The International Journal of Mycobacteriology

: 2020  |  Volume : 9  |  Issue : 2  |  Page : 121--137

Determination of potentially novel compensatory mutations in rpoc associated with rifampin resistance and rpob mutations in Mycobacterium tuberculosis Clinical isolates from peru

Ana Paula Vargas1, Angela A Rios1, Louis Grandjean2, Daniela E Kirwan3, Robert H Gilman4, Patricia Sheen1, Mirko J Zimic1,  
1 Department of Cellular and Molecular Sciences, Laboratory of Bioinformatics and Molecular Biology, Faculty of Science and Philosophy, Universidad Peruana Cayetano Heredia, Lima, Peru
2 Department of Paediatric Infectious Diseases, Imperial College, University of London, London, England
3 Infection and Immunity Research Institute, St. George's, University of London, London, England
4 Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA

Correspondence Address:
Mirko J Zimic
Laboratory of Bioinformatics and Molecular Biology, Faculty of Science and Philosophy, Universidad Peruana Cayetano Heredia, Av. Honorio Delgado 430, San Martin de Porres, Lima


Background: Rifampicin (RIF) resistance in Mycobacterium tuberculosis is frequently caused by mutations in the rpoB gene. These mutations are associated with a fitness cost, which can be overcome by compensatory mutations in other genes, among which rpoC may be the most important. We analyzed 469 Peruvian M. tuberculosis clinical isolates to identify compensatory mutations in rpoC/rpoA associated with RIF resistance. Methods: The M. tuberculosis isolates were collected and tested for RIF susceptibility and spoligotyping. Samples were sequenced and aligned to the reference genome to identify mutations. By analyzing the sequences and the metadata, we identified a list of rpoC mutations exclusively associated with RIF resistance and mutations in rpoB. We then evaluated the distribution of these mutations along the protein sequence and tridimensional structure. Results: One hundred and twenty-five strains were RIF susceptible and 346 were resistant. We identified 35 potential new compensatory mutations, some of which were distributed on the interface surface between rpoB and rpoC, arising in clusters and suggesting the presence of hotspots for compensatory mutations. Conclusion: This study identifies 35 putative novel compensatory mutations in the β' subunit of M. tuberculosis RNApol. Six of these (S428T, L507V, A734V, I997V, and V1252LM) are considered most likely to have a compensatory role, as they fall in the interaction zone of the two subunits and the mutation did not lead to any change in the protein's physical–chemical properties.

How to cite this article:
Vargas AP, Rios AA, Grandjean L, Kirwan DE, Gilman RH, Sheen P, Zimic MJ. Determination of potentially novel compensatory mutations in rpoc associated with rifampin resistance and rpob mutations in Mycobacterium tuberculosis Clinical isolates from peru.Int J Mycobacteriol 2020;9:121-137

How to cite this URL:
Vargas AP, Rios AA, Grandjean L, Kirwan DE, Gilman RH, Sheen P, Zimic MJ. Determination of potentially novel compensatory mutations in rpoc associated with rifampin resistance and rpob mutations in Mycobacterium tuberculosis Clinical isolates from peru. Int J Mycobacteriol [serial online] 2020 [cited 2020 Aug 7 ];9:121-137
Available from:

Full Text


Tuberculosis (TB) is the leading cause of death by a single infectious agent worldwide.[1] With an average of 27,000 new cases per year, Peru has the second highest rate of TB in the Americas; thanks to a national TB control program,[2] the incidence of TB has decreased since 2000 by 1.6% per year.[3]

Rifampicin (RIF) is one of the most important drugs in the treatment of TB, and resistance to RIF threatens the success of TB control programs globally.[4] In 2017, the WHO reported more than 159,000 cases of RIF-resistant TB including 1,508 in Peru.[5] In the same year, about 390,000 new cases of multidrug-resistant TB (MDR-TB) were reported worldwide.[3] In the preceding decade, the average number of reported cases per year in Peru exceeded 1,100, with a trend toward an increase in the later years.[6]

RIF causes bacterial cell death by binding to the β-subunit of the bacteria's RNA polymerase (RNApol), blocking RNA transcription, and inhibiting gene expression.[7],[8],[9],[10] Mutations in rpoB, the gene encoding the Mycobacterium tuberculosis RNApol β-subunit, have been shown to be the main cause of RIF-resistant microorganisms.[11] Moreover, the RIF resistance-determining region (RRDR) is an 81 bp region in rpoB within which 95% of all known RIF-resistant mutations are found.[11] Many studies have examined the structural variations of rpoB mutants and their association with drug binding.[12]Escherichia coli, distinct species of Salmonella, Thermus, and Mycobacterium have been the most studied models for RIF resistance to date.

Mutations in rpoB have been found to be associated with reduced fitness of the affected strain. Several studies have demonstrated that the fitness loss due to rpoB mutations can be compensated for by other mutations, which may consequently be positively selected. These mutations have been mainly identified in the rpoC and/or rpoA genes which are responsible for coding the β' and α-subunits of the RNApol complex, respectively, and which interact closely with rpoB.[8],[10],[13],[14],[15] Several compensatory[13],[14],[16],[17],[18],[19] and putative compensatory[20],[21],[22],[23] mutations in these two genes have been described, mostly in rpoC, and including the V483A and L516P mutations that have been extensively described in M. tuberculosis.[16],[19],[20],[23]

In this study, we analyzed rpoA, rpoB, and rpoC gene sequences from 469 M. tuberculosis clinical strains, recovered from unrelated TB patients from Peru, to identify compensatory mutations in rpoC/rpoA genes associated with RIF resistance. Characterization of these mutations will improve our understanding of the mycobacterial compensatory mechanisms that allow them to overcome the fitness cost associated with rpoB mutations and thereby enable the survival and ongoing transmission of RIF-resistant strains. Furthermore, this study could provide insight into evolutionary processes, leading to RIF resistance in TB.


Selection of strains and rifampicin-susceptibility testing

We selected 469 well-characterized MTB strains, isolated from unrelated TB patients in hospitals from Callao and South Lima, Peru, between 2009 and 2013. These were collected and selected as part of two large studies that have been described previously:[24] one was a population-level study and the other a household follow-up study. The MODS assay was used to confirm TB infection and determine RIF susceptibility. MODS is a highly sensitive (98%) and specific (98.6%) tool that can both diagnose TB and detect drug resistance by determining the presence or absence of bacterial growth under a fixed drug concentration, and is approved for routine clinical use.[24],[25],[26]

The genotypes of these isolates have previously been studied using spoligotyping,[24] which characterizes mycobacterial strains according to their hybridization patterns.[24],[27] Thus, in this study, we also evaluated the association between RIF resistance and the spoligotype.

Sequences and mutations

The rpoA, rpoB, and rpoC gene sequences, and mutations in these genes compared to the wild-type reference strain H37Rv (GeneBank: NC_000962) were obtained from the complete genome sequences attained previously.[24] Briefly, samples were sequenced on the Illumina MiSeq platform (150 bp paired-end reads) and aligned to the reference H37Rv genome using standard methodology.

Mutations in rpoA and rpoC associated with rifampicin-resistant strains

The 469 strains were classified as RIF susceptible or RIF resistant according to their MODS result. Each group was subclassified based on the presence or absence of at least one mutation in the rpoB gene. Each of these four groups was further subclassified based on the presence or absence one or more mutations in rpoA or rpoC.

To identify possible compensatory mutations in rpoC, we selected the RIF-susceptible strains with mutations in rpoC and a wild-type rpoB gene, regardless of the presence or absence of mutations in rpoA (set A). We also selected the RIF-resistant strains harboring mutations in rpoC and any mutation in rpoB, regardless of rpoA status (set B). After removing all rpoC mutations from set B that were also present in set A, we obtained a list of rpoC mutations exclusively associated with RIF resistance and mutations in rpoB. The same approach was applied to mutations in rpoA.

rpoA and rpoC compensatory mutations reported elsewhere

A literature search was performed to identify any mutations in rpoC or rpoA that have previously been associated with RIF resistance and rpoB mutations and have been shown experimentally to be compensating for a loss of fitness.[13],[14],[16],[17],[18],[19],[20],[21],[22],[23] These were compiled into a list, based on which we were able to classify the mutations identified in this study as either previously reported or as potentially novel.

Distribution of potentially novel compensatory rpoC mutations along the protein sequence

The distribution of rpoC mutations was represented in a linear histogram, which indicated an equitable amino acid spread along the complete sequence of the protein. The β' subunit (coded by rpoC) sequence comprises 1,300 amino acids, and the histogram was built on blocks of 100 amino acids. Previously reported rpoC mutations and mutations identified in this study were mapped separately. Peruvian mutations that had already been reported elsewhere were marked. This analysis was not performed for mutations in rpoA because of insufficient mutations, either previously reported or identified in this study.

Distribution of potentially compensatory rpoC mutations on the three-dimensional structure of Mycobacterium tuberculosis RNA polymerase

To better understand the mechanism of action of the potentially compensatory rpoC mutations, these were mapped onto the recently published crystal structure of the M. tuberculosis RNApol transcription initiation complex[28] (PDB ID: 5UHD). The structures were visualized and analyzed using VMD.[29] First, all of the putative compensatory rpoC mutations were mapped onto the structure. Then, the mutations unique to this study were differentiated from those reported previously.

We visualized the distribution of the mutations in the β' subunit (coded by rpoC) that fell within the region of the surface that interacts with the β subunit (coded by rpoB). To identify the residues of physical importance in the β subunit, we identified those that arise within 5.0–6.0 Š of these rpoC mutations.


Determination of mutations in rpoB, rpoA, and rpoC

Among the 469 genomes analyzed [Table 1], we found 125 (27%) to be RIF sensitive and 346 (74%) to be RIF resistant according to the MODS assay [Figure 1]. We schematically distributed the strains according to their sensitivity to RIF, mutations in rpoB, and mutations in rpoC/rpoA [Figure 1]. 237 (50.5%) strains harbored mutation (s) in rpoC and/or rpoA. 322 (93.1%) RIF-resistant strains harbored at least one mutation in rpoB.{Table 1}{Figure 1}

175 (54%) strains that were RIF resistant and harbored mutations in rpoB had at least one mutation in rpoA or rpoC, suggesting that these may be compensatory mutations. 41 (41.8%) RIF-susceptible strains with no mutations in rpoB had at least one mutation in rpoC or rpoA. The presence of these 41 mutations may suggest the occurrence of random mutations, not necessarily associated with bacterial fitness.

Strikingly, 11 RIF-susceptible strains of different spoligotypes harbored rpoB mutations in the critical region (RRDR) [Table 1]. These mutations represent 3.4% of the 327 strains harboring a mutation in the RRDR. Five of these 11 strains also harbored the G594E mutation in rpoC [Table 2]. Notably, G594E appears only in the Haarlem genotype among these strains. This suggests that a very small percentage (3.4%) of mutations in the RRDR may not be associated with RIF resistance.{Table 2}

Rifampicin susceptibility and genotyping

The distribution of RIF-resistant strains (i.e., the RIF-resistant/RIF-susceptible strains ratio) was similar (1.4–4.6) within the different spoligotype clades [Figure 2]. This suggests that our investigation is not biased toward a specific genotype. LAM was the predominant genotype (54.4%), followed by Haarlem (17.3%), other small clades and T (10.03%), Beijing (7.3%), Caprae (0.64%), and finally Bovis and EAI2-Manilla (0.21%).{Figure 2}

Mutations in rpoA and rpoC associated with rifampicin resistant strains

The first group generated for analysis [set A, [Supplementary Table 1] comprised a total of 43 strains, while the second group [set B, Supplementary [Table 2] comprised 176 strains. Both groups had a variety of rpoC mutations, as depicted in [Figure 3].[INLINE:1]{Figure 3}

Interestingly, G594E appeared in both groups, being related to strains associated with RIF resistance and mutations in rpoB, but also to strains with a wild-type rpoB and RIF susceptibility. Other studies also found this mutation and excluded it from their analyses.[14],[16],[17],[20],[22],[23]

Forty-four mutations found exclusively in set B [bold in [Figure 3] were considered to be potentially compensatory. Nine (22%) of these mutations have been previously described [Supplementary Table 3]: N698K,[17] I491V,[17] H525Q,[17] P1040R,[17] L516P,[16],[19] V1252M,[17] V483A,[16],[19] V483G,[16],[17],[19] and V1252 L.[17] The remaining 35 (78%) were novel.[INLINE:2][INLINE:3]{Table 3}

Only three mutations in the rpoA gene were found (T187A, G31S, and V183G), all of which have been reported elsewhere.[16],[17],[19] Our literature search identified a total of 29 compensatory mutations in rpoA.

Distribution of the potentially compensatory mutations along the rpoC protein sequence

A high frequency of mutations occurred between amino acids 400–500, 500–600, and 1000–1100 (a total of 21 mutations) [Figure 4], suggesting the existence of hotspots. 47% of all the potentially compensatory mutations proposed in this study arose in these regions. E49A and P54 L were isolated from the rest, located within the first 100 amino acids of the protein. The distribution of the mutations found here and those reported previously was found to be remarkably similar (P = 0.2318, Kolmogorov–Smirnov nonparametric test).{Figure 4}

To better analyze this distribution, we classified the amino acids into the following groups: basic (R, K, and H), acidic (D and E), polar (G, S, Y, C, Q, T, and N), and nonpolar (F, L, W, P, I, M, V, and A). Based on this, we found that 54% of the potentially compensatory mutations found in rpoC preserve their physical–chemical properties despite the amino acid change.

Distribution of the potential compensatory rpoC mutations on the three-dimensional structure of Mycobacterium tuberculosis RNApol

The potentially compensatory mutations that we described above [Supplementary Table 3] were mainly distributed on the surface of the three-dimensional (3D)-crystal structure of the β' subunit (82%), and 9 were buried below the surface [Figure 5]. Only one of the buried mutated sites, H525Q, has been previously reported to be compensatory.[17] E49A, which was visibly separated from the rest in the primary structure of rpoC, can be seen on the side of the subunit 3D-structure, distant from the main cluster of mutations [Figure 5].{Figure 5}

β' RNApol has a domain that resembles the shape of an “arm” that hosts a particularly large number of mutations. The “arm” is best visible on the right side of the structure [Figure 6]. Our data show that some of the putative compensatory mutations fall in similar spots as those that have been previously reported, in particular those located in this “arm”. Interestingly, the reported mutation P54 L appears hidden below the surface along with E49A [Figure 6]. We also found that two of the proposed mutations, G332R and W484S, occur in regions where other mutations have already been reported.{Figure 6}

The 3D-crystal structure of β' interacting with the β subunit confirms that several mutations fall in the regions of interaction (S428T, L507V, A734V, I997V, and V1252 LM) [Figure 7]. These represent 13.5% of the surface mutations. All of the putative compensatory mutations that we identified in the rpoC/rpoB contact surface conserved their physical–chemical properties during the amino acid change [Figure 7].{Figure 7}


This study describes 35 novel mutations in rpoC, associated with RIF resistance and rpoB mutations, detected in Peruvian clinical isolates of M. tuberculosis. These mutations may serve to compensate for any loss of fitness caused by a mutation in rpoB, as reported in other studies.[8],[10],[13],[14],[15] Of note, some were found to be associated with controversial mutations in rpoB that have been identified in RIF-susceptible MTB strains.

The similarity in the distribution of the potentially compensatory rpoC mutations found in Peru compared to those reported elsewhere is striking. The spatial distribution along the protein sequence reveals a similar, almost mirrored pattern, suggesting that compensatory mutations in rpoC preferentially occur in certain positions. These cluster-like regions could be related to functional aspects of the protein or may be structurally critical. According to the Pfam protein domain family database,[30],[31],[32] these high-frequency mutation zones correspond to specific domains. The first zone (amino acids 400–500, second domain) contains the protein's active site, the second zone (amino acids 500–600, third domain) contains the pore thought to act as a channel for nucleotides, and the third zone (amino acids 1000–1100, fifth domain) includes the discontinuous cleft required to form the channel where DNA is bound. Mutations occurring in these critical regions may cause significant changes in function.

Surface mutations are more likely to interact with other subunits at a quaternary structure level or with other proteins. Furthermore, a surface mutation would not significantly affect the tertiary structure of the protein, whereas a deep mutation may destabilize the whole subunit. Thus, it would be favorable for compensatory mutations to be located on the surface of the protein, even more on the interaction sites. In our study, the position of the novel mutations on the crystal structure revealed that 82.2% fall on the surface of the β' subunit.

Moreover, only six mutations were found in the region of interaction between the surfaces of β and β', which constitute our most likely candidates for compensation, as a mutation at this site would produce a conformational change sufficient to affect the binding between the subunits but not significant enough to disrupt the inner core. Two of these six mutations have already been reported, and although W484S was not found in the literature, W484G/C has in fact been previously described. Moreover, as noted before and according to our expectations that mutations would result in minimal (but significant) structural changes, every mutation that we reported as a candidate conserves the physical–chemical properties of the original amino acid, favoring the maintenance of the protein's stability.[33] Several mutations found in this study are very close to mutations reported elsewhere, and in some cases, even at the same position. This may also indicate the preference of certain locations for the emergence of compensatory mutations.

Only 20.5% of the mutations that we identified coincided with previously reported compensatory mutations.[16],[17],[19] This is likely to be due to the large genotypic diversity among Peruvian M. tuberculosis strains. Of note, in our study, most of the isolates (250) corresponded to the “LAM” spoligotype.

We identified just three mutations in rpoA, which is thought to be, along with rpoC, the most important site of RIF-resistant compensatory mutations. All three have been previously reported. This is in keeping with previous findings that the majority of compensatory mutations arise in rpoC,[16],[17],[19] suggesting that these may be more effective at recovering loss of fitness arising from mutations in rpoB.[16],[21],[22],[23]

Furthermore, we found mutations in rpoC that were not associated with any mutations in rpoB but were present in RIF-resistant strains. It is likely that the appearance of these mutations is random, but further studies are necessary to clarify their true nature.

We reported 11 RIF-susceptible strains harboring mutations in the RRDR, with 5 harboring rpoC G594E [Table 2]. This is unlikely to be due to incorrect DST results, as MODS is a highly sensitive (98%) and specific (98.6%) method.[26] Rather, this finding led us to further analyze our data in search of patterns in the distribution of rpoC G594E. We found it to be present in 17.3% of RIF-resistant and 39.8% of RIF-susceptible strains, being by far the most common mutation. G594E has been described in strains with no rpoB mutations and is a SNP present in all strains from a group within the Haarlem genotype, which would mean that it arises independently of rpoB mutations and RIF resistance. Taken together, these features make it unlikely that G594E is associated with drug resistance, as broadly stated before.[14],[16],[17],[20],[22],[23]

In view of this, we found G594E in both of our analysis groups: set A (rpoC mutations in RIF-susceptible strains without rpoB mutations) and set B (rpoC mutations in RIF-resistant strains harboring rpoB mutations). Furthermore, we found it in the five previously mentioned RIF-susceptible strains with mutations in the RRDR [Table 2]. Three of these harbor rpoB L452P, described elsewhere.[19],[34] The remaining two harbor rpoB S450M, which is similar to the widely-known S450 L[10],[17],[18],[19],[34] and their described variants S450F/Q/Y/W.[19],[21],[34],[35],[36],[37],[38] The detection of rpoB S450M is puzzling because a Ser to Met mutation would require changes in two nucleotides (TCG → ATG), making it unlikely to occur spontaneously. It is possible that S450F occurred first (TCG → TTG), followed by a change from TTG → ATG, resulting in the end of a Ser to Met change. It is notable that both amino acids are nonpolar; that is, they have similar physicochemical properties. We could not find any other study describing S450M, so we encourage further investigation of the role and evolution of this mutation, as our theory of its development is mere speculation.

Nonetheless, we cannot avoid discussing the dichotomy regarding the appearance of rpoC G594E. Considering all of the above, we propose another possibility toward its appearance, which is that under certain conditions some mutations in rpoB do not cause RIF resistance but may still decrease the fitness of the bacteria. This would subdue the bacteria to selective pressure, leading to the positive selection of a compensatory mutation in rpoC. Until now, most investigations have attributed the unlikeliness of G594E being associated with drug resistance to its uncharacteristic appearance, but believe that there may be other ways for this to happen and thus we cannot draw any conclusions about the true nature of this mutation until more is known.

Our study identifies mutations that may have a compensatory role, but there is a need for experimental confirmation of the effect of each mutation as some of these are likely to arise at random. Previous studies have used methods such as the measurement of the catalytic activity and rate of bacterial growth.[14],[15],[16],[17],[18] We suggest that the use of site-directed mutagenesis techniques, such as the emerging CRISPR-Cas9 technology,[39] would be useful to assess the fitness of mutant strains.


This study identifies 35 putative novel compensatory mutations in the β' subunit of M. tuberculosis RNApol. Six of these (S428T, L507V, A734V, I997V, and V1252 LM) are considered most likely to have a compensatory role, as they fall in the interaction zone of the two subunits and the mutation did not lead to any change in the protein's physical–chemical properties. Further studies assessing the fitness change in strains harboring these mutations are needed to confirm their potential compensatory role. In addition, it would be interesting to evaluate the evolutionary relationship of the various SNPs found in this study.

Financial support and sponsorship

This research was funded by the Wellcome Trust (Ref: 099805/Z/12/Z), the Grand Challenge Canada (GCC Number 0687-01-10), and the LOREAL-UNESCO-CONCYTEC 2014 award.

PS was supported by a Wellcome Trust Intermediate fellowship.

Conflicts of interest

There are no conflicts of interest.


1World Health Organization. World Health Statistics 2016: Monitoring Health for the SDGs, Sustainable Development Goals. Geneva: World Health Organization. Available from: [Last updated on 2020 Jan 13; Last accessed on 2019 Jan 20].
2Silva C, Bermúdez V, Arraiz N, Bermúdez F, Rodríguez M, Valdelamar L, et al. Fármacos de primera línea utilizados en el tratamiento de la tuberculosis. Arch Venez Farmacol Y Ter 2007;26:21-6.
3World Health Organization. Global Tuberculosis Report. Geneva: World Health Organization Available from: [Last updated on 2020 Jan 13; Last accessed on 2019 Jan 20].
4World Health Organization. Tuberculosis. Available from: [Last updated on 2020 Jan 13; Last accessed on 2019 Jan 20]
5World Health Organization. Drug resistant TB-Data by WHO Region. Available from: [Last updated on 2020 Jan 13; Last accessed on 2019 Jan 20]
6Ministry of Health. Lima: Analysis of the Epidemiological Situation in Peru. Available from: [Last updated on 2020 Jan 13; Last accessed on 2019 Jan 20]
7Mendoza L, Rodríguez H, Rodríguez E, Tarazona C. Conocimientos básicos sobre antibacterianos. 1st ed. Maracaibo: Editorial Astro Data; 1999. p. 73-5.
8Koch A, Mizrahi V, Warner DF. The impact of drug resistance on Mycobacterium tuberculosis physiology: What can we learn from rifampicin? Emerg Microbes Infect 2014;3:e17.
9Hughes D, Brandis G. Rifampicin resistance: Fitness costs and the significance of compensatory evolution. Antibiotics (Basel) 2013;2:206-16.
10Alifano P, Palumbo C, Pasanisi D, Talà A. Rifampicin-resistance, rpoB polymorphism and RNA polymerase genetic engineering. J Biotechnol 2015;202:60-77.
11Telenti A, Imboden P, Marchesi F, Lowrie D, Cole S, Colston MJ, et al. Detection of rifampicin-resistance mutations in Mycobacterium tuberculosis. Lancet 1993;341:647-50.
12Goldstein BP. Resistance to rifampicin: A review. J Antibiot (Tokyo) 2014;67:625-30.
13Brandis G, Hughes D. Genetic characterization of compensatory evolution in strains carrying rpoB Ser531Leu, the rifampicin resistance mutation most frequently found in clinical isolates. J Antimicrob Chemother 2013;68:2493-7.
14Song T, Park Y, Shamputa IC, Seo S, Lee SY, Jeon HS, et al. Fitness costs of rifampicin resistance in Mycobacterium tuberculosis are amplified under conditions of nutrient starvation and compensated by mutation in the β' subunit of RNA polymerase. Mol Microbiol 2014;91:1106-19.
15Mariam DH, Mengistu Y, Hoffner SE, Andersson DI. Effect of rpoB mutations conferring rifampin resistance on fitness of Mycobacterium tuberculosis. Antimicrob Agents Chemother 2004;48:1289-94.
16Comas I, Borrell S, Roetzer A, Rose G, Malla B, Kato-Maeda M, et al. Whole-genome sequencing of rifampicin-resistant Mycobacterium tuberculosis strains identifies compensatory mutations in RNA polymerase genes. Nat Genet 2011;44:106-10.
17Cohen KA, Abeel T, Manson McGuire A, Desjardins CA, Munsamy V, Shea TP, et al. Evolution of extensively drug-resistant tuberculosis over four decades: Whole genome sequencing and dating analysis of Mycobacterium tuberculosis isolates from kwazulu-natal. PLoS Med 2015;12:e1001880.
18Brandis G, Wrande M, Liljas L, Hughes D. Fitness-compensatory mutations in rifampicin-resistant RNA polymerase. Mol Microbiol 2012;85:142-51.
19Zhang H, Li D, Zhao L, Fleming J, Lin N, Wang T, et al. Genome sequencing of 161 Mycobacterium tuberculosis isolates from China identifies genes and intergenic regions associated with drug resistance. Nat Genet 2013;45:1255-60.
20Ali A, Hasan Z, McNerney R, Mallard K, Hill-Cawthorne G, Coll F, et al. Whole genome sequencing based characterization of extensively drug-resistant Mycobacterium tuberculosis isolates from Pakistan. PLoS One 2015;10:e0117771.
21Casali N, Nikolayevskyy V, Balabanova Y, Ignatyeva O, Kontsevaya I, Harris SR, et al. Microevolution of extensively drug-resistant tuberculosis in Russia. Genome Res 2012;22:735-45.
22de Vos M, Müller B, Borrell S, Black PA, van Helden PD, Warren RM, et al. Putative compensatory mutations in the rpoC gene of rifampin-resistant Mycobacterium tuberculosis are associated with ongoing transmission. Antimicrob Agents Chemother 2013;57:827-32.
23Li QJ, Jiao WW, Yin QQ, Xu F, Li JQ, Sun L, et al. Compensatory mutations of rifampin resistance are associated with transmission of multidrug-resistant Mycobacterium tuberculosis Beijing genotype strains in China. Antimicrob Agents Chemother 2016;60:2807-12.
24Grandjean L, Gilman RH, Iwamoto T, Köser CU, Coronel J, Zimic M, et al. Convergent evolution and topologically disruptive polymorphisms among multidrug-resistant tuberculosis in Peru. PLoS One 2017;12:e0189838.
25Caviedes L, Lee TS, Gilman RH, Sheen P, Spellman E, Lee EH, et al. Rapid, efficient detection and drug susceptibility testing of Mycobacterium tuberculosis in sputum by microscopic observation of broth cultures. The Tuberculosis Working Group in Peru. J Clin Microbiol 2000;38:1203-8.
26Minion J, Leung E, Menzies D, Pai M. Microscopic-observation drug susceptibility and thin layer agar assays for the detection of drug resistant tuberculosis: A systematic review and meta-analysis. Lancet Infect Dis 2010;10:688-98.
27Kamerbeek J, Schouls L, Kolk A, van Agterveld M, van Soolingen D, Kuijper S, et al. Simultaneous detection and strain differentiation of Mycobacterium tuberculosis for diagnosis and epidemiology. J Clin Microbiol 1997;35:907-14.
28Lin W, Mandal S, Degen D, Liu Y, Ebright YW, Li S, et al. Structural Basis of Mycobacterium tuberculosis Transcription and Transcription Inhibition. Mol Cell 2017;66:169-790.
29Humphrey W, Dalke A, Schulten K. VMD: Visual molecular dynamics. J Mol Graph 1996;14:33-8, 27-8.
30Finn RD, Coggill P, Eberhardt RY, Eddy SR, Mistry J, Mitchell AL, et al. The Pfam protein families database: Towards a more sustainable future. Nucleic Acids Res 2016;44:D279-85.
31Severinov K, Mustaev A, Kukarin A, Muzzin O, Bass I, Darst SA, et al. Structural modules of the large subunits of RNA polymerase. Introducing archaebacterial and chloroplast split sites in the beta and beta' subunits of Escherichia coli RNA polymerase. J Biol Chem 1996;271:27969-74.
32Cramer P, Bushnell DA, Kornberg RD. Structural basis of transcription: RNA polymerase II at 2.8 angstrom resolution. Science 2001;292:1863-76.
33Ladunga I, Smith RF. Amino acid substitutions preserve protein folding by conserving steric and hydrophobicity properties. Protein Eng. Des. Sel. 1997;10:187-96.
34Casali N, Broda A, Harris SR, Parkhill J, Brown T, Drobniewski F. Whole genome sequence analysis of a large isoniazid-resistant tuberculosis outbreak in London: A Retrospective Observational Study. PLoS Med 2016;13:e1002137.
35Salamon H, Yamaguchi KD, Cirillo DM, Miotto P, Schito M, Posey J, et al. Integration of published information into a resistance-associated mutation database for Mycobacterium tuberculosis. J Infect Dis 2015;211 Suppl 2:S50-7.
36Yang C, Luo T, Shen X, Wu J, Gan M, Xu P, et al. Transmission of multidrug-resistant Mycobacterium tuberculosis in Shanghai, China: A retrospective observational study using whole-genome sequencing and epidemiological investigation. Lancet Infect Dis 2017;17:275-84.
37Ullah I, Ahmad W, Shah AA, Shahzada A, Tahir Z, Qazi O, et al. Detection of rifampicin resistance of Mycobacterium tuberculosis using multiplex allele specific polymerase chain reaction (MAS-PCR) in Pakistan. Infect Genet Evol 2019;71:42-6.
38Miotto P, Cabibbe AM, Borroni E, Degano M, Cirillo DM. Role of Disputed Mutations in the rpoB Gene in Interpretation of Automated Liquid MGIT Culture Results for Rifampin Susceptibility Testing of Mycobacterium tuberculosis. J Clin Microbiol 2018;56:e01599-17.
39Rock JM, Hopkins FF, Chavez A, Diallo M, Chase MR, Gerrick ER, et al. Programmable transcriptional repression in mycobacteria using an orthogonal CRISPR interference platform. Nat Microbiol 2017;2:16274.