<%BANNER%>

Directed Evolution of a Sulfoxidation Biocatalyst


PAGE 1

DIRECTED EVOLUTION OF A SULFOXIDATION BIOCATALYST By ARIS A. POLYZOS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2003

PAGE 2

Copyright 2003 by Aris A. Polyzos

PAGE 3

This work is dedicated to the strong women in my life. Without you I would not be half of what I am now. My (soon to be) wife April Flanders, my mother Penelope Polyzou, my sister Cassandra Polyzou and all the little furry ones.

PAGE 4

ACKNOWLEDGMENTS I would like to give my gratitude to everyone who helped and encouraged me through these years and made the completion of this work possible. During my career as a graduate student in the Chemistry Department I was blessed with many mentors, colleagues and wonderful friends. I would like to thank Dr. J. Stewart for the opportunity to work in his group and for his advice and guidance with my doctoral research. The genes and plasmids graciously provided throughout the course of this work by Dr. Philpot, Dr. Williams and Dr. Cashman have also been invaluable as they have been the main subjects of this study. In providing help both by guidance and physically providing us with many of the supplies for this project I must make a special note of thanks to my friend Dr. Michael Thomson, one of the most capable molecular biologists I have come across, who is now in Cold Springs Harbor Laboratories. Among the small biochemistry group in the Chemistry Department of UF there have been several whom I am very glad to have met. Dr. B. Horenstein and Dr. Tom (Lyons) have provided insights at times of scientific consternation and have also been guides through the non-scientific problems one can find in a research project. Dr. R. Geyer, now a faculty member at the University of Saskatchewan, also helped with advice and expertise in questions of molecular biology and in reminding me of home. As a member of the Stewart group I have made many wonderful friendships with people still close to me, even if they are now dispersed around the globe. Sonia Rodriguez is an individual I will always respect, having been both a labmate and a guide in maintaining iv

PAGE 5

my beliefs and upstanding moral judgment and in exemplifying thorough scientific inquiry. Carlos Martinez, who I am overjoyed to say is very happy now, has always been the source of information on matters of science and an example of quiet and undaunted work no matter what the situation. For 4 years Jennifer Tonzello, who is also happily living out West now, was a big part of my life along with Sonia and Carlos. She was most evidently the one who brought heart and cheer into the laboratory when it was most lacking. Marko Mihovilovic, who should expect my visit sometime soon to his newly built home in Austria, was the impressive postdoc, having both a very thorough appreciation for scientific work balanced with enough of a zest for life to be always cheerful. Recently Despina Bougioukou, a compatriot of genes and of spirit, has been here beside me in the laboratory. I am completely confident that her scientific undertakings will be thorough and well thought-out and expect she will do very well in her research. I will also always be in touch and seriously invite her to the Great White North to visit us. I was fortunate to spend my time in the laboratory with many other people who made the research setting the eclectic mixture that it was. Kavitha Vedha-Peters, Erin Ringus, Mirella Stefans, Lisa Manning, Abhijit Roy-Chundry, Harch Ooi, Bart Neff, Kersten Schroeder, Matt Carrigan, Lee Raley, Stefan Lutz, Jun Zhu, Darwin Ang, Kim Millar, Debbie Burroughs, Catherine Charron, Tricia Pokey, Michael Sismour, Alonso Ricardo, Leslie Tuchman, Iwona Kaluzna, Brian Kyte, Brent Feske and many others have all played their part in this wonderful experience. Throughout my studies from almost the first year at UF I am thankful to have crossed paths with Benjamin Lopman, initially as the intelligent undergraduate in a chemistry lab I was overseeing, to later becoming my closest friend, fellow athlete and most recently as a fellow graduate student. He and his wife, Leah Garces, have been as close as family since we v

PAGE 6

first met and our relationship has fed off each others zeal and curiosity for all the wonders that this world has to offer. They have been close throughout my studies even from their nest way across the pond. My support comes from up north too and is also unwavering. I owe so much to my mother Penelope and sister Cassandra Polyzou as to not be able to fit my thanks into an entire novel much less in a few pages at the beginning of a thesis. They are as much a part of me as my own thoughts and I am glad to be moving closer to them soon. Vincent Vadez and Rachid Serraj have been the two French academics who breathed the joy of life into me in my early years here and their thoughtfulness and compassion are always remembered. I wish them luck in India and in the jungles of Bolivia where they are now using their academic knowledge for practical purposes. I would also thank the University of Florida cycling team. Winning the SEC Conference Championships and placing within the top teams nationally on whatever type of bike one could imagine these last two years are not the reasons. The reasons are the very good friends, the reaching of new physical limits almost on a weekly basis and the incredible scenery that I have shared with some very hard working people. The cycling team has also been the group that made possible my meeting of April Flanders, who is the main factor for me being here in one conscious piece. She is my partner and guide. She is the cycling art professor that stole me away and as we speak now I cannot imagine what it would have been like had we never met. It could never get this good. vi

PAGE 7

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...............................................................................................................x LIST OF FIGURES...........................................................................................................xi ABSTRACT.....................................................................................................................xvi CHAPTER 1 INTRODUCTION AND BACKGROUND.................................................................1 Enantiopure Sulfoxides in Synthesis.....................................................................2 Approaches Towards Sulfoxides...........................................................................9 Routes to Racemic Sulfoxides, Classical Organic Chemistry........................9 Routes to Non-racemic Sulfoxides, Classical Organic Chemistry...............10 Enzymatic Routes to Chiral Sulfoxides...............................................................15 Directed Evolution of Enzymes...........................................................................20 Theory and Approaches...............................................................................20 High Throughput Screening for Enantioselectivity.....................................23 Flavin-Monooxygenases.............................................................................................29 Background and Properties..................................................................................30 Catalysis by FMOs.............................................................................................34 Structure of FMOs..............................................................................................37 Strategy Towards A Stereoselective Sulfoxidation Catalyst, An Overview of this Work......................................................................................................................42 2 ENGINEERING OF MICROORGANISMS FOR FMO PRODUCTION................44 Expression Using Escherichia coli.............................................................................46 Expression of Active FMO Enzymes..................................................................47 Rabbit FMO1 and Rabbit FMO3.................................................................49 Rabbit FMO2................................................................................................50 Rhesus macaque FMO2...............................................................................51 Improving the Expression of Active FMO..........................................................53 3 GENERATING LIBRARIES OF MUTANT FMOS...............................................58 Error-Prone PCR; Controlled Random Mutagenesis..................................................58 vii

PAGE 8

DNA Shuffling and Recombination Mutagenesis......................................................66 4 BIOCATALYSIS REACTIONS AND SCREENING...............................................71 Sulfoxidation Using FMO Biocatalysts......................................................................72 Screening Biocatalysis Reactions...............................................................................77 Screening Biotransformations Using UV/VIS and OR Detectors.......................78 Developing Methodology.............................................................................78 Assessment of UV/ORD Screening for Enantioselectivity..........................84 Screening Biotransformations Using Chiral HPLC............................................85 Developing Methodology.............................................................................85 Nonenzymatic Racemic Sulfoxidation.........................................................88 5 DIRECTED EVOLUTION OF FMOS TOWARDS EFFICIENT (S)-SULFOXIDATION CATALYSTS............................................................................93 Results of Screening Assays.......................................................................................93 Conclusions and Future Work..................................................................................101 APPENDIX A EXPERIMENTAL PROCEDURES.........................................................................103 Reagents and Supplies..............................................................................................103 Purification of Pfu Polymerase.................................................................................104 Error-Prone PCR of FMO.........................................................................................105 DNA Shuffing...........................................................................................................106 Fragmentation of Genes....................................................................................106 Reassembly of DNA Fragments........................................................................107 Amplification of the Shuffled Genes.................................................................107 Assay for FMO Activity; Oxidation of Thiobenzamide...........................................108 Crude Membrane Preparation Activity Assay; Oxidation of Methimazole and MTS.....................................................................................................................109 Synthesis of Starting material and Sulfoxide Standards...........................................110 Aryl Alkyl Sulfides............................................................................................110 Racemic Methyl p-tolyl Sulfoxide (MTSO)......................................................111 Alkyl p-tolyl Sulfoxides....................................................................................111 PCR Primers.............................................................................................................112 Strains.......................................................................................................................113 Recombinant DNA Techniques................................................................................114 HPLC Analysis.........................................................................................................115 viii

PAGE 9

B EXPERIMENTAL DATA AND MISCELANEOUS..............................................122 C GLOSSARY OF TERMS AND ABBREVIATIONS..............................................130 LIST OF REFERENCES.................................................................................................131 BIOGRAPHICAL SKETCH...........................................................................................144 ix

PAGE 10

LIST OF TABLES Table page 1. Classical organic chemistry sythesis of enantiopure MTSO.........................................15 2. Methods for the stereoselective sulfoxidation of methyl p-tolyl sulfide (MTS). Various classical synthetic organic methods as well as isolated enzymes and whole cell systems could be used to produce MTSO................................................................17 3. List of identified and putative FMOs to date................................................................32 4. Various expression systems for rabbit and rhesus macaque FMOs in Escherichia coli and Saccharomyces cerevisiae.................................................................................48 5. Catalytic activity of Escherichia coli expressed rabbit FMO1 fused with various peptides at either the Nor Cterminal ends...........................................................55 6. Mutagenic frequency of EP-PCR reactions with various concentrations of Mn 2+ EP-PCR libraries (#C, D, E, F, G) generated from wild type rabbit FMO1..................62 7. Mutagenesis and sulfoxidation activities of EP-rFMO1-J library.................................64 8. Mutagenic frequency of EP-PCR reactions with various concentrations of Mn 2+ EP-PCR libraries (#L, M, N, O, P) generated from wild type rabbit FMO1.................64 9. Sulfoxidations of a variety of aryl-alkyl sulfides by purified FMOs (ip) and crude-membrane preparations of heterologously expressed FMOs (cmp).......................73 10. Sulfoxidation of an array of alkyl aryl sulfides catalyzed by crude membrane preparations of Escherichia coli expressing rabbit FMO1.......................................75 11. Various strategies to remove background racemic sulfoxidation of MTS..................90 12. The cost of screening a library of singly mutated FMO (10 000clones) for sulfoxidation activity (enantioselectivities and yields)............................................91 13. The strains of Escherichia coli used in this study and their genotypes.....................113 14. The strains of Saccharomyces cerevisiae used in this study and their genotypes.....114 x

PAGE 11

LIST OF FIGURES Figure page 1. Chirality of unsymmetrically substituted sulfoxides.......................................................2 2. The active (S) enantiomer of omeprazole........................................................................2 3. Chiral sulfoxide mediated route from esters to enantiomerically pure alcohols6...........3 4. Methyl methyl sulfoxide used in the sythesis of ketones................................................4 5. The retrosynthesis of mevinic acid-type hypocholestemic agents using enantiopure MTSO as a chiral auxiliary........................................................................................5 6. Total sythesis of the rubiginone antibiotic incorporating the chiral auxiliary (R)-MTSO.........................................................................................................................5 7. The retrosynthetic scheme towards Panamycin 607 which incorporates an asymmetric reduction of chiral -ketosulfoxide derived from (S)-MTSO....................................6 8. The retrosynthesis of (-)-Centrolobine, indicating the chiral sulfoxide mediated key diastereoselective ketone reduction............................................................................7 9. Stereoselective total synthesis of an enantiomerically pure fluorinated analogue of tetrahydroisoquinoline................................................................................................7 10. Retrosynthetic scheme for the synthesis of 2-fluoro analogues of frontalin indicating the stereodirecting effect of the chiral sulfoxide auxiliary in the formation of the epoxide 6....................................................................................................................8 11. A facile synthesis of D-erythrose employing (R)-MTSO for stereoselective alkylation of an aldehyde............................................................................................................8 12. The synthesis of enantiopure (S) benzyl alcohol using chiral (S) Methyl 1-naphthyl sulfoxide.....................................................................................................................9 13. The most prevalent current synthetic routes towards chiral sulfoxides.......................10 14. A modification of the Kagan-Sharpless oxidation of sulfides using (R)-(+)-binaphthol in place of DET........................................................................................................12 15. Chiral oxazidines as sulfoxidation catalysts................................................................12 xi

PAGE 12

16. The Andersen synthesis of chiral p-tolyl sulfoxides. The route towards the antipodal sulfoxide is identical but requires different separation techniques to afford the (R)-menthyl ester............................................................................................................13 17. Kinetic resolution in the Andersen synthesis, recrystallization in acetone-HCl..........13 18. Several chiral sulfinyl transfer reagents for the enantioselective synthesis of sulfoxides.................................................................................................................15 19.Sources of large libraries from which enantioselective catalysts can be explored.......20 20. Enzymatic reactions can produce highly chromophoric and fluorogenic molecules that can be used to screen for activity......................................................................24 21. An enantioselective mutant PAL catalyses the kinetic resolution of a racemic p-nitrophenol ester.......................................................................................................25 22. The reaction catalyzed by hydantoinase, which has distinct enantiomeric substrates (the inherent racemization between the 5-monosubstituted hydantoin isomers is slow).........................................................................................................................26 23. The enantioselective reduction of a coumarinyl derivative of 2-butyl ketone can be monitored using high-throughput fluorescence screening for the reverse oxidation reaction coupled with elimination to form a fluorophore product...........................27 24. Screening for enantioselectivity using competitive enzyme immunoassays...............28 25. The catalytic cycle postulated for pig liver FMOs.....................................................36 26. Model of the active site of rabbit and pig FMO1 and rabbit FMO2 proposed by Cashman and Rettie showing the orientation substrates (propyl 2-naphthyl sulfide, methyl 2-naphthyl sulfide, MTS, heptyl p-tolyl sulfide) would adopt in the active site............................................................................................................................38 27. The topography of the ~535 amino acid protein of rabbit FMO1, 2, 3.......................39 28. Differences in the enantioselectivity of sulfoxidation by purified FMOs from different sources with increasing alkyl chain length of alkyl p-tolyl sulfides.........40 29. An overview of the strategy for the directed evolution of an FMO enzyme as a catalyst for the efficient production of (S) MTSO, a molecule with many uses in organic syntheses......................................................................................................43 30. General schematic of an Escherichia coli expression plasmid vector.........................49 31. Expression of rabbit FMO1, expected at ~ 59kDa, from BL21(DE3)/pAAP11 analyzed with SDS-PAGE.......................................................................................50 32. Electron-micrograph evidence of lack of inclusion bodies..........................................52 xii

PAGE 13

33. Growth rates of recombinant Escherichia coli cells....................................................56 34. Error-prone PCR and cloning into expression systems. Indicated are the possible sources of wild-type gene which can carry over into final mutagenized library.....60 35. Activities of rabbit FMO1 clones (EP-rFMO1-J)........................................................63 37. Error-prone PCR and cloning into expression system. Strategy to remove wild-type contamination from the final mixtures. nm = non-methylated DNA.......................65 38. Multiple sequence alignment of Rabbit FMO1 and Rhesus macaque FMO2 genes (DNA) (ClustalW)....................................................................................................67 39. Multiple sequence alignment of Rabbit FMO1 and Rhesus macaque FMO2 proteins (ClustalW)................................................................................................................68 40. General overview of DNA Shuffling...........................................................................69 42. Biotransformation of MTS by FMO presented as crude membrane preparations from Escherichia coli based expression systems..............................................................76 43. Chiral amide, forms crystals with molecules of (R) MTSO (1 molecule amide: 1 molecule sulfoxide)..................................................................................................78 44. The addition of two auxochromic substituents para to each other and charge transfer resonance forms influence the absorption characteristic of phenyl rings................79 45. Comparative UV/VIS spectra of biotransformation reactions on various substrates..79 46. UV spectrum of MTS and MTSO, indicating the distinct absorption maxima; MTSO (233nm) and MTS (254nm).....................................................................................80 47. Cell extract background in the biotransformations......................................................81 48. A comparison of the UV and ORD signals of MTS and MTSO in various solvents..83 49. Overview of the method to screen, using UV and OR measurements for the enantioselectivity of biocatalysis reactions..............................................................84 50. Values of % ee. for samples measured using the UV/ORD screen.............................85 51. Chiral HPLC chromatographs showing separation of a series of racemic alkyl p-tolyl sulfoxides.................................................................................................................86 52. An overview of the use of HPLC adapted to determine reaction yields of biocatalysis reactions...................................................................................................................87 53. The chromatograms of MTS sulfoxidation reactions (0.6mole scale ) of lysed rabbit FMO1 expressing cells.............................................................................................88 xiii

PAGE 14

54. Background racemic sulfoxidation of MTS.................................................................89 55. Values of % ee. for samples measured using a chiral HPLC column..........................91 56. Determination of enantiomeric composition of reactions using a chiral HPLC column.92 57. Sequencing the mutant rabbit FMO1 genes from the EP-PCR generated library involves 3 primers flanking and internal to the gene itself......................................94 58. The activity of C) clone 176E70, compared to B) the wild type activity (rabbit FMO1) and the A) background air oxidation...........................................................95 59. Rescreening of 13 clones of rabbit FMO1 for stereoselectivities................................95 60. Extrapolated times required for each enzyme mutant to oxidize 2mM of MTS..........96 61. The activities of 4 clones with higher activities than the wild type.............................97 62. Sulfoxidation stereochemistries and yields catalyzed by ~4000 various singly mutated clones of rabbit FMO1 (library EP-PCR-FMO1-S+T)............................................99 63. Sulfoxidation stereochemistries and yields catalyzed by ~1100 various chimeras of rabbit FMO1 and Rhesus macaque FMO2 (library Sh-FMO1/2-C)......................100 64. Postulated tertiary ribbon structure of a dimer of human FMO3...............................102 65. Pfu polymerase (prepared) vs. commercial (Fisher Scientific) Taq polymerase amplification...........................................................................................................105 66. Fragmentation of the FMO gene with DNase I run on a 2.5% agarose gel...............107 67. Reassembled and amplified FMO gene chimeras loaded on a 0.8% agarose gel......108 68. Construction of pAAP20 (plasmid for expression of rabbit FMO2 under the control of P trc ) and pAAP15 (plasmid for expression of rabbit FMO2 under the control of P T7 )116 69. The construction of pAAP18 (plasmid for expression of rabbit FMO4 under the control of P T7 ) and pAAP14 (plasmid for expressio of rabbit FMO3 under the control of P T7 ).........................................................................................................117 70. The construction of pAAP12 (plasmid for the expression of rabbit FMO2 under the control of P tac ), pAAP10 (plasmid for expression of rabbit FMO2 in yeast) and pAAP21 (plasmid for the expression of rabbit FMO2 under the control of P T7 with kanamycin selection)..............................................................................................118 71. Construction of pAAP11 (plasmid for the expression of rabbit FMO1 under the control of P T7 with ampicillin selection), pAAP27 (plasmid for the expression of rabbit FMO1 under the control of P T7 with kanamycin selection) and pAAP38 (template for the error prone PCR reactions to generate clones of rabbit FMO1).119 xiv

PAGE 15

72. Construction of pAAP35 (plasmid for the expression of rhesus macaque FMO2 under the control of P T7 with ampicillin selection), pAAP36 (plasmid for the expression of rhesus macaque FMO2 under the control of P T7 with kanamycin selection) and pAAP38 (template for the error prone PCR reactions to generate clones of rhesus macaque FMO2)....................................................................................................120 73. Construction of pAAP46 (plasmid for the expression of rabbit FMO1 with a C-terminal His 6 tag under the control of P T7 ), pAAP41 (plasmid for the expression of rabbit FMO1 with a Cand N-terminal His 6 tag under the control of P T7 ) and pAAP29 (plasmid for the expression of rabbit FMO1 fused to MalE at the N-terminus, under the control of P tac )........................................................................121 74. Standardization measurement of UV 233nm for MTS and MTSO indicating the error apparent in each reading. Samples were in solutions of cell extract and the background was corrected from the relative absorbance at 280nm. The upper graph indicates the non-linearity of measurements of concentrations >0.1mM...............123 75. Standardization measurement of UV254nm for MTS and MTSO indicating the error apparent in each reading.........................................................................................123 76. Standardization measurement of OR for MTS and MTSO indicating the error apparent in each reading.........................................................................................124 78. The sequence of clone 170AA21 of rabbit FMO1. Upper sequence is DNA, beneath it is the predicted amino acid sequence.....................................................................125 79. The sequence of clone 176C34 of rabbit FMO1. Upper sequence is DNA, beneath it is the predicted amino acid sequence.....................................................................126 80. The sequence of clone 170D96 of rabbit FMO1. Upper sequence is DNA, beneath it is the predicted amino acid sequence.....................................................................127 81. The sequence of clone 170D62 of rabbit FMO1. Upper sequence is DNA, beneath it is the predicted amino acid sequence.....................................................................128 82. The sequence of clone 176E70 of rabbit FMO1. Upper sequence is DNA, beneath it is the predicted amino acid sequence.........................................................................129 xv

PAGE 16

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DIRECTED EVOLUTION OF A SULFOXIDATION BIOCATALYST By Aris A. Polyzos May 2003 Chair: J. Stewart Major Department: Chemistry The chirality of sulfoxides has been exploited in many organic syntheses. The persistent use of non-racemic sulfoxides directly or as chiral relay agents in organic synthesis has spurred the development of improved routes to each sulfoxide enantiomer. The long term goal of this project is to develop a biocatalytic system for quick and selective production of both stereoisomers of a chiral sulfoxide (methyl p-tolyl sulfoxide). The directed evolution of the enzyme flavin-containing monooxygenase was used to develop this biocatalytic system. FMO enzymes from rabbit (FMO1) and rhesus macaque (FMO2) have opposite stereochemical preference for the sulfoxidation of methyl p-tolyl sulfide. They have been cloned into an Escherichia coli based expression and biocatalytic system. Libraries of mutants of these FMOs have been developed for screening in order to identify an enzyme that produces high yields of the (S) sulfoxide. Genes with single amino acid mutations were generated from rabbit FMO1, which was also shuffled with the monkey enzyme to generate a library of chimeric genes. xvi

PAGE 17

These libraries were screened for enantioselectivity in the sulfoxidation of methyl p-tolyl sulfide. A screening method developed using a flow cell optical rotation and a UV detector proved slow (<100 clones/day) and had low sensitivity for identifying differences in enantiomeric compositions of reactions. Instead, using chiral HPLC the activities of 6000 clones were tested (6min/clone) for the sulfoxidation of methyl p-tolyl sulfide. From both the mutant libraries one clone (176E70) displayed an increase in selectivity for the (S) sulfoxide of 30% ee. and there were 4 clones which displayed up to 35% increased activity. Of the two directed evolution strategies the single mutation library of FMO1 yielded interesting clones while none were observed from the chimeric enzymes. xvii

PAGE 18

CHAPTER 1 INTRODUCTION AND BACKGROUND Chirality in chemical compounds has been known since the 1840s when Louis Pasteur observed that he could separate the crystals of the different isomeric forms of tartaric acid (which combined were known as racemic acid). The importance of organic chiral substances and their interactions with biological materials has become ever more significant. In 1999 the annual sales of chiral drug molecules topped $100 billion for the first time.1 There are several reasons chirality has affected the drug industry so greatly. In large part it is because biological messenger molecules and cell surface receptors are chiral and the drug molecules that medicinal chemists design and the pharmaceutical industry produces must have corresponding asymmetry. Another reason is simple cost effectiveness. The Food and Drug Administration in the U.S. requires that each enantiomer of a drug being sold be tested in detail separately1 which doubles the cost of screening racemic mixtures. There are also many examples of chiral drug molecules in which one enantiomer is pharmacologically active and others are inactive or even harmful.2 The (R)-(+) form of thalidomide can be used by pregnant women to treat morning sickness, whereas the (S)-(-) enantiomer is associated with startling numbers of baby deaths and birth defects as well as peripheral neuritis. Another example is ibuprofen, whose (S)-enantiomer is a common anti-inflammatory drug, whereas (R)-ibuprofen is inactive. 1

PAGE 19

2 Unsymmetrically Substituted Sulfoxides; Chiral Molecules There are steric and stereoelectronic differences between 4 substituents of an unsymmetrically substituted sulfoxide, a highly configurationally stable moiety (Figure 1).3 The lone electron pair, the oxygen and the two distinct carbon groups provide a chiral environment rich in its chemistry. Figure 1. Chirality of unsymmetrically substituted sulfoxides. Enantiopure Sulfoxides in Synthesis The best selling drug of all time, Astra-Zenecas anti-ulcer omeprazole (Prilosec/Losec), is itself a chiral sulfoxide.4 Initially marketed as a racemate, it is the (S) enantiomer (Figure 2) (with its own generic name of esomeprazole) that has pharmacoactivity and displays the least side-effects.1 NHN S+ O : H3CO N CH3 CH3 OCH3 esomeprazole Figure 2. The active (S) enantiomer of omeprazole. There are several examples of synthetic chiral sulfoxide targets5 although the interest in the chirality of sulfoxides to synthetic and medicinal chemists extends even further than the molecules themselves, into their use as chiral relay agents in syntheses.6

PAGE 20

3 The electron deficient character of the sulfoxide sulfur can be used to stabilize adjacent carbanions and many applications of sulfoxides in synthesis involve the reaction of sulfur-stabilized carbanions with electrophiles.7,8 Solladie and Carreno have thoroughly extended the usefulness of sulfoxides into asymmetric synthesis.9 -ketosulfoxides, which can be readily prepared from esters coupling to a chiral sulfoxide -carbanion, can be reduced to either diastereomer selectively by using DIBAL (diisobutylaluminum) or ZnCl 2 /DIBAL. Desulfurization with Raney nickel affords corresponding enantiomerically pure carbinols (Figure 3).6 R OCH3 O S+ Ar : O R O S+ Ar : O R S+ Ar : O OH R OH R S+ Ar : O OH R OH S(S) ZnCl2DIBAL (R) S(S) (S) Raney nickel DIBAL (S) S(S) (R) Raney nickel Figure 3. Chiral sulfoxide mediated route from esters to enantiomerically pure alcohols6. As mentioned above an unsymmetrically sulfoxide can act as a chiral reagent in total syntheses. In the past decade (1993-2003) chiral sulfoxides such as methyl p-tolyl sulfoxide (MTSO) have served as chiral auxiliriaries in several total sytheses of natural products and pharmacochemicals. Deprotonation to a sulfoxide can be accomplished with bases such as NaH (in DMSO) and produces a strong base and a reagent able to generate ylides (such as Wittig reagents) and which reacts with esters to produce enolates of -ketosulfoxides.

PAGE 21

4 Subsequent desulfurization of -keto sulfinyl compoundswith aluminum amalgam provides a synthesis of ketones (Figure 4). SCH3 CH3 O SCH3 CH2 O [Na]+ R CH3O O CH3S R O O [Na]+ R1X CH3S O R O R1 R O R1 HNa, DMSO Al(Hg), H2O Figure 4. Methyl methyl sulfoxide used in the sythesis of ketones. Non-racemic MTSO produced by biocatalysis using Saccharomyces cerevisiae (yeast) was used as a precursor in the total synthesis of mevinic acids (Figure 5).10 Considerable interest in these compounds arises from their ability to lower the level of cholesterol in blood plasma by inhibiting enzymes involved in the biosynthesis of cholesterol. In this approach the chiral sulfoxide directs the chirality of the -hydroxy--lactone moiety, the key structural feature in all HMG-CoA inhibitors.11 The chiral center at C-3 of the lactone ring is determined during an aldol condensation of the -cabanion of the sulfoxides with an aldehyde. The -sulfinyl ester can be later removed by desulfurization using aluminum amalgam.12

PAGE 22

5 R O H CH3 O CH3 O OH O O OH O OH OH OH O OH S+ O : OH OH O S+ O : S+ O : S :: R-MTSO R = H (+) compactinR = CH3 (+) mevinolinR = OH (+) provastatin Saccharomyces cerevisae 60% yield, 92% ee (R) COMPACTIN, MEVINOLIN, PROVASTATIN Figure 5. The retrosynthesis of mevinic acid-type hypocholestemic agents using enantiopure MTSO as a chiral auxiliary.10 The angucycline antibiotics such as the various rubiginone derivatives, also called fujianmycins, are claimed as useful in the treatment of AIDS and Alzheimers13 and exhibit potentiation of vincristine-induced cytotoxicity against multidrug-resistant tumour cells.14 Carreno et al.15 utilizes a chiral sulfoxide moiety twice in the total synthesis of these molecules (Figure 6). (R)-MTSO was incorporated into 1 and directed the highly chemoand diastereoselective conjugate addition of Al(CH 3 ) 3 to form 2. O O OCH3 OR CH3 OH O O OCH3 S+ : O pTol TBDMSO CH3 OCOiPr O CH3 OH S(O)pTol S+pTol O : O OH R = H, COiPr + Al(CH3)3, CH2Cl2 -78oC, 44hrs 65%, 98% d.e. RUBIGINONE 1 2 Figure 6. Total sythesis of the rubiginone antibiotic15 incorporating the chiral auxiliary (R)-MTSO.

PAGE 23

6 The diastereoselective reduction of -keto sulfoxide with DIBAL-H was incorporated by Solladie et al.16 into the total synthesis of panamycin (Figure 7) and in the enantioselective synthesis of centrolobine (Figure 8). The crystalline natural product centrolobine, isolated from the stem of the amazonian plant Brosinum potabile in 1964, required enantioselective synthesis to establish unequivocally its absolute configuration. The approach by Colobert et al.17 was the first strategy to achieve this synthesis. Condensation of (R)-MTSO with glutaric anhydride affords the -keto sulfoxide 3, which can in turn be reduced stereoselectively to the diastereomerically pure hydroxyketone 4, again directed by the incorporated sulfoxide. O O O CH3N O O O O OO S+ pTol O : O S+ pTol O : OH O O tBuO O OH PANAMYCIN 607 1) DIBAL-H2) oxalic acid, THF, H2O 80%, >95% d.e. Figure 7The retrosynthetic scheme towards Panamycin 60716 which incorporates an asymmetric reduction of chiral -ketosulfoxide derived from (S)-MTSO.

PAGE 24

7 O CH3O OH S+ OH O : pTol CH3O O S+ O : pTol CH3O O O DIBAL/ZnBr2 THF80%, 98% d.e. CENTROLOBINE 3 4 Figure 8. The retrosynthesis of (-)-Centrolobine17, indicating the chiral sulfoxide mediated key diastereoselective ketone reduction. Chiral sulfoxide auxiliaries can also exert stereodirecting effects in Pictet-Spengler reactions.18 Bravo and co-workers19 used (S) MTSO as an auxiliary to direct the ring closure of 5. The attack of the 3,4-dimethoxyphenyl group was directed to the less hindered Re face of the stabilized carbocation formed by imine protonation of 5 with TFA. An example of this stereocontrol was employed in the total synthesis of a novel non-racemic 1-trifluoromethyl analogues of tetrahydroisoquinoline (Figure 9).19 N CH3O CH3O CH3 CH3OH F3C N CH3O CH3O H F3C S+ pTol O : N CH3O CH3O F3C S+ pTol : O TFA, 0oC, CHCl3 74%, 71% e.e. 1-TRIFLUOROMETHYL TETRAHYDROISOQUINOLINE ALKALOID 5 Figure 9. Stereoselective total synthesis of an enantiomerically pure fluorinated analogue of tetrahydroisoquinoline.19

PAGE 25

8 In the synthesis of a natural product analogue Bravo and co-workers again applied the chirality of the (R)-MTSO but this time to direct methylene insertion from diazomethane into the carbonyl group of a -keto sulfoxide 7.20 Modest diastereoselectivities were improved by separation using flash chromatography to obtain the intermediate in the synthesis of 2-fluoro analogues of frontalin (Figure 10), an analogue of the bioactive component of the aggregation pheromone of pine beetles. O O CH3 CH3 F S+ H F O pTol : O S+ H F pTol : O O 95%, 85% d.e. CH2N2, CH3OH, 0oC 2-FLUORO ANALOGUE OF FRONTALIN 6 7 Figure 10. Retrosynthetic scheme for the synthesis of 2-fluoro analogues of frontalin20 indicating the stereodirecting effect of the chiral sulfoxide auxiliary in the formation of the epoxide 6. Arroyo-Gomez and coworkers used (R)-MTSO in the diastereoselective alkylation of aldehyde 8.21 Due to the matched asymmetric induction from chirality in both the aldehyde and the nucleophile such a reaction proceeded with complete diastereoselectivity affording a precursor for a facile synthesis of D-erythrose (Figure 11). O OH OBn H OH OO S+ pTol OH : O OO O H S+pTol : O CH3 2-O-Benzyl-()-D-erythrofuranose LDA + 70%, 99% d.e. D-ERYTHROSE DERIVATIVE 8 Figure 11. A facile synthesis of D-erythrose employing (R)-MTSO21 for stereoselective alkylation of an aldehyde. Addition of the carbanion of (S) Methyl 1-naphthyl sulfoxide to n-alkyl phenyl ketones occurs with remarkably high diastereoselective excess. Raney nickel is used to remove the chiral sulfoxide auxiliary and produce enantiopure (S) benzyl alcohols

PAGE 26

9 (Figure 12). The initial chiral sulfoxide is prepared by sulfoxidation of the parent aryl alkyl sulfide with -cyclodextrin and peracetic acid followed by crystallization.22 S+ CH3 O : R O S+ O : R OH OH CH3 R 1) LDA 2) 96-98% (S,S) hydroxysulfoxide (100% d.e.) Raney Ni (S) alcohol 100% e.e. R = n-alkyl Figure 12. The synthesis of enantiopure (S) benzyl alcohol using chiral (S) Methyl 1-naphthyl sulfoxide22. Approaches Towards Sulfoxides Routes to Racemic Sulfoxides, Classical Organic Chemistry The oxidation of sulfur bonded to carbon has a long precedence and can be done by many reagents. Many general oxidants perform sulfoxidation, these include nitric acid 23, H 2 O 2 24, dinitrogen tetroxide 25, chromic acid26, ozone27, manganese dioxide26, selenium dioxide28, NaIO 3 29 and oxone (or 2KHSO 5 KHSO 4 K 2 SO 4 ) which has provided sulfoxidations of some cyclic sulfides30. Modest stereoselectivities can be achieved with NaIO 4 -catalyzed S-oxygenation in the presence of BSA31. NaIO 4 usually used in methanol/water affords yields of 90%29 and can also be supported on silica32 or alumina. Control of oxidation to avoid overoxidation to the corresponding sulfone and sulfonic acid33 and control of the orientation of the oxygen atom added are often the

PAGE 27

10 most sensitive criteria for developing sulfoxidation chemistry for current synthetic purposes. Routes to Non-racemic Sulfoxides, Classical Organic Chemistry Several techniques are currently available to obtain enatiomerically pure sulfoxides. These include the asymmetric oxidation of thioethers, asymmetric synthesis by nucleophilic substitution on chiral sulfur derivatives, kinetic resolution and optical resolution (Figure 13). The selective complexation and crystallization of the enantiomers of ethyl p-tolyl sulfoxide was first achieved in 1996 by Cope and Caress34 with platinum(II) complexes containing either (+)--methylbenzylamine or (-)--methylbenzylamine. The complexes decompose with aqueous sodium cyanide to afford the enantiopure sulfoxide. A comprehensive review of optical resolutions used up to the late 1980s was prepared by Drabowicz et al.35 Figure 13. The most prevalent current synthetic routes towards chiral sulfoxides.

PAGE 28

11 The most widely used of the methods towards non-racemic sulfoxides5 is the oxidation of a parent sulfide. Several oxidative methods have been applied to this task, with varying selectivity both in the stereochemistry of the reaction and the range of substrates that are accomodated. Various organic peracids have been used by Overberger and Cummins.36 in sulfoxidations. Chiral peracids offer low stereoselectivities (in the order of 5-10 ee%)37 although they are highly reactive and lead to overoxidation to the corresponding sulfone unless there is careful monitoring of temperature and excess catalyst. Hydroperoxides and organic peroxides38 alone have also been used and in conjunction with a sulfur co-ordinating metal bearing chiral ligands, modest stereoselectivities can be achieved. Kagan39 and Modena40 modified the Sharpless catalyst and used this for sulfoxidation with some success. Sulfoxidation with Ti(OiPr) 4 /DET/ t BuOOH/H 2 O afforded moderate enantioselectivity although this is on a limited substrate range.41,42 Further increases in the stereoselectivity of this oxidation has been been achieved by replacing the oxidant with cumene hydroperoxide and by changing the chiral ligand. The use of (R)-(+)-binaphthol has increased the modest e.e.% (60-70%) of aryl methyl sulfoxidation of the Kagan modified Sharpless catalyst to 80-96%.43 This takes advantage of the kinetic resolution of the further oxidation selectively of one enantiomer of the sulfoxide to the sulfone by the titanium-binaphthol complex (Figure 14). This does not improve the yield of the reaction though, which requires further purification from the sulfone.44

PAGE 29

12 S CH3 : : CH3 OH OH S+ CH3 : O CH3 CH3 S CH3 O O S+ CH3 O CH3 Ti(Oi Pr)4, tBuOOH, 56% ee (R) ASYMMETRIC OXIDATION KINETIC RESOLUTION 96% ee (R) (44% overall) Figure 14. A modification of the Kagan-Sharpless oxidation of sulfides using (R)-(+)-binaphthol in place of DET.43 Chiral oxaziridines (such as the Davis oxaziridines) can also perform selected sulfoxidations often with high enantiospecificity (Figure 15).45-50 (Camphorsulfonyl)oxaziridines and their 8,8-dichloro derivatives are available in both antipodals allowing access to both enantiomers of the product sulfoxide with stereoselectivities of 84-96% ee.50,51 N O SO2 X X S CH3 : : S+ CH3 : O X = Cl, H OXAZIRIDINESULFOXIDATION CATALYST a = a (X=H)CH2Cl2, rt 85% (22% e.e. (R)) Figure 15 Chiral oxazidines as sulfoxidation catalysts.50,51 The Andersen method was the most widely used route to nonracemic sulfoxides prior to enantioselective sulfoxidations (Figure 16). It is general in scope and can be applied to acquire complex homochiral sulfoxides, although its major drawback is the necessity to obtain the optically pure menthyl p-toluenesulfinate precursor. Esterification of p-toluenesulfinyl chloride with 1-menthol is a racemic reaction that requires separation

PAGE 30

13 of the enantiomers, which is achieved by fractional crystallization.52,53 The maximal yield of each enantiomer is 50% but can be increased with concurrent epimerization using an acid catalyst such as hydrogen chloride (Figure 17). In such a way up to 90% of the (S) sulfoxide ester can be obtained. The reduced facility to obtain the antipodal menthyl sulfinate limits this method to production of only the (R) enantiomer of sulfoxides. S+ Cl CH3 O S+ OMenthyl CH3 : O S+ OMenthyl CH3 O : S+ OMenthyl CH3 O : S+ R CH3 : O (racemic) Menthol,Py, Et2O, rt + Selective Recrystallization acetone, HCl, -20oC Grignard reagent R-MgX (inversion) + Menthol Figure 16. The Andersen synthesis of chiral p-tolyl sulfoxides. The route towards the antipodal sulfoxide is identical but requires different separation techniques to afford the (R)-menthyl ester.9 pTolS+ OMenthyl : O pTolS+ OMenthyl : Cl Cl pTolS+ OMenthyl O : HCl HCl H2O H2O Crystallization (acetone-HCl) 90% (S) (R) Figure 17. Kinetic resolution in the Andersen synthesis, recrystallization in acetone-HCl. The reaction of the arenesulfinate with a Grignard carbanion equivalent is performed in benzene with moderate to high yield as a route to aryl p-tolyl sulfoxides. This requires a change of solvent to remove the menthol byproduct.9 Lithium cuprates

PAGE 31

14 can also be used to generate appropriate sulfoxides providing lower yields (16-59%) but reducing the purification steps.35 The Andersen method has also been adapted for solid phase sulfoxidation on a Wang resin.54 The initial stereoselectivity of the coupling to the resin cannot be controlled though and the reactions are racemic. The Andersen method is also limited to the preparation of alkyl-aryl sulfoxides, as the displacement of the menthol group does not occur in the corresponding alkylsulfinate esters55. To afford alkyl alkyl sulfoxides certain alkyllithium reagents can be used to displace the aryl group with inversion of stereochemistry.56,57 The synthesis of sulfoxides using the Andersen method employs the transfer of an already enantiomerically pure sulfinyl group to a nucleophilic center, often carbon. This technique has been expanded upon with more reactive sulfinates and with sulfoxides that allow controlled double displacement by carbon nucleophiles. This affords a choice in the selection of both substituents (Figure 18). Exclusive inversion at sulfur offers enantioselectivity to these syntheses, provided the initial sulfoxide substrate is enantiopure. The use of benzyl p-bromophenyl sulfoxide for a double displacement at sulfur is an example. It has yields of 50-90% product in >98% e.e. and is a feasible route towards chiral nonracemic alkyl alkyl sulfoxides.58There have been several thorough reviews covering extensions to the Andersen method and applications in synthesis.5,9

PAGE 32

15 R1S+ LG O : R1S+ R2 : O LG1S+ LG2 O : R1S+ LG2 : O R1S+ R2 O : R1S+ O : O CH3 CH3 CH3 O Ph N O Ph N O O Ph CH3 S+ O : NO CH3 CH3 Ph NHO Ph CH3 Ph ON CH3 C10H7 CH3 Br CHIRAL SULFINYL TRANSFER REAGENTS R2R1R2menthol(Andersen method) ephedrine Figure 18. Several chiral sulfinyl transfer reagents for the enantioselective synthesis of sulfoxides.5 Non-racemic MTSO has been produced using the aforementioned methodologies, with varying yields, enantioselectivities and complexity of methods (Table 1). Table 1. Classical organic chemistry sythesis of enantiopure MTSO. Method ee.% % yield (scale) Ref. Andersen synthesis 99% (R) 80% (204mmol, 60g) 59 Kagan-Sharpless with cumene hydroperoxide 96% (R) 44% (0.5mmol, 69mg), 30hrs 43 Kagan-Sharpless with binaphthol ligand 60% (R) 80% (1 mmol, 138mg) 60 Kagan-Sharpless 54% (R) 67% 41 t BuOOH, VO(acac) 2 menthol, benzene/toluene 10% (R) 100% 61 (S)1-phenyl-ethyl hydroperoxide, Ti(O i Pr) 4 80% (S) 71%, with 11% sulfone (400mol, 55mg) 44 Enzymatic Routes to Chiral Sulfoxides Stereochemistry is an additional intricacy in the assembly of atoms; it enriches the diversity of chemistry and the space of biological molecules is acutely sensitive to it. Proteins and large molecules in biological systems incorporate an environment

PAGE 33

16 surrounded by solid which can restrict smaller organic molecules and provide control and selectivity in chemistry. This chiral cavity itself has been used solely to direct stereoselective reactions. Prasanta et al. recently demonstrated that the sulfoxidation of alkyl aryl sulfides by H 2 O 2 which in solution is a racemic conversion, in the presence of the hydrolase -chymotrypsin becomes enantioselective.4 The selectivity of biological catalysts has also been investigated for the oxidation of sulfides. The oxidative metabolism of many biomolecules containing sulfur provides a large assortment of enzymes for this chemistry. Enzymes described in the oxygenation of MTS (Table 2), have been studied for an assortment of reasons, including the interest in their catalytic mechanisms and active site topography, their role in xenobiotic drug metabolism and their possible use as biocatalysts. Microbial sulfoxidation can provide highly enantiopure MTS of both antipodals in at least small scale reactions although they have not been widely used in synthetic applications. There are several factors that can reduce the attractiveness of isolated biological catalysts to organic synthesis. Oxygenation enzymes studied for biotransformations are often membrane-bound, require expensive co-factors, must be extensively purified from tissues and cells and their activity is often coupled with that of other enzymes.62 Many organisms possess more than one catalyst for oxidative metabolism, contaminating the chemistry and further complicating purification.63 This intractability has limited the use of isolated enzymes and instead many biocatalytic

PAGE 34

17 preparations are done with whole-cell systems. The synthetic application of whole cell systems however can be complicated by the pathogenicity of some microorganisms and the requirement in the laboratory for microbiological expertise. Overexpression of catalysts in a heterologuous host such as Escherichia coli can avoid competing reactivity, is relatively quick and has widely established culture methods. This has provided access to the study and application to several sulfoxidation enzymes.64 65 Table 2. Methods for the stereoselective sulfoxidation of methyl p-tolyl sulfide (MTS). Various classical synthetic organic methods as well as isolated enzymes and whole cell systems could be used to produce MTSO. Method e.e.% % yield (scale) Kinetic parameters Ref. Mortierella isabellina (wc) 100% (R) 66 Corynebacterium equi (wc) 100% (R) 67 rabbit FMO1 (cmp), heterologous expression >99% (R) (90nmol, 12.4g) 2.3 nmole/min/mg, K m <10M 64 rabbit FMO2 (cmp), heterologous expression >99% (R) (90nmol, 12.4g) 3.6nmole/min/mg, K m <10M 64 Chloroperoxidase (ip) 99% (R) 83% (1.25mmol, 172.5mg) 2hrs. 68 H 2 O 2 and peroxidase from Coprinus cinereus (ip) 96% (R) 82% (1mmol, 138mg) 8hrs. 69 Saccharomyces cerevisiae (wc) 92% (R) 60% (1.3mmol, 180mg) 70 Vanadium bromoperoxidase (ip) 82% (R) 18% (25mol, 3.45mg) 71 rabbit FMO3 (cmp), heterologous expression 50% (R) (1 mmol, 138mg) 10nmole/min/mg, K m <300M 64 Helminthosporium (wc) 27% (S) 10% 72 Rhesus macaque FMO2 (cmp), heterologous expression 30% (S) (1mmole, 138mg) 65

PAGE 35

18 Table 2 (continued) Method e.e.% % yield (scale) Kinetic parameters Ref. (1-phenyl)ethyl hydroperoxide and peroxidase from Coprinus cinereus (ip) 79% (S) 68% 4hrs. 73 Horseradish peroxidase (mutant) (ip) 77% (S) (100mol, 13.8mg) 19 nmole/min/mol, K m = 4.5mM 74 Pseudomonas putida expressing the naphthalene dioxygenase gene (wc) 94% (S) 11% 75 Pseudomonas putida expressing toluene dioxygenase (wc) 98% (S) 76 wc = whole cell biotransformation, cmp = crude membrane preparation of heterologously expressed protein, i p = isolated enzyme The sulfoxidation reactions on many different pro-chiral sulfur atoms catalyzed by a wide range of fungal and bacterial systems and by isolated enzymes has been reviewed in detail by Holland.77,78 Improving Catalysts The relevance of asymmetric catalysis was evidenced by the 2001 Nobel Prize for Chemistry to K. Barry Sharpless for his work on chirally catalyzed oxidation techniques, as well as Ryoji Noyori and William S. Knowles for their work on chirally catalysed hydrogenation reactions. The two major methods currently available to chemists can be classed as homogeneous transition metal catalysts and biocatalysts.79 The applicability of transition metal catalysts is based on tuning of the ligands from molecular modeling, knowledge of the reaction mechanism, trial and error and some degree of intuition. Within the last decade the technique of combinatorial asymmetric catalysis has allowed the screening of large numbers of catalysts. The production of large numbers of catalysts

PAGE 36

19 requires strategies for modular synthesis of chiral ligands coupled with development of high-throughput assays to monitor enantiomeric excess (ee%.).80 So far the size of the libraries generated has been limited to fewer than a hundred catalysts and the analysis is performed by analyzing reactions conventionally using chiral separation chromatography.81 For protein catalysts, often lacking structural data information and the molecular basis for enantioselectivity or kinetic resolution not being well understood, rational design has had very limited applicability in providing novel stereoselective activities. In directed evolution, a combination of the molecular biological control for changes in enzyme catalysts (mutagenesis) and their expression has led to large libraries, similar to the combinatorial asymmetric catalysts, of possible novel asymmetric catalysts. The source for differing natural enzymes is vast; it has been estimated that >99% of the naturally occurring enzymes have yet to be discovered.82 The other sources of interesting candidates could therefore be from panning metagenome DNA.82 Efficient screening methodologies to study these new libraries have been a relatively recent but exciting research venture.79,83

PAGE 37

20 Figure 19.Sources of large libraries from which enantioselective catalysts can be explored. Directed Evolution of Enzymes Theory and Approaches Whereas in nature enzymes have evolved as agents in the tasks required for life, the requirements of enzymes for our use outside that context are often very different. The properties of enzymes that are of interest to the synthetic organic chemist include regioselectivities, substrate specificities, stereoselectivities, activity in non-aqueous solvents, stability and enhanced reaction rates. There are many useful organic reactions which do not have a known or suitable enzyme catalyst. Arrangements of atoms through to larger assemblages is not only unique in the order of subunits (in the 1 st dimension) but extends into the 3 rd dimension. Uniqueness in the co-ordinates of shapes can be, as much as the composition of shapes, determinable in biological chemistry. Enzymes are conglometrates of atoms extending into space, the occupation of these molecules in the 3 rd dimension can alternatively reduce or mask the

PAGE 38

21 distinctiveness of the subunit composition, although both could influence the properties of the whole. Can this shape be molded, without prior knowledge of the relationship between the composition of the 1 st and 3 rd dimensions? Aside from rational design, the blunt probing of this relationship could possibly help elucidate these questions and concurrently might provide biological catalyst tools with increased effectiveness in chiral chemistry. The prediction of the chemistry of the known enzymes to substrates in organic synthesis is usually restricted. Little structural information is often available for many enzymes which limits substrate suitability and stereoselectivity predictions.77 One way to circumvent this has been through probing using substrates of various forms and stereoelectronic configurations62 to develop a model of the active site. Another method to increase the palette of available biocatalysts for a given reaction is development of new enzymes. Within the last 7 years several enzymes have been evolved artificially towards an industrial or biotechnological catalytic goal. Arnold et al. 84 and Stemmer and coworkers85 have pioneered the field known as the directed evolution of proteins. The enzymes that they have modified included many for which no previous structural data was available. Enzymes are long sequences of the more than 20 different amino acids. This complexity is reflected in variants of a protein, of which there are at least N 20 The in vitro generation of altered enzymes takes advantage of molecular biological techniques which control protein sequence at the DNA source to generate some of these variants.

PAGE 39

22 One approach generates randomly mutated descendants of proteins, each possessing a sequence with several alterations dispersed along the length of the gene. Larger libraries of these enzyme variants will undoubtedly have more likelihood of possessing improved enzymes. Large numbers are also important due to the observations that most of the generated mutants are inactive (deleterious mutations) or do not have altered properties (neutral mutations) and that only a small percentage have beneficial changes in the properties that are being tested (positive mutations).86,87 The size of the library of enzyme variants (mutants) is exponentially proportional to the number of changes allowed per mutant.88 Too many mutations per gene can easily create libraries much too large to screen in their entirety. It is therefore the screening or selection strategy that limits the size of libraries used. Even with these limited library sizes, which with current screening/selection methods are currently at 10 5 -10 6 members (covering about 1-2 mutations per clone), directed evolution has provided several significantly improved enzymes. Thermostability89, activity in acidic conditions90, substrate specificity91, activity in organic solvents91, stereoselectivities92 and kinetic resolutions93 have all been improved for various catalysts using the strategy of random mutagenesis and screening. The applications of in vitro evolution to develop enzymes with improved characteristics are numerous and beyond the scope of this thesis, although there are several extensive reviews94,95 on the subject. There are several requirements to successfully evolve a biocatalyst; a method to create altered descendants of enzyme, an efficient way to express and present the enzyme and a method to screen or select for the desired properties of the enzyme is essential.84

PAGE 40

23 High Throughput Screening for Enantioselectivity The screening of asymmetric catalyst libraries, whether homogeneous transition metal catalysts or biocatalysts, requires monitoring to provide the ee% of each reaction. The large numbers of samples in catalyst libraries also require methods that are fast, sensitive and can be used with high-throughput. A thorough synopsis of some of the clever recent techniques has been prepared by Reymond.80 The most convenient and prevalent method to screen large numbers of reactions rapidly involves measuring a chromogenic or fluorogenic signal in the products. There are several known chromophores and fluorophores which have been used to task (Figure 20). Many can even be taken up by cells and allow for monitoring of catalysis in vivo. X-gal hydrolized by -galactosidase produces insoluble blue dye indigo (Figure 20A) 96 visible through unlysed cells. A coupled reaction system using horseradish peroxidase to polymerize products of mutant cytochrome P450 also produces signals that can be detected through cell membranes (Figure 20B). Digital imaging of the cultures can detect the fluorescent polymers produced.97,98

PAGE 41

24 O OH OH O OH OH NH NH O Cl Br NH O Cl Br OH O O O OH OH O OH OH O CH3 O OH O CH3 O -Galactosidase Indigo blue A P450cam HRP, H2O2 (fluoroscent dimer) B -Galactosidase 7-hydroxy-4-methylcoumarine C Figure 20. Enzymatic reactions can produce highly chromophoric and fluorogenic molecules that can be used to screen for activity. Reaction screening for enantioselectivity has required an array of intricate methodologies. Some of the first catalyzed reactions to be screened were the hydrolysis of esters adapted to release the chromophore p-nitrophenol (Figure 21)93 or fluorophores such as coumarinyl derivatives (Figure 20C).99,100 The enantioselectivity of protein catalysts either is due to kinetic resolution in which the enzyme selectively converts one to the enantiomers of a racemic substrate into product or to a stereoselective reaction in which catalysis converts an achiral substrate into a chiral product. Kinetic resolution is usually expressed as an E-value, the ratio of activity the enzyme has towards each enantiomer101. At high dilutions of substrate (below the Michaelis-Menten constant K M for each of the enantiomers) this can be determined

PAGE 42

25 directly from the reaction rates for both enantiomers. A system was initially developed by Reetz and co-workers83 to monitor the kinetic resolution of lipase catalyzed reactions. The lipase catalyzed kinetic resolution of hydrolysis of p-nitrophenol esters was done for each enantiomer separately and analyzed pairwise using simple UV/Vis based 96-well plate readings. This simple screening (initial screening was done using spectrophotometric methods and interesting reactions were further analyzed using chiral GC) was done on large libraries (1000-5000 clones). Several rounds of library screenings discovered a lipase, improved from Pseudomonas aeruginosa (PAL), initially having essentially no enantioselectivity (E = 1.1) to preference for hydrolysis of the (S) isomer of 91% ee. at 18% conversion (E = 25.8) (Figure 21).93 This has been to date the only example of successful enantioselective evolution of a lipase. Wahler80 cautions on the disadvantages to comparing reactions monitored separately with each enantiomer. R O O CH3 NO2 R O OH CH3 OH NO2 R O O CH3 NO2 R = n-C8H17 PAL lipase, H2O + + (R) 18%, 91% ee. (S) Figure 21. An enantioselective mutant PAL catalyses the kinetic resolution of a racemic p-nitrophenol ester.83 Methods for the monitoring of improved kinetic resolution of other enzymes have also been developed. Arnold et al. improved the kinetic resolution of a hydantoinase towards the production of D-methionine by selecting for the best mutants of the enzyme in the hydrolysis of the enantiomers of 5-monosubstituted hydantoin (Figure 22). The enzyme is presented with each enantiomer separately and the formation of product is monitored in both reactions from the increase in acidity, using a simple pH indicator

PAGE 43

26 (cresol red). This affords a rapid screen of mutants, of which >20 000 were screened, for the improved activity towards the L enantiomer. Further characterization of initial positive clones was confirmed by chiral HPLC analysis. The evolved enzyme mutant identified produces 20% ee. of the D enantiomer compared to the 40% ee. L selectivity of the wild type enzyme. NHNH O O H R NHNH O O R H NHN2OH H O O H R NHNH2OH O O R H D-enantiomer L-enantiomer 5-monosubstitutedhydantoin H2O H2O HYDANTOINASE CATALYSED REACTION N-Carbamoylamino acid Figure 22. The reaction catalyzed by hydantoinase, which has distinct enantiomeric substrates (the inherent racemization between the 5-monosubstituted hydantoin isomers is slow102). Reetz and coworkers have also developed a monitoring system for kinetic resolution reactions using mass spectrometry.103 The technique requires that one enantiomer be isotopically labeled to distinguish it using electrospray mass spectrometry. With an eight-channel sprayer system samples can be run very quickly (70sec/sample) and library sizes of up to 10 000 can be screened per day. As opposed to kinetic resolutions directly screening enantioselective reactions that occur with prochiral substrates it is a bigger challenge.

PAGE 44

27 Using the principle of microscopic reversibility these reactions might be treated as kinetic resolutions, since many enzyme catalyze reversible processes. Microscopic reversibility predicts that the ratio of the rates of the reverse reactions indicates the enantioselectivity of the forward reaction. Catalysts are provided with the enantiopure products of the forward reactions and monitoring is done for the products of the reverse reaction ie. the natural substrates for the enzyme. Sensitive detection is required for the low yield of these products or as was done by Reymond et al. coupling with a second irreversible step can be used to help trap and detect the product (Figure 23).104 The reverse reaction of alcohol dehydrogenase on a fluorogenic derivative of 2-butanone is trapped by further reaction to the fluorescent product umbelliferone. This -elimination is catalyzed by bovine serum albumin in alkali media. CH3 O O O OH CH3 O O O OH CH3 O O O O OH O O BSA pH > 7 7-hydroxy-coumarin (fluorescent) Alcohol dehydrogenase NADP+ (R) (S) Figure 23. The enantioselective reduction of a coumarinyl derivative of 2-butyl ketone can be monitored using high-throughput fluorescence screening for the reverse oxidation reaction coupled with elimination to form a fluorophore product. Very recently examples of screening systems have been developed for high-throughput % ee., analysis of various reactions. These have been thoroughly reviewed by

PAGE 45

28 Reetz.105 An 1 H-NMR based screening assay allows up to 1400 ee determinations per day (using an appropriate flow-through cell and autosampler)82, capillary array electrophoresis (allowing in some cases 30 000 ee. determinations per day)106 and the use of circular dichroism spectroscopy (CD) (which can accommodate up to 10000 samples per day)81 have all been developed within the last 4 years. The work of Wagner and coworkers on immunoassays to selectively visualize chiral products107 is a screen for enantioselectivity which is powerfully adaptable. Antibodies can be raised to almost any compound of interest. An antibody recognizing both enantiomers can be used to measure yields of reactions and one specific for an isomer (in the analysis of mandelic acid prepared from the Ru-catalyzed hydrogenation of benzoyl formic acid) can be used to indicate the stereoselectivity (Figure 24). Figure 24. Screening for enantioselectivity using competitive enzyme immunoassays107. A colourimetric assay to quantify the racemate (yields of a reaction) and the stereoselectivity with two antibodies, providing a method to screen for yields and %ee. of reactions. One antibody is blue and recognizes both enantiomers and the other is red and is (S) specific. The aforementioned methods have been used to direct the evolution of the kinetic resolution of various enzymes and for the screening of transition metal asymmetric catalysts. These methods though have as yet not been employed in the screening of an enzyme library for the stereselectivity of a reaction that starts from a racemic prochiral

PAGE 46

29 substrate such as the sulfoxidation of an unsymmetrically substituted sulfide. In addition there has yet to be an example of an enzyme catalyzing such a reaction whose enantioselectivity has been inverted from preference towards one isomer to that of the antipodal. Flavin-Monooxygenases Flavin nucleotides have rich redox chemistry, in various protein environments, acting in conjunture with other functional groups they can be the catalytic part of reductases108, oxidases109,110 and monooxygenases.111 The first of the enzymes termed flavin-containing monooxygenases was purified from readily available pig liver in 1972 by Carolyn Mitchell (a student of Daniel M. Ziegler at the University of Texas at Austin).111 The isolated enzyme was associated with equimolar amounts of FAD but not with other co-factors or metals and was identified initially as an NADPH and O 2 dependent amine N-oxidase. It was later found that it catalyzes the oxygenation of a wide range of inorganic and organic substrates that contain soft nucleophilic heteroatoms such as Se, S and P and of N and B centers also.112 These substrates have a variety of structures and sizes, but are with few exceptions devoid of charged groups113. Oxygenation of sulfur atoms occurred with high enantioselectivities.114 FMOs have since been isolated from mammalian and fungal sources, each providing several forms of the enzyme. The liver form (hepatic), lung form (pulmonary) and kidneys form (renal) have all been characterized from rats, pigs, humans and rabbits. FMOs are microsomal enzymes whose main purpose is postulated to be in the oxidative metabolism of environmental toxicants, natural products, and therapeutics. In

PAGE 47

30 humans FMOs are responsible for metabolism of many xenobiotics and drugs.115 They promote detoxification by oxidizing nucleophilic centers116 and, along with the cytochrome P-450s, oxygenate primary and N-substituted amines into hydrophilic products, that could be more readily excreted from the body. Flavin containing monooxygenases are classified as Phase I or Oxidative Drug Metabolism Enzymes. Background and Properties Both flavin (FMOs) and heme-containing monooxygenases (cytP450) are to be found associated with the endoplasmic reticulum of mammalian cells. Both of these enzyme types are NADPH and O 2 dependent and are grouped within their respective family into members related by primary amino acid sequence. The genes are expressed in species and tissue specific levels but do not appear to be inducible 116 A standardized nomenclature for this family was advanced jointly in 1994 by some of the major researchers in the field117 based on the guidelines established for the cytochromes P450s 118. This nomenclature is based on similarities in primary amino acid sequence and replaces the previous laboratory specific classifications. Currently there are 5 agreed upon isoforms of the flavin-containing monooxygenase (FMO) gene family in each mammal extensively studied to date (Table 3). Homologs from different species are grouped into the 5 families of isoforms; isoforms (from different species) possess >80% identity whereas there is <60% identity between homologs within the family. Structure defines the classing irrespective of the enzymes distribution among tissues or functional or chemical properties. As part of the human genome effort, another FMO-like gene in humans, FMO6, was identified between FMO3 and FMO2 (GenBank

PAGE 48

31 accession no. ALO21026) although reverse transcriptase coupled polymerase chain reaction (DNA amplification) show that most of the gene transcripts would not code for functional protein (based on an analysis of possible open reading frames).119 FMO has yet to be used in organic synthesis, although FMO have been used in biosensors enzyme electrode and a bioelectronic nose (bio-nose) to detect methyl mercaptan (detecting 1-4000ppm).120 The drive to studying mammalian FMOs has mainly come from the medical need to further understand the mechanisms of oxidative metabolism. This has been an impetus for determining human FMO substrate specificity to anticipate the metabolism and interaction with drugs.121 There is developing though a significant gap in information pertaining to orthologs in insects, echinoderms, avian, reptilian and amphibian species.122 Very rare mutations of the human FMO3 gene that have been associated with deficient N-oxygenation of dietary trimethylamine.123 Defective trimethylamine N-oxygenation causes trimethylaminuria or "fish-like odor syndrome". In cows a nonsense mutation in FMO3 underlies a fishy off flavour in cows milk.124 The rabbit FMO4 gene has as yet no protein counterpart, either isolated or expressed heterologuously125. The genome of humans and mice have revealed that all isozymes of FMO reside in a single cluster of a chromosome 1a.121,126 The FMO2 from Rhesus macaque is 97% identical to human FMO2.127Due to the close proximity of the genes and the uniformity of differences in the sequences of the orthologs (ie. all members of a family are similar across species) Cashman126 has postulated that the genes for FMOs have arisen from duplication of a single ancestral gene 250-300 million years ago, 200 million years

PAGE 49

32 before mammalian speciation. A enzyme in Arabidopsis thaliana and one in yeast with similar amino acid identity have also been termed FMOs. The purpose of the yeast enzyme is postulated to be in maintaining the thiol-disulfide redox potential in the cell.128-130 Although enzymes in the family were identified based on sequence information a classification by Cashman126 and Ziegler114 of the family was also developed to describe their activity characteristics: Flavin-containing monooxygenase (EC 1.14.13.8) or FMO NADPH dependent oxidative catalyst, utilizing both O 2 and H + and which is associated with a flavin adenine dinucleotide (FAD) prosthetic group. a microsomal or membrane associated protein. multisubstrate enzyme which oxygenate many nucleophilic atoms. These include sulfur, selenium, phosphorus and nitrogen centers. they can form a kinetically stable 4-hydroperoxyflavin independently of the presence of substrate Table 3. List of identified and putative FMOs to date. Isoform Organism Tissue Level Genbank # Ref. FMO1 Pig liver cDNA M32031 131 Rabbit liver cDNA M32030 132 protein 133 Human kidney cDNA M64082 134 Rhesus macaque liver cDNA 135 Rat liver, kidney cDNA M84719 136 Mouse kidney cDNA P50285 137 Dog cDNA 138 Saccharomyces cerevisiae cDNA 139

PAGE 50

33 Table 3 (continued) FMO2 Rabbit lung cDNA M32029 132 lung protein 140 Guinea pig cDNA 141 Human Liver, kidney, lung cDNA 4503759 142 Rat cDNA 143 Isoform Organism Tissue Level Genbank # Ref. Gorilla cDNA 144 Mouse lung cDNA 145 Chimpanzee cDNA 144 Rhesus macaque lung cDNA P97501 146 FMO3 Rabbit liver cDNA L10391 147 Mouse liver, kidney cDNA U87147.1 148 Human liver cDNA M83772 149 Cow cDNA 124 Rat kidney, liver cDNA 150 Arabidopsis thaliana cDNA 151 FMO4 Rabbit kidney cDNA L10392 152 Human liver cDNA Z11737 142 Rat kidney cDNA 153 Mouse cDNA 154 FMO5 Rabbit liver, kidney cDNA L08449 155 mouse liver, kidney cDNA U90535 137 Human liver cDNA M83772 Rat liver, kidney cDNA 156 Guinea pig liver cDNA L37081 157 FMO6 Human liver DNA 119 AF458414.1 tissue of predominant localization putative FMO, from extrapolated primary sequence comparison, but not corroborated with studies of activity

PAGE 51

34 Catalysis by FMOs A catalytic cycle was postulated by Ziegler and Ballou based on kinetic and spectral studies with FMO from rabbit lung microsomes63 and pig liver microsomes158-160 (Figure 25). In the ordered sequential cycle of these FMOs, interaction between the xenobiotic substrate and the enzyme is preceded by binding of NADPH and O 2 The relatively stable 4-hydroperoxyflavin form of the enzyme, seen in the presence of NADPH and O 2 has been identified by similarities in its spectral characteristics to N 5 -Ethyl-4-hydroperoxyflavin.161 It is resistant to decomposition, long-lived and forms irrespective of substrate binding as compared to other flavoenzymes such as cyclohexanone monooxygenase (CHMO from Acinetobacter sp.). Attack of the terminal flavin peroxide oxygen by nucleophiles (such as N in dimethylaniline) occurs (Step A) rapidly in a bimolecular reaction (4700 M -1 s -1 ) forming the pseudobase 4-hydroxyflavin form of the enzyme and immediately releasing the oxygenated product. Only a single point of contact between the enzyme and the substrates studied is postulated to be required for formation of the product.158-160 Consequently any soft nucleophile not excluded from the active site will be oxidized and released in the first step. Conversely the active site is likely lipophilic and free of nucleophilic residues. The precise fit often needed to lower the activation energy of an enzyme-catalyzed reaction is not required since the energy to drive the reaction is already present in the 4-hydroxyflavin form of the enzyme before it encounters the substrate. This reduces the relevance of describing K m for FMO enzymes in terms of high or low affinity substrate binding. These enzymes do however follow the approximations of Michaelis-Menten saturation kinetics.158

PAGE 52

35 The next step (Step B), dehydration of the 4-hydroxyflavin was demonstrated by rapid kinetic studies 160 to be about 10 times slower than any other step (1.9M -1 sec -1 ). Solvent deuterium effects also suggest participation of a general acid catalyst162 which could be participating in the dehydration. In the enzymes studied (porcine liver and rabbit lung FMOs) this step becomes rate limiting. Since it occurs after product release, reactions following this mechanism should have similar V max values for all good substrates. The oxidized co-factor NADP + is a competitive inhibitor of NADPH and is released in the following step (Step C) prior to regeneration of the active 4-hydroperoxyflavin form of the enzyme. In the absence of bound NADP + FMOs becomes a NADPH oxidases; molecular oxygen reacts with the reduced form of the enzyme and produces significant levels of H 2 O 2 If this were to occur extensively naturally in FMOs it would make them a source of oxidative damage inside the cell as well as reduce cellular levels of NADPH. The fully oxidized flavin form of FMO reacts (53 M -1 ) with NADPH (Step D) to give an enzyme with bound NADP + and reduced flavin. The reduced flavin can then form (45 M -1 ) the 4-hydroperoxyflavin in the presence of molecular oxygen (Step E) regenerating the active form of the enzyme.

PAGE 53

36 Figure 25. The catalytic cycle postulated for pig liver FMOs. The 4--hydroperoxyflavin bound to the enzyme oxidizes the substrate in step (A) followed by release of water (the rate limiting step) (B). The active form of the enzyme is regenerated in steps C, D, E independent of the presence of substrate. This mechanism was developed using the porcine liver and rabbit lung FMOs which have been studied in much detail and whose oxidative activities are limited to small molecule N and S oxidation, but it is hoped that most other FMO isozymes follow similar mechanisms and that this scheme is applicable to various xenobiotic substrates.

PAGE 54

37 In fact rat and human liver enzymes (FMO1) defer in that the decomposition of the pseudo-base, and dehydration does not appear to be rate limiting.163 It is also known that rabbit FMO5, classed as an FMO based on primary structure164, does not freely form the 4-hydroperoxyflavin intermediate which is postulated to account for its restricted substrate specificity. Its catalytic cycle could be similar instead to other studied flavoenzymes such as to that of p-hydroxybenzoatehydroxylase.165 So far crude membrane preparations and the purified enzyme shows only N-oxidation activity of n-octylamine and n-nonylamine. Amine nitrogens in comparison to sulfur, selenium or other readily polarizable heteroatoms are not as nucleophilic. There is more to the active site in this reaction as the enzyme somehow activates these moieties, 4-hydroperoxyflavin bound to FMO oxidizes amines 10 6 times faster than in solution.166 Structure of FMOs These membrane-associated proteins have as yet not been crystallized. Active site models have been postulated based on substrate structure acceptance for catalysis.167 The individual substrate profiles of each FMO studied can be explained by and provide some insight into the active site channel. Some structural knowledge is also predicted from relatedness to protein with crystal structures. In particular the residues 12-23 of FMOs IGGGPGGLAAAR are 83% identical to the glutathione reductase sequence IGGGSGGLASAR which is a -sheet and turn region which forms the floor of the FAD binding site.168

PAGE 55

38 Based on substrate stereoselectivy profiles of sulfoxidation and the ability to distinguish between pro-S and pro-R sulfur atoms, Cashman126 and Rettie169 have proposed a structure for the active site of several FMOs (rabbit and pig FMO1 and rabbit FMO2). According to this model (Figure 26) the sulfide passes into the binding channel in a committed orientation. The substrate binding channel leads to two compartments, one pocket is larger (A) and prefers to bind large, planar aromatic residues and is less accessible but can accommodate short n-alkyl chains or a p-tolyl moiety. With this topography MTS is oriented so as to present the pro-R lone pair of electrons, exclusively to the hydroperoxide. Figure 26. Model of the active site of rabbit and pig FMO1 and rabbit FMO2 proposed by Cashman126 and Rettie169 showing the orientation substrates (propyl 2-naphthyl sulfide, methyl 2-naphthyl sulfide, MTS, heptyl p-tolyl sulfide) would adopt in the active site. There are two principal substrate binding pockets A and C adjacent to the flavin cofactor (below the surface of the active site) and the binding channel D open to the surface of the enzyme. At a critical length the n-alkyl chain of alkyl aryl sulfides preferentially occupies pocket A, reversing the stereoselectivity of the reaction. This critical length varies among

PAGE 56

39 several FMOs tested (Figure 28)64,135,170 and would be indicative of the differences between the compartments A and C in the each enzyme with the proposed model. From this analysis the substrate binding channel is clearly distinct among forms of FMO studied so far167,171,172 and provides enzymes with varying stereoand substrate selectivities. Larger substrates readily oxidized by pig, guinea pig and rabbit FMO1 are excluded from the active site of human FMO1173 indicating a large variation between homologs. FMO cDNA encode proteins generally 533-535 amino acids in size (Figure 27) although examples of proteins with additional 19174, 25142 amino acid at the C-terminal ends have been observed. Figure 27. The topography of the ~535 amino acid protein of rabbit FMO1, 2, 3.

PAGE 57

40 Rabbit FMO1304050607080901001234n-alkyl chain length (n)ee% (R) Rabbit FMO2304050607080901001234n-alkyl chain length (n)ee% (R) S (CH2)nH Rabbit FMO3304050607080901001234n-alkyl chain length (n)ee% (R) Rhesus macaque FMO1304050607080901001234n-alkyl chain length (n)ee% (R) Figure 28. Differences in the enantioselectivity of sulfoxidation by purified FMOs from different sources64,135 with increasing alkyl chain length of alkyl p-tolyl sulfides.

PAGE 58

41 Hydropathy profiles175 of FMO orthologs show similarities even in regions of very modest identity (25-30%). The strong membrane associations of all mammalian FMOs studied to date is revealed in the poor solubility and highly intractable nature of the purified proteins.140 Although the lipophilic membrane associated region is predicted by hydropathy profiles of primary sequence to lie in the 215 C-terminus amino acids, truncation of this region does not diminish membrane association176. Cashman has assumed that the N-terminal end of the peptide is also not likely to be inserted into cellular membranes due to the presence of binding sites for cytosolic cofactors (NADP + FAD).126 Fusion proteins of FMO joined at the N-terminal end to peptide fragments such as -galactosidase126, maltose binding protein177 and His 6 and His 10 tags show added solubility and stability to proteolysis and thermal inactivation177 as well as improved activity of the protein isolates. FAD binding domain (GxGxxG)178 is common in all FMOs studied, at amino acid position 9-14 as is the NADP + binding domain at position 186-196.126 Post-translational modifications has been detected in several FMOs (rabbit FMO1, FMO2, FMO3) from native tissues and include the proteolitic cleavage of the initiation amino acid (methionine) and the N-acetylation of the following (N-terminal) amino acid.179 The porcine FMO1 protein has been shown to be glycosylated at Asn 120180 and putative consensus site for glycosylation (Asn-Xaa-Ser/Thr) is present and highly conserved between all FMOs.

PAGE 59

42 Although this might be indicative of such modifications either structurally or catalytically, lack of glycosylation (as in proteins heterologously expressed in microbial systems) or other post-translational modification has no measurable effect on the substrate specificity, stereoselectivity, kinetic properties of catalysis or the robustness of the enzyme.126 Single amino acid changes have been observed previously to drastically affect the physical and chemical properties of FMOs181 such as their migration on SDS-PAGE gels. It has been proposed by Cashman that few or even single amino acid mutant of FMOs would possess markedly different catalytic properties.126 Strategy Towards A Stereoselective Sulfoxidation Catalyst, An Overview of this Work This project is attempting to use directed evolution to provide catalysts for the enantioselective synthesis of sulfoxides. Flavin-containing monooxygenase enzymes with a preference for one enantiomer in the sulfoxidation of MTS will be modified towards production of the opposite enantiomer. The resultant enzyme variants will hopefully possess the catalytic activities inherent in the native FMO and with the template gene would yield a pair of enzymes to provide the choice of routes to both antipodals of the sulfoxide MTSO. Catalytic activity will be modified in a random fashion with methods that have been established for other enzymes94,95 and screened for improved enantioselectivity. The general scheme is depicted in Figure 29.

PAGE 60

43 Figure 29. An overview of the strategy for the directed evolution of an FMO enzyme as a catalyst for the efficient production of (S) MTSO, a molecule with many uses in organic syntheses.

PAGE 61

CHAPTER 2 ENGINEERING OF MICROORGANISMS FOR FMO PRODUCTION FMOs from Oryctolagus cuniculus (rabbit) and Rattus norvegicus (rat) where the first full sets ever be isolated and extensively characterized. All 5 of the known rabbit FMO genes were graciously provided to us by Dr. R. Philpot. The FMO2 cDNA of rhesus macaque was a generous gift of Dr. D.M. Ziegler. They are all products of reverse transcription from the host animal mRNA, a process which removes the non-coding information that can be present in many mammalian genes.176,182 These genes were intended for cloning into heterologous expression systems. The extensive knowledge of its genetics and biochemistry has made Escherichia coli the organism predominantly chosen for cloning studies182. This prokaryote can express high levels of protein, has simple and inexpensive media requirements and relatively rapid growth.183 The mammalian proteins expressed from Escherichia coli however lack many of the post-translational modifications of higher organisms, which have been identified in the native FMO proteins, including acetylation of the N-terminal amino acid179 and amino acid side chain glycosylation.180 They might also require refolding into their active form.183 Further complications can occur due to the fact that the cDNA derived genes have not been optimized for the differing codon usage for whichever heterologuous expression system is used, such as Escherichia coli184 or Saccharomyces cerevisiae.185 44

PAGE 62

45 Cloning and expression of several of the rabbit FMO cDNA proteins has been recorded in work done previously in the labs of Philpot174,176,186 and Ziegler.63 The exception seems to be rabbit FMO4, which to date has not been expressed in an active form from cDNA.126 The expressed proteins were similar to those isolated from rabbit tissue based on their mobility on SDS-PAGE gels and from recognition by antibodies raised to the naturally occurring peptides.174,176,186 They agreed in other physical properties, including optimal pH for activity and thermal stability and their oxidation activity towards the test substrates: chlorpromazine126,187, trifluoroperazine187, methimazole174,176, n-octylamine174, thiourea176, dimethylaniline176 and cysteamine.176 The exception is the enzyme of rabbit FMO5 which has only been isolated in its heterologously expressed form and has a distinctly limited substrate range. Concurrently Saccharomyces cerevisiae (yeast) is the eukaryotic microorganism with the best known genetics and most extensive use as a cloning host.182 It possesses many of the post-translational processing systems present in higher organisms, has growth rates faster than many higher eukaryotic organisms, reduced media requirements, lower costs to maintain and is benign.183 Escherichia coli and Saccharomyces cerevisiae expressed FMO have been used as representatives of the naturally occurring enzymes in many studies to date.126 There are several advantages to these expressed forms. Purification in large amounts can be done without complication from the separation of the various FMO isoforms, which are often co-expressed in their native tissue.126 The E. coli expression system also allowed

PAGE 63

46 for the testing of synthetic variants of the FMOs: truncated proteins to probe for the location of the membrane binding domain176 and site specific mutant peptides to confirm the identity and efficacy of the binding sites for the NADPH and FAD cofactors.178 In our labs, expression of rabbit FMO1 and FMO2 as well as rhesus macaque FMO2 was of interest for the directed evolution of their stereoselectivities of catalysis. The rabbit enzymes catalyzing very highly selective sulfoxidation to the (R) sulfoxide (99% ee)135 while the enzyme from the primate possesses the inverse enantioselectivity, producing 30% ee. (S) product.65 These values are both for the native enzymes isolated from animal tissues. In the current study, the enzymes are being expressed in Escherichia coli and it is needed to corroborate these activities for these forms. These catalysts are relatively similar in sequence identity (60-85%), making them attractive targets for creating chimeras, or DNA shuffling of regions between them. The expression of enzymes for these studies has different requirements added to those of research done with FMOs in the past. Whereas a clean source of enzyme is the main prerequisite for the study of an enzymes properties, for biocatalysis done in parallel on large numbers the applications also require production of concentrated catalyst in small volumes. There are several factors that can be developed to optimize this. Expression Using Escherichia coli Biocatalytic reactions used in the high-throughput screening of large numbers of enzymes requires a reproducible presentation of the enzyme with the required co-factors with limited pre-reaction and post-reaction processing. Catalysis using intact cells of

PAGE 64

47 Escherichia coli expressing the membrane associated FMO proteins could provide encapsulated high concentrations of the necessary co-factors FAD and NADPH In minimal media containing glucose Escherichia coli cells can contain 200M of NADP/NADPH and 51M of FAD.188 Intact cells are also relatively simple to prepare for an aqueous reaction, which can be done in growth media itself or by transferring cells to a reaction buffer. If intact whole cells cannot be used, lysis of the cells and semi-isolation of the membrane associated proteins can de done to present expressed proteins. Expression of Active FMO Enzymes Escherichia coli expression plasmid vector based systems vary among several regulatory sequences. Bacterial plasmids adapted for expression of FMOs incorporate a low-stringency high copy number origin of replication (pBR322 or ColE1)182, an antibiotic resistance gene (such as ampicillin or kanamycin) and an IPTG controlled promoter-terminator couple (Figure 30). The strategies that were followed for the construction of the various expression plasmids are recorded in Appendix A.

PAGE 65

Table 4. Various expression systems for rabbit and rhesus macaque FMOs in Escherichia coli and Saccharomyces cerevisiae. Plasmid Strain Gene Prom Selectablemarker gene ori Expression Activity Notes pAAP10 15C (yeast) rabbit FMO1 GAL1 URA3 2 No protein visible 1.8 x 10 -8 (b) pAAP11 BL21(DE3) rabbit FMO1 T7 Amp R pBR322 High 8.4 x 10 -6 (a) 1.4 x 10-5 (c) pAAP27 BL21(DE3) rabbit FMO1 T7 Kan R pBR322 High 5.6 x 10 -5 (c) pAAP27 BL21(DE3)codon + rabbit FMO1 T7 Kan R pBR322 High 4.9 x 10 -5 (c) pAAP15 BL21(DE3) rabbit FMO2 T7 Amp R pBR322 High ND (a) Not active pAAP21 BL21(DE3) rabbit FMO2 T7 Kan R pBR322 High ND (a) Not active pAAP21 BL21(DE3)codon + rabbit FMO2 T7 Kan R pBR322 High ND (a) Not active pAAP12 XL1 Blue rabbit FMO2 tac Amp R pBR322 Low ND (a) Not active pAAP20 XL1 Blue rabbit FMO2 trc Amp R pUC Low ND (a) Not active pAAP14 BL21(DE3) rabbit FMO3 T7 Amp R pBR322 Low 5.3 x 10 -6 (a) pAAP18 BL21(DE3) rabbit FMO4 T7 Amp R pBR322 No protein visible ND (a) No protein produced pAAP35 BL21(DE3) rhesus macaqueFMO2 T7 Amp R pBR322 High 2.1 x 10 -5 (c) pAAP36 BL21(DE3) rhesus macaqueFMO2 T7 Kan R pBR322 High 7.8 x 10 -5 (c) pAAP13 15C (yeast) rabbit FMO2 GAL1 HIS3 2 No protein visible 2.3 x 10 -8 (a) 48 expression was determine from SDS-PAGE analysis of total cell extracts monitoring for presence of ~60Kda protein activity (mmole/min/mg protein) was measured as a) the consumption of NADPH in the methimazole coupled oxidation assay using crude membrane protein preparations or b) the oxidation of thiobenzamide c) the oxidation of MTS (Appendix A), ND = not detectable

PAGE 66

49 Figure 30. General schematic of an Escherichia coli expression plasmid vector. Rabbit FMO1 and Rabbit FMO3 In our work the initial cloning of the rabbit FMO genes was done into the commonly used pET based Escherichia coli expression system (Novagen). The system provides a high copy number of plasmids inside cells, controlled by the origin of replication derived from pBR322.189 IPTG induced production of other proteins using this highly selective T7 RNA polymerase have resulted in levels of expression as high as 50% of the total cellular protein.190 Protein production for rabbit FMO1, FMO2 and FMO3 gene products (~60KDa peptide bands not visible in the microbial protein background) was as high as 20, 30% and <10% of total cellular protein respectively (Figure 31). There was no visible protein expressed from the FMO4 clone, which is in agreement with published results.64

PAGE 67

50 Figure 31. Expression of rabbit FMO1, expected at ~ 59kDa132, from BL21(DE3)/pAAP11 analyzed with SDS-PAGE. Total protein samples run on 12% SDS-PAGE: A) 6hrs. post-induction BL21(DE3) cells (background), BL21(DE3)/pAAP11 post-induction: B) 0hrs., C) 2hrs., D) 4hrs., E) 6hrs., F) Protein standard. The cloned genes of rabbit FMO1 and FMO3 possessed identical nucleotide sequence with the open reading frames of the genes described by Rettie et al. (Appendix B), without further mutations which could have accumulated during cloning. A further characterization of expressed products from the FMO1 and FMO3 systems (pAAP11 and pAAP13) indicated that proteins had activities in the oxidation of methimazole and MTS (Table 4). Oxidation was monitored indirectly by the consumption of NADPH during the reactions (Appendix A). This corroborated the expression of both FMO enzymes (rabbit FMO1 and rabbit FMO3) with activities for known substrates.64 Rabbit FMO2 There was no active protein apparent in the rabbit FMO2 system although a large amount of protein of the expected size (~60KDa) is visible by SDS-PAGE analysis (pAAP12, Table 4). This is not due to any visible inclusion bodies, conglomerates of misfolded proteins, inside the cell since no distinction can be made between the cells

PAGE 68

51 expressing rabbit FMO1 or FMO2 and the native Escherichia coli strain in electron micrographs (Figure 32). An additional indication that overexpression-induced inclusion bodies are not responsible for the lack of active protein was the similar inactivity observed for FMO2 protein produced by a system with for lower protein expression levels. The pJLS plasmid system incorporates an analogue of the Escherichia coli trp promoter for the heterologuous expression of proteins at lower levels than the expected results from the highly selective expression in the pET (T7 promoter-driven) system.170 Sequencing of the cloned rabbit FMO2 to test for an inactivation mutation at the DNA level was done and agrees with that expected from the initial identification of the gene by Philpot and coworkers170 therefore the gene is not inactive due to mutations. Rhesus macaque FMO2 A similar expression system was initially constructed for rhesus macaque FMO2 (pAAP36) which indicates that a ~60KDa protein is produced that has NADPH coupled reaction activity. Sequencing indicated that there are 4 DNA mutations in this gene from the published wild-type sequence which also resulted in several amino acid changes (Arg298Ser, Cys302Ser and two silent mutations). NADPH oxidation was not coupled to the oxidation of substrate in this enzyme mutant since there was no apparent sulfoxidation activity of MTS whatsoever. It appears that this is an inactive mutant of the wild type enzyme but interestingly it is a substrate independent NADPH oxidase. Recloning of the original cDNA derived gene, using high fidelity Pfu polymerase provided the wild type rhesus macaque FMO2 gene (in pAAP45) (Appendix B) for

PAGE 69

52 subsequent directed evolution studies. The requirement for one (R) and one (S) FMO sulfoxidation catalyst for DNA shuffling is satisfied by the rhesus macaque FMO2 (pAAP35) and rabbit FMO1 isoform systems (pAAP11) and further investigation into the inactivity of rabbit FMO2 was abandoned. Instead, fine tuning to raise the expression of active rabbit FMO1 was investigated. This would produce more enzyme, allowing scale-up of the biocatalytic reactions and enabling more sensitivity in the screenings. Figure 32 Electron-micrograph evidence of lack of inclusion bodies. Transmission Electron micrographs x 25 000 of A) BL21(DE3), B) BL21(DE3)/pAAP11 rabbit FMO1, C) BL21(DE3)/pAAP15 rabbit FMO2, D) (and insert) E. coli BL21(DE3)/pREP4/pETMW2 overexpressing Orf4-protein into inclusion bodies (used with permission from images by Dr. G. Acker)191. Electron micrography of FMO expressing cells done by Dr. H. Aldrich (Microbiology and Cell Science, UF).

PAGE 70

53 Improving the Expression of Active FMO The antibiotic selection system was chosen based on maximal plasmid retentions. In parallel culture assays, otherwise identical systems displayed varying plasmid retentions. This indicates the use of kanamycin (83% plasmid retention after 20hrs. culture) is preferred over ampicillin (46% plasmid retention after 20hrs. culture) as a selective marker in the expression systems for rhesus macaque FMO2 and rabbit FMO1. Higher plasmid retention provided an apparent 4 fold increase in activity of the cell culture (Table 4: pAAP11 compared to pAAP27). Codon usage varies, with the preferences for members of a degenerate codon set being distinct among different organisms192. Prokaryotes such as Escherichia coli often have different levels of tRNAs than mammals from which genes are derived. There are commercially available strains of Escherichia coli with increased expression of these mammalian favouring tRNAs (argU, ileY and leuW) that might have increased the levels of protein production. Expression of FMO in one such strain (BL21(DE3) codon+, graciously provided by Dr. Michael Thomson) does not seem to increase the levels of active protein. The expression systems for yeast FMO have been improved previously with the aid of folding accessory proteins. Activity of FMO produced by an Escherichia coli expression system was increased (x 4) with the co-expression of the chaperonins GroES and GroEL.139 The expression of the inherent chaperonins of Escherichia coli can also be induced by brief heat shock of cultures at 42 o C193. For the expression systems of rabbit FMO1 and rhesus macaque FMO2 in this study though this does not increase the expression of active protein.

PAGE 71

54 As was mentioned earlier (Chapter 1) fusing peptide fragments to the N-terminal of FMOs has produced heterologuously expressed proteins with added solubility, stability and activity.126,177 Other N-terminal modifications also increase activity of cell culture isolates. Whether this increases levels of active protein or affects the activity directly has not been investigated. The N-terminal His 10 modification of yeast FMO has reportedly increases S-oxygenation activity 20 fold.139 Attempts to modify the various ends of expressed rabbit FMO1 proteins also indicate a preferential effect of adding peptide fragments to the N-terminal side. Cashman and co-workers177 used an expression system to produce human FMO3 protein fused to a maltose binding domain. Proteins were expressed from a system which fused the MalE signal sequence to the N-terminal end of the FMO gene during expression (pMal-c2x). It was reported that the protein was produced as a cytosolic soluble enzyme with activities identical to the wildtype FMO3. The level of active protein produced was much higher and additionally the fusions could be later purified on amylose resin. This clone was graciously provided to us by Dr. John Cashman. Insertion of the rabbit FMO1 in the place of the human gene into this plasmid vector (pAAP240) though did not produce active protein (Table 5). The activity of the FMO3 fusion clone corroborated the results from Cashman (data not shown) indicating this system might work with some but not necessary all FMO enzymes. Addition of multiple histidine amino acids to ends of the rabbit FMO1 sequence had more rewarding results. Although there was no investigation into verifying the actual presence of the His tags in the mature protein, an expression system designed to add a His 6 tag to the N-terminal of rabbit FMO1 (pAAP38) increase the activity of crude

PAGE 72

55 membrane preparations (x 2) (Table 5). The addition of a histidine tag to the C-terminal end of the protein alternatively leads to reduction in the amount of enzyme activity. Table 5. Catalytic activity of Escherichia coli expressed rabbit FMO1 fused with various peptides at either the Nor Cterminal ends. Strain/ Plasmid Promoter, selection, ori Expression Activity Notes XL1Blue/ pAAP24 P tac Amp R pBR322 Low -100Kda band ND (a)(b) expressed as MalE fusion protein (N-terminal) not catalytically active BL21(DE3)/ pAAP38 T7, Kan R PBR322 High 1.3 x 10 -4(a) His 6 tag (N-terminal) BL21(DE3)/ pAAP41 T7, Kan R PBR322 High 9.2 x 10 -6(a) His 6 tag (N-terminal) His6 tag (C-terminal) BL21(DE3)/ pAAP46 T7, Kan R PBR322 His 6 tag (C-terminal) not catalytically active no protein apparent expression was determine from SDS-PAGE analysis of total cell extracts, monitoring for the presence of ~60Kda protein activity (mmole/min/mg protein) was measured as consumption of NADPH in a) MTS coupled oxidation assay using crude membrane protein preparations b) activity of both the cytosolic and membrane associated fractions was assayed. Increasing the density of cell cultures is another strategy to increase the concentration of catalyst. Increasing the cell growth might therefore lead to greater sulfoxidation activity. Cell growth is already reduced due to the induction of FMO production, which poses a load on the transcriptional and translational resources of the cells. Induced cells reach 25% lower cell densities than those uninduced (Figure 33). Lowering the concentration of salt in the culture media increased the cell growth (Figure 33). Higher cell density in low salt medium (TB as opposed to LB) concurrently has increased the biocatalytic activity of each cell culture (x 1.5).

PAGE 73

56 For sample handling of cultures numbers required for library screening 96-microwell plates are commonly used, accommodating up to 1-2mL cultures per well. Growth of Escherichia coli cell cultures in microwell plates is not as rapid as in flask cultures. It is thought that the limited amount of aeration provided by agitation of small volumes in microwells to be the main cause.194 The growth of Escherichia coli cultures in shake flasks (250mL flasks with 100mL culture at 200rpm, 37 o C) reaches stationary phase faster and at higher cell densities than in multiwell plates (1.5mL culture/well at 200rpm, 37 o C) (Figure 33). This necessitates longer incubation times for the growth and protein expression of small cultures. Figure 33. Growth rates of recombinant Escherichia coli cells: A) comparison between cells, induced for FMO expression, grown in high (LB) and low (TB) salt media. B) comparison between growth in shake flasks and microwell plates at 37 o C and 200 rpm. Overall the sulfoxidation activity of a given culture of the Escherichia coli expression system for FMO was increased 14 fold from an initial cloning system and is optimized in low salt media using kanamycin antibiotic selection and with a N-terminal

PAGE 74

57 His 6 -tag. This improves biocatalysis reactions in microwell plates, which are needed for the easy handling of large numbers of samples in library screenings.

PAGE 75

CHAPTER 3 GENERATING LIBRARIES OF MUTANT FMOS Error-Prone PCR; Controlled Random Mutagenesis Error-prone PCR attempts to produce mutants of enzymes by subverting the natural fidelity of Taq polymerase in PCR DNA amplification reactions. Introduction of base pair substitutions by Taq polymerase are in turn translated into changes in the amino acid sequence of the gene product, effectively producing mutant enzymes. Increasing the concentrations of Mg 2+ adding Mn 2+ increasing and unbalancing the concentrations of the four dNTPs, increasing the extension time and the concentration of Taq polymerase all reduce the fidelity of PCR.In fact, the concentration of Mn 2+ in the reaction solution is directly proportional to a loss of Taq polymerase fidelity195. The concentrations of Mn 2+ added can then be used to control and fine tune the frequency of mutation. Caldwell and co-workers have altered the mutagenesis rates of Taq polymerase in increments of as little as 0.5% (mutation/base pair).196 Mutations can be deletions, insertions or base substitutions. In random mutagenesis base substitutions are preferred as all other mutations would result in a shift of the reading frame and an inactive enzyme.197 Specific reaction buffers and concentrations of Mg 2+ can be used to favour this. Within within base substitutions though there is also an undesirable preference in error-prone reactions, a strong bias towards transitions or purine-purine and pyrimidine-pyrimidine changes (specifically AG, TC) over transversions. This preference can be reduced somewhat although not 58

PAGE 76

59 completely by using unbalanced amounts of the dNTPs, specifically increased dCTP and dTTP concentrations.198 Since the effect of Mn 2+ on the mutagenic frequency of Taq is dependant on the composition and length of the template and on the primers used the mutagenesis of each gene must be carefully calibrated first. This can be done by subjecting the wild type gene to mutagenesis reactions with various concentrations of Mn 2+ and sequencing the clones that are produced to identify the desired mutagenic frequency conditions. The Taq polymerase that was used in the mutagenesis experiments was a generous gift from Dr. Michael Thomson. This polymerase had an added N-terminal His 6 tag to aid in its purification which did not seem to affect it in the non-mutagenic PCR reactions (including 5mM Mg 2+ pH 8.0). In the presence of Mn 2+ however this polymerase did not show any activity. It required specific conditions to polymerize DNA in the error-prone reactions, including : 2mM Mg 2+ pH 9.0, 0.01% Triton X-100, 0.1mg/mL BSA and 10mM (NH 4 ) 2 SO 4 Another change of conditions which which gave error-prone product was using 8mM Mg 2+ at pH 8.4. The added His 6 tag of the polymerase might be complexing Mg 2+ or Mn 2+ from solution. Rabbit FMO1 was mutated in an initial set of error-prone PCR experiments following the scheme in Figure 34. The error-prone PCR reactions on the wild-type template rabbit FMO1 was done in several reactions to yield enough DNA for the subsequent cutting and cloning into the expression plasmid (pET26(b+)). These ligated plasmids were then transformed into Escherichia coli cells. From here plasmids were purified in large scale and sequenced while the mutagenized genes were also expressed and tested for their sulfoxidation activity.

PAGE 77

60 Figure 34. Error-prone PCR and cloning into expression systems. Indicated are the possible sources of wild-type gene which can carry over into final mutagenized library.

PAGE 78

61 Ligated products must yield at least ~10 000 distinct clones in Escherichia coli to produce a library covering the total number of possible singly mutated variants (20 x 535). Transformation efficiencies of these libraries using electroporation ranged from 2000-4000 individual clones per reaction and several electroporations were performed to achieve a full library. To calibrate mutagenic frequency (EP-rFMO1-G through K) several different concentrations of Mn 2+ were initially used, ranging up to the highest reported in the literature (0.5mM197,198). There was however noticeably less PCR product at 0.3 and 0.5mM Mn 2+ than in the non-mutagenized reactions, possibly indicating that these concentrations inhibit the activity of the polymerase. After transformation and plating, individual colonies of Escherichia coli should each have a separate clone of rabbit FMO1. Sequencing a handful of these clones from each mutagenesis reaction gives an indication of the mutagenic frequency. To limit the amino acid substitutions to 1 per gene, mutation at the DNA level should be kept to about 0.125-0.3%.84 In a 1600 base pair gene that is about 2-4 base pair substitutions per clone. Sequencing of these clones does not however coincide with an increase in mutagenic frequency with Mn 2+ concentration as reported.198

PAGE 79

62 Table 6. Mutagenic frequency of EP-PCR reactions with various concentrations of Mn 2+ EP-PCR libraries (#C, D, E, F, G) generated from wild type rabbit FMO1. [Mn 2+ ] mM (library #) Clone # bp sequenced Deletions/ Insertions # DNA mutations Non-silent mutations Mutation Frequency (%) 0.00 (G) 1 564 0 1 0 0.18 2 453 0 0 0 0.00 3 632 0 0 0 0.00 4 463 0 0 0 0.00 5 732 0 1 1 0.14 0.06 0.10 (H) 1 536 0 1 1 0.19 2 610 0 1 1 0.16 3 375 0 0 0 0.00 4 701 0 0 0 0.00 5 0.07 0.20 (I) 1 425 0 0 0 0.00 2 463 0 2 1 0.43 3 455 0 1 1 0.22 4 631 0 0 0 0 5 234 1 0.13 0.30 (J) 1 645 0 2 1 0.31 2 423 0 1 0 0.24 3 566 0 3 3 0.53 4 512 0 0 0 0 5 289 0 0 0 0 0.22 0.50 (K) 1 443 0 0 0 0.00 2 623 0 2 2 0.32 3 530 0 0 0 0.00 4 584 0 3 2 0.55 5 433 0 0 0 0 0.17 It is noticeable that within the clones from each Mn 2+ concentration used there were several which did not have any mutations. This was taken to be carry-over contamination of the wild type gene.

PAGE 80

63 Figure 35. Activities of rabbit FMO1 clones (EP-rFMO1-J). NADPH consumption (measured at 340nm) during the sulfoxidation reactions of MTS with crude membrane fractions from (#1-5) cells expressing clones of EP-PCR library of rabbit FMO1 (EP-FMO1-J), cells expressing the wild type rabbit FMO1 and (-)ve: the native Escherichia coli strain. The activity of 5 of these cloned genes indicate (Figure 35) that 3 harbouring no mutation possess wild type activity, the remaining inactive enzymes are all mutants (Figure 35, Table 6). The resulting library of mutants even though they possess on average the correct mutagenic frequency, are likely composed of a subset of wild type and a sub set of mutant genes. This effectively increases the size of the library that must be screened to sample all the possible mutant variants. Such a problem in cloning of mutagenized genes has been reported once previously.199 To remove all the possible sources of wild-type contamination a different cloning strategy was adopted (Figure37). By changing the antibiotic selection the template wild type is no longer a viable plasmid if carried over to the final transformation. This is guaranteed by digestion of the non-methylated template with Dpn I. Another set of mutagenesis reactions using Mn 2+ were cloned into Escherichia coli using this strategy and the resulting mutagenic frequency appears to have a more linear progression with increasing mutagen concentration as expected (Table 7).

PAGE 81

64 Reactions of error-prone PCR using 0.15-0.2mM Mn 2+ cloned into the expression systems were prepared for screening of sulfoxidation activities. Table 7. Mutagenesis and sulfoxidation activities of EP-rFMO1-J library. Clone mutated bp/ bp sequenced Mut Freq.% Amino acids mutated from wt % activity* of wt no FMO (-ve) 0 wt rabbit FMO1 100 EP-rFMO1-J-1 2/271 0.74 2 (G49E,V81A) 5 EP-rFMO1-J-2 1/756 0.13 0 74 EP-rFMO1-J-3 0/652 0.00 0 88 EP-rFMO1-J-4 0/643 0.00 0 72 EP-rFMO1-J-5 2/398 0.50 2 (V46A,L99P) 3 EP-rFMO1-J-6 3/650 0.46 2 (S102R, L149P) ND activity from NADPH consumption assay of MTS sulfoxidation of wild-type rabbit FMO1 was (100%) 5.5 x 10-3 mmole/min/mg membrane protein, ND = assay not done Table 8. Mutagenic frequency of EP-PCR reactions with various concentrations of Mn 2+ EP-PCR libraries (#L, M, N, O, P) generated from wild type rabbit FMO1. [Mn 2+ ] mM (library #) Clone # bp sequenced Deletions/ Insertions # DNA mutations Non-silent mutations Mutation Frequency (%) 0.00 (K) 1 634 0 1 0 0.16 2 750 0 0 0 0.00 3 570 0 0 0 0.00 0.05 0.10 (L) 1 453 0 1 1 0.22 2 676 0 1 1 0.15 3 534 0 0 0 0.00 0.12 0.15 (M) 1 376 0 1 1 0.27 2 658 0 1 1 0.15 3 548 0 1 1 0.19 0.19 0.20 (N) 1 590 0 2 1 0.34 2 436 0 1 0 0.23 3 786 0 3 2 0.39 0.33 0.25 (O) 1 567 0 2 2 0.35 2 578 0 3 3 0.52 3 643 0 1 1 0.16 0.34

PAGE 82

65 Figure 37. Error-prone PCR and cloning into expression system. Strategy to remove wild-type contamination from the final mixtures. nm = non-methylated DNA.

PAGE 83

66 DNA Shuffling and Recombination Mutagenesis Another method to alter genes involves the combination of fragments from different full length sequences into a set of new genetic combinations. This DNA-shuffling method developed by Stemmer85 can be used to combine sequences with few variations (such as the products of error-prone PCR mutagenesis) or sequences of related genes as done in family shuffling.200 Reshuffling in this way allows one to combine beneficial mutations from mutants selected out of mutagenesis screening into one gene clone. The accumulation of beneficial mutations is believe to have an additive character88 allowing one to increase the improvement of the trait selected for. The probability of combining positive mutations by this method would be greater than from an attempt to introduce them sequentially by successive rounds of mutagenesis and selection. In family shuffling two related naturally occuring genes are shuffled with each other. Due to similarity in function and structure between even quite different sequences the exchange of segments is hoped to retain an overall active arrangement.201 Traits that are distinct to the two might be found combined in an active clone. With this strategy we shuffled the sequences of rabbit FMO1 and rhesus macaque FMO2, two isoforms with 56.5% identity in amino acid sequence (Figure 38 & 39) but with different catalytic properties. Rabbit FMO1 has higher activity for the catalysis of sulfoxidation of MTS whereas the monkey enzyme has a greater preference for the oxidation to the (S) sulfoxide (Table 2). It is hoped that DNA shuffling of these genes might produce a highly (S) selective catalyst with high activity.

PAGE 84

67 Figure 38Multiple sequence alignment of Rabbit FMO1 and Rhesus macaque FMO2 genes (DNA) (ClustalW). Identical residues are shown with a black background. These genes have 65% identity (1045/1608bp aligned identically). The FAD binding site is at location 25-42 (GGA GCT GGG GT G / C AG C / T GGC)178 and the putative NADP binding site at 571-588 (GGA ATG GG C / A AA T / G TC T / G GGC).

PAGE 85

68 Figure 39Multiple sequence alignment of Rabbit FMO1 and Rhesus macaque FMO2 proteins (ClustalW). Identical residues are shown with a black background for both proteins. There are 302/535 identical residues (56.5% identity). FAD binding site aa9-14 (GAGVSG)178 and NADP putative binding site aa 191-196 (GMGNSG) are both conserved between the two enzymes.

PAGE 86

69 DNA shuffling is the fragmentation and reassembly of genes into chimeras (Figure 40). There are several procedures which have been used to fragment and recombine full length sequences.85,91,202,203 A method using Dnase I digestion and self-priming reassembly was suggested by Dr. C. Martinez in personal correspondences (Appendix A). It is critical that fragmentation to 1-200bp fragments be complete to avoid carry over of the wild type sequences. The activity of Dnase I enzyme appeared to decreased rapidly with storage time and it was necessary to recalibrate the enzyme before each digestion. Figure 40. General overview of DNA Shuffling. Two or more isolated parent genes are initially digested into random 1-200bp fragments then reassembled using self-primed PCR into a variety of chimeric combinations. Recombination PCR is prone to add mutations. One way to circumvent this is through the use of high fidelity polymerases. Pfu polymerase is about 6 times less likely

PAGE 87

70 to incorporate a error during amplification than is Taq polymerase, with error rates of 1 out of every 770 000 and 1 out of every 125 000 base pairs respectively 204. Reassembly reactions were done with 1:1 mixtures of Pfu and Taq polymerases.202 The random recombination of the gene fragments was checked using a PCR time course (Figure 41). Amplification of the chimeras using all four combination of primers for the ends of both genes is predicted to proceed at similar rates for completely randomly shuffled genes. A parallel increase in the expected 1.6Kbp fragment of amplification was observed from all 4 primer sets. Figure 41. Confirmation of random assembly of gene fragments into chimeras of rabbit FMO1 and rhesus macaque FMO2. DNA shuffled reactions were amplified with a combination of primers A)rabFMO1for, rmFMO2 rev;B)rabFMO1for,rabFMO1rev; C)rmFMO2for,rabFMO1rev; and D)rmFMO2for, rmFMO2rev. PCR were sampled at different increments during the reactions and checked on agarose gel for FMO gene (~1.6Kbp).

PAGE 88

CHAPTER 4 BIOCATALYSIS REACTIONS AND SCREENING The successful directed evolution of enzymes is largely dependent on the screening or selection method employed.94 Although the speed and ease of the screening is important as it limits the size of libraries that can be investigated the sensitivity of the method will determine if improved mutants are detected at all. Each round of mutation produces a range of changes in the property that is being investigated. At the start of this project approximately 40 different enzymes had been improved using directed evolution in a variety of properties; thermostability, activity towards a certain substrate, activity in organic solvents and expression levels being some of them.205 At the time though there were few examples of altering the enantioselectivity of an enzyme. Bornsheuer and co-workers199 had used a powerful selection technique to screen over 100 000 clones of an esterase in the first round of mutagenesis and uncoverred a clone with a change of enantioselectivity from the wild type of 25% ee. Arnold et al.206 have screened for the kinetic resolution of the enzyme hydantoinase with a 20% increase in preference for the hydrolysis of the D enantiomer of the substrate.This resulted after screening of the first library of 10 000 clones. Reetz and co-workers produced a clone of a lipase with a change in relative rates from E = 1.1 to E = 2.1 towards preference for one enantiomer of a substrate from a library of 3000.79 With these few results it is difficult to predict what the minimum accuracy needed would be for the screening to detect improved variation in enantioselectivity from a library of singly mutated rabbit FMO1. 71

PAGE 89

72 The enantioselective oxidation of sulfur is an enzymatic ability which has as yet not been investigated using directed evolution. To probe the variants of an enantioselective sulfoxidation catalyst such as FMO, a method has been developed to produce and present the enzymes (Chapter 3) and use them in oxidation reactions. A tailored biocatalysis reaction must be developed which would be amenable to a rapid and sensitive screen of both the product yields and product stereochemistries. A screening procedure with high sensitivity and efficiency must also be developed. Sulfoxidation Using FMO Biocatalysts The Flavin monooxygenase enzyme have previously been observed to catalyse oxidation of organic sulfides. In fact FMOs oxygenate a wide array of aryl alkyl sulfides with varying stereoselectivities (Table 8). This chemistry was first described for the porcine isoform 1 purified by Ziegler and coworkers.111 Purification of the enzymes from their native source tissue is a labor intensive and lengthy procedure.111 Later, heterologuous expression systems were developed for rabbit FMO1 and FMO264 and rhesus macaque FMO2.65 These produced mammalian enzymes in microbial cells with identical properties to the native forms but with much more ease and speed of preparation.126 The activities of sulfoxidation of several aryl alkyl sulfides by various FMO enzymes were investigated using such expressed enzymes (Table 9).

PAGE 90

73 Table 9. Sulfoxidations of a variety of aryl-alkyl sulfides by purified FMOs (ip) and crude-membrane preparations of heterologously expressed FMOs (cmp). Substrate FMO Isoform e.e.% Enzyme Preparation Ref. pig FMO1 96% (R) ip 111 rabbit FMO1 98% (R) cmp rabbit FMO1 >99% (R) cmp 64 rabbit FMO2 >99% (R) cmp 64 rabbit FMO3 50% (R) cmp 64 rhesus macaque FMO2 30% (S) ip 65 S human FMO3 57% (R) cmp 207 rabbit FMO1 98% (R) cmp 64 rabbit FMO2 91% (R) cmp 64 rabbit FMO3 50% (R) cmp 64 S human FMO3 49% (R) cmp 207 rabbit FMO1 96% (R) cmp 64 rabbit FMO2 86% (R) cmp 64 rabbit FMO3 72% (R) cmp 64 S (CH2)2 human FMO3 75% (R) cmp 207 rabbit FMO1 96% (R) cmp 64 rabbit FMO2 70% (R) cmp 64 S pig FMO1 55% (R) cmp 170 rabbit FMO3 88% (R) cmp 64 S (CH2)3 human FMO3 88% (R) cmp 207 *cmp : crude membrane preparation of heterologuously expressed FMO ip : isolated enzyme .enantioselectivity determined by chiral GC-MS analysis of product sulfoxide. Although suitable for investigation of single reactions between an isoform of FMO and a single substrate, crude membrane preparations (100 000 x g particulates fraction) of heterologuously expressed enzyme do not lend themselves to rapid, multi-sample screening. A less complex strategy to present heterologously expressed enzyme for biocatalytic reactions is from intact, whole cells. Whole cell biotransformations with the E. coli expression systems (BL21(DE3)/pAAP11 and BL21(DE3)/pAAP27) for rabbit FMO1 though show no activity in the sulfoxidation of MTS. The reactions were monitored over 3 days at various

PAGE 91

74 temperatures (rt, 30 o C and 37 o C) for MTSO production by GC analysis. Protein is detectable from SDS-PAGE analysis and is active as microsomal (semi-isolated) preparations, whereas unlysed cells do not have catalytic activity. It should be noted that a similar system with a different sulfoxidation enzyme also did not show any activity in whole cell reactions with this substrate. Whole cell biocatalysis with E. coli (BL21(DE3)/pMM04) overexpressing cyclohexanone monooxygenase produced no sulfoxidation product of MTS after 19 hrs. incubation.208 It was also noted that for whole cell biotransformations in shake flasks there is a disappearance of 50-60% of the starting material (MTS) by the end of a 62hrs assay, although no extractable product (MTSO) is discernible. Since the disappearance follows a very linear rate over 62hr and occurs in the absence of cells it could be that the starting material is escaping perhaps being lost to vapour. Reactions with 100mmole MTS were performed using 100mL unlysed Escherichia coli cultures at 200rpm and at 25, 30, 37 o C. The expression system for rabbit FMO1 based on the pET26b(+) plasmid vector (BL21(DE3)pAAP27) does express active protein. This activity of this protein is not apparent unless the cells are lysed. There is no precedent for whole cell biocatalysis using FMOs in literature. For the sulfoxidation of MTS we have not been able to observed any reaction with various unlysed expression systems for either rabbit FMO1, rabbit FMO2 or rhesus macaque FMO2. In case the access of substrate to the enzyme can be overcome by permeabilization of the outer membrane several known Escherichia coli permeabilization agents have been used. Metal chelators209, DMSO210 and polyethyleneimine211 though

PAGE 92

75 did not seem to improve whole cell biocatalysis. This indicates that cell lysis and the subsequent provision of co-factors seems to be required for observable catalysis. Table 10. Sulfoxidation of an array of alkyl aryl sulfides catalyzed by crude membrane preparations of Escherichia coli expressing rabbit FMO1. Reaction done with 1mole substrate and yields after 20hrs reactions determined by GC. Substrate Yield % S 88% S 64% S 42% S 45% S Br 72% S CH3O 69% S OH npd S (CH2)5 43% S (CH2)9 32% *cmp : crude membrane preparation (100 000g particulate fraction), ip : isolated enzyme npd : no product detectable Crude precipitated membrane preparations of rabbit FMO1 do possess catalytic activity for the oxidation of several alkyl aryl sulfides (Table 9). Methods for cell lysis used during biocatalysis screening must be gentle enough to be non-disruptive to enzyme viability and rapid enough to be used on large numbers of samples simultaneously. Sonication lyses cells by liquid shear and cavitation,

PAGE 93

76 homogenization by pressure lysis, both can be used for active enzyme preparations212 although they cannot easily accommodate fast throughput. Enzymatic lysis using lysozyme or snap-freezing is preferred. Large number of cell cultures (in 96 microwell plates) initially washed of media and resuspended in reaction buffer (with 0.1mM EDTA) can be quickly lysed at.80 o C (15-30min). If supplied with enough co-factor (NADPH), 1.5mL of cell culture expressing rabbit FMO1 prepared in this way can oxidize 1.5mole of MTS in 30hrs to 80% sulfoxide. The addition of protease inhibitors (such as PMSF) does not seem to improve the activity of lysed cells. The biotransformation reaction of MTS using the flavin monooxygenases requires a buffered solution (pH 8-9.5) and a supply of reducing equivalents, either as NADPH (1.1 molar substrate equivalents) or NADP + (0.1 molar substrate equivalents) along with a suitable regeneration system (Figure 42). The rhesus macaque FMO2 and rabbit enzymes have pH optimums of 9.5127 and 8.264 respectively for the oxidation of MTS and n-alkyl amines. Figure 42. Biotransformation of MTS by FMO presented as crude membrane preparations from Escherichia coli based expression systems. A regeneration system reduces the amount of cofactor required per reaction, an important advantage in libraries of 10 000 reactions. A suitable system requires an enzyme to reduce the oxidize NADP + and provide FMO with a constant supply of

PAGE 94

77 reducing equivalents. A regeneration enzyme-substrate often used for this purpose is glucose-6-phosphate, glucose-6-phosphate dehydrogenase. To perform reactions in aqueous media substrate and product solubilities become a factor also. The substrate sulfide in this case is not water soluble and is visible as a layer on the surface of the solutions. To avoid volatilization of the substrate and increase reaction yields, reaction vessels are tightly capped and detergent (Triton X-100 or Brij-35) or cyclodextrins (-hydroxycyclodextrin) can be added. Screening Biocatalysis Reactions Sulfoxidation reactions by FMO mutant will possess possible differences in both yield (rates of reaction) and in enantioselectivity. In taking advantage of the physical and chemical properties of the products to design a screening, a method with the flexibility for application to other systems is desirable. Coupling to other reactions or modification of the reactants or products to produce an easily monitored characteristic could limit expansion of screening which would be too specific for the substrate initially used. There is also a lack of enzymatic or other chemistry that can be used to distinguish between (S) and (R) MTSO. There are chiral shift reagents for 1 H-NMR spectroscopy that can be used to distinguish the enantiomers of MTSO213 and there are also co-crystallization agents developed for the resolution of the two enantiomers. (R) enantiomer of the sulfoxide (MTSO) selectively will crystallize as a 1:1 complex with 9.214 The crystal structure indicates that 1 molecule of 9 is hydrogen bonded through NH to the SO group of the sulfoxide.

PAGE 95

78 NH H CH3 O NO2 NO2 9 Figure 43. Chiral amide, forms crystals with molecules of (R) MTSO (1 molecule amide: 1 molecule sulfoxide).214 Some of the distinguishing properties of the two sulfoxide enantiomers that could be of use in determining enantiomeric excess are: the different directions each rotates plane polarized light and the distinct interaction each has with chiral environments (the principle of chiral GC and HPLC chromatography). Short chain alkyl p-tolyl sulfoxides are water soluble molecules with relatively high specific rotation (in aqueous solution [] D 20 (R)MTSO = 145 o ). For determination of enantiomeric composition of a reaction with optical rotation though, both the rotation of light and the concentration of the substance must be determined. Screening Biotransformations Using UV/VIS and OR Detectors Developing Methodology Short chain alkyl p-tolyl sulfoxides are water soluble molecules with relatively high specific rotation (in aqueous solution [] D 20 (R)MTSO = 145 o ). For determination of enantiomeric composition of a reaction with optical rotation though, both the rotation of light and the concentration of the substance must be determined. The chromophoric molecule p-nitrophenol absorbs in the visible spectrum and can be screened in complex mixtures with sensitivity ( max = 10 000 max = 318nm), it has been incorporated into many screenings of enzymatic reactions194. This molecule has an increased intensity of a K band corresponding to a transition due to a charge

PAGE 96

79 transfer absorption (Figure 44).215 The two auxochromic groups (the OH group being electron donating, the NO 2 group being electron attracting) are parato each other. Similar orientation of auxochromic groups Br/-OCH 3 para to the sulfoxide electron attracting group on a phenyl ring though does not produce such a visible effect (Figure 45) although both are substrates for rabbit FMO1 (Table 9). A colorimetric assay would have been helpful in monitoring the concentration (yield) of sulfoxide product. OH N+O O N+ O O O N+ NH2 O O OCH3 S+ O CH3 Br S+ O CH3 S+ O CH3 CH3 max= 318nm max= 381nm max= 245nm max= 240nm max= 233nm Figure 44. The addition of two auxochromic substituents para to each other and charge transfer resonance forms influence the absorption characteristic of phenyl rings. These analogs do show stronger UV absorbance than MTS (Figure 45) in the region 200-250nm. The stereochemistry of the rabbit FMO1 sulfoxidation reactions when the para substituent is changed has though not been investigated. S CH3 A S CH3O B S Br C Figure 45. Comparative UV/VIS spectra of biotransformation reactions on various substrates. Starting material (s.m.) at the beginning of the reactions and the products (prod.) after 24hrs. biotransformations. Spectral analysis directly from cell lysate reaction mixtures of rabbit FMO1.

PAGE 97

80 The measurement of yield of the sulfoxidation reactions using UV was not that straightforward since there was no wavelength that could be used to monitor the product of the reaction uniquely at which the starting material didnt also possess some absorption. There will be some experimental variation between the amounts of starting material added over the large number of reactions to be screened. Determination of yield in this case requires knowing both the concentrations of unreacted substrate and product at the end of the reactions. By choosing two absorption wavelengths along the spectrums of both compounds (Figures 46) we can solve for the concentrations of each. Many solubilizers which could be used to help disperse the substrate in the reaction solutions (such as cyclodextrins and Triton X-100) produce too high signals in the chosen UV range to be practical. Instead, the non-ionic detergent Brij-35 (a saturated alcohol) was added to reactions and prior to sampling they were diluted with 3 volumes of EtOH. Figure 46. UV spectrum of MTS and MTSO, indicating the distinct absorption maxima; MTSO (233nm) and MTS (254nm). The concentration of MTSO can be determined for each reaction from measurements at both wavelengths and by solving the equation (Appendix B), concentrations of unreacted MTS can be calculated with a similar equation.

PAGE 98

81 Lysed cell suspensions contain a variety of particles and biomolecules. Several strategies can be used to reduce the possible background these might have on both the OR signal and the UV signal of the products. The addition of EtOH to 75% helps precipitate proteins from solution. Samples were also centrifuge and filtered (0.2m) to remove the majority of the suspended molecules. Even with these precautions there was considerable background at both 233nm 254nm (Appendix B) and in the OR signal (Appendix B). Since the growth of the cell cultures was a variable this background varied between samples. The spectra of a solution of cell extract displays a characteristic peak at 280nm, a region where neither the starting material nor product have any significant absorption (Figure 47). Using this wavelength as a handle the background in all the measurements (UV 233 UV 254 and OR) could be calibrated and taken into account for each sample individually. 020000040000060000080000010000001200000140000016000000.10.250.50.751dilution of cell extract Figure 47. Cell extract background in the biotransformations. UV spectra of cell extract in 75% EtOH with a peak at 280nm.Graph indicating the relationship between the amount of cell extract with the signal at 280nm. Optical rotation readings with the OR detector did not appear to be very sensitive and required a higher concentration of product (even through the small 40L flow cell) than was required for UV measurements. In fact the sensitivity of the UV detector is ~4000 x higher than ORD for MTSO (measured at 233nm). To accommodate this

PAGE 99

82 inequality dilutions were done after OR readings to each sample (100 fold dilution) before measurement by UV. Concentrations of MTSO above 0.05mM saturate the signal from the UV detector (Appendix B). Dilutions can be performed automatically using the Gilson 215 autoinjector/liquid handling system although they require extra time and present another source of errors. To increase the signal read by the ORD the reactions can be scaled up. Methods to increase the production of enzymes were discussed in Chapter 2. Another factor that could influence the signal would be change of the solvent environment216. In organic solvents (EtOAc, hexanes, Et 2 O) which could be used to extract the reaction products, the ORD signals are actually lower for non-racemic MTSO (Figure 48). Miscible solvents such as EtOH or MeOH do not affect the signal appreciably (data not shown).

PAGE 100

83 UV Spectrum Optical Rotation Signal Water: A B 01.02.03.04.05.06.0mVolts 00.51.00Minutes Diethyl ether: C D 01.02.03.04.05.06.0mVolts 00.51.00Minutes EtOAc: E F 01.02.03.04.05.06.0mVolts 00.51.00Minutes Hexane: G H 01.02.03.04.05.06.0mVolts 00.51.00Minutes Figure 48. A comparison of the UV and ORD signals of MTS and MTSO in various solvents. (A,C,E,G): UV spectra (from 200 to 300nm) of 0.04mM MTS (solid line) and MTSO (dashed line). (B,D,F,H) ORD chromatogram showing relative intensities of signals of 4mM (R)MTSO in each solvent.

PAGE 101

84 Assessment of UV/ORD Screening for Enantioselectivity The optimized method using OR and UV measurements to screen reactions of MTS sulfoxidation is presented in Figure 49. After biotransformation reactions of lysed cells in aqueous buffer, co-solvent (EtOH) is added, cell debris is removed by filtration and readings are taken. Optical rotation measurements are followed by dilution of the samples and consecutive measurements at different UV wavelengths. The readings are used to calculate the extent of cellular background, the concentration of product and substrate and the optical rotation of the samples. From this the enantiomeric composition of each reaction can be calculated. Figure 49. Overview of the method to screen, using UV and OR measurements for the enantioselectivity of biocatalysis reactions.

PAGE 102

85 Although the individual measurements had slight error attributable to them (appendix B), the combined errors resulting from the dilutions, the measurements, variable growth rates and enzyme expression gave variations of ~30-40% ee. (Figure 50). Screening times including dilution and measurements were about 12-13minutes/sample. There was also a need to periodically (every 20-30 samples) wash the system with acetone and clean the HPLC 0.22m filter to prevent clogging of the lines. A distribution of ee%012345460667278849096ee% (R)occurence B distribution of ee%012345460667278849096ee% (R)occurence Figure 50. Values of % ee. for samples measured using the UV/ORD screen. Distribution of measurements from A) 20 samples of >99% (R) MTSO and B) 20 samples of wild type rabbit FMO1 enzyme catalysed reactions. Screening Biotransformations Using Chiral HPLC Developing Methodology As mentioned earlier another means to determine the enantiomeric composition of the sulfoxidation products is by using chiral GC or HPLC. A chiral HPLC column (Regis Technologies, Whelk-01 (R,R)) was able to resolve the enantiomers of a series of alkyl p-tolyl sulfoxides with surprisingly fast sample runs (all <10min before optimization) (Figure 51). This is promising for adaption of the screening for other substrates and is rapid enough to be used as a screening for large numbers of samples.

PAGE 103

86 Figure 51. Chiral HPLC chromatographs showing separation of a series of racemic alkyl p-tolyl sulfoxides. Using the Whelk-01 (R R) column sample resolutions were each completed in <10min. For the use of chiral HPLC on aqueous biocatalysis reactions of lysed cells, each reaction is first extracted into an organic solvent and the organic layer is sampled automatically using a Gilson 215 liquid handler/autoinjector (Figure 52). Due to the dilution of intracellular co-factor during cell lysis, sulfoxidation reactions must be supplemented with NADPH, used by FMO in the catalytic cycle to regenerate active enzyme. The cost of supplying enough NADPH for the enzyme can be prohibitive with large number of reactions ($1025 /10 000 reactions). An alternative is to add a small amount (0.1 x substrate equivalents) of the less expensive oxidized form (NADP + $40/10 000 reactions) of the co-factor and a regeneration system, such as D-glucose-6-phosphate (G-6-P) and G-6-P dehydrogenase (G-6-Pdh). The endogenous G-6-Pdh present in Escherichia coli though presented enough activity to exclude this from the addition (Figure 53).

PAGE 104

87 Figure 52. An overview of the use of HPLC adapted to determine reaction yields of biocatalysis reactions.

PAGE 105

88 A. 0200400 mV ol t s 345Minutes B. 0200400 mV ol t s 345Minutes C. 0200400 mVolts 345Minutes FMO FMO + NADPH FMO + Figure 53. The chromatograms of MTS sulfoxidation reactions (0.6mole scale ) of lysed rabbit FMO1 expressing cells. A) cells B) cells and 0.6mole NADPH co-factor, C) cells, 0.06mole NADP + and 0.6mole D-glucose-6-phosphate. (R) MTSO ret time = 4.2min, (S) MTSO ret time = 5.0min. Nonenzymatic Racemic Sulfoxidation There is a surprising background of the MTS reactions (Figure 54), an apparent sulfoxidation that is racemic and not attributable to the activity of the FMO. Two peaks are visible in the chromatogram of an incubation of MTS in a reaction without FMO enzyme. These peaks have the same retention times as the (S) and (R) sulfoxide product, are similar in intensities and increase at a reproducible rate with time of reaction. This background could be due to a non-FMO racemic sulfoxidation of the substrate. This degradation is not present in the neat sample and is independent of the presence of the various components of the reactions, including cell fractions, co-factors, regeneration substrate (NADP + ), buffer and detergent. This oxidation was also observed although at slower rates in organic solutions (EtOH, hexanes, EtOAc), free of all other components of the biocatalysis reactions.

PAGE 106

89 A 0.000.200.400.600.801.001.201.401.601.8002040time (hrs)MTSO(mM) 6 0 B y = 0.0086x 0.01570.000.050.100.150.200.250.300.350.400.4502040time (hrs)S-MTSO(mM) 60 Without enz y me (R) MTSO With rabbit FMO1 (S) MTSO Figure 54. Background racemic sulfoxidation of MTS. A) 2mM MTS biocatalysis reaction of wild-type rabbit FMO1, background oxidation to produce the (S) enantiomer of product stops when all substrate is depleted. B) production of the (S) sulfoxide in the presence of FMO and in the absence of FMO. The initial screens of the expression systems for sulfoxidation activity (measured using chiral GC (Appendix B)) did not indicate any background oxidation, although these biocatalyses used a different stock of substrate (different lot#). The possibility that a contaminant was inherent in the substrate (purchased from Aldrich 99% purity) was tested by using a separate and freshly purchased stock of MTS was tested (from a different lot#). It also indicated oxidation decomposition. The substrates did not indicate any contaminant from analysis using GC-MS. Although it is still possible that it arose from impurities in the substrates the true source of this oxidation is still unknown. Several attempts to prevent the racemic oxidation of substrate using various agents were unsuccessful (Table 10).

PAGE 107

90 Table 11. Various strategies to remove background racemic sulfoxidation of MTS. Reagents were added (10mM ea.) to cell extracts and substrate and run on a Whelk-01(RR) chiral column to observe the extend of sulfoxide produced. No effect = identical to sample with no treatment, Interference in chromatogram = peaks in chromatogram obscure the product peaks. Method/Reagent Action Effect Adjusting pH (3-8) Reduce [OH ] No effect Ascorbic acid Antioxidant No effect EDTA Chelate metals that might be involved in redox reactions No effect D-sorbitol Antioxidant No effect 2,2-dipyridyl (bpy) Fe +++ chelator Interference in chromatogram 1,10-phenanthroline Fe +++ chelator Interference in chromatogram Biphenyl sulfide Compete for sulfur oxidation No effect Chelex beads Chelate metal ions No effect Inert atm. (Arg) Reduce O 2 in solution No effect Catalase Scavenge H 2 O 2 No effect DTT Antioxidant Interference in chromatogram 8-hydroquinoline Antioxidant Interference in chromatogram A background oxidation with a predictable and reproducible rate is not a hindrance to the screening and can be accounted and subtracted from the % ee. measurements (Figure 56). The screening of reactions using the chiral HPLC column was relatively rapid (<6min/sample, ie.~250 samples/day), inexpensive (Table 11) and reproducible (Figure 55).

PAGE 108

91 A distribution of ee%0123456786270788694102ee% (R)occurence B distribution of ee%012345662667074788286909498102106ee% (R)occurence Figure 55 Values of % ee. for samples measured using a chiral HPLC column. Distribution of measurements from A) 20 samples of >99% (R) MTSO and B) 20 samples of reactions of wild type rabbit FMO1 enzyme catalysed reactions. Table 12. The cost of screening a library of singly mutated FMO (10 000clones) for sulfoxidation activity (enantioselectivities and yields). Reagents Wt./vol. Cost (USD) Supplier Biocatalyst expression Tryptone 180g $15 Fisher Yeast Extract 90g $10 Fisher NaCl 5g negligible Fisher Kanamycin 0.72g negligible Fisher IPTG 3.57g $90 Fisher Biocatalysis Reactions and Screening Methyl p-tolyl sulfide 0.83g/0.808mL $20 Sigma-Aldrich Hexanes (HPLC grade) 45L $225 on site EtOH (100%) 45L $20 on site EtOAc 6L $30 on site NADP + 0.472g $40 Biocatalytics Glucose-6-phosphate 2.018g $40 Sigma-Aldrich Reaction buffer chemicals negligible Total < $500 negligible < $5.00

PAGE 109

92 Figure 56. Determination of enantiomeric composition of reactions using a chiral HPLC column. Each plate of 96 reactions includes three standards that are used to calibrate and account for the background of each sample.

PAGE 110

CHAPTER 5 DIRECTED EVOLUTION OF FMOS TOWARDS EFFICIENT (S)-SULFOXIDATION CATALYSTS Results of Screening Assays The directed evolution of an FMO enzyme to invert the stereoselectivity of sulfoxidation can be attempted through various experimental paths. Starting with rabbit FMO1, an enzyme selective for the sulfoxidation of MTS to the (R) sulfoxide, a strategy of single mutations followed by screening can explore the sequences within close similarity for an enzyme preferring the (S) sulfoxide. One such singly mutated clone library was produced and screened for sulfoxidation stereochemistry using a chiral HPLC column. Another approach involves creating chimeras of rabbit FMO1 with its closely related isoform from rhesus macaque. The rhesus macaque enzyme has less activity towards the sulfoxidation of MTS but possess in contrast stereoselectivity of making the (S) sulfoxide. It is hoped that an (S) stereoselective enzyme will result that has the higher activity attributable to rabbit FMO1. Both types of clone libraries were created and screened. Each enzyme gene was expressed from and maintained in an Escherichia coli expression plasmid vector. The interesting clones that came from the mutagenesis and screening were sequenced from this vector (Figure 57) to characterize the responsible mutations. 93

PAGE 111

94 Figure 57. Sequencing the mutant rabbit FMO1 genes from the EP-PCR generated library involves 3 primers flanking and internal to the gene itself. A library of 1323 chimeras of rabbit FMO1 and rhesus macaque FMO2 were screened to sample the DNA shuffling approach. Of these only 2.4% appeared to have any activity at all for this substrate. Their activity was also stereochemically uninteresting as it was identical to wild type (Figure 63). It is possible that since both enzyme genes differ by 560 different base pairs from each other the incompatibility of the majority of chimera assemblages is too great to be overcome with the screening numbers of this study. If only 2.4% of the clones are active, the chances of finding any activity of interest are greatly reduced. Initial screening of 4238 clones of the single mutation library of rabbit FMO1 indicates that most variants are either inactive (producing little to no MTSO) or are active and have similar preference as the wild type for producing the (R) sulfoxide (Figure 62). Within the variability of the screening an arbitrary cut off of under 50%ee (R) MTSO and over 50% yield of product sulfoxide culled 13 interesting clones (Figure 59 & 62). To more carefully screen stereoselectivites each clone was studied in triplicate using the same screening assay. Although yields of sulfoxidation were comparable to wild type ( >80% wild type activity, data not shown) the stereoselectivities of all but one clone were

PAGE 112

95 not any different (~97% ee (R) MTSO). One clone though (176E70), had reproducibly produced ~70% ee (R) sulfoxide (Figure 58 & 59). This is a decrease in enantioselectivity from the wild type of ~27% ee. From its sequence this appears to the a result of one non-silent mutation (Appendix B) of serine at amino acid position 227 into threonine. A) 0200400 mVolts 3.04.05.0Minutes phenanthrene R-mtso B) 0 2 00 4 00 mVo l ts 3.04.05.0Min utes phenanthrene R-mtso C) 0 2 00 4 00 3.04.05.0 Minutes phenanthrene R-mtso Figure 58. The activity of C) clone 176E70, compared to B) the wild type activity (rabbit FMO1) and the A) background air oxidation. Reactions were stopped well before completion (12hrs. reaction times) and chromatograms indicate the internal std. peak (~3.5min), the (R) (~4.3min) and (S) enantiomers of the MTSO product. 020406080100120012345678910111213clones(R)% ee MTSO Figure 59. Rescreening of 13 clones of rabbit FMO1 for stereoselectivities. Clones screened in triplicate. 1) mutant clone176E70, 2) 172B90, 3) 170AG67, 4) 170F72, 5) 170C81, 6) 170AP5, 7) 170H87, 8) 170R66, 9) 170Z3, 10) 170T41, 11) 170AI70, 12) 170O62, 13) 170Q72. Several mutant clones had less (S) MTSO at the end of the reactions than was apparent from the wild type FMO1 oxidations. They also produced more (R) MTSO than

PAGE 113

96 the wild type. Since wild type enzyme has ~99% ee selectivity for producing the (R) enantiomer, these clones must have a higher activity but at least the same enantioselectivity as rabbit FMO1. Since the air oxidation of MTS has a linear rate (0.0172mM/hr MTS oxidation) and is assumed in these mutant to produce all the (S) MTSO, we can use the (S) sulfoxide concentration to predict when the enzymatic reactions were completed. In the wild type reactions it takes about 32.2hrs. to completely deplete a solution of 2mM substrate. Substrate is oxidized by the enzyme and also self-oxidizes in air. When screened again in triplicate, 4 out of 5 of these faster enzymes clones were calculated to have depleted all substrate in less than ~20hrs. (Figure 60 & 61). All mutant enzymes produced ~100% ee of the (R) sulfoxide indicating that they have the same stereoselectivity as wild type but have improved activities. 170 D62170 D96170 C34170 AA21170 G47051015202530350123456clonesreaction time (hrs) Figure 60. Extrapolated times required for each enzyme mutant to oxidize 2mM of MTS. 1) mutant 170D62, 2) 170D96, 3) 176C34, 4) 170AA21, 5) 170G47.

PAGE 114

97 A) 025.050.0 mV ol ts 3.604.50Minutes phenanthrene R-mtso B) 176 E 1 : 1 0200400 mVo l ts 3.04.05.0Minutes phenanthrene R-mtso C) 0200 4 00 mVo l t s 3.04.05.0 Minutes phenanthrene R-mtso S-mtso D) 176 E 4 : 4 0 2 00 4 00 3.04.05.0Minutes phenanthrene R-mtso S-mtso E) 0 2 00 4 00 3.04.05.0Minutes phenanthrene R-mtso Figure 61. The activities of 4 clones with higher activities than the wild type: B) mutant 170D62, C) 170D96, D) 176C34, E) 170AA21 as compared to A) the background sulfoxidation in air. Reactions were taken to completion (35hrs.) and chromatograms indicate the internal std. peak (~3.5min), the (R) (~4.3min) and (S) enantiomers of the MTSO product. Improved perceived activities can be due to several factors, such as faster reaction rates for the enzymes or more active catalyst available from the cells. Arnold et al.217 postulate that increased activity observed in the directed evolution of horseradish peroxidase to be due to mutations which assisted proteins to fold more efficiently.217 This also was visualized as an increase in activity. The 4 faster clones have been sequenced and the mutations have been identified as (mutant 170D62) K164N and M408I, (mutant 170D96) Y428H, (mutant 176C34) I1138A and Y81F and (mutant 170AA21) N451D. The stability and reproducibility of the screening method over many thousands of samples might drift. The fear was that particulate debris would carry over into the HPLC system from the extractions of the reactions. Clogging of the 2m filter would have the

PAGE 115

98 deleterious effect of increasing the retention times and broadening the peaks in the chromatograms. To help prevent this the system was purged with acetone after every 10-20 plates. In each multiwell plate there were sample standards of 2mM (R) MTSO used to calibrate the screening for the concentration of product. Over the course of the screening of the libraries (~60 multiwell plates) the readings of these standards did not change significantly, peak areas and retention times remained relatively constant. This is an indication that the system is stable over many sample runs.

PAGE 116

-100.00%-50.00%0.00%50.00%100.00%150.00%-0.5000.0000.5001.0001.5002.0002.500total product MTSO yield (mM)% ee (R) 99 Figure 62. Sulfoxidation stereochemistries and yields catalyzed by ~4000 various singly mutated clones of rabbit FMO1 (library EP-PCR-FMO1-S+T). (2mM substrate concentration, reaction times of 9-26hrs, reactions analyzed using chiral HPLC).

PAGE 117

-250.00%-200.00%-150.00%-100.00%-50.00%0.00%50.00%100.00%150.00%200.00%-0.5000.0000.5001.0001.5002.000total product MTSO yield (mM) ee% (R) 100 Figure 63. Sulfoxidation stereochemistries and yields catalyzed by ~1100 various chimeras of rabbit FMO1 and Rhesus macaque FMO2 (library Sh-FMO1/2-C). (2mM substrate concentrations, reaction times of 9-12hrs, reactions analyzed using chiral HPLC).

PAGE 118

101 Conclusions and Future Work This work has attempted to use directed evolution to improve the sulfur oxidation activity of an FMO enzyme towards formation of the (S) sulfoxide. Two strategies were investigated; the incremental improvement of this trait by single mutation and screening from an (R) specific enzyme and the shuffling of FMO genes to combine high activity and the desired stereochemistry into a chimeric progeny. The screening methodology initially used involved optical rotation and UV measurements using flow cell detectors. The interference from FMO reactions in lysed cell biocatalysis was compounded from both detectors and resulted in a fairly high error in % ee. determinations. It is possible that using enzyme systems that can accomplish reactions using whole cells could reduce this interferences from cellular debris. UV-OR screening might then be applicable to screening large libraries of catalysts for stereoselectivity changes. Instead a simple chiral HPLC screen was used to screen the first generation of mutant FMO enzymes for altered enantioselectivity. 200-250 clones could be screened a day economically and reproducibly and further screening can be done for stereoselectivities using other substrates (such as other alkyl p-tolyl sulfides). Several mutations appear to increase the activity of FMO in the biocatalysis reactions of sulfide oxidations used in these tests. The method in which activity is increased is not known and could be due to many factors, including improving the ability of enzyme to fold into an active form or increasing the expression levels of the proteins. The substitution of an asparagines at position 451 with an aspartate residue in mutant 170E70 appears to be responsible for the increase of ~30% ee. (S) MTSO formation. From the active site model postulated by Cashman126 and Rettie169 this is

PAGE 119

102 predicted to affect a decrease in the accessibility of binding pocket A (Figure ?) to the p-tolyl rind substituent of the MTS substrate. The lack of any structural data as yet on the FMO enzymes hampers analysis of these individual mutations. Ziegler has postulated from comparison of primary sequences that FMOs are members of the flavocytochrome c dehydrogenase (41.7.1.1.1.1) subfamily of flavoproteins.166 There are several NADPH-dependent oxidases in this family that catalyse the two electron reduction of dioxygen and NADPH-dependent reductases that catalyse the two electron reduction of disulfides. Using the crystal structure of a member of this family, the glutathione reductase of Escherichia coli, as a template, Zielger substituted the sequence for a FMO enzyme (human FMO3) and predicted the structure in Figure 64 (reprinted with permission from the author). The regions of the dimer interface correspond to highly conserved sequence among FMO isoforms and the predicted substrate cleft incorporates the amino acid residues 397-431. The substitution pattern observed for the rabbit FMO1 variant mutants might provide some added information to future structural assessments of the enzymes. Figure 64. Postulated166 tertiary ribbon structure of a dimer of human FMO3. The dark space filling structures represent NADP + co-factors and the light space-filling models represent the bound FAD co-factors.

PAGE 120

APPENDIX A EXPERIMENTAL PROCEDURES Reagents and Supplies Taq DNA polymerase (both the native form and with a N-terminal 6 histidine tag) was supplied to us by Michael Thomson (Department of Anatomy and Cell Science, UF). Pfu polymerase was expressed in vivo by BL21(DE3)/pET28-Pfu and purified in our labs (refer to procedure later in this Appendix). Organic compounds and most reagents were supplied by Sigma-Aldrich and the Fisher Scientific company. Restriction endonucleases, DNA modifying enzymes and DNA Ligase were purchased from New England Biolabs. DNAse I was supplied by Roche Diagnostics, GmbH. Molecular biology grade agarose was procured from Shelton Scientific, Inc. His-Binding nickel column was purchased from Novagen. Media and buffers prepared as described in Sambrook, et al. (1989) except TYGP media (0.2% Bacto-Tryptone, 1% Bacto-Yeast Extract, 1% Glycerol, 1% Na 2 HPO 4 7H 2 O). Antibiotics and supplements were used in concentrations: kanamycin (40g/mL), ampicillin (200g /mL) and X-Gal (1.6mg in 40L DMSO plated on a 90mm petri plate). The PCR Clean-up Kit was purchased from QIAGEN, the WIZARD Plus MiniPrep kit was purchased from Promega. Epicurian Coli BL21-Codon Plus cells were provided by Stratagene. pCR2.1 TOPO PCR cloning vector kit was supplied by Invitrogen. 103

PAGE 121

104 The UV-VIS flowcell detector used in these studies was a Gilson model 618 and the optical rotations of samples were measured using a Shodex flow cell OR-2 OR Detector with a 40L flow cell volume. Purification of Pfu Polymerase Inoculated an overnight culture of BL21(DE3)/pET28-Pfu (3mL LB/kanamycin) into 500mL LB/kanamycin and grew at 37 o C, 200rpm until late lag phase (OD 600 ~ 0.5). Induced with 840mM IPTG (595L) and incubated ~16-18hrs. at 37 o C, 200rpm. Pelleted (8000xg, 10min, 4 o C) and washed cells with DTE (40mL, 50mM TrisCl pH 7.9, 50mM dextrose, 1mM EDTA). Lysed cells with Lysis Buffer (20mL, 0.1% Triton X-100, 0.4% lysozyme, 0.5M NaCl, 20mM TrisCl pH 7.9, 2mM imidazole) by shaking at rt. 45min. Denatured proteins at 80 o C 40min, spun out debris (8000xg, 10min, 4 o C), sonicated 8 x 10secs and re-peletted debris. Loaded S/N onto a His-Binding nickel column (which was washed with 6x bed volume of deIH 2 O, Binding Buffer (0.5M NaCl, 20mM TrisCl pH 7.9, 2mM imidazole), Charging Buffer (20mM nickel sulfate) and Binding Buffer). The column was subsequently eluted with 6x bed volume of Binding Buffer then Wash Buffer (0.5M NaCl, 20mM TrisCl pH 7.9, 60mM imidazole). The protein was eluted ~0.2mL/min with Elution Buffer (0.5mM NaCl, 20mM TrisCl pH 7.9, 1M imidazole) and fractions collected and analyzed with a conductivity meter. Pfu eluted after ~ 60 bed volumes (5-6hrs.) and the fractions (~30mL) were concentrated using the Amicon Filtration System to 2mL. Fractions were analyzed for protein by SDS-PAGE and for activity by amplification of a test template in parallel with Taq polymerase of known concentration (5U/L). The fidelity of the enzyme as checked by sequencing of amplified products was 0 mutations/ 1476bp. Final concentration ~1.5U/L, final yield 2700U.

PAGE 122

105 Figure 65 Pfu polymerase (prepared) vs. commercial (Fisher Scientific) Taq polymerase amplification. 1) Hae III-X174 2) Hind III-DNA, 3) Pfu polymerase (1L) amplified rabFMO1, 4) 0.5U (1L) Taq amplified rabFMO1, (rabFMO1 ~ 1600Kbp) Error-Prone PCR of FMO Error prone PCR can be tuned to create libraries of mutants with on average 1 amino acid mutation per gene. For a 535 amino acid gene that gives a library size of 535 x 20 = 10700. A single amino acid change per gene is achieved with a mutation rate of the DNA of 1-3bp substitutions per 1000bp gene84. For FMO (1600bp) this leads to a desired mutation of about 1.5-4.5bp per mutant gene. Mutation rates of 0.25 to 20bp substitutions per 1000bp can be achieved by varying the concentration of Mn 2+ 197,198,218 and mutagenic frequency can be determined by sequencing. Set up the following PCR reaction in 100L (final volume, with Mn 2+ and Taq polymerase (Taq polymerase without an N-terminal 6 histidine tag): 10mM TrisCl pH 9.0, 2mM MgSO 4 7H 2 O, 0.01mg BSA, 10mM KCl, 10mM (NH 4 ) 2 SO 4 0.1% triton X-100, 0.2mM of each dATP and dGTP, 1mM each of dTTP and dCTP, 16fmole of template gene, 15-20pmole of each primer. In a PCR thermocycler heated reaction for

PAGE 123

106 10min at 94 o C then added 1.5mM MnCl 2 and 5-10U Taq polymerase. The thermal cycler repeated the following steps for 25 cycles: 1min at 94 o C, 2min at 56 o C, 3min at 72 o C. This was followed by 7min at 72 o C to allow full extension and then the reaction was stored at 4 o C until further use. DNA Shuffing85,91,201,203,208,219,220 Shuffling of two genes together was done by digesting 1g of each gene to approximately 100bp fragments and reassembling them with self-primed PCR amplification using high-fidelity polymerases. Fragmentation of Genes A mixture of 1g of each gene fragment, 3.3mM TrisCl pH 7.5, 10mM MgCl 2 0.05% BSA and 0.10U DNase I (RNase free) in 100L was incubated at rt for 20min. The reaction was stopped with 20L of 180mM EDTA, 0.6% SDS and stored on ice. The reaction can be checked by loading onto a 2.5% agarose gel and comparing it to a know standard of fragment-sizes ~100bp (the activity of the DNase I varies with time in storage, often it is necessary to calibrate the reaction, adjusting time and concentration of DNase I to achieve ~100bp fragment sizes). The DNA was then cleaned with the PCR-clean up kit (QIAGEN).

PAGE 124

107 Figure 66. Fragmentation of the FMO gene with DNase I run on a 2.5% agarose gel. Lane 1) undigested 1:1 mixture of Rabbit FMO1 (1.6Kbp) and Rhesus macaque FMO2 (1.6Kbp) 2) after digestion with DNase I showing a smear of fragments ~1-200bp in size. Initial reassembly with high fidelity Pfu polymerase (fidelity ~ 1.3 in 10 6 bp errors204) is followed by amplification using a long PCR mixture of Pfu/Taq (fidelity is 3.9-5.2 errors in 10 6 bp204). Reassembly of DNA Fragments The self-primed re-assembly reaction was set up in 20L: 1g of DNA fragments, 25mM TrisCl pH 9.0, 5mM MgSO 4 7H 2 O, 0.025mg BSA, 25mM KCl, 25mM (NH 4 ) 2 SO 4 0.25% triton X-100, 10mM of each of dATP, dCTP, dGTP and dTTP and incubated at 96 o C for 3min in a thermal cycler. 2.5U of Pfu were added and the thermal ran the following steps for 40 cycles: 94 o C for 1min, 55 o C for 1min and 72 o C for 1min + 5sec/cycle. The reaction was left to finish elongation for 7min at 72 o C and stored at 4 o C. Amplification of the Shuffled Genes The following 100L PCR reaction was set up: 10mM TrisCl pH 9.0, 2mM MgSO 4 7H 2 O, 0.01mg BSA, 10mM KCl, 10mM (NH 4 ) 2 SO 4 0.1% triton X-100, 10mM

PAGE 125

108 each of dATP, dGTP, dCTP and dTTP, 15pmole of each primer (both sets of forward and reverse primers for each gene) and allowed to heat in a thermal cycler for 2min at 96 o C. The thermal cycler repeated 10 times the cycle: 30secs at 94 o C, 30secs at 55 o C and 72 o C for 45secs. This was followed with 14 repeats of the cycle: 30secs at 94 o C, 30secs at 55 o C and 72 o C for 45secs + 20secs/cycle. The reactions were allowed to elongate for 7min at 72 o C and stored at 4 o C. They were subsequently cloned into either pre-cloning or expression plasmids (note that they will need a further incubation at 72 o C for 15min with 5U Taq polymerase to be ligated into the T overhang vectors such as pCR2.1). Figure 67. Reassembled and amplified FMO gene chimeras loaded on a 0.8% agarose gel. Lane 1) amplified Rabbit FMO1 2) amplified chimera of Rabbit FMO1 and Rhesus macaque FMO2 (all major bands in lanes 1 & 2 are ~1.5Kbp) Assay for FMO Activity; Oxidation of Thiobenzamide136,221 An assay used for the characterization of active FMO protein from E. coli expression systems. Purified rabbit FMO1 and 2 are both reported to oxidize thiobenzamide.221

PAGE 126

109 Cultures of Escherichia coli induced for the expression of enzyme (100mL) are pelleted (1000 x g, 15min, 4-10 o C) and washed with wash buffer (50mM dextrose, 25 mM TrisCl (pH 8.0), 10 mM EDTA) resuspended in lysis buffer (50mM TrisCl (pH 8.0), 1 mM EDTA) and freeze lysed at o C, 15min. The resuspended cell debris is homogeneized at 10 000 psi in a French Press twice at 4 o C. The preparation is kept on ice in all subsequent steps to prevent denaturation of enzyme. Cell debris is spun off (5000 x g, 20min, 4 o C) and the supernatant is spun (100 000 x g, 30min, 4 o C) to pellet the remaining membrane-fractions. Concentration of protein in suspension can be determined using the Bradford ragents.222 0.5mg of a fresh suspension of E. coli crude membrane preparation of proteins, stored on ice was added to a 1mL solution of 0.2 mM NADP + 2mM glucose-6-phosphate, 0.008U/mL glucose-6-phosphate dehydrogenase, 50mM TrisCl pH 8.0, 7mM MgCl 2 and 1mM thiobenzamide (from a 20mM solution in EtOH). Once the thiobenzamide was added the solution was mixed gently and briefly centrifuged. The OD 370 of the S/N was monitored over 30min. The oxidized product has 370 = 1.93 AUmM -1 cm -1 Crude Membrane Preparation Activity Assay; Oxidation of Methimazole and MTS A pertinent activity assay for this work is determining if the engineered Escherichia coli cells express activity in the oxidation of MTS. Spectrophotometric assay for the consumption of the cofactor NADPH is a general assay for the activity of oxidative enzymes (those that require reducing equivalents of NADPH for regeneration). Crude membrane preparation of cells are prepared as in the thiobenzamide activity assay.

PAGE 127

110 0.2mg of a fresh suspension of E. coli crude membrane preparation of proteins, stored on ice was added to a 1mL solution of 0.5 moleNADPH, 50mM TrisCl pH 8.0, 7mM MgCl 2 0.1% Tween 40 and 1mole substrate (MTS or methimazole from a 20mM solution in EtOH). Once the substrate was added the solution was mixed gently and briefly centrifuged. The OD 340 of the S/N was monitored over 30min. The oxidization of NADPH ( 340 = 6300 AUmM -1 cm -1 ) is coupled to the oxidation of substrate. There are a few important control reactions to be able to decipher anything from this test: the consumption of NADPH by cells expressing enzyme (without substrate being added) and cells not containing the expression system (with substrate being added). Synthesis of Starting material and Sulfoxide Standards Aryl Alkyl Sulfides Alkyl-aryl sulfides substrates were synthesized using a modification of the general procedure by Rettie et al.170. Briefly, p-thiocresol (10mmole) was added to MeOH (10mL) and K 2 CO 3 (1.0g) and stirred until dissolved (5min at rt). The solution was put on ice and the corresponding alkyl-iodide (12.5mmole) was added dropwise. The solution was allowed to warm to rt with stirring overnight. The reactions were monitored by TLC (hexane:EtOAc 1:1 mobile phase) for completion. Alkylations were stopped with 10% NaOH (40mL) and the solutions were extracted with CH 2 Cl 2 (3 x 10mL) and the combined organic layers were dried over MgSO 4 washed with brine, filtered and concentrated via rotary evaporation. Silica chromatography (eluant hexane:EtOAc 1:1) afforded the following products: (ethyl p-tolyl sulfide) (80% yield) GC-MS m/e 153 (M+H) + ; 1 H-NMR (300MHz, CDCl 3 ): 7.25 (2H, d, J = 8.0Hz), 7.10 (2H, d, J = 8.0Hz), 2.90 (2H, q, J = 7.4Hz), 2.32 (3H, s), 1.28 (3H, t, J = 7.4Hz). (butyl p-tolyl sulfide) (92%

PAGE 128

111 yield) GC-MS m/e 181 (M+H) + ; 1 H-NMR (300MHz, CDCl 3 ): 7.27 (2H, d, J = 8.1Hz), 7.10 (2H, d, J = 8.1Hz), 2.93 (2H, t, J = 7.2Hz), 2.36 (3H, s), 1.62 (2H, sept., J = 7.2Hz), 1.43 (2H, sept., J = 7.2Hz), 0.92 (3H, t, J = 7.2Hz). (hexyl p-tolyl sulfide) (89% yield) GC-MS m/e 208 (M+H) + ; 1 H-NMR (300MHz, CDCl 3 ): 7.25 (2H, d, J = 8.0Hz), 7.08 (2H, d, J = 8.0Hz), 2.87 (3H, t, J = 7.1Hz), 1.88 (2H, t, J = 7.1), 2.3 (3H, s), 1.62 (2H, sept., J = 7.1Hz), 1.40 (2H, sept. J = 7.1Hz), 1.27 (4H, sept., J = 7.1Hz), 1.08 (2H, t, J = 7.1). (decyl p-tolyl sulfide) (95% yield) GC-MS m/e 265 (M+H) + ; 1 H-NMR (300MHz, CDCl 3 ): 7.24 (2H, d, J = 7.9Hz), 7.10 (2H, d, J = 7.9Hz), 2.85 (3H, t, J = 6.8Hz), 2.35 (3H, s), 1.62 (2H, t, J = 6.8Hz), 1.17 (18H, s), 0.88 (3H, t, J = 6.8). Racemic Methyl p-tolyl Sulfoxide (MTSO) To a cold (ice bath) stirred solution of NaIO 4 (684mg, 3.2mmole) in deionized water (8mL) was added methyl p-tolyl sulfide (MTS) (448mg, 3.2mmole). Solution was stirred 30min at 0 o C and let come to rt 16hrs.. The reaction was extracted with EtOAc (3x 30mL) and the combined organic phases were washed with brine (10mL), dried over anhydrous MgSO 4 filtered and concentrated via rotary evaporation. Silica chromatography (eluent EtOAc) afforded 479mg (96%) of methyl p-tolyl sulfoxide, clear, slightly oily crystals: GC-MS m/e 154 (M+H) + ; 1 H-NMR (300MHz, CDCl 3 ) : 7.55 (2H, d, J = 7.9Hz), 7.35 (2H, d, J = 7.9Hz), 2.72 (3H, s), 2.31 (3H, s), Alkyl p-tolyl Sulfoxides Other n-alkyl p-tolyl sulfoxides (ethyl-, n-butyl-, n-hexyland n-decyl-) were synthesized with the same method as MTSO using the corresponding n-alkyl p-tolyl sulfides.

PAGE 129

112 PCR Primers

PAGE 130

113 Strains223 Various Escherichia coli strains were used for the pre-cloning (JM109, XL1Blue, TOP10) and expression (XL1Blu, BL21(DE3)) of the FMO genes and proteins. Table 13. The strains of Escherichia coli used in this study and their genotypes E. coli strain Genotype XL1-Blue TOP10 BL21(DE3) JM109 F ::Tn 10 proA + B + lacI q (lacZ)M15/recA1 endA1 gyrA96 (NaI r ) thi hsdR17 (r K m K + ) glnV44 relA1 lac F mcrA (mrr-hsdRMS-mcrBC) 80lacZlacX74 recA1 deoR araD139(ara-leu)7697 galU galK rpsL (Str R ) endA1 nupG F ompT gal [dcm][Ion] hsdS B (r B m B ; an E. coli B strain) with DE3, a prophage carrying the T7 RNA polymerase gene F traD36 lacIq(lacZ)M15 pro A+B+/e14(McrA-) (lac-proAB) thi gyrA96 (NaIr) recA1 endA1 hsdR17 (rK mK +) supE44 relA1 The yeast strains used in this work for the expression of FMO proteins include InvScI (purchased from Invitrogen) and 15C (provided by Dr. A. Buchman).

PAGE 131

114 Table 14. The strains of Saccharomyces cerevisiae used in this study and their genotypes S. cerevisiae strain Genotype 15-C InvSc I MATa, leu2, ura3-52, trp1, his4-80, pep4-3 MATa/MAT, leu2, ura3-52, trp1, his3 Recombinant DNA Techniques Procedures for the manipulation of DNA was carried out essentially as is described by Sambrook et al.224. PCR amplification was performed on a Perkin Elmer GeneAmp PCR System 2400 using Pfu and Taq polymerase graciously provided by Dr. Mike Thomson and prepared in house respectively (see above). Genes and plasmids were cut with restriction enzymes for cloning manipulations, run on agarose gels and fragments purified using either methods described in Sambrook et al.224, the DNA Prep-A-Gene kit from Biorad or by using low melt agarose gels.224 Restriction enzyme and DNA modifying enzyme (klenow fragment, Dnase I, mung bean exonuclease and shrimp alkaline phosphatase) reactions were carried out as described in the literature from the supplier, New England Biolads. Mini-preparations of plasmid DNA were performed using alkali lysis, phenol extraction and EtOH or i PrOH precipitation.224 Large scale preparations of plasmids for sequencing and transformations were done with the CsCl density gradient ultracentrifugation in the presence of ethidium bromide or by means of Clontech or QIAGEN midiprep kits. DNA sequencing was done using the Big Dye Terminator reactions and chromatography of samples was run at the DNA Sequencing Core Facility of the University of Florida. Oligonucleotide primers required for the PCR reactions were obtained from Gemini Biotechnologies.

PAGE 132

115 HPLC Analysis Chiral HPLC chromatography can be used for the resolution of the enantiomers of MTSO. Sample from 96-well microwell plates were manipulated using a Gilson 215 Liquid Handler and injected into a Gilson 819 Injector fitted with a 20L injection loop. A Gilson HPLC pump assembly fed samples through a Regis Technologies (RR) Whelk-01 pirkle chiral column (250mm x 4.6mm, 5m bead diameter) which was connected in turn to a UV-VIS (200nm-600nm) detector. Samples in EtOAc (with 0.0035% phenanthrene internal standard) were injected at a flow rate of 1.5mL/min in hexane:EtOH (1:1) with runtimes lasting 6-7minutes.

PAGE 133

116 Figure 68. Construction of pAAP20 (plasmid for expression of rabbit FMO2 under the control of P trc ) and pAAP15 (plasmid for expression of rabbit FMO2 under the control of P T7 )

PAGE 134

117 Figure 69. The construction of pAAP18 (plasmid for expression of rabbit FMO4 under the control of P T7 ) and pAAP14 (plasmid for expressio of rabbit FMO3 under the control of P T7 ).

PAGE 135

118 Figure 70. The construction of pAAP12 (plasmid for the expression of rabbit FMO2 under the control of P tac ), pAAP10 (plasmid for expression of rabbit FMO2 in yeast) and pAAP21 (plasmid for the expression of rabbit FMO2 under the control of P T7 with kanamycin selection).

PAGE 136

119 Figure 71. Construction of pAAP11 (plasmid for the expression of rabbit FMO1 under the control of P T7 with ampicillin selection), pAAP27 (plasmid for the expression of rabbit FMO1 under the control of P T7 with kanamycin selection) and pAAP38 (template for the error prone PCR reactions to generate clones of rabbit FMO1).

PAGE 137

120 Figure 72. Construction of pAAP35 (plasmid for the expression of rhesus macaque FMO2 under the control of P T7 with ampicillin selection), pAAP36 (plasmid for the expression of rhesus macaque FMO2 under the control of P T7 with kanamycin selection) and pAAP38 (template for the error prone PCR reactions to generate clones of rhesus macaque FMO2).

PAGE 138

121 Figure 73. Construction of pAAP46 (plasmid for the expression of rabbit FMO1 with a C-terminal His 6 tag under the control of P T7 ), pAAP41 (plasmid for the expression of rabbit FMO1 with a Cand N-terminal His 6 tag under the control of P T7 ) and pAAP29 (plasmid for the expression of rabbit FMO1 fused to MalE at the N-terminus, under the control of P tac ).

PAGE 139

APPENDIX B EXPERIMENTAL DATA AND MISCELANEOUS Yield and Enantioselectivity Calculations From standard plots of the peak areas vs. concentration: a = peak area at 233nm / mM MTS b = peak area at 233nm / mM MTSO c = peak area at 254nm / mM MTS d = peak area at 254nm / mM MTSO The concentration of unreacted substrate and product in each sample can be determined by measuring its peak area at 233nm and 254nm and using: bdadPAcPAaMTSOnmnm233254][ and adbcPAdPAbMTSnmnm233254] [ 122

PAGE 140

123 0.0E+002.0E+074.0E+076.0E+078.0E+071.0E+081.2E+081.4E+081.6E+0800.20.40.60.811.2mMpeak area (233nm) y = 9E+08x + 2E+06R2 = 0.9921y = 8E+08x 228376R2 = 0.99130.0E+001.0E+072.0E+073.0E+074.0E+075.0E+076.0E+0700.010.020.030.040.050.06mMpeak area (233nm) MTSO MTS MTSO MTS Figure 74. Standardization measurement of UV 233nm for MTS and MTSO indicating the error apparent in each reading. Samples were in solutions of cell extract and the background was corrected from the relative absorbance at 280nm. The upper graph indicates the non-linearity of measurements of concentrations >0.1mM. y = 4E+08x + 98149R2 = 0.9719y = 2E+09x 1E+06R2 = 0.99320.0E+002.0E+074.0E+076.0E+078.0E+071.0E+081.2E+0800.010.020.030.040.050.06mM MTS MTSO Figure 75. Standardization measurement of UV254nm for MTS and MTSO indicating the error apparent in each reading. Samples were in solutions of cell extract and the background was corrected from the relative absorbance at 280nm.

PAGE 141

124 A y = 8586.7x 1826.3R2 = 0.99590100002000030000400005000060000700008000090000100000024681012mM (R)-MTSO B y = 8740.7x 1780.5R2 = 0.99050100002000030000400005000060000700008000090000100000024681012mM (R)-MTSO Figure 76. Standardization measurement of OR for MTS and MTSO indicating the error apparent in each reading. Samples were in A) water or in B) solutions of cell extract for which the background was corrected from the relative absorbance at 280nm. Figure 77. Stereochemistry of MTS sulfoxidation by crude membrane preparations of rabbit FMO1 expressed in E. coli. A) racemic MTSO B) R-MTSO (Sigma-Aldrich) C) reaction product of MTS with crude membrane preparation of BL21(DE3)/pAAP11. (Analyzed by chiral GC (Chirasil-Val, 20m)).

PAGE 142

125 1 ATG GCC AAG CGA GTG GCA ATT GTG GGA GCT GGG GTG AGC GGC CTG GCC TCC ATC AAG TGC TGT CTG 1 M A K R V A I V G A G V S G L A S I K C C L 67 GAG GAA GGC CTG GAG CCC ACC TGC TTT GAG AGA AGT GAT GAC CTT GGG GGA CTG TGG AGA TTC ACG 23 E E G L E P T C F E R S D D L G G L W R F T 133 GAA CAT GTT GAA GAA GGA AGA GCC AGT CTG TAC AAG TCT GTG GTC TCC AAC AGC TGC AAG GAG ATG 45 E H V E E G R A S L Y K S V V S N S C K E M 199 TCC TGT TAC TCA GAT TTT CCA TTC CCA GAA GAT TAC CCA AAC TAT GTG CCA AAT TCT CAA TTC CTG 67 S C Y S D F P F P E D Y P N Y V P N S Q F L 265 GAC TAT CTC AAA ATG TAT GCA GAT CGA TTC AGC CTT CTG AAA TCG ATT CAA TTC AAG ACT ACA GTC 89 D Y L K M Y A D R F S L L K S I Q F K T T V 331 TTC AGT ATA ACA AAA TGC CAA GAT TTT AAT GTC TCT GGC CAG TGG GAG GTA GTG ACT CTG CAT GAA 111 F S I T K C Q D F N V S G Q W E V V T L H E 397 GGG AAG CAA GAA TCA GCC ATC TTT GAT GCT GTC ATG GTC TGC ACT GGT TTT CTT ACG AAC CCT CAT 133 G K Q E S A I F D A V M V C T G F L T N P H 463 TTG CCA TTG GGT TGC TTT CCA GGC ATA AAA ACC TTT AAA GGC CAA TAC TTT CAC AGC CGA CAG TAT 155 L P L G C F P G I K T F K G Q Y F H S R Q Y 529 AAA CAT CCA GAT ATA TTT AAG GAC AAG AGA GTT CTT GTG GTT GGA ATG GGC AAT TCT GGC ACA GAC 177 K H P D I F K D K R V L V V G M G N S G T D 595 ATT GCT GTG GAG GCC AGC CAC GTG GCA AAA AAG GTG TTC CTC AGC ACT ACC GGA GGG GCG TGG GTG 199 I A V E A S H V A K K V F L S T T G G A W V 661 ATC AGC CGG GTC TTT GAC TCC GGG TAC CCA TGG GAC ATG GTG TTC ACG ACA CGA TTT CAG AAC TTC 221 I S R V F D S G Y P W D M V F T T R F Q N F 727 ATC AGA AAT TCT CTC CCA ACT CCA ATT GTG ACT TGG CTG GTG GCA AAA AAG ATG AAC AGC TGG TTC 243 I R N S L P T P I V T W L V A K K M N S W F 793 AAC CAT GCA AAT TAT GGC TTA GTA CCA AAA GAC AGG ATT CAA CTG AAA GAG CCT GTG CTA AAT GAT 265 N H A N Y G L V P K D R I Q L K E P V L N D 859 GAG CTC CCA GGC CGT ATC ATC ACG GGG AAA GTT TTT ATC AGG CCA AGC ATA AAG GAG GTG AAA GAA 287 E L P G R I I V T G K V F I R P S I K E V K 925 AAC TCT GTT GTA TTT GGC AAT GCA CAC AAC ACC CCA AGC GAA GAG CCC ATT GAT GTC ATT GTC TTT 309 E N S V V F G N A H N T P S E E P I D V I V 991 GCC ACT GGA TAC ACT TTT GCC TTC CCC TTC CTT GAT GAG TCT GTA GTG AAA GTT GAA GAT GGT CAA 331 F A T G Y T F A F P F L D E S V V K E D G Q 1057 GCA TCA CTG TAC AAG TAT ATA TTC CCT GCA CAT CTG CAA AAG CCA ACC TTG GCT GTT ATT GGC CTC 353 A S L Y K Y I F P A H L Q K P T L A V I G L 1123 ATT AAA CCC TTG GGG TCT ATG TTA CCC ACA GGA GAA ACA CAA GCT CGA TAT ACT GTT CAA GTT TTT 375 I K P L G S M L P T G E T Q A R Y T V Q V F 1189 AAA GGT GTA ATT AAA TTA CCA CCA ACA AGT GTC ATG ATA AAA GAA GTT AAT GAA AGG AAA GAA AAC 397 K G V I K L P P T S V M I K E V N E R K E N 1255 AAG CAC AAT GGG TTT GGC TTG TGC TAC TGC AAG GCC TTA CAA GCA GAT TAC ATC ACG TAC ATA GAT 419 K H N G F G L C Y C K A L Q A D Y I T Y I D G 1321 GAT CTC CTG ACC TCT ATC AAC GCA AAA CCC AAC CTG TTC TCT CTA CTC CTG ACT GAC CCA CTG CTG 441 D L L T S I N A K P D/ N L F S L L L T D P L L 1387 GCT TTG ACT ATG TTC TTT GGC CCA TAT TCA CCA TAC CAA TTC CGC TTG ACT GGA CCA GGA AAA TGG 463 A L T M F F G P Y S P Y Q F R L T G P G K W 1453 AAA GGA GCC AGA AAT GCC ATC ATG ACA CAA TGG GAT CGC ACA TTC AAG GTC ACC AAA ACT CGA ATT 485 K G A R N A I M T Q W D R T F K V T K T R I 1519 GTA CAA GAA TCG TCA TCT CCC TTT GAA AGT TTG CTT AAA CTC TTC GCC GTT CTG GCT TTG CTT GTG 507 V Q E S S S P F E S L L K L F A V L A L L V 1585 TCT GTT TTC CTG ATT TTC CTA 529 S V F L I F L Figure 78. The sequence of clone 170AA21 of rabbit FMO1. Upper sequence is DNA, beneath it is the predicted amino acid sequence. Wild type sequence is indicated underlined where there has been a mutation (N451D).

PAGE 143

126 1 ATG GCC AAG CGA GTG GCA ATT GTG GGA GCT GGG GTG AGC GGC CTG GCC TCC ATC AAG TGC TGT CTG 1 M A K R V A I V G A G V S G L A S I K C C L 67 GAG GAA GGC CTG GAG CCC ACC TGC TTT GAG AGA AGT GAT GAC CTT GGG GGA CTG TGG AGA TTC ACG 23 E E G L E P T C F E R S D D L G G L W R F T 133 GAA CAT GTT GAA GAA GGA AGA GCC AGT CTG TAC AAG TCT GTG GTC TCC AAC AGC TGC AAG GAG ATG 45 E H V E E G R A S L Y K S V V S N S C K E M A 199 TCC TGT TAC TCA GAT TTT CCA TTC CCA GAA GAT TAC CCA AAC TTT GTG CCA AAT TCT CAA TTC CTG 67 S C Y S D F P F P E D Y P N F/ Y V P N S Q F L 265 GAC TAT CTC AAA ATG TAT GCA GAT CGA TTC AGC CTT CTG AAA TCG ATT CAA TTC AAG ACT ACA GTC 89 D Y L K M Y A D R F S L L K S I Q F K T T V 331 TTC AGT ATA ACA AAA TGC CAA GAT TTT AAT GTC TCT GGC CAG TGG GAG GTA GTG ACT CTG CAT GAA 111 F S I T K C Q D F N V S G Q W E V V T L H E C 397 GGG AAG CAA GAA TCA GCA ATC TTT GAT GCT GTC ATG GTC TGC ACT GGT TTT CTT ACG AAC CCT CAT 133 G K Q E S A A/ I F D A V M V C T G F L T N P H 463 TTG CCA TTG GGT TGC TTT CCA GGC ATA AAA ACC TTT AAA GGC CAA TAC TTT CAC AGC CGA CAG TAT 155 L P L G C F P G I K T F K G Q Y F H S R Q Y 529 AAA CAT CCA GAT ATA TTT AAG GAC AAG AGA GTT CTT GTG GTT GGA ATG GGC AAT TCT GGC ACA GAC 177 K H P D I F K D K R V L V V G M G N S G T D 595 ATT GCT GTG GAG GCC AGC CAC GTG GCA AAA AAG GTG TTC CTC AGC ACT ACC GGA GGG GCG TGG GTG 199 I A V E A S H V A K K V F L S T T G G A W V 661 ATC AGC CGG GTC TTT GAC TCC GGG TAC CCA TGG GAC ATG GTG TTC ACG ACA CGA TTT CAG AAC TTC 221 I S R V F D S G Y P W D M V F T T R F Q N F 727 ATC AGA AAT TCT CTC CCA ACT CCA ATT GTG ACT TGG CTG GTG GCA AAA AAG ATG AAC AGC TGG TTC 243 I R N S L P T P I V T W L V A K K M N S W F 793 AAC CAT GCA AAT TAT GGC TTA GTA CCA AAA GAC AGG ATT CAA CTG AAA GAG CCT GTG CTA AAT GAT 265 N H A N Y G L V P K D R I Q L K E P V L N D A 859 GAG CTC CCA GGC CGT ATC ATC ACG GGG AAA GTT TTT ATC AGG CCA AGC ATA AAG GAG GTG AAA GAG 287 E L P G R I I V T G K V F I R P S I K E V K 925 AAC TCT GTT GTA TTT GGC AAT GCA CAC AAC ACC CCA AGC GAA GAG CCC ATT GAT GTC ATT GTC TTT 309 E N S V V F G N A H N T P S E E P I D V I V 991 GCC ACT GGA TAC ACT TTT GCC TTC CCC TTC CTT GAT GAG TCT GTA GTG AAA GTT GAA GAT GGT CAA 331 F A T G Y T F A F P F L D E S V V K E D G Q 1057 GCA TCA CTG TAC AAG TAT ATA TTC CCT GCA CAT CTG CAA AAG CCA ACC TTG GCT GTT ATT GGC CTC 353 A S L Y K Y I F P A H L Q K P T L A V I G L 1123 ATT AAA CCC TTG GGG TCT ATG TTA CCC ACA GGA GAA ACA CAA GCT CGA TAT ACT GTT CAA GTT TTT 375 I K P L G S M L P T G E T Q A R Y T V Q V F 1189 AAA GGT GTA ATT AAA TTA CCA CCA ACA AGT GTC ATG ATA AAA GAA GTT AAT GAA AGG AAA GAA AAC 397 K G V I K L P P T S V M I K E V N E R K E N 1255 AAG CAC AAT GGG TTT GGC TTG TGC TAC TGC AAG GCC TTA CAA GCA GAT TAC ATC ACG TAC ATA GAT 419 K H N G F G L C Y C K A L Q A D Y I T Y I D 1321 GAT CTC CTG ACC TCT ATC AAC GCA AAA CCC AAC CTG TTC TCT CTA CTC CTG ACT GAC CCA CTG CTG 441 D L L T S I N A K P N L F S L L L T D P L L 1387 GCT TTG ACT ATG TTC TTT GGC CCA TAT TCA CCA TAC CAA TTC CGC TTG ACT GGA CCA GGA AAA TGG 463 A L T M F F G P Y S P Y Q F R L T G P G K W 1453 AAA GGA GCC AGA AAT GCC ATC ATG ACA CAA TGG GAT CGC ACA TTC AAG GTC ACC AAA ACT CGA ATT 485 K G A R N A I M T Q W D R T F K V T K T R I 1519 GTA CAA GAA TCG TCA TCT CCC TTT GAA AGT TTG CTT AAA CTC TTC GCC GTT CTG GCT TTG CTT GTG 507 V Q E S S S P F E S L L K L F A V L A L L V 1585 TCT GTT TTC CTG ATT TTC CTA 529 S V F L I F L Figure 79. The sequence of clone 176C34 of rabbit FMO1. Upper sequence is DNA, beneath it is the predicted amino acid sequence. Wild type sequence is indicated underlined where there has been a mutation (I138A) (Y81F).

PAGE 144

127 1 ATG GCC AAG CGA GTG GCA ATT GTG GGA GCT GGG GTG AGC GGC CTG GCC TCC ATC AAG TGC TGT CTG 1 M A K R V A I V G A G V S G L A S I K C C L 67 GAG GAA GGC CTG GAG CCC ACC TGC TTT GAG AGA AGT GAT GAC CTT GGG GGA CTG TGG AGA TTC ACG 23 E E G L E P T C F E R S D D L G G L W R F T 133 GAA CAT GTT GAA GAA GGA AGA GCC AGT CTG TAC AAG TCT GTG GTC TCC AAC AGC TGC AAG GAG ATG 45 E H V E E G R A S L Y K S V V S N S C K E M 199 TCC TGT TAC TCA GAT TTT CCA TTC CCA GAA GAT TAC CCA AAC TAT GTG CCA AAT TCT CAA TTC CTG 67 S C Y S D F P F P E D Y P N Y V P N S Q F L 265 GAC TAT CTC AAA ATG TAT GCA GAT CGA TTC AGC CTT CTG AAA TCG ATT CAA TTC AAG ACT ACA GTC 89 D Y L K M Y A D R F S L L K S I Q F K T T V 331 TTC AGT ATA ACA AAA TGC CAA GAT TTT AAT GTC TCT GGC CAG TGG GAG GTA GTG ACT CTG CAT GAA 111 F S I T K C Q D F N V S G Q W E V V T L H E 397 GGG AAG CAA GAA TCA GCC ATC TTT GAT GCT GTC ATG GTC TGC ACT GGT TTT CTT ACG AAC CCT CAT 133 G K Q E S A I F D A V M V C T G F L T N P H 463 TTG CCA TTG GGT TGC TTT CCA GGC ATA AAA ACC TTT AAA GGC CAA TAC TTT CAC AGC CGA CAG TAT 155 L P L G C F P G I K T F K G Q Y F H S R Q Y T 529 AAA CAC CCA GAT ATA TTT AAG GAC AAG AGA GTT CTT GTG GTT GGA ATG GGC AAT TCT GGC ACA GAC 177 K H P D I F K D K R V L V V G M G N S G T D 595 ATT GCT GTG GAG GCC AGC CAC GTG GCA AAA AAG GTG TTC CTC AGC ACT ACC GGA GGG GCG TGG GTG 199 I A V E A S H V A K K V F L S T T G G A W V 661 ATC AGC CGG GTC TTT GAC TCC GGG TAC CCA TGG GAC ATG GTG TTC ACG ACA CGA TTT CAG AAC TTC 221 I S R V F D S G Y P W D M V F T T R F Q N F 727 ATC AGA AAT TCT CTC CCA ACT CCA ATT GTG ACT TGG CTG GTG GCA AAA AAG ATG AAC AGC TGG TTC 243 I R N S L P T P I V T W L V A K K M N S W F 793 AAC CAT GCA AAT TAT GGC TTA GTA CCA AAA GAC AGG ATT CAA CTG AAA GAG CCT GTG CTA AAT GAT 265 N H A N Y G L V P K D R I Q L K E P V L N D C 859 GAG CTC CCA GGC CGT ATC ATT ACG GGG AAA GTT TTT ATC AGG CCA AGC ATA AAG GAG GTG AAA GAA 287 E L P G R I I V T G K V F I R P S I K E V K 925 AAC TCT GTT GTA TTT GGC AAT GCA CAC AAC ACC CCA AGC GAA GAG CCC ATT GAT GTC ATT GTC TTT 309 E N S V V F G N A H N T P S E E P I D V I V 991 GCC ACT GGA TAC ACT TTT GCC TTC CCC TTC CTT GAT GAG TCT GTA GTG AAA GTT GAA GAT GGT CAA 331 F A T G Y T F A F P F L D E S V V K E D G Q 1057 GCA TCA CTG TAC AAG TAT ATA TTC CCT GCA CAT CTG CAA AAG CCA ACC TTG GCT GTT ATT GGC CTC 353 A S L Y K Y I F P A H L Q K P T L A V I G L 1123 ATT AAA CCC TTG GGG TCT ATG TTA CCC ACA GGA GAA ACA CAA GCT CGA TAT ACT GTT CAA GTT TTT 375 I K P L G S M L P T G E T Q A R Y T V Q V F 1189 AAA GGT GTA ATT AAA TTA CCA CCA ACA AGT GTC ATG ATA AAA GAA GTT AAT GAA AGG AAA GAA AAC 397 K G V I K L P P T S V M I K E V N E R K E N T 1255 AAG CAC AAT GGG TTT GGC TTG TGC CAC TGC AAG GCC TTA CAA GCA GAT TAC ATC ACG TAC ATA GAT 419 K H N G F G L C H/ Y C K A L Q A D Y I T Y I D 1321 GAT CTC CTG ACC TCT ATC AAC GCA AAA CCC AAC CTG TTC TCT CTA CTC CTG ACT GAC CCA CTG CTG 441 D L L T S I N A K P N L F S L L L T D P L L 1387 GCT TTG ACT ATG TTC TTT GGC CCA TAT TCA CCA TAC CAA TTC CGC TTG ACT GGA CCA GGA AAA TGG 463 A L T M F F G P Y S P Y Q F R L T G P G K W 1453 AAA GGA GCC AGA AAT GCC ATC ATG ACA CAA TGG GAT CGC ACA TTC AAG GTC ACC AAA ACT CGA ATT 485 K G A R N A I M T Q W D R T F K V T K T R I 1519 GTA CAA GAA TCG TCA TCT CCC TTT GAA AGT TTG CTT AAA CTC TTC GCC GTT CTG GCT TTG CTT GTG 507 V Q E S S S P F E S L L K L F A V L A L L V 1585 TCT GTT TTC CTG ATT TTC CTA 529 S V F L I F L Figure 80. The sequence of clone 170D96 of rabbit FMO1. Upper sequence is DNA, beneath it is the predicted amino acid sequence. Wild type sequence is indicated underlined where there has been a mutation (Y428H).

PAGE 145

128 1 ATG GCC AAG CGA GTG GCA ATT GTG GGA GCT GGG GTG AGC GGC CTG GCC TCC ATC AAG TGC TGT CTG 1 M A K R V A I V G A G V S G L A S I K C C L 67 GAG GAA GGC CTG GAG CCC ACC TGC TTT GAG AGA AGT GAT GAC CTT GGG GGA CTG TGG AGA TTC ACG 23 E E G L E P T C F E R S D D L G G L W R F T 133 GAA CAT GTT GAA GAA GGA AGA GCC AGT CTG TAC AAG TCT GTG GTC TCC AAC AGC TGC AAG GAG ATG 45 E H V E E G R A S L Y K S V V S N S C K E M 199 TCC TGT TAC TCA GAT TTT CCA TTC CCA GAA GAT TAC CCA AAC TAT GTG CCA AAT TCT CAA TTC CTG 67 S C Y S D F P F P E D Y P N Y V P N S Q F L 265 GAC TAT CTC AAA ATG TAT GCA GAT CGA TTC AGC CTT CTG AAA TCG ATT CAA TTC AAG ACT ACA GTC 89 D Y L K M Y A D R F S L L K S I Q F K T T V 331 TTC AGT ATA ACA AAA TGC CAA GAT TTT AAT GTC TCT GGC CAG TGG GAG GTA GTG ACT CTG CAT GAA 111 F S I T K C Q D F N V S G Q W E V V T L H E 397 GGG AAG CAA GAA TCA GCC ATC TTT GAT GCT GTC ATG GTC TGC ACT GGT TTT CTT ACG AAC CCT CAT 133 G K Q E S A I F D A V M V C T G F L T N P H A 463 TTG CCA TTG GGT TGC TTT CCA GGC ATA AAC ACC TTT AAA GGC CAA TAC TTT CAC AGC CGA CAG TAT 155 L P L G C F P G I N/ K T F K G Q Y F H S R Q Y 529 AAA CAT CCA GAT ATA TTT AAG GAC AAG AGA GTT CTT GTG GTT GGA ATG GGC AAT TCT GGC ACA GAC 177 K H P D I F K D K R V L V V G M G N S G T D 595 ATT GCT GTG GAG GCC AGC CAC GTG GCA AAA AAG GTG TTC CTC AGC ACT ACC GGA GGG GCG TGG GTG 199 I A V E A S H V A K K V F L S T T G G A W V 661 ATC AGC CGG GTC TTT GAC TCC GGG TAC CCA TGG GAC ATG GTG TTC ACG ACA CGA TTT CAG AAC TTC 221 I S R V F D S G Y P W D M V F T T R F Q N F 727 ATC AGA AAT TCT CTC CCA ACT CCA ATT GTG ACT TGG CTG GTG GCA AAA AAG ATG AAC AGC TGG TTC 243 I R N S L P T P I V T W L V A K K M N S W F 793 AAC CAT GCA AAT TAT GGC TTA GTA CCA AAA GAC AGG ATT CAA CTG AAA GAG CCT GTG CTA AAT GAT 265 N H A N Y G L V P K D R I Q L K E P V L N D 859 GAG CTC CCA GGC CGT ATC ATC ACG GGG AAA GTT TTT ATC AGG CCA AGC ATA AAG GAG GTG AAA GAA 287 E L P G R I I V T G K V F I R P S I K E V K 925 AAC TCT GTT GTA TTT GGC AAT GCA CAC AAC ACC CCA AGC GAA GAG CCC ATT GAT GTC ATT GTC TTT 309 E N S V V F G N A H N T P S E E P I D V I V 991 GCC ACT GGA TAC ACT TTT GCC TTC CCC TTC CTT GAT GAG TCT GTA GTG AAA GTT GAA GAT GGT CAA 331 F A T G Y T F A F P F L D E S V V K E D G Q 1057 GCA TCA CTG TAC AAG TAT ATA TTC CCT GCA CAT CTG CAA AAG CCA ACC TTG GCT GTT ATT GGC CTC 353 A S L Y K Y I F P A H L Q K P T L A V I G L 1123 ATT AAA CCC TTG GGG TCT ATG TTA CCC ACA GGA GAA ACA CAA GCT CGA TAT ACT GTT CAA GTT TTT 375 I K P L G S M L P T G E T Q A R Y T V Q V F G 1189 AAA GGT GTA ATT AAA TTA CCA CCA ACA AGT GTC ATA ATA AAA GAA GTT AAT GAA AGG AAA GAA AAC 397 K G V I K L P P T S V I/ M I K E V N E R K E N 1255 AAG CAC AAT GGG TTT GGC TTG TGC TAC TGC AAG GCC TTA CAA GCA GAT TAC ATC ACG TAC ATA GAT 419 K H N G F G L C Y C K A L Q A D Y I T Y I D 1321 GAT CTC CTG ACC TCT ATC AAC GCA AAA CCC AAC CTG TTC TCT CTA CTC CTG ACT GAC CCA CTG CTG 441 D L L T S I N A K P N L F S L L L T D P L L 1387 GCT TTG ACT ATG TTC TTT GGC CCA TAT TCA CCA TAC CAA TTC CGC TTG ACT GGA CCA GGA AAA TGG 463 A L T M F F G P Y S P Y Q F R L T G P G K W 1453 AAA GGA GCC AGA AAT GCC ATC ATG ACA CAA TGG GAT CGC ACA TTC AAG GTC ACC AAA ACT CGA ATT 485 K G A R N A I M T Q W D R T F K V T K T R I 1519 GTA CAA GAA TCG TCA TCT CCC TTT GAA AGT TTG CTT AAA CTC TTC GCC GTT CTG GCT TTG CTT GTG 507 V Q E S S S P F E S L L K L F A V L A L L V 1585 TCT GTT TTC CTG ATT TTC CTA 529 S V F L I F L Figure 81. The sequence of clone 170D62 of rabbit FMO1. Upper sequence is DNA, beneath it is the predicted amino acid sequence. Wild type sequence is indicated underlined where there has been a mutation (K164N) (M408I).

PAGE 146

129 1 ATG GCC AAG CGA GTG GCA ATT GTG GGA GCT GGG GTG AGC GGC CTG GCC TCC ATC AAG TGC TGT CTG 1 M A K R V A I V G A G V S G L A S I K C C L 67 GAG GAA GGC CTG GAG CCC ACC TGC TTT GAG AGA AGT GAT GAC CTT GGG GGA CTG TGG AGA TTC ACG 23 E E G L E P T C F E R S D D L G G L W R F T 133 GAA CAT GTT GAA GAA GGA AGA GCC AGT CTG TAC AAG TCT GTG GTC TCC AAC AGC TGC AAG GAG ATG 45 E H V E E G R A S L Y K S V V S N S C K E M T 199 TCC TGT TAC TCA GAT TTT CCA TTC CCA GAA GAC TAC CCA AAC TAT GTG CCA AAT TCT CAA TTC CTG 67 S C Y S D F P F P E D Y P N Y V P N S Q F L 265 GAC TAT CTC AAA ATG TAT GCA GAT CGA TTC AGC CTT CTG AAA TCG ATT CAA TTC AAG ACT ACA GTC 89 D Y L K M Y A D R F S L L K S I Q F K T T V 331 TTC AGT ATA ACA AAA TGC CAA GAT TTT AAT GTC TCT GGC CAG TGG GAG GTA GTG ACT CTG CAT GAA 111 F S I T K C Q D F N V S G Q W E V V T L H E 397 GGG AAG CAA GAA TCA GCC ATC TTT GAT GCT GTC ATG GTC TGC ACT GGT TTT CTT ACG AAC CCT CAT 133 G K Q E S A I F D A V M V C T G F L T N P H 463 TTG CCA TTG GGT TGC TTT CCA GGC ATA AAA ACC TTT AAA GGC CAA TAC TTT CAC AGC CGA CAG TAT 155 L P L G C F P G I K T F K G Q Y F H S R Q Y 529 AAA CAT CCA GAT ATA TTT AAG GAC AAG AGA GTT CTT GTG GTT GGA ATG GGC AAT TCT GGC ACA GAC 177 K H P D I F K D K R V L V V G M G N S G T D 595 ATT GCT GTG GAG GCC AGC CAC GTG GCA AAA AAG GTG TTC CTC AGC ACT ACC GGA GGG GCG TGG GTG 199 I A V E A S H V A K K V F L S T T G G A W V T 661 ATC AGC CGG GTC TTT GAC ACC GGG TAC CCA TGG GAC ATG GTG TTC ACG ACA CGA TTT CAG AAC TTC 221 I S R V F D T/ S G Y P W D M V F T T R F Q N F 727 ATC AGA AAT TCT CTC CCA ACT CCA ATT GTG ACT TGG CTG GTG GCA AAA AAG ATG AAC AGC TGG TTC 243 I R N S L P T P I V T W L V A K K M N S W F 793 AAC CAT GCA AAT TAT GGC TTA GTA CCA AAA GAC AGG ATT CAA CTG AAA GAG CCT GTG CTA AAT GAT 265 N H A N Y G L V P K D R I Q L K E P V L N D 859 GAG CTC CCA GGC CGT ATC ATC ACG GGG AAA GTT TTT ATC AGG CCA AGC ATA AAG GAG GTG AAA GAA 287 E L P G R I I V T G K V F I R P S I K E V K 925 AAC TCT GTT GTA TTT GGC AAT GCA CAC AAC ACC CCA AGC GAA GAG CCC ATT GAT GTC ATT GTC TTT 309 E N S V V F G N A H N T P S E E P I D V I V 991 GCC ACT GGA TAC ACT TTT GCC TTC CCC TTC CTT GAT GAG TCT GTA GTG AAA GTT GAA GAT GGT CAA 331 F A T G Y T F A F P F L D E S V V K E D G Q 1057 GCA TCA CTG TAC AAG TAT ATA TTC CCT GCA CAT CTG CAA AAG CCA ACC TTG GCT GTT ATT GGC CTC 353 A S L Y K Y I F P A H L Q K P T L A V I G L 1123 ATT AAA CCC TTG GGG TCT ATG TTA CCC ACA GGA GAA ACA CAA GCT CGA TAT ACT GTT CAA GTT TTT 375 I K P L G S M L P T G E T Q A R Y T V Q V F 1189 AAA GGT GTA ATT AAA TTA CCA CCA ACA AGT GTC ATG ATA AAA GAA GTT AAT GAA AGG AAA GAA AAC 397 K G V I K L P P T S V M I K E V N E R K E N 1255 AAG CAC AAT GGG TTT GGC TTG TGC TAC TGC AAG GCC TTA CAA GCA GAT TAC ATC ACG TAC ATA GAT 419 K H N G F G L C Y C K A L Q A D Y I T Y I D 1321 GAT CTC CTG ACC TCT ATC AAC GCA AAA CCC AAC CTG TTC TCT CTA CTC CTG ACT GAC CCA CTG CTG 441 D L L T S I N A K P N L F S L L L T D P L L 1387 GCT TTG ACT ATG TTC TTT GGC CCA TAT TCA CCA TAC CAA TTC CGC TTG ACT GGA CCA GGA AAA TGG 463 A L T M F F G P Y S P Y Q F R L T G P G K W 1453 AAA GGA GCC AGA AAT GCC ATC ATG ACA CAA TGG GAT CGC ACA TTC AAG GTC ACC AAA ACT CGA ATT 485 K G A R N A I M T Q W D R T F K V T K T R I 1519 GTA CAA GAA TCG TCA TCT CCC TTT GAA AGT TTG CTT AAA CTC TTC GCC GTT CTG GCT TTG CTT GTG 507 V Q E S S S P F E S L L K L F A V L A L L V 1585 TCT GTT TTC CTG ATT TTC CTA 529 S V F L I F L Figure 82. The sequence of clone 176E70 of rabbit FMO1. Upper sequence is DNA, beneath it is the predicted amino acid sequence. Wild type sequence is indicated underlined where there has been a mutation (S227T).

PAGE 147

APPENDIX C GLOSSARY OF TERMS AND ABBREVIATIONS FMO Flavin-containing monooxygenase FAD Flavin adenine dinucleotide DET Diethyl D or L -tartrate Ti(OiPr) 4 Titanium tetraisopropoxide DIBAL Diisobutylaluminum hydride TBDMSO t ButyldimethylsiloxyTHF Tetrahydrofuran IPTG Isopropyl -D-Thiogalactopyranoside DMSO Dimethyl Sulfoxide Oxone 2KHSO 5 KHSO 4 K 2 SO 4 DET diethyl tartrate e.e. enantiomeric excess d.e. diasteriomeric excess MTS Methyl p-tolyl sulfide MTSO Methyl p-tolyl sulfoxide wt wild-type (natural allele of a gene) X-Gal 5-Bromo-4-chloro-3-indolyl -D-glactopyranoside rt Room Temperature EtOH Ethanol s/n Supernatant UV/VIS Ultraviolet/Visible wavelength detector ORD Optical rotation detector GC Gas chromatography deIH 2 O De-ionized H 2 O NADP -Nicotinamide adenine dinucleotide phosphate (oxidized form) NADPH -Nicotinamide adenine dinucleotide phosphate (reduced form) PCR Polymerase Chain Reaction FAD Flavin adenine dinucleotide bp Base pairs dATP 2-deoxyadenosine 5-triphosphate dGTP 2-deoxyguanosine 5-triphosphate dCTP 2-deoxycytosine 5-triphosphate dTTP 2-deoxythymidine 5-triphosphate EDTA Ethylenediaminetetraacetic acid 130

PAGE 148

LIST OF REFERENCES (1) Stinson, S. C. Chem. Eng. News 2000, Oct. 23, 55-78. (2) DeSouza Pereira, R. Crit. Rev. Biotech. 1998, 18, 25-83. (3) Mislow, K.; Siegel, J. J. Am. Chem. Soc. 1984, 106, 3319-3324. (4) Prasanta, K. D.; Caaveiro, J.M.M.; Luque, S.; Klibanov, A.M. Biotech. Bioeng. 2002, 78, 104-109. (5) Carreno, M. C. Chem. Rev. 1995, 95, 1717-1760. (6) Solladie, G. Heteroatom. Chem. 2002, 13, 443-452. (7) Posner, G. In The Chemistry of Sulphones and Sulphoxides: The Chemistry of Functional Groups; Patai, S.; Rappoport, Z.; Stirling, C.; Ed.; Wiley: New York, 1988; pp 823-834. (8) Oae, S.; Uchida, Y. In The Chemistry of Sulphones and Sulphoxides; The Chemistry of Functional Groups; Patai, S.; Rappoport, Z.; Stirling, C.; Ed.; Wiley: New York, 1988; pp 583-597. (9) Solladie, G.; Carreno, M.C. In Organosulfur Chemistry; Synthetic Aspects; Page, P.; Ed.; Academic Press: New York, 1995. (10) Tang, J.; Brackenridge, I.; Roberts, S.M.; Beecher, J.; Willetts, A.J. Tetrahedron 1995, 51, 13217-13238. (11) Prugh, J. D.; Deana, A.A. Tetrahedron 1982, 23. (12) Fleming, I.; Trost, B.M. Comprehensive Organic Synthesis; Pergamon Press: London, 1991; Vol. 8. (13) Kimura, K.; Kano, F.; Kurosawa, K.; Yoshihama, M. Chem. Abstr. 1994, 122, 8155. 131

PAGE 149

132 (14) Oka, M.; Kamei, H.; Hamagishi, Y.; Tomita, K.; Miyaki, T.; Konishi, T.; Oki, T. J. Antibiot. 1990, 43, 967-976. (15) Carreno, M. C.; Ribagorda, M.; Somoza, A.; Urbano, A. Angew. Chem. Int. Ed. Engl. 2002, 41, 2755-2757. (16) Solladie, G.; Salmon-Roig, X.J.; Hanquest, G. Tetrahedron Lett. 2000, 41, 551-554. (17) Colobert, F.; Mazery, R.; Solladie, G.; Carreno, M.C. Org. Lett. 2002, 4, 1723-1725. (18) Pictet, A.; Spengler, T. Ber. Dtsch. Chem. Ges. 1911, 44, 2030-2036. (19) Bravo, P.; Crucianelli, M.; Farina, A.; Meille, S.V.; Volonterio, A.; Zanda, M. Eur. J. Org. Chem. 1997, 435-440. (20) Bravo, P.; Frigierio, M.; Ono, T.; Panzeri, W.; Pesenti, C.; Sekine, A.; Viani, F. Eur. J. Org. Chem. 2000, 1387-1389. (21) Arroyo-Gomez, Y.; Rodriguez-Amo, J.F.; Santos-Garcia, M.; Sanz-Tejedor, M.A. Tetrahedron Asymm. 2000, 11, 789-796. (22) Sakuraba, H.; Ushiki, S. Tetrahedron Lett. 1990, 31, 5349-5352. (23) Bordwell, F.; Boutan, P.J. J. Am. Chem. Soc. 1957, 79, 717-719. (24) Peak, D. A.; Watkins, T. I. J. Chem. Soc. 1950, 445-446. (25) Addison, C. C.; Sheldon, J.C. J. Chem. Soc. 1956, 2705-2707. (26) Edwards, D.; Stenlake, J.B. J. Chem. Soc. 1954, 3274-3276. (27) Horner, L.; Shaefer, H.; Ludwig, W. Chem. Ber. 1958, 75, 91-101. (28) Melnikov, N. N. Uspekhi Khim. 1936, 5. (29) Johnson, C. R.; Keiser, J.E. Organic Synthesis Collect. Vol V 1973, 791-812. (30) Evans, T. L.; Grade, M.M. Synth. Commun. 1986, 16, 1207-1208. (31) Sugimoto, T.; Kokubo, T.; Miyazaki, J.; Tanimoto, S.; Okan, M. J. Chem. Soc.; Chem. Commun. 1979, 402-404. (32) Lin, K. T.; Tong, Y.C. J. Org. Chem. 1978, 43, 2717-2721.

PAGE 150

133 (33) Solladie, G. In Comprehensive Organic Synthesis; Trost, B. M.; Fleming, I.; Ed.; Pergammon: Oxford, 1995; Vol. 2; pp 144-153. (34) Cope, A. C.; Caress, E.A. J. Am. Chem. Soc. 1966, 88, 1711-1713. (35) Drabowicz, J.; Kielbasinski, P.; Mikolajczyk, M. In The Chemistry of Sulfones and Sulfoxides; Patai, S.; Rappoport, Z.; Stirling, C.J.M.; Ed.; Wiley: New York, 1988; pp 233. (36) Overberger, C. G.; Cummins, R.W. J. Am. Chem. Soc. 1953, 75, 4250-4252. (37) Folli, U.; Larossi, D.; Montanon, F.; Tores, G. J. Chem. Soc. 1968, 1317-1320. (38) Hargrave, K. R. Proc. Royal Soc. London 1956, A-235, 55-57. (39) Pitchen, P.; Dunach, E.; Deshmukh, N.N.; Kagan, H.B. J. Am. Chem. Soc. 1984, 106, 8188-8192. (40) Di Furia, M.G.; Seraglia, R. Synthesis 1984, 325-329. (41) Pitchen, P.; Dunach, E.; Deshmukh, M.N.; Kagan, M.B. J. Am. Chem. Soc. 1984, 106. (42) Cashman, J. R.; Olsen, L.D.; Bornheim, L.M. J. Am. Chem. Soc. 1990, 112, 3191-3195. (43) Komatsu, N.; Hashizume, M.; Sugita, T.; Uemura, S. J. Org. Chem. 1993, 58, 4529-4533. (44) Adam, W.; Marion, N.K.; Roschmann, K.J.; Moller, C.R. J. Org. Chem. 1999, 63, 3423-3428. (45) Davis, F. A.; Chen, B.C. Chem. Rev. 1992, 92, 919-935. (46) Davis, F. A.; Sheppard, A.C. Tetrahedron 1989, 45. (47) Davis, F. A.; Towson, J.C.; Weismiller, M.C.; Lal, C.; Carroll, P.F. J. Am. Chem. Soc. 1988, 110, 8477-8482. (48) Davis, F. A.; Jenkins, R.H.; Awad, S.B.; Stringer, O.D.; Watson, W.H.; Galloy, J J. Am. Chem. Soc. 1982, 104, 5412-5418. (49) Davis, F. A.; MacCauley, J.P.; Chattopadhyay, S.; Harakal, M.E.; Towson, J.C.; Watson, W.H.; Travanaiepour, I. J. Am. Chem. Soc. 1987, 109, 3370-3374.

PAGE 151

134 (50) Davis, F. J.; Reddy, R.T.; Han, W.; Carroll, P.J. J. Am. Chem. Soc. 1992, 114, 1428-1434. (51) Davis, F. A.; Reddy, R.T.; Weismiller, M.C. J. Am. Chem. Soc. 1989, 111, 5964-5969. (52) Axelrod, M.; Bickart, P.; Jacobus, J.; Green, M.M.; Mislow, K. J. Am. Chem. Soc. 1968, 90, 4835-4839. (53) Mislow, K.; Ternay, A.; Melillo, J.T. J. Am. Chem. Soc. 1963, 85, 2329-2333. (54) Rolland, C.; Hanquet, G.; Ducep, J.B.; Solladie, G. Tetrahedron Lett. 2001, 42, 9077-9080. (55) Page, P. C. B.; Wilkes, R.D.; Reynolds, D. In Comprehensive Organic Functional Group Transformations; Pergamon: New York, 1995; Vol. 2; pp 144-153. (56) Durst, T.; LeBelle, M.J.; VenDerElzen, R.; Tin, K.C. Can. J. Chem. 1974, 52, 761-764. (57) Lockard, J. P.; Schroek, C.W.; Johnson, C.R. Synthesis 1973, 485-489. (58) Capozzi, M. A. M.; Cardellicchio, C.; Naso, F.; Rosito, V. J. Org. Chem. 2002, 67, 7289-7294. (59) Solladie, G.; Hutt, J.; Girardin, A Synthesis 1987, 173-175. (60) Green, S. D.; Monti, C.; Jackson, R.F.W.; Anson, M.S.; MacDonald, S.J.F. Chem. Commun. 2001, 2594-2595. (61) DiFuria, F.; Modena, G.; Curci, R. Tetrahedron Lett. 1976, 4737-4639. (62) Holland, H. L.; Allen, L.J.; Chernishenko, M.J.; Diez, M.; Kohl, A.; Ozog, J.; Gu. J.X. J. Mol. Catal. B: Enzymatic 1997, 3, 311-324. (63) Williams, D. E.; Ziegler, D.M.; Nordin, D.J.; Hale, S.E.; Masters, B.S.S. Bioche. Biophys. Res. Comm. 1984, 125, 116-122. (64) Rettie, A. E.; Lawton, M.P.; Sadeque, A.J.M.; Meier, G.P.; Philpot, R.M. Arch. Biochem. Biophys. 1994, 311, 369-377. (65) Krueger, S. K.; Yueh, M.; Martin, S.R.; Pereirra, C.B.; Williams, D.E. Drug Metab. Disp. 2001, 29, 693-700. (66) Holland, H. L.; Poopert, H.; Ninniss, R.N.; Chenchaiah, P.C. Can. J. Chem. 1985, 63, 1118-1123.

PAGE 152

135 (67) Ohta, H.; Okamoto, Y.; Tsuchihashi, G. Chem. Lett. 1984, 205-206. (68) Deurzen, M. P. J.; Remkes, I.J.; Rantwijk, F.; Sheldon, R.A. J. Mol. Catal. A: Chemical 1997, 117, 329-337. (69) Okrasa, K.; Falcimaigne, A.; Guibe-Jampel, E.; Therisol, M. Tetrahedron Asym. 2002, 13, 519-522. (70) Beecher, J.; Brackenridge, I.; Roberts, S.M.; Tang, J.; Willetts, A.J. J. Chem. Soc.; Perkins Trans. I 1995, 1641-46. (71) Brink, H.; Holland, H.L.; Schoemaker, H.E.; Lingen, H.; Wever, R. Tetrahedron Asym 1999, 10, 4563-4572. (72) Holland, H. L.; Bornmann, M.J.; Lakshmaiah, G. J. Mol. Cat. B: Enzymatic 1996, 1. (73) Adam, W.; Mock-Knoblauch, C.; Saha-Moller, C.R. J. Org. Chem. 1999, 64, 4834-4839. (74) Ozaki, S.; deMontellano, P.R.O. J. Am. Chem. Soc. 1995, 117, 7056-7064. (75) Boyd, D. R.; Sharma, N.D.; Kennedy, M.A.; Shepherd, S.D.; Malone, J.F.; Alves-Areias, A.; Holt, R.; Allenmark, S.G.; Lemurell, M.A.; Dalton, H.; Luckarift, H. Chem. Commun. 2002, 1452-1453. (76) van Deurzen, M. P. J.; Remkes, I.J.; van Rantwijk, F.; Sheldon, R.A. J. Col. Catal. A: Chem. 1997, 117. (77) Holland, L. H. Chem. Rev. 1988, 86, 478-485. (78) Holland, H. L. Nat. Prod. Rep. 2001, 18, 171-181. (79) Reetz, M. T. Angew. Chem. Int. Ed. 2002, 41, 1335-1338. (80) Wahler, D.; Reymond, J.L. Curr. Opin. Chem. Biol. 2001, 5, 152-158. (81) Reetz, M. T.; Kuhling, K.M.; Hinrichs, H.; Deege, A. Chirality 2000, 12, 479-482. (82) Reetz, M. T.; Elipper, A.; Tielmann, P.; Mynott, R. Adv. Synth. Catal. 2002, 344, 1008-1016. (83) Reetz, M. T.; Zonta, K.; Schimossek, K.; Liebeton, K.; Jaeger, K.E. Angew. Chem. Int. Ed. Engl. 1997, 36, 2830-2832.

PAGE 153

136 (84) Moore, J. C.; Arnold, F.H. Nat. Biotech. 1996, 14, 458-467. (85) Stemmer, W. P. C. Proc. Natl. Acad. Sci. USA 1994, 91, 10747-10751. (86) Suzuki, M.; Christian, F.C.; Kim, B.; Skandalis, A.; Black, M, Loeb, L.A. Mol. Diversity 1996, 2, 111-118. (87) Sahfikhani, S.; Siegel, R.A.; Eugenio F.; Schellenberger, V. Biotech. 1997, 23, 304-310. (88) Kuchner, O.; Arnold, F.H. Trends Biotech. 1997, 15, 523-530. (89) Liao, H.; McKenzie, T.; Hageman, R. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 576-580. (90) Chen, K. A.; F.H. Protein. Eng. 1996, 9, 77-83. (91) Stemmer, W. P. C. Nature 1994, 370, 389-391. (92) Matcham, G. W.; Bowen, A.R.S. Chem. Today 1996, 14. (93) Liebeton, K.; Zonta, A.; Schimossek, K.; Nardini, M.; Lang, D.; Dijkstra, B.W.; Reetz, M.; Jaeger, K.E. Chem. Biol. 2000, 7, 709-718. (94) Arnold, F. H. Nature Biotechnology 1998, 16, 617-618. (95) Moore, J. C.; Jin, H.; Kuchner, O.; Arnold, F.H. J. Mol. Biol. 1997, 272, 336-347. (96) Pearson, B.; Wolf, P.; Vasquez, J. Lab. Invest. 1963, 12, 1249-1259. (97) Joo, H.; Arisawa, A.; Lin, Z.; Arnold, F.H. Chem. Biol. 1999, 6. (98) Joo, H.; Lin, Z.; Arnold, F.H. Nature 1999, 399, 670-673. (99) Vankayalapati, H.; Singh, G. Angew. Chem. Int. Ed. Engl. 1999, 39, 1453-1455. (100) Gee, K. R.; Sun, W.C.; Bhalgat, M.K.; Upson, R.H.; Klaubert, D.H.; Latham, K.A.; Haulgand, R.P. Anal. Biochem. 1999, 273, 41-48. (101) Cehn, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C.J. J. Am. Chem. Soc. 1982, 104, 7294-7299. (102) Volkel, D. W.; F. Ann. NY. Acad. Sci. 1995, 750, 1-9. (103) Schrader, W.; Eipper, A.; Pugh, D.J.; Reetz, M.T. Can. J. Chem. 2002, 80, 626-632.

PAGE 154

137 (104) Klein, G.; Reymond, J.L. Helv. Chim. Acta. 1999, 82, 400-407. (105) Reetz, M. T. Angew. Chem. Int. Ed. Engl. 2002, 41, 1335-1338. (106) Reetz, M. T.; Kuhling, K.M.; Deege, A.; Hinrichs, H.; Belder, D Angew. Chem. Int. Ed. 2000, 39, 3891-3893. (107) Taran, F.; Gauchet, C.; Mohar, B.; Meunier, A.; Valleix, A.; Renard, C.; Creminon, J.; Grassi, A, Wagner, A.; Miokowski, C. Angew. Chem. Int. Ed. Engl. 2002, 41, 124-127. (108) McLane, K.; Fisher, J.; Ramakrishman, K. Drug. Metab. Rev. 1983, 14, 741-799. (109) Denny, R. M.; Fritz, R.R.; Patel, N.T. Science 1982, 215, 1400-1403. (110) Rajagopalan, K. V. Xanthine oxidase and aldehyde oxidase; Academic: New York, 1980. (111) Ziegler, D. M.; Mitchell, C.H. Arch. Biochem. Biophys. 1972, 150, 116-125. (112) Ganther, H. E.; Lawrence, J.R. Tetrehedron 1997, 53, 12299-12310. (113) Taylor, K. H.; Ziegler, D.M. Biochem. Pharma. 1987, 36, 141-146. (114) Ziegler, D. M. Annu. Rev. Pharmacol. Toxicol. 1993, 33, 179-199. (115) Zhou, S. F.; Kestell, P.; PaxtonJ.W. Eur. J. Drug Metab. Pharma 2002, 27, 179-183. (116) Hlavica, P. Drug Metab. 2002, 34, 451-477. (117) Lawton, M. P.; Cashman, J.R.; Cresteil, T.; Dolphin, C.T.; Elfarra, A.A.; Hines, R.N.; Hodgson, E.; Kimura, T.; Ozols, J.; Phillips, I.R.; Philpot, R.M.; Poulsen, L.L.; Rettie, A.E.; Shephard, E.A.; Williams, D.E.; Ziegler, D.M. Arch. Biochem. Biophys. 1994, 308, 254-257. (118) Nelson, D. R.; Kamataki, T.; Waxman, D.J.; Guengerich, F.P.; Estabrook, R.W.; Feyereisen, R.; Gonzalez, F.J.; Coon, M.J.; Gunsalus, I.C. DNA Cell. Biol. 1993, 12, 1-51. (119) Hines, R. N.; Hopp, K.A.; Franco, J.; Saeian, K.; Begun, F.P. Mol. Pharma. 2002, 62, 320-325. (120) Mitsubayashi, K.; Hashimoto, Y. Sens. Actuat. B Chem. 2002, 83, 35-40.

PAGE 155

138 (121) Cashman, J. R. Curr. Drug Metab. 2000, 1, 181-191. (122) Schlenk, D. Comp. Biochem. Physiol. CPharmac. Tox. Endoc. 1998, 121, 185-195. (123) Cholerton, S.; Smith, R.L. N-Oxidation of Drugs: Biochemistry, Pharmacology, Toxicology 1991, 107-131. (124) Lunden, A.; Markland, S.; Gustafsson, V.; Andersson, L. Genome Res. 2002, 12, 1885-1888. (125) Ripp, S.; Overby, L.H.; Philpot, R.M.; Elfarra, A.A. Mol. Pharma. 1997, 51. (126) Cashman, J. R. Chem. Res. Toxic. 1995, 8, 165-181. (127) Krueger, S. K.; Yueh, M.F.; Martin, S.R.; Pereira, C.B.; Williams, D.E. Drug Metab. Disp. 2001, 29, 693-700. (128) Suh, J. K.; Robertus, J.D. Methods Enz. 2002, 348, 113-121. (129) Suh, J. K.; Poulsen, L.L.; Ziegler, D.M.; Robertus, J.D. Arch. Biochem. Biophys. 2000, 381, 317-322. (130) Suh, J. K.; Poulsen, L.L.; Ziegler, D.M.; Robertus, J.D. Proc. Nat. Acad. Sci. U.S.A. 1999, 96, 2687-2691. (131) Gaser, R.; Tynes, R.E.; Lawton, M.P.;Korsmeyer, K.K.; Ziegler, D.M. Biochemistry 1990, 29. (132) Lawton, M. P.; Gasser, R.; Tynes, R.E.; Hodgson, E.; Philpot, R.M. J. Biol. Chem. 1990, 265, 5855-5861. (133) Ozols, J. J. Biol. Chem. 1990, 265, 10289-10299. (134) Dolphin, C.; Shephard, E.A.; Povey, S.; Palmer, C.N.; Ziegler, D.M. J. Biol. Chem. 1991, 266, 12379-12385. (135) Rettie, A. E.; Meier, G.P.; Sadeque, A.J.M. Chem. Bio. Interact. 1995, 96, 3-15. (136) Itoh, K. Biochim. Biophys. Acta 1993, 1173, 165-171. (137) Cherrington, N. J.; Falls, J.G.; Rose, R.L.; Clements, K.M.; Philpot, R.M.; Levi, P.E.; Hodgson, E. J. Biochem. Mol. Toxicol. 1998, 12, 205-212. (138) Lattard, V.; Longin-Sauvageon, C.; Lachuer, J.; Delatour, P.; Benoit, E. Drug Metab. Dispos. 2002, 30, 119-128.

PAGE 156

139 (139) Suh, J. K.; Poulsen, L.L.; Ziegler, D.M.; Robertus, J.D. Arch. Biochem. Biophys. 1996, 336, 268-274. (140) Guan, S.; Falick, A.M.; Williams, D.E.; Cashman, J.R. Biochemistry 1991, 30, 9892-9900. (141) Nikbakht, K. N.; Lawton, M.P.; Philpot, R.M. Pharmacogenesis 1993, 2, 207-216. (142) Dolphin, C. T.; Shephard, E.A.; Povey,S.; Smith, R.L.; Phillips, I.R. Biochem. J. 1992, 287, 261-267. (143) Lattard, V.; Longin-Sauvageon, C.; Krueger, S.K.; Williams, D.E.; Benoit, E. Biochem. Biophys. Res. Commun. 2002, 292, 558-563. (144) Furnes, B.; Somer, S.; Feng. J.; Schlenk, D. direct submission to GenBank, http://www.ncbi.nlm.nih.gov/Genbank/ 2002, 01/10/2003. (145) Karoly, E. D.; Hodgson, E.; Rose, R.L. direct submission to GenBank, http://www.ncbi.nlm.nih.gov/Genbank/ 1999, 01/10/03. (146) Yueh, M. F.; Krueger, S.K.; Williams, D.E. Biochim. Biophys. Acta. 1997, 1350, 267-271. (147) Ozols, J. Arch. Biochem. Biophys. 1991, 290. (148) Falls, J. G.; Cherrington, N.J.; Clements, K.M.; Philpot, R.M.; Levi, P.E.; Rose,. R.L.; Hodgson, E. Arch. Biochem. Biophys. 1997, 347, 9-18. (149) Lomri, N.; Gu, Q.; Cashman, J.R. Proc. Natl. Acad. Sci. USA 1992, 89, 1685-1689. (150) Lattard, V.; Buronfosse, T.; Lachuer, J.; Longin-Sauvageon, C.; Moulin, C.; Benoit, E. Arch. Biochem. Biophys. 2001, 391, 30-40. (151) Nguyen, M.; Karlin-Neumann, G.; Southwick, A.; Tripp, M.; Miranda, M.; Palm, C.J.; Bowser, L.; Jones, T.; Banh, J.; Carninci, P.; Chen, H.; Cheuk, R.; Chung, M.K.; Hayashizaki, Y.; Ishida, J.; Kamiya, A.; Kawai, J.; Kim, C.; Lin, J.; Liu, S.X.; Narusaka, M.; Pham, P. In direct submission to GenBank, http://www.ncbi.nlm.nih.gov/Genbank/ 2002, 01/10/03. (152) Burnett, V.; Lawton, M.P.; Philpot, R.M. FASEB J. 1992, 6. (153) Lattard, V.; Longin-Sauvageon, C.; Benoit, E. In direct submission to GenBank, http://www.ncbi.nlm.nih.gov/Genbank/ 2002, 01-10-03.

PAGE 157

140 (154) Hernandez, D.; Janmohamed, A.; Chandan, P.; Phillips, I.R.; Shephard, E.A. In direct submission to GenBank, http://www.ncbi.nlm.nih.gov/Genbank/ 2002, 01/10/03. (155) AttaAsafoAdjei, E.; Lawton, M.P.; Philpot, R.M. J. Biol. Chem. 1993, 268. (156) Lattard, V.; Benoit, E. In direct submission to GenBank, http://www.ncbi.nlm.nih.gov/Genbank/ 2001, 01/10/03. (157) Overby, L. H.; Buckpitt, A.R.; Lawton, M.P.; Atta-Asafo-Adjei, E.; Schulze, J.; Philpot, R.M. Arch. Biochem. Biophys. 1995, 317, 275-284. (158) Poulsen, L. L.; Ziegler, D.M. J. Biol. Chem. 1979, 254, 6449-6455. (159) Beaty, N. B.; Ballou, D.P. J. Biol. Chem. 1981, 256. (160) Beaty, M. B.; Ballou, D.P. J. Biol. Chem. 1981, 256. (161) Kemal, C.; Bruice, T. Proc. Natl. Acad. Sci. USA 1976, 73. (162) Fujimori, K.; Yaguchi, M.; Mikami, A.; Matsura, T.; Furukawa, N.; Oae, S.; Iyanagi, T. Tetrahedron Lett. 1986, 27, 1179-1182. (163) Kim, Y. M.; Ziegler, D.M. Drug. Metab. Dispos. 2000, 28, 1003-1006. (164) Philpot, R.; Overby, L.; Wyatt, M. ISSX Proc. 1996, 10. (165) Schreuder, H.; Prick, P.; Wierenga, W.; Vriend, G.; Wilson, K.; Hol, W.; Drenth, J. J. Mol. Biol. 1989, 208. (166) Ziegler, D. M. Drug Metab. Rev. 2002, 34, 503-511. (167) Nagata, T.; Williams, D.E.; Ziegler, D.M. Che. Res. Toxicol. 1990, 3, 373-376. (168) Mittle, P. R.; Schultz, G.E. Protein Sci. 1994, 3, 799-809. (169) Fisher, M. B.; Rettie, A.E. Tetrahedron Asymm. 1997, 8, 613-618. (170) Rettie, A. E.; Bogucki, B.D.; Lim, I.; Meier, G.P. Mol. Pharma. 1990, 37, 643-651. (171) Lomri, N.; Yang, Z.C.; Cashman, J.R. Chem. Res. Toxicol. 1993, 6, 425-429. (172) Lomri, N.; Yang, Z.C.; Cashman, J.R. Chem. Res. Toxicol. 1993, 6. (173) Kim, Y. M.; Ziegler, D.M. Drug Metab. Disp. 2000, 28, 1003-1006.

PAGE 158

141 (174) Adjei, E.; Lawton, M.P.; Philpot, R.M. J. Biol. Chem. 1993, 268, 9681-9689. (175) Plots of hydropathy profiles are of use in pinpointing hydrophobic pockets in protein sequences or more particularly in highlighting potential transmembrane domains. As with all analysis methods, it is wise to use a range of hydropathy parameters and use the general trend or consensus to guide your "prediction"; no single method is likely to give an unambiguous profile. (176) Lawton, M. P.; Philpot, R.M. J. Biol. Chem. 1993, 268, 5728-5734. (177) Brunelle, A.; Bi, Y.A.; Lin, J.; Russell, B.; Luy, L.; Berkman, C.; Cashman, J. Drug Metab. Dispos. 1997, 25, 1001-1007. (178) Kubo, A.; Itoh, S.; Itoh, K.; Kamataki, T. Arch. Biochem. Biophys. 1997, 345, 271-277. (179) Flinta, C.; Persson, B.; Jornvall, H.; von Heijne, G. Eur. J. Biochem. 1986, 154, 193-196. (180) Kosmeyer, K. K.; Poulsen, L.L.; Ziegler, D.M. Flavins and Flavoproteins; de Gruyter: Berlin, 1991. (181) Ozols, J. Biochemistry 1994, 33, 3751-3757. (182) Madigan, M. T.; Martinko, J.M.; Parker, J. Biology of Microorganisms, 8th ed.; Prentice Hall International, Inc.: New York, 1997. (183) Fernandez, J. M.; Hoeffler, J.P. Gene Expression Systems; Using Nature for the Art of Expression; Academic Press: San Diego, 1999. (184) Ikemura, T. J. Mol. Bio. 1981, 151, 389-390. (185) Bennetzen, J. L.; Hall, B.D. J. Biol. Chem. 1982, 257, 3026-3031. (186) Burnett, V. L.; Lawton, M.P.; Philpot, R.M. J. Biol. Chem. 1994, 269, 1413-14322. (187) Lomri, N.; Thomas, J.; Cashman, J.R. J. Biol. Chem. 1993, 268, 5048-5059. (188) Neuhard, J.; Nygaard, P. In Escherichia coli and Salmonella typhimurium; Neidhardt, F. C.; Ed.; ASM: Washington, D.C.; 1987. (189) Moffatt, B. A.; Studier, F.W. J. Mol. Biol. 1986, 189, 113-130. (190) Mierendorf, R.; Yaeger, K.; Novy, R. inNovations 1994, 1, 1-3.

PAGE 159

142 (191) Acker, G. http://www.uni-bayreuth.de/departments/emikroskopie/acker24.htm 2002, 02/06/2003. (192) Ikemura, T. J. Mol. Biol. 1981, 151, 389-409. (193) Cole, P. A. Structure 1996, 4, 239-242. (194) Duetz, W. A.; Ruedi, L. Appl. Environ. Micro. 2000, 66, 2641-2646. (195) Shafikhani, S. Biotechniques 1997, 23, 301-310. (196) Caldwell, R. C.; Joyce, G.F. PCR Meth. Appl. 1994, 3, 136-140. (197) Leung, D. W.; Chen, E.; Goeddel, D.V. Meth Cell Mol Bio 1989, 1, 11-15. (198) Caldwell, R. C.; Joyce, G.F. PCR Methods and Applications 1992, 2, 28-33. (199) Bornscheuer, U. T. m. A.; J.; Meyer, H.H. Biotech. Bioeng. 1997, 58, 554-559. (200) Crameri, A.; Raillard, S.A.; Bermudez, E.; Stemmer, W.P.C. Nature 1998, 391, 288-291. (201) Crameri, A. Chimia 1999, 53, 617-620. (202) Xirodimas, D. P.; Lane, D.P. J. Biol. Chem. 1999, 274, 28042-28049. (203) Lorimer, I. A. J.; Pastan, I. Nuc. Acids Res. 1995, 23, 3067-3068. (204) Cline, J.; Braman, J.C.; Hogrefe, H.H. Nuc. Acids Res. 1996, 24, 3546-3551. (205) Arnold, F. H. http://www.che.caltech.edu/groups/fha/ 2002, 01/10/2003. (206) May, O.; Nguyen, P.T.; Arnold, H. Nat. Biotech. 2000, 18, 317-320. (207) Haining, R. L.; Hunter, A.P.; Sadeque, J.M.; Philpot, R.M.; Rettie, A.E. Drug Metab. Disp. 1997, 25, 790-797. (208) Personal communication with Dr. Carlos Martinez, California Institute of Technology. (209) Leive, L. Ann. NY Acad. Sci. 1974, 235, 109-127. (210) Aubin, R. J. Somat. Cell. Mol. Genet. 1988, 14, 155-167. (211) Helander, I. M.; Alakomi, H.L.; LatuaKava, K.; Koski, P. Microbio. UK 1997, 143, 3193-3199.

PAGE 160

143 (212) Bollag, D. M.; Rozycki, M.D.; Edelstein, S.J. Protein Methods; 2nd ed.; Wiley-Liss: New York, 1996. (213) Fumio, T.; Koji, M.; Jyoji, O.; Manabu, N.; Akichika, I.; Kyoto, O.; Kaoru, F. Chem. Lett. 1988, 1, 131-134. (214) Charpin, P.; Dunach, E.; Kagan, H.B.; Theobald, F.R. Tetrahedron Lett. 1986, 27, 2989-2992. (215) Silverstein, R. M.; Bassler, G.C.; Morrill, T.C. Spectrometric Identification of Organic Compoounds; 5th ed.; John Wiley and Sons, Inc.: New York, 1991. (216) Majd, F.; Nickerson, T.A. J. Agri. Food Chem. 1978, 26, 207-210. (217) Lin, Z.; Thorsen, T.; Arnold, F. Biotech. Prog. 1999, 15, 467-471. (218) Eckert, K. A.; Kunkel, T.A. PCR Methods Applic. 1991, 1, 17-24. (219) Harayama, S. Trends Biotech. 1998, 16, 76-82. (220) Zhao, H.; Arnold, F.H. Nuc. Acids Res. 1997, 25, 1307-1308. (221) Cashman, J. R. Biochem. Biophys. Res. Com. 1981, 98. (222) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254. (223) New England Biolabs, I. Catalog and Technical Reference; New England Biolabs, Inc.; 2001. (224) Sambrook, J.; Fritsch, E.F.; Maniatis, T. Molecular Cloning; A Laboratory Manual (2nd ed) 1989.

PAGE 161

BIOGRAPHICAL SKETCH Aris Anup Polyzos was born in Berlin, West Germany, on July 14, 1972, to an East Indian father and Greek mother. Both his mother and father were students, his mother finishing her Diplom Engineer in organic chemistry at the time of his birth. Throughout his early years he got very accustomed to chemistry labs, miscellaneous lab glassware around the house and his mother in a white labcoat. It seemed to him a natural thing to progress throughout elementary school (in Canada and the United States) through high school (in Greece) and on to University. His undergraduate studies he did at the University of Waterloo in Ontario, Canada, in the field of biochemistry (with a minor in computer science). This allowed him to combine chemistry, which had a comforting familiarity to him, with biology, which connected him to a long fostered fascination with the natural living world. After a degree in which he learned organic chemistry and molecular biology in distinct separate spheres he chose to attend the University of Florida to pursue of a doctoral degree in the laboratories of Dr. J. Stewart. The research in those laboratories was a melding of current molecular biological techniques with the need in organic chemical synthesis for specific catalysts. Upon completion of his studies with Dr. J. Stewart Aris will be working as a postdoctoral fellow with Dr. Carole Yauk at the Ministry of Health in Ottawa, Canada. This upcoming research uses molecular biological techniques on questions concerning human health (identifying cancer-causing agents and in the assessment of consumables from new technologies). 144


Permanent Link: http://ufdc.ufl.edu/UFE0000769/00001

Material Information

Title: Directed Evolution of a Sulfoxidation Biocatalyst
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000769:00001

Permanent Link: http://ufdc.ufl.edu/UFE0000769/00001

Material Information

Title: Directed Evolution of a Sulfoxidation Biocatalyst
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000769:00001


This item has the following downloads:


Full Text












DIRECTED EVOLUTION OF A SULFOXIDATION BIOCATALYST


By

ARIS A. POLYZOS
















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2003




























Copyright 2003

by

Aris A. Polyzos



























This work is dedicated to the strong women in my life. Without you I would not be half
of what I am now. My (soon to be) wife April Flanders, my mother Penelope Polyzou,
my sister Cassandra Polyzou and all the little furry ones.















ACKNOWLEDGMENTS

I would like to give my gratitude to everyone who helped and encouraged me through

these years and made the completion of this work possible.

During my career as a graduate student in the Chemistry Department I was blessed

with many mentors, colleagues and wonderful friends. I would like to thank Dr. J. Stewart

for the opportunity to work in his group and for his advice and guidance with my doctoral

research. The genes and plasmids graciously provided throughout the course of this work by

Dr. Philpot, Dr. Williams and Dr. Cashman have also been invaluable as they have been the

main subjects of this study. In providing help both by guidance and physically providing us

with many of the supplies for this project I must make a special note of thanks to my friend

Dr. Michael Thomson, one of the most capable molecular biologists I have come across, who

is now in Cold Springs Harbor Laboratories. Among the small biochemistry group in the

Chemistry Department of UF there have been several whom I am very glad to have met. Dr.

B. Horenstein and Dr. Tom (Lyons) have provided insights at times of scientific

consternation and have also been guides through the non-scientific problems one can find in

a research project. Dr. R. Geyer, now a faculty member at the University of Saskatchewan,

also helped with advice and expertise in questions of molecular biology and in reminding me

of home.

As a member of the Stewart group I have made many wonderful friendships with

people still close to me, even if they are now dispersed around the globe. Sonia Rodriguez is

an individual I will always respect, having been both a labmate and a guide in maintaining









my beliefs and upstanding moral judgment and in exemplifying thorough scientific inquiry.

Carlos Martinez, who I am overjoyed to say is very happy now, has always been the source

of information on matters of science and an example of quiet and undaunted work no matter

what the situation. For 4 years Jennifer Tonzello, who is also happily living out West now,

was a big part of my life along with Sonia and Carlos. She was most evidently the one who

brought heart and cheer into the laboratory when it was most lacking. Marko Mihovilovic,

who should expect my visit sometime soon to his newly built home in Austria, was the

impressive postdoc, having both a very thorough appreciation for scientific work balanced

with enough of a zest for life to be always cheerful. Recently Despina Bougioukou, a

compatriot of genes and of spirit, has been here beside me in the laboratory. I am completely

confident that her scientific undertakings will be thorough and well thought-out and expect

she will do very well in her research. I will also always be in touch and seriously invite her to

the Great White North to visit us. I was fortunate to spend my time in the laboratory with

many other people who made the research setting the eclectic mixture that it was. Kavitha

Vedha-Peters, Erin Ringus, Mirella Stefans, Lisa Manning, Abhijit Roy-Chundry, Harch Ooi,

Bart Neff, Kersten Schroeder, Matt Carrigan, Lee Raley, Stefan Lutz, Jun Zhu, Darwin Ang,

Kim Millar, Debbie Burroughs, Catherine Charron, Tricia Pokey, Michael Sismour, Alonso

Ricardo, Leslie Tuchman, Iwona Kaluzna, Brian Kyte, Brent Feske and many others have all

played their part in this wonderful experience.

Throughout my studies from almost the first year at UF I am thankful to have crossed

paths with Benjamin Lopman, initially as the intelligent undergraduate in a chemistry lab I

was overseeing, to later becoming my closest friend, fellow athlete and most recently as a

fellow graduate student. He and his wife, Leah Garces, have been as close as family since we









first met and our relationship has fed off each other's zeal and curiosity for all the wonders

that this world has to offer. They have been close throughout my studies even from their nest

way across the pond.

My support comes from up north too and is also unwavering. I owe so much to my

mother Penelope and sister Cassandra Polyzou as to not be able to fit my thanks into an

entire novel much less in a few pages at the beginning of a thesis. They are as much a part of

me as my own thoughts and I am glad to be moving closer to them soon.

Vincent Vadez and Rachid Serraj have been the two French academics who breathed

the joy of life into me in my early years here and their thoughtfulness and compassion are

always remembered. I wish them luck in India and in the jungles of Bolivia where they are

now using their academic knowledge for practical purposes.

I would also thank the University of Florida cycling team. Winning the SEC

Conference Championships and placing within the top teams nationally on whatever type of

bike one could imagine these last two years are not the reasons. The reasons are the very

good friends, the reaching of new physical limits almost on a weekly basis and the incredible

scenery that I have shared with some very hard working people.

The cycling team has also been the group that made possible my meeting of April

Flanders, who is the main factor for me being here in one conscious piece. She is my partner

and guide. She is the cycling art professor that stole me away and as we speak now I cannot

imagine what it would have been like had we never met. It could never get this good.
















TABLE OF CONTENTS
page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES ..................................................................... .... .... x

LIST OF FIGURES ......... ........................................... ............ xi

ABSTRACT ........ .............. ............. ...... ...................... xvi

CHAPTER

1 INTRODUCTION AND BACKGROUND ..................................... ...............

Enantiopure Sulfoxides in Synthesis.................................. ....................... 2
A approaches Tow ards Sulfoxides...................................................... .................9
Routes to Racemic Sulfoxides, Classical Organic Chemistry........................9
Routes to Non-racemic Sulfoxides, Classical Organic Chemistry .............10
Enzymatic Routes to Chiral Sulfoxides............... ...................................... 15
D directed Evolution of Enzym es.................................... .................................... 20
Theory and A approaches ............................. ................. ............... .... 20
High Throughput Screening for Enantioselectivity ...................................23
Flavin-M onooxygenases......................................................................... 29
B background and Properties........................................... ........................... 30
Catalysis by FM O 's .................. ................................ .. .... .. .. ........ .... 34
Structure of FM O 's.................... .. .. ................................. .... .. .............. 37
Strategy Towards A Stereoselective Sulfoxidation Catalyst, An Overview of this
W o rk ..............................................................................4 2

2 ENGINEERING OF MICROORGANISMS FOR FMO PRODUCTION ................44

Expression U sing E scherichia coli ............................................................................ 46
Expression of Active FMO Enzymes ............ .............................................47
Rabbit FM O1 and Rabbit FM O3 ..... ......... ....................................... 49
Rabbit FMO2 ......... ... ........... ......... 50
R hesus m acaque FM O 2 ............................................. ....... ............... 51
Improving the Expression of Active FMO .............................................. 53

3 GENERATING LIBRARIES OF MUTANT FMO'S ................................ .........58

Error-Prone PCR; Controlled Random M utagenesis....................... .....................58









DNA Shuffling and Recombination Mutagenesis ............................................... 66

4 BIOCATALYSIS REACTIONS AND SCREENING................... ....... .........71

Sulfoxidation Using FM O Biocatalysts.............. ................................................... 72
Screening Biocatalysis R actions ............................... .................. ..... ............... 77
Screening Biotransformations Using UV/VIS and OR Detectors.......................78
D developing M ethodology.................. ................... ..... ... ............... 78
Assessment of UV/ORD Screening for Enantioselectivity ..........................84
Screening Biotransformations Using Chiral HPLC ...........................................85
Developing Methodology.... .................... ..................85
Nonenzymatic Racemic Sulfoxidation...............................................88

5 DIRECTED EVOLUTION OF FMO'S TOWARDS EFFICIENT (S)-
SULFOXIDATION CATALYSTS ................... ......... ........................ 93

R results of Screening A ssay s............................................................ .....................93
Conclusions and Future W ork ........................................................ ............... 101

APPENDIX

A EXPERIMENTAL PROCEDURES......................................................................103

Reagents and Supplies ....................................................... ... .. .... 103
Purification of Pfu Polym erase.......................................................... ............... 104
Error-Prone PCR of FM O ........................................................................ 105
D N A Shuffing...................... .......... .. .... ........ .............. .. 106
Fragmentation of Genes ..................................... ............... 106
Reassembly of DNA Fragments ........................................... ................... 107
Amplification of the Shuffled Genes...................................... ............... 107
Assay for FMO Activity; Oxidation of Thiobenzamide.....................................108
Crude Membrane Preparation Activity Assay; Oxidation of Methimazole and
M T S .................. .......... .... .............. ... ........ ............... ............ 109
Synthesis of Starting material and Sulfoxide Standards ................ ................110
A ryl A lkyl Sulfides ........................... ........ ........ ........ ........ .... 110
Racemic M ethyl p-tolyl Sulfoxide (M TSO)................... ..................................111
A lky l p -toly l Su lfox ides .................................................................................... 11
PCR Prim ers ....................................................................... ........ 112
S train s ................... .........................................................................1 13
Recombinant DNA Techniques ................................................... ............... ... 114
H PL C A analysis ............. ..... ........ ..........................................115









B EXPERIMENTAL DATA AND MISCELANEOUS......... .............................122

C GLOSSARY OF TERMS AND ABBREVIATIONS ................ ......... .........130

L IST O F R E F E R E N C E S ......... .................................... ............................................. 13 1

BIOGRAPH ICAL SKETCH .............. ............................ ................... ............... 144















LIST OF TABLES


Table p

1. Classical organic chemistry sythesis of enantiopure MTSO. ........................................15

2. Methods for the stereoselective sulfoxidation of methyl p-tolyl sulfide (MTS). Various
classical synthetic organic methods as well as isolated enzymes and whole cell
systems could be used to produce M TSO. ..................................... ............... 17

3. List of identified and putative FM O's to date ............... ........................ ............... 32

4. Various expression systems for rabbit and rhesus macaque FMO's in Escherichia coli
and Saccharomyces cerevisiae. ........................................ .......................... 48

5. Catalytic activity of Escherichia coli expressed rabbit FMO 1 fused with various
peptides at either the N- or C- terminal ends. .................................. .................55

6. Mutagenic frequency of EP-PCR reactions with various concentrations of Mn2+. EP-
PCR libraries (#C, D, E, F, G) generated from wild type rabbit FMO1..................62

7. Mutagenesis and sulfoxidation activities of EP-rFMO1-J library.............................64

8. Mutagenic frequency of EP-PCR reactions with various concentrations of Mn2+. EP-
PCR libraries (#L, M, N, O, P) generated from wild type rabbit FMO1. ...............64

9. Sulfoxidations of a variety of aryl-alkyl sulfides by purified FMO's (ip) and crude-
membrane preparations of heterologously expressed FMO's (cmp). ...................73

10. Sulfoxidation of an array of alkyl aryl sulfides catalyzed by crude membrane
preparations of Escherichia coli expressing rabbit FMO 1 ......................................75

11. Various strategies to remove background racemic sulfoxidation of MTS. ...............90

12. The cost of screening a library of singly mutated FMO (10 000clones) for
sulfoxidation activity (enantioselectivities and yields). ......................................... 91

13. The strains of Escherichia coli used in this study and their genotypes .....................113

14. The strains of Saccharomyces cerevisiae used in this study and their genotypes ..... 114















LIST OF FIGURES


Figure p

1. Chirality of unsymmetrically substituted sulfoxides. ................... ................... .......... 2

2. The active (S) enantiomer of omeprazole. ............................................ ............... 2

3. Chiral sulfoxide mediated route from esters to enantiomerically pure alcohols6. .........3

4. Methyl methyl sulfoxide used in the sythesis of ketones. .............................................4

5. The retrosynthesis of mevinic acid-type hypocholestemic agents using enantiopure
M TSO as a chiral auxiliary. ........................................................................... 5

6. Total sythesis of the rubiginone antibiotic incorporating the chiral auxiliary (R)-
M T S O ............. ........... .. .......... ... ................................................... 5

7. The retrosynthetic scheme towards Panamycin 607 which incorporates an asymmetric
reduction of chiral 3-ketosulfoxide derived from (S)-MTSO. ..................................6

8. The retrosynthesis of (-)-Centrolobine, indicating the chiral sulfoxide mediated key
diastereoselective ketone reduction ......................................................................... 7

9. Stereoselective total synthesis of an enantiomerically pure fluorinated analogue of
tetrahy droiso qu in olin e...................................................................... ................

10. Retrosynthetic scheme for the synthesis of 2-fluoro analogues of frontalin indicating
the stereodirecting effect of the chiral sulfoxide auxiliary in the formation of the
epoxide 6 ............................................................................. 8

11. A facile synthesis of D-erythrose employing (R)-MTSO for stereoselective alkylation
of an aldehyde. ......................................................................... 8

12. The synthesis of enantiopure (S) benzyl alcohol using chiral (S) Methyl 1-naphthyl
su lfo x id e ........................................................... ................ 9

13. The most prevalent current synthetic routes towards chiral sulfoxides....................10

14. A modification of the Kagan-Sharpless oxidation of sulfides using (R)-(+)-binaphthol
in place of D ET ........... .. .............. ......... .. .......... ..... ........ ............... 12

15. Chiral oxazidines as sulfoxidation catalysts. ..................................... ............... 12









16. The Andersen synthesis of chiral p-tolyl sulfoxides. The route towards the antipodal
sulfoxide is identical but requires different separation techniques to afford the (R)-
m enthyl ester. ...................................................... ................. 13

17. Kinetic resolution in the Andersen synthesis, recrystallization in acetone-HC ..........13

18. Several chiral sulfinyl transfer reagents for the enantioselective synthesis of
sulfoxides. ...........................................................................15

19.Sources of large libraries from which enantioselective catalysts can be explored.......20

20. Enzymatic reactions can produce highly chromophoric and fluorogenic molecules
that can be used to screen for activity. ........................................ ............... 24

21. An enantioselective mutant PAL catalyses the kinetic resolution of a racemic p-
nitrophenol ester. .......................................................................25

22. The reaction catalyzed by hydantoinase, which has distinct enantiomeric substrates
(the inherent racemization between the 5-monosubstituted hydantoin isomers is
slow ). ................................................................................ 2 6

23. The enantioselective reduction of a coumarinyl derivative of 2-butyl ketone can be
monitored using high-throughput fluorescence screening for the reverse oxidation
reaction coupled with elimination to form a fluorophore product .........................27

24. Screening for enantioselectivity using competitive enzyme immunoassays ..............28

25. The catalytic cycle postulated for pig liver FM O's. ............. .................... .........36

26. Model of the active site of rabbit and pig FMO 1 and rabbit FMO2 proposed by
Cashman and Rettie showing the orientation substrates propyll 2-naphthyl sulfide,
methyl 2-naphthyl sulfide, MTS, heptyl p-tolyl sulfide) would adopt in the active
site ........................................................ ................................. 3 8

27. The topography of the -535 amino acid protein of rabbit FMO1, 2, 3 ....................39

28. Differences in the enantioselectivity of sulfoxidation by purified FMO's from
different sources with increasing alkyl chain length of alkyl p-tolyl sulfides. ........40

29. An overview of the strategy for the directed evolution of an FMO enzyme as a
catalyst for the efficient production of (S) MTSO, a molecule with many uses in
org anic sy nth eses ............................................... ................. 4 3

30. General schematic of an Escherichia coli expression plasmid vector.......................49

31. Expression of rabbit FMO 1, expected at- 59kDa, from BL21(DE3)/pAAP11
analyzed with SDS-PAGE............. .... ......... ..... ............... 50

32. Electron-micrograph evidence of lack of inclusion bodies ................................. ...52









33. Growth rates of recombinant Escherichia coli cells..............................................56

34. Error-prone PCR and cloning into expression systems. Indicated are the possible
sources of wild-type gene which can carry over into final mutagenized library. ....60

35. Activities of rabbit FMO 1 clones (EP-rFMO -J)................................... ..............63

37. Error-prone PCR and cloning into expression system. Strategy to remove wild-type
contamination from the final mixtures. nm = non-methylated DNA.....................65

38. Multiple sequence alignment of Rabbit FMO 1 and Rhesus macaque FMO2 genes
(DN A ) (ClustalW ) .................. .......................................... ........ .... 67

39. Multiple sequence alignment of Rabbit FMO 1 and Rhesus macaque FMO2 proteins
(ClustalW )............................................................... .... ..... ........ 68

40. General overview of DNA Shuffling ........................................ ....................... 69

42. Biotransformation of MTS by FMO presented as crude membrane preparations from
Escherichia coli based expression systems. ....................................................... 76

43. Chiral amide, forms crystals with molecules of (R) -MTSO (1 molecule amide: 1
m molecule sulfoxide). .............................................. .... ........ ......... 78

44. The addition of two auxochromic substituentspara to each other and charge transfer
resonance forms influence the absorption characteristic of phenyl rings. ...............79

45. Comparative UV/VIS spectra of biotransformation reactions on various substrates.. 79

46. UV spectrum of MTS and MTSO, indicating the distinct absorption maxima; MTSO
(233nm ) and M TS (254nm ). ............................................ ............................ 80

47. Cell extract background in the biotransformations........................... ..................81

48. A comparison of the UV and ORD signals of MTS and MTSO in various solvents..83

49. Overview of the method to screen, using UV and OR measurements for the
enantioselectivity of biocatalysis reactions. ............. ...............................................84

50. Values of % ee. for samples measured using the UV/ORD screen ...........................85

51. Chiral HPLC chromatographs showing separation of a series of racemic alkyl p-tolyl
sulfoxides ........................................................................ .. ....... ....... 86

52. An overview of the use of HPLC adapted to determine reaction yields of biocatalysis
reactions. ............................................................................87

53. The chromatograms of MTS sulfoxidation reactions (0.6|jmole scale ) of lysed rabbit
FM O 1 expressing cells ............................................... .. ...... .. ............ 88









54. Background racemic sulfoxidation of M TS................................. ............. ........... 89

55. Values of % ee. for samples measured using a chiral HPLC column........................91

56. Determination of enantiomeric composition of reactions using a chiral HPLC column.92

57. Sequencing the mutant rabbit FMO 1 genes from the EP-PCR generated library
involves 3 primers flanking and internal to the gene itself. ...................................94

58. The activity of C) clone 176E70, compared to B) the wild type activity (rabbit
FM O 1) and the A) background air oxidation .............. ........................................95

59. Rescreening of 13 clones of rabbit FMO1 for stereoselectivities.............................95

60. Extrapolated times required for each enzyme mutant to oxidize 2mM of MTS..........96

61. The activities of 4 clones with higher activities than the wild type ..........................97

62. Sulfoxidation stereochemistries and yields catalyzed by -4000 various singly mutated
clones of rabbit FMO1 (library EP-PCR-FMO1-S+T)...........................................99

63. Sulfoxidation stereochemistries and yields catalyzed by -1100 various chimeras of
rabbit FMO 1 and Rhesus macaque FMO2 (library Sh-FMO 1/2-C)...................... 100

64. Postulated tertiary ribbon structure of a dimer of human FMO3.............................102

65. Pfu polymerase (prepared) vs. commercial (Fisher Scientific) Taq polymerase
amplification ............... .......... .................. ... ....... 105

66. Fragmentation of the FMO gene with DNase I run on a 2.5% agarose gel.............107

67. Reassembled and amplified FMO gene chimeras loaded on a 0.8% agarose gel......108

68. Construction of pAAP20 plasmidd for expression of rabbit FMO2 under the control of
Ptrc) and pAAP 15 plasmidd for expression of rabbit FMO2 under the control of PT7)116

69. The construction of pAAP18 plasmidd for expression of rabbit FMO4 under the
control of PT7) and pAAP 14 plasmidd for expression of rabbit FMO3 under the
control of PT7) ........................................................ ................... .............. 117

70. The construction of pAAP12 plasmidd for the expression of rabbit FMO2 under the
control of Pac), pAAP 10 plasmidd for expression of rabbit FMO2 in yeast) and
pAAP21 plasmidd for the expression of rabbit FMO2 under the control of PT7 with
k an am y cin selection) ......................... ........ .. .... .. .............. .... .............. 118

71. Construction of pAAP1 1 plasmidd for the expression of rabbit FMO 1 under the
control of PT7 with ampicillin selection), pAAP27 plasmidd for the expression of
rabbit FMO 1 under the control of PT7 with kanamycin selection) and pAAP38
(template for the error prone PCR reactions to generate clones of rabbit FMO1). 119









72. Construction of pAAP35 plasmidd for the expression of rhesus macaque FMO2 under
the control of PT7 with ampicillin selection), pAAP36 plasmidd for the expression
of rhesus macaque FMO2 under the control of PT7 with kanamycin selection) and
pAAP38 (template for the error prone PCR reactions to generate clones of rhesus
m acaque FM O 2). ........................ ........................ .. .... ..... .. ............120

73. Construction of pAAP46 plasmidd for the expression of rabbit FMO 1 with a C-
terminal His6 tag under the control of PT7), pAAP41 plasmidd for the expression of
rabbit FMO 1 with a C- and N-terminal His6 tag under the control of PT7) and
pAAP29 plasmidd for the expression of rabbit FMO 1 fused to MalE at the N-
terminus, under the control of P ). ............................................ ............... 121

74. Standardization measurement of UV233nm for MTS and MTSO indicating the error
apparent in each reading. Samples were in solutions of cell extract and the
background was corrected from the relative absorbance at 280nm. The upper graph
indicates the non-linearity of measurements of concentrations >0. lmM .............. 123

75. Standardization measurement of UV254nm for MTS and MTSO indicating the error
apparent in each reading............................................... .............................. 123

76. Standardization measurement of OR for MTS and MTSO indicating the error
apparent in each reading ........... .................................................. ............... 124

78. The sequence of clone 170AA21 of rabbit FMO1. Upper sequence is DNA, beneath it
is the predicted amino acid sequence. ....................................... ............... 125

79. The sequence of clone 176C34 of rabbit FMO 1. Upper sequence is DNA, beneath it
is the predicted amino acid sequence. ....................................... ............... 126

80. The sequence of clone 170D96 of rabbit FMO1. Upper sequence is DNA, beneath it
is the predicted amino acid sequence. ....................................... ............... 127

81. The sequence of clone 170D62 of rabbit FMO 1. Upper sequence is DNA, beneath it
is the predicted amino acid sequence. ....................................... ............... 128

82. The sequence of clone 176E70 of rabbit FMO1. Upper sequence is DNA, beneath it is
the predicted amino acid sequence ........................... ............. ............... 129















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

DIRECTED EVOLUTION OF A SULFOXIDATION BIOCATALYST
By

Aris A. Polyzos

May 2003
Chair: J. Stewart
Major Department: Chemistry


The chirality of sulfoxides has been exploited in many organic syntheses. The

persistent use of non-racemic sulfoxides directly or as chiral relay agents in organic

synthesis has spurred the development of improved routes to each sulfoxide enantiomer.

The long term goal of this project is to develop a biocatalytic system for quick and

selective production of both stereoisomers of a chiral sulfoxide (methyl p-tolyl

sulfoxide). The directed evolution of the enzyme flavin-containing monooxygenase was

used to develop this biocatalytic system.

FMO enzymes from rabbit (FMO1) and rhesus macaque (FMO2) have opposite

stereochemical preference for the sulfoxidation of methyl p-tolyl sulfide. They have been

cloned into an Escherichia coli based expression and biocatalytic system. Libraries of

mutants of these FMO's have been developed for screening in order to identify an

enzyme that produces high yields of the (S) sulfoxide. Genes with single amino acid

mutations were generated from rabbit FMO1, which was also shuffled with the monkey

enzyme to generate a library of chimeric genes.









These libraries were screened for enantioselectivity in the sulfoxidation of methyl

p-tolyl sulfide. A screening method developed using a flow cell optical rotation and a UV

detector proved slow (<100 clones/day) and had low sensitivity for identifying

differences in enantiomeric compositions of reactions. Instead, using chiral HPLC the

activities of 6000 clones were tested (6min/clone) for the sulfoxidation of methyl p-tolyl

sulfide. From both the mutant libraries one clone (176E70) displayed an increase in

selectivity for the (S) sulfoxide of 30% ee. and there were 4 clones which displayed up to

35% increased activity. Of the two directed evolution strategies the single mutation

library of FMO1 yielded interesting clones while none were observed from the chimeric

enzymes.














CHAPTER 1
INTRODUCTION AND BACKGROUND

Chirality in chemical compounds has been known since the 1840's when Louis

Pasteur observed that he could separate the crystals of the different isomeric forms of

tartaric acid (which combined were known as racemic acid). The importance of organic

chiral substances and their interactions with biological materials has become ever more

significant. In 1999 the annual sales of chiral drug molecules topped $100 billion for the

first time. There are several reasons chirality has affected the drug industry so greatly. In

large part it is because biological messenger molecules and cell surface receptors are

chiral and the drug molecules that medicinal chemists design and the pharmaceutical

industry produces must have corresponding asymmetry. Another reason is simple cost

effectiveness. The Food and Drug Administration in the U.S. requires that each

enantiomer of a drug being sold be tested in detail separately1 which doubles the cost of

screening racemic mixtures.

There are also many examples of chiral drug molecules in which one enantiomer

is pharmacologically active and others are inactive or even harmful.2 The (R)-(+) form of

thalidomide can be used by pregnant women to treat morning sickness, whereas the

(S)-(-) enantiomer is associated with startling numbers of baby deaths and birth

defects as well as peripheral neuritis. Another example is ibuprofen, whose (S)-

enantiomer is a common anti-inflammatory drug, whereas (R)-ibuprofen is inactive.









Unsymmetrically Substituted Sulfoxides; Chiral Molecules

There are steric and stereoelectronic differences between 4 substituents of an

unsymmetrically substituted sulfoxide, a highly configurationally stable moiety (Figure

1).3 The lone electron pair, the oxygen and the two distinct carbon groups provide a

chiral environment rich in its chemistry.











Figure 1. Chirality of unsymmetrically substituted sulfoxides.

Enantiopure Sulfoxides in Synthesis

The best selling drug of all time, Astra-Zeneca's anti-ulcer omeprazole

(Prilosec/Losec), is itself a chiral sulfoxide.4 Initially marketed as a racemate, it is the (S)

enantiomer (Figure 2) (with its own generic name of esomeprazole) that has

pharmacoactivity and displays the least side-effects.


H

H3CO" N CH3

H3C OCH3
esomeprazole
Figure 2. The active (S) enantiomer of omeprazole.

There are several examples of synthetic chiral sulfoxide targets5 although the

interest in the chirality of sulfoxides to synthetic and medicinal chemists extends even

further than the molecules themselves, into their use as chiral relay agents in syntheses.6










The electron deficient character of the sulfoxide sulfur can be used to stabilize

adjacent carbanions and many applications of sulfoxides in synthesis involve the reaction

of sulfur-stabilized carbanions with electrophiles.7,8


Solladie and Carreno have thoroughly extended the usefulness of sulfoxides into

asymmetric synthesis.9 p-ketosulfoxides, which can be readily prepared from esters

coupling to a chiral sulfoxide a-carbanion, can be reduced to either diastereomer

selectively by using DIBAL (diisobutylaluminum) or ZnC12/DIBAL. Desulfurization

with Raney nickel affords corresponding enantiomerically pure carbinols (Figure 3).6


O O
+ 0 ZnCI2
O /S'Ar O: DIBAL 0
II .- + OH -. + Raney nickel OH
R OCH, ^ K- -- -
R OCH3 R 'Ar R S'Ar R
S(S) (R) S(S) (S)




DIBAL OH # 0 Raney nickel OH
R S,'Ar R
(S) S(S) (R)

Figure 3. Chiral sulfoxide mediated route from esters to enantiomerically pure alcohols6.

As mentioned above an unsymmetrically sulfoxide can act as a chiral reagent in

total syntheses. In the past decade (1993-2003) chiral sulfoxides such as methyl p-tolyl

sulfoxide (MTSO) have served as chiral auxiliriaries in several total sytheses of natural

products and pharmacochemicals.


Deprotonation a to a sulfoxide can be accomplished with bases such as NaH (in

DMSO) and produces a strong base and a reagent able to generate ylides (such as Wittig

reagents) and which reacts with esters to produce enolates of P-ketosulfoxides.










Subsequent desulfurization of p-keto sulfinyl compounds with aluminum amalgam

provides a synthesis of ketones (Figure 4).

0 [Na]

11 HNa, DMSO CH 30' R 0 0
H3C ,CH3 H3C CH2 H3Cs R
[Na]

R1
SX


0 AI(Hg), H20 O
R I H3c y R
R1 R1

Figure 4. Methyl methyl sulfoxide used in the sythesis of ketones.

Non-racemic MTSO produced by biocatalysis using Saccharomyces cerevisiae

(yeast) was used as a precursor in the total synthesis of mevinic acids (Figure 5). 10

Considerable interest in these compounds arises from their ability to lower the level of

cholesterol in blood plasma by inhibiting enzymes involved in the biosynthesis of

cholesterol. In this approach the chiral sulfoxide directs the chirality of the p-hydroxy-8-

lactone moiety, the key structural feature in all HMG-CoA inhibitors. 11 The chiral center

at C-3 of the lactone ring is determined during an aldol condensation of the a-cabanion of

the sulfoxides with an aldehyde. The a-sulfinyl ester can be later removed by

desulfurization using aluminum amalgam. 12










COMPACTIN, MEVINOLIN, PROVASTATIN '
O O

] o C HO 0 K0

H O H O H
O, HO






R = H (+)- compactin Saccharomyces cerevisae
R= CH3 (+) mevinolin : S
R = OH (+) provastatin 60% yield, 92% ee (R) .,.S
I IR-MTSO

Figure 5. The retrosynthesis of mevinic acid-type hypocholestemic agents using
enantiopure MTSO as a chiral auxiliary.10

The angucycline antibiotics such as the various rubiginone derivatives, also called


fujianmycins, are claimed as useful in the treatment of AIDS and Alzheimers13 and

exhibit potentiation of vincristine-induced cytotoxicity against multidrug-resistant tumour


cells. 14 Carreno et al. 15 utilizes a chiral sulfoxide moiety twice in the total synthesis of

these molecules (Figure 6). (R)-MTSO was incorporated into 1 and directed the highly

chemo- and diastereoselective conjugate addition of Al(CH3)3 to form 2.


RUBIGINONE
OHO CH3 TBDMSO CH3

OR p>Tol "OCOiPr


OCH3 OCH3 O
R = H, CO'Pr i


AI(CH3)3, CH2CI2
0 :O -780C, 44hrs O CHS )pT
+^ S(O)pTol
OH \ 65%, 98% d.e. OH
1 pTol 2

Figure 6. Total sythesis of the rubiginone antibiotic15 incorporating the chiral auxiliary
(R)-MTSO.









The diastereoselective reduction of p-keto sulfoxide with DIBAL-H was

incorporated by Solladie et al. 16 into the total synthesis of panamycin (Figure 7) and in

the enantioselective synthesis of centrolobine (Figure 8). The crystalline natural product

centrolobine, isolated from the stem of the amazonian plant Brosinum potabile in 1964,

required enantioselective synthesis to establish unequivocally its absolute configuration.

The approach by Colobert et al.17 was the first strategy to achieve this synthesis.

Condensation of (R)-MTSO with glutaric anhydride affords the P-keto sulfoxide 3, which

can in turn be reduced stereoselectively to the diastereomerically pure hydroxyketone 4,

again directed by the incorporated sulfoxide.

PANAMYCIN 607



0 0 t" o'tBuO
CH3N O HO





0 : O 1) DIBAL-H
0 0 A\ + 2) oxalic acid, THF, H20 OH
VSpTo pTol
80%, >95% d.e.
Figure 7- The retrosynthetic scheme towards Panamycin 60716 which incorporates an
asymmetric reduction of chiral P-ketosulfoxide derived from (S)-MTSO.










CENTROLOBINE


0 OH 0

Jj~If ISpToI
CH30 OH CH30 4

DIBAL/ZnBr2
THF
80%, 98% d.e.


Spol

CH,30S

Figure 8. The retrosynthesis of (-)-Centrolobinel7, indicating the chiral sulfoxide
mediated key diastereoselective ketone reduction.

Chiral sulfoxide auxiliaries can also exert stereodirecting effects in Pictet-

Spengler reactions.18 Bravo and co-workersl9 used (S) MTSO as an auxiliary to direct

the ring closure of 5. The attack of the 3,4-dimethoxyphenyl group was directed to the

less hindered Re face of the stabilized carbocation formed by imine protonation of 5 with

TFA. An example of this stereocontrol was employed in the total synthesis of a novel

non-racemic 1-trifluoromethyl analogues of tetrahydroisoquinoline (Figure 9).19

1-TRIFLUOROMETHYL TETRAHYDROISOQUINOLINE ALKALOID



CH3O CH30 TFA, 0C, CHCI3 CH30

CH3O NCH3' CH30 N'H 74%, 71% e.e. CH30 I
F3 CH3OH F3 F3
S 5 \
O pol pTol

Figure 9. Stereoselective total synthesis of an enantiomerically pure fluorinated analogue
of tetrahydroisoquinoline. 19










In the synthesis of a natural product analogue Bravo and co-workers again applied

the chirality of the (R)-MTSO but this time to direct methylene insertion from

diazomethane into the carbonyl group of a P-keto sulfoxide 7.20 Modest

diastereoselectivities were improved by separation using flash chromatography to obtain

the intermediate in the synthesis of 2-fluoro analogues of frontalin (Figure 10), an

analogue of the bioactive component of the aggregation pheromone of pine beetles.

2-FLUORO ANALOGUE OF FRONTALIN


H O C H F CH2N2, CHO3H, 00C H F
O CH3 pTo, .: pTok, -
F 00 95%, 85% d.e. I o
0 6 0 7
Figure 10. Retrosynthetic scheme for the synthesis of 2-fluoro analogues of frontalin20
indicating the stereodirecting effect of the chiral sulfoxide auxiliary in the
formation of the epoxide 6.

Arroyo-Gomez and coworkers used (R)-MTSO in the diastereoselective

alkylation of aldehyde 8.21 Due to the matched asymmetric induction from chirality in

both the aldehyde and the nucleophile such a reaction proceeded with complete

diastereoselectivity affording a precursor for a facile synthesis of D-erythrose (Figure 11).

D-ERYTHROSE DERIVATIVE


oH 04- LDA
\OH ,> O pToI O- H + H ,, spTo
OH H n 70%, 99% d.e.
OH OBn OH O O O
8
2-O-Benzyl-(a+p)-D-erythrofuranose
Figure 11. A facile synthesis of D-erythrose employing (R)-MTSO21 for stereoselective
alkylation of an aldehyde.

Addition of the carbanion of (S) Methyl 1-naphthyl sulfoxide to n-alkyl phenyl

ketones occurs with remarkably high diastereoselective excess. Raney nickel is used to

remove the chiral sulfoxide auxiliary and produce enantiopure (S) benzyl alcohols









(Figure 12). The initial chiral sulfoxide is prepared by sulfoxidation of the parent aryl

alkyl sulfide with 3-cyclodextrin and peracetic acid followed by crystallization.22

1) LDA
2) R O
0
0o
S R
'" ^ s-( Raney Ni H3

CH3 96-98% OH


(S,S) p- hydroxysulfoxide (100% d.e.) (S) alcohol
100% e.e.
R = n-alkyl
Figure 12. The synthesis of enantiopure (S) benzyl alcohol using chiral (S) Methyl 1-
naphthyl sulfoxide22

Approaches Towards Sulfoxides

Routes to Racemic Sulfoxides, Classical Organic Chemistry

The oxidation of sulfur bonded to carbon has a long precedence and can be done

by many reagents. Many general oxidants perform sulfoxidation, these include nitric acid

23, H20224, dinitrogen tetroxide 25, chromic acid26, ozone27, manganese dioxide26,

selenium dioxide28, NaIO329 and oxone (or 2KHS05-KHS04-K2S04) which has

provided sulfoxidations of some cyclic sulfides30. Modest stereoselectivities can be

achieved with NaI04-catalyzed S-oxygenation in the presence of BSA31. NaI04 usually

used in methanol/water affords yields of 90%29 and can also be supported on silica32 or

alumina.


Control of oxidation to avoid overoxidation to the corresponding sulfone and

sulfonic acid33 and control of the orientation of the oxygen atom added are often the











most sensitive criteria for developing sulfoxidation chemistry for current synthetic

purposes.


Routes to Non-racemic Sulfoxides, Classical Organic Chemistry

Several techniques are currently available to obtain enatiomerically pure

sulfoxides. These include the asymmetric oxidation of thioethers, asymmetric synthesis

by nucleophilic substitution on chiral sulfur derivatives, kinetic resolution and optical

resolution (Figure 13). The selective complexation and crystallization of the enantiomers


of ethyl p-tolyl sulfoxide was first achieved in 1996 by Cope and Caress34 with

platinum(II) complexes containing either (+)-a-methylbenzylamine or (-)-a-

methylbenzylamine. The complexes decompose with aqueous sodium cyanide to afford

the enantiopure sulfoxide. A comprehensive review of optical resolutions used up to the


late 1980's was prepared by Drabowicz et al.35




S,
R1 R2

ASYMMETRIC SULFOXIDATION

Kagan-Sharpless modified catalyst
-chiral oxazindines
chiral (salen)manganese(lll)
biocatalysts
ect...


L R CHIRAL SULFINYL TRANSFER REAGENTS
OPTICAL RESOLUTION S
+ R1
selective crystallization R1 R2 R2 -ephedrine
-ect...
,st
R2- LG1 LG2
S+ 0O
R1 R2 0
R1 R2 /, Andersen synthesis
S/St -ect...
R1 LG2
Figure 13. The most prevalent current synthetic routes towards chiral sulfoxides.









The most widely used of the methods towards non-racemic sulfoxides5 is the

oxidation of a parent sulfide. Several oxidative methods have been applied to this task,

with varying selectivity both in the stereochemistry of the reaction and the range of

substrates that are accomodated. Various organic peracids have been used by Overberger

and Cummins.36 in sulfoxidations. Chiral peracids offer low stereoselectivities (in the

order of 5-10 ee%)37 although they are highly reactive and lead to overoxidation to the

corresponding sulfone unless there is careful monitoring of temperature and excess

catalyst. Hydroperoxides and organic peroxides38 alone have also been used and in

conjunction with a sulfur co-ordinating metal bearing chiral ligands, modest

stereoselectivities can be achieved. Kagan39 and Modena40 modified the Sharpless

catalyst and used this for sulfoxidation with some success. Sulfoxidation with

Ti(OiPr)4/DET/BuOOH/H20 afforded moderate enantioselectivity although this is on a

limited substrate range.41,42 Further increases in the stereoselectivity of this oxidation

has been been achieved by replacing the oxidant with cumene hydroperoxide and by

changing the chiral ligand. The use of (R)-(+)-binaphthol has increased the modest e.e.%

(60-70%) of aryl methyl sulfoxidation of the Kagan modified Sharpless catalyst to 80-

96%.43 This takes advantage of the kinetic resolution of the further oxidation selectively

of one enantiomer of the sulfoxide to the sulfone by the titanium-binaphthol complex

(Figure 14). This does not improve the yield of the reaction though, which requires

further purification from the sulfone.44










ASYMMETRIC OXIDATION KINETIC RESOLUTION

Ti(O' Pr)4, tBuOOH,








56%ee(R) S 96% ee (R)
HC -S-CH3 (44% overall)
0
Figure 14. A modification of the Kagan-Sharpless oxidation of sulfides using (R)-(+)-
binaphthol in place of DET.43

Chiral oxaziridines (such as the Davis oxaziridines) can also perform selected


sulfoxidations often with high enantiospecificity (Figure 15).45-50

(Camphorsulfonyl)oxaziridines and their 8,8-dichloro derivatives are available in both

antipodals allowing access to both enantiomers of the product sulfoxide with


stereoselectivities of 84-96% ee.50,51


OXAZIRIDINESULFOXIDATION
CATALYST
a (X=H)
CH, CH2CI2

a= N SCH3 CH3
a= N3

SSO2 85% (22% e.e. (R))

X=CI, H

Figure 15 Chiral oxazidines as sulfoxidation catalysts.50,51

The Andersen method was the most widely used route to nonracemic sulfoxides

prior to enantioselective sulfoxidations (Figure 16). It is general in scope and can be

applied to acquire complex homochiral sulfoxides, although its major drawback is the

necessity to obtain the optically pure menthyl p-toluenesulfinate precursor. Esterification

ofp-toluenesulfinyl chloride with 1-menthol is a racemic reaction that requires separation







13


of the enantiomers, which is achieved by fractional crystallization.52,53 The maximal

yield of each enantiomer is 50% but can be increased with concurrent epimerization

using an acid catalyst such as hydrogen chloride (Figure 17). In such a way up to 90% of

the (S) sulfoxide ester can be obtained. The reduced facility to obtain the antipodal

menthyl sulfinate limits this method to production of only the (R) enantiomer of

sulfoxides.


0 : 0. 0
Menthol, f ,
SCI Py, Et2, rt OMenthyl + S OMenthyl

HC H3C H 3C
racemicc)
Selective Recrystallization
acetone, HCI, -200C



Grignard reagent O
-, R-MgX S
S RM Ni S, OMenthyl
Menthol + SR COMenthyl
(inversion) H3
H3C
Figure 16. The Andersen synthesis of chiral p-tolyl sulfoxides. The route towards the
antipodal sulfoxide is identical but requires different separation techniques to
afford the (R)-menthyl ester.9


O. Cl O
\- / HCI I HCI /
pTol SOMenthyl S--OMenthyl pTol 'OMenthyl
H20 pTol I H20
Cl
(S) (R)
Crystallization (acetone-HCI)
90%
Figure 17. Kinetic resolution in the Andersen synthesis, recrystallization in acetone-HC1.

The reaction of the arenesulfinate with a Grignard carbanion equivalent is

performed in benzene with moderate to high yield as a route to aryl p-tolyl sulfoxides.


This requires a change of solvent to remove the menthol byproduct.9 Lithium cuprates









can also be used to generate appropriate sulfoxides providing lower yields (16-59%) but

reducing the purification steps.35 The Andersen method has also been adapted for solid

phase sulfoxidation on a Wang resin.54 The initial stereoselectivity of the coupling to the

resin cannot be controlled though and the reactions are racemic.

The Andersen method is also limited to the preparation of alkyl-aryl sulfoxides, as

the displacement of the menthol group does not occur in the corresponding alkylsulfinate

esters55. To afford alkyl alkyl sulfoxides certain alkyllithium reagents can be used to

displace the aryl group with inversion of stereochemistry.56,57

The synthesis of sulfoxides using the Andersen method employs the transfer of an

already enantiomerically pure sulfinyl group to a nucleophilic center, often carbon. This

technique has been expanded upon with more reactive sulfinates and with sulfoxides that

allow controlled double displacement by carbon nucleophiles. This affords a choice in the

selection of both substituents (Figure 18). Exclusive inversion at sulfur offers

enantioselectivity to these syntheses, provided the initial sulfoxide substrate is

enantiopure. The use of benzyl p-bromophenyl sulfoxide for a double displacement at

sulfur is an example. It has yields of 50-90% product in >98% e.e. and is a feasible route

towards chiral nonracemic alkyl alkyl sulfoxides.58There have been several thorough

reviews covering extensions to the Andersen method and applications in synthesis.5,9













O R2-
R1I LG R1' R2
0'S


CHIRAL SULFINYL TRANSFER REAGENTS


S R1 0 R2 0 O
LG SLG2 R. ,SLG R2 1S'R2
LG< LG2 RIS1LG2 R1iS-R2


n, menthol
O CH](Andersen method) N- O h
Ph CH, HC Ph
OH3
0 H ephedrine
N N CH3
Ph Ph PH L oHN Br
H3CO H3 Ph0



Figure 18. Several chiral sulfinyl transfer reagents for the enantioselective synthesis of
sulfoxides.5

Non-racemic MTSO has been produced using the aforementioned methodologies,

with varying yields, enantioselectivities and complexity of methods (Table 1).


Table 1. Classical organic chemistry sythesis of enantiopure MTSO.

Method ee. % % yield Ref.
(scale)
Andersen synthesis 99% (R) 80% (204mmol, 60g) 59
Kagan-Sharpless with 96% (R) 44% (0.5mmol, 69mg), 43
cumene 30hrs
hydroperoxide
Kagan-Sharpless with 60% (R) 80% (1 mmol, 138mg) 60
binaphthol ligand
Kagan-Sharpless 54% (R) 67% 41
tBuOOH, VO(acac)2, 10% (R) 100% 61
menthol,
benzene/toluene
(S)1-phenyl-ethyl 80% (S) 71%, with 11% sulfone 44
hydroperoxide, (400|jmol, 55mg)
Ti(O'Pr)4

Enzymatic Routes to Chiral Sulfoxides

Stereochemistry is an additional intricacy in the assembly of atoms; it enriches the

diversity of chemistry and the space of biological molecules is acutely sensitive to it.

Proteins and large molecules in biological systems incorporate an environment









surrounded by solid which can restrict smaller organic molecules and provide control and

selectivity in chemistry.

This chiral cavity itself has been used solely to direct stereoselective reactions.

Prasanta et al. recently demonstrated that the sulfoxidation of alkyl aryl sulfides by H202,

which in solution is a racemic conversion, in the presence of the hydrolase a-

chymotrypsin becomes enantioselective.4

The selectivity of biological catalysts has also been investigated for the oxidation

of sulfides. The oxidative metabolism of many biomolecules containing sulfur provides a

large assortment of enzymes for this chemistry.

Enzymes described in the oxygenation of MTS (Table 2), have been studied for

an assortment of reasons, including the interest in their catalytic mechanisms and active

site topography, their role in xenobiotic drug metabolism and their possible use as

biocatalysts.

Microbial sulfoxidation can provide highly enantiopure MTS of both antipodals in

at least small scale reactions although they have not been widely used in synthetic

applications. There are several factors that can reduce the attractiveness of isolated

biological catalysts to organic synthesis. Oxygenation enzymes studied for

biotransformations are often membrane-bound, require expensive co-factors, must be

extensively purified from tissues and cells and their activity is often coupled with that of

other enzymes.62 Many organisms possess more than one catalyst for oxidative

metabolism, contaminating the chemistry and further complicating purification.63 This

intractability has limited the use of isolated enzymes and instead many biocatalytic









preparations are done with whole-cell systems. The synthetic application of whole cell

systems however can be complicated by the pathogenicity of some microorganisms and

the requirement in the laboratory for microbiological expertise. Overexpression of

catalysts in a heterologuous host such as Escherichia coli can avoid competing reactivity,

is relatively quick and has widely established culture methods. This has provided access

to the study and application to several sulfoxidation enzymes.64 65

Table 2. Methods for the stereoselective sulfoxidation of methyl p-tolyl sulfide (MTS).
Various classical synthetic organic methods as well as isolated enzymes and
whole cell systems could be used to produce MTSO.

Method* e.e.% % yield Kinetic Ref.
(scale) parameters
Mortierella isabellina 100% (R) 66
(wc)
Corynebacterium 100% (R) -67
equi (wc)
rabbit FMO1 (cmp), >99% (R) (90nmol, 12.4ig) 2.3 64
heterologous nmole/min/mg,
expression Km<10MLM
rabbit FMO2 (cmp), >99% (R) (90nmol, 12.4ig) 3.6nmole/min/mg, 64
heterologous Km< 10pM
expression
Chloroperoxidase (ip) 99% (R) 83% (1.25mmol, 2hrs. 68
172.5mg)
H202 and peroxidase 96% (R) 82% (Immol, 8hrs. 69
from Coprinus 138mg)
cinereus (ip)
Saccharomyces 92% (R) 60% (1.3mmol, 70
cerevisiae (wc) 180mg)
Vanadium 82% (R) 18% (25imol, 71
bromoperoxidase (ip) 3.45mg)
rabbit FMO3 (cmp), 50% (R) (1 mmol, 138mg) 10nmole/min/mg, 64
heterologous Km<300|.M
expression
Helminthosporium 27% (S) 10% 72
(wc)
Rhesus macaque 30% (S) (Immole, 65
FMO2 (cmp), 138mg)
heterologous
expression









Table 2 (continued)
Method e.e.% % yield Kinetic Ref.
(scale) parameters
(1-phenyl)ethyl 79% (S) 68% 4hrs. 73
hydroperoxide and
peroxidase from
Coprinus cinereus
(ip)
Horseradish 77% (S) (100mol, 19 74
peroxidase (mutant) 13.8mg) nmole/min/tmol,
(ip) Km = 4.5mM
Pseudomonas putida 94% (S) 11% 75
expressing the
naphthalene
dioxygenase gene
(wc)
Pseudomonas putida 98% (S) -76
expressing toluene
dioxygenase (wc)
w = w le cell botransormation, cmp =crude membrane preparation of heterologously expressed protein, i = isolated enzyme

The sulfoxidation reactions on many different pro-chiral sulfur atoms catalyzed

by a wide range of fungal and bacterial systems and by isolated enzymes has been

reviewed in detail by Holland.77,78

Improving Catalysts

The relevance of asymmetric catalysis was evidenced by the 2001 Nobel Prize for

Chemistry to K. Barry Sharpless for his work on chirally catalyzed oxidation techniques,

as well as Ryoji Noyori and William S. Knowles for their work on chirally catalysed

hydrogenation reactions. The two major methods currently available to chemists can be

classed as homogeneous transition metal catalysts and biocatalysts.79 The applicability

of transition metal catalysts is based on tuning of the ligands from molecular modeling,

knowledge of the reaction mechanism, trial and error and some degree of intuition.

Within the last decade the technique of combinatorial asymmetric catalysis has allowed

the screening of large numbers of catalysts. The production of large numbers of catalysts









requires strategies for modular synthesis of chiral ligands coupled with development of

high-throughput assays to monitor enantiomeric excess (ee%.).80 So far the size of the

libraries generated has been limited to fewer than a hundred catalysts and the analysis is

performed by analyzing reactions conventionally using chiral separation

chromatography.81

For protein catalysts, often lacking structural data information and the molecular

basis for enantioselectivity or kinetic resolution not being well understood, rational

design has had very limited applicability in providing novel stereoselective activities. In

directed evolution, a combination of the molecular biological control for changes in

enzyme catalysts (mutagenesis) and their expression has led to large libraries, similar to

the combinatorial asymmetric catalysts, of possible novel asymmetric catalysts.

The source for differing natural enzymes is vast; it has been estimated that >99%

of the naturally occurring enzymes have yet to be discovered.82 The other sources of

interesting candidates could therefore be from panning metagenome DNA.82

Efficient screening methodologies to study these new libraries have been a

relatively recent but exciting research venture.79,83










combinatorial synthesis directed evolution of
to produce a variety enzymes for metagenome DNA panning
of transition metal enantioselective and expression of novel
based chiral catalysts reaction catalysts enzymes





Large libraries of
pote ntiallyenantioselective
catalysts



Screening




novel enantioselective catalysts
for use in organic synthesis

Figure 19.Sources of large libraries from which enantioselective catalysts can be
explored.

Directed Evolution of Enzymes

Theory and Approaches

Whereas in nature enzymes have evolved as agents in the tasks required for life,

the requirements of enzymes for our use outside that context are often very different. The

properties of enzymes that are of interest to the synthetic organic chemist include

regioselectivities, substrate specificities, stereoselectivities, activity in non-aqueous

solvents, stability and enhanced reaction rates. There are many useful organic reactions

which do not have a known or suitable enzyme catalyst.


Arrangements of atoms through to larger assemblages is not only unique in the

order of subunits (in the 1st dimension) but extends into the 3rd dimension. Uniqueness in

the co-ordinates of shapes can be, as much as the composition of shapes, determinable in

biological chemistry. Enzymes are conglometrates of atoms extending into space, the

occupation of these molecules in the 3rd dimension can alternatively reduce or mask the









distinctiveness of the subunit composition, although both could influence the properties

of the whole. Can this shape be molded, without prior knowledge of the relationship

between the composition of the 1st and 3rd dimensions? Aside from rational design, the

blunt probing of this relationship could possibly help elucidate these questions and

concurrently might provide biological catalyst tools with increased effectiveness in chiral

chemistry.

The prediction of the chemistry of the known enzymes to substrates in organic

synthesis is usually restricted. Little structural information is often available for many

enzymes which limits substrate suitability and stereoselectivity predictions.77 One way

to circumvent this has been through probing using substrates of various forms and

stereoelectronic configurations62 to develop a model of the active site. Another method

to increase the palette of available biocatalysts for a given reaction is development of new

enzymes.

Within the last 7 years several enzymes have been evolved artificially towards an

industrial or biotechnological catalytic goal. Arnold et al. 84 and Stemmer and

coworkers85 have pioneered the field known as the directed evolution of proteins. The

enzymes that they have modified included many for which no previous structural data

was available.

Enzymes are long sequences of the more than 20 different amino acids. This

complexity is reflected in variants of a protein, of which there are at least N20. The in

vitro generation of altered enzymes takes advantage of molecular biological techniques

which control protein sequence at the DNA source to generate some of these variants.









One approach generates randomly mutated descendants of proteins, each possessing a

sequence with several alterations dispersed along the length of the gene.

Larger libraries of these enzyme variants will undoubtedly have more likelihood

of possessing improved enzymes. Large numbers are also important due to the

observations that most of the generated mutants are inactive (deleterious mutations) or do

not have altered properties (neutral mutations) and that only a small percentage have

beneficial changes in the properties that are being tested (positive mutations).86,87

The size of the library of enzyme variants (mutants) is exponentially proportional

to the number of changes allowed per mutant.88 Too many mutations per gene can easily

create libraries much too large to screen in their entirety. It is therefore the screening or

selection strategy that limits the size of libraries used. Even with these limited library

sizes, which with current screening/selection methods are currently at 105-106 members

(covering about 1-2 mutations per clone), directed evolution has provided several

significantly improved enzymes. Thermostability89, activity in acidic conditions90,

substrate specificity91, activity in organic solvents91, stereoselectivities92 and kinetic

resolutions93 have all been 'improved' for various catalysts using the strategy of random

mutagenesis and screening. The applications of in vitro evolution to develop enzymes

with improved characteristics are numerous and beyond the scope of this thesis, although

there are several extensive reviews94,95 on the subject.

There are several requirements to successfully evolve a biocatalyst; a method to

create altered descendants of enzyme, an efficient way to express and present the enzyme

and a method to screen or select for the desired properties of the enzyme is essential.84









High Throughput Screening for Enantioselectivity

The screening of asymmetric catalyst libraries, whether homogeneous transition

metal catalysts or biocatalysts, requires monitoring to provide the ee% of each reaction.

The large numbers of samples in catalyst libraries also require methods that are fast,

sensitive and can be used with high-throughput. A thorough synopsis of some of the

clever recent techniques has been prepared by Reymond.80

The most convenient and prevalent method to screen large numbers of reactions

rapidly involves measuring a chromogenic or fluorogenic signal in the products. There

are several known chromophores and fluorophores which have been used to task (Figure

20). Many can even be taken up by cells and allow for monitoring of catalysis in vivo. X-

gal hydrolized by P-galactosidase produces insoluble blue dye indigo (Figure 20A) 96

visible through unlysed cells. A coupled reaction system using horseradish peroxidase to

polymerize products of mutant cytochrome P450 also produces signals that can be

detected through cell membranes (Figure 20B). Digital imaging of the cultures can detect

the fluorescent polymers produced.97,98












OH CI o
H
0A HO0 H P-Galactosidase Br Br

N -N Br
HO OH H H
OH O Cl
Indigo blue



B4 P450cam C 7 HRP, H202,
-- D,,,-- -
SOH o O

fluorescentt dimer)


OH
C O O"0 O P-Galactosidase HO O

HO OH
OH CH3 OH3
7-hydroxy-4-methylcoumarine

Figure 20. Enzymatic reactions can produce highly chromophoric and fluorogenic
molecules that can be used to screen for activity.

Reaction screening for enantioselectivity has required an array of intricate


methodologies. Some of the first catalyzed reactions to be screened were the hydrolysis


of esters adapted to release the chromophore p-nitrophenol (Figure 21)93 or fluorophores


such as coumarinyl derivatives (Figure 20C).99,100


The enantioselectivity of protein catalysts either is due to kinetic resolution in


which the enzyme selectively converts one to the enantiomers of a racemic substrate into


product or to a stereoselective reaction in which catalysis converts an achiral substrate


into a chiral product.


Kinetic resolution is usually expressed as an E-value, the ratio of activity the


enzyme has towards each enantiomerl01. At high dilutions of substrate (below the


Michaelis-Menten constant KM for each of the enantiomers) this can be determined









directly from the reaction rates for both enantiomers. A system was initially developed by

Reetz and co-workers83 to monitor the kinetic resolution of lipase catalyzed reactions.

The lipase catalyzed kinetic resolution of hydrolysis ofp-nitrophenol esters was done for

each enantiomer separately and analyzed pairwise using simple UV/Vis based 96-well

plate readings. This simple screening (initial screening was done using

spectrophotometric methods and interesting reactions were further analyzed using chiral

GC) was done on large libraries (1000-5000 clones). Several rounds of library screenings

discovered a lipase, improved from Pseudomonas aeruginosa (PAL), initially having

essentially no enantioselectivity (E = 1.1) to preference for hydrolysis of the (S) isomer

of 91% ee. at 18% conversion (E = 25.8) (Figure 21).93 This has been to date the only

example of successful enantioselective evolution of a lipase. Wahler80 cautions on the

disadvantages to comparing reactions monitored separately with each enantiomer.

NO N -0O
RHOH20 R R OH 0O. N2 -NO
OH + + 0
CH3 HO CH3

R = n-C8H17 18%, 91% ee. (S) (R)
Figure 21. An enantioselective mutant PAL catalyses the kinetic resolution of a racemic
p-nitrophenol ester.83

Methods for the monitoring of improved kinetic resolution of other enzymes have

also been developed. Arnold et al. improved the kinetic resolution of a hydantoinase

towards the production of D-methionine by selecting for the best mutants of the enzyme

in the hydrolysis of the enantiomers of 5-monosubstituted hydantoin (Figure 22). The

enzyme is presented with each enantiomer separately and the formation of product is

monitored in both reactions from the increase in acidity, using a simple pH indicator










cresoll red). This affords a rapid screen of mutants, of which >20 000 were screened, for

the improved activity towards the L enantiomer. Further characterization of initial

positive clones was confirmed by chiral HPLC analysis. The evolved enzyme mutant

identified produces 20% ee. of the D enantiomer compared to the 40% ee. L selectivity of

the wild type enzyme.


5-monosubstituted N-Carbamoyl
hydantoin amino acid

HYDANTOINASE
H CATALYSED REACTION O

HNR H20 RH OH
O N HN NH2

0
D-enantiomer


HH0' H20o
H,. o N 0 H..Y OH
N HN NH2
0 H "N-,n
L-enantiomer
Figure 22. The reaction catalyzed by hydantoinase, which has distinct enantiomeric
substrates (the inherent racemization between the 5-monosubstituted
hydantoin isomers is slowl02).

Reetz and coworkers have also developed a monitoring system for kinetic


resolution reactions using mass spectrometry.103 The technique requires that one

enantiomer be isotopically labeled to distinguish it using electrospray mass spectrometry.

With an eight-channel sprayer system samples can be run very quickly (70sec/sample)

and library sizes of up to 10 000 can be screened per day.


As opposed to kinetic resolutions directly screening enantioselective reactions

that occur with prochiral substrates it is a bigger challenge.









Using the principle of microscopic reversibility these reactions might be treated as

kinetic resolutions, since many enzyme catalyze reversible processes. Microscopic

reversibility predicts that the ratio of the rates of the reverse reactions indicates the

enantioselectivity of the forward reaction. Catalysts are provided with the enantiopure

products of the forward reactions and monitoring is done for the products of the reverse

reaction ie. the natural substrates for the enzyme. Sensitive detection is required for the

low yield of these products or as was done by Reymond et al. coupling with a second

irreversible step can be used to help trap and detect the product (Figure 23).104 The

reverse reaction of alcohol dehydrogenase on a fluorogenic derivative of 2-butanone is

trapped by further reaction to the fluorescent product umbelliferone. This p-elimination is

catalyzed by bovine serum albumin in alkali media.

Alcohol dehydrogenase
OH NADP







OH [lNr jO BSA pH > 7


LII
H3CHO 0 0 *

HO~O
7-hydroxy-coumarin
(fluorescent)
Figure 23. The enantioselective reduction of a coumarinyl derivative of 2-butyl ketone
can be monitored using high-throughput fluorescence screening for the
reverse oxidation reaction coupled with elimination to form a fluorophore
product.

Very recently examples of screening systems have been developed for high-

throughput % ee., analysis of various reactions. These have been thoroughly reviewed by









Reetz.105 An 1H-NMR based screening assay allows up to 1400 ee determinations per

day (using an appropriate flow-through cell and autosampler)82, capillary array

electrophoresis (allowing in some cases 30 000 ee. determinations per day)106 and the

use of circular dichroism spectroscopy (CD) (which can accommodate up to 10000

samples per day)81 have all been developed within the last 4 years. The work of Wagner

and coworkers on immunoassays to selectively visualize chiral productsl7 is a screen

for enantioselectivity which is powerfully adaptable. Antibodies can be raised to almost

any compound of interest. An antibody recognizing both enantiomers can be used to

measure yields of reactions and one specific for an isomer (in the analysis of mandelic

acid prepared from the Ru-catalyzed hydrogenation of benzoyl formic acid) can be used

to indicate the stereoselectivity (Figure 24).




YIELD CHO COOH HO COOH


Ph + Ph

STERESELECTIVITY OCOOH COOH HO' COOH


Figure 24. Screening for enantioselectivity using competitive enzyme immunoassays07.
A colourimetric assay to quantify the racemate (yields of a reaction) and the
stereoselectivity with two antibodies, providing a method to screen for yields
and %ee. of reactions. One antibody is blue and recognizes both enantiomers
and the other is red and is (S) specific.

The aforementioned methods have been used to direct the evolution of the kinetic

resolution of various enzymes and for the screening of transition metal asymmetric

catalysts. These methods though have as yet not been employed in the screening of an

enzyme library for the stereselectivity of a reaction that starts from a racemic prochiral









substrate such as the sulfoxidation of an unsymmetrically substituted sulfide. In addition

there has yet to be an example of an enzyme catalyzing such a reaction whose

enantioselectivity has been inverted from preference towards one isomer to that of the

antipodal.

Flavin-Monooxygenases

Flavin nucleotides have rich redox chemistry, in various protein environments,

acting in conjunture with other functional groups they can be the catalytic part of

reductases 108, oxidases109,110 and monooxygenases.111

The first of the enzymes termed flavin-containing monooxygenases was purified

from readily available pig liver in 1972 by Carolyn Mitchell (a student of Daniel M.

Ziegler at the University of Texas at Austin). 11 The isolated enzyme was associated

with equimolar amounts of FAD but not with other co-factors or metals and was

identified initially as an NADPH and 02 dependent amine N-oxidase. It was later found

that it catalyzes the oxygenation of a wide range of inorganic and organic substrates that

contain soft nucleophilic heteroatoms such as Se, S and P and of N and B centers

also. 112 These substrates have a variety of structures and sizes, but are with few

exceptions devoid of charged groups 113. Oxygenation of sulfur atoms occurred with high

enantioselectivities. 114 FMO's have since been isolated from mammalian and fungal

sources, each providing several forms of the enzyme. The liver form (hepatic), lung form

(pulmonary) and kidneys form (renal) have all been characterized from rats, pigs, humans

and rabbits.

FMO's are microsomal enzymes whose main purpose is postulated to be in the

oxidative metabolism of environmental toxicants, natural products, and therapeutics. In









humans FMO's are responsible for metabolism of many xenobiotics and drugs. 115 They

promote detoxification by oxidizing nucleophilic centers116 and, along with the

cytochrome P-450's, oxygenate primary and N-substituted amines into hydrophilic

products, that could be more readily excreted from the body. Flavin containing

monooxygenases are classified as Phase I or Oxidative Drug Metabolism Enzymes.

Background and Properties

Both flavin (FMO's) and heme-containing monooxygenases (cytP450) are to be

found associated with the endoplasmic reticulum of mammalian cells. Both of these

enzyme types are NADPH and 02 dependent and are grouped within their respective

family into members related by primary amino acid sequence. The genes are expressed in

species and tissue specific levels but do not appear to be induciblel


A standardized nomenclature for this family was advanced jointly in 1994 by

some of the major researchers in the field 17 based on the guidelines established for the

cytochromes P450's 118. This nomenclature is based on similarities in primary amino

acid sequence and replaces the previous laboratory specific classifications. Currently

there are 5 agreed upon isoforms of the flavin-containing monooxygenase (FMO) gene

family in each mammal extensively studied to date (Table 3). Homologs from different

species are grouped into the 5 families of isoforms; isoforms (from different species)

possess >80% identity whereas there is <60% identity between homologs within the

family. Structure defines the classing irrespective of the enzymes distribution among

tissues or functional or chemical properties. As part of the human genome effort, another

FMO-like gene in humans, FMO6, was identified between FMO3 and FMO2 (GenBank









accession no. ALO21026) although reverse transcriptase coupled polymerase chain

reaction (DNA amplification) show that most of the gene transcripts would not code for

functional protein (based on an analysis of possible open reading frames). 119

FMO has yet to be used in organic synthesis, although FMO have been used in

biosensors enzyme electrode and a bioelectronic nose (bio-nose) to detect methyl

mercaptan (detecting 1-4000ppm).120 The drive to studying mammalian FMO's has

mainly come from the medical need to further understand the mechanisms of oxidative

metabolism. This has been an impetus for determining human FMO substrate specificity

to anticipate the metabolism and interaction with drugs.121 There is developing though a

significant gap in information pertaining to orthologs in insects, echinoderms, avian,

reptilian and amphibian species. 122

Very rare mutations of the human FMO3 gene that have been associated with

deficient N-oxygenation of dietary trimethylamine.123 Defective trimethylamine N-

oxygenation causes trimethylaminuria or "fish-like odor syndrome". In cows a nonsense

mutation in FMO3 underlies a "fishy" off flavour in cow's milk.124 The rabbit FMO4

gene has as yet no protein counterpart, either isolated or expressed heterologuouslyl25.

The genome of humans and mice have revealed that all isozymes of FMO reside

in a single cluster of a chromosome la. 121,126 The FMO2 from Rhesus macaque is 97%

identical to human FMO2.127Due to the close proximity of the genes and the uniformity

of differences in the sequences of the orthologs (ie. all members of a family are similar

across species) Cashmanl26 has postulated that the genes for FMO's have arisen from

duplication of a single ancestral gene 250-300 million years ago, 200 million years









before mammalian speciation. A enzyme in Arabidopsis thaliana and one in yeast with

similar amino acid identity have also been termed FMOs. The purpose of the yeast

enzyme is postulated to be in maintaining the thiol-disulfide redox potential in the

cell.128-130


Although enzymes in the family were identified based on sequence information a

classification by Cashmanl26 and Zieglerl 14 of the family was also developed to

describe their activity characteristics:

Flavin-containing monooxygenase (EC 1.14.13.8) or FMO


NADPH dependent oxidative catalyst, utilizing both 02 and H+ and which is
associated with a flavin adenine dinucleotide (FAD) prosthetic group.

a microsomal or membrane associated protein.

multisubstrate enzyme which oxygenate many nucleophilic atoms. These include
sulfur, selenium, phosphorus and nitrogen centers.

they can form a kinetically stable 4a-hydroperoxyflavin independently of the
presence of substrate

Table 3. List of identified and putative FMO's to date.
Isoform Organism Tissue Level Genbank # Ref
FMO1 Pig liver cDNA M32031 131
Rabbit liver cDNA M32030 132
protein 133
Human kidney cDNA M64082 134
Rhesus liver cDNA 135
macaque
Rat liver, cDNA M84719 136
kidney
Mouse kidney cDNA P50285 137
Dogt cDNA 138
Saccharomyces cDNA 139
cerevisiae









Table 3 (continued)
FMO2 Rabbit lung cDNA M32029 132

lung protein 140
Guinea pig' cDNA 141
Human Liver, cDNA 4503759 142
kidney,
lung
Rat cDNA AF458414.1 143
Isoform Organism Tissue Level Genbank # Ref.
Gorillat cDNA 144
Mouse lung cDNA 145
Chimpanzeet cDNA 144
Rhesus lung cDNA P97501 146
macaque
FMO3 Rabbit liver cDNA L10391 147
Mouse liver, cDNA U87147.1 148
kidney
Human liver cDNA M83772 149
Cowt cDNA 124
Rat kidney, cDNA 150
liver
Arabidopsis cDNA 151
thaliana t
FMO4 Rabbit kidney cDNA L10392 152
Human liver cDNA Z11737 142
Rat kidney cDNA 153
Mouset cDNA 154
FMO5 Rabbit liver, cDNA L08449 155
kidney
mouse liver, cDNA U90535 137
kidney
Human liver cDNA M83772
Ratt liver, cDNA 156
kidney
Guinea pigt liver cDNA L37081 157
FMO6 Humant liver DNA 119
* tissue of predominant localization
. putative FMO, from extrapolated primary sequence comparison, but not corroborated with studies of activity









Catalysis by FMO's

A catalytic cycle was postulated by Ziegler and Ballou based on kinetic and

spectral studies with FMO from rabbit lung microsomes63 and pig liver microsomes 158-

160 (Figure 25).


In the ordered sequential cycle of these FMO's, interaction between the

xenobiotic substrate and the enzyme is preceded by binding of NADPH and 02. The

relatively stable 4a-hydroperoxyflavin form of the enzyme, seen in the presence of

NADPH and 02 has been identified by similarities in its spectral characteristics to N5-

Ethyl-4a-hydroperoxyflavin.161 It is resistant to decomposition, long-lived and forms

irrespective of substrate binding as compared to other flavoenzymes such as

cyclohexanone monooxygenase (CHMO from Acinetobacter sp.). Attack of the terminal

flavin peroxide oxygen by nucleophiles (such as N in dimethylaniline) occurs (Step A)

rapidly in a bimolecular reaction (4700 Mls-1) forming the pseudobase 4a-hydroxyflavin

form of the enzyme and immediately releasing the oxygenated product. Only a single

point of contact between the enzyme and the substrates studied is postulated to be

required for formation of the product. 15 8-160 Consequently any soft nucleophile not

excluded from the active site will be oxidized and released in the first step. Conversely

the active site is likely lipophilic and free of nucleophilic residues. The precise fit often

needed to lower the activation energy of an enzyme-catalyzed reaction is not required

since the energy to drive the reaction is already present in the 4a-hydroxyflavin form of

the enzyme before it encounters the substrate. This reduces the relevance of describing

Km for FMO enzymes in terms of high or low affinity substrate binding. These enzymes

do however follow the approximations of Michaelis-Menten saturation kinetics. 158









The next step (Step B), dehydration of the 4a-hydroxyflavin was demonstrated by

rapid kinetic studies 160 to be about 10 times slower than any other step (1.9AM'sec-1).

Solvent deuterium effects also suggest participation of a general acid catalystl2 which

could be participating in the dehydration. In the enzymes studied (porcine liver and rabbit

lung FMO's) this step becomes rate limiting. Since it occurs after product release,

reactions following this mechanism should have similar Vmax values for all good

substrates. The oxidized co-factor NADP+ is a competitive inhibitor ofNADPH and is

released in the following step (Step C) prior to regeneration of the active 4a-

hydroperoxyflavin form of the enzyme. In the absence of bound NADP+ FMO's becomes

a NADPH oxidases; molecular oxygen reacts with the reduced form of the enzyme and

produces significant levels of H202. If this were to occur extensively naturally in FMO's

it would make them a source of oxidative damage inside the cell as well as reduce

cellular levels of NADPH. The fully oxidized flavin form of FMO reacts (53 M1) with

NADPH (Step D) to give an enzyme with bound NADP+ and reduced flavin. The reduced

flavin can then form (45 M1) the 4a-hydroperoxyflavin in the presence of molecular

oxygen (Step E) regenerating the active form of the enzyme.






















OH
FAD


w


NADP


E


NADPH
FAD


NADP+


Figure 25. The catalytic cycle postulated for pig liver FMO's. The 4-a-hydroperoxyflavin
bound to the enzyme oxidizes the substrate in step (A) followed by release of
water (the rate limiting step) (B). The active form of the enzyme is
regenerated in steps C, D, E independent of the presence of substrate.

This mechanism was developed using the porcine liver and rabbit lung FMOs

which have been studied in much detail and whose oxidative activities are limited to

small molecule N and S oxidation, but it is hoped that most other FMO isozymes follow

similar mechanisms and that this scheme is applicable to various xenobiotic substrates.









In fact rat and human liver enzymes (FMO1) defer in that the decomposition of

the pseudo-base, and dehydration does not appear to be rate limiting. 163 It is also known

that rabbit FM05, classed as an FMO based on primary structure164, does not freely

form the 4a-hydroperoxyflavin intermediate which is postulated to account for its

restricted substrate specificity. Its catalytic cycle could be similar instead to other studied

flavoenzymes such as to that ofp-hydroxybenzoatehydroxylase.165 So far crude

membrane preparations and the purified enzyme shows only N-oxidation activity of n-

octylamine and n-nonylamine.

Amine nitrogens in comparison to sulfur, selenium or other readily polarizable

heteroatoms are not as nucleophilic. There is more to the active site in this reaction as the

enzyme somehow activates these moieties, 4a-hydroperoxyflavin bound to FMO

oxidizes amines 106 times faster than in solution. 166

Structure of FMO's

These membrane-associated proteins have as yet not been crystallized. Active site

models have been postulated based on substrate structure acceptance for catalysis. 167

The individual substrate profiles of each FMO studied can be explained by and provide

some insight into the active site channel. Some structural knowledge is also predicted

from relatedness to protein with crystal structures. In particular the residues 12-23 of

FMO's IGGGPGGLAAAR are 83% identical to the glutathione reductase sequence

IGGGSGGLASAR which is a P-sheet and turn region which forms the 'floor' of the FAD

binding site. 168










Based on substrate stereoselectivy profiles of sulfoxidation and the ability to

distinguish betweenpro-S andpro-R sulfur atoms, Cashmanl26 and Rettiel69 have

proposed a structure for the active site of several FMO's (rabbit and pig FMO 1 and rabbit

FMO2). According to this model (Figure 26) the sulfide passes into the binding channel

in a committed orientation. The substrate binding channel leads to two compartments,

one pocket is larger (A) and prefers to bind large, planar aromatic residues and is less

accessible but can accommodate short n-alkyl chains or ap-tolyl moiety. With this

topography MTS is oriented so as to present the pro-R lone pair of electrons, exclusively

to the hydroperoxide.






CH3


^ CH,
H3C


CH3








Figure 26. Model of the active site of rabbit and pig FMO 1 and rabbit FMO2 proposed by
Cashmanl26 and Rettie169 showing the orientation substrates propyll 2-
naphthyl sulfide, methyl 2-naphthyl sulfide, MTS, heptyl p-tolyl sulfide)
would adopt in the active site. There are two principal substrate binding
pockets A and C adjacent to the flavin cofactor (below the surface of the
active site) and the binding channel D open to the surface of the enzyme.

At a critical length the n-alkyl chain of alkyl aryl sulfides preferentially occupies

pocket A, reversing the stereoselectivity of the reaction. This critical length varies among










several FMO's tested (Figure 28)64,135,170 and would be indicative of the differences

between the compartments A and C in the each enzyme with the proposed model. From

this analysis the substrate binding channel is clearly distinct among forms of FMO

studied so far167,171,172 and provides enzymes with varying stereo- and substrate

selectivities. Larger substrates readily oxidized by pig, guinea pig and rabbit FMO1 are

excluded from the active site of human FMO 1173 indicating a large variation between

homologs.


FMO cDNA encode proteins generally 533-535 amino acids in size (Figure 27)

although examples of proteins with additional 19174, 25142 amino acid at the C-terminal

ends have been observed.


FMO protein topography


FAD NADP
0FAD NADP lipophilic portion
(Met) I ICOOH
H
I I I I I
100 200 300 400 500
Figure 27. The topography of the -535 amino acid protein of rabbit FMO1, 2, 3.
















Rabbit FMO1


1 2 3
n-alkyl chain length (r



Rabbit FM03


(CH2)nH Rabbit FMO2


100


90


80


70-


S60


50


40


30 -
4 1 2 3
n) n-alkyl chain length (n)



Rhesus macaque FMO1


100


90


80


70-


a 60


50


40


30 -


1 2 3 4 1 2 3 4
n-alkyl chain length (n) n-alkyl chain length (n)


Figure 28. Differences in the enantioselectivity of sulfoxidation by purified FMO's from

different sources64,135 with increasing alkyl chain length of alkyl p-tolyl
sulfides.









Hydropathy profiles175 ofFMO orthologs show similarities even in regions of

very modest identity (25-30%). The strong membrane associations of all mammalian

FMO's studied to date is revealed in the poor solubility and highly intractable nature of

the purified proteins.140 Although the lipophilic membrane associated region is predicted

by hydropathy profiles of primary sequence to lie in the 215 C-terminus amino acids,

truncation of this region does not diminish membrane association176. Cashman has

assumed that the N-terminal end of the peptide is also not likely to be inserted into

cellular membranes due to the presence of binding sites for cytosolic cofactors (NADP+,

FAD). 126 Fusion proteins of FMO joined at the N-terminal end to peptide fragments

such as 3-galactosidasel26, maltose binding protein177 and His6 and Hisio tags show

added solubility and stability to proteolysis and thermal inactivationl77 as well as

improved activity of the protein isolates.


FAD binding domain (GxGxxG)178 is common in all FMO's studied, at amino

acid position 9-14 as is the NADP+ binding domain at position 186-196.126

Post-translational modifications has been detected in several FMO's (rabbit

FMO1, FMO2, FMO3) from native tissues and include the proteolitic cleavage of the

initiation amino acid (methionine) and the N-acetylation of the following (N-terminal)

amino acid. 179


The porcine FMO1 protein has been shown to be glycosylated at Asn 120180 and

putative consensus site for glycosylation (Asn-Xaa-Ser/Thr) is present and highly

conserved between all FMO's.









Although this might be indicative of such modifications either structurally or

catalytically, lack of glycosylation (as in proteins heterologously expressed in microbial

systems) or other post-translational modification has no measurable effect on the

substrate specificity, stereoselectivity, kinetic properties of catalysis or the robustness of

the enzyme. 126

Single amino acid changes have been observed previously to drastically affect the

physical and chemical properties of FMO'sl81 such as their migration on SDS-PAGE

gels. It has been proposed by Cashman that few or even single amino acid mutant of

FMO's would possess markedly different catalytic properties. 126

Strategy Towards A Stereoselective Sulfoxidation Catalyst, An Overview of this
Work

This project is attempting to use directed evolution to provide catalysts for the

enantioselective synthesis of sulfoxides. Flavin-containing monooxygenase enzymes with

a preference for one enantiomer in the sulfoxidation of MTS will be modified towards

production of the opposite enantiomer. The resultant enzyme variants will hopefully

possess the catalytic activities inherent in the native FMO and with the template gene

would yield a pair of enzymes to provide the choice of routes to both antipodals of the

sulfoxide MTSO.

Catalytic activity will be modified in a random fashion with methods that have

been established for other enzymes94,95 and screened for improved enantioselectivity.

The general scheme is depicted in Figure 29.














mutagenesis clone into
or DNA expression
shuffling mutant system
library


FMO



(DNA)








culture,
express protein,
wash cells of media


+ NADPH






-i=-
BIOSULFOXIDATION
0"



yield ?
E r stereochemistry ?
SCREENING

determine
ee% of each
clones' reaction





select best clones







Figure 29. An overview of the strategy for the directed evolution of an FMO enzyme as a
catalyst for the efficient production of (S) MTSO, a molecule with many uses
in organic syntheses.














CHAPTER 2
ENGINEERING OF MICROORGANISMS FOR FMO PRODUCTION

FMO's from Oryctolagus cuniculus (rabbit) and Rattus norvegicus (rat) where the

first full sets ever be isolated and extensively characterized. All 5 of the known rabbit

FMO genes were graciously provided to us by Dr. R. Philpot. The FMO2 cDNA of

rhesus macaque was a generous gift of Dr. D.M. Ziegler. They are all products of reverse

transcription from the host animal mRNA, a process which removes the non-coding

information that can be present in many mammalian genes. 176,182 These genes were

intended for cloning into heterologous expression systems.

The extensive knowledge of its genetics and biochemistry has made Escherichia

coli the organism predominantly chosen for cloning studiesl8. This prokaryote can

express high levels of protein, has simple and inexpensive media requirements and

relatively rapid growth. 183 The mammalian proteins expressed from Escherichia coli

however lack many of the post-translational modifications of higher organisms, which

have been identified in the native FMO proteins, including acetylation of the N-terminal

amino acid179 and amino acid side chain glycosylation.180 They might also require

refolding into their active form. 183 Further complications can occur due to the fact that

the cDNA derived genes have not been optimized for the differing codon usage for

whichever heterologuous expression system is used, such as Escherichia coli184 or

Saccharomyces cerevisiae. 185









Cloning and expression of several of the rabbit FMO cDNA proteins has been

recorded in work done previously in the labs of Philpotl74,176,186 and Ziegler.63 The

exception seems to be rabbit FMO4, which to date has not been expressed in an active

form from cDNA.126

The expressed proteins were similar to those isolated from rabbit tissue based on

their mobility on SDS-PAGE gels and from recognition by antibodies raised to the

naturally occurring peptides. 174,176,186 They agreed in other physical properties,

including optimal pH for activity and thermal stability and their oxidation activity

towards the test substrates: chlorpromazinel26,187, trifluoroperazine87,

methimazole 74,176, n-octylaminel74, thioureal76, dimethylanilinel76 and

cysteamine. 176 The exception is the enzyme of rabbit FMO5 which has only been

isolated in its heterologously expressed form and has a distinctly limited substrate range.

Concurrently Saccharomyces cerevisiae (yeast) is the eukaryotic microorganism

with the best known genetics and most extensive use as a cloning host.182 It possesses

many of the post-translational processing systems present in higher organisms, has

growth rates faster than many higher eukaryotic organisms, reduced media requirements,

lower costs to maintain and is benign. 183

Escherichia coli and Saccharomyces cerevisiae expressed FMO have been used

as representatives of the naturally occurring enzymes in many studies to date. 126 There

are several advantages to these expressed forms. Purification in large amounts can be

done without complication from the separation of the various FMO isoforms, which are

often co-expressed in their native tissue. 126 The E. coli expression system also allowed









for the testing of synthetic variants of the FMO's: truncated proteins to probe for the

location of the membrane binding domainl76 and site specific mutant peptides to

confirm the identity and efficacy of the binding sites for the NADPH and FAD

cofactors. 178

In our labs, expression of rabbit FMO 1 and FMO2 as well as rhesus macaque

FMO2 was of interest for the directed evolution of their stereoselectivities of catalysis.

The rabbit enzymes catalyzing very highly selective sulfoxidation to the (R) sulfoxide

(99% ee)135 while the enzyme from the primate possesses the inverse enantioselectivity,

producing 30% ee. (S) product.65 These values are both for the native enzymes isolated

from animal tissues. In the current study, the enzymes are being expressed in Escherichia

coli and it is needed to corroborate these activities for these forms. These catalysts are

relatively similar in sequence identity (60-85%), making them attractive targets for

creating chimeras, or DNA shuffling of regions between them.

The expression of enzymes for these studies has different requirements added to

those of research done with FMO's in the past. Whereas a clean source of enzyme is the

main prerequisite for the study of an enzymes' properties, for biocatalysis done in parallel

on large numbers the applications also require production of concentrated catalyst in

small volumes. There are several factors that can be developed to optimize this.

Expression Using Escherichia coli

Biocatalytic reactions used in the high-throughput screening of large numbers of

enzymes requires a reproducible presentation of the enzyme with the required co-factors

with limited pre-reaction and post-reaction processing. Catalysis using intact cells of









Escherichia coli expressing the membrane associated FMO proteins could provide

encapsulated high concentrations of the necessary co-factors FAD and NADPH In

minimal media containing glucose Escherichia coli cells can contain 200M of

NADP/NADPH and 51 |M of FAD. 188 Intact cells are also relatively simple to prepare

for an aqueous reaction, which can be done in growth media itself or by transferring cells

to a reaction buffer. If intact whole cells cannot be used, lysis of the cells and semi-

isolation of the membrane associated proteins can de done to present expressed proteins.

Expression of Active FMO Enzymes

Escherichia coli expression plasmid vector based systems vary among several

regulatory sequences. Bacterial plasmids adapted for expression of FMO's incorporate a

low-stringency high copy number origin of replication (pBR322 or ColE1)182 an

antibiotic resistance gene (such as ampicillin or kanamycin) and an IPTG controlled

promoter-terminator couple (Figure 30).

The strategies that were followed for the construction of the various expression

plasmids are recorded in Appendix A.












Table 4. Various expression systems for rabbit and rhesus macaque FMO's in Escherichia coli and Saccharomyces cerevisiae.
Plasmid Strain Gene Prom Selectable ori Expression Activityt Notes
marker gene
pAAP10 15C (yeast) rabbit FMO1 GAL1 URA3 2i No protein 1.8 x 10 (b
visible
pAAP 11 BL21(DE3) rabbit FMO1 T7 AmpR pBR322 High 8.4 x 106 (a)
1.4 x 10-5 (c)
pAAP27 BL21(DE3) rabbit FMO1 T7 KanR pBR322 High 5.6 x 10-5 (c)
pAAP27 BL21(DE3) rabbit FMO1 T7 KanR pBR322 High 4.9 x 10-5 (c)
codon +
pAAP15 BL21(DE3) rabbit FMO2 T7 Am pBR322 High ND (a) Not active
pAAP21 BL21(DE3) rabbit FM02 T7 KanR pBR322 High ND (a) Not active
pAAP21 BL21(DE3) rabbit FMO2 T7 Kan pBR322 High ND (a) Not active
codon +
pAAP12 XL1 Blue rabbit FMO2 tac AmpR pBR322 Low ND (a) Not active
pAAP20 XL1 Blue rabbit FMO2 trc Amp pUC Low ND (a) Not active
pAAP14 BL21(DE3) rabbit FMO3 T7 AmpR pBR322 Low 5.3 x 10-6(a)
pAAP18 BL21(DE3) rabbit FMO4 T7 Amp" pBR322 No protein ND (a) No protein
visible produced
pAAP35 BL21(DE3) rhesus macaque T7 AmpR pBR322 High 2.1 x 10 "(c
FMO2
pAAP36 BL21(DE3) rhesus macaque T7 Kank pBR322 High 7.8 x 10- (c)
FMO2_____
pAAP13 15C (yeast) rabbit FMO2 GAL1 HIS3 2tL No protein 2.3 x 10- (a)
visible
expression was determine from SDS-PAGE analysis o total cell extracts monitoring for presence of 6da protein activy (mmole/min/mg protein) was
measured as a) the consumption of NADPH in the methimazole coupled oxidation assay using crude membrane protein reparations r b) the oxidation of thiobenzamide c)
the oxidation of MTS (Appendix A), ND = not detectable










promoter
(translation)
your preferred
gene inserted here


terminator
(translation)
expression vector



selection marker

origin of replication


Figure 30. General schematic of an Escherichia coli expression plasmid vector.

Rabbit FMO1 and Rabbit FMO3

In our work the initial cloning of the rabbit FMO genes was done into the

commonly used pET based Escherichia coli expression system (Novagen). The system

provides a high copy number of plasmids inside cells, controlled by the origin of


replication derived from pBR322.189 IPTG induced production of other proteins using

this highly selective T7 RNA polymerase have resulted in levels of expression as high as

50% of the total cellular protein. 190


Protein production for rabbit FMO 1, FMO2 and FMO3 gene products (-60KDa

peptide bands not visible in the microbial protein background) was as high as 20, 30%

and <10% of total cellular protein respectively (Figure 31). There was no visible protein


expressed from the FMO4 clone, which is in agreement with published results.64










A B C D E F
kD
200
--- 116
97.4
66




31
I



Figure 31. Expression of rabbit FMO1, expected at ~ 59kDa132, from
BL21(DE3)/pAAP11 analyzed with SDS-PAGE. Total protein samples run on
12% SDS-PAGE: A) 6hrs. post-induction BL21(DE3) cells (background),
BL21(DE3)/pAAP11 post-induction: B) Ohrs., C) 2hrs., D) 4hrs., E) 6hrs., F)
Protein standard.

The cloned genes of rabbit FMO1 and FMO3 possessed identical nucleotide

sequence with the open reading frames of the genes described by Rettie et al. (Appendix

B), without further mutations which could have accumulated during cloning. A further

characterization of expressed products from the FMO 1 and FMO3 systems (pAAP 11 and

pAAP13) indicated that proteins had activities in the oxidation of methimazole and MTS

(Table 4). Oxidation was monitored indirectly by the consumption of NADPH during the

reactions (Appendix A). This corroborated the expression of both FMO enzymes (rabbit


FMO1 and rabbit FMO3) with activities for known substrates.64


Rabbit FMO2

There was no active protein apparent in the rabbit FMO2 system although a large

amount of protein of the expected size (~60KDa) is visible by SDS-PAGE analysis

(pAAP12, Table 4). This is not due to any visible inclusion bodies, conglomerates of

misfolded proteins, inside the cell since no distinction can be made between the cells









expressing rabbit FMO 1 or FMO2 and the native Escherichia coli strain in electron

micrographs (Figure 32).

An additional indication that overexpression-induced inclusion bodies are not

responsible for the lack of active protein was the similar inactivity observed for FMO2

protein produced by a system with for lower protein expression levels. The pJLS plasmid

system incorporates an analogue of the Escherichia coli trp promoter for the

heterologuous expression of proteins at lower levels than the expected results from the

highly selective expression in the pET (T7 promoter-driven) system.170

Sequencing of the cloned rabbit FMO2 to test for an inactivation mutation at the

DNA level was done and agrees with that expected from the initial identification of the

gene by Philpot and coworkersl70 therefore the gene is not inactive due to mutations.

Rhesus macaque FMO2

A similar expression system was initially constructed for rhesus macaque FMO2

(pAAP36) which indicates that a -60KDa protein is produced that has NADPH coupled

reaction activity. Sequencing indicated that there are 4 DNA mutations in this gene from

the published wild-type sequence which also resulted in several amino acid changes

(Arg298Ser, Cys302Ser and two silent mutations). NADPH oxidation was not coupled to

the oxidation of substrate in this enzyme mutant since there was no apparent

sulfoxidation activity of MTS whatsoever. It appears that this is an inactive mutant of the

wild type enzyme but interestingly it is a substrate independent NADPH oxidase.

Recloning of the original cDNA derived gene, using high fidelity Pfu polymerase

provided the wild type rhesus macaque FMO2 gene (in pAAP45) (Appendix B) for









subsequent directed evolution studies.

The requirement for one (R) and one (S) FMO sulfoxidation catalyst for DNA

shuffling is satisfied by the rhesus macaque FMO2 (pAAP35) and rabbit FMO1 isoform

systems (pAAP 11) and further investigation into the inactivity of rabbit FMO2 was

abandoned. Instead, fine tuning to raise the expression of active rabbit FMO1 was

investigated. This would produce more enzyme, allowing scale-up of the biocatalytic

reactions and enabling more sensitivity in the screenings.

























Figure 32 Electron-micrograph evidence of lack of inclusion bodies. Transmission
Electron micrographs x 25 000 of A) BL21(DE3), B) BL21(DE3)/pAAP1 1
rabbit FMO1, C) BL21(DE3)/pAAP15 rabbit FMO2, D) (and insert) E. coli
BL21(DE3)/pREP4/pETMW2 overexpressing Orf4-protein into inclusion
bodies (used with permission from images by Dr. G. Acker)191. Electron
micrography of FMO expressing cells done by Dr. H. Aldrich (Microbiology
and Cell Science, UF).









Improving the Expression of Active FMO

The antibiotic selection system was chosen based on maximal plasmid retentions.

In parallel culture assays, otherwise identical systems displayed varying plasmid

retentions. This indicates the use of kanamycin (83% plasmid retention after 20hrs.

culture) is preferred over ampicillin (46% plasmid retention after 20hrs. culture) as a

selective marker in the expression systems for rhesus macaque FMO2 and rabbit FMO 1.

Higher plasmid retention provided an apparent 4 fold increase in activity of the cell

culture (Table 4: pAAP11 compared to pAAP27).

Codon usage varies, with the preferences for members of a degenerate codon set

being distinct among different organismsl92. Prokaryotes such as Escherichia coli often

have different levels of tRNA's than mammals from which genes are derived. There are

commercially available strains of Escherichia coli with increased expression of these

mammalian favouringg' tRNA's (argU, ileY and leuW) that might have increased the

levels of protein production. Expression of FMO in one such strain (BL21(DE3) codon+,

graciously provided by Dr. Michael Thomson) does not seem to increase the levels of

active protein.

The expression systems for yeast FMO have been improved previously with the

aid of folding accessory proteins. Activity of FMO produced by an Escherichia coli

expression system was increased (x 4) with the co-expression of the chaperonins GroES

and GroEL.139 The expression of the inherent chaperonins of Escherichia coli can also

be induced by brief heat shock of cultures at 42C 193. For the expression systems of

rabbit FMO 1 and rhesus macaque FMO2 in this study though this does not increase the

expression of active protein.









As was mentioned earlier (Chapter 1) fusing peptide fragments to the N-terminal

of FMO's has produced heterologuously expressed proteins with added solubility,

stability and activity. 126,177 Other N-terminal modifications also increase activity of cell

culture isolates. Whether this increases levels of active protein or affects the activity

directly has not been investigated. The N-terminal Hisio modification of yeast FMO has

reportedly increases S-oxygenation activity 20 fold. 139

Attempts to modify the various ends of expressed rabbit FMO1 proteins also

indicate a preferential effect of adding peptide fragments to the N-terminal side.


Cashman and co-workersl77 used an expression system to produce human FMO3

protein fused to a maltose binding domain. Proteins were expressed from a system which

fused the MalE signal sequence to the N-terminal end of the FMO gene during expression

(pMal-c2x). It was reported that the protein was produced as a cytosolic soluble enzyme

with activities identical to the wildtype FMO3. The level of active protein produced was

much higher and additionally the fusions could be later purified on amylose resin. This

clone was graciously provided to us by Dr. John Cashman. Insertion of the rabbit FMO1

in the place of the human gene into this plasmid vector (pAAP240) though did not

produce active protein (Table 5). The activity of the FMO3 fusion clone corroborated the

results from Cashman (data not shown) indicating this system might work with some but

not necessary all FMO enzymes.

Addition of multiple histidine amino acids to ends of the rabbit FMO 1 sequence

had more rewarding results. Although there was no investigation into verifying the actual

presence of the His tags in the mature protein, an expression system designed to add a

His6 tag to the N-terminal of rabbit FMO 1 (pAAP38) increase the activity of crude









membrane preparations (x 2) (Table 5). The addition of a histidine tag to the C-terminal

end of the protein alternatively leads to reduction in the amount of enzyme activity.

Table 5. Catalytic activity of Escherichia coli expressed rabbit FMO 1 fused with various
peptides at either the N- or C- terminal ends.
Strain/ Promoter,
Plasmid selection, Expression Activityt Notes
ori
XL1Blue/ Ptac, Low ND(a)(b) expressed as MalE fusion
pAAP24 AmpR, -100Kda protein (N-terminal)
pBR322 band not catalytically active
BL21(DE3)/ T7, High 1.3 x 104(a) His6 tag (N-terminal)
pAAP38 KanR
PBR322
BL21(DE3)/ T7, High 9.2 x 106(a) His6 tag (N-terminal)
pAAP41 KanR, His6 tag (C-terminal)
PBR322
BL21(DE3)/ T7, His6 tag (C-terminal)
pAAP46 KanR, not catalytically active
PBR322 no protein apparent
*expression was determine from SDS-PAGE analysis o total cell extracts, monitoring for the presence of
-60Kda protein
activity (mmole/min/mg protein) was measured as consumption ofNADPH in a) MTS coupled oxidation
assay using crude membrane protein preparations b) activity of both the cytosolic and membrane
associated fractions was assayed.

Increasing the density of cell cultures is another strategy to increase the

concentration of catalyst. Increasing the cell growth might therefore lead to greater

sulfoxidation activity. Cell growth is already reduced due to the induction of FMO

production, which poses a load on the transcriptional and translational resources of the

cells. Induced cells reach 25% lower cell densities than those uninduced (Figure 33).

Lowering the concentration of salt in the culture media increased the cell growth

(Figure 33). Higher cell density in low salt medium (TB as opposed to LB) concurrently

has increased the biocatalytic activity of each cell culture (x 1.5).










For sample handling of cultures numbers required for library screening 96-

microwell plates are commonly used, accommodating up to 1-2mL cultures per well.

Growth of Escherichia coli cell cultures in microwell plates is not as rapid as in flask

cultures. It is thought that the limited amount of aeration provided by agitation of small


volumes in microwells to be the main cause. 194 The growth of Escherichia coli cultures

in shake flasks (250mL flasks with 100mL culture at 200rpm, 37C) reaches stationary

phase faster and at higher cell densities than in multiwell plates (1.5mL culture/well at

200rpm, 37C) (Figure 33). This necessitates longer incubation times for the growth and

protein expression of small cultures.






A B
uninduced cells induced cells
4 (LB medium) (TB medium)
4 shake flask
3.5 induced cells 3.5
3 (TB medium)
E E 3
S 2.5 .induced cells o 2.5
2. (LB medium) induced cells
0 0 2 (TB medium)
1.5 15 microwell plate
1.5 1.5


0.5 0.5
0 0
0 5 10
time (hrs) 0 time hr) 10



Figure 33. Growth rates of recombinant Escherichia coli cells: A) comparison between
cells, induced for FMO expression, grown in high (LB) and low (TB) salt
media. B) comparison between growth in shake flasks and microwell plates at
37C and 200 rpm.

Overall the sulfoxidation activity of a given culture of the Escherichia coli

expression system for FMO was increased 14 fold from an initial cloning system and is

optimized in low salt media using kanamycin antibiotic selection and with a N-terminal






57


His6-tag. This improves biocatalysis reactions in microwell plates, which are needed for

the easy handling of large numbers of samples in library screenings.














CHAPTER 3
GENERATING LIBRARIES OF MUTANT FMO'S

Error-Prone PCR; Controlled Random Mutagenesis

Error-prone PCR attempts to produce mutants of enzymes by subverting the

natural fidelity of Taq polymerase in PCR DNA amplification reactions. Introduction of

base pair substitutions by Taq polymerase are in turn translated into changes in the amino

acid sequence of the gene product, effectively producing mutant enzymes.

Increasing the concentrations of Mg2+, adding Mn2+, increasing and unbalancing

the concentrations of the four dNTP's, increasing the extension time and the

concentration of Taq polymerase all reduce the fidelity of PCR.In fact, the concentration

of Mn2+ in the reaction solution is directly proportional to a loss of Taq polymerase

fidelity195. The concentrations of Mn2+ added can then be used to control and fine tune

the frequency of mutation. Caldwell and co-workers have altered the mutagenesis rates of

Taq polymerase in increments of as little as 0.5% (mutation/base pair). 196

Mutations can be deletions, insertions or base substitutions. In random

mutagenesis base substitutions are preferred as all other mutations would result in a shift

of the reading frame and an inactive enzyme. 197 Specific reaction buffers and

concentrations of Mg2+ can be used to favour this. Within within base substitutions

though there is also an undesirable preference in error-prone reactions, a strong bias

towards transitions or purine-purine and pyrimidine-pyrimidine changes (specifically

A-G, T->C) over transversions. This preference can be reduced somewhat although not









completely by using unbalanced amounts of the dNTP's, specifically increased dCTP and

dTTP concentrations. 198

Since the effect of Mn2+ on the mutagenic frequency of Taq is dependant on the

composition and length of the template and on the primers used the mutagenesis of each

gene must be carefully calibrated first. This can be done by subjecting the wild type gene

to mutagenesis reactions with various concentrations of Mn2+ and sequencing the clones

that are produced to identify the desired mutagenic frequency conditions.

The Taq polymerase that was used in the mutagenesis experiments was a

generous gift from Dr. Michael Thomson. This polymerase had an added N-terminal His6

tag to aid in its purification which did not seem to affect it in the non-mutagenic PCR

reactions (including 5mMMg2+, pH 8.0). In the presence of Mn2+ however this

polymerase did not show any activity. It required specific conditions to polymerize DNA

in the error-prone reactions, including : 2mMMg2+, pH 9.0, 0.01% Triton X-100,

0. Img/mL BSA and 10mM(NH4)2SO4. Another change of conditions which which gave

error-prone product was using 8mM Mg2+ at pH 8.4. The added His6 tag of the

polymerase might be completing Mg2+ or Mn2+ from solution.

Rabbit FMO 1 was mutated in an initial set of error-prone PCR experiments

following the scheme in Figure 34. The error-prone PCR reactions on the wild-type

template rabbit FMO1 was done in several reactions to yield enough DNA for the

subsequent cutting and cloning into the expression plasmid (pET26(b+)). These ligated

plasmids were then transformed into Escherichia coli cells. From here plasmids were

purified in large scale and sequenced while the mutagenized genes were also expressed

and tested for their sulfoxidation activity.

















EP-PCR AMPLIFICATION
Mn +

dGTP, dATP (0.2mM)
dCTP, dTTP (1.0mM)
Nde I FMO forward
1..... primer

FMO reverse EcoR I
primer .......
Semi-purify products


wt FMO
II c iFAI '


uncut
wt FMO



........ .. .. .. .....


SEcoR I


transform
EXPRESSION SYSTEM into E. coli .%.4

for singly mutated FMO
contaminated with wt -.............

Figure 34. Error-prone PCR and cloning into expression systems. Indicated are the
possible sources of wild-type gene which can carry over into final
mutagenized library.


Some /
wt FMO -




..........


Nde I EcoR I
library of
singly mutated
clones




~I



Cut with
Nde I &
EcoR I


Nde Ib EcoR I
library of
singly mutated
clones








Ligate into
expression vector
144*I









Ligated products must yield at least -10 000 distinct clones in Escherichia coli to

produce a library covering the total number of possible singly mutated variants (20 x

535). Transformation efficiencies of these libraries using electroporation ranged from

2000-4000 individual clones per reaction and several electroporations were performed to

achieve a full library.

To calibrate mutagenic frequency (EP-rFMO1-G through K) several different

concentrations of Mn2+ were initially used, ranging up to the highest reported in the

literature (0.5mM197,198). There was however noticeably less PCR product at 0.3 and

0.5mM Mn2+ than in the non-mutagenized reactions, possibly indicating that these

concentrations inhibit the activity of the polymerase.

After transformation and plating, individual colonies of Escherichia coli should

each have a separate clone of rabbit FMO 1. Sequencing a handful of these clones from

each mutagenesis reaction gives an indication of the mutagenic frequency. To limit the

amino acid substitutions to 1 per gene, mutation at the DNA level should be kept to about

0.125-0.3%.84 In a 1600 base pair gene that is about 2-4 base pair substitutions per clone.

Sequencing of these clones does not however coincide with an increase in mutagenic

frequency with Mn2+ concentration as reported. 198









Table 6. Mutagenic frequency of EP-PCR reactions with various concentrations of Mn2+
EP-PCR libraries (#C, D, E, F, G) generated from wild type rabbit FMO1.

[Mn ] Clone bp Deletions/ # DNA Non-silent Mutation
mM # sequenced Insertions mutations mutations Frequency
(library #)(%
0.00 (G) 1 564 0 1 0 0.18
2 453 0 0 0 0.00
3 632 0 0 0 0.00 0.06
4 463 0 0 0 0.00
5 732 0 1 1 0.14
0.10 (H) 1 536 0 1 1 0.19
2 610 0 1 1 0.16
3 375 0 0 0 0.00 0.07
4 701 0 0 0 0.00
5
0.20 (I) 1 425 0 0 0 0.00
2 463 0 2 1 0.43 0.13
3 455 0 1 1 0.22
4 631 0 0 0 0
5 234 1 -
0.30 (J) 1 645 0 2 1 0.31
2 423 0 1 0 0.24
3 566 0 3 3 0.53 0.22
4 512 0 0 0 0
5 289 0 0 0 0
0.50 (K) 1 443 0 0 0 0.00
2 623 0 2 2 0.32 0.17
3 530 0 0 0 0.00
4 584 0 3 2 0.55
5 433 0 0 0 0

It is noticeable that within the clones from each Mn2+ concentration used there

were several which did not have any mutations. This was taken to be carry-over


contamination of the wild type gene.










2.7
^ wt rab FMO1


22-2

o4


S1.25 -
I-I 3
B 15 3s 45 bo
MINUTES
Figure 35. Activities of rabbit FMO 1 clones (EP-rFMO -J). NADPH consumption
(measured at 340nm) during the sulfoxidation reactions of MTS with crude
membrane fractions from (#1-5) cells expressing clones of EP-PCR library of
rabbit FMO1 (EP-FMO1-J), cells expressing the wild type rabbit FMO1 and
(-)'ve: the native Escherichia coli strain.

The activity of 5 of these cloned genes indicate (Figure 35) that 3 harbouring no

mutation possess wild type activity, the remaining inactive enzymes are all mutants

(Figure 35, Table 6). The resulting library of mutants even though they possess on

average the correct mutagenic frequency, are likely composed of a subset of wild type

and a sub set of mutant genes. This effectively increases the size of the library that must

be screened to sample all the possible mutant variants. Such a problem in cloning of


mutagenized genes has been reported once previously. 199


To remove all the possible sources of wild-type contamination a different cloning

strategy was adopted (Figure37). By changing the antibiotic selection the template wild

type is no longer a viable plasmid if carried over to the final transformation. This is

guaranteed by digestion of the non-methylated template with Dpn I.


Another set of mutagenesis reactions using Mn2+ were cloned into Escherichia

coli using this strategy and the resulting mutagenic frequency appears to have a more

linear progression with increasing mutagen concentration as expected (Table 7).









Reactions of error-prone PCR using 0.15-0.2mM Mn2+ cloned into the expression

systems were prepared for screening of sulfoxidation activities.


Table 7. Mutagenesis and sulfoxidation activities of EP-rFMO1-J library.
Clone mutated bp/ Mut Amino acids % activity*
bp sequenced Freq.% mutated from wt of wt
no FMO (-'ve) 0
wt rabbit FMO1 100
EP-rFMO1-J-1 2/271 0.74 2 (G49E,V81A) 5
EP-rFMO -J-2 1/756 0.13 0 74
EP-rFMO1-J-3 0/652 0.00 0 88
EP-rFMO1-J-4 0/643 0.00 0 72
EP-rFMO1-J-5 2/398 0.50 2 (V46A,L99P) 3
EP-rFMO1-J-6 3/650 0.46 2 (S102R, L149P) ND
activity trom NADPH consumption assay ot MIS sultox nation of wild-type rabbit FMO1 was (100%) 5.5 x 10-3 mmole/mn/mg
membrane protein, ND = assay not done

Table 8. Mutagenic frequency of EP-PCR reactions with various concentrations of Mn2+. EP-
PCR libraries (#L, M, N, O, P) generated from wild type rabbit FMO1.
[Mn2+] Clone bp Deletions/ # DNA Non-silent Mutation
mM # sequenced Insertions mutations mutations Frequency
(library #) (%)
0.00 (K) 1 634 0 1 0 0.16
2 750 0 0 0 0.00 0.05
3 570 0 0 0 0.00
0.10(L) 1 453 0 1 1 0.22
2 676 0 1 1 0.15 0.12
3 534 0 0 0 0.00
0.15 (M) 1 376 0 1 1 0.27
2 658 0 1 1 0.15 0.19
3 548 0 1 1 0.19
0.20 (N) 1 590 0 2 1 0.34
2 436 0 1 0 0.23 0.33
3 786 0 3 2 0.39
0.25(0) 1 567 0 2 2 0.35
2 578 0 3 3 0.52 0.34
3 643 0 1 1 0.16

















EP-PCR AMPLIFICATION

Mn++

dGTP, dATP (0.2mM)

dCTP, dTTP (1.0mM)

Nd I FMO forward
.... primer

FMO reverse EcoR
primer ...... .

Semi-purify products


uncut
*wt FMO


SEcoR I


EcoR I


EXPRESSION SYSTEM

for singly mutated FMO


Ligate into
expression vector


Transform into
E. coli

SNde I


Digest with
Dpn I


Figure 37. Error-prone PCR and cloning into expression system. Strategy to remove wild-
type contamination from the final mixtures. nm = non-methylated DNA.


: Some "
wt FMO S F





4 -- -
+
Ndel EcoFII
N library of
singly mutated
clones








Cut with
Nde I &
EcoR I

Nde I EcoR I
library of
singly mutated
clones





I -I









DNA Shuffling and Recombination Mutagenesis

Another method to alter genes involves the combination of fragments from

different full length sequences into a set of new genetic combinations. This DNA-

shuffling method developed by Stemmer85 can be used to combine sequences with few

variations (such as the products of error-prone PCR mutagenesis) or sequences of related

genes as done in "family shuffling".200

Reshuffling in this way allows one to combine beneficial mutations from mutants

selected out of mutagenesis screening into one gene clone. The accumulation of

beneficial mutations is believe to have an additive character88 allowing one to increase

the improvement of the trait selected for. The probability of combining positive

mutations by this method would be greater than from an attempt to introduce them

sequentially by successive rounds of mutagenesis and selection. In family shuffling two

related naturally occurring genes are shuffled with each other. Due to similarity in

function and structure between even quite different sequences the exchange of segments

is hoped to retain an overall active arrangement.201 Traits that are distinct to the two

might be found combined in an active clone.

With this strategy we shuffled the sequences of rabbit FMO 1 and rhesus macaque

FMO2, two isoforms with 56.5% identity in amino acid sequence (Figure 38 & 39) but

with different catalytic properties. Rabbit FMO 1 has higher activity for the catalysis of

sulfoxidation of MTS whereas the monkey enzyme has a greater preference for the

oxidation to the (S) sulfoxide (Table 2). It is hoped that DNA shuffling of these genes

might produce a highly (S) selective catalyst with high activity.















rahMO 291 A T ,." '
rib PX02 2i A0aA iurB












Figure 38- Multiple sequence alignment of Rabbit FMO 1 and Rhesus macaque FM2 genes (DNA) (ClustalW). Identical residues are












shown with a black background. These genes have 65% identity (1045/1608bp aligned identically). The FAD binding site
is at location 25-42 (GGA GCT GGG GTG/c AGC/T GGC)178 and the putative NADP binding site at 571-588 (GGA ATG
GGC/A AAT/G TCT/G GGC).
rFl02 141 C U0VCAW i1 CatcE M.cT iA GA OCX T COCUUCI
rabPFMO 2I 1 ArO -AC U U .








runo2 1249 Ai4 AEE iS^U^ <.iTO jA AIA U riAir 'A riIA,;a :.... ...


rabFMOl 1113 5TAA.





rabtMOl 15l2 an (GeIAGCT- -GC nd pt Nin e -5 -1-- -58M8AT( T







GGC/A AAT/G TCT/ GGC).























rabFMO1 1 2
rmFM02 1I =


rabFMO1 91i M- i-imB a I'.'-i' E- .n E .*EE ..-, ;ig:
rmFMO2 91 F F Fl1 1Ei F-, M ELI- i ill IJ B 'B E i'n- ^


OO
rabFMO2 181 !li 1M 7fVqMM5TR H, ,_ [. T -jT -.l[M" -EPr M TM A F ,M- F in L.' I, i '-. lMIn i E '. t
rmFM02 181 ,'IILj .L IL hi.- E.i M L I L i M.E-,

rabFMO1 271 a L j
rmFM02 273 1 .; Ii L I : T i II Fri- *1 j ML A *jj *' .'M E1L L^ ilNi L
rabFMOl 361 LI.3(E LBI i, LIEi r .Mt i1 *:, Aff 1-BI T IBL. LTLnl mE

rabFMO1 451 lLMFjL i EU L3 *: al i M riLB TIjFE- LFfL LL j LIFL-
rmFM02 448 L:I VF. L L LLiL F : E^ LL :i


Figure 39- Multiple sequence alignment of Rabbit FMO 1 and Rhesus macaque FMO2 proteins (ClustalW). Identical residues are
shown with a black background for both proteins. There are 302/535 identical residues (56.5% identity). FAD binding site
aa9-14 (GAGVSG)178 and NADP putative binding site aa 191-196 (GMGNSG) are both conserved between the two
enzymes.


E3LEP TCLIE:Pi


r.jLWLMJ7jHffiE L !M- --jZq-Mi F-Umi FM-F r.
--FMMF"MFLMi' LIE1










DNA shuffling is the fragmentation and reassembly of genes into chimeras

(Figure 40). There are several procedures which have been used to fragment and


recombine full length sequences.85,91,202,203 A method using Dnase I digestion and

self-priming reassembly was suggested by Dr. C. Martinez in personal correspondences

(Appendix A). It is critical that fragmentation to 1-200bp fragments be complete to avoid

carry over of the wild type sequences. The activity of Dnase I enzyme appeared to

decreased rapidly with storage time and it was necessary to recalibrate the enzyme before

each digestion.


DNA SHUFFLING
rabbit FMO1 rhesus macaque FMO2
I I II r l I I I I I I [ I I I

Random fragmentation (Dnase I)




I


I denaturation


Sannealing, self-priming
Reasse


Polymerase extension



amplification with primers
for ends of both genes

chimeric genes to be cloned
into expression plasmids


mbly PCR


Figure 40. General overview of DNA Shuffling. Two or more isolated parent genes are
initially digested into random 1-200bp fragments then reassembled using self-
primed PCR into a variety of chimeric combinations.

Recombination PCR is prone to add mutations. One way to circumvent this is

through the use of high fidelity polymerases. Pfu polymerase is about 6 times less likely


\1










to incorporate a error during amplification than is Taq polymerase, with error rates of 1

out of every 770 000 and 1 out of every 125 000 base pairs respectively 204. Reassembly

reactions were done with 1:1 mixtures of Pfu and Taq polymerases.202


The random recombination of the gene fragments was checked using a PCR time

course (Figure 41). Amplification of the chimeras using all four combination of primers

for the ends of both genes is predicted to proceed at similar rates for completely

randomly shuffled genes. A parallel increase in the expected 1.6Kbp fragment of

amplification was observed from all 4 primer sets.




S 1 cycle 4 cycles 7 cycles 10 cycles
-I r ----- r i -------
ABCDABCDABCDABCD
FMO
S-1.6Kbp






Figure 41. Confirmation of random assembly of gene fragments into chimeras of

rabbit FMO 1 and rhesus macaque FMO2. DNA shuffled reactions were amplified with a

combination of primers A)rabFMO1 for, rmFMO2 rev;B)rabFMO 1for,rabFMO 1rev;

C)rmFMO2for,rabFMO rev; and D)rmFMO2for, rmFMO2rev. PCR were sampled at

different increments during the reactions and checked on agarose gel for FMO gene

(-1.6Kbp).














CHAPTER 4
BIOCATALYSIS REACTIONS AND SCREENING

The successful directed evolution of enzymes is largely dependent on the

screening or selection method employed.94 Although the speed and ease of the screening

is important as it limits the size of libraries that can be investigated the sensitivity of the

method will determine if improved mutants are detected at all. Each round of mutation

produces a range of changes in the property that is being investigated. At the start of this

project approximately 40 different enzymes had been 'improved' using directed evolution

in a variety of properties; thermostability, activity towards a certain substrate, activity in

organic solvents and expression levels being some of them.205 At the time though there

were few examples of altering the enantioselectivity of an enzyme.


Bornsheuer and co-workers 99 had used a powerful selection technique to screen

over 100 000 clones of an esterase in the first round of mutagenesis and uncovered a

clone with a change of enantioselectivity from the wild type of 25% ee. Arnold et a/.206

have screened for the kinetic resolution of the enzyme hydantoinase with a 20% increase

in preference for the hydrolysis of the D enantiomer of the substrate.This resulted after

screening of the first library of 10 000 clones. Reetz and co-workers produced a clone of

a lipase with a change in relative rates from E = 1.1 to E = 2.1 towards preference for one

enantiomer of a substrate from a library of 3000.79 With these few results it is difficult to

predict what the minimum accuracy needed would be for the screening to detect

improved variation in enantioselectivity from a library of singly mutated rabbit FMO1.









The enantioselective oxidation of sulfur is an enzymatic ability which has as yet

not been investigated using directed evolution. To probe the variants of an

enantioselective sulfoxidation catalyst such as FMO, a method has been developed to

produce and present the enzymes (Chapter 3) and use them in oxidation reactions. A

tailored biocatalysis reaction must be developed which would be amenable to a rapid and

sensitive screen of both the product yields and product stereochemistries. A screening

procedure with high sensitivity and efficiency must also be developed.

Sulfoxidation Using FMO Biocatalysts

The Flavin monooxygenase enzyme have previously been observed to catalyse

oxidation of organic sulfides. In fact FMO's oxygenate a wide array of aryl alkyl sulfides

with varying stereoselectivities (Table 8). This chemistry was first described for the

porcine isoform 1 purified by Ziegler and coworkers. 11 Purification of the enzymes

from their native source tissue is a labor intensive and lengthy procedure. 111 Later,

heterologuous expression systems were developed for rabbit FMO 1 and FM0264 and

rhesus macaque FMO2.65 These produced mammalian enzymes in microbial cells with

identical properties to the native forms but with much more ease and speed of

preparation. 126 The activities of sulfoxidation of several aryl alkyl sulfides by various

FMO enzymes were investigated using such expressed enzymes (Table 9).









Table 9. Sulfoxidations of a variety of aryl-alkyl sulfides by purified FMO's (ip) and
crude-membrane preparations of heterologously expressed FMO's (cmp).
Substrate FMO e.e.%t Enzyme* Ref.
Isoform Preparation
s pig FMO1 96% (R) ip 111
rabbit FMO1 98% (R) cmp
rabbit FMO 1 >99% (R) cmp 64
rabbit FMO2 >99% (R) cmp 64
rabbit FMO3 50% (R) cmp 64
rhesus macaque 30% (S) ip 65
FMO2
human FMO3 57% (R) cmp 207
sq rabbit FMO1 98% (R) cmp 64
rabbit FMO2 91% (R) cmp 64
rabbit FMO3 50% (R) cmp 64
human FMO3 49% (R) cmp 207
s / rabbit FMO 1 96% (R) cmp 64
H2)2 rabbit FMO2 86% (R) cmp 64
rabbit FMO3 72% (R) cmp 64
human FMO3 75% (R) cmp 207
,s rabbit FMO 1 96% (R) cmp 64
rabbit FMO2 70% (R) cmp 64
pig FMO1 55% (R) cmp 170
s / rabbit FMO3 88% (R) cmp 64
H2)3 human FMO3 88% (R) cmp 207

*cmp: crude membrane preparation of heterologuously expressed FMO' p: isolated
enzyme tenantioselectivity determined by chiral GC-MS analysis of product sulfoxide.

Although suitable for investigation of single reactions between an isoform of

FMO and a single substrate, crude membrane preparations (100 000 x g particulates

fraction) of heterologuously expressed enzyme do not lend themselves to rapid, multi-

sample screening. A less complex strategy to present heterologously expressed enzyme

for biocatalytic reactions is from intact, whole cells.

Whole cell biotransformations with the E. coli expression systems

(BL21(DE3)/pAAP 11 and BL21(DE3)/pAAP27) for rabbit FMO 1 though show no

activity in the sulfoxidation of MTS. The reactions were monitored over 3 days at various









temperatures (rt, 300C and 37C) for MTSO production by GC analysis. Protein is

detectable from SDS-PAGE analysis and is active as microsomal (semi-isolated)

preparations, whereas unlysed cells do not have catalytic activity. It should be noted that

a similar system with a different sulfoxidation enzyme also did not show any activity in

whole cell reactions with this substrate. Whole cell biocatalysis with E. coli

(BL21(DE3)/pMM04) overexpressing cyclohexanone monooxygenase produced no

sulfoxidation product of MTS after 19 hrs. incubation.208

It was also noted that for whole cell biotransformations in shake flasks there is a

disappearance of 50-60% of the starting material (MTS) by the end of a 62hrs assay,

although no extractable product (MTSO) is discernible. Since the disappearance follows a

very linear rate over 62hr and occurs in the absence of cells it could be that the starting

material is escaping perhaps being lost to vapour. Reactions with 100mmole MTS were

performed using 100mL unlysed Escherichia coli cultures at 200rpm and at 25, 30, 37C.

The expression system for rabbit FMO 1 based on the pET26b(+) plasmid vector

(BL21(DE3)pAAP27) does express active protein. This activity of this protein is not

apparent unless the cells are lysed. There is no precedent for whole cell biocatalysis using

FMO's in literature. For the sulfoxidation of MTS we have not been able to observed any

reaction with various unlysed expression systems for either rabbit FMO 1, rabbit FMO2

or rhesus macaque FMO2.

In case the access of substrate to the enzyme can be overcome by

permeabilization of the outer membrane several known Escherichia coli permeabilization

agents have been used. Metal chelators209, DMSO210 and polyethyleneimine211 though









did not seem to improve whole cell biocatalysis. This indicates that cell lysis and the

subsequent provision of co-factors seems to be required for observable catalysis.


Table 10. Sulfoxidation of an array of alkyl aryl sulfides catalyzed by crude membrane
preparations of Escherichia coli expressing rabbit FMO1. Reaction done with
1 mole substrate and yields after 20hrs reactions determined by GC.

Substrate Yield %
osY 88%

s 64%


42%



]Sv 45%

S 72%
Br
69%
OH O
O s npd



s., 43%

/"-sC, 32%

*cmp : crude membrane preparation (100 OOOg particulate traction), p : isolated enzyme
t npd : no product detectable

Crude precipitated membrane preparations of rabbit FMO 1 do possess catalytic

activity for the oxidation of several alkyl aryl sulfides (Table 9).


Methods for cell lysis used during biocatalysis screening must be gentle enough to

be non-disruptive to enzyme viability and rapid enough to be used on large numbers of

samples simultaneously. Sonication lyses cells by liquid shear and cavitation,










homogenization by pressure lysis, both can be used for active enzyme preparations212

although they cannot easily accommodate fast throughput. Enzymatic lysis using

lysozyme or snap-freezing is preferred. Large number of cell cultures (in 96 microwell

plates) initially washed of media and resuspended in reaction buffer (with 0. 1mMEDTA)

can be quickly lysed at.80C (15-30min). If supplied with enough co-factor (NADPH),

1.5mL of cell culture expressing rabbit FMO1 prepared in this way can oxidize 1.5ptmole

of MTS in 30hrs to 80% sulfoxide. The addition of protease inhibitors (such as PMSF)

does not seem to improve the activity of lysed cells.


The biotransformation reaction of MTS using the flavin monooxygenases requires

a buffered solution (pH 8-9.5) and a supply of reducing equivalents, either as NADPH

(1.1 molar substrate equivalents) or NADP+(0.1 molar substrate equivalents) along with a

suitable regeneration system (Figure 42). The rhesus macaque FMO2 and rabbit enzymes

have pH optimums of 9.5127 and 8.264 respectively for the oxidation of MTS and n-

alkyl amines.


H+ 02
NADPH or NADP+and
regeneration system

7


FMO crude membrane
preparation
Figure 42. Biotransformation of MTS by FMO presented as crude membrane
preparations from Escherichia coli based expression systems.

A regeneration system reduces the amount of cofactor required per reaction, an

important advantage in libraries of 10 000 reactions. A suitable system requires an

enzyme to reduce the oxidize NADP+ and provide FMO with a constant supply of









reducing equivalents. A regeneration enzyme-substrate often used for this purpose is

glucose-6-phosphate, glucose-6-phosphate dehydrogenase.

To perform reactions in aqueous media substrate and product solubilities become

a factor also. The substrate sulfide in this case is not water soluble and is visible as a

layer on the surface of the solutions. To avoid volatilization of the substrate and increase

reaction yields, reaction vessels are tightly capped and detergent (Triton X-100 or

Brij-35) or cyclodextrins (P-hydroxycyclodextrin) can be added.

Screening Biocatalysis Reactions

Sulfoxidation reactions by FMO mutant will possess possible differences in both

yield (rates of reaction) and in enantioselectivity. In taking advantage of the physical and

chemical properties of the products to design a screening, a method with the flexibility

for application to other systems is desirable.

Coupling to other reactions or modification of the reactants or products to produce

an easily monitored characteristic could limit expansion of screening which would be too

specific for the substrate initially used. There is also a lack of enzymatic or other

chemistry that can be used to distinguish between (S) and (R) MTSO.

There are chiral shift reagents for 1H-NMR spectroscopy that can be used to

distinguish the enantiomers of MTSO213 and there are also co-crystallization agents

developed for the resolution of the two enantiomers. (R) enantiomer of the sulfoxide

(MTSO) selectively will crystallize as a 1:1 complex with 9.214 The crystal structure

indicates that 1 molecule of 9 is hydrogen bonded through NH to the SO group of the

sulfoxide.











NN O2


9 NO2
Figure 43. Chiral amide, forms crystals with molecules of (R) -MTSO (1 molecule
amide: 1 molecule sulfoxide).214

Some of the distinguishing properties of the two sulfoxide enantiomers that could

be of use in determining enantiomeric excess are: the different directions each rotates

plane polarized light and the distinct interaction each has with chiral environments (the

principle of chiral GC and HPLC chromatography).

Short chain alkyl p-tolyl sulfoxides are water soluble molecules with relatively

high specific rotation (in aqueous solution [a]D20(R)MTSO = 1450). For determination of

enantiomeric composition of a reaction with optical rotation though, both the rotation of

light and the concentration of the substance must be determined.

Screening Biotransformations Using UV/VIS and OR Detectors

Developing Methodology

Short chain alkyl p-tolyl sulfoxides are water soluble molecules with relatively

high specific rotation (in aqueous solution [a]D20(R)MTSO = 1450). For determination of

enantiomeric composition of a reaction with optical rotation though, both the rotation of

light and the concentration of the substance must be determined.

The chromophoric molecule p-nitrophenol absorbs in the visible spectrum and can

be screened in complex mixtures with sensitivity (Emax = 10 000 max = 318nm), it has

been incorporated into many screenings of enzymatic reactionsl9. This molecule has an

increased intensity of a K band corresponding to a 7c-** transition due to a charge










transfer absorption (Figure 44).215 The two auxochromic groups (the -OH group being

electron donating, the -NO2 group being electron attracting) are para- to each other.

Similar orientation of auxochromic groups -Br/-OCH3 para to the sulfoxide electron

attracting group on a phenyl ring though does not produce such a visible effect (Figure

45) although both are substrates for rabbit FMO1 (Table 9). A colorimetric assay would

have been helpful in monitoring the concentration (yield) of sulfoxide product.


OH 0 NH2 OCH3 Br CH3




oN O- 0-NO O ~NO 0 -S'CH3 0 SCH3 0 SCH3

max= 318nm ,max= 381nm Xmax= 245nm hmax= 240nm Xmax= 233nm
Figure 44. The addition of two auxochromic substituentspara to each other and charge
transfer resonance forms influence the absorption characteristic of phenyl
rings.

These analogs do show stronger UV absorbance than MTS (Figure 45) in the

region 200-250nm. The stereochemistry of the rabbit FMO1 sulfoxidation reactions when

the para substituent is changed has though not been investigated.


A sz B sI C


Hc3 C CH3 0S Br) S

-Bprod. \ prod.

Ssm. sm.


Figure 45. Comparative UV/VIS spectra of biotransformation reactions on various
substrates. Starting material (s.m.) at the beginning of the reactions and the
products (prod.) after 24hrs. biotransformations. Spectral analysis directly from
cell lysate reaction mixtures of rabbit FMO 1.










The measurement of yield of the sulfoxidation reactions using UV was not that

straightforward since there was no wavelength that could be used to monitor the product

of the reaction uniquely at which the starting material didn't also possess some

absorption. There will be some experimental variation between the amounts of starting

material added over the large number of reactions to be screened. Determination of yield

in this case requires knowing both the concentrations of unreacted substrate and product

at the end of the reactions. By choosing two absorption wavelengths along the spectrums

of both compounds (Figures 46) we can solve for the concentrations of each.


Many solubilizers which could be used to help disperse the substrate in the

reaction solutions (such as cyclodextrins and Triton X-100) produce too high signals in

the chosen UV range to be practical. Instead, the non-ionic detergent Brij-35 (a saturated

alcohol) was added to reactions and prior to sampling they were diluted with 3 volumes

of EtOH.



UV233nm


.36
3 *7 mtso mtso UV2s4nm

.a mts


1.iz ax PA254nm -P
[MTSO]= PA23
a E4M 2. 20 B dxa b
d
UAUELENGTH
Figure 46. UV spectrum of MTS and MTSO, indicating the distinct absorption maxima;
MTSO (233nm) and MTS (254nm). The concentration of MTSO can be
determined for each reaction from measurements at both wavelengths and by
solving the equation* (Appendix B), concentrations of unreacted MTS can be
calculated with a similar equation.











Lysed cell suspensions contain a variety of particles and biomolecules. Several

strategies can be used to reduce the possible background these might have on both the

OR signal and the UV signal of the products. The addition of EtOH to 75% helps

precipitate proteins from solution. Samples were also centrifuge and filtered (0.2|m) to

remove the majority of the suspended molecules. Even with these precautions there was

considerable background at both 233nm 254nm (Appendix B) and in the OR signal

(Appendix B). Since the growth of the cell cultures was a variable this background varied

between samples. The spectra of a solution of cell extract displays a characteristic peak at

280nm, a region where neither the starting material nor product have any significant

absorption (Figure 47). Using this wavelength as a handle the background in all the

measurements (UV233, UV254 and OR) could be calibrated and taken into account for each

sample individually.



1600000
1400000
1200000
1000000
800000
600000
400000
200000

01 025 05 075 1
dilutionof cell extract

Figure 47. Cell extract background in the biotransformations. UV spectra of cell extract
in 75% EtOH with a peak at 280nm.Graph indicating the relationship between
the amount of cell extract with the signal at 280nm.

Optical rotation readings with the OR detector did not appear to be very sensitive

and required a higher concentration of product (even through the small 40L flow cell)

than was required for UV measurements. In fact the sensitivity of the UV detector is

-4000 x higher than ORD for MTSO (measured at 233nm). To accommodate this









inequality dilutions were done after OR readings to each sample (100 fold dilution)

before measurement by UV. Concentrations of MTSO above 0.05mM saturate the signal

from the UV detector (Appendix B). Dilutions can be performed automatically using the

Gilson 215 autoinjector/liquid handling system although they require extra time and

present another source of errors.

To increase the signal read by the ORD the reactions can be scaled up. Methods to

increase the production of enzymes were discussed in Chapter 2. Another factor that

could influence the signal would be change of the solvent environment216. In organic

solvents (EtOAc, hexanes, Et20) which could be used to extract the reaction products, the

ORD signals are actually lower for non-racemic MTSO (Figure 48). Miscible solvents

such as EtOH or MeOH do not affect the signal appreciably (data not shown).














UV Spectrum


Water:


Diethyl ether: C












EtOAc: E


Optical Rotation Signal


Hexane: G U- H
Figure 48. A comparison of the UV and ORD signals of MTS and MTSO in various
solvents. (A,C,E,G): UV spectra (from 200 to 300nm) of 0.04mMMTS (solid
line) and MTSO (dashed line). (B,D,F,H) ORD chromatogram showing
relative intensities of signals of 4mM (R)MTSO in each solvent.


i