|Home | About | Journals | Submit | Contact Us | Français|
Mycobacterium tuberculosis KatG enzyme functions both as catalase for removing hydrogen peroxide (H2O2) and as peroxidase for oxidating isoniazid (INH) to active form of anti-tuberculosis drug. Although mutations in M. tuberculosis KatG confer INH resistance in tuberculous patients, structural bases for INH-resistant mutations in the KatG gene remains poorly understood. Here, three M. tuberculosis KatG mutants bearing Arg418→ Gln, Ser315 → Thr, or Trp321 → Gly replacement were assessed for changes in catalase-peroxidase activities and possible structure bases relevant to such changes. These three M. tuberculosis KatG mutants exhibited a marked impairment or loss of catalase-peroxidase activities. The possible structural bases for the mutant-induced loss of enzyme activities were then analyzed using a three-dimensional model of M. tuberculosis KatG protein constructed on the basis of the crystal structure of the catalase-peroxidase from Burkholderia pseudomallei. The model suggests that three M. tuberculosis KatG mutants bearing Arg418 → Gln, Ser315 → Thr, or Trp321 → Gly replacement affect enzyme activities by different mechanisms, although each of them impacts consequently on a heme-associated structure, the putative oxidative site. Moreover, in addition to the widely accepted substrate-binding site, M. tuberculosis KatG may bear another H2O2 binding site. This H2O2 binding site appears to interact with the catalytic site by a possible electron-transfer chain, a Met255-Tyr229-Trp107 triad conserved in many catalase-peroxidases. The Ser315 → Thr mutant may have direct effect on the catalytic site by interfering with electron transfer in addition to the previously proposed mechanism of steric constraint.
Tuberculosis remains an important life-threatening disease with annual death of two million people worldwide.2 One of important reasons for the leading death is the epidemics of drug-resistant strains of Mycobacterium tuberculosis.8 While isoniazid (isonicotinic acid hydrazide, INH) maintains one of the most important and widely used anti-tuberculosis drugs, INH-resistant tuberculosis constitutes a major challenge for anti-tuberculosis treatment.12 Elucidation of mechanisms for INH resistance, therefore, is an important step for the global control of tuberculosis and multidrug-resistant tuberculosis.
Genetic studies have demonstrated that the catalase-peroxidase, encoded by M. tuberculosis katG, plays a role in resistance to INH.42 It has also been shown that molecular modifications of M. tuberculosis katG result in the abrogation or diminution of catalase activity and confer high-level resistance to isoniazid in more than 80% of isoniazid-resistant strains of M. tuberculosis. 10,43 As a bifunctional enzyme belonging to class I peroxidase superfamily, the catalase-peroxidase of M. tuberculosis (M. tuberculosis KatG) exhibits a predominant catalase activity and a substantial peroxidase activity.9,20,21,23,24,38,42 On one hand, KatG is a component of the oxidative defense system of M. tuberculosis and functions primarily as a catalase to remove hydrogen peroxide (H2O2).13 Such catalase activity is regarded as one of the virulence factors for M. tuberculosis,18,39 because this metabolism facilitates mycobacterial survival in macrophage.41 On the other hand, M. tuberculosis KatG functions as a peroxidase, oxidazing INH from a prodrug chemical to an electrophilic species, which binds to and inactivates InhA, an acyl carrier protein,1,28 and other targets.16,17 These pharmacological reactions result in the inhibition of the biosynthesis of cell wall mycolic acids (long-chain α-branched, β-hydroxylated fatty acids) in M. tuberculosis and make the mycobacterium susceptible to bactericidal effects derived from reactive oxygen radicals and other environmental factors.22 Multiple mutations interfering with KatG-mediated activation of INH can be selected by M. tuberculosis for the development of INH resistance.
Despite a decade-long investigation of M. tuberculosis KatG, detailed mechanisms by which M. tuberculosis KatG catalyzes H2O2 and activates INH have not been precisely elucidated. Specifically, it is not definitely known about where the substrate-binding site is located in M. tuberculosis KatG, how the electrons transfer between the substrate and the enzyme, and what steps of catalytic procedures are affected by the mutations of INH-resistant M. tuberculosis KatG mutants. Furthermore, questions remain unsolved as to which of structural requirements accounts for catalase activity, and what is responsible for peroxidase function in KatG catalase-peroxidases. Bifunctional catalase-peroxidases identified to date share sequence similarity with fungal cytochrome c (CCP) and plant ascorbate peroxidases (APXs), members of the class I peroxidase family.38 Crystallography analyses of bifunctional catalase-peroxidases and mono-peroxidases should shed a light on the structural bases for both catalytic requirements and a change in enzyme activities mediated by INH-resistant mutants.25,26 Nevertheless, such studies have been hindered by the absence of the crystal structure of M. tuberculosis KatG. Interestingly, crystal structures of two catalase-peroxidases from Haloarcula marismortui and Burkholderia pseudomallei, respectively (H. marismortui CP and B. pseudomallei KatG), have recently been reported.4,40 The crystal structures of these catalase-peroxidases may provide some information regarding the structure–function relation, because M. tuberculosis KatG shares 55% and 65% similarities in amino acid sequences to H. marismortui CP and B. pseudomallei KatG, respectively.4,40
SWISS-MODEL, an automated protein homology modeling server, has recently been used as a molecular simulation tool to construct three-dimensional structures of proteins using the known crystal structure of a protein homologue as a template.33 In the present study, SWISS-MODEL was employed to build model structures of M. tuberculosis KatG using the crystal structure of B. pseudomallei KatG as a template. On the basis model structures, three INH-resistant mutants of M. tuberculosis KatG were assessed for changes in enzyme activities and possible structural bases underlying those changes. Our results suggest that three mutations in the INH-resistant mutants may reduce the catalytic activities of M. tuberculosis KatG in different ways. The analyses also imply that besides the widely accepted putative substrate-binding site near Ser315 and the electron transfer chain connecting this site with the catalytic center,4,10,29,37,40,41 M. tuberculosis KatG may have additional substrate-binding site and another electron transfer chain. As far as we are concerned, our studies represent the first structural modeling of potential structures of M. tuberculosis KatG.
The three mutants of MtKatG, as well as the wild-type MtKatG, were studied. E. coli DH12s and BL21(DE3) were used as host cell for pET24b (Novagen) plasmid transformation and for expression of hexahistidine-tagged KatG proteins (His6-KatG), respectively. E. coli cells were grown in Luria-Bertani broth at 37°C in the presence of kanamycin (50 µg/ml).
The mutagenic oligonucleotides are listed in Table 1. The restriction nuclease fragments that were mutagenized following the Kunkel procedure,14 sequenced, and subsequently reincorporated into pET24b to generate the mutated katG gene are also listed. Sequence confirmation of all sequences was by the Sanger method30 on double-stranded plasmid DNA generated in E. coli DH12s.
Luria-Bertani broth (2 ml) containing kanamycin (50 µg/ml) was inoculated with a fresh colony of recombinant E. coli BL21(DE3)::His6-KatG and grown at 37°C overnight. The in-oculum was then transferred to 50 ml of the same medium in a 250-ml shaking flask. When the culture reached an A600 value of 0.6–0.8, isopropyl β-d-thiogalactoside (IPTG) was added to a final concentration of 1 mM. After incubation for 21 hr at 22°C, cells were harvested by centrifugation at 5,000 rpm for 10 min at 4°C. The pellet was resuspended in 1.5 ml of buffer A comprised of 10 mM pH 8.0 Tris-HCl, 50 mM NaCl, 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 0.5 mg/ml lysozyme. The mixture was incubated on ice for 30 min. Then, DNase and MgSO4 were added to the mixture to final concentration of 0.1 mg/ml and 10 mM, respectively. Subsequently, the mixture was incubated for 30 min at 4°C and then frozen at −70°C for 30 min. After thawing, the cycle of freezing and thawing was repeated for a total of three times. The insoluble debris was removed by centrifugation at 15,000 rpm for 10 min at 4°C. The supernatant was collected in a fresh tube and KatG proteins were purified by His·Bind® column chromatograph according to the manufacturer’s instructions (Novagen) and then Q sepharose F. F. (pre-equilibrated with 0.1 M sodium phosphate buffer pH 7.0). The fractions containing the recombinant MtKatG protein were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).
Catalase activity was determined spectrophotometrically by measuring the decrease in H2O2 concentration at 240 nm (ε240 = 0.0435mM−1·cm−1), at 25°C. The reaction mixture (1 ml) contained 50 mM sodium phosphate buffer pH 7.0 and 18 mM H2O2. Peroxidase activity was determined spectrophotometrically by measuring the rate of oxidation of 0.1 mM o-dianisidine at 460 nm (ε460 = 11.3mM−1·cm−1) at 25°C, in the presence of 23 mM t-butyl hydroperoxide in 50 mM sodium acetate buffer pH 5.2.15 Specific activities were expressed as enzyme units/mg of total protein. The protein concentration was determined by the Bradford method,3 using bovine serine albumin (BSA) as a standard.
The enzyme (1.22 µM) was in 50 mM Na2HPO4 pH 7.5 at 25°C. Fluorescence measurements were carried out with an Hitachi fluorescence spectrophotometer (Model F-4010). Fluorescence emission spectra were obtained by exciting at 280 nm and recording between 300 and 400 nm.7
A model of the M. tuberculosis KatG protein was constructed by SWISS-MODEL on the basis of the known three-dimensional structure of the B. pseudomallei KatG (BpKatG), which was the most related catalase-peroxidase found in the Protein DataBank (65% identity).4 Autodock software was used to dock the ligand, the heme, to its receptor.44
INH-resistant M. tuberculosis KatG mutants bearing Arg418 → Gln, Ser315 → Thr, and Trp321 → Gly replacements exhibit a loss or reduction of catalase/peroxidase activities and fluorescence emission. Recently, we examined mutations in M. tuberculosis KatG in clinical isolates obtained from patients with INH-resistant tuberculosis. Our studies identified three common mutations that result in amino acid replacements: Arg418 → Gln, Ser315 → Thr, and Trp321 → Gly.45 In fact, the mutation bearing Arg418 → Gln replacement in M. tuberculosis KatG was initially found in an INH-resistant population in China.45 Although the capacity of these mutants to resist to INH is well defined,19,27,35,36,45 the structural bases for explaining INH-resistant mutants bearing those amino acid replacements remain poorly investigated. One of the major reasons for this is the absence of crystal structures of M. tuberculosis KatG protein. As the first step to explore the structure– function correlate of INH-resistant mutants, we expressed and purified the M. tuberculosis KatG proteins, which carried a single mutation that leads to Arg418 → Gln, Ser315 → Thr, and Trp321 → Gly replacements, respectively (Fig. 1). These purified enzyme proteins were then assessed for catalase and peroxidase activities as well as tryptophan fluorescence spectrum. When compared with the wild-type KatG, both mutants bearing Arg418 → Gln and Trp321 → Gly replacements exhibited undetectable catalase activities and significantly impaired peroxidase activities (5.6% and 14.8% of the wild-type level, respectively). The mutant with Ser315 → Thr replacement showed 16.5% of the catalase activity and 14.5% of the peroxidase activity, respectively (Table 2).
Furthermore, analyses of fluorescence emission spectra for these wild-type and mutant enzymes showed that the maximum intensity of two KatG mutants bearing Arg418 → Gln and Ser315 → Thr replacements were reduced 4% and 15%, respectively, whereas the maximum emission of these two mutants exhibited neither blue shift nor red shift (749.8 a.u. for the mutant Ser315 → Thr and 672.8 a.u. for the mutant Arg418 → Gln) (Fig. 2). The mutant Trp321 → Gly was involved in the modification of the amino acid residue Trp, and, therefore, was unsuitable for the spectrum analyses. Thus, these results provide evidence that INH-resistant mutants bearing Arg418 → Gln, Ser315 → Thr, or Trp321 → Gly replacement exhibit a loss or reduction of catalase/peroxidase activities and fluorescence emission. Such findings would allow us to undertake crystallography studies of these mutants with respective amino acid replacements and to explore structure bases for a loss or reduction in catalytic function of the mutants.
To explore structure bases for malfunction of mutations in M. tuberculosis KatG, SWISS-MODEL was employed to examine the extent to which the crystal structure of M. tuberculosis KatG resembles that of its functional counterpart, B. pseudomallei KatG. SWISS-MODEL has been widely used for homology-building three-dimensional structure for any proteins that share >25% identities in amino acids with a comparative protein with known crystal structures.33 M. tuberculosis KatG shares 65% identity in amino acids to B. pseudomallei KatG (Fig. 3). The M. tuberculosis KatG three-dimensional structure created by SWISS-MODEL was extremely similar to the reported crystal structure of B. pseudomallei KatG (Fig. 4).32 When the carbon backbones of B. pseudomallei KatG and M. tuberculosis KatG were superimposed, there was apparent similarity in topology between these two proteins. Like B. pseudomallei KatG, M. tuberculosis KatG protein contains mostly α-helices with few β-sheets (Fig. 4). The subunit model is organized in two domains, which are connected by a large contact area. It is important to note that M. tuberculosis KatG, like its B. pseudomallei counterpart, bears the putative oxidative structure, heme, which was buried inside the amino-terminal domain of each of these two proteins (Fig. 4). Importantly, the model demonstrates that the amino acid Arg418 stands in an environment enriched in Tyr and Try residues, which are directly related to tryptophan fluorescence emission (Fig. 5). This fine structure suggests that the mutation causing Arg418 replacement is likely to cause a change in the configuration of the Tyr and Try residues, and to impact on Tyr/Try-related chemical characteristics. In fact, such a prediction is consistent with the finding that the mutant bearing Arg418 → Gln replacement displays a decrease in the maximum intensity of fluorescence emission spectra (Fig. 2). Thus, crystallography of M. tuberculosis KatG protein can be built based on the crystal structure of B. pseudomallei. Such a homology model can be used to explore molecular explanation for a loss or decrease of enzyme activities of INH-resistant mutants of the M. tuberculosis KatG protein.
Given the three-dimensional model of wild-type M. tuberculosis KatG, we sought to examine how the Arg418 interacts with the surrounding residues and to determine whether a mutant bearing the Arg418 → Gln replacement was able to alter the heme-associated structure topology. The amino acid residue Arg418 is situated at the bottom of a dramatic cleft, which is formed between the two domains but localized in the side of the subunit (Fig. 6). The environment around the Arg418 is very similar to that surrounding its counterpart Arg426 in the B. pseudomallei KatG (Fig. 7). Because Arg418 is located on the surface of the bottom, the side-chain of Arg418 contributes to the cationic center on an otherwise predominantly anionic surface. Such a striking and well-defined cavity is proposed to provide another perfect binding site for a substrate, which is different from the previously putative substrate-binding site nearby Ser315.4,10,29,37,40,41 Moreover, Arg418 abuts against the Trp107-Tyr229-Met255 adduct, a fine structure conserved in many catalase-peroxidases.40 The existing evidence suggests that the adduct may be a general feature of KatGs (Fig. 8).40 Trp 107 in the adduct is the important Trp residue of the conserved triad in the distal pocket next to the heme (Fig. 9). The kinetic study reported by Regelsberger et al.11,26 indicates that this specific Trp residue plays a crucial role in compound I reduction mediated by H2O2. Compound I reduction is the second stage of the catalase catalytic process, whereas compound I formation is the first stage of both the catalase catalytic and peroxidase catalytic processes (Fig. 10). Studies of Synechocystis and Escherichia coli catalase-peroxidases have shown that the variants with a mutation that replaces the corresponding conserved distal Trp with an amino acid without the indole ring exhibit very low-level or even undetectable catalase activity. 26 It is proposed that the indole ring of the conserved Trp residue is essential for the catalytic process but not for compound I formation nor for the peroxidatic reduction stage.11,26
Analyses of Arg418 and surrounding amino acid triads in the heme vicinity facilitate explaining the loss of catalase and peroxidase activities for the KatG mutant bearing Arg418 → Gln replacement. It is noted that the Arg418 may form an interaction with the Tyr229 and Met255 by slightly moving its side chain and ulteriorly influencing the Trp107 (Fig. 9). Such an interaction appears to be important for catalase activity, because Arg418 and the associated adduct comprised of Trp107, Tyr229, and Met255 are conserved in most dual catalase-peroxidase enzymes but not in the mono-peroxidases (CCP and APX). On the basis of these findings, we propose that the first H2O2 reacts with the KatG at the initial substrate-binding site nearby Ser315 and oxidates the latter to form compound I. The second H2O2, acting as the reducing agent, is localized at the bottom of the U-shaped cleft, the second substrate binding site, and subsequently oxidated through its interaction with the cationic cluster that includes the side chain of Arg418. As indicated by Carpena,4 the sulfate of Met255 in the Met255-Tyr229-Trp107 adduct is most likely carrying a positive charge and would provide a draw for electrons from the cationic groups.
Therefore, we presume that in addition to the electron transfer chain connecting the first substrate binding site and the catalytic center, the Arg418 and its association with the adduct might provide another direct pathway for transferring electron to the heme and for reducing the compound I back to the resting state. Impairing the pathway may eliminate the catalase activity but remain the peroxidase activity. This prediction is consistent with the mono-peroxidase activity of CCP and APX as well as the abrogation of catalase activity by a nonconserved replacement of Arg418.11,26 The Arg418 → Gln replacement may impair the cationic cluster at the cleft and subsequently block the electron-transfer chain from the second H2O2 to the heme. This presumption could readily explain the complete loss of catalase activity in the INH-resistant mutant bearing the Arg418 → Gln replacement (Table 2).
However, why was the peroxidase activity of the mutant Arg418 → Gln partially reduced at the same time? One possible explanation is that the Arg418 → Gln replacement may result in a change in the conformation of the Trp107, which is involved in the interaction between Arg418 and the Met255-Tyr229-Trp107 adduct. This change may in turn lead to conformation changes in the distal pocket. Such conformation changes can presumably prevent the peroxidase cycle initiator with bulky side chain (t-butyl hydroperoxide) from interacting with the catalytic site and, consequently, affect the first stage of the peroxidasing process. Given that a small increase in the size of the active site cavity due to the Trp → Phe replacement in the distal pocket can alter peroxidase activity,11 the change in size and shape of the distal pocket caused by the Arg418 → Gln replacement may impact on the peroxidasing process.
We then sought to examine the circumstance of the residue Ser315 to explore the molecular mechanisms underlying a loss of enzyme activity for the INH-resistant mutant bearing the Ser315 → Thr replacement. The potential mechanisms for Ser315Thr mutant-mediated loss of KatG enzyme activity remain poorly understood. Although some studies suggest that the poor peroxidase activity of the mutant is likely to be related to the steric constraint introduced by the Thr residue replacement, 29,37,41 others report that there are reduced Vmax in hydrolytic reactions toward both the catalase and peroxidase substrates. 29 On the basis of the three-dimensional model of M. tuberculosis KatG, the residue Ser315 corresponds to the Ser324 in B. pseudomallei KatG (Fig. 11).4 The Ser315 is located near a heme propionate side chain whose carboxylate group may form a hydrogen bond with the side-chain of Ser315 (Fig. 11). The Ser315 → Thr replacement in this position may interfere with the development of the hydrogen bond and there-after affect the orientation of the heme and the formation of the 6-c heme. This scenario is supported by a recent finding seen in spectroscopic studies.41 It has been shown that the Ser315 → Thr replacement in M. tuberculosis KatG results in the predominance of 5-c heme in enzyme reactions, and such a predominance may be due to an increased difficulty in the conversion of the 5-c form to the 6-c form.41 It is likely that even a subtle change in the heme orientation due to the Ser315 → Thr replacement can influence the interaction between the heme and Trp107 through a mechanism that is independent upon the compound I conformation. Because Trp107 plays an important role in the second stage of the catalytic process, the Ser315 → Thr replacement may hinder a transfer of the electron from the reducing agent, i.e., the second H2O2, to the heme by its effect on heme orientation, and consequently impair the catalase activity.
Further studies of the three-dimensional structure of M. tuberculosis KatG revealed that the region surrounding Ser315 is composed predominantly of hydrophobic amino acid residues including Pro136, Ala139, Ala281, Val284, Trp300, and Ile313 (Fig. 11). The environment around Ser315 in M. tuberculosis KatG is indeed quite similar to that around the Ser324 of B. pseudomallei KatG, suggesting that these two regional structures may share a same aspect of structure–function. The Ser315 may participate in the catalytic reaction by both interacting with the hydrazine portion of INH or –CH2OH of the o-dianisidine and providing a direct route for transferring an electron from the substrate to the radical of compound I or compound II.4 Thus, the Ser315 → Thr replacement in the INH-resistant mutant is likely to interfere with the electron-transfer chain and the interaction between the substrate and the Ser main chain.
Finally, we sought to examine if the fine structures of the residue Trp321 and the heme environment of M. tuberculosis KatG could provide an explanation for the loss of enzyme activity for the INH-resistant mutant bearing the Trp321 → Gly replacement. Because the catalytic mechanisms associated with the heme environment in CCP have been elucidated, we compare Trp321-related fine structures between the crystal structure of CCP and that of the three-dimensional model of M. tuberculosis KatG built on the basis of B. pseudomallei crystal structure. The three-dimensional model of M. tuberculosis KatG suggests that the amino acid Trp321 is situated in the proximal pocket next to the heme (Fig. 12). Trp321, together with His270 and Asp381, constitutes a triad on the proximal side of the heme, whereas the triad comprised of Trp107, His108, and Arg104 is located on the side distal to the heme. Such fine structures of Trp321 and its surroundings resembled what was seen in the corresponding regions of CCP (Fig. 12). In the proximal pocket, His270 is presumed to be the fifth ligand to the heme iron atom as postulated in the setting of CCP.6 In addition, His270 is closely associated with Asp381, and the Asp381 appeared to interact with the indole N atom from the proximal Trp321. Trp321 in M. tuberculosis KatG is indeed the counterpart of Trp191 in CCP (Fig.12). Given that Trp191 in CCP plays a role in the first but not the second stage of the catalytic process,5 the corresponding Trp321 in M. tuberculosis KatG may have this same function for the peroxidase activity. The remarkable similarity in heme environments between M. tuberculosis KatG and CCP suggests that the compound I radical in M. tuberculosis KatG may dissociate from the heme to the Trp321, a similar situation as seen for Trp191 in CCP.5,31,34 On the basis of these fine structures of Trp321 and its surroundings, the INH-resistant mutant bearing the Trp321 → Gly replacement may cause a change in the heme environment and compound I radical reaction. Such a change may impact on catalase and peroxidase activities.
In conclusion, on the basis of the three-dimensional structure model of M. tuberculosis KatG, three M. tuberculosis KatG mutants bearing Arg418 → Gln, Ser315 → Thr, or Trp321 → Gly replacement influence enzyme activities by different mechanisms underlying a change in catalase-peroxidase structures. Our data suggest that in addition to the widely accepted substrate-binding site, there may be another H2O2 binding site. This site appears to be connected with the heme-associated catalytic site by a possible electron transfer chain, which is a Met255-Tyr229-Trp107 triad conserved in many catalase-peroxidases. Moreover, the Ser315 → Thr mutant may have direct effect on the catalytic site by interfering with the electron-transfer chain composed of the Met255-Tyr229-Trp107 triad in addition to the previously proposed mechanism of steric constraint.
This work was supported in part by the 10th 5-year research grant “Quick Detection of Multi-drug Resistant Tuberculosis” from the Ministry of Technology and Science, China. W.H.Z. and Z.W.C. were supported by National Institutes of Health grant R01 HL64560 (to Z.W.C.).