The Crystal Structure of Human Geranylgeranyl Pyrophosphate Synthase Reveals a Novel Hexameric Arrangement and Inhibitory Product Binding
2006; Elsevier BV; Volume: 281; Issue: 31 Linguagem: Inglês
10.1074/jbc.m602603200
ISSN1083-351X
AutoresK.L. Kavanagh, James E. Dunford, G. Bunkóczi, R.G.G. Russell, Udo Oppermann,
Tópico(s)ATP Synthase and ATPases Research
ResumoModification of GTPases with isoprenoid molecules derived from geranylgeranyl pyrophosphate or farnesyl pyrophosphate is an essential requisite for cellular signaling pathways. The synthesis of these isoprenoids proceeds in mammals through the mevalonate pathway, and the final steps in the synthesis are catalyzed by the related enzymes farnesyl pyrophosphate synthase and geranylgeranyl pyrophosphate synthase. Both enzymes play crucial roles in cell survival, and inhibition of farnesyl pyrophosphate synthase by nitrogen-containing bisphosphonates is an established concept in the treatment of bone disorders such as osteoporosis or certain forms of cancer in bone. Here we report the crystal structure of human geranylgeranyl pyrophosphate synthase, the first mammalian ortholog to have its x-ray structure determined. It reveals that three dimers join together to form a propeller-bladed hexameric molecule with a mass of ∼200 kDa. Structure-based sequence alignments predict this quaternary structure to be restricted to mammalian and insect orthologs, whereas fungal, bacterial, archaeal, and plant forms exhibit the dimeric organization also observed in farnesyl pyrophosphate synthase. Geranylgeranyl pyrophosphate derived from heterologous bacterial expression is tightly bound in a cavity distinct from the chain elongation site described for farnesyl pyrophosphate synthase. The structure most likely represents an inhibitory complex, which is further corroborated by steady-state kinetics, suggesting a possible feedback mechanism for regulating enzyme activity. Structural comparisons between members of this enzyme class give deeper insights into conserved features important for catalysis. Modification of GTPases with isoprenoid molecules derived from geranylgeranyl pyrophosphate or farnesyl pyrophosphate is an essential requisite for cellular signaling pathways. The synthesis of these isoprenoids proceeds in mammals through the mevalonate pathway, and the final steps in the synthesis are catalyzed by the related enzymes farnesyl pyrophosphate synthase and geranylgeranyl pyrophosphate synthase. Both enzymes play crucial roles in cell survival, and inhibition of farnesyl pyrophosphate synthase by nitrogen-containing bisphosphonates is an established concept in the treatment of bone disorders such as osteoporosis or certain forms of cancer in bone. Here we report the crystal structure of human geranylgeranyl pyrophosphate synthase, the first mammalian ortholog to have its x-ray structure determined. It reveals that three dimers join together to form a propeller-bladed hexameric molecule with a mass of ∼200 kDa. Structure-based sequence alignments predict this quaternary structure to be restricted to mammalian and insect orthologs, whereas fungal, bacterial, archaeal, and plant forms exhibit the dimeric organization also observed in farnesyl pyrophosphate synthase. Geranylgeranyl pyrophosphate derived from heterologous bacterial expression is tightly bound in a cavity distinct from the chain elongation site described for farnesyl pyrophosphate synthase. The structure most likely represents an inhibitory complex, which is further corroborated by steady-state kinetics, suggesting a possible feedback mechanism for regulating enzyme activity. Structural comparisons between members of this enzyme class give deeper insights into conserved features important for catalysis. Synthesis of isoprenoids is intrinsic to all organisms and leads to a vast array of metabolites with diverse functions. In humans and other mammals, the products of this pathway include essential molecules such as cholesterol, heme A, ubiquinone, dolichol, and farnesoids (Fig. 1A). The latter products include farnesyl pyrophosphate (FPP) 3The abbreviations used are: FPP, farnesyl pyrophosphate; DMAPP, dimethylallyl pyrophosphate; FPPS, farnesyl pyrophosphate synthase; GGPP, geranylgeranyl pyrophosphate; GGPS, geranylgeranyl pyrophosphate synthase; GPP, geranyl pyrophosphate; IPP, isopentenyl pyrophosphate; N-BP, nitrogen-containing bisphosphonate; NCS, noncrystallographic symmetry; TCEP, tris(2-carboxyethyl)phosphine; TEV, tobacco etch virus; PDB, Protein Data Bank; PEG, polyethylene glycol. and geranylgeranyl pyrophosphate (GGPP), which are precursors for protein prenylation and might serve as nuclear receptor ligands for the receptors farnesoid X receptor or liver X receptor (1Murthy S. Tong H. Hohl R.J. J. Biol. Chem. 2005; 280: 41793-41804Abstract Full Text Full Text PDF PubMed Scopus (30) Google Scholar, 2Niesor E.J. Flach J. Lopes-Antoni I. Perez A. Bentzen C.L. Curr. Pharm. Des. 2001; 7: 231-259Crossref PubMed Scopus (51) Google Scholar). The post-transcriptional modification of proteins with isoprenoids consists of farnesylation and geranylgeranylation of proteins with a C-terminal CaaX motif (where a is any aliphatic residue) by protein prenyltransferases (3Sinensky M. Biochim. Biophys. Acta. 2000; 1529: 203-209Crossref PubMed Scopus (62) Google Scholar, 4Sinensky M. Biochim. Biophys. Acta. 2000; 1484: 93-106Crossref PubMed Scopus (208) Google Scholar). Typical examples of prenylated proteins are small GTPases such as Ras, which is farnesylated, and the Rho family of GTPases, which is geranylgeranylated (5Coxon F.P. Ebetino F.H. Mules E.H. Seabra M.C. McKenna C.E. Rogers M.J. Bone. 2005; 37: 349-358Crossref PubMed Scopus (92) Google Scholar, 6Coxon F.P. Helfrich M.H. Van't Hof R. Sebti S. Ralston S.H. Hamilton A. Rogers M.J. J. Bone Miner. Res. 2000; 15: 1467-1476Crossref PubMed Scopus (352) Google Scholar, 7Coxon F.P. Rogers M.J. Calcif. Tissue Int. 2003; 72: 80-84Crossref PubMed Scopus (84) Google Scholar). Prenylation has been shown to be crucial to the targeting and activity of GTPases that are involved in cell growth and survival, motility, cytoskeletal regulation, intracellular transport, and secretion (8Coxon F.P. Helfrich M.H. Larijani B. Muzylak M. Dunford J.E. Marshall D. McKinnon A.D. Nesbitt S.A. Horton M.A. Seabra M.C. Ebetino F.H. Rogers M.J. J. Biol. Chem. 2001; 276: 48213-48222Abstract Full Text Full Text PDF PubMed Scopus (152) Google Scholar, 9Molnar G. Dagher M.C. Geiszt M. Settleman J. Ligeti E. Biochemistry. 2001; 40: 10542-10549Crossref PubMed Scopus (64) Google Scholar). In mammals, as in most eukaryotes, isoprenoid synthesis proceeds through the mevalonate pathway starting from acetyl-CoA with the intermediates hydroxymethylglutaryl-CoA, mevalonate, isopentenyl pyrophosphate (IPP), dimethylallyl pyrophosphate (DMAPP), and FPP (Fig. 1A). Farnesyl pyrophosphate synthase (FPPS) resides at a key branch point of the pathway, because it produces precursors for all isoprenoids. Several enzymes in the pathway constitute important and well established drug targets, for example statins are used to lower cholesterol levels by inhibiting the rate-limiting enzyme in the pathway, hydroxymethylglutaryl-CoA reductase. Another class of drugs in clinical use are the nitrogen-containing bisphosphonates (N-BPs) that inhibit FPPS, used to treat disorders characterized by bone resorption such as osteoporosis, Paget disease, or multiple myeloma (10Russell R.G. Rogers M.J. Bone. 1999; 25: 97-106Crossref PubMed Scopus (768) Google Scholar). Further targets for drug development presently being explored are the protein prenyltransferases for the treatment of cancer (11Basso A.D. Kirschmeier P. Bishop W.R. J. Lipid Res. 2006; 47: 15-31Abstract Full Text Full Text PDF PubMed Scopus (268) Google Scholar) or FPPS from protozoan parasites for the treatment of malaria, Leishmaniasis, and Chagas disease (12Cheng F. Oldfield E. J. Med. Chem. 2004; 47: 5149-5158Crossref PubMed Scopus (51) Google Scholar, 13Garzoni L.R. Waghabi M.C. Baptista M.M. de Castro S.L. de Meirelles Mde N. Britto C.C. Docampo R. Oldfield E. Urbina J.A. Int. J. Antimicrob. Agents. 2004; 23: 286-290Crossref PubMed Scopus (96) Google Scholar). We recently determined the structure of human FPPS and were able to deduce the molecular mechanism of N-BP inhibition (14Kavanagh K.L. Guo K. Dunford J.E. Wu X. Knapp S. Ebetino F.H. Rogers M.J. Russell R.G.G. Oppermann U. Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 7829-7834Crossref PubMed Scopus (451) Google Scholar). In this study we describe the structure of human geranylgeranyl pyrophosphate synthase (GGPS), the enzyme producing the isoprenoid molecule essential for geranylgeranylation of proteins. The enzyme is a potential drug target for oncology and bone disorders. GGPS predominantly catalyzes the condensation of IPP with FPP to obtain the C20 product GGPP, although it can utilize DMAPP or GPP as alternate allylic substrates (Fig. 1B). The enzyme is 17% identical to human FPPS, sharing key consensus regions and possibly a common catalytic mechanism. Currently, structural information is only available for one archaeal (PDB code 1WY0), one fungal (PDB code 2DH4), and one bacterial ortholog (PDB code 1WMW) sharing ∼20-44% sequence identity and displaying different quaternary structures. Cloning, Expression, and Purification of Human GGPS—A clone encoding human GGPS encompassing residues 1-300 (derived from clone accession number gi 4758430) as an N-terminally His6-tagged fusion protein with a TEV protease cleavage site was expressed in Escherichia coli BL21(DE3). In brief, 10 ml of overnight culture were used to inoculate 1 liter of Terrific Broth containing 100 μg/ml kanamycin. Cells were grown at 37 °C to an A600 of 1 and were then cooled to 18 °C before being induced with 0.5 mm isopropyl 1-thio-β-d-galactopyranoside and cultured overnight. Cells were harvested by centrifugation, and the pellet was resuspended in 20 ml of binding buffer (500 mm NaCl, 5% glycerol, 50 mm HEPES, pH 7.5, 5 mm imidazole, 0.5 mm TCEP) with protease inhibitors (Complete, Roche Applied Science), followed by lysis using a high pressure cell disrupter. The lysate was cleared by centrifugation before applying to a pre-equilibrated nickel-nitrilotriacetic acid (Qiagen) column with a 3-ml bed volume. The column was washed with 20 column volumes of binding buffer, 10 column volumes of wash buffer (500 mm NaCl, 5% glycerol, 50 mm HEPES, pH 7.5, 30 mm imidazole, 0.5 mm TCEP), and eluted in 12 ml of the same buffer containing 250 mm imidazole. The hexahistidine tag was removed by incubation with His-tagged TEV protease (50 μg/mg of recombinant GGPS) for 48 h at 4 °C, followed by removal of His-tagged protein by passing the digest over nickel-nitrilotriacetic acid resin and collecting the unbound fraction. The TEV-cleaved protein was further purified by gel filtration chromatography using a Superdex 200 column on an ÄKTA purifier system (GE Healthcare). Purity and integrity of GGPS were confirmed by SDS-PAGE and liquid chromatography/mass spectrometry (Agilent). Selenomethionine Labeling—Selenomethionine-substituted protein was produced using cells grown in SelenoMet medium (Molecular Dimensions) in the presence of 75 mg/liter selenomethionine together with amino acids suppressing de novo synthesis of methionine (15Doublie S. Methods Enzymol. 1997; 276: 523-530Crossref PubMed Scopus (796) Google Scholar). Labeled GGPS was purified as described for the native protein, and the incorporation of selenomethionine was confirmed by liquid chromatography/mass spectrometry. Crystallization and Data Collection—Crystals of native protein were grown at 20 °C in sitting drops by mixing 200 nl of 90 mg/ml protein in 10 mm HEPES, pH 7.5, 500 mm NaCl, 5% glycerol with 100 nl of precipitant solution consisting of 25% PEG 3350, 200 mm magnesium formate, pH 5.5, and equilibrating against 100 μl of the precipitant solution. A single crystal was transferred to a cryo-protectant prepared with 20% glycerol, 80% well solution and flash-cooled in liquid nitrogen. A native data set was collected at a wavelength of 1.008 Å at the Swiss Light Source PXII beamline. Data processing indicated the space group was either P41212 or P43212, and calculation of a Matthews coefficient of 2.5 implied six monomers per asymmetric unit. Attempts at molecular replacement using models with ∼20% sequence identity were unsuccessful (after this structure was deposited, the Saccharomyces cerevisiae form of GGPS with 44% identity became available). Selenomethionine-labeled protein was crystallized by suspending a 3-μl drop containing 26 mg/ml protein, 333 mm NaCl, 0.67 mm MgCl2, 0.67 mm GGPP, 3% glycerol, 15% 2-methyl-2,4-pentanediol, 1.7% PEG 10,000, 6.7 mm HEPES, pH 7.5, over a 1-ml reservoir containing 45% 2-methyl-2,4-pentanediol, 5% PEG 10,000. A single crystal was flashed-cooled in liquid nitrogen. Diffraction data were collected for the selenomethionine derivative at the Swiss Light Source PXII beamline at 0.9791 Å (peak wavelength determined from a fluorescence scan). Data sets were processed using XDS (16Kabsch W. Arnold E. Rossmann M.G. International Tables for Crystallography. Kluwer Academic Publishers Group, Dordrecht, Netherlands2001: 218-225Google Scholar), and data statistics are shown in Table 1. Data processing confirmed a P1 unit cell for the derivative crystal, and analysis of solvent content predicted 12 monomers in the unit cell.TABLE 1Data processing and refinement statistics Values in parentheses are for data in highest shell. r.m.s.d. indicates root mean square deviation.NativeSelenomethionineSpace groupP41212P1Unit cell (Å and degrees)a = b = 141.2, c = 211.7a = 82.05, b = 102.8, c = 139.7, α = 69.36, β = 81.51, γ = 70.27No. of reflections458,740730,472No. of unique data59,48296,958Resolution (Å)47.1-2.70 (2.80-2.70)49.1-2.80 (2.90-2.80)Completeness99.9% (100.0%)98.0% (97.4%)〈I/σ(I) 〉12.1 (2.8)23.2 (5.7)Rint0.1365 (0.5221)0.0718 (0.2543)Final modelRcryst/Rfree0.21/0.25R.m.s.d. bond length (Å)0.014R.m.s.d bond angle (°)1.454Protein atoms (6 chains)13,723GGPP atoms164Magnesium ions12Water molecules24 Open table in a new tab Structure Solution, Model Building, and Refinement—Before merging the selenomethionine data set, the anomalous signal was assessed by calculating ΔF/σ using the program XPREP (Bruker AXS, 2005) indicating significant signal to 4.0 Å. Forty eight selenium sites (corresponding to 12 chains) were easily found using SHELXD (17Sheldrick G.M. Hauptman H.A. Weeks C.M. Miller M. Usón I. Arnold E. Rossmann M.G. International Tables for Crystallography. F. Kluwer Academic Publishers Group, Dordrecht, Netherlands2001: 333-351Google Scholar) with convincing statistics. Noncrystallographic symmetry (NCS) analysis was performed manually by assigning selenium sites to their NCS equivalents, and NCS matrices were calculated using LSQKAB (18Kabsch W. Acta Crystallogr. 1976; A32: 922-923Crossref Scopus (2354) Google Scholar). Initial phases were calculated using SHELXE (19Sheldrick G.M. Z. Kristallogr. 2002; 217: 644-650Crossref Scopus (360) Google Scholar), and NCS averaging was performed with dm (20Cowtan K. Joint CCP4 and ESF-EACBM Newsletter on Protein Crystallography. 1994; 31: 34-38Google Scholar) using an NCS mask calculated from the selenium positions (a 15-Å radius sphere was used around each selenium atom to construct the mask). Visual inspection of the electron density clearly indicated there were two hexamers in the P1 unit cell. An initial model for a single chain was built into the density-modified P1 maps using the program COOT (21Emsley P. Cowtan K. Acta Crystallogr. Sect. D Biol. Crystallogr. 2004; 60: 2126-2132Crossref PubMed Scopus (23370) Google Scholar). Subsequently, molecular replacement was performed with PHASER (22McCoy A.J. Grosse-Kunstleve R.W. Storoni L.C. Read R.J. Acta Crystallogr. Sect. D Biol. Crystallogr. 2005; 61: 458-464Crossref PubMed Scopus (1599) Google Scholar) on the native data set testing both enantiomeric space groups in the tetragonal unit cell. A solution was found for six monomers in the P41212 unit cell with similar hexameric arrangement as seen in the P1 cell. Before refinement commenced, 5% of the data were flagged for calculation of Rfree. Iterative rounds of refinement using REFMAC5 (23Murshudov G.N. Acta Crystallogr. Sect. D Biol. Crystallogr. 1997; 53: 240-255Crossref PubMed Scopus (13868) Google Scholar) and manual fitting in COOT converged to the final model for which statistics are shown in Table 1. Tight main chain and medium side chain NCS restraints were used throughout the refinement process. Sequence Alignment—Sequences were extracted from NCBI and were aligned using the program ICM (Molsoft, San Diego) with the implemented alignment tool. The four GGPS crystal structures (from human, Thermus thermophilus, Pyrococcus horikoshi, and S. cerevisiae) were superimposed and used as template to compare all primary structures. Obvious misalignments in the resulting output file were manually adjusted to obtain the final alignment. Determination of Molecular Weight in Solution—The molecular weight of GGPS was measured by gel filtration chromatography using an ÄKTA purifier system. GGPS (3 mg) was applied to a calibrated Superdex 200 10/300 GL (GE Healthcare) column and developed with 100 mm NaCl, 10 mm HEPES, pH 7.5, 1 mm MgCl2, 0.5 mm TCEP at a flow rate of 0.5 ml/min. The molecular mass in solution was estimated by comparing the retention time of GGPS to the standard curve obtained with molecular weight markers (Sigma). Kinetics of Recombinant Human GGPS—GGPS activity was analyzed by the method of Reed and Rilling (24Reed B.C. Rilling H.C. Biochemistry. 1976; 15: 3739-3745Crossref PubMed Scopus (69) Google Scholar) with the following modifications. In brief, 80 ng (2 pmol) was assayed in a final volume of 100 μl of buffer containing 50 mm Tris, pH 7.7, 2 mm MgCl2, 1 mm TCEP, 5 μg/ml bovine serum albumin, 0.2% Tween 20. The concentrations of FPP and IPP ([14C]IPP, 400 kBq/μmol) were as indicated and were typically between 0.2 and 20 μm. For inhibition studies, the concentration of GGPP varied from 0.4 to 40 μm. Reactions were initiated by addition of enzyme and were allowed to proceed at 37 °C. Assays were terminated by the addition of 0.2 ml of HCl/methanol (1:4) and incubated for 10 min at 37 °C. The reaction mixtures were extracted with 0.4 ml of ligroin and, after thorough mixing, the amount of radioactivity in the upper phase was determined using a Packard Tricarb 1900CA scintillation counter by adding 0.2 ml of the ligroin to 4 ml of general purpose scintillant. Data were fitted by nonlinear regression to the Michaelis-Menten equation using the Graphpad Prism software package or to a competitive inhibition model using the enzyme kinetics module in Sigmaplot. Analysis of Reaction Products—Enzyme reactions were performed as described above except that [14C]IPP (2.18 GBq/mmol) was employed at a concentration of 17 μm, and the reactions were carried out in 50-μl volumes. Reactions were initiated by the addition of 1 μg of protein and stopped after 5 min by the addition of 2 μl of 0.5 m EDTA. Enzyme reaction products were analyzed using thin layer chromatography by spotting 5 μl of the reaction mixture onto Silica Gel 60 TLC plates (Merck) that were developed in propan-2-ol/ammonia/water (6:3:1). Standards were visualized by staining with iodine vapor, and radioactivity was visualized using a Storm860 PhosphorImaging system (GE Healthcare). Biochemical Analysis of Recombinant Human GGPS—Size-exclusion chromatography of recombinant GGPS revealed a molecular mass of 193 kDa (Fig. 2A) which, given a monomeric molecular mass of 34.96 kDa, suggests there are 5-6 monomers associated per molecule in solution. This is consistent with previous studies that estimate that GGPS from bovine brain is a 195-kDa homo-oligomer (25Sagami H. Morita Y. Ogura K. J. Biol. Chem. 1994; 269: 20561-20566Abstract Full Text PDF PubMed Google Scholar) but deviates, however, from a study showing that human GGPS is a 280-kDa octamer (26Kuzuguchi T. Morita Y. Sagami I. Sagami H. Ogura K. J. Biol. Chem. 1999; 274: 5888-5894Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar). Importantly, these data confirm that the crystallographic model containing six monomers (see below) is the biologically relevant unit. Human GGPS expressed in E. coli as an N-terminal His-tagged protein was used to study steady-state kinetics and to ensure that the enzyme used for crystallization was active. The tagged and TEV-cleaved forms of the enzyme were tested for activity, and it was found that the presence of the tag had no significant effect on enzyme activity (data not shown). GGPS generated 14C-labeled acid-labile products using DMAPP, GPP, or FPP as a substrate. This activity was stimulated by the addition of 0.2% Tween 20 to the reaction (Fig. 2B), which is consistent with the finding that octyl glucoside increases activity for the bovine form of the enzyme (25Sagami H. Morita Y. Ogura K. J. Biol. Chem. 1994; 269: 20561-20566Abstract Full Text PDF PubMed Google Scholar). Analysis of reaction mixtures by thin layer chromatography showed that the enzyme has a strong preference for FPP as the allylic substrate in agreement with previous observations (26Kuzuguchi T. Morita Y. Sagami I. Sagami H. Ogura K. J. Biol. Chem. 1999; 274: 5888-5894Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar). Significant to the structural results discussed below, the ultimate product of the GGPS-catalyzed reaction is GGPP, and reactions set up using GGPP as allylic substrate showed no further chain elongation. Steady-state kinetic constants were calculated by varying the concentration of one substrate while holding the concentration of the second substrate constant (Table 2 and Fig. 2, C and D). The enzyme catalyzes the production of GGPP with a kcat = 0.204 s−1 and apparent Km values of 3.0 μm (IPP) and 4.2 μm (FPP). The Michaelis constants are in general agreement with reported values for the bovine brain (Km, IPP = 2 μm and Km, FPP = 0.74 μm) and yeast enzymes (Km, IPP = 0.8 μm and Km, FPP = 3.2 μm) (25Sagami H. Morita Y. Ogura K. J. Biol. Chem. 1994; 269: 20561-20566Abstract Full Text PDF PubMed Google Scholar, 27Chang T.H. Guo R.T. Ko T.P. Wang A.H. Liang P.H. J. Biol. Chem. 2006; 281: 14991-15000Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar). Although the kcat is 2-fold lower than that determined for human FPPS (kcat = 0.42 s−1), it is an order of magnitude higher than the reported value for S. cerevisiae GGPS (kcat = 0.025 s−1) and is therefore within the range expected for this class of enzymes (14Kavanagh K.L. Guo K. Dunford J.E. Wu X. Knapp S. Ebetino F.H. Rogers M.J. Russell R.G.G. Oppermann U. Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 7829-7834Crossref PubMed Scopus (451) Google Scholar, 27Chang T.H. Guo R.T. Ko T.P. Wang A.H. Liang P.H. J. Biol. Chem. 2006; 281: 14991-15000Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar). Unlike FPP synthase (28Barnard G.F. Popjak G. Biochim. Biophys. Acta. 1981; 661: 87-99Crossref PubMed Scopus (39) Google Scholar), no substrate inhibition was observed at IPP concentrations up to 100 μm. However, the product GGPP was found to inhibit the enzyme competitively with respect to FPP with a Ki of 25 μm.TABLE 2Steady-state kinetic parameters for recombinant human GGPSConstantIPPFPPKm (μm)3.0 ± 0.24.2 ± 0.3Vmax (nmol/min/nmol)12.7kcat (FPP as substrate)0.204 s−1 Open table in a new tab Overall Structure of Human GGPS—The crystallographic asymmetric unit for the native protein contains six protein chains (A-F), each associated with two Mg2+ ions and one GGPP molecule (Fig. 3A). Except for subunit F in which the GGPP molecule is disordered past C-10, the refined B-factors for the ligand atoms are comparable with the B-factors for the surrounding protein atoms. Similar to FPPS, each chain adopts the all α-helical trans-prenyltransferase fold, and the monomers associate into dimers. The crystal structures of GGPS from the bacteria T. thermophilus (PDB code 1WMW), 4K. Suto, K. Nishio, Y. Nodake, K. Hamada, M. Kawamoto, N. Nakagawa, S. Kuramitu, and K. Miura, unpublished results. the Archaea P. horikoshii (PDB code 1WY0), 5M. Sugahara and N. Kunishima, unpublished results. and the yeast S. cerevisiae (27Chang T.H. Guo R.T. Ko T.P. Wang A.H. Liang P.H. J. Biol. Chem. 2006; 281: 14991-15000Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar) also exhibit this dimeric quaternary structure. In fact, all-trans-prenyltransferases that have previously had their x-ray structures determined are dimeric. However, in both crystal forms that we have characterized three dimers join together to form a hexamer in a 3-bladed propeller arrangement (Fig. 3A). Size-exclusion chromatography (see above) confirms that the protein is hexameric in solution. As previously seen in other trans-prenyltransferases, the monomers are composed of α-helices joined together by loop regions. Ten of these helices form a helical bundle that surrounds a central cavity where the active site is located. The N-terminal helix is perpendicular to the core helices and contributes residues that are involved in dimerization as well as hexamerization. Helices α9-α11 form a lid over the active site, and in human GGPS this region is also involved in inter-subunit interactions as discussed below. Each chain associates with its dimer partner and a monomer from each of the other two dimers (Fig. 3B). The average surface area per monomer is 13,573 Å2 with the total surface area buried by hexamer formation equal to 16,750 Å2. The largest contact area occurs at the monomer to dimer level with ∼1600 Å2 per monomer buried upon dimer formation. On average, an additional 1200 Å2 per monomer is buried upon hexamerization. To investigate whether the regions involved in inter-dimer contacts are conserved in other species, a sequence alignment of trans-prenyltransferases was performed and is shown in Fig. 4. This family of enzymes consists of α-helical proteins, and the secondary structural elements observed in human GGPS are drawn as red cylinders below the alignment and are labeled α2-α13 (Fig. 4). Regions that are involved in inter-dimer contacts in the human GGPS structure are boxed in blue and labeled A-C (Fig. 4). The following discussions concerning equivalent amino acids refer to the residue numbering for human GGPS. The inter-dimer interface between chains A and E (Fig. 3B) consists of residues from the N-terminal helix α2 (region A in Fig. 4) and loop 75-83 (region B) on chain A interacting with residues 226-254 from chain E (region C) on the adjacent dimer (Fig. 3B). At the core of this interface is a hydrophobic patch consisting of Tyr-18, Phe-76, Pro-77, Ile-82, and Tyr-83 of chain A contacting Ile-233, Ile-243, and Tyr-246 of chain E. These contacts are mirrored with Ile-233, Ile-243, and Tyr-246 of chain A interacting with Tyr-18, Phe-76, Pro-77, Ile-82, and Tyr-83 on chain D of the third dimer to form a ring-like structure. Similar interactions are observed between chains B, C, and F. These regions are largely conserved in mammalian and Drosophila GGPS but not in bacterial, archaeal, fungal, or plant GGPS or in mammalian FPPS indicating that this hexameric quaternary structure may be limited to a subset of eukaryotic GGPS, including mammalian and insect orthologs. Analysis of Ligand Pocket—The site where GGPP is bound is a ∼25-Å long channel surrounded by mainly aliphatic and aromatic side chains of residues Arg-28, Leu-31, Phe-35, His-57, Leu-122, Leu-155, Phe-156, Ala-159, Val-160, and Phe-184 (Fig. 5, A and B). This pocket is capped by charged and polar residues, including the aspartate-rich motifs on helices α4 and α8 (64DDIED68 and 188DDYAN192) that ligate the magnesium ions and the pyrophosphate moieties, and residues Arg-73, Lys-151, Gln-185, Lys-202, and Lys-212 that are also involved in phosphate binding. A similar arrangement of magnesium ions mediating the interaction between acidic side chains and phosphate groups occurs in ligand-bound FPPS structures (14Kavanagh K.L. Guo K. Dunford J.E. Wu X. Knapp S. Ebetino F.H. Rogers M.J. Russell R.G.G. Oppermann U. Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 7829-7834Crossref PubMed Scopus (451) Google Scholar, 29Hosfield D.J. Zhang Y. Dougan D.R. Broun A. Tari L.W. Swanson R.V. Finn J. J. Biol. Chem. 2004; 279: 8526-8529Abstract Full Text Full Text PDF PubMed Scopus (237) Google Scholar, 30Gabelli S.B. McLellan J.S. Montalvetti A. Oldfield E. Docampo R. Amzel L.M. Proteins. 2006; 62: 80-88Crossref PubMed Scopus (126) Google Scholar, 31Rondeau J.M. Bitsch F. Bourgier E. Geiser M. Hemmig R. Kroemer M. Lehmann S. Ramage P. Rieffel S. Strauss A. Green J.R. Jahnke W. Chem. Med. Chem. 2006; 1: 267-273Crossref Scopus (206) Google Scholar, 32Tarshis L.C. Proteau P.J. Kellogg B.A. Sacchettini J.C. Poulter C.D. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 15018-15023Crossref PubMed Scopus (315) Google Scholar). By analogy with FPPS with which GGPS shares a common fold and conserved motifs known to be involved in catalysis (Fig. 4), the reaction is proposed to proceed by an ionization-condensation-elimination mechanism (33Poulter C.D. Argyle J.C. Mash E.A. J. Biol. Chem. 1978; 253: 7227-7233Abstract Full Text PDF PubMed Google Scholar, 34Poulter C.D. Satterwhite D.M. Biochemistry. 1977; 16: 5470-5478Crossref PubMed Scopus (75) Google Scholar). In this scheme (Fig. 1B), the enzyme-bound allylic substrate undergoes cleavage at the C-1-O bond. The resulting carbocation intermediate is proposed to be stabilized by the accompanying negatively charged pyrophosphate and by residues in the active site, most notably Thr-152 or its equivalent (motif IV, Fig. 4). Condensation of IPP with the first carbocation intermediate results in a second positively charged intermediate, and the final product results from stereospecific elimination of a proton. In the alignment shown in Fig. 4, the five conserved regions pr
Referência(s)