Family ofEscherichia coliDr Adhesins: Decay‐Accelerating Factor Receptor Recognition and Invasiveness

In 1984, Väisänen-Rhen [1] and Labigne-Roussel et al. [2] described a novel Escherichia coli adhesin of unknown receptor specificity. Due to its association with uropathogenic strains of serotype O75 and because of the unknown receptor, the adhesin was named O75X [1]. Between 1984 and 1989, 3 groups independently cloned novel adhesins that were later found to have very similar genetic organization: The clones were Dr (cloned by Nowicki et al. [3]), afimbrial adhesins (Afa)-I and Afa-III (cloned by Labigne-Roussel and colleagues [2, 4]), and F1845 (cloned by Bilge et al. [5]). In 1990, Nowicki et al. [6] provided evidence that the novel adhesins, including O75X, recognize a common receptor, Dr (a+) blood group antigen. Dr (a+) blood antigen was later localized to the key complement regulatory molecule, decay-accelerating factor ([DAF]; CD55), which protects tissues from damage by autologous complement attack Most uropathogenic adhesins traditionally were named after the receptor they recognized; hence, we proposed to rename O75X the “Dr adhesin,” with the remaining adhesins to be included in the Dr family of E. coli adhesins. In the 1990s, various investigators described several adhesins that were cloned from E. coli of human and animal origin and that display similar genetic organization and/or DAF receptor recognition. Of interest, most of the tested adhesins appeared to have invasive capacity, with probably at least two genes contributing to the internalization process [7, 8]. We therefore propose that the prototypes and the new related adhesins constitute members of the Dr family of adhesins (table 1) Prototypes and new members of the family of Dr adhesins of uropathogenic and diarrhea-associated Escherichia coli Epidemiology of Dr adhesinsThe most frequent colonization factors among uropathogenic E. coli include P and type 1 fimbriae. The family of Dr adhesins seems to be the third most often observed group of colonization factors. Epidemiologic studies of clinical isolates from urinary tract infections (UTIs) and diarrhea samples suggest that children and pregnant women are predisposed to infection by the E. coli–expressing Dr adhesins. As many as 50% of isolates from children with protracted diarrhea and 25%–50% of cystitis isolates from children may express Dr adhesins [4]. In contrast, in adults, especially young, nonpregnant women, Dr+E. coli isolates are less frequent in cystitis and even less frequent in pyelonephritis cases. Of interest, in pregnant women, Dr+E. coli accounted for nearly 30% of pyelonephritis cases (up to 40% during the third trimester of gestation) [9]. It has been suggested that UTI caused by Dr+E. coli increases the risk of recurrent UTI [10], a suggestion that supports our hypothesis of a possible association of Dr adhesins with recurrent or chronic UTI [11]. In addition, 15% of healthy adults have colon colonization with Dr adhesin–bearing E. coli compared with 12% of pregnant females (unpublished data) E. coli Dr fimbriaeDr fimbriae were initially described as an O75X fimbria–like adhesin by Väisänen-Rhen [1], who, after observing them in high incidence in uropathogenic E. coli serotype O75, proposed that they represent a clonal group. The Dr (O75X) adhesin was cloned from the clinical pyelonephritis strain E. coli IH11128 [3]. In renal tropism studies, the tissue binding sites for the Dr adhesin were localized in the interstitial compartments of the kidney. Purified recombinant Dr adhesin binds to the basement membrane of the renal tubules and Bowman’s capsule of the glomerulus. A panel of human erythrocytes lacking specific antigens was used in an attempt to identify the tissue receptor that mediates binding of Dr adhesin and led to the finding that the Dr blood group antigen (Dra) is the receptor for O75X adhesin. This adhesin was termed Dr hemagglutinin. Unlike other members of the Dr adhesin family, Dr hemagglutinin in its interaction with Dra antigen was inhibited by chloramphenicol and modified tyrosine; therefore, it was termed chloramphenicol-sensitive hemagglutination (CSHA) Patients with paroxysmal nocturnal hemoglobinuria have red blood cells that are lower or deficient for DAF, and these cells are not hemagglutinated by Dr+E. coli. This observation led to the hypothesis that DAF harbors the Cromer blood group antigens, including the Dra, and may serve as the epithelial cell receptor for the Dr adhesins. Indeed, in a recombinant adhesin-receptor cell–cell interaction model, all of the Dr adhesins bound to DAF. The Dr receptor has been identified in the digestive, urinary, genital, and respiratory tracts, providing possible sites for colonization and infection by Dr adhesins. Dr hemagglutinin adhered to the brush borders of colonic and ileal enterocytes. In the kidneys, Dr hemagglutinin bound preferentially to basement membranes and Bowman’s capsule, apparently due to the ability of Dr fimbriae to bind an additional receptor (type IV collagen), as described by Westerlund et al. [12] Characterization of dra genes involved in the biogenesis of Dr fimbriae indicated that draA (the adhesin subunit) and three other genes (draC, draD, and draE) were required for the hemagglutination phenotype. The amino terminal portion of the draA fimbrial subunit was responsible for the CSHA DAF–binding phenotype, which is unique to Dr hemagglutinin among members of the Dr adhesin family. The dra genes were later renamed, and draE was called the adhesin subunit in accordance with the daaE (encoding F1845 fimbriae), afaE-1 (encoding Afa-I), and afaE-3 (encoding Afa-III) adhesin subunits. Mutational analysis of the draE adhesin has demonstrated the importance of a negatively charged amino acid at position 54 of the DraE adhesin for CSHA and suggested the involvement of a conformational binding epitope in type IV collagen and DAF binding Dr-fimbriated E. coli was virulent in a model of chronic tubulointerstitial nephritis in C3H/HeJ mice [11] and in pregnant rats (unpublished data). The Dr+E. coli IH11128 established a 1-year colonization of renal tissue, while its isogenic Dr− mutant was eliminated from 50% of animals within 20 weeks and from 100% of animals between weeks 32 and 52. Complementation of the Dr− mutant with plasmid carrying Dr operon restored the capacity of the mutant to cause persistent infection, fulfilling Koch’s molecular postulates for virulence factors [11]. In vivo, Dr+E. coli colonized the renal interstitium, a histologically unavailable site for a bacterium infecting by the ascending route, suggesting that the strain may employ an unknown invasion mechanism Goluszko et al. [7] provided evidence for Dr-fimbriated E. coli entry into HeLa epithelial cells that abundantly express DAF. Along with Goluszko et al., we found that F1845 and Afa-I internalize into the HeLa cell line (unpublished data). Of interest, an independent study by Jouve et al. [8] showed the invasive capacity of Afa-III. Selvarangan et al. [13] demonstrated that the short consensus repeat–3 (SCR-3) of DAF was required for internalization to occur and that glycosylphosphatidylinositol anchorage was more efficient than chimeric transmembrane anchorage of DAF in providing entry of Dr+E. coli into epithelial cells. We proposed that this novel entry mechanism facilitates access of Dr+E. coli to renal interstitium, which is rich in Dr receptors, DAF, and type IV collagen We also observed that the ability of Dr+E. coli to bind to type IV collagen provides a functional basis for the persistence of infection in a mouse model of chronic pyelonephritis (Selvarangan et al., unpublished study). Johnson et al. [14] showed that adherence of Dr adhesins to polymorphonuclear neutrophils resulted in minimal bacterial killing when compared with other adhesins of uropathogenic strains (e.g., type 1 and P fimbriae). This novel mechanism of resistance to phagocytic killing may be a significant factor in promoting persistence of infection by Dr-fimbriated E. coli. The lethal effect of nitric oxide–enhanced Dr E. coli and of Afa-III on pregnant rats may result from unique gestational tropism and reflect urovirulence of Dr E. coli to the pregnant female (Le Bouguenec, Yallampalli, and Nowicki; unpublished observations) E. coli F1845 fimbriaeF1845, a fimbrial adhesin originally cloned by Bilge et al. [5] from a diarrheal E. coli isolate, causes diffuse Hep-2 cell adherence. A probe designed from the daaC gene (a conserved region of F1845/Dr operons) to detect diffuse-adhering E. coli resulted in association of F1845/Dr adhesins with a significant proportion of isolates from children with protracted diarrhea. Bilge et al. [5] described the daa genes encoding the F1845 fimbriae that were present both on the plasmid and chromosome of different isolates, suggesting that they may be carried by mobile DNA elements. The operon involved in the biogenesis of the F1845 fimbriae consists of at least five genes (daaA–E) encoding products similar to other Dr adhesin systems, and daaE has been termed the adhesin subunit. Sequences downstream of daaE encode a stem-loop structure that was proposed to determine the stability of the daaE mRNA. A novel gene (daaP) may be involved in processing of daaE Similar to other members of the Dr family, F1845 recognized Dr blood group antigen that was mapped to the SCR-3 domain of DAF [6]. F1845 mediates adherence both to the human intestinal undifferentiated and differentiated fluid-transporting cells and to mucus-secreting cells. The binding of a clinical diarrheal E. coli strain, C1845, to the epithelial intestinal cells promotes microvillar injury lesions resulting from disassembly of the actin network. Peiffer et al. [15] subsequently demonstrated that the microvillar lesions were initiated by signal-transduction events mediated by DAF and bound the F1845-positive E. coli E. coli Afa-IAfa-I was originally isolated by Labigne-Roussel et al. [2] in 1984 from a uropathogenic E. coli that exhibited mannose-resistant, P fimbriae–independent hemagglutination. Afa-I exists on the bacterial surface as a macromolecular aggregate not associated with fimbrial morphology. The clinical strain KS52 of the O2 serotype expressed Afa-I that was composed of a single, repeating 16-kDa polypeptide subunit that was identified as the hemagglutinin. The Afa-I operon has a genetic organization that is similar to that of the other members of the Dr adhesins and consists of at least five genes involved in the biogenesis of the Afa-I adhesin. A 750-bp fragment of the afaB region (chaperone protein) was used to identify related sequences in E. coli isolated from UTIs and diarrhea samples [4] Similar to F1845, Afa-I–expressing E. coli adhere to human intestinal undifferentiated and differentiated fluid-transporting cells and to mucus-secreting cells. Although Afa-I binds the Dr antigen on DAF, it seems likely that the epitope recognized by Afa-I is distinct from other Dr adhesins since trypsin treatment of erythrocytes abolishes hemagglutination and additional sequences on SCR-4 may also be needed for binding [6] E. coli Afa-IIIThe Afa-III gene cluster was initially cloned by Labigne-Roussel and Falkow [4] in 1988 from E. coli A30, a cystitis isolate. The Afa-III operon is carried on large, 100-kb plasmids of an E. coli isolate of urinary or diarrheal origin. The afa gene cluster contains six open-reading frames, designated afaA–F, and the entire afa-3 gene cluster is flanked by two directly oriented insertion sequence elements, which may allow translocation of the operon from the plasmid to the chromosome. On the basis of homologies with other adhesin systems, putative functions were attributed to the individual Afa products. AfaA and AfaF proteins were found to be homologous to the PapI-PapB transcriptional regulatory proteins, while AfaB and AfaC were recognized as periplasmic chaperone and outer membrane anchor proteins, respectively. AfaE-III protein was described as the structural adhesin subunit. Sequence comparison between AfaE-III and DraE adhesin of the Dr fimbriae revealed high homology (98.1%), with a difference of only 3 amino acids at positions 54, 90, and 113 of the mature adhesin protein Despite the high homology, structural and functional differences exist between Dr and Afa-III adhesins. Dr (DraE) adhesin is associated with fimbrial structure and significant DAF clustering, while AfaE-III is an afimbrial adhesin and clusters DAF less efficiently. Dr binds to DAF and type IV collagen, while Afa-III binds only to DAF. Binding of Dr fimbriae to the receptors is blocked by chloramphenicol (2 μM) or modified tyrosine molecules, while AfaE-III binding is not affected by these molecules. Of most interest, Selvarangan et al. [13] demonstrated that purified recombinant Dr fimbriae alone bind and mediate internalization of the Dr+E. coli through a cholesterol-dependent glycolipid pathway. Mutation of afaD-3 (Jouve et al. [8]) and draD (Goluszko et al. [7]) also decreased invasion, suggesting that there may be at least two invasins encoded by Dr operon. Jouve et al. [8] proposed that AfaE-III allows binding to epithelial cells and AfaD-III mediates the internalization of adherent bacteria. The exact mechanisms explaining invasive differences between DraE and AfaE-III remain to be investigated but may result in part from the 3 amino acid sequence differences E. coli Dr-II adhesinPham et al. [9] screened 64 pyelonephritis isolates from a population of pregnant women for Dr adhesins and found a 750-bp afaB–polymerase chain reaction (PCR) fragment in 17 of them. Twelve of these PCR-positive isolates were classified as Dr adhesins because they did not agglutinate Dr(a−) erythrocytes, a simple phenotypic test for Dr adhesin. Of interest, the remaining 5 isolates agglutinated Dr(a−) erythrocytes, suggesting the presence of novel DAF-binding epitopes and/or a novel fimbria. The gene cluster from 1 such isolate was subsequently cloned and called Dr-II adhesin [9]. Despite the low homology (20%) of Dr-II adhesin with Dr, the binding site of Dr-II was similar to that of Dr adhesin, which was mapped to the SCR-3 region of DAF; possible interaction of Dr-II with SCR-4 of DAF requires further investigations. Similar to the case with Dr fimbriae, the interaction of Dr-II with DAF leads to internalization of the bacterium into HeLa cells New members of Dr familyThe E. coli Dr operons involved in the biogenesis of the Dr family of adhesins share a similar genetic organization. The products encoded by the Dr operon are involved in the assembly of Dr adhesins, which recognize DAF as their cellular receptor. Recently, many adhesin systems encoded by genetic clusters similar to the organization of the Dr operons have been described. These new adhesins include afimbrial adhesins Afa-VII and Afa-VIII [16], nonfimbrial adhesin Nfa-I [17, 18], and aggregative adherence fimbriae Aaf-I and Aaf-II [19] The AfaE-VIII adhesin seems to be very similar to the M agglutinin of strains uropathogenic to humans. It is interesting to note that the amino acid composition of the adhesin subunits encoded by the newly identified adhesion systems are different from the previously described Dr adhesins. Therefore, it was of interest to see if these adhesins, similar to Dr adhesins, mediated binding to DAF. We analyzed E. coli expressing adhesins Afa-V, Aaf-I, or Aaf-II and Nfa-I–related adhesin (from Marrs and colleagues, School of Public Health, Department of Epidemiology, University of Michigan) for their ability to mediate binding to recombinant CHO cells expressing DAF. Results from the binding assay indicate that Afa-V and Nfa-I belong to the family of Dr adhesins that binds to CHO cells expressing DAF (CHO-DAF+) but not to control CHO cells (CHO-DAF−) (unpublished data). Although Aaf-I and Aaf-II showed similar genetic organization to Dr adhesins, they mediated an aggregative adherence pattern both to the CHO-DAF+ and CHO-DAF− cells. This suggests that they either do not employ DAF as their cellular receptor or that their aggregating phenotype masks detection of the DAF binding capacity The family of Dr adhesins appears to be one of the most frequent among uropathogenic E. coli and it is surprisingly frequent in diarrheal strains. Their unusual urodiarrheal virulence and distribution require further studies but may result from two common features: recognition of a common receptor and invasive capacity. Published studies indicate that most of the Dr family adhesins represent a growing superfamily of adhesins with similar genetic organization of the gene cluster involved in the biogenesis of the adhesin and with recognition of DAF and/or the capacity to invade epithelial cells, a feature that may contribute to E. coli survival and virulence. The invasion capacity may be encoded by two genes (draE, draD) or their afa homologues. The heterogeneity of the members of the Dr superfamily may account for a broad spectrum of virulence, with some strains predisposed to cause UTI in pregnant females or children, some prone to cause protracted diarrhea or both diarrhea and UTI in children, and perhaps some predisposed to cause chronic and/or recurrent UTI. Further epidemiologic and molecular studies are necessary to understand virulence of Dr adhesins, which despite their frequent occurrence are still underinvestigated Due to the publication requirements regarding the number of references, we were unable to cite all the excellent work and relevant papers in the field. We thank Carla Pairett for manuscript preparation