EP0241546A4 - Method for producing heterologous proteins. - Google Patents

Abstract

Method for producing a heterologous protein in a bacterial host cell such that the protein is exported from the host cell into the culture medium. The method involves culturing in a bacterial culture medium a genetically engineered bacterial strain containing a fusion DNA sequence comprising a first nucleotide sequence encoding at least an N-terminal portion of a flagellin protein and a second nucleotide sequence encoding the heterologous protein. The first nucleotide sequence is linked via its 3' terminus to the 5' terminus of the second nucleotide sequence, and the fusion DNA sequence is itself linked to an expression control sequence. In certain embodiments the first and second nucleotide sequences are linked by means of a linking nucleotide sequence encoding a selectively cleavable polypeptide. In those embodiments the resulting exported fusion protein will contain a selectively cleavable site at which the fusion protein may be selectively cleaved by chemical or enzymatic methods to produce the heterologous protein encoded for by the second nucleotide sequence of the fusion DNA sequence. The heterologous protein may then be separately recovered from any polypeptide fragment of flagellin or other proteinaceous material.

Description

METHOD FOR PRODUCING H_=TEROLOGOUS PROTEINS

Field of the Invention

This invention relates to a novel method for producing a heterologous protein in a bacterial host cell such that the protein is exported from the host cell into the extracellular medium.

Throughout this application various publications are referenced. Full citations for these publications may be found at the end of the specification. The disclosure of these publications are hereby incorporated by reference in order to more fully describe the state of the art to which this invention pertains. Background of the invention

Advances in cellular and molecular biology have made it possible, in certain cases, to identify a gene encoding a desired protein, to isolate the gene, to insert the gene into a host cell and to express the inserted gene in the host cell to produce the desired protein. Bacteria, especially Escherichia coli and Bacillus subtilis. have been intensively studied as host cells. When bacteria are used as host cells for this heterologous gene expression, however, two problems frequently have been encountered. Most bacterial expression systems produce proteins intracellulariy. When high level expression is achieved, the protein is often found to be insoluble (Marston et al., 1984; Williams et al., 1982; Schone et al., 1985). Production of an active protein from this insoluble material requires solubilization and refolding protocols which are often prohibitively expensive. If the protein is produced in an active, soluble form within the cell, its isolation requires cell lysis which releases hundreds of other soluble intracellular proteins. This can present a formidable problem in purification of the desired product.

Both the problem of production of insoluble, inactive proteins and of difficulty of purification may be overcome by having the bacteria secrete the desired protein into the growth medium. One particularly well documented method of directing the secretion of proteins is the use of a secretory signal sequence (Randall and Hardy, 1984; Silhavy et al., 1983; Wic ner, 1979). When a signal peptide is fused to the amino-terminal end of a heterologous protein, it directs the heterologous protein to the secretory machinery at the cell membrane. The heterologous protein is then translocated across the membrane and a specific protease, sometimes referred to as "signal peptidase", removes the signal peptide and releases the heterologous protein; In E. coli. secretion results in the accumulation of the heterologous protein in the periplasmic space, while in a gram positive bacterium, such as B. subtilis. secretion results in the accumulation of the product in the culture medium. This method has been used to direct the secretion of heterologous proteins in bacteria (Fraser and Bruce, 1978; Palva et al., 1983; Talmadge et al., 1981 ). As a result of these and other studies, problems, both potential and realized, have been discovered in the use of this particular approach. Cleavage of the signal peptide by signal peptidase may not be efficient or even accurate. Consequently, the secreted population of heterologous protein may contain unprocessed or misprocessed subpopulations. In addition, the amount of heterologous protein secreted is usually very small, and because both E. coli and B. subtilis also secrete proteases, a significant amount of the heterologous protein can be degraded after it is secreted.

Because of this latter point, the secretion and accumulation of heterologous proteins in the culture medium bv B. subtilis is vitiated unless the host cell expression and secretion of proteolytic enzymes is minimized or eliminated. One method for minimizing the effect of protease degradation of secreted proteins is to utilize mutant strains deficient in protease production. Mutations have been isolated in both the alkaline and neutral protease structural genes by recombinant methods (Stahl and Ferrari, 1984; Yang et al., 1984; Kawamura and Doi, 1984). Other protease deficient mutations isolated, to date, are pleitropic and also block the formation of mature endospores (Michel and Millet, 1970). Many of these mutations cause the cells to lyse when the culture is in the stationary phase of growth, thus may not be desirable for use in _L subtilis for the expression and secretion of heterologous proteins. While the use of existing protease deficient mutants may reduce the problem of- product instability, it may be necessary to isolate mutations in other protease genes to obtain maximum product stability.

In addition to using mutants of B. subtilis. the onset of endospore development and the secretion of proteases can be reduced significantly simply by adding to the medium a substance, such as glucose, which blocks the onset of secondary metabolism (Hoch, 1976). In the presence of glucose, the secretion of many proteases and cell lysis are inhibited. Cell lysis is to be avoided since release of intracellular proteins, of which some could be proteases, could result in additional degradation of the product and make it more difficult and costly to purify.

We have now discovered a new method for microbial production and export of a desired protein which avoids some of the problems associated with secretion via a signal peptide and secretion during stationary phase of growth. The method of this invention results in the transport of protein out of a flagellated bacterium and does so during the logarithmic growth phase and in the presence of a repressive substance such as-glucose. Products thus secreted are likely to be spared the problem of degradation by some proteases. Combining this secretion method with protease deficient mutants may improve product stability even more. This method harnesses the export system normally used by the host cell in exporting the protein flagellin.

Before describing the subject invention in detail, it may be helpful to set forth briefly further background information concerning flagellin.

Flagellin, which is the monomeric protein component of the flagellar filament, is a major extracellular protein product in many bacteria. Specifically, it is the predominant extracellular protein in logarithmic and early stationary phase of growth when Bacillus is grown- in minimal salts and glucose. The mechanism by which flagellin is exported is unknown. It does not seem to be exported by using a signal sequence which is cleaved from the amino-terminus of the protein (Silhavy et al., 1983). The amino-terminus of purified flagellin from Caulobacter crescentus. for example, has a sequence which corresponds to the putative translation start of its cloned structural gene (Gill and Aggbian, 1982,1983). The amino-terminus of purified flagellin from Salmonella tvphimurium begins with alanine which corresponds to the second amino acid following the translation start ©f its cloned structural gene (Joys and Rankis, 1972; Zieg and Simon, 1980). It is therefore unlikely that a processed leader sequence mediates transport of flagellin in bacteria such as Bacillus. Salmonella or Caulobacter.

Flagellin and several other proteins seem to exit the cell through the central core of the flagellum (lino, 1977; Silverman and Simon, 1977). These proteins can be as large as about 60 Kd so the physical size of the organelle core does not seem to limit this system unduly. The mechanism of secretion and the structural necessities of proteins to be exported by this system are not known, but much information about this system and the related system in £_ £__!_ has been collected and reviewed by lino (1977) and Silverman and Simon (1977). One notable feature of the system is its efficiency. It suffices to note that a flagellated £. coϋ cell has some 60,000 flagellin molecules (Komeda, 1982), thus a culture containing 1 x 10⁹ cells per ml exports approximately 5 mg per liter of flagellin.

To date, a minimum of 40 genes have been identified in E. coli which are apparently involved in bacterial motility and 29 involved in the ^• synthesis of the flagellar organelle (lino, 1977; Silverman and Simon, 1977). A pathway for the assembly of a flagellum was proposed by Suzuki and Komeda (1981). The central dogma in flagellar assembly is that the structure is assembled from the ceil membrane outward and the new components are derived from proteins that are transported through the core of the organelle and are assembled on the tip of the growing organelle. The flagellin structural gene is one of the last flagellar genes to be transcribed and translated during the synthesis of the flagellar organelle. Thus, a strain deleted for the flagellin gene should have an intact basal body and hook structure but would lack the filament. A mutation of interest to this invention is the ci≥ mutation, which has a phenotype of constitutive flagellar synthesis when this strain is grown in the presence of glucose (Silverman and Simon, 1977). £. cQJl strains carrying this particular mutation also produce five-fold more flagellin than wild-type strains.

Grant and Simon, (1969), isolated temperature sensitive (ts) fia mutants of subtilis 168 by isolating mutants resistant to bacteriophage PBS1 at high but not at lower temperatures . To date, 3 alleles of the i___g locus (encoding the so-called "h-antigen" which is the flagellin protein) in E. subtilis have been described. Wild-type E_ subtilis 168 contains the hag-1 allele, E. subtilis W23 has hag-2. and hag-3 is a "straight" mutant of hag-1. Another mutation of interest to this invention is the jfm mutation, which has a phenotype of higher motility and increased flagellin production (Grant and Simon, 1969; Pooley and Karamata, 1984).

In reducing to practice the present invention we have isolated and determined the sequence of the E- subtilis hag gene; deleted, in certain embodiments, part or all of this gene from the genome of the host cell; identified essential .elements of the sequence involved in transport of the protein to the outside of the cell; inserted into the host cell a heterologous gene encoding a desired protein at some site within the genome of the bacterium or within a flagellin gene locus of the host cell genome or as an extrachromosomal plasmid and expressed and exported fusion proteins containing the desired protein fused to that portion of flagellin essential for export. Methods and materials for the execution of this strategy are disclosed in detail hereinafter.

UBSTITUTE SHEET Summary of the Invention

This invention concerns a method for producing a heterologous protein in a bacterial host cell such that the protein is exported from the host cell into the culture medium. The method involves culturing in a bacterial culture medium a genetically engineered bacterial strain containing a fusion DNA sequence comprising a first nucleotide sequence encoding at least an N-terminal portion of a flagellin protein and a second nucleotide sequence encoding the heterologous protein. The first nucleotide sequence is linked via its 3' terminus to the 5' terminus of the second nucleotide sequence, and the fusion DNA sequence is itself operatively linked to an expression control sequence. The two linked nucleotide sequences making up the fusion DNA sequence are linked to each other "in frame" such that the coding region of the entire fusion DNA sequence is translated to produce the encoded protein. In certain embodiments the first and second nucleotide sequences are linked by means of a linking nucleotide sequence encoding a selectively cleavable polypeptide. In those embodiments the resulting exported fusion protein may be selectively cleaved by chemical or enzymatic methods to produce the heterologous protein encoded for by the second nucleotide sequence of the fusion DNA sequence. The heterologous protein may then be separately recovered from any polypeptide fragment of flagellin or other proteiπaceous material. Brief Description of the Tables and Figures

Figure 1 depicts restriction maps of clones p4A and p8A and the extent of nucleotide sequencing of clone p4A.

Table i depicts the available nucleotide sequence data for clone p4A.

Table 2 depicts the nucleotide and amino acid sequence of the Δ5M proinsulin gene and corresponding protein.

Table 3 depicts the nucleotide sequence of the £. __ς_Ii flagellin gene.

8

Detailed Description of the Invention

The invention relates to a method for producing a heterologous protein in a bacterium of a flagellate species such that the heterologous protein is exported by the bacterium into the bacterial growth medium. The method involves culturing in a suitable bacterial growth medium a bacterial strain containing as part of its genetic material a "fusion" DNA sequence which includes a nucleotide sequence encoding at least a portion of the N-terminus of a flagellin protein linked to a heterologous gene, i.e., a gene encoding a protein other than flagellin. The fusion DNA sequence is operatively linked to an expression control sequence, preferably that of the flagellin gene of the host bacterium, and contains a translational terminating signal 3' to the heterologous gene component.

Suitable host cells may be selected from a wide range of flagellate bacterial species including for example Escherichia coli. Caulobacter crescentus and Bacillus subtilis. The host cell must contain a known or identifiable nucleotide sequence encoding a flagellin protein. It should be noted that bacteria in which flagellin-encoding DNA has not been identified heretofore may also be useful in the practice of this invention. In that case the appropriate nucleotide sequence may be identified and characterized by using conventional techniques to recover and appropriately purify a suitable amount of flagellin from the bacteria for protein sequencing, determine the amino acid sequence of a portion of the flagellin, prepare oiigonucleotide probes corresponding to the amino acid sequence so determined, screen a DNA library derived from the bacteria for the presence of a nucleotide sequence capable of hybridizing to the probe(s) and determine the nucleotide sequence of the DNA so identified and/or its location in the bacterial genome. For example, the flagellin gene of E- subtilis may be routinely obtained from the B. subtilis genome as a 2.5 Kb PstI fragment by purely conventional means using an oligonucleotide probe complementary to part or all of the sequence depicted in Table 1. Similarly, the

E. coli flagellin gene may be obtained from the E. coli

Genetic Stock Center, (Barbara Bachmann, Curator,

Department -of Human Genetics, Yale University, 333 Cedar

Street, New Haven, Conn.), on a Clark and Carbon library plasmid, pLC24-16. Part of all of the gene may be routinely identified by hybridization to an oligonucleotide complimentary to the sequence depicted in Table 3.

The wild-type host cell must contain at least one flagellum and preferably, as in the case B. subtilis or E. coli. a plurality of flagella. In one embodiment the host cell is an increased flagellin and motility (ifm) strain of B . subtilis. . Strains carrying ifm mutations produce and export significantly more flagellin than wild-type host cells and may be conveniently obtained by iteratively selecting from cultured colonies those cells which migrate furthest away from the spot of inoculation on a semisolid medium referred to as "motility agar". An ifm strain of B. subtilis, for example, has been so obtained which produces and exports about twenty times as much flagellin as does the wild-type B. subtilis. After appropriate insertion into the genome of the B. subtilis ifm strain of a fusion DNA sequence, as disclosed in detail hereinafter, the genetically engineered ifm strain produced and exported about twenty times as much heterologous protein as a similarly treated wild-type strain.

In the practice of this invention the DNA sequence encoding the N-terminal portion of flagellin, e.g. a portion of the hag gene of B. subtilis. is operatively linked to an. expression control sequence, including for example, a promoter, a ribosome binding site and a translation start codon. Preferably the expression control sequence used is the host cell's expression control

SUBSTITUTE SHEET 10A

Table 1

1 GATCTCCGCATTATCCTCACAAAAAAAGTGAGGATTTTTTTATTTTTGTATTAACAAAATCAGCAGACAAT 72 CCGATATTAATGATGTAGCCGGGAGGAGGCGCAAAAGACTCAGCCAGTTACAAAATAAGGGCACAAGGACG

^'fMet Arg lie Asn His Asn lie Ala Ala 143 TGCCTTAACAACATATTCAGGGAGGAACAAAACA ATG AGA ATT AAC CAC AAT ATT GCA GCG

Leu Asn T r Leu Asn Arg Leu Ser Ser Asn Asn Ser Ala Ser Gin Lys Asn Met 204 CTT AAC ACA CTG AAC CGT TTG TCT TCA AAC AAC AGT GCG AGC CAA AAG AAC ATG

Glu Lys Leu Ser Ser Gly Leu Arg lie Asn Arg Ala Gly Asp Asp Ala Ala Gly 258 GAG AAA CTT TCT TCA GGT CTT CGC ATC AAC CGT GCG GGA GAT GAC GCA GCA GGT

Leu Ala lie Ser Glu Lys Met Arg Gly Gin lie Arg Gly Leu Glu Met Ala Ser 312 CTT GCG ATC TCT GAA AAA ATG AGA GGA CAA ATC AGA GGT CTT GAA ATG GCT TCT

Lys Asn Ser Gin Asp Gly lie Ser Leu lie Gin T r Ala Glu Gly Ala Leu Thr 366 AAA AAC TCT CAA GAC GGA ATC TCT CTT ATC CAA ACA GCT GAG GGT GCA TTA ACT

Glu Thr His Ala lie Leu Gin Arg Val Arg Glu Leu Val Val Gin Ala Gly Asn 420 GAA ACT CAT GCG ATC CTT CAA CGT GTT CGT GAG CTA GTT GTT CAA GCT GGA AAC

Thr Gly Thr Gin Asp Lys Ala Thr Asp Leu Gin Ser lie Gin Asp Glu He Ser 474 ACT GGA ACT CAG GAC AAA GCA ACT GAT TTG CAA TCT ATT CAA GAT GAA ATT TCA

Ala Leu Thr Asp Glu He Asp Gly He Ser Asn Arg Thr Glu Phe Asn Gly Lys 528 GCT TTA ACA GAT GAA ATC GAT GGT ATT TCA AAT CGT ACA GAA TTC AAT GGT AAG

Lys Leu Leu. Asp Gly Thr Tyr Lys Val Asp Thr Ala Thr Pro Ala Asn Gin Lys 582 AAA TTG CTC.GAT.GGC ACT TAC AAA GTT GAC ACA GCT ACT CCT GCA AAT CAA AAG

Asn Leu Vai Phe Gin He Gly Ala Asn Ala Thr Gin Gin He Ser Val Asn He 636 AAC TTG GTA TTC CAA ATC GGA GCA AAT GCT ACA CAG CAA ATC TCT GTA AAT ATT

Glu Asp Met Gly Ala Asp Ala Leu Gly He Lys Glu Ala Asp Gly Ser He Ala 690 GAG GAT ATG GGT GCT GAC GCT CTT GGA ATT AAA GAA GCT GAT GGT TCA ATT GCA

Ala Leu His Ser Val Asn Asp Leu Asp Val Thr Lys Phe Ala Asp Asn Ala Ala 744 GCT CTT CAT TCA GTT AAT GAT CTT GAC GTA ACA AAA TTC GCA GAT AAT GCA GCA

Asp Thr Ala Asp He Gly Phe Asp Ala Gin Leu Lys Val Val Asp Glu Ala He 798 GAT ACT GCT GAT ATC GGT TTC GAT GCT CAA TTG AAA GTT GTT GAT GAA GCG ATC

Asn Gin Val Ser Ser Gin Arg Ala Lys Leu Gly Ala Val Gin Asn Arg Leu Glu 852 AAC CAA GTT TCT TCT CAA CGT GCT AAG CTT GGT GCG GTA CAA AAT CGT CTA GAG

His Thr He Asn Asn Leu Ser Ala Ser Gly Glu Asn Leu Thr Ala Ala Glu Ser 906 CAC ACA ATT AAC AAC TTA AGC GCT TCT GGT GAA AAC TTG ACA GCT GCT GAG TCT

Arg He Arg Asp Val Asp Met Ala Lys Glu Met Ser Glu Phe Thr Lys Asn Asn 960 CGT ATC CGT GAC GTT GAC ATG GCT AAA GAG ATG AGC GAA TTC ACA AAG AAC AAC

He Leu Ser Gin Ala Ser Gin Ala Met Leu Ala Gin Ala Asn Gin Gin Pro Gin 1014 ATT CTT TCT CAG GCT TCT CAA GCT ATG CTT GCT CAA GCA AAC CAA CAG CCG CAA

Asn Val Leu Gin Leu Leu Arg Oc

1068 AAC GTA CTT CAA TTA TTA CGT TAA TTTTAAAAAAGACCTTGGCGTTGCCAGGGTCTTTTAATT

1131 TAAATTTCTATCTCCTAATCATTCCTCATCCTGTCACTAACTCATGATATAATAACCGGATTCTCCACTAA

1202 CTTTTTATAAATGTATTTCCATACAAGAAATCTAAAACAGAAGATTTTTTTCCAAAAATATGTGTAATCTT

1273 ATCTCGACTTAGTCGATATAAACGATAGATTGGGGCATAGGGGATGATCAATTGAACATTGAAAGGCTCAC

1344 TACGTTACAACCTGTTTGGGATCGTTATGATACTCAAATACATAATCAGAAAGATAATGATAACGAGGTTC

1415 CTGTTCATCAAGTTTCATATACCAATCTTGCTGAAATGGTGGGGGAAATGAACAGCTT 10B Table 3 flagellin coding sequence is underlined

1 cccgactccc agcgatgaaa tacttgccat gcgatttcct tttatctttc

51 gacacgtaaa acgaataccg gggttatcgg tctgaattgc gcaaagttta

101 cgtttaattg ttttttttaa tagcgggaat aaggggcaga gaaaagagta

151 tttcggcgac taacaaaaaa tggctgtttt tgaaaaaaat tctaaaggtt

201 gttttacgac agacgataac agggttgacg gcgattgagc cgacgggtgg 251 aaacccaata cgtaatcaac gacttgcaat ataggataac gaatcatggc

301 acaagtcatt aataccaaca gcctctcgct gatcactcaa aataatatca 351 acaagaacca gtctgcgctg tcgagttcta tcgagcgtct gtcttctggc 401 ttgcgtatta acagcgcgaa ggatgacgca gcgggtcagg cgattgctaa 451 ccgtttcacc tctaacatta aaggcctgac tcaggcggcc cgtaacgcca 501 acgacggtat ctccgttgcg cagaccacca ccgaaggcgc gctgtccgaa 551 atcaacaaca acttacagcg tgtgcgtgaa ctgacggtac aggccactac 601 cggtactaac tctgagtctg atctgtcttc tatccaggac gaaattaaat 651 cccgtctgga tgaaattgac cgcgtatctg gtcagaccca gttcaacggc

701 gtgaacgtgc tggcaaaaaa tggctccatg aaaatccagg ttggcgcaaa 751 tgataaccag actatcacta tcgatctgaa gcagattgat gctaaaactc

801 ttggccttga tggttttagc gttaaaaata acgatacagt taccactagt 851 gctccagtaa ctgcttttgg tgctaccacc acaaacaata ttaaacttac 901 tggaattacc ctttctacgg aagcagccac tgatactggc ggaactaacc 951 cagcttcaat tgagggtgtt tatactgata atggtaatga ttactatgcg

1001 aaaatcaccg gtggtgataa cgatgggaag tattacgcag taacagttgc

1051 taatgatggt acagtgaσaa tggcgactgg agcaacggca aatgcaactg

1101 taactgatgc aaatactact aaagctacaa ctatcacttc aggcggtaca

1151 cctgttcaga ttgataatac tgcaggttcc gcaactgcca accttggtgc

SUBSTITUTE SHEET IOC

Table 3 (cont' d)

1201 tgttagctta gtaaaactgc aggattccaa gggtaatgat accgatacat

1251 atgcgcttaa agatacaaat ggcaatcttt acgctgcgga tgtgaatgaa

1301 actactggtg ctgtttctgt taaaactatt acctatactg actcttccgg

1351 tgccgccagt tctccaaccg cggtcaaact gggcggagat gatggcaaaa

1401 cagaagtggt cgatattgat ggtaaaacat acgattctgc cgatttaaat 1451 ggcggtaatc tgcaaacagg tttgactgct ggtggtgagg ctctgactgc 1501 tgttgcaaat ggtaaaacca cggatccgct gaaagcgctg gacgatgcta 1551 tcgcatctgt agacaaattc cgttcttccc tcggtgcggt gcaaaaccgt 1601 ctggattccg cggttaccaa cctgaacaac accactacca acctgtctga 1651 agcgcagtcc cgtattcagg acgccgacta tgcgaccgaa gtgtccaata 1701 tgtcgaaagc gcagatcatc cagcaggccg gtaactccgt gttggcaaaa 1751 gctaaccagg taccgcagca ggttctgtct ctgctgcagg gttaatcgtt 1801 gtaacctgat taactgagac tgacggcaac gcaaattgcc tg-atgcgctg 1851 cgcttatcag gcctacaagt tgaattgcaa tttattgaat ttgcacattt 1901 ttgtaggccg gataaggcgt ttacgcgcat ccggcaacat aaagcgcaat 1951 ttgtcagcaa cgtgcttccc gccaccggcg gggttttttt ctgcctggaa 2001 tttacctgta acccccaaat aacccctcat ttcacccact aatcgtccga 2051 ttaaaaaccc tgcagaaacg gataatcatg ccgataactg ctataacgca 2101 gggctgttt

UB 11 sequence for flagellin. Thus in the ifm embodiment the preferred expression control sequence is the expression control sequence of the^' hag gene.

Depending on the amount and nature of flagellin DNA which is fused to the heterologous gene, the heterologous gene, the heterologous protein which is produced and exported will usually be a fusion protein comprising at least a portion of the flagellin protein linked to the protein encoded for by the heterologous gene. In certain embodiments of the invention the fusion DNA sequence contains a full-length flagellin-encoding nucleotide sequence linked via its 3'terminus to the 5' terminus of the heterologous gene. In other embodiments the flagellin- encoding sequence is truncated at its 3 ' terminus. Thus, in one embodiment the fusion DNA sequence contains nucleotides 1-633 of the flagellin-encoding gene linked via nucleotide 633 to the 5' terminus of the heterologous sequence. In another embodiment a shorter portion of the flagellin gene is used which contains nucleotides 1-432. Other embodiments may contain deletions of various lengths within the 432-912 nucleotide region of'the flagellin gene. Sequences containing further deletion of nucleotides 5' to nucleotide - 432 are also expected to be useful in the practice of this^" invention although the exact length of the remaining flagellin sequence which permits or optimizes export of the fusion protein has not yet been precisely determined. Indeed, in specific cases the desired flagellin-encoding sequence may be only about 75, 50, 25 or 10 codons in length. Even shortier flagellin-encoding sequences may be useful in this invention, and it is possible that the 5¹ untranslated region alone of the flagellin gene, with no flagellin-encoding nucleotide sequence, will permit export of the heterologous protein in certain cases. By "heterologous" as the term is used herein is meant a protein or DNA sequence other than a flagellin protein or a DNA sequence encoding a flagellin

SUBSTITUTE SHEET 12 protein, respectively.

In one embodiment the fusion DNA sequence contains an additional nucleotide sequence which links the flagellin gene portion and the heterologous gene. Preferably the linking sequence encodes a polypeptide which is selectably cleavable or digestable by conventional chemical or enzymatic methods. The fusion protein of this embodiment will thus contain an engineered cleavage site at which it may be selectably cleaved. Cleavage of the fusion protein yields the "mature" protein which is encoded by the heterologous gene. The mature protein may in turn be obtained in purified form, free from any polypeptide fragment of flagellin to which it was previously linked.

Preferably, the engineered host cells produce and export the heterologous. protein during a growth phase when protease secretion is at a minimum. Such is the case with B_. subtilis , in which production and export of the heterologous protein occurs during the logarithmic/early stationary growth phase. It is also preferred that the engineered host cells produce and export the heterologous protein in the presence of a substance which tends to further reduce the level of exported proteases e.g. glucose, in -the case of B. subtilis.

As this invention is not limited to any specific type of heterologous DNA a wide variety of heterologous proteins may be produced by this method including, for example, proteins useful for human or veterinary therapy or diagnostic applications, such as hormones, cytoxins, growth or inhibitory factors, etc., fu ctional enzymes, and modified natural or wholly synthetic proteins.

Furthermore, it should be understood that a variety of recombinant genetic constructions will be useful in achieving the primary objective of this invention, namely the utilization of the bacterial machinery normally used in the bacterial production and export of flagellin to effect the production and export of a heterologous protein from a 13 flagellate bacterium. Indeed, several illustrative recombinant approaches are presented hereinafter.

Accordingly, it should also be understood that this invention is not limited to any one particular recombinant method for achieving its objectives.

One approach for producing a genetically engineered bacterium of this invention involves deleting a portion or all of the flagellin gene from the chromosome of the host bacterium and inserting into the flagellin deletion locus or into another chromosomal locus, a plasmid-borned heterologous gene via a single recombination event. The replacement of the host flagellin gene with a deleted version constructed in vitro is performed by established methods (Stahl and Ferarri, 1984, Yang et al . , 1984;

Kawamura and Doi, 1984). The use of an "integrable plasmid" or an "integration vector" in B. subtilis is well documented (Ferrari et al . , 1983) . This particular integration vector is comprised of a selectable antibiotic resistance gene and a plasmid origin that allows extrachromosal replication in 32. coli . but not in B. subtilis. In addition, this vector must include a sequence which is homologous to a sequence within the host genome; this may be- a portion of the flagellin gene that has not been deleted from the host genome, or the sequence could be a portion or all of another host gene. The plasmid also includes a heterologous gene fused to a. portion of the flagellin gene to allow expression and export of a heterologous protein. When an integration vector such as described above is transformed in B}. subtilis, transformed cells carrying the plasmid-borne antibiotic resistance gene are selected. This plasmid cannot replicate extrachromosomally, therefore the plasmid integrates into the genome via a single recombination event between the homologous sequences on the plasmid and the chromosome.

The resulting chromosomal structure contains the plasmid flanked by directly duplicated copies of the homologous

SUBSTITUTE SHEET 14 sequence. As long as antibiotic selection is maintained, the plasmid-derived sequences are replicated- and stably inherited as part of the bacterial genome. In some cases, perhaps depending on which antibiotic resistance gene is placed on this plasmid, the integrated plasmid can be

"amplified", or the number of integrated plasmid copies can be increased, by growth of the strain carrying the integrated plasmid in higher levels of the antibiotic used to select for the initial integration (Gutterson and

Koshland, 1983) . This results in amplification of the number of heterologous gene copies which may result in increased expression and export of heterologous protein.

Further increases in expression and export of heterologous protein may be accomplished by transforming, with or without amplification, the plasmid into a host strain carrying the ifm mutation.

A second approach involves stably inserting a plasmid into a flagellin deletion strain, preferably one that contains the ifm mutation, wherein the plasmid contains a fusion DNA sequence as previously described and in addition, a functional origin that allows extrachromosomal replication in B. subtilis. The plasmid must also contain a selectable gene, such as an antibiotic resistance gene, which can be used to select for the inheritance of the plasmid by transformation and to insure maintenance of the plasmid during culture growth. To maximize the expression and export of heterologous protein, it may be useful to adjust heterologous gene dosage, or copy number, by placing the gene into the different plasiώds. For example, the plasmid pUBllO, which is a Staphylococcus aureus plasmid that is often used in B. subtilis molecular biological applications, is a potentially useful high copy number plasmid (Gryczan, et al., 1978). This particular plasmid has a copy number of approximately 40 per cell. Another plasmid, pE194, may be useful as a low copy plasmid in B. subtilis (Gryczan and Dubnau, 1978) . When this plasmid is 15 transformed into B. subtilis it maintains a copy number of approximately 5-10 per cell.

A third approach for producing a genetically engineered bacterium of this invention is to integrate a plasmid, which is comprised of a heterologous gene fused to the 3 ' end of a portion of the flagellin gene that lacks the transcription and translation control sequence and in addition may lack a portion of the gene encoding the N- terminal region of the gene, into a B. subtilis host containing an intact flagellin gene and preferably the ifm mutation. This integrable plasmid also contains a selectable antibiotic resistance gene and a plasmid origin that allows extrachromosomal replication in E. coli, but not in B. subtilis. When transformed into B. subtilis, selection is for the inheritance of the antibiotic resistance gene and integration into the chromosome is mediated by . a single recombination event between the flagellin sequence on the plasmid and the corresponding homologous sequence within the flagellin gene in the chromosome. As a result of integration, the heterologous gene is fused to the transcription and translation regulatory sequences and all or part of the encoding sequences of the host flagellin gene. The fusion junction between flagellin and the heterologous gene must be a codon that is 3 ' of those flagellin sequences required for export. If so, the integration of this plasmid generates one copy of a completely functional gene that codes for the expression and export of a heterologous protein. The integration also generated two truncated and nonfunctional genes, a flagellin gene that lacks transcription and translation control sequences and may or may not contain sequences encoding for a portion of the N-terminus, and a flagellin-heterologous gene fusion that lacks the same sequences. With this particular integration scheme the latter truncated gene may be amplified by amplifying the plasmid sequences.^' Thus transformation of this plasmid

SUBSTITUTE SHEET 16 into B. subtilis interrupts the host flagellin gene and at the same time introduces the desired gene fusion between flagellin and the heterologous gene at a copy number of one per chromosome.

Numerous aspects and advantages of the invention will be apparent to those skilled in the art upon consideration of the preceding in view of the illustrative experimental examples, results, and discussion which follow.

1 6 A

EXPERIMENTAL EXAMPLES

MATERIALS AND METHODS Bacterial strains and plasmids. Escherichia col MM294 (F, supE44. endAI . thi-1. hsdR4) was used as a host for plasmid constructions and for screening the pUC18 based Bacillus subtilis 168 genomic library. E. coli was transformed by the procedure of Dagert and Ehrlich (1979), with selection on_.L agar plates containing 15 μg/ml neomycin, 15 μg/ml chloramphenicol, or 50 μg/ml ampicillin. B. subtilis strains were transformed by the procedure of Anagnostopoulos and Spizizen (1961), with selection on L agar plates containing 5 μg/ml neomycin or 5 μg/ml chloramphenicol. Auxotrophic markers were selected on minimal glucose plates supplemented with the appropriate amino acids at 50 μg/ml (Spizizen, 1958). B. subtilis G1B1 was constructed by transforming E_ subtilis 168 trpC2 with E_ subtilis W23 DNA and selecting for Trp⁺ transformants. An ifm mutation was selected in this strain by repeated selection for hypermotility on motility agar by the method of Grant and Simon (1969).

The plasmids pBR322, pJH101 , pUC18, pUC19, and pUB110 have all been described previously (Bolivar et al., 1977; Yanisch-Perron et al., 1985; Ferrari et al., 1983; Gryczan et al., 1978). The plasmid pALlΔ5M contains the human proinsulin gene that has been specifically mutagenized to encode a proinsulin that can be processed in vitro to insulin by enzymatic and chemical means (U. S. Serial No. 646,573 and International Application No. PCT/US 85/01673; see figure 3).

Reagents and media. Restriction enzymes, T4 polynucleotide kinase, Bal-31 exonuclease, and the Klenow fragment of £. ______ DNA polymerase I were purchased from commercial sources and used according 1 7

to the suppliers' conditions. Motility deficient mutants were screened and tested on motility agar (Grant and Simon, 1969). For the expression and export of homologous and heterologous proteins, cultures were grown in expression medium, which contained minimal salts (Spizizen, 1958) supplemented with 2 % glucose, 0.1 % technical grade casamino acids (Difco), and the appropriate amino acids supplemented at 50 μg/ml. In some experiments, total protein was labeled with L-[³⁵S]-methionine (>400 Ci/mmol; New England Nuclear) by adding 10 μCi/ml to the above medium.

DNA and protein characterization. Plasmid DNA was prepared from f ______ transformants by the alkaline lysis method of Birnboim and Doly

(1979). B. subtilis chromosomal DNA was prepared by the method of Marmur (1961 ). The separation of restriction fragments on polyacrylamide and agarose gels and the electroelution of DNA fragments were performed as previously described (Lawn et al., 1981). All plasmid constructions were made with DNA fragments purified by electroelution from gels. Restriction fragments were ligated into appropriate sites of M13 phage vectors mp18 or mp19 (Vieira and Messing, 1982; Yanisch-Perron et al., 1985) in preparation for sequence determination by dideoxy methods (Sanger et al., 1977). DNA restriction fragments were prepared as probes by labeling [alpha-³ P] CTP by nick-translation (Rigby et al., 1971 ). Synthetic oligonucleotides were synthesized by the phosphotriester method (Crea and Horn, 1980), and end labeled with [gamma-³²P] ATP and T4 polynucleotide kinase (Richardson, 1971). Hybridization conditions for the labeled oligonucleotide pools were at 37 C in a solution of 1 X Denhardt solution, 0.1 mM ATP, 1 mM NaCI, 0.5 % Nonidet® P-40-, (a nonionic detergent; Sigma), 200 ng/ml soluble type XI bakers yeast RNA (Sigma), 90 18

mM Tris-OH pH 7.5, and 6 mM EDTA. Washing was at 37 C in 6X SSC (1 X SSC is 0.15 M NaCI plus 0.015 M sodium citrate). For southern hybridization analysis, digested DNA fragments were separated on 1 % agarose and depuriπated as described by Wahl et al. (1979) and transferred to nitrocellulose by the method of Southern (1975). Hybridizations and washings for southern blots with nick-translated probes were performed as described by Maniatis et al. (1978).

For the expression and export of homologous or heterologous proteins, isolated colonies were picked from streak plates or transformation plates and inoculated into expression medium with or without L- S-methionine. The culture was grown to mid-logarithmic

stage of growth (ODssonm ⁼ °-⁵) ^{ancl at tnιs time} P^heny^|methylsulfonyl fluoride (PMSF) and EDTA were added to the culture each at final concentrations of 1 mM. PMSF and EDTA are serine protease and metallo-protease inhibitors respectively and their addition increases the stability of heterologous proteins in the medium. One hour after the addition of the protease inhibitors one ml aliquots were removed; if the strain being examined contains the wild-type flagellin gene intact, the culture sample is heated at 80 C for 10 iπ to depolymerize the flagellar filament into flagellin monomers; if a flagellin-heterologous fusion protein is being expressed and exported, the heat treatment is not needed. The culture aliquot is then centrifuged for 3 minutes in an Eppendorf centrifuge, in 1.5 ml eppendorf tubes, and 900 μL of supernatant is removed and added to another tube containing 100 μL of 100 % trichloro acetic acid (TCA). The TCA precipitations were allowed for 20 min. on ice, then are centrifuged for 5 min. and the pellet washed three 1 9

times with one ml aliquots of cold acetone. The cell pellet was washed in one ml of wash buffer (l OO. mM tris pH8, 150 mM NaCI, 1 mM EDTA) and resuspended in 50 μL of TE buffer (10 mM tris pH 8, 1 mM EDTA). The cells were then lysed by sonic disruption. The proteins from the cell pellet and supernatant fractions were then separated on SDS-polyacrylamide gels according to Laemmli (1970) and transferred to nitrocellulose electrophoretically for western blot analysis by the method of Burnette (1981 ).

E_ subtilis 168 flagellin was purified by the method of Martinez (1963). Once isolated, the material was separated from minor contaminants on a preparative SDS-polyacrylamide gel and the band containing flagellin was cut out, lyophilized and used as an antigen in rabbits for the production of flagellin specific antibodies. This protocol resulted in the production of highly specific antibodies for the detection of flagellin and flagellin-heterologous fusion proteins by western blot analysis.

SUBSTITUTE SHEET 20

RESULTS

Characterization of the ifm mutation, E_ subtilis GIBI and _ subtilis GIB1 ifm were grown in expression medium plus L-³^S-methionine to mid-logarithmic phase of growth. Samples from the culture were processed as described in the methods section to compare the levels of flagellin produced in the two strains. There was approximately 10-fold more flagellin exported in the strain carrying the ifm mutation. The western blot with aπtiflagellin antibody confirmed that this protein is flagellin.

Cloning of the B. subtilis hag gene. The 17-mer oligonucleotide probe pool for the cloning, by hybridization, of the hag gene of B. subtilis GIB1 was designed and based on the published amino acid sequence of flagellin (Delange et al., 1976). Two pools of 12 17-mer oligonucleotides completely covered the degeneracy of amino acids 170-174 and, in addition, the first two bases of the glycine codon at amino acid 175 of the sequence (Asn-lle-Glu-Asp-Met-Gly). The sequences of the oligonucleotides in pool number 1 are δ'-A-A-T/C-A-T-T/C/A-G-A-A/ G-G-A-T-A-T-G-G-G-3^* and pool number 2 are 5'-A-A-T/C- A-T-T/C/A-G- A- A/G-G-A-C-A-T-G-G-G-3^* .

A genomic library was prepared in pUC18 using DNA from E_ - subtilis GIB1. The vector was digested with Eall and the first two bases complementary to the 5' overlapped ends were filled in using the Klenow fragment of DNA polymerase I and dTTP and dCTP. The bacterial DNA was partially digested with £__il3A and sized on a preparative agarose gel. DNA fragments ranging in size from 2-5 Kb were cut put and electroeluted from the gel and then treated with the Klenow fragment and dGTP and dATP to 21

fill in the first two bases complementary to the overlapped ends. The insert and vector DNAs were then ligated with T4 DNA ligase. This strategy allowed only one insert per vector and prevented tandem ligations of two or more insert DNA fragments or religation of vector DNA fragments (Hung and Wensik, 1984). E. coli MM294 was transformed with the above ligated DNA and the screening of bacterial colonies for plasmids with inserts containing the flagellin gene was by transfer to nitrocellulose according to Grunstein and Hogness (1975). Crude restriction maps of two clones identified as hybridization positives, p4A and p8A, are shown in figure 1. The complete sequence of an open reading frame contained in both p4A and p8A was found to encode a protein that is 304 amino acids; all but two amino acids are homologous to the published protein sequence of E. subtilis 168 flagellin (Delange et al., 1976). The exception was a pair of amino acids, glycine-101 and threonine-102, which are inverted in the published sequence. The extent of clone p4A that is sequenced is shown in figure 1 and the sequence itself is shown in

Table 1 .

Construction of E. coli - B. subtilis shuttle vectors. The EL <___ii - subtilis shuttle vector, pBE3, contains the pUC18 polylinker (147 bp EcoRI - Pvull restriction fragment), the pBR322 origin of replication (1166 bp P___U.II - Ahalll restriction fragment), and the neomycin nucleotidyl traπsferase gene and origin of replication from pUB110 (3,529 bp Pvull - EcoRI restriction fragment). The order of these fragments in a clockwise direction on a circular map is EcoRI--polylinker--Pvull/Pvull--pBR322 origin-- Ahalll/Pvull— pUB110 origin-neomycin gene--E__ς_RI. This plasmid replicates autonomously and confers neomycin resistance in both J £2Ji and E_ subtilis. 22

The integration vector, plEVI , is a derivative of pJH101 that replicates autonomously in E. coli. but when transformed into B. subtilis. must integrate into the chromosomal flagellin locus. The plasmid contains the chloramphenicol acetyl transferase (CAT) gene and origin of replication from pJH101 (3,224 bp £stl - Aval restriction fragment), part of the pUC18 polylinker (200 bp Pvull - Xbal restriction fragment) and a 400 bp Hindi 11 - £2tl restriction fragment from the E. subtilis chromosome just 5' of the Jbag, promoter region (see Fig. 1). The 5' overlaps of the Aval. Xbal. and Hindlll ends were filled in by the Klenow fragment of DNA pol I with all four dNTPs before ligation. The order of these restriction fragments in a clockwise direction on a circular map is ____IJ--origin-CAT gene--Aval/Pvull-polylinker--Xbal/Hindlll--40Q bp chromosome fragmeπt-Pstl.

Construction of plEV1fla304PlΔC. The plasmid plEV1fla304PlΔC is a derivative of plasmids, pBE3, pALIΔ5M, p4A, and plEVI which contains the pBR322 origin of replication, the CAT gene which confers functional resistance to chloramphenicol in both £, coli and E. subtilis. and a sequence which encodes amino acids 144 - 304 of flagellin (see Table i) , four junction amino acids (Gly-Met-Gln-Ala), and the Δ5M proinsulin gene (see Table 2) . The latter encoding sequence does not contain regulatory sequences for the initiation of transcription and translation. When transformed into B. subtilis GIB1 ifm. it integrates via a single recombination event between the homologous plasmid-bome and chromosomal flagellin sequences apd results in the reconstitution of a functional gene which encodes a fusion protein containing 1 - 304 amino acids of flagellin, the 4 junction amino acids, and the Δ5M proinsulin sequence. This gene includes the host transcription and translation start

SUBSTITUTE SHEET 22A

Table 2

Met Phe Val Asn Gin His Leu Cys Gly Ser His Leu Val Glu Ala Leu Tyr Leu ATG TTT GTG AAC CAA CAC CTG TGC GGC TCA CAC CTG GTG GAA GCT CTC TAC CTA

Val Cys Gly Glu Arg Gly Phe Phe Tyr Thr Pro Lys Thr Asp Leu Gin Val Gly GTG TGC GGG GAA CGA GGC TTC TTC TAC ACA CCC AAG ACC GAT CTG CAG GTG GGG

Gin Val Glu Leu Gly Gly Gly Pro Gly Ala Gly Ser Leu Gin Pro Leu Ala Leu CAG GTG GAG CTG GGC GGG GGC CCT GGT GCA GGC AGC CTG CAG CCC TTG GCC CTG

Glu Gly Ser Leu Gin Lys Arg Met Gly He Val Glu Gin Cys Cys Thr Ser He GAG GGG TCC CTG CAG AAG CGT ATG GGC ATT GTG GAA CAA TGC TGT ACC AGC ATC

Cys Ser Leu Tyr Gin Leu Gin Asn Tyr Cys Asn Am TGC TCC CTC TAC CAG CTG GAG AAC TAC TGC AAC TAG

SUBSTIT 23

sequences of the flagellin gene. This plasmid was constructed as follows. The 4750 bp Hindlll - Pvull restriction fragment from pBE3, (the first three bases of the Hindlll 5' overlap were filled in by the Klenow fragment with dATP, dGTP, and dCTP), was ligated to the 470 bp f&hl - Ndel restriction fragment from pALIΔ5M, (the 3' overlap of the Sphl site was chewed back by the Klenow fragment and the first base of the Ndel 5' overlap was filled in by the Klenow fragment with dTTP), to construct pFPH . The 5200 bp BamHl - E≤ll restriction fragment from pFPH , (the 3' overlap of Rsil was removed using the Klenow fragment), was ligated to the 2632 bp BamHl - Ahalll restriction fragment from p4A to construct pFPIfla304. The Ahalll end of the fragment from p4A was treated with "slow" bal-31 exonuclease before ligatidn, and the proper pFPIfla304 construction was screened by colony hybridization with an oligonucleotide (5'-T-T-A-T-T-A-C-G-T-G-G-C-A-T-G-C-A-A-3^*) that spans the correct ligation juntion. The sequences of the hybridization positives were determined to confirm the proper construction. The 1621 bp BamHl - Boil restriction fragment from pFPIfla304, (the Bail 5' overlap was filled in with the Klenow fragment and all four dNTPs), was ligated to the 3827 bp BamHl - EcoRI restriction fragment from plEVI ( the EcoRI 5' overlap was filled in with the Klenow fragment and all four dNTPs) to construct the plasmid plEV1fla304PI. The plasmid plEV1fla304PlΔC was constructed by digesting plEV1fla304PI with CJal, purifying the 4500 bp fragment and religatiπg the same fragment.

Expression and export of flagellin-proinsulin fusion protein in B. subtilis GIB1 ifm. The plasmid plEV1fla304PIΔC was transformed into E. subtilis GIB1 ifm and an isolated colony was used to inoculate 10 ml of expression medium plus L-[ ^S]-methionine in a 250 ml baffled 24

erlenmeyer flask. The culture was incubated at 37 C on a gyratory shaker operating at 250 revolutions per minute. At the mid-logarithmic stage of growth (OD₅5_0nm = 0.5), protease inhibitors were added and one hour later

samples were removed and processed as described in the Methods section. After examination of the ⁵S-methionine total labeling and western blot autoradiograms, the flagellin-proinsulin fusion protein was identified as a band that bound antiflagellin antibody and migrated at the expected molecular weight when compared to the migration of flagellin. The appearance of this band in the supernatant fraction of the culture aliquot confirms that a significant amount of flagellin-proinsulin fusion protein was exported into the medium.

25

DISCUSSION

Flagellin in B. subtilis G1B1 ifm is exported at levels up to 10 - 20 % of the total cell protein during logarithmic stage of growth, in the presence of glucose, where the secretion of extracellular proteases is minimized. In this invention the flagellin export pathway has been utilized to export heterologous fusion proteins into the culture medium. In a specific demonstration of the potential for this system a recombinant flagellin - proinsulin fusion protein was exported via the flagellin export pathway. This same experimental approach was successfully used to export another flagellin - heterologous fusion protein, namely flagellin - TEM β-lactamase fusions. This particular β-lactamase is from the plasmid pUC18 (Yanisch-Perron et al., 1985), and confers ampicillin resistance to various gram negative bacteria including £. coli. Flagellin - β-lactamase gene fusions were expressed in Bacillus which resulted in the accumulation of flagellin - β-lactamase fusion protein in the culture medium. This fusion protein has β-lactamase activity and also cross reacts with antiflagellin and anti β-lactamase antibodies. In addition, strains carrying the flagellin - β-lactamase gene fusions were resistant to ampicillin. These results indicate that the flagellin export system may be useful for the production of many homologous fend heterologous proteins.

The flagellin - proinsulin fusion protein contains a methionine residue at the junction between the flagellin amino acid residues and the proinsulin residues thus the latter could be cleaved from flagellin with cyanogen bromide. Active and properly folded insulin may thus be obtained by combined treatment of the fusion protein with cyanogen bromide and a

SUBSTITUTE SHEET 26

specific protease from Psedomonas fragii. Accordingly, the strategy for the export of a variety of homologous or heterologous proteins via the flagellin pathway is to fuse the coding sequence for that protein "X" to a portion or all of the flagellin coding sequence, and at the junction, introduce a specific cleavage site so that the desired sequence may be removed by chemical or enzymatic means. In addition to cyanogen bromide, which cleaves on the carboxy side of methionine residues, formic acid may be used to cleave between aspartic acid and proline residues (Nilsson et al., 1985). There are highly specific proteases which also may be useful for site specific cleavages. Two examples are porcine enteropeptidase, which cleaves on the carboxy side of the sequence (Asp)₄-Lys ( Maroux et al., 1971), and factor X_a, which cleaves on the

carboxy side of the sequence lle-Glu-Gly-Arg (Nagai arid Thøgersen, 1984). A nucleotide sequence that encodes for either of the specific recognition sites for these or other specific proteases may be placed, by conventional recombinant methods, at the junction of flagellin - protein "X" encoding sequences. The use of specific proteases to cleave fusion proteins exported via the flagellin pathway would result in the release of protein "X" without an f-Met or Met residue at the N-terminus.

The fact that export via the flagellin pathway may require a portion or all of the flagellin coding sequence may be advantageous with respect to purification of flagellin - protein "X' fusion proteins. Flagellin can be purified easily and is highly antigenic, consequently fusion proteins may be purified by affinity chromatography with flagellin antibody, then processed by the appropriate chemical or enzymatic means.

Many homologous or heterologous proteins exported as flagellin 27

fusion proteins would require specific processing to a mature, active form by specific chemical or enzymatic means as described above. Examples of these types of proteins include insulin, colony stimulating factors, human growth hormone, or other pharmaceuticals destined for human use. Other proteins, for example, enzymes such as proteases, amyiases or proteins such as animal growth hormones, may be active and suitable for use as flagellin fusion proteins. In cases such as these the specific chemical or enzymatic processing step required for removal of the flagellin encoding sequences would be unnecessary.

The export of homologous or heterologous proteins via the flagellin export pathway may be further improved by modifications in host cell development, vectors, and promoter vector combinations. At least two general catagories of host cell mutations may further increase the final yield of flagellin - protein "X" fusion protein obtained in this process. To increase the stability of exported proteins in the culture medium, host mutations that decrease protease activity may be used. Most protease activity can be minimized simply by growing the culture in the presence of excess glucose, but further improvements may be obtained by isolating mutations in regulatory genes, such as spoQ mutations, which are pleitropic and result in decreased expression of some proteases (Michel and Millet, 1970; Hoch, 1976). Recombinant methods may be used to isolate in vitro - derived mutations in other protease structural genes as has been accomplished with the alkaline and neutral protease genes (Stahl and Ferrari, 1984; Yang et al., 1984; Kawamura and Doi, 1984).

Mutations within the coding sequence for flagellin itself may increase the efficiency by which some flagellin - protein "X" fusion proteins are exported. Presumably these mutations would be in sequences 28

that encode for that portion of flagellin that is important for directing the transport of the fusion protein. ι

Should the co-presence in the same cell of the gene encoding the desired fusion protein and the host flagellin gene result in competition between the fusion protein and flagellin for the same export site machinery, the host flagellin gene may be inactivated to provide for more efficient export of flagellin fusion proteins. In the example described in the methods and results section, this was accomplished by integrating the expression vector, plEV1fla304PIΔC,.into the host flagellin gene. The integration event generated an active flagellin - proinsulin gene fusion, and simultaneously, inactivated the resident flagellin gene. The inactivation of the host flagellin gene can also be accomplished by replacing the gene with an in vitro - derived deletion mutation (Stahl and Ferrari, 1984; Yang et al., 1984; Kawamura ^'and Doi, 1984). This would increase the flexibility of using alternate vector - promoter combinations which may ultimately increase the yield of the desired product. The following are examples where the use of a host strain, from which all or part of the flagellin gene has been deleted, may be useful for increasing the product yield. The regulatory sequences for the initiation of transcription and translation of flagellin - gene "X" gene fusions in these examples may be those from the flagellin gene or.may be from another gene where transcription and translation is constitutive; or these sequences may be from a gene that is regulated and thus could be controlled. The latter type of regulatory sequence may be used where it is desired to prevent gene expression until the culture density is high, at which point transcription and translation may then be initiated to yield product accumulation in the culture medium. Expression of genes encoding 29

heterologous or homologous proteins controlled by any one of the above regulatory sequences may be on low-copy vectors such as integrable plasmids '(Ferrari et al., 1983) or plasmids such as pE194 that replicate extrachromosomally (Gryczan and Dubnau, 1978) or high-copy vectors such as pUB110 (Gryczan et al., 1978) and pBE3 which replicate extrachromosomally. An integration vector may be inserted into any gene in the chromosome. A particularly attractive insertion site is a gene that - is dispensible for normal growth, such as the neutral protease structural gene (Yang et al., 1984). This gene may be cloned and a portion of the coding sequence could be used as the homologous sequence on an integrable plasmid that is required for integration by recombination.

Genes or portions thereof for other proteins of the flagellum, e. g. the hook or basal body proteins, may be used in place of the flagellin gene to achieve production and export of the heterologous protein. In such cases the protein may be recovered, purified and sequenced, in whole or part, and the gene encoding the protein identified by hybridization with oligonucleotide probes, for example. Identification and use of such genes in accordance with this invention may be accomplished in analagous fashion to the methods disclosed herein for flagellin-related embodiments.

£. cjQϋ is certainly an attractive host for use in the flagellin export system. The flagellin gene can be easily cloned as described previously in this document and flagellin - heterologous gene fusions can be expressed as a part of low or high copy plasmid vectors or as sequences integrated into the chromosome. A mutation, cj≥, has been isolated which when introduced into a strain results in a five-fold overproduction of flagellin and renders the strain constitutively motile (Silverman and

SUBSTITUTE SHEET 30

Simon, 1977). Five-fold more flagellin - heterologous fusion protein may be produced if the appropriate vector containing the gene fusion is introduced into this mutant strain.

30A

LITERATURE CITED

Anagnostopoulos, C. and J. Spizizen. 1961. Requirements for transformation in Bacillus subtilis. J. Bacteriol. 81.741-746.

Birnboim, H. C, and J, Doly. 1979. A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res. 7:1513-1523.

Bolivar, F., R. L. Rodriguez, P. J. Greene, M. C. Betiachy, H. L. Heynecker, H. W. Boyer, J. H. Crosa, and S. Falkow. 1977. Construction and characterization of new cloning vehicles II. A multipurpose cloning system. Gene 2:95-113.

Burπette,.W. N. 1981. "Western blotting": Electrophoretic transfer of proteins from sodium dodecyl sulfate-polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibody and radioiodiπated protein A. Anal. Biochem. 112:195-203.

Crea, R., and T. Horn. 1980. Synthesis of oligonucleotides on cellulose by a phosphotriester method. Nucleic Acids Res. 8:2331-2348.

Dagert, M., and S. D. Ehriich. 1979. Prolonged incubation in calcium chloride I improves the competence of Escherichia coli cells. Gene 6:23-28.

Delange, R. J., J. Y. Chang, J. H. Shaper, and A. M. Glazer. 1976. Amino acid sequence of flagellin of Bacillus subtilis 168. J. Biol. Chem.

254:705-711. Ferrari, F. A., A. Nguyen, D. Lang, and J. A. Hoch. 1983. Construction and properties of an integrable plasmid for Bacillus subtilis. J.

Bacteriol. 154:1513-1515.

SUBSTITUTE SHEET 31

Fraser, T. H., and B. J. Bruce. 1978. Chicken ovalbumin is synthesized and secreted by Escherichia coli. Proc. Nati. Acad. Sci. U.S.A.

75:5836-5940. Gill, P. R., and N. Agabian. 1982. A comparitive structural analysis of the flagellin monomers of Caulobacter crescentus indicated that these proteins are encoded by two genes. J. Bacteriol. 150:925-933. Gill, P. R., and N. Agabian. 1983. The nucleotide sequence of the Mr = 28,500 flagellin gene of Caulobacter crescentus. J. Biol. Chem.

258:7395-7401. Grant, G. F., and M. I. Simon. 1969. Synthesis of bacterial flagella II. PBS1 transduction of flagella-specific markers in B. subtilis. J.

Bacteriol. 99:116-124. Grunstein, M., and D. Hogness. 1975. Colony hybridization: a method for the isolation of cloned DNAs that contain a specific gene. Proc. Natl.

Acad. Sci. U.S.A. 72:3961-3965. Gryczan, T. J., S. Contente, and D. Dubnau. 1978. Characterization of

Staphylococcus aureus plasmids introduced by transformation into

Bacillus subtilis. J. Bacteriol. 134:318-329. Gutterson, N. I., and D. E. Koshland, Jr. 1983. Replacement and amplification of bacterial genes with sequences altered in vitro. Proc. Natl.

Acad. Sci. U.S.A. 80:4894-4898. Hoch, J. A. 1976. Genetics of bacterial sporulation. Adv. Genet. 18:69-98. Hung, M-C, and P. C. Wensick. 1984. Different restriction enzyme-generated sticky DNA ends can be joined in vitro. Nucleic

Acids Res. 12:1863-1874. lino, T. 1977. Genetics of structure and functions of bacterial flagella. 32

Ann. Rev. Genet. 11 :161-182. Joys, T. M., and V. Roukis. 1972. The primary structure of the phase-1 flagellar protein of Salmonella tvphimurium. I. the tryptic peptides. J. Biol. Chem. 247:5180-5193. Kawamura, F., and R. H. Doi. 1984. Construction of a Bacillus subtilis double mutant deficient in extracellular alkaline and neutral proteases. J. Bacteriol. 160:442-444. Komeda, Y. 1982. Fusion of flagellar operons to lactose genes on a Mud lac bacteriophage. J. Bacteriol. 150:16-26. Laemmii, U. K. 1970. Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature 227:680-685. Lawn, R. M., J. Adelman, S. C. Bock, A. E. Franke, C. M. Houck, R. C. Najarian,

P. H. Seeburg, and K. L. Wion. 1981. The sequence of human serum albumin cDNA and its expression in E. coli. Nucleic Acids Res.

9:6103-6114. Maniatis, T., R. C.Ηardison, E. Lucy, J. Laner, C. O'Connell, D. Quan, G. K. Sim, and A. Efstratiadis. 1978. The isolation of structural genes from libraries of eucaryotic DNA. Cell 15:687-701.

Marmur, J. 1961. A procedure for the isolation of deoxyribonucleic acid i from microorganisms. J. Mol. Biol. 3:208-218.

Maroux, S., J. Baratti, and P. Desnuelle. 1971. Purification and specificity of porcine enterokinase. J. Biol. Chem. 246:5031 -5039.

Marston, F. A. O., P. A. Lowe, M. T. Doel, J. M. Schoemaker, S. White, and S. Aπgal. 1984. Purification of calf prochymosin (prorenniπ) synthesized in Escherichia coli. Biotechnology 2:800-804.

Martinez, R. J. 1963. A method for the purification of bacterial flagella by

SUBSTI 33

ion exchange chromatography. J. Gen. Microbiol. 33:115-120.

34

Michel, J. F., and J. Millet. 1970. Physiological studies on early-blocked sporulation mutants of Bacillus subtilis. J. Appl. Bacteriol.

33:220-227. Millet, J. 1970. Characterization of proteinases excreted by Bacillus subtilis Marburg strain during sporulation. J. Appl. Bacteriol.

33:207-219. Nagai, K., and H. C. Thøgersen. 1984. Generation of β-globin by sequence-specific proteolysis of a hybrid protein produced in

Escherichia coli. Nature 309:810-812. - Nilsson, B., E. Holmgren, S. Josephson, S. Gatenbeck, L. Philipson, and M.

Uhlen. 1985. Efficient secretion and purification of human insulin-like growth factor I with a gene fusion vector in staphylococci. Nucleic Acids Res. 13:1151 -1162. Palva, I., P. Lehtovaara, L. Kaariainen, M. Sibakov, K. Cantell, C. H. Schein, K.

Kashiwagi, and C. Weissman. 1983. Secretion of interferon by

Bacillus subtilis. Gene 22:229-235. Perlmaπ, D., and H. O. Halvorson. 1983. A putative signal peptidase recognition site and sequence in eucaryotic and procaryotic signal peptides. J. Mo Biol. 167:391-409.

I

Pooley, H. M., and D. Karamata. 1984. Genetic analysis of autolysin-deficient and flagellaless mutants of Bacillus subtilis.

J. Bacteriol. 160:1123-1129. Randall, L. L. and S. J. S. Hardy. 1984. Export of protein in bacteria.

Microbiol. Rev. 48:290-298. Richardson, C. C. 1971. Polynucleotide kinase from Escherichia coli infected with bacteriophage T4. Proc. Nucleic Acid Res. 2:815-828. 35

Rigby, P. W. J., M. Dieckmann, C. Rhodes, and P. Berg. 1977. Labeling deoxyribonucleic acid to high specific activity in vitro by nick-translation with DNA polymerase I. J. Mol. Biol.

113:237-251. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. U.S.A.

74:5463-5467. Scherer, S., and R. W. Davis. 1979. Replacement of chromosome segments with altered DNA sequences constructed in vitro. Proc. Natl. Acad.

Sci. U.S.A. 76:4951-4955. Schoner, R. G., L. F. Ellis, and B. E. Schoner. 1985. Isolation and purification of protein granules from Escherichia coli cells overproducing bovine growth hormone. Biotechnology 3:151 -154. Shortle, D., J. E. Haber, and D. Botstein. 1982. Lethal disruption of the yeast - actin gene by integrative DNA transformation. Science.

217:371 -373. Silhavy, T. J., S. A. Benson, and S. D. Emr. 1983. Mechanisms of protein localization. Microbiol. Rev. 47:313-344. Silverman, M., and M. Simon. 1977. Bacterial flagella. Ann. Rev. Microbiol.

31 :397-419. Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol.

98:503-517. Spizizen, J, 1958. Transformation of biochemically deficient strains of

Bacillus subtilis bv deoxyribonucleate. Proc. Natl. Acad. Sci. U.S.A.

44:1072-1078. 36

Stahl, M. L., and E. Ferrari. 1984. Replacement of the Bacillus subtilis subtilisin structural gene with an in vitro-derived deletion mutation. J. Bacteriol. 158:411-418. Suzuki, T.. and Y. Komeda. 1981. Incomplete flagellar structures in

Escherichia coli mutants. J. Bacteriol. 145:1036-1041. Talmadge, K., J. Brosius, and W. Gilbert. 1981. An 'internal' signal sequence directs secretion and processing of proinsulin in bacteria. Nature.

294:176-178. Vieira, J., and J. Messing. 1982. The pUC plasmids, and M13mp7-derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene. 19:259-268. Wahl, G. M., M. Stem, and G. R. Stark. 1979. Efficient transfer of large DNA fragments from agarose gels to diazobeπzyloxymethyl-paper and rapid hybridization by using dextran sulfate. Proc. Natl. Acad. Sci.

U.S.A. 76:3683-3687. Watson, M. F. F. T984. Compilation of published signal sequences. Nucleic

Acids Res. 12:5145-5164. Wickner, W. 1979. The assembly of proteins into biological membranes: the membrane trigger hypothesis. Annu. Rev. Biochem. 48:23-45. i

Williams, D. C, R. M. VanFrank, W. L Muth, and J. P. Burnett. 1982. Cytoplasmic inclusion bodies in Escherichia coli producing biosynthetic human insulin proteins. Science 215:687-688.

Yang, M. Y., E. Ferrari, and D. J. Henner. 1984. Cloning of the neutral protease gene of Bacillus subtilis and the use of the cloned gene to create an in-vitro-derived deletion mutation. J. Bacteriol. 160:15-21. 37

Yanisch-Perron, C, J. Vieira, and J. Messing. 1985. Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33:103-119.

Zeig, J., and M. Simon. 1980. Analysis of the nucleotide sequence of an invertable controlling element. Proc. Natl. Acad. Sci. U.S.A. 77:4196-4200.

Claims

37A

What is claimed is:

1. A method for producing a heterologous protein in a bacterial host cell such that the protein is exported from the host cell into the culture medium, the method comprising culturing in a bacterial culture medium a genetically engineered bacterial strain containing a fusion DNA sequence which comprises a first nucleotide sequence encoding at least an N-terminal portion of a flagellin protein and a second nucleotide sequence encoding the heterologous protein, said first nucleotide sequence being linked via its 3' terminus to the 5' terminus of the second DNA sequence and said fusion DNA sequence being operatively linked to an expression control sequence.

2. A method according to claim 1 which further comprises recovering the exported protein from the culture medium.

3. A method according to claim 1 , wherein the first and second nucleotide sequences of the fusion DNA sequence are linked by a linking nucleotide sequence which encodes a selectably cleavable polypeptide such that the exported protein contains a selectably cleavable site.

4. A method according to claim 3 which further comprises cleaving the expqrted protein at the selectably cleavable site to produce the heterologous protein encoded for by the second nucleotide sequence of the fusion DNA sequence.

5. A method according to claim 4 which further comprises recovering the heterologous protein from any polypeptide fragment of flagellin or other proteinaceous material.

6. A method according to claim 1 , wherein the engineered bacterial cells are cultured in the presence of a substance which represses the

SUBSTITUTE SHEET 38

production or export of proteases.

7. A method according to claim 1 , wherein the fusion DNA sequence is integrated into the chromosome of the host cell.

8. A method according to claim 1 , wherein the fusion DNA sequence is contained in an extrachromosomal plasmid within the host cell.

9. A method according to claim 1 , wherein the engineered host cell lacks a functional gene for native flagellin production.

10. A method according to claim 2, wherein the exported protein is recovered by immunoaffinity chromatography using anti-flagellin antibody.

11. A protein produced by the method of claim 1.

12. A protein produced by the method of claim 3.