RESUMO
In this study, we tested the capability of enterobacterial repetitive intergenic consensus (ERIC) and repetitive extragenic palindromic (REP) polymerase chain reaction (PCR) to detect genetic diversity among Escherichia coli strains isolated from chickens bearing clinical signs of colibacillosis and compared the genotypes so obtained with the O:H serotypes and virulence of those strains. The DNAs from 50 avian E. coli strains and from E. coli ATCC 25922 were used to amplify ERIC and REP sequences. DNA from avian strains produced from 8 to 17 bands by ERIC-PCR and from 6 to 20 bands by REP-PCR; E. coli ATCC produced 11 bands by both methods. ERIC and REP-PCR showed good discriminating power, and the dendograms based on the different patterns revealed extensive genetic diversity among the avian strains. Those strains were allocated into four major clonal clusters, each one with 60% of similarity by ERIC and REP-PCR, and those clusters corresponded to strains with different degrees of pathogenicity. However, 56% of the pathogenic strains (28/50) belonged to two out of three major clonal clusters, and 86% of the nonpathogenic strains tended to group in one cluster and one subgroup. The 32 serotypes detected were distributed in all clusters, and within a serogroup, different DNA fingerprints were observed; however, strains with same serotypes tended to form clusters with similarity coefficients greater than 80%. These results suggest that no specific serotype and genotype is responsible for colibacillosis and that REP and ERIC-PCR are reproducible techniques that can improve the studies needed to clarify the pathways to the pathogenesis of colibacillosis.