Appl Microbiol Biotechnol (2014) 98:2573–2583 DOI 10.1007/s00253-013-5468-7

APPLIED GENETICS AND MOLECULAR BIOTECHNOLOGY

Elimination of diaminopeptidase activity in Pichia pastoris for therapeutic protein production Daniel Hopkins & Sujatha Gomathinayagam & Heather Lynaugh & Terrance A. Stadheim & Stephen R. Hamilton

Received: 7 November 2013 / Revised: 9 December 2013 / Accepted: 10 December 2013 / Published online: 14 February 2014 # Springer-Verlag Berlin Heidelberg 2014

Abstract Yeast are important production platforms for the generation of recombinant proteins. Nonetheless, their use has been restricted in the production of therapeutic proteins due to differences in their glycosylation profile with that of higher eukaryotes. The yeast strain Pichia pastoris is an industrially important organism. Recent advances in the glycoengineering of this strain offer the potential to produce therapeutic glycoproteins with sialylated human-like N- and O-linked glycans. However, like higher eukaryotes, yeast also express numerous proteases, many of which are either localized to the secretory pathway or pass through it en route to their final destination. As a consequence, nondesirable proteolysis of some recombinant proteins may occur, with the specific cleavage being dependent on the class of protease involved. Dipeptidyl aminopeptidases (DPP) are a class of proteolytic enzymes which remove a two-amino acid peptide from the N-terminus of a protein. In P. pastoris, two such enzymes have been identified, Ste13p and Dap2p. In the current report, we demonstrate that while the knockout of STE13 alone may protect certain proteins from N-terminal clipping, other proteins may require the double knockout of both STE13 and DAP2. As such, this understanding of DPP activity enhances the utility of the P. pastoris expression system, thus facilitating the production of recombinant therapeutic proteins with their intact native sequences.

Electronic supplementary material The online version of this article (doi: 10.1007/s00253-013-5468-7 ) contains supplementary material, which is available to authorized users. D. Hopkins : S. Gomathinayagam : H. Lynaugh : T. A. Stadheim : S. R. Hamilton (*) GlycoFi, Inc. (a wholly owned subsidiary of Merck & Co., Inc.), Biologics Discovery, Merck Research Laboratories, 16 Cavendish Court, Lebanon, NH 03766, USA e-mail: [email protected]

Keywords Diaminopeptidase . STE13 . DAP2 . P. pastoris . Proteolysis . Recombinant protein

Introduction One of the fastest growing classes of therapeutic compounds in the pharmaceutical industry is that of therapeutic proteins (Walsh 2010). To date, a number of protein expression systems have been developed, including bacterial, yeast, plant, insect, and mammalian cell lines (Demain and Vaishnav 2009). Since many therapeutic proteins need to undergo posttranslational modifications (such as glycosylation), mammalian cell lines have become the predominant expression system for such proteins, with Chinese hamster ovary (CHO) cell lines being the most widely utilized in the industrial setting (Zhu 2012). However, there are several limitations with using mammalian cell culture to generate therapeutic proteins. These include high production cost, risk of viral contaminants, and heterogeneous glycosylation of proteins. Furthermore, the use of nonhuman-derived cell lines can lead to the introduction of nonhuman glycosylation modifications, including the addition of α-1,3-galactose or N-glycolylneuraminic acid (Hokke et al. 1990). Such modifications can reduce the therapeutic efficiency of the molecule by reducing key attributes such as half-life (Ghaderi et al. 2010). Furthermore, these nonhuman glycoforms have been shown to elicit an immune response in humans (Higashi et al. 1977; Chung et al. 2008). As such, these drawbacks have led researchers to develop alternative expression systems for posttranslationally modified proteins. Yeast and other filamentous fungi have emerged as alternative expression systems to mammalian cell culture. The benefits of yeast cell lines as expression systems for the production of therapeutic proteins include the ability of growth in defined media, ease of scale-up, and high yields

2574

of secreted proteins (Punt et al. 2002). The methylotrophic yeast, Pichia pastoris, is an important organism used for the production of recombinant proteins. Over the last decade, P. pastoris has been extensively engineered to produce humanized N- and O-linked glycosylation profiles, thus making glycoengineered P. pastoris a more suitable host for the production of human glycoproteins (Choi et al. 2003; Hamilton et al. 2003, 2006, 2013; Bobrowicz et al. 2004; Hopkins et al. 2011). Similar to higher eukaryotes, P. pastoris produces a number of proteases. Proteases play important roles in protein maturation and degradation of undesirable protein products, but in an expression system, protease activity may lead to undesirable proteolysis of recombinant proteins. The production of a full-length protein is of importance since proteolysis of a recombinant protein can alter its stability and/or function. For example, removal of three or more N-terminal amino acids from the peptide hormone extendin-4 results in an antagonistic function, as opposed to the agonistic behavior of the full-length peptide (Montrose-Rafizadeh et al. 1997). Dipeptidyl aminopeptidases (DPP), also referred to as diaminopeptidases, are a class of proteases that cleave two amino acids from the N-terminus of a protein. These enzymes have been identified in a number of eukaryotic organisms, including yeast (Hong et al. 1989; Misumi et al. 1992; Matoba et al. 1997; Kumagai et al. 2003). In Saccharomyces cerevisiae, two gene products, Ste13p and Dap2p, have been identified as having DPP activity (Julius et al. 1983; SuarezRendueles and Wolf 1987). Mutations in S. cerevisiae STE13 produce incompletely processed alpha-mating factor, resulting in yeast strain sterility (Julius et al. 1983). The second DPP, Dap2p, was identified in a screen of S. cerevisiae mutants deficient in Ste13p activity (Suarez-Rendueles and Wolf 1987). Ste13p is a type II membrane serine protease anchored in the Golgi, while Dap2p is localized to the vacuole (Roberts et al. 1992). The preferred substrates for Ste13p and Dap2p have been shown to be the N-terminal motif of X-P/A, where X can be any amino acid and the second position is either a proline or alanine (Julius et al. 1983). The homologs of S. cerevisiae Ste13p and Dap2p have been identified and knocked out in P. pastoris (Prabha et al. 2009). In this study, Prabha et al. disrupted the STE13 gene in P. pastoris and found that this prevented N-terminal cleavage of an insulinotropic peptide which possessed a novel DPP substrate motif (histidine-glycine) at the N-terminus (Prabha et al. 2009). By contrast, disruption of the DAP2 gene did not prevent N-terminal clipping (Prabha et al. 2009). In a recent study, our group observed N-terminal clipping of a recombinant fusion protein composed of the tumor necrosis factor 2 ectodomain fused to the crystallizable fragment of human IgG1 (TNFR2:Fc), when expressed in glycoengineered P. pastoris possessing functional Ste13p and Dap2p (Hamilton et al. 2013). The TNFR2:Fc molecule possessed a

Appl Microbiol Biotechnol (2014) 98:2573–2583

leucine-proline motif at its N-terminus, of which about 90 % was cleaved. Unlike Prabha et al. (2009), we found that genetic disruption of STE13 alone did not completely protect the N-terminus of TNFR2:Fc in this glycoengineered background. In the current report, we investigate the N-terminal processing of TNFR2:Fc in a glycoengineered P. pastoris background through the knockout of both the STE13 and DAP2 genes. Furthermore, we investigate if the outcome was an artifact of expressing the therapeutic protein in a highly engineered P. pastoris strain. Finally, we go on to characterize our observations in both STE13 and DAP2 knockout backgrounds using a number of other reporter proteins which are processed by the N-terminal removal of two amino acid residues.

Methods and materials Strains, culture conditions, and reagents Escherichia coli strains TOP10 or XL10-Gold were used for recombinant DNA work. Restriction and modification enzymes were obtained from New England BioLabs (Ipswich, MA) and used as directed by the manufacturer. Oligonucleotides were obtained from Integrated DNA Technologies (Coralville, IA). Salts and buffering agents were from Sigma (St. Louis, MO). The P. pastoris yeast strain NRRL-Y11430 was obtained from ATCC (Manassas, VA). The glycoengineered P. pastoris yeast strain YGLY7406 was generated at GlycoFi and will be described in more detail later in the section entitled “Generation of STE13 and DAP2 knockout strains.” Minimal medium is 1.4 % yeast nitrogen base, 2 % dextrose, 1.5 % agar, and 4×10−5 % biotin and amino acids supplemented as appropriate. YSD-rich media is 1 % yeast extract, 2 % soytone, and 2 % dextrose, with the addition of 1.5 % agar for YSD plates. BSGY consists of the following: 40 g/L glycerol, 20 g/L soytone, 10 g/L yeast extract, 11.9 g/L KH2PO4, 2.3 g/L K2HPO4, 18.2 g/L sorbitol, 13.4 g/L YNB with ammonium sulfate without amino acids, and 8 mg/L biotin. Generation of knockout vectors Generation of STE13::Ura5 knockout vector DNA fragments corresponding to 5′ and 3′ flanking regions of the STE13 open reading frame were amplified using PfuUltra™ DNA polymerase (Stratagene, La Jolla, CA) and genomic DNA from the P. pastoris strain NRRL-Y11430 as template. The primer pairs SH774+SH775 (see Table 1 for all primer sequences) and SH776+SH777 were used to amplify 771 and 949 bp fragments for STE13 5′ and 3′, respectively. Following incubation with ExTaq™ (TaKaRa, Bio. Inc., Japan) for 10 min at 72 °C, the amplified fragments were cloned into pCR2.1 (Invitrogen,

Appl Microbiol Biotechnol (2014) 98:2573–2583

2575

Table 1 Primers used in this study Primer

Sequence (5′ to 3′)

Description

SH85 SH86 SH379 SH380 SH491 SH558 SH774 SH775 SH776 SH777 SH778 SH779 SH780 SH781 SH782 SH783 SH784

GGCTCGAGGATCTGTTTAGCTTGCCTCGTCC GGCTCGAGGGAGCTCGTTTTCGACACTGGATGG CATGCCCCTGAGCTGCGCACGTCAAG CAGAAAGTAATATCATGCGTCAATCG GGCGATTACCGTTGATGTTGAAGTGGCGAG CATCCAGAGGCACTTCACCGCTTGCCAGCG GGAATTCGGCCTTGGGGGCCTCCAGGACTTGCTG GGAATTCCTCGAGCTGTTTGAATCTGGAACGTACTCG GAAGCTTCTCGAGCTACTGGGAACCACGAGACATCAC GCAAGCTTGGCCCATTAGGCCCACCTACAATCATTACC CAAGGCACATTAAAAGTCCGCCAAAGG GTGGCCCTTGTATTGATAGAAGTATTCAG CACGTCTATCGTTGAACCAAAACAGAC GTAACCAATGGTATCTCCAACGACAG GGAATTCGGCCACCTGGGCCTGTTGCTGCTGGTACTG CGAATTCCTCGAGCGTTGTAAGTGATTGTAGACTCG GAAGCTTCTCGAGGGCAGCAAAGCCTTACGTTG

Nat cass XhoI for Nat cass XhoI rev pTEF (Nat) outwards TEF tt (Nat) outwards LacZ 5′–3′ screen out LacZ 3′–5′ screen out PpSTE13 5′ EcoRI for PpSTE13 5′ EcoRI rev PpSTE13 3′ HindIII for PpSTE13 3′ HindIII rev PpSTE13 pre 5′ PpSTE13 post 3′ PpSTE13 KO for PpSTE13 KO rev PpDAP2 5′ EcoRI for PpDAP2 5′ EcoRI rev PpDAP2 3′ HindIII for

SH785 SH786 SH787 SH788 SH789 SH805

GCAAGCTTGGCCTAGGTGGCCGACCCATTTTTAGAGG CACTTTCATCCTGAGGATCTTGGTCCTG CATATACCAAAGCAATTGATATCTGGTC CGGATAAGAGACATAATTGGCGCCATTC CTTTCTATTGAGGATTTCTTGGTTGCTG CGCCATCCAGTGTCGAAAACGCGTTGTAAGTGATTGTAGACTCGTTG

PpDAP2 3′ HindIII rev PpDAP2 pre 5′ PpDAP2 post 3′ PpDAP2 KO for PpDAP2 KO rev DAP2 5′ (Nat) rev

SH806 SH807 SH808

CAACGAGTCTACAATCACTTACAACGCGTTTTCGACACTGGATGGCG CAACGTAAGGCTTTGCTGCCTGTTTAGCTTGCCTCGTCCCCG CGGGGACGAGGCAAGCTAAACAGGCAGCAAAGCCTTACGTTG

Nat (DAP2 5′) for Nat (DAP2 3′) rev DAP2 3′ (Nat) for

Carlsbad, CA) and transformed into TOP10 competent cells. DNA sequencing confirmed that the STE13 5′ and 3′ flanking regions were correct and the resultant vectors designated as pGLY4511 and pGLY4512, respectively. A 759-bp STE13 5′ flanking region fragment was digested from pGLY4511 using EcoRI and subcloned into a P. pastoris Ura5-blaster vector, similar to pJN396 (Nett and Gerngross 2003), previously digested with the same restriction enzyme and treated with calf intestinal alkaline phosphatase (CIAP). Following transformation into XL10-Gold competent cells and confirmation by restriction analysis, the resultant vector was designated as pGLY4518. The vector pGLY4512 was digested with HindIII to release a 936-bp fragment encoding the STE13 3′ flanking region and subcloned into pGLY4518 previously digested with the same enzyme and CIAP-treated. The ligation product was transformed into XL10-Gold competent cells and designated as pGLY4520 following restriction analysis. This final STE13::Ura5 knockout vector is illustrated in Fig. 1a. Generation of DAP2::Ura5 knockout vector The DAP2 5′ and 3′ flanking regions were amplified from P. pastoris genomic

DNA as described above using the primer sets SH782+ SH783 and SH784+SH785, to generate 1,003 and 1,142 bp fragments, respectively. Following cloning into pCR2.1 and sequencing, the vectors were designated as pGLY4513 and pGLY4514, encoding the DAP2 5′ and 3′ regions, respectively. Following a similar strategy outlined above, the 991-bp DAP2 5′ region was subcloned into the EcoRI site in the Ura5blaster vector, giving the intermediate construct pGLY4519. Subsequently, the 1,126-bp DAP2 3′ region was subcloned into the HindIII site of pGLY4519, giving the DAP2::Ura5 knockout vector pGLY4521, illustrated in Fig. 1b. Generation of dominant marker DAP2 knockout vector PCR fusion was used to generate the DAP2 knockout vector. Briefly, the DAP2 5′ and 3′ fragments were amplified from pGLY4521 with the primer pairs SH782 + SH805 and SH808+SH785 using PfuUltra™ DNA polymerase. Likewise, the nourseothricin (Nat) marker cassette was amplified from pAG25 (Goldstein and McCusker 1999) using the primers SH806+SH807. The PCR reactions were run on a DNA agarose gel and the 1,011-, 1,151-, and 1,248-bp

2576

Appl Microbiol Biotechnol (2014) 98:2573–2583

A

ƒFig. 1

EcoRI EcoRI(395) (395) SfiI SfiI (401) (401)

STE13 5'

XhoI(1152) (1152) XhoI

pUC19 EcoRI (1158) EcoRI (1158)

Vectors generated to knockout P. pastoris STE13 and DAP2 genes. Represented are the two URA5-marked vectors pGLY4520 (a) and pGLY4521 (b), used to knockout STE13 and DAP2, respectively. Also represented is the nourseothricin-marked vector pGLY5019 (c), used to knockout DAP2. STE13 and DAP2 flanking regions are highlighted in black with SfiI restriction sites used to excise the knockout fragments prior to yeast transformation being underlined

lacZ repeat

pGLY4520 (6748 bp)

HindIII HindIII(4508) (4508)

fragments, corresponding to DAP2 5′, 3′, and the Nat marker, respectively, were isolated. Subsequently, 20 ng of each of these products were combined and fused together using PfuUltra™ DNA polymerase and the primer pair SH782+ SH785. Following incubation for 10 min at 72 °C with ExTaq™ DNA polymerase (TaKaRa, Bio. Inc., Japan), the amplified 3,321-bp fragment was cloned into pCR2.1 (Invitrogen, Carlsbad, CA) and transformed into TOP10 competent cells. DNA sequencing confirmed that the DAP2::Nat fusion was correct and the resultant vector designated as pGLY5019. This vector is illustrated in Fig. 1c.

PpURA5

STE13 3'

SfiI(4495) (4495) SfiI

lacZ repeat

XhoI (3574) XhoI (3574) HindIII (3568) HindIII (3568)

B

EcoRI (395) SfiI (401) I (401) Sfi

PpDAP2 5'

pUC19

XhoI (1384) XhoI (1384) EcoRI (1390) EcoRI (1390)

pGLY4521

lacZ repeat

(7173 bp)

HindIII(4933) (4933) HindIII PpURA5

SfiI SfiI(4920) (4920) PpDAP2 3' lacZ repeat

XhoI(3806) (3806) XhoI HindIII (3800) HindIII (3800)

C

EcoRI (282) (282) EcoRI EcoRI (294) EcoRI (294) SfiI(300) (300) SfiI

pUC ori DAP2 5'

Amp

TTEF

pGLY5019

NatR

(7227 bp) Kan PTEF

f1 ori

DAP2 3'

SfiI (3593) (3593) HindIII (3606) HindIII (3606) EcoRI (3619) EcoRI (3619)

Generation of reporter protein expression vectors The TNFR2:Fc expression vector pGLY3465 was previously described in Hamilton et al. (2013). Briefly, this vector integrated into the P. pastoris genome at the TRP2 loci using the bleomycin resistance cassette to confer resistance against Zeocin™. This vector possessed a single TNFR2:Fc expression cassette under the control of the inducible AOX1 promoter. The TNFR2:Fc was fused C-terminal to the human serum albumin secretion signal (HSAss) to facilitate secretion of the TNFR2:Fc from the cell. The granulocyte macrophage colony-stimulating factor (GMCSF) expression vector (pGLY13860) was generated by replacing the sequence encoding the TNFR2:Fc amino acid sequence in pGLY3465 with sequence encoding amino acids 18 to 144 of GM-CSF (GenBank accession number NP_000749.2), a Gly-Gly-Gly-Gly-Ser linker and 6xHis tag codon-optimized for P. pastoris expression by GeneArt® (Regensburg, Germany) using the GeneOptimizer® software. The resultant GM-CSF open reading frame possessed an N-terminal HSAss. The rat erythropoeitin (rEPO) expression vector (pGLY4510) generated was similar to the TNFR2:Fc expression vector pGLY3465 except that a synthetic open reading frame encoding amino acids 1–89 of the Saccharomyces cerevisiae α-mating factor pre-pro signal sequence (αMFpp), a Glu-Phe motif, amino acids 27 to 192 of rEPO (GenBank accession number NP_058697.1), a 17-amino acid stuffer region, and a C-terminal 6xHis tag was introduced into the AOX1 expression cassette. This synthetic rEPO ORF was codon-optimized for P. pastoris expression using the GeneOptimizer® software.

Appl Microbiol Biotechnol (2014) 98:2573–2583

2577

Generation of STE13 and DAP2 knockout strains Generation of glycoengineered STE13 and DAP2 knockout strains A P. pastoris auxotrophic glycoengineered cell line YGLY7406 [Δoch1, Δpno1, Δmnn4B, Δbmt2, Δura5, Kluyveromyces lactis and Mus musculus UDP-GlcNAc transporters, M. musculus α-1,2-MnsI, Homo sapiens β-1,2GlcNAc transferase I, Rattus norvegicus β-1,2-GlcNAc transferase II, Drosophila melanogaster MnsII, Schizosaccharomyces pombe Gal epimerase, D. melanogaster U D P - G a l t r a n s p o r t e r, a n d H . s a p i e n s β - 1 , 4 galactosyltransferase] expressing GS5.0 glycans (Bobrowicz et al. 2004) was used as the starting strain for all manipulations. This strain expresses the human tumor necrosis factor receptor II fused to the Fc domain of IgG1 (TNFR2:Fc) as a reporter protein and was generated by transforming p G LY 3 4 6 5 l i n e a r i z e d w i t h S p eI i n t o a n e m p t y glycoengineered parent strain (Hamilton et al. 2013). This host strain was transformed with pGLY4520 and 4521 to knockout STE13 or DAP2, respectively, and selected on Ura minus minimal plates. Successful knockouts of each gene were confirmed using the 5′, 3′ and knockout primer sets listed in Table 2. The STE13 and DAP2 knockout strains were named YGLY8084 and YGLY8090, respectively. Subsequently, the double STE13/DAP2 knockout strain was generated by transforming the STE13 knockout strain YGLY8084 with pGLY5019, previously digested with SfiI. Transformants were plated on 100 μg/mL Nat YSD plates and successful double knockouts confirmed using the 5′, 3′, and knockout primer sets listed in Table 2. A representative double knockout strain was designated as YGLY8096. Generation of wild-type STE13 and DAP2 knockout strains The P. pastoris wild-type strain NRRL-Y11430 was used as the starting strain for assessment of DPP activity in a nonglycoengineered P. pastoris background. A URA5 auxotrophic derivative of this strain was generated and designated as YGLY1-3. The vector pGLY4520 was linearized with SfiI and transformed by electroporation into YGLY1-3. Following

Table 2 Primers used to screen yeast strains for knockouts

One kilobase pair product obtained with the presence of wildtype loci

Expression and purification of reporter protein Recombinant TNFR2:Fc for the purpose of this study was generated in bioreactors using inoculum seed flasks as described in Hopkins et al. (2011). Briefly, an inoculum seed flask of each strain was grown for 48 h at 24 °C. This was subsequently used to inoculate 1 L (fedbatch-pro, DASGIP BioTools) bioreactors charged with 0.54 L 4 % BSGY media. Subsequently, fermentations were performed through batch and fed-batch stages prior to being induced with methanol for 36–60 h at 24 °C as described previously (Hopkins et al. 2011). Following expression, the cells were removed by centrifugation at 2,000 rpm for 15 min. The TNFR2:Fc fusion protein was captured by affinity chromatography from the

Knockout

Vector

Region

Primer pair

Product size (kb)

STE13::Ura5

pGLY4520

DAP2::Ura5

pGLY4521

5′ crossover 3′ crossover Knockouta 5′ crossover 3′ crossover

SH778+SH558 SH779+SH491 SH780+SH781 SH786+SH558 SH787+SH491

1.0 1.1 No product 1.2 1.4

Knockouta 5′ crossover 3′ crossover Knockouta

SH788+SH789 SH786+SH380 SH787+SH379 SH788+SH789

No product 1.2 1.4 No product

DAP2::Nat a

selection on uracil minus plates and PCR confirmation, a representative strain was designated as YGLY19659. The vector pGLY4521 was linearized with SfiI and transformed by electroporation into YGLY1-3. Following selection on uracil minus plates and PCR confirmation, a representative strain was designated as YGLY20330. A double STE13/ DAP2 knockout strain was generated by linearizing pGLY5019 with SfiI and transforming into YGLY19659. Following selection on 100 μg/mL nourseothricin YSD plates and PCR confirmation, a representative strain was designated as YGLY20341. The nonglycoengineered reporter protein expression strains were generated by transforming the empty host strains generated in the previous paragraph with the GM-CSF and rEPO expression vectors. Briefly, pGLY13860 was digested with SpeI and transformed by electroporation into wild type ( N R R L - Y 11 4 3 0 ) , Δs t e 1 3 ( Y G LY 1 9 6 5 9 ) , Δd a p 2 (YGLY20330), and Δste13/Δdap2 (YGLY20341) to produce the strains YGLY34145, YGLY34146, YGLY34152, and YGLY34157, respectively. Likewise, rEPO expression strains were generated by linearizing pGLY4510 with SpeI and transforming the same empty parent strains to generate YGLY35715, YGLY35718, YGLY35722, and YGLY35724, respectively.

pGLY5019

2578

supernatant using Streamline rProtein A resin as described previously (Hamilton et al. 2013). Recombinant GM-CSF and rEPO were generated using microreactors as described previously by Choi et al. (2012). Briefly, each strain was inoculated with 3.5 mL of 4 % BSGY in a Whatman 24-well Uniplate and incubated for 65–72 h at 24 °C. One milliliter of this inoculum was used to inoculate 4 mL of 4 % BSGY in a Micro24 reactor plate. The Micro24 fermentation process was performed through batch and methanol induction stages at 24 °C for 48 h as described previously (Choi et al. 2012). Following centrifugation to remove the cells, recombinant GM-CSF or rEPO were purified using nickel column affinity purification as described previously (Hamilton et al. 2006). Reporter protein proteolytic digestion TNFR2:Fc samples (400 μg) were reduced and denatured in 10 mM dithiothreitol (DTT), 4 M guanidine hydrochloride (GuHCl), and 25 mM NH4HCO3 pH 7.8 for 60 min at 37 °C. The reduced Cys residues were alkylated in 50 mM iodoacetic acid (IAA). Alkylation reaction proceeded in the dark for 45 min at room temperature. Residual chemicals (DTT, GuHCl, and IAA) were removed by buffer exchange to 25 mM NH4HCO3 pH 7.8 using 10 KD MWCO spin tubes. Trypsin was added to the sample at a ratio of 1:100 (w/w) and incubated at 37 °C overnight. Subsequently, N-terminal peptides were mapped by liquid chromatography-mass spectrometry. Liquid chromatography-mass spectrometry (LC-MS) Peptide mapping was performed on a Thermo LTQ Orbitrap XL mass spectrometer (Thermo Electron, San Jose, CA) equipped with an Acquity UPLC (Waters, Milford, MA). Digested protein was injected onto an Acquity BEH C18 1.7 μm (2.1×100 mm) reversed phase column. Solvent A was 0.1 % FA in water and solvent B was 0.1 % FA in acetonitrile. The peptide elution gradient was 5 % B from 0 to 5 min, then 5 to 30 % B in 40 min at a flow rate of 0.2 mL/ min. The LTQ was operated in positive ion mode and the settings are as follows: spray voltage 4.5 kV, capillary voltage 47 V, capillary temperature 275 °C, and tube lens 90 V. Sheath gas was 25 arb, Aux gas was 2 arb, and no sweep gas was used. The mass spectrometer was operated only in ion trap mode with a mass range from 920 to 1,038 Da. Spectra were collected in the centroid mode. The top three strongest ions were selected for MS/MS CID. Isolation width was 1 Da, normalized collision energy 35 %, activation Q 0.25, and activation time was set to 30 ms. Dynamic exclusion was enabled with two repeated counts during 30 s. Collected spectra were analyzed with Xcalibur or Proteome Discoverer software.

Appl Microbiol Biotechnol (2014) 98:2573–2583

Quadruple time-of-flight mass spectrometry (Q-TOF) Mass spectrometric analysis was done in positive ion mode on Accurate-Mass Q-TOF LC/MS 6520 (Agilent technologies, Santa Clara, CA). The protocol used was as previously described (Choi et al. 2012), except that only intact glycoprotein analysis was performed.

Results Generation and confirmation of DPP knockout strains In an attempt to elucidate the roles of DPPs in the Nterminal clipping of TNFR2:Fc secreted from P. pastoris, a series of knockout vectors were generated to disrupt STE13 and/or DAP2 in the P. pastoris genome (Fig. 1). Each vector was constructed by cloning the 5′ and 3′ regions of the genes for STE13 or DAP2 into vectors in combination with either URA5 or NatR markers so that transformation would result in disruption of the open reading frame (ORF) present in each gene (as depicted in Fig. 2). Specifically, the STE13 knockout vector (pGLY4520, Fig. 1a) was designed to knockout amino acids 112 to 697 of the 869-amino acid STE13 ORF (GenBank accession CCA38024), using the URA5 marker. To knockout DAP2, the vector pGLY4521 (Fig. 1b) was designed to knockout amino acids 136 to 663 of the 816amino acid ORF (GenBank accession XP_002493132), using the URA5 marker. A second DAP2 knockout vector (pGLY5019, Fig. 1c) was generated by replacing the URA5 marker in pGLY4521 with the nourseothricin marker to facilitate the knockout of DAP2 in previously generated Δste13 strain background. Partial ORF knockouts were chosen for STE13 and DAP2 due to the juxtapositioning on neighboring genes, which we did not wish to disrupt through significant truncation of their promoters or terminators. The knockout vectors were used to knockout STE13 and/or DAP2 in a P. pastoris strain expressing TNFR2:Fc, which had previously been engineered to produce bi-antennary N-linked glycans terminating in galactose. For each strain, PCR amplifications were performed for the relevant 5′ and 3′ crossover products and the original DPP locus to confirm successful disruption of each ORF. Figure 3 confirms the genotype of the native and the knockout status of four strains representing DPP wild type, Δste13, Δdap2, and Δste13/Δdap2. The first lane represents the DPP wild-type strain (YGLY6646) and demonstrates that both STE13 and DAP2 loci are intact (Fig. 3a, b, lane 1). The second lane represents the Δste13 strain (YGLY8084), confirming crossover of the 5′ and 3′ regions at the STE13 locus, and the elimination of the internal ORF fragment from the wild-type locus (Fig. 3a, lane 2), thus

Appl Microbiol Biotechnol (2014) 98:2573–2583

2579 EcoRI EcoRI(395) (395) SfiI SfiI (401) (401)

Fig. 2 Illustrative representation of DPP knockout strategy. The vectors illustrated in Fig. 1 are used to knockout the DPP genes in P. pastoris. Briefly, each vector is digested with the restriction enzyme SfiI (underlined) to produce a linear deletion allele, which possesses 5′ and 3′ flanking regions homologous to the genomic locus of the target gene to be knocked out. Transformation of this linear deletion allele into P. pastoris is followed by genetic recombination, resulting in the replacement of the endogenous locus with the linear deletion allele, and in doing so, knocking out the target gene. Here, the STE13 knockout vector (pGLY4520) is used to exemplify the knockout of STE13 locus using URA5

STE13 5'

EcoRI (1158) EcoRI (1158)

lacZ repeat

pGLY4520 (6748 bp)

HindIII HindIII(4508) (4508)

PpURA5

STE13 3'

SfiI(4495) (4495) SfiI

lacZ repeat

XhoI (3574) XhoI (3574) HindIII (3568) HindIII (3568)

5’-overlap region for STE13

5’- region of STE13

Δ dap2

Δ ste13

A

Δ ste13/ Δdap2

5’- region of STE13

WT

XhoI(1152) (1152) XhoI

pUC19

5’ cross -over

URA5-blaster marker

STE13 open reading frame

URA5-blaster marker

3’-overlap region for STE13

Linear deletion allele

3’- region of STE13

Genomic locus

3’- region of STE13

Knockout locus

confirming the successful knockout of STE13. Like the wildtype strain, the locus for the DAP2 in the Δste13 strain remains unaltered (Fig. 3b, lane 2). The Δdap2 strain (YGLY8090) by contrast shows an intact STE13 locus while the DAP2 locus has been successfully disrupted (Fig. 3a, b, lane 3). Finally, the disruption of both STE13 and DAP2 loci is confirmed in the double knockout strain (YGLY8096) (Fig. 3a, b, lane 4).

3’ cross -over WT loci STE13 Loci

B 5’ cross -over 3’ cross -over WT loci DAP2 Loci Fig. 3 PCR confirmation of DPP knockout. The primer sets described in Table 2 were used to confirm the genetic knockout of STE13 and DAP2 in P. pastoris strains. Represented are the 5′ and 3′ crossover PCR amplifications, along with PCR amplifications of the original DPP locus. (a) The STE13 loci in a DPP wild-type strain (YGLY6646), a STE13 knockout strain (YGLY8084), a DAP2 knockout strain (YGLY8090), and a double STE13/DAP2 knockout strain (YGLY8096). (b) The DAP2 loci of the same strains assessed in (a)

Characterization of N-terminal clipping of TNFR2:Fc in glycoengineered P. pastoris strains Following expression and purification of the TNFR2:Fc from the four strains representing DPP wild type, Δste13, Δdap2 and Δste13/Δdap2, LC-MS peptide analysis was used to analyze the degree of N-terminal processing (Fig. 4). The elution profile of TNFR2:Fc expressed from the wild-type strain (YGLY6646) demonstrated that approximately 94 % of the total peptide population had the first two amino acids removed (represented by−LP), while only approximately 6 % was fully intact peptide. By contrast, the elution profile of TNFR2:Fc expressed in the Δste13 strain (YGLY8084) clearly demonstrated that Ste13p played a major role in the proteolysis of the N-terminus of TNFR2:Fc. Of the total peptide population from this strain, approximately 78 % was fully intact. By contrast, the role of the DAP2 knockout was not nearly as important in

2580

Appl Microbiol Biotechnol (2014) 98:2573–2583

TNFR2:Fc TNFR2:Fc (–LP)

Determination if the requirement for the double DPP knockout is TNFR2:Fc specific

100 80 60

Wild-type

40 20 100

Relative Abundance

80 60

Δste13

40 20 100 80 60

Δdap2

40 20 100

The N-terminal processing of two additional reporter proteins, for which we had previously observed N-terminal clipping, was expressed from the four nonglycoengineered strain backgrounds and assessed for DPP activity. Recombinant granulocyte/macrophage-colony stimulating factor (GMCSF) was transformed into wild-type, Δste13, Δdap2, and Δste13/Δdap2 knockout strains. Likewise, recombinant rat erythropoietin (rEPO) was also transformed into the same set of empty base strains. Of particular note is that the intact secreted rEPO possessed an artificial glutamine-alanine repeat at the N-terminus of the molecule, as a consequence of fusion to the S. cerevisiae α-mating factor prepro (αMFpp) secretion sequence. As such, this rEPO variant was distinguished by designating the variant as rEPO(+EAEA). For the study, both of the GM-CSF and rEPO (+EAEA) reporter proteins were PNGase-treated to remove N-glycans and then analyzed by Q-TOF. Figure 5 shows the Q-TOF analysis of GM-CSF

80 60

GMCSF (–AP)

Δste13/Δ dap2

40

GMCSF

20 6

24

26

28

30

32

x10 3

34

Time (min)

2 1

Wild-type

-H 15264

0 6

x10 2

15569

Δ ste13

15401

1 Counts

Fig. 4 LC-MS N-terminal peptide analysis of TNFR2:Fc expressed in glycoengineered P. pastoris. TNFR2:Fc was isolated from representative DPP wild-type (YGLY6646), STE13 knockout (YGLY8084), DAP2 knockout (YGLY8090), and double STE13/DAP2 knockout (YGLY8096) strains. Following tryptic digestion, the N-terminal peptide was identified using LC-MS and characterized for N-terminal intactness. The intact peptide elutes at 31 min, while the truncated peptide (−LP) minus the N-terminal Leu-Pro dipeptide elutes at 27 min

15401

0 x10

6

15401

protecting the N-terminus from proteolysis since only 10 % of TNFR2:Fc was intact. However, when STE13 and DAP2 where both knocked out in the same strain (YGLY8096), no significant N-terminal truncation was detected. This result was in contrast to the findings of Prabha et al. (2009), which reported that a recombinant protein expressed in a wild-type P. pastoris strain required only the STE13 knockout to fully prevent N-terminal dipeptidyl aminopeptidase activity. To determine if our result was an artifact of expressing TNFR2:Fc in a glycoengineered P. pastoris background, we generated Δste13, Δdap2, and Δste13/Δdap2 in the wild-type NRRL-Y11430 strain background. When TNFR2:Fc was expressed in these knockout backgrounds and wild-type NRRL-Y11430, the same pattern of Nterminal proteolysis was observed, i.e., only the STE13/ DAP2 double knockout strain prevented N-terminal clipping (Supplementary Fig. S1).

2 1

Δ dap2

-H 15264

0 6

x10 4 2

15569

-H

Δ ste13/ Δ dap2

15432 0 15200 15300 15400 15500 15600 15700 Deconvoluted Mass (amu)

Fig. 5 Q-TOF analysis of GM-CSF expressed in nonglycoengineered P. pastoris. GM-CSF was isolated from DPP wild-type (YGLY34145), STE13 knockout (YGLY34146), DAP2 knockout (YGLY34152), and double STE13/DAP2 knockout (YGLY34157) strains. Following PNGase treatment to remove N-linked glycans, the samples were analyzed by Q-TOF. Intact GM-CSF peptide is observed with a deconvoluted mass of 15,569 mass units, while GM-CSF lacking the N-terminal Ala-Pro motif (−AP) has a reduced mass of 15,401 mass units. Peptides lacking one of the His residues from the C-terminal 6xHis tag are indicated on the profile by “–H”

Appl Microbiol Biotechnol (2014) 98:2573–2583

2581

expressed from each of the DPP knockout strains. Using this reporter protein, we observed a similar DPP knockout requirement as for TNFR2:Fc. Briefly, in the wild-type strain (YGLY34145), all of the GM-CSF population had the Nterminal alanine-proline di-peptide motif clipped, as depicted by “−AP.” A similar observation was also made when GMCSF was expressed in the DAP2 knockout background (YGLY34152). By contrast, expression of GM-CSF from the STE13 knockout strain (YGLY34146) indicated that the majority of the peptide was intact, though some clipped product could still be detected. However, when GM-CSF was expressed from the STE13/DAP2 double knockout strain (YGLY34157), complete protection of the N-terminus was observed. DPP processing of the recombinant rEPO(+EAEA) Nterminus can produce three variants of this protein: rEPO plus four additional amino acids (+EAEA), rEPO plus two amino acids (+EA), or rEPO sequence in isolation. Figure 6 shows the Q-TOF analysis of rEPO expressed from the various DPP rEPO

rEPO rEPO (+EA) (+EAEA)

6

x10 1

0.5

21660

Wild-type -H 21322

21459

-H

21860

knockout strains. Consistent with the observations for TNFR2:Fc and GM-CSF, rEPO(+EAEA) expressed in a wildtype strain (YGLY35715) had undergone significant proteolysis at the N-terminus. While the majority of the secreted protein had the first two amino acids (EA) removed, a fraction of the protein also had had both “EA” motifs removed, with only the rEPO sequence remaining. A similar observation was made when rEPO(+EAEA) was isolated from the DAP2 knockout strain (YGLY35722). Interestingly, rEPO(+EAEA) isolated from the STE13 only knockout strain background (strain YGLY35718) demonstrated full protection, with the entire population of secreted rEPO possessing all four additional amino acids (+EAEA). As expected, the STE13/DAP2 double knockout background (strain YGLY35724) also showed complete protection of the N-terminus. Interestingly, both TNFR2:Fc and GM-CSF possessed the human serum albumin (HSA) secretion signal, while the rEPO possessed the αMFpp secretion signal. To assess if the difference in DPP knockout requirement was a consequence of altered processing of the different signal sequences, we exchanged signal sequences. Analyses of the N-termini of the TNFR2:Fc and rEPO reporter proteins indicated that the DPP knockout requirement remained the same for each protein, i.e., TNFR2:Fc with the αMFpp secretion signal still required the double DPP knockout, while rEPO with the HSA secretion sequence only required the single STE13 knockout (Supplementary Figs. S2 and S3, respectively).

0 4

x10

21860

Δ ste13

Counts

5

0 5

x10 1.5

21660

Δdap2

1 0.5

21460

21860

0 5

x10

21860

3

Δ ste13/ Δ dap2

2

-H

1

21723

0 21300

21500

21700

21900

22100

Deconvoluted Mass (amu)

Fig. 6 Q-TOF analysis of N-terminally modified rEPO expressed in nonglycoengineered P. pastoris. Recombinant rEPO(+EAEA) was isolated from DPP wild-type (YGLY35715), STE13 knockout (YGLY35718), DAP2 knockout (YGLY35722), and double STE13/DAP2 knockout (YGLY35724) strains. Following PNGase treatment to remove Nlinked glycans, the samples were analyzed by Q-TOF. Intact rEPO(+EAEA) peptide is observed with a deconvoluted mass of 21,860 mass units, while (+EAEA) lacking a single (−EA) or double (−EAEA) N-terminal GlurEPO Ala motif have reduced masses of 21,660 and 21,460 (+/−1 amu) mass units. Peptides lacking one of the His residues from the C-terminal 6xHis tag are indicated on the profile by “–H”

Discussion Therapeutic proteins are an important class of compounds in the pharmaceutical industry. Although mammalian cell lines are the industrially preferred platform for producing glycosylated therapeutic proteins, a number of alternative expression platforms have been developed, including the methylotrophic yeast P. pastoris (Betenbaugh et al. 2004; Sethuraman and Stadheim 2006; Hamilton and Gerngross 2007). Expression of therapeutic proteins in the P. pastoris platform, however, has inherent difficulties. Besides the requirement for significant engineering of the glycosylation pathway, the existence of endogenous proteases can exhibit deleterious effects on a secreted protein. An understanding of the action of these proteases is an important aspect of optimizing the P. pastoris platform further. Recently, while studying the expression of recombinant TNFR2:Fc in P. pastoris, we observed clipping of two amino acids from the N-terminus of the molecule (Hamilton et al. 2013). A class of proteases that is involved in such cleavage is that of the dipeptidyl aminopeptidases. In S. cerevisiae, two such proteins have been identified, Ste13p and Dap2p. Homologs of these diaminopeptidases have been identified in P. pastoris (Prabha et al. 2009). In this study, we investigated the roles of Ste13p and Dap2p in the N-terminal

2582

proteolysis of TNFR2:Fc and two other reporter proteins (GM-CSF, and rEPO(+EAEA)), for which we had previously detected N-terminal cleavage of the first two amino acids. Each reporter protein was expressed in a P. pastoris wildtype strain, a STE13 knockout strain, a DAP2 knockout strain, or a STE13/DAP2 double knockout strain. In the wild-type background, each reporter protein showed complete or near-complete DPP processing of the first two amino acids. More specifically, GM-CSF demonstrated complete N-terminal cleavage, while residual amounts of intact TNFR2:Fc and rEPO(+EAEA) were detected. The presence of varying levels of intact recombinant protein secreted from wild-type P. pastoris was not a surprising observation. Previous examples of partial or no DPP activity on proteins secreted from wild-type P. pastoris have been reported in the literature. For example, expression of interferon-alpha 2b fused to the αMFpp secretion signal, with the glutamine-alanine repeats following the secretion signal, resulted in a heterologous mixture of secreted protein, with and without the glutaminealanine repeats (Ghosalkar et al. 2008). In another example, when an anti-B-type natriuretic peptide single chain variable fragment (scFv), also fused to the αMFpp secretion signal with the glutamine-alanine repeats, was expressed in wildtype P. pastoris, no cleavage of the glutamine-alanine repeats was observed (Maeng et al. 2012). While reasons for the varying levels of DPP activity observed have been discussed by Maeng et al., it is clearly difficult to predict DPP activity in wild-type P. pastoris. When recombinant TNFR2:Fc was expressed in the STE13 knockout background, we observed that the majority of the protein was now protected from N-terminal proteolysis, though some clipping did remain. This was an interesting observation since it had previously been shown in P. pastoris that the single knockout of STE13 was able to protect the recombinant protein extendin-4 from DPP activity (Prabha et al. 2009). Indeed, it was only when we expressed TNFR2:Fc in a STE13/DAP2 double knockout background that we observed full protection of the N-terminus of this molecule. A possible factor in the different requirements for DPP knockout may have been due to the TNFR2:Fc being expressed in a highly glycoengineered P. pastoris strain background. To determine if our previous engineering efforts had altered protease activity, we assessed TNFR2:Fc expression in a wild-type P. pastoris background. However, once again we observed the requirement for the double knockout for full protection of the TNFR2:Fc N-terminus. Having concluded that indeed both STE13 and DAP2 knockouts were needed for protection of the N-terminus of TNFR2:Fc, we went to investigate if TNFR2:Fc was unique in this requirement. For this purpose, GM-CSF and rEPO(+EAEA), both of which had previously demonstrated N-terminal clipping, were assessed. Interestingly, these two reporter proteins demonstrated differing requirements for DPP knockout. Like TNFR2:Fc, the double

Appl Microbiol Biotechnol (2014) 98:2573–2583

knockout of STE13 and DAP2 was required to fully protect the N-terminus of GM-CSF. By contrast, the expression of intact recombinant rEPO(+EAEA) only required the single knockout of STE13 for full protection. This latter example agreed with what was observed when exendin-4 was expressed in P. pastoris (Prabha et al. 2009). Together these observations clearly show that while Ste13p plays a major role in the DPP processing of proteins secreted from P. pastoris, Dap2p also plays a significant role in particular cases. In summary, the data presented in this study further elucidates the role of Ste13p and Dap2p in the posttranslational processing of recombinant proteins secreted from P. pastoris. As such, this understanding of DPP activity enhances the utility of the P. pastoris expression system, thus facilitating the production of recombinant therapeutic proteins with the desired peptide sequences. Acknowledgments The authors would like to thank the members of the Strain Development, Purification, Analytical, High Through-put Screening and Fermentation groups of GlycoFi, Inc. who supported this study. Conflict of interest The authors declare that they have no conflict of interest.

References Betenbaugh MJ, Tomiya N, Narang S, Hsu JT, Lee YC (2004) Biosynthesis of human-type N-glycans in heterologous systems. Curr Opin Struct Biol 14:601–606 Bobrowicz P, Davidson RC, Li H, Potgieter TI, Nett JH, Hamilton SR, Stadheim TA, Miele RG, Bobrowicz B, Mitchell T, Rausch S, Renfer E, Wildt S (2004) Engineering of an artificial glycosylation pathway blocked in core oligosaccharide assembly in the yeast Pichia pastoris: production of complex humanized glycoproteins with terminal galactose. Glycobiology 14:757–766 Choi BK, Bobrowicz P, Davidson RC, Hamilton SR, Kung DH, Li H, Miele RG, Nett JH, Wildt S, Gerngross TU (2003) Use of combinatorial genetic libraries to humanize N-linked glycosylation in the yeast Pichia pastoris. Proc Natl Acad Sci U S A 100:5022–5027 Choi BK, Warburton S, Lin H, Patel R, Boldogh I, Meehl M, d’Anjou M, Pon L, Stadheim TA, Sethuraman N (2012) Improvement of Nglycan site occupancy of therapeutic glycoproteins produced in Pichia pastoris. Appl Microbiol Biotechnol 95:671–682 Chung CH, Mirakhur B, Chan E, Le QT, Berlin J, Morse M, Murphy BA, Satinover SM, Hosen J, Mauro D, Slebos RJ, Zhou Q, Gold D, Hatley T, Hicklin DJ, Platts-Mills TA (2008) Cetuximab-induced anaphylaxis and IgE specific for galactose-alpha-1,3-galactose. N Engl J Med 358:1109–1117 Demain AL, Vaishnav P (2009) Production of recombinant proteins by microbes and higher organisms. Biotechnol Adv 27:297–306 Ghaderi D, Taylor RE, Padler-Karavani V, Diaz S, Varki A (2010) Implications of the presence of N-glycolylneuraminic acid in recombinant therapeutic glycoproteins. Nat Biotechnol 28:863–867 Ghosalkar A, Sahai V, Srivastava A (2008) Secretory expression of interferon-alpha 2b in recombinant Pichia pastoris using three different secretion signals. Protein Expr Purif 60:103–109 Goldstein AL, McCusker JH (1999) Three new dominant drug resistance cassettes for gene disruption in Saccharomyces cerevisiae. Yeast 15: 1541–1553

Appl Microbiol Biotechnol (2014) 98:2573–2583 Hamilton SR, Gerngross TU (2007) Glycosylation engineering in yeast: the advent of fully humanized yeast. Curr Opin Biotechnol 18:387–392 Hamilton SR, Bobrowicz P, Bobrowicz B, Davidson RC, Li H, Mitchell T, Nett JH, Rausch S, Stadheim TA, Wischnewski H, Wildt S, Gerngross TU (2003) Production of complex human glycoproteins in yeast. Science 301:1244–1246 Hamilton SR, Davidson RC, Sethuraman N, Nett JH, Jiang Y, Rios S, Bobrowicz P, Stadheim TA, Li H, Choi BK, Hopkins D, Wischnewski H, Roser J, Mitchell T, Strawbridge RR, Hoopes J, Wildt S, Gerngross TU (2006) Humanization of yeast to produce complex terminally sialylated glycoproteins. Science 313:1441– 1443 Hamilton SR, Cook WJ, Gomathinayagam S, Burnina I, Bukowski J, Hopkins D, Schwartz S, Du M, Sharkey NJ, Bobrowicz P, Wildt S, Li H, Stadheim TA, Nett JH (2013) Production of sialylated Olinked glycans in Pichia pastoris. Glycobiology 23:1192–1203 Higashi H, Naiki M, Matuo S, Okouchi K (1977) Antigen of “serum sickness” type of heterophile antibodies in human sera: identification as gangliosides with N-glycolylneuraminic acid. Biochem Biophys Res Commun 79:388–395 Hokke CH, Bergwerff AA, Vandedem GWK, Vanoostrum J, Kamerling JP, Vliegenthart JFG (1990) Sialylated carbohydrate chains of recombinant human glycoproteins expressed in Chinese-hamster ovary cells contain traces of N-glycolylneuraminic acid. FEBS Lett 275: 9–14 Hong WJ, Petell JK, Swank D, Sanford J, Hixson DC, Doyle D (1989) Expression of dipeptidyl peptidase IV in rat tissues is mainly regulated at the mRNA levels. Exp Cell Res 182:256–266 Hopkins D, Gomathinayagam S, Rittenhour AM, Du M, Hoyt E, Karaveg K, Mitchell T, Nett JH, Sharkey NJ, Stadheim TA, Li H, Hamilton SR (2011) Elimination of beta-mannose glycan structures in Pichia pastoris. Glycobiology 21:1616–1626 Julius D, Blair L, Brake A, Sprague G, Thorner J (1983) Yeast alpha factor is processed from a larger precursor polypeptide: the essential role of a membrane-bound dipeptidyl aminopeptidase. Cell 32:839– 852 Kumagai Y, Yajima A, Konishi K (2003) Peptidase activity of dipeptidyl aminopeptidase IV produced by Porphyromonas gingivalis is

2583 important but not sufficient for virulence. Microbiol Immunol 47: 735–743 Maeng BH, Choi J, Sa YS, Shin JH, Kim YH (2012) Functional expression of recombinant anti-BNP scFv in methylotrophic yeast Pichia pastoris and application as a recognition molecule in electrochemical sensors. World J Microbiol Biotechnol 28:1027–1034 Matoba S, Morano KA, Klionsky DJ, Kim K, Ogrydziak DM (1997) Dipeptidyl aminopeptidase processing and biosynthesis of alkaline extracellular protease from Yarrowia lipolytica. Microbiology 143: 3263–3272 Misumi Y, Hayashi Y, Arakawa F, Ikehara Y (1992) Molecular cloning and sequence analysis of human dipeptidyl peptidase IV, a serine proteinase on the cell surface. Biochim Biophys Acta 1131:333–336 Montrose-Rafizadeh C, Yang H, Rodgers BD, Beday A, Pritchette LA, Eng J (1997) High potency antagonists of the pancreatic glucagonlike peptide-1 receptor. J Biol Chem 272:21201–21206 Nett JH, Gerngross TU (2003) Cloning and disruption of the PpURA5 gene and construction of a set of integration vectors for the stable genetic modification of Pichia pastoris. Yeast 20:1279–1290 Prabha L, Govindappa N, Adhikary L, Melarkode R, Sastry K (2009) Identification of the dipeptidyl aminopeptidase responsible for Nterminal clipping of recombinant Exendin-4 precursor expressed in Pichia pastoris. Protein Expr Purif 64:155–161 Punt PJ, van Biezen N, Conesa A, Albers A, Mangnus J, van den Hondel C (2002) Filamentous fungi as cell factories for heterologous protein production. Trends Biotechnol 20:200–206 Roberts CJ, Nothwehr SF, Stevens TH (1992) Membrane protein sorting in the yeast secretory pathway: evidence that the vacuole may be the default compartment. J Cell Biol 119:69–83 Sethuraman N, Stadheim TA (2006) Challenges in therapeutic glycoprotein production. Curr Opin Biotechnol 17:341–346 Suarez-Rendueles P, Wolf DH (1987) Identification of the structural gene for dipeptidyl aminopeptidase yscV (DAP2) of Saccharomyces cerevisiae. J Bacteriol 169:4041–4048 Walsh G (2010) Biopharmaceutical benchmarks 2010. Nat Biotechnol 28:917–924 Zhu J (2012) Mammalian cell protein expression for biopharmaceutical production. Biotechnol Adv 30:1158–1170