HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 22, 20033-20040, May 31, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/22/20033    most recent M200864200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Nunn, C. M. Articles by Ortiz de Montellano, P. R. Search for Related Content PubMed PubMed Citation Articles by Nunn, C. M. Articles by Ortiz de Montellano, P. R. Social Bookmarking                      What's this?

The Crystal Structure of Mycobacterium tuberculosis Alkylhydroperoxidase AhpD, a Potential Target for Antitubercular Drug Design^*

Christine M. Nunn, Snezana Djordjevic^§, Patrick J. Hillas^¶, Clinton R. Nishida^¶, and Paul R. Ortiz de Montellano^¶

From the  Department of Biochemistry and Molecular Biology, University College, Gower Street, London WC1E 6BT, United Kingdom and the ^¶ Department of Pharmaceutical Chemistry, University of California, San Francisco, California 94143-0446

Received for publication, January 28, 2002, and in revised form, March 15, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The resistance of Mycobacterium tuberculosis to isoniazid is commonly linked to inactivation of a catalase-peroxidase, KatG, that converts isoniazid to its biologically active form. Loss of KatG is associated with elevated expression of the alkylhydroperoxidases AhpC and AhpD. AhpD has no sequence identity with AhpC or other proteins but has alkylhydroperoxidase activity and possibly additional physiological activities. The alkylhydroperoxidase activity, in the absence of KatG, provides an important antioxidant defense. We have determined the M. tuberculosis AhpD structure to a resolution of 1.9 Å. The protein is a trimer in a symmetrical cloverleaf arrangement. Each subunit exhibits a new all-helical protein fold in which the two catalytic sulfhydryl groups, Cys-130 and Cys-133, are located near a central cavity in the trimer. The structure supports a mechanism for the alkylhydroperoxidase activity in which Cys-133 is deprotonated by a distant glutamic acid via the relay action of His-137 and a water molecule. The cysteine then reacts with the peroxide to give a sulfenic acid that subsequently forms a disulfide bond with Cys-130. The crystal structure of AhpD identifies a new protein fold relevant to members of this protein family in other organisms. The structural details constitute a potential platform for the design of inhibitors of potential utility as antitubercular agents and suggest that AhpD may have disulfide exchange properties of importance in other areas of M. tuberculosis biology.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Tuberculosis is a leading pathogen worldwide, infecting an estimated 8 million and killing 2-3 million people each year (1). Furthermore, it is estimated that ~2 billion people have been exposed to Mycobacterium tuberculosis and thus are at risk of developing the active disease. Although most of the infected individuals reside in less industrialized countries, the rates of infection in areas undergoing rapid social change, such as the countries of the former Soviet Union, are also increasing at an alarming rate (2). In part because of the symbiotic relationship between human immunodeficiency virus and tuberculosis, and the increasingly rapid displacement of populations, the incidence of multidrug-resistant tuberculosis is an important worldwide concern. Thus, in a recent survey of 35 countries, 12.6% of M. tuberculosis isolates were found to be resistant to at least one drug, and 2.2% were resistant to both isoniazid and rifampin, the two primary drugs used to treat tuberculosis (3). This resurgence of tuberculosis has led to renewed efforts to find new drugs for the treatment of this dread disease, particularly agents that exhibit activity against drug-resistant strains, completely sterilize the infection, or shorten the duration of drug therapy and thus promote drug compliance.

One of the primary drugs used to treat tuberculosis is isoniazid, a prodrug that is oxidized by the KatG catalase-peroxidase to an activated form that inhibits cell wall synthesis (4, 5). Isoniazid resistance commonly results from mutation of the KatG to a protein that is either inactive or has an impaired ability to activate isoniazid (6-9). The physiological function of the KatG catalase-peroxidase has not been completely defined, but includes protection of the mycobacterium against H₂O₂ and other reactive oxygen species produced by the microbe and its host (10, 11). The attenuated virulence of KatG null strains of M. tuberculosis has been definitively established by experiments in which resistance to H₂O₂, and virulence, are restored by transfection with plasmid-encoded KatG (12-15). A conundrum therefore exists, in that loss or attenuation of KatG activity, an activity important for virulence, is required for resistance to isoniazid.

A possible solution to this conundrum is offered by the finding that suppression of KatG activity is paralleled by increased expression of the alkylhydroperoxidase AhpC (4, 11, 16-18). AhpC is a member of the widespread non-heme peroxiredoxin family that catalyzes the reduction of alkylhydroperoxides to alcohols (19), although additional functions have been ascribed to individual members of this protein family. For example, human peroxiredoxins, in addition to their alkylhydroperoxidase activity, have phospholipase A2 activity (20) and regulate NF-B¹ activation (21). In microorganisms, the ahpC gene is usually under control of the oxyR locus, but this locus in M. tuberculosis is not functional (22, 23). Enhanced expression of AhpC in M. tuberculosis involves elevated expression of the protein due to mutations in its promoter region (11, 17). The M. tuberculosis AhpC has been expressed, purified, and shown to have peroxidase activity toward a broad range of substrates (24). This activity was measured using either dithiothreitol or the Salmonella typhimurium AhpF flavoprotein as a surrogate electron donor, because the native electron donor partner in M. tuberculosis has not been identified.

The genes for a number of flavoproteins of unknown function are present in M. tuberculosis, but none of them correspond to the S. typhimurium ahpF gene (25). The position immediately downstream of ahpC in the S. typhimurium genome is occupied by ahpF, but in the M. tuberculosis genome this position is taken by a gene termed ahpD because of its position relative to ahpC rather than because of any sequence or functional evidence. Homologous ahpD genes have so far been found in a few other organisms, including other mycobacteria, Streptomyces viridosporus,² Streptomyces coelicolor (27), Brucella melitensis (28), and Ralstonia solanacearum.³ We recently expressed the M. tuberculosis AhpD and demonstrated that, despite the absence of any sequence identity between AhpD and the AhpC family of proteins, it also exhibits alkylhydroperoxidase activity with the same surrogate electron transfer partners as utilized by AhpC (24). The specific role of AhpD in detoxifying peroxides and other reactive oxygen and nitrogen species, and any possible additional functions of the protein, remain undetermined. Unlike ahpC, ahpD appears not to have been among the genes examined in microarray studies of changes in M. tuberculosis gene expression caused by drugs and hypoxia (18, 30). However, its position relative to ahpC suggests that AhpD expression may be enhanced under the same conditions that enhance AhpC expression.

The ahpD (Rv2429) gene encodes a protein of 177 amino acids and a molecular mass of 18,781 Da.⁴ Size-exclusion chromatography of the protein expressed in Escherichia coli indicates that AhpD is present as an oligomeric, possibly dimeric, species (24). The AhpD sequence reveals the presence of two cysteines (Cys-129 and Cys-132, or Cys-130 and Cys-133 if the initial methionine that is absent in the heterologously expressed protein is counted). The two cysteines are separated by a serine and a histidine. Site-specific mutagenesis of the cysteines has shown that both, but particularly Cys-133, are important for alkylhydroperoxidase activity. We report here crystallization of M. tuberculosis AhpD and determination of its crystal structure to a resolution of 1.9 Å. The crystal structure reveals a homotrimeric protein in which each of the subunits has an identical but novel protein fold. The structure suggests a mechanism for the alkylhydroperoxidase activity of AhpD.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- M. tuberculosis AhpD was heterologously expressed in E. coli and purified as previously reported (24). All chemical reagents were purchased from Sigma Chemical Co. (St. Louis, MO). E. coli strain DL41, auxotrophic in methionine biosynthesis due to a metA disruption (32), was obtained from the Yale University E. coli Genetic Stock Center (CGSC). Selenomethionine was from Acros. Trisodium citrate was from Aldrich (St. Louis, MO). LB and agar were from Difco. Acetic acid, chloramphenicol, NaCl, NaOH, SDS, and MOPS were from Fisher. Protein molecular weight standards were from Invitrogen (Gaithersburg, MD). Thymine, guanosine, and adenine hemisulfate dihydrate were from ICN Biomedical (Aurora, OH). Purified proteins were concentrated using Millipore YM10-regenerated cellulose ultrafiltration membranes. Q-Sepharose was from Amersham Biosciences, Inc. (Peapack, NJ). IPTG was from Promega (Madison, WI). SC-Met was from Qbiogene (Carlsbad, CA). Polyethyleneimine (10% solution) was from Research Biochemicals International (Natick, MA). Ammonium sulfate, SigmaUltra dithiothreitol, EDTA, KCl, glycerol, magnesium sulfate heptahydrate, d-(+)-glucose, potassium phosphate (mono- and di-basic), uracil, d-biotin, thiamine hydrochloride, and amino acids (Phe, Thr, Val, Leu, Ile, Lys, and Met) were from Sigma Chemical Co.

Selenomethionine-labeled Protein-- Attempts to express the selenomethionine-labeled protein in the methionine auxotrophic E. coli strain DL41 (32) were unsuccessful. The labeled protein was therefore expressed in BL21 cells under conditions that minimize methionine biosynthesis. Thus, two 5-ml LB/chloramphenicol cultures were inoculated with a single colony from a freshly streaked plate and grown at 37 °C for 8 h or less. Cells spun down at 2500 rpm for 15 min were resuspended in A medium to minimize the addition of methionine from the rich LB medium. The A₆₀₀ of the resuspension was measured, and cells were diluted to A₆₀₀ ~ 0.001 into 1 liter of A medium in a 3-liter Fernbach flask. Each liter of growth medium was supplemented with 2 g of glucose (10 ml of 20% glucose) and 34 mg of chloramphenicol. Growth at 37 °C, 250 rpm, was performed until the A₆₀₀ was 0.7, a process that took at least 10-12 h. To each liter of culture was added 100 mg each of Phe, Lys·HCl, and Thr and 50 mg each of Ile, Leu, and Val. For native or SeMet expression, 30 mg/liter l-methionine or l-selenomethionine also was added, respectively. After 15 min of further growth, IPTG was added to 0.2 mM (200 µl of 1 M IPTG), expression was continued for 6-8 h, and the cells were harvested and stored frozen at 70 °C overnight. If desired, expression can be done for as long as 14 h to increase cell yield.

The selenomethionine-labeled protein was purified according to the protocol of Hillas et al. (24), except that 5 mM dithiothreitol and 0.2 mM EDTA were added to all the buffers, and all the buffers used in protein purification were degassed. The degassing, EDTA, and high concentration of dithiothreitol were employed to minimize the oxidation of selenomethionine to the selenoxide (33). The buffers were degassed by subjecting them, with stirring, to a minimum of three 15-min cycles of house-line vacuum followed by 2 min of bubbling with nitrogen using a bubbler with a fine frit to optimize mixing of the gas into the buffer. In the case of buffers to be applied to chromatography columns, the final cycle should be done sufficiently in advance to allow dispersion of the gas bubbles, because this prevents the introduction of bubbles into the column.

Per each gram of cells was added 5 ml of buffer B (50 mM potassium phosphate, pH 7.0, 5% glycerol, 0.2 mM EDTA, and 5 mM dithiothreitol) plus 0.5 mg/ml lysozyme, 44 µg/ml phenylmethylsulfonyl fluoride, 2 µg/ml leupeptin, and 2 µg/ml pepstatin. The suspended cells were incubated on ice for 60 min and then, while still on ice, were sonicated with a Branson sonifier 1/2" tip on setting 3 (~20-watt output) for 4 cycles of alternating 30-s continuous pulses followed by 30 s of rest. To the supernatant obtained by ultracentrifugation for 30 min at 100,000 × g (4 °C) was added polyethyleneimine to a concentration of 0.005%. The mixture was incubated on ice for 15 min to precipitate the nucleic acids, and the mixture was then centrifuged as above to produce the crude lysate supernatant.

The crude lysate was loaded onto a 2.5- × 17-cm (~80 ml) Q-Sepharose column at 1 ml/min. The column was washed with 10-20 column volumes of buffer B at 1 ml/min, and the protein was then eluted at 1 ml/min using a linear gradient of 375 ml of buffer B and 375 ml of buffer B containing 100 mM KCl. Fractions of 7.5 ml resulted in elution of the AhpD protein in fractions 1-45. Every other fraction from 1 through 43 was analyzed via SDS-PAGE. Pure fractions were pooled and concentrated to 20 mg/ml or greater using a Millipore concentrator equipped with a YM10 (10,000 molecular weight cutoff) regenerated cellulose ultrafiltration membrane. The concentration was determined using a molar extinction coefficient of 15,720 M¹ cm¹ (24). For volumes of 0.5-3.0 ml, the concentrated solution was loaded into a Pierce Slide-A-Lyzer cassette and dialyzed three times against 1 liter of 50 mM MOPS, pH 7.2, 100 mM KCl, 20% glycerol, 5 mM dithiothreitol, and 0.2 mM EDTA. The protein was then aliquoted into 25- or 100-µl volumes, frozen on dry ice, and stored until used at 80 °C.

The protein was subjected to analysis on 20% homogeneous gels using the PhastSystem (Amersham Biosciences, Inc.). The protein molecular weight standards (Invitrogen) were prestained insulin ( and  chains, usually unresolvable), bovine trypsin inhibitor, lysozyme, -lactoglobulin, carbonic anhydrase, and ovalbumin, which have apparent molecular weights of 3,145, 6,060, 14,840, 19,875, 29,435, and 45,695 g/mol.

Crystallization-- Crystallization was carried out by the hanging drop method. To each well in a 24-well plate was added 1.0 ml of 100 mM sodium citrate buffer, pH 5.6, containing 200 mM ammonium acetate, and 26% polyethylene glycol 4000. A freshly thawed solution of AhpD (4.5 mg/ml) in 25 mM MOPS buffer, pH 7.2, containing 50 mM KCl, 10% glycerol, 0.1 mM EDTA, stored at 80 °C, was brought to a concentration of 5.0 mM dithiothreitol with an ice-cold stock solution of the reducing agent. The final solution was kept at 4 °C, and the crystallization drops were laid down as quickly as possible. In the case of the selenomethionine-labeled protein, the protein stock solution already contained 5.0 mM dithiothreitol but a further 5 mM was added. A 2-µl aliquot of the well solution was placed on each coverslip followed by a 2-µl aliquot of the protein solution. The coverslip was then inverted over the well and the plate was allowed to stand at room temperature (between 17 and 21 °C) until appropriate crystals appeared. Crystallization usually occurred within a period of 3-4 days. The crystals appeared as large, rhombohedral prisms. The crystals were of space group C2 with cell dimensions a = 186.38 Å, b = 117.28 Å, c = 88.99 Å,  = 113.97°.

Data Collection and Processing-- Multi-wavelength x-ray data were collected at the beamline BM14, European Synchrotron Radiation Facility, Grenoble. An x-ray fluorescence spectrum was recorded and used to select the wavelength optima for the subsequent MAD data collections. All data sets were collected using the same single, flash-frozen crystal of SeMet-AhpD of size 0.35 × 0.45 × 0.2 mm. The four data sets were collected at  = 0.97877 Å (f" maximum),  = 0.97890 Å (inflection point),  = 0.91840 Å (remote high energy wavelength), and at  = 0.8855 Å. Each data set was individually processed, integrated, and scaled using the HKL and CCP4 suite of programs with unit cell dimensions a = 99.334 Å, b = 58.636 Å, c = 88.989 Å,  = 120.980° (34, 35). This cell is consistent with the cell parameters, which had been obtained from our home source. The high resolution data set was also processed in the C2 cell with dimensions a = 186.38 Å, b = 117.28 Å, c = 88.99 Å,  = 113.97° and solvent content of 37%; the Matthews coefficient was 1.97 Å³ Da¹. Data statistics for the MAD data collections are given in Table I (see below).

Multi-wavelength Phasing, Model Building, and Refinement-- Within the smaller C2 cell a protein trimer exists within the crystallographic asymmetric unit. Selenium sites were located from Patterson search procedures using the program CNS (36). Six selenium sites were located with two selenium sites in each AhpD protein, at positions 78 and 80 of the protein sequence. After density modification in CNS (36) initial chain tracing was carried out using the program TURBO-FRODO (37). ARP/wARP (38) was run in parallel and located some of the -helical regions of the structure (approximately one-third of the structure was located using ARP/wARP). 10% of the reflections were used for cross-validation analysis (39), and the behavior of R_free was used to monitor the refinement strategy. For the structure building and subsequent refinement, data to 1.9 Å were used. Model building was carried out using programs O (40) and Turbo, and refinement was achieved using CNS. Refinement included simulated annealing with torsion angle dynamics, anisotropic scaling, energy minimization, individual isotropic B-factor refinement, and bulk-solvent correction against the maximum-likelihood target.

The structure refinement proceeded well until the later stages when it became clear that some parts of the structure (loops primarily) were poorly fitted. A re-inspection of the data frames and re-indexing displayed a larger unit cell to be the correct unit cell for this structure with a volume four times the size of the initial cell. This cell was not identified in the first instance due to weak diffraction for the additional reflections and could only be seen fully using the synchrotron source. After reprocessing of the data molecular replacement was carried out using co-ordinates of the refined AhpD trimer within CNS. The cross-rotation function produced three distinct peaks, and the translation functions identified the positions for three further AhpD trimers. The trimers are related by translation only. Following molecular replacement, refinement of the four independent trimers was carried out using CNS, with a similar protocol to that described above. Because the starting trimer was already reasonably well refined in the smaller cell, at this point use of the non-crystallographic symmetry (NCS) constraints or restraints did not lead to improvement of the model and, therefore, the refinement was continued without NCS. Refitting of the previously poorly fitting regions of the protein showed significant improvement as the refinement proceeded. During the final stages of the refinement well-defined residual electron density peaks in difference Fourier maps were assigned as solvent water molecules and included in the model. At the end of the refinement four residues were omitted from the model due to a lack of electron density in these regions (residues 61 from chains B and E, and residue 99 from chains G and L), and no residual electron density was unaccounted for. Root mean square deviations between the trimer-comprising protein chains ABC and the other three protein trimers is 0.58 Å (ABC/DEF), 1.12 Å (ABC/GHI), and 0.91 Å (ABC/JKL). Higher R_factor values reflect the non-crystallographic symmetry that gave rise to systematically weaker reflections that were not clearly detectable until the crystals were diffracted by a stronger x-ray source (synchrotron radiation). These weaker reflections are associated with the larger unit cell. It has been recently noted that these types of structures do not refine to give low R_factors (41). While this report was under review, another report on the AhpD structure was published in which the protein was crystallized under different conditions in a different crystal form (42). Comparison of the two structures shows that the root mean square deviation (r.m.s.d.) for the main-chain atoms is 0.5 Å, providing independent confirmation of the validity of both structures. Refinement of the structure was completed with 12 protein molecules in the crystallographic asymmetric unit. Diffraction data for the correct C2 data is shown in Table II (see below), and the final refinement parameters are in Table III (see below).

Preparation of Figures-- The figures were prepared with MOLSCRIPT (43), GRASP (44), TURBO-FRODO (37), and DINO (Visualizing Structural Biology (2001); available at www.dino3d.org).

Coordinates-- The atomic coordinates have been deposited in the Protein Data Bank and are available under PDB identity code 1gu9.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Overall Structure-- Electron density maps calculated with solvent-flattened MAD phases revealed that the protein is built entirely of -helices. These maps were used both for tracing the C chains and for the initial side-chain assignments that were assisted by locating the positions of the selenium atoms in the asymmetric unit. In contrast to the proposed homodimeric composition of AhpD suggested by analogy to its functional analogs and by gel-filtration data (24), the crystal structure showed that AhpD exists as a trimer (Fig. 1A). Subsequent equilibrium centrifugation experiments have confirmed that AhpD also exists as a trimer in solution.⁵ The trimer has an overall cloverleaf shape with a local 3-fold symmetry axis running through the center. The two faces of the cloverleaf-like structure are not identical; they are assembled as in a propeller from the different surfaces of the monomeric subunits. The width of the central opening varies from ~8 Å on one to ~12 Å on the other side of the trimer. The molecular surface is further characterized by the presence of three shallow grooves, each associated with one of the monomers (Fig. 1B). The monomers are tightly associated with each other through hydrophobic interactions between adjacent helices as well as through hydrogen bonds closer to the surface of the molecule. Each of the monomers interacts with two other subunits with ~2800 Ų of average surface area being buried between the two subunits. A total of ~8500 Ų of surface area is therefore buried in formation of the trimer, a number to be compared with the 8900 Å² of accessible surface area of each individual monomer. The trimer is further stabilized by hydrogen bonds within the central cavity, primarily between the Arg-86 residue of each of the subunits and the backbone carbonyl groups of residues 126 and 127 from an adjacent subunit. Specifically, Arg-86A interacts with the carbonyls of chain C, Arg-86B with the carbonyls of chain A, and Arg-86C with the carbonyls of chain B. 

View larger version (51K): [in this window] [in a new window]   Fig. 1.   AhpD forms an all-helical trimer. A, ribbon representation of the trimer organization of AhpD with each of the subunits colored separately in green, orange, and pink. B, electrostatic potential of the trimer molecular surface with regions of negative potential shown in red and regions of positive electrostatic potential shown in blue. An atomic model of AhpD is also shown. Green arrows point to the positions of the Cys-130 residues located near the central opening. The white arrow points to the shallow groove that is postulated to be an alkylhydroperoxide binding site.

The final refined model contains 12 subunits (four trimers), labeled A to L, with residues 4A-175A, 3B-175B, 3C-170C, etc. making a total of ~2100 residues (Table III). Data collection and model statistics are given in Tables I-III. A portion of the final difference Fourier map showing the central region of the trimer is shown in Fig. 2. Each of the subunits in the trimer ABC was built independently, and subsequently the initial models for the other three trimers were generated from the ABC trimer by applying local symmetry matrixes. Furthermore, each of the subunits was manually adjusted and refined independently. For the purpose of this manuscript we will be referring only to trimer ABC in the figures and discussion.

View larger version (145K): [in this window] [in a new window]   Fig. 2.   Final weighted 2Fo  Fc electron density map calculated by CNS (36) for the central region of the AhpD trimer. Overlaid with the electron density is a refined model of AhpD. The positions of the catalytic Cys-130 residue and Arg-86, a residue involved in intersubunit interaction, are labeled.

The B-factors within the trimer indicate that the ends of the protein chains, particularly the N terminus, and the solvent-exposed loops that connect the helices have the highest temperature factors. These are the structural elements that would normally be expected to have the highest mobility.

AhpD Forms a Novel Fold-- A single subunit of AhpD is built of eight helices that fold in a pseudo-symmetrical topology (Fig. 3) with the core of the protein built of four central helices (3, 4, 6, 7) flanked by two long helices (5 and 8) on either side. A helix-turn-helix motif of the N-terminal residues 3-27 (1 and 2) forms the top of the double figure of eight-like fold whereas the corresponding residues (92-114) on the other side of the central region are less structured with only one 3₁₀ and one alpha helical turn. Some of the residues facing the core of the protein in this region of the molecule are involved in intersubunit interactions, whereas other residues in the same region are more flexible and solvent-exposed. The structure was compared against all the structures in the Protein Data Bank (September 2001) using the algorithm GRATH⁶ and representatives from the CATH data base (45). The GRATH algorithm ranks all the pairwise comparisons using an optimized scoring function, and similar folds are found in the top 10 matches. For AhpD no global, significant structural match was found to any protein already in the Protein Data Bank. Within the trimer, intermonomer interactions mostly involve the two long helices 5 and 8 such that, for example, helix 5 of subunit A interacts with helix 8 of subunit C on one side whereas helix 8 of subunit A (located on the opposite face of the central four-helix array) interacts with helix 5 of subunit B. This interaction is hydrophobic in nature and involves the N-terminal two-thirds of helix 5. Importantly, the C-terminal part of helix 5B and several of the following residues of the long connection to helix 6 are in close contact with the N-terminal part of 7A, the helix that bears the active site residues. In this area, the two subunits A and C are also held together by a salt bridge between Asp-97A and Arg-148C. In addition, helix 2 also contributes to this interface: Specifically, there is a hydrogen bond between Asp-16C (2) and nitrogen atoms of the backbone residues 105A and 106A.

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Schematic ribbon representation of a monomer fold. The N and C termini and each of the eight alpha helices are labeled.

The Active Site-- The catalytically active residues Cys-130 and Cys-133 are located at the N-terminal end of helix 7, with Cys-130 bordering on the central cavity of the trimer. Structurally the Cys-Ser-His-Cys motif of AhpD resembles that of the thioredoxin family of proteins in which the two active site Cys residues are similarly located toward the N terminus of the helix. However, in thioredoxin the helix is part of a structure with adjacent parallel strands, and the Asp, implicated as the catalytically relevant acid-base residue, is located on one of the -strands. In AhpD, which is an all-helical molecule, no -strands are involved in formation of the active site; however, the position that would spatially coincide with the location of the aspartate carboxylate group is occupied by the imidazole ring of His-137 (Fig. 4).

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Stereo drawing of the active site. The figure shows the three-dimensional arrangement of the AhpD residues involved in the reaction mechanism, Glu-118, Cys-130, Cys-133, His-132, and His-137, as well as the water molecule (represented by a sphere). In addition, the Asp-118 residue (labeled by an asterisk) corresponding to the base in the thioredoxin active site is shown. This residue was placed by overlapping the three-dimensional arrangement of the helical turn containing di-cysteine motifs in the active sites of thioredoxin and AhpD.

The active sites appear to be structurally independent, because all of the residues surrounding the two cysteines belong to the same subunit. However, it is possible that the individual subunits are unstable and that intersubunit interactions preserve the active site fold of the individual monomers. Although Cys-130 is located at the edge of the central cavity, the distance between the closest Cys residues of neighboring subunits is ~15 Å (Fig. 3), which clearly excludes the possibility of intersubunit disulfide bond formation and further suggests that the active sites function independently.

Proposed Reaction Mechanism-- A cysteine residue in alkylhydroperoxidase (or peroxiredoxin) enzymes generally reacts with a hydroperoxide to give a sulfenic acid and the reduced hydroxy product. The sulfenic acid then reacts with a second sulfhydryl to give a disulfide bond. This second sulfhydryl is usually that of another cysteine residue in the same polypeptide or in a second subunit within an oligomeric complex, although it can also be provided by a small thiol molecule.

In AhpD from M. tuberculosis, each of the three subunits in the trimer has two cysteine residues, Cys-130 and Cys-133. The thiol group of Cys-130, which borders the central cavity, is 3.5 Å from that of Cys-133. The geometric disposition of these cysteine sulfhydryls, despite the absence of sequence similarity with the peroxiredoxins, suggests that a related mechanism involving the two cysteine residues is responsible for the reduction of alkylhydroperoxides by AhpD. Because the protein was crystallized in the presence of dithiothreitol, which is able to reduce the enzyme (24), no disulfide bond is present in the enzyme. However, the relative orientation of the Cys-130 and Cys-133 thiol groups and the distance of 3.5 Å between them are clearly suitable for the formation of a disulfide bond between the two cysteines in the oxidized protein.

Mutagenesis studies have shown that mutation of Cys-133 to a serine residue completely suppresses the alkylhydroperoxidase activity of the enzyme supported by the Salmonella typhimurium AhpF redox partner, whereas similar mutation of Cys-130 leaves a low residual activity (24). This suggests, but does not prove, that Cys-133 is the residue converted to a sulfenic acid by reaction with alkylhydroperoxidase substrates. Two histidine residues appear poised to facilitate the catalytic process. The nitrogen of His-132 is 4.7 Å from the sulfur of Cys-130 and 3.7 Å from that of Cys-133 and thus could serve as a base for deprotonation of either cysteine. The second histidine is His-137, the nitrogen of which is 5.1 Å from the sulfur of Cys-133 and 8.3 Å from that of Cys-130. His-137 is too far from Cys-130 to interact productively with it, but it can readily interact with the sulfhydryl of Cys-133 through an intervening water molecule that is located 3.3 Å from the sulfur atom of that residue and 2.5 Å from the nitrogen of His-137. Furthermore, the N-H hydrogen of His-137 is hydrogen-bonded to Glu-118, which is located 2.6 Å from the histidine nitrogen. This hydrogen-bonding interaction should markedly increase the basicity of His-137, because protonation of the N nitrogen of the histidine could be coupled to transfer of the N-H proton to the glutamic acid carboxylate group. Another interesting observation is that the relation of His-137 to Cys-133 is similar to that between the catalytic aspartate and the cysteine in thioredoxins, in which the aspartate assists in deprotonation of the cysteine via the relay action of an intervening water molecule (46). Furthermore, spacing of the cysteines by two intervening residues is the same as that found in the thioredoxins (47). We therefore propose that deprotonation of Cys-133 facilitated by His-137 leads to nucleophilic attack of the cysteine thiolate on the peroxide to give the sulfenic acid intermediate (Fig. 5). Hydrogen bonding of the incipient thiolate with the amide hydrogen of Asn-82 may assist its formation, because the amide nitrogen of Asn-82 is only 3.4 Å away from the cysteine sulfhydryl group. Furthermore, His-137 is so positioned that the proton transferred to it from the cysteine can be used to protonate the alkoxy anion that is formed as the hydroperoxide oxygen-oxygen bond is broken by attack of Cys-133.

View larger version (12K): [in this window] [in a new window]   Fig. 5.   Schematic drawing of the proposed mechanism for the alkylhydroperoxidase activity of AhpD.

The second step of the catalytic cycle, formation of the disulfide bond between Cys-133 and Cys-130 with elimination of the sulfenic hydroxyl group as a water molecule, is likely to be triggered by deprotonation of Cys-130 by His-132, although a small movement of the residues toward each other would be necessary due to the distance of 4.7 Å between the nitrogen and sulfur atom. The incipient Cys-130 thiolate anion would receive some stabilization from hydrogen bonding with the backbone nitrogen of Arg-86, which is 3.5 Å from the cysteine sulfur atom. Attack of the cysteine thiolate on the sulfenic acid would give a disulfide bond between Cys-130 and Cys-133. The proton transferred to His-132 in this reaction step could be used to protonate the sulfenic hydroxyl group, facilitating its elimination as a water molecule, or could be retained to facilitate the disulfide cleavage reaction in the final step of the catalytic cycle. This final step involves reaction with a thiol-disulfide exchange protein that can be replaced in catalytic assays by the AhpF protein of S. typhimurium (24).

Putative Substrate Binding Site-- The molecular surface of AhpD exhibits a very polar character, as shown by its calculated electrostatic potential. The central cavity is particularly positively charged and filled with water molecules and is therefore an unlikely binding site for the lipophilic compounds that were shown to be substrates for AhpD (24). In addition to the central cavity of the cloverleaf trimeric structure there is a shallow hydrophobic groove associated with each of the subunits that extends from the outer, solvent-exposed edge to the active site cysteine residues. We propose that this groove is the binding site for alkylhydroperoxide substrates (Fig. 6). This binding site is formed by the aliphatic side chains of the residues Ile-77 and Met-78 from helix 5, Met-104, Ile-106, and Ile-107 from the connecting loop leading to helix 6, and Phe-117 from the helix 6. As already mentioned, helix 5 and the beginning of the following loop are involved in interactions with the neighboring subunit, and therefore the trimeric assembly may be of importance in maintaining the structure of the active site. Simple docking experiments with cumene hydroperoxide, one of the known substrates, confirmed that this compound could bind to the putative substrate-binding site with the phenyl group in a hydrophobic environment and the peroxide moiety accessible to Cys-133. The trimeric surface of the AhpD molecule together with the central cavity might, on the other hand, provide a binding site for the reducing protein. This would create an efficient interaction in which a single molecule would be able to interact with any of the three disulfide bonds without precluding access of the lipid peroxides to the active Cys residues.

View larger version (71K): [in this window] [in a new window]   Fig. 6.   Molecular surface showing the putative substrate binding site. The three subunits in the trimer are colored orange, turquoise, and pink. Within the binding cavity, hydrophobic residues described in the text are colored green. The blue surface at the bottom of the cavity is due to Cys-133. The reaction starts by deprotonation of this cysteine. The enlarged portion of the figure, on the top, is denoted within the whole trimer surface.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The crystal structure of AhpD from M. tuberculosis together with mutagenesis data allows us to identify some of the catalytic residues and to describe a reaction mechanism for the alkylhydroperoxidase activity of this enzyme. Analysis of the proteins now known to have sequence identity with the M. tuberculosis AhpD indicates that the catalytic motif is largely preserved (27, 28).^2,3 This catalytic motif consists of two cysteines separated by two amino acid residues. The two spacer residues differ, being Ser-His in the mycobacterial proteins, Gly-Met in the Brucella melitensis and Streptomyces proteins, and His-Phe in the Ralstonia solanacearum protein. Thus, the equivalent of His-132 is not present in some of the proteins and its catalytic role, if any, would have to be played by alternative residues. However, the equivalent of His-137 is conserved in all of the proteins, as is a carboxylate group equivalent to Glu-118. It is therefore likely that the basic mechanism proposed for the activation of Cys-133 is preserved in this new and growing family of proteins.

The two cysteine residues in AhpD more closely resemble the cysteines in thioredoxin than those in the peroxiredoxin family of enzymes, although apart from the spacing of the two cysteines there is little sequence identity between AhpD and the thioredoxins. The catalytic cysteines in the AhpC proteins are located at opposite ends of the polypeptide (48), whether their catalytic action involves disulfide bond formation within the same polypeptide (26) or between two polypeptides in an oligomeric structure (29). In AhpD, the cysteine residues of one subunit are at least 15 Å from those of the adjacent subunit and are therefore clearly too far apart for intersubunit disulfide bond in the absence of a massive conformational change. On the other hand, as in the thioredoxins, the disulfide pair within a single AhpD chain is optimally placed for intramolecular disulfide bond formation.

A further resonance with the structure of the thioredoxins is provided by the mechanism of activation of Cys-133 in AhpD. In the thioredoxins, a carboxylate oxygen is located ~6 Å from the sulfur of one of the two catalytic cysteines (46). However, through an intervening water molecule, it is proposed that the carboxylic acid serves as the base that deprotonates the cysteine residue. Comparison of the location in protein space of the carboxylic acid in the thioredoxins shows that the corresponding site is occupied in AhpD by a histidine, specifically His-137. This histidine is activated by hydrogen bonding to Glu-118, and a water molecule provides a bridge between the carboxylate oxygen atoms and the thiolate of Cys-133. Thus, a mechanism for deprotonation of the cysteine can be written that closely resembles that of thioredoxin, except that the carboxylate is even further from the cysteine proton and its function as an acid-base catalyst is relayed through a histidine and a water molecule.

Completion of the catalytic cycle requires reduction of the disulfide bond that is formed between Cys-130 and Cys-133. In our earlier catalytic studies, reduction of the double bond was achieved using either dithiothreitol or the S. typhimurium AhpF as a surrogate reducing partner (24). In each instance, a disulfide exchange reaction is responsible for regeneration of the reduced form of AhpD. During review of this report, an endogenous system was reported in M. tuberculosis that is able to efficiently reduce AhpD (42). This system consists of SucB, a dihydrolipoamide-containing protein that undergoes a disulfide exchange reaction with AhpF, and dihydrolipoamide dehydrogenase, which regenerates the dihydrolipoamide cofactor of SucB.

	ACKNOWLEDGEMENTS

We thank Kinkead Reiling of the Robert Stroud laboratory at the University of California, San Francisco, for help with the initial crystallization trials. We also thank F. Pearl, A. Harrison, and C. Porter for their help with structure analysis.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grant GM56531 and the Higher Education Council for England United Kingdom.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The atomic coordinates and the structure factors (code 1gu9) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/).

^§ To whom correspondence may be addressed: Dept. of Biochemistry and Molecular Biology, University College, Gower St., London WC1E 6BT, UK. Tel.: 44-020-7679-2230; Fax: 44-020-7679-7193; E-mail: snezana@biochemistry.ucl.ac.uk.

To whom correspondence may be addressed: School of Pharmacy, S-926, University of California, San Francisco, CA 94143-0446. Tel.: 415-476-2903; Fax: 415-502-4728; E-mail: ortiz@cgl.ucsf.edu.

Published, JBC Papers in Press, March 25, 2002, DOI 10.1074/jbc.M200864200

² S. Ramachandran, T. S. Magnuson, and D. L. Crawford (1999) GenBank^TM accession number AAD33341.

³ M. Salanoubat, S. Genin, F. Artiguenave, J. Gouzy, S. Manguenot, M. Arlat, A. Billault, P. Brottier, J. C. Camus, L. Cattolico, C. Gaspin, M. Lavie, A. Moisan, C. Robert, W. Saurin, T. Schiex, P. Siguier, P. Thebault, M. Whalen, P. Wincker, M. Levy, J. Weissenbach, and C. A. Boucher (2001) GenBank^TM accession number NP519766.

⁴ Available at www. sanger.ac.uk/cgi-bin/yeastpub/get_tb_orf_working.perl.

⁵ G. Knudsen and P. R. Ortiz de Montellano, unpublished results.

⁶ A. Harrison, F. Pearl, I. Sillitoe, T. Slidel, R. Mott, J. Thornton, and C. Orengo, submitted for publication.

	ABBREVIATIONS

The abbreviations used are: NF-B, nuclear factor B; IPTG, isopropyl-1-thio--D-galactopyranoside; MOPS, 3-morpholinepropanesulfonic acid; MAD, multiwavelength anomalous diffraction; NCS, non-crystallographic symmetry; r.m.s.d., root mean square deviation.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				K. Ito, R. Arai, E. Fusatomi, T. Kamo-Uchikubo, S.-i. Kawaguchi, R. Akasaka, T. Terada, S. Kuramitsu, M. Shirouzu, and S. Yokoyama Crystal structure of the conserved protein TTHA0727 from Thermus thermophilus HB8 at 1.9 A resolution: A CMD family member distinct from carboxymuconolactone decarboxylase (CMD) and AhpD Protein Sci., May 1, 2006; 15(5): 1187 - 1192. [Abstract] [Full Text] [PDF]