Introduction

Tuberculosis (TB) remains one of the major causes of death around the world, especially due to the emergence of multidrug-resistant TB (MDR-TB) and extensively drug-resistant TB (XDR-TB). There were approximately 440 000 cases of MDR-TB in 2008, which were resistant to both rifampicin (RIF) and isoniazid (INH). Almost 50.0% of MDR-TB cases worldwide are estimated to occur in China and India.1 Approximately 8.0% of the patients with MDR-TB have XDR-TB, which is defined as resistance to at least RIF and INH, plus resistance to any fluoroquinolone and at least one of three injectable anti-TB drugs (capreomycin, kanamycin or amikacin).2 Therefore, controlling tuberculosis caused by drug-resistant M. tuberculosis has become an urgent public health problem in many regions of the world, particularly in developing countries.

RIF and INH are the principal first-line drugs used in combination in tuberculosis chemotherapy. Therefore, it is important to understand the molecular basis for this resistance.3 RIF, as a broad spectrum derivative of RIF, exerts its anti-tuberculosis role by binding to the bacterial RNA polymerase, thereby inhibiting RNA synthesis. Ninety-five percent of the resistance to RIF in M. tuberculosis is due to single mutation in an 81-bp core region, termed the RIF resistance-determining region (RRDR) of the rpoB gene, which encodes the beta subunit of DNA-dependent RNA polymerase.4 INH is the first-line drug responsible for the initial dramatic decrease in actively metabolizing bacilli during treatment of TB, yet its mechanism of resistance remains only partially understood. Unlike RIF, INH resistance is associated with mutations in multiple loci, such as the catalase-peroxidase gene (katG), the enoyl-ACP reductase gene (inhA) and its promoter, the alkyl hydroperoxide reductase gene (ahpC), and the intergenic region between the oxyR and ahpC genes.5, 6

Traditional culture-based methods of testing for drug resistance can take at least 4 weeks, as M. tuberculosis is a slow-growing organism, leading to higher mortality and the further spread of MDR infections. Thus, various molecular methods have been used to detect MDR-TB by identifying mutations associated with drug resistance.7, 8 Among these methods, DNA sequencing is the most direct and reliable, and can rapidly detect both known and novel mutations.

It has been reported that some mutations have a geographical distribution.8, 9, 10 However, little data have been available on the molecular characterization of multidrug-resistant clinical isolates of M. tuberculosis from South-central region in China. In this study, the type and frequency of mutations in the rpoB, katG, mabA-inhA regulatory region, the structural gene, ahpC, and the oxyR-ahpC intervening region were investigated. The MICs of RIF and INH for MDR strains were determined. These studies identified genetic markers of MDR M. tuberculosis isolates that are unique to this geographical region and will further the understanding of the molecular mechanism of drug resistance, particularly multi-drug resistance in M. tuberculosis. Currently, the mechanism of multidrug resistance in M. tuberculosis still needs to be elucided. Based on the analysis of the relationship between the gene typing and MIC, the findings of present study further reveal the molecular mechanism of MDR-TB strains in China. Also, the results help to supplement and optimize the current molecular detection techniques, which assist in the personalized treatment for the MDR-TB patients. In addition, the molecular characterization identified recent transmission patterns of multidrug-resistant isolates of the Beijing genotype in South-central China.

Large numbers of reports have confirmed that the Beijing genotype strains had a deletion in RD105, which was a unique characteristics of Beijing genotypes isolates. Therefore, the deletion of RD105 sequence can serve as a simple and rapid method to identify Beijing genotypes strains. Analyzing the distribution and transmission of MTB in this area can provide a theoretical basis for controlling and preventing infectious diseases in Hubei Province.11, 12

Materials and Methods

Clinical isolates and drug susceptibility testing (DST)

A total of 1733 clinical isolates of M. tuberculosis was obtained from 5299 patients with pulmonary tuberculosis in the Wuhan Medical Treatment Center, Hubei, South-central China, from August 2009 to July 2010. All isolates were routinely cultured with egg-based Löwenstein-Jensen medium at 37 °C and were evaluated by proportion method for DST.13

Proportion method was performed on DSTs use the following drug concentrations in the modified Löwenstein-Jensen medium: RIF, 40.0 μg ml−1; INH, 0.2 μg ml−1; ethambutol, 2.0 μg ml−1; streptomycin, 4 μg ml−1. In addition, some second-line drugs, widely used in China, were also performed on DSTs with the following concentrations: kanamycin, 30.0 μg ml−1; paramycin, 1.0 μg ml−1; protionamide (Pto), 40.0 μg ml−1; ofloxacin, 2.0 μg ml−1; thiophen-2-carboxylic acid hydrazine (TCH), 5.0 μg ml−1; para-nitrobenzoic acid (PNB), 500 μg ml−1.

The sodium 2,3-bi(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide inner salt (XTT) colorimetric method14 was performed in 7H9-S medium containing Middlebrook 7H9 broth, 10% oleic acid albumin dextrose catalase and 0.5% glycero and 0.1% Casitone. By using 96-well plates, the MIC was determined with 7H9-S medium containing twofold dilution series of the RIF ranging from 0.125 to 8 μg ml−1, the INH ranging from 0.063 to 4 μg ml−1.

Sixty-five clinically resistant isolates, including 60 MDR strains and 7 extensively drug resistant (XDR) strains, were selected from 664 (38.3%, 664/1733) drug-resistant strains, and 20 pan-susceptible stains were also randomly chosen as control (Table 1).

Table 1 Prevalence of drug-resistance phenotypes among clinical M. tuberculosis isolates (n=85)
Full size table

DNA extraction and isolates identified

Genomic DNA of M. tuberculosis isolates was extracted as previously described.15 M. tuberculosis culture isolates previously identified by conventional biochemical methods were confirmed by 16S rRNA gene sequencing.16

PCR amplification and sequencing

Oligonucleotide primers were designed from the M. tuberculosis H37Rv genome sequence17 using Primer Premier 5.0 (PREMIER Biosoft, Palo Alto, CA, USA) and Oligo 6 software (Molecular Biology Insights, Inc., Cascade, CO, USA). The uniqueness of the sequences of the primers designed was analyzed with the Blast search (http://www.ncbi.nlm.nih.gov). The nucleotide sequences for the mutant genes obtained in the present study were deposited in the GenBank database (Table 2). Primers were synthesized by Invitrogen Bio Co. (Shanghai, China). Six fragments were amplified and sequenced: rpoB (including the 81-bp RRDR), katG (including codon 315), inhA, ahpC gene, mabA-inhA regulatory region and the oxyR-ahpC intergenic region. Sequence data were assembled and analyzed by CLUSTAL W software (European Bioinformatics Institute). PCR amplification was performed with the EDC-810 model thermal cycler (Eastwin Life Sciences, Inc., Beijing, China), and the Pfu Taq DNA polymerase with high fidelity was used to avoid false priming. The amplicons were purified with the E.Z.N.A. MicroElute Cycle-Pure Kit (Omega, Lilburn, GA, USA). Sequence analysis was determined on both strands by direct sequencing of the PCR products on an automated model 377 DNA sequencer (Applied Biosystems, Foster City, CA, USA) with fluorescence-labeled dideoxynucleotide terminators (ABI Prism Big Dye terminator cycle sequencing ready reaction kit; Sangaon Biotech Co., Ltd, Shanghai, China).

Table 2 Primers used in this study and resulting PCR products
Full size table

Genotyping method

The identification of Beijing genotypes strains was performed using the RD105 deletion-targeted amplified by PCR.18 Primers were designed flanking this 3467 base pair (bp) deletion polymorphism P1 and P2 to detect the presence or absence of this deletion in the M. tuberculosis genome. A third primer P3, which is internal to deletion RD105, was also utilized. A χ2 test was performed with SPSS17.0 software (SPSS China, Shanghai, China) to analyze the distribution characteristics of Beijing genotypes.

Results

All 85 strains chosen for analysis were confirmed as M. tuberculosis by 16S rRNA gene sequencing. Specific resistant genes or regulatory regions were sequenced, and the sequencing results were compared with the DST results. No mutation was found in the pan-susceptible isolates.

Mutation patterns in MDR and XDR isolates

Forty-nine of the sixty MDR were found to have a mutation in the six characterized genetic regions. Forty-one MDR isolates (68.3%,41/60) were found to carry mutations in the rpoB gene and katG gene, and 21 isolates (35.0%, 21/60) contained mutations in both rpoB531 and katG315. Sequencing of three MDR isolates failed to identify any mutations in the six regions investigated. Therefore, these six genetic regions account for 95.0% (57/60) of the MDR isolates characterized.

Of the 49 MDR isolates, 44 (89.8%) were Beijing genotypes and the other 5 were non-Beijing genotypes. Among the 36 non-MDR, 34 (94.4%) were Beijing genotypes and 2 isolates were non-Beijing genotypes (χ2=0.593, P=0.441).

Among the seven XDR isolates, four were found to possess alterations in rpoB 531 and katG 315, and one of them had a third mutation in the mabA-inhA regulatory region (C-15T; Table 3). Two of the remaining three XDR isolates carried mutations in three fragments, one had mutations in rpoB H526L (CAC→CTC), katG S315T (AGC→ACC) and inhA A190S (GCC→TCC); and the other isolate had a mutation in rpoB S531L (TCG→TTG) with a seven amino acids (GKTHGAG) insertion mutation in katG 279-280, plus base substitutions C-15T in mabA-inhA regulatory region. The last isolate harbored mutations in both rpoB Q517(CAG deletion) and mabA-inhA regulatory region C-15T. All of the seven XDR were non-Beijing genotypes.

Table 3 Mutations in genes associated with INH and RIF of the XDR-TB clinical isolates (n=7)
Full size table

RIF resistances and rpoB gene

Of the 60 RIF resistance isolates, 55 (91.7%, 55/60) carried mutations in RRDR of the rpoB gene, 45 isolates had a single-point mutation, 8 had double mutations (13.3%, 8/60), 2 had a deletion that was the same in both isolates and 5 have no mutation (8.3%, 5/60, MIC>2–8 μg ml−1; Table 4). The most frequently mutated rpoB codons were codon 531, 526 and 516, with mutation frequencies of 51.7% (31/60, MIC⩾2–8 μg ml−1), 18.3% (11/60, MIC>2–8 μg ml−1) and 10.0% (6/60, MIC>2–8 μg ml−1), respectively. Consistent with previous reports, the mutation S531L was the most common mutation in rpoB gene. Codon 526 harbored the most variable mutations. Six isolates exhibited codon 516 mutations, including D516V and D516G. Eight isolates with seven different types of double mutations were identified. In addition, one deletion was detected at codon 517 (CGC, MIC⩾4 μg ml−1) in two stains. Of these mutations, three have not previously been described, E458A (GCG→GAG; GenBank: JF812083), S509R (AGC→AGA) and P535S (CCC→TCC). Five isolates (8.3%) had no mutation in the RRDR of rpoB gene. There are 55 high-level (MIC⩾4–8 μg ml−1) and 5 low-level (MIC⩽0.5–4 μg ml−1) RIF resistance strains among 60 MDRs.

Table 4 Distrbution of mutations in the rpoB gene among MDR-TB clinical isolates (n=60)
Full size table

Of the 55 RIF mutant strains, 51 (92.7%) were Beijing genotypes and 4 (7.3%) were non-Beijing genotypes. Besides, among the 30 RIF-susceptible strains, 27 (90.0%) were Beijing genotypes and 3 (10.0%) isolates were non-Beijing genotypes (χ2=0.191, P=0.662).

INH resistances and katG, inhA, ahpC mabA-inhA and oxyR-ahpC

In 84.4% (54/64) of the INH resistances strains, at least one mutation was found in the katG, inhA, ahpC, mabA-inhA regulatory region and oxyR-ahpC intergenic region, which included mutations at katG codon 315 in isolates. And no mutation was found in 10 MDR isolates (15.6%, MIC⩾2–4 μg ml−1; Table 5). Of the 64 INH resistances, 70.3% (45/64) had genetic alterations in katG nucleotide fragment. Mutations at codon 315 in the katG region contributed the majority of INH resistance, with high frequency of 62.5% (40/64, MIC>1–4 μg ml−1) isolates, and the common base substitution mutation S315T (AGC→ACC) was the most frequently alteration. Three isolates carried other exchanges in codon 315. In addition, we found three novel types of mutation in katG gene, G273S (GGT→AGT, MIC>4 μg ml−1; GenBank: HQ913569), I266T (ATC→ACC, MIC=4 μg ml−1; GenBank: HQ913567) and P232S (CCG→TCG, MIC=4 μg ml−1; GenBank: HQ913568). Interestingly, a 10.8-kb fragment deletion in WH56 isolate comprising the katG with seven upstream genes and one downstream gene was found in one MDR isolate (Figure 1). In addition, one novel insert mutation was identified in the katG gene, resulting in a seven-amino-acid, GKTHGAG, insert between codons 279 and 280 (MIC⩾4 μg ml−1).

Table 5 Mutations in the genes of katG, inhA, ahpC mabA-inhA and oxyR-ahpC INH-resistant isolates (n=64)
Full size table
Figure 1
figure 1

A 10.8-kb gene deletion fragment in WH56 isolates. White box represents for the genes deleted.

Full size image

Mutations in inhA and ahpC with their noncoding regions were identified in 21.9% (14/64) isolates. No alteration was found in the inhA structural gene, whereas one mutation was detected in ahpC structural gene: A190S (GCC→TCC). In contrast, 12.5% (8/64) of the INH-resistant isolates harbored base substitutions C-15T in the mabA-inhA regulatory region. Three of them had an additional mutation in the katG gene. 9.4% (6/64) isolates had nucleotide substitutions at loci of −46,−52 and -54 in the oxyR-ahpC intergenic region, and two of isolates had an additional mutation in katG gene. There are 56 high-level (MIC⩾2–4 μg ml−1) and 8 low-level (MIC⩽0.25–2 μg ml−1) INH resistance strains among 64 MDRs.

Of the 64 INH mutant strains, 59 (92.2%) were Beijing genotypes and 5 (7.8%) were non-Beijing genotypes. Among the 30 INH-susceptible strains, 19 (90.5%) were Beijing genotypes and 2 (9.5%) isolates were non-Beijing genotypes (χ2=0.061, P=0.804).

Genotyping of the Beijing genotypes strains

PCR was performed on the 85 clinical isolates of M. tuberculosis to identify the Beijing genotype. We obtained a 1466-bp PCR amplified product for the non-Beijing and H37Rv, and a 761-bp for the Beijing genotype (Figure 2). PCR on 78 (91.8%) obtained a 761-bp product, identifying a Beijing isolate and 7 strains (8.2%) did not generate the expected PCR products, identifying a non-Beijing isolate. The H37Rv strain was identified as a non-Beijing isolate (Table 1).

Figure 2
figure 2

RD105 amplified by PCR. Lane M: DL 2000 Marker; Lane 1–20: 20 clinical stains; Rv: H37Rv (non-Beijing family strain).

Full size image

Discussion

It has been reported that the frequency of rpoB mutations range from 76.0% (19/25)19 to 99.0% (102/103),20 suggesting that mutations in the rpoB gene are possible molecular markers for MDR-TB. In this study, the mutation frequency of rpoB gene was 91.7% (55/60) consistent with the observed resistance prevalence around the world.8, 21, 22, 23 The most frequently mutated rpoB codons were 531, 526 and 516 (MIC⩾2–8 μg ml−1), which have been reported elsewhere,10, 20, 21 indicating that there is no geographic difference in the main mutation types and mutation rates of rpoB gene in South-central China. Double mutations were observed in 8/60 (13.3%) isolates in this study, which is higher than reported in other studies from Shanghai, 7.4% (18/242),23 and Germany, 2.0%,20 but lower than the 76.1% reported from isolates in Afghanistan.24 Nucleotide changes in codon 531: TCG→TTG, 51.7% (31/60), were more predominantly observed among isolates, and correlating to a higher frequency level of resistance to RIF MIC⩾8 μg ml−1 in our study (Table 4). It is unclear whether novel mutation types E458A (GAG→GCG, MIC>8 μg ml−1), P535S (CCC→TCC, MIC>8 μg ml−1) and S509R (AGC→AGA, MIC⩾4 μg ml−1) are associated with RIF resistance because these isolates also carried mutations commonly associated with RIF resistance.25 Five isolates (8.3%, MIC>2–8 μg ml−1) had no mutation in the rpoB gene, in accordance with the previous reports,20 which suggest that the mutations may be elsewhere or that there are unknown mechanisms of drug resistance.

It has been reported that M. tuberculosis resistance to INH was the first step in the evolution of a MDR phenotype.26 71.7% (43/60) MDR isolates in our study were observed to have mutations in the katG fragments, consistent with the reports from many different geographical regions.10, 23, 27 The most common mutation in the katG gene was at codon 315, 62.5% (40/64, MIC⩾4 μg ml−1), similar to the mutation rates previously reported as 72.7% in Shanghai, China,23 68.4% in Cairo, Egypt,16 and 68.7% in California, USA.10 However, this is lower than the mutation rates reported in the Russian Federation (93.6%) and in Brazil (87.1%).23 It appears that mutation in katG 315 can show geographical differences in frequency and may not be a suitable independent marker for INH resistance. In this study, nucleotide changes in codon 315: AGC→ACC, 62.5%(40/64), were more predominantly observed among isolates, and correlating to a higher frequency level of resistance to INH MIC⩾4 μg ml−1 (Table 5). However, it remains to be determined if the novel mutations G273S (MIC>4 μg ml−1), I266T (MIC=4 μg ml−1), P232S (MIC=4 μg ml−1) and the 10.8-kb deletion affect susceptibility to INH.

We also sequenced the structure genes of inhA and ahpC, the mabA-inhA regulatory region and oxyR-ahpC intergenic region. No mutation was detected in the structure gene of inhA in all 85 isolates. However, 12.5% (8/64) of the INH-resistant isolates harbored base substitutions C-15T in the mabA-inhA regulatory region, consistent with the prior investigation.28, 29

In this study, only one mutation, A190S, in the inhA structural gene was found, coupled with S315T mutation in the katG gene, a mutation frequency much lower than previous reports.30 Among all mutations in oxyR-ahpC intergenic region, one isolate also had a G273S (GGT→AGT) novel mutation in katG gene. An association between oxyR-ahpC intergenic region mutations and mutations in katG other than at codon 315 has been reported previously.31 Our results implied that the relationship between mutations of ahpC structural gene and oxyR-ahpC intergenic region, and INH resistance still is not very clear, and is worthy of further investigation.

Of particular importance, in this study, were seven XDR isolates, accounting for 11.7% (7/60) of the identified MDR isolates, which is much higher than previous reports.32, 33 Mutations in the XDR isolates were focused on rpoB 531 or rpoB 526, and katG 315. Recently, however, Müller et al. reported that the percentage of isolates exhibiting mabA-inhA regulatory region C-15T mutations increased significantly from 48.4% and 62.4%, respectively, in MDR isolates to 85.5% and 91.9% in XDR isolates.34 The results of this study showed that only 42.9% (3/7) of XDR isolates were detected to have C-15T mutations, and these isolates also had an accompanying mutation in katG gene. Hence, more studies need to be performed to investigate if the mabA-inhA regulatory region was associated with the XDR isolates.

At the genetic level, gene alterations can change the structure of a target protein via mutations in the coding region or the amount of the protein expressed by modulating gene regulation, both of which ultimately cause the anti-drug resistance in M. tuberculosis.

In this study, the Beijing genotypes (91.8%) were the major epidemic genotypes among the 85 clinical isolates. There are no visible differences between the Beijing genotype and drug resistance. Beijing genotypes account for 91.8% of all isolates and non-Beijing genotype account for 7.2%. As the isolates in different regions had different characteristics, it is necessary to carry out studies on the relationship between Beijing genotype and drug resistance, pathogenicity and the ability to transmit.

In conclusion, this is the first report on the simultaneous genotypic analysis of the rpoB, katG, inhA, ahpC, mabA-inhA regulatory region and oxyR-ahpC intergenic region in South-central China M. tuberculosis isolates. These six gene regions account for 95.0% (57/60) of the MDR M. tuberculosis isolates. These information illustrate the need for further investigations to develop a more rapid and specific assay for the detection of MDR M. tuberculosis to be used as a screening method in areas. These observations also suggested that the drug resistance mechanisms of these first-line antituberculosis drugs were not completely clear and still need further research.