Herd prevalence of Enterobacteriaceae producing CTX-M-type and CMY-2 β-lactamases among Japanese dairy farms



Mamoru Ohnishi, Veterinary Clinical Laboratory, Nemuro District Agricultural Mutual Aid Association, 11-5 Minami, 5 Nishi, Nakashibetsu-cho, Shibetsu-gun, Hokkaido 086-1105, Japan. E-mail: monishi@kind.ocn.ne.jp



To determine the herd prevalence of Enterobacteriaceae producing CTX-M-type extended-spectrum β-lactamases (ESBLs) among 381 dairy farms in Japan.

Methods and Results

Between 2007 and 2009, we screened 897 faecal samples using BTB lactose agar plates containing cefotaxime (2 μg ml−1). Positive isolates were tested using ESBL confirmatory tests, PCR and sequencing for CTX-M, AmpC, TEM and SHV. The incidence of Enterobacteriaceae producing CTX-M-15 (= 7), CTX-M-2 (= 12), CTX-M-14 (= 3), CMY-2 (= 2) or CTX-M-15/2/14 and CMY-2 (= 4) in bovine faeces was 28/897 (3·1%) faecal samples. These genes had spread to Escherichia coli (= 23) and three genera of Enterobacteriaceae (= 5). Herd prevalence was found to be 20/381 (5·2%) dairy farms. The 23 E. coli isolates showed clonal diversity, as assessed by multilocus sequence typing and pulsed-field gel electrophoresis. The pandemic E. coli strain ST131 producing CTX-M-15 or CTX-M-27 was not detected.


Three clusters of CTX-M (CTX-M-15, CTX-M-2, CTX-M-14) had spread among Japanese dairy farms.

Significance and Impact of the Study

This is the first report on the prevalence of multidrug-resistant CTX-M-15–producing E. coli among Japanese dairy farms.


Since 2000, Escherichia coli and other Enterobacteriaceae species producing CTX-M-type extended-spectrum β-lactamases (ESBLs) (CTX-M) have been commonly isolated from community-acquired extraintestinal infections in humans and their companion animals (Harada et al. 2012; Kuroda et al. 2012) and faeces of food-producing animals worldwide (Carattoli 2008). The blaCTX-M-related genes were transferred separately to genetic mobilization units (i.e. plasmids) from chromosomes of different Kluyvera species (i.e. K. ascorbata, K. georgiana, K. cryocrescens) that live in water, soil, human and animal intestinal tract; and thereby, CTX-M has derived in five CTX-M clusters (CTX-M-1, CTX-M-2, CTX-M-8, CTX-M-9 and CTX-M-25) from base sequence homology. This has been facilitated by genetic mobilization units such as insertion sequences (i.e. ISEcp1 or ISCR1) (Bonnet 2004; Cantón et al. 2012). CTX-M confers resistance against penicillins, oxyimino-cephalosporins (i.e. cefuroxime, cefotaxime, ceftriaxone, cefpodoxime, ceftazidime, ceftiofur, cefquinome) and monobactams (Bonnet 2004).

Plasmidic CMY-2, derived from the AmpC β-lactamase of Citrobacter freundii, became the dominant AmpC β-lactamase among Enterobacteriaceae, especially Salmonella enterica, in humans and animals throughout the world (Tamang et al. 2012). CMY-type β-lactamases are cephamycinases that confer resistance to cephamycins and oxyimino-cephalosporins. Recently, a high incidence of blaCTX-M-positive E. coli was found in retail chicken meat. (Overdevest et al. 2011). However, few studies have reported the prevalence of Enterobacteriaceae producing CTX-M and CMY on dairy farms (Hartmann et al. 2012). The aims of the present study were to determine the herd prevalence of Enterobacteriaceae producing CTX-M and CMY among numerous dairy farms in Japan and to evaluate the genetic relatedness among E. coli isolates derived from dairy cattle by serotyping, multilocus sequence typing (MLST) and pulsed-field gel electrophoresis (PFGE).

Materials and methods

Faecal samples

A total of 897 faecal samples were obtained from each of 897 cattle affected by diarrhoea (caused by i.e. colibacillosis, rotavirus infection, cryptosporidiosis, coccidiosis, parasitic gastroenteritis) or dictyocauliasis; the 897 cattle consisted of 550 calves (<12 months old), 80 heifers (≥12 months old to first parturition) and 267 cows. Faecal samples were submitted daily for clinical diagnostic tests to the clinical laboratory of the Nemuro District Agricultural Mutual Aid Association by nine-member veterinary clinics. The faecal samples were obtained from a total of 381 dairy farms localized in Nemuro Subprefecture of Hokkaido Prefecture, Japan; 475 samples were obtained from 244 dairy farms between May and December in 2007 and 422 samples from 214 dairy farms between July and December in 2009.

Screening, identification, PCR and sequence typing for ESBLs and AmpC producers

The faecal samples were cultured as the first screening for ESBL producers on the day they were submitted to the laboratory. Approximately, 50 μg faeces was streaked directly onto a BTB-lactose agar plate containing 2 μg ml−1 cefotaxime and 8 μg ml−1 vancomycin (Shiraki et al. 2004) and then incubated at 35°C for 40 h.

A total of 250 isolates were tested using the Clinical and Laboratory Standards Institute (CLSI) combination disc ESBL confirmatory tests (CLSI 2008) and a chromogenic oxyimino-cephalosporins hydrolysis test (Cica-β-test I; Kanto Chemical, Tokyo, Japan) to detect ESBLs, plasmidic AmpC β-lactamases and metallo-β-lactamases. All positive isolates were screened for metallo-β-lactamases by the sodium mercaptoacetic acid (SMA) double-disc synergy test using two Kirby–Bauer discs containing ceftazidime and one disc containing SMA (Eiken Chemical, Tokyo, Japan). Enterobacteriaceae isolates were identified using the ID 32 E API system (Sysmex bioMérieux, Tokyo, Japan).

All isolates that were positive for the ESBL confirmatory tests and/or the Cica-β-test were analysed by multiplex polymerase chain reaction (PCR) for the presence of blaCTX-M genes (Xu et al. 2007) and plasmid-mediated AmpC β-lactamase genes (i.e. CMY, ACC, FOX, MOX, DHA, CIT and EBC groups) (Pérez-Pérez and Hanson 2002). CTX-M-positive isolates were analysed by bidirectional sequencing using group-specific PCR primers for blaCTX-M-1 group (Mena et al. 2006), blaCTX-M-2 group and blaCTX-M-9 group (Kojima et al. 2005). The AmpC-positive isolates were analysed using type-specific PCR primers (Kojima et al. 2005). The blaTEM and blaSHV genes were analysed using primers (Kojima et al. 2005) and bidirectionally sequenced (Table S1).

Antimicrobial susceptibility testing

The minimum inhibitory concentrations (MICs) of 23 antimicrobial agents were determined for isolates positive for CTX-M and/or AmpC genes by broth microdilution using a customer-designed, commercially prepared microtiter panel (Opt Panel MP; Kyokuto Pharmaceutical, Tokyo, Japan), and additional susceptibility test for cefoxitin, kanamycin, chloramphenicol and levofloxacin was performed by disk diffusion according to the CLSI guidelines (CLSI 2008, 2011).

The breakpoints for veterinary pathogens were used for 12 antimicrobial agents (CLSI 2008), and the breakpoints for human Enterobacteriaceae isolates were used for the other 12 antimicrobial agents (CLSI 2011). Escherichia coli ATCC25922 and Pseudomonas aeruginosa ATCC27853 were used as quality-control strains.

Serotyping, MLST and PFGE of Escherichia coli

The 23 E. coli isolates were serotyped according to O and H antigens using the pathogenic E. coli antisera ‘SEIKEN’ Set 1 for 50 O antigens and Set 2 for 22 H antigens (Denka Seiken, Tokyo, Japan) according to the manufacturer's instructions.

We carried out the amplification and sequencing of seven housekeeping genes and determined the allelic profile and sequence types (STs) using the protocols and database on the E. coli MLST website (http://mlst.ucc.ie/mlst/dbs/Ecoli).

Pulsed-field gel electrophoresis was conducted according the PulseNet standardized laboratory protocol (CDC 2004). The DNA restriction fragments digested with Xba I or Bln I (Roche Applied Science, Mannheim, Germany) were separated using the CHEF-DR III electrophoresis systems (Bio-Rad, Hercules, CA, USA). Dendrogram of the combined PFGE patterns of Xba I and Bln I was constructed using BioNumerics software version 5.1 (Applied Maths, Austin, TX, USA) with Dice coefficient and the unweighted pair group method, and the similarity of the average from the two experiments. PFGE types were distinguished at a cut-off of ≥90% similarity.


Incidence and herd prevalence of CTX-M and CMY-2 producers

A total of 30 isolates were positive for the ESBL confirmatory tests and/or the Cica-β-test. However, no isolate was SMA test-positive. Twenty-eight CTX-M and/or CMY-2 producers (E. coli,= 23; Klebsiella pneumoniae, n = 3; C. freundii,= 1; Enterobacter cloacae,= 1) were isolated from 28 faecal samples obtained from 17 calves, one heifer and 10 cows from 20 dairy farms between 2007 and 2009. The other two isolates identified as Stenotrophomonas maltophilia or Burkholderia cepacia were negative for CTX-M, AmpC, TEM and SHV genes.

Seven E. coli isolates (25·0%) harboured blaCTX-M-15, 12 isolates (42·9%) (E. coli,= 7; K. pneumoniae, n = 3; E. cloacae,= 1; C. freundii, n = 1) harboured blaCTX-M-2 and three E. coli isolates (10·7%) harboured blaCTX-M-14. Two E. coli isolates (7·1%) harboured blaCMY-2 and four E. coli isolates (14·3%) harboured blaCMY-2 and blaCTX-M-2 or blaCTX-M-15 or blaCTX-M-14. In addition to CTX-M and CMY-2, 12 E. coli isolates (42·9%) and one C. freundii isolate (3·6%) harboured blaTEM-1 or blaSHV-11 and one K. pneumoniae isolate (3·6%) harboured blaTEM-1 and blaSHV-11. Four (50%) of eight CTX-M-15–producing E. coli isolates also harboured blaTEM-1 (Tables 1 and 2).

Table 1. ST and ST complex (multilocus sequence typing), pulsed-field gel electrophoresis (PFGE) and serotype of 23 β-lactamase–producing Escherichia coli isolates obtained from faeces
IsolateYear, Farm, CattleCTX-M, CMYTEM SHVST (ST complex)PFGE typeO-, H-serotype
  1. a

    O-antigen untypable: H-antigen untypable.

  2. b

    ST complex undefined.

FE107, IA, Calf1CTX-M-2TEM-158 (155)OOUTa: HUTa
FE207, IB, Cow1CTX-M-2 10 (10)EO8: H7
FE307, M, Cow2CTX-M-15TEM-11167 (−)bKOUT: H19
FE407, M, Cow3CTX-M-15TEM-1540 (−)UOUT: HUT
FE507, M, Cow4CTX-M-15 1167 (−)KO28ac: HUT
FE607, IC, Calf2CTX-M-2 10 (10)GOUT: H51
FE707, IC, Calf3CTX-M-2 10 (10)GOUT: HUT
FE807, ID, Calf4CTX-M-14TEM-169 (69)LO15: HUT
FE907, ID, Calf6CTX-M-14TEM-110 (10)COUT: H9
FE1007, IF, Cow5CMY-2 1284 (−)DOUT: HUT
FE1107, IG, Cow6CTX-M-15, CMY-2TEM-1744 (−)HOUT: HUT
FE1207, IH, Calf7CTX-M-15 394 (394)SO44: H18
FE1309, IJ, Heifer1CTX-M-15 1266 (−)ROUT: H34
FE1409, IL, Calf9CMY-2TEM-12438 (−)AOUT: H16
FE1509, IM, Calf10CTX-M-14TEM-157 (350)JOUT: H6
FE1609, M, Calf11CTX-M-15TEM-12325 (−)MOUT: HUT
FE1709, IN, Calf12CTX-M-2 2324 (−)POUT: HUT
FE1809, II, Calf13CTX-M-2SHV-112437 (−)TO164: HUT
FE1909, IP, Cow9CTX-M-2, CMY-2SHV-1144 (10)IOUT: HUT
FE2009, IN, Calf14CTX-M-14, CMY-2TEM-188 (23)BOUT: HUT
FE2109, IQ, Cow10CTX-M-2 48 (10)QO124: H4
FE2209, IR, Calf16CTX-M-2, CMY-2 46 (46)NOUT: HUT
FE2309, IS, Calf17CTX-M-15 744 (−)FOUT: HUT
Table 2. Origins of CTX-M, TEM and SHV type β-lactamase–producing Enterobacteriaceae isolates other than Escherichia coli
IsolateIsolation year, Farm, CattleBacterial speciesCTX-MTEM, SHV
FK107, IE, Calf5 Klebsiella pneumoniae CTX-M-2TEM-1, SHV-11
FK209, II, Calf8 K. pneumoniae CTX-M-2 
FK309, IK, Cow7 K. pneumoniae CTX-M-2 
FCf109, IO, Cow8 Citrobacter freundii CTX-M-2TEM-1
FEn109, II, Calf15 Enterobacter cloacae CTX-M-2 

The incidence of Enterobacteriaceae producing CTX-M and/or CMY-2 in bovine faeces was 28 (3·1%) of the 897 faecal samples, and the incidence of E. coli producing these β-lactamases was 23 of 897 (2·6%). The incidence of Enterobacteriaceae was 17 (3·1%) of the 550 faecal samples from calves and 10 (3·7%) of the 267 faecal samples from cows; however, this difference was not significant (P > 0·05, chi-square test). The herd prevalence of these Enterobacteriaceae was 20 (5·2%) of the 381 dairy farms. Four E. coli isolates from three cows and one calf on farm M harboured blaCTX-M-15. (Tables 1 and 2).

Antimicrobial susceptibility and antimicrobial use

Isolates producing CTX-M exhibited high resistance to oxyimino-cephalosporins. However, isolates producing CTX-M-2 and CTX-M-14 were highly susceptible to ceftazidime, cefmetazole, moxalactam, carbapenems, gentamicin, amikacin and fluoroquinolones. MICs for ceftazidime, cefmetazole, enrofloxacin and ciprofloxacin were higher for the E. coli isolates producing CTX-M-15 and CMY-2 than for the isolates producing CTX-M-2 and CTX-M-14 (Table 3). Most of isolates producing CTX-M, especially CTX-M-15, exhibited resistance to kanamycin, oxytetracycline, chloramphenicol and trimethoprim-sulfamethoxazole (Table 3).

Table 3. Antimicrobial susceptibilities of 23 bovine faecal Escherichia coli isolates and five Enterobacteriaceae isolates other than E. coli that produce CTX-M type and/or CMY-2 β-lactamases, as assessed by broth microdilution or disk diffusion method
Antimicrobial agentE. coli CTX-M-15 n=7E. coli CTX-M-2 n=7E. coli CTX-M-14 = 3E. coli CMY-2 and CTX-M- 2/15/14 = 4E. coli CMY-2 = 2Klebsiella pneumoniae,= 3 Citrobacter freundii, n = 1 Enterobacter cloacae, n = 1; CTX-M-2Breakpointa μg ml−1; mm
  1. CAZ, ceftazidime; CLA, clavulanic acid; CTX, cefotaxime; OTET, oxytetracycline; SXT, trimethoprim-sulfamethoxazole; KAN, Kanamycin; CHL, chloramphenicol; LVX, levofloxacin.

  2. a

    Breakpoint for resistance in accordance with CLSI document M31-A3 (2008) for veterinary pathogens.

  3. b

    Antimicrobial agent approved for cattle in Japan; the other 17 antimicrobials in this table are unapproved for cattle.

  4. c

    Breakpoint for resistance in accordance with CLSI document M100-S21 (2011) for isolates from human infections with Enterobacteriaceae.

  5. d

    Breakpoint for resistance is not defined in CLSI documents M31-A3 (2008) and M100-S21 (2011).

  6. e

    Susceptibilities were tested by disk diffusion method according to CLSI documents M31-A3 (2008) and M100-S21 (2011).

Ampicillinb256 to >512>512>512>512>512512 to >512≥32
Cefuroximeb64 to >256>256>256>256>256>256≥32c
CAZ16–644–32≤2 to 432–6416–64≤2 to 32≥16c
CAZ/CLA0·5 to >80·5≤0·25 to 1>88 to >8≤0·25 to >8 
CTX32 to >51264 to >512128–512512 to >51264 to >512≤16 to >512≥4c
CTX/CLA≤1 to 8≤1 to 8≤18 to >8>8≤1 to >8 
Cefpodoxime64 to >64>64>64>64>64>64≥8
Ceftriaxone64 to >512>512512 to >512512 to >51264 to >512128 to >512≥4c
Ceftiofurb64 to >512512 to >512>512512 to >512128 to >512256 to >512≥8
Cefquinomeb16 to >12832 to >12864 to >128128 to >12832 to >12816 to >128-d
Cefepime8 to >64≤8 to >6416 to >6464 to >64≤8 to >64≤8 to >64≥32c
Cefmetazole≤4 to 32≤4 to 32≤43232≤4 to >32≥64c
Moxalactam≤8≤8 to 32≤8≤8 to 16≤8≤8 to >32≥64c
Meropenem≤2≤2 to 4≤2≤2≤2≤2≥4c
Aztreonam≤8 to >6416 to >64≤864 to >64≤8 to >64≤8 to >64≥16c
Gentamicinb≤2 to >16≤2 to 4≤2 to >16≤2 to 16>16≤2≥16
Amikacin≤4 to 8≤4 to 8≤4≤4 to >16≤4≤4 to 8≥64
OTETb>16>16≤4 to >16>16>168 to >16≥16
SXTb≤0·5/9·5 to >4/76≤0·5/9·5 to >4/76≤0·5/9·5 to >4/76>4/76>4/76≤0·5/9·5 to >4/76≥4/76
Enrofloxacinb≤0·25 to >2≤0·25≤0·250·5 to >2≥2≤0·25≥2
Ciprofloxacin≤0·5 to >2≤0·5≤0·5≤0·5 to >2>2≤0·5≥4c
CefoxitineS to RS to RSRRS to R≤14 mmc
KANb,eS to RS to RS to RS to RRS to R≤13 mm
CHLeS to RS to RS to IS to RS or RS to R≤12 mm
LVXeS to RSSS to RS or RS≤13 mmc

According to the clinical charts and the purchase volume of antimicrobial agents provided by the member veterinary clinics, subcutaneous injections of enrofloxacin (4 mg kg−1 body weight) were administered to one cow and one calf with diarrhoea once daily for 2–4 days. The other 26 calves, heifer and cows were not treated with oxyimino-cephalosporins or fluoroquinolones. Between 2006 and 2010, the member veterinary clinics used 2016 daily doses of ceftiofur and 1408 daily doses of enrofloxacin per year for approximately 130 000 cows and heifers of 1000 dairy farms in Nemuro Subprefecture. Daily doses were based on the weight of an adult cow. Neither cefquinome nor the other fluoroquinolones were used.

MLST, serotyping and PFGE of 23 Escherichia coli isolates

The 23 E. coli isolates were grouped into 18 STs (Table 1). The eight CTX-M-15–producing E. coli isolates belonged to six STs and were non-ST131. Four isolates belonged to ST10, and two isolates belonged to ST44 and ST48 [ST 10 complex (STC10)]. Five isolates belonged to ST46 (STC46), ST57 (STC350), ST58 (STC155), ST69 (STC69) and ST88 (STC 23) (Table 1). Most of the E. coli isolates were O-antigen untypable (OUT), but the serotypes of six isolates were O8:H7, O15:H-antigen untypable (HUT), O28ac:HUT, O44:H18, O124:H4 and O164:HUT (Table 1).

Pulsed-field gel electrophoresis analysis with Xba I and Bln I divided the 23 E. coli isolates into 21 genetic profiles. Nineteen of the genetic profiles consisted of a single strain. Isolates FE03 and FE05, isolated from two different cows from farm M, showed PFGE type K, ST1167 and blaCTX-M-15; and FE06 and FE07, isolated from two different calves from farm IC, showed PFGE type G, ST10 and blaCTX-M-2, but showed distinct serotypes. Conversely, isolates FE2 and FE9, isolated from two different cows from different farms, showed the ST10 but differed on PFGE type and serotype. The isolates FE11 and FE23, isolated from two different cows from different farms, showed ST744 but differed on PFGE type (Table 1, Fig 1).

Figure 1.

Pulsed-field gel electrophoresis (PFGE) types using Xbl 1 and Bln-I enzymes of 23 Escherichia coli isolates producing CTX-M and CMY-2 from faeces. PFGE analysis with Xba I and Bln I divided the 23 E. coli isolates into 21 genetic profiles. Nineteen of the genetic profiles consisted of a single strain. Dendrogram of the combined PFGE patterns of Xba I and Bln I was constructed with Dice coefficient and the unweighted pair group method, and the similarity of the average from the two experiments. PFGE types were distinguished at a cut-off of ≥90% similarity.


The incidence and herd prevalence of ESBL-positive E. coli among dairy farms in the present study were lower than those of French cattle (Hartmann et al. 2012). In agreement with Japanese studies of broiler chickens (Hiroi et al. 2012), three CTX-M clusters (CTX-M-15, CTX-M-2 and CTX-M-14) were detected in the present study. Among human and animal isolates in Western European countries, Japan, China, Taiwan and South Korea, the most common CTX-M types were the CTX-M-1 cluster (CTX-M-1, CTX-M-3, CTX-M-15, CTX-M-35, CTX-M-55) and CTX-M-9 cluster (CTX-M-9, CTX-M-14, CTX-M-27); however, CTX-M-15 predominated among human isolates in India (Carattoli 2008; Hawkey 2008; Suzuki et al. 2009; Overdevest et al. 2011; Harada et al. 2012; Kuroda et al. 2012; Tamang et al. 2012). Except for the dominance of CTX-M-2, our results are similar to these previous reports other than the Indian study.

Most of the CTX-M-15 producers also harboured blaTEM-1 and were resistant to ceftazidime and fluoroquinolones. CTX-M types other than CTX-M-15, CTX-M-16 and CTX-M-27 efficiently hydrolyse cefotaxime and ceftriaxone but not ceftazidime (Bonnet 2004). The blaTEM-1 and blaSHV-11 genes detected in the present study encode non-ESBL enzymes (Paterson and Bonomo 2005).

Annual ceftiofur and enrofloxacin use was low in the large dairy cattle population evaluated by our clinics, and their use does not appear to be a major risk factor for ESBL-producing Enterobacteriaceae infections. An earlier human study (Colodner et al. 2004) identified the previous use of fluoroquinolones, penicillins and second- and third-generation cephalosporins as risk factors for ESBL-producing E. coli and K. pneumoniae infections. Thus, the spread of blaCTX-M and blaCMY-2 may be due in part to the frequent use of mastitis preparations consisting of first- and second-generation cephalosporins in member veterinary clinics in Nemuro Subprefecture and using waste milk mixed with cephalosporins to feed calves. Previous studies (Lowrance et al. 2007; Daniels et al. 2009) revealed that ceftiofur use had a transient effect on the selection of ceftiofur resistance in commensal E. coli at the individual calf level. However, most CTX-M-15 producers are resistant to both oxyimino-cephalosporins and fluoroquinolones.

Consistent with our findings, the E. coli clones ST10, ST44, ST48, ST57, ST58, ST69 and ST88, producing CTX-M-1, CTX-M-2, CTX-M-9, CTX-M-14, CTX-M-15, SHV-5 and SHV-12 were isolated from chicken caeca in the UK (Randall et al. 2011), and retail chicken meat in the Netherlands (Overdevest et al. 2011). The clonal diversity of STs and PFGE types observed in our study among bovine E. coli strains harbouring the same type of blaCTX-M is in agreement with a recent canine study (Tamang et al. 2012). Both PFGE and MLST were unable to differentiate isolates FCE03 and FCE05 and isolates FCE06 and FCE07, which differed according to serotype, indicating that the two clusters of two isolates each were closely related (Table 1, Fig 1). However, four ST10 isolates showed three distinct PFGE types, and two ST744 isolates showed two distinct PFGE types. It should be noted that PFGE showed a greater discriminatory ability than MLST in the present study, differentiating strains from different farms and strains of different serotypes. In general, housekeeping genes have low mutation rates, but PFGE is thought to easily detect mutations in E. coli.

In this study, we did not detect the CTX-M-15–producing E. coli ST131 (O25:H4), which is a worldwide pandemic multidrug-resistant strain that infects humans (Kuroda et al. 2012), nor epidemic E. coli strains ST38 (O86:H18) or ST131 (O25:H4), which produce the CTX-M-9 group of ESBLs that infect humans, canines and felines in Japan (Suzuki et al. 2009; Harada et al. 2012). Madec et al. (2012) revealed that bla CTX-M-15–carrying plasmids from cattle-derived non-ST131 E. coli isolates were highly similar to those found in human-derived ST131 E. coli isolates.

Recently, a low incidence of blaCTX-M-positive E. coli was reported in retail beef (Overdevest et al. 2011), suggesting that bovine blaCTX-M-positive E. coli may pose a relatively low risk to public health.

In conclusion, the incidence of E. coli and three genera of Enterobacteriaceae producing CTX-M-15, CTX-M-2, CTX-M-14 or CMY-2 in bovine faeces was 3·1% of the 897 faecal samples. The herd prevalence was 5·2% of the 381 dairy farms in Nemuro Subprefecture, Japan. The 23 E. coli isolates showed clonal diversity, as assessed by MLST analysis and PFGE typing. The eight E. coli isolates producing CTX-M-15 belonged to six STs and were non-ST131; most were resistant to ceftazidime and fluoroquinolones. The present results indicate that dairy cattle are a potential reservoir of ESBL-producing bacteria. Thus, routine monitoring for CTX-M (especially CTX-M-15) and CMY-2 producers derived from faeces and mastitis milk and the prudent use of antimicrobial agents are necessary to prevent their clonal spread.