Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

Structural and Kinetic Analyses of the Interaction of Anthrax Adenylyl Cyclase Toxin with Reaction Products cAMP and Pyrophosphate

2004; Elsevier BV; Volume: 279; Issue: 28 Linguagem: Inglês

10.1074/jbc.m402689200

ISSN

1083-351X

Autores

Qing Guo, Yuequan Shen, Natalia L. Zhukovskaya, Jan Florián, Wei‐Jen Tang,

Tópico(s)

Poxvirus research and outbreaks

Resumo

Anthrax edema factor (EF) raises host intracellular cAMP to pathological levels through a calcium-calmodulin (CaM)-dependent adenylyl cyclase activity. Here we report the structure of EF·CaM in complex with its reaction products, cAMP and PPi. Mutational analysis confirmed the interaction of EF with cAMP and PPi as depicted in the structural model. While both cAMP and PPi have access to solvent channels to exit independently, PPi is likely released first. EF can synthesize ATP from cAMP and PPi, and the estimated rate constants of this reaction at two physiologically relevant calcium concentrations were similar to those of adenylyl cyclase activity of EF. Comparison of the conformation of adenosine in the structures of EF·CaM·cAMP·PPi with EF·CaM·3·dATP revealed about 160° rotation in the torsion angle of N-glycosyl bond from the +anti conformation in 3·dATP to -syn in cAMP; such a rotation could serve to distinguish against substrates with the N-2 amino group of purine. The catalytic rate of EF for ITP was about 2 orders of magnitude better than that for GTP, supporting the potential role of this rotation in substrate selectivity of EF. The anomalous difference Fourier map revealed that two ytterbium ions (Yb3+) could bind the catalytic site of EF·CaM in the presence of cAMP and PPi, suggesting the presence of two magnesium ions at the catalytic site of EF. We hypothesize that EF could use a "histidine and two-metal ion" hybrid mechanism to facilitate the cyclization reaction. Anthrax edema factor (EF) raises host intracellular cAMP to pathological levels through a calcium-calmodulin (CaM)-dependent adenylyl cyclase activity. Here we report the structure of EF·CaM in complex with its reaction products, cAMP and PPi. Mutational analysis confirmed the interaction of EF with cAMP and PPi as depicted in the structural model. While both cAMP and PPi have access to solvent channels to exit independently, PPi is likely released first. EF can synthesize ATP from cAMP and PPi, and the estimated rate constants of this reaction at two physiologically relevant calcium concentrations were similar to those of adenylyl cyclase activity of EF. Comparison of the conformation of adenosine in the structures of EF·CaM·cAMP·PPi with EF·CaM·3·dATP revealed about 160° rotation in the torsion angle of N-glycosyl bond from the +anti conformation in 3·dATP to -syn in cAMP; such a rotation could serve to distinguish against substrates with the N-2 amino group of purine. The catalytic rate of EF for ITP was about 2 orders of magnitude better than that for GTP, supporting the potential role of this rotation in substrate selectivity of EF. The anomalous difference Fourier map revealed that two ytterbium ions (Yb3+) could bind the catalytic site of EF·CaM in the presence of cAMP and PPi, suggesting the presence of two magnesium ions at the catalytic site of EF. We hypothesize that EF could use a "histidine and two-metal ion" hybrid mechanism to facilitate the cyclization reaction. cAMP, a key intracellular second messenger, is primarily regulated at the level of synthesis by adenylyl cyclase, the enzyme that converts ATP to cAMP and pyrophosphate. Adenylyl cyclase can be categorized into five classes (1Linder J.U. Schultz J.E. Cell. Signal. 2003; 15: 1081-1089Crossref PubMed Scopus (143) Google Scholar, 2Sunahara R.K. Dessauer C.W. Gilman A.G. Annu. Rev. Pharmacol. Toxicol. 1996; 36: 461-480Crossref PubMed Scopus (732) Google Scholar). Enzymes within a class share sequence similarity but have no homology with members from the other classes. Class II adenylyl cyclase consists of several bacterial toxins that are secreted by pathogenic bacteria and activated upon their entry into host cells (3Yahr T.L. Vallis A.J. Hancock M.K. Barbieri J.T. Frank D.W. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 13899-13904Crossref PubMed Scopus (353) Google Scholar, 4Ladant D. Ullmann A. Trends Microbiol. 1999; 7: 172-176Abstract Full Text Full Text PDF PubMed Scopus (221) Google Scholar, 5Shen Y. Lee Y.S. Soelaiman S. Bergson P. Lu D. Chen A. Beckingham K. Grabarek Z. Mrksich M. Tang W.J. EMBO J. 2002; 21: 6721-6732Crossref PubMed Scopus (82) Google Scholar). These include edema factor (EF) 1The abbreviations used are: EF, edema factor; EF3, catalytic domain of edema factor (amino acids 291-800); CaM, calmodulin; EF3·-CaM, catalytic domain of edema factor complexed with calmodulin; 3′dATP, 3′-deoxy-ATP; 2′d-3′ANT-ATP, 2′-deoxy-3′-anthraniloyl ATP; ATPαS, adenosine α-thio-5′-triphosphate; cAMPαS, adenosine 3′,5′-cyclic monophosphorothioate; HPLC, high pressure liquid chromatography.1The abbreviations used are: EF, edema factor; EF3, catalytic domain of edema factor (amino acids 291-800); CaM, calmodulin; EF3·-CaM, catalytic domain of edema factor complexed with calmodulin; 3′dATP, 3′-deoxy-ATP; 2′d-3′ANT-ATP, 2′-deoxy-3′-anthraniloyl ATP; ATPαS, adenosine α-thio-5′-triphosphate; cAMPαS, adenosine 3′,5′-cyclic monophosphorothioate; HPLC, high pressure liquid chromatography. from Bacillus anthracis (anthrax), CyaA from Bordetella pertussis (whooping cough), and ExoY from Pseudomonas aeruginosa (various nosocomial infections). Class III is the largest group, which includes adenylyl cyclases from bacteria, yeasts, parasites, insects, and vertebrates. Class III includes enzymes responsive to a plethora of extracellular signals such as hormones, neurotransmitters, odorants, and chemokines. These enzymes control diverse physiological responses such as sugar and lipid metabolism, fight or flight responses, and learning and memory. The other three classes are found in various prokaryotes including Gram-negative bacteria (class I), Aeromonas hydrophila (class IV), and Prevotella ruminicola (class V). The molecular structures of the catalytic domain of EF (class II) and mammalian adenylyl cyclase (class III) reveal no structure similarity between these two members, suggesting the converging evolution of these two classes of enzymes (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar, 7Tesmer J.J. Sunahara R.K. Gilman A.G. Sprang S.R. Science. 1997; 278: 1907-1916Crossref PubMed Scopus (670) Google Scholar).EF, a key virulence factor for anthrax pathogenesis, has two functional domains (8Mock M. Fouet A. Annu. Rev. Microbiol. 2001; 55: 647-671Crossref PubMed Scopus (851) Google Scholar, 9Mourez M. Lacy D.B. Cunningham K. Legmann R. Sellman B.R. Mogridge J. Collier R.J. Trends Microbiol. 2002; 10: 287-293Abstract Full Text Full Text PDF PubMed Scopus (67) Google Scholar). The N-terminal 30-kDa domain of EF binds anthrax protective antigen with high affinity (5-10 nm), enabling its entrance into the intracellular space (10Elliott J.L. Mogridge J. Collier R.J. Biochemistry. 2000; 39: 6706-6713Crossref PubMed Scopus (110) Google Scholar). The C-terminal 58-kDa domain of EF is a calmodulin (CaM)-dependent adenylyl cyclase, and its activity is modulated by physiological calcium concentrations (5Shen Y. Lee Y.S. Soelaiman S. Bergson P. Lu D. Chen A. Beckingham K. Grabarek Z. Mrksich M. Tang W.J. EMBO J. 2002; 21: 6721-6732Crossref PubMed Scopus (82) Google Scholar). This domain can be further divided into two functional entities. The N-terminal 43-kDa portion of the adenylyl cyclase domain of EF forms the catalytic core that shares 34 and 29% sequence similarity to CyaA and ExoY, while the C-terminal 17-kDa helical domain has no catalytic activity but facilitates CaM activation of EF (11Drum C.L. Yan S.Z. Sarac R. Mabuchi Y. Beckingham K. Bohm A. Grabarek Z. Tang W.J. J. Biol. Chem. 2000; 275: 36334-36340Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar). Structures of the 58-kDa domain of EF alone and in complex with CaM reveal that one of the catalytic loops of EF is disordered in the absence of CaM (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar). CaM has N- and C-terminal globular domains, each binding two Ca2+ ions (12Hoeflich K.P. Ikura M. Cell. 2002; 108: 739-742Abstract Full Text Full Text PDF PubMed Scopus (579) Google Scholar). NMR and mutational analyses suggests that the N-terminal CaM initiates its contact with the C-terminal 17-kDa helical domain of EF, leading to the insertion of C-terminal CaM between the catalytic core and helical domains of EF (5Shen Y. Lee Y.S. Soelaiman S. Bergson P. Lu D. Chen A. Beckingham K. Grabarek Z. Mrksich M. Tang W.J. EMBO J. 2002; 21: 6721-6732Crossref PubMed Scopus (82) Google Scholar, 13Ulmer T.S. Soelaiman S. Li S. Klee C.B. Tang W.J. Bax A. J. Biol. Chem. 2003; 278: 29261-29266Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar). The binding of CaM induces the conformational changes to stabilize the disordered catalytic loop, leading to over 1000-fold increase in the catalytic rate (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar).EF has a relatively high catalytic rate with a turnover number around 1000-2000 s-1. With a Km around 0.2-1 mm, the catalytic efficiency (kcat/Km of EF·CaM) approaches 107·m-1·s-1, a catalytic rate that is at least 100-fold higher than mammalian adenylyl cyclases (mACs) (5Shen Y. Lee Y.S. Soelaiman S. Bergson P. Lu D. Chen A. Beckingham K. Grabarek Z. Mrksich M. Tang W.J. EMBO J. 2002; 21: 6721-6732Crossref PubMed Scopus (82) Google Scholar, 6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar, 14Dessauer C.W. Gilman A.G. J. Biol. Chem. 1997; 272: 27787-27795Abstract Full Text Full Text PDF PubMed Scopus (118) Google Scholar, 15Tesmer J.J. Sunahara R.K. Johnson R.A. Gosselin G. Gilman A.G. Sprang S.R. Science. 1999; 285: 756-760Crossref PubMed Scopus (277) Google Scholar). Structures of EF3·CaM in complex with several non-cyclizable ATP analogs together with mutational analyses have provided a starting point in building a model of EF catalysis (5Shen Y. Lee Y.S. Soelaiman S. Bergson P. Lu D. Chen A. Beckingham K. Grabarek Z. Mrksich M. Tang W.J. EMBO J. 2002; 21: 6721-6732Crossref PubMed Scopus (82) Google Scholar, 6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar, 16Shen Y. Guo Q. Zhukovskaya N.L. Drum C.L. Bohm A. Tang W.J. Biochem. Biophys. Res. Commun. 2004; 317: 309-314Crossref PubMed Scopus (23) Google Scholar). The adenine moiety is recognized by a main chain carbonyl, while the ribose is held in position by an asparagine (Asn-583). The triphosphate moiety is coordinated by several positively charged residues, including Arg-329, Lys-346, Lys-353, and Lys-372. His-351 is near the putative 3′-OH. The homologous residue in CyaA (His-63) is postulated to act as a catalytic base (17Munier H. Bouhss A. Krin E. Danchin A. Gilles A.M. Glaser P. Barzu O. J. Biol. Chem. 1992; 267: 9816-9820Abstract Full Text PDF PubMed Google Scholar). This is based on the observation that the mutation of His-63 to arginine shifted the pH dependence toward a more alkaline optimum. Thus, His-351 is proposed to serve as a catalytic base to generate 3′-oxy anion. EF also has two aspartates, Asp-491 and Asp-493, that could coordinate the catalytic metals similar to mACs and many DNA and RNA polymerases (15Tesmer J.J. Sunahara R.K. Johnson R.A. Gosselin G. Gilman A.G. Sprang S.R. Science. 1999; 285: 756-760Crossref PubMed Scopus (277) Google Scholar, 18Steitz T.A. Steitz J.A. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 6498-6502Crossref PubMed Scopus (1019) Google Scholar, 19Ling H. Boudsocq F. Woodgate R. Yang W. Cell. 2001; 107: 91-102Abstract Full Text Full Text PDF PubMed Scopus (536) Google Scholar, 20Doublie S. Tabor S. Long A.M. Richardson C.C. Ellenberger T. Nature. 1998; 391: 251-258Crossref PubMed Scopus (1099) Google Scholar, 21Huang H. Chopra R. Verdine G.L. Harrison S.C. Science. 1998; 282: 1669-1675Crossref PubMed Scopus (1354) Google Scholar, 22Franklin M.C. Wang J. Steitz T.A. Cell. 2001; 105: 657-667Abstract Full Text Full Text PDF PubMed Scopus (487) Google Scholar, 23Sawaya M.R. Prasad R. Wilson S.H. Kraut J. Pelletier H. Biochemistry. 1997; 36: 11205-11215Crossref PubMed Scopus (571) Google Scholar, 24Pelletier H. Sawaya M.R. Kumar A. Wilson S.H. Kraut J. Science. 1994; 264: 1891-1903Crossref PubMed Scopus (756) Google Scholar).Little is known about how EF binds and releases reaction products cAMP and PPi. Here we report the structure determination of EF·CaM in complex with reaction products cAMP and PPi as well as a kinetic analysis of EF. These analyses suggest a mechanism for the binding and releasing of reaction products in EF. The structure of EF·CaM in complex with cAMP and PPi also offers evidence suggesting a "histidine and two-metal ion" hybrid mechanism of catalysis.EXPERIMENTAL PROCEDURESMaterials—The QuikChange kit was purchased from Biocrest; Bradford reagent was from Bio-Rad. Nickel-nitrilotriacetic acid resin and anti-H5 antiserum were from Qiagen. Pyrophosphate, cAMP, (Rp)-cAMPαS, (Sp)-cAMPαS, ITP, and GTP were from Sigma. The racemic mixture of ATPαS was from Jena Bioscience. The purified (Rp)-ATPαS and (Sp)-ATPαS diastereomers were a gift from Fritz Eckstein at Max-Planck Institute. Hexokinase and type XI glucose-6-phophate dehydrogenase were purchased from Roche Applied Science and Sigma, respectively.Protein Expression and Purification—The plasmids for the expression of mutant forms of the catalytic domain of EF (EF3) were constructed by site-direct mutagenesis and confirmed by DNA sequencing (11Drum C.L. Yan S.Z. Sarac R. Mabuchi Y. Beckingham K. Bohm A. Grabarek Z. Tang W.J. J. Biol. Chem. 2000; 275: 36334-36340Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar). The recombinant proteins expressed and purified from Escherichia coli including EF3, EF3 mutants, and CaM were performed as described previously (5Shen Y. Lee Y.S. Soelaiman S. Bergson P. Lu D. Chen A. Beckingham K. Grabarek Z. Mrksich M. Tang W.J. EMBO J. 2002; 21: 6721-6732Crossref PubMed Scopus (82) Google Scholar, 6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar, 11Drum C.L. Yan S.Z. Sarac R. Mabuchi Y. Beckingham K. Bohm A. Grabarek Z. Tang W.J. J. Biol. Chem. 2000; 275: 36334-36340Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar).Structure Determination of EF3·CaM·cAMP·PPi Complex—To determine the structure of EF3·CaM·cAMP·PPi, crystals of EF3·CaM complex were grown using vapor diffusion, soaked with 1 mm cAMP and 1 mm PPi during cryoprotection for overnight, and frozen in liquid nitrogen as described previously (25Drum C.L. Shen Y. Rice P.A. Bohm A. Tang W.J. Acta Crystallogr. Sect. D Biol. Crystallogr. 2001; 57: 1881-1884Crossref PubMed Scopus (21) Google Scholar). Data were collected at 100 K at the Advanced Photon Source BioCars 14-BM-C and Structural Biology Center ID19 and processed with the programs DENZO and SCALEPACK (26Otwinowski Z. Minor W. Methods Enzymol. 1997; 276: 307-326Crossref Scopus (38361) Google Scholar). The initial phase was obtained by difference Fourier method using the software program CNS and the model of EF3·CaM complex (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar). The model was refined and built using the programs CNS and O (27Brunger A.T. Adams P.D. Clore G.M. DeLano W.L. Gros P. Grosse-Kunstleve R.W. Jiang J.S. Kuszewski J. Nilges M. Pannu N.S. Read R.J. Rice L.M. Simonson T. Warren G.L. Acta Crystallogr. Sect. D Biol. Crystallogr. 1998; 54: 905-921Crossref PubMed Scopus (16929) Google Scholar). The coordinates for EF3·CaM·cAMP·PPi are available from the Protein Data Bank (accession code 1SK6).Enzymatic Assays for the Forward Reaction of Adenylyl Cyclase—The activities were measured at 30 °C in the presence of 10 mm MgCl2, the indicated ATP concentrations, and a trace amount of [α-32P]ATP for 10 min (28Farndale R.W. Allan L.M. Martin B.R. Milligan G. Signal Transduction, a Practical Approach. Oxford Press, Oxford1992: 75-103Google Scholar). The reaction was buffered by 100 mm Hepes, pH 7.2, and free calcium concentration was controlled by 10 mm EGTA to 0.1 and 2 μm free Ca2+ based on calculations using the MAXC program. 2See www.stanford.edu/~cpatton/max.html. cAMP was separated from ATP by Dowex and alumina columns as described previously (28Farndale R.W. Allan L.M. Martin B.R. Milligan G. Signal Transduction, a Practical Approach. Oxford Press, Oxford1992: 75-103Google Scholar). Initial velocities were linear with time, and less than 10% of the ATP was consumed at the lowest substrate concentrations.Enzymatic Assays for the Reverse Reaction of Adenylyl Cyclase—Synthesis of ATP from cyclic AMP and PPi by EF3 was measured spectrophotometrically in the presence of glucose, hexokinase, NADP, and glucose-6-phosphate dehydrogenase (14Dessauer C.W. Gilman A.G. J. Biol. Chem. 1997; 272: 27787-27795Abstract Full Text Full Text PDF PubMed Scopus (118) Google Scholar). Reaction velocities were calculated from the linear increase in A340 resulting from the reduction of NADP. Reactions contained 100 mm Na-Hepes (pH 7.2), 50 mm glucose, 0.8 mm NADP, 10 mm free MgCl2, 2.5 units of hexokinase, and 0.5 units of glucose-6-phosphate dehydrogenase in a volume of 500 μl. PPi was always added last to avoid precipitation. Reactions were typically started by the addition of adenylyl cyclase toxin to the reaction mixtures. The reaction was monitored based on the changes in A340 for 15-20 min at 30 °C in a Beckman DU640 spectrophotometer with a temperature-controlled cuvette holder. The background for the change in A340 in the absence of adenylyl cyclase was subtracted, and optical densities of greater than 1.5 were excluded from analysis.Non-isotopic Adenylyl Cyclase Assays Using HPLC—Adenylyl cyclase assays of EF were carried out using 10 μm CaM, 10 mm MgCl2, 1.1 μm free CaCl2 as calculated using the MAXC program, and the indicated concentrations of EF and nucleotide triphosphate analog at pH 7.2. The reaction was incubated at 30 °C (unless stated otherwise) and was stopped with 50 mm EDTA. 0.4 mm GTP or cAMP was added as a tracer to monitor the recovery. Phenol extraction was used to separate the nucleotides from the proteins. The aqueous phase was then loaded on a reverse-phase C18 HPLC column and eluted with a linear gradient of 0.1 m triethylammonium acetate and acetonitrile.RESULTS AND DISCUSSIONStructure of EF3·CaM in Complex with cAMP and PPi—We have determined the structure of EF3 and CaM with and without the non-cyclizable ATP analogs (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar, 16Shen Y. Guo Q. Zhukovskaya N.L. Drum C.L. Bohm A. Tang W.J. Biochem. Biophys. Res. Commun. 2004; 317: 309-314Crossref PubMed Scopus (23) Google Scholar). To better understand the catalytic mechanism, we solved the structure of EF3·CaM in complex with its reaction products, cAMP and PPi. To do so, EF3·CaM crystals were soaked with cAMP and PPi (1 mm each), and the crystal was diffracted at best to 3.2-Å resolution (Table I). The structure model of EF and CaM in the EF3·CaM·cAMP·PPi structure is similar to those in EF3·CaM. EF3 consists of the catalytic core (CA and CB) and helical domains, and the extended conformation of CaM is inserted between these two domains of EF3 (Fig. 1A). There are three EF·CaM molecules in each asymmetric unit of I222 crystal lattice, and the cAMP and PPi molecules are clearly visible in the active site of all three EF molecules based on the simulated annealing omit map contoured at 3.5 σ (Fig. 1B). We also soaked EF3·CaM crystals with 1 mm cAMP. However, the simulated annealing omit map revealed no visible electronic density of cAMP in the catalytic site of EF, suggesting that PPi is required for the high occupancy of cAMP.Table IStatistics of the EF3·CaM·cAMP·PPi complex dataData collectionBeamlineAPS SBC ID 19aAdvanced Photon Source Structural Biology Center ID19.Space groupI222Unit cell (Å)a116.9b166.4c342.1Resolution (Å)15-3.2Completeness (%)98.4 (96.1)bThe outer resolution shell.RedundancycNobs/Nunique.5.5 (4.3)bThe outer resolution shell.Rsym (%)dRsym = ∑j|I - Ij|/∑I where Ij is the intensity of the jth reflection and I is the average intensity.11.2 (50.2)bThe outer resolution shell.I/σ16.6 (2.3)bThe outer resolution shell.RefinementRcryst (%)eRcryst = ∑hkl|Fobs - Fcalc|/∑hklFobs.0.250Rfree (%)fRfree, calculated the same as for Rcryst but on the 10% data excluded from the refinement calculation.0.307r.m.s.bond (Å)gr.m.s., root mean square.0.009r.m.s.angle (°)1.35a Advanced Photon Source Structural Biology Center ID19.b The outer resolution shell.c Nobs/Nunique.d Rsym = ∑j|I - Ij|/∑I where Ij is the intensity of the jth reflection and I is the average intensity.e Rcryst = ∑hkl|Fobs - Fcalc|/∑hklFobs.f Rfree, calculated the same as for Rcryst but on the 10% data excluded from the refinement calculation.g r.m.s., root mean square. Open table in a new tab The conformation of cAMP in all three EF3·CaM·cAMP·PPi models of the asymmetry unit is roughly similar. The ribose of cAMP is best fit to 2′-C-exo puckering with the torsion angle of N-glycosidic bond approximately -40° so that adenine is in the -syn conformation relative to ribose. The adenosine ring of cAMP forms van der Waals contacts with the main chains of Thr-548, Gly-578, Thr-579, Asp-582, and Asn-583. The N-6 atom of adenine is within hydrogen bonding distance from the main chain carbonyl of Thr-548. The O-4′ atom of ribose forms a hydrogen bond with the side chain of Asn-583; such interaction is postulated to hold ribose in place during the catalysis (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar). The 3′-O of cAMP is within hydrogen bonding distance (2.7 Å in model B) from the side chain of His-351, which is proposed to serve as a catalytic base (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar). PPi is about 4.3Å-5.1 Å away from cAMP and is coordinated by salt bridges with several positively charged residues including Lys-372 and Lys-346.Ytterbium ion is one of the additives that promotes the growth of the EF·CaM crystal. We have found that only one ytterbium ion occupies the catalytic site of EF in the structure of EF·CaM in complex with the non-cyclizable ATP analog 3′dATP (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar). This ion is coordinated by Asp-491 and Asp-493. Surprisingly the anomalous difference Fourier map revealed the presence of more than one ytterbium ion (see Supplemental Fig. 1) in the active site of EF·CaM·cAMP·PPi. In particular, the alternation of one Yb3+ ion and two Yb3+ ions in a ratio of 4 to 1 reduces the residual electron density (Fo - Fc) to less than 2.5 σ. This suggests that two metal binding states exist in the structure of EF·CaM·cAMP·PPi: one with a single Yb3+ ion and the other with two Yb3+ ions. In the single Yb3+ binding state, the metal ion is coordinated by Asp-491, Asp-493, and His-577. This coordination pattern is similar to that of the Yb3+ ion in the structures of EF·CaM·3′dATP and EF·CaM·2′d-3′ANT-ATP complexes (5Shen Y. Lee Y.S. Soelaiman S. Bergson P. Lu D. Chen A. Beckingham K. Grabarek Z. Mrksich M. Tang W.J. EMBO J. 2002; 21: 6721-6732Crossref PubMed Scopus (82) Google Scholar, 6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar). In the state with two Yb3+ ions, the two ions are about 4 Å apart from each other. The first ion is coordinated by His-577 and Asp-493 and also interacts with 3′-O of cAMP (3.7 Å in model B). The second ion is coordinated by Asp-491 as well as the phosphate of both cAMP and PPi.Mutational Analysis of EF to Validate the Crucial Interactions of EF with Its Products—Mutational analysis was used to evaluate whether the EF3·CaM·cAMP·PPi model accurately depicts the interaction of EF with PPi and cAMP. Lys-372 forms a salt bridge with 3′dATP (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar) and also appears to make a crucial contact with PPi in the structure of the reaction product. Thus, we made the EF3-K372A mutant in which Lys-372 is mutated to alanine. This mutation reduced the catalytic rate constant of EF3-K372A 30-fold and increased the Km value of ATP 3-fold (EF3, 6.0 ms-1, 0.6 mm; EF3-K372A, 0.2 ms-1, 2.0 mm) with minimal effect on the EC50 value for CaM activation (EF3, 12 nm; EF3-K372A, 6 nm). We then tested the ability of cAMP and PPi to inhibit EF3-K372A. Consistent with the structural model, EF3-K372A had at least a 20-fold increase in the IC50 value for the inhibition by PPi, while its ability to be inhibited by cAMP was not affected (Fig. 2, A and B).Fig. 2The inhibitions of EF3 mutants by cAMP and PPi. The activity was measured in the presence of 0.3 nm EF3, 10 μm CaM, 2 mm ATP, and 0.1 μm free Ca2+ with variable concentrations of cAMP and PPi. The specific activities of wild type EF3 (WT) and EF3 mutants EF3-K353A, EF3-K372A, EF3-N583A, EF3-E588A, and EF3-D590A were 1197, 2, 30, 8, 50, and 188 s-1, respectively. Mean ± S.E. values are representative of at least two experiments.View Large Image Figure ViewerDownload (PPT)Lys-353 forms a salt bridge with Glu-588. This salt bridge forms a "lid" over the catalytic site of EF. In addition, Lys-353 is in proximity to form a salt bridge with PPi and the phosphate of cAMP. We have reported previously that the mutation of Lys-353 to alanine resulted in 500-fold reduction in catalytic rate and 7-fold increase in Km value of ATP without affecting the interaction of EF with CaM (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar). Our present kinetic analysis revealed that EF3-K353A also exhibits a 20-fold increase in IC50 value for the inhibition by PPi. In contrast, the propensity of this mutant to be inhibited by cAMP was unaltered (Fig. 2, A and B). Lys-346 also forms a salt bridge with PPi. The mutation of Lys-346 to alanine resulted in a reduction of catalytic rate greater than 4 orders of magnitude, making a more accurate kinetic analysis impractical.In the EF3·CaM·cAMP·PPi structure model, the adenosine moiety participates in numerous main chain interactions with EF. Here the most prominent interaction is the hydrogen bonding of its O-4′ atom with Asn-583. Our previous analysis has shown that the mutation of Asn-583 to alanine decreased the catalytic rate constant 150-fold. The same mutation has only a minimal effect on the Km value of ATP or the IC50 value for CaM activation (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar). Consistent with our structural model, EF3-N583A increased the IC50 value for the inhibition by cAMP about 10-fold, while the sensitivity of this mutant to the inhibition by PPi was the same as that of the wild type enzyme (Fig. 2, C and D). Our structures show that Glu-588 and Asp-590 contribute to the organization of the catalytic site of EF (6Drum C.L. Yan S.Z. Bard J. Shen Y. Lu D. Soelaiman S. Grabarek Z. Bohm A. Tang W.-J. Nature. 2002; 415: 396-402Crossref PubMed Scopus (348) Google Scholar), but they are not directly involved in the binding of cAMP and PPi. The mutation of these residues to alanine resulted in minimal alteration of the inhibition by cAMP or PPi (Fig. 2, C and D). Thus, our mutational data confirms the structural model of EF3·CaM·cAMP·PPi.Product Inhibition of the EF3·CaM Complex—Patterns of inhibition of enzymatic activity by products can be used to determine whether the release of product is an ordered or random event. To do so, we examined the inhibition of adenylyl cyclase activity of EF3 by cAMP and PPi (Fig. 3). As reported, calcium not only affects the binding of CaM to EF to facilitate activation but also binds directly to EF to inhibit catalysis (5Shen Y. Lee Y.S. Soelaiman S. Bergson P. Lu D. Chen A. Beckingham K. Grabarek Z. Mrksich M. Tang W.J. EMBO J. 2002; 21: 6721-6732Crossref PubMed Scopus (82) Google Scholar). Thus, we performed our assays in the presence of two free Ca2+ concentrations, 0.1 μm and 2 μm (Fig. 3, A and B). With the large excess of 10 μm CaM, most of EF should be tightly associated with CaM in both calcium concentrations. However, the activity of EF·CaM was minimally affected by calcium ion at 0.1 μm free Ca2+, while its activity was significantly reduced at 2 μm free Ca2+. At 0.1 μm free Ca2+, kinetic data could be interpreted as an ordered product release with PPi being released first. This is because the kinetics of inhibition of EF activity by cAMP was competitive, while that by PPi was mixed. However, such ordered release became random at 2 μm free Ca2+ when the kinetics of inhibition of enzymatic activity by either cAMP or PPi were

Referência(s)