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

J. Biol. Chem., Vol. 277, Issue 20, 18134-18142, May 17, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/20/18134    most recent M200912200v1 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 Fu, Z. Articles by Kim, J.-J. P. Search for Related Content PubMed PubMed Citation Articles by Fu, Z. Articles by Kim, J.-J. P. Social Bookmarking                      What's this?

The Structure of a Binary Complex between a Mammalian Mevalonate Kinase and ATP

INSIGHTS INTO THE REACTION MECHANISM AND HUMAN INHERITED DISEASE^*

Zhuji Fu, Ming Wang, David Potter, Henry M. Miziorko, and Jung-Ja P. Kim

From the Department of Biochemistry, Medical College of Wisconsin, Milwaukee, Wisconsin 53226

Received for publication, January 28, 2002, and in revised form, February 26, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Mevalonate kinase catalyzes the ATP-dependent phosphorylation of mevalonic acid to form mevalonate 5-phosphate, a key intermediate in the pathways of isoprenoids and sterols. Deficiency in mevalonate kinase activity has been linked to mevalonic aciduria and hyperimmunoglobulinemia D/periodic fever syndrome (HIDS). The crystal structure of rat mevalonate kinase in complex with MgATP has been determined at 2.4-Å resolution. Each monomer of this dimeric protein is composed of two domains with its active site located at the domain interface. The enzyme-bound ATP adopts an anti conformation, in contrast to the syn conformation reported for Methanococcus jannaschii homoserine kinase. The Mg²⁺ ion is coordinated to both - and -phosphates of ATP and side chains of Glu¹⁹³ and Ser¹⁴⁶. Asp²⁰⁴ is making a salt bridge with Lys¹³, which in turn interacts with the -phosphate. A model of mevalonic acid can be placed near the -phosphoryl group of ATP; thus, the C5 hydroxyl is located within 4 Å from Asp²⁰⁴, Lys¹³, and the -phosphoryl of ATP. This arrangement of residues strongly suggests: 1) Asp²⁰⁴ abstracts the proton from C5 hydroxyl of mevalonate; 2) the penta-coordinated -phosphoryl group may be stabilized by Mg²⁺, Lys¹³, and Glu¹⁹³; and 3) Lys¹³ is likely to influence the pK_a of the C5 hydroxyl of the substrate. V377I and I268T are the most common mutations found in patients with HIDS. Val³⁷⁷ is located over 18 Å away from the active site and a conservative replacement with Ile is unlikely to yield an inactive or unstable protein. Ile-268 is located at the dimer interface, and its Thr substitution may disrupt dimer formation.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Mevalonate kinase (MK, ATP:mevalonate 5-phosphotransferase, EC 2.7.1.36)¹ catalyzes the transfer of the -phosphoryl group from ATP to the C5 hydroxyl oxygen of mevalonic acid to form mevalonate 5-phosphate, a key intermediate in the biosynthetic pathway for isoprenoids and sterols from acetate. Although the enzyme was discovered in the late 1950s (1, 2), it suffered over three decades of neglect, as research on the isoprenoid pathway was focused on HMG-CoA reductase, which catalyzes the previous step in the pathway, i.e. the formation of mevalonic acid from HMG-CoA. However, interest in MK has been revived recently, because it was recognized that this enzyme, together with HMG-CoA synthase and HMG-CoA reductase, is involved in coordinate regulation of this pathway and therefore may represent a secondary control point. The recent recognition of the involvement of the diverse non-sterol isoprenoid metabolites in various cellular functions (e.g. protein prenylation, protein glycosylation, and cell cycle regulation) has also increased interest in this enzyme. In addition, the significance of MK has been further highlighted by the implication of the enzyme in human inherited diseases, such as mevalonic aciduria and hyperimmunoglobulinemia D/periodic fever syndrome (HIDS, Mendelian Inheritance in Man 260920).

The enzyme is found in eukaryotes, archaebacteria, and some eubacteria. The mammalian enzyme is reported to be a homodimer with a subunit molecular mass of 42 kDa (pig (3), rat (4), and human (5)). Kinetic studies suggest that the enzyme catalyzes an ordered sequential reaction, with mevalonic acid binding first to the enzyme and phosphomevalonate the first product released from the enzyme (3). Farnesyl pyrophosphate and geranyl pyrophosphate, both of which are downstream products of the isoprenoid biosynthetic pathway, are potent inhibitors of the enzyme with respect to ATP (6). Site-specific mutagenesis studies have been employed to identify catalytic residues and to investigate the mechanism of the enzyme action. Asp²⁰⁴ has been suggested as the catalytic base that abstracts the proton from the C5-OH group of mevalonic acid (5); Ser¹⁴⁶, located in a conserved glycine-rich region, has been implicated in the binding of Mg-ATP (7); and Lys¹³ has shown to be involved in the binding of ATP as well as facilitating catalysis (8).

Mevalonate kinase is a member of the GHMP kinase family, one of the three distinct sugar kinase families classified by sequence comparison and originally, including galactokinase, homoserine kinase, mevalonate kinase, and phosphomevalonate kinase (9, 10). The list keeps growing, now including mevalonate 5-diphosphate decarboxylase (11), isopentenyl monophosphate kinase (12), and 4-diphosphocytidyl-2C-methylerythritol kinase (13). Sequence comparisons revealed that these kinases share three highly conserved glycine-rich motifs that are implicated in the binding of ATP. The structure of homoserine kinase (HSK) from Methanococcus jannaschii has been solved with and without substrates/analogs (14, 15). The M. jannaschii HSK shares ~20% sequence identity with rat MK but is smaller by about 100 residues than the mevalonate kinase. Similar to other known kinase structures, the HSK structure is composed of two domains, the smaller N-terminal domain and the larger C-terminal domain, and the active site is located at the cleft between the two domains. The bound ATP adopts an unusual syn conformation, and a conserved glycine-rich motif (PXGXGLGSSAA) forms the phosphate-binding loop. However, no strong nucleophile that might act as a catalytic base can be found near the phosphoryl acceptor hydroxyl group. Very recently, the structure of M. jannaschii MK has been reported (16). As expected from sequence homology, the overall polypeptide folding is similar to that of HSK; however, the rare syn ATP-binding mode could not be confirmed, because the structure determined is that of the apo form of the enzyme.

We here report the three-dimensional structure of rat MK in complex with Mg-ATP determined at 2.4-Å resolution. The detailed disposition of the residues involved in the active site, together with previously determined kinetic and mutagenesis studies, allows us to propose a plausible mechanism of the enzyme action. Because the amino acid sequence of the rat enzyme is 82% identical to that of the human enzyme, it is expected that structural conclusions drawn from the rat enzyme will apply to the human enzyme. Furthermore, comparison of the rat MK structure with those of HSK and MK both from M. jannaschii reveals the differences as well as similarities in structure and catalytic mechanism between the two members of the GHMP kinases. In addition, the conclusions derived here are likely to be applicable to other members of the kinase family, including galactokinase and phosphomevalonate kinase.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Purification and Crystallization of Rat MK-- Recombinant rat MK (RMK) was expressed and purified as previously described (8). Briefly, cDNA encoding rat mevalonate kinase was inserted into pET-11a vector, expressed in Escherichia coli, BL21, and purified by chromatography using a Fast-Q anion exchange column followed by a hydrophobic column (phenyl-agarose). Crystals of RMK were grown at 4 °C using the sitting drop method by mixing equal volumes of an RMK solution (13 mg/ml protein, 1 mM ATP, and 2 mM MgCl₂) and a precipitant solution containing 100 mM Hepes buffer, pH 7.5, 17.5% polyethylene glycol-5000 monomethylester. The crystals appeared within 3 days and grew to an approximate size of 0.1 × 0.1 × 0.3 mm in a week. The crystals belong to the orthorhombic space group P2₁2₁2 with cell dimensions a = 77.31, b = 118.64, and c = 42.30 Å. Assuming one monomer per asymmetric unit, the calculated V_m value (17) is 2.4 Å³/Da, corresponding to a solvent content of 54%. The mercury derivative was obtained by cocrystallization with 1 mM thimerosal in the above conditions, and all other derivatives were obtained by soaking the native crystals in an artificial mother liquor solution containing each respective heavy atom compound. A thimerosal cocrystal was used for the multiwavelength anomalous dispersion (MAD) data collection.

Data Collection and Processing-- The MAD data set, one native (NAT1) set, and the osmium derivative set were collected at 180 °C on a synchrotron beamline (SBC-19ID) at the Advanced Photon Source, Argonne National Laboratory. The second native data set (NAT2) and two other derivative sets, mercury and platinum, were obtained at 4 °C on an R-axis IIC imaging plate detector system equipped with Osmic mirrors. For cryo-data collection, crystals were soaked for 2 min in the mother liquor containing 18% ethylene glycol prior to flash freezing in liquid nitrogen. Each data set was obtained with an oscillation angle of 0.5° per frame, and the exposure times were 2 s for the synchrotron data and 10 min for the R-AXIS data. All data were processed with the program package DENZO and SCALEPACK (18). Data collection statistics are given in Table I.

Phasing, Model Building, and Refinement-- The initial phasing was carried out using NAT1 as the native. The heavy atom positions (one mercury, one osmium, and one platinum) were determined using the program SOLVE (19). Difference and anomalous Patterson syntheses confirmed the mercury position, and cross difference Fourier synthesis using the mercury phasing also confirmed the osmium and platinum sites. Phase refinements employing both MIR and MAD techniques were carried out using the program package SHARP (20), yielding an overall figure of merit of 0.52 at 2.3-Å resolution. The resulting phases were further improved by solvent flattening, and at this point, well-defined  helices and  strands were easily identified. The model building was carried out using the interactive graphics program TURBO-FRODO (21). After building the initial model of RMK containing 337 residues (out of a total 395 residues), electron density maps were generated by combining the experimental phases with those of the model by using the SIGMAA algorithm in the CCP4 program package. The resulting map allowed all but 17 residues (16 residues from #73 to #88 and the last C-terminal residue, #395) to be traced. The model was subjected to several rounds of simulated annealing procedure using the CNS package (22) with alternating manual fitting. Despite further rounds of refinement with CNS and manual re-checking of the model using omit maps, the R-values remained at 31% for R_crystal and 36.7% for R_free. At this point, the NAT2 data set, which was collected on an R-AXIS at 4 °C, was introduced as the native and was used in all subsequent analyses. After a few rounds of CNS refinement, the R_crystal and R_free values dropped to 25 and 31%, respectively. At this stage, ATP was clearly visible and in the subsequent cycles, a magnesium ion and water molecules were added. The final model was refined using 2.4-Å resolution data and has an R_crystal of 21.7% and R_free of 26.6%. The C-terminal residue and residues 73-88 were disordered and were not included in the final model. The final refinement statistics are listed in Table II.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Overall Polypeptide Folding-- The overall monomer structure (Fig. 1) is dumbbell-shaped (84 × 45 × 43 Å) with two domains. The N-terminal domain (blue in Fig. 1) is composed of a large, mixed, concave seven-stranded -sheet (1, 4-8, and 15) packed with seven helices (2-7 and 13) and a small, antiparallel, four-stranded -sheet (2, 3, 9, and 10) extending upward from the large seven-stranded sheet. The convex side of the large sheet is mostly exposed to the solvent and covered only by a short C-terminal helix (13). Residues 73-88, located between the 2 and 3 helices, are disordered and lie at the surface of the molecule. The C-terminal domain (green in Fig. 1A) is composed of another antiparallel, four-stranded -sheet, facing the N-terminal small sheet on one side, and five long  helices (8-12) crowning the solvent side of the monomer. Thus, the two antiparallel -sheets make up the interface of the two domains. The interface area is about 1150Ų per domain and is about 17% of the total surface of the smaller C-terminal domain. Because of this relatively tight interface and because the polypeptide traverses twice between the domains, the linkage between the two domains (the dumbbell handle) appears to be rather rigid. Therefore, the molecule is less likely to undergo a large conformational change (i.e. to bend the dumbbell handle) to allow the cleft between the two domains to change in shape and size. This is different from the case of the catalytic domain of cAMP-dependent protein kinase, in which its cleft is open or closed depending on whether the substrate is bound or not (23). However, the structure of the catalytic domain of cAMP-dependent protein kinase is also bi-lobal with a deep cleft between the lobes, which are connected by a flexible linker.

View larger version (52K): [in this window] [in a new window]   Fig. 1.   The structure of RMK in complex with MgATP. A, structure of a monomer. The N-terminal domain is shown in blue, and the C-terminal domain is in green. Helices are numbered 1 through 13 and -strand labeled 1 through 14. MgATP is depicted with a ball-and-stick model (atom-colored), and Val377 located at the beginning of the 15 strand is represented by a large red ball. Both N and C termini are indicated, and the disordered region (residue 72-89) is shown with a dotted line. B, structure of a dimer is shown. The view for the green monomer is 90° rotated along the y axis from the view shown in panel A. The arrow indicates the molecular 2-fold axis. The graphics were generated by using Molscript (35) and rendered with Raster3D (36).

Dimer Structure-- The enzyme is reported to be a homodimer in solution, with no unusual hydrodynamic properties that would indicate its asymmetric shape (5, 6). In the crystal structure, however, the 2-fold axis is located at the tip of the N-terminal domain (Fig. 1B), making the dimer molecule extremely elongated, with its longest and shortest dimensions of 120 and 40 Å, respectively. The dimer interface area is 2450 Å² per dimer, which is only 7.6% of the total surface area. It is composed of residues from two  helices, residues 259-270 (helix 9) and residues 300-308 (helix 10) of each monomer, making a classic four-helix bundle with hydrophobic side chains interdigitating at the core of the bundle. It is unlikely that these hydrophobic interactions are due to the crystal packing. Therefore, the dimer structure observed in the crystalline state is most likely the same as that which exists in solution. However, rigorous hydrodynamic characterizations of the enzyme are necessary to confirm this. It is interesting to note that the crystalline M. jannaschii MK protein is a monomer, whereas the protein is reported to be a dimer in solution (16). Thus, it is not surprising that a majority of the hydrophobic residues involved in the dimer interface in the rat and human MK proteins are not conserved in M. jannaschii MK; most of them are replaced by hydrophilic residues.

Comparison with the Structure of Homoserine Kinase-- The overall polypeptide folding of rat MK (RMK) is very similar to that of homoserine kinase (HSK) from M. jannaschii (Zhou et al. (14)), as predicted from their sequence homology. However, there are some significant differences between the two structures. First of all, RMK (395 residues) is significantly larger than HSK (300 residues), and consequently there are many additional parts in the rat MK structure. There are three large insertions, all of them lying in the N-terminal domain. The largest one (about 50 residues) lies between 5 and 6, forming 6-2-loop-3-loop. This insertion is situated at the surface of the molecule (the concave side of the seven-stranded -sheet) and includes the disordered region, residues 73-88. The next insertion is a long excursion, including 8, lying between 5 and 6. Both 6 and 8 are parts of the concave -sheet, expanding it to a seven-stranded sheet (one strand on each side of the sheet), as compared with the five-stranded sheet found in the structure of HSK. The third insertion is the C-terminal tail (residues 381-395) lying on the convex side of the -sheet. Because all of these insertions lie at the surface of the molecule, they are probably not directly involved in either the catalytic function or the dimer formation of the enzyme. However, these insertions must play a role in the stability of the RMK protein. Excluding these three regions and other minor loop areas, the root mean square deviation (r.m.s.d.) of the main-chain atoms between the two structures is 4.1 Å for 265 residues. However, the r.m.s.d. values for the individual domains are much smaller, 1.1 Å for the N-terminal domain (105 residues) and 1.2 Å for the C-terminal domain (160 residues), suggesting that the relative orientations of the two domains in the two structures are different. Fig. 2 shows an overlay of the two structures using only the two N-terminal domains for superposition. The RMK structure is straighter, whereas that of HSK is bent about 15° from the vertical axis, making its active-site cleft, located at the interface between the two domains, more closed than that of the RMK structure. Both structures are of binary complexes with the nucleotide substrate/product, RMK with ATP and HSK with ADP. Unlike the M. jannaschii MK protein, which is a monomer in the crystalline state, the HSK protein is also a dimer. The dimer interface in the HSK protein involves -strands in addition to the corresponding helices that are involved in the rat MK dimer formation. Consequently, the overall dimer shape in the HSK structure is less elongated than that of the RMK structure. In addition to these differences in the overall structure, a few minor, but functionally significant insertions/deletions exist between the two structures, as discussed below. Using the structures of both RMK and HSK, a structure-based sequence alignment of representative members of the GHMP family has been constructed (Fig. 3).

View larger version (59K): [in this window] [in a new window]   Fig. 2.   An overlay of the structures of rat MK and M. jannaschii HSK using only their N-terminal domain superimposed. MK with its ATP is shown in dark gray and HSK with its ADP in light gray. The C-terminal domain of MK sits straight on "top" of its N-terminal domain, whereas the N-terminal domain of HSK is bent about 15° from the vertical axis depicted by a solid line. The N-terminal domain in the MK structure has extra regions that are lacking in HSK, one region on the "left " side and another on the "right " side of the core.

View larger version (78K): [in this window] [in a new window]   Fig. 3.   Structure-based sequence alignment of rat MK and M. jannaschii HSK together with other representative members of the GHMP family. MK, mevalonate kinase; GAL, galactokinase; PMK, phosphomevalonate kinase; HSK, homoserine kinase, RAT, Rattus norvegicus; HUM, Homo sapiens; AF, Archaeoglobus fulgidus; MJ, M. jannaschii; YE, S. cerevisiae. The invariant residues are highlighted in dark gray and mostly conserved residues are in light gray. The secondary structural elements in the rat MK structure are indicated above the sequence: -strands with arrows and -helices with cylinders; the N-terminal domain is in dark gray and the C-terminal domain is in light gray. Residues involved in ATP binding in the rat MK structure are marked with gray dots; those implicated in mevalonic aciduria or HIDS are marked with #.

ATP Binding Site-- Crystals of RMK were obtained in the presence of ATP and MgCl₂. A difference Fourier map calculated at a later stage of the refinement yielded extra electron densities at the cleft between the two domains. These densities can be fitted with an ATP molecule and a magnesium ion (Fig. 4). The nucleotide in the enzyme-bound ATP has an anti conformation. In contrast, in the structure of HSK complexed with either ADP or ATP analogs, the nucleotide adopts a syn conformation (14). As in the case of the HSK structure, the enzyme-bound ATP in the RMK structure also appears to have multiple conformations at its triphosphate moiety. Although the electron density for the adenosine portion of the bound nucleotide is well defined, that of the triphosphate moiety was best fitted with two different conformations: a major (estimated occupancy, 70%, shown as thick sticks in Fig. 4) and a minor conformation (30% occupancy, thin sticks in Fig. 4). For clarity and simplicity, in the following discussion the major conformation of ATP is used, unless noted otherwise. Fig. 5 shows residues involved in the nucleotide binding. Both the amide carbonyl group of Asn-55 and the hydroxyl group of Ser¹³⁵ are hydrogen-bonded to the amino group of the adenine ring, and the 2'-OH and 3'-OH of the ribose moiety are bound to Ser¹⁰⁸ and Asn¹⁰⁴, respectively. The magnesium-triphosphate moiety is surrounded by a glycine-rich loop, Motif 2 (residues ¹⁴⁰GAGLGSSAA¹⁴⁸, see Fig. 3), and the magnesium ion is coordinated to both - and - (- in the minor conformer) phosphates. It is also coordinated to the carboxyl of Glu¹⁹³ and the hydroxyl group of Ser¹⁴⁶, both of which are hydrogen-bonded to each other. The hydroxyl group of Ser¹⁴⁶ is also hydrogen-bonded to -PO₄ in the minor conformer and is about 3.7 Å away from a -phosphate oxygen, suggesting that they might form a hydrogen bond in solution. Of particular interest is the interaction between Lys¹³ and Asp²⁰⁴, making a salt bridge. Lys¹³ also makes a close interaction with the -PO₄ of ATP. This is consistent with the results of the pyridoxal phosphate inactivation of the RMK and its protection by ATP (8). It is also consistent with the mutagenesis studies by same authors, in which the replacement of Lys¹³ with methionine (K13M) diminished its K_d for ATP by 56-fold. In addition to binding and stabilizing the -phosphate group, it is reasonable to assume that Lys¹³ may contribute to influencing the pK_a value of Asp²⁰⁴ by making a salt bridge. It is interesting to note that, in the M. jannaschii MK structure in its apo form, no corresponding interaction between Asp²⁰⁴ and Lys¹³ (Asp¹⁵⁵ and Lys⁸, respectively, in M. jannaschii MK) is apparent (16). However, because the current structure of rat MK is that of a complex with MgATP, it is conceivable that some side-chain conformations (including Lys¹³ and Asp²⁰⁴) and/or minor local structural movements might occur upon binding of a substrate, MgATP.

View larger version (56K): [in this window] [in a new window]   Fig. 4.   Electron density map of the enzyme-bound MgATP. An omit map was calculated and contoured at the 2.5- level. The magnesium ion is shown with a large ball. The triphosphate moiety was best fitted in two different conformations; a major conformation (estimated relative occupancy, 70%) is shown with a thick gray ball-and-stick model and a minor conformation (estimated relative occupancy, 30%) with thin sticks.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Vicinity of the ATP binding site. The magnesium ion (large gray ball) is coordinated to both - and -PO4. It is also coordinated to the side chains of Glu193 and Ser146, which form a hydrogen bond with each other. Lys13 is bound to the -phosphate and makes a salt bridge with Asp204. The Motif 2 glycine-rich loop (gray twisted coil), containing residues 140GAGLGSSAA148, wraps around the phosphate moiety. Hydrogen bonds are indicated by dotted lines, and coordination bonds are represented by dashed lines.

The Putative Mevalonate Binding Site-- At the junction between the two domains of the RMK structure exists a large cavity, which is adjacent to the binding site of the -phosphoryl group of the bound ATP. This cavity is surrounded by residues conserved in various MK proteins, and these residues are also mostly conserved in other GHMP family members (Fig. 3). We have modeled a mevalonic acid molecule in the cavity such that the C5 hydroxyl group lies close to and in line with the -phosphoryl group of ATP (Fig. 6). The carboxyl of the substrate can be placed near Arg²⁴¹, Thr²⁴³, and Ala³³⁴. Arg²⁴¹ makes a salt bridge with the carboxyl group, thus stabilizing the binding of the substrate. Both Ala³³⁴ and Thr²⁴³ are implicated in mevalonic aciduria (24, 25). The A334T mutation has been identified in patients with MK deficiency. Kinetic characterization indicated that the mutant protein exhibits about 1.4% of wild type activity and >30-fold increase in K_m for mevalonate, consistent with Ala³³⁴ involvement in mevalonate binding. Ala³³⁴ is located in the third glycine-rich loop (Motif 3, the loop between 12 and 13, see Fig. 3), and in our model the main-chain amide nitrogen of Ala³³⁴ makes a hydrogen bond with the carboxyl group of the substrate. Replacement of Ala³³⁴ with Thr makes the binding cavity crowded, where the Motif 3 loop is making a tight turn. The T243I mutant found in patients with mevalonic aciduria exhibits less than 2% of the wild type activity (25). Even a conservative substitution of an alanine for Thr²⁴³ (7) resulted in a moderately significant increase in the apparent K_m values (~20-fold) for mevalonic acid, but a relatively minimal decrease in V_max (50% of the wild type value). In this model of mevalonic acid (Fig. 6), the C3 hydroxyl group of the substrate is situated within a hydrogen bonding distance from the main-chain carbonyl oxygen of Gly¹⁴², which is located in the phosphate-binding loop (Motif 2). On the other hand, His²⁰ lines the active site cavity, and the H20P mutation is implicated in mevalonic aciduria (26) and HIDS (27). Houten et al. (26) reported that the H20P-GST fusion protein expressed in E. coli had no detectable MK activity, although its expression level was comparable to the control. However, the MK protein level in fibroblast lysate from a patient with the H20P mutation was significantly reduced, suggesting that the H20P protein is unstable. His²⁰ is located at the "roof" of the active site cavity and is about 8 Å away from the closest atom of the modeled mevalonic acid (carboxylate atoms). This residue is also in the ¹⁸GEHA²¹ loop in Motif 1, which forms part of the cavity entrance. Replacement of His²⁰ with proline would alter this loop conformation, which would in turn change the cavity size and/or shape, resulting in an inactive or less active enzyme molecule. In addition, the preceding residue Glu¹⁹ forms a salt bridge with Lys³³⁰, which lies at the beginning of the Motif 3 loop, thus stabilizing the interface between the domains of the RMK structure (part of this interface also forms the "back wall" of the cavity). Therefore, a conformational change in the Motif 1 loop, caused by the proline substitution, would disrupt the salt bridge, thereby yielding an unstable protein. In the structure of the RMK·MgATP binary complex, the active site cavity is too large to fit a mevalonic acid molecule snugly and, therefore, the bound mevalonate is rather exposed to the solvent; such an environment do not seem optimized for efficient catalysis. Two ways can be envisioned to change the size of the binding cavity and to shield the bound substrate from the solvent. As stated earlier, the interface between the two domains in the MK structure is relatively rigid, and therefore, a large structural rearrangement (such as the entire domain rotation observed in the structure of the catalytic domain of cAMP-dependent protein kinase) is unlikely to occur. On the other hand, it is entirely possible that the MK molecule undergoes local conformational changes upon binding of the mevalonate substrate. Side-chain reorientations and loop movements near the active site may result in tightening of the binding cavity. In the structure of HSK, the "upper lip" helix (residues 181-189) moves down closer to the substrate, and the side chain of Arg¹⁸⁷ flips almost 180° making a salt bridge with Asp²³, thus completely shielding the substrate binding cavity (15). The corresponding segment in the MK structure is the beginning of the 8 helix and its preceding loop (residues 238-246), including Arg²⁴¹ and Thr²⁴³. Therefore, in the MK structure, it is conceivable that similar local conformational changes would occur when the enzyme forms the ternary complex. Then the active site would become tighter, and additional close interactions could be made between the bound mevalonate and the polypeptide chain. Therefore, it is also possible that both the Motif 1 loop (containing His²⁰) and the Motif 3 loop (containing Ala³³⁴) become closer and interact with the carboxyl group of the substrate. In our model of the enzyme-bound mevalonic acid, the C5 hydroxyl group is located within 4 Å from the carboxyl group of Asp²⁰⁴, -amino group of Lys¹³, and the -phosphoryl group of ATP. This constellation of residues is consistent with our previously proposed mechanism of the MK reaction (7). In addition, the current structure allows us to refine the mechanism and assign the roles of the residues previously studied by site-specific mutagenesis, including Asp²⁰⁴, Lys¹³, Ser¹⁴⁶, and Glu¹⁹³.

View larger version (64K): [in this window] [in a new window]   Fig. 6.   The putative mevalonate binding site. A, a ribbon diagram of the RMK structure showing the orientation of the molecule. The two domains are shown in blue and green as in Fig. 1A, and the segment (7-loop-5) containing the ATP-binding loop (Motif 2) is shown in light blue. B, molecular surface of the binding site for the area within the red square in panel A. A mevalonic acid molecule (MEV) is modeled in a cavity adjacent to the -phosphoryl group of the bound ATP. Residues in the active site are shown with ball-and-stick models, and the molecular surface is rendered semitransparent. The ATP binding loop (Motif 2) is depicted with a light blue coil. The electrostatic potential is mapped onto the molecular surface. Surface color codes are blue for positive (10 kbT, where kb is Boltzmann constant and T is absolute temperature), red for negative (10 kbT), and white for neutral. This graphic was generated by using GRASP (37).

Catalytic Mechanism-- Fig. 7 shows a schematic of a possible mechanism of mevalonate kinase, based on the structure reported here and on previous mutagenesis studies. In the rat MK structure, Asp²⁰⁴ makes a salt bridge with Lys¹³, which in turn is making a close interaction with the -phosphate group of ATP. The C5-OH group of the modeled mevalonate is within about 4 Å from both the aspartate and the -phosphate group, suitably situated to donate its proton to the aspartate and to accept the phosphoryl group from ATP. This arrangement is very similar to that of the corresponding groups in the structure of phosphofructokinase (28) and hexokinase (29), in which an aspartic acid functions as a catalytic base to abstract the proton from the acceptor hydroxyl of the sugar molecule. Mutation of Asp²⁰⁴ in human MK decreased the activity by 10⁴-fold compared with wild type, consistent with its role as the catalytic base (5). The penta-coordinated -phosphate transition state may be stabilized by the magnesium ion, the side chains of Glu¹⁹³ and Lys¹³, and the main-chain carbonyl group of Gly¹⁴². Both Glu¹⁹³ and Ser¹⁴⁶ stabilize the magnesium ion position. In our model, Lys¹³ is also about 3-4 Å away from the C5 hydroxyl group. The pK_a of the C5 hydroxyl of mevalonic acid in free solution is estimated to be ~13 (30). Thus, the pK_a of the enzyme-bound mevalonic acid must be lowered by at least several pH units for an efficient catalysis to occur. The lowering of the pK_a may partially come from its proximity to Lys¹³, which stabilizes the deprotonated C5 alkoxide group. In the studies with rat MK, elimination of the basic group, K13M, resulted in a ~60-fold diminution in the V_max value, an observation compatible with the influence of Lys¹³ on the pK_a of the C5 hydroxyl group. Both Asp²⁰⁴ and Lys¹³ are conserved in other members of the GHMP family, including galactokinase and phosphomevalonate kinase (in galactokinase, an arginine replaces Lys¹³). Therefore, it is reasonable to assume that the reaction mechanism of these two members of the GHMP is similar to that of mevalonate kinase. An aspartate residue functions as the catalytic base and a basic residue (Lys or Arg) near the acceptor hydroxyl group lowers its pK_a to facilitate the proton abstraction. In contrast, the corresponding residues in the homoserine kinase structure are Asn (Asn²⁴¹) for Asp²⁰⁴ and Thr (or Ser in the yeast enzyme) for Lys¹³. It has been suggested that HSK does not involve a catalytic base for activating the acceptor hydroxyl group, and instead the -phosphoryl group is directly involved in the proton abstraction, as in the case of nucleotide monophosphate kinase and fructose-6-phosphate 2-kinase (15).

View larger version (21K): [in this window] [in a new window]   Fig. 7.   A schematic of the proposed catalytic mechanism of mevalonate kinase. The carboxyl group of Asp204 functions as the catalytic base that abstracts the proton from the C5 hydroxyl of mevalonate. Lys13, Glu193, and Ser146, are involved in stabilization of the transition state of the penta-coordinated -phosphoryl group of ATP. Lys13 may also influence the pKa of the C5 hydroxyl of mevalonate. All of these interactions facilitate transfer of the -PO4 group of ATP to form mevalonate 5-phosphate and ADP. Hydrogen bonds are shown with thin dotted lines; coordination bonds, thick dotted lines; possible bonds with modeled mevalonate, dashed line.

V377I and I268T versus HIDS-- A number of residues implicated in mevalonic aciduria are located at or near the active site of the molecule. H20P introduces straining of the local secondary structure in the Motif 1 loop; the T243I mutation results in a hydrophobic residue in place of Thr²⁴³, which is likely to make a hydrogen bond with the C3 hydroxyl of the mevalonate substrate; A344T introduces a bulkier side chain causing a steric hindrance at the active site. Among other mutations that correlate with MK deficiency, V377I and I268T are of particular interest. Cuisset et al. (27) reported that V377I is the most common mutation among patients suffering with HIDS. Although genetic studies indicate that most HIDS patients are compound heterozygotes, in a few cases cDNA analysis detects only a single mutation corresponding to V377I. Drenth et al. (31) point out that, in many such instances, examination of genomic DNA suggests heterozygosity. Thus, an undetected mutation may result in inhibition of gene expression, mRNA instability, or poor mRNA amplification by reverse transcription-PCR. Indeed, although cDNA analysis (27) of several patients suggested homozygosity for the V377I substitution, more detailed scrutiny of parent DNA samples ruled out such an assignment, because one was heterozygous for the V377I substitution while the other was homozygous for Val³⁷⁷ (i.e. wild type). There has been a recent claim (32) of an HIDS case in which homozygosity for the V377I allele has been confirmed by analysis of both patient and parental genomic DNA. On this basis, the V377I substitution has been proposed to account for the HIDS phenotype. Such a proposal must be balanced by the observation that, in other eukaryotes, Val³⁷⁷ is replaced by cysteine (Saccharomyces cerevisiae, also see Fig. 3), alanine (Arabadopsis thaliana), or leucine (Drosophila melanogaster). In addition, the MK protein in a variety of prokaryotes (including the M. jannaschii protein, which has been functionally expressed and characterized (33)) terminates prior to any residue that would correspond to Val³⁷⁷. Val³⁷⁷ is located on the sixth strand (15) of the seven-stranded -sheet in the N-terminal domain and is about 18 Å away from the -phosphoryl group (Fig. 1), indicating that Val³⁷⁷ is not directly involved in the catalytic action of the enzyme nor in substrate binding. Its side chain points toward the concave side of the -sheet, surrounded by hydrophobic residues and protected from the solvent by residues in the 6 helix. We have modeled in an isoleucine residue in place of valine at this position, and there appears to be enough room to accommodate the extra methylene group of the isoleucine side chain. This, together with the fact that both valine and isoleucine are hydrophobic, suggests that the V377I mutation is not detrimental to the protein folding. These are consistent with the results of the in vitro experiments, in which a recombinant human V377I protein was characterized (34). The mutant protein exhibits only a modest decrease in enzyme activity (<20%) compared with the wild type enzyme, in contrast to the drastic decrease of activity (>95%) observed with cultured lymphocytes of HIDS patients with the V377I mutation. Therefore, the V377I mutation is unlikely to be the major reason for the depressed catalytic activity and/or stability of mevalonate kinase observed in HIDS patients. I268T is the second most common mutation among HIDS patients. Ile²⁶⁸ is located in the 9 helix, which together with 10 forms a four-helix bundle at the dimer interface of the molecule. It is conceivable that a substitution of Thr (hydrophilic residue) for Ile (hydrophobic residue) may weaken the dimerization interaction, thus, making the mutant protein somewhat less stable. This is consistent with the studies of Hinson et al. (25). They reported that <50% of recombinant I268T protein in comparison with wild type MK was detectable by immunoblot analysis. Thus, the reduced MK activity and the protein instability observed in HIDS patients is most likely due partially to the Ile²⁶⁸ mutation and not due to the V377I mutation.

In conclusion, the structure of rat MK in complex with MgATP presented here reveals overall polypeptide folding similar to those of MK and HSK (both from M. jannaschii), confirming the structural similarity among members of the GHMP kinase family. However, the nucleotide binding mode in RMK is in an anti conformation in contrast to the syn conformation observed in the structure of HSK. Other notable contrasts with HSK include MK's salt bridge between Lys¹³ and Asp²⁰⁴, as well as the positioning of these residues to support: 1) a function for Lys¹³ in influencing ionization of substrate mevalonate and the Asp²⁰⁴ side chain carboxyl and 2) a function for Asp²⁰⁴ in general base catalysis of phosphoryl transfer.

	ACKNOWLEDGEMENTS

We thank Advanced Photon Source personnel, especially Dr. Sergey Korolev, for assistance in MAD data collection.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grant DK53766. Use of the Argonne National Laboratory Structural Biology Center beamlines at the Advanced Photon Source was supported by the U.S. Department of Energy, Office of Biological and Environmental Research, under Contract W-31-109-ENG-38.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 1KVK) 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 should be addressed. Tel.: 414-456-8479; Fax: 414-456-6510; E-mail: jjkim@mcw.edu.

Published, JBC Papers in Press, February 27, 2002, DOI 10.1074/jbc.M200912200

	ABBREVIATIONS

The abbreviations used are: MK, mevalonate kinase; RMK, rat mevalonate kinase; HSK, homoserine kinase; HID, hyperimmunoglobulinemia; HIDS, hyperimmunoglobulinemia/periodic fever syndrome; HMG-CoA, 3-hydroxy-3-methylglutaryl-coenzyme A reductase; GHMP, galactokinase, homoserine kinase, mevalonate kinase, phosphomevalonate kinase; MAD, multiwavelength anomalous dispersion; r.m.s.d., root mean square deviation.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				J. L. Andreassi II, P. W. Bilder, M. W. Vetting, S. L. Roderick, and T. S. Leyh Crystal structure of the Streptococcus pneumoniae mevalonate kinase in complex with diphosphomevalonate Protein Sci., May 1, 2007; 16(5): 983 - 989. [Abstract] [Full Text] [PDF]

				J. B. Thoden and H. M. Holden The Molecular Architecture of Human N-Acetylgalactosamine Kinase J. Biol. Chem., September 23, 2005; 280(38): 32784 - 32791. [Abstract] [Full Text] [PDF]

				S. S. Doun, J. W. Burgner II, S. D. Briggs, and V. W. Rodwell Enterococcus faecalis phosphomevalonate kinase Protein Sci., May 1, 2005; 14(5): 1134 - 1139. [Abstract] [Full Text] [PDF]

				J. B. Thoden, D. J. Timson, R. J. Reece, and H. M. Holden Molecular Structure of Human Galactokinase: IMPLICATIONS FOR TYPE II GALACTOSEMIA J. Biol. Chem., March 11, 2005; 280(10): 9662 - 9670. [Abstract] [Full Text] [PDF]

				L. Wang and K. M. J. Menon Regulation of Luteinizing Hormone/Chorionic Gonadotropin Receptor Messenger Ribonucleic Acid Expression in the Rat Ovary: Relationship to Cholesterol Metabolism Endocrinology, January 1, 2005; 146(1): 423 - 431. [Abstract] [Full Text] [PDF]

				D. Krepkiy and H. M. Miziorko Identification of active site residues in mevalonate diphosphate decarboxylase: Implications for a family of phosphotransferases Protein Sci., July 1, 2004; 13(7): 1875 - 1881. [Abstract] [Full Text] [PDF]