The crystal structure of isoniazid-bound KatG catalaseperoxidase from Synechococcus elongatus PCC7942 Saori Kamachi1, Kei Hirabayashi2,3, Masahiro Tamoi4, Shigeru Shigeoka4, Toshiji Tada1 and Kei Wada2 1 2 3 4
Graduate School of Science, Osaka Prefecture University, Sakai, Osaka, Japan Organization for Promotion of Tenure Track, University of Miyazaki, Japan Graduate School of Science, Osaka University, Toyonaka, Osaka, Japan Faculty of Agriculture, Kinki University, Nakamachi, Nara, Japan
Keywords antituberculous drug; catalase-peroxidase; crystal structure; isoniazid; KatG Correspondence T. Tada, Graduate School of Science, Osaka Prefecture University, Sakai, Osaka 599– 8531, Japan Fax: +81 72 254 9935 Tel:+81 72 254 9820 E-mail: [email protected] and K. Wada, Organization for Promotion of Tenure Track, University of Miyazaki, Miyazaki 889–1692, Japan Fax: +81 985 85 0873 Tel:+81 985 85 0873 E-mail: [email protected] (Received 5 September 2014, revised 26 September 2014, accepted 6 October 2014) doi:10.1111/febs.13102
Isoniazid (INH) is one of the most effective antibiotics against tuberculosis. INH is a prodrug that is activated by KatG. Although extensive studies have been performed in order to understand the mechanism of KatG, even the binding site of INH in KatG remains controversial. In this study, we determined the crystal structure of KatG from Synechococcus elongatus resolution. Three PCC7942 (SeKatG) in a complex with INH at 2.12-A INH molecules were bound to the molecular surface. One INH molecule was bound at the entrance to the e-edge side of heme (designated site 1), another was bound at the entrance to the c-edge side of heme (site 2), and another was bound to the loop structures in front of the heme propionate side chain (site 3). All of the interactions between KatG and the bound INH seemed to be weak, being mediated mainly by van der Waals contacts. Structural comparisons revealed that the identity and configuration of the residues in site 1 were very similar among SeKatG, Burkholderia pseudomallei KatG, and Mycobacterium tuberculosis KatG. In contrast, sites 2 and 3 were structurally diverse among the three proteins. Thus, site 1 is probably the common KatG INH-binding site. A static enzymatic analysis and thermal shift assay suggested that the INH-activating reaction does not proceed in site 1, but rather that this site may function as an initial trapping site for the INH molecule. Database The atomic coordinates and structure factors have been deposited in the Protein Data Bank under the accession number 3WXO.
Introduction Tuberculosis remains an important life-threatening disease, and is responsible for 1.5–2 million annual deaths worldwide [1]. Isoniazid (INH) is a widely used antituberculous prodrug that is activated by Mycobacterium tuberculosis itself via the catalase-peroxidase KatG. During INH activation, the INH molecule is cleaved by KatG to generate an isonicotinyl radical
(IN) and diazene, and subsequently the IN radical species is combined with NAD+ to generate an isonicotinyl-NAD (IN–NAD) covalent adduct in the presence of superoxide, which is formed by O2 reduction by KatG [2]. The IN–NAD adduct inhibits the longchain enoyl acyl carrier protein reductase (InhA) involved in the synthesis of mycolic acid [3], which is a
Abbreviations BpKatG, Burkholderia pseudomallei KatG; INH, isoniazid; IN, isonicotinyl radical; MPD, 2-methyl-2,4-pentanediol; MtKatG, Mycobacterium tuberculosis KatG; SeKatG, Synechococcus elongatus PCC7942; WT, wild-type.
54
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
S. Kamachi et al.
major component of the mycobacterial cell wall. Although the mechanism of inhibition of InhA by the IN–NAD adduct has been clearly demonstrated [4,5], the mechanism by which KatG forms IN species, including the exact binding sites for INH, remains unclear. Extensive structural studies of KatGs from several bacteria have been performed; the crystal structures of KatG from Haloarcula marismortui [6], Burkholderia pseudomallei (BpKatG) [7], M. tuberculosis (MtKatG) [8] and Synechococcus elongatus PCC7942 (SeKatG) [9] have been determined. However, an INH-bound structure has only been reported for BpKatG [2]. In that study, the INH-binding site was located on the opposite side of the BpKatG protein from the entrance to the heme cavity. Interestingly, NMR analyses of typical peroxidases, which had a very similar heme cavity to that of KatG, showed that the INH molecule could bind to the heme cavity [10]. Furthermore, in ascorbate peroxidase bearing the point mutation H42A, INH directly interacted with the heme iron [11]. More recently, the structure of BpKatG bearing the point mutation D141A demonstrated that the INH molecule could also enter into/bind to the heme cavity after removal of the ‘lid residue’ (Asp141) at the entrance to the funnel-shaped heme cavity, and a metadynamics analysis revealed that the mutation of Asp141 removed an energy barrier preventing entry INH into the heme cavity [12]. In addition, spectroscopic analyses with wild-type (WT) BpKatG also strongly implied that the INH molecule binds around the heme cavity [2]. More than half of the INH-resistant clinical isolates of M. tuberculosis contain a mutation in MtKatG, whereby the strictly conserved serine at position 315 (Fig. 1) is replaced with threonine (S315T MtKatG) [13–15]. Biochemical analyses have shown that S315T MtKatG retains only ~ 30–50% of the WT activity for IN–NAD formation [2]; this reduced rate of activity is sufficient to confer INH resistance in M. tuberculosis infections. To date, using S315T MtKatG and the corresponding BpKatG mutant (S324T), biochemical [2,16–18], spectroscopic [19–21] and structural [22,23] investigations have been carried out to find the factor (s) underlying the reduction in IN–NAD formation activity. Unfortunately, these studies yielded no apparent differences between the WT and mutant KatGs. One likely reason for this is that there has been no consensus concerning the INH-binding sites on KatG, probably owing to their low affinities, because the primary function of KatG in vivo is that of scavenging hydrogen peroxide. FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
Crystal structure of isoniazid-bound KatG
All KatG orthologs so far used in structural studies, which share at least 53% sequence identity (Fig. 1), have very similar 3D/4D structures. Here, we investigated INH-binding/IN–NAD formation characteristics by using SeKatG and the crystal structure of SeKatG resolution. in complex with INH molecules at 2.12-A In SeKatG, three INH-binding sites were characterized on the molecular surface, all of which were completely different sites from those observed in the INH-bound structure of BpKatG. However, one site was located at the entrance to the e-edge side of heme, where Ser308 [the residue corresponding to Ser315 in MtKatG (Fig. 1)] is located. A structural comparison between INH-bound SeKatG and BpKatG indicated that the entrance to the e-edge side of heme may be the common INH-binding site; however, the IN formation reaction seems not to proceed at this site. Therefore, the common INH-binding site found in this study may function as an initial trapping/recruitment site for INH molecules.
Results and Discussion Enzymatic IN–NAD formation by SeKatG and its mutant protein It has been reported that not only MtKatG but also BpKatG can enhance the conversion of INH and NAD+ to IN–NAD in the presence of Mn2+ and O2. To determine whether SeKatG also has the ability to induce IN–NAD formation, we performed an enzymatic assay (Table 1). WT SeKatG showed a specific activity of 33.3 nmolmin1mg1 protein for IN–NAD formation. In order to compare SeKatG and MtKatG, we also measured MtKatG activity under the same conditions; MtKatG had an activity of 35.1 nmolmin1mg1 protein. Thus, SeKatG possessed substantial activity for the conversion of INH to IN–NAD, and this activity was equivalent to that of MtKatG. When the S308T mutation, which is equivalent to the S315T mutation in MtKatG, was introduced into SeKatG, the IN–NAD formation Vmax was decreased by 42% as compared with the WT. Nevertheless, the Km values for INH and NAD+ were only subtly different between the WT protein (5.6 mM and 17.6 lM, respectively) and S308T protein (10.8 mM and 23.3 lM, respectively). The enzymatic nature of SeKatG was very similar to those of MtKatG and BpKatG; previous biological studies have shown that in MtKatG the S315T mutation results in an ~ 51% reduction and that in BpKatG the S324T mutation results in an ~ 41% reduction in in vitro IN–NAD formation, 55
Crystal structure of isoniazid-bound KatG
S. Kamachi et al.
Fig. 1. Multiple sequence alignment of KatGs from several representative organisms. The sequence numbering of SeKatG is shown. Identical residues are highlighted in light brown, and similar residues are boxed in blue. The residues of the catalytic Met-Tyr-Trp adduct are highlighted in black. The residues involved in INH binding in SeKatG are indicated by blue (site 1), green (site 2) and yellow (site 3) triangles, respectively. The figure was prepared with CLUSTALW [33] and ESPRIPT [34]. KatG orthologs are shown from S. elongatus PCC7942 (NCBI accession no. WP_011244741), M. tuberculosis (WP_003899075), B. pseudomallei (WP_011205232), H. marismortui (WP_011223478), and Synechocystis PCC6803 (WP_010872602).
56
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
S. Kamachi et al.
Crystal structure of isoniazid-bound KatG
Table 1. Rates of IN–NAD formation by WT and S308T SeKatG. The specific activities were evaluated at 37 °C in a reaction mixture containing 0.4 mM NAD+, 1.0 mM INH, 2 lM MnCl2, and 50 mM Tris/HCl (pH 7.8). SD, standard deviation.
WT SeKatG S308T SeKatG MtKatG
Specific activity (nmolmin1mg1 protein SD)
Vmax (nmolmin1mg1 protein)
INH (mM)
NAD+ (lM)
33.3 6.3 7.3 2.4 35.1 3.6
278.9 115.7 –
5.6 10.8 –
17.6 23.3 –
Km
without strongly affecting the Km values for INH and NAD+ [2]. Insights into why IN–NAD formation is reduced upon introduction of the S308T mutation came from a thermal shift assay (Fig. 2). S308T SeKatG showed a 4 °C decrease in its thermal denaturation temperature (Tm) relative to that of the WT, indicating that the A
B
S308T mutation affected protein stability. We postulate that the mutation at the Ser308 position, which is in the vicinity of the heme propionate side chain, may alter the stability of the heme cavity and/or the flexibility of the heme pocket entrance, thus leading to a decrease in IN–NAD formation activity. In other words, the invariant serine (Ser308 in SeKatG) is probably not involved directly in INH binding, but may contribute to recruiting the INH molecule. Indeed, it has been reported that INH may bind to the heme cavity of the BpKatG D141A mutant protein [12] and a typical peroxidase [11]. Thus, if IN–NAD formation occurs in the heme cavity, flexibility of the heme entrance is probably necessary to achieve incorporation of the INH molecule into the cavity. Conducting the thermal shift assay in the presence of INH revealed an additional effect of INH binding on stability, in that the Tm values of both the WT and Ser308 SeKatG proteins were decreased by 2.5 °C relative to their unbound Tm values. This result indicated that the INH molecule undoubtedly binds to SeKatG, and that its effect is large enough to change the thermostability. It should be noted that this method could not determine whether the bound INH was involved in IN–NAD formation; also, the number of occupied INH-binding site(s) was not determined. When the INH molecule(s) are bound to SeKatG, the protein thermostability deteriorates, raising the possibility that the INH molecule may trigger an increase in the flexibility of KatG. Overall structure of SeKatG in complex with INH molecules
Fig. 2. A thermal shift assay revealed that the S308T mutation and INH binding both affect the melting temperature of SeKatG. (A) Inverse first-derivative plots of WT (blue) and S308T (red) SeKatG. (B) Inverse first-derivative plots of WT (blue) and S308T (red) SeKatG in the presence of 1 mM INH. n = 5 for each assay; the Tm values and standard errors are shown below the curves.
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
The structure of INH-bound SeKatG was refined at resolution to R and Rfree values of 0.158 and 2.12-A 0.197, respectively. The asymmetric unit contained one subunit of an SeKatG molecule, including a protoporphyrin IX heme moiety, four cations (modeled as sodium ions), and 487 water molecules. The crystallographic equivalent subunit is created by a two-fold symmetry operation to form the functional dimer, as shown in Fig. 3A. Although the electron density for
57
Crystal structure of isoniazid-bound KatG
S. Kamachi et al.
A
B
C
D
SeKatG was mostly continuous and of high quality, the densities for the N-terminal segment (residues 1– 10) were poorly defined; accordingly, these residues were not included in the model. When the structure of INH-bound SeKatG was superposed on that of substrate-free SeKatG [9], no significant structural changes were observed in the enzyme; the rmsd for the Ca atoms was 0.15 A. Binding environments of the INH molecules The electron density map of INH-bound SeKatG revealed that three INH molecules were bound to the molecular surface (Fig. 3). One INH molecule was bound to the entrance of the e-edge side of heme, where Ser308 was located (designated site 1), another was bound to the entrance of the c-edge side of heme (site 2), and another was bound to the loop structures in front of the heme propionate side chain (site 3). All of the INH-binding sites were located in close proximity to heme, but there was no direct interaction with heme, as INH molecules did not enter the heme cavity. The INH-binding site is located at the entrance to the e-edge side of heme (site 1), where the nearest neighbor distances between INH and heme is ~ 5.6 A. This site consisted of Pro123, Asp124, Val193, Leu220, Asn224, Pro225, and Ser308 (Fig. 3B); this represents the first structural evidence that an INH molecule can interact with the invariant serines during IN–NAD 58
Fig. 3. SeKatG has three low-affinity INH-binding sites. (A) Overall structure of INH-bound SeKatG, and the locations of the INH-binding sites. The SeKatG dimeric structure was generated via a symmetry operation using the crystallographic two-fold axis, as the asymmetric unit contains one monomeric SeKatG. The individual subunits are shown in green or purple. The bound INH molecules are shown in blue, and heme is shown in red. (B–D) The electron density map showing details of the SeKatG INH-binding sites. An Fo – Fc map omitting the INH molecules, contoured at 3.0r (green), is overlain with a stick model of the bound INH.
formation. The bound INH in this site mainly interacted through van der Waals contacts, and hydrogen bonding was not observed. The Ser308 Oc side chain was located on the opposite side of the INH molecule; the Ca and Cb atoms of Ser308 might contribute to INH binding. The configuration of the bound INH implies that the IN radical formation reaction is not directly related to Ser308, because there is no candidate for the nucleophile residue to generate the IN radical. Also, the INH-binding site at the entrance to the c-edge side of heme (site 2), where the distance consisted of Trp77, between INH and heme is ~ 20 A, Trp78, Lys130, Ile294, Gly300, and Ile301 (Fig. 3C). At this site, the INH molecule was also mainly bound through van der Waals contacts. It seems likely that the bulky side chain of Trp78 hampered the entrance of INH into the heme cavity in the crystalline state, resulting in the trapping of the INH molecule at this site. Finally, the loop structures in front of the heme propionate side chain also constituted an INH-binding site (site 3), where the distance between INH and heme The side chains of Asn271, Arg303, Asn304, is ~ 18 A. Glu311, Pro340, and Arg367 interacted with the INH molecule (Fig. 3D). Of particular importance were the p–p stacking interactions between the guanidyl group of Arg367 and the pyridine ring of INH, which might contribute to the retention of INH at this site. This structural analysis clearly demonstrated the binding sites and the relevant interactions between protein and small molecule at each location. All of the FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
S. Kamachi et al.
interactions between KatG and INH appeared to be weak, as the interactions did not involve hydrogen bonds. Indeed, the crystallographic average temperature factors (B-factors) of the bound INH molecules were 78.6 A3 (site 1), 46.3 A3 (site 2), and 66.4 A3 (site 3), respectively; the occupancies of the bound INH probably differed between sites, and these values probably reflected the affinity of the INH molecule for each INH-binding site in the case of SeKatG. The B-factors of the different INH molecules were considerably higher than the overall B-factor (37.8 A3), but are consistent with the low-affinity Km values for INH. Structural comparison of the INH-binding sites among SeKatG, BpKatG, and MtKatG We identified three novel INH-binding sites on the molecular surface of SeKatG (Fig. 3). However, the INH molecules seem to bind only weakly to SeKatG, as there were no hydrogen-bonding or electrostatic interactions. So far, the crystal structure of INHbound BpKatG has shown that one INH molecule is bound per homodimeric BpKatG in the asymmetric unit. In BpKatG, the INH-binding site is located near distant from the entrance the dimer interface, ~ 20 A to the heme cavity (Fig. 4, left panel). In other words, the INH-binding sites of SeKatG and BpKatG are completely different. Furthermore, the configurations of residues involved in the binding site in BpKatG were also completely different from those in the corresponding region in SeKatG (Fig. 4, right panel), although in BpKatG the INH molecules also seem to bind through weak interactions without the hydrogen
Crystal structure of isoniazid-bound KatG
bond seen in the SeKatG. Superposition of the bound INH in BpKatG on SeKatG revealed that, in SeKatG, the INH molecule is not able to bind at the same position, owing to the steric hindrance between the INH molecule and the side chains of Asp183 or Asn475 in (Fig. 4, right panel). This raises SeKatG (> 1.6 A) questions regarding why each KatG has different INH-binding sites, and where the INH-binding site is in MtKatG. To answer these questions, we compared the structures of SeKatG, BpKatG, and MtKatG. Of the three INH-binding sites observed in SeKatG, we focus here on the site located at the entrance to the e-edge side of heme (site 1) (Fig. 5A), because the configurations and residue compositions at the other two INH-binding sites were very dissimilar across SeKatG, BpKatG, and MtKatG (Figs 5B,C). When INH-bound SeKatG was compared with BpKatG, the residues located in site 1 (see the blue triangles in Fig. 1) were identical, with the exception that Val193 in SeKatG was altered to the homologous leucine (Leu209) in BpKatG (Fig. 5A, left panel), and the shape and size of the binding site were highly conserved between the two proteins. However, for BpKatG, the corresponding site has not been identified as an INH-binding site. Instead, this site in INH-bound BpKatG was occupied by 2-methyl-2,4-pentanediol (MPD) (Fig. 5A, left panel), which might have been used as a cryoprotectant or additive reagent during the crystallization of BpKatG. Moreover, in the INH-free BpKatG structure, 2-methylpentane-2,4-diol (chain A) and MPD (chain B) were observed to be bound in this site. That is, the INH-binding site in SeKatG (site 1) was assigned as an organic solvent-binding site in BpKatG.
Fig. 4. A structural comparison of the INH-binding site in BpKatG and the corresponding region in SeKatG: superimposition of the homodimeric structures of INH-bound BpKatG (pink) and SeKatG (light purple). The bound INH molecules are shown by the space-filling model, and the hemes are shown in red. The INH-binding site in BpKatG is shown in the square box, and the close-up view corresponding to this region is shown in the right panel to provide details of the site.
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
59
Crystal structure of isoniazid-bound KatG
S. Kamachi et al.
A
B
C Fig. 5. A structural comparison of the INH-binding sites in SeKatG, BpKatG, and MtKatG. (A) Comparison of site 1 in SeKatG, site 1 in BpKatG, and site 1 in MtKatG. Left panel: superimposed structures of SeKatG (light purple) and BpKatG (pink). Right panel: superimposed structures of SeKatG (light purple) and MtKatG (yellow). INH bound to SeKatG and MPD bound to BpKatG are shown in blue and red, respectively. (B, C) Structural comparison of site 2 (B) and site 3 (C) in SeKatG.
INH binding is probably disrupted by organic solvent; thus, if these reagents had not been used in the BpKatG crystallization, it seems likely that INH could have bound to this site. In the corresponding site in MtKatG (Figs 1 and 5A), the configuration and residues were quite similar to those found in SeKatG and BpKatG. Together, these findings make this site a promising candidate for the common, but low-affinity, KatG INH-binding site. 60
In this study, we identified three INH-binding sites in SeKatG, one of which was located near the invariant serine. However, the Km values for IN–NAD formation were only subtly different between the WT and S308T SeKatG proteins. The bound INH configurations in WT SeKatG implied that there is no candidate for the nucleophile residue for the reaction. Therefore, the reaction leading to IN radical species formation might not occur in this position. Another possibility is that FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
S. Kamachi et al.
the INH-binding site that we observed here may be the first molecular trapping site leading to an association between the INH molecule and the protein. Subsequently, the INH molecule might enter the heme cavity (e.g. by the thermodynamic motion of the entrance of the funnel-shaped heme cavity) to form the IN radical species. This concept is consistent with the results of the thermal melting experiments, which showed that binding of the INH molecules probably increases KatG flexibility (Fig. 2). It should be noted that the formation of the IN–NAD adduct requires NAD+ and superoxide, but in this study we could not obtain evidence for adduct formation, probably because adduct formation occurs independently of KatG. Furthermore, EPR studies have identified the two electron pathways for the reaction in BpKatG [2,24], so one possibility is that KatG has several INH-binding sites for the reaction. Further mutational, biochemical and structural studies are needed to elucidate the reaction mechanisms of IN radical and IN–NAD adduct formation by KatG.
Experimental procedures Expression and purification of SeKatG The expression and purification of SeKatG were performed as described previously [25]. Briefly, an Escherichia coli BL21(DE3)pLysS strain transformed with plasmid pET3a– SeKatG was grown at 37 °C to an attenuance of 0.6 at 600 nm (D600 nm) in 3 L of liquid LB broth containing ampicillin (50 lgmL1). Then, SeKatG expression was induced by the addition of 1 mM IPTG. After induction, the cells were cultured at 37 °C for 18 h, collected by centrifugation (2560 g), and lysed. SeKatG protein was € purified on an AKTA purifier system (GE Healthcare, Buckinghamshire, UK) with a HiLoad 26/10 Q Sepharose column (GE Healthcare), followed by a HiLoad Phenyl Sepharose column, and a Mono Q HR 5/5 column (GE Healthcare). Purified SeKatG showed a UV–visible spectrum indistinguishable from that previously reported, and was concentrated with a Vivaspin filter (GE Healthcare).
Preparation of an S308T SeKatG mutant Site-directed mutagenesis was performed on SeKatG by use of the KOD-Plus-mutagenesis kit (Toyobo, Osaka, Japan), with plasmid pET3a–SeKatG as the template. The oligonucleotides 50 -GCGgcgtcttacgcaggtgatc-30 and 50 -gtcggcatat gacaaatgcat-30 were used to introduce the S308T mutation into SeKatG. Capital letters in the sequence indicate the codons corresponding to the replacement residues. Sequence analysis verified that the new constructs were free of errors. The mutant proteins were expressed and purified as described above for WT SeKatG.
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
Crystal structure of isoniazid-bound KatG
Expression and purification of MtKatG The MtKatG gene was constructed by use of a synthetic oligonucleotide (Eurofins Genomics, Brussels, Belgium), based on the published amino acid sequence optimized for codon usage in E. coli. The resulting MtKatG gene was cloned into a pET-15b vector (Novagen, Madison, WI, USA) and expressed in E. coli strain C41(DE3) [26] to produce N-terminal His-tagged MtKatG in high yield. The His-tagged MtKatG protein was purified with Ni2+– nitrilotriacetic acid affinity resin (Nacalai Tesque, Kyoto, Japan), according to the manufacturer’s protocol. After elution from the resin, the His-tagged MtKatG was fur€ ther purified by gel filtration with an AKTA explorer 10S system equipped with a HiPrep 16/60 Sephacryl S200 HR column (GE Healthcare). Purified MtKatG was concentrated and desalted with a Vivaspin filter (GE Healthcare).
Enzymatic IN–NAD formation by SeKatG IN–NAD formation activity was assayed at 37 °C in a reaction mixture containing 0.4 mM NAD+, 1.0 mM INH, 2 lM MnCl2, 50 mM Tris/HCl (pH 7.8), and 50 lgmL1 enzyme. IN–NAD formation was monitored at 326 nm (e = 6900 M1 cm1) with a V630BIO spectrometer (JASCO). The nonenzymatic reaction was assayed with the same reaction mixture without enzyme, and the result was used for the background value in calculations of enzymatic activity. Km values for INH were determined by using reaction mixtures containing various concentrations of either INH (0.5, 1.25, 2.5, 5, 10, 20 and 40 mM) or NAD+ (0.0025, 0.05, 0.1, 0.2, 0.4, 0.8 and 2 mM), and the enzyme kinetic parameters were estimated by nonlinear least-squares curve fitting in Microsoft Excel Solver.
Thermal shift assays Thermal shift assays were performed on a CFX96 RealTime PCR Cycler (Bio-Rad Laboratories, Hercules, CA, USA). In a typical experiment, 1 lL of SYPRO Orange [Sigma-Aldrich, St. Louis, MO, USA; diluted from 9 5000 stock into 50 mM Tris/HCl, pH 7.8, 4 lL of protein (0.4 mgmL1), and 5 lL of buffer (50 mM Tris/ HCl, pH 7.8)] was mixed on ice in a white 96-well PCR plate (Bio-Rad Laboratories). To evaluate the effect of INH binding, 1 lL of 9 5000 SYPRO Orange, 4 lL of protein (0.4 mgmL1), 1 lL of INH (10 mM) and 4 lL of Tris/HCl buffer were mixed. Fluorescence was measured from 25 °C to 85 °C in 0.5 °C steps (excitation, 450–490 nm; detection, 560–580 nm). All measurements were performed five times. Data evaluation and melting point determination were conducted with Bio-Rad CFX MANAGER software.
61
Crystal structure of isoniazid-bound KatG
Preparation of INH-bound SeKatG crystals Purified SeKatG was desalted by repeated concentration with Vivaspin filters (GE Healthcare), and then diluted with 50 mM Hepes buffer (pH 7.0). Crystallization conditions were optimized with the sitting-drop vapor-diffusion method based on previously reported conditions [25]. Drops were prepared by mixing 2 lL of protein (12 mgmL1) with 1 lL of reservoir solution (4.3 M sodium formate and 100 mM sodium citrate, pH 6.4). The crystals grew in 1 week to an average size of 0.5 9 0.25 9 0.25 mm. INH-bound crystals were obtained by soaking the crystals for 24 h in crystallization solution containing 100 mM INH.
X-ray crystallography The INH-bound crystals were flash-cooled with a nitrogen gas stream at –180 °C. Notably, the crystals did not require any cryoprotection, because the crystallization conditions included high salt concentrations. Diffraction data were collected by the use of synchrotron radiation and the MX300HE detector (Rayonix, Evanston, IL, USA) of beamline BL44XU at Spring-8 (Harima, Japan). The crystal was oscillated by 0.5° between image collections, and a total of 400 images were processed. Integrated intensities were merged and scaled with the HKL2000 suite [27]. The results of the data collection are summarized in Table 2.
S. Kamachi et al.
Table 2. Crystallographic data and refinement statisticsa. PDB, Protein Data Bank. Crystallographic data Space group Cell parameters ( A) Resolution range ( A) Observed reflections Unique reflections Mean I/r (I) Redundancy Completeness (%) Rsym (%)b Refinement statistics Rcryst (%)c Rfree (%)d Number of molecules INH Heme Na+ Water Rmsd from ideal values Bond length ( A) Bond angle (°) Average B-factor ( A3 ) Ramachandran plot Most favored (%) Additionally allowed (%) Generously allowed (%) Disallowed (%) PDB entry
P41212 a = b = 108.7, c = 202.9 50.0–2.12 (2.20–2.12) 504 680 126 347 9.9 (3.1) 4.0 (3.9) 95.2 (97.3) 8.0 (51.5) 15.8 19.7 3 1 4 487 0.020 1.95 37.8 89.6 10.1 0.3 0.0 3WXO
a
Structure determination and refinement of INH-bound SeKatG Crystals of INH-bound SeKatG were nearly isomorphous with those of native KatG, and the structure of INHbound SeKatG was determined with the structure of native KatG (Protein Data Bank ID: 3WNU) as a model [9]. The structure was subjected to rigid-body refinement to 3.0-A resolution with REFMAC5 [28] from the from 25-A CCP4 package [29]. The structure was further refined at resolution by restrained refinement in REFMAC5, 2.12-A and manual model revision was carried out with COOT [30]. Ordered water molecules were added to the model by the use of ARP/WARP [31]. The electron density map at the final stage was clear enough for exact assignment of the orientations of the bound INH molecules. The INH molecules were unambiguously fitted into the Fo – Fc map on the protein surface. Structure refinement statistics are summarized in Table 2. Figures 2 and 3 were prepared with PYMOL [32].
Acknowledgements We thank Drs S. Baba and E. Yamashita for their assistance with data collection at the SPring-8 synchrotron radiation facility (Hyogo, Japan). The 62
Values in parentheses are for the outermost shell. Rsym = ΣhklΣi|Ii(hkl) |/ΣhklΣiIi(hkl), where is the average intensity over equivalent reflections. c Rcryst = Σ||Fobs(hkl)| |Fcalc(hkl)||/Σ|Fobs(hkl)|. d Rfree is the R-value calculated for 5% of the dataset not included in the refinement. b
synchrotron radiation experiments were performed on BL38B1 and BL44XU at SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute (Proposal Nos. 2011B2066, 2012A6726, 2012B6725, and 2012B6726). This work was supported by the Program to Disseminate Tenure Tracking (to K. Wada) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), by MEXT Grants-in-Aid of Scientific Research 26440035 (to T. Tada) and 5840023 (to K. Wada), by a grant from the Japan Foundation for Applied Enzymology (to K. Wada), and by a Grantin-Aid for JSPS Fellows 24 1292 (to K. Hirabayashi). We also thank Professor T. Inoue and Dr H. Matsumura (Osaka University) and Dr K. Hirata (RIKEN) for invaluable suggestions, and Y. Motoyama and N. Kaseda (University of Miyazaki) for technical assistance. FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
S. Kamachi et al.
Author contributions Planned experiments: TT and KW; Performed experiments: SK, KH, TT and KW; Analyzed data: SK, KH, TT and KW; Contributed reagents or other essential material: MT, and SS; Wrote the paper: SK, KH, MT, SS, TT and KW.
References 1 Bloom BR & Murray CJ (1992) Tuberculosis: commentary on a reemergent killer. Science 257, 1055– 1064. 2 Wiseman B, Carpena X, Feliz M, Donald LJ, Pons M, Fita I & Loewen PC (2010) Isonicotinic acid hydrazide conversion to Isonicotinyl-NAD by catalaseperoxidases. J Biol Chem 285, 26662–26673. 3 Rozwarski DA, Grant GA, Barton DH, Jacobs WR Jr & Sacchettini JC (1998) Modification of the NADH of the isoniazid target (InhA) from Mycobacterium tuberculosis. Science 279, 98–102. 4 Rawat R, Whitty A & Tonge PJ (2003) The isoniazidNAD adduct is a slow, tight-binding inhibitor of InhA, the Mycobacterium tuberculosis enoyl reductase: adduct affinity and drug resistance. Proc Natl Acad Sci U S A 100, 13881–13886. 5 Wang F, Jain P, Gulten G, Liu Z, Feng Y, Ganesula K, Motiwala AS, Ioerger TR, Alland D, Vilcheze C et al. (2010) Mycobacterium tuberculosis dihydrofolate reductase is not a target relevant to the antitubercular activity of isoniazid. Antimicrob Agents Chemother 54, 3776–3782. 6 Yamada Y, Fujiwara T, Sato T, Igarashi N & Tanaka crystal structure of catalaseN (2002) The 2.0 A peroxidase from Haloarcula marismortui. Nat Struct Biol 9, 691–695. 7 Carpena X, Loprasert S, Mongkolsuk S, Switala J, Loewen PC & Fita I (2003) Catalase-peroxidase KatG resolution. J Mol of Burkholderia pseudomallei at 1.7 A Biol 327, 475–489. 8 Bertrand T, Eady NA, Jones JN, Jesmin Nagy JM, Jamart-Gregoire B, Raven EL & Brown KA (2004) Crystal structure of Mycobacterium tuberculosis catalase-peroxidase. J Biol Chem 279, 38991– 38999. 9 Kamachi S, Wada K, Tamoi M, Shigeoka S & Tada T resolution structure of the catalase(2014) The 2.2 A peroxidase KatG from Synechococcus elongatus PCC7942. Acta Crystallogr F Struct Biol Commun 70, 288–293. 10 Pierattelli R, Banci L, Eady NA, Bodiguel J, Jones JN, Moody PC, Raven EL, Jamart-Gregoire B & Brown KA (2004) Enzyme-catalyzed mechanism of isoniazid activation in class I and class III peroxidases. J Biol Chem 279, 39000–39009.
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
Crystal structure of isoniazid-bound KatG
11 Metcalfe C, Macdonald IK, Murphy EJ, Brown KA, Raven EL & Moody PC (2008) The tuberculosis prodrug isoniazid bound to activating peroxidases. J Biol Chem 283, 6193–6200. 12 Vidossich P, Loewen PC, Carpena X, Fiorin G, Fita I & Rovira C (2014) Binding of the antitubercular prodrug isoniazid in the heme access channel of catalaseperoxidase (KatG). A combined structural and metadynamics investigation. J Phys Chem B 118, 2924– 2931. 13 Kochi A, Vareldzis B & Styblo K (1993) Multidrugresistant tuberculosis and its control. Res Microbiol 144, 104–110. 14 Heym B, Alzari PM, Honore N & Cole ST (1995) Missense mutations in the catalase-peroxidase gene, katG, are associated with isoniazid resistance in Mycobacterium tuberculosis. Mol Microbiol 15, 235–245. 15 Huyen MN, Cobelens FG, Buu TN, Lan NT, Dung NH, Kremer K, Tiemersma EW & van Soolingen D (2013) Epidemiology of isoniazid resistance mutations and their effect on tuberculosis treatment outcomes. Antimicrob Agents Chemother 57, 3620–3627. 16 Wengenack NL, Uhl JR, St Amand AL, Tomlinson AJ, Benson LM, Naylor S, Kline BC, Cockerill FR 3rd & Rusnak F (1997) Recombinant Mycobacterium tuberculosis KatG(S315T) is a competent catalaseperoxidase with reduced activity toward isoniazid. J Infect Dis 176, 722–727. 17 Wengenack NL, Todorovic S, Yu L & Rusnak F (1998) Evidence for differential binding of isoniazid by Mycobacterium tuberculosis KatG and the isoniazidresistant mutant KatG(S315T). Biochemistry 37, 15825– 15834. 18 Rickman KA, Swancutt KL, Mezyk SP & Kiddle JJ (2013) Isoniazid: radical-induced oxidation and reduction chemistry. Bioorg Med Chem Lett 23, 3096– 3100. 19 Lukat-Rodgers GS, Wengenack NL, Rusnak F & Rodgers KR (2000) Spectroscopic comparison of the heme active sites in WT KatG and its S315T mutant. Biochemistry 39, 9984–9993. 20 Kapetanaki SM, Chouchane S, Yu S, Magliozzo RS & Schelvis JP (2005) Resonance Raman spectroscopy of Compound II and its decay in Mycobacterium tuberculosis catalase-peroxidase KatG and its isoniazid resistant mutant S315T. J Inorg Biochem 99, 1401–1406. 21 Kapetanaki SM, Chouchane S, Yu S, Zhao X, Magliozzo RS & Schelvis JP (2005) Mycobacterium tuberculosis KatG(S315T) catalase-peroxidase retains all active site properties for proper catalytic function. Biochemistry 44, 243–252. 22 Deemagarn T, Carpena X, Singh R, Wiseman B, Fita I & Loewen PC (2005) Structural characterization of the Ser324Thr variant of the catalase-peroxidase (KatG) from Burkholderia pseudomallei. J Mol Biol 345, 21–28.
63
Crystal structure of isoniazid-bound KatG
23 Zhao X, Yu H, Yu S, Wang F, Sacchettini JC & Magliozzo RS (2006) Hydrogen peroxide-mediated isoniazid activation catalyzed by Mycobacterium tuberculosis catalase-peroxidase (KatG) and its S315T mutant. Biochemistry 45, 4131–4140. 24 Colin J, Wiseman B, Switala J, Loewen PC & Ivancich A (2009) Distinct role of specific tryptophans in facilitating electron transfer or as [Fe(IV)=O Trp(*)] intermediates in the peroxidase reaction of Bulkholderia pseudomallei catalase-peroxidase: a multifrequency EPR spectroscopy investigation. J Am Chem Soc 131, 8557– 8563. 25 Wada K, Tada T, Nakamura Y, Kinoshita T, Tamoi M, Shigeoka S & Nishimura K (2002) Crystallization and preliminary X-ray diffraction studies of catalase-peroxidase from Synechococcus PCC 7942. Acta Crystallogr D Biol Crystallogr 58, 157–159. 26 Miroux B & Walker JE (1996) Over-production of proteins in Escherichia coli: mutant hosts that allow synthesis of some membrane proteins and globular proteins at high levels. J Mol Biol 260, 289–298. 27 Otwinowsky Z & Minor W (1997) Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol 276, 307–326.
64
S. Kamachi et al.
28 Murshudov GN, Vagin AA & Dodson EJ (1997) Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr D Biol Crystallogr D53, 240–255. 29 Collaborative Computational Project N (1994) The CCP4 suite: programs for protein crystallography. Acta Crystallogr D Biol Crystallogr D50, 760–763. 30 Emsley P & Cowtan K (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr D60, 2126–2132. 31 Cohen SX, Ben Jelloul M, Long F, Vagin A, Knipscheer P, Lebbink J, Sixma TK, Lamzin VS, Murshudov GN & Perrakis A (2008) ARP/wARP and molecular replacement: the next generation. Acta Crystallogr D Biol Crystallogr D64, 49–60. 32 DeLano WL (2002) PyMOL. DeLano Scientific, San Carlos, CA. 33 Thompson JD, Higgins DG & Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22, 4673–4680. 34 Robert X & Gouet P (2014) Deciphering key features in protein structures with the new ENDscript server. Nucleic Acids Res 42, W320–W324.
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
Graduate School of Science, Osaka Prefecture University, Sakai, Osaka, Japan Organization for Promotion of Tenure Track, University of Miyazaki, Japan Graduate School of Science, Osaka University, Toyonaka, Osaka, Japan Faculty of Agriculture, Kinki University, Nakamachi, Nara, Japan
Keywords antituberculous drug; catalase-peroxidase; crystal structure; isoniazid; KatG Correspondence T. Tada, Graduate School of Science, Osaka Prefecture University, Sakai, Osaka 599– 8531, Japan Fax: +81 72 254 9935 Tel:+81 72 254 9820 E-mail: [email protected] and K. Wada, Organization for Promotion of Tenure Track, University of Miyazaki, Miyazaki 889–1692, Japan Fax: +81 985 85 0873 Tel:+81 985 85 0873 E-mail: [email protected] (Received 5 September 2014, revised 26 September 2014, accepted 6 October 2014) doi:10.1111/febs.13102
Isoniazid (INH) is one of the most effective antibiotics against tuberculosis. INH is a prodrug that is activated by KatG. Although extensive studies have been performed in order to understand the mechanism of KatG, even the binding site of INH in KatG remains controversial. In this study, we determined the crystal structure of KatG from Synechococcus elongatus resolution. Three PCC7942 (SeKatG) in a complex with INH at 2.12-A INH molecules were bound to the molecular surface. One INH molecule was bound at the entrance to the e-edge side of heme (designated site 1), another was bound at the entrance to the c-edge side of heme (site 2), and another was bound to the loop structures in front of the heme propionate side chain (site 3). All of the interactions between KatG and the bound INH seemed to be weak, being mediated mainly by van der Waals contacts. Structural comparisons revealed that the identity and configuration of the residues in site 1 were very similar among SeKatG, Burkholderia pseudomallei KatG, and Mycobacterium tuberculosis KatG. In contrast, sites 2 and 3 were structurally diverse among the three proteins. Thus, site 1 is probably the common KatG INH-binding site. A static enzymatic analysis and thermal shift assay suggested that the INH-activating reaction does not proceed in site 1, but rather that this site may function as an initial trapping site for the INH molecule. Database The atomic coordinates and structure factors have been deposited in the Protein Data Bank under the accession number 3WXO.
Introduction Tuberculosis remains an important life-threatening disease, and is responsible for 1.5–2 million annual deaths worldwide [1]. Isoniazid (INH) is a widely used antituberculous prodrug that is activated by Mycobacterium tuberculosis itself via the catalase-peroxidase KatG. During INH activation, the INH molecule is cleaved by KatG to generate an isonicotinyl radical
(IN) and diazene, and subsequently the IN radical species is combined with NAD+ to generate an isonicotinyl-NAD (IN–NAD) covalent adduct in the presence of superoxide, which is formed by O2 reduction by KatG [2]. The IN–NAD adduct inhibits the longchain enoyl acyl carrier protein reductase (InhA) involved in the synthesis of mycolic acid [3], which is a
Abbreviations BpKatG, Burkholderia pseudomallei KatG; INH, isoniazid; IN, isonicotinyl radical; MPD, 2-methyl-2,4-pentanediol; MtKatG, Mycobacterium tuberculosis KatG; SeKatG, Synechococcus elongatus PCC7942; WT, wild-type.
54
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
S. Kamachi et al.
major component of the mycobacterial cell wall. Although the mechanism of inhibition of InhA by the IN–NAD adduct has been clearly demonstrated [4,5], the mechanism by which KatG forms IN species, including the exact binding sites for INH, remains unclear. Extensive structural studies of KatGs from several bacteria have been performed; the crystal structures of KatG from Haloarcula marismortui [6], Burkholderia pseudomallei (BpKatG) [7], M. tuberculosis (MtKatG) [8] and Synechococcus elongatus PCC7942 (SeKatG) [9] have been determined. However, an INH-bound structure has only been reported for BpKatG [2]. In that study, the INH-binding site was located on the opposite side of the BpKatG protein from the entrance to the heme cavity. Interestingly, NMR analyses of typical peroxidases, which had a very similar heme cavity to that of KatG, showed that the INH molecule could bind to the heme cavity [10]. Furthermore, in ascorbate peroxidase bearing the point mutation H42A, INH directly interacted with the heme iron [11]. More recently, the structure of BpKatG bearing the point mutation D141A demonstrated that the INH molecule could also enter into/bind to the heme cavity after removal of the ‘lid residue’ (Asp141) at the entrance to the funnel-shaped heme cavity, and a metadynamics analysis revealed that the mutation of Asp141 removed an energy barrier preventing entry INH into the heme cavity [12]. In addition, spectroscopic analyses with wild-type (WT) BpKatG also strongly implied that the INH molecule binds around the heme cavity [2]. More than half of the INH-resistant clinical isolates of M. tuberculosis contain a mutation in MtKatG, whereby the strictly conserved serine at position 315 (Fig. 1) is replaced with threonine (S315T MtKatG) [13–15]. Biochemical analyses have shown that S315T MtKatG retains only ~ 30–50% of the WT activity for IN–NAD formation [2]; this reduced rate of activity is sufficient to confer INH resistance in M. tuberculosis infections. To date, using S315T MtKatG and the corresponding BpKatG mutant (S324T), biochemical [2,16–18], spectroscopic [19–21] and structural [22,23] investigations have been carried out to find the factor (s) underlying the reduction in IN–NAD formation activity. Unfortunately, these studies yielded no apparent differences between the WT and mutant KatGs. One likely reason for this is that there has been no consensus concerning the INH-binding sites on KatG, probably owing to their low affinities, because the primary function of KatG in vivo is that of scavenging hydrogen peroxide. FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
Crystal structure of isoniazid-bound KatG
All KatG orthologs so far used in structural studies, which share at least 53% sequence identity (Fig. 1), have very similar 3D/4D structures. Here, we investigated INH-binding/IN–NAD formation characteristics by using SeKatG and the crystal structure of SeKatG resolution. in complex with INH molecules at 2.12-A In SeKatG, three INH-binding sites were characterized on the molecular surface, all of which were completely different sites from those observed in the INH-bound structure of BpKatG. However, one site was located at the entrance to the e-edge side of heme, where Ser308 [the residue corresponding to Ser315 in MtKatG (Fig. 1)] is located. A structural comparison between INH-bound SeKatG and BpKatG indicated that the entrance to the e-edge side of heme may be the common INH-binding site; however, the IN formation reaction seems not to proceed at this site. Therefore, the common INH-binding site found in this study may function as an initial trapping/recruitment site for INH molecules.
Results and Discussion Enzymatic IN–NAD formation by SeKatG and its mutant protein It has been reported that not only MtKatG but also BpKatG can enhance the conversion of INH and NAD+ to IN–NAD in the presence of Mn2+ and O2. To determine whether SeKatG also has the ability to induce IN–NAD formation, we performed an enzymatic assay (Table 1). WT SeKatG showed a specific activity of 33.3 nmolmin1mg1 protein for IN–NAD formation. In order to compare SeKatG and MtKatG, we also measured MtKatG activity under the same conditions; MtKatG had an activity of 35.1 nmolmin1mg1 protein. Thus, SeKatG possessed substantial activity for the conversion of INH to IN–NAD, and this activity was equivalent to that of MtKatG. When the S308T mutation, which is equivalent to the S315T mutation in MtKatG, was introduced into SeKatG, the IN–NAD formation Vmax was decreased by 42% as compared with the WT. Nevertheless, the Km values for INH and NAD+ were only subtly different between the WT protein (5.6 mM and 17.6 lM, respectively) and S308T protein (10.8 mM and 23.3 lM, respectively). The enzymatic nature of SeKatG was very similar to those of MtKatG and BpKatG; previous biological studies have shown that in MtKatG the S315T mutation results in an ~ 51% reduction and that in BpKatG the S324T mutation results in an ~ 41% reduction in in vitro IN–NAD formation, 55
Crystal structure of isoniazid-bound KatG
S. Kamachi et al.
Fig. 1. Multiple sequence alignment of KatGs from several representative organisms. The sequence numbering of SeKatG is shown. Identical residues are highlighted in light brown, and similar residues are boxed in blue. The residues of the catalytic Met-Tyr-Trp adduct are highlighted in black. The residues involved in INH binding in SeKatG are indicated by blue (site 1), green (site 2) and yellow (site 3) triangles, respectively. The figure was prepared with CLUSTALW [33] and ESPRIPT [34]. KatG orthologs are shown from S. elongatus PCC7942 (NCBI accession no. WP_011244741), M. tuberculosis (WP_003899075), B. pseudomallei (WP_011205232), H. marismortui (WP_011223478), and Synechocystis PCC6803 (WP_010872602).
56
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
S. Kamachi et al.
Crystal structure of isoniazid-bound KatG
Table 1. Rates of IN–NAD formation by WT and S308T SeKatG. The specific activities were evaluated at 37 °C in a reaction mixture containing 0.4 mM NAD+, 1.0 mM INH, 2 lM MnCl2, and 50 mM Tris/HCl (pH 7.8). SD, standard deviation.
WT SeKatG S308T SeKatG MtKatG
Specific activity (nmolmin1mg1 protein SD)
Vmax (nmolmin1mg1 protein)
INH (mM)
NAD+ (lM)
33.3 6.3 7.3 2.4 35.1 3.6
278.9 115.7 –
5.6 10.8 –
17.6 23.3 –
Km
without strongly affecting the Km values for INH and NAD+ [2]. Insights into why IN–NAD formation is reduced upon introduction of the S308T mutation came from a thermal shift assay (Fig. 2). S308T SeKatG showed a 4 °C decrease in its thermal denaturation temperature (Tm) relative to that of the WT, indicating that the A
B
S308T mutation affected protein stability. We postulate that the mutation at the Ser308 position, which is in the vicinity of the heme propionate side chain, may alter the stability of the heme cavity and/or the flexibility of the heme pocket entrance, thus leading to a decrease in IN–NAD formation activity. In other words, the invariant serine (Ser308 in SeKatG) is probably not involved directly in INH binding, but may contribute to recruiting the INH molecule. Indeed, it has been reported that INH may bind to the heme cavity of the BpKatG D141A mutant protein [12] and a typical peroxidase [11]. Thus, if IN–NAD formation occurs in the heme cavity, flexibility of the heme entrance is probably necessary to achieve incorporation of the INH molecule into the cavity. Conducting the thermal shift assay in the presence of INH revealed an additional effect of INH binding on stability, in that the Tm values of both the WT and Ser308 SeKatG proteins were decreased by 2.5 °C relative to their unbound Tm values. This result indicated that the INH molecule undoubtedly binds to SeKatG, and that its effect is large enough to change the thermostability. It should be noted that this method could not determine whether the bound INH was involved in IN–NAD formation; also, the number of occupied INH-binding site(s) was not determined. When the INH molecule(s) are bound to SeKatG, the protein thermostability deteriorates, raising the possibility that the INH molecule may trigger an increase in the flexibility of KatG. Overall structure of SeKatG in complex with INH molecules
Fig. 2. A thermal shift assay revealed that the S308T mutation and INH binding both affect the melting temperature of SeKatG. (A) Inverse first-derivative plots of WT (blue) and S308T (red) SeKatG. (B) Inverse first-derivative plots of WT (blue) and S308T (red) SeKatG in the presence of 1 mM INH. n = 5 for each assay; the Tm values and standard errors are shown below the curves.
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
The structure of INH-bound SeKatG was refined at resolution to R and Rfree values of 0.158 and 2.12-A 0.197, respectively. The asymmetric unit contained one subunit of an SeKatG molecule, including a protoporphyrin IX heme moiety, four cations (modeled as sodium ions), and 487 water molecules. The crystallographic equivalent subunit is created by a two-fold symmetry operation to form the functional dimer, as shown in Fig. 3A. Although the electron density for
57
Crystal structure of isoniazid-bound KatG
S. Kamachi et al.
A
B
C
D
SeKatG was mostly continuous and of high quality, the densities for the N-terminal segment (residues 1– 10) were poorly defined; accordingly, these residues were not included in the model. When the structure of INH-bound SeKatG was superposed on that of substrate-free SeKatG [9], no significant structural changes were observed in the enzyme; the rmsd for the Ca atoms was 0.15 A. Binding environments of the INH molecules The electron density map of INH-bound SeKatG revealed that three INH molecules were bound to the molecular surface (Fig. 3). One INH molecule was bound to the entrance of the e-edge side of heme, where Ser308 was located (designated site 1), another was bound to the entrance of the c-edge side of heme (site 2), and another was bound to the loop structures in front of the heme propionate side chain (site 3). All of the INH-binding sites were located in close proximity to heme, but there was no direct interaction with heme, as INH molecules did not enter the heme cavity. The INH-binding site is located at the entrance to the e-edge side of heme (site 1), where the nearest neighbor distances between INH and heme is ~ 5.6 A. This site consisted of Pro123, Asp124, Val193, Leu220, Asn224, Pro225, and Ser308 (Fig. 3B); this represents the first structural evidence that an INH molecule can interact with the invariant serines during IN–NAD 58
Fig. 3. SeKatG has three low-affinity INH-binding sites. (A) Overall structure of INH-bound SeKatG, and the locations of the INH-binding sites. The SeKatG dimeric structure was generated via a symmetry operation using the crystallographic two-fold axis, as the asymmetric unit contains one monomeric SeKatG. The individual subunits are shown in green or purple. The bound INH molecules are shown in blue, and heme is shown in red. (B–D) The electron density map showing details of the SeKatG INH-binding sites. An Fo – Fc map omitting the INH molecules, contoured at 3.0r (green), is overlain with a stick model of the bound INH.
formation. The bound INH in this site mainly interacted through van der Waals contacts, and hydrogen bonding was not observed. The Ser308 Oc side chain was located on the opposite side of the INH molecule; the Ca and Cb atoms of Ser308 might contribute to INH binding. The configuration of the bound INH implies that the IN radical formation reaction is not directly related to Ser308, because there is no candidate for the nucleophile residue to generate the IN radical. Also, the INH-binding site at the entrance to the c-edge side of heme (site 2), where the distance consisted of Trp77, between INH and heme is ~ 20 A, Trp78, Lys130, Ile294, Gly300, and Ile301 (Fig. 3C). At this site, the INH molecule was also mainly bound through van der Waals contacts. It seems likely that the bulky side chain of Trp78 hampered the entrance of INH into the heme cavity in the crystalline state, resulting in the trapping of the INH molecule at this site. Finally, the loop structures in front of the heme propionate side chain also constituted an INH-binding site (site 3), where the distance between INH and heme The side chains of Asn271, Arg303, Asn304, is ~ 18 A. Glu311, Pro340, and Arg367 interacted with the INH molecule (Fig. 3D). Of particular importance were the p–p stacking interactions between the guanidyl group of Arg367 and the pyridine ring of INH, which might contribute to the retention of INH at this site. This structural analysis clearly demonstrated the binding sites and the relevant interactions between protein and small molecule at each location. All of the FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
S. Kamachi et al.
interactions between KatG and INH appeared to be weak, as the interactions did not involve hydrogen bonds. Indeed, the crystallographic average temperature factors (B-factors) of the bound INH molecules were 78.6 A3 (site 1), 46.3 A3 (site 2), and 66.4 A3 (site 3), respectively; the occupancies of the bound INH probably differed between sites, and these values probably reflected the affinity of the INH molecule for each INH-binding site in the case of SeKatG. The B-factors of the different INH molecules were considerably higher than the overall B-factor (37.8 A3), but are consistent with the low-affinity Km values for INH. Structural comparison of the INH-binding sites among SeKatG, BpKatG, and MtKatG We identified three novel INH-binding sites on the molecular surface of SeKatG (Fig. 3). However, the INH molecules seem to bind only weakly to SeKatG, as there were no hydrogen-bonding or electrostatic interactions. So far, the crystal structure of INHbound BpKatG has shown that one INH molecule is bound per homodimeric BpKatG in the asymmetric unit. In BpKatG, the INH-binding site is located near distant from the entrance the dimer interface, ~ 20 A to the heme cavity (Fig. 4, left panel). In other words, the INH-binding sites of SeKatG and BpKatG are completely different. Furthermore, the configurations of residues involved in the binding site in BpKatG were also completely different from those in the corresponding region in SeKatG (Fig. 4, right panel), although in BpKatG the INH molecules also seem to bind through weak interactions without the hydrogen
Crystal structure of isoniazid-bound KatG
bond seen in the SeKatG. Superposition of the bound INH in BpKatG on SeKatG revealed that, in SeKatG, the INH molecule is not able to bind at the same position, owing to the steric hindrance between the INH molecule and the side chains of Asp183 or Asn475 in (Fig. 4, right panel). This raises SeKatG (> 1.6 A) questions regarding why each KatG has different INH-binding sites, and where the INH-binding site is in MtKatG. To answer these questions, we compared the structures of SeKatG, BpKatG, and MtKatG. Of the three INH-binding sites observed in SeKatG, we focus here on the site located at the entrance to the e-edge side of heme (site 1) (Fig. 5A), because the configurations and residue compositions at the other two INH-binding sites were very dissimilar across SeKatG, BpKatG, and MtKatG (Figs 5B,C). When INH-bound SeKatG was compared with BpKatG, the residues located in site 1 (see the blue triangles in Fig. 1) were identical, with the exception that Val193 in SeKatG was altered to the homologous leucine (Leu209) in BpKatG (Fig. 5A, left panel), and the shape and size of the binding site were highly conserved between the two proteins. However, for BpKatG, the corresponding site has not been identified as an INH-binding site. Instead, this site in INH-bound BpKatG was occupied by 2-methyl-2,4-pentanediol (MPD) (Fig. 5A, left panel), which might have been used as a cryoprotectant or additive reagent during the crystallization of BpKatG. Moreover, in the INH-free BpKatG structure, 2-methylpentane-2,4-diol (chain A) and MPD (chain B) were observed to be bound in this site. That is, the INH-binding site in SeKatG (site 1) was assigned as an organic solvent-binding site in BpKatG.
Fig. 4. A structural comparison of the INH-binding site in BpKatG and the corresponding region in SeKatG: superimposition of the homodimeric structures of INH-bound BpKatG (pink) and SeKatG (light purple). The bound INH molecules are shown by the space-filling model, and the hemes are shown in red. The INH-binding site in BpKatG is shown in the square box, and the close-up view corresponding to this region is shown in the right panel to provide details of the site.
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
59
Crystal structure of isoniazid-bound KatG
S. Kamachi et al.
A
B
C Fig. 5. A structural comparison of the INH-binding sites in SeKatG, BpKatG, and MtKatG. (A) Comparison of site 1 in SeKatG, site 1 in BpKatG, and site 1 in MtKatG. Left panel: superimposed structures of SeKatG (light purple) and BpKatG (pink). Right panel: superimposed structures of SeKatG (light purple) and MtKatG (yellow). INH bound to SeKatG and MPD bound to BpKatG are shown in blue and red, respectively. (B, C) Structural comparison of site 2 (B) and site 3 (C) in SeKatG.
INH binding is probably disrupted by organic solvent; thus, if these reagents had not been used in the BpKatG crystallization, it seems likely that INH could have bound to this site. In the corresponding site in MtKatG (Figs 1 and 5A), the configuration and residues were quite similar to those found in SeKatG and BpKatG. Together, these findings make this site a promising candidate for the common, but low-affinity, KatG INH-binding site. 60
In this study, we identified three INH-binding sites in SeKatG, one of which was located near the invariant serine. However, the Km values for IN–NAD formation were only subtly different between the WT and S308T SeKatG proteins. The bound INH configurations in WT SeKatG implied that there is no candidate for the nucleophile residue for the reaction. Therefore, the reaction leading to IN radical species formation might not occur in this position. Another possibility is that FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
S. Kamachi et al.
the INH-binding site that we observed here may be the first molecular trapping site leading to an association between the INH molecule and the protein. Subsequently, the INH molecule might enter the heme cavity (e.g. by the thermodynamic motion of the entrance of the funnel-shaped heme cavity) to form the IN radical species. This concept is consistent with the results of the thermal melting experiments, which showed that binding of the INH molecules probably increases KatG flexibility (Fig. 2). It should be noted that the formation of the IN–NAD adduct requires NAD+ and superoxide, but in this study we could not obtain evidence for adduct formation, probably because adduct formation occurs independently of KatG. Furthermore, EPR studies have identified the two electron pathways for the reaction in BpKatG [2,24], so one possibility is that KatG has several INH-binding sites for the reaction. Further mutational, biochemical and structural studies are needed to elucidate the reaction mechanisms of IN radical and IN–NAD adduct formation by KatG.
Experimental procedures Expression and purification of SeKatG The expression and purification of SeKatG were performed as described previously [25]. Briefly, an Escherichia coli BL21(DE3)pLysS strain transformed with plasmid pET3a– SeKatG was grown at 37 °C to an attenuance of 0.6 at 600 nm (D600 nm) in 3 L of liquid LB broth containing ampicillin (50 lgmL1). Then, SeKatG expression was induced by the addition of 1 mM IPTG. After induction, the cells were cultured at 37 °C for 18 h, collected by centrifugation (2560 g), and lysed. SeKatG protein was € purified on an AKTA purifier system (GE Healthcare, Buckinghamshire, UK) with a HiLoad 26/10 Q Sepharose column (GE Healthcare), followed by a HiLoad Phenyl Sepharose column, and a Mono Q HR 5/5 column (GE Healthcare). Purified SeKatG showed a UV–visible spectrum indistinguishable from that previously reported, and was concentrated with a Vivaspin filter (GE Healthcare).
Preparation of an S308T SeKatG mutant Site-directed mutagenesis was performed on SeKatG by use of the KOD-Plus-mutagenesis kit (Toyobo, Osaka, Japan), with plasmid pET3a–SeKatG as the template. The oligonucleotides 50 -GCGgcgtcttacgcaggtgatc-30 and 50 -gtcggcatat gacaaatgcat-30 were used to introduce the S308T mutation into SeKatG. Capital letters in the sequence indicate the codons corresponding to the replacement residues. Sequence analysis verified that the new constructs were free of errors. The mutant proteins were expressed and purified as described above for WT SeKatG.
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
Crystal structure of isoniazid-bound KatG
Expression and purification of MtKatG The MtKatG gene was constructed by use of a synthetic oligonucleotide (Eurofins Genomics, Brussels, Belgium), based on the published amino acid sequence optimized for codon usage in E. coli. The resulting MtKatG gene was cloned into a pET-15b vector (Novagen, Madison, WI, USA) and expressed in E. coli strain C41(DE3) [26] to produce N-terminal His-tagged MtKatG in high yield. The His-tagged MtKatG protein was purified with Ni2+– nitrilotriacetic acid affinity resin (Nacalai Tesque, Kyoto, Japan), according to the manufacturer’s protocol. After elution from the resin, the His-tagged MtKatG was fur€ ther purified by gel filtration with an AKTA explorer 10S system equipped with a HiPrep 16/60 Sephacryl S200 HR column (GE Healthcare). Purified MtKatG was concentrated and desalted with a Vivaspin filter (GE Healthcare).
Enzymatic IN–NAD formation by SeKatG IN–NAD formation activity was assayed at 37 °C in a reaction mixture containing 0.4 mM NAD+, 1.0 mM INH, 2 lM MnCl2, 50 mM Tris/HCl (pH 7.8), and 50 lgmL1 enzyme. IN–NAD formation was monitored at 326 nm (e = 6900 M1 cm1) with a V630BIO spectrometer (JASCO). The nonenzymatic reaction was assayed with the same reaction mixture without enzyme, and the result was used for the background value in calculations of enzymatic activity. Km values for INH were determined by using reaction mixtures containing various concentrations of either INH (0.5, 1.25, 2.5, 5, 10, 20 and 40 mM) or NAD+ (0.0025, 0.05, 0.1, 0.2, 0.4, 0.8 and 2 mM), and the enzyme kinetic parameters were estimated by nonlinear least-squares curve fitting in Microsoft Excel Solver.
Thermal shift assays Thermal shift assays were performed on a CFX96 RealTime PCR Cycler (Bio-Rad Laboratories, Hercules, CA, USA). In a typical experiment, 1 lL of SYPRO Orange [Sigma-Aldrich, St. Louis, MO, USA; diluted from 9 5000 stock into 50 mM Tris/HCl, pH 7.8, 4 lL of protein (0.4 mgmL1), and 5 lL of buffer (50 mM Tris/ HCl, pH 7.8)] was mixed on ice in a white 96-well PCR plate (Bio-Rad Laboratories). To evaluate the effect of INH binding, 1 lL of 9 5000 SYPRO Orange, 4 lL of protein (0.4 mgmL1), 1 lL of INH (10 mM) and 4 lL of Tris/HCl buffer were mixed. Fluorescence was measured from 25 °C to 85 °C in 0.5 °C steps (excitation, 450–490 nm; detection, 560–580 nm). All measurements were performed five times. Data evaluation and melting point determination were conducted with Bio-Rad CFX MANAGER software.
61
Crystal structure of isoniazid-bound KatG
Preparation of INH-bound SeKatG crystals Purified SeKatG was desalted by repeated concentration with Vivaspin filters (GE Healthcare), and then diluted with 50 mM Hepes buffer (pH 7.0). Crystallization conditions were optimized with the sitting-drop vapor-diffusion method based on previously reported conditions [25]. Drops were prepared by mixing 2 lL of protein (12 mgmL1) with 1 lL of reservoir solution (4.3 M sodium formate and 100 mM sodium citrate, pH 6.4). The crystals grew in 1 week to an average size of 0.5 9 0.25 9 0.25 mm. INH-bound crystals were obtained by soaking the crystals for 24 h in crystallization solution containing 100 mM INH.
X-ray crystallography The INH-bound crystals were flash-cooled with a nitrogen gas stream at –180 °C. Notably, the crystals did not require any cryoprotection, because the crystallization conditions included high salt concentrations. Diffraction data were collected by the use of synchrotron radiation and the MX300HE detector (Rayonix, Evanston, IL, USA) of beamline BL44XU at Spring-8 (Harima, Japan). The crystal was oscillated by 0.5° between image collections, and a total of 400 images were processed. Integrated intensities were merged and scaled with the HKL2000 suite [27]. The results of the data collection are summarized in Table 2.
S. Kamachi et al.
Table 2. Crystallographic data and refinement statisticsa. PDB, Protein Data Bank. Crystallographic data Space group Cell parameters ( A) Resolution range ( A) Observed reflections Unique reflections Mean I/r (I) Redundancy Completeness (%) Rsym (%)b Refinement statistics Rcryst (%)c Rfree (%)d Number of molecules INH Heme Na+ Water Rmsd from ideal values Bond length ( A) Bond angle (°) Average B-factor ( A3 ) Ramachandran plot Most favored (%) Additionally allowed (%) Generously allowed (%) Disallowed (%) PDB entry
P41212 a = b = 108.7, c = 202.9 50.0–2.12 (2.20–2.12) 504 680 126 347 9.9 (3.1) 4.0 (3.9) 95.2 (97.3) 8.0 (51.5) 15.8 19.7 3 1 4 487 0.020 1.95 37.8 89.6 10.1 0.3 0.0 3WXO
a
Structure determination and refinement of INH-bound SeKatG Crystals of INH-bound SeKatG were nearly isomorphous with those of native KatG, and the structure of INHbound SeKatG was determined with the structure of native KatG (Protein Data Bank ID: 3WNU) as a model [9]. The structure was subjected to rigid-body refinement to 3.0-A resolution with REFMAC5 [28] from the from 25-A CCP4 package [29]. The structure was further refined at resolution by restrained refinement in REFMAC5, 2.12-A and manual model revision was carried out with COOT [30]. Ordered water molecules were added to the model by the use of ARP/WARP [31]. The electron density map at the final stage was clear enough for exact assignment of the orientations of the bound INH molecules. The INH molecules were unambiguously fitted into the Fo – Fc map on the protein surface. Structure refinement statistics are summarized in Table 2. Figures 2 and 3 were prepared with PYMOL [32].
Acknowledgements We thank Drs S. Baba and E. Yamashita for their assistance with data collection at the SPring-8 synchrotron radiation facility (Hyogo, Japan). The 62
Values in parentheses are for the outermost shell. Rsym = ΣhklΣi|Ii(hkl) |/ΣhklΣiIi(hkl), where is the average intensity over equivalent reflections. c Rcryst = Σ||Fobs(hkl)| |Fcalc(hkl)||/Σ|Fobs(hkl)|. d Rfree is the R-value calculated for 5% of the dataset not included in the refinement. b
synchrotron radiation experiments were performed on BL38B1 and BL44XU at SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute (Proposal Nos. 2011B2066, 2012A6726, 2012B6725, and 2012B6726). This work was supported by the Program to Disseminate Tenure Tracking (to K. Wada) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), by MEXT Grants-in-Aid of Scientific Research 26440035 (to T. Tada) and 5840023 (to K. Wada), by a grant from the Japan Foundation for Applied Enzymology (to K. Wada), and by a Grantin-Aid for JSPS Fellows 24 1292 (to K. Hirabayashi). We also thank Professor T. Inoue and Dr H. Matsumura (Osaka University) and Dr K. Hirata (RIKEN) for invaluable suggestions, and Y. Motoyama and N. Kaseda (University of Miyazaki) for technical assistance. FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
S. Kamachi et al.
Author contributions Planned experiments: TT and KW; Performed experiments: SK, KH, TT and KW; Analyzed data: SK, KH, TT and KW; Contributed reagents or other essential material: MT, and SS; Wrote the paper: SK, KH, MT, SS, TT and KW.
References 1 Bloom BR & Murray CJ (1992) Tuberculosis: commentary on a reemergent killer. Science 257, 1055– 1064. 2 Wiseman B, Carpena X, Feliz M, Donald LJ, Pons M, Fita I & Loewen PC (2010) Isonicotinic acid hydrazide conversion to Isonicotinyl-NAD by catalaseperoxidases. J Biol Chem 285, 26662–26673. 3 Rozwarski DA, Grant GA, Barton DH, Jacobs WR Jr & Sacchettini JC (1998) Modification of the NADH of the isoniazid target (InhA) from Mycobacterium tuberculosis. Science 279, 98–102. 4 Rawat R, Whitty A & Tonge PJ (2003) The isoniazidNAD adduct is a slow, tight-binding inhibitor of InhA, the Mycobacterium tuberculosis enoyl reductase: adduct affinity and drug resistance. Proc Natl Acad Sci U S A 100, 13881–13886. 5 Wang F, Jain P, Gulten G, Liu Z, Feng Y, Ganesula K, Motiwala AS, Ioerger TR, Alland D, Vilcheze C et al. (2010) Mycobacterium tuberculosis dihydrofolate reductase is not a target relevant to the antitubercular activity of isoniazid. Antimicrob Agents Chemother 54, 3776–3782. 6 Yamada Y, Fujiwara T, Sato T, Igarashi N & Tanaka crystal structure of catalaseN (2002) The 2.0 A peroxidase from Haloarcula marismortui. Nat Struct Biol 9, 691–695. 7 Carpena X, Loprasert S, Mongkolsuk S, Switala J, Loewen PC & Fita I (2003) Catalase-peroxidase KatG resolution. J Mol of Burkholderia pseudomallei at 1.7 A Biol 327, 475–489. 8 Bertrand T, Eady NA, Jones JN, Jesmin Nagy JM, Jamart-Gregoire B, Raven EL & Brown KA (2004) Crystal structure of Mycobacterium tuberculosis catalase-peroxidase. J Biol Chem 279, 38991– 38999. 9 Kamachi S, Wada K, Tamoi M, Shigeoka S & Tada T resolution structure of the catalase(2014) The 2.2 A peroxidase KatG from Synechococcus elongatus PCC7942. Acta Crystallogr F Struct Biol Commun 70, 288–293. 10 Pierattelli R, Banci L, Eady NA, Bodiguel J, Jones JN, Moody PC, Raven EL, Jamart-Gregoire B & Brown KA (2004) Enzyme-catalyzed mechanism of isoniazid activation in class I and class III peroxidases. J Biol Chem 279, 39000–39009.
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
Crystal structure of isoniazid-bound KatG
11 Metcalfe C, Macdonald IK, Murphy EJ, Brown KA, Raven EL & Moody PC (2008) The tuberculosis prodrug isoniazid bound to activating peroxidases. J Biol Chem 283, 6193–6200. 12 Vidossich P, Loewen PC, Carpena X, Fiorin G, Fita I & Rovira C (2014) Binding of the antitubercular prodrug isoniazid in the heme access channel of catalaseperoxidase (KatG). A combined structural and metadynamics investigation. J Phys Chem B 118, 2924– 2931. 13 Kochi A, Vareldzis B & Styblo K (1993) Multidrugresistant tuberculosis and its control. Res Microbiol 144, 104–110. 14 Heym B, Alzari PM, Honore N & Cole ST (1995) Missense mutations in the catalase-peroxidase gene, katG, are associated with isoniazid resistance in Mycobacterium tuberculosis. Mol Microbiol 15, 235–245. 15 Huyen MN, Cobelens FG, Buu TN, Lan NT, Dung NH, Kremer K, Tiemersma EW & van Soolingen D (2013) Epidemiology of isoniazid resistance mutations and their effect on tuberculosis treatment outcomes. Antimicrob Agents Chemother 57, 3620–3627. 16 Wengenack NL, Uhl JR, St Amand AL, Tomlinson AJ, Benson LM, Naylor S, Kline BC, Cockerill FR 3rd & Rusnak F (1997) Recombinant Mycobacterium tuberculosis KatG(S315T) is a competent catalaseperoxidase with reduced activity toward isoniazid. J Infect Dis 176, 722–727. 17 Wengenack NL, Todorovic S, Yu L & Rusnak F (1998) Evidence for differential binding of isoniazid by Mycobacterium tuberculosis KatG and the isoniazidresistant mutant KatG(S315T). Biochemistry 37, 15825– 15834. 18 Rickman KA, Swancutt KL, Mezyk SP & Kiddle JJ (2013) Isoniazid: radical-induced oxidation and reduction chemistry. Bioorg Med Chem Lett 23, 3096– 3100. 19 Lukat-Rodgers GS, Wengenack NL, Rusnak F & Rodgers KR (2000) Spectroscopic comparison of the heme active sites in WT KatG and its S315T mutant. Biochemistry 39, 9984–9993. 20 Kapetanaki SM, Chouchane S, Yu S, Magliozzo RS & Schelvis JP (2005) Resonance Raman spectroscopy of Compound II and its decay in Mycobacterium tuberculosis catalase-peroxidase KatG and its isoniazid resistant mutant S315T. J Inorg Biochem 99, 1401–1406. 21 Kapetanaki SM, Chouchane S, Yu S, Zhao X, Magliozzo RS & Schelvis JP (2005) Mycobacterium tuberculosis KatG(S315T) catalase-peroxidase retains all active site properties for proper catalytic function. Biochemistry 44, 243–252. 22 Deemagarn T, Carpena X, Singh R, Wiseman B, Fita I & Loewen PC (2005) Structural characterization of the Ser324Thr variant of the catalase-peroxidase (KatG) from Burkholderia pseudomallei. J Mol Biol 345, 21–28.
63
Crystal structure of isoniazid-bound KatG
23 Zhao X, Yu H, Yu S, Wang F, Sacchettini JC & Magliozzo RS (2006) Hydrogen peroxide-mediated isoniazid activation catalyzed by Mycobacterium tuberculosis catalase-peroxidase (KatG) and its S315T mutant. Biochemistry 45, 4131–4140. 24 Colin J, Wiseman B, Switala J, Loewen PC & Ivancich A (2009) Distinct role of specific tryptophans in facilitating electron transfer or as [Fe(IV)=O Trp(*)] intermediates in the peroxidase reaction of Bulkholderia pseudomallei catalase-peroxidase: a multifrequency EPR spectroscopy investigation. J Am Chem Soc 131, 8557– 8563. 25 Wada K, Tada T, Nakamura Y, Kinoshita T, Tamoi M, Shigeoka S & Nishimura K (2002) Crystallization and preliminary X-ray diffraction studies of catalase-peroxidase from Synechococcus PCC 7942. Acta Crystallogr D Biol Crystallogr 58, 157–159. 26 Miroux B & Walker JE (1996) Over-production of proteins in Escherichia coli: mutant hosts that allow synthesis of some membrane proteins and globular proteins at high levels. J Mol Biol 260, 289–298. 27 Otwinowsky Z & Minor W (1997) Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol 276, 307–326.
64
S. Kamachi et al.
28 Murshudov GN, Vagin AA & Dodson EJ (1997) Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr D Biol Crystallogr D53, 240–255. 29 Collaborative Computational Project N (1994) The CCP4 suite: programs for protein crystallography. Acta Crystallogr D Biol Crystallogr D50, 760–763. 30 Emsley P & Cowtan K (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr D60, 2126–2132. 31 Cohen SX, Ben Jelloul M, Long F, Vagin A, Knipscheer P, Lebbink J, Sixma TK, Lamzin VS, Murshudov GN & Perrakis A (2008) ARP/wARP and molecular replacement: the next generation. Acta Crystallogr D Biol Crystallogr D64, 49–60. 32 DeLano WL (2002) PyMOL. DeLano Scientific, San Carlos, CA. 33 Thompson JD, Higgins DG & Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22, 4673–4680. 34 Robert X & Gouet P (2014) Deciphering key features in protein structures with the new ENDscript server. Nucleic Acids Res 42, W320–W324.
FEBS Journal 282 (2015) 54–64 ª 2014 FEBS
