CA2090969C - Production of insulin-like growth factor-1 in methylotrophic yeast cells - Google Patents

Abstract

Insulin-like growth factor-1 (IGF-1), a naturally occurring, relatively short, single chain polypeptide, is prepared by grow-ing methylotrophic yeast transformants containing in their genome at least one copy of DNA encoding IGF-1, in operational lin-kage with DNA encoding a signal sequence, which is effective for directing secretion of proteins from the host cells and which also includes the proteolytic processing syte lys-arg and may include one or more glu-ala sequences. In preferred embodiments the signal sequence is the S. cerevisiae alpha mating factor pre-pro sequence.
Expression of both the DNA encoding IGF-1 and the pre-pro signal sequence are regulated by a promoter region derived from a methanol responsive gene of a methylotrophic yeast. DNA constructs and recombinant methylotrophic yeast strains used for the expression and secretion of IGF-1 are also pro-vided. For preferred embodiments, protease deficient Pichia pastoris strains are provided.

Description

i~4'(~::y'04363 PCT/US91 /06452 -,- 20909b9 PRODUCTION OF INSULIN-LIRE GROWTH
FACTOR-1 IN HETHYLOTROPHIC YF;AST CE1.L8 FIELD OF THE INVENTION
The invention described herein relates to processes for producing authentically folded biologically active insulin-like growth factor-1 (IGF-1) using methylotrophic yeast host cells, such as Pichia gastoris (P. pastoris).
The invention further relates to the methylotrophic yeast that produce IGF-1, to DNA fragments for preparing 1~ expression vectors that include DNA encoding IGF-1 and that are used for transforming the methylotrophic yeast cells, to the expression vectors, and to cultures containing the transformed yeast cells.
BACKGROUND OF THE INVENTION
Insulin-like growth factor-1 belongs to a heterogeneous family of peptides which share some of the biological and chemical properties of insulin, but which are antigenically distinct from insulin. Currently available experimental evidence suggests that IGF-1 promotes growth by mediating the effects of growth hormone. Thus, such processes as skeletal growth, cell replication and other growth related processes are A

WO 92/04363 PCT/US91/0645~,....
zu~u9s~

affected by IGF-1 levels. Physiological concentrations of IGF-1 have been shown to be influenced by such conditions as thyroid disease, diabetes and malnutrition (see Preece (1983) Horm. Blood, 4: 108). IGF-1 has also been shown to act synergistically with other growth factors, for example, in accelerating the healing of soft and mesenchymal tissue wounds (see Lynch et al. (1989) J.
Clin. Periodontol., 16: 545; and Lynch et al. (1987) Proc. Natl. Acad. Sci. USA, 84: 7696), and in enhancing the growth of mammalian cells. in serum-free tissue culture medium (see Burleigh et al. (1986) American Biotech. Lab., 4: 48).
Considering the many clinical and research applications of IGF-1, a ready supply of IGF-1 would be of great value to the medical and biotechnology fields.
Since isolation from natural sources is technically difficult, expensive, and time consuming, recent efforts have centered on the development of efficient recombinant methods for the production of IGF-1.
Among host cells that have been used for the production of heterologous proteins, ~ co ' and Saccharomyces cerevisiae (Baker's yeast) are probably the best characterized. Insulin-like growth factor-1 (IGF-1), which is a polypeptide of 70 amino acids with a molecular weight of 7648 daltons, is a single chain protein that has three intrachain disulfide bridges.
These disulfide bonds, along with numerous hydrogen bonds and hydrophilic interactions, maintain the compact tertiary structure of this molecule. E. coli, however, does not possess the ability to produce disulfide bonds in proteins, so that proteins, such as IGF-1, that include disulfide bonds, when cloned into and expressed in E. coli, frequently are not stable and tend to aggregate into inactive complexes. In addition, IGF-1 produced in ~ coli has to be extracted and treated with s oxidizing agents to produce the disulfide bonds. Upon reduction and reoxidation, IGF-1 refolds in a variety of ways, forming as many as 15 monomeric configurations (Meng et al. (1988) J. Chrom., 443: 183) because cell breakage and too rapid formation of disulfide linkages results in random disulfide bond formation. In order to produce biologically active IGF-1, the resulting mixture of 15 different forms of IGF-1 must be separated.
Consequently, the yield of purified product is very low (Grossgian (1985) Gene, 18: 199).
Furthermore, since E. coli is a prokaryote, in order to produce IGF-1 molecules which contain the authentic N-terminal glycine, and not the initiating methionine present on the primary translation product, it is necessary to express IGF-1 in E. coli as a fusion protein. Cleavage of mature IGF-1 from the initially produced fusion protein necessitates an additional step in the production process. Consequently, attempts to produce this peptide by recombinant means in _E. coli host expression systems results in a complex mixture of product forms which must be separated for further use (see, Grossgian (1985) Gene, 18: 199).
Eukaryotic host cells, including yeast cells, thus, are the host cells of choice for the expression of many eukaryotic proteins. Yeast host cells offer clear advantages over bacteria in the production of heterologous proteins, including their ability to properly process pre-pro-heterologous proteins and secrete heterologous proteins into the culture medium.
Secretion of proteins from cells is often superior to production of proteins in the cytoplasm because secreted products are obtained in a higher degree of initial purity and further purification of the secreted products is made easier by the absence of cellular debris. In addition, the secretory pathway of the cell and the WO 92/04363 ~ ~ ~ PCT/US91/06452 extracellular medium tend to be oxidizing environments which support disulfide bond formation necessary for proper folding of many proteins (Smith, et al. (1985) Science 229: 1219); whereas, the cytoplasm is a reducing environment in which disulfide bonds do not form. Thus, for production of sulfhydryl-rich proteins that rely on disulfide bonds to maintain the correct tertiary structures, there is a compelling need to develop eukaryotic hosts capable of secreting such proteins into the culture medium. Therefore, production of sulfhydryl-rich proteins, such as IGF-1, that contain appropriately formed disulfide bonds, can best be achieved by transit through the secretory pathway.
IGF-1 has been cloned into and expressed using S.
cerevisiae host cells by introducing DNA encoding IGF-1 on autonomously replicating extrachromosomal elements.
Gellerfors et al. ((1989) J. Biol. Chem., 264: 11444-11449) describes the production of IGF-1 in S. cerevisiae under the control of the S. cerevisiae actin promoter.
The IGF-1 product is encoded by autonomously replicating plasmid-borne DNA. In a similar study, Bayne et al.
((1988) Gene 66: 235-244) describes the production of IGF-1 in S. cerevisiae under the control of the S.
cerevisiae alpha mating factor promoter. The latter, however, reports yields of IGF-1 of only about 2 mg of IGF-1 per liter of fermentation broth.
In view of this low yield and the problems generally encountered with up-scaling the production of heterologous proteins in autonomous plasmid-based yeast systems, such as loss of selection for plasmid maintenance and problems concerning plasmid distribution, copy number and stability in fermentors operated at high cell density, there is a need to develop more efficient means for producing large quantities of biologically active IGF-1.

pC?IUS 91/06452 _ -5-IpEA~US 30 JUL1992 Therefore, it is an object of this invention to provide host cells and expression vectors that stably express IGF-1 and that secrete high concentrations of biologically active IGF-1.
It is another object of this invention to provide an expression system for the production of biologically active IGF-1 that, not only secretes high concentrations of biologically active IGF-1, but that can be readily scaled up to produce large quantities of such IGF-1.
SUMMARY OF THE INVENTION
. Expression systems and methods using the expression systems for the production of biologically active insulin-like growth factor-1 (IGF-1) using methylotrophic yeast host cells are provided. The methods of production are readily scaled up from shake-flask cultures to large scale fermentors with no loss in IGF-1 productivity and without the need for making major changes in the fermentation conditions used for the growth of the transformed strains. Methods for isolation and purification of the IGF-1 product are also provided.
The expression systems and methods provided herein avoid the problems encountered with heterologous protein expression in ,~ cerevisiae in which high level expression can only be achieved by the introduction of 25. multicopy plasmids into the host cells.
The,expression system described herein uses .'methylotrvphic yeast host cells, such as for example, P.
pastoris, for the expression of IGF-1. Key features of the system include the ability to stably integrate and express multiple copies of the DNA encoding IGF-1 and the DNA encoding the signals that direct secretion and processing of the IGF-1 and the ability to properly process mature IGF-1 from the expressed precursor form IGF-1 and to secrete the resulting mature IGF-1 product.
Another feature of the system resides in selection g~ggnTUTE SHEET

of the promoter that has been used to control expression of the DNA encoding IGF-1. The promoter, which is derived from a methanol-responsive gene, such as AOX1, of a methylotrophic yeast, is tightly regulated and provides for high-level regulated expression of genes placed under its control (see, eg., European Patent Application No. 85113737.2, published June 4, 1986, under No. 0 183 071, now issued in the United States as U.S. Patent No. 4,855,231).
Expression and secretion of high levels of IGF-1 peptide has been accomplished by transforming a methylotrophic yeast host with a DNA construct that contains at least one copy, but may contain as many as six or more copies, of DNA
encoding an IGF-1 peptide in which the DNA is operably linked with DNA encoding a signal sequence that is effective for directing the processing and secretion of the IGF-1 peptide product. The DNA construct also includes a promoter region, which directs expression of the DNA encoding the signal sequence and IGF-1 peptide, and a transcription terminator functional in a methylotrophic yeast.
The DNA construct provided here, also includes sequences of nucleotides that have sufficient homology with a target gene in the methylotrophic yeast host cell genome to effect stable integration. Integration takes place by addition or replacement at the site of the target gene. Alternatively, the DNA construct is provided as part of a circular plasmid that integrates by addition at a site of homology between the host and the plasmid.
Thus one aspect of the invention provides a DNA
construct, comprising: (a) DNA encoding an insulin-like growth factor-1 (IGF-1) ; (b) a promoter region from a ., '.q. r,M,":
p ': ' i~.
...iY' ""~'~

-----. 6 a methanol-responsive gene of a methylotrophic yeast operably linked to said DNA that encodes IGF-1, said methylotrophic yeast is a member of the genus Pichia or Hansenula; (c) DNA
encoding the Saccharomyces cerevisiae Oc-mating factor (ocMF) pre-pro signal sequence, wherein the DNA that encodes said signal sequence is operably linked via DNA encoding one or more processing sites lys-arg that are recognised and cleaved by methylotrophic yeast cell proteases to the DNA that encodes IGF-1; and (d) a transcription terminator functional in a methylotrophic yeast host cell operably linked to the DNA that encodes IGF-l; wherein the methylotrophic yeast is a species that utilises methanol as a sole carbon source and is a member of the genera selected from the group consisting of Candida, Hansenula, Pichia, and Torulopsis.
In accordance with other embodiments, expression vectors containing the DNA construct, which includes at least one copy of an expression cassette, are provided.
According to another aspect, there are provided ..r~;

20909b9 P~TtUS 91 / 0 6 4 5 2 _, _ methylotrophic yeast cells containing in their genome at least one copy of the above-described DNA fragment. In preferred embodiments, the cells are deficient in one or more proteases that are not essential for proper cleavage of the precursor form of IGF-1.
In preferred embodiments, the host cell is P.
~aastoris, the promoter is the AOX1 promoter, and the signal sequence is the ;~ cerevisiae alpha-mating factor (aMF) pre-pro sequence, which includes a processing sequence of the formula lys-arg-(glu-ala)x in which x is - preferably between 0 and 3. Methylotrophic yeast cell-s in which at least one copy of this DNA construct has been introduced efficiently produce and secrete biologically active IGF-1 peptides into the medium. In preferred embodiments, IGF-1 has been very efficiently produced in and secreted from the methylotrophic yeast P. pastoris.
The polypeptide product produced by the methylotrophic host cells is secreted into the culture medium at high concentrations; the level of IGF-1 peptides that are secreted is many times the concentrations that are reported for the secretion of IGF-1 from ~ cerevisiae (see, e.g_, Bayne a a (1988) Gene, 66: 235-244). The IGF-1 peptides are produced by methylotrophic yeast cells that contain one or more 25- integrated copies of DNA encoding IGF-1 peptides operably ~~linked w-ith DNA encoding a signal sequence, such as the ~~,5,~ cerevisiae a-mating factor (aMF) pre-pro sequence that includes at least one site, such as lys-arg, that is sufficient for proper proteolytic cleavage of mature IGF-1 from the pre-pro-protein, under the regulation of a promoter region of a methanol-responsive gene of a methylotrophic yeast. In addition, the DNA encoding the signal sequence may encode one or more glu-ala sequences, which also serve to direct proper .. 35 processing. The number of glu-ala sequences is s ~iUBSTITUfE SHEEN

_8-preferably from 0 to 3, but may include more copies. The number- of copies is limited by the ability of the selected host to properly process the pre-pro-protein.
According to a still further embodiment, there is provided a process for producing IGF-1 peptides by growing methylotrophic yeast transformants containing in their genome at least one copy of a DNA sequence operably encoding an IGF-1 peptide, operably associated with DNA
encoding the S. cerevisiae aMF pre-pro sequence. This pre-pro sequence includes a lys-arg processing site and may also include from 0 to 3 glu-ala spacer segments.
Transcription of the DNA encoding the signal and the DNA
encoding the IGF-1 peptide is under the regulation of a promoter region of a methanol-responsive gene of a methylotrophic yeast. Under conditions that induce the promoter, transcription of the DNA is induced, thereby resulting in expression and processing of the precursor IGF-1, and, ultimately, secretion of IGF-1 peptides into the culture medium.
Cultures of viable methylotrophic yeast cells capable of producing IGF-1 peptides are also provided.
In preferred embodiments, the yeast host cells that express and secrete IGF-1 have been modified by disruption of one or more host cell genes that directly or indirectly influence host cell protease expression.
This disruption results in a reduction of some protease activities of the host cell, for example, proteinase A
and carboxypeptidase Y activities.
A preferred process for recovering and purifying IGF-1 peptides from P. pastoris fermentation broth is described in copending Canadian Application No.
2,098,183. The preferred form of IGF-1 is a polypeptide product which exhibits substantially the same biological activities as natural insulin-like growth factor-1, as measured in recognized bioassays (see, e.Q., -,7718-6 2090969 Takano et al. (1976) Acta Endocr. 82: 449-459; Schoenie et al.
(1982) Nature 296: 252-253; Copeland et al. (1983) Am. J.
Primatology 5: 161-169 and Buul-Offers et al. (1986) Ped.Res.
20: (825-827) having substantially the same amino acid sequence as native IGF-l, Sequence ID No. 1, in the sequence listings. It is understood that polypeptides having some variation in amino acid sequences such as lacking one or more amino acids, or containing additional amino acids, or having some substituted amino acids are within scope of the invention, provided that these peptides exhibit the functional activity of IGF-1 according to known bioassays. Such IGF-1 may include, not only authentically folded IGF-l, but also the misfolded and multimeric forms of IGF-1 that exhibit the ability to detestably bind to IGF-1 receptors and that demonstrate at least one bioactivity associated with IGF-1. In addition, biologically inactive forms may be converted to active forms by reduction and oxidation.
BRIEF DESCRIPTION OF THE FIGURES
Figure 1 provides the nucleotide sequence of a synthetic insulin-like growth factor gene.
Figure 2 is a restriction map of plasmid pIGF201.
Figure 3 is a restriction map of plasmid pIGF202.
Figure 4 is a restriction map of plasmid pIGF204.
Figure 5 is a restriction map of plasmid pIGF206.
Figure 6 is a restriction map of plasmid pA0815.

''7718-6 DETAILED DESCRIPTION OF THE INVENTION
Definitions:
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as is 5 commonly understood by one of skill in the art to which this invention belongs.
As used herein, IGF-1 or an IGF-1 peptide is intended to include all the allelic variations of IGF-1. Moreover derivatives obtained by simple modification of the amino acid 10 sequence of the naturally occurring product, such as by way of site-directed mutagenesis or other standard procedures, are included within the scope of the present invention. Forms of IGF-1 that exhibit similar biological activities to naturally occurring IGF-1 are also encompassed by the present invention.
IGF-1 is intended to encompass all forms of IGF-1 present in the culture medium. Such IGF-1 may be mixture of several different three-dimensional forms. In addition to authentic IGF-1, which is the intact, the monomeric, correctly folded material, other forms may be present in the extracellular medium of the host cells that express IGF-l, such as misfolded proteins, multimeric forms such as dimers, and trimers. It is intended that the IGF-1 peptide, as used herein, includes any peptide that has the ability to bind to IGF-1 receptors and that exhibits at least one activity, such as the ability to promote cell proliferation or cell growth in a standard activity assay, that is associated with the biological activity of IGF-1 (see, eg., Takano et al. (1976) Acta Endocr. 82:
449-459; Schoenie et al. (1982) Nature 296: 252-253; Copeland et al. (1983) Am. J. Primatology 5: 161-169 and Buul-Offers et al. (1986) Ped.Res. 20: 825-827).

10a 2090969 As used herein, an IGF-1 peptide that has biological activity is a peptide that specifically binds to IGF-1 receptors, as detected by a standard assay, and that exhibits a biological activity associated with IGF-1 as measured in a recognized assay. Such assays are known to those of skill in this art. As used herein, authentic IGF-1 refers to IGF-1 that includes disulfide bonds as they exist in naturally-2090969 - ~ ~ -p~~S 91/Ob452___ -11~P~~~S 3 0 J U l 1992 occurring IGF-1. Authentic IGF-1 may be efficiently recovered and purified from P. pastoris fermentation broth.
As used herein, mature IGF-1 refers to processed IGF-1 from which the signal and processing sequences have been cleaved. Mature IGF-1 includes authentic IGF-1 and any other forms of processed IGF-1 that are secreted into the medium. IGF-1 is intended to encompass peptides that have the ability to bind to IGF-1 receptors and to promote cell growth-or proliferation as measured by any means known to those of skill in the art.
As used herein, pre-pro-IGF-1 refers to a polypeptide that includes a leader or signal sequence that effects secretion of mature IGF-1 into the extracellular space of said host and one or more processing signals that direct processing of the pre-pro-IGF-1 to produce the mature IGF-1. As used herein, secreted IGF-1 ref8rs to processed IGF-1 that does not include the signal or leader sequence. Processed IGF-1 or processed protein refer to IGF-1 or protein from which the leader signal has been cleaved.
As used herein, a signal or leader sequence, which expressions are used interchangeably, refers to a sequence of amino acids that effects transport of a 25- linked polypeptide through the cell membrane. A signal ~equence_refers to a sequence of hydrophobic amino acids _. : _ .at the amino terminus of the protein to which it is linked. DNA encoding a signal sequence is located downstream (3' in the direction of transcription) from the ATG start codon and upstream (5') from the DNA that encodes the structural gene. In addition, the signal sequence includes one or more processing sites, sequences of amino acids that are recognized by one or more host . cell proteases, interposed between the signal sequence and the protein. Cleavage by a host cell protease s suBS~nTU~ SNEET

~1/06452,_ -12-ipEAIUS 3 0 JUL1992 effects removal of the signal sequence. The signal sequence, processing sites and protein are referred to as a pre-pro-protein.
The signal sequences and processing sites contempla-ted for use herein are those that effect transport of IGF-1 through the cellular membrane of a methylotrophic yeast host cell, such as P. pastoris. In preferred embodiments the signal sequence and processing sites, lys-arg and (glu-ala)x, in which x is an integer, preferably between 0 and 3, are derived from the S. cerevisiae aMF gene. Any signal sequence and processing site known to those of skill in the art that are effective for secreting mature IGF-1 into the extracellular space of a methylotrophic host may be used. In preferred embodiments, the signal sequences are selected so that fermentation cultures of methylotrophic host cells that have a cell density of about 300 g/L
secrete at least about 9 mg/L of authentic IGF-1 into the broth. To select suitable signal sequences, DNA constructs encoding a signal sequence and processing sites) linked to DNA encoding IGF-1 may be prepared and introduced into the genome of a methylotrophic host by methods known to those of skill in the art or described herein and tested for their ability to direct secretion of mature IGF-1 into the medium. Any signal sequence that leads to of secretion 25- relatively high concentrations, preferably more than '~':lOmg/ml 9f mature IGF-1, is contemplated for use herein.
As used herein, heterologous DNA includes DNA that does not occur naturally as part of the genome in which it is present or which is found in a location or locations in the genome that differs from that in which it occurs in nature. Heterologous DNA is generally not endogenous to the cell into which it is introduced, but has been obtained from another cell. Generally, although not necessarily, such DNA encodes RNA and proteins that are not normally produced by the cell in which the DNA or RNA is b SUBSTITUTE SHEE"~' 2090969 ' 9I/06~+52 -1=- ~pEA~US 30 JUL1992 expressed. Heterologous DNA may also be referred to as foreign DNA. Any DNA that one of skill in the art would recognize or consider as heterologous or foreign to the cell in which is expressed is herein encompassed by heterologous DNA. Examples of heterologous DNA include, but are not limited to, DNA that encodes IGF-1, transcriptional and translational regulatory sequences, and selectable or traceable marker proteins, such as a protein that confers drug resistance.
As used herein, expression cassette refers to a DNA
construct that includes sequences functional for both the expression and the secretion of IGF-1. Accordingly, an expression cassette includes DNA encoding a promoter region, DNA encoding a transcription terminator region, and sequences sufficient for translation, secretion and proper processing of the expressed peptide. In addition, in preferred embodiments, the expression cassette is on a fragment that includes sequences at 5' and 3' ends that are homologous to a target locus in the host cell genome, where-by, upon introduction into the host cell, the expression cassette is stably integrated into the host cell genome.
As used herein, the term DNA construct embraces expression cassettes and also includes DNA fragments that include more than one expression cassette.
25- As used herein, the term operative linkage or ''- ~apexabl~associated -refers to the relationship among .elements of a DNA construct in which the elements are arranged whereby regulatory sequences of nucleotides that are part of the construct directly or indirectly control expression of the DNA in the construct, including DNA
encoding a protein or a peptide.
As used herein, the term "a DNA fragment operably encoding IGF-1 peptides" includes DNA fragments encoding IGF-1 or any other "IGF-1 peptide" as defined hereinabove. DNA encoding IGF-1 is known in the art and ,, __ p~'t'~tl~ :1 l 0 6 4 5 2 _ .
-14-1PEA~us 3 0 ~U~ X992 may be obtained by chemical synthesis or by transcription of messenger RNA (mRNA) corresponding to IGF-1 into complementary DNA (cDNA) and converting the latter into a double stranded cDNA. Chemical synthesis of a gene for human IGF-1 is, for example, disclosed by Niwa et a (1986) Annals of the NY Academy of Science, 469: 31-52, and Buell et a (1985) Nucleic Acids Research, 13: 1923-1938. The requisite DNA sequence can also be removed, for example, by restriction enzyme digestion of known vectors harboring the IGF-1' gene. Examples of such - vectors and the means for their preparation are well known to those of skill in the art. See, e.a., Niwa a~ (1986) Annals of the NY Academy of Science, 469: 31-52, and Buell et al. (1985) Nucleic Acids Research, 13:
1923-1938. The nucleotide sequence of a presently preferred IGF-1 gene is illustrated in FIG. 1 and is further elucidated in the examples.
As used herein, the term expression vector is intended to include vectors capable of expressing DNA
that are in operational association with other sequences capable of effecting their expression, such as promoter sequences, in a selected host cell. In general, expression vectors usually used in recombinant DNA
technology are often in the form of "plasmids", which are 25- circular, double-stranded DNA loops, extrachromosomal ''- ale~nents-.
-As used herein, the terms "vector" and "plasmid" are used interchangeably and are not intended to be limited, but to include any expression vectors or means that permit heterologous DNA to be expressed in a particular host cell.
As used herein, the term "culture" means a propagation of cells in a medium conducive to their growth, and all sub-cultures thereof. The term "subculture" refers to a culture of cells grown from ;;
SUBSTITUTE SHEET

20909b9 . ..
PGTI~IS 91 / Q 6 4 5 ~ .

cells of another culture (source culture), or any subculture of the source culture, regardless of the number of times subculturing has been performed between the subculture of interest and the source culture.
The amino acids which occur in the various sequences of amino acid set forth in the specification have their usual, three- and one-letter abbreviations, routinely used in the art:
Amino Acid Abbreviation L-Alanine ~ Ala A
L-Arginine Arg R
L-Asparagine Asn N
L-Aspartic acid Asp D
L-Cysteine Cys C
L-Glutamine Gln Q
L-Glutamic Acid Glu E
L-Glycine Gly G
L-Histidine His H
L-Isoleucine Ile I
L-Leucine Leu L
L-Lysine _ Lys K
L-Methionine ~ Met M
L-Phenylalanine Phe F
L-Proline Pro P
L-Serine Ser S
L-Threonine Thr T
L-Tryptophan Trp W
L-Tyrosine Tyr Y
L-Valine Val V
Host c~lis.
Yeast species contemplated for use herein are methylotxophic yeast that are able to grow on methanol as ~a carbon source. Species intended for use herein have the biochemical pathways necessary for methanol utilization and fall into four genera, Candida, Hansenula, Pichia, and Torulopsis. A substantial amount is known about the molecular biology of members of the species Hansenula polymor~ and Pichia pastoris.
~ pastoris is the presently preferred yeast species. ~ pastoris is a known industrial yeast strain SUBSTITUTE SHEET

2090969 ~ -~ 91 /Ob 4 52--16- ~PE~~US 3 0 JUL 1992 that is capable of efficiently utilizing methanol as the sole carbon and energy source.
DNA constructs.
The DNA constructs and expression cassettes used for transforming methylotrophic yeast cells contain a methanol-responsive promoter from a methylotrophic yeast gene, DNA encoding IGF-1 (IGF-1 gene), DNA encoding a signal sequence in-reading frame with the gene, and and a transcription terminator functional in a l0 methylotrophic yeast. The-signal sequence may be any - such sequence that functions to direct the~secretion from and proper processing of IGF-1 in the yeast host cell.
The S. cerevisiae aMF pre-pro sequence is the preferred signal sequence. The aMF pre-pro sequence includes a DNA sequence encoding a processing sequence lys-arg, and (glu-ala)X spacer sequences in which x is 0, 1, 2 or 3.
The S. cerevisiae alpha-mating factor is a 13-residue peptide, secreted by cells of the "alpha" mating type, that acts on cells of the opposite "a" mating type to promote efficient conjugation between the two cell types and thereby formation of "a-alpha" diploid cells (Thorner et al. (1981) The Molecular Biology the Yeast Saccharomyces, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 143). The a-factor is synthesized as a 165 25- amino acid long precursor that contains an 83 residue 1'':leader and four a-factor peptide regions. Each region is .preceded by a short spacer peptides of sequences:lys-arg-glu-ala-asp-ala-glu-ala; lys-arg-(glu-ala)3 or lys-arg-(glu-ala)2. The leader and spacer includes the amino acid sequences for proteolytic processing and secretion (see, ea., Brake stet al., Proc. Natl. Acad. Sci. USA, 8~: 4642 (1984)).
There are a number of methanol-responsive genes in methylotrophic yeast. The expression of each is controlled by methanol-responsive regulatory regions, SUBSTITUTE SHEET

2090969 pC'1'~,ls 91106 452 also referred to as promoters. Any of such methanol-responsive promoters are suitable for use in the DNA
constructs. Examples of specific regulatory regions include, but are not limited to, the promoter for the primary alcohol oxidase gene from Pichia pastoris AOX1, the promoter for the secondary alcohol oxidase gene from P. pastoris AOX2, the promoter for the dihydroxyacetone synthase gene (DAS) from ~ pastoris, the promoter for the P40 gene from P. pastoris the promoter for the catalase gene from P. pastoris. Selection of suitable promoter regions and other regulatory regions is, in light of this disclosure, within the level of skill in the art.
The presently preferred promoter region employed to drive IGF-1 gene expression is derived from a methanol-regulated alcohol oxidase gene of P. pastoris. P.
pastoris expresses two functional alcohol oxidase genes:
alcohol oxidase I ~(AOX1) and alcohol oxidase II (AOX2) genes. The coding portions of the two AOX genes are closely homologous at both the DNA and the predicted amino acid sequence levels and share common restriction sites. The proteins expressed from the two genes have similar enzymatic properties but the promoter of the AOX1 gene is more efficient and provides for higher levels of 25. gene expression; therefore, its use is preferred for IGF
1'' 1 e~pres~ion. The AOXl gene, including its promoter, has .'been isolated and thoroughly characterized (see, Ellis et a . (1985) Mol. Cell. Biol. 5: 1111 and U.S. Patent No.
4,855,231).
The DNA construct that is introduced into methylotrophic yeast cells contains, in addition to a.methanol-responsive promoter of a methylotrophic yeast gene, DNA encoding IGF-l, DNA encoding an in-reading frame signal and processing sequence, and a transcription terminator functional in a methylotrophic ,.
su8s~ sHEEr ~~r~ ~ 91 ~ 0 6 ~ 5 z yeast. The transcription terminator functional in a methylotrophic yeast intended for use herein has either (a) a subsegment that encodes a polyadenylation signal and polyadenylation site in the transcript, and/or (b) a subsegment that provides a transcription termination signal for transcription from the promoter used in the expression cassette. The entire transcription terminator may be obtained from a protein-encoding gene that may be the same or different from the gene that is the source of the promoter.
The DNA construct may also contain a selectable marker gene. For this purpose, any selectable marker gene functional in methylotrophic yeast may be employed, including, but not limited to, any gene which confers a selectable phenotype upon methylotrophic yeast permitting such yeast cells to be identified and selectively grown from among a vast majority of untransformed cells.
Appropriate selectable marker systems, for example, would be an auxotrophic mutant P. pastoris host strain and a wild type biosynthetic gene which complements the host s defect. For example, the ~ cerevisiae or P. pastoris HIS4 gene may be used to complement His4~ P. pastoris strains, or the ~ cerevisiae A~tG4 gene or the P.
pastoris ARG4 gene may be employed to complement Arg4 25- mutants.
In addition, the DNA constructs may additionally i .include selectable marker genes that are functional in bacteria. Thus, any gene can be used which confers a phenotype on bacteria that allows transformed bacterial cells to be identified and selectively grown from among a vast majority of untransformed cells. This additional selectable marker permits such DNA to be introduced into and amplified in bacterial host cells, such as E. coli.
Suitable selectable marker genes include the ampicillin ,_ 35 resistance gene (Amp'), tetracycline resistance gene i SUBSTITUTE SHEET

~

J i . -2090969 p~'~~ 91106452__.
ipEaus 3 o JuL 1992 (Tcr), and the like.
The DNA construct may include sequences allowing for replication and selection in bacteria, especially E.
coli. In particular, the DNA construct may be inserted into a plasmid or it may be circularized to form a plasmid that includes sequences for replication and extrachromosomal maintenance in bacteria. In this way, large quantities of the DNA fragment can be produced by replication in bacteria. When such amplification is desired, the DNA coristruct~should include a bacterial ~ origin of replication to ensure the maintenance of the DNA construct from generation to generation of the bacteria. Exemplary bacterial origins of replication include the fl-ori, colisin, col E1, and others that are known to or that can be identified by those of skill in the art.
The DNA construct may include multiple copies of the DNA encoding IGF-l~operatively linked to the signal and processing signals, and other transcriptional regulatory signals, i-e., the DNA construct may include multiple copies of the expression cassette.
Transformation of m~thylotrophic y~ast host cells ~rith the DNA constructs.
Methods for transforming methylotrophic yeast, such 25. as, for example, ~ ~~astoris, as well as methods for culturing methylotrophic yeast cells containing in their .~genome a gene encoding a heterologous protein, are generally known in the art. The expression cassettes may be introduced into the methylotrophic yeast cells by any method known to those of skill in the art. Preferred methods include the spheroplast technique, described by Cregg gt ~ (1985) Mol. Cell. Biol. ~: 3376 and U.S.
Patent No. 4,879,231, and the whole-cell lithium chloride yeast transformation system (Ito et al. (1984) Aqric.
Biol. Chem. 48: 341), with modification necessary for SUBSTITUTE SHEE'~

_ . ., 2090969 P~T~I~ 91/6 x+52_ _ _ IPEA~US 3 0 JUL 1992 adaptation to methylotrophic yeast such as P. pastoris (See European Patent Application No. 312,934; also available as U.S. Patent No. 4,929,535).
If the yeast host is transformed with a linear DNA
fragment containing the IGF-1 gene and DNA encoding aMF
pre-pro sequences necessary for processing and secretion under the regulation of a promoter region of a P.
taastoris gene, the expression cassette is integrated into the host genome by any of the gene replacement techniques known in the art, such as by one-step gene replacement (see e.g., Rothstein (1983) Methods Enzymol. 101: 202;
Cregg et ,ate. (1987) Bio/TechnoloQV 5: 479; and U.S.
Patent No. 4,882,279) or by two-step gene replacement methods (see e.g., Scherer and Davis (1979) Proc. Natl.
Acad. Sci. USA, 76: 4951). The linear DNA fragment is directed to the target gene by flanking DNA sequences having sufficient homology with the target gene to effect integration of the~DNA fragment therein. One-step gene disruptions are usually successful if the DNA to be introduced has as little as 0.2 kb homology with the fragment locus of the target gene; it is preferable however to maximize the degree of homology for efficiency.
In preferred embodiments, multiple copies of these 25- expression cassettes are included on one DNA fragment, preferably in a head-to-tail orientation.
If the DNA fragment is included in a circular plasmid, the plasmid may be integrated by addition rather than gene disruption into the genome. One or more copies of the plasmid can be integrated at the same or different loci of the genome. Integration into the genome is facilitated by linearization of the plasmid by means of a suitable restriction endonuclease.
DNA fragments in the expression cassettes) are said to be "operably associated" with one another when the 8UBSTtTUTE SHEET

- . ..
2090969 p~,~~ g 1 /Ob 4 52_ _ ._ ~p~~us 3 0 ~u~~992 DNA sequence encoding the IGF-1 peptide is positioned and oriented functionally with respect to the promoter, the DNA sequence encoding the S. cerevisiae aMF pre-pro sequence and the transcription terminator. Thus, the polypeptide-encoding segment is transcribed under regulation of the promoter region into a transcript capable of providing, upon translation, the desired polypeptide. Because of the presence of the signal sequence, such as the aMF pre-pro sequence, the expressed IGF-1 product is secreted into the culture medium.
Appropriate reading frame positioning and orientation of the various segments of the expression cassette are within the knowledge of persons of skill in the art and are specifically described in the Examples.
Positive transformants may be isolated, identified and characterized by any method known to those of skill in the art, including Southern blot analysis (Maniatis et (1982) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, USA) to determine, for example, the site of DNA
integration; Northern blots (Maniatis et al. (1982) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, USA ; R.S. Zitomer and B.D. Hall (1976) J. Biol. Chem, 25- 25~,: 6320), for example, to confirm methanol-responsive ~I~Fri gene_expression; and product analysis to detect the -pre$ence of secreted IGF-1 peptides in the growth media.
The presently preferred host cells for transforma tion with multiple expression cassettes are P. pastoris cells having at least one mutation that can be complemented with a marker gene present on a transforming DNA- fr.~t.. Preferably His4- (GS115) or Arg4- (GS190) auxotrophic mutant P. pastoris strains are employed. The most preferred strains are additionally deficient in expression of one or more proteases that are not required b SUBSTITUTE SHEET

fir proper processing of the selected leader sequence.
'flee deficiency may be manifested as an absence of or a decrease in the activity of one or more proteases by virtue of point mutations, insertions, or deletions in genes that encode proteases or genes that directly or indirectly modulate expression of genes that encode proteases.
In preferred embodiments, DNA fragments containing one or more expression cassettes) are inserted into a vector containing a marker gene complementing the host's defect, and optionally containing additional sequences such as bacterial marker genes, yeast sequences which direct vector integration, and the like. Plasmids based on pBR322, such as pA0815, are preferred as vectors.
Insertion of one or more copies of the IGF-1 expression/secretion acassette into parent plasmid pAO815 produces plasmids such as pIGF201, pIGF202, pIGF204, pIGF206, and pIGF816.
Protease-deficient strains of P. ~astoris, known as pep4~ or pep4~ prb-1- strains, described below and in co-pending Cdn. Appln. No. 2,105,064, are among the preferred host cells. Protease-deficient P. ~astoris strains are generated by disruption of P. pastoris genes, such as the PEP4 gene or PRB-1 gene, that encode a protein that directly or indirectly affects the protease activity of the cell. This disruption is preferably accomplished by the insertion of a plasmid or DNA
fragment, unrelated to the constructs for expression of IGF-1, into the PEP4 gene or PBR-1 gene. This disruption results in a reduction of some of the protease activities in the cells, such as, for example, proteinase A, proteinase B and carboxypeptidase Y activities. This host cell strain preferably also contains a mutation that can be complemented by a marker gene carried on the expression cassette, as discussed below in the Examples.
A

20909 69 pCTf~J~ ~ 1 / 06 4 52_ _ ..
~pr~us 3 0 ~uL~992 The preferred host cells may be transformed with the expression vectors that are introduced into targeted specific sites in the host cell genome in order to generate Mut' or Mut+ expression strains of P. pastoris.
"Mut" refers to the methanol-utilization phenotype. In certain embodiments, Mut' strains are prepared by first digesting the selected expression vector with an appropriate enzyme to yield a linear DNA fragment with ends homologous to the 5' and 3' ends of the AOX1 gene.
The linearized DNA construct expression cassettes) is then integrated into the host genome at the AOX1 site by a one-step gene replacement technique. As a result of gene replacement of the AOXl, Mut' strains are obtained.
Mut' strains exhibit decreased ability to utilize methanol. This is manifestested as a slow growth rate of the strain on methanol, which is maintained by expression of the AOX2 gene product. Mut' transformants in which the expression cassette has integrated into the AOX1 locus by site-directed recombination can be identified by first screening for the presence of a complementing marker gene present in the expression cassette. This is preferably accomplished by growing the cells in medium lacking the complementing gene product and identifying those cells which are able to grow by virtue of the expression of the 25- complementing gene. Next, the selected cells are yscreene~ for their "Mut" phenotype by growing them in the r presence of methanol and monitoring their growth rate.
Mut' phenotypes have much slower growth on methanol than the.. parent strain.
To develop Mut+ IGF-1-expressing strains, the fragment that contains one or more expression cassettes) preferably is integrated into the host genome by transformation of the host with a circular plasmid or a linearized plasmid containing the DNA construct.
., 35 The DNA is integrated into a locus or loci having QiJBSTiTUT'E SHEET

2090969 pL'1'~U~ 91/06452._. _ IPE,4IUS 3 0 ~u~ X992 homology with one or more sequences present in the fragment.
Positive transformants may be characterized by Southern analysis to identify, for example, the site of DNA integration; by Northern analysis to detect, for example,r methanol-responsive IGF-1 gene expression; and by product analysis to detect the presence of secreted IGF-1 peptides in the growth medium.
Culturing transformed yeast host cells that harbor the DNA constructs.
The transformed host cells are cultured under conditions in which the DNA encoding IGF-1, signal and processing signals, is expressed, and the pre-pro-peptide is processed and secreted. Transformed strains, of the desired phenotype and genotype, may be grown in large quantities by any means known to those of skill in the art. In preferred embodiments, the strains are cultured in fermentors. For the-large-scale production of recombinant DNA-based products in methylotrophic yeast, a three-stage, high cell-density, fed batch fermentation system is normally the preferred fermentation protocol.
In the first, or growth stage, expression hosts are cultured in defined minimal medium with an excess of a non-inducing carbon source, such as but not limited 25, to, glycerol. when grown on such carbon sources, -.- heterologous gene expression is completely repressed, which allows the generation of cell mass in the absence of heterologous protein expression. It is presently preferred, during this growth stage, to maintain the pH of the medium at about 5. The second stage, referred to as a period of growth under limiting conditions, refers to a short period of non-inducing carbon source limitation growth. In this stage, cell mass continues to increase and the methanol-responsive suBS~tTUTE SHE~t 2090969 ~S 91106 4 52.-- ~.
PCT
t P EA~US 3 0 J U L 1992 promoter is derepressed. Although a pH of about 5 is preferred for optimal growth of P. pastoris, undetectable or low levels of IGF-1 are obtained from the broth of proteolytically intact IGF-1-expressing strains that are cultured entirely at pH 5. High concentrations of IGF-1 are produced, however, if the pH of the medium during the limitation growth period is adjusted to about 4 or less, preferably in the range of about 2 - 3.5. This pH is maintained during the production phase.
Expression of IGF-1 from protease deficient strains is less sensitive to pH than expression from proteolytically intact strains. Protease deficient strains may be cultured entirely at pH 5.
Alternatively, and preferably, protease deficient strains are cultured at pH 5 during the first two stages, after which the pH is lowered to about 2.5 to about 3.0 during the third stage.
In the third stage; the production stage, either methanol alone, referred to herein as "methanol excess fed-batch mode", or a limiting amount of a non-inducing carbon source plus methanol, referred to herein as "mixed-feed fed-batch mode", are added in the fermentor.
Addition of methanol induces expression of genes, including IGF-1, that are in operative linkage with 25- methanol-responsive promoters.
According to a preferred embodiment, the a _ .~heterologous protein expression system used for IGF-1 production utilizes the promoter derived from the methanol-regulated AOX1 gene of ~ pastoris. This promoter is tightly regulated and provides for efficient gene expression. The AOX1 gene can be the source of the transcription terminator as well. The presently preferred expression cassette comprises, in operational association with one another, the P. pastoris AOX1 promoter, DNA encoding the Sue. cerevisiae aMF pre-pro ,, SUBSTITUTE SHEET

2090969 p~~~ g 1/06452-.
-26- ip~~US 3 0 JUL 1992 sequence including DNA encoding the processing sequence lys-arg-(glu-ala)x, in which x may vary from 0 to 3, DNA
encoding IGF-1 polypeptides, and a transcription terminator derived from the P. pastoris AOX1 gene.
Preferably, two or more of such expression cassettes are contained on one DNA fragment, in head-to-tail orientation, to yield multiple expression cassettes on a single contiguous DNA fragment.
Selected methylotrophic yeast transformants that have the desired genotype and phenotype are cultured in fermentors under conditions in which IGF-l~is expressed and secreted into the culture medium. It is presently preferred to use the three-step production process described above. The level of IGF-1 secreted into the media can be determined by western blot analysis of the medium in parallel with an IGF-1 standard, using anti-IGF-1 antisera; by radioimmunoassay (RIA); by radio-receptor assay; or by HPLC after suitable pretreatment of the medium.
As noted above, IGF-1 as used herein is intended to embrace all the allelic variations of IGF-1. Moreover, as noted above, derivatives obtained by simple modification of the amino acid sequence of the naturally occurring product, e.a., by way of site-directed 25- mutagenesis or other standard procedures, are included ~vwithin the scope of the present invention. Forms of .. : _ .IGF-1 produced by proteolysis by host cell proteases that exhibit similar biological activities to naturally occurring IGF-1 are also encompassed by IGF-1.
The following examples are included for illustrative purposes only and are not intended to limit the scope of the invention.
EBAMPLEB
~ pastoris is described herein as a model system for the use of methylotrophic yeast hosts. Methylotrophic SUBSTITUTE SHEET

V~"O 92/04363 PCT/US91/06452 -2?-yeast host cells from four genera: Candida, Hansenula, Pichia and Torulopsis may also be used. Host cells from any species that demonstrably grows on methanol as the sole carbon nutrient may be used (see, e-a., Gleeson et al. (1988) Yeast 4: 1). Such species may be used as described herein for P. pastoris.
EBAMPhE 1 USED IN THE PREPARATION OF THE IGF-1 E8PRE88ION VECTORS.
A. Construction of plasmid pA0203.
Plasmid pA0203 contains a 5' EcoRI-3'HindIII DNA
fragment encoding the aMF pre-pro region, the amino acids of protease processing sites, lys-arg and (glu-ala)2 and the AOX1 transcription terminator.
The AOX1 transcription terminator was isolated from pPG2Ø Plasmid pPG2.0 had been prepared by inserting the BamHI-HindIII fragment of pG4.0 (NRRL 15868), which contains the AOX1 terminator, into pBR322. 20 ~g of pPG2.0 was digested with StuI, followed by the addition of 0.2 ~Cg SalI linkers (GGTCGACC). The plasmid was subsequently digested with HindIII and the 350 by HindIII-SalI fragment was isolated from a 10% acrylamide gel and subcloned into pUCl8 (Boehringer Mannheim) that had been digested with HindIII and Sa I. The ligation mix was transformed into JM103 cells and AmpR colonies were selected. The correct construction was verified by HindIII and Sa I digestion, which yielded a 350 by fragment, and was called pA0201.
5 ~,g of pA0201 was digested with HindIII, filled in using E. coli DNA Polymerase I Klenow fragment, and 0.1 ~Cg of BalII linkers (GAGATCTC) were added. After digestion of the excess BalII linkers, the plasmid was recl~sed and transformed into MC1061 cells. AmpR cells were selected, DNA was prepared, and the correct plasmid was verified by BalII, SalI double digests, which yielded 20909fi9 the expected 350 by fragment, and by a HindIII digest, which showed loss of a HindIII site. This plasmid was designated pA0202.
An alpha factor-GRF fusion was isolated as a 360 by BamHI-PstI partial digest from pYSV201. Plasmid pYSV201 had been prepared by inserting the EcoRI-BamHI fragment of GRF-E-3, which is described in European Patent Application No. 206,783, into M13mp18 (New England Biolabs). 20 ~,g of pYSV201 plasmid was digested with BamHI and partially digested with PstI. The following oligonucleotides were added to this partial digest:
5' AATTCGATGAGATTTCCTTCAATTTTTACTGCA 3' (Seq. ID No. 8) 3' GCTACTCTAAAGGAAGTTAAAAATG 5' (Seq. ID No. 9).
Only the antisense strand of the oligonucleotide was kinase labelled so that the oligonucleotides did not polymerize at the 5'- end. A 350 by fragment was isolated by electroelution following acrylamide gel electrophoresis (10%). This EcoRI- BamHI fragment of 385 by was cloned into pA0202 which had been cut with EcoRI
and BamHI. MC1061 cells were transformed and Amps cells were selected. The resulting plasmid, pA0203, was identified by cutting with EcoRI and BQ1II to yield a fragment of greater than 700 bp.
B. Construction of plasmid pA0807.
i. Preparation of fl-on DNA.
Bacteriophage fl DNA (50 ~Cg) was digested with 50 units of RsaI and DraI to release the 458 by DNA
fragment containing the fl origin of replication (ori).
The digestion mixture was extracted with an equal volume of phenol:chloroform (V/V) followed by extracting the aqueous layer with an equal volume of chloroform and finally the DNA in the aqueous phase was precipitated by adjusting the NaCl concentration to 0.2M and adding 2.5 volumes of absolute ethanol. The mixture was allowed to stand on ice (4°C) for 10 minutes and the DNA precipitate 91 /Ob x+52.__..
2090969 P~TN
iP~us ~ o ~uL~s92 was collected by centrifugation for 30 minutes at 10,000 x g in a microfuge at 4°C.
The DNA pellet was washed 2 times with 70% aqueous ethanol. The washed pellet was vacuum dried and dissolved in 25 ~1 of TE buffer (1.0 mM EDTA in 0.01 M
(pH 7.4) Tris buffer). This DNA was electrophoresed on 1.5% agarose gel and the gel portion containing the X458 by fl-on fragment was excised out and the DNA in the gel was electroeluted onto DE81 (Whatman) paper and eluted from the paper in 1M NaCl. ~ The DNA solution was precipitated as detailed above and the DNA~precipitate -was dissolved in 25 ~1 of TE buffer (fl-on fragment).
2. Cloning of fl-on into DraI sites of pHR322.
Plasmid pBR322 (2 fig) was partially digested with 2 units of DraI . The reaction was terminated by phenol:chloroform extraction followed by precipitation of DNA as detailed-in-step-1 above. The DNA pellet was dissolved in 20 ~1 of TE buffer. About 100 ng of this DNA was ligated with 100 ng of fl-on fragment (step 1) in 20 ,ul of ligation buffer by incubating at 14°C for overnight with 1 unit of T4 DNA ligase. The ligation was terminated by heating to 70°C for 10 minutes and then used to transform ~ coli strain JM103 (Janisch-Perron et 25. al. (1983) Gene ~2: 103). AmpR transformants were pooled a~d~superinfected with helper phage R4o8. Single ,'stranded phage were isolated from the media and used to reinfect JM103. AmpR transformants contained pBRfl-on which contains fl-on cloned into the Dra I sites (nucleotide positions 3232 and 3251) of pBR322.
3. Construction of the plasmid pA0807.
Plasmid pBRfl-on (l0 ~cg) was digested for 4 hours at 37°C with 10 units each of Pst I and Nde I. The digested DNA was phenol: chloroform extracted, precipitated and dissolved in 25 ~1 of TE buffer, as ,, sues~rru~ sHEEr 2090969 ~~s 91/06452 _~gEA~US 30 JUL1992 detailed in step 1 above. This material was electrophoresed on a 1.2% agarose gel and the Nde I - Pst I fragment (approximately 0.8 kb) containing the fl-on was isolated and dissolved in 20 dal of TE buffer as detailed in step 1 above. About 100 ng of this DNA was mixed with 100 ng of pA0804, which had been prepared as described in published International Patent Application No. WO 89/04320, May 18, 1989 and below and which had been digested with PstI and NdeI and phosphatase-treated.
This mixture was ligated iri 20 ~1 of ligation buffer by incubating overnight at 14°C with 1 unit o.f T4 DNA
ligase. The ligation reaction was terminated by heating at 70°C for 10 minutes. This DNA was used to transform E. coli strain JM103 to obtain pA0807.
C. Construction of plasmid pA0804:
Plasmid pA0804 has been described in published International Patent Application No. WO 89/04320, May 18, 1989. Construction of this plasmid involved the following steps:
Plasmid pBR322 was modified as follows to eliminate the coRI site and insert a c~llI site into the PvuII
site. Plasmid pBR322 was digested with EcoRI, the protruding ends were filled in with Klenow Fragment of E.
coli DNA polymerase I, and the resulting DNA was 25- recircularized using T4 ligase. The recircularized DNA
._-v~eas~used to transform ~ co i MC1061 to ampicillin-r resistance and transformants were screened for having a plasmid of about 4.37 kbp in size without an EcoRI site.
One. such transformant was selected and cultured to yield a plasmid, designated pBR322nRI, which is pBR322 with the SRI site replaced with the sequence:
- 5'-GAATTAATTC-3' 3'-CTTAATTAAG-5'.
_ Plasmid pBR322nRI was digested with PvuII, and the linker having the sequence:
suBSV~ sH~r 2090969 p~'~~ 91 /06 45~_ 5'-CAGATCTG-3' 3'-GTCTAGAC-5' was ligated to the resulting blunt ends employing T4 ligase. The resulting DNAs were recircularized, also with T4 ligase, and then digested with BalII and again recircularized using T4 ligase to eliminate multiple B~lII sites due to ligation of more than one linker to the ~vuII-cleaved pBR322nRI. The DNAs, treated to eliminate multiple BalII sites, were used to transform E.
coli MC1061 to ampicillin-resistance. Transformants were screened for a plasmid of about 4.38 kbp with a BalII _ site. One such transformant was selected and cultured to yield a plasmid, designated pBR322oRIBGL, for further work. Plasmid pBR322nRIBGL is the same as pBR322nRI
except that pBR322nRIBGL has the sequence 5'-CAGCAGATCTGCTG-3'(Sequence ID No. 11) 3'-GTCGTCTAGACGAC-5' (Sequence ID No. 12) in place of the ,III site in pBR322nRI.
Plasmid pBR322nRIBGL was digested with a SalI and ~q,~,II and the large fragment (approximately 2.97 kbp) was isolated. Plasmid pBSAGISI, which is described, for example, in European Patent Application Publication No.
0 226 752, was digested completely with _Bg~II and ~,oI
and an approximately 850 by fragment from a region of the 25- P. ~astoris AOX1 locus downstream from the AOX1 gene ~,t~anscr~-ption terminator (relative to the direction of ~tr~nscription from the AOX1 promoter) was isolated. The Bg~II-X~I fragment from pBSAGISI and the approximately 2.9.? kbp, Sa I-Bc~II fragment from pBR322nRIBGL were combined and subjected to ligation with T4 ligase. The ligation mixture was used to transform E. coli MC1061 to ampicillin-resistance and transformants were screened for a plasmid of the expected size (approximately 3.8 kbp) with a $g,~II site. This plasmid was designated pA0801.
The overhanging end of the Sa I site from the sussTrruTE sHEEt 2090969 ~~~~ 9llOb-4~~-.

pBR322oRIBGL fragment was ligated to the overhanging end of the Xh~cI site on the 850 by pBSAGISI fragment and, in the process, both the Sa I site and the XhoI site in pAO8o1 were eliminated.
Plasmid pBSAGISI was then digested with ClaI and the approximately 2.0 kbp fragment was isolated. The 2.0 kbp fragment has an approximately 1.0-kbp segment which comprises the P. pastoris AOX1 promoter and transcription initiation site, an approximately 700 by segment encoding the hepatitis B virus surface antigen (HBsAg) and an approximately 300 by segment which comprises the P.
bastoris AOX1 gene polyadenylation signal and site-encoding segments and transcription terminator. The HBsAg coding segment of the 2.0 kbp fragment is terminated, at the end adjacent the 1.0 kbp segment with the AOX1 promoter, with an coRI site and, at the end adjacent the 300 by segment with the AOX1 transcription terminator, with-a-Stul-site, and has its subsegment which codes for HBsAg oriented and positioned, with respect to the 1.0 kbp promoter-containing and 300 by transcription terminator-containing segments, operatively for expression of the HBsAg upon transcription from the AOX1 promoter. The coRI site joining the promoter segment to the HBsAg coding segment occurs just upstream 25. (with respect to the direction of transcription from the AOXl promoter) from.the translation initiation signal-,'encoding triplet of the AOX1 promoter.
Plasmid pA0801 was cut with C aI and mixed with T4 ligase and the approximately 2.0 kbp C aI-site-terminated fragment from pBSAGISI. The ligation mixture was used to transform ~ coli MC1061 to ampicillin resistance, and t~ansformants were screened for a plasmid of the expected size (approximately 5.8 kbp) which, upon digestion with ~I and $g~II, yielded fragments of about 2.32 kbp (with the origin of replication and ampicillin-resistance gene sues~rru~ sH~r 2090969 ~ 1 ~p6 x+52_ _ .
p~TN ~
tpEA~US 3 0 JUL 1992 from pBR322) and about 1.9 kbp, 1.48 kbp, and 100 bp. On digestion with BalII and coRl, the plasmid yielded an approximately 2.48 kbp fragment with the 300 by terminator segment from the AOX1 gene and the HBsAg coding segment, a fragment of about 900 by containing the segment from upstream of the AOX1 protein encoding segment of the AOX1 gene in the AOX1 locus, and a fragment of about 2.42 kbp containing the origin of replication and ampicillin resistance gene from pBR322 and an approximately 100 by ClaI-BalII segment of the - AOX1 locus (further upstream from the AOX1-encoding segment than the first mentioned 900 by EcoRI-BalII
segment). Such a plasmid had the ClaI fragment from pBSAGISI inserted in the desired orientation. Had it been inserted in the opposite undesired orientation, there would have been coRI-Bc~II fragments of about 3.3 kbp, 2.38 kbp and 900 bp. One of the transformants harboring the desired plasmid, designated pA0802, was selected for further work and was cultured to yield that plasmid.
Plasmid pA0802 was then treated to remove the HBsAg coding segment terminated with an EcoRI site and a StuI
site. The plasmid was digested with StuI and a linker of sequence:
25- 5'-GGAATTCC-3' 3'-CCTTAAGG-5' ,~ .: _ .was,ligated to the blunt ends using T4 ligase. The mixture was then treated with coRI and again subjected to ligating using T4 ligase. The ligation mixture was used to transform ~ coli MC1061 to ampicillin resistance and transformants were screened for a plasmid of the expected size (5.1 kbp) with coRI-BalII fragments of about 1.78 kbp, 900 bp, and 2.42 kbp and Bg,~II-ClaI
fragment of about 100 bp, 2.32 kbp, 1.48 kbp, and 1.2 kbp. A transformant with the desired plasmid was SUBSTITUTE SHEET

PCTIU~ 91 /06 x+.52.__.
tPEAIUS 3 0 JUL 1992 selected for further work and was cultured to yield pA0803.
Plasmid pA0804 was prepared from pA0803 by inserting, into the nHI site from pBR322 in pA0803, an approximately 2.75 kbp BalII fragment from the P.
pastoris HIS4 gene. See, e.a., Cregg et al. (1985) Mol.
Cell. Biol. ~, 3376 and European Patent Application Publication Nos. 180,899 and 188,677. Plasmid pA0803 was digested with BamHI and combined with the HIS4 gene-containing $g~II site-terminated fragment and the mixture - subjected to ligation using T4 ligase. The ligation mixture was used to transform E. coli MC1061 to ampicillin-resistance and transformants were screened for a plasmid of the expected size (7.85 kbp), which is cut by SalI. One such transformant was selected for further work, and the plasmid it harbors was designated pA0804.
Plasmid pA0804 has one SalI-ClaI fragment of about 1.5 kbp and another of abut 5.0 kbp and a C aI-C aI
fragment of 1.3 kbp, indicating that the direction of transcription of the S4 gene in the plasmid is the same as the direction of transcription of the ampicillin resistance gene and opposite to the direction of transcription from the AOX1 promoter.
The orientation of the HIS4 gene in pA0804 is not 25- critical to the function of the plasmid or any ~-derivat~ves that include heterologous DNA fragments .- i ~ir~serted at the coRI site between the AOX1 promoter and terminator fragments. Thus, a plasmid with the HIS4 gene in the orientation opposite to that of the HIS4 gene in pA0804 could be used in place of pA0804.
D. Construction of plasmid pAOSiS
Plasmid pAO815 was constructed by mutagenizing plasmid pA0807 to change the C aI site downstream from the AOX1 transcription terminator in pA0807 into a BamHi site. The oligonucleotide used for mutagenizing pA0807 s SUBSTITUTE SHEET

2090969 ~'CTtU~ 9 flab 452_ -35- ~P~~~S 3 0 JUL1992 had the following sequence:
5'-GAC GTT CGT TTG TGC GGA TCC AAT GCG GTA GTT TAT-3' (Sequence ID No. 10).
The mutagenized plasmid was designated pA08o7-Bam.
Plasmid pA0804 (described above and in International Patent Application WO 89/04320) was digested with Bg~II
and 25 ng of the 2400 by fragment were ligated to 250 ng of the 5400 by Bc~II fragment from BalII-digested pA0807-Bam. The ligation mix was transformed into MC1061 cells and the correct construct was verified by digestion with PstI/.~amHI to identify 6100 and 2100 by sized bands. The correct construct was called pA0815. The restriction map of the expression vector pA0815 is shown in Figure 6.

CONBTROCTION OF VECTORS FOR EBPRESSION OF IGF-1 IN _P.
PASTORIB.
The expression vector constructions were carried out using standard procedures, as described, for example by Maniatis et al. ((1982) Molecular Cloning: A Laboratorv anual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, USA) and Davis e~ ~, ((1986) Basic Methods in Molecular Bioloav, Elsevier Science Publishing, Inc., New York).
A. Construction of expression vectors that contain DNA oncoding aMF pr~pro-lys-arg-IGF-i.
Expression vectors were prepared that contain one or -..--,fnulfiplWcQpies of the IGF-1 gene expression cassette, 5' ~AO~1-aMF prepro-lys-arg-IGF-1 - 3' OX , in which the aMF
prepro-encoding element contains the aMF prepro-lys-arg but-~-lacks glu-ala spacer-encoding nucleotides.
i. Construction of pIGF2oi which contains a single copy of the expression cassette.
The synthetic gene encoding human IGF-1 shown in Figure 1 as a HindIII-BamHI fragment, was incorporated into vector pUCl8 and was used to transform E. coli strain MC1061. Ampicillin-resistant transformants were SUBSTITUTE SHEET

WO 92/04363 ~ ~ ~ ~ ~ PCT/US91/0645~...

selected and screened by examination of restriction enzyme-digested DNA for the presence of a HindIII-BamHI
insert of approximately 240 bp, the size expected for the IGF-1 gene. One transformant with an insert of this size was used to prepare plasmid DNA designated pIGF101.
The 240 by HindIII-F3amHI fragment (250 ng) isolated from pIGFl0l, containing the IGF-1 gene, was inserted into the HindIII-BamHI site of plasmid pA0203 (10 ng), which contains DNA encoding an EcoRI site at the 5' end, the aMF pre-pro region, followed by nucleotides which encode the amino acids for three processing sites, lys-arg and (glu-ala)Z, and including a HindIII site at the 3' end. The resulting plasmid was used to transform E. coli MC1061 cells. Ampicillin-resistant colonies were selected and screened with an oligonucleotide complementary to sequence 162 to 132 of the IGF-1 gene, Sequence ID No. 1. One colony which was positive in this screen was selected and its plasmid designated pIGF102.
The EcoRI-BamHI fragment from pIGF102 (250 ng), containing the aMF pre-pro region and proteolytic processing sites, and IGF-1 gene, was cloned into M13mp19 (10 ng) and used to transform E. coli JM103 cells. The resulting transformants were screened by analysis of restriction enzyme-digested DNA, and one clone ~(pIGF103) with an insert of the correct size (480 by coRl-BamHI
fragment) was used to prepare single-stranded DNA for site-directed mutagenesis. Site-directed mutagenesis of the single-stranded DNA was performed to delete the (glu-ala)2 spacer sites, the HindIII cloning site, the polylinker attached to the synthetic gene, and the codon for the initial methionine of IGF-1. Mutagenesis was accomplished using standard procedures and oligonucleotides having the following sequences:
mutagenizing oligonucleotide (Sequence ID No.
2):

2090969 p~/~ 91/06452 .
_3,_ iP~us 3 0 ~uLt992 5'-GTATCTTTGGATAAAAGAGGACCGGAGACGCTCTGC-3' screening oligoncul~otide (Sequence ID No. 3):
5'-ATAAAAGAGGACCGGA-3'.
Removal of the above noted sequences yielded a fusion gene containing DNA encoding the aMF pre-pro region and lys-arg processing site fused directly to the coding region of the IGF-1 gene.
The selected clone, designated pIGF104, was sequenced to verify the changes and then subjected to a second site-directed mutage~nesis to insert an EcoRI site immediately following the translation termination codon of the IGF-1 gene. Oligonucleotides having sequences as follows were used in this mutagenesis:
mutageni$ing oligonucleotide (sequence ID No.
4) 5'-AGTCAGCTTGATAAGAATTCAAATGAGTCGACCTGCAGGC-3' screening oligonculeotide (Sequence ID No. 5):
5'-TAAGAATTCAAATGAGT-3'.
The mutagenized clone was designated pIGF105.
After the second mutagenesis was confirmed by DNA
sequencing, the aMF-IGF-1 gene fusion was isolated on a 450 by coRI fragment, then 250 ng of this coRI fragment was inserted into 10 ng of the P. pastoris expression vector pA0815, which had been previously digested with 25- SRI and treated with calf alkaline phosphatase. The ''' resylting single-copy expression vector, pIGF201, .contains one copy of the aMF-IGF-1 fusion gene under the transcriptional control of the Pichia pastoris AOX1 promoter and regulatory regions, as well as the AOX1 transcription termination and polyadenylation signals.
In addition, the vector includes the Pichia pastoris H S4 gene used for selection in His- hosts and additional 3' AOX1 sequences which can be used to direct integration of the vector into the host genome. Plasmid pIGF201 is shown in Figure 2.
SUBSTITUTE SHEET

2090969 p~'~'~j~ 91/06 452 IPEA~US 3 0 JUL 1992 The entire aMF-IGF-1 fusion gene and approximately 50 nucleotides each of the promoter and termination regions of pIGF201 were sequenced to verify that the nucleotide sequences were not altered during the cloning process.
2. Construction of sxprsssion vectors pIGF202, pIGF204, and pIGF206, which contain multiple copies of the s~cprsssion cassette.
The expression cassette containing the AOX1 promoter and regulatory region, the aMF-IGF-1 fusion gene, and the AOX1 transcription termination and polyadenylation signals was isolated from pIGF201 as a 1700 by BQ1II- -BamHI fragment. The Bg~II-BamHI expression cassette (250 ng) was inserted back into the unique BamHI site of pIGF201 (10 ng of BamHI digested, calf alkaline phosphatase-treated pIGF201). MC1061 cells were transformed with the ligation. AmpR colonies were selected and plasmid was characterized by restriction digest. Analysis of restriction enzyme digests of the resulting plasmid, pIGF202 shown in Figure 3, verified that the two expression cassettes were joined as tandem-repeat units rather than inverted-repeat units: SalI
digest yielded -2100, 1750, and 6100 by bands; ClaI/BamHI
digest yielded -3800 and 6450 by bands.
The III- a$~mHI fragment from plasmid pIGF202, containing two copies of the expression cassette, was ~i'soiatedw(250 ng) and inserted back into the unique BamHI
'side in calf alkaline phosphatase-treated pIGF202 (10 ng) to yield vector pIGF204 shown in Figure 4, containing four copies of the expression cassette. MC1061 cells ' were transformed with the ligation. AmpR colonies were selected and plasmid was characterized by restriction digest. Correct plasmid demonstrated -6600 and 6900 by bands upon digestion with C' aI and a~HI.
To construct expression vector pIGF206, which contains six copies of the expression cassette, the SUBSTITUTE SHEET

p~~~ 91J06~+52 __.
'PEA~US 3 0 JUL 1992 $g~,II-BamHI fragment from pIGF202 (250 ng) was cloned into the BamHI site of pIGF204 (10 ng; pIGF204 had been previously treated with calf alkaline phosphatase).
MC1061 cells were transformed with the ligation. AmpR
colonies were selected and plasmid was characterized by restriction digest. Restriction digests of the vector DNA
were examined to verify the number of expression cassettes and that the expression cassettes were joined as tandem-repeat units. Correct plasmid yielded bands of -6600 and 10300 by upon digestion with ClaI and BamHI.
Plasmid pIGF206 is shown in Figure 5.
H. Construction of expression vectors containing aMF prspro-lys-arg-glu-ala-IGF-1.
IGF-1 expression vectors containing DNA encoding the aMF prepro sequence, including the lys-arg and one glu-ala processing sequence, fused to the IGF-1 coding sequence were prepared from pIGF105, which contains the aMF prepro-lys-arg-IGF-1 fusion gene in M13mp19 (see Example 2.A.1). Site-directed mutagenesis of single-stranded DNA from pIGF105 was performed to insert codons for a glu-ala spacer site between the aMF prepro-lys-arg and IGF-1 DNA sequences. Mutagenesis was accomplished using standard procedures and oligonucleotides having the following sequences:
25- Mutageni$iag oligonucleotids (Sequence ID No. 6) ~.- 5'-TCTTTGGATAAAGAGAGGCTGGACCGCAGACGCTC-3' Bc~.saing oligonuclsotiae (Sequence ID No. 7) 5'-AAAAGAGAGGCTGGACCGC-3'.
The selected clone was sequenced to verify that the DNA
encoding the glu-ala residues was added correctly. The aMF prepro-lys-arg-glu-ala-IGF-1 gene fusion was isolated from the selected clone on a 465-by coRI fragment and ligated to coRI-digested, alkaline phosphatase-treated vector pA0815. The resulting single copy expression vector, pIGF816, contains one copy of the aMF prepro-lys-SUBSTITUTE SHEET

p~~~ 91/06+52 ..
I P EA~US 3 0 J U L 1992 arg-glu-ala-IGF-1 gene fusion under transcriptional control of the ~. pastoris AOXl promoter and regulatory regions, as well as the AOXl transcription termination and polyadenylation signals. The vector also includes the P. pastoris HIS4 gene for selection in His' hosts and additional AOX1 sequences. The entire aMF prepro-lys-arg-glu-ala-IGF-1 gene fusion and approximately 30 nucleotides each of the promoter and termination regions of pIGF816 were sequenced to verify that the sequences were not altered during the cloning process.
Expression vectors containing two or more copies of the aMF prepro-lys-arg-glu-ala-IGF-1 gene cassette can be prepared as described for construction of vectors containing multiple copies of the aMF prepro-lys-arg-IGF-1 gene expression cassette (see Example 2.A.2.).

E8PRE88ION VECTORB_INTO STRAINS OF PICHIA PA8TORI8.
A. Mut+ strains Expression vectors pIGF201, pIGF202, pIGF204, and pIGF206 and pIGF816 were used to develop IGF-1-expressing Mut+ strains of Pichia pastoris. The Mut phenotype refers to the methanol utilization ability of the strain.
Mut* strains consume methanol at a rate similar to that of wild-type strains. A His' mutant of Pichia pastoris, ~~-- GS115 (ATCC No. 20864), was used as the host for all .'transfo~'mations, which were accomplished by either the spheroplast method (performed as described in U.S. Patent No. 4,879,231) or by the whole-cell lithium chloride yeast transformation system (Ito, et al. (1984) A ric.
Biol. C m. g$: 341) with modification necessary for adaptation to methylotrophic yeast, such as _P. pastoris (see European Patent Application No. 312,934; also available as U.S. Patent No. 4,929,535).
Mut'' strains were generated by integration of the SUBSTITUTE SHEET

2~p9~p9G9 entire expression vector into either the AOX1 or HIS4 locus of the host genome by an additive homologous recombination event. Plasmid pIGF201 was transformed into GS115 as an undigested circular vector and allowed to integrate randomly into the OX1 locus at either the 5' or 3' regions homologous to sequences contained in the plasmid, or into the HIS4 locus. For site-directed addition to the HIS4 locus, the multi-copy expression vectors pIGF202, pIGF204, and pIGF206 and the single copy vector pIGF816 were digested with StuI, which linearizes the plasmids by cutting the vector within the HIS4 region. Additive integration at either the AOX1 or HIS4 locus does not disturb the AOX1 gene.
The Mut+ transformants resulting from additive integration of the expression plasmids, either randomly (pIGF201) or by site-direction (pIGF202, pIGF204, pIGF206 and pIGF816), were initially screened for histidine prototrophy. Prototrophic strains generated by separate transformation of GS115 with each of the four plasmids were selected for further analysis.
Ten of the Hisa transformants resulting from transformation of GS115 with pIGF201 were analyzed by Southern blot hybridization to verify the site of integration of the expression plasmid and number of copies of the plasmid integrated. Chromosomal DNA from the 10 transformants was separately digested with EcoRI
and BalII, separated by agarose gel electrophoresis, and transferred to nitrocellulose. The EcoRI digests were probed with pBR322-based plasmids containing either the AOX1 5' and 3' regions or the Pichia pastoris HIS4 gene.
The BalII digest was probed with an oligonucleotide homologous to the IGF-1 gene.
Sixteen of the His'" transformants generated by transformation of GS115 with pIGF202, 16 generated by transformation with pIGF204, and 26 generated by 2U~U969 transformation with pIGF206 were also analyzed by Southern blot hybridization to verify the site of integration and integrity of the multi-copy expression vector. Chromosomal DNA was digested with BalII and probed with plasmids containing the AOX1 5' and 3' regions or HIS4 gene, and also separately digested with StuI and probed with the oligonucleotide homologous to the IGF-1 coding sequence.
Analysis of DNA from ten pIGF201 transformants by Southern blot hybridization indicated that four transformants contained a single copy of the expression vector integrated at the AOX1 locus and two transformants contained multiple copies of the plasmid integrated at the HIS4 locus. The four other strains contained the plasmid integrated at unknown loci. It was not possible to determine the exact number of copies of the plasmid integrated at HIS4. Southern analysis of DNA revealed that 10 of the pIGF202 transformants, 9 of the pIGF204 transformants, and 13 of the 26 pIGF206 transformants contained a single copy of the respective expression vector integrated at HIS4. The other transformants contained plasmids integrated at unknown loci.
The following representative strains generated by transformation of GS115 with vectors pIGF201, pIGF202, pIGF204 or pIGF206 were selected for further analysis:
Number of Expression Plasmid Site of Plasmids Cassette Strain Name Integrated Integration Integrated Copy Number G+IGF201S1 pIGF201 AOX1 1 1 G+IGF206S2 pIGF206 HIS4 1 6 G+IMB202S2 pIGF202 HIS4 1 2 G+IMB204S14 pIGF204 HIS4 1 4 G+IMB206S1 pIGF206 HIS4 1 6 In a i ion, our s rains genera a y transformation with pIGF816, strains G+IGF816S1, G+IGF816S2, G+IGF816S9 and G+IGF816S11, were also analyzed further during growth in one-liter fermentations.
B. Mut- strains In Mut~ strains, the expression vector integrates into the AOX1 locus and disrupts it. Thus, Mut- strains consume methanol at a much slower rate than Mut' strains.
To generate Mut- strains, plasmid pIGF206 was digested with BalII. This liberates a fragment comprised of the six IGF-1 expression cassettes, the HIS4 gene for selection, and the AOX1 3' region. Both ends of this fragment contain long sequences which are homologous to the 5' and 3' ends of the AOX1 locus. Upon transformation of GS115 hosts with the BalII-ended fragment, integration into the AOX1 locus by homologous recombination results in substitution of the ,c~lII-ended fragment for the AOX1 structural gene. Positive transformants were selected by their His+ phenotype and by their Mut- phenotype, slow growth on methanol.
Selection is accomplished by plating His' transformants on minimal glucose (2%) master plates to obtain colonies originating from single cells. After overnight incubation at 30°C, the masters were replica-plated to minimal glucose plates and plates containing no carbon source to which methanol was added in vapor phase. This is accomplished by adding approximately 200 ~1 of methanol to the underside of the top of a covered petri dish. The plates were incubated at 30°C for 4-6 days with additional methanol added in the vapor phase every two days. Colonies showing visible growth were scored as Mut' and those with no visible, or slow, growth were scored as Mut-. Approximately 15% were slow growers, which is indicative of disruption of the AOX1 gene.
Following the initial screening for methanol 2090969 PG'C~ls 91 ~~~ X52 _ _ .
-44- ~P~~US 3 0 JUL t992 utilization and histidine prototrophy, chromosomal DNA
from three Mut- transformants was analyzed by three different Southern blots to verify the site of plasmid integration and number of copies integrated.
Chromosomal DNA from Mut- transformants was digested with coRI and probed with pBR322-based plasmids containing either the AOX1 5' and 3' regions or the Pichia pastoris H S4 gene. An additional blot of Bg,~II-digested DNA was probed with the plasmid containing the IGF-1 gene. Southern analysis of the three Mut' transformants generated using plasmid pIGF206 revealed that all three contained a single copy of the BalII
fragment, containing six expression cassettes, integrated by replacement at the AOX1 locus.
The strains were characterized as follows:
Number of Expression Site of Plasmids cassette Strain Name Integration Integrated copy number C. Breparation of protease-deficient strains for e3cpressioa and secretion of Ic3F-1.
The use of protease -deficient strains of P.
pastoris hosts for recombinant expression of heterologous proteins susceptible to degradation by ~ pastoris -.- proteases is described in U.S. Patent Application No.
r0'T~~678,916~ Protease-deficient ~ pastoris strains have been generated by disrupting P. pastoris genes, such as PEP4 and PRB-1, that encode proteins that directly or indirectly affect protease activity of the cell.
Disruption of these genes results in a reduction of at least a portion of the protease activities in the cell, including proteinase A, carboxypeptidase Y and proteinase 8 activities. Recombinant Pep4' (also called pep4_) and Pep4-Prb-1- (also called pep4 prb-1) IGF-1-secreting SUBSTITUTE SHEET

wc> azi~a~~~ ~~crvus9mo~as2 strains of P. pastoris_ have been prepared.
Four p~.~o';.ease-d~~f icient IGF-1 expressing strains of P, pastoris, strains M+IMB206S1, M+IGF816S1, M+IGF816S4 and C+IGF816S1 were generated using two different approaches. Strain M+IMB206S1 was developed by transformation.of a P, pastoris strain containing six copies of an aMF prepro-lys-arg-IGF-1 gene expression cassette (strain G+IMB206S1, see Example 3.A.) with a PEP4 gene disruption vector, pDR421. Vector pDR421 (described below and in detail in copending Canadian Application No. 2,105,064 contains an internal portion of the _P. pastoris PEP4 gene which integrates into the host genome at the PEP4 locus to generate two incomplete and nonfunctional copies of the PEP4 gene.
Strains M+IGF816 S1, M+IGF816S4 and C+IGF816S1 were generated by transforming protease deficient P. pastoris host strains with the aMF prepro-lys-arg-glu-ala-IGF-1 gene construct. Strains M+IGF816S1 and M+IGF816S4 were generated by transformation of a pep4 host and C+IGF816S1 by transformation of a pep4, prb-11 host.
1. Construction of the P. oastoris PEP4 gene disruption vector pDR421.
Plasmid pDR421 was constructed for use in the preparation of PEP4-deficient (Pep4-) strains of P.
pastoris in which the host PEP4 gene is disrupted by addition of an incomplete PEP4 gene to the endogenous PEP4 locus. Plasmid pDR421 contains an internal portion of the PEP4 gene, which, when introduced into the host genome at the PEP4 locus, results in two incomplete and nonfunctional copies of the PEP4 gene.
To construct pDR421, the URA3 gene of P. pastoris was cloned into vector pEP205. Plasmid pEP205 contains pUCl9 sequences and the portion of the ~EP4 gene encoded by the approximately 450 by BamHI fragment derived from r.

. -, 2090969 p~"~S 91106 ~+~2 IpEMUS 3 0 JUL1992 pEP202. Plasmid pEP202 was prepared by inserting a 10.6 kb EcoRI fragment of the P. ~astoris genome, that had been identified as containing the PEP4 gene by Southern blot hybridization of EcoRI-digested ~ gastoris genomic DNA with the homologous S. cerevisiae PEP4 gene as a probe, into pUCl9. The 10.6 kb fragment was isolated and approximately 200 ng of it was ligated with an equal amount of pUCl9 which had been cut with EcoRI and dephosphorylated. The ligation mixture was used to transform E. coli strain MC1061. Ampicillin-resistant . colonies were selected and screened by analysis of restriction enzyme digests of colony DNA for the presence of the diagnostic 10.6 kb Fc~RI fragment. A large-scale plasmid preparation was made from a colony containing the correct plasmid, which was named pEP202. Plasmid pEP202 contains the complete P. pastoris PEP4.
The (~R 3 gene was isolated from Spel/S_~hI digested pPU205 as a 2 kb fragment and ligated to XbaI/SphI-digested pEP205 and used to transform E. co i strain MC1061 to ampicillin resistance. AmpR colonies were screened by analysis of $~nHI/S~hI digested colony DNA
for the presence of 2.7kb, 0.4 kb and 1.9 kb fragments.
A selected transformant that contained the correct plasmid construct was identified and the plasmid was 25- designated pDR421.
2. Pr~paration of M+IMH20681 M+IMB206S1, a peb4 IGF-1 expressing strain, was generated by transforming a pep4+ ~ Pastoris strain that contained six copies of an aMF prepro-lys-arg-IGF-1 construct with the P.EP4 disruption vector pDR421, which contains an internal portion of the ~ gastoris PEP4 gene and which integrates into the host genome at the PEP4 locus, thereby disrupting the gene locus by generating two incomplete and nonfunctional copies of the gene.
SUBSTITUTE SHEET

2090969 PGTtuS 9I/061~52 - li~~A~US 3 0 JUL 1992 In order to facilitate identification of transformants that incorporated pDR421, an auxotrophic marker selection system was established in strain G+IMB206S1 prior to transformation. This was accomplished by isolation of a colony of G+IMB206S1 that, through spontaneous mutation, became auxotrophic for uracil (Ura-). Because pDR421 contains a P. pastoris UR~3 marker gene, in addition to the PEP4 fragment, it was possible to initially select those cells of the Ura-derivative of G+IMB206S1 that had incorporated pDR421 on the basis of uracil prototrophy. The isolation of a Ura-IGF-1-expressing P. pastoris strain and its use in generating a pep4 IGF-1-expressing P. pastoris strain are described in this section.
a. Isolation of the vra- Ic3F-1 expressing strain IGF-v.
Resistance to a 5-fluoro-orotic acid (5-FOA) was used as a means-of~identifying Ura~ derivatives of strain G+IMB206S1. 5-FOA is an analog of a uracil biosynthetic pathway intermediate that, when metabolized by Ura+
strains, yields a toxic compound. Therefore, Ura+
strains cannot survive on 5-FOA-containing medium. In contrast, the uracil biosynthetic pathway of Ura- strains is blocked at certain steps, and such strains do not 25_ metabolize 5-FOA and are unaffected by its toxic effects.
-~- Consequently, plating cells on 5-FOA-containing medium ,can be used as a method to detect Ura- strains generated by spontaneous mutation.
A Ura derivative of the IGF-1-producing strain G+IMB206S1 was isolated by direct plating of approximately 5 X 10' cells into a 5-FOA-containing medium supplemented with uracil (0.67% yeast nitrogen base, 2%
agarose, 2% glucose, 750 mg/L of 5-FOA and 45 mg/L of uracil). After one week of incubation at 30°C, a colony growing on the plate was isolated. This colony, which SUBSTITUTE SHEET' 2090969 i'~'I'I~ 91/06 452 . .
~ s ~r~.: :~
0 J U 1.1992 required uracil to grow, was unable to complement a ura3 strain of Pichia pastoris and was designated IGF-U.
b. Transformation of IGF-U with pDRl21 Plasmid pDR421 contains the 'c is UR~3 gene and a small portion of the Pichia PEP4 gene. BalII-digested plasmid pDR421 was used to transform IGF-U using standard spheroplast transformation procedures. Digestion of pDR421 with BalII linearized the vector by cleaving it within the PEP4 gene fragment. Thus, the resulting linear DNA had 5' and 3' ends containing sequences from the PEP4 gene and a URA3 gene. Because the ends of this B~,~II fragment are homologous to adjacent sequences of the PEP4 gene, integration of the fragment in the genome of IGF-U was directed at the PEP4 locus via homologous recombination that resulted in disruption of the PEP4 locus. The URA3 gene of the integrated DNA fragment conferred a Ura+ pheno-type on the transformants that provided a marker for strains which were transformed.
Ura+ transformants were subsequently analyzed for carboxypeptidase Y (CPY) activity using a colony overlay colorimetric screening procedure (Jones, E. (1977) Genetics 85: 23-33). Colonies which appeared to have low carboxypeptidase Y (CPY) activity based on the results of this assay were isolated, subcultured, and then 25- rescreened using the overlay assay. One of these strains ''' ~wasr designated M+IMB206S1.
~3. Pr~paration of M+IGF81681 and M+IGF8168~
Protease-deficient strains containing one or more copies of the aMF-prepro-lys-arg-glu-ala-IGF-1 gene expression cassette were generated by transformation of a Pep4- His4- strain of ~. pastoris, designated MGP21, with pI~GF816 .
a. Gen~ration of MGP21 The protease-deficient host strain MGP21 was generated by gene replacement methods, as described in :, V1'()~92/04363 PCT/US91 /06452 U.S. Patent Application Serial No. 678,916, in which the P_EP4 4ene is replaced with a 3efective~ pe 4 gene.
Briefly, a His4- Ura- P. pastoris strain (GS4-2), selected from cells of a His4~ strain (GS115) that had been plated onto 5-FOA, was transformed with a linear fragment of a PEP4 disruption vector designated pDR602 (see copending Cdn. Appln. No. 2,105,064). The linear fragment of pDR602 -consisted of the P. ~astoris URA3 gene flanked on each side with DNA coding for a portion of the P.
l0 pastoris PEP4 gene. Homology between the ends of the fragment and the endogenous PEP4 gene of GS4-2 stimulated integration of the fragment at the PEP4 locus resulting in a gene replacement event in which the defective URA3-containing pe_p4 gene replaced the endogenous PEP4 gene.
The resulting strain, MGP21, is His4- Ura+ Pep4-.
b. Transformation of MGP21 Strain MGP21 was transformed with StuI-digested pIGF816 via the lithium chloride method. Because StuI
cleaves pIGF816 in the HIS4 gene, the linear fragment was directed to integrate into the his4 locus of strain MGP21. Resulting transformants were screened for histidine prototrophy. Two of the strains were designated M+TGF816S1 and M+IGF816S4.
Preparation of strain C+IGF81681 -A pep4 prb-1 IGF-1-expressing strain of P. pastoris containing one copy of the aMF prepro-lys-arg-glu-ala-IGF-1 expression cassette was generated by transformation of a pep4 -prb-1 ura3 his4 P. pastoris host strain, designated MG18, with pIGF816.
a. Generation of MG18 Strain MG18 was generated by disruption of the PRB-1 gene of the pep4 ura3 his4 P. pastoris strain GS4-2521-4-5 (see copending Canadian Application No. 2,105,064 for a detailed description of the generation of GS4-2521-4-5 from the ura3 his4 strain GS4-2 via the intermediate 20909b9 strain GS4-2521-4). This was accomplished by transformation of GS4-2521-4-5 with the PRB-1 disruption vector pDR911. Vector pDR911 (described in copending Canadian Application No. 2,105,064) contains an internal portion of the P_. pastoris PRB-1 gene, which, when used to transform PRB-1 strains of g. Qastoris, integrates into the host genome at the PRB-1 locus to generate two incomplete and non-functional copies of the PRB-1 gene.
Vector pDR911 also contains a complete functional ~.
pastoris URA3 gene for use as a selectable marker in ura3 host strains of P. pastoris, such as GS4-2521-4-5.
Plasmid pDR911 was linearized by cleavage with BqlII
which cuts the plasmid within the fragment of the PRB-11 gene. The linear DNA was used to transform GS4-2521-4-5 according to the spheroplast method. Because the ends of this linear DNA are homologous to adjacent sequences of the _P_RB-11 gene, integration of the DNA into the genome of GS4-2521-4-5 was directed at the P B- locus via a homologous recombination event resulting in the disruption of the PRB-1 locus. The URA3 gene of the integrated DNA fragment conferred a Ura+ phenotype on the transformants that provided a marker for strains that had been transformed. Genomic DNA from transformants that were able to survive on media lacking uracil was analyzed by Southern blot hybridization. The DNA was digested with coRV, seperated by electrophoresis on a 0.75%
agaros~~gel and transferred to nitrocellulose. The filter was probed with random-primed pDR911. The band pattern expected for a strain in which the PRB-1 gene had been disrupted was a loss of the 5 kb fragment representing the endogenous undisrupted PRB-1 gene and the appearance of an approximately 10 kb fragment representing the endogenous 5 kb RP B-1 fragment plus the 5 kb linear vector DNA pDR911. DNA from 3 out of 40 transformants exhibited this band pattern. One of these A

~ strains was MG18.
b. Transformation of MG18 with pIGF816 To develop ,pep4 prb-11 strains of P.
~astoris, strain MG18 was transformed with StuI-digested pIGF816 by the lithium chloride method. Because StuI
cleaves pIGF816 in the HIS4 gene, the linear fragment was directed to integrate into the 's4 locus of strain MG18.
Resulting transformants were screened for histidine prototrophy. One strain was designated C+IGF816S1.

GROWTH OF 8TRAINB IN ONE- AND TEN-LITER FERMENTATIONS.
Medium employed in fermentations described herein had the following composition:
A. lOx BASAL SALTS
Chemical Grams/liter Phosphoric acid, 85% 42.0 ml Calcium Sulfate.2H20 1.8 Potassium Sulfate 28.6 Magnesium Sulfate.7H20 23.4 Potassium Hydroxide 6.5 B . PTM~ TRACE 8ALT8 Chemical Grams/liter Cupric Sulfate~5H20 6.0 Sodium Iodide 0.08 Manganese Sulfate~H20 3.0 Sodium Molybdate~2H20 0.2 Boric Acid 0.02 Cobalt Chloride 0.5 Zinc Chloride 20.0 Ferrous Sulfate~7Hz0 65.0 Biotin 0.20 Sulfuric Acid 5.0 ml ._.. ~~s _91 ~ 0 6 ~ ~.~_ .
2090969 ip~US
3 0 J U l 1992 A. One-Liter Fermentations In order to maximize expression and secretion of IGF-1, the effects of glycerol concentration and pH on IGF-1 production in one-liter fermentations have been investigated.
1. Mut+ protocol:msthanol fsd-batch fermentation.
The Mut+ fermentation protocol included three separate phases: (1) cells are initially cultured on glycerol in a batch_mode; (2) following exhaustion of the glycerol, a limited glycerol feed is initiated such that glycerol does not accumulate, the AOX1 promoter is derepressed and cell mass increases: and (3) a methanol feed is initiated for production of heterologous protein in a methanol fed-batch mode.
Prior to use the fermentor was autoclaved with about one liter of medium containing 500 ml of lOx basal salts medium (final basal salts concentration of 5X, see above) and 30-50 g glycerol. After sterilization, 4 ml PTM, trace salts (see above) were added to.the fermentor and the pH was adjusted to 5 with concentrated NH40H. In runs in which the pH was maintained at 5, pH was adjusted by addition of 50% NHdOH containing 0.1% Struktol J673 antifoam. Dissolved oxygen was maintained above 20% of saturation by regulating agitation and aeration, or -.- supplementing the air feed with oxygen. In low pH
'fermentations in which the pH was maintained between 2.8 and 3.5, the pH controller was adjusted to the desired pH
either at the initiation of the limited glycerol feed or at the initiation of the methanol feed and then allowed to decrease by virtue of cellular metabolism to the desired pH.
Inocula were prepared from buffered YNB glycerol plates and grown overnight at 30°C in phosphate-buffered YNB (11.5 g/L KHzP04, 2.66 g/L K2HP04, 0.67% yeast nitrogen SUBSTITUTE SHEE?

base, pH 6)) containing 2% glycerol. The fermentor was inoculated with these cultured cells, which had grown to an OD6oo of 1-6, and the batch growth regimen was continued for 18 to 24 hours. At the point of glycerol exhaustion, indicated by increased dissolved oxygen, a glycerol feed (50% glycerol plus 12 ml/1 of PTM~) was initiated at a rate in the range of about 5-20 ml/h.
This period of growth on limited glycerol was continued until a total volume of approximately 40-300 ml of feed was added to the fermentor. After addition of the glycerol feed, a methanol feed (100% methanol plus 12 ml/L PTM~) was started. The initial rate of methanol addition was approximately 1-2 ml/h. The methanol feed rate was increased for 3 to 8 hours to 5-6 ml/h. In some fermentations the feed rate was directly increased to a rate of 6 ml/h after three hours of methanol feeding.
The vessel was harvested 40-95 hours following initiation of methanol induction.
2. Mut_ protocol methaaol fed-batch fermentation (Run 657).
The first two phases of the methanol fed-batch fermentations of the Mut' strains were conducted as described for the Mut' strain fermentations in Example 4(A)(1). However, the methanol induction phases of the Mut+ and Mut' fermentation protocols differed in terms of the manner in which the methanol feeds were added to the cultures. In standard fermentations of the Mut' strain, the methanol feed rate was adjusted to maintain an excess of methanol in the medium which did not exceed 0.3% (as determined by gas chromatography). The methanol feed was initiated at 1 ml/hr and after two hours was increased in 10% increments every 30 minutes to a rate of 3 ml/hr which was maintained for the duration of the fermentation. The vessel was then harvested after 101 hours of growth of the strain on methanol.

~CT~JS 91 / d 6 4 5 2 2090969 -- - ~~ - - --' a p ~~,n!S ~ 0 J U L 1992 B. Ten-Liter Fermentations 1. Fad-batch a. Standard methanol teed rate (Runs 843, 959, 960, 962, 968 and 999).
A 15-liter fermentor containing 3.5 liters lOX basal salts and 220 g glycerol brought to a total volume of 5.5 liters by the addition of water was sterilized. After the fermentor had cooled, 24 ml PTM~ trace salts were added and the pH was adjusted to 5.0 with the addition of 28% ammonium hydroxide. The pH of the fermentation was maintained by addition of the same solution and foaming was controlled with the addition of a 5% solution of Struktol J673. Temperature was maintained at 30°C, and dissolved oxygen was maintained above 20% of saturation by regulating agitation, aeration, and reactor pressure or by supplementing the air feed with oxygen.
Inocula were prepared from cells grown overnight in bufferEd YNB- (1135 g/L KHZP04, 2.66 g/1 KZHP04, 6.7 g/L
yeast nitrogen base, pH 6) containing 2% glycerol. The fermentor was inoculated with 500-700 ml of the cultured cells which had grown to a concentration of 2-8 ODD, and the batch growth regimen was continued for 18-24 hours until glycerol was exhausted. At the point of glycerol exhaustion, indicated by an increase in dissolved oxygen, a glycerol feed (50$ w/v glycerol plus 12 ml/L PTM,) was initiated at 100 ml/hr. After four hours, the glycerol feed wa3 terminated and a methanol feed (100% methanol ~pl~s 12 ml/L PTM~) was initiated at a rate of 20 ml/hr.
After 3-4 hours of methanol feeding, the methanol feed rate was increased to 60 ml/hr and maintained at this rate for the remainder of the fermentation. The vessel was harvested approximately 72 hours after the initiation of the methanol feed.
In pH 5 fermentations, the pH of the culture was - 35 maintained at 5 throughout the fermentation. In low pH
v SUBSTITUTE SHEET

'~'O 92/04363 fermentations (i.e., pH 2.8-3.0), the set point of the pH
controller was adjusted to the target pH and the pH of the culture was allowed to decrease to the new set point as a result of cellular metabolism. In some fermentations, the set point of the pH controller was adjusted at the initiation of the glycerol feed, whereas in other fermentations, the pH set point was not lowered until initiation of the methanol feed.
b. Modified methanol feed rates (Runs 789, 810 and 906).
Additional 10-liter fermentations were conducted essentially as described in Example 4.B.l.a. except with modified methanol feed rates. In Run 789, the methanol feed rate was steadily increased as the cell density increased during the fermentation. After the initial three hours of methanol feeding at 20 ml/hr, the feed rate was increased to 30 ml/hr. The feed rate was then increased twice daily in proportion to the increase in cell mass.
In Run 810, the methanol feed rate was increased to 60 ml/hr after an initial four hours at 20 ml/hr. After 21 hours of methanol feeding, the feed rate was then increased to 100 ml/hr and maintained at the rate for the remainder of the fermentation (51 hours).
In Run 906, a methanol feed rate of 40 ml/hr was maintained for the first 60 hours of feeding, and then increased to 60 ml/hr for the remainder of the fermentation (90 hours total).
2. Continuous culture (Run 929).
In Run 929, the fermentation process was switched from a fed-batch to a continuous mode after 20 hours of methanol feeding. This was accomplished by initiating a 3X basal salts feed along with the methanol feed and maintaining a constant volume in the fermentor by continuous removal of whole broth (including cells). The V'U 92104363 P~.T/US91/06452 methanol feed was maintained at~a constant rate of 70 ml/hr and the be~s31 salts feed rate was set ;~t 70% of the total feed (i.e., methanol plus basal salts) rate. Thus, a total feed of 230 ml/hr, containing 30% methanol, was fed into the fermentor with a constant total volume of 8 liters. Continuous culture was carried out for 139 hours (total of 159 hours on methanol).
C. Results Samples (15-ml aliquots) of the fermentor culture were removed from the fermentor at various times throughout the course of the fermentation. Aliquots of each sample were centrifuged at 6500 x g for 5 min to separate broth and cells. The levels of the NH40H, antifoam, glycerol and methanol reservoirs were recorded at these time points. Methanol and ethanol concentrations in the supernatant were determined by gas chromatography using a PorapaicQ column (Alltech, Deerfield, IL).
In addition, the wet weight of the culture was determined as an indicator of cell growth in the fermentor. For this purpose a one-ml aliquot of the fermentor culture was centrifuged for four minutes in a microfuge, the supernatant was decanted, and the wet pellet was weighed.
The IGF-1 levels of the cell-free broth were determined by RIA, quantitative immunoblot or reverse-phase HPLC (see Example 5 for a description of each of these methods). The results of fermentations of the IGF-1-expressing strains of P. pastoris are provided in the Tables I-V.
TABLE I
Results of Quantitative Immunoblot Analyses of Cell-Free Broth from One-Liter Fermentations *Trade-mark A

2090969 PCT~~ 91 /06 4 52 .

EXPRESSION CELL DENSITY
CASSETTE GLYCEROLWET WEIGHT HOURS
Ig/L) RUN NAME NUMBER ,pH Ig) GLYCEROL MeOH MaOH (mg/L) 573 G +IGF201S11 5 100 224 308 49 11 57B G +IGF208S28 5 100 198 280 49 117 578 G+IGF208S26 3.5'150 328 360 101 284 598 G+IGF208S26 3 150 306 385 48 555 598 G+IGF208S28 3.5 150 298 363 64 455 605 G+IGF201S11 3 150 306 415 64 21 613 G+IGF208S26 3 150 304 344 70 489 ' The pH was decreased from 5 to 3.5 after 26 hours of growth on methanol.
The pH was decreased from 5 to the indicated pH at the initiation of the glycerol fed-batch phase.
TABhE II
Results of IncStar Antisera-Based RIAs of Cell-Free Broth from One-Liter Fermentations EXPRESSION CELL DENSITY
-CASSETTE GLYCEROLWET WEIGHT HOURS
- (g/Ll RUN NAME NUMBER pH Ig) GLYCEROL MeOH MeOH (mg/L) 598 G +IGF208S28 3 150 308 385 48 1850 805 G +IGF201S11 3 150 308 415 64 167 881 G +IMB208S18 3 150 267 320 73 1280 880 G +IM8202S22 3 150 270 430 73 740 891 G +IMB204S144 3 150 265 350 92 1400 894 G +IM8208S16 5 150 254 325 94 306 SUBSTITUTE SHEET

209 PCTIU~ 91 /06 4 5 2 ._ _ .

~PEA~US 30 JUL1992 TABLE III
Results of Nichols RIAs of Cell-Free Broth from One-Liter Fermentations EXPRESSION CELL DENSITY
CASSETTE GLYCEROLWET WEIGHT HOURS
(glLl RUN NAME NUMBER pH (g) GLYCEROL MeOH MeOH (mglL) 881 G +IMB20BS16 3 150 267 320 73 140 I, G +IM8202S22 3 150 270 430 73 64 891 G +IM8204S144 3 150 265 350 92 174 894 G +IM8206S16 5 150 254 325 94 15 755 G+IM8204S144 2.8 50 160 460 72 300 758 G+IMB204S144 2.8 175 314 445 72 315 757 G +IM8204S144 2.8 80 201 491 72 300 772 G+IM8206S1 8 2.8 85 181 401 71 335 i G+IM8204S144 2.8 90 200 377 70 250 775 G + IM8204S4 2.8 40 104 409 70 280 778 G+IM820BS1 8 2.8 45 118 490 71 355 -.-,j SUBSTITUTE SHEET

2090969 PCT~US 91106 452 ~ -tPEAIUS 3 0 J U L 1992 Results of HPLC Analyses of Cell-Free Broth from One-Liter Fermentations°
EXPRESSION CELL
CASSETTE GLYCEROLDENSITY HOURS
WET WEIGHT
(g/ll RUN NAME NUMBER pH (g) GLYCEROLMeOH MeOH (mg/U

805 G +IGF201S11 3 150 306 415 64 14 680 G +IMB202S22 3 150 270 430 73 39 891 G +IMB204S144 3 150 265 350 92 101 894 G +IMB206S1B 5 150 254 325 94- 3 757 G+IMB204S144 2.8 80 201 491 72 74 772 G+IMB208S18 2.8 85 181 401 71 121 934 M+IM8206S16 2.8 50 132 450 72 139 935 M+IM8208S16 5.0 50 123 363 72 167 943 G+IMB204S144 5.0 50 125 330 72 0 1039 G+IGF818S11' 3.0 50 158 320 49 16 1047 M +IGF818S11' 3.0 50 123 282 69 16 1048 G+IGF818S2- 1' 3.0 50 134 447 69 17 1049 G+IGF818S9multi' 3.0 50 131 380 69 44 1050 G+IGF818S11multi' 3.0 50 131 386 94 9 1053 M+IGF818S4multi' 3.0 50 143 393 70 33 1055 C + IG 1 ' 3.0 50 143 312 70 20 1088 C+IGF818S11 5 50 123 234 70 20 1087 M+IGF818S4multi' 5 50 132 331 70 50 1089 G+IGF818S11multi' 5 50 132 323 70 0 ' Cell-tree broth was peased through a cation exchange chromatography column prior to analysis by HPLC~sss Example 4C1.
Concentletion of authentic IGF-1 (i.e., intact, correctly folded monomeric IGF-1).
' The number of expression cassettes contained in these strains is estimated based on IGF-1 expression levels.
SUBSTITUTE SHEET

PCT~IJS 91 / 0 6 ~+ 5 2 IpEA~US 3 0 JUL1992 TABhE 0 Ten-Liter Fermentations ExPRESSION CELL
CASSETTE G~YCERO~DENSITY HOURS
AooED WET WEIGHT
igll) STRAIN NuINeER ON IGF-1 RUN NAME pH GLYCEROLMeOH MeOH (mg/L) 789 G + IMB204S4 2.8'420 107 584 60 290'190 810 G+IM8204S144 2.8'420 181 595 51 305'/75 843 G+IMB204S144~ 2.8'420 177 482 72 308'/103 908 G+IMB204S144 2.8'420 178 471 90 300'!93 929 G + IMB204S144 2.8'420 180 400-480159 55-70 959 M+IM820BS1 6 5.0 420 217 383 75 54 960 M+IM8206S1 6 2.8'420 184 387 72 138 962 G+IMB204S144 2.8'420 188 536 72 77 999 M+IMB206S1 6 3.0'420 191 398 72 131 ' The pH was decreased from 5 to 2.8 et the initiation of the glycerol fed-batch phase.
Continuous culture.
Determined by Nichola RIA of cell-free broth (see Example 5A).
Authentic IGF-1 levels determined by HPLC analysis of pretreated cell-free broth (see Example 5A).
' The pH was decreased from 5 to the indicated pH at the initiation of the methanol fed-batch phase.
As shown in Tables I- V, the four different assays used to measure IGF-1 levels yielded inconsistent results. As discussed in Example 5, the different assays most likely detect different combinations of the several 5~- species of IGF-1 present in varying concentrations in P.
'pastoris broth, including: intact correctly folded monomeric IGF-1, referred to as authentic IGF-1;
multimeric IGF-1: nicked IGF-1; and incorrectly folded IGF-1; thereby providing differing values for IGF-1 concentrations of the broth. It is possible, however, to compare the IGF-1 levels of different fermentations when they are measured by the same assay. Such analyses yield the following conclusions.
SUBSTITUTE SHEET

~fJ 92/04363 PGT/US91/06452 20909~9 The results of quantitative immunoblot analyses of fermentations of the six-copy strain G+IGF206S2 (Runs 576, 578, 596, 598 and 613) and the one-copy strain G+IGF201S1 (Runs 573 and 605) are shown in Table I. All samples quantified by immunoblot were first reduced with dithiothreitol (DTT), which disrupts disulfide bonds holding together multimeric or nicked IGF-1. Therefore, because only the protein that co-migrated with IGF-1 standard was quantified in this assay (see Example 5A), this method detected, but did not distinguish authentic monomer IGF-1, misfolded monomer IGF-1 and multimer IGF-1. The results shown in Table I demonstrate that expression cassette copy number has an effect on the level of immunoreactive IGF-1 contained in the broth, IGF-1 levels increase as the number of copies of the expression cassette contained in the strain increases (compare run 605 of the one-copy strain (21 mg/1) and Run 613 of the six-copy strain (489 mg/1)). A comparison of the IGF-1 levels of broth from fermentations conducted entirely at pH 5 (Runs 573 and 576) and the levels from fermentations conducted at lower pH (pH decreased from 5 to 3-3.5 at either initiation of glycerol feed (Runs 598, 605 and 613) or during methanol fed batch phase (Run 578)) reveals that the greatest amount of immunoreactive IGF-1 was obtained when the fermentation was conducted at lower pH. Furthermore, the time at which the pH was lowered from 5 to <3.5 also had an effect on IGF-1 levels. Lower IGF-1 levels were obtained when the pH of the fermentation was decreased later in the fermentation, i.e., during the methanol fed batch phase (compare Runs 598 and 578). The results of several fermentations of the six-copy strain conducted according to the same protocol (addition of 150 g of glycerol, 60-70 hours of methanol feeding, and decrease in pH from 5 to 3 at the initiation of glycerol feeding) demonstrated the WO 92/04363 PCT/US91/0645'' w 2090969.

reproducibility of the IGF-1 production process, since all fermentations typically yielded between approximately 350-480 mg/1).
The results of Incstar antisera-based RIAs (Table II; see Example 5A for a description of this assay) of broth from fermentations of the six-copy strain G+IMB206S1 (Run 661), the four-copy strain G+IMB204S14 (Run 691), the two-copy strain G+IMB202S2 (Run 680) and the one-copy strain G+IGF201S1 (Run 605) further illustrate the dramatic effect of expression cassette copy number on IGF-1 expression by P. pastoris. The four-and six-copy strains produced similar levels of RIA-reactive IGF-1 (1400 and 1280 mg/1, respectively) and 2-and nearly 10-fold more IGF-1 than the two- and one-copy strains (740 and 167 mg/l, respectively). Furthermore, the Mut+ six-copy strain G+IMB206S1 appeared to be superior to the Mut- six-copy strain (G-IMB206S3) in terms of IGF-1 production capability (compare Runs 661 and 657). Incstar antisera-based RIAs of broth from Runs 661 and 694 (strain G+IMB206S1) were in agreement with the finding that pH 5 fermentations yield significantly lower levels of IGF-1 than low pH fermentations.
Although Nichols RIAs consistently yielded lower values for IGF-1 expression levels than the Incstar antisera-based RIA, the relative IGF-1 levels from different fermentations, as measured by Nichols RIA, were in agreement with those as measured by Incstar antisera-based RIAs. For instance, a comparison of the IGF-1 levels of fermentations of the two-, four-, and six-copy strains G+IMB202S2 (Run 680), G+IMB204S14 (Run 691) and G+IMB206S1 (Run 661), respectively, as determined by Nichols RIA, further demonstrated the effect of expression cassette copy number on IGF-1 expression levels. Furthermore, the level of IGF-1 product determined by Nichols RIA was significantly higher in - ~ r PCT~J~ 91 /06 4 52 . _ 2 0 9 0 9 6 9 ~p~US 3 0 J U L 1992 fermentations conducted at lower pH: compare Run 694 (pH
5; 15 mg/1) and Run 661 (pH 3; 140 mg/1) and Run 776 (pH
2.8; 355 mg/1). Finally, it appears that changing the amount of glycerol added during the glycerol batch and fed batch phases had little effect on the final IGF-1 concentration from fermentations of the four-and six-copy strains (compare Runs 772 (85 g glycerol added) and 776 (45 g glycerol added) which both produced approximately 345 mg/1; also compare Runs 755, 756, 757, 773 and 775 (glycerol added carried from 40-175 g) which yielded 250-300 mg/1.
HPLC analysis (see Example 5C) of fermentation broth permitted distinction and accurate measurement of each species of IGF-1 present in the broth. Table IV lists the levels of authentic IGF-1, as determined by reverse-phase HPLC of broth from one-liter fermentations of IGF-1-expressing strains of _P. eastoris. The results of these HPLC analyses are consistent with those of RIA and immunoblot analyses of broth in terms of the relative IGF-1 levels in broth from different strains grown under different conditions. As shown in Table IV, expression cassette copy number clearly influences IGF-1 expression levels (compare IGF-1 levels of broth from a one-copy strain, G+IGF201S1 (14 mg/1, Run 605), a two-copy strain, G+IMB202S2 (39 mg/l, Run 680), a four-copy strain, G+IMB204S14 (101 mg/1, Run 691) and a six-copy strain, ;G+~M8206S1-(121 mg/1, Run 772)). In addition, the results .in'Table IV demonstrate the effect of pH on production of IGF-1 in ~. pastoris; fermentations of strains G+IMH206S1 and~G+IMB204S14 conducted at pH 5 (Runs 694 and 943, respectively) yielded essentially no authentic IGF-1 (or any~form of IGF-1); whereas 121 and 74 mg of IGF-1/1 were produced in low pH fermentations of these strains (Runs 772 and 757, respectively, pH was decreased from 5 to 2.8 at the initiation of the glycerol fed batch phase). The SUBS'fiTUTE SHEET

Y~ ~2 S ~ ~ tS o~~ I-~ ~GC a nir+ ~J~SiS
-__11_ _r ..rW ~ _~_1__~_~ L~~Lt~ r~~__ ~~___~_1 _!_r~ 1 low pH fermentations of strain G+IMB204S14 in which the amount of glycerol added to the fermentor varied between 40 and 175 g confirmed the finding that differing glycerol levels in this range had no effect on IGF-1 production; all fermentations yielded 70-110 mg/1.
The broth from one-liter fermentations of the protease-deficient (pep4) strain M+IMB206S1 was also analyzed by HPLC. As shown in Table IV, pH 5 and low pH
fermentations (Runs 935 and 934, respectively) of this strain yielded similar high levels (167 and 139 mg/1, respectively) of authentic IGF-1. Furthermore, the level of authentic IGF-1 produced in these fermentations of strain M+IMB206S1 was approximately 60% higher than that produced in pH 2.8 fermentations of the four-copy strain PEP4 G+IMB204S14 (Run 757; 74 mg/1) and approximately 15%
higher than that produced in pH 2.8 fermentations of the six-copy PEP4 strain G+IMB206S1 (Run 772; 121 mg/1).
The results of HPLC analysis of broth from one-liter fermentations of strains containing one or more aMF
prepro-lys-arg-glu-ala-IGF-1 expression cassettes are also shown in Table IV. These strains showed two different levels of IGF-1 expression. Strains G+IGF816S1 and G+IGF816S2 yielded similar levels of authentic IGF-1 (approximately 16 mg/1) in two identical fermentations (Runs 1039 and 1048). In contrast, strains G+IGF816S9 and G+IGF816S11 produced three-fold more authentic IGF-1 (approximately 45 mg/1) in Runs 1049 and 1050.
Therefore, it appears that strains G+IGF816S9 and G+IGF816S11 fortuitously incorporated 2-3 copies of the aMF prepro-lys-arg-glu-ala-IGF-1 expression cassette into their genomes; whereas strains G+IGF816S1 and G+IGF816S2 apparently contain a single copy of the expression cassette.
Broth from fermentations of protease-deficient ~'.a, 92/04363 PCT/US91/06452 strains M+IGF816S1 (pep4), M+IGF816S4 (pep4) and C+IGF816S1 ( e~,4, prb=1) containing one or more aMF
prepro-lys-arg-glu-ala-IGF-1 expression cassettes (see Example 2C) also was analyzed by HPLC. Based on the authentic IGF-1 levels of the broth from low pH
fermentations of strains M+IGF816S1 and C+IGF816S1 (Runs 1047 and 1055, 16-20 mg/1) as compared to those of the broth from low pH fermentations of strain M+IGF816S4 (Run 1053, 33 mg/1), it appears that strains M+IGF816S1 and C+IGF816S1 contain one copy of the expression cassette;
whereas M+IGF816S4 contains two or more expression cassettes. It is noteworthy that pH 5 fermentations of these strains (Runs 1066 and 1067) yielded at least as much authentic IGF-1 as low pH fermentations of these strains. In contrast, pH 5 fermentations of a PEP4 strain containing multiple copies of the aMF-lys-arg-glu-ala-IGF-1 expression cassette (Run 1069) yielded no authentic IGF-1.
As shown in Table V, high-level production of IGF-1 2o by P_. pastoris was readily scaled up to the 10-liter fermentation level. A comparison of low pH 10-liter fermentations of the four-copy strain G+IMB204S14 conducted with varying methanol feed rates (Runs 789, 810, 843, and 906) revealed similar levels of authentic IGF-1 production. A significant level of authentic IGF-1 (54 mg/1) was produced in a pH 5.0 10-liter fermentation of the pep4 strain M+IM8206S1 (Run 959). This level was increased in 10-liter fermentations of this strain that were initiated at pH 5 and then conducted at pH 2.8 or 3.0 during the entire methanol fed batch phase (Runs 960 and 999, respectively). A similar fermentation of strain G+IMB204S14 (Run 962) yielded IGF-1 levels equivalent to those obtained in a fermentation in which the initial pH
of 5 was decreased to 2.8 upon initiation of the glycerol fed batch phase (Runs 789, 810, 843 and 906). Continuous wo yz~oa3~~ ~crius9tio~,asz culture of strain G+IMB204S14 (Run 929) produced between 5!~ and 70 mg/1 of authentic IGF-1 at a rate of 230 ml of culture media per hour. Therefore, the productivity of continuous culture fermentations was slightly higher than that of fed-batch fermentations.
EuAMPLE 5 MEDIUM.
A. RIA Assay.
A 1:5000 final dilution of rabbit anti-human IGF-1 antisera (Incstar anti-somatomedin C antibody, catalog #22275), 10,000-12,000 cpm of ~zSI-IGF-1 (Incstar catalog #22303), and various dilutions of recombinant human IGF-1 standard (purchased from Imcera and quantified by amino acid analysis) or the unknown broth solution were incubated overnight at 4°C in a final volume of 0.5 ml in 12 x 75 mm polystyrene tubes. At the end of the incubation, 100 ~1 of Pansorbin (working dilution of 1:40) was added to the tubes and incubated for 15 minutes at room temperature. Two milliliters of RIA buffer (50 mM NaP04, 0.1% BSA, 0.1% NaN3, and 0.1% Triton X-100, pH

7.4) were added to each tube before centrifugation at 3200 rpm for 68 minutes at 4°C in a Beckman J6M
centrifuge. Following centrifugation, the supernatant was decanted and the radioactivity associated with the pelleted material was determined in a gamma counter.
Alternatively, some of the .samples were assayed using a commercial RIA (Nichols Institute Diagnostic; San Juan Capistrano, CA). The Nichols assay consistently measured lower levels for IGF-1 than the RIA described above (see data above from Runs 661, 680, 691 and 694).
This apparent discrepancy probably arises because the antibody used in the RIA described above detects both monomer and multimer IGF-1, whereas the antibody used in the Nichols assay most likely detects monomer IGF-1 only.
*Trade-mark ,.
A

PCTIU~ 91/06452. _ 2 0 9 0 9 6 9 ~pFAIUS 3 4 JUL 1992 H. Western blot Immunoblot analysis of P. pastoris fermentation broth was used to qualitatively and quantitatively evaluate IGF-1 production in ~ pastoris.
Several dilutions of recombinant human IGF-1 standard (Imcera) and samples of cell-free broth obtained at the conclusion of selected fermentations of the Mut+
and Mut- IGF-1-expressing strains of Pichia pastoris were analyzed by quantitative western blots to estimate the amount of immuno-reactive IGF-1 contained in the broth.
- The proteins were first separated by Tricine SDS-PAGE
(Schagger, H. and von Jagow, G. (1987) Anal. Biochem.
~: 304-310) and then transferred to 0.1 ~,m nitrocellulose by electroblot in a solution of Towbin buffer (25 mM Tris-HC1, pH 8.3, 190 mM glycine, 20%
methanol) for at least 90 minutes at 20 V/cm. After the proteins were transferred onto nitrocellulose, the filter was incubated for Dne hour in blocking buffer (0.25%
gelatin, phosphate-buffered saline, 0.05% Tween 20, 0.02%
sodium azide). Rabbit anti-IGF-1 antisera l0A was diluted 1:2000 with blocking buffer and incubated with the filter for a minimum of two hours. Antibody l0A was raised against a synthetic peptide corresponding to the last 14 amino acids of the carboxy terminus of human IGF-25. 1, which was conjugated to human a-globulin. The 1''~~ant~sera was adsorbed to a-globulin prior to use. The .filter wa~ washed with blocking buffer for an hour and incubated with ~~I-Protein A (0.02 ~cCi/ml) for 45 minutes.
After one hour of washing with blocking buffer, the filter was air dried and exposed to X-ray film with an intensifying screen at -75°C. To estimate the IGF-1 . content of broth, the intensities of bands corresponding to IGF-1 in lanes containing broth were compared to the intensities of bands corresponding to different known amounts of IGF-1 standard. Autoradiographs were used as .;
SUBSTITUTE SHEET

PCTfiJS 91 / 0 6 4 5 2 .
2090969 ~p~~US
3 0 J U l 1992 templates and the portions of the gels corresponding to bands were excised and the radioactivity in the bands counted in a scintillation counter. A standard curve was prepared by plotting the cpm of bands containing known amounts of IGF-1 versus IGF-1 concentration and the IGF-1 content of the broth was estimated by comparison with the standard curve.
C. Quantitative Rev~rse-Phase High-Performance Liquid Chromatography (HPLC) 1. HPLC System A Waters (Bedford, MA) 600 solvent delivery system, Waters Model 481 Lambda Max variable wavelength detector, Wisp 7108 auto-injector and a Schimadzu Crom-Pac integrator (Cole Scientific, Moorepark, CA) constituted the HPLC system. A Vydac C4 column (0.46 X 5 cm) with a guard column was used to resolve components of the Pichia pastoris-produced IGF-1 preparations. Broth samples that had been pretreated as described in Example 6.C.2. below were loaded onto the column at a flow rate of 1 ml/minute and were eluted in a trifluoroacetic acid (TFA)/acetonitrile-TFA gradient. The eluant was prepared by using mobile phase A (0.1% TFA) to dilute mobile phase B (95% acetonitrile, 5% water, 0.1% TFA). A 1%/minute gradient of 25% to 42% mobile phase B was passed through the column during a period of 17 minutes at a flow rate of 1 ml/minute to elute the material that had been loaded :onto the column. The column was then regenerated with 100% mobile phase B at a flow rate of 2 ml/minute for 4 minutes followed by 25% mobile phase B for 4 minutes at 2 ml/iainute. The flow rate was then reduced to 1 ml/minute and the column was equilibrated at the initial starting conditions (25% mobile phase B) for 2 minutes before ,, suBSmv~ sH~' reinjection with another sample to t~e analyzed. The detector was set at 0.05 absorbance units fu:.l scale (AUFS) and a wavelength of 215 nm was used for maximum sensitivity.
2. Pretreatment of crude fermentation broth.
P. Pastoris-produced IGF-1 exists as several forms in the broth of fermentations of IGF-1-expressing P.
pastoris strains. HPLC analysis of crude cell-free broth from fermentations of IGF-1-expressing P. pastoris does not adequately resolve the various IGF-1 species. In order to distinguish these IGF-1 species by HPLC, the authentic P. pastoris proteins must be removed from the broth by a small-scale cation exchange chromatography step. Several cation exchange systems were tested for this purpose: sulfylpropyl cation exchange capsules (FMC
(Pinebrook, NJ) and Cuno (Meriden, CT)) and the use of a bulk cation exchanger (e. g., S-Sepharose Fast Flow (Pharmacia, Uppsala, Sweden), SP-Spherodex (IBF, Columbia, MD), or Toyopearl SP650M and SP550C (Toso Haas, Woburn, MA)) in a 2-ml disposable polypropylene column (0.8 x 4 cm, BioRad) with an integrated 10-ml reservoir. Any of these systems yielded satisfactory results. The two systems routinely used to pretreat crude broth for quantitative HPLC analysis of the IGF-1 levels employed the Cuno cation exchange capsule or the SP-Spherodex or SP550C cation exchangers in a column format.
The cr'Gde broth was cleaned up by cation exchange chromatography and injected directly onto an HPLC column.
The resulting chromatogram clearly resolved the different IGF-1 species in broth.
a. Pretreatment using catioa exchange capsules.
The Cuno cation exchange capsule is a 25-mm disk.
It was first washed with a 4 ml 0.2 M acetic acid, then *Trade-mark A

2090969 PCTRJS 91 /06 4 52 .
n Eau s 3 0 ~ a ~ X992 equilibrated with 4 ml 0.02 M acetic acid. A volume of crude broth (1-10 ml) was diluted 1:2 with 0.02 M acetic acid and loaded onto the disk. After loading, the disk was washed with 1.5 ml 0.02 M acetic acid and the IGF-1 was eluted with 4 ml 0.02 M sodium acetate, pH 5.5, plus 1 M NaCl. The first 1.5 ml of eluate contained 75-80% of the total IGF-1 and was usually the only elution volume collected. The capsule could then be regenerated by washing with 4 ml of 100% methanol.
b. Pr~-trsatment using bulk cation exchanger in a column format.
To the disposable column, 0.25 ml of prehydrated cation exchanger was added. The absorbent was first washed with 2 ml of 0.2 M acetic acid, then equilibrated with 2 ml of 0.02 M acetic acid. A volume of broth (1 ml) was loaded onto the column which was then washed with 1 ml-o_f 0.02 M acetic acid. All buffers and broth were carefully added to the column in an effort to avoid disruption of the column bed. Although some suspension of the absorbent usually occurred during addition of liquid to the column, it was not detrimental to sample binding or eluting. Broth samples and buffer were allowed to flow through the column by gravity. The IGF-1 was eluted with 2 ml of a 0.05 M sodium acetate, pH
~- 5.5, buffer containing 1 M NaCl. The first milliliter of r °eluateWontained 80-90% of the total IGF-1, eluted by the .2 inl of elution buffer. Periodically (approximately every 5-10 samples), the column was regenerated after the salt elution with a 50% methanol wash. Less often, the column was also regenerated with 0.5 M NaOH. These SUBSTITUTE SHEET

V1'U 92/04363 PCT/US91/OG452 columns retain their selective binding properties through many_ successive u:~es.
3. Measurements of IGF-1 concentrations.
In order to measure the concentrations of Pichia-produced IGF-1 by HPLC, known amounts of standard IGF-1 (Amgen, Thousand Oaks, CA) and authentic Pichia-produced IGF-1, which was purified as described in copending Canadian Application No. 2,098,183) and quantified by amino acid composition analysis were injected onto the l0 HPLC column and the area under the corresponding peaks in the chromatogram was measured. A standard curve was generated by plotting area versus micrograms of IGF-1 loaded onto the HPLC column. A conversion factor for use in converting the area under HPLC chromatogram peaks to IGF-1 concentration was calculated from the standard curve. Using this information, it was possible to determine the concentration of authentic IGF-1 in a pretreated broth sample by measuring the area under the corresponding peak on the chromatogram from HPLC analysis of the sample. This conversion factor was also used to estimate the approximate concentration of other IGF-1 species as well. The absolute concentrations of each of these other species, however, may vary depending on differences in their specific conversion factors.

MEDIOM
Pichia-produced IGF-1 was characterized by immunoblot, SDS-PAGE, HPLC, and protein sequence analyses. These methods and the results of these analyses are described in this example.
A. Immunoblot Analysis Several characteristics of the IGF-1 produced in fermentations of IGF-1-expressing ~ pastoLris were A' WO 92/04363 PCT/US91 /0645' ~'""' ~09U969 ascertained from immunoblot analysis of reduced (by adding dithiothreitol (DTT)) and non-reduced (no added DTT) samples of broth (performed as described in Example 5B) .
First, if the broth samples were not reduced prior to electrophoresis, the IGF-1 migrated as several forms which appeared to be a monomer, dimer and various multimers of IGF-1. The ratio or profile of monomer to multimer forms of secreted IGF-1 did not seem to be affected by copy number. Second, the immunoblot of reduced samples revealed the absence of most of the higher molecular weight immunoreactive species that were evident in the nonreduced broth samples, and revealed the presence of a protein that co-migrates with IGF-1 standard (monomer) as well as immunoreactive protein which migrates to a position slightly below that of standard IGF-1. N-terminal protein sequence analysis (Example 6.d.) of this lower molecular weight immunoreactive species demonstrated that it begins with residue 25 of mature human IGF-1 and, thus, is a proteolytic fragment of IGF-1. Since this lower molecular weight species is not seen under non-reducing conditions, the IGF-1 molecule may be nicked only between residues 24 and 25, i.e., not completely fragmented, with the disulfide bonds at Cys-6/Cys-48 and Cys-18/Cys-61 of IGF-1 holding the peptide fragments that contain amino acids 1-24 and 25-70 together. Further, the decreased glycerol fermentation (Run 755) produced wholly intact IGF-1 molecules (no degradation), while the extended glycerol feed fermentation (Run 756) demonstrated a low level of degraded product. The degradation appeared to be slightly higher with the six-copy strain (Run 776) than with the four-copy strain (Run 775).

W'O 92/04363 PCT/US91 /06452 IGF-1 produced in the fermentation of the six-copy Mut~ strain (Run 657) w~3~ also analyzed by SDS-PAGE and Western blot. The IGF-1 band patterns of reduced and non-reduced broth from the fermentation of the six-copy Mut- strain are similar to the band patterns of reduced and non-reduced broth from the fermentations of the six-copy and four-copy Mut~ strains.
B. Qualitative HPLC Analysis Crude _P. pastoris fermentation broth was also analyzed by HPLC (see Example 5.C.1 for a description of the HPLC system used in this analysis). Direct injection of crude broth onto the HPLC column did not result in a chromatogram with distinct peaks. In order to analyze the components of crude broth by HPLC, it was necessary to remove endogenous P. pastoris contaminants from the broth using small-scale cation exchange chromatography, as described in Example S.C.
The eluate obtained after small-scale can on exchange chromatography of cell-free broth from a low pH
fermentation of an IGF-1-expressing strain of ~ ~astoris contains all the IGF-1 polymers, including nicked, misfolded, multimeric and authentic IGF-1 which can be resolved by the HPLC protocol described in Example 5C.
The different peaks detected in the chromatogram from HPLC analysis of broth that had been subjected to canon exchange chromatography correspond to different forms of IGF-1 produced in recombinant ~ pastoris fermentations.
The identity of the proteins corresponding to these peaks was established in HPLC, SDS-PAGE, immunoblots, gel filtration, cation exchange and hydrophobic interaction chromatography (performed as described in copending Canadian Application No. 2,098,183).
A chromatogram from HPLC analysis (conducted as described in Example 5C.) of broth from a fermentation of an IGF-1-expressing P. pastoris strain contains a peak A

corresponding to protein that eluted from the HPLC c~>lumn at approximately 10 minutes, which represents correc~:ly folded, intact IGF-1 monomer. The identity of this protein was confirmed initially on the basis of its elution time by HPLC, which is identical to that of standard recombinant IGF-1 (Amgen, Thousand Oaks, CA).
Furthermore, the protein with an elution time of approximately l0 minutes was purified (see copending Canadian Application No. 2,098,183) and subjected to additional analyses. SDS-PAGE analysis of reduced and non-reduced samples of the purified protein yielded identical results, revealing that it was a 7.7-kDa intact protein that co-migrated with IGF-1 standard. Immunoblot analysis of the purified protein demonstrated that it is reactive with an antibody generated against the last 14 amino acids of IGF-1. Gel filtration chromatography of the purified protein revealed that it elutes as expected for an IGF-1 monomer of the correct size. Finally, amino acid analysis confirmed that the amino acid ratios of the purified protein correspond to those of standard IGF-1.
Protein sequence analysis showed that the complete amino acid sequence of the purified protein is identical to that of authentic IGF-1.
The protein that elutes from the HPLC column at approximately 8.6 minutes was tentatively identified as misfolded IGF-1. This protein was isolated by HPLC and hydrophobic interaction chromatography and characterized by SDS-PAGE, immunoblot and protein sequence analysis.
SDS-PAGE analysis of reduced and non-reduced samples of this protein demonstrated that this form migrated with authentic monomeric IGF-1 and that it was not a nicked form of IGF-1. Immunoblot analysis of this protein using an antibody directed against the C-terminus of IGF-1 A

N,o 92ina36~ rcrms9 mobasz showed 'that it was immunoreactive. Amino-terminal protein sequencing of this protein also confirmed that the molecule was intact since only one amino-terminal sequence could be identified. These results suggest that this form is a misfolded species of IGF-1.
The proteins that elute from an HPLC column at 10.5-11.5 minutes were identified as nicked or degraded forms of IGF-1 (i.e., IGF-1 molecules containing two or more peptide fragments, generated by cleavage of one or more peptide bonds and held together by disulfide bonds).
There appears to be at least two peaks by HPLC analysis of cleaned up broth that correspond to the nicked IGF-1.
The protein represented by the major peak (eluting at 10.7 minutes) was isolated during the S-5epharose cation exchange step of the IGF-1 purification process (see co-pending Canadian Application No. 2,098,183). SDS-PAGE
analysis of non-reduced samples of this isolated species revealed that it co-migrated with IGF-1 standard and appeared as a single band. Gels of reduced samples of this protein, however, exhibited a doublet representing two peptides of approximately 3-4 kDa each (approximately half the size of intact IGF-1). The position of this doublet in the gel corresponded to that of the lower of the two bands detected below the band representing intact IGF-1 in gels of reduced samples of crude broth. These results indicate that this molecule is monomeric IGF-1 cleaved or nicked and held together by disulfide bonds.
Amino-terminal protein sequence analysis of the protein confirmed that the molecule is nicked prior to residue 40 of IGF-1. Immunoblot analysis of reduced and non-reduced samples of this isolated nicked IGF-1 molecule revealed that it is less reactive than intact IGF-1 with the antibody directed against the C-terminus of IGF-1.
Two additional nicked species were identified in protein sequence analysis of IGF-1 recovered from the A

first cation exchange chromatography step of the purification process (see copening Canadian Application No. 2,098,183). Either or both of these species could correspond to the minor peak in the HPLC chromatogram of cleaned up cell-free broth (protein eluting at 11 minutes). The amino-terminus of the C-terminal fragment of one of these nicked molecules begins at residue 25 of IGF-1. The amino terminus of the C-terminal fragment of the other nicked species detected in the broth begins at residue 14 of IGF-1.
The last set of proteins detected in HPLC analysis of cell-free broth, which elute from the HPLC column after 11.5-16 minutes, appears to be disulfide-bonded multimer forms of IGF-1. The presence of disulfide-bonded IGF-1 multimers in P. gastoris broth was indicated in SDS-PAGE gels of broth and immunoblots of the gels.
The putative multimers migrated as IGF-1 dimers and trimers on non-reduced SDS-PAGE gels and were reactive with antibody directed against the C-terminus of IGF-1.
When these multimers were reduced, they co-migrated with standard monomer IGF-1 in SDS-PAGE gels, which indicates that they contain disulfide-bonded IGF-1 monomers. The amino acid sequence generated in protein sequence analysis of gel-isolated dimer was identical to the N-terminus of authentic IGF-1. Furthermore, multimer IGF-1 (apparent dimer and trimer species) were isolated on a gel filtration column and analyzed by HPLC. The isolated multimer eluted from the column at 12-14 minutes, which corresponds to the elution times of the proteins in broth that were proposed to be multimers of IGF-1.

CHARl~CTERIZATIOId OF PURIFIED lrUTHENTIC PICBIA-PRODUCED
IGF-1.
Correctly folded, intact, monomeric IGF-1 (authentic IGF-1) was purified from the broth of IGF-1-expressing A

_" _ strains G+IMB204S14 and'M+IMB206S1 using a combination of cation exchange chromatography, hydrophobic interaction chromatography and gel filtration chromatography methods.
The purification procedure is described in detail in co-y pending Canadian Application No. 2,098,183. The purified material was characterized by HPLC, gel filtration chromatography, SDS-PAGE and immunoblot analysis, and amino acid composition and sequence analysis.
A. HPLC Analysis Several different dilutions of the purified preparation were subjected to HPLC using the system and procedure described in Example SC. A single major peak was detected in the resulting chromatogram which represented a protein with an elution time identical to that of IGF-1 standard.
B. Gel Filtration Chromatography 1. Procedure A Superdex*75 HR 10/30 gel filtration column (Pharmacia) with a bed volume of 24 ml (10 x 300 mm) was used in gel filtration chromatography methods of characterizing IGF-1 on the basis of size. The HPLC
system described in Example 5C was modified for use with this column. The detector was set at a wavelength of 280 nm and a sensitivity of 0.02 AUFS. The gel filtration chromatography column buffer contained 0.05 M ammonium acetate, pH 6. Buffer was flowed through the column at a flow rate of 0.5 ml/minute.
2. R~aults The chromatogram from gel filtration chromatography analysis of 200 ug of purified IGF-1 yielded a single peak representing protein that co-eluted with IGF-1 standard.
C. SD8-PAGE and Immunoblot Analysis.
The purified material was examined by SDS-PAGE and immunoblot methods as described in Example 5B. Only a *Trade-mark A}

WO 92/04363 PCT/ US91 /064 r"~"' -78_ 2pgfl969 single band was detected in both silver-stained gels and immunoblots of reduced and non-reduced samples of the material containing 1 ~g of protein or less. This protein co-migrated with IGF-1 standard.
D. Amino Acid composition Purified IGF-1 was acid hydrolyzed, and the amino acids were characterized on a Beckman (Palo Alto, CA) 6300 Amino Acid Analyzer. To acid hydrolyze the IGF-1 protein, carefully measured volumes of purified IGF-1 solution were added to 6 x 50 mm glass tubes and dried in a Savant (Farmingdale, NY) Speed Vac. These tubes were placed in a reaction flask containing ~0.5 ml 6 N HC1.
Oxidation was minimized by applying a vacuum and sealing the flask. The flask was placed overnight in a 110°C
oven, and the protein was hydrolyzed by the HC1 vapors.
Following hydrolysis, the reaction flask was cooled to room temperature, and the hydrolyzate was removed.
Any HC1 that may have condensed in the tubes was removed by drying the tubes in a Speed Vac. The free amino acids were dissolved in a minimum of 100 ul Beckman Amino Acid Sample Dilution Buffer, Na-S, for loading in the 50-ul loop of the analyzer. A Nelson (Cupertino, CA) 3000 Series Chromatography Data System was used to measure concentrations by comparing the integrated chromatogram of amino acid standard solutions and of the resuspended, hydrolyzed sample.
Table VI shows the amino acid ratios estimated for the purified IGF-1 and the actual published amino acid ratios for human IGF-1. The estimated and published ratios are in close agreement, and slight deviations in the estimated and published compositions are within the expected limits of this analysis.

20909 b9 p~'~'~S 91 /06 4 52 r TABLE VI
Purified IGF-1 Amino Acid Composition Analysis Data Published Amino Acid Composition' Experimental Asp (+Asn) 5 5.8 Thr 3 3.1 Ser 5 4.9 Glu(+Gln) 6 6.9 Pro 5 3.0 Gly 7 7.4 Ala 6 6.8 Cys 6 5.5 Val 3 ~ 2.6 _ Met 1 0.9 Ile 1 0.66 Leu 6 7.1 Tyr 3 2.1 Phe 4 3.6 His 0 0 Lys 3 3.3 Arg 6 6.7 Trp 0 'Derived from nucleotide by Rotwein et al.
sequence published (1986) 3 0 J. Biol. Chem. 261:4828 .

E. Amino Ac id 8~qu~nc~

i. Pro ceduro To determine the N-terminal amino acid sequence of the IGF-1 purified from the broth strain M+IMB206S1, of samples of this ma terial were loadeddirectly onto an Applied (Foster City, CA) 470/120 Gas Phase Biosystems }Protein~Sequencer. Sequencing was performed according to the methods descri bed by Hunkapillerand Hood (Science x:650 (1983)).

~ To determine the entire amino acid sequence of IGF-1 purified from the broth of strain IMB204S14, the G+

material was seque nced from the N-terminal amino acid through as much of the remainder of the protein sequence as' possible. This analysis generated the sequence of the first 59 residues of the purified protein. Because the k SUBSTITU'T'E SHEET

2090969 pC'~',~US 91/06452--~ _ IPEA/US 3 0 J U l 1992 amino acid at residue 59 is a methionine residue, and cyanogen bromide cleaves proteins after methionine residues, it was possible to isolate the peptide con-taining the C-terminal 11 amino acids (residues 60-70) of purified IGF-1 for use in completing the protein sequence analysis of the purified material. The C-terminal 11 amino acids of the purified IGF-1 were obtained as a peptide fragment isolated from cyanogen bromide-treated IGF-1 by HPLC, using the same C4 column as described in Example 5C. This fragment was loaded onto the protein sequences to generate the sequence of the C-terminal amino acids (amino acids 60-70).
The N-terminal amino acid sequences of authentic IGF-1 from the broth of fermentations of strains G+IGF816S1 and G+IGF816S11 were also determined. The broth was pre-treated as described in Example 5C and sub-jected to HPLC (also as described in Example 5C). The fraction that eluted from the HPLC column with an elution time identical to that of standard IGF-1 was applied to an SP-Spherodex column to remove the HPLC solvent (acetoni-trile and TFA). The eluate fraction containing the IGF-1 was then loaded directly onto the protein sequences.
2. Results The first 30 amino acids of the IGF-1 purified from the broth of strain M+IMB206S1 and the entire amino acid sequence of IGF-1 purified from broth of strain ~G~-IMB204S14 were identical to the corresponding amino acids of human IGF-1, Sequence ID No. 1 (i-e., residues 2-31) .
~ Similarly, the first 15 amino acids of the IGF-1 from the broth of strains G+IGF816S1 and G+IGF816S11 were identical to the corresponding amino acids of human IGF-1. These results confirm that mature IGF-1 is correctly processed from the aMF prepro-lys-arg-IGF-1 or aMF
- 35 prepro-lys-arg-glu-ala-IGF-1 precursor in ~ pastoris.
,, SUBSTITUTE SHEET

20909b9 Since modifications will be apparent to those of skill in the art, it is intended that this invention be limited only by the scope of the appended claims.

SEQUENCE LISTING
(1) GENERAL INFORMATION:
(i) Applicant: Brierley, Russell A.
Davis, Geneva R.
Holtz, Gregory C.
Gleeson, Martin A.
Howard, Bradley D.
(ii) TITLE OF INVENTION: Production of Insulin-Like Growth Factor-1 in Methylotrophic Yeast Cells (iii) NUMBER OF SEQUENCES: 12 (iv) CORRESPONDENCE ADDRESS:
(A) ADDRESSEE: Fitch, Even, Tabin & Flannery (B) STREET: 135 South LaSalle Street, Suite 900 (C) CITY: Chicago (D) STATE: Illinois 2O (E) COUNTRY: USA
(F) ZIP: 60603 (v) COMPUTER READABLE FORM:
(A) MEDIUM TYPE: Floppy disk (B) COMPUTER: IBM PC* compatible (C) OPERATING SYSTEM: PC-DOS*/MS-DOS
(D) SOFTWARE: PatentIn* Release #1.0, Version #1.25 3O (vi) CURRENT APPLICATION DATA:
(A) APPLICATION NUMBER:
(B) FILING DATE:
(C) CLASSIFICATION:
(vii) PRIOR APPLICATION DATA:
(A) APPLICATION NUMBER: US 07/578,728 (B) FILING DATE: 04-SEP-1990 82a 209 09 69 (viii) ATTORNEY/AGENT INFORMATION:
(A) NAME: Seidman, Stephanie L.
(B) REGISTRATION NUMBER: 33,779 (C) REFERENCE/DOCKET NUMBER: 51874 (ix) TELECOMMUNICATION INFORMATION:
(A) TELEPHONE: (619)552-1311 (B) TELEFAX: (619)552-0095 (2) INFORMATION FOR SEQ ID N0:1:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 240 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: double (D) TOPOLOGY: unknown (E) (ii) MOLECULE TYPE: cDNA
(iii) FEATURE:
2O (A) NAME/KEY: CDS
(B) LOCATION: 14..232 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:1:

Met Gly Pro Glu Thr Leu Gys Gly Ala Glu Leu Val Asp Ala Leu Gln Phe Val Cys Gly Asp Arg Gly Phe Tyr Phe Asn Lys '~'!7 92/04363 ~Q~Q969 Pro Thr Gly Tyr Gly Ser Ser Ser Arg Arg Ala Pro Gln Thr Gty Ile llal Asp Glu Cys Cys Phe Arg Ser Cys Asp Leu Arg Arg Leu Glu Met Tyr Cys Ala Pro Leu Lys pro Ala Lys Ser Ala A
(2) INFORMATION FOR SEQ ID N0:2:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 36 base pairs (B) TYPE: nucleic acid (C) STRANDEONESS: single (D) TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic DNA) (xi) SEQUENCE DESCRIPTION: SEO ID N0:2:

(2) INFORMATION FOR SEQ ID N0:3:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 16 base pairs (B) TYPE: rxicleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic DNA) (xi) SEQUENCE DESCRIPTION: SEQ ID N0:3:
ATMMGAGG ACCGGA

(2) lNFORIIATION FOR SE0 ID N0:4:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 40 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic DNA) (xi) SEQUENCE DESCRIPTION: SEO ID N0:4:

C2) INFORMATION FOR SEO ID N0:5:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 17 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic DNA) PCT/US91/064~' ~a9o~s~

(xi) SEDUENCE DESCRIPTION: SEG ID N0:5:

TMGAATTCA MTGAGT
(2) INFORMATION FOR SEQ ID N0:6:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 35 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: unknown (ii) lIOLECULE TYPE: DNA (genomic DNA) (xi) SECUENCE DESCRIPTION: SEO ID N0:6:
TCTTTGGATA MGAGAGGCT GGACCGCAGA CGCTC
(2) 1NFORhIATION FOR SEG ID N0:7:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 18 base Pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic DNA) (xi) SEQUENCE DESCRIPTION: SEC ID N0:7:

AAMGAGAGG CTGGACCG
(2) INFORMATION FOR SEG 1D N0:8:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 33 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: double (D) TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic DNA) (xi) SEQUENCE DESCRIPTION: SEO ID N0:8:

(2) INFORMATION FOR SEG ID N0:9:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 25 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: double (D) TOPOLOGY: unknown (ii) IAOLECULE TYPE: DNA (genomic DNA) (xi) SEQUENCE DESCRIP1ION: SEO ID N0:9:
GTAAAMTTG MGGAMTCT CATCG
(2) INFORMTION FOR SEG ID N0:10:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 36 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single '~3Ln 92/04363 PCT/US91/06452 (D) TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic DNA) (xi) SECUENCE DESCRIPTION: SEO ID N0:10:

(2) INFORMATION FOR SEO ID N0:11:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 14 base pairs (8) TYPE: nucleic acid (C) STRANDEDNESS: double (D) TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic DNA) (xi) SEQUENCE DESCRIPTION: SEA ID N0:11:

(2) INFORMATION FOR SEO ID N0:12:
(i) SEQUENCE CNARACTERISTICS:
(A) LENGTH: 14 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: double (D) TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic DNA) (xi) SEQUENCE DESCRIPTION: SEO ID N0:12:

Claims (68)

CLAIMS:

1. A DNA construct, comprising:
(a) DNA encoding an insulin-like growth factor-1 (IGF-1);
(b) a promoter region from a methanol-responsive gene of a methylotrophic yeast operably linked to said DNA that encodes IGF-1, said methylotrophic yeast is a member of the genus Pichia or Hansenula;
(c) DNA encoding the Saccharomyces cerevisiae .alpha.-mating factor (.alpha.MF) pre-pro signal sequence, wherein the DNA
that encodes said signal sequence is operably linked via DNA
encoding one or more processing sites, lys-arg that are recognised and cleaved by methylotrophic yeast cell proteases to the DNA that encodes IGF-1; and (d) a transcription terminator functional in a methylotrophic yeast host cell operably linked to the DNA that encodes IGF-1; wherein the methylotrophic yeast is a species that utilises methanol as a sole carbon source and is a member of the genera selected from the group consisting of Candida, Hansenula, Pichia, and Torulopsis.

2. A methylotrophic yeast host cell, comprising the DNA
construct of claim 1, wherein said DNA construct is integrated into the genome of said host cell, and said host cell is a member of the genus Pichia or Hansenula.

3. A process for producing IGF-1, comprising culturing the yeast host cell of claim 2 under conditions whereby authentic IGF-1 is produced and secreted into extracellular medium at concentrations of 9 mg/L or greater.

4. A DNA construct, comprising one or more copies of an expression cassette that contains:
(a) DNA encoding insulin-like growth factor-1 (IGF-1) ;
(b) a promoter region from a methanol responsive gene of a methylotrophic yeast operably linked to said DNA that encodes IGF-1, said methylotrophic yeast is a member of the genus Pichia or Hansenula;
(c) DNA encoding the S. cerevisiae .alpha.-mating factor (.alpha.MF) pre-pro sequence;
(d) a transcription terminator functional in a methylotrophic yeast host cell operably linked to said DNA that encodes IGF-1, wherein the DNA that encodes said pre-pro sequence is operably linked via DNA encoding the processing site lys-arg-(glu-ala)x to the DNA that encodes IGF-1;
x is 0, 1, 2 or 3;
the promoter region directs methanol-induced transcription in a methylotrophic yeast host of the DNA
encoding the pre-pro sequence and the IGF-1 peptide; and the methylotrophic yeast is a species that utilizes methanol as a sole carbon source and is a member of a genus selected from the group consisting of Candida, Hansenula, Pichia, and Torulopsis.

5. The DNA construct of claim 4, wherein x is 1, 2 or 3.

6. The DNA construct of claim 4, further comprising at least one selectable marker gene and a bacterial origin of replication.

7. A plasmid, comprising the DNA construct of claim 6.

8. The plasmid of claim 7, that is pIGF816.

9. The DNA construct of claim 4, wherein the DNA
encoding an IGF-1 peptide encodes an IGF-1 peptide that has the amino acid sequence set forth in Figure 1.

10. The DNA construct of claim 4, wherein the promoter is derived from a strain of Pichia pastoris.

11. The DNA construct of claim 10, wherein said promoter region from a methanol-responsive gene of a methylotrophic yeast and the transcription terminator are both derived from the P. pastoris AOX1 gene.

12. The DNA construct of claim 11, wherein the 3'- and 5'-ends of the construct have sufficient homology with a target gene of a yeast host to effect site directed integration of said construct into said target gene.

13. The DNA construct of claim 12, comprising multiple copies of said expression cassette.

14. The DNA construct of claim 4, wherein the 3'- and 5'-ends of the construct have sufficient homology with a target gene of a yeast host to effect site directed integration of said construct into said target gene.

15. The DNA construct of claim 4, comprising multiple copies of said expression cassette.

16. The DNA construct of claim 15, wherein said multiple copies of said expression cassette are oriented in head-to-tail orientation.

17. The DNA construct of claim 4, wherein said construct is included in the DNA in plasmids selected from the group consistng of pIGF201, pIGF202, pIGF204, and pIGF206.

18. A methylotrophic yeast cell, comprising the DNA
construct of claim 4, wherein the methylotrophic yeast cell is a species that utilizes methanol as a sole carbon source and is a member of the genus Hansenula or Pichia.

19. A methylotrophic yeast cell, comprising the DNA
construct of claim 9, wherein the methylotrophic yeast cell is a species that utilizes methanol as a sole carbon source and is a member of the genus Hansenula or Pichia.

20. The methylotrophic yeast cell of claim 18, wherein said yeast is a strain of Pichia pastoris.

21. A methylotrophic yeast cell, comprising the DNA
construct of claim 15, wherein the methylotrophic yeast cell is a species that utilizes methanol as a sole carbon source and is a member of the genus Hansenula or Pichia.

22. A Pichia pastoris cell, comprising the DNA construct of claim 12.

23. The P. pastoris cell of claim 22, wherein said cell is selected from the group consisting of strains G+IGF201S1, G+IGF201S2, G+IGF201S6 and G+IGF201S10.

24. A P. pastoris cell, comprising the DNA construct of claim 13.

25. The P. pastoris cell of claim 24, wherein said cell is selected from strains G+IGF202S3, G+IGF202S5, G+IGF204S2, G+IGF204S8, G+IGF206S2, G+IGF206S5, G+IGF206S8, G+IGF206S9, G+IMB202S2, G+IMB204S14, G+IMB206S1, G+IMB206S3.

26. The P. pastoris cell of claim 24, wherein said cell is selected from the group consisting of strains G-IMB206S1, G-IMB206S2 and G-IMB206S3.

27. A culture of viable methylotrophic yeast cells, comprising yeast cells of claim 18.

28. A culture of viable P. pastoris cells, comprising yeast cells of claim 22.

29. A culture of viable P. pastoris cells, comprising a single strain of yeast cells of claim 23.

30. A culture of viable P. pastoris cells, comprising yeast cells of claim 24.

31. A culture of viable P. pastoris cells, comprising a single strain of yeast cells of claim 25.

32. A process for producing insulin-like growth factor-1 (IGF-1) peptides, comprising culturing the cells of claim 18 under conditions whereby said cells express and secrete mature IGF-1 peptide into the culture medium.

33. The process of claim 32, wherein said methylotrophic yeast is a strain of Pichia pastoris.

34. The process of claim 32, wherein said cells are cultured in a medium containing methanol as a carbon source.

35. The process of claim 32, wherein said cells have the Mut+ phenotype.

36. The process of claim 32, wherein said cells have the Mut- phenotype.

37. The process of claim 32, wherein the cells are proteolytically intact and the initial pH of the culture medium is about 5, which decreases, and is maintained between 2 and about 4 prior to induction of the methanol responsive promoter.

38. The process of claim 37, wherein the pH is maintained at less than about 3.

39. The process of claim 32, wherein said cells are protease deficient Pichia cells and the initial pH of the culture medium is about 5 and is maintained between about 2 and prior to induction of the methanol responsive promoter.

40. A process for producing insulin-like growth factor-1 (IGF-1) peptides, comprising culturing the cells of claim 21 under conditions whereby said cells express and secrete mature IGF-1 peptide into the culture medium.

41. The process of claim 40, wherein said methylotrophic yeast is a strain of Pichia pastoris.

42. The process of claim 40, wherein said cells are grown in a medium containing methanol as a carbon source.

43. The process of claim 40, wherein said cells have the Mut+ phenotype.

44. The process of claim 40, wherein said cells have the Mut- phenotype.

45. The process of claim 40, wherein the initial pH of the culture medium is about 5 and decreases and is maintained at a pH of between about 2 and 4 prior to and during induction of the methanol responsive promoter.

46. The process of claim 45, wherein the pH is maintained at less than about 3.

47. The process of claim 40, wherein said cells are protease deficient Pichia cells that are deficient in proteinase B activity or are deficient in proteinase A and carboxypeptidase Y activities; and the initial pH of the culture medium is about 5 and is maintained at a pH of between about 2 and 5 prior to induction of the methanol responsive promoter.

48. The methylotrophic yeast cell of claim 18, wherein the methylotrophic yeast cell is a species of the genus Pichia, said species is deficient in protease B activity or is deficient in proteinase A and carboxypeptidase Y activities.

49. The strain of claim 20, wherein said strain is a protease deficient Pichia pastoris strain that is deficient in proteinase B activity or deficient in proteinase A and carboxypeptidase Y activities.

50. The strain of claim 49, wherein said deficiency results in an absence or reduction in the concentration of one or more proteinases selected from the group consisting of proteinase A, carboxypeptidase Y, and proteinase B.

51. The strain of claim 50, which is M+IGF816S1 or M+IGF816S4.

52. The strain of claim 50 that is a PEP4-IGF-1-expressing strain.

53. The strain of claim 50 that is a PEP4-,PRB-1- IGF-1-expressing strain.

54. The strain of claim 53 that is C+IGF816S1.

55. The strain of claim 49 that is M+IMB206S1.

56. A DNA construct, comprising one or more copies of an expression cassette that contains:
(a) a sequence of nucleotides encoding insulin-like growth factor-1 (IGF-1);

(b) a promoter region from a methanol-responsive gene of a methylotrophic yeast operably linked to the sequence of nucleotides that encodes IGF-1, said methylotrophic yeast is a member of the genus Pichia or Hansenula (c) a sequence of nucleotides encoding the Saccharomyces cerevisiae .alpha.-mating factor (.alpha.MF) pre-pro sequence;
(d) a transcription terminator functional in a methylotrophic yeast host cell operably linked to the sequence of nucleotides that encodes IGF-1, wherein:
the sequence of nucleotides that encodes the pre-pro sequence is operably linked via a sequence of nucleotides encoding one or more processing sites selected from the group consisting of lys-arg and lys-arg-(glu-ala)x to the sequence of nucleotides that encodes IGF-1;
x is 1, 2 or 3;
the promoter region directs methanol-induced transcription in a methylotrophic yeast host of the sequence of nucleotides encoding the pre-pro sequence and the IGF-1 peptide; and the methylotrophic yeast is a species that utilizes methanol as a sole carbon source and is a member of the genus Candida, Hansenula, Torulopsis or Pichia.

57. The DNA construct of claim 4, wherein x is 3.

58. The DNA construct of claim 4, wherein x is 1.

59. The DNA construct of claim 4, wherein x is 2.

60. The DNA construct of claim 4, wherein x is 0.

61. The DNA construct of claim 4, wherein the promoter region is selected from the group of Pichia pastoris promoter regions of methanol responsive genes consisting of the promoter for the primary alcohol oxidase gene AOX1, the promoter region of the secondary alcohol oxidase gene AOX2, the promoter region for the dihydroxyacetone synthase gene DAS, the promoter for the P40 gene and the promoter for the catalase gene promoters that direct methanol-induced transcription in Pichia pastoris of the sequence of nucleotides encoding the polypeptide.

62. A P. pastoris cell selected from the group consisting of strains G+IGF816S1, G+IGF816S2, G+IGF816S9 and G+IGF816S11.

63. A culture of viable P. pastoris cells, comprising yeast cells of claim 62.

64. A culture of viable Pichia pastoris cells, comprising the cells of claim 20.

65. The process of claim 33, wherein said cells are cultured in a medium containing methanol as a carbon source.

66. The process of claim 33, wherein said cells have the Mut+ phenotype.

67. The process of claim 33, wherein said cells have the Mut- phenotype.

68. The process of claim 33, wherein the cells are proteolytically intact and the initial pH of the culture medium is about 5, which decreases, and is maintained between 2 and about 4 prior to induction of the methanol responsive promoter.