Harry L. T. Mobley is a Professor in the Department of
Microbiology and Immunology at the University of Maryland School of
Medicine in Baltimore.
Links to Other ASM Pages:
- Virulence of the Two Primary
Uropathogens
Escherichia coli and Proteus mirabilis exhibit distinct mechanisms of pathogenesis when causing urinary tract infections
Harry L. T. Mobley
The urinary tract is among the most common sites of bacterial infection. Individuals at high risk for symptomatic urinary tract infection (UTI) include neonates, preschool girls, sexually active women, and elderly women and men. In 1991, the last year for which a large data base is available for the U.S. population, UTIs were the cause of 9.6 million physician visits and were noted in 1.5 million hospital case records. This high attack frequency places UTIs first among kidney and urologic diseases in terms of total cost, exceeding that of chronic renal failure with its attendant dialysis and transplant expenses.
Escherichia coli is by far the most common cause of community-acquired infection in those with normal urinary tracts. Proteus mirabilis, on the other hand, is not a common cause of UTI in the normal urinary tract, but is found frequently in individuals with structurally abnormal urinary tracts or with devices such as an indwelling urinary catheter. To a lesser extent, other microorganisms also cause UTIs, including Staphylococcus saprophyticus, S. epidermidis, Pseudomonas aeruginosa, Providencia stuartii, Morganella morganii, and Klebsiella, Enterococcus, Enterobacter, Citrobacter, Serratia, and Candida spp.
E. coli Is the Most Common Agent of UTIs
E. coli is the predominant cause of uncomplicated cystitis, or bladder infections, and acute pyelonephritis, in which one or both kidneys also become infected. Of women going to physicians for a UTI, 95% do so for symptoms of cystitis. About 40% of adult women will suffer symptoms of cystitis during their lifetime, and E. coli is likely to be identified as the etiologic agent in 75–80% of these cases. Women, who have shorter urethras, are 15 times more likely to acquire a UTI than men.
Although less frequent, acute pyelonephritis, the more serious UTI, is also commonly caused by E. coli. This infection in one or both kidneys usually results from ascent of organisms from the bladder via the ureter, and is distinguished from other UTIs clinically and pathologically. Patients with acute pyelonephritis typically present with the triad of fever, flank pain, and bacteriuria with or without profuse perspiration, rigors, abdominal or groin pain, and nausea and vomiting.
Laboratory tests typically show leukocytosis, bacteriuria, pyuria, diminished renal concentrating ability, and elevated C-reactive protein. On pathological examination, wedge-shaped areas of inflammation containing predominantly polymorphonuclear leukocytes (PMNs) extend from papillae to cortex; tubules are filled with PMNs; and necrosis of proximal tubular epithelial cells is evident. However, glomeruli tend to be spared, even in areas of intense inflammation. Localized inflammation may coalesce to form a renal abscess and additional complications, particularly when stones obstruct the urinary outflow. The infection may spread beyond the urinary tract if bacteria enter the bloodstream directly.
Virulent strains of E. coli that cause such infections display specific phenotypic traits and represent a heterogenous group of isolates, restricted to a small number of O-serogroups with different phenotypes. Isolates typically carry large blocks of genes, called pathogenicity-associated islands (PAI), not found in fecal isolates.
Although many UTI isolates appear to be clonal, the phenotypic profile that causes UTI apparently involves several subclasses of uropathogenic E. coli. Specific adhesins including P fimbriae and type 1 fimbriae appear to aid in colonization. Moreover, these strains produce
several toxins, including the pore-forming hemolysin, cytotoxic necrotizing factor, and an autotransported protease, Sat. Completion of the sequencing of the chromosome of E. coli CFT073 in 2000 by Fred Blattner, Rodney Welch, and their colleagues at the University of Wisconsin, Madison, and efforts by other investigators to identify virulence genes by signature-tagged mutagenesis and other methods should help toward developing a comprehensive model of pathogenesis for uropathogenic E. coli.
Specific UTI-Associated E. coli Phenotypes
Certain E. coli phenotypes are found more frequently in urine of patients with acute pyelonephritis than in urine from patients with cystitis or asymptomatic bacteriuria or in feces of normal individuals (Fig. 1). E. coli from a small number of O-serogroups (six O-groups cause three-fourths of UTIs) have phenotypes epidemiologically associated with acute pyelonephritis in the normal urinary tract; typical traits among such strains include P fimbriae, hemolysin, aerobactin, serum resistance, and encapsulation.
For example, about 80% of E. coli strains from patients with pyelonephritis produce P fimbriae, six times that of fecal strains. From studying the phenotypes of virulent uropathogenic strains and commensal fecal strains, experts agree that E. coli strains causing cystitis and acute pyelonephritis are genetically distinct from the "base model" fecal strain represented by E. coli K-12.
E. coli strains that cause pyelonephritis typically express P fimbriae, a trait that long has been thought to explain the virulence of these strains in the urinary tract. For instance, this cell surface structure, which confers the adherence phenotype, seems to explain the affinity of these strains for uroepithelium. Although the genetics of the pap operon are now well understood and numerous epidemiologic studies implicate P fimbriae and hemolysin in the pathogenesis of acute pyelonephritis, there only two studies in which true isogenic P fimbrial mutants of virulent clinical isolates were constructed to test this explanation in animal models of infection.
For example, in my laboratory, we inoculated the bladders of CBA mice with strain CFT073 (a virulent wild-type strain) or its isogenic P fimbria-negative mutant. After one week, we could not detect any substantive differences in organism concentration or histological findings between parent and mutant in urine, bladder, or kidney at any of five bacterial inoculant concentrations. Thus, adherence by P fimbriae of uropathogenic E. coli strain CFT073 plays only a subtle role in the development of acute pyelonephritis in the CBA mouse model. Similarly, James Roberts of the Tulane Regional Primate Research Center, in Covington, La., and his collaborators subjected monkeys to infections with a papG mutant of E. coli that fails to bind the a -galb 1–4-gal-receptor on epithelial cells. Although this mutant shows a modest reduction in virulence compared to normal E. coli, the effect is not so dramatic to implicate P fimbria as the sole virulence determinant.
Nevertheless, P fimbrial genes (pap) that are found preferentially in urovirulent strains encode an adhesin, and it likely contributes to the ability of E. coli to colonize the urinary tract. These genes also may serve as a marker for additional virulence genes.
Pathogenicity-Associated Islands
Investigators are now finding many examples of clustered virulence factors, virulence cassettes, or pathogenicity-associated islands (PAI) in various pathogenic bacterial species. For example, strains of enteropathogenic, enterohemorrhagic, and meningitis-associated E. coli, Citrobacter freundii, Helicobacter pylori, Salmonella typhimurium, and Yersinia pestis all possess clustered virulence genes not found in less virulent strains within these species.
Jörg Hacker and his colleagues at the University of Würzburg in Würzburg, Germany, discovered and coined the phrase for such a PAI in uropathogenic E. coli 536. Other uropathogenic strains of E. coli also contain PAIs consisting of large stretches of DNA encoding genes not found in the genome of fecal E. coli. Moreover, three strains—536, J96, and CFT073—each contain multiple PAIs (Fig. 2).
In general, these PAIs vary from 25 kb to 190 kb, carry virulence genes, are inserted near or within tRNA genes, contain insertion sequences, and have a G+C content that differs from that in the rest of the chromosome. My lab identified a distinct PAI in highly virulent pyelonephritogenic E. coli CFT073 and determined its precise insertion point within the genome. Genes encoding P fimbriae, hemolysin, and iron uptake are found within this PAI, but other nearby genes were not previously identified in E. coli, and, in some cases, have no recognized homologs.
We also analyzed members within our large collection of pyelonephritis-, cystitis-, catheter-associated bacteriuria-causing, and also fecal strains. This analysis shows that PAI sequences are widespread among urovirulent isolates but only occasionally found in fecal isolates.
Secreted Proteins from E. coli Are Part of UTI Pathogenesis
Secreted proteins often are key elements in pathogenesis for a variety of bacteria, including Yersinia, Salmonella, Shigella, and enteropathogenic and enterohemorrhagic E. coli. All of them use a type III secretion pathway to release proteins that subvert normal host cell functions. For example, enteropathogenic E. coli secrete proteins that elicit cytoskeletal rearrangements in target host cells. Secreted proteins typically are encoded on pathogenicity islands as blocks of contiguous genes.
Uropathogenic E. coli use a more recently described "autotransporter" system to secrete proteins. In this system, proteins are translocated across the inner membrane via the sec system and across the outer membrane through a b -barrel porin structure formed by the carboxyterminus autotransporter domain.
For example, Debra Guyer in my laboratory identified a 107-kDa protein, designated Sat (secreted autotransporter toxin), that is expressed significantly more often in E. coli strains associated with the clinical syndromes of acute pyelonephritis than in fecal strains. The protein isolated from E. coli CFT073 closely resembles proteins, called SPATE (serine protease autotransporter of Enterobacteriaceae), that are produced by diarrheagenic E. coli. These proteins are serine proteases and represent a new class of cytotoxins, according to James Nataro and Ian Henderson at the University of Maryland School of Medicine, Baltimore.
Type 1 Fimbriae Appear Essential for Colonizing the Bladder
Type 1 fimbriae, which are a common feature of most E. coli, appear essential for bladder colonization and virulence in the urinary tract. Thus, E. coli expressing type 1 fimbriae, whether it be phase variants or expression from a plasmid, are more likely to colonize the bladder than are mutants lacking type 1 fimbriae. Moreover, type 1 fimbriae bind directly to uroplakin receptors on bladder epithelial cells and trigger exfoliation, according to Scott Hultgren and his colleagues at Washington University School of Medicine in St. Louis, Mo. However, such fimbriae apparently do not play a role in the development of acute pyelonephritis.
Expression of type 1 fimbriae undergoes phase variation. Transcription of the fimbrial genes is controlled by a promoter that is carried on a 314-bp invertible element. In one orientation, the promoter allows transcription but, once inverted, blocks transcription. Using a PCR-based assay to measure the percentage of an infecting bacterial population with this element oriented to permit transcription, Nereus Gunther and colleagues at the University of Maryland School of Medicine in Baltimore tracked the expression of type 1 fimbriae during the course of a urinary tract infection. They then collected urine from mice infected with an inoculum in the off orientation at various times to determine how the element was oriented within the bacterial population.
Four hours after challenge with a pyelonephritis strain, 12% of the bacteria in urine had turned the switch on; after 24 hours, that proportion of the population increased to 32% and was maintained at that level at 48 hours, before being turned completely off by 96 hours. In contrast, when a cystitis strain was used, 85% of the bacterial population had turned the switch on by 24 hours; the portion of the population that was on remained very high throughout the remainder of the experiment. These results suggest that expression of type 1 fimbriae is more critical in the early stages of UTI and may be required for colonization of the bladder, particularly by cystitis strains.
Soloman Langermann and colleagues at MedImmune, Inc. in Gaithersburg, Md., are developing and evaluating a vaccine that uses the type 1 fimbrial adhesin protein (FimH), which binds to mannosylated receptors in the urinary tract, to protect against UTIs. This vaccine was successful in protecting mice and primates against such infections and is being readied for clinical trials.
Proteus mirabilis Also Causes UTIs but Favors "Complicated" Urinary Tracts
P. mirabilis, named for the Greek god Proteus, who changed shape to avoid detection, is a member of the Enterobacteriaceae. It can appear as a typical fimbriated bacterial cell with a few polar flagella when cultured in broth, or it can differentiate into highly elongated hyperflagellated swarming cells when cultured on agar.
Unlike E. coli, P. mirabilis, which is a motile, urease-positive, lactose-negative, indole-negative, and gram-negative rod, is not a common cause of UTI in hosts whose urinary tracts are normal. Instead, it preferentially infects patients with urinary tracts having functional, anatomical, or other abnormalities, including indwelling urinary catheters. For example, John Warren and his colleagues at the University of Maryland School of Medicine, Baltimore, cultured P. mirabilis at 105 CFU/ml urine from 44% of long-term catheterized patients.
While this microorganism can cause cystitis, it has a propensity for the upper urinary tract, where it can cause severe renal scarring even after a single infection. A hallmark of P. mirabilis infection is the formation of struvite or apatite stones in the kidney; urease, induced by urea, catalyzes production of such stones. While it also causes cystitis and acute pyelonephritis, the hallmark of infection with P. mirabilis is the development of bladder and kidney stones composed of struvite, which consist of magnesium ammonium phosphate, or apatite, which consists of calcium phosphate. Occasionally, these stones completely fill the renal pelvis, forming "staghorn calculi."
Several P. mirabilis proteins play important roles in virulence, as revealed by experiments involving mice with ascending urinary tract infections. When mice are infected with strains carrying mutations for genes encoding either flagella, urease, hemolysin, mannose-resistant Proteus-like fimbriae, Proteus mirabilis fimbriae (PMF), or Zap protease, the symptoms of those urinary tract infections are attenuated to varying degrees. For instance, certain mutations in the flagella genes block motility, thereby reducing the fitness of the microorganism and making it a less effective pathogen. Another mutation that abolishes urease activity prevents the formation of stones in the bladder or kidney of mice following infection, whereas inactivation of MR/P fimbriae genes reduces the numbers of colonizing organisms in both bladder and kidneys.
Urease, MR/P Fimbriae Are Critical Virulence Factors for P. mirabilis
The P. mirabilis version of the enzyme urease consists of three copies of three distinct subunits. The intact enzyme catalyzes hydrolysis of urea, which typically reaches concentrations of 400 mM in urine, releasing ammonia and CO2, and raising the pH of the urine. Thus, ions that are soluble at lower pH values precipitate, sometimes around the bacterial cells, to form infection-induced stones.
P. mirabilis readily triggers stone formation when applied to agar surfaces containing urine (Fig. 3). After overnight incubation, in addition to development of an objectionable odor, struvite crystals form within the agar itself, mimicking stone formation in vivo. Urea induces the bacterial urease genes through a positive transcriptional activator, UreR, that acts on a multigene operon. In addition to the three genes that encode the enzyme itself, four other genes within this operon encode polypeptides that chaperone nickel ions into the active site of the apoenzyme. Indeed, two nickel ions are coordinated into each of three active sites in the enzyme and are required for catalysis. Urease is one of only a very few nickel-metalloenzymes found in nature.
The bacterial urease also is induced in mice that are infected with P. mirabilis. When those infecting bacteria carry a version of urease whose gene is fused with that encoding the green fluorescent protein, fluorescent fimbriated bacteria are readily observed in both the bladder and kidneys of infected mice (Fig. 4A). Stones that are removed from the bladders of such mice typically harbor hundreds of fluorescing P. mirabilis bacteria (Fig. 4B).
Meanwhile, MR/P fimbriae, encoded by the mrp genes, appear to represent another critical virulence determinant for P. mirabilis. An invertable element, similar to that which controls expression of E. coli type 1 fimbriae, controls the mrp genes. Based on a PCR assay, that switch and, thus, MR/P fimbrial gene expression are always on during an infection, suggesting there is selective pressure to retain expression of MR/P fimbriae in the host. Consistent with this finding, we readily observe the short fimbriated form of the bacterium by confocal microscopy of samples from the bladder, kidney pelvis, proximal tubules, and even capillaries of the kidney parenchyma. The elongated swarmer cell has not yet been observed and thus its role in the host is not known.
Assembly of MR/P fimbriae appears to follow many of the conventions of the model of fimbrial biogenesis proposed by Scott Hultgren at Washington University, St. Louis, Mo., for P fimbriae of E. coli. The adhesin of the MR/P fimbriae, MrpH, is present at the tip of the fimbrial shaft. During infection, there is no detectable immune response against this protein, which is present in only vanishingly small amounts. However, when the MrpH adhesin protein is overexpressed, purified, and presented with mucosal adjuvant in an intranasally administered vaccine, mice are protected against homologous transurethral challenge with P. mirabilis. This protection gives some optimism for the development of a vaccine to prevent stone formation and complicated or catheter-associated UTI in a well-defined patient population.
ACKNOWLEDGMENTS
I thank my colleagues John Warren, Mike Donnenberg, Dave Johnson, Bob Belas, Rich Hebel, Jim Kaper, and Jim Nataro for maintaining a stimulating environment in which to carry out our work. I also thank Chris Coker, Nereus Gunther, Deb Guyer, Susan Heimer, Carrie Poore, Xin Li, Jill Phillips, Hui Zhao and all past members of my lab for their contributions to UTI-related projects. Thanks to Deb Guyer for preparation of Fig. 1.
This work was supported in part by Public Health Service Grants A123328, AI43363, A107540, and DK49720 from the National Institutes of Health.
SUGGESTED READING
Donnenberg, M. S., and R. A. Welch. 1996. Virulence determinants of uropathogenic Escherichia coli, p.135–174. In H. Mobley and J. Warren (ed.), Urinary tract infections: molecular pathogenesis and clinical management. ASM Press, Washington, D.C.
Guyer, D. M., J.-S. Kao, and H. L. T. Mobley. 1998. Genomic analysis of a pathogenicity island in uropathogenic Escherichia coli CFT073: distribution of homologous sequences among pyelonephritis, cystitis, catheter-associated bacteriuria, and fecal isolates. Infect. Immun. 66: 4411–4417.
Hacker, J., G. Blum-Oehler, B. Janke, G. Nagy, and W. Goebel. 1999. Pathogenicity islands of extraintestinal Escherichia coli, p.59–76. In J. B. Kaper and J. Hacker (ed.), Pathogenicity islands and other mobile virulence elements. ASM Press, Washington, D.C.
Jones, B. D., C. V. Lockatell, D. E. Johnson, J. W. Warren, and H. L. T. Mobley. 1990. Construction of a urease negative mutant of Proteus mirabilis: analysis of virulence in a mouse model of ascending urinary tract infection. Infect. Immun. 58:1120–1123.
Langermann, S., S. Palaszynski, M. Barnhart, G. Auguste, J. S. Pinkner, J. Burlein, P. Barren, S. Koenig, S. Leath, C.H. Jones, and S. J. Hultgren. 1997. Prevention of mucosal Escherichia coli infection by FimH-adhesin-based systemic vaccination. Science 276:607–611.
Lim, J. K., N. W. Gunther IV, H. Zhao, D. E. Johnson, S. K. Keay, and H. L. T. Mobley. 1998. In vivo phase variation of Escherichia coli Type 1 fimbrial genes in women with urinary tract infection. Infect. Immun. 66:3303–3310.
Mobley, H. L. T., K. G. Jarvis, J. P. Elwood, D. I. Whittle, C. V. Lockatell, R. G. Russell, D. E. Johnson, M. S. Donnenberg, and J. W. Warren. 1993. Isogenic P fimbrial mutants of pyelonephritogenic E. coli: the role of adhesin loci in a -gal(1–4) b -gal binding by a wild type strain. Mol. Microbiol. 10:143–155.
Mulvey, M. A., Y. S. Lopez-Boado, C. L. Wilson, R. Roth, W. C. Parks, J. Heuser, and S. J. Hultgren. 1998. Induction and evasion of host defenses by type 1-piliated uropathogenic Escherichia coli. Science 282: 1494–1497.
Svanborg-Eden, C., U. Jodal, L. A. Hanson, U. Lindberg, and A. S. Akerlund. 1976. Variable adherence to normal human urinary-tract epithelial cells of Escherichia coli strains associated with various forms of urinary-tract infection. Lancet 2:490–492.
Warren, J. W., J. H. Tenney, J. M. Hoopes, and H. L. Muncie. 1982. A prospective microbiologic study of bacteriuria in patients with chronic indwelling urethral catheters. J. Infect. Dis. 146:719–723.
Zhao, H., R. B. Thompson, V. Lockatell, D. E. Johnson, and H. L. T. Mobley. 1998. Use of green fluorescent protein to assess urease gene expression by uropathogenic Proteus mirabilis during experimental ascending urinary tract infection. Infect. Immun. 66:330–335.
Email: webmaster@asmusa.org