Separation of Mycobacterium abscessus into subspecies or genotype level by direct application of peptide nucleic acid multi-probe- real-time PCR method into sputa samples

Background Recently, we introduced a novel peptide nucleic acid (PNA) multi-probe real time PCR method targeting the hsp65 gene (hsp65 PNA RT-PCR) to distinguish Mycobacterium abscessus groups. Methods Here, we evaluated the usefulness of the hsp65 PNA RT-PCR for the direct identification of the M. abscessus group at the subspecies and genotype levels from sputa samples. The method was applied to total sputa DNA from 60 different patients who were identified as having mycobacterial infections via rpoB PCR restriction analysis of the same cultures. Results The hsp65 PNA RT-PCR method had higher sensitivity than the multi-probe real-time PCR assay targeting hsp65 (HMPRT-PCR) for the detection of M. abscessus from sputum [96.7 % (29/30 samples) vs. 70 % (21/30 samples); 100 % specificity]. Conclusions These results suggest that the PNA-based method is feasible for the detection of M. abscessus members not only from cultures but also directly from sputa. Electronic supplementary material The online version of this article (doi:10.1186/s12879-015-1076-8) contains supplementary material, which is available to authorized users.

The taxonomic status of the M. abscessus group remains undetermined. Recent advances in taxonomic approaches revealed that the M. abscessus group could be further divided into three closely related taxa [i.e., M. abscessus subsp. abscessus (hereafter referred to as M. abscessus), M. abscessus subsp. massiliense (hereafter referred to as M. massiliense) and M. abscessus subsp. bolletii (hereafter referred to as M. bolletii)] [6][7][8]. Recently, it was reported that M. massiliense could be further subdivided into two genotypes (hereafter referred to as Type I and Type II) based on hsp65 sequence analysis. All of the strains belonging to Type II are distinct in Korean patients due to a rough colony morphotype without any exceptions [9]. A recent complete genome study revealed that the rough colony phenotype of the Type II strains may be due to a large deletion event in the glycopeptidolipid (GPL)-related genes [10,11]. Furthermore, a recent report based on pulsed-field gel electrophoresis (PFGE) and multi-locus sequence typing (MLST) also revealed taxonomic separation between M. massiliense Type II-1 and Type II-2 [12].
M. massiliense has been increasingly recognized as an emerging pathogen that causes postsurgical wound infection outbreaks [13]. Recently, it was identified as the causative agent of respiratory outbreaks in two cystic fibrosis centers and showed evidence of patient-to-patient transmission [14,15]. Infections with the M. abscessus group are difficult to treat, because these mycobacteria exhibit intrinsic resistance to various antibiotics. Susceptibility to clarithromycin (CLA) varies among members of the M. abscessus group; for example, while the majority of M. abscessus and M. bolletii exhibit resistance to CLA, M. massiliense is susceptible. Resistance is characterized by the presence of the erm(41) gene that encodes the erythromycin ribosomal methylases (ERM) [16].
Despite their close genetic relatedness, disparities in pathogenic potential, transmission mode, and antibiotic susceptibility were found among members of the M. abscessus group. Therefore, accurate identification of members of the M. abscessus group is important for patient treatment and epidemiological purposes [17].
Peptide nucleic acids (PNAs) are artificially synthesized DNA analogues with an uncharged peptide backbone. PNAs have more favorable hybridization properties and chemical, thermal, and biological stability due to their uncharged nature and peptide bond-linked backbone [18]. Due to these favorable characteristics, PNA has been widely applied as a diagnostic tool in molecular biology [19]. A PNA probe-based real-time PCR assay has been developed for mycobacteria diagnosis, particularly for the simultaneous separation of M. tuberculosis and NTM in clinical specimens [20,21]. Recently, we introduced a novel peptide nucleic acid (PNA) multi-probe real-time PCR method that targeted the hsp65 gene (hsp65 PNA RT-PCR) to distinguish between the four types within the M. abscessus groups [M. abscessus and the 3 M. massiliense types (Type I, Type II-1 and Type II-2)] using 3 PNA probes. For this evaluation, we applied 27 reference strains and 228 clinical isolates belonging to the M. abscessus groups. With the exception of one clinical isolate, most of the samples (227/ 228 isolates, 99.6 % sensitivity) were clearly separated at the subspecies or genotype levels, thereby demonstrating the technique's feasibility for the detection of M. abscessus at the mycobacterial culture level. To the best of our knowledge, this was the first report to use a PNA-based multiprobe approach for bacterial diagnosis [22].
The aim of this study was to evaluate the usefulness of the hsp65 PNA RT-PCR in directly identifying the M. abscessus group at the subspecies or genotype levels from sputum samples. We applied the technique to sputum DNA samples from 60 different patients who were previously diagnosed with mycobacterial infections via rpoB PCR restriction analysis of the same cultures (30 samples were culture positive for M. abscessus complex strains, and 30 samples were culture positive for other mycobacteria). These results were compared with the results from the multi-probe real-time PCR assay targeting hsp65 (HMPRT-PCR) that was previously developed by our laboratory [23]. The sputa were digested, decontaminated, and concentrated as recommended by the WHO [24]. The processed sediment was stained using the Ziehl-Neelsen method. The results of the AFB smears were graded according to the recommendations of the American Thoracic Society and the Center for Disease Control and Prevention [25]. Sputa with trace AFB smear results (1-2 bacilli in 300 fields) were also included in this study. The protocol for this study including the documentation for waiver of informed consent was approved by the institutional review board of Seoul National University Hospital (C-1503-058-655) and Asan Medical Center (AMC IRB 2007-0331).

DNA extraction
Chromosomal DNA was extracted from the sputum samples using the bead beater-phenol extraction method as previously described [26].

hsp65 PNA RT-PCR
For the detection of M. abscessus from sputum samples, the hsp65 PNA RT-PCR method was applied to 60 sputum DNA samples, as previously described [22]. Briefly, a total of three reporter dyes were used for the specific simultaneous detection of the four types of the M. abscessus group (M. abscessus, Type I, Type II-1, and Type II-2) in a single reaction.: FAM for the detection of M. abscessus and M. massiliense at the species level, Hex for the discrimination of M. massiliense Type I and Type II, and Texas Red for the discrimination of M. massiliense Type II-1 and Type II-2. The sequences of the probes and primers are provided in Table 1. The probes were purchased from Panagene (Daejeon, Korea) and the primers from Macrogen (Seoul, Korea). A LightCycler (version 96; Roche Life Science, Mannheim, Germany) system was used for the real-time PCR, and three channels were used for the experiment. The optimal reaction mixture was established for the sensitive and specific detection of target sequences. A 10 μl reaction mixture was prepared for each sample as follows: 1 μl PCR reaction buffer for FastStart Taq  The cycling conditions were 300 s at 95°C and 45 cycles of 20 s at 95°C, 15 s at 62°C (single acquisition of fluorescence signals), and 40 s at 74°C. Melting curve analysis was performed using the following cycles: 120 s at 95°C and 180 s at 37°C with a ramping speed of 1.1°C/s. Then, the temperature was increased from 37°C to 80°C at a temperature transition rate of 0.07°C/s during which time the fluorescence signal was continuously acquired. Duplicate experiments were performed to determine the melting temperatures of the probes designed for the target M. abscessus group by real-time PCR, and DNA from a total of 60 sputum samples (Additional file 1) was subsequently tested for species identification by T m analysis.

HMPRT-PCR
The HMPRT-PCR was applied to the same 60 sputum DNA samples as previously described for comparison with the hsp65 PNA RT-PCR [23]. Briefly, a total of 4 channels (CH.) were used for probes specific for 7 mycobacteria spe-  M. massiliense strains. The assay was also able to differentiate all 11 M. massiliense strains into three genotypes: Type I (5 sputa), Type II-1 (4 sputa) and Type II-2 (2 sputa) ( Fig. 1a-c, Table 2 and Additional file 1).
To verify the consistency of our hsp65 PNA RT-PCR method in a clinical setting, we analyzed the range of T m s of the clinical samples at the intra-subspecies or intra-genotypic level in each channel. In the FAM channel, differences in T m s among the 18 M. abscessus sputum samples and 11 M. massiliense strains were 2.9°C (49.0-52.3°C) and 0.7°C (61.4-62.1°C), respectively, which was less than the inter-subspecies T m s differences between M. abscessus and M. massiliense (11.5-11.9°C). In the Hex channel, differences in T m s among the 4 M. massiliense Type I and 6 M. massiliense Type II sputum samples were 1.0°C (60.0-61.0°C) and 1.1°C (68.5-69.6°C), respectively, which was less than the intersubspecies T m s difference between M. massiliense Type I and Type II (8.8-9.1°C). In the Texas Red channel, the differences in T m s among the 4 M. massiliense Type II-1 and 2 M. massiliense Type II-2 sputum samples were 1.9°C (51.5-53.4°C) and 0.3°C (46.4-46.7°C), respectively, which was less than the inter-genotypic T m s difference between M. massiliense Type II-1 and Type II-2 (4.8-7°C) ( Table 3) in the FAM channel that was specific for M. abscessus group species detection, thereby validating the high specificity of the hsp65 PNA RT-PCR method for the detection of the M. abscessus group from sputum samples. Only one sputum sample (M. fortuitum) formed an insignificant T m in the Hex channel (Fig. 1d).
Comparison of hsp65 PNA RT-PCR versus HMPRT-PCR for the direct detection of the M. abscessus group from sputum samples Previously, we introduced a multiprobe real-time PCR assay targeting hsp65 (HMPRT-PCR) to detect and identify pathogenic mycobacteria directly from sputum specimens. In this study, we compared the 2 methods (hsp65 PNA RT-PCR and HMPRT-PCR) for the direct detection of the M. abscessus group from sputum samples. When the 2 methods were applied to the same 30 sputum DNA samples, the results showed that the hsp65 PNA RT-PCR had higher sensitivity compared to the HMPRT-PCR in detecting M. abscessus from sputum [96.7 % (29/30 samples) vs. 70 % (21/30samples)] ( Table 2). A discrepancy between the two methods was found in nine samples. One sputum sample (S1380) identified as M. massiliense Type II-1 infection by hsp65 PNA RT-PCR was identified as M. avium infection by HMPRT-PCR; its culture identification was confirmed to be a coinfection of M. abscessus and M. avium. HMPRT-PCR failed to detect M. abscessus in this coinfection. Another sample (S1479) identified as M. massiliense Type II-2 infection by hsp65 PNA RT-PCR was identified as a coinfection of M. massiliense and M. avium HMPRT-PCR; its culture identification was M. abscessus infection. HMPRT-PCR could not detect 5 samples identified as M. abscessus infections by hsp65 PNA RT-PCR or two samples identified as M. massiliense Type I and Type II-1 infections by hsp65 PNA RT-PCR (Additional file 1).

Discussion
Recently, we reported that phylogenetic analysis based on the sequence of the 604-bp hsp65 gene enabled taxonomic separation among members of the M. abscessus groups at the genotype and subspecies levels despite representing a single target gene [9]. We also reported that the separation of members of the M. abscessus group based on pulsed-field gel electrophoresis (PFGE) and multi-locus sequence typing (MLST) was nearly in agreement with the results obtained using sequencebased phylogenetic analysis of the 604-bp hsp65 gene [12], which resulted in the phylogenetic separation of four members of the M. abscessus groups [M. abscessus and the three types of M. massiliense (I, II-1 and II-2)].
SNPs have been widely used for the diagnostic targeting of pathogens [27,28] and human diseases [29]. For use in bacterial identification, an SNP should show not only intra-species or genotypic conservation but also   (Table 2), thereby demonstrating its feasibility in identifying the M. abscessus group not only from cultured isolates but also directly from sputum samples. PNA-based molecular beacons were reported to be superior to conventional molecular probes due to their faster hybridization kinetics, high signal to background ratio and improved specificity [30]. Indeed, PNAs have been successfully applied for the highly sensitive detection of anthrax DNA and HIV RNA [31,32]. PNA-based molecular beacons were reported to be advantageous for genotyping short sequences when high sequence specificity was required. The typing of an SNP with DNA probes such as the TaqMan or LightCycler probes (which are usually at least 23 nucleotides in length) can be problematic due to the limited discriminating power of long DNA probes. PNA molecular probes are significantly shorter than the TaqMan or LightCycler probes, making probe design and genotype discrimination easier [33]. Despite the short sequence length (13-15 bp) of the probes, our hsp65 PNA RT-PCR could be successfully applied to the separation of members of the M. abscessus group from sputum samples ( Table 1) as well as clinical isolates [22]. The presence of a mismatch in a PNA/DNA duplex is reported to be more destabilizing than a mismatch in a DNA/DNA duplex, suggesting that the use of PNA probes may lead to more pronounced difference in the average T m in SNP detection compared to the use of DNA probes [34]. In this study, our PNA probes showed relatively large differences (4.8°C to 11.9°C) in T m s in SNP detection despite one nucleotide mismatch. Thus, our probes provided improved discriminative power for the separation of members of the M. abscessus group compared to DNA probes, which are always higher compared to the probes used for intrasubspecies or intra-genotype analyses. This result demonstrates the feasibility and reproducibility of our hsp65 PNA RT-PCR method in a clinical setting. The comparison of hsp65 PNA RT-PCR versus HMPRT-PCR in this study demonstrated the higher sensitivity of the former compared to the latter [96.7 % (29/30 samples) vs. 70 % (21/30 samples)] in detecting M. abscessus from sputum samples. These findings strongly support the hypothesis that PNA-probe based technology may be superior to technologies based on other probes (i.e., FRET-based dual probes for the direct detection of mycobacteria from sputum samples).
Notably, the T m s measured from sputum samples showed small decreasing shifts (1-2.7°C) compared to our previously reported T m s of clinical isolates of the M. abscessus genotypes [22] without compromising the discrimination of the genotypes. This outcome may be due to the presence of PCR-inhibitory substances in the DNA extracts from the sputum specimens because inhibitory substances can cause T m shifts [35,36].