Real-time PCR for detection of blaOXA-48 genes from stools

Thierry Naas*, Garance Cotellon, Ayla Ergani and Patrice Nordmann

Service de Bacteriologie-Virologie, INSERM U914: ‘Emerging Resistance to Antibiotics’, LabEx LERMIT, Hopital de Bicetre,94275 Le Kremlin-Bicetre, Assistance Publique-Hopitaux de Paris, Faculte de Medecine Paris-Sud, France

*Corresponding author. Tel: +33-1-45-21-29-86; Fax: +33-1-45-21-63-40; E-mail. thierry.naas@bct.aphp.fr

Received 22 May 2012; returned 10 July 2012; revised 25 July 2012; accepted 27 July 2012

Objectives: Outbreaks of OXA-48-like carbapenemase producers are increasingly reported in many Europeancountries and are often the result of difficulties in detection, especially for isolates with MICs of carbapenemsthat remain in the susceptibility range.

Methods: An in-house real-time quantitative PCR (qPCR) assay using TaqMan chemistry to detect blaOXA-48-like

genes was compared with bacterial culturing on ChromID ESBL and SUPERCARBA media of spiked stool sampleswith several species producing OXA-48 variants.

Results: qPCR amplification using plasmid DNA was linear over 10 log dilutions (r2¼0.998 and slope¼23.14),with an amplification efficiency of 1.10, and the detection limit of the assay was reproducibly estimated at10 plasmid molecules/PCR. No cross-reaction was detected with DNA extracted from several multidrug-resistant bacteria harbouring other b-lactam resistance genes. The blaOXA-48 qPCR assay was capable of detecting10–50 cfu of OXA-48 producers/100 mg of faeces. ChromID ESBL was capable of detecting OXA-48 producers(1×101 to 3×102 cfu/100 mg of faeces), as long as the isolates exhibited a high level of resistance to cepha-losporins due to an associated extended-spectrum b-lactamase. The SUPERCARBA screening medium wascapable of detecting all the OXA-48-like producers (1–3×101 cfu/100 mg of faeces), except those producingOXA-163, a variant lacking carbapenem-hydrolysing activity.

Conclusions: The qPCR is likely to shorten the time for blaOXA-48 detection from 48 to 4 h and will be a valuabletool for outbreak follow-up in order to rapidly isolate colonized patients and assign them to cohorts.

Keywords: carbapenemases, carriers, qPCR

IntroductionOXA-48-like carbapenemases are increasingly identified in manyEuropean countries, such as France, Germany and the UK,through transfer of hospitalized patients from endemic areas,but also from likely autochthonous diffusion.1,2 The main reser-voir of OXA-48-like producers is likely to be countries locatedon the southern to eastern borders of the Mediterranean Sea(Morocco to Turkey) and the Indian subcontinent.1,2 Several out-breaks have been described that have often been the conse-quence of difficulties in detection, especially with isolates withlow MICs of carbapenems that may remain in the susceptibilityrange.1,2

Several OXA-48 variants have been described: OXA-48,OXA-162, OXA-181, OXA-204 and OXA-232 hydrolyse carbape-nems and hydrolyse broad-spectrum cephalosporins veryweakly, while OXA-163 identified from Argentinean isolateslacks carbapenem-hydrolysing activity, but hydrolyses expanded-spectrum cephalosporins.2 The level of resistance to carbapenems

provided by OXA-48-like-carbapenemase producers may vary sig-nificantly, making their detection difficult when based just on highlevels of carbapenem resistance. Therefore, detection of OXA-48producers is particularly challenging in a routine laboratory, espe-cially in the course of managing an outbreak, when rapid identifi-cation and cohorting of patients is required. Several media havebeen proposed for the detection of carbapenemase producers.The ChromID ESBL medium has good sensitivity, its main disad-vantage being its lack of detection of OXA-48-like producers thatare susceptible to cefpodoxime in the absence of co-productionof an extended-spectrum b-lactamase (ESBL).3 The ChromAgarKPC medium has lower sensitivity due to lack of detection of bac-teria with low MICs of carbapenems.3 A Drigalski agar-basedculture medium containing a carbapenem, cloxacillin and zincsulphate (SUPERCARBA medium) has recently been describedto identify accurately all carbapenemase-producing, includingOXA-48-producing, Enterobacteriaceae.4

Recently, the development of real-time quantitative PCR(qPCR) for detecting carbapemenases directly from rectal

# The Author 2012. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.For Permissions, please e-mail: journals.permissions@oup.com

J Antimicrob Chemotherdoi:10.1093/jac/dks340

1 of 4

Journal of Antimicrobial Chemotherapy Advance Access published September 11, 2012 by guest on D

ecember 7, 2015

http://jac.oxfordjournals.org/D

ownloaded from

swabs or stool samples with comparable, if not superior, detec-tion limits compared with culture has shown its usefulnessdue to a faster turnaround time of results (4 versus 48 h).5,6

Here, we describe the development of a qPCR assay combinedwith a fully automated extraction system for the detection ofblaOXA-48-like genes from culture or from spiked stool samples.

Materials and methods

Bacterial strains and plasmidsEleven well-characterized enterobacterial OXA-48-like producers (7OXA-48, 2 OXA-181 and 2 OXA-163 producers), along with 14non-OXA-48-like producers used as negative control strains, were usedfor direct detection in culture or for spiking experiments, in order to val-idate the novel detection assay (see Tables 1 and 2). A recombinantplasmid, pOXA-48, consisting of a pCR-II-TOPO backbone (Invitrogen,Saint Aubin, France) carrying an �1 kb insert expressing the OXA-48 car-bapenemase, was used for qPCR optimization. The concentration of theplasmid solution was 1011 plasmid molecules/mL. MICs were determinedusing the Etest (bioMerieux, Marcy l’Etoile, France) and interpretedaccording to CLSI.7

Serial dilutions and spiking experimentsBacterial suspensions of strains (see Table 2) with an optical density at600 nm of 0.5 were serially diluted in PBS (bioMerieux); nine 10-fold dilu-tions were made. To quantify the viable bacteria in each dilution step, ablood agar plate (bioMerieux) was inoculated with 100 mL of a suspen-sion and incubated overnight at 378C; the number of colonies thatgrew was counted the following day.

Spiked Escherichia coli cultures were prepared by adding 100 mL ofeach dilution in PBS was added to 900 mL of fresh overnight E. coliEc1.6 Spiked faecal samples were made by adding 100 mL of each dilutionin PBS to 900 mL of a faecal suspension that was obtained by suspending6 g of freshly pooled faeces from three healthy volunteers in 60 mL ofdistilled water, as previously described.6 A faecal suspension withoutthe addition of a bacterial strain was used as a negative control. Aliquotsof 200 mL of spiked stool samples were subjected to DNA extraction. Eachfaecal suspension was plated onto the ChromID ESBL and SUPERCARBAmedia. Viable bacteria were counted after 24 h at 378C and growth onselective media was compared with qPCR results. All experiments wereperformed in triplicate.

DNA extraction from bacterial cultures and from stoolsamplesFresh overnight bacterial cultures were used for DNA extraction with aQIAamp DNA Minikit (Qiagen, Les Ulis, France) according to the protocolsuggested by the manufacturer. Extracted bacterial DNA was eluted fromthe columns in 200 mL of elution buffer and stored at 2208C.

Spiked E. coli Ec1 cultures and spiked stool samples were extractedwith the EasyMag automated DNA extraction system (bioMerieux),which enables simultaneous DNA extractions from 24 samples in ,2 h.

OXA-48 qPCR assayThe qPCR Rotor-Gene 6000 amplification/detection system (Qiagen)was used for the amplification and detection of the blaOXA-48 amplicon(100 bp) by TaqMan technology. The forward PCR primer (OXA-48-For; 5′-GTAGCAAAGGAATGGCAA-3′) and the reverse PCR primer(OXA-48-Rev; 5′-CCTTGCTGCTTATTGTCA-3′) specific for the detection ofblaOXA-48-like genes and the blaOXA-48-specific probe [OXA-48-Probe;

5′-FAM-TCC(+A)GA(+G)CA(+C)AA(+C)TACG-TAMRA-3′] [bracketednucleotides with+signs are locked nucleic acid (‘LNA’) nucleotides]were designed in-house. The concentrations of the primers and probeused in this study that gave the best detection limits were 400 nM forblaOXA-48 primers and 200 nM for the blaOXA-48 probe. The QuantitectVirus Kit (Qiagen) was used as recommended by the manufacturer.One PCR experiment (including mix preparation and amplification)takes ,2 h. Simplexa

TM

from Focus Diagnostics (Eurobio, Courtaboeuf,France) was used as an internal control, for extraction and amplification.6

ResultsThe analytical sensitivity of the blaOXA-48 qPCR assay was deter-mined after serially diluting plasmid pOXA-48. The amplificationwas linear over 10 log dilutions (r2¼0.998 and slope¼23.14),and the amplification efficiency was 1.10. The detection limitof the assay was reproducibly estimated at 10 plasmid mole-cules/PCR.

None of the OXA-48-negative bacterial pathogens was posi-tive in the blaOXA-48 qPCR assay (Table 1). Of the 14 isolates nega-tive for the blaOXA-48 gene, 3 were imipenem susceptible and 11were intermediate or resistant, by means of another carbapene-mase gene. Of the 11 isolates positive for the blaOXA-48-like genes,5 were susceptible to imipenem (MIC ,1 mg/L) and 6 were inter-mediate or resistant (MIC .4 mg/L). All were identified using thisassay, with Ct values of 19–20. Overall, there was 100% con-cordance between genotype of bacteria and the blaOXA-48 qPCRassay results (Table 1), suggesting excellent specificities of theprimers and probe. The assay identified OXA-48, OXA-181 andOXA-163 equally well.

The blaOXA-48 qPCR assay on extracted DNA from seriallydiluted OXA-48 producers into E. coli Ec1 and subsequent auto-mated DNA extraction was linear over 9 log dilutions(r2¼0.994 and slope¼22.92), and the amplification efficiencywas 1.09. The limit of the linear range of the qPCR was repro-ducibly 2×101 cfu. The qPCR allowed detection of 1–5 cfu;however, the Ct values were out of the linear range of the qPCR.

The blaOXA-48 qPCR assay was capable of detecting 10–50 cfu/100 mg of faeces. Below 10 cfu (,1 equivalent cfu/PCR), resultswere not reproducible. The analytical sensitivity of the assay didnot change upon the use of different blaOXA-48-like-positive strainswith different carbapenem susceptibility patterns and differentassociated b-lactamases: Citrobacter freundii, Enterobactercloacae, E. coli and Klebsiella pneumoniae blaOXA-48--positivestrains (Table 2).

The qPCR results were compared with those obtained withtwo selective culture media: the chromogenic ChromID ESBL(bioMerieux) and SUPERCARBA. With serially diluted OXA-48 pro-ducers co-producing an ESBL or with OXA-163, which can hydro-lyse expanded-spectrum cephalosporins, the lowest limit ofdetection ranged from 1×101 to 3×102 cfu/mL for ChromIDESBL (Table 2). However, for isolates lacking an ESBL thedetection limit was .105 cfu/mL (Table 2). ChromID ESBL,which contains cefpodoxime as a selector and is routinely usedfor screening ESBL producers,8 was capable of detectingcarbapenemase-producing Enterobacteriaceae of the OXA-48type, as long as a high level of resistance to cephalosporinswas present due to an associated ESBL.3 For SUPERCARBAmedium, the lowest limit of detection of OXA-48 producersranged between 1×101 and 3×101 cfu/mL (Table 2). The

Naas et al.

2 of 4

by guest on Decem

ber 7, 2015http://jac.oxfordjournals.org/

Dow

nloaded from

Table 2. Spiking of faecal samples with OXA-48-like-producing enterobacterial isolates

Isolate Bacteria OXA/NDM ESBLIPM MIC(mg/L)a

Lowest limit of detectionin stool (cfu/mL of

stoola), ChromID ESBL

Lowest limit of detection instool (cfu/mL of stoola),

SUPERCARBA

Lowest limit of detectionof OXA-48 qPCR

(cfu/mL of stoola)

1 K. pneumoniae OXA-163b 0.5 3×102 NG 2×101

2 E. cloacae OXA-48 SHV-5 0.5 2×101 3×101 5×101

3 K. pneumoniae UCL OXA-48 1 NG 1×101 3×101

4 K. pneumoniae HPA-1 OXA-48 CTX-M-15 1.5 1×101 1×101 3×101

5 E. cloacae OXA-163b 0.5 1×102 NG 101

6 E. coli OXA-48 TEM-150 24 2×101 2×101 101

7 E. coli OXA-48 CTX-M-24 24 2×101 2×101 2×101

8 P. rettgeri OXA-48 TEM-101 .32 1×102 1×102 2×102

9 C. freundii OXA-181 .32 NG 2×101 4×101

10 E. cloacae OXA-48 0.5 NG 2×101 101

11 K. pneumoniae OXA-181 0.5 NG 1×101 3×101

16 E. coli A NDM-1 6 2×101 1×101 —c

IPM, imipenem; NG, no growth with 105 cfu/mL of stool.acfu/mL of stool. One millilitre of stool contains 100 mg of stool. Values were calculated using the standard curve made with serially diluted OXA-48-producing E. coli isolates diluted in 108 E. coli WT.bOXA-163, unlike OXA-48 or OXA-181, can hydrolyse expanded-spectrum cephalosporins, but not carbapenems.cNo amplification.

Table 1. Bacterial isolates used in this study

Isolate Bacteria b-Lactamases expressed

MIC (mg/L)

OXA-48 qPCR ReferenceIPM MEM ETP

1 K. pneumoniae OXA-163, TEM-1 0.5 0.25 0.38 + 32 E. cloacae OXA-48, SHV-5 0.5 0.5 3 + 33 K. pneumoniae UCL OXA-48 1 1 4 + 34 K. pneumoniae HPA-1 OXA-48, CTX-M-15, SHV-1 1.5 12 32 + 45 E. cloacae OXA-163, TEM-1 0.5 0.5 2 + 36 E. coli OXA-48, TEM-150 24 32 24 + 37 E. coli OXA-48, CTX-M-24 24 32 24 + 38 Providencia rettgeri OXA-48, TEM-101 .32 .32 .32 + 39 C. freundii OXA-181 .32 .32 .32 + 310 E. cloacae OXA-48 0.5 0.5 2 + 411 K. pneumoniae OXA-181 0.5 0.5 2 + 412 K. pneumoniae A28006 KPC-2, CTX-M-2, TEM-1, SHV-1 16 32 24 2 413 E. coli KPC-2, CTX-M-9, TEM-1 24 16 32 2 314 E. cloacae HGM KPC-2 12 12 12 2 315 K. pneumoniae A NDM-1, CTX-M-15, TEM-1, SHV-1, CMY-2 .32 .32 .32 2 416 E. coli A NDM-1, CTX-M-15, TEM-1 6 4 4 2 417 E. cloacae A NDM-1 8 6 6 2 418 K. pneumoniae VIM-19, TEM-1 8 4 16 2 419 E. coli DIH-1 VIM-19, CTX-M-3, TEM-1 8 4 16 2 420 K. pneumoniae VIM-1, SHV-5 .32 .32 .32 2 421 E. cloacae IMP-4 .32 .32 .32 2 422 E. coli IMP-1, TEM-1 0.5 0.5 3 2 423 K. pneumoniae 6852 IMP-1 1 8 2 2 424 E. coli Ec1 WT 0.06 0.016 0.006 2 625 K. pneumoniae KPS1 OKP-A 0.12 0.12 0.032 2 6

IPM, imipenem; MEM, meropenem; ETP, ertapenem.

qPCR detection of OXA-48 producers in stools

3 of 4

JAC by guest on D

ecember 7, 2015

http://jac.oxfordjournals.org/D

ownloaded from

SUPERCARBA screening medium, which may detect not onlyKPC b-lactamases and metallo-b-lactamases but also OXA-48producers,4 was capable of detecting all the carbapenem-hydrolysing variants of OXA-48-like, even those with low MICs.OXA-163 was not detected on this medium, due to lack ofcarbapenem-hydrolysing activity.

DiscussionThe prevention of spread of carbapenemase producers relies onearly detection of carriers. Screened patients should be kept instrict isolation prior to obtaining the screening results, meaningthat these screening techniques have to be sensitive, specificand rapid. Since the reservoir of carbapenemase producersremains the intestinal flora, stools and rectal swabs are ad-equate samples for performing this screening. ChromID ESBL isa reliable culture medium for screening most carbapenemaseproducers, except many OXA-48 producers, directly from faecalsamples.3 However, this medium is not capable of distinguishingbetween ESBL and carbapenemase producers. The SUPERCARBAmedium represents a significant improvement compared withthe available screening media to detect carbapenemase produ-cers, particularly for the detection of OXA-48 producers that donot co-express an ESBL. In addition, this medium is useful forspecifically selecting carbapenemase producers in stools thatalso contain a large amount of ESBL producers, the growth ofwhich will be inhibited.4 This property is particularly relevant,since high rates of ESBL carriage are now reported worldwide.9

Alternative strategies, such as qPCR, have proven their useful-ness for detecting KPC and for the management of epidemicsituations.8 Hindiyeh et al.5 described the development and veri-fication of a qPCR assay for detection of blaKPC genes directly inperianal swabs. Direct detection of blaKPC by PCR shortened thetime to identification of patients colonized or infected withcarbapenem-resistant organisms and was more sensitive thanculture. Similarly, blaNDM-1 qPCR showed excellent sensitivityand specificity, with detection limits lower than those of theculture techniques, and with a significant gain in time toresults (4 versus 48 h).6 For OXA-48 qPCR, a similar time toresults was observed: 2 h for automated extraction and ,2 hfor qPCR. The OXA-48 qPCR thus yielded results in ,4 h, versus2–3 days using conventional methods. This qPCR assay iscapable of detecting OXA-48, OXA-181 and OXA-163, anOXA-48 variant capable of hydrolysing third-generation cepha-losporins but not carbapenems.2 The newly described variants,OXA-162, OXA-204 and OXA-232, should also be detectedusing this qPCR assay, since the nucleotide changes betweenthe variants are located outside the primer and probe bindingsequences.2 We found good sensitivity and specificity, and excel-lent agreement with culture screening techniques. Routine

screening for carriage of OXA-48 producers should be based onselective culture media, while the molecular techniques forscreening directly from stools should reveal their usefulness inan outbreak situation, for preventing further spread of OXA-48producers.

AcknowledgementsWe thank Gaelle Cuzon for technical support.

FundingThis work was funded by grants from INSERM, by the Ministere de l’Edu-cation Nationale et de la Recherche and by the European Community’sSeventh Framework Programme FP7/2007-2013 under grant agreementno. 241742.

Transparency declarationsNone to declare.

References1 Nordmann P, Naas T, Poirel L. Global spread of carbapenemase-producing Enterobacteriaceae. Emerg Infect Dis 2011; 17: 1791–8.

2 Poirel L, Potron A, Nordmann P. OXA-48-like carbapenemases: thephantom menace. J Antimicrob Chemother 2012; 67: 1597–606.

3 Carrer A, Fortineau N, Nordmann P. Use of ChromID ESBL medium fordetecting carbapenemase-producing Enterobacteriaceae. J Clin Microbiol2010; 48: 1913–4.

4 Nordmann P, Girlich D, Poirel L. Detection of carbapenemase producersin Enterobacteriaceae by use of a novel screening medium. J ClinMicrobiol 2012; 50: 2761–6.

5 Hindiyeh M, Smollen G, Grossman Z et al. Rapid detection of blaKPC

carbapenemase genes by real-time PCR. J Clin Microbiol 2008; 46:2879–83.

6 Naas T, Ergani A, Carrer A et al. Real-time PCR for detection of NDM-1carbapenemase genes from spiked stool samples. Antimicrob AgentsChemother 2011; 55: 4038–43.

7 Clinical and Laboratory Standards Institute. Performance Standardsfor Antimicrobial Susceptibility Testing: Twenty-second InformationalSupplement M100-S22. CLSI, Wayne, PA, USA, 2012.

8 Reglier-Poupet H, Naas T, Carrer A et al. Performance of chromID ESBL,a chromogenic medium for detection of Enterobacteriaceae producingextended-spectrum b-lactamases. J Med Microbiol 2008; 57: 310–5.

9 Rogers BA, Sidjabat HE, Paterson DL. Escherichia coli O25b-ST131: apandemic, multiresistant, community-associated strain. J AntimicrobChemother 2011; 66: 1–14.

Naas et al.

4 of 4

by guest on Decem

ber 7, 2015http://jac.oxfordjournals.org/

Dow

nloaded from