Methods of using Salmonella enterica presenting C. jejuni N-glycan or derivatives thereof

Abstract

The present invention relates to Salmonella enterica comprising at least pgl operon of Campylobacter jejuni or a functional derivative thereof and presenting at least one N-glycan of Campylobacter jejuni or N-glycan derivative thereof on its cell surface and, in particular, to medical uses and pharmaceutical compositions thereof as well as methods for treating and/or preventing Campylobacter and optionally Salmonella infections and methods for producing these Salmonella strains.

Claims

1. A method of using Salmonella enterica, including but not limited to live Salmonella enterica, for preparing a medicament, including but not limited to a vaccine, wherein the Salmonella enterica comprises at least one pgl operon of Campylobacter jejuni and further comprises at least one N-glycan of Campylobacter jejuni on its cell surface, wherein one or more genes for bacillosamine biosynthesis are inactivated by mutation or partial or complete deletion.

2. The method of use of claim 1 for preparing a medicament, including but not limited to a vaccine for the prevention or treatment of Campylobacter jejuni and optionally Salmonella infections.

3. The method of use of claim 2 wherein the infections to be treated are infections in livestock, including but not limited to cattle and poultry.

4. A method for treating or preventing C. jejuni and optionally Salmonella infections, comprising administration of Salmonella enterica, via a delivery medium including but not limited to pharmaceutical composition, food or feed, to a human or animal in need thereof in a physiologically active amount, wherein the Salmonella enterica comprises at least one pgl operon of Campylobacter jejuni and further comprises at lest one N-glycan of Campylobacter jejuni on its cell surface, and wherein one or more genes for bacillosamine biosynthesis are inactivated by mutation or partial or complete deletion.

Description

FIGURES

(1) FIG. 1 is a schematic picture of a C. jejuni N-glycan display on S. enterica sv. Typhimurium. A) shows the transfer of C. jejuni N-glycan to an S. Typhimurium lipid A core in a strain producing O-antigen and featuring the pgl.sub.mut operon (“mut” means PglB is inactivated by 2 point mutations; B) shows an S. Typhimurium ΔwbaP strain without any O-antigen and featuring a pgl3.sub.mut operon in which the genes for bacillosamine biosynthesis are deleted; C) illustrates the deletions in the pgl3.sub.mut operon.

(2) FIG. 2 demonstrates the display of C. jejuni N-glycans on S. enterica sv. Typhimurium A) shows an anti-C. jejuni N-glycan immunoblot of a SDS-PAGE of S. Typhimurium wild type and ΔwbaP proteinase K-treated whole cell extracts of strains carrying the denoted plasmids and demonstrates display of C. jejuni N-glycan on S. Typhimurium lipid A core. B) is a silver-stained SDS-PAGE (left panel) and an anti-Salmonella group B O-antigen immunoblot (right panel) of a SDS-PAGE of S. Typhimurium wild type and ΔwbaP whole cell extracts treated with proteinase K. It confirms the lack of polymeric O-antigen in the ΔwbaP strain. C) shows an anti-C. jejuni N-glycan immunoblot of a SDS-PAGE of S. Typhimurium ΔwbaP strain with an integrated empty vector (control) or an integrated pgl3.sub.mut operon and proves expression of C. jejuni N-glycan on S. Typhimurium ΔwbaP lipid A core with an integrated pgl3.sub.mut operon. D) depicts in the left panel an immunoblot using serum from a mouse infected intravenously with heat-killed S. Typhimurium ΔwbaP displaying C. jejuni N-glycan with GlcNAc at the reducing end and encoded by pgl3.sub.mut. Recognition of C. jejuni wild type but not of C. jejuni 81-176pglB cells is evident. The right panel_shows a Coomassie-stained SDS-PAGE of the samples used in the immunoblot analysis of the mouse sera.

(3) FIG. 3 depicts the in vitro tests used to demonstrate the attenuation of S. Typhimurium ΔwbaP A) shows an increased sensitivity of S. Typhimurium ΔwbaP to complement in human serum: Complement-mediated killing of kanamycin-resistant serovar Typhimurium wild type strain, M939, O-antigen negative ΔwbaP::cat (SKI11) and complemented mutant ΔwbaP::pKI9 (SKI33) was tested by incubating a 1:1:1 mixture of wild type, ΔwbaP and ΔwbaP::pKI9 (SKI33) Salmonella for the indicated time points together with 20% human serum or 20% heat-inactivated human serum. Survival was analysed by plating on differentiating media. B) depicts the the result of the experimental setting of A) but differing in the use of heat-inactivated serum instead. None of the strains are affected in survival. C) illustrates the defect of S. Typhimurium ΔwbaP in swimming motility compared to S. Typhimurium wild type and non-motile strain fliGHI:Tn10.

(4) FIG. 4 demonstrates a reduced colonisation ability for S. Typhimurium ΔwbaP in a co-infection experiment with S. Typhimurium wild type. A) grafically presents the competitive indices (CI; (mutant/wild type) output/(mutant/wild type) input) of serovar Typhimurium ΔwbaP (SKI12) and wild type determined at days 1-3 post infection in feces and at day 4 post infection in the cecal content demonstrating a reduced colonisation ability of S. Typhimurium ΔwbaP when compared to wild type. B) CI in the mLN, spleens and livers at day 4 post infection.

EXAMPLES

Bacterial Strains and Growth Conditions

(5) A summary of bacterial strains used for the experiments listed in the examples is provided in table 1. Bacteria were grown in Luria-Bertani (LB) medium (10 g/l Bacto tryptone, 5 g/l Bacto yeast extract, 5 g/l NaCl). LB agar plates were supplemented with 1.5% (w/v) agar. Antibiotics were used in the following final concentrations: Ampicillin (amp) 100 μg/ml, kanamycin (kan) 50 μg/ml, chloramphenicol (cam) 25 μg/ml, streptomycin (strep) 50 μg/ml, tetracycline (tet) 10 μg/ml.

Example 1—Display of C. jejuni N-glycan on the Salmonella enterica sv Typhimurium Lipid A Core

(6) Wzy-dependent O-antigen biosynthesis and C. jejuni N-glycan biosynthesis are homologous processes (Feldman et al., Proc. Natl. Acad. Sci. USA.; 102(8):3016-21, 2005) which both start with the assembly of an oligosaccharide structure on an undecaprenylpyrophosphate linker. The homology of the two pathways as well as the relaxed substrate specificity of the S. enterica sv. Typhimurium O-antigen ligase WaaL (Fait et al., Microbial Pathogenesis 20:11-30, 1996; De Qui Xu et al., Vaccine 25: 6167-6175, 2007) were explored for the possibility of combining the pathways to display the C. jejuni N-glycan on Salmonella lipid A core.

(7) A plasmid containing the C. jejuni pgl.sub.mut operon with inactivated PglB (pACYCpgl.sub.mut; Wacker et al 2002) was introduced into a Salmonella enterica serovar Typhimurium strain by electroporation. As negative control the corresponding empty vector pACYC184 was used.

(8) The glycoconjugates of the transformants were tested for display of the C. jejuni N-glycan by SDS-PAGE and subsequent immunoblot with an anti-C. jejuni N-glycan antiserum (Amber 2008). Samples were prepared as follows: The equivalent of 2 OD.sub.600/ml of log phase growing cultures of S. enterica sv Typhimurium containing either pACYC184 or pACYpgl.sub.mut was spun down at 16,000 g for 2 min and the supernatant was discarded. Cells were resuspended in 100 μl Lämmli sample buffer (0.065 M Tris-HCl pH 6.8, 2% SDS (w/v), 5% β-Mercaptoethanol (v/v), 10% Glycerin (v/v), 0.05% Bromophenol blue (w/v)) and lysed for 5 min at 95° C. After cooling to room temperature, proteinase K (Gibco/Life Technologies) was added (final concentration 0.4 mg/ml) and incubated 1 h at 60° C. before loading equal amounts on a 15% sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE). To detect C. jejuni N-glycan, a rabbit polyclonal antiserum against the C. jejuni N-glycan was used (S. Amber, PhD.—thesis, ETH Zürich, Department of Biological Science. Zurich, 2008). Visualisation of signals was carried out with goat-anti-rabbit-IgG-HRP conjugate (Santa Cruz) and ECL (Amersham) as recommended by the manufacturer.

(9) C. jejuni N-glycan could be detected on S. enterica sv. Typhimurium lipid A core when pACYCpgl.sub.mut was present in the cells (FIG. 2A lane 2) but not if the empty vector had been introduced into the cells (FIG. 2A lane 1). This shows that S. enterica sv Typhimurium WaaL transfers C. jejuni N-glycan from undecaprenylpyrophosphate to lipid A core.

Example 2—Construction of a wbaP Deletion in Salmonella enterica sv Typhimurium and Increased Display of C. jejuni N-Glycan in the O-Antigen Negative Strain

(10) Deletion of O-antigen biosynthesis was assumed to abolish competition between the O-antigen biosynthesis pathway and biosynthesis of the C. jejuni N-glycan for the lipid carrier undecaprenylphosphate.

(11) Construction of a wbaP deletion mutant of S. Typhimurium wild type SL1344 was carried out as described (Datsenko and Wanner, PNAS USA 97(12): 6640-5, 2000). Primers RfbP H1P1 (for sequence see table 1) and RfbP H2P2 annealing to template DNA from plasmid pKD3, which carries a chloramphenicol-resistance gene flanked by FRT (FLP recognition target) sites were synthesised. These primers also contain 40 to 45 additional nucleotides corresponding to regions directly upstream and downstream of the wbaP gene. They were used to amplify a gene cassette for in frame deletion of wbaP as described (Datsenko and Wanner, see above). After arabinose-induced expression of the λ Red recombinase from plasmid pKD46 in S. Typhimurium wildtype strain SL1344 the recombinase exchanged the target gene with the chloramphenicol cassette of the PCR product introduced by electroporation. Transformants were selected by plating on chloramphenicol plates at 37° C. overnight and presence of the cat gene in the correct position in the genome was confirmed by PCR. The chloramphenicol resistant resulting clone (wbaP::cat) was termed SKI11. Removal of the chloramphenicol resistance cassette was possible by using pCP20 encoding the FLP recombinase recognising the flanking FRT regions and the resulting strain was termed SKI12 after verification by PCR (also see IIg, Endt et al., Inf. Immun., 77, 2568, June 2009).

(12) Phenotypic analysis of the glycoconjugates of the resulting strain was performed by SDS-PAGE followed by subsequent staining of the glycoconjugates by silver. For SDS-PAGE, samples were prepared as follows: The equivalent of 2 OD.sub.600/ml of log phase growing cultures of S. Typhimurium wild type or S. Typhimurium ΔwbaP (SKI12) was spun down at 16,000 g for 2 min and the supernatant was discarded. Cells were resuspended in 100 μl Lämmli sample buffer (0.065 M Tris-HCl pH 6.8, 2% SDS (w/v), 5% R-Mercaptoethanol (v/v), 10% Glycerin (v/v), 0.05% Bromophenol blue (w/v)) and lysed for 5 min at 95° C. After cooling down to room temperature, proteinase K (Gibco/Life Technologies) was added (final concentration 0.4 mg/ml) and incubated 1 h at 60° C. before loading equal amounts on a 12% sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE). To detect S. Typhimurium O-antigen, Salmonella 0 Antiserum Group B factors 1, 4, 5, 12 (Difco) was used. Visualisation of signals was carried out with goat-anti-rabbit-IgG-HRP conjugate (Santa Cruz) and ECL (Amersham) as recommended by the manufacturer. For staining, the method from Tsai and Frasch was used (Tsai and Frasch, Anal. Biochem. 119(1): 115-9, 1982).

(13) The deletion of the gene encoding for the phosphogalactosyltransferase WbaP in S. enterica wild type led to an abolishment of O-antigen biosynthesis as visible in FIG. 2B. SDS-PAGE with subsequent staining of the glycoconjugates by silver as well as an SDS-PAGE followed by an immunoblot with Salmonella group B specific anti-O-antiserum show the typical lipopolysaccharide ladder pattern of the polymeric O-antigen for the S. enterica sv. Typhimurium wild type strain and the absence of this patter in the ΔwbaP strain.

(14) This O-antigen negative S. enterica sv. Typhimurium ΔwbaP SKI12 was tested for its ability to display C. jejuni N-glycan on its cell surface. Plasmids pACYCpglmut or pACYC184 were introduced by electroporation. The glycoconjugates of the transformants were analysed as described in example 1. C. jejuni N-glycan could be detected in higher intensities in the lane containing the ΔwbaP strain compared to wildtype (FIG. 2A lane 4 vs lane 2). No C. jejuni N-glycan could be detected when the empty vector pACYC184 was present in the S. enterica sv. Typhimurium ΔwbaP SKI12. This demonstrates that in the ΔwbaP strain more C. jejuni N-glycan is transferred to lipid A core.

Example 3—Construction of an Altered C. jejuni pglmut Operon Leading to Increased C. jejuni N-Glycan Display on Salmonella enterica sv. Typhimurium

(15) In C. jejuni, the N-glycan is synthesised as the heptasaccharide GalNAc5(Glc)-Bac, where Bac, the sugar at the reducing end, is 2,4-diacetamido-2,4,6-trideoxy-glucopyranose. In E. coli and S. Typhimurium Bac is not synthesised unless the C. jejuni N-glycan biosynthesis machinery is heterologously expressed. It was shown that in E. coli wild type cells co-expressing the C. jejuni N-glycan biosynthesis machinery two different kinds of N-glycan are synthesised, one with Bac at the reducing end and one with GlcNAc. This phenomenon could be attributed to the action of WecA, an UDP-GlcNAc: undecaprenylphosphate GlcNAc-1-phosphate transferase involved in glycolipid biosynthesis (Linton D. et al., Mol. Microbiol., 55(6):1695-703, 2005). As it is known that Salmonella enterica sv Typhimurium O-antigen ligase WaaL can transfer GlcNAc containing structures to lipid A core it was speculated that a GlcNAc-containing N-glycan could be a better substrate for WaaL than a Bac-containing N-glycan. A pgl.sub.mut operon was constructed that was deleted in the genes for bacillosamine biosynthesis, namely pglD,E,F,G. The genes encoding for PglE, F, G were completely deleted while the one encoding for PglD was partially deleted. The pglD open reading frame (ORF) in the altered pgl operon terminates after 270 base pairs while the full length ORF contains 612 base pairs. The procedure to construct this altered pgl.sub.mut operon was carried out using E. coli DH5a as host strain for plasmid propagation and is as follows: pACYCpglmut DNA was digested with Alw44I and SmaI, then the Alw44I overhang was filled in with the DNA polymerase I Klenow fragment and religated. The resulting operon was termed pACYCpgl3mut and was transformed into the ΔwbaP strain. The glycoconjugates of the resulting transformants were analysed as described in example 1. C. jejuni N-glycan could be detected in higher intensities in the lane containing the ΔwbaP strain with the pgl3mut operon than in the lane containing the ΔwbaP strain with the pglmut operon, when compared to wildtype (FIG. 2A lane 5 vs lane 4). All in all, the ΔwbaP strain with the pgl3mut shows the highest intensities when probed with the anti-C. jejuni N-glycan antiserum and therefore demonstrates the highest levels of C. jejuni N-glycan displayed on Salmonella enterica sv Typhimurium lipid A core.

Example 4—Integration of the pgl3mut Operon into the Genome of the O-Antigen Negative Salmonella enterica sv Typhimurium ΔwbaP Strain

(16) For ensuring continuous display of the C. jejuni N-glycan on the Salmonella enterica sv Typhimurium ΔwbaP strain lipid A core in vivo, the pgl3mut operon was integrated into the genome of ΔwbaP strain SKI12 downstream of the pagC gene.

(17) All cloning steps involving a suicide plasmid with the oriR6K were performed in E. coli CC118λpir. The final integrative suicide plasmid pK115 was constructed in the following way: A 512 bp sequence homologous to the target region in the Salmonella genome was amplified by PCR with the primers 3′ PagC Fw NotI and 3′ PagC Rev SacII (sequence see table 1). The resulting DNA fragment was inserted with SacII and NotI into pSB377 and the plasmid was termed pKI14 after verification of the insert sequence. PKI15 was constructed by digesting pACYCpgl3mut DNA with BamHI and EheI while digesting pKI14 with BamHI and SmaI. The 11083 bp fragment cut from pACYCpgl3mut was then ligated with the pKI14 backbone. As electroporation of suicide plasmids into Salmonella strains is very inefficient, pKI15 or pKI14 were first introduced into E. coli Sm10λpir for conjugation by electroporation. Sm10λpir containing pKI15 or pKI14 was then conjugated with SKI12. For conjugation the equivalents of 4 OD600 of late log-phase cultures of Sm10λpir containing pKI15 and SKI12 were spun down and washed three times with 1 ml LB. The pellets were resuspended in 100 μl LB, combined and spread with a diameter of 3 cm onto an LB agar plate that was then incubated over night at 37° C. The bacteria were washed off the plate in the following morning with 1 ml LB and several dilutions plated on LB (+strep+tet) selecting for conjugants. The resulting strains were called SKI34 (SKI12::pK114) and SKI35 (SKI12::pK115).

(18) To test for C. jejuni N-glycan on lipid A core of the O-antigen-negative strains containing either the integrated pgl3mut cluster or the integrated empty vector as a negative control, whole cell extracts of SKI34 and SKI35 were prepared and analysed as described in example 1. FIG. 2C is an immunoblot detected with anti-C. jejuni N-glycan antiserum which shows intense signals in lane 2 containing SKI35 but no signal for lane 1 containing SKI34. This demonstrates efficient transfer of C. jejuni N-glycan to Salmonella enterica sv Typhimurium lipid A core from an integrated pgl3mut operon.

Example 5—Immunogenicity of the Glycan Encoded by the Pgl3mut Operon

(19) In order to investigate the immunogenicity of the pgl3.sub.mut-encoded glycan mice were infected with heat-inactivated bacteria SKI12+pMLpgl3.sub.mut and their sera were tested for anti-C. jejuni N-glycan antibodies. The experiment was carried out as follows:

(20) Mouse Infection Experiments

(21) Salmonella infections were performed in individually ventilated cages at the RCHCI, Zurich, as previously described (Stecher, Hapfelmeier et al., Infection Infect Immun. 2004 July; 72(7):4138-50 2004). For the intravenous infection mice were injected into the tail vein with 5×10.sup.5 CFU of heat-inactivated S. Typhimurium SL1344ΔwbaP (SKI12) carrying pMLBAD (control) or pMLpgl3.sub.mut. After analysis of the sera at day 29 post infection mice were re-injected with the same bacterial strains at day 36 and sera were analysed on day 50.

(22) Analysis of Mouse Sera

(23) Mouse sera were analysed for production of anti-C. jejuni N-glycan antibodies by immunoblot against whole cell extracts of C. jejuni 81-176 and 81-176pglB (negative control). C. jejuni 81-176 μg/B does not produce glycosylated proteins and served as negative control. Whole cell extracts were prepared by harvesting C. jejuni from plates of confluent bacterial growth with 1 ml PBS. After adjusting the samples with PBS to the same optical density cells were collected by centrifugation for 2 min at 16000×g at room temperature. Cells were lysed for 5 min at 95° C. in Lämmli sample buffer (0.065 M Tris-HCl pH 6.8, 2% SDS (w/v), 5% β-Mercaptoethanol (v/v), 10% Glycerin (v/v), 0.05% Bromophenol blue (w/v)) added to the same final volumes as determined before with PBS to give the same amount of cells in each sample. This was confirmed by separating equal volumes of each sample by SDS-PAGE followed by staining of proteins with Coomassie blue. Additionally, glycosylated and unglycosylated protein AcrA was used for visualising an immune response against C. jejuni N-glycan. For analysis of mouse sera equal volumes of the whole cell extracts as well as equal amounts of glycosylated and unglycosylated AcrA were separated by SDS-PAGE followed by transfer of the proteins to a polyvinylidenfluorid membrane for immunoblot detection. The mouse sera served as primary antisera in the first incubation step. Bound IgG were identifed by an anti-mouse-IgG-HRP conjugate (Bethyl Laboratories). Detection was performed with ECL (Amersham) according to the manufacturer.

(24) FIG. 1D) shows the presence of anti-C. jejuni N-glycan-IgG in mouse sera 61 days after re-infection. The antibodies did not recognise unglycosylated AcrA or unglycosylated protein extracts from C. jejuni and thereby prove specificity for the glycan. No C. jejuni N-glycan-specific reaction could be observed with sera of mice infected with the control strains (data not shown).

Example 6—Attenuated Phenotype of S. Typhimurium ΔwbaP

(25) The attenuation of S. Typhimurium ΔwbaP was tested in several in vitro and an in vivo approach. The in vitro approach consisted of testing the mutant as well as the wildtype for their serum resistance, motility and resistance to the antimicrobial peptide mimic polymyxin B. The colonisation ability of ΔwbaP was analysed in an in vivo co-infection experiment.

(26) Analysis of Serum Resistance

(27) Bactericidal activity of complement was tested essentially as described (Bengoechea, Najdenski et al. 2004). In brief, serovar Typhimurium wbaP::cat (SKI11), M939, a kanamycin-resistant derivative of serovar Typhimurium wild type SL1344 strain (aph integrated downstream of sopE) and cells from serovar Typhimurium ΔwbaP::pKI9 (SKI33) taken from exponentially growing cultures were mixed in equal amounts (3×10.sup.8 cfu/ml for M393; 4×10.sup.8 cfu/ml for SKI11 and SKI33) and diluted 5×10.sup.4 fold before use in sterile 1×PBS. This diluted bacterial culture was mixed 1:1 with 20% human serum containing no antibodies against serovar Typhimurium LPS and incubated at 37° C. with slight agitation. Aliquots were taken at 0, 15 and 30 min after mixing and complement activity was quenched by adding Brain Heart Infusion Broth. The aliquots were kept on ice until plating on LB (strep, kan) selecting for wild type, LB (Sm, Cam) selecting for wbaP::cat and LB (Sm, Tet) to determine ΔwbaP::pKI9 CFU. The same experiment was carried out using serum where complement was heat-inactivated at 56° C. for 30 min. Data is shown as means of log CFU±standard deviations. FIG. 3A shows the decreased serum resistance of S. Typhimurium ΔwbaP when compared to wildtype: After 30 min of incubation with 20% human serum the counts for ΔwbaP were sixty times less than at the beginning of the incubation period. FIG. 3B depicts the same strains incubated with heat-inactivated serum as negative control.

(28) Swimming Motility Assay

(29) Because motility of bacteria is a known virulence factor motility of bacteria was tested on soft agar plates (0.3% (w/v) agar, 5 g/l NaCl, 10 g/l Bacto tryptone). 1 μl of overnight cultures of serovar Typhimurium wild type (SL1344), serovar Typhimurium ΔwbaP (SKI12), serovar Typhimurium ΔwbaP::pKI9 (SKI33) or serovar Typhimurium fliGHI::Tn10 (M933) were spotted in the middle of plates and motility was quantified by measuring the diameter of the halo visible after 4.75 h and 9.5 h of incubation at 37° C. Each experiment was carried out in triplicate on two different occasions and data are shown as means±standard deviations. As visible in FIG. 3C motility was strongly decreased in ΔwbaP (SKI12) when compared to wildtype but still higher than in the non-motile control fliGHI::Tn10.

(30) Analysis of Polymyxin B Resistance

(31) The equivalent of 1 OD.sub.600/ml of exponentially growing cultures from serovar Typhimurium wild type SL1344 strain or serovar Typhimurium ΔwbaP (SKI12) was spun down, resuspended in 150 μl cold sterile 1×PBS and diluted 5×10.sup.6 fold before use. For the assay 45 μl of the diluted cultures were mixed with 5 μl of Polymyxin B (Sigma, 1 μg/ml final concentration) or 5 μl PBS and incubated for 1 h at 37° C. under slight agitation. After addition of 80 μl LB bacteria were plated on LB-agar plates containing streptomycin. The survival efficiency was calculated by dividing the CFU (colony forming units) of peptide-treated culture by the CFU of untreated culture multiplied by 100. The assay was performed in triplicate on two independent experiments and data are shown as means±standard deviations. Decreased polymyxin B resistance of S. Typhimurium ΔwbaP compared to wildtype is evidenced in FIG. 3D.

(32) Colonisation Ability of ΔwbaP in a Co-Infection Experiment

(33) The colonisation ability of S. Typhimurium ΔwbaP was tested in a co-infection experiment in which mice were infected intragastrically with the ΔwbaP mutant as well as the wild type strain. C57BU6 mice (SPF; colony of the RCHCI, Zurich) were pretreated by gavage with 20 mg of streptomycin. 24 h later the mice were inoculated with 5×10.sup.7 CFU of serovar Typhimurium strain or mixtures of strains as indicated. Bacterial loads (CFU) in fresh fecal pellets, mesenteric lymph nodes (mLNs), spleen, and cecal content were determined by plating on MacConkey agar plates (50 μg/ml streptomycin) as previously described (Barthel, Hapfelmeier et al. 2003). The competitive indices (CI) were determined according to the formula CI=(mutant/wild type) output/(mutant/wild type) input after plating. A co-infection experiment of serovar Typhimurium wild type (M939) and ΔwbaP strain (SKI11) was performed. 5 streptomycin-treated mice were infected with a 1:2 mixture (total 5×10.sup.7 CFU) intragastrically of the ΔwbaP strain (SKI11) and wild type strain. The ratio of the 2 strains (CI; competitive index, see Materials and Methods) was determined in the feces at day 1, 2 and 3 p.i. A decrease of ΔwbaP counts compared to wild type was detected (one log scale per day) and proved that the ΔwbaP strain (SKI11) had indeed a severe competitive defect in comparison to wild type serovar Typhimurium strain in the intestinal tract (p>0.05; FIG. 4A). Moreover, the CI of the two strains at systemic sites (mLN, liver, spleen) at day 4 p.i. also demonstrated a significant competitive defect of serovar Typhimurium ΔwbaP (SKI12). Nevertheless, the defect was less pronounced than in the intestine (FIG. 4B).

(34) TABLE-US-00001 TABLE 1 Strains, plasmids and primers for wbaP deletion used in this work Genotype and Strain phenotype Source or reference Salmonella enterica sv Typhimurium strains SL1344 wild type; strep.sup.R Hoiseth, S. K. and B. A. Stocker, Nature 291:238-239, 1981 SKI11 SL1344ΔwbaP::cat; this study strep.sup.R, cam.sup.R SKI12 SL1344ΔwbaP; strep.sup.R this study SKI34 SKI12::pKI14; strep.sup.R, this study tet.sup.R SKI35 SKI12::pKI15; strep.sup.R, this study tet.sup.R Escherichia coli strains DH5a SupE44 ΔlacU169 Hanahan, D., J. Mol. Biol., (Φ80lacZΔM15) hsdR17 5,166(4):557-80, 1983 recA1 endA1 gyrA96 thi- 1 relA1 CC118 λpir Δ(ara-leu), araD, Herrero, M., V. de Lorenzo, ΔlacX74, galE, galK, and K. N. Timmis. J Bacteriol phoA20, thi-1, rpsE, 172:6557-6567. rpoB, argE(Am), recA, λpir Sm10λpir thi thr leu tonA lacY Miller, V. L. and J. J. supE recA::RP4 2- Mekalanos. J. Bacteriol. Tc::Mu λpir, kan.sup.R 170:2575-2583, 1988. Plasmids Plasmid Genotype Source or reference pSB377 tet.sup.R oriR6K Mirold et al., Proc. Natl. Acad. Sci. USA, 96:9845-9850, 1999. pKD3 bla FRT cat FRT PS1 Datsenko, K. A. and B. L. PS2 oriR6K Wanner, Proc. Natl. Acad. Sci. USA, 97:6640-6645, 2000. pKD46 bla P.sub.BAD gam bet exo Datsenko, K. A., and B. L. pSC101 oriTS Wanner, Proc. Natl. Acad. Sci. USA, 97:6640-6645, 2000. pCP20 bla cat cI857 λP.sub.R flp Datsenko, K. A., and B. L. pSC101 oriTS Wanner, Proc. Natl. Acad. Sci. USA, 97:6640-6645, 2000 pACYC184 Cm.sup.R, Tc.sup.R, ori p15A New England Biolabs pACYCpgl.sub.mut Cm.sup.R, ori p15A; C. jejuni Science, 298(5599):1790-3, pgl cluster with 29. Nov. 2002 pglB.sup.W458A,D459A cloned in pACYC184 pACYCpgl3.sub.mut Cm.sup.R, on p15A; C. jejuni This study pgl cluster with PglB.sup.W458A,D459A cloned in pACYC184, deletion of pglE, F, G and 3′- half of pglD pKI14 Tet.sup.R, oriR6K, 500 bp This study region 3′ of PagC cloned in pSB377 pKI15 Te.sup.R, oriR6K, C. jejuni This study pgl3mut cluster with pglB.sup.W458A,D459A cloned into pKI15 Primers for wbaP deletion RfbP H1P1 CTTAATATGCCTATTTTATTTACATTATGCAC GGTCAGAGGGTGAGGATTAAGTGTAGGCTGGA GCTGCTTC (SEQ ID NO: 4) RfbP H2P2 GATTTTACGCAGGCTAATTTATACAATTATTA TTCAGTACTTCTCGGTAAGCCATATGAATATC CTCCTTAGTTCCTATTCC (SEQ ID NO: 5) Primers for pgl3.sub.mut Integration 3′ PagC Fw AAGCGGCCGCGCATAAGCTATG CGGAAGGTTC NotI (SEQ ID NO: 6) 3′ PagC Rev ACCGCGGGACACTGAGGTAATA ACATTATACG SacII (SEQ ID NO: 7)