A

aarKiU5b©

K

1

isfstpb�am

0d

International Journal of Antimicrobial Agents 29 (2007) 332–337

Short communication

Characterisation of KLUA-9, a �-lactamase from extended-spectrumcephalosporin-susceptible Kluyvera ascorbata,

and genetic organisation of blaKLUA-9

Marıa Margarita Rodrıguez a, Pablo Power a, Cedric Bauvois b, Jose Di Conza a,Juan A. Ayala c, Moreno Galleni b, Gabriel Gutkind a,∗

a Catedra de Microbiologıa, Facultad de Farmacia y Bioquımica, Universidad de Buenos Aires, Junın 954 (1113), Buenos Aires, Argentinab Centre d’Ingenierie des Proteines, Universite de Liege, Institut de Chimie, B6 (B4000), Sart Tilman, Liege, Belgium

c Centro de Biologıa Molecular ‘Severo Ochoa’, CSIC-UAM, Campus de Cantoblanco 28049, Madrid, Spain

Received 21 July 2006; accepted 15 September 2006

bstract

This study characterised the genetic environment of the chromosomally encoded blaKLUA-9 gene from a clinical Kluyvera ascorbata isolatend performed a kinetic characterisation of KLUA-9. Purified KLUA-9 showed the highest catalytic efficacies towards benzylpenicillin,mpicillin, piperacillin, first-generation cephalosporins, cefuroxime and cefoperazone; like other ‘cefotaximases’, it showed a much higherate of hydrolysis of cefotaxime than ceftazidime, whilst dicloxacillin, cefoxitin and imipenem behaved as poor substrates. A 9 kb insert from. ascorbata was cloned (Escherichia coli KK68C1) and sequenced. blaKLUA-9 and its 266 bp upstream flanking region (almost identical to the

ntegron-associated blaCTX-M-2) are preceded by an aspat variant, a ypdABC-like operon and two open reading frames with unknown functions.

nlike ISCR1-associated blaCTX-M-2 genes, we failed to detect the putative orf513 recombination sites. Instead, we were able to localise thebp target sites for insertion of ISEcp1B, suggesting that this element could be responsible for future (or still undetected) mobilisation oflaKLUA-9 to more efficiently transferred elements.

2006 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.

MedsM[

ieg

eywords: Oxyiminocephalosporinase; ESBL; CTX-M; ISEcp1B; orf513

. Introduction

Most plasmid-borne �-lactamases are believed to orig-nate from chromosomal genes. Although not known inome cases, these genes are probably transferred by dif-erent recombination events to more efficiently transferredtructures. For example, some CTX-M-derived enzymes,he so-called ‘cefotaximases’ (CTX-M-ases), have been pro-osed to derive from soil, environmental or commensalacteria such as Kluyvera species. To date, the CTX-M

-lactamases (and their putative progenitor counterparts)ccount for more than 50 representatives clustered in fiveain groups, with CTX-M-1, CTX-M-2, CTX-M-8, CTX-

∗ Corresponding author. Tel.: +54 11 4964 8285; fax: +54 11 4964 8274.E-mail address: ggutkind@ffyb.uba.ar (G. Gutkind).

tarsep

924-8579/$ – see front matter © 2006 Elsevier B.V. and the International Societyoi:10.1016/j.ijantimicag.2006.09.015

-9 and CTX-M-25 being the first described members ofach group [1]. The CTX-M-2 subgroup appears to beerived from KLUA-1 of Kluyvera ascorbata [2], CTX-M-8ubgroup from KLUG-1 of Kluyvera georgiana [3] and CTX--9 subgroup from KLUY enzymes Kluyvera georgiana

4].Recently, we reported a K. ascorbata strain harbour-

ng a chromosomally encoded CTX-M-3 and proposed thisnzyme to be the probable origin of the CTX-M-1 sub-roup [5], which was previously shown to be also relatedo Kluyvera cryocrescens KLUC-1 [6]. We also proposeddoption of the term ‘born oxyiminocephalosporinases’ to

efer to pre-existing (natural) chromosomal cefotaximasesuch as CTX-M-3 in order to differentiate them from otherxtended-spectrum �-lactamases (ESBLs) that emerge byoint mutation from narrower-spectrum variants [5].

of Chemotherapy. All rights reserved.

rnal of A

hcmamcob

Kigr

2

2

t(fi(ipgpbpar

2

tCNuwcct

2

PCotTHi

swcc(GSa

amGKEcwpHcaiiMcT

2

aitbwaaE

2

aLdair(p

me

M.M. Rodrıguez et al. / International Jou

Although CTX-M-derived enzymes show more efficientydrolysis towards cefotaxime or ceftriaxone than towardseftazidime, resistance to the latter may arise from pointutations, trading off some activity against the first [7],

lthough activity towards ceftazidime may still be one orore orders lower than for cefotaxime. This preference for

efotaxime is a usual difference from most of the TEM-r SHV-derived ESBLs, which are commonly considered asetter ceftazidime-hydrolysing enzymes [8].

In this report, we performed a biochemical analysis onLUA-9, a �-lactamase produced by a K. ascorbata clinical

solate, and studied the genetic environment of the blaKLUA-9ene to search for putative target sites that could lead to itsecruitment and plasmid-mediated dissemination.

. Materials and methods

.1. Bacterial strains and plasmids

Kluyvera ascorbata strain 68 was isolated from a spu-um sample at Hospital de Clınicas ‘Jose de San Martın’Buenos Aires, Argentina) in 1995. The strain was identi-ed using standard biochemical criteria, commercial systemsAPI 20E; bioMerieux, Marcy l’Etoile, France) and sequenc-ng of 16S rDNA as a genetic criterion for identification, asreviously described [9]. Escherichia coli Top10F’ (Invitro-en, Carlsbad, CA) was used as the recipient for recombinantlasmids. Plasmid vectors used were: the high-copy num-er pK19 vector (kanamycin resistance) [10]; and a modifiedBAD/Myc-His expression vector (Invitrogen) in which thempicillin resistance marker was replaced by kanamycinesistance.

.2. Antimicrobial agents and susceptibility tests

Minimum inhibitory concentrations were determined byhe agar dilution method following the guidelines of thelinical and Laboratory Standards Institute (formerly theational Committee for Clinical Laboratory Standards) [11]sing a Steers’ multipoint inoculator. Detection of ESBLsas performed by double-disk synergy test using ampi-

illin/clavulanic acid disks (10/10 �g) placed between 30 �gefotaxime and ceftazidime disks. All disks were from Bri-ania, Buenos Aires, Argentina.

.3. Molecular biology techniques

Plasmid DNA extraction was performed using GFX Microlasmid Prep kit (GE Healthcare, Buenos Aires, Argentina).hromosomal DNA was extracted by a modification of theriginal method described by Gerhardt et al. [12] and par-

ially digested with HindIII (Gibco BRL, Rockville, MD).he resulting fragments were purified and ligated into aindIII-digested pK19 vector. Recombinant plasmids were

ntroduced by transformation and the transformants were

dLeB

ntimicrobial Agents 29 (2007) 332–337 333

elected on Luria–Bertani (LB) agar plates supplementedith 20 mg/L kanamycin and 0.5 mg/L cefotaxime. Clones

ontaining the blaKLUA gene were detected by polymerasehain reaction (PCR) using specific primers for blaCTX-M-2BlaI: 5′-TTAATGATGACTCAGAGCATTC-3′; BlaII: 5′-ATACCTCGCTCCATTTATTG-3′). pS21 plasmid fromalmonella enterica serovar Infantis S21 strain was used aspositive control for blaCTX-M-2 [13].

To purify KLUA-9, the blaKLUA-9 gene (974 bp) wasmplified by PCR from the K. ascorbata 68 chromoso-al DNA using modified primers (KLUK-Dir, 5′-CGAGGGTACCACTAATAGAGG-3′, KpnI restriction site; andLUE-Rev, 5′-AAGACAAGACTGGAATTCAGGAGC-3′,coRI restriction site), in a Biometra T-Gradient thermocy-ler (Biometra Inc., Tampa, FL). Amplicons were digestedith EcoRI and KpnI (New England Biolabs, Beverly, MA),urified and subsequently ligated in a modified pBAD/Myc-is expression vector and used to transform competent E.

oli Top10F’ cells. Recombinant clones were selected on LBgar plates supplemented with 30 mg/L kanamycin, 30 �Msopropyl �-d-1-thiogalactopyranoside (IPTG; Sigma Chem-cal Co., St Louis, MO) and 80 mg/L X-Gal (Boehringer,

annheim, Germany). Presence of the blaKLUA gene wasonfirmed by sequencing. One selected clone, called E. coliBK-1, was used for further purification of KLUA-9.

.4. DNA sequencing and computer analysis

The DNA sequence on both strands was determined by theutomated Sanger dideoxy chain termination method [14]n an ABI PRISM 3700 DNA Analyser (Applied Biosys-ems, Foster City, CA). Sequence analyses were performedy National Center for Biotechnology Information (http://ww.ncbi.nlm.nih.gov/) and EBI (http://www.ebi.ac.uk/)

nalysis tools. Theoretical signal peptide, molecular weightnd isoelectric point (pI) for KLUA-9 were determined byxPASy Proteomic tools (http://www.expasy.org).

.5. Protein electrophoresis techniques

Crude extracts were obtained from log-phase culturesfter centrifugation at 10 000 × g for 15 min at 4 ◦C. �-actamases were analysed by isoelectric focusing usingifferent �-lactamases with known pI values (SHV-2, TEM-1nd CTX-M-2) and pI markers (Pharmacia isoelectric focus-ng calibration kit) as standards. �-Lactamase activity wasevealed by the agar iodometric system using ampicillin500 mg/L) and ceftriaxone (1000 mg/L) as substrates, asreviously described [15].

The molecular weight of the �-lactamases was deter-ined by sodium dodecyl sulphate polyacrylamide gel

lectrophoresis (SDS-PAGE) in 15% polyacrylamide gels, in

uplicate, employing molecular weight standards (Bio-Radaboratories, Tecnolab S.A., Buenos Aires, Argentina). Afterlectrophoresis, one gel was stained with Coomassie Brilliantlue R250, whilst the other one was renatured with 20 mg/mL

3 rnal of A

baoA

2

tcoa3ddscweacG0lup1p

2

m�msCoayutwKsiswaet

2

n

3

cE

ctpcaa

h((s

lefwrtacmati(aetrsl(w[peaof both antibiotics. The most stable antibiotics appear to becefoxitin, ceftazidime, cefepime, aztreonam and imipenem.For some of these drugs, such as cefoxitin and imipenem,a high affinity was observed (low Km values), but very low

34 M.M. Rodrıguez et al. / International Jou

enzylpenicillin in 0.05 M phosphate buffer (PB) (pH 7.0)nd �-lactamase activity was revealed by the iodometricverlay system using ampicillin (500 mg/L) as substrate (G.micosante, personal communication).

.6. Expression and purification of KLUA-9 enzyme

Overnight cultures of recombinant E. coli TBK-1 (con-aining the blaKLUA-9 gene) were diluted (1/50) in LB brothontaining 30 mg/L kanamycin and grown at 37 ◦C until anptical density at 600 nm of 0.5. l(+)-arabinose was addedt a final concentration of 0.2% and cultures were grown at7 ◦C for 3 h for induction. Crude extracts were obtained asescribed above. Escherichia coli TBK-1 crude extract wasialysed overnight at 4 ◦C against 0.05 M PB (pH 8.5). Theample was loaded on a DEAE-Sephadex column (Pharma-ia, Uppsala, Sweden) equilibrated in the same buffer. Afterashing the column, bound proteins were eluted with a lin-

ar NaCl gradient (0 M to 1 M) in PB (pH 8.5). �-Lactamasectivity of collected fractions was tested with 100 �M nitro-efin. Active fractions were pooled and loaded on a Sephadex-100 column (Pharmacia). Elution was performed with.01 M PB (pH 7.0) and the �-lactamase activity of col-ected fractions was detected by an iodometric overlay systemsing 500 mg/L ampicillin as substrate. Active fractions wereooled and concentrated by ultrafiltration using a Centricon-0 (Amicon, Beverly, MA). The purity of the final enzymaticreparation was estimated at 90% by SDS-PAGE.

.7. Determination of the main kinetic parameters

Hydrolysis of the antibiotics by purified KLUA-9 wasonitored by following the absorbance variation of the-lactam solution in 0.01 M PB (pH 7.0). All the measure-ents were evaluated on a UV-2101 PC, UV-VIS scanning

pectrophotometer (Shimadzu Scientific Instruments Inc.,olumbia, MD). Reactions were performed in a total volumef 1 mL at 25 ◦C. The steady-state kinetic parameters (Kmnd kcat) were determined by analysing the complete hydrol-sis time courses as described by De Meester et al. [16], orsing the Hanes’ linearization of the Michaelis–Menten equa-ion. In the case of low values of Km, or under conditionshen the antibiotic behaved as a poor substrate, apparentm values were determined as competitive inhibition con-

tants (Ki), monitoring the residual activity of the enzymen the presence of the drug and nitrocefin 75 �M as reporterubstrate. For suicide inhibitors, IC50 values (instead of Ki)ere determined by interpolating the inhibitor concentration

ble to reduce the initial activity to 50% under the samexperimental conditions, without any further mathematicalransformation.

.8. Nucleotide sequence accession numbers

Sequence data were deposited in the GenBank/EMBLucleotide databases under the accession no. AM086668.

Faaa(

ntimicrobial Agents 29 (2007) 332–337

. Results and discussion

No plasmid could be detected in K. ascorbata 68. Afterloning fragments of HindIII-digested chromosomal DNA,. coli KK68C1 recombinant clone was selected for analysis.

Antimicrobial susceptibilities of K. ascorbata 68 and E.oli KK68C1 are shown in Table 1. Double-disk synergyests were positive in E. coli KK68C1, as expected for strainsroducing plasmid-encoded class A ESBLs (not shown). Thislone harboured an insert of ca. 9 kb, which was sequencednd the presence of putative open reading frames (orfs) wasnalysed (Fig. 1).

The insert included an 876 bp orf, located at the 3′-end,aving >99% nucleotidic identity with the reported blaKLUA-9accession number AJ427466), with a single base differenceG831A), and 100% identity with the deduced amino acidequence of KLUA-9.

Along with other enzymes found in K. ascorbata iso-ates, KLUA-9 shares strong similarity with those CTX-Mnzymes included in the CTX-M-2-group [2]. We found onlyour and five amino acid substitutions in KLUA-9 comparedith CTX-M-2 (98.5% identity) and Toho-1 (98.1% identity),

espectively. Only one enzyme was detected after isoelec-ric focusing using crude extracts both from K. ascorbata 68nd E. coli KK68C1. The enzyme had a pI of 7.7 (theoreti-al pI of 7.2) and a molecular mass of 28.2 kDa (theoreticalolecular weight 28 117) and was active against 500 mg/L

mpicillin used as substrate in the iodometric overlay sys-em. The main kinetic parameters of KLUA-9 are presentedn Table 2. KLUA-9 exhibited the highest catalytic efficacieskcat/Km) towards benzylpenicillin, ampicillin, piperacillinnd first-generation cephalosporins. However, a more het-rogeneous behaviour of KLUA-9 against cephalosporinshan penicillins occurred. Like other CTX-M-ases, the mostemarkable feature is the significantly higher rate of hydroly-is of cefotaxime than ceftazidime. In fact, hydrolysis of theatter could not be observed even after prolonged reactions>30 min). It is noteworthy that some intriguing differencesere noted with the published kinetic parameters of Toho-1

17]. As no comprehensive analysis of CTX-M-2 has beenublished, it is difficult to understand these differences; how-ver, affinities for ceftazidime and cefotaxime in our caseppear to correlate much better with microbiological effects

ig. 1. Schematic representation of the insert of pKK68C1 from Kluyverascorbata 68. The sequence comprises: blaKLUA-9, located at the 3′ extreme;spat, encoding an aspartate aminotransferase; the ypdABC-like genes;nd two open reading frames (orfs) encoding hypothetical proteins of 100orf100) and 418 (orf418) amino acids. See text for details.

M.M. Rodrıguez et al. / International Journal of Antimicrobial Agents 29 (2007) 332–337 335

Table 1Minimum inhibitory concentrations (MICs) for Kluyvera ascorbata 68 strain and derived Escherichia coli clones

Antibiotic MIC (mg/L)

K. ascorbata 68 E. coli KK68C1 E. coli Top10F’/pK19 E. coli TBK-1a E. coli Top10F’/pBAD/Myc-His (KANr)

E. coli Top10F’

Ampicillin 16 >1 024 2 4 2 2Ampicillin/clavulanic

acid (2:1)1/0.5 4/2 2/1 2/1 2/1 2/1

Piperacillin 4 >256 2 2 1 1Cephalothin 16 >1 024 4 4 2 2Cefoxitin 2 4 2 2 2 2Cefotaxime 0.063 64 0.032 0.032 ≤0.016 ≤0.016Cefotaxime/clavulanic

acidb≤0.016 0.063 N.D. N.D. N.D. N.D.

Ceftazidime 0.032 2 0.125 0.063 0.063 0.125Ceftazidime/clavulanic

acidb≤0.016 0.125 N.D. N.D. N.D. N.D.

Cefepime 0.016 8 0.016 0.016 0.016 0.016Imipenem 0.063 0.063 0.063 0.125 0.063 0.063Aztreonam 0.016 16 0.032 0.032 0.032 0.032Kanamycin 1 >256 >256 >256 >256 1

N

k0sOht

slt

Csw

fCcoyaa

9b

TK

S

BAPCNCCCCCCCAI

N

.D., not determined.a No arabinose induction.b Fixed concentration of lithium clavulanate (4 mg/L).

cat values led to very poor catalytic efficiencies (0.007% and.012% relative kcat/Km, respectively); this behaviour corre-ponds to that for poor substrates or transient inactivators.n the other hand, ceftazidime and cefepime display veryigh Km values, making these drugs hardly recognised byhe active site.

All tested class A �-lactamase inhibitors appear to haveimilar activities against KLUA-9, with IC50 values forithium clavulanate of 338 nM, sulbactam of 294 nM andazobactam of 179 nM.

A wide range of variability was observed between theTX-M-type and other class A enzymes, with kcat/Km values

panning several orders of magnitude among them, especiallyith cephalosporins. As previously stated, significant dif-

tgdr

able 2inetic parameters of KLUA-9

ubstrate Km (�M) kcat (s−1)

enzylpenicillinb 16 ± 1 10 ± 0.1mpicillinb 15 ± 1 7 ± 0.1iperacillinb 11 ± 1 5 ± 0.1arbenicillinb 17 ± 1 1 ± 0.1itrocefin 47 ± 6 29 ± 2ephalothin 49 ± 3 56 ± 2efoxitinb 10 ± 1 (4.3 ± 0.1) × 10−4

efuroximeb 19 ± 1 2.5 ± 0.1efotaximeb 43 ± 1 3.3 ± 0.1eftazidimeb >1000 N.D.efoperazoneb 1.2 ± 0.1 0.22 ± 0.02efepimeb >1000 N.D.ztreonamb 390 ± 40 0.24 ± 0.02

mipenemb 3.4 ± 0.4 (2.4 ± 0.1) × 10−4

.D., not determined.a Relative to benzylpenicillin.b Km determined as Ki.

erences among the available kinetic parameters of differentTX-M-ases [18] make it even more difficult to establish alear relationship between the kinetic parameters and groupr family of �-lactamases. Currently, a comprehensive anal-sis of at least a few ‘already reported’ kinetic data is missingnd it may be interesting to have some of these data re-nalysed.

Downstream of blaKLUA-9, a 385 bp sequence having8.9% identity with the corresponding region downstream oflaKLUA-1 was found (accession no. AJ272538). As expected,

here was also a high similarity compared with the homolo-ous region in the blaCTX-M-2-harbouring unusual integrons,epicting 98.9% and 99.2% identity with In116 and InS21,espectively [13,19].

kcat/Km (�M−1 s−1) Relative kcat/Km (%)a

0.63 ± 0.04 1000.47 ± 0.04 770.46 ± 0.04 770.061 ± 0.005 100.61 ± 0.11 1021.15 ± 0.13 192(4.3 ± 0.4) × 10−5 0.0070.13 ± 0.01 220.077 ± 0.003 13– –0.181 ± 0.006 30– –(6.2 ± 0.6) × 10−4 0.10(6.9 ± 0.9) × 10−5 0.012

3 rnal of A

aiChukt‘Ntu

sMtas

mIr

sgrmIr2ss

tblbo(r

ptt

egIMaoml

br

awcs

lhbtao

A

ddC(La(ft

R

36 M.M. Rodrıguez et al. / International Jou

Different orfs were detected upstream of blaKLUA-9: (i)spat, encoding an aspartate aminotransferase having 99%dentity with the enzyme of K. ascorbata (accession no.AC03612); (ii) a ypdABC-like operon, encoding threeypothetical proteins with 80–88% identity with similar reg-latory and metabolic proteins from E. coli (LytS, histidineinase (ypdA, accession no. NP 416881); LytT, a signalransducer (ypdB, accession no. NP 416882); and a putativeHTH’ ARAC-type regulatory protein (ypdC, accession no.P 416883)); and (iii) two orfs encoding hypothetical pro-

eins of 100 (orf100) and 418 (orf418) amino acids, withnknown function.

ISEcp1-like elements appear to act as a factor in the dis-emination of blaCTX-M-9- and blaCTX-M-1-like genes [20].

oreover, it has recently been shown that ISEcp1B is ableo mobilise a blaCTX-M gene from a K. ascorbata strain toconjugative plasmid location and be further transferred to

usceptible strains [21].On the other hand, orf513 recombinase appears to have

ediated the incorporation of blaCTX-M-2 and blaCTX-M-9 inn6/In7-related integrons, probably by an ISCR1-mediated,olling circle replication [22].

A 266 bp region immediately upstream of blaKLUA-9howed 99.6% and 99.2% identity with the homolo-ous regions upstream of InS21 and In116 blaCTX-M-2,espectively [13,19], reinforcing the hypothesis of a com-on origin of these �-lactamase-encoding genes. In the

SCR1-containing integrons, a putative orf513-mediatedecombination site is present in a second region spanning40 bp upstream of the 266 bp sequence [13,19], which ineveral cases is also associated with an ISEcp1-type insertionequence [20].

It is noteworthy that in the chromosome of K. ascorbata,he 240 bp region associated with orf513 (present in plasmid-orne homologous genes) is missing. In turn, we were able toocalise 5 bp target sites for ISEcp1B insertion, as describedy Lartigue et al. [21], at exactly the same positions upstreamf the blaKLUA/CTX-M-2 ATG: TACTA, AATAC and TAATAat 19 bp, 21 and 43 bp, and 22 bp upstream of the bla genes,espectively).

The presence of these putative IS target sites at similarositions from blaKLUA-9 suggests that the mobilisation ofhis gene could occur by an ISEcp1B-mediated recombina-ion event.

We could therefore speculate at least two independentvolutionary pathways of these genes: whereas blaCTX-Menes in our region could have been mainly recruited bySCR1-mediated events (owing to the prevalence of CTX-

-2), some chromosomally encoded bla genes, such as K.scorbata 68 blaKLUA-9, could have served as the sourcef other plasmid-borne blaCTX-M-2-like genes, probablyobilised by an ISEcp1B-mediated recombination in other

ocations.Questions arise about the apparent absence of plasmid-

orne blaCTX-M-2-like genes associated with ISEcp1 in ouregion: whether there is an underestimation of these vari-

[

ntimicrobial Agents 29 (2007) 332–337

nts (probably owing to failures in their detection), orhether ISCR1-mediated blaCTX-M-2 genes are more effi-

iently selected by antibiotic pressure in local clinicalettings.

Plasmid-borne blaCTX-M-2 genes so far detected in iso-ates from Argentina and neighbouring countries appear toave been acquired by an orf513-mediated (ISCR1) recom-ination. However, in these genes, the proposed 5 bp ISEcp1Barget sites located at 19, 21 and 22 bp upstream of blaCTX-M-2re also present, but not the 43 bp site identified upstream ofther blaCTX-M genes [21].

cknowledgments

This work was supported in part by grants from: Secretarıae Ciencia y Tecnologıa (PICT 2004/14234), Universidade Buenos Aires and Ministerio de Salud (Beca Onativia-arrillo), Argentina, to G.G.; the European Commission

Targeted Project COBRA) to M.G.; and the FP6 projectSHM-CT-503335 of the European Union to J.A.A. G.G.nd P.P. are members of Carrera del Investigador CientıficoCONICET). We are grateful to Dr Ferraro and Dr Martinorom Catedra de Farmacognosia, FFyB, UBA, for allowinghe use of their facilities.

eferences

[1] Bonnet R. Growing group of extended-spectrum �-lactamases: theCTX-M enzymes. Antimicrob Agents Chemother 2004;48:1–14.

[2] Humeniuk C, Arlet G, Gautier V, Grimont P, Labia R, PhilipponA. �-Lactamases of Kluyvera ascorbata, probable progenitors ofsome plasmid-encoded CTX-M types. Antimicrob Agents Chemother2002;46:3045–9.

[3] Poirel L, Kampfer P, Nordmann P. Chromosome-encoded Ambler classA �-lactamase of Kluyvera georgiana, a probable progenitor of a sub-group of CTX-M extended-spectrum �-lactamases. Antimicrob AgentsChemother 2002;46:4038–40.

[4] Olson AB, Silverman M, Boyd DA, et al. Identification of a progen-itor of the CTX-M-9 group of extended-spectrum �-lactamases fromKluyvera georgiana isolated in Guyana. Antimicrob Agents Chemother2005;49:2112–5.

[5] Rodrıguez MM, Power P, Radice M, et al. Chromosome-encoded CTX-M-3 from Kluyvera ascorbata: a possible origin of plasmid-borneCTX-M-1-derived cefotaximases. Antimicrob Agents Chemother2004;48:4895–7.

[6] Decousser JW, Poirel L, Nordmann P. Characterization of a chromoso-mally encoded extended-spectrum class A �-lactamase from Kluyveracryocrescens. Antimicrob Agents Chemother 2001;45:3595–8.

[7] Welsh KJ, Barlow M, Tenover FC, et al. Experimental prediction of theevolution of ceftazidime resistance in the CTX-M-2 extended-spectrumbeta-lactamase. Antimicrob Agents Chemother 2005;49:1242–4.

[8] Bradford PA. Extended-spectrum �-lactamases in the 21st century:characterization, epidemiology and detection of this important resis-tance threat. Clin Microbiol Rev 2001;14:933–51.

[9] Weisburg WG, Barns SM, Pelletier DA, Lane DJ. 16S ribosomalDNA amplification for phylogenetic study. J Bacteriol 1991;173:697–703.

10] Pridmore RD. New and versatile cloning vectors with kanamycin-resistance marker. Gene 1987;56:309–12.

rnal of A

[

[

[

[

[

[

[

[

[

[

[

M.M. Rodrıguez et al. / International Jou

11] National Committee for Clinical Laboratory Standards. Performancestandards for antimicrobial disk susceptibility tests. 8th ed. Approvedstandard. M2-A8. Fourteenth Informational Supplement, vol. 24.Wayne, PA: NCCLS; 2004.

12] Gerhardt P, Murray RGE, Wood WA, Krieg NR. Methods for generaland molecular bacteriology. Washington, DC: ASM Press; 1994.

13] Di Conza J, Ayala JA, Power P, Mollerach M, Gutkind G. Novel class1 integron (InS21) carrying blaCTX-M-2 in Salmonella enterica serovarInfantis. Antimicrob Agents Chemother 2002;46:2257–61.

14] Sanger I, Nicklen S, Coulson A. DNA sequencing with chain terminat-ing inhibitors. Proc Natl Acad Sci USA 1977;74:5463–7.

15] Rossi A, Lopardo H, Woloj M, et al. Non-typhoid Salmonella spp.resistant to cefotaxime. J Antimicrob Chemother 1995;36:697–702.

16] De Meester F, Joris B, Reckinger G, Bellefroid-Bourguignon C, Frere

JM, Waley SG. Automated analysis of enzyme inactivation phenomena.Application to �-lactamases and DD-peptidases. Biochem Pharmacol1987;36:2393–403.

17] Ishii Y, Ohno A, Taguchi H, Imajo S, Ishiguro M, Matzusawa H.Cloning and sequence of the gene encoding a cefotaxime-hydrolyzing

[

ntimicrobial Agents 29 (2007) 332–337 337

class A �-lactamase isolated from Escherichia coli. Antimicrob AgentsChemother 1995;39:2269–75.

18] Walther-Rasmussen J, Hoiby N. Cefotaximases (CTX-M-ases), anexpanding family of extended-spectrum �-lactamases. Can J Microbiol2004;50:137–65.

19] Power P, Galleni M, Conza JD, Ayala JA, Gutkind G. Description ofIn116, the first blaCTX-M-2-containing complex class 1 integron found inMorganella morganii isolates from Buenos Aires, Argentina. J Antimi-crob Chemother 2005;55:461–5.

20] Eckert C, Gautier V, Arlet G. DNA sequence analysis of the geneticenvironment of various blaCTX-M genes. J Antimicrob Chemother2006;57:14–23.

21] Lartigue M-F, Poirel L, Aubert D, Nordmann P. In vitro analysis ofISEcp1B-mediated mobilization of naturally occurring �-lactamase

gene blaCTX-M of Kluyvera ascorbata. Antimicrob Agents Chemother2006;50:1282–6.

22] Toleman MA, Bennett PM, Walsh TR. ISCR elements: novel gene-capturing systems of the 21st century? Microbiol Mol Biol Rev2006;70:296–316.