Dehyrdogenation of indole and indoline compounds by cytochrome P450 enzymes.

3-Substituted indoles like 3-methylindole, zafirlukast and MK-0524 are dehydrogenated by cytochrome P450s to form reactive 3-methyleneindolenine electrophiles, which covalently conjugate with protein and/or DNA nucleophilic residues to cause toxicities. Likewise, the dehydrogenation of indolines to form indoles is a potentially important transformation. However, the mechanisms that govern the dehydrogenation reaction by selected P450s have not been fully established. The goal of this dissertation was to evaluate the ability of P450s to catalyze the dehydrogenation of indole and indoline compounds. In the present study, the in vitro metabolism of another 3-substituted indole, SPD-304, a newly discovered small-molecule TNF antagonist, together with indoline and indoline-containing drugs including indapamide, DW2282 (anti-cancer), SB-206553 (5-HT 2C/2B antagonist), SB-224289 (5-HT 1B inverse agonist) and pyroquilon, and several synthetic indoline derivatives, were investigated. We found SPD-304 was bioactivated through a dehydrogenation mechanism identical to 3-methylindole, to produce an electrophilic 3-methyleneindolenine, which was trapped by the nucleophile glutathione. This potentially important new drug caused the selective mechanism-based inactivation of CYP3A4. In addition, we also found that all the indolines were aromatized to indoles through a novel dehydrogenation pathway, which appeared to be initiated by C-H bond breakage instead of nitrogen oxidation. The metabolic profiles of the substrates were also assessed by in silico molecular docking. Dehydrogenation of SPD-304 appeared to be a major metabolic pathway, and the arginine 212 residue in the active site of CYP3A4 played an important role in hydrogen bonding with all indoline nitrogens by positioning the C-2 and C-3 carbons of indolines close to the heme iron for dehydrogenation. Using this in silico docking strategy, in combination with site-directed mutagenesis and homology modeling techniques, pulmonary cytochrome P450 2F3 and mutants were constructed and docked with substrate 3-methylindole. Substrate orientation favoring dehydrogenation was identified and verified by enzyme kinetic studies, with additional critical residues in the active site of CYP2F3 identified that could direct the selective dehydrogenation reaction. This research has significantly broadened our current understanding about the mechanisms of cytochrome P450-mediated dehydrogenation of xenobiotic indole and indoline compounds.

DEHYDROGENATION OF INDOLE AND INDOLINE COMPOUNDS BY CYTOCHROME P450 ENZYMES by Hao Sun A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy In Pharmacology Department of Pharmacology and Toxicology The University of Utah December 2007 Copyright Hao Sun 2007 All Rights Reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITTEE APPROVAL of a dissertation submitted by Hao Sun This dissertation has been read by each member of the following supervisory committee and by majority vote has been found to be satisfactory. J (t 7: Loo1- I I THE UNIVERSITY OF UTAH GRADUATE SCHOOL FINAL READING APPROVAL To the Graduate Council of the University of Utah: I have read the dissertation of Hao Sun in its final fonn and have found that (l) its fonnat, citations, and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the supervisory committee and is ready for submission to The Graduate School. Approved for the Major Department Approved for the Graduate Council ...:D~ (C Q~ __ . David S. Cha an Dean of The Graduate School ABSTRACT 3-Substituted indoles like 3-methylindole, zafirlukast and MK-0524 are dehydrogenated by cytochrome P450s to form reactive 3-methyleneindolenine electrophiles, which covalently conjugate with protein andlor DNA nucleophilic residues to cause toxicities. Likewise, the dehydrogenation of indo lines to form indoles is a potentially important transformation. However, the mechanisms that govern the dehydrogenation reaction by selected P450s have not been fully established. The goal of this dissertation was to evaluate the ability of P450s to catalyze the dehydrogenation of . indole and indo line compounds. In the present study, the in vitro metabolism of another 3-substituted indole, SPD-304, a newly discovered small-molecule 1NF antagonist, . together with indo line and indo line-containing drugs including indapamide, DW2282 (anti-cancer), SB-206553 (5-HT2C/2B antagonist), SB-224289 (5-HTlB inverse agonist) and pyroquilon, and several synthetic indoline derivatives, were investigated. We found SPD-304 was bioactivated through a dehydrogenation mechanism identical to 3-methylindole, to produce an electrophilic 3-methyleneindolenine, which was trapped by the nucleophile glutathione. This potentially important new drug caused the selective mechanism-based inactivation of CYP3A4. In addition, we also found that all the indolines were aromatized to indoles through a novel dehydrogenation pathway, which appeared to be initiated by C-H bond breakage instead of nitrogen oxidation. The metabolic profiles of the substrates were also assessed by in silico molecular docking. Dehydrogenation of SPD-304 appeared to be a major metabolic pathway, and the arginine 212 residue in the active site of CYP3A4 played an important role in hydrogen bonding with all indoline nitrogens by positioning the C-2 and C-3 carbons of indolines close to the heme iron for dehydrogenation. Using this in silico docking strategy, in combination with site-directed mutagenesis and homology modeling techniques, pulmonary cytochrome P450 2F3 and mutants were constructed and docked with substrate 3-methylindole. Substrate orientation favoring dehydrogenation was identified a~d verified by enzyme kinetic studies, with additional critical residues in the active site of CYP2F3 identified that could direct the selective dehydrogenation reaction. This research has significantly broadened our current understanding about the mechanisms of cytochrome P450-mediated dehydrogenation of xenobiotic indole and indoline compounds. v TABLE OF CONTENTS ABSTRACT ....................................................................................................................... iv ACKNOWLEDGMENTS ............................................................................................... viii CHAPTER 1. BACKGROUND AND SIGNIFICANCE .......................................................... 1 P450 Heme and Catalytic Cycle ................................................................. 2 Cytochrome P450 Catalyzed Reactions ...................................................... 5 P450-Catalyzed Dehydrogenation ............................................................ 12 Indole and Indoline Compounds with Pharmacological Activities .......... 20 In Silico Modeling in Drug Discovery ...................................................... 29 Thesis Summary ........................................................................................ 36 References ................................................................................................. 37 2. HOMOLOGY MODELING OF CYP2F3 ........................................................ 55 Homology Modeling ................................................................................. 56 Small Molecule Docking .......................................................................... 60 CYP2F3 Site-Directed Mutagenesis ......................................................... 65 Materials and Methods ...................... ; ....................................................... 66 Results and Discussion ............................................................................. 68 References .................... ".. .......... 0 ................. 0 ............................. 0 ................ 81 3. METABOLIC ACTIVATION OF A NOVEL 3-SUBSTITUTED INDOLE-CONTAINING TNF-a INHIBITOR: DEHYDROGENATION ANDINACTIV ATION OF CYP3A4 .................................................................. 87 Materials and Methods .............................................................................. 90 Results ....................................................................................................... 96 Discussion ............................................................................................... 114 References ............................................................................................... 119 4. DEHYDROGENATION OF INDOLINE BY CYTOCHROME P450 ENZYMES: A NOVEL "AROMATASE" PROCESS ..................................... 123 Abstract ................................................................................................... 124 Materials and Methods ............................................................................ 125 Results ..................................................................................................... 127 Discussion ............................................................................................... 130 References ............................................................................................... 131 5. MECHANISTIC STUDIES ON THE CYP3A4-CATALYZED DEHYDROGENATION OF INDOLINES ........................................................ 133 Materials and Methods..... ....................... ......................................... ....... 137 Results ..................................................................................................... 145 Discussion ............................................................................................... 166 References ............................................................................................... 1 72 6. CONCLUSIONS AND FUTURE DIRECTIONS .......................................... 176 References ............................................................................................... 1 79 vii ACKNOWLEDGMENTS I wish to express my sincere gratitude to my mentor Professor Garold S. Yost for his confidence in me and for giving me the opportunity to work in his lab during the last five years. My graduate career at the University of Utah has been a wonderful and rewarding experience. This would have not been possible without his constant encouragement, support, and guidance. His mentorship was paramount in providing a well rounded experience consistent my long-term career goals. He encouraged me to not only grow as an experimentalist and a chemist but also as an independent thinker. He has been incredibly patient and has given me the time to grow professionally and mature personally. I could not have had a better mentor and I will never be able to thank him enough for all that he has given me. I would like to thank all the members of the Yost lab, who provided supportive and inspiring environment for working and learning. I especially want to thank Diane Lanza for keeping the lab running smoothly and helping out ,wherever needed in all my projects. She has always been extremely generous with her expertise and talents. I thank Dr. Christopher Reilly for serving on my dissertation committee, for supplying advice and perspective on mass spectrometry throughout the years. I thank Drs. William Nichols and Shane Cutler for their time and energy spent as members of my candidacy committee, and for being so helpful and for many pleasant conversations. I also would like to thank Dr. Konstantine Skordos for his advice in chromatography and Dr. LaHoma Easterwood for her effort in synthesizing chemicals for my project. I consider all of these people to be great friends and terrific scientists, and I look forward to future interactions with everyone. Finally, Brian Carr, Cassandra Deering, Frederick Henion, Jaya Kartha, Chad Moore, Aaron Rowland, Ashwini.Sabnis, Mohammad Shadid, Kiumars Shahrokh, Karen Thomas and Jie Wan have been wonderful labmates and friends, who have provided a lighthearted atmosphere that has made working in the lab pleasure. I would like to acknowledge all members of the Yost lab for their help, advice, generosity and friendship. I would like to thank my dissertation committee members Drs. Douglas E. Rollins, C. Dale Poulter and Gregory A. Voth. Their advice, support, suggestions, and critical review of this project were most appreciated. I thank Dr. Eric F. Johnson at the Scripps Research Institute, for his teaching and advice in homology modeling and small molecule docking, and for his invaluable contribution to the in silico projects of my dissertation. He is an excellent mentor who taught me most of what I know about cytochrome P450 modeling and substrate docking. Once again, I must thank Dr. Yost for giving me this wonderful opportunity and supporting me to visit Dr. Johnson's lab in La Jolla. I would like to thank Drs. William J. Ehlhardt and Palaniappan KuJanthaivel at Eli Lilly, for their exceptional technical support on the NMR determination of indo line metabolites. I also want to acknowledge Drs. James R. Halpert and Santosh Kumar at the University of Texas, for providing CYP3A4 mutants for my study. There are no words to express my gratitude to my parents. I owe a huge debt of gratitude to them for their constant love and encouragement throughout not only the course of my graduate studies, but throughout my entire life. It was under their watchful ix eye that I gained so much drive and an ability to tackle challenges head on. They brought science into my life, encouraged me to pursue my dreams, and have been with me every step of the way. Thanks, Mom and Dad. Finally, and most importantly, I would like to thank my wife Shunyu Fan. Her support, encouragement, quiet patience and unwavering love were undeniably the bedrock upon which the past ten years of my life have been built. Her love and support is unfailing and her patience and understanding during many long nights working was unselfish and appreciated more than she knows. I'll never thank her enough for helping me realize this part of my dream, completing studies with a Ph.D. degree in the United States. This work, and my life, is dedicated to her. Chapter 4 of this dissertation is reprinted with permission from the American Society for Pharmacology and Experimental Therapeutics, which could be found in the Journal of Pharmacology and Experimental Therapeutics, 322: 843-851,2007. x CHAPTER 1 BACKGROUND AND SIGNIFICANCE Cytochrome P450 monooxygenases are a diverse superfamily of heme-thiolate proteins, which could originate from an ancestral gene that existed over 3 billion years ago (Danielson, 2002). P450 enzymes are the most versatile biological catalysts involved in numerous cellular functions, including xenobiotic oxygenation as well as the synthesis of steroid hormones, bile acids, vitamins and cholesterol (pikuleva and Waterman, 1999; Coon, 2005; Guengerich, 2006). They are found in the bacteria, euglenozoa, mycetozoa, alveolates, stramenopiles, fungi, plants and animals with a total of over 700 families and about 8000 species. The name P450 derives from the characteristic exhibition of a UV maximum absorption at 450 nm upon binding of carbon monoxide by the reduced enzyme (Danielson, 2002). Human P450 enzymes are located in the membrane of the endoplasmic reticulum of liver and many other tissues such as lung, kidney, brain, small intestine, nasal tissue, peripheral blood cells, platele~s, neutrophils, seminal vesicles, aorta, adrenals and other steroidogenic tissues (Guengerich, 2005). Fifty-seven human P450s exist and are divided into 18 families and 43 subfamilies. The catalytic activities of P450s include steroidogenesis (CYP1B1, 7A11B1, 8B1, 11AlIBI1B2, 17, 19, 21A2, 27A1, 39,46 and 51), metabolism of vitamins (CYP 24, 26A11B1 and 27B1), biotransformation of fatty 2 acids (CYP 2J2, 4All, 4Bl and 4F12), eicosanoid metabolism (C~ 4F2/3/8, 5Al and 8AI), as well as metabolism of xenobiotics/drugs (CYPIA1I2, 2A6/13, 2B6, 2C8/9/18/19, 2D6, 2El, 2Fl, 3A4/5/7) (Guengerich, 2005). Only drug metabolism cytochrome P450s are discussed in this dissertation. P450 Heme and Catalytic Cycle As shown in Figure 1.1, a tetrapyrrole ring system linked by methene bridges and complexedwith iron forms the P450 heme, a highly conjugated and planar porphyrin. The heme iron is octahedrally coordinated by six heteroatoms. The four porphyrin nitrogens are equatorially orientated, and the remaining two axial ligands lie above and below the plane of the heme. The proximal ligand (below the heme) is the thiolate sulfur atom of cysteine for P450s and chloroperoxidase, but is the nitrogen atom of histidine for all the other peroxidases and the oxygen atom of tyrosine for catalase. The distal ligand (sixth coordination site, above the heme) is occupied by an oxygen atom from either O2 eOOH eOOH Figure 1.1. Structure ofP450 Heme: Protoporphyrin IX Complexed with Iron (III) 3 or H20 for P450s (Shaik and De Visser, 2005). The activity of P450s requires a cofactor reducing agent NADPH, nicotinamide adenine dinucleotide phosphate, as well as molecular oxygen (paine et aI., 2005). NADPH provides the coupled and step-wise supply of electrons as a protein electron transport system or redox partner. Microsomal P450s receive electrons from a \ membrane-bound enzyme,NADPH cytochrome P450 oxidoreductase, which contains FAD (flavin adenine dinucleotide) and FMN (flavin mononucleotide) cofactors. Reductase catalyzes the electron transfer from NADPH to FAD to FMN and then to P450s, in which reductase changes its conformation to control the transfer process in the chain, for example, using the "open" conformation state for the efficient electron transfer to P450s, but "closed" conformation state for inter-flavin electron transfer. In addition, several novel types of P450 redox partner fusion enzymes have been identified (Munro et ai., 2007), such as bypassing the requirement for a protein redox partner or using hydrogen peroxide to provide electrons, protons and oxygen. Cytochrome bs, located on the cytosolic side of endoplasmic reticulum, frequently enhances the rates of P450 catalysis, especially play an important role in CYP3A4-catalyzed reactions (paine et aI., 2005). In the P450 catalytic cycle as shown in Figure 1.2, one or several short-lived, highly oxidizing intermediates were formed at the heme iron and near the bound substrate to execute the catalysis activity (Makris et aI., 2005),. In the resting-state P450s without any substrate binding, ferric iron is in the low-spin six-coordinate form with water conjugated as the sixth ligand. The catalytic cycle starts with the substrate binding that results in the removal of water from the active site so the heme ion becomes five- 4 R H ROH H H RH '0/ -1 '0/ RH flt~7 flt~V .. ~r~V SCys SCys SCys r Resting State ! e- RH RH 0 ~_I_Il e(~ Z.+ P450CYCLE ~F~ SCys SCys B++lIzO ! Oz OHRH 2- 0- If RH d IRH 0 0 flt? .. H+ flt~7 .. e flt~7 SCys SCys SCys Figure 1.2. Catalytic Cycle of Cytochrome P450s coordinate, and hence the heme iron changes to predominantly high-spin state to increase its reduction potential and improve the coupling efficiency of electron transfer. At the same time, the first electron transfer from NADPH reduces the iron to ferrous state. Oxygen binding to ferrous P450 results in an unstable oxy-P450 complex that then accepts the second electron from NADPH. These sequential electron transfers are believed to be the rate-limiting steps. 5 Following the formation of the peroxo-ferric intermediate, the oxygen could be protonated to yield a hydroperoxo-ferric intermediate, from which a second protonation at the distal oxygen atom followed by a heterolysis or cleavage of the 0-0 bond to form the ion-oxo "Compound I" and water. However, characterization of "Compound I" state has not yet been achieved. The iron-oxo species is the most widely accepted active intermediate in P450-catalyzed reactions but recent studies also indicated that multiple intermediates affect substrate turnover. It has been proposed that multiple intermediates are all active oxygenating species, with varying electrophilic or nucleophilic properties that could contribute to the versatility of P450 enzymes. An example would be an iron­oxo for hydroxylation but hydroperoxo-iron for epoxidation reactions. Finally the substrate is oxygenated to a product complex and then released from the active site of P450s (Makris et aI., 2005). Cytochrome P450 Catalyzed Reactions The biotransformation of xenobiotics is catalyzed by both phase I and phase II enzymes. Cytochrome P450 catalyzed oxidations belongs to phase I reactions, which also include hydrolysis catalyzed by carboxylesterases, peptidase and epoxide hydrolase, as well as reduction reactions. Phase I metabolism often introduces polar functional groups to xenobiotics such as hydroxyl, carboxyl and amino, which thereby could be conjugated by phase II enzymes such as UDP-glucUronosyltransferases and sulfotransferases. P450s can catalyze common oxidation reactions such as hydroxylation (both aliphatic and aromatic), epoxidation, N- or O-dealkylation, N-oxidation and 6 dehalogenation~ and also "uncommon" oxidative biotransformation such as dehydrogenation. Hydroxylation P450-catalyzed carbon hydroxylation generally forms an alcohol, or further oxidation of the alcohol to a carbonyl or an aldehyde that could subsequently form. a carboxylic acid. The mechanism of hydroxylation is generally considered to be a classic hydrogen abstraction by the ferryl oxygen~ followed by oxygen rebound re-combination of the resulting carbon radicals with iron-bound hydroxyl radical (Guengerich, 2001a; Ortiz de Montellano and De Voss, 2005). .These mechanisms have been studied extensively using kinetic deuterium isotope effect experiments~ but they have been challenged by studies with radical clock probes that could estimate the rate of the oxygen rebound step and the lifetime of the radical intermediates formed during hydroxylation (Ortiz de Montellano and De Voss, 2005). These results from radical clock probe studies indicated the radical intermediate formation might not be a mandatory step in hydrocarbon hydroxylation, or the oxygen can be inserted into the C-H bond through a concerted and nonradical mechanism. Almost all the major drug metabolism P450 enzymes can catalyze the hydroxy lation of xenobiotics. Some selected e.xamples that have been studied in recent years are listed as follows. CYPIAI can catalyze the 2-, 4-, and 16a.-hydroxylation of 17p-estradiol (E2), an estrogen (Spink et al.~ 1992). CYPIA2 is an important enzyme involved in the metabolism of 2-amino-3,5-dimethylimidazo[4,5 .. f]quinoline (MeIQ) and other heterocyclic arylamines that could be bioactivated to corresponding mutagens such 7 as hydroxylamines and nitroso compounds (Kim and Guengerich, 2004). CYF2A6 is one of the major enzymes involved in the 7-hydroxylation of coumarin and 2'-hydroxylation of nicotine as well as metabolism of other carcinogenic nitrosamines (Hecht et aI., 2000; Kim et aI., 2005a). B uprop ion, an atypical antidepressant that acts as a norepinephrine reuptake. inhibitor and dopamine reuptake inhibitor, was hydroxylated by CYP2B6 (Turpeinen et aI., 2004). In addition, an endothelin ETA receptor antagonist was found to be hydroxylated by CYP2C8 (Ma et aI., 2004). CYP2C9 can catalyze the 5'-hydroxylation of lornoxicam (!ida et aI., 2004), a nonsteroidal anti-inflammatory agent that decreases prostaglandin synthesis by directly inhibiting cyclooxygenase. Anticonvulsant mephobarbital was hydroxylated by CYF2C19 stereoselectively. A major metabolite of several psychoactive drugs, 1-(2- pyrimidinyl)-piperazine, was found to be catalyzed by CYF2D6 through 5-hydroxylation (Raghavan et aI., 2005). CYP2E1 is the enzyme for the hydroxylation of salicylate into 2,3 and 2,5-dihydroxybenzoic acids, 00-1 hydroxylation of oleic acid as well as hydroxylation of skeletal muscle relaxant chlorzoxazone (Jayyosi et aI., 1995; Adas et aI., 1998; Dupont et aI., 1999). Moreover, CYP3A4 catalyzes .the hydroxylation of many different xenobiotics with various pharmacological activities such as the hydroxylation of bupropion, a second-generation antidepressant agent used in the management of smoking cessation (Faucette et aI., 2001). Epoxidation Chemicals such as olefins, heterocyclics, vinyl halides, aromatic hydrocarbons, vinyl nitrosamines and ethyl carbamate can be biotransformed by P450s through 8 epoxidation reactions to form corresponding epoxides (or oxiranes) (Guengerich, 2003). One of the 1t-bonds in the aromatic ring such as benzene could also form benzene oxide through epoxidation, but it is unstable and one of the epoxide C-O bond undergoes heterolytic cleavage to give a ketone intermediate followed by the tautomerization to form a phenol product (Ortiz. de Montellano and De Voss, 2005). In addition, epoxidation of olefin, acetylene and other aliphatic unsaturated carbons was also through the oxidation of their 1t-bonds. As shown in Figure 1.3, acrylanlide was found to be biotransformed through the epoxidation by CYP2E1 to glycidamide that could conjugate glutathione, DNA and hemoglobin (Ghanayem et aI., 2005). Other examples include the epoxidation of thiophene derivatives (Dansette et al., 2005), olefins (Jin et al., 2003) and styrenes (Estavillo et aI., 2003). The epoxidation pathway often causes toxicities through the formation of the electrophilic intermediate epoxides (Lau and Zannoni, 1979; Guengerich, 2003). Our previous studies of 3-methylindole oxidation by P450s demonstrated that several (A) (B) cd ,,/ NH 0 0 0 CYP2El H2N~O oj H2NJV -----+- ~ I NH ~ roo ~ I NH Figure 1.3. P450-Catalyzed Epoxidation Reaction 9 epoxidation products were formed including 2,3-epoxy-3-methylindole (Skordos et aI., 1998) and two epoxides on the benzene ring of 3-methylindole (Yan et aI., 2007). The ring opening of 2,3-epoxide could form 3-hydroxy-3-methylindoline reactive intermediate that was found to be further oxidized to 3-hydroxy-3-methyloxindole. Epoxides can be stable from less than one second to several hours and these epoxide reactive intermediates can ultimately cause toxicities by covalently binding to tissues, alkylating proteins and nucleic acids, as well as the overall damage to biological systems. N-dealkylation and O-dealkylation Both reactions belong to cytochrome P450-catalyzed heteroatom oxidation that . also include S-dealkylation and oxidative dehalogenation. If a hydroxyl group is introduced onto the carbon atom attached to the heteroatom such as nitrogen or oxygen, the compound becomes unstable and an elimination step will be followed to form a carbonyl and an amine/alcohol (Ortiz de Montellano and De Voss, 2005). P450-mediated O-dealkylation reactions were found with various substrates such as 6-methoxyquinoline, 7-alkoxycoumarin (Desta et aI., 2002), fluoxetine, pentoxyresorufm and 7-benzyloxy-4-trifluoromethylcoumarin, as shown in Figure 1.4. The mechanisms of O-dealkylation of several of these compounds were also investigated with deuterium isotope studies and found that the C-H bond breakage was the rate­limiting step for the oxidation of these substrates. Domperidone is a dopamine D2 receptor antagonist indicated as an antiemetic drug for the prevention of nausea and vomiting caused by chemotherapy and antiparkinsonian drug therapy (Ward et aI., 2004). An N-dealkylation of domperidone >=Aoz ~~~~N )0l NIl '---i-A < ~ CYP!A4 o HNAN ............... COOH o + ~~ J ~tr Cl Domperidone Cl JJ:l ~ , I CBaCBaO ~ 0 0 CYP1A2 • ~ m~oAo + CH]CHO 7 -Ethoxycoumarin Figure 1.4. P450-Catalyzed N-Dealkylation and O-Dealkylation Reactions ...... o 11 catalyzed by CYP3A4 will break the whole molecule into two separate benzimidazole moieties: one is an amine and the other a carboxylic acid through the further oxidation of an intermediate aldehyde, as shown in Figure 1.4. Other cytochrome P450-catalyzed N­de alkylation studies recently include the formation of amphetamine by N-dealkylation of fenproporex (Kraemer et aI., 2004), N-dealkylation of cyclopropylamine to cyclopropanone (Shaffer et aI., 2002), as well as inhibition studies on the pimozide N­dealkylation (Desta et aI., 2002). Since nitrogen has a low electronegativity, a direct nitrogen electron abstraction (to produce nitrogen radical cation) could be the rate­limiting step of N-dealkylation (Ortiz de Montellano and De Voss, 2005). Other oxidation reactions Other "uncommon" reactions such as reductions, oxidative ester cleavage, ring expansions or formation, aldehyde scissions, dehydration, coupling reactions, isomerization as well as rearrangement of fatty acid and prostaglandin hydroperoxides were also observed in P450-catalyzed oxidation (Guengerich, 2001b; Guengerich, 2001a; Ortiz de Montellano and De Voss, 2005; Isin and Guengerich, 2007). Dehydrogenation reactions are discussed below. Some examples of these "uncommon" reactions include the oxidation of dihydrobenzoxathiin by CYP3A4 producing a new cyclic ring (Zhang et aI., 2005), cleavage of a 1,2,4-oxadiazole ring to generate amidine, amine, and carboxylic acid ring­opened products (Dalvie et aI., 2002), cleavage of amino oxazoles such as ditazole oxidative deamination reactions (Maurer and Kleff, 1988), ring contraction through the conversion of 2,2,6,6-tetramethylpiperidines to 2,2-dimethylpyrrolidines by CYP3A4 12 (Yin et al., 2004), as well as a novel ring formation product of the anti-cancer drug YH3945 through the conjugation of a diene with an iminium (Lee et aI., 2004). P450-Catalyzed Dehydrogenation P450-mediated dehydrogenation of xenobiotics initiates from the hydrogen atom abstraction or other one electron oxygenation, to from a radical intermediate, the same process as the hydroxylation reactions as mentioned above. However, instead of the following oxygen rebound in hydroxylation, the second step of dehydrogenation is another one-electron oxidation (Yost, 2001) to form a dehydrogenated product. In the last twenty years, many studies have demonstrated that dehydrogenation is not an "uncommon" pathway any more (Ortiz de Montellano and De Voss, 2005). Some selected studies on the P450-catalyzed dehydrogenation are listed below: a cholesterol-lowering drug lovastatin was found to be dehydrogenated to 6'-exomethylene (Vyas et aI., 1990); quinone methides were found through the dehydrogenation of butylated hydroxytoluene (Bolton et aI., 1990), eugenol (Bolton et aI., 1995) and 4- hydroxytamoxifen (Fan et aI., 2000). by various P450s; valproic acid was catalyzed by CYP4B 1 to form ~4,5-valproic acid through dehydrogenation (Rettie et aI., 1995); a terminal olefin, 2-ethyl-5-hexenoic acid was formed during the dehydrogenation of 2- ethylhexanoic acid (pennanen et aI., 1996); N-dehydrogenation and aliphatic dehydrogenation of capsaicin by CYP2E1 and CYP3A4 formed imines and alkenes, respectively (Reilly et aI., 2003); metabolic activation of zafirlukast (Kassahun et aI., 2005) and MK-0524 (Levesque et aI., 2007) through dehydrogenation formed o:,~­unsaturated iminiums; dehydrogenation of testosterone at the 6 and 7 positions (Nagata 13 et aI., 1986}; dehydrogenation of lauric acid formed 11-dodecenoic acid by CYP4B 1 (Guan et aI., 1998); an imine methide, 3-methylenindolenine, was formed by dehydrogenation of the pneumotoxicant 3-methylindole by CYP2F enzymes (Skiles and Yost, 1996). In summary, P450-mediated dehydrogenation generally could transform saturated hydrocarbons to unsaturated ones, amines to imines, or alcohols to carbonyl compounds such as quinone and quinone methide. 3-Methylindole, a potent pneumotoxin found in animal and human digestive systems through tryptophan degradation and also in the cigarette smoke, can cause severe lung injury (Nocerini et aI., 1984; Yost, 1989). Our studies demonstrated that 3- methylindole was bioactivated through dehydrogenation by cytochrome P450s with two steps, formation of a 3-methyl radical, followed by a second one-electron oxidation, to produce a reactive intermediate, 3-methylenindolenine, an electrophilic imine methide species, which could covalently bind to protein and/or DNA nucleophilic residues to cause toxicities (Skiles and Yost, 1996; Thornton-Manning et aI., 1996; Regal et aI., 2001) (Figure 1.5). In another study on the dehydrogenation of indoline compounds, we found a novel indo line aromatization mechanism (Sun et aI., 2007). The dehydrogenation-mediated aromatization, reputedly occurring through two one-electron oxidation steps that are the same as 3-methylindole dehydrogenation, is different from the CYP 19 .. catalyzed aromatization, in which a sequence of two carbon oxidations and an oxidative carbon-carbon bond cleavage were found (Fishman, 1982). Some other P450- catalyzed dehydrogenation aromatization reactions were also investigated in recent years, which include 9,1 O-dihydrobenzo[ e ]pyrene, 1,2-dihydronaphthalene, 1,2- dihydroanthracene, 1,4-dihydro- pyridines, as well as N-alkyl-1 ,2,3,4-tetrahydroquinoline fudroxvlation 00" OR OH I "\ .oC-- ::::::..... NH Rebound ~. U-N~ Indole-3-c.arbinol Dehydrogenation /e U~-rJ 3-Metbyleneindolenine RSH ~JI ~I ::::::..... NH P450 ~ -H ~ U-N~ 3MI P~ 00 .. 0 ?"'I ::::::..... NH 2,3-Epoxide =1 ci2I = 0 :::::-.... NH P450 .. 3-Metbyloxindole P450. R ~ O-V-N~ ~OH ?'" I G)Y' ::::::..... NH Hydroxyindolenine .. o Qlygenation mH ?'" I 0 ::::::..... NH 3-Hydroxy-3-metbyloxindole Figure 1.5. Metabolic Pathways of 3-Methylindole by Cytochrome P450s ....... ~ 15 pyridines, as well as N-alkyl~1,2,3,4-tetrahydroquinoline (Wood et aI., 1979; Guengerich, 1988; Boyd et aI., 1993; Gu et aI., 2006). These biotransfonnation reactions were not believed to occur through hydroxylation followed by dehydration (Boyd et aI., 1993). Reactive Intennediates P450-catalyzed dehydrogenation often produces highly reactive electrophilic intermediates that could conjugate the nucleophiles such proteins and/or DNA, which could ultimately cause toxicities (Guengerich and Kim, 1991; Regal et aI., 2001; Evans et aI., 2004). As mentioned above, the dehydrogenation of 3-methylindole produced an electrophilic intennediate 3-methylenindolenine. Two other 3-methylindole-containing drugs, zafirlukast (Kassahun et aI., 2005) and . MK-0524 (Dean et aI., 2007; Levesque et aI., 2007) as shown in Figure 1.6 were also found to be dehydrogenated by P450s to fonn (A) (6) -0 NJ-D 8 - cO 0'0,)( 0 ~ N (C) (0) cr)N/ \N F 9' \ / OH I ~ 0=8=0 ~O-r: I ~ F F Figure 1.6. Structures of3-Methylindole, Zafirlukast, MK-0524 and SPD-304 16 similar electrophilic intermediates, a,~-unsaturated iminiums. Several other dehydrogenation-caused reactive intermediates have been mentioned above, which include a quinone imide formed by capsaicin dehydrogenation, imine or hydroxyl imine formed by eYP2B6-catalyzed dehydrogenation of phencyclidine, dehydrogenation of troglitazone by eYP3A4 eYP2e and eYP2D6 to form a quinone methide, L\4,5-valproic acid formed by eYP4B I-catalyzed dehydrogenation of valproiC acid, quinone methides formed by dehydrogenation of eugenol, isoeugenol and hydroxytamoxifen, as well as quinones formed by eYP3A4-catalyzed dehydrogenation of equine estrogens. Electrophilic reactive intermediates can also be formed through other P450 reactions. eYPlAl, lA2 and lBl are enzymes that metabolize polycyclic aromatic hydrocarbons such as benzo[a]pyrene to form phenols and oxides. The diol-epoxides produced are highly electrophilic intermediates that covalently bind to DNA and proteins. Other pathways, such as the conversion of radical cations through one-electron oxidation processes, or further oxidation of diols to reactive o-quinones, can also react with DNA to form stable depurinating adducts (Shimada, 2006). Quinone imines were found from the metabolism of a catechol-O-methyltransferase inhibitor, tolcapone(Smith et aI., 2003). It is a reactive intermediate that can ultimately covalently bind to hepatic proteins resulting in liver damage that could be associated with numerous cases of reported hepatotoxicity. In other similar studies, nontricyclic antidepressant trazodone and nefazodone were also found to produce reactive quinone-imine species by eVP3A4 (Kalgutkar et aI., 2005a; Kalgutkar et aI., 2005b) 17 The reactive intermediates could be detoxified by phase II enzymes and converted to more polar metabolites. The phase II enzymes that catalyze the detoxification reactions are glutathione transferases, UDP-glucuronosyltransferases, sulfotransferases, N-acetyltransferases, thioether methyltransferases and NADPH quinone oxidoreductase 1. Glucuronidation and sulfation are two major routes of detoxification elimination pathways that are catalyzed by UDP-glucuronosyltransferases and sulfotransferases to form polar and water soluble conjugates that are eliminated from the body in urine or bile (Wells et al., 2004). N-Acetylation is another major rote. of biotransformation for drugs containing aromatic amines that could be either activated (to form hydroxylamines) or deactivated (to form amide) by N-acetyltransferases. The most important conjugation reaction that have been used for trapping reactive intermediates in this dissertation is the glutathione conjugation reaction, which will form a thioether through glutathione nucleophilic attack of electrophilic carbons formed during phase I metabolism. Several minor phase II enzyme pathways are listed as follows: the methylation by thioether methyltransferases, as well as the a,p-hydrogenation of 2-alkenals to detoxify the lipid peroxide-derived reactive aldehydes by NADPH quinone oxidoreductase 1 (Mano et aI., 2002). Phase I enzyme epoxide hydrolase could also detoxify the epoxides to alcohols. Mechanism-Based Inactivation of Enzymes Reactive electrophilic intermediates formed by dehydrogenation in the active site of cytochrome P450s could be responsible for the irreversible inactivation (or the mechanism-based inhibition) of the enzyme directly. Other P450-catalyzed reactions 18 such as epoxidation could also foim reactive intermediates to cause the mechanism-based inactivation of enzymes. Xenobiotics inactivate P450s by either covalent binding to enzymes apoproteins such as sulfur and halogenated compounds, olefins and acetylenes, or by quasi-irreversible coordinating to the prosthetic heme such as methylenedioxy compounds, amines and 1, I-disubstituted and acyl hydrazines, or by covalently binding to the prosthetic heme such as terminal olefins, acetylenes, dihydropyridines, dihydroquinolines as well as arylhydrazines and hydrazones, or by other apoproteinlheme modification and degradation modes (Correia and Ortiz de Monte llano, 2005). Mechanism-based inhibitors have also been used as probes to characterize the active site of P450 enzymes and mechanism of catalysis such as identifying the critical amino· acid residues for binding and catalysis to design new drugs (Kent et aI., 2001). A variety of compounds including acetylenes, isothiocyanates, xanthates, aminobenzotriazoles, phencyclidine and furanocoumarins inactivate human cytochrome P450s and in many cases also cause toxicities (Kent et aI., 2001). Mechanism-based inactivators are often enzyme-specific. N-aralkyl-l-aminobenzotriazoles and rhapontigenin were found to be mechanism-based inhibitors of CYPIAI (Knickle et aI., 1993; Chun et aI., 2001). CYPIA2 is responsible for its own irreversible inactivation by furafylline through the formation of an imidazomethide intermediate and CYPIA2 is also inactivated by 2-ethynylnaphthalene, dihydralazine, isoniazid and zileuton, a 5-. lipoxygenase inhibitor (Yun et aI., 1992; Racha et aI., 1998; Masubuchi and Horie, 1999; Wen et aI., 2002; Lu et aI., 2003). Menthofuran, 8-methoxypsoralen, furanocoumarins and isoniazid are responsible for the mechanism-based inactivation of CYP2A6 (Koenigs et aI., 1997; Khojasteh-Bakht et aI., 1998; Koenigs and Trager, 1998; Wen et aI., 2002). 19 Several compounds such as phenyl diaziridines, phencyclidine, 2-phenyl-2-(l­piperidinyl) propane, a series of xanthates and n-propylxanthate could cause the irreversible inactivation of CYP2B6 (Kent et aI., 1999; Yanev et aI., 1999; Fan et aI., 2003; Jushchyshyn et aI., 2003; Sridar et aI., 2006). For CYP2C subfamily enzymes, the nonsteroidal anti-inflammatory drug suprofen, tienilic acid and thiophene derivatives are responsible for CYP2C9 inactivation (Lopez-Garcia et aI., 1994; Jean et aI., 1996; O'Donnell et aI., 2003), as well as isoniazid, tic10pidine and tienilic acid for CYP2C19 (Jean et aI., 1996; Ha-Duong et ai., 2001; Wen et aI., 2002). In addition, alkamides from piper nigrum and 1-[(2-ethyl-4-methyl-1H-imidazol- 5-yl)methyl]- 4-[4-(trifluoromethyl) .. 2-pyridinyl]piperazine are two categories of compounds for the mechanism-based inactivation of CYP2D6 (Hutzler et aI., 2004; Subehan et aI., 2006); and some tert-butyl acetylenes and 5-phenyl-l-pentyne for CYP2El. Many clinically important compounds were identified as mechanism-based CYP3A4 inhibitors for their irreversible inactivation of the enzyme (Chan et aI., 1998; He et aI., 1999; Masubuchi and Rorie, 1999; Lightning et aI., 2000; Lin et aI., 2002; Wen et aI., 2002; Alvarez-Diez and Zheng, 2004), such as selective peptide leukotriene receptor antagonist zafirlukast (Kassahun et aI., 2005) , selective estrogen receptor modulator tamoxifen (Zhao et aI., 2002), natural cytotoxin 4-ipomeanol (Alvarez-Diez and Zheng, 2004), a major constituent of oral contraceptives 17a.-ethynylestradiol (Lin et aI., 2002) as well as a potential HIV protease inhibitor L-754,394 (Lightning et aI., 2000). The total numbers of these human cytochrome P450 mechanism-based inactivators have dramatically increased in recent years. 20 Indole and Indoline Compounds with Pharmacological Activities Indole and 3-methylindole are two major products from the degradation of tryptophan that originates from the diet or materials derived from the upper part of the intestine. They are formed by gut bacteria and are absorbed into the body in substantial amounts (Jensen et a!., 1995; Gillam et aI., 2000). Indole dyes indigo and indirubin have been found in human urine, both of which could be formed through indole metabolism by cytochrome P450s (Gil~am et aI., 2000). Most of other indoles are natural alkaloid products as discussed below. Indoline, 2,3-dihydroindole, is a "saturated" structural analogue of indole. There are no endogenous indolines but indo line-containing drugs have been utilized as drug candidates in a variety of therapeutic fields, and are more frequently employed in new therapeutic entities in recent years. Indoles As mentioned above, P450-catalyzed dehydrogenation of 3-methylindole has been studied extensively in the past 30 years. In this dissertation study, both 3- substituted indoles and analogue indo lines were investigated. Zafirlukast (Accolate), a leukotriene D4 receptor antagonist used for the treatment of mild to moderate asthma, structurally is a 3-substituted indole. It was found that zafirlukast was dehydrogenated by CYP3A4 to form a similar electrophilic iminium intermediate as 3-methylindole, which could inactivate CYP3A4 in a mechanism-based process. Glutathione was added during the incubation· to trap this imine intermediate that was released by CYP3A4 and demonstrated that an adduct was formed on the 3-methylene group of zafirlukast (Kassahun et aI., 2005). MK-OS24, another 3-substituted indole, was also investigated 21 recently and the results showed the glutathione conjugation with the electrophilic iminium intermediate though the same dehydrogenation mechanism. The 3-substitued indole that was chosen for this dissertation is newly discovered and the first successful small-molecule TNF antagonist, SPD-304 (Figure 1.6). 1NF antagonists are very difficult to develop because the target is a protein-protein interaction between TNF and the TNF receptor that are relatively large and flat, and does not possess a substantial concavity to converge binding energy on a small-molecule antagonist (Whitty and Kumaravel, 2006). Structurally SPD-304 is 6,7-dimethyl-3-{[methyl-(2- {methyl-[i -(3-trifluoromethyl-phenyl)-lH-indol-3-ylmethylJ. .. amino }-ethyl)-amino]~ methyl}-chromen-4-one (He et aI., 2005). SPD-304 executes its anti ... TNF function by forming a "TNF dimer-antagonist complex", in which the trifluoromethylphenyl indole moiety binds to one monomer and the dimethyl chromone moiety binds to the other monomer of TNF through hydrophobic and shape-driven interactions between SPD-304 and mainly six tyrosine residues located at TNF monomers. Considering the structural similarities of SPD-304 with 3-methylindole, zafirlukast and MK-0524, we speculated that SPD-304 could also be dehydrogenated by P450s and thus is highly probable to cause toxicities. For. SPD-304, there is a structural difference from zafirlukast and MK-0524, instead of a carbon connected to the 3- methylene group, it has a nitrogen atom. Therefore, the electron donating effect of an amine group in SPD-304 will facilitate the hydrogen atom abstraction at 3-methylene carbon. Tryptophan is an essential amino acid from diet for protein biosynthesis, which is also the biochemical precursor for serotonin and melatonin. All three compounds are 3- 22 substituted indoles. Serotonin is an important neurotransmitter in the central nervous system and enterochromaffm cells in the gastrointestinal tract of humans that plays an important role in the regulation of mood, sleep, body temperature, sexuality and appetite. Many disorders are associated with serotonin depletion such as depression, migraine, irritable bowel syndrome, tinnitus, bipolar and anxiety. In recent years, it was found that serotonin agonist and/or antagonist also play significant roles in the pathogenesis of epilepsies (Bagdy et aI., 2007). Melatonin is a hormone produced through the activation of melatonin receptors with a particular role in the protection of nuclear and mitochondrial DNA. Melatonin was found to be metabolized by cytochrome P450s such as CYP2C19 to 6-hydroxymelatonin that could be further conjugated with sulfate and excreted in urine (Ma et aI., 2005; Huuhka et aI., 2006). Other tryptamines are also 3- substituted indoles such as N,N-dimethyltryptamine, a psychedelic tryptamine; mononlethyltryptamine, a tryptamine alkaloid found in the bark, shoots and'leaves of Virola with psychoactive effects; and psilocybin, a' psychedelic alkaloid found in psilocybin mushrooms, all of which could be catalyzed either by monoamine oxidase or cytochrome P450s through deamination reactions (Yu et aI., 2003). Other important 3-substituted indoles with pharmacological activities are listed below. Indomethacin is a nonsteroidal anti-inflammatory drug that could reduce the production of prostaglandins. Previous study showed that CYP2C9 catalyzed 0- demethylation of indomethacin (Nakajima et aI., 1998). In addition, ergolines are all 3- substitued indoles such as ergine and ergotamine. Ergine is responsible for the psychedelic activity and ergotamine is a vasoconstrictor used for the prevention of migraine. Reserpine is also an indole alkaloid indicated as an antipsychotic and 23 antihypertensive drug. For othe~ alkaloids, yohimbine is a selective adrenergic receptor antagonist for the treatment for erectile dysfunction. Yohimbine was found to be hydroxylated by CYP2D6 to 11-hydroxy-yohimbine (Le Corre et aI., 2004). In particular, vincristine is a very big molecule with molecule weight around 1000 as shown in Figure 1.7 but still was found to be metabolized to a secondary amine metabolite by CYP3A5 that could be the result of oxidative cleavage of the piperidine ring of dihydro­hydroxycatharanthine moiety of vincristine. Vincristine is a chemotherapy drug used for the treatment of leukemia, lymphoma, breast and lung cancers (Dennison et aI., 2006; Dennison et aI., 2007). HO Figure 1.7. Structure of Vincristine 24 As shown in Figure 1.8, rutaecarpine is a major active alkaloid component of the herbal medicine, Evodia rutaecarpa. Several hydroxylation metabolites at aromatic rings were found in vitro by human and rat microsomes and recombinant P450s. In vivo data showed that another hydroxylation product was formed at its aliphatic moiety which suggested the possibility of 3-methyleneindolenine formation (Lee et aI., 2005; Ueng et aI., 2006). In addition, rutaecarpine was also found to be a mechanism based inhibitor of CYP3A4 (Iwata et aI., 2005) and a potent inhibitor of CYPIA2 in both mouse and human liver microsomes (Ueng et aI., 2002). Another alkaloid, evodiamine, is an analogue of rutaecarpine that has antitumor activity through interleukin-1 mediated pathway to cause (A) {B} ~J':Q0~ VLN~N / - (C) V~Ltr:NQ °-~ Figure 1.8. Structures of Harm alan (A), Evodiamine (B) and Rutaecarpine (C) 25 apoptosis and necrosis (Pearce et aI., 2004; Zhang et aI., 2004; Liao et ai., 2005; Wang et aI., 2005). It is also an agonist for the vanilloid receptor TRPVl. The compound harmalan has a similar structure. Clonidine-displacing substances are considered as the endogenous ligands for imidazoline binding. A study on the purified active component of clonidine-displacing substances from bovine lung revealed the presence of L-tryptophan and l-carboxy-l-methyltetrahydrocarboline, harmane and harmalan, and the later two compounds exhibited a high affinity for both type 1 and type 2 imidazoline binding sites (Parker et aI., 2004). Several other derivatives of 3-methylindole, alkaloids or synthetic drugs, were also studied for their pharmacological activities with their structure shown on Figure 1.9. (A) /\-0- 05N N "--I '7 ~ ~ I NH (C) (B) (0) / N ~") d Figure 1.9. Structures of L-745,870 (A), Gramine (B), PD146176 (C) and Mebhydrolin (D) 26 For these indoles, instead of a carbon connected to the 3-methylene group, it is an electron donating nitrogen or sulfur atom. Gramine is an indole alkaloid found in the barley leaves and was indicated as a 5-hydroxytryptamine antagonist (5-HT2A) that could act as a vasorelaxing agent to cause concentration-dependent relaxation in pre contracted arterial rings (Froldi et aI., 2004; Larsson. et aI., 2006). Mebhydrolin is an HI antihistamine for the treatment of allergic rhinitis and allergic conjunctivitis (Kulthanan et aI., 2003). L-745,870 is a selective dopamine D4 receptor antagonist (Patel et aI., 1997; Gazi et aI., 1998). PD146176 is a 15-lipoxygenase inhibitor, which can limit the progression and monocyte-macrophage enrichment of hypercholesterolemia-induced atherosclerosis (Bocan et aI., 1998; Kelavkar et ai., 2001; Kim et aI., 2005b; Sordillo et ai., 2005). Indo lines Indolines have been widely investigated as serotonin receptor agonists or antagonists such as SB-242084, 6-chloro-5-methyl-l-[[2-[(2-methyl-3-pyridyl)oxy]-5- pyridyl]carbamoyl]- indo li ne, a selective 5-HT2c receptor antagonist with anxiolytic activity (Bromidge et aI., 1997); SB-224289, 1 '-methyl-5-[[2'-methyl-4'-(5-methyl-l ,2, 4- oxadiazol-3-yl)biphenyl-4-yl]carbonyl]-2,3,6,7-tetrahydro- spiro [furo[2,3-f]indole-3,4'­piperidine], a selective 5-HTlB receptor antagonist with negative intrinsic activity (Selkirk et aI., 1998); SB-206553, 5-methyl-1-(3-pyridy1carbamoyl)-1,2,3,5- tetrahydropyrrolo[2,3-1]indole, a selective 5-HT2C/2B receptor antagonist (Bromidge et aI., 1997) ; as well as 1-(l-indolinyl)-2-propylamine, a 5-HT2C receptor agonist for the treatment of obesity (Bentley et aI., 2004), as shown in Figure 1.10. W Indoline o SB224289 {~\ (~ tIN) D 0)..1 ~ SB206553 (Q--~ ~ fH o=c h NII2IIoKI Indapamide Figure 1.10. Structures 'of Indoline and Indoline Derivatives -. HN1'° Q-ZN,lJ oI ~I W ~ N DW2282 bo ~ cao Pyroquilon N .....J 28 A series of potent factor Xa inhibitors, the indo line derivatives of DX -9065a, (+)- 2S-2-[4-[[(3S)-I-acetimidoyl-3-pyrrolidinyl]oxy]phenyl]-3-[7-amidino-2-naphthyl]-pro­panoic acid, could be novel antithrombotics for the treatment and prevention of thromboembolic diseases (Noguchi et aI., 2006). In addition, an indoline derivative, 5- amino-l-(3,5-dimethylphenyl)-indoline was reported as a selective cyclooxygenase-l inhibitor for its anti-angiogenic property (Sano et aI., 2006). Some other indo line derivatives have been also investigated recently. Examples are: DW2282, (S)-(+)-4- phenyl-l-[I-( 4-aminobenzoyl)-indoline-5-sulfonyl]-4,5-dihydro-2-imidazolone hydro­chloride was found to be an anticancer agent (Hwang et aI., 1999); I-hexyl­indoline1actam- V, is a protein kinase C selective activator (Nakagawa et aI., 2006); as well as indo line methotrexate was investigated as an antirheumatic agent (Matsuoka et aI., 1996). A fascinating observation about indo line-containing therapeutic agents is that all of these drugs are under development, not commercially available yet. The exception are indapamide, 4-chloro-N-(2-methyl-2,3-dihydroindol-l-yl)- 3-sulfamoyl-benzamide, which is an indo line diuretic drug used to treat edema and hypertension (Robinson and Wellington, 2006) and pyroquil'on, 1,2,5,6- tetrahydro-4H-pyrrolo[3,2,I-ij]quinolin-4- one, which as used as a fungicide (Liao et aI., 2001), with structures shown in Figure 1.10. These indo line derivatives all share the same "saturated" dihydropyrrole structure at the C-2 and C-3 position of the indo line ring, but different electron donating or withdrawing groups (weak or strong) on the nitrogen of the indoline ring. A potential consequence of indo line aromatization is that the indole metabolites might have 29 significantly different pharmacological activities, when compared to their original indo line compounds (Bromidge et aI., 1997; Noguchi et aI., 2006). In Silica Modeling in Drug Discovery It requires 14 years on average for pharmaceutical companies to bring a successful drug into the market (Good, 2001; Lyne, 2002; Ghosh et aI., 2006), but the side effects of new drug entities are often exposed in a relative late stage, which imparts a big burden on research and development budgets of pharmaceutical companies (Figure 1.11). A recent termination of clinical testing of the cholesterol lowering drug, torcetrapib, cost Pfizer over $800 million (Cutler, 2007). The same situation was reported for the discontinuation of phase III trials of MK -7 6 7 (for the treatment of Lack of efficacy 30% Pharmacokinetics 39% Figure 1.11. Reasons for Attrition in Drug Development 30 diabetes) by Merck and GAL IDA (for the treatment of the glucose and lipid abnormalities associated with type 2 diabetes) by AstraZeneca (Calkin et aI., 2003). Recent technical improvements of drug screening include the automation of in vitro assays with robotics and miniaturization, the high-throughput screening of compounds with the development of new techniques in LCIMS and LCINMR, and utilization of combinatorial chemistry methods to synthesize huge libraries of related chemicals. However, high-throughput screening suffered from limitations with respect to available bioassays. Especially recent years, pharmaceutical companies are testing the drug metabolism, pharmacokinetics and toxicities much earlier than before. Therefore, in silico approaches could reduce the total hit-to-drug time line, accelerate drug discovery,· and also increase the quality of original hit-to-Iead process to select better candidates. In addition, in silico models have also been utilized to help choose appropriate assays for high-throughput screening. These predictive models could ultimately become sophisticated enough to replace in vitro assays and/or in vivo experiments (Langer and Hoffmann, 2001). ADMET and Modeling Tools Drugs enter into the human body either enterrally (oral, sublingual and rectal) or parenterally (subcutaneous, intramuscular and intravascular). Comparing to drugs given parenterally, oral drugs in tablets, capsules and soft gels are more convenient with lower costs and higher patient compliance rates (Iyer et aI., 2007). The ADMET models discussed here are only for oral drugs which are absorbed through g~strointestinal tract. 31 Modeling absorption. Three major factors to be considered for oral drug absorption, are drug dissolution, diffusion (crossing the intestinal membrane), and perfusion into the blood stream (lyer et aI., 2007). Various in silico models such as GastroPlus and Idea (Norris et aI., 2000; Agoram et aI., 2001) were developed for the prediction of oral absorption and lead optimization. For poorly soluble drugs, the rate­limiting step could be dissolution, but it is a perfusion process for soluble drugs. When the drug absorption is controlled by diffusion, some in vivo rat intestinal absorption models or in vitro Caco-2 or Madin-Darby canine kidney monolayer cell models are generally used to provide experimental data for in silico modeling (van de Waterbeemd and Gifford, 2003). Another important factor that affects drug absorption is the drug efflux pump p­glycoprotein, a member of the ATP-binding cassette transporter, which could transport various drugs in/out of cells for the uptake and elimination of drugs (Pleban and Ecker, 2005). Modeling of p-glycoprotein will help expose the target activity in early discovery stage when an efficient in vivo model is unavailable (Raub, 2006). MolSurf is a program to generate p-glycoprotein models, which was designed to predict A TPase activity according to the molecular surface, polarizability and hydrogen-bonding potentials (Osterberg and Norinder, 2000). The other program, Catalyst, allows the prediction of ICso values for p-glycoprotein inhibitors (Ekins et aI., 2002b; Ekins et aI., 2002a). Furthermore, the metabolizing enzymes in the gut wall also affect drug absorption. CYP3A4 expressed in the gut wall can metabolize some drugs before they reach the liver for major metabolism, but models that could combine all these factors would be very difficult to build. It is very complicated to construct a model to predict the bioavailability 32 of a drug, which includes the complicated sequence of two processes: absorption and liver metabolism (also including biliary excretion) (vande Waterbeemd and Gifford, 2003). Modeling distribution. The rate of distribution of drugs to target organs is determined primarily by blood flow, volume of distribution and the affinity of xenobiotics to target tissues, as predicted by several constructed models (Lombardo et al., 2002; Lobell and Sivarajah, 2003; Lombardo et aI., 2004). In addition, when drugs bind to plasma proteins such as red blood cells, leukocytes and platelets, as well as albumin, glycoproteins and lipoproteins, they cannot cross membranes to enter tissues. Sever~l techniques are then applied to model drug plasma binding, such as using a genetic function approximation algorithm, mUltiple computer-automated structure evaluation, pharmacophoric-similarity concept, or a 4D-fingerprint model (Beaudry et aI., 1999; Chung et aI., 2001; Colmenarejo et aI., 2001; Liu et al., 2006; Tang and Mayersohn, 2006). Moreover, several blood-brain barrier penetration models for CNS drugs were . also developed (Crivori et al., 2000; Iyer et at, 2002; Ooms et aL, 2002; Rose et aI., 2002; Rou and Xu, 2003; Winkler and Burden, 2004), which were used for the prediction of drug distribution into the brain. These models were based on multiple linear regression approaches including the octanol/water partition coefficient, hydrogen-bond acceptor number and the polar surface area calculation, physicochemical properties as well as ADME-tailored properties. Modeling metabolism. Basically, two different methods are currently used for the prediction of metabolism: data modeling and molecular modeling. Data modeling implements a quantitative structure-activity relationship (QSAR) method but molecular 33 modeling assesses the possibilities of the interactions between the substrates and proteins (cytochrome P450s). These models have played increasingly important roles as an alternative to in vitro metabolism studies (Waszkowycz, 2002; Schuster et aI., 2006; Jolivette and Ekins, 2007). In QSAR studies, several programs have been developed to predict metabolism using databases like Metabolite from MDL and Metabolism from Accelrys (Rodrigues et aI., 2001), in which several parameters such as the rate and extent of metabolism, and the enzymes and products formed are utilized. Additionally, drug­drug interactions could also be screened in silico by quantitative drug interaction prediction systems that integrate physiological, anatomical and genetic information derived from in vitro data (Bonnabry et aI., 1999; Bonnabry et aI., 2001). Interestingly, in silico studies of binding between substrates and the pregnane X receptor as well as the constitutive androstane receptor (CAR) can predict the induction of drug metabolism in an early phase of development (Liddle and Goodwin, 2002; Willson and Kliewer, 2002; Lamba et aI., 2005). In addition to QSAR-based metabolism prediction, pharmacophore-based, structure-based, reactivity-based and rule-based methods are also used to predict metabolites. The MetaSite program, a pharmacophore-based method, can perform semiempirical calculations, pharmacophoric recognition, descriptor handling and similarity computation (Cruciani et aI., 2005; Caron et aI., 2007). For structure-based molecular modeling studies (or docking), either the crystal structure of the drug metabolizing enzymes or a homology model based on the crystal structure template(s) were utilized. Docking approaches have been widely used in today's "hit to lead" drug discovery process (Kitchen et aI., 2004) and remain' a key step in virtual screening of 34 small-molecule drugs as a crucial component in contemporary drug discovery and development (Roberts, 2001). With new techniques developed in molecular biology, genomics and high-throughput X-ray crystallography, more and more three-dimensional protein structures have been elucidated including many of the major human drug metabolism enzymes, which provide us reliable in silico templates for modeling drug metabolism. Modeling excretion. The kidney is the most important organ for the excretion of xenobiotic drugs, but other excretion routes such as though feces, gases via the lungs, and biliary excretion are also very important. The in silico models of excretion are rather limited due to a lack of experimental data. Some programs have been designed to simulate the biliary excretion from human and animal hepatocyte data (van de Waterbeemd and Gifford, 2003). Modeling toxicity. The most direct and important parameter to evaluate a drug is its toxicity that is responsible for 20-40% drug failures. Computer-aided toxicity models are based on quantitative structure-toxicity relationships (DEREK and Multi Case ). There are two different approaches to predict toxicity: using models based on abstracting and codifying knowledge from the. scientific literature, using models based on descriptors of chemical structures arid statistical analysis of the relationships between toxicological end­points and descriptors (van de Waterbeemd and Gifford, 2003). Current in silico toxicity prediction programs are primarily focused on carcinogenicity and mutagenicity (Durham and Pearl, 2001; Cavalli et aI., 2002; Greene, 2002), with other toxicities such as teratogenicity, irritation, sensitization, immunotoxicology and neurotoxicity also considered but rather limited due to the lack of public-domain toxicology data. 35 Cytochrome P450 Models and Small Molecule Docking As mentioned above, cytochrome P450 homology modeling and small molecule docking belong to molecular modeling paradigms that evaluate the potential binding positions of small molecules within the active site of P450s, and predict substrate conformations and orientations. However, P450 molecular modeling and docking requires three-dimensional structures of proteins. It has been very exciting in the past three years that the structures of every major human P450 enzymes such as . CYP3A4 (Yano et aI., 2004), CYPIA2 (Sansen et aI., 2007), CYP2A6 (Yano et aI., 2005), CYP2D6 (Rowland et aI., 2006) and CYP2CS/9 (Schoch et aI., 2004; Wester et aI., 2004) have been determined by X-ray crystallography. For other human P450s that do not have three-dimensional structures, homology modeling of related P450 structures could be applied, which has provided stable and reliable prediction in the past (Kopp and Schwede, 2004). Since more and more three-dimensional protein structures have been known, in silica drug docking has become a crucial component for lead optimization in drug discovery. The docking and scoring techniques are often used in concert with other computational methods, such as data-based modeling of ADMET, to extend traditional approaches to structure-based design of the next generation drugs (Kitchen et aI., 2004). Overall, to predict ADMET parameters accurately in silico, both mathematical approaches and sufficient experimental data are required. Most of the data modeling models to date are rule-based instead of mechanism-based, and the later should be the next generation predictive models to more successfully predict and simulate ADMET' properties with an 'automated decision-making engine' for drug discovery. 36 Thesis Summary The research described in this dissertation has provided important, novel insight into cytochrome P450-catalyzed dehydrogenation of xenobiotic indole and indo line compounds. Chapter 1 provides an introduction to P450-catalyzed oxidation of xenobiotics, pharmacologically important indole and indo line compounds, as well as the in silico ADME modeling. Chapter 2 describes an in silico study in which a three­dimensional homology model of CYP2F3 and its mutants were built using other P450 structures determined by X-ray crystallography, and its prototypical dehydrogenation substrate 3-methylindole was docked into the active site. This study was both a means of testing the results from site-directed mutagenesis experiments and identifying important residues within the active site of CYP2F3 that could potentially control the dehydrogenation of 3-methylindole. The dehydrogenation and oxygenation of another 3- substituted indole SPD-304 was characterized using in vitro techniques as described in Chapter 3. In Chapter 4, a novel aromatization pathway through the dehydrogenation of indo line to form indole was described, and a mechanistic study on several indo line derivatives is provided in Chapter 5. Changes in electron density on the nitrogen of indo line did not abolish dehydrogenation activity. Therefore, the mechanism of indo line dehydrogenation appeared to be initiated by C-H bond breakage. Studies of the exact mechanism of indo line and indole dehydrogenation remain a major obstacles because we lacked appropriate deuterated compounds, but in silico molecular docking studies provided potentially important molecular interaction (described in Chapter 4 and 5). Arginine 212 in the active site of CYP3A4 may playa vital role in hydrogen bonding with the indo line nitrogen and hence positioning the C-2 and C-3 carbons of indoline 37 close to the heme iron to dramatically lower C-H bond breakage and facilitate selective dehydrogenation. References Adas F, Berthou F, Picart D, Lozac'h P, Beauge F and Amet Y (1998) Involvement of cytochrome P450 2El in the (omega-I)-hydroxylation of oleic acid in human and rat liver microsomes. J Lipid Res 39:1210-1219. Agoram B, Woltosz WS and Bolger MB (2001) Predicting the impact of physiological and biochemical processes on oral drug bioavailability. Adv Drug Deliv Rev 50 Suppl1:S41-67. ' Alvarez-Diez TM and Zheng J (2004) Mechanism-based- inactivation of cytochrome P450 3A4 by 4-ipomeanol. Chem Res ToxicoI17:150-157. , Bagdy G, Kecskemeti V, Riba P and Jakus R (2007) Serotonin and epilepsy. J Neurochem 100:857-873. Beaudry F, Coutu M and Brown NK (1999) Determination of drug-plasma protein binding using human serum albumin chromatographic column and multiple linear regression model. Biomed Chromatogr 13:401-406. Bentley 1M, Adams DR, Bebbington D, Benwell KR, Bickerdike MJ, Davidson JE, Dawson CE, Dourish CT, Duncton MA, Gaur S, George AR, Giles PR, Hamlyn RJ, Kennett GA, Knight AR, Malcolm CS, Mansell HL, Misra A, Monck NJ, Pratt RM, Quirk K, Roffey JR, Vickers SP and Cliffe IA (2004) Indoline derivatives as 5-HT(2C) receptor agonists. Bioorg. Med. Chem. Lett. 14:2367- 2370. Bocan TM, Rosebury WS, Mueller SB, Kuchera S, Welch K, Daugherty A and Comicelli JA (1998) A specific 15-lipoxygenase inhibitor limits the progression and monocyte-macrophage enrichment of hypercholesterolemia-induced atherosclerosis in the rabbit. Atherosclerosis 136:203-216. Bolton JL, Comeau E and Vukomanovic V (1995) The influence of 4-alkyl substituents 'on the formation and reactivity of 2-methoxy-quinone methides: evidence that extended pi-conjugation dramatically stabilizes the quinone methide formed from eugenol. Chem Bioi Interact 95:279-290. Bolton JL, Sevestre H, Ibe BO and Thompson JA (1990) Formation and reactivity of alternative quinone methides from butylated hydroxy toluene: possible explanation for species-specific pneumotoxicity. Chem Res ToxicoI3:65-70. 38 Bonnabry P, Sievering J, Leemann T and Dayer P (1999) Quantitative drug interactions prediction system (Q-DIPS): a computer-based prediction and management support system for drug metabolism interactions. Eur J Clin Pharmacol 55:341- 347. Bonnabry P, Sievering J, Leemann T and Dayer P (2001) Quantitative drug interactions prediction system (Q-DIPS): a dynamic computer-based method to assist in the choice of clinically relevant in vivo studies. Clin Pharmacokinet 40:631-640. Boyd DR, Sharma ND, Agarwal R, McMordie RA, Bessems JG, van Ommen B and van Bladeren PJ (1993) Biotransformation of 1,2-dihydronaphthalene and 1,2- dihydroanthracene by rat liver microsomes and purified cytochromes P-450. Formation of arene hydrates of naphthalene and anthracene. Chem. Res. Toxicol. 6:808-812. Bromidge SM, Duckworth M, Forbes IT, Ham P, King FD, Thewlis KM, Blaney FE, Naylor CB, Blackburn TP, Kennett GA, Wood MD and Clarke SE (1997) 6- Chloro-5-methyl-1-[[2-[(2-methyl-3-pyridyl)oxy ]-5-pyridyl]carbamoyl]- indo line (SB-242084): the first selective and brain penetrant 5-HT2C receptor antagonist. J. Med. Chem. 40:3494-3496. Calkin AC, Thomas MC and Cooper ME (2003) MK-767. KyorinlBanyulMerck. Curr Opin Investig Drugs 4:444-448. Caron G, Ermondi G and Testa B (2007) Predicting the Oxidative Metabolism of Statins: An Application of the MetaSite(R) Algorithm. Pharm Res. Cavalli A, Poluzzi E, De Ponti F and Recanatini M (2002) Toward a pharmacophore for drugs inducing the long QT syndrome: insights from a CoMF A study of HERG K(+) channel blockers. J Med Chem 45:3844-3853. Chan WK, Nguyen LT, Miller VP and Harris RZ (1998) Mechanism-based inactivation of human cytochrome P450 3A4 by grapefruit juice and red wine. Life Sci 62:PL135-142. . Chun YJ, Ryu SY, Jeong TC and Kim MY (2001) Mechanism-based inhibition of human cytochrome P450 lA1 by rhapontigenin. Drug Metab Dispos 29:389-393. Chung SM, Park EJ, Swanson 8M, Wu TC and Chiou WL (2001) Profound effect of plasma protein binding on the polarized transport of furosemide and verapamil in the Caco-2 model. Pharm Res 18:544-547. Colmenarejo G, Alvarez-Pedraglio A and Lavandera JL (2001) Cheminformatic models to predict ~inding affmities to human serum albumin. J Med Chem 44:4370-4378. 39 Coon MJ (2005) Cytochrome P450: nature's most versatile biological catalyst. Annu Rev Pharmacol ToxicoI45:1-25. Correia MA and Ortiz de Montellano PR (2005) Inhibition of cytochrome P450 enzymes, in Cytochrome P450: structure, mechnism, and biochemistry (Ortiz de Montellano PR ed) pp 247-322, Kluwer AcademicIPlenum Publishers. Crivori P, Cruciani G, Carrupt PA and Testa B (2000) Predicting blood-brain barrier penneation from three-dimensional molecular structure. J Med Chem 43:2204- 2216. Cruciani G, Carosati E, De Boeck B, Ethirajulu K, Mackie C, Howe T and Vianello R (2005) MetaSite: understanding metabolism in human cytochromes from the perspective of the chemist. J Med Chem 48:6970-6979. Cutler DM (2007) The demise of the blockbuster? N Engl J Med 356:1292-1293. Dalvie DK, Kalgutkar AS, Khojasteh-Bakht SC, Obach RS and O'Donnell JP (2002) Biotransfonnation reactions of five-membered aromatic heterocyclic rings. Chem Res ToxicoI15:269-299. Danielson PB (2002) The cytochrome P450 superfamily: biochemistry, evolution and drug metabolism in humans. Curr Drug Metab 3:561-597. Dansette PM, Bertho G and Mansuy D (2005) First evidence that cytochrome P450 may' catalyze both S-oxidation and epoxidation of thiophene derivatives. Biochem Biophys Res Commun 338:450-455. Dean BJ, Chang S, Silva Elipe MY, Xia YQ, Braun M, Soli E, Zhao Y, Franklin RB and Karanam B (2007) Metabolism of MK-0524, a prostaglandin D2 receptor 1 antagonist, in microsomes and hepatocytes from preclinical species and humans. Drug Metab Dispos 35:283-292. Dennison JB, Jones DR, Renbarger JL and Hall SD (2007) Effect of CyP3A5 expression on vincristine metabolism with human liver microsomes. J Pharmacol Exp Ther 321:553-563. Dennison JB, Kulanthaivel P, Barbuch RJ, Renbarger JL, Ehlhardt WJ and Hall SD (2006) Selective metabolism of vincristine in vitro by CYP3A5. Drug Metab Dispos 34:1317-1327. Desta Z, Soukhova N and Flockhart DA (2002) In vitro inhibition of pimozide N­dealkylation by selective serotonin reuptake inhibitors and azithromycin. J Clin Psychopharmacol 22: 162-168. 40 Dupont I, Berthou F, Bodenez P, Bardou L, Guirriec C, Stephan N, Dreano Y and Lucas D (1999) Involvement of cytochromes P-450 2E1 and 3A4 in the 5·hydroxylation of salicylate in humans. Drug Metab Dispos 27:322-326. Durham SK and Pearl GM (2001) Computational methods to predict drug safety liabilities. Curr Opin Drug Discov DeveI4:110-115. Ekins S, Kim RB, Leake BF, Dantzig AH, Schuetz EG, Lan LB, Yasuda K, Shepard RL, Winter MA, Schuetz JD, Wikel JH and Wrighton SA (2002a) Application of three-dimensional quantitative structure·activity relationships of P-glycoprotein inhibitors and substrates. Mol PharmacoI61:974·98·1. Ekins S, Kim RB, Leake BF, Dantzig AH, Schuetz EG, Lan LB, Yasuda K, Shepard RL, Winter MA, Schuetz JD, Wikel JH and Wrighton SA (2002b) Three-dimensional quantitative structure-activity relationships of inhibitors of P-glycoprotein. Mol Pharmacol,61 :964-973. Estavillo C, Lu Z, Jansson I, Schenkman JB and Rusling JF (2003) Epoxidation of styrene by human cyt P450 1A2 by thin film electrolysis and peroxide activation compared to solution reactions. Biophys Chern 104:291-296. Evans DC, Watt AP, Nicoll-Griffith DA and Baillie TA (2004) Drug-protein adducts: an industry perspective on minimizing the potential for drug bioactivation in drug discovery and development. Chern Res ToxicoI17:3-16. ' Fan PW, Gu C, Marsh SA and Stevens JC (2003) Mechanism-based inactivation of cytochrome P450 2B6 by a novel tenninal acetylene inhibitor. Drug Metab Dispos 31:28-36. Fan PW, Zhang F and Bolton JL (2000) 4-Hydroxylated metabolites of the antiestrogens tamoxifen and toremifene are metabolized to unusually stable quinone methides. Chem Res ToxicoI13:45-52. Faucette SR, Hawke RL, Shord SS, Lecluyse EL and Lindley CM (2001) Evaluation of the contribution of cytochrome P450 3A4 to human liver microsomal bupropion hydroxylation. Drug Metab Dispos 29:1123-1129. Fishman J (1982) Biochemical mechanism of aromatization. Cancer Res. 42:3277s- 3280s. Froldi G, Silvestrin B, Dorigo Pand Caparrotta L (2004) Gramine: a vasorelaxing alkaloid acting on 5-HT(2A) receptors. Planta Med 70:373-375. Gazi L, Bobimac I, DanzeisenM, Schupbach E, Bruinvels AT, Geisse S, Sommer B, Hoyer D, Tricklebank M and Schoeffler P (1998) The agonist activities of the 41 putative antipsychotic agents, L-745,870 and U-101958 in HEK293 cells expressing the human dopamine D4.4 receptor. Br J PharmacoI124:889-896. Gelb MH, Heimbrook DC, Malkonen P and Sligar SG (1982) Stereochemistry and deuterium isotope effects in camphor hydroxylation by the cytochrome P450cam monoxygenase system. Biochemistry 21:370-377. Ghanayem BI, McDaniel LP, Churchwell MI, Twaddle NC, Snyder R, Fennell TR and Doerge DR (2005) Role of CYP2E1 in the epoxidation of acrylamide to glycidamide and formation of DNA and hemoglobin adducts. Toxicol Sci 88:311- 318. Ghosh S, Nie A, An J and Huang Z (2006) Structure-based virtual screening of chemical libraries for drug discovery. Curr Opin Chem Bioi 10:194-202. Gillam EM, Notley LM, Cai H, De Voss JJ and Guengerich FP (2000) Oxidation of indole by cytochrome P450 enzymes. Biochemistry 39:13817-13824. Good A (2001) Structure-based virtual screening protocols. Curr Opin Drug Discov DeveI4:301~307. Greene N (2002) Computer systems for the prediction of toxicity: an update. Adv Drug Deliv Rev 54:417-431. Groves IT and McClusky GA (1978) Aliphatic hydroxylation by highly purified -liver microsomal cytochrome P-450. Evidence for a carbon radical intermediate. Biochem Biophys Res Commun 81:154-160. Gu C, Collins R, Holsworth DD, Walker GS and Voorman RL (2006) Metabolic aromatization ofN-alkyl-1,2,3,4-tetrahydroquinoline substructures to quinolinium by human liver micro somes and horseradish peroxidase. Drug Metab Dispos 34:2044-2055. Guan X, Fisher MB, Lang DH, Zheng YM, Koop DR and Rettie AE (1998) Cytochrome P450-dependent de saturation of lauric acid: isoform selectivity and mechanism of formation of 11-dodecenoic acid. Chem Bioi Interact 110:103-121. Guengerich FP (2001a) Common and uncommon cytochrome P450 reactions related to metabolism and chemical toxicity. Chem Res ToxicoI14:611-650. Guengerich FP (2001b) Uncommon P450-catalyzed reactions. Curr Drug Metab 2:93- 115. Guengerich FP (2003) Cytochrome P450 oxidations in the generation of reactive electrophiles: epoxidation and related reactions. Arch. Biochem. Biophys. 409:59- 71. 42 Guengerich FP (2005) Human cytochrome P450 enzymes, in Cytochrome P450: structure, mechnism, and biochemistry (Ortiz de Montellano PR ed) pp 377-530, Kluwer AcademicfPlenum Publishers. Guengerich FP (2006) Cytochrome P450s and other enzymes in drug metabolism and toxicity. Aaps J8:E101-111. Guengerich FP and Bocker RH (1988) Cytochrome P-450-catalyzed dehydrogenation of 1,4-dihydropyridines. J BioI Chem 263:8168-8175. Guengerich FP and Kim DH (1991) Enzymatic oxidation of ethyl carbamate to vinyl carbamate and its role as an intermediate in the formation of 1,N6- ethenoadenosine. Chem Res ToxicoI4:413-421. Ha-Duong NT, Dijols S, Macherey AC, Goldstein JA, Dansette PM and Mansuy D (2001) Tic10pidine as a selective mechanism-based inhibitor of human cytochrome P450 2C19. Biochemistry 40:12112-12122. Han DH, Pearson PG, Baillie TA, Dayal R, Tsang LH and Gescher A (1990) Chemical synthesis and cytotoxic properties ofN-alkylcarbamic acid thioesters, metabolites of hepatotoxic formamides. Chem Res ToxicoI3:118-124. He K, Woolf TF and Hollenberg PF (1999) Mechanism-based inactivation of cytochrome P-450-3A4 by mifepristone (RU486). J Pharmacal Exp Ther 288:791-797. He MM, Smith AS, Oslob JD, Flanagan WM, Braisted AC, Whitty A, Cancilla MT, Wang J, Lugovskoy AA, Yoburn JC, Fung AD, Farrington G, Eldredge JK, Day ES, Cruz LA, Cachero TO, Miller SK, Friedman JE, Choong IC and Cunningham BC (2005) Small-molecule inhibition oflNF-alpha. Science 310:1022-1025. Hecht SS, Hochalter JB, Villalta PW and Murphy SE (2000) 2f-Hydroxylation of nicotine by cytochrome P450 2A6 and human liver microsomes: formation of a lung carcinogen precursor. Proc Natl Acad Sci USA 97:12493-12497. Hou TJ and Xu XJ (2003) ADME evaluation in drug discovery. 3. Modeling blood-brain barrier partitioning using simple molecular descriptors. J Chem In! Com put Sci 43:2137-2152. Hutzler 1M, Steenwyk RC, Smith EB, Walker GS and Wienkers LC (2004) Mechanism­based inactivation of cytochrome P450 2D6 by 1-[(2-ethyl-4-methyl-1H­imidazol- 5-yl)methyl]- 4-[4-(trifluoromethyl)-2-pyridinyl]piperazine: kinetic characterization and evidence for apoprotein adduction. Chem Res Taxicol 17:174-184. 43 Huuhka K, Riutta A, Haataja R, Ylitalo P and Leinonen E (2006) The effect of CYP2C 19 substrate on the metabolism of melatonin in the elderly: A randomized, double­blind, placebo-controlled study. Methods Find Exp Clin PharmacoI28:447-4S0. Hwang HS, Moon EY, Seong SK, Choi CH, Chung YH, Jung SH, Lee DK and Y oon SJ (1999) Characterization of the anticancer activity of DW2282, a new anticancer agent. Anticancer Res. 19:5087-5093. Iida I, Miyata A, Arai M, Hirota M, Akimoto M, Higuchi S, Kobayashi K and Chiba K (2004) Catalytic roles of CYP2C9 and its variants (CYP2C9*2 and CYP2C9*3) in lomoxicam 5'-hydroxylation. Drug Metab Dispos 32:7-9. Isin EM and Guengerich FP (2007) Complex reactions catalyzed by cytochromeP4S0 enzymes. Biochim Biophys Acta 1770:314-329. Iwata H, Tezuka Y, Kadota S, Hiratsuka A and Watabe T (2005) Mechanism-based inactivation of human liver microsomal CYP3A4 by rutaecarpine and limonin from Evodia fruit extract. Drug Metab Pharmacokinet 20:34-45. Iyer M, Mishru R, Han Y and Hopfinger AJ (2002) Predicting blood-brain barrier partitioning of organic molecules using membrane-interaction QSAR analysis. Pharm Res 19:1611-1621. Iyer M, Tseng YJ, Senese CL, Liu J and Hopfinger AJ (2007) Prediction and mechanistic interpretation of human oral drug absorption using MI -QSAR analysis. Mol Pharm 4:218-231. Jayyosi Z, Knoble D, Muc M, Erick J, Thomas PE and Kelley M (1995) Cytochrome P- 450 2E1 is not the sole catalyst of chlorzoxazone hydroxylation in rat liver microsomes. off. J Pharmacol Exp Ther 273: 1156-1161. Jean P, Lopez-Garcia P, Dansette P, Mansuy D and Goldstein JL (1996) Oxidation of tienilic acid by human yeast-expressed cytochromes P-4S0 2C8, 2C9, 2C18 and 2C 19. Evidence that this drug is a mechanism-based inhibitor specific for cytochrome P-4S0 2C9. Eur J Biochem 241:797-804. Jensen MT, Cox RP and Jensen BB (1995) 3-Methylindole (skatole) and indole production by mixed populations of pig fecal bacteria. Appl Environ Microbiol 61 :3180-3184. Jin S, Makris TM, Bryson T A, Sligar SG and Dawson JH (2003) Epoxidation of ole fins by hydroperoxo-ferric cytochrome P4S0. J Am Chem Soc 125:3406-3407. Jolivette LJ and Ekins S (2007) Methods for predicting human drug metabolism. Adv Clin Chem 43:131-176. 44 Jushchyshyn MI, Kent UM and Hollenberg PF (2003) The mechanism-based inactivation of human cytochrome P450 2B6 by phencyclidine. Drug Metab Dispos 31:46-52. Kalgutkar AS, Helme KR, Lame ME, Vaz AD, Collin C, Soglia JR, Zhao SX and Hop CE (2005a) Metabolic activation of the nontricyclic antidepressant trazodone to electrophilic quinone-imine and epoxide intermediates in human liver microsomes and recombinant P4503A4. Chem BioI Interact 155: 10-20. Kalgutkar AS, Vaz AD, Lame ME, Henne KR, Soglia J, Zhao SX, Abramov YA,' Lombardo F, Collin C, Hendsch ZS and Hop CE (2005b) Bioactivation of the nontricyclic antidepressant nefazodone to a reactive quinone-imine species in human liver microsomes and recombinant cytochrome P450 3A4. Drug Metab Dispos 33:243-253. Karasevich E1, Shestakov AF and Shilov AE (1999) Novel mechanism of aliphatic hydroxylation in enzymatic and biomimetic systems: problems of the reaction dynamics. Exp Toxicol PathoI51:335-341. Kassahun K, Skordos K, McIntosh I, Slaughter D, Doss GA, Baillie TA and Yost GS (2005) Zafirlukast metabolism by cytochrome P450 3A4 produces an electrophilic alpha,beta-unsaturated iminium species that results in the selective mechanism-based inactivation of the enzyme. Chem. Res. Toxicol. 18:1427-1437. Kelavkar UP, Nixon JB, Cohen C, Dillehay D, Eling TE and Badr KF (2001) Overexpression of 15-lipoxygenase-1 in PC-3 human prostate cancer cells increases tumorigenesis. Carcinogenesis 22:1765-1773. Kent UM, Juschyshyn MI and Hollenberg PF (2001) Mechanism-based inactivators as probes of cytochrome P450 structure and function. Curr Drug Metab 2:215-243. Kent UM, Yanev S and Hollenberg PF (1999) Mechanism-based inactivation of cytochromes P450 2B 1 and P450 2B6 by n-propylxanthate. Chem Res Toxicol 12:317-322. Khojasteh-Bakht SC, Koenigs LL, Peter RM, Trager WF and Nelson SD (1998) (R)-(+)­Menthofuran is a potent, mechanism-based inactivator of human liver cytochrome P450 2A6. Drug Metab Dispos 26:701-704. Kim D and Guengerich FP (2004) Selection of human cytochrome P450 1A2 mutants with enhanced catalytic activity for heterocyclic amine N-hydroxylation. Biochemistry 43:981-988. Kim D, Wu ZL and Guengerich FP (2005a) Analysis of coumarin 7-hydroxylation activity of cytochrome P450 2A6 using random mutagenesis. J BioI Chem 280:40319-40327. 45 Kim JS, Baek SJ, Bottone FG, Jr., Sali T and Eling TE (2005b) Overexpression of 15- lipoxygenase-1 induces growth arrest through phosphorylation of p53 in human colorectal cancer cells. Mol Cancer Res 3:511-517. Kitchen DB, Decornez H, Furr JR and Bajorath J (2004) Docking and scoring in virtual screening for drug discovery: methods and applications. Nat Rev Drug Discov 3:935-949. Knickle LC, Webb CD, House AA and Bend JR (1993) Mechanism-based inactivation of cytochrome P450 1A1 by N-aralkyl-1-aminobenzotriazoles in guinea pig kidney in vivo and in vitro: minimal effects on metabolism of arachidonic acid by renal P450-dependent monooxygenases. J Pharmacol Exp Ther 267:758-764. Koenigs LL, Peter RM, Thompson SJ, Rettie AE and Trager WF (1997) Mechanism­based inactivation of human liver cytochrome P450 2A6 by 8-methoxypsoralen. Drug Metab Dispos 25:1407-1415. Koenigs LL and Trager WF (1998) Mechanism-based inactivation of P450 2A6 by furanocoumarins. Biochemistry 37: 10047 -10061. Kopp J and Schwede T (2004) Automated protein structure homology modeling: a progress report. Pharmacogenomics 5:405-416. Kraemer T, Pflugmann T, Bossmann M, Kneller NM, Peters FT, Paul LD, Springer D, Staack RF and Maurer HH (2004) F enproporex N -dealky lation to amphetamine-­enantioselective in vitro studies in human liver microsomes as well as enantioselective in vivo studies in Wi star and Dark Agouti rats. Biochem PharmacoI68:947-957. Kulthanan K, Tiprungkorn P and Linpiyawan R (2003) Cutaneous reaction to oral antihistamine. Clin Exp DermatoI28:229-230. Lamba J, Larrlba V and Schuetz E (2005) Genetic variants of PXR (NR1I2) and CAR (NR1I3) and their implications in drug metabolism and pharmacogenetics. Curr Drug Metab 6:369-383. Langer T and Hoffmann RD (2001) Virtual screening: an effective tool for lead structure discovery? Curr Pharm Des 7:509-527. Larsson KA, Zetterlund I, De1p G and Jonsson LM (2006) N-Methyltransferase involved in gramine biosynthesis in barley: cloning and characterization. Phytochemistry 67:2002-2008. Lau SS and Zannoni va (1979) Hepatic microsomal epoxidation of bromobenzene to phenols and its toxicological implication. Toxicol. Appl. Pharmacol. 50:309-318. 46 Le Corre P, Parmer RJ, Kailasam MT, Kennedy BP, Skaar TP, Ho H, Leverge R, Smith DW, Ziegler MG, Insel PA, Schork NJ, Flockhart DA and O'Connor D T (2004) Human sympathetic activation by alpha2-adrenergic blockade with yohimbine: Bimodal, epistatic influence of cytochrome P450-mediated drug metabolism. Clin Pharmacol Ther 76:139-153. Lee J, La S, Ahn BR, Jeong TC and Kim DH (2004) Metabolism of 1-(3-[3-(4- cyanobenzyl)-3H-imidazol-4-yl]-propyl)-3-(6-methoxypyridin-3-yl )-1-(2- trifluoromethylbenzyl)thiourea (YH3945), a novel anti-cancer drug, in rats using the 14C-Iabeled compound. Rapid Commun Mass Spectrom 18:1901-1910. Lee SK, Lee DW, Jeon TW, Jin CH, Kim GH, Jun IH, Lee DJ, Kim SI, Kim DH, Jahng Y and Jeong TC (2005) Characterization of the Phase II metabolites of rutaecarpine in rat by liquid chromatography-electrospray ionization-tandem mass spectrometry. Xenobiotica 35:1135-1145. Levesque JF, Day SH, Chauret N:, Seto C, Trimble L, Bateman KP, Silva JM, Berthelette C, Lachance N, Boyd M, Li L, Sturino CF, Wang Z, Zamboni R, Young RN and Nicoll-Griffith DA (2007) Metabolic activation of indole-containing prostaglandin D2 receptor 1 antagonists: Impacts of glutathione trapping and glucuronide conjugation on covalent binding. Bioorg Med Chem Lett. Liao CH, Pan SL, Guh JH, Chang YL, Pai HC, Lin CH and Teng CM (2005) Antitumor mechanism of evodiamine, a constituent from Chinese herb Evodiae fructus, in human multiple-drug resistant breast cancer NCI/ADR-RES cells in vitro and in vivo. Carcinogenesis 26:968-975. Liao D, Basarab GS, Gatenby AA, Valent B and Jordan DB (2001) Structures of trihydroxynaphthalene reductase-fungicide complexes: implications for structure­based design and catalysis. Structure 9:19-27. Liddle C and Goodwin B (2002) Regulation of hepatic drug metabolism: role of the nuclear receptors PXR and CAR. Semin Liver Dis 22:115-122. Lightning LK, Jones JP, Friedberg T, Pritchard MP, Shou M, Rushmore TH and Trager WF (2000) Mechanism-based inactivation of cytochrome P450 3A4 by L- 754,394. Biochemistry 39:4276-4287. Lin HL, Kent UM and Hollenberg PF (2002) Mechanisn1-based inactivation of cytochrome P450 3A4 by 17 alpha-ethynylestradiol: evidence for heme destruction and covalent binding to protein. J Pharmacol Exp Ther 301:160-167. Liu J, Yang L, Li Y, Pan D and Hopfinger AJ (2006) Constructing plasma protein binding model based on a combination of cluster analysis and 4 D-fingerprint molecular similarity analyses. Bioorg Med Chem 14:611-621. 47 Lobell M and Sivarajah V (2003) In silico prediction of aqueous solubility, human plasma protein binding and volume of distribution of compounds from calculated pKa and AlogP98 values. Mol Divers 7:69-87. Lombardo F, Obach RS, Shalaeva MY and Gao F (2002) Prediction of volume of distribution values in humans for neutral and basic drugs using physicochemical measurements and plasma protein binding data. J Med Chem 45:2867-2876. Lombardo F, Obach RS, Shalaeva MY and Gao F (2004) Prediction of human volume of distribution values for neutral and basic drugs. 2. Extended data set and leave­class- out statistics. J Med Chem 47:1242-1250. Lopez-Garcia MP, Dansette PM and Mansuy D (1994) Thiophene derivatives as new mechanism-based inhibitors of cytochromes P-450: inactivation of yeast­expressed human liver cytochrome P-450 2C9 by tienilic acid. Biochemistry 33: 166-175. Lu P, Schrag ML, Slaughter DE, Raab CE, Shou M and Rodrigues AD (2003) Mechanism-based inhibition of human liver microsomal cytochrome P450 1A2 by zileuton, a 5-lipoxygenase inhibitor. Drug Metab Dispos 31:1352-1360. Lyne PD (2002) Structure-based virtual screening: an overview. Drug Discov Today 7:1047-1055. Ma B, Subramanian R, Schrag ML, Rodrigues AD and Tang C (2004) Cytochrome P450 2C8 (CYP2C8)-mediated hydroxylation of an endothelin ETA receptor antagonist in human liver microsomes. Drug Metab Dispos 32:473-478. Ma X, Idle JR, Krausz KW and Gonzalez FJ (2005) Metabolism of melatonin by human cytochromes p450. Drug Metab Dispos 33:489-494. Makris TM, Denisov I, Schlichting I and Silgar SG (2005) Activation of molecular oxygen by cytochrome P450, in Cytochrome P450: structure, mechnism, and biochemistry (Ortiz de Montellano PR ed) pp 149-182, Kluwer AcademiclPlenum Publishers. . Mano J, Torii Y, Hayashi S, Takimoto K, Matsui K, Nakamura K, Inze D, Babiychuk E, Kushnir S and Asada K (2002) The NADPH:quinone oxidoreductase P1-zeta­crystallin in Arabidopsis catalyzes the alpha,beta-hydrogenation of 2-alkenals: detoxication of the lipid peroxide-derived reactive aldehydes. Plant Cell Physio/ 43:1445-1455. Masubuchi Y and Horie T (1999) Mechanism-based inactivation of cytochrome P450s 1A2 and 3A4 by dihydralazine in human liver microsomes. Chem Res Toxicol 12: 1 028-1 032. 48 Matsuoka H, Kato N, Tsuji K, Maruyama N, Suzuki H, Mihara M, Takeda Y and Yano K (1996) Antirheumatic agents. I. Novel methotrexate derivatives bearing an indo line moiety. Chem. Pharm. Bull. (Tokyo) 44:1332-1337. Maurer Hand Kleff I (1988) On the metabolism of ditazole In man. Arzneimitteljorschung 38: 1843-1845. Munro AW, Girvan HM and McLean KJ (2007) Cytochrome P450--redox partner fusion enzymes. Biochim Biophys Acta 1770:345-359. Nagata K, Liberato DJ, Gillette JR and Sasame HA (1986) An unusual metabolite of testosterone. 17 beta-Hydroxy-4,6-androstadiene-3-one. Drug Metab Dispos 14:559-565. Nakagawa Y, Irie K, Yanagita RC, Ohigashi H, Tsuda K, Kashiwagi K and Saito N (2006) Design and synthesis of 8-octyl-benzolactam-V9, a selective activator for protein kinase C epsilon and eta. J. Med. Chem. 49:2681-2688. Nakajima M, Inoue T, Shimada N, Tokudome S, Yamamoto T and Kuroiwa Y (1998) Cytochrome P450 2C9 catalyzes indomethacin O-demethylation in human liver microsomes. Drug Metab Dispos 26:261-266. Nocerini MR, Carlson JR and Yost GS (1984) Electrophilic metabolites of 3- methylindole as toxic intermediates in pulmonary oedema. Xenobiotica 14:561- 564. Noguchi T, Tanaka N, Nishimata T, Goto R, Hayakawa M, Sugidachi A, Ogawa T, Asai F, Matsui Y and Fujimoto K (2006) Indoline derivatives I: synthesis and factor Xa (FXa) inhibitory activities. Chem. Pharm. Bull. (Tokyo) 54:163-174. Norris DA, Leesman GD, Sinko PJ and Grass GM (2000) Development of predictive pharmacokinetic simulation models for drug discovery. J Control Release 65:55- 62. O'Donnell JP, Dalvie DK, Kalgutkar AS and Obach RS (2003) Mechanism-based inactivation of human recombinant P450 2C9 by the nonsteroidal anti­inflammatory drug suprofen. Drug Metab Dispos 31:1369-1377. Ooms F, Weber P, Carrupt PA and Testa B (2002) A simple model to predict blood-brain barrier permeation from 3D molecular fields. Biochim Biophys Acta 1587:118- 125. Ortiz de Montellano PR and De Voss JJ (2005) Substrate oxidation by cytochrome P450 enzymes, in Cytochrome P450: structure, mechnism, and biochemistry (Ortiz de Montellano PR ed) pp 183-~45, Kluwer AcademiclPlenum Publishers. 49 Osterberg T and Norinder U (2000) Theoretical calculation and prediction of P­glycoprotein- interacting drugs using MolSurf parametrization and PLS statistics. Eur J Pharm Sci 10:295-303. Paine MJI, Scrutton NS, Munro A W, Gutierrez A, Roberts GCK and Wolf CR (2005) Electron transfer partners of cytochrome P450, in Cytochrome P450: structure, mechnism, and biochemistry (Ortiz de Montellano PR ed) pp 115-148, Kluwer AcademiclPlenum Publishers. Parker CA, Anderson NJ, Robinson ES, Price R, Tyacke RJ, Husbands SM, Dillon MP, Eglen RM, Hudson AL, Nutt DJ, Crump MP and Crosby J (2004) Harmane and harmalan are bioactive components of classical clonidine-displacing substance. Biochemistry 43: 16385-16392. Patel S, Freedman S, Chapman KL, Emms F, Fletcher AE, Knowles M, Marwood R, McAllister G, Myers J, Curtis N, Kulagowski JJ, Leeson PD, Ridgill M, Graham M, Matheson S, Rathbone D, Watt AP, Bristow LJ, Rupniak NM, Baskin E, Lynch JJ and Ragan CI (1997) Biological profile of L-745,870, a selective antagonist with high affmity for the dopamine D4 receptor. J Pharmacol Exp Ther 283:636-647. Pearce LV, Petukhov PA, Szabo T, Kedei N, Bizik F, Kozikowski AP and Blumberg PM (2004) Evodiamine functions as an agonist for the vanilloid receptor TRPV1. Org Biomol Chem 2:2281-2286. Pennanen S, Kojo A, Pasanen M, Liesivuori J, Juvonen RO and Komulainen H (1996) CYP enzymes catalyze the formation of a terminal olefin from 2-ethylhexanoic acid in rat and human liver. Hum Exp ToxicoI15:435-442. Pikuleva I and Waterman M (1999) Cytochromes P450 in synthesis of steroid hormones, bile acids, vitamin D3 and cholesterol. Mol Aspects Med20:33-42, 43-37. Pleban K and Ecker GF (2005) Inhibitors of p-glycoprotein--lead identification and optimisation. Mini Rev Med Chem 5:153-163. Racha JK, Rettie AE and Kunze KL (1998) Mechanism-based inactivation of human cytochrome P450 lA2 by furafylline: detection of a 1: 1 adduct to protein and evidence for the formation of a novel imidazomethide intermediate. Biochemistry 37:7407-7419. Raghavan N, Zhang D, Zhu M, Zeng J and Christopher L (2005) Cyp2D6 catalyzes 5- hydroxylation of 1-(2-pyrimidinyl)-piperazine, an active metabolite of several psychoactive drugs, in human liver microsomes. Drug Metab Dispos 33 :203-208. Raub TJ (2006) P-glycoprotein recognition of substrates and circumvention through rational drug design. Mol Pharm 3:3-25. 50 Regal KA, Laws GM, Yuan C, Yost GS and Skiles GL (2001) Detection and characterization of DNA ad ducts of 3-methylindole. Chem Res Toxicol 14: 1014- 1024. Reilly CA, Ehlhardt WJ, Jackson DA~ Kulanthaivel P, Mutlib AE, Espina RJ, Moody DE, Crouch DJ and Yost GS (2003) Metabolism of capsaicin by cytochrome P450 produces novel dehydrogenated metabolites and decreases cytotoxicity to lung and liver cells. Chem Res ToxicoI16:336-349. Rettie AE, Sheffels PR, Korzekwa KR, Gonzalez FJ, Philpot RM and Baillie TA (1995) CYP4 isozyme specificity and the relationship between omega-hydroxylation and terminal de saturation ofvalproic acid. Biochemistry 34:7889-7895. Roberts SA (2001) High-throughput screening approaches for investigating drug metaboHsm and pharmacokinetics.Xenobiotica 31:557-589. Robinson DM and Wellington K (2006) Indapamide sustained release: a review of its use in the treatment of hypertension. Drugs 66:257-271. Rodrigues AD, Winchell GA and Dobrinska MR (2001) Use of in vitro drug metabolism data to evaluate metabolic drug-drug interactions in man: the need for quantitative databases. J Clin PharmacoI4i:368-373. Rose K, Hall LH and Kier LB (2002) Modeling blood-brain barrier partitioning using the electrotopological state. J Chem In! Com put Sci 42:651-666. Rowland P, Blaney FE, Smyth MG, Jones JJ, Leydon VR,Oxbrow AK, Lewis CJ, Tennant MG, Modi S, Eggleston DS, Chenery RJ and Bridges AM (2006) Crystal structure of human cytochrome P450 2D6. J BioI Chem 281:7614-7622. Sano H, Noguchi T, Miyajima A, Hashimoto Y and Miyachi H (2006) Anti-angiogenic activity of basic-type, selective cyclooxygenase (COX)-1 inhibitors. Bioorg. Med. Chem. Lett. 16:3068-3072. Sansen S, Yano JK, Reynald RL, Schoch GA, Griffin KJ, Stout CD and Johnson EF (2007) Adaptations for the Oxidation of Polycyclic Aromatic Hydrocarbons Exhibited by the Structure of Human P450 1A2. J BioI Chem 282:14348-14355. Schoch GA, Yano JK, Wester MR, Griffin KJ, Stout CD and Johnson EF (2004) Structure of human microsomal cytochrome P450 2C8. Evidence for a peripheral fatty acid binding site. J BioI Chem 279;9497-9503. Schuster D, Steindl TM and Langer T (2006) Predicting drug metabolism induction in silico. Curr Top Med Chem 6:1627 ... 1640. 51 Selkirk N, Scott C, Ho M, Burton MJ, Watson J, Gaster LM, Collin L, Jones BJ, Middlemiss DN and Price GW (1998) SB-224289--a novel selective (human) 5- HT1B receptor antagonist with negative intrinsic activity. Br J Pharmacol 125:202-208. Shaffer CL, Harriman S, Koen YM and Hanzlik RP (2002) Formation of cyclopropanone during cytochrome P450-catalyzed N-dealkylation of a cyclopropylamine. J Am Chem Soc 124:8268-8274. Shaik . S and De Visser SP (2005) Computational approaches to cytochrome P450 function, in Cytochrome P450: structure. mechnism. and biochemistry (Ortiz de Montellano PR ed) pp 45-85, Kluwer AcademiclPlenum Publishers. Shimada T (2006) Xenobiotic-metabolizing enzymes involved in activation and detoxification of carcinogenic polycyclic aromatic hydrocarbons. Drug Metab Pharmacokinet 21:257-276. Skiles GL and Yost GS (1996) Mechanistic studies on the cytochrome P450-catalyzed dehydrogenation of3-methylindole. Chem Res ToxicoI9:291-297. Skordos KW, Skiles GL, Laycock JD, Lanza DL and Yost GS (1998) Evidence supporting the formation of 2,3-epoxy-3-methylindoline: a reactive intermediate of the pneumotoxin 3-methylindole. Chem. Res. Toxicol. 11:741-749. Smith KS, Smith PL, Heady TN, Trugman JM, Harman WD and Macdonald TL (2003) In vitro metabolism oftolcapone to reactive intermediates: relevance to tolcapone liver toxicity. Chem Res ToxicoI16:123-128. Sordillo LM, Weaver JA, Cao YZ, Corl C, Sylte MJ and Mullarky IK (2005) Enhanced 15-HPETE production during oxidant stress induces apoptosis of endoth,elial cells. Prostaglandins Other Lipid Mediat 76:19-34. Spink DC, Eugster HP, Lincoln DW, 2nd, Schuetz JD, Schuetz EG, Johnson JA, Kaminsky LS and Gierthy JF (1992) 17 beta-estradiol hydroxylation catalyzed by human cytochrome P450 1A1: a comparison of the activities induced by 2,3,7,8- tetrachlorodibenzo-p-dioxin in MCF-7 cells with those from heterologous expression of the cDNA. Arch Biochem Biophys 293:342-348. Sridar C, Kobayashi Y, Brevig H, Kent UM, Puppali SG, Rimoldi JM and Hollenberg PF (2006) Synthesis of substituted phenyl' diaziridines and characterization as mechanism-based inactivators of human cytochrome P450 2B6. Drug Metab Dispos 34:1849-1855. Subehan, Usia T, Kadota Sand Tezuka Y (2006) Mechanism-based inhibition of human liver microsomal cytochrome P450 2D6 (CYP2D6) by alkamides of Piper nigrum. Planta Med 72:527~532. 52 Sun H, Ehlhardt WJ, Kulanthaivel P, Lanza DL, Reilly CA and Yost GS (2007) Dehydrogenation of indo line by cytochrome P450 enzymes: a novel "aromatase" process. J Pharmacol Exp Ther 322:843-851. Tang Hand Mayersohn M (2006) On the observed large interspecies overprediction of human clearance ("vertical allometry") of UCN-O 1: further support for a proposed model based on plasma protein binding. J Clin Pharmacol 46:398-400. Thornton-Manning J, Appleton ML, Gonzalez FJ and Yost GS (1996) Metabolism of 3- nlethylindole by vaccinia-expressed P450 enzymes: correlation of 3- methyleneindolenine formation and protein-binding. J Pharmacol Exp Ther 276:21-29. Turpeinen M, Nieminen R, Juntunen T, Taavitsainen P, Raunio Hand Pelkonen 0 (2004) Selective inhibition of CYP2B6-catalyzed bupropion° hydroxylation in human liver microsomes in vitro. Drug Metab Dispos 32:626-631. Ueng YF, Don MJ, Jan WC, Wang SY, Ho LK and Chen CF (2006) Oxidative metabolism of the alkaloid rutaecarpine by human cytochrome P450. Drug Metab Dispos 34:821-827. Ueng YF, Jan WC, Lin LC, Chen TL, Guengerich FP and Chen CF (2002) The alkaloid rutaecarpine is a selective inhibitor of cytochrome P450 lA in mouse and human liver microsomes. Drug Metab Dispos 30:349-353. van de Waterbeemd H and Gifford E (2003) ADMET in silico modelling: towards prediction paradise? Nat Rev Drug Discov 2:192-204. Vyas KP, Kari PH, Prakash SR and Duggan DE (1990) Biotransformation of lovastatin. II. In vitro metabolism by rat and mouse liver microsomes and involvement of cytochrome P-450 in dehydrogenation of lovastatin. Drug Metab Dispos 18:218- 222. Wang C, Wang MW, Tashiro S, Onodera Sand Ikejima T (2005) Evodiamine induced human melanoma A3 75-S2 cell death partially through interleukin 1 mediated pathway. BioI Pharm Bull 28:984-989. Ward BA, Morocho A, Kandil A, Galinsky RE, Flockhart DA and Desta Z (2004) Characterization of human cytochrome P450 enzymes catalyzing domperidone N­dealkylation and hydroxylation in vitro. Br J Clin PharmacoI58:277-287. Waszkowycz B (2002) Structure-based approaches to drug design and virtual screening. Curr Opin Drug Discov DeveI5:407-413. Wells PG, Mackenzie PI, Chowdhury JR, Guillemette C, Gregory PA, Ishii Y, Hansen AJ, Kessler FK, Kim PM, Chowdhury NR and Ritter JK (2004) Glucuronidation 53 and the UDP-glucuronosyltransferases in health and disease. Drug Metab Dispos 32:281-290. Wen X, Wang JS, Neuvonen PJ and Backman JT (2002) Isoniazid is a mechanism-based inhibitor of cytochrome P450 lA2, 2A6, 2C 19 and 3A4 isoforms in human liver microsomes. Eur J Clin Pharmacol 57 :799-804. Wester MR, Yano JK, Schoch GA, Yang C, Griffin KJ, Stout CD and Johnson EF (2004) The structure of human cytochrome P450 2C9 complexed with flurbiprofen at 2.0-A resolution. J Bioi Chem 279:35630-35637. Whitty A and Kumaravel G (2006) Between a rock and a hard place? Nat Chern Bioi 2:112-118. Willson TM and Kliewer SA (2002) PXR, CAR and drug metabolism. Nat Rev Drug Discov 1:259-266. Winkler DA and Burden FR (2004) Modelling blood-brain barrier partitioning using Bayesian neural nets. J Mol Graph Model 22:499-505 . Wood AW, Levin W, Thakker DR, Vagi H, Chang RL, Ryan DE, Thomas PE, Dansette PM, Whittaker N, Turujman S, Lehr RE, Kumar S, Jerina DM and Conney AH (1979) Biological activity of benzo[ e ]pyrene. An assessment based on mutagenic activities and metabolic profiles of the polycyclic hydrocarbon and its derivatives. J. Bioi. Chern. 254:4408-4415. Yan Z, Easterwood LM, Maher N, Torres R, Huebert N and Yost GS (2007) Metabolism and bioactivation of 3-methylindole by human liver microsomes. Chern Res Toxico120:140-148. Yanev S, Kent UM, Pandova B and Hollenberg PF (1999) Selective mechanism-based inactivation of cytochromes P-450 2B 1 and P-450 2B6 by a series of xanthates. Drug Metab Dispos 27:600-604. Yano JK, Hsu MH, Griffin KJ, Stout CD and Johnson EF (2005) Structures of human microsomal cytochrome P450 2A6 complexed with coumarin and methoxsalen. Nat Struct Mol Bioi 12:822-823. Yano JK, Wester MR, Schoch GA, Griffin KJ, Stout CD and Johnson EF (2004) The structure of human microsomal cytochrome P450 3A4 determined by X -ray crystallography to 2.05-A resolution. J Bioi Chern 279:38091-38094. Yin W, Mitra K, Steams RA, Baillie TA and Kumar S (2004) Conversion of the 2,2,6,6- tetramethylpiperidine moiety to a 2,2-dimethylpyrrolidine by cytochrome P450: evidence for a mechanism involving nitroxide radicals and heme iron. Biochemistry 43:5455-5466. 54 Yost GS (1989) Mechanisms of 3-methylindole pneumotoxicity. Chem Res Toxicol 2:273-279. Yost GS (2001) Bioactivation of toxicants by cytochrome p450-mediated dehydrogenation mechanisms. Adv Exp Med Bioi 500:53-62. Yu AM, Granvil CP, Haining RL, Krausz KW, Corchero J, Kupfer A, Idle JR and Gonzalez FJ (2003) The relative contribution of monoamine oxidase and cytochrome p450 isozymes to the metabolic deamination of the trace amine tryptamine. J Pharmacol Exp Ther 304:539-546. Yun CH, Hammons GJ, Jones G, Martin MV, Hopkins NE, Alworth WL and Guengerich FP (1992) Modification of cytochrome P450 1A2 enzymes by the mechanism­based inactivator 2-ethynylnaphthalene and the photoaffinity label 4- azidobiphenyl. Biochemistry 31: 1 0556-10563. Zhang Y, Wu LJ, Tashiro S, Onodera Sand Ikejima T (2004) Evodiamine induces tumor cell death through different pathways: apoptosis and necrosis. Acta Pharmacol Sin 25:83w 89. Zhang Z, Chen Q, Li Y, Doss GA, Dean BJ, Ngui JS, Silva Elipe M, Kim S, Wu JY, Dininno F, Hammond ML, Steams RA, Evans DC, Baillie TA and Tang W (2005) In vitro bioactivation of dihydrobenzoxathiin selective estrogen receptor modulators by cytochrome P450 3A4 in human liver microsomes: formation of reactive iminium and quinone type metabolites. Chem Res ToxicoI18:675-685. Zhao XJ, Jones DR, Wang YH, Grimm SW and Hall SD (2002) Reversible and irreversible inhibition of CYP3A enzymes by tamoxifen and metabolites. Xenobiotica 32:863-878. CHAPTER 2 HOMOLOGY MODELING OF CYP2F3 There are approximately 1,200,000 protein sequences but only 20,000 three­dimensional structures were determined by X -ray crystallography or NMR methods (Sanchez and Sali, 2000). To facilitate experimental studies of proteins, in silico homology modeling has been extensively used in recent years (Tanaka et aI., 2004a; Tanaka et aI., 2004b; Pogozheva et aI., 2005; Chavez-Gutierrez et aI., 2006; Jorgensen et aI., 2007). Cytochrome P450 homology modeling can be applied to identify important residues within the active site of P450s to direct further site-directed mutagenesis work (Szklarz and Halpert, 1997; Gonzalez-Segura et aI.,. 2005; Miura et aI., 2006), for searching ligand binding confonnations and orientations within the active site of P450s by further molecular docking studies (Bathelt et aI., 2002; Hamza, 2002; Chae et aI., 2006; Liu et aI., 2006), as well as for refining models determined from NMR or X-ray crystallography but in relatively low resolution. For proteins in the same family or subfamily, they are rather conserved in their three-dimensional structures; therefore, it is straightforward to choose a template within the same subfamilies for homology modeling. 56 Homology Modeling Generally there are four steps for homology modeling: searching for related protein structures and selecting one or more templates, target and template alignment, model building, and model evaluation. If the results from the evaluation are not satisfactory, the above steps can be repeated until a satisfactory model is constructed. Template Selection Templates could be searched from online databases such as NCBI and PDB-Blast. There are several factors to be considered when selecting a template: the quality of a template (higher resolution), the higher sequence similarity (within the same subfamilies), and the similarity between the "environment" of the template and target (such as a close size of active site). In addition, mUltiple templates (or multiple template segments) could enhance the model accuracy (Fiser and Sali, 2003). Alignment There are several protocols for protein sequence alignment. Pairwise alignment will define scores for every possible amino acid match and then search for the highest score available to build a comparison matrix based on the two sequences to be aligned, but the pitfall is the highest score is not necessary to be the best alignment. Other factors such as insertions and gaps either long or short in the sequence have to be considered using the "gap open penalties" ,"gap extension penalties" as well as "affine gap penalties" methods (Sali and Blundell, 1990; Mullan, 2006). Another protocol called multiple sequence alignments can increase the accuracy and enhance the alignment 57 quality. An example is CLUSTAL W (Ogden and Rosenberg, 2007), by which the alignment errors could be reduced by approximately one third relative to those that were performed with standard sequence alignment techniques, especially as the number of gaps increases and the similarity between the sequences decreases. For target-template sequences with high similarities, such as homology modeling between CYP2 subfamilies, the alignment is relatively simple using standard sequence­sequence alignment methods, as above. For target-template sequences with low similarities, however, the alignment is complicated and will affect the quality of the resulting model. Generally, gaps should be avoided in the secondary structure elements (a-helix and p-sheet), in buried regions, or between two residues that are far apart in space (Fiser and Sali, 2003). Multiple template methods could be used between templates to avoid gaps and other significant errors. Model Building Various methods have been used to construct three-dimensional structures of target proteins, which can be divided into three groups: the rigid-body assembly, segment-based matching, and spatial restraints satisfaction. The rigid-body assembly is the oldest homology modeling method, which relies on only a few core regions from loops and side chains, such as WHAT IF (Vriend, 1990), which can calculate solvent­accessible surfaces, contact surfaces, van der Waals exclusion volumes and Connolly surfaces. In another program, 3D-JIGSAW (Bates et aI., 2001), all loops are considered for replacement via database searches combined with torsion angle adjustments, and the 58 removal of steric clashes by energy minimization using the program CHARMM (patel and Brooks, 2004; Patel et aI., 2004). The segment-based matching method considers the approximate positions of conserved atoms fIrst, from which the coordinates of other atoms are calculated. SWISS­MODEL (Schwede et aI., 2003) is a segment-based modeling program, in which the best loop and flanking residues are selected using a scoring scheme based on GROMOS96 force field energy, steric hindrance and favorable interactions such as hydrogen bond formation. In addition, side chains can also be reconstructed based on the weighted positions of corresponding residues in the template structures by evaluating favorable hydrogen bonds and disulfide bridge interactions. The spatial restraints satisfaction me~od is the newest, in which all the spatial restraints such as nonbonded atomic distances and dihedral angles ~s well as bond length and bond angle preferences as derived from the alignment are optimized, a similar protocol that used for the structural determination by NMR spectroscopy. Modeller (Sali and Blundell, 1993; Sanchez and Sali, 1997; Fiser et aI., 2000; Sanchez and Sali, 2000) used in this dissertation, belongs to this group that uses a molecular mechanics force field of CHARMM-22 to optimize the target model with minimal violations, using both conjugate gradients and molecular dynamics methods. In detail, from template structures, spatial restraints such as atom-atom distances and dihedral angles are extracted first, and followed by the alignment of the target protein and the template to determine equivalent residues. The alignment accuracy usually affects the model accuracy, but for P450 modeling in this study, this is not true because the sequence identity between target P450s and templates are very high (over 50%). In addition, 59 stereochemical restraints are also included. Finally, the target model is optimized to best satisfy the spatial restraints. Two loop modeling methods are also included in Modeller, the ab initio loop prediction based on the enumeration of conformations using a energy function, and the database search approach through a database of many known protein structures, which can select, superpose and anneal the chosen loop onto the stem region in the main chain, followed by energy-based refinement. Model Evaluation and Postmodification Several programs were developed to test the quality of a homology model for both internal and external evaluations. Internal evaluation checks model self-consistency or stereochemistry factors such bond angles atom-atom distances and dihedral angles. Programs Pro Check and WhatCheck are often used. Errors are most important when a residue is within the active site. External evaluation tests if the template used is correct, when different templates are used. The Prosa Z score measures the compatibility between the sequence and the structure. In addition, the unreliable regions in the model could be predicted by Prosa II by calculating the "pseudo-energy" profile of each residue in the model. Some regions of cytochromes P450, such as the F -G loop, are difficult to model because they could undergo large conformational changes when substrates are bound. However, this has little effect on substrate binding to the heme region, especially when the molecular size of substrates is relatively small. Another factor that could affect the model accuracy is the sequence identity but this is not a problem for P450 modeling because of their high similarities. We can predict that a CYP3A5 homology model that 60 was based on a CYP3A4 x-ray structure would be very close to the experimentally determined structure, because the two enzymes have super high sequence identities. Experimental observations such as in vitro metabolism and site-directed mutagenesis should also be used to evaluate the model. In addition, multiple templates can improve the construction of a model because the model tends to inherit the best regions from each template. Small Molecule Docking The molecular docking studies were applied in this investigation and also other studies in subsequent chapters to predict the favored substrate binding conformation and orientation in the active site of cytochrome P450s. Low resolution of crystallographic proteins, high-degree of ligand flexibility, participation of water molecules in protein­ligand interactions are major obstacles to be considered when performing these molecular docking studies. Search Methods for Flexible Substrates There are three major categories of searching methods: systematic, stochastic and simulation methods. A systematic search explores all the degrees of freedom of a substrate, and hence, has the limitation in combinatorial explosion. Practically, with programs like DOCK 4.0 and FlexX, substrates are usually incrementally grown into the active site of P450s as a de novo ligand-design strategy (fragmentation) with rigid cores docked first and then the flexible parts of the substrate (Kramer et aI., 1999; 61 Schellhammer and Rarey, 2004; Cross, 2005). The search is based on libraries of pre­generated conformations. Stochastic algorithms apply random changes to the substrate and evaluate a pre~ defined probability function. Monte CarlQ and genetic algorithms are two major stochastic algorithms, both of which were used in Autodock (Morris et aI., 1996; Osterberg et aI., 2002). DOCK and GOLD (Verdonk et aI., 2003) implemented genetic algorithms. Molecular dynamics generally only accommodate ligands in a local minima of the energy surface, because it could not cross high-energy barriers within feasible simulation time periods. Several strategies are applied in molecular dynamics to avoid local minima problems. A strategy might be to simulate only a part of a P450~substrate system, or calculate different substrate positions. In addition, search methods such as molecular dynamics with Monte Carlo calculations, rotamer libaries (a dead-end elimination algorithm) and protein ensemble grids are used for flexible P450s (Paulsen et aI., 1996; Bathelt et aI., 2002). Scoring Functions A score function is required to rank the favored ligand conformations in the active site of P450s. Fundamentally these scoring functions in docking programs make various assumptions; therefore, the evaluation process is rather simplified and only could partially and roughly simulate the biological process in the body. Researchers have tried to decrease the assumptions and combine as many as different energy parameters to make 62 the modeling program more reliable. Three types of functions have been used in today's docking programs: force-field-based, empirical and knowledge-based scoring functions. 1. Force-field-based scoring. In force-filed-based scoring functions the internal protein energy is not considered, and usually only two energies are calculated: the receptor-ligand interaction energy and internal ligand energy induced by the binding process. The van ier Waals energy is generally given by a Lenn~rd-Jones potential function such as a 12-6 Lennard-Jones potential in D-Score (Baak et a!., 2001), 8-4 Lennard-lones potential in G-score and 12-10 Lennard-Jones potential in Autodock. Electrostatic energy is calculated using Coulombic fonnulation with a distance-dependent dielectric function. Hydrogen bonding, torsion and other energies are also considered in some docking programs (Kitchen et at, 2004; Verdonk et at, 2004). The Autodock program is based on the AMBER force to calculate these two energies as follows, where for two atoms i and j, A andB are van der Waals parameters, C and D are hydrogen bond parameters, d is the interatomic distance, q is atomic partial charge, e is a distance-dependent dielectric function, and E(t) is a angular weight factor. There are different schemes in Autodock that could consider more parameters such as torsion energy and solvation terms. (Receptor-ligand energy in Autodock) (Internal ligand energy in Autodock) 63 G-Score is another popular force-field-based function in Tripos force field and it includes different hydrogen-bondi,ng and torsional entropy terms, with equations as follows, where for two atoms i andj, A and B are van der Waals parameters for receptor-ligand energy, C and D are van der Waals parameters for internal ligand energy, d is the interatomic distance. (Receptor-ligand energy in G-Score) (Internal ligand energy in G-Score) ( c.. D .. J 1 [" 11 (I )~ E wn!' + Etorsion = ~ l~ - ~"u + ~ - V 1 + -I I cos~nlco hg d ij d ij lTg 2 n Gold is another force-field-based function included in Insight II, with equations as follows, where for two atoms i and j, A and B are van der Waals. parameters, q is atomic partial charge, g is a distance-dependent dielectric function, and d is the interatomic distance. (Receptor-ligand energy in Gold) E,ww +Eel~~talk = ~i:[[ ~ + :; )+332.0 E ~:~if] (Internal ligand energy in Gold) E E = ~[(.Aij + By J+332.0 qiqj ] wlW + eJectrostalic f;: d; d: E (d ij )d fj 2. Empirical scoring functions depend on the molecular data sets used for regression analyses and fitting that were obtained from experimentally determined 64 binding energies such as using X-ray crystallography, and hence causes conflict between differently fitted scoring functions and new scoring function. The scoring function in empirical scoring is usually simpler than that of force-field scoring function. A popular empirical-based scoring function called ChemScore (Kitchen et aI., 2004; Verdonk et aI., 2004), does not differentiate between different types of hydrogen bonds, but implement ligand rotational entropy, which considers the molecular environment surrounding each rotatable bond. It is calculated in the following form, where the free energy of binding includes the sum of hydrogen bonding, ionic, hydrophobic and ligand rotational entropy four terms. 6.Gbind =ilGH-bond Lf(M,~a)+AGmetal D(M~Aa)+AGJipoD(M)+AGrotor D(l!I,P,;J+AGo H-bond metal lipo ,'otor 3. Knowledge-based scoring. The function in knowledge-based scoring is designed to reproduce experimental structures rather than binding energies, such as DrugScore (Kitchen et aI., 2004; Verdonk et aI., 2004) includes solvent-accessibility corrections to pair-wise potentials. The advantage of the function is its computational simplicity, but a disadvantage is that it is only based on limited sets of protein-ligand complex structures. The function in DrugScore is shown as follows, where SAS is the solvent accessible surface area terms and W is the distance dependent pairwise potential. LlW = yL:LLiWv(r)+ (1- r)X[L~W';(SAS, SASo) + L A~(SAS~SASo)] prot fig itg prOf Overall, to avoid the shortcomings of each individual docking scoring, a recent trend in drug discovery is to combine different scoring schemes, which could improve the probability of identifying 'true' lead or most favored binding conformation. 65 CYP2F3 Site-Directed Mutagenesis CYP2F3 is a goat lung specific P450 enzyme that selectively catalyzes the dehydrogenation of pneumotoxin 3-methylindole to an electrophilic· 3- methyleneindolenine reactive intermediate that could conjugate with protein and DNA to cause toxicities (Wang et ai., 1998). In our previous site-directed mutagenesis study (unpublished data, Table 2.1), we found several mutants within substrate recognition sites (SRS) (Gotoh, 1992) 5 and 6 that changed this dehydrogenation activity. In fact, we found double mutants in SRS 5 (A362I, D363T) and SRS 6 (S476H, S4771) introduced oxygenation activity to CYP2F3 by producing 3-methyloxindole (pathway as described in the introduction section). Additional point mutation studies were also conducted to produce mutants D362T, S476H and S4771, which all introduced the oxygenation activity of CYP2F3, and at the same time the dehydrogenation activity of S4771 mutant was abolished completely (unpublished data). Table 2.1 Kinetic Parameters for the Formation of 3-Methylindole Metabolites by CYP2F3 and its Mutants * CYP2F3 D362T S476H 54771 3-MI Dehydrogenation forming3-MINAC Km{IlM) 73 32 55 - Vmax (min-1) 3.80 6.00 5.50 - V/k 0.05 0.20 0.10 NO 3-MI Oxygenation forming 3-Methyloxindol e KmhJ.M} - 120 102 110 Vmax(min-1} - 3.50 1.00 5.00 V/k ND 0.03 0.01 0.05 * Data listed in this table were determined by Dr. Jaya Kartha in her PhD. dissertation at the University of Utah, Summer 2007 66 In this study, we describe an in silico. method to construct a three-dimensional homology model of CYP2F3 and the mutants that were produced and evaluated, using other P450 structures determined by X-ray crystallography, and followed by docking the substrate 3-methylindole to the active site of CYP2F3. This study was both a means of testing the results from site-directed mutagenesis experiments and a way to identify new important residues within the active site of CYP2F3 that could also control the dehydrogenation of 3-methylindole. Materials and Methods Building CYP2F3 Model The available three-dimensional structures of P450 enzymes are CYP3A4 (Yano et aI., 2004), CYP1A2 (Sansen et aI., 2007), CYP2A6 (Yano et aI., 2005), CYP2D6 (Rowland et aI., 2006) and CYP2C8/9 (Schoch et aI., 2004; Wester et aI., 2004) which all were determined by X-ray crystallography. Considering the sequence similarities have important effects on the quality of the model, only CYP2 subfamily structures were chosen as the templates to build CYP2F3 structure using a homology modeling method. The templates used in this study include a single CYP2C9 template, a single CYP2A6 template and a multiple template combination of CYP2C8, CYP2C9 and CYP2A6. CYP2F3 model was generated using Modeller8v2 (Fiser and Sali, 2003). In general an alignment of 2F3 sequence (Wang et aI., 1998) with above three template selections was performed by Modeller using either MALIGN2D or MALIGN3D functions. These functions are based on dynamic programming algorithms, and they are different from standard sequence-sequence alignment methods since they also take into 67 account structural infonnation from the template. The alignment of CYP2F3 sequence with these templates was achieved through a variable gap penalty function, as discussed above. Inspection of the sequence alignments of CYP2F3 with these single or multiple templates did not indicate that one was superior. The alignment results were then applied as the parameter files in Modeller to generate CYP2F3 models. The structure minimization was perfonned with a combination of both topological and CHARMM potential energetic constraint fUnctions. Since there is no constraint parameters included in CHARMM on the distance between cysteine sulfur atom and heme iron, a patch was written to define the distance between them. Additional constraint patches were accordingly written to fix residues within 6 A of this proximal cysteine. CYP2F3 models were therefore constructed using an energetic/topological optimization algorithm in Modeller. A total of 20 CYP2F3 models for each template were built and the one with the lowest energy was selected for further docking analysis. The stereochemical quality of the model was evaluated by Modeller and Pro Check (Laskowski et aI., 1996). In addition, the volume of the active site was measured by VOIDOO (Kleywegt and Jones, 1994). The mutants of CYP2F3 were constructed using these same methods. Docking 3-Methylindole Autodock program version 3.05 (Goodsell et aI., 1996) was acquired from Scripps Research Institute (La Jolla, CA) and the docking experiments were run on a Dell Precision 690 workstation with two 64 .. bit Dual-Core Intel Xeon processors and the Red Hat Enterprise Linux WS4 operating system. The three-dimensional structure of 3- 68 methylindole was constructed using Chem3D Intra 1 0 (CambridgeSoft Corporation, Cambridge, MA) and its energy was minimized with a molecular mechanics method. The AutoDockTools program (Scripps Research Institute, La Jolla, CA) was used to prepare all docking parameter files for 3-methylindole as well as for CYP2F3 and its mutants. Oasteiger atomic charges were assigned and flexible torsions were defined for 3-methylindole (actually there are no flexible bonds in the structure of 3-methylindole). In addition, polar hydrogens, Kollman partial charges, and solvation parameters were added to the CYP2F3 template and its mutants. A grid box with sufficient space to cover the whole active site of. CYP2F3 was defined using AutoOrid 3.06 (Scripps Research Institute, La Jolla, CA), in which Autodock searched the optimal 3-methylindole conformations and orientations using the Lamarckian genetic algorithm (LOA), which is a hybrid of genetic algorithm (GA) with an adaptive local search (LC) method. Five hundred LOA were executed for 3-methylindole in each template. All docking confirmations were clustered with a RMSD of 0.5, and both energetically and specially favored clusters were selected for analysis. Results and Discussion CYP2F3 Model with CYP2A6 Template From our previous studies (Wang et aI., 19