OUP user menu

Identification of S, F1C and three PapG fimbrial adhesins in uropathogenic Escherichia coli by polymerase chain reaction

Kenji Mitsumori, Akito Terai, Shingo Yamamoto, Osamu Yoshida
DOI: http://dx.doi.org/10.1111/j.1574-695X.1998.tb01173.x 261-268 First published online: 1 August 1998


S and F1C fimbrial adhesins often expressed by uropathogenic Escherichia coli are genetically homologous. A multiply primed polymerase chain reaction (PCR) was developed for discriminating the S (sfa) and F1C (foc) fimbrial operons. A total of 270 uropathogenic E. coli strains and 80 fecal isolates were examined. PCR specifically detected the sfa and foc alleles in 105 (93%) of 113 sfa/foc+ strains by DNA hybridization. Furthermore, 87% of sfa+ uropathogenic E. coli simultaneously possessed the genes encoding the class III P fimbrial adhesin (prsGJ96), α-hemolysin and cytotoxic necrotizing factor 1. Statistical analysis showed the class II P fimbrial adhesin (papGIA2) and F1C fimbria to be associated with high relative virulence in pyelonephritis and cystitis, respectively. The multiply primed PCR developed should be useful for assessing the contribution of the S and F1C fimbriae in the pathogenesis of urinary tract infections.

  • Escherichia coli
  • P fimbria
  • S fimbria
  • F1C fimbria
  • Polymerase chain reaction

1 Introduction

Escherichia coli causes about 80% of acute urinary tract infections (UTIs) [1]. The ability of the bacteria to adhere to uroepithelial cell receptors through specific fimbrial adhesins is critical for the initiation of infections [1]. The adhesins can be differentiated on the basis of their receptor binding specificities. Uropathogenic E. coli most often expresses P fimbriae (encoded by the pap operon) which bind to digalactoside-containing receptors and, to a lesser extent, S fimbriae (encoded by the sfa operon) which bind to sialic acid-containing receptors [1]. Recent genetic investigations have shown that P and S fimbriae have several homologous operons encoding adhesins with different binding specificities. P fimbriae have three defined PapG adhesin variants; class I variant encoded by the papGJ96 gene, class II variant (P adhesin) encoded by the papGIA2 gene and class III variant (F adhesin) encoded by the prsGJ96 gene [2]. It has been shown that P adhesin predominates in acute pyelonephritis strains, while F adhesin is most frequent in cystitis strains [3]. F1C fimbriae (encoded by the foc operon) are genetically related to S fimbriae and their nucleotide sequences share a high degree of homology [4, 5]. F1C fimbriae can bind to the uroepithelial cells although their receptor specificity has not yet been determined [5]. However, the role of S and F1C fimbriae in UTIs has not been established.

Recently, epidemiological studies on the fimbrial adhesins have been performed by DNA hybridization assay, which is independent of gene expression and can avoid the ambiguities of phenotypic assay. The three papG variant genes can be distinguished by colony hybridization using the specific DNA probes [6]. Although the oligonucleotide probes that bind specifically to the sfa gene were reported, no specific DNA probes have been described for detecting the foc operons [7]. This may be due to the fact that F1C fimbriae have been examined in only a few studies [8]. Recently, Johnson et al. [9] developed a multiply primed PCR assay for simultaneous detection of the three papG variant genes, which is simpler and more practical than colony hybridization. Our previous investigations on the urovirulence determinants of uropathogenic E. coli strains using DNA hybridization test and multiply primed PCR [10, 11] have not been intended to differentiate the three papG variants or sfa/foc genes. In this study, we developed a new multiply primed PCR assay for differentiating the sfa and foc genes and validated this assay by comparing with our previous results using other genotyping methods. We then analyzed the relationship between the papG variant and sfa/foc genotypes, other urovirulence determinants and O:K:H serotypes in 270 uropathogenic E. coli strains.

2 Materials and methods

2.1 Bacterial strains and plasmids

We studied 270 uropathogenic E. coli strains, consisting of 194 isolates from women with acute uncomplicated cystitis and 76 isolates from women with acute uncomplicated pyelonephritis, and 80 isolates from normal fecal flora. These strains were described previously [10, 11].

The recombinant plasmids used in this study include pANN801-13 which carries the complete sfa operon (kindly provided by J. Hacker, Institute für Genetik und Mikrobiologie, Universität Würzburg, Germany), pPIL110-54 coding for the complete foc operon (kindly provided by G.B. Oehler, Institute für Molekulare Infektionsbiologie, Universität Würzburg, Germany) and pSH2 which codes for the complete pil operon (kindly supplied by P.E. Orndorff, College of Veterinary Medicine, North Carolina State University, USA). pANN801-13 and pPIL110-54 were used as positive controls and pSH2 as a negative control, respectively, for sfa- and foc-specific PCR assay.

2.2 DNA colony hybridization

Colony hybridization test was performed under high stringency conditions as described previously [10, 11]. In brief, the pap probe was a 2.4-kb SmaI fragment of pRHU845 containing papEFG genes. pRHU845 was kindly provided by S.I. Hull, Department of Microbiology and Immunology, Baylor College of Medicine, USA. The sfa/foc probe was a 1.8-kb EcoRV-NcoI fragment of pANN801-13 spanning sfaSH genes. The pap probe cannot discriminate between P fimbrial operons encoding class I, II and III papG adhesins because the papF sequences are highly homologous irrespective of the class of papG adhesin [2]. The sfa/foc probe reacts with both S and F1C fimbrial operons since a portion of the sfaH and focH sequences are highly homologous (overall homology, 81%) [4, 5].

2.3 PCR primers

The primers used in the multiply primed PCR to simultaneously detect six urovirulence determinants including the pap and sfa/foc operons were described previously [10, 12]. The PCR product of the pap primers is a 336-bp segment spanning the papEF region and for the sfa/foc primers, a 410-bp segment spanning the sfaDE region. We earlier reported that, for 194 cystitis isolates, the results of the multiply primed PCR were in complete agreement with those obtained by DNA colony hybridization test [10].

Primer pairs that specifically discriminate the sfa and foc genes were selected on the basis of the published sequences [4, 5] (GenBank accession numbers; X16664 and S68237, respectively), while avoiding primer dimer formation and giving the PCR products of different size to allow resolution by agarose gel electrophoresis. The sfa-specific primers, sfa711f and sfa861r, flanked a 151-bp DNA segment within the sfaS gene. Their sequences were as follows: sfa711f, 5′-GGATGTTTCTTTGGGTAATCTG-3′; sfa861r, 5′-CATTCCCTGTATTCGCATAG-3′. The foc-specific primers, foc844f and foc1231r, flanked a 388-bp DNA segment spanning the focGH genes, and their sequences were as follows: foc844f, 5′-GGTGGAACCGCAGAAAATAC-3′; foc1231r, 5′-GAACTGTTGGGGAAAGAGTG-3′. These sfa- and foc-specific primers will be designated as the sfaS and focGH primers, respectively, whenever necessary to avoid confusion with the previously reported multiply primed PCR for detecting both sfa and foc genes (sfa/foc primers) [10].

PCR primers for identification of three variant papG genes (papGJ96, papGIA2 and prsGJ96) were described previously [9].

2.4 Amplification procedures

Bacteria were collected by centrifugation from overnight broth cultures and resuspended in 1/5 vol of sterile water. After boiling for 15 min, bacterial suspension was centrifuged at 13 000×g and supernatant (10 00B5;l for papG variant-specific PCR assay and 5 00B5;l for sfaS/focGH-specific PCR assay) was used as template DNA. Amplification was done in a 30 00B5;l reaction mixture containing template DNA, 2 mM MgCl2, 0.2 mM each of the four dNTPs, 5% DMSO, 100 pmol of each primer and 2 U of AmpliTaq Gold (Perkin Elmer, NJ, USA) in 1×PCR buffer II (Perkin Elmer, NJ, USA). The samples were heated at 95°C for 9 min for the initial denaturation and activation of the Taq polymerase by the thermal cycler PC-800 (Astec, Fukuoka, Japan) and this step was followed by 30 cycles of denaturation (94°C, 1 min), annealing (68°C, 1 min) and extension (72°C, 2 min), and a final extension (72°C, 10 min). Samples were electrophoresed in 2% agarose gels, then stained with ethidium bromide and photographed using UV transillumination.

2.5 Determination of nucleotide sequence

Nucleotide sequence of the sfaS/focGH-specific PCR products was determined with direct sequencing using Sequencing Pro (Toyobo, Tokyo, Japan).

2.6 Statistical methods

Logistic regression analysis was used to evaluate the effect of the fimbriae on the probability of a strain to cause cystitis and pyelonephritis. Both the separate analysis assessing each of the fimbriae and multivariate analysis including all the fimbriae simultaneously were performed.

All statistical analyses were made using the SPSS software (SPSS Inc., Chicago, USA).

3 Results

3.1 Validation of sfaS- and focGH-specific PCR assay

Each of the sfaS- and focGH-specific primers yielded a single PCR product of the expected size (151 bp and 388 bp, respectively) for the corresponding positive control strains, HB101 (pANN801-13) and HB101 (pPIL110-54). When the two primer pairs were used together in combination, the results of the multiply primed PCR were consistent with the sum of the singly primed PCR (Fig. 1). To confirm that the amplification products obtained by this multiply primed PCR truly represented the expected sfaS and focGH sequences, the nucleotide sequence of each PCR product was determined. Each PCR product showed the identical sequence to the corresponding region of the sfaS or focGH gene, respectively (data not shown).

Figure 1

The results of multiply primed sfaS/focGH-specific PCR and sfa/foc PCR with three control strains and two clinical isolates from urinary tract infection. Lane designations are as follows: M, molecular mass markers (fX174 DNA digested by HaeIII); Ps, HB101 (pANN801-13); Pf, HB101 (pPIL110-54); NC, HB101 (pSH2); A and B, UTI isolates.

3.2 Prevalence of sfaS/focGH and papG variant genotypes

Prevalence of the sfaS/focGH and papG variant genes among E. coli strains isolated from acute UTIs and normal fecal flora is shown in Table 1. With regard to the sfa/foc operon, the results of colony hybridization test and multiply primed PCR were identical. Of 113 sfa/foc+ strains, however, seven cystitis and one fecal strains were negative with the sfaS- and focGH-specific PCR. No strains possessed the sfa and foc genes simultaneously. Prevalence of the sfa and foc genes was not significantly different between the cystitis and pyelonephritis strains. Of 206 strains which were pap+ with colony hybridization test, all but one were pap+ with multiply primed PCR and all but two were positive for at least one papG variant gene. The most frequent genotype was papG+IA2 (127 strains), followed by prsG+J96 (77 strains), and the papGJ96 gene was present only in a fecal strain. Among the pyelonephritis strains, 33 of 36 (92%) papGIA2 genes and 22 of 25 (88%) prsGJ96 genes existed as the only papG adhesin gene in the genome although their copy numbers could not be determined.

View this table:
Table 1

Prevarence of sfa/foc and pap genotypes of E. coli isolates from each category

MethodGenotypesAcute cystitis (n=194)Acute pyelonephritis (n=76)Normal fecal flora (n=80)
No. strains (%)No. strains (%)No. strains (%)
sfaS- and focGH-specific PCRsfa+ alone37(19)16(21)6(8)
foc+ alone27(14)16(21)3(4)
sfa+ or foc+64(33)32(42)9(11)
Colony hybridizationsfa/foc+71(37)32(42)10(13)
sfa/foc PCRsfa/foc+71(37)32(42)10(13)
papG variant-specific PCRpapG+IA2 alone65(34)33(43)8(10)
prsG+J96 alone48 (25)22 (29)7 (9)
positive for at least one papG variant125(64)58(76)21(26)
colony hybridizationpap+125(64)59(78)22(28)
pap PCRpap+125(64)59(78)21(28)
  • Numbers are cited from our previous studies [10, 11].

  • PCR with no variant-specific primers used in our previous studies.

3.3 Effect of fimbrial adhesin on the probability of a strain to cause cystitis and pyelonephritis

The results of statistical analysis were shown in Table 2. Except for the PrsGJ96 adhesin, separate and multivariate analysis showed the similar results in both cystitis and pyelonephritis. In the strains causing cystitis, F1C fimbria and AFA-I showed high relative virulence. In the strains isolated from pyelonephritis, only the PapGIA2 adhesin showed high relative virulence. S fimbria was associated with lower virulence in cystitis. The PrsGJ96 adhesin and type 1 fimbria did not show predictive value in each category.

View this table:
Table 2

Relative virulence (odds ratios and significance) of each fimbria

Cystitis (n=194) vs. fecal (n=80) strains
Separate analysisMultivariate analysis
factorsOdds ratioSignificanceOdds ratioSignificance
Pyelonephrititis (n=76) vs. fecal (n=80) strains
Separate analysisMultivariate analysis
factorsOdds ratioSignificanceOdds ratioSignificance

3.4 Relationship of sfaS/focGH and papG variant genotypes to O serogroups and other urovirulence determinant genes among uropathogenic E. coli

Table 3 shows the distribution of the sfaS/focGH and papG variant genotypes among ten O serogroups commonly encountered in uropathogenic E. coli. It is evident the O serogroups were related to the papG variant and sfaS/focGH genotypes. Thus the majority of the O1 and O16 strains possessed the papGIA2 gene, whereas 31 of 33 O18:K1:H7 strains were sfa+prsG+J96.

View this table:
Table 3

Relationship between O serogroup and genotypes of sfa/foc and papG variants in uropathogenic E. coli

O serogroupnsfa variantspapG variants
sfa+foc+papG+IA2 onlyprsG+J96 onlypapG+IA2 and prsG+J96

Then we investigated the relationship of the papG variant and sfa/foc genes to the hly and cnf1 genes (Table 4). Both hly and cnf1 genes were present in 87 of 270 (32%) uropathogenic E. coli strains. The hly+cnf1+ strains occupied 49 of 53 (92%) sfa+ strains, 74 of 85 (87%) prsG+J96 strains, 25 of 43 (58%) foc+ strains and 19 of 113 (17%) papG+IA2 strains. Of 46 sfa+prsG+J96hly+cnf1+ strains, 31 belonged to O18:K1:H7, six to O6, one each to O4 and O22, two and four were rough and nontypeable, respectively. In contrast, 63 of 113 (56%) papG+IA2 strains were lacking in all of the sfa, foc, prsGJ96, hly and cnf1 genes.

View this table:
Table 4

Relationship between sfa/foc genotypes and other virulence factor genotypes in uropathogenic E. coli

sfa/foc genotypespapGIA2prsGJ96hlycnf1No. of strains
sfa and foc++++3
  • + and − indicate that the indicated gene is present and lacking, respectively.

4 Discussion

For the purpose of molecular epidemiologic studies, PCR assay seems to be more useful than traditional DNA hybridization assay because of the simplicity, specificity and rapidity. Oligonucleotide primers should be deliberately designed to discriminate homologous genes such as P and S fimbrial adhesin variants. A multiply primed PCR assay for simultaneous detection of three papG variant genes recently developed by Johnson et al. [9] seems a useful tool for molecular epidemiologic studies. S and F1C fimbrial adhesins have been investigated by DNA hybridization [13] or multiply primed PCR [10, 12], both of which detect foc as well as sfa sequences. Phenotypic assay using monoclonal or polyclonal antibodies may be subject to a variation in expression under laboratory conditions [8, 13]. Therefore, the role of S and F1C fimbriae in UTIs has not been established. Our previous epidemiologic studies using multiply primed PCR as well as DNA hybridization test have not been intended to differentiate the three papG variant genes or between the sfa and foc operons [10, 12]. Recently, Sokolowska-Köhler et al. [7] described an oligonucleotide probe consisting of three oligonucleotides that bind specifically to sfa operon. However, F1C fimbrial determinants could not be detected directly with this method and hybridization procedure takes rather long time. In this study, we developed and validated a multiply primed PCR for differentiating the sfa and foc operons.

SfaS- and focGH-specific PCR showed that 59 (52%) of 113 sfa/foc+ strains were sfa+ and 46 (41%) were foc+, whereas eight (7%) showed negative results with this PCR assay. No strains possessed the sfa and foc genes simultaneously. Other researchers have recently reported that S fimbrial (sfa) operon has other genetically homologous variants such as sfr, sfaII and f1652 operons [1416]. Although the precise reason for this discordance between the results of various genotyping methods is unclear, such variant genes might be responsible for a part of the discordance. Schönian et al. [13] reported that 92 of 246 (37%) UTI isolates were sfa/foc+ and 61 (25%) were S-fimbriated by dot enzyme immunoassay using monoclonal antibody. Siitonen et al. [8] showed that 52 of 383 (14%) UTI isolates were F1C-fimbriated by immunofluorescence using monoclonal antibody. Sokolowska-Köhler et al. [7] reported that 83 of 261 (38%) UTI isolates were sfa/foc+ and 27 (12%) had sfa operon. Therefore, our result that 53 (20%) and 43 (16%) of 270 UTI isolates were genotypically sfa+ and foc+, respectively, are comparable to these reports. The logistic regression analysis indicated that F1C fimbria had higher relative virulence than S fimbria in cystitis. These data suggest that F1C fimbria, as well as S fimbria, may have some role in UTI.

PapG variant-specific PCR successfully discriminated the three papG variant genes in 204 of 206 (99%) pap+ strains. Therefore, our study confirmed this multiply primed papG PCR assay to be very useful for molecular epidemiologic purposes. The most frequent genotype was papG+IA2 (127 strains, 62%), followed by prsG+J96 (77 strains, 37%). The papGJ96 gene was present only in one fecal strain. Prevalence of the papGIA2 gene (class II variant) and prsGJ96 gene (class III variant) was not significantly different between the cystitis and pyelonephritis strains, unlike other studies showing that the papGIA2 gene predominates in pyelonephritis strains while the prsGJ96 gene is most frequent in cystitis strains [3]. However, the statistical analysis of our data showed that papGIA2 was associated with higher relative virulence than prsGJ96 in the strains isolated from pyelonephritis.

Another interesting finding from our genotypic study is the associations between the virulence determining genes. Forty-six of 53 (87%) sfa+ uropathogenic strains were prsG+J96hly+cnf1+. Furthermore, 31 of 33 (94%) O18:K1:H7 strains were sfa+prsG+J96hly+cnf1+. Compared with the sfa gene, however, the foc gene did not show such strong associations. Twenty-five of 43 (58%) foc+ strains possessed both the hly and cnf1 genes while 24 of 43 (56%) were papG+IA2. Neither foc+hly+cnf1+ strains or foc+papG+IA2 strains were associated with the specific serotypes. Recently, it has been shown that large segments of the chromosome called pathogenicity islands (PAIs) accommodate many virulence genes of uropathogenic E. coli[17, 18]. It is unclear, however, that the associations among the virulence determining genes in our results reflect such genetic linkage, functional association or epidemiological association.

In conclusion, the development of multiply primed PCR assay for detecting the sfa and foc genes should be useful for large-scale epidemiologic studies addressing the pathogenic significance of the S and F1C fimbrial adhesin variants.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
View Abstract