Identification of antimicrobial resistance and class 1 integrons in Shiga toxin-producing Escherichia coli recovered from humans and food animals

Abstract

Introduction

The emergence of antimicrobial-resistant bacterial pathogens is a public health problem. The resistance phenotypes most commonly observed for Escherichia coli have been those to tetracycline, sulphonamides, ampicillin and streptomycin. However, resistance to clinically relevant, front-line antimicrobials such as fluoroquinolones, expanded-spectrum β-lactams, and third-generation cephalosporins has emerged among E. coli.^1,2 This in turn has led to antimicrobial treatment failure of infections caused by E. coli in humans, examples of which include fluoroquinolone-resistant bacteraemia in persons undergoing cancer chemotherapy³ and trimethoprim/sulfamethoxazole-resistant acute uncomplicated cystitis in women.⁴ Furthermore, the isolation of antimicrobial-resistant E. coli from the intestinal flora of healthy humans and animals, together with data which show E. coli readily transfer their plasmids to other E. coli strains⁵ and to strains of different genera (e.g. Hafnia alvei⁶ ), suggests commensal E. coli strains may be an important reservoir of transferable antimicrobial resistance genes. Shiga toxin-producing E. coli (STEC) have been an important cause of food-borne illness worldwide. Currently, research is urgently needed to determine the antimicrobial resistance profiles among STEC.

Integrons are DNA elements that may contain transferable antimicrobial resistance gene cassettes. At least five classes of integrons have been described, the majority of those from clinical isolates belonging to class 1. Class 1 integrons have been identified in generic E. coli strains recovered from food animals, water, food and humans. Relatively little is known about the occurrence of class 1 integrons in STEC.

The objective of this study was to identify antimicrobial resistance and class I integrons among STEC isolates recovered from humans and food animals.

Materials and methods

E. coli isolates

Two hundred and seventy-four STEC strains recovered from sick animals (n = 193) and human patients (n = 81) during 1985 to 2000 and consisting of serogroups O26 (n = 57), O103 (n = 72), O111 (n = 63), O128 (n = 7) and O157 (n = 75) were from the Gastroenteric Disease Center at The Pennsylvania State University. The isolates were grown on trypticase soy agar (Difco, Detroit, MI, USA) supplemented with 5% defibrinated sheep's blood (Becton Dickinson Microbiology Systems, Cockeysville, MD, USA) and stored in trypticase soy broth (Difco) containing 15% glycerol at −80°C.

Antimicrobial susceptibility determination

Antimicrobial susceptibility profiles were determined by broth microdilution with the PASCO MIC/ID (Becton Dickinson, Sparks, MD, USA) or Sensititre (Trek Diagnostic Systems, Westlake, OH, USA) systems. The following antimicrobials were included in the panels: cefoxitin, chloramphenicol, tetracycline, ceftriaxone, amoxicillin/clavulanic acid, sulfamethoxazole, gentamicin, trimethoprim/sulfamethoxazole and streptomycin. E. coli ATCC 25922, E. coli ATCC 35218, Enterococcus faecalis ATCC 51299 and Pseudomonas aeruginosa ATCC 27853 were used as quality controls. All experiments were conducted, and the results interpreted, according to guidelines of the National Committee for Clinical Laboratory Standards (NCCLS).⁷

Identification of class 1 integrons and associated resistance genes

DNA templates, oligonucleotide primers and PCR conditions (including positive and negative controls) to detect the presence of class 1 integrons have been described previously.⁶ All PCR amplicons were purified using a PCR purification kit (Boehringer-Mannheim, Indianapolis, IN, USA) and sequenced using an ABI Prism 3700 DNA analyzer (Applied Biosystems). DNA sequences were analysed by searching the GenBank database using the BLASTn algorithm (www.ncbi.nlm.nih.gov).

Conjugation

Integron-positive STEC strains CVM 9320, CVM 9574, CVM 9279, CVM 9530 and CVM 9514, each of which was nalidixic acid-susceptible, were used as donors in conjugation experiments. A nalidixic acid-resistant strain (CVM 19752), which did not contain an integron, was used as the recipient. Experiments were performed by filter mating.⁸ Transconjugants from each mating were characterized by antimicrobial susceptibility testing (see above). The presence of class 1 integrons among transconjugants was confirmed by PCR (see above).

Results and discussion

Ninety-three (34%) of the 274 STEC isolates were resistant to streptomycin, followed by 89 (32%) to sulfamethoxazole, 83 (30%) to tetracycline, 48 (18%) to ampicillin, 29 (11%) to cefalothin, 22 (8%) to trimethoprim/sulfamethoxazole, 18 (7%) to gentamicin, 13 (5%) to chloramphenicol, and 10 (4%) to cefoxitin. These results agree with those from previous studies from our laboratory in which E. coli, including STEC, have been found resistant to penicillins, sulphonamides and tetracyline.^1,2 Forty-three (16%) of the STEC isolates contained class 1 integrons (Table 1). The integrons ranged from 0.65 to 2.0 kb in length (Table 1), with a 1.0 kb gene cassette found in 41 (95%) of the integron-containing isolates. Integrons were found in isolates of serogroups O26, O103 and O111. Forty-one (95%) of the 43 class 1 integron-containing STEC isolates were resistant to one or more antimicrobial.

The aadA gene cassette, which putatively encodes resistance to streptomycin, was found in class 1 integrons from 41 (95%) of the 43 integron-containing STEC isolates. We observed structural variants of the aadA gene, including aadA1, aadA2, aadA7, aadA12, aadA21 and aad23b which matched the sequence data reported in GenBank with accession numbers listed in Table 1; however, none of these variants were novel. That aadA was found in STEC isolates of various serogroups suggests dissemination by a common genetic element. Eight of the aadA-containing integrons also carried the dfrA12 cassette, which putatively encodes resistance to trimethoprim. These findings are similar to results from previous studies indicating cassettes for aminoglycoside and trimethoprim resistance are common in class 1 integrons.^9,10 Three isolates, each from cattle, contained integrons that carried the sat1 gene cassette. sat1 encodes putative resistance to streptothricin, a drug that has never been approved for therapeutic use in the USA.

The presence of class 1 integrons did not necessarily correlate to antimicrobial resistance phenotype. For instance, though sulphonamide resistance is typically associated with the presence of class 1 integrons, four of the integron-containing STEC isolates (three of serogroup O111 and one of serogroup O26) were susceptible to sulfamethoxazole. sul1 may not have been present or expressed. Similarly, two isolates containing aadA were susceptible to streptomycin. The sequence data indicated that no frame-shift mutations were detected. It has been shown, however, that silent integron-borne aadA genes may be expressed when transferred to a new host by conjugation.⁶aadA is not known to be regulated by attenuation, thus, the differential expression may result from the lack of a transcriptional activator in the donor, or a repressor in the E. coli recipient strains. Three strains carried one gene cassette (aadA) yet were resistant to eight or more antimicrobials, suggesting elements besides class 1 integrons were involved in antimicrobial resistance among the STEC isolates. Seven strains carried two aadA genes, which may have contributed to high-level streptomycin resistance.

Each of five E. coli donor isolates transferred class 1 integrons to an E. coli recipient strain in vitro (Table 2). The observation that resistance phenotypes for ampicillin, chloramphenicol, cefalothin, gentamicin, tetracycline, trimethoprim, sulfamethoxazole and streptomycin were transferred suggests resistance genes in addition to aadA were present on the same plasmid or on a different plasmid that was co-transferred. One of the transconjugants (19752/9279) showed an MIC of cefalothin of ≥32 mg/L, which was greater than those of the donor (8 mg/L) and recipient (8 mg/L) strains. We believe that this might be due to the synergy with resident β-lactamases present in the recipient strain, since chromosome-encoded β-lactamases are present in most Gram-negative bacteria and these enzymes are often expressed at low levels. Another possibility might be due to the lack of a transcriptional activator in the donor, or a repressor in the E. coli recipient cells. Yu et al.⁹ have shown that among E. coli strains isolated from urinary tracts from patients in Korea, the prevalence of dfrA17 (putative trimethoprim resistance) was due to horizontal transfer of class 1 integrons through conjugative plasmids. Data demonstrating E. coli transfer plasmids to other genera, including Hafnia⁶ raise further concern about dissemination of class 1 integrons identified here. We conclude that class 1 integrons have facilitated emergence and dissemination of antimicrobial resistance among STEC in humans and food animals. The observation that the class 1 integrons identified here (i) contained multiple antimicrobial resistance cassettes and (ii) were horizontally transferable underscores the need for further research designed to limit the spread of these elements among bacterial populations.

The study was made possible by grant USDA/NRI 2000-02600 from the U.S. Department of Agriculture.

References

Author notes

1Division of Animal and Food Microbiology, Office of Research, Center for Veterinary Medicine, U.S. Food & Drug Administration, 8401 Muirkirk Road, Laurel, MD 20708, USA; 2Department of Nutrition and Food Science, University of Maryland, College Park, MD, USA; 3Gastroenteric Disease Center, The Pennsylvania State University, University Park, PA, USA