. .. 3‘ T... .yflw:.r2 . . 3’0ri 1.. I .m... 6M3? .. . A. ‘.\ 9: 14:3nt 0" 6.5.15» . ... , kruwnb; . . . . “wwfi . .Crfiwy‘rn .u... , .3. .. u , I. V . if. : v. C growth-.1: d, . Harri}: . .8 Ecstl‘: s l .V] 3":CL-Il i\ o. .v 1...! “and I» on 333......uvimh fifimfi 3?. nan. ‘ v.5. ‘ . if. .. :. .. LIBRARY Michigan State University This is to certify that the dissertation entitled ROLE OF C-TERMINAL 18 AMINO ACIDS FOR THE BIOLOGIGAL ACTIVITY OF . PROSTAGLANDIN ENDOPEROXIDE H SYNTHASE-2 presented by Hui-yuan Tang has been accepted towards fulfillment of the requirements for the PhD degree in Biochemistry Molecular Biology_ :D c) 7)qu Major Professor's Signature $5404- Date MSU is an affirmative-action. equal-opportunity employer . ._--_.-.—.—.- 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 5108 K:IProi/Acc&Pres/ClRCIDateDue.indd ROLE OF C-TERMINAL 18 AMINO ACIDS FOR THE BIOLOGIGAL ACTIVITY OF PROSTAGLANDTN ENDOPEROXIDE H SYNTHASE-Z By Hui-yuan Tang A DI S SERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry and Molecular Biology 2007 ABSTRACT ROLE OF C-TERMINAL l8 AMINO ACIDS FOR THE BIOLOGICAL ACTIVITY OF PROSTAGLANDIN ENDOPEROXIDE H SYNTHASE-2 By Hui—yuan Tang C yclooxygenases-I and —2 are membrane-localized heme-containing homodimers that catalyze the committed step in prostaglandin synthesis. COX-2 protein has a unique I8-amino acid cassette located four residues from the carboxyl terminus of all COX-2 species that is not found in COX-1. The hypothesis that I investigated was that this 18-amino acid cassette is responsible in part for the distinct biological activity of COX-2 by mediating protein-protein interactions. 293 T-Rex cell lines that inducibly express Flag-tagged native COX-2, COX-2 d81581-598 (mutant with amino acids 581-598 deleted), COX-1 and COX-1 ins580-598 (mutant with amino acids 580-598 of COX-2 inserted near the amino terminal), or 293 Freestyle cells that transiently express Flag-tagged native COX-2 were used to identify potential protein partners for the cyclooxygenase isoforms. Two proteins were identified by proteomic analyses that reproducibly co-purified with F lag-tagged COX-2, FAM44A (GI/Q8NFC6) and Heat shock-induced protein (01/ 188492). FAM44A has been reported as a 33OKDa protein with an AT-hook DNA binding domains and FAM44A protein can be phosphorylated upon DNA damage. The interaction between COX-2 and FAM44A could be an alternative way to regulate cell cycle progression. Heat shock proteins, like Heat shock-induced protein usually act like chaperones to guide protein folding. Thus, the Heat shock-induced protein we identified may help process COX-2 or be involved in its degradation. No reproducible protein partners were identified for COX-2 (lo/581698 or COX-I. Interestingly, the FAM44A protein was also purified with COX-1 ins580—598, which contains the 18- amino acid insert, providing further evidence that it may interact specifically with the 18-amino acid cassette of COX-2. One role of the IS—amino acid cassette is to mediate the degradation of COX-2 via the ER-associated degradation (ERAD) system. Kifnusenin, an inhibitor of terminal fucosidases that increases the COX-2 stability, was applied to 293 T-Rex cells expressing Flag-tagged COX-2 in an attempt to trap protein partners of COX-2 that might be involved in protein degradation. No additional proteins were identified with this treatment. Yet other COX-2 protein partners may exist whose interactions are too transient or have too low affinity to survive purification. Untransfected 293 cells do not express endogenous COX-2, and may not also express other protein partners for COX-2. The role of the 18-amino acid cassette on cell biology was also investigated. Our results demonstrated that 18-amino acid insert had no effect on cyclooxygenase activity, but reduced the number of colonies that could be detected that stably expressed COX-2. This may be due to reduce protein stability of COX-2 compared to COX-2 dc1581-598. The I8 amino acid insert had no effect on COX-2 protein subcellular location or cell growth. ACKNOWLEDGEMENTS I would like to express my gratitude to my advisor Dr. David L. DeWitt for his wonderful tutoring throughout my PhD study. His incredible guidance, technically and intellectually, has helped me accomplish this challenging thesis project. Without his continuous encouragement, I wouldn’t have been able to successfully complete my graduate career. Also I would like to make a grateful acknowledgement for my committee members. Their helpful advice and innovative ideas were extremely useful in my research. I am appreciative for all of the lab members, present or past. The friendship, technical help, and career advice they offered has helped me through my graduate study. I would like to thank Robin Goodwin for all the kind support I received ever since I joined the lab. I have constantly turned to her for suggestions on many issues and have learned a lot from her. I would like to thank Christi Hemming for giving me useful advice on both work and life. She has been a good friend that I can rely on and I enjoyed working with her. I would like to thank Doug Whitten for teaching me how to analyze the Mass Spectrometry data and all the helpful discussion. I would like to thank Drs. Uri R. Mbonye, Jiayan Liu and Christine Harman for all the thought-provoking discussion and advice. Thank you all for providing me such a pleasant environment to practice my research study. Finally I would like to thank my family for their selfless support and love. It is impossible for me to achieve anything without them. iv TABLE OF CONTENTS LIST OF TABLES ..................................................................................... vi LIST OF FIGURES .................................................................................. viii ABBREVIATIONS ................................................................................... xi CHAPTER 1: LITERATURE REVIEW Introduction ................................................................... 1 Reaction Mechanism of C yclooxygenase ................................. 5 Crystal Structures of Cyclooxygenase .................................... 7 C yclooxygenase Inhibitors .................................................. 9 Regulation of Cyclooxygenase gene expression ........................ 13 Different Biological Activity of Cyclooxygenase ...................... 16 CHAPTER II: PUTATIVE PROTEIN PARTNERS FOR CYCLOOXYGENASE PROTEINS Summary ..................................................................... 18 Introduction .................................................................. 1 9 Experimental Procedures .................................................... 21 Resultsil Discussion ................................................................... 57 CHAPTER III: EFFECT OF EXPRESSION OF CYCLOOXYGENASE PROTEINS ON CELL BIOLOGY Summary ..................................................................... 61 Introduction .................................................................. 62 Experimental Procedure ..................................................... 64 Results ........................................................................ 68 Discussion ................................................................... 90 CONCLUSION ...................................................................................... 93 APPENDIX .......................................................................................... 98 BIBLIOGRAPHY ................................................................................. l 14 Table 1. Table 2. Table 3a-l. Table 3a-2. Table 3a-2. Table 3a-2. Table 3a-3. Table 3b. Table 30-1. Table 3c-2. Table 3d-l. Table 3d-2. Table 36. Table 3f—1. LIST OF TABLES Comparison of the expression levels of different COX-2 mutants in Sf21 insect cells .................................................... 31 Luciferase Assay for individual colonies of stably transfected MCF-7 cells ...................................................................... 36 Proteins co-purified with Flag-tagged C OX-2 in 293 Flp-in cells (experiment l)39 Proteins co-purified with F lag-tagged COX-2 in 293 Flp-in cells (experiment 4) ...................................................... 44 Proteins co-purified with Flag-tagged COX-2 in 293 F lp-in cells (experiment 5) ..................................................... 44 Proteins co-purified with Flag-tagged COX-2 in 293 Flp-in cells (experiment 6) ..................................................... 44 Proteins co-purified with Flag-tagged COX-2 in 293 F lp-in cells (experiment 7) ..................................................... 44 Proteins co-purified with Flag-tagged COX-2 de1581-598 in 293 F lp-in cells ........................................................................ 45 Proteins co-purified with F lag-tagged COX-1 in 293 Flp-in cells (experiment 1) ...................................................... 46 Proteins co—purified with Flag-tagged COX-1 in 293 F lp-in cells (experiment 2) ...................................................... 46 Proteins co-purified with Flag-tagged COX-l in.s'580-598 in 293 F lp-in cells (experiment 1) ..................................................... 47 Proteins co-purified with Flag-tagged COX-1 insSSO-598 in 293 Flp-in cells (experiment 2) ..................................................... 48 Proteins non-specifically isolated on anti-flag agrose beads from 293 Flp-in cells ................................................................... 49 Proteins co-purified with Flag-tagged C OX-2 in 293 Freestyle cells (experiment 1) ................................................. 52 vi Table 3f-2. Proteins co-purified with Flag-tagged C OX-Z in 293 Freestyle cells (experiment 2) ................................................. 53 Table 3f—3. Proteins reproducibly co-purified with F lag-tagged COX-2 ............... 55 Table 3g. Proteins non-specially isolated on anti-flag agrose beads from in 293 Freestyle cells ................................................................ 56 Table 4. Transfection efficiency for COX proteins ................................... 69 Table 5. Colony forming Agar Assay for COX and prostaglandin E synthase expressing transfectants ......................................................... 82 Table 6. Half-Life of COX proteins ...................................................... 89 LIST OF TABLES (APPENDIX) Table 7. Plasmids constructed by QuickChangelM site-directed mutagenesis ...................................................................... 98 Table 8. Plasmids constructed by subcloning ........................................ 100 Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 103. Figure 10b. Figure 1 I. Figure 12. Figure 13. Figure 14. Figure 15. LIST OF FIGURES Biosynthetic pathway for the formation of prostanoids derived from arachidonic acid ................................................... 2 Sequence alignment of COX enzymes .......................................... 3 Domain Structures ofCOX-l and COX-24 Cyclooxygenase and peroxidase catalysis and suicide inactivation of COXs .............................................................. 6 Structure of ovine prostaglandin endoperoxide H synthase-1 . .8 Hypothetical orientation of arachidonate within the substrate binding pocket ofovine COX-110 Structure of representative nonselective and COX-2 selective NSAle ........................................................ 1 1 Comparison of the accessible volume of the sheep COX-1 and human COX-2 hydrophobic substrate binding pockets. . . . . 12 Regulatory elements in the human COX-2 promoter. . . . . . . . . . . . . . . 1 4 Western blot analysis of Flag-tagged COX—2 and C OX-2 de1581-598 eluted with 3XF lag peptide ............................. 32 Coomassie Blue Stained PAGE gel of Flag-tagged COX-2 and COX-2 del58l-598 eluted with 3X Flag peptide ................................... 33 Coomassie Blue Stained PAGE gel of Flag-tagged COX-2 eluted with TEV protease ............................. 34 Western blot analysis of the inducible expression of COX isoforms in Flp-In 293 cells38 Sequence alignment of FAM44 protein ...................................... 41 Western blot analysis of F lag-tagged C OX-2 expression in 293 Freestyle cells.. ..................................................................... 51 Western blot examining stability of expression of COX in transfected cell lines .............................................. 68 viii Figure 16a(1-3). Growth curve for T-Rex 293 cell lines expressing Flag-tagged COX-2 ............................................................ .73 Figure l6b(1-3). Growth curve for T-Rex 293 cell lines expressing Flag-tagged COX-2 dc]58l-598 ................................ 74 Figure l6c(1-3). Growth curve for T-Rex 293 cell lines expressing Flag-tagged COX-l insS80-598 75 Figure 16d(1-3). Growth curve for T—Rex 293 cell lines expressing Flag-tagged COX-1 ............................................... 76 Figure 17. Western blot examining stability of expression of mPGES in COX expressing Flp-ln 293 cell lines ....................................... 77 Figure 18a(1-3). Growth curve for T-Rex 293 cell lines expressing Flag-tagged COX-2 and mPGES ............................... 78 Figure 18b(1-3). Growth curve for T-Rex 293 cell lines expressing Flag-tagged COX-2 de1581-598 and mPGES79 Figure 18c(1-3). Growth curve for T-Rex 293 cell lines expressing Flag-tagged COX-1ins580-598 and mPGES80 Figure 18d(l-3). Growth curve for T-Rex 293 cell lines expressing Flag-tagged COX-1 and mPGES ............................................. 81 Figure 19(a-e). Immunocytochemistry of Flp-In 293 cells line stably transfected with Flag-tagged COX-2, Flag-tagged COX-2 de1581-598, Flag-tagged COX-2 Y371F, Flag-tagged COX-1 and Flag-tagged COX-l insS80-598 ......................................... 83 Figure 203. Determination of Ty; for Flag-tagged COX-2 ............................ 84 Figure 20b. Determination Ole/z for Flag-tagged COX-2 (16]58I-598 . ........84 Figure 20c. Determination ome. for Flag-tagged COX-1 ............................. 84 Figure 20d. Determination of T1,]; for Flag-tagged COX-l insS80-598 ..............84 Figure 20c. Determination of Tm for Flag-tagged COX-2 N5 80A ................... 84 Figure 213. Determination of Tm for COX-2 ........................................... 86 Figure 21b. Determination of T“: for COX-2 (Jar/58 l -598 ............................. 86 Figure 21c. Figure 21d. Figure 21c. Figure 22. Figure 23. Figure 24. Figure 25. Figure 26. Figure 27. Figure 28a. Figure 28b. Figure 28c. Figure 28d. Figure 29. Figure 30. Figure 31. Figure 32. Determination of Tm for COX-2 N580A .................................. 86 Determination of Tm for COX-l ............................................ 86 Detemiination Olezj for COX-1 {115580-598 ............................. 86 LIST OF FIGURES (APPENDIX) Map of COX-2 cloned into pFastBac (Invitrogen) at Not I site. . . . . 102 Map of His tagged-COX-I cloned into pIND (Invitrogen) at Hind [I] Site .................................................................. 103 Map of F lag-tagged COX-2 and Flag-tagged COX-2 dc1581-598 subcloned into pcDNAS/FRT/TO ............................................. 104 Map of F lag-tagged COX-2 subcloned into pOSML vector. . . . . . ....105 Map of pOG44 vector expressing Flp recombinase ....................... 106 F lp-In-T-Rex Tetracycline Inducible System .............................. 107 AP inducible expressing system ............................................. 108 AP2 l 967 ........................................................................ 108 Target gene vector pLH-lel-PL ............................................. 109 Transcription factor vector pLzNz-RHS3H/ZF3 ........................... 109 Bioluminescent reaction catalyzed by firefly luciferase .................. 1 10 Map ofvector pOTB7 that was source omeGES.........................11l Map of mPGES cloned into Blunt TOPO 11 after PCR ................... 1 12 Map of mPGES cloned into pBABE-puro vector at EcoR I site .......... 1 13 Images in this thesis/dissertation are presented in color. 18 aa 3’-UTR ADP ATM ATP ATP ARE C HX COX CREB EGF ER ERAD FBS FLAP HEME HMG Hu iNOS KIF MBD mPGES NSAID Ov PBS PEI PGHS PG POX PPAR TEV Tet Tl/fi do ABBREVIATIONS C-terminal 18-amino acid cassette of COX-2 3’ untranslated region arachidonic acid adenosine diphosphate ataxia telangiectasia mutated adenosine triphosphate ATM and Rad3 related AU-rich motif cycloheximide cyclooxygenase CAMP regulatory binding protein epidermal growth factor endoplasmic reticulum ER-associated degradation fetal bovine serum S-Iipoxygenase activating protein ferric iron protoporphyrin IX non-histone chromosomal protein (high mobility group) human inductible nitric oxide synthase kifunensine membrane binding domain microsomal prostaglandin E synthase non-steroidal anti-inflammatory drug ovine phosphate buffered saline polyethylenimine prostaglandin endoperoxide H synthase prostaglandin peroxidase peroxisome proliferator-activated receptors tobacco etch virus tetracycline half-life xi CHAPTER I LITERATURE REVIEW Introduction Prostaglandin endoperoxide H synthase-I and 2 (PGHS-1 and PGHS-2; also cyclooxygenase-l and 2, COX-1 and COX-2) are membrane-localized heme- containing homodimers that localize to the luminal side of the endoplasmic reticulum (ER) membrane and to the inner and outer nuclear membranes (1). They catalyze the committed steps in prostaglandin synthesis: a cyclooxygenase reaction in which arachidonate plus two molecules of 02 are converted to Prostaglandin G2 (PGGz) and a two electron peroxidase reduction of PGGz to PGH; (Fig. 1). Various prostaglandin synthases then catalyze the isomerization or reduction of PGHz to the biological active prostaglandins such as prostaglandin E3 (PGEg), prostaglandin F3 (PGFg), prostacyclin (P013) (Fig. 1). Prostaglandins are unsaturated carboxylic acids, consisting of a 20-carbon skeleton with a five member ring. Prostaglandins are ubiquitous autocrine/paracrine regulators that mediate a wide range of physiological functions, such as control of cell growth, contraction of smooth muscle, and modulation of inflammation (24).. Prostaglandin receptors are a subfamily of cell surface seven-transmembrane G- protein-coupled receptors, and are classified into 9 subgroups: DPl-2, EPl-4, FP, IP, and TP (5), that respond to the prostaglandins D, E, F2“, prostacyclin and thromboxane, respectively. Subtypes of PG-specific receptors couple with different signaling pathways. For example, the 4 PGE receptor subtypes individually regulate intracellular Call mobilization, as CAMP cyclase activity, and phosphodiesterase activity (6, 7). Thus the same prostaglandin can have opposing biological functions, depending on which subtype is expressed in a cell or the concentration of PG product. arachidonic acid C cloox enase 202 y yg OHM-m WCOOH Oil-w»- M PGG: 50H Thaw f‘v W COO H 01mm PGHZ 0H Figure 1. Biosynthetic pathway for the formation of prostanoids derived from arachidonic acid. Phospholipid is cleaved by the phospholipase A2 to generate the arachidonic acid. The COX enzymes convert arachidonic acid to form PGHg by two steps. Various terminal synthases then convert PGHg to different prostanoids. The primary structures of COX-1 and -2 from many species are known (I) (Fig 2). COX-1 and 2 contain different lengths of signal peptides that are cotranslationally cleaved from the nascent polypeptide by microsomal signal peptidases. At the COX-1 and -2 carboxyl termini is a four amino acid sequence STEL, thought to be an analogue to the KDEL retention sequence necessary for retention of proteins in the endoplasmic reticulum (8). Mature COX-1 and COX-2 contain 576 and 587 amino hCOX—l hCOX-Z hCUX—l hCUX-Z hCOX-l hCOX—Z hCUX—l hCUX-Z hCOX-l hCOX-Z hCOX-l hCOX-Z hCOX~l hCUX—2 hCUX—l hCOX-2 hCOX—l hCOX—Z hCOX—l hCOX-Z hCOX-l hCOX~2 IO 20 110 ‘10 50 6(1 l l l l l l MSR-SLLLRFLLFLLLL---PPLPVLLADP WLA-—-RALLLCAVL 70 80 90 100 110 120 l l l I l l GAPTPV\PCCYYPCQHQG[CVRFGLDRYQCDCIRIGYSGPNCIlPGLWTWLRVSLRPSPS ALSHTANPCCSHPCQVRGVCMSVGFDQYKCDCTRIGFYGEVCSTPEFLTRlKLFLKPTPN 130 I40 150 160 170 180 i l I i l l FTHFLLTHGRWFWEFVV-ATFIREMLMRLVLTVRSVLIPSPPTYNSAHDYISWESFSNVS TVHYILTHFKGFWNVVNNIPFLRNAIMSYVLTSRSHLIDSPPTYNADYGYKSWEAFSNLS 190 200 210 220 230 240 i l l l i l f . I l YYTRILPSVPKDCPTPMGTKCKKQLPDAQLLARRFLLRRKFlPDPQGTVLMFAFFAQHFT YYTRALPPVPDDCPTPLGVKGKKQLPDSNElVEKLLLRRKFIPDPQGSNMMFAFFAQHFT 250 260 270 280 290 300 l i l l l HQFFKTSGKMGPGFTKALGHGVDLGHIYGDNLERQYQLRLFKDGKLKYQVLDGEWYPPSV HQFFKTDHKRGPAFTNGLGHGVDLXHIYGETLARQRKLRLFKDGKMKYQIIDGEMYPPTV 310 320 330 340 350 360 l l l l l l EEAPVLMHYPRGIPPQSQMAVGQEVFGLLPGLMLYATLWLREHNRVCDLLKAEHPTWGDE KDTQAEMIYPPQVPEHLRFAVGQEVFGLVPGLMMYATIWLREHVRVCDVLKQEHPEWGDE 37(1 Ii80 1390 4(X) 41(1 {120 i l l l l l QLFQTTRLILIGETIKlVlEEYVQQLSGYFLQLKFDPELLFGVQFQYRVRIAMEFVHLYH QLFQTSRLILIGETIKIVlEDYVQHLSGYHFKLKFDPELLFVKQFQYQNRlAAEFNTLYH 430 440 450 460 470 480 l l l l i l WHPLMPDSFKVGSQEYSYEQFLFNTSMLVDYGVEALVDAFSRQIAGRIGGGRNMDHHILH WHPLLPDTFQIHDQKYNYQQFIYNNSILLEHGITQFVESFTRQIAGRVAGGRNVPPAVQK 490 500 510 520 530 540 l l i l l l VAVDVIRESREMRLQPFVEYRKRFGMKPYTSFQELVGEKEMAAELEELYGDIDALEFYPG VSQASlDQSRQMKYQSFVEYRKRFMLKPYESFEELTGEKEMSAELEALYGDIDAVELYPA 550 560 570 580 590 600 l l l i l . LLLEKCHPVSIFGESMIEIGAPFSLKGLLGNPICSPEYWKPSTFGGEVGFNIVKTATLKK LLVEKPRPDAIFGETMVEVGAPFSLKGLMGNVICSPAYWKPSTFGGEVGFQIINTASIQS 610 620 630 640 650 l l l l l LVCLVTKTCPYVSFRVPDA SQDDGPAVERP——STEL LICNVVKGCPFTSFSVPDPELIKTVTIVASSSRSGLDDIVPTVLLKERSTEL Figure 2. Sequence alignment of Human COX-1 and -2 COX-1 lSignal leer [MBD I Catalytic Domain COX-2 I Signal leer [MBD 1 Catalytic Domain Ilsaa I 566FSVPDPELIKTVTINimlid lollir‘llrl'a'l litlsRSTEL‘S‘” Human FNVQDPQPTKTATIN , :_ l l rm 1 ‘\ HKRRSTEL Mouse FSVQDAHLTKTV'HN .: «all :i-I"~'l“il l l'lLERSTEL Ovine FNVQDPQATKTATIN \f‘< iii 3‘ l t'Lll-:.RRSTEL Rat Figure 3. Domain Structures of COX-1 and COX-2. Both COX isoforms have N- terminal signal peptides, EGF-like domains, membrane binding domains and catalytic domains. The most significant difference in primary structure between the two is the presence of a unique lS-residue cassette near the C-terminal end of COX-2. acids and share 60%—65% amino acid sequence homology (9). The major sequence differences between COX isoforms occur in the membrane binding domains and at the carboxyl termini, where a unique 18—amino acid sequence is present at the C terminus of COX-2 and not in COX-1. The function of this 18-amino acid is the topic of these studies (Fig. 3). COX-1 and COX-2 are encoded by separate genes, Pigs-1 and Pigs-2, respectively. While the proteins share similar structures and kinetic properties, COX-1 and COX-2 have different expression pattern and are involved in different physiological processes. COX-1 is constitutively expressed in most tissues and is thought to produce prostaglandins that serve housekeeping functions (10. l 1). COX- 2, however, is not expressed in most unstimulated tissues but is rapidly induced by growth factors, cytokines, and tumor promoters (12, I3). Prostaglandins produced by COX-2 regulate inflammation, differentiation, mitogenesis and angiogenesis (14). Reaction Mechanism of Cyclooxygenase The reaction mechanism of cyclooxygenase is shown in Fig 4. The peroxidase activity is required for cyclooxygenase activity; however peroxidase activity itself is independent. For the first step of the reaction, peroxide reacts with the heme group. Then a two-electron oxidation occurs, forming compound I and alcohol (15-18) (Fig. 4). Compound I can be rearranged by a single electron oxidation of Tyr385 to form intermediate II which contain an oxyferryl group (Fe(IV)=O), a neutral protoporphyrin IX, and a Tyr385 tyrosyl radical which is required for cyclooxygenase activity (15, 16, I9, 20). As mentioned before, compound I can continue to cycle through the peroxidase reaction independently from cyclooxygenase turnover (21). When the cyclooxygenase site is occupied by a fatty acid substrate, intermediate II can abstract the hydrogen atom from C13 in arachidonate to form an radical (22) (Fig. 1 and 4). The fatty acid radical then reacts with molecular O: to produce an ll-hydroperoxyl radical which in turn cyclizes to form a C ll-Co endoperoxide bridge. A second 0; molecule adds at C .5 to produce PGGg. PGG; is reduced to PGHz by the peroxidase activity of COX. POX suicide 0H ROH / inactivation :7 i441“? iE'FQ / Compoundl Resting enzyme Tyr-385 Tyr-385 POX \vze m\ o 0“ Pot;2 \\ OzAAOO‘ . . “F335 lntermediated ll ' / Tyr-385 ‘\ 0 OH / AA. COX l 202 AA. cox suicide inactivation Tyr—385 Figure 4. 3Cyclooxygenase and peroxidase catalysis and suicide inactivation of COXs. Fe3 PPIX, ferric iron protoporphyrin IX (heme); ROOH, alkyl hydroperoxide; ROH, alcohol, AA, arachidonic acid, Fe4 =PPD(, oxyferryl heme. Compound I can form intermediate 11 or alternatively undergo a one electron reduction by an exogenous electron donor, yielding compound 11. Intermediate 111 is a spectral intermediate of unknown structure but perhaps involving a heme group with a protein radical located on an amino acid side chain other than Tyr3 85. One interesting point is that both the cyclooxygenase or peroxidase activity of COX-1 and -2 rapidly catalytically inactivate in the presence of sufficient substrate. One probable mechanism for this suicide inactivation is shown in Figure 4. It involves the production of a Tyrosine38 ' radical (intermediate 11) (23) and migration of the radical on Tyr385 to another tyrosine within the protein that leads to protein damage by side chain cross linking or some other poorly understood mechanism. Consistent with this idea are findings that protein tyrosyl radicals in oPGHS-l are observed on tyrosines other than 'l‘yr385 (24n26), and that an intermediate III has been detected whose time course of production parallels peroxidase inactivation (23). It should be mentioned that the rates of both peroxidase and cyclooxygenase inactivation can be noticeably slowed by peroxidase-reducing cosubstrates (23, 27, 28). Reducing cosubstrates may increase the efficiency of conversion of intermediate II to compound II, and reduce production of intermediate Ill. Suicide inactivation is an interesting chemical phenomenon whose biological relevance is unclear. Crystal Structures of Cyclooxygenase Crystal structures of COX-1 and COX-2 have been determined by the Garavito and Luong laboratories (29, 30) (Fig. 5) and are very similar. In his initial report on the structure of COX-1, Garavito defined separate and distinct protein domains involved in dimerization, membrane binding and catalysis (29). The C-terminal tails are not visible in the crystal structures, presumably because of their flexibility. COX-1 and COX-2 dimers are held together via hydrophobic interactions, hydrogen bonding, and salt bridges between the dimerization domains of each monomer. Three disulfide bonds hold the EGF domain together. The catalysis domain of COX enzymes includes a cycloxygenase and a peroxidase active site. For the two-step conversion of arachidonate to prostaglandin H2, the first step takes place in the cyclooxygenase active site to form the peroxide intermediate, prostaglandin GZ. PGG; then next diffuses out of the cyclooxygenase site to the peroxidase active site on the opposite side of the protein, where it is reduced to prostaglandin H2. The COX enzymes have an unusual and unique mechanism of interaction with the membrane (Fig. 5). Figure 5. Structure of ovine prostaglandin endoperoxidase H synthase-l (oCOX- l) (Picot et aL, 1994). (A) oCOX-l homodimer associates with the luminal face of the ER membrane. Three major folding domains are: epidermal growth factor domain for dimerization (EGF; green), membrane binding domain (MBD; gold) and globular catalytic domain (blue) which contains peroxidase (red) and cyclooxygenase (yellow) active sites. (B) The oCOX-l monomer with the locations of the peroxidase (POX) and cyclooxygenase (COX) active sites and the EGF and membrane binding (MBD) domains. The color scheme is the same as in (A). Although these enzymes are integral membrane proteins, they do not have any transmembrane sequences; instead four amphipathic helices form a hydrophobic surface that anchors these enzymes on the membrane. These amphipathic helices not only form the base of the molecule but they also form the entrance to the cyclooxygenase active site, which is a hydrophobic pocket that projects inward from the membrane surface of the enzyme. These helices interact at the membrane solvent interface. It is likely that fatty acid substrates and NSAIDs (Non-steroidal anti- inflammatory drugs) also partition to this interfaced region where they are accessible to the cyclooxygenase active site (30). Cyclooxygenase Inhibitors Inhibitors of cyclooxygenase belong to the class of drugs referred to as non- steroidal anti-inflammatory drugs (NSAIDs). All NSAIDs inhibit arachidonate binding to the cyclooxygenase active site of COX-1 and 2. Non selective NSAIDs inhibit both COX-1 and -2, while most selective inhibitor preferentially inhibits COX- 2. NSAIDS use can have adverse side effects including peptic ulceration and dyspepsia. These side effects are believed to be due to the inhibition of prostaglandin synthesis by COX-l, which produces prostaglandins that mediate protective reaction in the gastrointestinal mucosa. Drugs that selectively inhibit COX-2 have reduced gastrointestinal toxicities. Celecoxib and rofecoxib are the two COX-2 selective NSAIDs. To understand the mechanism for nonselective and selective inhibition of COX enzymes, it is necessary to examine how arachidonate and nonselective NSAle bind within the active sites of COX-1 and COX-2. Arg'20 is one of the few charged amino acids at the cyclooxygenase active site of COX. Crystal structure indicates that the guanidinium group of Arg'20 can form a salt bond with the carboxylic moiety of arachidonate (31). The carboxyl of arachidonate anchored near the mouth of the hydrophobic pocket and the hydrophobic tail of arachidonate insert into the hydrophobic pocket, forming a hairpin turn between carbons 9 and ll (Fig. 6). This orientation allows for the addition of two molecules of oxygen at carbon 9, 11 and 15 resulting in the formation of prostaglandin C13. Evidence from mutagenesis has also shown that Arg'20 is required for COX-1 activity and for the binding and inhibition by acidic cyclooxygenase inhibitors, which constitute the largest group of nonselective NSAIDs (32) (Fig. 7). Interestingly, nonacidic (COX-2 selective) NSAIDs, such as Dup697 and L-746, inhibit the R120E COX-1 mutant about 10 times more efficiently than the native cox-1, suggesting that Arg‘20 may actually discourage binding of He 523 Figure 6. Hypothetical orientation of arachidonate within the substrate binding pocket of ovine COX-l (R. Kurumbail, Second International Workshop on COX-2, 1998). Bind' goof arachidonate within the COX-1 active site is dependent on coordination with Argl (green). nonacidic inhibitors in COX-l (33). In contrast to its function in COX-l, Arg120 plays only an accessory role for NSAIDs and fatty acid binding in COX-2 (33, 34). The COX-2 isozyme also has a larger cyclooxygenase pocket than COX-l (30). This increased size allows fatty acids and inhibitors to more readily to access to the COX-2 active site, which diminishes the relative importance of ionic interactions with Arg'zo. Another effect of the larger COX-2 active site is that steric crowding causing by Arg'30 at the entrance of the pocket may be reduced and thereby increases the access N02 1, R = cyclohexyl 3, R = phenyt Figure 7. Structure of representative nonselective and COX-2 selective NSAIDs NS-398(l), DuP 697(2), nimesulide(3), celecoxib(4), rofecoxib(5), Flurbiprofen (6). of nonacidic inhibitors in COX-2. Therefore Argm may play an indirect role in the drug selectivity by discriminating against binding of nonacidic NSAIDs more efi'ectively in the COX-1 binding site than in the larger COX-2 binding site. Besides the Argmresidue, several other amino acids in cyclooxygenase binding site influence the volume and chemical environment of the COX-2 active site, and contribute to the selective inhibition. The subsituition of visitinc’23 in cox-2 with an Isoleucine in COX-l (35, 36) produced a much larger binding site in COX-2 than in COX-1 (Fig. 8) because of the shorten side chain of Valine. This amino acid contacts directly with the inhibitors. Substitution of Phenlyalanines03 with Leucine in COX-2 has added effect of allowing to a larger binding pocket, filled by most COX-2 inhibitor (36, 37). Figure 8. Superposition of COX-1 (yellow) to COX-2 (purple) around SC-558 (Kurumbail et al., 1996). The larger binding pocket is COX-2 is clearly visible. Regulation of Cyclooxygenase Gene Expression Transcription Regulation Because COX-1 is expressed constitutively in most tissues, and expression levels of this enzyme do not vary greatly in adult animals, it has been difficult to study transcriptional regulation of the COX-1 gene. The COX-l gene has a TATA- less promoter that contains multiple start sites for transcription (38). Gel shift assays have demonstrated that Spl cis-regulatory elements in the human COX-l promoter, at positions -1 l l/-105 and-610/-604, bind the trans-activating Spl protein (39). Deletion of either site leads to a reduction of about 50% in basal transcription, and deletion of both sites results in a reduction of about 75% (3 8). To date these Spl sites are the only cis-acting elements documented to regulate transcription of COX-1. Although numerous possible regulatory elements have been identified in the COX-2 genes, only five have been rigorously demonstrated to regulate transcription: overlapping E-box and ATF/CRE sequences near the TATA box, an NF/IL6 CAAT enhancer binding sequence upstream, and two NF KB binding sites at -214 and 427 (Fig. 10). The most critical of these regulatory sequences is the ATF/CRE, a regulatory element that typically is activated by hetero- and homodimers of the c-Fos, c-Jun, and ATF families of bZIP proteins (AP-l) (40), and the CAMP regulatory binding protein (CREB). The ATF/CRE and NF/IL6 regulatory elements will also cooperate to activate the COX-2 expression in human epithelial cells (41). The E-box is required for hormonal regulation of COX-2 in rat granulosa cells (42) and for the elevated expression of COX-2 in the murine carcinoma cell line JWF2 (43). NFKB signaling has been implicated variously in the expression of COX-2 stimulated by TNFCX, hypoxia, endothelin, and lL—lB in osteoblastic cells ( l2), synoviacytes (44, 45), I3 epithelial cells (46, 47), endothelial cells (48), and hepatocytes (49). Each of these . effectors, as well as LPS, can activate the NF KB signaling pathway (50, 51). Another important signal pathway: the MAPK pathways have been shown to contribute to the increased expression of COX-2 in one or more cultured cell systems in response to inflammatory stimuli, including IL-lB, TNFOL, and LPS, as well as the phorbol ester TPA (52, 53). Dependence on kinase signaling for COX-2 expression has been demonstrated by overexpressing active kinases, or conversely by using dominant negative mutant kinases (54-56), and by using small molecule inhibitors that selectively block one or more of the ERK1/2, JNK/SAPK, and p38/RK/Mpk2 pathways (56-65). 445/447 -270/-256 -223/-214 -132/-124 -58/-53 -55/-50 -3l/-25 Start NF KB Spl NFKB NF/lL6 ATF/CRE E Box TATA BOX Figure 9. Regulatory elements in the human COX-2 promoter (Y amamoto, 1995) Post-transcriptional Regglation COX-2 mRNA has a very short half life due to the presence of multiple copies of the AUUUA motif within the 3’-UTR of COX-2 mRNA that are known to regulate mRNA stability (66, 67). Deletion of these motifs in COX-2 stabilizes the transcript (66, 68). COX-1 mRNA lacks these AU-rich motif (ARES) and is very stable (68). The mechanism by which ARES promote mRNA degradation in mammalian cells is not clearly understood. There is general agreement that ARES promote deadenylation of the polyA tail which precedes 5’ to 3’ or 3’ to 5’ exonuclease cleavage of the mRNA (69-72). The ARE binding protein(s) responsible for initiating ARE-mediated decay of COX-2 mRNA are yet to be identified. However many proteins have been found that can bind to the COX-2 3’-UTR and stabilize the transcript. For example, HuR- mediated COX-2 transcript stabilization has been reported in colon cancer cells where COX-2 is aberrantly over-expressed (73). Cyclooxygenase protein turnover Recent data from the Smith’s laboratory has indicated that degradation of COX- 2 occurs via the ER-associated degradation (ERAD) system in 293 cells (74). Kifnusenin, an inhibitor of terminal fucosidases stabilizes the protein (74). Consistent with this idea, Rockwell et al., has observed the accumulation of COX-2 in its native form and as polyubiquitin conjugates in HT4 neuronal-like cells treated with inhibitors of proteasomal degradation, while COX-l protein levels were unchanged by this treatment (75). These experimental results suggest that COX-2 may be selectively regulated by the ubiquitin-proteasome pathway, which has been implicated in the degradation of intracellular proteins with short half-lives. Rapid turnover of COX-2 in tissues where the enzyme is transiently expressed may serve a significant physiological role in regulating the levels of prostanoids whose synthesis is attributed to this COX isoforrn. Different Biological Activity of Cyclooxygenases COX-1 and COX-2 are involved in different physiological processes. Mouse knock out studies clearly Show primary roles for COX-1 in platelet aggregation (76) and parturition (77, 78), and for COX-2 in ovulation, implantation (79), and neonatal development (80, 81). COX-2 knock out mice have multiple defects in ovulation, implantation, and decidualization (82); in contrast COX-l knock-out mice have lengthened bleeding time and difficulty with parturition (83). To date, most studies have focused the role of COX-2 in carcinogenesis, less well studied are the role of COX—1 (84, 85). Increased COX-2 expression is sufficient to cause formation of breast tumors in transgenic mice (84). Non-steroidal anti-inflammatory drugs (NSAIDS) reduce carcinogen-induced mammary tumors in rats (85). In Apc‘s761 knockout mice, a model for inherited colorectal cancer (FAP), genetic disruption of COX-2, or NSAIDS that selectively inhibit COX-2, suppress polyp formation (86). These studies confirm an obligate role for COX-2 overexpression in the transformation process. It is not clear how COX-2 facilitates transformation. COX-2‘s roles in oncogenesis have variously been attributed to the stimulatory effects of prostaglandins on cell growth (87-89), to inhibition of apoptosis (90), and to activation of the nuclear PPAR receptors (91-93). However, deficiency in either COX-l or COX-2 reduces polyp formation in Min'“ mice almost equally (94, 95). So an unanswered question is whether these two contribute to carcinogenesis via the same or different mechanisms. One plausible explanation for the COX isozymes roles in colon cancer is that COX-1 provides prostaglandins that protect carcinogen initiated stem cells from DNA damage-induced apoptosis, whereas COX-2 promotes transformation after loss of heterozygosity of the Apc gene. There has been no good explanation for the apparent redundancy of COX-1 and 2. These two enzymes are so similar that it is difficult to explain why their biological activities are different. Difference in regulation, protein turnover, and the subtle catalytic differences towards arachidonic acid favors COX-2 at low substrate levels may be important, but other factors may also be involved. My thesis will attempt to investigate the biochemical mechanism for the unique biological activity of COX-2. CHAPTER II PUTATIVE PROTEIN PARTNERS FOR CYCLOOXYGENASE PROTEINS Summary There has been no good explanation for the apparent redundancy of COX-1 and 2. These two enzymes are so Similar that it is difficult to explain why their biological activities are different. The COX-2 protein has a unique l8-amino acid cassette located four residues from the carboxyl terminus of all COX-2 Species that is not found in COX-l. Experiments in our lab have demonstrated that deletion of this 18 amino acid sequence does not affect catalytic activity. The hypothesis that I investigated was that this 18-amino acid cassette iS responsible in part for the distinct biological activity of COX-2 by mediating protein-protein interactions. Two proteins were identified by proteomic analyses that reproducibly co-purified with Flag-tagged COX-2 in 293 cells: FAM44A (GI/Q8NFC6) and Heat shock-induced protein (GI/188492). FAM44A was identified in 5 out of the 7 experiments in protein complexes with COX-2 purified from T-Rex 293 cells, and l in 2 experiments where COX-2 was expressed transiently in Freestyle 293 cells. FAM44A protein was also identified once with COX-l ins580-598, which provided further evidence for its specific interaction with the COX-2 18-amino acid cassette. FAM44A is a 330kDa protein that has been identified previously from cDNA data. FAM44A can be phosphoylated during DNA damage. The interaction between COX-2 and F AM44A could be involved in DNA damage induced cell cycle checkpoint control. The heat shock-induced protein was only identified in 1 out of the 7 experiments in protein complexes with COX-2 purified from T-Rex 293 cells, and l in 2 experiments where COX-2 was expressed transiently in Freestyle 293 cells. It may function as a chaperone in the process of COX-2 or be involved in its degradation. Kifnusenin, an inhibitor of terminal fucosidases which increases the stability of COX-2, was applied to 293 T-Rex cells expressing Flag-tagged COX-2 in an attempt to trap protein partners of COX-2 that might be involved in protein degradation. No new unique proteins were identified even with this treatment. No noteworthy protein partners were identified for COX-2 de1581-598 or COX-1. Other COX-2 protein partners may exist that were missed due to the fact that their interactions are transient, or have low affinity and do not survive purification. Furthermore untransfected 293 cells do not express endogenous COX-2, and may not also express the normal protein partners for COX-2. Introduction That protein-protein interactions may modify the activity of COX-2 is a speculative hypothesis, but there is circumstantial evidence to support its validity. The position of the 18 amino acid insert near the carboxyl terminal would be ideal to allow interactions with other proteins. The carboxyl terminus of COX-2 cannot be observed in crystal structures; usually an indication of regions that are very flexible, suggesting that the COX-2 tail would be free to interact with other proteins. In addition, cytokine-stimulated prostaglandin synthesis proceeds selectively via newly expressed COX-2 in many cell systems, even when COX-1 is present. Such synthesis is often channeled through a coordinately-expressed membrane associated PGE: synthase (96- 99). The mechanism whereby COX-2 selectively provides PGH: to prostaglandin synthases may be via protein-protein interactions of COX-2 with a channel protein. 5- 19 lipoxygenase, the enzyme that catalyzes the initial step in leukotriene biosynthesis requires the 5-lipoxygenase activating protein (FLAP) for the arachidonate substrate (100, 101). Nuclear receptors, PPARs, may also mediate the biological effects of COX-2-derived prostaglandins by direct regulation of gene expression by prostanoids (91-93). Nevertheless, such a mechanism requires the nuclear transport of prostaglandins. Since no cyclooxygenase is present in the nucleus, accessory proteins would be needed to import prostaglandins. Nitric oxide produced by inducible NO synthase (iNOS) and prostaglandins generated by COX-2 have been reported to be involved in inflammation (102, 103). Since the inducible expression of the two enzymes by inflammatory stimuli has the similar time course, suggests that the two systems may interact. Recently direct interaction of iNOS and COX-2 has been revealed (104). Endogenous COX-2 and iNOS can be induced by LPS in RAW264.7 cells and co-immunoprecipitation confirmed their interaction. GST pull down assay carried out in HEK293 cells that transiently co-expressing iNOS and COX-2 also proved the direct binding between these two proteins. Mutagenesis showed that the binding occurred at the C-terminal domain of COX-2 which was in accordance with our hypothesis. The interaction seems to bring NO in proximity to COX-2, S-nitrosylate COX-2 and enhance the catalytic activity of COX-2. The molecular synergism between iNOS and COX-2 may represent a major mechanism of inflammatory responses. However iNOS is soluble and found predominantly in the cytosol while COX-2 is localized to the ER and nuclear membrane. In order to further confirm the interaction, their subcellular localization should be examined to verify that they can localize together. It has been demonstrated before that COX—2 may be degraded by the ubiquitin proteasome system (UPS) (74, 75). However, how ubiquitination is accomplished and regulated was unclear. An important regulator of the UPS is the COP9 signalosome (CSN), which controls the stability of many proteins. Recent data in Dubiel’s lab have confirmed that COX-2 can physically interact with the CSN using density gradient centrifugation and immunoprecipitation (105). Pull down experiments with Flag- tagged COX-2 revealed that COX-2 was associated with large complexes consisting of the CSN, cullin-RING Ub ligases and the 268 proteasome which further proved the proteasome-dependent degradation of COX-2 in HeLa cell lysate. Our understanding of the different biological roles of COX-1 and COX-2 is only beginning to emerge. The functions of these two isozymes in apoptosis, particularly as it relates to the development of a variety of cancers (e.g. colon, lung, and breast), angiogenesis, respiration, inflammation, pain, and reproduction are currently being studied in cultured cells and whole animals. If specific protein interactions were identified, this would provide a more complete explanation for COX-2 signal transduction and its role in transformation. In addition, a better understanding of signaling through the COX pathway could also lead to more specific chemoprevention for a variety of cancers. Experimental Procedures Materials. FreeStyleTM 293 cell line and F lp-ln T-Rex 293 cell line were from Invitrogen. Dulbecco’s modified Eagle medium (DMEM), Lipofectamine 2000, cellfectin, tetracycline, blasticidin, zeocin, hygromycin, and penicillin-streptomycin were from Invitrogen/Gibco. Fetal bovine serum (PBS) was purchased from Atlas Biologicals. Cycloheximide and puromycin were obtained from Sigma-Aldrich. Kifnusenin was purchased from Calbiochem. 21 Construction of the F lau_g;_tagged COX-2 a__r_1_d COX-2 de1581-598 in the [$133139 vector for Infection. FLAG (DYKDDDDK) and TEV (CENLYFQG) sequences were inserted into the N-terminus of human (h) COX-2 cloned into the pFastBac vector (Appendix, Fig. 22). F lag-tagged COX-2-pFastbac was used as a template to make the deletion mutant (ASSSRSGLDDINPTVLLK) by PCR mediated mutation. The Flag tag binds to the anti-F lag peptide, which when coupled to agarose beads, can be used for the affinity purification; TEV (Tobacco Etch Virus) sites are recognized and cleaved by the TEV protease which allows protease-mediated elution from the Flag affinity column. The plasmids were used to produce baculoviruses, which were used to infect Sf21 insect cells. High levels of expression in insect cells allowed us to confirm that if the deletion and the insertion of the tag affected hCOX-2 activity. The primers, vector and PCR conditions used to design the above mutants are shown in the ‘Appendix’ section (Table 7). Construction of Flag-tagged COX-1 and COX-1 in3580-598 in the pIND vector. The histidine-tagged ovine (o) COX-1 and COX-1 ins580-598 subcloned into the pIND vector at Hind 111 site were a gift from the laboratory of Dr. William Smith. FLAG (DYKDDDDK) sites were inserted by PCR amplification into the N-terminus of ovine COX-1 to replace the His tag (HHHHHH) in the pIND vector using specific primers (Appendix, Table 7) (Appendix, Fig. 23). Construction of LCDNA expression plasmid_§ for the Flag-tagged COX-2. COX-2 de1581-598, COX-2 N580A, COX-2 Y371F, COX-1 and COX-1 ins580-598. Flag-tagged COX-2 and COX-2 delS81-598 were subcloned from pFastbac into a pcDNAS/FRT/TO vector (Appendix, Fig. 24). COX-1 and COX-1 insS80-598 were also subcloned from pIND into a pcDNAS/FRT/TO vector (Appendix, Table 8). N580A is a mutation that removes a glycosylation site from the 18 amino acid insert f») I») of COX-2 (Table 1a) (106). The Y371F mutant does not have cyclooxygenase activity because the tyrosine involved in the initial step in hydrogen abstraction has been changed to phenylalanine (Table 1a) (107). All the plasmids were used for stable transfection in F lp-In 293 cell lines. Construction of Fngiagged COX-2 in the pOSML vector. Flag-tagged COX-2 in pcDNAS/FRT/TO was directly subcloned into the Not I site of the pOSML vector (Appendix, Fig. 25) (Appendix, Table 8) and was used for the transient transfection of 293 Freestyle cells. Cell culture. Frozen Sf21 cells were stored in liquid nitrogen until ready to use. Frozen vials were thawed in a 37°C water bath and transferred into a 125 mL shake flask containing 27 mL of pre-warmed Sf-900 II SFM, and incubated in a 27°C :t 05°C non-humidified, ambient air-regulated incubator in flasks. Sf21 cells were grown in shake flasks to Z 2 x 106 viable cells/ml, and infected at an M.O.I of 1:1, with baculovirus constructs. The cells were then grown for 72 h and harvested. The Flp-In T-Rex -293 cell line constitutively expresses the lacZ-zeocin fusion gene, the Tet repressor and contains a single integrated Flp Recombination Target (F RT) site. The cell line can be used to generate tetracycline-inducible cell lines with high frequency by co-transfecting the pcDNAS/FRT/TO expression vector containing a gene of interest together with the Flp recombinase expression plasmid, pOG44 (Appendix, Fig. 26). Flp recombinase mediates insertion of the pcDNAS/FRT/TO expression construct into the genome at the FRT integration site through site-specific DNA recombination (Appendix, Fig. 27). Following the transfection, the Flp-1n- TREx-293 expression clones should become sensitive to zeocin and should be selected with 100ug/ml hygromycin B to generate a stable cell line. Expression of the transfected gene can be induced with tetracycline. Flp-In. T-REXi-293 cells (Invitrogen) were cultured in 100mm plates containing 10 ml of complete medium (DME high-glucose medium containing 10% fetal bovine serum, with 1% Pen-Strep containing lOOug/ml zeocin and lSug/ml blasticidin) at 37°C with 6% C02. The cells were subcultured when 80-90% confluent (2-5 days). All medium was removed and the cells were washed once with 10 ml PBS. One ml 0.25% trypsin in versene (0.14 M NaCl, 1.5 mM KH2PO4, 2.7 mM KCl, 8.0 mM Na2HPO4, 0.5 mM EDTA) solution was then added and the cells were incubated for 1-5 minutes at room temperature until the cells detached. Nine ml of complete medium was then added and the solution was pipetted up and down to break up clumps of cells. One ml of the 10 m1 cell suspension was transferred to a new 100mm plate and 10 ml fresh, complete medium containing zeocin. and blasticidin was added. The HEK 293 cell line is a permanent line established from primary embryonic human kidney transformed with sheared human adenovirus type 5 DNA. The ElA adenovirus gene is expressed in these cells and participates in transactivation of some viral promoters, allowing these cells to produce very high levels of protein. The FreeStyleTM 293 cell line is a variant of the 293 cell line that has been adapted to suspension growth in FreeStyleTM 293 Expression Medium. Cells were subcultured at approximately 1-1.5 x106 viable cells/ml, and were subcultured approximately every 48-72 hrs into new shaker flasks at 0.1-0.2 x 106 viable cells/ml. Flasks were incubated in a 37°C incubator containing a humidified atmosphere of 6% C02 in air on an orbital shaker platform rotating at 135 rpm. Isolation viruses from insect Sf21 cells with F lag-tagged COX-Z-QFastBac and COX-2 d81581-598-9FastBac. The BAC-to-BAC Baculovirus Expression System from Invitrogen was used to construct viruses to express the COX mutants in insect cells (Sf21, Spodoptera frugiperda). The pFastBac recombinants were first 24 transformed into MAX efficiency DHIOBac cells, and the recombinant bacmid DNA was isolated. 9 x 105 Sf 21 cells were plated in 2 ml of Sf-900 II SFM containing 50unit/ml penicillin and 50ug/ml streptomycin. Five ul of baculovirus DNA was diluted into 100 pl Sf-900 II SFM without antibiotics. Cellfectin was mixed and 6 ul was added separately into 100 pl of Sf-900 II SFM without antibiotics. The diluted DNA was combined with diluted Cellfectin, mixed gently and incubated for 15-45 minutes at room temperature. The growth medium was removed and the cells were rinsed with Sf-900 II SFM without antibiotics. The wash medium was removed and 0.8 ml of Sf- 900 II SFM was added to the DNA-Cellfectin complex, mixed gently and added to the cells. Cells were incubated at 27°C for 5 hrs. The transfection mixture was removed and replaced with 2 ml of Sf-900 II SFM containing antibiotics. Cells were incubated at 27°C. After three days, the recombinant baculoviruses were harvested and used to infect S0] cells. These cells were harvested after three or four days by centrifugation at 5000 rpm for 10 minutes (Beckman TJ-6 centrifirge). The virus-containing supernatant was saved for amplification and the cell pellet was examined for protein expression. Cyclooxygenase Assay Using the 07 Electrode. Cyclooxygenase activity was . measured polarmetrically as described previously (108). Cuvettes used to measure 0; consumption were loaded with 3m] 0.1M Tris-HCl pH 8.0 containing 1001.11 2mg/ml arachidonic acid, 1mM Phenol and 25p] 3.4mg/ml hemoglobin. The reaction was initiated by injecting homogenates of baculovirus-infected Sf21 cells. One unit of activity is the amount of enzyme required to convert lnmole 03/ min at 37°C under the assay conditions described. 25 Stable transfection of different forms of cyclooxygenase in 293 Flp-In T-Rex m pCDNAS/FRT/TO plasmids containing for the COX mutants were cotransfected with the F lp-recombinase expression vector pOG44 into Flp-In 293 cells (Invitrogen). One day before transfection, a 60mm plate was plated with 1 x 106 cells/4 ml in growth medium without antibiotics. Cells reached 90-95% confluence by the time of transfection. For each transfection, 8 ug DNA (7.2pg COX-pCDNA: 0.8pg pOG44) was diluted into 0.5 ml of Opti-MEM® Medium without serum and mixed gently. LipofectamineTM 2000 (20 ul) was gently mixed in 0.5 ml Opti-MEM® I Medium and incubated for 5 minutes at room temperature. After 5 minutes incubation, the diluted DNA and diluted LipofectamineTM 2000 were combined, mixed gently and incubated for 20 minutes at room temperature. The mixture was then added to a plate containing cells and medium and mixed gently by rocking the plate back and forth. Cells were incubated at 37°C in a CO: incubator for 24 hrs, and then the old medium was removed and fresh medium was added. After 48 hrs, cells 1 were divided 1:5 into new plates with fresh growth medium. Selective medium containing 15ug/m1 blasticidin and 100ug/ml hygromycin B was added after 2-3 hrs. Clones formed within two weeks and individual colonies were isolated with sterile Q— tips (Fisher). Cell lines were tested for inducible expression of COX proteins by treating the selected colony cells with lug/m1 tetracycline for 24 hrs, and followed by Western blot analysis. Transient tansfection of Flag-tagged COX-2 into FreeSgle 293 Suspension Cultures. Twenty four hrs before transfection, FreeStyleTM293 cells were subcultured to 6~7 x 105 cells/ml in fresh FreeStyleTM 293 Expression Medium on an orbital shaker platform rotating at 135 rpm at 37°C at 6% C02. On the day of transfection, the cell density is 1.2-1.5 x 106/m1. Twenty five ml of cells were added into each 125 26 ml shaker flask, and plasmid DNA (62.5 ug) was diluted into 150mM NaCl to a total volume of 1.25 ml. In a separate tube, 125 pl of polyethylenimine (PEI, 1mg/ml) (Polysciences, Inc) was diluted into 150mM NaCl to a total volume of 1.25 ml and mixed gently by inverting the tube. The plasmid DNA and PEI were immediately mixed and vortexed for 10 seconds. The DNA mixture was then incubated for 10 minutes at room temperature to allow complexes to form before adding to the cells. Transfected cells were incubated for 5 hrs at 37°C in 6% CO; atmosphere on an orbital shaker platform rotating at 135 rpm, and then 25ml of fresh FreeStyleTM 293 Expression Medium was added. Cells were harvested after 24 to 48 hrs and protein expression was examined by Western blot analysis. Purification. COX-1 and COX-2 protein complexes were purified from stably- transfected 293 T-Rex cell lines that had been induced with tetracycline or from transiently transfected FreeStyle 293 cells. Cells were washed once with PBS, collected by centrifugation and homogenized in 0.5 m1 lysis buffer (50mM Tris HCl, pH 7.4, 150 mM NaCl, 1mM EDTA) containing a cocktail of protease inhibitors (Complete-Mini Protease Inhibitors, Roche). The lysates were sonicated twice for 20 sec, and 1% Tween 20 was added. The homogenate was centrifuged at 3000>< g for 15 min, and the supematants were clarified by passing them through a 0.22 um filter to remove particulate matter. Cleared cellular lysates were incubated with anti-Flag M2 agarose gel (Sigma) (10w per 100mm plate) previously equilibrated in TBS (50mM Tris HCl, pH 7.4, 150 mM NaCl) with gentle mixing for 90 min. The resin was collected by centrifugation at 1000>< g for 5 min and was washed with TBS buffer twice for 10 min. Flag-COX and associates proteins were eluted from the resin by incubation with 300 ng/ul 3xF lag peptide (Sigma) for 60 min at 4°C in lml TBS or 10 unit TEV protease (Invitrogen) in 150w 1M Tris HCl (pH 8.0) with 10mM EDTA 27 overnight. In the initial experiments, beads were washed by Gly-HCl (pH 3.5) for 5 minutes to elute all the rest of the proteins that were attached in order to test the efficiency of elution. Coomassie Staining of SDS-PAGE gels. Electrophoresis on 4-12% Bis-Tris NuPAGE polyacrylamide gels (Invitrogen) was used to fractionate the proteins purified by FLAG affinity chromatography and Coomassie Blue Stain was used to visualize the isolated proteins. Coomassie Brilliant Blue R250 (0.25% w/w) was dissolved in 90% methanol with 10% glacial acetic acid. The gel was immersed in five volumes of staining solution and placed on a slowly rotating platform overnight. The gel was destained by soaking in a 40% methanol/10% acetic acid solution on a slowly rocking platform until the background stain was gone and bands were clearly visible. Western Blotting. Proteins samples for the Western blot were separated on a 4- 12% Bis-Tris NuPAGE gels and transferred to nitrocellulose membranes. Visualization was performed by incubating with anti-flag (1:3000 dilution) and anti- COX (1:2000 dilution) primary antibody in 5% milk TBST (TBS plus 1% (v/v) Tween-20) solution for two hrs. The membranes were then washed four times with TBST for 10 minutes. Membranes were next incubated with horseradish peroxidase- conjugated anti-mouse or anti-rabbit IgG antibodies in 5% milk TBST solution (1:2000 dilution) for one hr. After four 10 minutes washes with TBST, immunoreactive proteins were visualized using the Western Lighting Chemiluminiscent Kit (Perkin Elmer, Boston, MA) and exposure to X-ray film (Amersham). For the actin controls, the same blots were blocked a second time in 5% milk TBST overnight, and then incubated with anti-actin antibody (1:2000 dilution) in 28 TBST for two hs. The membrane was washed four times with TBST for 10 minutes, and the membranes were then incubated with horseradish peroxidase-conjugated anti- mouse IgG antibody in 5% milk TBST solution (1:2000 dilution) for one hr. After four 10 minutes washes with TBST, actin was visualized as described above. Mass spectrometry. Proteins isolated by a FLAG chromatography were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel (Bio-Rad). The gels were fixed with a 40% methanol/20% acetic acid solution for 2 hrs, and then stained with Coomassie blue overnight. The next day, the gel was destained with a 10% acetic acid solution. The protein band was then excised by using a razor blade and sliced into 1-2mm pieces, all of which were placed into a single Eppendorf tube. The gel pieces were washed with 100 111 100mM NH4HCO3 for 5 mins. The buffer was then replaced by 50111 of 100% acetonitrile and the gel pieces were dehydrated at room temperature for 15 minutes. Acetonitrile was then removed and the samples were dried completely in a Speedvac. The gel pieces were rehydrated with 50111 10mM DTT in 100mM NH4HCO3 at 56°C for 30 minutes to reduce the proteins. DTT was then replaced with 100% acetonitrile and the gel pieces were incubated at room temperature for 5 minutes. This step was repeated, and the acetonitrile was removed and the samples were dried again in a Speedvac. 50111 of 55mM Iodoacetic acid (1AA) in 100mM NH4HCO3 was added. The samples were kept in the dark for 20 minutes at room temperature. The supernatant was discarded and the samples were washed briefly with 50111 100mM NH4HC03. The samples were then washed with new 100mM NH4HCO3 for 15 minutes at room temperature. The liquid was decanted and 50111 100% acetonitrile was added at room temperature for 15 minutes. Next, the acetonitrile was removed and the samples were dried in a Speedvac. The samples were then rehydrated in 20111 digestion buffer (15ng/111 lyophilized trypsin in 50mM 29 NH4HCO3) and incubated for 45 minutes on ice. The excess digestion buffer was replaced with 20111 50mM NH4HC03 and incubated at 37°C overnight. The samples were then centrifuged at 15,000>< g for 5 minutes. The liquid was collected in a new 1.5m] tube and set aside. 20111 of 60% ACN/ 1% TFA was added to each gel piece and sonicated for 10 minutes. The gel pieces were centrifuged and the supernatant was added to the previous supernatant. The gel pieces were washed twice and all of the liquid was combined and dried by using a Speedvac to less than 2111. 18 111 MS buffer (1% Trifluoroacetic acid and 98% H20) was then added to the tubes and the tubes were sonicated for 5 minutes. The tryptic peptides were then injected onto a Paradigm Platinum Peptide Nanotrap (C18, 0.15 x 50mm). The bound peptides were eluted onto a 10 cm x 75 um New Objectives Picofrit column packed with Microm Magic C18 AQ packing material and eluted over 30 minutes with a flow rate of 250 nl/min and a gradient of 5% to 90% Acetonitrile, with constant 10% 1% formic acid in the first 24 minutes using a Michrom Paradigm liquid chromatography attached to a ThermoElectron LTQ Linear Ion trap mass spectrometer. The top five ions in each survey scan are then subjected to data-dependent low energy collision induced dissociation (CID). The resulting MS/MS spectra were converted to peak lists using BioWorks Browser v 3.2. All protein entries were downloaded from the National Center for Biotechnology Information web page (http://wwwncbinlmhih.gov/, downloaded 01/26/2006) and the peak lists were searched against this library using Mascot. Identifications were considered positive if 2 peptides per protein were identified with a significant Mascot score (p< 0.05). Results Expression of Flag-tagged COX2 and COX-2 de1581-598 in S121 cells. To determine if deletion of the 18 amino acid C-tenninal cassette and the insertion of FLAG-TEV sequence at the N-terminus of cyclooxygenase-2 affected its activity, and to optimize methods for the affinity purification of Flag-tagged COX protein using FLAG affinity chromatography, Flag-tagged COX-2 and COX-2 de1581-598 were expressed in S121 cells. Polarrnetric oxygen electrode assays of cyclooxygenase activity in cell crude lysates from individual infections demonstrated that each of these enzymes had activities similar to the untagged COX-2 N580A (a mutant form of COX-2 lacks the N580 C-terminal glycosylation site but has the similar activity as COX-2) (Table 1) (109). Neither the insertion of the FLAG-TEV sites or the deletion of the 18 amino acids affected the COX activity, validating the use of Flag-tagged COX-2 and COX-2 de1581-598 in the mammalian cell studies. Proteins Activity (11mole Ozlmin/mg sample) COX-2 N580A 248 FLAG-TEV-COXZ 2 1 5 FLAG-TEV-COX-Z de1581-598 223 SD] cells only 0 Table 1. Comparison of the expression level of different COX-2 mutants in 8121 insect cells. COX-2 N580A, a mutant form of COX—2 with similar activity as native COX-2 was used as a positive control. 31 Since our goal was to affinity-purify COX-2 complexes using the Flag-tagged COX-2, we first optimized the purification of Flag-tagged COX-2 from insect cells using anti-Flag agarose with 3X Flag peptide to elute COX-2. Western blotting with anti-Flag antibody and Coomassie staining was performed to check the efficiency of the purification (Fig 10a, b). Both Western blotting and Coomassie staining revealed bands migrated at approximately 72 kDa, which is the molecular weight of COX-2. No significant difference was observed in the migration of the Flag-tagged and un- tagged COX—2. Bands visible at molecular weights above 100 kDa are likely dimers and trimers of COX-2. The results showed that Flag-tagged COX-2 and COX-2 COX-2 trimer COX-2 dimer COX-2 Figure 10(a). Western blot of Flag-tagged COX-2 and COX-2 de1581-598 eluted with 3XFlag peptide. 15 pg of proteins were loaded onto each lane of a 10% NuPAGE Bis-Tris gel and proteins were separated , by electrophoresis. Afier transferring to nitrocellulose membrane, F lag-tagged COX-2 and COX-2 de1581-598 were detected by a specific anti-flag antibody. Lane1-3 are purification fractionation samples isolated from COX-2 N580A infected insect cells: eluted with 3X Flag peptide in pH 7.4 TBS, flow through and eluted by pH 3.5 Gly-HCl respectively (see purification in methods). COX-2 N580A doesn’t have a Flag-tag so no protein was observed in LaneI-3. Lanes 4-6 are purification fractionation samples isolated from Flag-tagged hCOX2 infected insect cells: eluted with 3X Flag peptide in pH 7.4 TBS, flow through and eluted by pH 3.5 Gly-HCl respectively. Lanes 7-9 are purification fractionation samples isolated from COX—2 delS81-598 infected insect cells: eluted with 3X Flag peptide in pH 7.4 TBS, flow through and eluted by pH 3.5 Gly-HCl respectively. l9lKD 97KD 64KD 51KD 39KD 28KD Figure 10(b). Coomassie Blue Stained PAGE gel of Flag-tagged COX-2 and COX-2 de1581-598 eluted with 3X Flag peptide. 15 pg samples were loaded onto each lane of a 10% NuPAGE Bis-Tris gel and proteins were separated by electrophoresis. Lane] is Marker. Lane2-4 are fractionation samples isolated from COX-2 N580A infected insect cells: eluted with 3X Flag peptide in pH 7.4 TBS, flow through and eluted by pH 3.5 Gly-HCl respectively (see purification in methods). COX-2 N580A doesn’t have a Flag-tag so protein was only observed in Lane 3. Lanes 5—7 are fractionation samples isolated from Flag-tagged COX-2 infected insect cells: eluted with 3X Flag peptide in pH 7.4 TBS, flow through and eluted by pH 3.5 Gly- HCl respectively. Lanes 8—10 are fractionation samples isolated from COX-2 de1581- 598 infected insect cells: eluted with 3X Flag peptide in pH 7.4 TBS, flow through and eluted by pH 3.5 Gly-HCl respectively. 1 9 1 KD 97KD v COX-2 64KD 5 l KD 39KD 28KD Figure 11. Coomassie Blue Stained PAGE gel of Flag-tagged COX-2 eluted with TEV protease. 25 pg of samples were loaded onto each lane of a 10% NuPAGE Bis- Tris gel and proteins were separated by electrophoresis. Lane] is Marker. Lanes 2-5 are fractionation samples isolated from COX-2 N580A infected insect cells: crude lysate, flow through, eluted with TEV protease and eluted by pH 3.5 Gly—HCl respectively (see purification in methods). Lanes 6-10 are fractionation samples isolated from Flag-tagged COX—2 infected insect cells: crude lysate, flow through, eluted with 251tg TEV protease, eluted with 3011g TEV protease and eluted by pH 3.5 Gly-HCl respectively. de158l-598 expressed to high levels in Sf21 cells and eluted efficiently from the anti- Flag agarose using 3X Flag peptide. Although the constructs contained TEV protease sites adjacent to the Flag-tag, no COX-2 elution was observed upon treatment with TEV protease. However, the COX-2 could be released at low pH using Glycine-HCI buffer (pH 3.5) (Fig 11). Longer incubations with TEV protease and increasing the 34 concentration of the protease didn’t result in any significant release of the COX-2. The Flag sites binding to the anti-Flag beads maybe sterically block the TEV site. Expression of Flag-tagged COXZfl COX-2 de1581-598 in MCF-7 cells. When trying to identify protein partners for COX-2, we chose cell lines that might express COX-2 under normal physiological conditions and for which COX-2 might have an important biological or pathological fiinction. Because of our interest in COX-2 for breast cancer, we chose the MCF-7 mammalian breast cancer-derived epithelial cell line. Previous studies in this laboratory were unable to constitutively express COX-2 stably in transfected cells. This was believed to be due to the effects of COX-2 on cell growth, so inducible systems were evaluated. The AP-inducible expression system (ARIAD Pharmaceuticals) was the first system we attempted to use to construct an inducible COX-2 expressing cell line. This system uses a chemical inducer (AP) to link the activation and DNA binding domains of a bipartite transcription factor to form a functional transcription factor that stimulates transcription of a target gene from a synthetic promoter (Appendix, Fig 28a). AP (AP21967) is a chemically modified derivative of rapamycin that can be used to induce heterodimerization of engineered proteins. AP21967 has two separate motifs that can each bind with high affinity to the transcription activation and DNA binding protein modules (Appendix, Fig 28b). This system is one of the most tightly controlled regulated expression systems yet developed (110). Flag-tagged COX-2 and COX-2 de1581-598 were cloned into the target Ariad gene expression vector pLH-Zlgl-PL (Appendix, Fig 280) and transfected into a MCF- 7 cell line expressing the Ariad bipartite transcription factor (Appendix, Fig 28d). 35 Colony number Activity/100mm plate l 1.17 2 0.92 3 1 1.45 4 25.48 5 333.70 6 0.90 7 0.93 8 0.98 9 141.20 10 0.89 11 0.78 12 1107.00 13 12.15 14 0.06 15 0.96 16 0.71 17 0.91 18 1.29 19 388.60 20 1.06 21 1658.00 22 1.09 Cells only 0.34 Table 2. Luciferase Assay for individual colonies of stably transfected MC F-7 cells 36 This parent cell line was developed by the Gallo and Conrad laboratories at Michigan State University and has been used for expression of a F lag-tagged version of MLK-3 (111) and estrogen receptor (unpublished data). To verify that the system worked in our hands, luciferase was transfected as a positive control. Luciferase activity (Appendix, Fig 29) was tested according to the manufacturer’s protocol (Sigma). Significant inducible luciferase activity was detected in 8 of 22 colonies (Table 2) which verified the integrity of the parent cell line and our experimental technique. COX-2 expression was also screened in stable colonies by Western blot analysis. Surprisingly, no expression could be detected for Flag-tagged COX-2 or COX-2 del581-598 in any of the antibiotic resistant colonies. Numerous transfections were tested as well as several different COX-2 antibodies, all with negative results. There is no simple explanation as to why Flag-tagged COX-2 and COX-2 de1581-598 could not be expressed in the AP system. The high sensitivity of the luciferase assay may simply allow detection of lower level of luciferase compared to the Western detection of COX—2 in this cell system. As this cell system is very tightly regulated, it does not seem likely that leaky expression of COX-2 played a role in our inability to isolate protein-expressing colonies. Isolation of colonies th_at inducibliexpress cyclooxygenase in 293 Flp-In T- Rex cell system and identifying possible partners. Since we were unsuccessful using the AP system to express COX-2, we next tried the tetracycline-inducible 293 Flp-In T-Rex system. Colonies selected from different transfections were checked by Western blot for protein expression. Although the efficiency of recovering antibiotic- resistant colonies that expressed cyclooxygenase varied widely depending on the isozyme (COX-1 > COX-2) and the construct (Flag-tagged COX-2 < Flag-tagged COX-2 mutants), multiple cell lines that inducibly expressed Flag-tagged COX-2, 37 COX-2 de1581-598, COX-1 and COX-1 ins580-598 were successfully isolated (Fig 12) (See Chapter 111, Table 4). Respective colonies were grown (2 X 108 cells) and harvested 24 hr after tetracycline induction. A corresponding number of uninduced cells were isolated as negative controls. Protein homogenates from induced and uninduced cells were incubated with anti-Flag agarose beads to purify COX as described in the Methods section. The proteins isolated following the affinity purification were electrophoresed into the stacking gel of an 8-16% Tris—HCl SDS- PAGE gel (Bio-Rad), not to fiactionate the proteins, but to remove any detergent that might interference with mass spectral analysis. The entire protein band was then excised and the proteins were digested from the gel with trypsin and then analyzed by mass spectrometry. Proteins were identified from mass spectral data by analysis using the Mascot program using the NCBI-NR database. Several criteria were used to filter those proteins identified by Mascot that might be relevant protein partners for COX-2. COX-l COX-2 COX-l ins580-598 COX-2 delS8 1-598 - + - + Tetracycline l‘r' c .0! h‘a‘- - ': .‘f‘! _:Tf—r|—_-“"—9_.-.i—ll-zi .‘_: run—.1 5 will. ~T~COX isoforms r— "ah-“i fins)" .“ km ’- C W in» -'."‘ ' ' 67 8 10 Figure 12. Western blot analysis of the inducible expression of COX isoforms in Flp-1n 293 cells. Cells were treated with or without tetracycline. Lysates were separated on a 4-12% NuPAGE Bis-Tris gel, transferred to nitrocellulose membrane and visualized with a specific anti-flag antibody or an anti-COX antibody. Lanes 1 and 2 was lysates isolated from Flag-tagged COX-1 expressing Flp-In 293 cells. Lanes 3 and 4 was lysates isolated from Flag-tagged COX-l insS80-598 expressing Flp-In 293 cells. Lane5 shows a purified COX-1 protein standard. Lanes 6 and 7 was lysates isolated from F lag-tagged COX-2 expressing Flp-In 293 cells. Lanes 8 and 9 was lysates isolated from Flag-tagged COX-2 de1581-598 expressing Flp-In 293 cells. LaneIO shows a purified Flag-tagged COX-2. 38 Proteins identified in both control and induced cells were considered non-specific proteins that either bind non-specifically to the Flag-agarose beads, or were antibodies that detached from the anti-Flag agarose beads. For the proteins that were identified only in the isolated samples from cyclooxygenase-expressing cells, Mascot scores >37 were considered reliable using the NCBI-NR database. The higher the score, the more likely the protein identification is correct. Although multiple spectra are often assigned to a single protein identity, Mascot uses a system of red and bold typefaces to mark the difference. The first time Mascot assigns a peptide to a protein that peptide is shown in bold face. Whenever a spectrum of a peptide is the top ranking peptide match for a protein, it is shown in red. 0n the other hand, if the peptide is not shown in bold, this spectrum may also represent a peptide of another protein. If the peptide is black this indicates a lower fitness of match. The more bold and red peptide matches a protein is identified with, the more reliable it becomes. For our analyses, only protein having a mascot score above 37 using NCBI-NR database with single peptide identified more than once or multiple peptide identifications, at least one of which was typefaced in bold red, were considered as reliable protein identifications. Protein partners for Flag-tagged COX—2 in 293 Flp-In T-Rex cells. Since the expression level of native Flag-tagged COX-2 in 293 Flp-In T-Rex cells was low, it was difficult to obtain sufficient protein for Mass spectrometry analysis. Of seven experiments conducted to identify COX-2 protein partners, five of them successfully identified COX-2. Only proteins that were identified in the tetracycline—induced cells but not in the un-induced cells and met our criteria were considered the potential partners. BiP was identified as the second highest scoring protein (219) in the first experiment (COX-2 with the highest Mascot score (629) and 29 peptides) (Table 3 (a- 39 Proteins Score Peptides Unique NC Bl- Function Peptides G.l. COX-2 629 29 9 181254 Catalyze the committed step in prostaglandins synthesis BiP 219 4 3 6470150 Facilitate the assembly of multimeric protein complexes inside the ER Heat shock- 130 2 2 188492 Molecular chaperones induced protein Heat shock 95 2 2 67462296 Molecular chaperone protein 90 Heat shock 70 64 2 2 1346317 HSP70 family kDa protein 7 FAM44A 53 2 l QSNFC6 Containing AT-hook DNA binding motif Chaperonin 47 3 2 306890 HSP60 family: posttranslational modification, protein turnover, chaperones Table 3(a—1). Proteins co-purified with Flag-tagged COX-2 in 293 Flp-in cells (experiment 1). Protein homogenates were purified with anti-Flag agarose beads. The proteins isolated were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel. The entire protein band was excised and digested out of the gel with trypsin to be analyzed by mass spectrometry. Only proteins had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications and at least one of which was bold red were listed. 1)). BiP is a 78 kDa glucose-regulated protein homolog precursor. This protein is a chaperone-like protein that facilitates the assembly of multimeric ER protein complexes (112). COX—2 and BiP protein both localize to the ER membrane 40 indicating that BiP could be involved in assisting the folding of COX-2. With three unique peptides and a high Mascot score, the identification seems likely to be correct. But this was the only experiments in which this protein was identified. It could be an artifact due to over-expression of COX-2 in this cell line, or an anomaly of the isolation in this experiment. Four heat shock proteins were identified in the induced cells in the first experiment: Heat shock-induced protein, Heat shock protein 90, Heat shock 70 kDa protein 7 and Chaperonin. The scores varied fi'om 47-130 and only one or two unique peptides were identified for each protein. Normally heat shock proteins act like chaperones for protein folding (113). However, only the Heat shock-induced protein (GI/188492) was identified in more than one experiment. Given that different but similar heat shock proteins were also identified in the proteins purified from the control cells, it may be that these proteins are purified non-specifically and their similar sequences lead to variable identification by the Mascot program. 1 185 1248 1710 2872-2884 3051 FAM44A L11 'i‘.,._ . 1,. , Serine AT-hook DNA Bindingdomain 1 185 FAM44B [j 1 173 FAM44C [:1 Figure 13. Sequence alignment of FAM44 proteins. Ser1710 is targeted for phosphorylation by ATM and ATR upon DNA damage. AT-hook DNA binding domain starts at site 2872 containing a GRP tripeptide. 4| FAM44A was another protein that was identified only in the induced cells with a score of 53 and 1 unique bold red peptide. Its cDNA was isolated from human chromosomes 4 (114). FAM44 protein has 3 members: FAM44A, FAM44B and FAM44C which are coded from chromosome 4, 5, 18 respectively. FAM44B has been reported to be involved in chromosome segregation and mitosis (115). Protein sequence alignment demonstrates that FAM44B and FAM44C are most closely related (75% identity). The N terminus of FAM44A is very similar to FAM44B and FAM44C. However F AM44A is much larger than other family members with a large C-terminal extension (Fig 13). F AM44A protein has AT-hook DNA binding domain which is prevalent in many eukaryotic nuclear proteins (116). A number of experiments have demonstrated that AT-hook-containing proteins like HMG-1(Y) play important roles in chromatin structure and act as transcription factor cofactors (117-119). AT-hook containing proteins ELYS/MEL-28 have also been reported as nuclear envelope proteins for nuclear pore assembly and proper cell division (120- 122). So it is possible that FAM44A protein may localize to the nuclear membrane and function in the gene transcription regulation. (Subcellular location of COX-2 is also in the nuclear envelope.) FAM44A has also been reported that can be phosphorylated upon DNA damage, probably by ATM (ataxia telangiectasia mutated) or ATR (ATM and Rad3 related) (123,124). Proteins phosphorylated during DNA damage can serve as a cell cycle check-point control (125,126). For examples, tumor suppressor protein p53 was phosphorylated during DNA damage which prevented the degradation of the protein. The activated form of p53 can induce cell cycle arrest, activate DNA repair, or initiate apoptosis (125). Thus the interaction between COX-2 and F AM44A could be an alternative way to affect cell cycle. After binding, COX-2 may prevent FAM44A binding to its target DNA, disrupt its DNA repairing function 42 up DNA damage, promotes tumor growth. Another possibility is that COX-2 may interact with FAM44A for transcription regulation. As mentioned before, nuclear receptors, PPARs, may regulate gene expression by binding to prostanoids (91-93). , However such a mechanism requires the nuclear transport of prostaglandins. FAM44A may serve as a chaperon for the COX-2-derived prostanoids to import into the nuclear to bind its target receptors for different gene expression regulation, or FAM44A may couple with COX-2-derived prostanoids to regulate gene transcription. In experiments 4 and 5, only COX-2 protein met our criteria for a reliable protein identification (FAM44A was also identified but did not meet our cut-off criteria) (Table. 3(a-2)). In experiment 6, COX-2 had a score of 186 and three unique bold red peptides, no other proteins were identified that met our criteria (FAM44A was also identified but did not meet our cut-off criteria)) (Table. 3(a-2)). An experiment in which Kif was added to cell along with tetracycline, to stabilize COX-2 also yielded no new protein identification. In this experiment, COX- 2 did have a Mascot score of 1402 with 56 peptides isolated and FAM44A was identified that met our criteria (Table. 3(a-3)). 43 Experiment Proteins Score Peptides Unique NCBI- Function Number Peptides GJ. 4 hCOX-Z 92 2 2 181254 Catalyze the committed step in prostaglandins synthesis 5 hCOX-2 53 2 2 181254 Catalyze the committed step in prostaglandins synthesis 6 hCOX-2 186 6 3 181254 Catalyze the committed step in prostaglandins synthesis Table 3(a-2). Proteins co-purified with Flag-tagged COX-2 in 293 Flp-in cells (experiment 4-6). Protein homogenates were purified with anti-F lag agarose beads. The proteins isolated were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel. The entire protein band was excised and digested out of the gel with trypsin to be analyzed by mass spectrometry. Only proteins that had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications and at least one of which was bold red were listed. Proteins Score Peptides Unique NCBI- Function Peptides G.l. hCOX-2 1402 56 15 181254 Catalyze the committed step in prostaglandins synthesis FAM44A 68 5 2 Q8NFC6 Containing AT-hook DNA binding motif Table 3(a-3). Proteins co-purified with Flag-tagged COX-2 in 293 Flp-in cells (experiment 7). Protein homogenates were purified with anti-F lag agarose beads. The proteins isolated were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel. The entire protein band was excised and digested out of the gel with trypsin to be analyzed by mass spectrometry. Only proteins that had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications and at least one of which was bold red were listed. 44 Proteins Score Peptides Unique NC Bl- Function Peptides G.l. hCOX-2 1010 34 14 181254 Catalyze the committed step in prostaglandins synthesis Table 3(b). Proteins co—purified with Flag-tagged COX-2 de1581-598 in 293 Flp- in cells. Protein homogenates were purified with anti-F lag agarose beads. The proteins isolated were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel. The entire protein band was excised and digested out of the gel with trypsin to be analyzed by mass spectrometry. Only proteins that had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications and at least one of which was bold red were listed. Protein partners for COX-2 de1581-598 in 293 Fh)-In T-Rex cells. This experiment was also repeated for three times. COX-2 was the only unique protein identified (Table 3-b). The identification of proteins that specifically co-purified with COX-2 and not with COX-2 de1581-598 suggests that the 18 amino acids may be important in protein-protein interaction. Protein partners for COX-1 in 293 Flp-In T-Rex cells. Potential protein partners for native oCOX-l were also identified by mass spectrometry analysis. Experiments were conducted twice and COX-1 was the only protein identified in the experiment one that met our criteria (Table. 3(c-1)). For experiments 2, several other proteins were identified besides the COX-1 protein. However all the other proteins identified had low scores (most peptides identified were black) and no common proteins were identified for COX-1 and COX-2 (Table. 3(c-2)). 45 Proteins Score Peptides Unique NCBI- Function Peptides G.I. COX-1 1 14 10 4 249624 Catalyze the committed step in prostaglandins synthesis Table 3(c-l). Proteins co-purified with Flag-tagged COX-1 in 293 Flp-in cells (experiment 1). Protein homogenates were purified with anti-Flag agarose beads. The proteins isolated were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel. The entire protein band was excised and digested out of the gel with trypsin to be analyzed by mass spectrometry. Only proteins that had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications, at least one of which was bold red were listed. Proteins Score Peptides Unique NCBI- Function Peptides G.l. COX-1 571 52 12 55958152 Catalyze the committed step in prostaglandins synthesis 46 10 4 21427632 Belong to the ASC- MLL3 2/NCOA6 complex (ASCOM), a coactivator complex of nuclear receptors Protocadherin 46 4 4 7407146 Similar to the secretin Flamingo 1 family of G-protein linked receptors KIAA0543 protein 38 13 3 51466599 Regulation of transcription, DNA- dependent Table 3(c-2). Proteins co-purified with Flag-tagged COX-l in 293 Flp-in cells (experiment 2). Protein homogenates were purified with anti-Flag agarose beads. The proteins isolated were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel. The entire protein band was excised and digested out of the gel with trypsin to be analyzed by mass spectrometry. Only proteins that had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications, at least one of which was bold red were listed. 46 The Protocadherin Flamingo 1 had a Mascot score of 46. This protein is a seven- transmembrane receptor located on the cell surface (127), which is a different cell location compared to the location of COX-1, so it seems like this protein does not interact with COX—1 in viva. Myeloid/lymphoid or mixed-lineage leukemia protein 3 (MLL3, Histone-lysine N-methyltransferase, H3 lysine-4 specific MLL3) belongs to the ASC-2/NCOA6 complex (ASCOM) (128), a coactivator complex of nuclear receptors, involved in transcriptional coactivation. Since the subcellular location of COX-1 is not in the nuclear, this protein does not interact with COX-1. KIAA0543 protein has a pretty low score of 38, but of 13 peptides, three unique ones were identified (129). The protein was present in the nuclear and may be involved in the transcriptional regulation. Again as with MLL3, it seems like the interaction does not occur in vivo and is an artifacts of the isolation. Proteins Score Peptides Unique NCBI- Function Peptides G.l. COX-1 42 2 1 249624 Catalyze the committed step in prostaglandins synthesis Table 3(d-1). Proteins co-purified with Flag-tagged COX-1 ins580-S98 in 293 Flp-in cells (experiment 1). Protein homogenates were purified with anti-Flag agarose beads. The proteins isolated were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel. The entire protein band was excised and digested out of the gel with trypsin to be analyzed by mass spectrometry. Only proteins that had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications, at least one of which was bold red were listed. 47 Proteins Score Peptides Unique NCBI- Function Peptides G.l. COX-l 363 25 8 249624 Catalyze the committed step in prostaglandins synthesis COX-2 101 6 2 181254 Catalyze the committed step in prostaglandins synthesis FAM44A protein 52 8 l Q8NFC6 Containing AT-hook DNA binding motif Homology of FAM44A GR1A2 protein 47 5 3 14714846 Receptor for glutamate Signal-induced 44 3 2 55664135 Containing a PDZ proliferation- domain and a Rap-Gap associated 1 like2 domain Zinc finger protein 44 21 3 51473106 May be involved in 469 transcriptional regulation Table 3(d-2). Proteins co-purified with Flag-tagged COX-1 ins580-598 in 293 F lp-in cells. Protein homogenates were purified with anti-Flag agarose beads. The proteins isolated were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel. The entire protein band was excised and digested out of the gel with trypsin to be analyzed by mass spectrometry. Only proteins that had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications, at least one of which was bold red were listed. 48 Proteins Score Peptides Unique NCBI- Function Peptides G.l. Beta actin variant 493 14 62897409 Essential for the structural integrity Anti-colorectal 345 15 425518 lmmunoglobulin C carcinoma heavy chain region Chain M, Crystal 326 23 24158784 lmmunoglobulin Structure Of Fab domain variable region Fragment Complexed With Gibberellin A4 Methylates specific Protein arginine N- 299 6 2323410 arginine residues in the methyltransferase 5 small nuclear ribonucleoproteins Sm D1 and Sm D3 lg kappa 138 19 227564 lmmunoglobulin V region lmmunoglobulin kappa 138 19 2970528 lmmunoglobulin V light chain region Methylosome Protein 63 2 13559060 May regulate an early 50 step in the assembly of U snRN Ps, possibly the transfer of Sm proteins to the SMN-complex Table 3(e). Proteins non-specifically isolated on anti-flag agrose beads for 293 Flp-in cells. Protein homogenates were purified with anti-Flag agarose beads. The proteins isolated were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel. The entire protein band was excised and digested out of the gel with trypsin to be analyzed by mass spectrometry. This list contains the proteins that were isolated and identified in both samples, and represent what we assume are proteins that non-specifically bind anti-flag agrose beads. Only proteins that had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications, at least one of which was bold red were listed. 49 Protein partners for COX-1 ins580-598 in 293 Flp-In T-Rex cell svstem. Potential protein partners for native oCOX—l were also identified by mass spectrometry analysis. Experiments were conducted twice and COX-1 insS80-598 was the only protein identified in the experiment one that met our criteria (Table. 3(d- 1)). For experiments 2, several unique proteins were identified only in the induced cell samples (Table 3-(d-2)). None of these proteins were co-isolated for both COX-1 and COX-1 insS80-598. Since COX-1 insS80-598 containing the 18-amino acids of COX- 2 at its amino terminal, during the mass spectrometry analysis this peptide was recognized and identified as a match to COX-2 by Mascot. Interestingly one protein identified in COX-1 ins580-598 complex was FAM44A with a Mascot score of 53 and one unique peptide identified eight times. FAM44A was identified with COX-2 repeatedly, and once with COX-1 in.s580- 598, but never with COX-2 de1581—598 or COX-1, suggesting its interaction may dependent on the 18-amino acids insert. Similar to oCOX-l, some proteins that were involved in DNA binding and transcription were also identified: such as Zinc finger protein 469. As stated before the interaction is probably not real. Common artifactual protein identified in 293 Flp-In T-Rex cells. Common proteins present in all the COX expressing cells in the tetracycline-induced and uninduced cells are shown in Table 3-e. Actin is one major protein that is identified. This protein is important for the cell structural. It could bind directly to the beads, causing an artifact (130). A protein identified as anti-colorectal carcinoma heavy chain, 1g kappa, lmmunoglobulin kappa light chain are probably eluted from the anti-F lag agarose beads. 50 Protein arginine N-methyltransferase 5 and methylosome protein 50 are both nuclear proteins. They likely bind to the agarose beads non-specifically. Mient expression of Flpgiagged COX-2 in 293 Freestyle cell svstem an_c_l_ identimng possible partners. Transient transfections in the 293 Freestyle cells were also carried out to identify the possible partners for Flag-tagged COX-2 (Fig 14). In this system cells can be grown in the shaker flask instead of the plate, thus it is much easier to scale up more cells for the experiment. After transient transfection, cells were harvested, protein was purified with anti-Flag beads and analyzed by mass spectrometry (Table. 3-(f-1)). Figure 14. Western blot analysis of Flag-tagged COX-2 expression in 293 Freestyle cells. Lysates were separated on a 4-12% NuPAGE Bis-Tris gel, transferred to nitrocellulose membrane and visualized with a specific anti-flag antibody Lanes 1 shows a purified Flag-tagged COX-2 standard. Lane 2 was lysates isolated from transient transfection of Flag-tagged COX-2 in Freestyle 293cells. Two heat shock proteins were identified only in the transfected cells: heat shock-induced protein and heat shock cognate 71 kDa protein. Heat shock-induced protein had the second highest score (521) and 20 peptides were identified. The heat shock-induced protein was the same protein identified in the T-Rex cells (Table. 3-a (1)). Since this protein was identified twice in different cell system both with high credibility, it could be a possible partner of COX-2. Recent work has proven that this 51 Proteins Score Peptides Unique NC Bl- Function Peptides G.l. COX-2 546 36 10 181254 Catalyze the committed step in prostaglandins synthesis Heat shock-induced 521 20 7 188492 In cooperation with protein other chaperones KIAA0139 237 28 14 40788877 Motifin proteasome subunits, Int—6, Nip-1 and TRIP-15 DN A-binding protein 80 10 3 9802306 Ribosomal protein L6 TAXREB 107 Heat shock cognate 71 75 4 2 57085907 Chaperones kDa protein DEAD-box protein 3, 63 5 4 6014945 Inhibits the binding of Y-chromosomal survival motor neuron protein (SMN) to Sm proteins Mammary tumor- 62 2 2 2695641 Motif in proteasome associated protein subunits, Int-6, Nip-1 1NT6 and TRIP-15 Splicing factor 38 54 4 2 19863446 Subunit ofthe splicing subunit 3 factor SF3B FAM44A protein 53 5 3 QSNFC6 Containing AT-hook DNA binding motif ElF3S9 protein 52 4 4 12654669 Eukaryotic translation initiation factor 3 subunit 9 Table 3(f-1). Proteins co-purified with Flag-tagged COX-2 in 293 Freestyle cells (experiment 1). Protein homogenates were purified with anti-Flag agarose beads. The proteins isolated were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel. The entire protein band was excised and digested out of the gel with trypsin to be analyzed by mass spectrometry. Only proteins that had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications, at least one of which was bold red were listed Proteins Score Peptides Peptides Unique NCBI- Function COX-2 ATP synthase subunit beta, mitochondrial precursor Calmodulin ATP synthase subunit alpha, mitochondrial precursor Transmembrane protein 109 precursor (Mitsugumin-23) 1194 254 128 48 44 46 10 181254 28940 71664 4757810 13129092 Catalyze the committed step in prostaglandins synthesis Produces ATP from ADP in the presence of a proton gradient across the membrane. The beta chain is the catalytic subunit. calcium-binding protein that can bind to and regulate a multitude of different protein targets Produces ATP from ADP in the presence of a proton gradient across the membrane. The alpha chain is a regulatory subunit. May interacts with cytoplasmic protein(s) and participates in a housekeeping function Table 3(f-2). Proteins co-purified with Flag-tagged COX-2 in 293 Freestyle cells (experiment 2). Protein homogenates were purified with anti—Flag agarose beads. The proteins isolated were electrophoresed into the stacking gel of an 8-16% Tris-HCl SDS-PAGE gel. The entire protein band was excised and digested out of the gel with trypsin to be analyzed by mass spectrometry. Only proteins that had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications, at least one of which was bold red were listed. protein can inhibit LPS-induced NF-KB signaling cascade activation and subsequently decrease COX-2 expression (131). Heat-shock induced protein can bind directly with TRAP-6 protein to inhibit its ubiquition. TRAF-6 will affect lKk so it won’t be activated, and thus NF-KB signaling cascade. So the interaction between COX-2 and the heat shock-induced protein could be an alternative way to affect the expression of COX-2. Heat shock cognate 71 kDa protein had a Mascot score of 76, although the four peptides were not identified with high confidence. This was the only time this protein was identified so the presence of the protein could simply be stimulated by the over-expresSion of COX-2 protein and thus be an artifact. KIAA0139 isolated from 293 freestyle cells transfected with COX-2 have motifs involved in protein degradation. KIAA0139 had a score of 237 and 28 peptides, eleven of which were unique, indicated with high probability that this protein is present (although this was the only time this protein was identified) (132). COX-2 is degraded via the proteasome pathway. If the interaction was genuine, this protein could mediate the degradation of COX-2, especially if it is expressed at high level. FAM44A protein was identified again with a score of 53, with five peptides, of which 3 were unique. This protein was identified in both the T-Rex and Freestyle cell system. FAM44A has been identified four times in the T-Rex cells. It was the only protein that was identified in both the native COX-2 complex and the COX-1 containing the COX-2 insert protein complex. Several proteins that were involved in the transcription were identified: DNA- binding protein TAXREB107, DEAD-box protein 3, Y-chromosomal. Since the subcellular location of COX was not in the nuclear, it is a high probability that the interactions were not real. 54 The experiment was repeated once with twice the cells. Surprisingly no overlap proteins were found between the two transfections. Two proteins were newly identified as Mitochondrial F -type ATP synthase alpha and beta subunits respectively. The proteins are mitochondrial proteins located on the inner membrane and are - involved in ATP synthesis (133). Both alpha and beta subunit belongs to the catalytic core F1. Thus, it is not clear why this protein might associate with COX-2. Common artifactual protein identified in the 293 Freesgle cells. lmmunoglobulin, anti-colorectal carcinoma heavy chain and immunoglobulin kappa chain, actin and the ribonucleoproteins are some common proteins that are both present in the transfected and control 293 cells (Table. 3g). New artifactual proteins also showed up in this expression system: Ubiquitin and Histone 1. Why different artifactual proteins occur when COX-2 is purified using the 293 cells transient expression is not clear. Proteins Identification Identification Total NCBI- Function in Freestyle identifications (3.1. 293 cells/ Total experiments experiments , Containing AT- FAM44A 1/2 6 08NFC6 hook DNA binding motif Heat shock- 1/2 2 18 8492 Molecular induced chaperones protein Table 3(f-3). Proteins reproducibly co-purified with Flag-tagged COX-2. 55 Proteins Score Peptides Unique NCBI- Function Peptides G.l. Heterogeneous nuclear 1 109 47 10 4758544 Binds pre-mRNA and ribonucleoprotein C nucleates the assembly of 405 hnRNP particles Heterogeneous nuclear 1087 40 1 1 16923998 Binds pre-mRNA and ribonucleoprotein K nucleates the assembly of 40S hnRNP particles Heterogeneous nuclear 1027 43 9 133254 Binds pre-mRNA and ribonucleoprotein A1 nucleates the assembly of 408 hnRNP particles lmmunoglobulin kappa 732 29 3 1572705 lmmunoglobulin V chain region Histone 1, H2aj 568 30 5 7264004 Core component of nucleosome Anti-colorectal 527 30 5 425518 lmmunoglobulin C carcinoma heavy chain region HNRPU protein 488 43 12 16041796 Binds to pre-mRNA Cytoskeletal protein, Vimentin 488 22 9 37852 assembles into 10nm filaments Heat shock protein 260 9 4 386785 HSP70 family Helicase superfamily Dead box, X isoform 237 12 5 2580552 c-terminal domain Beta actin variant 181 8 4 62897625 Essential for the structural integrity Ubiquitin 139 7 3 340062 Mark other proteins for destruction Table 3(g). Proteins non-specifically isolated on anti-flag agrose beads for 293 Freestyle cells. This list contains the proteins that were isolated and identified in both samples, and represent what we assume are proteins that non-specifically bind anti— flag agrose beads. Only proteins that had a Mascot scores above 37 with single peptide identified more than once or multiple peptides identifications, at least one of which was bold red were listed. 56 Discussion FAM44A is the strongest candidate protein partner for COX-2. It was identified in 5 out of the 7 experiments in protein complexes with COX-2 purified from T-Rex 293 cells, and 1 in 2 of the experiments where COX-2 was expressed transiently in Freestyle 293 cells. FAM44A was also identified in protein complexes with COX-1 ins580-598 but never in the uninduced 293 cells, or in protein complexes with COX-1 or COX-2 de1581-598, suggesting association requiring a specific interaction with the 18-amino acid cassette of COX-2. FAM44A protein have AT-hook DNA binding domain which is prevalent in many eukaryotic nuclear proteins (116) and a number of experiments have demonstrated that AT-hook-containing proteins play important roles in chromatin structure and act as transcription factor cofactors (117-119). AT- hook containing proteins ELYS/MEL-28 have also been reported as nuclear envelope proteins for nuclear pore assembly and proper cell division (120-122). So it is possible that FAM44A may localize to the nuclear membrane and function in the gene transcription regulation. F AM44A has also been reported that can be phosphorylated upon DNA damage, probably by ATM (ataxia telangiectasia mutated) or ATR (ATM and Rad3 related) (123,124). Proteins phosphorylated during DNA damage can serve as a cell cycle check-point control (125,126). Thus the interaction between COX-2 and FAM44A could be an alternative way to affect cell cycle. COX-2 has been previously reported involving in cell cycle control. NSAIDS can inhibit cell growth by inducing apoptosis and cell cycle arrest in either human hepatocellular carcinoma cells or human oesophageal adenocarcinima cells (134,135). Serum induced cell growth in NIH 3T3 cells was related to Go to G1 cell cycle control by expressing COX-2 (136). Another possibility is that FAM44A may serve as a chaperon for import of COX-2-derived prostanoids to nucleus. For future study, co- 57 immunoprecipitation and GST pull down assay need to be carried out to confirm the interaction. The heat shock-induced protein was another protein that was identified in both T-Rex and Freestyle systems associating with COX-2. It was identified 1 out of the 7 experiments in protein complexes with COX-2 purified from T—Rex 293 cells, and 1 in 2 experiments where COX-2 was expressed transiently in Freestyle 293 cells. Since heat shock-induced protein was not present in complexes with the COX-1 ins580-598 or the COX-2 de1581-598, the interaction may require COX-2 protein domains in addition to the 18-amino acid cassette of COX-2. Heat shock-induced protein may aid in the processing of COX-2 or be involved in its degradation. Recent work has proven that this protein can inhibit LPS-induced NF-KB signaling cascade activation and subsequently decrease COX-2 expression (131), thus the interaction between COX-2 ' and heat shock-induced protein may also affect the expression of COX-2. Although heat shock-induced protein has been identified in both systems, overall it was only co- purified twice with COX-2 in 9 experiments. It could be an artifact due to over- expression of COX-2, or an anomaly of the isolation in a specific experiment. BiP, KIAA0139 and Mitochondrial F-type ATP synthase were also co-purified once with COX-2 complex in all 9 experiments. BiP protein identified in T-Rex cells is a possible partner for COX-2 as may be involved in COX-2 folding in ER as a chaperone. KIAA0139 protein isolated fi'om 293 freestyle cells transfected with COX- 2 could be involved in the degradation of COX-2. ATP synthase is a mitochondrial enzyme and its interaction with COX-2 has no defined functions. Since all three proteins were not present in complexes with COX-l ins580-598 or COX-2 de158l- 598, their interaction may require COX-2 protein domains in addition to the 18-amino acid cassette of COX-2. Although they were all identified with high credibility (high 58 Mascot score and many bold red peptides identified) one time with COX-2, considering it's the only time they were found, it was possible that they were just an artifact due to over-expression of COX-2. Several other heat shock proteins were also identified once in the 293 T-Rex or Freestyle system. Given that different but similar heat shock proteins were also identified in the proteins purified from the control cells, it may be that these proteins are purified non-specifically and their similar sequences lead to variable identification by the Mascot program. For COX-2 de1581-598 construct, 3 experiments were conducted, but no unique proteins were ever identified in the T-Rex cells. The identification of proteins that specifically co-purified with COX-2 and not with COX-2 de1581-598 suggests that the 18 amino acids may be important in protein-protein interaction. Two experiments were carried out to identify potential protein partners for COX-l in T-Rex cells. The first experiment failed to capture any proteins that met our criteria except COX-1. When the experiment was repeated, several unique proteins co-purified with COX-1. However the Mascot scores of these three proteins, MLL3, Protocadherin Flamingo 1 and KIAA0543, identified are near 37 (cut-off score) and most peptides identified were low credibility. Moreover they don’t share subcellular location with COX-1. It is reasonable to assume that none of the proteins were real partners for COX-1 in viva. No protein partners were identified in cormnon for COX- 1 and COX-2. Since COX-1 and COX-2 have very different biological fimctions, it is reasonable to assume that the specific interaction identified in COX-2 may cause different functions between COX isozyme. Among the proteins co-purified with COX-l ins580-598, F AM44A protein was the most probable protein partner candidate. As stated above, this protein was also co- 59 purified with COX-2 complex and may association with COX-2 specifically via the 18-amino acid cassette. Although several other proteins were also identified, none of them were co-isolated with both COX-1 and COX-1 in5580-598. Thus, these proteins are likely artifacts. The explanation for why not many reliable candidates were identified could be due to the followings: the interactions are transient, or that the interaction of these proteins have low affinity and do not survive purification. Furthermore 293 cells do not normally express endogenous COX-2, and may not express protein partners for COX-2. The failure to obtain any new protein partners even when using the Kifnusenin suggested that either no proteins interact with COX-2 during the degradation or the interaction has low affinity so no potential partners survive during purification. One limitation of this experimental approach is the non-specific purification of protein via agarose affinity chromatography. A large number of proteins were identified in both COX expressing and control samples). These non-specific proteins had high Mascot scores with multiple peptides identified. Because the signals of these proteins are so intense, they may mask the MS spectral signals from the true COX-2 protein partners. To overcome these problems, different cell lines could be used instead of the 293 cells. The cells should endogenously express COX isozymes to guarantee that potential protein partners exist. Secondly, alternative tag could be employed for COX protein purification. 60 CHAPTER HI EFFECTS OF EXPRESSION OF CYCLOOXYGENASE PROTEINS ON CELL BIOLOGY Summary Our investigation also included exploring the role of the 18-amino acid cassette on cell biology. COX-2 variant without the 18-amino acid cassette or which were catalytically inactive were expressed in a higher percentage of transfectants and at higher protein levels than native COX-2, indicating that both catalytic activity and the 18-amino acid cassette were important mediators of the protein expression. Immunocytochemistry of COX isoforms was carried out to examine whether the 18 amino acids affects the subcellular localization of COX enzymes, and no differences were observed. All COX proteins localized to the nuclear envelope and ER membrane. Growth curves and agar assays were also conducted on T-REX 293 cell lines stably expressing Flag-tagged COX-2, COX-2 de1581-598, COX-1 and COX-1 insS80-598, but no significant difference was observed cells expressing the different COX isoforms. The T172 of Flag-tagged COX-2, COX-2 de1581-598, COX-1 and COX-1 ins580-598 were determined and found to be similar to each other, indicating that the activity and the 18-amino acid cassette may not be important in the degradation process. However results in Smith’s laboratory later showed that the FLAG tag at the N-terminus affected the turnover rate of COX enzymes. Proteins without FLAG tags were constructed and expressed in the 293 T-REX cell line. Surprisingly the non- tagged COX-2 delS81-598 protein has a much longer T1,: (compared to COX-2) while the COX-1 ins580-598 protein has a shorter Tm (compared to COX-1), indicating that l8-amino acid cassette is involved in the protein turnover regulation. Furthermore 61 COX-2 N580A also has a much longer T1 /2 than native COX-2, suggesting that the C- terrninal glycosylation site (just before the 18-amino acid cassette) was also important in the turnover regulation of COX proteins. Introduction Prostaglandin Endoperoxide H synthase-1 and 2 (PGHS-l and PGHS-2; also known as cyclooxygenase-1 and 2, COX-1 and COX-2) catalyze the committed step in prostaglandins synthesis. Cyclooxygenase-1 and 2 have many features in common while many studies show that the biological function is quite different between the two. COX-1 is constitutively expressed in most tissues and knock-out mouse studies have confirmed its responsibility in platelet aggregation (76) and parturition (77, 78). COX-2, however, is not expressed in most unstimulated tissues but is rapidly induced by growth factors, cytokines, and tumor promoters. COX-2 is responsible for ovulation, implantation (79), and neonatal development (80, 81). To date, many evidences have indicated a primary role for COX-2, rather than COX-1 in carcinogenesis (84, 85). New data from the Smith laboratory has indicted that one role of the l8-amino acid cassette is to mediate entry of COX-2 into the ER-associated degradation (ERAD) system that transports ER proteins to the cytoplasm (74). Their studies found that the insertion of the 18-amino acid cassette into COX-1 can destabilize the protein while deletion of the 18-amino acid cassette in COX-2 stabilized the protein. Thus, the 18- amino acid cassette has a critical role in regulation of COX-2 stability. The Smith group has also proposed that the glycosylation of N580 and subsequent trimming of the core sugar GlC3ManoGlCNAC3 complex is necessary for the entry of COX-2 into the 62 ER-associated degradation (ERAD) system based in part on the effects of Kifnusenin, an inhibitor of terminal fucosidases that stabilized the protein (74). The different regulation of COX protein stability could be important for their different biological activity. In chapter 11 I provided evidence that it is difficult to isolate cell lines that express COX-2, but COX-1 can be expressed without problem. If the 18-amino acid cassette is involved in regulating the protein expression, the deletion of the cassette should allow COX-2 expression and insertion of the cassette should prevent such expression. > Recent work demonstrated that cotransfection of COX-2 and mPGES-1 . (microsomal PGE synthase) into HEK293 cells can stimulate the production of PGE2, and thus cell proliferation (137). Co-expression of COX-2 and mPGES-1 in HEK293 cells can also resulted in cellular transformation and tumor formation. This was accompanied by changes in the expression of a variety of genes related to proliferation, morphology, adhesion and the cell cycle such as RhoA and TRAF-l. We attempts to replicate their experiments and also to determine process by examining the cell growth in transfected cell lines expressing different forms of COX- 1 and -2. There has been no good explanation for the apparent redundancy of COX—1 and 2. These two enzymes are so similar that it is difficult to explain why their biological activities are different. 18-amino acid cassette of COX-2 has been proved to be an important factor in regulation the protein stability. The study of this Chapter is to verify the results and determine if it is also responsible for other distinct biological activities of COX-2. 63 Experimental Procedures Materials. F lp-In-T-Rex-293 cell line was from Invitrogen. Dulbecco’s modified Eagle medium (DMEM), Lipofectamine 2000, tetracycline, blasticidin, zeocin, hygromycin, and penicillin-streptomycin were from Invitrogen/Gibco. Fetal bovine serum (FBS) was from Atlas Biologicals. Cycloheximide and puromycin were obtained from Sigma-Aldrich. Construction of flDNA expression plasmids for COX-2, COX-2 de1581-598, COX-2 N580A, oCOX-l aid COX-l in3580-598. Untagged version of the COX-2- pCDNA, COX-2 de1581-598-pCDNA and COX-2 N580A-pCDNA were constructed using FLAG-TEV-plasmids as templates by PCR mediated mutation. The Flag-tagged COX-l-pCDNA and COX-1 in3580-598-pCDNA were used as templates to make untagged version. All the plasmids were used for the stable transfection in F LP-In 293 cell lines. The primers, vector and the PCR conditions used to design the above mutants are shown in the ‘Appendix’ section (Table. 7). Construction of the human microsomal PGES (PGE: synthase) in the pBABE vector. cDNA of human prostaglandins E synthase was amplified from pOTB7 vector (Appendix Fig. 30) by PCR. The fragment was cloned into Zero blunt TOPO vector (Invitrogen) (Appendix Fig. 31) and then into pBABE-puro vector (Appendix Fig. 32) at the EcoRI site. This plasmid was used to transfect Flp-In 293 cells that inducibly co-express one of the COX mutants. S_table transfection of mPGES in Flp-In 293 cells with COX-1 a_r_1_d COX-2 expressing. The pBABE-puro plasmid coding for the PGE synthase enzyme was transfected into COX expressing Flp-In 293 cells (Invitrogen). One day before transfection, l x 10° cells were plated into a 60mm tissue culture plate in 4 ml of growth medium. Cells were 90-95% confluent at the time of transfection. For each 64 transfection, 8 pg DNA (811g mPGES-pBABE) was diluted in 0.5 ml of Opti-MEM® 1 Medium without serum and mixed gently. Transfections were carried out as described previous. Clones expressing target proteins were selected by treating the cells with 1511ng blasticidin, 10011g/m1 hygromycin B and 111ng puromycin. Constitutive expression of the mPGES was analyzed by Western blot analysis. Western Blotting. Proteins samples for the Western blot were separated on a 4- 12% Bis-Tris NuPAGE gels and transferred to nitrocellulose membranes. Visualization was performed by incubating with anti-flag (1:3000 dilution), anti-COX (1:2000 dilution) or anti-mPGES (1:500 dilution) primary antibody in 5% milk TBST (TBS plus 1% (v/v) Tween-20) solution for two hs. The membranes were then washed four times with TBST for 10 minutes. Membranes were next incubated with horseradish peroxidase-conjugated anti-mouse or anti-rabbit IgG antibodies in 5% milk TBST solution (1:2000 dilution) for one h. After four 10 minutes washes with TBST, immunoreactive proteins were visualized using the Western Lighting Chemiluminiscent Kit (Perkin Elmer, Boston, MA) and exposure to x-ray film (Amersham). For the actin controls, the same blots were blocked a second time in 5% milk TBST overnight, and then incubated with Anti-actin antibody (1:2000 dilution) in TBST for two hs. The membrane was washed four times with TBST for 10 minutes, and the membranes were then incubated with horseradish peroxidase-conjugated anti- mouse IgG antibody in 5% milk TBST solution (1:2000 dilution) for one h. After four washes with TBST for 10 minutes, actin was visualized as described above. Cell Growth Curvmd Semi-soft Agar Assay. Cells were seeded at 1.0X 10 4 or 5.0>< 10 4 cells/ ml (6.0x 10 4 or 3.0x 10 5 cells/ 20cm2 plate) and cultured for 10 days. 65 Individual plates were harvested every 2 days and cell number was determined by counting with a hemocytometer. Semi-soft Agar Assay: Cells ( 104 cells/ml ) were suspended in cell culture medium containing 1% (w/v) low-melt agarose at 37°C and plated on 60-mm culture dishes. After culturing for 10 days at 37 °C in a 6% C02 incubator, colonies were counted. Colony size was determined by measuring five random colonies in each plate using a calibrated lens ruler. Immunocflochemisfl of COX-1 and COX-2 mutants. Stably-transfected HEK 293 T-REX cells were plated on glass coverslips in 6 well plates in medium with 2mg/ml tetracycline overnight. For staining, the cells were washed with phosphate buffered saline (PBS), fixed by incubating with 2% paraformaldehyde in PBS for 10 minutes, and then washed with PBS containing 10% calf serum in (138). The cells were next incubated with a Flag antibodies diluted 1:20 in PBS containing 0.2% saponin and 10% calf serum for 30 min. After washing in PBS containing 10% calf serum, the samples were incubated for 30 min with a 1:40 dilution of FITC- conjugated goat anti-mouse IgG in PBS containing 0.2% saponin and 10% calf serum. The samples were washed with PBS containing 10% calf serum and then rinsed with PBS. Finally, coverslips containing the stained cells were mounted with PermaFluor on slides and examined by fluorescence microscopy. Determination of COX protein T1,». Stably-transfected HEK 293 TREX cells were cultured in 100mm>< 20mm plates. Eight plates of each cell line, 107 cells per plate were Seeded. One plate was used as a no tetracycline negative control, the rest were induced for cyclooxygenase expression by adding 211g/m1 tetracycline for 24 hrs. The cells were washed then with PBS and incubated with the medium containing 5011m cycloheximide for 0, 2, 4, 6, 8, 12, 24 hrs before harvesting. Cells were 66 dissolved in 0.5m] PBS containing protease inhibitors (Complete Mini Protease Inhibitor, Roche) and sonicated for 3 times 10 seconds and expression of COX was determined by Western blotting. Expression levels were quatitated by analyzing the band density of the scanned X-ray film, using the Scion Image program (hpp://www.scioncogpcom/pages/scion image windows.htm). Tm was calculated by integrated first-order velocity equation; 2.3log [S] o/[S] =kt. [S] 0 represents the band density of the time 0; [S] represents the band density of the time 2 to 24 hrs respectively. All the band density was normalized against the actin control. The value of log [S] versus the time t (0-24 hr) was used to plot a straight line via Microsoft excel program, and the slope of the curve is —k/2.3 for equation log [S] = log [S] 0- kt/2.3. When [S] equals to 1/2 [S]o, then t in the equation is the T172 of the protein T 1/2 , and T 172 =-2.3log2/k. Since k equals to -2.3slope from the curve, T [/2 can be calculated by log2/ slope. Ratio of expression and loss of eflnession. Cell clones expressing native COX-2 were much rarer than clones expressing other COX-2 mutants. To examine the reasons for lower native COX-2 expressing clones, different constructs of COX-1 and COX-2 were transfected, and the ratio of antibiotic resistant colonies expressing and not expressing each protein was determined. Twenty four or more antibiotic resistant colonies resulting from transfections of Flp-In 293 cells with expression plasmid constructs for COX-1 and COX-2 (Flag-tagged COX-2, COX-2 de1581-598, COX-2 Y371F, COX-l in3580-598, COX-1) were randomly selected. The effect of NS-398, a COX-2 selective inhibitor was also used to determine whether cyclooxygenase inhibitors would increase recovery of COX-2 expressing cell lines. Cells from each transfection were harvested and checked for expression by Western blotting. 67 The long term ability of stably transfected cells to express native COX-2 was also examined. Cells were cultured in the absence and presence of tetracycline to determine whether constitutive expression of COX-2 lead to loss of enzyme expression. After each harvest, cells were divided into collected. After 10 generations, protein expressions were checked by Western blot. Results mo of expressiorflnd loss of expression. Based on our difficulty expression COX-2 in MCF-7 cells and HEK 293 derived cell lines, we attempted to determine what structure or catalytic features of COX-2 prevented its expression. Transfection experiments. with different forms of COX protein were carried out and the relative efficiency of isolating clones expressing different mutants was determined by Western blot (Table. 4). COX-1 and COX-1 insS80-598 were expressed in 100% of 24 randomly picked colonies. COX-2 had the lowest expression ratio, and was detectable in 26.7% of 'colonies. Expression ratios of COX-2 de158l-598 and COX-2 Y371F COX isoforms Figure 15. Western blot examining stability of expression of COX in transfected cell lines. Samples were separated on a 4-12% NuPAGE Bis-Tris gel, transferred to nitrocellulose membrane and visualized with a specific anti-flag antibody. Lane I is standard Flag-tagged COX-2. Lane 2 was lysates isolated fi'om constitutive Flag- tagged COX-2 expressing Flp-in 293 cells. Lanes 3 was lysates isolated form Flag- tagged COX-2 expressing F lp-in 293 cells respectively. 68 (an inactiveimutant) were 41.7 and 45.8% respectively, higher than native COX-2 but lower than native COX-1 and COX-1 ins580-598. Since inactive COX-2 (COX-2 Y371F) was expressed more readily than native COX-2, it is possible that COX-2 activity has a negative effect on cell growth. To test this possibility, the effect of addition of a COX-2 specific inhibitor NS398 (139) during the transfection process was tested to see if it increased our ability to isolate Protein Expressed colonies in all % Expressing colonies F lag-tagged COX-1 24/24 100.0 Flag-tagged COX-1 insS80- 24/24 100.0 598 F lag-tagged COX-2 8/30 26.7 Flag-tagged COX-2+NS 398 11/12 91.7 Flag-tagged COX-2 de1581- 10/24 41.7 598 F lag-tagged COX-2 de1581- 21/24 87.5 598+NS 398 Flag-tagged COX-2 Y371F 1 1/24 45.8 F lag-tagged COX-2 Y371F 21/24 87.5 +NS 398 Table 4. Transfection efficiency for COX proteins. Stably transfected Flp-1n 293 cells line expressing different COX mutants were used to check the ratio of expression. Samples from different colonies were separated on a 4-12% NuPAGE Bis-Tris gel, transferred to nitrocellulose membrane and visualized with a specific anti-flag antibody. The ratio number of colonies that expressed the target protein was calculated. 69 COX-2 expression colonies. When NS398 was added during the selection process, 92% of colonies expressed native COX-2. Similar results were also obtained for the COX-2 de1581-598 when treated with the drug. Our results have shown that either deletion of the 18-amino acid cassette or inactivation of the enzyme could increase protein expression ratio as compared to the native COX-2, indicating both activity and the 18-amino acid cassette were important for the protein expression. Experiments were also carried out to determine whether constitutive expression of COX-2 was detrimental to cell growth. If so, we expected to see a gradual loss of COX-2 expression overtime. As shown in Fig 15, expression of COX-2 disappeared in cells grown for 14 sub-culturing in the presence of tetracycline while non- constitutively COX-2 expressing cells maintained the normal expression. This further confirmed that COX-2 expression was tightly controlled in the cells. Cell Growth Curve and Semi-soft Agar Assay for F lag-tagged COX-2, COX-2 de158l-598LCOX-l and COX-1 ins580-598. Evidence has indicated that NSAIDS can reduce the incidence of colorectal cancers by inhibiting colon cancer cell growth and COX-2 was related to this effect (140). To determine if the COX enzymes can stimulate cell growth and if the 18-amino acid insert has a role in it, growth curve were carried out on T-Rex 293 cell lines stably expressing Flag—tagged COX-2, COX- 2 de1581-598, COX-1 and COX-1 in5580-598 (Fig 16(a-d)). Three separate experiments were conducted and similar results were obtained. No significant difference was observed either between the growths of uninduced and induced cells expressing the COX isoforms or between cells expressing the different COX isoforms. As anchorage-independent growth is considered to be an in vitro test for tumorigenesis, we examined the growth of COX-transfected HEK293 cells in a 70 semisoft agar medium (Table. 4). No difference was found in the size and number of colonies between control cells and cells induced with tetracycline. Also, no difference was observed between different COX isoforms. As stated in Kudo et a1 (2003), cells expressing COX-2 alone formed small colonies when HEK293 parental cells did not grow appreciably in soft agar. In our case, 293 F lp-in cells can also form similar size and number of colonies as COX-2 expressing cells. Since no differences were found in this case, we wonder if the difference was - not significant enough to be observed. In order to magnify the growth difference mPGES was co-expressed with all COX isoforms in the next attempt. Cell Growth Curve and Semi-soft Agar Ass_av for F lag-tagged COX-2, COX-2 de1581-598. COX-l gig COX-l in_3580-598 with mPGES-1. The membrane-bound form of the PGE synthase (mPGES-l) has been previously reported to couple with COX-2, not COX-1, to produce PGEz. PGEz can increase growth and motility of colorectal carcinoma cells through the EP4 receptor signaling pathway (141, 142). Expressing of PGES has also been reported to increase cell growth when co-expressed with COX-2 (134). To confirm these observations and also to evaluate the role of 18- amino acid insert in this process, we transfected mPGES into 293 T-Rex cell lines expressing different COX isoforms (Fig 17). The Flag-tagged COX-2 and COX-2 de1581-598 expressing cells grew faster than the control cells when co-expressed with mPGES (Fig 18(a-b)), probably by increasing PGE; production. On the other hand, even in cell lines constitutively expressing mPGES, no difference was observed in cell growth for COX-1 and COX-1 insS80-598 expressing cells compared to the control cells (Fig l8(c-d)). This could confirm the coupling between COX-2 and mPGES not COX-1. These results were in accordance to the results shown in Kudo et a1 (2003). 71 Although the cells expressing mPGES and COX-2 isoforms grew faster, both Flag-tagged COX-2 and COX-2 de158l—598 had the same phenotype indicating that the 18 amino acids itself may not be involved. A colony-forming test was also done to check the phenotype of cells expressing different COX isoforms (Table 5). According to Kudo et al (2003), COX-2 and mPGES-1 coexpressing cells exhibited marked anchorage-independent growth, as manifested by the appearance of a number of large colonies. On the other hand, cells expressing COX-2 alone or cells co-expressing COX-1 and mPGES-1 formed fewer and smaller colonies (the size and number is about 5 times difference). However the results in our lab showed that no significant difference was found in the colony size and number between uninduced and induced cells expressing the COX isoforms or between cells expressing the different COX isoforms. This assay also confirmed that 18 amino acids were not involved in the cell growth. Immunocytochemical localization of F lag-tagged COX-2, COX-2 de1581-598, COX-2 Y371F, COX-1 and COX-1 in8580-598. To check if the l8-amino acid affects the subcellular localization of COX enzymes, immunocytochemistry of COX isoforms in transfected cells were conducted (Fig 19(a-e)). No differences were observed for any of the COX proteins: similar staining of the perinuclear envelope and ER membrane were observed for all COX isoforms. The 18-amino acid cassette does not seem to regulate the subcellular localization of COX enzymes. Determinption ttunover 111164 of Flag-tagged COX-2, COX-2 de1581-598. COX- 2 N580A. COX-lgandfi COX-1 in3580-598. Stably-transfected HEK 293 TREX cells were cultured and induced for cyclooxygenase expression by adding 211g/ml ' tetracycline for 24 hrs. The cells were then incubated with the medium containing 72 300 250 200 150 100 cell number(x104) 50 c011 numhcr(x104) 1000 800 600 400 200 cell number(x101) 0 (1 2 4 6 8 10 days Figure l6((a)-l-3). Growth curve for 293 T-Rex cell lines expressing F lag-tagged COX-2 (Solid line is tetracycline induced cells and dash line is cell control. 1)( 10 or 5X10 4 cells were inoculated at day 0). Results are shown in three independent experiments. 73 cell number(x104) cell number(x104) l()()() ti()() 1i()() Z100 cell number(x10') 22()() (lery's Figure 16((b)-1-3). Growth curve for 293 T-Rex cell lines expressing Flag-tagged COX-2 de1581-598 (Solid line is tetracycline induced cells and dash line is cell control. 1X10 4 or 5X10 4cells were inoculated at day 0). Results are shown in three independent experiments. '741 250 200 cell number(x104) cell number(x104) 1000 800 600 400 cell number(x104) 200 days Figure 16((c)-l-3). Growth curve for 293 T-Rex cell lines expressing Flag-tagged COX-1 in3580-598 (Solid line is tetracycline induced cells and dash line is cell control. 1X 10 4 or 5X 10 4 cells were inoculated at day 0). Results are shown in three independent experiments. 75 300 [\3 01 O 200 150 100 cell number(x10 I) 300 250 200 150 100 cell number<10 4 or 5X10 4 cells were inoculated at day 0). Results are shown in three independent experiments. 79 1600 1400 1200 1000 800 600 400 200 cell number(x104) 1400 1200 1000 800 600 400 200 0 cell numhcr(x104) 1600 1400 1200 1000 800 600 400 200 0 cell number (x10 4) 0 2 4 6 8 l 0 days Figure 18((c)-1-3). Growth curve for 293 T-Rex cell lines expressing COX-1 insS80-598 and mPGES (Solid line is tetracycline induced cells and dash line is cell control. 1>< 10 4 or 5X10 4 cells were inoculated at day 0). Results are shown in three independent experiments. 80 1600 1400 1200 1000 800 600 400 200 cell number(x104) l6(Xl 1400 1200 1000 800 600 400 200 0 cell number(x104) coll numhcr(x104) 0 2 4 6 8 10 days Figure 18((d)-1-3). Growth curve for 293 T-Rex cell lines expressing COX-1 and mPGES (Solid line is tetracycline induced cells and dash line is cell control. 1X10 4 or 5X10 4 cells were inoculated at day 0). Results are shown in three independent experiments. 81 COX Alone COX + mPGES Cell lines Colonies SIZE Colonies SIZE (mm) (mm) No tetracycline Flag-tagged 44 0.35 65 0.35 COX-2 cells only control Tetracycline induced 50 0.46 70 0.35 Flag-tagged COX-2 expressing cells No tetracycline F lag-tagged 40 0.28 52 0.42 COX-2 del581-598 cells only control Tetracycline induced Flag-tagged 38 0.35 48 0.37 C OX-2 de1581-598 expressing cells No tetracycline COX-1 cells only 40 0.26 38 0.33 control Tetracycline induced COX-1 46 0.23 42 0.28 expressing cells Table 5.Colony forming Agar assay for COX expressing transfectants. Stably transfected Flp-1n 293 cells line expressing different COX mutants and the prostaglandin E synthase were tested for their ability to grow in the soft agar. Each experiment was repeated three times and similar results were obtained. 5011m cycloheximide for 0, 2, 4, 6, 8, 12, 24 hrs before harvesting. Expression of COX was determined by Western blotting and expression levels were quantitated by analyzing the band densities using Scion Image software (Fig 20(a-f)). Tm was calculated by integrated first-order velocity equation; 2.3log [S] o/[S] =kt (Table 6). 82 (d) (6) Figure 19(a-e). Immunocytochemistry of Flp-In 293 cells line stably transfected with Flag-tagged COX-2(a), COX-2 de158l-598 (b), COX-2 Y371F(c), COX-1(d), COX-1 ins580-598 (e). Anti-F LAG antibody was used as the primary antibody and FITC-conjugated goat anti—mouse IgG was used as the secondary antibody. Stained cells were then examined under the microscope. Three experiments were conducted, similar results were obtained and average Tm was calculated. The Tm for the COX-1 and COX-1 in5580—598 were both over 24 hr. COX-2 N580A, COX-2 de1581-598, and COX-2 Y371F (the inactive COX-2 enzyme) had the similar Tug compared to Flag-tagged COX-2. Since the entire turnover rates of COX-1 and COX-2 mutants were similar to each other respectively, this indicated that the activity and the 18 amino acids may not be important in the degradation process. Determination turnover rates of untagged COX—2. COX-2 del58l-598, COX-2 N580A COX-1 and COX-1 ins580-598. Results in Smith’s laboratory indicated (Uri’ thesis) that the FLAG tag at N-terminus affected the turnover rate of COX enzymes, so the half-lives of untagged proteins were next determined. Stably—transfected HEK eg'v. wanna-m “mm-311mm wanna. Am lfi m a... ._.n , f" COX-2 1" 1. E Egan-‘1'. - . .‘ ‘ - -I‘ ":51 i W4 4 . 1 . Actin T1/2=2.9h Fig 20(a). Tm for Flag-tagged COX-2 U024681224 COX-2 de1581-598 Actin T1/3=4.5h Fig 20(b). Tm for Flag-tagged COX-2 de158l-598 U024681224 a“ .' "T. » (in? Sfi'lmaikrfiii ‘fifim‘thbfifis’fi?a'lsl J's. 31.11} V . .,, . *1 . 't 9. M, , COX-1 E4: . , , . a . _ . -1.‘ Hannfnaxrr. .--:.?.>. r -1 2.11111 ' '1 '41:? ‘ "41."; «.1 9." -1“. ‘ .1 ~. "he ’15 "1. 81r'l'-"~.‘s'l ' ix M A ' 3" Ctln . 1:. ~ .. . .p’ . .n :‘u “ 4 ... ‘ . .. .. .1 E: . is“ a”): ‘nx’v. 1.47.1."- . ‘ L‘L' 3*.- ~ n :AL' ....‘.£ T1/2>24h Fig 20(c). Tm for Flag-tagged COX-l U024681224 4?.”‘l4—‘L' ‘ . ." '71 r r 4.4 , ,, . 4" E i- .‘s . 1; . ~ ,;,-d-.u ... 1" ‘. - . ‘13. L .. ,1-..“ -- g. _ . A. . . .—. i! COX-l in3580-598 .Ln‘mgw _.._._ n—um a... a"fl."$‘~.‘a “bw— ‘Inu\ 2.. .--.. ...-...»- . ~ , .' _. -... ‘-. '.‘ ..-. ",.,.:.«‘ ,‘P c ‘ . , ,. . .,y. ,s .. 4. 1 W Actin 1‘: ' 1..., . . . . m --- ,c . up ,.- . ~r~.' .u. finihsfl‘bafiwall‘firkh .- nu‘ ‘ ;~ 1"‘1'1. .r-r u“ .kct- r1"! T1/2>24h Fig 20(d). Tm for Flag-tagged COX-l ins580-598 84 Fig 20 (cont’d) U024681224 -4_-,._-P...._‘..,.. .,...._..‘,.. “1...? "“7"“ . - K's" .‘..a' ' _...: m :5'3'15. -.P.. S ‘. .‘id'. ‘ ' $11 111-é .. COX-2 N580A WW7” s: a P 1‘. M W '. 5": 1].-..4‘ . '3- — 1 :11; 1,. f‘ j C..- -..-_;.-_-_1__:1.‘; T1/3=2.3h Fig 20(c). Tm for Flag-tagged COX-2 N580A Figure 20(a-e). Determination of Tm for different COX mutants. Samples were loaded onto each lane of a 4-12% NuPAGE Bis-Tris gel and proteins were separated by electrophoresis. After transferring to nitrocellulose membrane, COX isoforms were detected by a specific anti-flag antibody or an anti-COX antibody. From left to right is cell only control and O, 2, 4, 6, 8, 12, 24 hr (after CHX treatment) cell samples from F lp-In 293 cells line stably transfected with Flag-tagged COX-2(a), COX-2 de1581- 598(b), COX-1(c), COX-l ins580-598(d), COX-2 N580A(e) respectively. Actin control was also included. Each experiment was repeated three times and similar results were obtained, average T112 was calculated and presented. 85 U024681224 COX-2 F:a:m1«:.-~.«1:;-.w.m11.21161: 1.11:1: 1:11.:1 11.9.1111: -1.~11. 11 1: 121 h. '1 ‘— '___ w . 0 fl ’ ACtm 1‘. 1 . 11% m. 1 ' i ' , 1 ‘ .1 1- - u , o , ' .51 u 'fi'l' ._. I 7 _ - n! "I r'" T1/2=6.5h Fig 21(a). Tm for COX-2 U024681224 COX-2 de1581-598 w .11.! .ugi.1t_¢'.cms.e-e.-m-.. . J-u-LE.'.F:'E:M‘4'MLHH"flu-Q” unis—11:1. ' - - . .._| 1 , , . . .- a l. I ‘ t. . 1 1;. ~~"~~~wm-M~ 1; Actln -i . |. I; ' T‘h’a ‘Zuuin'f- '-. 1A!!! 717x111. '. 11.1 .-"- '-.'*'.'i"-". -'~.A- L1. .‘z'u i'"_- ’- n... hum”. T1/2>24h Fig 21(b). Tm for COX-2 d21581-598 U 0 2 4 6 8 12 24 ' “11* «PI -‘ “W .111" 3“" ‘11-? 71:1"... ELF-”FA“. agfi COX-2 N58OA :fiY-marnffg “:3... . 7 T 1 1 “1“"; “mafia? -."\ - ‘ ‘ ' 1' '1, 1; CI.” "WM-1W. -- if; Actin T1 /2>24h Fig 21(c). Tm for cox-2 N580A U0 24681224 .,.... .~« ~._. j . r1. . ‘ , . _ 1 Eu“. , , 1 1 '1 ’f 1“ €31.44- -. ‘- ‘ *é.Ef¥~fi5-xi~ 1&5- &.~.\. .. , COX-l ' -. . .. . A 3.311;: l.’~Xt-Eif:§=.'.j.‘:£;“é‘£_.ilfi;zgl.£z.-')L!i.."i.£' .sggc..;1_;.4.;-._nfl Actin T1/2>24h Flg 21(d). Tl/z fOI' COX-1 86 Fig 21 (cont’d) 3’ ‘ ”W M" . COX-1 m3580-598 Actin Tm=21h Fig 21(e). Tm for COX-1 in5580-598 Figure 21(a-e). Determination of Ty; for different COX mutants. Samples were loaded onto each lane of a 4-12% NuPAGE Bis-Tris gel and proteins were separated by electrophoresis. After transferring to nitrocellulose membrane, COX isoforms were detected by a specific anti-flag antibody or an anti-COX antibody. From left to right is cell only control and O, 2, 4, 6, 8, 12, 24 hr (after CHX treatment) cell samples from Flp-In 293 cells line stably transfected with COX-2(a), COX-2 N580A(b), COX-2 de1581-598(c), COX-1(d), COX-1 in5580-598(e) respectively. Actin control was also included. Each experiment was repeated three times and similar results were obtained, average T1/3_WaS calculated and presented. 293 T-Rex cells were cultured and induced for cyclooxygenase expression by adding Zug/ml tetracycline for 24 hrs. The cells were then incubated with the medium containing 50pm cycloheximide for O, 2, 4, 6, 8, 12, 24 hrs before harvesting. Expression of COX was determined by Western blotting and expression levels were quantitated by analyzing the band densities using Scion Image software (Fig 21(a-t)). Tl/z was calculated by integrated first-order velocity equation; 2.3log [S] o/[S] =kt (Table 6). Three experiments were conducted with similar results and average Tm I was calculated. The tagged oCOX and native oCOX-l had similar Tm. Untagged COX-l in3580-598 had a slightly shorter T1 /2 than the tagged protein (21hr verses >24hr). The half life for. Tl/ZCOX-Z was similar to the F lag-tagged COX-2 (2.9hr verses 6.5hr). 87 Surprisingly, COX-2 de1581-598 and COX-2 N580A had a much longer Tug (>24hr) compared to their tagged protein (24hr verses 4.5 and 2.3hr). Overall COX-1 containing 581-598 amino acids had a shorter Tm than COX-1 and Tm of COX-2 delS8l-598 was much longer than the COX-2. Adding the 18 amino acids increased the turnover rate of COX-1, and deleting the 18 amino acids decreased the turnover ' rate of COX-2, suggesting that the l8-amino acids insert promotes degradation. Furthermore COX-2 N580A had a much longer Tm than COX-2 which suggests this C-terminal glycosylation site is also important in the turnover regulation of COX proteins. 88 Protein Tm (hr) TI/z (hr) (AVG) Flag-tagged COX-l ins580- 598 Flag-tagged COX-l Flag-tagged COX-2 Flag-tagged COX-2 de1580- 598 Flag-tagged COX-2 N580A COX-l [[13580-598 COX-l COX-2 COX-2 delS80-598 C OX-2 N580A >24.0, >24.0, >24.0 >24.0, >240, >24.0 3.2, 2.5, 3.0 4.0, 4.5, 5.0 2.8, 1.9. 2.2 21221.9, 19.9 >24.0, >24.0, >24.0 7.5, 6.4, 5.6 >24.0, >24.0, >24.0 >24.0, >24.0, >24.0 >24.0 >24.0 2.9 4.5 2.3 21.0 >24.0 6.5 >24.0 >24.0 Table 6. Half-life of COX proteins. Stably transfected Flp-In 293 cells line expressing different COX mutants were used to test protein tumover rate. Each experiment was repeated three times and average Tl/z was calculated and presented. 89 Discussion To determine what structure or catalytic features of COX-2 may prevent its expression, transfection experiments with different forms of COX protein were carried out and the relative efficiency of isolating clones expressing different mutants was determined. The results indicated that COX-2 had the lowest expression ratio while deletion of the 18-amino acid cassette or inactivation of the enzyme could increase protein expression ratio, indicating that both activity and the 18-amino acid cassette were important for the protein expression. The Flag-tagged COX-2 de1581- 598 had a longer half life (TI/2=4.5h) compared to the native Flag-tagged COX-2 (Tl/2:2.9h), the small difference between protein stability could in part explain the difficult to isolating the native COX-2 expressing colony. Results in Smith’s lab indicated that inactive COX-2 mutant resisted to the AA induced degradation of COX-2 which clarified why COX-2 Y371F had a higher ratio of expression compared to native COX-2 (Smith’s lab, Uri’s thesis). Meanwhile NS398 was also able to increase the expression ratio of the native COX-2. The outcome was much more dramatic than the inactive mutant (COX-2 Y371F) since it increased the ratio from 27% to 92% compared to only 46%. These results indicated additional potential role of NS398 other than simply inhibiting COX- 2. Results in Smith’s lab showed that NS398 could inhibit the substrate induced degradation of COX-2 (Smith’s lab, Uri’s thesis). The possible mechanism is that after binding to the enzyme, the conformation could be altered so the NS398-COX-2 complex is resistant to any other modification. F lag-tagged COX-l ins580-598 had 100% expression in all the colonies, same as Flag-tagged COX-1. The reason that 18 amino acids insert had no influence could due to the fact that both protein had the similar half life (T1;2>24h) because of the Flag 90 tag. To verify the role of 18 amino acids insert in this case, expression ratio of untagged COX-l in3580-598 needs to be determined. Evidence has indicated that NSAIDS can inhibit colon cancer cell growth by inhibiting COX-2 (140). To determine if the COX enzymes can stimulate cell growth and if the 18-amino acid insert has a role in the process, growth curve and colony forming agar assay were conducted. We were unable to detect any difference when cells were expressing COX proteins alone. When COX proteins were co-expressing with the prostaglandin E synthase, our data confirmed the results in Kudo’s laboratory that the coupling happened between COX-2 and mPGES, not COX-l. This coupling led to the increasing PGE; production which in turn can stimulate cell growth. However both F lag-tagged COX-2 and COX-2 delS8l-598 expressing cells (with or without mPGES) had the same phenotype in both experiments indicating that the 18 amino acids-itself may not be involved in cell growth. The role of 18-amino acid cassette in subcellular location has been studied by immunocytochemistry and no differences were observed for any of the COX proteins: similar staining of the perinuclear envelope and ER membrane were observed for all COX isoforms. The results indicated that 18-amino acid cassette does not affect the subcellular localization of the protein When half-life study of COX protein was first studied, the Flag-tagged COX proteins were used and the entire turnover rates of COX-1 and COX-2 mutants were ' similar to each other respectively, suggesting that the 18 amino acids may not be important in the degradation process. However the results from Smith’s laboratory indicated that the N-terminal tag (His tag or Flag tag) may significantly affect the turnover rate of COX enzymes (unpublished data). Since either His or Flag tag was just a small peptide, how they had such a striking effect on the protein degradation 9] remains mysterious. When focusing on the untagged COX protein, results in Smith’s were similar to ours. COX-l containing 581-598 amino acids had a shorter Tm than COX-1 and T1,; of COX-2 de1581-598 was much longer than the COX-2. Since adding the 18 amino acids could increase the turnover rate of COX-1 and deleting the 18 amino acids could decrease the turnover rate of COX-2, it indicated the role of 18 amino acids in the degradation pathway. Furthermore COX-2 N580A had a much longer T1 ,2 than COX-2 which means this C-terminal glycosylation site was also important in the turnover regulation of COX proteins. Subtle difference in the catalytic activities of COX-2, and its different stability and expression profile may explain the different biological functions between the two isozymes. In summary, our results demonstrated that l8-amino acid insert had no effect on cyclooxygenase activity, but reduced the number of colonies that could be detected that stably expressed COX-2. The 18 amino acid insert had no effect on COX-2 protein subcellular location and cell growth. The 19 amino acids at the C-terminus of COX-2 were important in the turnover regulation of COX proteins. The Flag-tag at the N-terminal of COX protein can significantly affect the protein turnover rate by unknown reasons. Whether the Flag-tag also affect other features of COX proteins are not clear yet. To solve this problem, untagged protein or different Tags should be used instead of F lag-tag. 92 CONCLUSION My research focused on identifying putative protein partners for cyclooxygenase proteins by mass spectrometry analysis and determining the role of 18-amino acid cassette on cyclooxygenase cell biology. To fulfill the first goal, different COX proteins (Flag-tagged native COX-2, COX-2 de1581-598, COX-1 and COX-1 ins580-598) were expressed and purified for the proteomic analysis to identify the potential partners. Only two proteins were reproducibly identified, FAM44A (GI/QSNFC6) and Heat shock-induced protein (GI/188492). FAM44A was identified in 5 out of the 7 experiments in protein complexes with COX-2 purified from T-Rex 293 cells, and 1 in 2 of the experiments where COX-2 was expressed transiently in Freestyle 293 cells. FAM44A was also identified in protein complexes with COX-1 ins580-598 but never in the uninduced 293 cells, or in protein complexes with COX-1 or COX-2 de1581-598, suggesting association requiring a specific ‘ interaction with the 18-amino acid cassette of COX-2. FAM44A has been reported as a 330KDa protein. Its cDNA was isolated from human chromosomes 4 (114). FAM44A protein has AT-hook DNA binding domain which is prevalent in many eukaryotic nuclear proteins (116) and FAM44A can be phosporylated during DNA damage. Thus the interaction between COX-2 and FAM44A could be an alternative way to affect cell cycle. Another possibility is that COX-2 may interact with FAM44A for gene transcription regulation. FAM44A may serve as a chaperon for the COX-2-derived prostanoids to import into the nuclear to bind its target receptors for different gene expression regulation. The heat shock-induced protein was another protein that was identified in both T-Rex and Freestyle systems associating with COX-2. It was identified 1 out of the 7 experiments in 93 protein complexes with COX-2 purified from T-Rex 293 cells, and 1 in 2 experiments where COX-2 was expressed transiently in Freestyle 293 cells. Since heat shock-induced protein was not present in complexes with the COX-1 in3580-598 or the COX-2 de1581- 598, the interaction may require COX-2 protein domains in addition to the 18-amino acid cassette of COX-2. Heat shock-induced protein may aid in the processing of COX—2 or be involved in its degradation. Recent work has proven that this protein can inhibit LPS- induced NF-KB signaling cascade activation and subsequently decrease COX-2 expression (140), thus the interaction between COX-2 and heat shock-induced protein may also affect the expression of COX-2. Other proteins like BiP and KIAA0139 were only identified once in all 9 experiments indicating they could be an artifact due to over- expression of COX-2, or an anomaly of the isolation in a specific experiment. No unique proteins were ever identified for COX-2 de1581-59in the T-Rex cells. The identification of proteins that specifically co-purified with COX-2 and not with COX-2 de1581-598 suggests that the 18 amino acids may be important in protein-protein interaction. No interesting proteins were ever identified for COX-l in the T-Rex cells. Also no protein partners were identified in common for COX-1 and COX-2. Since COX-1 and COX-2 have very different biological functions, it is reasonable to assume that the specific interaction identified in COX-2 may cause different functions between COX isozyme. Among the proteins found in the COX-1 ins580-598 construct, FAM44A protein was the most probable protein partner candidate. As stated above, this protein was also co-purified with COX-2 complex and may association with COX-2 specifically via the 94 18-amino acid cassette. Although several other proteins were also identified, none of them were co-isolated for both COX-1 and COX-1 ins580-598. Thus, these proteins could just be the artifacts of over-expression. To determine the role of 18-amino acid cassette on cyclooxygenase cell biology, _ the following aspects were checked: expression in 293 Flp-in cells, cell growth, subcellular localization and protein stability. The ratio of expression results indicated that COX—2 had the lowest expression ratio while deletion of the lS-amino acid cassette or inactivation of the enzyme could increase protein expression ratio, indicating that both activity and the 18-amino acid cassette were important for the protein expression. That COX-2 activity has a negative effect on cell growth was also confirmed by addition of a COX-2 specific inhibitor NS398 during the transfection process We were able to confirm the coupling between COX-2 and mPGES, not COX-1. However both Flag-tagged COX-2 and COX-2 de1581-598 expressing cells had the same phenotype indicating that the 18 amino acids itself may not be involved in cell growth. No differences were observed for any of the COX proteins subcellular localization: similar staining of the perinuclear envelope and ER membrane were observed for all COX isoforms. The results indicated that l8-amino acid cassette does not affect the subcellular localization of the protein The results in our lab confirmed the insertion of the l8-amino acid cassette into COX-1 can destabilize the protein while deletion of the l8-amino acid cassette in COX-2 stabilized the protein, indicating the role of 18 amino acids in the degradation pathway. 95 Furthermore COX-2 N580A had a much longer Tm than COX-2 which means this C- terrninal glycosylation site was also important in the turnover regulation of COX proteins. The explanation for why not many reliable candidates were identified could be due to the followings: the interactions are transient, or that the interaction of these proteins have low affinity and do not survive purification. Furthermore 293 cells do not normally express endogenous COX-2, and may not express protein partners for COX-2. Another problem we are having is the high background of non-specific (isolated and identified in both COX expressing and control samples) proteins. We assume that they could non-specifically bind to anti-flag agrose beads during purification. It is possible that when these proteins bind to the agrose beads it may affect even block the binding of potential protein partners. Besides the signals of these proteins are so intense, it may mask the signal from the real protein partners which is less redundant during the mass analysis, thus miss the identification of protein partner. To overcome these problems, first of all new cell should be used instead of the 293 cells. The cells should endogenously express COX isozymes to guarantee that potential protein partners exist. Secondly, new tag instead of Flag-tag should be used on COX protein for purification. When the protein complexes were purified with F lag-tag, the mass date contains high background because many non-specific proteins were attaching to the Flag agrose beads. Another problem with the Flag-tag is that it significantly affected the turnover rate of COX enzymes (Smith’s lab, unpublished data). Since Flag-tag was just a small peptide, how it had such a striking effect on the protein degradation remains mysterious. We already proved that Flag-tag would not affect the catalytic activity of COX-2. Whether 96 the Flag-tag can affect other features of COX proteins is not clear yet. To solve this problem, untagged protein or different Tags should be used instead of Flag-tag. 97 APPENDIX Table 7. Plasmids constructed by QuicltChangeTM site-directed mutagenesis PARENT PRODUCT PRIMERS SEQUENCES PLASMID PLASMID 5 ’GCGCTCAGCCATACAGCAAATTGCGAGAACTTA , TACTTTCAGGGACCTTGCTGTTCCCACCCATGT 3’ hCOX-2- Flag-tagged pFastBac COX-2- pFastBac 5’ACATGGGTGGGAACAGCAAGGTCCCTGAAAGTA TAAGTTCTCGCAATTTGCTGTATGGCTGAGCGC 3 ’ 5’AGCCATACAGCAAATGATTACAAAGACGATGAC F lag- FLAG-TEV GATAAGTGCGAGAACTTATACTTTCAG 3’ tagged ' COX-2- hCOXZ- pFastBac 5’CTGAAAGTATAAGTTCTCGCACTTATCGTCATCGT pFastBac CTTTGTAATCATTTGCTGTATGGCT 3’ 5 ’ACAGTCACCATCAATGAACGTTCGAC 3 ’ FLAG- FLAG- TEV- TEV- S’CTACAGTTCAGTACGTTCATTG 3’ COX-Z- . COX-2 pFastBac de1581-598- pFastBac 5’CGGGGCGCCCCGGGCTTATCGTCATCGTCTTTGTA His-tagged Flag-tagged ATCGTCCGCTGAGAAGACG 3’ COX-l- COX-1- pIND pIND 5’TCTTCTCAGCGGACGATTACAAAGACGATGACGATA AGCCCGGGGCGCCCGCGC3 ’ 98 Table 7 (cont’d) His- Flag- 5 ’CGGGGCGCCCCGGGCTTATCGTCATCGTCTTTGTA tagged tagged ATCGTCCGCTGAGAAGACG 3’ ins580- insS80- 598 598 5’TCTTCTCAGCGGACGATTACAAAGACGATGACGAT COX-1 COX-l AAGCCCGGGGCGCCCGCGC3’ FLAG- FLAG- 5’CATTAAAACAGTCACCATCGCTGCAAGTTCTTCCCG TEV TEV- CTCC 3’ COX-Z- N580A- pCDNAS hCOX-Z- 5 ’GGAGCGGGAAGAACTTGCAGCGATGGTGACTGTTT pCDNAS TAAT 3’ FLAG- , FLAG- 5’TTAACACCCTCTTCCACTGGCATCCCCTTC 3’ TEV TEV- COX-Z— Y371F- 5’ GGGATGCCAGTGGAAGAGGGTGTTAAATTC 3’ pCDNAS hCOX-Z- pCDNAS FLAG- 5’CAGCCATACAGCAAATCCTTGCTGTTCCCACC 3’ TEV hCOX-Z- COX-Z- ‘pCDNAS 5’ GGTGGGAACAGCAAGGATTTGCTGTATGGCTG 3’ pCDNAS FLAG- COX-2 5’CAGCCATACAGCAAATCCTTGCTGTTCCCACC 3’ TEV de1581- de1581- 598- 5’ GGTGGGAACAGCAAGGATTTGCTGTATGGCTG 3’ 598 pCDNAS COX-2- pCDNAS FLAG- N580A- 5’CAGCCATACAGCAAATCCTTGCTGTTCCCACC 3’ TEV hCOX-Z- N580A- pCDNAS 5’ GGTGGGAACAGCAAGGATTTGCTGTATGGCTG 3’ hCOX-Z- pCDNAS 99 Table 7 (cont’d) pOSML Flag-tagged 5’TTCTCAGCGGACCCCGGGGCGC 3’ COX-l- COX-1- pCDNAS pCDNAS 5’CGCCCCGGGGTCCGCTGAGAAG 3’ Flag-tagged insS80-598 5’ TTCTCAGCGGACCCCGGGGCGC 3’ insS 80-598 COX-1- COX-l- pCDNAS S’CGCCCCGGGGTCCGCTGAGAAG 3’ pCDNAS PARENT PLASMID PRODUCT RESTRICTION . PLASMID SITE F lag-tagged COX-2 pFastBac Flag-tagged COX-Z- Not I pCDNAS F lag-tagged COX-2 de1581- COX-2 de1581-598- Not I 598-pFastBac pCDNAS Flag-tagged COX-l-pIND Flag-tagged COX-l - Hind III pCDNAS Flag-tagged COX-l insS 80— Flag-tagged COX-1 Hind III 598-pIND insS80-598-pIND -pCDNA5 Flag-tagged COX-2-pCDNA5 F lag-tagged COX-2- Not I Table 8. Plasmids constructed by subcloning 100 PCR-mediated mutation-Quick change PCR-mediated introduction of mutations, inserts or deletions used either: 200ng and 20ng template vector, 5111 10 x PCR reaction buffer (Stratagene), 2ul 10mM dNTP, 1 pl polymerase (Strategene), 125 ng each primer and water was added to a total volume of 50111. PCR conditions were adjusted for each prime set. Starting conditions were: 95°C for 1 minute, and 20-40 cycles of 95°C for 30 second, 50°C for 1 minute, and 68°C for various minutes depending on the size of plasmid (2 min/kb). A final extension step was employed at 68°C for 10 minutes and the reaction was lowered to 4°C. Linker-Restriction Site Insertion To change a restriction site, a single complementary oligo containing the new restriction site and a complementary to restriction site were ligated into cognate restriction digested plasmid. The self complementary oligo was first dissolved into Tris- EDTA containing 100mM NaCl at a concentration of 1mg/ml. The oligo was then heated to 80°C of water and allowed to cool to room temperature to from a double strand dimer. The hybridized oligo was stored at ~20 °C. A typical ligation included In] hybrid oligo (lug/til), lul 10x ligation buffer (New England), 1111 plasmid (lOOng/ul) digested at complementary site, 11.11 T4 ligase (New England) and H20 to 101.11. The reaction was incubated at room temperature overnight, transformed the next day and analyzed for new restriction site by digestion of mini-prep samples of recombinant plasmids. 101 Figure 22. Map of COX-2 cloned into pFastBac (Invitrogen) at Not I site ( htto: // invitrogen.com/content/ sf s/vectors/pfastbac 1 mappdf) 102 Nhe I Pme I Afl II Hind III Asp718 l Kpn I Sac I BamH I Spe I BstX I EcoR I EcoR V BstX I Not I Xho I Xba I Apa I Pme I DUC ori Figure 23. Map of His tagged-COX-l cloned into pIND (Invitrogen) at Hind III site (http://invitrogen.com/content/sfs/vectors/pind mappdf) 103 Ec00109 I Apa I [me I E X FCDNAS/FRT/TOQ g 5 9.5 I" - Id DUC ori Figure 24. Map of Flag-tagged COX-2 and Flag-tagged COX-2 delS8l-598 subcloned into pcDNAS/FRT/TO (Invitrogen) vector from pFastbac at Not I site. Flag-tagged COX-1 and Flag-tagged COX-1 ins580-598 subcloned into pcDNAS/FRT/TO vector from pIND at Hind III site. (http://invitrogen.com/content/sfg/vectors/pcdn35frtto mm 104 50M) SV40 ori A Q 1000 NI 4000 ~ P J Po’yionk?’ cloning Site pOSML-l (1055-1098) (pMT2+polylinker) Pst I, Xho I, Kpn 5005 bp 1, Sal I, Xba I, 9 Not 1 and EcoR I Figure 25. Map of Flag-tagged COX-2 subcloned into pOSML vector from pcDNAS/FRT/TO at Not I site. (http://invitrogen.com/content/sfs/vectors/mdnaSfrtto magmt) 105 Figure 26. Map of pOG44 vector expressing Flp recombinase (https://www.invitrogen.com/content/sfs/vectors/pog44 mappdf) 106 _’ Expression of lacZ and Zeocin Psv40 AMP pUC orl ATG >‘I IacZ-Zeocin" FRT + pOG44 + pcDNASIFRT —-‘ . AMP IUC I / ”V40 ATG >l IacZ-Zeocm" or / I Expression Vector / / \ I pcDNAS/FRT V I POW \\_‘// Expression of your gene xpression of hygromycin ATG Hygro GOI LacZ-Zeocin Figure 27. Flp-In-T-Rex Tetracycline Inducible System. The Flp-In T-Rex system exhibit tetracycline-inducible gene expression after specific integration of the transfected gene into the genome via F lp recombinase-mediated DNA recombination. (www.invitrogen.com/content/sfs/manuaIs/flpinsystem manpdf ). 107 ti ——*ét» tléitnermivrlmivr i‘.’flill"‘tilli‘ivilfi ’* w W \\ \\ \'\ \\ \\ “i5. \\ \ Figure 28(a). AP inducible expressing system. The transcription of a target gene under the control of a small molecule “dimerizer”. htt ://www.ariad.com/ df/Re Tx-Retrovirus. f#search=%22AP%20inducible °n7nemres inrt%203vstem%20Ariad0/020Ph... " ‘ "n20%22). : ”'F‘I‘V’JL’ Figure 28(b). AP21967. It is a chemically modified derivative of rapamycin that can be used to induce heterodimerization of FKBP and FRB1-309XL -containing fusion proteins. 108 12 ZFH DI . binding sites Polylmker IL-2 promoter \ m / Figure 28(c). Target gene vector pLH—ZnI-PL. Vector contains 12 ZFHDI binding sites and a minimal human interleukin-2 gene promoter (lel), upstream of a polylinker (PL) inserted downstream of an LTR-driven hygro resistance gene (Hygro'). Insertion of the gene of interest into the polylinker places its expression under control of the dirnerizer-regulated transcription factors. NLS Stop NLS Stop Psve rkalpss I nsr ' zmn man"l T - IRES LTR Activation domain DNA binding domain Neo'r LTR ' L__T__J Figure 28(d). Transcription factor vector pLzNz-RHS3H/ZF3. An activation domain fusion (RHS3H) which contains the FRB fragment of human FRAP (RH), fused to a highly potent chimeric activation domain called S3H. S3H consists of amino acids 281 to 551 from the p65 subunit of human NFKB (S3) and amino acids 406-530 from human heat shock factor 1(H). The FRB domain consists of amino acids 2021-2113 of FRAP, in which the threonine at amino acid 2098 is mutated to leucine. This mutation allows the protein to bind to rapamycin analogs (e. g. AP21967) which no longer bind appreciably to endogenous FRAP. A DNA binding domain fusion (ZF3) which consists of the ZFHDI DNA binding domain (Z) and three tandemly repeated copies of human FKBP12 (F3). 109 Beetle Luciierin Firefly Luciferase p. Mg2r UN) (ST + AMP + PPl + CO2 + Light Oxyluciierin Figure 29. Bioluminescent reaction catalyzed by firefly luciferase. (http://www.p_romega.com/tb§/tb281/tb281.Ddfi 110 pOTB7 181 5 bp Tet promoter Polylinker r—‘vg‘ Figure 30. Map of vector pOTB7 that was source of mPGES (Invitrogen). cDNA made by oligo-dT primimg. Directionally cloned into EcoR l/Xho 1 sites using the following 5’ adaptor: GGCACGAG (G) (www.rzpd.de/info/vectorjs/DOTB7_pic.shtml) lll 71 N :3 m 3 ~< 2. a Figure 31. Map of mPGES cloned into Blunt TOPO 11 after PCR. (http://www.invitrogen.com/content/sfs/manuals/zeroblunttopo manflf) 112 PstI (6) Seal (4735)., XbaI (184) Ford (4625)\_: hi7; 5'er FspI (4477), /\Kpn1 (371) AmpRM’ "/ Mutant splice donor NotI (4134)" " 1. ‘Z‘ A PstI (878) pBéfigEliuro 21;, PstI (1054) [/f % Truncated gag ORIM/ ““ BamHI (1356) KpnI(2935),-_h 3.22.. .1 13710310370) 3'LTR 34L \ 2 EcoRI (1380) XbaI (2748) 2,:SalI(1398) lacs ’SV40IEP Puro F igure 32. Map of mPGES cloned into pBABE-puro vector at EcoR I Site 113 BIBLIOGRAPHY 1. Smith WL, DeWitt DL and Garavito RM 2000 Annu Rev Biochem 69:145-82 2. Castellone MD, Teramoto H, Williams BO, Druey KM and Gutkind JS 2005 Science 31021504 3. Kao HW, Hyman PE, Finn SE and Snape W] Jr 1988 Am J Physiol Gastrointest Liver Physiol 255:G807-G812, 4. Raud J, Dahle'n SE, Sydbom A, Lindbom L and Hedqvist P 1989 Inflammation Research 26:42-44 5. Narumiya S, Sugimoto Y and Ushikubi F 1999 Physiol Rev 79:1193-1226 6. Schaefer M, Hofmann T, Schultz G and Gudermann T 1998 Proc Natl Acad Sci 95: 3008-13 7. Negishi M, Ito S and Hayaishi O 1989 J. Biol. Chem 264(7):3916-3923 8. Andres DA, Dickerson [M and Dixon JE 1990 J. Biol. Chem 265(l.1):5952-5955 9. Xie WL, Chipman JG, Robertson DL, Erikson RL and Simmons DL 1991 Proc Natl Acad Sci 88(7):2692-6 10. Kulmacz RJ 1998 FEBS Lett 430:154—157 11. Smith WL and Langenbach R 2001 J Clin Invest 10721491—1495 12. Yamamoto K, Arakawa T, Ueda N and Yamamoto S 1995 J Biol Chem 270:31315—20 114 13. Subbaramaiah K, Chung WJ, Michaluart P, Telang N and Tanabe T 1998 J Biol Chem 273:21875—82 14. DeWitt DL 1999 Mol Pharmacol 4:625—31 15. Landino LM, Crews BC, Gierse JK, Hauser SD and Marnett L1. 1997 J Biol Chem 272:21565—74 16. Dietz R, Nastainczyk W and Ruf HH 1988 Eu J Biochem 171:321-28 17. Lambeir AM, Markey CM, Dunford HB and Marnett LJ 1985 J Biol Chem 260: 14894—96 18. Lu G, Tsai AL, VanWart HE and Kulmacz RJ 1999 J Biol Chem 274:16162—67 19. Karthein R, Dietz R, Nastainczyk W and Ruf HH 1988 EurJ Biochem 171:313- 20 20. Tsai A, Wu G, Palmer G, Bamba B and Koehn JA 1999 J Biol Chem 274: 21695—700 21. Smith WL, Lands WEM. 1972. Biochemistry] 1:3276—82 22. Tsai A, Kulmacz RJ and Palmer G 1995 J Biol Chem 270:10503—8 23. Wu G, Wei C, Kulmacz RJ, Osawa Y and Tsai AL 1999 JBiol Chem 274:9231—37 24. Tsai A, Hsi LC, Kulmacz RJ, Palmer G and Smith WL 1994 J Biol Chem 269:5085—91 25. Hsi LC, Hoganson CW, Babcock GT, Garavito RM and Smith WL 1995 Biochem Biophys Res Commun 207:652—-60 115 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. Shi W, Hoganson CW, Espe M, BenderCJ and Babcock GT 2000 Biochemistry 39(14):41 12-21 Koshkin V, Dunford HB 1999 Biochim Biophys Acta 1430:341—48 Hemler ME, Lands WEM. 1980 JBiol Chem 255:6253—61 Picot D, Loll PJ, Garavito RM 1994 Nature 367 1243—49 Luong C 1996 Nat Struct Biol 32927-933. Kurubail R 1998 Second International Workshop on Cox-2 DeWitt DL 1999 Molecular Pharmacology 55:625—631 Smith WL and Reike CJ personal communication Greig GM, Francis DA, Falgueyret J -P, Ouellet M, Percival MD, Roy P, Bayly C, Mancini JA and O'Neill GP 1997 Mol Pharmacol 52: 829-838 Gierse JK, McDonald 1], Hauser SD, Rangwala SH, Koboldt CM and Seibert K 1996 JBiol Chem 271:15810—15814 Guo Q, Wang LH, Ruan KH and Kulmacz RJ 1996 J Biol Chem 271:19134—19139 Kurumbail RG, Stevens AM, Gierse J K, McDonald JJ, Stegeman RA, Pak JY, Gildehaus D, Miyashiro JM, Penning TD, Seibert K, Isakson PC and Stallings WC 1996 Nature London 384:644-648. Kraemer SA, Meade SA and DeWitt DL 1992 Arch Biochem Biophys 293:391—400 Xu XM, Tang JL, Chen X, Wang LH and Wu K 1997JBioI Chem 272(11): 6943-50 116 40. Xie W and Herschman HR 1996 J Biol Chem 271:31742—48 41. Subbaramaiah K, Chung WJ, Michaluart P, Telang N and Tanabe T 1998 J Biol Chem 273:21875—82 42. Morris JK and Richards JS 1996 J BiolChem 271:16633—43 43. Gietz RD, Schiestl RH, Willems AR and Woods RA 1995 Yeast 11:355-60 44. Roshak AK, Jackson JR, McGough K, Chabot-Fletcher M and Mochan E 1996 J Biol Chem 271(49):31496-501 45. Crofford LJ, Tan B, McCarthy CJ and Hla T 1997 Arthritis Rheum 40:226—36 46. Miller C, Zhang MK, He YL, Zhao J and Pelletier JP 1998 J Cell Biochem 69:392—41 3 47. Jobin C, Morteau O, Han DS, Balfour-Sartor R 1998 Immunology 95:537—43 48. Schmedtje JF Jr, J i YS, Liu WL, DuBois RN, Runge MS 1997 J Biol Chem 272:601—8 49. Gallois C, Habib A, Tao J, Moulin S and Maclouf J 1998 J Biol Chem 273(36):23183—90 50. Ghosh S, May MJ and Kopp EB 1998 Annu Rev Immunol 16:225—60 51. Zhang FX, Kirschning CJ, Mancinelli R, Xu XP and Jin Y 1999 J Biol Chem 274:7611—14 52. Su B and'Karin M 1996 Curr Opin Immun018:402—1 1 117 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. Wallach D,Varfolomeev EE, Malinin NL,Goltsev W and Kovalenko AV 1999 Annu Rev Immunol 17:331—67 Xie W and Herschman HR 1995 J Biol Chem 270:27622—28 Subbaramaiah K, Chung WJ and Dannenberg AJ 1998 J Biol Chem 273:32943-49 Guan ZH, Buckman SY, Pentland AP, Templeton DJ and Morrison AR 1998 J Biol Chem 273:12901—8 Scherle PA, Jones EA, Favata MF, Daulerio AJ and Covington MB 1998 J Immunol 161:5681-86 Dean J L, Brook M, Clark AR and Saklatvala J 1999 J Biol Chem 274:264—69 Adderley SR and Fitzgerald DJ 1999 J BiolChem 274:503 8—46 Ridley SH, Dean J L, Sarsfield SJ, BrookM and Clark AR 1998 FEBS Lett 439:75—80 Guan Z, Baier LD and Morrison AR 1997 J Biol Chem 272:8083—89 Guan ZH, Buckman SY, Miller BW, Springer LD and Morrison AR 1998 J Biol Chem 273:28670—76 Sheng B, Williams CS, Shao J, Liang P and DuBois RN1998 J Biol Chem 273:22120—27 Ridley SH, Sarsfield SJ, Lee JC, Bigg HF and Cawston TE 1997 J Immunol 158:3165—73 Buckman SY, Gresham A, Hale P, Hruza G and Anast J 1998 Carcinogenesis 19:723—29 118 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. Tanabe T and Tohnai N 2002 Prostaglandins Other Lipid Mediat 68:95-114 Chen CY and Shyu AB 1994 Molecular and cellular biology 14(12):8471-8482 DeWitt DL and Meade EA 1993 Arch Biochem Biophys 306(1):94-102 Chen CY and Shyu AB 1995 Trends in biochemical sciences 20(11):465470 Chen CY, Gherzi R, Ong SE, Chan EL, Raijmakers R, Pruijn GJ, Stoecklin G, Moroni C, Mann M and Karin M 2001Cell 107(4):451-464 Mukherjee D, Gao M, O'Connor JP, Raijmakers R, Pruijn G, Lutz CS and Wilusz J 2002 J EMBO 21(1-2):165-174 Stoecklin G, Mayo T and Anderson P 2006 EMBO reports 7(1):72-77 Dixon DA, Tolley ND, King PH, Nabors LB, McIntyre TM, Zimmerman GA and Prescott SM 2001 J clinical investigation 108(11):1657-1665 Mbonye UR, Wada M, Rieke CJ, Tang HY, DeWitt DL and Smith WL 2006 JBiol Chem 281: 35770—35778 Figueiredo-Pereira ME, Li Z, Jansen M and Rockwell P 2002 J Biol Chem 277(28): 25283-25289 Sugimoto Y, Narumiya S, and Ichikawa A 2000 Prog Lipid Res 39:289-314 Langenbach R 1995. Cell 832483-492 Lim H 1997 Cell 91:197-208 119 79 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. Dinchuk JE 1995 Nature 378:406-409 Komhoff M 2000 Kidney Int. 57:414-422 Loftin CD 2001 Proc Natl Acad Sci USA 98: 1059-1064 Charkraboty I, Das SK, Wang J, Dey SK 1996 J.Nol. Endocrine] 16: 197-208. Langenbach R, Lofiin C, Lee CA, Tiano H 1999 Biochemical Pharmacology 58: 1237-1246 Liu CH, Chang SH, Narko K, Trifan OC, Wu MT, Smith E, Haudenschild C, Lane TF, Hla T 2001 J Biol Chem 276: 18563-18569 Giardiello FM, Offerhaus GJ, DuBios RN 1995 EurJ Cancer 7-8: 1071-1076 Oshima M, Dinchunk JE, Kargman SL, Oshima H, Hancock B, Kwong E, Trzaskos JM, Evans JF, Taketo MM 1996 Cell 87:4803-809 Chinery R, Robert JC, Ramona G, Kirkland SC, Stephanie CS, William EZ, John A0, and Morrow JD 1999 Cancer Res 59: 2739-2746 Chinery R, Beauchamp RD, Shyr Y, Kirkland SC, Coffey RJ, and Morrow JD 1998 Cancer Res 58: 2323-2327 Castano E, Bartrons R, Gil J Pharmacol Exp Ther. 2000 293:509-13 Tsuj ii M and DuBois RN Cell 1995 83(3):493-501 Bakhle YS Brit J Pharm 2001 134:1137-1150 DuBois RN, Gupta R, Brockman J, Reddy BS, Krakow SL, Lazar MA 1998 Carcinogenesis. 19:49-53 93. He TC, Chan TA, Vogelstein B and Kinzler KW 1999 cell 99:335 94. Chulada PC, Thompson MB, Mahler JF, Doyle CM, Gaul BW, Lee C, Tiano HF, Morham SG, Smithies O and Langenbach R 2000Cancer Res 60: 4705-8 95. Kitamura T, Kawamori T, Uchiya N, Itoh M, Noda T, Matsuura M, Sugimura T and Wakabayashi K 2002 Carcinogenesis 23: 1463-6 96. Murakami M, Naraba H, Tanioka T, Semmyo N, Nakatani Y, Kojima F, Ikeda T, Fueki M, Ueno A, Oh S and Kudo 12000 J Biol Chem 275: 32783—32792 97. Jakobsson PJ, Thorén S, Morgenstem R and Samuelsson B 1999 Proc Natl Acad Sci USA 96(13): 7220—7225. 98. Huang M, Stolina M, Sharma S, Mao JT, Zhu L, Miller PW, Wollman J, Herschman H and Dubinett SM 1998 Cancer Res 58(6):1208-16 99. Yoshimatsu K, Golijanin D, Paty PB, Soslow RA, Jakobsson PJ, DeLellis RA, Subbaramaiah K and Dannenberg AJ 2001 Clin Cancer Res 7:3971-3976 100. Abramovitz M, Wong E, Cox ME, Richardson CD, Li C and Vickers PJ 1993 Eur J _ Biochem 215(1):]05-11 101. Mancini JA, Abramovitz M, Cox ME, Wong E, Charleson S, Perrier H, Wang Z, Prasit P and Vickers PJ 1993 FEBS Lett 318(3):277-81. 102. Turini ME and DuBois RN 2002 Annu Rev Med 53:35. 103. Moncada s 1999 .1 R Soc Med 92:164. 104. Kim SF, Huri DA and Snyder SH 2005 Science 310(5756):l966-70 121 105. Neuss H, Huang XH, Hetfeld BKJ, Deva R, Henklein P, Nigam S, Mall JW, Schwenk W and Dubiel W 2007 J Mol Med 85: 961-970 106. Simmons DL, Botting RM and Hla T 2004 Pharmacol Rev 56:387—437 107. Tsai A, Palmer G, Xiao G, Swinney DC and Kulmacz RJ 1998 J Biol Chem 273:3888-3894 108. Noden PA, Smith WL, DeWitt DL and Roux JF 19813iol Reprod 24(5):1042-7 109. Otto JC, DeWitt DL and Smith WL 1993 J Biol Chem 268:18234—18242 1 10. Pollock R, Issner R, Zoller K, Natesan S, Rivera VM and Clackson T 2000 Proc Natl Acad Sci USA 97(24):13221-6. 1 11. Zhang H, Wu W, Du Y, Santos SJ, Conrad SE, Watson JT, Grammatikakis N and Gallo KA 2004 J Biol Chem 279(19):]9457-63 112. Ting J, Lee AS 1988 DNA 7: 275-286 1 13. Milner CM and Campbell RD 1990 Immunogenetics 32: 242-251 114. Hillier LW, Graves TA, Fulton RS, Fulton LA, Pepin KH, Minx P, Wagner McPherson C, Layman D, Wylie K, Sekhon M, Becker MC, Fewell GA, Delehaunty KD, Miner TL, Nash WE, Kremitzki C, Oddy L, Du H, Sun H, Bradshaw-Cordum H, Ali J, Carter J, Cordes M, Harris A, Isak A, van Brunt A, Nguyen C, Du F, Courtney L, Kalicki J, Ozersky P, Abbott S, Armstrong J, Belter EA, Caruso L, Cedroni M, Cotton M, Davidson T, Desai A, Elliott G, Erb T, Fronick C, Gaige T, Haakenson W, Haglund K, Holmes A, Harkins R, Kim K, Kruchowski SS, Strong CM, Grewal N, Goyea E, Hou S, Levy A, Martinka S, Mead K, McLellan MD, Meyer R, Randall-Maher J, Tomlinson C, Dauphin— Kohlberg S, Kozlowicz-Reilly A, Shah N, Swearengen-Shahid S, Snider J, Strong 122 JT, Thompson J, Yoakum M, Leonard S, Pearman C, Trani L, Radionenko M, Waligorski JE, Wang C, Rock SM, Tin-Wollam AM, Maupin R, Iatreille P, Wendi MC, Yang SP, Pohl C, Wallis JW, Spieth J, Bieri TA, Berkowicz N, Nelson JO, Osborne J, Ding L, Meyer R, Sabo A, Shotland Y, Sinha P, Wohldmann PE, Cook LL, Hickenbotham MT, Eldred J, Williams D, Jones TA, She X, Ciccarelli FD, Izaurralde E, Taylor J, Schmutz J, Myers RM, Cox DR, Huang X, McPherson JD, Mardis ER, Clifion SW, Warren WC, Chinwalla AT, Eddy SR, Marra MA, Ovcharenko I, Furey TS, Miller W, Eichler EE, Bork P, Suyama M, Torrents D, Waterston RH and Wilson RK 2005 Nature 434(7034):724-31 115. Porter 1M, McClelland SE, Khoudoli GA, Hunter CJ, Andersen J S, McAinsh AD, Blow JJ and Swedlow JR 2007 J Cell Biol 179(2): 1 87-97. 116. Aravind L and Landsman D 1998 Nucleic Acids Res 26(19):4413-21. 117. Onate SA, Prendergast P, Wagner JP, Nissen M, Reeves R, Pettijohn DE and Edwards DP 1994 Mol Cell Biol 14:3376-3319 118. Falvo JV, Thanos D and Maniatis T1995 Ce1183:1101-1111 119. Girard E, Bello B, Laemmli UK and Gehring WJ 1998 EMBOJ 17:2079-2085 120. Rasala BA, Orjalo AV, Shen Z, Briggs S and Forbes DJ 2006 Proc Natl Acad Sci USA 103(47):17801—17806 121. Gillespie PJ, Khoudoli GA, Stewart G, Swedlow JR and Blow JJ 2007 Curr Biol 17(19): 1 657-62 122. Galy V, Askjaer P, Franz C, Lopez-Iglesias C and Mattaj [W 2006 Curr Biol 16(17): 1 748-56. 123. Olsen JV, Blagoev B, Gnad F, Macek B, Kumar C and Mortensen P 2006 Cell 127: 635-648 124. Matsuoka S, Ballif BA, Smogorzewska A, McDonald ER 111, Hurov KB and Luo J 2007 Science 316:1160-1 166 123 125. Kastan MB, Onyekwere O, Sidransky D, Vogelstein B and Craig RW 1991 Cancer Res 51:6304-631 1 126. Cortez D, Wang Y, Qin J, Elledge SJ 1999 Science 286: 1162-6. 127. Nakayama M, Nakaj ima D, Nagase T, Nomura N, Seki N and Ohara O 1998 Genomics 51: 27-34 128. Ruault M, Brun M-E, Ventura M, Roizes G and De Sario A 2002 Gene 284: 73-81 129. Nagase T, Ishikawa K, Miyajima N, Tanaka A, Kotani H and Nomura N 1998 DNA Res 5:31-39 130. Nakajima-Iijima S, Hamada H, Reddy P and Kakunaga T 1985 Proc Natl Acad Sci USA 82:6133-6137 131. Chen H, Wu Y, Zhang Y, Jin L, Luo L, Xue B, Lu C, Zhang X and Yin Z 2006 FEBS Lett 580(13):3145-52 132. Nagase T, Seki N, Tanaka A, Ishikawa K and Nomura N 1995 DNA Res 2: 167-174 133. Wang ZG, White PS and Ackerman SH 2001 J Biol Chem 276:30773-30778 134. Fodera D, D'Alessandro N, Cusimano A, Poma P, Notarbartolo M, Lampiasi N, Montalto G and Cervello M 2004 Ann N Y Acad Sci 1028:440-9 135. Cheong E, Ivory K, Doleman J, Parker ML, Rhodes M and Johnson IT 2004 Carcinogenesis 25(10):1945-52 136. DeWitt DL and Meade EA 1993 Arch Biochem Biophys 306(1):94-102 124 137. Kamei D, Murakami M, Nakatani Y, Ishikawa Y, Ishii T and Kudo I. 2003 J Biol Chem 278(21): 19396-405 138. Morita I, Schindler M, Regier MK, Otto JC, Hori T, DeWitt DL and Smith WL 1995 J Biol Chem 270(18): 10902-8 139. Futaki N, Yoshikawa K and Hamasaka Y 1993 Gen Pharmacol 24:105—1 10 140. Sheng H, Shao J, Kirkland SC, Isakson P, Coffey RJ, Morrow J, Beauchamp RD and DuBois RN 1997 J Clin Invest 99(9): 2254—2259 141. Subbaramaiah K, Telang N, Ramonetti JT, Araki R, DeVito B, Weksler, BB and Dannenberg AJ 1996 Cancer Res. 56:4424—4429 142. Kargman SL, O'Neill GP, Vickers PJ, Evans JF, Mancini JA and Jothy S 1995 Cancer Res 55:2556—2559 125 5 2221222122222222