Arachidonic Acid Oxygenation by COX-1 and COX-2
1999; Elsevier BV; Volume: 274; Issue: 33 Linguagem: Inglês
10.1074/jbc.274.33.22903
ISSN1083-351X
AutoresLawrence J. Marnett, Scott W. Rowlinson, Douglas C. Goodwin, Amit S. Kalgutkar, Cheryl A. Lanzo,
Tópico(s)Steroid Chemistry and Biochemistry
Resumocyclooxygenase protein or the gene that codes for it non-steroidal anti-inflammatory drugs prostaglandin G2 Prostaglandins were discovered in human semen in 1930, but their low concentrations and instability precluded identification for nearly 30 years (for a brief historical review, see Ref. 1Bindra J.S. Bindra R. Prostaglandin Synthesis. Academic Press, New York1977: 7-22Crossref Google Scholar). Once they were identified, it was clear they arose from polyunsaturated fatty acids by a complex series of reactions involving oxygenation, cyclization, and the generation of five chiral centers from an achiral substrate. The mechanism of prostaglandin biosynthesis was outlined in 1967 by Hamberg and Samuelsson (2Hamberg M. Samuelsson B. J. Biol. Chem. 1967; 242: 5336-5343Abstract Full Text PDF PubMed Google Scholar), and the basic tenets have been confirmed in subsequent studies. The key step in their proposed mechanism was the formation of bicyclic peroxides (endoperoxides) as the initial products of polyunsaturated fatty acid oxygenation (Fig. 1). The term cyclooxygenase (COX)1 2The term cyclooxygenase is used to describe the enzyme activity or to refer to the active site for that activity on the protein.was coined to describe the enzyme that carried out this complex chemical transformation, and its role was confirmed by the isolation of prostaglandin endoperoxides in 1973 (3Hamberg M. Samuelsson B. Proc. Natl. Acad. Sci. U. S. A. 1973; 70: 899-903Crossref PubMed Scopus (570) Google Scholar, 4Nugteren D.H. Hazelhof E. Biochim. Biophys. Acta. 1973; 326: 448-461Crossref PubMed Scopus (635) Google Scholar). In addition to catalyzing a fascinating metabolic transformation, COX is an enormously important pharmacological target. Vane reported in 1971 (5Vane J.R. Nat. New Biol. 1971; 231: 232-235Crossref PubMed Scopus (7348) Google Scholar) that non-steroidal anti-inflammatory drugs (NSAIDs) inhibit prostaglandin formation and demonstrated that their relative inhibitory potency in vitro correlates to their anti-inflammatory activity in vivo . This not only explained the beneficial activity of NSAIDs but also their side effects such as gastrointestinal toxicity and bleeding because prostaglandins and related molecules (i.e. thromboxane) are involved in a very broad range of physiological and pathophysiological responses. The importance of these molecules as autocrine and paracrine mediators has been confirmed recently by the phenotypes of mice bearing targeted deletions inCOX genes or prostaglandin receptor genes. The discovery of a second gene (COX-2 ) coding for cyclooxygenase and the demonstration that its protein product is distributed differently from the originally discovered enzyme (COX-1) raised the possibility that some of the beneficial effects of NSAIDs may be separable from their side effects by development of isoform-selective inhibitors (6Fu J.-Y. Masferrer J.L. Seibert K. Raz A. Needleman P. J. Biol. Chem. 1990; 265: 16737-16740Abstract Full Text PDF PubMed Google Scholar, 7Xie W. Chipman J.G. Robertson D.L. Erikson R.L. Simmons D.L. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 2692-2696Crossref PubMed Scopus (1686) Google Scholar, 8Kujubu D.A. Fletcher B.S. Varnum B.C. Lim R.W. Herschman H.R. J. Biol. Chem. 1991; 266: 12866-12872Abstract Full Text PDF PubMed Google Scholar, 9O'Banion M.K. Sadowski H.B. Winn V. Young D.A. J. Biol. Chem. 1991; 266: 23261-23267Abstract Full Text PDF PubMed Google Scholar). This hypothesis has been dramatically validated by the demonstration that selective COX-2 inhibitors are anti-inflammatory and analgesic but lack the gastric toxicity associated with all currently available NSAIDs (10Masferrer J.L. Zweifel B.S. Manning P.T. Hauser S.D. Leahy K.M. Smith W.G. Isakson P.C. Seibert K. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 3228-3232Crossref PubMed Scopus (1290) Google Scholar, 11Simon L.S. Lanza F.L. Lipsky P.E. Hubbard R.C. Talwalker S. Schwartz B.D. Isakson P.C. Geis G.S. Arthritis Rheum. 1998; 41: 1591-1602Crossref PubMed Scopus (490) Google Scholar). Substantial evidence supports the hypothesis that COX oxygenates arachidonic acid by a free radical mechanism (Fig. 1). Thus, COX appears to have co-opted the process that gives rise to isoprostanes to generate prostaglandins. The major differences between COX-catalyzed and spontaneous oxidation of arachidonic acid are the increased rate and high degree of stereochemical control of the enzymatic reaction (1 of 64 possible isomers predominates). Thus, the overall role of COX is rather simple: stereospecifically remove the 13-pro-S -hydrogen and control the stereochemistry of oxygenation. How does it do this? A protein tyrosyl radical appears to be the oxidizing agent that initiates arachidonic acid oxygenation (12Karthein R. Dietz R. Nastainczyk W. Ruf H.H. Eur. J. Biochem. 1988; 171: 313-320Crossref PubMed Scopus (231) Google Scholar). A tyrosyl radical is formed during COX turnover, and although there has been debate over the identity of the spectroscopically detected radicals, they appear capable of oxidizing arachidonic acid (13DeGray J.A. Lassmann G. Curtis J.F. Kennedy T.A. Marnett L.J. Eling T.E. Mason R.P. J. Biol. Chem. 1992; 267: 23583-23588Abstract Full Text PDF PubMed Google Scholar, 14Tsai A.-L. Palmer G. Kulmacz R.J. J. Biol. Chem. 1992; 267: 17753-17759Abstract Full Text PDF PubMed Google Scholar, 15Tsai A.-L. Palmer G. Xiao G. Swinney D.C. Kulmacz R.J. J. Biol. Chem. 1998; 273: 3888-3894Abstract Full Text Full Text PDF PubMed Scopus (81) Google Scholar). The crystal structures of both COX-1 and COX-2 reveal that Tyr-385 is positioned perfectly to react with the fatty acid substrate (Fig.2) (16Picot D. Loll P.J. Garavito R.M. Nature. 1994; 367: 243-249Crossref PubMed Scopus (1151) Google Scholar, 17Luong C. Miller A. Barnett J. Chow J. Ramesha C. Browner M.F. Nat. Struct. Biol. 1996; 3: 927-933Crossref PubMed Scopus (561) Google Scholar, 18Kurumbail R.G. Stevens A.M. Gierse J.K. McDonald J.J. Stegeman R.A. Pak J.Y. Gildehaus D. Miyashiro J.M. Penning T.D. Seibert K. Isakson P.C. Stallings W.C. Nature. 1996; 384: 644-648Crossref PubMed Scopus (1600) Google Scholar). Indeed, the Y385F mutant is not catalytically active and does not oxidize arachidonic acid when it is treated with peroxide (19Shimokawa T. Kulmacz R.J. DeWitt D.L. Smith W.L. J. Biol. Chem. 1990; 265: 20073-20076Abstract Full Text PDF PubMed Google Scholar). Incubation of wild-type enzyme with arachidonate in the presence of nitric oxide quenches the EPR signal of the tyrosyl radical and leads to the formation of nitrotyrosine at position 385 in the protein (20Gunther M.R. Hsi L.C. Curtis J.F. Gierse J.K. Marnett L.J. Eling T.E. Mason R.P. J. Biol. Chem. 1997; 272: 17086-17090Abstract Full Text Full Text PDF PubMed Scopus (163) Google Scholar, 21Goodwin D.C. Gunther M.H. Hsi L.H. Crews B.C. Eling T.E. Mason R.P. Marnett L.J. J. Biol. Chem. 1998; 273: 8903-8909Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar). Protein radicals require an oxidant for their formation, which in most cases is a metal-containing prosthetic group (22Stubbe J.S. van der Donk W.A. Chem. Rev. 1998; 98: 705-762Crossref PubMed Scopus (1364) Google Scholar). COX is a homodimer of 70-kDa subunits that each contain one molecule of heme (16Picot D. Loll P.J. Garavito R.M. Nature. 1994; 367: 243-249Crossref PubMed Scopus (1151) Google Scholar). The iron is ferric in the resting enzyme and is likely thermodynamically incapable of oxidizing Tyr-385 (E 12 = 0.9 V for Tyr⋅ → Tyr and E 12 = −0.2 to +0.2 V for Fe3+ → Fe2+ for most hemes) (22Stubbe J.S. van der Donk W.A. Chem. Rev. 1998; 98: 705-762Crossref PubMed Scopus (1364) Google Scholar, 23Kulmacz R.J. Ren Y. Tsai A.-L. Palmer G. Biochemistry. 1990; 29: 8760-8771Crossref PubMed Scopus (84) Google Scholar). Reaction of the heme of COX with peroxides generates a ferryl-oxo complex analogous to compound I of classic heme peroxidases (24Lambeir A.M. Markey C.M. Dunford H.B. Marnett L.J. J. Biol. Chem. 1985; 260: 14894-14896Abstract Full Text PDF PubMed Google Scholar). The redox potential of such higher oxidation states is typically on the order of +1 V so the compound I of COX is capable of oxidizing Tyr-385. Ruf and co-workers (25Dietz R. Nastainczyk W. Ruf H.H. Eur. J. Biochem. 1988; 171: 321-328Crossref PubMed Scopus (197) Google Scholar) demonstrated some time ago that oxidation of COX with organic hydroperoxides or fatty acid hydroperoxides generates a spectroscopically detectable tyrosyl radical, and they postulated that the tyrosyl radical oxidizes arachidonic acid. Support for this hypothesis is provided by the existence of significant lag phases for the induction of cyclooxygenase activity of Mn-protoporphyrin IX-reconstituted enzyme or site-directed mutants that exhibit diminished rates of reaction with hydroperoxide (26Smith W.L. Eling T.E. Kulmacz R.J. Marnett L.J. Tsai A. Biochemistry. 1992; 31: 3-7Crossref PubMed Scopus (132) Google Scholar). The identity of the hydroperoxide activator has been uncertain. Our laboratory has reported that peroxynitrite, the coupling product of nitric oxide and superoxide anion, is an excellent oxidant for the heme of COX and activates the enzyme even in the presence of concentrations of glutathione peroxidase and glutathione that inhibit activation by fatty acid hydroperoxides (27Landino L.M. Crews B.C. Timmons M.D. Morrow J.D. Marnett L.J. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 15069-15074Crossref PubMed Scopus (398) Google Scholar). Activation is inhibited by superoxide dismutase, which scavenges peroxynitrite or prevents its formation from NO and O⨪2. Lipophilic superoxide dismutase mimetic agents inhibit prostaglandin biosynthesis by cultured mouse macrophages, which is consistent with a role for peroxynitrite in cyclooxygenase activation in intact cells. These findings provide a biochemical link between NO biosynthesis and prostaglandin biosynthesis and may explain the finding that NO synthase inhibitors reduce prostaglandin biosynthesis in inflammatory lesions in vivo (Equation 1) (28Salvemini D. Settle S.L. Masferrer J.L. Seibert K. Currie M.G. Needleman P. Br. J. Pharmacol. 1995; 114: 1171-1178Crossref PubMed Scopus (251) Google Scholar). Peroxynitrite activation of cyclooxygenase may be especially important in activated macrophages because inducible NO synthase and COX-2 are immediate early genes that are dramatically expressed in response to exposure to inflammatory stimuli such as lipopolysaccharide. The identity of the cyclooxygenase activator in non-inflammatory cells remains to be determined. There has been considerable experimental debate about whether the Tyr-385 radical is regenerated at the end of each cyclooxygenase catalytic cycle or requires reoxidation by another peroxidase catalytic cycle (29Bakovic M. Dunford H.B. Biochemistry. 1994; 33: 6475-6482Crossref PubMed Scopus (37) Google Scholar, 30Wei C. Kulmacz R.J. Tsai A.-L. Biochemistry. 1995; 34: 8499-8512Crossref PubMed Scopus (49) Google Scholar). Detailed kinetic investigations confirm observations extant at the onset of the debate that strongly support the regeneration of the catalytic tyrosyl radical at the end of each cycle of arachidonic acid oxidation (4Nugteren D.H. Hazelhof E. Biochim. Biophys. Acta. 1973; 326: 448-461Crossref PubMed Scopus (635) Google Scholar, 31Tsai A.-L. Wu G. Kulmacz R.J. Biochemistry. 1997; 36: 13085-13094Crossref PubMed Scopus (23) Google Scholar, 32Hamberg M. Svensson J. Wakabayashi T. Samuelsson B. Proc. Natl. Acad. Sci. U. S. A. 1974; 71: 345-349Crossref PubMed Scopus (872) Google Scholar). Thus, the final step in each round of arachidonic acid oxygenation is reduction of the peroxyl radical precursor of PGG2 by Tyr-385, which regenerates the Tyr-385 radical for the next round of cyclooxygenase catalysis (Fig.1). This leads to multiple turnovers per activation event and allows the accumulation of PGG2 (4Nugteren D.H. Hazelhof E. Biochim. Biophys. Acta. 1973; 326: 448-461Crossref PubMed Scopus (635) Google Scholar, 32Hamberg M. Svensson J. Wakabayashi T. Samuelsson B. Proc. Natl. Acad. Sci. U. S. A. 1974; 71: 345-349Crossref PubMed Scopus (872) Google Scholar). How does COX ensure that a single stereoisomer of PGG2 is produced from arachidonic acid? One can predict on purely chemical grounds that the enzyme must bind arachidonate in a conformation similar to that illustrated in Fig. 1. Removal of the 13-pro-S hydrogen followed by reaction with O2, serial cyclization, and reaction with the second O2 could occur with minimal motion of the reaction intermediates to produce PGG2 with all the correct stereocenters. We and others have docked arachidonate into the cyclooxygenase active site of sheep COX-1 to test whether such a conformation can be accommodated. The carboxylate was positioned adjacent to Arg-120, which plays a crucial role in binding arachidonate and arylalkanoic acid-type inhibitors (33Bhattacharyya D.K. Lecomte M. Rieke C.J. Garavito R.M. Smith W.L. J. Biol. Chem. 1996; 271: 2179-2184Abstract Full Text Full Text PDF PubMed Scopus (170) Google Scholar), and the 13-pro-S hydrogen was placed near the phenolic hydroxyl of Tyr-385. The ω-end of the fatty acid was inserted into a channel at the top of the cyclooxygenase active site that eventually leads to the surface of the protein. This end of arachidonate straddles the α-helix containing Ser-530, the site acetylated by aspirin. Acetylation by aspirin blocks arachidonate binding to COX-1 (34DeWitt D.L. El-Harith E.A. Kraemer S.A. Andrews M.J. Yao E.F. Armstrong R.L. Smith W.L. J. Biol. Chem. 1990; 265: 5192-5198Abstract Full Text PDF PubMed Google Scholar). The arachidonate-enzyme complex was then minimized to obtain the final conformation displayed in Fig.3. It is clear that the fatty acid substrate can be nicely accommodated within the active site in a conformation expected to yield PGG2. One prediction of this model is that O2 molecules diffuse up the central channel to couple to the solvent-exposed sides of the carbon radical intermediates to form PGG2. This prediction is consistent with the fact that the 13-pro-S hydrogen is on the opposite side of the fatty acid from the direction of both O2coupling reactions. This model was developed using coordinates for sheep COX-1, but a similar model can be developed using the coordinates for mouse COX-2. The active site structures are quite similar for the two forms of the enzyme, with a few exceptions that are detailed below. Despite this similarity, clear differences in substrate specificities exist between the two enzymes. COX-2 appears much more accommodating than COX-1 in that it oxidizes 18 carbon polyunsaturated fatty acids with much higher efficiency than COX-1. Also, COX-2 oxidizes the hydroxyethylamide derivative of arachidonic acid (anandamide) whereas COX-1 does not (35Laneuville O. Breuer D.K. Xu N. Huang Z.H. Gage D.A. Watson J.T. Lagarde M. DeWitt D.L. Smith W.L. J. Biol. Chem. 1995; 270: 19330-19336Abstract Full Text Full Text PDF PubMed Scopus (208) Google Scholar,36Yu M. Ives D. Ramesha C.S. J. Biol. Chem. 1997; 272: 21181-21186Abstract Full Text Full Text PDF PubMed Scopus (350) Google Scholar). Consistent with the latter observation, the R120Q mutant of COX-1 demonstrates a 100-fold increase in K m for arachidonate whereas the corresponding mutant of COX-2 displays aK m for arachidonate that is quite similar to the wild-type enzyme (33Bhattacharyya D.K. Lecomte M. Rieke C.J. Garavito R.M. Smith W.L. J. Biol. Chem. 1996; 271: 2179-2184Abstract Full Text Full Text PDF PubMed Scopus (170) Google Scholar). 3Rieke, C. J., Mulichak, A. M., Garavito, R. M., and Smith, W. L. (1999) J. Biol. Chem. 274,17109–17114. The molecular basis for these differences is not understood. Another difference between the two enzymes is the ability of aspirin-acetylated COX-2 to oxygenate arachidonic acid to 15-(R )-hydroxyeicosatetraenoic acid (37Lecomte M. Laneuville O. Ji C. DeWitt D.L. Smith W.L. J. Biol. Chem. 1994; 269: 13207-13215Abstract Full Text PDF PubMed Google Scholar, 38Xiao G. Tsai A.-L. Palmer G. Boyar W.C. Marshall P.J. Kulmacz R.J. Biochemistry. 1997; 36: 1836-1845Crossref PubMed Scopus (111) Google Scholar). Acetylated COX-1 is unable to carry out this transformation. The importance of the channel at the top of the active site for binding the ω-end of arachidonate was recently confirmed by mutagenesis of Gly-533. Bulky substitutions at this position inhibit the oxidation of arachidonic acid but not fatty acids with three less carbons at their ω-end. 4Rowlinson, S. W., Crews, B. C., Lanzo, C. A., and Marnett, L. J. (1999) J. Biol. Chem. , in press. The anticipated pot-of-gold at the end of the rainbow awaiting a selective COX-2 inhibitor triggered a world class race to develop candidate drugs. Some of the efforts have been been successful and are chronicled in several recent reviews (39Marnett L.J. Kalgutkar A.S. Curr. Opin. Chem. Biol. 1998; 2: 482-490Crossref PubMed Scopus (101) Google Scholar, 40Talley J.J. Expert Opin. Therapeutic Patents. 1997; 7: 55-62Crossref Scopus (27) Google Scholar, 41Prasit P. Riendeau D. Annu. Rep. Med. Chem. 1997; 32: 211-220Crossref Scopus (96) Google Scholar, 42Munroe D.G. Lau C.Y. Chem. Biol. 1995; 2: 343-350Abstract Full Text PDF PubMed Scopus (29) Google Scholar). The most extensively represented class of inhibitors is diarylheterocycles; other classes of inhibitors include acidic sulfonamides, indomethacin analogs, zomepirac analogs, and di-t -butylphenols. All appear to be slow, tight binding inhibitors in which the selectivity is manifest in the second step (Equation 2) (43Copeland R.A. Williams J.M. Giannaras J. Nurnberg S. Covington M. Pinto D. Pick S. Trzaskos J.M. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 11202-11206Crossref PubMed Scopus (409) Google Scholar). This slow step involves the conversion of the (E ·I) complex to the (E *·I) complex, in which the inhibitor is bound more tightly to the enzyme. Formation of the E *·I occurs in seconds to minutes and may reflect the induction of a subtle protein conformational change. The time-dependent change is not associated with covalent modification for most inhibitors. Aspirin is the only COX inhibitor that covalently modifies the protein (44Roth G.J. Stanford N. Majerus P.W. Proc. Natl. Acad. Sci. U. S. A. 1975; 72: 3073-3076Crossref PubMed Scopus (903) Google Scholar). It acetylates Ser-530, which is juxtaposed to Arg-120, and it is more potent against COX-1 than COX-2 (34DeWitt D.L. El-Harith E.A. Kraemer S.A. Andrews M.J. Yao E.F. Armstrong R.L. Smith W.L. J. Biol. Chem. 1990; 265: 5192-5198Abstract Full Text PDF PubMed Google Scholar, 45Van Der Ouderaa F.J. Buytenhek M. Nugteren D.H. Van Dorp D.A. Eur. J. Biochem. 1980; 109: 1-8Crossref PubMed Scopus (135) Google Scholar). Recently, an aspirin-like molecule (acetoxyphenylheptynyl sulfide) was developed that exhibits 20-fold selectivity for COX-2 and acetylates only Ser-530 (46Kalgutkar A.S. Crews B.C. Rowlinson S.W. Garner C. Seibert K. Marnett L.J. Science. 1998; 280: 1268-1270Crossref PubMed Scopus (269) Google Scholar, 47Kalgutkar A.S. Kozak K.R. Crews B.C. Hochgesang Jr., G.P. Marnett L.J. J. Med. Chem. 1998; 41: 4800-4818Crossref PubMed Scopus (120) Google Scholar). Crystal structures of complexes of sheep COX-1, mouse COX-2, and human COX-2 with non-selective and with selective inhibitors have been solved at 3–3.5-Å resolution (16Picot D. Loll P.J. Garavito R.M. Nature. 1994; 367: 243-249Crossref PubMed Scopus (1151) Google Scholar, 17Luong C. Miller A. Barnett J. Chow J. Ramesha C. Browner M.F. Nat. Struct. Biol. 1996; 3: 927-933Crossref PubMed Scopus (561) Google Scholar, 18Kurumbail R.G. Stevens A.M. Gierse J.K. McDonald J.J. Stegeman R.A. Pak J.Y. Gildehaus D. Miyashiro J.M. Penning T.D. Seibert K. Isakson P.C. Stallings W.C. Nature. 1996; 384: 644-648Crossref PubMed Scopus (1600) Google Scholar). Carboxylic acid-containing NSAIDs ion pair with the guanidinium group of Arg-120, which also ion pairs with the carboxylate of arachidonate (18Kurumbail R.G. Stevens A.M. Gierse J.K. McDonald J.J. Stegeman R.A. Pak J.Y. Gildehaus D. Miyashiro J.M. Penning T.D. Seibert K. Isakson P.C. Stallings W.C. Nature. 1996; 384: 644-648Crossref PubMed Scopus (1600) Google Scholar, 48Loll P.J. Picot D. Garavito R.M. Nat. Struct. Biol. 1995; 2: 637-642Crossref PubMed Scopus (468) Google Scholar, 49Loll P.J. Picot D. Ekabo O. Garavito R.M. Biochemistry. 1996; 35: 7330-7340Crossref PubMed Scopus (170) Google Scholar). Site-directed mutagenesis of the arginine residue in COX-1 to glutamine or glutamate renders the protein resistant to inhibition by carboxylic acid-containing NSAIDs (33Bhattacharyya D.K. Lecomte M. Rieke C.J. Garavito R.M. Smith W.L. J. Biol. Chem. 1996; 271: 2179-2184Abstract Full Text Full Text PDF PubMed Scopus (170) Google Scholar, 50Mancini J.A. Riendeau D. Falgueyret J.-P. Vickers P.J. O'Neill G.P. J. Biol. Chem. 1995; 270: 29372-29377Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar). Arg-120 is part of a hydrogen-bonding network with Glu-524 and Tyr-355, which stabilizes substrate/inhibitor interactions and closes off the upper part of the cyclooxygenase active site from the spacious opening at the base of the channel, which we call the lobby (Fig. 4) (16Picot D. Loll P.J. Garavito R.M. Nature. 1994; 367: 243-249Crossref PubMed Scopus (1151) Google Scholar). Disruption of this hydrogen-bonding network opens the constriction and enables substrate/inhibitor binding and release to occur (17Luong C. Miller A. Barnett J. Chow J. Ramesha C. Browner M.F. Nat. Struct. Biol. 1996; 3: 927-933Crossref PubMed Scopus (561) Google Scholar). In addition, Tyr-355 sterically hinders the mouth of the COX active site, which accounts for the preferential inhibition exhibited byS -stereoisomers of α-methyl-substituted arylalkanoic inhibitors (e.g. the profens) (51So O.-Y. Scarafia L.E. Mak A.Y. Callan O.H. Swinney D.C. J. Biol. Chem. 1998; 273: 5801-5807Abstract Full Text Full Text PDF PubMed Scopus (88) Google Scholar). Opening and closing of the Arg-120–Glu-524–Tyr-355 constriction may contribute to the time dependence of all COX inhibitors. Structures of COX-inhibitor complexes presumably reflect the (E *·I) complex in which the inhibitor is bound tightly to the enzyme. That COX-2-selective inhibitors, especially diarylheterocycles, bind to regions accessible in COX-2 but not COX-1 is consistent with the hypothesis that these structures reveal the molecular basis for their selectivity. Fig. 4 A demonstrates that the sulfonamide moiety in SC-558 wedges into a hydrophobic "side pocket" of COX-2 bordered by Val-523 (18Kurumbail R.G. Stevens A.M. Gierse J.K. McDonald J.J. Stegeman R.A. Pak J.Y. Gildehaus D. Miyashiro J.M. Penning T.D. Seibert K. Isakson P.C. Stallings W.C. Nature. 1996; 384: 644-648Crossref PubMed Scopus (1600) Google Scholar) and hydrogen bonds to Arg-513 and the peptide bond of Phe-518 (18Kurumbail R.G. Stevens A.M. Gierse J.K. McDonald J.J. Stegeman R.A. Pak J.Y. Gildehaus D. Miyashiro J.M. Penning T.D. Seibert K. Isakson P.C. Stallings W.C. Nature. 1996; 384: 644-648Crossref PubMed Scopus (1600) Google Scholar). A similar hydrophobic side pocket off the main channel in COX-1 is not accessible because of the presence of an isoleucine instead of valine at position 523, which sterically hinders inhibitor approach. Other changes between COX-2 and COX-1 that contribute to rigidification of this side pocket include the substitutions R513H and V434I. The COX-2 mutant V523I is resistant to time-dependent inhibition by diarylheterocycles but not arylalkanoic acid-type NSAIDs (52Gierse J.K. McDonald J.J. Hauser S.D. Rangwala S.H. Koboldt C.M. Seibert K. J. Biol. Chem. 1996; 271: 15810-15814Abstract Full Text Full Text PDF PubMed Scopus (389) Google Scholar, 53Wong E. Bayly C. Waterman H.L. Riendeau D. Mancini J.A. J. Biol. Chem. 1997; 272: 9280-9286Abstract Full Text Full Text PDF PubMed Scopus (130) Google Scholar, 54Guo Q.P. Wang L.H. Ruan K.H. Kulmacz R.J. J. Biol. Chem. 1996; 271: 19134-19139Abstract Full Text Full Text PDF PubMed Scopus (118) Google Scholar). Conversely, the COX-1 mutant I523V is sensitive to time-dependent inhibition by diarylheterocycles (53Wong E. Bayly C. Waterman H.L. Riendeau D. Mancini J.A. J. Biol. Chem. 1997; 272: 9280-9286Abstract Full Text Full Text PDF PubMed Scopus (130) Google Scholar). Movement of Val-523 and insertion of the sulfonamide or methylsulfone moiety into the side pocket may contribute to the time dependence of inhibition by diarylheterocycles. The structural basis for selectivity by indomethacin analogs is dependent on substitutions at the top rather than the side of the cyclooxygenase active site. The structure of human COX-2 complexed with a 4-bromobenzyl indomethacin analog reveals that the 4-bromobenzyl group is in van der Waals contact with Leu-503 at the apex of the COX-2 active site. 5B. M. McKeever, S. R. Pandya, M. D. Percival, M. Ouellet, C. Bayly, G. P. O'Neill, L. Bastien, B. P. Kennedy, M. Adam, W. Cromlish, P. Roy, W. C. Black, D. Guay, and Y. Leblanc, submitted for publication. Position 503 in COX-1 is substituted with phenylalanine, which is not as easily displaced by the bromobenzyl group as leucine. Decreased flexibility at the top of the COX-1 active site may reduce the affinity of the protein for the indomethacin analog, thereby accounting for its COX-2 selectivity. Not all structures of COX-2 inhibitor complexes provide insights into the mechanism of selectivity. The NS-398·COX-2 complex demonstrates that the sulfonamide group ion pairs to Arg-120 in a manner similar to carboxylate-containing NSAIDs rather than inserting into the Val-523 side pocket (55Kurumbail R. Pawlitz J.L. Stevens A.M. Stageman R.A. Gierse J.K. Kobolt C.M. Seibert K. Isakson P.C. Stallings W.C. Honn K.V. Nigam S. Marnett L.J. Dennis E. Eicosanoids and Other Bioactive Lipids in Cancer, Inflammation and Related Diseases. Plenum Publishing Corp., New York1999Google Scholar). However, the reason that NS-398 preferentially interacts with COX-2 is not evident from the structure. Likewise, crystallographic analysis of zomepirac-derived COX-2-selective inhibitors does not provide a rationale for their selective COX-2 inhibition. The zomepirac-acylsulfonamide inhibitor breeches the constriction at the mouth of the COX active site and projects into the sterically uncongested lobby region (Fig. 4 B ) (17Luong C. Miller A. Barnett J. Chow J. Ramesha C. Browner M.F. Nat. Struct. Biol. 1996; 3: 927-933Crossref PubMed Scopus (561) Google Scholar). The sulfonamide moiety of the inhibitor hydrogen bonds to Arg-120, Glu-524, and Tyr-355 in COX-2 in a manner similar to the arylalkanoic acid inhibitors. Because these three residues are common to COX-1 and COX-2, their importance in determining the selectivity of this inhibitor for COX-2 is uncertain. The cyclooxygenase mechanism represents a beautiful marriage of peroxidase chemistry and free radical chemistry. The involvement of protein radicals in arachidonate oxidation is now generally accepted, but there is a dearth of details on the electron transfer that generates them. Furthermore, there are proposals of non-peroxidatic mechanisms for tyrosyl radical generation that need to be explored (56Tang M.S. Copeland R.A. Penning T.M. Biochemistry. 1997; 36: 7527-7534Crossref PubMed Scopus (16) Google Scholar). A static picture of substrate-protein and inhibitor-protein interactions is emerging from crystallographic and mutagenesis studies, but the dynamics of these interactions is poorly understood. Fluorescence quenching should be especially useful for studying inhibitor-COX interactions in real time, and preliminary studies indicate diarylheterocycle binding is more complex than the simple two-step model suggests (57Lanzo C.A. Beechem J. Talley J. Marnett L.J. Biochemistry. 1998; 37: 217-226Crossref PubMed Scopus (34) Google Scholar). Two selective COX-2 inhibitors are now on the market and if they are as safe in the general population as they have been in clinical trials (11Simon L.S. Lanza F.L. Lipsky P.E. Hubbard R.C. Talwalker S. Schwartz B.D. Isakson P.C. Geis G.S. Arthritis Rheum. 1998; 41: 1591-1602Crossref PubMed Scopus (490) Google Scholar), they will represent a major public health advance by reducing mortality from bleeding ulcers. Equally exciting is the potential selective inhibitors and COX knockout mice have for uncovering this enzyme's involvement in a wide range of physiological and pathophysiological responses in human beings (58DuBois R.N. Abramson S.B. Crofford L. Gupta R.A. Simon L.S. Van de Putte B.A. Lipsky P.E. FASEB J. 1998; 12: 1063-1073Crossref PubMed Scopus (2225) Google Scholar). In addition, it is likely that a broader range of structures of COX-2 inhibitors will emerge over the next few years as exemplified by a survey of the recent patent literature. 6Information regarding compounds for which patents have been filed recently is available at the IBM patent server web site. The reference numbers for chromenes, pyranoindoles, and substituted phenylacetic acids are PCT WO9847890, US 5776967, and PCT WO9709977, respectively. Finally, although this review has not covered regulation, this topic remains an extremely important and exciting area of investigation. Despite intense study, the reason for the existence of distinct COX genes (sometimes expressed in the same cell type) is not understood; the source of arachidonic acid for the two enzymes remains incompletely defined, and the mechanism by which glucocorticoids inhibit COX-2 expression is not certain. Elucidating the answer to these questions not only will provide important fundamental knowledge but also may uncover new molecular targets for pharmacological intervention against a range of human diseases. We are grateful to R. Kurumbail and M. Browner for COX-2-inhibitor coordinates and helpful discussions; to B. McKeever and W. Smith for communication of submitted manuscripts; and P. Isakson for helpful discussions. Download .pdf (.13 MB) Help with pdf files
Referência(s)