Background

Tuberculosis (TB) is a devastating infectious disease causing high mortality and morbidity worldwide with 8 million new TB cases and 2–3 million deaths annually. The situation of TB is made even worse by the rising emergence of drug resistant strains of Mycobacterium tuberculosis. Multi-drug resistant TB (MDR-TB) is defined as resistant to at least isoniazid (INH) and rifampin (RMP), the two most active first-line drugs against TB. MDR-TB treatment takes up to 2 years with second line drugs, which are expensive and have side effects. In 2006 US Centers for Disease Control and Prevention (CDC) and the World Health Organization (WHO) drew attention to the emergence of M. tuberculosis with extensive drug resistance to second-line antituberculosis drugs (XDR). XDR-TB is resistant to at least INH and RMP among the first-line drugs and to at least one of three injectable second-line anti-tuberculosis drugs used in TB treatment (capreomycin, kanamycin, amikacin) [1]. Thus, the treatment of such tuberculosis is becoming seriously limited, sometimes returning TB control to the pre-antibiotic era [1]. Tuberculosis chemotherapy started in 1944, when streptomycin (SM) was administered for the first time to a critically ill TB patient. Later, TB treatment was enriched with paraaminosalicylic acid (PAS-1949), INH (1952), pyrazinamide (PZA-1954), ethambutol (EMB-1962) and RMP (1963). It was identified that monotherapy generates drug-resistant mutants within a few months, endangering the success of antibiotic treatment. This problem was overcome by using combinations of drugs with as many as four drugs recommended nowadays by CDC and WHO [2].

The key antituberculosis drug commonly used in the treatment of tuberculosis is RMP. The loss of RMP as an effective drug leads to a need for a longer duration of therapy and often to a lower cure rate [3–6]. Drug resistance in M. tuberculosis is caused by mutations of various chromosomal genes, as identified for MDR occurrence due to the sequential accumulation of mutations in different genes that provide resistance to individual drugs. The individual molecular mechanisms of resistance have been identified for all first-line drugs and the majority of second-line drugs [7]. In M. tuberculosis, resistance to RMP results from mutations in the β-subunit of RNA polymerase, which is encoded by the rpo B gene [8]. Approximately 95% of RMP-resistant strains carry mutations within an 81-bp region containing codons 507 through 533 of the rpoB gene [8–10]. The single mechanism of resistance and narrow distribution of mutations make rpoB-81 bp region very attractive for molecular detection of resistance to RMP [11, 12]. However, within several dozen different mutations detected in the rpoB-81 bp region of RMP-resistant M. tuberculosis strains [for review see [13]], very few were tested by cloning and complementation assays. Mutated rpoB genes (S531L; H526Y; D516V) were introduced into the RMP sensitive M. tuberculosis H37Rv strain, resulting in acquired drug resistance of the host strain [14]. These authors observed that the level of acquired resistance was higher for mutants carrying mutations in codons 531 and 526 compared to mutation in codon 516. In this paper a genetic model was constructed allowing for a relatively simple verification of the relationship between the presence of a given mutation in rpoB-81 bp region and the RMP resistance of the host strain carrying such a mutation. Some rpoB mutations revealed drug-resistance only in selected M. tuberculosis strains suggesting that genetic background of the host is important for the development of resistance to RMP.

Methods

Bacterial strains and growth conditions

The M. tuberculosis strains examined for this study were isolated from TB patients in Poland in 2000 during the second national survey of drug resistance [12, 15]. Eight clinical strains identified as drug resistant, carrying different mutations in the rpoB gene, and two susceptible strains identified as drug sensitive, which did not carry any mutation in rpoB, were selected. Moreover, a control laboratory strain M. tuberculosis H37Ra, was included in this study. Primary isolation, differentiation, and drug susceptibility testing were performed with Lowenstein-Jensen (LJ) medium and the BACTEC 460-TB system (Becton-Dickinson, Sparks, Md.), as reported earlier [15]. All mycobacterial strains used in this study were cultured in Middlebrook 7H9 broth supplemented with OADC (albumin-dextrose-sodium chloride) and with kanamycin (25 μg/ml), or hygromycin (10 μg/ml), when required. Mycobacterial transformants were selected on Middlebrook 7H10 agar plates enriched with OADC containing kanamycin (Km) or hygromycin (Hyg).

Gene cloning strategies

Standard molecular biology protocols were used for all cloning procedures [16]. All PCR products were obtained using thermostable ExTaq polymerase (Takara) and cloned initially into pGemT vector (Promega), sequenced, and then released by digestion with appropriate restriction enzymes before cloning into the final vectors. Some restriction enzymes recognition sites were incorporated into the sequence of primers. The primers and plasmids used in this work are listed in Table 1 and 2, respectively. To engineer various rpoB genes of M. tuberculosis controlled by a natural promoter, a basal pRpoZero vector was constructed (Fig. 1). The vector contained the 5' end of rpoB until a natural Bst EII restriction enzyme recognition site (681 plus 950 bp of upstream region) which was connected to the 3' fragment of the gene starting with a natural Bst EII restriction enzyme recognition site (1122 plus 218 bp of downstream region). The resultant construct was used for cloning of the inner Bst EII-Bst EII fragment (1716 bp) of rpoB genes from various M. tuberculosis clinical strains resistant to RMP. The correct orientation of cloned Bst EII fragments was verified by digestion with PvuII endonuclease. Next, the cloned genes controlled by their natural promoter, carrying given mutations or wild type sequence in the hot spot region were relocated into the pMV306 integration vector. The resultant constructs (pMRP1-9) were electrotransformed into RMP susceptible strains, and the integration of DNA was monitored by Km selection and verified by PCR. Alternatively, the investigated rpoB genes were relocated without putative promoter sequence into pMV306P hsp integration vector under control of strong promoter (Phsp 65). The resultant constructs (pMHRP1-9) were electrotransformed into RMP susceptible strains, and the integration of DNA was monitored by Hyg selection and verified by PCR.

Figure 1
figure 1

Construction strategy of integration (pMRP1-9; pMHRP1-9) and self-replicating (pMERP1-9) plasmids carrying wild type and mutated rpoB genes under control of own (pMRP1-9) and heat shock (pMHRP1-9; pMERP1-9) promoter. Description in the text.

Full size image
Table 1 Primer sequences used for PCR amplification
Full size table
Table 2 Plasmids used in this study
Full size table

Susceptibility testing

Susceptibility testing was conducted using the proportion method on Youmans' liquid medium supplemented with 10% OADC with seven concentrations of RMP (50, 25, 12.5, 6.2, 1.5, 0.75, 0.37 μg/ml). The growth was determined after 21 days of incubation. The results were verified by Alamar Blue Assay [17–19] and by plating bacteria on Middlebrook 7H10 supplemented with OADC and various concentrations of RMP.

Results

The level of RMP resistance depends on the site and kind of substitution identified in the rpoB gene

The epidemiological studies carried out in many clinical laboratories worldwide have revealed several dozen mutations present in the rpoB gene of RMP resistant M. tuberculosis strains [12, 14, 20–23]. According to our knowledge, only three specific mutations of rpoB have been verified so far by molecular cloning techniques [14]. The complementation of RMP sensitive M. tuberculosis strain with rpoB gene carrying given mutation is not simply due to the gene length (3519 bp). One step amplification of gene together with its putative promoter based on M. tuberculosis genomic DNA as a template and its cloning is rather tough for investigators. To avoid this problem we have engineered pRpoZero vector carrying a 950 bp putative promoter region followed by 5'(721 bp) and 3' (1258 bp) rpoB gene fragments of an RMP-sensitive M. tuberculosis H37Ra strain (Fig. 1). The missing inner part of the rpoB gene flanked with natural Bst EII restriction sites contains an 81-bp mutable region. The Bst EII fragment (1716 bp) of rpoB gene can be easily amplified based on genomic DNA isolated from investigated M. tuberculosis RMP-resistant strains and cloned in frame to complete the rpoB gene in the pRpoZero system.

In this study we have selected eight M. tuberculosis RMP-resistant clinical strains carrying different mutations in rpoB gene [12] (Table 3). The PCR generated Bst EII inner fragments of the rpoB gene were verified by sequencing and were cloned into the pRpoZero vector. The correct orientation of insert was confirmed by Pvu II restriction analysis. Subsequently, the constructed genes, together with the putative promoter region, were relocated into the pMV306 integration vector using Xba I and Hind III restriction enzymes. The resultant constructs carrying wild type or mutated rpoB genes under control of a natural promoter, were electroporated into an RMP-sensitive M. tuberculosis H37Ra host. The integration of plasmid DNA into the attB site of chromosomal DNA was verified by PCR using MVs and MVr primers.

Table 3 Rifampin resistance of clinical and control M. tuberculosis strains
Full size table

The wild type clinical strains and engineered M. tuberculosis H37Ra mutants were subjected to RMP-resistance analysis using the proportional method. Each strain was encoded by number and analyzed at least three times by standard procedure at the National Reference Center for Mycobacteria in Poland. The results obtained by the proportional method were verified using Alamar Blue Assay and by plating bacteria on Middebrook 7H10 supplemented with OADC and various concentrations of RMP (data not shown). The results obtained for clinical strains and engineered mutants are summarized in Table 3 and 4, respectively. Only three out of eight analyzed mutations (H526D; D516V; S531L) revealed the same level of RMP-resistance in clinical strains and engineered H37Ra mutants. Introduction of other mutations identified in RMP-resistant M. tuberculosis clinical strains into the H37Ra host did not result in resistance to RMP or the level of MIC was very low in comparison with clinical strains. Mutation of codon 516 substituting D with V resulted in a high level of RMP resistance. This effect was not observed when D was substituted with Y or G, even when an extra mutation was present in codon 510, 512 or 515.

Table 4 Rifampin resistance of M. tuberculosis recombinant clones
Full size table

Some rpoB mutations are able to cause RMP resistance only in a particular M. tuberculosis host

The observed different levels of resistance of M. tuberculosis clinical strains and H37Ra strain carrying rpoB genes mutated at the same positions lead to the conclusion that some mutations in the rpoB gene can reveal drug-resistant phenotype only in a specific genetic background of the host. To verify this hypothesis integration vectors carrying mutated rpoB genes under natural promoters were introduced by electroporation into two M. tuberculosis clinical strains (KL463; KL1936) sensitive to RMP. The selected transformants were verified by PCR amplification as described above. The resultant clinical strains carrying mutated rpoB genes were subjected to RMP resistance analysis by the proportional method. The results obtained were compared to the RMP-resistance of clinical strains carrying the same mutations and to the H37Ra recombinants described above (Table 4). The mutated rpoB genes generating high RMP-resistance level in M. tuberculosis H37Ra (H526D; D516V; S531L) were also responsible for high level of resistance of both clinical strains when introduced into their chromosomal DNA. On the other hand, mutation Q513L identified in an M. tuberculosis strain with resistance to a high level of RMP (MIC up to 50 μg/ml) which did not cause significant resistance of M. tuberculosis H37Ra (MIC up to 6.2 μg/ml), was responsible for RMP-resistance of KL463 and KL1936 strains at the level depending on the host (up to 12.5 and 50 μg/ml, respectively). The double mutation of rpoB in positions 510 (Q/H) and 516 (D/Y) identified in a highly resistant M. tuberculosis strain (MIC 25 μg/ml) which did not reveal resistance in H37Ra (MIC 1.5 μg/ml) was responsible for low level of resistance of both clinical tubercle bacilli hosts (MIC 6.2 μg/ml).

The overproduction of mutated RpoB does not cause high level of resistance to RMP

We could not exclude that the different resistance of M. tuberculosis hosts carrying identical mutations in rpoB depends on different expression of RpoB controlled by unknown regulatory proteins. For example, the raised expression of target molecule (InhA) due to accumulations of mutations in promoter region is one of the known mechanisms of resistance to INH. As questions arose as to whether expression of mutated rpoB genes under control of the heat shock promoter (Phsp 60) resulted in increased resistance of M. tuberculosis to RMP, the wild type rpoB and its mutated copies were cloned under control of the heat shock promoter as described in Methods. Although we did not have antibodies to test the level of expression for RpoB, the expression system is known to be very efficient [24, 25]. The self-replicating constructs (pMERP1-9, Fig. 1) appeared to be very unstable when introduced into M. tuberculosis host (data not shown). Therefore the vectors (pMHRP1-9), which are able to integrate into attB site of mycobacterial chromosomal DNA, carrying wild type and mutated rpoB under Phsp 60promoter were constructed and electroporated into M. tuberculosis H37Ra. The presence of the relevant DNA introduced into the attB site of chromosomal DNA was verified by PCR amplification. The resultant recombinant strains were subjected to RMP resistance analysis by the proportional method. The results of RMP-MIC analysis obtained for strains carrying mutated rpoB genes under control of the strong Phsp 60promoter were similar to strains carrying the same rpoB genes under control of their natural promoter (Table 4). These observations suggest that the RMP-resistance of M. tuberculosis strains carrying rpoB mutated genes was not dependent on the rpoB expression level but resulted from the host genetic background that influence the drug-resistance phenotype.

Discussion

All bacteria achieve resistance to RMP by mutations in a defined region of the RNA polymerase subunit β. In M. tuberculosis, approximately 95% of RMP resistant clinical isolates carry a mutation in the rpoB gene [8]. On the other hand, many isolates from M. avium and M. intracellulare present a natural resistance to RMP as a result of an efficient permeability and exclusion barrier [26, 27]. Mutations in rpoB generally result in high level resistance to RMP. However, specific mutations in codons 511, 516, 518 and 522 can result in a lower resistance to RMP [14, 28, 29]. The role of some rpoB mutations (H526Y, S531L, D516V) in causing resistance was confirmed by genetic transformation experiments [14, 30]. Several dozen other mutations identified in the rpoB gene of RMP-resistant M. tuberculosis clinical isolates have never been confirmed by genetic cloning [12, 31–35]. Nowadays, when many genetic techniques are well developed, the knowledge about mutations connected to RMP-resistance is becoming used in the rapid identification of drug resistance [11, 12, 36, 37]. However, the utility of these techniques depends on the precise information about the role of any given mutation in RMP resistance.

In this study we have engineered a genetic system which is helpful in the verification of the relationship between the presence of a given mutation in rpoB and RMP resistance. We have found that rpoB gene carrying either D516V or S531L mutation causes resistance to RMP when introduced into the M. tuberculosis hosts what was in agreement with previous investigations [14]. On the other hand, when mutated rpoB was introduced into drug sensitive M. tuberculosis laboratory or clinical strains, the other substitutions in position 516 (D/Y; D/G), even when supported with Q510H, M515I or S512I identified in RMP-resistant M. tuberculosis clinical strains, did not result in a significant increase of RMP-resistance. Other authors previously reported the identification of D516Y substitutions of rpoB in M. tuberculosis resistance to a high level of RMP [21, 38], low level of RMP [14] and in strains sensitive to RMP [39]. Taken together, this suggests that D516Y/G substitutions in rpoB are not sufficient to result in RMP-resistance of M. tuberculosis. The substitutions in codon 526 (H/Y, D, R, L, P) were usually identified in M. tuberculosis clinical isolates highly resistant to RMP [14, 23, 38]. In this paper we have provided direct evidence that mutation H526D in rpoB is responsible for RMP-resistance when introduced into M. tuberculosis host. This result is similar to a previous finding verified by genetic transformation of rpoB carrying mutation H526Y [14]. By contrast, the contribution of rpoB carrying Q513L mutation to RMP-resistance was not that evident. The insertion of this gene into an M. tuberculosis H37Ra laboratory strain did not result in a significant level of RMP-resistance, however the insertion of the same gene was responsible for resistance to RMP of two M. tuberculosis clinical strains (MIC 12.5 and 50 μg/ml) when used as hosts. As identified in various clinical studies, the level of RMP-resistance of M. tuberculosis isolates carrying the Q513L mutation varies from 2 to 200 μg/ml [14, 20, 21, 23, 38]. The collected results suggest that rpoB carrying Q513L mutation is able to cause resistance to RMP only in selected tubercle bacilli. It is likely that this mutation can result in RMP-resistance in strains with low cell wall permeability since this exclusion barrier is responsible for natural resistance of some MAIC strains [26, 27]. We also cannot exclude the possibility that other mechanisms support RMP-resistance of strains carrying Q513L mutation.

The drug resistance of M. tuberculosis can be also connected to the overproduction of a drug target due to accumulation of point mutations in a promoter region [40–42]. To test whether overproduction of rpoB carrying a given mutation result in higher MIC for RMP compared to a strain expressing the same gene under control of the natural promoter, rpoB genes were cloned under control of the P hsp promoter and introduced into M. tuberculosis host. The P hsp promoter, commonly used in genetics studies of mycobacteria controlling the groEL gene (Rv0440) in M. tuberculosis, has already been reported as highly active in mycobacterial cells growing in vitro [24, 25]. A recent microarray study showed that the expression level of groEL in M. tuberculosis cells growing in log phase is high, but not higher than rpoB [43]. However, the arresting of M. tuberculosis growth results in 3.6-fold induction of groEL with a decrease of rpoB expression in the same conditions [44]. We have not observed higher RMP resistance when mutated rpoB genes were expressed under control of P hsp promoter in comparison to the natural promoter. It is possible that the natural level of RpoB is high enough to saturate RMP (if its concentration in cell is low). On the other hand, the extra expression of rpoB cannot help in cells accumulating high RMP level. However, to elucidate this problem an alternative expression system and precise control of protein expression would be required.

The natural resistance to RMP in some M. avium and M. intracellulare strains is known to be as a result of an efficient cell wall permeability and exclusion barrier [26, 27], suggesting that these elements may be also important in M. tuberculosis. Changes in cell wall composition could affect permeability [45] decreasing the intracellular concentration of drug.

Conclusion

Among several dozen amino acid substitutions identified in an rpoB 81-bp region only a few are directly responsible for M. tuberculosis resistance to rifampin. Many others require a specific genetic background to develop resistance. Our findings lead to the conclusion that direct, molecular identification of rifampin resistant M. tuberculosis clinical isolates is possible only for strains carrying selected mutations in RpoB. The identification of other mutations suggests that investigated strains might be resistant to this drug.