Rapid Detection and Identification of Metallo-{beta}-Lactamase-Encoding Genes by Multiplex Real-Time PCR Assay and Melt Curve Analysis

Metallo-ß-lactamase enzymes (MßL) are encodedby transferable genes, which appear to spread rapidly amonggram-negative bacteria. The objective of this study was to developa multiplex real-time PCR assay followed by a melt curve stepfor rapid detection and identification of genes encoding MßL-typeenzymes based on the amplicon melting peak. The reference sequencesof all genes encoding IMP and VIM types, SPM-1, GIM-1, and SIM-1were downloaded from GenBank, and primers were designed to obtainamplicons showing different sizes and melting peak temperatures(T_m). The real-time PCR assay was able to detect all MßL-harboringclinical isolates, and the T_m-assigned genotypes were 100% coincidentwith previous sequencing results. This assay could be suitablefor identification of MßL-producing gram-negativebacteria by molecular diagnostic laboratories.

INTRODUCTION

Top

Introduction

Since the first report of acquired metallo-ß-lactamase(MßL) in Japan in 1994 (15), genes encoding IMP- andVIM-type enzymes have spread rapidly among Pseudomonas spp.(1, 5, 10, 13, 14, 16, 18, 22-24), Acinetobacter spp. (3, 4,17, 21, 29), and strains of Enterobacteriaceae (6, 8, 11, 12,20, 28). Moreover, new MßL types have been described,such as SPM (25), GIM (2), and, more recently, SIM (9).

The prevalence of MßL-producing gram-negative bacillihas increased in some hospitals, particularly among clinicalisolates of Pseudomonas aeruginosa and Acinetobacter spp. (21,23, 27). Since MßL production may confer phenotypicresistance to virtually all clinically available ß-lactams,the continued spread of MßL is a major clinical concern(26). The aim of this study was to develop a multiplex real-timePCR assay followed by a melt curve step for rapid detectionand identification of genes encoding the MßL-typeenzymes so far described. The MßL type identificationwas based on the characteristic amplicon melting peak.

MATERIALS AND METHODS

Top

Materials and Methods

MßL-harboring clinical isolates and MßL-negative control strains. The strains used in this study are listed in Tables 1 and 2.The MßL genotypes of the clinical isolates of gram-negativenonfermentative and fermentative bacteria harboring MßLwere previously characterized by PCR and sequencing. When applicable,these clinical isolates were also previously molecularly typedto ensure that genetically unrelated strains were used. Additionally,several American Type Culture Collection (ATCC; Manassas, VA)reference strains and laboratory strains were used as MßL-negativecontrols (Table 2).

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

TABLE 1. MßL-harboring clinical isolates used during the validation process

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

TABLE 2. ATCC reference and laboratory strains of gram-negative bacteria used during the validation process as MßL-negative control strains^a

DNA preparation. The microorganisms were grown on blood agar plates overnightat 37°C to ensure colony purity. Three or four bacterialcolonies were taken from the blood agar plates and suspendedin 200 µl of DNase/RNase-free distilled water (Invitrogen,CA). Two microliters of this suspension was used as templatesfor further amplification.

Primer design. The currently available reference sequences of the MßL-encodingIMP- and VIM-type (http://www.lahey.org/studies/), SPM-1 (AJ492820),GIM-1 (AJ620678), and SIM-1 (AY887066) genes were downloadedfrom GenBank (National Center for Biotechnology Information,National Institutes of Health, Bethesda, MD). Based on the comprehensiveanalyses and alignments of each MßL type, primerswere designed to yield amplicons showing different sizes andmelting peak temperatures (T_m) separated by at least 1°C.Predicted amplicon sizes and T_m were determined by the Lasergenesoftware package (DNASTAR, Madison, WI).

Additionally, a primer pair targeting the consensus region ofthe bacterial 16S rRNA gene was included in the reaction mixtureas a PCR internal-control target. Primer pairs were evaluatedin a single format (using a primer concentration of 0.5 µM)to ensure that they correctly amplified their respective lociand that the amplicons showed the expected T_m. Subsequently,the multiplex format was optimized by assaying different primerpair concentrations. The primer sequences, positions, and concentrations,and the sizes of the corresponding amplicons, are given in Table3.

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

TABLE 3. Primers used in this study

Multiplex real-time PCR. Amplification was performed in a 48-µl mixture containing25 µl of Platinum SYBR Green qPCR SuperMix (Platinum TaqDNA polymerase, SYBR Green I dye, Tris-HCl, KCl, 6 mM MgCl₂,400 µM dGTP, 400 µM dATP, 400 µM dCTP, 800µM dUTP, uracil DNA glycosylase, and stabilizers) (Invitrogen,CA), six pairs of primers at their respective concentrations(Table 3), and 2 µl of the template by using the DNA EngineOpticon 2 system (Bio-Rad Laboratories, CA). The PCR conditionswere as follows: initial denaturation at 94°C for 5 min;35 cycles of 94°C for 20 s, 53°C for 45 s, and 72°Cfor 30 s; and a melt curve step (from 68°C, gradually increasing0.5°C/s to 95°C, with acquisition data every 1 s). Meltcurves were then converted into melting peaks by plotting thenegative derivative of fluorescence versus temperature (–dF2/dTversus T and –dF3/dT versus T).

Multiplex real-time PCR validation. In order to assess the accuracy of the assay, 44 bacterial strainswere blindly tested after real-time PCR optimization (Tables1 and 2).

Multiplex real-time PCR sensitivity. The sensitivity of the reaction was estimated by dilution experiments.Briefly, one representative of each MßL-harboringclinical isolate was suspended in DNase/RNase-free distilledwater to a density corresponding to a McFarland turbidity standardof 1.0 (3 x 10⁸ CFU/ml). These suspensions were used to prepareserial 10-fold dilutions using DNase/RNase-free distilled water.

RESULTS AND DISCUSSION

Top

Results and Discussion

When different strains were submitted to the real-time PCR assay,differences in the T_m of the amplicons were observed for strainsharboring bla_IMP-type allelic variants (from 76.0°C to 77.5°C)as well as for those harboring bla_VIM-type allelic variants(from 87.5°C to 88.5°C) (Table 1). These differencesin T_m will be observed mainly for amplicons generated from bla_IMP-typegenes, since the GC contents of the amplicons generated willbe more divergent than those for bla_VIM-type genes (Table 1).

Allelic variants for the remaining MßL types (bla_SPM-1,bla_GIM-1, and bla_SIM-1) have not been found yet. For this reason,only one clinical isolate harboring bla_GIM-1 and one harboringbla_SIM-1 were used during the validation process. The theoreticaland practical T_m obtained were very similar, and no T_m differenceswere observed when several genetically unrelated bla_SPM-1-harboringP. aeruginosa isolates were submitted to the assay (Table 1).

When the negative-control ATCC reference strains and laboratorystrains were submitted to the assay, the melt curve analysisshowed only one melting peak varying from 85.5°C to 86.5°C(Table 2). These melting peaks were consistent with the T_m ofthe amplicon generated by the primers targeting the conservedsequences of the 16S rRNA gene. This internal-control primerpair was used at a lower concentration than the primers targetingthe MßL genes; thus, the latter would have preferenceduring the amplification reaction. This strategy was employedto avoid double amplification, which could compromise the meltcurve analysis.

The real-time PCR sensitivity experiment showed that the assaywas capable of detecting the16S rRNA target gene at a dilutioncorresponding to 6 x 10³ CFU per reaction; bla_SPM, bla_VIM, bla_SIM,and bla_IMP at 6 x 10² CFU per reaction; and bla_GIM at 6 x 10¹CFU per reaction (data not shown). Additionally, the lowestdetection limits of the target genes were represented by thecycle threshold values of 34.62, 34.11, 33.66, 32.69, 32.58,and 28.86 for the16S rRNA, bla_SPM, bla_GIM, bla_IMP, bla_SIM, andbla_VIM genes, respectively. This suggests that the assay asdeveloped is sufficiently robust, even when the bacterial cellssuspended in water are used as the template.

Although the assay was developed to detect all MßL-encodinggenes, we could not submit strains harboring all the bla_IMPand bla_VIM allelic variants, since we do not have access toall of them. We also acknowledge the possibility of future assaylimitations once more MßL types or newly emergingMßL allelic variants are detected, requiring a possibleassay reconfiguration.

The assay was able to detect and identify all MßL-harboringstrains evaluated. It is a single-tube reaction, technicallysimple, performed in only 2 h after colony selection. The T_m-assignedMßL genotypes are easily interpreted (Fig. 1a) andmay be suitable for the detection of MßL-producinggram-negative bacteria by molecular diagnostic laboratories.Furthermore, the assay may also be performed through a conventionalamplification reaction, followed by visualization of the ampliconsby using a UV light box after electrophoresis on a 1.5% agarosegel containing 0.5 µg/ml ethidium bromide (Fig. 1b).

View larger version (59K):
[in this window]
[in a new window]

FIG. 1. (a) Characteristic melting peaks (colored lines) of the amplicons generated by primers targeting the five MßL types so far described when MßL-harboring clinical isolates were submitted to the real-time PCR assay. Colors and genes targeted, from left to right, are as follows: blue, bla_GIM-1 (T_m, 72.0°C); red, bla_IMP-type genes (T_m, 76.5°C); green, bla_SIM-1 (T_m, 80.5°C); pink, bla_SPM-1 (T_m, 83.5°C); orange, bla_VIM-type genes (T_m, 89.0°C). (b) Amplicons generated by primers targeting the five MßL types and the internal-control gene (16S rRNA). Visualization was performed in a UV light box after electrophoresis on a 1.5% agarose gel containing 0.5 µg/ml ethidium bromide. Lane 1, SPM-1 amplicon (798 bp; strain 48-1997A); lane 2, SIM-1 amplicon (569 bp; strain 03-9-T104); lane 3, VIM-type amplicon (382-bp; strain 7-406); lane 4, IMP-type amplicon (188 bp; strain 48-696D); lane 5, GIM-1 amplicon (72 bp; strain 73-5671); lanes 6 to 9, the internal-control amplicon (1,499 bp; strains A. calcoaceticus ATCC 33305, P. aeruginosa ATCC 27853, Klebsiella pneumoniae ATCC 700603, and Enterobacter aerogenes ATCC 13048, respectively); lane 10, negative control; lanes M, molecular size markers (50-bp DNA ladder; Invitrogen).

The rapid detection of MßL-producing isolates couldbe helpful for epidemiological purposes and for monitoring theemergence of MßL-producing isolates in clinical settings.The detection of such isolates could help rapidly establishstandards for hospital infection control measures to minimizethe spreading of these resistant determinants.

ACKNOWLEDGMENTS

We thank Timothy R. Walsh, Mark A. Toleman, Yoshichika Arakawa,and Yunsop Chong for providing some of the MßL-harboringclinical isolates included in this study.

FOOTNOTES

* Corresponding author. Mailing address: Division of Infectious Diseases, Federal University of São Paulo, Rua Leandro Dupret, 188, São Paulo, SP, Brazil CEP 04025-010. Phone: (55-11) 5081-2819 or -2965. Fax: (55-11) 5571-5180. E-mail: rodrigo.mendes{at}lemc.com.br.

Published ahead of print on 8 November 2006.

REFERENCES

Top