The study included all patients with diarrhea who visited the Tropical Medicine Unit of Hospital Clinic in Barcelona, Spain, during 2005 and 2006. Patients with diarrhea that started during or shortly after (<5 days) a stay in a developing country were eligible. After the participants provided informed consent, clinical and epidemiologic data were collected.

Among all eligible participants, infection with EAEC and no other enteropathogen was found for 51. Of these 51 EAEC isolates, 5 from patients who had traveled to India were resistant to third-generation cephalosporins. Resistance phenotypes indicated ESBL production. MICs for antimicrobial agents and susceptibility class were determined by using the Clinical and Laboratory Standards Institute breakpoints guideline (Table 1). All strains were resistant to penicillins; second-, third-, and fourth-generation cephalosporins; and all β-lactamase–inhibitor combinations except piperacillin/tazobactam. Apart from β-lactam susceptibility, the strains showed resistance to other classes of antimicrobial agents, such as fluoroquinolones, tetracyclines, and monobactams (aztreonam). Positive amplification with specific primers and sequencing for the bla_CTX-M-15 gene provided positive genotypic confirmatory test results for ESBL production.

The epidemiologic relationships among the 5 strains were studied by repetitive sequence–based PCR, pulsed-field gel electrophoresis, and multilocus sequence typing (10,11). The PCR and pulsed-field gel electrophoresis genomic fingerprinting showed that the 5 strains were not epidemiologically related (Figure 1). However, multilocus sequence typing identified 2 clonal complexes: ST38 (3 strains) and ST10 (1 strain). The fifth strain could not be classified into any clonal complex (Table 2).

E. coli strains were classified into phylogenetic groups by multiplex PCR, described by Clermont et al. (12). The 3 strains in clonal complex ST38 belonged to the potentially virulent phylogenetic group D; the other 2 belonged to group B2 (Table 2).

A PCR method was used to detect genes encoding for typical EAEC virulence factors (2). These genes include aggA and aafA (encoding for adhesions); aap (for dispersin); aatA (for TolC); aggR (for regulation of aggregation); astA, set1A, and sen (for toxins), fyuA (for iron recruitment); agn43 (for antigen 43); and genes encoding for serine protease autotransporter toxins such as pet and sat. Gene aatA was detected in the 5 strains, whereas aap, aggR, and aggA had positive amplification for only 2 of the strains belonging to ST38. The other genes detected are shown in Table 2. EAEC was also identified by typical adherence to HEp-2 cells.

To determine the genetic environment of the bla_CTX-M-15 gene, we designed an inverse PCR. We designed the primers by studying the gene sequence and were directed outside the gene. The ISEcp1 insertion sequence was upstream from the bla_CTX-M-15 gene, which was also confirmed by PCR of the specific insertion sequence. To confirm the possible relationship between ISEcp1 and the resistance bla_CTX-M-15 gene_, we conducted a PCR with the forward primer for the ISEcp1 and the reverse primer for the bla_CTX-M-15 gene.

For plasmid extraction of the 5 isolates, we used the method of Kado and Liu (13). Only 3 strains had plasmids ranging from 93 kb to 170 kb (Figure 2, panel A). To confirm the absence of plasmids in the 2 strains, we conducted S1 digestion of the strains, resolving chromosomal DNA from plasmidic DNA. Southern blot of this digestion showed that the bla_CTX-M-15 gene was chromosomally located in these 2 strains, as was the aatA gene (usually found in the plasmid contained in EAEC strains) (data not shown). Finally, the location of the bla_CTX-M-15 gene in the 3 plasmid-containing strains was analyzed by using Southern blot from the plasmid extraction. The bla_CTX-M-15 gene was located in a plasmid in the 3 strains. The size of the plasmid containing CTX-M-15 varied in each strain (Figure 2, panel B). Plasmids with specific known molecular weight were used to provide a range of the size of the plasmids studied.

We identified several features concerning the molecular epidemiology of CTX-M-15–producing EAEC isolates in India. First, all strains belonged to phylogenetic groups D and B2, the 2 groups most commonly found with E. coli infections (14). Second, not finding ST131 suggests that ST131 might not be the most common ST among EAEC strains from India and that clonal complex ST38 might play a large role in causing infectious intestinal diseases. Third, the bla_CTX-M-15 gene is not only located in the plasmid but may also be in the chromosome. However, previous reports have shown that bla_CTX-M-15 is consistently linked with ISEcp1, which means that the chromosomal location might have originated from a previous plasmid location that was part of either a transposon or a cassette within an integron (9). It is also worth noting that the size of the plasmids containing the bla_CTX-M-15 gene was not the same in all strains, indicating that this gene may be located in different types of plasmids.

This evidence of widespread distribution and flexibility of the bla_CTX-M-15 gene highlights the need to develop appropriate means to control dissemination of this gene and associated resistance genes. Epidemiologic surveillance and correct use of antimicrobial agents will help prevent the steady increase of antimicrobial drug resistance worldwide.