Single Turnover of Transient of Reactants Supports a Complex Interplay of Conformational States in the Mode of Action of Mycobacterium tuberculosis Enoyl Reductase

: The enoyl reductase from Mycobacterium tuberculosis ( Mt InhA) was shown to be a major target for isoniazid, the most prescribed ﬁrst-line anti-tuberculosis agent. The Mt InhA (EC 1.3.1.9) protein catalyzes the hydride transfer from the 4 S hydrogen of β -NADH to carbon-3 of long-chain 2-trans -enoyl thioester substrates (enoyl-ACP or enoyl-CoA) to yield NAD + and acyl-ACP or acyl-CoA products. The latter are the long carbon chains of the meromycolate branch of mycolic acids, which are high-molecular-weight α -alkyl, β -hydroxy fatty acids of the mycobacterial cell wall. Here, stopped-ﬂow measurements under single-turnover experimental conditions are presented for the study of the transient of reactants. Single-turnover experiments at various enzyme active sites were carried out. These studies suggested isomerization of the Mt InhA:NADH binary complex in pre-incubation and positive cooperativity that depends on the number of enzyme active sites occupied by the 2-trans -dodecenoyl-CoA (DD-CoA) substrate. Stopped-ﬂow results for burst analysis indicate that product release does not contribute to the rate-limiting step of the Mt InhA-catalyzed chemical reaction. The bearings that the results presented herein have on function-based anti-tuberculosis drug design are discussed.


Introduction
The Global Tuberculosis Report 2022 published by the World Health Organization (WHO) was based on data reported by 202 countries and territories accounting for more than 99% of the world's population [1]. It has been estimated that approximately one-fourth of the world's population is infected with Mycobacterium tuberculosis, the main causative agent of tuberculosis (TB) in humans [1]. Deaths from TB of HIV-negative and HIV-positive patients in 2021 accounted for 1.4 million and 187,000 people, respectively. Moreover, drug resistance continues to be a public health threat [1]. Approximately 85% of people who develop TB can be successfully treated with a 6-month drug regimen for drug-susceptible TB (two months with isoniazid, rifampicin, pyrazinamide and ethambutol, followed by four months with isoniazid and rifampicin). On the other hand, treatment of MDR/RR-TB cases requires a course of second-line drugs, which are more expensive (≥USD 1000 per person), and ought to be supported by counselling and monitoring for adverse events [2]. The development of new oral chemotherapeutic agents is needed to further reduce the course of treatment of TB, and to reduce both the governmental costs of treating TB patients and patient-incurred costs (including travel, nutrition, loss of productivity, physical and mental issues and delayed career growth in young patients). measurements of MtInhA activity showed that it is a β-NADH-dependent enoyl-ACP (acyl carrier protein) reductase enzyme and thus a member of the FAS-II system [12]. MtInhA was also shown to be specific for long-chain (C 18 > C 16 ) enoyl thioester substrates [12]. MtInhA was shown to be a major target for isoniazid (INH) [13][14][15], which is the most prescribed first-line chemotherapeutic agent for both the treatment of active TB and for prophylaxis of this disease.
The MtInhA (EC 1.3.1.9) enzyme catalyzes the hydride transfer from the 4S hydrogen of β-NADH to carbon-3 of long-chain 2-trans-enoyl thioester substrates (enoyl-ACP or enoyl-CoA) and induces enolate formation, followed by protonation, yielding NAD + and acyl (enoyl-ACP or acyl-CoA) products [12,16,17] (Figure 1). Studies on steady-state kinetics suggested that wild-type MtInhA follows a sequential kinetic mechanism in which β-NADH binds first, followed by enoyl-CoA binding to form the catalytically competent ternary complex [12]. On the other hand, primary kinetic deuterium isotope effects on enzyme activity suggested a random-order mechanism of substrate binding to wild-type MtInhA [16]. The latter mechanism has been supported by measurements of quench in intrinsic protein fluorescence at equilibrium and stopped-flow rates for 2-trans-dodecenoyl-CoA (DD-CoA) substrate binding to wild-type MtInhA [18]. A thorough description of chemical and kinetic mechanisms, structural data and inhibitors of MtInhA enzyme activity has recently been reported [19]. The graphical abstract of our contribution strives to provide both a summary of past enzyme kinetic data and the conclusions reached by the studies described herein. The hyperbolic equilibrium binding of NADH to free wild-type MtInhA, which was assessed by measuring the enhancement in nucleotide fluorescence, suggested that the initial nucleotide binding process is followed by a conformational change in the binary complex, yielding an overall dissociation constant value of 0.57 µM [14]. The kinetics of NADH binding to wild-type MtInhA and isoniazid-resistant mutant proteins have been investigated via fluorescence spectroscopy with stopped-flow equipment [20]. The stopped-flow traces showed a biphasic enhancement in nucleotide fluorescence for NADH binding to wild-type and isoniazid-resistant MtInhA mutant proteins [20]. The proposed mechanism included a fast phase for the rapid association of NADH to one conformer of an enzyme and a slow phase that is rate-limited by the slow conversion of two pre-existing conformers of free enzymes in solution [20]. Notwithstanding, the typical hyperbolic decrease in the slow-phase rate constant values with increasing nucleotide concentration could not be observed for wild-type MtInhA [20], as detection of this process also depends on whether or not the equilibrium is displaced to the conformational state of free enzymes that bind NADH. On the other hand, the overall dissociation constant derived from analysis of the total amplitude changes for the fast and slow phases of the stopped-flow biphasic enhancement in nucleotide fluorescence was in agreement with the values obtained from the analysis of rate constants, thereby supporting the proposed mechanism by which free MtInhA enzymes exist in equilibrium between two conformers and only one of them binds NADH [20].
Binding of the DD-CoA substrate to free MtInhA was monitored by measuring the quench in intrinsic protein fluorescence upon binary complex formation [18]. Equilibrium fluorescence spectroscopy data were sigmoidal, suggesting positive cooperativity upon DD-CoA binding to free enzymes [18]. As studies on steady-state kinetics and equilibrium binding cannot provide insights into isomerization of central complexes or individual rates of substrate binding to free enzymes, stopped-flow measurements of partial reactions or elementary steps were employed to remove these limitations as transient kinetics observe the changes occurring in the enzyme molecule itself and thus help clarify the elementary steps of enzyme reaction. Stopped-flow traces of quench in intrinsic protein fluorescence for MtInhA:DD-CoA binary complex formation were biphasic [18]. The observed apparent rate for the fast phase showed a linear dependence on increasing DD-CoA concentration, suggesting a single-step reversible bimolecular association process. On the other hand, the observed apparent rate for the slow phase displayed a hyperbolic increase, suggesting a bimolecular association process followed by a slow unimolecular isomerization of the MtInhA:DD-CoA binary complex [18].
On one hand, the studies on NADH binding suggested two forms of free MtInhA in solution prior to nucleotide binding, followed by a conformational change in the binary complex with no cooperativity. On the other hand, enoyl-CoA binding results suggest a bimolecular association process followed by a slow isomerization of the binary complex and that there is positive homotropic cooperativity. There are two generally invoked allosteric mechanisms for sigmoidal binding data: (1) the symmetry or concerted model, which predicts that two forms of free enzymes exist at equilibrium in solution [21], and (2) the sequential or induced-fit model, which suggests one form of free enzymes in solution and a slow isomerization of the binary complex upon ligand binding to the enzyme [22]. The isomerization steps are necessarily slower than the ligand binding steps in both the symmetry and sequential models [21,22]. The symmetry model for cooperative transitions can be divided into two systems: (1) K systems, in which the transition from low to high or from high to low affinity is triggered by substrate binding to an allosteric site, and (2) V systems, in which the substrate has the same affinity for the two states and acceleration or deceleration can be observed only if the two states of the protein differ in their catalytic activity, and the substrate can act as an effector if binding to an allosteric site of active (positive V system) or inactive (negative V system) state occurs [21]. Whether or not free MtInhA exists in two interconvertible forms and only one of them binds NADH (symmetry model), in which no cooperativity can be observed at equilibrium, or the enoyl-CoA substrate binds to one form of free enzymes in solution followed by binary complex conformational change with positive cooperativity (induced-fit model) observed at equilibrium, remains to be solved. It has long been recognized that records of time courses following either the formation of products or detection of transient intermediates during the first turnover of enzymes are uniquely suited for showing either the independence of sites or interactions between sites of oligomeric polypeptides (quaternary structure) during catalysis [23]. Thus, the equivalence or non-equivalence of the enzyme active sites may be investigated by measuring rates of product formation under conditions in which enzyme active-site concentration is in excess over substrate concentrations (single-turnover experiments) with varying degrees of occupancy of the enzyme active sites of oligomeric proteins. For instance, single-exponential records, regardless the degree of site occupancy, suggest a lack of interaction between the enzyme active sites. It appears appropriate to distinguish between the transient of reactants and the transient of intermediates [23]. Following the time course of total concentration of product (enzyme-bound and free) informs on the transient of reactants, thereby yielding purely kinetic information [23]. Detecting the changes in concentration of the different forms of enzymes (free enzyme, enzyme substrate and enzyme product complexes) informs on the transient of intermediates [23]. Accordingly, the single-turnover data presented herein are those of studies on the transient of reactants. The relevance of our contribution to the target-based rational design of functional inhibitors of MtInhA enzyme activity as potential anti-TB agents is discussed in Section 4 (Conclusions).

Burst
To determine whether product release is part of the rate-limiting step, pre-steadystate kinetic measurements of the reaction catalyzed by MtInhA were performed using an Applied Photophysics SX 18MV-R stopped-flow spectrofluorometer on absorbance mode. The decrease in absorbance was monitored at 340 nm (1 mm slit width = 4.65 nm spectral band) and 25 • C, using a time course of 500 ms, 400 data points for each time base and an optical path of 10 mm. The experimental conditions were 10 µM MtInhA, 300 µM NADH and 225 µM DD-CoA in 100 mM Pipes pH 7.0 (mixing chamber concentrations). The experimental conditions for the control experiment were 300 µM NADH and 225 µM DD-CoA in 100 mM Pipes pH 7.0 (mixing chamber concentrations). The dead time of the stopped-flow equipment was 1.37 ms. The stopped-flow trace shown represents the average of ten individual reactions of MtInhA.

Results and Discussion
The subheadings presented below strive to provide a concise description of the experimental results, data analysis and interpretation for the transient of reactants under single-turnover experimental conditions and for burst in product formation.

Single Turnover
The decrease in NADH concentration upon double-bond reduction of 2-trans-dodecenoyl-CoA (DD-CoA) by MtInhA under single-turnover conditions was followed in the stoppedflow instrument. The large excess of the second substrate (either DD-CoA or NADH) was Future Pharmacol. 2023, 3 384 employed to ensure that binding to the binary complex would not play any role in limiting enzyme catalysis, which should hopefully simplify the reaction model.
A quadratic equation (Equation (1)) was used to estimate the occupancy of MtInhA substrate binding sites. Equation (1) represents the solution of a quadratic equation for a simple binding process [25].
The pre-steady-state time course of decreasing absorbance at 340 nm upon NADH conversion to NAD + was fitted to a single-exponential decay equation (Equation (2)), in which A is the absorbance at time t, A 0 is the absorbance at time zero, and k obs is the apparent first-order rate constant for product formation (Figure 2). The k obs values for each NADH concentration are given in Table 1. Here, it is assumed that conversion of substrates to products goes to completion.
As pointed out above, single-turnover experiments can be used to test whether the sites of oligomeric enzymes are kinetically identical. The results for the single-turnover experiment without incubation show a linear relationship between k obs values and an increasing number of MtInhA sites occupied by NADH (Figure 3). These results suggest that the active sites of tetrameric MtInhA are kinetically equivalent, which is also borne out by the single-exponential records for any degree of site occupancy. However, it is also possible that NADH binding occurs after DD-CoA substrate binding to form the ternary complex and the ensuing turnover as there was no pre-incubation of enzymes with a reduced dinucleotide substrate.
The pre-steady-state time course of decreasing absorbance at 340 nm upon NADH conversion to NAD + was fitted to a single-exponential decay equation (Equation (2)), in which A is the absorbance at time t, A0 is the absorbance at time zero, and kobs is the apparent first-order rate constant for product formation (Figure 2). The kobs values for each NADH concentration are given in Table 1. Here, it is assumed that conversion of substrates to products goes to completion.

=
(2)  As pointed out above, single-turnover experiments can be used to test whe sites of oligomeric enzymes are kinetically identical. The results for the single-t experiment without incubation show a linear relationship between kobs values increasing number of MtInhA sites occupied by NADH (Figure 3). These results that the active sites of tetrameric MtInhA are kinetically equivalent, which is als out by the single-exponential records for any degree of site occupancy. However, possible that NADH binding occurs after DD-CoA substrate binding to form the complex and the ensuing turnover as there was no pre-incubation of enzyme reduced dinucleotide substrate. To try remove this ambiguity, another experiment was carried out in w enzyme MtInhA (15 μM of tetramers) was pre-incubated with NADH (5, 10 or 14 10 min before starting the reaction with a large excess of DD-CoA (225 μM) in the s flow instrument (mixing chamber concentrations). The pre-steady-state time cour reaction with the pre-incubation of MtInhA and NADH was fitted to a double-exp To try remove this ambiguity, another experiment was carried out in which the enzyme MtInhA (15 µM of tetramers) was pre-incubated with NADH (5, 10 or 14 µM) for 10 min before starting the reaction with a large excess of DD-CoA (225 µM) in the stopped-flow instrument (mixing chamber concentrations). The pre-steady-state time course of the reaction with the pre-incubation of MtInhA and NADH was fitted to a double-exponential  (3)), yielding values for the apparent rate constant for the fast (k obs1 ) and slow (k obs2 ) phases of the reaction.
The results for the number of active sites of tetrameric enzymes occupied by NADH prior to catalysis are given in Table 1. The k obs1 decreased as a function of increasing NADH concentration (Table 1, Figure 4). These data indicate an isomerization step between two forms of the MtInhA:NADH binary complex in solution and binding of DD-CoA to one of the conformers [20]. The fairly constant values for k obs2 (Figure 4) are in agreement with the k cat values (9 s −1 ) for the MtInhA enzyme [24]. Notwithstanding, the k obs1 values are larger than k cat and hence appear not to play a role in limiting the enzyme-catalyzed chemical reaction. Interestingly, the double-exponential traces suggest that the active sites are not equivalent, which is in agreement with the dependence of k obs1 on the degree of site occupancy, suggesting an equilibrium in the solution of two conformers of the MtInhA:NADH binary complex. The studies on the transient of reactants described herein are in agreement with the hyperbolic NADH binding at equilibrium [14] and the pre-steadystate transient found in intermediate studies [20].

= + (3)
The results for the number of active sites of tetrameric enzymes occupied by NADH prior to catalysis are given in Table 1. The kobs1 decreased as a function of increasing NADH concentration (Table 1, Figure 4). These data indicate an isomerization step between two forms of the MtInhA:NADH binary complex in solution and binding of DD-CoA to one of the conformers [20]. The fairly constant values for kobs2 (Figure 4) are in agreement with the kcat values (9 s −1 ) for the MtInhA enzyme [24]. Notwithstanding, the kobs1 values are larger than kcat and hence appear not to play a role in limiting the enzyme-catalyzed chemical reaction. Interestingly, the double-exponential traces suggest that the active sites are not equivalent, which is in agreement with the dependence of kobs1 on the degree of site occupancy, suggesting an equilibrium in the solution of two conformers of the MtInhA:NADH binary complex. The studies on the transient of reactants described herein are in agreement with the hyperbolic NADH binding at equilibrium [14] and the presteady-state transient found in intermediate studies [20]. For the single-turnover experiments with MtInhA pre-incubated with a DD-CoA substrate in single-turnover kinetics, attempts were made to fit the data to the doubleexponential decay equation, as was performed for NADH. However, single and double equations yielded statistically indistinguishable results. The values for kobs1 and kobs2 for the double equation were fairly similar (results not shown). Accordingly, singleexponential decay (Equation (2)) was employed, and the kobs values are given in Table 1. The results demonstrate a non-linear increase in kobs as a function of increasing DD-CoA concentration ( Figure 5). Although the single-exponential traces indicate equivalent binding sites, the non-linear dependence of kobs suggests positive cooperativity that depends on the number of enzyme sites occupied by DD-CoA that increase the binding and/or turnover rate upon NADH association with the MtInhA:DD-CoA binary complex. A slow isomerization of the MtInhA:DD-CoA binary complex prior to NADH binding to form the ternary complex and induce catalysis can be ruled out as kobs values would decrease as a function of increasing active site occupancy. The results presented herein For the single-turnover experiments with MtInhA pre-incubated with a DD-CoA substrate in single-turnover kinetics, attempts were made to fit the data to the doubleexponential decay equation, as was performed for NADH. However, single and double equations yielded statistically indistinguishable results. The values for k obs1 and k obs2 for the double equation were fairly similar (results not shown). Accordingly, single-exponential decay (Equation (2)) was employed, and the k obs values are given in Table 1. The results demonstrate a non-linear increase in k obs as a function of increasing DD-CoA concentration ( Figure 5). Although the single-exponential traces indicate equivalent binding sites, the non-linear dependence of k obs suggests positive cooperativity that depends on the number of enzyme sites occupied by DD-CoA that increase the binding and/or turnover rate upon NADH association with the MtInhA:DD-CoA binary complex. A slow isomerization of the MtInhA:DD-CoA binary complex prior to NADH binding to form the ternary complex and induce catalysis can be ruled out as k obs values would decrease as a function of increasing active site occupancy. The results presented herein are in agreement with positive cooperativity upon DD-CoA binding to free MtInhA detected by fluorescence spectroscopy at equilibrium (K' = 8.2 µM; n = 2) [18]. Moreover, the studies on the transient of reactants described herein are also in agreement with those in which analysis of total stopped-flow signal amplitude showed that the changes in biphasic quench in protein fluorescence for binary complex formation were sigmoidal (K' = 14.1 µM; n = 2.5), and isomerization of the MtInhA:DD-CoA binary complex was proposed [18]. Unfortunately, the experimental conditions precluded the possibility of achieved higher subunit occupancy and thus providing a more thorough description of the dependence of k obs on DD-CoA concentration. At any rate, data for the initial part of a sigmoidal increase in k obs as a function of DD-CoA appear to have been captured ( Figure 5). are in agreement with positive cooperativity upon DD-CoA binding to detected by fluorescence spectroscopy at equilibrium (K' = 8.2 μM; n = 2) [1 the studies on the transient of reactants described herein are also in agreem in which analysis of total stopped-flow signal amplitude showed that t biphasic quench in protein fluorescence for binary complex formation were = 14.1 μM; n = 2.5), and isomerization of the MtInhA:DD-CoA binary proposed [18]. Unfortunately, the experimental conditions precluded the achieved higher subunit occupancy and thus providing a more thorough the dependence of kobs on DD-CoA concentration. At any rate, data for the i sigmoidal increase in kobs as a function of DD-CoA appear to have been cap 5).

Burst
Transient kinetics of the forward reaction catalyzed by MtIn concentrations of enzymes and substrates were performed to determine wh release is part of the rate-limiting step of catalysis. The pre-steady-state tim reaction ( Figure 6) was fitted to an equation describing single-expo (Equation (2)). This analysis yielded a value of 0.680 ± 0.003 s −1 for the appar rate constant. The value of 0.68 s −1 for the change in absorbance at 34 exponential decay in the pre-steady-state experiment for the MtInhA corr s −1 (using Δε 6220 M −1 cm −1 , optical path 1 cm and MtInhA subunit conce μM). This result is in reasonably good agreement with the kcat determined velocity experiment in steady-state kinetics, which ranged from 2.8 s −1 in pH 6.8 at 25 °C [12] and 4.6 s −1 in 30 mM PIPES 150 mM NaCl pH 6.8 at 25 in 100 mM Pipes 100 mM pH 7.0 at 25 °C [24]. The observation of burst course in the pre-steady state phase of the reaction is evidence of a build-up the active site prior to being released into solution. If a burst is observ transient phase, and the concentration of NAD + produced is approximatel enzyme subunit concentration at time zero, it would indicate that the prod slower step compared to the chemical step of the reaction [26]. The stoppe indicate that product release does not contribute to the rate-limiting step o catalyzed chemical reaction as no burst in product formation could be detec

Burst
Transient kinetics of the forward reaction catalyzed by MtInhA at large concentrations of enzymes and substrates were performed to determine whether product release is part of the rate-limiting step of catalysis. The pre-steady-state time course of the reaction (Figure 6) was fitted to an equation describing single-exponential decay (Equation (2)). This analysis yielded a value of 0.680 ± 0.003 s −1 for the apparent first-order rate constant. The value of 0.68 s −1 for the change in absorbance at 340 nm for the exponential decay in the presteady-state experiment for the MtInhA corresponds to 3.8 s −1 (using ∆ε 6220 M −1 cm −1 , optical path 1 cm and MtInhA subunit concentration of 10 µM). This result is in reasonably good agreement with the k cat determined by the initial velocity experiment in steady-state kinetics, which ranged from 2.8 s −1 in 30 mM PIPES pH 6.8 at 25 • C [12] and 4.6 s −1 in 30 mM PIPES 150 mM NaCl pH 6.8 at 25 • C [16] to 9 s −1 in 100 mM Pipes 100 mM pH 7.0 at 25 • C [24]. The observation of burst during a time course in the pre-steady state phase of the reaction is evidence of a build-up of product in the active site prior to being released into solution. If a burst is observed during the transient phase, and the concentration of NAD + produced is approximately equal to the enzyme subunit concentration at time zero, it would indicate that the product release is a slower step compared to the chemical step of the reaction [26]. The stopped-flow results indicate that product release does not contribute to the rate-limiting step of the MtInhA-catalyzed chemical reaction as no burst in product formation could be detected ( Figure 6).

Conclusions
The process by which biological macromolecules (mostly proteins) transmit the effect of binding at one site to another, often distal and functional site, is referred to as allostery, which is important for the regulation of biological activity [27]. As the experimental technologies have improved in sophistication, the concepts and models of allostery have evolved. The concept of allostery evolved from the pre-allostery Bohr effect to allostery from conformational change in a two-state model (1965,1966), dynamic allostery (1984), ensembles of multiple states (1999), energetic connectivity between residues of proteins and allosteric networks, to a unified view including thermodynamics, population shift and three-dimensional structure [28]. The largely qualitative, static images of end point protein structures (e.g., X-ray crystallography to provide structural information of the protein before and after perturbation) have been replaced with more quantitative, dynamic views of allostery [27]. Computational and experimental methods to study protein dynamics provide access to protein shape-shifting processes that may reveal mechanisms of allosteric communication and features such as cryptic pockets [29]. These studies may offer new therapeutic opportunities, including a better understanding of biological systems and diseases, and informing allosteric drug design efforts [30]. These opportunities include, but are not limited to, enhancement (rather than inhibition) of protein function, displacement of equilibrium to lower substrate affinity conformers, avoidance of off-target effects of inhibitors of function of proteins belonging to large families that show a high degree of conservation (e.g., kinases, NAD(P)-dependent enzymes) and targeting functional sites that are apparently undruggable (e.g., proteinprotein interactions, flat and extensively solvent-exposed sites) [29]. Allosteric sites for drug discovery serve as an alternative to enzyme active sites since their lower conservation may be translated into higher selectivity, lower metabolic interference and lower undesirable off-targets [29]. Moreover, allosteric sites can offer new opportunities for optimization of pharmacokinetic properties whenever it is needed [29]. These objectives are thought to be difficult to achieve because the rational-based approach is currently limited to employing intermolecular interactions between ligands and key

Conclusions
The process by which biological macromolecules (mostly proteins) transmit the effect of binding at one site to another, often distal and functional site, is referred to as allostery, which is important for the regulation of biological activity [27]. As the experimental technologies have improved in sophistication, the concepts and models of allostery have evolved. The concept of allostery evolved from the pre-allostery Bohr effect to allostery from conformational change in a two-state model (1965,1966), dynamic allostery (1984), ensembles of multiple states (1999), energetic connectivity between residues of proteins and allosteric networks, to a unified view including thermodynamics, population shift and three-dimensional structure [28]. The largely qualitative, static images of end point protein structures (e.g., X-ray crystallography to provide structural information of the protein before and after perturbation) have been replaced with more quantitative, dynamic views of allostery [27]. Computational and experimental methods to study protein dynamics provide access to protein shape-shifting processes that may reveal mechanisms of allosteric communication and features such as cryptic pockets [29]. These studies may offer new therapeutic opportunities, including a better understanding of biological systems and diseases, and informing allosteric drug design efforts [30]. These opportunities include, but are not limited to, enhancement (rather than inhibition) of protein function, displacement of equilibrium to lower substrate affinity conformers, avoidance of off-target effects of inhibitors of function of proteins belonging to large families that show a high degree of conservation (e.g., kinases, NAD(P)-dependent enzymes) and targeting functional sites that are apparently undruggable (e.g., protein-protein interactions, flat and extensively solvent-exposed sites) [29]. Allosteric sites for drug discovery serve as an alternative to enzyme active sites since their lower conservation may be translated into higher selectivity, lower metabolic interference and lower undesirable off-targets [29]. Moreover, allosteric sites can offer new opportunities for optimization of pharmacokinetic properties whenever it is needed [29]. These objectives are thought to be difficult to achieve because the rationalbased approach is currently limited to employing intermolecular interactions between ligands and key amino acids to sterically occlude key functional sites to inhibit undesirable biological activities [29].
An example for tuberculosis includes a synthetic azetidine derivative that is an allosteric inhibitor of tryptophan synthase (an allosterically regulated enzyme) and shown to kill M. tuberculosis [31]. These authors suggested that, as the inhibitor binds to α-β-subunit interface of tryptophan synthase and affects multiple steps in the overall enzyme-catalyzed chemical reaction, the resulting inhibition would not be overcome by changes in the metabolic environment [31]. This study also highlights the effectiveness of allosteric inhibition for dynamic protein targets that are essential in vivo despite being dispensable under in vitro conditions [31]. The single-turnover results described herein suggest that MtInhA displays a typical symmetry model [21] with free enzymes in equilibrium between two conformers, followed by NADH binding to one conformer and conformational change in this binary complex. On the other hand, a sequential model [22] with only one conformer of free MtInhA in solution to which DD-CoA binds with homotropic positive cooperativity was invoked. Incidentally, molecular dynamics simulations have shown a normal distribution of free MtInhA enzymes in closed and open conformations [24]. Moreover, classical molecular dynamics simulations of tetrameric MtInhA showed protein flexibility that was borne out by the conformational space sampled by apo, binary and ternary complex forms of enzymes [32]. Molecular dynamics has been employed to study the influence of restrictions of structural flexibility on the A loop (F97-H121) and B loop (D150-A167) of the substrate-binding pocket to propose a monomeric structure that mimics the biologically active tetrameric MtInhA enzyme [33]. These efforts were made to propose a restrained force constant monomeric model system that describes the flexibility of the quaternary structure of MtInhA to evaluate ligands identified via virtual screening without losing dynamics information at a lower computational cost [33]. Future efforts to develop MtInhA enzyme inhibitors may include modeling of ligands to either the protein:DD-CoA binary complex or free enzymes with a larger dissociation constant for the formation of the protein:NADH binary complex.