Predicted Michaelis-Menten Complexes of Cocaine-Butyrylcholinesterase
2001; Elsevier BV; Volume: 276; Issue: 12 Linguagem: Inglês
10.1074/jbc.m006676200
ISSN1083-351X
AutoresHong Sun, Jamal El Yazal, Oksana Lockridge, Lawrence M. Schopfer, Stephen Brimijoin, Yuan‐Ping Pang,
Tópico(s)Protein Structure and Dynamics
ResumoButyrylcholinesterase (BChE) is important in cocaine metabolism, but it hydrolyzes (−)-cocaine only one-two thousandth as fast as the unnatural (+)-stereoisomer. A starting point in engineering BChE mutants that rapidly clear cocaine from the bloodstream, for overdose treatment, is to elucidate structural factors underlying the stereochemical difference in catalysis. Here, we report two three-dimensional Michaelis-Menten complexes of BChE liganded with natural and unnatural cocaine molecules, respectively, that were derived from molecular modeling and supported by experimental studies. Such complexes revealed that the benzoic ester group of both cocaine stereoisomers must rotate toward the catalytic Ser198 for hydrolysis. Rotation of (−)-cocaine appears to be hindered by interactions of its phenyl ring with Phe329 and Trp430. These interactions do not occur with (+)-cocaine. Because the rate of (−)-cocaine hydrolysis is predicted to be determined mainly by the re-orientation step, it should not be greatly influenced by pH. In fact, measured rates of this reaction were nearly constant over the pH range from 5.5 to 8.5, despite large rate changes in hydrolysis of (+)-cocaine. Our models can explain why BChE hydrolyzes (+)-cocaine faster than (−)-cocaine, and they suggest that mutations of certain residues in the catalytic site could greatly improve catalytic efficiency and the potential for detoxication. Butyrylcholinesterase (BChE) is important in cocaine metabolism, but it hydrolyzes (−)-cocaine only one-two thousandth as fast as the unnatural (+)-stereoisomer. A starting point in engineering BChE mutants that rapidly clear cocaine from the bloodstream, for overdose treatment, is to elucidate structural factors underlying the stereochemical difference in catalysis. Here, we report two three-dimensional Michaelis-Menten complexes of BChE liganded with natural and unnatural cocaine molecules, respectively, that were derived from molecular modeling and supported by experimental studies. Such complexes revealed that the benzoic ester group of both cocaine stereoisomers must rotate toward the catalytic Ser198 for hydrolysis. Rotation of (−)-cocaine appears to be hindered by interactions of its phenyl ring with Phe329 and Trp430. These interactions do not occur with (+)-cocaine. Because the rate of (−)-cocaine hydrolysis is predicted to be determined mainly by the re-orientation step, it should not be greatly influenced by pH. In fact, measured rates of this reaction were nearly constant over the pH range from 5.5 to 8.5, despite large rate changes in hydrolysis of (+)-cocaine. Our models can explain why BChE hydrolyzes (+)-cocaine faster than (−)-cocaine, and they suggest that mutations of certain residues in the catalytic site could greatly improve catalytic efficiency and the potential for detoxication. butyrylcholinesterase three-dimensional molecular dynamics Cocaine overdose is a leading cause of death among urban-dwelling young adults (1Schrank K.S. Natl. Inst. Drug Abuse Res. Monogr. 1992; 123: 110-128Google Scholar, 2Marzuk P.M. Tardiff K. Leon A.C. Hirsch C.S. Stajic M. Portera L. Hartwell N. Iqbal M.I. N. Engl. J. Med. 1995; 332: 1753-1757Crossref PubMed Scopus (101) Google Scholar). A new idea for treating cocaine overdose is to accelerate metabolic clearance by administering an appropriate hydrolase. Plasma butyrylcholinesterase (BChE)1 can hydrolyze cocaine to ecgonine methyl ester and reportedly accounts for all the blood-borne cocaine hydrolysis activity in humans (3Stewart D.J. Inaba T. Tang B.K. Kalow W. Life Sci. 1977; 20: 1557-1563Crossref PubMed Scopus (196) Google Scholar, 4Inaba T. Stewart D.J. Kalow W. Clin. Pharmacol. Ther. 1978; 23: 547-552Crossref PubMed Scopus (221) Google Scholar, 5Stewart D.J. Inaba T. Lucassen M. Kalow W. Clin. Pharmacol. Ther. 1979; 25: 464-468Crossref PubMed Scopus (253) Google Scholar). Pretreatment with human BChE provides rats with substantial protection against cocaine-induced cardiac arrhythmia, hypertension, and locomotor hyperactivity (6Mattes C. Bradley R. Slaughter E. Browne S. Life Sci. 1996; 58: L257-L261Crossref Scopus (48) Google Scholar). Exogenous BChE is also able to rescue rats given an overdose of cocaine (7Mattes C.E. Lynch T.J. Singh A. Bradley R.M. Kellaris P.A. Brady R.O. Dretchen K.L. Toxicol. Appl. Pharmacol. 1997; 145: 372-380Crossref PubMed Scopus (73) Google Scholar). However, the kcat of BChE for (−)-cocaine is only 3.9 min−1,versus 33,900 min−1 for the optimal substrate, butyrylthiocholine. We estimate that timely detoxication of a cocaine overdose, with serum levels up to 20 mg/l (8Wetli C.V. Am. J. Forensic Med. Pathol. 1987; 8: 1-2Crossref PubMed Scopus (23) Google Scholar), would require at least 100 mg of enzyme. An A328Y mutant of BChE developed by the trial and error approach (9Xie W. Altamirano C.V. Bartels C.F. Speirs R.J. Cashman J.R. Lockridge O. Mol. Pharmacol. 1999; 55: 83-91Crossref PubMed Scopus (120) Google Scholar) showed 4–5-fold improvement in (−)-cocaine hydrolysis, but it is still not satisfactory for treatment of cocaine overdose. Because human BChE hydrolyzes synthetic (+)-cocaine (see Fig. 1) about 2,000-fold faster than the natural (−)-cocaine (see Fig. 1) (9Xie W. Altamirano C.V. Bartels C.F. Speirs R.J. Cashman J.R. Lockridge O. Mol. Pharmacol. 1999; 55: 83-91Crossref PubMed Scopus (120) Google Scholar), elucidation of the structural factors responsible for the different catalytic rates may offer a rational strategy for engineering BChE mutants that can hydrolyze (−)-cocaine rapidly enough to treat cocaine overdose. Theoretical 3D structures of BChE (10Harel M. Sussman J.L. Krejci E. Bon S. Chanal P. Massoulie J. Silman I. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 10827-10831Crossref PubMed Scopus (292) Google Scholar, 11Millard C.B. Broomfield C.A. Biochem. Biophys. Res. Commun. 1992; 189: 1280-1286Crossref PubMed Scopus (31) Google Scholar, 12Felder C.E. Botti S.A. Lifson S. Silman I. Sussman J.L. J. Mol. Graph. Model. 1998; 15: 318-327Crossref Scopus (68) Google Scholar), of its complex with cocaine (9Xie W. Altamirano C.V. Bartels C.F. Speirs R.J. Cashman J.R. Lockridge O. Mol. Pharmacol. 1999; 55: 83-91Crossref PubMed Scopus (120) Google Scholar), and of the cocaine molecule itself have all been reported (13Singh S. Chem. Rev. 2000; 100: 925-1024Crossref PubMed Scopus (254) Google Scholar, 14Carroll F.I. Lewin A.H. Abraham P. Parham K. Boja J.W. Kuhar M.J. J. Med. Chem. 1991; 34: 883-886Crossref PubMed Scopus (74) Google Scholar, 15Sherer E.C. Yang G. Turner G.M. Shields G.C. Landry D.W. J. Phys. Chem. 1997; A101: 8526-8529Crossref Scopus (27) Google Scholar, 16Pati R. Das T.P. Sahoo N. Ray S.N. J. Phys. Chem. A. 1977; 101: 6101-6106Crossref Scopus (5) Google Scholar). Nevertheless, the literature provides no explanation of how cocaine forms its Michaelis-Menten complex with BChE or why BChE hydrolyzes the unnatural stereoisomer much faster than the natural cocaine although their Km values are similar (9Xie W. Altamirano C.V. Bartels C.F. Speirs R.J. Cashman J.R. Lockridge O. Mol. Pharmacol. 1999; 55: 83-91Crossref PubMed Scopus (120) Google Scholar, 17Gatley S.J. Biochem. Pharmacol. 1991; 41: 1249-1254Crossref PubMed Scopus (90) Google Scholar,18Berkman C.E. Underiner G.E. Cashman J.R. Biochem. Pharmacol. 1997; 54: 1261-1266Crossref PubMed Scopus (15) Google Scholar). Addressing these questions with a computational approach should reduce the need to construct and evaluate many mutant enzymes in a random search for an effective cocaine hydrolase. We chose to predict the Michaelis-Menten complex of BChE with (+)- and (−)-cocaine via docking and molecular dynamics (MD) simulation studies. We reasoned that the overall hydrolysis rate constants (k3) for both natural and unnatural cocaine molecules are relatively low as compared with that for the good substrate, butyrylcholine. It was therefore plausible to assume that the binding affinities of the cocaine molecules would approximately reflect theirKm values. An advantage of this strategy was that predictions of the Michaelis-Menten complexes avoided computationally intensive ab initio calculations to address bond formation, as would have been required in predicting transition state enzyme-substrate complexes. In addition, Michaelis-Menten complexes offer crucial information regarding the initial stage of substrate binding that would not be available from the transition state complexes. From this information, we could compare calculated intermolecular interaction energies and Km values. A correlation between experimental and calculated values would provide insights into the catalytic mechanisms for hydrolysis of cocaine isomers by BChE. The Michaelis-Menten complexes would also provide a basis for predicting transition state complexes via ab initio and potential of mean force calculations (19Tobias D.J. Brooks C.L. J. Chem. Phys. 1988; 89: 5115-5127Crossref Scopus (86) Google Scholar, 20Pang Y.P. Miller J.L. Kollman P.A. J. Am. Chem. Soc. 1999; 121: 1717-1725Crossref Scopus (47) Google Scholar), which in turn should help identify active site residues that hinder formation of the transition state (−)-cocaine-BChE complex. Here, we report the following: 1) the predicted Michaelis-Menten complexes of BChE liganded with (−)- and (+)-cocaine stereoisomers, respectively, 2) correlation of the calculated intermolecular interaction energies of such complexes with experimentally determined Km values, and 3) pH dependence studies in support of the predicted complexes and the structural factors responsible for different hydrolysis rates of cocaine stereoisomers. The natural and unnatural cocaine structures in the protonated state (see Fig. 1) were built by employing the PREP, LINK, EDIT, PARM, and SANDER modules of the AMBER 5.0 program (21Pearlman D.A. Case D.A. Caldwell J.W. Ross W.S. Cheatham III, T.E. Debolt S. Ferguson D. Seibel G. Kollman P.A. Comput. Phys. Commun. 1995; 91: 1-41Crossref Scopus (2736) Google Scholar) with the force field by Cornell et al. (22Cornell W.D. Cieplak P. Bayly C.I. Gould I.R. Merz Jr., K.M. Ferguson D.M. Spellmeyer D.C. Fox T. Caldwell J.W. Kollman P.A. J. Am. Chem. Soc. 1995; 117: 5179-5197Crossref Scopus (11765) Google Scholar) and additional force field parameters provided in the supporting information. The RESP charges of these molecules were generated by calculating electrostatic potentials using the GAUSSIAN 98 program (23Frisch M.J. Trucks G.W. Schlegel H.B. Gill P.M.W. Hohnson B.G. Robb M.A. Raghavachari K. Al-Laham M.A. Zakrzewski V.G. Ortiz J.V. Foresman J.B. Cioslowski J. Stefanov B.B. Nanayakkara A. Challacombe M. Peng C.Y. Ayala P.Y. Chen W. Wong M.W. Andres J.L. Replogle E.S. Gomperts R. Martin R.L. Fox D.J. Binkley J.S. Defrees D.J. Baker J. Stewart J.P. Head-Gordon M. Gonzales C. Pople J.A. Gaussian 98. Gaussian, Inc., Pittsburgh, PA1999Google Scholar) with the HF/6–31G*//HF/6–31G* method followed by a two-stage fitting using the RESP module of the AMBER 5.0 program (24Cieplak P. Cornell W.D. Bayly C. Kollman P.A. J. Comput. Chem. 1995; 16: 1357-1377Crossref Scopus (895) Google Scholar). All such charges are available in the supporting information. The BChE structure was modified from a theoretical 3D structure of human BChE, derived by homology modeling on the basis of the x-ray structure of Torpedo acetylcholinesterase and confirmed by site-directed mutagenesis studies (10Harel M. Sussman J.L. Krejci E. Bon S. Chanal P. Massoulie J. Silman I. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 10827-10831Crossref PubMed Scopus (292) Google Scholar). The modification procedure included: 1) protonation or deprotonation of the Arg, Lys, Asp, Glu, His, and Cys residues; 2) addition of counter ions to neutralize the charged residues; and 3) an energy minimization of the resulting structure. To determine the protonation state, all Arg, Lys, Asp, Glu, His, and Cys residues were visually inspected. Asp and Glu were treated as deprotonated unless they were located in a hydrophobic environment. One Na+ cation was placed in the vicinity of a deprotonated, anionic residue if this residue was more than 8 Å away from a cationic residue. Arg and Lys were treated as protonated unless they were surrounded by hydrophobic residues. One Cl−anion was introduced next to the protonated, cationic residue if this residue was more than 8 Å away from an anionic residue. His was treated as protonated if it was less than 8 Å away from an acidic residue, whereas the His438 residue, which was next to an acidic residue but constituted the catalytic triad, was treated as neutral. In the structure of the neutral His, one proton was attached to the δ nitrogen atom of the imidazole ring if the resulting tautomer formed more hydrogen bonds in the protein; otherwise the proton was attached to the ε nitrogen atom. Cys was treated as deprotonated when it formed a disulfide bond. The location of every counter-ion was determined by an energy minimization, with a positional constraint applied to all atoms of the system except for the counter-ion. Such energy minimizations were performed with a nonbonded cutoff of 8 Å and a dielectric constant of 1.0. Specifically, in the BChE structure, Cys65, Cys92, Cys252, Cys263, Cys400, and Cys519 were deprotonated; His77 and His423 were protonated; His126, His214, His207, and His438 were assigned as HID (Nδ-H) tautomer, whereas His372 was assigned as HIE (Nε-H); all the Glu and Asp residues were deprotonated except that Glu441 was treated as neutral; and all the Arg and Lys residues were deprotonated. Arg40, Arg135, Arg138, Arg219, Arg465, Arg470, Arg509, Arg515, Lys9, Lys12, Lys44, Lys51, Lys60, Lys105, Lys131, Lys180, Lys190, Lys248, Lys313, Lys314, Lys458, Lys494, Asp2, Asp70, Asp87, Asp91, Asp268, Asp295, Asp301, Asp324, Asp375, Asp391, Asp454, Glu80, Glu276, Glu308, Glu363, and Glu404 were each neutralized by a nearby counter-ion (Na+ or Cl−), respectively. In the MD simulation, to neutralize the substrate-bound BChE structure, one Na+ ion was added at the center of the mouth of the catalytic gorge of BChE, and its final location was determined by energy minimization. In the docking studies, the two Na+ions added to neutralize Glu197 and Glu325 were removed because they interfered with access of cationic substrate to the active site. Conformational searches were performed for (+)- and (−)-cocaine molecules employing the CONSER program (devised by Y.-P. Pang). This program first generated conformations by specifying all discrete possibilities at 60° of arc increment in a range of 0 to 360 for five rotatable torsions of cocaine specified in Fig. 1. It then optimized such conformers with the RESP charges and the Cornell et al.(22Cornell W.D. Cieplak P. Bayly C.I. Gould I.R. Merz Jr., K.M. Ferguson D.M. Spellmeyer D.C. Fox T. Caldwell J.W. Kollman P.A. J. Am. Chem. Soc. 1995; 117: 5179-5197Crossref Scopus (11765) Google Scholar) force field. Afterward, it performed a cluster analysis to delete duplicates. These duplicates include those caused by C2 symmetry of the phenyl ring. In the cluster analysis, two conformers were judged different if at least one of the defined torsions differed by more than 30° of arc. The chiralities of the molecules were preserved during energy minimizations by applying constraints on the chiral atoms and their attached atoms. All docking studies were performed by using the EUDOC program (25Pang Y.P. Kozikowski A.P. J. Comput.-Aided Mol. Des. 1994; 8: 669-681Crossref PubMed Scopus (119) Google Scholar, 26Perola E. Xu K. Kollmeyer T.M. Kaufmann S.H. Prendergast F.G. Pang Y.P. J. Med. Chem. 2000; 43: 401-408Crossref PubMed Scopus (143) Google Scholar). 2Y. P. Pang, E. Perola, K. Xu, and F. G. Prendergast, submitted for publication.This program systematically translates and rotates a ligand in a putative binding pocket of a receptor to search for energetically favorable orientations and positions of the ligand. A box is defined within the binding pocket to confine the translation of a ligand. The energy used to judge the preferred orientation and position of the ligand is termed intermolecular interaction energy and is defined as the potential energy of the complex relative to the potential energies of the ligand and receptor in their free states. The potential energies are calculated with the additive, all atom force field by Cornellet al. (22Cornell W.D. Cieplak P. Bayly C.I. Gould I.R. Merz Jr., K.M. Ferguson D.M. Spellmeyer D.C. Fox T. Caldwell J.W. Kollman P.A. J. Am. Chem. Soc. 1995; 117: 5179-5197Crossref Scopus (11765) Google Scholar). A distance-dependent dielectric function was used to calculate the electrostatic interactions (28McCammon J.A. Gelin B.R. Karplus M. Nature. 1977; 267: 585-590Crossref PubMed Scopus (1429) Google Scholar). No cutoff for steric and electrostatic interactions was used in calculating the intermolecular interaction energies. In this work, a box of 5.5 × 4.0 × 10.0 Å3 was defined and surrounded by residues Trp82, Ile442, Glu197, Tyr128, Gly439, His438, Met437, Ser198, Gly115, Gly116, Gly117, Thr120, Gln119, Asn68, Val288, Asp70, Tyr440, Ala328, Phe329, Tyr332, Gly333, Trp430, Leu286, Phe398, and Trp231 in the active site of BChE. The translational and rotational increments were set at 1.0 Å and 10o of arc, respectively. All MD simulations were performed by employing the AMBER 5.0 program with the Cornell et al. (22Cornell W.D. Cieplak P. Bayly C.I. Gould I.R. Merz Jr., K.M. Ferguson D.M. Spellmeyer D.C. Fox T. Caldwell J.W. Kollman P.A. J. Am. Chem. Soc. 1995; 117: 5179-5197Crossref Scopus (11765) Google Scholar) force field and additional force field parameters available in the supporting information. The MD simulation used 1) the SHAKE procedure for all bonds of the system (NTC = 3 and NTF = 3) (29Ryckaert J.P. Ciccotti G. Berendsen H.J.C. J. Comput. Phys. 1977; 23: 327-341Crossref Scopus (17655) Google Scholar); 2) a time step of 0.1 fs (DT = 0.001); 3) a dielectric constant, ε, of 1.0 (IDIEL = 1.0); 4) the Berendsen coupling algorithm (NTT = 1) (30Berendsen H.J.C. Postma J.P.M. van Gunsteren W.F. Di Nola A. Haak J.R. J. Chem. Phys. 1984; 81: 3684-3690Crossref Scopus (24698) Google Scholar); 5) the Particle Mesh Ewald method (31Darden T.A. York D. Pedersen L. J. Chem. Phys. 1993; 98: 10089-10092Crossref Scopus (22291) Google Scholar) used to calculate the electrostatic interactions (BOXX = 91.4869, BOXY = 77.6960, BOXZ = 69.3721, ALPHA = BETA = GAMMA = 90.0, NFFTX = 81, NFFTY = 81, NFFTZ = 64, SPLINE_ORDER = 4, ISCHARGED = 0, EXACT_EWALD = 0, DSUM_TOL = 0.00001); 6) the nonbonded atom pair list updated at every 20 steps (NSNB = 20); 7) a distance cutoff of 8.0 Å used to calculate the nonbonded steric interaction (CUT = 8.0); and 8) defaults of other keywords not described here. The most energetically stable Michaelis-Menten complex of (−)-cocaine-BChE generated by the aforementioned docking study was used as the initial structure for the MD simulation. The complex structure was simulated in a TIP3P (32Jorgensen W.L. Chandreskhar J. Madura J.D. Impey R.W. Klein M.L. J. Chem. Phys. 1982; 79: 926-935Crossref Scopus (31601) Google Scholar) water box with a periodic boundary condition at constant temperature (298 K) and pressure (1 atm) for 1.0 ns (NCUBE = 20, QH = 0.4170, DISO = 2.20, DISH = 2.00, CUTX = CUTY = CUTZ = 8.2, NTB = 2, TEMPO = 298, PRESO = 1, TAUTP = 0.2, TAUTS = 0.2, TAUP = 0.2, NPSCAL = 0, and NTP = 1). The solvated complex consisting of 50,941 atoms was first energy-minimized for 500 steps to remove close van der Waals contacts in the system. The minimized system was then slowly heated to 298 K (10 K/ps; NTX = 1) and equilibrated for 100 ps before a 1.0-ns simulation. The time-average structure of 1000 instantaneous structures of the complex at 1-ps intervals was generated by using the CARNAL module of the AMBER 5.0 program. Hydrolysis of the two cocaine stereoisomers by BChE was measured over a range of substrate concentrations between 0.3 × Km and 3 ×Km. Enzymatic reaction of BChE with (−)-cocaine was measured by using a radiometric method developed in the Brimijoin group to achieve adequate sensitivity. Details of assay performance and validation will be published separately. Briefly, [3H](−)-cocaine (222,000 dpm) diluted with varying amounts of unlabeled (−)-cocaine was mixed with BChE in 0.1m sodium phosphate buffer (pH 7.0) to a final volume of 100 μl. To get maximal signal, the reaction was incubated at 37 °C for 1 h. The enzymatic reaction was stopped by addition of 1 ml of 0.05 m HCl, and the product [3H]benzoic acid was extracted with 4 ml of toluene for quantitation by scintillation counting. Reaction with (+)-cocaine at 25 °C was simply monitored by UV spectrophotometry (17Gatley S.J. Biochem. Pharmacol. 1991; 41: 1249-1254Crossref PubMed Scopus (90) Google Scholar), because modest amounts of BChE caused readily measured hydrolysis. Vmax andKm values were calculated by direct nonlinear fitting to the Michaelis-Menten equation using Sigma Plot for Macintosh (Jandel Scientific). To calculate the catalytic rate constant,kcat, Vmax was divided by the concentration of active sites, previously determined by titration with echothiophate or diisopropyl fluorophosphate. Vmax andKm were measured as a function of pH over the range of 5.5 to 8.5 in 0.1 m sodium phosphate buffer (pH values were determined before and after reaction to ensure that conditions were stable). Hydrolysis of (−)-cocaine was measured by the radiometric method described above, whereas that of (+)-cocaine was measured by UV spectrophotometry at 240 nm. Spontaneous hydrolysis rates in the absence of BChE (general base catalysis) were subtracted from the observed rates. Data were fitted to Equations 1 and 2. kcat app =kcat 1+H+Ka(Eq. 1) kcal /Kmapp =kca /Km1+H+2Ka(Eq. 2) These equations assume that only the enzyme-substrate complex with the catalytic His438 in its neutral state can proceed to hydrolysis. In Equation 1, Ka is the equilibrium constant for protons binding to the enzyme-substrate complex. In Equation 2, Ka is the equilibrium constant for protons binding to the free enzyme. Direct estimates of these parameters, along with associated standard errors, were calculated by the curve-fitting routine of Sigma Plot. Altogether, 15 different conformations of (−)-cocaine (derived from a conformational search) were docked to the active site of BChE by utilizing the EUDOC program. The docking results suggested that the ammonium nitrogen atom of (−)-cocaine interacts with Trp82 via cation-pi interaction in the most energetically stable complex (Fig. 2). The distance between the ammonium nitrogen atom and the midpoint of the indole ring of Trp82 was 4.4 Å. The ammonium nitrogen atom of (−)-cocaine also favorably interacted with the anionic residue Glu197, in addition to the electrostatic interactions with other anionic residues of the enzyme. The distance between the ammonium nitrogen atom and the side chain carbonyl carbon atom of Glu197 was 4.3 Å. The phenyl ring of (−)-cocaine engaged in pi-pi interactions with Tyr332 and Phe329(Fig. 2). The distances of the midpoint of the phenyl ring of (−)-cocaine to those of Tyr332 and Phe329 were 5.0 and 5.8 Å, respectively. The intermolecular interaction energy of this complex was −54.1 kcal/mol, with a van der Waals contribution of −35.4 kcal/mol and an electrostatic contribution of −18.7 kcal/mol (TableI). The carbonyl carbon atom of the benzoic ester is 5.9 Å away from the oxygen atom of the catalytic Ser198, indicating that the benzoic ester group must rotate toward Ser198 for hydrolysis. This rotation may involve pivoting about the ammonium nitrogen atom that engaged in cation-pi interaction with Trp82.Table IKinetic constants and relative intermolecular interaction energies of human BChE for (−)- and (+)-cocaine, respectivelyEnzymeKmkcatEEvdwEeleμmmin−1kcal/molkcal/molkcal/mol(−)-Cocaine9.0 ± 0.33.9 ± 0.8−54.1 ± 0.5−35.4 ± 0.5−18.7 ± 0.5(+)-Cocaine8.5 ± 0.56423 ± 24−56.7 ± 0.5−30.0 ± 0.5−26.7 ± 0.5E, total interaction energy; Evdw, interaction energy contributed by van der Waals interactions;Eele, interaction energy contributed by electrostatic interactions. Km andkcat values were derived from UV spectrophotometric observations of reactions run in pH 7.0 sodium phosphate buffer, at 25 °C for (+)-cocaine and 37 °C for (−)-cocaine. Open table in a new tab E, total interaction energy; Evdw, interaction energy contributed by van der Waals interactions;Eele, interaction energy contributed by electrostatic interactions. Km andkcat values were derived from UV spectrophotometric observations of reactions run in pH 7.0 sodium phosphate buffer, at 25 °C for (+)-cocaine and 37 °C for (−)-cocaine. Using the EUDOC program, the active site of BChE was also docked with 15 different conformations of (+)-cocaine derived from a conformational search. The results for the most energetically stable complex with this unnatural cocaine isomer predicted a cation-pi interaction between the ammonium nitrogen atom of (+)-cocaine and Trp82 (Fig. 2). The distance between the ammonium nitrogen atom and the midpoint of the indole ring of Trp82 was 5.1 Å. The ammonium nitrogen atom of (+)-cocaine also favorably interacts with Glu197, in addition to the electrostatic interactions with other anionic residues in the catalytic gorge. The distance between the ammonium nitrogen atom and the side chain carbonyl carbon atoms of Glu197 was 4.4 Å. In addition, the phenyl ring of (+)-cocaine weakly interacted with Tyr332 and Phe329 (Fig. 2). Thus in many respects the Michaelis-Menten complexes of the two cocaine isomers appear similar. However, the distances of the midpoint of the phenyl ring of (+)-cocaine to those of Tyr332 and Phe329 were 6.9 and 7.7 Å, respectively. The distances of the midpoint of the phenyl ring of (−)-cocaine to those of Tyr332 and Phe329 were 5.0 and 5.8 Å, respectively. Thus, these two aromatic residues of BChE have weaker interactions with (+)-cocaine than with (−)-cocaine. The intermolecular interaction energy of the (+)-cocaine-BChE complex was about 2.0 kcal/mol higher than that of the complex with (−)-cocaine: −56.7 kcal/mol, with a van der Waals contribution of −30 kcal/mol and an electrostatic contribution of −26.7 kcal/mol (Table I). The carbonyl carbon atom of the benzoic ester is 6.6 Å away from the oxygen atom of the catalytic Ser198, indicating that the benzoic ester group must rotate toward Ser198 for hydrolysis. This rotation may involve pivoting about the ammonium nitrogen atom engaged in the cation-pi interaction with Trp82. Separate 1.0-ns (1.0 fs time step) MD simulations of the (−)- and (+)-cocaine-BChE complexes were performed for the following reasons. First, although the 3D structure of BChE used in this study had been confirmed by site-directed mutagenesis studies (10Harel M. Sussman J.L. Krejci E. Bon S. Chanal P. Massoulie J. Silman I. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 10827-10831Crossref PubMed Scopus (292) Google Scholar), it had not been tested with MD simulation in water. That kind of simulation was deemed desirable to test self-consistency of the model. In a correct model, there would be no large root mean square deviation between the initial static and time-average structures (a tendency of unfolding) during the MD simulation. A 1.0-ns MD simulation is certainly not long enough to observe folding or unfolding of a protein. However, it is long enough to observe regional backbone conformational changes involving a partial unfolding of a 3D structure that was incorrectly folded by homology modeling. In other words, a "sustained" 1.0-ns MD simulation does not necessarily validate a model, but a failed simulation calls for a trial of alternatives (10Harel M. Sussman J.L. Krejci E. Bon S. Chanal P. Massoulie J. Silman I. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 10827-10831Crossref PubMed Scopus (292) Google Scholar, 11Millard C.B. Broomfield C.A. Biochem. Biophys. Res. Commun. 1992; 189: 1280-1286Crossref PubMed Scopus (31) Google Scholar, 12Felder C.E. Botti S.A. Lifson S. Silman I. Sussman J.L. J. Mol. Graph. Model. 1998; 15: 318-327Crossref Scopus (68) Google Scholar). A second reason for MD simulation is that no alternative conformations of BChE were used in the docking studies to account for the molecular flexibility of BChE. MD simulations of the cocaine-bound BChE complexes can help to convert the most energetically stable conformation of BChE in the substrate-free state to the most energetically stable conformation of BChE in the substrate-bound state. In the present case, comparing all non-hydrogen, protein backbone atoms in the initial (−)-cocaine-BChE complex and the corresponding time-average of 1,000 instantaneous structures saved at 1.0-ps intervals during the 1.0-ns MD simulation, we found a root mean square deviation of 1.4 Å (2,104 matched atoms) (Fig.3). The corresponding root mean square deviation for the (+)-cocaine-BChE complex was 1.5 Å (2,104 matched atoms) (Fig. 3). The only regions that show significant differences between the MD structure and the homology structure are L1 (Ile69–Gly78), L2 (Ala277–Thr284), L3 (Ile356–Ser362), and L4 (Tyr373–Arg381) on the protein surface (Fig.3). These results suggest that the theoretical 3D model of BChE used in the docking studies was reliable. The time-averaged structure of the (−)-cocaine-BChE complex derived from the 1.0-ns MD simulation was consistent with the initial structure derived from the docking study. The only difference between the two structures was that the tilted T-shaped pi-pi interaction between the phenyl ring of (−)-cocaine and Tyr332 in the EUDOC-generated structure was changed to an off-center pi-pi interaction in the time-averaged structure (Fig.4). This conformational change was promoted by
Referência(s)