General Mechanisms Of Drug Resistance Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

To control the disease, the World Health Organization (WHO) launched a strategy named Directly Observed Treatment and Short-course drug therapy (DOTS) in early 90s . Under DOTS strategy, four drug regimens with different combinations of the five first-line anti-TB drugs (including isoniazid, rifampicin, ethambutol, streptomycin and pyrazinamide) are suggested for treatment of TB caused by drug-susceptible MTB . Although the aim of introduction of DOTS strategy is to prevent emergence of drug resistance, unorganized management of TB control and failure of patients to complete whole TB therapy lead to emergence of multi-drug resistance (MDR). MDR-TB is defined as resistance to at least isoniazid and rifampicin among all first-line drugs . According to the report of WHO on Global TB Control in 2009, 0.5 million cases of MDR-TB were reported in 2007 and most cases occur in developing countries like India and China . Treatment of MDR-TB requires the use of second-line drugs, however, they are less effective and cause more toxic side effects than fist-line drugs . Besides, most of them are much more expensive and unavailable in developing countries where MDR-TB occurs frequently . The situation is even worse due to the synergy between HIV and MDR-TB when immune system is suppressed by HIV . Hence, it is urgent to elucidate the molecular mechanisms involved in drug resistance for effective control of MDR-TB.

The purpose of this review is to introduce the recently proposed molecular mechanisms conferring INH resistance in MTB that contributes to MDR-TB. In addition, current research approaches for elucidating the resistance are also discussed.

2. Drug Resistance

General Mechanisms of Drug Resistance

Mechanisms of drug resistance can be caused by intrinsic or acquired means . Intrinsic resistance is caused by factors other than genetic mutations while acquired resistance is caused by gene mutations. Examples of intrinsic resistance includes: the protection of MTB from antibiotics by their permeability barrier formed by hydrophobic, waxy, lipid-rich cell. Influx of drugs is inhibited by the barrier and intracellular drug accumulation can be decreased . Besides, resistance may also be conferred by suppressing uptake/stimulating efflux of antibiotics out of the cells. Acquired resistance is also common in MTB, as observed by inactivating or modifying enzymes involved in drug action and inhibiting pro-drug activation .

Isoniazid (INH) Resistance

INH, also known as isonicotinyl hydrazine, has been used since 1952 . It is one of the first-line drugs involved in the regimen for treating TB due to the high sensitivity of INH susceptible MTB to it. The minimum inhibitory concentration (MIC) of INH required to inhibit visible growth of MTB is about 0.02μg/ml . Forty years before the first usage of INH for TB treatment, INH had been synthesized from ethyl isonicotinate and hydrazine by two chemists in Prague for their doctoral work . However, medical value of INH had not yet been discovered until the recognition of anti-TB activity of thiosemicarbazones and nicotinamide . After trials of combining these two compounds together to form isonicotinaldehyde thiosemicarbazone, it was found that the intermediate (INH) formed in the synthesis was itself a drug with potent anti-TB activity . After the study for more than 50 years, more information is now available about INH working mechanism. It is generally accepted that INH kills MTB by inhibiting synthesis of cell wall mycolic acids . In spite of high anti-TB activity of INH, incidence of INH resistance increases. It has been reported that about 28% of MTB strains are resistant to INH . Among previously treated TB cases, 60% MTB strains are found INH resistant and about 6 in 100 new patients suffering from TB are resistant to INH treatment. These figures urge the studies on INH action and underlying mechanisms that cause INH resistance.

2.3 Mechanism of INH activation and action

In this section, INH activation, INH-NAD(P) adduct formation, potential targets of the adduct in MTB and their relevance to drug actions are discussed.

Activation of INH

INH enters MTB by passive diffusion .Several evidences suggest that INH exists as pro-drug form, and it is required to be activated by an enzyme called catalase peroxidase (KatG) first before it exerts its action. The evidences include the observation that some INH-resistant isolates loss catalase activity , INH resistance in MTB is resulted from deletion of katG gene , and sensitivity towards INH in INH-resistant MTB is restored when wild-type katG gene is transferred to INH-resistant MTB in form of multicopy plasmid . It is believed that the pro-drug is oxidized and activated by KatG coded by katG gene to form isonicotinoyl radicals .

KatG is a heme enzyme that belongs to class I superfamily of peroxidases . It is homodimeric or homotetrameric and the molecular mass of each subunit is about 80-kDa . KatG has been proposed to protect MTB from toxic molecules like hydroperoxides and hydroxyl radicals formed during aerobic respiration . After analyzing the crystal structure of a recombinant form of KatG, in addition to comparative analyses of protein sequences and structures of other class I peroxidases, active site of KatG to INH is determined to be close with δ-meso edge of heme which is commonly known as a hydrophobic pocket . The key active site residues in distal pocket as shown in figure 1 are suggested as Arg104, Trp107, His108, while in proximal pocket, His270, Trp321 and Asp381 play a major role for catalytic action. These six residues are also well conserved in other class I peroxidases and their roles in regulating peroxidation are different .

During INH activation by KatG, isonicotinoyl radicals are expected to form. It has been proposed that the radicals can be formed by several reactions as shown in figure 2 . In process A, wild-type KatG can first form Compound I, an oxidized ferryl porphyrin π-cation radical (shown as (Por.)FeIV=O in figure 2) by reacting with peroxide, then in process B, it is reduced by INH to form isonicotinoyl radical (INH.) and Compound II, that is wild-type KatG with a protein radical and Fe3+ at heme of it (shown as (KatG.)FeIII in figure 2). In process C, Compound II is further reduced by INH to form another isonicotinoyl radical. Besides, in KatGs consisting of mutations at active site, an intermediate of iron (IV)-oxo Compound II (shown as (KatG.) (Por.)FeIV=O in figure 2) are also formed from Compound II through process D and it is found to facilitate formation of isonicotinoyl radical when it is reduced back to compound I through process E. Another compound named Compound III (shown as (KatG)FeIII-(O2-) in figure2) can be formed by binding dioxygen or excessive hydrogen peroxide to active site of ferric heme of KatG as shown in process F. Compound III may also be involved in activation of INH, as shown in process G and B.

Although several mechanisms of isonicotinoyl radical formation have been proposed, Compound I is the critical component in traditional catalase cycle . The molecular mechanism of radical formation involving conversion of Compound I to Compound II is shown in figure 3 . In figure 3, it can be seen that after an electron is transferred from INH to heme group of Compound 1, the hydrazide moiety of INH loses a proton which can be accepted by His-108. After that, diazene is formed by splitting C-N bond of hydrazide. The diazene is in an orientation such that the carbonyl group of INH is directly below acid group of Asp-137 in active site. As a result, diazene can be stabilized and radicals formed may diffuse out from active site.

Formation of INH-NAD(P) adduct

The isonicotinoyl radical formed after activation of INH reacts non-enzymatically with oxidized NAD(P)+. It is then covalently bound to nicotinamide ring of from NAD(P)+ moiety . The schematic description of conversion of INH and NAD+ into INH-NAD adduct is shown in figure 4 . Six types of isonicotinoyl-NAD(P) (INH-NAD(P)) adducts can be formed as shown in figure 5 .

Potential targets of INH-NAD adduct and their relevance to drug action

As a drug target with antimicrobial activity, it should be an enzyme of bacteria that can be bound and inhibited by the drug and the inhibition leads death of bacteria . Several proposed targets of INH-NAD adduct are discussed here:

Enoyl-ACP reductase (InhA)

This enzyme is coded by inhA gene and it is well documented as a major target of INH-NAD adduct . One of the adducts acyclic 4S isomer (compound 1 shown in figure 5 ) has high affinity to InhA which is involved in type II fatty acid synthase (FAS) pathway for production of mycolic acids . Mycolic acids help constitute MTB cell wall which protects the bacteria from chemical damage, dehydration and antibiotics. As shown in figure 6 , InhA catalyzes NADH-dependent reduction of double bond between C2 and C3 of trans-2-enoyl-acyl carrier proteins (ACPs). At the beginning of each cycle, two-carbon subunits from malonyl-ACP are added for elongation of C16-C18 ACPs. Eventually meromycolates composing of about 50-60 carbons are formed. The meromycolates are then condensed with fatty acids composed of 22-24 carbons to form mycolic acids, a α-branched fatty acids, to form mycobacterial cell walls. When InhA is inhibited by INH-NAD adduct, synthesis of mycolic acid will also be inhibited. Consequently, cellular integrity is decreased and MTB will be more susceptible to oxygen radicals and other environmental factors that can trigger mycobacterial cell death .

There are several evidences suggesting InhA is a target of INH-NAD adduct.

Firstly, in terms of statistics, mutations within promoter or structural regions of inhA gene have been found in about 25% of clinical INH-resistant isolates of MTB but not in INH-susceptible isolates .

Secondly, after artificially introducing mutation like S94A found in INH-resistant isolates to INH susceptible isolates by allelic exchange, INH resistance could be conferred . After the transduction, transductants with or without mutated inhA were screened by gene sequencing and their MIC was also tested. It was found that the mutation was 100% correlated with INH resistance with increased MIC. This further supported that InhA is a target for INH and mutations in inhA gene play an important role in INH resistance, which will be discussed later.

In vitro InhA inhibition by INH-NAD adduct has also been studied to illustrate InhA as a target of activated INH . Kinetic studies between interactions of INH-NAD adduct with InhA are also available . The enzymatic activity of InhA in the presence or absence of INH-NAD adduct can be followed by oxidation of NADH to NAD+ at 340nm measured by spectrophotometer. Inhibition of InhA by adduct was found to decrease when concentration of acyl-ACP substrates increased . This showed that inhibition of adduct took place at substrate binding region of InhA in a competitive manner. It has also been demonstrated that InhA is inhibited in a two-step process in vitro . From the progress curve for inhibition of InhA by the adduct obtained in the study, it can be observed that under different concentrations of adduct, the InhA activity dropped exponentially with time and finally became steady. The higher the concentrations of adduct, the lower the turnover rate of NADH to NAD+ and the earlier the InhA activity became steady. This result suggested that the adduct first bound to InhA weakly to form an initial enzyme-inhibitor complex rapidly, then it converted to a final inhibited complex slowly. When the substrate binding region was blocked by the adduct, enoyl-ACP substrate could no longer be catalyzed to form mycolic acid and this manifested antimicrobial action of activated INH.

Blocking action of adduct at substrate binding site of InhA was further supported by electron density map derived from x-ray crystallography of InhA inhibited by INH, indicating there is covalent binding between INH-NAD adduct within active site of InhA . At the position where hydride of NADH is transferred to reduce enoyl-ACP substrates during normal type II FAS pathway, the 4S hydrogen of NADH is replaced by acyl group of INH-NAD adduct. Active site of InhA binding with (i) fatty acyl substrates with NADH and (ii) INH-NAD adduct is shown in figure 7 . The crystal structure of InhA bound with adduct also shows that location and orientation of isonicotinic acyl group fits to the active site .

In addition, after comparing mass spectra of (i) InhA alone, (ii) InhA in presence of NADH and (iii) InhA inhibited by INH , a novel peak is found in mass spectrum of (iii) inhibited InhA when compared with that of (i) InhA alone and (ii) InhA with NADH, revealing that a compound with molecular mass of 770 daltons is formed when InhA is inhibited by INH, that agrees with chemical structure of INH-NAD adduct with InhA.

Moreover, over-expression of inhA in MTB result in more than ten fold increase in INH-resistance . This indicates INH drug action is closely related with InhA level.

Because of consistent conclusions drawn from the studies mentioned above, InhA is now generally accepted as a primary target of activated INH. However, whether the adduct is formed before binding to InhA active site or NADH binding in InhA's active site precedes adduct formation is still in a debate. Based on some previous studies showing that inhA mutations (S94A) in resistant isolates caused NADH affinity to InhA decrease , it was suggested that NADH binding to InhA was a prerequisite for adduct formation that inhibited InhA afterwards. On the other hand, it was also found that during in vitro inactivation of InhA, the enzyme inhibition rate constant obtained when adduct formation was allowed before adding to InhA was similar to the rate constant for conversion of initial enzyme-inhibitor complex to a final inhibited complex . This suggested that INH-NAD adduct might not be formed in enzyme in vivo. In addition, it was demonstrated that inactivation rate of InhA was similar when NADH was substituted with NAD+. As NAD+ has much lower affinity for InhA than NADH, this further argued that cofactor binding to InhA may not be required before adduct formation in InhA .

Beta-ketoacyl-ACP synthase (KasA)

Similar to InhA, KasA is also involved in synthesis of mycolic acids in FASII pathway in MTB as shown in figure 6. It was first suggested as a primary target instead of InhA in 1998, due to the discovery of over-expression of KasA and a 80kDa covalent complex formed by INH-NAD adduct with KasA and ACP, the acyl carrier protein in FASII pathway in INH-resistant isolates . In addition, accumulation of a saturated C26 fatty acid on ACP was also observed after INH treatment . It was suggested to be a consequence after inhibition of KasA by INH-NAD adducts that leads to cessation of fatty acid elongation. In that study in 1998, among 28 INH-resistant isolates, two of them were identified with missense mutations in kasA gene but not other common mutations associated with INH resistance (katG, inhA and ahpC). However, the sample size of that study was small and therefore not representative. Besides, only three common mutated genes associated with INH resistance were sequenced and other feasible mechanisms underlying resistance were completely ruled out. Hence, the study could not provide strong evidence to show KasA as a target of INH. In addition, several follow up studies also revealed that when compared with InhA, the role of KasA as a drug target was not significant. Based on the mutations found in INH-resistant isolates, three kasA mutations (G269S, G312S and F413L) were introduced to INH-susceptible MTB by specialized linkage transduction . No transductants had higher MIC than susceptible isolates. This result opposed the claimant before that only mutated KasA could confer INH resistance in MTB. Moreover, there were numerous studies against the conclusion drawn in the study in 1998. For example, accumulation of saturated C26 fatty acid could be caused by temperature sensitive mutation instead of KasA inactivation , over-expression of kasA in INH-susceptible MTB could not confer them with INH resistance . These evidences implied that roles of kasA mutations might be indirect to INH resistance and further investigation of KasA as drug target is required

Dihydrofolate reductase (DHFR)

DHFR is an enzyme involved in nucleic acid synthesis . The possibility of DHFR being one of the drug targets is based on previous reports showing that MTB nucleic acid synthesis is inhibited by INH . In contrast to InhA and KasA, it is the INH-NADP+ adduct instead of INH-NAD adduct that may target on DHFR . Among the six types adducts, according to the crystal structure of complex formed between the adduct and DHFR, the acyclic 4R isomer of INH-NADP+ was shown to be a subnanomolar (Kiapp = 1nM.) bisubstrate inhibitor of the enzyme by an in vitro study . It was suggested that the adduct acted as a potent linear competitive (with respect to NADPH) inhibitor of DHFR by fully occupying the active site . Because of structural similarity, the adduct could displace the NADPH tightly bound with enzyme and behaved as a bisubstrate analog. It was also observed that when dfrA, the gene encoding DHFR in MTB was over-expressed in a surrogate host, M. smegmatis, a 2-fold increase of INH resistance was conferred . These observations were claimed to support DHFR as a target of INH in addition to InhA . However, there are several arguments regarding to this hypothesis drawn from the above observations. Firstly, there is a lack of information of formation of 4R INH-NADP adduct in vivo . Whether KatG or other enzymes involved for 4R INH-NADP adduct production and the corresponding mechanism are still not well documented. Secondly, importance of dfrA gene to MTB also affects its relevance to INH action. Despite the fact that it is essential for nucleic acid synthesis in other organisms, it was found that MTB with disrupted dfrA could still infect mice . This implied other genes in addition to dfrA may play a more significant role for MTB replication and survival. Moreover, the maximal INH resistance conferred by over-expression of dfrA in M. smegmatis found so far was only 2 fold higher than the wild type , which was not significant enough to be a strong evidence supporting DHFR as a drug target. Besides, surrogate host Mycobacterium smegmatis but not disease-causing MTB was used for over-expression study. Strain difference might contribute to different resistance level and the result obtained from it might not be strongly relevant for study of INH resistance in MTB, given the fact that wild type M. Mycobacterium smegmatis is more resistant to INH by 100 folds than MTB . An evaluation of MIC after over-expression of dfrA in MTB showed that MIC was the same between dfrA over-expressed strain and wild-type strain. Whole-genome sequencing was also used to show that no polymorphisms were found in dfrA in six INH-resistant clinical isolates lacking mutations in inhA and katG . These demonstrated DHFR is not a direct drug target.

Although genetic and biochemical studies strongly suggest that InhA is a primary target of INH, there are still possibilities of other targets. The study of DHFR's role in drug target and resistance raises attention to other compounds that react with adducts as they may be potential drug targets. Identification of these compounds provides a new approach for the study of new potential drug targets and their related drug resistance mechanism.

2.4 Mechanism of INH resistance

Although INH resistance first occurred in 1950's after introducing for TB treatment , the underlying mechanism is still not fully known. To control TB and drug resistance aroused, elucidation of mechanisms for the resistance is essential. Studies of genetic mutations, efflux pumps and their expressions have been done to explain the mechanism.

2.4.1 Genetic mutations

Early researches of molecular mechanism of INH resistance mainly focused on genetic mutations. The genetic mutations proposed to associate with resistance are summarized in Table1. katG gene

Role of katG in INH resistance was first recognized when it was shown to transfer drug susceptibility to INH resistant MTB . Association between INH resistance and substitutions, deletions, insertions or complete loss of katG gene has been proposed . katG encoding catalase peroxidase in MTB is the hottest spot of mutations found in resistant isolates. About 50% of all INH-resistant isolates carry mutated katG . Mutations frequently occur between codons 138 and 328 and different mutations in katG confer both high and low level of INH resistance . Among all INH-resistant isolates carrying katG mutations, 50-95% of them have serine at position 315 substituted by threonine .

katG-mediated INH resistance can be affected by variety of factors, like oxidant used by KatG for INH activation, affinity of enzyme with INH and NADH, KatG/heme stability and hydrogen bondings altered by mutations. Positions of mutations and even different residues at the same position can elicit different mechanisms for INH resistance. Relationships between some katG mutations and INH resistance have been proposed based on biochemical assays, enzymology and structural analysis. Some of the examples are summarized into three groups based on their locations in KatG:

Mutations at substrate access channel

S315T and INH resistance

Mutations at Ser315 can confer high-level INH resistance by increasing MIC up to 200 folds . Studying the structural changes before and after these mutations provide hints for elucidating origins of resistance. As shown in figure 8 , serine at 315 position is at the edge of INH binding pocket near the bottom of substrate access channel . The hydroxyl group of Ser315 forms a hydrogen bond with carboxylate group of the heme propionate side chain .When serine is substituted by threonine, threonine has larger side chain and more steric bulky than serine, the side chain of threonine is located closer than allowed Van der Waal's contact distances , so steric hindrance of binding INH to KatG increases . This mutant has been reported to have a substrate access channel constricted from 6 Å to 4.7 Å . Affinity of enzyme for INH decreases , hence activation rate of INH decreases and confers INH resistance to MTB. It should be noted that although S315T KatG has lower affinity to INH, it has same INH oxidation rate with wild type KatG . This implies threonine at 315 position does not affect the activation of INH.

S315G and INH resistance

S315G INH-resistant mutants have similar affinity of INH with KatG as glycine is smaller than serine and it does not cause steric hindrance for access of substrate to KatG. However, it has been observed that INH is not a good substrate for reducing Compound I intermediate during INH activation for these mutants. As mentioned in earlier part discussing INH activation, the exact way for INH activation is not yet confirmed. It was suggested that this mutant may shunt the catalytic reactions that are involved for INH activation to other pathways, and the catalytic intermediates involved for INH activation was redirected into nonproductive reactions, so that normal KatG function was retained but the formation of INH-NADH dropped . When hydrogen peroxide was used as substrate for initiation for activation of INH, levels of INH-NADH adduct formed by S315G mutant and wild-type KatG is similar. However, when superoxide was used instead of hydrogen peroxide, there was no adduct formation. As the physiological substrate for KatG is still unknown, it may be possible that S315G inhibits superoxide-dependent pathway for INH activation and this may lead to INH resistance . The underlying mechanism for this observation is still unknown and further investigation is required.

In addition to threonine and glycine, other amino acids including asparagines, isoleucine and arginine have also been reported be a substitute of serine at 315 position . The predominance of S315T may be due to the fact that S315T provides survival advantage, but not other substitutions . For some S315 mutants, they lose catalase, peroxidase and INH oxidase activity. In contrast, S315T mutant retains all KatG activity . It is well known that KatG is essential for protection of MTB from toxic metabolites under aerobic environment. Therefore, S315T mutants are naturally selected and become dominant.

Mutations at C-terminal domain and region connecting it with N-terminal domain

D419H, M420T

Isolates with these mutations have MIC of 1μg/ml INH. In a homodimeric KatG, C-terminal domain and N-terminal domain are found in each monomer, with the heme binding site found in N-terminal domain only . D419H and M420T are found in region connecting the two domains and that region is essential for interdomain interactions and thus formation of functional homodimer . The 3D conformation of KatG may be altered in D419H, M420T mutants and the enzymatic activity may lose , this inhibits activation of INH and confer high-level INH resistance in MTB.

D542H, R632C

D542H and R632C are found in C-terminal domain of KatG. KatG with these mutations were found to lose enzymatic activity . The underlying mechanism is still unknown as the functional role of C-terminal domain is still not clear. However, it was found that KatG is inactivated after deletion of whole C-terminal domain , indicating its possible roles in stabilizing enzyme or facilitating enzymatic activity in N-terminal domain. Hence, mutations in C-terminal domain may confer resistance by affecting its roles.

Mutations at/near active site

Mutations at active site: R104L, W107R, H108E/Q

All these mutations suppressed INH-NAD adduct formation when compared with wild-type KatG . His108 is one of the active site residues in distal pocket of KatG and it plays an important role in INH activation. Mutation of His108 confers high INH resistance. Mutated isolates have MIC of 50μg/ml INH . As shown in figure 3 , after an electron is transferred from INH to the heme group of Compound 1, the hydrazide moiety of INH loses a proton which can be accepted by His-108. When His108 is substituted by glutamic acid or glutamine, they are not expected to accept proton from INH . Hence, the activation of INH is hindered and INH resistance is conferred. Other mutations that affect local conformation of His108 like A110V also confer INH resistance by altering interaction between His108 and INH, and thus the INH activation. In addition, it was suggested that activation of heme group in active site and reaction of its intermediates might be affected by these mutations, hence INH activation decreased and INH resistance was conferred .

Mutations near active site and INH resistance

A number of mutations near Asp-137 have been found in INH resistant isolates, including N138S/D, A139P, S140N and D142A . During INH activation, the diazene formed is stabilized by orientating the carbonyl group of INH directly below acid group of Asp-137 near active site as shown in figure 3 . Besides, Asp137 is a proton donor specific for catalase peroxidase . Catalytic function, stabilization effect, and thus INH binding may be hindered by mutations at this position. However, no mutations have been reported at Asp137 . Instead, mutations nearby Asp137 have been reported. These mutations may alter position of carboxyl group and down-regulate INH activation but meanwhile, catalase activity of KatG can be retained to protect MTB from bacteriocidal factors . inhA gene

inhA codes enoyl acyl carrier protein reductase, a proposed target for INH-NAD adduct . Compared with katG, mutations in inhA confer lower level of resistance and their frequency is less than that in katG, they are only found in 25% of INH-resistant isolates . Mutations are found in both promoter and structural region of inhA, in spite of higher frequency of mutations in promoter region . The relationship between these mutations and INH resistance is summarized as follows:

Mutations at inhA promoter

The frequent mutations occur at inhA promoter found in INH-resistant isolates include G-24T, A-16G, T-8G/A and C-15T . These mutations may cause over-expression of inhA observed in INH resistant isolates. This will increase drug target levels and INH resistance can be conferred via a titration mechanism . Among these mutations, C-15T is the most common one and it confers low level of INH resistance (0.2μg/ml) in the absence of katG mutations . mc24914, a strain of MTB carrying C-15T mutation, has been proven to induce over-expression of inhA by increasing inhA mRNA level by 20 folds . In vitro over-expression of InhA has also been done by cloning inhA to multicopy plasmids with strong promoter and transforming susceptible MTB with these plasmids . The result showed that increased InhA level correlated to increase INH resistance in originally INH susceptible MTB. Combined with these results, it can be deduced that mutations-mediated over-expression of inhA may play a role in INH resistance.

Mutations at inhA structural region

Several mutations have been reported at inhA structural region, including S94A, I21T, I21V, V78A and I95P . Mutations at InhA structural region confer low level INH resistance and they usually occur with mutations in katG . Among them, S94A is relatively well studied on the mechanism underlying INH resistance.

The S94A InhA activity was measured under different concentrations of INH-NAD adduct that inhibits the enzyme . It was found that the mutated InhA was 17 times more resistant than wild type to inhibitory effect by the adduct. Crystal structures of complex formed between the adduct with S94A InhA or wild-type InhA were also available as shown in figure 9 . The INH-NAD adduct orients in a similar manner in active sites of S94A InhA and wild type InhA. However, substitution of serine with alanine disrupts hydrogen bond network by altering position of ordered water molecule. In the wild-type InhA-adduct structure, the oxygen O9 of phosphate from NAD of adduct forms hydrogen bond with main chain nitrogen of Ile21 to orient itself for hydrogen bond interaction with the ordered water molecule . That water molecule forms hydrogen bond network with oxygen from side chain of Ser94, oxygen from main chain of Gly14, oxygen atoms O3 and O9 of phosphate group from adduct . When Ser94 is substituted with alanine, as seen in InhA(S94A)-adduct structure, hydrogen bond network between the water molecule and residues nearby is disrupted, as alanine does not contribute a hydroxyl group to form hydrogen bond with the water molecule. In addition, distance between water molecule and O3 increase and therefore becomes too far away to form hydrogen bond. Meanwhile, water molecules become closer to Ala22 and Ile21. This allows formation of hydrogen bonds between water molecules and main chain nitrogen atoms of Ala22 and Ile21 . Such alterations reduce affinity between NADH to S94A InhA, and lead Michaelis constant (Km) for NADH increases . Consequently, fatty acyl-ACP substrates may be promoted to bind with InhA before NADH. This protects NADH in active site from reacting with activated INH and blocking InhA (under the hypothesis that adduct forms in InhA active site) or INH-NAD adduct entering into active site (under the hypothesis that adduct formed before entering active site). The INH-NAD adduct in InhA may also be promoted to release due to lower affinity between InhA and NAD. As a result, inhibition of InhA by adducts decreases and low level INH resistance is conferred .

Although mechanisms of INH resistance caused by mutations occurring in InhA structural region other than S94A are not well understood, based on the molecular contacts between adduct and active site of InhA shown in figure 7(b) , it can be seen that both Ile21 and Ile95 are involved in hydrogen bond network for proper orientation of INH-NAD adduct in active site of InhA. Hence, I21T, I21V and I95P may confer low level INH resistance by alteration of hydrogen bond network distance and lowering affinity of NADH to InhA. ndh gene

ndh gene codes for the NADH dehydrogenase . It binds to active site of InhA and form ternary complex with activated INH . About 9.5% of INH resistant isolates acquire ndh mutations, and the hottest spots for mutations are T110A and R268H . It has been proposed that the mutated NADH dehydrogenase maybe defective in oxidizing NADH to NAD. Consequently, the NADH/NAD+ ratio increases . Under the hypothesis that INH-NAD adduct formed before they occupy active site of InhA, NADH competes with INH-NAD adduct to bind active site. When NADH level increases, they inhibit binding of adduct to InhA and this facilitates catalysis action of InhA. As a result, INH resistance is conferred . This has been proved by both in vivo and in vitro studies using Mycobacterium smegmatis and Mycobacterium bovis ndh mutants. An in vivo study showed that ndh mutation caused defective NADH dehydrogenase activity in these ndh mutants . Increased intracellular [NADH] was also detected while [NAD+] was unaffected and similar to that of wild type. Subsequently, NADH/NAD+ ratio increased in mutants. Given that InhA is inhibited by INH-NAD adduct , the inhibition is measured in vitro under different concentrations of NADH. Irrespective of the order of incubation of InhA with NADH, INH-NAD adduct and substrates, it was found that [NADH] was negatively related to inhibition of InhA by the adduct. One limitation of that study was that MTB ndh mutants were not included. It is known that MTB have slower growth rate and 100-fold more sensitive than Mycobacterium smegmatis , the metabolic differences between them may alter the resistance level conferred by ndh mutations or underlying mechanism. Hence slow growing Mycobacterium bovis ndh mutants were used to mimic the case of MTB. Mycobacterium bovis have similar INH sensitivity with MTB. Sequence alignments of NADH dehydrogenases from MTB and Mycobacterium bovis are also highly similar. However, the mutation positions between MTB ndh mutants and Mycobacterium bovis ndh mutants are quite different. Hence, the result from Mycobacterium bovis may be only able to partially provide a possible explanation to INH resistance conferred by ndh mutations in MTB and the actual mechanism in MTB requires more investigations. ahpC gene

ahpC codes for alkyl hydroperoxide reductase (AhpC) . About 10% of INH-resistant isolates are detected with point mutations in the promoter region of this gene, in addition to katG mutations . Mutations in promoter lead to over-expression of AhpC. This may act as a compensatory mechanism to maintain peroxide homeostasis by protecting MTB which has defective KatG after some katG mutations from oxidative stress and detoxifying organic peroxides . The mutations in ahpC promoter region that have been reported in INH-resistant isolates result in induction of promoter activity . Over-expression of aphC has been proven to associate with these mutations by comparing AhpC protein level of INH-resistant isolates with these mutations and control with immunoblot analysis . Despite of detection of ahpC mutations in INH-resistant isolates, it should be noted that they do not confer resistance directly, based on following observations : (1) over-expression of ahpC in INH-susceptible isolates by multicopy plasmids did not confer significant increase in MIC; (2) ahpC promoter mutations were only detected together with katG mutations that known for resulting in low (or even loss) KatG activity but not common katG mutations that retain certain degree of KatG activity e.g. S315T . This further suggests that ahpC mutations act as compensatory role when KatG activity decreases. kasA gene

kasA codes for β-ketoacyl-ACP synthase for synthesis of mycolic acids . Relationship between kasA mutations and INH resistance had been suggested based on its involvement in mycolic acid synthesis and detection of covalent binding between activated INH and ACP by radioactive INH . Four mutations in kasA gene have been detected INH-resistant isolates together with mutations in other genes , though three of these mutations have also been found in INH-susceptible isolates . As the nature of binding between KasA-ACP and activated INH is still unknown , the possibility of kasA mutations being a resistance conferring mechanism is still unknown.

2.4.2 Efflux pump

Although gene mutations may explain some mechanisms of drug resistance of MTB, there are still about 20 to 30 % clinical INH-resistant isolates which do not have mutations in known genes related with INH resistance . In fact, interactions between different resistance mechanisms instead of individual mechanism are required for conferring high-level resistance. The interactions may explain resistance in MTB with no known resistant gene mutations, and also the varied resistance level in MTB with mutations in genes known for causing resistance . To acquire more in-depth information about mechanisms underlying INH resistance, more studies now focus on possibility of efflux-related mechanisms . It has been predicted that MTB encode multiple putative efflux proteins after publication of whole genome sequence of MTB laboratory strain H37Rv in 1998 , despite the fact that most of them are lack of characterization . It is believed that efflux pumps help MTB maintain intracellular homeostasis through removing toxins or drugs . Specificities of efflux pumps vary. Some efflux pumps are specific for removal of particular drugs while some pumps like multidrug resistance efflux pumps may facilitate efflux of drugs with unrelated structure or functions . However, number of pumps are limited, so long as the pump activity is saturated with increasing drug concentration, pump cannot confer resistance anymore . Hence, the number of functional efflux pumps, the efficiency of pumps and the mechanism that regulate amounts of pumps also determine drug resistance level conferred by pumps.

There are several evidences showing the possible role of efflux pumps in INH resistance in MTB.

Firstly, it has been demonstrated efflux systems of MTB can be induced after prolonged exposure to INH and consequently INH-susceptible MTB has higher INH resistance level . In that study , MIC of MTB isolates were measured first and their susceptibility to INH was confirmed (MIC was equal to 0.03mg/L). Then, these isolates were subjected to medium containing 0.1mg/L INH. The incubation period was extended until they showed visible growth level similar to their controls in drug-free medium. After that, MIC was measured again and it was found to increase from 0.1mg/L to 0.2mg/L. The cells passages were repeated with serial increase in INH concentration in medium which corresponded to MIC for previous passage. The maximum MIC induced in that study reached 40mg/L INH. This showed INH resistance could be induced by serial increase in INH concentrations in growth medium. Later, those cells induced with MIC of 40mg/L INH were subjected to drug free medium and their MIC was found to decrease gradually. At last, their MIC decreased to a similar level to their parent strain. They became INH susceptible and they could be used to start repeating the cycle of induction again. To confirm whether the transient induction of INH resistance was due to mutations associated with drug resistance, the most common resistance-related gene katG was sequenced and no mutations at that gene were detected. In addition, the efflux pump activity of isolates with induced INH resistance was also determined with known inhibitors of bacterial efflux pumps, including reserpine. It was observed that for the isolates with induced INH resistance, the higher the concentrations of reserpine present in their growing medium, the lower their resistance level. This implied reserpine-sensitive efflux pumps might be involved in acquiring INH resistance in MTB. Although only katG gene was sequenced in that study and it might be argued that mutations in other resistance-associated genes could confer resistance. However, based on the observation of reversal resistance in that study, it can be concluded that genetic mutations did not play an essential role. If mutations mediate resistance, when MTB were subjected to drug-free medium, their MIC should not decrease and back to level before induction, given that the probability of reverse mutations was very low. In addition, prolonged passages in drug-free medium were required for reversal resistance. This argued with the possibility that resistance was caused by up-regulation of certain enzymes to inactivate or degrade INH as seen in other species of bacteria by drugs, given that the drug resistance caused by gene up-regulation is short-lasting .

The successful induction of INH resistance by INH encourages more studies on gene expression levels of putative efflux pumps. It has been found that clinical isolates MTB can be induced by INH to over-express several MFS (major facilitator superfamily) efflux pump genes, including efpA, pstB and ABC transporters (ATP-binding cassette-type multidrug transporters) like Rv1819c and Rv2136c . However, the mechanisms underlying pump-mediated resistance are still unclear. Most of the putative protein pumps are not yet characterized . It is still unknown whether the induced resistance is the result of activation of pumps, increasing amount of function efflux pumps or coordination of different pumps if any. The mechanisms for induction of pumps are unclear. In addition, identification of pumps is restricted to MTB laboratory strains with induced expression of efflux pumps, but very limited studies about clinical isolates are available.

Summary points:

The molecular mechanism involved in emergence of INH resistance of MTB is very complex. Activation of INH to form adducts is required for its action to inhibit mycolic acid synthesis, that finally causes MTB cell lysis and death. The activation is believed to be mediated by KatG. Several targets have been proposed for the action of adduct, and InhA is the most well-accepted one.

As KatG and InhA are involved in drug activation and action, katG and inhA gene mutations are the most common mutations found in clinical INH resistant isolates. Other gene mutations may play complementary roles to mutations in katG and inhA. The mutations reported to be associated with INH resistance and their actions are summarized in figure10.

There are also growing evidences to show involvement of efflux pumps in INH resistance. However, most of them lack characterization and more investigations are required.

Future issues:

Firstly, little is known about the metabolic pathways of MTB. Present studies mainly focus on pathways that are directly related to INH activations or actions, however, there are limited data related to drug degradation/modifications or other intrinsic compensatory mechanism in MTB . The lack of knowledge of metabolic pathways of MTB renders the inability to explain over-expression of some genes. One cannot elucidate whether over-expression of particular genes lead to drug resistance or genes over-expression acts as an intrinsic metabolic regulatory role for loss of function of other mutated gene products. Although several genes have been shown to be induced in presence of INH and some of them acquire mutations, the functional role of their over-expression should be investigated before drawing the conclusion that they are related to drug resistance. Thanks to the publication of whole genome sequence of MTB laboratory strain H37Rv, it is expected that more potential resistance-related metabolic pathways and proteins can be identified and characterized.

Secondly, as mentioned before, there are still about 20 to 30 % clinical INH-resistant isolates which do not have mutations in known genes related with INH resistance. This implies some other mutations in unknown genes may contribute to the drug resistance. Whole genome sequencing may be an approach to identify new mutations. Meanwhile, it should be noted that polymorphisms may not be related to drug resistance, but reflect different geographical prevalence of specific genotypes. This situation is particular common for mutations found with low frequency rate. Hence, high throughput sequencing, a new technique for whole genome sequencing by massively parallel sequencing process, can be applied to tackle with this. Multiple copies of sequencing products (reads) can be produced simultaneously during high throughput sequencing and this enables majority voting on bases which are polymorphisms unrelated to drug resistance.

Although relationships between genetics and INH resistance have been studied for long time, relationship between epigenetics and drug resistance are still not well investigated. Some studies on gene over-expressions like genes coding efflux pumps under INH treatment are on-going. However, most of them remain at the stage of gene identifications and the variations in epigenome (i.e. the mechanism causing change in gene expressions) are still not yet well studied. This provides a large area for exploration of new molecular mechanism involved for INH resistance in MTB.