Available online at www.sciencedirect.com

Journal of Biotechnology 133 (2008) 424–432

High-level expression of the native barley ␣-amylase/subtilisin inhibitor in Pichia pastoris Pernille Ollendorff Micheelsen a,∗ , Peter Rahbek Østergaard b , Lene Lange c , Michael Skjøt d a

Copenhagen Biocenter, Department of Molecular Biology, University of Copenhagen, Ole Maaløes Vej 5, 2200 Copenhagen N, Denmark b Protein Biochemistry, Novozymes A/S, Krogshøjvej 36, 2880 Bagsværd, Denmark c Institute of Molecular Biology, University of Copenhagen, Ole Maaløes Vej 5, 2200 København N, Denmark d Protein Diversity, Novozymes A/S, Brudelysvej 26, 2880 Bagsværd, Denmark Received 16 August 2007; received in revised form 15 October 2007; accepted 28 November 2007

Abstract An expression system for high-level expression of the native Hordeum vulgare ␣-amylase/subtilisin inhibitor (BASI) has been developed in Pichia pastoris, using the methanol inducible alcohol oxidase 1 (AOX1) promoter. To optimize expression, two codon-optimized coding regions have been designed and expressed alongside the wild-type coding region. To ensure secretion of the native mature protein, a truncated version of the alpha mating factor secretion signal from Saccharomyces cerevisiae was used. In order to be able to compare expression levels from different clones, single insertion transformants generated by gene replacement of the AOX1 gene was selected by PCR screening. Following methanol induction, expression levels reached 125 mg L−1 from the wild-type coding region while expression from the two codon-optimized variants reached 65 and 125 mg L−1 , respectively. The protein was purified and characterized by Edman degradation, liquid chromatography mass spectrometry and insoluble blue starch assay, and was shown to posses the same characteristics as wild-type protein purified from barley grains. © 2007 Elsevier B.V. All rights reserved. Keywords: Barley ␣-amylase/subtilisin inhibitor (BASI); Pichia pastoris; Codon optimization; Heterologous expression; Truncated alpha mating factor secretion signal

1. Introduction Over the past two decades, the enzyme industry has become a billion dollar industry; one of the reasons for this massive expansion is the desire to use cleaner and more environmentally friendly biological processes and products. Efforts have been put into developing high yield expression systems in various hosts. Pichia pastoris is one of the eukaryotic expression hosts that have been developed into a highly successful expression system (Macauley-Patrick et al., 2005). Research conducted so far on P. pastoris has focused primarily on expression of proteins of vertebrate origin (Cereghino and Cregg, 2000). HowAbbreviations: BASI, barley ␣-amylase/subtilisin inhibitor; AOX1, alcohol oxidase 1; PCR, polymerase chain reaction; AMY2, ␣-amylase 2; WASI, wheat ␣-amylase/subtilisin inhibitor; RASI, rice ␣-amylase/subtilisin inhibitor; CDR, coding region. ∗ Corresponding author. Tel.: +45 44460675; fax: +45 44980246. E-mail address: [email protected] (P.O. Micheelsen). 0168-1656/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jbiotec.2007.11.012

ever, the plant kingdom possesses a number of unique proteins of great industrial interest, including starch modifying enzymes (Mohnen, 1999), cytochromes P450 (Morant et al., 2003), sweettasting proteins (Faus, 2000) and proteins with pharmacological properties (Kong et al., 2003). Previously most plant proteins expressed in P. pastoris resulted in low yields (Cereghino and Cregg, 2000). Expression of plant proteins in filamentous fungi has similarly proven to be very difficult (Gouka et al., 1997). However, within the last couple of years, plant proteins have successfully been expressed in P. pastoris at high levels (Fierens et al., 2004; Yoneyama et al., 2007; Pan et al., 2007). It has previously been shown that organisms display a nonrandom pattern of synonymous codon usage (Grantham et al., 1980), and that the codon usage in yeasts correlates with the tRNA abundance (Ikemura, 1982; Sharp et al., 1986). Additionally, GC content in plant genes varies, dicots have a narrow and symmetrical GC-distribution centred around 46% GC, while the GC-distribution in Gramineae (essentially barley, maize and wheat) is broad and asymmetrical between 60 and 70% (Montero

P.O. Micheelsen et al. / Journal of Biotechnology 133 (2008) 424–432

et al., 1990). Therefore, by optimizing codon usage and GC content of plant genes, to that preferred for P. pastoris, one might overcome some of the limitations of heterologous expression of plant proteins. In this study we present successful high-level expression of the Hordeum vulgare ␣-amylase/subtilisin inhibitor (BASI) in P. pastoris. BASI has previously been expressed in P. pastoris, but yields were very low (Bønsager et al., 2003). BASI is a doubleheaded inhibitor, which acts both on ␣-amylase 2 (AMY2) from barley, and on serine proteases of the subtilisin family. It belongs to the Kunitz-type trypsin inhibitor family (Leah and Mundy, 1989; Mundy et al., 1983; Svendsen et al., 1986). BASI is a single chain protein consisting of 181 amino acids and contains two disulphide bridges, which are conserved in the structure of Kunitz inhibitors. BASI shares 92 and 58% sequence identity with inhibitors from wheat (WASI) and rice (RASI) (Mundy et al., 1984; Ohtsubo and Richardson, 1992), respectively. The biological role of BASI is still unclear; it has been suggested to have two functions: control degradation of starch during premature sprouting (Mundy et al., 1983) and protect the seed from subtilisin-type serine proteases from pathogens (Pekkarinen and Jones, 2003), for review see Bellincampi et al. (2004). The prospects of controlling the activity of both proteases and amylases are significant: ␣-amylases constitute the first step in the enzymatic degradation of starch, glycogen and related polysaccharides, and are therefore very important for carbohydrate metabolism in microorganisms, plants and animals. Inhibition of insecticidal ␣-amylases could spare the agricultural production for severe crop looses, while inhibition of ␣-amylases in humans reduces the post-prandial glucose peak, and have been proposed as a therapeutic agent for both diabetes and obesity (Sørensen et al., 2004). Virus-encoded proteases have emerged as novel targets for antiviral intervention, and small-molecule protease inhibitors have become a major research area (reviewed by Patick and Potts, 1998). Proteinaceous protease inhibitors could also be valuable in this context. Proteases are extensively used in various industries, and within some of these applications the activity of the proteases occasionally needs to be controlled. However, in order to be able to characterize and evaluate the potential effects of proteinaceous protease inhibitors in these diverse applications, purified protein is needed, an expression system therefore needs to be developed. In this paper a high-level expression system for the bifunctional ␣-amylase/subtilisin inhibitor from barley is presented. Compared to previous attempts to express BASI in P. pastoris (Bønsager et al., 2003), various conditions have been optimized in this study, including signal sequence, codon usage, transformation technique and growth conditions. Two codon-optimized coding regions (CDRs) have been designed and used for expression studies alongside with the wild-type gene. The mature part of the CDRs have been fused to a truncated version of the alpha mating factor secretion signal from S. cerevisiae, resulting in expression levels of 125 mg L−1 from single-copy transformants harbouring the wild-type CDR, whereas expression from the two codon-optimized variants yielded 65 and 125 mg L−1 , respectively.

425

2. Materials and methods 2.1. Strains, culture media and growth conditions P. pastoris (Invitrogen, USA) GS115 was used for heterologous expression of BASI. It was grown and maintained in YPD medium (1% (w/v) yeast extract, 2% (w/v) peptone, 2% (w/v) dextrose). Minimal plates (1.34% (w/v) YNB (Yeast Nitrogen Base with ammonium sulphate without amino acids), 4 × 10−5 % (w/v) biotin, 2% (w/v) dextrose) were used for His4 selection of transformants. To induce expression P. pastoris was grown in 500 mL BMSY media (1% yeast extract, 2% peptone, 100 mM potassium phosphate buffer, pH 6.0, 1.34% YNB, 4 × 10−5 % biotin, 1% sorbitol) at 28 ◦ C, 280 rpm. After 24 h growth, expression was induced by adding 50% methanol to a concentration of 0.5% (v/v). To maintain induction 50% methanol was added to 0.5% (v/v) every 24 h. Every 24 h 1 mL samples were taken, and the amount of secreted protein was determined by SDS-PAGE, using 12% Criterion XT Bis-Tris gels (BIO-RAD, USA). After 4–5 days cells were removed by centrifugation (5 min, 2000 × g, 4 ◦ C). Escherichia coli TOP 10 (Invitrogen, USA) cells were used for all plasmid constructions. 2.2. Cloning A new polylinker (5 -GGATCCATCCGCGGTAGGTCACCACTTAAGTACTAGTACGTAGAATTCCCGGGCGGCCGC3 ) was introduced into pPIC9K by replacing the BamHI–NotI fragment, including the alpha mating factor secretion signal, creating pPIC9Kp (Fig. 1). Wild-type BASI cDNA was received from Birte Svensson, DTU, Denmark. A fully codon-optimized CDR (arBASI) and a CDR where all arginine codons have been optimized (RBASI) were synthesized by PCR-based gene synthesis (Hoover and Lubkowski, 2002). All oligos were purchased from Invitrogen. BASI, arBASI and RBASI cDNA without signal sequence and the alpha mating factor secretion signal was amplified and combined by Splicing by Overlapping Extension PCR (SOE-PCR). All constructs were subcloned in the TOPO vector, using TOPO Blunt TOPO PCR Cloning Kit (Invitrogen, USA). 2.3. Preparation of competent cells A 10 mL YPD (1% (w/v) yeast extract, 2% (w/v) peptone, 2% (w/v) dextrose) culture was inoculated with a single P. pastoris GS115 colony and grown over night at 30 ◦ C, 180 rpm. A 500 mL YPD culture was inoculated to OD600 ∼ 0.01 and grown over night at 30 ◦ C, 180 rpm to OD600 ∼ 1.3–1.5. Cells were harvested by centrifugation (5 min, 2000 × g, 4 ◦ C). Cells were resuspended in 100 mL YPD medium with 0.2 M HEPES pH 8.0. When all cells were resuspended, 2.5 mL 1 M DTT was added. Cells were incubated at 30 ◦ C for 15 min without shaking. Four hundred-milliliter cold water was added, and cells were pelleted by centrifugation (5 min, 2000 × g, 4 ◦ C). Cells were washed in 250 mL cold water by centrifugation (5 min, 2000 × g, 4 ◦ C). Cells were washed in 20 mL cold 1 M sorbitol by centrifugation (5 min, 2000 × g, 4 ◦ C). All cells were resuspended between

426

P.O. Micheelsen et al. / Journal of Biotechnology 133 (2008) 424–432

Fig. 1. Schematic overview of the polylinker region in pPIC9Kp, modified from pPIC9K.

each washing step. After washing, all cells were resuspended in 0.5 mL 1 M sorbitol and 40 ␮L aliquots were prepared. The competent cells were either used right away, or stored at −80 ◦ C for later use.

of 400 ␮L for 10 min at 37 ◦ C in reaction buffer (40 mM Tris pH 8.0; 5 mM CaCl2 ; 0.05% BSA). Assay was performed as previously described (Bønsager et al., 2003). 2.7. Southern blotting

2.4. Transformation Prior to transformation all plasmids were linearized by BglII digestion, following gel purification using MinElute QIAGEN (Hilden, Germany). For all transformations, approximately 25 ng linearized DNA in 1 ␮L TE buffer and 40 ␮L competent cells were used. Transformations were done by electroporation using a Gene Pulser (BIO-RAD, USA), voltage: 1500 V, capacitance: 25 ␮F and resistance: 200 . Transformations were carried out in 2-mm electroporation cuvettes. Following electroporation, cells were suspended in 1 mL cold 1 M sorbitol and spread on YNB plates (Yeast Nitrogen Base with ammonium sulphate without amino acids), 4 × 10−5 % (w/v) biotin, 2% (w/v) dextrose) for His4 selection. Plates were incubated at 30 ◦ C for 2–4 days. 2.5. Screen for single-copy transformants Single colonies were picked from the initial YNB plates and resuspended in 10 ␮L water. Cells were lysed by microwaving for 1 min. PCR mastermix, containing Phusion DNA polymerase (Finnzymes, Finland), and the following primers—AOX1 for: 5 -GACTGGTTCCAATTGACAAGC-3 ; AOX1 rev: 5 GCAAATGGCATTCTGACATCC-3 ; pPIC9Kp junction for: 5 -CATCAGAGATTTTGAGACACAACG-3 and pPIC9Kp junction rev: 5 -GGCCGTTAGCATTTCAACGAACC-3 were added to the lysed cells, and 30 cycles of PCR were performed with annealing at 58 ◦ C. PCR products were analysed by agarose gel electrophoresis. 2.6. Insoluble blue starch assay AMY2 (3.75 nM) was preincubated with BASI at various concentrations (ranging from 0 to 50 nM) in a total volume

Total DNA was purified by using the MasterPureTM Yeast DNA Purification kit from Epicentre (Madison, USA), followed by phenol:chloroform extraction. The quality and concentration of the DNA was determined by agarose gel electrophoresis and spectrophotometry. Fifty-microgram genomic DNA was digested with EcoRI over night, and separated by electrophoresis on a 0.7% agarose gel. DNA was transferred to a Zeta-Probe GT blotting Membrane (BIO-RAD, USA) over night and the filter was hybridized following the manufacturer’s protocol. For probing, a randomly labelled 898 bp AOX1 promoter fragment from pPIC9Kp, amplified by PCR using the following primers—AOX for: 5 -GGTTGAATGAAACCTTTTTGCC-3 and AOX1 rev: 5 -CAACTAATTATTCGAAGGATCC-3 was used. Probe-labelling was performed using P32 -labelled ATP, following Zeta-Probe GT Blotting Membrane manufacturer’s protocol (BIO-RAD, USA). To develop, the filter was exposed to a phosphoimager screen over night. 2.8. Protein purification BASI was precipitated from the supernatant by adding solid ammonium sulphate ((NH4 )2 SO4 ) to 3.2 M final concentration, and the solution was stirred for 30 min. The precipitate was collected by centrifugation (20,000 × g, 20 min) and dissolved in a minimal volume of Buffer A (20 mM CH3 COOH/NaOH, pH 5.0). The dissolved BASI was transferred to Buffer A by gelfiltration (desalting) on a G25 sephadex column. The desalted BASI solution was slightly turbid, and this turbidity was removed by 0.45 ␮m filtration. The filtrate was applied to a SP-sepharose HighLoad column equilibrated in Buffer A. After washing the column with Buffer A, the column was eluted with a linear NaCl gradient (0–0.5 M) in Buffer A over 10 column volumes. Fractions from the column with BASI were pooled and diluted 10×

P.O. Micheelsen et al. / Journal of Biotechnology 133 (2008) 424–432

with deionized water to reduce the conductivity. The diluted BASI pool was applied to a SOURCE S column equilibrated in Buffer A. After washing the column with Buffer A, the column was eluted with a linear NaCl gradient (0–0.5 M) in Buffer A over 10 column volumes. BASI containing fractions from the column were analysed by SDS-PAGE and pure fractions were pooled as the product of the purification. G25 sephadex, SPsepharose HighLoad and SOURCE S are column materials from GE Healthcare.

427

tein shares 72% sequence identity with the wild-type gene (153 of 546 nucleotides have been altered), GC content was lowered from 69 to 44%. Another gene encoding BASI was designed by exchanging all arginine-encoding codons into AGA. Arginine codons were selected for modification, as P. pastoris very consequently uses AGA to encode arginine (Zhao et al., 2000). An alignment of the nucleotide sequences for BASI, arBASI and RBASI is seen in Fig. 2. 3.2. Cloning of BASI, arBASI and RBASI

3. Results 3.1. Design of codon-optimized genes To investigate whether alteration of the codon usage in the BASI CDR, could result in higher expression levels, two alternative BASI genes were designed: one fully codon optimized (arBASI) and one where only arginine codons were altered (RBASI). Monocot genes generally have GC content, about 60–70%, while the GC content of endogenous Pichia genes normally is below 50% (Montero et al., 1990). Furthermore, a decreasing GC gradient has been observed in genes from monocots (Wong et al., 2002). A new CDR encoding BASI was designed, by changing all codons into the most widely used codons in P. pastoris, using available codon usage tables (Zhao et al., 2000). The fully codon-optimized gene (arBASI) encoding the mature pro-

Both synthetic CDRs were synthesized by PCR-based gene synthesis, as described previously (Hoover and Lubkowski, 2002). All three CDRs were cloned into the P. pastoris expression vector pPIC9Kp (modified from pPIC9K, see Section 2). The original BASI signal peptide was replaced by a truncated version of the alpha mating factor secretion signal from S. cerevisiae, in order to ensure secretion and authenticity. Processing of the alpha mating factor secretion signal happens in three steps; initially, the pre-sequence is removed by a signal peptidase in the ER. Following, Kex2 specifically cleaves C-terminally to the Lys-Arg site, this is rapidly followed by cleavage of Glu-Ala repeats by the dipeptidyl endopeptidase Ste13 (Brake, 1990). In S. cerevisiae it has been noted that Glu-Ala repeats are not necessary for Kex2 cleavage, also in some cases Ste13 cleavage of Glu-Ala repeats is not sufficient, and Glu-Ala repeats are therefore left on the N-terminal of the expressed protein (Almeida et

Fig. 2. Nucleotide alignment of the mature parts of BASI, arBASI and RBASI. Altered nucleotides are marked with gray background.

428

P.O. Micheelsen et al. / Journal of Biotechnology 133 (2008) 424–432

Fig. 3. Schematic overview of the truncated alpha mating factor secretion signal sequence used. Five residues have been removed in order to remove the two Ste13 sites (marked with arrows) and thereby ensure a clean N-terminal on the secreted protein.

al., 2001; Brake et al., 1984). To avoid excessive residues at the N-terminal of heterologous expressed BASI, a truncated version of the alpha mating factor secretion signal missing the Glu-Ala repeats, was used in this study (Fig. 3) (Cabral et al., 2003). To enhance the initial binding of the small subunit of the ribosome to the mRNA template, the consensus Kozak sequence (Kozak, 1987) (ACC) was inserted just upstream of the start codon in all constructs. In addition, all constructs mentioned above were prepared with the native BASI signal, and expression from transformants harbouring these was analysed. 3.3. Screening for single-copy transformants To be able to compare expression levels from the generated constructs, it is essential that all transformants are generated in the same way, inserts are integrated in the same position and with the same frequency. In this study single insertion transformants, generated by gene replacement of the AOX1 gene, was preferred, as this type of transformants is relatively easily screened for, and multiple insertion events rarely take place. Therefore, all constructs were digested with BglII prior to transformation, which results in a fragment, with ends homologous to the 5 and 3 ends of the AOX1, thereby targeting replacement of the AOX1 gene by double homologous recombination. For selection of clones, where replacement of the AOX1 gene had taken place, a colonyPCR screen was developed using two primer sets. One primer set was specific for the AOX1 gene, and one primer set was vector specific. The AOX1 primers were designed to give a 2.2 kb band if the AOX1 gene was present, and a 1 kb band if the AOX1 gene had been replaced. The vector specific primers were designed to amplify the vector junctions created if multiple insertions had occurred, indicated by a 1.5 kb band (Fig. 4). For each transformation 100 clones were screened, approximately 25% of these were single insertion transformants, generated by replacement of the AOX gene (Table 1). All single-copy transformants carrying either of the BASI CDRs, with the alpha mating factor secretion signal, tested showed equal expression levels, while transformants carrying either of the BASI CDRs with the native signal, expressed no protein (data not shown). One transformant of each type carrying the alpha mating factor secretion signal was selected for further analysis. To verify the exact integration site, a Southern blot analysis was performed, using a randomly labelled AOX1 promoter fragment as probe. As illustrated in Fig. 5, the selected clones were all single insertion transformants, generated by gene replacement of the AOX1 gene, thus

making comparison of the expression levels from the various CDRs, in these particular transformants, possible. 3.4. Expression, purification and characterization of BASI The selected single-copy AOX1 gene replacement transformants were all grown in 3 mL BMSY media at 28 ◦ C. Maximum expression levels were reached after 4 days for all constructs (Fig. 6A). If cultures were grown for more than 4 days, cell lysis was observed. Expression levels from the wild-type CDR reached 125 mg L−1 while the fully codonoptimized CDR and the arginine-optimized CDR produced 65 and 125 mg L−1 , respectively (Fig. 6B). In addition, expression

Fig. 4. PCR screen of selected clones. Lane 1 is BASI, lane 2 is arBASI, lane 3 is RBASI, lane 4 is empty pPIC9Kp and lane 5 is a control using a multiple insertion transformant not generated by replacement of the AOX1 gene. ␭ DNA-BstEII digest was used as marker.

Table 1 Overview of integration events

BASI arBASI RBASI pPIC9Kp empty

Single

Single GRa

Multiple

Multiple GRa

ND

53 64 59 57

26 25 24 18

1 1 3 1

0 0 0 0

20 10 14 –

Number of different integration events. One hundred clones transformed with alpha mating factor secretion signal-BASI, -arBASI or -RBASI constructs, and 76 clones transformed with the empty expression vector, were screened with PCR to find single transformants generated by gene replacement of the AOX1 gene. PCR reactions that were unsuccessful are categorized as ND. a GR: gene replacement of the AOX1 gene.

P.O. Micheelsen et al. / Journal of Biotechnology 133 (2008) 424–432

429

Fig. 5. (A) Schematic overview of the predicted integration event via double homologous recombination. The labelled AOX1 promotor fragment used as a probe is indicated as a line. Sizes of predicted EcoRI fragments appear in kilobases (kb). (B) Southern blot analysis on genomic DNA from different Pichia pastoris clones. Fifty-microgram chromosomal DNA was digested with EcoRI and filter hybridized with an 898 bp AOX1 promotor fragment probe. The 1.9 and 2.7 kb bands correspond to the AOX1 promotor fragment carried on the vector integrated and the 5.7 kb band to the chromosomal AOX1 promotor. Lane 1 is BASI, lane 2 is arBASI, lane 3 is RBASI, lane 4 pPIC9Kp empty vector and lane 5 is GS115. Labelled ␭ DNA-BstEII digest was used as marker.

Fig. 6. (A) Graph showing expression levels following methanol induction. (B) Coomassie blue stained SDS-polyacrylamide gel of 16 ␮L culture supernatant. Lane 1 is BASI, lane 2 is arBASI, lane 3 is RBASI and lane 4 is pPIC9Kp empty vector. Marker is MultiMark® Multi-colored Standard from Invitrogen.

levels were analysed following induction at 23 and 30 ◦ C. The amount of protein produced at 23 and 30 ◦ C was about half compared to 28 ◦ C (data not shown). In order to characterize the recombinant protein, the selected RBASI clone was grown in larger scale (8× 0.5 L cultures grown in 2 L shake flasks), and the expressed protein was purified (see Section 2). The purified protein was sequenced by Edman degradation (Edman, 1970). The N-terminal sequence was ADPPPVHDTD, which is identical to the mature protein from barley (Mundy et al., 1984). In order to analyse the recombinant BASI protein for the presence of post-translational modifications, liquid chromatography mass spectrometry (Domon and Aebersold, 2006; Mitulovic and Mechtler, 2006) was performed, using a single quadruple ESI instrument coupled to a HPLC instrument. As expected, one major peak was observed at 19875.05 Da (theoretical Mw = 19875.09 Da). Activity of the recombinant protein was assessed using the insoluble blue starch assay (Rodenburg et al., 1995). The assay was performed with BASI purified from P. pastoris and recombinant His6 -tagged BASI (rBASI) puri-

fied from E. coli (kindly provided by Dr. Birgit Bønsager, DTU, Denmark), which has previously been shown to posses the same properties as BASI purified from barley grain (Bønsager et al., 2003). BASI purified from P. pastoris and from E. coli showed identical AMY2 inhibition properties (Fig. 7). 4. Discussion During the past two decades P. pastoris has become a widely used heterologous expression system. Transformants can be generated in two different ways. Homologous recombination, resulting in a single crossover event, in either the His4 gene or the AOX1 gene, with rather high transformation efficiency (50–80%) or double homologous recombination, resulting in replacement of the AOX1 gene. The latter results in transformants with the methanol utilization slow (Muts ) phenotype, efficiency is about 10–25%. Remaining transformants have undergone gene conversion, where only the His4 gene has been integrated (Cereghino and Cregg, 2000; Cregg et al., 1993; Daly

430

P.O. Micheelsen et al. / Journal of Biotechnology 133 (2008) 424–432

Fig. 7. Inhibition of AMY2 by BASI expressed in P. pastoris and His6 -tagged rBASI expressed in E. coli. Assay was carried out at pH 8.0, all experiments were repeated four times.

and Hearn, 2005). In this study all constructs were digested with BglII prior to transformation, which should result in gene replacement of the AOX1 gene by double homologous recombination. In theory this should lead to 10–25% Muts transformants. However, all clones derived from these transformations showed the Muts phenotype, when grown on minimal plates with and without methanol, even though in about 50% of the clones, the AOX1 gene could be amplified with PCR. In the majority of these clones, the presence of the expression cassette could also be verified by PCR. This indicates that the AOX1 gene has been disrupted in all clones, but replaced in only about 25% of the clones. In vivo ligation of the vector has previously been suggested to explain unexpected integration events (Clare et al., 1991). However, the phenotypes seen here cannot be explained by in vivo ligation. Clones that did not secrete protein, but showed the Muts phenotype, could arise from conversion events, where HIS4 is integrated into the 3 end of the AOX1 gene by a single crossover event, without integration of the expression cassette. Furthermore it was very consistently observed that only transformants generated by gene replacement expressed the protein. The fact that approximately 25% of the transformants are generated by gene replacement is consistent with the literature (Cereghino and Cregg, 2000). BASI has been successfully expressed in E. coli with various tags, where the highest yield obtained was with a His6 -tag, resulting in 25 mg L−1 , an untagged version has been expressed at 3 mg L−1 (Bønsager et al., 2003). The yield of secreted protein produced in this study was estimated from SDS-PAGE to be in the range of 125 mg L−1 for both the wild-type CDR and the CDR with altered arginine codons. Apart from the level of expression, one of the advantages of the Pichia-based system is secretion of the mature protein. Pichia secretes very few endogenous proteins (Fig. 6B), purification can therefore be performed in fewer steps, and thereby less protein is lost during purification, compared to E. coli expression. Remarkably, the yield of the fully codon-optimized gene was estimated to be about half compared to wild-type CDS

and RBASI CDS. arBASI was designed by altering all codons into the most highly used codon in P. pastoris coding sequences (Zhao et al., 2000). This design strategy results in a very narrow codon usage compared to the wild-type gene. tRNA abundance varies for different species, and it has previously been shown that highly expressed genes tend to use a narrow set of codons, corresponding to the tRNA abundance in order to avoid tRNA depletion (Ikemura, 1981a,b, 1985). Even though codon optimization have been shown useful to optimize expression levels in a broad variety of organisms, including both prokaryotes and eukaryotes (Chang et al., 2006; Hu et al., 1996; Park et al., 2002; Sinclair and Choy, 2002; Woo et al., 2002), consistently using only one codon per amino acid, might make tRNA concentration the limiting factor for expression. As numerous bottlenecks in heterologous expression besides tRNA depletion is know, there may be an array of other factors influencing the expression level from arBASI (Wu et al., 2004). When BASI and arBASI mRNA is submitted to the mRNA secondary structure prediction servers (Reeder et al., 2006) mFOLD (Zuker and Stiegler, 1981) and RNAfold (Hofacker, 2003), it is seen that the free energy of the BASI mRNA is higher than that of arBASI. This might indicate that the arBASI mRNA is less stable, thus eventually leading to inferior BASI expression (Garneau et al., 2007). The fact that the fully codon-optimized CDR results in lower expression levels than the wild-type and that only constructs with the alpha mating factor secretion signal results in secreted protein, shows that in this study, the choice of signal sequence has greater impact on the expression levels than the choice of codon usage. In this study, three genes using different codons, but all encoding the same ␣-amylase/subtilisin inhibitor from barley, have been heterologously expressed in P. pastoris resulting in 65–125 mg L−1 native protein. The expression levels achieved in the current study does not however allow for bulk applications. Heterologous expression yields may be further improved by expression in filamentous fungi. Further, in order for BASI to be of both pharmaceutical and industrial interest as a protease inhibitor, more information is needed about the function. So far, the structure of BASI has been solved in complex with AMY2 from barley (Vallee et al., 1998) and the Protein Disulphide Reductase Thioredoxin also from barley (Maeda et al., 2006). The structure of BASI in complex with a subtilisin can be modelled from the known structure of the highly homologous wheat ␣-amylase inhibitor in complex with proteinase K (Pal et al., 1994). But in order to do specific engineering on BASI, a structure of BASI in complex with a subtilisin would be preferred. Acknowledgements We are grateful to Professor Birte Svensson and Dr. Birgit Bønsager, BioCentrum, Technical University of Denmark for kindly donating BASI cDNA and purified rBASI and AMY2 protein and for help with the insoluble blue starch assay. The work was financially supported by the STF programme of the Danish Research Agency, Ministry of Science, Technology and Innovation.

P.O. Micheelsen et al. / Journal of Biotechnology 133 (2008) 424–432

References Almeida, M.S., Cabral, K.S., De Medeiros, L.N., Valente, A.P., Almeida, F.C.L., Kurtenbach, E., 2001. cDNA cloning and heterologous expression of functional cysteine-rich antifungal protein Psd1 in the yeast Pichia pastoris. Arch. Biochem. Biophys. 395, 199–207. Bellincampi, D., Camardella, L., Delcour, J.A., Desseaux, V., D’Ovidio, R., Durand, A., Elliot, G., Gebruers, K., Giovane, A., Juge, N., Sorensen, J.F., Svensson, B., Vairo, D., 2004. Potential physiological role of plant glycosidase inhibitors. Biochim. Biophys. Acta 1696, 265–274. Bønsager, B.C., Praetorius-Ibba, M., Nielsen, P.K., Svensson, B., 2003. Purification and characterization of the beta-trefoil fold protein barley alphaamylase/subtilisin inhibitor overexpressed in Escherichia coli. Protein Expr. Purif. 30, 185–193. Brake, A.J., Merryweather, J.P., Coit, D.G., Heberlein, U.A., Masiarz, F.R., Mullenbach, G.T., Urdea, M.S., Valenzuela, P., Barr, P.J., 1984. Alpha-factordirected synthesis and secretion of mature foreign proteins in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. U.S.A. 81, 4642–4646. Brake, A.J., 1990. Alpha-factor leader-directed secretion of heterologous proteins from yeast. Methods Enzymol. 185, 408–421. Cabral, K.M., Almeida, M.S., Valente, A.P., Almeida, F.C., Kurtenbach, E., 2003. Production of the active antifungal Pisum sativum defensin 1 (Psd1) in Pichia pastoris: overcoming the inefficiency of the STE13 protease. Protein Expr. Purif. 31, 115–122. Cereghino, J.L., Cregg, J.M., 2000. Heterologous protein expression in the methylotrophic yeast Pichia pastoris. FEMS Microbiol. Rev. 24, 45–66. Chang, S.W., Lee, G.C., Shaw, J.F., 2006. Codon optimization of Candida rugosa lip1 gene for improving expression in Pichia pastoris and biochemical characterization of the purified recombinant LIP1 lipase. J. Agric. Food Chem. 54, 815–822. Clare, J.J., Rayment, F.B., Ballantine, S.P., Sreekrishna, K., Romanos, M.A., 1991. High-level expression of tetanus toxin fragment C in Pichia pastoris strains containing multiple tandem integrations of the gene. Bio/technology (Nature Publishing Company) 9, 455–460. Cregg, J.M., Vedvick, T.S., Raschke, W.C., 1993. Recent advances in the expression of foreign genes in Pichia pastoris. Bio/technology (Nature Publishing Company) 11, 905–910. Daly, R., Hearn, M.T., 2005. Expression of heterologous proteins in Pichia pastoris: a useful experimental tool in protein engineering and production. J. Mol. Recognit. 18, 119–138. Domon, B., Aebersold, R., 2006. Mass spectrometry and protein analysis. Science 312, 212–217. Edman, P., 1970. Sequence determination. Mol. Biol. Biochem. Biophys. 8, 211–255. Faus, I., 2000. Recent developments in the characterization and biotechnological production of sweet-tasting proteins. Appl. Microbiol. Biotechnol. 53, 145–151. Fierens, K., Geudens, N., Brijs, K., Courtin, C.M., Gebruers, K., Robben, J., Van Campenhout, S., Volckaert, G., Delcour, J.A., 2004. High-level expression, purification, and characterization of recombinant wheat xylanase inhibitor TAXI-I secreted by the yeast Pichia pastoris. Protein Expr. Purif. 37, 39–46. Garneau, N.L., Wilusz, J., Wilusz, C.J., 2007. The highways and byways of mRNA decay. Nat. Rev. Mol. Cell. Biol. 8, 113–126. Gouka, R.J., Punt, P.J., van den Hondel, C.A.M.J., 1997. Efficient production of secreted proteins by Aspergillus?: progress, limitations and prospects. Appl. Microbiol. Biotechnol. 47, 1–11. Grantham, R., Gautier, C., Gouy, M., 1980. Codon frequencies in 119 individual genes confirm consistent choices of degenerate bases according to genome type. Nucleic Acids Res. 8, 1893–1912. Hofacker, I.L., 2003. Vienna RNA secondary structure server. Nucleic Acids Res. 31, 3429–3431. Hoover, D.M., Lubkowski, J., 2002. DNAWorks: an automated method for designing oligonucleotides for PCR-based gene synthesis. Nucleic Acids Res. 30, e43. Hu, X., Shi, Q., Yang, T., Jackowski, G., 1996. Specific replacement of consecutive AGG codons results in high-level expression of human cardiac troponin T in Escherichia coli. Protein Expr. Purif. 7, 289–293.

431

Ikemura, T., 1981a. Correlation between the abundance of Escherichia coli transfer RNAs and the occurrence of the respective codons in its protein genes: a proposal for a synonymous codon choice that is optimal for the E. coli translational system. J. Mol. Biol. 151, 389–409. Ikemura, T., 1981b. Correlation between the abundance of Escherichia coli transfer RNAs and the occurrence of the respective codons in its protein genes. J. Mol. Biol. 146, 1–21. Ikemura, T., 1982. Correlation between the abundance of yeast transfer RNAs and the occurrence of the respective codons in protein genes: differences in synonymous codon choice patterns of yeast and Escherichia coli with reference to the abundance of isoaccepting transfer RNAs. J. Mol. Biol. 158, 573–597. Ikemura, T., 1985. Codon usage and tRNA content in unicellular and multicellular organisms. Mol. Biol. Evol. 2, 13–34. Kong, J.M., Goh, N.K., Chia, L.S., Chia, T.F., 2003. Recent advances in traditional plant drugs and orchids. Acta Pharmacol. Sin. 24, 7–21. Kozak, M., 1987. An analysis of 5 -noncoding sequences from 699 vertebrate messenger RNAs. Nucleic Acids Res. 15, 8125–8148. Leah, R., Mundy, J., 1989. The bifunctional alpha-amylase/subtilisin inhibitor of barley: nucleotide and patterns of seed-specific expression. Plant Mol. Biol. 12, 673–682. Macauley-Patrick, S., Fazenda, M.L., McNeil, B., Harvey, L.M., 2005. Heterologous protein production using the Pichia pastoris expression system. Yeast 22, 249–270. Maeda, K., Hagglund, P., Finnie, C., Svensson, B., Henriksen, A., 2006. Structural basis for target protein recognition by the protein disulfide reductase thioredoxin. Structure 14, 1701–1710. Mitulovic, G., Mechtler, K., 2006. HPLC techniques for proteomics analysis—a short overview of latest developments. Brief. Funct. Genomics Proteomics 5, 249–260. Mohnen, D., 1999. Biosynthesis of pectins and galactomannans. In: Pinto, B.M. (Ed.), Carbohydrates and their Derivatives Including Tannins, Cellulose, and Related Lignins. Elsevier, pp. 497–527. Montero, L.M., Salinas, J., Matassi, G., Bernardi, G., 1990. Gene distribution and isochore organization in the nuclear genome of plants. Nucleic Acids Res. 18, 1859–1867. Morant, M., Bak, S., Møller, B.L., Werck-Reichhart, D., 2003. Plant cytochromes P450: tools for pharmacology, plant protection and phytoremediation. Curr. Opin. Biotechnol. 14, 151–162. Mundy, J., Svendsen, I., Hejgaard, J., 1983. Barley alpha-amylase/subtilisin inhibitor. Isolation and characterization. Carlsberg Res. Commun. 48, 81–90. Mundy, J., Hejgaard, J., Svendsen, I., 1984. Characterization of a bifunctional wheat inhibitor of endogenous alpha-amylase and subtilisin. FEBS Lett. 167, 210–214. Ohtsubo, K.I., Richardson, M., 1992. The amino acid sequence of a 20 kDa bifunctional subtilisin/alpha-amylase inhibitor from brain of rice (Oryza sativa L.) seeds. FEBS Lett. 309, 68–72. Pal, G.P., Kavounis, C.A., Jany, K.D., Tsernoglou, D., 1994. The threedimensional structure of the complex of proteinase K with its naturally occurring protein inhibitor, PKI3. FEBS Lett. 341, 167–170. Pan, H.Y., Whittaker, M.M., Bouveret, R., Berna, A., Bernier, F., Whittaker, J.W., 2007. Characterization of wheat germin (oxalate oxidase) expressed by Pichia pastoris. Biochem. Biophys. Res. Commun. 356, 925–929. Park, S.J., Lee, S.K., Lee, B.J., 2002. Effect of tandem rare codon substitution and vector–host combinations on the expression of the EBV gp110 C-terminal domain in Escherichia coli. Protein Expr. Purif. 24, 470–480. Patick, A.K., Potts, K.E., 1998. Protease inhibitors as antiviral agents. Clin. Microbiol. Rev. 11, 614–627. Pekkarinen, A.I., Jones, B.L., 2003. Purification and identification of barley (Hordeum vulgare L.) proteins that inhibit the alkaline serine proteinases of Fusarium culmorum. J. Agric. Food Chem. 51, 1710–1717. Reeder, J., Hochsmann, M., Rehmsmeier, M., Voss, B., Giegerich, R., 2006. Beyond Mfold: recent advances in RNA bioinformatics. J. Biotechnol. 124, 41–55. Rodenburg, K.W., Varallyay, E., Svendsen, I., Svensson, B., 1995. Arg-27, Arg127 and Arg-155 in the beta-trefoil protein barley alpha-amylase/subtilisin inhibitor are interface residues in the complex with barley alpha-amylase 2. Biochem. J. 309, 969–976.

432

P.O. Micheelsen et al. / Journal of Biotechnology 133 (2008) 424–432

Sharp, P.M., Tuohy, T.M., Mosurski, K.R., 1986. Codon usage in yeast: cluster analysis clearly differentiates highly and lowly expressed genes. Nucleic Acids Res. 14, 5125–5143. Sinclair, G., Choy, F.Y., 2002. Synonymous codon usage bias and the expression of human glucocerebrosidase in the methylotrophic yeast Pichia pastoris. Protein Expr. Purif. 26, 96–9105. Sørensen, J.F., Kragh, K.M., Sibbesen, O., Delcour, J., Goesaert, H., Svensson, B., Tahir, T.A., Brufau, J., Perez-Vendrell, A.M., Bellincampi, D., D’Ovidio, R., Camardella, L., Giovane, A., Bonnin, E., Juge, N., 2004. Potential role of glycosidase inhibitors in industrial biotechnological applications. Biochim. Biophys. Acta 1696, 275–287. Svendsen, I., Hejgaard, J., Mundy, J., 1986. Complete amino acid sequence of the alpha-amylase/subtilisin inhibitor from barley. Carlsberg Res. Commun. 51, 43–50. Vallee, F., Kadziola, A., Bourne, Y., Juy, M., Rodenburg, K.W., Svensson, B., Haser, R., 1998. Barley alpha-amylase bound to its endogenous protein ˚ resolution. Structure inhibitor BASI: crystal structure of the complex at 1.9 A 6, 649–659.

Wong, G.K.-S., Wang, J., Tao, L., Tan, J., Zhang, J., Passey, D.A., Yu, J., 2002. Compositional gradients in Gramineae genes. Genome Res. 12, 851–856. Woo, J.H., Liu, Y.Y., Mathias, A., Stavrou, S., Wang, Z., Thompson, J., Neville, D.M., 2002. Gene optimization is necessary to express a bivalent anti-human anti-T cell immunotoxin in Pichia pastoris. Protein Expr. Purif. 25, 270–282. Wu, X., Jornvall, H., Berndt, K.D., Oppermann, U., 2004. Codon optimization reveals critical factors for high level expression of two rare codon genes in Escherichia coli: RNA stability and secondary structure but not tRNA abundance. Biochem. Biophys. Res. Commun. 313, 89–96. Yoneyama, T., Taira, M., Suzuki, T., Nakamura, M., Niwa, K., Watanabe, T., Ohyama, T., 2007. Expression and characterization of a recombinant unique acid phosphatase from kidney bean hypocotyl exhibiting chloroperoxidase activity in the yeast Pichia pastoris. Protein Expr. Purif. 53, 31–39. Zhao, X., Huo, K.K., Li, Y.Y., 2000. Synonymous codon usage in Pichia pastoris. Sheng Wu Gong Cheng Xue Bao 16, 308–311. Zuker, M., Stiegler, P., 1981. Optimal computer folding of large RNA sequences using thermodynamics and auxiliary information. Nucleic Acids Res. 9, 133–148.