Hydrophilic domain II of Escherichia coli Dr fimbriae facilitates cell invasion

Uropathogenic and diarrheal Escherichia coli strains expressing adhesins of the Dr family bind to decay-accelerating factor, invade epithelial cells, preferentially infect children and pregnant women, and may be associated with chronic or recurrent infections. Thus far, no fimbrial domain(s) that facilitates cell invasion has been identified. We used alanine scanning mutagenesis to replace selected amino acids in hydrophilic domain II of the structural fimbrial subunit DraE and evaluated recombinant mutant DraE for attachment, invasion, and intracellular compartmentalization. The mutation of amino acids V28, T31, G33, Q34, T36, and P40 of DraE reduced or abolished HeLa cell invasion but did not affect attachment. Electron micrographs showed a stepwise entry and fusion of vacuoles containing Escherichia coli mutants T36A and Q34A or corresponding beads with lysosomes, whereas vacuoles with wild-type Dr adhesin showed no fusion. Mutants T31A and Q34A, which were deficient in invasion, appeared to display a reduced capacity for clustering decay-accelerating factor. Our findings suggest that hydrophilic domain II may be involved in cell entry. These data are consistent with the interpretation that in HeLa cells the binding and invasion phenotypes of Dr fimbriae may be separated.     

P-antigen-recognizing fimbriae from human uropathogenic Escherichia coli strains.

P-antigen-recognizing fimbriae (P fimbriae) from four pyelonephritogenic Escherichia coli strains and type 1 fimbriae from an E. coli strain and a Salmonella typhimurium strain were purified. The P fimbriae were morphologically similar to type 1 fimbriae. The purified P fimbriae agglutinated neuraminidase-treated human P1 and P2k erythrocytes but not p erythrocytes, which lack all P-blood group-specific glycosphingolipids. However, coating of neuraminidase-treated p erythrocytes with globoside rendered such erythrocytes agglutinable by the P fimbriae. The hemagglutinations were in all instances fully inhibited by the synthetic alpha-D-Galp-(1-4)-beta-D-Galp-1-O-Me glycoside. The binding specificity of the P fimbriae could also be demonstrated by using fimbriae coated onto latex particles and nontreated erythrocytes. It was thus concluded that the P fimbriae recognize and bind to the alpha-D-Galp-(1-4)-beta-D-Galp carbohydrate sequence occurring in the series of P-blood group antigen-specific glycosphingolipids. In contrast to both type 1 fimbriae, all four P fimbriae preparations showed multiple bands in sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Antisera were raised in rabbits against the various E. coli fimbriae. In enzyme-linked immunosorbent assays each one of the antisera to the P fimbriae reacted to titers of log 4 to 7 with both the homologous and the heterologous P fimbriae, but not with the type 1 fimbriae of E. coli. In a reciprocal fashion, the antiserum to the type 1 fimbriae of one E. coli strain reacted only with the homologous type 1 but not with any of the P fimbriae preparations.     

Mannose-resistant and eluting haemagglutinins and fimbriae in Escherichia coli

Twenty-three strains of Escherichia coli with mannose-resistant and eluting (MRE) haemagglutinins with eight different patterns of substrate specificity were examined by a variety of electronmicroscope methods. In 12 strains, the presence of MRE haemagglutinins with broad-spectrum patterns 1, 3 and 4 was correlated with that of fimbriae. In 11 strains with MRE haemagglutinins with the less common patterns 6, 7 var., 8, 9 and 10, fimbriae were not found on bacteria in MRE+ cultures, indicating that the latter haemagglutinins are non-fimbrial.     

Functional heterogeneity of type 1 fimbriae of Escherichia coli.

Escherichia coli and other members of the family Enterobacteriaceae express surface fibrillar structures, fimbriae, that promote bacterial adhesion to host receptors. Type 1 fimbriae possess a lectinlike component, FimH, that is commonly thought to cause binding to mannose-containing oligosaccharides of host receptors. Since adhesion of type 1 fimbriated organisms are inhibited by mannose, the reactions are described as mannose sensitive (MS). We have studied the adhesion of the type 1 fimbriated CSH-50 strain of E. coli (which expresses only type 1 fimbriae) to fibronectin (FN). E. coli CSH-50 does not bind detectable amounts of soluble FN but adheres well to immobilized plasma or cellular FN. This adhesion was inhibited by mannose-containing saccharides. By using purified domains of FN, it was found that E. coli CSH-50 adheres primarily to the amino-terminal and gelatin-binding domains, only one of which is glycosylated, in an MS fashion. Binding of the mannose-specific lectin concanavalin A to FN and ovalbumin was eliminated or reduced, respectively, by incubation with periodate or endoglycosidase. Adhesion of E. coli CSH-50 to ovalbumin was reduced by these treatments, but adhesion to FN was unaffected. E. coli CSH-50 also adheres to a synthetic peptide copying a portion of the amino-terminal FN domain (FNsp1) in an MS fashion. Purified CSH-50 fimbriae bound to immobilized FN and FNsp1 in an MS fashion and inhibited adhesion of intact organisms. However, fimbriae purified from HB101 (pPKL4), a recombinant strain harboring the entire type 1 fim gene locus and expressing functional type 1 fimbriae, neither bound to FN or FNsp1 nor inhibited E. coli adhesion to immobilized FN or FNsp1. These novel findings suggest that there are two forms of type 1 MS fimbriae. One form exhibits only the well-known MS lectinlike activity that requires a substratum of mannose-containing glycoproteins. The other form exhibits not only the MS lectinlike activity but also binds to nonglycosylated regions of proteins in an MS manner.     

Coordinate expression of fimbriae in uropathogenic Escherichia coli

Uropathogenic Escherichia coli is the most common etiological agent of urinary tract infections. Bacteria can often express multiple adhesins during infection in order to favor attachment to specific niches within the urinary tract. We have recently demonstrated that type 1 fimbria, a phase-variable virulence factor involved in adherence, was the most highly expressed adhesin during urinary tract infection. Here, we examine whether the expression of type 1 fimbriae can affect the expression of other adhesins. Type 1 fimbrial phase-locked mutants of E. coli strain CFT073, which harbors genes for numerous adhesins, were employed in this study. CFT073-specific DNA microarray analysis of these strains demonstrates that the expression of type 1 fimbriae coordinately affects the expression of P fimbriae in an inverse manner. This represents evidence for direct communication between genes relating to pathogenesis, perhaps to aid the sequential occupation of different urinary tract tissues. While the role of type 1 fimbriae during infection has been clear, the role of P fimbriae must be further defined to assert the relevance of coordinated regulation in vivo. Therefore, we examined the ability of P fimbrial isogenic mutants, constructed in a type 1 fimbrial-negative background, to compete in the murine urinary tract over a period of 168 h. No differences in the colonization of these mutants were observed. However, comparison of these results with previous studies suggests that inversely coordinated expression of adhesin gene clusters does occur in vivo. Interestingly, the mutant that was incapable of expressing either type 1 or P fimbriae compensated by synthesizing F1C fimbriae.     

Monoclonal antibodies for serotyping the P fimbriae of uropathogenic Escherichia coli.

Monoclonal antibodies (MAbs) against seven serologically different P fimbriae (F7(1), F7(2), F8, F9, F11, F12, and F13) of uropathogenic Escherichia coli were tested for their ability to detect the P fimbriae on wild-type strains. In a plate agglutination test the MABs could detect the fimbriae on strains which expressed cloned fimbriae but not on wild-type strains. In a coagglutination test and in a whole-bacterium enzyme-linked immunosorbent assay the MAbs recognized the fimbriae on strains with cloned fimbriae and on wild-type strains. However, the coagglutination test has some disadvantages: only immunoglobulin G MAbs can be used, and the results cannot be read in an objective way. From these results, we concluded that the whole-bacterium enzyme-linked immunosorbent assay is the most convenient method for the determination of P fimbriae on wild-type E. coli strains. With this fast and easy method it is possible to do epidemiological studies on the distribution of P fimbriae among clinical isolates of uropathogenic E. coli and to extend the O:K:H serotype with the F serotype.     

Molecular structure of adhesin domains in Escherichia coli fimbriae

Crystal structures of FimH, PapG, GafD, and DraE fimbrial adhesin subunits or lectin domains have been resolved. These adhesins bind to different targets and are only distantly related in amino acid sequence. The overall structures of the fimbrial lectins, however, appear similar, suggesting that the fimbrial lectins have diverged from a common scaffold. FimH, PapG and GafD are two-domain structures connected by a flexible linker, and the N-terminal adhesin domains have an elongated beta-barrel jelly roll fold that contains the receptor-binding groove. The adhesin domains differ in disulfide patterns, in size and location of the ligand-binding groove, as well as in mechanism of receptor binding. Minor sequence variations that can be either distant from, near to, or at the ligand-binding groove have profound effects on receptor binding by the fimbriae; this is particularly apparent with FimH. The existing structures give insight into the molecular basis of the diversity in fimbrial lectins.     

Production of type 1 fimbriae by Escherichia coli HB101.

Escherichia coli HB101 is frequently used as a host in the cloning of bacterial virulence genes because of its reported lack of virulence determinants such as fimbriae, adhesins and haemagglutinins. However, passage of HB101 in standing broth culture rapidly induced the production of fimbriae which mediated adhesion to HEp-2 cells and mannose-sensitive haemagglutination of human and guinea-pig erythrocytes. Fimbrial serology, morphology and pilin molecular mass of 18 kDa were consistent with those of type 1 fimbriae.     

Binding of Escherichia coli S fimbriae to human kidney epithelium.

Purified S fimbriae and an Escherichia coli strain carrying the recombinant plasmid pANN801-4 that encodes S fimbriae were tested for adhesion to frozen sections of human kidney. The fimbriae and the bacteria bound to the same tissue domains, and in both cases the binding was specifically inhibited by the receptor analog of S fimbria, sialyl(alpha 2-3)lactose. S fimbriae bound specifically to the epithelial elements in the kidneys; to the epithelial cells of proximal and distal tubules as well as of the collecting ducts and to the visceral and parietal glomerular epithelium. In addition, they bound to the vascular endothelium of glomeruli and of the renal interstitium. No binding to connective tissue elements was observed. The results suggest that the biological function of S fimbriae is to mediate the adhesion of E. coli to human epithelial and vascular endothelial cells.     

Identification of F11 fimbriae on chicken Escherichia coli strains.

Fimbriae were purified from Escherichia coli strains isolated from chickens with septicemia or colibacillosis. When grown on solid media, these strains expressed fimbriae with an apparent subunit molecular mass of 18 kDa. Morphological, biochemical, serological, functional, and molecular characterization revealed that these 18-kDa fimbriae are identical to F11 fimbriae, which were previously found to be involved in the pathogenesis of human urinary tract infection. Screening of a large strain collection showed that 78% of chicken E. coli strains expressed F11 fimbriae, whereas this percentage increased to 96% when the only strains taken into account were those with the serotypes most commonly encountered in avian colibacillosis (O1:K1, O2:K1, O35, and O78:K80). The prevalence of F11 fimbrial expression appeared to be independent of the country of isolation of the strains, except for the United States, where the prevalence seemed higher. Expression of F11 fimbriae by chicken E. coli strains could not be correlated with adherence to chicken tracheal or pharyngeal cells.     

Rapid screening method for enterotoxigenic Escherichia coli.

A rapid screening method for detection of Escherichia coli producing heat-labile enterotoxin is described. Single colonies are transferred directly from primary culture plates into 96-well microculture plates containing 0.3 ml of brain heart infusion broth in each well. After 24 h at 37 degrees C, each brain heart infusion broth culture is assayed by the miniculture method in the corresponding well of a microculture plate in which Y-1 mouse adrenal tumor cells have been grown. All enterotoxigenic isolates detected by this method were confirmed in the assay but with culture supernatants.     

Epithelial invasion by Escherichia coli bearing Dr fimbriae is controlled by nitric oxide-regulated expression of CD55

We previously reported that inhibition of nitric oxide (NO) increases the rate of bacteremia and maternal mortality in pregnant rats with uterine infection by Escherichia coli expressing the Dr fimbria (Dr⁺). Epithelial binding and invasion by Dr⁺ E. coli has also been shown to be dependent upon the expression level of the cellular receptor decay-accelerating factor (DAF; CD55). Here, we hypothesize that NO-related severity of infection could be mediated by changes in DAF expression and in the rate of epithelial invasion. The cellular basis of NO effects on epithelial invasion with Dr⁺ E. coli was studied using Ishikawa endometrial carcinoma cells as an in vitro model of the human endometrial epithelium. Initially, we show that Ishikawa cells produce NO and express both NO synthase enzymes, NOS II and NOS III, and DAF protein. We next tested the abilities of both Dr⁺ E. coli and a Dr⁻ E. coli mutant to invade Ishikawa cells, and invasion was seen only with Dr⁺ E. coli. Invasion by Dr⁺ E. coli was decreased by elevated NO production and increased by NO inhibition. Elevated NO production significantly decreased DAF protein and mRNA expression in Ishikawa cells in a time- and dose-dependent manner. Here, we propose that in vitro invasion of an epithelial cell line is directly related to NO-regulated expression of DAF. The significance of NO-regulated receptor-ligand invasion is that it may represent a novel unrecognized phenomenon of epithelial defense against infection.     

Role of Escherichia coli P fimbriae in intestinal colonization in gnotobiotic rats.

Adherence via P fimbriae is associated with long-term persistence of Escherichia coli in the human large intestine, but a causal relationship has not been proven. In the present study, germfree rats were colonized with a mixture of two isogenic E. coli strains, one P fimbriated and the other type 1 fimbriated. Both types of fimbriae conferred adherence to rat colonic epithelial cells. With two mutant strains from a pyelonephritogenic isolate of serotype O75:K5:H-, the P-fimbriated strain 824 attained much higher numbers than its type 1-fimbriated counterpart when colonized in vivo for 2 weeks (10(10) versus 10(6) bacteria per g, respectively; P < 0.0001). The expression of P fimbriae by 824 was also retained during colonization. With transformant isogenic strains obtained from a normal fecal isolate incapable of phase variation, no benefit of P fimbriae was seen and most bacteria lost their plasmids during in vivo colonization. When the pyelonephritogenic mutant and fecal transformant strains were combined, the former colonized at high levels while the latter were suppressed. In contrast, no suppression was seen when the transformant E. coli strains colonized in combination with Lactobacillus acidophilus or Peptostreptococcus sp. The results indicate that P fimbriae, but also other bacterial traits linked to uropathogeneicity, could play an important role for persistence in the gut normal microbiota. Neither P nor type 1 fimbriae seemed to contribute to the ability to translocate to the mesenteric lymph nodes.     

Localization of binding sites for purified Escherichia coli P fimbriae in the human kidney.

Binding sites in the human kidney for purified P fimbriae of pyelonephritogenic Escherichia coli were determined. The purified KS71A (F7(1)) fimbriae bound only to epithelial elements of the kidney, i.e., to the apical aspect of proximal and distal tubular cells, as well as to the apical and cytoplasmic sites of collecting ducts. In addition, binding was seen at the vascular endothelium throughout the kidney and at the parietal epithelium of the glomeruli. The binding was specifically inhibited by the receptor analog of E. coli P fimbriae, globotriose. The binding sites identified suggested a possible pathogenetic mechanism for the invasion of P-fimbriated bacteria into the renal parenchyma, as well as for their subsequent spread into the circulatory system.     

Receptor structure for F1C fimbriae of uropathogenic Escherichia coli

F1C fimbriae are correlated with uropathogenic Escherichia coli strains. Although F1C fimbriae mediate binding to kidney tubular cells, their receptor is not known. In this paper, we demonstrate for the first time specific carbohydrate residues as receptor structure for F1C-fimbria-expressing E. coli. The binding of the F1C fimbriated recombinant E. coli strain HB101(pPIL110-54) and purified F1C fimbriae to reference glycolipids of different carbohydrate compositions was evaluated by using thin-layer chromatography (TLC) overlay and solid-phase binding assays. TLC fimbrial overlay analysis revealed the binding ability of purified F1C fimbriae only to glucosylceramide (GlcCer), beta1-linked galactosylceramide 2 (GalCer2) with nonhydroxy fatty acids, lactosylceramide, globotriaosylceramide, paragloboside (nLc(4)Cer), lactotriaosylceramide, gangliotriaosylceramide (asialo-GM(2) [GgO(3)Cer]) and gangliotetraosylceramide (asialo-GM(1) [GgO(4)Cer]). The binding of purified F1C fimbriae as well as F1C fimbriated recombinant E. coli strain HB101(pPIL110-54) was optimal to microtiter plates coated with asialo-GM(2) (GgO(3)Cer). The bacterial interaction with asialo-GM(1) (GgO(4)Cer) and asialo-GM(2) (GgO(3)Cer) was strongly inhibited only by disaccharide GalNAcbeta1-4Galbeta linked to bovine serum albumin. We observed no binding to globotetraosylceramide or Forssman antigen (Gb(5)Cer) glycosphingolipids or to sialic-acid-containing gangliosides. It was demonstrated that the presence of a GalCer or GlcCer residue alone is not sufficient for optimal binding, and additional carbohydrate residues are required for high-affinity adherence. Indeed, the binding efficiency of F1C fimbriated recombinant bacteria increased by 19-fold when disaccharide sequence GalNAcbeta1-4Galbeta is linked to glucosylceramide as in asialo-GM(2) (GgO(3)Cer). Thus, it is suggested that the disaccharide sequence GalNAcbeta1-4Galbeta of asialo-GM(2) (GgO(3)Cer) which is positioned internally in asialo-GM(1) (GgO(4)Cer) is the high-affinity binding epitope for the F1C fimbriae of uropathogenic E. coli.     

Loss of lectin-like activity in aberrant type 1 fimbriae of Escherichia coli.

Growth of a streptomycin-resistant strain of Escherichia coli (VL-2) in the presence of 30 microgram of streptomycin per ml resulted in the production by these bacteria of structurally altered, nonfunctional type 1 fimbriae. This strain, when grown in this subinhibitory concentration of streptomycin, became incapable of producing mannose-sensitive hemagglutination (<1% of that of the control). Adhering ability to epithelial cells and human leukocytes was also diminished (42 and 7% of that of the control, respectively). Although these streptomycin-treated bacteria were as heavily fimbriated as untreated bacteria, their fimbriae were significantly longer. Furthermore, in contrast to the fimbriae of the untreated bacteria, those isolated from the drug-treated bacteria were found to lack mannose binding activity as measured by hemagglutination. It appears, therefore, that streptomycin can cause even resistant bacteria to produce an aberrant fimbrial protein, possibly by causing misreading of messenger RNA. These studies indicate that the use of sublethal doses of certain antibiotics whose mode of action is well known may shed light on the genetic and chemical modulation of bacterial factors involved in mucosal colonization.     

Serological response to the P fimbriae of uropathogenic Escherichia coli in pyelonephritis.

Uropathogenic Escherichia coli strains isolated from four patients with pyelonephritis were characterized by their O:K serotype, hemolysin production, mannose-resistant hemagglutination, and the serotype of the P fimbriae. These P fimbriae were serotyped with specific monoclonal antibodies. Serum samples from the patients were analyzed for the presence of specific antibodies to the P fimbriae. In all cases antifimbrial antibodies were found, strongly suggesting that these P fimbriae are expressed in vivo. However, the antibodies in the patient sera were not able to inhibit the mannose-resistant hemagglutination. This finding suggests that these antibodies react with the fimbrial components and not with the minor components which are responsible for adhesion.     

Expression of P fimbriae of uropathogenic Escherichia coli strains

The occurrence of P fimbriae in a total of 222 uropathogenic Escherichia coli (UPEC) strains was investigated. Out of the total, 31 (14%) were P fimbriated. Of 24 pyelonephritogenic E. coli strains, three (13%) with P fimbriae occurred in children with clinical pyelonephritis, and of 198 E. coli strains 29 (15%) occurred in children with cystitis. Prevalence of P fimbriae of E. coli strains was found to be quite similar in patients with cystitis and pyelonephritis     

Binding of Escherichia coli S fimbriae to cultured human endothelial cells.

The binding of Escherichia coli adhesins to human umbilical vein endothelial cells was studied by a cell monolayer enzyme-linked immunosorbent assay. S fimbriae displayed a concentration-dependent and saturable binding to the endothelial cells which was mediated by their sialylgalactoside-specific lectin activity. P fimbriae exhibited only low binding, and type 1 fimbriae exhibited no binding to these cells.     

Increased frequency of intestinal Escherichia coli carrying genes for S fimbriae and haemolysin in IgA-deficient individuals.

Persons with selective IgA deficiency carry an increased risk of coeliac disease, inflammatory bowel disease and perhaps also gastrointestinal malignancies. Inflammatory bowel disease is associated with an increased carriage of adherent and haemolytic Escherichia coli in the intestinal microflora. This study was designed to investigate whether IgA-deficient individuals carry E. coli with virulence-associated properties in their gut flora. The last free-lying colony of E. coli isolates obtained from rectal flora of 25 IgA-deficient and 20 age-matched control individuals was assayed by multiplex PCR for genes for the following adhesins or virulence determinants: P, type 1 and S fimbriae, Dr haemagglutinin, haemolysin, aerobactin and the capsular types K1 and K5. E. coli strains from the intestinal microflora of IgA-deficient individuals more often had the gene for S fimbriae (36% of the strains compared with 0% in control subjects, P=0.003) as well as for haemolysin (40 vs 10% of the strains, P=0.040). IgA-deficient individuals had instead lower frequencies of E. coli carrying genes for type 1 fimbriae in their microflora (68 vs 90%, P=0.14). The results suggest that IgA-deficient individuals carry an increased frequency of E. coli with potentially inflammatogenic properties in their microflora, which may contribute to the development of gastrointestinal disorders such as inflammatory bowel diseases.