INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Many virulence functions that contribute to in vivo fitness are encoded on pathogenicity islands, sequences presumed to have been acquired by horizontal transfer as evidenced by their atypical base composition and codon usage (17). We have recently used in vivo expression technology (20, 34, 35) to identify several Salmonella enterica serovar Typhimurium in vivo-induced (ivi) acquired sequences, which are distinct within and between Salmonella serovars and contribute to fitness within the host (10). One of these acquired virulence regions, iviXXII, is composed of a mosaic of Salmonella-specific sequences that is also serovar specific (present in a distinct set of serovars of Salmonella but not in E. coli or several other enteric pathogens). Removal of a 17-kb portion of the iviXXII region conferred a greater than 200-fold defect in virulence in a murine model of typhoid fever (10).

ssrA resides in or near the junction point of this mosaic of Salmonella-specific sequence (10) and is also known to serve as the insertion site for acquired sequences such as the cryptic phage CP4-57 in Escherichia coli (27) and pathogenicity islands in Vibrio cholerae (25, 30) and Dichelobacter nodosus (2), the causative agents of cholera and ovine foot rot, respectively. ssrA encodes a small, stable RNA molecule (~350 nucleotides), tmRNA, which may serve as both a tRNA and an mRNA, tagging partially synthesized proteins for degradation (26). Additionally, tmRNA plays other diverse roles in E. coli including regulating an alternative protease activity (27), modulating DNA binding protein activity (41), and contributing to the growth of -P22 hybrid phage (40, 55). We investigated whether ssrA contributed to the pathogenesis of Salmonella since ssrA mapped within a 17-kb region that is required for pathogenesis (10) and RNA molecules are also known to play a key role in the virulence of many organisms (17, 39, 43).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and phage. All S. enterica serovar Typhimurium strains used in this study were derived from strain ATCC 14028 (CDC 6516-60) and are listed in Table 1. The high-frequency generalized transducing bacteriophage P22 mutant HT105/1, int-201, was used for all transductional crosses (47), and phage-free, phage-sensitive transductants were isolated as previously described (5). The isolations of all ivi fusions used in this study were described previously (19). Bacteriophage P22-3657 and S. enterica serovar Typhimurium LT2 strain MST3357 were kindly provided by Stan Maloy. P22-3657 contains a deletion in gene 9, encoding phage P22 tail protein (15). MST3357 constitutively expresses P22 tail protein, which is necessary for P22-3657 to form plaques.

Nomenclature. The nomenclature is generally as described previously (4, 7, 13). Unless otherwise specified, the mutation designations used in this study are those of Sanderson et al. (45). The nomenclature z--::Tn10 refers to a Tn10 insertion in a "silent" DNA region; the z-- describes the map position of the insertion (44). The nomenclature used for chromosomal rearrangements is as described previously (8, 21, 33, 48).

Media and chemicals. Luria broth (LB) (11) was the laboratory medium used in this study except for -galactosidase assays, for which 2-(N-morpholino)ethanesulfonic acid (MES) buffered to pH 5.5 and supplemented with 0.05 mM Mg²⁺ was used. Unless otherwise specified, the final concentrations of antibiotics were as follows: ampicillin, 50 µg/ml (LB) or 16 µg/ml (MES); kanamycin, 50 µg/ml; and tetracycline, 20 µg/ml. Mitomycin C and o-nitrophenyl--D-galactopyranoside (ONPG) were purchased from Sigma.

Construction of ssrA mutations. (i) Kn^r deletions. Deletions within the ssrA region were constructed by a PCR-based strategy as previously described (23). The resulting mutations have defined deletion endpoints that contain a Kn^r (kanamycin) determinant at the deletion joint point. The endpoints and primers used to generate the deletions were as follows: MT2055, 5'-TTTGATGCGCGGAAACAC-3' and 5'-GCAATGCACCTTGGGTTT-3'; MT2094, 5'-TACTCATCTCCTGAGTGG-3' and 5'-GCAATGCACCTTGGGTTT-3'; MT2095, 5'-GCGAGCGGTAAATCGTTG-3' and 5'-CGGTGTATCAGGCTATCG-3'; MT2096, 5'-GAATGCGCAGTTAACGGC-3' and 5'-CGATAGCCTGATACACCG-3'; and MT2194, 5'-TTTCGGACGCGGGTTCAA-3' and 5'-GCAATGCACCTTGGGTTT-3'.

(ii) Tn10d-Tc insertions. Tn10d-Tc insertions in the ssrA region were isolated using localized mutagenesis to a genetically linked ivi fusion, iviXXII (10). Briefly, MT1219 (iviXXII) was used as a recipient of phage P22 grown on a genomic pool of Tn10d-Tc insertions. Tetracycline-resistant transductants (Tc^r) were selected and screened for loss of lacZ (inherent in ivi fusions) on LB plates containing X-Gal (5-bromo-4-chloro-3-indolyl--D-galactopyranoside). These Lac Tn10d-Tc insertions were pooled, and a P22 lysate grown on these insertions was used to transduce S. enterica serovar Typhimurium harboring a recombinant plasmid containing a 3.4-kb DNA insert of the ssrA region (cloned within pWKS30 [53], Ap^r) to tetracycline resistance. The Tc^r transductants were pooled, and plasmid DNA was isolated and used to electroporate E. coli DH5 pir to tetracycline resistance. These Tc^r Ap^r transformants contain Tn10d-Tc insertions in the 3.4-kb ssrA region. Transposon insertion joint points on the plasmid were determined by sequence analysis using a primer (5'-ACCTTTGGTCACCAACGC-3') which hybridizes to the right end of Tn10d-Tc. To move these Tn10d-Tc^r transposon insertions into the S. enterica serovar Typhimurium chromosome, phage P22 was grown on the strains containing the Tn10d-Tc-bearing plasmids and used to transduce wild-type S. enterica serovar Typhimurium 14028 to tetracycline resistance. Transductants containing a Tn10d-Tc insertion which had recombined into the chromosome were confirmed by screening for Ap^s (loss of pWKS30) and by verifying linkage to the parental ivi fusion, iviXXII.

Competitive index assay in BALB/c mice. Mutant and wild-type cells were grown overnight in LB medium. BALB/c mice were intraperitoneally infected with a total of 10³ cells of a 1:1 ratio of mutant to wild type (MT2057). The mice were sacrificed after 5 days, and the bacterial cells were recovered from the spleen; the competitive index is the ratio of the number of mutant to wild-type bacteria recovered.

Phage P22 assays. (i) EOP. To determine the relative efficiency of plating EOP for ssrA strains, serial dilutions of phage P22 (HT105/1 int-201) were mixed with wild-type and ssrA strains and plated on LB medium containing 0.75% top agar. The number of PFU was determined on both wild-type and ssrA hosts. The EOP was calculated by dividing the number of PFU per milliliter obtained on the ssrA host strain by the number of PFU per milliliter obtained on the wild-type host.

(ii) Phage adsorption. Phage P22 adsorption assays were performed by adding approximately 10⁸ overnight-grown bacteria and 10⁵ phage P22 to an Eppendorf tube containing prewarmed LB medium (37°C) in a total volume of 1 ml. A no-cell control was used to calculate the efficiency of adsorption. After a 15-min incubation, the cells were centrifuged at low speed for 30 s, and then at full speed spin 90 s, after which the supernatant was immediately transferred to a fresh tube. To determine the remaining PFU, supernatants were subjected to titer determination on wild-type S. enterica serovar Typhimurium. The percent adsorption is calculated as follows: [(number of phage PFU recovered from the supernatant of phage without cells  number of PFU recovered from the supernatant of phage with cells)/number of phage PFU recovered from the supernatant of phage without cells] × 100.

(iii) Lysogen formation. The frequency of P22 lysogen formation was determined as follows. Overnight cultures of wild-type and ssrA S. enterica serovar Typhimurium were diluted 1:50 and incubated for 2.5 h at 37°C. At this point, the number of viable cells was determined and 4 × 10⁷ bacteria were mixed with 1.5 × 10⁹ phage P22-3657 (which contains a kanamycin resistance gene in place of the phage mnt gene [15]), for a multiplicity of infection (MOI) of approximately 40, in a total volume of 0.4 ml. Following a 1-h standing incubation at 37°C, the cell-phage mixture was serially diluted and plated on LB plates containing kanamycin to select for lysogens. The percent lysogeny was determined as follows: (number of lysogens/viable-cell count) × 100.

(iv) Lysogen induction. Induction of viable phage from a lysogen was quantified by a single-burst assay in which overnight-grown wild-type and ssrA lysogens were subcultured 1:50 and incubated for 2.5 h at 37°C, at which time an aliquot was removed from the culture (time = 0). Mitomycin C (Sigma) was added to a final concentration of 2 µg/ml, and the cultures were returned to 37°C. Aliquots (1 ml) were removed at various times for the determination of phage production and viable cells. Phage production was determined by removing an aliquot, lysing the cells by vortexing them in chloroform, centrifuging them for 2 min, and transferring the supernatant to a fresh tube. Because phage P22-3657 is defective in tail protein production, P22 phage tails were added to each lysate in vitro by adding 6 µl of P22 tail stock prepared as described previously (37), after which the lysates were incubated overnight at room temperature. PFU were determined by subjecting the lysates to titer determination on strain MST3357, which constitutively expresses P22 gene 9 tail protein.

-Galactosidase assays. To determine -galactosidase levels, a single colony was inoculated into 1 ml of MES medium buffered to pH 5.5 and containing 0.05 mM Mg²⁺ and was grown overnight at 37°C with shaking for 16 h. -Galactosidase activities were determined by the method of Slauch and Silhavy (49). The activity of -galactosidase is reported as 10³ U per optical density at 600 nm unit per milliliter of cell suspension, where units are micromoles of o-nitrophenol (ONP) formed per minute (n = 3 trials; standard deviations were ±10% of the mean).

DNA sequencing and analysis. DNA was sequenced by the dideoxy chain termination method (46). Homology searches were conducted using the FASTA program (Wisconsin Package version 10.0; Genetics Computer Group, Madison, Wis.) and the BLASTN and BLASTX algorithms available at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

ssrA is required for Salmonella virulence. Previously, we have shown that a deletion of 17 kb of Salmonella-specific DNA at min 60 conferred a greater-than-200-fold virulence defect in BALB/c mice (10). To identify the gene(s) responsible, a series of defined deletions throughout the 17-kb region were constructed and tested for the ability to confer defects in virulence in a murine model for typhoid fever. Table 2 shows that removal of 3.4 kb of sequence (MT2094) conferred the same virulence defect as did the 17-kb deletion (MT2055). Sequence analysis of this region (Fig. 1) revealed three genes: ssrA, a predicted open reading frame with DNA homology (55%) to E. coli slpA integrase (27), and a predicted open reading frame of unknown function. Tn10d-Tc transposon insertions were isolated in each of these genes and retested for virulence. Table 2 shows that ssrA1::Tn10d-Tc (MT2121) conferred a virulence defect (70-fold) whereas all other downstream insertion (MT2122 to MT2124) and upstream deletion (MT2095 and MT2096) strains exhibited wild-type virulence. Additionally, MT2194, which contains a deletion of sequences entirely within ssrA, reproduced the 200-fold defect, suggesting that ssrA was responsible for the virulence phenotype (Table 2).

View larger version (11K): [in this window] [in a new window]   FIG. 1.   Genetic organization of the S. enterica serovar Typhimurium ssrA region. Thick black and gray lines represent two distinct Salmonella-specific regions (10); thin black lines represent sequences shared between S. enterica serovar Typhimurium and E. coli (i.e., ssrA and smpB); dashed lines represent the extent of deleted DNA. The enlarged region denotes the material deleted in MT2094; Tn10d-Tc insertions in this region are designated by triangles; the arrows depict the predicted direction of transcription. Salmonella-specific sequences upstream of ssrA encode predicted proteins with similarity to sequences encoded within SPI4, an unlinked S. enterica serovar Typhimurium pathogenicity island (56). Protein similarities include HlyD and LktD, which are involved in the secretion of hemolysin in E. coli and Pasteurella haemolytica, respectively (32% identity over 176 amino acids and 26% identity over 146 amino acids to SPI4 protein D), ABC transport proteins (HlyB and LktB family, involved in the transport of hemolysin) (32% identity over 177 amino acids and 26% identity over 175 amino acids to SPI4 protein R), AggA and PrtF, which are involved in Pseudomonas putida adherence to root tip cells and protease secretion in Erwinia chrysanthemi, respectively (21% identity over 375 amino acids and 18% identical over 401 amino acids to SPI4 protein C).

ssrA is required for expression of ivi genes. The effect of ssrA on ivi gene expression was determined by transducing an ssrA deletion (ssrA2::Kn) into our collection of more than 200 ivi fusion strains (19). Figure 2 shows that ssrA affected the expression of six ivi genes, three of which (vacB, asnS, and thiI) were induced by ssrA and three of which (f304, dedE, and yoaA) were repressed by ssrA. Several of these genes are required for or implicated in virulence, including vacB, which is required for the expression of the invasion plasmid antigen proteins of Shigella flexneri (28); f304, an open reading frame of unknown function which resides within a uropathogenic E. coli pathogenicity island (16); and dedE (cvpA), which is required for the production of the colicin V toxin (14). Other genes under ssrA control include asnS, which encodes asparaginyl-tRNA synthase (57); thiI, which is involved in the production of thiazole, a precursor of thiamine synthesis (54); and yoaA, a homologue of the dinG family of DNA helicases involved in the bacterial SOS response (31).

View larger version (25K): [in this window] [in a new window]   FIG. 2.   ssrA regulation of ivi genes. -Galactosidase activities were determined following overnight incubation in MES minimal medium buffered to pH 5.5 and containing 0.05 mM Mg2+. -Galactosidase activities are given as 103 micromoles of ONP formed per minute per optical density unit at 600 nm per milliliter of cell suspension. Values represent the average of three independent cultures, with standard deviations (±10%) shown as error bars.

Salmonella-specific sequence adjacent to ssrA. Previously, we have shown that ssrA resides near the junction of two different regions of Salmonella-specific sequence (10). To ascertain whether ssrA was at the junction point of either of these unique DNA regions, the ssrA region was sequenced. S. typhimurium ssrA shows 96.1% homology to the E. coli ssrA gene. However, homology to E. coli is lost immediately after the 3' end of ssrA, which contains the P4-like bacteriophage attachment site, attL (27). attL serves as the attachment site for the E. coli cryptic prophage, CP4-57 (27), and pathogenicity islands in Vibrio cholerae (30) and Dichelobacter nodosus (2), the causative agents of cholera and sheep foot rot, respectively. These data suggest that ssrA attL sequences serve as the attachment site for one of the two regions of Salmonella-specific sequences at min 60. Moreover, Salmonella-specific sequence with limited or no homology to E. coli continues for several kilobases 3' to the ssrA attL attachment site. The first 180 nucleotides downstream of attL have no homology to the GenEMBL DNA database, after which limited homology to the E. coli DNA sequence (55%) extends for 1.2 kb (Fig. 1). This coincides with the location of E. coli slpA, which encodes the cryptic phage CP4-57 integrase SlpA.

ssrA is required for support of the growth of phage P22 and for the induction of P22 lysogens. Table 3 shows that S. enterica serovar Typhimurium ssrA hosts are severely restricted (nearly 10,000-fold) in their EOP of phage P22. Moreover, the rare phage plaques produced on an ssrA host appeared to be mutant by two criteria: (i) the plaques produced on an ssrA host have a clear plaque morphology relative to the turbid plaques produced by phage P22 on wild-type S. enterica serovar Typhimurium (data not shown); and (ii) clear-plaque phage showed a wild-type EOP when plated on either ssrA or wild-type hosts (data not shown).

View larger version (12K): [in this window] [in a new window]   FIG. 3.   Effect of ssrA on the induction of phage P22 lysogens. Phage production in wild-type (MT2207) and ssrA (MT2208) strains [represented as the log (PFU/ml) on the y axis] is shown as a function of time after mitomycin C induction of a P22-3657 lysogen. P22-3657 contains a deletion of P22 tail gene 9, and thus single-burst progeny phage in the supernatant were subjected to titer determination on strain MST3375, which constitutively expresses the P22 tail protein. Values represent the average of three independent experiments, where the standard deviation is <20%.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

ssrA encodes a multifunctional RNA molecule that plays many diverse roles in the bacterial cell, ranging from tagging abnormal proteins for degradation to supporting phage growth to modulating the activity of DNA binding proteins. Here we show that ssrA (i) is required for full virulence in S. enterica serovar Typhimurium, (ii) affects the expression of several genes induced during infection, (iii) is the integration site for a large region of Salmonella-specific sequence, and (iv) is required to support the growth of bacteriophage P22 and for the induction of P22 lysogens.

The role of ssrA in Salmonella virulence may be attributed to its known role as a tmRNA that tags aberrant polypeptides for degradation. Keiler et al. (26) proposed that tmRNA initially functions as a tRNA and then as an mRNA, resulting in the release of the parental mRNA template from the ribosome and translocation of the nascent peptide to tmRNA. The subsequent addition of an tmRNA-encoded tag marks the nascent polypeptide for degradation (26). In the macrophage phagosome, reactive oxidants are a major source of DNA and/or protein damage and play a major role in bacterial killing (1). Accordingly, the repair of damaged DNA/or protein has been implicated in Salmonella pathogenesis (3). ssrA may counter, in part, the oxidative damage by preventing the buildup of abnormal proteins which may be toxic to the cell.

Another role of ssrA in Salmonella pathogenesis may be its effect on the expression of virulence functions since RNA molecules are known to control protein production at the transcriptional (e.g., transcriptional silencing [50]) and posttranscriptional (e.g., antisense [12, 22]), anti-antisense [36], and mRNA decay [32, 43] mechanisms) levels. Moreover, small RNA molecules which control the expression of virulence genes and functions have been identified for several pathogens including (i) tRNA molecules implicated for translation of virulence genes within pathogenicity islands of uropathogenic E. coli (17, 42) and type 1 fimbriae in Salmonella spp. (9, 52); (ii) RNAIII, a global Staphylococcus aureus regulator involved in the expression of several extracellular toxins and cell surface proteins (39); and a putative Erwinia carotovora RNA molecule, AepH, an activator of extracellular proteases involved in tissue maceration in plant hosts (38).

Similarly, a subset of the ssrA-regulated genes identified in this work are required for or implicated in virulence: vacB is an RNase R which regulates the production of proteins required for invasion, intracellular dissemination, and macrophage apoptosis in Shigella flexneri (6, 28); f304 is contained within a pathogenicity island in uropathogenic E. coli strain CFT073, and its presence has been correlated with more acute clinical symptoms (16, 24); dedE is required for the production of the toxin colicin V (14); and yoaA encodes a homologue of the dinG family of DNA helicases involved in the bacterial SOS response (31). Additionally, ssrA may affect the expression of other virulence genes in Salmonella.

The integration sites for pathogenicity islands are often within genes encoding small RNA molecules, e.g., tRNA genes (17). Similarly, a portion of the bacteriophage P4 attachment site, attL, resides within the 3' end of ssrA, which serves as the insertion site for acquired sequences in several organisms, including the cryptic phage CP4-57 in E. coli (27), and pathogenicity islands in Vibrio cholerae (25, 30) and Dichelobacter nodosus (2). Here we show that the ssrA attL is also the attachment site for a region of acquired sequence in Salmonella. Additionally, another region of acquired sequence maps 5' to ssrA, integrating between the corresponding ssrA and smpB genes in E. coli. Sequence analysis of the ssrA-proximal junction region did not reveal bacteriophage attachment sites, IS sequences, or sequence with significant repeats, and thus the mechanism of inheritance of this region remains unclear.

The temperate bacteriophage P22 is subject to complex regulatory mechanisms (51). Here we have shown that Salmonella ssrA mutants were deficient in the ability to support the growth of phage P22, as evidenced by a severe reduction in the EOP (nearly 10,000-fold). ssrA mutants also showed a delay in the production of viable P22 phage when induced from a lysogen, which may arise from inefficient phage P22 excision or from an inability to effectively initiate and/or sustain phage replication. Retallack et al. (40) have shown that ssrA mutants impair the growth of hybrid /P22 phage in an E. coli host. However, phage mutants that lack the C1 gene product, which activates transcription of the P22 repressor (c2), produce phage at wild-type levels in E. coli ssrA hosts. These researchers propose a model in which tmRNA directly or indirectly affects the expression of C1-controlled genes/functions required for phage growth. Their model suggests that the lack of tmRNA alters the affinity of a phage-regulatory protein(s) for its cognate DNA binding site (40) and/or leads to a deficiency of aminoacylated tmRNA molecules required for translation of phage mRNA (55). In addition to the defects in phage P22 growth presented here for S. enterica serovar Typhimurium, tmRNA-mediated alterations in DNA-protein interactions and/or translation may contribute to the role of ssrA in bacterial gene expression and in the virulence of Salmonella and other pathogens.