Characterization of bla_SHV Genes on Plasmids from Escherichia coli and Salmonella enterica Isolates from Canadian Food Animals (2006-2007)

Abstract

bla_SHV genes from Escherichia coli and Salmonella enterica isolates from chicken (n = 19) and pork (n = 1) were identified as bla_SHV-2 (n = 5) or bla_SHV-2a (n = 15). Eighteen were on plasmids of the incI1 (n = 15), incP (n = 2), and incFIB (n = 1) incompatibility groups. These plasmids were all transferable by conjugation between E. coli and S. enterica.

TEXT

Although SHV enzymes remain among the major extended-spectrum β-lactamases (ESBLs) in bacteria from humans (1), data on their prevalence in bacteria from food animals are sparse (2), particularly in North America. Broad-spectrum cephalosporin resistance seen in bacteria from farm animals in North America, and Canada in particular, has been dominated by cephamycinase resistance determinants (3, 4). However, in addition to numerous bla_CMY-2-positive isolates, we recently detected Escherichia coli and Salmonella enterica isolates carrying bla_SHV genes in Canadian farm animals. Therefore, the objective of this study was to investigate bla_SHV and associated plasmids from these E. coli and S. enterica isolates, obtained by the Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS) in 2006 to 2007 (5, 6), to provide a baseline for comparison with isolates from later years and of human origin.

PCR screening for bla_SHV (7) was performed on 675 E. coli isolates (448, 187, and 40 isolates of chicken, porcine, and bovine origins, respectively) and 205 Salmonella enterica isolates (142, 63 isolates of chicken and porcine origins, respectively) resistant to ampicillin recovered by CIPARS in 2006 and 2007 from retail meat and abattoir cecal samples across Canada. Eighteen E. coli isolates (2.7%) of chicken (n = 17) and porcine (n = 1) origins and two S. enterica isolates (0.9%) of chicken origin were positive. These 20 isolates were screened for the presence of bla_TEM and bla_CMY-2 (7). One Salmonella isolate (SA18) was positive for bla_TEM. The isolates were serotyped at the Laboratory for Food-Borne Zoonoses (LFZ; Guelph, Ontario, Canada) and tested in duplicate on different days by broth microdilution for susceptibility to a panel of 27 antimicrobial agents, including 16 β-lactams (8). Each isolate belonged to a different serotype. The two S. enterica isolates had the highest MICs for most β-lactams (Table 1).

Table 1.

β-Lactam MICs in the original E. coli and Salmonella isolates and in E. coli DH10B transformants and Salmonella Typhimurium LB500 transconjugants

β-Lactama	MIC (μg/ml) forb:
	Original isolates		Transformants		Original isolates		E. coli transformants	Salmonella Typhimurium transconjugants
	blaSHV-2	blaSHV-2a	blaSHV-2	blaSHV-2a	E. coli	Salmonella
CZO	20 (≤8–32)	16 (≤8–32)	16 (≤8–32)	32 (16–32)	16 (≤8–32)	32 (32–32)	32 (≤8–32)	32 (≤8–32)
FEP	≤1 (≤1–≤1)	≤1 (≤1–2)	≤1 (≤1–2)	≤1 (≤1–≤1)	≤1 (≤1–≤1)	2 (2–8)	≤1 (≤1–2)	≤1 (≤1–2)
CTX	0.75 (0.25–4)	1 (0.25–8)	0.75 (0.25–1)	2 (1–2)	1 (0.25–4)	8 (4–16)	1 (0.25–2)	4 (0.5–4)
FOX	≤4 (≤4–16)	≤4 (≤4–8)	8 (≤4–8)	8 (≤4–16)	≤4 (≤4–16)	≤4 (≤4–≤4)	8 (≤4–16)	≤4 (≤4–≤4)
CPD	4 (1–16)	4 (1–>32)	6 (4–8)	12 (4–32)	4 (1–16)	32 (32–64)	8 (4–32)	16 (4–32)
CAZ	0.5 (0.25–1)	0.5 (0.25–2)	1 (0.25–2)	2 (1–2)	0.5 (0.25–2)	4 (2–8)	2 (0.5–2)	2 (0.5–4)
CCV	≤0.12 (≤0.12–≤0.12)	≤0.12 (≤0.12–0.25)	0.25 (0.25–0.25)	0.25 (0.25–0. 5)	≤0.12 (≤0.12–0.25)	0.25 (≤0.12–0.5)	0.25 (≤0.12–0.5)	0.25 (≤0.12–0.25)
CRO	≤1 (≤1–4)	≤1 (≤1–8)	≤1 (≤1–2)	2 (≤1–4)	≤1 (≤1–4)	8 (4–16)	1.5 (≤1–4)	4 (≤1–8)

^aCZO, cefazolin; FEP, cefepime; CTX, cefotaxime; FOX, cefoxitin; CPD, cefpodoxime; CAZ, ceftazidime; CCV, ceftazidime-clavulanic acid; CRO, ceftriaxone.

^bNumbers in parentheses represent the MIC ranges (μg/ml) observed in each respective group. Data for ampicillin, amoxicillin-clavulanic acid, piperacillin-tazobactam, cephalothin, cefotaxime-clavulanic acid, and carbapenems (imipenem and meropenem) are not shown because the MICs were monomorphic and outside the testing range.

Plasmid preparations from each original isolate (Qiagen plasmid midikit; Qiagen, Hilden, Germany) were electroporated into ElectroMAX E. coli DH10B cells (Invitrogen, Carlsbad, CA). Transformants were recovered with the plasmid preparations from 18 isolates on Mueller-Hinton agar containing 50 μg/ml ampicillin or 25 μg/ml tetracycline. However, despite repeated attempts, transformants were not recovered with plasmid preparations of two E. coli isolates. Southern blot hybridizations (9) of total DNA (DNeasy blood and tissue kit; Qiagen) and of plasmid preparations with a digoxigenin-labeled bla_SHV probe (PCR DIG probe synthesis kit; Roche, Mannheim, Baden-Württemberg, Germany) confirmed the bla_SHV from these isolates was chromosomal. DNA sequencing (described below) showed that these two isolates carried bla_SHV-2a, and they were not investigated further. The bla_SHV genes from the remaining 18 isolates were carried on plasmids of sizes ranging from approximately 95 kbp to 200 kbp.

DNA sequencing (Laboratory Services, University of Guelph) with a combination of PCR and primer walking and comparison with DNA sequences in NCBI's GenBank (10) and with SHV amino acid sequences (http://www.lahey.org/studies/webt.asp#SHV) was used to identify bla_SHV variants. Five and 13 plasmids carried bla_SHV-2 and bla_SHV-2a, respectively. All bla_SHV-2 and bla_SHV-2a genes were identical to GenBank accession no. AF148851 and X53817, respectively (11), and matched the SHV-2 and SHV-2a amino acid sequences in the Lahey Clinic database. The median MICs of the original isolates for β-lactams were generally identical for SHV-2a and SHV-2. However, they were 1 dilution higher for SHV-2a than for SHV-2 with most β-lactams when transformants were compared (Table 1). Thus, the slight differences in β-lactam resistance phenotypes provided by the substitution between SHV-2 and SHV-2a and in their promoter regions may be detectable when expressed under a standardized genetic background, but not readily in natural bacterial populations.

As the two S. enterica isolates had median MICs at least 2 dilution steps higher than the E. coli isolates for most β-lactams (Table 1), bla_SHV promoter regions of the two S. enterica isolates (SA17 and SA18) and of two E. coli isolates (EC7 and EC15) were characterized by PCR (12). The promoter regions for bla_SHV-2 for E. coli EC7 and Salmonella SA18 were identical in sequence, and so were bla_SHV-2a promoter sequences for E. coli EC15 and Salmonella SA17, thus suggesting the observed difference in MICs between bacterial host species was not due to promoter variants affecting SHV expression levels.

Although some minor variation resulting in seven antimicrobial susceptibility profiles was observed, transformants were similar overall, with most bla_SHV plasmids associated with resistance to sulfonamides, tetracycline, streptomycin, and occasionally trimethoprim or gentamicin (Fig. 1). PCR screening (7) of transformants confirmed the presence of associated resistance genes (Fig. 1) for sulfonamides (n = 17), tetracycline (n = 15), and streptomycin (n = 17). PCR screening for intI1 and intI2 genes (13) indicated that 17 of the 18 bla_SHV plasmids carried class 1 integrons. Sequencing of PCR products (14) confirmed integron variable regions of 1,200 bp (n = 10) and 900 bp (n = 7), consisting of dfrA1 for trimethoprim resistance plus aadA1 for streptomycin resistance or aadA1 alone (Fig. 1). The presence of these additional resistance determinants suggests a potential for coselection and persistence of bla_SHV plasmids in the absence of cephalosporin use. The plasmids belonged to three different replicon types (2), but the vast majority were of the incI1 type (Fig. 1). EcoRI restriction analysis of transformant plasmids was conducted following standard protocols (3) and analyzed using BioNumerics v5.1 software (Applied Maths, Austin, TX). Only partial correlations between plasmid clusters, replicon types, antimicrobial susceptibility, integron size, antimicrobial resistance genes, and bla_SHV variants were visible (Fig. 1), suggesting numerous recombinations in the history of these plasmids. However, the S. enterica isolates and some E. coli isolates shared highly related replicon type I1 plasmids, and two pairs of E. coli isolates (EC3 and EC16 and EC10 and EC14) from spatially and temporally distinct sources carried indistinguishable I1 plasmids (Fig. 1). This suggests the ability of bla_SHV plasmids to move across bacterial populations and from commensal E. coli to Salmonella. To confirm this, the 2 S. enterica and 13 E. coli isolates were conjugated with a nalidixic acid-resistant derivative of S. enterica serovar Typhimurium strain LB5000 (15) by overnight mating in Luria-Bertani (LB) broth using a 1:1 ratio of donor and recipient cells. Transconjugants were selected using LB agar containing 50 μg/ml ampicillin and 50 μg/ml nalidixic acid. Three E. coli isolates (EC5, EC15, and EC16) were resistant to nalidixic acid and could not be tested. The bla_SHV plasmids of all 15 isolates tested were transferable by conjugation. Efficiencies of 7.36 × 10⁻² and 1.41 × 10⁻⁶ per donor, respectively, were recorded for the transfer of the bla_SHV plasmid between two randomly selected E. coli isolates (EC2, replicon type I1, and EC9, replicon type P) and S. enterica serovar Typhimurium LB5000. Overall, these results demonstrate widespread horizontal transfer of bla_SHV plasmids, in particular of replicon type incI1, rather than clonal spread of a few SHV-2- and SHV-2a-producing strains. Furthermore, these plasmids appear to move easily between E. coli and S. enterica.

Fig 1.

EcoRI restriction analysis of bla_SHV plasmids and characteristics of bla_SHV isolates. The scale on the top left of the figure indicates the percentage of similarity for the EcoRI restriction profiles of the bla_SHV plasmids. Footnotes: a, isolate identification number (all isolates are E. coli, except SA17 and SA18, which are Salmonella enterica); b, year isolate was collected; c, antimicrobial for which the original isolates' MICs were in the intermediate or resistant range; d, antimicrobial resistance genes detected in transformants. Genes that were shown to be part of a class 1 integron variable region are underlined. Antimicrobial abbreviations: AMP, ampicillin; CEP, cephalothin; CPD, cefpodoxime; CRO, ceftriaxone; CTX, cefotaxime; CZO, cefazolin; GEN, gentamicin; NAL, nalidixic acid; SSS, sulfonamides; STR, streptomycin; SXT, sulfonamide-trimethoprim; TET, tetracycline; TIO, ceftiofur; TZP, piperacillin-tazobactam. Reduced susceptibilities that were transferred to the respective transformants are in boldface and underlined.

Previous reports have suggested that lower cephalosporin MIC values should be used for E. coli than for S. enterica to trigger investigations of the presence of ESBLs (16). Although not extensive enough for statistical evaluation, the results from the present study strongly support these findings in the case of bla_SHV. The two S. enterica isolates had median MICs at least 2 dilution steps higher than those of the E. coli isolates for cefotaxime, ceftriaxone, ceftazidime, and cefpodoxime and 1 dilution step higher for most other β-lactams (Table 1). Similar trends were visible when comparing the bla_SHV plasmids in the S. Typhimurium LB5000 host with the E. coli DH10B host or with the original E. coli isolates, whereas transfer from the original S. enterica isolate into S. Typhimurium LB5000 resulted in minor changes only (Table 1).

In conclusion, although their prevalence was low (3.2% and 0.4% of ampicillin-resistant isolates from chicken and porcine origins, respectively), the detection of plasmid-borne bla_SHV-2 and bla_SHV-2a in bacteria from Canadian food animals and retail meat with a high potential for horizontal transfer between E. coli and S. enterica is a public health concern. It is certainly not a coincidence that the first bla_SHV characterized in a Canadian non-Typhi Salmonella isolate of human origin was also a bla_SHV-2a (17). Despite their high mobility, the vast majority of bla_SHV plasmids were found in bacteria originating from chicken. Further testing of 12 E. coli isolates from 2010 (9 chicken and 2 bovine isolates and 1 porcine isolate) and of two Salmonella isolates from 2008 (S. enterica serovar Kiambu from chicken) and 2010 (S. enterica serovar Derby from swine) from the CIPARS collection confirmed that bla_SHV-2a (n = 11) and bla_SHV-2 (n = 2) are still persisting in bacteria from food animals in Canada (P. Boerlin, G. Chalmers, and L. Martin, unpublished data), in particular from chickens. Concerns have been expressed previously regarding use of broad-spectrum cephalosporins and the presence of cephamycinases in E. coli and Salmonella isolates from chickens (4). The data presented here suggest that these concerns may also extend to the selection of ESBLs, such as bla_SHV.