CA2237581A1 - Salmonella secreted proteins and uses thereof - Google Patents

Abstract

Substantially pure Salmonella secreted proteins (Ssp), the secretion of which is dependent upon the expression of PrgH; methods of diagnosing Salmonella infection; and live attenuated vaccine strains in which Ssp secretion is decreased.

Description

CA 02237~81 1998-0~-13 SATM~NELLA SECRETED PROTEINS AND USES ~HEREOF
Statement as to FederallY Sponsored Research This invention was made with Government support 5 under AI34504 and AI30479 awarded by the National Institutes of Health. The Government has certain rights in the invention.
Backqround of the Invention The invention relates to virulence factors of 10 Salmonella typhi 7n77rium ~
Salmonella ty~h; rium (S. tyrhir ~rium) enter epithelial cells by a process termed bacterial~ ted endocytosis ~ s ~ tyrh i ~77rium stimulates these normally nonphagocytic cells to undergo significant cytoskeletal 15 rearrangements that are visualized as localized membrane ruffling ad}acent to the bacteria. Bacteria are then internalized via membrane-bound vacuoles formed from the membrane ruffles.
Several S. tyrhi rium loci have been identified 20 that are re~uired for the induction of bacterial-mediated endocytosis (BME) by epithelial cells. Many of these epithelial-cell signaling loci have a similar chromosomal location, clustered within a 40 kb ~Ivirulence island"
located between 59 and 60 minutes on the S. typhim~7rium 25 chromosome (Mills et al., Mol. Microbiol. 15:749-759, 1995). InvJ is a S. tymphimurium gene which is thought to encode a secreted protein necessary for BME (Collazo et al., Mol. ~icrobiol. 15:25-38, 1995).

s~ ry of the Invention The invention features proteins involved in Salmonella ty~hi rium virulence and/or bacterial-mediated endocytosis. The genes encoding these proteins have now been cloned and their corresponding gene products characterized. Accordingly, the invention CA 02237~81 1998-0~-13 wo97/l822s PCT~S96/18504 features a substantially pure DNA encoding a Salmonella secreted protein (Ssp). By the term 'ISalmonella secreted protein" is meant a Salmonella-derived protein, the secretion of which is dependent on the expression of 5 PrgH. In preferred embodiments the invention features substantially pure DNA encoding a Salmonella ty~hi~7~ium secreted protein. BY SA 7 ~ne7 7 a ty~h i ~ium secreted protein is meant as SA7 ~n~7 7~ typhim~rl~m derived protein, the secretion of which is dependent on the lO expression of PrgH.
One aspect of the invention features a substantially pure DNA molecule which includes the SspB
gene; preferably, the DNA includes the DNA sequence of SEQ ID NO: 1, or degenerate variants thereof encoding the 15 amino acid sequence of SEQ ID NO: 5. In another aspect the invention features a substantially pure DNA molecule which includes the SspC gene; preferably, the DNA
includes the DNA sequence of SEQ ID NO: 2, or degenerate variants thereof encoding the amino acid sequence of SEQ
20 ID NO: 6. In another aspect the invention features a substantially pure DNA molecule which includes the SspD
gene; preferably, the DNA includes the DNA sequence of SEQ ID NO: 3, or degenerate variants thereof encoding the amino acid sequence of SEQ ID NO: 7. In another aspect 25 the invention features a substantially pure DNA molecule which included the SspA gene; preferably, the DNA
includes the DNA sequence of SEQ ID NO: 4, or degenerate variants thereof encoding the amino acid sequence of SEQ
ID NO: 8. The invention also features a substantially 30 pure DNA molecule which includes the SspB, SspC, SspD, and SspA genes; preferably, the DNA includes the DNA
se~uence of SEQ ID NO: 15. The invention also features a substantially pure DNA molecule which includes the SspH
gene; preferably, the DNA includes the DNA sequence of 35 SEQ ID NO: 13, or degenerate variants thereof encoding CA 02237~8l l998-0~-l3 wo97ll822s PCT~S96/18504 the amino acid sequence of SEQ ID NO: 14. The invention also features a substantially pure DNA molecule which includes the Salmonella tyrosine phosphatase A (stpA) gene; preferably, the DNA includes the DNA sequence of 5 SEQ ID NO: lO, or degenerate variants thereof encoding the amino acid sequence of SEQ ID NO:12.
The invention also features a cell into which has been introduced substantially pure DNA encoding an Ssp (or a mutant variant thereof). The substantially pure lO DNA can be introduced as a portion of a plasmid or other autonomously replicating molecule. In addition the substantially pure DNA can be introduced by homologous recombination. Preferably, the bacterial cell is a Salmonella cel-; more preferably the bacterial cell is a 15 Salmonella ty~; ~ium cell. Cells into which have been introduced substantially pure DNA encoding an Ssp (or mutant variant thereof) can be used as a source of purified Ssp.
The invention includes a substantially pure Sspc 20 polypeptide, e.g., a polypeptide which includes an amino acid sequence substantially identical to the amino acid sequence of SEQ ID NO: 6 or an active fragment thereof and a substantially pure SspD polypeptide, e.g., a polypeptide which includes an amino acid se~uence 25 substantially identical to the amino acid sequence of SEQ
ID NO: 7 or an active fragment thereof. The invention includes a substantially pure SspB polypeptide, e.g., a polypeptide which includes an amino acid sequence substantially identical to the amino acid sequence of SEQ
30 ID NO: 5 (incomplete protein sequence) or an active fragment thereof and a substantially pure SspA
polypeptide, e.g., a polypeptide which includes an amino acid sequence substantially identical to the amino acid sequence of SEQ ID NO: 8 (incomplete protein sequence) or 35 an active fragment thereof. The invention includes a CA 02237~81 1998-0~-13 WO97/1822s PCT~S96/18504 substantially pure full-length SspB polypeptide, e.g., a polypeptide which includes an amino acid sequence substantially identical to the amino acid sequence of SEQ
ID NO: 5 (incomplete protein sequence) and the r~in~er 5 of the SspB sequence. Full-length SspA and SspB genes can be isolated by those skilled in the art using the partial DNA sequences disclosed herein. The invention also includes a substantially pure full-length SspA
polypeptide, e.g., a polypeptide which includes an amino lO acid sequence substantially identical to the amino acid sequence of SEQ ID NO: 8 (incomplete protein sequence) and the remainder of the SspA seguence. The invention also features An active fragment of an Ssp B polypeptide or an SspC polypeptide or an SspD polypeptide is defined 15 as an SspB, SspC, or an SspD polypeptide, respectively, at least 50 amino acids, preferably at least 25 amino acids, more preferably at least lO amino acids in length having the ability to induce BME in the absence of the full-length version of the corresponding protein. In 20 other preferred embodiments the SspB, SspC, SspD or SspA
polypeptide is able to translocate into an epithelial cell, preferably a human epithelial cell. Translocation can be assayed using any suitable assay, e.g., the assay of Sogy et al. ~Molecular Microbiol . 14:583:94, 1994).
The invention also includes a substantially pure SspH polypeptide, e.g., a polypeptide which includes an amino acid sequence substantially identical to the amino acid sequence of SEQ ID NO:14, or a biologically active fragment thereof.
The invention also includes a substantially pure IagB polypeptide, e.g., a polypeptide which includes an amino acid sequence substantially identical to the amino acid sequence of SEQ ID NO:ll, or a biologically active fragment thereof.

CA 02237~81 1998-0~-13 WO97/1822S PCT~S96/18504 Also within the invention is an antibody which binds to a Ssp, e.g., a polyclonal or monoclonal antibody which specifically binds to an epitope of Ssp.
Polyclonal and monoclonal antibodies pro~llc~ against the 5 polypeptides of the invention can be used as diagnostic or therapeutic agents. The invention ~ncomr~ses not only an intact monoclonal antibody, but also an immunologically-active antibody fragment, e.g., a Fab or (Fab)2 fragment; an engineered single chain Fv molecule.
lO In preferred embodiments, the antibody may be l;nke~ to a detectable label, e.g. a radioactive label, fluorescent label, paramagnetic label, or colorimetric label.
The invention also includes a method of detecting a Salmonella infection in a mammal which includes the 15 steps of contacting a biological sample derived from the mammal, e.g., a human patient, with a Ssp-specific antibody and detecting the binding of the antibody to a Ssp in the sample. Antibody binding indicates that the r ~1 is infected with Salmonella. The presence o~
20 Salmonella in a biological sample may also be detected using a method which includes the steps of contacting the sample with a Ssp-encoding DNA, or the complement thereof, under high stringency conditions and detecting the hybridization of the DNA to nucleic acid in the 25 sample. Hybridization indicates the presence of Salmonella in the biological sample. By "high stringencyl' is meant DNA hybridization and wash conditions characterized by high temperature and low salt concentration, e.g., wash conditions of 65~C at a salt 30 concentration o~ approximately O.l x SSC. For example, high stringency conditions may include hybridization at about 42~C in the presence of about 50% ~ormamide; a first wash at about 65~C with about 2 x SSC cont~;n;ng 1%
SDS; followed by a second wash at about 65~C with about 35 O.l x SSC.

CA 02237~81 1998-0~-13 wO97/1822s PCT~S96/18504-The invention also features a method for detecting the presence of antibodies to an Ssp using all or part o~
an Ssp protein. The method includes contacting a biological sample with the Ssp protein and measuring the 5 binding of the Ssp protein to an antibody present in the sample.
The invention also features a method of targeting an antigen to an epithelial cell in a ~ ~l which includes the steps of linking the antigen to an Ssp, 10 e.g., SspC or SspD, or active fragment thereof, to produce a Ssp oh; -~iC antigen and a~ ; n; ~tering the chimeric antigen to the ~ ~l.
A method of inducing a cytotoxic T cell ;
response in a ~ ~l is also within the invention. This 15 method includes the steps of linking the antigen to an Ssp or active fragment thereof to produce a Ssp ~hime~ic antigen and contacting an antigen-presenting cell, e.g., a Class I major histocompatibility complex (MHC) antigen-expressing cell, with the ch; -ric antigen.
The invention also features a vaccine which includes a bacterial cell, the virulence of which is attenuated by decreased secretion of a Ssp, and a method of vaccinating an - =l, e.g., a h on patient, against a Salmonella infection by a~ ; n; stering such a vaccine.
25 Preferably, the bacterial cell is a Salmonella ty~hi ~ium cell, e.g., a Salmonella enteriditis cell, or a Salmonella typ~i cell. A live Salmonella cell in which a gene encoding a heterologous antigen is inserted into a Ssp-encoding gene is also included in the invention.
Also within the invention is a substantially pure StpA polypeptide and a method of dephosphorylating a protein which includes the steps of contacting the protein, e.g., a protein at least one tyrosine of which is phosphorylated, with a StpA polypeptide or an active 35 fragment thereof. An active fragment of StpA is defined CA 02237~8l l998-0~-l3 WO97/18225 PCT~S96/18sO4 as a Salmonella-derived polypeptide at least lO amino acids in length which is capable of removing a phosphate group from a tyrosine residue.
The invention feature live Salmonella (particularly Salmonella typhimurium) vaccines in which one or more gene required for BME is mutated so as reduce their activity. Among the genes which can be mutated are SspB, SspC, and SspD. Although SspA appears not to be re~uired for BME, it may be useful to mutate this gene as lO well (preferably in combination with mutation of one or more of the other Ssp genes). Any mutation of these genes which decreases function, including complete or partial deletion and one or more point mutations may be useful. In addition, function of Ssp gene may be 15 impaired by altering its control region.
The invention provides a Salmonella vaccine which does not cause transient bacteremia. In general, the invention features a bacterial cell, preferably a Salmonella cell, e.g., a S. typhi, S. enteritidis 20 ty~hi~7~ium, or S. cholerae-suis cell, the virulence of which is attenuated by a first mutation in an Ssp gene.
The preferred mutations are loss of function mutations.
However, functions causing partial loss of function may - be useful if virulence is ade~uately reduced. Such a 25 bacterial cell can be used as a vaccine to ; ~ln; ze a ~1 against salmonellosis.
The Salmonella cell may be of any serotype, e.g., S. typhimurium, S. paratyphi A, S. paratyphi B, S.
paratyphi C, S. pylorum, S. dublin, S. heidelberg, S.
30 newport, S. inne.sota, S. in~antis, s. virchow, or S.
pAnA --, The first mutation may be a non-revertible null mutation in one or more of the following genes: SspB, SspC, or SspD . Preferably, the mutation is a deletion 35 of at least lOO nucleotides; more preferably, the CA 02237~81 1998-0~-13 WO97/18225 PCT~S96/18504 mutation is a deletion of at least 500 nucleotides; even more preferably, the mutation is a deletion of at least 750 nucleotides. Mutations in the prg~ gene or the prg~
operon can be used ~or the same purpose.
In preferred embodiments loss or function (partial or complete) is due to decreased expression as a result of a change or mutation, e.g., a deletion, (preferably a non-revertible mutation) at the promoter or other regulatory element of SspB, SspC, or SspD (or some lO combination thereof).
In another aspect, the invention features a vaccine including a bacterial cell which is attenuated by decrease of expression of a Ssp virulence gene.
The invention also features a live Salmonella 15 cell, or a substantially purified preparation thereof, e.g., a S. typhi, S. enteriditis typhi ~ium, or S. cholerae-suis cell, in which there is inserted into a virulence gene, e.g., an Ssp gene, a gene encoding a heterologous protein, or a regulatory element thereof.
In another aspect the invention includes a method of vaccinating an ~n;m~l, e.g., a ~ 1, e.g., a human, against a disease caused by a bacterium, e.g., S~7 -n~lla~ including a~;n;~tering a vaccine of the invention.
By Uvaccine" is meant a preparation including materials that evoke a desired biological response, e.g., an immune response, in combination with a suitable carrier. The vaccine may include live organism, in which case it is usually a~~ ; ni ~tered orally, or killed 30 org~n; ~ or components thereof, in which case it is usually al ; n i ~tered parenterally. The cells used for the vaccine of the invention are preferably alive and thus capable of colonizing the intestines of the inoculated animal.

CA 02237~8l l998-0~-l3 WO97/18225 PCT~S96/18504 By Umutation~ is meant any change (in comparison with the appropriate parental strain) in the DNA sequence ~ of an org~ . These changes can arise e.g., spontaneously, by chemical, energy e.g., X-ray, or other 5 forms of mutagenesis, by genetic engineering, or as a result of mating or other forms of exchange of genetic information. Mutations include e.g., base changes, deletions, insertions, inversions, translocations or duplications.
A mutation attenuates virulence if, as a result of the mutation, the level of virulence of the mutant cell is decreased in comparison with the level in a cell of the parental strain, as measured by (a) a significant (e.g., at least 50%) decrease in virulence in the mutant 15 strain compared to the parental strain, or (b) a significant (e.g., at least 50%) decrease in the amount of the polypeptide identified as the virulence factor in the mutant strain compared to the parental strain.
A non-revertible mutation, as used herein, is a 20 mutation which cannot revert by a single base pair change, e.g., deletion or insertion mutations and mutations that include more than one lesion, e.g., a mutation composed of two separate point mutations.
Heterologous protein, as used herein, is a protein 25 that in wild type, is not expressed or is expressed from a different chromosomal site, e.g., a heterologous protein is one encoded by a gene that has been inserted into a second gene.
A substantially purified preparation of a 30 bacterial cell is a preparation of cells wherein cont~ ;n~ting cells without the desired mutant genotype constitute less than 10%, preferably less than 1%, and more preferably less than 0.1% of the total h~ of cells in the preparation.

CA 02237~81 1998-0~-13 WO97/l8225 PCT~S96/18504-A substantially pure DNA, as used herein, refers to a nucleic acid sequence, segment, or fragment, which has been purified from the sequences which ~lank it in a naturally occurring state, e.g., a DNA which has been 5 removed from the sequences which are normally adjacent to the fragment, e.g., the seguences adjacent to the fragment in the genome in which it naturally occurs. The term also applies to DNA which has been substantially purified from other components which naturally ac~ _Any 10 the DNA, e.g., DNA which has been purified from proteins which naturally accompany it in a cell.
Other features and advantages of the invention will be apparent ~rom the ~ollowing description of the preferred embodiments and from the claims.
By "polypeptide" is meant any chain of amino acids, regardless of length or post-translational modification (e.g., glycosylation or phosphorylation).
By "substantially identical" is meant a polypeptide or nucleic acid exhibiting at least 50%, 20 preferably 85%, more preferably 90%, and most preferably 95% sequence identity to a reference amino acid or nucleic acid sequence. For polypeptides, the length of c _-rison sequences will generally be at least 10 amino acids, preferably at least 20 amino acids, more 25 preferably at least 25 amino acids, and most preferably 35 amino acids. For nucleic acids, the length of c~ _A~ison sequences will generally be at least 50 nucleotides, preferably at least 60 nucleotides, more preferably at least 75 nucleotides, and most preferably 30 100 nucleotides.
Sequence identity is typically measured using sequence analysis software (e.g., Sequence analysis software package of the genetics ~ _~Ler group, university of Wisconsin biotechnology center, 1710 35 university avenue, Madison, WI 53705). Such software CA 02237~81 1998-0~-13 wo97/l822s PCT~S~6/18504-matches similar sequences by assigning degrees of homology to various substitutions, deletions, substitutions, and other modifications. Conservative substitutions typically include substitutions within the 5 following groups: glycine, alanine; valine, isoleucine, leucine; aspartic acid, glutamic acid, asparagine, glutamine; serine, thr~on;n~; lysine, arginine; and phenylalanine, tyrosine.
By a "substantially pure polypeptide" is meant a lO Ssp polypeptide which has been separated from components which naturally accompany it. Typically, the polypeptide is substantially pure when it is at least 60% Ssp by weight. Preferably, the preparation is at least 75%, more preferably at least 90%, and most preferably at 15 least 99~, by weight, Ssp polypeptide. A substantially pure Ssp polypeptide may be obtA;ne~, for example, by extraction from a natural source (e.g., Salmonella bacterium); by expression of a recombinant nucleic acid encoding a Ssp polypeptide; or by chemically synthesizing 20 the protein. Purity can be measured by any appropriate method, e.g., using column chromatography, polyacrylamide gel electrophoresis, or by ~PLC analysis.
A protein is substantially free of naturally associated components when it is separated from those 25 contaminants which acc_ any it in its natural state.
Thus, a protein which is chemically synthesized or produced in a cellular system different from the cell from which it naturally originates will be substantially free from its naturally associated components.
30 Accordingly, substantially pure polypeptides include those derived from one type of prokaryotic organism, e.g., S. typhimurium , but synthesized in E. coli or another prokaryotic organism.
By "substantially pure DNA" is meant DNA that is 35 free of the genes which, in the naturally-occurring CA 02237~81 1998-0~-13 WO 97/18225 PCT/USg6/18504 genome of the organism from which the DNA of the invention is derived, flank the gene. The term therefore i~cludes, for example, a recombinant DNA which is incorporated into a vector; into an autonomously 5 replicating plasmid or virus; or into the genomic DNA of a prokaryote or eukaryote; or which exists as a ~eparate molecule (e.g., a cDNA or a genomic or cDNA ~ragment produced by PCR or restriction endonuclease digestion) independent of other sequences. It also includes a 10 recombinant DNA which is part of a hybrid gene encoding additional polypeptide se~uence, e.g, a hybrid gene encoding a ~h;~iC antigen.
Other feat~res and advantages of the invention will be apparent from the following description of the~5 preferred embodiments thereof, and from the claims.
Petailed DescriPtion Fig. 1 is a diagram of the a genetic map of the 59-60 min region of the S. ty~i ~ium chromosome and partial physical map of the restriction endonuclease 20 sites of the prgH chromosomal region within the hil locus and related plasmids. The horizontal arrows indicate the direction of transcription of the orfl, prgHIJR, and org genes and of the neomycin promoter of the Tn5B50 insertions within the hil locus. The vertical arrows 25 indicate and the location of the predicted start of transcription of the prgHTJR operon (small arrow) and the location of the two Tn5B50 insertions that define the hil locus (large arrows). The open triangle indicates the location of the prgHl: :TnphoA insertion. Restriction 30 endonuclease sites are as follows: B, BamHI; E, EcoRI; H, HindIII; S, SacI; Ss, SspI; V, EcoRV; X, XhoI.
Fig. 2 is a photograph of a Northern blot assay in which the prgHIJK and org transcripts are identified.
Blot hybridization of a prgH (A), prgI-J (B) prgR (C), 35 org (D), and pagC (E) DNA probe to RNA purified from CA 02237~8l l998-0~-l3 WO97/18225 PCT~S96118504 wild-type (Wt) and phoP constitutive (PC) S. ty~hi~7~ium strains were grown aerobically to 0.5 optical density units. The bars indicate the RNA markers and are 9488, 6255, 3911, 2800, 1898, and 872 nucleotides (NT~ in size ~ 5 from top to bottom.
Fig. 3 is a photograph of a primer extension analysis of RNA isolated from wild-type and phoPC
5. ty~hi ~ium strains by using an oligonucleotide primer IB08 corresponding to nucleotides 1217 to 1199 of the 10 prgN se~uence. Lanes labeled "AGCT" represent dideoxy DNA se~uencing reactions. The lane labeled "wt"
represents the products of a primer extension reaction initiated with primer IB08 and wild-type RNA as a template, and the lane labeled "pc" represents the 15 products of a primer extension reaction initiated with the same primer and phoPC RNA as a template. Reverse transcription of wild-type RNA with primer IB08 resulted in an approximately 270-nucleotide product corresponding to a predicted transcriptional start at nucleotide 949 of 20 the prgH se~uence. Abbreviations: wt, wild type strain 14028s; pc, phoPC strain CS022.
Fig. 4A is a diagram showing the similarity and alignment of prgI, mxiH, and yscF predicted gene products.
Fig. 4B is a diagram showing the similarity and ali~_ -nt of prgJ and mxiI predicted gene products.
Fig. 4C is a diagram showing the similarity and ali~_ -nt of prgR, mxiJ, and yscJ predicted gene products. For Figs. 4A-4C, residues conserved among each 30 of the predicted gene products are indicated with a plus (+); residues conserved among the prgI and either the mxi~ or yscF predicted gene products and between the prgR
and either the mxiJ or yscJ predicted gene products are indicated with an asterisk (*). The location of the 35 lipoprotein processing sites (Leu-Xaa-Gly-Cys) of the CA 02237~8l l998-0~-l3 wo97/l822s PCT~S96/18504 prgR, mxiJ, and ysc~ predicted gene products are indicated by underlining. Predicted protein se~l~nc~
were compared using the GCG BLAST network service and ALIGN ~ Gy ~ (Feng et al., ~. Mol . Evol . 35:351-360, 5 1987; ~iggins et al., CABIOS 5:151-153, 1989).
Fig. 5 is a photograph of a SDS-PAGE gel.
Salmonella proteins found in the culture supernatant of stationary-phase S. tyrhi ~ium 14028s were compared to proteins isolated from lysed whole cells or cellular 10 fractions (membranes or intracellular soluble proteins).
TCA precipitable material from 2 ml of supernatant ~rom cultures of OD600 = 2.2 was used. The whole cell, membrane, and soluble lanes con~A;ne~ material from 0.10 ml, 0.35 ml, and 0.15 ml of cells, respectively.
15 Proteins were fractionated in a 12~ polyacrylamide gel by SDS-PAGE and stained with Coomassie Brilliant Blue ~-250.
The molecular masses of protein st~n~ds are indicated on the side of the gel as kDa.
Fig. 6 is a photograph of a SDS-PAGE gel showing a 20 comparison of culture supernatant proteins from S. ty~h;~ium 14028s and culture supernatants from mutants which are defective in eucaryotic signaling. TCA
precipitable material from 2 ml of bacterial culture supernatant was isolated at different times following 25 inoculation: mid-log, OD600 = 0.6; late-log /
early-stationary, OD600 = 1.1; stationary, OD600 = 2.2.
Proteins were fractionated in a 12% polyacrylamide gel by SDS-PAGE and stained with coomassie Brilliant Blue R-2sO.
The molecular masses of protein s~n~ds are indicated 30 on the side o~ the gel as kDa. wt, wild type (14028s);
pc, phoPC (CS022); P~, PhoP~ (CS015); ~hil (CS451), deleted for the hil locus.
Fig. 7 is a photograph of a SDS-PAGE gel showing an analysis of prgH: :TnphoA and complementation of the 35 insertion mutation by pWKSH5. TCA precipitable material CA 02237~8l l998-0~-l3 WO97/18225 PCT~S96/18sO4 from 2 ml of supernatant from stationary phase cultures was fractionated in a 10% polyacrylamide gel by SDS-PAGE.
Protein was StA i n~ with Coomassie Brilliant Blue R-250.
The molecular masses of protein standards are indicated - 5 on the side o~ the gel as kDa. wt, wild-type (14028s);
IB040, prg~1: :TnphoA; IB043, prg~l: :Tnp~oA with plasmid pWKSH5 containing a 5.1 kb insert of S. ty~hi~7~ium DNA
including prgHIJR. Supernatant protein bands complemented by pWKSH5 are indicated by arrows (87 kDa 10 and 65 k~a) and a bracket (three bands in the 35-40 kDa range).
Fig. 8 is a photograph of a SDS-PAGE gel showing Salmonella secreted proteins (Ssp) concentrated from supernatants of different strains. Each lane contains 15 Ssp collected from 2 ml of culture supernatant. Lanes l:
wild-type S. typhimurium SL1344; 2: EE638 (lacZY11-6); 3:
EE633 (lacZY4); 4: VB122 (~ilA::kan-112); 5: EE637 (invF::lacZY11-5); 6: IB040 (prgHl: :TnphoA) St: molec weight standard. Sizes of protein bands are given in 20 kDa. * marks a protein band which was variably present in different preparations of Ssp from the same strains.
Fig. 9 is a diagram showing the chromosomal organization of the sspBCDA genes and phenotypes of mutants sspC::~acZY4 (EE633) and sspA::lacZY11-6 (EE638).
25 The chromosomal location of ssp with respect to spaT and prgH is shown. An asterisk (*) indicates partially se~lenc~ genes. Restriction sites in parentheses have only been mapped in the left region of the 11 kb EcoRI
fragment. Abbreviations of restriction sites are: E:
30 EcoRI, B: BamHI, P: PvuII, N: NcoI . Invasion of epithelial cells by different 5. tyrhi ~ium strains is given as the percentage of the bacterial inoculum surviving gentamicin treatment. Values represent means and st~n~d errors o~ the means of three independent 35 experiments, each performed in triplicate. Presence or CA 02237~81 1998-0~-13 wo97/l822s PCT~S96/18504 absence of .~A7~o~e7la secreted proteins SspA, SspC and SspD in culture supernatants of different strains is indicated by + or -, respectively. The molecular weights in ~Da of these Ssp are shown in parentheses.
Fig. lO is a diagram showing a complementation analysis of EE638. Complementing fragments of chromosomal DNA in a low-copy plasmid are shown according to the chromosomal map. Designations of the plasmids are given in brackets on the left. The positions of the lac lO promoter (PlaC) are indicated. A indicates a deletion.
Fig. ll is a photograph of an immunoblot analysis of various strains for expression and secretion of Ssp87.
Total cellular proteins from bacteria collected from 0.2 ml of cultures were loaded in lanes designated "C", 15 supernatant proteins from 0.2 ml bacterial culture supernatants were loaded in lanes designated "S". l: wild type S. ty~h;m77~ium; 2: CS022 (PhoPC); 3: IBO40 (prgHl : :TnphoA); 4 : CS451 ~hil : :Tn5-428); 5: EE638 (sspC: :lacZY11--6); 6: EE633 (sspA: :lacZY4) .
Fig. 12 is a diagram showing a comparison of the deduced partial amino acid sequence of SspB with the S.
flexneri homologue IpaB. Bars indicate identical residues, dots indicate gaps introduced in order to ~ e similarity according to the GAP program of the 25 GCG package.
Fig. 13 is a diagram showing a comparison of the deduced amino acid sequences of SspC with the S. flexneri homologues IpaC. Bars indicate identical residues, dots indicate gaps introduced in order to ~;~;ze similarity 30 according to the GAP program of the GCG package.
Fig. 14 is a diagram showing a c~ p-~ison of the deduced amino acid sequences of SspD with the S. flexneri homologues IpaD. Bars indicate identical residues, dots indicate gaps introduced in order to maximize similarity 35 according to the GAP ~oy r am of the GCG package.

CA 02237~8l l998-0~-l3 WO97/18225 PCT~S96/18S04-Fig. 15 is a diagram of the amino-terminal sequence derived from the 5'-region of sspA . Amino acids determined by amino-terminal sequencing of SspC and SspA
are underlined.
- 5 Fig. 16 is a photograph of a SDS-PAGE gel showing total soluble Ssp collected from 2 ml of culture supernatants of wild type S. typhimurium SL1344 and EE638 (sspC: :lacZY11 - 6) transformed with various plasmids.
Lanes 1: SLl344 [pWSK29]; 2: EE638 [pWSK2~]; 3: EE638 [pCH004 (sspC) ]; 4: EE638 [pCH005 (sspCD) ]; 5: EE638 [pCH006 (sspD) ]; 6: EE638 [pCH002 (sspCDA) ]; 7: SL1344 [pCH002 (sspC~A) ]. Lanes 8 and 9 contain soluble Ssp from SL1344 [pWSK29] and EE638 [pWSK29], respectively.
The sizes of the protein bands are given in kDa. An 15 asterisk (*) indicates a protein band which was variably present in different preparations of Ssp from the same strains.
Fig. 17 is a photograph of an SDS-PAGE gel showing insoluble Ssp precipitates collected from 2 ml of culture 20 supernatants of wild type S. ty~him~ium SL1344 and EE638 (sspC: :lacZY11 - 6) transformed with various plasmids.
Lanes 1: SL1344 [pWSK29J; 2: EE638 [pWSK29~; 3: EE638 [pCH004 (sspC) ]; 4: EE638 [pCH005 (sspCD) ]; 5: EE638 [pCH006 (sspD) ]; 6: EE638 [pCH002 (sspCDA) ]; 7: SL1344 [pCH002 (sspCDA) ]. Lanes 8 and 9 contain soluble Ssp ~rom SL1344 [pWSK29] and EE638 [pWSK29], respectively.
The sizes of the protein bands are given in kDa. An asterisk (*) indicates a protein band which was variably present in different preparations of Ssp from the same 30 strains.
Fig. 18 is a diagram showing the genetic organization of the invasion gene clusters ~rom S. tyrhi i~ium and S. flexneri . The relative positions of each gene are indication and the directions of gene 35 transcription are indicated by arrows. Arrows are not CA 02237~8l l998-0~-l3 Wos7/l822s PCT~S96/18504 drawn to scale. Gene clusters conserved in sequence and gene order are indicated ~y stippling (inv-spa/mxi-spa), crosshatching (prgl~/mx7HI] ), and dark arrows (ssp/ipa) .
Genes with no homologues within the respective regions 5 are shown as open arrows.
Fig. 19 is a depiction of the nucleic acid sequence of SspB (missing part o~ the 5' end) tSEQ ID NO:
1~ .
Fig. 20 is a depiction of the nucleic acid lO sequence of SspC ( SEQ ID NO: 2).
Fig. 21 is a depiction of the nucleic acid sequence of SspD ( SEQ ID NO: 3).
Fig. 22 is a depiction of the nucleic acid sequence of SspB (missing part of the 3' end) (SEQ ID NO:
15 4) and the predicted amino acid sequence SspB (partial c-terminal) (SEQ ID NO: 5).
Fig. 23 is a depiction of the predicted amino acid sequences of SspC (SEQ ID NO: 6), SspD (SEQ ID NO: 7), and SspA (partial animo te~ ;n~l ) (SEQ ID NO: 8).
Fig. 24 is a depiction of the nucleic acid sequences of iagB (SEQ ID NO: 9) and stpA (SEQ ID NO:
10) .
Fig. 25 is a depiction of the predicted amino acid sequences of iagB (SEQ ID NO: 11) and stpA (SEQ ID NO:
25 12).
Fig. 26 is a depiction of the nucleic acid sequence of prgH (SEQ ID NO: 13).
Fig. 27 is a depiction of the predicted amino acid se~uences of prgB (SEQ ID NO: 14).
Fig. 28 is a depiction of the nucleic acid sequence of ~spBCDA (truncated at 3' and 5' ends) (SEQ ID
NO: 15).
Fig. 29 is a depiction of the nucleic acid sequence of prg~ and 5' and 3' flanking sequences (SEQ ID
35 NO: 16).

CA 02237~81 1998-0~-13 WO 97/18225 PCT/US96/18~;04 Ss~ Proteins and Genes The Salmonella secreted proteins (Ssp) of the ~ invention have a variety of uses. For example, they can be used as diagnostic reagents, therapeutic agents, and - 5 research products. The genes encoding Ssp also have a variety of uses. For example, they can be used as diagnostic reagents. They can also be used to create vaccines including live attenuated vaccines.
Because Salmonella infection is a significant 10 health problem and because Ssp proteins are soluble proteins that are found on the surface of Sal monel l a, various Ssp, DNA encoding various Ssp, and antibodies directed against various Ssp are useful in diagnostic assays. Because Ssp are required for optimal virulence, 15 ~NA encoding a mutant Ssp having decreased function can be used to create strains of Sal monel l a with reduced virulence. Such strains are useful as live vaccines.
An Ssp (or a portion thereof which can gain entry into the cytoplasm) can be used to translocate a second 20 molecule, e.g., a polypeptide, into the cytoplasm of a cell. This approach can be use~ul ~or the induction or priming of cytotoxic lymphocytes (CTL) directed against the second molecule. An Ssp (or a portion thereof capable of translocating an attached second molecule) can 25 be used to introduce a second molecule into the cell cytoplasm for the purpose of drug delivery. Often the second molecule is a polypeptide which is covalently linked to an Ssp (or a portion thereof), e.g., by a peptide bond. Such molecules can be readily produced 30 first preparing a ~h; ~~~iC gene encoding the Ssp (or portion thereof) and the second molecule as a single polypeptide chain. This gene can be used to prepare the fusion protein for administration to a patient.
Alternatively, the Ch; - ~iC gene can be introduced into a CA 02237~81 1998-0~-13 WO97/18225 PCT~S96/18504 strain of Salmonella which can then be used as either a live vaccine or drug delivery system.
S8p as Diaqnostic Reagents An Ssp can be used as a diagnostic tool for the 5 detection of Sa7monella infection in a patient or to evaluate status of an immune response to Salmonella. For example, one or more Ssp can be used an antigen in an ~rT,~A assay to detect the presence of Salmonella-specific antibodies in a bodily fluid, e.g., blood or plasma, lO obtained from an infected patient or an individual suspected of being infected with Salmonella. Ssp can also be used to test i ~ cell activation, e.g., T or B
cell proliferation or cytokine production, in a sample of patient-derived cells, e.g., peripheral blood mononuclear 15 cells, to detect the presence of a cellular immune response to Salmonella.
Polynucleic acids (e.g., primers and probes) encoding all or part of an Ssp can be used in hybridization assays to detect the presence Salmonella 20 infection, e.g., using a PCR assay or other probe or primer based assay designed to detect particular DNA
sequences.
Antibodies capable of selectively binding a particular Ssp can be used to detect the presence of 25 SA7~ne71a in a biological sample. Such antibodies can be produced using st~n~d methods.
Thera~eutic A~lications of Ss~ Fusion Proteins Fusion proteins comprising all or part of an Ssp and a second protein or polypeptide are useful for a 30 variety of therapeutic applications such as vaccines (e.g., recombinant Salmonella vaccines or vaccines against heterologous pathogens), cell targeting agents for delivery of drugs (e.g., cytotoxic agents), and adjuvants, (e.g., to boost an immune response to a co-3 5 ~m in;~tered antigen).

CA 02237~81 1998-0~-13 WO97/18225 pcT~s96/l8so4 To produce a recombinant Salmonel 7a vaccine, a gene encoding an Ssp fusion protein can be introduced into a Salmonella vaccine. Because Ssp are involved in bacterial mediated endocytosis, the Ssp fusion protein ~ 5 will cause the second polypeptide or protein to be internalized by epithelial cells (or other cells to which the Ssp binds) of the individual to which the vaccine is a~~ i n; ~tered. This internalization can trigger a Type I
MHC-mediated response to the second protein or lO polypeptide. The induction of this response will lead to the induction of CTL (or the priming of CTL) specific for the second protein or polypeptide. The induction or priming of antigen-specific CTL can provide therapeutic or prophylactic benefits.
Purified fusion proteins can be used as recombinant vaccines. Proteins fused to Ssp are specifically targeted to epithelial cells or other cell types to which the Ssp bind; the fusion proteins are then internalized by the targeted cells. Thus, Ssp fusion 20 proteins are useful to generate an ; ~ response to the antigen to which the Ssp is linked or to deliver a therapeutic compound, e.g., a toxin ~or the treatment of cancer or autoimmune diseases in which the killing of specific cells, i.e., the cells to which a Ssp binds, is 25 desired. Delivery of a toxin linked to a SspC or SspD
polypeptide is especially useful in cancer therapy because man types of cancers are of epithelial cell origin.
Ssp fusion proteins which contain all or part of a 30 Ssp linked to a heterologous protein can be made using methods known in the art. Two or more polypeptides may be linked together via a covalent or non-covalent bond, or both. Non-covalent interactions can be ionic, hydrophobic, or hydrophilic.

CA 02237~81 1998-0~-13 wo97/l822s PCT~S96/18504 A covalent linkage may take the form of a disulfide bond. For example, the DNA encoding one of the polypeptides can be engineered to contain a unigue cysteine codon. The second polypeptide can be 5 derivatized with a sulfhydryl group reactive with the cysteine o~ the first component. Alternatively, a sul~hydryl group, either by itself or as part o~ a cysteine residue, can be introduced using solid phase polypeptide techniques.
A number of other covalent crosslinking agents, e.g., photoreactive crosslinkers, water-soluble crosslinkers, which are ~ ~rcially available may be used to join a heterologous polypeptide to a Ssp to create a fusion protein. If the fusion protein is 15 produced by expression of fused genes, a peptide ~ond serves as the link between the components of the fusion protein. Such fusion proteins are produced by expression of a ~h i ~iC gene in which se~uences encoding all or part o~ an Ssp are in ~rame with seguences encoding the 20 second protein or polypeptide. In some circums~nc~ it may be useful to include a linker polypeptide between the Ssp and second protein of polypeptide.
Internalization of the fusion protein may not require the presence of a complete Ssp protein. A
25 internalization-competent portion of an Ssp will be adequate in many circumstances. Whether a particular portion of a selected Ssp is sufficient for internalization can be tested as follows. The selected portion of an Ssp is fused to a ~1 -d~lin-dependent 30 adenylate cyclase. If this test fusion protein ii internalized, it will be exposed to calmodulin and the cylcase will be activated. The presence of adenylate cyclase activity can then be used as a measure o~
internalization. This general approach is described by 35 Sorg et al. (Nolecular Mcrobiol. 14:583-94, 1994).

CA 02237~81 1998-0~-13 wos7/l822s PCT~S96/18504 Ssp are virulence factors that alter the ability of bacteria to be internalized by specific populations of host cells and to induce an immune response. Sa 7monella with mutations in genes encoding Ssp are useful in the - 5 manufacture of live Salmonella vaccines with altered cell tropism.
Deletion or overexpression of Ssp in Salmonella can be used to target strains or fusion proteins to various ~ ~lian cell types. Invasion of epithelial lO cells or macrophages can be selected depending on the Ssp mutated. For example, use of Salmonella as an antigen or drug delivery vehicle can be optimized by deleting part or all of a gene encoding a Ssp involved in bacterial mediated endocytosis (or mutating such a gene to impair 15 Ssp function), thereby minimizing the ability of Salmonella to invade epithelial cells (and therefor ~; ;zing antigen delivery to antigen presenting cells such as macrophages). In this ~nne~, strains with mutated Ssp genes can be used to modulate the host immune 20 system. Deletion of Ssp genes in Salmonella can also be used to alter the ability of Salmonella to stimulate IL-8 secretion by epithelial cells.
Fusions of antigens to Ssp genes can be used to facilitate an immune response to the linked antigens for 25 the purpose of generating an antigen-specific cytotoxic T
cell response in a patient. For example, Ssp fusions to viral antigens are useful as therapeutic vaccines for diseases such as AIDS and Herpes genitalis in which the generation of a cytotoxic T cell (CTL) response is 30 desired. Delivery of antigens in this ~nn~ favors the generation of an antigen-specific CTL response because the Ssp portion of Ssp fusion protein mediates translocation of the fusion protein across eucaryotic cell membranes into the intracellular compa~t -nts in the 35 cytoplasm of cells which participate class I MHC-mediated CA 02237~81 1998-0~-13 2~ PCTIUS96/18504 antigen processing and presentation, i.e., the generation of class I MHC-restricted antigen--specific CTLs.
Fusion proteins which include all or part of a Ssp linked to a cytotoxic molecule can be used to target a 5 cytotoxic molecule to a specific cell type, e.g., an epithelial cell--derived cancer cell, which would then by killed by the cytotoxic agent. Cytotoxic fusion proteins can be synthetically or recombinantly produced and a~ ; n; stered directly to a patient. Alternatively, live lO Salmonella expressing a cytotoxic Ssp fusion protein can be administered and allowed to produce and secrete the fusion protein in vivo.
Ssp are also useful as adjuvants to boost the immunogenicity of antigens with which they are delivered 15 or to which they are chemically or recombinantly linked.
Ssp that have enzymatic effects, e.g., phosphatase activity, on certain types of eucaryotic cells can be used to promote specific types of i c responses such as TH2 or THl T cell responses. Since these proteins are 20 secreted and are likely taken up in the cytoplasm of eucaryotic cells, gene fusions to these proteins are likely to be more; -- ogenic and more efficient in inducing the development of an immune response, particularly a class I MHC--restricted CTL response.
Various oral and parenteral delivery systems are known in the art and can be used to deliver the Ssp polypeptides and/or ~.him-~ic antigens of the invention, such as encapsulation in liposomes, or controlled release devices. The compositions of the invention can be 30 formulated in a pharmaceutical excipient in the range of approximately lO ,ug/kg and lO mg/kg body weight The compositions and methods of the invention provide the tools with which to construct better vaccines against Salmonella infection and for the prevention and 35 treatment of other diseases, e.g., cancer and AIDS, by CA 02237~81 1998-0~-13 WO97/1822s PCT~S96/18504' using Salmonella secreted proteins as carriers of heterologous antigens, e.g., tumor antigens or viral antigens, either as purified components or as hybrid proteins produced in live Salmonella vaccine strains.
5 Ssp and Attenuated Bacterial Str~; n~
Deletion or mutation of one or more Ssp genes can be used to attenuate vaccine strains. For instance deletion of Ssp genes leads to lack of neutrophil transmigration across epithelial cell lO monolayers (a model system that correlates well with the ability of certain strains to cause gastroenteritis).
Vaccine strains are usually a~; n; ~tered at doses of l x 105 to l x lOlO cfu/single oral dose. Those skilled in the art can determine the correct dosage using 15 stAn~d t~c~n;ques.
Research Products Ssp with enzymatic activity, e.g., Salmonella tyrosine phosphatase (stpA), can be used as reagents for protein modification. StpA catalyzes the release of 20 phosphate groups from tyrosine residues in proteins, and thus, is especially useful in the field of signal transduction. Since a ~,~ h~ of eucaryotic and procaryotic signal transduction proteins are regulated by the phosphorylation and dephosphorylation of tyrosine 25 residues, stp can be used to deactivate or activate these proteins, thereby altering intracellular signal transduction. Thus, Stp can be used as a research tool to study and evaluate phosphorylation-regulated signal transduction pathways.
30 ~odi~ication of Ssp and SsP Variants When an Ssp is being used to translocate a second molecule into a eukaryotic cell, it may be use~ul increase expression of the Ssp (or Ssp fusion protein) so that BME is increased. Increased expression of sspC, 35 sspD and other ssp genes may be accomplished using CA 02237~81 1998-0~-13 WO97/1822s PCT~S96/185~4 methods known in the art, e.g., by introducing multiple copies of the gene(s) into the bacterial cell or cloning the Ssp-encoding DNA under the control of a strong promoter.
Under other circumstances it may be desirable to increase uptake of a bacterial strain, e.g., a Salmonella strain, by a macrophage in a ~ -1 by impairing the normal invasion m~ch~n;~m of the strain. This can be accomplished by decreasing expression of the DNA encoding lO the SspC and/or SspD (and thereby decreasing secretion of Ssp and/or SspD polypeptides) and a~ n; ~tering the cell to the mammal. Ssp expression may be reduced using methods known in the art, e.g., insertion of a transposon (Tn) into the gene, deletion of some or all of the gene, 15 mutating a gene upon which SspC and/or SspD expression depends, e.g., prgH , e.g., a deletion or Tn insertion in the prgHIJK operon. Instead of decreasing the expression of sspC and/or sspD, the method may include the step of impairing the function of one or both of the gene 20 products, e.g., by Tn insertion, deletion mutagenesis, o~
by impairing the secretory pathway by which the gene products are secreted such that the gene products are produced but not effectively transported to the extracellular enviro; ~nt.

25 Exam~le l: PhoP/PhoO TranscriPtional Repression of S. tyrhi~1~rium Invasion Genes: Evidence for a Role in Protein Secretion The PhoP-repre~sed prg~ locus of S. ty~h i m~'~ium may ~e important for signaling epithelial cells to 30 endocytose S. tyrhi ~ium. The following series o~
experiments demonstrate that the prgH locus is an operon of four genes encoding polypeptides of 392 amino acids (prgH), 80 amino acids (prgI), lOl amino acids (prgJ), and 252 amino acids (prgK). These experiments also CA 02237~8l l998-0~-l3 WO97tl8225 PCT~S96/18504 demonstrate that expression of the 2.6-kb prgHI~
transcript is reduced when PhoP/PhoQ is activated, - suggesting that PhoP/Pho~ regulates prgHIJK by transcriptional repression. Further, analysis of the 5 culture supernatants from wild-type S. tyrhi ~ium revealed the presence of at least 25 polypeptides larger than 14 kDa. Additional experiments demonstrated that prgH1: :Tnrh~, phoP constitutive (PhoPC), and hil deletion mutants have significantly defective supernatant lO protein profiles. A further set of experiments described below demonstrate that both the invasion and supernatant protein profile defects of the prg~1: :TnphoA mutant can be complemented by a 5.1 kb plasmid that included prgHI~K. Taken together these results suggest that 15 PhoP/PhoQ regulates extracellular transport of proteins by transcriptional repression of secretion deter ;nAntS
and that secreted proteins are likely involved in signaling epithelial cells to endocytose bacteria.
The following reagents and procedures were used to 20 evaluate the prg~ locus.
Bacterial Strains Growth and Conditions S. ty~hi ~ium strain ATCC 14028s (American Ty-pe Culture Collection, Bethesda, MD) is a virulent wild-type parent strain from which all other ~A7 one7 7a strains 25 described in Example 1 were derived. Bacterial strains and plasmids are described in Table 1. Luria-Bertani broth (LB) was used as rich bacterial growth medium.
Antibiotics were added to LB broth or agar in the following concentrations: ampicillin, 25 ~g/ml;
30 chloramphenicol, 50 ~g/ml; k~nA y~in, 45 ~g/ml.
DNA se~uencin~ and analYsis Double-strand templates were sequenced by the dideoxy-chain termination method known in the art as ~ modified ~or use with SequenaseT~ (US Biochemicals, Corp.) 35 and ~-35S]dATP. Computer analysis of the DNA se~uence was accomplished with the GENEPR0 (Riverside Scientific, Riverside, CA) and Wisconsin package ~GCG, version 7) programs. The nucleotide seguence of the prgHIJR locus has been deposited in GeneBank under accession number U21676.

RNA extraction. RNA blot analyses and ~rimer extension ~na 1 Yses RNA was isolated from mid-log phase cultures (OD600 ~ 0.5) of aerobically-grown (with shakin~) and 10 microaerophically-grown (without ~h~king) SA7~0ne7 7a strains using a stAn~d hot phenol procedure (Pulkkinen et al., J. Bacteriol. 173:86-93, 1993). For RNA blots, 20 ,ug of RNA was diluted in H20 and incubated for 15 minutes at 55~C in 50% formamide, 17.5% formaldehyde in 1 15 x Northern buffer (0.36 M Na2HP04-7H20, 0.04 M
NaH2P04-H20). Samples were run on 1% agarose gels cont~inin~ 6% formaldehyde and 1 x Northern buffer and were transferred to Gene Screen Plus membranes (NEN/Dupont). RNA was crosslinked to the membrane using 20 a StratalinkerTM W crosslinker (Stratagene, La Jolla, CA). Membranes were hybridized and washed according to the manufacturerrs protocol.
The DNA probes for RNA-DNA and DNA-DNA blot hybridization were obtained from recombinant plasmid DNA
25 by restriction endonuclease digestion or by polymerase chain reaction (PCR) using the G~nf~ TM PCR kit (Perkin-Elmer/Cetus). The following DNA probes were synthesized: a 841-bp prgH probe from the oligonucleotide primers IB07 (5'-CCAGGTGGATACGGA-3'; SEQ ID N0: 17;
30 nucleotides 1198 to 1212) and IB19 (5'-TAGCGTCCTCCCCATGTGCG-3'; SEQ ID N0: 18; nucleotides 2039 to 2021); a 433-bp prgI - prgJ probe from the primers IB26 (5'-CCGGCGCTACTGGCGGCG-3'; SEQ ID N0: 19), nucleotides 2304 to 2321) and DP04 (5'AGCGTTTCAACAGCCCCG-3'; SEQ ID N0: 20), nucleotides CA 02237~8l l998-0~-l3 WO97/18225 PCT~S96/18504 2737 to 2719); a 341-bp prg~ probe from primers DP03 (5'-CGGGGCTGTTGAAACGC-3'; SEQ ID N0: 21), nucleotides 2720 to 2736) and DP08 (5'-AACCTGGCCTTTTCAG-3'; SEQ ID
N0: 22), nucleotides 3060 to 3045); a 724-bp org probe ~ 5 from primers DP15 (5'-GGCAGGGAGCCTTGCTTGG-3'; SEQ ID N0:
23), nucleotides 3774 to 3792) and DP17 (5'-GTGCCTGGCCAGTTCTCCA-3'; SEQ ID N0: 24); and a 608-bp pagC probe from a Psi and StuI restriction-endonuclease digest of pWPL4 that contains the wild-type pagC gene.
10 DNA probes were radiolabelled using a s~An~rd method of random priming with [Q-32P~dCTP.
For primer extension analyses, oligonucleotide primers (0.2 picomoles) were end-labelled with [y-32P~dATP
(NEN/Dupont), annealed to S. typhimurium RNA (20~g) and 15 extended with reverse transcriptase (Gibco BRL, St.
Louis, M0). Reactions were electrophoresed in 6%
polyacrylamide, 8 M urea gels ad~acent to sequencing reactions initiated with primers used for cDNA synthesis.
DNA blot hybridization analysis Chromosomal DNA was isolated, restriction endonuclease digested, size fractionated in agarose gels, and transferred to GeneScreen Plus membranes (NEN/Dupont). For dot blot hybridization experiments, high stringency hybridization was performed according to 25 s~An~Ard methods at 65~C using radiolabelled probes.
Protein isolation and analYsis Bacteria were grown in LB, with shAking at 37~C.
Bacterial cultures were chilled to 4~C and centrifuged at 154,000 x g for 1.7 hours. The supernatant was carefully 30 removed and trichloroacetic acid (TCA) was added to a final concentration of 10%. The precipitates were collected by centrifugation at 69,000 x g for 1 hour, rinsed with cold acetone, dried and stored at 4~C. The bacterial cell pellet was fractionated to obtain 35 periplasmic, cytoplasmic, and membrane fractions.

CA 02237~8l l998-0~-l3 W097/~8225 PCT~S96/18S04 Samples were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on a 10-12~ polyacrylamide (0.1 M Tris pH 8.45, 0.1% SDS) gel using a st~n~Ard Tris-glycine buffer system or 5 st~n~rd Tris-tricine buffer system. TCA precipitates were ~;~ with sample buffer (250 mM Tris pH 6.8, 2%
SDS, 0.0025% bromophenol blue, 5.0~ ~-mercaptoethanol, 10~ glycerol) and heated to 100~C for 5 minutes.
Proteins were visualized by st~in;ng with Coomassie 10 Brilliant Blue R-250.
Enzvme assays Presence of the marker enzymes, alkaline phosphatase (periplasm) and ~-galactosidase (cytoplasm) were used to assess fraction purity. A plasmid, pPOS3, 15 cont~; n; ng an arabinose-inducible phoA gene, was inserted into wild-type strain 14028s by transformation and moved into other strains using P22 bacteriophage-mediated transduction. Addition of arabinose (0.02%) to the culture medium induced transcription of the phoA gene.
20 Dete~ tion of alkaline phosphatase activity of strains contA; n;ng pPOS3 was performed using the substrate p-Nitrophenyl phosphate according to st~n~rd methods.
The results were expressed in st~n~rd units for ~-galactosidase (Miller, J.H., 1972, Experiments in 25 Molecular Genetics, Cold Spring Harbor Laboratory, Cold Spring Har~or, N.Y., pp. 352-355). ~-galactosidase was produced from a strain with a mudJ-generated gene fusion of msg and lacZ. The gene, msg, is constitutively expressed and not PhoP regulated. ~-galactosidase 30 activity of strains carrying msg: :MudJ was measured using routine methods (Miller et al., supra).

CA 0223758l l998-05-l3 W097118225 PCT~S96/18S04 Table 1. Bacterial strainsr plasmids and relevant ~roperties .

5 . ty~h i ~ium Relevant genotYPe ATCC 14028s Wild Type 5 CS002 pho-24 CS019 phoN2zxx::6251~nlOd-Cm IB040 CSOl9 with prgHl::TnphoA
IB043 IB040 with pWKSH5 CS015 phoP-102::TnlOd-Cm 10 CS451 14028s derivative of EE451 with hhi Esch~ichia coli DH5~ F~~8~dlacZ~M15~ (lacZYA-argF)U169endAl re~AlhcdRl7deoRthi-lsupE44AgyrA96relA
Plasmids 15 pIBOl: pUC19 (ampR) cont~;~;ng a 10.7-kb EcoRV fragment with prg~l: :TnphoA (kan~) pVB3 pUCl9 cont~;n;ng a 5.9-kb ~indIII- EcoRI fragment of the prgN locus pWKSH5 pWKS30 (ampR) contA;n;ng a 5.1-kb ~indIII fragment of prgH locus pWPL4 pUC19 cont~;ning a 5.0-kb EcoRV fragment of the pagC locus pPOS5 pBR322 cont~;n;ng arabinose-inducible PhoA
Cloning and sequencing of pra~
The DNA cont~;n;ng the prgHl: :TnphoA gene fusion was cloned based upon information derived from the physical map of restriction endonuclease sites surrounding the transposon insertion (Fig. 1) (Behlau et al., J. Bacteriol. 175:4475-4484, 1993, hereby 30 incorporated by reference). Chromosomal DNA from strain IB040 cont~;n;ng the prg~l: :TnphoA insertion was digested with the restriction endonuclease EcoRV and ligated into SmaI digested pUC19 to generate a library of recombinant plasmids. These recombinant plasmids were transformed 35 into Escherichia coli (E. coli) DH5~. A recombinant plasmid con~in;ng a 10.7 kb EcoRV fragment was identified by selecting for kanamycin resistance (Tnpho~

CA 02237~8l l998-0~-l3 wo97/l822s PCT~S96/18~04 encoded) and was designated pIB01 (Fig. 1). DNA
hybridization analysis of strain IB040 with a radiolabelled 1.5-kb HindIII-SacI-generated DNA fragment of pIB01 resulted in hybridization to an approximately 5 10.7-kb EcoRV DNA fragment. This was approximately 7.7 kb (the size of TnphoA) larger than the 3-kb fragment present in the wild-type strain ATCC 14028s. This probe also hybridized strongly to plasmid pVB3 that cont~;n~
the 5.9 kb HindIII-EcoRI fragment of the hil locus (Fig.
10 1), confirming the location of the prgH locus within this region. This data indicated DNA cont~; n i~g the prgHl::TnphoA insertion had been cloned.
The DNA sequence of the 4,034-bp Hin~III-SspI
fragment (within which the TnphoA insertion in prgH was 15 localized) was dete~ ;ne~ ~y sequencing plasmid pIB0 cont~in;ng the cloned prgHl: :Tnpho~ allele. This sequence was confirmed by DNA se~uencing of pWKSH5 cont~;n;ng the wild-type prgH allele (Fig. 1).
Information from DNA sequence of the prgHl::phoA fusion 20 junction was used to determine the direction of transcription and correct reading frame of prgH. TnphoA
was inserted after nucleotide 1548 within an open reading frame that ext~n~e~ from nucleotides 981 to 2156. prgH
was predicted to encode a 392 amino acid polypeptide with 25 a calculated Mr ~f 44,459 daltons and pI of 5.86. The N-terminal portion of prgH was found to have a stretch of nonpolar residues followed by the motif Leu-Xaa-Gly-Cys at residues 24 to 27 (corresponding to nucleotides 1050 to 1061) characteristic of the processing site of 30 bacterial lipoproteins. There was a strong hydrophobic domain (amino-acid residue 144 to 154, corresponding to nucleotides 1410 to 1433) upstream of the TnphoA
insertion.
Analysis of the nucleotide se~uence located 35 upstream of prgH revealed an additional open reading CA 02237~8l l998-0~-l3 W097/18225 PCT~S96/18504 frame from nucleotides 665 to 222, termed or~l, likely to be oppositely transcribed from prgH. The intergenic region between or~l and prgH was 216 nucleotides. orfl was predicted to encode a gene product of 148-amino-acid 5 residues with a calculated Mr of 17,186. The start codon of orfl was preceded by a potential ribosome b;n~ing site at 7 to 11 nucleotides 5' to the predicted start of translation (5'-AAAGG-3', nucleotides 676 to 672) suggesting that this open reading frame was translated.
10 The orfl predicted gene product had no signal sequence nor any strong hydrophobic dl_ ~;n~.
Identification of praI, prgJ, and praR
Analysis of the nucleotide sequence located downstream from prgH revealed four additional open 15 reading frames that were predicted to be transcribed in the same direction and form an operon: (a) nucleotides 2184 to 2423; (b) nucleotides 2445 to 2747; (c) nucleotides 2747 to 3502; and (d) nucleotide 3476 to beyond the 3' SspI site. The first three of these four 20 open reading frames identified were designated prgI, prgJ, and prgR respectively. prgI, prgJ, and prgR were predicted to encode gene products of 80 amino acids (Mr~
8865 daltons), 101 amino acids (Mr~ 10,929 daltons), and 252 amino acids (Mr~ 28,210 daltons). The predicted gene 25 products encoded by prgI and prgJ did not contain a signal sequence or strong hydrophobic domains. The predicted gene product encoded by prgR contained a N-terminal hydrophobic region followed by a potential lipoprotein processing site from amino-acid residue 15 to 30 18 (corresponding to nucleotides 2788 to 2800). The fourth open reading frame corresponded in DNA seauence to the S. tyrhi ~ium oxygen-regulated gene ( org).
~r~H-R transcriPtion is neqatively regulated by PhoP/PhoO
To determine whether prg~ was negatively regulated 35 by PhoP/PhoQ, RNA isolated from wild-type ~ATCC 14028s) CA 02237~81 1998-0~-13 W097/18225 PCT~S96/1850 and phoPC (CS022) strains of S. typhimurium were analyzed. In numerous RNA blot analyses, the prgH-specific DNA probe hybridized with an approximately a 2600-nucleotide RNA from the wild-type ~train (Fig. 2).
5 The size of the RNA that hybridized to the prgH probe was similar to that of the prgH - R open reading frame predicted from the DNA sequence ~i.e., 2600 vs. 2522 nucleotides). In contrast, no transcript was seen when equal amounts and similar guality of RNA (as assessed by lO methylene blue stA; n; ng) isolated from the phoPC strain was probed with prgH-specific DNA (Fig. 3). In comparison, when the same RNA preparations were hybridized with a pagC-specific probe, an approximately llO0-nucleotide pagC transcript was highly expressed in 15 the phoPC strain (Fig. 2), consistent with the constitutive phenotype of pag gene expression in the phoPC mutant (plllkk;nen et al., J. Bacteriol. 173:86-93, l99l, here~y incorporated by re~erence). These results indicate that regulation of prgH occurs at the level of 20 transcription.
Primer extension analysis was performed to obtain information on the possible initiation site of prgH
transcription. Based on this analysis, the start of prgH
transcription was predicted to begin approximately 25 32 nucleotides upstream from the prgH translational start (Fig. 3). Several different primers were used that resulted in primer extension products of differing lengths, but all predicted that transcription initiated at this site. The predicted -lO (5'-TAATCT-3') and -35 (5'-TTCA$C-3') regions are similar to the consensus sequences for typical ~70 E. coli promoters. Similar to the results of RNA blot hybridization analysis, a primer extension product was detected only with RNA isolated from wild-type S. ty~hi ~ium and not with RNA isolated 35 from the phoPC strain (Fig. 3).

CA 02237~8l l998-0~-l3 WO97/18225 PCT~S96/18504-The size of the RNA that hybridized to the prgH-specific probe suggested that prgH-R could form a transcriptional unit. Therefore, to determine whether prgI-R formed an operon that was regulated by PhoP/PhoQ, 5 RNA blot hybridization and primer extension analysis were performed using DNA probes and primers specific to the prgI, prg~, and prgR open reading frames. Similar to the results with prgH, the prgI-J - and prgR-specific DNA
probes hybridized with an approximately 2600-nucleotide 10 RNA isolated from wild-type S. typhimurlum and not with RNA from the phoPC strain (Fig. 2). No primer extension products less than 350 nucleotides were detected using RNA isolated from either the wild-type or phoPC strains using prgI, prgJ, and prgR primers. These primers were 15 from 1662 to 2332 nucleotides downstream from the predicted start of prgH transcription. These fin~;n~s indicated that prgH-R were transcribed as an operon, heretofore referred to as prgHIJR. Furthermore, this operon was likely to be transcribed from the prgH
20 promoter and was negatively regulated by PhoP at the level of transcription.
org is not requlated by PhoP/PhoO
Although the above results suggested that the prgNI~R transcriptional unit did not include org, 25 experiments were performed to test this possibility.
Blot hybridization analysis was performed with RNA
isolated from wild-type S. typhimurium and an org-specific DNA probe. As shown in Fig. 2, two distinct transcripts hybridized to the org probe: an approximately 30 1400-nucleotide abundant RNA and a minor RNA of approximately 3800 nucleotides. The size of the smaller RNA was similar to that of the org open reading frame ~1400 vs. 1236 nucleotides~. In c~--p~ison, only the major 1400-nucleotide RNA was seen when RNA from the 35 phoPC strain was hybridized with the org-specific DNA

CA 02237~81 1998-0~-13 W097/18225 PCT~S96/18504-probe, suggesting that the 3800-nucleotide RNA was PhoP
repressed.
A minor RNA of approximately 3800 nucleotides also was detected in long exposure of wild-type RNA blots that 5 were hybridized with either the prgH, prgI-J, or prgR
probes, suggesting possible cotranscription of prgHIJ~
and org. However, both the major (1400 nucleotide) and minor (3800 nucleotide) transcripts were detected when RNA isolated from the prgHl::TnphoA strain was lO hybridized with the org probe, indicating that the prgHl::TnphoA insertion was not polar on either of the org transcripts. Because expression of an org: :lacZY
fusion was shown to be increased approximately fourteen fold in low-oxygen ~ _-red with high-oxygen tension, RNA
15 from wild-type and phopc strains that were grown aerobically or microaerophically to an optical density at 260 nm of 0.5 were compared by blot hybridization with the org-specific DNA probe. No substantial difference was seen in the relative i ~lLs of RNA transcripts 20 detected in wild-type or phoPC strains grown under these conditions. These data indicate that org did not form part of the prgHI3R operon and that it was not regulated by PhoP/PhoQ.
The ~rgI. prqJ and PrqK ~redicted polyPePtides are 2~ 5;m; lar to S. flexneri Mxi and Y. enterocolitica Ysc proteins The sequences of the five predicted polypeptide~
(PrgH, PrgI, PrgJ, PrgK, and Orfl were compared with the protein seguences translated ~rom the GeneBank ~ibrary 30 using BLAST network software. This ~. ~ison revealed similarity between the predicted products o~ prgI, prgJ, and prgR and the ~xiE, MxiI, and MxiJ proteins of S.
~lexneri . Each of the these polypeptide sequences were similar over their entire length, with 65% (PrgI vs.
35 MxiH~, 38~ (PrgJ vs. MxiI), and 46% (PrgK vs. MxiJ) of CA 02237~81 1998-0~-13 WO97/1822~ PCT~S96/18504-positions occupied by identical residues (Figs. 4A-4C).
The prgI and prgK predicted gene products were also similar to the YscF and YscJ proteins, respectively, of Y. enterocolitica, with 28% and 308 of positions occupied 5 by identical residues. The Poisson probabilities were highly significant for each of these c _~isons. No protein similar to the prgH or orfl predicted polypeptides was detected in the protein sequence library.
lO Isolation of ~roteins from S . typh i ~ium culture supe~natants The role of prgHI~ in S. ty~ ~ium protein secretion was analyzed by ~ ;n~tion of the proteins present in cell culture supernatant. Culture media of 15 wild-type bacteria was collected for protein analysis by centrifuging stationary phase cultures at 154,000 x g for l.7 hours. From 6-8 ~g/ml of protein was precipitated by addition of trichloroacetic acid (TCA) to overnight culture supernatants. The TCA-precipitable material in 2 20 ml of supernatant then was f ractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 5). Approximately 25 protein bands, ranging in molecular mass from 18-87 kDa, were detected by Coomassie brilliant blue st~; ni ng.
To rule out the possibility that the supernatant protein bands represented proteins released from lysed cells, soluble and membrane fractions of whole cells and whole cell lysates were c~ p~ed with proteins from the supernatant by SDS-PAGE (Fig. 5). Many of the major 30 polypeptides in the supernatant (e.g., the polypeptide with molecular mass of 87 kDa) were not the major proteins in the other cellular fractions. Conversely, major intracellular soluble proteins and membrane proteins (e.g., the 36 kDa ompC porin) were not detected 35 in the supernatant in this analysis. In addition, CA 02237~81 1998-0~-13 W097/18225 PCT~S96/18504-following centrifugation, the overnight culture media ~rom bacteria expressing alkaline phosphatase (a periplasmic protein) and ~-galactosidase (a cytoplasmic protein) always ContA i n~ less than 9~ and 1%, 5 respectively, of the whole-cell activity of these enzymes. Although some of the supernatant protein bands may represent degradation products of larger protein species, these data indicate that S. typ~imurium was capable of significant protein secretion.
To determine the timing of release of polypeptides in to the supernatant and to test for an effect of PhoP
regulon mutations on secretion, supernatants from CS015, with a null mutation in PhoP (PhoP~), CS022 (PhoPC), and wild-type bacteria (ATCC 14208s) were compared. As shown 15 in Fig. 6, the guantity of protein increased for each strain when supernatant samples taken from mid-log-phase (OD600 = 0.6), late-log/early-stationary-phase (OD600 =
l.l), and stationary-phase (OD600 = 2.2) were ~ ~ed.
However, the pattern of major protein bands detected for 20 each strain was unchanged from mid-log to stationary phase (Fig. 6).
Altered supernatant protein ~rofiles of mutants defective in signaling epithelial cells At each phase of growth ~m; n~, a s;~;l~
25 pattern and guantity of protein was detected in the culture supernatants of PhoP~ strain CS015 and wild-type bacteria (Fig. 6). In contrast, the protein level of 2 ml of phopc strain CS022 supernatant was 24% of wild type levels. At least lO major protein bands seen in the 30 wild-type supernatant were greatly reduced or undetectable by Coomassie blue st~;n;ng of the CS022 supernatant, especially those of higher molecular weight (Fig. 6). In addition, four major protein bands appeared to be increased in amount in CS022 ~. -~ed with 35 wild-type supernatant (31.5 kDa, 30 kDa, 23 kDa, and 20 CA 02237~81 1998-0~-13 wo97ll822s PCT~S96/18504 kDa) (Fig. 6). Although this result could be due to degradation of higher molecular weight polypeptides, these data suggest that the phoPC mutant likely was defective in synthesis or secretion of Ssp.
The defect observed with the phoPC mutant was consistent with prg gene products having a role in protein secretion. Therefore, the Ssp of strains having transposon insertion or deletion of prgHIJR were co~r~ed to wild-type bacteria (ATCC 14028s) by SDS-PAGE. As 10 observed for the phoPC mutant, IB040 (prgNl::TnphoA) and CS451, cont~ining a 10-kb deletion of hil locus (~hil) DNA, each had a pronounced defect in their Ssp profile compared with the wild-type strain (Fig. 6 and 7). IB043 and CS451 culture supernatants contained 100% and 62%, 15 respectively, of wild-type protein levels. At least 5 and 11 major protein bands seen in the wild-type supernatant were greatly reduced or undetectable by Coomassie blue st~;n;ng of the IB040 and CS451, respectively. Five protein bands [87 kDa, 65 kDa, and 20 three in the 35-40 kDa range (Fig. 7), two of which run as a doublet under different electrophoretic conditions (Fig. 6)] were undetectable in the supernatants of CS022, IB040, and CS451. These findings indicated that the presence of at least some of the products of the prgNI~K
25 operon were nec~s~y for synthesis or secretion of Ssp.
The defect in bacterial mediated endocytosis associated with prgEl::TnphoA was complemented by a low-copy number plasmid, pWKS~5, cont~; n ing a 5.1-kb fragment including prgNIJR, org, and orfl. Consistent 30 with this observation, the prgNl::TnphoA mutant carrying pWRS~5 (strain IB043) had a supernatant protein profile similar to that of wild type (Fig. 7). Of the five protein bands undetectable or greatly reduced in culture supernatants of prgNl: :TnphoA, each was detected in IB043 35 and three of them were increased in amount (87 kDa, 65 CA 02237~81 1998-0~-13 W097/18225 PCT~S96/18504 kDa, and 35 kDa) compared with wild-type supernatants.
This f; n~; ng demonstrates a correlation between the ability to secrete proteins and induction of epithelial cell bacterial mediated endocytosis.
5 The prgH locus is im~ortant for 5 . tY~ 7 ~ ~ium to induce endoc~tosis by epithelial cells The defect in BME of the prgH1:: TnphoA mutant is complemented by a plasmid cont~; n; ~g 5.l kb of DNA from this region, indicating that the gene or genes disrupted 10 by the prgE1: :TnphoA insertion are important for BME.
Analysis of the DNA sequence of this region identified six potential open reading frames that could be affected by this transposon insertion. As depicted in Fig. l, five of these open reading frames, namely those 15 designated prgH-R are either disrupted (i.e., prgH) or are 3' to the prgHl::TnphoA insertion. The orfl translational start is 884 nucleotides u~t~eam from the Tnph~ insertion and that orfl is predicted to be oppositely transcribed from the prgHIJK operon.
An approximately 2600 nucleotide PhoP-repressed transcript was detected when RNA was hybridized with prgH-, prgI-~-, or prgR-specific DNA probes. In contrast, the pred~ ;n~nt transcripts detected with org was smaller (approximately 1400 nucleotides), was not 25 altered in the prgHl::TnphoA mutant, and was not repressed by PhoP. Primer extension analysis of the potential start site of transcription, the size of the prgHIJR transcript, and the pre~nc~ of a potential transcriptional terminator i ~i~tely downstream of prgR
30 also were consistent with transcription terminating before org.
In addition to the major transcripts of prgHI~K
and org, a minor PhoP-repressed transcript of approximately 3800 nucleotides also was detected in 35 multiple RNA blots hybridized with the org and prgH, ~ =
CA 02237~8l l998-0~-l3 WO97/18225 PCT~S96/18504 prgI-~, or prgR DNA probes. This minor RNA was similar in size to the combined prgHIJR and org open r~A~ i ng frames (i.e., 3731 nucleotides) and, thus, could represent cotranscription of prg~IJ~ and org. However, 5 both the 3800- and 1400-nucleotide transcripts were detected in RNA from the prgHl: :TnphoA mutant suggesting that the 3800-nucleotide RNA did not represent cotranscription of prgHIJR and org. These data indicate that one or more genes in the prgHIJR operon are 10 important to BME of epithelial cells.
A PhoP constitutive mutation repressed the synthesis of approximately 20 prg-encoded cell-associated protein species (Miller et al., J. Bacteriol. 172:2485-2490, 1990, herein incorporated by reference). Although 15 PhoP/PhoQ has been shown to transcriptionally activate pag (Miller et al., Proc. Natl. Ac~d. Sci. USA 86:5054-5058, 1989, herein incorporated by reference; Pulkkinen et al., supra, herein incorporated by reference), the ?ch~ni~ of protein repression by PhoP/PhoQ had not been 20 characterized prior to the present studies. No transcript was detected when RNA from the PhoP
constitutive mutant was probed with prgH-, prgI-~-, or prgR-specific DNA, indicating that the prgHI~K operon was negatively regulated by PhoP/PhoQ at the level of 25 transcription. Thus, PhoP/PhoQ can both activate and repress transcription of virulence genes.
Consistent with the role of one or more of the products of prg~IJK in bacterial mediated endocytosis and possibly in protein secretion, a low-copy plasmid 30 cont~;n;ng 5.1 kb of DNA (IB043), including prg~IJR, org, and orfl, complemented both the bacterial mediated endocytosis defect and the supernatant protein profile defect of the prgHl: :Tnr~o~ mutant. Based upon its similarity to MxiJ and YscJ, which are 35 membrane-associated lipoproteins that are nec~C~y for CA 02237~81 1998-0~-13 WO97/18225 PCT~S96/18504 export and secretion of Ipa and Yops protein respectively, the prgK gene product is most likely to have such a role in bacterial ~ ted endocytosis and protein secretion. Similar to PrgK, PrgH was predicted 5 to be a membrane lipoprotein. However, in contrast to prgI-R, which are similar to plasmid-encoded genes of Shigella and Yersi~ia spp., a prgH DNA probe hybridized to chromosomal DNA but not virulence-plasmid DNA from Shigella spp.
Neither they nor the prgI or prg~ predicted gene products have signal se~l~nc~ or long hydrophobic d.- ~i n~ that suggest their cellular localization.
However, the location of these genes within operons that encode secretion determinants suggests that they may have 15 a role in this process.
The predicted gene products of the prgHIJR operon were found to be similar to gene products required for protein secretion in other bacterial species. An analysis of proteins present in culture supernatants of 20 S. tyrh; ~ium was per~ormed. These experiments revealed that the supernatants of wild-type cultures contained a large 1ll h~r of protein bands, whereas strains with mutations affecting the prgH locus, including prg~l: :TnphoA, ~hil and phoPC were each defective in 25 protein secretion as assessed by Ssp profiles. This analysis suggested that PhoP/PhoQ could control protein secretion, at least in part, by repressing prgHIJ~ whose products could form part of a secretion machinery.
Furthermore, the finding that phoPC and ~hil mutants were 30 associated with greater defects in their Ssp profile compared with the prg~l: :Tnpho~ mutant suggested that more than one ~Gh~n;~ may be involved in protein secretion and that gene products encoded by the lO kb region that is deleted in the hil mutant also contribute 35 to secretion of Ssp.

CA 02237~81 1998-0~-13 wo97ll822s PCT~S96/18504 Since the strains with altered Ssp profiles were each impaired in signaling epithelial cells, these data suggest that Ssp are involved in signaling such cells to initiate BME. The finding that five Ssp were missing 5 from culture supernatants of the prgN mutant suggested that one or more of these proteins were specifically involved in BME, e.g., Ssp and/or prgNI~R gene products may form a structure on the surface Of 5~ tyrhi ~ium which induces bacterial mediated endocytosis.
S. ty~hi ~ium strains with transposons inserted between prgN and spa that result in reduced bacterial mediated endocytosis were also missing a subset of the Ssp missing from the prgNIJR mutant. DNA sequence analysis of the regions fl~nk;ng the transposon 15 insertions revealed deduced protein sequences that were similar to IpaB and IpaD of S. flexneri. These data suggest that the transposon insertions define an operon in S. ty~hi ~ium that encodes Ipa homologues.

Example 2: Salmonella tyFhi r-7~ium Secreted Invasion 20 Determinants Two Sal~onella typh; ~ium secreted protein (Ssp) mutants with transposon insertions located between spa~
and prgH were identified. One mutant lacks the 87 kDa Ssp, while the other lacks Ssp of 87, 42, and 36 kDa.
25 The invasiveness of these mutants implicates the 42 and 36 kDa Ssp, but not the 87 kDa Ssp in invasion. DNA
sequencing of this region identified two complete and two partial open reading frames (designated sspB, sspC, sspD, and sspA).
The deduced amino acid sequences of sspBCDA are homologous to Shigella flexner} secreted proteins IpaB, IpaC, IpaD, and IpaA. Complementation analyses and amino-te, ~ n~1 sequencing showed that sspc and sspA
encode the 42 kDa and the 87 kDa Ssp and that both CA 02237~81 1998-0~-13 Wo97/1822~ PCT~S96/18504-proteins are secreted without amino-terminal processing.
SspA is abundantly secreted by wild type bacteria but is completely retained within the cellular fraction of a mutant in prg~IJR encoding part of the Ssp secretion 5 apparatus. A precipitate cont~;n;ng SspC and three major Ssp of 63, 59, and 22 kDa was isolated from culture supernatants of wild type bacteria. These data indicate that major secreted invasion deteL ;nAntS of S.
ty~h;m~7~ium are structurally and functionally homologous lO to S. f l exneri Ipa proteins.
The following reagents and experimental procedures were used to characterize Ssp.
Construction of plasmids and strains:
To construct pCH002, pW8-l was cut with EcoRI, 15 the ll kb fragment eluted from a 1% agarose gel and cloned into the EcoRI site of pWSK29. In pC~002, transcription of sspCDA is driven ~rom the l ac promoter.
pCH004 was constructed by cloning the 3 kb RA '~T fragment from pCH002 into the BamHI site of pWSK29. pCH005 20 contains the 4 kb EcoRI-PvuII fragment from pCH002 cloned into EcoRI-HincII restricted pWSK29. pCH006 was constructed by restriction of pCH005 with NcoI and religation of the l.7 kb and the 5 kb fragment. The correct orientations of the cloned inserts were confirmed 25 by appropriate restriction analyses.
PCR of a chromosomal fragment of EE638 comprising the 5'-region of Tn51acZY and adjacent DNA was performed in three independent experiments by using primers Oh l (5' CGCGGATCCATTATGGGATGTATCGG 3 ~; SEQ ID NO: 25) and 30 OL2 (5' CCGGCAG~AAA~GTTGCAG 3r; SEQ ID NO: 26). The 0.8 kb amplified DNA fragments were then restricted with Bam~I and cloned into pWSK29 for sequencing. All th~ee sequences were identical.
Strain VB122 (hilA::kan-112) was constructed as 35 follows: the mutation was originally constructed on a CA 02237~81 1998-0~-13 WO 97/1822~; PCT/US96/18504 -plasmid by inserting a kan cassette (Pharmacia Biotech, Piscateway, NJ) in a EincII site in the 5' region of the hilA coding sequence. The plasmid-encoded hilA::~an-112 mutation was recombined into the chromosome, and the 5 chromosomal mutation was confirmed by PCR analysis.
Mutant EE633 (lacZY4) was isolated by screening for oxygen regulated gene fusions created by random Tn51acZY insertions in S. typhimurium W 114 (hilA::kan-114) and further selection for insertions 10 linked to a hilA::kan-114 by P22 transduction into S.
ty~hi~rium SL1344 and selection for TetR and KanR.
Media and growth conditions for bacterial cultures:
Bacteria were grown in LB broth at 37~C. If 15 necessary, selection was carried out using 50 ~g/ml ampicillin, 10 ~g/ml tetracycline, or 25 ~g/ml kanamycin.
Preparation and analvsis of 5. ty~hi 7rium supernatant prot~i n R
Bacterial cultures were grown for 16 to 17 hours 20 in 12 ml LB in 1.5 x 14 cm glass tubes at 37~C on a TC-7 roller (New Brunswick, Edison, NJ) at 50 rev./min.
Soluble proteins from culture supernatants were obtained as described above. Precipitates in the culture were retrieved, rinsed 5 times with 1 ml H20, dissolved in 25 sample buffer (4% SDS, 12% glycerol, 5%
~-mercaptoethanol, 0.05 M Tris-HCl pH 6.8, 0.01%
bromphenol blue), and resolved in 10% polyacrylamide gels using SDS-PAGE and a Tris-Tricine buffer.
Immunoblotting:
Whole cell samples were prepared from overnight cultures using st~n~d methods with the additional step of ~iltering the culture through a Whatman 1 qualitative paper filter (Whatman International, Maidstone, ~ent, England) before centrifugation. The proteins were 35 resolved by SDS-PAGE and transferred to nitrocellulose by CA 02237~8l l998-0~-l3 wos7/l822s PCT~S96/l8504 electroblotting using a conventional trans~er buffer.
Western blots were incubated with polyclonal rabbit serum prepared against the 87 kDa Ssp. The immunogen was purified by SDS-PAGE and injected into New Zealand White rabbits (Charles River, Wilmington, MA). Serum was collected after two booster injections and subse~uently absorbed with an acetone powder prepared from S.
tyrhim~J~ium strain EE63. Horseradish peroxidase-labelled goat anti-rabbit antibodies were used to label the lO primary antibodies and were visualized using ~h~ ~ luminescence (ECL, Amersham, International, Buckingh - h ire, England) Invasion assays:
Invasion of HEp-2 epithelial cells was carried out 15 according to the method of Behlau et al. ( J. Bacteriol.
175:4475-4484, 1993). To ;n; ;ze epithelial cell detachment from the bottom of the assay wells after bacterial uptake, the following modifications were introduced: invasion time was reduced from 9O to 60 min 20 and gentamicin treatment was performed for 15 min with lOO ~g/ml gentamicin, conditions which were shown to kill 99% of a bacterial culture of 2 x 1O8 cells/ml.
~-terminAl ~rotein seauencinq:
Proteins separated by SDS-PAGE were blotted on 25 PVDF membranes ~Bio-Rad, Hercules, CA) and stained with Ponceau-S. Blotted proteins were sequenced using an ABI
470A protein sequencer with 120A PTH-AA analyzer.

WO97/18225 PCT~S96/18504-Table 7: Strains and plasmids used in this example Bacterial strain Marker E. coli DH5~ F-~80dlacZ~M15~(lacZYA-argF)U169endAl recAlhsdR17(rK~, 5 mK+)deoRthi-lsupE44AgyrA96relA1 S. typhimurium SL1344 wild type W 114 hil::kan-114 VB122 KanR, hilA::kan-112 EE637 TetR, invF: :lacZY11-5 10 EE633 TetR, sspA: :lacZY4 EE638 TetR, sspC::lacZY11-6 S. typhimurium wild type (ATCC14028s) CS451 14028sAhil::Tn5-428 15 CS022 pho-24 (PhoPC) lB04 prgH1:: TnphoA
Plasmid Marker pWSK29 AmpR
p W8-1 TetR
20 p W 71 AmpR
pCH002 AmpR, sspCDA, hilA
pCH004 AmpR, sspC
pCH005 AmpR, sspCD
pCH006 AmpR, sspD
25 Identification of S. typh;m~rium Mutants with TransPoson Insertions in Genes Encodinq Ss~
To identify genes encoding Ssp, Tn51acZY mutants of S. tyFhi rium SL1344 with transposon insertions located within the 40 kb ~Ivirulence island'l (59-60 min~
30 of the S. typhimurium chromosome) were analyzed for changed patterns in Ssp. An insertion in invF
(invF::lacZY11-5), the first gene of the inv-spa operon, and a hilA::~an-122 insertion in VB122 led to major defects in the pattern of Ssp similar to a mutation in 35 the prgHlJ~ operon (prg~l::TnphoA) which has implicated in S. tyrhi~n~rium protein secretion (see Example 1).
Specifically, all three mutants lack 5 major Ssp of 36, 38, 42, 63 and 87 kDa, while the hllA::kan-112 insertion leads to loss of some lower molecular weight protein 40 bands in addition to these 5 Ssp (Fig. 8, lanes 4, 5, 6).

CA 02237~8l l998-0~-l3 WO97/18225 PCT~S96/18504-Two other mutants exhibited detectable loss of only one and of three Ssp, respectively. The supernatants from the mutant strain EE633 contA;n;ng the fusion lacZY4 were missing a protein of 87 kDa, while supernatants from the 5 mutant strain EE638, cont~;n;ng fusion 7acZY11-6, were missing protein species of 87, 42 and 36 kDa. In addition, supernatants from EE638 showed an increased abundance of a 63 kDa Ssp (Fig. 8, lanes 2, 3). Tn51acZY
in EE638 maps approximately 2.5 kb downstream from spaT
lO while in EF633 the transposon maps ~.5 kb downstream from spaT as determined by Southern hybridization and PCR
analyses. Both transposons were inserted in the same orientation (Fig. 9). A degenerate pool of oligonucleotides synthesized according to the sequence of 15 the 12 amino-terminal amino acids of the 87 kDa protein (VTSVRTQPPVIM; SEQ ID NO: 27), hybridized specifically to a 5.5 kb BamHI fragment in p W71 which comprises sequences between hil A and spaT (Fig. 9). These data indicate that the 87 kDa Ssp is encoded in the 20 chromosomal region adjacent to the transposon insertions.
Tn51acZY in ~E633 is likely to be directly within the gene encoding the 87 kDa Ssp, while Tn51acZY in EE638 is likely to be inserted within one of the genes encoding the 42 and the 36 kDa ssp having a polar effect on the 25 synthesis of the other two Ssp missing in supernatants of this mutant.
Secretion of the 87 Ssp kDa Ssp is De~endent on prgHIJK
Since it was possible that the absence of the 87 kDa Ssp (Ssp87) in supernatants of EE633 and EE638 was 30 due to impaired secretion rather than expression, whole cell lysates and supernatants of various strains were analyzed by immunoblotting with antiserum raised against partially purified Ssp87. Fig. 11 shows that Ssp87 of wild type S. ty~hi ~ium is found mainly in the 35 supernatant, although some of the protein is detected in CA 02237~81 1998-0~-13 the cellular fraction (Fig. 11, lane 1). In contrast to wild type bacteria, all of the protein is found in the cell~ ~ fraction of the prgHl: :Tnpho~ mutant IB040 (lane 3). Ssp87 could not be detected in the cellular 5 fractions nor in supernatants of various invasion and secretion mutants: CS022 (PhoPC), a mutant which constitutively represses PhoP regulated genes (Miller et al., J. Bacteriol . 172:2485-2490, 1990, hereby incorporated by reference) (lane 2), CS451 lO (~hi7: :Tn5-428) carrying a 10 kb chromosomal deletion of the hil locus between 59 and 60 min. (lane 4), EE638 (7acZY11-6) (lane 5), and EE633 (lacZY4) (lane 6). The signal at 51 kDa in the supernatant fraction of wild type bacteria might represent a degradation product of Ssp87, 15 while the faint band at 34 kDa is likely nonspecific hybridization since it is present in all bacteria analyzed. These results demonstrate that lack of Ssp87 in supernatants of EE633 and EE638 is due to impaired expression while lack of Ssp87 in supernatants of IB040 (prg~l: :TnphoA) is caused by impaired secretion of the protein. These results further show that expression of the gene encoding Ssp87 is affected by the hil deletion and that expression of Ssp87, either directly or indirectly, is repressed by PhoP.
25 Strain ~638. but not EE633. Is Markedly Deficient in Invasion To determine the function of the 87, 42, and 36 kDa Ssp in invasion of epithelial cells, the ability of strain EE638 and EE633 to invade HEp-2 cells was 30 analyzed. EE638 showed more than a 100-fold reduction in invasiveness when c~ _~ed to wild type bacteria, while EE633 exhibited invasion levels comparable to wild type bacteria ~Fig. 9). These results suggested that the 36 and/or the 42 kDa Ssp but not the 87 kDa Ssp are required 35 for epithelial cell invasion. In addition, observation CA 02237~8l l998-0~-l3 WO97/18225 PCT~S96/18504 of interactions between these mutants and PtK2 cells by time-lapse videomicroscopy indicated that the ability of EE638 to induce epithelial cell membrane ruffling is also markedly reduced, while EE633 in~llcr~ localized membrane 5 ruffles at a fre~uency similar to wild type S.
ty~hi~7~ium ~
The Tn51acZY Insertions in EE638 and EE633 Define a Chromosomal Reqion Encodinq S5p 5. ty~h; rium Homologues of the Shiaella ipaBCDA Operon To determine the gene(s) affected by the transposon insertions in EE638 and EE633, part of a 11 kb EcoR1 subclone of p W8-1 was se~uenced. Two complete and two partial open reading frames (ORFs), positioned in the same transcriptional direction, were identified (Fig. 9).
15 The deduced gene products of the complete ORFs exhibit similarity to Shigella secreted proteins IpaC and IpaD
(31% identity, 47% similarity; 37% identity, 56%
similarity) respectively, and therefore were designated ssp~ and sspD (see Fig. 13 and Fig. 14). The gene 20 products of the complete open reading frames were designated sspC and sspD. The amino acid sequence derived from the 5' -end of sspC was identical to the amino-terminal sequence of the 42 k~a Ssp (underlined in Fig. 13). The deduced gene product of the partial ORF
25 located im ~~;~tely upstream from sspC shows 47% identity (67% similarity) to the carboxyte. ;n~l portion of S.
flexneri secreted protein IpaB and was designated sspB
(Fig. 12). The ORF starting i ~;ately downstream of sspD was designated sspA. The amino acid se~uence 30 deduced from the 5' end of an ORF starting immediately downstream from sspD did not exhibit similarity to IpaA.
~owever, DNA sequencing of internal parts of the gene predicted that the protein encoded by this gene, designated sspA, is similar to IpaA. Nevertheless, the 35 se~uence of amino acids 2-13 (underlined in Fig. 15) was CA 02237~81 1998-0~-13 WO97/1822s PCT~S96/18sO4~

identical to the amino - teL ~ n~ 1 sequence of the 87 kDa Ssp (see above). sspB, sspC, sspD, and sspA are ~ separated by 27, 70, and 15 bp, respectively, and putative ribosome b;n~;ng sites precede sspC, sspD, and 5 sspA .
The amino acid similarities of Ssp to Ipas do not extend over the entire lengths of the proteins. The similarities between SspC/IpaC and SspD/IpaD are highest in the carboxy-terminal regions, while the central parts lO of SspB and IpaB are conserved (see Fig. 12, 13, and 14).
These similarities could reflect conservation in regions of the proteins required for secretion and/or invasion.
Although both SspC and SspD appear to be secreted by the same mech~n; ~m, no obvious similarities or motifs common 15 to these proteins were detected, thus implying conformational rather than sequential features in the secretion of proteins by type III secretion pathways.
The precise insertion of Tn51acZY in EE638 was determined by cloning and sequencing of a PCR product 20 comprising the 5' region of the transposon and upstream chromosomal sequences and was shown to be located 189 bp downstream from the ATG start codon of sspC . The order of the ssp genes and the Ssp profile of EE638 indicate that the transposon insertion in sspC is polar on 25 expression of sspD and sspA and that these genes are likely to be organized in a singly transcribed unit.
Both sspC and sspD are Necessary for S. ty~h; ~ium Tnvasion of Epithelial Cells A complementation analysis was carried out to 30 determine the m;n; ~1 fragment necessary for complementation of the epithelial cell invasion defect of EE638 (sspC: :lacZY11-6) as well as for reconstitution of Ssp. All analyzed fragments were cloned downstream from the lac promoter in the 6-8 copies/chromosome vector 35 pWSK29. As shown in Fig. lO, a 3.9 kb EcoRI - PvuII

CA 02237~81 1998-0~-13 WO97/18225 PCT~S96/~8504 fragment comprising sspC and sspD in pCH005 was sufficient to complement the invasion defect of EE638 to wild type levels. When analyzed for Ssp, EE638 tpCH005]
showed a pattern of Ssp similar to the wild type strain [pWSK2g] except ~or the missing 87 kDa protein (SspA) (Fig. 16, lane 4). EE638 transformed with pCH002 carrying an 11 kb EcoR1 fragment was partially complemented for invasion as well as ~or all 3 missing Ssp (Fig. 10 and Fig. 16, lane 6). In contrast, EE638 10 transformed with plasmids that contained either sspC or sspD alone (pCH004 and pCH006, respectively) were not complemented for invasion but showed reconstitution of the 42 kDa Ssp (SspC) or the 36 kDa Ssp (SspD), respectively (Fig. 10 and Fig. 16, lanes 3 and 5). In 15 addition, the ablln~ncy of a 63 kDa Ssp, which was found to be more abundant in supernatants of EE638, was reduced in supernatan~s of strains EE638 [pCH005], EE638 [pCH006], and EE638 [pCH002] and of SL1344 [pCH002].
These results demonstrate that both SspC and Ss pD are 20 necessary for invasion of epithelial cells and indicate that SspC encodes the 42 kDa Ssp while the 36 kDa Ssp is likely to be encoded by SspD. In addition, complementation of the invasion defect of EE638 with pC~005 indicates that invasiveness is not influenced by 25 the observed changes in the ab~ ncy of the 63 kDa Ssp.
A Preci~itate Found in S. typhimurium Culture Su~ernatants Contains Hiqhly Abundant Ss~C and Other Prot~i n c Supernatants from S. ty~hi ~ium wild type 30 cultures contained a precipitate that, when solubilized in reducing SDS sample buffer, separates into at least ~our highly abundant protein bands of 63, 59, 42 and 22 kDa on SDS-PAGE (see Fig. 17, lane 1). Protein precipitates were also found in culture supernatants of 35 EE638 and EE633, but not in supernatants of S.

CA 02237~8l l998-0~-l3 WO97/18225 PCT~S96/18504 ty~i~77rium mutants with global defects in protein secretion [CS022 (PhoPC), IB040 (prgH: :TnphoA~, CS451 (~hil::Tn5-428) and VB122 (hilA: :kan - 112) . S.
tyrhi~7~}um 14028s, the wild type parent of CS022 and ~ 5 IB040, showed the same protein pattern of precipitated material as SL1344]. The precipitate from EE633 cultures showed a similar composition to that of wild type precipitate by SDS-PAGE analysis. In contrast, a major protein band of 42 kDa was absent from the precipitate 10 isolated from cultures of EE638 (Fig. 17, lane 2).
Amino-terminal sequencing of this 42 kDa Ssp identified it as encoded by sspc. The identity of the amino-terminal protein sequence (MLISNVGINPAAYLN; SEQ TD
NO: 28) with the amino acid sequence derived from the 15 5'-region of SspC (Fig. 13) shows that no amino - teL ;n~l processing of SspC occurs prior to its release into the supernatant.
SDS-PAGE analyses of precipitated material from culture supernatants of EE638 [pCH004 (SspC) ] and EE638 [pCH005 (SspCD) ] showed a pattern similar to wild type [pWSK29] material (Fig. 17, lane 3 and 4), confirming that the respective plasmids complemented the mutant for secretion of SspC. Protein patterns of soluble Ssp and precipitates isolated from untransformed cultures of 25 SL1344 or EE638 were identical to those shown in Fig. 16, 17, lane 1 and 2, respectively. Precipitate of EE638 [pCH006 (SspD) ] was found to be similar to precipitate from EE638 [pWSK29~ except for reduced abundancy of a 63 kDa protein band (Fig. 17, lane 5). The precipitate from 30 EE638 [pCH002 (SspCDA) ] con~n~ an additional major protein band of approximately 51 kDa, which was also present in precipitate from SL1344 ~pCH002] (Fig. 17, lanes 6, 7). Comparison of precipitated proteins to soluble Ssp on SDS-PAGE (Fig. 17, lanes 8, 9) showed that 35 SspC in the precipitate has the same electrophoretic CA 02237~81 1998-0~-13 WO97/18225 PCT~S96/18sO4 mobility as the 42 kDa soluble Ssp. These data suggest that the 42 kDa soluble Ssp is identical to precipitated SspC .
SspC and SspA are secreted proteins of 42 and 87 5 kDa, as demonstrated by amino-te r ; n~ 1 sequencing and by complementation analyses. It is further likely that the 36 kDa protein encoded by SspD is secreted, since lack of a 36 kDa ssp in supernatants of EE638 (lacZY11-6) was complemented by transformation of this mutant with lO plasmids containing SspD. The 63 kDa Ssp is the protein likely to be encoded by SspB.
SspA, SspB, SspC, and SspD appear to be targets of the inv-spa-prgEI~K encoded secretion apparatus, since these Ssp are missing in supernatants of mutants 15 affecting expression or regulation of inv-spa and prgHI~K
(Fig. 8). Typical for proteins secreted by type III
secretion pathways, no amino-ter ; n~l processing of SspA
and SspC was observed. The dependency of Ssp secretion on prgHIJ~ was further proven by demonstrating that SspA
20 is abundantly secreted by wild type cells, while it is completely re~ine~ in the cellular fraction of the prgH1 ::TnphoA mutant IBO40 (Fig. ll). The 38 kDa Ssp of the ~ive major Ssp dependent on the inv-spa-prgHIJR secretion apparatus may be the product of the invJ invasion locus.
The immunoblot analysis of SspA secretion suggests that expression of the gene encoding SspA is negatively controlled by the virulence two component regulatory system PhoP/PhoQ. PhoP/Q has a global ef~ect on protein secretion which is partially due to negative 30 transcriptional regulation of prgHI~R (see Example l).
The SspBCDA genes are located between the large inv-spa gene cluster and prgHI~ at 59 minutes on the S.
ty~hi ~ium chromosome. Fig. 18 shows the relative positions of the invasion genes in S. ty~hi ~ium in 35 c. _A~ison to their S. flexneri homologues, which are CA 02237~81 1998-0~-13 WO97/1822~ PCT~S96/18504-clustered in a 3l kb region of a large virulence plasmid.
The invasion genes cluster in three groups (inv-spa/mxi-spa, Ssp/ipa, and prgIJK/mxlHIJ) which exhibit conserved gene structure and organization, 5 suggesting that these genes were acquired by horizontal gene transfer. Acquisition by horizontal gene transfer is further supported by the fact that these S.
typhi ~ium invasion genes are within a 40 kb "virulence island" which, despite the otherwise high overall genetic lO similarity between S. typhimurium and E. coli K-12, is unique to S. typhimurium. However, the three invasion gene clusters from S. flexneri and S. typhi ~ium are in different relative positions to each other and are interspersed between non-homologous genes, thus implying 15 multi-recombinational events in the evolution of these genetic regions.
In addition to soluble Ssp the ~upernatants of S.
ty~him77~ium cultures cont~;ne~ a flocculent precipitate consisting of SspC and three other major protein species 20 of 63 (Ssp 63), 59 (Ssp 59), and 22 (Ssp 22) kDa. The combination and abundancy of Ssp in the precipitate from S ~ tyrh i ~ium cultures is strikingly different from that in the soluble fraction (see Fig. 17). Though Ssp, including SspC, are found in both the precipitate and the 25 soluble fraction, SspD, even when overproduced, was not detected in the precipitate. This : h~izes the difference in composition of precipitate and soluble fraction and supports the possibility of specific protein-protein interactions between the four Ssp leading 30 to precipitate formation.

OTHER EMBODIMENTS
Using reagents derived from partial cDNA clones of an Ssp, e.g., SspA, the isolation of a full-length cDNA
encoding the Ssp is well within the skill of those CA 02237~8l l998-0~-l3 W097/18225 PcT~s96/l8so4 skilled in the art of molecular biology. For example, a radiolabelled probe made from a known partial cDNA
sequence can be used to identify and isolate from a library of recombinant plasmids cDNAs that contain 5 regions with identical to the previously isolated cDNAs.
The screening of cDNA libraries with radiolabelled cDNA
probes is routine in the art of molecular biology (see Sambrook et al., 1989, Molecular Cloning: a Laboratory MAn77~7, second edition., Cold Spring Harbor Press, Cold lO Spring Harbor, N.Y). The cDNA can be isolated and subcloned into a plasmid vector, and the plasmid DNA
purified by st~n~Ard techniques. The cDNA insert is sequenced using the dideoxy chain termination method well known in the art (Sambrook et al, supra).
15 Oligonucleotide primers corresponding to bordering vector regions as well as primers prepared from previously isolated cDNA clones can be employed to progressively determine the se~uence of the entire gene.
Similar methods can be used to isolate Ssp which 20 are related to SspA, SspB, SspC, or SspD. To isolate related Ssp, a probe having a sequence derived from (or identical to) all or a portion of SspA, SspB, SspC, or SspD can be used to screen a library of Salmonella DNA
(or cDNA). DNA encoding a related Ssp will generally 25 hybridize at greater stringincy than DNA encoding other proteins. This approach can be used to identify Salmonella ty~h; ~ium Ssp as well as Ssp of other Salmonella.
Generation of Monoclonal Antibodies:
Monoclonal antibodies can be generated to purified native or recombinant gene products, e.g., Ssp, by st~n~rd procedures, e.g., those described in Coligan et al., eds., Current Protocols in Immunology, 1992, Greene Publishing Associates and Wiley-Interscience). To 35 generate monoclonal antibodies, a mouse is ; ized with CA 02237~81 1998-0~-13 WO97/18225 PCT~S96/18504 the recombinant protein, and antibody-secreting B cells isolated and immortalized with a non-secretory myeloma cell fusion partner. Hybridomas are then screened for production of specific ant; ho~; es and cloned to obtain a - 5 homogenous cell population which produces a monoclonal antibody. For example, hybridomas secreting the desired antibodies can be screened by ELISA. Specificities of the monoclonal antibodies can be determined by the use of different ~rotein or peptide antigens in the ELISA.
lO Useful quantities of antibodies can be produced by either the generation of ascites fluid in mice or by large scale in vitro culture of the cloned antibody-producing hybridoma cell line. Antibodies can be purified by various chromatographic procedures known in the art, such 15 as affinity chromatography on either immobilized Protein A or Protein G.
The invention also includes DNA encoding other Ssp (e.g., Ssp 54, Ssp 42, and Ssp 22) found in cell supernatants. Those skilled in the art can readily clone 20 the corresonding genes based on the amino terminal sequence or the corresponding protein. The amino tel ; n~1 sequence of Ssp54 is MNNLTL~X~XKVG (SEQ ID NO:
29). The amino te~ ;n~l sequence of Ssp42 is MLISNVGINPAAYLN (SEQ ID NO: 30). The amino terminal 25 sequence o~ Ssp 22 is TKITLSPQNFFI (SEQ ID NO: 31).

Claims

-- 58 --1. Substantially pure DNA encoding a Salmonella secreted protein (Ssp).

2. The DNA of claim 1, wherein said DNA comprises the SspB gene.

3. The DNA of claim 2, wherein said DNA comprises the DNA sequence of SEQ ID NO: 1 or degenerate variants thereof encoding the amino acid sequence of SEQ ID NO: 5.

4. The DNA of claim 1, wherein said DNA comprises the SspC gene.

5. The DNA of claim 4, wherein said DNA comprises the DNA sequence of SEQ ID NO: 2 or degenerate variants thereof encoding the amino acid sequence of SEQ ID NO: 6.

6. The DNA of claim 1, wherein said DNA comprises the SspD gene.

7. The DNA of claim 6, wherein said DNA comprises the DNA sequence of SEQ ID NO: 3 or degenerate variants thereof encoding the amino acid sequence of SEQ ID NO: 7.

8. The DNA of claim 1, wherein said DNA comprises the SspA gene.

9. The DNA of claim 8, wherein said DNA comprises the DNA sequence of SEQ ID NO: 4, or degenerate variants thereof encoding the amino acid sequence of SEQ ID NO: 8.

10. The DNA of claim 1, wherein said DNA
comprises the SspB gene, the SspC gene, the SspD gene and the SspA gene.

11. The DNA of claim 10, wherein said DNA
comprises the DNA sequence of SEQ ID NO: 15.

12. The DNA of claim 1, wherein said DNA
comprises the SspH gene.

13. The DNA of claim 12, wherein said DNA
comprises the DNA sequence of SEQ ID NO: 13, or degenerate variants thereof encoding the amino acid sequence of SEQ ID NO: 14.

14. The DNA of claim 1, wherein said DNA
comprises the stpA gene.

15. The DNA of claim 14, wherein said DNA
comprises the DNA sequence of SEQ ID NO: 10 or degenerate variants thereof encoding the amino acid sequence of SEQ
ID NO: 12.

16. A cell which contains the DNA of claim 1.

17. A method of inducing uptake of a bacterial cell by an epithelial cell in a mammal, comprising increasing expression of the DNA of claim 4 or 6 in said cell and administering said cell to said mammal.

18. The method of claim 17, wherein said bacterial cell is a Salmonella cell.

19. A method of inducing uptake of a bacterial cell by a macrophage in a mammal, comprising decreasing expression of the DNA of claim 4 or 6 and administering said cell to said mammal.

20. A substantially pure SspC polypeptide.

21. The polypeptide of claim 20, comprising an amino acid sequence substantially identical to the amino acid sequence of SEQ ID N0: 6.

22. An active fragment of the polypeptide of claim 21.

23. A substantially pure SspD polypeptide.

24. The polypeptide of claim 23, comprising an amino acid sequence substantially identical to the amino acid sequence of SEQ ID N0: 7.

25. An active fragment of the polypeptide of claim 24.

26. A substantially pure SspH polypeptide.

27. The polypeptide of claim 26, comprising an amino acid sequence substantially identical to the amino acid sequence of SEQ ID N0: 14.

28. An active fragment of the polypeptide of claim 27.

29. A substantially pure IagB polypeptide.

30. The polypeptide of claim 29, comprising an amino acid sequence substantially identical to the amino acid sequence of SEQ ID NO: 11.

31. An active fragment of the polypeptide of claim 40.

32. An antibody which binds to a Ssp.

33. A method of detecting a Salmonella infection in a mammal comprising contacting a biological sample derived from said mammal with the antibody of claim 32 and detecting the binding of said antibody to a Ssp in said sample, wherein said binding indicates that said mammal is infected with Salmonella.

34. A method of detecting the presence of Salmonella in a biological sample comprising contacting said sample with a Ssp-encoding DNA under high stringency conditions and detecting the hybridization of said DNA to nucleic acid in said sample, wherein hybridization indicates the presence of Salmonella in said biological sample.

35. A method of targeting an antigen to an epithelial cell in a mammal, comprising linking said antigen to an Ssp or active fragment thereof to produce a Ssp chimeric antigen and administering said chimeric antigen to said mammal.

36. The method of claim 35, wherein said Ssp is SspC or SspD.

37. A method of inducing a cytotoxic T cell immune response in a mammal, comprising linking said antigen to an Ssp or active fragment thereof to produce a Ssp chimeric antigen and contacting an antigen-presenting cell with said chimeric antigen.

38. A vaccine comprising a bacterial cell the virulence of which is attenuated by decreased secretion of a Ssp.

39. The vaccine of claim 38, wherein said bacterial cell is a Salmonella typhimurium cell.

40. The vaccine of claim 39, wherein said bacterial cell is a Salmonella enteriditis cell.

41. The vaccine of claim 38, wherein said bacterial cell is a Salmonella typhi cell.

42. A live Salmonella cell in which a gene encoding a heterologous antigen is inserted into a Ssp-encoding gene.

43. A method of vaccinating an animal against a Salmonella infection comprising administering the vaccine of claim 38.

44. A substantially pure StpA polypeptide.

45. A method of dephosphorylating a protein, comprising contacting said protein with the polypeptide of claim 44 or an active fragment thereof.