Selective delivery of antigens by recombinant bacteria.

The means to attenuate Salmonella and to endow such avirulent strains with the ability to express colonization and virulence antigens from other pathogens has achieved considerable progress during the past several years. One can therefore begin to design and construct strains with specificity to a given animal host and to express in a defined way specific colonization and virulence antigens in a manner to stimulate long-lasting immunity to the Salmonella and to the pathogen supplying the genetic information for the colonization and virulence antigens. Since most pathogens colonize on or invade through mucosal surfaces, the use of recombinant bivalent Salmonella vaccine strains to stimulate a mucosal immune response would induce the development of a first line of defense against a diversity of pathogens. Mucosal immunity should therefore reduce contagious spread of many pathogens since the dose to overcome the mucosal immune barrier would be increased to result in a diminished likelihood of infection. The fact that the recombinant Salmonella vaccine strains also induce humoral and cellular immune responses justifies their use for induction of long-lasting immunity. Although considerable progress has been made in targeting antigens to the GALT by use of avirulent Salmonella, a similar strategy for delivery of antigens to the BALT has yet to be discovered and developed. In addition to constituting a system for induction of immunity against a diversity of pathogens, the recombinant avirulent Salmonella system should provide a means to explore parameters of the mucosal immune response. This would include investigation of the location and duration of memory, the age dependence of induction of mucosal immunity, and the means for the possible induction of oral tolerance with regard to either the mucosal or humoral response to an antigen expressed by the recombinant Salmonella. It is also possible to contemplate using the avirulent Salmonella to target expression of various modulators of the immune system such as interleukin-2 and interferon-gamma to the GALT and thus further enhance the immune response. Lastly, one can introduce into avirulent Salmonella strains genes for putative colonization antigens in order to investigate whether induction of an immune response against the putative colonization antigen does or does not interfere with infection. This system, therefore, permits another means to analyze the relative importance of various bacterial surface attributes in conferring pathogenicity to the microbe.

Role of the Salmonella abortusovis virulence plasmid in the infection of BALB/c mice.

Following oral inoculation of BALB/c mice, Salmonella abortusovis strain SS44 was recovered in lower numbers from the Peyer's patches and mesenteric lymph nodes compared with S. typhimurium strain SL1344, whereas splenic infections were equivalent between the two serovars. SS44 was cured of its virulence plasmid or subjected to mutagenesis of the spv genes, and the Spv(-) derivatives were tested for virulence in mice. Plasmid-cured S. abortusovis SU40 retained virulence in BALB/c mice when inoculated intraperitoneally. On the other hand, mice infected orally with SU40 had greatly reduced splenic infection compared to those infected with wild-type SS44. Similar results were obtained after Tn5 insertion mutagenesis of the spvR gene or deletion of the spvABCD locus. These results suggest that in the gut-associated lymphoid tissues S. abortusovis may replicate less than S. typhimurium and that the S. abortusovis virulence plasmid primarily affects systemic infection after oral inoculation but not after intraperitoneal administration in the mouse model.

Hybridization studies with a DNA probe derived from the virulence region of the 60 Mdal plasmid of Salmonella typhimurium.

Plasmid DNA of 68 strains of Salmonella that belonged to 18 serovars and exhibited 48 different plasmid profiles was examined for hybridization with a 32P-labelled DNA probe which consisted of a 3750 base pairs (bp) HindIII-HindIII fragment derived from the virulence region of the 60 megadalton (Mdal) plasmid of Salmonella typhimurium. The 32 Mdal plasmid of S. cholerae-suis, the 50 Mdal plasmid of S. dublin, the 36 Mdal plasmid of S. enteritidis, the 60 Mdal plasmid of S. gallinarum, the 60 Mdal plasmid of S. pullorum, and the 60 Mdal plasmid of S. typhimurium, plasmids that have been associated with virulence, all hybridized with the probe. Digestion of plasmid DNA of these strains with PvuII and hybridization with the probe revealed that the plasmids of strains of all six serovars contained fragments of approximately 2520 and 1520 bp that hybridized with the probe. Similarly, hybridization with BglI digests of DNA of the virulence-associated plasmids of strains of these six serovars showed that all six plasmids contained a fragment of approximately 3690 bp that hybridized with the probe. No other plasmids of these strains nor any plasmids of 12 other Salmonella serovars hybridized with the probe. Chromosomal DNA did not hybridize with the probe. The 60 Mdal plasmids of S. gallinarum and S. pullorum showed similar digestion patterns with restriction endonucleases BglI, BglII and PvuII.

Virulence plasmids of Salmonella typhimurium and other salmonellae.

Related high molecular weight plasmids of several serotypes and species of Salmonella have been associated with virulence in a variety of animal models of infection. The primary virulence plasmid phenotype is in the ability of salmonellae to spread beyond the initial site of infection, the intestines. The mechanism of this plasmid-mediated invasive infection has not been identified, but may be a complex interaction in the host-pathogen relationship. A common region of the salmonella plasmids has been associated with virulence, and specific virulence genes and their products are now being identified; however, much is yet to be accomplished in this field. The combined analysis of pathogenesis and genetics associated with the salmonella virulence plasmids may identify new systems of bacterial virulence and the genetic basis for this virulence.

Genetic and DNA sequence analysis of the Salmonella typhimurium virulence plasmid gene encoding the 28,000-molecular-weight protein.

We have confirmed that the 28,000-molecular-weight (28K) protein encoded by the virA gene of the 90-kilobase Salmonella typhimurium virulence plasmid is a virulence factor. It was previously shown that a Tn5 insertion, vir-22::Tn5, located in the virulence plasmid greatly attenuated virulence for mice and inhibited the production of a 28K protein (P.A. Gulig and R. Curtiss III, Infect. Immun. 56:3262-3271, 1988). Plasmid pYA426 fully complemented vir-22::Tn5 to virulence by increasing splenic infection after oral inoculation and encoded the 28K protein. To identify the virulence gene(s) of pYA426 mutated by vir-22::Tn5, we constructed nested deletions in pYA426 and examined deletion derivatives for their abilities to complement vir-22::Tn5. Only derivatives still producing the 28K protein complemented vir-22::Tn5. Furthermore, the smallest complementing derivative encoded only the 28K protein, as determined by DNA sequence analysis. Therefore, the 28K protein is sufficient for complementation of the attenuating mutation vir-22::Tn5 and must be the virulence factor inhibited by the insertion. We determined the nucleotide sequence of the 1.2-kilobase BamHI-EcoRI fragment encoding the 28K protein and identified the structural gene, virA. A 723-base-pair open reading frame which encodes a peptide with a molecular weight of 27,572 was found.

Cloning and transposon insertion mutagenesis of virulence genes of the 100-kilobase plasmid of Salmonella typhimurium.

We have cloned regions of the 100-kilobase (kb) plasmid, pStSR100, of Salmonella typhimurium SR-11 that confer virulence to plasmid-cured S. typhimurium. Cells carrying recombinant plasmids that conferred virulence were selected by inoculating mice orally with recombinant libraries in virulence plasmid-cured S. typhimurium and harvesting isolates that infected spleens. Three plasmids, pYA401, pYA402, and pYA403, constructed with the cosmid vector pCVD305 conferred wild-type levels of virulence to plasmid-cured S. typhimurium and had a common 14-kb DNA insert sequence. Another recombinant plasmid, pYA422, constructed with the vector pACYC184, conferred to plasmid-cured S. typhimurium a wild-type 50% lethal dose (LD50) level, but mice died more slowly than when infected with wild-type S. typhimurium. Furthermore, pYA422 conferred the ability to cause a higher, but not a wild-type, level of splenic infection on plasmid-cured S. typhimurium. pYA422 had a 3.2-kb insert sequence which mapped to the center of the 14-kb common sequence of the cosmid clones. Transposon Tn5 insertion mutations in pYA403 inhibited virulence to various degrees, and when transduced into the native virulence plasmid of S. typhimurium, these Tn5 insertions decreased virulence to degrees similar to those observed when the Tn5 insertions were present in pYA403. vir-22::Tn5 in pStSR100 greatly lowered infection of spleens relative to unmutagenized virulence plasmid, while vir-26::Tn5 and vir-27::Tn5 lowered splenic infection to lesser degrees. At least three proteins were encoded by pYA403 containing 23 kb of insert sequence and subclone pYA420, containing the 14-kb common insert sequence present in all of the cosmid clones. One of these proteins, with an apparent molecular weight of 28,000, was also encoded by pYA422. The Tn5 insertion that most attenuated virulence, vir-22::Tn5, inhibited synthesis of the 28,000-molecular-weight protein. The vir-22::Tn5 insertion was complemented by recombinant plasmids encoding only the 28,000-molecular-weight protein, suggesting a role of this protein in virulence. However, recombinant plasmids, exemplified by pYA422, that encoded only the 28,000-molecular-weight protein did not confer full virulence.

Coprecipitation of lipopolysaccharide and the 39,000-molecular-weight major outer membrane protein of Haemophilus influenzae type b by lipopolysaccharide-directed monoclonal antibody.

The major outer membrane protein of Haemophilus influenzae type b (Hib) with an apparent molecular weight of 39,000 (39K) was purified from three different Hib strains and was shown to be free from detectable contamination with other proteins. However, these purified 39K protein preparations were found to contain Hib lipopolysaccharide (LPS). Immunization of rats with these 39K protein preparations resulted in the production of antisera containing both 39K protein-directed and LPS-directed antibodies, as determined by Western blot analysis. The reactivity pattern of the LPS-directed serum antibodies with different Hib strains was identical to the reactivity of these Hib strains with a set of monoclonal antibodies (mabs) previously shown to immunoprecipitate the 39K protein in a radioimmunoprecipitation (RIP) system. Examination of the antigenic specificities of the 39K protein-immunoprecipitating mabs by using Western blot analysis showed that these mabs were actually directed against Hib LPS. RIP analysis of 125I-labeled Hib cells and 32P-labeled Hib cells revealed that the 39K protein and LPS existed as a complex in a RIP system, which resulted in the coprecipitation of both antigens by LPS-directed mabs. The interaction between LPS and the 39K protein was highly selective for this protein and did not involve other outer membrane proteins. The LPS/39K protein complex could be reconstituted by mixing purified LPS and purified 39K protein; it could also be reconstituted with 39K protein from one Hib strain and LPS from another Hib strain. These findings have necessitated the reinterpretation of previous studies involving the 39K protein-immunoprecipitating mabs. Of primary importance is the fact that the demonstrated immunoprotective ability of a 39K protein-immunoprecipitating mab (E. J. Hansen, S. M. Robertson, P. A. Gulig, C. F. Frisch, and E. J. Haanes, Lancet i:366-368, 1982) must now be regarded as evidence that antibody directed against Hib LPS can be protective against experimental Hib disease.

Plasmid-associated virulence of Salmonella typhimurium.

We investigated the role of the 100-kilobase (kb) plasmid of Salmonella typhimurium in the virulence of this organism for mice. Three strains, LT2-Z, SR-11, and SL1344, which possessed 100-kb plasmids with identical restriction enzyme digestion profiles, were cured of their respective 100-kb plasmids after Tnmini-tet was used to label plasmids. Curing wild-type virulent strains SR-11 and SL1344 raised peroral 50% lethal doses from 3 x 10(5) and 6 x 10(4) CFU, respectively, to greater than 10(8) CFU. Both wild-type strains had intraperitoneal 50% lethal doses of less than 50 CFU, whereas the intraperitoneal 50% lethal doses for cured SR-11 and SL1344 were less than 50 and 400 CFU, respectively. Reintroduction of the Tnmini-tet-labeled, 100-kb plasmid restored wild-type virulence. Invasion from Peyer's patches to mesenteric lymph nodes and spleens after peroral inoculation was the stage of pathogenesis most affected by curing S. typhimurium of the 100-kb plasmid. Wild-type S. typhimurium replicated in spleens of mice inoculated intravenously to a greater extent than did plasmid-cured derivatives. Wild-type and cured strains equally adhered to and invaded Henle-407, HEp-2, and CHO cells; furthermore, the presence of the 100-kb plasmid was not necessary for replication of S. typhimurium within CHO cells. The 100-kb plasmid had no effect on phagocytosis and killing of S. typhimurium by murine peritoneal macrophages in vitro and in vivo. Similarly, wild-type and plasmid-cured strains were resistant to killing by 90% normal human, rabbit, and guinea pig sera. All wild-type and plasmid-cured S. typhimurium strains possessed complete lipopolysaccharide, as determined by silver staining solubilized cells in sodium dodecyl sulfate-polyacrylamide gels. We have confirmed the role of the 100-kb plasmid of S. typhimurium in virulence, primarily in invasion to mesenteric lymph nodes and spleens after peroral inoculation of mice. Involvement of the 100-kb plasmid in infection of mesenteric lymph nodes and spleens suggests a role for the plasmid in the complex interaction of S. typhimurium with cells of the reticuloendothelial system.

The Salmonella typhimurium virulence plasmid encodes a positive regulator of a plasmid-encoded virulence gene.

The 90-kb virulence plasmid of Salmonella typhimurium is necessary for invasion beyond the Peyer's patches to the mesenteric lymph nodes and spleens of orally inoculated mice. Two Tn5 insertions located on the left side of a previously identified 14-kb virulence region (P. A. Gulig and R. Curtiss III, Infect. Immun. 58:3262-3271, 1988) and mapping 272 bp from each other exhibited opposite effects on splenic infection of mice after oral inoculation. spvR23::Tn5 decreased splenic infection by 1,000-fold, whereas a spv-14::Tn5 mutant outcompeted wild-type S. typhimurium for splenic infection by 27-fold in mice fed mixtures of mutated and wild-type S. typhimurium. spvR23::Tn5 was complemented by a virulence plasmid subclone with an insert sequence encoding only an 891-bp open reading frame specifying a 33,000-molecular-weight protein. The amino acid sequence of this open reading frame had significant homology to members of the LysR family of positive regulatory proteins; thus, the gene was named spvR (salmonella plasmid virulence). To examine the possible regulatory effects of spvR on other virulence genes, we constructed a lacZ operon fusion in a downstream virulence gene, spvB. When spvR subcloned behind the lac promoter was provided on a separate plasmid in trans to the spvB-lacZ operon fusion, transcription of spvB increased 15-fold. spv-14::Tn5, which conferred a competitive advantage to S. typhimurium, increased the expression of a spvR-lacZ operon fusion in cis. spvR is therefore a positive regulator of spvB and an essential virulence gene of S. typhimurium. As opposed to having spvR subcloned behind the lac promoter, the wild-type spvR gene present on the virulence plasmid did not function to positively regulate spvB-lacZ in trans when salmonellae were grown to the log phase in L broth, suggesting that this regulatory system is activated in vivo during infection.