Abstract
The gyrase mutations and efflux pumps confer fluoroquinolones (FQ) resistance in Mycobacterium tuberculosis. However, the contribution of two mechanisms in FQ mono-resistant M. tuberculosis is still unclear. Here, we investigated the contribution of gyrase mutations and efflux pumps to FQ resistance among 17 clinical FQ mono-resistant strains. Our data showed that gyrase mutations in gyrA QRDR were only responsible for four FQ mono-resistant strains. Mutations located in Ala90 and Asp94 of GyrA confer high-level LFX resistance, which can be explained by 3D modeling affinity change between GyrA and LFX. In addition, we found that a high level of efflux pump pstB transcripts may confer FQ resistance in two high-level FQ-resistant isolates (MIC ≥ 4 μg mL−1). The recombinant Escherichia coli with pstB revealed greatly increased MIC level from < 0.125 μg mL−1 to 2 μg mL−1. For the two isolates harboring high-level pstB transcripts, the presence of CCCP reduced LFX resistance to 1.0 μg mL−1. The transcriptional levels of pstB showed no significant difference among 10 clinical M. tuberculosis isolates with different drug susceptibility profiles. In conclusion, our findings demonstrate that both QRDR mutation and efflux pump mechanisms are responsible for monoresistance to FQ. PstB may serve as FQ-related efflux pumps in M. tuberculosis.
Introduction
An ancient infectious killer, tuberculosis (TB), remains a major cause of morbidity and mortality worldwide (Morens et al., 2004; Lienhardt et al., 2012; Pang et al., 2012). The introduction of effective antitubercular drugs in the 1960s seemed to bring tuberculosis under control. However, with the spread of HIV infection and emergence of drug-resistant Mycobacterium tuberculosis strains, TB has become an increasingly prevalent health problem (Sullivan et al., 1995; Espinal et al., 2001; Mariam et al., 2004). Of particular concern is the increasing incidence of multidrug-resistant TB (MDR-TB), resistant to at least isoniazid and rifampin, leading to treatment failure and the growing epidemic of drug resistance (Nachega & Chaisson, 2003; Donald & Van Helden, 2009; Zhao et al., 2012). Fluoroquinolones are one of the most important second-line drugs used against drug-resistant TB, especially MDR-TB (Gillespie & Kennedy, 1998; Chan et al., 2004; Wade & Zhang, 2004; Jacobson et al., 2010; Kim et al., 2010; Pantel et al., 2012).
Fluoroquinolones bind to DNA gyrase, which is responsible for DNA supercoiling, thereby inhibiting DNA supercoiling of pathogens. In addition to generate double-stranded breaks in DNA, fluoroquinolones may lead to accumulation of toxic hydroxyl radicals in the target organism (Kohanski et al., 2007). The DNA gyrase includes two A and two B subunits, encoded by gyrA and gyrB, respectively (Drlica & Malik, 2003). High-level resistance to fluoroquinolones is associated with mutations in the quinolone resistance-determining region (QRDR) of A subunits of the gyrase (gyrA), including codons 88, 90, 91, and 94, which are the most common (Takiff et al., 1994; Cheng et al., 2004). In contrast, a mutation of the gyrB gene is responsible for the lower level of resistance. Unlike the high relativity between rpoB mutation with rifampin-resistant phenotype, the mutation of the gyrase is detected in only 70–90% of FQ-resistant strains (Takiff et al., 1994; Ginsburg et al., 2003; Campbell et al., 2011; Maruri et al., 2012), which suggests that the resistance of the strains with no mutation in the gyrase may be due to an active drug promoting the efflux pump mechanism (De Rossi et al., 2006). Although drug efflux may play an important role in the mechanisms of FQ resistance, little is known about the function of FQ efflux in vivo or in clinical isolates of M. tuberculosis (De Rossi et al., 2006; Wang et al., 2007; Xu et al., 2009). Two major facilitator family (MFS) efflux pumps, including LfrA efflux pump of Mycobacterium smegmatis and the Rv1634 efflux pump of M. tuberculosis, are associated with FQ transport in mycobacteria (De Rossi et al., 2006; Da Silva et al., 2011). In addition to the two efflux pumps mentioned above, numerous efflux determinants of fluoroquinolone resistance in mycobacteria have been described (Banerjee et al., 1996; Pasca et al., 2004; Siddiqi et al., 2004; Poole, 2005). Hence, it is necessary to gain insight into the efflux mechanism of fluoroquinolone resistance in mycobacteria, especially M. tuberculosis.
This study investigated the contribution of gyrase mutation and efflux pump mechanism to FQ resistance among 17 clinical FQ mono-resistant strains. The entire genes of gyrA/gyrB were scanned by sequencing to determine the role of gyrase mutation in the FQ mono-resistant mechanism. In addition, we used expression profile analysis of efflux pumps for the FQ mono-resistant strains without mutations to conduct preliminary studies to characterize efflux pump-mediated FQ resistance in M. tuberculosis. Furthermore, we used Escherichia coli as the model organism to confirm the function of the candidate efflux pumps.
Materials and methods
Bacterial strains and culture condition
Bacterial strains obtained from drug resistance baseline surveillances in China (Pang et al., 2012) were stored in Trypticase soy broth containing 10% glycerol at −70 °C. The strains were cultured on Lowenstein-Jensen medium for 4 weeks at 37 °C prior to drug susceptibility testing and genotype analysis.
Traditional drug susceptibility testing and Mycobacterium species identification
According to the recommendation by the World Health Organization and the International Association Against Lung Disease, four first-line anti-TB drugs (isonazid, rifampin, ethambutol, and streptomycin) and two second-line anti-TB drugs (kanamycin and ofloxacin) were enrolled in traditional species identification and testing of clinical tuberculosis strains for drug susceptibility. Drug susceptibility was determined by means of the proportion method with the use of Löwenstein-Jensen medium. The concentrations of drugs in media were isonazid 0.2 μg mL−1, rifampin 40 μg mL−1, ethambutol 2 μgmL−1, streptomycin 4 μg mL−1, kanamycin 30 μg mL−1, and ofloxacin 2 μg mL−1. We used media supplied separately with paranitrobenzoic acid (500 mg mL−1) and thiophen-2-carboxylic acid hydrazide (5 mg mL−1) to identify Mycobacterium species. The critical growth proportion for resistance was 1% for all drugs. All drugs were purchased from Sigma–Aldrich (St. Louis, MO).
Genomic DNA extraction
Genomic DNA was extracted from freshly cultured bacteria. After transfer into a microcentrifuge tube containing 500 μL Tris-EDTA (TE) buffer, the cells were centrifuged at 16 000 g for 2 min. The supernatant was discarded, and the pellet was resuspended in 500 μL TE buffer, followed by heating in water bath at 95 °C for 1 h. After centrifugation at 16 000 g for 5 min to separate cellular debris, DNA in the TE buffer was stored at −20 °C until the PCR amplification reactions were preformed (Pang et al., 2011a, b).
PCR amplification, molecular species identification, and sequencing of gyrA and gyrB genes
The 16S rRNA gene was used to perform the species identification as previously reported (Pang et al., 2011a, b). In addition, the entire gyrA and gyrB, two quinolone resistance-related genes, were amplified by PCR with the primers as a previous report (Supporting information, Table S1; Devasia et al., 2012). After purification with QIAquick Gel Extraction Kit, the amplicons were sequenced with Applied Biosystems ABI Prism Big Dye terminator cycle sequencing kit with an ABI Prism 3130 genetic analyzer.
Determination of minimal inhibitory concentration
To determine minimal inhibitory concentrations (MIC) of FQ mono-resistant M. tuberculosis and E. coli strains, a microplate Alamar Blue assay (MABA) was performed as described elsewhere (Franzblau et al., 1998; Baysarowich et al., 2008). The results were obtained after 24 h for E. coli, and after 7–14 days for M. tuberculosis. Due to superior antibacterial activity and clinical outcome, the LFX, instead of OFLX, was selected as the candidate FQ to carry out MIC experiment (Long et al., 2012). The final drug concentrations of LFX were 0.125–64 μg mL−1. In addition, carbonyl cyanide m-chlorophenylhydrazone (CCCP) was added to the assay mixture at 100 μM. The MICs were defined as the lowest concentration of antibiotic that reduced the viability of the culture by at least 90% as determined by fluorescence measurements at room temperature in top-reading mode; the excitation wavelength and emission wavelength were 530 and 590 nm, respectively (Pang et al., 2013). The high-level FQ resistance was defined as the strains with MICs ≥ 4 μg mL−1.
Modeling of three-dimensional structure for GyrA
The modeling of protein structures and analysis of protein–ligand interaction were carried out using the discovery studio 3.1 software (Acclerys, San Diego, CA). The wild-type protein structure with LFX and a magnesium ion (MG) was modeled from residues Arg14 to Ala500 using a Build Homology Modeling protocol with Protein Data Bank (PDB) (http://www.rcsb.org/pdb/) file 3RAE as the template. After optimization by loop refinement (MODELER), we used the Smart Minimizer algorithm for energy minimization, performing 5000 steps at RMS gradient of 0.001 kcal/(mol × angstrom) with the protein backbone atoms fixed. We then used a Build Mutants module to produce the structure models of Ala90Val, Asp94Ala, and Asp94Gly. We used an Align and Superimpose Proteins model to compare the mutant structures with wild-type structure. The interactions between ligands (LFX or MG) and wild-type or mutant GyrA were analyzed by the Structure Monitor model.
RNA isolation and quantitative real-time RT-PCR (qRT-PCR)
RNAs from fresh-cultured strains were isolated by Trizol (Invitrogen, Grand Island, NY) according to the manufacturer's instructions. First-strand cDNA was synthesized from 1 μg total RNA using the SuperScript III First-Strand Synthesis System for RT-PCR according to the manufacturer's instructions, plus an additional step of DNase I treatment (1 unit/reaction for 15 min at 25 °C) (Invitrogen). Twenty-one pairs of gene-specific RT-PCR primers, including 20 putative drug efflux pump genes and polA gene as reference, were synthesized by Sangon Company (Shanghai, China) as listed in Table S1. qRT-PCR was performed using the SYBR Green PCR kit (New England BioLabs, Ipswich, MA). The qRT-PCR cycles were as follows: initiation with a 10-min denaturation at 95 °C, followed by 42 cycles of amplification with 10 s of denaturation at 94 °C, 15 s of annealing at 59 °C, 20 s of extension at 72 °C, and reading the plate for fluorescence data collection at 78 °C. A melting curve was performed from 65 °C to 95 °C (1-s hold per 0.3 °C increase) to check the specificity of the amplified product. All qRT-PCRs were performed in triplicate using three independent RNA samples.
Construction of expression vector
The open reading frames of Rv0933 (pstB) or Rv2938 were amplified using the primers listed in Table S1. They were purified with the Qiagen PCR purification kit (Qiagen, Valencia, CA) and separately ligated into pEASY-E1, an E. coli expression vector, resulting in generation of plasmids pEASY-E1-Rv0933 and pEASY-E1-Rv2938. The nucleotide sequences of Rv0933 and Rv2938 from both vectors were confirmed with DNA sequencing. The constructs were transferred into E. coli BL21 and selected with LB medium supplanted with ampicillin (50 μg mL−1).
Western blotting
Escherichia coli BL21 harboring pEASY-E1-Rv0933, pEASY-E1-Rv2938, or control plasmid pEASY-E1 were induced with 100 μM isopropyl β-D-1-thiogalactopyranoside (IPTG) when cultures reached an OD600 nm of ∼0.6 and harvested 2 h later. The whole crude lysate was separated by SDS–polyacrylamide gel (SDS–PAGE) electrophoresis, and the separated proteins were transferred onto a polyvinylidene difluoride (PVDF) membrane. The membrane was blocked 2 h in 5% dry milk in TBS (20 mM Tris, pH 7.5; 150 mM NaCl) at room temperature, followed by incubation with commercial anti-His6 tag antibody (Zhongshan Jinqiao Biotech., Beijing, China) for 2 h at room temperature. Then the membrane was hybridized with horseradish peroxidase (HRP)-linked second antibody. The corresponding substrate diaminobenzidine (DAB) was used to detect the antigen-antibody complex (Towbin et al., 1979).
Results
Mutation and MIC analysis of QRDR
Of 3634 M. tuberculosis isolates obtained from the National Drug Resistance Survey, 4.0% (145/3634) of M. tuberculosis isolates had resistance to OFLX (Pang et al., 2012), while only 17 isolates were mono-resistant to OFLX. All the 17 isolates were further identified as M. tuberculosis by 16S rRNA gene sequencing. Analysis of the gyrA/gyrB QRDRs in 17 FQ mono-resistant strains revealed mutations in QRDR of gyrA were in four strains (23.5%), and one mutation was observed in the QRDR of gyrB (5.9%). The gyrA mutations associated with FQ resistance were identified in codons 90 and 94 (including one strain with Ala90Val, two strains with Asp94Ala, and one strain with Asp94Gly; Table 1). In addition, 52.9% (9/17) and 70.6% (12/17) of resistant isolates had a polymorphism outside of gyrA or gyrB QRDR, respectively, and these polymorphisms did not confer to FQ resistance in M. tuberculosis (Devasia et al., 2012). The LFX MICs of four strains with mutations in gyrA QRDR ranged from 4 μg mL−1 to 8 μg mL−1 (median, 5 μg mL−1), while the LFX MICs of resistant strains without mutation in gyrA QRDR ranged from 0.5 μg mL−1 to 8 μg mL−1 (median, 1.9 μg mL−1). The statistical analysis showed that the MICs of gyrA-mutated strains were significantly higher than those of strains without mutation (P = 0.02, Table 1).
Mutation and MIC detected in single FQ-resistant Mycobacterium tuberculosis study isolates
Minimal inhibitory concentration against LFX.
QRDR, quinolone resistance-determining region, gyrA QRDR, codons 74–113; gyrB QRDR, codons 500–538.
Synonymous mutation of gyrA: Leu(CTG)591Leu(CTA) for TB2, Ala(GCC)581Ala(GCA) for TB6, Ala(GCC)581Ala(GCT) for TB10, and Leu(TTG)549Leu(TTA) for TB14.
Synonymous mutation of gyrB: Ser(TCT)55Ser(TCA) for TB2, Glu(GAG)662Glu(GAA) for TB12, and Thr(ACC)664Thr(ACG) for TB13.
Mutation and MIC detected in single FQ-resistant Mycobacterium tuberculosis study isolates
Minimal inhibitory concentration against LFX.
QRDR, quinolone resistance-determining region, gyrA QRDR, codons 74–113; gyrB QRDR, codons 500–538.
Synonymous mutation of gyrA: Leu(CTG)591Leu(CTA) for TB2, Ala(GCC)581Ala(GCA) for TB6, Ala(GCC)581Ala(GCT) for TB10, and Leu(TTG)549Leu(TTA) for TB14.
Synonymous mutation of gyrB: Ser(TCT)55Ser(TCA) for TB2, Glu(GAG)662Glu(GAA) for TB12, and Thr(ACC)664Thr(ACG) for TB13.
3D Structure modeling
For further structural analysis, the structure of wild-type TB GyrA complexed with LFX and MG was modeled using a discovery studio 3.1 template of the crystal structure of levofloxacin-DNA cleavage complex of type IV topoisomerase from Streptococcus pneumoniae (PDB code 3RAE). The predicted structure of TB GyrA was similar to the crystal structure of S. pneumoniae DNA topoisomerase 4 subunit A; the root-mean-square deviation (RMSD) value of the main chain atoms between these two structures was 2.463 Å (Fig. 1a). As shown in Fig. 1b, residue Ala90 had intermolecular interactions with LFX and MG, and residue Asp94 had intermolecular interactions only with MG.
Structure model of wild-type TB GyrA and detailed interactions of GyrA-LFX-MG complex. (a) Superimposition of wild-type TB GyrA (yellow) and Streptococcus pneumoniae DNA topoisomerase 4 subunit A (cyan, PDB code 3RAE). (b) Intermolecular interactions between wild-type GyrA and ligands (MG and LFX). The intermolecular interactions were indicated by green lines. (c) Structural comparisons between wild-type GyrA (yellow) and mutants Ala90Val (orange). The intermolecular structural bumps between mutant GyrA and ligands were indicated by magenta dashed lines. (d) Structural comparisons between wild-type GyrA (yellow) and mutants Asp94Ala (blue) and Asp94Gly (magenta). The intermolecular interactions between wild-type GyrA and MG were indicated by green lines. (b, c, d) The residues of GyrA involved in intermolecular interaction were shown in ball-and-stick style, and the MG (green) and LFX were shown in CPK style.
The structure comparisons between wild-type and mutant-type GyrA are shown in Fig. 1c and d. Wild-type GyrA lacks a structural bump between the side chain of alanine residue and ligands. In mutant Ala90Val, the replacement of Ala90 by Val90 led to structure bumps between the mutant valine residue and LFX molecule (Fig. 1c). Asp94 in wild-type GyrA had negative charge opposite to nearby MG and could interact with MG. After replacement of Asp94 by Ala94 or Gly94, the interaction with MG was lost (Fig. 1d).
Identification of efflux pump
To investigate the FQ-resistant mechanism in the strains without gyrA QRDR mutation, we carried out real-time PCR to explore the relative expression of genes encoding efflux pumps (Cole et al., 1998). As shown in Fig. 2, transcriptional level of two efflux pump genes Rv2938 and Rv0933 was significantly upregulated in the high-level resistant strains as compared to low-level resistant strains, while transcription of the remaining 18 was unchanged. Hence, Rv2938 and Rv0933 were selected as the candidate genes related to FQ resistance.
Relative expression levels of 20 different putative efflux pump genes in 13 Mycobacterium tuberculosis isolates without QRDR mutation. qRT-PCR was performed in triplicate using independent RNA samples prepared from different strains. The points on the line chart indicate fold increase relative to the H37Rv value (arbitrarily set to 1) of the same efflux pump gene.
We evaluated the in vivo efficiency of Rv2938 and Rv0933 by determining the MICs of LFX required to arrest cell growth. As shown in Fig. S1, Western blot analysis confirmed the presence of bacterial expressed Rv2938 and Rv0933 using anti-His6 antibody. The recombinant E. coli with pstB showed greatly increased MIC level, and under conditions where the genes were induced with isopropyl-beta-D-thiogalactopyranoside, MICs increased to 2 μg mL−1. In contrast, MIC of wild-type E. coli was lower than 0.125 μg mL−1, the E. coli clones transformed with pEASY-E1-Rv2938 revealed no effect in RIF resistance, and the MIC was also lower than 0.125 μg mL−1. When the efflux pump inhibitor CCCP was supplied in the media, the MIC of E. coli transformed with pEASY-E1-Rv0933 was significantly decreased to 0.25 μg mL−1 (Table 2).
In vitro levofloxacin susceptibility of overexpressing Rv2938 and Rv0933
| Construct in E. coli | MIC (μg mL−1) |
| pEASY-E1 | < 0.125 |
| pEASY-E1-Rv2938 | < 0.125 |
| pEASY-E1-Rv0933 | 2 |
| pEASY-E1 + CCCP | < 0.125 |
| pEASY-E1-Rv2938 + CCCP | < 0.125 |
| pEASY-E1-Rv0933 + CCCP | 0.25 |
| Construct in E. coli | MIC (μg mL−1) |
| pEASY-E1 | < 0.125 |
| pEASY-E1-Rv2938 | < 0.125 |
| pEASY-E1-Rv0933 | 2 |
| pEASY-E1 + CCCP | < 0.125 |
| pEASY-E1-Rv2938 + CCCP | < 0.125 |
| pEASY-E1-Rv0933 + CCCP | 0.25 |
In vitro levofloxacin susceptibility of overexpressing Rv2938 and Rv0933
| Construct in E. coli | MIC (μg mL−1) |
| pEASY-E1 | < 0.125 |
| pEASY-E1-Rv2938 | < 0.125 |
| pEASY-E1-Rv0933 | 2 |
| pEASY-E1 + CCCP | < 0.125 |
| pEASY-E1-Rv2938 + CCCP | < 0.125 |
| pEASY-E1-Rv0933 + CCCP | 0.25 |
| Construct in E. coli | MIC (μg mL−1) |
| pEASY-E1 | < 0.125 |
| pEASY-E1-Rv2938 | < 0.125 |
| pEASY-E1-Rv0933 | 2 |
| pEASY-E1 + CCCP | < 0.125 |
| pEASY-E1-Rv2938 + CCCP | < 0.125 |
| pEASY-E1-Rv0933 + CCCP | 0.25 |
The MICs of LFX were also determined in the presence of efflux pump inhibitor in the clinical isolates. For TB16 and TB17 isolates, harboring high-level Rv0933 transcriptional level, the presence of CCCP reduced LFX resistance to 1.0 μg mL−1. In contrast, the MICs of LFX among the TB isolates were not affected by efflux pump inhibitor CCCP significantly (Fig. 3).
Effect of CCCP on MICs of LFX for 13 clinical FQ-resistant Mycobacterium tuberculosis isolates without QRDR mutation in gyrA and gyrB.
To investigate whether Rv0933 conferred other drug resistance in M. tuberculosis, we analyzed the transcriptional profiles among 10 isolates with different drug-resistant patterns. As shown in Fig. 4, the transcriptional levels of Rv0933 showed no significant difference among those clinical isolates.
Relative expression levels of Rv0933 in 10 clinical Mycobacterium tuberculosis isolates with different drug susceptibility profiles. qRT-PCR was performed in triplicate using independent RNA samples prepared from different strains. The points on the line chart indicate fold increase relative to the H37Rv value (arbitrarily set to 1).
Discussion
Previous studies have demonstrated that the FQ-resistant phenotype of M. tuberculosis is associated with gene mutations. Several mutations located in the FQ-resistance-determining region are correlated with about 70–90% of FQ-resistant strains, while the mutation of the other 10–30% of FQ-resistant strains could not be detected in QRDR (Von Groll et al., 2009; Surcouf et al., 2011). Hence, studying the molecular mechanism of FQ drug resistance will increase our knowledge of how the strains without mutation become FQ resistant and will improve the reliability of the rapid molecular assay for detecting FQ resistance in M. tuberculosis.
In the present study, we first identified the genotypic mutations in the QRDR of both gyrA and gyrB. To our surprise, only 23.5% of FQ mono-resistant isolates were associated with gyrA mutations, significantly less than the portion reported previously (Long et al., 2012). Our data indicated that other molecular alterations, such as overexpression of efflux pumps, are mainly responsible for FQ monoresistance. In agreement with previous study, the strains with mutation in Ala90 and Asp94 were associated with high-level resistance (Von Groll et al., 2009), while most of other strains without mutations were associated with low-level resistance.
Amino acid alterations at position 88, 90, and 94 in GyrA are involved in acquired resistance to FQ (Aubry et al., 2006). The recent study on the 3D structure of wild-type DNA gyrase allowed us to obtain several structural insights into the mechanism of intrinsic resistance to FQ. The analysis showed that A90 and R482 of DNA gyrase are the core region for drug target, which directly affects drug resistance level (Piton et al., 2010). Our structural modeling analysis revealed that the replacement of Ala90 by Val90 led to structure bumps, which reduced the affinity between LFX and GyrA. However, the substitution of amino acid at position 94, the most common single nucleotide mutation site conferring FQ resistance, revealed a paradoxical effect between the resistance level and the protein 3D structure described previously (Piton et al., 2010). In this study, the modeling results indicated Asp94 might interact with MG while the mutants Asp94Ala and Asp94Gly lost this interaction. Results indicated that MG could mediate the binding of quinolone to DNA but showed no direct evidence that MG could mediate or facilitate the binding of LFX to gyrase. Thus, further studies are needed to reveal the role of the interaction between Asp94 and MG.
In our study, just under 25% of FQ mono-resistant strains carried mutations in the QRDR of the gyrA/gyrB genes, while more than 70% of drug-resistant isolates could not be explained by the classical mutations in the target genes, consistent with a previous study in a Chinese cohort (Zhu et al., 2012). Some reports have shown that efflux pumps serve as important contributors to drug resistance in M. tuberculosis (Almeida Da Silva & Palomino, 2011; Pang et al., 2013). To investigate the involvement of efflux pumps in conferring resistance to FQ among the single-drug-resistant strains, we analyzed the expression level of 20 putative efflux pump genes among the strains without QRDR mutations. Increased expression of PstB seems to raise the FQ resistance level in two FQ-resistant isolates without genomic mutation in QRDR regions of gyrA and gyrB, and further transgenic experiments revealed that encoding phosphate-specific transporter Rv0933 may account for the FQ resistance among the FQ mono-resistant strains. Rv0933 belongs to the ABC family of transporters, members of which confer drug resistance in M. tuberculosis (Pasca et al., 2004). Pasca et al. (2004) revealed that Rv2686c-Rv2687c-Rv2688c, an ABC transporter, served as an active efflux pump by pumping out FQ from M. tuberculosis cells. The Pst system, one of the phosphate importers in mycobacteria, operates in phosphate-limiting conditions (Sarin et al., 2001). As a member of the Pst system, PstB plays a vital role in the ATP-consuming import of phosphate (Sarin et al., 2001). In contrast to other subunits of this system, only a single copy of the pstB gene is present in the M. tuberculosis genome, which indicates that PstB gains additional importance in the transporter system (Sarin et al., 2001). An early study reported that pstB overexpressed as well as amplified in an FQ-resistant colony of rapidly growing M. smegmatis, indicating that pstB may confer FQ resistance in Mycobacterium (Banerjee et al., 1998, 2000; Bhatt et al., 2000). Our experiments first indicated that increased transcript levels of pstB might contribute to FQ resistance in some M. tuberculosis isolates. By combining theories and data from previous reports, we propose two potential mechanisms that could explain how PstB promotes such FQ drug resistance. First, PstB is an ATPase responsible for catalyzing ATP hydrolysis. Hence, the M. tuberculosis PstB may actively pump out FQ drug by ATP hydrolysis as an energy source. In addition, phosphate is a vital nutrient for M. tuberculosis during the rapidly changing infectious stage of the host environment (Braibant et al., 2000). The overexpression of PstB may create a more congenial nutritional environment for M. tuberculosis survival by importing phosphate. However, this hypothesis must be tested by further research.
In conclusion, our findings demonstrate that both QRDR mutation and efflux pump mechanism are responsible for monoresistance to FQ. PstB may serve as FQ-related efflux pumps, conferring high-level resistance to LFX. Further studies need to be performed in M. smegmatis to confirm the function of PstB.
Acknowledgements
This study was supported by the National Grant Research Program of China (2012ZX10004209). We thank Ms. Sandy Marvinney from PATH for her help in editing the language of this manuscript.
Author's contribution
J.L and M.L. contributed equally to this paper.
References
Supporting Information
Additional Supporting Information may be found in the online version of this article:
Fig. S1. Western blot analysis of recombinant proteins produced in E. coli cells.
Table S1. Primers for sequencing and qRT-PCR.
Table S2. Primers used in this study for MIRU-VNTR amplification.
Author notes
Editor: Roger Buxton
