RpoB is the target of RIF, a derivative of rifamycin. RpoB catalyzes the transcription of DNA into mRNA by using the four ribonucleoside triphosphates as substrates. In M. tuberculosis RIF binds with β-subunit of the RNA polymerase, encoded by rpoB, and inhibits the elongation of messenger RNA, thus interfering with transcription (Carlos and Martin, 2013; Piccaro et al., 2014). RpoB has a gene length of 3519 bp and conformational changes caused by mutations in rpoB can lead to RIF-resistance (Telenti et al., 1993). Ninety-six percent of RIF-resistance occurs within the “hot-spot region” (of 81 bp), also known as RIF-resistance determining region (RRDR), covering codons 507–533 of rpoB gene (Ramaswamy et al., 2003). Several studies have reported mutations in codons 516, 526 and 531 are most commonly found in RIF-resistant isolates (Ocheretina et al., 2014; Thirumurugan et al., 2015).

In particular, mutation in the rpoB gene at 531 codon (serine to leucine) is highly significant as it confers cross-resistance to rifabutin (Mboowa et al., 2014; Thirumurugan et al., 2015; Aye et al., 2016). whereas mutations at codons 516, 518, 526, and 529 are associated with low-level resistance to RIF and conserved susceptibility to other rifamycins, e.g., rifabutin or rifalazil (Cavusoglu et al., 2004; Tan et al., 2012). However, compensatory mutations were also identified in rpoA and rpoC encoding respectively for α and β' subunits of RNA polymerase (Comas et al., 2012). These compensatory mutations could be important for reinstating the fitness and emergence of MDR strains, and their transmissibility in vivo (Brandis and Hughes, 2013). Further studies are required to better understand their specific roles.

Mutations in katG and inhA are the main cause of resistance to INH in M. tuberculosis. KatG has a molecular mass of 80,572.8 Da and gene size of 2,223 bp while InhA has 28,527.8 Da molecular mass and 810 bp gene size, respectively. katG encodes multifunctional enzymes that exhibit both catalase-peroxidase and peroxynitritase activities, the former being important for activating the prodrug, and the latter in association with pathways involving reactive nitrogen and oxygen intermediates (Zhang et al., 1992). On the other hand, AhpC, SodC, KatG, and TpX are recognized as essential for virulence of the M. tuberculosis (Forrellad et al., 2013). NADH-dependent enoyl-acyl carrier protein (ACP)-reductase, encoded by inhA, is involved in the biosynthesis of mycolic acids and participates in the second reductive step in fatty acid biosynthesis (Rozwarski et al., 1998; Vilchèze et al., 2006). Mutational changes in these two genes, katG and inhA, are mainly associated with the resistance mechanism of INH (Ramaswamy et al., 2003).

Conversely, mutations in Rv0340-0343, fadE24, efpA, and kasA were identified in both INH-resistant and INH-susceptible strains, so their association with INH-resistance requires further clarification (Vilchèze et al., 2007). Among the katG mutations, S315T is considered as the most pervasive mutation that accounts for 40~94% resistance in MDR strains and results in reducing the ability of KatG to convert INH into isonicotinic acid, a precursor for the formation of INH-NAD adduct (Bantubani et al., 2014; Seifert et al., 2015; Aye et al., 2016). In addition, another newly reported katG mutation L101R identified in clinical isolates, changing an hydrophobic leucine to hydrophilic arginine, possibly involved in alteration of the conformation of binding protein near the active site and inhibited its bio-activation (Datta et al., 2016).

The second main cause of resistance to INH is due to mutations in the promoter region of inhA, resulting in the overexpression of inhA. The most frequently observed mutation in inhA regulatory region at position −15C/T is more generally associated with low level resistance to INH (MIC < 1 μg/ml) (Banerjee et al., 1994; Fenner et al., 2012; Aye et al., 2016). Moreover, double mutations at −8T/C, −15/T (Zhang and Yew, 2015), and −17C/T (Müller et al., 2011) in the inhA promoter region were also found to be associated with INH-resistance. Some recent studies have described that a mutation in the regulatory region of inhA, together with a mutation in inhA coding region, results in high-level resistance against INH as well as cross-resistance against the structurally related ethionamide (Machado et al., 2013).

EmbB and EmbC have molecular masses of 118,021 Da, and 117,490 Da, respectively, and gene sizes of 3,297 bp, 3,285 bp, respectively. Mutations in embB and embC cause resistance against ethambutol [EMB; dextro-2,2′-(ethylenediimino)-DI-1-butanol], through restricting the action of drug to cease the biosynthesis of mycobacterial cell wall. In M. tuberculosis, the genes embCAB, are organized as an operon that encodes arabinosyl transferases, involved in the synthesis of arabinogalactan. (Mikusová et al., 1995). EMB plays intrusive role in biosynthesis of arabinogalactan in the cell wall to convert multiplying bacilli into a bacteriostatic phase (Wang et al., 2010). Mutations in the embCAB operon were described to cause alterations in the drug-protein binding site (Xu et al., 2015). Numerous studies concluded that a mutation at codon306 in embB was the cause of EMB-resistance (Shi et al., 2011a; Yoon et al., 2013; Moure et al., 2014).

It was also reported that mutations in genes of the decaprenylphosphoryl-beta-D-arabinose (DPA) biosynthesis and utilization pathway genes, Rv3806c and Rv3792, combined with mutations in embB and embC genes increased the MICs ranges of EMB, depending on mutation type and number (Shi et al., 2011a; Safi et al., 2013). A strong association was also noticed between the Met306Val and Met306Leu at emb306 and EMB-resistance (Zhang et al., 2014), which suggests the usefulness of the embB306 mutation to serve as potential molecular markers for EMB resistance. However, in 30% EMB-resistant M. tuberculosis strains there was no mutation in embB, alluding to the need to intensively search for other mechanisms of EMB resistance (Shen et al., 2007; Perdigão et al., 2009). Interestingly, in another study, 70% mutations were found in 306, 406 or 497 codons, 13% mutations outside the three regions between codons 296 and 426 and 15% mutations in the embC-embA intergenic region among the total 98% of mutations in the embCAB locus (Brossier et al., 2015). Recently, mutations in embB along with mutations in ubiA, encoding for a decaprenylphosphoryl-5-phosphoribose (DPPR) synthase associated with arabinogalactan synthesis pathway, were associated with high level of resistance to EMB (Tye et al., 2015).

Pyrazinamide (PZA) is an important front-line anti-TB drug. In M. tuberculosis, resistance to PZA is mainly associated with pncA, rpsA, and panD. However, some other proteins such as ppsA encoding polyketide synthase involved in phthiocerol dimycocerosate (PDIM) synthesis, cell division protein FtsH, TetR family transcriptional regulator (3R)-hydroxyacyl-ACP dehydratase subunit HadC, phosphate ABC transporter permease protein PstC2 and transmembrane transport protein MmpL4 found to be responsible for PZA resistance (Zhang et al., 2017). Pyrazinamidase (Pzase), encoded by the 561-bp pncA, converts the pro-drug PZA into pyrazinoic acid (POA, its active form). PZA resistance in M. tuberculosis is most commonly associated with mutations in the whole open reading frame or 82–262 bp regulatory region of pncA gene (Juréen et al., 2008).

POA interrupts the bacterial membrane activities. POA and its n-propyl ester may have the ability to constrain the fatty acid synthase type I in replicating M. tuberculosis bacilli, ineffective efflux pump also may result in accumulation of protonated POA during acidic conditions resulting in bacterial cell damage (Scorpio and Zhang, 1996; Zimhony et al., 2007; Shi et al., 2011b, 2014; Njire et al., 2016). The 481-amino acid-long 30S ribosomal protein S1 is encoded by rpsA. It plays a significant role in mRNA translation (involving a short shine-dalgarno (SD) purine-rich sequence), and in trans-translation, a unique process that utilizes transfer-messenger RNA (tmRNA) molecule to rescue stalled ribosomes. POA was also associated with disruption of the trans-translation process through binding to the 30S ribosomal protein S1 (RpsA) (Shi et al., 2011b). In the work of Shi et al. (2011b), deletion of alanine, resulting from a GCC deletion at 438 bp (C-terminus) of rpsA, was found in PZA-resistant strains lacking mutations in pncA gene. Such mutation was thought to induce resistance against POA and, accordingly, increased level of resistance were observed after overexpression of rpsA (Shi et al., 2011b).

Later, many studies presented disparate results about the mutations in rpsA in both PZA-resistant and PZA-susceptible isolates (Bhuju et al., 2013; Maslov et al., 2015). However, only synonymous mutations were reported in rpsA (A636C and G960A) for both PZA-resistant and PZA-susceptible clinical strains (Alexander et al., 2012). In another study, mutations in rpsA gene were found in 3/52 PZA-resistant and 1/108 PZA-susceptible clinical isolates (Tan et al., 2014). This absence of clinical isolates presenting rpsA mutations may indicate a strong fitness cost associated with such mutations, although more investigations are required to fully understand the link between RpsA and the activity of PZA. The third gene panD, of 420 bp in size, is putatively involved in pantothenate biosynthesis through converting L-aspartate into beta-alanine (Shi et al., 2014). Some studies found mutations in panD gene of PZA-resistant isolates lacking mutations in pncA and rpsA (Pandey et al., 2016).

Among the three genes identified so far as potentially linked with resistance to PZA, i.e., pncA, rpsA, and panD, the most frequent mutations (>80%) were found to occur in the pncA gene (Xia et al., 2015; Xu et al., 2016). Recently, a new target of PZA clpC1 (Rv3596c) was identified, encodes an ATP-dependent ATPase which is involved in protein degradation by forming a complex with protease ClpP1 and ClpP2 (Zhang et al., 2017). ClpC1 is an 848-amino acids containing protein and the mutation (G296T) change amino acid G99D cause resistance to PZA. Interestingly, mutations in clpC1 has also been reported as a resistance factor against three new cyclic peptide antibiotics including cyclomarin A (Schmitt et al., 2011), lassomycin at the N-terminal repeat region at Q17R, R21S, and P79T (Gavrish et al., 2014) and ecumicin at L92S or F or L96P sites (Gao et al., 2015). The three new drug candidates appear to bind at the different sites of the ClpC1. The gene Rv2783c encoding a bifunctional enzyme was indicated as a new target very recently (Njire et al., 2017). Rv2783 was proved to be able not only to catalyze metabolism of RNA and single-stranded DNA, but also to metabolize ppGpp, an important signal transducer involved in the stringent response in bacteria.

The rpsL gene is very small, only 375-bp, while rrs gene is much larger, with 1,537 bp but both are associated with resistance to STR. The rpsL encodes 30S ribosomal protein S12 RpsL involved in the initiation step of RNA translation, while rrs encodes 16S rRNA. Mutations in rpsL and rrs are known to cause resistance against STR (Finken et al., 1993), the first anti-TB drug discovered. STR is an aminoglycoside and targets the actively growing bacteria by hindering the translation of protein synthesis (Chakraborty et al., 2013). More precisely, STR interacts with formyl-methionyl-tRNA to bind the 30S subunit of the ribosome at the ribosomal protein S12 encoded by rpsL gene and also at 16S rRNA encoded by rrs gene, preventing normal functioning of the ribosome (Sharma et al., 2007).

The most commonly reported mutation in rpsL is the replacement of lysine into arginine at positions 43 and 88. Regarding rrs, mutations around nucleotides 530–915 make M. tuberculosis strains highly resistant to STR (Jagielski et al., 2014; Zhao et al., 2014). Additionally, a 675 bp gene, gidB, encoding for a conserved 7-methylguanosine methyltransferase specific for the 16S rRNA, was found to confer intermediate-level of resistance toward STR by conferring the A80P mutation to the gene product GidB (Perdigão et al., 2014). Generally, mutations in the rpsL gene covers around 80% of STR-resistance (Jnawali et al., 2013). More precisely, the mutations at codons 43 and 88 in rpsL gene were identified in more than 60% of STR-resistance cases, while 17% contained rrs mutations and 14% were observed with mutations in gidB, in a panel of 161 clinical isolates (Smittipat et al., 2016).

Some STR-resistant M. tuberculosis strains also conferred cross-resistance with kanamycin (KAN) and amikacin (AMK). The mutation at the codon A1401G in the rrs gene in particular was associated with high-level resistance to KAN and AMK, along with cross-resistance to capreomycin (CAP). Whereas mutations C1402T or G1484T were also associated with CAP resistance in addition cross-resistance with KAN or viomycin (VIM) (Du et al., 2013; Sowajassatakul et al., 2014). Similarly, mutations at position 1,400, 1,401, and 1,483 bp in the rrs gene were linked with high-level resistance to both KAN and AMK in KAN-resistant strains (Ajbani et al., 2011; Yuan et al., 2012). Moreover, mutations in the promoter region of the whiB7 gene, with a role in transcriptional mechanisms, were shown to be associated with resistance to an array of antibiotics, and increased expression of the efflux enhanced by atpE and the eis promoter region was also observed (Reeves et al., 2013).

Further, eis and tlyA (1,209 bp and 807 bp in size, respectively) were also found to have putative roles in resistance. eis encodes an aminoglycoside acetyltransferase and mutation at position −10 and −35 of the eis promoter has been associated with the low-level resistance to KAN (Ajbani et al., 2011; Yuan et al., 2012; Du et al., 2013; Reeves et al., 2013; Sowajassatakul et al., 2014). tlyA encodes rRNA methyltransferase, specific for 2′-O-methylation of ribose in rRNA. Expression of tlyA gene (commonly A1408G) affects both drug susceptibility and fitness cost of drug resistance through methylation mechanisms (Freihofer et al., 2016). However, another study did not find any mutation in the tlyA gene (Du et al., 2013). Besides, overexpression of Rv0148 protein play a vital role in three-fold increase in MIC of AMK and two-fold of KAN (Sharma et al., 2015). While overexpression of ferritin (Rv3841/bfrB) showed an alternate behavior, there was two-fold increase in MIC of AMK and three-fold of KAN (Sharma et al., 2016). Regarding to above discussed genes and their role in development of resistance, it is important to note that KAN and AMK are aminoglycosides and CAP and VIM are cyclic peptide antibiotics, they have similar structures and bind on the same location in the ribosome, at the interface of the small and large subunits (Stanley et al., 2010). The similarity of the mechanisms of actions and their resistance development simplifies understanding to some extent.

gyrA and gyrB are the two main genes with mutations linked to resistance against fluoroquinolones (FQ's). gyrA is 2517-bp long while gyrB gene is 2,028-bp. Mutations in the quinolone resistance-determining region (QRDR) of gyrA (74–113 codons) and gyrB (codon 500–540) in M. tuberculosis are largely responsible for generating the resistance to FQ's (Pantel et al., 2012; Nosova et al., 2013). Two subunits α and β each of gyrA and gyrB, that encode topoisomerase II (DNA gyrase) catalyzing the supercoiling of DNA, furnishes the most important target of FQ action (Aubry et al., 2004). The most commonly found mutations at position Ala-74, Gly-88 Ala-90, Ser-91, and Asp-94 of gyrA generally result in high-level resistance against FQ (Li et al., 2014).

A natural polymorphism at position 95 (Ser or Thr) in gyrA is not related to FQ's-resistance since it is also found in FQ-sensitive strains (Nosova et al., 2013). Usually, mutations in gyrB (10–15%), being less commonly found among clinical isolates, induce low-level resistance to FQ's (Cui et al., 2011). However, the synchronized mutations in both gyrA and gyrB, such as Ala543Val (gyrB)-Asp94Asn/Asp94Gly (gyrA) and Asn538lle (gyrB)-Asp94Ala (gyrA) persuaded very high resistance against FQ's (Long et al., 2012). An interested study identified that presence of mutations T80A and A90G in gyrA led to hypersusceptibility to several quinolones (Aubry et al., 2006). Similarly, most of the mutations in the gyrA Asn538Asp and Asp500His in gyrB are shown to be associated with cross-resistance among the FQs, whereas in gyrB the mutation Arg485His does not confer any resistance (Nosova et al., 2013). One intriguing study found a M. tuberculosis strain with Asn533Thr mutation in gyrB showing susceptibility to ofloxacin (OFX) but resistance to moxifloxacin (MOX) and gatifloxacin (Von Groll et al., 2009).

Proteomic comparative study enlightened the 11 proteins (Rv1080c, Rv1827, Rv2623, Rv1636, Rv0952, Rv1932, Rv0009, Rv0054, Rv2889c, Rv3418c, and Rv3914) that were overexpressed in the presence of OFX and MOX. Among them, Rv1636, Rv2623, and Rv1827 were reported with unidentified functions. Molecular docking and InterProScan investigation explored the interaction of conserved domain of hypothetical proteins with these both drugs which possibly directs the inhibition of functions of these proteins that might be overexpressed to repress this effect (Lata et al., 2015b).

In another study, the same author reported 14 proteins with more intensities in OFX resistant isolates as compare to susceptible. Interaction of conserved domains and motifs of hypothetical proteins (Rv3551, Rv2744c, Rv0148, and Rv2140c) and OFX was revealed through molecular docking, these findings indicated the role of these proteins in resistance mechanisms as well as prospective drug targets (Lata et al., 2015a). Drug efflux mechanisms or changes in the membrane permeability to agents might constitute alternative resistance mechanisms regarding FQ resistance in M. tuberculosis (Escribano et al., 2007). A study corroborating such ideas has found that 30% of FQ's-R strains do not harbor any mutation in the QRDRs of both gyrA and gyrB genes (Alvarez et al., 2014). In a recent study by Pucci et al., M. tuberculosis GyrB inhibitors (SPR-720, SPR-750) have proved to be attractive for developing novel drugs against TB without cross-resistance to known quinolones and showing very good activity in vivo.

ethA and ethR are associated with resistance to ethionamide (ETH) (2-ethylpyridine-4-carbothioamide) and prothionamide (2-propylpyridine-4-carbothioamide) which are important drugs for the treatment of MDR-TB and TB meningitis in both adults and children (Thee et al., 2016). The 1,470-bp long ethA encodes the monooxygenase EthA which bio-activates the pro-drug ETH into its active form (Grant et al., 2016). EthA is negatively regulated by the transcriptional repressor EthR (Carette et al., 2011). ETH is derived from nicotinic acid having a similar structure to INH. The action of ETH is to interrupt mycolic acid synthesis by establishing an adduct with NAD that inhibits the enoyl-ACP reductase InhA (Mori et al., 2017). Mutations in ethA/ethR, as well as mutations in inhA or its promoter cause the resistance to both INH and ETH (Carette et al., 2011). Accordingly, −15C to −15T mutation in the promoter region of inhA and S94A (serine to alanine) and I194T (isoleucine to threonine) mutations in the gene product InhA were found to confer high-level resistance against INH and ETH in MDR-TB (Machado et al., 2013). Furthermore, mshA, a gene encoding a glycosyltransferase involved in mycothiol biosynthesis, has also been considered as a possible resistance target for ETH (Vilchèze et al., 2008).

The 792-bp long thyA, encoding thymidylate synthase A, has an important role in deoxyribonucleotide biosynthesis, by providing the sole de novo source of dTMP through conversion of deoxyuridine monophosphate (dUMP) to deoxythymidine monophosphate (dTMP). Interacting with folate metabolism is essential, especially involving tetrahydrofolate as a “methyl” donor. In an interesting study, mutations in thyA gene were identified by transposon mutagenesis were associated with resistance to PAS (Rengarajan et al., 2004). Some studies have affirmed Thr202Ala in thyA as the most common mutation associated with PAS-resistance, though a few susceptible isolates containing the same mutation have also been observed (Meumann et al., 2015). According to recent studies, the 1464-bp long folC gene (encoding FolC-dihydrofolate synthase/ folylpolyglutamate synthase), with various missense mutations were found to confer resistance to PAS in laboratory isolates of M. tuberculosis. The main function of FolC, the bifunctional enzyme, is to convert folates into polyglutamate derivatives, in simplistic terms. Bacteria require folate for the biosynthesis of glycine, methionine, formyl-met-tRNA, thymidylates, purines and pantothenate. In 85 MDR-TB clinical isolates, mutations in folC were identified in five PAS-resistant isolates (Zhao et al., 2014). In another study, only 37% of PAS-resistant strains had mutations in thyA, suggesting that other mechanisms of resistance to PAS would need to be explored (Mathys et al., 2009).

rplC and rrl genes have been associated with different levels of phenotypic resistance to Linezolid (LZD) and Sutezolid (PNU-100480, SZD) in M. tuberculosis (Makafe et al., 2016; Zhang et al., 2016). rplC gene has 654 bp in length and encodes the 50S ribosomal L3 protein, and may contribute in the synthesis of the ribosomal peptidyltransferase. Whereas, rrl is 3138 bp long and encodes 23S ribosomal RNA. LZD and SZD belonging to the oxazolidinone class, act by fixation of an early step in protein synthesis, through binding of the assembly initiator protein directly near the 3′-end of 23S rRNA, where nucleation of the assembly of the 50S subunit is performed (Williams et al., 2009). However, recently, the mutation T460C in rplC of LZD-resistant strain (Cys154Arg) has been repeatedly shown to be of considerable importance (Williams et al., 2009; Beckert et al., 2012; Makafe et al., 2016; Zhang et al., 2016). Moreover, in vitro selected LZD-resistant mutants with G2576T and G2061T mutations in rrl gene had MICs of 16–32 μg/ml, while susceptible strains without mutations showed MICs of 4–8 μg/ml (Hillemann et al., 2008; Zhang et al., 2016). The mutations in these two genes are associated with only 29.4% of LZD-resistance in M. tuberculosis (Islam et al., 2017). So, further resistance mechanisms need to be explored.

The 498-bp Rv0678 encodes a conserved protein with a molecular mass of 18,346.7 (Da). Some recent studies recommended that mutations in the transcriptional regulator of Rv0678 up-regulated MmpL5, a multi-substrate efflux pump, causing resistance not only to CLO (a riminophenazine compound) but also to BDQ, a diarylquinoline (Andries et al., 2014). Some studies suggested the outer membrane of M. tuberculosis as the possible target of CLO (Yano et al., 2011). The main factor for resistance to CLO other than mutations in the transcriptional regulator of Rv0678 is not yet fully characterized (Hartkoorn et al., 2014). The atpE gene is known to encode the subunit C of the ATP synthase, AtpE (lipid-binding protein) (dicyclohexylcarbodiimide-binding protein), a complex structure that generates the ATP needed by the mycobacterial cell.

Mutations in the 246-bp long atpE gene (Rv1305) are thought to be related to the resistance to BDQ with the most frequently detected mutations including A63P and I66M (Andries et al., 2005; Koul et al., 2007). BDQ has a favored specificity toward the mycobacterial ATP synthase, as compared to mitochondrial ATP synthase in exerting its therapeutic action (Haagsma et al., 2009), but the mutations in AtpE (A63P and I66M) disturbs such inhibition against bacterial ATP synthase C. However, out of 53 strains of M. tuberculosis only 15 displayed mutations while 38 strains lacked mutations in atpE or even in the F0, F1 operons (Huitric et al., 2010), which strongly implicates that some other resistance mechanisms against BDQ might be present. Recently, mutations in pepQ (Rv2535c, a putative Xaa-Pro aminopeptidase) were proposed to be able to confer cross-resistance between BDQ and CLO. Whereas mutations in Rv0678 and pepQ were noticed at a similar rate in a wild-type population treated with BDQ or CLO in mice. Mutations in Rv0678 and pepQ regulate the significant mechanisms of clinical resistance to BDQ and CFZ (Almeida et al., 2016).

alrA encodes D-alanine racemase AlrA, required for the conversion of L-alanine to D-alanine (Palomino and Martin, 2014). D-alanine ligase inhibits the production of peptidoglycan conductive to cell wall formation in bacteria. D-cycloserine (DCS) is a time-honored oral bacteriostatic anti-TB drug used in treatment of TB, especially MDR-TB and XDR-TB (Zhang and Yew, 2015). The absolute target of DCS in M. tuberculosis is still not fully identified. However, according to some prior studies, overexpression of AlrA triggered resistance to DCS in recombinant mutants of Mycobacterium smegmatis (Cáceres et al., 1997). Interestingly, a very recent study reported that loss-of-function mutations in ald (Rv2780), encoding L-alanine dehydrogenase, were also associated with resistance to DCS (Desjardins et al., 2016).

Another gene cycA, of 1,671 bp in size, encodes a D-alanine transporter. The point mutation in cycA has been considered a possible cause for resistance in M. bovis against DCS (Chen et al., 2012). Low cell wall permeability and the expression of emr37, a gene that collates methylase at a specific site in the 23S rRNA, hindering the binding of the antibiotic has been ascribed as the key factor of resistance to macrolides leading four to eight-fold rise in MIC values (Andini and Nash, 2006).

ddn has a gene size of 456 bp and encodes deazaflavin-dependent nitroreductase, while fgd1 with a size of 1,011 bp encodes F420-dependent glucose-6-phosphate dehydrogenase Fgd1. fbiA/B/C is a complex/operon of three genes with 996, 1,347, and 2,571 bp gene sizes, respectively. These three genes are mainly required for coenzyme F420 production for the biosynthesis of protein FbiA, FbiB, and FbiC sequentially. ddn, fgd1, and fbiA/B/C gene products were proposed as the targets of DMD and pretomanid (PA-824, PRM). DMD is a derivative of nitro-dihydro-imidazooxazole and acts by impeding the synthesis of mycolic acid in a distinct fashion compared to INH, as it only inhibits methoxy- and keto-mycolic acid synthesis while INH also inhibits α-mycolic acid (Matsumoto et al., 2006; Palomino and Martin, 2014). Mutations in ddn, fgd1, and fbiA/B/C have been proposed to be a mechanism of resistance to DMD and PRM in M. tuberculosis (Shimokawa et al., 2014).

A very recent study reported mutations in the fbiA and fgd1 genes associated with M. tuberculosis resistance to DMD in a patient with formidable drug-resistant TB (Bloemberg et al., 2015). PRM is a nitroimidazole derivative with activity against both latent TB and active TB as tested under aerobic and anaerobic conditions (Stover et al., 2000). PRM is a prodrug that needs a metabolic activation principally by a deazaflavin (cofactor F420)-dependent nitroreductase (Ddn) (Manjunatha et al., 2006). It also acts by inhibiting mycolic acid synthesis. A very recent assay of drug metabolism genes found potential resistance to PRM, and detected mutations frequency higher than those previously reported (Stover et al., 2000; Haver et al., 2015). However, further studies are essential to better delineate the mechanisms underlying bacillary resistance.

Mutations in mmpL3 were attributed to resistance in M. tuberculosis against SQ-109 (1,2-ethylenediamine), a synthetic analog of EMB (Tahlan et al., 2012). The 2,835-bp MmpL3 is a transmembrane transport protein presumably involved in fatty acid transport. SQ109 acts by interfering with the assembly of mycolic acids into the bacterial cell wall, resulting in accumulation of trehalose monomycolate, a precursor of the trehalose dimycolate. Similarly other cell wall inhibitors such as INH and EMB, SQ-109 induces the transcription of iniBAC operon required for efflux pump functioning (Boshoff et al., 2004). Its synergistic effects with BDQ in vitro and interactions with SZD were also observed. Additionally, mutations in the mmpL3 gene of spontaneously generated SQ-109-resistant mutants has suggested mmpL3 as the target of SQ-109 by virtue of the transporter mechanism discussed above (Grzegorzewicz et al., 2012). In M. tuberculosis strains resistant to INH, EMB and SQ109, there is an up-regulation of ahpC, signifying a possible role of this gene in the development of resistance against these drug(s) (Jia et al., 2005).

dprE1 (Rv3790) and dprE2 (Rv3791) encode decaprenylphosphoryl-beta-D-ribose 2′-oxidase and decaprenylphosphoryl-D-2-keto-beta-erythro-pentose reductase, respectively. These genes encode proteins that catalyze the two-step epimerization of decaprenylphosphoryl ribose (DPR) to decaprenylphosphoryl arabinose (DPA) in the arabinan synthesis pathway, which is essential for building the bacterial cell wall (Kolly et al., 2014). dprE1 and dprE2 were the possible targets of 1,3-benzothiazin-4-one or benzothiazinone (BTZ) (Makarov et al., 2014). BTZ is activated in the bacteria by reduction of an essential nitro group to a nitroso derivative, which can react with a key cysteine residue in DprE1 and form a covalent adduct (Trefzer et al., 2010). In spontaneous BTZ-resistant mutants, amino-acid Cys387 of DprE1 was replaced by Ser or Gly. On the other hand, in M. avium, which is naturally resistant to BTZ, amino-acid Cys387 was replaced by an Ala. An alternative BTZ-resistance mechanism has recently been described in M. smegmatis. Overexpression of nitroreductase NfnB indeed inactivated the drug by reducing its critical nitro-group into an amino-group. This was facilitated by a common amino acid stretch between NfnB and DprE1 (Manina et al., 2010). M. tuberculosis, however, seems to lack nitroreductases which enables it to inactivate this drug. Nevertheless, this finding could be important for development of new BTZ analogs with improved activity.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.