TEXT

Top Abstract Text References

P-fimbrial adhesins in Escherichia coli enable the colonization of host tissues. By mediating attachment to P-blood group antigens on uroepithelial cells (6, 12, 16), P fimbriae play a critical role in the development of urinary tract infections (UTIs). P fimbria production is regulated by a chromosomal pap operon, containing 11 genes (4). The P-fimbrial-tip adhesin, which is encoded by papG, attaches directly to host cells (7, 9). The three adhesin classes (papG_J96 [class I], papG_AD/IA2 [class II], and prsG_J96 [class III]) were characterized based on their capacities for binding to specific Gal(1-4)Gal-containing glycolipids (14). prf (pap-related fimbriae) is generalized nomenclature that includes all gene clusters encoding P fimbriae. The prf probe contains the most conserved genes and thus hybridizes to all of these gene clusters (1). Class I adhesins have 45% identity at the amino acid level to class II and 46% identity to class III, while class II adhesins have 56% identity to class III. Minor papG variants have been reported (GenBank accession numbers AAF61952 [J. R. Johnson, N. Kaster, T. T. O'Bryan, and A. L. Stell, unpublished data], AAF61956 [5], AAD13607 [3], and AAA59216 [10]) with homology to the three papG alleles, ranging in identity from 89 to 96%. We present the discovery of a new G allele of P fimbriae with less than 65% identity to known adhesin classes.

Three E. coli collections were studied: 313 isolates from college women aged 18 through 39 at the University of Michigan or University of Texas at Austin between 1992 and 1995 with a first-time UTI (first UTI), 51 isolates from a subset of these same women reporting a second UTI within 6 months of the first (second UTI), and 377 fecal and 74 periurethral isolates from healthy women presenting to the University of Michigan Student Health Service for gynecological exams between February and March 1996. All E. coli isolates were cultured and processed as previously described (1, 17).

A total of 815 E. coli strains were screened by dot blot hybridization for the presence of prf, a cluster of gene sequences specific to P-related fimbriae, as described previously (1). Sequence homology to prf was detected in 332 (41%) of the strains tested. The P-fimbrial adhesin class was also determined by screening those strains positive for prf with the three class-specific DNA probes using dot blot hybridization (Table 1). The adhesin class-specific probes (papG_J96 [class I], papG_AD [class II], and prsG_J96 [class III]) were derived from published sequences (8, 13, 14) and isolated from control strain J96 or C1212 by PCR (1, 2). Strain J96 contains papG_J96 and prsG_J96, while control strain C1212 contains papG_AD/IA2. We confirmed dot blot results with PCR using unique primers for each adhesin class. Identical PCR conditions, except the annealing temperatures, were used for each adhesin class (30 cycles of 94°C for 60 s and 73°C for 40 s, with an annealing time of 35 s). Table 2 lists the annealing temperatures and PCR primers. Based on these results, 20 strains positive for prf and negative for all three adhesin classes were identified (Table 1), suggesting the presence of papG variants.

Pulsed-field gel electrophoresis analysis showed that all 20 strains differ by three or more bands (15) and therefore do not represent a clonal grouping (data not shown). Because papF is conserved among the three known adhesin classes (13), we assessed whether it could be amplified using PCR in the 20 strains (Fig. 1). A 502-bp fragment was amplified in control strains J96 and C1212 and in 11 of 20 (55%) strains with an unknown papG adhesin class.

View larger version (85K): [in this window] [in a new window]   FIG. 1.   PCR amplification of papF. Lane 1, DNA lambda digested with HindIII; lane 2, J96; lanes 3 to 12, unknown-adhesin strains BF31, BF56, BF1163, BF141, BF192, BF115, BF1160, BF166, and BF268; lane 13, negative control. PCR amplification of papF was detected using primers 5'-ATCGTTGCTTCTGACATCGG-3' and 5'-GTCAATAAGTAATCCCATA CTG-3' (30 cycles of 94°C for 60 s, 56°C for 30 s, and 74°C for 30 s).

In order to determine whether any of these strains contained novel P-fimbrial G alleles, we cloned and sequenced papG from two randomly selected strains (BF1163 and BF31). Southern blot hybridization using a 502-bp papF PCR probe labeled with digoxigenin (Genius System kit; Boehringer Mannheim, Indianapolis, Ind.) detected DNA fragments of 6.5 kb for fecal strain BF31 (Fig. 2), and of 4 and 2.3 kb for UTI strain BF1163 digested with BsaB1 and Psp14061 (data not shown), respectively. Both the 6.5- and the 4-kb DNA fragments were purified following gel electrophoresis and cloned using the pZErO-1 vector (Invitrogen, San Diego, Calif.) with TOP10F' (Invitrogen) as the recipient strain by methods described previously (17). Plasmid DNA was isolated using a plasmid preparation kit (Qiagen, Chatsworth, Calif.). Restriction enzyme digestion, nuclease treatment, and ligation were performed according to standard protocols (11).

View larger version (51K): [in this window] [in a new window]   FIG. 2.   Restriction fragment length polymorphisms of E. coli strains digested with BsaB1 and probed with papF (502 bp). Lanes 1 and 15;  DNA markers digested with HindIII and BstEII; lanes 2, 3, 4, 5, 7, 8, 9, l0, and 12, unknown-adhesin fecal strains BF6, BF31, BF54, BF56, BF164, BF166, BF191, BF224, and BF370; lanes 6, 11, and 13, unknown-adhesin UTI strains BF115, BF268, and BF1009; lane 14, control strain J96.

The double-stranded DNA sequences of both clones were determined at the University of Michigan Biology Core Facility with an Applied Biosystems model 373A automated sequencer using primers T7 and SP6. Fecal strain BF31 contained a novel papG allele (papG_BF31), whereas UTI strain BF1163 contained a variant of papG with a deletion. papG_BF31 had amino acid sequence identities of 65% to papG_J96 (class I), 46% to papG_AD/IA2 (class II), and 45% to prsG_J96 (class III) (Fig. 3). We refer to papG_BF31 as the P-fimbrial class IV adhesin gene. BF1163 was most similar (70%) to papG_J96; however, the open reading frame was truncated at bp 290. 

View larger version (62K): [in this window] [in a new window]   FIG. 3.   Amino acid comparison of the new class IV adhesin (encoded by papGBF31) with the three known adhesin classes. DNAStar (Madison, Wis.) and Genetics Computer Group (Madison, Wis.) software was used for DNA and amino acid analyses.

In order to estimate the prevalence of papG_BF31 in other E. coli strains, we screened a sample of strains (n = 308) positive for prf by dot blot hybridization using a 371-bp probe specific to papG_BF31. papG_BF31 occurred with similar frequency in each collection. The numbers of strains positive for papG_BF31 were as follows: 21 (15%) among the first UTI collection (n = 144), 36 (15%) among the second UTI collection (n = 20), 20 (17%) among the fecal collection (n = 120), and 3 (13%) among the periurethral collection (n = 24). papG_BF31 is positively associated with aer and drb and is negatively associated with prsG_J96, hly, cnfl, ompT, and sfa (Table 3). Among isolates positive for papG_BF31, papG_AD was present in 55%.

PCR was performed on the remaining 19 prf-positive, class I-, II-, and III-negative strains using primers 5'-GACTATTCTGGTTATGATTC-3' and 5'-CAATGAATTAAGGTTTAG-3' (30 cycles of 95°C for 60 s, 46°C for 40 s and 73°C for 23 s), taken from a unique coding region of papG_BF31. A 371-bp fragment was amplified in 8 of the 19 (42%) strains, suggesting that other novel G allele variants may exist.

The novel class IV adhesin gene shows 45 to 65% similarity at the amino acid level to the three adhesin classes, thereby representing a unique adhesin class that is found equally among UTI and fecal E. coli strains. Thus, class IV adhesins are not exclusively associated with UTIs, although they could be associated with the pathogenesis of other important diseases.

The large prf probe used in this study hybridized to strains containing novel papG alleles as well as inactive papG variants or variants with deletions. Because a PCR fragment specific to the class IV gene was not amplified in 11 of the strains without class I, II, or III adhesins, it is possible that other novel molecular variants of papG exist. Future work should include hemagglutination assays to determine whether papG_BF31 is functional and to identify other novel papG variants and assess their role in UTIs or other diseases.

Nucleotide sequence accession number. The GenBank accession number for the papG_BF31 nucleotide sequence is AF304159.