9. . .n .3. . . . a vhf» fitfi... PE. . :2. , . $.firhfimflafiu .. . . . n u. ‘ ; . . . ; M #3 1%.. .. 3.... Lu a...» ‘ . 4.1%. . . .m... bawmflwe. nnw .. féfi. .5. .3 {Jar 5.5%”.th v“. a. k gm 59.... w . an w :, .. g4 wmeNQA. .St a ,, , ‘ haw .mfi. .9. a V uxfiffiam a. fin. . . . 3? 17.13 {. V . L ax? "r3. . , A . mmfia . b... .4- ...t ‘1 “a 9. r ‘ . ,, “ma. :5 I. . \h‘ . ... ., a. _ fixrfiv m. V . 1‘ . . . , , _. :wmkfimfi Lump“, . . r H .hh; ‘ . . a v ‘ , . ‘ . . . . .. a ‘ . u ERMA” 5‘3 . . v . . . ‘ H , . ‘ V . fa. , . :2. .. .. , ‘ . V . . 5&3”? . , , . It»??? .53 . . :3 , . u : .. .. in... 1‘: u? . ‘ ‘ .. syfiuéwfiy , .nsm‘cfikmn Ltnfifiwk : ., $qu 31min. 1| rim 3...}. 55.3,.”"W... . 2 M. 3...‘ m ¢ i. I . . ,1 s... 3 . Du ‘- \ . 1; t!’ ?. bum .. , 1.? . , ) .: v . . ...\.4h .9. in...) 1:1...- 3. :ixi ‘ 13.2. THESIS 2 2'0?! This is to certify that the dissertation entitled Fatty Acid Substrate Interactions with Cyclooxygenase Active Site Residues in Ovine Prostaglandin Endoperoxide H Synthase-l presented by Elizabeth Dager Thuresson has been accepted towards fulfillment of the requirements for Ph.D. Biochemistry degree in Waxy/AA Major professor Date pv/Ll/fl MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 LIBRARY MIchlgan State Unlverslty PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE ' DATE DUE DATE DUE moo WWI-$869.14 rm' At' ACTIVE $1: FATTY ACID SUBSTRATE INTERACTIONS WITH CYCLOOXYGENASE ACTIVE SITE RESIDUES IN OVINE PROSTAGLANDIN ENDOPEROXIDE H SYNTHASE—l By Elizabeth Dager Thuresson A DISSERTATION Submitted to Michigan State University In partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry 2000 tandem: acad step In the 11m: also ethm: w: 3005. l lR-h SZ.SZ\l lZ.l3 *Y" I . chiral-phase h Hméd iII‘m . 553mm am with both nut acid {Oman QLIllLdi \' ABSTRACT FATTY ACID SUBSTRATE INTERACTIONS WITH CYCLOOXYGENASE ACTIVE SITE RESIDUES IN OVINE PROSTAGLANDIN ENDOPEROXIDE H SYNTHASE-l By Elizabeth Dager Thuresson Prostaglandin endoperoxide H synthases-l and —2 catalyze the conversion of arachidonic acid and two molecules of 02 to prostaglandin 62. This is the committed step in the formation of prostaglandins and thromboxane A2. Prostaglandin synthases also exhibit some lipoxygenase activity producing small amounts of the monohydroxy acids, llR-hydroxy-SZ,BZ,12E,l4Z-eicosatetraenoic acid and 15S/R-hydroxy- SZ,SZ,IlZ,13E-eicosatetraenoic acid. Using thin layer chromatography and reverse- and chiral-phase high performance liquid chromatography, we have examined the products formed from AA and a cyclooxygenase active site mutant V349L ovine prostaglandin synthase-l and have determined the kinetics for the formation of each individual product. With both native and V349L oPGHS-l, the KM values for prostaglandin G; and hydroxy acid formation were different. These results establish that AA can assume at least three catalytically productive arrangements within the cyclooxygenase site of oPGHS-l leading to prostaglandin G2, 11R- and 155- and/or lSR-hydroxyeicosatetraenoic acids, respectively. ICSO values for inhibition of formation of the individual products by the competitive inhibitor, ibuprofen, were determined and found to be the same for a given enzyme form, suggesting that the various substrate/enzyme arrangements occur afier the binding of substrate in the cyclooxygenase active site. I V I .u v :2 iii mum‘tu - l m _. . . 90-9} )4“be OLA-K I u» so .‘v.' '1 };?.‘u1huul(\n I L \ u as: inImCIZI‘f-‘ 3? “actldt‘i‘iJi-C 7mm with at. great as a 3’" Hgiru' " ‘ ° u I \:\LIU'\1.‘ Law 'I‘ -, !VT\I\‘.\ M ‘ Inhiomng 1hr 59 -l . ‘ updmalx alw' 311d Leu534 \x 'L 1 _ .. 1| JAm mkh‘ld‘ ‘r I mQSurabl Cb; We have also performed mutational analyses to determine the roles of residues lining the hydrophobic pocket of the prostaglandin synthase-l cyclooxygenase active site in (1) arachidonate binding and oxygenation and (2) the oxygenation of fatty acid substrates other than arachidonate. Tyr348, when substituted with phenylalanine, exhibited catalytic properties similar to that of the native enzyme, while a leucine substitution resulted in a complete loss of cyclooxygenase activity. These results suggest that interactions between the phenyl ring of Tyr348 are essential for positioning of C-13 of arachidonate for hydrogen abstraction. Substitutions of Va1349 and Leu534, when reacted with arachidonate, resulted in catalytically active enzymes which produced as great as a 20-fold abundance of the monohydroxy acid products, 11R- and 155- hydroxyeicosatetraenoic acids, respectively. These results establish that Va1349 and Leu534 provide hydrophobic interactions with arachidonate which contribute to positioning the substrate such that when hydrogen abstraction occurs the fatty acid is optimally aligned to yield prostaglandin G2 rather than monohydroxide products. Va1349 and Leu534 were also found to play significant roles in oxygenation of fatty acids other than arachidonate. Conversion of Va1349 to alanine caused >800 fold decreases in the efficiencies (Vmax/KM) of oxygenation of 20:3n-6 and 20:5n—3 compared with only a 35% decrease in the efficiency of arachidonate oxygenation. Thus, Va1349 is particularly critical in positioning the carboxyl half of these alternative substrates for proper alignment of C-13. A L534V mutant oxygenated C20 but not C13 substrates, suggesting an important role in positioning of the l8-carbon fatty acids for catalysis. The remaining residues comprising the cyclooxygenase active site channel were determined to provide measurable but lesser contributions to optimizing catalysis. Iiit‘QU-i I ' . aimxtfi I ' ‘ :mthsmv ‘ u ~ arxxmlc, For“ m l - 0- L ..CT."‘C.'\ 0'. LIL' 333 “\‘v n‘ t . LUI. lat. ‘1' H lartzra 1'?" RAILEiUC:C\\' I. T L {.TH-Qn ~uuull‘n II", C an ., H L:- 1' ‘ Admit-"P We. ‘tu: |,l hm? .MUSQQ {lied to the m4 Mflhw “MUD v7»: ~4sz SUn I" hma Rai- Kitten. Q (I) V l I 5’ 5 1.60“? \ paid for ?}~ \‘I Ppiiin 0f m \ aWEQmQ. ‘3 II ACKNOWLEDGMENTS I would first like to thank my advisor, Bill Smith, for sharing his wealth of knowledge with me and teaching me a tremendous amount about science and scientific thinking. I also want to thank him for opening his lab to me when I knew very little and providing me with a great number of scientific opportunities. He has made my journey through this program and, more importantly, my successful completion of this program, possible. For this, I owe him a debt of gratitude. I have enjoyed working with the members of the Smith and DeWitt labs, past and present, for all of their help and humor and color and friendship during what can sometimes be a very intense experience. I am grateful to Joseph Leykam of the MSU Macromolecular Structure Facility for his tireless assistance with the HPLCs, to Dr. Honggao Yan for his invaluable instruction in enzyme kinetics and to Mike Malkowski for providing countless crystallographic figures and help with computer modeling, as well as being a very beneficial scientific sounding board. Also to Julie Oesterle who is worth her weight in gold to the graduate students of Biochemistry, thank you for taking some of the stress out of all of our lives. I would like to thank some of my Biochemistry companions for all of their support and friendship through all the years I’ve been here including Jill, Tonya, Raj, Karen, Colleen and Wayne. Also, I thank my parents for their support and faith, my sister for her sanity and advice and my brother for providing me with the thick skin required for this task. Finally, I never could have done any of this without the love and support of my wonderful husband Magnus and of course, our daughter Anna, who added a sweetness to this part of my life which would have been otherwise absent. lSiOFlAh. LISIOF not I LSTOFABB CHAPTERI:' kmdtt \hih Ham; (filth - Smcd; PtfliSF PffliSI (0371px Actlxc 3 Sam: I l IPh'D'H I‘M“ ”5‘ (Wham CHAPTER 11: ; CICLI )( illil'udug Matti. Result». Dlx‘cd“E ‘.\\L\n‘ I“ CHAPTER III; RESIDI OFARi 1mm“. TABLE OF CONTENTS LIST OF TABLES .................................................................................. vii LIST OF FIGURES ................................................................................ viii LIST OF ABBREVIATIONS ....................................................................... x CHAPTER 1: LITERATURE REVIEW Introduction ................................................................................... l Mobilization of Arachidonic Acid and Prostanoid Biosynthesis ....................... 6 Peroxidase Catalysis by PGHS ........................................................... 10 Cyclooxygenase Catalysis by PGHS ..................................................... 13 Suicide Inactivation ........................................................................ 13 PGHS Structure and Biochemical Properties ........................................... 16 PGHS Crystal Structures .................................................................. 19 Comparison of PGHS—1 and PGHS-2 Crystal Structures ............................. 22 Active Sites of PGHSS ..................................................................... 23 Structural Comparison between PGHSS and other Fatty Acid Binding Proteins..28 Inhibition of PGHSs ....................................................................... 31 Cyclooxygenase Substrates of PGHSs .................................................. 33 CHAPTER II: ARACHIDONIC ACID CONFORMERS IN THE CYCLOOXYGENASE ACTIVE SITE OF OVINE PGHS-l Introduction ................................................................................. 37 Materials and Methods ..................................................................... 39 Results ....................................................................................... 45 Discussion ................................................................................... 60 Acknowledgment ........................................................................... 66 CHAPTER III: THE FUNCTIONS OF CYCLOOXYGENASE ACTIVE SITE RESIDUES IN THE BINDING, POSITIONING AND OXYGENATION OF ARACHIDONIC ACID Introduction ................................................................................. 67 Materials and Methods ..................................................................... 70 Results and Discussion .................................................................... 80 ffiAPlER l\ I REST.) E0139 Intuit, Mater, A .J RN‘H.».‘ 99wa A Imam»- CHAPTER IV: THE FUNCTIONS OF CYCLOOXYGENASE ACTIVE SITE RESIDUES IN SUBSTRATE SPECIFICITY OF PROSTAGLANDIN ENDOPEROXIDE H SYNTHASE-l Introduction ................................................................................ 1 03 Materials and Methods ................................................................... 107 Results and Discussion ................................................................... 11 1 APPENDIX A: Derivations for Kinetic Equations ........................................... 128 REFERENCES ..................................................................................... 13 7 vi Table 1. L351: \fl LIST OF TABLES Table 1. Catalytic constants for the formation of prostaglandin and monohydroxy acid products from arachidonate by native and V349L oPGHS-l ............................................................ 55 Table II. Oligonucleotide primers used for preparing oPGHS-l Mutants. . . . . . . . ....75 Table III. Contacts between Arachidonic Acid (AA) and Cyclooxygenase Active Site Residues .............................................................. 76 Table IV. Interactions Between Leu53l and other Cyclooxygenase Active Site Area Residues ................................................................ 77 Table V. Kinetic Properties and Product Analyses for oPGHS-l Cyclooxygenase Active Site Mutants .......................................... 78 Table VI. Contacts between Dihomo-y-linolenic acid (DHLA) and Cyclooxygenase Active Site Residues ........................................ 1 14 Table VII. Kinetic Properties for Oxygenation of Various Fatty Acids by Native and Mutant PGHS-l and PGHS-2 .................................... 115 Table VIII. Oxygenase Activity for oPGHS-l Cyclooxygenase Active Site Mutants with Various Fatty Acid Substrates ................................. 1 19 vii . put" l 14.n- l. b p 5‘? ’8‘ . Lute -. . f‘ s PfLY'" ‘ I. a. _ . v Figure 4. Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 10. Figure l 1. Figure 12. Figure 13. Figure 14. Figure 15. Figure 16. LIST OF FIGURES The arachidonate cascade ......................................................... 2 Biosynthetic pathway for the formation of prostanoids derived from arachidonic acid .................................................... 4 Model for peroxide-dependent activation of cyclooxygenase activity of PGH synthase via formation of an intermediate tyrosyl radical ......... l 1 Mechanism model for conversion of arachidonic acid to PGG2 by the cyclooxygenase activity of PGHSs ......................................... 14 Comparison of the PGHS-l and PGHS-2 proteins ............................ 17 Crystal structure of the oPGHS-l dimer ....................................... 20 Model of the active site for ovine PGHS-1 .................................... 24 Cyclooxygenase active site from the Coy-heme oPGHS-l crystal structure complexed with arachidonic acid ........................... 26 Autoradio gram of thin layer chromatograms of products formed from [l-MC] arachidonate by native and V349L oPGHS-l ................. 46 Chiral HPLC of the methyl esters of the l 1- and lS-HETES produced by native oPGHS-l ................................................... 49 Chiral HPLC of the methyl esters of the 15-HETES produced by native oPGHS-l at increasing arachidonate concentrations ................. 50 Effect of arachidonic acid concentration on the formation of PGGZ, 1 l-HETE and 15-HETE by native oPGHS-l ................................. 52 Effect of arachidonic acid concentration on the formation of PGGZ, and 15-HETE by V349L oPGl-lS-l ............................................. 53 Autoradiogram of thin layer chromatograms from IC50 determinations for inhibition of native (upper) and V349L (lower) oPGHS-l by ibuprofen ............................................................................ 57 Inhibition of oPGHS-l by ibuprofen ........................................... 58 Inhibition of V349L oPGHS-l by ibuprofen .................................. 59 viii . .Lua D F‘r'm 1Q A .nhnS l e ‘71 u" ‘f! *# (1. I.) 1” Figure 17. Figure 18. Figure 19. Figure 20. Figure 21. Figure 22. Figure 23. Figure 24. Figure 25. Kinetic schemes for the binding of arachidonate to a single enzyme form and for the binding of arachidonate to a mixture of different enzyme forms ..................................................................... 63 Schematic representation of interactions between arachidonic acid and amino acid residues lining the cyclooxygenase active site channel ........ 69 Western blot analysis of Va1349 oPGHS-l mutant proteins ................ 87 Thin layer chromatogram of products formed from [l-MC] arachidonic acid by native and mutant oPGHS-l enzymes .................. 9O Chiral HPLC analysis of the methyl ester of 1 l-HETE produced by V349A oPGI-IS-l .............................................................. 92 Effect of arachidonic acid concentration on the formation of PGGz and l l-HETE by V349A oPGHS-l .................................... 93 Chiral HPLC analysis of the methyl ester of the 15-HETE produced by L534A oPGHS-l .............................................................. 96 Chiral HPLC analysis of the methyl ester of the 15-HETE produced by S53OT oPGHS-l ............................................................. 100 Structures of and products formed from various n-6 and n-3 fatty acid substrates of PGHSs ....................................................... 105 ix fiHS :?CHSl “CH8: COX . ‘“ht “L lint g 3313 c 33m} (1 l iiHPETE 1 ELHHTre 1 lHODE s EEHODE ] 1m} 1 K“ . R K; POX ELI. \SAJD HPLC PGHS oPGHS-l hPGHS-2 COX. AA DHLA 20:4n-6 20:3n-6 20:5n—3 18:2n-6 18:3n-3 l l-HETE l l-HPETE lS-HETE lS-HPETE 12-HHTre 9-HODE l3-HODE PBS SDS-PAGE TMPD Aliphatic and aromatic carbons LIST OF ABBREVIATIONS prostaglandin endoperoxide H synthase ovine prostaglandin endoperoxide H synthase-l human prostaglandin endoperoxide H synthase-2 cyclooxygenase arachidonic acid dihomo-y-linolenic acid arachidonic acid dihomo-y-linolenic acid eicosapentaenoic acid linoleic acid a—linolenic acid 1 l-hydroxy-5,8, 12, l 4-eicosatetraenoic acid 1 l-hydroperoxy-5,8, l 2, 14-eicosatetraenoic acid 1 5-hydroxy-5,8,1 1,1 3-eicosatetraenoic acid 1 5-hydroperoxy-5,8,l l, l 3-eicosatetraenoic acid 12-hydroxy-5,8, 1 O-heptadecatrienoic acid 9-hydroxy-10, l 2-octadecadienoic acid l3-hydroxy-9, 1 l -octadecadienoic acid maximum velocity of a reaction Michaelis-Menton constant dissociation constant for enzyme-inhibitor complex inhibitor concentration for 50% inhibition peroxidase cytosolic 85 kDa phospholipase A2 secretory phospholipase A2 non-steroidal anti-inflammatory drug high performance liquid chromatography thin layer chromatography epidermal grth factor membrane binding domain thromboxane prostaglandin Dulbecco’s modified Eagle medium phosphate buffered saline sodium dodecyl sulfate — polyacrylamide gel elecrophoresis 3 ,3 ,3 ’ ,3 ’-tetramethylenephenyledediamine CA CZ CG CB CG CB CE CD CD v- o .l P.\‘\..‘._ k .. .. _ , ‘ .!‘\ V ',.n H SI... . \“\..J‘\ v» {16"}.1. 1‘ n‘ I . shmu'g' ""1“ § : Ag“ '\ vii . ‘ “1 q f‘ms‘l I "\w l 5.6. v. u CHAPTER I LITERATURE REVIEW Introduction Prostaglandin endoperoxide H sythases-l and —2 (PGHS-l and —2) or cyclooxygenases (COX-1 and -2) are the key enzymes in the synthesis of prostanoids from arachidonic acid [1-3]. Prostanoids, along with leukotrienes (LTs) and epoxyeicosatetraenoic acids (EETs) belong to a family of oxygenated 20-carbon lipid signaling molecules called eicosanoids (Figure l). The prostanoids (prostaglandins, prostacyclins and thromboxanes) act in an autocrine or paracrine fashion through the actions of G-protein linked receptors to regulate a broad range of physiological processes including vascular homeostasis, platelet aggregation, renal water homeostasis, inflammation and reproductive functions [4-6]. They have also been implicated in pathologies such as the development of colon cancer [7, 8] and Alzheimer’s disease [9]. Prostanoid synthesis is initiated by the interaction of various hormones with their cell surface receptors, resulting in the activation of one or more phospholipase Azs (PLAzs). PLAzs mobilize arachidonic acid from phospholipid stores, after which free arachidonate can then be acted upon by PGHSs and converted to various prostanoid products. There are two PGHS isozymes, known as PGHS-1 and PGHS—2. Both PGHSs catalyze the first two steps in the biosynthesis of prostaglandins via a cyclooxygenase activity in which arachidonic acid is oxidized to the hydroperoxy endoperoxide PGGz and P.~ T'} yin lkxm'iu“: “$031M Figure 1 U.“ . 71 ~ ( . (I). .1} Mac /V\/=\/=\/=\/=\/ 0m“ ARACHIDONIC ACID 202 PGH Synthase 02 5,12, and 15- Oz P-450 (Cyclooxygenase) Lipoxygenases Epoxygenase Prostaglandins (PGs) Leukotrienes (LTs) Epoxy Acids (EETS) Thromboxanes (I'Xs) Hydroxy Acids (HETEs) Hydroxy Acids Figure l. The arachidonate cascade. Arachidonic acid is metabolized by three different enzymes systems: the cyclooxygenase, the lipoxygenase and the P450 epoxygenase pathways 2mm .; TIN mo 3;: ‘2}. JCIIRZUu Fifi-033d t‘ ‘ ' ‘Q- ,. ‘ ' Fiiixu-IIFAULJ‘. l l I .‘ sizzmt \m d. . {0:511:ij c: \ int: 1 inc placid .1: :R\.\ and p' u“; iii-L‘lmdiwn. W by PC. H u, “M". 4': M mL'Thh “man. V Hing an T358]. The b R l . T‘n - ‘ ”d- r- a» ‘Uxi'n. [1‘ ( W» -\h width} r l)‘ a peroxidase activity by which PGG2 is reduced to the hydroxy endoperoxide PGH2. These two activities occur at distinct albeit interacting sites on the PGHS molecule and both activities require heme [IO-13]. PGH2, the final product of PGHSs, is then transformed by a range of enzymes and nonenzymatic mechanisms into the primary prostanoids, prostaglandins (PG) E2, F20, D2, 12 and thromboxane (Tx)A2 [14-16] (Figure 2). PGHS-l and —2 are considered to be the constitutive and inducible PGHSs, respectively. Despite numerous catalytic and structural similarities, PGHS-l and —2 exhibit very different patterns of expression in mammalian tissues. PGHS-1 is expressed constitutively in nearly all tissues [17] and forms prostanoids central to several housekeeping functions including water reabsorption in the kidney, vascular homeostasis and platelet aggregation. PGHS-2 is normally undetectable in most cells, but both mRNA and protein can be rapidly induced in many cell types upon treatment with inflammatory cytokines, growth factors and tumor promoters [17]. Thus, prostanoids formed by PGHS-2 have been implicated in the inflammatory response as well as in cell replication and differentiation. Other notable differences between the two isozymes include membrane binding domains of different sequence and an 18-amino acid cassette containing an additional glycosylation site found only near the C-terminus of PGHS-2 [18]. The cyclooxygenase activity of PGHSs is importantly the therapeutic target of nonsteroidal anti-inflammatory drugs (N SAIDs) including aspirin, ibuprofen and naproxen. [19, 20]. These drugs interact with the cyclooxygenase active site competing with arachidonate for binding. In the case of aspirin, the IC50 is about 20mM in the HO' HORMONE / ( CELI- MEMBRANE PHOSPHOLIPID(PI,PC,PE), / D TRIGchERIOE, OR I “Hum" CHOLESTEROL ESTER I i LIPASE(S) ’ __ ARACHIDONIC ACID CYCLOOXYGENASE PGH SYNTHASE 0' - \ PEROXIDASE 0‘ - \ PGH2 6” H 0‘ / \HOOC SYNTHASES , PGI2 few / Q, PG°2 bH 6H WW I O OH ~./'=\/\/ COOH O ’\=/\/\ coo” HO' PGE2 I0H b" "A2 Figure 2. Biosynthetic pathway for the formation of prostanoids derived from arachidonic acid. resent“ 01 1" 5:17.11 15 acct .l ‘ tultKfl‘ * .‘CT‘JA v x ' ' . .rxfix’fi ed. hc’.‘ . .._,,_ I .xx..t"~\ iidi 1.17 presence of lOOuM arachidonate [21]. However, an important time-dependent effect of aspirin is acetylation of the hydroxyl group of Ser530 and concomitant irreversible cyclooxygenase inactivation [22, 23]. Although all residues important for catalysis are conserved between PGHS-l and PGHS-2, the PGHS-2 cyclooxygenase active site is somewhat larger than that of PGHS-l and has recently become the target for development of novel drugs such as celecoxib and rofecoxib which selectively inhibit PGHS-2 and appear to lack the gastrointestinal side effects associated with more classical NSAIDS. ' .. .l' . ' pagmhuspd ~ Y"“.".'\ P‘..\ 'V‘ .ms.\?..\ \? ... Z‘Ci'lifi a fie-'1» .-' \‘HSIS tdk’ l‘ lii‘li ‘ ‘I tdnhirfma?c§ \ ._‘ t Mobilization of Arachidonic Acid and Prostanoid Biosynthesis The biosynthesis of prostaglandins requires the release of arachidonic acid from phospholipid stores. Arachidonic acid is preferentially found at the sn-2 position of glycerophospholipids and is cleaved by one of a number of phospholipase A2s (PLA2s) to generate a lysophospholipid and free arachidonate available to PGHSs and lipoxygenases. The number of identified PLA2s is growing steadily (there are at least 16) and all are classified according to molecular weight, amino acid sequence, calcium dependence and structural homology [24, 25]. Despite their common function of releasing arachidonic acid from membranes, the physiological roles of different PLA2s appear to be distinct, and different PLA2s may be independently coupled to PGHS-1 or PGHS-2 [26]. Cytosolic PLA2 (called cPLA2 or Group IV PLA2) is a large (85-100 kDa) enzyme that, when activated, preferentially hydrolyzes arachidonate from the sn—2 position of phosphoglycerides. Agonist-induced increases in intracellular calcium levels (uM levels) activate cPLA2 by promoting its association with membranes [27, 28] and cPLA2 has been shown to mediate honnonally-induced arachidonic acid release. Activity of cPLA2 is also potentiated by phosphorylation of Ser-SOS by kinases of the mitogen- activated protein kinase cascade [29-31]. cPLA2 can be activated by a wide variety of stimuli including the proinflammatory cytokines interleukin 1 (IL-1) [32, 33], tumor necrosis factor (TNF) [34], thrombin [32, 35] and lipopolysaccharide (LPS), among others. When cells are stimulated by agents that mobilize intracellular Ca2+, cPLA2 translocates from the cytosol to the membrane where it interacts with phospholipid HEM? “I (1' . y ' lfuii)‘ but “is“ [8.11%. “C. \' ‘sfl ~ ‘F I " .~t..“"..ll.\ 3.; R. I‘. I l and at the up. ape? effluent: The :3. nos: lt‘t‘t‘nil} ' final 13‘ 3y “PM: Lil: renaming pl‘. modifitrriun. ii is L Winding 3L indicating th. il'll in gen“ dimming or .4 _ «Md-cm (n substrates [36]. This translocation is mediated through the CaLB or C2 domain, a calcium-dependent phospholipid-binding regulatory domain of the phospholipase. The other major group of PLA2s comprises the l3-18kDa secretory PLA2s (sPLA2s or Groups I, IIA, IIC, V and X). These enzymes also require calcium (mM levels) but use it for the catalytic mechanism of arachidonate release rather than for membrane association. The mammalian sPLA2s which have been characterized (Groups IB, IIA, IIC, V and X) are distinguished by their number of disulfide bridges, expression patterns and physiological roles [25]. These enzymes are not specific for arachidonic acid at the sn-2 position and will hydrolyze linoleic acid from the sn-2 position with equal efficiency. The calcium independent PLA2s (iPLA2 or Groups VI, VII and VIII PLA2s) have most recently been identified, but only the Group VI enzyme has been characterized in detail [37, 38]. It shares some characteristics with both the sPLA2s and others with the cPLA2. Like the sPLA2s, the iPLA2 exhibits no substrate specificity for arachidonate- containing phospholipids and appears not to be subjected to posttranslational covalent modifications, but like the cPLA2 the iPLA2 is large (80-85kDa) and cytosolic [3 8]. It is unclear as to which group of phospholipases is primarily responsible for providing arachidonate to PGHSs. However, there is a growing body of evidence indicating that cPLA2 acts in concert with various sPLA2s (particularly Groups V and 11A) in generating arachidonic acid for oxygenation by PGHS-l, PGHS—2 or both, depending on the stimulation and cell type [26] [39]. With the PLA2s as well as the two PGHS isozymes, the contribution to overall prostaglandin synthesis appears to be largely dependent on cell type and stimulus used. Activation of the cPLA2 appears to be the .I ) "'- QdeLlL’S .1:le «a v’ ‘ O W it...3.l\‘fl- ..- lake'l‘flfid 3'7 |.L | «.inL “1“ ‘ 0: i T ."‘l ofstzrnflatiur. for hours :45- mzjom of d.’ m“. an MA; CPLi I“. 4 appears to be rachidunstc mmnmd ‘m .hi l‘ lab ilIlL’TLiiC C '_, v 1,) (’_“ . t4) (3 J... 1.; Olive “GEMS of P( principal event [40-42] and may be mediated by several signals such as phosphorylation cascades and intracellular Ca2+ levels [28, 43]. The cytosolic phospholipase, upon activation, translocates to the perinuclear and endoplasmic reticular membranes [27, 28], liberating arachidonic acid where PGHSs are co-localized [44]. cPLA2 has been shown to play an important role in immediate eicosanoid biosynthesis occurring within minutes of stimulation as well as in the delayed, PGHS-2-dependent prostanoid generation lasting for hours [45-48]. In cells not expressing sPLAzs, cPLA2 appears to account for the majority of arachidonate mobilized during cellular activation. However, in cells that contain sPLA2, the bulk of arachidonate release appears to be mediated by sPLA2, not the cPLA2 [41, 49, 50], however, sPLA2 action, particularly that of Groups IIA and V, appears to be dependent on an active cPLA2 [41, 45, 49, 50]. Thus cPLA2 is key for arachidonate signaling even when sPLA2 is the major source of arachidonate used for eicosanoid biosynthesis. sPLA2s, once secreted, associate with the outer surface of cells and liberate arachidonate which can be used by surrounding cells to produce eicosanoids [30,39,41,45,51] Once the arachidonate is released, it is then converted to PGH2 through the actions of PGHSs. Release of PGH2 intracellularly allows for subsequent interaction with relevant synthases for conversion to prostaglandin products. Cell-specific expression of the different prostaglandin synthases dictates the final products formed (Figure 2). Immediately following their synthesis, prostaglandins are released from the cell via at least one known prostaglandin transporter [52, 53] and act in either an autocrine or paracrine fashion on cells via several specific G protein-linked prostanoid receptors on the plasma membrane [54]. These receptors are specific for PGE, PGI, PGD, PGF and vi ‘ ‘ F ‘2'.“ 1" u“ “i ., \ .u ‘ I TU"; 'l‘ ‘A t... I‘ .b. - Ab— ‘ ' I i .i.\ la; 1 .15 011‘“ thromboxane (the EP, 1P, DP, FP and TP receptors) and are coupled to heterotrimeric G- proteins which, in turn, modulate the activities of effector proteins to control cellular levels of cAMP, calcium and phosphatidyl inositol [53]. peas- "" b- ‘ mam. T-‘L it’dt‘iirfln 0Cyl: C-CJ‘E'U CUE-1‘ 312:0. lilt‘ PCB l m mi“? 13‘ O u" 8‘ =.*\.‘c.1.!r.'I‘z 0i i.’ A C\ K“ 5 up Pins 3 r;-_ ral,, ' Wutllnn bV . His 3828' and fir ‘1]!‘ t “5 a neutral - l a» WI radial L I CHJP' , :1' Q — — ,. — Peroxidase Catalysis by PGHS PGHSs catalyze both a cyclooxygenase (bis-oxygenase) and a peroxidase reaction. The cyclooxygenase reaction involves the addition of two molecules of O2 to arachidonic acid to produce prostaglandin G2 (PGG2). PGG2 then undergoes a two- electron reduction in the peroxidase reaction to produce PGH2. These two reactions occur at distinct but structurally and functionally interconnected sites. The peroxidase reaction occurs at a heme-containing active site located near the protein surface. The cyclooxygenase reaction occurs in a hydrophobic channel in the core of the enzyme. In vitro, the peroxidase activity can operate independently of the cyclooxygenase [55, 56]. In contrast, the cyclooxygenase reaction is peroxide-dependent [57] and requires the oxidation of the heme group at the peroxidase site [58]. A combination of electron paramagnetic resonance (EPR) spectroscopy, 3H- labeling and spectrophotometric studies has led to a branched-chain, mechanistic model for peroxide-dependent activation of the cyclooxygenase activity first proposed by Ruf (Figure 3; [59-62]. This model proposes that the initial reduction of a hydroperoxide results in a two-electron oxidation of the PGHS heme group to produce an oxyferryl group plus a radical-cation protoporphyrin IX (Compound I) and an alcohol derived from the oxidizing peroxide [13, 58, 59, 63]. Compound I next undergoes an intramolecular reduction by a single electron traveling from Tyr385l to the proximal heme ligand, His388 and finally, to the heme group. The result is Intermediate 11, an oxyferryl group plus a neutral protoporphyrin IX and a Tyr385 tyrosyl radical [13, 59, 64, 65]. This tyrosyl radical is thought to engage in abstraction of the l3-proS hydrogen from ' Amino acid numbers correspond to the sequence of ovine PGHS- l. 10 CYCLOOXYGENASE A f 202 [Ee4+ =0 PPIX] 335Tyr E [Fe4+ =o PPIX] 335Tyr m... ..>\ X: We” PPIxr 335m X445.“ =0 pmx+1 385w. __. 5.4+ =0 Pplx] 3851,. . H+ COMPOUND r INTERMEDIATE u e- A/e/ Ty'#¢ IFe“ =0 PPIX] 335“" [Fe4+ =0 PPIX] 385m + M”- L COMPOUND 11 INTERMEDIATE Ill cnoss- ‘ V LINKING PEROXIDASE SUICIDE INACTIVATION Figure 3. Model for peroxide-dependent activation of cyclooxygenase activity of PGH synthase via formation of an intermediate tyrosyl radical. Fe3 PPIX, ferric iron protoporphyrin IX (heme); ROOH, alkyl hydroperoxide; ROH, alcohol; AA, arachidonic acid, Fe4+ =0 PPIX, oxyferryl heme. Compound I, an oxyferryl group (Fe(IV)= 0) plus a protoporphyrin IX radical cation; Intermediate II, an oxyferryl group plus a neutral protoporphyrin IX plus a Tyr385 tyrosyl radical; Compound 11, an oxyferryl group plus a neutral protoporphyrin IX. ll 1" g {h .\n1< \q *0; I, Katalm'irilg. 1.x (0 acme ox inc- ”TERI mL‘CllJEI‘ urban kit-Oil [M] arachidonate, the rate-determining step in cyclooxygenase catalysis. Solution of the Coy-heme ovine (o)PGHS-l crystal structure complexed with arachidonate has led to detailed mechanistic implications for the cyclooxygenase reaction presented in the next section [66]. 12 .r-g ' .0. melt»- ic’giis of the formation of ti: 5.56 ofthc Qt‘i abstaczion to 3" 1:09 to rm I must occur in 1' at a \I .r t} MUDCT‘TIQT Cyclooxygenase Catalysis by PGHS The initial step in the cyclooxygenase reaction is the stereospecific removal of the l3-proS hydrogen from arachidonate [61]. This hydrogen is abstracted as a hydrogen atom [67-70] by a tyrosyl radical at Tyr385, generating an arachidonyl radical at C-1 1. (Figures 3,4). Examination of the Coy-heme oPGHS-l/arachidonate crystal structure combined with knowledge derived from previous mechanistic studies have led to the details of the proposed cyclooxygenase mechanism that follow [66]. Following formation of the C-11 arachidonyl radical, molecular oxygen can then migrate from the base of the cyclooxygenase site and add to the substrate from the side opposite hydrogen abstraction to yield an llR-peroxyl radical. This radical then reacts with the double bond at C-9 to form the endoperoxide. Afier endoperoxide formation, a conformational change must occur in order to position C-8 and C-12 at the appropriate distance for formation of the cyclopentane ring. This movement, in turn, allows the llR-peroxyl group to react with O9. Formation of the cyclopentane ring results in a repositioning of the 0) end of the fatty acid molecule such that C-15 becomes optimally positioned for the addition of a second 02 molecule coming from the base of the cyclooxygenase site to form PGG2. Once formed, PGG2 cycles back into the peroxidase reaction and is quickly reduced at the peroxidase active site to form PGH2. Suicide Inactivation Both PGHS-l and —2 undergo a poorly understood phenomenon called suicide inactivation. Both the peroxidase and cyclooxygenase activities are inactivated within two minutes of the start of catalysis. On average, each enzyme molecule can turn over 13 Figure 4. \ qclooxiaen r C -H . +02 COOH COOH o n \ .1 .I O O \ \ COOH COOH 0 ’M 0 ’w \ I \ I ., ———> 44‘! ' COOH COOH Figure 4. Mechanism model for conversion of arachidonic acid to PGG2 by the cyclooxygenase activity of PGHSs. l4 '11:". 3 TCJCIIUI“. 2.": mhrdonamodcrfi . ' Q ~ 121“» - - 9 .. “1mm 13 .‘11 111‘... I . \zUC‘ll E‘ ‘ ‘3“, q.“ ‘1 a lLane .‘. .‘uHI - if!” LT'I‘ ........-0.'1 0i .1 I‘le oPGHS-l ‘W c be .Iu‘ Litmedinc 111 '3‘- .3.th63111‘. 11V 2 0 [‘1‘ g mm‘vaurrn. P11 ‘9. ~45. . Jusflonn 2 T“T ‘ . 1115 the mimic about 400 times before becoming suicide inactivated. Suicide inactivation originates with a reaction intermediate [71]. This intermediate does not directly involve an arachidonate-derived radical because the rate of fatty acid attachment to PGHS during catalysis is 30 times slower than that of suicide inactivation [72, 73]. As depicted in Figure 3, suicide inactivation likely proceeds from Intermediate II [71] and involves the formation of a tyrosyl radical other than the Tyr385 radical. Protein tyrosyl radicals in oPGHS-l have been shown to be localized to tyrosines other than Tyr3 85 [74-76], and an Intermediate 111 has been detected in association with peroxidase inactivation [71]. Interestingly, one mutation (H386A) results in an enzyme that does not undergo suicide inactivation. The His 386 residue may be involved in the transport of the radical from neighboring Tyr3 85 that may somehow trigger an aberrant cross linking reaction that kills the enzyme. 15 #53115 for hurt The Drotczns. h far PGHS-II .1? . A ~. C~ . Ltm‘ndnfn .k‘r K PGHSK membranes 1?: R1, L I W".- .l-L, L} ‘10-ka I" ‘ ‘I Do it 14 , w ht} hem c: 1.- 1‘ . ‘5 .nc absence . @561106 01‘ 11 PGHS Structure and Biochemical Properties The primary structures of PGHSs-l and -—2 from numerous species are known [1]. They are encoded by different genes with PGHS—2 being a small, immediate early gene (8.3kb for human PGHS-2) and PGHS-1 originating from a much larger, 22-kb gene. The proteins, however, have similar molecular weights (72kDa for PGHS-1 and 74kDa for PGHS-2) and share a 60% sequence identity within species and >85% identity among individual isoforms from different species. Additionally, all residues identified as being important for catalysis are conserved between the two isoforms. PGHSs are integral membrane proteins which have been proposed to interact with membranes through a novel membrane binding domain of about 50 amino acids in which hydrophobic residues of amphipathic helices interdigitate into one leaflet of the lipid bilayer. It is in this region that the major sequence differences occur with only 38% identity between PGHS-1 and -2. Another notable difference between the two isoforms is the absence of a sequence of 17 amino acids from the N terminus and the insertion of a sequence of 18 amino acids at the C terminus of PGHS-2 in comparison to PGHS-l [77]. The significance of this 18-amino acid insert in PGHS-2 has not been identified, but may be involved in targeting, protein degradation or protein-protein interactions (Figure 5). PGHSs function as homodimers and are localized to the luminal side of the endoplasmic reticulum and the inner and outer nuclear envelope [78, 79]. These enzymes contain an epidermal grth factor (EGF) homology domain of about 50 amino acids which is present at the dimer interface, neighboring the membrane binding domain. The globular catalytic domain consists of about 460 amino acids [80-82] and is located in the C-terminal of the enzyme. PGHS-1 PGHS-2 EGF] MED] Catalytic mc 59% 38% 70% PGHS-2 N EGF] MED] Catalytic Figure 5. Comparison of the PGHS-1 and PGHS-2 proteins. Percentages denote the sequence identity shared between PGHS-l and PGHS-2 in the various structural domains. Shaded regions depict short inserts in PGHS-l and -—2 that do not appear in the other isozyme. EGF = endothelial grth factor domain; MBD = putative membrane binding domain; Catalytic = catalytic domain, which includes the cyclooxygenase and peroxidase active sites. '11- i“ ; " .I|.l. 9“, mb LL tree sztcs IS 9‘“ 'U (*1 I. (I1 | 1) (a p i 'J $137.1.1110I1l‘rc 1‘ pt 1 Marv-me stir .‘espectix 61V. 11 1 Tu PL..Sd1mc-r mi PGHSs are heme-containing proteins with one molecule of heme per monomer [10, 83] and are glycosylated at three sites (Asn68, 144 and 410). Glycosylation of all three sites is essential for enzyme activity in both isozymes [84]. The 74 kDa mass of PGHS-2 (compared to 72 kDa for PGHS-1) is attributed to an inefficient glycosylation at a unique fourth site, Asn580, whose glycosylation has no effect on the enzymatic activity of PGHS-2 [84]. Kinetic as well as structural properties of both isozymes have been shown to be similar for the primary substrates, arachidonic acid and molecular oxygen. The KM’s for arachidonic acid and 02 have been determined to be in the range of 1-5 11M and 5 11M, respectively, with cyclooxygenase turnover rates at 3,500 moles of arachidonate/mo] of PGHS dimer/minute [85-88]. 18 Tn hk‘m \StID 113513“: 5:31.388 C0111“ $115313 \11 Pkll Email} $9677 77‘- PGHS: h 113211x112m‘1s. T‘ 1121:1111 of the matures. ma. 3301131 factor 11‘ and a mem'nram The \-t badge: and m; domains encun 1‘16 heme site CORSQCUIIVE 3H PGHS Crystal Structures The homodimeric crystal structure of ovine (o)PGHS-l complexed with the NSAID flurbiprofen was solved in 1994 and is shown in Figure 6 [80]. Since then, the structures Coy-heme oPGHS-l complexed with arachidonate [66] and of human [81] and murine [82] PGHS-2 complexed with NSAIDs have been solved and were found to be virtually superimposable on the original oPGHS-l structure. PGHSs have a large helical content with very little beta sheet and function as homodimers. The reason that dimerization is necessary for catalysis is unknown. The solution of the crystal structures confirmed the implications of the known primary structures, indicating three distinct folding regions for each monomer: an epidermal growth factor (EGF)-like domain at the dimer interface, a large globular catalytic domain and a membrane binding domain [80-82]. The N-terminal EGF homology domain of PGHSs is tethered by three disulfide bridges and may play a role in the dimerization of the monomers. The globular catalytic domains encompass both the cyclooxygenase and peroxidase active sites which neighbor the heme site on opposite sides. The membrane binding domain contains four short, consecutive amphipathic Ot-helices and creates a hydrophobic patch that interacts with the lipid bilayer. These helices also surround an opening to the cyclooxygenase active site with the fourth helix (Helix D) protruding into the catalytic domain. The cyclooxygenase active site exists as a long hydrophobic channel which opens at the outer surface of the membrane binding domain and can bind fatty acid substrates and NSAIDs. The peroxidase active site is a shallow cleft containing the Fe- protoporphyrin IX prosthetic group which is coordinated by two histidines. This site is 19 figure 6. ( $110M in gr 117%}er C11 and aracmd. I1115111$Sena Figure 6. Crystal structure of the oPGHS—l dimer. The major folding units are shown in green (EGF homology domain), orange (membrane binding domain) and blue (catalytic domain). The heme group is shown in red, bound in the peroxidase active site and arachidonate is shown in yellow, bound in the cyclooxygenase active site. Images in this dissertation are presented in color. 20 I .w : Vt ow 3. 1 1.11:3... at: \1 V5 ‘4 structurally related to the peroxidase site of myelOperoxidase [80]. PGHS structures also contain several water channels, one of which extends from the cyclooxygenase site near Gly533 to the dimer interface. It is not clear if the water channels play a direct role in catalysis. 21 (”t In 19%. 115152105. \1 as c smmmmno 011.16 PGHS-l .1 1'1: caillmc sztc 1-1tPGllS-2 t 122:1: 111 P0115 sew-gen... ‘. ‘ ;-_ _ x “L: A- “ i‘rk‘l’ N o . 1".1‘ duerence at f ‘wz-‘mcg film Chianti-5 , . :L hub d]\ mm “ Cele \ kn 903111011 of 1111 PGHS-21h“ . the membrant Comparison of PGHS-1 and PGHS-2 Crystal Structures In 1996, the crystal structure of human PGHS-2 complexed with two different inhibitors, was determined and found to have a tertiary and quaternary structure very similar to that of PGHS-l [81, 82]. The amino acid sequences of human PGHS-2 and ovine PGHS-l are 67% identical. The residues that form the substrate binding channel, the catalytic sites and the residues immediately adjacent are all identical between PGHS- l and PGHS-2 except for three small variations. Isoleucine in PGHS-l is exchanged for valine in PGHS-2 at positions 434 and 523 and histidine is exchanged for arginine at residue 513. The PGHS-2 structure revealed a second internal pocket extending off the cyclooxygenase site. In PGHS-2, the volume of the cyclooxygenase site is calculated to be 394 A, whereas the volume of the cyclooxygenase site of PGHS-1 is 316 A [81]. The 25% increase in the size of the cyclooxygenase site in PGHS-2 is primarily due to the difference at position 523 and possibly explains some of the differences that the two isozymes exhibit with respect to interaction of the various NSAIDs and substrates. This change has also allowed for the development of a number of PGHS-2-specific inhibitors such as celecoxib and rofecoxib [89-91]. The other major structural difference is in the position of the last of the four helices of the membrane binding domain, helix D. In PGHS-2, this helix is tilted toward the body of the enzyme, providing a larger opening in the membrane binding domain. 22 PGHS-1 I ' ’ ‘ , 11.7.11‘nul‘1t J ' 1 .N 073111165 .11 Ill: denim Fill-Q: Orly one dzf’r‘crt grsrtinn 523 in fiwnflms Th? ' ' 3.1-1. toptmh\n I". 01 the err/mk- .4“”‘ i I 7 ~, "01.12% [9‘ ‘ - affimcm .1: drt‘pertl \ 1 d c 5 mam,“ of ( 1701117 d 10 136 It. Refidu 0‘ .- . 11111114110114] Rif- ‘ and CD‘“. 1 $11,; ( K1,- ‘ .131 \ 01 the i" 71- - a e-limjtim 0.. Active Sites of PGHSs PGHS-1 and —2 each contain a cyclooxygenase active site in the form of a hydrophobic channel into which fatty acid substrates and NSAIDs bind. This channel originates at the membrane binding domain and extends into the core of the globular domain [80-82]. Nineteen residues line the hydrophobic cylooxygenase active site with only one difference between the isozymes — Ile at position 523 in PGHS-1 and Val at position 523 in PGHS-2. Otherwise, all residues important for catalysis are conserved between PGHS-1 and —2. Both PGHS-1 and —2 also contain one iron (FeIII) protoporphyrin IX group per subunit that binds a heme group in the peroxidase active site of the enzyme [83, 92, 93] and is essential for both cyclooxygenase and peroxidase activities [92]. As mentioned earlier, the heme group is coordinated by two histidines, His207 and His388, which were identified by mutational studies. His388 is the proximal heme ligand and is required for both cyclooxygenase and peroxidase activities. Hi3207 acts as the distal histidine and is predicted to be important in deprotonation of the hydroperoxide substrate and subsequent protonation to form the alcohol during generation of Compound I (Figures 3 and 7) [58, 80-82]. Another histidine, His386, was found to be required for peroxidase but not cyclooxygenase activity [94]. Residues critical to PGHS activity were initially identified through a combination of mutational and elecron paramagnetic resonance (EPR) spectroscopy studies. In 1988, Ruf and co-workers used EPR to establish the existence of a tyrosyl radical in PGHS catalysis which was coincident with the formation of Intermediate II [59, 64] (Figure 3). Most of the evidence suggests that Tyr385 is the protein radical species involved in the rate-limiting abstraction of the l3-proS hydrogen of arachidonate. Tyr385 is essential for 23 —-_- _ mm {1 Figure 7. \IOC ' me9011181“ Figure 7. Model of the active site for ovine PGHS-1. Residues shown are conserved between the two isozymes and are described in the text. 24 reoridesc‘ 1nd Q 310111315 I“ I SUMMER-1‘6 5‘" reedes 1101" {EC Citii‘x‘WI‘gm‘bc J Side opposite I)?- The argfr. gmxylate grnu; itcprtifen ‘11 Q" meme in the | 15.12:} arachidur remit linkage bet interaction berm R1200 mutation ("xiecnation al‘h kl H‘HPETE funn T}T‘~5 s LII .71?st ' at {IP11 ' A ‘sd I - 11h “131.991 enzyme activity [95] and is found neighboring the heme group and the l3-proS hydrogen of arachidonate bound in the cyclooxygenase active site [66, 80]. The proximity of these peroxidase and cyclooxygenase active site residues to one another and, more importantly, the proximity of Tyr385 to the heme group and peroxidase active site confirmed that the cyclooxygenase and peroxidase activities, although separate, were both functionally and structurally coupled. Other important residues in the cyclooxygenase active site include Arg120 which resides near the opening of the active site, Tyr355, also near the mouth of the cyclooxygenase active site channel directly across from Arg120, Ser530 residing at the side opposite Tyr385 and Ile523 (Figure 8). The arginino group of Arg120 has been shown to serve as a counter ion for the carboxylate group of fatty acid substrates and certain NSAIDs such as flurbiprofen and ibuprofen [80, 96]. Replacement of Arg120 with a neutral glutamine causes a 1000 fold increase in the KM for arachidonate indicating that this residue is essential for high affinity arachidonate binding to the cyclooxygenase site [96, 97] and that this involves an ionic linkage between Arg120 and the carboxylate of arachidonate. In contrast, an ionic interaction between Arg120 and arachidonate is not important with PGHS-2 where an R120Q mutation has no detectable effect on either the Vmax or the KM for arachidonate oxygenation although this mutation does cause a small increase in the relative amount of ll-HPETE formed [98]. Tyr355 has been shown to play a role in the effectiveness of both instantaneous and time-dependent PGHS inhibitors and Ser530 is the site of acetylation of PGHSs by aspirin [21, 99]. Lastly, Ile523 which is a valine in PGHS-2 appears to play a role in the 25 Figure 8 complex R120 Y355 Figure 8. Cyclooxygenase active site from the Coy-heme oPGHS-l crystal structure complexed with arachidonic acid [66]. Images in this dissertation are presented in color. 26 .jgf'erfm‘t‘ [‘L‘ \SXHDSJndr izx‘bé‘ét‘n “31 Rt‘c‘L‘T‘.‘ wnrmeank 163'} similar It properties of a More imports: mine intent mmmsmm, neumflmfiL- 154.05,] and afferent residt he first shell u Emmi acids; bu mutational anal l . ength m Cham differences that the two isozymes exhibit with respect to interaction of the various NSAIDs and substrates. These three residues will be discussed in more detail in the discussion of PGHS inhibition. Recently, the crystal structure of Co3+-heme oPGHS-l complexed with the substrate, arachidonic acid was solved by Garavito and co-workers [66]. The structure is very similar to the inhibitor-bound complexes and confirms positions and functional properties of active site residues indicated by the earlier structures and mutational studies. More importantly, this structure was able to illuminate the nature of specific substrate- enzyme interactions, allowing for a much more thorough understanding of residue functions and catalysis in the cyclooxygenase active stie. Arachidonic acid is bound in an extended L-shaped conformation and makes a total of 48 hydrophobic contacts (i.e. 2.5-4.0 A) and two hydrophilic contacts with the protein, involving a total of nineteen different residues (see Figure 8). Additionally, there are several amino acids which are in the first shell of the cyclooxygenase hydrophobic tunnel and contact other first shell amino acids but lie outside of van der Waals distance to arachidonic acid. A large scale mutational analysis of these cyclooxygenase active site area residues is discussed at length in Chapter III. 27 Structural U There it's intimation of! generns in the '9 Riding proteif nitrates t1 it t agents [lint], m‘gutt‘riuni 1. iiexxgemtior Some acid substratt the binding 5 303K 01’ NS charged or rfildut§ ant Structural Comparison between PGHSs and other Fatty Acid Binding Proteins There are many proteins known to bind mono- and polyunsaturated fatty acids. Examination of the crystal structures of three of these proteins reveals some fundamental patterns in the protein-substrate interactions also observed in PGHS. The adipocyte lipid- binding protein (ALBP) serves as an intracellular receptor for numerous fatty acid substrates for transport to the appropriate site for use as metabolic fiaels and regulatory agents [100]. The cytochrome p450BM-3 is a fatty acid hydroxylase from Bacillus megaterium [101] and the soybean and human lS-lipoxygenases catalyze the dioxygenation of polyunsaturated fatty acids [102]. Some common characteristics among these proteins are: (l) folding of the fatty acid substrate is required for accommodation of the entire length of the fatty acid chain in the binding site, (2) all of these proteins, much like PGHS, require at least one stabilizing ionic or hydrogen-bonding interaction between the carboxyl end of the fatty acid and a charged or polar residue and (3) several van der Waals contacts between active site residues and areas along the fatty acid chain serve to stabilize substrate binding. ALBP has been crystallized in a complex with arachidonic acid [103]. The binding affinity (KD) for arachidonate in ALBP is 4.4uM, in the same range as that observed for PGHSs (2-5uM). Another similarity is the requirement for an ionic/hydrogen bonding interaction between the carboxylate of arachidonate, two arginines (one of which is linked through a water molecule) and one tyrosine. There are some very major differences between these proteins, however. ALBP is a small protein (130 amino acids) with a B-barrel structure and the substrate binding cavity is water- filled, in contrast to the hydrophobic pocket of PGHSs. Also, the carboxyl end of 28 RF amortize is l PGHS. Aithoug attiie szte. URN generous 11313" eaten. mono; PGHS. is a 1.11 some B-s‘neet etidenced b} phenomenon interaction b1 the probabzli transient in h:"¢'(1ph\ibi \ C}1(1Chl'k)m Q 111111 PGH K The arachidonate is buried in the binding pocket, “upside down” from that observed with PGHS. Although there are a number of other substrate stabilizing forces in the ALBP active site, these include not only hydrophobic interactions, as with PGHS, but also numerous hydrogen bonds between water molecules and other intervening active site residues. The cytochrome p450BM-3 structure has been solved complexed with the 16- carbon, monounsaturated fatty acid, palmitoleic acid [101]. This protein, similar to PGHS, is a larger heme-containing fatty acid hydroxylase, largely helical in content w/ some B-sheet. Additionally, there is evidence for multiple substrate conformations evidenced by the formation of multiple products (hydroxy acid or epoxide formation), a phenomenon observed for PGHSS [104]. Other similarities include a hydrogen bonding interaction between the arachidonate carboxylate and an active site tyrosine, as well as the probability of an ionic interaction with one arginine, although this interaction may be transient in nature. The substrate is also stabilized by multiple interactions with hydrophobic active site residues. Again, one important difference between PGHS and the cytochrome p450BM-3 structure is that the fatty acid is bound “upside down,” compared with PGHS, with the arachidonate carboxyl end buried in the binding pocket. The soybean lS-lipoxygenase structure has been solved in the absence of any bound substrate and inhibitor [105]. Implications for substrate binding has been determined through docking of arachidonic acid in the soybean lipoxygenase putative binding site and applied to both the soybean and human 15-lipoxygenases [102]. Both soybean and human lipoxygenases catalyze the formation of hydroperoxides from arachidonic acid and, although the plant and mammalian forms share only 28% identity, 29 r... if can 0 ifiese are 1: sheet. ..e neighlwrs a and is duck-t 3:111 its cart» (u U: E: :3 "/1 one can assume structural similarity due to their similar activities. Similar to PGHSs, these are iron-containing proteins which are largely helical in structure with some B- sheet. The iron is coordinated by three histidines, an isoleucine and an asparagine and neighbors a solvent-accessible site thought to be the substrate binding site. Arachidonic acid is docked in the putative substrate binding site in close proximity to the catalytic iron with its carboxylate in an ionic linkage with a lysine residue in the soybean enzyme and a corresponding arginine in the human in a manner similar to that of oPGHS-l. Additionally, numerous hydrophobic residues interact with areas along the fatty acid chain, including a Trp, Phe, Ile, Leu, Met and Thr and the KM for arachidonate binding is, again, similar to that of PGHSs at 7.5 uM. 30 2K THC C}! £11051 Common \ minty of the e'.‘ terrible inhib Eitipttifen am but produce a ti and indomethst gnnidi .ium g zachidtinate [ hzghaftinit)‘ .s 3} SIUdlCS m \ fielded an en 1.96.91, The i QCIOOXE‘ECI- 367531? )_ bur a11:81.1'1Ciicd P inhthUI-S \\ {3111]} 10 P1 The r“ ‘ d‘tnbutcd II t~ . U \ a] "sub\_ Inhibition of PGHSs The cyclooxygenase activities of both PGHS-1 and PGHS-2 can be inhibited by most common NSAIDs such as aspirin, ibuprofen and naproxen, leaving the peroxidase activity of the enzymes unaffected. NSAIDs such as ibuprofen and naproxen cause reversible inhibition by competing with the substrate for the active site of the enzyme. F lurbiprofen and indomethacin also compete with substrate for the cyclooxygenase site, but produce a time-dependent, slowly reversible inhibition of PGHSs. Both flurbiprofen and indomethacin are carboxylic acid-containing NSAIDs which ion pair with the guanidinium group of Arg120, also known to interact with the carboxylate of arachidonate [82, 99, 106]. This interaction has been shown not only to be essential for high-affinity substrate binding, but also for time-dependent inhibition. This is evidenced by studies in which replacement of Arg120 in PGHS-1 with gluatmine or glutamate yielded an enzyme which was resistant to time-dependent inhibition by these NSAIDs [96, 97]. The inhibition by aspirin is due to the irreversible acetylation of the cyclooxygenase site of PGHS-1 at Ser530. PGHS-2 is also acetylated by aspirin at Ser530, but will still oxidize arachidonic acid to make lSR-HETE, whereas similarly acetylated PGHS-l will not oxidize arachidonic acid at all [107-109]. Additionally, inhibitors will differentiate between PGHS-l and PGHS-2 and, in general, bind more tightly to PGHS-l [110]. The biochemical differences between the two PGHS isoforms have been attributed to the larger and more accommodating active site of PGHS-2 resulting from Ile to Val substitutions at positions 523 and 434 and a His to Arg substitution at position 513 31 . ll; 1 3.1151105. All 0 60311-11 in 3.12s. .trgllll TI] ese ramihle meek. 1 nethatism [3. - PGHS-l and —.‘ Lzlibiznrs hem tL fats PGHS-I~ [$1536 mutant . emit in ptoti'. cnsui stricture-x Armin inhibit Mich is stencil in PGHS-2. These differences are exploited in the development of PGHS-2 specific inhibitors. All of the new PGHS-2 inhibitors are larger than other classical NSAIDs and contain an arylsulfur group rather than a carboxylate group and thus, do not bind to Arg120. These inhibitors selectively inhibit PGHS-2 by a time-dependent, slowly reversible mechanism while they inhibit PGHS-1 by a rapid, competitive and reversible mechanism [3, 20]. Swapping of cyclooxygenase active site area residues unique to PGHS-l and —2 effectively exchanges the inhibitory pattern of PGHS-Z-specific inhibitors between the isoforms. In particular, mutation of Ile 523 to Val in PGHS-l allows PGHS-Z-specific inhibitors to bind and inhibit prostaglandin formation, and the reverse mutant of PGHS-2 in which Va1523 is exchanged for Ile shows inhibitor selectivity profiles comparable to native PGHS-1 [111, 112]. Examination of PGHS-2 crystal structures [81, 82] has indicated that the smaller size of Va1523 allows the PGHS- 2-specific inhibitor access to a side pocket off the main substrate channel in PGHS-2 which is sterically denied by the longer side chain of Ile in PGHS-l. 32 Aracl prostanoids l ailii'nits 3 tr. same conditi- pikwgnxaturd‘. PGHS acids in 115771 7-iinolenic at pmstaglandim ‘8'195 pn‘uxzdgj products by P established tha' film the mg Cd is. abstracted .11 l Alfie‘hltlt .713 jOl’ prostanoi: 1116]. The m, indicating that d m - 0' Dlht‘imw. l l gand from “'h' l\ 51,117] I Cyclooxygenase Substrates of PGHSs Arachidonic acid is mobilized by phospholipases A2 for conversion to the prostanoids by PGHSs. However, only the cytosolic Cay-dependent phospholipase A2 exhibits a marked selectivity for arachidonic acid [113]. Thus, it is likely that under some conditions, PGHS-l and PGHS-2 will be exposed to a pool of a mixture of different polyunsaturated fatty acids. PGHSs have been shown to oxygenate a variety of C13 and C20 n-3 and n-6 fatty acids in vitro [114, 115]. The 20-carbon, n-6 fatty acids, arachidonic (20:4) and dihomo- y-linolenic acids (20:3) are used most efficiently, producing the 2- and l-series prostaglandins, respectively. Eicosapentaenoic acid (20:5n-3) is converted to the 3- series prostaglandins and the l8-carbon fatty acids are oxidized to monohydroxy acid products by PGHSs. Product characterization from various fatty acid substrates has established that, in most cases oxygenation is initiated following abstraction of hydrogen from the n-8 carbon. One exception is a-linolenate (18:3,n-3) in which the n-5 hydrogen is abstracted [115]. Arachidonic acid (20:4n-6) is converted in vivo to the 2-series prostanoids -- the major prostanoids present in plasma, endothelial cells and renal collecting tubule cells [116]. The major prostanoids found in seminal fluid, however, are of the 1-series, indicating that dihomo-y-linolenic acid (20:3n-6) can also serve as a PGHS substrate in viva. Dihomo-y-linolenic acid is the primary C20 fatty acid found in sheep vesicular gland from which some of the earliest work on prostanoid biosynthesis was performed [61,117]. 33 _ 4n|'.l‘ 1“- u“ h:.'df1_‘@t’.'\ 1\ {if b} prodt 011-3131131101 hydropero x i eicosapentat Eicosapentat binding to P( “hen ingeste- through this health and pn lihlblllun of binds efficient 11} PGHSS. The n-3 and n-9 fatty acids have been shown to act as relatively poor substrates, but effective inhibitors of the oxygenation reaction [114, 115]. Eicosapentaenoic acid (20:5n-3) is converted, albeit in small amounts, to prostanoid products of the 3—series by PGHSS when added exogenously to intact cells [118, 119] or when mobilized from cellular phosphoglycerides [118, 120-123]. Oxygenation of eicosapentaenoic acid by PGHS-l has been shown to be more sensitive than arachidonate oxygenation to hydroperoxide concentrations and is enhanced by addition of exogenous hydroperoxide or by production of hydroperoxy lipoxygenase products in cells [124, 125]. Curiously, oxygenation of this n-3 substrate by PGHS-2 is not reliant on the same levels of hydroperoxide activator and, although a poor substrate for both isozymes, eicosapentaenoate appears to be a slightly better substrate for PGHS-2 [115]. Eicosapentaenoic acid has been shown to compete effectively with arachidonate for binding to PGHSs and has been shown to inhibit the formation of 2-series prostaglandins when ingested or added to human platelets or endothelial cells [120, 121, 126, 127]. It is through this mode of action that n-3 fatty acids are thought to promote cardiovascular health and protect against thrombosis and atherogenesis. Similar arguments apply to the inhibition of arachidonate oxygenation by docosahexaenoic acid (22:6n-3) which also binds efficiently to the cyclooxygenase site [114], but does not appear to by oxygenated by PGHSs. The 18-carbon fatty acids are oxygenated to monohydroxy acids by PGHSS. Linoleic acid (18:2n-6) is an essential nutrient because it is converted to arachidonic acid, the primary substrate in formation of the eicosanoids. In addition, a number of tissues synthesize oxygenated products directly from linoleic acid, suggesting that it may have 34 L‘i'jg’r V1131 II”;01’1\‘Z".}1§1’L1‘ 9.. i'C’CEJLiCi exjsgenatzur. 013311311011 be produced . .111 61113 [13; It‘fi‘clii cf} . 5511:135113‘. g .11 flit the H01; Wu- 1 ' .1t..lfllllt71i'l (11‘: A; and lf-Hli is ‘r‘hldmqu 1; [1. La) Slibstratc 16113.1,u,”L \\ \-\ 2b1111101 PGH Hum-cm SS the [1 other vital functions. PGHSs have been shown to convert 18:2 to a mixture of the monohydroxy acids 9-hydroxy-10,l2-octadecadienoic acid (9-HODE) and 13-hydroxy- 9,11-octadecadienoic acid (l3-HODE) with a catalytic efficiency that is 5% that for oxygenation of arachidonate for human PGHS-1 and 15% the efficiency of arachidonate oxygenation by human PGHS-2 [l 15]. These linoleate metabolites have been shown to be produced under both basal and stimulated conditions from endogenous linoleic acid in alveolar macrophages [123] and human umbilical vein endothelial cells [119], respectively, and were inhibited by the cyclooxygenase inhibitor, indomethacin, suggesting an in vivo role for linoleic acid in cyclooxygenase catalysis. Suggested roles for the HODE products include maintainance of the water barrier in the skin [128], inhibition of 5-lipoxygenase activity in leukocytes [129] and modulation of thromboxane A2 and lZ-HETE formation in platelets [130]. a-linolenic acid (18:3n-3) is converted to 12-hydroxy-(9Z,13E/Z,lSZ)-octadecatrienoic acids (lZ-HOTrE) by PGHSs [115]. This n-3 substrate is an exceedingly poor substrate for PGHS-1, but markedly better for PGHS-2, suggesting the possibility that a-linolenate could serve as a substrate for PGHS- 2 but not PGHS-l in vivo. Similar to linoleate (18:2n-6), a-linolenate is an important nutrient as the precursor to other n-3 fatty acids such as eicosapentaenoic acid (20:5) and docosahexaenoic acid (22:6). Other less common fatty acids can also serve as cyclooxygenase substrates including adrenic acid (22:4n-6) [131], the Mead acid 52,82,11Z-eicosatrienoic acid [132], columbinic acid (5E,9Z,l2Z-octadecatrienoic acid) [133, 134], and 5,6-oxido-eicosatrienoic acid [132, 135]. As suggested by the preceding discussion, PGHS-2 will accept a wider range of fatty acids as substrates than will PGHS-1. Although both enzymes can use arachidonic 35 1 2L» ,.-. an; dunno _ 1 _ -\ >. SLICK 5b CISA‘\\. Izzoleic aezd 1 PCHS'l Coup 3m een Arg 1 ; and dihomo-y-linolenic acid equally well, PGHS-2 oxygenates other fatty acids substrates such as eicosapentaenoic acid (20:5,n-3), y- and or-linolenic acids (18:3,n-6 and n-3) and linoleic acid (18:2,n-6) more efficiently than does PGHS-1 [115]. This can most likely be attributed to the larger, more accommodating active site of PGHS-2 compared with PGHS-l coupled with the fact that PGHS-2 is not dependent on the ionic interaction between Arg120 of the enzyme and substrate carboxylate groups [98]. 36 P71“: .‘(ifi'ic’fslfll'l o prostanoids t :5 often ret‘en isteihle ism functions as considerable membrane pri gorilall'llll‘l. "' prototi'pe 0113 1 leaflet ot‘the 111 not throueh m1" PGHSs reaction in uhie PGGg undergoe- L"i'clooxyueria\t' :5". 114 ' ~ . 15Hj) 1: (£21. 5111!; {88. CHAPTER II ARACHIDONIC ACID CONFORMERS IN THE CYCLOOXYGENASE ACTIVE SITE OF OVINE PGHS-1 Introduction Prostaglandin endoperoxide H synthases-l and -2 (PGHS-l and -2) catalyze the conversion of arachidonic acid and O; to PGH 2 -- the committed step in the formation of prostanoids (prostaglandins, thromboxane A2 [1-3, 20, 136, 137]). PGHS-l (or COX-1) is often referred to as the constitutive enzyme whereas PGHS-2 (COX-2) is known as the inducible isoform [138-146]. Apart from their important biological roles and their functions as targets of nonsteroidal anti-inflammatory drugs [3, 20], PGHSs are of considerable interest in the context of the structural biology and enzymology of membrane proteins. These enzymes are homodimeric (ca. 72 kDa/subunit), heme- containing, glycoproteins with two catalytic sites; moreover, PGHSs represent a prototype of a new class of integral membrane proteins that appear to be anchored to one leaflet of the lipid bilayer through the hydrophobic surfaces of amphipathic helices and not through more typical transmembrane domains [2, 147-149]. PGHSs catalyze two separate reactions: a cyclooxygenase (bis-oxygenase) reaction in which arachidonate is converted to P66; and a peroxidase reaction in which PGGz undergoes a two electron reduction to PGH2. PGHS-1 and PGHS-2 have similar CYelooxygenase turnover numbers (ca. 3500 moles of arachidonate/min/mole of dimer; [87, 124, 150]) and KM values for arachidonate (ca. 5 11M; [86, 87, 115, 150]) and for O; (03- 5 uM; [88, 151]) and exhibit similar fatty acid substrate specificities [115]. The 37 1”“- xiv-13?"): 51m gratiilh 01:11:67- add \JRSICTU-id-ii the cyciNH .35. peroxidase PGHS lining-82.1. from 5.1 1.14-1 arachidt‘inie ae sanity. smtht acculated PGl Orientations in : HETE [155]. \' 101PGHS-l ank delefinlned [he I 111)“mined the Collective] y, ou attire site ofop alidngcmems 01 attire site cyclooxygenase reaction begins with a rate-limiting abstraction of the 13-proS hydrogen from arachidonate to yield an arachidonyl radical [61, 70]. This is followed by sequential oxygen additions at C-11 and C-15 to yield the prostaglandin endoperoxide PGG2. Nonsteroidal anti-inflammatory drugs compete directly with arachidonate for binding to the cyclooxygenase site and inhibit cyclooxygenase activity, but have little or no effect on peroxidase catalysis [55, 152, 153]. PGHSs exhibit some lipoxygenase activity producing small amounts of l 1- hydroxy-8Z,l2E,14Z-eicosatrienoic acid and lS-hydroxy-82,l lZ,13E-eicosatrienoic acid from 8,11,14-eicosatrienoic acid [61] and the corresponding 11- and lS-HETEs fi'om arachidonic acid [154, 155]. Aspirin-acetylated PGHS-2, which has no cyclooxygenase activity, synthesizes 15R-HETE [108, 155]. Studies comparing native and aspirin- acetylated PGHS-2 have raised the possibility that arachidonate can bind in distinct orientations in the PGHS-2 active site to produce either PGG2 plus llR-HETE or 15R- HETE [155]. We have examined the products formed from arachidonate by native ovine (o) PGHS-1 and a cyclooxygenase active site mutant, V349L oPGHS-l, and have determined the kinetics for the formation of each individual product. We have also determined the ICso values for the inhibition of formation of each product by ibuprofen. Collectively, our results indicate that arachidonate can be bound in the cyclooxygenase active site of oPGHS-l in at least three different, catalytically competent arrangements that lead to PGGz, llR-HETE, and 15-H ETE, respectively, and that these three arrangements of arachidonate occur subsequent to its entry into the cyclooxygenase active site. 38 194': Asher. Nil rzzimd .\U boxine 58m? 5:51:1an \\ 6“ 31'tR3d-C Oligonuclentz L'nii'ersity M. from commi it: Prt-‘pur .1113mp19-op( PGHS-l [79. 1 1111112 L‘nic 01131 Materials and Methods Materials. Arachidonic acid was purchased from Cayman Chemical Co., Ann Arbor, MI. [1-14C]arachidonic acid (40-60 mCi/mmol) was purchased from New England Nuclear. F lurbiprofen was from Sigma. Restriction enzymes and Dulbecco’s modified Eagle’s Medium (DMEM) were purchased from GIBCO. Calf serum and fetal bovine serum were purchased from HyClone. Primary antibodies used for Western blotting were raised in rabbits against purified oPGHS-l and purified as an IgG fraction [44, 79, 149]; goat anti-rabbit IgG horseradish peroxidase conjugate was purchased from BioRad. C10E7 detergent used for solubilization of oPGHS-l was from Anatrace, Inc. Oligonucleotides used as primers for mutagenesis were prepared by the Michigan State University Macromolecular Structure and Sequencing Facility. All other reagents were from common commercial sources. Preparation of V349L oPGHS—l. V349L oPGHS-l was prepared starting with Ml3mpl9—oPGHS-l, which contains a 2.3 kb SalI fragment encoding native ovine PGHS-l [79, 149], using a BioRad Muta-Gene kit and the manufacturer’s protocol. The mutagenic Oligonucleotide was 5'-1 I27AGAGGAGTAT®CAGCAGCTGA1'49-3'. Single-stranded phage samples were sequenced using Sequenase (ver. 2.0, US. Biochemical Corp.) and the protocol described by the manufacturer. The replicative form of Ml3mp19-oPGHS-l containing the V349L mutation was prepared from phage cultures, digested with Sal], isolated by gel electrophoresis and electroluted into dialysis tubing using standard protocols. The resulting 2.3 kb oPGHS-l cDNA fragment was purified and subcloned into pSVT7, followed by digestion with Pstl to verify the orientation of the insert [44, 79, 149]. Plasmids used for transfection were purified by 39 CsCl gradient ultracentrifugation, and the mutation was reconfirmed by double-stranded sequencing of the pSVT7 construct as described above. T ransfection of COS-1 cells with recombinant oPGHS-I. COS-1 cells (ATTC CRL-l650) were grown in DMEM containing 8% calf serum and 2% fetal bovine serum until near confluence (~3x106 cells/ 10 cm dish) [44, 79, 149]. Cells were then transfected with a pSVT7 plasmid containing cDNA coding for native or V349L oPGHS- 1 using the DEAE dextran/chloroquine transfection method as reported previously [44, 79, 149]. Forty hours following transfection, cells were harvested in cold phosphate- buffered saline (PBS), collected by centrifugation and resuspended in 0.1 M Tris-HCl, pH 7.5. Sham-transfected cells were collected in an identical manner. Microsomal preparations were used for Western blotting and for cyclooxygenase and peroxidase assays. Purification of oPGHS—I from seminal vesicles. Microsomes were solubilized with 0.1% C10E7, centrifuged, and the supernatant loaded onto an equilibrated DEAE- cellulose column. Fractions fi'om the DEAE-cellulose column were assayed for peroxidase activity and protein concentration. Desired fractions were pooled, concentrated, and loaded onto an equilibrated S-300 column overnight. Fractions from the DEAE and S-300 columns were used for assays with partially purified enzyme. The specific cyclooxygenase activity was 2-10 units/mg for DEAE fractions and 15-30 units/mg for S300 fractions. One unit of enzyme is defined as that amount of enzyme which will catalyze the oxygenation of l umol of arachidonate/min. Characterization of arachidonate derived products. Forty hours following transfection, COS-1 cells were collected, sonicated and resuspended in 0.1M Tris-HCl, 40 pH 7.5. Aliquots of the cell suspension (75 ug of protein) or 5 cyclooxygenase activity units of partially purified oPGHS-l were incubated for l min at 37° C with various concentrations of [1-I4C]arachidonic acid with and without 200 uM flurbiprofen. Radioactive products were extracted and separated by thin layer chromatography as described previously [57]. Products were visualized by autoradiography and quantitated by liquid scintillation counting. Negative control values from samples incubated with 200 uM flurbiprofen were subtracted from the experimental values observed for each sample in the absence of flurbiprofen. Inhibition of product formation by ibuprofen was measured similarly using different concentrations of ibuprofen in reaction mixtures containing 10 [1M arachidonate. Reverse phase (RH-HPLC. Native or V349L oPGHS-l (75 ug cell protein) prepared from transfected COS cells or semi-purified oPGHS-l (S cyclooxygenase activity units) were incubated with 2-100 11M arachidonic acid for 1 minute at 37° C. The products were extracted as described above and resuspended in HPLC buffers (1/ 1; v/v). 15- and ll-HETEs were separated by RP-HPLC using a C18 column from Vydac (Hesperia, CA) and detected with a Waters Model 600 HPLC equipped with a 990 photo diode array detector set at 234 nm. The polar component of the mobile phase was 0.1% aqueous acetic acid, and the eluting solvent was acetonitrile. The flow rate was 1 ml/min. The elution profile used was: 0-30 min, 30% acetonitrile; 30—100 min, 50% acetonitrile; 100-125 min, 75% acetonitrile; 125-130 min, 100% acetonitrile. The retention times of 15-HETE and ll-HETE were 36 and 38 minutes, respectively. Chiral HPLC of methyl I5—HETE and methyl II-HETE. Material obtained by RP-HPLC was esterified by treatment with excess diazomethane and subjected to chiral 41 .iibt Ilium Iain. .1133: "gm“ .1 . 5w. T111. dll {fluid 1 Wm.“ H‘dtillr l phase HPLC. Diazomethane was prepared from Diazald (N-methyl-N-nitroso-p— toluenesulfonamide) (Aldrich Chemical Company) as described in the company’s technical bulletin #AL-l80. Chiral-phase HPLC was performed on a Chiralcel OC column (250 X 4.6 mm) from Daicel Chemical Industries (Osaka, Japan) using hexane/2- propanol (98/2; v/v) at a flow rate 0.5 ml/min. The retention times of the methyl esters of 15R- and 15S-HETE were 25 minutes and 27 minutes, respectively. Retention times for the methyl esters of 1 IR- and llS—HETE were 26 minutes and 28 minutes, respectively. Cyclooxygenase and peroxidase assays. Cyclooxygenase assays were performed at 37°C by monitoring the initial rate of 02 uptake using an oxygen electrode [98]. Each standard assay mixture contained 3.0 ml of 0.1 M Tris-HCl, pH 8.0, 1 mM phenol, 85 ug of bovine hemoglobin and from 2 to 100 11M arachidonate. Reactions were initiated by adding approximately 250 pg of microsomal protein prepared from COS-1 cells or partially purified oPGHS-l in a volume of 20-50 ul to the assay mixture. Peroxidase activity was measured spectrophotometrically as described previously [57]. The reaction mixtures contained 0.1 M Tris-HCl, pH 8.0, 0.1 mM 3,3,3',3’- tetramethylphenylenediamine (TMPD), approximately 100 ug of microsomal protein and 1-7 11M hematin in a total volume of 3 m1. Reactions were initiated by adding 100 pl of 0-3 mM H202, and the absorbance at 610 nm was monitored with time. Kinetic Schemes and Derivations. As presented later in the Discussion section, two different kinetic schemes were developed to model mechanisms by which oPGHS-l could produce different products with different kinetic properties. Rate and ICso equations were derived for each scheme using the general rate equations: v = Vmax[S]/{KM + [5]} and V = Vmax[S]/{KM (1 + [ll/Ki) + [3]} 42 .- 17'. ' " PD 1“ 0 1119‘ “(1 fir“, ‘ 11.111 - h 111111 Both schemes were developed for the formation of two separate products to illustrate the principles; however, the schemes are easily expandable to accommodate situations where there are multiple products. Derivations are included in Appendix A of this manuscript. The relevant rate equations for Scheme 1: v(l) no inhibitor: k5 [E10,][So] K.(K0+ [So] + [Sol/K2) + [S0] v(2) no inhibitor: k6[Emt][S0] K2(Ko + [301+ [Sol/KI) + [50] V“) W/ inhibitor _ k5 [E10,][SO]/ (Kl + K] /K2+ l) (KoKi/ (Ki + Ki /K2+ ”X 1 ‘1 [ll/Kl) + [50] v(2) w/ inhibitor = k6 [Eioi][So]/ (K2+ [(2 /K. + l) (KoKz/ “(2+ K2 /K1 + ”X 1 + [ll/Kl) + [30] ICso(1) = Ki(1 + [Sol/K0 + [Sol/KiKo + [So]/K2Ko) IC50(2) = Kl (1 + [SO]/KO + [So]/K1Ko + [So]/K2Ko) ICso(1) = ICso(2) The relevant rate equations for Scheme 2 are: v(l) no inhibitor = k2 [E101] [80] K1“ + l/K0+ [S()]/ K0 [(2)-1‘ [So] v(2) no inhibitor = k; [EWMSO] K2(l + Ko+ Kolsol/ Kl) + [30] v(l) w/ inhibitor = k2[Em.][So]/(l + K./K.,K2) Kl“ + [101/ Klm+ l/Ko+ [Iol/Ku21Ko) + [30] (1 ‘1 KI/KoKz) v(2) w/ inhibitor = k4[E1m][So]/(l + Kng/ K1) K2(l + [lolKo/Kiu)+ Ko+ [lol/ K1121) + [S ] o (1 +K0K2/Ki) 43 L" 11" .x A. \ “350(1): 1+ l/K0'1' [Sela/K019 +1/K1) l/Kmy’r‘ l/ K1(2)K0 1C50(2) = 1 + K0 + [301(Ko/Ki + l/Kz) l/Kuz)+ Ko/ K111) K350“) 1“ I(30(2) Statistical analysis. Results are expressed as the mean +/- S.E.M. for a minimum of four experiments using different preparations of partially purified oPGHS-l or membrane-associated native or V349L oPGHS-l. Statistical significance was determined using a two-sample t-test assuming equal variances. 44 R v0 in: V \.. ° 1 l l ‘I '; 5w '1'» e~\ . .1. p1 5\l Vfl 5N “ . ’- um .\. i) Results Microsomal preparations of oPGHS-l and V349L oPGHS-l from COS-1 cells transfected with plasmids encoding native oPGHS-l or V349L oPGHS-l as well as partially purified native oPGHS-l from seminal vesicles were used to determine overall KM values for the oxygenation of arachidonate using an oxygen electrode assay. The KM values were 3 uM for both solubilized, partially purified and for membrane-associated, recombinant native oPGHS-l; as expected, this value is in the range (2-8 uM) reported previously for recombinant and purified forms of ovine, murine and human PGHS-l [86, 87, 115, 150, 156]. The KM for arachidonate with the V349L oPGHS-l mutant was 7 uM. Western transfer blotting indicated that V349L oPGHS-l was expressed in COS-1 cells at the same level as native oPGHS-l (data not shown) and had 65% of the cyclooxygenase activity and 100% of the peroxidase activity of native oPGHS-l expressed in parallel transfections. As illustrated in Figure 9, oPGHS-l and V349L oPGHS-l formed the same products from [1-'4C] arachidonate including those products derived from PGG2 (HHTre, PGD2, PGE2 and PGF2oc) and the monohydroxy acids 1l-hydroxy-SZ,8Z,12E,14Z- eicosatetraenoic acid (1 l-HETE) and lS-hydroxy—SZ,8Z,1lZ,13E-eicosatetraenoic acid (lS-HETE) which are derived from reduction of the corresponding hydroperoxides. However, V349L oPGHS-l produced substantially larger amounts of lS-HETE and relatively less PGG2 and ll-HETE (Figure 9) than native enzyme. With 2 11M arachidonate and 5 cyclooxygenase units of either partially purified oPGHS-l or broken cell preparations of COS-1 cells expressing oPGHS-l, 95% of the products were derived fi'om PGG2, 3% was ll-HETE and 2% was 15-HETE; with 2 uM arachidonate and broken cell preparations of COS-1 cells expressing V349L oPGHS-l, 70% of the product was PGG2, <0.5% was ll-HETE and 30% was lS-HETE. 45 Native V349L oPGHS-l oPGHS-l (—-AA 15-I-IETE—') (— lS-HETE ll-HETE“'—) (— ll-HETE HHTre —) (-— HHTre PGan 1: PGan Figure 9. Autoradiogram of thin layer chromatograms of products formed from [1- "Ci arachidonate by native and V349L oPGHS-l. Partially purified oPGHS-l (5 units) or broken cell preparations of COS-1 cells expressing V349L oPGHS-l (75 ug of cell protein) were incubated for one min with 50 uM [l-MC] arachidonate, and the products were extracted and separated by thin layer chromatography as described in Methods. The locations of the various chromatographic standards are as noted. See Figure 2 for PG structures. 46 Extracts from incubation mixtures containing partially purified oPGHS-l, recombinant oPGHS-l or recombinant V349L oPGHS-l were incubated with different concentrations of arachidonate and then separated by RP-HPLC. 11- and lS-HETE were isolated, converted to their methyl esters and examined by chiral HPLC. This is illustrated in Figure 10 for partially purified native oPGHS-l incubated with 100 uM arachidonate. l l-HETE was exclusively 1 lR-HETE. lS-HETE was 70% lSS-HETE and 30% lSR-HETE. The ratio of 15S-HETE to lSR-HETE was the same for both purified and microsomal, recombinant oPGHS-l. Similarly, the lS-HETE fraction derived from incubation of arachidonate with V349L oPGHS-l contained 70% lSS-HETE and 30% lSR-HETE; only trace amounts of ll-HETE were formed by V349L oPGHS-l, and the chirality of this latter product was not determined. Interestingly, the enantiomeric composition of the lS-HETE product was 65-70% lSS-HETE and 30-35% lSR-HETE at all arachidonate concentrations tested with both native and V349L oPGHS-l (i.e. when 2, 5, 10, 20, 35, 50 or 100 uM arachidonate was used as the substrate with purified oPGHS- 1, when 10 and 100 uM arachidonate was used as the substrate for membrane-associated, recombinant oPGHS-l or when 2, 5, 10, 20, 35, 50 or 100 uM arachidonate was used as the substrate for membrane preparations of recombinant V349L oPGHS-l) (Figure 11). We noted in the studies outlined above that there were obvious differences in the relative ratios of the monohydroxy fatty acid and PGG2-derived products at different arachidonate concentrations. Accordingly, we performed a series of measurements to determine the KM values for the formation of the different products by native oPGHS-l (Figure 12) and V349L oPGHS-l (Figure 13). The results are summarized in Table I. It should be noted that the amounts of enzyme (5 units for partially purified native enzyme or 75 ug of broken COS-l cell protein for V349L oPGHS-l) used in these assays were adjusted so that <25% of the total substrate was consumed even with the lowest 47 1211110. ( 11111 1P0] :rm" u 0‘ M' 101 “l” ‘ I ‘e . ”Ck hi?» “14.1 Figure 10. Chiral HPLC of the methyl esters of the 11- and lS-HETEs produced by native oPGHS-l. Arachidonate (100 uM) was incubated with partially purified oPGHS—l for one min at 37°C. 1 l- and lS-HETE fractions were separated by RP-HPLC and esterified by treatment with diazomethane. Chiral-phase HPLC was performed as described in Methods using a Chiralcel OC column with hexane/2-propanol (98/2 v/v) as the solvent and a flow rate of 0.5 ml/min. The UV detector was set to monitor absorbance at 234 nm. A. l l(R/S)-HETE methyl ester standard; B. ll-HETE methyl ester from incubation of oPGHS-l with arachidonate; C. Coinjection of 1 1-HETE methyl ester from incubation of oPGHS-l with 100 uM arachidonate and ll(R/S)-HETE methyl ester standard; D. 15(R/S)-HETE methyl ester standard; E. lS-HETE methyl ester from incubation of oPGHS-l with arachidonate; F. Coinjection of lS-HETE methyl ester from incubation of oPGHS-l with 100 uM arachidonate and 15(R/S)-HETE methyl ester standard. 48 A234 IlR-HETE i [its-m . \A (I '1 1 l [LR-BETH ‘ ‘W { l lS-HEI'E U o 81 Time (min) A234 49 L l 1; 1512-11511: ”smart . . n i , 4 t r 1 r r 4 h . lSS-HETE - ‘ lSR-HEI'E 1 1‘ D t P fi‘ r i ‘F L 1 . l i lSS-HETE . ‘ lSR-HETE / ’ , . l P 1 P l . + I T r r r 30 Time (min) ii I mun-'1'“ -‘ A\..n - Flgu 0P(,‘ iflttl HPL. 15131415115] b l t . . (Q ‘ l‘ 1 .1 e i a; .......... . ..... i ............. L ~- s s i . i 'r r. ; ...... ..- . . .1 r _ e g . ; 15(R)-HEIE : l 7 I u-n: > I“ 3: t bf: V 1 : i l t t 1 . ‘0 9515)-}!er :- . , i“ = . datum-21.5 ... . O... .......... .; .. . Ill! its 111 3b 35 2uM 2024 10uM 20:4 50uM 2024 Time (min) Figure 11. Chiral HPLC of the methyl esters of the lS-HETEs produced by native oPGHS-l at increasing arachidonate concentrations. Arachidonate (2-100 uM) was incubated with partially purified oPGHS-l for one min at 37°C. RP- and chiral phase HPLC was carried out as described in Figure 10. 50 .Axvmhm:-m~ Ea AOer—(mm-: .m 28 ADV 3269a wo>to©k00m .< .358595 8833 Ewfi mo omfigm SN 80¢ 0.8 338% .mEfiwSmEoEo him— EE :o 92:55 038868 mo 3888a 23 oEEcoHow 8 new: $3 @5550 5:22:38 2:34 .2855 NmOm no $525 8505— 55, cemcwafioo 98 308 xmoq wEws 38328 803 $260K mo 355:8 2: .0133 8m 25502 use flutes—2 5 contact mm xzmfiwofifiofio 8b: 55 068 5:23 Canada 3 BEA—mam 325 3260.5 .EE oco com Dobm 3 2235538 22828.8 mo maoumbcoocoo 3:885 05 53» @2335 33 3E: ommcowxxoofixo mv TmEOmo coming 333:5 .Tmluvmo 25a: 3 mhm:-m— can aka—.7: .«UUA— 3 55.258 2: :e not—«.5538 Eon 3.31322: we Subm— .N~ «...-ME 5| A23 Eizoemuéfi fio Nd - md .. To 52 u vod r mod ELLVH (uiw/ionpmd SQIOIIIU) A _ .S E ‘3 m “o F‘ e g5 o. m 2 o E _ 5 KM—S $ 80 100 Figure 13. Effect of arachidonic acid concentration on the formation of PGG2 and 15-HETE by V349L oPGHS-l. V349L oPGHS-l (75ug protein from transfected COS- ] cells) was incubated with the indicated concentrations of arachidonate at 37°C for one min. Products were analyzed by thin layer chromatography and quantitated by liquid scintillation counting of radioactive products from thin layer chromatograms. Results are an average from four separate transfections. PGG2-derived products (0) and 15- HETE(O). 53 “A "1'.” ““51: "Vh.- We... ‘37:; concentration of arachidonate tested (2 pM). Product formation was analyzed after 1 min incubations by radio thin layer chromatography and/or by RP-HPLC. The key finding of these experiments is that the KM values for the formation of PGGz, ll-HETE and 15- HETE by native oPGHS-l and the KM values for PGG; and lS-HETE formation by V349L oPGHS-l were different for each product (Table 1). Because the KM values were different for the formation of each product, it was of interest to determine if the IC50 values for the inhibition of P662, ll-HETE and 15- HETE synthesis would be different or the same. Accordingly, we determined IC50 values for inhibition by ibuprofen of each of the products formed by partially purified native oPGHS-l (Figures 14 and 15) and by the V349L oPGHS-l mutant (Figures 14 and 16). Ibuprofen was chosen because it is a simple competitive, reversible cyclooxygenase inhibitor [157]. The [€50 values for inhibition of PGG2, ll-HETE and lS-HETE were the same (170 pM) for native oPGHS-l; moreover, the IC50 values for the inhibition of P662 and lS-HETE formation by V349L oPGHS-l were also identical (15 pM). 54 Ta ....... I. Ilianj.» .Il'r . I l I Table 1. Catalytic Constants for the Formation of Prostaglandin and Monohydroxy Acid Products from Arachidonate by Native and V349L oPGHS-l. Partially purified oPGHS-l (5 cyclooxygenase units) or membrane-associated V349L oPGHS-l (75 pg of cell protein) was incubated with increasing concentrations of arachidonate at 37°C for one min. KM and Vmax values for PGGz-derived and HETE products were calculated from RP-HPLC and radio thin layer chromatography product analyses as described under Experimental Procedures and illustrated in Figures 12 and 13. Data are from an average of eight and four separate experiments for native and V349L oPGHS-l, respectively and are expressed +/- S.E.M. ND, not determined. ap< .01 versus PGG; for oPGHS-l; bp<.05 versus ll-HETE for oPGHS-l; cp<.001 versus PGGz for V349L. oPGHS-l V349L oPGHS-l Products I(M Vmax KM Vmax (uM) (nmoles (pM) (nmoles product/min) product/min) PGG2-derived 5.5 +/- 1.7 0.9 +/- .05 11 +/- 0.54 0.43 +/- 0.04 products 1 l-HETE 12 +/- 1.43 0.04 +/- .001 ND ND lS-HETE 19 +/- 4.3” 0.03 +/- .006 5 +/- 0.7“ 0.12 +/- 0.04 55 Figure 14. Autoradiogram of thin layer chromatograms from ICso determinations for inhibition of native (upper) and V349L (lower) oPGHS-l by ibuprofen. Partially purified oPGHS-l (5 units) or broken cell preparations of COS-1 cells expressing V349L oPGHS-l (75 pg of cell protein) were incubated for one min with 10 uM [l-MC] arachidonate in the presence of increasing concentrations of ibuprofen. Products were extracted, separated by thin layer chromatography and quantitated as described in Methods. 56 Native oPGHS-l E T E H. U E T E H. J PGG2- derived products [IBP] (mM) V349L oPGHS-l E T. E H w. E T E H. 1 P662- derived products [1131’] (PM) 57 120 - 100 1”,”--- .Q‘ 30 ‘ .2 *6 < 60 - o\° 40 _ 20 _ O T T FT 1 IIIII l T T I Figure 15. Inhibition of oPGHS-l by ibuprofen. Partially purified oPGHS—l (5 cyclooxygenase units) was incubated with 10 pM arachidonate and the indicated concentrations of ibuprofen at 37°C for one min. ICSO values for inhibition of each product were calculated from quantitation of radioactive products from thin layer chromatograms. Results are from four independent experiments. PGG2-derived products (0); ll-HETE(O); and 15-HETE(><). 58 1207 100— 15\ / \ \ 80- \ / .Q‘ .2. ‘ *5 <1 60- .\° . 40- == 20‘ \a 4.— _J_. §g o ., . -....n, . “4-2., . Wm... .---.m, 10'8 10'7 10’6 10‘5 10‘ 1113?} (M) Figure 16. Inhibition of V349L oPGHS-l by ibuprofen. V349L oPGHS-l (75pg of protein from transfected COS-1 cells) was incubated with 10 pM arachidonate and the indicated concentrations of ibuprofen at 37°C for one min. IC50 values for inhibition of each product were calculated from quantitation of radioactive products from thin layer chromatograms. Results are from four independent experiments. PGGz-derived products (a) and lS-HETE(O). 59 Discussion There are now several x-ray crystallographic structures of PGHS-l and -2. All of the structures reported to date have nonsteroidal anti-inflammatory drugs bound [80-82]. These crystallographic studies have defined the cyclooxygenase active site as a tunnel lined with the side chains of hydrophobic amino acids that protrudes into the core of a globular catalytic domain; the globular catalytic domain itself closely resembles that of mammalian peroxidases. We are currently addressing the question of how arachidonate itself binds within the cyclooxygenase active site at the time the rate determining step in the reaction, abstraction of the 13-proS hydrogen [61], occurs. These ongoing studies involve a combination of determining the crystal structure of arachidonate bound to a catalytically inactive form of oPGHS-l with Coy-heme substituted for Fey-heme and analyzing the functional effects of amino acid substitutions of cyclooxygenase active site residues. In establishing a baseline from which to interpret the results of these crystallographic and mutagenic analyses, it was essential to determine precisely which products are formed by oPGHS-l under different experimental conditions. Accordingly, we characterized the products generated when arachidonic acid was incubated with native oPGHS-l and an active site mutant, V349L oPGHS—l, and determined kinetic constants for the formation of the different products. Experiments were performed comparing partially purified with recombinant oPGHS-l to verify that the results are applicable to solubilized and membrane-associated forms of the enzyme. Our key finding is that the KM values for the formation of PGG2, llR-HETE, and lSS/R-HETE by native oPGHS—l differ from each other. Similarly, with V349L oPGHS- 1, which forms a quantitatively different mix of products, primarily PGG2 and 15-HETE 6O but little ll-HETE, the KM values for formation of these products differed fiom one another. If the rate determining step in the formation of the various products is, as expected [61, 158], abstraction of the 13-proS hydrogen from arachidonate, differences in KM values would only occur if there are different enzyme substrate complexes leading to the different products. The simplest interpretation of our findings is that arachidonate can assume up to three catalytically competent "arrangements" in the cyclooxygenase active site of native oPGHS-l in which the l3-proS hydrogen is removed in the rate determining step, but from which oxygenation and cyclization proceed differently to form PGG2, 11R- HETE or a combination of 15S- and 15R-HETE depending on the arrangement of arachidonate at the instant of hydrogen abstraction. These arrangements could involve subtle differences in carbon chain conformations, differences in electrostatics along the carbon chain of arachidonate or in active site residues or perhaps, most likely, a combination of these factors. For example, a conformation of arachidonate having an optimal orientation of C-9 with respect to GM would result in formation of the C-9 to C- 11 endoperoxide and proceed to PGGz; alternatively, if the orientation between C-9 and OH is not optimal-—either too great or too near--the incipient ll-peroxyl radical could not proceed to the endoperoxide but would yield the ll-hydroperoxide as the abortive product. Previous studies have shown that ll-hydroperoxide product, ll-HPETE is not converted to PGG2 by oPGHS-l. There are two general possibilities which could account for the fact that different products are formed with different KM and Vmax (Figure 17, Schemes 1 and 2). One possibility is that there is essentially one form of the enzyme (E) which binds arachidonate and that, secondarily, rearrangements of the substrate (and/or enzyme) occur 61 afier the substrate becomes bound which then proceed to product (Scheme 1). The basic assumption of this scheme is that, changes which occur either in substrate position, enzyme position or a combination of the two, occur only after substrate has entered the cyclooxygenase active site. That is, the enzyme does not undergo any conformations changes in the free form. The other general possibility is that different starting forms of the enzyme occur (E1 and E2) which bind substrate differently and yield different products (Scheme 2). The basic assumption of Scheme 2 is that changes are occuring only in the free enzyme form and it is this conformational change in the enzyme which results in variant binding of the substrate. As a test of these two possibilities, we determined the [€50 values for inhibition of the formation of each product by ibuprofen. In Scheme 1, ibuprofen would be expected to inhibit the formation of each product with the same ICso. Indeed, this is the experimental observation with both native oPGHS-l and V349L oPGHS-l which leads us to favor the assumptions of Scheme 1 which dictate that conformational changes, whether in substrate alone, enzyme alone, or the substrate/enzyme complex, occur only as a result of substrate binding and not before. In Scheme 2, one would expect ibuprofen to bind different forms of the enzyme differently and yield different IC50 values for each product. The experimental data suggest that this scenario is less likely. Also, structural data which is available from multiple crystal structure determinations of both PGHS-1 and PGHS-2 indicate that the free enzyme conformation is the same, not only among determinations of the same structure, but also between determinations of PGHS-l compared with PGHS-2. Of course, we cannot rule out the possibility that different forms of the enzyme bind 62 751 1(1 K E111 SIC-l- E1+Soé E|S| A E+P1 Kati K 2 K E212 $1.4 E2 + s0 :3 E282 —’k“ E + P2 V Figure 17. Kinetic schemes for the binding of arachidonate to a single enzyme form and for the binding of arachidonate to a mixture of different enzyme forms. A. Scheme 1: arachidonate enters a static cyclooxygenase active site and then can assume alternative conformations or arrangements leading to different products (e. g. P602, 1 l- HETE or lS-HETE). This scheme and its equations also apply to situations where arachidonate binding induces changes in the active site. B. Scheme 2: arachidonate binds to different forms of the cyclooxygenase active site which catalyze the formation of different products; a variation of this scheme applies to situations in which different forms of the enzyme bind different forms of the substrate to produce different products. 63 ibuprofen identically but that the different enzyme forms bind the substrate in different arrangements. Also, we realize the limitations of crystallography, in that certain enzyme forms may very well crystallize more easily than others. Thus, while the results of the experiments with ibuprofen are consistent with Scheme 1, they do not rule out Scheme 2 and there are situations during which ICSO values for the inhibition of all products could be equal under the assumptions of Scheme 2. PGHS-1 and -2 have been shown previously to form small amounts of the monooxygenation products, 11— and lS-HETEs [61, 154, 155]. We found that oPGHS-l forms an additional product, lSS—HETE, which is apparently not produced from arachidonate by PGHS-2 [155]. In fact, lSS—HETE is the predominant lS-HETE enantiomer formed by oPGHS-l whereas both PGHS-2 and aspirin-acetylated PGHS-2 form almost exclusively lSR-HETE [108, 155]. Both ISS-HETE and lSR-HETE would be expected to be formed under conditions favoring a combination of formation of lS-arachidonyl radical and 02 access to C-lS. Formation of lSR-HETE by aspirin-acetylated PGHS-2 involves abstraction of the l3-proS hydrogen as does formation of PGG; (and presumably llR-HETE and 15S- HETE) [61, 154]. We assume, therefore, that the lSR-HETE formed in our experiments involves abstraction of this same prochiral hydrogen. The 15S-/15R-HETE ratio remained constant for lS-HETE synthesized by oPGHS-l at arachidonate concentrations between 2 and 100 pM arachidonate (Figure 11). This was also true with the V349L oPGHS-l mutant. Collectively, this information suggests that 15S— and lSR-HETE derive from the same arrangement of arachidonate in the cyclooxygenase site and that the 15S—/15R-HETE ratio simply reflects positional access of 02 to a 15-arachidonyl radical 64 formed by removal of the l3-proS hydrogen. Although antarafacial 02 addition is observed in most lipoxygenase reactions, there is recent precedent for a suprafacial 02 addition with a fungal lipoxygenase [159]. The sidechain of Va1349 protrudes into the cyclooxygenase active site approximately opposite the sidechain of Ser530, the site of aspirin acetylation [21, 80, 99]. Our results indicate that replacing Va1349 with the larger leucine has the overall effects of decreasing the amounts of products formed as a result of 02 insertion at C-ll (P662 and ll-HETE) and increasing the amount of product involving 02 insertion at C- 15 (IS-HETE). That is, with the V349L mutant, arrangements of arachidonate in the active site which lead to 02 insertion at C-ll become relatively less favorable than in native oPGHS-l. The KM for P062 formation is increased and P662 and ll-HETE formation occurs to a significantly decreased extent with V349L oPGHS-l. In contrast, 02 access to C-15 is relatively enhanced with a relative increase in the rate of monooxygenation at C-15 and a relative decrease in the KM for 15-HETE production. We have emphasized in this report the fact that oPGHS-l forms products other than PGGz. However, in concluding we point out that the KM and Vmax values favor the synthesis of this endoperoxide over monohydroxy acids by native oPGHS-l. At arachidonate concentrations (52 pM) expected to be most relevant physiologically, PGGZ represents >95% of the total oxygenated product. 65 Acknowledgment This work was previously published in the Journal of Biological Chemistry as Thuresson, ED, Lakkides, KL, and Smith WL (2000) “Different catalytically competent arrangements of arachidonic acid within the cyclooxygenase active site of prostaglandin endoperoxide H synthase-l lead to the formation of different oxygenated products.” J.BioI.Chem., 275, 8501-8507., and is reprinted with permission. 66 CHAPTER III THE FUNCTIONS OF CYCLOOXYGENASE ACTIVE SITE RESIDUES IN THE BINDING, POSITIONING AND OXYGENATION OF ARACHIDONIC ACID Introduction Prostaglandin endoperoxide H synthases-l and -2 (PGHS-l and -2) catalyze the conversion of arachidonic acid (AA), two molecules of Oz and two electrons to PGH2. This is the committed step in the formation of prostaglandins and thromboxane A2 [2, 3, 137]. PGHS-l (or COX-1 (for cyclooxygenase-1)) is a constitutive enzyme while PGHS- 2 (COX-2) is the inducible isoform [2, 3, 137] [138-140, 142-146, 160]. PGHSs catalyze two separate reactions including a cyclooxygenase (bis— oxygenase) reaction in which AA is converted to PGG2 and a peroxidase reaction in which PGGz undergoes a two- electron reduction to PGH2 [2, 3, 137]. These reactions occur at physically distinct but interactive sites within the cyclooxygenase structure. The cyclooxygenase reaction begins with a rate-limiting abstraction of the 13-proS hydrogen from AA to yield an arachidonyl radical [61, 70]. This is followed by sequential oxygen additions at C-11 and C-15 producing PGGZ. NSAIDs compete directly with AA for binding to the cyclooxygenase site [80, 82]thereby inhibiting cyclooxygenase activity (but not peroxidase activity; [55, 152, 153]). Crystallographic studies of enzyme-inhibitor complexes have suggested that the cyclooxygenase active site exists in the form of a hydrophobic channel that protrudes into the body of the major globular domain of the protein [80-82]. More recently, the structure of AA bound within the cyclooxygenase active site of ovine (o) PGHS-l was 67 determined [66]. AA is bound in an extended L-shaped conformation and makes a total of 48 hydrophobic contacts (i.e. 2.5-4.0 A) and two hydrophilic contacts with the protein, involving a total of nineteen different residues (Figure 18). Additionally, there are several amino acids which are in the first shell of the cyclooxygenase hydrophobic tunnel and contact other first shell amino acids but lie outside of van der Waals distance to AA. Although AA can assume some 107 low energy conformations [161], only three of these conformations are catalytically competent [104]; one conformation leads to PGGz, one, to llR-HPETE and another to 15R- plus lSS-HPETE. Previous mutational studies have demonstrated that the guanidinium group of Arg120 is important for high affinity binding of AA to PGHS-l [96, 97](but not PGHS-2; [98, 162]), that Tyr385 is involved as a tyrosyl radical in abstracting the l3-proS—hydrogen fiom AA [74, 95], and that Ser530 and Ile523 are determinants of inhibitor specificity [21, 111, 112, 156, 163]. Other than for Arg120 and Tyr385 relatively little is known about the functions of residues located in the core of the hydrophobic cyclooxygenase tunnel [95, 164]. In the study reported here we have performed mutational analyses of a number of the residues that line the hydrophobic active site channel to determine their functional importance in arachidonate binding and oxygenation. Our results suggest that individually and collectively these residues function primarily to position arachidonate in a specific conformation that optimizes its conversion to P002. 68 Figure 18. Schematic representation of interactions between arachidonic acid and amino acid residues lining the cyclooxygenase active site channel. All dashed lines represent interactions of g 4.0 A between specific side chain atoms of the protein and carbon or oxygen atoms of AA [66]. The distances between various carbons of AA and interacting sidechain residues are listed in Tables III and IV. 69 Materials and Methods Materials. Fatty acids were purchased from Cayman Chemical Co., Ann Arbor, MI. [l-MC] arachidonic acid (40-60 mCi/mmol) was purchased fiom New England Nuclear. Flurbiprofen was purchased from Sigma. Diazald (N-methyl-N-nitroso-p- toluenesulfonamide) was from Aldrich Chemical Company. Restriction enzymes and Dulbecco’s modified Eagle’s medium (DMEM) were purchased from GIBCO. Calf serum and fetal bovine serum were purchased from HyClone. Primary antibodies used for Western blotting were raised in rabbits against purified oPGHS-l and purified as an IgG fraction [149] and goat anti-rabbit IgG horseradish peroxidase conjugate was purchased from BioRad. Oligonucleotides used as primers for mutagenesis were prepared by the Michigan State University Macromolecular Structure and Sequencing Facility. All other reagents were from common commercial sources. Preparation of oPGHS-I mutants. Mutants were prepared either starting with Ml3mpl9-PGHS0v which contains a 2.3-kilobase Sal] fragment coding for native oPGHS-l and employing the BioRad Muta-Gene kit and the protocol as described by the manufacturer [149] or by site-directed mutagenesis of oPGHS-l in the pSVT7 vector [95] employing the Strategene QuikChange mutagenesis kit and the protocol of the manufacturer. Oligonucleotides used in the preparation of various mutants are summarized in Table II. Plasmids used for transfections were purified by CsCl gradient ultracentrifugation and mutations were reconfirmed by double-stranded sequencing of the pSVT7 constructs using Sequenase (ver. 2.0, US. Biochemical Corp.) and the protocol described by the manufacturer. 70 T ransfectt'on of C OS-I cells with oPGHS-I constructs. COS-1 cells (ATTC CRL- 1650) were grown in DMEM containing 8% calf serum and 2% fetal bovine serum until near confluence (~3x10° cells/10 cm dish). Cells were then transfected with pSVT7 plasmids containing cDNAs coding for native oPGHS-l or mutant oPGHS-l using the DEAE dextran/chloroquine transfection method as reported previously [149]. Forty hours following transfection, cells were harvested in ice cold phosphate-buffered saline (PBS), collected by centrifugation and resuspended in 0.1 M Tris-HCI, pH 7.5. The cells were disrupted by sonication and microsomal membrane fractions were prepared at 0- 4°C, as described previously [149]. Membranes were isolated from sham-transfected cells in an identical manner. Protein concentrations were determined using the method of Bradford [165] with bovine serum albumin as the standard. Microsomal preparations were used for Western blotting and for cyclooxygenase and peroxidase assays. C yclooxygenase and peroxidase assays. Cyclooxygenase assays were performed at 37°C by monitoring the initial rate of 02 uptake using an oxygen electrode [94, 96]. Each standard assay mixture contained 3.0 m1 of 0.1 M Tris-HCI, pH 8.0, 1 mM phenol, 85 pg of bovine hemoglobin and 100 pM arachidonic acid. Reactions were initiated by adding approximately 250 pg of microsomal protein in a volume of 20-50 p1 to the assay chamber. KM values were measured using 0.5-500 pM arachidonate. Inhibition of cyclooxygenase activity was measured by adding aliquots of microsomal suspensions to assay mixtures containing 100 pM arachidonate and 200pM flurbiprofen. Peroxidase activity was measured spectrophotometrically with N,N,N',N‘- tetramethylphenylenediamine (TMPD) as the reducing cosubstrate [166]. The reaction mixture contained 0.1 M Tris-HCl, pH 8.0, 0.1 mM TMPD, approximately 100 pg of 71 microsomal protein and 1.7 pM hematin in a total volume of 3 ml. Reactions were initiated by adding 100 ml of 0.3 mM H203 and the absorbance at 610 nm was monitored with time. Inhibition of cyclooxygenase activity. IC50 values were determined using cyclooxygenase assays as described above and rate measurements of oPGHS-l enzymes with 50pM arachidonate and increasing concentrations of inhibitors. Western blot analysis. Microsomal samples (ca. 5 pg of protein) were resolved by one-dimensional SDS-PAGE and transferred electrophoretically to nitrocellulose membranes using a Hoeffer Scientific Semi-Dry Transfer apparatus. Membranes were blocked for 12 hr in 3% non-fat, dry milk, 0.1% Tween20 and Tris-buffered saline, followed by a two-hour incubation with a peptide-directed antibody against oPGHS-l [84] in 1% dry milk, 0.1% Tween-20 and Tris-buffered saline at room temperature. Membranes were washed and incubated for one hr with a 1:2000 dilution of goat anti— rabbit IgG-horseradish peroxidase, afier which they were incubated with Amersham ECL reagents and exposed to film for chemiluminescence. Characterization of arachidonic acid-derived products. A general protocol for product analysis is as follows. Forty hr following transfection, COS-1 cells were collected, sonicated and resuspended in 0.1 M Tris-HCI, pH 7.5. Aliquots of the cell suspension (100-250 pg of protein) were incubated for 1-10 min at 37° C in 0.1 M Tris- HCl, pH 7.5, containing 1 mM phenol and 6.8 pg bovine hemoglobin in a total volume of 200 pl. Reactions were initiated with 35 pM [l-MC] arachidonic acid and were performed with or without 200 pM flurbiprofen and stopped by adding 1.4 ml of CHCl3zMeOH (1 :1; v/v). Insoluble cell debris was removed by centrifugation and 0.6 m1 of CHC13 and 0.32 72 ml of 0.88% formic acid were added to the resulting supernatant. The organic phase was collected, dried under N2, redissolved in 50 pl of CHCI; and spotted on a Silica Gel 60 thin layer chromatography plate; the lipid products were chromatographed for 1 h in benzene:dioxane:formic acidzacetic acid (82:14:12]; v/v/v/v). Products were visualized by autoradiography and quantified by liquid scintillation counting. Negative control values from samples incubated with 200 pM flurbiprofen were subtracted from the experimental values observed for each sample in the absence of flurbiprofen. For RP-HPLC analyses of products, native or mutant oPGHS-l (1 mg of microsomal protein) was reacted with 100 pM arachidonic acid for 30 minutes at 37° C, and products were collected as described previously. Products were dried under N2 and resuspended in HPLC buffers (1:1; v/v). 15- and ll-HETEs were separated by reverse phase HPLC using a C18 column (Vydac); the Waters Model 600 HPLC was equipped with a 990 photo diode array detector set to 200 and 234 nm. The strong component of the mobile phase was 0.1% acetic acid and the eluting solvent was acetonitrile. The flow rate was 1 ml/min. The following elution profile was used: Initial conditions were 30% acetonitrile, increased linearly over 30 min to 50% acetonitrile, then increased linearly from 30-100 min to 70% acetonitrile, then increased linearly from 100-125 min to 100% acetonitrile and held for 5 min before returning to initial conditions. The retention times of lS-HETE and ll-HETE averaged 36 and 38 min, respectively. Material obtained by RP-HPLC was esterified by treatment with excess diazomethane and subjected to chiral- phase HPLC. Chiral-phase HPLC separations of the methyl esters of l l- and lS-HETEs were performed with a Chiralcel OC column (250 X 4.6 mm; Daicel Chemical Industries, Osaka) using hexane/Z-propanol (90:10; v/v) as the solvent and a flow rate of 0.5 ml/min. 73 Diazomethane was prepared from Diazald and distilled in ether per the Aldrich Chemical Company Technical Bulletin #AL-180. Molecular modeling . Mutations were modeled and analyzed in the program CHAIN [167] using the coordinates from the crystal structure of Co3+-0PGHS-l complexed with AA (PDB entry lDIY) and Fe3+-0PGHS-l complexed with flurbiprofen (PDB entry lCQE). 74 Table II. Oligonucleotide Primers Used for Preparing oPGHS-l Mutants Mutants were prepared using two different procedures. For those made using the BioRad Mutagene kit, short single-stranded Oligonucleotide primers were used. For PCR mutagenesis, double standed primers were used; in the latter cases, designed with an asterisk (*), only a single strand is shown. Mutation sites are underlined. The numbering refers to the location in cDNA for oPGHS-l [22]. Mutation Oligonucleotide Primer F205A* 5'-°°7CCTTCTITGCTCAGCACG_C§ACCCATCAGTTC'ITCAAAACTI'CC-B' 1:20er 5'-°°7CCTI‘CTITGCTCAGCACQIQACCCATCAGTTC'ITCAAAACTTCC-3' F209A“ 5'-°°7CCTTCTITGCTCAGCACTTCACCCATCAGQTTCAAAACTICCfi' F209L“ 52687CCTTCTITGCTCAGCACTTCACCCATCAGCT_CTTCAAAAC”ITCC-3' Y348L 5'3'”TGATAGAGGAGC_‘T_TGTGCAGCAG-3' Y348W 5'-"”GATAGAGGAGIQQGTGCAGCAGC-B' V349A 52' '27 AGAGGAGTATQQQCAGCAGCTGA-T V349L 5'-' '”AGAGGAGTATCJQCAGCAGCTGA-s V3498 5'-"27AGAGGAGTATT_C_GCAGCAGCTGA-3' V349T 5'-"”AGAGGAGTATAQCAGCAGCTGA-s S353A" 5'- "”ccAGCAGCTGQQGGCTACWCCTGCAGC-3' s353G* 5'-"”GCAGCAGCTGQQQGGCTACTTCCTGCAGC-3' 8353T“ 5'-"”GCAGCAGCTGA_CCGGCTACTTCCTGCAce-3' r377v* 5'-'2°5CCAGTACCGCAACCGCQLCGCGATGGAGTTCAACC-3' F381A“ 5'-'2'°GCATCGCCATGGAQQQAACCAGCTGTACC-T F381L“ 5'-'2'°GCATCGCCATGGACJCAACCAGCTGTACC—3' 1523A“ 5'-'“SCCATCTITGGGGAGAGTATGQLAGAAATGGGGGCTC-3" S530v 5'-'°7°GGCTCCT1TfflCTTAAGGGCC-3' L531A 5'-'°73rccrrrrrCC_G_C_TAAGGGCCTCT-3' L531V 5'-'°73TCCTT'I'ITCCG__TFAAGGGCCTCT-3' L534A" 5'-'°°2CCTTAAGGGC§flCTTGGAAACCCCATCTGTTCTCC-3' L534V* 5'-'°°2CCTTAAGGGCG_TCCTTGGAAACCCCATCTGTTCTCC-3' 75 Table III. Contacts between Arachidonic Acid (AA) and Cyclooxygenase Active Site Residues Van der Waals and hydrogen bond interactions were calculated using CHAIN [167]. Van der Waals contacts within 4A are listed. See list of abbreviations for carbon designations. Residue Atom Residue Atom Distance (A) Phe 205 CE2 AA C14 4.0 CE2 C15 3.4 CZ C15 3.9 CZ C18 3.9 Phe 209 CD1 AA C19 3.9 CEl C19 3.6 CE2 C17 3.9 CZ C17 3.7 CZ C18 3.9 CZ C19 3.7 Tyr 348 CE2 AA C12 3.2 CE2 C13 3.9 CE2 C14 3.6 Val 349 CGI AA C3 3.4 CGl C4 3.1 Ser 353 CB AA C3 4.0 "e 377 CD1 AA C20 3.7 Phe 381 CE2 AA C17 3.8 CE2 C20 3.9 CZ C16 3.9 CZ C17 3.5 Ile 523 CG2 AA C2 3.8 CG2 CS 3.3 CG2 C6 3.4 Ser 530 CA AA C16 3.7 CB C10 3.5 CB C16 3.8 Leu 534 CD1 AA C18 3.9 CD2 C15 3.5 CD2 C16 3.6 76 Table IV. Interactions Between Leu531 and other Cyclooxygenase Active Site Area Residues Leu531 does not contact arachidonic acid, but is involved in a number of hydrogen bonding interactions with other active site area residues. Hydrogen bonds were calculated using CHAIN [167]. Hydrogen bonding was assigned for O-H-"N, C=O°--H and H-N—C bond angles > 90° and distances not exceeding 3.6A. Residue Atom Residue Atom Distance (A) Comments [cu 531 N Ala 527 O 2.8 main chain H-bond 0 Gly 533 N 3 main chain H-bond O Leu 534 N 3.1 main chain H-bond 0 Leu 535 N 2.9 main chain H-bond 77 Table V. Kinetic Properties and Product Analyses for oPGHS-l Cyclooxygenase Active Site Mutants Peroxidase activity was measured spectrophotometrically using 0.2 mM H202 and 100 pM TMPD as substrates and oxygenase activity was measured with an oxygen electrode as described in the text. Values are calculated for arachidonic acid turnover and are corrected for the percentage of mono- and bis-oxygenated products formed with the mutants. A value of 100% is assigned for peroxidase and oxygenase activity of native oPGHS-l. Vmax and KM values represent the means from a minimum of four separate determinations with standard deviations within 10% of all values reported. Relative Vmax values reported for mutant enzymes for which KM values were not determined represent rate measurements taken using 100 pM arachidonate. ND, not determined. ENZYME OXYGENASE PEROXIDASE l l-HETE 1 S-HETE (% of Control) (% of Total (% of Total Vmax (%) KM Vim/KM Products) Products) (11M) NATIVE 100 2 50 100 2.5 (11R) 2.5 (155') SHAM 0 -- -- 0 -- -- F 205A 28 3 9.3 90 l l 5 F 205L 65 ND ND 106 5 2 F209A 15 10 1.5 55 13 3 F209L 43 ND ND 104 5 2 Y348L 0 -- -- 9 -- .. Y348F 83 2.9 29 76 2.1 2.2 [74.75.95] Y348W 0 -- -- 0 -- -_ V349A 55 1.7 32 52 53 (11R) 0 V349A/ 3 ND ND 49 27 12 S353T V3495 43 14 3.2 94 41 (11R) 0 V349T 39 13 3.0 136 5.5 (11R) 0.6 V349L 63 7.1 8.9 117 0 13(15S), l 1( 15R) S353G 61 2 30.5 87 4 2 S353A 56 ND ND 60 3 3 S353T 42 ND ND 46 14 12 I377V 72 1.1 65 110 10 2 78 Table V (cont’d) F381A 4 6 0.7 101 9 4 F381L 21 ND ND 74 7 1523A 64 ND ND 112 12 3 1523v 70 ND ND 57 2.5 2.5 S530A 58 2 29 68 2.5 2.5 [21, 156] S53OT 17 13 1.3 30 0 10 (15R) [21, 156] ss30v 0 -- -- 94 -- -- L531A 4.6 54 .08 100 2.5 2.5 L531D 13 2 6.5 26 2.5 2.5 [156] L531N 7 3 2.3 9 2.5 2.5 [156] L531V 7.7 1.6 4.8 75 2.5 2.5 L53ll 19 .4 48 74 ND ND [156] L531K 0 -- -- 0 -— -- [156] L534A 59 8 7.4 72 2 56 (95 5, 5R) L534V 98 ND ND 102 8 47 (96 s, 4R) 79 Results and Discussion Overview. As illustrated in Figure 18, arachidonic acid (AA) is bound in an extended L-conformation in the AA/Co3+-heme oPGHS-l co-crystal structure [66]. The carboxylate group of AA interacts with Arg120, the (0 end abuts Ile 377 and Gly533, the l3-proS hydrogen is appropriately aligned with Tyr385, and there is ample space for the first 02 insertion at C-11 and facile bridging of the incipient ll-hydroperoxyl radical to O9 to form the endoperoxide. Formation of the cyclopentane ring is proposed to involve rotation about the C-lO/C-ll bond and consequent movement of the 0) terminus so that C-12 can react with the C-8 radical that is produced upon formation of the endoperoxide group; this movement, in turn, positions 015 adjacent to Tyr385 for a second antarafacial O2 insertion and a one electron reduction of the lS-hydroperoxyl radical by Tyr3 85 to regenerate the Tyr385 radical [66]. In the AA/Co3+-heme PGHS-l co-crystal structure AA makes two hydrophilic and 48 hydrophobic contacts involving a total of 19 residues in the first shell of the cyclooxygenase active site (Figure 18, Table III; [66]). These studies involve mutations of 10 of the residues that are putatively involved in direct interactions with AA as well as one other residue (Leu531) which is in the first shell of the active site but does not directly contact the substrate (Figure 18; Tables III and IV). In analyzing the various mutants, we identified the AA oxygenation products, determined kinetic constants for AA oxygenation and measured peroxidase activity (Table V). Cyclooxygenase and peroxidase activities for native and all mutant enzymes were normalized to levels of protein expression determined from western blot analysis and densitometric quantitation. The results of western transfer blotting indicated that the native enzyme as well as all 80 mutants tested were expressed at similar levels by COS-1 cells as shown in a representative figure (Figure 19). A discussion of the implications of the results are presented below for individual residues. These studies identify residues which fall into three functional groups: (a) residues essential for positioning C-l3 of AA for hydrogen abstraction (Tyr348); (b) residues critical for conforming AA such that when hydrogen abstraction does occur the AA is appropriately arranged to yield PGG2 (Va1349 and L534); and (0) various other active site residues, which individually make lesser but measurable contributions to optimizing catalytic efficiency. The simplifying assumptions that serve as the basis for this categorization are as follows. First, residues, which when altered yield mutant enzymes without cyclooxygenase activity are considered to be critical for positioning C- 13 for hydrogen abstraction by Tyr385. Second, AA can exist in three distinct, catalytically competent conformations in native oPGHS-l which yield PGG2 (95%), 11- HPETE (2.5%) and lSR/S—HPETE (2.5%), respectively [104]; as a first approximation, amino acids residues, mutation of which causes a major change in the composition of the oxygenation products, are considered to be those critical for positioning AA in the conformation that yields PGG2. These categories are meant to draw attention to those residues that exhibit the most dramatic effects on catalytic constants and product profiles. Mutations of other residues have similar effects, but to lesser degrees, and are discussed in subsequent sections of the text. We also emphasize that in interpreting the results of the mutational analyses, we have in most instances made the simplifying assumption that amino acid substitutions alter only interactions with the AA chain and have largely ignored potential effects on interactions among adjoining residues. This is clearly an 81 oversimplification which can only be justified broadly by pointing out that most of the mutant enzymes retain both peroxidase or cyclooxygenase activities; however, subtle changes in inter-residue interactions brought about by various mutations certainly may have important effects on kinetic properties and product compositions. Future studies on the crystal structures of AA complexed with mutant forms of oPGHS-l will be necessary to help resolve the relative influences of mutations on interactions with AA versus those with neighboring amino acid residues. Residues essential for positioning GB of AA for hydrogen abstraction (T yr348). Tyrosine 348. Examination of the crystal structures of oPGHS-l complexed with nonsteroidal anti-inflammatory drugs [80, 99] and the Coy-heme oPGHS-l/AA complex (Figure 18; [66]) suggests that Tyr348 may be involved in the appropriate positioning of Tyr385 and/or AA. The distance between the phenolic oxygens of Tyr348 and Tyr385 suggest that there is a hydrogen bond between these two atoms that is important for positioning Tyr385; however, there are no hydrophobic contacts between Tyr348 and Ty385. The CE2 phenyl ring carbon of Tyr348 is within van der Waals distance of C-12, C-13 and C-14 of AA suggesting that hydrophobic interactions between Tyr348 and substrate could be important in positioning C-13 of AA. Our results indicate that these latter contacts between Tyr348 and AA are essential in positioning C-13 of AA but that there is no functionally important hydrogen bonding between Tyr348 and Tyr385. Results which argue against a significant hydrogen bonding interaction between Tyr348 and Tyr385 are as follows. Previous studies have established that substitution of Tyr348 with phenylalanine has little effect on either the cyclooxygenase or peroxidase activity of the enzyme [74, 75, 95]. A more detailed kinetic analysis of Y348F oPGHS-l 82 indicates that the Vmax/KM is 60% of that of native oPGHS-l and that there are no significant differences in product formation between native and Y348F oPGHS-l (Table V). Additionally, Y348F oPGHS-l forms a tyrosyl radical with properties similar to that of native oPGHS-l [74,75]. The concept that Tyr348 is essential for positioning AA comes from comparing results obtained with native oPGHS-l and the Y348F and Y348L mutants (Table V). Y348F oPGHS-l has properties similar to those of native enzyme. In contrast, Y348L oPGHS-l lacks cyclooxygenase activity while retaining 9% native peroxidase activity; no cyclooxygenase products were detected with Y348L oPGHS-l even with a sensitive radio thin layer chromatographic assay using [l-MC] arachidonic acid as the substrate. Examination of the Coy-heme oPGHS-l/AA complex (Figure 18; [66]) suggests that substituting leucine at position 348 would permit C-13 of AA to move away from Tyr385 increasing the distance between Tyr385 and 013 such that hydrogen abstraction could not occur. There may also be significant pi bond interactions between the phenyl group of Tyr348 and the substrate which are eliminated in the substitution with a leucine residue. The drop in peroxidase activity for this mutant also suggests the possibility of structural changes affecting the peroxidase as well as the cyclooxygenase active site. Tyr348 lies directly below Tyr385 which resides in one of the helices contacting the heme group of the peroxidase active site. Substitution of Tyr348 with a leucine could create a space which might allow movement of this helix and perturbation of the peroxidase active site and could eliminate critical pi bond interactions holding the substrate in a catalytically favorable configuration. Thus, we conclude that the phenyl 83 ring of Tyr348 is essential in positioning C-l3 of AA with respect to Tyr385 but could play a role in the preservation of the structural integrity of the peroxidase site as well. A Y348W mutation disrupted both peroxidase and cyclooxygenase functions; addition of lSS-HPETE to a reaction mixture containing AA and Y348W oPGHS-l failed to activate cyclooxygenase activity. The lack of peroxidase activity observed with the Y348W mutant suggests that this substitution has an effect on the peroxidase active site. We speculate that when a tryptophan group is inserted at position 348, it causes a change in the position of Trp387 and/or Tyr385 which, in turn, leads to movement of His388, the proximal heme ligand, and results in a change in the binding of the heme group so that peroxidase catalysis does not occur. Residues critical for conforming AA such that when hydrogen abstraction does occur the AA is appropriately arranged to yield PGG2 (Va1349, Leu534). Valine 349. One of the methyl groups (CGI) of Va1349 lies within van der Waals distance of both C-3 (3.36 A) and G4 (3.14 A) of AA in the AA/Co3+-heme oPGHS-l crystal structure (Figure 18; [66]). To assess the role of Va1349 in cyclooxygenase catalysis, we substituted this residue with alanine, serine, threonine and leucine and evaluated these mutants. All of these mutant proteins were expressed at comparable levels in COS-1 cells (Figure 19). All of the mutants retained oxygenase and peroxidase activity. The catalytic efficiencies (Vmax/KM values) for the oxygenase reaction ranged from 6-60% of native oPGHS-l (Table V). Lipid soluble products synthesized from [l-MC] arachidonic acid by sham- transfected COS-1 cells and COS-1 cells transfected with native oPGHS-l or various oPGHS-l mutants were separated by thin layer chromatography (Figure 20) or RP-HPLC 84 and the amounts of each product quantified (Table V). Mutants in which Va1349 was replaced with a smaller residue (i.e. V349A, V349S and V349T oPGHS-l) all produced an abundance of ll-HETE versus PGG2-derived products and little or no detectable 15- HETE; the II-HETE/PGG2 ratio for these mutants increased as the size of the group at position 349 was decreased. In contrast, as reported previously [104], replacing Va1349 with a larger leucine residue (V349L oPGHS-l) led to the generation of a relative abundance of lS-HETE and only trace amounts of ll-HETE. Sham-transfected COS-1 cells did not transform [1-14C]arachidonic acid to products. Chiral HPLC analysis of ll-HETEs formed from arachidonic acid by V349A, V349S and V349T oPGHS-l mutants established that the product was exclusively 11R- HETE. A representative chromatographic profile for the chiral analysis of the ll-HETE formed by the V349A oPGHS-l mutant is presented in Figure 21. As reported previously, the V349L oPGHS-l mutant formed a 65/35 mixture of lSS/R-HETE, the same enantiomeric profile obtained with native oPGHS-l [104]. We also examined the interaction of various cyclooxygenase inhibitors with the V349A and V349L mutants. IC50 values for cyclooxygenase inhibition were determined for docosahexaenoic acid (22:n-3), flurbiprofen and flufenamic acid. Inhibition was studied using the O2 electrode assay with 50 pM AA as substrate. The V349A mutant was more sensitive than native oPGHS-l to inhibition by both flurbiprofen (IC50 values of 0.28 pM and 2.9 pM, respectively) and flufenamic acid (ICso values of 0.9 pM and 6.0 pM, respectively); the substrate analogue, docosahexaenoic acid was a less potent inhibitor of the V349A mutant than the native enzyme (IC50 values of 350 pM vs. 100 pM, respectively). 85 Flurbiprofen and flufenamate have similar structures, including two phenyl rings and a carboxylate moiety. The carboxyl groups of flurbiprofen and other carboxylate- containing NSAIDs such as iodosuprofen and iodoindomethacin all form salt bridges with Arg120 at the mouth of the cyclooxygenase active site [80, 96], and the structurally similar flufenamate is thought to be bound in the same manner. Examination of the flurbiprofen-bound crystal structure of oPGHS- shows extensive van der Waals contacts between the C6] of Va1349 and the lower phenyl ring of flurbiprofen [80]. Substitution of Va1349 with an alanine results in more potent inhibition by both flurbiprofen and flufenamate, suggesting that the contacts made by Va1349 are preventing the binding of these inhibitors in the most thermodynamically stable conformation. Thus, additional space created by V349A seems to create a cyclooxygenase active site which can better accommodate the extra bulk of the 2-phenyl inhibitor compounds. In contrast, doxosahexaenoic acid (22:6n-3) is structurally similar to the substrate, arachidonic acid, and was a somewhat less potent inhibitor of the V349A oPGHS-l mutant compared with the native enzyme. 22:6 is a relatively poor inhibitor of native oPGHS-l (ICSO = IOOpM) and an even less potent inhibitor of PGHS-2 which is known to have a slightly larger and more accommodating active site [98]. Because it is elongated and contains two extra double bonds compared with arachidonic acid, 22:6 may be too long and rigid to thread all the way into the cyclooxygenase active site, making it a poor binder. Va1349 makes multiple contacts at the carboxyl end of arachidonate (Figure 18, Table III) and would be expected to contact the same region of 22:6. Stabilization of 22:6 apparently relies somewhat on Va1349 contacts, removal of which creates a larger active site which further reduces the already poor binding affinity of this inhibitor. 86 Native V349AV349L V3498 V349T $53OV SS3OT Sham oPGHS1 3"" 5 ...... Figure 19. Western blot analysis of Va1349 oPGHS-l mutant proteins. Cell protein (5 pg) from COS-1 cells transfected with native or mutant oPGHSs-l was resolved by SDS-PAGE, transferred to nitrocellulose and visualized using a polyclonal antibody raised against oPGHSs-l as described in the text under Methods. Densitometric analysis was performed to determine relative levels of protein expression. 87 We reported previously that AA can assume at least three catalytically competent arrangements in the cyclooxygenase active site of oPGHS-l [104]. These arrangements, occurring at the time of hydrogen abstraction, lead to different products--PGG2, 11R- HPETE or 15R/S-HPETE. With native oPGHS-l the kinetically most favorable arrangement of AA is that which yields PGG2. With V349A oPGHS—l the arrangements yielding PGG2 and ll-HPETE appear to be equally favorable. ll-HPETE would be expected to be formed from an arrangement of AA in which C-9 and C-11 are slightly misoriented such that the endoperoxide bridge between these carbons cannot be formed. The kinetics of both PGG2 and ll-HPETE formation were found to be similar to one another with the V349A mutant (Figure 22), and the Vmax/KM ratios are similar for V349A and native oPGHS-l (Table V). These observations suggest that the valine to alanine substitution does not appreciably influence the positioning of C- 1 3 with respect to Tyr385; moreover, formation of both PGG2 and ll-HPETE presumably proceeds in the same way via abstraction of the 13proS hydrogen and formation of an ll-hydroperoxyl radical [168]. Accordingly, we suggest that Va1349 plays a major role in positioning the carboxyl half of AA and especially in positioning C-9 without appreciably affecting the location of the 0) half of AA within the cyclooxygenase site. Obviously, the effect of Va1349 on the position of C-9 must occur indirectly because Va1349 contacts only C-3 and C-4 of the substrate (Figure 18). Placing smaller amino acids at position 349 apparently permits the carboxyl half of AA greater flexibility, which in turn, translates into a small shift in the orientation of 09 with respect to C-1 1. Leucine 534. The CD1 and CD2 methyl groups of Leu534 are within van der Waals distance of C-15 (3.54 A), C-16 (3.6 A) and C-18 (3.91 A) of AA in the 88 AA/Co3+-heme oPGHS-l crystal structure [66]. Interestingly, substitution of Leu534 with either alanine or valine yields mutant enzymes which produce large amounts of 15- HETE (Table V) and the 15-HETE which is produced is almost exclusively (395%) 15S- HETE (Figure 23); in contrast, with native oPGHS-l lSS-HETE comprises 65% of the total 15-HETE. Apparently, having a hydrophobic residue smaller than leucine at position 534 enlarges the cyclooxygenase active site near C-15 providing for relatively greater access of O2 antarafacial to the site of hydrogen abstraction. Thus, Leu534 is important in facilitating formation of PGG2 versus 15-HPETE. lSS-HETE is the only 15- HETE formed by native human PGHS-2 [155], and the cyclooxygenase site of PGHS-2 is somewhat larger and more accommodating than that of PGHS-1 [137]. 89 Native V349A V349L V3498 V349T oPGHS1 Sham $53OT Arachidonic Acid —’ 15-HETE —> 11-HETE —§ 12-HHTre —) PGD2 —-> PGE2 —> PGlea Figure 20. Thin layer chromatogram of products formed from [l-"Clarachidonic acid by native and mutant oPGHS-l enzymes. Cell protein (250 pg) from COS-1 cells transfected with native or mutant oPGHSs-l was incubated for 10 min with 35 pM [l- l4C]arachidonic acid, and the products were extracted, separated by thin layer chromatography and visualized by autoradiography as described in Methods. The appearance of variability of various product levels between samples shown in the figure is not representative of relative VMAx values among enzymes (determined by oxygen electrode measurements (Table V)) because of the variability in extraction efficiencies among the samples. The locations of the various chromatographic standards are as noted. 90 Figure 21. Chiral HPLC analysis of the methyl ester of ll-HETE produced by V349A oPGHS-l. Arachidonate (100 pM) was incubated for ten min at 37°C with cell protein (1 mg) from COS-1 cells transfected with V349A oPGHS-l. Thell-HETE fraction was isolated by RP-HPLC and esterified by treatment with diazomethane. Chiral-phase HPLC of the methyl ester was performed as described in Materials and Methods using a Chiralcel OC column with hexane/2-propanol (98:2; v/v) as the solvent and a flow rate of 0.5 ml/min. The UV detector was set to monitor absorbance at 234 nm. A. ll(R/S)-HETE methyl ester standard; B. ll-HETE methyl ester derived from incubation of V349A oPGHS-l with arachidonate; C. coinjection of the ll(R/S)-HETE methyl ester standard and the ll-HETE methyl ester from incubation of V349A oPGHS- l with 100 pM arachidonate. 91 FIFDFDKPD- I lFlrbbplhilL>pan-I>LF.I P..Lbhb1pbrhbp , F .P A ->.Vl .k..—7.L..—. LuPrr P r! IerF p».... ..PPtprP» .leh... P Llrphrfivh.«PrththPlktrbE ». AAAAAA-nn‘mA [IR A S l l f ” m llS Vvvfiv vvatV'VY‘Vf‘Y‘ 1" <14thb p>>->pup-pP-brLbbbhhrrbhhlhb-thhbbP-p-nh nub-prh—rPrL-p.hbhhhb—hbhhbhbbP—upbp-h 41 40 Time (min) 35 96 ITT'I'I'I'ITTTITI'I' by I r. i I F V S ”I. S I S r. 5 . 5 r 5 1 II v III. F I 4 .1. W 4 I‘ll Tl ll] n. v 111 1 1111“! W v n I I v .1 R 1 R\ v R .1 5 1 5 l 1 5 . . r 1 .l f I 4.31“ 11 “4“-4141q{41311.11114q1“53d d113““d4‘<414114l1<3‘41 i‘id‘d‘ 1441.4114111‘ ‘11‘41““~111411I1q1‘114‘q1111111“1Id‘d<‘4‘11‘1|1‘ddl A B 4 C contributor to substrate positioning and stabilization at this point in the AA molecule whereas the C-14 contact made by Phe205 is relatively unimportant. Also, there are 34 contacts between active site residues and the methyl half of AA compared with 16 contacts with the carboxyl end; thus, eliminating a contact with the methyl end of AA may be relatively less important than eliminating a contact with the carboxyl half of AA. The positioning of AA seems to be most critically dependent on having enough space at the methyl end of the substrate (e.g. Gly533) to permit the correct positioning of the 13- proS hydrogen with respect to Tyr385. Serine 353. The C0 carbon of Ser353 makes a van der Waals contact with C-3 of AA at the side opposite the C-3 and C-4 contacts made by Va1349 (Figure 18, Table III; [66]). When this residue was mutated to glycine, alanine or threonine, all of the mutants retained substantial cyclooxygenase activity. The products formed by the 83536 and S353A oPGHS-l mutants were the same as those formed by native oPGHS-l; 8353T oPGHS-l formed larger amounts of both 11- and lS-HETE. A V349A/S353T double mutant, designed to remove all contacts between 03 and C-4, had only 3% of the cyclooxygenase activity of the native enzyme compared with the single mutants which had 55% (V349A) and 42% (S353T) of native activity. This suggests that a C-3 contact is particularly important, albeit indirectly, in positioning C-13 with respect to Tyr385. Isoleucine 3 77. Ile377 lies at the distal end of the cyclooxygenase active site channel with one of its delta carbons within 3.7 A of C-20 of AA (Figure 18, Table III; [66])). Substitution of Ile377 with a smaller valine residue yields a mutant enzyme with kinetic properties very similar to native oPGHS-l. Substitution of a valine at this POSition will result in the removal of the delta carbons and increase the distance between 97 this residue and 020 of AA minimally to 5.2 A, a distance too great to support van der Waals interactions. This indicates that the contact made by Ile377 makes little contribution to either substrate binding or positioning of C-13 with respect to Tyr3 85, but does not address the possible necessity for a space-filling residue at this position. 13 77V oPGHS-l forms somewhat more ll-HETE than native oPGHS-l (Table 111) suggesting that this residue does play a small role in orienting C-1 1 with respect to C-9. Serine 530. Ser530 is the site of acetylation of oPGHS-l by aspirin [21, 99], whereby cyclooxygenase activity is completely eliminated. In the AA/Co3+-heme oPGHS-l crystal structure the C01 and CB carbons of Ser530 make van der Waals contacts with C-10 and C-16 (Figure 18; [66]); these contacts are on the side of AA directly opposite that Tyr385. Earlier studies had shown that substitution of Ser530 with alanine has relatively little effect on AA oxygenation, that replacement of SerS30 with threonine causes a 95% drop in catalytic efficiency, and that replacement of Ser530 with glutamine prevents catalysis apparently by blocking substrate access to the Tyr385 radical [21, 156]. Here we have extended these findings showing (a) that a S530V mutant is catalytically inactive (Table V) and (b) that the S530T mutant forms significant amounts of lSR-HETE but not lSS—HETE (Figure 24). Recent studies have indicated that aspirin-acetylated PGHS-2 forms exclusively lSR-HPETE [108, 155] and that formation of this product involves removal of the l3-proS hydrogen [168]. Assuming that lSR-HPETE synthesis by SS30T oPGHS-l occurs in a similar manner, this would imply that 15R-HPETE formation involves suprafacial addition of O2 to 015. Observation of Figure 18 suggests that a residue slightly larger than serine at position 530 might be expected to block antarafacial O2 addition without preventing abstraction of 98 Figure 24. Chiral HPLC analysis of the methyl ester of lS-HETE produced by SS30T oPGHS-l. The 15-HETE fraction produced by SS30T was analyzed as described in Figure 21. A. 15(R/S)-HETE methyl ester standard; B. lS-HETE methyl ester from incubation of SS30T oPGHS-l with arachidonate; C. coinjection of the 15(R/S)-HETE methyl ester standard and the lS-HETE methyl ester from incubation of S530T oPGHS-l with 100 pM arachidonate. 99 I-Fp.Lbllrw>0kn—thPhthr.br.rhrl—fh>>.vllr nlblrlbbphprbr1>>hb>>1rubbbbhbPbibelrlPl irrbhnphb—rlbpbebl—sth.—rflf—Ibu- llllJlLJZLALlllAlLll "I'TYT'TYVYTTV'Y‘f 'I‘VTTjTU'VVVV‘ITTr‘r' S 'IV'I'VVIVVUV'V'IYI flql‘idill‘1434-43‘4‘1‘14‘_44341—14111‘414 A Idl! B I-144‘1-13444-1445q11‘1‘1111l!_4‘1 vm~< 114441 4&44 4141-114 4‘11! 44.1. 1411 4‘1 4 _ C 40 Time (min) 30 100 the l3-proS hydrogen. Moreover, a slightly larger valine residue would be expected to narrow the cyclooxygenase channel so that C-l3 would be mispositioned with respect to Tyr385 such that hydrogen abstraction could not occur. Leucine 531. Leu53l is present in the first shell of the cyclooxygenase site in close proximity to Arg 120 on the side opposite the carboxylate group of AA and outside of van der Waals contact distance from the carbon skeleton of AA (Figure 18). Substitutions of Leu531 with similarly sized, hydrophilic residues (Asp and Asn) decreased the Vmax values for AA but with little effect on the KM's [156]. In contrast, replacement of Leu53l with alanine caused a 25 fold increase in the KM. The relative amounts of oxygenated products formed with Leu53l mutants are similar to those formed with the native enzyme (Table V). Apparently, replacement of Leu53l with hydrophilic residues results in stable, however, improper liganding of the substrate carboxylate to the enzyme, resulting in a decreased Vmax and slight mispositioning of C-13 of arachidonate with respect to Tyr385. Reduction of the residue size at position 531, however, is damaging to high affinity substrate binding of the substrate carboxylate to the enzyme, supporting the notion that proper AA binding to Arg120 is critical for PGHS-l activity and even modest strutural changes in the immediate environment are not tolerated. Summary. Nineteen residues in the cyclooxygenase channel of oPGHS-l are predicted to make contacts with arachidonic acid (Figure 18, Table III; [66]). Three of the residues critical in permitting oPGHS-l to convert arachidonic acid to PGG2 are discussed in this section. Tyr348 is involved in hydrophobic interactions with AA that are necessary to position 013 appropriately with respect to Tyr385. Without the interaction between the Tyr348 phenyl ring and AA, catalysis cannot occur at all. Va1349 and 101 Leu534 provide hydrophobic interactions with AA which contribute, not primarily to initiation of catalysis, but to positioning AA such that when hydrogen abstraction occurs the fatty acid molecule is optimally aligned to yield PGG2 rather than monohydroperoxide products. Va1349 stabilizes the carboxyl half of AA to promote proper positioning of C-9 with respect to OH for efficient endoperoxide formation and Leu534 provides steric hindrance which blocks premature oxygenation at C-15. The remaining residues comprising the cyclooxygenase active site channel provide measurable but lesser contributions to optimizing catalysis. 102 CHAPTER IV THE FUNCTIONS OF CYCLOOXYGENASE ACTIVE SITE RESIDUES 1N SUBSTRATE SPECIFICITY OF PROSTAGLANDIN ENDOPEROXIDE H SYNTHASE-l Introduction Prostaglandin endoperoxide H synthases (PGHSs) catalyze the conversion of arachidonic acid to prostaglandin H2 (PGH2) [2, 3, 137]. This is the committed step in the biosynthesis of prostaglandins and thromboxanes. These compounds are local hormones that act at or near their sites of synthesis to coordinate intercellular responses evoked by circulating hormones and perhaps intracellular events associated with cell replication and differentiation [137]. There are two PGHS isozymes designated PGHS-l and PGHS-2 (or cyclooxygenase-l and -2 (COX-1 and -2)). PGHS-1 and -2 are often referred to as the constitutive and inducible forms of the enzyme, respectively. These isozymes are closely related structurally [66, 80-82] and mechanistically [3, 137, 169] although there are some subtle kinetic differences between the two isoforms with respect to hydroperoxide activator requirements [170, 171] and substrate [115] and inhibitor [20, 172] specificities. Both PGHS isozymes catalyze two different reactions—a cyclooxygenase reaction in which arachidonate is converted to an intermediate prostaglandin endoperoxide PGG2 and a peroxidase reaction in which the 15- hydroperoxyl group of PGG2 is reduced to an alcohol yielding PGH2 [2, 3, 137]. Although the peroxidase activity can function independently of the cyclooxygenase aCtivity [56], activation of the cyclooxygenase requires a functional peroxidase. 103 X-ray crystallographic studies indicate that the cyclooxygenase reaction occurs in a hydrophobic channel that protrudes from the membrane binding domain of the enzyme into the core of the globular domain [66]. The fatty acid substrate is positioned in this site in an extended L-shaped conformation. Cyclooxygenase catalysis begins with abstraction of the 13-proS hydrogen from arachidonate in the rate determining step to generate an arachidonate radical [61, 70]. A tyrosyl radical positioned on Tyr385 abstracts this hydrogen from the substrate fatty acid [3, 70, 95, 137]. The tyrosyl radical is formed as a result of oxidation of the heme group at the peroxidase site on the enzyme [3,137,169] Most studies of the cyclooxygenase have utilized arachidonate as the substrate. Although arachidonate is the best substrate, both enzymes will oxygenate n-3 and n-6 and C13 and C20 fatty acids in vitro with catalytic efficiencies in the range of 005-07 of that of arachidonate [173]. These substrates including 9,12-octadecadienoate (18:2n-6) , 8,11,14-eicosatrienoate (20:3n-6) and 5,8,1 1,l4,l7-eicosapentaenoate (20:5n-3) are also oxygenated via cyclooxygenase activity when added exogenously to intact cells [118, 119] or when mobilized from cellular phosphoglycerides [120-123, 174] (Figure 25). In tissues such as vesicular gland which have low levels of A5 desaturase activity and thus low levels of arachidonate, 20:3n-6 is converted efficiently to the l-series prostanoid products found in semen [121]. Other less common fatty acids can also serve as Cyclooxygenase substrates including adrenic acid (22:4n-6) [131], the Mead acid 52,8Z,112_eicosatrienoic acid [158], columbinic acid (5E,9Z,122_octadecatrienoic acid) [133, 134], and 5,6-oxido-eicosatrienoic acid [132, 135]. 104 Arachidonic acid (20:4, n-6) Dihomo-y-linolenic acid Eicosapentaenoic acid (20:3, n-6) (20:5, 11-3) _ coon _ _ coon _ _ N QCQL—VA V V l-series Prostaglandins 3-series Prostaglandins V 2-series Prostaglandins <:/\/\/\/ COOH m COOH _ N _ __ Linoleic acid (18:2, n-6) a-Linolenic acid (18:3, 113) Monohydroxy acids Figure 25. Structures of and products formed from various n-6 and n-3 fatty acid substrates of PGHSs 105 Substrates other than arachidonate typically have KM values comparable to those of arachidonate and compete with arachidonate for the cyclooxygenase active site thereby inhibiting arachidonate oxygenation. This was first documented by Lands and coworkers in vitro [114] but this form of inhibition appears to occur in vivo as well [126, 127, 175]. Here we report studies of PGHS-l that were designed to identify active site residues which are determinants of cyclooxygenase fatty acid substrate specificity. 106 Materials and Methods Materials. Fatty acids were purchased from Cayman Chemical Co., Ann Arbor, MI. [l-MC] arachidonic acid (40-60 mCi/mmol), [l-l4C]homo-y-linolenic (8,11,14- eicosatrienoic acid) (50-60 mCi/mmol), [l-I4C]linoleic acid (50-60 mCi/mmol) and [1- l4C]eicosapentaenoic acid (50-60 mCi/mmol) were from New England Nuclear. Flurbiprofen and soybean lipoxygenase-l were from Sigma Chemical Company. Restriction enzymes and Dulbecco’s modified Eagle’s medium (DMEM) were purchased from GIBCO. Calf serum and fetal bovine serum were from HyClone. Primary antibodies used for Western blotting were raised in rabbits against purified oPGHS-l and purified as an IgG fraction [149] and goat anti-rabbit IgG horseradish peroxidase conjugate was purchased from BioRad. Oligonucleotides used as primers for mutagenesis were prepared by the Michigan State University Macromolecular Structure and Sequencing Facility. All other reagents were from common commercial sources. Preparation of oPGHS-I mutants. Mutants were prepared by site-directed mutagenesis of oPGHS-l in the pSVT7 vector employing the Strategene QuikChange mutagenesis kit and the protocol of the manufacturer [149]. Oligonucleotides used in the preparation of various mutants are found in Table II and as follows: I434V 5'-'375GCCAGCCTGCAGGCCGGGTTGGTGGGGGTAGG-3' H513R 5'-""7CTTGAGAAGTGTCGACCGAACTCCATCrrrGG-3' L352V 5'-‘ '3°ATGTGCAGCAGGTGAGCGGCTAC 3' Plasmids used for transfections were purified by CsCl gradient ultracentrifugation and mutations were reconfirmed by double-stranded sequencing of the pSVT7 constructs using Sequenase (ver. 2.0, US. Biochemical Corp.) and the protocol described by the manufacturer as described previously. 107 T ransfection of C OS-I cells with oPGHS-I constructs. COS-1 cells (ATTC CRL- 1650) were grown in DMEM containing 8% calf serum and 2% fetal bovine serum, transfected with pSVT7 plasmid constructs containing cDNAs coding for native oPGHS- l or mutant oPGHS-l using the DEAE dextran/chloroquine transfection method. Forty hours following transfection, cells were harvested and microsomal membrane fractions prepared as described previously [149]. Protein concentrations were determined using the method of Bradford [165] with bovine serum albumin as the standard. Microsomal preparations were used for Western blotting and for cyclooxygenase and peroxidase assays. C yclooxygenase and peroxidase assays. Cyclooxygenase assays were performed at 37°C by monitoring the initial rate of O2 uptake using an oxygen electrode. Reactions were initiated by adding approximately 250 pg of microsomal protein in a volume of 20-50 pl to the assay chamber. The oxygenation of all fatty acids by oPGHS- 1 enzymes was completely inhibited by the addition of 0.2 mM flurbiprofen to the reaction mixture. The addition of exogenous peroxide (5 pM lSS-HPETE) to the reaction mixture was necessary to sustain cyclooxygenase activity with eicosapentaenoic acid (EPA) [124]. Addition of this peroxide to the reaction mixture had no effect on the conversion rates of other substrates. All data were normalized to the relative levels of PGHS protein expression as determined by western blotting and densitometry. Fatty acid substrate concentrations of 100 pM were used to estimate Vmax values. KM values for the different fatty acids were measured using concentrations of fatty acid substrates between 0.5 and 500 pM. Peroxidase activities were measured spectrophotometrically with N,N,N',N'-tetramethylphenylenediamine (TMPD) as the reducing cosubstrate [166] as 108 reported previously [104]. Reactions were initiated by adding 100 ml of 0.3 mM H202 and the absorbance at 610 nm was monitored with time. Western blot analysis. Microsomal samples (ca. 5 pg of protein) were resolved by one-dimensional SDS-PAGE and transferred electrophoretically to nitrocellulose membranes using a Hoeffer Scientific Semi-Dry Transfer apparatus. Membranes were blocked for 12 hr in 3% non-fat, dry milk, 0.1% Tween20 and Tris-buffered saline, followed by a two-hour incubation with a peptide-directed antibody against oPGHS-l [84] in 1% dry milk, 0.1% Tween-20 and Tris-buffered saline at room temperature. Membranes were washed and incubated for one hr with a 1:2000 dilution of goat anti- rabbit IgG-horseradish peroxidase, after which they were incubated with Amersham ECL reagents and exposed to film for chemiluminescence. Characterization of fatty acid oxygenation products. A general protocol for product analysis is as follows. Forty hr following transfection, COS-1 cells were collected, sonicated and resuspended in 0.1 M Tris-HCl, pH 7.5. Aliquots of the cell suspension (100-250 pg of protein) were incubated for 1-10 min at 37° C in 0.1 M Tris- HCl, pH 7.5, containing 1 mM phenol and 6.8 pg bovine hemoglobin in a total volume of 200 pl. Reactions were initiated by adding l-MC-labelled fatty acid (35 pM final concentration) and were performed with or without 200 pM flurbiprofen and stopped by adding 1.4 ml of CHCl3zMeOH (1:1; v/v). Insoluble cell debris was removed by centrifugation and 0.6 m1 of CHC13 and 0.32 ml of 0.88% formic acid were added to the resulting supernatant. The organic phase was collected, dried under N2, redissolved in 50 [.11 of CHCl3 and spotted on a Silica Gel 60 thin layer chromatography plate; the lipid products were chromatographed for l h in benzene:dioxane:formic acid:acetic acid 109 (82:14:l:1; v/v/v/v). Products were visualized by autoradiography and quantified by liquid scintillation counting. Negative control values from samples incubated with 200 pM flurbiprofen were subtracted from the experimental values observed for each sample in the absence of flurbiprofen. Preparation of 1 5S-HPE T E. 15S-HPETE was prepared according to Graff [176] with modifications. Arachidonic acid (50 pmol) was suspended in 100 ml of 0.2 M borate buffer, pH 9.0. The reaction was initiated by adding soybean lipoxygenase (50 mg) and the incubation continued for five min at 30°C. The reaction was quenched by adding enough 0.24 M HCl to bring the pH to 3.0 (ca. 20 ml) and extracted three times with an equal volume of ice-cold petroleum ether:ethyl acetate (1:1). Extracts were pooled, dried under a nitrogen stream and resuspended in 1 ml of ethanol. The concentration of lS-HPETE was determined spectrophotometrically by measuring cis- trans conjugated double bond formation (6234 = 3 x 104 mM'l cm") and by measuring peroxide levels using a ferrous sulfate-xylenol orange reagent as described by Gupta [177]. 110 Results and Discussion Overview. As discussed in the Introduction, both PGHS-l and PGHS-2 can oxygenate a variety of C13 and C20 polyunsaturated fatty acids substrates. Arachidonic acid is the preferred substrate for the cyclooxygenase activity of both isozymes (Table VII; [173, 178]). Crystallographic [66] and mutagenic analyses of the interaction of arachidonic acid within the cyclooxygenase active site of ovine (o) PGHS-1 has led to the assignment of active site residues to five fimctional categories [179]: (a) residues directly involved in abstraction of the 13 proS hydrogen (Tyr385); (b) residues involved in positioning C-13 for hydrogen abstraction (Tyr348 and Gly533); (b) residues essential for high affinity binding of arachidonate (Arg120) and ((1) residues critical for positioning arachidonate such that following hydrogen abstraction the arachidonyl radical is converted to PGG2 (Va1349, Trp387 and L534). In the following paragraphs we describe the effect of active site substitutions on the oxygenation of fatty acids other than arachidonate. Substitutions of most active site residues caused approximately parallel changes in the oxygenation of all fatty acids. However, there were two significant exceptions--Va1349 and Leu534. Va1349 interacts with the C-3 and C-4 of arachidonate. Substitution of Va1349 with alanine caused 175-1000 fold decreases in the efficiencies (Vmax/KM) of oxygenation of 18:2n-6, 20:3n-6, and 20:5n-3 with only a 35% decrease in the efficiency of arachidonate oxygenation; kinetic analyses indicate that the large decrease in Vmax/KM values for 20:3n-6 with the Val349Ala mutant results from mispositioning of C-13 for hydrogen abstraction. The crystal structure of a 20:3n-6/Co3+-heme oPGHS-l complex lll has been determined and compared to that of the 20:4n-6/Co3+-heme oPGHS-l complexz. The regions involving O3 to C-7 are differently positioned for arachidonate versus eicosatrienoate and suggest that Va1349, which interacts with both substrates in this region of the carbon chain, is particularly critical in positioning the carboxyl half of eicosatrienoate in a way that provides for proper alignment of C-13 for hydrogen abstraction by the Tyr385 radical. A L534V oPGHS-l mutant oxygenated C20 but not C13 substrates; Leu534 interacts with the (1) end of C20 substrates (i.e. C-15, C-16 and C- 18 of arachidonic acid [66] and dihomo-y-linolenic acid) and thus this residue appears to be particularly important for the positioning and oxygenation of C13 substrates. All mutant enzymes were expressed at protein levels similar to that of the native enzyme and retained >50% the peroxidase activity of native oPGHS-l. Valine 349. Va1349 plays an important role in positioning arachidonate for PGG2 formation [104]; substitutions with smaller residues lead to substantial increases in the formation of ll-HPETE versus PGG2 whereas substitution of Va1349 with a larger leucine residues leads in increases in lS-HPETE synthesis. For example, V349A oPGHS-l forms 55% ll-HPETE from arachidonate versus 2.5% ll-HPETE by native oPGHS-l. In contrast substitutions of Va1349 with either alanine or leucine caused less than two-fold increases in hydroperoxy acid formation from [1-'4C]dihomo-y-linolenic acid and produced oxygenated products in the same ratios as native enzyme when tested with [1-'4C]eicosapentaenoic acid or [1-‘4C]linoleic acid (data not shown). However, Va1349 appears to be critical for the oxygenation of substrates other than arachidonate. As shown in Table VII the oxygenation of linoleic (18:2n-6), eicosadienoic (20:2n-6), 2 Personal communication, M.G. Malkowski and RM. Garavito 112 dihomo-y-linolenic (20:3n-6), eicosapentaenoic (20:5n-3) a-linolenic (18:3n-3) acids by native oPGHS-l occurs with 0.6-50% of the efficiencies (Vmax/KM values) observed with arachidonate. Substitutions of Va1349 led to dramatic decreases in the Vmax/KM values for substrates other than arachidonate. For example, while V349L oPGHS-l was six fold less efficient than native oPGHS-l in the oxygenation of arachidonate, this mutant showed a 350 fold decline in the Vmax/KM value with 18:2n-6 as the substrate. V349A oPGHS-l oxygenates arachidonate with a catalytic efficiency (Vmax/KM) that is 65% that of native oPGHS-l. However, the Vmax/KM value for the V349A oPGHS-l mutant was nearly 200 fold less when tested with 18:2n-6, 1000 fold less when tested with 20:5n-3, and >800 fold less when tested with 20:3n-6 (Table VII). The only difference between 20:3n-6 and 20:4n-6 is the A5 double bond. Va1349 has been shown previously to make van der Waals contacts with C-3 and C-4 of the arachidonate molecule [66]. Removal of these contacts by substitution of an alanine at position 349 does not significantly decrease the enzyme’s catalytic efficiency in the oxygenation of arachidonate (20:4n-6). The crystal structure of oPGHS-l bound with dihomo—y-linolenic acid (20:3n-6) shows a structure in which the carboxyl half of the substrate is stretched due to the loss of the C- 5--C-6 double bond. This is the area which is in direct contact with Va1349 at C-5 (Table VI). Va1349 and Ile523 both make contacts at C-5, but from opposites sides of the substrate. Removal of the C-5 contact made by Ile523 by replacement with an alanine has relatively little effect on the oxygenation of 20:3 (Vmax is ~ 60% that of native enzyme) while removal of the C-5 contact made by Va1349 decreases 20:3 oxygenation dramatically (Vmax is only 4% that of native enzyme and the Vmax/KM is reduced by >800 fold) (Table V11). With the removal of the CGl Va1349 contact, the lower portion of 113 Table VI. Contacts between Dihomo-y-Iinolenic acid (DHLA) and Cyclooxygenase Active Site Residues* Van der Waals and hydrogen bond interactions were calculated using CHAIN [167]. Van der Waals contacts within 4A are listed. See list of abbreviations for carbon designations. Residue Atom Residue Atom Distance (A) Phe 205 CE2 DHLA C15 3.6 CZ C18 3.9 Phe 209 CEl DHLA C19 3.9 CE2 C19 3.9 CZ C17 3.9 CZ C18 3.9 CZ C19 3.6 Tyr 348 CD2 DHLA C14 3.9 CE2 C12 3.6 CE2 C13 4.0 CE2 C14 3_4 Val 349 CGI DHLA C5 3.4 Leu 352 CD2 DHLA C6 3.8 CD2 C7 3.2 CD2 C11 3.3 CD2 C12 3.9 Phe 381 CE2 DHLA C17 3.8 CZ C16 3.9 CZ C17 3.5 116 523 CA DHLA C8 3.9 CG2 C3 3.3 CGZ C4 3.3 CGZ C5 3.7 €02 C6 3.2 CG2 C7 3.9 (:02 C8 3.9 Ser 530 CA DHLA C16 3.3 CA C17 3.9 CB C10 3.6 CB C13 3.6 CB C15 4.0 (33 C16 3.1 Leu 534 CD1 DHLA C18 3.6 CD2 C15 3.8 CDZ C16 3.7 CD2 C18 3.9 *Personal communication from M.G. Malkowski and RM. Garavito. 114 ...Efim .135 com .8on ._..0 80¢ oosmomoooooooo Roofing... oz oz oz oz oz 8 oz oz oz oz oz 8 oz oz 3 43:92 29> oz oz oz oz oz 3. oz oz m 2 2 S we a 0m 3-2.6.2 <89 2 a. E M: R S Z: 2 9. ma 2 m: M: 9m 2: 13:95 oz oz 3 8. c? S 5. N9. 2. 2 3 mm o e 8 :49 oz oz o oz oz 4 oz oz 2 oz oz 3. oz oz cm 2.33%: 322239 oz oz 0.0 oz oz 4 oz oz N oz oz 3. oz oz 8 39342.; oz oz N «o. S m Sod 8m 2 8.0 a 1 a: 3 mm 5 mm 2: <3; no 2 ON e 3 3. E N 2 2 3 ; om N 2: o>fiNZo 2x KNE> 2M KNE> «(g me> uLVm XNE> Suv— xflE> 1 m-.. 2: a-.. 3: n... m5" 5... m3 5... Ex Eo< 9:33:58 23.. 22°51. Eu< o_o:05=oaamee_m Eat. o_:o_o:__->-oEe_=D Eu< 3.52.355. fieEEoBoc Ho: .02 .652 ca 8 958 33 3-5 New 5; TmEOmo 2:3: 51* ax; 85> 98593 Eon boom 21 coo momma gazebo $585808 BE E323.— woomooooew Ho: 20>» 82? 2M :05? 5m 8535 E82: com Cotonou moo—g 55> gum—om Cotonou 82? EN mo o\oo_ ESE» mooumgoc chateau £3, mooumfiotooow 8838 58 «o EEEEE a 80¢ women: 2.: E8892 moo—g 2v— com 85> .28 2:02:88 53> N- com TmEOm 032E mo 53:8 ommoomzxo o8 coowfimm a $02 mo 02? < 605:8 $2605 Bomoowbooéfi com -oooEmo owfloooooo 9.: 8m @8858 can com 5388 Box beam How @23ng com 83.3 .oxoo on“ E Contact 8 355020 combs on at? @8335 mm? .0558 ommoowmxo N1mlUn— can Twin: «55:2 HE: 253 Z .3 mEo< baa...— m=c€n> he oouaoowhxo ...... move—onen— 385 .=> 93.; 115 20:3 (03 through 05) is supported solely by van der Waals interactions on the opposite side of the substrate with Ala527, Gly526 and Ile523. Thus, Va1349 may provide the one critical interaction with the carboxyl end of 20:3 responsible for proper positioning of the l3-proS hydrogen for abstraction by Tyr3 85. Although there is not currently an oPGHS- 1 crystal structure bound with 18:2n-6 or 20:5n-3 available, it is likely that the carboxyl end of 18:2 is bound similarly to that of 20:3, causing it to be reliant on similar, if not the same, carboxyl end interactions with Va1349, particularly at C-5. As shown in Table VII, V349A, oxygenates arachidonate (20:4) with 65% the efficiency of the native enzyme but oxygenates the similar substrate, 20:3, with only 0.1% native efficiency although it is able to bind 20:3 relatively easily (K. = 2pM). This suggests that dihomo-y-linolenic acid can bind to the V349A oPGHS-l mutant with about the same affinity as binding to native oPGHS-l, but that V349A oPGHS-l binds dihomo- y-linolenate in a nonproductive conformation such that abstraction of the l3proS hydrogen cannot occur. Thus, we conclude that Va1349 is a key determinant of the cyclooxygenase substrate specificity of oPGHS-l. Although Val349 is an important determinant of the substrate specificity for PGHS-1, it is not an important determinant for PGHS-2. The corresponding mutation in PGHS-2 results not only in efficient oxygenation of arachidonate but also in efficient oxygenation of 20:3 (61% the efficiency of native enzyme) (Table VII). This discrepancy between substrate specificities of identical active site mutants (V349A) in PGHS-l and PGHS-2 has led to the construction of two oPGHS-l enzymes containing multiple mutations. PGHS-2 has been shown to accommodate a wider range of substrates than PGHS-l [l 15], possibly because of its slightly larger cyclooxygenase site 116 [82]. The cyclooxygenase active site differences between PGHS-l and -2 are the result of three residue substitutions in PGHS-2. Ile523 is a valine in PGHS-2 and is the only residue in the hydrophobic lining of the cyclooxygenase active site which is different between PGHS-1 and -2. Two other residues immediately outside the cyclooxygenase active site, in the outer shell, are His513, an arginine in PGHS-2 and Ile434, a valine in PGHS-2. V349A/1523V and V349A/1523V/H513R/l434V oPGHS-l were constructed in order to make the area surrounding the oPGHS-l cyclooxygenase active site identical to that of PGHS-2 and possibly restore the ability of V349A to oxygenate 20:3 efficiently. V349A has been shown to produce an abundance of llR-HPETE when reacted with arachidonic acid. It should be noted that both the double and quadruple V349A mutants exhibited the same pattern of oxygenated product distribution as the single V349A mutant (data not shown). When tested with lOOuM substrate, the double mutant showed a slight increase in oxygenation of all substrates tested. The quadruple mutant showed a further slight increase in oxygenation of 20:3, but had little effect on other substrates and was unable to oxygenate 18:3n-3 at all. In general, the effects of these changes were not largely significant, with a maximal increase in Vmam of 1.5 fold. By making changes of this nature, we have not succeeded in mimicking the activity of V349A PGHS-2 with 20:3. This may be due in part to the difference between the necessity of the ionic bond between Arg120 and the carboxyl group of fatty acid substrates for oPGHS-l activity and the lack of this necessity in PGHS-2 [98]. This difference could potentially result in significant deviations between the positioning of substrates in each active site (oPGHS-l and PGHS-2) and the substrate-stabilizing influences, particularly at the carboxyl end of the fatty acid molecules. 117 Residues involved in stabilization of n-6 vs. n-3 fatty acid substrates (Phe381, Phe205 and Leu531). Phenylalanine 381. Phe381 makes four contacts with three carbons (C-16, C-17 and 020) at the methyl end of arachidonic acid [66] and three contacts with two carbons (C- l 6 and C-1 7) at the methyl end of dihomo-y-linolenic acid, both n-6 substrates (Tables III and VI). F381A was not able to oxygenate any of the substrates well, but had the most detrimental effect on the n-6 fatty acids, 20:4, 20:3 and 18:2, using these substrates only 2-4% as well as the native enzyme, while using the n-3 fatty acid, 20:5 with 17% the activity of native oPGHS-l (Table VIII). The contacts made by the phenylalanine are probably somewhat maintained by the leucine mutation, but nearly nonexistent when changed to the smaller alanine. These contacts, apparently, are very important in keeping the 13 proS hydrogen aligned for abstraction by Tyr385, thus having a significant effect on the oxygenation of all substrates. The n-6 substrates will be much more acutely affected by mutations made near the methyl end of fatty acids because of the lack of rigidity at the methyl end compared with n-3 substrates, thus making contacts at that end more important for substrate stabilization. In particular, n-6 fatty acids with fewer double bonds such as 20:3 and 18:2 will be more destabilized by the removal of three methyl end contacts at 016 and C-17, which, in the case of 20:3 and 18:2, result in a 98% decrease in the enzyme’s ability to oxygenate either of these substrates. These fatty acids are not only more flexible than n-3 substrates, but also more flexible than arachidonate. Phenylalanine 205. Phe205 contacts C-l4, C-15 and C-18 of arachidonate [66] and C-15 and C-18 of dihomo-y-linolenic acid (Tables III and VI). F205A showed a 118 Table VIII Oxygenase Activity for oPGHS-l Cyclooxygenase Active Site Mutants with Various Fatty Acid Substrates Oxygenase activity was measured with an oxygen electrode as described in the text. Values are calculated for fatty acid turnover and a value of 100% is assigned for oxygenase activity of native oPGHS—l. Vmax and KM values represent the means from a minimum of four separate determinations with standard deviations within 10% of all values reported. Relative Vmam values reported for mutant enzymes for which KM values were not determined represent rate measurements performed using 100 uM fatty acid substrate. ND, not determined. ENZYME 20:4 (n-6) 20:2 (n-6) 20:3 (n-6) 20:5 (n-3) 18:2 (n-2) 18:3 (n-3) SHAM 0 0 0 0 0 O FZOSA 28 ND 17 52 8 29 F2051. 65 ND 32 118 49 115 F209A 15 ND 6 2 4 3 F2091, 43 ND 39 33 21 27 Y348L 0 ND 0 0 0 0 V349A/S353T 3 ND 0 0 0 0 L352V 34 ND 30 57 21 ND 83530 61 ND 60 34 13 28 8353A 56 ND 43 35 33 96 S353T 42 ND 26 73 32 19 I377V '72 ND 52 6O 56 37 F38|A 4 ND 2 l7 2 6 F381 L 21 ND 14 21 ll 5 1523A 64 ND 58 50 7 33 1523V 70 ND ND ND 28 ll SS30A [21. 156] 70 (Km 2) 109 97 56 ND (Km 8.6) (Km = 48) (Km = 16) 62 ssxor {21. 156] 17 3.3 2 l6 51 (Km :13) ND (Km= 100) (Km=115) (Km=305) SS30V 0 0 O O O 0 L53“ 4-6 20 7 46 11 36 (Km = 54) L53IV 7.7 18 I6 63 IS 18 (Km 5 1.6) (Km = 8.4) L534A 59 ND 43 50 41 I4 L534V 98 ND 34 10 O 0 119 similar pattern to F381A, using 20:5n-3, 52% as well as native, while retaining only 8- 28% the activity with the n-6, 20-carbon fatty acids. This mutant also retained 30% native activity with the n-3 fatty acid, 18:3, while retaining only 8% activity with 18:2,n— 6. F205 L, while quite active with all substrates (32-118% native activity) highly favored the n-3's, using both 20:5 and 18:3 with greater than 100% native activity (Table VIII). Similar to Phe381, the contacts made by Phe205 are also near 013 of both arachidonate (C-l4, C-15 and C-18) and dihomo-y-linolenante (C-15 and C-l8), thus the reasoning is similar to that for Phe381. F205L actually oxygenates the n-3 fatty acids, 20:5 and 18:3 more efficiently than native. PGHS-2 has been shown to oxygenate both 20:5 and 18:3 significantly more efficiently than PGHS-1 [115]. In an inhibitor-bound crystal structure of murine PGHS-2, the cyclooxygenase pocket has been shown to be slightly larger than that of oPGHS-l [82]. This extra space created by changing Phe205 to the smaller leucine may provide just enough extra space to provide the n-3 double bond with enough flexibility so that the (08 hydrogen in the case of 20:5 and the (05 hydrogen in the case of 18:3 can be better aligned for abstraction. Leucine 531. Leu53l is present in the first shell of the cyclooxygenase site in close proximity to both the carboxyl and methyl ends of arachidonate and dihomo-y- linolenate, but outside of van der Waals contact distance from the carbon skeleton of either substrate (Table IV). Again, when Leu531 is mutated to the smaller residues (L531V or L531A), oxygenation of n-3 substrates remains relatively unaffected, while oxygenation of n-6 substrates decreases considerably (Table VIII). For instance, L531A oxygenates the n-3 substrates 20:5 and 18:3 with 46 and 36% the efficiency of the native enzyme while oxygenating n-6 substrates with only 5-11%. Leu53l makes four main 120 chain hydrogen bonding interactions with active site residues, two of which are involved in interactions with the methyl ends of both arachidonate and dihomo-y-linolenate, Gly533 and Leu534. Changes in Leu531 could adversely affect the interactions between these adjoining active site residues and the methyl ends of substrates, resulting in decreased stabilization of the more flexible n-6 versus n-3 substrates. Residues selectively involved in stabilization of l 8:2n-6 (Ser353, Ile523) Serine 353. The CB of Ser353 makes one contact with arachidonic acid [66] at C- 3 but is greater than 4A from C-3 of dihomo-y-linolenate. However, shortening of this residue by replacement with a glycine results in an enzyme which no longer oxygenates 18:2n-6 efficiently at 13% the rate of the native, compared with 61% the rate of the native enzyme with arachidonate (Table VIII). Because this mutant retains high cyclooxygenase activity and 87% the native peroxidase activity, we can assume that we have not perturbed the structural integrity of the enzyme overall. There is not currently a structure of oPGHS-l bound with 18:2, but these results suggest the possibility of a Ser353 contact with the 18:2 substrate, which appears to play a role in stabilization of the carboxyl end of this 18-carbon substrate for pr0per positioning and efficient hydrogen abstraction. Isoleucine 523. One of the gamma carbons of Ile523 makes van der Waals contacts with C2, C5 and C6 of arachidonate [66], but makes six contacts with C-3 through C-8 of dihomo-y-linolenate as well as an additional contact between its Ca and C-8 of 20:3. This isoleucine is a valine in PGHS-2. Not surprisingly, substitution of Ile523 with a valine has very little effect on the oxygenation of all fatty acid substrates tested. Substitution with an alanine eliminates the gamma carbon contacts made between 121 Ile523 and multiple carboxyl end carbons of 20:4 and 20:3, but still has relatively little effect on the oxygenation of 20:4, 20:3 or the n-3 fatty acid 20:5 (Vmax values are 64%, 58% and 50% that of native enzyme, respectively). As mentioned earlier, the carboxyl end of 20:3 is more flexible than that of 20:4 and most likely relies more heavily on carboxyl end interactions, however, there are numerous other carboxyl end contacts with the same side of the 20:3 substrate as Ile523 including contacts at C-3 and C-4 made by Ala527, a contact at Gly526 at C-8 and the one Ca contact at C-8 retained by I523A which could compensate for the loss of Ile523 interactions. In contrast, removal of one C—5 contact made by Va1349 on the side of the substrate opposite Ile523 results in almost a complete loss of 20:3 oxygenation, most likely because of the lack of compensatory interactions on that side. The 1523A substitution did, however, exhibit a significantly decreased ability to oxygenate 18:2n-6 (Table VIII). Again, 18:2 is likely to be more heavily reliant on stabilizing van der Waals interactions at its carboxyl end due to the absence of the C-5 and C-8 double bonds, thus interactions made by Ile523 are more significant for the stabilization of 18:2 versus 20:3, 20:4 or 20:5. Residues involved in stabilization of n-3 vs. n-6 fatty acid substrates (Phe209, Ser530) Phenylalanine 209. Phe209 resides near the methyl end of arachidonate and dihomo-y-linolenate and makes numerous contacts with the C- l 7 through C-l9 regions of both substrates [66]. Mutation of P116209 to an alanine resulted in significantly decreased oxygenation of all substrates tested, while the leucine mutant was much more active and showed little selectivity among substrates. This is most likely because some of the original phenylalanine contacts were maintained by the leucine substitution. The n-3 122 substrates were somewhat more negatively affected by the F209A change than their n-6 counterparts. Phe209 is the only residue near the methyl end of substrates for which this is true, suggesting that the methyl end contacts made by Phe209 may be the primary contributors to stabilization of n-3 fatty acids at the methyl ends of these substrates. Serine 530. Ser530 is the site of acetylation of oPGHS-l by aspirin [21, 99]. Van der Waals contacts between Ser530 and arachidonic acid and dihomo-y-linolenic acid occur at C-10 and C-16 for arachidonate and at C-10, C-13, C-15, C-16 and C-17 for dihomo-y-linolenate (Table VI) [66]. In both cases, these contacts are on the side of the substrate directly opposite Tyr385. Earlier studies had shown that substitution of SerS30 with alanine has relatively little effect on arachidonate oxygenation, that replacement of Ser530 with threonine causes a 95% drop in catalytic efficiency, and that replacement of Ser530 with glutamine or valine prevents catalysis apparently by blocking substrate access to the Tyr385 radical [21, 156]. Here we have extended these findings showing that S530A has very little effect on oxygenation of substrates other than arachidonate while S530T exhibits poor catalytic efficiency when tested with all substrates (Table VIII), confirming that having a small residue at position 530 is important for Optimal positioning of all fatty acids for hydrogen abstraction by Tyr385. Also, SS30T seems to have a more negative effect on the n-3 versus the n-6 substrates, with a Vmax which is ~l 7% that of native oPGHS-l for oxygenation of 20:4n-6 and 18:2n-6 and only a Vmax of 2 and 1% for oxygenation of 20:5n-3 and 18:3n-3, respectively. Positioning of Ser530 near the mid-chain carbons of fatty acids combined with the methyl end contacts made by Phe209 appear to play a significant role in situating the n-3 substrates optimally for hydrogen abstraction by Tyr385. 123 Residues involved in stabilization of 18-carbon vs. 20-carbon fatty acid substrates (Leu534) Leucine 534. Leu534 makes van der Waals contacts between one or more of its delta carbons and the C-15 through C-18 regions of both arachidonate and dihomo-y- linolenate (Table VI) [66]. Although active with all 20-carbon fatty acids tested, a L534V mutant is unable to oxygenate either of the 18-carbon fatty acids tested, 18:2n-6 or 18:3n-3. Interestingly, the L534A mutant is able to oxygenate both 18:2 and 18:3 with 41 and 14% the rate of native oPGHS-l, respectively. Computer modeling of these mutants with 18:2 bound in the cyclooxygenase active site suggests that this 18-carbon substrate may be bound differently at the methyl end compared with both 20:3 and 20:4, possibly in closer proximity to Leu534 rather than passing over Gly533 as is observed for 20:3 and 20:4. Although interpretation of these data is based solely on a computer model, one possibility may be that the gamma carbons of the L534V are configured differently than those of the leucine and could possibly engage in an aberrant contact which is detrimental to a catalytically favorable orientation of the 18-carbon substrates. Residues equally involved in stabilization of all substrates tested (T yr348) Tyrosine 348. Phenyl ring carbons of Tyr348 are within van der Waals distance of C-12, C-13 and 014 of both arachidonic and dihomo-y-linolenic acids (Table VI) [66]. These interactions have been shown to be important in positioning C-l3 of arachidonate as evidenced by a Y348L oPGHS-l mutant which lacks cyclooxygenase activity while retaining significant peroxidase activity as discussed in Chapter III. Here we show that this mutant is also unable to oxygenate any fatty acid substrates tested (Table VIII), indicating that substitution of leucine at position 348 permits the 124 mispositioning of C-13 of most n-6 and n-3 substrates with respect to Tyr385 such that hydrogen abstraction cannot occur. Residues which oxygenate substrates with a pattern similar to native oPGHS-l (Leu352, 1163 7 7) All of the residues in and around the cyclooxygenase active site, when mutated, lessen the ability of oPGHS-l to efficiently oxygenate arachidonate and other substrates to some extent. Leu352 and Ile377, when mutated, result in enzymes which negatively affect the oxygenation of all substrates tested to roughly the same degree. One delta carbon of Leu352 makes multiple contacts with the upper carboxyl to mid region of both arachidonate (C-7, C-11 and C-12) and dihomo-y-linolenate (C-6, C-7, C-11 and C-12). Substitution of this residue with a valine should, in all likelihood, eliminate these contacts. L352V, however, retains 21-57% the Vmax of the native enzyme when tested with 20:4, 20:3, 20:5 and 18:2. The mid regions of both arachidonate and dihomo-y- linolenate are supported by numerous van der Waals interactions, both on the side of Leu3 52 and the side opposite. Some of the residues contributing in these interactions are Ile523, Phe518, Gly526, Ala527, Ser530, Trp387 and Tyr348. The interactions made by Leu352 appear to be complementary to the many others in place in this region of the substrate One delta carbon of Ile377 contacts C-20 of arachidonate but is out of van der Waals range of dihomo-y-linolenate [66]. Again, mutation of this residue to a valine should, in all likelihood, eliminate this one contact. I377V oxygenates 20:4, 20:3, 20:5, 18:2 and 18:3 with 40-70% the rate of the native enzyme and does not appear to be selectively involved in the stabilization of certain substrates over others (Table VIII). 125 The best known substrate for ovine Prostaglandin Endoperoxide H Synthase is arachidonic acid [115]. However, both dihomo-y-linolenic acid (20:3n-6) and linoleic acid (18:2n-6) have also been shown to be substrates for PGHSs in vivo [61, 1 17,123,119]. Here we have determined that two of the most significant determinants of substrate specificity for oPGHS-l are Va1349 in the case of 20:3 and Leu534 in the case of 18:2. Va1349 plays a crucial role in stabilization of the carboxyl ends of substrates other than arachidonate, particularly in the case of substrates lacking one or more double bonds toward the carboxyl end of the molecule, while Leu534 appears to play a more significant role in the methyl-end stabilization of l8-carbon fatty acids such as 18:2. Ongoing crystallographic and mutational analyses with PGHS-1 and PGHS-2 will help to further illuminate the key determinants in substrate usage by both PGHSs. 126 APPENDIX 127 APPENDIX A Derivations for Kinetic Equations Scheme 1 Derivations (no inhibitor): Term definitions v(l) = kS[ESI] K0: [Ellsol/ 1530] K2 = [5501/ 1532] V(2) = kleszl KI = [5301/ [E81] [Stot] = [80} [Etot]=[E] +1530] +IESII+15321 [ESQ] = 1511301 [532] z [5301/ K2 [5321/1531] = K2/ K1 [E311 = [5501/ K1 Formation of Product 1: [581] = 1E][SO]/K0Kl : [Ellsol/ KOKI [Etot] [E] + [1380] + [E81] + [1352] [E] + [Ellsol/Ko + [Ellsol/KOKI + [Ellsol/KoKz : 1501/ KOKI * (KoKi/ KoKi) 1 + 1501/ K0 + [301/ K0K1+ [S0]/ KoKz = [So] KOKI + [$01K] + [So] + [S0]K. /K2 Therefore: [ESI] : [Etot][SO] Since V“) = k51ESl] and K1(Ko+ [501 + [Sol/K2)+ [30] V“) = vmax(l)[S]/KM(1)+ [S]. V(1) = k51EtotIIS01 ; Vmaxa): k51Etot] and K1(K0 + [so] + 1501/ K2) + 150] KM“) = K1(Ko + 1501 + 1501/ K2) Formation of Product 2: E37 = E S / K 1 -1 1 11 0] K0 2 * (KoKz/ KoKz) [Etot] [E] + [Ellsol/ K0 + [E][So]/ KoK1+ [Ellsol/ KoKz 128 = [So] K2(K0 + [80] + [So]/K.) + [S0] Therefore: v(2) = R6 [Etot] [SO] ; Vmax(2) = k6 [Etotl and K2(Ko + [so] + [Sol/K1) + [Sol Km(2) = K2(Ko + [Sol + [Sol/Kl) Scheme 1 Derivations (with inhibitor): Term Definitions Same as above with the following additions and changes: K. = [Ellll/ [El] [Em] = [E] + [580] + [E81] + [1552] + [El] [131} = [Ellll/ K. Formation of Product 1 in the presence of inhibitor: [E81] = [Ellsol/KOK. [Em] [E] +[E][So]/Ko+IE][So]/K0Ki+ [Ellsol/K0K2+ [Ellll/Ki : 1501/ KOKI * (KOKI/ KOKI) 1 + [Su]/ [(0 + [Soy KoKl '1' [S()]/ KoKz + [I]/K| : [so] KOK| + [S()]K| + [So] '1' [So]K. /K2+ [1] KoKl/ K1 ThCI’CfOI’CI [ES.] = [5.0111301 K0K1+ [SO]K.+ [So] + [S(.]K. /K2+ [I] KOK./ K. Since v(1)= k5[ES.] and V(1) = Vmax(1)[s] Km(1)(1 + [ll/KI) + 15] 311611 V“) = k5 [510.1150] * (l/ K. + K. /K2+ l) K0K1(1+[Il/KI)+1301(K1+ KI/K2+1) (1/ K1 + Kl/K2'l'1) 129 Rearranges to: v(1)= k5 [Emllsoll (K. + K. M; + 1) (KOKI/ (K1+ K1 "Q + ”X 1 + 111/ KI) + [Sol Vmax(1) = k5 [Etotl/ Kl + K1 le + 1 and KM“) = KoKi/ (Kl + Kl /K2+ 1) Formation of Product 2 in the presence of inhibitor: [532] = [Ellsol/ KoKz [Etot] [El+[E11301/K0+[Ellsol/K0KI+ [Ellsol/K0K2+ [Elm/K. = s / K [ 0] K0 2 *(KoKz/KOKZ) 1 + 1301/ K0 + [301/ K0K1+ [501/ KoKz + [ll/Kl : [so] KoKz + [50119 + [so] + [song/K. + [I] K0K2/ K. Therefore: [E32] = [Etot][SO] KoKz + [Sole + [30] + [30] Kz/K1+ [I] K01(2/ Kl Since v(2) = k6[ES2] and V(2)= vmax(2)[s] Km(2)(l + [ll/K.) + [S] ; then v(2) = k6 [1510111301 * (1/ K2+ K2/KI + 1) K0K2(1 +1Il/KI)+1301(K2+K2/Ki+ 1) (“19+ Kz/K1+1) Rearranges to: v(2) = k6 [E.,.|[s.,]/ (K; + K; /K. + 1) (KoKz/ (K2+ K2 ”Q + ”X 1 + 111/ KI) ‘1' 1301 v...,,(2)= k,[E.,.]/ K2+ K2 /K.+1 and KM(2) = KoK2/(K2+K2/K.+ 1) Relationship of KLto ICfQ values for Scheme 1: V( 1) + inhibitor = 0.5V( 1) — inhibitor gives the relationship: 130 Ki = KM (ICSO) KM + [3] Inhibition of Product 1 by a competitive inhibitor: KI = KIKO (IC50(1))/ (K1 + K1 /K2+ 1) (KIKO/ (KI + KI/K2+1))+1501 KIKIKO + KIKIISOI + K1130] + K1150] K1 /K2 = KlKo (IC50(1)) ICsoU) = KI+ Kllsol/ K0+ Kllsol/ K1K0+ K11:SO]/K2 K0 “350(1) = Kl(1 + 1301/ Ko‘l' 1501/ KiKo+ 1501/ K2 K0) Inhibition of Product 2 by a competitive inhibitor: KI = K2 K0 (1C50(2))/ (K2+ Kz/KI + 1) (K2 1(0/(K2+ Kz/K1+1))+1501 K.K2K0 + K.K2[So] + K.[S..] + K.[S0] Kz/K. = KZKO (IC50(2)) 1C5..(2) = K.+ K.[S..]/ K0+ K.[s0]/ K2K0+ K.[S..]/ K. K0 1C5..(2) = K.(1 + [sol/ K0 + [soy K.K0 + [501/ K2 K0) 1C5..(1) = “250(2) Scheme 2 Derivations (no inhibitor): Term definitions V“) = kzlEisll KO: [511/ [52] K2 = [5211301/15232] v(2) = 1(4[E282] K. = [E.][So]/ [E.S.] [Stot] = [SO] [Etot]=[E1] +[Ez] +1531] + [5252] [El] = [Eleo [Elsi] = 15111301/ KI [52] = [Ed/K.) [13252] = [13211301/ K2 Formation of Product 1 [Elsi] = [Bullsol/ K1 [Etot] [E.] + [Ell/K0 + [Billsol/ K1 + [13111501/ K0 K2 131 1301/ K1 * (K./ K.) 1 + 1/K0+ 1301/ K1 + [S()]/ K0 K2 = [So] Since v(l) = k2[E.S.] and K1 + Kr/K0+ [301+ [30] KI/Ko K2 V“) = Vmax(1)[S]/KM(1) + [S], V(1) = k2 [Etotl [$01 ; Vmax(1) = R2 [Etot] Kr(1 + l/Ko + 1301/ K0 K2) + 1501 Km(1) = K.(l + UK)“ 1501/ K0 K2) Formation of Product 2 [5232] = [Ezllsol/ K2 [Etot] 1132] + [5211(0+ [521130] K0/ K1 + [Ezllsol/ K2 = [So]/ K2 * (K2/ K2) 1 + Ko+ Kolsol/ KI + 1301/ K2 = [So] Since v(2) = k4[E282] and K2+ KzKo + KoKzlsol/ K1 + [30] V(2) = Vmax(2)[s]/ KMQ) + [S], v(2) = R4 [Etotl 1501 3 an(2) = k4 [Etot] K2(l + Ko+ Kolsoll Ki) '1' 1501 KM(2) = K2(1 ‘1' K0 + Kolsoll Ki) Scheme 2 Derivations (with inhibitor): Term Definitions Same as above with the following additions and changes: Klrl)=lEllllol/[EI11] KHZ) = [Ezlllol/ [3212] [E101] = [Br] + [E21 + [EISI] + [13232] + [Elli] + [13212] 1E1111=lEllllol/KI(I) [13212] = [Ezlllol/ K10): [Ellllol/ KOKKZ) Formation of Product 1 in the presence of inhibitor 132 [E181] = [Eu][So]/ K. [Em] [Ell + [E.]/K0+ [13111301/ K1 + [EIHSol/ K0 K2+ [Ellllol/KI<1)+ [Ellllol/KoKuzi = [301/ K. 1 + 1/K..+ [soy K. + [801/ K. K2+ [Io]/ K....+ [101/K0Ku2) * (Kl/ KI) 130] K. ‘1' K|/K()+ [8.)] + [80] 1(1/ KOK2+ [10]K|/ K|(|)+ [10]K|/ K0K|(2) ; ThCI'CfOI'CI V“) = k21Etot][SO] * 1/(1 + Ki/KoKz) K|(1 + [10]/K|(|)+ 1/K()+ [1()]/ K0K1(2)) + [So](l + K|/K0K2) ll“ '1' Kl/KoKz) W” = kZIEtotHSOI/(l + Kr/KoKz) Kr(1 + 1101/ K|(r)+ l/Ko+ 1101/ KunKo) (1 + Kr/KoKz) + 1301 KM“) = Kl“ + 1101/ KI(I)+ l/Ko+ 1101l KI(2)K0) (1 + Kr/KoKz) Vmax(1) = kzlEror1/(l + Kr/KoKz) Formation of Product 2 in the presence of inhibitor [13252] = [Ezllsol/ K2 [Etot] [52] + 11321180+ [E2] KO [301/ K. + 15211501/ K2 + [EZlUOl/ K1(2)+ [5211(01101/ KI(I) = [30] K2 + KoKz + K0K2[S0]/ K. + [S..] + [10]K2/ K..2,+ [10]K0K2/ K...) ; Therefore: v(2)= k4[E.,.][s0] * 1/(1 +K0K2/ K.) Kzl1 + [IO]K0/Kl(l)+ Ko+110]/KI(2))+1301(1 + KoKz/ KI) 1/(1 + KoKz/ Kr) v(2) = k41Etor11501/(1 + KoKz/ KI) K2(1 + [lolKo/KI(1)+ K0+ 1101/ KI(2)) (1 + KoKzl Kr) + 1501 133 KMa) = K2(1 + [lolKo/KI(I)+ Ko+ 1101/ Kl(2)) (1 + KoKz/ Kr) Vmax(2) = k41Etot1/(1 + KOKZ/ Kl) Relationship of K. to IC fl values for Scheme 2: v( 1) + inhibitor = 0.5v( 1) ~ inhibitor Inhibition of Product 1 by a competitive inhibitor k2[Etot][SO]/(l + KI/KoKz) = 0.5 k2[Etot][SO] K.“ + [10]/ K|(|)+ 1/K0+ [10]/K1(2)K0) + [S ] K1“ '1' l/K0'1' [So]/ K0 K2) ‘1' [So] 0 (1 + KI/KoKz) Multiply left side by (l + K./K0K2)/ (1 + K./K..K2), rearrange and multiply both sides by 2 to give: 2K1+2K|/K0+2KI[Sol/K0K2+2lsol = K1+KI[101/Kiri)+K1/K0+K1[101/K0Kr(2)+130]+[SolKi/KoKz Subtract K., K./K0, [S0] and [So]K./K0K2 from both sides, divide by K. and rearrange to give: 1150(1)] = 1 +1/Ko +1501(1/K0K2 +1/KI) 1/K|(|)+ l/ K|(2) K0 Inhibition of Product 2 by a competitive inhibitor k41Etot][SO]/(1 + KoKz/ K1) = 0-5 k41Etot][SO] K2(1 + 1101/ KI(2)+ K0+ K01101/ Kl(l)) + [S ] K2(1 + K0+ Ko[So]/ K1) + [30] 0 (l + KoKz/ K1) Multiply left side by (1 + K0K2/ K.)/ (1 + K0K2/ K.), rearrange and multiply both sides by 2 to give: 2K2+2K0K2+21301K0K2/K1 +2130] = K2+K0K2+K21101/KI(2)+ K0K21101/ KI(I)+130]+[SO]K0K2/KI Subtract K2, K0K2 , [S0] and [S0]K0K2/K. from both sides, divide by K2 and rearrange to give: 134 1150(2)] = 1 + K0 + 1501(K0/Kr ‘1’ 1/K2) 1/K|(2)+ Ko/ Kl“) “350(1) ¢1Cso(2) The only situation in which 1C50(1) = IC50(2) is that in which K. = 1 135 REFERENCES 136 10. ll. 12. 13. References Smith, W.L. and D.L. DeWitt, Prostaglandin endoperoxide H synthases-I and -2, in Advances in Immunology, F.J. Dixon, Editor. 1996, Academic Press: San Diego, CA. p. 167-215. Smith, W.L., R.M. Garavito, and D.L. DeWitt, Prostaglandin endoperoxide H synthases (cyclooxygenases)-l and -2. Journal of Biological Chemistry, 1996. 271(52): p. 33157-33160. Mamett, L.J., et al., Arachidonic Acid Oxygenation by COX-1 and COX-2. Journal of Biological Chemistry, 1999. 274: p. 22903-22906. Oates, J.A., et al., Clinical implications of prostaglandin and thromboxane A2 formation. New England Journal of Medicine, 1988. 319(11): p. 689-698. Oates, J .A., et al., Clinical implications of prostaglandin and thromboxane A2 formation: second of two parts. New England Journal of Medicine, 1988. 319(12): p. 761-767. Smith, W.L. and LJ. Mamett, Prostaglandin endoperoxide synthase: structure and catalysis. Biochimica et Biophysica Acta, 1991. 1083: p. 1-17. Eberhart, C.E., et al., Up-regulation of cyclooxygenase 2 gene expression in human colorectal adenomas and adenocarcinomas. Gastroenterology, 1994. 107(4): p. 1183-1188. Kargrnan, S.L., et al., Expression of Prostaglandin G/H synthase-1 and -2 protein in human colon cancer. Cancer Research, 1995. 55: p. 2556-2559. Stephenson, J ., More evidence links NSAID, estrogen use with reduced Alzheimer risk [news]. Jama, 1996. 275(18): p. 1389-90. Van der Ouderaa, F.J., et al., Purification and characterisation of prostaglandin endoperoxide synthetase from sheep vesicular glands. Biochimica et Biophysica Acta, 1977. 487(2): p. 315-331. Kulmacz, R.J. and WE. Lands, Prostaglandin H synthase. Stoichiometry of heme cofactor. Journal of Biological Chemistry, 1984. 259(10): p. 6358-6363. Roth, G.J., E.T. Machuga, and P. Strittmatter, The heme-binding properties of prostaglandin synthetase fi'om sheep vesicular gland. Journal of Biological Chemistry, 1981. 256(19): p. 10018-10022. Lambeir, A.M., et al., Spectral properties of the higher oxidation states of prostaglandin H synthase. Journal of Biological Chemistry, 1985. 260(28): p. 14894-14896. 137 14. 15. l6. 17. 18. 19. 20. 21. 22. 23. 24. Hara, 8., et al., Isolation and molecular cloning of prostacyclin synthase from bovine endothelial cells. Journal of Biological Chemistry, 1994. 269(31): p. 19897-19903. Kuwamoto, S., et al., Inverse gene expression of prostacyclin and thromboxane synthases in resident and activated peritoneal macrophages. F EBS Letters, 1997. 409(2): p. 242-246. Jakobsson, P., et al., Identification of human prostaglandin E synthase: A microsomal glutathione-dependent, inducible enzyme, constituting a potential novel drug target. Proceedings of the National Academy of Science U. S. A., 1999. 96: p. 7220-7225. O'Neill, GP. and A.W. Ford-Hutchinson, Expression of mRNA for cyclooxygenase-1 and cyclooxygenase-2 in human tissues. FEBS Letters, 1993. 330(2): p. 156—160. Kraemer, S.A., S.A. Meade, and D.L. DeWitt, Prostaglandin endoperoxide synthase gene structure: identification of the transcriptional start site and 5’- flanking regulatory sequences. Archives of Biochemistry and Biophysics, 1992. 293(2): p. 391-400. Smith, W.L. and D.L. DeWitt, Biochemistry of prostaglandin endoperoxide synthases-I and -2 and their differential suceptibility to nonsteroidal anti- inflammatory drugs. Seminars in Nephrology, 1995. 15: p. 179-194. DeWitt, D.L., Cox-2-selective inhibitors: the new super aspirins. Molecular Pharmacology, 1999. 55(4): p. 625-631. DeWitt, D.L., et al., The aspirin and heme-binding sites of ovine and murine prostaglandin endoperoxide synthases. Journal of Biological Chemistry, 1990. 265(9): p. 5192-5198. DeWitt, D.L. and W.L. Smith, Primary structure of prostaglandin G/H synthase from sheep vesicular gland determined from the complementary DNA sequence. Procedings of the National Academy of Science USA, 1988. 85: p. 1212-1416. Van der Ouderaa, F.J., et al., Acetylation of prostaglandin endoperoxide synthetase with acetylsalicylic acid. European Journal of Biochemistry, 1980. 109(1): p. 1-8. Leslie, C.C., Properties and regulation of cytosolic phospholipase A2. Journal of Biological Chemistry, 1997. 272(27): p. 16709-16712. 138 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. Tischfield, J .A., A reassessment of the low molecular weight phospholipase A2 gene family in mammals. Journal of Biological Chemistry, 1997. 272(28): p. 17247-17250. Johansen, B., et al., Expression of cytosolic and secreted forms of phospholipase A(2) and cyclooxygenases in human placenta, fetal membranes, and chorionic cell lines. Prostaglandins and Other Lipid Mediators, 2000. 60(4-6): p. 119-25. Schievella, A.R., et al., Calcium-mediated translocation of cytosolic phospholipase A2 to the nuclear envelope and endoplasmic reticulum. Journal of Biological Chemistry, 1995. 270(51): p. 30749-30754. Glover, S., et al., T ranslocation of the 85-kDa phospholipase A2 from cytosol to the nuclear envelope in rat basophilic leukemia cells stimulated with calcium ionophore or IgE/antigen [published erratum appears in J Biol Chem 1995 Sep I;270(35):20870]. Journal of Biological Chemistry, 1995. 270(25): p. 15359-67. Borsch-Haubold, A.G., et al., Identification of the phosphorylation sites of cytosolic phospholipase A2 in agonist-stimulated human platelets and HeLa cells. Journal of Biological Chemistry, 1998. 273(8): p. 4449-4458. Lin, L.L., et al., cPLA2 is phosphorylated and activated by MAP kinase. Cell, 1993. 72(2): p. 269-78. Nemenoff, R.A., et al., Phosphorylation and activation of a high molecular weight form of phospholipase A2 by p42 microtubule—associated protein 2 kinase and protein kinase C. Journal of Biological Chemistry, 1993. 268(3): p. 1960-4. Lin, L.-L., A.Y. Lin, and D.L. DeWitt, Il-Ialpha induces the accumulation of cPLA2 and the release of Prostaglandin E2 in human fibroblasts. Journal of Biological Chemistry, 1992. 267: p. 23451-23454. Schalkwijk, C.G., et al., Interleukin-l beta-induced cytosolic phospholipase A2 activity and protein synthesis is blocked by dexamethasone in rat mesangial cells. FEBS Letters, 1993. 333(3): p. 339-343. Hoeck, W.G., et al., Cytoplasmic phospholipase A2 activity and gene expression are stimulated by tumor necrosis factor: dexamethasone blocks the induced synthesis. Proceedings of the National Academy of Science U. S. A., 1993. 90(10): p. 4475-4479. Kramer, R.M., et al., Ca(2+)-sensitive cytosolic phospholipase A2 (cPLA2) in human platelets. Journal of Lipid Mediators, 1993. 6(1-3): p. 209-216. 139 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. Clark, J.D., et al., A novel arachidonic acid-selective cytosolic PLA2 contains a Ca2+-dependent translocation domain with homology to PKC and GAP. Cell, 1991. 65: p. 1043-1051. Ackermann, E.J. and EA. Dennis, Mammalian calcium-independent phospholipase A2. Biochimica et Biophysica Acta, 1995. 1259(2): p. 125-36. Balsinde, J. and BA. Dennis, Function and inhibition of intracellular calcium- independent phospholipase A2. Journal of Biological Chemistry, 1997. 272(26): p. 16069-16072. Murakami, M., et al., T he functions of five distinct mammalian phospholipase A2S in regulating arachidonic acid release. Type 11a and type V secretory phospholipase A2S are functionally redundant and act in concert with cytosolic phospholipaseAZ. Journal of Biological Chemistry, 1998. 273(23): p. 14411-23. Balsinde, J. and EA. Dennis, Distinct roles in signal transduction for each of the phospholipase A2 enzymes present in P388D1 macrophages. Journal of Biological Chemistry, 1996. 271(12): p. 6758-6765. Balsinde, J ., M.A. Balboa, and EA. Dennis, Functional coupling between secretory phospholipase A2 and cyclooxygenase-2 and its regulation by cytosolic group I V phospholipase A2. Proceedings of the National Academy of Science U. S. A., 1998. 95(14): p. 7951-7956. Balsinde, J., et al., Arachidonic acid mobilization in P388D1 macrophages is controlled by two distinct Ca(2+)-dependent phospholipase A2 enzymes. Proceedings of the National Academy of Science U. S. A., 1994. 91(23): p. 1 1060-11064. Qiu, Z.H., et al., The role of calcium and phosphorylation of cytosolic phospholipase A2 in regulating arachidonic acid release in macrophages. Journal of Biological Chemistry, 1998. 273(14): p. 8203-11. Spencer, A.G., et al., Subcellular localization of prostaglandin endoperoxide H synthases-1 and -2 by immunoelectron microscopy. Journal of Biological Chemistry, 1998. 273(16): p. 9886-9893. Reddy, ST. and HR. Herschman, Prostaglandin synthase-1 and prostaglandin synthase-2 are coupled to distinct phospholipases for the generation of prostaglandin D2 in activated mast cells. Journal of Biological Chemistry, 1997. 272(6): p. 3231-3237. Murakami, M., et al., Prostaglandin E2 amplifies cytosolic phospholipase A2- and cyclooxygenase—2-dependent delayed prostaglandin E2 generation in mouse 140 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. osteoblastic cells. Enhancement by secretory phospholipase A2. Journal of Biological Chemistry, 1997. 272(32): p. 19891-19897. Roshak, A., G. Sathe, and LA. Marshall, Suppression of monocyte 85-kDa phospholipase A2 by antisense and effects on endotoxin-induced prostaglandin biosynthesis. Journal of Biological Chemistry, 1994. 269(42): p. 25999-26005. Marshall, L.A., et al., Depletion of human monocyte 85-kDa phospholipase A2 does not alter leukotriene formation. Journal of Biological Chemistry, 1997. 272(2): p. 759-765. Naraba, H., et al., Segregated coupling of phospholipases A2, cyclooxygenases, and terminal prostanoid synthases in different phases of prostanoid biosynthesis in rat peritoneal macrophages. Journal of Immunology, 1998. 160(6): p. 2974- 2982. Kuwata, H., et al., C ytosolic phospholipase A2 is required for cytokine-induced expression of type ”A secretory phospholipase A2 that mediates optimal cyclooxygenase—2-dependent delayed prostaglandin E2 generation in rat 3Y1 fibroblasts. Journal of Biological Chemistry, 1998. 273(3): p. 1733-1740. Reddy, ST. and HR. Herschman, T ranscellular prostaglandin production following mast cell activation is mediated by proximal secretory phospholipase A2 and distal prostaglandin synthase 1. Journal of Biological Chemistry, 1996. 271(1): p. 186-191. Kanai, N., et al., Identification and characterization of a prostaglandin transporter. Science, 1995. 268(5212): p. 866-869. Lu, R., et al., Cloning, in vitro expression, and tissue distribution of a human prostaglandin transporter cDNA(hPGT). Journal of Clinical Investigation, 1996. 98(5): p. 1142-9. Kennedy, 1., et al., Studies on the characterization of prostanoid receptors. Advances in Prostaglandin, Thromboxane and Leukotriene Research, 1983. 11: p. 327-32. Mizuno, K., S. Yamamoto, and W.E.M. Lands, Effects of non-steroidal anti- inflammatory drugs on fatty acid cyclooxygenase and prostaglandin hydroperoxidase activities. Prostaglandins, 1982. 23(5): p. 743-757. Koshkin, V. and H.B. Dunford, Coupling of the peroxidase and cyclooxygenase reactions of prostaglandin H synthase. Biochimica et Biophysica Acta, 1999. 1430(2): p. 341-348. 141 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. Smith, W.L. and W.E.M. Lands, Oxygenation of polyunsaturated fatty acids during prostaglandin biosynthesis by sheep vesicular glands. Biochemistry, 1972. 11: p. 3276-3282. Landino, L.M., et al., Mutational analysis of the role of the distal histidine and glutamine residues of prostaglandin-endoperoxide synthase-2 in peroxidase catalysis, hydroperoxide reduction, and cyclooxygenase activation. Journal of Biological Chemistry, 1997. 272: p. 21565-21574. Dietz, R., W. Nastainczyk, and H.H. Ruf, Higher oxidation states of prostaglandin H synthase. Rapid electronic spectroscopy detected two spectral intermediates during the peroxidase reaction with prostaglandin 02. European Journal of Biochemistry, 1988. 171(1-2): p. 321-328. Wei, C., R.J. Kulmacz, and A.L. Tsai, Comparison of branched-chain and tightly coupled reaction mechanisms for prostaglandin H synthase. Biochemistry, 1995. 34(26): p. 8499-8512. Hamberg, M. and B. Samuelsson, 0n the mechanism of the biosynthesis of prostaglandins E-I and F-I-alpha. Journal of Biological Chemistry, 1967. 242(22): p. 5336-5343. Smith, W.L. and L.J. Mamett, Prostaglandin endoperoxide synthases, in Metal Ions in Biological Systems, H. Sigel and A. Sigel, Editors. 1994, Marcel Dekker: New York. p. 163-199. Lu, 6., et al., Comparison of the peroxidase reaction kinetics of prostaglandin H synthase-I and -2. Journal of Biological Chemistry, 1999. 274: p. 16162-16167. Karthein, R., et al., Higher oxidation states of prostaglandin H synthase. EPR study of a transient tyrosyl radical in the enzyme during the peroxidase reaction. European Journal of Biochemistry, 1988. 171(1-2): p. 313-320. Tsai, A., et al., Rapid Kinetics of Tyrosyl Radical Formation and Heme Redox State Changes in Prostaglandin H Synthase -I and -2. Journal of Biological Chemistry, 1999. 274: p. 21695-21700. Malkowski, M.G., et al., The x-ray crystal structure of prostaglandin endoperoxide H synthase-I complexed with arachidonic acid. Science, 2000. Mason, R.P., et al., A carbon-centered free radical intermediate in the prostaglandin synthetase oxidation of arachidonic acid. Spin trapping and oxygen uptake studies. Journal of Biological Chemistry, 1980. 255(11): p. 5019-5022. Schreiber, J ., T.E. Eling, and RP. Mason, The oxidation of arachidonic acid by the cyclooxygenase activity of purified prostaglandin H synthase: spin trapping of 142 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. a carbon-centered free radical intermediate. Archives of Biochemistry and Biophysics, 1986. 249(1): p. 126-136. Kwok, P.-Y., F.W. Muellner, and J. Fried, Enzymatic conversion of 10,10- difluoroarachidonic acid with PGH synthase and soybean lipoxygenase. Journal of the American Chemical Society, 1987. 109: p. 3692-3698. Tsai, A., R.J. Kulmacz, and G. Palmer, Spectroscopic evidence for reaction of prostaglandin H synthase-I tyrosyl radical with arachidonic acid. Journal of Biological Chemistry, 1995. 270(18): p. 10503-10508. Wu, G., et al., A mechanistic study of self-inactivation of the peroxidase activity in prostaglandin H synthase-1. Journal of Biological Chemistry, 1999. 274: p. 9231- 9237. Kulmacz, R.J., Attachment of substrate metabolite to prostaglandin H synthase upon rection with arachidonic acid. Biochemical and Biophysical Research Communications, 1987. 148: p. 539-545. Lecomte, M., et al., Covalent binding of arachidonic acid metabolites to human platelet proteins. Journal of Biological Chemistry, 1990. 265: p. 5178-5187. Tsai, A., et al., Characterization of the tyrosyl radicals in ovine prostaglandin H synthase-l by isotope replacement and site-directed mutagenesis. Journal of Biological Chemistry, 1994. 269(7): p. 5085-5091. Hsi, L.C., et al., An examination of the source of the tyrosyl radical in ovine prostaglandin endoperoxide synthase-1. Biochemical and Biophysical Research Communications, 1995. 207(2): p. 652-660. Shi, W., et al., EPR and ENDOR spectroscopic identification and characterization of the tyrosyl radicals in prostaglandin H synthase-1. Journal of the American Chemical Association, 2000. in press. Herschman, H.R., Prostaglandin synthase 2. Biochimica et Biophysica Acta, 1996. 1299(1): p. 125-140. Regier, M.K., et al., Subcellular localization of prostaglandin endoperoxide synthase-2 in murine 3T3 cells. Archives of Biochemistry and Biophysics, 1993. 301: p. 439-444. Otto, J.C. and W.L. Smith, The orientation of prostaglandin endoperoxide synthases-I and -2 in the endoplasmic reticulum. Journal of Biological Chemistry, 1994. 269(31): p. 19868-19875. 143 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. Picot, D., P.J. L011, and M. Garavito, The X-ray crystal structure of the membrane protein prostaglandin H2 synthase-I. Nature, 1994. 367: p. 243-249. Luong, C., et al., Flexibility of the NSAID binding site in the structure of human cylcooxygenase—Z. Nature Structural Biology, 1996. 3: p. 927-933. Kurumbail, R.G., et al., Structural basis for selective inhibition of cyclooxygenase-2 by anti-inflammatory agents. Nature, 1996. 384(6610): p. 644- 648. Hemler, M., W.E.M. Lands, and W.L. Smith, Purification of the cyclooxygenase that forms prostaglandins. Demonstration of two forms of iron in the holoenzyme. Journal of Biological Chemistry, 1976. 251: p. 5575-5581. Otto, J.C., D.L. DeWitt, and W.L. Smith, N-glycosylation of prostaglandin endoperoxide synthases-l and -2 and their orientations in the endoplasmic reticulum. Journal of Biological Chemistry, 1993. 268(24): p. 18234-18242. Meade, E.A., W.L. Smith, and D.L. DeWitt, Expression of the murine prostaglandin (PGH) synthase-.1 and PGH synthase-2 isozymes in cos-I cells. Journal of Lipid Mediators, 1993. 6: p. 119-129. Meade, E.A., W.L. Smith, and D.L. DeWitt, Diflerential inhibition of prostaglandin endoperoxide synthase (cyclooxygenase) isozymes by aspirin and other non-steroidal anti-inflammatory drugs. Journal of Biological Chemistry, 1993. 268: p. 6610-6614. Barnett, J ., et al., Purification, characterization and selective inhibition of human prostaglandin G/H synthase 1 and 2 expressed in the baculovirus system. Biochimica et Biophysica Acta, 1994. 1209(1): p. 130-139. Lands, W.E.M., J. Sauter, and G.W. Stone, Oxygen requirement for prostaglandin biosynthesis. Prostaglandins and Medicine, 1978. 1: p. 117-120. Seibert, K., et al., Pharmacological and biochemical demonstration of the role of cylooxygenase 2 in inflammation and pain. Procedings of the National Academy of Science USA, 1994. 91: p. 12013-12017. Penning, T.D., et al., Synthesis and biological evaluation of the 1.5-diarylpyrazole class of cyclooxygenase-2 inhibitors- identification of 4—[5-(4-methylphenyl)-3- (trifluoromethyl)-IH-pyrazol-1 -yl]benze nesulfonamide (SC-58635, celecoxib). Journal of Medicinal Chemistry, 1997. 40(9): p. 1347-1365. Prasit, P. and D. Riendeau, Selective Cyclooxygenase-2 Inhibitors, in Annual Reports in Medicinal Chemistry, Hagmann, Editor. 1997, Academic Press. 144 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. Van der Ouderaa, F.J., M. Buytenhek, and F .J. Slikkerveer, 0n the haemoprotein character of prostaglandin endoperoxide synthetase. Biochimica et Biophysica Acta, 1979. 572: p. 29-42. Ruf, H.H., D. Schuhn, and W. Nastainczyk, EPR titration of ovine prostaglandin H synthase with hemin. F EBS Letters, 1984. 165(2): p. 293-296. Shimokawa, T. and W.L. Smith, Essential histidines of prostaglandin endoperoxide synthase. Journal of Biological Chemistry, 1991. 266(10): p. 6168- 6173. Shimokawa, T., et al., Tvrosine 385 of prostaglandin endoperoxide synthase is required for cyclooxygenase catalysis. Journal of Biological Chemistry, 1990. 265(33): p. 20073-20076. Bhattacharyya, D.K., et al., Involvement of arginine 120, glutamate 524, and tyrosine 355 in the binding of arachidonate and 2-phenylpropionic acid inhibitors to the cyclooxygenase active site of ovine prostaglandin endoperoxide H synthase-I. Journal of Biological Chemistry, 1996. 271(4): p. 2179-2184. Mancini, J .A., et al., Arginine 120 of prostaglandin G/H synthase-1 is required for the inhibition by nonsteroidal anti-inflammatory drugs containing a carboxylic acid moiety. Joumal of Biological Chemistry, 1995. 270(49): p. 29372- 29377. Rieke, C.J., et al., The role of arginine I 20 of human prostaglandin endoperoxide H synthase-2 in the interaction with fatty acid substrates and inhibitors. Journal of Biological Chemistry, 1999. 274(24): p. 17109-17114. Loll, P.J., D. Picot, and R.M. Garavito, The structural basis of aspirin activity inferred from the crystal structure of inactivated prostaglandin H2 synthase. Nature Structural Biology, 1995. 2: p. 637-643. Bemlohr, D.A., et al., Intracellular lipid-binding proteins and their genes. Annual Review of Nutrition, 1997. 17: p. 277-303. Li, H. and TL. Poulos, The structure of the cytochrome p45OBM-3 haem domain complexed with the fatty acid substrate, palmitoleic acid. Nature Structural Biology, 1997. 4(2): p. 140-6. Gan, Q.F., et al., Defining the arachidonic acid binding site of human 15- lipoxygenase. Molecular modeling and mutagenesis. Journal of Biological Chemistry, 1996. 271(41): p. 25412-8. 145 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. LaLonde, J .M., et al., Adipocyte lipid-binding protein complexed with arachidonic acid. Titration calorimetry and X-ray crystallographic studies. Journal of Biological Chemistry, 1994. 269(41): p. 25339-47. Thuresson, E.D., K.M. Lakkides, and W.L. Smith, Different catalytically competent arrangements of arachidonic acid within the cyclooxygenase active site of prostaglandin endoperoxide H synthase-I lead to the formation of different oxygenated products. Journal of Biological Chemistry, 2000. 275: p. 8501-8507. Boyington, J .C., B.J. Gaffney, and L.M. Amzel, The three-dimensional structure of an arachidonic acid 15-lipoxygenase. Science, 1993. 260(5113): p. 1482-6. Loll, P.J., et al., Synthesis and use of iodinated nonsteroidal antiinflammatory drug analogs as crystallographic probes of the prostaglandin H2 synthase cyclooxygenase active site. Biochemistry, 1996. 35(23): p. 7330-7340. Mancini, J.A., et al., Mutation of serine-516 in human prostaglandin G/H synthase-2 to methionine or aspirin acetylation of this residue stimulates I5-R- HE T E synthesis. FEBS Letters, 1994. 342(1): p. 33-37. Lecomte, M., et al., Acetylation of human prostaglandin endoperoxide synthase-2 (cyclooxygenase-2) by aspirin. Journal of Biological Chemistry, 1994. 269(18): p. 13207-13215. O'Neill, G.P., et al., Overexpression of human prostaglandin G/H synthase-1 and -2 by recombinant vaccinia virus: inhibition by nonsteroidal anti-inflammatory drugs and biosynthesis of 15-hydroxyeicosatetraenoic acid. Molecular Pharmacology, 1994. 45(2): p. 245-254. Griswold, DE. and J.L. Adams, Constitutive cyclooxygenase (COX-1) and inducible cyclooxygenase (COX- 2): rationale for selective inhibition and progress to date. Medicinal Research Reviews, 1996. 16(2): p. 181-206. Gierse, J .K., et al., A single amino acid dijference between cyclooxygenase-1 (COX-1) and -2 (COX-2) reverses the selectivity of COX-2 specific inhibitors. Journal of Biological Chemistry, 1996. 271(26): p. 15810-15814. Guo, Q., et al., Role of Val509 in time-dependent inhibition of human prostaglandin H synthase-2 cyclooxygenase activity by isoform-selective agents. Journal of Biological Chemistry, 1996. 271(32): p. 19134-19139. Nalefski, E.A., et al., Delineation of two functionally distinct domains of cytosolic phospholipase A2, a regulatory Ca(2+)-dependent lipid-binding domain and a Ca(2+)-independent catalytic domain. Journal of Biological Chemistry, 1994. 269(27): p. 18239-18249. 146 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. Lands, W.E.M., et al., Inhibition of prostaglandin biosynthesis. Advances in the Biosciences, 1973. 9: p. 15-28. Laneuville, O., et al., Fatty acid substrate specificities of human prostaglandin- endoperoxide H synthase-I and -2. Formation of IZ-hydroxy-(9Z, 13E/Z, I5Z)- octadecatrienoic acids from alpha-linolenic acid. Journal of Biological Chemistry, 1995. 270(33): p. 19330-19336. Willis, A.L., in CRC Handbook of Eicosanoids: Prostaglandins and Related Lipids, A.L. Willis, Editor. 1987, CRC Press, Inc.: Boca Raton, FL. p. 3-46. Hamberg, M. and B. Samuelsson, Oxygenation of unsaturated fatty acids by the vesicular gland of sheep. Journal of Biological Chemistry, 1967. 242(22): p. 5344-5354. Baer, A.N., P.B. Costello, and EA. Green, Stereospecificity of the hydroxyeicosatetraenoic and hydroxyoctadecadienoic acids produced by cultured bovine endothelial cells. Biochimica et Biophysica Acta, 1991. 1085(1): p. 45-52. Kaduce, T.L., et al., Formation of 9-hydroxyoctadecadienoic acid fi'om linoleic acid in endothelial cells. Journal of Biological Chemistry, 1989. 264(12): p. 6823- 30. Abeywardena, M.Y., et al., In vivo formation of metabolites of prostaglandins 12 and [3 in the marmoset monkey (Callithrix jacchus) following dietary supplementation with tuna fish oil. Biochimica et Biophysica Acta, 1989. 1003(2): p. 161-6. Knapp, H.R., Prostaglandins in human semen during fish oil ingestion: evidence for in vivo cyclooxygenase inhibition and appearance of novel trienoic compounds. Prostaglandins, 1990. 39(4): p. 407-23. Leaver, H.A., A. Howie, and NH. Wilson, The biosynthesis of the 3-series prostaglandins in rat uterus after alpha-linolenic acid feeding: mass spectroscopy of prostaglandins E and F produced by rat uteri in tissue culture. Prostaglandins, Leukotrienes and Essential Fatty Acids, 1991. 42(4): p. 217-24. Engels, F., H. Willems, and PP Nijkamp, Cyclooxygenase-catalyzed formation of 9-hydroxylinoleic acid by guinea pig alveolar macrophages under non- stimulated conditions. FEBS Letters, 1986. 209(2): p. 249-53. Kulmacz, R.J., R.B. Pendleton, and W.E.M. Lands, Interaction between peroxidase and cyclooxygenase activities in prostaglandin-endoperoxide synthase. Interpretation of reaction kinetics. Journal of Biological Chemistry, 1994. 269: p. 5527-5536. 147 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. Morita, I., et al., Eflects of eicosapentaenoic acid on arachidonic acid metabolism in cultured vascular cells and platelets: species difference. Thrombosis Research, 1983. 31(2): p. 211-7. Spector, A.A., et al., Eicosapentaenoic acid and prostacyclin production by cultured human endothelial cells. Journal of Lipid Research, 1983. 24(12): p. 1595-604. Lagarde, M., et al., Uptake and effect on arachidonic acid oxygenation of some icosaenoic acids in human platelets. Biomedica Biochimica Acta, 1984. 43(8-9): p. S319-22. Nugteren, DH. and GA. Kivits, Conversion of linoleic acid and arachidonic acid by skin epidermal lipoxygenases. Biochimica et Biophysica Acta, 1987. 921(1): p. 135-41. Camp, RD. and NJ. Fincham, Inhibition of ionophore-stimulated leukotriene B4 production in human leucocytes by monohydroxy fatty acids. British Journal of Pharmacology, 1985. 85(4): p. 837-41. Yamaja Setty, B.N., M. Berger, and M.J. Stuart, 13-Hydroxyoctadeca-9,I 1- dienoic acid (I3—HODE) inhibits thromboxane A2 synthesis, and stimulates 12- HE T E production in human platelets. Biochemical and Biophysical Research Communications, 1987. 148(2): p. 528-33. Larsen, L.N., E. Dahl, and J. Bremer, Peroxidative oxidation of leuco- dichlorofluorescein by prostaglandin H synthase in prostaglandin biosynthesis from polyunsaturated fatty acids. Biochimica et Biophysica Acta, 1996. 1299(1): p. 47-53. Oliw, E.H., Metabolism of 5(6)Oxidoeicosatrienoic acid by ram seminal vesicles. Formation of two stereoisomers of 5—hydroxyprostaglandin 11. Journal of Biological Chemistry, 1984. 259(5): p. 2716-21. Elliott, W.J., et al., The metabolic transformations of columbinic acid and the effect of topical application of the major metabolites on rat skin. Journal of Biological Chemistry, 1985. 260(2): p. 987-92. Nugteren, DH. and E. Christ Hazelhof, Naturally occurring conjugated octadecatrienoic acids are strong inhibitors of prostaglandin biosynthesis. Prostaglandins, 1987. 33(3): p. 403-17. Balazy, M., Metabolism of 5,6-epoxyeicosatrienoic acid by the human platelet. Formation of novel thromboxane analogs. Journal of Biological Chemistry, 1991. 266(35): p. 23561-7. 148 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. Smith, W.L., L.J. Mamett, and D.L. DeWitt, Prostaglandin and thromboxane biosynthesis. Pharmacology and Therapeutics, 1991. 49(3): p. 153-179. Smith, W.L., D.L. DeWitt, and R.M. Garavito, Cyclooxygenases: Structural, Cellular and Molecular Biology. Annual Review of Biochemistry, 2000. 69: p. 149-182. Raz, A., et al., Regulation of prostanoid synthesis in human fibroblasts and human blood monocytes by interleukin-1, endotoxin, and glucocoticoids, in Advances in Prostaglandin, T hromboxane, and Lekotriene Research, B. Samuelsson, Editor. 1990, Raven Press, Ltd: New York. p. 22-27. Jones, D.A., et al., Molecular cloning of human prostaglandin endoperoxide synthase type II and demonstration of expression in response to cytokines. Journal of Biological Chemistry, 1993. 268: p. 9049-9054. Hulkower, K.I., et al., Interleukin-I beta induces cytosolic phospholipase A2 and prostaglandin H synthase in rheumatoid synovial fibroblasts. Evidence for their roles in the production of prostaglandin E2. Arthritis and Rheumatism, 1994. 37(5): p. 653-661. Matijevic-Aleksic, N., et al., Diflerential expression of thromboxane A synthase and prostaglandin H synthase in megakaryocytic cell line. Biochimica et Biophysica Acta, 1995. 1269(2): p. 167-175. Foegh, M.L., Obstretrics and Gynecology, in Prostaglandins in Clinical Pratice, W.D.e.a. Watkins, Editor. 1989, Raven Press, Ltd: New York. p. 131-140. Evett, G.E., et al., Prostaglandin G/H isoenzyme 2 expression in fibroblasts: regulation by dexamethasone, mitogens, and oncogenes. Archives of Biochemistry and Biophysics, 1993.306: p. 169-177. Kujubu, D.A., et al., T [S] 0, a phorbol ester tumor promoter inducible mRNA from Swiss 3T 3 cells, encodes a novel prostaglandin synthase/cyclooxygenase homologue. Journal of Biological Chemistry, 1991. 266(20): p. 12866-12872. Xie, W. and HR. Herschman, v-src induces prostaglandin synthase 2 gene expression by activation of the c-Jun N-terminal kinase and the c-Jun transcription factor. Journal of Biological Chemistry, 1995. 270(46): p. 27622- 27628. Xie, W. and HR. Herschman, Transcriptional regulation of prostaglandin synthase 2 gene expression by platelet-derived growth factor and serum. Journal of Biological Chemistry, 1996. 271(49): p. 31742-31748. 149 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. Wendt, K.U., K. Poralla, and GE. Schulz, Structure and function of a squalene cyclase. Science, 1997. 277(5333): p. 1811-1815. Li, Y., et al., The membrane association sequences of the prostaglandin endoperoxide synthases-1 and -2 isozymes. Journal of Biological Chemistry, 1998. 273(45): p. 29830-29837. Spencer, A.G., et al., The membrane binding domains of prostaglandin endoperoxide H synthase-I and -2.' Peptide mapping and mutational analysis. Journal of Biological Chemistry, 1999. 274: p. 32936-32942. Gierse, J .K., et al., Expression and selective inhibition of the constitutive and inducible forms of human cyclo-oxygenase. Biochemistry Journal, 1995. 305(Pt 2): p. 479-484. Juranek, 1., H. Suzuki, and S. Yamamoto, Aflinities of various mammalian arachidonate lipoxygenases and cyclooxygenases for molecular oxygen as substrate. Biochimica et Biophysica Acta, 1999. 1436(3): p. 509-518. Rome, L.H. and W.E.M. Lands, Structural requirements for time-dependent inhibition of prostaglandin biosynthesis by anti-inflammatory drugs. Procedings of the National Academy of Science USA, 1975. 72(12): p. 4863-4865. Marshall, P.J. and R.J. Kulmacz, Prostaglandin H synthase: distinct binding sites for cyclooxygenase and peroxidase substrates. Archives of Biochemistry and Biophysics, 1988. 266(1): p. 162-170. Hecker, M., et al., Identification of novel arachidonic acid metabolites formed by prostaglandin H synthase. European Journal of Biochemistry, 1987. 169: p. 113- 123. Xiao, G., et al., Analysis of hydroperoxide-induced tyrosyl radicals and lipoxygenase activity in aspirin-treated human prostaglandin H synthase-2. Biochemistry, 1997. 36(7): p. 1836-1845. Shimokawa, T. and W.L. Smith, Prostaglandin endoperoxide synthase: The aspirin acetylation region. Journal of Biological Chemistry, 1992. 267: p. 12387- 12392. Laneuville, O., et al., Differential inhibition of human prostaglandin endoperoxide H synthases-I and -2 by nonsteroidal anti-inflammatory drugs. The Journal of Phamacology and Experimental Therapeutics, 1994. 271(2): p. 927- 934. Oliw, E.H., et al., Oxygenation of 5,8,11-eicosatrienoic acid by prostaglandin endoperoxide synthase and by cytochrome P450 monooxygenase: structure and 150 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. mechanism of formation of major metabolites. Archives of Biochemistry and Biophysics, 1993. 305(2): p. 288-297. Hamberg, M., C. Su, and E. Oliw, Manganese lipoxygenase. Discovery of a bis- allylic hydroperoxide as product and intermediate in a lipoxygenase reaction. Journal of Biological Chemistry, 1998. 273(21): p. 13080-13088. Guan, Z., et al., Interleukin-Ibeta-induced cyclooxygenase-2 expression requires activation of both c-Jun NH2-terminal kinase and p38 MAPK signal pathways in rat renal mesangial cells. Journal of Biological Chemistry, 1998. 273(44): p. 28670-28676. Gund, P. and T.Y. Shen, A model for the prostaglandin synthetase cyclooxygenation site and its inhibition by antiinflammatory arylacetic acids. Journal of Medicinal Chemistry, 1977. 9: p. 1146-1152. Greig, G.M., et al., The interaction of arginine 106 of human prostaglandin G/H synthase-2 with inhibitors is not a universal component of inhibition mediated by nonsteroidal anti-inflammatory drugs. Molecular Pharmacology, 1997. 52(5): p. 829-838. Wong, E., et al., Conversion of prostaglandin G/H synthase-I into an enzyme sensitive to PGHS-Z-selective inhibitors by a double His513 to Arg and Ile523 to Val mutation. Journal of Biological Chemistry, 1997. 272: p. 9280-9286. Rowlinson, S.W., et al., The Binding of Arachidonic Acid in the Cyclo-oxygenase Site of Mouse Prostaglandin Endoperoxide Sythase-2 (C OX-2). Journal of Biological Chemistry, 1999. 274. Bradford, M.M., A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry, 1976. 72: p. 248-254. Kulmacz, R.J., Prostaglandin 02 levels during reaction of prostaglandin H synthase with arachidonic acid. Prostaglandins, 1987. 34(2): p. 225-240. Sack, J .S., Chain — A Crystallographic Modeling Program. Journal of Molecular Graphics, 1988. 6: p. 224-225. Schneider, C. and AR. Brash, Stereospecificity of hydrogen abstraction in the conversion of arachidonic acid to 1 5R-HE T E by aspirin-treated cyclooxygenase- 2: Implications for the alignment of substrate in the active site. Journal of Biological Chemistry, 2000. 275: p. 4743-4746. 151 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. Chen, W., TR. Pawelek, and R.J. Kulmacz, Hydroperoxide Dependence and Cooperative C yclooxygenase Kinetics in Prostaglandin H Synthase-I and -2. Journal of Biological Chemistry, 1999. 274: p. 20301-20306. Capdevila, J .H., et al., The catalytic outcomes of the constitutive and the mitogen inducible isoforms of prostaglandin H2 synthase are markedly aflected by glutathione and glutathione peroxidase(s). Biochemistry, 1995. 34(10): p. 3325- 3337. Kulmacz, R.J. and L.H. Wang, Comparison of hydroperoxide initiator requirements for the cyclooxygenase activities of prostaglandin H synthase-1 and -2. Journal of Biological Chemistry, 1995. 270(41): p. 24019-24023. Mamett, L.J. and AS. Kalgutkar, Cyclooxygenase 2 inhibitors: discovery, selectivity and the future. Trends in Pharmacological Sciences, 1999. 20: p. 465- 469. Laneuville, 01., et al., Fatty acid substrate specificities of hyuman prostaglandin H synthases-I and -2.' Formation of 12 hydroxy-(9Z1 3E/Z,I 5Z)-Octadecatrienoic acids from alpha-linolenic acid. Journal of Biological Chemistry, 1995. 270: p. 19330-19336. Baer, A.N., P.B. Costello, and F .A. Green, Stereospecificity of the products of the fatty acid oxygenases derived from psoriatic scales. Journal of Lipid Research, 1991. 32(2): p. 341-7. Needleman, P., et al., Manipulation of platelet aggregation by prostaglandins and their fatty acid precursors: pharmacological basis for a therapeutic approach. Prostaglandins, 1980. 19(1): p. 165-81. Graff, G., Preparation of I5—L-Hydroperoxy-5,8,l 1,13-eicosatetraenoic acid (15- HPETE). Methods in Enzymology, 1982. 862p. 386-392. Gupta, B.L., Microdetermination Techniques for H202 in Irradiated Solutions. Microchemical Journal, 1973. 18: p. 363-374. Smith, W.L., et al., Fatty Acid Substrate Interactions with Cyclooxygenases, in Advances in Eicosanoid Research, C.N. Serhan and H.D. Perez, Eds. 2000, Springer-Verlag: Berlin. 312p. 53-65 Thuresson, E.D., et al., Prostaglandin Endoperoxide H Synthase-l: The Functions of C yclooxygenase Active Site Residues in the Binding, Positioning and Oxygenation of Arachidonic Acid. Journal of Biological Chemistry, 2000 submitted. 152