186

MICROBIAL DRUG RESISTANCEVolume 13, Number 3, 2007© Mary Ann Liebert, Inc.DOI: 10.1089/mdr.2007.726

Extended-Spectrum �-Lactam Resistance Due to AmpCHyperproduction and CMY-2 Coupled with the Loss ofOMPK35 in Malaysian Strains of Escherichia coli and

Klebsiella pneumoniae

SELVI PALASUBRAMANIAM,1 GEETHA SUBRAMANIAM,1 SEKARAN MUNIANDY,2

and NAVARATNAM PARASAKTHI3

ABSTRACT

In this report, we describe the detection of AmpC and CMY-2 �-lactamases with the loss of OmpK35 porinamong seven sporadic strains of ceftazidime-resistant Klebsiella pneumoniae and ceftazidime-resistant Es-cherichia coli. Cefoxitin, which was used as a marker of resistance toward 7-�-methoxy-cephalosporins, ex-hibited high minimum inhibitory concentration (MIC) values ranging between 128 �g/ml and �256 �g/ml in all the strains. The presence of hyperproducing AmpC enzymes was indicated by the positive three-di-mensional test. Isoelectric focusing (IEF) study confirmed the presence of AmpC enzymes in all the strains.The ampC gene was detected by PCR in all the strains and confirmed by DNA sequencing. Large plasmidsin all the strains, ranging from 60 kb to 150 kb in size, most likely encode the ampC gene. Two E. coli strainsout of the seven strains showed positive amplification of the blaCMY-2 gene, an AmpC variant, and was con-firmed by DNA sequence analyses. DNA hybridization confirmed the blaCMY-2 gene to be plasmid-mediatedin both of these strains. However, one of these two strains also mediated a chromosomal CMY gene. All thestrains showed an absence of OmpK35 by sodium dodecyl sulfate–polyacrylamide gel electrophoresis(SDS/PAGE) and was confirmed by western blot analyses using raised polyclonal anti-OmpK35 antiserum.This suggests that, apart from CMY production, absence of OmpK35 porin also contributed to cefoxitin re-sistance resulting in extended-spectrum �-lactam resistance among these isolates.

INTRODUCTION

KLEBSIELLA PNEUMONIAE and Escherichia coli are majornosocomial pathogens. In addition to extended-spectrum

�-lactamase (ESBL) production, transferable resistance to thecephamycins has been reported in these organisms. With thecontinuing use of 7-�–methoxy-cephalosporins and the useof �-lactamase inhibitor combinations, plasmids encodingclass C �-lactamases have been described worldwide,6 mostfrequently in K. pneumoniae and E. coli. Such resistance is commonly due to the production of plasmid-mediatedAmpC enzymes; however, impermeability due to loss of porin expression may also result in increase resistance to

cephamycins. Two major porins, OmpF and OmpC, havebeen described, in E. coli that are also homologous to theOmpK35 and OmpK36 porins, respectively, from K. pneu-moniae. The loss of either of these two porins have been re-lated to antibiotic resistance.14 It is important for laborato-ries to be able to detect for such novel mechanisms ofresistance because they also play a major role in the emer-gence of ESBL-resistant strains of K. pneumoniae and E. coli.In this report, we describe the detection of AmpC and CMY-2 �-lactamases with the loss of OmpK35 porins in three iso-lates of K. pneumoniae and four isolates of E. coli obtainedfrom routine clinical samples of inpatients from UniversityMalaya Medical Centre.

1Department of Medical Microbiology, University of Malaya, 50603 Kuala Lumpur, Malaysia.2Department of Molecular Medicine, University of Malaya, 50603 Kuala Lumpur, Malaysia.3School of Medicine and Health Sciences, Monash University Malaysia, 46150 Petaling Jaya, Malaysia.

MATERIALS AND METHODS

Bacterial strains

A total of seven ESBL-resistant strains of K. pneumoniaeand E. coli isolated from routine clinical samples from April,2001, until September, 2002, from inpatients of UniversityMalaya Medical Centre, were screened for possible AmpC hy-perproduction. All strains were identified using API 20E kit(bioMerieux, France).

Detection of ESBL production

Screening of ESBL production by the double-disc synergytest (DDST) was carried out as described by Jarlier et al.7

Screening test for AmpC hyperproduction

Cefoxitin resistance was screened as described by the National Committee for Clinical Laboratory Standards (NCCLS).13 The presence of hyperproducing AmpC enzymeswas suggeted by the three-dimensional test, as described byVercauteren et al.18 Briefly, a 4-mm cylindrical plug of agarwas removed near the cefoxitin disc (Oxoid, UK) placed 2-mmaway, and the resulting well was filled with 30 �l of test in-oculum. The test inoculum was adjusted to McFarland no. 5turbidity standard (Bio Merieux, France). E. coli ATCC 25922strain was lawned onto the plate prior to overnight incubationat 37°C.

Minimum inhibitory concentration

Minimum inhibitory concentration (MICs) of strains to var-ious antibiotics including cefoxitin (FOX), cefepime (FEP),imipenem, ceftazidime (CAZ), ceftriaxone (CRO), cefotaxime(CTX), and ceftazidime (CAZ) in combination with clavulanicacid (CA) were determined using the agar dilution method de-scribed by the NCCLS.12 E. coli ATCC 25922 and ATCC35218 were used as control strains. All the antibiotics that werein the powder form with known potency were purchased fromSigma, with the exception of cefepime and clavulanate, whichwere generously provided by Bristol-Myers Squibb (Italy) andSmith-Kline and Beecham Ltd. (UK), respectively.

Isoelectric focusing

The test strain was cultured in 50 ml Luria–Bertani (LB)medium with overnight aeration at 37°C. Cells were disruptedby sonication, and the resulting supernatant was analyzed forthe presence of �-lactamases by isoelectric focusing (IEF) onPhast Gels using the PhastSystem (Pharmacia, Sweden).

Plasmid analyses

Plasmid analyses was carried out by the technique of Kadoand Liu8 with modification. E. coli 39R861 containing plas-mids with known sizes was used to estimate the size of the plas-mids present in E. coli and K. pneumoniae.

DNA hybridization

An intragenic biotin-labeled CMY probe, which is a PCRproduct of 647 bp, was hybridized on to a nylon membrane

containing the plasmids derived from the strains. Hybridizedmembrane was reacted with strepavidin alkaline phosphataseconjugate followed by the addition of BCIP/NBT (KPL,USA) substrate to detect the presence of plasmid-mediatedCMY genes. A � DNA/HindIII marker (Promega, USA) andthe PCR product of the the CMY gene were used as positivecontrols.

Purification of plasmid DNA from gel

A total of 200–300 �l (depending on the band intensity) ofplasmid product was separated on a 0.7% agarose gel, afterwhich the plasmid band of interest was excised carefully usinga sterile scapel. The blade of the scapel was flamed with 70%ethanol and cooled before being used. The gel containing theDNA of interest was excised and purified using the QiaquickPlasmid Purification Kit (Qiagen, Germany) as described by themanufacturer. A total of 2 �l of purified plasmid product wasanalyzed on 1% agarose gel to ensure that the product was suc-cessfully purified (without chromosomal contamination), afterwhich it was used as DNA template in the PCR reaction forPCR analyses.

Polymerase chain reaction

Specific primers used to amplify the 396-bp sequence of theampC gene were: forward, 5�-ATTCGTATGCTGGATCT-CGCCACC-3�, and reverse, 5�-CATGACCCAGTTCGCCA-TATCCTG-3�.4 Specific primers were also used to amplify theentire coding region of the CMY gene: forward, 5�-ATGAT-GAAAAAATCGTTATGCTGC-3�, and reverse, 5�-TTAT-TGCAGC TTTTCAAGAATGCGCCA-3�.1 Primers were syn-thesized by Genemed Biotechnologies Inc (USA). All of thePCR products were sized using a 100-bp marker (Promega,USA). One of the isolates, a K. pneumoniae strain, 48551 whichwas confirmed by PCR (using the same primers) and DNA se-quencing, was used as an AmpC-positive control. This controlstrain is an AmpC hyperproducer that produces �-lactamasesof pI � 8.0 that correlates to the pI value of AmpC enzymes.5

DNA sequencing

DNA sequencing of the 396-bp PCR product of the ampCgene from two representative strains, 64615 and 48551 andblaCMY genes from strains 15 and 9462 was carried out usingan automated DNA sequencer, ABI PRISM 377 (Perkin Elmer,USA).

Outer membrane protein extraction

Outer membrane proteins (OMPs) were prepared using themodified method of Matsuyama et al.10 Briefly, 500 ml ofexponential-phase bacterial cells grown in nutrient broth(Amersham, USA) were harvested. The cells were resus-pended in 20 ml of 10 mM phosphate buffer pH 7.2 (Oxoid,UK) and subjected to six rounds of sonications at 30 KHzoutput for 30 sec with intermittent 30-sec cooling periods.The resulting supernatant was treated with 2% Triton X-100(Sigma, USA) before final centrifugation at 12,000 rpm for90 min. The pelleted OMPs were solubilized in Laemmli’ssample buffer at 96°C. Samples were then separated on a12.5% sodium dodecylsulfate–polyacrylamide gel elec-

AMPC AND CMY-2 HYPERPRODUCTION WITH LOSS OF OMPK35 187

trophoresis (SDS-PAGE) gel, after which OMPs were de-tected with Coomassie Blue.

Identification of porins

To detect the specific porins apart from the SDS/PAGEanalyses, identification of porins in low-osmolarity mediumwas carried out as described by Wu et al.19 A single colonyof bacteria was grown in high-osmolarity medium, such asMH broth and low-osmolarity medium, nutrient broth, re-spectively. The OMPs were extracted as described previouslyand separated on the SDS/PAGE gels. The OMP profiles ex-pressed in both media were compared and identified from thegel, and the porins were sized with the K. pneumoniae ATCC13883. This control strain expresses OmpK35 and OmpK36porins as well as OmpA, which is not a porin. In low-osmo-larity medium, both the porins in the control strain are highlyexpressed when compared to their level of expression in high-osmolarity medium. On the basis of differences in osmolar-ity of the mediums in which the OMPs were extracted from,the porins from clinical samples could be identified and sizedusing the control strains.

Western blot analyses

SDS/PAGE gels containing the OMP proteins from all of thestrains and the positive control were transferred onto nitrocel-lulose membrane using the semidry transfer system as describedby Towbin et al.17 with modification. The membrane was in-cubated overnight with specific antiserum diluted 1:1,000 (anti-OMPK35 raised in rabbit). Following overnight incubation, thebound antiserum was reacted with biotin conjugate-labeled anti-rabbit immunoglobulin G (IgG) (1:400,000) (Sigma, USA).Strepavidin alkaline phosphatase conjugate (Sigma, USA) wasadded to the membrane followed by the addition of BCIP/NBT(KPL, USA) substrate to detect the presence of the specificporins.

RESULTS

The presence of AmpC enzymes was confirmed in all ofthe isolates by the positive three-dimensional test and IEF

analyses, which showed that all of the strains had a single pIvalue of �8.0, which correlated to the pI of AmpC �-lacta-mases. The DDST tests were negative for all the strains sug-gesting the absence of ESBL enzymes. Negative ESBL pro-duction was further confirmed by a � four-fold decrease inthe MIC value of ceftazidime in combination with clavulanatewhen compared with the MIC value of ceftazidime alone2

(Table 1). In addition to the DDST and MIC values, whichconfirm the strains to be non-ESBL producers, the suscepti-bility of the strains to cefepime on the basis of the low MICvalue of cefepime also suggests the strains to be non-ESBLproducers. ESBL-producing strains have been associated withintermediate resistance or high-level resistance to cefepime.Cefoxitin, which was used as a marker of resistance toward7-�-methoxy-cephalosporin, exhibited high MIC values rang-ing between 128 �g/ml and �256 �g/ml in all strains (Table1). MIC values of imipenem in all the seven strains rangedbetween 0.5 �g/ml and 1 �g/ml. All of the strains carried theampC gene, which was detected by PCR producing an am-

PALASUBRAMANIAM ET AL.188

TABLE 1. MICS (�G/ML) OF THE STRAINS TOWARD �-LACTAM ANTIBIOTICS

AND CEFTAZIDIME IN COMBINATION WITH CLAVULANATE

Strains FOX CAZ CTX CRO FEP CAZ/CA IPM

K. pneumoniae (48551) �256 �256 �256 �256 4 �256 1K. pneumoniae (64615) �256 �256 �256 64 4 �256 0.5K. pneumoniae (1/A40) 64 256 128 �256 8 128 0.5E. coli (9462) 128 �256 �256 32 2 �256 1E. coli (15) 128 256 64 32 0.25 128 1E. coli (2/B13) 128 256 256 32 0.125 128 1E. coli (1/A8) �256 256 128 256 2 128 0.5

FOX, Cefoxitin; CRO, ceftriaxone; CAZ, ceftazidime; FEP, cefepime; CTX, cefotaxime; IPM, imipenem; CAZ/CA, ceftazidimein combination with clavulanate.

FIG. 1. DNA hybridization of the CMY gene to plasmidpreparations from strains 15 and 9462. Lane 1, � DNA/HindIIImarker; lane 2, strain 15; lane 3, strain 0462; lane 4, CMY PCRproduct; lane A, � DNA/HindIII marker; lane B, strain 15; laneC, strain 9462; lane D, CMY PCR; lane E, E. coli 39R861strain.

plicom size of 396 bp. DNA sequencing analyses of this PCRproduct from both of the representative strains 48551 and64615 showed a 99% homology with the ampC gene of En-terobacter cloacae P99 (accession number X07274). Onlytwo (15 and 9462) out of the seven strains showed positiveamplification of the CMY gene, producing an amplicon sizeof 1,145 bp. The presence of large plasmids in all the strainsranging from 60 kb to 150 kb in size most likely encode theampC gene. Conjugation and hybridization methods were notdone in the remaining five strains (apart from strains 15 and9462, which mediated the CMY-2 gene) because positive am-plification of this gene from purified plasmid (without chro-mosomal contamination) was obtained. This method, as de-scribed, showed that the AmpC gene in these strains wasplasmid borne. Reference strain E. coli 39R861 was used asa plasmid control to size the unknown plasmids from the iso-lates characterized. This control was not used in the hy-bridization process. Figure 1 shows the plasmid profiles ofE. coli 39R861 on an agarose gel. The CMY probe hybridizedto the large plasmids of 150 kb and 100 kb molecular weightin strain 15, indicating that the CMY gene was plasmid en-coded (Fig. 1). However, in strain 9462, the CMY probe notonly hybridized to the large plasmid of 60 kb in size but alsoto the chromosomal DNA, indicating that the CMY gene inthis strain was plasmid and chromosomally mediated (Fig. 1).DNA sequencing of the 1,145-bp amplicon of the blaCMY

gene from strain 15 (accession number DQ364596) and 9462(accession number DQ364597) showed a 100% homologywith the CMY-2 gene of K. pneumoniae (accession numberX91840) and 100% homology with its deduced amino acidsequence. All of the extracts showed �-lactamases with pIvalues �8.0, which correlated to the pI value of AmpC en-zymes.5 The extracted OMPs were separated on a 12.5%SDS-PAGE gel and detected using Coomassie Blue. Porinprofiles from all the strains were compared. K. pneumoniaeATCC 13883, which expressed OmpK35 and OmpK36porins, was used as the positive control (Fig. 2). All of thestrains had the constitutively expressed OmpA protein, whichis not a porin and does not contribute to antibiotic resistancephenotype.20 All of the strains showed the absence ofOmpK35 by SDS-PAGE analyses (Fig. 2) and was confirmedby western blot analyses. As depicted in Fig. 3, representa-tive strains 15 and 9462 showed loss of OmpK35 porin bySDS-PAGE and western blot analyses.

DISCUSSION

It should be noted that AmpC �-lactamases can be encodedon chromosomally or plasmid-mediated genes or both, as wasseen in one of the E. coli isolates encoding a blaCMY-2 gene.This was possible because E. coli could also possess a chro-mosomal AmpC unlike Klebsiella spp.16 This also suggeststhe possibility that ampC genes escaped onto plasmids. Thereare a wide range of plasmid-mediated AmpC variants pro-duced in E. coli and K. pneumoniae to date. This includesACT-1, MIR-1, CMY-2, CMY-11, and FOX-5.3,9,11 In addi-tion to the production of class C �-lactamases, resistance tothe 7-�–methoxy-cephalosporins among enteric pathogensmay also be due to porin loss, as was observed in our study.All of the strains showed the absence of OmpK35 porin bySDS-PAGE analyses, and this absence may be a contributingfactor to their resistance to the ESBLs. Moreover, OmpK35is an important porin for �-lactam penetration.15 This sug-gests that, apart from AmpC hyperproduction, absence ofOmpK35 porin also contributed to cefoxitin resistance re-sulting in ESBL resistance among these isolates. In conclu-sion, resistance to ESBLs in these K. pneumoniae and E. coliisolates is mainly due to the production of plasmid-mediatedAmpC of the CMY-2 enzyme and loss of OmpK35 porin. Theemergence of plasmid-mediated non-ESBL enzymes, apartfrom the prevalent ESBLs, may increase the horizontal spreadof �-lactam resistance among E. coli and K. pneumoniae. Fur-thermore, their emergence may pose a problem for detectionand selection of therapy.

ACKNOWLEDGMENTS

This study was funded by IRPA grants (no. 06-02-03-0017EA 017 and 06-02-03-0109 PR0047/19-06) provided bythe Ministry of Science, Technology and Environment,Malaysia.

REFERENCES

1. Bauernfeind, A., I. Stemplinger, R. Jungwirth, and H. Gia-marellou. 1996. Characterization of the plasmidic �-lactamase

AMPC AND CMY-2 HYPERPRODUCTION WITH LOSS OF OMPK35 189

FIG. 3. SDS-PAGE and western blot analyses of represen-tative E. coli strains 15 and 9462, showing loss of OmpFporin suggestive of cefoxitin resistance. Lane 1, K. pneumo-niae ATCC 13883; lane 2, strain 15; lane 3, strain 9462; lane4, K. pneumoniae ATCC 13883; lane 5, strain 15; lane 6,strain 9462.

FIG. 2. SDS-PAGE of the remaining five E. coli K. and pneu-moniae strains showing loss of OmpF porin suggestive of ce-foxitin resistance. Lane 1, Strain 48551; lane 2, strain 64615;lane 3, 1/A40; lane 4, strain 2/B13; lane 5, 1/A8; and lane 6,K. pneumoniae ATCC 13883.

CMY-2, which is responsible for cephamycin resistance. Antimi-crob. Agents Chemother. 40:221–224.

2. Cormican, M.G., S.A. Marshall, and R.N. Jones. 1996. Detec-tion of extended-spectrum �-lactamase (ESBL)-producing strainsby the E-Test Screen. J. Clin. Microbiol. 34:1880–1884.

3. Coudron, P.E., N.D. Hanson, and M.W. Climo. 2003. Occur-rence of extended-spectrum and AmpC beta-lactamases in blood-stream isolates of Klebsiella pneumoniae: isolates harbor plasmid-mediated FOX-5 and ACT-1 AmpC beta-lactamase. J. Clin.Microbiol. 41:772–777.

4. Galleni, M., F. Lingberg, S. Normark, S. Cole, N. Honore, B.Joris, and J.M. Frere. 1988. Sequence and comparative analysesof three Enterobacter cloacae ampC �-lactamase genes and theirproducts. Biochem. J. 250:753–760.

5. Hanson, N.D., K.S. Thomson, E.S. Moland, C.C. Sanders, G.Berthold, and R.G. Penn. 1999. Molecular characterization of amultiply resistant Klebsiella pneumoniae encoding ESBLs and aplasmid-mediated AmpC. J. Antimicrob. Chemother. 44:377–380.

6. Jacoby, G.A., and P. Han. 1996. Detection of extended-spectrum�-lactamases in clinical isolates of Klebsiella pneumoniae and Es-cherichia coli. J. Clin. Microbiol. 34:908–911.

7. Jarlier, V., M.H. Nicolas, G. Fournier, and A. Philippon. 1988.Extended broad-spectrum beta-lactamases conferring transferable re-sistance to newer beta-lactams in Enterobacteriaceae: hospital preva-lence and susceptibility patterns. Rev. Infect. Dis. 10:867–878.

8. Kado, C.I., and S.T. Liu. 1981. Rapid procedure for detection andisolation of large and small plasmids. J. Bacteriol. 145:1365–1373.

9. Lee, S.H., J.Y. Kim, G.S. Lee, S.H. Cheon, Y.J. An, S.H. Jeong,and K.J. Lee. 2002. Characterization of blaCMY-11, an AmpC-type plasmid-mediated beta-lactamase gene in a Korean clinicalisolate of Escherichia coli. J. Antimicrob. Chemother. 49:269–273.

10. Matsuyama, S., K. Inokuchi, and S. Mizushima. 1984. Promoterexchange between OmpF and OmpC, genes for osmoregulated ma-jor outer membrane proteins of Escherichia coli K 12. J. Bacteriol.158:1041–1047.

11. Mulvey, M.R., E. Bryce, D.A. Boyd, M. Ofner-Agostini, A.M.Land, A.E. Simor, and S. Paton. 2005. Molecular characteriza-tion of cefoxitin-resistant Escherichia coli from Canadian hospi-tals. Antimicrob. Agents Chemother. 49:358–365.

12. National Committee for Clinical Laboratory Standards. 2001.Performance standards for antimicrobial susceptibility testing.NCCLS document, eleventh informational supplement. M100-S11,vol. 21, no.1.

13. National Committee for Clinical Laboratory Standards. 1997.Performance standards for antimicrobial disk susceptibility tests,6th edition. NCCLS, Villanova, PA. Approved Standard M2-A6.

14. Nikaido, H. 1989. Outer membrane barrier as a mechanism of antimicrobial resistance. Antimicrob. Agents Chemother. 33:1831–1836.

15. Nikaido, H., W. Liu, and E.Y. Rosenberg. 1990. Outer mem-brane permeability and beta-lactamase stability of dipolar ioniccephalosporins containing methoxyimino substituents. Antimicrob.Agents Chemother. 34:337–342.

16. Thomson, K.S. 2001. Controversies about extended-spectrum andAmpC �-lactamases. Emerg. Infect. Dis. 7:333–336.

17. Towbin, H., T. Staehelin, and G.R. Stark. 1979. Transfer of pro-tein from gels to diazobenzyl paper and detection with antisera: amethod for studying antibody specificity and antigen structure.Proc. Natl. Acad. Sci. USA 76:4350.

18. Vercauteren, E., P. Descheemaeker, M. Ieven, C.C. Sanders,and H. Goossens. 1997. Comparison of screening methods for de-tection of extended-spectrum �-lactamases and their prevalenceamong blood isolates of Escherichia coli and Klebsiella spp. in aBelgian teaching hospital. J. Clin. Micro. 35:2191–2197.

19. Wu, T.L., L.K. Siu, L.H. Su, T.L. Lauderdale, F.M. Lin, H.S.Leu, T.Y. Lin, and M. Ho. 2001. Outer membrane protein changecombined with co-existing TEM-1 and SHV-1 �-lactamases leadto false identification of ESBL-producing Klebsiella pneumoniae.J. Antimicrob. Chemother. 47:755–761.

20. Yigit, H., A.M. Queenan, G.J. Anderson, A. Domenech-Sanchez, J.W. Biddle, C.D. Steward, S. Alberti, K. Bush, andF.C. Tenover. 2001. Novel carbapenem-hydrolyzing �-lactamase,KPC-1, from a carbapenem-resistant strain of Klebsiella pneumo-niae. Antimicrob. Agents Chemother. 45:1151–1161.

Address reprint requests to:Dr. N. Parasakthi

School of Medicine and Health SciencesMonash University Malaysia

No. 20 & 22, Jalan PJS 11/5Bandar Sunway

46150 Petaling Jaya, Malaysia

E-mail: parasakthi@med.monash.edu.my

PALASUBRAMANIAM ET AL.190