EP0449839A1 - Improved bacterial strains for heterologous gene expression - Google Patents

Abstract

La présente invention concerne une nouvelle souche bactérienne permettant d'améliorer l'expression de gènes hétérologues.The present invention relates to a new bacterial strain for improving the expression of heterologous genes.

Description

IMPROVED BACTERIAL STRAINS FOR HETEROLOGOUS GENE EXPRESSION BACKGROUND OF THE INVENTION 1. Field of the Invention

This invention relates to novel bacterial strains, and to methods of producing and using such novel strain. More particularly, this invention relates to such a strain which has been genetically modified to have improved characteristics for heterologous gene expression. 2. Prior Art

Recombinant DNA technology developed within the past two decades has provided a number of methods for identifying, isolating and expressing genes of research or commercial value. Advances in our understanding of gene structure and function in procaryotes and eucaryotes in particular have allowed the deliberate design and implementation of gene expression strategies for overproduction of single proteins in a variety of systems such as bacterial cells.

The most widely used bacterial gene expression systems are those in which E. coli cells are used as hosts and heterologous gene expression occurs by transcription and subsequent translation from mRNA of genes carried on a plasmid or bacteriophage cloning vector. E. coli gene expression hosts desirably meet one or more of several genetic and biochemical criteria for optimal production of foreign proteins. These criteria include provisions for controlling heterologous gene expression in a regulatable manner for any foreign DNA inserted into an expression vehicle, a requirement for a recombination-deficient phenotype, stabilization of heterologous mRNA, and stabilization of heterologous protein product. Such criteria allow one to optimize the yield of heterologous protein produced by the host strains. Transcription of heterologous genes in E. coli has been controlled by any of a variety of genetic promoter elements placed in front of the foreign DNA on the relevant expression vectors. A popular choice of promoter element in this regard has been the leftward promoter p of bacteriophage lambda [ref. R.W. Hendrix, J. . Roberts, F.W. Stahl and R.A. Weisberg (eds.), Lambda II Cold Spring Harbor Laboratory (1983)]. This promoter can be repressed by the protein product of the lambda _cl_gene; this protein binds to the p^ promoter and" prevents the bacterial RNA polymerase from initiating transcription. For some regulatable systems, a thermolabile form of the cl protein due to the c,l857 mutation [(M. Rosenberg et al., Methods in Enzymology, Vol. 101, R. u et al. (eds.), pp. 123-138 (Academic Press, 1983)], is sufficient to control expression from the p_L promoter by changing the temperature^* of the bacterial culture. For the scope of the present invention, the cl857 gene can be either on the expression vector, on another plasmid replicon, or contained in the chromosome of the host strain, usually by lysogenization with a lambda bacteriophage carrying the cl857 gene. For example, see S. Gottesman et al. , J. Mol. Biol. 140, pp. 57-75 (1980). Representative vectors contained a p^ promoter which are compatible with controlled expression through use of the cl857 gene include those found on plasmid pHUB2 [N.C. Franklin, J. Mol. Biol. 89, pp. 33-48 (1974)], pKC3θ^'_F. Sanger et al., J. Mol. Biol. 162, pp^'. 729-733 (1982)], pAS1 (M. Rosenberg et al., supra) and pJLβ [J.A. Lautenberger et al., Gene 23, pp. 75-84 (1983)]; also see U.S. Patent Application 6,511,108, filed July 6, 1983 by T.S. Papas et al.

The successful combined application of the cl857 gene on a defective lambda prophage and plasmid derivatives of pJL6 or similar plasmid constructs for heterologous gene expression has been previously shown (See M. Rosenberg et al. and references therein). However, none of this literature specifically discusses the cloning and expression of genes coding for structural proteins or genes with highly repetitive DNA sequences. Any advantages of the use of E. coli recombination- deficient bacterial hosts such as those lacking a functional RecA protein for gene expression from any of the recombinant plasmids claimed within the scope of the present invention is also not disclosed nor discussed in the above-referenced literature. For example,

Lau^'tenberger et al. utilize E. coli strain N4830 which is commercially available and has a recA^"1" genotype. The RecA protein is a regulatory protein in E. coli that is also required for homologous recombination. For review see A. J. Clark, Annu. Rev. Genetics 7, pp. 67-86

(1973). A functional RecA protein can be undesirable in bacterial gene expression hosts because cloned DNA sequences may contain regions of homology which in turn might serve as substrates for RecA-dependent recombination. This process may thus alter and/or delete part of the cloned sequence. Three recA mutations are commonly found in bacterial cloning strains: recA56, recA1 , and recA13. All three mutations were isolated by random mutagenesis procedures and have been shown to represent point mutations. See P. Howard-Flanders, Genetics 53, pp. 1137-1150 (1966). Although these mutations can be commonly found in various E. coli cloning or gene expression hosts (e.g., the commercial strains HB101 or JM109), they do not necessarily completely abolish homologous recombination activity. Another problem encountered with point mutations in the recA gene is the tendency to revert back to a RecA⁺ phenotype. Such revertants would be expected to no longer stabilize repetitive DNA sequences such as those included within the scope of the present invention. It is therefore preferred to select and use a mutated recA gene in E. coli expression hosts of the class covered by the present invention that eliminates RecA recombination activity as measured by some applicable sensitive method and in addition that is nonrevertable. Mutations which best fulfill these conditions are deletions of the recA structural gene which functionally inactivate RecA protein activity. D.K. Willis et al., Mol. Gen. Genet. 183, pp. 497-504 (1981) have described a pair of mutations designated srlR3Q1 : ;Tn10 Δ(srlR-recA)3-6. The end result of these two mutations which were constructed sequentially is a 90-100? deletion of the recA gene which cannot be separated by recombination from Tn10 using P1 transduction. Previous reports in the open literature or disclosure in relevant patent applications, such as those summarized in U.S. Patent Application Serial No. 001,292, filed Jan. 7, 1987 by J.I. Williams et al., which described applications of thermoinducible gene expression promoter systems such as the lambda p^ -cl857 binary combination, rarely emphasize any need for or desirable properties obtained from carrying out gene expression manipulations in recA hosts.

The construction of gene expression plasmid pAC1 is described in detail in U.S. Patent Application Serial No. 001,292, filed January 7, 1987. This plasmid is a derivative of pJLδ which contains a 335-base pair (bp) insert of the repeating DNA sequence GGTCCGCCG which was cloned into the Clal restriction enzyme recognition site. This insert was prepared from two overlapping synthetic oligodeoxynucleotides each 18 bp long, which were ligated to each other in a tandem array prior to insertion into pJLδ. The final construct codes for a fusion protein consisting of the first 14 amino acids of the lambda ell protein followed by approximately 33 repeats of a gene sequence encoding (Gly-Pro-Pro). This fusion protein is referred to in subsequent sections of the description of the present invention embodied herein as a collagen analog peptide. This collagen analog peptide has been expressed as described in the aforementioned patent application of J.I. Williams et al.. in E. coli strain DC1139A. The strain DC1139A harbors a defective lambda prophage containing the cl857 gene and a recA mutation to inhibit homologous recombination, but no genetic modification of DC1139A to inhibit endogenous proteolysis nor discussion of advantageous qualities to be obtained by such genetic modification has been performed or discussed prior to the present disclosure.

It has been shown in the literature that the stability of bacterial proteins can be affected by their susceptibility to bacterial proteases. Protease La is one of several proteases in E. coli, the product of the Ion gene. See C.H. Chung and A.L. ^* Goldberg, Proc. Natl Acad. Sci. U.S.A. 78, pp. 4931-4935 (1981). S. Gottesman et al., Cell 24, pp. 225-233 (1981) determined that protease La plays a major role in the degradation of abnormal proteins in E. coli. However, protease La is not the only ATP-dependent protease capable of degrading abnormal proteins in E. coli. (See M.R. Maurizi et al., J. Bacteriol. 164, pp. 1124-1135 (1985)). Also, protease La does not degrade small peptides such as insulin. Small peptides such as insulin are degraded by a second class of endoproteases which are metalloproteases and have no activity against larger proteins like casein or globin. Still a third class of endoproteases which requires no ATP is active on large proteins such as casein and globin. (For review see A. L. Goldberg and S.A. Goff, In Maximizing Gene Expression pp. 287-314, Reznikoff and Gold (eds.), Butterworth Publishers, Stoneham MA (1986).

It was later discovered that protease La is one of a class of proteins whose expression is modulated by temperature; these proteins are collectively known as heat shock proteins. See F. Neidhardt et al. , Ann. Rev. Genet_.. 18, pp. 295-330 (1984). Transcription of the genes encoding these proteins is coordinately regulated by the product of the htpR (rpoH) gene. See T. Yamamori and T. Yura, Proc. Natl. Acad. Sci. U.S.A. 79, pp. 860- 864 (1982) and F. Neidhardt et al. J. Bacteriol 153, pp. 597-603 (1983). It is unclear, however, whether any other proteases are among the proteins coordinately controlled by the htpR locus. These results led to the speculation that degradation of abnormal proteins should be signi icantly reduced in htpR mutant strains. T.A. Baker et al., Proc. Natl Acad. Sci. U.S.A. 81 , pp. 6779-6783 (1984) describe a mutation in the htpR locus called htpR165. This allele ^*is commonly used in the study of bacterial proteolysis and has been maintained in a supC (Ts) host strain. Cells harboring these two mutations are viable at 30°C but eventually die at 42°C. Baker et al. showed for two unstable bacterial proteins that degradation is almost completely inhibited in the presence of the htpRI65 allele (protein half- lives greater than 60 min.). Similar observations are summarized in Table II of J.F. Kane and D.L. Hartley, Trends in Biotechnology 6, pp. 95-101 (1988). In particular, G. Buell et al. , Nucleic Acids Res. 13, PP« 1923-1938 (1985) disclosed the increased expression of a synthetic gene encoding somatomedin-C (IGF-1) utilizing a lambda P_j_-cl857 promoter-repressor system, in hosts defective in Ion and/or htpR. This increased expression of somatomedin-C in protease La-deficient strains contrasts with the reported lack of protease La activity against the structurally analogous and similarly-sized protein insulin (Goldberg and Goff and references therein). Furthermore, Baker et al. reported that the htpRt65 allele does not inhibit the degradation of every unstable protein. For example, htpRI 65 does not affect the rate of degradation of the unstable lambda ell gene product. It therefore cannot be predicted whether heterologous or hybrid protein expressed from a lambda p^ promoter in E. coli will be stabilized by the htpRI65 allele, nor whether genetic modification to include the htpRI65 allele in an E. coli gene expression host that is recA^" and utilizes a lambda PL "________ promoter-repressor system for foreign gene expression will lead to stabilized production or accumulation of the foreign protein.

SUMMARY OF THE INVENTION This invention relates to novel bacterial strains for heterologous gene expression. More particularly, this invention relates to a novel bacterial strain comprising: a. One or more heterologous genes which code for the production of polypeptides composed of repeating amino acid sequences, or unique amino acid sequences encoded by heterologous gene(s) containing recombinogenic regions of DNA; b. Controlling means for controlling the activity of said polypeptide production genes; c. Bacterial proteolysis retarding means for retarding the proteolysis of polypeptides produced by said heterologous genes; and d. Gene stabilization means for stabilizing the heterologous genes. This invention has provided a novel strain of _E_. coli which includes a combination of one or more heterologous genes having internally repetitive or quasi-repetitive DNA sequences which code for the production of polypeptides having repeating amino acids together with three host attributes which particularly favor expression of such heterologous genes. These three elements consist of (1) a controlling means to elevate transcription of the polypeptide-encoding DNA sequence under the desired conditions, (2) bacterial proteolysis retarding means for retarding peptide proteolysis and, (3) a heterologous gene stabilization means.

As a result, this strain is especially suitable for expression of heterologous genes, notably natural, synthetic or semi-synthetic genes with exact or homologous but not exact internally repetitive DNA sequences which code for the production of polypeptide having repeating or quasi-repeating amino acid sequences or for certain heterologous genes coding for unique protein sequences wherein said genes contain recombinogenic DNA sequences.

BRIEF DESCRIPTION OF THE FIGURES Figure 1 shows the synthesis of collagen analog peptide in E_. coli IG110 (NRRL No. B-18352). Cellular proteins were labeled with [ 1 C]pr'oline and electrophoresed on a 12.5? SDS-polyacrylamide gel. The band representing the collagen analog peptide is indicated by an arrow and has an apparent molecular weight of 22 kilodaltons (kDa). Molecular weight markers were albumin (M_r 69,000), ovalbumin (M_r 46,000), carbonic anhydrase (M_r 30,000), lactoglobulin A (M_r. 18,367) and cytochrome C (M_r 12,300). Figure 2 illustrates the fate of collagen analog peptide in _E_. coli IG109(pAC1) and in _E_. coli IG110 (NRRL No. B-18352) (pAC1). Cells were pulsed with 5 μCi of [ C]proline for 3 min. and chased for various periods of time with 1 mg of unlabeled proline. The cells were processed as described in Comparative Example 1 and aliquots were electrophoresed on 12.5? SDS - polyacrylaraide gels. The intensity of the band corresponding to the 22 kDa collagen analog peptide at each time point was determined by densitometry. For both strains, the amount of collagen analog peptide observed after 5 min. of chase time was designated as 100? protein remaining.

DESCRIPTION OF THE PREFERRED EMBODIMENTS This invention relates to a novel strain of bacterial host organism having unique characteristics. The bacterial host organism may vary widely. Illustrative of useful bacterial strains are _E_. coli, Pseudomonas aeruginosa, Bacillus subtilis, Bacillus stearothermophilus, Salmonella typhimurium and the like. The preferred bacterial organism is E_. coli.

One feature of the novel strain of this invention is one or more heterologous genes which code for the production of polypeptides composed of repeating or quasi-repeating amino acid units. The nature of the polypeptide-encoding genes can vary widely, the only requirement is that the DNA sequences code for the production of polypeptides composed of repeating or quasi-repeating amino acid sequences, and that the genes contain the same or substantially the same internally repetitive DNA sequence. For the purposes of the invention described herein, the internally repeated DNA sequences share homologies equal to or greater than about 65? to about 70?. The degree of repetitiveness can be judged by DNA or protein sequence homology using various theoretical techniques in molecular biology. See, for example, S.B. Needleman and CD. Wunsch, Journal of Molecular Biology 48, pp. 443-453 (1970), A.D. McLachlan, Journal of Molecular Biology 61 , pp. 409-424 (1971), and D. Eisenberg et al. , Proc. Natl. Acad. Sci. (U.S.A.) 81, pp. 140-144 (1984).

In the preferred embodiments of the invention, polypeptide production genes of choice are those which code for the production of primary amino acid sequences which confer secondary structures characteristic of α-helix, polyproline helix, β-sheet and/or 3-turn or combinations thereof. Illustrative of polypeptide production genes which are useful in the practice of this invention are naturally-occurring genes or gene fragments such as those which code for part or all of any form or isolate of the proteins collagen, trematode egg shell dopa- proteins (e.g., Fasciola hepatica and Schistosoma mansoni), insect salivary gland silk/adhesive proteins, bioadhesive proteins from marine crustaceans, such as Mytilis edulis, M. californianus, Geukensia demissa, elastin, keratin, troponin C, any other intermediate filament protein [cf. E. Lazarides, Nature 283, pp. 249- 256 (1980)] or silk fibroin and which includes most or all of an amino acid sequence which exhibits some degree of repetitiveness within the protein sequence. Useful natural genes and gene fragments include complementary DNA resulting from reverse transcription and DNA strand copying from messenger RNA by an appropriate reverse transcription process and DNA strand copying process wherein the messenger RNA is transcribed from genes coding for proteins such as collagen, bioadhesive proteins from marine crustaceans, such as Mytilis edulis, M_. californianus, Geukensia demissa, elastin, keratin, troponin C, any other intermediate filament, or silk fibroin. These illustrative natural gen^*es and gene fragments useful in the practice of this invention are representative and are not meant to be inclusive of all useful naturally-occurring genes. Natural genes for use in this invention will preferably be prepared for isolation using a restriction enzyme such as BamHI, Bglll, EcoRI, Hindlll, Xbal, and the like, which leaves cohesive termini on the natural DNA fragments compatible with the cohesive termini on other DNA fragments. Alternatively, the ends of any natural DNA fragments preferably may be adapted or modified with an appropriate DNA linker or linkers which subsequent to attachment to the natural DNA fragments can either be uniquely cleaved with one or more restriction enzymes to reveal or intrinsically has one or more cohesive termini compatible with the cohesive termini of one or more other DNA fragments.

Also useful as polypeptide production genes in the practice of this invention are synthetic sequences. Suitable synthetic genes may vary widely depending on the desired repeating polypeptide. Illustrative of useful synthetic genes are those which code for the production of such polypeptides which comprise one or more blocks of recurring units of the formula -(Gly)-, -(Ala)-, -(Gly-Ala)-, -(Ala-Lys)-, -(Gly-Ala-Gly-Ala- Gly-Ser)-, -(Gly-Ala-Pro)-, -(Gly-Pro-Ala)-, -(Gly-Pro- Pro)-, -(Gly-Val-Gly-Val-Pro)-, -(Gly-Lys-Leu-Glu-Ala- Leu-Glu)-, -(Ala-Lys-Pro-Thr-Tyr-Lys)-, -(Ala-Lys-Pro- Ser-Tyr-Pro-Pro-Thr-Tyr-Lys)-, and the like wherein each amino acid residue has the L-amino acid conformation. Hydroxylated forms of any of these sequences are also preferable within certain embodiments of this invention. In the preferred embodiments of the invention, synthetic polypeptide production genes of choice are those which code for the production of poly(Gly-X-Y) , poly(Gly-Pro-X) , poly(Gly-X-Pro) , poly(X-Pro-Gly-Y-Gly) , poly(X-Pro-Gly-Gly) , poly(X-Pro-Gly-Val-Gly-Y) , Poly[(Ala)i_|-Lys-Ala-Ala-Lys-(Phe/Tyr)-Gly-Ala], polyt(Ala)₂-Lys-(Ala^-Lys-CAla)₂], poly(Gly-Ala-Gly- Ala-Gly-Ser) , and poly(Ala-Lys-Pro-Ser-Tyr-Pro-Pro-Thr- Tyr-Lys) where X and Y are the same- or different and each is an amino acid. In the particularly preferred embodiments, the synthetic polypeptide production genes of choice are those which code for the production of poly(Gly-Pro-Pro) , poly(Gly-Val-Gly-Val-Pro) , poly(Ala- Lys-Pro-Ser-Tyr-Pro-Pro-Thr-Tyr-Lys) and sequential permutations thereof. Semi-synthetic genes can also be used in the practice of this invention. Semi-synthetic genes are natural genes or gene fragments connected to synthetic genes to give chimeric genes encoding blocks of repeating or quasi-repeating amino acids. For example, naturally-occuring gene fragments such as those which code for part or all of any form or isolate of the proteins collagen, trematode egg shell dopa-proteins, insect salivary gland silk/adhesive proteins, - bioadhesive proteins from marine crustaceans, elastin, keratin, troponin C, any other intermediate filament protein or silk fibroin can be jointed directly or by the use of DNA linkers or adaptors to synthetic genes such as those which code for the production of poly(Gly), poly(Ala), poly(Gly-Ala) , poly(Ala-Lys) , poly(Gly-Ala-Gly-Ala-Gly-Ser) , poly(Gly-Ala-Pro) , poly(Gly-Pro-Ala) , poly(Gly-Pro-Pro) , poly(Gly-Val-Gly- Val-Pro), poly(Gly-Lys-Leu-Glu-Ala-Leu-Glu) , poly(Ala- Lys-Pro-Thr-Tyr-Lys) , poly(Ala-Lys-Pro-Ser-Tyr-Pro-Pro- Thr-Tyr-Lys) and the like. Construction of semi- synthetic genes is constrained by the requirement for maintenance of proper reading frame among the various natural and synthetic gene or gene fragment components and may require the use of DNA adaptors or linkers to satisfy this requirement. Useful DNA linker or adaptor sequences are usually less than about 30 bp in length and may or may not encode peptides similar in primary or secondary protein structure to the adjoining peptides. In the preferred embodiments of the invention, semi-synthetic polypeptide production genes of choice are those which code for the production of polypeptides having one or more synthetic blocks selected from the group consisting of poly(Gly-X-Y) , poly(Gly-Pro-X) , poly(Gly-X-Pro) , poly(X-Pro-Gly-Y-Gly) , poly(X-Pro-Gly- Gly), poly(X-Pro-Gly-Val-Gly-Y) , poly[(Ala)iι-Lys-Ala- Ala-Lys-(Phe/Tyr)-Gly-Ala], poly[(Ala)₂~Lys-(Ala)3-Lys- (Ala)₂3, poly(Gly-Ala-Gly-Ala-Gly-Ser) , and poly(Ala- Lys-Pro-Ser-Tyr-Pro-Pro-Thr-Tyr-Lys) where X and Y are the same or different and each is an amino acid, and having one or more naturally-occurring blocks selected from the group consisting of collagen, elastin, trematode egg shell dopa-proteins, silk fibroin, insect salivary gland silk/adhesive proteins and bioadhesive proteins.

In the particularly preferred embodiments, the semi-synthetic polypeptide production genes of choice are those which code for the production of polypeptides having one or more blocks selected from the group consisting of poly(Gly-Pro-Pro) , poly(Gly-Val-Gly-Val- Pro), poly(Ala-Lys-Pro-Ser-Tyr-Pro-Pro-Thr-Tyr-Lys) and sequential permutations thereof, and having one or more blocks selected from the above-listing of naturally- occurring polypeptides. The size of the polypeptide production gene may vary widely and depends on the molecular weight of the desired polypeptide. In the preferred embodiments of the invention, the length of the polypeptide production gene is at least about 75 bp in length, and in the particularly preferred embodiments, the gene is at least about 100 bp in length. In the most preferred embodiments of this invention, the size of the poly- peptide production gene 'is at least about 500 bp in length.

As a second essential feature, the novel strain of _■_• - ---- provided by this invention includes a controlling system for initiating the activity of the heterologous gene. Controlling systems useful in the practice of this invention may vary widely. Illustrative of useful controlling systems are those which are responsive to temperature change, changes in nutrients, addition of a foreign RNA polymerase, presence or absence of antibiotics, changes in the levels of intracellular compounds involved in intermediary metabolism and the like. For example, the lambda p^ (^ΛPL^ promoter in conjunction with a cl857 mutation which codes for a temperature-sensitive repressor can be used to form a genetic switch which is responsive to changes in temperature. Lambda p_L is usually present on expression vectors containing ColE1 or Rl origins of replication. Such plasmids have a convenient restriction site following the λ p_L promoter into which the heterologous gene can be inserted using conventional genetic engineering techniques. See G.D. Stormo, T.D. Scheider and L.M. Gold, Nucleic Acids Research 10, pp. 2971-2996 (1982); A. Shatzmans Y.S. Ho and M. Rosenberg in Experimental Manipulation of Gene Expression, M. Inouye, (eds.), pp. 1-14 (Academic Press, 1983); A. Rattray, S. Altuvia, G. Mahagna, A.B. Oppenheim and M. Gottesman, Journal of Bacteriology 159, pp. 238-242 (1984). A ribosome binding site and an AUG initiator codon may follow the p_L promoter which allows for correct initiating of translation of the heterologous gene(s). The controlling system and the plasmid can be introduced into E. coli using conventional genetic engineering techniques. See J.G. Sutcliffe and F.M. Ausubel in Genetic Engineering, A.M. Chakrabarty (eds.), pp. 83-111 (CRC Press, 1978) and R. Wu, L-H. Guo and R. C. Scarpella in Genetic Engineering Techniques, P.C. Huang, T.T. Kuo and R. Wu, (eds.), pp. 3-21, (Academic Press, 1982). The temperature-senstive cl857 mutation can be introduced into suitable E_. coli hosts in a variety of ways. For example, the mutation can either be introduced on a compatible plasmid or episome (such as p15A-derived plasmids for ColE1 or Rl- derived plasmids carrying the λp_L promoter) or be present directly on the cloning vector. Alternatively, the temperature-sensitive cI587 mutation can be supplied on the bacterial chromosome by, for example, a lysogenic lambda phage. For the establishment of a good expression system, this phage should have the following properties. It should be unable to excise from the bacterial chromosome either spontaneously or at high temperature, it should produce enough cl repressor so that its concentration is not titrated out by the λp_L promoter carried on the plasmid, and it should not synthesize any significant amounts of phage-related proteins under inducing conditions which could reduce the yield of heterologous protein. It may be desirable for the phage to express a functional N protein and for the plasmid to contain an N utilization sequence between the PL promoter and the heterologous gene when transcription termination signals are present in the heterologous gene sequence. Using this arrangement, the antitermination function of N leads to significantly higher expression of the heterologous protein(s). See M. Rosenberg et al. , supra.

Illustrative of a controlling system which depends on the presence of an inducer is plasmid pKK233-2 [E. Amann and J. Brosius, Gene 40, pp. 183-190 (1985)]. This controlling system allows for high-level production of heterologous proteins in an unfused state. The plasmid pKK233-2 contains the trc promoter, which is a trp-lac fusion promoter and differs from the tac promoter by addition of 1 bp between the -10 and -35 regions in order to arrive at the consensus 17-bp spacing. The trc promoter is followed by the lacZ ribosome-binding site and an (ATG) translation initiation codon. The pKK233-2 plasmid is normally maintained in a lacl^*^ strain for tight repression of the trc promoter. Instead of high temperature, this promoter is induced by addition of galactosides such as isopropyl thiogalactopyranoside (IPTG) to the bacterial culture. This controlling system can be introduced into suitable _E_. coli hosts by standard transformation techniques, for example D. Hanahan, J. Mol. Biol. 166, pp. 557-580 (1983), or by electroporation.

Illustrative of switches which rely on the depletion of a nutrient in the medium in order to control the heterologous gene is plasmid pWT121. [M.T. Doel et al. , Nucleic Acids Research, _8_, pp. 4574-4592 (1980)]. This plasmid is responsive to the absence of tryptophan. This genetic switch becomes activated upon depletion of tryptophan from the culture medium, or by the addition of β-indoleacrylic acid. The genetic switch can be introduced into suitable E_. coli hosts by the same transformation techniques shown above. A controlling system which depends on the availablity of a foreign RNA polymerase is exemplified by the T7 expression system [A.H. Rosenberg et al. , Gene 56, pp. 125-135 (1987)]. The DNA sequences defining _E. coli and T7 promoters are quite different, and a T7 promoter is not recognized by _E_. coli RNA polymerase. Vectors have therefore been constructed, like pET-3a, where the heterologous gene can be placed under the control of the T7 promoter. The gene would not be transcribed unless T7 RNA polymerase is supplied. This is generally achieved in any of three ways: the gene for the T7 polymerase can be present on the host chromosome under control of a regulatable _E_. coli promoter; it can reside on a compatible plasmid under control of a regulatable _E_. coli promoter; or it can be introduced on a phage λderivative which carries the gene for T7 RNA polymerase under control of a promoter in λDNA.

In the practice of this invention, preferred controlling systems include but are not limited to those discussed above.

As a third essential feature, the E_. coli strain of this invention includes bacterial proteolysis-retarding means. The nature of such means may vary widely. Illustrative of suitable proteolysis-retarding means is to utilize strains defective in the production of one or more proteases by way of mutation in appropriate loci which may be responsible for encoding proteases, and/or for regulating the formation, activity or degradation of these proteases. Such an approach has been disclosed in PCT Publication No. 85 03949 and U.S. Patent No.

4,758,512 to A. L. Goldberg et al. wherein the use of E. coli strains containing mutant alleles of Ion, htpR or combinations of both to prevent the degradation of expressed heterologous proteins is described. Another means of retarding proteolysis is to produce within the host cell an antagonist or inhibitor of proteases or the means for their specific formation or activation. An example of this type of proteolysis retarding is disclosed in European Patent No. 72925 filed March 2, 1983 by L. D. Simon and R. B. Fay. Described is the use of the cloned pin gene from bacteriophage T4 to inhibit protease activity in E. coli and thus stabilize certain expressed heterologous proteins such as human fibroblast interferon. Both of these general methods are preferred proteolysis-retarding means.

In the most preferred embodiments of this invention, protelysis-retarding means is provided by introducing a mutation in a gene which regulates an enzyme(s) involved in proteolysis. One of the major bacterial proteases is protease La, the product of the Ion gene. This gene is part of the heat-shock system and therefore under control of the htpR (rpoH) positive regulatory gene. Decreases in proteolysis can therefore be achieved by introducing mutations into either the Ion or htpR loci. Mutations in the Ion gene are available, for example the lonR9 and lonM00 mutations . See S.A. Goff et al. Proc. Natl. Acad. Sci. U.S. 81, pp. 6647- 6651 (1981). Mutations in the Ion gene can be conveniently introduced into the _E_. coli host by standard molecular genetic techniques such as bacteriophage Pi transduction. Although a number of htpR mutants have been isolated [A.D. Grossman et al. J. Bacteriol 161, pp. 939-943 (1985)], only htpRl65 has proven itself to be a workable mutation (Baker et al. , supra) . The htpRI65 allele contains an amber nonsense codon. In order for a strain containing the htpRI65 allele to be viable at

30°C (not at 41 °C) , it is generally to be maintained in conjunction with the temperature-sensitive amber suppressor supC (Ts) .

The mutations have a profound effect on proteolysis in _E_. coli. Under certain conditions, the htpRI65 mutation leads to a larger decrease in proteolysis than a Ion mutation. Therefore, if only one mutation is used, htpRl65 would be the preferred mutation. Other advantages to using an _E_. coli strain only containing an htpR mutation have been claimed (Goldberg et al. , supra) including no overproduction of polysaccharide, no defective cell division, and no abnormal sensitivity to UV light. Thus, under fermentation conditions, htpRI 65 mutants are more viable strains than strains containing Ion mutations. In other cases, however, it seems that a Ion htpRl65 double mutant has the lowest rate of protein degradation (Goldberg et al., supra) . It may therefore be desirable to construct a double mutant for expression studies of certain unstable proteins. The htpRI65 mutation can be transferred among bacterial strains using standard genetic techniques, provided the recipient strain already harbors a supC(Ts) mutation. One way of transferring the mutation is by P1 transduction simultaneously with a mutation in malT, since htpR and alT are about 5? linked.

The fourth essential feature of this invention is a means for preventing the spontaneous deletion of the heterologous polypeptide-producing genes. A number of factors can affect the stability of heterologous peptide-producing genes including the gene sequence itself, the plasmid replicon in which it resides, and various host factors. Host factors affecting the stability of heterologous genes can include but are not limi-ted to gene products involved in replication, repair, and recombination (both general homologous and site-specific recombination) of DNA.

Gene products such as RecB, RecC, RecD and particularly RecA usually play a prominent role in general homologous recombination of most types of DNA substrates. In the preferred embodiments of this invention, a recA mutation is used to inhibit the spontaneous deletion of highly repetitive heterologous genes via general homologous recombination. A recA allele is desired which is not only stable but also inactivates the recombination function of the RecA protein completely rather than partially. Alleles of the recA gene which fulfill these criteria are commonly available and include deletions of the recA gene as well as transposon insertions. [See D.K. Willis et al. supra) . Other commonly used recA alleles are recA1 and recAI3 which are present in commercially avail-able strains. Transfer of these mutations among strains is facilitated by a selectable marker. This marker is generally either an antibiotic resistance or a metabolic marker. Transposons are often used such as Tn10 which harbors resistance to the antibiotic tetracycline. Other transposons such as Tn5 and Tn9 which harbor resistance to kanamycin and chloramphenicol, respectively, are also available as is the Mud phage which harbors resistance to ampicillin. Alternatively, the gene for sorbitol utilization, srl, is located closely to recA on the _E_. coli genetic map and mutations in this locus are also available. The markers can be located directly in the recA gene such as a transposon insertion or can be located in a nearby gene and used in conjunction with a nonselectable mutation in recA.

These mutations can be transferred among strains by a variety of standard genetic techniques. One of the most common is P1 transduction. One of two available P1 mutants is often used, PI_{v r} or PIcml, clrlOO. Both have several technical advantages over wild-type P1 phage. The use of P1_vir prevents the establishment of fortuitous P1 lysogens in the recipient strain. P1cml, crl100 can be easily stored as a lysogen by selection for chloramphenicol resistance in the host strain. High-titer lysates can be made by heat induction and transduction can be carried out at high temperature (no lysogens are formed) or low temperature (accidental lysogens which form can be identified by their chloramphenicol resistance). Other genetic techniques are also available for transfer of mutations among strains. These include interrupted matings between Hfr or F' and F~ cells and the use of specialized transducing phages. It is important to note that the desired recA mutation is usually introduced as the last step in strain construction. Since these marker transfer techniques generally depend on homologous recombination to integrate the appropriate mutations into the _E_. coli chromosome, stable integration can be very dif icult in a recA-deficient strain. A subculture of a most preferred embodiment of this invention, E_. coli strain IG110, has been deposited in the permanent collection of the Northern Regional Research Laboratories, Agricultural Research Services, U.S. Department of Agriculture, Peoria, 111. U.S.A., under the accession number NRRL B-18352. The permanency of the deposit of this culture, and ready accessibility thereto by the public are afforded throughout the effective life of the patent in the event the patent is granted. Access to the culture is available during pendency of the application under 37 C.F.R. 1.14 and 35 U.S.C. 112. All restrictions on the availability to the public of the deposited culture will be irrevocably removed upon granting of a patent.

The novel strain of E_. coli provided by this invention can be used in the process of this invention to make or create bacteria which produce many useful polypeptide products. Illustrative of such products are analogs to naturally-occurring proteins such as collagen, elastin, keratin, protein or glycoprotein elements of thick, intermediate or thin filaments in higher organisms, silk fibroin, tropomyosin, troponin C, resilin, egg shell proteins, insect cuticle proteins or other architectural proteins containing a repetitive or quasi-repetitive primary structure.

In this process, the novel strain of _E_. coli is cultivated under suitable conditions for a time sufficient to produce the desired amount of poly- peptide. Cultivation of free-living or immobilized _E_. coli can be carried out in both liquid and solid nutrient media at a temperature of 22° to 30°C. It is to be understood also that for the preparation of limited amounts of the microorganism surface culture and bottles can be employed. The organism is grown in a nutrient medium containing a carbon source, for example, an assi ilatable carbohydrate, and a nitrogen source, for example, an assimilatable nitrogen compound or proteinaceous material. Preferred carbon sources include glucose, brown sugar, sucrose, glycerol, starch, cornstarch, lactose, dextrin, molasses, and the like. Preferred nitrogen sources include corn steep liquor, yeast, autolyzed brewer's yeast with milk solids, soybean meal, cottonseed meal, cornmeal, milk solids, pancreatic digest of casein, distillers' solids, animal peptone liquors, fishmeal, meat and bone scraps, inorganic salts such as NH^Cl or NH4SO4 and the like. Combinations of these carbon and nitrogen sources can be used advantageously. Trace metals, for example, zinc, magnesium, manganese, cobalt, iron, and the like, usually need not be added to the fermentation media since tap water and unpurified ingredients containing such trace metals are used as media components.

After sufficient polypeptide has been produced, the product can be isolated from the growth media using conventional purification techniques well known to those skilled in the art. The techniques can vary widely for application to a particular protein but for non-secreted proteins usually employ a means for concentrating and subsequently disrupting the microbial cells, followed by one or a combination of the following: precipitation, phase partitioning, gel filtration, ion-exchange chromatography, affinity or immunoaffinity chromatography,FPLC, HPLC, gel electrophoresis, and others too numerous to mention. The reader is directed to the following sources for reviews on this subject: R. Scopes, Protein Purification: Principles and Practice, Springer-Verlag New York, Inc. (1982); F.A.O. Marston, Biochem. J. 240, pp. 1-12 (1986); M. Ratafia and I. Keenan. American Biotechnology Laboratory 4, pp. 40-47 (1986):

The following is meant to be generally illustrative of approaches employed for protein purification from fermentation cultures. A process for the purification of recombinant human interferon- A is disclosed by K. Kitano and F. Shigeru in U.S. Patent No. 4,656,131 (1987) wherein the _E_. coli culture is collected by centrifugation, disrupted by sonication subsequent to dissolving the _E_. coli pellet in a buffer containing protease inhibitors and lysozyme, and centrifuged to give a pellet and supernatant phase. The interferon- αA activity present in the supernatant phase is purified by application of the diluted supernatant phase to an anti- interferon- αA immunoaffinity column. In this case, the desired human interferon- αA recombinant protein product remains in the supernatant of the cell lysate following centrifugation. However, many recombinant protein products partition with the pellet following centrifugation of the cell lysate. This can be attributed to the condensation or aggregation of the recombinant protein into refractile bodies or inclusions (See Kane and Hartley, supra for review of factors involved in the formation of inclusion bodies). Illustrative of this case is purification of recombinant human interferon- β from cultures of _E_. coll disclosed by M.W. Konrad and L.S. Lin in U.S. Patent No. 4,450,103 (1984). In this protocol, E. coli is collected by cross-flow filtration, disrupted by passage through a Manton-Gaulin homogenizer, and then centrifuged to give pellet and supernatant fractions. The inclusion bodies are solubilized in SDS and the human interferon- β is phase-partioned by extraction with 2-butanol. Human interferon- β is then precipitated from the butanol solution by addition of aqueous buffer at acid pH. The precipitate is dissolved in aqueous SDS solution and purified by chromatography on a molecular sieve column. For some applications, it may be preferable to use the protein without extensive purification. For example, see P.J. Rockwell, Single cell proteins from cellulose and hydrocarbons, Noyes Data Corporation (1976).

The bacterial strain of this invention may be used to prepare various useful polypeptides. For example, the polypeptides may be synthetic analogs to naturally- occurring fibrous proteins such as collagen, elastin, insect salivary gland, silk protein, silk fibroin, troponin C, tropomysin, and the like which can be used in the manufacture of fibrous products, synthetic skin and additives to cosmetics. Similarly, the bacterial strain of this invention can be used to prepare synthetic analogs of naturally-occurring, adhesives such as insect salivary gland adhesive protein, bioadhesive proteins from marine crustaceans such as Mytilis edulis, M. californianus, and Geukensia demissa, trematode egg shell dopa-proteins, and the like which can be used in the manufacture of adhesives. Likewise, the bacterial strain of this invention can be used in the preparation of synthetic analogs to naturally-occurring architectural proteins such as egg shell proteins, insect cuticle proteins, and the like which can be used in the fabrication of parts.

The following examples are presented to more particularly illustrate the invention and are not to be construed as limitations thereon.

EXAMPLE l^' Preparation of Plasmid pACI Having a Synthetic Gene for a Collagen Analog Without DNA linkers.

The following complementary and overlapping oligodeoxynucleotides were prepared using solid phase phosphoramidite chemistry as disclosed in S. L. Beaucage and M. H. Caruthers, Tetrahedron Letters 22; pp. 1859- 1862 (1981) on an Applied Biosystems model 380 DNA synthesizer: A. 5' - CG GGT CCG CCG GGT CCG C - 3'

B.^' 3' - GGC CCA GGC GGC CCA GGC - 5'

Each oligodeoxynucleotide was isolated from shorter chain-elongation failure products by electrophoresis on and elution from 20? polyacrylamide gels containing 8 M urea. The final product was greater than 95? pure as determined by densitometry of autoradiograms prepared from end-labeled oligodeoxynucleotide products separated by analytical gel electrophoresis. Phosphate -was added to the 5' ends of oligodeoxynucleotides A and B in separate reactions that contained 8.6 nmol oligodeoxy¬ nucleotide and 20 units T4 polynucleotide kinase in 35- 45 ul buffer (66 mM Tris-HCl, pH 7.6, 1 mM spermidine, 10 mM MgCl₂, 15 mM dithiothreitol, 200 ug/ml bovine serum albumin (BSA), and 1 mM [ Υ - ³²P]ATP with a specific activity of 0.2 Ci/mmol). These reaction mixtures were incubated for 2 hours at 37°C, then they were combined and were incubated at 14°C overnight. During this time, oligodeoxynucleotides A and B were annealing, presumably to form 17 base pair hetero- duplexes with one base pair overhanging 3' ends or 10 base pair heteroduplexes with 8 base pair overhanging 5' ends. T4 DNA ligase (40 units) was added and incubation was continued at 14°C for three days to polymerize the annealed oligodeoxynucleotides into long repetitive heteroduplex DNA coding for multiple repeats of the tripeptide (Gly-Pro-Pro) . These synthetic genes were dialyzed against TE buffer (10 mM Tris-HCl, pH 7.5, 1 mM EDTA) to remove unincorporated oligodeoxynucleotides and buffer components. The ends of the synthetic genes were then blunt-ended by using three units of the Klenow fragment of E. coli DNA polymerase I in a reaction (50 ul total volume) containing the following: 600 uM each of dCTP, dGTP, dATP and TTP; 50 mM Tris-HCl, pH 7.8; 9mM MgCl₂; 10 mM 2-mercaptoethanol; and 50 ug/ml BSA. This reaction mixture was incubated at 14°C for 30 min., then Na^EDTA was added to 10 mM and 1 0 ul of TE buffer were also added. The synthetic genes were purified on a DE- 52 column, then ethanol precipitated. These synthetic genes were combined with the excluded fraction of another batch of synthetic genes prepared in substantially like manner that had previously been passed over a Sepharose 6B (Pharmacia) column. The combined synthetic genes were size fractionated on a Sepharose 4B (Pharmacia) column. The size distribution of synthetic genes was determined by electrophoresis on a 5? polyacrylamide gel.

The relative molecular weight distribution of fractions enriched for highly polymerized synthetic genes was compared on denaturing (i.e., containing 8 M urea) and non-denaturing 5? polyacrylamide gels. These gels showed the molecular weight distribution of single- stranded synthetic genes to be smaller than expected from the molecular weight distribution of heteroduplex synthetic genes, suggesting that nicks and/or gaps were present in the double-stranded heteroduplex DNA. The nicks and/or gaps in 1.2 ug of synthetic genes were nick-translated in vitro using one unit of E. coli DNA polymerase I in the presence of 167 uM of each of dCTP, dGTP, dATP and TTP (with other buffer components as described in the blunt-ending reaction above) at 10°C for 20 min. (15 ul total volume).

Synthetic genes (0.5 ug heteroduplex DNA) were ligated without further manipulation to Clal-digested and blunt-ended pJLδ plasmid DNA (2.0 ug) using five units of T4 DNA ligase in the buffer described above for the kinasing and ligation reactions (10 ul total volume). The reaction mixture was incubated overnight at 14°C, diluted to 200 ul in TE buffer, and used directly to transform E. coli strain MH01 according to the Hanahan (1983) procedure. Colonies containing plasmids carrying synthetic gene inserts were identified by colony hybridization using radiolabeled oligodeoxynucleotide A as a probe. Insert-bearing plasmids were isolated and subjected to physical analysis by restriction enzyme mapping to size inserts. Plasmids were selected with collagen analog gene inserts of 200 to 350 bases and the inserts and adjacent plasmid regions were sequenced by a combination of chemical and dideoxy methods. (A.M. Maxam and W. Gilbert, Method Enzymol. 65, pp. 499-560 1980) (R.J. Zagursky et al. , Gene Analytical Techniques 2, pp. 289- 294 (1985). On the basis of proper reading frame and correct coding information at both the 5' and 3' plasmid/insert junctions, one of these plasmids was designated pAC1 and used for analysis of poly(Gly-Pro- Pro) expression.

EXAMPLE 2 Construction of Expression Strain IG110 he parent strain for this construction is CAG456 (lacZ(Am) trp(Am) pho(Am) rpoHl65 supC (Ts) mal rpsL phe rel) and is described in T.A Baker, A.D. Grossman and C.A. Gross, Proc. Natl. Acad. Sci. U.S.A. _8_1_, pp. 6779- 6783 (1984). This strain grows slowly at 30°C and is unable to grow at 42°C. CAG456 was obtained by T. Patterson who introduced the defective lambda phage λ ΔBamrex: :Km^rcl857 Δ(cro-bioB). In the process, the strain became Bio^" and has been named TAP130. The λdef phage used in this step is a new construct and has not yet been described in the literature.

TAP130 was sent to us and a stable recA mutation was introduced by bacteriophage P1 transduction. Donor of the recA allele was _E_. coli strain DC1138 [r~ m⁺ pro (leu ( ΔsrlR-recA)306 : :Tn10 ( λ)] which was obtained from T. Patterson and is described in U. S. Patent application 001,292, filed January 7, 1987 by J.I. Williams et al. The recA mutation in this strain is 100? cotransducible with the Tet^r phenotype and is not known to revert spontaneously. In order to perform the P1 transduction, the procedure of J. H. Miller, Experiments in Molecular Genetics (Cold Spring Harbor, 1972) was used. The generalized transducing page ^p1_vι_r was chosen in this case because it is unable to form any fortuitous lysogens during the course of the experiment. A P1 lysate grown on DC1138 was prepared as follows: One drop of DC1138 overnight culture was transferred into 5 ml of LB broth containing 5 x 10~3 M CaCl₂. The cells were then incubated until they reached a density of 2 x 10° cells/ml. P1 phage were preadsorbed to the cells by adding 10' phage to 1 ml of this culture and incubating for 20 min. at 37°C. At the end of this time period, 3 ml R-top agar were added to the tube and the contents were plated onto an LB plate. The phage were harvested after overnight incubation at 37°C. The soft agar layer was diluted with 5 ml LB and transferred to a centrifuge tube. Five drops of chloroform were added and the suspension was vortexed to lyse the cells and release all phage particles. After centrifugation at 10,000 rpm for 10 min., the supernatant, which contains the P1 lysate, was saved.

This lysate was used to transduce the recA mutation into TAP130. Five ml of a fresh overnight culture of TAP130 were resuspended in an equal volume of MC buffer (0.1 M MgS0_|, 5 mM CaCl₂) . The cells were then aerated at 30°C for 20 min. A 0.1 ml aliquot of the cells was added to 0.1 ml of a 10^"1 and 10~² dilution of the P1 lysate. The phage were preadsorbed by incubating at 30°C for 30 min. Then 0.2 ml 1M sodium citrate were added to each tube to prevent any additional phage attachment. The contents were plated onto LB plates containing 12.5 ug/ml of tetracycline using 3 ml R-top agar and the plates were incubated at 30°C for 48 hour^'s. Several tetracycline-resistant colonies were obtained.

The presence of the recA mutation was confirmed by testing for sensitivity to UV light. Each potential bacterial transductant as well as the parent strain TAP130 was streaked across an LB plate and different sections of the streaks were exposed to UV light for 0, 5, or 10 seconds, respectively. The plate was then incubated at 30°C overnight. One strain which was highly UV sensitive relative to its parent (as demonstrated by growth only in the zero exposure section of the streak while TAP130 grew at all UV exposures) was chosen and designated IG110.

EXAMPLE 3 Peptide Expression from the

Synthetic Collagen Contained in Plasmid pAC1

Using IG110 as the Host Strain The in vivo expression of a collagen analog peptide encoded by pAC1 was demonstrated using a whole-cell labeling protocol. An overnight culture of IG110(pAC1) was grown at 25°C with vigorous aeration. The next morning, 1 ml of overnight culture was inoculated into 20 ml LB broth containing 50 ug/ml of ampicillin. The culture was grown to at 25°C. A 1 ml sample was then taken and washed twice in M63 salt solution. The pellet was resuspended in 1 ml M63 medium including 0.2? glucose, 1 ug/ml of vitamin B1 , and 100 ug/ml of all amino acids except proline. The culture was preincubated at 42° for 20 min. Next, 10 uCi of [¹ C]proline were added and the incubation was continued for 3 min. About 1 mg of unlabeled proline was then added, and the incubation was carried out for an additional 3 min. The chase period was terminated by pelleting the cells. The cell pellet was washed once in 1 ml M63 salts to remove any residual unincorporated [' C]proline. The final cell pellet was resuspended in 50 ul SDS loading buffer (80 mM Tris-HC1 , pH 6.8, 100 mM dithiothreitol, 2? SDS, 10? glycerol, 100 ug/ml of bromophenol blue) and was immediately heated in a boiling water bath for 5 min. A 20 ul aliquot was then electrophoresed on a 12.5? SDS-polyacrylamide gel. The gel was treated with En^Hance (New England Nuclear) and exposed to X-ray film overnight. The resulting autoradiogram showed a strong protein band with a molecular weight of 22 kDa. This experiment demonstrates that the collagen analog peptide is efficiently expressed in strain IG110 (see Figure 1). COMPARATIVE EXAMPLE 1

Stability of the Collagen Analog Peptide in E. coli strains IG109 (htpR*) and IG110 htpRI65 The procedure for labeling the cellular proteins with [' C]proline is essentially as described in Example 3. Six 1 ml samples of each strain were prepared and labeled with 5 uCi of [¹^C]proline for 3 min. Then 1 mg of unlabeled proline was added to each culture and the incubation was continued at 42°C. The chase periods were terminated after 5, 10, 20, 40, 60 or 120 min. and the samples were processed as in Example 3. Aliquots of 20 ul were electrophoresed on 12.5? SDS-polyacrylamide gels. Following fluorography, the gels were exposed to X-ray film overnight.

A band corresponding to the 22 kDa collagen analog peptide was observed in all cases. The relative amount of collagen peptide in each lane was quantitated by first scanning the autoradiogram with a densitometer and then weighing the corresponding peak cut out of the tracing. The weight of the peak obtained for 5 min. after the beginning of the chase period was arbitrarily designated to represent 100? protein remaining and all other peaks were analyzed in relationship to it (see Figure 2). The results showed that in strain IG109 (htpR⁺) 40? of the collagen analog peptide was degraded by 20 min. after addition of unlabeled proline. By 40 min., only 20? of the original amount of collagen peptide still remained. In the case of IG110, however, no decrease in the amount of collagen analog peptide occurred over the 120 min. of the experiment. The presence of an htpRI 65 mutation therefore greatly increases the stability of this foreign protein in _E_. coli.

Claims

WHAT IS CLAIMED IS:

1. A novel strain of bacterial host organism comprising: a. one or more heterologous genes which code for the ₅ production of polypeptides composed of repeating amino acid sequences; b. controlling means for initiating the activity of said polypeptide production genes; c. bacterial proteolysis retarding means for

,_Q retarding the proteolysis of polypeptides produced by said heterologous genes; and d. gene stabilization means for stabilizing the heterologous genes.

2. The strain of claim 1 wherein said genes are ₁₅ naturally-occurring genes.

3. The strain of claim 1 wherein said genes are synthetic genes or semi-synthetic genes.

4. The strain of claim 1 wherein said controlling means is selected from the group consistinq of means which are responsive to temperature change, changes in nutrients, presence or absence of antibiotics and changes in the levels of intracellular compounds involved in intermediary metabolism.

5. The strain of claim 1 wherein said bacterial ₂₅ proteolysis retarding means is selected from the group consisting of defective protease-encoding gene means for regulating the formation, activity or degradation of proteases, and means which is antagonistic or_. inhibitory to proteases or antagonisitic or inhibitory to the means ₃₀ for the formation or activation of proteases.

6. The strain of claim 1 wherein said bacterial proteolysis means is a defective protease-encoding gene(s) resulting from the introduction of a mutation in one or more genes which regulate one or more enzymes involved in

-₅ proteolysis.

7. The strain of claim 1 wherein said gene stabilization means is a recA mutation.

8. A novel strain of E_. coli according to claim 1 comprising: a. one or more heterologous genes which code for the production of polypeptides composed of repeating amino

_ acid sequence; b. controlling means for initiating the activation of said genes selected from the group consisting of the lambda p. promoter in conjunction with a CI857 mutation, the trc promoter, the trc promoter and the T7 promoter in conjunction with T7 RNA polymerase;

- c. bacterial proteolysis retarding means selected from the group consisting of mutated Ion gene, a mutated htpR gene or a combination thereof; and d. gene stabilization means for stabilizing said gene selected from the group consisting of recA.

9. The E. coli strain NRRL ΔB-18352.

10. A method for producing a polypeptide having^* repeating amino acid sequences which comprises culturing a bacterial strain comprising: a. one or more heterologous genes which code for the production of polypeptides composed of repeating amino acid sequences; b. controlling means for initiating the activity of said polypeptide production genes; c. bacterial proteolysis retarding means for retarding the proteolysis of polypeptides produced by said heterologous genes; and d. gene stabilization means for stabilising the heterologous genes.