• Human;
  • Dog;
  • Cat;
  • papG allele;
  • Escherichia coli


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results and discussion
  6. References

The distribution of alleles I, II and III of the P adhesin gene papG among Escherichia coli isolated from urinary tract infections in humans, dogs and cats was studied by PCR. Allele I was present in 6% and 5% of the human and cat isolates. Allele II as such was present in 30% and 22%, or in association with allele III in 12% and 2% of the human and canine isolates, respectively. Allele III was present in 33% of the human strains and predominated largely over allele II in E. coli isolates from cystitis of animal origin (72% in dog and 95% in cat strains). The three different classes of the PapG adhesin have been suggested to play a role in host specificity, for example human versus canine specificity. Recent studies, however, showed papG III positive human and dog cystitis isolates to be largely indistinguishable. We found the Class II allele in animal isolates and detected for the first time in Europe the Class I allele in a different genetic background than the J96-like clonal group. Our findings show that uropathogenic E. coli isolates from different species can have the same papG alleles and thus may have zoonotic potential.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results and discussion
  6. References

Escherichia coli is the major causative agent of urinary tract infection (UTI) in humans, dogs and cats [1]. Adherence to uroepithelial cells by these uropathogenic E. coli strains is an important mechanism in the pathogenesis of UTI. Of the adhesion factors involved, a central role has been demonstrated for P fimbriae [1,2].

P fimbriae mediate attachment to host urinary epithelial surfaces via binding to Gal (α1-4) Gal-containing isoreceptors on host tissues [1,3]. This binding is mediated by PapG adhesin molecules located at the fimbrial tips [4]. PapG occurs in three known molecular variants (Classes I–III). The three PapG adhesin molecules exhibit subtle differences in their preference for substituents in the consensus Gal (α1-4) Gal core receptor [5]. Differences in hemagglutination patterns have been used in epidemiological studies to classify wild-type E. coli strains according to their papG alleles [5,6]. A role for the three different classes of the PapG adhesin has been suggested in host specificity [3]. However, recently papG III positive human and dog cystitis isolates were found to be largely indistinguishable [7,8]. Also the type of UTI caused by the infecting strain has been suggested to be determined by the PapG adhesin. For example, data from the literature associate acute cystitis with the papG allele III and acute pyelonephritis with allele II [9–11]. The exclusive involvement of allele II in E. coli bacteremia has been suggested also [11]. However, in subsequent studies a substantial prevalence of allele III was found as well, independent of the source of bacteremia [12].

At present only papG III alleles have been reported in canine E. coli uropathogenic isolates [3,7,8]. The presence of other papG variants in isolates of animal source and of the combination of papG Class I and III in human wild-type E. coli strains in Europe is unknown. The minor subunit proteins E and F, present in the tip of P fimbriae, have been shown to exhibit binding capability for certain matrix proteins [13]. Polymorphisms in the papE and papF genes of the P fimbrial operon therefore may also affect the adherence phenotype. However, to our knowledge this has not been described.

In view of the proposed potential zoonotic properties of animal UTI isolates, the present study to define the distribution of papG alleles among E. coli isolates from humans, dogs and cats was undertaken. Confirmation of this supposed zoonotic potential might have influence on the use of antibiotics to treat animal UTI.

2Materials and methods

  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results and discussion
  6. References

2.1Bacterial strains

One hundred and six uropathogenic E. coli strains shown to possess a P fimbrial operon were selected from three previously characterized strain collections [14,15]. These collections comprised 208 uropathogenic E. coli strains isolated from humans, dogs and cats with significant bacteriuria and clinical symptoms of UTI. The first collection comprised 90 uropathogenic E. coli isolated at the National Institute of Health in Lisbon from community-acquired UTIs in human patients, including children and adults of both sexes. The second collection comprised 68 E. coli strains isolated from samples collected at the Veterinary Teaching Hospital of the Faculty of Veterinary Medicine and at veterinary practices in the Lisbon area. The third collection comprised 25 dog and 25 cat hemolytic E. coli strains isolated at the Veterinary Microbiological Diagnostic Center of Utrecht University [15]. The detection of the P fimbrial operon in the 208 uropathogenic E. coli strains was done by PCR amplification of a 336-bp fragment between the papE and papF genes [14,15]. Fifty-one dog, 33 human and 22 cat papEF+E. coli strains isolated from human, canine and feline patients with UTI were examined in this study (Table 1). Control strains for the papG multiplex were prototype uropathogenic E. coli strains J96 and IA2 [16,17].

Table 1.  Characterization of E. coli strains causing UTI in humans, dogs and cats
Host speciesE. coli (n)papEF positive strains (n)Host-associated clinical syndrome
  1. aNo information was available concerning the nature of the localization of UTI.


2.2DNA preparation

Total DNA extraction was performed by a rapid boiling procedure as described [15].

2.3Multiplex PCR for papG alleles

Genotypes of the papG allele were determined using a multiply primed papG PCR [18], with modifications. Amplification was done in a 25-μl reaction mixture containing template DNA (2 μl of the boiled lysate), 4 mM MgCl2, 0.2 mM each of the four dNTPs, 0.45 μM of each primer (j96-193f, j96-653r; ia2-383f, ia2-572r; prs-198f, prs-455r), and 2.5 U of Taq DNA polymerase in 1× PCR buffer (MBI Fermentas, BioPortugal, Portugal). The reaction mixture was overlaid with 20 μl of mineral oil. PCR amplification comprised the following steps: heating at 94°C for 3 min; 25 cycles of denaturation at 94°C for 1 min, annealing and extension at 72°C for 4 min; and a final extension step for 10 min at 72°C (Mastercycler, Eppendorf, Germany). Ten μl of the reaction mixtures were analyzed by electrophoresis for the amplicons obtained, as described [15]. Negative (sterile water) and positive controls were included in each assay.

3Results and discussion

  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results and discussion
  6. References

The papG gene was detected in the multiplex PCR in 27 (82%), 49 (96%) and 22 (100%) papEF positive uropathogenic E. coli strains isolated from humans, dogs and cats, respectively (Table 2). Six (18%) papEF positive strains of human origin and two (4%) of canine origin were negative in the PCR for the papG gene. The six papEF+/papG human strains were found to be positive in the papC PCR (J. Machado, unpublished data), which is an additional indication for the presence of an intact P fimbrial operon. Of these six strains, four were negative in the F-type-specific papA PCR [15]. The remaining two strains were positive for the fimbrial serotypes F13 and F15, and F12 and F13, respectively (C. Féria, unpublished data). In the strains where both the major subunit gene (papA) and the adhesin gene (papG) PCR are negative, truncated P fimbrial operons may be present. Since we did not perform Southern hybridization experiments it is also possible that these strains have a papA and papG allele which is unknown until now and to which our primers do not anneal. One papEF+/papG dog isolate was also found to be F-type negative. A papA allele specific for the F13 serotype was detected in the second papEF+/papG canine strain [15].

Table 2.  Distribution of papG alleles among papEF positive E. coli strains causing UTI in humans, dogs and cats
Host speciespapEF positive E. coli (n)Number (%) of strains with papG allele(s):
  I onlyII onlyII+IIIIII onlyAny IIaAny IIIaNone
  1. aSums of ‘any II’ plus ‘any III’ are greater than total number of papG positive strains, since strains with both alleles II and III are counted both as ‘any II’ and ‘any III’.

Humann=332 (6)10 (30)4 (12)11 (33)14 (42)15 (45)6 (18)
Dogn=51011 (22)1 (2)37 (72)12 (23)38 (74)2 (4)
Catn=221 (5)0021 (95)021 (95)0

The Class I allele is associated with the J96-like clonal group of E. coli strains of serotype O4:H5:F13 [19–21]. In these strains the Class I allele is always found together with a Class III allele and both alleles are associated with a papA molecule of the F13 serotype [19,20]. Among our strains, the Class I allele was detected in two (6%) and one (5%) of the human and cat strains, respectively. It was not detected in canine E. coli isolates. The two papG Class I isolates of human source did not have other papG alleles nor had they hemolytic properties (either genotypic or phenotypic). Furthermore, one of these two isolates was F-non-typable and the other had an F16 serotype allele. The papG Class I allele positive feline strain did also not have other papG alleles. However it did have other urovirulence genes (sfaDE+, cnf1+ and hlyA+) and a papA gene of the F14 serotype [15]. Thus, these papG Class I alleles are present in a different genetic background than that characteristic for the J96-like clonal group [19,20]. These findings indicate that the Class I papG variant is not unique to the prototype strain J96 and the J96-like clonal group and can occur in non-hemolytic isolates of human origin (not associated with the J96 pathogenicity island VI) [22], and can also occur in isolates of animal origin.

Ten human strains (30%) and 11 (22%) canine strains had allele II only. Allele II in combination with allele III was found in four (12%) and one (2%) isolate, respectively. Allele II was absent in feline E. coli strains. Allele III predominated largely over allele II in E. coli isolates of animal origin, occurring in 37 (72%) of the canine isolates and 21 (95%) of the feline isolates. E. coli strains isolated from human hosts were positive for allele III in 11 (33%) isolates.

In humans, allele II predominates over allele III (71% of the strains versus 7%) in isolates from cases of urosepsis. Allele III is significantly associated with urosepsis and host-compromised in humans [23]. Our collection of human E. coli strains unfortunately was not characterized with respect to the clinical syndromes of the hosts, which may explain the relatively equal distribution observed for alleles II and III only (30% and 33%, respectively). Detection of both alleles II and III in one strain occurred in four of our isolates of human origin (12%). A similar association has been reported for a collection of urosepsis isolates in 1% of the strains and in a collection of E. coli isolates from unselected bacteremia in 5% of the strains [12,23].

The two E. coli collections of animal origin were largely from lower UTI (cystitis). The predominance of allele III in animal source isolates (72% in dogs and 95% in cats) is in agreement with the data that associate cystitis with papG allele III in women (74% of strains) [9].

At present, only papG class III has been described in uropathogenic E. coli of canine origin. In this study we detected the presence of the Class II gene also in canine uropathogenic E. coli strains. We also found Classse I and III of the papG gene in E. coli causing UTIs in cats. The Class I gene was detected in our study in human E. coli isolates with a different genetic background than that of the characteristic J96 prototype uropathogenic strain and the J96-like clonal group.

We have shown that uropathogenic E. coli isolates from different species share common papG alleles. These findings are in contrast to earlier studies on the distribution of the different papG alleles which lead to the concept of host specificity for human and dog E. coli isolates from UTI [4]. Our findings suggest that human, canine and feline UTI E. coli strains are highly similar. Therefore it cannot be excluded that UTI E. coli isolates from pets are a potential zoonotic risk. This new concept has important implications for disease prevention and antimicrobial resistance surveillance in veterinary practice


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results and discussion
  6. References
  • [1]
    Donnenberg, M.S. and Rodney, A.W. (1996) Virulence determinants of uropathogenic Escherichia coli. In: Urinary Tract Infections, Molecular Pathogenesis and Clinical Management (Mobley, H.L.T. and Warren, J.W., Eds.), pp. 135–174. ASM Press, Washington, DC.
  • [2]
    Johnson, J.R. (1991) Virulence factors in Escherichia coli urinary tract infection. Clin. Microbiol. Rev. 4, 80128.
  • [3]
    Marklund, B.-I., Tennent, J.M., Garcia, A.H., Båga, M., Lindberg, F., Gaastra, W., Normark, S. (1992) Horizontal gene transfer of the Escherichia coli pap and prs pili operons as a mechanism for the development of tissue specific adhesive properties. Mol. Microbiol. 6, 22252242.
  • [4]
    Kuehn, M.J., Heuser, J., Normark, S., Hultgren, S.J. (1992) P pilli in uropathogenic E. coli are composite fibres with distinct fibrillar adhesive tips. Nature 356, 252255.
  • [5]
    Strömberg, M., Marklund, B.-I., Lund, B., Ilver, D., Hamers, A., Gaastra, W., Karlsson, K.A., Normark, S. (1990) Host-specificity of uropathogenic Escherichia coli depends on differences in binding specificity to Galα 1-4Gal-containing isoreceptors. EMBO J. 9, 20012010.
  • [6]
    Jones, C.H., Dodson, K. and Hultgren, S.J. (1996) Structure, function, and assembly of adhesive P pili. In: Urinary Tract Infections, Molecular Pathogenesis and Clinical Management (Mobley, H.L.T. and Warren, J.W., Eds.), pp. 175–219. ASM Press, Washington, DC.
  • [7]
    Johnson, J.R., O'Bryan, T.T., Low, D.A., Ling, G., Delavari, P., Fasching, C., Russo, T.A., Carlino, U., Stell, A.L. (2000) Evidence of commonality between canine and human extraintestinal pathogenic Escherichia coli strains that express papG allele III. Infect. Immun. 68, 33273336.
  • [8]
    Johnson, J.R., Stell, A.L., Delavari, P., Murray, A.C., Kuskowski, M., Gaastra, W. (2001) Phylogenetic and pathotypic similarities between Escherichia coli isolates from urinary tract infections in dogs and extraintestinal infections in humans. J. Infect. Dis. 183, 897906.
  • [9]
    Johnson, J.R., Russo, T.A., Brown, J.J., Stapleton, A. (1998) papG alleles of Escherichia coli causing first episode or recurrent cystitis in adult women. J. Infect. Dis. 177, 97101.
  • [10]
    Johanson, I.M., Plos, K., Marklund, B.I., Svanborg, C. (1993) pap, papG and prsG DNA sequences in Escherichia coli from fecal flora and the urinary tract. Microb. Pathog. 15, 121129.
  • [11]
    Otto, G., Sandberg, T., Marklund, B.I., Ulleryd, P., Svanborg, C. (1993) Virulence factors and pap genotype in Escherichia coli isolates from women with acute pyelonephritis, with or without bacteremia. Clin. Infect. Dis. 17, 448456.
  • [12]
    Johnson, J.R., Brown, J.J., Maslow, J.N. (1998) Clonal distribution of the three alleles of the Gal(α1-4)Gal-specific adhesin gene papG among Escherichia coli strains from patients with bacteremia. J. Infect. Dis. 177, 651661.
  • [13]
    Westerlund, B., van Die, I., Kramer, C., Kuusela, P., Holthofer, H., Tarkkanen, A.M., Virkola, R., Riegman, N., Bergmans, H., Hoekstra, W. et al (1991) Multifunctional nature of P fimbriae of uropathogenic Escherichia coli: mutations in fsoE and fsoF influence fimbrial binding to renal tubuli and immobilized fibronectin. Mol. Microbiol. 5, 29652975.
  • [14]
    Féria, C., Correia, J.H., Gonçalves, J. and Machado, J. (2000) Detection of virulence factors in uropathogenic Escherichia coli isolated from humans, dogs and cats in Portugal. In: Genes and Proteins Underlying Microbial Urinary Tract Virulence (Emödy, L. et al., Eds.), pp. 305–308. Kluwer Academic/Plenum Publishers.
  • [15]
    Féria, C., Machado, J., Duarte Correia, J.H., Gonçalves, J., Gaastra, W. (2001) Virulence genes and P fimbriae PapA subunit diversity in canine and feline uropathogenic Escherichia coli. Vet. Microbiol. 82, 8189.
  • [16]
    Low, D.A., Lark, D., Schoolnik, G., Falkow, S. (1984) Gene clusters governing the production of hemolysin and mannose-resistance hemagglutination are closely linked in Escherichia coli serotype O4 and O6 isolates from urinary tract infections. Infect. Immun. 43, 353358.
  • [17]
    Clegg, S. (1982) Cloning of genes determining the production of mannose-resistant fimbriae in a uropathogenic strain of Escherichia coli belonging to serogroup O6. Infect. Immun. 38, 739744.
  • [18]
    Johnson, J.R., Brown, J.J. (1996) A novel multiply primed polymerase chain reaction assay for identification of variant papG genes encoding the Gal(α1-4)Gal-binding PapG adhesins of Escherichia coli. J. Infect. Dis. 173, 920926.
  • [19]
    Johnson, J.R., Russo, A.T., Scheutz, F., Brown, J.J., Zhang, L., Palin, K., Rode, C., Bloch, C., Marrs, C.F., Foxman, B. (1997) Discovery of disseminated J96-like strains of uropathogenic Escherichia coli O4:H5 containing genes for both PapGJ96 (class I) and PrsGJ96 (class III) Gal(α1-4)Gal-binding adhesins. J. Infect. Dis. 175, 983988.
  • [20]
    Johnson, J.R., Stapleton, A.E., Russo, T.A., Sheutz, F., Brown, J.J., Maslow, J.N. (1997) Characteristics and prevalence within serogroup O4 of a J96-like clonal group of uropathogenic Escherichia coli O4:H5 containing the class I and class III alleles of papG. Infect. Immun. 65, 21532159.
  • [21]
    Johnson, J.R., Orskov, I., Orskov, F., Goullet, P., Picard, B., Moseley, S.L., Roberts, P.L., Stamm, W.E. (1994) O, K, and H antigens predict virulence factors, carboxylesterase B pattern, antimicrobial resistance, and host compromise among Escherichia coli causing urosepsis. J. Infect. Dis. 169, 119126.
  • [22]
    Swenson, D.L., Bukanov, N.O., Berg, D.E., Welch, R.A. (1996) Two pathogenicity islands in uropathogenic Escherichia coli J96: cosmid cloning and sample sequencing. Infect. Immun. 64, 37363743.
  • [23]
    Johnson, J.R. (1998) papG alleles among Escherichia coli strains causing urosepsis: associations with other bacterial characteristics and host compromise. Infect. Immun. 66, 45684571.