A single-nucleotide-polymorphism real-time PCR assay for genotyping of Mycobacterium tuberculosis complex in peri-urban Kampala

Background Accurate and high-throughput genotyping of Mycobacterium tuberculosis complex (MTBC) may be important for understanding the epidemiology and pathogenesis of tuberculosis (TB). In this study, we report the development of a LightCycler® real-time PCR single-nucleotide-polymorphism (LRPS) assay for the rapid determination of MTBC lineages/sublineages in minimally processed sputum samples from TB patients. Method Genotyping analysis of 70 MTBC strains was performed using the Long Sequence Polymorphism-PCR (LSP-PCR) technique and the LRPS assay in parallel. For targeted sequencing, 9 MTBC isolates (three isolates per MTBC lineage) were analyzed for lineage-specific single nucleotide polymorphisms (SNPs) in the following three genes to verify LRPS results: Rv004c for MTB Uganda family, Rv2962 for MTB lineage 4, and Rv0129c for MTB lineage 3. The MTBC lineages present in 300 smear-positive sputum samples were then determined by the validated LRPS method without prior culturing. Results The LSP-PCR and LRPS assays produced consistent genotyping data for all 70 MTBC strains; however, the LSP-PCR assay was 10-fold less sensitive than the LRPS method and required higher DNA concentrations to successfully characterize the MTBC lineage of certain samples. Targeted sequencing of genes containing lineage-specific SNPs was 100 % concordant with the genotyping results and provided further validation of the LRPS assay. Of the 300 sputum samples analyzed, 58 % contained MTBC from the MTBC-Uganda family, 27 % from the MTBC lineage 4 (excluding MTBC Uganda family), 13 % from the MTBC lineage 3, and the remaining 2 % were of indeterminate lineage. Conclusion The LRPS assay is a sensitive, high-throughput technique with potential application to routine genotyping of MTBC in sputum samples from TB patients. Electronic supplementary material The online version of this article (doi:10.1186/s12879-015-1121-7) contains supplementary material, which is available to authorized users.


Introduction
Mycobacterium tuberculosis (MTB) is an acid-fast bacillus that causes tuberculosis (TB) a chronically debilitating disease with a mortality rate approaching 2 million deaths per year [1][2][3]. The disease primarily develops in 5-10 % individuals following inhalation of air droplets containing Mycobacterium tuberculosis complex (MTBC) bacilli, but may also occur following reactivation of a latent infection [4]. In Kampala, Uganda, 3 dominant MTBC genotypes have been identified namely MTBC Uganda family that accounts for 63 % of TB cases, followed by other MTBC lineage 4 genotypes other than Uganda genotype and then MTBC lineage 3 [5,6]. These genotypes present with diverse clinical outcomes for instance MTBC Uganda family genotypes are less prone to drug-resistance, less virulent, and not associated with extra pulmonary TB [5,[7][8][9][10]. The MTBC lineage 4 genotypes progress rapid to disease compared to other genotypes [11,12], while the MTBC lineage 3 genotypes cause severe disease [13]. Therefore accurate determination of the MTBC strain diversity within a population like Kampala can lead to the design of intervention strategies that more effectively target circulating strains.
The currently available MTBC genotyping assays are challenging to implement in areas with endemic TB and are limited in their ability to discriminate MTBC strains present in clinical isolates. For example robust techniques such multi-locus sequence typing (MLST) [14] and whole genome sequencing (WGS) [15,16], are difficult to adopt in resource-limited countries because they are prohibitively expensive [17]. Other techniques, such as MIRU-VNTR, IS6110-RFLP, PGRS-RFLP, and CRISP [18,19], can erroneously classify MTBC lineages [16,20] due to homoplasy and are technically cumbersome. Furthermore, some of these methods typically require prior culturing of MTB from sputum samples, a process that takes 1-2 months [21]. For samples containing a mixed MTBC population, this culturing step may skew strain diversity by promoting growth competition between different strains [22]. Thus, there is a need for a more robust genotyping assay that is fast, sensitive, and can be applied directly to processed sputum samples without prior culturing.
To mitigate the aforementioned flaws a real-time PCR (RT-PCR) assay-the LightCycler® 480 RT-PCR SNP (LRPS) assay-was developed to genotype MTBC directly from processed sputum samples using hybridization probes. This assay was evaluated for the ability to accurately identify MTBC lineages in peri-urban Kampala and subsequently used to analyze 300 smear-positive sputum samples from individual patients.

Identification of lineage-specific SNPs for genotyping MTBC
The MTBC lineage-specific SNPs used in this study were obtained from whole genome sequencing data as previously described [14 16] with reference to the first MTBC (i.e., H37Rv) genome [23] to be sequenced. A SNP corresponding to a specific MTB lineage/sublineage was annotated by showing its position in the corresponding gene (ORF) and the associated nucleotide change (See Additional file 1: Table S1).

Design of primers and probes for LRPS assay
Primers and probes for typing MTBC Uganda family (MTB L4-U) MTBC lineage 4 excluding the MTBC Uganda family (MTB L4-NU), and MTBC lineage 3 (MTB L3) were chosen based on the list of lineage-specific SNPs described previously [14,16] (See Additional file 2: Table S2). Light-Cycler® Probe Design Software 2 (Roche Applied Science, Germany) was used for the design of assay primers and probes. In brief, RT-PCR enables the quantitative detection of a particular segment of DNA by coupling a fluorescent signal with DNA amplification. The fluorescence produced during amplification is directly proportional to the amount of DNA present in a given sample, amplification efficiency of the primer and probe combination. In order to distinguish the DNA of different MTBC lineages, hybridization probes were designed to recognize unique SNPs that are specific to particular MTBC lineages/sublineages. To identify MTBC L4-U and MTBC L3, hybridization probes were designed to perfectly complement wild type (H37Rv) DNA for MTBC L4-NU probes were designed to complement the mutant DNA. Thus, for MTBC L4-U and MTBC L3 probes will produce lower melting temperature (T m ) values (due to single base mismatch) than samples with wild type DNA (due to perfect match), whereas the MTBC L4-NU probes will produce higher T m values with mutant DNA (due to perfect match). We used MTBC lineage 3 (CAS), Uganda family and H37Rv DNA as controls.

MTBC DNA extraction
The genomic DNA from stored isolates was extracted as described by Wampande et al. Stucki et al. [5,24]. Frozen isolates that were earlier characterized as MTBC by IS6110-PCR were thawed, centrifuged, and the pellet washed twice with phosphate buffer saline (PBS). Pellets were subsequently reconstituted in 100 μl PCR water, heated at 90°C for 1 h, and sonicated for 15 min to completely lyse the bacilli and release the genomic DNA. The latter was recovered in the supernatant following centrifugation at 13000 g for 30 min, quantified by Qubit® 2.0 fluorometer (Invitrogen, USA) and used immediately or stored at −20°C for future use.

Long sequence polymorphism (LSP) -PCR analysis
In order to ascertain whether the LRPS assay correctly identifies MTBC lineage we compared the LRPS results with LSP-PCR data using genomic DNA extracted from 70 MTBC stored isolates. LSP-PCR was performed using RD 724 deletion primers (specific for MTB Uganda family) and RD750 deletion primers (Specific for MTB lineage 3) as described by Gagneux et al. [25] and Tsolaki et al. [26]. The 10 μl reaction volume PCR was containing 5.5 μl water, 1 μl (10 μM final concentration) forward primer (RD 724 or RD 750) and 1 μl (10 μM final concentration) reverse primer (Reverse RD 724 or RD 750), 1 μl of 10 x Thermo Fischer Scientific Custom PCR Master mix, 1 μl template DNA (at least 50 ng) and 0.5 μl (0.5 unit) DNA polymerase. The reaction was run in a standard thermocycler programmed at 95°C for 10 min 35 cycles of 95°C for 1 min, 64°C for 30 s and 72°C for 30 s. PCR products were analyzed by gel electrophoresis.

PCR and targeted sequencing
To further validate the LRPS assay target sequencing of three ORFs (Rv004c for MTBC L4-U Rv2962 for MTBC L4 and Rv0129c for MTBC L3) which contain lineagespecific SNPs from 9 MTBC isolates was performed using primers in Table 1. The PCR was run in a 20 μl reaction volume with 12 μl water 1 μl forward primer and 1 μl reverse primer (0.5 μM of each of the primers), 4 μl of 5 x Roche genotyping master mix (containing Taq polymerase and dNTPs), and 2 μl template DNA (at least 50 ng). The reaction was run in a standard thermocycler (95°C for 10 min 35 cycles of 95°C for 10 s primer(s) annealing (57°C for Rv004c ORF or 53°C for Rv0129c ORF or 51°C for Rv2962 ORF) for 10 s 72°C for 10 s. PCR products were analyzed by gel electrophoresis purified by Qiagen PCR purification kit and commercially sequenced using the primers designed for the specific ORFs (See Table 1). To confirm the presence of the SNP in the respective ORFs, BioEdit software (Ibis Biosciences, USA) was used to align the corresponding H37Rv ORF with the sequenced fragments.

Patient recruitment and processing sputum samples
Patients were recruited from Mulago hospital TB clinic (ward 5 & 6) which serves as the main referral TB centre in Uganda. Sample processing confirmatory microscopy (both ZN and auramine staining) and culturing was done at Mycobacteriology Laboratory, Department of Medical Microbiology, College of Health Sciences, Makerere University. 300 sputum samples were processed in a Biosafety cabinet class II following standard procedures [27]. The final sediment was suspended in 2.5 ml PBS buffer (pH 6.8) part of this sample was used to inoculate Middlebrook 7H10 supplemented with 10 % (v/v) glycerol and OADC for culturing following standard procedures [28] and the remainder was resuspended in 50 μl of PCR water for DNA extraction. The latter was heat killed at 95°C for 1 h and later sonicated for 15 min to completely lyse the bacilli and release the genomic DNA. The genomic DNA was obtained in the supernatant following centrifugation at 13000 rpm and stored at −20°C or used immediately in the LRPS assay.

LRPS genotyping assay
Genotyping of MTBC from cultured and frozen isolates or processed sputum samples was performed using ORF-specific primers and probes designed based on SNP positions (See Table 1 and Additional file 2: Table  S2) in Roche LightCycler ® RT-PCR 480 machine (Roche Applied Science Germany). Briefly, the LRPS assay was run in 20 μl reaction containing 11.2 μl of PCR water, 1 μl (0.5 μM final concentration) reverse primer, 1 μl x Roche genotyping master mix, and 2 μl (containing at least 5-50 ng) of extracted genomic DNA. The Roche LightCycler® 480 (Roche Applied Science, Germany) was programmed for PCR amplification and a melting curve stage. For each of the three uniplex assays, the amplification stage consisted of a pre-PCR stage performed at 95°C for 10 min, an amplification stage with denaturation at 95°C for 10 s, primer annealing (57°C for Rv004c ORF or 53°C for Rv0129c ORF or 51°C for Rv2962 ORF) for 10 s with a single acquisition mode to allow capture of the fluorescence, and extension at 72°C for 10 s for 45 cycles. The melting curve analysis  consisted of denaturation of amplicons at 95°C for 1 min to produce single stranded DNA (ssDNA), probe annealing temperature at 40°C (allows hybridization of the probes to the complimentary sites of the ssDNA) for 10 s, probe melting temperature ranging from 40-80°C (allows the probe to detach from the ssDNA) with a continuous mode of acquisition at a rate of 1 acquisition/s that allows capture of the fluorescence.

Data analysis
From the melting curves the LightCycler® 480 software version 1.2 (Roche Applied Science, Germany) was used to derive probe melting temperature (Tm), which is lineage-specific. For MTBC Uganda family and MTB Lineage 3, the Tm was lower (due to mismatch) than the wild type (perfect match) yet for MTBC lineage 4 the Tm (due to perfect match) was higher than the wild type.

Ethical consideration
This study obtained ethical approval from Makerere University institutional review board and the Uganda National Council for Science and Technology. Written informed consent was obtained from all the study participants.

Primers and probes used in genotyping MTBC
A total of eight primer/probe sets were successfully designed using the LightCycler® probe design (Roche Applied Science Germany) software 2 (Additional file 2: Table S2). During the LRPS optimization step primer probe set Rv0006 a and Rv0407 b failed to give signals (no amplification), primer/probe set Rv3133c C and Rv2959c were giving results that were conflicting with the positive (Central Asian Strain, CAS) and negative controls (H37Rv), thus these sets were excluded. Also Primer/ probe set Rv2949c a was excluded from further sample analysis since it required twice the probe concentration as the counterpart. Therefore primer/probe sets Rv004c a was used to analyze samples for presence of MTBC Uganda family; Rv2962 b and Rv0129c c primer sets were used to identify MTBC lineage 4 and MTBC lineage 3 respectively (Table 1).

Identification of MTBC Uganda family using LSP-PCR and LRPS assay
LSP-PCR and targeted sequencing reactions were run in parallel to validate the LRPS assay. A total of 70 MTBC isolates (confirmed by IS6110 PCR) were genotyped using LSP-PCR with primers specific for the RD 724 deletion in parallel with LRPS assay using Rv004c a primer/ probes set (Table 1). While both assays were equally capable of identifying MTBC Uganda ( Fig. 1 and Additional file 5 Fig S2(a)) only the LRPS assay was able to genotype all 70 samples the LSP-PCR assay failed to identify the MTBC lineages of several isolates (Fig. 1: lane 11 13, 14, 15, 21, 22, 26, 27, 33, 39, 43, 45, 56, 61, 62). With that result, both assays were re-evaluated using serially diluted H37Rv genomic DNA (10 ng-100 ng for LSP-PCR and 1 ng-10 ng for LRPS per PCR reaction) extracted by the Enzyme/CTAB method [29]. The detection limit of the LSP-PCR assay was 10-fold higher than that of the LRPS assay (30 ng approximately 7 × 10 6 copies/reaction and 3 ng approximately 7 × 10 5 copies/reaction) respectively (Additional file 4: Figure S1). Later, DNA from those samples that previously produced no LSP-PCR product (Fig. 1: lane 11

Identification of MTBC Lineage 3 genotype using LSP-PCR and LRPS
A total of 33 MTBC isolates (See Additional file 3: Table  S3) that were not MTBC Uganda (non-Uganda family MTBC) by RD 724 analysis (See Fig. 1 Fig. 2b peak with 57°C and Additional file 5 Fig S2b).

Sequencing PCR products to identify MTB lineage-specific SNP
To ascertain the accuracy of LRPS in genotyping MTBC lineages PCR products of 9 MTBC isolates (3 isolates for each lineage) were sequenced. The resulting sequences of the gene containing the lineage-specific SNP were compared with the corresponding H37Rv sequences using Bio Edit software (Ibis Biosciences, USA) the data confirmed that the sequenced PCR products contained lineage-specific SNPs (See Fig. 3).

Discussion
A number of useful SNPs for robust genotyping of MTBC have been made available following the interrogation of whole genome sequence (WGS) data from global MTBC strains [14,16,30,31]. Comparisons of SNPs with other markers used in molecular epidemiological studies of MTBC have proved their superiority since they are able to discriminate closely unrelated (no homoplasy) genotypes of MTBC [16]. Therefore, this study evaluated the use of novel single nucleotide polymorphic (SNP) markers in a LightCycler ® 480 (Roche, Germany) real-time PCR (LRPS) assay to genotype MTBC isolates using heat inactivated samples. First, SNP-pecific primers/probes were designed to accurately delineate MTBC lineages by real time PCR (RT-PCR). Secondly, the optimized and validated LRPS was used to sub-type MTBC lineages present in 300 smear-positive sputum samples from different individuals. By and large, these data suggest that LRPS can be used to accurately identify MTBC genotypes using heat killed crude lysates of processed sputum samples without prior culturing. The findings show the successful use of 3 sets of primers/hybridization probes containing MTBC lineagespecific SNPs in RT-PCR (LRSP) to accurately delineate MTBC lineages (MTBC L4-U MTBC L4-NU and MTBC L3). The LRSP assay is based on coupling a fluorescent dye to an amplified segment of DNA and use of MTBC lineage specific probes-conjugated to dyes that ensures real-time identification of the genotype in 2 h for every 92 samples. This is contrary to other MTBC genotyping assays that require different step(s) for detection of the genotype, hence increasing the turnaround time 1-3 days depending on the method [17]. Comparison of LSP-PCR and targeted sequencing data of genes containing these lineage-specific SNPs as a step to validate the SNP assay indicate 100 % concordance. This agreement was not surprising since LSP-PCR, LRPS and sequencing methodologies have been used before to accurately genotype MTBC [14,16,32]. However, the detection limit of LSP-PCR was 10-fold higher than the SNPbased assay, thus rendering LRPS more sensitive (3 ng/ assay or 7 × 10 5 copies/reaction) enough to be used with low DNA concentration as well as with heat-inactivated samples. The advantage of using heat lysates eliminates the long steps in DNA extraction which is labor intensive and time consuming. Thus, the LRPS may be better suited for genotyping MTBC from processed sputum samples, which can have a low bacillary load. The enhanced ability to detect small amounts of DNA in a sample by LRPS can likely be attributed to the more sensitive fluorescence detection system of the Roche Light-Cycler® 480 (Roche Applied Science, Germany) machine [33]. In contrast, conventional PCR-based assays are limited by quantifiable DNA, the relatively poor sensitivity of gel-based DNA detection systems, and have a long turnaround time [34]. To support this observation, PCRbased genotyping methods such as spoligotyping and MIRU-VNTR require 20-50 ng of template DNA for a successful run if the detection system is modified as seen in luminex spoligotyping (Luminex Technology, TX, USA) or automated MIRU-VNTR, the sensitivity increases and the turnaround time is reduced, but these methods are still prone to misclassification of MTB lineages [17]. Taken together, this data suggest that the LRPS is more sensitive than the LSP-PCR approach and fast in identifying MTBC in clinical samples since the culturing step is eliminated, making it more suitable for early TB diagnosis, genotyping applications that involve samples with low bacillary loads for instance in TB/ HIV coinfected patients and smear negative TB patients.
Unlike the LRPS most available genotyping methods rely on prior culturing of MTBC, which is laborintensive, time-consuming, and introduces the risk of selective growth in cases of mixed infections [35]. While previous efforts have been made to sub-type MTBC isolates directly from sputum using MTB lineage-specific PCR, MIRU-VNTR and spoligotyping, these approaches have seen limited success [35,36] due to their requirement for relatively high amounts of DNA and/or the presence of inhibitors in the sample and a long turnaround time. In the current study, validated novel SNPbased genetic markers were evaluated to genotype MTBC isolates directly from processed sputum by LRPS without prior culturing. The assay successfully genotyped 300 MTBC isolates from sputum samples of these 58 % were classified as MTB L4-U, 27 % as MTB L4-NU, 13 % as MTB L3 and 2 % as unknown MTBC lineage, and these proportions did not significantly differ from the work published by Wampande et al., 2013 [6]. Notably, the LRPS assay was able to detect more than one genotype in certain isolates (11/300) in contrast the MIRU-VNTR method failed to reveal mixed infections in these samples, presumably due to the lack of sufficient DNA. Overall, these data indicate that the LRPS assay can be used directly on smear-positive, processed sputum samples to genotype MTBC. Due to its high sensitivity and the use of 2 probes with distinct melting curves, this assay has the potential to detect mixed infections in clinical isolates.

Limitations
While the probes used in this study were designed to rapidly identify the three MTBC lineages circulating in peri-urban Kampala [5] additional MTBC lineagespecific probes would need to be developed to genotype other MTBC lineages/sublineages. The current assay is robust in defining deep phylogeny, but is alone not suitable for transmission studies in such circumstances MIRU-NVTR could be used in tandem with the LRPS assay. PCR inhibition was not observed in the LRPS assay, but, it could be relevant and impact negatively on the assay due to minimal buffering. Furthermore, only smear-positive sputum samples, which typically contain a high MTB DNA concentration were analyzed in this study thus further studies will be required to evaluate assay performance on smear-negative samples. The maintenance and initial cost of Roche LightCycler® 480 (Roche Applied Science, Germany) machines are very high, however LRPS is robust, of high throughput, and fast to perform: it has diverse applications, for instance, mRNA display studies, HIV viral load studies and as an ordinary PCR machine.

Conclusion
The LRPS assay is a sensitive rapid, simple and highthroughput technique for detecting and/or genotyping MTBC from minimally processed, smear positive sputum and should be broadly applicable to genotyping SNPs in other microorganisms.

Additional files
Additional file 1: Table S1.