin resistance gene of the pPICZ(A–C), pPICZα(A–C), pGAPZ(A–C) and pGAPZα(A–C) vectors (Invitrogen) with the hgh gene using the procedure described above.
Construction and microscopy of GFP expression strains
To construct GFP expression plasmids, a 0.8 kb KpnI–SpeI fragment from pEGFP (Clontech) was cloned into pGHYB and pAHYB, using the KpnI and XbaI sites. The resulting plasmids, pGHYB–GFP and pAHYB–GFP, were linearized with PmeI or BlnI and transformed into GS115 cells by electroporation. Hygromycin-resistant transformants were selected on YPD plates with 200 mg/l hygromycin and colonies were randomly picked for PCR confirmation (Thor et al., 2005). Two strains with correct gene integration were designated G20 and G41. Empty pGHYB and pAHYB vectors were also transformed into GS115, following the same procedure as controls, yielding strains G00 and A00.
For visualization of GFP, yeast cells were grown for 48 h in MM or MGY medium supplemented with histidine. Fluorescence microscopy was performed at ×400 magnification using an Olympus Fluoview FV1000 microscope, with excitation at 488 nm and emission at 510 nm.
Construction of HSA expression strains
The HSA coding region fragment was synthesized using the HSA sequence (GenBank NM_000477.5) with BamHI–NotI sites added. A 2 kb fragment containing the HSA coding region was cloned into pPIC9 (Invitrogen) at BamHI–NotI sites. The resulting plasmid was digested with AsuII–NotI and the resulting fragment cloned into pAHYB at the same sites, generating pAHYB–HSA, which contains the HSA gene under control of the AOX1 promoter.
The plasmid pAHYB–HSA was linearized with PmeI and transformed into X33 cells by electroporation. Five transformants were tested for HSA expression. The strain with the highest HSA yield was named H30. HSA was also cloned into the pZAB (Yang et al., 2009) plasmid and transformants selected using Zeocin.
PTVA was performed by culturing strains with low plasmid copy number on YPD plates supplemented with increasingly higher concentrations of hygromycin (200, 700, 1500 and 4000 mg/l). Each round of selection was carried out for 48–72 h. PTVA-enriched strains were analysed by PCR. Five clones that demonstrated better growth, as larger colonies, at 4 g/l hygromycin were investigated further. The strain with the largest colonies was named H33. After the first round of PTVA, the strain with the highest yield of HSA was used in the second round of PTVA with higher concentrations of hygromycin (4000 and 5000 mg/l). Colonies were analysed by PCR for their genotype and shake-flask culture for the yield of HSA. Primers aoxup500 and HSAR350 were used to analyse HSA integration. Primer aoxup500 binds upstream of the AOX1 promoter and HSAR350 binds to the HSA ORF. Primers ColE_F and HSAR350 were used to analyse vector amplification. Primer ColE_F binds to the vector backbone (Figure 3). Genetic stability of the strains was determined by 20 rounds of streaking on YPD plates without antibiotic.
HSA expression was determined by SDS–PAGE and HPLC. HSA proteins were separated on a TSK gel G3000SWxL (Tosoh Corp.) column, using 200 mm phosphate 1% isopropanol buffer. Elution of analytes was recorded spectrophotometrically by absorbance at 280 nm. HSA activity was determined using affinity for aspirin and digoxin (Sun Yat-sen University, Guangzhou, China).
Introducing SAM2, GSH1 and GSH2 into P. pastoris
To co-express the S. cerevisiae SAM2, GSH1 and GSH2 genes in P. pastoris, three expression plasmids with different selection makers were constructed. A vector with the SAM2 expression cassette and the Zeocin resistance marker was generated by amplifying a 1.3 kb fragment of SAM2 from genomic DNA from S. cerevisiae BY4742, using the primer pair SAMUP–SAMDN (Table 2), and subcloned into pMD18-T (Takara). A 1.3 kb AsuII/EcoRI fragment containing SAM2 was inserted into pGAPZB to yield pGZ-SAM.
A vector with GSH1 and the hygromycin resistance marker was generated by amplifying a 2 kb DNA fragment with the GSH1 gene, using the primer pair GSH1UP–GSH1DNM (Table 2) from S. cerevisiae, and subcloning into pMD18-T (Takara). A 2 kb AsuII–SalI fragment containing GSH1 was inserted into pGHYB to yield pGHYB-GSH1.
For the GSH2 expression vector, a G418/Kan resistance vector was constructed as described previously (Papakonstantinou et al., 2009). A fragment containing the G418/Kan resistance gene was amplified from pUG6 (Guldener et al., 1996), using the primer pair kanMXup/Ttef_dn (Table 2), and used to replace the NcoI–EcoRV region containing the ZeoR gene of pGAPZB, yielding the expression vector pGKB. A 1.5 kb GSH2 fragment, amplified using the primer pair GSH2UP/GSH2DNE (Table 2), was inserted into pGKB following the strategy described above for cloning SAM2 and GSH1.
The three plasmids with SAM2, GSH1 or GSH2 were transformed into GS115 cells by electroporation. Strains resistant to all three antibiotics were selected and the largest colony on YPD with G418, Zeocin and hygromycin was selected and named S12.
To determine S-adenosyl-l-methionine (SAM) and glutathione levels, a single colony of S12 was used to inoculate 5 ml liquid YPD and cultured at 30ºC and 250 rpm for 18 h. From this seed culture, 500 µl was used to inoculate a 500 ml shake flask, containing 50 ml BMGY medium with 0.01 m cysteine and 0.01 m methionine, that was cultivated at 30ºC and 250 rpm for 96 h. Every 24 h, glycerol was added to a final concentration of 1%, cysteine was added to 0.01 m and methionine was added to 0.01 m. Yields of SAM and glutathione from cultures were analysed by HPLC, as described previously (Fei et al., 2009; Yu et al., 2003).
New P. pastoris expression vectors were generated (Figure 1). The plasmids contain an AOX1 or GAP promoter cassette, the hph gene from K. pneumoniae under the control of the TEF1 promoter as a dominant selection marker, and the colE origin for high-copy replication and plasmid maintenance in E. coli. The hph gene encodes hygromycin B phosphotransferase, which confers resistance to the antibiotic hygromycin B in both E. coli and yeast.
Multiple restriction sites between the AOX1 or GAP promoter and the AOX1 terminator can be used for further cloning. We constructed several plasmids based on these vectors and used them for heterologous protein expression. Heterologous protein was expressed under the control of the inducible AOX1 promoter or the constitutive GAP promoter.
Intracellular expression of GFP
We first tested whether the hygromycin vectors could be used for intracellular expression of the model protein GFP. Two plasmids containing expression cassettes were constructed, one with GFP under control of the inducible AOX1 promoter, and one with GFP under control of the constitutive GAP promotor. The plasmids were transformed into the P. pastoris GS115 strain. Vector integration was confirmed by PCR for hygromycin-resistant transformants grown on YPD. Transformants were transferred to minimal medium for GFP expression. Two strains transformed with empty vector acted as negative controls.
Fluorescence microscopy demonstrated constitutive GFP expression from the strain with GFP under GAP promoter control and methanol-inducible GFP expression from the strain with GFP under AOX1 promoter control (Figure 2). No fluorescence was observed in control strains grown in glycerol or methanol medium.
Secretion of HSA
HSA was used as a reporter to determine whether the hygromycin vectors could be used for secretion of heterologous protein (Ohtani et al., 1998; Zhao et al., 2008). Yeast transformed with an expression vector for HSA secreted HSA into the culture medium after induction (Figure 3C).
We also tested whether the hygromycin resistance gene was suitable for amplification by PTVA. An initial transformant strain, H30, with a single copy of a hygromycin-resistance vector, was resistant to no more than 500 µg/ml hygromycin before PTVA. PTVA yielded strains with resistance to hygromycin at > 4 mg/ml. PCR was used to analyse integration and vector amplification. All strains with the hygromycin-resistance plasmid produced a 2 kb band, before or after PTVA, that demonstrated that HSA integrated into the AOX1 locus of the P. pastoris genome. Only strains after PTVA produced a band of 1.7 kb, demonstrating vector amplification. (Figure 3A, B). Before and after PTVA, five strains were tested by HPLC for HSA yield. HSA yield increased after each round of PTVA using the hygromycin resistance marker. These results were similar to results for PTVA using the Zeocin resistance marker (Figure 3C). All HSA samples showed approximately the same affinity for aspirin and digoxin.
Strain stability was determined by consecutive streak cultivation of strains H33 on YPD. After 20 rounds of streaking, 30 single colonies were tested by PCR, with 28 producing a band of 1.7 kb with primers ColE_F and HSAR350, which indicated at least two copies of the HSA expression vector integrated in the genome. Five colonies that gave positive PCR results were tested by HPLC and shown to produce HSA at levels similar to those of strain H33 (data not shown).
Transformation of P. pastoris with SAM2, GSH1 and GSH2 vectors using three dominant selectable markers
To determine whether the new hygromycin-resistance plasmids could be used in with Zeocin-resistance and G418-resistance plasmids to introduce multiple genes into P. pastoris, we transformed the S. cerevisiae genes SAM2, GSH1 and GSH2 into P. pastoris GS115 and selected using the three dominant selectable markers. The genes encode enzymes in the synthesis of the sulphur amino acid derivatives SAM and glutathione (Figure 4A) (Fei et al., 2009; Thomas and Surdin-Kerjan, 1997; Yu et al., 2003). Co-expression of the genes should result in increased production of glutathione and SAM when a strain with the genes is cultured in medium supplemented with methionine and cysteine (Wang et al., 2012).
The integration of SAM2, GSH1 and GSH2 into the P. pastoris genome was confirmed by PCR (Figure 4B). The level of glutathione and SAM increased approximately two-fold compared to the initial strain GS115. In 72 h, glutathione was produced at 0.29 g/l and SAM at 0.19 g/l (Figure 4C, D).
Tools for genetically modifying P. pastoris are widely available, but the number of dominant selectable markers is currently limited. The aim of this study was to expand the current range of vectors and provide a means for co-expressing several genes in the same strain.
We constructed new P. pastoris expression vectors that confer hygromycin resistance using the K. pneumoniae hph gene. The new plasmids contain an AOX1 or GAP promoter cassette, the hph gene under control of the TEF1 promotor as a dominant selection marker, and the colE origin for high-copy replication and maintenance of the plasmid in E. coli. The vectors were used for both intracellular GFP and secreted HSA protein expression.
GFP from the jellyfish Aequorea victoria is a widely used reporter protein that is species-independent and readily visualized in both prokaryotic and eukaryotic cells by its fluorescence. GFP can be expressed in P. pastoris (Lin-Cereghino et al., 2008; Papakonstantinou et al., 2009; Yang et al., 2009). In this work, GFP expression from the hygromycin-resistance vector was similar to expression from a Zeocin-resistance vector or expression as an unmarked knock-in gene (Yang et al., 2009).
HSA is the most abundant protein in plasma and recombinant HSA is approved as a medicine. HSA is efficiently secreted from P. pastoris. In this study, we constructed several strains with two or more copies of HSA integrated in the P. pastoris genome. HSA produced by these strains was active, as determined by affinity for aspirin and digoxin. In ongoing work, we are screening for strains with increased HSA yield.
To demonstrate the compatibility of the hygromycin-resistance marker with other antibiotic-resistance markers, we co-introduced GSH1, GSH2 and SAM2 into P. pastoris using three dominant drug-resistance markers. Co-transformation resulted in a strain with a two-fold increase in glutathione and SAM production compared to the parental strain.
We also demonstrated that the hygromycin-resistance expression vectors could be used in PTVA to increase plasmid copy number, similar to reports for vectors containing genes for Zeocin and G418 resistance (Sunga et al., 2008). In metabolic engineering and synthetic biology, a challenge is balancing the expression of proteins with optimal flux of an engineered heterologous metabolic pathway for high yield and productivity. PTVA offers the potential to produce strains with optimal protein expression, by using media with different concentrations of drugs to screen for strains with different copy numbers of vectors containing the target gene.
Multiple copies of expression vectors for a single protein could be transformed into P. pastoris by multiple rounds of transformation with vectors with different dominant antibiotic-resistance markers. Several rounds of PTVA with higher increasing concentrations of multiple antibiotic drugs might yield a strain with a very high copy number of the target gene. This could be a potential method for increasing the expression of already well-expressed proteins and poorly expressed proteins.
Our data demonstrated that the new hygromycin-based vectors presented here are useful tools for protein expression in P. pastoris. The vectors can be used on their own or in conjunction with existing vectors.
The authors thank Professor Sheng Yang (Institute of Plant Physiology and Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences) for supporting this research. We thank Pavel Shliaha and Daniel Nightingale (University of Cambridge, UK) for English language editing. We thank Professor Benjamin S. Glick (University of Chicago) for providing SnapGene software for producing plasmid maps. We also thank the anonymous referees and editor for their comments and constructive suggestions. This study was supported by the National Major Scientific and Technological Special Project for ‘Significant New Drugs Development’ of China (Grant No. 2011ZX09202-301), the National Basic Research Programme (973 Programme) of China (Grant No. 2014CB745100) and the National Key Technologies R&D Programme of China (Grant No. 2012AA022101).