In vivo gene expression analysis identifies genes required for enhanced colonization of the mouse urinary tract by uropathogenic Escherichia coli strain CFT073 dsdA.

Deletional inactivation of the gene encoding d-serine deaminase, dsdA, in uropathogenic Escherichia coli strain CFT073 results in a hypermotile strain with a hypercolonization phenotype in the bladder and kidneys of mice in a model of urinary tract infection (UTI). The in vivo gene expression profiles of CFT073 and CFT073 dsdA were compared by isolating RNA directly from the urine of mice challenged with each strain individually. Hybridization of cDNAs derived from these samples to CFT073-specific microarrays allowed identification of genes that were up- or down-regulated in the dsdA deletion strain during UTI. Up-regulated genes included the known d-serine-responsive gene dsdX, suggesting in vivo intracellular accumulation of d-serine by CFT073 dsdA. Genes encoding F1C fimbriae, both copies of P fimbriae, hemolysin, OmpF, a dipeptide transporter DppA, a heat shock chaperone IbpB, and clusters of open reading frames with unknown functions were also up-regulated. To determine the role of these genes as well as motility in the hypercolonization phenotype, mutants were constructed in the CFT073 dsdA background and tested in competition against the wild type in the murine model of UTI. Strains with deletions of one or both of the two P fimbrial operons, hlyA, fliC, ibpB, c0468, locus c3566 to c3568, or c2485 to c2490 colonized mouse bladders and kidneys at levels indistinguishable from wild type. CFT073 dsdA c2398 and CFT073 dsdA focA maintained a hypercolonization phenotype. A CFT073 dsdA dppA mutant was attenuated 10- to 50-fold in its colonization ability compared to CFT073. Our results support a role for d-serine catabolism and signaling in global virulence gene regulation of uropathogenic E. coli.

Role of the lipopolysaccharide-CD14 complex for the activity of hemolysin from uropathogenic Escherichia coli.

Bacterial pathogens produce a variety of exotoxins, which often become associated with the bacterial outer membrane component lipopolysaccharide (LPS) during their secretion. LPS is a potent proinflammatory mediator; however, it is not known whether LPS contributes to cell signaling induced by those microbial components to which it is attached. This is partly due to the common view that LPS present in bacterial component preparations is an experimental artifact. The Escherichia coli exotoxin hemolysin (Hly) is a known inducer of proinflammatory signaling in epithelial cells, and the signal transduction pathway involves fluctuation of the intracellular-Ca(2+) concentration. Since LPS is known to interact with Hly, we investigated whether it is required as a cofactor for the activity of Hly. We found that the LPS/Hly complex exploits the CD14/LPS-binding protein recognition system to bring Hly to the cell membrane, where intracellular-Ca(2+) signaling is initiated via specific activation of the small GTPase RhoA. Hly-induced Ca(2+) signaling was found to occur independently of the LPS receptor TLR4, suggesting that the role of LPS/CD14 is to deliver Hly to the cell membrane. In contrast, the cytolytic effect triggered by exposure of cells to high Hly concentrations occurs independently of LPS/CD14. Collectively, our data reveal a novel molecular mechanism for toxin delivery in bacterial pathogenesis, where LPS-associated microbial compounds are targeted to the host cell membrane as a consequence of their association with LPS.

Regulation of type 1 fimbriae by unlinked FimB- and FimE-like recombinases in uropathogenic Escherichia coli strain CFT073.

Genomic DNA sequence analysis of the uropathogenic Escherichia coli strain CFT073 revealed that besides the fimB and fimE recombinase genes that control the type 1 pilus fim phase switch, there are three additional fimB- and fimE-like genes: ipuA, ipuB, and ipbA. Alignment of the predicted amino acid sequences showed that the five recombinases range in sequence similarity from 63 to 70%. An epidemiological survey indicates that ipuA and ipuB are present and linked next to the dsdCXA locus in 24 of 67 uropathogenic E. coli strains but are found in only 1 of 15 normal human fecal isolates. The ipbA sequence located next to the betABIT locus was found in 42 of 67 uropathogenic isolates and 8 of 15 of the commensal strains. We show that two of these recombinases, those encoded by ipuA and ipbA, can function at the type 1 pilus fim switch. In a CFT073 deletion mutant lacking all five recombinase genes, recombinant ipuA or ipbA provided in trans inverted the fim element from the off state to the on state. When a fim OFF CFT073 DeltafimBE mutant was used to infect the urinary tracts of mice, a switch to the fim on state was detected within 24 h in bacteria recovered from urine, the bladder, and the kidneys. A fim OFF CFT073 DeltafimBE ipuB ipbA mutant also demonstrated the ability to switch from the fim off state to the on state during mouse infection. CFT073 recombinase mutants derived from isolates in either the fim on or off state showed a reciprocal relationship for motility. Switches from a nonmotile to a motile phenotype and from a fim on to off genotype were observed in fim ON CFT073 DeltafimBE ipuAB ipbA mutants when ipuA or fimB was provided in trans. Together these results indicate that ipuA has fimB-like on-to-off and off-to-on fim switching activity and that ipbA has the ability to switch fim from the off to the on orientation.

Uropathogenic Escherichia coli use d-serine deaminase to modulate infection of the murine urinary tract.

Although once thought to be unique to bacteria, d-amino acids are also produced by mammals. For example, d-serine is excreted in human urine at concentrations ranging from 3.0 to 40 micro g ml-1. An epidemiological survey demonstrated that urine isolates of E. coli are more likely to catabolise d-serine via expression of d-serine deaminase, DsdA than enteric disease isolates. The urosepsis strain, CFT073, and an isogenic dsdA mutant have similar growth kinetics in minimal or complex media. However, relative to the wild type, the dsdA mutant has a pleiomorphic cell shape and a prolonged, 4-6 h lag phase when grown in human urine. This suggests that d-serine catabolism provides a growth advantage in the urinary tract. Unexpectedly, in a direct competition model of urinary tract infection, the dsdA mutant was recovered 300-times more frequently than the wild type in the bladders of mice 48 h after infection. A new model of E. coli uropathogenesis is proposed where growth and gene expression are modulated in response to environmental d-serine levels. In support of this, the CFT073 dsdA mutant is hyperflagellated and more motile than the wild type indicating that intracellular levels of d-serine may directly or indirectly influence the expression of regulons associated with E. coli uropathogenesis.