Replacing Reverse Line Blot Hybridization Spoligotyping of the Mycobacterium tuberculosis Complex

Abstract

Spoligotyping is a tool for the molecular characterization/typing of Mycobacterium tuberculosis complex (MTBC) strains based on target sequences (spacers) in the direct repeat (DR) region (14). The standard spoligotyping assay involves the hybridization of amplified sample DNA to nylon membrane-immobilized oligonucleotides whose sequences are representative of 43 spacer regions. Variations in the number of spacers as a result of deletions of adjacent blocks of repetitive units allow the differentiation of clinical isolates. In the present study, we developed a new multiplexed primer extension-based spoligotyping assay using automated matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) that improves the classical reverse line blot hybridization assay with respect to reproducibility, throughput, process flow, ease of use, and data analysis. Validation of the MALDI-TOF MS-based spoligotyping assay with two sample sets with a total of 326 samples resulted in 96.6% concordance (315/326) when the full spoligotype patterns were compared with the results of standard spoligotyping and 99.9% concordance when the results were compared with those of individual primer extension assays. Ten strains (including two Mycobacterium canettii strains) showed discordant results with one or two spacer differences from the membrane-based spoligotyping result. Most discordant samples were identified to be the result of ambiguities in the interpretation of weak hybridization signals in the reverse line blot assay and sequence variations in the spacer regions. We established a new automated primer extension assay and successfully validated it for use for the routine typing of MTBC strains in the research and public health laboratory environments. The present multiplex levels of up to 30 are extendable and allow the additional incorporation of controls and antibiotic resistance markers.

The detection of an estimated 9.3 million new tuberculosis (TB) cases in 2007 (26) stresses that Mycobacterium tuberculosis is still a major world health problem and a great epidemiologic concern. In this context, molecular typing of M. tuberculosis complex (MTBC) isolates has become a powerful tool used to understand and predict ongoing TB transmission patterns at the regional and national scales. The current standardized and most widely used methods are IS6110 restriction fragment length polymorphism (RFLP) typing (23) and, to a growing extent, a combination of interspersed repetitive-unit-variable-number tandem-repeat (MIRU-VNTR) typing and spacer oligonucleotide typing (spoligotyping) (22). The last two methods provide equal specificities and sensitivities for the detection of recent transmission chains (18). They are PCR based and are thus amenable to the detection of trace amounts of material, uncultured cells, or extracts of clinical samples. Extracted DNA, heat-killed cell suspensions from growth medium, and even primary specimens have successfully been used as templates in PCRs (13).

As a first PCR-based method for the molecular typing of MTBC strains, spoligotyping has been used worldwide and enables the rapid provision of typing results for epidemiological purposes and also first insights into the population structure of MTBC (3).

Spoligotyping targets a direct repeat (DR) region located in the chromosomes of members of MTBC (13). The DR region is composed of numerous identical 36-bp direct repeats, interspersed by a variable number of nonrepetitive short sequences called spacers. This region most probably evolved through the deletion of the direct variable repeat sequences by unidirectional events of homologous recombination, single nucleotide mutations, and the integration of IS6110 elements (1, 6, 8, 24, 25).

Spoligotyping assays are currently performed by one of two methods (5). The first, “gold standard” method, established about 10 years ago, involves the detection of the 43 extremely well conserved spacer sequences through amplification of the DR region and subsequent hybridization of the resulting PCR products to nylon membranes with immobilized synthetic oligonucleotide probes specific to each spacer (reverse line blot hybridization). The presence of the spacers is detected by chemiluminescent staining of the membrane in spots where the PCR product hybridizes to the probes. Strains vary in the number of DRs and the presence or absence of particular spacers, resulting in strain-specific hybridization patterns on the membrane. Tens of thousands of isolates analyzed with this method enabled the creation of large international databases and established a global picture of the diversity of M. tuberculosis (2, 3).

A second method replaces the time-consuming membrane step, which requires extensive hands-on time and manual data interpretation and which utilizes a high-throughput, multianalyte flow system (Luminex) that permits the analysis of higher numbers of strains (4). The synthetic spacer oligonucleotide probes are immobilized on microspheres via a conjugation process, and detection is achieved via fluorochromes attached to the beads and hybridized PCR product. This method allows the simultaneous analysis of 96 samples, as opposed to 45 samples for the membrane-based method, while producing a digital output that is readily portable.

Described herein is a novel alternative which improves upon the throughput, process flow, ease of use, and data analysis of the previously used spoligotyping assays. We adopted automated matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) for spoligotype detection and replaced the hybridization step with a multiplexed primer extension assay set, as shown in Fig. 1 (20). The assays were implemented with a test set of 100 strains and validated with two sets comprising a total of 326 strains for which classical membrane-based spoligotyping results were available.

FIG. 1.

Spoligotyping by MALDI-TOF MS targeting the DR region in the chromosomes of members of the MTBC composed of numerous identical 36-bp direct repeats, interspersed with up to 43 nonrepetitive short sequences called spacers. The result for a strain of the M. tuberculosis Beijing family is shown as an example. (A) Amplification of the DR region by PCR, resulting in a mixture of different-size fragments. (B) Single-base extension of oligonucleotides targeting the 43 spacer regions. (C) Mass spectra for the M. tuberculosis Beijing family spoligotype. Spacer oligonucleotides (oligos) that bind to the amplification product result in an extension product and a corresponding mass signal in the spectrum. Spacer oligonucleotides for absent spacers do not get extended and are represented by their mass in the spectrum. (D) Gold standard reverse line blot hybridization assay result.

We have shown that a homogeneous assay format of PCR and a multiplexed primer extension assay followed by MALDI-TOF MS detection on a MassARRAY system (Sequenom, Inc.) streamlines sample processing by avoiding extensive washing steps and microsphere conjugation.

MATERIALS AND METHODS

Bacterial study collection.

Purified DNA of MTBC strains was obtained from the Health Protection Agency (HPA) in London, United Kingdom (100 strains; the HPA test set), and the Research Center in Borstel, Germany (326 strains, the validation set). The DNA from Borstel included DNA from strains collected over 1 year (between 1 January and 31 December 2003) of a population-based epidemiological study carried out in Hamburg, Germany (152 strains; the Hamburg collection) (18), as well as the reference strain set (174 strains; the MIRU-VNTRplus collection) of the MIRU-VNTRplus database (http://www.miru-vntrplus.org/), which comprises the main phylogenetic lineages of the MTBC (2). For all samples, the spoligotypes obtained by the gold standard reverse line blot hybridization method were available from the source agencies.

Assay design.

Oligonucleotides DRa (5′-gttggatgGGTTTTGGTCTGACGAC-3′) and DRb (5′-acgttggatgCCGAGAGGGGACGGAAAC-3′), which target the repetitive unit, were used to amplify the direct repeat region (13). An additional two oligonucleotides 16S_F (aacgttggatgTGGCGGCGTGCTTAAC) and 16S_R (acgttggatgCTACCCGTCGTCGCCTTG) targeting the 16S rRNA region were used as internal controls. The lowercase letters in the primer sequences are sequence tags used to increase the molecular weights of the PCR primers so that any leftover primers from the PCR do not interfere with any signals in the mass window of the spectrum.

Extension assays were designed by the use of QGE assay design software (version 1.0; Sequenom, Inc.). Three different sets of assays (a 23- and 25-plex assay, a 18- and 30-plex assay, and a 20- and 28-plex assay) covering all 43 spacer regions, an additional spacer (spacer 0), as well as two mycobacterial control assays were performed. Oligonucleotides were purchased from Integrated DNA Technologies.

PCR.

Approximately 10 ng of mycobacterial genomic DNA was used per 10-μl PCR mixture. The reaction mixtures contained 1× DNA polymerase buffer [50 mM Tris-HCl, 10 mM KCl, 5 mM (NH₄)₂SO₄, 3.5 mM MgCl₂; Sequenom, Inc.], each deoxynucleotide triphosphate (dNTP; Sequenom, Inc.) at 500 μM, each oligonucleotide (DRa, DRb,16S_F, and 16S_R) at 400 nM, and 0.3 U of polymerase (Sequenom, Inc.). The PCR was initiated at 94°C for 5 min, followed by 20 cycles of denaturation at 94°C for 60 s, annealing at 55°C for 60 s, and extension at 72°C for 30 s and a final extension at 72°C for 3 min. Subsequently, the remaining dNTPs were dephosphorylated with shrimp alkaline phosphatase (SAP) (Sequenom, Inc.) by addition of 0.24× SAP buffer and 1 U of SAP (Sequenom, Inc.). The reaction mixtures were incubated at 37°C for 40 min, followed by inactivation at 85°C for 10 min.

Multiplexed primer extension reactions.

The 9-μl extension reaction mixtures contained 7 μl of the PCR/SAP product, 0.22× TypePLEX buffer plus (Sequenom, Inc.), 1× TypePLEX termination mixture (Sequenom, Inc.), 1× TypePLEX enzyme (Sequenom, Inc.), and 0.625 μM or 1.25 μM each extension oligonucleotide. The cycling parameters included 30 s at 94°C, followed by 40 cycles of 5 s at 94°C, 5 s at 52°C, and 5 s at 80°C and an additional loop of 5 cycles between the 52°C and the 80°C steps. Final extension was carried out at 72°C for 3 min. After conditioning of the reaction mixture by the addition of 16 μl of deionized water and desalting of each reaction mixture by addition of 6 mg Clean resin (Sequenom, Inc.), 15 nl product was applied to a 384-spot of a SpectroCHIP solid support (Sequenom, Inc.). Data were acquired on a MassARRAY compact analyzer (Sequenom, Inc.).

Data analysis.

Analysis of the spoligotyping patterns was performed by the use of MassARRAY Typer analyzer software (version 3.4; Sequenom, Inc.).

Nucleotide sequence accession numbers.

The M. tuberculosis H37Rv complete genome sequence (GenBank accession number Z48304.1) was used for extension primer design. The M. canettii nucleotide sequence (GenBank accession number AF190853) was used for the alignment of the extension primers for spacers 30 and 36.

RESULTS

Three sets of multiplexed primer extension spoligotyping assays (a 23- and 25-plex assay, a 18- and 30-plex assay, and a 20- and 28-plex assay) were designed on the basis of all 43 spacer sequences in the informative DR region of MTBC (13). Target amplification was performed as previously described for the membrane-based spoligotyping approach (13). The amplification products were subsequently interrogated by the use of oligonucleotide extension primers designed to bind to the spacer regions. Extension primers were carefully designed and multiplexed to avoid primer cross-hybridization. Upon binding to the target region, the oligonucleotides were elongated by single-base extension. The assays were designed in such a way that the extension oligonucleotides and their subsequent extension products occupy unique separated masses in the resulting spectrum. The concentrations of the extension primers were adjusted to compensate for differences in the sequences and molecular weights and to equalize the signal intensities in the spectrum. Adjustments were made on the basis of mass spectral data.

The assay principle is shown in Fig. 1. The presence of a specific spacer region in the sample is inferred by a single-base extension product and its corresponding signal in the mass spectrum. Up to two signals can appear in the spectrum, one for the unextended oligonucleotide and the second one for the corresponding extension product, depending upon the extension efficiency. The absence of a spacer results in only one signal at the mass of the unextended oligonucleotide. Data were gathered in electronic form and were automatically extracted by the MassARRAY Typer analyzer software.

A set of 100 samples provided by HPA was used to optimize the PCR and the performance of the assay. PCR amplification was tested at 20, 40, and 45 cycles. The 20-cycle program was found to be superior to the 40- and 45-cycle programs, which revealed unspecific primer extension products, especially for mycobacterial strains outside the MTBC (data not shown).

As an additional quality measure for the detection of members of the MTBC, an internal control based on a 249-bp region of the 16S rRNA gene was added to each assay mixture. The region was coamplified during PCR and allowed the single-base extension of two oligonucleotides, one specific to mycobacterial species in general and the other one specific to members of the MTBC. No extension of any of the two control extension primers indicates no or weak PCR performance and suggests the need for a repeat test of the sample with, for example, increased DNA template amounts. Internal controls were tested as part of all spoligotyping assays by using two species outside the MTBC (M. gordonae and M. smegmatis), as well as M. bovis BCG, a member of the MTBC. Template DNA was titrated from 0.7 ng/PCR mixture to 5.6 ng/PCR mixture. Spoligotyping pattern and positive internal controls were detected for M. bovis BCG at all DNA concentrations tested. M. gordonae and M. smegmatis showed positive reactions by the mycobacterial species control assay when their DNA was used at template concentrations greater than 1.4 ng/PCR mixture and no signal for the MTBC-specific control reaction when their DNA was used at any template concentration (data not shown).

After optimization, primer extension and MALDI-TOF MS of the HPA sample set resulted in spoligotyping results for 89 strains. We identified 56 unique groups generated by 41 unique and 15 shared spoligotypes. Ten samples gave no spoligotyping results and were identified as mycobacterial species outside the MTBC (M. avium, M. kansasii, M. gordonae, M. smegmatis, M. xenopi). In addition, one of the samples repeatedly failed processing and was excluded from analysis.

Validation of spoligotyping by use of the MassARRAY system was performed with two strain collections from the Research Center Borstel (the Hamburg and the MIRU-VNTRplus collections). The MIRU-VNTRplus collection comprises the maximum diversity of the MTBC, ranging from various M. tuberculosis lineages such as Haarlem, Beijing, and East African Indian to M. africanum and M. bovis (see www.MIRU-VNTRplus.org). All samples had previously been spoligotyped by the reverse line blot hybridization membrane-based assay.

All three assay sets (the 23- and 25-plex, 18- and 30-plex, and 20- and 28-plex assays) were run with 326 DNA samples. After filtering of the samples throughout all three sets which repeatedly failed PCR processing, PCR amplification, single-base extension of the 43 spacer regions, and MALDI-TOF MS analysis provided spoligotypes for 297 samples. Seventy-six unique groups and individual spoligotypes were identified for the samples in the Hamburg collection, and 93 unique groups and individual spoligotypes were identified for the samples in the MIRU-VNTRplus collection. This generated 57 unique and 19 shared spoligotypes and 63 unique and 30 shared spoligotypes for the two collections, respectively.

To evaluate the performance of the assay sets, we looked at the concordance of the spoligotypes identified with those identified by the membrane-based method. The 23- and 25-plex assay achieved a concordance of 96.6% (287/297 spoligotypes), while the 18- and 30-plex and the 20- and 28-plex assays achieved concordances of 94.3% (280/297 spoligotypes) and 95.6% (284/297 spoligotypes), respectively. When the spacers were considered independently, the overall rate of concordant calls for the corresponding individual primer extension assays was greater than 99.8% (on average, 12,750/12,771 spacers) for all three assay sets.

On the basis of these results, the 23- and 25-plex assay was chosen as the one with the superior assay set. The primer sequences of the extension primers, their expected masses, as well as the extension products and corresponding masses are listed in Table S1 in the supplemental material for the 25-plex assay and in Table S2 in the supplemental material for the 23-plex. The extension primer concentrations for the 23- and 25-plex assay set are listed in Tables S3 and S4 in the supplemental material.

The 23- and 25-plex assay was used to repeat the tests with all failed and discordant samples. Spoligotypes were obtained for all but 1 sample, so spoligotypes were obtained for a total of 325 samples. The corresponding spoligotyping patterns are summarized in Fig. S1 in the supplemental material. The final concordance of the spoligotyping pattern obtained by the 23- and 25-plex assay with that obtained by the membrane-based approach increased to 96.9% (315/325), and the number of concordant primer extension assays was 99.9% (13,964/13,975).

Overall, 10 strains (4 of 152 of the Hamburg collection and 6 of 173 of the MIRU-VNTRplus collection) showed a MALDI-TOF MS-derived spoligotyping pattern with minor differences (one or two spacers) from the membrane-based results. Discordant samples were further evaluated and repeat tested by membrane-based spoligotyping. Figure 2 shows an example of a discordant spoligotype. Spacer 26 was found to be absent from the mass spectra of the MALDI-TOF MS approach, while it was present on the first spoligotyping membrane but absent on the second membrane. Three repeat tests of the sample by MALDI-TOF MS reproducibly confirmed the absence of spacer 26 (data not shown). In the MALDI-TOF MS approach, the presence of a spacer in the sample is inferred by an extension product and its corresponding signal in the mass spectrum, while the absence of a spacer results in the presence of an unextended primer and its corresponding signal in the mass spectrum. There are thus confirmatory signals both for the absence and the presence of a spacer. In the autoradiograph of the traditional membrane-based spoligotyping approach, the presence of the spacer is inferred by the shade of the signal in comparison to that of a control signal for the equivalent region. The absence of the spacer does not produce a confirmatory signal event. Weak hybridization signals due to sequence variations in the spacer regions are encountered and result in ambiguities in interpretation and discordance of repeated sample runs, as presented.

FIG. 2.

Spoligotyping pattern of the same sample obtained by the membrane-based spoligotyping method on two different membranes. In comparison, the mass spectra of the same sample on three different days run over the 23- and 25-plex assay show no extension of the primer for spacer 26. One representative spectrum is displayed. Extended products are labeled by the corresponding spacer numbers. The unextended missing primer signal for spacer 26 is boxed.

Two consistently discordant samples were identified as M. canettii strains. M. canettii, M. microti, and M. pinnipedii have spoligotyping patterns very different from those of the members of the MTBC and hybridize to few, if any, of the 43 traditional spoligotyping spacers (3, 21). For M. canettii, the hybridization of spacers 30 and 36 has been observed by membrane-based spoligotyping, although the signals are weak and can be seen only if the standard PCR conditions are changed. No primer extension of this spacer region was detected in the MALDI-TOF MS approach, while both internal control assays were positive. Sequence analysis of the M. canettii direct repeat region and the extension primers revealed significant sequence polymorphisms and thus no extension products in the reactions. As shown in Fig. 3, this is especially apparent for spacer 36, for which the mismatches on the 3′ end of the primer prevent extension.

FIG. 3.

Mycobacterium canettii (GenBank accession number AF190853) direct repeat regions for spacers 30 and 36 with alignments of the MassARRAY extension primer sequences. Matching and mismatching bases are indicated. Due to the amount of mismatched bases throughout the primer sequences as well as at the 3′ end, neither primer showed any hybridization or extension by the MassARRAY spoligotyping approach.

DISCUSSION

Worldwide, spoligotyping has become the most widely used method for the molecular typing of clinical MTBC isolates, because it is inexpensive, rapid, and easy to perform. Recently, the combination of spoligotyping and MIRU-VNTR typing has been proposed as a new standard scheme for molecular epidemiological studies (22). As a crucial point, comparison of the spoligotyping results for the large data sets necessary for longitudinal epidemiological studies requires highly standardized experimental and computerized procedures, including digital data formats. However, the reproducibility of traditional membrane-based spoligotyping has been shown to be suboptimal, mainly due to variations in membrane production, variations in the intensities of the hybridization signals for particular spacers, and the need for manual interpretation of the hybridization results (14).

In the study described here, we showed that spoligotyping by the use of PCR, primer extension, and MALDI-TOF MS is a superior alternative to traditional membrane-based spoligotyping. The membrane-based reverse line blot hybridization process involves four steps, consisting of PCR target amplification, hybridization, posthybridization wash steps, and chemiluminescent detection. The number of samples that can be tested simultaneously is limited, as the Miniblotter apparatus (Immunetic, Inc.) allows the simultaneous testing of only 45 samples. As a major drawback, the method also requires manual data inspection and transfer from the membrane into a digital format, resulting in the clear potential for individual differences in data interpretation. This can be complicated, as data interpretation is heavily dependent upon the quality of the membranes, and it can be challenging to differentiate background noise from nonspecific hybridization versus weak signals. The use of a validation procedure to control the proper manufacture of a new spoligotyping membrane by including a series of previously characterized strains is recommended for each new membrane lot (5).

We have utilized an established primer extension and MALDI-TOF MS detection approach and applied it to the routine typing of MTBC strains for rapid, reproducible, and easy-to-perform spoligotyping in the research and public health laboratory environments. Spoligotyping by MALDI-TOF MS yields portable results for interlaboratory comparisons in an electronic form, forgoing any need to transfer blot patterns manually to electronic form. The throughput is greatly improved, while the hands-on time is reduced. The use of robotic liquid handling for the PCR and primer extension setup as well as the MALDI-TOF MS method enables the analysis of 48 to 192 samples per microtiter plate in about 7 h, including data analysis. Multiple microtiter plates can be processed in parallel. For comparison, a spoligotyping analysis of 45 samples under optimal conditions by reverse line blot hybridization requires >24 h of labor, and the results depend on careful manual interpretation of the developed X-ray films. Data analysis by the MassARRAY Typer analyzer software provides an automated output of the spoligotypes found to be present in the sample as well as cluster plot analysis for all spoligotyping assays. Cluster plots were used to make manual calls for samples with spoligotyping assay results showing weak extension rates. An example of a representative cluster plot is given in Fig. S2 in the supplemental material.

We demonstrate that the new MALDI-TOF MS approach produces results highly concordant with those derived from the traditional membrane-based method. The variations detected in 8 of 323 strains (the 2 M. canettii strains excluded) were a result of minor differences of one spacer (for 7 strains) and two spacers (for 1 strain), while all other spacers for these eight strains were concordant between the two methods.

In one case, we demonstrated that the result for a particular spacer was also variable when different membranes were used (Fig. 2). In this example of discordant results between the membrane-based and the MALDI-TOF MS approaches, we show reproducible and consistent automated MALDI-TOF MS spoligotyping calls between all three plex combinations tested over three repeats each. Since the three MALDI-TOF MS-specific plex combinations tested in the study use different extension primers on both DNA strands to target the same spoligotyping region, great confidence can be attributed to the result when the results for all three are in agreement.

Although the MassARRAY assay design (Sequenom, Inc.) can achieve a multiplex level as high as 60, we use a substantially lower level of multiplexing, so that the primer extension assay design for the MALDI-TOF MS system presented here will remain sufficiently flexible. Additional assays such as control assays, assays with additional spacer sequences, M. canettii-specific assays, and antibiotic resistance assays can be incorporated with ease. This can allow resistance detection and analysis of mixtures of mutants and the wild type, reflected as corresponding peak ratios with quantitative application.

An advantage of the MassARRAY system-based spoligotyping approach that has already been implemented is an internal amplification control. We show a strong correlation of the overall reaction failure and internal control failure for the validated 23- and 25-plex assay. The control implemented is helpful for the detection of whether mycobacteria other than MTBC or members of the MTBC with divergent spacer sequences like M. canettii are present. Our incorporation of the 16S rRNA control is a conservative approach to data quality, as it helps to identify weaker reactions that require an increase in the amount of template DNA used in the PCR.

Spoligotyping by the MassARRAY system with the 23- and 25-plex assay presented here offers many benefits, such as higher throughput and decreased labor, whereas the membrane-based method requires hybridization, wash steps, and manual data analysis. The assay is homogeneous and does not require the conjugation or wash steps needed for the membrane and the bead-based Luminex method. Internal controls ensure appropriate quality control for each reaction and eliminate the need for the positive-control samples that are required for the membrane-based method. The results are automatically obtained by the analysis software in digital format and are not affected by subjective interpretation of the hybridization signals. The MassARRAY system produces numerical data. Each spoligotype is interpreted as 43 binary characters plus controls. Each binary character denotes the presence or the absence of a spacer in the DR locus of the MTBC. The final spoligotypes can thus be easily reported by simple Excel spreadsheet calculation without manual data entry, which makes the results readily shareable among laboratories and enables comparison of the results to those that exist in international databases, such as the MIRU-VNTRplus (2) and SpolDB (3) databases.

Repetitive units are a powerful tool for the identification and characterization of multifold organisms. The rapid increase in the number of genomic sequences available provides new opportunities for comparative genomics and the development of effective genotyping methods. In this context, it was found that the direct repeat region of the MTBC is among a family of repetitive genome sequences called clustered regularly interspersed short palindromic repeats (CRISPRs), which are present in many different species of bacteria and archaea (7, 12, 15, 17). Methods similar to spoligotyping for the typing of other bacterial isolates with CRISPRs are being developed and can utilize the MassARRAY system. In general, the technology and the automated assay design tool are flexible and able to replace dot blot detection assays and membrane-based approaches, surpassing previous MALDI-TOF MS applications with respect to high-throughput single nucleotide polymorphism (SNP) genotyping (19) and variant detection (16, 27). Likewise, the MassARRAY platform has already been shown to have the ability to be used for bacterial pathogen identification through comparative sequencing (9-11).

In conclusion, spoligotyping by use of the MassARRAY system was able to detect the 43 spoligotype spacers in two multiplex reactions that are accurate and whose results are concordant with those of the current standard method. Due to the flexible assay design capabilities of the platform, extension of the assays, e.g., for the detection of phylogenetically informative SNPs or SNPs involved in drug resistance, is desirable. The MassARRAY system is ideal as a technology complementary to the other molecular typing methods currently employed in pharmaceutical and public health studies.