Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Whole Genome Sequencing Based Characterization of Extensively Drug-Resistant Mycobacterium tuberculosis Isolates from Pakistan

  • Asho Ali ,

    ‡ These authors are joint first authors on this work.

    Affiliations Department of Pathology and Microbiology, Aga Khan University Hospital, Karachi, Pakistan, School of Nursing and Midwifery, The Aga Khan University Hospital, Karachi, Pakistan

  • Zahra Hasan ,

    zahra.hasan@aku.edu

    ‡ These authors are joint first authors on this work.

    Affiliation Department of Pathology and Microbiology, Aga Khan University Hospital, Karachi, Pakistan

  • Ruth McNerney,

    Affiliation Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, United Kingdom

  • Kim Mallard,

    Affiliation Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, United Kingdom

  • Grant Hill-Cawthorne,

    Affiliations Pathogen Genomics Laboratory, King Abdullah University of Science and Technology (KAUST), Thuwal, Kingdom of Saudi Arabia, Sydney Emerging Infections and Biosecurity Institute and School of Public Health, Sydney Medical School, University of Sydney, Sydney, Australia

  • Francesc Coll,

    Affiliation Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, United Kingdom

  • Mridul Nair,

    Affiliation Pathogen Genomics Laboratory, King Abdullah University of Science and Technology (KAUST), Thuwal, Kingdom of Saudi Arabia

  • Arnab Pain,

    Affiliation Pathogen Genomics Laboratory, King Abdullah University of Science and Technology (KAUST), Thuwal, Kingdom of Saudi Arabia

  • Taane G. Clark,

    Affiliation Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, United Kingdom

  • Rumina Hasan

    Affiliation Department of Pathology and Microbiology, Aga Khan University Hospital, Karachi, Pakistan

Whole Genome Sequencing Based Characterization of Extensively Drug-Resistant Mycobacterium tuberculosis Isolates from Pakistan

  • Asho Ali, 
  • Zahra Hasan, 
  • Ruth McNerney, 
  • Kim Mallard, 
  • Grant Hill-Cawthorne, 
  • Francesc Coll, 
  • Mridul Nair, 
  • Arnab Pain, 
  • Taane G. Clark, 
  • Rumina Hasan
PLOS
x

Abstract

Improved molecular diagnostic methods for detection drug resistance in Mycobacterium tuberculosis (MTB) strains are required. Resistance to first- and second- line anti-tuberculous drugs has been associated with single nucleotide polymorphisms (SNPs) in particular genes. However, these SNPs can vary between MTB lineages therefore local data is required to describe different strain populations. We used whole genome sequencing (WGS) to characterize 37 extensively drug-resistant (XDR) MTB isolates from Pakistan and investigated 40 genes associated with drug resistance. Rifampicin resistance was attributable to SNPs in the rpoB hot-spot region. Isoniazid resistance was most commonly associated with the katG codon 315 (92%) mutation followed by inhA S94A (8%) however, one strain did not have SNPs in katG, inhA or oxyR-ahpC. All strains were pyrazimamide resistant but only 43% had pncA SNPs. Ethambutol resistant strains predominantly had embB codon 306 (62%) mutations, but additional SNPs at embB codons 406, 378 and 328 were also present. Fluoroquinolone resistance was associated with gyrA 91–94 codons in 81% of strains; four strains had only gyrB mutations, while others did not have SNPs in either gyrA or gyrB. Streptomycin resistant strains had mutations in ribosomal RNA genes; rpsL codon 43 (42%); rrs 500 region (16%), and gidB (34%) while six strains did not have mutations in any of these genes. Amikacin/kanamycin/capreomycin resistance was associated with SNPs in rrs at nt1401 (78%) and nt1484 (3%), except in seven (19%) strains. We estimate that if only the common hot-spot region targets of current commercial assays were used, the concordance between phenotypic and genotypic testing for these XDR strains would vary between rifampicin (100%), isoniazid (92%), flouroquinolones (81%), aminoglycoside (78%) and ethambutol (62%); while pncA sequencing would provide genotypic resistance in less than half the isolates. This work highlights the importance of expanded targets for drug resistance detection in MTB isolates.

Introduction

Mycobacterium tuberculosis (MTB) caused 9 million cases of tuberculosis (TB) and 1.5 million death worldwide in 2013 [1]. Drug resistance in MTB increases the burden due to the disease and complicates both diagnosis and treatment. Multidrug-resistant (MDR)-TB is caused by MTB resistant to at least the two first line drugs i.e. rifampicin and isoniazid. Globally, in 2013, an estimated 480,000 people developed TB while, there were about 210,000 deaths due to TB [1]. Extensively drug resistant (XDR)-TB is caused by MDR-TB strains resistant to fluoroquinolone and to any one of the injectable aminoglycosides; amikacin, kanamycin or capreomycin. In Pakistan, MDR-TB comprises 4–5% of all TB cases, and XDR-TB is 4–5% of the MDR-TB cases [2,3]. Appropriate MDR-TB management requires drug susceptibility testing to guide treatment regimens. The usual turnaround time for phenotypic MTB drug susceptibility testing (DST) is about 8–10 weeks, often leading to delays in initiation of appropriate therapy. In high TB burden low resource settings facilities for mycobacterial culture and DST are limited. Therefore, rapid genotypic methods for detection of drug resistant MTB strains are required for more effective management of TB.

The Xpert MTB/RIF assay (Cepheid, USA) assay was rolled out at a global level since 2012; it detects rifampicin resistance by targeting the ribosomal polymerase rpoB gene [4]. The LiPA MTBDRplus (Hain Lifescience GmbH, Germany) assay targets rpoB for rifampicin resistance and katG and inhA genes for isoniazid resistance. The LiPA MTBDRsl (Hain Lifescience GmbH, Germany) targets gyrA, rrs and embB for fluoroquinolone, aminoglycoside and ethambutol resistance respectively [5]. Whilst these assays have significant value and allow rapid diagnosis of MDR-TB it is apparent that further improvements in molecular DSTs are required. MTB is largely clonal [6,7]. The TBDReaM database (http://www.tbdreamdb.com), catalogues greater than 1400 SNPs associated with drug resistance [8]. Studies of drug resistant MTB have shown that mutations in hot- spot regions, and those occurring beyond hypervariable regions vary according to both strain type and geographical location [9,10,11]. Thus, it is important for all possible resistance conferring mutations or single nucleotide polymorphisms (SNPs) related to each anti-tuberculous drug to be identified. Whole genome sequencing (WGS) facilitates the identification of SNPs, insertions and deletions in genomic DNA [12]. Thus, WGS of MTB can reveal new SNPs in genes previously associated with drug resistance and also identify outbreak strains [13].

By providing information on multiple genes simultaneously WGS facilitates the understanding of relationships between genes and their interactions such as in the case of fitness conferring compensatory mutations which occur due to the presence of drug induced selection pressure. Compensatory mutations are believed to play a role in conferring fitness to drug resistant MTB strains by affecting the growth rate of the isolates [14,15,16]. Compensatory mutations in rpoB genes which aid fitness of rifampicin resistant strains are well documented [17]. Knowledge of additional resistance mutation and compensatory mutations in MTB will be important for determining appropriate treatment for MTB strains. Here we have used WGS analysis to describe XDR-TB strains from Pakistan and identify SNPs in genes known to be associated with drug resistance to particular drugs.

Materials and Methods

Ethics statement

This study was approved by the Ethical Review Committee, The Aga Khan University.

Strain selection

The M. tuberculosis strains studied were those collected from 2004–2009. Mycobacterial strains were obtained from the strain bank of Aga Khan University Clinical Microbiology laboratory. Demographic information on patients was obtained from clinical laboratory records and collected as part of good clinical practice. The hospital and its clinical laboratory are accredited by the Joint Commission of International Accreditation (JCIA, USA). MTB isolates were identified through a random selection method. Each strain had its unique identification number which at the time of testing was anonymised and unlinked with patient demographic information. The XDR-TB (n = 37) and drug susceptible (n = 5) isolates studied here were spoligotyped previously [3].

Spoligotyping

Spoligotype patterns for each strain were obtained as described previously [18], and also confirmed in silico from the sequencing reads using SpolPred software [19].

Culture and drug susceptibility testing

M. tuberculosis strains were isolated from Lowenstein-Jensen media and MGIT (Becton Dickinson, Franklin Lakes, NJ, and USA). M. tuberculosis was identified using BACTEC NAP TB differentiation test (Becton Dickinson), growth on para-nitrobenzoic acid containing media, nitrate reduction, and niacin accumulation [20]. DST of these isolates was performed using an agar proportion method on enriched Middlebrook 7H10 medium (BBL Microbiology Systems, Cockeysville, MD, USA) at the following concentrations: rifampicin 1 μg/mL, isoniazid 0.2 μg/mL, streptomycin 2 μg/mL, and ethambutol 5 μg/mL. Pyrazinamide sensitivity was determined by using BACTEC 7H12 medium, pH 6.0, at 100 μg/mL (BACTEC PZA test medium, Becton Dickinson). MDR TB strains were further tested with ciprofloxacin 2 μg/mL, capreomycin 10 μg/mL, amikacin 5μg/mL, kanamycin 6μg/mL and ethionamide 5 μg/mL. The reference strain M. tuberculosis H37Rv was used as a control with each susceptibility testing batch [21].

DNA extraction and Whole Genome Sequencing of MTB isolates

DNA was extracted by the cetyl-trimethyl ammonium bromide (CTAB) method [22]. One microgram of DNA was used for sequencing. All samples underwent WGS with 76-base paired end fragment sizes, using Illumina paired end HiSeq2000 technology, and the raw sequence data is available in the European nucleotide archive (http://www.ebi.ac.uk/ena/data/view/PRJEB7798). For each sample trimmomatic software (http://www.usadellab.org/cms/?page=trimmomatic) was used to remove low quality reads and trim low-quality 3’ ends of reads. Nucleotide positions in the reads with a quality score lower than Q20 were removed. High quality reads were then mapped to the H37Rv reference genome (Genbank accession: AL123456.3) using BWA-MEM software (http://bio-bwa.sourceforge.net). SAMtools (http://samtools.sourceforge.net) and GATK (https://www.broadinstitute.org/gatk) were used to call single nucleotide polymorphisms (SNPs) and small indels. Variants (with quality at least Q30) were then selected as the intersection dataset between those obtained from both programs. Sample genotypes were called using the majority allele (minimum frequency 75%) in positions supported by at least 20-fold total coverage; otherwise they were classified as missing. Samples with a proportion of missing genotype calls greater than 15% were filtered out [21]. Similarly, we excluded positions in the genome with more than 15% missing genotypes across samples. Larger indels were called using a consensus from paired end mapping distance or split read approaches [23,24,25,26]. For more detail, the data processing pipeline has been described [27]. Spoligotypes were inferred from the sequencing reads SpolPred software [19]. Variation density maps were generated using Circos software (http://www.circos.com).

Single nucleotide polymorphisms (SNPs)

The SNP data obtained was analysed according to known resistance conferring genes for each anti-tuberculous drug. For rifampicin resistance, rpoB, rpoA and rpoC were analysed. For isoniazid resistance, katG, inhA,oxyR-ahpC, fpbC, Rv1592C, Rv1772, Rv2242, fabD, kasA, accD, oxyR, ndh, fadE24, and nat were investigated. For pyrazinamide resistance, the pncA and rpsA genes were studied. For ethambutol resistance, embC, embA, embB, embR, iniA, iniB, iniC, Rv3124, manB, PPE49 and rmlD genes were studied. For streptomycin resistance, rpsL, rrs and gidB genes were investigated. For flouroquinolone resistance, gyrA and gyrB genes were studied. For aminoglycoside resistance, rrs and tlyA, and for ethionamide resistance ethA and fabG1 genes were analysed. SNPs were divided into synonymous (sSNPs) and non-synonymous (ns-SNPs) mutations. Each SNP was compared with that listed in TBDReaMDB (http://www.tbdreamdb.com) database and other published reports of drug resistant MTB genotypes. For each strain, genotypic drug resistance findings were compared with phenotypic drug susceptibility results.

Phylogenetic analysis

Phylogenetic analysis was performed in PHYLIP using a RAxML tree (Maximum likelihood phylogenetic software) with bootstrap to 1000 replicated [28]. Phylogenetic tree was visualized using Dendroscope software [29].

Results

Demographic information, phenotypic drug susceptibility and genotypic profile

The XDR strains comprised mainly CAS family isolates: CAS1-Delhi (n = 20), CAS (n = 2) and CAS2 (n = 1). Other strains belonged to EAT3-IND (n = 2), X3 (n = 1), T1 (n = 3) and Orphan (n = 8) lineages. Drug susceptible isolates were CAS1-Delhi (n = 2) and CAS (n = 3) strains.

The XDR strains were from the four provinces of Pakistan; Sindh (n = 21), Punjab (n = 16), Khyber Pakhtunkhwa Province (n = 4) and Baluchistan (n = 1), S1 Dataset. Sixty-two percent (n = 26) of MTB isolates were from males and 32% (n = 16) from females. The mean and median ages of patients was 34.2 and 31 years respectively (males: mean, 36.2, median, 34.5; female: mean 31.1, median, 26 years).

Isolates from Sindh comprised Central Asian Strain (CAS), T1, East African Indian (EAI), X and Orphan types and thus had more diverse lineages as compared to those from other provinces. However, within each province the CAS lineage was predominant, as described previously [30].

Shared type spoligotypes and unique MTB isolates were linked to their respective principal genetic group (PGG) based on the combinations of katG463-gyrA95 polymorphism [31], S1 Dataset and S2 Dataset. Most of the strain types (90.5%, 38/42) except, X3 (n = 1) and T1 (n = 3) belonged to the PGG1 lineage (based on katG codon 463 CTG (Leu) and gyrA codon 95 ACC (Thr) polymorphisms). PGG1 strains included CAS, EAI and Orphan strains. One, X3 isolate was grouped in PGG2 as it showed katG 463 CGG (Arg) and gyrA codon 95 ACC (Thr) polymorphism. Three T1 isolates with katG 463 CGG (Arg) and gyrA codon 95 AGC (Ser) mutations were grouped in PGG3.

Polymorphisms in genes associated with drug resistances

A visual description of SNPs densities throughout the genome of MTB isolates was compiled by a circos plot (Fig. 1). The maximum likelihood analysis is demonstrated by the circular phlyogenetic tree based on the polymorphism in the drug resistance genes using Dendroscope software version 3.2.8 [32], (Fig. 2). The clustering pattern of the strains demonstrated that T1 and X lineage strains clustered together (PGG3 and PGG2 groups respectively), and were distinct from CAS and EAI strains which clustered in a similar node. However, EAI remained distant from other PGG1 isolates in the phylogenetic tree.

thumbnail
Fig 1. Comparative drug resistance genes information of Mycobacterium tuberculosis study isolates.

Circos plot of Mycobacterium tuberculosis highlighting the genomic regions (Rv) where drug resistance associated SNPs were identified (colour coded in blue). The 5 inwards tracks represent SNP density, Non-synonymous/synonymous densities, small insertion/deletions densities, large deletion density and uniqueness track.

https://doi.org/10.1371/journal.pone.0117771.g001

thumbnail
Fig 2. Circular phylogram based on the polymorphisms in the drug resistance genes of Mycobacterium tuberculosis isolates.

Maximum likelihood analyses were performed with single nucleotide polymorphism identified in the drug resistance genes using Dendroscope software, version 3.2.8. Scale bar indicates 1% nucleotide divergence in the genes. Branch length corresponds to number of genetic polymorphisms inferred.

https://doi.org/10.1371/journal.pone.0117771.g002

By comparison with the H37Rv reference genome, SNPs were identified in 40 genes associated with drug resistance (Fig. 3). All synonymous SNPs (sSNPs) were removed from the analysis so that only non-synonymous SNPs (nsSNPs) are displayed in S2 Dataset. This nsSNP dataset is further described in Tables 15, which focus on the known resistance-determining regions in particular MTB genes.

thumbnail
Fig 3. Process for identification of SNPs in genes associated with drug resistance.

The flow diagram describes the manner in which SNPs were identified for inclusion in the prepared Tables.

https://doi.org/10.1371/journal.pone.0117771.g003

thumbnail
Table 1. SNPs leading to rifampicin, isoniazid and pyrazinamide resistance in XDR strains.

https://doi.org/10.1371/journal.pone.0117771.t001

thumbnail
Table 3. Combination of SNPs in genes conferring resistance to streptomycin.

https://doi.org/10.1371/journal.pone.0117771.t003

thumbnail
Table 4. Combination of SNPs in genes conferring resistance to fluoroquinolone.

https://doi.org/10.1371/journal.pone.0117771.t004

thumbnail
Table 5. SNPs present in XDR strains related to aminoglycoside resistance.

https://doi.org/10.1371/journal.pone.0117771.t005

Polymorphisms conferring rifampicin and isoniazid resistance

Resistance to rifampicin is associated with mutations in the RNA polymerase gene sub-unit B, rpoB. However, there is a fitness cost associated with the acquistion of rpoB mutation and rifampicin resistant MTB strains often acquire compensatory mutations in rpoC and rpoA genes to their improve growth [33]. Similarly, there is a fitness cost associated with acquisition of isoniazid resistance, and while katG followed by the inhA gene mutations are most common, a number of other genes are thought to contribute to resistance [34]. We first evaluated SNPs in the context of MDR by studying the rifampicin-resistance determining region (RRDR) in rpoB, and compensatory mutations in rpoC and rpoA. For isoniazid resistance we analysed the katG and inhA genes, and also investigated sequences of oxyR-ahpC, fbpC, Rv1592C, ndh, Rv2242, fabD, accD, fad E24 and nat. Combinations of mutations observed in the XDR strains related to rifampicin and isoniazid are illustrated in Table 1. All the XDR isolates exhibited mutations in the RRDR of rpoB gene; codons 531 (n = 24), 516 (n = 12) and 526 (n = 1). The rpoB codon 531 mutation was present in PGG1, PGG2 and PGG3 isolates whereas, a mutation at codon 526 was observed in one PGG1 group EAI3-IND isolate. Mutations at codons 515 and 516 were present concurrently in each case, in CAS and Orphan strains.

Twenty-two XDR-TB isolates with compensatory mutations in rpoC and rpoA genes had the low fitness cost mutation rpoB S531L. One isolate with the rpoB H526R mutation had two rpoC SNPs. In all, we observed 15 different rpoC SNPs and one rpoA SNP. The most common was rpoC A172V which present in 5 isolates and rpoC P601L which was present in 3 isolates.

Polymorphisms at codon 463 of katG were observed among all the PGG1 isolates. Resistance to isoniazid could be attributed to the katG codon in 34/37 of cases and most strains (86%) had low fitness cost mutation, katG S315T while, two strains had katG S315N. Two strains had a inhA S94A mutation in the absence of katG mutations. We did not observe any inhA promoter mutations amongst the XDR isolates studied. One EAI3-lineage strain had a katG Y413H change.

In summary, if rifampcin and isoniazaid resistance were tested in these XDR strains using only MDR hot-spot region based targets (rpoB, katG and -15 region inhA), 100% of rifampicin and 89% of isoniazid resistance would be detected.

When additional isoniazid resistance associated genes were analysed we found the oxyR g142a nucleotide substitution to be present in 4 strains together with a katG S315T mutation (S2 Dataset). The EAI3 lineage strain (with katG Y413H) had the mutations oxyR g38a, accD M439I and A142V; in addition to rpoB H526R, and rpoC A172V and P601L mutations. We also observed additional SNPs present only in resistant strains with a katG S315T mutation; Rv1592C (G9D, n = 4), ndh L104F and E360K, Rv2242 (A156V), fabD (A3T), accD (E224D and A243V), fadE24 (Q85R), nat (T175A).

Polymorphisms conferring resistance to pyrazinamide

All of the XDR strains were resistant to pyrazinamide treatment (Table 1). Non-synonymous SNPs in the pncA gene were observed in 14 (38%) strains, whereby 11 different SNPs were observed (S2 Dataset). Twenty-three pyrazinamide resistant isolates did not reveal any nsSNP in pncA; two such strains had the rpsA mutation Q1228R.

Both pyrazinamide and isoniazid are pro-drugs which are converted to their active forms by enzymatic activity within the mycobacterium. We explore an association between isoniazid and pyrazinamide resistance by comparing katG and pncA mutations within strains. Of the XDR strains, thirteen had pncA and katG S315T mutations, while one strain with a pncA mutation had inhA S94A (Table 1).

Polymorphisms in the genes for ethambutol resistance

All of the XDR strains were resistant to ethambutol. The ethambutol-resistance determining region (ERDR) is primarily in the embB gene and we found that 23 strains to have embB codon 306 mutations, where M306I was most common (Table 2). Based on this data, commercial molecular diagnostic assays targeting embB codon 306 would detect ethambutol resistance in 62% of the XDR strains studied here.

Other ethambutol resistant strains had SNPs at embB codon 1024 (n = 4), codon 497 (n = 3), codon 406 (n = 1), codon 378 (n = 3), codon 328 (n = 1) and codon 72 (n = 1), S2 Dataset.

We analyzed additional genes associated with ethambutol resistance; embC, embA, Rv3124, rmlD, iniA, iniB, iniC and PPE49. Different combinations of SNPs were observed in these ethambutol-resistance associated genes possibly, as compensatory mutations and or lineage associated mutations. Two EAI3-IND isolates (X5 and X32) had a distinct combination of mutations; embB M306V, E378A and G406A, embA P913S, embC T270I, N394E, rmlD S257P and iniA H481Q (S2 Dataset). Two Orphan isolates (X17 and X57) also had a distinct combination of mutations; embB E378A, embA P913S, embC T270I, N394E, rmlD S257P, iniA H481Q, iniC R481C and PPE49 M67L. Another isolate had a SNP in Rv3124 (V273I) (S2 Dataset) while one strain displayed a single nucleotide (T) deletion at codon 173 of iniA gene at 519 position.

Polymorphisms conferring resistance to streptomycin

The mechanism of streptomycin resistance in the XDR strains was investigated by analyzing sequences of the ribosomal genes rpsL, rrs (500 region) and gidB. The rpsL mutation K43R was found in 41% of streptomycin resistant isolates; this SNP was found independently (n = 11) and also in combination with rrs A514C (n = 2), Table 3. The rpsL mutation K88R was found in combination with rrs A906G and gidB P75S. Three strains had rrs A514C mutations only. There were eight different SNPs in gidB. Eight streptomycin resistant strains had gidB mutations only, without rpsL and rrs (500 region) SNPs. However, six streptomycin resistant and five sensitive strains were without mutations in rrs, rpsL or gidB (S2 Dataset).

Polymorphisms in the genes for fluoroquinolone resistance

Mutations in gyrA and gyrB genes were analysed to investigate flouroquinolone resistance in the XDR strains. The most common mutation in the quinolone-resistance determining region (QRDR) of gyrA gene covering codons 90–98 was A94G, followed by A90V and A94Y (Table 4). All CAS family strains had gyrA QRDR mutations. One X3 strain had gyrA D94Y together with gyrB T500N. In total, 81% of isolates had SNPs in the QRDR of gyrA, which is the target region in currently available commercial diagnostic assays for flouroquinolone resistance.

Polymorphism at codon 95 of gyrA was observed among PGG1 and PGG2 isolates (S2 Dataset). The gyrA, G668D polymorphism was present in PGG3 (n = 3) isolates belonging to T1 lineage.

Of the eight strains which did not have gyrA QRDR SNPs, four had SNPs in the gyrB gene. One T1 strain only had the gyrB S447F mutation, two EAI3 strains only had the M291I mutation and an Orphan isolate had a gyrB M291I and A432V mutations. In one strain, a three nucleotide deletion was observed at codon 683 of gyrB gene at 7158 position. In all, four CAS1-Delhi XDR strains without gyrA or gyrB mutations.

Polymorphisms conferring resistance to aminoglycosides

To investigate the mechanism for aminoglycoside resistance codon 1400 region of the rrs gene was analysed. Thirteen XDR strains were resistant to all three aminoglycosides (amikacin, kanamycin and capreomycin). Twenty-four XDR strains were resistant to amikacin and kanamycin but sensisitive to capreomycin. The most common rrs mutation across lineages was A1401G. Of strains resistant to amikacin, kanamycin and capreomysin, eleven had the SNP rrs A1401G and one had G1484T, whilst 2 strains had no mutation in the 1400 region of rrs (Table 5). Of the strains resistant to amikacin and kanamycin, eighteen had the SNP A1401G whilst five did not have a SNP in the rrs 1400 region. As current commercial diagnostic assays for aminoglycoside resistance target rrs nt 1401, 78% of resistance would be detected in these XDR isolates.

In all, 6 aminoglycoside resistant isolates did not reveal any polymorphism in hyper-variable 1400 region of rrs gene (S2 Dataset).

Polymorphisms in the genes for ethionamide resistance

Five XDR isolates resistant to ethionamide but none showed any polymorphism in ethA gene, but one one ethionamide resistant isolate had the mutation ethA R239G which was also present in an ethionamide sensitive isolate (S2 Dataset).

Discussion

Our results present the first WGS based characterization of XDR-TB strains from Pakistan. We focused on SNPs in genes known to confer resistance to first- and second-line drugs and present data relevant to molecular diagnostic assays for drug susceptibility testing.

Consistent with published reports all the rifampicin resistant MTB isolates in this study revealed at least one rpoB gene codon mutation i.e. 531, 526, 516 or 515, in the RRDR region [35,36,37,38]. The most common mutation was the rpoB S531L, corroborating with previous reports [33]. In 22/23 isolates, compensatory mutations in rpoC and rpoA genes were coincident with the low fitness cost rpoB531 mutation, supporting previous reports which demonstrate increased fitness in rifampicin resistant isolates as a consequence of these compensatory mutations [39,40,41]. One EAI lineage isolate had the rpoB H526R mutation together with two rpoC mutations. As the rpoB codon 526 mutation has been shown to have a higher fitness cost than the codon 531 mutation [17] therefore, it would be an advantage for such an isolate to accumulate the rpoC compensatory mutations.

Based on the katG463-gyrA95 polymorphism 91% of the MTB isolates belonged to PGG1 of which the majority were CAS family isolates, as previously reported [30]. PGG1 is considered the oldest of MTB lineages and has been previously been reported from Asia [42]. The predominance of CAS strains suggests that they are well adapted to human populations in this region [43,44,45,46].

The most frequently observed isoniazid resistance associated mutation observed in our strains was katG S315T, which has previously been shown to be associated with a growth advantage in strains [47]. Three isolates with inhA codon 94 mutations did not reveal katG315 mutations, supporting a causal relationship of this mutation for isoniazid resistance [48,49].

Pyrazinamide is important as a first-line anti-tuberculous drug but phenotypic and genotypic discordance exists regarding testing its drug susceptibility. Consistent with previous reports, we found that more than 50% of resistant isolates did not reveal any mutation in pncA gene [50]. Mutations in pncA gene are thought to be the primary mechanism for pyrazinamide resistance, but our data consistent with others indicates that other resistance mechanisms may exist [51]. We found rpsA mutations in two XDR isolates which did not have pncA mutation, suggesting that this gene may play a role in resistance to pyrazinamide [52]. Additional pyrazinamide resistant strains had neither pncA nor rpsA mutations, and further supports the suggestion that mechanisms such an efflux pump for pncA wild type resistant MTB isolates may be relevant [53].

The embB306 SNP was found to be the most common mutation associated with ethambutol resistance in MTB strains, as reported previously [54]. The embB306 mutation has been shown to be associated with high-level ethambutol resistance [55]. In addition, we also observed embB codon 378, 406 and 497 mutations; which have reported previously [56] [55,57]. Our observations of distinct combinations of embB and compensatory mutations in XDR isolates of EAI-IND and Orphan types suggest that some of these may lineage associated mutations. Especially, we observed a particular pattern of embB, embA, embC, rmlD and ini mutations was in four Orphan types, of which three were resistant and one was sensitive to ethambutol.

The most common streptomycin resistance associated mutation we observed was rspL K43R. This mutation has been shown to play a role in reduced fitness cost and enhanced growth rate of resistant MTB isolates [47]. The K43R mutation has also been associated with high-level streptomycin resistance [58]. In the rrs gene we identified point mutations at positions 514 and 906, encoding for streptomycin resistance [58,59]. Previous studies have reported role of gidB polymorphism for low level of streptomycin resistance in MTB [60]. Here we found eight different gidB SNPs. Three strains had compensatory gidB SNPs with rpsL and/or rrs mutations. Six streptomycin resistant isolates had gidB SNPs but did not have either rpsL or rrs gene mutations. The occurrence of gidB mutations in strains reduces drug pressure for selection of highly resistant isolates, therefore it has been suggested that high-level streptomycin resistant strains (e.g. rpsL mutants) arise more frequently in gidB mutants than in wild-type cells [61]. We observed six streptomycin resistant strains without mutations in rpsL, rrs or gidB. The aminoglycoside antibiotics inhibit protein translation therefore it is likely that in additional mutations may be relevant for inactivating the effect of streptomycin in inhibiting tRNA translocation in mycobacteria.

In the strains resistant to amikacin, kanamycin and capreomycin we found 16S rRNA gene rrs1401 and 1484 mutations to be the most common. These rrs mutations have previously been associated kanamycin and capreomycin [39,56,62]. In addition, it has been shown that the tlyA and rrs gene mutations jointly play a role for capreomycin resistance [63]. However, we did not observe any tlyA mutations amongst our strains.

We found that across MTB lineages, the most common fluoroquinolone resistance associated mutation was at gyrA codon 94, as shown previously [56,64]. Lineages other than CAS (X3, T1, EAI and Orphan) had gyrB SNPs in the 400 region. The gyrB gene has previously been shown to result in flouroquinolone resistance [65]. Four XDR strains did not have gyrA or gyrB mutations. This may be explained by the fact that additional mechanisms such as efflux pumps may be involved in a generating resistance flouroquinolone [66]. The predominant lineage amongst our XDR isolates was CAS and there is less representation of the other strain types. However, this data is suggestive of the association of flouroquinolone resistance in CAS lineages with gyrA QRDR region.

In the case of ethionamide, we found that none of the isolates had ethA SNPs which were particular to ethionamide resistance, as one isolate had ethA R239G mutation which was also present in an ethionamide sensitive isolate. A low frequency of mutation in ethA gene of ethambutol resistant isolates have also been reported earlier and may suggest existence of some other mechanisms such as efflux pump for ethionamide resistance [56,67].

Overall, the most frequently seen drug resistance mutations observed for isoniazid, rifampicin, streptomycin, ethambutol, aminoglycosides and ofloxacin were katG315, rpoB531, rpsL43, embB306, rrs1401 and gyrA94 respectively as previously reported [35,36,37,38,55,56,68,69]. Currently available commercial assays for detection of drug resistance (MTBDRplus and MTBDRsl) target the rpoB531, 516 and 526, katG315, gyrA94, 91 and 90 and rrs1401 regions. Based on our data the frequency of drug resistance detection using such assays for these strains would have been 100% for rifampicin resistance, 92% for isoniazid resistance, 81% for fluoroquinolone resistance, 78% for aminoglycoside and 62% for ethambutol resistance. Our data indicates that SNPs in drug resistance genes vary with different MTB lineages. Therefore, it highlights the relevance for WGS data from prevalent global strain types in order to develop improved molecular diagnostic assays for resistance in MTB. Additional knowlegde of mutations that are compensatory to those that confer drug resistance will be important for prediction of levels of drug resistance and therefore, to guide treatment of drug resistance isolates.

Supporting Information

S1 Dataset. Drug susceptibility and demographic information of Mycobacterium tuberculosis isolates.

https://doi.org/10.1371/journal.pone.0117771.s001

(DOCX)

S2 Dataset. An overview of SNPs observed in genes associated with drug resistance in XDR TB isolates.

https://doi.org/10.1371/journal.pone.0117771.s002

(XLSX)

Acknowledgments

We thank the Aga Khan University Hospital Clinical Laboratory and its staff for providing samples used in the study. This work was made possible through the generous support for whole genome sequencing provided by Dr. Arnab Pain at KAUST, Saudi Arabia.

Author Contributions

Conceived and designed the experiments: RH RM ZH. Performed the experiments: AA GHC MN. Analyzed the data: FC KM TGC AA ZH. Contributed reagents/materials/analysis tools: RM AP FC KM TGC. Wrote the paper: RH ZH AA TGC KM.

References

  1. 1. WHO (2014) Global Tuberculosis Report. Geneva: World Health Organization. Available: http://wwwwhoint/tb/publications/global_report/en/. p. pmid:25637818
  2. 2. Ayaz A, Hasan Z, Jafri S, Inayat R, Mangi R, et al. (2013) Characterizing Mycobacterium tuberculosis isolates from Karachi, Pakistan: drug resistance and genotypes. Int J Infect Dis 16: e303–309.
  3. 3. Hasan R, Jabeen K, Ali A, Rafiq Y, Laiq R, et al. (2010) Extensively drug-resistant tuberculosis, Pakistan. Emerg Infect Dis 16: 1473–1475. pmid:20735937
  4. 4. Steingart KR, Sohn H, Schiller I, Kloda LA, Boehme CC, et al. (2013) Xpert(R) MTB/RIF assay for pulmonary tuberculosis and rifampicin resistance in adults. Cochrane Database Syst Rev 1: CD009593. pmid:23440842
  5. 5. Hillemann D, Rusch-Gerdes S, Richter E (2009) Feasibility of the GenoType MTBDRsl assay for fluoroquinolone, amikacin-capreomycin, and ethambutol resistance testing of Mycobacterium tuberculosis strains and clinical specimens. J Clin Microbiol 47: 1767–1772. pmid:19386845
  6. 6. Supply P, Warren RM, Banuls AL, Lesjean S, Van Der Spuy GD, et al. (2003) Linkage disequilibrium between minisatellite loci supports clonal evolution of Mycobacterium tuberculosis in a high tuberculosis incidence area. Mol Microbiol 47: 529–538. pmid:12519202
  7. 7. Hirsh AE, Tsolaki AG, DeRiemer K, Feldman MW, Small PM (2004) Stable association between strains of Mycobacterium tuberculosis and their human host populations. Proc Natl Acad Sci U S A 101: 4871–4876. pmid:15041743
  8. 8. Sandgren A, Strong M, Muthukrishnan P, Weiner BK, Church GM, et al. (2009) Tuberculosis drug resistance mutation database. PLoS Med 6: e2. pmid:19209951
  9. 9. Borrell S, Espanol M, Orcau A, Tudo G, March F, et al. (2009) Factors associated with differences between conventional contact tracing and molecular epidemiology in study of tuberculosis transmission and analysis in the city of Barcelona, Spain. J Clin Microbiol 47: 198–204. pmid:19020067
  10. 10. Borrell S, Gagneux S (2011) Strain diversity, epistasis and the evolution of drug resistance in Mycobacterium tuberculosis. Clin Microbiol Infect 17: 815–820. pmid:21682802
  11. 11. Coll F, McNerney R, Guerra-Assuncao JA, Glynn JR, Perdigao J, et al. (2014) A robust SNP barcode for typing Mycobacterium tuberculosis complex strains. Nat Commun 5: 4812. pmid:25176035
  12. 12. Ford C, Yusim K, Ioerger T, Feng S, Chase M, et al. (2012) Mycobacterium tuberculosis—heterogeneity revealed through whole genome sequencing. Tuberculosis (Edinb) 92: 194–201. pmid:22218163
  13. 13. Walker TM, Ip CL, Harrell RH, Evans JT, Kapatai G, et al. (2012) Whole-genome sequencing to delineate Mycobacterium tuberculosis outbreaks: a retrospective observational study. Lancet Infect Dis 13: 137–146. pmid:23158499
  14. 14. Andersson DI, Hughes D (2010) Antibiotic resistance and its cost: is it possible to reverse resistance? Nat Rev Microbiol 8: 260–271. pmid:20208551
  15. 15. Gagneux S (2009) Fitness cost of drug resistance in Mycobacterium tuberculosis. Clin Microbiol Infect 15 Suppl 1: 66–68. pmid:19220360
  16. 16. Gagneux S, Long CD, Small PM, Van T, Schoolnik GK, et al. (2006) The competitive cost of antibiotic resistance in Mycobacterium tuberculosis. Science 312: 1944–1946. pmid:16809538
  17. 17. Mariam DH, Mengistu Y, Hoffner SE, Andersson DI (2004) Effect of rpoB mutations conferring rifampin resistance on fitness of Mycobacterium tuberculosis. Antimicrob Agents Chemother 48: 1289–1294. pmid:15047531
  18. 18. Kamerbeek J, Schouls L, Kolk A, van Agterveld M, van Soolingen D, et al. (1997) Simultaneous detection and strain differentiation of Mycobacterium tuberculosis for diagnosis and epidemiology. J Clin Microbiol 35: 907–914. pmid:9157152
  19. 19. Coll F, Mallard K, Preston MD, Bentley S, Parkhill J, et al. (2012) SpolPred: rapid and accurate prediction of Mycobacterium tuberculosis spoligotypes from short genomic sequences. Bioinformatics 28: 2991–2993. pmid:23014632
  20. 20. Koneman EW AS, Dowell VR, Sommers HM. (1983) Color Atlas and Text book of Diagnostic microbiology; 1983. Publisher; JB Lippincott Company, Philadelphia. pmid:25121320
  21. 21. Woods GL B-EB, Desmond EP, Hall GS, Heifets L, Pfyffer GE, et al. (2003) Susceptibility testing of mycobacteria, nocardiae, and other aerobic actinomycetes: approved standard Wayne, PA: NCCLS; 2003. Document No M24-A.
  22. 22. Honore-Bouakline S, Vincensini JP, Giacuzzo V, Lagrange PH, Herrmann JL (2003) Rapid diagnosis of extrapulmonary tuberculosis by PCR: impact of sample preparation and DNA extraction. J Clin Microbiol 41: 2323–2329. pmid:12791844
  23. 23. Chen K, Wallis JW, McLellan MD, Larson DE, Kalicki JM, et al. (2009) BreakDancer: an algorithm for high-resolution mapping of genomic structural variation. Nat Methods 6: 677–681. pmid:19668202
  24. 24. Rausch T, Zichner T, Schlattl A, Stutz AM, Benes V, et al. (2012) DELLY: structural variant discovery by integrated paired-end and split-read analysis. Bioinformatics 28: i333–i339. pmid:22962449
  25. 25. Wang J, Mullighan CG, Easton J, Roberts S, Heatley SL, et al. (2011) CREST maps somatic structural variation in cancer genomes with base-pair resolution. Nat Methods 8: 652–654. pmid:21666668
  26. 26. Ye K, Schulz MH, Long Q, Apweiler R, Ning Z (2009) Pindel: a pattern growth approach to detect break points of large deletions and medium sized insertions from paired-end short reads. Bioinformatics 25: 2865–2871. pmid:19561018
  27. 27. Coll F, Preston M, Guerra-Assuncao JA, Hill-Cawthorn G, Harris D, et al. PolyTB: a genomic variation map for Mycobacterium tuberculosis. Tuberculosis (Edinb) 94: 346–354. pmid:24637013
  28. 28. Stamatakis A (2014) RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30: 1312–1313. pmid:24451623
  29. 29. Huson DH, Richter DC, Rausch C, Dezulian T, Franz M, et al. (2007) Dendroscope: An interactive viewer for large phylogenetic trees. BMC Bioinformatics 8: 460. pmid:18034891
  30. 30. Tanveer M, Hasan Z, Siddiqui AR, Ali A, Kanji A, et al. (2008) Genotyping and drug resistance patterns of M. tuberculosis strains in Pakistan. BMC Infect Dis 8: 171. pmid:19108722
  31. 31. Sreevatsan S, Pan X, Stockbauer KE, Connell ND, Kreiswirth BN, et al. (1997) Restricted structural gene polymorphism in the Mycobacterium tuberculosis complex indicates evolutionarily recent global dissemination. Proc Natl Acad Sci U S A 94: 9869–9874. pmid:9275218
  32. 32. Huson DH, Scornavacca C (2012) Dendroscope 3: an interactive tool for rooted phylogenetic trees and networks. Syst Biol 61: 1061–1067. pmid:22780991
  33. 33. Brandis G, Hughes D (2013) Genetic characterization of compensatory evolution in strains carrying rpoB Ser531Leu, the rifampicin resistance mutation most frequently found in clinical isolates. J Antimicrob Chemother 68: 2493–2497. pmid:23759506
  34. 34. Jagielski T, Bakula Z, Roeske K, Kaminski M, Napiorkowska A, et al. (2014) Detection of mutations associated with isoniazid resistance in multidrug-resistant Mycobacterium tuberculosis clinical isolates. J Antimicrob Chemother 69: 2369–2375. pmid:24855126
  35. 35. Ajbani K, Rodrigues C, Shenai S, Mehta A (2011) Mutation detection and accurate diagnosis of extensively drug-resistant tuberculosis: report from a tertiary care center in India. J Clin Microbiol 49: 1588–1590. pmid:21289142
  36. 36. Ali A, Hasan R, Jabeen K, Jabeen N, Qadeer E, et al. (2011) Characterization of Mutations Conferring Extensive Drug Resistance to Mycobacterium tuberculosis Isolates in Pakistan. Antimicrob Agents Chemother 55: 5654–5659. pmid:21911575
  37. 37. Ali A, Hasan Z, Moatter T, Tanveer M, Hasan R (2009) M. tuberculosis Central Asian Strain 1 MDR isolates have more mutations in rpoB and katG genes compared with other genotypes. Scand J Infect Dis 41: 37–44. pmid:19012077
  38. 38. Campbell PJ, Morlock GP, Sikes RD, Dalton TL, Metchock B, et al. (2011) Molecular detection of mutations associated with first- and second-line drug resistance compared with conventional drug susceptibility testing of Mycobacterium tuberculosis. Antimicrob Agents Chemother 55: 2032–2041. pmid:21300839
  39. 39. Casali N, Nikolayevskyy V, Balabanova Y, Ignatyeva O, Kontsevaya I, et al. (2012) Microevolution of extensively drug-resistant tuberculosis in Russia. Genome Res 22: 735–745. pmid:22294518
  40. 40. Comas I, Borrell S, Roetzer A, Rose G, Malla B, et al. (2011) Whole-genome sequencing of rifampicin-resistant Mycobacterium tuberculosis strains identifies compensatory mutations in RNA polymerase genes. Nat Genet 44: 106–110. pmid:22179134
  41. 41. de Vos M, Muller B, Borrell S, Black PA, van Helden PD, et al. (2013) Putative compensatory mutations in the rpoC gene of rifampin-resistant Mycobacterium tuberculosis are associated with ongoing transmission. Antimicrob Agents Chemother 57: 827–832. pmid:23208709
  42. 42. Arjomandzadegan M, Titov LP, Surkova LK, Farnia P, Sheikholeslami F, et al. (2012) Determination of principal genotypic groups among susceptible, MDR and XDR clinical isolates of Mycobacterium tuberculosis in Belarus and Iran. Tuberk Toraks 60: 153–159. pmid:22779936
  43. 43. Gagneux S, DeRiemer K, Van T, Kato-Maeda M, de Jong BC, et al. (2006) Variable host-pathogen compatibility in Mycobacterium tuberculosis. Proc Natl Acad Sci U S A 103: 2869–2873. pmid:16477032
  44. 44. Shang S, Harton M, Tamayo MH, Shanley C, Palanisamy GS, et al. (2011) Increased Foxp3 expression in guinea pigs infected with W-Beijing strains of M. tuberculosis. Tuberculosis (Edinb) 91: 378–385. pmid:21737349
  45. 45. de Steenwinkel JE, ten Kate MT, de Knegt GJ, Kremer K, Aarnoutse RE, et al. (2012) Drug susceptibility of Mycobacterium tuberculosis Beijing genotype and association with MDR TB. Emerg Infect Dis 18: 660–663. pmid:22469099
  46. 46. Ioerger TR, Feng Y, Chen X, Dobos KM, Victor TC, et al. (2010) The non-clonality of drug resistance in Beijing-genotype isolates of Mycobacterium tuberculosis from the Western Cape of South Africa. BMC Genomics 11: 670. pmid:21110864
  47. 47. Spies FS, von Groll A, Ribeiro AW, Ramos DF, Ribeiro MO, et al. (2013) Biological cost in Mycobacterium tuberculosis with mutations in the rpsL, rrs, rpoB, and katG genes. Tuberculosis (Edinb) 93: 150–154. pmid:23276692
  48. 48. Vilcheze C, Wang F, Arai M, Hazbon MH, Colangeli R, et al. (2006) Transfer of a point mutation in Mycobacterium tuberculosis inhA resolves the target of isoniazid. Nat Med 12: 1027–1029. pmid:16906155
  49. 49. Valvatne H, Syre H, Kross M, Stavrum R, Ti T, et al. (2009) Isoniazid and rifampicin resistance-associated mutations in Mycobacterium tuberculosis isolates from Yangon, Myanmar: implications for rapid molecular testing. J Antimicrob Chemother 64: 694–701. pmid:19710078
  50. 50. Stoffels K, Mathys V, Fauville-Dufaux M, Wintjens R, Bifani P (2012) Systematic analysis of pyrazinamide-resistant spontaneous mutants and clinical isolates of Mycobacterium tuberculosis. Antimicrob Agents Chemother 56: 5186–5193. pmid:22825123
  51. 51. Zhang Y, Wade MM, Scorpio A, Zhang H, Sun Z (2003) Mode of action of pyrazinamide: disruption of Mycobacterium tuberculosis membrane transport and energetics by pyrazinoic acid. J Antimicrob Chemother 52: 790–795. pmid:14563891
  52. 52. Tan Y, Hu Z, Zhang T, Cai X, Kuang H, et al. (2013) Role of pncA and rpsA gene sequencing in detection of pyrazinamide resistance in Mycobacterium tuberculosis isolates from southern China. J Clin Microbiol 52: 291–297. pmid:24131688
  53. 53. Sheen P, Lozano K, Gilman RH, Valencia HJ, Loli S, et al. (2013) pncA gene expression and prediction factors on pyrazinamide resistance in Mycobacterium tuberculosis. Tuberculosis (Edinb) 93: 515–522. pmid:23867321
  54. 54. Sirgel FA, Warren RM, Streicher EM, Victor TC, van Helden PD, et al. (2012) embB306 mutations as molecular indicators to predict ethambutol susceptibility in Mycobacterium tuberculosis. Chemotherapy 58: 358–363. pmid:23171739
  55. 55. Srivastava S, Garg A, Ayyagari A, Nyati KK, Dhole TN, et al. (2006) Nucleotide polymorphism associated with ethambutol resistance in clinical isolates of Mycobacterium tuberculosis. Curr Microbiol 53: 401–405. pmid:16972132
  56. 56. Liu CH, Li HM, Lu N, Wang Q, Hu YL, et al. (2012) Genomic sequence based scanning for drug resistance-associated mutations and evolutionary analysis of multidrug-resistant and extensively drug-resistant Mycobacterium tuberculosis. J Infect.
  57. 57. Chan RC, Hui M, Chan EW, Au TK, Chin ML, et al. (2007) Genetic and phenotypic characterization of drug-resistant Mycobacterium tuberculosis isolates in Hong Kong. J Antimicrob Chemother 59: 866–873. pmid:17360809
  58. 58. Jnawali HN, Yoo H, Ryoo S, Lee KJ, Kim BJ, et al. (2013) Molecular genetics of Mycobacterium tuberculosis resistant to aminoglycosides and cyclic peptide capreomycin antibiotics in Korea. World J Microbiol Biotechnol.
  59. 59. Meier A, Kirschner P, Bange FC, Vogel U, Bottger EC (1994) Genetic alterations in streptomycin-resistant Mycobacterium tuberculosis: mapping of mutations conferring resistance. Antimicrob Agents Chemother 38: 228–233. pmid:8192448
  60. 60. Okamoto S, Tamaru A, Nakajima C, Nishimura K, Tanaka Y, et al. (2007) Loss of a conserved 7-methylguanosine modification in 16S rRNA confers low-level streptomycin resistance in bacteria. Mol Microbiol 63: 1096–1106. pmid:17238915
  61. 61. Verma JS, Gupta Y, Nair D, Manzoor N, Rautela RS, et al. (2014) Evaluation of gidB alterations responsible for streptomycin resistance in Mycobacterium tuberculosis. J Antimicrob Chemother 69: 2935–2941. pmid:25074855
  62. 62. Sirgel FA, Tait M, Warren RM, Streicher EM, Bottger EC, et al. (2011) Mutations in the rrs A1401G gene and phenotypic resistance to amikacin and capreomycin in Mycobacterium tuberculosis. Microb Drug Resist 18: 193–197. pmid:21732736
  63. 63. Maus CE, Plikaytis BB, Shinnick TM (2005) Mutation of tlyA confers capreomycin resistance in Mycobacterium tuberculosis. Antimicrob Agents Chemother 49: 571–577. pmid:15673735
  64. 64. Sirgel FA, Warren RM, Streicher EM, Victor TC, van Helden PD, et al. (2012) gyrA mutations and phenotypic susceptibility levels to ofloxacin and moxifloxacin in clinical isolates of Mycobacterium tuberculosis. J Antimicrob Chemother 67: 1088–1093. pmid:22357804
  65. 65. Yin X, Yu Z (2010) Mutation characterization of gyrA and gyrB genes in levofloxacin-resistant Mycobacterium tuberculosis clinical isolates from Guangdong Province in China. J Infect 61: 150–154. pmid:20452372
  66. 66. El Garch F, Lismond A, Piddock LJ, Courvalin P, Tulkens PM, et al. (2010) Fluoroquinolones induce the expression of patA and patB, which encode ABC efflux pumps in Streptococcus pneumoniae. J Antimicrob Chemother 65: 2076–2082. pmid:20709735
  67. 67. Wang K, Pei H, Huang B, Zhu X, Zhang J, et al. (2013) The expression of ABC efflux pump, Rv1217c-Rv1218c, and its association with multidrug resistance of Mycobacterium tuberculosis in China. Curr Microbiol 66: 222–226. pmid:23143285
  68. 68. Dalla Costa ER, Ribeiro MO, Silva MS, Arnold LS, Rostirolla DC, et al. (2009) Correlations of mutations in katG, oxyR-ahpC and inhA genes and in vitro susceptibility in Mycobacterium tuberculosis clinical strains segregated by spoligotype families from tuberculosis prevalent countries in South America. BMC Microbiol 9: 39. pmid:19228426
  69. 69. Spies FS, von Groll A, Ribeiro AW, Ramos DF, Ribeiro MO, et al. (2012) Biological cost in Mycobacterium tuberculosis with mutations in the rpsL, rrs, rpoB, and katG genes. Tuberculosis (Edinb) 93: 150–154.