Multiclonal Outbreak of Klebsiella pneumoniae Producing Extended-Spectrum {beta}-Lactamase CTX-M-2 and Novel Variant CTX-M-59 in a Neonatal Intensive Care Unit in Brazil

An outbreak of cephalosporin-resistant Klebsiella pneumoniaeoccurred in a neonatal intensive care unit in São Paulo,Brazil. Of the 10 pulsotypes identified during the outbreakand follow-up periods, nine produced CTX-M-2 or its new variantCTX-M-59 and one produced SHV-5. bla_CTX-M-2_/₅₉ genes were locatedon closely related plasmids that were transferable.

INTRODUCTION

Top

INTRODUCTION

Klebsiella pneumoniae has long been known to cause outbreaksof infections in neonatal intensive care units (NICUs). Sincethe advent of extended-spectrum β-lactamases (ESBLs), manysuch outbreaks have been associated with ESBL-producing strains(7). Most outbreak investigations of ESBL-producing K. pneumoniaein NICUs have attributed the epidemics to dissemination of singleclones, whereas only a few studies reported involvement of morethan one strain in an outbreak setting (2, 6). We here describethe molecular epidemiology of a multiclonal outbreak in a BrazilianNICU caused by K. pneumoniae that produced CTX-M-type ESBLsincluding CTX-M-59, a novel variant of CTX-M-2.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Bacterial strains and plasmids. The outbreak occurred in the NICU of a hospital near SãoPaulo in October 2003. A total of 10 cephalosporin-resistantK. pneumoniae strains were isolated at the time. Five were obtainedfrom clinical specimens (one each from blood, cerebrospinalfluid, catheter tip, urine, and oropharynx), four from perirectalswabs, and one from a mother's milk. Subsequently, 37 nonrepetitive,cephalosporin-resistant K. pneumoniae strains were isolatedfrom perirectal, axillary, and oropharyngeal swabs during afollow-up investigation between May and August 2004. Additionally,two more strains were identified from a catheter tip and bloodculture during the follow-up period. Thus, a total of 49 strainswere included in the study.

Escherichia coli DH10B was used as the host for cloning experiments.Chloramphenicol-resistant pBC-SK(–) plasmid (Stratagene,La Jolla, CA) was used as the cloning vector. E. coli XL1-BlueRif^r NA^r was used as the recipient for conjugation and transformationexperiments. Bacterial cultures were routinely grown in Luria-Bertani(LB) broth at 37°C.

PFGE. Pulsed-field gel electrophoresis (PFGE) analysis was performedusing restriction enzyme XbaI (New England Biolabs, Ipswich,MA) and a CHEF III DR electrophoresis system (Bio-Rad, Hercules,CA). Relatedness of the strains was determined according tothe criteria of Tenover et al. (9).

Susceptibility testing. MICs of cefotaxime with and without clavulanic acid, ceftazidimewith and without clavulanic acid, cefepime, aztreonam, ertapenem,ciprofloxacin, and gentamicin were determined for all studystrains by use of Etest strips according to the instructionsfrom the manufacturer (AB Biodisk, Solna, Sweden). Productionof ESBL was confirmed by the disk diffusion method endorsedby the Clinical and Laboratory Standards Institute (CLSI) (3).

PCR analysis and sequencing. Primer sets to detect bla_TEM (5), bla_SHV (10), and bla_CTX-M(4) were used for PCR analysis to detect the ESBL genes in thestudy strains. Nucleotide sequencing was performed with an ABI3100genetic analyzer. The full sequences of bla_CTX-M genes wereobtained using a set of primers flanking the genes (Table 1).

View this table:
[in this window]
[in a new window]

TABLE 1. Primers used for the study

PCR cloning of bla_CTX-M-2/59. To assess whether the amino acid substitution observed withCTX-M-59 affected the level of resistance to β-lactams,E. coli clones producing either CTX-M-2 or CTX-M-59 were constructed.The structural genes corresponding to the enzymes were amplifiedwith primers CTXM2F-XbaI and CTXM2R-BamHI (Table 1). The PCRproduct was digested with XbaI and BamHI (New England Biolabs)and ligated with pBC-SK(–). E. coli DH10B was then transformedby these recombinant plasmids by electroporation. Transformantswere selected on LB agar plates containing 50 µg/ml ofampicillin and 30 µg/ml of chloramphenicol. After confirmationof the nucleotide sequences, E. coli DH10B(pCTX-M-2) producingCTX-M-2 and E. coli DH10B(pCTX-M-59) producing CTX-M-59 wereused to determine the MICs of β-lactams.

Conjugation and transformation. The standard broth mating method was employed for conjugationexperiments (10). Transconjugants were selected on LB agar platescontaining 50 µg/ml of ampicillin, 50 µg/ml of rifampin,and 50 µg/ml of nalidixic acid. For the strains that didnot yield transconjugants, the plasmids were extracted by themodified alkaline lysis method described previously (8). E.coli XL1-Blue Rif^r NA^r was transformed with the plasmids byelectroporation. Transformants were selected on LB agar platescontaining 50 µg/ml of ampicillin.

DNA hybridization. Plasmids were extracted from the transconjugants and transformantsby the modified alkaline lysis method (8) and digested withEcoRI, HindIII, or PstI (New England Biolabs). The plasmidswere then subjected to electrophoresis in an 0.8% agarose gel.The plasmids digested with PstI were hybridized with a digoxigenin-labeledDNA probe specific for bla_CTX-M-2/59 using the PCR DIG detectionsystem (Roche Diagnostics, Indianapolis, IN).

Nucleotide sequence accession numbers. The nucleotide sequences of bla_CTX-M-59 and bla_SHV-85 have beendeposited in the GenBank/EMBL/DDBJ database under accessionno. DQ408762 and DQ322460, respectively.

RESULTS AND DISCUSSION

Top

RESULTS AND DISCUSSION

PFGE. PFGE revealed the presence of 10 pulsotypes (Table 2; Fig. 1).Pulsotypes A and J were detected solely during the outbreakand were possibly related to each other, while pulsotype B waspresent both during and after the outbreak. Pulsotypes C andD were commonly recovered during the follow-up period. Takentogether, these five pulsotypes accounted for 82% of the studystrains.

View this table:
[in this window]
[in a new window]

TABLE 2. MICs and ESBL types of K. pneumoniae strains according to PFGE pulsotype^a

View larger version (115K):
[in this window]
[in a new window]

FIG. 1. PFGE profiles of the 10 pulsotypes identified in the study. M, marker.

Susceptibilities of the study strains. All of the study strains were confirmed as ESBL producers bythe disk diffusion method. Strains belonging to all pulsotypesexcept D had median cefotaxime MICs of greater than 256 µg/ml.Strains with pulsotype D had a median ceftazidime MIC of 256µg/ml but a lower median cefotaxime MIC (48 µg/ml).All strains were susceptible to ertapenem and ciprofloxacin.They were variably resistant to gentamicin.

PCR analysis and sequencing of β-lactamase genes. All PFGE pulsotypes except D gave amplicons consistent withbla_CTX-M. The sequences corresponded to CTX-M-2 for pulsotypesA, B, F, H, I, and J and CTX-M-59 for pulsotypes C, E, and G(Table 2). CTX-M-59 is a novel variant of CTX-M-2 with an H89Lsubstitution according to the Ambler numbering scheme (1). PulsotypeD carried bla_SHV, which was confirmed to encode SHV-5 upon sequencingof the full open reading frame. All the other pulsotypes exceptF were also positive for bla_SHV, which encoded various non-ESBLSHV enzymes including SHV-1, -11, -77, and -85. SHV-85 is anovel variant of SHV-11 with an L19M substitution, which islocated within the signal peptide. All pulsotypes except D werepositive for bla_TEM. The deduced amino acid sequences were consistentwith that of TEM-1.

Phenotype conferred by CTX-M-59. The β-lactam MICs of E. coli DH10B(pCTX-M-59) did not differsignificantly from those of E. coli DH10B(pCTX-M-2) (Table 3).Both clones displayed susceptibility patterns typical for organismsproducing CTX-M-type enzymes, with higher levels of resistanceto cefotaxime than to ceftazidime. These results suggested thatthe kinetic properties of CTX-M-59 were not likely altered bythe H89L substitution compared with its progenitor CTX-M-2.

View this table:
[in this window]
[in a new window]

TABLE 3. MICs of E. coli DH10B with recombinant plasmids encoding bla_CTX-M-2/59

Mating experiments. Transconjugants were obtained from representative strains ofall pulsotypes at frequencies of 2 x 10^–5 or higher perrecipient cell except for pulsotypes A and J, for which no transconjugantwas obtained. All transconjugants except for that from pulsotypeD were resistant to cefotaxime but not ceftazidime. PCR analysisfor these transconjugants yielded amplicons with primers specificfor bla_CTX-M-2 and bla_TEM, indicating successful transfer ofbla_CTX-M-2/59 and bla_TEM-1. The transconjugant from pulsotypeD was resistant to ceftazidime but not cefotaxime and was positivefor bla_SHV by PCR analysis, indicating conjugal transfer ofbla_SHV-5. The plasmids harboring bla_CTX-M-2 in representativestrains from profiles A and J were successfully transferredto the recipient strain by transformation, which was confirmedby PCR analysis likewise. The MICs for the transconjugants andtransformants are shown in Table 4.

View this table:
[in this window]
[in a new window]

TABLE 4. MICs and PCR results for the transconjugants/transformants from each PFGE pulsotype

Plasmid analysis and DNA hybridization. Plasmids obtained from the transconjugants and transformantsencoding bla_CTX-M-2/59 shared identical or very similar bandingpatterns with any of the three restriction enzymes, whereasthat from the transconjugant encoding bla_SHV-5 was distinct(results not shown). The sizes of these plasmids were estimatedto be ca. 50 kb. The probe for bla_CTX-M-2/59 hybridized withtwo bands with similar sizes, ca. 1.5 kb and 0.5 kb, for theplasmids from all pulsotypes except D (Fig. 2). Two bands weregenerated because bla_CTX-M-2/59 contains a PstI restrictionsite. This finding suggested that the genetic support of bla_CTX-M-2/59in this outbreak was likely uniform regardless of pulsotypes.Taken together, the outbreak phase was initiated by three genomicclones carrying similar plasmids, all encoding bla_CTX-M-2. Itappears that one of the plasmids then further disseminated toother clones and produced a variant plasmid encoding bla_CTX-M-59.Thus, we speculate that these ESBL-encoding plasmids persisted,evolved, and disseminated among different clones of K. pneumoniaein the confined environment of a NICU over the 10-month period.

View larger version (84K):
[in this window]
[in a new window]

FIG. 2. DNA hybridization with a bla_CTX-M-2-specific probe. bla_CTX-M-2/59 contains a PstI restriction site. All the plasmids hybridized with the probe, except for that from pulsotype D, which is bla_CTX-M-2/59 negative and bla_SHV-5 positive.

Conclusion. We described a multiclonal outbreak of K. pneumoniae producingCTX-M-2 and its variant CTX-M-59 ESBLs which took place in aBrazilian NICU. In addition to the spread of each clonal group,dissemination of identical or related plasmids harboring theCTX-M-type ESBL genes among different clonal groups was responsiblefor the persistence of ESBL-producing K. pneumoniae over time.

ACKNOWLEDGMENTS

We thank Hanna Sidjabat for expert technical assistance.

D.D.O.G. and part of this study were supported by a FogartyInternational Center Global Infectious Diseases Research TrainingProgram grant, National Institutes of Health (D43TW006592; principalinvestigator, Lee H. Harrison). Y.D. is supported by NIH traininggrant T32AI007333. D.L.P. has received prior research fundingfrom Pfizer, Elan, Merck, and AstraZeneca and is supported inpart by NIH grant R01AI070896.

FOOTNOTES

* Corresponding author. Mailing address: Division of Infectious Diseases, University of Pittsburgh Medical Center, Falk Medical Building, Suite 3A, 3601 Fifth Avenue, Pittsburgh, PA 15213. Phone: (412) 648-6401. Fax: (412) 648-6399. E-mail: doiy{at}dom.pitt.edu

Published ahead of print on 17 March 2008.

REFERENCES

Top