Letters to the editor should be sent to:

placey@asmusa.org

ASM News Article Links:

• The Path to Salmonella

Links to Other ASM Pages:

• Table of Contents
• ASM News Issues

In the article "The Path to Salmonella'' (ASM News, January 2000, p. 21-27), Groisman and Ochman describe the evolution and divergence of Salmonella enterica from its close relative Escherichia coli. The authors succinctly present the contribution of horizontally acquired pathogenicity islands to the pathogenic lifestyle of S. enterica. However, in their description of the separate paths taken by Salmonella and E. coli from their common ancestor, the authors inaccurately include P fimbriae along with metabolic genes as elements that are present in commensal strains of E. coli. Strictly speaking, nearly all uropathogenic E. coli causing uncomplicated urinary tract infections first reside in the intestines of their hosts and become pathogens following a niche switch from the intestines to the urinary tract. However, P fimbriae certainly are not present in most E. coli strains residing among the human normal intestinal flora. Unlike metabolic genes encoding ArgC, GuaA, lactose utilization, or curli and Type 1 fimbrial adhesins, which are present in the majority of both innocuous and pathogenic E. coli strains, genes encoding P fimbriae are present in only certain E. coli strains. For example, among strains of the E. coli reference (ECOR) collection, P fimbrial genes are mostly limited to two phylogenetic clusters (E. F. Boyd and D. L. Hartl, J. Bacteriol. 180:1159-1165, 1998).

Since the pathogenesis of E. coli extraintestinal infections is complex, the direct contribution of P fimbriae to pathogenesis is not as clear as is, for example, the role of the locus of enterocyte effacement (LEE) in the pathogenesis of diarrheal disease due to enteropathogenic or enterohemorrhagic E. coli. However, there is strong evidence that supports a role for P fimbriae in the pathogenesis of extraintestinal disease. P fimbriae are highly associated with E. coli isolates causing pyelonephritis/cystitis in humans compared to commensal E. coli isolates (J. R. Johnson, Clin. Microbiol. Rev. 4:80-128, 1991; A. Siitonen, J. Infect. Dis. 166:1058-1065, 1992), and are associated with E. coli strains causing extraintestinal infections in animals. Animal model studies using isogenic knockout P fimbrial mutants suggest a direct role for P fimbriae in renal (J. A. Roberts et al., Proc. Natl. Acad. Sci. USA 91:11889-11893, 1994) and systemic infections (M. Ngeleka et al., Infect. Immun. 61:836-843, 1993). In fact, as with the five major Salmonella PAIs described by the authors, genes encoding P and related adhesins were likely acquired by horizontal gene transfer to certain subgroups of E. coli. Genes encoding P fimbriae are often linked to other suspected extraintestinal virulence genes such as hly, cnf, and kps, and are located on the pathogenicity islands of uropathogenic E. coli strains 536 (PAIs 1 and 2), J96 (PAIs 4 and 5), and CFT073 (J. Hacker, Mol. Microbiol. 23:1089-1097, 1997; C. M. Dozois and R. Curtiss III, Vet. Res. 30:157-179, 1999).

In addition to being able to promote adherence to urinary tract cells, E. coli P fimbriae were described by Groisman and Ochman as adhesins that increase persistence of E. coli strains in the colon. This is perhaps a misplaced emphasis, since there is little solid published evidence to support this claim. In vitro studies demonstrated that purified P fimbriae or class I type G adhesin bind to surfactant-like particles from small intestinal and colonic cells (G. S. Goetz et al., Infect. Immun. 67:6161-6163, 1999), and E. coli strains producing P and other adhesins adhered to ileal and colonic epithelial cells (I. Adlerberth et al., Microb. Pathog. 18:373-385, 1995; A. E. Wold et al., Infect. Immun. 56:2531-2537, 1988). However, in vitro data and the results of studies involving gnotobiotic rats (M. V. Her|fias et al., Infect. Immun. 65:531-536, 1997; 63:4781-4789, 1995) are of uncertain relevance to actual colonization of the human intestine. Further, observational studies of the prevalence of P fimbriae among E. coli strains from the feces of healthy neonates have yielded conflicting results (K. Tullus et al., J. Hosp. Infect. 11:349-356, 1988; Epidemiol Infect. 108:415-421, 1992). Wold et al. (A. E. Wold et al., J. Infect. Dis. 165:46-52, 1992) reported that resident colonic strains expressed P fimbriae, although the samples were taken from girls that already had bacteriuria. Thus, it is currently unknown whether the persistence of certain E. coli strains in the colonic microflora of ``normal'' hosts is promoted by P fimbriae, and, even if so, whether P fimbriae mediate increased colonization or persistence in the human colon as compared with other E. coli adhesins such as Type 1, S/F1C, and Dr adhesins. Of note, in a cynomolgous monkey model in which P fimbriae were shown to contribute to pyelonephritis (J. A. Roberts et al., Proc. Natl. Acad. Sci. USA 91:11889-11893, 1994), P fimbriae conferred no advantage to colonization of the colon (J. Winberg et al., J. Exp. Med. 182:1695-1702, 1995).

Certainly, in a feature review that focuses primarily on the evolution of Salmonella pathogenesis, the amount of space the authors can devote to explaining the divergence of E. coli and Salmonella is of necessity limited. However, what is stated by Groisman and Ochman regarding E. coli P fimbriae could give a false impression regarding the phylogenetic associations of these E. coli adhesins and their contribution to disease caused by extraintestinal E. coli pathotypes. We deemed it necessary to present our views and clarify these points.

Charles M. Dozois
Washington University
Saint Louis, Mo.

James R. Johnson
VA Medical Center
Minneapolis, Minn.