Differential Identification of Mycobacterium tuberculosis Complex and Nontuberculous Mycobacteria by Duplex PCR Assay Using the RNA Polymerase Gene (rpoB)

Abstract

A novel duplex PCR method that can amplify the 235- and 136-bp rpoB DNAs of Mycobacterium tuberculosis complex and nontuberculous mycobacteria (NTM), respectively, with two different sets of primers was used to differentially identify 44 reference strains and 379 clinical isolates of mycobacteria in a single-step assay. Showing 100% sensitivity and specificity, the duplex PCR method could clearly differentiate M. tuberculosis complex and NTM strains. In addition, restriction fragment length polymorphism analysis and direct sequencing of the amplicon of NTM could be used to supplement species identification.

With the recent global resurgence of mycobacterial infections, especially of tuberculosis, attributed to increased human immunodeficiency virus infection, there is an increasing demand for rapid, sensitive, and specific diagnostic methods for the detection and identification of Mycobacterium tuberculosis and nontuberculous mycobacteria (NTM) in a clinical setting (2, 3, 4). NTM infection can cause clinical problems, as its pathogenic potential and susceptibilities to antituberculosis treatments vary (22). In addition, mixed infections of M. tuberculosis and NTM have been reported (16). Therefore, it has become important to be able to differentiate between the two during the early stage of the diagnostic procedure.

The diagnosis of mycobacterial infection is accomplished by culture-based identification. Primary culture of slowly growing mycobacteria, without using the BACTEC culture system, usually takes 4 to 6 weeks or longer (10). However, recent methodological advances in molecular biology have provided alternative rapid approaches, e.g., the PCR and PCR-linked methods. For the rapid detection or identification of M. tuberculosis, target genes specific to mycobacteria are used in a PCR (7, 8, 17, 19).

Because the incidence of NTM infection is increasing, any methods capable of simultaneously determining the presence of M. tuberculosis and/or NTM would be useful. For this purpose, multiplex PCR, which simultaneously uses two or three different genes, has been frequently used, as the technique can specifically detect and identify different species of the genus Mycobacterium (6, 15, 18) and differentiate members of the M. tuberculosis complex (6, 9) in the routine diagnostic laboratory by using Mycobacterium genus- and species-specific genes. However, some of these genes have been found to lack specificity for M. tuberculosis. In addition, IS6110 PCR has been reported to produce false-negative (23) and false-positive (11) results, and the mtp40 gene is not present in all M. tuberculosis strains (21). These reports suggest that the multiplex PCR targeting of these genes has associated problems.

In the present study, we used a simplified multiplex PCR assay, basically a duplex PCR (DPCR) assay, to differentiate M. tuberculosis complex and NTM by using a single gene, the RNA polymerase β-subunit-encoding gene (rpoB). To demonstrate the efficiency and usefulness of the DPCR assay in this context, we used it to identify reference strains and clinical isolates of mycobacteria.

Forty-four mycobacteria and 17 non-mycobacteria were used as reference strains (Table 1). Three hundred seventy-nine clinical isolates (193 M. tuberculosis complex [183 M. tuberculosis and 10 M. bovis] and 186 NTM isolates), which had been isolated by the Korean Institute of Tuberculosis and the Seoul National University Hospital, were identified by conventional biochemical tests and provided for blind testing. An IS6110 PCR assay and analysis of partial 16S rRNA gene (rDNA) sequences were performed separately, and the results were compared with those of the DPCR assay.

TABLE 1.

Mycobacteria and non-mycobacteria used for DPCR and their PCR amplification products

Species	Type straind	No. of clinical isolatesa	Size of amplicon (bp)		Species	Type straind	No. of clinical isolatesa	Size of amplicon (bp)
Mycobacteria
M. abscessus	CAP97E-03T	8b	136
M. africanum	ATCC 25420T		235
M. asiaticum	ATCC 25276T		136
M. aurum	ATCC 23366T		136
M. avium	ATCC 25291T	21	136
M. bovis	ATCC 19210T	10	235
M. bovis BCG	French strain 1173P2T		235
M. celatum Type I	ATCC 51131T		136
M. celatum Type II	ATCC 51130T		136
M. chelonae	ATCC 35749T	4b	136
M. chitae	ATCC 19627T		136
M. fallax	ATCC 35219T		136
M. flavescens	ATCC 14474T		136
M. fortuitum	ATCC 6841T	18	136
M. fortuitum 49403	ATCC 49403		136
M. gastri	ATCC 15754T		136
M. genavense	ATCC 51233T		136
M. gordonae	ATCC 14470T	25	136
M. haemophilum	ATCC 29548T		136
M. interjectum	ATCC 51457T		136
M. intermedium	ATCC 51848T		136
M. intracellulare	ATCC 13950T	47	136
M. kansasii	ATCC 12478T	30	136
M. leprae	Thai53 strainT		136
M. malmoense	ATCC 29571T		136
M. marinum	ATCC 927T		136
M. neoaurum	ATCC 25795T		136
M. nonchromogenicum	ATCC 19530T		136
M. paratuberculosis	ATCC 19698T		136
M. phlei	ATCC 11758T		136
M. peregrinum	ATCC 14467T		136
M. scrofulaceum	ATCC 19981T	15	136
M. senegalense	ATCC 35796T		136
M. shimoidei	ATCC 27962T		136
M. simiae	ATCC 25275T		136
M. smegmatis	ATCC 19420T		136
M. szulgai	ATCC 35799T	18	136
M. terrae	ATCC 15755T		136
M. thermoresistibile	ATCC 19527T		136
M. triviale	ATCC 23292T		136
M. tuberculosis H37Rv	ATCC 27294T	183c	235
M. ulcerans	ATCC 19423T		136
M. vaccae	ATCC 15483T		136
M. xenopi	ATCC 19250T		136
Non-mycobacteria
Tsukamurella inchonensis	KCTC 9866T		136
Tsukamurella paurometabola	KCTC 9821T		136
Rhodococcus equi	IMSNU 20114T		136
Rhodococcus erythropolis	IMSNU 20115		136
Rhodococcus rhodochrous	IMSNU 20349T		136
Nocardia brasiliensis	KCTC 9136		136
Nocardia brevicatena	KCTC 9370		136
Nocardia globerula	KCTC 9001		136
Nocardia nova	IMSNU 21197T		136
Nocardia otitidiscaviarum	IMSNU 21221T		136
Corynebacterium diphtheriae	MSNU		NAe
Corynebacterium glutamicum	IMSNU 21196T		NA
Staphylococcus aureus	MSNU		NA
Bacillus subtilis	MSNU		NA
Escherichia coli	MSNU		NA
Neisseria meningitidis	MSNU		NA
Haemophilus influenzae	MSNU		NA

^aClinical isolates were identified by biochemical tests and analysis of partial 16S rDNA sequences.

^bDifferentiation of M. abscessus and M. chelonae was performed by rpoB (306-bp) sequence analysis (13).

^cComposed of 40 rifampin-resistant and 143 rifampin-susceptible strains.

^dATCC: American Type Culture Collection; KCTC: Korean Collection for Type Cultures; MSNU: Department of Microbiology, Seoul National University College of Medicine; IMSNU: Institute of Microbiology, Seoul National University. Superscript T indicates type reference strain. There were a total of 379 clinical isolates.

^eNA, not amplified.

Two different kinds of DNA extraction protocols were used for type strains and clinical isolates. The DNAs of the reference strains were purified by using the bead beater-phenol extraction method as previously described (12, 13). The DNAs of clinical isolates were prepared by the boiling method without a purification step to reduce the DNA preparation time and to minimize the risk of cross contamination (1). A loopful culture of each sample was suspended in a screw-cap tube with 50 μl of TEN buffer (10 mM Tris-HCl, 1 mM EDTA, 100 mM NaCl [pH 8.0]), and the tube contents were then boiled on a hot plate for 10 min. Without further purification, 5 μl of the boiled mycobacterial suspension was used directly as a template for the DPCR assay.

For construction of the two DPCR primer sets, rpoB sequences of 44 mycobacteria (GenBank accession no. AF057449 to AF057493) were aligned by using the multiple alignment algorithm in the MegAlign package (Windows, version 3.12e; DNASTAR, Madison, Wis.). The specific nucleotides of the M. tuberculosis or NTM strains were located at the 3′-hydroxyl end of each primer, as previously reported (12). The developed Tbc1 (5′-CGT ACG GTC GGC GAG CTG ATC CAA-3′)-TbcR5 (5′-C CAC CAG TCG GCG CTT GTG GGT CAA-3′) and M5 (5′-G GAG CGG ATG ACC ACC CAG GAC GTC-3′)-RM3 (5′-CAG CGG GTT GTT CTG GTC CAT GAA C-3′) primer sets amplify a 235-bp DNA sequence from the M. tuberculosis complex and a 136-bp DNA sequence from NTM, respectively (Fig. 1A). Primers (10 pmol of Tbc1-TbR5 and 20 pmol of M5-RM3) and 5 μl of bacterial DNA were added to a PCR mixture tube (AccuPower PCR PreMix; Bioneer, Daejeon, Korea) containing 2 U of Taq polymerase, 250 μM each deoxynucleoside triphosphate, 10 mM Tris-HCl (pH 8.3), and 1.5 mM MgCl₂, and water was added to a final volume of 20 μl per reaction mixture. PCR was performed with an initial denaturation of 95°C for 5 min, 30 cycles of amplification (30 s at 95°C, 60 s at 72°C), and a final elongation at 72°C for 5 min (model 9600 thermocycler; Perkin-Elmer Cetus). Denaturation was extended to 15 min for the clinical isolates. Escherichia coli DNA was used as a negative control. The PCR products were analyzed by agarose gel (1.5%) electrophoresis.

FIG. 1.

Locations (A) and sequences (B) of the DPCR primers on the rpoB gene. DPCR products amplified by an M. tuberculosis-specific primer set (Tbc1-TbcR5) and an NTM-specific primer set (M5-RM3) were 235 and 136 bp long, respectively. rif^r indicates the region associated with the rifampin resistance of M. tuberculosis. Nucleotides of four mycobacteria differing from those of the primers are shown in bold letters. Numbers indicate the amino acid position of the RNA polymerase β subunit of E. coli. Note that TbcR5 and RM3 are displayed in reverse direction. TB, M. tuberculosis H37Rv (GenBank accession no. AF057454); AVI, M. avium (GenBank accession no. AF057457); INT, M. intracellulare (GenBank accession no. AF057472); KAN, M. kansasii (GenBank accession no. AF057473).

After completion of the DPCR assay, all 186 NTM isolates showing a 136-bp DNA amplicon were further analyzed by restriction fragment length polymorphism (RFLP) and by direct sequencing for species identification. On the basis of the rpoB sequences of mycobacteria (GenBank accession no. AF057449 to AF057493), two restriction enzymes, MspI and HaeIII were selected by using MapDraw (version 3.14; DNASTAR). Two enzymes, MspI (TaKaRa, Shiga, Japan) and HaeIII (TaKaRa) were independently applied to the PCR products. Ten microliters of the PCR products, 2 U of each enzyme, and restriction buffer were transferred to a fresh microcentrifuge tube, and water was added to a final volume of 20 μl per reaction mixture. Digestion was performed for 2 h at 37°C. Following digestion, the mixtures were electrophoresed in a 3% agarose gel.

NTM isolates were separately identified by determining 87-bp amplicon sequences, which excluded the primer sequences, and by comparing these with sequences in the GenBank database. Sequencing reactions with primers M5 and RM3 were performed as previously described (13). Analysis of the partial 16S rDNA sequence was performed separately to identify the NTM strains, and the results were compared with the results of the rpoB sequence analysis. Briefly, 16S rDNA fragments were amplified by using forward primer 285 and reverse primer 264 and then directly sequenced by using sequencing primer 244, as previously described (14).

The specificity of each primer set was assessed by a separate PCR. When a PCR using each specific primer set (Tbc1-TbcR5 or M5-RM3) was applied to the 44 reference strains at the same annealing temperature, an amplicon of either 235 or 136 bp of DNA was observed from the M. tuberculosis complex or NTM, respectively. When this method was applied to 17 non-mycobacteria, only 10 strains of Tuskamurella, Rhodococcus, and Nocardia species, which are phylogenetically close to mycobacteria, produced a 136-bp DNA amplicon (Table 1). Finally, a DPCR assay with a mixture of the two primer sets was performed on the reference strains. While the 235-bp DNAs were amplified from only four strains of M. tuberculosis complex, the 136-bp DNAs were amplified from all of the 40 NTM strains (Fig. 2). Nothing was amplified from the negative control, E. coli DNA. Therefore, DPCR assay allowed the differential identification of M. tuberculosis complex and NTM in a single reaction. Despite the mismatching of several nucleotides with those of the M5 or RM3 primer in the rpoB sequences of NTM (Fig. 1B), all of the NTM strains tested produced 136-bp amplicons.

FIG. 2.

DPCR assay results performed with the reference strains of M. tuberculosis complex and NTM. Two amplicons of different sizes (235 and 136 bp) were amplified from M. tuberculosis complex (lanes 1 to 4) and NTM (lanes 5 to 14) strains by a single DPCR. Lanes: M, marker DNA (50-bp ladder); 1, M. tuberculosis; 2, M. africanum; 3, M. bovis; 4, M. bovis BCG; 5, M. avium; 6, M. paratuberculosis; 7, M. scrofulaceum; 8, M. intracellulare; 9, M. terrae; 10, M. nonchromogenicum; 11, M. triviale; 12, M. gordonae; 13, M. asiaticum; 14, M. thermoresistibile; 15, E. coli (negative control).

In order to determine the lower limit of the sensitivity of the DPCR method, rpoB DNAs were amplified from the serially diluted DNAs (10 ng to 10 fg) of M. tuberculosis and M. avium. The amplification product was obtained from 10 pg of M. tuberculosis DNA, whereas in the case of M. avium DNA, as little as 1 pg was detected on an ethidium bromide-stained gel (data not shown). Taking 5 fg of DNA as a mycobacterial cell equivalent (5), this method could detect 2,000 bacillus equivalents.

The usefulness of the DPCR assay was demonstrated by applying the technique to the identification of 379 clinical isolates. Of the 379 culture isolates examined, 193 strains were identified as M. tuberculosis complex and the other 186 strains were identified as NTM (Table 1). These results were completely concordant with those obtained by conventional culture testing (Table 2). Although primer TbcR5 is located in the rif^r region (20) (Fig. 1A), in which mutations are related to the rifampin resistance of M. tuberculosis, all of the 40 rifampin-resistant M. tuberculosis isolates produced 235-bp amplicons without PCR interference.

TABLE 2.

Differentiation of M. tuberculosis complex and NTM clinical isolates by DPCR and conventional biochemical testing

DPCR product (size [bp])	No. of isolates identifieda as:
	M. tuberculosis (n = 193)	NTM (n = 186)
M. tuberculosis specific (235)	193	0
NTM specific (136)	0	186

^aIdentification was separately performed by niacin and heat-stable catalase tests (pH 7.0, 68°C).

Since DPCR produces only one (136-bp) rpoB DNA from NTM, restriction analysis and sequencing could be used for further species identification. An algorithm for species identification by DPCR-linked restriction analysis with MspI and HaeIII was developed (Fig. 3). All of the 186 NTM isolates analyzed in this study were tentatively identified in accordance with this scheme. However, NTM isolates, except M. abscessus and M. chelonae, which have identical sequences in the 87-bp rpoB region, could be exactly identified by comparing their determined rpoB sequences to those of reference strains. These results are concordant with those obtained by 16S rDNA analysis and culture-based biochemical testing.

FIG. 3.

Algorithm for species identification by DPCR-RFLP (136-bp amplicons) with MspI and HaeIII.

In the present study, we developed a novel DPCR assay based on rpoB sequences. Unlike previous methods, two differently sized DNAs were amplified from a single target gene of M. tuberculosis and NTM. The advantages of this DPCR assay over other multiplex PCRs are as follows. (i) rpoB nucleotides specific for M. tuberculosis complex or NTM are invariably constant. Therefore, false-positive or -negative results due to sequence variations do not occur. (ii) DPCR yields only one product, irrespective of the Mycobacterium species tested. Even the coexistence of M. tuberculosis and NTM can be detected by the presence of two different PCR products in a single reaction mixture. (iii) Moreover, the (136-bp) rpoB DNAs of NTM can be further analyzed by RFLP or by direct sequencing to supplement species identification.

In conclusion, the DPCR assay based on rpoB provides a rapid and reliable means for the differential identification of M. tuberculosis and NTM in culture with a single reaction.