CA1299120C - Stable dna constructs for expression of alpha-1-antitrypsin - Google Patents

Abstract

ABSTRACT
Methods are provided for producing alpha-1-antitrypsin in host cells and for selecting transformed cells comprising the step of transforming the host cell with a DNA molecule comprising a gene which complements a deficiency in the host cell. The host cell is a strain having a deficiency in a function necessary for normal cell growth. The gene in the DNA molecule, such as a plasmid, which complements the deficiency serves as a selectable marker whereby the growth conditions for selection may comprise a conventional complex medium.

Description

!L2~

STABL~ DNA CONSTRUCTS FOR EXPRESSION

Background of the Invention The use of microorganisms for the production of useful polypeptide products through recombinant DNA technology is becoming established as an industry. Foreign genetic material may be introduced into a culture of microorgan-isms, and, given the proper intracellular and extracellular conditions, the desired protein product(s) may be synthe-10 sized from the foreign gene(s). Such genetic material iscommonly introduced into microorganisms in the form of plasmids, which are autonomously replicating extrachromo-somal elements. In order to ensure the maintenance of plasmids within a culture of transformed cells, it has been 15 necessary to grow those cells under special conditions. In the absence of such conditions, the plasmids, which may be inherently unstable; will not be maintained, and the cell population will revert to the untransformed state.

Increased plasmid stability and copy number are important 20 to the biotechnology industry as a means of maintaining the production of plasmid-encoded proteins at a consistently high level. Previously reported attempts to increase plasmid stability do not appear to be optimal for commer~
cial application. The introduction of yeast centromeres 25 into _ -bearing plasmids, while enhancing stability, has been shown to markedly decrease plasmid copy number (Clarke and Carbon, Nature 287: 504-509, 1980 and Stinchcomb, et al., J. Molec. Biol. 158; 157~179, 1982). Linear centro-meric yeast plasmids similarly show an inverse relationship between stability and copy number ~Murray and Szostak, Nature 305 189-193 1983~
_ , .

Plasmids typically contain gene sequences, known as select-able markers, which encode antibiotic resistance or comple-ment nutritional requirements of the host cell. To selectfor the presence of such plasmids, transformed cells must thus be grown in special media which contain a selective drug or which are depleted for specific nutrients. These media requirements may be both expensive and prohibitive of 10 optimal cell growth rates during the large-scale fermenta-tion process. ~any such plasmids have been reported in the literature. Those comprising antibiotic drug resistance genes include pBR322 (Bolivar, et al., Gene 2: 95-113, 1977) and its derivatives, such as the pUC vectors IVieira 15 and Messing, Gene lg: 259-268, 1982) which carry a gene for ampicillin resistance; and pBR325 (Prentki, et al., Gene 14: 289, 1981) which carries resistance genes for ampicil-lin, tetracycline, and chloramphenicol. Plasmids which complement host nutrient requirements include the yeast 20 vectors YEpl3 (Broach, et al., Gene 8: 121-133, 1979), which carries the LEU2 gene; and YRp7' (Stinchcomb, et al., Nature 2 : 39, 1979), which carries the TRP1 gene.

Alpha-l-antitrypsin is a protease inhibitor, the principal function of which is to inhibit elastase, a broad spectrum 25 protease. Lung tissue in mammals is particularly vulner-able to attack by elastase, therefore alpha-l-antitrypsin deficiency or inactivation may lead to loss of lung tissue elasticity and subsequently to emphysema. Loss or reduc-tion of alpha-l-antitrypsin activity may be a result of 30 oxidation of alpha-l-antitrypsin due to environmental pollutants, including tobacco smoke. Deficiency of alpha-l-antitrypsin may result from one of several genetic disorders. See Gadek, James E., and R. D. Crystal, "Alpha-l-Antitrypsin Deficiency", The Metabolic Basis of Inherited 35 Disease, Stanbury, J. B., et al., Ed. McGraw-Hill, New York ~g~120 (1982) pp. 1~50-1467; and Carroll, et al., Nature 2988, 329-334 (1982).

It is therefore an object of the present invention to provide DNA constructs containing a DNA sequence encoding 5 alpha-1-antitrypsin and, as selectable markers, gene sequences whose products are essential for the viability or normal growth of the host cell on complex media.

It is another object of the present invention to provide transformant strains of microorganisms containing plasmids 10 which are selectable by growth on complex media and which are capable of expressing alpha-l-antitrypsin.

It is a further object of the present invention to provide strains of microorganisms that are deficient in essential functions which may act as hosts for DNA constructs 15 carrying gene sequences which complement these defective essential functions and are capable of expressing alpha-antitrypsin.

It is yet another object of the present invention to provide methods for producing alpha-l-antitrypsin in trans-20 formed microorganisms, wherein the alpha-l-antitrypsin is a product of a gene carried on a DNA construct which con-tains, as a selectable marker, a gene sequence which complements a deficiency in an essential gene in the host microorganism.

25 Other objects of the invention will become apparent to those skilled in the art.

Summary of the Invention According to the present invention, there are provided DNA
constructs and appropriate host cells such that the con-30 structs are capable of expressing alpha-l-antitrypsin and are maintained at high copy number without the need for special selective media. Growth in such conditions may ~zg9~20 result in faster growth, greater cell density, and reduced production costs.
According to another aspect of the present inventlon there is provided a method for producing alpha-1-antitrypsin in a microorganism host cell having a de~iciency in a function necessary for normal cell growth on complex media comprising the steps of:
(a) transforming said microorganism host cell with a DNA molecule comprlsing a gene which complements said deficiency and a sequence coding for said alpha-1-antitrypsin;
(b) culturing the transformants from s~ep (a) in a growth medium which need not contain antibiotics or heavy metals, and need not be depleted of specific nutrlents, under conditions whereby said gene functions as a selectable marker for transformant cells The present invention provides a method for producing alpha-1-antitrypsin in a host cell having a deficiency in a function necessary for normal cell growth in complex media, with the step of transformlng the host cell with a DNA molecule com-prising a gene which complements the deficiency and a sequencecoding for alpha-1-antitrypsin.
In some preferred embodiments: said gene is a gene required for host cell division, cell wall biosynthesis, membrane biosynthesis, organelle biosynthesis, protein synthesis, carbon source utilization, RNA transcription, or DNA replication; said gene is a gene required for host cell division, cell wall bio-synthesis, membrane biosynthesis, organelle biosynthesis, protein synthesis, carbon source utilization, RNA transcription, or DNA

~;Z9~2~
- 4a - 69140--31 replication and sai.d gene i5 selected from the group consisting of genes of the yeast cell division cycle and genes of the yeast glycolytic pathway; and said gene is selected from the group con-sisting of genes of the yeast cell division cycle and genes of the yeast glycolytic pathway.
As used herein the term "DNA construct" means any DNA
molecule which has been modified by man in a manner such that the nucleotide sequences in the molecule are not identical to a sequence which is produced naturally. The term "DNA construct"
also includes clones of DNA molecules which have been so modified.
The term "expression vector" is defined as a DNA construct which includes an autonomous site of replication, a site of transcript-ion initiation and at least one structural gene coding for a protein which is to be expressed in the host organism. The expression vector will usually also contain appropriate control regions such as a promoter and terminator which control the expression of the protein in the host organism. Expression vectors according to the present invention will also contain a selection marker comprising an essential gene as described herein.
The term "plasmid" wlll have its commonly accepted meaning, i.e., autonomously replicating, usually closed-looped, DNA.
In the accompanying drawings:
Figure 1 illustrates the construction of plasmid pB4.
Figure 2 illustrates the construction of plasmid pB5.
Figure 3 illustrates the construction of plasmid pB15L.

Figure 4 shows a Southern blot of DNA from S. cerevisiae strain A2.7.c co-transformed with plasmids pB5 and pB15L.
The blot was probed with a 2.5 kb BamHI-HindIII fragment from the 5' flanking region of CDC4 in order to test for disruption of the genomic CDC4 locus. Lane a contains DNA
from cells transformed with pB5 alone; Lane b, untrans-formed cells; Lanes c-h, co-transformants. Arrows indicate the genomic fragments hybridizing to the probe.

Figure 5 shows the sequences of the S. pombe POTl and S.
10 cerevisiae TPIl genes together with the respective inferred protein sequences. The entire S. ~ombe TPI protein sequence is given. The sequence of the S. cerevisiae protein is given only where it differs from the S. pombe sequence. The methionine at position 1 in the S. cerevis-15 lae protein sequence is not present in the mature protein.

Figure 6 illustrates the construction of the plas~id pCPOT.

Figure 7 illustrates the construction of the plasmidpFATPOT.

Figure 8 illustrates the construction of the plasmid 20 pTPI-LEU2.

Detailed Description The present invention is based in part upon the discovery that essential genes may be used as selectable markers on DN~ constructs such as plasmids which are capable of 25 expressing alpha-l-antitrypsin. An "essential gene" is defined as any gene that codes for a function necessary for cell viability or normal growth on complex media. Complex media are those media in which the nutrients are derived from products whose composition is not well defined, such 30 as crude cell extracts, meat extracts, fruit juice, serum, protein hydrolysates, etc. Hence, to select for a desired transformant according to the present invention, the selection growth medium will be merely a conventional ~2~

complex growth medium, not a special medium containing a relatively expensive antibiotic, metal antagonist, or other agent lethal to the untransformed host cell, or lacking one or more specific nutrients required by the untransformed host. Essential genes include, but are not limited t~, genes required for cell division, membrane biosynthesis, cell wall biosynthesis, organelle biosynthesis, protein synthesis, carbon source utilization, RNA transcription, and DNA replication.

10 In order to use an essential gene as a selectable marker on a DNA construct, such as a plasmid, it is necessary to provide an appropriate mutant host cell strain. Using the one-step gene disruption method of Rothstein (Meth. ln Enzymology 101: 202-210, 19~3) or the co-transformation 15 procedure described herein, suitable host strains may be constructed which carry deletions in an appropriate essen-tial gene in the genome. Such deletion mutants grow when the mutation is complemented by a function coded by plas-mid-borne genetic material. It is preferred that the 20 deletions in the essential gene or genes of the genome of the host comprise substantial segments of the coding region and/or flanking regions. If the mutation or mutations in the essential gene are accomplished in a manner to achieve only point mutations, then there is a likelihood that the 25 mutant host cell will revert to wild-type by mutation or a recombination repair mechanism, thereby reducing or eliminating the selectivity achievable by use of the plasmid-borne gene.

Essential genes often exist in multiple copies (such as 30 histone or ribosomal RNA genes) and/or in multiple, related forms called gene families (such as different hexokinase genes, or different DNA polymerase genes). In such cases, these redundant functions may be sequentially mutated to make a host cell which is multiply deficient for a given 35 essential function. ~owever, by using a high copy number plasmid to increase the activity of the gene, a single lZ9912~

essential gene on a plasmid may complement multiple host cell deficiencies. A high copy number plasmid is desirable because an increase in copy number of a cloned foreign gene may result in an increase in the production of the protein product encoded by said gene.

The selection for transformants containing high copy numbers of plasmids with essential genes may be accom-plished by reducing the expression levels of each plas-mid-borne essential gene and/or by reducing the activities 10 of the gene products encoded by the plasmid-borne select-able marker. One approach is to mutate the essential genes such that the transcription and/or translation rates of the genes are reduced or the gene products are altered to have lower specific activities. Another method for decreasing 15 the expression levels of essential genes used as selectable markers is to use a gene from another organism to comple-ment defects in the host cell. Such foreign genes may be naturally defective for expression in a host cell because the signals for transcription and/or translation may be 20 suboptimal in a different species or the gene product may have decreased activity or stability because it is in a foreign cellular milieu.

A broad range of functions necessary for cell viability or normal growth on complex media exists. A defect or dele-25 tion in an essential gene may result in lethality, adecrease in the rate of cell division, cessation of cell division, termination of DNA, RNA, or protein synthesis, termination of membrane synthesis, termination of cell wall synthesis, termination of organelle synthesis, defects in 3~ sugar metabolism, etc. Examples of essential genes include the CDC ~cell division cycle) genes of the yeast Sac-charomyces cerevisiae (for review see Pringle and Hartwell, "The Saccharomyces cerevisiae Cell Cycle", in Strathern, et al., eds., The Molecular Biology of the Yeast SaccharomYces 35 Life C~ycle and Inheritance, 97-142, Cold Spring ~larbor, 1981), the genes coding for functions of the S. cerevisiae ~Z99~120 and E. coli glycolytic pathways, and the SEC (Novick and Schekman, Proc. Nat. Acad. Sci. USA 76: 1856-1862, 1975 and Novick, et al., Cell 21: 205-215, 1980) and INO (Culbertson and Henry, Genetics 80: 23-40, 1975) genes of S. cerevis-lae.

One preferred class of essential yene-deficient host cells contains defects in CDC genes known as cdc mutations, which lead to stage-specific arrests of the cell division cycle.
~ost cdc mutations produce complete blockage of events 10 essential to the cell cycle by affecting either the synthe-sis or function of the particular CDC gene products. Such mutations may be identified by their effects on events which can be monitored biochemically or morphologically.
Most known cdc mutations are conditionally lethal (l.e., 15 temperature sensitive) mutations, which result in the cessation of normal development of mutant cells grown under restrictive conditions. However, the primary defect resulting from a cdc mutation need not be a defect in a stage-specific function per se. For example, continu-20 ously-synthesized gene products may have stage specific functions; a defect in the yeast glycolytic gene PYK1 (for the enzyme pyruvate kinase) is allelic to the cell division cycle mutation cdcl9 (Kawasaki, Ph.D. Thesis, University of Washington, 1979). This mutation results in cell cycle 25 arrest at the Gl phase of cells incubated in the typical yeast complex medium YEPD (1% yeast extract, 2% bactopep-tone, and 2~ dextrose). Thus, whether the cdc mutation results in a defect in a stage-specific function, or whether it causes an inhibition or disabling mutation of a 30 gene product having a stage-specific function, the effect of the defect may be monitored.

Pringle and Hartwell (ibid.) describe the function of some 51 CDC genes. For use in carrying out the present invention, such genes may be isolated irom gene libraries by complementation in a strain carrying the desir~d mutation. Gene libraries may be constructed by commonly lZ99~
g known procedures (for example, Nasmyth and Reed, Proc.
Natl. Acad. Sci. ~SA 77: 2119-2123, 1980; and Nasmyth and Tatchell, Cell _: 753-764, 198G). Strains carrying the desired cdc mutation may be prepared as described herein, or may be obtained from depositories accessible to the public, such as, the American Type Culture Collection and the Berkeley Yeast Stock Center.

A second preferred class of essential genes are those encoding products involved in the glycolytic pathway, 10 including genes coding for metabolic enzymes and for regulatory functions. Examples of glycolytic pathway genes in S. cerevisiae which have been identified are the glycolysis regulation gene GCRl and the genes coding for the enzymes triose phosphate isomerase, hexokinase 1, 15 hexokinase 2, phosphoglucose isomerase, phosphoglycerate klnase, phosphofructokinase, enolase, fructose 1, 6-bis-phosphate dehydrogenase, and glyceraldehyde 3-phosphate dehydrogenase. As noted above, the pyruvate kinase gene has been identified and described by Kawasaki. A plasmid 20 containing a yeast phosphoglycerate kinase gene and accompanying regulatory signals has been described by Hitzeman, et al. (J. Biol. Chem. 225: 12073-12080, 1980).
Isolation and sequencing of the yeast triose phosphate isomerase gene TPIl has been described by Alber and 25 Kawasaki (J. Mol. ~ . Genet. 1: 419-434, 1982) and by Kawasaki and Fraenkel (Biochem. Biophys. Res. Comm. 108:
1107-1112, 1982).

A particularly preferred glycolytic gene is TPIl, which codes for the yeast triose phosphate isomerase, an enzyme 30 which catalyzes the interconversion of glyceraldehyde-3-phosphate and dihydroxyacetone-3-phosphate and is therefore essential for glycolysis and gluconeogenesis. In S.
cerevisiae the single genetic locus, TPIl, codes for this function. Cells carrying mutations in TPIl do not grow on 35 glucose and grow poorly on other carbon sources.

~Z~91~0 The S. cerevisiae TPI1 gene was isolated by complementation of the tpil mutation (Alber and Kawasaki, i_ ., and Kawasaki and Fraenkel, ibid.). The triose phosphate isomerase gene from the fission yeast Schizosaccharomyces ~ (POT1) has been isolated by complementation of the same S. cerevisiae mutation, and has been sequenced as shown in FIG. 5. Sequencing of the S. ~ombe gene, designated POT1, has demonstrated that the S. ~ombe TPI
protein is homologous to the TPI protein of S. cerevisiae.

10 While in the usual case the essential gene which is utilized in the DNA construct (plasmid) will be a wild-type gene from the host species, in some cases it will be preferable to use an essential gene which is foreign to the host cell because the foreign gene may be naturally 15 defective, and thereby selectable to high plasmid copy number. As an example of such a foreign essential gene being used, one of the examples herein shows that the S.
~ POT1 gene may be effectively used as a selectable marker in an S. cerevisiae host.

20 The DNA constructs according to the present invention containing essential genes as selectable markers will be transformed into mutant host cells which are defective in the function of the essential gene. Properly mutated host cells must either be prepared or, may be readily available 25 from a public depository. Mutation of the wild-type cell to obtain a proper mutant may be accomplished according to conventional procedures. For example, wild-type cells may be treated with conventional mutagenizing agents such as ethane methyl sulfonate and transformed with a plasmid 30 containing an essential gene to identify the colonies where complementation occurs. Alternatively, the genome ma~ be disrupted to create a specific mutation (Rothstein, ibid).

The stability of the plasmid containing the essential gene in the host cell may be dependent on the absence of 35 homologous essential gene sequences in the host cell. The .

~25'9120 genetic defects in the host ensure that the plasmid will be maintained since growth of the host cell will not occur or will be severely limited by the lack of the essential gene function. Additionally, the integrity of the plasmid itself may be dependent upon the absence of homology between the plasmid-borne essential gene and the corresponding locus in the host genome, because recom-bination between respective plasmid and genomic loci may cure the cell of both the mutation and the plasmid. Thus, 10 it is preferred that mutation in the host cell genome which inactivates the genomic essential gene be of a substantial nature, i.e., deletions be made from the DNA sequences of the coding section and/or flanking regions of the chromosomal gene. Once this is accomplished, curing of the 15 genomic mutation by recombination is less likely to occur.

The plasmids of the present invention are unexpectedly stable when maintained in the appropriate mutant host cells. A preferred host cell is yeast; however, other eukaryotic cells may be utilized, as well as prokaryotic 20 cells. In the case of yeast cells, the stability of the plasmids according to the present invention appears to exceed even that of yeast plasmids containing centromeres.
Circular centromere plasmids are among the most stable plasmids previously reported for yeast, but suffer from an 25 extremely low copy number (Clarke and Carbon, ibid. and Stinchcomb, et al., 1982, ibid.). Linear centromeric yeast plasmids are either unstable or present at low copy number, depending on plasmid length (Murray and Szostak, ibid.).
It is therefore an unexpected advantage that improved 30 stability of plasmids bearing an essential gene is achieved.

The POTl and CDC4 genes are two examples of the utility of essential genes as selectable markers on expression vectors. These two genes belong to a broad class of genes 35 that are required for cell proliferation on complex media.
The use of other essential genes may allow for plasmid ~Z~9~

selection in plant Dr animal tissue culture which involves complex growth conditions and at the extreme may allow for the maintenance of plasmids in cells receiving nutrition from blood, serum, or sap of living animals or plants.

Data obtained from experiments using plasmids described herein show that human alpha-1-antitrypsin (AT) productiGn is doubled by the use of the S. pombe POTl gene as the selectable marker, when compared to AT production obtained with similar plasmids bearing a traditional auxotrophic 10 selectable marker, LEU2. These results indicate that POTl containing plasmids are functionally greater in copy number than the non-POTl plasmids from which they are derived.

The techniques used to produce the DNA constructs, l.e., in particular the plasmids, according to the present inven-15 tion, involve conventional methods. The essential gene tobe utilized in the DNA construct may be isolated from a library by using a labeled DNA probe if the structure of the gene is known, or identified by ligating segments of the DNA library to conventional vectors, transforming the 20 vectors into a mutant cell deficient in the particular essential gene and searching for colonies which are comple-mented. Once an appropriate DNA fragment containing the essential gene is identified it will be ligated to a vector which contains a DNA sequence coding for the structural 25 protein which will be expressed. The essential gene may be utilized together with its own promoter and other controls necessary for expression within the host organism. Alter-natively, a heterologous promoter may be utilized to increase or decrease expression of the essential gene.
30 Methods of ligation of DNA fragments are amply described and are well within the skill of those of ordinary skill in the art to perform.

After preparation of the DNA construct it will be trans-formed into the host organism under transforming condi-35 tions. Techniques for transforming prokaryotes and ~Z99~20 eukaryotes (including tissue culture cells) are known in the literature.

As described above the host organism must be deficient in the essential function for selection of the essential gene on a plasmid. Mutant host strains are available from conventional depositoriec or may be made by conventional means from wild-types by mutagenesis and screening for the mutant carrying the proper mutation.

The transformed host may then be selected by growth on lO conventional complex medium. In the case of yeast, a conventional medium such as YEPD (1% yeast extract, 2 bactopeptone, and 2~ dextrose) may be used. The selectable markers comprising essential genes according to the present invention may be used as markers wherever appropriate in 15 any DNA construction and thus it will be recognized that constructs containing the essential gene selection markers according to the present invention have many uses. The following examples are offered by way of illustration of such use, not by way of limitation.

20 Unless otherwise indicated, standard molecular biology methods were used. Enzymes were obtained from Bethesda Research Laboratories, New England BioLabs, and Boehringer Mannheim Biochemicals, and were used as directed by the manufacturer or as described by Maniatis, et al. (Molecu-25 lar Cloning: A Laboratory Manual, Cold Spring HarborLaboratory, 1982). E. coli cultures were transformed by the calcium chloride method, also disclosed in Maniatis, et al. (ibid.). Yeast cultures were transformed by the method of Beggs (Nature 275: 104-108, 1978), with modifications as 30 described herein.

1;~99~ZO

~xample 1 The S. cerevisiae CDC4 qene as selectable marker A. Construction of a stable CDC4-containing plasmid A yeast genomic library was constructed by partial diges-tion of yeast DNA with Sau3A, size selection on sucrose gradients, and insertion of the selected fragments into the yeast vector YRp7 which had been digested with BamHI
(Nasmyth and Reed, Proc. Natl. Acad. Sci. USA. 77: 2119-10 2123, 1980). A recombinant plasmid containing the CDC4gene was isolated by transformation of yeast strains GEB5 (MATa cdc4-4 leu2 trpl lysl ural) and GEB7 (MATa cdc4-3 leu2 trpl lysl) with the library. These strains were derived from strains A364A cdc4-3 and A364A cdc4-4 (Hart-15 well, et al., Genetics 74: 267-286, 1973) by crossing with a strain known to transform at high frequency (K79 [MAT~
leu2 trpl] (Nasmyth, et aI., Nature 289: 244-250, 1981;
Tatchell, et al., Cell 27:25-35, 1981) followed by back-crossing to high transformlng strains (K79 and K80 [MATa 20 leu2 trpl lysl]) to obtain the cdc4-3 and cdc4-4 mutations in the desired genetic background (leu2 ~ ). Selection of transformants for tryptophan prototrophy and the ability to grow at the restrictive temperature (37) identified one such plasmid (designated pJY35) which was shown to inte-25 grate into the genome and map to the CDC4 locus. Spontan-eous plasmid integrants were identified on the basis of their selective growth advantage. This growth advantage is due to the presence, on the original plasmid, of a CDC4-linked gene which is deleterious to cell growth when 30 present at high copy number (l.e., when the plasmid is not integrated into the host genome). In the integrants, the TRPl plasmid marker was shown to be genetically linked to SUPll, which is linked to CDC4 on chromosome VI (Mortimer and Schild, "Cenetic Map of Saccharomyces cerevisiae" in 35 Strathern, et al., eds., The Molecular Biology of the Yeast Saccharomyces cerevisiae Life Cycle and Inheritance, 641-651, Cold Spring Harbor, 1981). The cdc4-3 ~LZ991~:0 complementing region was purified from pJY35 as a 6.4kb BamHI fragment and was joined, using T4 DNA ligase, to the vector YRp7 (Struhl, et al., Proc. Natl. Acad. Sci. USA 76:
1035-1039, 1979) which had been cleaved with BamHI. This construct is known as pJY51, and is illustrated in FIG. 1.

Referring to FIG. 1 the CDC4 coding region was purified away from flanking genomic DNA sequences in the following manner. Plasmid pJY51 was cleaved with HindIII and the 3.6kb fragment comprising the CDC4 region was subcloned in lO the bacterial plasmid pBR322. This construct was digested to completion with BamHI, partially digested with HincII, and the ca. 2.3 kb CDC4-containing fragment was purified.
The HincII fragment end was converted to a BamHI end by the addition of linker sequences (sequence:5'CCGGATCCGG3') 15 (obtained from Collaborative Research) and subsequent digestion with BamHI to remove excess linkers. The result-ing fragment, comprising approximately 1.9 kb of the CDC4 gene, was inserted into the BamHI site of YRp7 to produce plasmid pJY70. This plasmid was shown to complement the 20 cdc4-3 mutation as described above. Although the 1.9 kb fragment lacks small portions of both the 5'- and 3'-coding regions of the CDC4 gene, it surprisingly complements the temperature-sensitive defect. Presumably, transcription and translation of the CDC4 sequence is controlled by se-25 quences located in the pBR322 regions of the plasmid,allowing for production of a functional gene product.

Plasmid pJY70 was cleaved with EcoRI to remove the yeast TRPl and ARS1 sequences and was re-ligated, yielding a hybrid plasmid comprising pBR322 and CDC4 sequences. This 30 plasmid is known as pJY71, and is illustrated in Figure 1.

The l.9 kb yeast sequence was purified from pJY71 as a BamHI-HindIII fragment. This fragment was joined to pBR322 which had been linearized by digestion with BamHI and HindIII, to produce the plasmid pB4, and is illustrated in 35 Figure 1.

~29S~ZO

The CDC~ region was re-isolated from pB4 for insertion intG
a high copy number yeast vector. Such a vector will contain an origin of replication of the yeast 2~ plasmid, and one or more restrlction enzyme cleavage sites which will serve as cloning sites for foreign genes of interest.
Preferably such sites will be unique sltes on the plasmid.
A preferred vector is MW5, which comprises the yeast 2~
plasmid replication origin and unique EcoRI and BamHI
cloning sites. Referring to FIG. 2 plasmid MW5 was derived 10 from plasmid YRp7' (Stinchcomb, et al., Nature 282: 39-43, 1979) by partial digestion with EcoRI to cleave, on aver-age, one of the two EcoRI sites per molecule. The result-ing unpaired ends of the linear molecules were filled in using DNA polymerase I (Klenow fragment) and the resulting 15 blunt ends were re-joined using T4 DNA ligase. The result-ing plasmid which retained the EcoRI site adjacent to the ARS1 sequence was then selected. The ARS1 sequence was removed by digestion with PstI and EcoRI, and replaced with the PstI-EcoRI fragment of plasmid YEpl3 (Broach, et al., 20 Gene 8: 121-133, 1979) which comprises the replication origin of yeast 2~ DNA. The resulting plasmid, designated MW5, is illustrated in FIG. 2.

To construct the final CDC4-containing stable plasmid, MW5 was cleaved with EcoRI and BamHI. The CDC4 fragment was 25 purified from plasmid pB4 by digesting the plasmid with BamHI and EcoRI. The two fragments were joined, using T4 DNA ligase, and the chimeric molecules so produced were transformed into E. coli strain RRI (Nasmyth and Reed, ibid.) with selection for ampicillin-resistant, tetracy-30 cline-sensitive colonies. Plasmid pB5 (shown in FIG. 2), isolated from one such colony, comprises the yeast 2~
replication origin, pBR322 plasmid sequences, the select-able marker TRP1, 1.9 kb of the yeast CDC4 coding sequence, and a unique EcoRI cloning site.

~Z99120 B. Construction of a plasmid for disruption of host CDC4 gene The stability, in a transformed host, of the CDC4-contain-ing plasmid according to the present invention is dependent on the lack of a functional CDC4 gene in the host. It is further desirable that no homology exists between the host genome and the CDC4-containing stable plasmid in order to prevent recombination between plasmid and chromosomal D~A's. To obtain a yeast strain having a suitably deleted lO CDC4 locus, a yeast host containing the wild-type CDC4 gene mav be transformed with a linearized plasmid fragment having a "disrupted" CDC4 gene (Rothstein, ibid.). The linearized plasmid fragment is a preferred transforming agent because the free ends of the fragment may enhance 15 recombination within the CDC4 region. Such a plasmid fragment will have intact CDC4 flanking regions at its ends to facilitate recombination with the intact genomic CDC4 locus. The genetic material inserted between the CDC4 flanking regions of the plasmid fragment will code for a 20 phenotypic characteristic which can be selected in the transformed host (a selectable marker such as TRPl or LEU2). The disrupting plasmid will preferably also lack a yeast origin of replication in order to select for the integration of the disrupted CDC4-selectable marker 25 sequence into the host genome. Following transformation with the linearized plasmid, genetic recombination results in the substitution of the disrupted sequence for the genomic sequence of the host. Cells in which the _DC4 gene has now been deleted are then selectable according to the 30 marker used in the disruption.

A method for a one-step disruption of a host genome is described by Rothstein (ibid.). As described above, disruption is performed with the added improvement of co-transforming a host strain with an intact stable plasmid 35 and a linearized plasmid such that in a-ddition to achieving ~299120 disruption of the host genome, transformation of the host with the stable plasmid is also effected.

A preferred plasmid for disruption of the host CDC4 locus is pBl5L, shown in Figure 3. It comprises the yeast LEU2 gene inserted between the flanking regions of CDC4, and the vector pUC13 (Vieira and Messing, Gene 19: 259-268, 1982 and Messing, Meth. in Enzymology 101: 20-77, 1983). When linearized at the junctions of yeast and vector sequences and transformed into a suitable yeast host strain, the lO plasmid produces a deletion of CDC4 in the host genome resulting from the substitution of the LEU2 sequence for the C _ region. In a host strain auxotrophic for leucine, disrupted transformants may then be selected on the basis of leucine prototrophy.

15 To construct plasmid pB15L, a 6.4 kb fragment comprising the CDC4 gene and its 5'- and 3'-flanking regions was purifled from a BamHI digest of pJY51. This fragment was inserted into BamHI-digested pUC13 to produce the plasmid pB14. Most of the CDC4 coding region was removed by 20 digesting pB14 with ClaI and purifying the larger fragment which comprises the pUC13 and CDC4 flanking sequences. The fragment ends were modified by the addition of XhoI (BglII) "smart" linkers (Worthington Diagnostic), and the 2.8 kb BglII LEU2 fragment of YEpl3 (Broach, et al., Gene 8:121-25 133, 1979) was joined to the resultant cohesive termini~DNA so prepared was used to transform E. coli strain RRI.
Transformants were selected on the basis of leucine proto-trophy, since the yeast LEU2 sequence complements the leuB
defect in the E. coli host. Plasmid pB15L was purified 30 from one such transformed colony.

Plasmid pB15L comprises only about 50 base pairs of the 5' end of the CDC4 coding sequence in addition to the 5' and 3' flanking sequences. A comparision of the maps of plasmids pB5 and p~15L shows a lack of homology between 35 their respective CDC4 sequences as the junction points of lZ~20 the CDC4-LEU2 gene fusion of pB15L are located outside the region of the CDC4 fragment present in pB5. This lack of homology prevents recombination between pB5 and the disrup-ted CDC4 locus in the host cell.

C. Co-transformation of S. cerevisiae To simultane~usly delete the genomic CDC4 gene and intro-duce plasmid pB5, yeast cells were co-transformed wi~h BamHI-cleaved pB15L and intact plasmid pB5. The host strain to be used in the transformation should be auxotro-10 phic for tryptophan and leucine in order to select simul-taneously for plasmid pB5 and the genomic CDC4 disruption.
Strain A2.7.c (MAT~ cdc4-3 trpl leu2-2,112 lysl his3-11,15 canl obtained from a cross of strain A2 (MAT~ leu2-2,112 _s3-11,15 canl; see Szostak, Meth. ln Enzymology 101:
15 245-252, 1983) with strain GEB7 (see Example lA) was used.

In a typical co-transformation experiment, lOml of a culture of S. cerevisiae A2.7.c in log phase growth were transformed with approximately 6~g of BamHI-digested pB15L, l~g pB5, and lO~g calf thymus DNA as carrier. Transforma-20 tion conditions were as described by Beggs (ibid.). Cellswere plated on a medium lacking leucine and tryptophan.
They were grown overnight at 22 and shifted to 37.
Approximately 30 colonies were obtained. The control transformation with pB5 alone and selection for tryptophan 25 prototrophy produced approximately 1,000 transformants.

Six co-transformed colonies were analy~ed to verify the disruption of the CDC4 locus and to test the stability of the pB5 plasmid. Genomic DNA was isolated from co-trans-formants by the method of Abraham, et al. (Cold Spring 30 Harbor Symposium Quant. Biol. 47: 989-998, 1983) and was digested with EcoRI and BamHI, electrophoresed on an agarose gel, and transferred to nitrocellulose (Southern, J. Mol. Biol. 98: 503-517, 1975). The blot was probed with the 2.5 kb BamHI-HindIII fragment from the 5' flanking 9~ZO

region of CDC4 present in pB15L but absent from pB5.
Figure 4 shows that the probe hybridized to a 6.4 kb fragment of DNA from untransformed cells (lane b); there is no EcoRI site within this 6.4 kb BamHI fragment. As the LE~2 sequence contalns an EcoRI site, disruption of the CDC4 locus will result in a reduction in size of the hybridizing band (indicated by arrows in Figure 4). This is the case for the transformants represented in lanes c, d, f, g, and h. Lane e shows a somewhat different pattern 10 and retains the genomic-size hand~ indicating that deletion of the genomic CDC4 did not occur. (The smaller bands seen in lanes c through h are due to contamination of the gel-purified probe, as shown by the patterns of the con-trols in lanes a and b.) 15 The six co-transformants were tested for plasmid stability by growing on complex medium (YEPD). Cells were grown for 30 generations in liquid YEPD at 25, then plated on YEPD
at 25, and replica plated onto YEPD at 37, tryptophanless medium, and leucineless medium. Results summarized in 20 Table 1 indicate that all co-transformants except #3 were 100% stable for the plasmid markers on complex media.
(Isolate number 3 is the same co-transformant represented in lane e of Figure 4).

Further stabllity tests were performed Oll two co-trans-25 formants, numbers 1 and 2. Testing was performed on 663 and 681 colonies respectively. After growth for 30 gener-ations on YEPD at 30, all colonies were prototrophic for tryptophan and leucine.

Co-transformant #1 was tested for growth rate at 22 and 30 was found to grow at ~he same rate as an untransformed A2.7.c control.

Co-transformant #l has been designated BELLl. It has been deposited with ATCC under accession number 20698.

lZ9gl~0 Example 2 Schizosaccharomyces pombe POT1 gene A. S. pombe POT1 gene as a selectable marker The Saccharomyces _revisiae TPI1 gene codes for the triose phosphate isomerase protein and has been obtained by complementing the ~1 deLiciency (Kawasaki and Fraenkel, ibid.; Alber and Kawasaki, ibid.). Surprisingly, the homologous gene from S. pombe has been isolated by comple-menting the same _. cerevisiae tpil mutation. The S. pombe 10 TPI gene, designated as POTl (for pombe triose phosphate isomerase), has been cloned from a library described by Russell and Hall (J. Biol. Chem. 25~: 143-149, 1983) which contains genomic S. pombe DNA that has been partially digested with Sau3A and inserted into the vector YEpl3. A
15 preliminary DNA sequence (by the method of Maxam and Gilbert, Meth. in Enzymology 65: 497-559, 1980) has demon-strated that the POTl gene codes for the TPI protein and said protein is homologous with TPI proteins from other organisms (see Alber and Kawasaki, ibid.). This POTl DNA
20 sequence is given in Figure 5, together with the S. cere-visiae TPI1 DNA sequence and the respective protein sequences.

The S. pombe POTl gene is preferred in this example over the S. cerevisiae TPIl gene as a selectable marker in S.
25 cerevisiae. Foreign genes, such as POTl in S. cerevisiae, may not function well in an alien host cell and therefore may necessitate a higher copy number to complement a host cell defect. Also the selectable POT1 gene on a yeast plasmid allows for the use of the endogenous TPIl promoter 30 and TPIl terminator (control regions that show no homology with POTl) for expression of commercially important genes on the same vector. Because POTl and the flanking regions of TPIl show no homology, intramolecular recombination and subsequent plasmid instability are reduced. Finally, the 35 POTl gene is not likely to recombine with the S. cerevisiae chromosomal DNA because it shares little homology at the lZ99120 DNA level with the TPIl sequence and much of the TPI1 gene has been deleted in the host strains. Thus, POTl contain-ing plasmids may remain at high copy numbers which are desirable for the elevated expression of foreign genes of commercial interest in yeast.

A plasmid comprlsing the POTl gene was identified from the S. ~ library of Russell and Hall (ibid.) by complemen-tation of the ~1 mutation in S. cerevisiae strain N587-2D
(~awasaki and Fraenkel, ibid.).

10 A restriction map of this plasmid, pPOT, is depicted in Figure 6. Because pPOT contains the vector YEpl3, it is inherently unstable, since it lacks replication functions necessary for the maintenance of 2-micron plasmids in yeast. Therefore, the POTl gene may be moved into more 15 competent vectors, such as Cl/l and related vectors that contain the entire 2-micron plasmid sequences. Plasmid Cl/l was derived from pJDB248 (Beggs, Nature 275: 104-109, 1978) and pBR322 as described in Example 3 herein. It contains all of the yeast 2-micron plasmid DNA, a select-20 able LEU2 gene, and pBR322 sequences.

The POTl gene was isolated from pPOT as a BamHI-XbaI
restriction fragment of nearly 3,400 base pairs and was inserted into the corresponding polylinker sites of pUC13.
The resulting plasmid is pUCPOT, a partial restriction map 25 of which is shown in Figure 6.

The pUCPOT plasmid was cut with SalI and religated to delete about 1,800 base pairs of S. pombe and S. cerevisiae DNA. This resulting pUCPOT-Sal plasmid is illustrated in Figure 6.

30 The POTl gene was put into Cl/l in the following manner.
As both Cl/l and pUCPOT-Sal have a BglI site in the ampi-cillin resistance gene and a unique BamHI site at some other location, the POTl fragment of pycpoT-sal may be , .

~299~20 substituted for a portion of the pBR322 region of Cl/l.
Cl/l was cut with BglI and BamHI to liberate a large fragment of nearly 7,700 base pairs that contains part of the ampr gene, all 2-micron DNA, and the LEU2 gene.
Likewise, pUCPOT-Sal was cut with BglI and BamHI to liber-ate a fragment of nearly 3,400 base pairs that contains the other portion of the ampr gene and the POTl gene. These two fragments were ligated to form pCPOT, which contains a "restored" selectable ampr gene, the POTl gene, the LEU2 10 gene, all 2-micron DNA, and the bacterial origin of repli-cation region from pUC13 (the bacterial origin region from pUC13 allows for a higher copy number of plasmids in E.
coli than does the origin region of pBR322).

E. coli strain HB101 transformed with pCPOT has been 15 deposited with ATCC under accession number 39685.

The POTl gene may also be inserted into Cl/l-derived vectors by a similar construction. For example, the plasmid pFAT5 (FIG. 7) contains an expression unit for the production of human alpha-l-antitrypsin (AT) inserted into 20 Cl/l. This expression unit, prepared as described in Example 4 consists of the TPIl promoter, the AT cDNA
sequence, and the TPIl transcription terminator. A re-striction map of pFAT5 is given in Figure 7.

pFAT5 was cut with BglI and BamHI to liberate a fragment 25 (2,200 base pairs) that contains the AT gene and the TPIl terminator. Also liberated is a BglI-BamHI fragment which is identical to the Cl/l BglI-BamHI fragment described above, except that the fragment from pFAT5 contains an additional 900 base pairs that comprise the TPIl promoter.
30 This latter pFAT5 piece and the pUCPOT-Sal 3400 bp Bgll-BamHI fragment (described above) are ligated to form the plasmid pFPOT, which has the restriction map shown in Figure 7.

~Z~9~20 The vector pFPOT was cut at the unique BamHI site to allow for the insertion of the 2,200 base pair AT gene and TPIl terminator fragment from pFAT5. The cloning of the 2,200 base pair fragment in the proper orientation into pFPOT
allows for the expression of human AT in this yeast vector.
The properly ligated product is designated pFATPOT, whose restriction map is given in Figure 7.

B. Disruption of host TPI gene The Saccharomyces cerevisiae TPIl gene has been cloned and 10 sequenced (Kawasaki and Fraenkel, ibid. and Alber and Kawasaki, ibid.). The plasmid pTPIC10, comprising the structural gene for the TPI protein, has been described in Alber and Kawasaki (ibid.). A BglII site exists at DNA
position 295 in the coding region of TPIl, and another 15 BglII site is located approximately 1,200 base pairs away in the 5' flanking resion. These BglII sites are conven-ient cloning sites for deleting part of the TPIl gene and for inserting another gene, such as the yeast LEU2 gene.
Such a construct can be used to produce a disruption of the 20 genomic TPIl locus in a transformed host.

At approximately -1800 in the 5' flanking region of TPIl is a Pstl site. In pTPICl0, therefore, the TPIl gene is flanked by a PstI site on the 5' side and by a SalI site (in the tetr gene) on the 3' side. This PstI-SalI fragment 25 which contains TPIl was inserted into pUC13 at the PstI and SalI sites to produce pUCTPI. A restriction map of the PstI-SalI insert (into pUC13) is given in Figure 8.

The plasmid pUCTPI was then cut with BglII and the two DNA
fragments were separated by electrophoresis. The larger fragment was purified and phosphatased to prevent self-li-gation. Into the BglII sites of this DNA was ligated the yeast LEU2 gene, which was removed from the plasmid YEpl3 (Broach, et al., Gene 8: 121-133, 1979~ as a BglII frag-ment. The resulting plasmid was pUCTPI-LEU2, which carries ,.

~Z9~120 a partial deletion of TPIl and an insertion of LEU2. pUCTPI-LEU2 is depicted in Figure 8.
The plasmid pUCTPI-LEU2 was cut with PstI and BamHI to linearize the DNA. The yeast sequences were then isolated from the pUC13 sequences by electrophoresis and gel purification. The yeast DNA portion depicted in Figure 8 was used to transform S.
cerevisiae strain E2-7B (ATCC No. 20689), which is deficient for LEU2, in order to "disrupt" the TPIl chromosomal gene (Rothstein, ibid.). Leu+ transformants were selected on a synthetic (modi-fied Wickerham's) medium (Mortimer and Hawthorne, in Rose andHarrison, eds., The Yeasts vol. 1, 385-460, Academic Press, 1969) which contained 3% glycerol and 1% lactate (neutralized to pH 7), lM Sorbitol, and no leucine. The transformants were screened for a TPI deficiency by their inability to grow on YEP-Dextrose. One tpi- transformant was found among the first 99 transformants screened. This strain was designated as E2-7B~tpi#29 (hereinafter ~tpi#29). ~tpi#29 grew on YEP-3% Glycerol-l~ Lactate but not on YEP-Dextrose. Enzyme assays (Clifton, et al., Genetics 88: 1-11, 1980) were run on crude cellular extracts and confirmed that ~tpi#29 was lacking detectable levels of triose phosphate iso-merase activity.
~tpi#29 may be crossed to other yeast strains to form diploids that are heterozygous for the tpi- deletion. Such diploids may be sporulated so that other strains deficient for triose phosphate isomerase can be generated. For example, ~tpi#29 has been crossed to E8-lOA (MAT~ eu2) (a spore segregant of the cross E2-7BxGK100[ATCC 20669]) to form the diploid, Ell. This diploid has been sporulated to generate the haploid descendant, E11-3C, which has the following genotype: MAT~ pep4-3 ~ .
E11-3C has been crossed back to ~tpi#29 to form a diploid, E18, that is homozygous for the ~ deletion. E18 may be preferred over ~tpi#29 as a host strain for a plasmid because it has no amino acid requirements, has larger cells, and grows ~;,z99~z~

faster. These tpi strains are deleted for the genetic material which codes for the glycolytic functlon and are, therefore, es~pected to be nonreverting (i.e., stable) mutants.

C. Transformation of the POTl gene into S. cerevisiae tpi deletion strains.

The plasmids pFPOT and pFATPOT were trans~ormed into ~tpi#29 and related tpi deletion strains. The yeast mutants were grown aerobically overnight to late log phase 10 in YEP-2~ Galactose at 30. Transformation conditions were as described by Beggs (ibid.), except that the cells were allowed to recover at 30 for 1-2 hours in lM Sorbitol containing YEP-3% Glycerol-1% Lactate or YEP-2% Galactose, instead of YEP-Dextrose, before plating the cells in top 15 agar. The top agar and plates contained synthetic, modi-fied Wickerham's medium with lM Sorbitol and 2% Dextrose.
After three days at 30, transformants were visible and were picked out of the agar for replating onto YEPD.
Thereafter, the transformants were maintained on YEPD or 20 other complex media containing dextrose.

Strain E18 transformed with pFATPOT was designated ZYM-3.
It has been deposited with ATCC under accession number 20699.

Stability of pFPOT and pFATPOT on complex media. To study 25 plasmid stability, colonies from a single celi were inocu-lated into tubes containing YEPD and allowed to grow to a total population of 109 cells (approximately 30 divisions).
The yeast cells were sonicated to break up clumps, diluted to appropriate numbers, and plated onto YEP-2% Galactose or 30 YEP-2% Glycerol-1% Lactate, which allows the growth of tpi cells (with or without the plasmids carrying the POT1 gene). The colonies which arose on YEP-Galactose were then replica plated onto YEPD to screen for the loss of the plasmid (l.e., tpi cells which have lost the . ~

lZ9~1~0 POT1-containing plasmid will not grow on dextrose). The xesults, summarized in Table 2, indicate that the pFPOT and pFATPOT plasrnids are stable in the yeast tpi deletiGn strains. They are surprisingly much more stable than yeast plasmids containing centromeres. Centromere-bearing plasmids (which are low in copy number) are among the mGst stable plasmids reported for yeast and are generally lost at a frequency of around 1~ of cells per division on complex media (see Murray and Szostak, ibid., for a review 10 of centromere plasmid stability). As Table 2 indicates, the POTl plasmids described herein are lost at a frequency of less than 1% after 30 divisions on complex media in tpi deletion strains.

D. Expression of human alpha-1-antitrypsin in S. cerevis-15 lae using POTl plasmids To test the use of the POT1 plasmids for enhancing expres-sion of foreign proteins in a transformed yeast, plasmids pFATPOT and pFAT5 were used to transform S. cerevisiae strains ~tpi#29 and E2-7B respectively. Transformed cells 20 were selected in leucineless media containing dextrose.
Cultures were grown at 30 to an O.D.600 of 3-4. Celi extracts were prepared and assayed for AT as described in Example 5.

AT produced by pFATPOT/~tpi#29 represented 4-6% of total 25 soluble protein. AT produced by pFAT5/E2-7B represented 2-3% of total soluble protein.

Although plasmid copy numbers are difficult to accurately measure and represent a population average, empirical observations of gene product quantities provide an indica-30 tion of relative plasmid levels, given that the expressionunit (promoter, gene of interest, terminator) remains the same. pFATPOT therefore appears to be functionally greater in number than pFAT5, from which it was derived. Because the two transformed strains are nearly identical lZ9~0 genetically (~tpi#29 being derived from E2-7B by plasmid-directed mutagenesis) and were grown under the same conditions, these results are indicative of the value of the herein~ described stable plasmid expression system over previously described vectors.

TABLE 1: STABILITY OF CDC4 PLASMIDS

Isolate Cv onies CDC4 Trp Leu l(BELL 1) 123 123 123 123 3 ~3 80 80 83 ~8 88 88 88 Cells were grown in liquld complex medium (YEPD) at 25 for 30 generations, then plated on YEPD at 25.
bCells were replica plated to YEPD at 37. Cells lacking an intact CDC4 gene failed to grow at this (restrictive) temperature.
20 CCells were replica plated to medium lacking tryptophan.

dCells were replica plated to medium lacking leucine.

~299~2o TABLE 2: STABILITY OF POTl PLASMIDS VS. pTPIC10 Total Experiment Plasmid/Strain Coloniesa TPI ~ Loss 1 pTPIC10/Qtpi#29 234163 30.3 2 pFPOT/Qtpi#29 308 308 0 3 pFATPOT/Qtpi#29 471471 0 4 pFATPOT/E18(ZYM-3) 1104 1104 0 pFATPOT/E18(~Y~l-3l 634 632 0.32 6 pFATPOT/Qtpi#29 426426 0 2-6 pooled data 2943 2941 0.07 aThe plasmid/strain combinations were grown on YEPD plates until easily visible colonies of approximately lQ4 to 105 cells were seen. These colonies were used to inoculate 6ml 15 Of YEPD liquid medium. The cultures were grow~ aerobically overnight to a cell density of 1-3x10~ cells/ml and were plated onto YEP-2%Glycerol-1%Lactate or YEP-2~Galactose.
Each of these media would allow tpi strains to grow, although the resulting tpi colonies arose more slowly than 20 tpi colonies. Only 100-300 cells were distributed on each plate so that each colony (whether tpi or tpi+) would be countable.

bThe colonies were replica plated onto synthetic media containing dextrose at a 2% final concentration. Cells 25 which had lost the triose phosphate isomerase gene on the plasmids were unable to grow.

CThe "% Loss" represents the frequency of ce]ls that had lost the plasmid after nearly 30 divisions in YEPD. The pooled data for experiments 2 to 6 indicate that the POTl 30 plasmids are extremely stable over these many divisions and are lost at a combined frequency well below 1% in 30 cell doublings.

1;~99~20 Example 3 Preparation of Plasmid C1/_ C1/1 was constructed from plasmid pJDB248 (Beggs, J., Nature 275, 104-109 (1978)). The pMB9 sequences were removed from pJDB248 by partial digestion with Eco RI and were replaced by pBR322 DNA which was cut with Eco RI. The restriction map of Cl/1 is given in FIG. 6. The C1/1 plasmid contains the entire 2-micron DNA from yeast (S.
cerevisiae), with a pBR322 insertion at an EcoRI site. It 10 also contains the LFU2 gene.

Example 4 PreDaration of Plasmid pFAT5 _ The gene coding for the predominant form of human alpha-1-antitrypsin (AT) was isolated from a human liver cDNA
15 library by conventional procedures using the baboon sequence (Kurachi et al., Proc. Natl. Acad. Sci. USA 78:
6826-6830, 1980; and Chandra et al., Biochem. Biophys. Res.
Comm. 103: 751-758, 1981) as a DNA hybridization probe.
The library was constructed by inserting human liver cDNA
20 into the PstI site of the plasmid pBR322 (Bolivar et al., Gene 2: 95-113, 1977). The AT gene was isolated from the library as a 1500 base pair (bp) PstI fragment. This fragment was inserted into the PstI site of pUC13 to produce the plasmid pUC~1. In pUC1, the AT sequence is 25 flanked on the 3' end by XbaI and EcoRI sites in the polylinker.

The TPI terminator was purified from plasmid pFG1 (Alber and Kawasaki, ibid) as a XbaI-EcoRI fragment of approxi-mately 700 bp and inserted into pUCl which had been 30 cleaved with XbaI and EcoRI. This construct was then cut with EcoRI, and oligonucleotide linkers (sequence:
AATTCATGGAG
5TACCTCCTAG) were added, in multiple linked copies, to provide a BamHI site to the 3' end of the TPI terminator.
35 The resultant plasmid is known as BAT5.

,.

'1299~20 The TPI promoter fragment was obtained from plasmid pTPIC10 (Alber and Kawasaki, ibid). This plasmid was cut at the unique KpnI site, the I'PI coding region was removed with Bal31 exonuclease, and an EcoRI linker (sequence: GGAATTCC) was added to the 3' end of the promoter. DigestiGn with BglII and EcoRI yielded a TPI promoter fragment having BglII and EcoRI sticky ends. This fragment was then joined to plasmid YRp7' (Stinchcomb, et al. Nature 282: 39-43, 1979) which had been cut with BglII and EcoRI. The result-lO ing plasmid, TE32, was cleaved with EcoRI and BamHI toremove a portion of the tetracycline resistance gene. The linearized plasmid was then recircularized by the addition of the above described EcoRI-BamHI linker to produce plasmid TEA32. TEA32 was then cleaved with ~glTI and 15 BamHI, and the TPI promoter was purified as a fragment of approximately 900 bp.

To construct plasmid pFAT5, plasmid Cl/l was linearized with BamHI, and was joined to the 900 bp TPI promoter fragment from TEA32. The resulting construct, known as 20 plasmid F, has a unique BamHI site located at the 3' end of the TPI promoter. This plasmid was cut with BamHI and a 2200 bp BamHI fragment, comprising the AT coding sequence and TPI terminator, was purified from BAT5 and inserted into the BamHI site. The resulting plasmid, known as 25 pFAT5, is illustrated in Figure 7.

Example 5 Assay for Alpha-l-Antitrypsin As a control, 10 microliters (1 microgram) of a solution of 100 microgram/ml trypsin, 100 microgram (lOO microliters) of bovine serum albumin and 100 microliters of 0.05 molar TRIS, pH 8.0 buffer containing lmM benzoylargininoyl-p-ni-troanilide were mixed, and the increase in absorbance at 405 nm was measured over time in a spectrophotometer. The absorbance value of this solution was used as a standard for 100~ trypsin activity. All assayed samples contain equal concentrations of substrate and bovine serum albumin.

Claims (16)

1. A method for producing alpha-1-antitrypsin in a microorganism host cell having a deficiency in a function necessary for normal cell growth on complex media comprising the steps of:
(a) transforming said microorganism host cell with a DNA molecule comprising a gene which complements said deficiency and a sequence coding for said alpha-1-antitrypsin;
(b) culturing the transformants from step (a) in a growth medium which need not contain antibiotics or heavy metals, and need not be depleted of specific nutrients, under conditions whereby said gene functions as a selectable marker for transformant cells.

2. A method according to Claim 1 wherein said gene is a gene required for host cell division, cell wall biosynthesis, membrane biosynthesis, organelle biosynthesis, protein synthesis, carbon source utilization, RNA transcription, or DNA replication.

3. A method according to Claim 1 wherein said gene is a gene required for host cell division, cell wall biosynthesis, membrane biosynthesis, organelle biosynthesis, protein synthesis, carbon source utilization, RNA transcription, or DNA replication and said gene is selected from the group consisting of genes of the yeast cell division cycle and genes of the yeast glycolytic pathway.

4. A method according to Claim 3 wherein said gene is a yeast CDC4 gene.

5. A method according to Claim 3 wherein said gene is a Schizosaccharomyces pombe triose phosphate isomerase gene.

6. A method according to Claim 1 wherein said gene is from a cell species different from said host cell.

7. A DNA construct comprising a gene which complements a deficiency in a microorganism host cell, said deficiency being in a function necessary for normal cell growth on complex media, and a DNA sequence coding for alpha-1-antitrypsin.

8. A DNA construct according to Claim 7 wherein said gene is a gene required for host cell division, cell wall biosynthesis, organelle biosynthesis, protein synthesis, carbon source utilization, transcription or replication.

9. A DNA construct according to Claim 8 wherein said gene is selected from the group consisting of genes of the yeast cell division cycle and genes of the yeast glycolytic pathway.

10. A DNA construct according to Claim 9 wherein said gene is a yeast CDC4 gene.

11. A DNA construct according to Claim 9 wherein said gene is a Schizosaccharomyces pombe triose phosphate isomerase gene.

12. A transformant microorganism strain containing a DNA
construct according to Claim 7, 8 or 9.

13. A transformant microorganism strain containing a DNA
construct according to Claim 10 or 11.

14. A method according to Claim 1 wherein said host cell is a yeast cell.

15. A DNA construct according to Claim 7 wherein said host cell is a yeast cell.

16. A transformant strain according to Claim 12 wherein said strain is a yeast.