|Home | About | Journals | Submit | Contact Us | Français|
Macrolide treatment failure in syphilis patients is associated with a single point mutation (either A2058G or A2059G) in both copies of the 23S rRNA gene in Treponema pallidum strains. The conventional method for the detection of both point mutations uses nested PCR combined with restriction enzyme digestions, which is laborious and time-consuming. We initially developed a TaqMan-based real-time duplex PCR assay for detection of the A2058G mutation, and upon discovery of the A2059G mutation, we modified the assay into a triplex format to simultaneously detect both mutations. The point mutations detected by the real-time triplex PCR were confirmed by pyrosequencing. A total of 129 specimens PCR positive for T. pallidum that were obtained from an azithromycin resistance surveillance study conducted in the United States were analyzed. Sixty-six (51.2%) of the 129 samples with the A2058G mutation were identified by both real-time PCR assays. Of the remaining 63 samples that were identified as having a macrolide-susceptible genotype by the duplex PCR assay, 17 (27%) were found to contain the A2059G mutation by the triplex PCR. The proportions of macrolide-susceptible versus -resistant genotypes harboring either the A2058G or the A2059G mutation among the T. pallidum strains were 35.6, 51.2, and 13.2%, respectively. None of the T. pallidum strains examined had both point mutations. The TaqMan-based real-time triplex PCR assay offers an alternative to conventional nested PCR and restriction fragment length polymorphism analyses for the rapid detection of both point mutations associated with macrolide resistance in T. pallidum.
Penicillin remains the preferred treatment for syphilis to date, while doxycycline and macrolides (azithromycin and erythromycin) are alternative antibiotics for the treatment of patients allergic to penicillin (1). The convenience of single-dose oral administration of azithromycin, its long half-life in tissue, and its efficacy, which is equivalent to that of penicillin (2, 3), have led to the frequent use of this antibiotic for the treatment of syphilis, particularly outside the United States (4). Macrolides are a diverse class of antibiotics that inhibit bacterial protein synthesis by binding to the 23S rRNA of the 50S ribosomal subunit (5). A macrolide-binding pocket is formed mainly by 23S rRNA domain V nucleotides 2058 and 2059. Alteration of these two key contact sites could cause conformational rearrangements of the binding site to sterically hinder the binding of macrolides (6).
The first point mutation associated with macrolide treatment failure in Treponema pallidum strains was originally reported as an adenine-to-guanine transition (A2058G), cognate to the position in Escherichia coli (7, 8), in both copies of the 23S rRNA gene (9), and the functional macrolide resistance was confirmed in vivo by using a rabbit model infected with an azithromycin- or erythromycin-resistant T. pallidum strain containing the mutation (7). Recently, a second point mutation (A2059G) in the same gene was also found to be associated with macrolide treatment failure in syphilis patients (10). Azithromycin treatment failures have been reported in Canada, Europe, Asia, and the United States (7, 8, 10–15). The percentages of T. pallidum-positive clinical specimens containing the A2058G mutation have increased significantly over time in San Francisco and Seattle, from 4 and 9% in 2002 to 77 and 56% in 2005, respectively (8, 16–18). While the mutation has been shown to be prevalent among T. pallidum strains from primary syphilis cases in Shanghai, China (11), reports from other countries (such as Taiwan, Madagascar, Tanzania, Uganda, and southern Africa) have either not found the mutation or shown it to be relatively uncommon (16, 19–23). However, an increasing prevalence of macrolide-resistant T. pallidum can be expected because of the spread of these resistant strains and the increasing use of macrolides for the treatment of sexually transmitted and other, unrelated, infections.
The conventional method for detection of the A2058G mutation uses a nested PCR followed by MboII restriction enzyme digestion and agarose gel electrophoresis (7). An additional restriction enzyme (BsaI) is needed to detect the A2059G mutation. Therefore, two separate enzymatic digestions, in addition to nested PCRs, are required to identify both mutations, which is both labor-intensive and time-consuming (10). A LightCycler (LC) real-time PCR assay using fluorescence resonance energy transfer (FRET) probes and melting curve analysis has also been developed, but the assay detects only the A2058G point mutation (24). The objectives of our study were to develop a real-time multiplex PCR assay for the rapid screening of a large number of T. pallidum-positive specimens for the presence of both the A2058G and A2059G mutations and to determine the relative prevalence of the mutations among T. pallidum strains in the United States.
DNA samples from 129 genital ulcer specimens that tested positive for T. pallidum by a real-time PCR targeting the polA gene were analyzed (25). These specimens were obtained from patients with primary or secondary syphilis presenting to sexually transmitted disease clinics at 11 sites in the United States during 2007 to 2009, as part of a study to determine the prevalence of the A2058G mutation (26). This study was considered routine disease surveillance and was exempt from CDC institutional review board approval. An additional 107 ulcer specimens from the same study that tested positive for herpes simplex virus but negative for T. pallidum were included as negative controls for the TaqMan-based real-time allelic discrimination assays. DNA was extracted from the ulcer specimens with the QIAamp DNA minikit (Qiagen Inc., Valencia, CA) by following the manufacturer's instructions.
In order to construct DNA with the A2058G mutation and the wild-type (WT) sequence for use as positive controls, both copies of the 23S rRNA gene were PCR amplified from genomic DNA of the T. pallidum Street 14 (contains the A2058G mutation) and Nichols (2058A, hereinafter referred to as the WT) strains with the same forward primer and two different reverse primers as described previously (7). The 1,593-bp amplicons were subjected to nested PCRs to produce a 629-bp fragment (7), with antisense primers that are unique to T. pallidum genes to avoid amplification of the 23S rRNA gene from other bacteria and to amplify both copies of the 23S rRNA gene. These PCR products were subsequently cloned into a TOPO TA cloning vector (pCR2.1-TOPO; Invitrogen, Carlsbad, CA) and transformed into a competent TOP10 E. coli strain. Plasmid DNA from multiple clones was purified with the Qiagen plasmid minikit (Qiagen Inc.), and the 629-bp inserts were sequenced in both directions with the BigDye Terminator v3.1 cycle sequencing kit and an Applied Biosystems 3130xl genetic analyzer (Life Technologies Corp., Carlsbad, CA). The A2059G mutation has been identified only in clinical specimens; thus, a 629-bp insert of the 23S rRNA gene fragment containing this point mutation was custom synthesized and cloned into pUC19 (Integrated DNA Technologies, Coralville, IA). The 629-bp insert from the pUC19 DNA was subsequently subcloned into the pCR2.1-TOPO vector so that this construct and the two described above were all in the same plasmid vector. The DNA sequences of the inserts were confirmed by sequencing as described above.
The A2058G point mutation was detected with a real-time duplex PCR that comprised a forward primer, a biotinylated reverse primer, and two competing TaqMan probes that differed by a single nucleotide (either A or G) at position 2058; the triplex PCR incorporated one additional TaqMan probe to simultaneously detect the A2058G and A2059G mutations. The sequences of the primers and probes used for the real-time PCR are listed in Table 1. Both real-time multiplex PCR assays were designed to amplify a 185-bp region encompassing the mutations at positions 2058 and 2059 in both copies of the 23S rRNA gene.
The duplex PCR amplification was performed with a 10-μl DNA sample in a 25-μl reaction volume containing a final concentration of 1× PCR buffer, 4 mM MgCl2, 400 μM deoxynucleoside triphosphates, 5 U of AmpliTaq Gold polymerase (all from Applied Biosystems), and 200 nM primers and two probes (Probe-WT and Probe-2058). The following thermocycling conditions were used for PCR amplification: initial hold at 95°C for 10 min, followed by 50 cycles of 95°C for 20 s and 64°C for 1 min and a final extension at 64°C for 10 min. The Rotor-Gene Q real-time PCR instrument (Qiagen Inc.) was used to perform all PCR assays. A previously reported LC real-time PCR (24) that allows the discrimination of WT (2058A) and resistant (2058G) strains of T. pallidum with FRET probes and melting curve analysis was used to evaluate the TaqMan-based duplex PCR assay. The FRET assay and melting curve analysis were performed with LC FastStart DNA Master HybProbe mix and the LC 480 Real-Time PCR instrument (Roche Diagnostics Corp., Indianapolis, IN).
The reaction mixture for the triplex PCR was the same as that for the duplex assay, except that it had an additional TaqMan probe (Probe-2059) and the annealing temperature was increased from 64 to 65°C. Fluorescent signals for the triplex assay were acquired after the first five cycles. Both positive (either purified genomic DNA or cloned plasmids) and no-template controls were included to standardize the threshold cutoff in each run. The analytic sensitivity of the real-time triplex PCR assay for allelic discrimination was determined with serially diluted purified plasmid DNA containing the 629-bp insert with different constructs and purified genomic DNA from the T. pallidum Nichols and Street 14 strains. The specificity of the triplex PCR assay was determined with DNA extracted from an in-house specificity panel comprising closely related T. pallidum subsp. endemicum and T. pallidum subsp. pertenue, nonpathogenic treponemes (T. denticola, T. refringens, and T. phagedenis), and pathogenic and commensal genital tract microorganisms, as well as normal skin flora (Table 2).
In order to verify that the real-time multiplex PCR assays were detecting the A2058G and A2059G point mutations, the biotinylated PCR amplicons were captured and the biotin-labeled strand was separated with streptavidin-coated beads. The resulting single-stranded DNA was used as a template for pyrosequencing with a sequencing primer located upstream of A2058G and A2059G. The oligonucleotide used for pyrosequencing is shown in Table 1. Pyrosequencing was performed with PyroMark Q96 reagents and the PyroMark Q96 ID instrument (Qiagen Inc.) in accordance with the manufacturer's instructions. The A2058G and A2059G mutations were analyzed by using the SQA (sequencing) mode of the PSA96MA software (version 2.1).
The LC assay discriminates between the WT (2058A) and azithromycin-resistant genotypes (2058G) on the basis of FRET and melting-curve analyses (24). In contrast, the detection of a single nucleotide polymorphism (SNP) at position 2058 by the real-time duplex PCR was achieved with two TaqMan probes that compete for hybridization to the SNP polymorphic site. The 6-carboxyfluorescein (FAM)-labeled probe is a perfect match to the target containing the WT sequence (2058A), while the CalRed610-labeled probe is a perfect match to DNA sequences containing the 2058G mutation. In this study, the CalRed610-labeled probe hybridized only to the azithromycin-resistant strain containing the A2058G mutation and produced fluorescence signals in the corresponding channel, whereas the FAM-labeled probe hybridized to both the WT and A2058G mutant strains, presumably because of mispairing of the adenine on the WT probe to the SNP guanine on the mutant allele, despite weaker affinity. Figure 1 illustrates the T. pallidum WT strain with positive fluorescence signals only in the FAM channel, while the mutant strain with 2058G exhibited positive fluorescence signals in both the FAM and CalRed610 channels.
The performance of the real-time duplex PCR was evaluated against that of the LC assay with 82 ulcer specimens that were T. pallidum polA positive by PCR and 107 PCR-negative specimens. There was a high level of agreement between the two assays (99.5%, 188/189 specimens; data not shown). Among the ulcer specimens polA positive by PCR, both assays identified 38 with the azithromycin-susceptible genotype (2058A) and 43 with the resistant genotype (2058G), with only one discrepant specimen that was identified as the WT by the LC assay but negative by the duplex PCR assay.
Representative real-time triplex PCR amplification curves with DNA from the T. pallidum WT strain (Nichols) or strains with either the A2058G (Street 14) or the A2059G mutation, and the test results are shown in Fig. 2. The FAM-labeled probe with the WT sequence cross-hybridized to T. pallidum strains with either the A2058G or the A2059G mutant allele, as observed with the duplex PCR assay; however, a weaker fluorescence signal was seen in strains with the A2059G mutation. Conversely, both Probe-2058 and Probe-2059 had higher affinity for the corresponding SNP sequences, allowing the discrimination of individual mutant alleles. In the CalRed610 channel, only DNA from T. pallidum strains with the A2058G mutation exhibited positive fluorescence signals whereas in the Quasar670 channel, positive fluorescence signals were observed only from strains with the A2059G mutation.
A total of 129 specimens PCR positive for T. pallidum were used to compare the performance of the duplex and triplex PCR assays. The results indicate that the real-time duplex PCR assay identified 63 (48.8%) of the 129 strains as having WT genotypes and 66 (51.2%) as having resistant genotypes with the A2058G mutation (Table 3). Those T. pallidum strains with the A2058G mutation were also confirmed by the triplex PCR assay. In contrast, of the 63 strains with macrolide-susceptible genotypes by the duplex PCR assay, 17 (27%) were found to have resistant genotypes containing the A2059G mutation by the triplex PCR assay. The proportions of macrolide-susceptible versus -resistant genotypes harboring either the A2058G or the A2059G mutation among the T. pallidum strains were 35.6, 51.2, and 13.2%, respectively. Of the 107 ulcer specimens that were negative for T. pallidum polA by PCR, none was reactive by either assay. In addition, none of the T. pallidum strains examined had both point mutations.
The use of a biotinylated reverse primer in the real-time triplex PCR assay enabled the verification of mutations at positions 2058 and 2059 of the 23S rRNA gene by pyrosequencing. The sequencing primer and the dispensation order were designed and programmed to readily distinguish the WT from azithromycin-resistant alleles in approximately 10 min. The individual point mutations were confirmed by performing pyrosequencing with the amplicons generated by the real-time PCR. There was a 100% concordance between pyrosequencing and the real-time triplex PCR in identifying the A2058G and A2059G mutations (data not shown). Both pyrosequencing and the real-time triplex PCR results showed consistently that none of the strains examined to date had both point mutations.
The limit of detection of the triplex PCR assay was 1 to 10 genomic copies per reaction mixture with serial dilutions of purified genomic DNA or 629-bp 23S rRNA gene fragments that were cloned into the plasmid pCR2.1-TOPO (data not shown). The analytical sensitivity and PCR efficiency (93 to 104%) were similar regardless of whether a plasmid with different 629-bp constructs or genomic DNA was used as the template. The triplex PCR assay did not amplify DNA from nonpathogenic treponemes (T. denticola, T. refringens, and T. phagedenis), other pathogenic or commensal genital tract microorganisms, or normal skin flora. However, the assay identified the closely related T. pallidum subsp. pertenue strains (CDC-1 and CDC-2) and T. pallidum subsp. endemicum strains (Bosnia A and Iraq B) as having azithromycin-susceptible genotypes because of the hybridization to the FAM-labeled WT probe alone (data not shown).
Azithromycin treatment failures and associated resistance in T. pallidum have been documented in different parts of the world; however, those reports are based solely on the presence of the A2058G mutation in T. pallidum strains, the only intrinsic macrolide resistance marker identified at the time. A recent report from the Czech Republic described a novel A2059G mutation that was associated with spiramycin treatment failure in a case of secondary syphilis (10). We initially developed a TaqMan-based real-time duplex PCR assay to specifically detect the A2058G point mutation in the 23S rRNA gene of T. pallidum. The appearance of the A2059G point mutation prompted us to expand the existing allelic discrimination assay into a triplex PCR format to simultaneously detect both point mutations.
Using the real-time triplex PCR assay, we found proportions of the WT sequence and the A2058G and A2059G mutations of 35.6, 51.2, and 13.2%, respectively, among 129 T. pallidum strains from 11 sites within the United States. With existing assays that detect only the A2058G mutation, 13.2% of the T. pallidum strains with the A2059G mutation would have been misidentified as being macrolide susceptible. Although the sample size in this study was small, T. pallidum strains that carry either the A2058G or the A2059G mutation were present at many sites in the United States (data not shown). We first reported the detection of the A2059G mutation among T. pallidum strains in the United States at the 19th International Society for Sexually Transmitted Disease Research meeting in Quebec, Canada, in 2011. Recently, Grimes et al. (27) showed that 10% of the T. pallidum strains obtained from men who have sex with men in Seattle, WA, during the past decade harbored the A2059G mutation. Although this mutation was initially associated with spiramycin treatment failure (10), the clinical relevance of this mutation and its association with azithromycin treatment failures in the United States are unclear.
The conventional method of using a nested PCR combined with restriction enzyme digestions can identify both mutations; however, this method is time-consuming and not suitable for the screening of a large number of samples. Also, separate TaqMan-based real-time duplex PCR or LC assays can be designed to specifically detect either the A2059G or the A2058G mutation; however, because both mutations are becoming more common, it was feasible to develop a triplex PCR assay that can rapidly detect and differentiate between the two mutations. The performance of the triplex PCR assay in identifying the A2058G mutation was identical to that of the duplex assay, and because the duplex assay does not detect A2059G, we confirmed the presence of this mutation by pyrosequencing. Although a pyrosequencing method can be developed to detect both mutations, this approach requires a specialized pyrosequencing instrument in addition to a PCR thermocycler, which might not be available in most laboratories. Because the pyrosequencing results were concordant with the real-time triplex PCR, it is no longer necessary to incorporate a biotin-labeled primer in the real-time triplex assay.
The use of macrolides to treat patients with early syphilis and their sexual contacts can pose a public health challenge if close monitoring of drug resistance is not performed. Azithromycin should be used with caution for syphilis treatment, only when treatment with penicillin and doxycycline is not feasible, and clinicians and public health practitioners should remain vigilant for treatment failures (1).
The proportion of T. pallidum strains harboring point mutations in the 23S rRNA gene is likely to be higher than previous estimates that were based on the detection of the A2058G mutation alone. The real-time triplex PCR assay described here is a fast, highly sensitive, and specific test that can serve as a screening tool for macrolide susceptibility among T. pallidum strains. Given the potential for undetected A2059G mutations among T. pallidum isolates in the United States, the use of this method to more extensively study the prevalence of the A2058G and A2059G mutations and their association with macrolide treatment failure is warranted. The potential utility of the triplex PCR assay in identifying azithromycin resistance markers and/or treatment failure among the nonvenereal treponematoses, especially yaws, is timely considering the WHO's campaign to use mass treatment with azithromycin to eradicate the disease worldwide.
We thank Michael Grabenstein, John Beltrami, Donna Helms, Deidra D. Parrish, Hillard S. Weinstock (CDC, Atlanta, GA), Edward Hook (University of Alabama, Birmingham, AL), Khalil G. Ghanem (Johns Hopkins University School of Medicine, Baltimore, MD), William Wong (Chicago Department of Public Health, Chicago, IL), Dawn Jackson (Detroit Department of Health and Wellness Promotion, Detroit, MI), Lewis D. Smith (New Mexico Department of Health, Albuquerque, NM), Elaine Pierce (San Diego Department of Health Services, San Diego, CA), Susan Philip, Jeff Klausner (San Francisco Department of Public Health, San Francisco, CA), Steven Wilson (Dallas County Health and Human Services, Dallas, TX), Matthew Golden (University of Washington, Seattle, WA), Kimberly Workowski (Emory University/CDC, Atlanta, GA), Kerry Kenney (Arizona Department of Health Services, Phoenix, AZ), Emily Erbelding (Johns Hopkins Bayview Medical Center, Baltimore, MD), Sara Valway (New Mexico Department of Health, Albuquerque, NM), and James Lee (Texas Department of State Health Services, Austin, TX).
The findings and conclusions in this report are ours and do not necessarily represent the views of the Centers for Disease Control and Prevention. We have no conflicts of interests to declare.
Published ahead of print 2 January 2013