Protein Expression and Purification 22, 381–387 (2001) doi:10.1006/prep.2001.1460, available online at http://www.idealibrary.com on High-Yield Expression and Purification of Human Interferon a-1 in Pichia pastoris Philip T. Liu, Tuan V. Ta, and Lorelie H. Villarete1 Pepgen Corporation, 1255 Harbor Bay Parkway, Suite B, Alameda, California 94502 Received December 18, 2000, and in revised form April 16, 2001; published online July 2, 2001 For several years, interferon a-1, also known as interferon a-D, has been studied for treatment of various viral diseases, such as hepatic fibrosis caused by hepatitis B, herpes simplex virus keratitis, and bovine respiratory diseases in calves. Currently, recombinant human interferon a-D (rHuIFNaD) is expressed intracellularly in Escherichia coli or secreted by Bacillus subtilis and Saccharomyces cerevisiae. In this report, we describe the process of obtaining a relatively highyield secretion of biologically active recombinant rHuIFNaD using the Pichia pastoris system. The process produced as high as 0.7 mg of purified protein per 20 ml of shake culture of rHuIFNaD with better bioactivity than the commercially available rHuIFNaD molecule produced in E. coli. q 2001 Academic Press Key Words: Pichia pastoris; interferon; protein expression; protein purification; antiviral. Interferon a-1 (IFNa1),2 also known as interferon aD (IFNaD), is a type I interferon of wide research and clinical interest. Recent reports have demonstrated the efficacy of recombinant IFNaD in the treatment of various viral diseases in humans as well as in animals (1–5). Due to the clinical and research interest in human IFNaD, different expression systems have been developed and employed. Expression of recombinant human IFNaD (rHuIFNaD) in 1982 was described for a Methylophilus methylotrophus system and an Escherichia coli system which both utilized the lac promoter 1 To whom correspondence should be addressed. Fax: (510) 4730005. E-mail: loreliev@pepgen.com. 2 Abbreviations used: IFNa1, interferon a-1; IFNaD, interferon aD; rHuIFNaD; recombinant human interferon a-D; IEF, isoelectric focusing; TCA, trichloroacetic acid; oIFNt, ovine IFNt; BSA, bovine serum albumin; PVDF, polyvinylidene difluoride. 1046-5928/01 $35.00 Copyright q 2001 by Academic Press All rights of reproduction in any form reserved. (6). In 1984, Genentech scientists reported a Saccharomyces cerevisiae system in which the IFNaD gene fused with the a-factor prepro signal sequence yielded a secreted IFNaD protein that had relatively low biological activity (7). Several years later, in 1989, secretory expression of human interferon genes in E. coli and Bacillus subtilis, using the staphylokinase heterologous expression–secretion signal, was developed (8). In these studies, only the B. subtilis system, and not the E. coli system, was able to secret rHuIFNaD into the culture medium. However, a significant improvement on the intracellular expression of rHuIFNaD in E. coli was reported in 1990 by the use of a defined medium in a fed batch mode during fermentation that allowed for a more efficient expression of the protein at reduced specific growth rates in E. coli (9). In this report, we describe yet another step in the development of the secreted expression of rHuIFNaD using the methylotrophic yeast, Pichia pastoris. P. pastoris is capable of metabolizing methanol as its sole carbon source by inducing the production of alcohol oxidase (10). Although P. pastoris codes for two alcohol oxidase genes, AOX1 and AOX2, the AOX1 gene is responsible for 85% of alcohol oxidase activity in the yeast cell. The expression of the AOX1 gene is tightly regulated by the AOX1 promoter. In addition, the use of the secretion signal sequence from the S. cerevisiae a-factor prepro peptide has been successful in the secreted expression of heterologous proteins in the P. pastoris system. Consequently, by inserting the rHuIFNaD gene downstream to the highly inducible AOX1 promoter and the prepro a-factor signal sequence, we were able to efficiently produce rHuIFNaD protein in relatively high yields with greater biological activity when compared to the commercially available rHuIFNaD molecule. 381 382 LIU, TA, AND VILLARETE MATERIALS AND METHODS Materials. SacI restriction endonuclease was purchased from Roche Molecular Biochemicals (Indianapolis, IN). All P. pastoris plasmid and yeast expression reagents were purchased from Invitrogen (Carlsbad, CA). All oligonucleotide primers, including primers for mutagenesis and the 58AOX1 and 38AOX1 primers for colony PCR screening and sequencing were synthesized by Life Technologies (Gaithersburg, MD) (58AOX1 primer sequence, 58-GACTGGTTCCAATTGACAAGC38; 38AOX1 primer sequence, 58-GCAAATGGCATTCTGACATCC-38). All PCR reagents excluding the primers were purchased from Perkin–Elmer/Roche (Branchburg, NJ). Microcon-100, Centricon-20, and MicropureEZ filter units were from Millipore Corporation (Bedford, MA). ZymoLyase was from Seikagaku, America (Ijamsville, MD). rHuIFNaA and rHuIFNaD standards were purchased from PBL (New Brunswick, NJ) and Biosource International (Camarillo, CA), respectively. DNA sequencing was performed by the dideoxy chaintermination method using ABI 373 DNA sequencer and ABI Prism BigDye terminator cycle sequencing kit (PE Biosystems Inc., Foster City, CA). The N-terminal sequence of the protein was determined by using Edman degradation on an ABI 494 Procise Sequenator (PE Biosystems Inc., Foster City, CA). DNA sequencing and N-terminal protein sequencing were both performed at the Protein and Nucleic Acid (PAN) Facility at Stanford University. Yeast genomic DNA was isolated using the method outlined in the 5 Prime → 3 Prime PerfectYeast gDNA isolation kit (5 Prime → 3 Prime, Inc., Boulder, CO). Strains and plasmids. INVaF8 One Shot chemically competent E. coli strain (genotype, F8 endA1 recA1 hsdR17 (rk2, mk+) supE44 thi-1 gyrA96 relA1 f80lacZDM15D(lacZY A-argF)U169g), plasmid pPICZa, and P. pastoris X-33 yeast strain were purchased from Invitrogen. Plasmid pNLVg001 contains a modified HuIFNaD synthetic gene inserted into the XhoI and NotI sites of the pPICZa vector (L. H. Villarete, unpublished data). In vitro site-directed mutagenesis. Site-directed mutagenesis was performed using a PCR-based in vitro site-directed mutagenesis method outlined in the QuickChange site-directed mutagenesis kit (Stratagene, San Diego, CA). Four primer pairs were synthesized and used in four consecutive mutagenesis reactions to change plasmid pNLVg001 to pPEPhIFNaD. The first primer pair used is 43 bases in length, while the rest of the primer pairs used are 44 bases. Clones containing plasmids with the desired mutation for each mutagenesis reaction were identified by bacterial colony PCR screening using the 58 and 38 AOX1 primers, followed by DNA sequence analysis of the resulting PCR products. The colony PCR reaction mixture was done in a total volume of 50 ml consisting of 20 ml of resuspended bacterial colony, 10 pmol each of the 58 and 38 AOX1 primers, 2.5 mM MgCl2, 125 mM dATP, 125 mM dTTP, 125 mM dCTP, 125 mM dGTP, 13 PCR Buffer II (50 mM KCl, 10 mM Tris–HCl, pH 8.3). PCR amplification was done for 30 cycles (each for 1 min at 958C, 1 min at 548C and 1 min at 728C) with a final extension of 7 min at 728C on a Perkin–Elmer DNA thermal cycler 480. P. pastoris transformation and expression experiments. The procedure for the transformation and expression experiments are as outlined in the Invitrogen EasySelect manual. Selected positive transformants were grown overnight in BMGY medium, and then the cells were induced using an OD600 of 2 in BMMY medium. The cultures were fed with 1 ml of 10% methanol at 24 and 48 h postinduction. After 72 h of culture, the entire culture supernatant was harvested, filter sterilized, and analyzed by one-dimensional SDS– polyacrylamide gel electrophoresis using a 14% Tris– glycine gel. Proteins were visualized by Colloidal Coomassie staining (Novex, San Diego, CA). A duplicate gel was transblotted onto a PVDF membrane (Millipore Corporation) for Western blotting. The rHuIFNaD protein was detected on the blot by incubation with an anti-human IFNa mouse monoclonal antibody (clone MMHA-3), followed by a biotinylated goat anti-mouse antibody (Biosource International) and a colorimetric detection system using streptavidin–alkaline phosphatase enzyme conjugate, nitroblue tetrazolium, and 58bromo-4-chloro-3-indolylphosphate. Purification of rHuIFNaD from P. pastoris supernatants. Culture supernatants (20 ml volume) were buffer exchanged with 10 mM Tris, 150 mM NaCl and concentrated to a final volume of 2 ml with a Centricon Plus-20 (Millipore Corporation) prior to loading onto a HiPrep Sephacryl 26/60 S-100 high-resolution sizeexclusion column at 48C (Pharmacia, Peapack, NJ). The first 120-ml flowthrough following the injection of the sample were collected at 1 ml/min and designated fraction 1. Three fractions of 20 ml each were collected and designated fractions 2, 3, and 4, respectively. The fractions were concentrated to 1 ml and were run on 14% SDS–PAGE gels and stained using the Novex colloidal blue staining kit. The AlphaImager 2000 software (Alpha Innotech, San Leandro, CA) was used to determine the band density ratios for densitometric analysis and estimations of isoelectric points (pI ) from isoelectric focusing gels. The Lasergene software (DNAStar, Madison, WI) was used to do all DNA analysis. Sample total protein concentration was determined using the bicinchoninic acid assay kit (Pierce, Rockford, IL). Isoelectric focusing. Isoelectric focusing (IEF) was performed using 1.0-mm thick precast IEF vertical gels (5% acrylamide with 2% ampholytes, pH 3–7) (Novex). HuIFNa1 EXPRESSED IN Pichia pastoris 383 The upper chamber cathode buffer used was 40 mM lysine and the lower chamber anode buffer was 7 mM phosphoric acid. The gel was run for 1 h at a 100-V constant and then 1 h for 200V, followed by 500 V for 30 min. Following the run, the gel was fixed using 12% TCA, 3.5% sulfosalicylic acid for 1 h and then stained with colloidal Coomassie stain. Antiviral assay. A standard cytopathic effect inhibition assay using vesicular stomatitis virus (Indiana strain) challenge of Madin Darby bovine kidney cells was performed to quantitate the antiviral activity of the rHuIFNaD in the sample preparations (11). RESULTS In previous experiments, we constructed a series of synthetic hybrid interferon genes using combinations of HuIFNa1 and ovine IFNt (oIFNt) sequences for structure–function analyses (unpublished data; patent pending). The hybrid genes were designed to include P. pastoris preferred codon usage and were inserted into the pPICZa expression vector (Invitrogen) using the XhoI and NotI restriction enzyme sites. The pPICZa vector features the AOX1 gene promoter to drive the expression of the heterologous protein, the a-mating factor prepro signal sequence to target the protein to the secretory pathway, and the Zeocin resistance gene for positive selection of the recombinant clones in E. coli and P. pastoris. One of the hybrid genes, pNLVg001, differed from HuIFNa1 sequence by only five amino acids. Four consecutive rounds of in vitro site-directed mutagenesis were performed in order to back-mutate FIG. 1. Map of pPEPhIFNaD expression construct. The map shows the different features and relevant restriction sites of plasmid pPEPhIFNaD. FIG. 2. Complete DNA and amino acid sequence of HuIFNaD. (A) The complete amino acid sequence was translated from the HuIFNaD cDNA sequence obtained from GenBank Accession No. J00210. (B) The DNA sequence shown here was back-translated from the amino acid sequence with Pichia preferred codon usage. the hybrid interferon gene in plasmid pNLVg001 to HuIFNaD. The resulting plasmid, called pPEPhIFNaD, includes all the elements of the pPICZa vector as well as the HuIFNaD gene sequence (Fig. 1). DNA sequence of the gene insert in the pPEPhIFNaD was confirmed by DNA sequence analysis (Fig. 2). Plasmid pPEPhIFNaD was linearized with SacI restriction endonuclease prior to electroporation into the X-33 strain of P. pastoris. A SacI-linearized pPICZa vector with no gene insert was also used as a control. In this system, homologous recombination is generally expected to occur within the upstream 58 sequence of the AOX1 promoter region since only a small 38AOX1 transcription termination fragment is included in the pPICZa expression vector (Fig. 1). Zeocin-resistant transformants with at least one integrated copy of the expression cassette can be easily distinguished from an untransformed colony by performing colony PCR using the 58AOX1 and 38AOX1 PCR primers and comparing the resulting PCR product lengths. Colony PCR of positive transformants resulted in the visualization of a 1029-bp band and an approximately 2200-bp band (Fig. 3). These bands represent genomic amplification products that include the HuIFNaD gene insert or the intact endogenous copy of the AOX1 gene, respectively. The presence of an approximately 650-bp band on a 6% PAGE rather than the 1029-bp band in the pPICZa 384 LIU, TA, AND VILLARETE amino acids and 19.4 kDa, respectively. These theoretical protein molecular weight values were slightly higher than the estimations based upon the prestained molecular weight markers (Figs. 4A and 4B, lane 1). In our hands, the prestained molecular weight markers have often demonstrated faster migration rates than predicted. The dyes linked to the molecular weight standards most likely account for this discrepancy. Using an anti-human IFNa-specific monoclonal antibody, Western blot analysis after electrotransfer of a duplicate gel from Fig. 4A confirmed that the putative rHuIFNaD protein band was indeed an interferon molecule and has the same electrophoretic mobility as the rHuIFNaD standard reference molecule (Fig. 4B, lanes 6 and 2, respectively). Two additional bands migrating at slightly slower rates than the major band also reacted with the anti-human IFNa antibody. Figure 4B also shows that immunoblotting did not detect any degradation products of the secreted rHuIFNaD. This indicates FIG. 3. Polyacrylamide (6% TBE) gel of PCR products from recombinant Pichia colonies. Lane 1, 100- to 2000-bp DNA ladder; lanes 2–4, colonies from pPEPhIFNaD-transformed yeast cells; lanes 5–7, colonies from pPICZa vector control-transformed yeast cells; lane 8, a PCR-positive control using the pPEPhIFNaD plasmid as the template. vector control transformed colonies confirmed the absence of the gene insert which is 501 bp long (Fig. 3). Note that the approximately 650-bp PCR product derived from the pPICZa transformants is actually 593 bp and migrates as such on an agarose gel (data not shown). This anomalous migration pattern may be due in part to the variance in resolving power between polyacrylamide and agarose gels (12). In these experiments, eight of the pPEPhIFNaD-positive colonies and one control pPICZa colony were subsequently used for smallscale expression studies. The selected recombinant colonies were induced for expression by growth in BMMY, a methanol-containing medium. The supernatants were harvested after 3 days of induction. Figure 4A shows a Coomassie-stained SDS–PAGE gel with protein samples that are representative of recombinant pPICZa control and HuIFNaD day 3 supernatants (Fig. 4A, lanes 5 and 6, respectively). The HuIFNaD supernatant shows the presence of a protein band that migrates slightly above the rHuIFNaA standard and below the roIFNt standard. This band is absent in the pPICZa control supernatant. This result corresponds to the amino acid lengths and theoretical molecular weights of these proteins. The amino acid lengths and theoretical molecular weights for rHuIFNaA, roIFNt, and rHuIFNaD are 165 amino acids and 19.2 kDa, 172 amino acids and 19.9 kDa, and 166 FIG. 4. SDS–PAGE gel and Western blot analysis of rHuIFNaD protein expressed in P. pastoris. Samples were run on duplicate gels. One gel was Coomassie stained (A) and the other was transblotted and probed with an anti-HuIFNa-specific antibody (B). Lanes 1–4, prestained molecular weight markers (pMWM), rHuIFNaD standard, rHuIFN-aA standard, and roIFNt, respectively; lanes 5 and 6, the harvested day 3 supernatant of the expression cultures from a representative pPICZa vector control transformant and pPEPhIFNaD, respectively; lanes 7–10, fractions collected following size-exclusion purification. HuIFNa1 EXPRESSED IN Pichia pastoris that rHuIFNaD is resistant to P. pastoris proteinases during expression under shake flask conditions. Note that the presence of additional nonimmunoreactive bands in the rHuIFNaD standard lane (Fig. 4A, lane 2) are most likely BSA protein bands, since the rHuIFNaD standard is in a buffered solution with 0.1% BSA (Fig. 4B, lane 2). All eight rHuIFNaD colonies tested secreted one major band and two minor bands that reacted with the HuIFNa antibody (data not shown). The Coomassie-stained SDS–PAGE gel, shown on Fig. 4A, also includes purified fractions of a rHuIFNaD day 3 supernatant from the best producing transformant, following one round of molecular sieve chromatography (Fig. 4A, lanes 7 to 10). Most of the highmolecular-weight proteins in fraction 3 and fraction 4 were no longer detectable by Coomassie staining (Fig. 4A) nor silver staining (data not shown) of SDS–PAGE gels (N 5 4). As observed with the unpurified supernatants, Western blot analysis of purified fraction 3 verified three immunoreactive proteins—one major band and two other bands with slightly higher migration patterns (Fig. 4B, lane 9). Although fraction 4 showed only two bands that were detected by Coomassie staining and immunoblotting (Figs. 4A and 4B, lane 10), the third band was clearly evident in silver-stained SDS– PAGE gels (data not shown). Densitometric analysis demonstrated that the combined band density ratios of the three HuIFNa-specific bands amounted to $99% of the total protein in the sample fraction 3 and sample fraction 4. The band density ratio of just the major band in fraction 3 and fraction 4 was about 62 and 89%, respectively. N-terminal microsequencing analysis of the two higher molecular weight bands revealed an additional 9- or 11-amino-acid N-terminal extension from the a-mating factor signal sequence, whereas the major band revealed correct processing of the signal sequence and showed an exact match at amino acid positions 2 through 20 with that of the native HuIFNaD sequence (Fig. 5). As expected, the highly reactive cysteine at the N-terminus could not be confirmed because the sample was neither reduced nor alkylated prior to N-terminal sequencing. The cysteine at amino acid FIG. 5. N-terminal sequence of the three anti-human IFNa antibody-immunoreactive bands. Proteins from purified fraction 3 were separated on a 14% SDS–PAGE gel, transferred onto a PVDF membrane, and Coomassie stained. The bands were excised and the sequence at the N-terminal end was obtained by Edman degradation using an automated Perkin–Elmer/ABI-494 Procise sequenator. 385 FIG. 6. IEF gel analysis of rHuIFNaD purification fraction 3. This Coomassie-stained IEF gel gives an estimation of the pI of the major and minor bands from the size exclusion gel purification fraction 3. Lanes 1 and 3, pI standards and fraction 3, respectively; lane 2, an empty lane. The theoretical pI of HuIFNaD is 5.077. The band indicated with the black arrow is the major band and has an estimated pI of 5.06 based upon calculations using the AlphaImager 2000 software (Alpha Innotech, San Leandro, CA). position 1 was indirectly verified by DNA sequencing of the HuIFNaD gene integrated in the genome of the recombinant HuIFNaD P. pastoris clones used for the expression studies. The IEF gel shown in Fig. 6 depicts the relative band mobilities representing the pI of the three bands present in fraction 3. The theoretical pI of rHuIFNaD is 5.077. Computer-generated calculations of the pI of the major band in purified fraction 3, based upon the band mobility on the IEF gels, was estimated to be about 5.06. The amount of biologically active rHuIFNaD secreted by P. pastoris was determined by measuring the antiviral activity in the purified fractions using a standard cytopathic protection assay. Typically, the total protein concentration in these purified fractions ranged from approximately 0.3 to 0.7 mg/ml per 20 ml of shake culture. Measuring the antiviral activity of the rHuIFNaD in the same fractions demonstrated that the rHuIFNaD purified from P. pastoris supernatants had full biological activity with specific activity of about 1.7 3 108 U/ mg. This is higher than the antiviral activity determined for the commercially available rHuIFNaD with specific activities ranging from 5.0 to 7.5 3 107 antiviral units/mg (PBL, New Brunswick, NJ). Clearly, these results further substantiate our findings that the rHuIFNaD secreted by P. pastoris is similar structurally as well as functionally to the commercially available rHuIFNaD. 386 LIU, TA, AND VILLARETE DISCUSSION One of the major advantages of expressing rHuIFNaD as a secreted protein in P. pastoris using the afactor prepro signal is that P. pastoris secretes very low levels of contaminating native proteins in the culture medium. This characteristic, combined with the use of the minimal Pichia growth medium with low protein content, led to the efficient secretion of a biologically active rHuIFNaD that constitutes the vast majority of the total protein in the medium. Although previous work has demonstrated the secretion of biologically active rHuIFNaD from S. cerevisiae using the a-factor prepro signal, the antiviral activity in the S. cerevisiae recombinant culture medium from shake flask studies was 100-fold less compared to the activity of the rHuIFNaD P. pastoris recombinant culture medium described in this report (105 and 107 U/ml, respectively) (7). This discrepancy may in part be due to the difference in the relative strength of the promoters used in the two systems. In this report, rHuIFNaD expression was driven by the methanol inducible AOX1 promoter, whereas the rHuIFNaD in the S. cerevisiae system used the a-factor promoter. In yeast, the a-factor promoter normally drives the expression of the a-factor or apheromone precursor proteins that are posttranslationally cleaved into oligopeptides of 12–13 amino acid residues. Although the a-factor promoter is responsible for driving the expression of the a-factor, which is one of the limited number of proteins found in the S. cerevisiae culture medium, the AOX1 promoter is capable of driving the expression of the AOX1 protein or other recombinant proteins to very high levels, typically resulting in $30% of the total soluble protein in P. pastoris grown with methanol as the sole carbon source (10). This highlevel expression results in a straightforward and simple procedure for further purification using techniques such as molecular sieve chromatography. In this report, one round of molecular sieve chromatography essentially eliminated the non-HuIFNaD higher molecular weight contaminating proteins leaving behind a rHuIFNaD that is $99% pure with higher antiviral activity than the commercially available molecule. The two rHuIFNaD minor bands that migrated slightly higher than the major band are rHuIFNaD molecules with either a 9 amino acid or a 11-amino-acid N-terminal extension from the a-factor signal sequence. This result is consistent with previous work on heterologous proteins secreted by S. cerevisiae that used the a-factor mating prepro signal sequence (7, 13). These particular studies demonstrated the presence of additional bands representing the protein of interest with either an extra 11- or a 14-amino-acid N-terminal extension derived from the signal sequence. Further studies are ongoing to produce a rHuIFNaD preparation without the additional immunoreactive minor bands. It is worthy to note, however, that the presence of these two additional bands did not seem to negatively affect the biological activity of the rHuIFNaD preparations from P. pastoris. The small-scale expression experiments described in this report used 10-ml shake cultures with an initial OD600 of 2. These experiments produced a total of 0.3 to 0.7 mg of purified, biologically active rHuIFNaD per 20 ml of culture supernatant. However, the ability of P. pastoris to grow to high densities in a fermentor creates an even greater potential to produce large amounts of relatively pure and functionally active rHuIFNaD that can easily be recovered from the culture supernatant fractions. Large-scale fermentation conditions have already been developed for the P. pastoris expression system for a number of different proteins, including roIFNt. These conditions can conceivably serve as the starting point for optimizing the fermentation process for other interferon proteins, such as rHuIFNaD. ACKNOWLEDGMENTS We thank Jackie Campos for her technical support. Our gratitude also extends to Shigeo Yagi at the State Viral and Rickettsial Laboratory at Berkeley, California, for the antiviral work, and Dick Winant and Martin Pentony at the Stanford PAN Facility for the N-terminal sequencing and DNA sequencing work, respectively. We thank Michael Meagher from the University of Nebraska at Lincoln for his constructive review of the manuscript. REFERENCES 1. Cheng, M. L., Wu, Y. Y., and Huang, K. F. (1997) The clinical study on treatment of hepatic fibrosis of hepatitis B by IFNalpha 1 and Chinese medical preparation. Chung Kuo Chung His I Chieh Ho Tsa Chih 17(8), 453–455. 2. Shepherd, F. A., Beaulieu, R., Gelmon, K., Thuot, C. A., Sawka, C., Read, S., and Singer, J. (1998) Prospective randomized trial of two dose levels of interferon alfa with zidovudine for the treatment of Kaposi’s sarcoma associated with human immunodeficiency virus infection: A Canadian HIV Clinical Trials Network Study. J. Clin. Oncol. 16(5), 1736–1742. 3. Jin, X. Y. (1992) A clinical investigation of rHuIFN alpha-1 in the treatment of herpes simplex virus keratitis. Chung Hua Yen Ko Tsa Chih 28(3), 134–137. 4. Akiyama, K., Sugii, S., and Hirota, Y. (1993) A clinical trial of recombinant bovine interferon alpha 1 for the control of bovine respiratory disease in calves. J. Vet. Med. Sci. 55(3), 449–452. 5. Babiuk, L. A., Lawman, M. J., and Gifford, G. A. (1987) Use of recombinant bovine alpha interferon in reducing respiratory disease induced by bovine herpes virus type 1. Antimicrob. Agents Chemother. 31(5), 752–757. 6. De Maeyer, E., Skup, D., Prasad, K. S., De Maeyer-Guignard, J., Williams, B., Meacock, P., Sharpe, G., Pioli, D., Hennam, J., Schuch, W., and Atherton, K. (1982) Expression of a chemically synthesized human alpha 1 interferon gene. Proc. Natl. Acad. Sci. USA 79(14), 4256–4259. 7. Singh, A., Lugovoy, J. M., Kohr, W. J., and Perry, L. J. (1984) Synthesis, secretion and processing of alpha-factor-interferon fusion proteins in yeast. Nucleic Acids Res. 12(23), 8927–8938. 8. Breitling, R., Gerlach, D., Hartmann, M., and Behnke, D. (1989) Secretory expression in Escherichia coli and Bacillus subtilis of HuIFNa1 EXPRESSED IN Pichia pastoris human interferon alpha genes directed by staphylokinase signals. Mol. Gen. Genet. 217(2–3), 384–391. 9. Riesenberg, D., Menzel, K., Schulz, V., Schumann, K., Veith, G., Zuber, G., and Knorre, W. A. (1990) High cell density fermentation of recombinant Escherichia coli expressing human interferon alpha 1. Appl. Microbiol. Biotechnol. 34(1), 77–82. 10. Cregg, J. M., Vedvick, T. S., and Raschke, W. C. (1993) Recent advances in the expression of foreign genes in Pichia pastoris. Bio/Technology 11, 905–910. 11. Pontzer, C. H., and Johnson, H. M. (1995) Measurement of interferons. Methods Neurosci. 24, 3–9. 387 12. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) “Molecular Cloning: A Laboratory Manual,” Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 13. Brake, A. J., Merryweather, J. P., Coit, D. G., Heberlein, U. A., Masiarz, F. R., Mullenbach, G. T., Urdea, M. S., Valenzuela, P., and Barr, P. J. (1984) a-Factor-directed synthesis and secretion of mature foreign proteins in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 81, 4642–4646. 14. Van Heeke, G., Ott, T., Strauss, A., Ammaturo, D., and Bazer, F. W. (1996) High yield expression and secretion of the ovine pregnancy recognition hormone interferon-t by Pichia pastoris. J. Interferon Cytokine Res. 16, 119–126.