New Multiplex PCR for Rapid Detection of Isoniazid-Resistant Mycobacterium tuberculosis Clinical Isolates

Abstract

In this study, we describe a multiplex PCR to detect a AGC→ACC (serine to threonine) mutation in the katG gene and a −15 C-to-T substitution (inhA^C−15T) at the 5′ end of a presumed ribosome binding site in the promoter of the mabA-inhA operon. These mutations have been reported in the majority of previous studies as the most frequent mutations involved in the resistance to isoniazid (INH) of Mycobacterium tuberculosis clinical strains with high levels of resistance. The method was optimized and validated after an analysis of 30 M. tuberculosis clinical isolates with known sequences of the relevant part of the katG gene and the regulatory region of the mabA-inhA operon. We analyzed 297 INH-resistant M. tuberculosis isolates collected in Spain from 1996 to 2003 by PCR-restriction fragment length polymorphism (using the katG gene), DNA sequencing, and the newly developed multiplex PCR. The results were concordant for all 297 isolates tested. The analysis revealed that 204 (68.7%) of the isolates carried one or both of the mutations. This finding suggests that with further development this multiplex PCR will be able to detect the majority of the INH-resistant M. tuberculosis clinical isolates from Spain and other countries where a high frequency of similar mutations occur.

Currently, throughout the world, isoniazid (INH) and rifampin (RIF) together represent the backbone of short-course chemotherapy for Mycobacterium tuberculosis infections. The number of multidrug-resistant strains of M. tuberculosis, defined as resistant to INH and RIF, has been increasing over the years, and several outbreaks have been reported (5, 9, 24). The development of resistance to these two drugs reduces the efficacy of standard antituberculosis (anti-TB) treatment to 77%. It is very important, therefore, to identify these strains as soon as possible to allow for adjustments in treatment and to minimize the transmission of drug-resistant strains. Phenotypic drug susceptibility testing by conventional methods on solid media (6, 8) requires 10 to 30 days after the primary culture has been isolated. This time can be reduced by the use of rapid methods such as BACTEC, which requires 5 to 10 days.

Resistance to RIF has been shown to be caused by an alteration of the β subunit of RNA polymerase, which is encoded by the rpoB gene. More than 95% of RIF-resistant strains are associated with mutations within an 81-bp region of the rpoB gene (encoding the RNA polymerase β subunit). Specific mutations, insertions, and deletions have been described in several countries by several authors, and this 81-bp region has been termed the rifampin resistance determinant region (7, 14, 20, 26, 27, 30). Numerous methods exist to detect resistance to rifampin (10, 12, 18, 23).

In contrast, resistance to INH is more complicated, as mutations in several genes can lead to drug resistance. For most INH-resistant strains, mutations have been found in two genes, i.e., the katG gene, encoding catalase-peroxidase (31), and the mabA-inhA regulon (4), encoding a target of activated prodrug, enoyl-acyl carrier protein reductase (1-3, 11, 13, 15, 17, 21, 27). For some other INH-resistant strains, however, mutations in the ahpC promoter region (located in the 105-bp oxyR-ahpC intergenic region) or within the β-ketocyl acyl carrier protein synthase gene kasA have also been reported (19, 25). Most studies have examined the mutations present in these genes by DNA sequencing or analyses of a portion of the katG gene after PCR amplification and digestion with the restriction enzyme MspI or SatI (2, 3, 17).

Molecular methods have been developed to detect resistance to INH and RIF as an alternative to conventional tests because of their ability to provide results rapidly. Upon the elucidation of the genes involved in resistance to RIF and INH, several studies describing various PCR-based molecular genetic techniques for the detection of resistance were published (12).

In the present study, we report a simple, rapid, and inexpensive assay based on allele-specific PCR methodology targeting an AGC→ACC mutation in the katG gene and an inhA^C−15T mutation in the regulatory region of the mabA-inhA operon to detect INH-resistant M. tuberculosis strains and to identify the M. tuberculosis complex in the same PCR tube for each sample.

MATERIALS AND METHODS

Description of study samples. (i) Set one.

Thirty M. tuberculosis strains with known sequences for the relevant part of the katG gene and the regulatory region of the mabA-inhA operon served as samples for optimization of the PCR assay conditions. The strains had the following distribution of katG and mabA-inhA operon alleles: wild type, 6 strains; AGC→ACC mutation, 10 strains; inhA^C−15T mutation, 10 strains; and both the AGC→ACC and inhA^C−15T mutations, 4 strains.

(ii) Set two.

A total of 297 isolates collected from 1996 to 2003 in Spain and identified as INH-resistant M. tuberculosis strains by standard biochemical methods served to validate the developed multiplex PCR assay. Susceptibility testing with INH (0.2 and 1.0 μg/ml) was performed with Lowenstein-Jensen medium by the proportional method of Canetti et al. (6). A total of 50 M. tuberculosis strains that were susceptible to isoniazid were used as controls.

Extraction of DNAs for PCR.

DNAs were extracted from cultures grown in Lowenstein-Jensen medium. A loop of culture was suspended in 500 μl of distilled water and boiled for 5 min, followed by centrifugation for 5 min at 12,000 × g, and 5 μl of the supernatant was used as a source of genomic DNA for amplification of the katG gene and the mabA-inhA operon.

PCR-RFLP.

katG restriction fragment length polymorphism (RFLP) analysis to detect codon 315 mutations was performed by the use of previously described protocols and reaction components (23). A 620-bp region of the katG gene was amplified by use of the following primer set described by Uhl et al. (28): katG904, 5′-AGCTCGTATGGCACCGGAAC; and katG1523, 5′-TTGACCTCCCACCCGACTTG. After amplification, 10 μl of the product was digested with MspI (MBI Fermentas, Madrid, Spain) and analyzed by 2% MS8 agarose (8066; Pronadisa, Madrid, Spain) gel electrophoresis. Ten microliters of the same katG-specific PCR product was digested with SatI (MBI Fermentas) and analyzed in the same run with products generated by restriction with MspI (Fig. 1). SatI contains an overlapping GC sequence at a single recognition site of the triplet AGC at position 315, and any mutation at this position eliminates cleavage of the katG fragment.

FIG. 1.

Schematic presentation of multiplex allele-specific PCR to detect the AGC→ACC mutation in codon 315 of the katG gene (A) and the −15 C-to-T substitution (inhA^C−15T) at the 5′ end of a presumed ribosome binding site in the promoter of the mabA-inhA operon (B). Short arrows indicate the primers, and long double-headed arrows depict the resulting PCR fragments. The targeted sequences are shown in shaded boxes. The mutated bases are underlined and in bold, and the mutated bases are underlined.

PCR and DNA sequencing.

For PCRs and DNA sequencing, specially designed primers (for regions in the katG gene and the regulatory region of the mabA-inhA operon that are known to confer resistance to M. tuberculosis) were used to amplify genomic DNAs. These primers were designed from a published sequence of the M. tuberculosis H37Rv genome (GenBank) by the use of PCGENE software (IntelliGenetics, Mountain View, Calif.).

PCRs were performed with a pureTaq Ready-To-Go PCR bead kit (Amersham Biosciences, Piscataway, N.J.). PCRs were performed in a 9600 thermocycler (Perkin-Elmer) under the same PCR conditions for both katG and mabA-inhA (an initial denaturation step of 5 min at 94°C, followed by 30 cycles of 1 min at 94°C, 1 min at 65°C, and 90 s at 72°C and ending with a final elongation step of 10 min at 72°C). Efficient PCR amplifications were confirmed by gel electrophoresis in 0.7% MS8 agarose gels (8066; Pronadisa). PCR products were purified by use of a GFX PCR DNA and gel band purification kit (Amersham Biosciences) and were sequenced by use of a Big Dye Terminator cycle sequencing ready reaction kit (Applied Biosystems Inc., Foster City, Calif.) in an ABI Prism 377 automated DNA sequencer (Applied Biosystems). The following thermocycler parameters were used: denaturation at 94°C for 3 min, followed by 25 cycles of denaturation at 96°C for 10 s and primer annealing at 60°C for 4 min. The primer sequences are shown in Table 1.

TABLE 1.

Primer sequences

Primer target or type	Primer designation	Sequence
Sequencing primers
katG	katGF	GTGCCCGAGCAACACCCACC
	katGR	CGCACGTCGAACCTGTCGAGG
mabA-inhA	mabAF	CGAAGTGTGCTGAGTCACACCG
	inhAR	CCCCACCGAAATGCAGGTCG
Multiplex PCR primers
gyrB	MTUBf	TCGGACGCGTATGCGATATC
	MTUBr	ACATACAGTTCGGACTTGCG
katG	katg0F	GCAGATGGGGCTGATCTACG
	R315mut	TCCATACGACCTCGATGCCAG
mabA-inhA	mabAF	CGAAGTGTGCTGAGTCACACCG
	inhARmut	AGTCACCCCGACAACCTATTA

All sequence data were assembled and edited electronically with the EDITSEQ, ALIGN, and MEGALIGN programs (DNASTAR, Madison, Wis.) and were compared with the published sequences for the katG gene and the mabA-inhA operon (GenBank accession number NC_000962).

Multiplex PCR.

Specially designed primers were used to detect the AGC→ACC and inhA^C−15T mutations. The reverse primer R315mut was positioned so that its 3′-OH end paired with two changed bases, i.e., a GC-to-CT substitution (Fig. 1A). Consequently, in the presence of the AGC→ACC mutation at this position in katG, a 296-bp fragment was amplified by the forward primer katg0F (described previously by Mokrousov et al. [22]). On the other hand, the reverse primer inhARmut was positioned so that its 3′-OH end paired with two changed bases, a −15 C-to-T substitution (inhA^C−15T) and a −14 G-to-A substitution at the 5′ end of a presumed ribosome binding site in the promoter of the mabA-inhA operon (Fig. 1B). If the inhA^C−15T mutation occurred, a 146-bp fragment was amplified by the forward primer mabAF. In the absence of a mutation, the result would be two mismatches at the 3′ end of the reverse primers (R315mut and inhARmut). Thus, under appropriate PCR conditions, resistant strains with these two mutations resulted in 296-bp and/or 146-bp PCR products. The MTUBf and MTUBr primers (previously reported by Kasai et al. [16]) produced a 1,020-bp fragment by amplification of a partial sequence of the gyrB gene. This fragment was the individual positive control and needed to be present in every sample. The absence of this amplification product would show that the PCR was inhibited and that the result should be ignored.

The PCR mix consisted of 10 mM Tris-HCl (pH 9), 50 mM KCl, 1.5 mM MgCl₂, a 200 μM concentration of each deoxynucleoside triphosphate, 2.5 U of pureTaq DNA polymerase and reaction buffer, which were included in the pureTaq Ready-To-Go PCR bead kit (Amersham Biosciences), and PCR primers to a final volume of 25 μl. The concentrations of the primers were 200 mM for katG0F and 400 mM for R315mut (to detect the AGC→ACC mutation), 400 mM for both mabAF and inhARmut (to detect the inhA^C−15T mutation), and 400 mM for both MTUBf and MTUBr. All primer sequences are shown in Table 1.

For PCRs, a GenAmp PCR system 9600 thermocycler (Applied Biosystems) was used, beginning with a 5-min denaturation at 95°C and followed by 30 cycles of 1 min at 95°C, 1 min at 68°C, and 45 s at 72°C. The final cycle was followed by an extension time of 10 min at 72°C. Efficient PCR amplifications and the sizes of the products were estimated after electrophoresis in 2% agarose gels by staining with ethidium bromide.

RESULTS AND DISCUSSION

An analysis of 297 INH-resistant and 50 INH-susceptible M. tuberculosis clinical isolates from TB cases by katG PCR-RFLP and automated DNA sequencing showed that none of the 50 INH-susceptible strains had the AGC→ACC mutation (S315T) at codon 315 of katG or the less frequent mutations in that codon. The AGC→ACC missense mutation was observed in 130 (43.8%) INH-resistant strains by PCR-RFLP with an MspI endonuclease treatment (Fig. 2, lanes 2 and 3), and 17 INH-resistant strains showed a different mutation in codon 315 by PCR-RFLP with a SatI endonuclease treatment (Fig. 2, lanes 10 and 11). The use of the MspI endonuclease only detected the AGC→ACC mutation, whereas SatI was able to detect any mutation in codon 315 that affected the GC nucleotides at that position. Of the 17 strains with different mutations, 12 (4.0%) showed an AGC→AAC mutation at codon 315 of katG and 5 (1.7%) showed an AGC→ACA mutation by automated DNA sequencing. The results are summarized in Table 2.

FIG. 2.

Profiles generated by PCR-restriction digestion with MspI and SatI endonucleases. Lanes 1 to 6, strains digested with MspI; lane 1, wild-type strain; lane 2, AGC→ACC mutation (S315T); lane 3, S315T mutation and R463L polymorphism; lane 4, AGC→AAC mutation at katG codon 315; lane 5, AGC→ACA mutation at katG codon 315; lane 6, wild type; lanes 7 to 12, the same strains digested with the SatI endonuclease; M, 50-bp DNA ladder (Amersham Biosciences).

TABLE 2.

Comparison of INH susceptibility test results obtained by the different methods used in this study

No. of strains	Susceptibility of strains by indicated methoda
	Proportional methodb	PCR-RFLP		DNA sequencingc	Multiplex PCR
		MspI	SatI
130	R	AGC > ACC	R	AGC > ACC	AGC > ACC
12	R	S	R	AGC > AAC	S
5	R	S	R	AGC > ACA	S
70	R	S	S	inhAC−15T	inhAC−15T
4	R	AGC > ACC	R	AGC > ACC + inhAC−15T	AGC > ACC + inhAC−15T
76	R	S	S	S	S
50	S	S	S	S	S

^aR, resistant; S, susceptible. When indicated, mutations were detected by the various methods.

^bPerformed on Lowenstein-Jensen medium.

^cDNA sequencing of the katG gene and the promoter of the mabA-inhA operon.

The inhA^C−15T mutation was found in 70 (23.6%) INH-resistant isolates, with 4 of these strains showing the AGC→ACC mutation as detected by PCR-RFLP, and in none of the 50 INH-susceptible strains (Table 2).

The multiplex PCR that we developed was evaluated for its detection of the AGC→ACC mutation in codon 315 of the katG gene and the inhA^C−15T mutation in the promoter of the mabA-inhA operon in the same PCR tube for each sample. The 130 INH-resistant strains with the AGC→ACC mutation produced a 296-bp band (Fig. 3, lanes 1 to 4), while the 70 strains with the inhA^C−15T mutation produced a 146-bp band (Fig. 3, lanes 5 to 12). The four strains with both mutations produced two bands, one at 296 bp and the other at 146 bp (Fig. 3, lanes 13 and 14). All of the strains, including the susceptible strains (Fig. 2, lanes 16 to 18), produced a 1,020-bp band by amplification of a partial sequence of the gyrB gene. The multiplex PCR results were concordant with those generated by PCR-RFLP analysis and DNA sequencing for all of the strains tested (Table 2). The specificity of the multiplex PCR for the detection of INH resistance in resistant strains carrying the AGC→ACC mutation at codon 315 of katG and/or the inhA^C−15T mutation in our setting was 100%. The sensitivity of the multiplex PCR for the detection of INH resistance depended on the prevalence of these two mutations among the INH-resistant strains in each geographic area. In our setting, the sensitivity was 68.7%. The negative predictive value was 35%, and the positive predictive value was 100%. In Spain, the prevalence rates of the AGC→ACC allele at codon 315 of katG and of the inhA^C−15T mutation are 47.2 and 25%, respectively (unpublished data), whilst other mutations in codon 315 are found in 6.6% of INH-resistant strains. Consequently, 72.2% of the INH-resistant strains in our country could be detected with the multiplex PCR.

FIG. 3.

Profiles generated by multiplex PCR assay. Lanes 1 to 4, strains with the AGC→ACC mutation at katG codon 315; lanes 5 to 13, strains with the −15C→T substitution in the promoter of the mabA-inhA operon; lanes 14 to 15, strains with both mutations; lanes 16 to 18, INH-susceptible strains; M, 100-bp DNA ladder (Bio Tools).

Mutations at the Ser315 codon of katG and at the regulatory region of the mabA-inhA operon occur most frequently in INH-resistant isolates (1-3, 11, 13, 15, 17, 21, 27). These mutations, AGC→ACC and inhA^C−15T, respectively, may be responsible for 70% of INH-resistant M. tuberculosis cases.

The prevalence of mutations in codon 315 of katG varies depending on the geographical region studied, with percentages ranging from 35% in Beirut (2) to 91% in Latvia and Russia (21, 27), while the prevalence of the inhA^C−15T mutation varied from 32.7% in a study by Bakonyte et al. (3) to 3.3% in a study by Van Rie et al. (29), with several studied strains being similar to those in Bakonyte et al.'s study. Torres et al. (26) reported that this mutation was found with a frequency of 4.3% in strains isolated from Seville, Spain, but the present study shows a much higher frequency of 23.6%. The frequency of other mutations in codon 315 (5.7%) was found to be lower than those in other studies (1, 11).

The high prevalence of the AGC→ACC mutation in the katG gene and of the inhA^C−15T mutation in the promoter of the mabA-inhA operon in the strains isolated from Spain suggests that the multiplex PCR technique can be used as a single, rapid tool for detecting INH resistance in M. tuberculosis clinical strains with a high probability. The assay is easy to perform and interpret and could be implemented as part of the routine practices of clinical laboratories in areas with a high prevalence of multidrug-resistant TB strains. For easier and unambiguous interpretations of multiplex PCR profiles, each run should include a wild-type strain (as a positive control for no amplification due to mutation) and one strain with a known mutation (as a positive control of amplification due to mutation). Furthermore, the procedure is inexpensive and requires only standard PCR and electrophoresis equipment. However, all negative results should be confirmed by conventional methods based on cell culture.