Previous Article | Next Article ![]()
Journal of Clinical Microbiology, November 2004, p. 5327-5331, Vol. 42, No. 11
0095-1137/04/$08.00+0 DOI: 10.1128/JCM.42.11.5327-5331.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Robert Koch Institute, Wernigerode Branch, Wernigerode, Germany
Received 2 March 2004/ Returned for modification 27 May 2004/ Accepted 5 July 2004
|
|
|---|
|
|
|---|
8 mg/liter are defined as resistant (6). Identification of the resistance genotype is complicated by the various numbers of 23S rRNA alleles among the genomes of these bacteria, for example, five copies in Staphylococcus aureus, four in Enterococcus faecalis, and six in Enterococcus faecium (2, 15, 19, 23). In vitro studies showed that one out of six mutated alleles in E. faecium and two out of four mutated alleles in E. faecalis were sufficient to confer linezolid resistance (14, 15). A correlation between the number of mutated alleles and the MIC was described for E. faecalis and E. faecium (14, 15). Homozygous susceptible or homozygous resistant isolates (all alleles mutated) can be detected by molecular tests, such as DNA sequencing of PCR-amplified fragments (14, 15, 25) or a restriction digestion following PCR amplification (15, 32). However, heterozygous linezolid-resistant isolates could be confirmed only by time-consuming and laborious methods (15). But these are exactly the isolates encountered in clinical practice: during linezolid therapy, primarily susceptible strains acquire resistance by stepwise mutation (1, 7, 8, 9) and probably by subsequent recombination (14). For rapid molecular detection and molecular confirmation of such isolates, we chose a 5' nuclease real-time PCR assay with Taqman probes, a method which allows rapid, sensitive, and quantitative detection of single-nucleotide polymorphisms.
Seventy-two strains were included; 10 of them were linezolid resistant (MIC
8 µg/ml). They all emerged during linezolid therapy (one strain from the United States, E. faecium 3819 [8]; five from Austria, E. faecalis 3932 and E. faecium 3935, 3936, 3938, and 3939 [9]; and four from Germany isolated from a single patient, E. faecalis 3696 and 3697 and E. faecium 3695 and 3698 [E. Halle, J. Padberg, S. Rousseau, I. Klare, G. Werner, and W. Witte, Correspondence, Infection 32:182-183, 2004]). The 10 enterococcal strains were partly clonally related or identical (e.g., strains 3935, 3936, 3938, and 3939; strains 3695 and 3698; and strains 3696 and 3698 [data not shown]), but the linezolid MICs for them were different, suggesting a different number of mutated 23S alleles. The 62 linezolid-susceptible isolates (MICs of 0.5 to 2 µg/ml) were clonally diverse (data not shown). DNA extraction and purification were done using standard procedures and commercial kits (QIAGEN, Hilden, Germany), and DNA was quantified by fluorescence labeling (Pico Green kit; Molecular Probes, Leiden, The Netherlands). Classical PCR was performed with DNA beads (Amersham Pharmacia, Freiburg, Germany). Real-time PCR was done with an ABI 7000 using a SYBR Green kit and a Taqman kit (Applied Biosystems, Darmstadt, Germany). The assay was first evaluated with primers 23S_TQF and 23S_TQR and the SYBR Green kit, and a 100 pM concentration of each primer and 1 ng of purified PCR product (later with 2 ng of genomic DNA) amplified in a classical approach with primers 23S_F and 23S_R were then added (Table 1). The specificity of products was confirmed by melting-curve analysis. The assay design was then applied to the Taqman kit system, including two labeled Taqman probes possessing 3' MGB (minor groove binder) VIC-LIZ-TQ-S (detecting a wild-type or susceptible allele) and FAM-LIZ-TQ-R (detecting a mutated or resistant allele) probes (Table 1). Optimization included 5' nuclease assays with various concentrations of primers (1, 10, and 100 pM), genomic DNA (0.066, 0.125, 0.25, 0.5, 1, and 2 ng), and probes (25, 50, 100, and 200 nM). An alternative FAM-LIZ-TQ-R probe was also tested (Table 1). All samples were assayed at least in triplicate.
|
View this table: [in a new window] |
TABLE 1. Primers and probes used in this study
|
TAG) or NheI (G
CTAGC) (mutated nucleotides are underlined) (15, 32). After endonuclease treatment, linezolid-susceptible enterococci still showed a nondigested fragment, and homozygous linezolid-resistant enterococci showed two fragments (however, a 43-bp fragment was not detectable) (Fig. 1). The band pattern in linezolid-resistant enterococci with a heterozygous genotype revealed three fragments (43 bp not detectable). Restriction was done with 5 µl of purified 23S_R-23S_F PCR product, 1x buffer, and 20 U of NheI at 37°C for 0.5 to 2 h. Separation of fragments was done with a 2% agarose gel and with a LabChip 1000 kit. LabChip technology requires a Bioanalyzer 2100 (Agilent Technologies, Waldbronn, Germany), allowing quick (3-min/sample) and easy-to-perform separation of DNA, RNA, or proteins in specialized microglass capillary chips. Each lane possesses two internal standards which are scaled to the external standard running on each chip. Inherent BioSizing software automatically calculates the size and quantity of each fragment in relation to the internal and external standards (Fig. 1 and Table 2).
![]() View larger version (79K): [in a new window] |
FIG. 1. Detection of 23S alleles of linezolid-resistant E. faecium and E. faecalis isolates by NheI digestion of purified pooled PCR fragments and subsequent separation in a LabChip 1000 measured in an Agilent Bioanalyzer 2100. The lowest and highest bands correspond to internal size markers. Lanes: M, external size markers; 1, E. faecium ATCC 19434; 2, E. faecium 3698; 3, E. faecium 3695; 4, E. faecium 3819; 5, E. faecium 3936; 6, E. faecium 3938; 7, E. faecium 3935; 8, E. faecium 3939; 9, E. faecalis V583; 10, E. faecalis 3932; 11, E. faecalis 3697; 12, E. faecalis 3696.
|
|
View this table: [in a new window] |
TABLE 2. Results of NheI-digested 23S ribosomal DNA-pooled PCR products resolved with LabChip 1000 technology in an Agilent Bioanalyzer 2100
|
Taqman PCR.
Homozygous linezolid-susceptible and -resistant isolates, including E. faecium ATCC 19434 (susceptible), E. faecium 3939 (resistant), E. faecalis V583 (susceptible), and E. faecalis 3696 (resistant; see also above), were chosen to establish a 5' nuclease assay. Primer concentrations of 100 pM each and concentrations of 25 nM for the probe FAM-LIZ-TQ-R, detecting the resistant allele, and 100 nM for the probe VIC-LIZ-TQ-S, detecting the susceptible allele, led to the best results (data not shown in detail). An alternative probe detecting the resistant allele did not perform better (data not shown). As a threshold for all experiments, a value of 0.2 could be assigned: all nonspecific signals were beneath this value, and all specific signals were above it. Based on these results, gene dosage experiments were performed as follows. Genomic DNA from homozygous linezolid-susceptible and -resistant test isolates (E. faecium and E. faecalis) was quantified and mixed in a manner simulating DNA from wild-type isolates. All possible ratios of wild-type to mutated alleles were covered (for E. faecalis, 0:4, 1:3, 2:2, 3:1, and 4:0; for E. faecium, 0:6, 1:5, 2:4, 3:3, 4:2, 5:1, and 6:0). Existence of the appropriate alleles was precisely detected by the corresponding probes, which means that a single mutated allele was detected among four in E. faecalis and among six in E. faecium (Fig. 2). Detection with the FAM-labeled probe revealed that the change in the
CT value was smaller the more alleles had mutated (Fig. 2). Similarly, when using the VIC-labeled probe, the change in the
CT value was smaller the less alleles had mutated (data not shown). However, even under optimized test conditions, the exact number of mutated versus wild-type alleles could not be quantitated. Nevertheless, the definite detection of a single mutated allele convinced us to prove the test scheme with wild-type isolates. All 72 isolates were investigated by using the approach described above (data not shown). All 10 linezolid-resistant isolates were precisely identified, allowing differentiation between isolates of homozygous and heterozygous linezolid resistance genotypes based on a signal from one or both of the Taqman probes used (data not shown).
![]() View larger version (26K): [in a new window] |
FIG. 2. Detection of the G or T allele type of 23S ribosomal DNA by real-time PCR using two labeled probes. In vitro gene dosage experiments vary the number of wild-type versus mutated alleles in E. faecium (A) and E. faecalis (B) (see text for details). Only results for the FAM-LIZ-TQ-R probe are shown. For better visibility, only one representative per allele mix is shown. Labels indicate the numbers of mutated alleles relative to the overall number of 23S ribosomal DNA copies per genome (E. faecium, 6; E. faecalis, 4). Delta Rn, difference of fluorescence signals of a given template and the no-template control.
|
A 5' nuclease assay using Taqman probes is a modern, time-saving PCR technique allowing online detection and quantification of amplified DNA (5, 11). Usage of distinctly designed probes, including an MGB motif, enables detection of single DNA nucleotide polymorphisms with great specificity and sensitivity. This is useful, for example, for detecting resistance characters based on single nucleotide polymorphisms such as fluoroquinolone or rifampin resistance (13, 28, 30). Even more sophisticated is the molecular detection of linezolid resistance, in which in Staphylococcus and Enterococcus, four to six gene copies code for 23S rRNA targeted by this antibiotic. Woodford and coworkers (32) described a real-time PCR using LightCycler technology, which is different from Taqman technology (two probes per allele versus one probe per allele; sustaining of probes in LightCycler technology versus degradation of probes during Taqman PCR). Mutated and wild-type alleles were detected by a single probe and distinguished by different melting curves (32). This assay design cannot be applied to other real-time PCR cyclers. We established a Taqman assay using two probes differing by a single nucleotide. Both probes are independently and in combination capable of detecting the susceptible and resistant allele types. The probes did not show any cross-hybridization: there were no nonspecific signals for the opposite allele types (Fig. 2). Our assay detected a single mutated 23S allele among four to six copies in the genomes of E. faecalis and E. faecium in vitro and in the in vivo-generated resistant isolates. This detection allows prediction of future linezolid resistance during therapy even before it is detected phenotypically, since a mutation in a single 23S allele might not in all cases be sufficient to confer linezolid nonsusceptibility (14, 15). The findings for E. faecium 3698 demonstrated that even in clinical practice, a single mutated allele is capable of mediating linezolid resistance (MIC, 8 µg/ml). There is reasonable concern whether such isolates would be unambiguously identified by melting-curve analysis after real-time PCR using the LightCycler (32) or by sequencing pooled PCR products of all 23S alleles (15).
In conclusion, we established a Taqman PCR assay with two labeled probes detecting linezolid-resistant and -susceptible 23S alleles. Our assay design could easily be applied to other genera, like Staphylococcus, but slightly different oligonucleotides would have to be used (e.g., for S. aureus, one nucleotide mismatch for the two probes). With the help of LabChip technology, we were able to address the expected number of mutated alleles, which correlated well with the corresponding linezolid MICs for E. faecalis, E. faecium, and S. aureus (14, 15; T. A. Wichelhaus, S. Besier, V. Brade, and A. Ludwig, abstr. KMP021 from the 55th Annu. Mtg. of the DGHM, Int. J. Med. Microbiol. 293[Suppl. 36]:375-376, 2003) (Table 2). This result illustrates again that modern techniques like multiplex PCR, real-time PCR, and DNA chip technology are appropriate tools to predict and/or confirm corresponding resistance phenotypes (3, 16, 26).
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»