Abstract
The main aim of this study was to determine the frequency of antibiotic resistance among Escherichia coli isolates recovered in Levine agar plates from 54 fecal samples of captive ostriches from a farm in the South of Portugal. Fifty-four nonselected E. coli isolates were obtained (one/sample) and the phenotypes and genotypes of antibiotic resistance were characterized. The following numbers of isolates showed antibiotic resistance: ampicillin (nine), tetracycline (seven), streptomycin (three), amoxicillin–clavulanic acid, cefoxitin, or gentamicin (one), and cefotaxime, ceftazidime, azthreonam, imipenem, nalidixic acid, ciprofloxacin, and trimethoprim/sulfamethoxazole (zero). The bla TEM gene was identified in six out of nine ampicillin-resistant isolates, and the tet(A) or tet(B) genes in five out of seven tetracycline-resistant isolates. Mutations at positions −42, −18, −1, and +58 of ampC promoter region were identified in one cefoxitin-resistant isolate. Further, the occurrence of extended-spectrum beta-lactamase (ESBL)–producing E. coli isolates was estimated in the 54 fecal samples of ostriches using cefotaxime-supplemented Levine agar plates for ESBL-positive E. coli recovery. Three samples contained ESBL-positive E. coli isolates of which one isolate/sample was characterized, leading to the detection of the following beta-lactamases: bla CTX-M-14a + bla TEM-1b (two isolates) and bla TEM-52c (one isolate). The three ESBL-positive isolates were classified into the phylogroup B1, and contained class 1 integrons with the gene cassettes dfrA17 + aadA5 (one isolate) and aadA1 (two isolates). This study adds to our knowledge about the wide dissemination of ESBL-producing E. coli isolates in different ecosystems, including captive ostriches, that could be transferred to humans through the food chain.
Introduction
E
In Portugal there are three farms that produce and commercialize ostriches, and there has been an increasing demand of this type of meat in the last few years. The purpose of this study was to estimate the frequency of antibiotic resistance in fecal E. coli isolated from captive ostriches in one of these farms. In addition, the mechanisms of resistance and the occurrence of ESBL-positive E. coli isolates and their genetic characteristics in these fecal samples were determined. The frequency and type of integrons in all of these isolates will be investigated in this study.
Materials and Methods
Fifty-four fecal samples of captive ostriches were recovered from one farm from Alentejo, in the South of Portugal. At this time of the study it included 24 breeding trio of ostriches and 3 different individual parks with young ostriches of different ages. Thus, during sample collection, there were 84–87 individual animals in the farm. Two fecal samples were obtained from each of the 27 breeding and individual parks, making a total collection of 54 samples. Samples were seeded on two types of media: (i) Eosin-Methylene-Blue-agar (Levine agar; Oxoid Limited, Basingstoke, United Kingdom) for recovery of E. coli isolates; (ii) Levine agar supplemented with cefotaxime (2 μg/mL) (Levine-CTX) for recovery of ESBL-positive E. coli isolates. Plates were incubated for 24 h at 37°C, and colonies with typical E. coli morphology were selected and identified by classical biochemical methods and by the API 20E system (BioMérieux, La Balme Les Grottes, France).
Antimicrobial susceptibility testing was carried out by the agar disk diffusion method on Müeller–Hinton agar plates according to CLSI standard criteria (CLSI, 2009). The following antibiotics were tested: ampicillin, amoxicillin–clavulanic acid, cefoxitin, ceftazidime, cefotaxime, aztreonam, imipenem, gentamicin, amikacin, tobramycin, streptomycin, nalidixic acid, ciprofloxacin, trimethoprim/sulfamethoxazole (SXT), tetracycline, and chloramphenicol.
The presence of bla TEM was checked in ampicillin-resistant isolates. Cefotaxime-resistant isolates were screened for ESBL production according to the CLSI criteria (CLSI, 2009). The minimal inhibitory concentration (MIC) to cefotaxime and ceftazidime was determined in all ESBL-positive E. coli isolates. The presence of genes encoding for TEM-, SHV-, CTX-M-, and CMY-type β-lactamases was studied by polymerase chain reaction (PCR) in all ESBL-positive isolates (Poeta et al., 2008; Vinué et al., 2008). The resulting DNA amplicons were sequenced on both strands, and sequences were compared with those included in the GenBank database to identify the specific β-lactamase gene. The genetic environment of bla CTX-M genes was also tested by PCR and sequencing (Eckert et al., 2006). The following genes were also studied by PCR: tet(A) and tet(B) in tetracycline-resistant isolates, and sul1, sul2, and sul3 in SXT-resistant isolates (Sáenz et al., 2004). The presence of intI1 and intI2 genes, encoding class 1 and 2 integrases, and their variable region were also analyzed by PCR and sequencing (Sáenz et al., 2004). Positive and negative controls from the bacterial collection of the University La Rioja (Spain) were used in all assays.
The major phylogenetic groups of the E. coli isolates were identified by PCR, using a combination of three genes: chuA, yjaA, and TspE4.C2 (Clermont et al., 2000). Genes encoding virulence factors, often found in pathogenic E. coli, namely, stx1-stx2 (shiga toxin 1 and 2), fimA (encoding type 1 fimbriae), papG allele III (adhesin PapG class III), cnf1 (cytotoxic necrotizing factor), papC (P fimbriae), and aer (aerobactin iron uptake system), were amplified by PCR as previously reported (Bastian et al., 1998; Ruiz et al., 2002), and if the E. coli isolates corresponded to the serotype O157 we used an allele-specific PCR of the 5′ region of rfb locus (Clermont et al., 2007).
The clonal relationship among E. coli isolates was determined by pulsed-field gel electrophoresis using XbaI enzyme as previously described (Sáenz et al., 2004). XbaI macrorestriction patterns were visually analyzed and interpreted according to previously reported criteria (Tenover et al., 1995).
Results and Discussion
Antibiotic resistance in E. coli isolates recovered in non-antibiotic-supplemented Levine agar plates
Fifty-four fecal E. coli isolates were obtained in nonsupplemented agar media from the 54 samples (one isolate/sample), and they were tested for antibiotic resistance. The numbers of isolates that showed antibiotic resistance were as follows: ampicillin (nine), tetracycline (seven), streptomycin (three), amoxicillin–clavulanic acid, cefoxitin, and gentamicin (one), and none of the isolates were resistant to cefotaxime, ceftazidime, azthreonam, imipenem, nalidixic acid, ciprofloxacin, and SXT. Higher percentages of tetracycline and streptomycin resistance have previously been detected in E. coli isolates in the intestinal track of ostrich carcasses in a separate study (Ley et al., 2001). The bla TEM gene was identified in six of our nine ampicillin-resistant isolates, and the tet(A) or tet(B) gene in five of seven tetracycline-resistant isolates. The promoter and attenuator region of the chromosomal ampC gene was amplified by PCR and sequenced in the cefoxitin-resistant isolate, being later compared with the same region of the E. coli K12 ampC gene. Mutations at positions −42 (C→T), −18 (G→A), −1 (C→T), and +58 (C→T) were detected in that isolate, and resistance could be associated with hyperproduction of chromosomal AmpC β-lactamase, as reported by other authors (Mulvey et al., 2005; Vinué et al., 2008). This isolate also contained a class 1 integron with the aadA gene cassette in its variable region.
Occurrence of ESBL-positive E. coli isolates in the fecal samples
Cefotaxime-resistant E. coli isolates were not detected in the studied fecal samples when the recovery of isolates was performed in non-antibiotic-supplemented agar media. To determine the occurrence of ESBL-positive E. coli isolates in the fecal samples of ostriches, these samples were also inoculated into cefotaxime-supplemented Levine agar plates to detect cefotaxime-resistant isolates, even if they were in a low proportion respect to the total E. coli fecal population (probably undetectable in nonsupplemented media). Cefotaxime-resistant E. coli isolates were recovered on Levine-CTX plates in 3 of the 54 analyzed fecal samples of ostriches (5.6%), and the 3 E. coli isolates obtained on these samples were ESBL positive. These isolates exhibited resistance to nalidixic acid, ciprofloxacin, tetracycline, and streptomycin, in addition to beta-lactams, and two of them also to chloramphenicol or STX (Table 1). They showed high MIC values to cefotaxime (≥128 μg/mL) and moderate resistance to ceftazidime (MIC 1–16 μg/mL). The following genes encoding ESBLs were detected in these isolates: bla CTX-M-14a + bla TEM-1b (two isolates) and bla TEM-52c (one isolate). The presence of ISEcp1 and IS903 surrounding the bla CTX-M-14a gene was identified in the two isolates that contained this gene, being this genetic environment previously reported (Poeta et al., 2008; Vinué et al., 2008). The CTX-M-14 and TEM-52 beta-lactamases have also been previously identified in E. coli isolates from different types of animals, as food-producing animals and wild animals in Portugal (Poeta et al., 2008; Costa et al., 2009), as well as in other countries (Carattoli, 2008).
MIC, minimal inhibitory concentration; ESBL, extended-spectrum beta-lactamase; CAZ, ceftazidime; CTX, cefotaxime; NAL, nalidixic acid; CIP, ciprofloxacin; TET, tetracycline; STR, streptomycin; SXT, trimethoprim/sulfamethoxazole; CHL, chloramphenicol.
The presence of class 1 integrons was confirmed in the three ESBL-positive isolates (they contained the intI1, qac EΔ1, and sul1 genes), and the following gene cassettes were identified in the variable region of these integrons: aadA1 (two isolates) and dfrA17 plus aadA5 (one isolate). These genes are associated with trimethoprim (dfrA17) and streptomycin (aadA1 or aadA5) resistance, and similar structures have also been previously reported in human and animal E. coli isolates (Vinué et al., 2008). The three ESBL-positive isolates were classified into the phylogenetic group B1, and contained the aer and fimA virulence genes. The three isolates showed unrelated pulsed-field gel electrophoresis patterns when their chromosomal DNA were digested with XbaI enzyme, indicating the absence of clonal relationship among them. The CTX-M-14-type beta-lactamase has been frequently detected in E. coli isolates implicated in human infections and in some animal species in Portugal and in its neighboring country Spain, in E. coli isolates implicated in human infections, and in some types of animals (Mendonça et al., 2007; Vinué et al., 2008; Costa et al., 2009). Another remarkable finding of this study was the detection of the gene encoding for TEM-52, which is a frequently detected ESBL in humans and animals in the referred country (Machado et al., 2008; Poeta et al., 2008; Costa et al., 2009). The phylogenetic group B1 is usually associated with commensal bacteria (Johnson et al., 2005), as is the case with the studied microorganisms that were recovered from feces of food-producing animals.
Our results show a wide variety of resistance genes, including the presence of ESBLs and the inclusion of resistant genes in integrons, in multiresistant nonpathogenic E. coli recovered from ostriches. These results add to our knowledge about the dissemination of ESBL-positive bacteria in non-hospital-associated environments, which could have implications in human health because these resistant bacteria can be transferred to humans through the food chain. Only one farm and a limited number of E. coli isolates were included in this study, which is a limitation; thus, more extensive studies should be carried out in the future to know whether this situation is similar in other farms in different countries. In addition, ostrich meat should be tested in the future for the presence of resistant bacteria.
Footnotes
Acknowledgments
We are grateful to the staff of Ostrichland Producing Ostrichs for helping us in sample collection.
Disclosure Statement
No competing financial interests exist.