|Home | About | Journals | Submit | Contact Us | Français|
We characterized 67 Escherichia coli isolates with reduced susceptibility to cefotaxime or ceftiofur obtained from healthy broilers housed in five Italian farms. The blaCTX-M-1, blaCTX-M-32 and blaSHV-12 β-lactamase genes were identified on IncI1, IncN, or IncFIB plasmids. Considerable genetic diversity was detected among the extended-spectrum β-lactamase (ESBL)-producing isolates, and we identified indistinguishable strains in unrelated farms and indistinguishable plasmids in genetically unrelated strains. The detection of highly mobile plasmids suggests a potential animal reservoir for β-lactamase genes.
Escherichia coli resistant to extended-spectrum cephalosporins through the production of extended-spectrum β-lactamases (ESBLs) is isolated at an increasing rate among different hosts and countries, thus representing a potential public health risk (2, 3). Typically, ESBLs are borne on mobile genetic elements and can therefore spread both clonally and horizontally among different lineages of bacteria (5).
To determine the role of animals as potential reservoirs for ESBL-producing bacteria, there is a need for improved knowledge of the factors that promote dissemination of resistance determinants, including the role of the intestinal flora (4). In this study, we characterized a collection of 67 E. coli isolates showing reduced susceptibility to cefotaxime (MIC ≥ 2 μg/ml) or ceftiofur (MIC ≥ 8 μg/ml). The isolates were obtained during July 2007 by processing fecal samples collected by the “sock method” (1, 22) in a broiler-parent farm and four broiler farms located in northeastern Italy. The farms were ordinary with respect to farm management and biosecurity, thus implying that no contact among the animals reared in different farms could have taken place. One flock (i.e., a group of chickens sharing the same airspace) per farm was sampled (Table (Table1).1). All broiler flocks (here identified as flocks 1, 2, 4, and 5) hatched from eggs incubated in the same hatchery, and the parent flock (here identified as flock 6) was among the flocks that provided eggs to that hatchery. Indication of ESBL presence resulted from the ESBL phenotypic detection test performed with the disk method (7, 10). Amplification of the blaTEM, blaSHV, and blaCTX-M genes was done by PCR using primers and conditions previously described (15, 16) on whole-genome bacterial DNA (GenElute bacterial genomic DNA kit; Sigma-Aldrich, Brøndby, Denmark). PCR products were purified (exonuclease I, E. coli, and FastAP thermosensitive alkaline phosphatase; Fermentas, Helsingborg, Sweden) and sequenced (Macrogen Inc., Seoul, South Korea). Comparison of the nucleotide sequences and the derived amino acid sequences with previously described sequences (GenBank database http://www.ncbi.nlm.nih.gov/ and www.lahey.org/studies/webt.html, respectively) showed that the blaTEM-1, blaSHV-12, blaCTX-M-1, and blaCTX-M-32 genes were present in this collection (Fig. (Fig.1).1). The genetic relatedness of the E. coli isolates was determined by amplified fragment length polymorphism (AFLP) analysis as previously described (1, 6). Using a similarity cutoff value of ≥94% (1), 15 different clonal types were identified (A to O in Fig. Fig.1).1). Eighteen strains were selected for plasmid analysis by selecting one strain per clonal type from each flock. Within one AFLP type, however, two strains from the same flock were examined to determine the intraclonal diversity. Plasmid DNA (GeneJET plasmid miniprep kit; Fermentas, Helsingborg, Sweden) was used to transform electrocompetent Genehog E. coli (Invitrogen, Taastrup, Denmark). Transformants were selected on brain heart infusion (BHI) agar (Merck, Glostrup, Denmark) plates containing 2 μg/ml of cefotaxime and tested by colony PCR (15, 16) to confirm the presence of ESBL-encoding genes. Plasmid DNA (GeneJET plasmid miniprep kit; Fermentas, Helsingborg, Sweden) from the transformants was analyzed by restriction fragment length polymorphism (RFLP) analysis and PCR-based replicon typing (PBRT) (3). IncN (n = 10), IncI1 (n = 5), or IncFIB (n = 3) plasmids were identified (Fig. (Fig.11 and Table Table2).2). Twelve different plasmid profiles were visualized on a 0.8% agarose gel after enzymatic digestion with ClaI and PstI (FastDigest; Fermentas, Helsingborg, Sweden). Notably, indistinguishable band patterns were obtained in (i) two blaCTX-M-1 IncN plasmids found in two closely related E. coli strains (with 91.6% similarity according to the analysis of the AFLP patterns) from flock 6, (ii) five blaCTX-M-32 IncN plasmids found in five unrelated E. coli strains from flock 4, and (iii) two blaSHV-12 IncFIB plasmids found in two clonally related E. coli strains from flocks 1 and 2 (Fig. (Fig.11 and Table Table2).2). The disk diffusion test (9, 10) showed that all donors were resistant to at least two additional antimicrobial classes (Table (Table2).2). Seventeen (94%) isolates were resistant to tetracycline, and 15 (83%) isolates were resistant to nalidixic acid, 8 (44%) of which were ciprofloxacin resistant. Sulfonamide resistance alone or in combination with trimethoprim resistance was detected in 4 (22%) and 10 (55%) donor isolates, respectively. No resistance against amikacin, colistin, and imipenem was detected. Sulfonamide resistance occurred in seven (39%) transformants and was associated with trimethoprim resistance in one (5%) case (Table (Table2).2). The cotransfer of sul2 (n = 1), sul3 (n = 5), and sul1 and sul3 (n = 1) genes was determined in such transformants by the use of PCR as previously described (Table (Table2)2) (14). As measured by agar dilution (8), the MIC of ciprofloxacin for all the transformants was <0.03 μg/ml, indicating that possible qnr-mediated resistance to this fluoroquinolone was not cotransferred by the ESBL-producing plasmids.
We isolated ESBL-producing E. coli in flocks that did not receive treatments with cephalosporins (Table (Table1).1). Treatments with other antimicrobials (Table (Table1)1) could have coselected for the bacteria described. Although a limited number of flocks and bacterial isolates was included in the study, the genetic environment for the ESBL-encoding genes was highly variable. The blaSHV-12 gene was associated with three major clones (A, H, and I) detected in four flocks. These strains carried IncI1 (H and I) or IncFIB (A) plasmids. The blaCTX-M-1 gene and the closely related blaCTX-M-32 gene were mainly associated with IncN plasmids found in 10 different clonal types in three flocks. As clearly shown in flock 4 (Table (Table2),2), IncN plasmids were broadly distributed among genetically unrelated strains. The horizontal mobility of IncN plasmids has already been reported by Moodley and Guardabassi, who described the spread of a blaCTX-M-1 IncN plasmid among multiple E. coli lineages of porcine and human origin (18). The finding of IncN plasmids that could disseminate by horizontal transmission is worrisome. Our findings concerning the association of ESBL-encoding genes and plasmid replicons are similar to previous findings in humans, chickens, pigs, and dogs (5, 11, 12, 13, 17, 18, 19, 21, 23). The similarity between the ESBL variants we detected and those present in humans in Italy (20) indicates a possible role of poultry in the dissemination of these resistance determinants.
This study was supported by a grant from the Danish Poultry Council, the Graduate School FOOD, and the University of Copenhagen.
We thank Henrik Hasman and Arshnee Moodley for assistance with ESBL and plasmid characterization.
Published ahead of print on 25 January 2010.