Extended-Spectrum β-Lactamase CTX-M-1 in Escherichia coli Isolates from Healthy Poultry in France

Abstract

Genes encoding extended-spectrum β-lactamase CTX-M-1 were detected in 12 Escherichia coli isolates recovered over a 7-month period from the ceca of healthy poultry in seven districts in France in 2005. Eleven of those strains were not clonally related and had a bla_CTX-M-1 gene located on transferable plasmids of different sizes and structures.

Most of the clavulanic acid-inhibited extended-spectrum β-lactamases (ESBL) are either derivatives of narrow-spectrum TEM- and SHV-type β-lactamases or CTX-M, PER, VEB, and GES/IBC-type β-lactamases (1, 35, 37). The CTX-M-type β-lactamases confer resistance to expanded-spectrum cephalosporins, such as cefotaxime, ceftazidime, and cefepime (3). Escherichia coli strains harboring CTX-M-type ESBL genes have been detected increasingly in humans since the beginning of the 1990s (3, 4, 5, 23-26, 46) and represent a real threat mostly in community-acquired infections (37). There are now more than 60 CTX-M β-lactamases (www.lahey.org/studies/other.asp), and they may be classified according to subgroups CTX-M-1, CTX-M-3, CTX-M-8, CTX-M-9, and CTX-M-25 (3).

Whereas plasmid-mediated cephalosporinases (CMY type) have been extensively reported in animal isolates (15, 42, 47), the aim of the present study was to search for CTX-M producers in healthy animals recovered in slaughterhouses of several districts located in the western part of France, which is the main region for broiler production.

From April to November 2005, 112 samples were collected from the ceca from healthy poultry in 10 slaughterhouses located in seven districts in France (i.e., Côtes-d'Armor, Finistère, Landes, Maine-et-Loire, Mayenne, Morbihan, and Vendée). These samples were plated on MacConkey agar plates containing ceftazidime (1 μg/ml) or cefotaxime (1 μg/ml). MICs of β-lactams were determined by an agar dilution technique (12). Out of the 112 samples, 32 nonduplicate E. coli isolates were resistant or intermediate to ceftazidime and/or to cefotaxime. Two isolates were resistant to amoxicillin and ticarcillin and susceptible to ticarcillin and clavulanate, suggesting the production of a penicillinase (PC) (Table 1). Eighteen isolates had a phenotype consistent with the expression of a cephalosporinase (CS), since they were resistant to amoxicillin, amoxicillin-clavulanate, and cephalothin and had reduced susceptibility to cefuroxime (Table 1). A double-disk synergy test for detection of ESBL carried out as described previously (29) revealed synergy between clavulanate and cefotaxime or ceftazidime-containing disks for 12 isolates, suggesting production of an ESBL in 10.7% of the samples (Table 1). The MICs of β-lactams for those isolates evidenced decreased susceptibility or resistance to expanded-spectrum cephalosporins, including ceftiofur that has been approved in France in 2003 for treatment of animals, including cattle, horses, and pigs (Table 1).

TABLE 1.

MICs of β-lactams for E. coli isolates and transformants

β-Lactama	MIC (μg/ml) for E. coli with the following phenotype:
	Penicillinase (TEM-1) (n = 2)	Cephalosporinase “overexpression” (n = 18)	ESBL + TEM-1 (n = 3)	ESBL (n = 12)	E. coli TOP10(pCTX-M-1)	E. coli TOP10
Amoxicillin	>512	>512	>512	>512	>512	4
Amoxicillin-CLA	512	512	8-16	8-16	8	4
Ticarcillin	>512	16-32	>512	>512	>512	4
Piperacillin	16-32	16	256-512	256-512	512	2
Piperacillin-TZB	8	8	1	1	2	2
Cephalothin	>512	>512	>512	>512	>512	4
Ceftazidime	2	4	0.5-2	0.5-2	2	0.5
Cefotaxime	1	1-2	32	32	32	0.06
Cefuroxime	32-64	16-64	>512	>512	>512	4
Cefoxitin	32-64	16-64	4-8	4-8	4	4
Cefepime	0.06	0.06	4-8	4-8	4	0.06
Ceftiofur	0.06	0.06-0.12	32-256	32-256	256	0.06
Aztreonam	4	1-4	4-8	4-8	8	0.12
Imipenem	0.25	0.25	0.25	0.25	0.25	0.25

^aCLA, clavulanic acid at a fixed concentration of 2 μg/ml; TZB, tazobactam at a fixed concentration of 4 μg/ml.

Detection of several β-lactamase genes, including bla_TEM, bla_SHV, bla_CTX-M, bla_CMY-2, bla_FOX, bla_OXA-1, and amplification of the entire ampC gene and its promoter regions were carried out by PCR as described previously (17, 24, 36). The PCR products were sequenced on both strands on an Applied Biosystems sequencer (ABI 377). Mutations in the ampC gene and promoter region were compared to those of E. coli K-12 strain. A TEM-1 β-lactamase gene was detected in the two PC-producing E. coli isolates (Table 1). Amplification and sequencing of the promoter region of the ampC gene were carried out on 5 isolates among the 18 CS-producing E. coli isolates and the PC-producing E. coli isolates. Mutations were detected at positions −42 (C→T), −18 (G→A), −1 (C→T), and +58 (C→T) of the ampC gene. Mutations at these locations could be associated with AmpC hyperproduction and thus explain the β-lactam resistance phenotype made of resistance to inhibitors (11, 22). Hyperproduction of the AmpC enzyme was confirmed by quantitative determination of cephalothin hydrolysis in culture extracts of the PC- and CS-producing and E. coli K-12 isolates as described previously (39). A 40-fold-higher (ca. 320 versus 8 mU/mg of protein) β-lactamase activity was noticed in the PC and CS culture extracts compared to that of E. coli K-12. Sequencing of the ampC structural gene did not identify specific mutations responsible for the extension of the hydrolysis spectrum towards expanded-spectrum cephalosporins as described by Mammeri and Nordmann (30). All ESBL-producing E. coli isolates had the same bla_CTX-M-1 gene, and three out of these isolates had an additional bla_CTX-M-1 β-lactamase gene.

The surrounding genetic structures of the bla_CTX-M-1 gene were characterized by PCR as reported previously (26). ISEcp1 sequence was identified 80 bp upstream of the start codon of the bla_CTX-M-1 gene in all cases (16, 43). ISEcp1 is an insertion sequence frequently associated with any of the bla_CTX-M genes belonging to three out of the five known clusters (CTX-M-1, -M-2, and -M-9 clusters) (2, 20, 25, 38, 43) and may contribute to expression of genes located in its right-hand extremity, including bla_CTX-M genes. ISEcp1 possesses peculiar transposition properties (23, 40) and may explain mobilization of the plasmid-located bla_CTX-M-1 gene from its chromosomal origin in Kluyvera cryocrescens (14).

Pulsed-field gel electrophoresis (PFGE) analysis, performed as described previously with XbaI endonuclease (18), showed that 11 out of the 12 bla_CTX-M-1-positive E. coli isolates belonged to distinct genotypes (Fig. 1). Despite several attempts, isolates 21 and 34 could not be typed by PFGE due to DNA self-denaturation. However, these strains had different enterobacterial repetitive intergenic consensus-PCR patterns from those of the other E. coli isolates (data not shown) (17). Only E. coli isolates 12 and 16 had very similar patterns. Thus, spread of the bla_CTX-M-1 gene did not result from the spread of a single clone. The characteristics of the bla_CTX-M-1-positive E. coli strains are shown in Table 1 and 2. The number of coresistance markers were limited compared to those identified in most CTX-M-producing E. coli human isolates (37).

FIG. 1.

PFGE of XbaI-digested genomic DNAs from 12 E. coli strains. Lanes: 1, E. coli isolate 3; 2, isolate 5; 3, isolate 12; 4, isolate 16; 5, isolate 21; 6, isolate 22; 7, isolate 33; 8, isolate 34; 9, isolate 36; 10, isolate 39; 11, isolate 47; 12, isolate 48; 13, E. coli K-12 used as an unrelated strain. Lane M contains molecular size markers (bacteriophage lambda DNA ladder), and the positions of the molecular size markers are shown to the left of the gel.

TABLE 2.

Features of CTX-M-1-positive E. coli isolates

Isolate	Slaughterhousea	Date of isolation in 2005	Non-β-lactam antibiotic resistanceb	bla gene(s) as determined by PCR and sequencing	Plasmid size (kb)	Incc	Antibiotic resistanceb cotransferred with blaCTX-M-1
3	A	May	Su, Te, Tp	blaCTX-M-1	100	I1	Su, Tp
5	A	June	Su, Te, Tp	blaTEM-1, blaCTX-M-1	100	I1	Su, Tp
12	B	October	Su, Te, Tp	blaTEM-1, blaCTX-M-1	100	I1	Su, Tp
16	B	October	Su, Te, Tp	blaTEM-1, blaCTX-M-1	100	I1	Su, Tp
21	C	June	Su, Tp	blaCTX-M-1	100	I1	Su, Tp
22	C	April	Su, Te, Tp	blaCTX-M-1	100	I1	Su, Tp
33	E	June	Su, Tp	blaCTX-M-1	100	I1	Su, Tp
34	E	June	Su, Te, Na	blaCTX-M-1	80	I1
36	E	June	Su, Tp	blaCTX-M-1	100	I1	Su, Tp
39	F	May	Su, Te	blaCTX-M-1	100	I1	Su, Te
47	H	May	Te, Tp, Ra	blaCTX-M-1	55	NT
48	H	June	Su, Te, Tp	blaCTX-M-1	100	I1	Su, Tp

^aSlaughterhouses and districts in France: A and G are located in Maine et Loire, B and D in Vendée, C in Côtes d'Armor, E in Mayenne, F in Finistère, H in Landes and J in Morbihan.

^bSusceptibility to non-β-lactam agents were determined by the disk diffusion method. Antibiotics tested included nalidixic acid (Na), rifampin (Ra), sulfonamide (Su), tetracycline (Te), and trimethoprim (Tp). In E. coli transformants 5 and 12, the bla_TEM-1 gene was not cotransferred with bla_CTX-M-1.

^cPlasmid incompatibility group (Inc). NT, nontypeable plasmid by PCR-based replicon typing.

Since bla_CTX-M-1 genes are usually located on large plasmids (20), plasmid DNA was extracted (21) and used for transformation into E. coli TOP10 (Invitrogen) (20). Large plasmids of ca. 100, 80, and 55 kb (Table 2) were identified (Fig. 2). PCR-based replicon typing of the major plasmid incompatibility groups (9) showed that the bla_CTX-M-1-positive plasmids belonged to the same IncI1 incompatibility group, with one exception (E. coli isolate 47) (data not shown). The plasmid of E. coli isolate 47 gave negative results with all the Inc/rep primers tested (9) and cannot be classified in a major plasmid incompatibility group as several other replicons as described previously (9). MICs of β-lactams for the E. coli transformants showed susceptibility profiles that mirrored those observed for clinical strains (Table 1). Then, plasmid DNA from transformants was digested with the EcoRI restriction enzyme and subjected to electrophoresis in a 1% agarose gel as described previously (44). Plasmids with similar sizes and structures were detected in 5 out of the 11 nonrelated isolates withdrawn from four different slaughterhouses (Fig. 3). This result suggests a possible spread of similar plasmids between different slaughterhouses. Direct transfer of these plasmids into rifampin-resistant E. coli C600 was attempted by liquid mating-out assays as described previously (23). Transconjugants resistant to amoxicillin (100 μg/ml) and rifampin (200 μg/ml) were selected at a frequency of 2.5 to 4 × 10⁻⁵ per donor cell. All transconjugants were CTX-M-1 positive, demonstrating self-transferability of these plasmids.

FIG. 2.

Plasmid DNA electrophoresis profile of CTX-M-1-producing E. coli isolates and E. coli TOP10(pCTXM-1) transformants (Tf). Lane 1, E. coli NTCC 50192 used as a molecular size marker; lane 2, isolate 3; lane 3, Tf 3; lane 4, isolate 5; lane 5, Tf 5; lane 6, isolate 12; lane 7, Tf 12; lane 8, isolate 21; lane 9, Tf 21; lane 10, isolate 22; lane 11, Tf 22; lane 12, isolate 33; lane 13, Tf 33; lane 14, isolate 34; lane 15, Tf 34; lane 16, isolate 36; lane 17, Tf 36; lane 18, isolate 39; lane 19, Tf 39; lane 20, isolate 47; lane 21, Tf 47; lane 22, isolate 48; lane 23, Tf 48.

FIG. 3.

Plasmid DNA fingerprinting of E. coli TOP10(pCTX-M-1) transformants (Tf) restricted with EcoRI. Lane M, molecular size markers (1-kb DNA ladder); lane 1, Tf 3; lane 2, Tf 5; lane 3, Tf 12; lane 4, Tf 21; lane 5, Tf 22; lane 6, Tf 33; lane 7, Tf 34, lane 8, Tf 36; lane 9, Tf 39; lane 10, Tf 47; lane 11, Tf 48.

Several reports identified CTX-M-s also from animal isolates (2, 6, 7, 8, 13, 19, 45, 46) including the very first CTX-M enzyme that was from an E. coli isolate of a Japanese dog (31). The bla_CTX-M-1 gene was identified in a single E. coli isolate (from milk cattle) out of 10 ESBL-producing E. coli isolates from sick animals in Spain in 2003 with CTX-M-14 being predominantly found in that case (6). Spread of E. coli carrying ESBL CMY-2, SHV-12 and CTX-M-1 has been identified in companion animals in Italy (10). Our work identified CTX-M-1 in 12 (10.7%) of 112 poultry fecal samples corresponding to a 10-fold-higher rate than that of the nation-based surveys performed in Japan for CTX-M-2 E. coli producers in cattle (22, 45). This rate mirrors the increasing and high percentage of CTX-M carriers reported in humans (32, 33) that may be related in part to contamination during abattoir slaughtering and food processing (15, 27). As described recently, E. coli strains of food-producing animal farms may be a reservoir for ESBL-producing organisms (2). Recent studies report bla_CTX-M genes on self-transferable plasmids in nontyphoid Salmonella in poultry in Spain, Greece, and France (6, 41, 46), suggesting a possible transfer of antibiotic resistance conjugative plasmids in E. coli in feces. Interestingly, we have shown that culture at 40°C may increase ISEcp1-mediated transfer of bla_CTX-M genes (23), a temperature that corresponds to the internal temperature of poultry. In addition, we identified here the bla_CTX-M-1 gene mostly located on IncI1 plasmids, whereas this ESBL gene was recently identified on IncN plasmids in human isolates in Italy (34).

Our study reports a widespread diffusion of CTX-M-1 in E. coli in food-producing poultry in France, although the most important resistance mechanism to extended-spectrum cephalosporins was overexpression of the nontransferable cephalosporinase. Noteworthy, CTX-M-1 belongs to the CTX-M group (including CTX-M-1, CTX-M-3, and CTX-M-15) that is predominantly identified in human isolates in France as well as worldwide (16, 17, 26, 28, 37; P. Nordmann, personal data). This study suggests a possible reservoir for ESBL genes on transferable plasmids in poultry.