|Home | About | Journals | Submit | Contact Us | Français|
We have developed a time- and cost-efficient one-step closed-tube assay for genotyping of aac(6′)-Ib-cr that is capable of distinguishing between the two genetic aac(6′)-Ib-cr variants. Our genotyping assay uses the combined information of simultaneously acquired high-resolution melting data from an unlabeled probe and the full-length amplicon.
There are several variants of the plasmid-borne gene encoding aminoglycoside acetyltransferase aac(6′)-Ib, and of greatest clinical concern is the aac(6′)-Ib-cr variant, which extends the enzyme targets to include fluoroquinolones in addition to aminoglycosides (2, 4, 6, 7, 9). The aac(6′)-Ib-cr gene has spread rapidly among Enterobacteriaceae, and although only conferring a low-level resistance, it may create an environment facilitating the selection of more highly resistant determinants, especially those harboring one topoisomerase mutation. The aac(6′)-Ib-cr variant is characterized by amino acid changes at codon 102 (Trp→Arg) and codon 179 (Asp→Tyr) (9). We have designed an assay which distinguishes between the three known alleles found in the aac(6′)-Ib gene at codon 102. There are two described aac(6′)-Ib-cr variants which differ from the wild-type aac(6′)-Ib (accession no. AF322577) (hereafter denoted wt) (12) at nucleotide 304 by a T→C change (accession no. EF100892 and EU675686) (hereafter denoted cr-C) (1, 11) or T→A change (accession no. EF636461, EF443074, and EU54272) (hereafter denoted cr-A) (5, 8, 10). Both nucleotide changes result in a Trp→Arg amino acid substitution.
On a high-resolution melting capable real-time PCR platform, single-nucleotide polymorphism (SNP) class 1 mutations (C/T or G/A) are easily detected due to the drastic change in the number of hydrogen bonds, which in short amplicons results in significant changes in melting characteristics. However, due to the subtle nature of SNP class 4 mutations (A/T), for which no changes in number of hydrogen bonds occur, these are notoriously difficult to reliably detect and differentiate from the normal variations of the instrument. We have addressed this problem by combining information from two simultaneous melting events obtained in a single reaction.
A high-resolution melt of a short amplicon containing nucleotide 304 enables easy identification of cr-C, which has a significantly higher melting temperature, but cr-A cannot be distinguished from wt. An unlabeled probe with a perfect match to the wt allele was included, which upon melting allowed discrimination between the wt and the two cr variants, where the latter two have lower melting temperatures due to the mismatch at the polymorphic site. The reaction mixture was changed to contain asymmetric concentrations of primers in order to promote amplification of the DNA strand complementary to the probe. The combined information from the two separate melting events results in an unambiguous genotype call.
The assay was developed using a set of 12 isolates, 3 isolates for each known aac(6′)-Ib genotype: wt, cr-C, cr-A, and wt/cr-C (heterozygote). The genotype was determined by restriction fragment length polymorphism (RFLP) analysis using BtsCI (New England Biolabs) (6) and confirmed by sequencing. Primers and probes were designed in Primer3 (http://frodo.wi.mit.edu/primer3/) using AF322577 as template. Primers were aacIb_HRM_F (5′-AGTCGTACGTTGCTCTTGGA-3′) and aacIb_HRM_R (5′-CCTGGATCGGTTTCTTCTTC-3′), which amplified a 58-bp product. The probe, denoted aacIb_HRM_probe (5′-AGCGGGGACGGATGGTGGGAAGAAG-3′; 25 bp), was phosphorylated at the 3′ end to prevent extension. Amplification was carried out in a 10-μl reaction mixture consisting of 1× LC480 HRM master mix (Roche), 3 mM MgCl2 (Roche), 5% Q-solution (Qiagen), with primer and probe concentrations as described below. Through testing of different primer concentrations (0.01 to 0.1 μM and 0.1 to 0.5 μM, respectively) and probe concentrations (0.5 to 1.0 μM) and degrees of primer asymmetry (1:5 and 1:10), it was determined that optimal visualization of both melting events was achieved with 0.05 μM aacIb_HRM_F, 0.25 μM aacIb_HRM_R (1:5 ratio), and 0.5 μM aacIb_probe. A higher concentration of probe disrupted the amplification and masked the amplicon melting, while a 1:10 ratio of primers led to poor amplification and low amplicon yield. Different dilutions of template DNA (1:50 and 1:100) were also tested, and a 1:100 dilution of the template DNA enhanced the reproducibility of the assay by reducing the effect of varying salt concentrations between different DNA preparations. PCR and subsequent high-resolution melt was carried out on a LightCycler 480 apparatus, with the following temperature profile: 95°C for 10 min; 45 cycles of [95°C for 10 s, 57°C for 20 s, and 72°C for 20 s] 95°C for 1 min; and 40°C for 1 min followed by high-resolution melting from 50°C to 99°C, with 25 acquisitions/°C. The results were analyzed in the Gene Scanning mode of the LightCycler 480 software package.
Two melting events were observed, the first around 65 to 75°C, describing the disassociation of the probe. The second, occurring around 79 to 84°C, showed the melting of the whole amplicon. The derivate melting curves for each genotype are presented in Fig. Fig.1.1. In brief, the probe and amplicon melting points were 72.0°C and 81.9°C for wt, 69.7°C and 82.7°C for cr-C, and 68.8°C and 81.7°C for cr-A. Heterozygote isolates containing both wt and cr alleles showed a distinct shoulder on the amplicon melting peak, which we attributed to the melting of heteroduplex fragments. In the case of the wt/cr-A heterozygote the probe melting event showed two distinct peaks representative of the two individual allele peaks, and the wt/cr-C heterozygote showed a flat-topped peak stretching across the individual allele peaks. All genotypes were easily identified and could be unambiguously called. This assay was also shown to perform successfully on the Corbett 6000 real-time platform (Qiagen).
The assay was validated on a cohort of isolates (n = 211) composed of Klebsiella pneumoniae (n = 92), Escherichia coli (n = 59), Enterobacter cloacae (n = 35), Serratia marcescens (n = 12), Proteus mirabilis (n = 7), Citrobacter freundii (n = 4), Enterobacter aerogenes (n = 1), and Providencia stuartii (n = 1), for which the aac(6′)-Ib status (91 positive and 120 negative) and genotype as determined by RFLP were known (32 wt, 37 cr, and 22 wt/cr) but to which the operator was blinded. Seven samples previously negative in the RFLP assay were positive in the high-resolution melting analysis (HRM) probe assay. This was determined by sequencing to be due to a partial deletion of the gene destroying one of the primer binding sites in the RFLP assay. The results for the remaining 204 isolates were 100% concordant with the previously determined genotype. The assay was subsequently used to test a cohort of isolates with an unknown aac(6′)-Ib genotype (n = 732), composed of Enterobacter cloacae (n = 308), Klebsiella pneumoniae (n = 168), Escherichia coli (n = 144), Enterobacter aerogenes (n = 53), Klebsiella oxytoca (n = 38), Enterobacter sakazakii (n = 4), Enterobacter hormaechei (n = 3), Enterobacter asburiae (n = 2), Enterobacter amnigenus (n = 1), Enterobacter gergoviae (n = 1), Enterobacter spp. (n = 1), Acinetobacter baumannii (n = 4), and Acinetobacter haemolyticus (n = 1). The genotype distribution in this cohort was 84 wt (11.5%), 84 cr-C (11.5%), 10 cr-A (1.4%), 10 wt/cr-C (1.4%), and 7 wt/cr-A (1.0%). The remaining 537 were negative for the aac(6′)-Ib gene. Sequencing traces of heterozygote isolates (wt/cr-C and wt/cr-A) showed double peaks of equal amplitude at nucleotide 304, indicating that the two alleles were present in equal quantities in the isolate. One strain showed peaks similar to that of wt but with 0.5°C-higher melting points. Five isolates displayed amplicon peaks with shoulders indicative of heterozygotes but with uneven probe melting peaks. We believe this can be explained by an asymmetric ratio of the two alleles. Among our isolates, coexistence of wt and cr alleles in an isolate was a surprisingly common phenomenon, an observation which was very recently supported in a French study (3). We hypothesize that these represent isolates where a proportion of the cells have acquired a cr allele, and under sustained antibiotic pressure these cells would quickly dominate the population.
We have developed a one-step closed-tube assay for the detection and genotyping of aac(6′)-Ib-cr, capable of differentiating between the two genetic variants responsible for the aac(6′)-Ib-cr phenotype, which from an epidemiological perspective is of importance. The assay is rapid, at <2 h, and costs less than US$1 per isolate. This assay addresses the need for an efficient detection system for the monitoring of the growing prevalence of clinically significant aac-Ib-cr variants.
We gratefully acknowledge the generosity of Roche Pharmaceuticals during the extended trial period of the LightCycler 480 instrument which enabled this assay development.
Published ahead of print on 14 December 2009.