A MECHANISTIC INVESTIGATION OF MIO-BASED AMINOMUTASES By Udayanga Wanninayake A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Chemistry Doctor of Philosophy 2013 ii ABSTRACT A MECHANISTIC INVESTIGATION OF MIO-BASED AMINOMUTASES By Udayanga Wanninayake -Amino acids are biologically active compounds of i nterest in medicinal chemistry, which are used as precursors for the biosynthesis of seve ral biologically active compounds such as taxol, andrimid, chondromides and C-1027. Also they are used as important precursors for the synthesis of -lactams and -peptides. A class I lyase-like family of aminomuta ses isomerizes (S)- -arylalanines to the corresponding -amino acids by exchange of the NH 2/H pair. This family uses a 3,5-dihydro-5-methylidene-4 H-imidazol-4-one (MIO) group within the active site to initiate the reaction. The structures of a pheny lalanine aminomutases from Taxus canadensis (Tc PAM) and from Pantoea agglomerans (Pa PAM) have been determined at 2.4 Å and 1.7 Å resolution, respectively. Tc PAM catalyzes the isomerization of ( S)- - to ( R)- -phenylalanine, making ( E)-cinnamate (~10%) as a by-product at steady state. In contrast, when ( S)-styryl- -alanine is used as a substrate, Tc PAM produces (2 E,4 E)-styrylacrylate as the major product (>99%) and ( R)-styryl- -alanine (<1%). Comparison of the rates of conversi on of the natural substrate ( S)- -phenylalanine and ( S)-styryl- -alanine to their corresponding products ( kcat values of 0.053 ± 0.001 and 0.082 ± 0.002 s Œ1 , respectively) catalyzed by Tc PAM suggests that the amino group resides in the active site longer than styrylacryla te. To demonstrate this principle, inhibition iii constants ( KI) for selected acrylates ranging from 0.6 to 106 M were obtained, and each had a lower KI compared to that of (2 E,4 E)-styrylacrylate (337 ± 12 M). Evaluation of the inhibition constants and the rates at which both the /-amino acids (between 7 and 80% yield) and styrylacrylate were made from a corresponding aryla crylate and styryl- -alanine, respectively, by Tc PAM catalysis revealed that the reaction progress w as largely dependent on the KI of the acrylate. Bicyclic amino donor substrates also tran sferred their amino groups to an arylacrylate, demonstrating for the first time that ring-fused am ino acids are productive substrates in the Tc PAM-catalyzed reaction. Burst-phase kinetic analysis was used to evaluate t he deamination rate of the aminatedŒ methylidene imidazolone (NH 2ŒMIO) adduct of Tc PAM. The kinetic parameters were interrogated by a non-natural substrate ( S)-styryl- -alanine that yielded a chromophoric styrylacrylate product upon deamination by the amin omutase. Transient inactivation of the enzyme by the NH 2ŒMIO adduct intermediate resulted in an initial bur st of product, with reactivation by deamination of the adduct. This stu dy validated the rate constants of a kinetic model, demonstrating that the NH 2ŒMIO adduct and cinnamate intermediate are sufficie ntly retained in the active site to catalyze the natural - to -phenylalanine isomerization. iv I dedicate this dissertation of my valuable graduat e research to my loving wife Nadeesha for her wonderful support and caring during my graduate career. v ACKNOWLEDGEMENTS First, I should thank Prof. Kevin Walker for his va luable advice and academic guidance. He has a remarkable ability to point out weaknesses an d emphasize my strengths that helped advanced me to the level of an outstanding scholar. His dedication and determination to be involved in science at all times was a benefit towa rd helping me achieve my goals. His scientific curiosity and inquiry have improved my cognitive an d problem solving skills. Finally, I really appreciate our friendly relationship that eased our personality and culture differences, which in my opinion help us engage in valuable scientific di scussions. Therefore, I acknowledge Prof. Walker for being the ideal mentor for me. I also th ank Prof. Dana Spence for always believing in my potential as a graduate student and for always e ncouraging me to achieve the best. I really appreciate his friendly anecdotal discussions about life and his research in the hallways of the Chemistry building. I thank Prof. James Geiger for his collaboration, valuable insights, and contributions to my thesis research. It was a great opportunity to work side-by-side with him. I appreciate him recognizing the value of my contribu tions towards our productive collaboration. Further, I thank Dr Geiger's laboratory members for sharing their instruments for protein purification. I thank Prof. Babak Borhan for being a wonderful advisor from the beginning of my graduate career. His advice for finding the right p ath in graduate school helped me during my graduate tenure. I also appreciate him as a teacher , since my success in his courses was predicated on understanding scientific concepts. Fu rthermore, I appreciate the valuable insights he provided during my second year organic chemistry seminar (CEM 958), committee meetings, and second year oral exam. I appreciate the generou s guidance offered by Dr. Ardeshir Azadnia towards me becoming an outstanding teacher. He alwa ys gave subtle clues that help improve my teaching. Furthermore, I appreciate him understandi ng the needs of international students to vi travel home; he demonstrated this by adjusting the teaching assistant schedule to accommodate international travel itineraries. Because of his ge nerosity, I was able go to my home country in the middle of the semester and get married. I alway s give special thanks to those mentioned above for writing recommendation letters on my beha lf to help me find a suitable job. I thank Dr. Brad Cox for supervising and teaching me molecular biology and biotechnology techniques during my first year. He trained me how to conduct biochemical assays and other important methods needed for my graduate research. I thank Pr of. Xuefei Huang for letting me sit in his lab in my first semester of graduate school. He seeming ly always had a friendly demeanor, and thus it was wonderful to talk whenever we met. I thank D r. Kathryn G. Severin for being a wonderful advisor during my first semester when I was working in the department of chemistry as a TE under her supervision. She was one of the friendlie st and caring supervisors I ever met. She also helped me with the UV-VIS and IR spectrometers duri ng my graduate career. I appreciate Dr. Chrysoula Vasileiou for helping me understand the P ymol software; this program became a centerpiece of my graduate studies. I thank Prof. L eslie Kuhn and Dr. Kaillathe (Pappan) Padmanabhan for their instructions on the use of th e SLIDE software. I thank Mrs. Zeynep Altinsel for being an effective instructor in the E nglish-speaking and listening class for international teaching assistants. My fluency in En glish improved because of her valuable guidance. I also thank the late Prof. Greg Baker fo r giving a lecture on "how to write a résumé," just before he passed away. This lecture helped me craft an effective résumé that, in part, enabled me to find a job before I graduated. Dr. Baker was a wonderful person to be around. In addition, Prof. Greg Baker's laboratory members graciously pr ovided an H 2 reactor, distilled solvents, and H2 gas for my reactions. I take this opportunity to fu rther acknowledge Prof. Honggao Yan in the Department of Biochemistry & Molecular Biology at M SU for allowing me to use his UV-VIS vii stopped-flow spectrophotometer and lab space to con duct a burst-phase kinetics experiment. Access to this instrument enabled me to obtain valu able data for one of my Biochemistry manuscripts. I greatly appreciate the help from Dr. Daniel Holmes and Mr. Kermit Johnson, the NMR specialists, who helped me understand the princ iples of multinuclear NMR and execute NMR experiments. I extend thanks to Dr. Thomas P. C arter for the help he provided to print my posters and design publication quality graphics for my manuscripts. Dr. Richard Staples was also instrumental in obtaining crystal structure data fo r several compounds needed as supporting information for my publications. His insights on re crystallization techniques were impressive, and we successfully obtained suitable crystals for each of my compounds of interest. I extend my gratitude to Prof. Gavin Reid for providing exact m ass data for my synthetic compounds. I also thank Prof. Gary Blanchard, Mrs. Joni Tucker, Mrs. Debbie Roper, and Mrs. Cherie Nelson for the wonderful guidance and advice to successfully c omplete the Ph.D. program as smoothly as possible. I thank the building engineer, Mr. Robert Rasico for his wonderful support and help to fix various apparati in the laboratory. I thank Mr. Bill Flick, Mr. Tom Geissinger, Dr. David Voss, Ms. Mellissa Parsons, and members of the Rese arch Technology Support Facility (RTSF) at MSU for their technical help during my graduate career. My thank goes to Prof. Mitch Smith and his laboratory members for providing us equipme nt to fix our GC-MS instrument, and to the laboratory of Prof. Rob Maleczka for providing dry distilled THF. I thank Prof. Bill Wulff's laboratory members for providing a GC column, Prof. David Weliky's laboratory personnel for providing the Nanodrop TM instrument for protein quantification. Most importantly, I thank my former lab mates: Dr. Danielle Nevarez, Dr. Irosha Nawarathne, Dr. Lei Feng, Dr. Mark Ondari, Dr. Susan Strom, Dr. Washington Mutatu, Mr. Yemane Mengistu, M.S., Dennis Quist, M.S., Getrude Dibo, M .S., and Sean Sullivan, M.S., and viii undergraduates Joshua Bilsborrow, Ebony Love and Yv onne DePorre for their valuable instructions and contributions to my research. I al so thank my current lab members, Chelsea Thornburg, Dilini Ratnayake, and Ruth Muchiri for t heir valuable advice during group meetings on my research projects. I acknowledge my colleague s Meisam Nosrati for helping me crystallize my enzyme, Amila Dissanayake, Heyi Hu, Iwan Setiwan, Kumar Ashtekar, Rahul Banerjee, Rosario Amado Sierra, and Quanxuan Zhang for their wonderful help during my graduate career. I thank Ajith Karunaratne, Lasanth a Ratnayake, Salinda Wijeratne, Tharanga Wijetunga, and my roommate Damith Perera for their emotional support and friendship. I also thank my undergraduate supervisor Prof. Veranja Kar unarathne, and my teachers, Prof. Gamini Rajapaksha, Prof. Namal Priyantha, and Prof. Rathna yake Bandara for their guidance, recommendation letters, and assistance to get me in to a graduate school program in USA. Lastly, I acknowledge my loving parents Mr. Piyasen a Wanninayake and Mrs. Muthumenika Dissanayake for raising me, educating me, and givin g me courage and strength to come to US for advanced professional training. I thank my loving p arents Tikiribandara Rajapaksha and Chandra Rathnayake for wishing me good luck and being with me during happy and sad moments. I thank my loving sisters Iresha Rajapaksha, Ruvini Rajapak sha and Thejani Wanninayake for the wonderful support and care. Finally, but certainly not the least, I offer my sincere gratitude to my loving, patient, caring, kind and beautiful wife an d life partner Nadeesha Rajapaksha ("The Trouble Maker") for providing courage, strength, an d most importantly, love. My greatest achievements came after she entered into my life. I really appreciate her dedication and resolve during stressful moments of my day-to-day life as w ell as during the challenges of my graduate work. ix TABLE OF CONTENTS LIST OF TABLES xiii LIST OF FIGURES xiv KEY TO ABBREVIATIONS xxiii LIST OF SCHEMES xxv 1. INTRODUCTION 1 1.1. -Amino Acids as Important Precursors in Biologicall y Active Molecules 1 1.2. Synthesis of -Amino Acids 2 1.3. Biosynthesis of -Amino Acids 4 1.3.1. Biosynthesis of -Amino Acids by Aminomutases 4 1.4. Aromatic Amino Acid MIO-dependent Aminomutases 9 1.4.1. 3,5-Dihydro-5-methylidene-4 H-imidazol-4-one (MIO) Prosthesis 9 1.4.2. Reaction Mechanism of MIO-Based Aminomutases 11 1.4.3. Stereochemistry of the Phenylalanine Aminomu tase Catalysis 14 2. INSIGHTS INTO THE MECHANISM OF MIO-BASED PHENYLALANINE AMINOMUTASE CATALYSIS 22 2.1. Introduction 22 2.2. Experimental 25 2.2.1. Chemicals 25 2.2.2. Expression and Purification of Tc PAM 25 2.2.3. Expression and Purification of Pa PAM. 27 2.2.4. PAM Activity Assays 27 2.2.5. GC/EIMS Analysis of PAM Catalyzed Products 2 8 2.2.6. Mutagenesis of Tc PAM cDNA for Expression of L104A Mutant. 29 2.2.7. Analysis of Kinetic Parameters of Tc PAM and PAMeLA_104 30 2.2.8. Analysis of Products Formed from the Incubat ion of Tc PAM with [15 N]Phenylalanine and [ring, -C- 2H6]- trans -Cinnamate Acid 30 2.2.9. Assessing the Product Distribution of Pa PAM and Tc PAM with 2'-Methyl-( S)- -Phenylalanine 31 2.2.10. Conversion of Cinnamate to - and -Phenylalanine by Pa PAM 31 2.3. Results 32 2.3.1. X-ray Crystal Structure of Tc PAM 32 2.3.2. Site-Directed Mutation of an Active Site Leu of Tc PAM 36 2.3.3. Inter/Intramolecularity Analysis of the Tc PAM Reaction 39 2.3.4. X-ray Crystal Structure of Pa PAM 40 2.3.5. Distribution of Products in the Catalysis of 2 -Methyl-( S)- -Phenylalanine by Tc PAM and Pa PAM 42 x 2.3.6. Formation of - and -Phenylalanines from Cinnamate and NH 3 by Pa PAM 43 2.4. Discussion 43 2.4.1. Stereoselectivity of the Aminomutase Reactio n 52 2.4.2. Effects of L104A Point Mutation on Enzymatic Activity of Tc PAM 54 2.5. Conclusion 57 3. (S)- -STYRYLALANINE USED TO PROBE THE INTERMOLECULAR MECHANISM OF AN INTRAMOLECULAR MIO-AMINOMUTASE 59 3.1. Introduction 59 3.2. Experimental 63 3.2.1. Chemicals 63 3.2.2. Instrumentation 63 3.2.3. Expression of the tcpam and Purification of Tc PAM 64 3.2.4. Identification of an Amine Donor Substrate 64 3.2.5. Assessing the Optimal Concentration of Amino Group Donor 6 65 3.2.6. Calculation of the Inhibition Constants of V arious Acrylates in the Tc PAM Reaction 65 3.2.7. Time Course Assays for Intermolecular Amino Group Transfer 67 3.2.8. Assessing the Absolute Stereochemistry of - and -Phenylalanine Product (8f) by the Tc PAM Transaminase Pathway 67 3.2.9. Assessing the Effects of Maintaining the Ste ady-State Conversion of 6 to 7 on the Production of 8f 68 3.2.10. Biosynthesis and Characterization of a 3,4- Dihydronaphthalene-2-carboxylic Acid (16-Acr) from 16 69 3.3. Results 70 3.3.1. Overexpression of Tc PAM 70 3.3.2. Calculation of KI for Various Acrylates and Time Course Studies for Optimal Amino Transfer 70 3.3.3. Relationship between the Rates of Formation of 7 and 8 and the KI values of 9Œ 14 73 3.3.4. Titration of the Amino Group Donor 6 to Main tain Steady-State Conversion to 7 78 3.3.5. Other Amino Donor Substrates 79 3.4. Discussion 82 3.5. Conclusion 85 4. ASSESSING THE DEAMINATION RATE OF NH 2-MIO ADDUCT BY BURST PHASE ANALYSIS 87 4.1. Introduction 87 4.2. Experimental 91 4.2.1. Chemicals 91 4.2.2. Enzyme Preparation 92 4.2.3. Quantification of Biosynthetic Styrylacrylat e during Kinetic Progressions 92 4.3. Results 93 4.4. Discussion 100 4.5. Conclusion 103 xi 5. A BACTERIAL TYROSINE AMINOMUTASE PROCEEDS THROUG H RETENTION OR INVERSION OF STEREOCHEMISTRY TO CATALYZE ITS ISOMERIZATION REACTION 105 5.1. Introduction 105 5.2. Experimental 110 5.2.1. Chemicals and Reagents 110 5.2.2. Instrumentation 110 5.2.3. Subcloning, Expression, and Purification of Cc TAM 111 5.2.4. Assessing the Activity and Stereochemistry o f the Cc TAM Reaction 112 5.2.5. Synthesis of Authentic - (1) and -Tyrosine (2) and ( E)-4-Hydroxycinnamic Acid Derivatives 113 5.2.5.1. 4'- O-Ethoxycarbonyl-( E)-coumaric Acid Methyl Ester (i.e., 4'- O-Ethylcarboxy- (E)-4'-Hydroxycinnamic Acid Methyl Ester) 113 5.2.5.2. The 4'- O,2- N-Di(ethoxycarbonyl)- -tyrosine Methyl Ester. 114 5.2.5.3. 4'- O,3- N-Di(ethoxycarbonyl)- -tyrosine Ethyl Ester 115 5.2.6. Synthesis of [Rh(NBD) 2]ClO 4 Complex. 116 5.2.7. Synthesis of [Rh(NBD)(( R)-Prophos)]ClO 4 Complex. 116 5.2.8. Synthesis of [ 2H]-Labeled (2 S)-1 Isotopomers 117 5.2.8.1. Synthesis of ( Z)-2-Benzamido-3-(4'-hydroxyphenyl)acrylic Acid 117 5.2.8.2. Synthesis of ( Z)-2-Benzamido-[3- 2H]-3-(4'-hydroxyphenyl)acrylic Acid 118 5.2.8.3. Synthesis of (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)-[3- 2H]-1 119 5.2.9. Characterization of (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)-[3- 2H]-1 120 5.2.10. Synthesis of Authentic 4'- O,3- N- Di(( S)-2-methylbutanoyl) Methyl Esters of (R)- and ( S)-2 121 5.2.11. Assessing the Stereospecificity of the C -Hydrogen Abstraction Catalyzed by Cc TAM 122 5.2.12. Assessing the Stereospecificity of the Hydr ogen Rebound at C Catalyzed by Cc TAM 122 5.2.13. Assessing the Intramolecular Proton Transfe r Step of the Cc TAM Reaction 123 5.2.14. Assessing the Effect of pH on Cc TAM Stereoselectivity 124 5.2.15. Synthesis of ( R)-2 Methyl Ester 124 5.3. Results 125 5.3.1. Cc TAM Activity and Stereochemistry 125 5.3.2. Assignment of the Prochiral Hydrogens of ( R)-2 by 1H-NMR 125 5.3.3. Using NMR to Assess the Mechanism of the Hyd rogen Transfer at C in the Cc TAM Reaction 128 5.3.4. Assessing the Mode of the Amino Group Attach ment at C by Cc TAM. 129 5.3.4.1. Synthesis of (2 S,3 S)-[2,3- 2H2]- and (2 S, 3R)-[3- 2H]-( S)-1 129 5.3.5. Analysis of 2 Made by Cc TAM Catalysis from (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)- [3- 2H]-1 129 5.3.6. Assessing the D H Exchange Rate during Cc TAM Catalysis 132 xii 5.3.7. Re-evaluation of the Stereoisomeric Product Distribution Catalyzed by Cc TAM 133 5.3.8. pH Effect on the Stereoselectivity of the Re action Catalyzed by Cc TAM 134 5.4. Discussion 135 5.4.1. Retention of Configuration at the Migration Termini during the Cc TAM Reaction. 135 5.4.1.1. Amino Group Migration 135 5.4.1.2. Hydrogen Migration Stereochemistry 136 5.4.2. Hydrogen Exchange during Migration 139 5.4.3. Diastereomeric Product Ratio Catalyzed by Cc TAM 141 5.5. Conclusion 146 APPENDIX 147 REFERENCES 201 xiii LIST OF TABLES Table 2.1 Kinetic Parameters for Tc PAM and PAMeLA_104 with Various Substrates 37 Table 2.2 GC/EIMS Analysis: Diagnostic Ions of Bios ynthetic [ 15 N]- -Phenylalanine a 39 Table 3.1 Kinetic Parameters of Various Arylacrylat es and Their Conversion to Aminoo Acids Using ( S)-Styryl- -alanine in the Tc PAM Reaction 72 Table 3.2 Steady-State Rate of Formation of 7 from 6 and of 8 from an Acrylate-+NH 2-MIO Complex in the Intermolecular Tc PAM Reaction. a 74 Table 3.3 Relative Steady-State Rates of Transfer o f an Amino Group from Non-Natural Amino Acids (6 and 15Œ17) to trans-3 -Methylcinnamate (14) by TcPAM Catalysis 80 Table 5.1 EI-MS Fragmentation of 4'- O,3- N- Di((ethoxycarbonyl)) Ethyl Ester Derivatives of Labeled and [ 2H]-Labeled Biosynthesi zed Isotopomers of 2 130 xiv LIST OF FIGURES Figure 1.1 -amino acids, found in some of the important bioact ive compounds, are highlighted in red 1 Figure 1.2 a) 2,3-isomerization of the analogs of ( S)- -phenylalanine to corresponding analogs of ( R)- -phenylalanine by PAM. 38, 39, 42 b) Ammonia addition to substituted trans -cinnamate analogs to form corresponding analogs of ( S)- -phenylalanine and ( R)- -phenylalanine by PAM. 40 c) Ammonia addition to substituted trans -cinnamate analogs to form corresponding ( S)- -phenylalanine analogs by PAL. 44 5 Figure 1.3 Transaminase reaction of Tc PAM. 7 Figure 1.4 Sequence homology model of Tc PAM, Pa PAM, Sg TAM, Cc TAM, PAL from Petroselinum crispum ( Pc PAL), TAL from Rhodobacter sphaeroides (Rs TAL), and HAL from Pseudomonas putida ( Pp HAL), showing conserved residues in shaded areas. 8 Figure 1.5 The crystal structure of Sg TAM with a bound mechanism based inhibitor, a) 2,2-difluoro-(3 R)- -tyrosine (PDB 2QVE) and b) (3 R)-amino-2,2- difluoro-3-(4'-methoxyphenyl)propanoic acid (PDB 2R JS) 13 Figure 1.6 a) Crystal structure of Tc PAM with bound cinnamic acid intermediate. b) The proposed orientation of the cinnamic acid inter mediate which leads to the formation of ( R)- -phenylalanine with a retention of configuration at both migration termini, by Wu. et al. (Modeled usin g Pymol from PDB# 3NZ4) 19 Figure 2.1 The Tc PAM active site-cinnamate complex (magenta) a) indi cating the displacement of plane of the aromatic ring of the c innamate relative to the -bond plane of the propeonate carboncarbon double bond, and (b) the topside view. The PAM active site residues are colo red by atom as follows: C (green), O (red), and N (blue). 33 Figure 2.2 (b) the 3 -methyl-( S)- -phenylalanine substrate is modeled into the mutant PAMeLA_104 active site where the steric volume is i ncreased through the mutation of Leu-104 to Ala; now, the closest distan ce between the bound substrate and residue 104 is estimated at ~ 4.4 Å, a nd The Tc PAM- cinnamate complex is used to approximate the trajec tory of non-natural substrates. (a) The 3 -methylphenyl- -alanine substrate is modeled into the active site of Tc PAM showing the distance (~2 Å) between the 3 -methyl group of the substrate and Leu-104. 34 xv Figure 2.3 a) Electron density (2F oF c map, blue mesh) calculated at 1.0 around the - and -phenylpropanoid that is covalently bound to the MI O found in monomer fiBfl (C and C are indicated). Atoms are color-coded as C (green), O (red), N (blue). b) Active site of Pa PAM in complex with a MIO-bound ( S)- -phenylalanine-type ligand (orange carbon atoms). Active site residues contributed by three monomers are colored accordingly (C: cyan, yellow, or green for each mon omer; oxygen: red; nitrogen: blue). 23 41 Figure 2.4 CLUSTAL 2.1 multiple sequence alignment 38 of representative class I lyase-like PAL, PAM and TAM enzymes whose structure s are known and five sequences highly homologous to Pa PAM that were found in the GenBank database by BLAST search: Pa PAM, PAM from Pantoea agglomerans ; V.bact , possible PAM from Vibrionales bacterium SWAT-3 ; S.marit , PAL/PAM EncP from Streptomyces maritimus ; B.subtl , possible PAM from Bacillus Subtlis ; K.pneu , possible PAM from Klebsiella pneumoniae 342 ; B.rhiz , possible PAM from Burkholderia rhizoxinica ; Sg TAM, TAM from Streptomyces globisporus ; Av PAL Phenylalanine ammonia lyase 39 from Anabaena variabilis ; Tc PAM, PAM from Taxus Canadensis . Enzymes that have a concerved phenylalanine resid ue inside the active which correspond to F455 residue in Pa PAM, produce ( S)- -phenylalanine (designated with asterisk (*)). 48 Figure 2.6 Cinnamate diastereoisomers modeled into Tc PAM. a) A space-filled rendering and b) a skeletal structure of cis -cinnamate are modeled into Tc PAM by preserving the tight salt bridge with Arg325 . The phenyl ring of the cinnamate is oriented toward the interior of the active site pocket. The collision distances between active site residue s and the phenyl ring of the cisoid diastereoisomer are given: Asn231 - 0.8 Å, Leu227 - 2.7 Å, Phe371 - 2.7 Å, Asn355 (main chain) - 2.7 Å. c) A s pace-filled rendering of trans -cinnamate is modeled into Tc PAM by orienting the carboxylate toward the salt bridge with Arg325; occlusions by a ctive site residues on the transoid structure are absent. d) Shown is an o verlay of the space-filled structures of ( cis )- and trans -cinnamates modeled in the Tc PAM active site with the Arg325 salt bridge preserved. 13 50 Figure 3.1 Ratio [(v o8)/(v o7)] of the steady-state rates for the conversion of acrylates 9Œ14 to 8 and of the formation of the NH 2ŒMIO complex plotted vs KI(A) for 9Œ14 (shown in parentheses). 75 Figure 3.2 (a) Time course assay. Amounts of ( S)-styryl- -alanine ( , 6), (2 E,4 E)- styrylacrylate ( , 7), ( S)-3 -methyl- - and ( R)-3 -methyl- -phenylalanine (, 8f- and 8f- ), and trans -3 -methylcinnamate ( , 14) in an aminotransferase reaction catalyzed by Tc PAM over 12 h. (b) Steady-state xvi conversion of 6 to 7 of 8f/ and 8f/ by transfer of an amino group from 6 to 14 (5 mg) by Tc PAM catalysis: (2 E,4 E)-styrylacrylate (7, ), ( S)-styryl- -alanine (6, ), trans -3 -methylcinnamate (14, ), and total of - and -isomers of 3 -methylphenylalanine (8f, ). 77 Figure 3.3 Modeled in the Tc PAM active site are natural substrates (a) ( S)- -phenylalanine and (b) ( R)- -phenylalanine for reference, (c) ( S)-styryl- -alanine, (d) ( S)-2 -furyl- -alanine, (e) (3 R)-3-aminotetralin-(2 R)-2- carboxylate, and (f) ( S)-2-aminotetralin-2-carboxylate. PyMOL (Schrödinger LLC, Cambridge, MA) was used for the s ubstrate modeling by preserving the key interactions with the active site residues. 83 Figure 4.1 Evaluation of the kinetic model (Scheme 4.3, shaded inset) for Tc PAM burst kinetics (Eqn 4.1) was used to globally fit e xperimental progress curves (Kaleidagraph 4.0) spanning six different ( S)-styryl- -alanine concentrations incubated with Tc PAM (5.5 M). Release of (2 E,4 E)- styrylacrylate was measured in a stopped-flow cell by A340 monitoring. Each time point is an average of three progression curves. 94 Figure 4.2 Hanes-Woolf analysis of the steady-state rates. The progress curves (Figure 4.1) were individually fit to the burst equ ation (Eqn 4.1) to evaluate the steady-state velocities ( A), for each concentration of styryl- -alanine, of the Tc PAM burst kinetics. The average value for each data point A was used. The dependence [S] o/A on [S] o with S.D. for the triplicate measurements of A. Linear regression fit ([S] o/A = 554.7 + 5.309[S] o; R 2 = 0.9913) to the data (solid line). 96 Figure 4.3 The progression curves (Figure 4.1) were individually fit to the burst equation (Eqn 4.1) to evaluate the steady-state vel ocity A and the burst amplitude B for the Tc PAM burst kinetics. The average value for each data point A was used. A2 dependence on B with S.D. for the triplicate measurements. Linear regression fit ( A2 = 0.001037 + 0.009153 B; R 2 = 0.9914) to the data (solid line). 97 Figure 4.4 The progress curves (Figure 4.1) were in dividually fit to the burst equation (Eqn 4.1) to evaluate the steady-state velocity A and the burst amplitude B for the Tc PAM burst kinetics. The average value for each data point A was used. The dependence of A/B on 1/[S] o with S.D. for the triplicate measurements. Linear regression fit ( A/B = 0.049558 + 6.88821/[S] o; R 2 = 0.97933) to the data (solid line). 99 Figure 5.1 Partial 1H-NMR profile of unlabeled 2 and the 3J coupling constants for the ABX spin system of ( R)-2. 126 xvii Figure 5.2 a) 2H-NMR (after solvent exchange into CH 3OH); the relative area of the peaks at 4.41 and 2.61 are shown and b) 1H-NMR (after solvent exchange into CD 3OD) of a mixture containing the remaining substrate (2 S)-[3,3- 2H2]-1 and the biosynthetic (2 S,3 R)-[2,3- 2H2]-2 after a Cc TAM- catalyzed reaction. c) 1H-NMR (in CD 3OD) of authentic ( R)-2. The signals for the prochiral protons of authentic ( R)-2 are aligned (boxes) with signals for the deuterium labeled product in t he biosynthetic sample. 128 Figure 5.3 Plotted are the D ®H exchange (×) and (2 S)-[3,3- 2H2]-1 ( ), [2,3- 2H2]-2 (), and [3- 2H]-4'-hydroxycinnamic acid ( ) (as mol %) during the Cc TAM conversion of (2 S)-[3,3- 2H2]-1 to labeled 2. 133 Figure 5.4 Intramolecular salt-bridge between the a mmonium ion and carboxylate group of ( R)-2 in methanol. Dihedral angles ( 1 and 2) between H A and HX and between H B and H X, respectively, in the pseudo six-membered ring formed by 2 are shown in Newman Projection. 13 6 Figure 5.5 Analysis of the diastereomeric mixture o f products catalyzed by Cc TAM. Plotted are mol % of ( R)-2 ( ) and ( S)-2 ( ) relative to amount of ( S)-1 added. The amount of ( R)-2 (as %) ( ) relative to the total amount of ( R)- and ( S)-2 made at steady state. (Average of duplicate ass ays is plotted). 138 Figure 5.6 ( S)-2 (as % of total 2) measured after substrate ( S)-1 was depleted by 13%, 35%, and 56% at pHs 7, 8, and 9 while incubated wit h Cc TAM. 139 Figure 5.7 Comparison of the Tc PAM, Pa PAM, and Sg TAM active site structures co- crystallized with phenylpropanoid adducts or comple xes and the Cc TAM active site with 4'-hydroxycinnamate was modeled on the Tc PAM crystal structure (PDB# 3NZ4). The orientation of phenylpro panoid (center of each diagram) relative to the Arg residue (at right of each diagram) is shown. Also shown are key non-catalytic residues in volved in binding and positioning the substrate; the catalytic tyrosine r esidue is above the plane in each drawing and is not shown. 142 Figure A 1 The linear relationship between the reci procal of the steady-state rate (1/ vo) and the concentrations of competitive inhibitors ([I] o in eq. 4 above): Top Left) trans -4'-Chloro- (9), Top Right) trans -4'-Methyl- (10), Middle Left) trans -4'-Fluorocinnamate (11), Middle Right) trans -Cinnamate (12). 148 Figure A 2 Double Reciprocal Plots of the Rate of C onversion of 8b/ to 8b/ and Concentration of Substrate ( S)-4'-Methyl- -phenylalanine (8b/ ) in xviii Tc PAM Reactions Containing (2 E, 4E)-Styrylacrylate (7) at 0 ( ), 50 ( ), 100 ( ), 200 ( ) M to Assess the KI of 7. 150 Figure A 3 a) 1H-NMR spectrum ((500 MHz, CDCl 3) d: 7.68 (s, 1 H), 7.32 - 7.17 (m, 4 H), 2.91 (t, J = 8.4 Hz, 2 H), 2.64 (dt, J = 1.5, 8.5 Hz, 2 H))of and the b) 13 C-NMR spectrum ((126 MHz, CDCl 3) d: 172.2, 138.7, 137.2, 132.3, 129.9, 128.8, 128.4, 127.7, 126.8, 27.5, 21.9) of 1 6-Acr 151 Figure A 4 Overlay of gas chromatography profiles o f; a) N-[(1' S)-camphanoyl] methyl esters of (2 S)- -4'-fluorophenylalanine (13.72 min) derived from intermolecular amino group transfer 152 Figure A 5 GC EI/MS fragments of a) N-[(1' S)-camphanoyl] methyl ester of (2 S)- -4'- fluorophenylalanine and b) N-[(1' S)-camphanoyl] methyl ester of (3 R)- -4'-fluorophenylalanine 154 Figure A 6 EI-MS fragmentation of 4'- O-(ethoxycarbonyl) methyl ester derivative of authentic 4'-hydroxycinnamic acid. Diagnostic fragm ent ions are m/z 178 and 147. 155 Figure A 7 a) 1H NMR of 4'- O-(ethoxycarbonyl) methyl ester of 4'-hydroxycinnami c acid 1H NMR (500 MHz, CDCl 3) d: 7.68 (d, J = 16.1 Hz, 1 H), 7.55 (d, J = 8.3 Hz, 2 H), 7.22 (d, J = 8.3 Hz, 1 H), 6.41 (d, J = 16.1 Hz, 1 H), 6.41 (d, J = 16.1 Hz, 1 H), 4.34 (q, J = 7.3 Hz, 2 H), 3.82 (s, 3 H), 1.41 (t, J = 7.3 Hz, 3 H) 156 Figure A 8 a) 13 C NMR of 4'- O-(ethoxycarbonyl) methyl ester of 4'-hydroxycinnami c acid 13 C NMR (126 MHz, CDCl 3) d: 167.2, 153.2, 152.4, 143.6, 132.2, 129.2, 121.6, 118.1, 65.1, 51.7, 14.2 157 Figure A 9 EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) methyl ester derivatives of unlabeled ( S)-1. Diagnostic fragment ions are m/z 250, 178, and 107. 158 Figure A 10 a) 1H NMR of 4'- O,2- N-di(ethoxycarbonyl) methyl ester of -tyrosine. 1H NMR (500 MHz, CDCl3) = 7.21 - 6.98 (m, 4 H), 5.10 (d, J = 7.3 Hz, 1 H), 4.67 - 4.55 (m, 1 H), 4.28 (q, J = 7.1 Hz, 2 H), 4.08 (q, J = 6.7 Hz, 2 H), 3.69 (s, 3 H), 3.10 (dd, J = 6.1, 14.0 Hz, 1 H), 3.05 (dd, J = 6.1, 14.0 Hz, 1 H), 1.36 (t, J = 7.0 Hz, 3 H), 1.20 (t, J = 6.7 Hz, 3 H). 159 Figure A 11 a) 13 C NMR of 4'- O,2- N-di(ethoxycarbonyl) methyl ester of -tyrosine. 13 C NMR (126 MHz, CDCl3) = 171.9, 155.8, 153.5, 150.2, 133.6, 130.3, 121.0, 64.8, 61.2, 54.5, 52.3, 37.6, 14.4, 1 4.1 160 xix Figure A 12 1H NMR of 4'- O,2- N-di(ethoxycarbonyl) ethyl ester of -tyrosine. 1H NMR (500 MHz, CDCl3) = 7.17 (d, J = 8.6 Hz, 2 H), 7.14 - 7.11 (m, J = 8.6 Hz, 2 H), 5.14 (d, J = 7.8 Hz, 1 H), 4.63 (dd, J = 5.6, 13.4 Hz, 1 H), 4.33 (q, J = 7.1 Hz, 2 H), 4.17 (q, J = 7.2 Hz, 2 H), 4.12 (q, J = 7.1 Hz, 2 H), 3.11 (d, J = 5.1 Hz, 2 H), 1.40 (t, J = 7.1 Hz, 3 H), 1.25 (t, J = 7.3 Hz, 3 H), 1.24 (t, J = 7.2 Hz, 3 H) 161 Figure A 13 13 C NMR of 4'- O,2- N-di(ethoxycarbonyl) ethyl ester of -tyrosine. 13 C NMR (126 MHz, CHCl 3) d: 171.7, 156.1, 153.8, 150.4, 133.9, 130.6, 121.3, 65.1, 61.8, 61.4, 54.9, 38.0, 14.7, 14.4, 14 .3 162 Figure A 14 a) GC profile of derivatized -tyrosine (with ethylchloroformate and CH 2N2 163 Figure A 15 EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester derivatives of authentic ( R)-2. Diagnostic fragment ions are m/z 280, 266, and 194. 165 Figure A 16 a) 1H NMR of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester of -tyrosine. 1H NMR (500 MHz, CDCl 3) d: 7.34 (d, J = 8.5 Hz, 2 H), 7.16 (d, J = 8.8 Hz, 2 H), 4.33 (q, J = 7.3 Hz, 2 H), 4.12 (q, J = 7.1 Hz, 2 H), 4.09 (q, J = 7.3 Hz, 2 H), 2.92 - 2.78 (m, 2 H), 1.40 (t, J = 7.3 Hz, 3 H), 1.24 (t, J = 7.2 Hz, 3 H), 1.19 (t, J = 7.3 Hz, 3 H) 166 Figure A 17 a) 13 C NMR of 4'- O, 3- N-di(ethoxycarbonyl) ethyl ester of -tyrosine. 13 C NMR (126 MHz, CDCl 3) d: 170.7, 155.8, 153.5, 150.3, 138.8, 127.4, 121.3, 64.8, 61.0, 60.7, 51.0, 40.6, 14.5, 14.1, 14 .0 167 Figure A 18 1H NMR of 4'- O,3- N-di(ethoxycarbonyl) methyl ester of -tyrosine. 1H NMR (500 MHz, CDCl 3) = 7.29 (d, J = 8.5 Hz, 2 H), 7.12 (d, J = 8.5 Hz, 1 H), 5.66 (br. s., 1 H), 5.17 - 5.07 (m, 1 H), 4.28 (q, J = 6.9 Hz, 2 H), 4.08 (q, J = 7.3 Hz, 2 H), 3.60 (s, 3 H), 2.90 - 2.73 (m, 2 H ), 1.36 (t, J = 7.3 Hz, 3 H), 1.20 (t, J = 7.0 Hz, 3 H) 168 Figure A 19 13 C NMR of 4'- O,3- N-di(ethoxycarbonyl) methyl ester of -tyrosine. 13 C NMR (126 MHz, CD 3COCD 3) d: 170.7, 155.9, 153.4, 150.5, 140.3, 128.4, 121.7, 65.0, 64.4, 60.0, 51.0, 40.2, 14.1, 1 3.6 169 Figure A 20 Crystal structure of a) ( Z)-2-benzamido-[3- 2H]-3-(4'-hydroxyphenyl) acrylic acid and Color coding: Carbon (black), Hydr ogen (Cyan), Deuterium (green) Oxygen (red), and Nitrogen (blue) , 170 xx Figure A 21 1H NMR spectrum of ( Z)-2-Benzamido-[3- 2H]-3-(4'- hydroxyphenyl)acrylic Acid. 1H NMR (500 MHz, DMSO-d 6) d: 12.49 (br. s., 1 H), 9.91 (s, 1 H), 9.77 (s, 1 H), 7.99 ( d, J = 7.1 Hz, 2 H), 7.59 (t, J = 7.6 Hz, 1 H), 7.56 - 7.49 (m, 4 H), 6.77 (d, J = 8.3 Hz, 2 H) 172 Figure A 22 13 C NMR spectrum of ( Z)-2-Benzamido-[3- 2H]-3-(4'- hydroxyphenyl)acrylic Acid. 13 C NMR (126 MHz, DMSO-d 6) d: 166.6, 165.9, 158.8, 133.7, 133.4, 131.9, 131.8, 128.5, 12 8.3, 127.7, 127.5, 124.6, 123.9, 115.5, 115.4 173 Figure A 23 1H NMR spectrum of ( Z)-2-Benzamido-3-(4'-hydroxyphenyl)acrylic Acid. 1H NMR (500 MHz, DMSO-d 6) d: 12.49 (br. s., 1 H), 9.91 (br. s., 1 H), 9.82 - 9.72 (m, 1 H), 7.99 (d, J = 7.3 Hz, 2 H), 7.59 (t, J = 7.1 Hz, 1 H), 7.56 - 7.47 (m, 4 H), 7.41 (s, 1 H), 6.77 (d, J = 8.3 Hz, 2 H) 174 Figure A 24 13 C NMR spectrum of ( Z)-2-Benzamido-3-(4'-hydroxyphenyl)acrylic Acid. 13 C NMR (126 MHz, DMSO-d 6) d: 166.6, 165.9, 158.8, 134.0, 133.8, 131.9, 131.8, 128.5, 128.4, 127.7, 127.5, 12 4.6, 124.0, 115.5, 115.4 175 Figure A 25 1H NMR spectrum of [Rh(NBD) 2]ClO 4. 1H NMR (500 MHz, CDCl 3) d: 5.20 (q, J = 2.2 Hz, 8 H), 4.15 - 4.11 (m, 4 H), 1.51 (t, J = 1.6 Hz, 4 H) 176 Figure A 26 1H NMR spectrum of [Rh(NBD)( R)-Prophos]ClO 4Ł0.5 CH 2Cl 2. Trace amounts of THF, H 2O, and silicon grease are present as impurities. 1H NMR (500 MHz, CDCl 3) = 7.79 - 7.73 (m, 2 H), 7.72 - 7.67 (m, 3 H), 7.65 - 7.61 (m, 2 H), 7.61 - 7.58 (m, 6 H), 7.58 - 7.55 (m, 6 H), 7.47 - 7.41 (m, 2 H), 7.35 - 7.29 (m, 2 H), 5.42 (br. s., 2 H), 5.31 (s, 1 H), 4.87 (br. s., 1 H), 4.28 (br. s., 1 H), 4.16 (br. s., 1 H), 2.71 - 2.59 (m, 2 H), 2.04 (td, J = 7.4, 12.5 Hz, 1 H), 1.84 - 1.76 (m, 2 H), 1.21 (dd, J = 6.5, 12.5 Hz, 3 H) 177 Figure A 27 13 C NMR spectrum of [Rh(NBD)( R)-Prophos]ClO 4Ł0.5 CH 2Cl 2. Trace amounts of THF, is present as impurities. 13 C NMR (126 MHz, DMSO- d6) = 143.1, 135.1, 135.1, 134.8, 134.5, 134.1, 133.9, 132.8, 132.7, 132.0, 131.9, 131.7, 131.1, 131.0, 130.9, 130.1, 12 9.3, 129.3, 128.9, 128.8, 128.8, 128.6, 128.5, 128.2, 127.8, 125.9, 125.5, 6 3.4, 48.2, 34.2, 33.0, 14.7 178 Figure A 28 EI-MS fragmentation of 4'- O,2- N-di(ethoxycarbonyl) methyl ester derivatives of (2 S,3 S)-[2,3- 2H2]-1. Diagnostic fragment ions are m/z 252, 180, and 108. 180 xxi Figure A 29 EI-MS fragmentation of 4'- O,2- N-di(ethoxycarbonyl) methyl ester derivatives of (2 S,3 R)-[3- 2H]-1. Diagnostic fragment ions are m/z 251, 179, and 108. 181 Figure A 30 EI-MS fragmentation of 4'- O,2- N-di(ethoxycarbonyl) methyl ester derivatives of (2 S)-[3,3- 2H2]-1. Diagnostic fragment ions are m/z 251, 179, and 109. 182 Figure A 31 Partial NMR spectra of isotopomers of ( S)-1. 183 Figure A 32 1H NMR spectrum of -tyrosine. 1H NMR (500 MHz, D 2O) d: 7.22 (d, J = 8.5 Hz, 2 H), 6.92 (d, J = 8.5 Hz, 2 H), 3.96 (dd, J = 5.1, 7.8 Hz, 1 H), 3.23 (dd, J = 5.0, 14.8 Hz, 1 H), 3.08 (dd, J = 7.8, 14.6 Hz, 1 H) 185 Figure A 33 a) 1H NMR spectrum of (2 S,3 S)-[2,3- 2H2]- -tyrosine. 1H NMR (500 MHz, D 2O) d: 7.22 (d, J = 8.5 Hz, 2 H), 6.92 (d, J = 8.5 Hz, 2 H), 3.08 (s, 1 H). b) 2H NMR spectrum of (2 S,3 S)-[2,3- 2H2]- -tyrosine. 2H NMR (77 MHz, H 2O) d: 3.97 (bs, 1 2H) 3.22 (bs, 1 2H). 186 Figure A 34 a) 1H NMR spectrum of (2 S,3 R)-[3- 2H]- -tyrosine. 1H NMR (500 MHz, D2O) d: 7.22 (d, J = 8.5 Hz, 2 H), 6.92 (d, J = 8.5 Hz, 2 H), 3.97 (d, J = 5.0 Hz, 1 H), 3.22 (d, J = 5.0 Hz, 1 H) b) 2H NMR spectrum of (2 S,3 R)- [3- 2H]- -tyrosine. 2H NMR (77 MHz, H 2O) d: 3.08 (bs, 1 2H). 187 Figure A 35 1H NMR spectrum of -tyrosine. 1H NMR (500 MHz, CD 3OD) d: 7.25 (d, J = 8.8 Hz, 2 H), 6.81 (d, J = 8.8 Hz, 2 H), 4.41 (dd, J = 4.2, 10.0 Hz, 1 H), 2.73 (dd, J = 10.0, 16.6 Hz, 1 H), 2.61 (dd, J = 4.2, 16.6 Hz, 1 H) 188 Figure A 36 a) 1H NMR and b) 2H NMR spectra of the reaction mixture of biosynthesized products catalyzed by Cc TAM from (2 S)-[3,3- 2H2]- -tyrosine. 189 Figure A 37 1H NMR spectrum of -tyrosine methyl ester. 1H NMR (500 MHz, CD 3OD) d: 7.29 (td, J = 2.2, 8.5 Hz, 2 H), 6.85 (td, J = 2.2, 8.5 Hz, 2 H), 4.63 (dd, J = 6.3, 7.8 Hz, 1 H), 3.09 (dd, J = 7.8, 16.6 Hz, 1 H), 2.98 (dd, J = 6.3, 16.7 Hz, 1 H) 190 Figure A 38 GC trace of 4'- O,3- N-di(( S)-2-methylbutanoyl) methyl ester derivatives of authentic ( S)-2 (18.68 min) and ( R)-2 (19.06 min) (a). EI-MS fragmentation of 4'- O,3- N-di(( S)-2-methylbutanoyl) methyl ester derivatives of authentic ( R)-2. Diagnostic fragment ions are m/z 278 and 194. 191 xxii Figure A 39 GC trace of 4'- O,3- N-di(( S)-2-methylbutanoyl) methyl ester derivatives of (S)-2 (18.68 min) and ( R)-2 (19.06 min) (a) and EI-MS fragmentation of 4'- O,3- N-di(( S)-2-methylbutanoyl) methyl ester derivatives of ( R)-2 biosynthesized by Cc TAM from unlabeled ( S)-1. 193 Figure A 40 EI-MS fragmentation of 4'- O-(ethoxycarbonyl) methyl ester derivative of 4'-hydroxycinnamic acid biosynthesized by Cc TAM from unlabeled ( S)-1. 194 Figure A 41 EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester derivatives of 2 biosynthesized by Cc TAM from unlabeled ( S)-1. Diagnostic fragment ions F1, F2, and F3 (m/z 280, 2 66, and 194) are highlighted. 196 Figure A 42 EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester derivatives of isotopomers of 2 biosynthesized by Cc TAM from (2 S,3 S)- [2,3- 2H2]-1. Diagnostic fragment ions F1, F2, and F3 (m/z 2 82, 266, and 194) are highlighted. 197 Figure A 43 EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester derivatives of isotopomers of 2 biosynthesized by Cc TAM from (2 S,3 R)- [3- 2H]-1. Diagnostic fragment ions F1, F2, and F3 (m/z 281, 267, and 195) are highlighted. 198 Figure A 44 EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester derivatives of isotopomers of 2 biosynthesized by Cc TAM from (2 S)-[3,3- 2H2]-1. Diagnostic fragment ions F1, F2, and F3 (m/z 2 82, 267, and 195) are highlighted. 199 Figure A 45 EI-MS fragmentation of 4'- O-(ethoxycarbonyl) methyl ester derivative of biosynthesized [3- 2H]-4'-hydroxycinnamic acid biosynthesized by Cc TAM from (2 S)-[3,3- 2H2]-1. Diagnostic fragment ions are m/z 179 and 148. 200 xxiii KEY TO ABBREVIATIONS 4'-HOCinn 4'-hydroxycinnamic acid AA1 Hypothetical amino acid that serves as an amino group donor in the double- displacement reaction catalyzed by Tc PAM AA2 Hypothetical amino acid derived from the aminat ion of an exogenously supplied arylacrylate in the double-displacement reaction ca talyzed by Tc PAM AC1 Hypothetical arylacrylate derived from an amino acid after elimination of ammonia in the double-displacement reaction catalyz ed by Tc PAM AC2 Hypothetical exogenously supplied arylacrylate in the double-displacement reaction catalyzed by Tc PAM Acr Any Arylacrylic acid AMs Aminomutases Av PAL Phenylalanine ammonia lyase from Anabaena variabilis B.rhiz PAM from Burkholderia rhizoxinica B.subtl PAM from Bacillus subtlis Cc TAM Tyrosine aminomutase from Chondromyces crocatus (R)- -Dopa 3,4-dihydroxy- -phenylalanine Eqn Equation E. coli Escherichia coli GC-EIMS Gas Chromatography Electron Impact Mass Spe ctrometer HALs Histidine ammonia lyases K.pneu PAM from Klebsiella pneumoniae 342 MIO 3,5-dihydro-5-methylidene-4 H-imidazol-4-one MWCO Molecular weight cutoff filters NBD Norbornadiene (Bicyclo[2.2.1]hepta-2,5-diene) Ni-NTA Nickel nitrilotriacetic acid xxiv PALs Phenylalanine ammonia lyases PAM Phenylalanine aminomutase PAMeLA104 Phenylalanine aminomutase exchange Leu Ala104 Pa PAM Phenylalanine aminomutase from Pantoea agglomerans Pc PAL Phenylalanine ammonia lyases from Petroselinum crispum PDB Protein Data Bank Pp HAL Histidine ammonia lyases from Pseudomonas putida Rg PAL Phenylalanine ammonia lyases from Rhodotorula glutinis Rs TAL Tyrosine ammonia lyases from Rhodobacter sphaeroides SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis Sg TAM Tyrosine aminomutase from Streptomyces globisporus S.marit PAL/PAM EncP from Streptomyces maritimus TALs Tyrosine ammonia lyases TAM Tyrosine aminomutase Tc PAM Phenylalanine aminomutase from Taxus canadensis TMSCH 2N2 (Trymethylsilyl) diazomethane V. bact PAM from Vibrionales bacterium SWAT-3 xxv LIST OF SCHEMES Scheme 1.1 Synthesis of -Amino Acids. 14 (a) Michael addition of ammonia to 3- methylbut-2-enoic acid derivatives. (b) Hydrolysis of 6,6-disubstituted dihydrouracils. (c) Three component Mannich reactio n of a ketone, NH 3, and a malonic acid derivative. (d) Ritter transform ation of 3- hydroxycarboxylates with nitriles in the presence o f conc. H 2SO 4. (e) Cycloaddition of chlorosulfonyl isocyanate with all enes to give an alkylidene -lactam. (f) Reaction of substituted cyclopropanes with chlorosulfonyl isocyanate to yield -lactams. (g) Cycloaddition of disubstituted alkenes and chlorosulfonyl isocyanate , reductive cleavage of the chlorosulfonyl group. (h) Indium-mediated react ion of enamines with methyl bromoacetate in the presence of acid. 14 3 Scheme 1.2 a) Formation of MIO, and b) The Structur e of the Chromophore of Green Fluoroscence Protein. 10 Scheme 1.3 Two Proposed Mechanisms for the Conversi on of Substrate to Product in a Generic Aminomutase: a) Amino-group Alkylation Pa thway and b) FriedelŒCrafts Aryl-Alkylation Pathway. 12 Scheme 1.4 Stereochemistry of the Migrating Amino G roup and the Proton of the Substrate and the Product during the Catalysis of a ) Tc PAM and (The Configurations at both C and C Positions are Retained) b) Pa PAM (The Configurations at both C and C Positions are Inverted) is Shown.* 15 Scheme 2.1 The Proposed Mechanism of the Aminomutas e Catalysis. Tyr is the Presumed Catalytic Base, Situated fiAbovefl the Inter mediate, and the NH 2-MIO Complex is Shown fiBelowfl the Intermediate. In Sg TAM and Pa PAM, After the Initial E 2-Type Elimination to Form the Intermediate, the Amino Group is Rebound at the C Position from the Si Face (Same Face), of the Intermediate while in Cc TAM and Tc PAM, the Intermediate Under Go 180º Rotation About the C 1C and C C ipso Bonds in order for the Amino Group to Attack at the C Position from the Re Face (Opposite Face) of the Intermediate. 53 Scheme 3.1 Mechanism of the Transaminase Reaction C atalyzed by Tc PAM with Its Natural Substrate 61 Scheme 4.1 Mechanism of MIO-Dependent Aminomutases 88 xxvi Scheme 4.2 a) Intermolecular Amino Group Transfer f rom [ 15 N]- -Phenylalanine to [2H6]-Cinnamate by Tc PAM Catalysis. b) Predominantly Ammonia Lyase Behavior of Tc PAM with ( S)-styryl- -alanine. 90 Scheme 4.3 Kinetic Model for Transaminase Reaction Catalyzed by Tc PAM a 91 Scheme 4.4 Kinetic Model for the Proposed Mechanism of the Tc PAM-Catalyzed Conversion of - to -Phenylalanine 102 Scheme 5.1 General MIO-dependent Aminomutase Mechan ism 107 Scheme 5.2 The Cc TAM Reaction on the Chondramides A-D Pathway 108 Scheme 5.3 Inversion and Retention of Configuration Pathways Catalyzed by Cc TAM a 141 1 1. INTRODUCTION 1.1. -Amino Acids as Important Precursors in Biologicall y Active Molecules Aryl- -Amino acids belong to an important class of compou nds that are present in medicinally and pharmaceutically important natural products (Figure 1.1). Figure 1.1. -amino acids, found in some of the important bioact ive compounds, are highlighted colored (red). (For interpretation of t he references to color in this and all other figures, the reader is referred to the electronic v ersion of this dissertation) Some examples are the antineoplastic agent Taxol, 1 from plant Taxus sp., the pigment, 2 Fe(III)-catecol complex (( R)- -dopa (3,4-dihydroxy- -phenylalanine)) from mushroom Cortinarius sp., 2 the antifungal agent Jasplakinolide from marine sp onge Jaspis sp., 3 antibiotic enediyne C-1027 from bacteria Streptomyces gloibisporus ., 4 the cytotoxic agents Chondramides from bacteria Chondromyces crocatus ., 5 and the antibiotic Andrimid from bacteria Pantoea agglomerans .6 Other amino acids have been used as building block s toward the synthesis of complex bioactive molecules, including -lactams, 7 and -peptides as mimics of -peptide hormones and as antimicrobial compounds. 8 -peptides demonstrate an increased metabolic stability, 9 a higher structural diversity, 10 and well-defined formation of secondary structure, 11 compared to their -analogues. 1.2. Synthesis of -Amino Acids An increasing demand for -amino acids has resulted in the development of var ious synthetic routes to synthesize -amino acids. 12-16 Such as Arndt-Eistert homologation, 17 Stereoselective synthesis of -amino acids starting from aspartic acid, 18 asparagine, and derivatives, 19 Curtius rearrangement, 20 through the intermediary preparation of perhydropy rimidin-4-ones, 21 conjugate addition to ,-unsaturated esters and imides, 22 hydrogenation, 23 reductive amination, 24 amino hydroxylation, 25 and through asymmetric -lactam synthesis. 26 The resolution of enantiomers from racemic mixtures is performed by synthesizing the diastereomers with chiral bases, such as (-)-ephedrine. 15 3 Scheme 1.1. Synthesis of -Amino Acids. 14 (a) Michael addition of ammonia to 3- methylbut-2-enoic acid derivatives. (b) Hydrolysis of 6,6-disubstituted dihydrouracils. (c) Three component Mannich reaction of a ketone, NH 3, and a malonic acid derivative. (d) Ritter transformation of 3-hydroxycarboxylates with nitriles in the presence of conc. H2SO 4. (e) Cycloaddition of chlorosulfonyl isocyanate wi th allenes to give an alkylidene -lactam. (f) Reaction of substituted cyclopropanes w ith chlorosulfonyl isocyanate to yield -lactams. (g) Cycloaddition of disubstituted alkenes and chlorosulfonyl isocyanate, reductive cleavage of the chlorosulfonyl group. (h) Indium-mediated reaction of enamines with methyl bromoacetate in the presence of acid. 14 However, these approaches have their inherent draw backs; involvement of multistep synthesis schemes ( Scheme 1.1 ), use of less greener chemicals, and production of by-products. 4 Due to the difficulty of resolving -amino acid enantiomers, kinetic resolution of race mic -amino acid derivatives or precursors (such as ester s, nitriles, amides, -lactams, and dihydrouracils) are performed using enzymatic resol ution, employing such enzymes as aspartases 27 and aminotransferases. 28 1.3. Biosynthesis of -Amino Acids Due to the inherent drawbacks of chemical synthesis and semi-synthesis of enantiopure -amino acids, greener, novel, biocatalytic asymmetri c approaches, with fewer steps were introduced. 29 Therefore, in vivo biocatalysts using enzymes expressed in host cells like E. coli , and yeast have been investigated. 29 Aminomutases expressed in E. coli , utilized for the asymmetric synthesis of enantiopure -amino acids, can potentially function as a biocata lytic system. 30 Semi-synthesis of chiral amino acids uses hydrolyt ic enzymes such as lipases, 31, 32 acylases and hydantoinases. 33 In addition, new methods have been developed, whic h are based upon asymmetric transformations, 30 such as resolution of racemic amines using amine oxidases, 34 and also transaminases. 35, 36 1.3.1. Biosynthesis of -Amino Acids by Aminomutases A phenylalanine aminomutase from Taxus canadensis ( Tc PAM) catalyzes the 2,3- isomerization of ( S)- -phenylalanine to ( R)- -phenylalanine. 37 The Tc PAM reaction also has broad substrate specificity ( Figure 1.2 a). 38, 39 In addition to catalyzing its forward reaction, Tc PAM catalyzes an amination reaction in the presence of NH 3 and substituted analogs of trans - 5 cinnamate ( Figure 1.2 b) to form enantiomerically pure ( S)- -arylalanine and ( R)- -arylalanine. 40 In an independent studied by another group, this " reverse" reaction was further modified through mutation (Q319M) to increase the -regioselectivity. 41 a b c Figure 1.2. a) 2,3-isomerization of the analogs of ( S)- -phenylalanine to corresponding analogs of ( R)- -phenylalanine by PAM. 38, 39, 42 b) Ammonia addition to substituted trans -cinnamate analogs to form corresponding analogs of ( S)- -phenylalanine and ( R)- -phenylalanine by PAM. 40 c) Ammonia addition to substituted trans -cinnamate analogs to form corresponding (S)- -phenylalanine analogs by PAL. 44 6 Furthermore, a phenylalanine aminomutase (also know n as AdmH) from Pantoea agglomerans ( Pa PAM) produces ( S)- -phenylalanine from ( S)- -phenylalanine. 6, 42 Therefore, phenylalanine aminomutases are efficient catalysts for the enantiopure synthesis of -amino acids. Also, phenylalanine ammonia lyase from Rhodotorula glutinis ( Rg PAL) was reported to catalyze the production of ( S)- -phenylalanine, in the presence of trans -cinnamate analogs and high concentrations (8M) of NH 3.43 This ammonia addition reaction has broad substrate specificity ( Figure 1.2 c). 44 These -phenylalanine analogs are further isomerized to -phenylalanine analogs by phenylalanine aminomutase at >99% ee. 39 In addition, tyrosine aminomutases SgcC4 from Streptomyces globisporus ( Sg TAM), 4 and CmdF from Chondromyces crocatus ( Cc TAM) 5 are also known to produce -tyrosine from -tyrosine. Furthermore, we recently found that Tc PAM can act as a transaminase enzyme in the presenc e of an amino group donor (substrates showing predominan t lyase activity) and an amino group acceptor (cinnamate analog) to form enantiomericall y pure - and -arylalanines. Several amino group donors, including bicyclo compounds, along wi th several amine group acceptors were found to be functional in the transaminase reaction catalyzed by Tc PAM ( Figure 1.3 ). 35 A more detailed explanation will be found in the Chapter 3 of this thesis. 7 Figure 1.3 . Transaminase reaction of Tc PAM. 8 * 120 * 140 * TcPAM : VQRKAEDGADI YGV TTGFG ACSSRRT--NQLSE LQ ES LI RCLLA GV FTK G : 116 PaPAM : LESMVSDERVI YGV NTSM GGFVNYIVPIAKASE LQ NN LI NAVATN V--- G : 113 PcPAL : VMDSMNKGTDS YGV TTGFG ATSHRRT--KQGGA LQ KE LI RFLNA GI FGN G : 146 SgTAM : FEGIAEQNIPI YGV TTGYG EMIYMQVDKSKEVE LQ TN LV RSHSA GV --- G : 100 CcTAM : TQAWGEAQHPI YGV NTGFG ELVPVMIPRQHKRE LQ EN LI RSHAA GG--- G : 86 RsTAL : LGAVIREARHV YGL TTGFG PLANRLISGENVRT LQ AN LV HFLAS GV --- G : 95 PpHAL : VEQIIAEDRTA YGI NTGFG LLASTRIASHDLEN LQ RS LV LSHAA GV --- G : 89 YG6 Tg G LQ L6 g G * 220 * 240 * TcPAM : VP LR GSV SASGDL IPL AY IAGL LIGKPSVIARI GDDVEVPAPE-A LSRV G : 215 PaPAM : IP EK GSL GT SGDL GPL AA IALVCT GQWKAR-YQ G---EQMSGAMA LEKA G : 204 PcPAL : LP LR GTI TASGDL VPL SY IAGL LTGRPNSKA-V GPTGVILSPEEAFKLA G : 242 SgTAM : IP EI GSL GASGDL APL SH VAST LIGEGYVL-RD G---RPVETAQV LAER G : 191 CcTAM : IP QQ GSL GASGDL SPL SH IALA LIGEGTVS-FK G---QVRKTGDV LREE G : 177 RsTAL : VP SR GTV GASGDL TPL AH MVLC LQGRGDFLDRD G---TRLDGAEG LRRGR : 187 PpHAL : IP LK GSV GASGDL APL AH MSLV LLGEGKAR-YK G---QWLPATEA LAVA G : 180 6P G36 aSGDL PL 6 l G G l g 260 * 280 * 300 TcPAM : LRP--FK LQA KEGLALVNGT SFA TALASTVMYDANV LLLLVETLCG MFC E : 263 PaPAM : ISP--ME LSF KEGLALINGT SAMVGLGVLLYDEVKR LFDTYLTVTS LSI E : 252 PcPAL : VEGGFFE LQP KEGLALVNGT AVG SGMASMVLFEANI LAVLAEVMSA IFA E : 292 SgTAM : IEP--LE LRF KEGLALINGT SGM TGLGSLVVGRALEQAQQAEIVTA LLI E : 239 CcTAM : LKP--LE LGF KGGL TLINGT SAM TGAACVALGRAYH LFRLALLATADFV Q : 225 RsTAL : LQP--LD LSH RDA LALVNGT SAM TGIALVNAHACRH LGNWAVALTA LLA E : 235 PpHAL : LEP--LT LAA KEGLALLNGT QAS TAYALRGLFQGED LYAAAIACGG LSV E : 228 6 p L 4 gLaL6NGT l 2 * 420 * 440 * TcPAM : SAN DNPI ID-HANDRAL HG ANF QGSA VGFYM DYVRIAV AG LGK LLFAQFT : 399 PaPAM : SSN DNPL ID-QTTEEVF HNGH FHGQY VSMAM DHLNIAL VT MMN LANRRID : 397 PcPAL : SVN DNPL ID-VSRNKAI HG GNF QGTP IGVSM DNTR LAI AA IGK LMFAQFS : 428 SgTAM : SAN DNPL FF-E-GKEIF HG ANF HGQP IAFAM DFVTIAL TQ LGV LAERQIN : 386 CcTAM : STN DNPL IF-DVPEQTF HG ANF HGQY VAMAC DYLNIAV TE IGV LAERQLN : 372 RsTAL : AVT DNPV FPPDGSVPAL HG GNF MGQH VALTS DALAT AV TV LAG LAERQIA : 378 PpHAL : AVS DNPL VF-AAEGDVIS GGNF HAEP VAMAA DNIALAI AE IGS LSERRIS : 358 DNP6 hg nF g 6 D A6 6 L Figure 1.4. Sequence homology model of Tc PAM, Pa PAM, Sg TAM, Cc TAM, PAL from Petroselinum crispum ( Pc PAL), TAL from Rhodobacter sphaeroides ( Rs TAL), and HAL from Pseudomonas putida ( Pp HAL), showing conserved residues in shaded areas. 9 1.4. Aromatic Amino Acid MIO-dependent Aminomutases Phenylalanine- and tyrosine aminomutases (PAMs 6, 45 and TAMs, 4, 46 respectively) catalyze the 2,3-isomerization of ( S)- -arylalanine to ( R)- or ( S)- -arylalanine. These aminomutases (AMs) are members of the class I lyase -like family and show high sequence homology with its members: phenylalanine ammonia ly ases (PALs), 47 tyrosine ammonia lyases (TALs), 48 and histidine ammonia lyases (HALs) (Figure 1.4). 49 PALs, TALs, and HALs produce aryl acrylates from the corresponding amino -acid substrate by the elimination of ammonia. 1.4.1. 3,5-Dihydro-5-methylidene-4 H-imidazol-4-one (MIO) Prosthesis By analogy, PALs, HALs, and TALs, aminomutases, Sg TAM, 4 Cc TAM, 46 Tc PAM, 45 and Pa PAM 6 all use an electrophilic prosthetic group 3,5-dihy dro-5-methylidene-4 H-imidazol-4-one (MIO), which initiates the reaction. 50 However, early studies suggested that the active s ite prosthesis was a dehydroalanine, which was thought to derive by dehydration of a conserved Ser residue in the histidine ammonia lyase from Pseudomonas putida ( Pp HAL). 51 When the ammonia lyase was treated with of NaB 3H4, and the enzyme was cleaved hydrolytically, [ 3H]- labeled alanine was recovered. This evidence strong ly suggested the identity of a dehydroalanine group in the active site of the enzyme. Further sup port that the dehydroalanine served as an active site moiety came from a substrate-shielding experiment; preincubation of the HAL with histidine prevented the NaB 3H4 reduction. 51 10 It was not until the crystal structure of Pp HAL was solved that the MIO group was first characterized. 50 The MIO group is proposed to form post-translation ally by a tandem of active site residues, typically Ala-Ser-Gly ( Scheme 1.2a ). 50, 52 The amide nitrogen of the Gly residue presumably attacks the carbonyl carbon of Ala to cr eate the five membered imidazolidin-4-one ring precursor, followed by two dehydration steps t o form the MIO moiety. The first dehydration step resembles that found for the formation of the chromophore in the green fluorescent protein (Scheme 1.2b ), 53 yet the overall mechanism of formation is slightly different from that of the MIO. Scheme 1.2: a) Formation of MIO, and b) The Structu re of the Chromophore of Green Fluoroscence Protein. a b The MIO is believed to function as an electrophilic /-unsaturated keto functional group (i.e., a 1,4-Michael acceptor). The nucleophile ami no group of the substrate purportedly attacks 11 the exocyclic methylidene carbon of MIO moiety. Thi s addition reaction is proposed to be driven by the aromatization of the MIO to an imidazole rin g system ( Figure 1.2 ). 50, 52, 54 1.4.2. Reaction Mechanism of MIO-Based Aminomutases Two mechanisms for MIO-based aminomutase were propo sed in early reports and the debates continued until 2010. 38 In one mechanism, the amino group of the amino aci d substrate acts as a nucleophile and attacks the methylidene o f the MIO through conjugate addition, forming an N-alkyl adduct. 55 The N-alkyl group is subsequently expelled from substrat e through /-elimination to form an acrylate reaction intermedi ate. Notably, in MIO-dependent ammonia lyase reactions, this acrylate intermediate is rele ased as the product (Figure 1.2 a).56 However, in the MIO-aminomutase reactions the acrylate remains in the active site for amino group rebound to form the -amino acid product. 38, 57 A second proposed mechanism suggests that -electrons of the phenyl ring of the substrate attack the MIO, by FriedelŒCrafts-like activation, to form a -complex through the ortho -carbon of the substrate ( Figure 1.2 b). 47, 58, 59 The second process was principally assigned to ammonia lyase reactions, 47, 49, 58 but was extended to include the aminomutase reactions. 60 12 Scheme 1.3: Two Proposed Mechanisms for the Convers ion of Substrate to Product in a Generic Aminomutase: a) Amino-group Alkylation Path way and b) FriedelŒCrafts Aryl- Alkylation Pathway. a b It should be noted that Tc PAM and Pa PAM enzymes catalyze residual ammonia lyase activity and produce trans -cinnamic acid as a minor product during steady-sta te catalysis. 38, 45, 61 Likewise, the Sg TAM and Cc TAM enzyme show similar relictual chemistry, produc ing trans -coumarate and -tyrosine. 46, 62 13 a b Figure 1.5 . The crystal structure of Sg TAM with a bound mechanism based inhibitor, a) 2,2- difluoro-(3 R)- -tyrosine (PDB 2QVE) and b) (3 R)-amino-2,2-difluoro-3-(4'- methoxyphenyl)propanoic acid (PDB 2RJS) The mechanism by which these enzymes catalyze their reactions remained ambiguous until now. Two Sg TAM, crystal structures were resolved with separate ly bound mechanism-based inhibitors, 2,2-difluoro-(3 R)- -tyrosine (PDB 2QVE), 63 and (3 R)-amino-2,2-difluoro-3-(4'- 14 methoxyphenyl)propanoic acid (PDB 2RJS), 63 where the amino group of the inhibitor was bound to the MIO group ( Figure 1.5 ). However, in the present study described in this the sis, in collaboration with crystallographer Prof. James Geiger, we resolved the crystal structu re of Pa PAM forming a covalent adduct partially with both ( S)- -phenylalanine and ( S)- -phenylalanine (PDB 3NUV). Both ligands were attached to the MIO group by the amine group ( cf. chapter 2), which showed definitively that the MIO formed a covalent adduct during the ca talysis of MIO-dependent aminomutases (Figure 1.2 a). 64 The deamination rate of the NH 2-MIO formed after removal of NH 3 from the substrate was determined for the first time, for an y MIO-dependent enzyme, in this study. Burst phase kinetic analysis of Tc PAM catalyzing the lyase reaction on ( S)- -styrylalanine was used to calculate the residence time of the NH 2-MIO adduct. 65 This study is explained in further detail in Chapter 4. 1.4.3. Stereochemistry of the Phenylalanine Aminomutase Ca talysis Due to their dependence on the MIO moiety, ALs and MIO-based AMs likely follow similar mechanistic courses. Thus, trans- stereoisomers of cinnamate and coumarate represent intermediates on the reaction pathway of the phenyl alanine and tyrosine aminomutases, respectively. However, Sg TAM and Pa PAM produce predominantly the corresponding ( S)- -amino acids with stereochemistry opposite to that o f the ( R)- -amino acids products made by Tc PAM and Cc TAM. However, the mechanism by which these enzymes control their stereoselectivity was further elucidated and valida ted for the first time in this thesis work. To provide detailed mechanistic information, the stere ochemistry of the migrating NH 2/H pair upon 15 reattachment to the phenylpropanoid skeleton to for m the -amino acid isomer was conducted for two PAM enzymes, Tc PAM, 66 and Pa PAM. 61 The studies showed that both PAMs initiate the catalysis by abstracting pro -(3 S) proton and eliminating the NH 2 of the substrate to form a trans -cinnamic acid intermediate. 61, 66 Scheme 1.4: Stereochemistry of the Migrating Amino Group and the Proton of the Substrate and the Product during the Catalysis of a ) Tc PAM and (The Configurations at both C and C Positions are Retained) b) Pa PAM (The Configurations at both C and C Positions are Inverted) is Shown.* a b * Color Coding of the Protons: Pro (3S) Proton (Red ), Pro (3R) Proton (Blue) and C Proton. The abstracted proton and the amino group, dwelling momentarily in the active site, exchange positions and are rebounded to C and C positions respectively to form the final -phenylalanine product. Stereospecific deuterium lab eling studies revealed that Tc PAM proceeds 16 through retention of configuration at both C and C (Scheme 1.4a), 66 whereas Pa PAM through an inversion of configuration at both terminal carb ons (Scheme 1.4b). 61 Even though the stereochemical studies reveal the different modes ( retention or inversion of configuration) of the stereochemical reattachment of the NH 2/H pair, we also provided preliminary data showing how these enzymes create a microenvironment that can in fluences the mode of retention or inversion of configuration. We propose that the si or re face selectivity of the intermediate trans -cinnamic acid governs the observed stereochemistries during catalysis of all MIO dependent aminomutases. After the initial E2-type anti elimination step, the resultant intermediate can reattach at C to the amino group bound by the MIO amino group on the same face or the opposite face. If the rebound approaches on the sam e ( si ) face, ( S)- -stereochemistry results from inversion of configuration at C and C of Pa PAM and Sg TAM. Alternatively, if rebound approaches from the opposite ( re ) face, the observed ( R)- - stereochemistry results from retention of configuration at both migration termin i of Tc PAM and Cc TAM. 38, 61, 64, 67 There is a debate over how the Tc PAM and Cc TAM present the opposite face of the intermediate to the MIO-bound NH 2 group. Wu et al., in a separate account, proposed (by inference from empirical data from our group) that after the initial E2-type anti elimination step in the Tc PAM reaction sequence, a 180° rotation of the C C ipso bond of the cinnamate intermediate, in complex through a strong bidentate salt bridge with Arg325 and H-bonding with surrounding residues, is preferred. This rotation i s purported to re-orient the carboxylate of the cinnamate, after breaking the salt bridge, into a f avorable trajectory to form a monodentate interaction and H-bonding network with Arg325 and f urther H-bond interactions with Gln319, 17 Asn355, and Asn231 ( Figure 1.6 ). It should be noted that during this proposed re- orientation, to place the C at the correct geometry for the amino group attack , the phenyl ring is forced into a sterically crowded active-site pocket that is elect ronically unfavorable (Glu455, Lys427, and Ile431) for binding an aromatic ring. Therefore, ro tation about the C C ipso bond alone followed by lateral translation of the phenylpropanoid skele ton into a proper trajectory seems relatively unlikely as a mechanism to present the re face of the cinnamate intermediate to the NH 2-MIO adduct. The authors of the independent study with Tc PAM continue and suggest that the loss of bidentate salt bridge and the formation of H-bonds lower the energy of the LUMO of the carboxylate to promote the conjugate addition of th e NH 2 bound to the MIO at C . Thus, the Q319M mutant made Wu et al. should lose an H-bondin g partner, and thus increase the LUMO of the carboxylate group compared to the wild type enzyme. Accordingly, the preference for conjugate addition should decrease, based on the au thors' explanation. Interestingly, the Q319M mutant shows higher -regioselectivity, compared to the wild type Tc PAM during the catalyzed addition of NH 3 to cinnamic acid, used in this case as a substrate . The authors continue to suggest that interactions between cinnamic acid and the residues in the Q319M mutant oriented the cinnamate at a trajectory that disfavored the f ormation of -phenylalanine ( Figure 1.6 ). 41 The latter observation seems paradoxical; the Q319M mutant, with one less H-bonding interaction, should likely favor rotation of the ci nnamate back to the stronger bidentate salt bridge interaction with Arg325, and thus preferenti ally make -phenylalanine. Also, the authors investigate an Arg325K mutant, showing increased -regioselectivity. Lysine, like arginine, is 18 known to form strong electrostatic interactions (po int charges) and H-bond interactions with counter-charged carboxylate anions. In our view, th is electrostatic complex would increase the LUMO energy level of the cinnamate carboxylate grou p and should therefore dissuade conjugate addition of the amino group at C of the /-unsaturated carboxylate in the PAM reaction. Furthermore, the authors argue that the carboxylate group of the cinnamate intermediate is an electron sink. They conclude that during the conjug ate addition, the -bond electrons between CC , shift to form a -bond between C C 1, placing a negative charge on both of the carboxylate oxygens. They continue to surmise that the migratory pro -(3S) proton (which is partially exchanged with solvent protons as dissect ed in our laboratory) could transfer to one of the carboxylate oxygens, which then would be vulner able for exchange with protons within the enzyme. If this mechanistic consideration is evalua ted across all PAM enzymes, then the Pa PAM reaction mechanism should proceed similarly. Ho wever, 100% of the migratory pro -(3 S) proton transfers from C to C in the Pa PAM-catalyzed isomerization reaction, 61 an observation that contradicts the one posited by the other research group. An additional piece of evidence shows that Pa PAM and Sg TAM, which form the ( S)- -amino acid isomers, do not form a bidentate salt bridge with the Arg residue positi oned analogously as that found in Tc PAM and Cc TAM. 63, 64 Instead, Pa PAM and Sg TAM lock their substrates in a monodentate interact ion through sterics and/or electrostatics. These trajec tories are analogous to those proposed by Wu, et al. for Tc PAM after the enzyme rotates the cinnamate intermed iate into a favorable position for amino group attack at C . Thus, Pa PAM and Sg TAM, according to the rationale posited by Wu, et al., regarding the cinnamate trajectory (abo ve), should catalyze the addition of NH 3 to the 19 cinnamate substrate in the "reverse" and produce pr edominantly the -amino acid isomer. However, our group showed that the "reverse" reacti on catalyzed by Pa PAM produces a 50:50 distribution of - and -phenylalanine products. 64 a b Figure 1.6. a) Crystal structure of Tc PAM with bound cinnamic acid intermediate. b) The proposed orientation of the cinnamic acid intermedi ate which leads to the formation of ( R)- -phenylalanine with a retention of configuration at both migration termini, by Wu. et al. (Modeled using Pymol from PDB# 3NZ4) After consideration of our empirical data and the t heoretical speculations offered by Wu, et al., we support a mechanism for aminomutase that ma ke ( R)- -amino acid isomers that proceed via 180º rotation of the transient cinnamate intermedi ate about the C 1C and C C ipso bonds. 20 In view of a molecular model, this crankshaft rotat ion results in minimal translational displacement of the phenyl ring, the rotating carbo ns, and the carboxylate group. Also, this model keeps the bidentate salt bridge intact betwee n the carboxylate and active site arginine. This rotation mode also minimizes unfavorable steri c and electronic interactions between the phenyl ring and active site residues. Chapter 2 pro vides details on the monodentate salt bridge between the ( S)- -amino acid ligands in Pa PAM and Sg TAM, formed without rotation of the cinnamate intermediate. The adducts shown in the cr ystal structures of these enzymes are contrasted to the bidentate salt bridge interaction s between the ligand found in Tc PAM and purported to reside in Cc TAM. 38, 61, 64 Prior to the work studied in this body of work, onl y the stereochemistry of the PAM reactions was known. This thesis explores the complete direct stereochemical analysis for a TAM enzyme. The stereochemistry of Sg TAM was indirectly explored using an epoxide ((2 S,3 R)-3-(4- fluorophenyl)oxirane-2-carboxyate) as suicide subst rate. The epoxide analog formed a covalent interaction with the MIO through the epoxide oxygen followed by an epoxide ring opening by a water molecule. Inversion at C 3 of the substrate in the active site resulted in a single diastereomer (2 R,3 S)-2,3-dihydroxy-3-(4-fluorophenyl)propanoic acid ob served in the crystal structure. The observed diol was proposed to be gen erated either through an S N1 or S N2-type mechanism, yet with net inversion of configuration at C 3.63 While this was an elegant inference of the reaction stereochemistry of Sg TAM, we sought to understand the complete stereochemistry of a homolo gous MIO-dependent tyrosine aminomutase ( Cc TAM). So far, five MIO-dependent aminomutases are k nown to isomerize either ( S)- -phenylalanine (EncP from Streptomyces maritimus ,68 AdmH ( Pa PAM) from 21 Pantoea agglomerans ,6, 61 and Tc PAM from Taxus plants 45, 66 ) or ( S)- -tyrosine (( S)-1) (SgcC4 (Sg TAM) from Streptomyces globisporus ,57, 69 and CmdF ( Cc TAM) from Chondromyces crocatus ,46 further described in Chapter 2 and 5) to their resp ective -amino acids. Efforts to optimize the turnover of this family of aminomutase s provide a potentially alternative means toward scaleable biocatalytic production of novel e nantiomerically pure -amino acids as synthetic building blocks in medicinal chemistry. E arlier substrate specificity studies showed that Tc PAM can convert substituted aromatic and heteroarom atic -alanines to the corresponding -alanines. Apparently, since TAMs requires the 4'-hy droxyl group on the substrate for catalysis, their use to biosynthesize -tyrosine analogs is limited. 46, 57 By employing stereospecifically deuterium-labeled tyrosine substrates we were able to use NMR and GC/EIMS analyses to dissect the absolute stereochemistry of the Cc TAM reaction, which is explained further in Chapter 5. 67 22 2. INSIGHTS INTO THE MECHANISM OF MIO-BASED PHENYLALANINE AMINOMUTASE CATALYSIS Reproduced in part with permission from [Feng, L.; Wanninayake, U.; Strom, S.; Geiger, J.; Walker, K. D., Mechanistic, mutational, and structu ral evaluation of a Taxus phenylalanine aminomutase. Biochemistry 2011, 50, (14), 2919-2930], and from [Ratnayake, N. D.; Wanninayake, U.; Geiger, J. H.; Walker, K. D., Ster eochemistry and mechanism of a microbial phenylalanine aminomutase. J. Am. Chem. Soc. 2011, 133, (22), 8531-8533] Copyright © 2011 American Chemical Society. Reproduced in part with permission from [Strom, S.; Wanninayake, U.; Ratnayake, N. D.; Walker, K. D.; Geiger, J. H., Insights into the mec hanistic pathway of the Pantoea agglomerans phenylalanine aminomutase. Angew. Chem., Int. Ed. Engl. 2012, 51, 11-17] Copyright © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. 2.1. Introduction As described in Chapter 1 of this thesis, methylide ne imidazolone (MIO)-dependent aminomutases (AMs) 4, 6, 45 , 46 and ammonia lyases 47 -49 are members of the class I lyase-like family and use phenylalanine, tyrosine, and histidi ne amino acids as substrates. The ALs produce arylacrylates from the corresponding amino acid sub strate by the elimination of ammonia, 47 -49 while the AMs isomerize the -amino acid substrates to their -amino acids. 4, 6, 45 , 46 These enzymes use an active site prosthesis MIO to cataly ze their reactions. In one proposed mechanism for this group of enzymes, the amino grou p of the amino acid substrate acts as a nucleophile 55 and in the second, the -electrons at the ortho -carbon atom of the phenyl ring of 23 the substrate attack the Lewis acidic MIO by Friede lŒCrafts-like activation (see Scheme 1.3). 47 , 58 , 59 Production of trans -cinnamate analogs as a result of residual ammonia lyase activity of the MIO-dependent aminomutases during steady-state cata lysis, 38 , 46 , 61 , 62 suggests that the lyases and aminomutases follow analogous mechanistic cours es and that trans- cinnamate analogues represent intermediates on the reaction pathway of the aminomutases. The structures of several enzymes of the MIO-depend ent family were characterized in earlier reports. 50 The structure of Rhodobacter sphaeroides tyrosine ammonia lyase ( Rs TAL) in complex with the competitive inhibitor 2-aminoindan -2-phosphonic acid, which is covalently bound by its amino group to the MIO in the active s ite (PDB 2O7E) was determined. 70 Yet, despite acquiring the structure, neither MIO-based mechanism was discussed for the Rs TAL reaction. 70 Also available are structures of the tyrosine amin omutase from Streptomyces globisporus ( Sg TAM) in complex with inhibitors 2,2-difluoro-( S)- -tyrosine (PDB 2QVE) and separately with (3 S)-amino-2,2-difluoro-3-(4'-methoxyphenyl)propanoic acid (PDB 2RJS) (see Chapter 1). 63 These -amino acids were alkylated at their amine groups b y the MIO, which is consistent with the aminoalkyl mechanism that invol ves the elimination of an alkylamine (see Chapter 1). However, automated docking and molecula r dynamics simulation studies of phenylalanine ammonia lyase from Petroselinum crispum ( Pc PAL, PDB 1W27) 71 suggested that the FriedelŒCrafts mechanism is nevertheless used b y both PAL and PAM enzymes. 60 A debate about the mechanism used by these MIO-based enzymes remained open. In addition, the crystal structure Sg TAM in complex with ( S)- -tyrosine (PDB 3KDZ) showed that Y63 was in close proximity to the geminal protons of the substrate. A non-functional Y63F mutant suggested the 24 important catalytic function of the highly conserve d tyrosine residue towards abstracting the C proton during the ,-elimination step of the aminomutase reaction. 54 In addition to acquiring crystal structure data to inform on the mechanism o f the PAM reactions, experiments were conducted to assess whether the isomerization occur red inter- or intramolecularly. Earlier studies on the Tc PAM mechanism using cell free assays of Taxus cambial bark 37 and recombinantly expressed Tc PAM 66 inferred that the amino group of the ( S)- -phenylalanine substrate was transferred to C intramolecularly, while the proton was reciprocall y transferred partially intramolecularly. It was clear from these earlier s tudies that the migratory amino group and hydrogen reattached to the phenylpropanoid skeleton with retention of configuration at C and C. It should be noted, that assays with the purified , recombinant Tc PAM contained a cleaner metabolite background with fewer contaminating enzy mes than the assays derived from the crude cell free extracts of Taxus bark. 37 While this early study elucidated the mode of the NH 2-group transfer, the influence of cinnamate and othe r cinnamate modifying enzymes in the crude plant enzyme extract on the NH 2-transfer process was not considered. Therefore, in the current study, the intramolecular mode of NH 2-transfer was re-examined by incubating purified, recombinant Tc PAM with a mixture of [ring, -C- 2H6]- trans -cinnamate and [ 15 N]phenylalanine. This confirmatory study was important since a purif ied, recombinantly expressed Taxus PAM was recently shown to catalyze the formation of -phenylalanine from cinnamate and ammonia co-substrates. 72 In this scenario, it could be imagined that cinnam ate molecules could bind and de-bind the holo (NH 2-MIO) form of Tc PAM before the NH 2-rebound step. 25 To provide further information on the mechanism of this class of catalysts, the X-ray crystal structures of Pa PAM 64 and Tc PAM 38 were determined. The Tc PAM structure along with modeled-in substrates also showed that the 3'-carbo n of the aromatic ring of the bound cinnamate intermediate/substrate is proximate to and makes a direct hydrophobic interaction with Leu104. This residue and several other distal hydrophobic i nteractions between the aromatic ring of cinnamate and Leu179, Leu227, and Val230 likely con tribute to substrate binding and to defining the topology of the active site. In additi on, the mechanism and stereochemistry of how (inter- or intramolecularly) the migratory groups a ttach at their receiving carbon centers (with retention or inversion of stereochemistry) during t he reaction was evaluated. 2.2. Experimental 2.2.1. Chemicals (S)- -Amino acids ( R)/( S)- -amino acids were commercially available from PepTe ch Corporation (Burlington, MA), except for styryl-( R)- -alanine, which was synthesized by modification of a described procedure. 73 5-Phenyl-(2 E,4 E)-pentadienoic acid (( E,E)- styrylacrylate) was purchased from Alfa Aesar (Ward Hill, MA). [ring, -C- 2H6]- trans -cinnamic acid and trans -2 methylcinnamic acid were obtained from Sigma-Aldric h (St. Louis, MO), and [15 N]phenylalanine was obtained from Cambridge Isotope Laboratories (Andover, MA). 2.2.2. Expression and Purification of Tc PAM Codon-optimized wild-type 74 or a mutant Tc PAM derived from the optimized clone was overexpressed in Escherichia coli BL21 (DE3) cells. The cells (six 1-L cultures) wer e grown in LuriaBertani medium supplemented with kanamycin (5 0 g/mL), induced for expression with 26 isopropyl- -D-thiogalactopyranoside (100 M) at 16 °C, and, after 16 h, were harvested by centrifugation (4,000 g for 20 min). To the cell pellet was added 100 mL o f resuspension buffer (50 mM sodium phosphate containing 10 mM imidazole, 5% (v/v) glycerol, and 300 mM NaCl, pH 8.5), the suspension was lysed by sonication, an d the cellular debris and light membranes were removed by centrifugation at 9,700 g (45 min) then at 102,000 g (1 h), respectively. The resultant crude aminomutase in the soluble fraction was purified by nickel-nitrilotriacetic acid affinity chromatography according to the protocol d escribed by the manufacturer (Qiagen, Valencia, CA); Tc PAM was eluted in 250 mM imidazole (3 mL total volu me, at >90% purity by SDSPAGE with Coomassie Blue staining). The fraction of protein that eluted from the nickel -nitrilotriacetic acid affinity column containing active Tc PAM (76 kDa) was exchanged with elution buffer (20 mM Tris-HCl, pH 7.5, 50 mM NaCl, and 5% glycerol) through several c oncentration/dilution cycles using a Centriprep centrifugal filter (30,000 MWCO, Millipo re) to remove the 250 mM imidazole. The protein, concentrated to 1 mL, was loaded onto a pr e-calibrated gel filtration chromatography column (Superdex 200 prep grade, GE Healthcare Life Sciences) connected to a Pharmacia FPLC system (comprising a Biotech Recorder 102, Pha rmacia Pump P-500, and Liquid Chromatograph Controller LCC-500). Tc PAM was eluted at 1 mL/min with elution buffer and was isolated in two peaks, one with an elution volu me consistent with that of a 150 kDa MW standard, considered to be a dimer, and a larger pe ak consistent with protein >300 kDa, likely a multimeric aggregate. 27 2.2.3. Expression and Purification of Pa PAM. Pa PAM was expressed according to a previously describ ed procedure, and is described briefly here. 61 E. coli cells (twelve 1-L cultures) were grown in LuriaBe rtani medium supplemented with kanamycin (50 µg/mL), induced for expression with isopropyl- -D- thiogalactopyranoside (100 µM) at 16 °C and harvest ed by centrifugation after 16 h. To the cell pellet was added 125 mL of resuspension buffer (50 mM sodium phosphate containing 5% (v/v) glycerol, 300 mM NaCl and 10 mM imidazole, pH 8.0), the cells were lysed by sonication, and the cellular debris and light membranes were remove d by centrifugation. The resultant crude aminomutase in the soluble fraction was purified by nickel-nitrilotriacetic acid affinity chromatography according to the protocol described by the manufacturer (Qiagen, Valencia, CA); Pa PAM eluted in 250 mM imidazole. Fractions containin g active soluble Pa PAM (59 kDa) were combined, and the buffer was exchanged with 50 mM sodium phosphate buffer (pH 8.0) containing 5% (v/v) glycerol through several concen tration/dilution cycles using a Centriprep centrifugal filter (30,000 MWCO, Millipore) to a fi nal concentration of 7 mg/mL (calculated by the Bradford protein assay). The purity of the conc entrated enzyme was assessed at >95% by SDSPAGE with Coomassie Blue staining. 2.2.4. PAM Activity Assays An aliquot (50 L) of the fraction containing Tc PAM (~ 80 g) from the nickel-nitrilotriacetic acid affinity column elusions was added to 50 mM ph osphate buffer solution (1 mL final volume, pH 8.5) containing 5% glycerol and ( S)- -phenylalanine (100 M), and the solution was incubated at 31 °C for 1 h. This procedure was repe ated for Pa PAM. 28 Equimolar amounts of protein from each fraction (40 g of the 150 kDa fraction and 80 g of the >300 kDa fraction) from the gel extraction chro matography were separately incubated with (S)- -phenylalanine (100 M) at 31 °C for 1 h. Products were analyzed using G C/EIMS (described in section 2.2.5 ). The aminomutase in the 150 kDa fraction had 3-fo ld higher activity than in the >300 kDa fraction; therefore, the fract ion containing the dimer was used for the crystallographic study, loaded into a size-selectiv e centrifugal filtration unit (30,000 MWCO), and concentrated to 10 mg/mL. The purity (>95%) of the concentrated enzyme was assessed by SDSPAGE with Coomassie Blue staining, and the quan tity was determined by the Bradford protein assay. 2.2.5. GC/EIMS Analysis of PAM Catalyzed Products The reactions were quenched with 6 N HCl, the pH wa s adjusted to pH 2, and then internal standards 3'-fluoro-( R)- -phenylalanine (at 20 M) and trans -2 -methylcinnamic acid (at 20 M) were added. The cinnamic acids were extracted into diethyl ether, the organic solvent was evaporated in vacuo , the residue was dissolved in ethyl acetate/methan ol (3:1, v/v) (200 L), and the solution was treated with a TMS-diazomethane di ssolved in ether (~ 5 L) to convert the acids to their methyl esters. Each sample was separ ately analyzed by GC/EIMS and quantified by linear regression analysis. The remaining aqueou s fraction was adjusted to pH >10 with 6 N NaOH, and the amino acids were derivatized to their N-di(ethoxycarbonyl) esters, as before. Each sample was separately analyzed/quantified by G C/EIMS; in brief, the relative amounts of the - or -amino acid were determined by linear regression an alysis of the area of the base peak ion of the derivatized - and -amino acids generated in the EIMS. The peak area w as converted to concentration by solving the corresponding linea r equation, derived by plotting the area of the 29 base peak ion (produced by the corresponding authen tic standard) against concentration ranging from 0 to 1.5 mM. 2.2.6. Mutagenesis of Tc PAM cDNA for Expression of L104A Mutant. Point mutations of Tc PAM were performed using the QuikChange II XL Site- Directed Mutagenesis Kit (Stratagene, La Jolla, CA). The oli gonucleotide primers used to exchange residue L104 to alanine (Ala) in the mutagenesis re actions were as follows: L104A (forward primer): 5 ŒGCAGGAGAGC GCC ATCCGCTGTCŒ3 (mutations are bold and underlined). The corresponding reverse complement primer (5 ŒGACAGCGGAT GGC GCTCTCCTGCŒ3 ) was paired with the respective forward primer. Expressi on vector pET-28a(+) containing the wild- type tcpam cDNA (10 ng) was used as the DNA template in the P CR reactions, the corresponding forward and reverse primers (150 ng e ach), dNTP mix (1 µL), 10X reaction buffer (5 µL), QuickSolution reagent (3 µL), PfuUltra HF DNA polymerase (1 µL at 2.5 U/µL) and ddH2O to bring the volume to 50 µL. The PCR program conditions were as follows: initial denaturing at 95 °C for 2 min followed by 30 cycles at 95 °C for 50 s, 55 °C for 50 s, and 68 °C for 7 min, and finally, the reactions were held at 68 °C for 7 min. The reactions were place on ice, to each was added restriction enzyme Dpn I (1 µL at 10 U/µL), and the reactions were incubated at 37 °C for 1 h to digest the template D NA. An aliquot (2 µL) of the plasmid solution from each PCR reaction was used to transform XL10-G old ultracompetent cells (provided in the QuikChange II XL Site-Directed Mutagenesis Kit). Th e resultant plasmids encoding the L104A mutation in the tcpam cDNA (designated as PAMeLA_104: phenylalanine amino mutase exchange Leu Ala_104) were confirmed by sequencing the correspond ing tcpam 104A cDNA. The mutant gene was expressed as described for the wild-type. 30 2.2.7. Analysis of Kinetic Parameters of Tc PAM and PAMeLA_104 The substrates ( S)- -phenylalanine, 3'- and 4'-methyl-( S)- -phenylalanine, and styryl-( S)- -alanine were incubated separately with wild-type Tc PAM and PAMeLA_104 to establish steady- state conditions with respect to a fixed protein co ncentration and time at 31 °C. Under steady- state conditions, each substrate at 10, 20, 40, 80, 150, 300, 500, and 750 M was separately incubated with Tc PAM or PAMeLA_104 in triplicate, single stopped-tim e assays. Each of the products in the reaction mixture was derivatized to their N-(ethoxycarbonyl) methyl ester and then quantified by GC/EIMS analysis, as described p reviously. The kinetic parameters ( KM and kcat ) were determined from the Hanes-Woolf plot ( R2 was typically >0.98), and the stereochemistry of enzyme-catalyzed products was as sessed by chiral GC/EIMS analysis (Chirasil-D-Val column, Varian). 39 2.2.8. Analysis of Products Formed from the Incubation of Tc PAM with [15 N]Phenylalanine and [ring, -C- 2H6]- trans -Cinnamate Acid Purified Tc PAM (100 g) was incubated with a mixture of [ 15 N]-phenylalanine and [ring, -C- 2H6]- trans -cinnamate acid (10 mmol of each) for 1 h. The - and -phenylalanine isotopomers were derivatized to their N-(ethoxycarbonyl) methyl esters, and cinnamic acid was converted to its methyl ester and analyzed by GC/EIMS. 31 2.2.9. Assessing the Product Distribution of Pa PAM and Tc PAM with 2'-Methyl-( S)- -Phenylalanine Pa PAM and Tc PAM, each at 20 µg/mL in separate 1-mL assays, were incubated with 2'- methyl-( S)- -phenylalanine (at 1 mM) at 31 °C for 1 h. Internal standards 3'-fluoro-( R)- -phenylalanine and trans -cinnamic acid (each at 20 µM) were added. The reac tions were acidified (pH 2, 6 N HCl) and the cinnamic acid analogues wer e separately extracted (3 × 1 mL) into ethyl acetate. An aliquot of a TMS-diazomethane solution was added to convert the acids to their methyl esters. The remaining amino acids were deriv atized to their ethoxycarbonyl methyl esters according to a previously described method separate ly analyzed by GC/EIMS, and the analytes were quantified by comparison against a standard cu rve. 2.2.10. Conversion of Cinnamate to - and -Phenylalanine by Pa PAM Pa PAM (7 mg/mL) was incubated for 48 h at 31°C with c innamic acid (1.25 mM) in phosphate buffer (pH 8.5) containing 60 µM of ammon ia in triplicate 1-mL assays. The reactions were quenched with 6 M NaOH and 3'-fluoro-( R)- -phenylalanine (20 nmol) was added as the internal standard. To each assay was added ethyl ch loroformate (100 µL) to N-carbonylate the amino acids. The assay solutions were acidified (6 N HCl) to pH 2 on ice, the putative, derivatized amino acids were extracted with ethyl a cetate (3 × 1 mL), and TMS-diazomethane dissolved in methanol was added drop wise to the or ganic fractions until the yellow color of diazomethane persisted. The organic solvent was rem oved, and the residue from each sample was re-dissolved in 3:1 (v/v) ethyl acetate:methano l (200 µL) and analyzed by GC/EIMS, and the base peak ions were evaluated. The amount of th e derivatized - and -amino acid was quantified as follows. Ratio 1 (abundance of the ba se peak ion of the internal standard at 20 µM: 32 abundance of the base peak ion of the analyte) was used to calculate product concentration. The abundance of the base peak ion produced by the corr esponding authentic standard at varying concentrations in a dilution series was plotted aga inst the abundance of the base peak ion produced by the internal standard (at 20 µM). The l inear equation derived from this plot was used to convert Ratio 1 to analyte (i.e., product) concentration. 2.3. Results 2.3.1. X-ray Crystal Structure of Tc PAM In the Tc PAM structure (PDB 3NZ4), A trans -cinnamate molecule is bound in the active site, lying above the MIO and under a loop region that in cludes residues 8097, which define the top of the active site ( Figure 2.1a ). The trans -cinnamate molecule lies about 3.4 Å above the methylidene carbon of the MIO moiety. The carboxyla te of the cinnamate makes a salt bridge interaction with a strongly conserved Arg325, which serves to position the product in the active site. The plane of the aromatic ring of the cinnama te was observed to be displaced ~ 40° relative to the -bond plane of the propeonate C=C double bond ( Figure 2.1a ). The aromatic ring is bound relatively loosely in the active site, making only one direct hydrophobic interaction with Leu-104 (3.3 Å, not shown). The only other close co ntact between cinnamate and an active site residue is a 2.9 Å interaction between the hydroxyl of Tyr80 and C of the cinnamate ( Figure 2.1b ). 38 33 a b Figure 2.1 . The Tc PAM active site-cinnamate complex (magenta) a) indi cating the displacement of plane of the aromatic ring of the cinnamate rela tive to the -bond plane of the propeonate carboncarbon double bond, and (b) the topside view . The PAM active site residues are colored by atom as follows: C (green), O (red), and N (blue ). 34 a b Figure 2.2 . (b) the 3 -methyl-( S)- -phenylalanine substrate is modeled into the mutant PAMeLA_104 active site where the steric volume is i ncreased through the mutation of Leu-104 to Ala; now, the closest distance between the bound substrate and residue 104 is estimated at ~ 4.4 Å, and The Tc PAM-cinnamate complex is used to approximate the tr ajectory of non-natural substrates. (a) The 3 -methylphenyl- -alanine substrate is modeled into the active site of Tc PAM showing the distance (~2 Å) between the 3 -methyl group of the substrate and Leu-104. 35 Figure 2.2 (cont'd). c d c) styryl-( S)- -alanine modeled in the Tc PAM active site showing the hydrophobic interaction with Leu 104 at a distance of ~ 2.2 Å. . (d) 3 - and 4 -methyl-( S)- -phenylalanine are superimposed and modeled in the Tc PAM active site. The relative positions of the meth yl substituents on the phenyl ring of each substrate t o Leu-104 are shown along with the distances (~ 2 and ~ 3.8 Å, respectively) between the C of Leu-104 and 3 -methyl and 4 -methyl. 36 2.3.2. Site-Directed Mutation of an Active Site Leu of Tc PAM As noted earlier, the Tc PAM-cinnamate complex shows that the 3'-carbon of t he bound intermediate is close to residue L104, likely invol ved in a van der Waals interaction with the substrate ( Figure 2.2 ). L104 of Tc PAM was changed to an A104 mutant (designated as PAMeLA_104) to assess how this would affect the ami nomutase reaction with sterically demanding non-natural substrates compared to the re action with the wild-type enzyme. The kinetic parameters were obtained separately for Tc PAM with substrates 3 -methyl-( S)- -phenylalanine, styryl-( S)- -alanine, and 4 -methyl-( S)- -phenylalanine and compared to those obtained for PAMeLA_104 in parallel enzyme assays w ith the same substrates ( Table 2.1 ). Notably, ( R)- -arylalanines and the various trans- cinnamate ( Table 2.1 ) are products of the Tc PAM reaction when ( S)- -arylalanines are used as substrates; 45 therefore, the sum of their production rates at steady-state was used to calcul ate the turnover rate ( kcat ) ( Table 2.1 ). The catalytic efficiency ( kcat /KM) of Tc PAM and PAMeLA_104 for the natural substrate ( S)- -phenylalanine 1 is 1100 M 1 ·s 1 and 560 M 1 ·s 1 , respectively. The product distribution of the Tc PAM reaction, after 30 min, was dominated by ( R)- -phenylalanine, made at 0.053 s 1 , and cinnamate was made more slowly at 0.012 s 1 . Reciprocally, the product pool of the PAMeLA_104 reaction with 1 principally contained cinnamate, which was made at a rate of 0.073 s 1 , while the -amino acid was made fractionally at 0.003 s 1 . 37 Table 2.1 . Kinetic Parameters for Tc PAM and PAMeLA_104 with Various Substrates Substrate Enzyme KM (M) × 10 3 kcat (s -1 )a cin kcat (s -1 )a kcat /KM (M -1 Łs -1 ) Tc PAM 0.057 (±0.004) 0.053 (±0.001) 0.012 (±0.002) 1100 (±100) 1 PAMeLA 0.136 (±0.002) 0.003 (<±0.001) 0.073 (±0.003) 560 (±10) Tc PAM 0.397 (±0.030) 0.022 (±0.002) 0.017 (±0.002) 100 (±20) 2 PAMeLA 0.083 (±0.008) 0.028 (±0.003) 0.032 (±0.003) 720 (±20) Tc PAM 0.091 (±0.005) 0.030 (±0.002) 0.005 (±0.002) 380 (±20) 3 PAMeLA 0.073 (±0.005) 0.020 (±0.002) 0.017 (±0.002) 510 (±10) Tc PAM 0.250 (±0.004) <0.0002 0.082 (±0.002) 330 (±10) 4 PAMeLA 0.120 (±0.004) 0.003 (±0.002) 0.12 (<±0.01) 1030 (±20) a. The terms kcat and cin kcat represent the kinetic constants for the formation of the -arylalanines and trans-arylacrylates, respectively. b. The kinetic constant k cat = kcat + cin kcat . (S)- -Phenylalanine ( 1), 3 -methyl-(S)- -phenylalanine ( 2), 4 -methyl-(S)- -phenylalanine ( 3), and styryl-(S)- -alanine ( 4) at steady-state. Standard errors are in parentheses. 38 The catalytic efficiency of Tc PAM and PAMeLA_104 for 3 -methyl-( S)- -phenylalanine ( 2) is 48 M 1 ·s 1 and 720 M 1 ·s 1 , respectively. The increase in catalytic efficienc y for substrate 2 is due largely to the ~ 5-fold decrease in KM of PAMeLA_104 (83 M) compared to the KM of Tc PAM (397 M). The distribution of the 3 -methyl-( R)- -phenylalanine and 3 -methylcinnamate was at ~1:1 for both PAMeLA_104 and Tc PAM catalysis; however, the combined rate of formation of both 3 -methyl-( R)- -phenylalanine and 3 -methylcinnamate by PAMeLA_104 (0.060 s 1 ) is slightly increased compared to the rate of the same reaction by Tc PAM (0.039 s 1 ) ( Table 2.1 ). Interestingly, both Tc PAM ( kcat = 0.035 s 1 ) and PAMeLA_104 (kcat = 0.037 s 1 ) were kinetically similar when 4 -methyl-( S)- -phenylalanine ( 3) was used as the substrate, and their KM values, 91 M and 73 M, respectively, were comparable as well as their catalytic efficiency values (380 M 1 ·s 1 and 510 M 1 ·s 1 ( Table 2.1 )). However, the rates of formation of the 4 -methyl-( R)- -phenylalanine and 4 -methylcinnamate products (0.030 s 1 and 0.005 s 1 , respectively) catalyzed by Tc PAM from 3 were significantly different from the respective distribution catalyzed by PAMeLA_104 (0. 020 s 1 and 0.017 s 1 ). Tc PAM was incubated with styryl- -alanine ( 4) and rapidly catalyzed the exclusive conversion of 4 to the corresponding 5-phenyl-(2 E,4 E)-pentadienoate (i.e., ( E,E )-styrylacrylate) at 0.082 s 1 and styryl- (R)- -alanine at <0.0002 s 1 ( Table 2.1 ). Comparatively, PAMeLA_104 converted 4 to the corresponding ( E,E)-styrylacrylate faster, at 0.12 s 1 ; however, styryl-( R)- -alanine was produced significantly faster by PAMeLA_104 than by Tc PAM, yet still at a slow rate of 0.003 s1 . It is worth noting that the Michaelis constants f or both Tc PAM and PAMeLA_104 are 250 M and 120 M, respectively, for the styryl-( S)- -alanine substrate, suggesting that the L104A 39 mutation is able to enhance the binding affinity fo r 4 and thus contributes toward increasing the overall catalytic efficiency from 330 M 1 ·s 1 for Tc PAM to 1000 M 1 ·s 1 for PAMeLA_104. 2.3.3. Inter/Intramolecularity Analysis of the Tc PAM Reaction Table 2.2. GC/EIMS Analysis: Diagnostic Ions of Bio synthetic [ 15 N]- -Phenylalanine a [M] Ł+ = 252 m/z 163 & 192 m/z 179 m/z 185 m/z a The asterisk indicates the 15 N-atom . Authentic standards of N-(ethoxycarbonyl)- -phenylalanine methyl ester eluted from the GC column at 9.87 min. The mass spectrum of the deriva tized biosynthetic ( R)- -phenylalanine (~ 200 nmol) isolated from the incubation allowed for quantitative analysis of the isotope enrichment and distribution. A molecular ion (M +) of m/z = 252 indicated that the biocatalyzed product contained one extra mass unit compared to t he mass of the unlabeled isotopomer ( m/z = 251). The molecular ion and diagnostic fragment ion s (base peak [M CH 3CH 2CO 2]+ = 179, and lesser abundant ions m/z = 192 [M HC(O)OCH 3]+ and m/z = 163 [ m/z 192 CH 3CH 2]+ (Table 2.2)) indicated that the additional mass unit was d erived from the 15 N-atom. In addition, the ratio (10:1) of the ion abundance for the base peak m/z = 178 and m/z = 179 for authentic N-(ethoxycarbonyl)- -phenylalanine methyl ester is identical to the cal culated ratio of the base peak m/z = 179 and m/z = 180 of the N-(ethoxycarbonyl) methyl ester of the biosynthesize d [15 N]-( R)- -phenylalanine made after the incubation of Tc PAM with [ 15 N]-( S)- -phenylalanine 40 and [ring, -C- 2H6]- trans -cinnamate acid. This suggested a 15 N enrichment of 98% and that no unlabeled ( R)- -phenylalanine derivative was present. A small perc entage (3%) of the biosynthetically derived ( R)- -phenylalanine contained seven additional mass unit s, according to a base peak fragment ion m/z = 185; the [ 15 N]-( R)- -phenylalanine derivative was present at 97%. The isotopomer containing seven additional mas s units was derived from the [ring, -C- 2H6]-trans -cinnamate acid in the mixed substrate assay with Tc PAM and was not present in control assays where the labeled cinnamate acid was left out. 2.3.4. X-ray Crystal Structure of Pa PAM Pa PAM has similar architecture to the class of MIO-de pendent aminomutases and ammonia lyases. 38 , 50 , 69 -71 The monomer consists of mostly helices that run parallel to one another and form a four helix bundle at the center. The catalyt ically relevant species is a dimer of dimers, in which the two monomers in the asymmetric unit are r elated by a crystallographic twofold axis to the other two monomers that comprise the catalytica lly functional Pa PAM tetramer, and each subunit contains an active site. 64 At the end of this bundle is the active site, whic h resides at the interfaces between three of the monomers in the tet ramer, and includes residues from all three protomers. Pa PAM, similar to other bacterially derived members o f this family whose structures are known, lacks the C-terminal capping domain that is present in the related plant enzymes, such as Tc PAM. 38 The inner-loop region, which rests just above the active site, is packed tighter towards MIO than that of Tc PAM, and is well ordered. However, the Pa PAM MIO is made autocatalytically from an amino acid tandem T-S-G 64 instead of from the common A-S-G sequence. 41 a b Figure 2.3 . a) Electron density (2F oF c map, blue mesh) calculated at 1.0 around the - and -phenylpropanoid that is covalently bound to the MIO found in monomer fiBfl (C and C are indicated). Atoms are color-coded as C (green), O ( red), N (blue). b) Active site of Pa PAM in complex with a MIO-bound ( S)- -phenylalanine-type ligand (orange carbon atoms). A ctive site residues contributed by three monomers are colored accordingly (C: cyan, yellow, or green for each monomer; oxygen: red; nitrogen: blue). 23 The active sites of the two monomeric structures in the asymmetric unit differ in what is bound to them. In monomer fiAfl, a molecule that is i ndistinguishable (by electron density) from (R)- -phenylalanine is found covalently attached by its amino group to the methylidene carbon atom of MIO. This complex is consistent with an enz yme-bound ( R)- -phenylalanine-type 42 intermediate along the conjugate amino-addition pat hway (cf. Scheme 1.3a ) In contrast, the electron density of monomer fiBfl suggests partial oc cupancy of two ligand types. The electron density for a ( S)- -phenylalanine-type complex is evident, as seen in monomer fiAfl. However, electron density shows a structural feature consist ent with a ( S)- -phenylalanine molecule covalently attached by its amino group to the methy lidene carbon atom of MIO ( Figure 2.3a ). 64 These complexes are, to our knowledge, the first st ructural identification of naturally occurring pathway intermediates from an unmodified MIO-depend ent enzyme, and provide additional evidence that reactions catalyzed by Pa PAM do not proceed through a FriedelŒCrafts pathway , as recently proposed. 60 2.3.5. Distribution of Products in the Catalysis of 2 -Methyl-( S)- -Phenylalanine by Tc PAM and Pa PAM Evidence to support the proposed pathways for the Pa PAM and Tc PAM reactions was provided by incubating 2 -methyl-( S)- -phenylalanine (1 mM) at 31 °C for 1 h separately w ith each enzyme. The distribution of 2 -methyl-( S)- -phenylalanine and trans -2 -methylcinnamate made from 2 -methyl-( S)- -phenylalanine by Pa PAM catalysis was 98:2, while a reciprocal distribution (~1:99) was observed for Tc PAM catalysis. Comparison of the kinetic parameters for Pa PAM ( kcat = 0.061 s Œ1 , -amino acid production; KM = 0.05 mM) and Tc PAM ( kcat = 0.002 sŒ1 , cinnamate production; KM = 0.01 mM) with 2 -methyl-( S)- -phenylalanine showed that the catalytic efficiency ( kcat /KM = 1.2 s Œ1 mM Œ1 ) of Pa PAM was 6-fold greater than the efficiency of Tc PAM, due largely to the superior kcat . 43 2.3.6. Formation of - and -Phenylalanines from Cinnamate and NH 3 by Pa PAM Presumably, the - and -phenylalanines were formed by the reaction of a ci nnamate ion with the covalent aminoŒMIO adduct, which was made when the recombinant protein was expressed in E. coli and retained during protein purification. Notably, the Luria-Bertani 76 media (pH 7.3) that was used to grow the bacteria for this crystal lography study was estimated to contain ammonium salts at a concentration of 2.4 m M, as assessed in an earlier study. 76 Therefore, LB media was a practical source of ammonia that could bind to the MIO of Pa PAM that was overexpressed in E. coli . To assess whether Pa PAM could transfer the amino group to the cinnamate ions that were added to the crystallizati on buffer, in a manner similar to that of other MIO-dependent enzymes, 77 Pa PAM (7 mg in 1 mL assay), cinnamic acid (1.25 m M), and ammonia (60 M) were incubated together for 48 h at 31 °C. Analysis of the products showed that ( S)- -phenylalanine (6.6 ± 0.9 nmol) and ( R)- -phenylalanine (8.7 ± 0.9 nmol) were produced. These results indicate that a concentrati on as low as 60 M of ammonia in the bulk media was sufficient to aminate the MIO of Pa PAM and to catalyze the reverse reaction. Similarly, MIO adducts of either amino 56 , 70 or hydroxy 50 groups of MIO were detected in previous crystallographic studies of enzymes in thi s class. 2.4. Discussion Structures have been determined for several members of this lyase class I-like family, including Pp HAL, 50 Rs TAL, 70 PALs from Anabaena variabilis ,78 Nostoc punctiforme ,78 Rhodosporidium toruloides ,56 and Petroselinum crispum ,71 respectively, and Sg TAM from Streptomyces globisporus .69 All have very similar overall folds, contain MIOs imbedded in an 44 active site of similar overall architecture, and ex ist as homotetramers. In most structures, two of the subunits significantly contribute active site r esidues, and a third subunit encapsulates one side of the active site, while in Sg TAM, this third subunit contributes a Tyr residue t hat interacts with the 4'-hydroxy group of the substrate. In recent st udies, - and/or -tyrosine mimics were covalently trapped in the active sites of Sg TAM, 63 and Rs TAL, 70 and their structures were determined, demonstrating that the amino group of t he substrate attacks the MIO of this aminomutase. The Sg TAM structures bound to the substrate or product mi mics show that a nucleophile at either C or C will attack the MIO moiety. 57 , 63 These results have shed considerable light on the mechanism of the MIO-base d catalysts, countering previous work inferring that the MIO couples as an electrophile t o the aromatic ring of the substrate to activate catalysis. 79 In earlier studies, the stereochemical evaluation o f the ammonia lyase reactions included an assessment of the stereoselectivity, showing univer sally that the trans -isomer of the acrylate product is made exclusively upon elimination of amm onia 80 and that one prochiral hydrogen at C is stereospecifically removed from the substrate d uring the process. 81 , 82 Sg TAM and Pa PAM of bacterial origin, and Tc PAM of plant origin are the only aminomutases in th e class I ammonia lyase-like family whose structures are known, and t he reaction stereochemistry of the aminomutases have been evaluated. The Sg TAM and Pa PAM reactions were found to make ( S)- -amino acids; the mode of transfer for Sg TAM (inversion or retention of configuration and intra- or intermolecular group transfer) was not ev aluated. 57 , 69 However, evaluation of the cryptic components of the Tc PAM and Pa PAM reactions has shown that the amino group and hydrogen are removed from the substrate, their posi tions exchanged, and then both are reattached 45 intramolecularly to the original carbon skeleton wi th retention of configuration in Tc PAM 66 and with inversion of configuration in Pa PAM. 61 The results of the mixed substrate assay containin g [ring, -C- 2H6]-trans -cinnamate and [ 15 N]-( S)- -phenylalanine indicate that the nitrogen migrates quantitatively from the - to the -carbon with negligible exchange to exogenously supplied cinnamate molecules at a concentration (10 mM) likely much higher than physiological levels. Tc PAM maintains high fidelity with the natural phenyl alanine substrate through tight binding of the substrate and reaction intermediate, at the exclusion of competitive substrates, in order to achieve the observed reaction efficiency. Despite the wealth of structural data collected on this family of enzymes with overall very similar active site architecture, none sheds any li ght to explain how the stereochemistry of the Tc PAM reaction proceeds with retention of configurati on and makes the (3 R)- -phenylalanine isomer, which is opposite to the (3 S)-stereochemistry found in the product of the Sg TAM and Pa PAM reaction. Therefore, a favorable binding confor mation of the trans -cinnamate substrate in the Tc PAM-cinnamate complex ( Figure 2.1a ) and Pa PAM- /-amino acid complex ( Figure 2.3a ) were used as basis to understand how these aminom utases control product stereoselectivity. The aminomutase reaction is proposed to be made fro m an arylalanine via a concerted Hoffman- like elimination reaction during the reaction cycle . 45 The catalytic base Tyr is situated above the CC bond of the bound phenylalanine substrate (in a st aggered conformation), with the pro -(3 S) hydrogen closest to Tyr ( Figure 2.3b ); the MIO moiety is positioned below the C C bond of the substrate, with the amino group proximate to the methylidene of the MIO. The formation of the transoid acrylate product supports the stere oselective removal of the pro -(3 S) proton by an anti elimination. 37 Furthermore, evaluation of the - and -phenylalanine complexes of monomer 46 fiBfl ( Figure 2.3a ) suggests that during the course of the amino grou p isomerization catalyzed by Pa PAM, the phenylpropanoid carbon backbone remains mo stly stationary above the amino group that is attached to the MIO moiety. These con figurations are consistent with the mechanism of stereoselectivity for this enzyme that proceeds with inversion of configuration at each migration terminus. 61 However, for the production of ( R)- -phenylalanine, the mechanism of catalysis must adh ere to an intramolecular amino group transfer, where th e amino group does not exchange with that from another substrate molecule. In addition, the s tereochemistry must account for the retention of configuration at the migration termini. 66 These data suggest that the amino group and hydrogen displaced from the (2 S)- -phenylalanine substrate must rebind, respectively, at C and C of the trans -cinnamate before it diffuses from the active site. The constraints of the intramolecularity and stereochemistry of the Tc PAM reaction present a challenge because the amino group must attach to the cinnamate on the sid e facing opposite to that of the NH 2-MIO intermediate (cf. Figure 2.1a ) to produce the ( R)-product. Since this stereochemistry is opposite to the ( S)- -phenylalanine product made by Pa PAM, the covalent intermediate derived by reacting Pa PAM with a product analog having the configuration at C as the natural ( S)- -product is not a valid representation of the Tc PAM reaction stereochemistry. The Pa PAM active site contains key catalytic residues tha t are found in other structurally characterized enzymes in the class I lyase-like fam ily ( Figure 2.3b ). Tyr78 (Tyr80 in Tc PAM and Tyr63 in Sg TAM) is positioned above and within 3.5 Å of both t he - and -carbon atoms of the phenylalanine complexes, and is poised to de- a nd re-protonate the intermediate phenylalanine complexes at both the and positions. Tyr320 (Tyr322 in Tc PAM and Tyr308 47 in Sg TAM) is only 2.6 Å from the amino group of the MIO- bound ( S)- -phenylalanine, and is thought to facilitate proton transfers when the ami no group is removed from the ( S)- -phenylalanine substrate and added to MIO. In contra st, differences in the residues near the active site of Pa PAM are notable when compared with those of other M IO-dependent catalysts ( Figure 2.4 ). For example, Phe455 (Asn458 in Tc PAM) is a glutamine or an asparagine in almost all other enzymes in the family. When the Tc PAMŒcinnamate structure 38 is overlaid onto the Pa PAMŒphenylalanine structure, the Tc PAMŒcinnamate ligand clashes with the sterically bulkier Phe455. In the Pa PAM structure, however, the covalently-bound - and -phenylalanine ligands avoid the steric clash with Phe455 by alter ing their trajectory through the active site, which avoids the collision (( Figure 1.5a ). The new trajectory results in a weaker mono dent ate salt bridge with Arg323 (versus the bi dentate salt bridge that is present in Tc PAM) and is almost identical to the trajectory of tyrosine in the Sg TAM structure. This similar alignment of ligands is reinforced by the superimposition of the bridging a toms (likely nitrogen) of the - and -phenylpropanoidŒMIO adducts in Pa PAM and the nitrogen and oxygen atoms, respectively , of the -difluoro- -tyrosine and 2,3-dihydroxycoumarate inhibitorŒMIO adducts in the two Sg TAM structures ( Figure 1.5b ). 57 , 63 48 PaPAM SNGLPPFLCAE-NAGLRLGLMGGQ FMTASITAESRASCMPMSIQSLS 450 *V.bact SNGLPAFLCAE-NAGLRLGLMGGQ FMTASITAESRASCMPMSIQSLS 450 *S.marit SNGLPAFLCRE-DPGLRLGLMGGQ FMTASITAETRTLTIPMSVQSLT 426 *B.subtl SNGLPAFLTKE-NPGLRLGLMGGQ FMSTSLTAENRSLCTPLSIQTLT 432 *K.pneu SNGLPSFLCAE-NGGLRFGLMGGQ FMSSSVTAENRSLATPVSIQTLT 438 *B.rhiz SNGLPPFLCAN-EQGIRLGLMGGQ FMSASLASENRSLCVPVSIHSLP 429 Sg TAM SYGLPEFLVSG-DPGLHSGFAGAQYPATALVAENRTIG-PASTQSV P 438 Av PAL SNGLPPSLLGNRERKVNMGLKGLQICGNSIMPLLTFYGNSIADRFP T 445 Tc PAM SNGLPGNLSLGPDLSVDYGLKGLDIAMAAYSSELQYLANPVTTHVH S 453 Figure 2.4 . CLUSTAL 2.1 multiple sequence alignment 38 of representative class I lyase-like PAL, PAM and TAM enzymes whose structures are known and five sequences highly homologous to Pa PAM that were found in the GenBank database by BLAS T search: Pa PAM, PAM from Pantoea agglomerans ; V.bact , possible PAM from Vibrionales bacterium SWAT-3 ; S.marit , PAL/PAM EncP from Streptomyces maritimus ; B.subtl , possible PAM from Bacillus Subtlis ; K.pneu , possible PAM from Klebsiella pneumoniae 342 ; B.rhiz , possible PAM from Burkholderia rhizoxinica ; Sg TAM, TAM from Streptomyces globisporus ; Av PAL Phenylalanine ammonia lyase 39 from Anabaena variabilis ; Tc PAM, PAM from Taxus Canadensis . Enzymes that have a concerved phenylalanine residue inside the active which correspond to F455 residue in Pa PAM, produce ( S)- -phenylalanine (designated with asterisk (*)). Sg TAM, however, contains a nonconserved His93 (Val108 in Pa PAM) that forms a hydrogen bond with the hydroxy group of the tyrosine substra te. 54 This hydrogen-bonding interaction, which is absent in Pa PAM, enforces the trajectory of tyrosine in the Sg TAM active site. Thus, Sg TAM and Pa PAM catalyze equivalent stereochemistries, presumab ly by orienting their 49 substrates identically in their active sites, but u se a distinct set of enzymeŒsubstrate interactions to accomplish this. Pa PAM mutants F455A and F455N (F455N is analogous to N458 of Tc PAM) each form the same products, but at approximat ely 2 % of the rate of Pa PAM, in a 40:60 ratio, where trans -cinnamate prevails. These data are consistent with the hypothesis that Phe455 is important for the proper trajectory of th e substrate in the active site for transfer of the amino group from the to the position. 64 Intrigued by the possibility that other class I lya se-like enzymes that catalyze the formation of the same ( S)- -phenylalanine produced by Pa PAM may exist, a BLAST 85 search was performed, and five other MIO-dependent enzymes, each from a d istinct organism, were found to have a Phe residue that is equivalent to Phe 455 ( Figure 2.4 ). In each case, all of the other residues in the vicinity of the active site were also identical to those of Pa PAM. Of these enzymes, EncP from Streptomyces maritimus , was characterized as a slow phenylalanine ammonia lyase ( kcat =0.0061 s1 ) that is required for the biosynthesis of the anti biotic enterocin in this organism. 78 , 86 , 87 Subsequent studies, however, showed that this enzym e has much higher phenylalanine aminomutase activity (with ( S)- -phenylalanine as the product) than lyase activity below 50 °C. 88 Furthermore, another organism ( Vibrionales bacterium SWAT-3 ) identified by the BLAST search produces andrimid, which contains an amino acid moi ety that is derived from ( S)- -phenylalanine. 89 50 a b c d Figure 2.5 . Cinnamate diastereoisomers modeled into Tc PAM. a) A space-filled rendering and b) a skeletal structure of cis -cinnamate are modeled into Tc PAM by preserving the tight salt bridge with Arg325. The phenyl ring of the cinnamat e is oriented toward the interior of the active site pocket. The collision distances between active site residues and the phenyl ring of the cisoid diastereoisomer are given: Asn231 - 0.8 Å, L eu227 - 2.7 Å, Phe371 - 2.7 Å, Asn355 (main chain) - 2.7 Å. c) A space-filled rendering o f trans -cinnamate is modeled into Tc PAM by orienting the carboxylate toward the salt bridge wi th Arg325; occlusions by active site residues on the transoid structure are absent. d) Shown is a n overlay of the space-filled structures of ( cis )- and trans -cinnamates modeled in the Tc PAM active site with the Arg325 salt bridge preserved. 13 51 The trajectory of the ( S)- -amino acid in the active site of Tc PAM is determined by the direction of the carboxylate group when it forms a salt bridge with the -guanidinium group of a proximate Arg325, as in the Tc PAM-cinnamate complex and all other bound structure s of enzymes from the lyase-like family. 50 , 56 , 69 -71 , 78 In previous investigations, assessing the stereochemical course of the Tc PAM reaction in the absence of structural data, it was suggested that the differences in the stereochemistry of the Tc PAM reaction compared to that of the Pa PAM reaction resulted from fundamental differences in the active sites of these isomerases. The substrate was hypothesized to bind the Tc PAM active site with the carboxylate and phenyl ring in a syn -periplanar trajectory to position the migrating H and NH 2 groups of the ( S)- phenylalanine on the same side of the molecule to a ccount for the retention of configuration at the reaction termini. 66 Shown herein, however, the active site of Tc PAM is arranged similar to the other MIO-dependent enzymes, where the catalyti c base and MIO (amino group acceptor) are antipodal with respect to the substrate; thus, the migration of the H and NH 2 groups must also occur on opposite sides. Consequently, a syn -periplanar trajectory or cis -cinnamate configuration of a reactive intermediate would confound the obser ved retention of stereochemical configuration of the Tc PAM reaction. More importantly, a cisoid phenylprop anoid (either as S-cis -phenylalanine or cis -cinnamate) in a modeled complex with Tc PAM would be sterically occluded by several active site residues ( Figure 2.5 ). Further, electron density consistent with a trans -cinnamate is clearly observed in the Tc PAM structure and provides good evidence that the reaction proceeds via a transoid intermediate. 38 52 2.4.1. Stereoselectivity of the Aminomutase Reaction In the Pa PAM and Sg TAM reactions, the amino group can migrate from C to C across the same face ( Si face) of the arylacrylic acid intermediate to isom erize ( S)- - to ( S)- -arylalanine (Scheme 2.1 ). In contrast, for the Tc PAM reaction, and possibly for the Cc TAM reaction the C -amino group must be removed from the Si -face (using C of the arylacrylic acid intermediate as reference), then reattach at C on the Re face. A possible explanation involves approximatel y 180° rotations of the already-bound arylacrylate in termediate about both the C 1C and C C ipso bonds prior to rebound of the amino group to C by the NH 2-MIO, without breaking the salt bridge to Arg-325, and with minor displacement of t he aromatic ring from its original position (Scheme 2.1 ). This rotamer positions C for nucleophilic attack by the NH 2-MIO moiety to form the enzyme/product covalent intermediate. Catalytic Tyr residue is in position on the opposite side to reprotonate the phenylpropanoid at C , resulting in overall retention of configuration i n the product. The nitrogen linking the MIO to the pr oduct is likely protonated by Tyr-322 that initiates the departure of the product from the MIO . Concomitant rotation about the C 1C and CC ipso bonds does not result in any steric clashes in the active site, nor in breaking direct interactions between enzyme and substrate. The fore going proposed mechanism is consistent with all of the stereochemical and mechanistic find ings for Tc PAM. In addition, the Keq = 1 for the Tc PAM reaction 66 suggests that the rotamers are energetically equiv alent. 53 Scheme 2.1: The Proposed Mechanism of the Aminomuta se Catalysis. Tyr is the Presumed Catalytic Base, Situated fiAbovefl the Intermediate, and the NH 2-MIO Complex is Shown fiBelowfl the Intermediate. In Sg TAM and Pa PAM, After the Initial E 2-Type Elimination to Form the Intermediate, the Amino Group is Reboun d at the C Position from the Si Face (Same Face), of the Intermediate while in Cc TAM and Tc PAM, the Intermediate Under Go 180º Rotation About the C 1C and C C ipso Bonds in order for the Amino Group to Attack at the C Position from the Re Face (Opposite Face) of the I ntermediate. 54 To further confirm the above phenomena the product distribution of the catalysis of 2'- methyl-( S)- -phenylalanine by Tc PAM and Pa PAM was analyzed. The 2'-methyl substituent at ortho position of the cinnamic acid likely affected the r otation of the intermediate to access Re face, and thus only trans -2 -methylcinnamate was made by Tc PAM. In contrast, Pa PAM presumably utilized a single 2 -methylcinnamate rotamer ( Si face) to make 2 -methyl-( S)- -phenylalanine without encountering the torsional ba rrier 61 . The difference in product distribution between the two enzymes supports a model consistent with steric and torsional strain between the 2 -methyl substituent and the C -hydrogen of the intermediary trans -2 -methylcinnamate in the Tc PAM reaction. Notably, alternative or additive ster ic interactions between the 2 -methyl substrate and active-site residues can also prevent interchange between s i and re face and abort the Tc PAM reaction ( Scheme 2.1 ). Conversely, it can be imagined that the single r otamer si face on the Pa PAM pathway need not encounter the same active-site interactions to proceed from 2 -methyl-( S)- - to 2 -methyl-( S)- -phenylalanine. The underlying mechanism responsibl e for the proposed rotational dynamics of Pa PAM and Tc PAM is not fully understood and is intriguing since the positions of most of the catalytic amino acids, the presumed H-bonding residues, and van der Waals interactions in each active site are conserved. 2.4.2. Effects of L104A Point Mutation on Enzymatic Activi ty of Tc PAM In this study, the effect of a targeted point mutat ion on substrate selectivity and enzymatic reactivity were investigated for the Taxus phenylalanine aminomutase. The structure of Tc PAM revealed that Leu-104 makes a direct hydrophobic in teraction with the aromatic ring (nearest the 3-carbon) of the presumed reaction intermediate cinn amate. Mutation of this Leu-104 to a sterically smaller alanine residue was proposed to increase the active site volume, reduce a steric 55 interaction between the substrate and the active si te, and thus increase the catalytic efficiency of the catalyst for arylalanine substrates bearing a s ubstituent on the 3 -carbon of the ring. For proof of principle, sterically demanding substrates 3 -methyl-( S)- -phenyl-, 4 -methyl-( S)- -phenyl-, and styryl-( S)- -alanine were chosen to evaluate the effects of the L104A mutation in Tc PAM (PAMeLA_104). In a previous investigation, Tc PAM was found to generally isomerize -arylalanines to their corresponding -arylalanines; however, the catalytic efficiency of Tc PAM decreased markedly with increasing steric bulk on the -amino acid substrate. 39 Thus, the steric hindrance of the active site seemingly limited its catalytic efficie ncy. Kinetic parameters of the PAMeLA_104 enzyme with non-natural aryl amino acids were compa red to the Tc PAM catalyst, demonstrating that Leu-104 has significant influence on substrate binding (approximated by KM), product distribution, and kcat . Notably, the KM of PAMeLA_104 increases over 2-fold with the phenylalanine substrate ( 1) compared to the KM of Tc PAM, suggesting that Arg-325 and Leu- 104 participate in substrate docking via a salt bri dge and hydrophobic interaction, respectively. PAMeLA_104 released total product (( R)- -phenylalanine and trans -cinnamate) at a rate ( kcat = 0.076 s 1 ) comparable to that of Tc PAM ( kcat = 0.065 s 1 ) but produced ( ~ 6-fold) more cinnamate than -phenylalanine (cf. Table 2.1). Likely, an accessib le trajectory of phenylalanine in the mutant causes the reaction to stall after th e first reaction step and release the acrylate product. Apparently, this trajectory is not achieva ble in Tc PAM, which preferentially makes -phenylalanine. In contrast, the KM of PAMeLA_104 was nearly 5-fold lower with 3 -methyl-( S)- -phenylalanine ( 2) compared to the KM of Tc PAM, suggesting that the 3 -methyl substituent of the substrate enhanced substrate binding in the mut ant. While the Leu Ala exchange likely 56 reduced unfavorable steric strain (cf Figure 2.2a ), the smaller alanine residue, however, could still make a constructive hydrophobic interaction w ith the 3 -methyl group of 2 (cf. Figure 2.2b ). Remarkably, the catalytic efficiency of PAMeLA_104 for 2 increased ~ 7-fold compared to Tc PAM catalysis, which is primarily a reflection of t he considerable (5-fold) reduction in KM (i.e., better binding) of PAMeLA_104 compared to th at of Tc PAM. While Tc PAM and PAMeLA_104 displayed dramatic differences i n their kinetic parameters with -phenylalanine and 3 -methyl-( S)- -phenylalanine, they did not display significant differences with 4 -methyl- -phenylalanine ( 3). The KM of Tc PAM with substrate 3 was 1.6-fold higher than that for Tc PAM with its natural substrate 1, suggesting that the 4 -methyl group of the substrate only modestly affected binding. Inter estingly, the KM of PAMeLA_104 for 3 was slightly lower than that of Tc PAM, suggesting that the Leu Ala replacement likely enabled the 4-methylphenyl functional group to adopt a suitable conformation to interact favorably with other distal hydrophobic residues (Leu-179, Leu-227 , and Val-230) of the active site pocket (cf. Figure 2.2d ). Moreover, the structural data for Tc PAM shows no active site residues proximate to the 4 -carbon of the natural substrate that would interfe re sterically with a 4 -alkyl substituent of 3. Tc PAM converted styryl- -alanine ( 4) almost exclusively to (2 E,4 E)-styrylacrylate, while PAMeLA_104 made the corresponding -amino acid from 4, albeit slowly, at about 3% of the styrylacrylate production rate at 0.12 s 1 . The wild-type and mutant enzymes made the styrylacrylate product from 4 faster than they were able to convert any of the o ther aryl amino acid substrates to their corresponding -amino acids and acrylate products combined (cf. Table 2.1 ). These data suggest that the ammonia lyase functi on of both enzymes remains largely efficient with 4. The stability of the conjugated -bonds and the larger steric volume of the product, likely affected the rapid release of the i ntermediate, reducing the residence time needed 57 to isomerize the - to -amino acid. PAMeLA_104 noticeably made more styryl -( R)- -alanine than does Tc PAM under steady-state conditions, indicating that the mutation likely increased the residence time of the substrate and ensuing interme diate in the active site. The 2-fold lower KM of PAMeLA_104 compared to that of Tc PAM for substrate 4 demonstrates that the mutant can accommodate 4 better. (cf. Figure 2.2c ). Clearly, the hydrophobic region of the Tc PAM active site surrounding the aromatic ring of the purported reaction intermediate (cinnamate) plays a major, but mechanistically unknown, role in substra te selectivity and in governing the distribution of the intermediate acrylic acid that is released and is converted to the -amino acid. 2.5. Conclusion The structure of Pa PAM was solved as - and -phenylpropanoid adducts, presumably with (S)- - and ( S)- -phenylalanine. These intermediates provide strong evidence that Pa PAM reacts by an alkylamine elimination pathway (such as a Hof fmann-type or E2-type elimination process), which involves covalent attachment betwee n the amino group of the substrate and the product as well as the MIO cofactor, as demonstrate d previously for Sg TAM. 63 The results indicate that the carbon skeleton of the ( S)-phenylalanine substrate remains in one rotameric conformation ( Si face) while the exocyclic CŒN bond of the NH 2ŒMIO adduct rotates into position below the - and -carbon atoms to complete the isomerization reactio n ( Figure 2.3a ). Thus, the structure also confirms the inversion of configuration at each migration terminus during the isomerization of the -amino acid substrate into its -isomer. 61 The structure of the phenylalanine aminomutase on the Taxol biosynthetic pathway 90 , 91 has been presented. The Tc PAM active site was observed to be arranged similar to that of other members of the MIO- dependent family of enzymes. The ( S)-product stereochemistry catalyzed by the bacteria l 58 Sg TAM and Pa PAM 4, 6 is opposite to the ( R)-product stereochemistry catalyzed by Tc PAM of plant origin and Cc TAM of bacterial origin. Conceptually, the stereoch emistry of the Tc PAM and Cc TAM reaction can be achieved by rotation of the int ermediate cinnamate in the active site by approximately 180° about the C 1C /C ipso C bonds prior to rebinding of the amino group at the -position on the trans -cinnamate intermediate. Comparing the active sites of Tc PAM, Pa PAM and Sg TAM showed subtle structural differences that may a ccount for the significant changes in the trajectory of the substrate, possibl y causing stereodifferentiation. On the basis of the Tc PAM crystal structure complex, the PAMeLA_104 mutan t was constructed and demonstrated superior catalytic eff iciencies for substrates 3 -methyl-( S)- -phenylalanine and styryl- -alanine possessing larger molecular steric volume. The L104A mutation likely reduced unfavorable steric clash th at conceivably created an altered alignment of the substrate and/or the ensuing acrylate intermedi ate within the active site that changed the kinetic parameters of PAMeLA_104 compared to the wi ld-type Tc PAM. 59 3. (S)- -STYRYLALANINE USED TO PROBE THE INTERMOLECULAR MECHANISM OF AN INTRAMOLECULAR MIO-AMINOMUTASE Reproduced with permission from [Wanninayake, U.; D ePorre, Y.; Ondari, M.; Walker, K. D., ( S)-Styryl-a-alanine used to probe the intermolecular mechanism of an intramolecular MIO- aminomutase. Biochemistry 2011, 50, (46), 10082-10090] Copyright © 2011 American Chemical Society 3.1. Introduction A Taxus canadensis phenylalanine aminomutase ( Tc PAM) requires no cofactors to intramolecularly transfer the amino group of ( S)- -phenylalanine to form ( R)- -phenylalanine, utilizing the same carbon skeleton. 38 This aminomutase activates the amino group of the substrate as an alkylammonium leaving group via ligation with a 3,5-dihydro-5-methylidene-4 H-imidazol-4-one (MIO) prosthetic group, formed by au tocatalytic post-translational condensation of active site residues (Ala-Ser-Gly). 69 This MIO purportedly serves as an electrophilic si nk (a 1,4-Michael acceptor) and is nucleophilically attac ked by the amino group of the substrate to form the alkylammonium complex to facilitate Hofman n-like elimination. 69 The pro -3 S hydrogen and an alkylamine are transiently eliminat ed to form a cinnamate intermediate and then rebound to interchanged positions on the phenylprop enoid skeleton. Under steady-state reaction conditions, Tc PAM catalyzes the production of a significant propo rtion of ( R)- -phenylalanine compared to trans -cinnamate, the latter is ordinarily made between 1 0 and 20 times slower. 38 60 A recent study showed that when Tc PAM was co-incubated with both [ 15 N]-( S)- -phenylalanine ( 1) and [ring,3- 2H6]- trans -cinnamate ( 2) (each at 10 mM), the [ 15 N]-amino group incorporated (97%) intramolecularly into ( R)- -phenylalanine. However, slight intermolecular amino group transfer occurred, as well. Analysis of the isolated -phenylalanine revealed a crossover reaction, in which the amino group of 1 was incorporated partially into precursor 2 to form [ 15 N, 2H6]- -phenylalanine ( 5) in ~ 3% yield (Scheme 3.1). 38 Thus, the unlabeled cinnamate intermediate 3 derived from [ 15 N]- -phenylalanine did not appreciably dissociate from the active site or exchange with exogenous 2H-labeled cinnamate (only a 3% occurrence), before the amino group of the 15 NH 2ŒMIO complex rebounded. A few labeled cinnamate mol ecules did, however, competitively bind the active site, accounting for the dearth of doubly labeled -amino acid product observed. The foregoing intermolecular data are supported by an earlier study demonstrating that a tyrosine aminomutase (TAM) transferred the amino gr oup from 3 -chlorotyrosine to 4- hydroxycinnamate to form 3 -chloro-4 -hydroxycinnamate and a mixture of - and - -tyrosine. 62 The results of the latter intermolecular process we re paradoxically used to explain that the TAM reaction proceeds intramolecularly. This argument w as further thought to be supported by the faster transfer of the amino group in the cross rea ction compared to the rate when exogenous NH 4+ was used as the amino group source. Notably, the h igh molar concentrations of ammonia used in the catalysis of the reverse reaction likel y approached conditions typically used to precipitate and denature proteins, 92 and thus affected the catalytic efficiency of the ammonia transfer reaction. In this earlier study, it also r emained unclear whether the amino group removed 61 from -tyrosine was then transferred intramolecularly to the same 4 -hydroxycinnamate reaction intermediate en route to -tyrosine, on the natural reaction pathway. 62 Scheme 3.1: Mechanism of the Transaminase Reaction Catalyzed by Tc PAM with Its Natural Substrate (Exogenously supplied cinnamate i s color coded in red) ˘ˇˆ ˆ Moreover, a homologous phenylalanine aminomutase (P AM) isolated from Taxus chinensis , in an earlier study, was incubated with an exogenou s supply of 6 M NH 4+ salts (pH 10) to provide ammonia to the reaction. The ensuing NH 2ŒMIO complex (or other amine complex) likely formed, and the hydrogen and amino group (fr om the MIO complex) were transferred to various trans -arylacrylates to form a mixture of the correspondi ng enantiopure - and -amino 62 acids in the reverse reaction. 40, 93 This latter study was identical in technique to ano ther earlier work demonstrating that mechanistically similar MIO -dependent enzymes also catalyzed their reverse reactions. This earlier compendium of work includes a description of how a phenylalanine ammonia lyase added a hydrogen/amino group pair from 6 M NH 4+ salts (pH 10) to trans -arylacrylates to produce non-natural ( S)- -amino acids, 94 while a tyrosine aminomutase analogously added the same pair from NH 4+ salts to 4-hydroxycinnamate to form a mixture of - and -tyrosine. 62 Furthermore, while Tc PAM converts ( S)- -phenylalanine principally to ( R)- -phenylalanine at a rate of 0.053 ± 0.001 s Œ1 , it also converts ( S)-styryl- -alanine to (2 E,4 E)- styrylacrylate as the major product (99%) at approx imately the same rate (0.082 ± 0.002 s Œ1 ). 38 Consequently, the styrylacrylate is released from t he active site before the amino group can rebound appreciably to form styryl- -alanine. These kinetic data along with the propose d MIO- dependent mechanism suggest that when ( S)-styryl- -alanine is converted to (2 E,4 E)- styrylacrylate, the transient amino group remains a s the NH 2ŒMIO complex, likely for the same duration as the --phenylalanine isomerization reaction. A hypothesis emerged from the aforementioned intra- and intermolecular mechanistic evaluations 38 and observations. 40 , 62 , 94 Conceptually, a weakly binding acrylate (AC1) intermediate derived from an amino acid (AA1) in th e Tc PAM reaction was replaced in the active site with a tighter binding, competitive acr ylate (AC2). The amino group removed from AA1, but now in complex with the enzyme, could rebi nd to AC2 and form - and -amino acids (AA2) different from those derived from AA1. Thus, guided by the partial intermolecular process observed for Tc PAM 38 and by the details of the kinetic parameters descr ibed herein, we explored four non-natural amino acids as amino grou p donor substrates. The amino group of 63 these donor substrates was transferred intermolecul arly to another arylacrylate skeleton by Tc PAM catalysis to form an - and -arylalanine mixture. 3.2. Experimental 3.2.1. Chemicals (R)- -Phenylalanine, 3 -methyl-( S)- -, 3 -methyl-( R)- -, 4 -methyl-( S)- -, 4 -methyl-( R)- -, 4-fluoro-( S)- -, 4 -fluoro-( R)- -, and 3 -fluoro-( R)- -phenylalanine and ( S)-styryl- -, ( S)-2 -thienyl- -, ( R)-2 -thienyl- -, and ( S)-2 -furyl- -alanine were purchased from Peptech Inc. (Burlington, MA). (1 S)-Camphanoyl chloride, ( S)- -phenylalanine, ( S)-2-aminotetralin-2- carboxylic acid, (3 R)-3-aminotetralin-(2 R)-2-carboxylic acid, -4 -chloro- -phenylalanine methyl ester, -4 -chloro- -phenylalanine, trans -2 -methyl- and trans -4 -methylcinnamate, trans -2 -furyl- and trans- 2-thienylacrylate, and trans -cinnamic acid were purchased from Sigma-Aldrich-Fl uka (St. Louis, MO). trans -4 -Chloro-, trans -3 -methyl-, and trans- 4-fluorocinnamate and (2 E,4 E)- styrylacrylate were acquired from Alfa Aesar (Ward Hill, MA). All chemicals were used without further purification, unless noted. 3.2.2. Instrumentation The GC oven (Agilent, model 6890N) conditions were as follows. The column temperature was programmed from 70 to 250 °C at a rate of 10 °C /min (18 min total run time); splitless injection was selected, and helium was used as the carrier gas (1.2 mL/min). The GC oven was coupled to a mass selective detector (Agilent, mode l 5973 inert ) in ion scan mode from 100 to 300 atomic mass units at a 70 eV ionization voltage . 64 3.2.3. Expression of the tcpam and Purification of Tc PAM Codon-optimized tcpam cDNA was previously ligated into expression vector pET28a(+), and the recombinant plasmid encoded an N-terminal His 6 tag. 74 The tcpam clone was overexpressed in six 1 L cultures of Escherichia coli BL21(DE3) cells by induction with isopropyl -D-thiogalactopyranoside. The overproduced protein was isolated from the bacteria and purified to 95% by Ni affinity chromatography to yield 5 mg of protein, as described previously. 74 Routine assays for assessing enzyme function were conducted with ( S)- -phenylalanine at saturation (1 mM) and Tc PAM (100 g, 1.3 nmol) in 50 mM phosphate buffer (pH 8.5) in 1 mL assays. 3.2.4. Identification of an Amine Donor Substrate (S)-Styryl- -alanine ( 6), 2 -furyl-( S)- -alanine ( 15 ), (3 R)-aminotetralin-(2 R)-carboxylic acid (16 ), and ( S)-2-aminotetralin-2-carboxylic acid ( 17 ) (each at 1 mM) were incubated separately with Tc PAM (250 g, 3.3 nmol) and 3 -methylcinnamate ( 14 ) (1 mM) in 5 mL assays at 31 °C. To calculate the initial steady-state rates, aliquo ts (1 mL) were withdrawn at 10, 20, 30, and 40 min from the reaction mixture containing 6, at 15, 30, 45, and 60 min from the reaction mixtu re containing 15 , and at 1, 2, 3, and 4 h from the reaction mixture containing 16 or 17 . An internal standard ( R)-3 -fluoro- -phenylalanine (20 nmol) was added; the reactions w ere quenched by increasing the pH to 10 (6 N NaOH), and the amino a cids were immediately N-carbonylated by adding ethyl chloroformate (100 L). After 10 min, the reaction mixtures were acidif ied to pH 2 (6 N HCl), the derivatives were extracted with diet hyl ether (2 × 1 mL), and the solvent was removed in vacuo. To the remaining residue was adde d a (trimethylsilyl) diazomethane solution dissolved in an ethyl acetate/methanol mixture (3:1 , v/v) (methanol was used to liberate the 65 diazomethane), until the yellow color of diazometha ne persisted, to convert the N-acyl amino acids to their methyl esters. 3.2.5. Assessing the Optimal Concentration of Amino Group Donor 6 After compound 6 had been identified as an amino group donor, 6 was incubated at 400, 600, 800, 1000, 2000, and 3000 M in the presence of Tc PAM (50 g, 0.7 nmol) in 1 mL reaction mixtures for 30 min at 31 °C. An aliquot (1 mL) was withdrawn from the reaction mixture and added to a 1.5 mL polystyrene cuvette (General Labo ratory Supply, Pasadena, TX) and analyzed by UVŒvisible spectroscopy (Beckmann DU 640, Beckma nn Coulter, Brea, CA) with A305 monitoring of the sample to quantify the product (2 E,4 E)-styrylacrylate ( 7). The absorbance values obtained from the samples were compared agai nst those of a concentration series (ranging from 0.3 to 80 M) made from authentic (2 E,4 E)-styrylacrylate dissolved in 50 mM phosphate buffer (pH 8.5), analyzed by the same method. A sam ple blank of 6 ( max = 275 nm; A305 = 0.007 at 0.1 mM) was used to subtract the background abso rbance. A nonlinear regression plot of the steady-state production rate of 7 versus the concentration of 6 was used to calculate the maximal steady-state velocity ( Vmax ) and to assess which concentrations of 6 were at Vmax . 3.2.6. Calculation of the Inhibition Constants of Various Acrylates in the Tc PAM Reaction Kinetic parameters and inhibition constants for var ious substrates and inhibitors in the Tc PAM reaction were acquired by first establishing th e linearity of enzyme activity with respect to both time and protein concentration for the subs trate without inhibitor. (2 E,4 E)-Styrylacrylate (at 50, 100, and 200 M) (inhibitor) was incubated with Tc PAM (30 g, 0.4 nmol) containing 4 -methyl-( S)- -phenylalanine as the substrate at 10, 20, 40, 80, 150, 300, 500, 750, and 1000 M 66 (in separate dilution series) in 50 mM phosphate bu ffer (pH 8.5, 1 mL) for 30 min. To each reaction mixture were added 3 -fluoro-( R)- -phenylalanine and trans -2 -methylcinnamic acid (at final concentrations of 20 M) as internal standards at 0 °C. The reaction mixt ures were basified to pH 10 (6 N NaOH). Ethyl chloroformate (100 L) was added. After 5 min, the pH of the samples was readjusted to 10. A second batch of eth yl chloroformate (100 L) was added, and the solution was mixed for 5 min. The samples were acidified to pH 2 (6 N HCl), and the arylacrylic acids and N- (ethoxycarbonyl) amino acids were extracted into di ethyl ether (2 × 1 mL). The organic fractions were combined, and the s olvent was removed in vacuo. To the remaining residue was added a (trimethylsilyl) diaz omethane solution dissolved in an ethyl acetate/methanol mixture until the yellow color of diazomethane persisted to produce the methyl esters of the N- (ethoxycarbonyl) amino acids and arylacrylates. The derivatized samples were quantified by GC/EIMS, wherein the relative abundan ces of the base peak fragment ions of the amino acid derivatives present in the samples were compared to those of authentic standards at various concentrations. From the LineweaverŒBurk pl ots of the data, the MichaelisŒMenten constants, inhibition constants, and rates were cal culated. trans -Cinnamate ( 3), trans -4 -chloro-, trans -4 -methyl-, trans- 4-fluoro-, and trans -3 -methyl- cinnamate ( 9Œ11 and 14 , respectively), and trans- 2-thienylacrylate ( 13 ) (each at 50, 100, 150, 300, 500, 1000, and 2000 M in separate dilution series) were incubated with Tc PAM (50 g, 0.7 nmol) containing 6 (1 mM) as a cosubstrate in 50 mM phosphate buffer (pH 8.5, 1 mL) for 30 min. The reaction mixtures in each series were p laced on ice and acidified to pH 2 (6 N HCl); trans -2 -methylcinnamic acid (at a final concentration of 2 0 M) was added to each as an internal standard, and the carboxylic acids were ex tracted into diethyl ether (2 × 1 mL). The organic fractions were combined, and the solvent wa s removed in vacuo. To the remaining 67 residue was added a (trimethylsilyl) diazomethane s olution dissolved in an ethyl acetate/methanol mixture until the yellow color of diazomethane persisted to produce the methyl esters of the acrylate. The derivatized samples wer e quantified by GC/EIMS, wherein the relative abundances of the base peak fragment ions of methyl styrylacrylate in the samples were compared to those of an authentic standard (see App endix) at various concentrations. From the LineweaverŒBurk plots of these data, the MichaelisŒ Menten constants and rates were calculated. 3.2.7. Time Course Assays for Intermolecular Amino Group T ransfer Tc PAM (300 g, 3.9 nmol) was incubated separately with various acrylates at 50, 100, 150, 300, 500, 1000, and 2000 M and 6 (1 mM) in 6 mL assays in 50 mM phosphate buffer (p H 8.5). Aliquots (1 mL) were withdrawn from each reaction m ixture at 0.5 h intervals over 1 h, and then at 2, 4, 8, and 12 h. Internal standards trans -2 -methylcinnamic acid and 3 -fluoro-( R)- -phenylalanine (each at 20 M) were added to each aliquot, and the reactions we re quenched by increasing the pH to 10 (6 N NaOH). The amino acids and acrylates were derivatized, as described earlier, and quantified by GC/EIMS. 3.2.8. Assessing the Absolute Stereochemistry of - and -Phenylalanine Product (8f) by the Tc PAM Transaminase Pathway trans -4 -Fluorocinnamate (500 M) was incubated with Tc PAM (200 g, 2.6 nmol) and 6 (1 mM) in 1 mL assays in 50 mM phosphate buffer (pH 8. 5) at 31 °C. After 3 h, the amino acids were derivatized to their N-[(1 S)-camphanoyl] methyl esters. The derivatized - and -amino acids were identified by GC/EIMS analysis and compa red against the retention time and mass spectrometry fragmentation of authentic N-[(1 S)-camphanoyl]-4 -fluoro-(2 S)- - and -4 -fluoro- (3 R)- -phenylalanine methyl esters (Figure A 4 & Figure A 5 in Appendix). 68 3.2.9. Assessing the Effects of Maintaining the Steady-Sta te Conversion of 6 to 7 on the Production of 8f To a solution of 6 (31 mol) in 50 mM phosphate buffer (pH 8.5, 30 mL) were added trans -3-methylcinnamic acid ( 14 ) (31 mol) and Tc PAM (1.5 mg, 20 nmol) at 31 °C. The reaction mixture was shaken slowly in a water bath. An aliqu ot (20 L) was withdrawn from the reaction flask every hour and diluted to 200 L in 50 mM phosphate buffer. The entire sample was added to a 450 L cuvette (Quartz, Hellma GmbH & Co. KG, Müllheim, Germany) and analyzed by UVŒvisible spectroscopy (Beckmann DU 640, Beckmann Coulter) with A305 monitoring of the sample to quantify the product (2 E,4 E)-styrylacrylate ( 7) by comparison against a concentration series (ranging from 1 to 100 M) made from authentic (2 E,4 E)-styrylacrylate dissolved in 50 mM phosphate buffer (pH 8.5), analyzed by the same method. A sample blank of 6 ( max = 275 nm; A305 = 0.007 at 0.1 mM) was used to subtract the backgr ound absorbance. Every hour, amino group donor 6 was added in an amount equal to that of 7 produced, to keep 6 at a concentration of ~1 mM during reaction. Aliquots (100 L) were withdrawn at 0, 4, 7, 10, 15, and 20 h, and to each were added internal standards 2 -methylcinnamate and 3 -fluoro-( R)- -phenylalanine (100 nmol of each) and 6 N HCl, until the pH was 2. The samples were extra cted with diethyl ether (2 × 1 mL). The organic fractions were combined. The solvent was ev aporated, and the remaining residue was treated with a dilute diazomethane solution to conv ert the acrylic acids to their methyl esters. Methyl esters of the biosynthetically derived 7 and the unused acrylates 14 in the reaction mixture were quantified by GC/EIMS. The abundances of their base peak fragment ions were compared to those of the same fragment ions generat ed from authentic samples of the methyl 69 esters of 7 and 14 , analyzed by an identical method. The remaining aq ueous fraction was basified to pH 10 (6 N NaOH), and the amino acids were deriv atized to their N-(ethoxycarbonyl) methyl ester analogues by reaction with ethyl chloroformat e and subsequently with a dilute diazomethane solution, as described before. The est ers were quantified by GC/EIMS analysis where the abundance of the base peak fragment ion d erived from each ester was compared to that of the same ion generated by identically analyzed a uthentic samples of N-acyl 3 -methyl-( S)- - and ( R)- -phenylalanine methyl esters at concentrations betw een 1 and 100 M. 3.2.10. Biosynthesis and Characterization of a 3,4-Dihydron aphthalene-2-carboxylic Acid (16-Acr) from 16 (1 R)-1-Amino-1,2,3,4-tetrahydronaphthalene-2-carboxyli c acid ( 16 ) (5 mM) was incubated with Tc PAM (1 mg, 13 nmol) in 50 mM phosphate buffer (20 m L, pH 8.5) at 31 °C. After 3 days, the reaction mixture was acidified to pH 2 (6 N HCl), and the acrylate derived by deamination of 16 (designated 16-Acr ) was extracted into diethyl ether (3 × 20 mL). The ether layer was dried (NaSO 4) and evaporated to dryness. The residue (~ 10 mg) w as recrystallized with a 0.1 M HCl/ethanol mixture (10:1, v/v), and t he solvent was decanted and removed in vacuo. The resulting crystals were characterized by 1H and 13 C NMR and GC/EIMS analyses. The exact mass was assessed on a Q-TOF Altima API m ass spectrometer (Waters, Milford, MA): 1H NMR (500 MHz, CDCl 3) : 7.65 (bs, 1H, HŒC C, olefin), 7.22 (m, 4H, aromatic-H), 2.88 (t, 2H, 3JHH = 8.5 Hz, CH 2, benzylic), 2.61 (dt, 2H, 3JHH = 8.5 Hz, 4JHH = 1.0 Hz, CH 2, allylic); 13 C NMR (125 MHz, CDCl 3) : 172 (C O), 138.8 (HŒC CŒCO 2H), 126.8 (HŒ CCŒ CO 2H), 137.2, 132.3, 129.9, 128.8, 128.4, 127.7 (aroma tic-C), 27.5 (CH 2 benzylic), 21.8 (CH 2, 70 allylic); exact mass [M Œ H] observed 173.0599, calculated 173.0603 for C 11 H9O2 (see Figure A 3). 3.3. Results 3.3.1. Overexpression of Tc PAM Codon-optimized tcpam was overexpressed from a pET vector by induction w ith isopropyl -D-thiogalactopyranoside in six 1 L cultures of E. coli BL21(DE3) cells engineered to express tcpam . The overproduced Tc PAM was isolated from the bacteria as an N-terminal His 6 fusion, purified to 95% by nickel affinity chromatography, as described previously. 74 A 1 mL stock of this enzyme at 5 mg/mL was used as the source of Tc PAM in the assays described herein. 3.3.2. Calculation of KI for Various Acrylates and Time Course Studies for Optimal Amino Transfer Preliminary guidelines were established to examine the intermolecular transfer of the amino group from ( S)-styryl- -alanine ( 6) to an arylacrylate catalyzed by Tc PAM. First, under steady- state conditions and in the absence of an inhibitor , 6 was optimally converted to 7 at Vmax when the concentration of 6 was 1 mM, but conversely, the rate plummeted when the concentration of 6 exceeded 1 mM (data not shown), likely because of substrate inhibition by a so far unknown mechanism. Therefore, 1 mM 6 was used in reaction mixtures with Tc PAM to assess the dynamics of the transamination reaction in the pres ence of various arylacrylates at varying concentrations. Next, arylacrylates ( 9Œ14 ) were each separately incubated with Tc PAM during the catalysis of 6 to 7 to calculate the competitive dissociation constant s [ KI(A) ] of 9Œ14 . A 71 relationship between 1/ vo vs [I o] was derived from Eqn 3.1 and KI(A) was extracted from the slope (m) of the linear regression curves shown in figure A 1 (See also Eqn 3.2 , Eqn 3.3 , and Eqn 3.4 ). [][] 00 00[]1 0ESk cat vISK MKI= ++ Eqn 3.1 [] 10[] 0[][][][] 00000 SK KMMIvESkESkK catcatI + =+ Eqn 3.2 [][] 00 KMmESkK catI = Eqn 3.3 () [][] 00 KMKImESk cat = Eqn 3.4 The KI(A) values were lower for arylacrylate inhibitors 9Œ14 (Table 3.1 and Figure A1 of the Appendix) than for (2 E,4 E)-styrylacrylate ( 7) ( KI = 337 ± 12 M) (Figure A 2 of the Appendix), suggesting that in the amine exchange reactions, 9Œ14 would bind Tc PAM better than the styrylacrylate product 7, derived from 6. Amino donor substrate 6 at 1 mM (initial concentration) was incubated with each acrylate ( 9Œ14 ) between 50 and 2000 M and co-incubated with Tc PAM (50 g, 0.7 nmol) to evaluate the steady-state rate para meters and to assess the point at which 6 could out compete 9Œ14 for the active site. 72 Table 3.1. Kinetic Parameters of Various Arylacryla tes and Their Conversion to Aminoo Acids Using ( S)-Styryl- -alanine in the Tc PAM Reaction 1 1 4-9 2a Arylacrylate KI(A) ( M) b nmol c 2 (nmol) 2 (%) d 2(:) 3 0.60 (± 0.04) 150 72 ( 2a ) 48 ( 2a ) 49:51 4 1.70 (± 0.05) 150 120 ( 2b ) 80 ( 2b ) 17:83 5 12.0 ( ± 0.5) 150 (8 h) 120 ( 2c ) 80 ( 2c ) 33:67 6 23.0 (± 0.7) 300 170 ( 2d ) 57 ( 2d ) 41:59 7 47.0 (± 2.6) 1000 73 ( 2e ) 7.3 ( 2e ) 75:25 8 106 (± 8) 1000 200 ( 2f ) 20 ( 2f ) 25:75 aR- Substituents are inferred from structures. bTcPAM catalyzed reaction converting 8b/ to 8b/ was inhibited for the calculation of K I values. cInitial concentration of acrylate in the 12 h transamination assays containing TcPAM (0.1 mg/mL ). dYield is with respect to the arylacrylate substrate. The standard deviations are shown and were calculated from triplicate 73 The production of arylalanines ( 8aŒf ) for each reaction was measured over time, and the yield was found to be maximal at 12 h, typically when the corresponding arylacrylate precursor was between 150 and 1000 M (Table 3.1). Then 8aŒf were converted to their N-(ethoxycarbonyl) methyl ester derivatives and verified by GC/EIMS an alysis. The stereochemistry of 8c was assessed, as an example, by converting the amino ac ids to their N-[(1 S)-camphanoyl] methyl ester derivatives. The retention times and mass spe ctrometry fragment ions determined by GC/EIMS analysis of the derivatized biosynthetic am ino acid diastereomers were compared to those of diastereomerically pure authentic compound s; the stereochemistries of the biosynthetic - and -amino acids were identified as 2 S and 3 R, respectively. These stereochemical data supported earlier findings reported elsewhere for t he Tc PAM reaction. 39 , 74 3.3.3. Relationship between the Rates of Formation of 7 an d 8 and the KI values of 9Œ14 The approximate steady-state production rate of 8aŒf (designated as vo8), reflecting the overall reaction flux after 1 h, was compared to th at of the formation of the NH 2ŒMIO complex, estimated by the observed rate at which 6 (at 1 mM) was converted to 7 (designated as vo7) (Table 3.2). The ( vo8)/( vo7) ratio was charted versus the inhibition constants for each acrylate ( 9Œ14 ) [ KI(A) ], yielding a logarithmic relationship ( Figure 3.1 ) ( Eqn 3.5 ). This set of arylacrylates 9Œ14 was selected for this study because the dissociati on constants (0.6Œ106 M) of these inhibitors were wide-ranging, thus allowing t he logarithmic dependence between KI and reaction rates to be observed ( Figure 3.1 ). The regression fit of the ratio of the steady state rates for the formation of 8 from acrylates 9 Œ 14 and of the NH 2-MIO complex ( vo8)/( vo7) plotted against KI(A) for each compound 9 Œ 14 74 is defined by the following Eqn 3.5 . Theoretically, if KI(A) were increased to 438 M, the (vo8)/( vo7) ratio would approach zero as vo8 was slowed because of the poor binding acrylate (amine acceptor). Meanwhile, vo7 would effectively approach the maximum rate at which 6 can be converted to 7; i.e., the conditions could resemble those as if t he acrylate were absent. By contrast, as KI(A) hypothetically approached 0 M ( Figure 3.1 ) (representative of an irreversibly bound acrylate), the ( vo8)/( vo7) ratio mathematically approached a maximum of ~ 3.8 under steady-state conditions ( Eqn 3.5 ). Thus, an amine acceptor that bound tightly and nonproductively to Tc PAM yet also bound productively to the acrylŒ +NH 2ŒE complex (Scheme 3.1) would theoretically increase the overall react ion rate ( vo8) ~ 3.8-fold relative to vo7. Table 3.2: Steady-State Rate of Formation of 7 from 6 and of 8 from an Acrylate-+NH 2-MIO Complex in the Intermolecular Tc PAM Reaction. a Arylacrylate KI(A) vo8 × 10 4 s -1 vo7× 10 4 s -1 9 0.60 (± 0.04) 8a: 1.7 (± 0.1) 1.7 (± 0.1) 10 1.70 (± 0.05) 8b: 3.5 (± 0.1) 4.4 (± 0.3) 11 12.0 (± 0.5) 8c: 18 (± 1) 29 (± 3) 12 23.0 (± 0.7) 8d: 26 (± 2) 60 (± 5) 13 47.0 (± 2.6) 8e: 46 (± 6) 150 (± 3) 14 106 (± 8) 8f: 140 (± 13) 650 (± 20) a The steady-state rate for the conversion of 6 to 7 (0.082 (± 0.002) s -1 ) and the KI (337 M) of 7 in the Tc PAM reaction are provided for comparison. 75 () 800.346log 70vKIA v ® = ® Eqn 3.5 Figure 3.1 . Ratio [( vo8)/( vo7)] of the steady-state rates for the conversion of acrylates 9Œ14 to 8 and of the formation of the NH 2ŒMIO complex plotted vs KI(A) for 9Œ14 (shown in parentheses). 76 Understandably, the rate of production of 8aŒf cannot exceed the rate at which the amine complex is formed in Tc PAM (estimated by vo7). Therefore, vo8 can never exceed vo7, and consequently, ( vo8)/( vo7) approaches 1 as the KI(A) approaches 0 M ( Figure 3.1 ) in the ping-pong-like reaction mechanism (Scheme 3.1). Thi s was evident when the steady-state production rate of 8a [ vo8a = (1.7 ± 0.1) × 10 Œ4 s Œ1 ] was compared to that of 7 [ vo7 = (1.7 ± 0.1) × 10 Œ4 s Œ1 ] (Table 3.2). These data indicated that the rate o f transfer of an amino group to acrylate 9, producing 8a , matched the amination rate for formation of the N H 2ŒE complex (Scheme 3.1), suggesting that 100% of the amino gro up was transferred from the enzyme to the acrylate. Further evaluation of the steady-state ra tes demonstrated that the amino group was likely lost from the NH 2ŒMIO complex as an inherent process of the reaction . To illustrate, the steady-state production rate of 8f [ vo8f = (140 ± 13) × 10 Œ4 sŒ1 ] was compared to that of 7 [vo7 = (650 ± 20) × 10 Œ4 s Œ1 ] in a separate transamination reaction (Table 3.2) . The comparison indicated that the rate of transfer of the amino fr om the enzyme to acrylate 14 , producing 8f , was 4.6-fold slower than the rate of amination to form the modified enzyme (NH 2ŒE complex of Tc PAM (Scheme 3.1)) suggesting that only ~ 20% of the amino group was productively transferred from the enzyme to the acrylate. The re maining NH 2ŒE complex must revert to free enzyme trans (Scheme 3.1) by competitive loss of NH 3 to keep the conversion of 6 to 7 at steady state, as observed. 77 a 0250 500 750 1000 036912 Concentration ( mM) Time (h) b 0500 1000 1500 05101520 Concentration ( mM) Time (h) Figure 3.2 . (a) Time course assay. Amounts of ( S)-styryl- -alanine ( , 6), (2 E,4 E)-styrylacrylate (, 7), ( S)-3 -methyl- - and ( R)-3 -methyl- -phenylalanine ( , 8f- and 8f- ), and trans -3 -methylcinnamate ( , 14 ) in an aminotransferase reaction catalyzed by Tc PAM over 12 h. (b) Steady-state conversion of 6 to 7 of 8f/ and 8f/ by transfer of an amino group from 6 to 14 (5 mg) by Tc PAM catalysis: (2 E,4 E)-styrylacrylate ( 7, ), ( S)-styryl- -alanine ( 6, ), trans -3 -methylcinnamate ( 14 , ), and total of - and -isomers of 3 -methylphenylalanine ( 8f , ). 78 3.3.4. Titration of the Amino Group Donor 6 to Maintain St eady-State Conversion to 7 Also observed during the time course study, when a mixture of 6 and 14 (each at 1 mM) and Tc PAM (300 g, 3.9 nmol in 6 mL of buffer) were incubated (12 h ), the production of 8f was maximal at ~ 20% converted yield (36 ng). Beyond 12 h, 6 was depleted, and consequently, the rate of production of 8f rapidly approached equilibrium ( Figure 3.2a ). Thus, as an alternative, the conversion of 6 to 7 was monitored by UV spectroscopy at A305 during the reaction, and 6was added accordingly to maintain its concentration at 1 mM in the presence of 14 (5 mg, 31 mol) and Tc PAM (1.5 mg, 20 nmol). The production of 8f was measured over 20 h and was obtained at 42% converted yield (2.3 mg, 13 mol) with respect to 14 ( Figure 3.2b ). Therefore, when the reaction mixture was titrated w ith 6, vo7 was apparently kept at the steady-state rate over a longer time frame, resulti ng in the greater production yields of 8f . As mentioned previously, 6 M ammonium salts were used in previous studies 77 to provide a hydrogen/amino group pair that was added across the double bond of an acrylate substrate in reverse reactions catalyzed by various MIO-dependen t enzymes. These reverse reactions were typically conducted at pH 10 at high ammonium salt concentrations, which likely affected the catalytic activity. In this study, the metered addi tion of the amino group donor 6 likely allowed the pH of the reaction to remain optimal at 8.5 and , more importantly, prevented an excessive surplus of ammonium ions from accumulating in the r eaction mixture that could potentially affect the rate of catalysis. 79 3.3.5. Other Amino Donor Substrates Non-natural amino acids, ( S)-2 -furyl- -alanine ( 15 ), (3 R)-aminotetralin-(2 R)-carboxylic acid (16 ), and ( S)-2-aminotetralin-2-carboxylic acid ( 17 ), were incubated in separate assays to assess their utility as amino group donors. These amino ac ids were chosen because, like 6, they were nearly exclusively converted by Tc PAM to their corresponding acrylates with only mino r (<10%), if any, isomeric amino acid made (Table 3.3 ). This suggested that the acrylates from 15Œ17 were likely derived by a route mechanistically sim ilar to that of 7 (i.e., the acrylates dissociate from Tc PAM and leave the NH 2ŒMIO enzyme complex behind). Compounds 15Œ17 indeed transferred their amino group to 14 , but 6 did so faster (0.56 ± 0.02 nmol/min) (Table 3.3). 80 Table 3.3. Relative Steady-State Rates of Transfer of an Amino Group from Non-Natural Amino Acids (6 and 15Œ17) to trans-3 -Methylcinnamate (14) by TcPAM Catalysis Donor Substrate KM (M) a kcat (min -1 )a Acryl: -A.a. b vo × 10 2 (nmol (8f) Łmin -1 )c 6 250 (± 4) 4.9 (± 0.1) 99:1 56 (± 2) 15 130 (± 6) 2.5 (± 0.1) 91:9 22 (± 1) 16 341 (± 6) 1.7 (± 0.3) 100:0 1.0 (± 0.3) 17 352 (± 9) 0.7 (± 0.1) 100:0 0.4 (>± 0.1) a Acryl, acrylate derived from elimination of H/NH 2 from the amino donor; -A.a., -amino acid made from either 6, 15 , or 17 via TcPAM catalysis. b Compound 14 was used as the amino group acceptor. c This ratio represents the proportion of acrylate to -amino acid. Supplied arylacrylates 9Œ14 were the sole means of amino group exchange to prod uce (S)- - and (R)- -arylalanine. Standard deviations are in parentheses and were cal culated from triplicate assays. 81 To simplify the kinetic evaluation, the rate of tra nsfer of an amino group to 14 was considered identical in each reaction; thus, the different ste ady-state rates of production of 8f by use of fisacrificialfl substrates 6 and 15Œ17 were reflective of the NH 2ŒMIO complex loading rate, e.g., 2.5-fold faster with 6 (0.56 ± 0.02 nmol of 8f /min) than with 15 (0.22 ± 0.01 nmol of 8f /min). This difference paralleled the 2-fold difference in catalytic efficiency ( kcat /KM) between 6 (0.020 ± 0.001 min Œ1 MŒ1 ) and 15 (0.010 ± 0.001 min Œ1 MŒ1 ). By contrast, the catalytic efficiencies of tetralins 16 (0.005 ± 0.001 min Œ1 MŒ1 ) and 17 (0.0020 ± 0.0003 min Œ1 MŒ1 ) were only 4- and 10-fold lower, respectively, than that of 6, whereas the transamination rates differed 56- and 140- fold, respectively (Table 3.3).Apparently, the cata lytic efficiencies of Tc PAM, with amino donors 6 and 15Œ17 , trend with the transamination rate, yet the two kin etic parameters are not directly proportional. The identity of the 3,4-dihy dronaphthalene-2-carboxylic acid (designated 16-Acr ) biosynthesized by Tc PAM from both 16 and 17 , in the amino group transfer reactions, was compared to an authentic standard derived biosy nthetically in a large-scale reaction that converted 16 to 16-Acr . NMR and mass spectrometry analysis verified the p roduct as authentic 3,4-dihydronaphthalene-2-carboxylic acid. 82 3.4. Discussion Overall, the intermolecular transamination reaction catalyzed by Tc PAM required that the amino donor substrate yield an acrylate intermediat e with a binding affinity for the active site lower than that of the acrylate serving as the amin o acceptor. ( S)-Styryl- -alanine served as a suitable amino donor for probing the kinetic parame ters of the transaminase reaction catalyzed by Tc PAM in the presence of a series of arylacrylates, w ith varying enzyme binding affinities that served as amino group acceptor substrates and as inhibitors in a quasi-ping-pong exchange mechanism. These results suggested that the nearly exclusive intramolecular transfer process observed for Tc PAM and its natural substrate 38 represented an extraordinary balance between both the retention of the cinnamate reaction interm ediate and the migratory amino group in the active site. Moreover, the transamination reaction followed a course on which 6 and acrylates 9Œ 14 engaged in a sequential ping-pong exchange in the active site. Because 9Œ14 competitively inhibited the catalysis of the conversion of 6 to 7 in the first step in which modified Tc PAM (the NH 2ŒMIO complex) was produced, the transaminase mechan ism accordingly deviated from true ping-pong (double-displacement) exchange. Generally , the substrate of the second step does not inhibit the first reaction step in a ping-pong mech anism. 95 Thus, strong inhibition of the first step in the Tc PAM transaminase reaction by a tight binding acryla te (amino group acceptor) resulted in a slower than expected reaction flux to the - and -arylalanine products 8 as the amount of acrylate substrate was increased. 83 a b c d e f Figure 3.3 . Modeled in the Tc PAM active site are natural substrates (a) ( S)- -phenylalanine and (b) ( R)- -phenylalanine for reference, (c) ( S)-styryl- -alanine, (d) ( S)- 2-furyl- -alanine, (e) (3 R)-3-aminotetralin-(2 R)-2-carboxylate, and (f) ( S)-2- aminotetralin-2-carboxylate. PyMOL (Schrödinger LLC , Cambridge, MA) was used for the substrate modeling by preserving the key intera ctions with the active site residues. 84 Interestingly, the Tc PAM active site can apparently accommodate the bicy clic tetralin amino acids 16 and 17 in an orientation that produces the corresponding dihydronaphthalene derived from each. The structure of Tc PAM in complex with cinnamate was determined in a p revious report 38 and is used here to approximate the trajectory of each non-natural amino acid substrate used as an amino group donor ( Figure 3.3 ). Key substrate docking interactions occur between the natural substrate phenylalanine and active site residues. The carboxylate and the aromatic ring of the substrate form a salt bridge with Arg32 5 and a hydrophobic contact with Leu104 of Tc PAM, respectively. 38 The amino group of ( S)- -phenylalanine is held above and proximate to the methylidene carbon of the methylidene imidazolo ne (MIO) group by a hydrogen bond interaction between the amino group of the substrat e and the hydroxyl of Tyr322 (not shown) in the active site ( Figure 3.3a ). The trajectory of the natural substrate phenylal anine essentially traces the carbon skeleton of the cocrystallized ci nnamate scaffold, and the -phenylalanine traces the carbon configuration of the presumed cin namate rotamer needed to position C for the amino group rebound ( Figure 3.3b ). 38 Tc PAM seemingly accommodates ( S)-styryl- -alanine (Figure 3.3c ) and ( S)-2 -furyl- -alanine ( Figure 3.3d ) in a conformation similar to the trajectory of the modeled natural substrate ( Figure 3.3a ), and these substrates exhibited rates of amino group transfer higher than those of the amino tetra lins. The furyl ring ( Figure 3.3d ) is oriented so that the heteroatom is pointed at Glu455 and tow ard the solvent-exposed entry point of the active site, where it could engage in hydrogen bond ing. By contrast, the dimethylene bridge of (3 R)-3-aminotetralin-(2 R)-2-carboxylate ( Figure 3.3e ) causes the aromatic ring of the substrate to reside at a displaced angle compared to the -phenylalanine congener model ( Figure 3.3b ); this displacement of the aromatic ring is more pron ounced in the model for ( S)-2-aminotetralin- 2-carboxylate [the presumed productive enantiomer t hat is analogous to the (2 S)-antipode of the 85 natural substrate] ( Figure 3.3f ). The distorted docking conformations of the latte r likely contribute to their poor catalytic efficiencies ( kcat /KM) in the transamination reaction catalyzed by Tc PAM. In addition, on the basis of the structure of a Tc PAMŒcinnamate complex, 38 the dimethylene bridge likely prevented the tetralins f rom adopting an optimal conformation for binding to the active site and transferring their a mino group. Further, the bridged 16 and 17 likely also, in part, precluded the reversible /-interchange via a two-bond rotation of the dihydronaphthalene intermediate, as proposed in pre vious accounts. 38 , 57 , 61 3.5. Conclusion Tc PAM was employed as an amino acid:arylacrylate tran saminase, and an interesting mechanistic property of the reaction was elucidated . The enzyme was originally characterized as principally producing its natural product ( R)- -phenylalanine from ( S)- -phenylalanine with nearly exclusive retention of the amino group and t he carbon skeleton during the reaction. 38 A distinguishing feature of the Tc PAM reaction was revealed in this study when ( S)-styryl- -alanine ( 6) was used as a substrate. The release of the ensui ng 7 from Tc PAM was apparently significantly faster than the release of the amino group from the enzyme. Thus, a significant proportion of Tc PAM existed as the NH 2ŒMIO enzyme complex, and intermolecular transfer of the amino group to the exogenously supplied arylacr ylates to produce - and -arylalanines was observed as the principal route of amino group tran sfer. In addition, it was demonstrated that under steady-state conditions, an appreciable amoun t of the amino group is lost nonproductively, likely as NH 3, from the NH 2ŒMIO complex to reset Tc PAM for the next round of catalysis. This loss of the amino group was observed to occur preva lently when the exogenously supplied acrylate (amino group acceptor) bound the enzyme mo re weakly in the reaction mixture. With an 86 improved understanding of the mechanism of Tc PAM and knowledge of how to employ surrogate substrates in examining cryptic aspects o f the aminomutase chemistry and kinetics, it may become feasible to measure the rate at which th e amino group is released from the enzyme complex. In addition, bicyclic tetralin amino acids 16 (a bicyclic -amino acid) and 17 (a bicyclic -amino acid) were shown for the first time to functi on as surrogate substrates in the Tc PAM reaction, or in any MIO-dependent enzyme-catalyzed reaction, to the best of our knowledge. The product pools derived from substrates 16 and 17 were comprised exclusively of the same corresponding acrylate ( 16 -Acr ). This suggested that a so far unknown impediment of the reaction stalled the progress of Œ isomerization, such as the ring fusion of the subs trates preventing access to a productive rotamer of the di hydronaphthalene intermediate. Moreover, the amino acids produced during the transamination reac tion were made concurrently with significant (2 E,4 E)-styrylacrylate ( 7) derived from 6 at >99% de. 38 Thus, Tc PAM could be a tractable resource of conjugated dienic carbonyl de rivatives such as 7 (and its analogues). Dienes of this type are typical structural subunits for th e synthesis of natural products and useful as synthetic precursors. 96 87 4. ASSESSING THE DEAMINATION RATE OF NH 2-MIO ADDUCT BY BURST PHASE ANALYSIS Reproduced with permission from [Wanninayake, U.; W alker, K. D., Assessing the deamination rate of a covalent aminomutase adduct b y burst phase analysis. Biochemistry 2012, 51, (26), 5226-5228] Copyright © 2012 American Chem ical Society 4.1. Introduction The subfamily of enzymes that includes aminomutases 38 , 57 , 64 and ammonia lyases 59 depends on the function of a 3,5-dihydro-5-methylidene-4 H-imidazol-4-one (MIO) cofactor within the active site. The recently solved structures of a ph enylalanine aminomutase from Pantoea agglomerans (Pa PAM) 64 and a tyrosine aminomutase from Streptomyces globisporus (Sg TAM) 69 support a mechanism where the amino group of the s ubstrates attacks the MIO moiety ( Scheme 4.1). The alkyl ammonium group is presumably removed by an elimination (E2-like; E2 = bimolecular, concerted elimination r eaction) mechanism that is initiated by removal of the pro -(3 S) proton from the substrate by a catalytic tyrosine residue. 38 , 57 , 64 For Tc PAM, the resulting cinnamate intermediate is princi pally trapped in the active site for the entire isomerization reaction 38 and rotates 180° about the C 1ŒC and C ŒC ipso bonds. The amino group of the aminated-MIO attacks C , and the pro -(3 S) proton is recovered by C to complete the isomerization. In the Tc PAM reaction, the original stereochemical configura tion at both C and C is retained after the re-addition of the NH 2 and proton. In contrast, the bacterial isozyme 88 Pa PAM and the related catalyst Sg TAM invert the stereochemistry at the migration ter mini to make the corresponding -amino acid product (cf. Chapter 2). 61 , 69 Scheme 4.1: Mechanism of MIO-Dependent Aminomutases ˙ ˙ ˙ ˝˝˝˝˝˝˛˚˜ ˝˛˚ The Tc PAM reaction was deemed predominantly intramolecula r; the amino group and proton from -phenylalanine rebound exclusively to the same carb on skeleton to make -phenylalanine. Stable-isotope labeling studies revealed ~ 7% inter molecular amino group transfer from [ 15 N]- -phenylalanine to [ 2H6]-cinnamate (Scheme 4.2a). 38 This observation suggested the cinnamate intermediate occasionally diffused from the active site while the ammoniaŒenzyme (NH 2ŒMIO; aminated 3,5-dihydro-5-methylidene-4 H-imidazol-4-one) (cf. Scheme 4.1 ) adduct remained intact. Thus, the lifetime of the NH 2ŒMIO adduct was longer than the residence time of t he 89 cinnamate complex in the active site. Reciprocally, under the same reaction conditions, Pa PAM did not transfer any label from [ 15 N]- -phenylalanine to [ 2H6]-cinnamate, 61 suggesting that the cinnamate remained in the active site longer than t he lifetime of the NH 2ŒMIO. Another earlier study showed that during the Sg TAM catalytic cycle, the transient amino group was transferred intermolecularly from 3'-chlorotyrosine to 4'-hydro xycinnamate (4'-HOCinn). This data suggested that the transient amino group remained a ttached to the enzyme during the course of the isomerization. This earlier study, however, did not evaluate an isotopically labeled-4 -HOCinn (an isotopomer of the natural pathway interm ediate) to evaluate whether the amino group could transfer from -tyrosine to 4 -HOCinn. Thus, it was unclear if the pathway intermediate, after elimination of NH 3 from -tyrosine, could occasionally exchange intermolecularly with an exogenous source of 4 -HOCinn already in solution. 62 An earlier study showed that Tc PAM converted 99% of ( S)-styryl- -alanine to (2 E,4 E)- styrylacrylate (Scheme 4.2b) at 0.082 s -1 steady state rate ( kcat ) which is at the same order as the steady state rate of natural ( S)- - to ( R)- -phenylalanine isomerization (0.052 s -1 ). 38 This suggested that during the conversion of styryl- -alanine to styrylacrylate, the transient amino group likely remained as the NH 2ŒMIO adduct for the same duration as the - to -phenylalanine conversion. Therefore, deamination rate of the enzy me bound amino adduct should provide mechanistic details of the aminomutase reaction. Th is hypothesis was interrogated by a mixture of ( S)-styryl- -alanine (amino group donor) and an arylacrylate (a mino group acceptor), both at steady state concentration. Acceptor substrates wer e selected, based on KI, to estimate the relative binding affinities compared to that of (2 E,4 E)-styrylacrylate. A model scheme for the transaminase reaction was proposed in order to quan tify the deamination (Scheme 4.3). 35 90 Scheme 4.2: a) Intermolecular Amino Group Transfer from [ 15 N]- -Phenylalanine to [ 2H6]- Cinnamate by Tc PAM Catalysis. b) Predominantly Ammonia Lyase Behav ior of Tc PAM with ( S)-styryl- -alanine. a OO15 NH 3OONH 3DD5OO15 NH 3OOD92.8 % Tc PAM D5OODD5NH 3+3.6 % 3.6 % b OONH 3OOOONH 3<1% >99% Tc PAM + 91 Scheme 4.3: Kinetic Model for Transaminase Reaction Catalyzed by Tc PAM a aShaded inset: Kinetic model to evaluate the burst k inetics of the deamination of (S)-styryl- -alanine. 4.2. Experimental 4.2.1. Chemicals (2 E,4 E)-styrylacrylate was acquired from Alfa Aesar (Ward Hill, MA), ( S)Œ Œphenylalanine was purchased from Sigma-Aldrich, and ( R)Œ Œphenylalanine was obtained from PepTech Corp. (Burlington, MA). 92 4.2.2. Enzyme Preparation Codon-optimized cDNA from Taxus canadensis was previously ligated into expression vector pET28a(+), and the recombinant plasmid encoded an N -terminal His6 tag. 74 The tcpam clone was overexpressed in six 1 L cultures of Escherichia coli BL21(DE3) cells by induction with isopropyl -D-thiogalactopyranoside. The overproduced protein was isolated from the bacteria and purified to 95% by Ni affinity chromatography t o yield 5 mg of protein, as described previously (1). The purity of the enzyme was determ ined by the SDS-PAGE method. Routine assays for assessing enzyme function were conducted with ( S)- -styrylalanine at apparent saturation (1 mM) and Tc PAM (100 g, 1.3 nmol) in 50 mM phosphate buffer with 5% glyc erol at pH 8.5 in 1 mL assays. The assays were analyzed by UVŒvisible spectroscopy (Beckmann DU 640, Beckmann Coulter, Brea, CA) with A340 monitoring of the sample to quantify the product (2 E,4 E)-styrylacrylate. 4.2.3. Quantification of Biosynthetic Styrylacrylate durin g Kinetic Progressions The molar absorptivity constant ( 340 = 4 × 10 3 M -1 Łcm -1 , pH 8.5, 23 °C) for (2 E,4 E)- styrylacrylate was calculated from standard curve f ormed by UV-visible spectroscopy (Beckmann DU 640, Beckmann Coulter Brea, CA). The a bsorbance at A340 was measured for a series of concentrations of (2 E,4 E)-styrylacrylate (1-100 M) dissolved in 50 mM phosphate buffer (pH 8.5). A sample blank of 50 mM phosphate buffer was used to subtract the background absorbance. It should be noted, the 1 6 mM ( S)-styryl- -alanine substrate ( max = 275 nm) does not have an absorbance signature at A340 . Beer's law was used to calculate the molar absorp tivity constant, which was then used to convert the absorb ance of styrylacrylate produced during the 93 stopped flow experiment to concentration. The forma tion of (2 E,4 E)-styrylacrylate was monitored at A340 ( 340 = 4 × 10 3 M Œ1 cm Œ1 , pH 8.5) with a stopped flow spectrometer (model SX.18MV-R, dead-time = 2 ms, 1 cm optical path, the rmostatically controlled at 23 °C). Aliquots (125 L) of both Tc PAM (5.5 M final concentration) and styryl- -alanine (final concentrations between 50 and 750 M) ( Figure 4.1 ) were separately mixed, and the absorbance was measured at 0.12 s intervals over 50 s. 4.3. Results The Tc PAM catalysis of the ( S)-styryl- -alanine ammonia lyase reaction shows a pre-steady burst at the initial progression of the reaction (F igure 4.1). The resulting absorbance data at each concentration at each time point were fit by nonlin ear least-squares regression ( Eqn 4.1 ) to obtain the burst amplitude ( B) of the presteady-state and the velocity of the r eaction at steady- state ( A). is used generally to define the kinetic paramete rs when an enzyme-bound intermediate slowly dissociates to form the free enzyme. 97 This equation was applied to calculate the rate constants and burst phase parameters of the Tc PAM reaction. 94 '[](1) kt PAtBe -=+- Eqn 4.1 Figure 4.1 . Evaluation of the kinetic model (Scheme 4.3, shad ed inset) for Tc PAM burst kinetics (Eqn 4.1) was used to globally fit experimental pro gress curves (Kaleidagraph 4.0) spanning six different ( S)-styryl- -alanine concentrations incubated with Tc PAM (5.5 M). Release of (2 E,4 E)-styrylacrylate was measured in a stopped-flow cel l by A340 monitoring. Each time point is an average of three progression curves. 95 Where the terms are; [ P]: product concentration; A: steady state velocity (Eqn 4.2); B: burst amplitude (Eqn 4.3); t: time; and k: apparent first order rate constant of the pre-ste ady-state. [][] 2300 [] 23 M0 kkES Aapp kk KS = ++ Eqn 4.2 22[] 02[] 0[] 23 M0 SkBE app kk KS = + + Eqn 4.3 Where the terms are [ E]o: total-enzyme concentration; [S] o: initial substrate concentration; and the apparent KM ( KMapp ) defined in Eqn 4.4 and the kcat is defined in Eqn 4.6. 3M23 kapp KK Mkk = + Eqn 4.4 12 1kk KMk +- = Eqn 4.5 23 23 kk kcat kk =+ Eqn 4.6 The linear relationship with [S] o/A to [S] o was derived from Eqn 4.2 via a series of rearrangements and simplifications (Eqn 4.7, Eqn 4. 8, & Eqn 4.9). The KMapp of Tc PAM for styryl- -alanine was determined from the ratio of Intercept to the gradient of the linear regression fit of [S] o/A vs [S] o (Figure 4.2). [] 230 [] [] 023 M0 kkE Aapp Skk KS = ++ Eqn 4.7 96 [][] 023230 M[][] 023230 app SkkkkS KAEkkkkE ++ =+ Eqn 4.8 [] 10M[] 0[][] 00 app SKSAEkEk catcat =+ Eqn 4.9 01000 2000 3000 4000 5000 0100200300400500600700800 y = 554.7 + 5.3086x R 2= 0.99134 [S] o/A (s) [S] o ( mM) Figure 4.2 . Hanes-Woolf analysis of the steady-state rates. T he progress curves) were individually fit to the burst equation to evaluate the steady-state velocities ( A), for each concentration of styryl- -alanine, of the Tc PAM burst kinetics. The average value for each data point A was used. The dependence [S] o/A on [S] o with S.D. for the triplicate measurements of A. Linear regression fit ([S] o/A = 554.7 + 5.309[S] o; R 2 = 0.9913) to the data (solid line). 97 22[][] 2300 2[] 23 M0 22[] 02[] 0[] 23 M0 kkES app kk KS ABSkEapp kk KS + + = + + Eqn 4.10 22 [] 03 ABEk = Eqn 4.11 00.005 0.01 0.015 0.02 0.025 0.03 00.511.522.533.5 y = 0.0010366 + 0.0091535x R 2= 0.99137 A2 ( mM2s-2)B ( mM s -1 ) Figure 4.3 : The progression curves were individually fit to t he burst equation to evaluate the steady-state velocity A and the burst amplitude B for the Tc PAM burst kinetics. The average value for each data point A was used. A2 dependence on B with S.D. for the triplicate measurements. Linear regression fit ( A2 = 0.001037 + 0.009153 B; R 2 = 0.9914) to the data (solid line). 98 Terms A2 and B were simplified and then rearranged (Eqn 4.10) to define a linear relationship between A2 and B at each substrate concentration (Eqn 4.11). Since [ E]o was known, k3 was thus calculated from the slope ([ E]ok32) of the plot of A2 against B (Figure 4.3). The steady state kinetic parameters KMapp (Eqn 4.4) and kcat of Tc PAM for styryl- -alanine were calculated from a Hanes-Woolf plot of [ S]o/A against [S] o (Figure 4.2), where A was extracted from the steady- state formation of styrylacrylate ( Figure 4.4 . The progress curves were individually fit to the burst equation to evaluate the steady-state velocit y A and the burst amplitude B for the Tc PAM burst kinetics. The average value for each data poi nt A was used. The dependence of A/B on 1/[S] o with S.D. for the triplicate measurements. Linear regression fit ( A/B = 0.049558 + 6.88821/[S] o; R 2 = 0.97933) to the data (solid line). Figure 4.4 ). Rate constant k2 was calculated from kcat that relates k3 and k2 (Eqn 4.6). A linear relationship between A/B and 1/[S] o was established from Eqn 4.12 where k3(k3 + k2)/ k2 is the intercept (Figure 4.4). The experimentally determined values for k2 (0.19 ± 0.01 s Œ1 ) and k3 (0.041 ± 0.002 s Œ1 ) were substituted into k3(k3+k2)/ k2, and the resulting value (0.050 ± 0.002 s Œ1 ) was comparable to the intercept (0.049 ± 0.004 s Œ1 ) from Eqn 4.12. Thus, the burst phase kinetic anal ysis of styryl- -alanine deamination by Tc PAM was considered reliable. ()()1323323 [] 220 app kkkkkkK AMBkkS ++ =+ Eqn 4.12 99 00.02 0.04 0.06 0.08 0.1 0.12 00.0020.0040.0060.0080.010.012 y = 0.049558 + 6.8882x R 2= 0.97933 A/B (s -1)1/[S] o ( mM-1) Figure 4.4 . The progress curves were individually fit to the burst equation to evaluate the steady- state velocity A and the burst amplitude B for the Tc PAM burst kinetics. The average value for each data point A was used. The dependence of A/B on 1/[S] o with S.D. for the triplicate measurements. Linear regression fit ( A/B = 0.049558 + 6.88821/[S] o; R 2 = 0.97933) to the data (solid line). Terms A2 and B were simplified and then rearranged (Eqn 4.10) to define a linear relationship between A2 and B at each substrate concentration (Eqn 4.11). Since [ E]o was known, k3 was thus calculated from the slope ([ E]ok32) of the plot of A2 against B (Figure 4.3). The steady state kinetic parameters KMapp (Eqn 4.4) and kcat of Tc PAM for styryl- -alanine were calculated from a Hanes-Woolf plot of [ S]o/A against [S] o (Figure 4.2), where A was extracted from the steady- 100 state formation of styrylacrylate. Rate constant k2 was calculated from kcat that relates k3 and k2 (Eqn 4.6). A linear relationship between A/B and 1/[S] o was established from Eqn 4.12 where k3(k3 + k2)/ k2 is the intercept (Figure 4.4). The experimentally d etermined values for k2 (0.19 ± 0.01 s Œ1 ) and k3 (0.041 ± 0.002 s Œ1 ) were substituted into k3(k3+k2)/ k2, and the resulting value (0.050 ± 0.002 s Œ1 ) was comparable to the intercept (0.049 ± 0.004 s Œ1 ) from Eqn 4.12. Thus, the burst phase kinetic analysis of styryl- -alanine deamination by Tc PAM was considered reliable. 4.4. Discussion The KMapp of Tc PAM for styryl- -alanine was 105 ± 10 M and the kcat was 0.034 ± 0.002 s Œ1 at 23 °C (pH 8.5). These values were agreeable ( KMapp = 250 M; kcat = 0.082 s Œ1 ) to those reported in an earlier study for the same reaction, at 31 °C. 35 The KM (588 ± 37 M) of Tc PAM was calculated from KMapp (see Eqn 4.4 and Eqn 4.5) for styryl- -alanine in the ammonia elimination reaction that produced styrylacrylate ( Scheme 4.3, shaded inset). The burst amplitude (B) and steady-state velocity ( A) was dependent on [S] o. The rate constant ( k3 = 0.041 ± 0.002 s Œ1) for the release of NH 3 (that reset Tc PAM for another catalytic cycle) was similar to kcat (0.034 ± 0.002 s Œ1 ) for the conversion of styryl- -alanine to styrylacrylate. In contrast, the calcul ated rate constant ( k2 = 0.19 ± 0.01 s Œ1 ) for the release of styrylacrylate was 5-fold grea ter than kcat . Therefore, kcat for the overall reaction was slowed 5-fold likely by the slower deamination rate of enzyme. 101 Tc PAM normally deaminates -phenylalanine in the first committed step, and the ensuing cinnamate complex, the protonated catalytic residue , and the NH 2ŒMIO adduct (see Scheme 4.1) are retained long enough to complete the intramolec ular isomerization. 38 Under steady-state conditions, the kcat / for this conversion was 0.014 ± 0.001 s Œ1 at 23 °C, corresponding to a transit time (1/ kcat /) of 71 s. In addition, -phenylalanine and cinnamate were produced at a 9:1 ratio. This kinetics data suggested that the transi t time (1/ k7 + 1/ k9 + 1/ k3) from the -PheŒ +NH 2ŒMIOŒE complex to the release of NH 3, via the cinnamate-production pathway (Scheme 4.4) , was 9 times longer (639 s) than 1/ kcat /. Since 1/ k3 (24 s) on this pathway is known, then (1/ k7 + 1/ k9), the estimated transit time to cinnamate (Scheme 4.4), was therefore 615 s. The latter transit time supported the notion that the dwell time of ci nnamate in the active site is sufficient to preferentially promote the intramolecular amino gro up rebound pathway to -phenylalanine. 102 Scheme 4.4: Kinetic Model for the Proposed Mechanis m of the Tc PAM-Catalyzed Conversion of - to -Phenylalanine The efficiency of the amino group transfer from the NH 2ŒMIO adduct in the described exchange reaction was largely dependent on the bind ing affinity of the acceptor arylacrylate for Tc PAM. The amino group transferred at nearly 100% eff iciency to a tighter-binding acceptor, while the efficiency decreased exponentially for we aker-binding acceptors when the - and -amino acid mixtures were made. 35 This suggested that the transit time (1/ k4 + 1/ k5) progressing from the amination of the acceptor to the release o f the amino acids was significantly less than the transit time (1/ k3) for the deamination of the NH 2ŒMIO adduct to the apo -MIO cofactor (Scheme 4.3). Reciprocally, for the weakest binding acceptor, the transit times were exchanged, where (1/ k4 + 1/ k5) > (1/ k3), and now the deamination of the NH 2ŒMIO adduct predominated. To further understand why the amine transfer efficienc y decreased with weaker binding acceptors in 103 the transamination reaction, and to infer the kinet ics of the mechanistically similar MIO- dependent catalysts, the dissociation rate ( k3) of the NH 2ŒMIO adduct in Tc PAM was calculated herein. Since Tc PAM forms a transient covalent NH 2ŒMIO adduct in its interaction with amino acid substrates, a burst kinetic method was employed to establish the decay rate of the adduct. ( S)- Styryl- -alanine (Peptech Inc., Burlington, MA) was used as an adventitious substrate in the Tc PAM-catalyzed reaction to show direct kinetic evide nce of the accumulation of the purported (NH 2ŒMIO)-modified enzyme. The burst phase of Tc PAM was evident in assays when ( S)- styryl- -alanine was converted to chromophoric product (2 E,4 E)-styrylacrylate in a fast step ( k2), followed by slower release ( k3) of the second product NH 3 (Scheme 4.2, shaded inset). The slower NH 3-release step led to the observed pre-steady state burst. In stopped-flow experiments, styrylacrylate was produced from styrylalanine init ially along an exponential burst phase prior to reaching a steady-state progression. The burst phas e parameters were then used to calculate the rate constants k2 and k3 (Scheme 4.3). 4.5. Conclusion Tc PAM undergoes finite inactivation during turnover o f ( S)-styryl- -alanine to styrylacrylate, resulting in burst kinetics with the steady-state r ate being a dynamic balance between the inactive modified enzyme (NH 2ŒMIO adduct) and reactivation by deamination of the adduct. To our knowledge there are no reports on the direct kineti c evaluation of this deamination for any MIO- dependent enzyme. Furthermore, above study provided kinetic data to indirectly obtain the dwelling time of cinnamic acid intermediate in the active site of Tc PAM during the aminomutase 104 reaction. It will be in great interest to perform a similar study on PAL enzymes to understand the effect on deamination rate and cinnamic acid dwelli ng time to the PAM and PAL activity. In this study, the calculated deamination rate constant for the dissociation of the NH 2ŒMIO adduct contributes information to further understand the m echanism of the Tc PAM reaction. 105 5. A BACTERIAL TYROSINE AMINOMUTASE PROCEEDS THROUGH RETENTION OR INVERSION OF STEREOCHEMISTRY TO CATALYZE ITS ISOMERIZATION REACTION Reproduced with permission from [Wanninayake, U.; W alker, K. D., A bacterial tyrosine aminomutase proceeds through retention or inversion of stereochemistry to catalyze its isomerization reaction. J. Am. Chem. Soc. 2013, 135, (30), 11193-11204] Copyright © 2013 American Chemical Society 5.1. Introduction -Amino acids are emerging as an important class of compounds that are present in bioactive natural products, such as the antineoplastic pharma ceutical paclitaxel isolated from Taxus plants, 1, 98 the aminopeptidase inhibitor bestatin obtained fro m Streptomyces olivoreticuli ,99 an antibacterial blasticidin S from Streptomyces griseochromogenes ,100 the antibiotic agent enediyne C-1027 from Streptomyces globisporus ,101 the anti-tuberculosis agent viomycin from Streptomyces vinaceus ,102 the antibiotic andrimid from Pantoea agglomerans , and cytotoxic agents chondramides A-D from Chondromyces crocatus. 5 In addition, single -aryl- -alanines show anti-epileptogenesis activity. 103 Other -aryl- -alanines have been used as building blocks toward the synthesis of complex bioactive molecules , including -lactams, 104 -peptides as mimics of -peptide, 105 , 106 and antimicrobial compounds. 107 106 Several chiral synthetic strategies have been devel oped for variously substituted -arylalanines that include tandem one-pot processes 108 and conjugate addition of homochiral lithium amides to ,-unsaturated acceptors. 109 The Knoevenagel condensation of benzaldehyde and malonic acid in the presence of NH 4OAC produced a series of racemic -arylalanines. 110 By contrast, there are only a few reports on the bioca talysis of asymmetric -arylalanines from the corresponding readily available natural and nonnatu ral -amino acids through aminomutase catalysis. So far, five 3,5-dihydro-5-methylidene-4 H-imidazol-4-one (MIO)-dependent aminomutases are known to isomerize either ( S)- -phenylalanine (EncP from Streptomyces maritimus ,68 AdmH (Pa PAM) from Pantoea agglomerans ,6, 61 and Tc PAM from Taxus plants 45 , 66 ) or ( S)- -tyrosine (( S)- 1) (SgcC4 ( Sg TAM) from Streptomyces globisporus 69 , 111 and CmdF ( Cc TAM) from Chondromyces crocatus 46 , further described herein) to their respective -amino acids. Efforts to optimize the turnover of this family of aminomutase s provide a potentially alternative means toward scalable biocatalytic production of novel en antiomerically pure -amino acids as synthetic building blocks in medicinal chemistry. E arlier substrate specificity studies showed that Tc PAM can convert substituted aromatic and heteroarom atic -alanines to the corresponding -alanines. Apparently, since TAMs requires the 4'-hy droxyl group on the substrate for catalysis, their use to biosynthesize -tyrosine analogues is limited. 46 , 111 These aminomutases belong to a class I lyase-like f amily (comprised of ammonia lyases 52 , 112 , 113 and aminomutases 6, 46 , 63 , 66 ) where, mechanistically, the amino group of the ar ylalanine 107 substrate nucleophilically attacks the MIO to form an amino acid-MIO adduct. Removal of a -proton via a tyrosine residue, and concommitant eli mination of the NH 2-MIO adduct from the substrate produces an intermediary arylacrylate pro duct. Interchange and rebound of the transient hydrogen and NH 2-MIO to the intermediate, and release of the -amino acid complete the conversion (Scheme 5.1). Stereochemical evidence sh ows that these enzymes can be sorted by their enantioselectivity for the -amino acid product; EncP, Pa PAM and Sg TAM make ( S)- -arylalanines, and Cc TAM and Tc PAM make ( R)- -arylalanines. Further, the properties of the MIO aminomutases segregate according to whether the cryptic stereochemistry at C and C of the product is inverted or retained. The different enantioselectivities of these isozymes is determined by the fate of the intermediate formed o n distinct reaction pathways. Scheme 5.1: General MIO-dependent Aminomutase Mecha nism For example, both Pa PAM and Sg TAM presumably bind their respective substrates and displace the NH 2-MIO adduct and pro -(3 S) hydrogen by anti -elimination. For both enzymes, the 108 NH 2-MIO and hydrogen interchange positions and rebound to the same face of the cinnamate intermediate from which they were removed to form t he ( S)- -amino acids, with inversion of configuration at C and C . Therefore, the intermediate on this reaction sequ ence must reside as a single rotamer. 61 , 63 By contrast, for the isozyme Tc PAM, the cinnamate intermediate is proposed to rotate 180° about the C 1ŒC and C ŒC ipso bonds. Then the labile NH 2 and pro -(3 S) hydrogen exchange positions and reattach to the opposite fac e of the intermediate from which they were removed. This mechanism results in retention of con figuration at C and C to form ( R)- -phenylalanine. 38 , 61 , 66 , 114 Scheme 5.2: The Cc TAM Reaction on the Chondramides A-D Pathway (S)- 1 ( R)- 2 Chondramides A: R 1 = OMe, R 2 = H B: R 1 = OMe, R 2 = Cl C: R 1 = R 2 = H D: R 1 = H, R 2 = Cl 109 Interestingly, the tyrosine aminomutases (TAMs) are less enantioselective than the PAMs. The former make both enantiomers of the -tyrosine ( 2) at steady state and show significantly lower enantiomeric excess after reaching equilibrium, 46 , 115 while the PAM enzymes make product at >99% ee, even after reaching equilibrium, and racem ization is not observed. 38 , 45 , 61 , 114 It should be noted that earlier mutational studies on Cc TAM (E399K) improved the enantioselectivity for (R)- 2 from 69% to 97% ee. 46 Despite this earlier study to correlate active sit e residues with the stereoselectivity of Cc TAM, the basis for why the -amino acid enantioselectivity was enhanced remained unexplored. In addition, the mode of attac hment (i.e., retention or inversion of configuration) of the reciprocally migratory NH 2 group and the hydrogen at C and C during the reaction is unknown. To understand and ultimately improve the substrate specificity profile, turnover rate, and enantioselectivity of Cc TAM, its mechanisms must be fully understood. To th is end, we add further mechanistic detail to Cc TAM, which converts ( S)- 1 to ( R)- 2 on the biosynthetic pathway of the cytotoxic cyclodepsipeptide chondramides A-D in C. crocatus myxobacteria (Scheme 5.2). 5, 46 , 116 The ( R)-product stereochemistry catalyzed by Cc TAM suggested that its stereochemical course is related to that catalyzed by Tc PAM. Herein, we used deuterium-labeled isotopomers of -tyrosine to evaluate the cryptic stereochemistry o f the Cc TAM mechanism and compare it to other aminomutases. 110 5.2. Experimental 5.2.1. Chemicals and Reagents (S)- -Tyrosine, 4'-hydroxycinnamic acid, unlabeled- and [ -2H]-4'-hydroxybenzaldehyde, hippuric acid, ethyl chloroformate, ( S)-2-methylbutyric anhydride, acetic anhydride, pyri dine, bicyclo[2.2.1]hepta-2,5-diene (norbornadiene), 2H2 gas, (2 S)-[3,3- 2H2]- -tyrosine ((2 S)-[3,3- 2H2]-1) and Dowex 50W (100-200 mesh, H + form) ion exchange resin were purchased from Sigma Aldrich (St. Louis, MO). ( R)-3-Amino-3-(4'-hydroxyphenyl)propanoic acid and ( S)-3- amino-3-(4'-hydroxyphenyl)propanoic acid were purch ased from Peptech Inc. (Bedford, MA). (R)-1,2-Bis(diphenylphosphino)propane (( R)-Prophos) was purchased from Alfa Aesar (Ward Hill, MA). Bicyclo[2.2.1]hepta-2,5-diene-rhodium(I) chloride dimer ([Rh(NBD)Cl] 2) and silver perchlorate were purchased from Strem Chemicals (Ne wburyport, MA). Hydrogen gas (99.995% purity) was obtained from Airgas Great Lakes (Indep endence, OH 44131). 5.2.2. Instrumentation 1H-NMR (500 MHz), 2H-NMR (76.7 MHz) and 13 C NMR (126 MHz) spectra were obtained on a Varian NMR-Spectrometer using standard acquisi tion parameters. X-ray crystallographic data were collected using a Bruker CCD (charge coup led device)Œbased diffractometer equipped with an Oxford Cryostream low-temperature apparatus operating at 173 K. The biosynthetic products were quantified and analyzed by gas chroma tography/electron-impact mass spectrometry (GC/EIMS): GC (model 6800N, Agilent, S anta Clara, CA) was coupled to a mass analyzer (model 5973 inert , Agilent, Santa Clara, CA) in ion scan mode from 5 0-400 atomic mass units. The GC conditions were as follows: colu mn temperature was held at 200 °C for 1 111 min and then increased linearly at 20 °C/min to 250 °C with a 1-min hold. Splitless injection was selected, and helium was used as the carrier gas. 5.2.3. Subcloning, Expression, and Purification of Cc TAM The cctam cDNA was codon-optimized by GenScript (Piscataway, NJ 08854) in the pUC57 vector. The clone was ligated into the pET-28a(+) v ector between the Nde I/ Bam HI cloning sites. Recombinant plasmids were used to transform Escherichia coli BL21(DE3) cells, which were grown in 1 L LuriaBertani medium supplemented with kanamycin (50 µg/mL). Overexpression of Cc TAM was induced by the addition of isopropyl- -D-thiogalactopyranoside (100 µM) to the medium, and the cells were grown at 16 °C for 16 h. The cells were then harvested by centrifugation, and the resulting pellet was resusp ended in buffer (50 mL of 50 mM sodium phosphate containing 5% (v/v) glycerol, 300 mM NaCl , and 10 mM imidazole, pH 8.0). The cells were lysed by sonication, and the cellular de bris and light membranes were removed by centrifugation. The crude, functionally soluble ami nomutase was purified by nickel nitrilotriacetic acid (Ni-NTA) affinity chromatogra phy according to the protocol described by the manufacturer (Qiagen, Valencia, CA). Fractions that eluted from the column, containing active Cc TAM (62 kDa) in 250 mM imidazole, were combined. The buffer was exchanged with 50 mM sodium phosphate (pH 8.0) containing 5% (v/v) glyce rol through several concentration/dilution cycles, using a Centriprep centrifugal filter (30,0 00 MWCO, Millipore). The purity of the concentrated Cc TAM (15 mg/mL in 4 mL, estimated by the Bradford me thod) was >95% by SDSPAGE with Coomassie Blue staining. 112 5.2.4. Assessing the Activity and Stereochemistry of the Cc TAM Reaction (S)- 1 (1 mM) was incubated with Cc TAM (0.1 mg) at 31 °C in 50 mM phosphate buffer (1 mL, pH 8.5) for 1 h. The assay mixtures were treate d in twice (5 min each time) with pyridine (0.6 mmol) and ethyl chloroformate (0.5 mmol). This step, containing excess ethyl chloroformate, caused partial ethyl esterification. Afterwards, the reaction was acidified to pH 2 (6M HCl), and the 4'- O, 2- N- and 4'- O,3- N- di(ethoxycarbonyl) derivative of - and -tyrosine, respectively, were extracted into ether and treated with a slight excess of diazomethane. The resulting sample contained a mixture of ethyl (~20 m ol%) and methyl (80 mol%) esters. 4'- Hydroxycinnamic acid by-product was converted to it s 4'- O- ethoxycarbonyl-( E)-coumaric acid ethyl (10 mol%) and methyl (90 mol%) esters under t hese conditions. The methyl esters of the -tyrosine derivative and the 4'- O- ethoxycarbonyl-( E)-coumaric acid, and the ethyl ester of the -tyrosine derivative were analyzed by GC/EIMS. To confirm the stereochemistry of the -isomer product, 46 another sample of Cc TAM (0.1 mg) was incubated at 31 °C with ( S)- 1 (1 mM) in phosphate buffer (1 mL, pH 8.5) for 1 h. To this solution were added pyridine (50 L, 0.64 mmol) and ( S)-2-methylbutyric anhydride (10 L, 0.05 mmol), and the mixture was stirred vigorously for 5 min. Another batch of pyridine (0.64 mmol) and ( S)-2-methylbutyric anhydride (0.05 mmol) was added, and the reaction was stirred for 5 min. The solution was acidified to pH 2 (6M H Cl) and extracted with diethyl ether (3 × 2 mL). The ether fractions were combined and dried. T he resulting residue was dissolved in diethyl ether (100 L) and the solution was titrated with a dilute solu tion of diazomethane dissolved in diethyl ether until the yellow color persisted. The se samples were analyzed by GC/EIMS and compared to the retention time and mass fragmentati on of authentic standards (Figure A 38-40) 113 5.2.5. Synthesis of Authentic - (1) and -Tyrosine (2) and ( E)-4-Hydroxycinnamic Acid Derivatives To ( S)Œ -, (R)Œ-tyrosine or (E)Œ 4'Œhydroxycinnamic acid (0.1 mmol of each) dissolve d in 50 mM phosphate buffer (1 mL, pH 8.5) were added pyrid ine (200 L, 2.4 mmol) and ethyl chloroformate (200 L, 1.6 mmol). The reactions were stirred for 5 min and treated with another batch of pyridine and ethyl chloroformate (both at 200 L) for 5 min with stirring. The ethanolic mixtures caused partial ethyl esterification. Each mixture was acidified to pH 2 (6M HCl), and extracted into diethyl ether (3 × 2 mL). The ether fractions were combined, dried under vacuum, and the residue was dissolved in methanol (100 L). To this solution was added a dilute diazomethane solution dissolved in diethyl ether, u ntil the yellow color persisted. This procedure resulted in a mixture of ethyl and methyl esters fo r each sample, as described earlier. 5.2.5.1. 4'-O-Ethoxycarbonyl-(E)-coumaric Acid Methyl Ester (i.e., 4'-O-Ethylcarboxy- (E)-4'-Hydroxycinnamic Acid Methyl Ester) The coumarate ester was recrystallized from ethanol and isolated at 80% yield (20 mg). Exact mass [M + H] + observed 251.0890 & calculated 251.0919 for [C 13 H15 O5]+. 1H NMR (500 MHz, CDCl 3) : 7.68 (d, J = 16.1 Hz, 1 H), 7.55 (d, J = 8.3 Hz, 2 H), 7.22 (d, J = 8.3 Hz, 1 H), 6.41 (d, J = 16.1 Hz, 1 H), 4.34 (q, J = 7.3 Hz, 2 H), 3.82 (s, 3 H), 1.41 (t, J = 7.3 Hz, 3 H). 13 C NMR (126 MHz, CDCl 3) : 167.2 (C (O)OCH 3), 153.2 (OC(O)O), 152.4 (C4'), 143.6 (C ), 132.2 (C1'), 129.2 (C2'), 121.6 (C3'), 118.1 (C ), 65.1 (OC(O)OC H2CH 3), 51.7 (C(O)OC H3), 14.2 (OC(O)OCH 2CH3). (Figure A 6-Figure A 8 for GC/EIMS fragmentation dataand NMR Spectra) 114 5.2.5.2. The 4'-O,2-N-Di(ethoxycarbonyl)- -tyrosine Methyl Ester. A mixture of ethyl (10 mol%) and methyl (90 mol%) e sters of 4'- O,2- N-di(ethoxycarbonyl)- -tyrosine was obtained at 74% yield (25 mg). The methyl ester of the -tyrosine derivative was used in the GC/EIMS analyses of labeled and unlabel ed -tyrosine substrates. Therefore, the mixed ester sample, dissolved in chloroform (500 L), was loaded onto a preparative silica gel TLC plate and eluted with 90:10 hexane:ethyl acetat e. Authentic 4- O,2- N-di(ethoxycarbonyl)- -tyrosine methyl ester was isolated ( Rf = 0.65) at 67% yield (22.5 mg). Exact mass [M + H + Pyridine] + observed 419.1781 & calculated 419.1818 for [C 21 H27 N2O7]+. 1H NMR (500 MHz, CDCl 3) : 7.21 - 6.98 (m, 4 H), 5.10 (d, J = 7.3 Hz, 1 H), 4.67 - 4.55 (m, 1 H), 4.28 (q, J = 7.1 Hz, 2 H), 4.08 (q, J = 6.7 Hz, 2 H), 3.69 (s, 3 H), 3.10 (dd, J = 6.1, 14.0 Hz, 1 H), 3.05 (dd, J = 6.1, 14.0 Hz, 1 H), 1.36 (t, J = 7.0 Hz, 3 H), 1.20 (t, J = 6.7 Hz, 3 H). 13 C NMR (126 MHz, CDCl 3) : 171.9 (C (O)OCH 3), 155.8 (OC(O)NH), 153.5 (OC(O)O), 150.2 (C4'), 13 3.6 (C1'), 130.3 (C2'), 121.0 (C3'), 64.8 (CH 3CH2OC(O)O), 61.2 (CH 3CH2OC(O)NH), 54.5 (C(O)OC H3), 52.3 (C ), 37.6 (C ), 14.4 (C H3CH 2OC(O)O), 14.1 (C H3CH 2OC(O)NH). (See Figure A10-12 for GC/EIMS fragmentation data and NMR spectra). 4'-O,2-N-Di(ethoxycarbonyl)- -tyrosine Ethyl Ester . The corresponding ethyl ester was isolated ( Rf = 0.50) from the TLC plate at 7% yield (2.5 mg). Exact mass [M + H + Pyridine] + observed 433.1933 & calculated 433.1974 for [C 22 H29 N2O7]+. 1H NMR (500 MHz, CDCl 3) : 7.17 (d, J = 8.6 Hz, 2 H), 7.14 Œ 7.11 (m, J = 8.6 Hz, 2 H), 5.14 (d, J = 7.8 Hz, 1 H), 4.63 (dd, J = 5.6, 13.4 Hz, 1 H), 4.33 (q, J = 7.1 Hz, 2 H), 4.17 (q, J = 7.2 Hz, 2 H), 4.12 (q, J = 7.1 Hz, 2 H), 3.11 (d, J = 5.1 Hz, 2 H), 1.40 (t, J = 7.1 Hz, 3 H), 1.25 (t, J = 7.3 Hz, 3 H), 1.24 (t, J = 7.2 Hz, 3 H). 13 C NMR (126 MHz, CHCl 3) : 171.7 115 (C (O)OCH 3), 156.1 (OC(O)NH), 153.8 (OC(O)O), 150.4 (C4'), 13 3.9 (C1'), 130.6 (C2'), 121.3 (C3'), 65.1 (CH 3CH2OC(O)O), 61.8 (CH 3CH2OC(O)), 61.4 (CH 3CH2OC(O)NH), 54.9 (C ), 38.0 (C ), 14.7 (C H3CH 2OC(O)O), 14.4 (C H3CH 2OC(O)), 14.3 (C H3CH 2OC(O)NH). (See Figure A 12-14 for NMR spectra) 5.2.5.3. 4'-O,3-N-Di(ethoxycarbonyl)- -tyrosine Ethyl Ester A mixture of ethyl (33 mol%) and methyl (67 mol%) e sters of 4'- O,3- N-di(ethoxycarbonyl)- -tyrosine was isolated ( Rf = 0.38) at 68% yield (24 mg). The ethyl ester of the -tyrosine derivative was used in the GC/EIMS analyses of labe led and unlabeled biosynthetic -tyrosines. Thus, authentic ethyl ester was purified by silica gel TLC (90:10 hexane:ethyl acetate). 4'- O,2- N-Di(ethoxycarbonyl)- -tyrosine ethyl ester was isolated at 23%yield (8 m g). Exact mass [M + H + Pyridine] + observed 433.1922 & calculated 433.1974 for [C 22 H29 N2O7]+. 1H NMR (500 MHz, CDCl 3) : 7.30 (d, J = 8.8 Hz, 2 H), 7.12 (d, J = 8.8 Hz, 2 H), 5.67 (d, J = 6.8 Hz, 1 H), 5.12 (br. s., 1 H), 4.28 (q, J = 7.3 Hz, 2 H), 4.08 (q, J = 6.8 Hz, 2 H), 4.05 (q, J = 7.0 Hz, 2 H), 2.91 - 2.69 (m, 2 H), 1.36 (t, J = 7.1 Hz, 3 H), 1.20 (t, J = 6.8 Hz, 3 H), 1.14 (t, J = 7.3 Hz, 3 H). (See Supporting Information Figures S17 Œ S19 for NMR sp ectra and GC/EIMS fragmentation data). 13 C NMR (126 MHz, CDCl 3) : 170.7 (C (O)OCH 2CH 3), 155.8 (OC(O)NH), 153.5 (OC(O)O), 150.3 (C4'), 138.8 (C1'), 127.4 (C2'), 121.3 (C3'), 64.8 (CH 3CH2OC(O)O)), 61.0 (CH 3CH2OC(O)), 60.7 (CH 3CH2OC(O)NH), 51.0 (C ), 40.6 (C ), 14.5 (C H3CH 2OC(O)O), 14.1 (C H3CH 2OC(O)), 14.0 (C H3CH 2OC(O)NH). (Figure A 15-18 for GC/EIMS fragmentation data and NMR spectra). 4'-O,3-N-Di(ethoxycarbonyl)- -tyrosine Methyl Ester was isolated ( Rf = 0.54) at 46% yield (16 mg). Exact mass [M + H + Pyr idine] + observed 419.1770 & calculated 116 419.1818 for [C 21 H27 N2O7]+. 1H NMR (500 MHz, CDCl 3) : 7.29 (d, J = 8.5 Hz, 2 H), 7.12 (d, J = 8.5 Hz, 1 H), 5.66 (br. s., 1 H), 5.17 - 5.07 (m, 1 H), 4.28 (q, J = 6.9 Hz, 2 H), 4.08 (q, J = 7.3 Hz, 2 H), 3.60 (s, 3 H), 2.90 - 2.73 (m, 2 H), 1.36 (t, J = 7.3 Hz, 3 H), 1.20 (t, J = 7.0 Hz, 3 H). 13 C NMR (126 MHz, CD 3COCD 3) : 170.7 (C (O)OCH 3), 155.9 (OC(O)NH), 153.4 (OC(O)O), 150.5 (C4'), 140.3 (C1'), 128.4 (C2'), 121.7 (C3'), 65.0 (CH 3CH2OC(O)O), 64.4 (CH 3CH2OC(O)NH), 60.0 (C(O)OC H3), 51.0 (C ), 40.2 (C ), 14.1 (C H3CH 2OC(O)O), 13.6 (C H3CH 2OC(O)NH). (See Figure A 18-20 for NMR spectra). 5.2.6. Synthesis of [Rh(NBD) 2]ClO 4 Complex. The [Rh(NBD) 2]ClO 4 complex was synthesized according to a described p rocess. 117 , 118 Briefly, a mixture of dimeric [Rh(NBD)Cl] 2 (0.46 g, 1 mmol) and norbornadiene (0.19 g, 2 mmol) dissolved in CH 2Cl 2 (15 mL) was added to silver perchlorate (0.42 g, 2 mmol) under N 2, and stirred for 1 h. The suspension was filtered to remove the white precipitate, and the filtrate was diluted with dry THF (15 mL). The sample was co ncentrated under vacuum until the orange needles of [Rh(NBD) 2]ClO 4 appeared. The crystals were collected, washed with ice-cold, dry THF, and dried under vacuum to obtain 0.65 g (85% i solated yield) of rust brown crystals. 1H NMR (500 MHz, CDCl 3) : 5.20 (q, J = 2.0 Hz, 4 H), 4.13 (br s, 2 H), 1.51 (t, J = 1.6 Hz, 2 H). 5.2.7. Synthesis of [Rh(NBD)(( R)-Prophos)]ClO 4 Complex. The Rh-Prophos catalyst was prepared by an establis hed procedure. 117 , 118 To the orange-red solution of [Rh(NBD) 2]ClO 4 (0.54 g, 1.4 mmol) and ( R)-Prophos (0.57 g, 1.4 mmol) dissolved in 117 a mixture of dry CH 2Cl 2 and THF (5 mL of each) was added hexane (5 mL) dro pwise under N 2. The solution stood, undisturbed at room temperature for 5 h, and then at 5 °C for 16 h. Orange- red crystals were collected by vacuum filtration an d washed with ice-cold, dry THF and then with hexane. The crystals were dried under N 2 to obtain 0.8 g (80% isolated yield) of the Rh- catalyst. 1H NMR (500 MHz, CDCl 3) : 7.79 - 7.73 (m, 2 H), 7.72 - 7.67 (m, 3 H), 7.65 - 7.61 (m, 2 H), 7.61 - 7.58 (m, 6 H), 7.58 - 7.55 (m, 6 H ), 7.47 - 7.41 (m, 2 H), 7.35 - 7.29 (m, 2 H), 5.42 (br. s., 2 H), 5.33 (br. s., 1 H), 5.31 (s, 1 H), 4.87 (br s, 1 H), 4.28 (br s, 1 H), 4.16 (br s, 1 H), 3.84 - 3.67 (m, 1 H), 2.71 - 2.59 (m, 2 H), 2.5 7 (d, J = 2.2 Hz, 1 H), 2.04 (td, J = 7.4, 12.9 Hz, 1 H), 1.88 - 1.84 (m, 1 H), 1.84 - 1.76 (m, 2 H ), 1.20 (dd, J = 6.5, 12.3 Hz, 4 H). 13 C NMR (126 MHz, DMSO-d 6) : 143.1, 135.1, 135.1, 134.8, 134.5, 134.1, 133.9, 132.8, 132.7, 132.0, 131.9, 131.7, 131.1, 131.0, 130.9, 130.1, 129.3, 12 9.3, 128.9, 128.8, 128.8, 128.6, 128.5, 128.2, 127.8, 125.9, 125.5, 67.0, 63.4, 48.2, 25.1, 14.8, 14.6. 5.2.8. Synthesis of [ 2H]-Labeled (2 S)-1 Isotopomers 5.2.8.1. Synthesis of (Z)-2-Benzamido-3-(4'-hydroxyphenyl)a crylic Acid According to a described procedure, 119 a mixture of 4-hydroxybenzaldehyde (1.9 g, 12.5 mmol), K 2HPO 4 (2.2 g, 12.5 mmol), and acetic anhydride (3.8 mL, 40 mmol) was stirred and heated at 80 °C under N 2 for 5 min. To the mixture was added hippuric acid (2.3 g, 12.5 mmol) in one lot, and the reaction was stirred at 80 °C f or 2 h. Yellow crystals were collected by vacuum filtration and washed with water to obtain a n oxazolone intermediate (3.1 g, 81% yield) that was used without further purification. To the oxazolone (3.07 g, 10 mmol) was added 2% NaOH in 70% aqueous ethanol (100 mL), and the suspe nsion was refluxed for 12 h. The reaction 118 mixture was cooled to room temperature, diluted wit h distilled water (~50 mL) and titrated with 12 M HCl until precipitation of the product ceased. The mixture was vacuum filtered, washed with distilled water, dried, and recrystallized fro m ethanol:water (70:30, v/v) to obtain 3 g (85% yield) of the desired product. 1H NMR (500 MHz, DMSO-d 6) : 12.49 (br s, 1 H), 9.91 (s, 1 H), 9.77 (s, 1 H), 7.99 (d, J = 7.1 Hz, 2 H), 7.59 (t, J = 7.3 Hz, 1 H), 7.54 (dd, J = 3.4, 8.8 Hz, 3 H), 7.51 (d, J = 7.3 Hz, 1 H), 6.77 (d, J = 8.7 Hz, 2 H, 13 C NMR (126 MHz, DMSO-d 6) : 166.6, 165.9, 158.8, 134.0, 133.7, 131.9, 131.8, 128.5, 12 8.3, 127.7, 127.5, 124.6, 124.0, 115.5, 115.4. (See Figure A 20 for crystallographic data). 5.2.8.2. Synthesis of (Z)-2-Benzamido-[3- 2H]-3-(4'-hydroxyphenyl)acrylic Acid The [3- 2H]-acrylic acid isotopomer was synthesized analogou sly to the unlabeled isomer (above), except [ -2H]-4'-Hydroxybenzaldehyde (0.62 g, 5 mmol) was used . Acetic anhydride (1.6 mL, 17 mmol), K 2HPO 4 (0.9 g, 5 mmol), and hippuric acid (0.9 g, 5 mmol) were varied to make the intermediate oxazolone (1.26 g, 82 % yield ). The oxazolone (1.23 g, 4 mmol) was saponified under reflux with ethanolic NaOH as befo re. The mixture was diluted with distilled water (25 ml), and the product was precipitated by the adding 12 M HCl at room temperature. The suspension was worked up and the product was re crystallized as described previously to obtain 1 g (88% yield) of product. 1H NMR (500 MHz, DMSO-d 6) : 12.49 (s, 1 H), 9.91 (s, 1 H), 9.77 (s, 1 H), 7.99 (d, J = 7.1 Hz, 1 H), 7.60 (t, J = 7.3 Hz, 1 H), 7.56 - 7.48 (m, 4 H), 6.77 (d, J = 8.8 Hz, 1 H). 13 C NMR (126 MHz, DMSO-d 6) : 166.6, 165.9, 158.8, 133.8, 131.9, 131.8, 131.6, 128.5, 128.4, 127.7, 127.5, 124.6, 123.9, 11 5.6, 115.4. (See Figure A 20 for crystallographic data). 119 5.2.8.3. Synthesis of (2S,3S)-[2,3- 2H2]- and (2S,3R)-[3- 2H]-1 The following procedure is based on previously desc ribed methods. 117 , 118 , 120 ( Z)-2- Benzamido-[3- 2H]-3-(4'-hydroxyphenyl) acrylic acid (1 g, 3.5 mmol ) and 10 mg of [Rh(NBD)(( R)-Prophos)]ClO 4 were dissolved in dry THF (25 mL) in a Parr reacto r. The reactor was successively evacuated and filled with H 2 gas and kept under H 2 gas (2.0 bar) for 16 h. The solvent was evaporated, followed by azeotropic remo val of residual THF with methanol. To the resultant yellow residue dissolved in methanol (15 mL) was added dry Dowex 50 W (100-200 mesh) cation exchange resin (1.5 g). The mixture wa s stirred until a clear solution was obtained, and then filtered. The retained resin was washed wi th warm methanol. All methanol filtrates were combined and dried under vacuum to obtain the crude N- benzoyl-(2 S,3 R)-[3- 2H]-tyrosine (0.9 g). The product was then refluxed with 40% HBr (aqueous) (10 mL) for 3 h. The mixture was cooled to room temperature, and the benzoic aci d crystals were removed by filtration. The filtrate was washed with diethyl ether (3 × 20 mL) to remove residual benzoic acid. The aqueous layer was lyophilized, and the crude product was di ssolved in 0.2 M NaOH (10 mL) and acidified with acetic acid to yield crystals after 1h at 0 °C. The product was isolated by vacuum filtration and the retentate was washed with ice co ld water to yield (2 S, 3R)-[3- 2H]- 1 as its zwitterionic amino acid (0.41 g, 70% yield). (2 S,3 S)-[2,3- 2H2]- 1 was synthesized by a procedure analogous to that de scribed for the synthesis of (2 S, 3R)-[3- 2H]- 1 with the following exceptions. ( Z)-2-Benzamido-3-(4'- hydroxyphenyl)acrylic acid was used instead of the unlabeled isotopomer, and D 2 gas was used 120 in place of H 2 gas for the reduction step to obtain (2 S,3 S)-[2,3- 2H2]- 1 as the zwitterionic amino acid (038 g, 65% yield). 5.2.9. Characterization of (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)-[3- 2H]-1 (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)-[3- 2H]- 1 (1 µmol of each) were separately derivatized to th eir 4'- O,2- N-di(ethoxycarbonyl) derivatives in water (1 mL) wit h ethyl chloroformate (50 µL, 0.5 mmol) and pyridine (50 µL, 0.6 mmol), and the sampl es were acidified to pH 2 (6M HCl) and extracted into diethyl ether (1 mL). After treatmen t of the ethanolic extract with diazomethane, the derivatives were isolated as a mixture of ethyl (10 mol%) and methyl (90 mol%) esters, as described before. The methyl esterified amino acid derivative was analyzed by GC/EIMS. The [2H]-labeled fragment ions originating from the ident ically derivatized (2 S)-[3,3- 2H2]- 1 were used to interpret the structures of various ions. T he relative ratio of diagnostic [ 2H]-labeled ions was compared to that of identical fragment ions of the unlabeled ( S)- 1 derivative to calculate the deuterium enrichment of labeled isotopomers: (2 S,3 S)-[2,3- 2H2]- 1 and (2 S,3 R)-[3- 2H]- 1 (Table A 1, and Figure A 28-31). The coupling constants ( J) for the geminal protons (H A and H B) against H X observed by 1H NMR for authentic ( S)- 1 in D 2O were JAX = 5.0 Hz at 3.23 (H A) and JBX = 7.8 Hz at 3.08. 121 In a previous study, the preferred conformation of (S)- 1 in water was established by NMR using stereospecifically deuterium isotopomers of 1 to correlate chemical shift and the magnitude of coupling constants with the conformation of the pro chiral hydrogens. 122 These earlier findings put the dihedral angle between the carboxy group an d the aromatic ring at 180°. In this 121 conformation, H A and H X were separated by ~ 60°, and H B and H X by 180°. Using this earlier data along with the magnitude of the J-values 122 observed for ( S)- 1, we assigned H A as pro -(3 S) and H B as pro -(3 R). This data was used to assign the absolute stereo chemistry at C of the synthetic (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)-[3- 2H]- 1. Authentic ( S)- 1 (10 mg) was dissolved in D 2O (1 mL), and transferred to a 5 mm diameter NMR tube, and analyzed by 1H NMR [(500 MHz, D 2O, 32 scans, 25 °C) : 7.23 (d, J = 8.5 Hz, 2 H), 6.93 (d, J = 8.5 Hz, 2 H), 3.97 (dd, J = 5.1, 7.8 Hz, 1 H), 3.23 (dd, J = 5.0, 14.7 Hz, 1 H), 3.08 (dd, J = 7.8, 14.7 Hz, 1 H)]. (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)-[3- 2H]- 1 (10 mg of each) were separately dissolved in D 2O (and H 2O) and analyzed by 1H NMR (and 2H NMR). (2 S,3 S)-[2,3- 2H2]- 1: 1H NMR (500 MHz, D 2O, 32 scans, 25 °C) : 7.23 (d, J = 8.5 Hz, 2 H), 6.93 (d, J = 8.5 Hz, 2 H), 3.07 (s, 1 H). 2H NMR (77 MHz ,H 2O, 512 scans, 25 °C) : 3.97 (bs, 1 H), 3.21 (bs, 1 H). (2 S,3 R)-[3- 2H]- 1: 1H NMR (500 MHz, D 2O, 32 scans, 25 °C) : 7.23 (d, J = 8.3 Hz, 2 H), 6.93 (d, J = 8.8 Hz, 2 H), 3.97 (d, J = 4.9 Hz, 1 H), 3.22 (d, J = 4.9 Hz, 1 H). 2H NMR (77 MHz, H2O, 512 scans, 25 °C) : 3.07 (bs, 1 H) 5.2.10. Synthesis of Authentic 4'- O,3- N- Di(( S)-2-methylbutanoyl) Methyl Esters of ( R)- and (S)-2 To a sample of ( R)-and ( S)- 2 (0.5 mol of each) dissolved in 50 mM phosphate buffer (1 mL) were added pyridine (50 L, 0.6 mmol) and ( S)-2-methylbutyric anhydride (10 L, 5 mol ) at 0 °C. The reactions were stirred for 5 min and treate d with another batch of pyridine (50 L) and (S)-2-methylbutyric anhydride (10 L) and stirred for 5 min. The mixtures were each ac idified 122 (pH 2 with 6 M HCl) and extracted with diethyl ethe r (3 × 2 mL). The ether fractions were combined, dried under vacuum, and the residue was d issolved in methanol (100 L). To this solution was added a dilute diazomethane solution d issolved in diethyl ether, until the yellow color persisted. The resultant 4'- O,3- N-di(( S)-2-methylbutanoyl)- -tyrosine methyl ester diastereoisomers were analyzed by GC/EIMS; their re tention times and fragment ions are noted in Figure A 38-40 5.2.11. Assessing the Stereospecificity of the C -Hydrogen Abstraction Catalyzed by Cc TAM (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)-[3- 2H]- 1 (each at 1 mM) were separately incubated with Cc TAM (0.1 mg) at 31 °C in 50 mM phosphate buffer (1 mL, pH 8.5) for 1 h. The tyrosine isomers were derivatized in situ with ethyl chloroformate (50 µL, 0.5 mmol) and pyr idine (50 µL, 0.6 mmol) to their 4'- O,N-di(ethoxycarbonyl) derivatives. The samples were a cidified to pH 2 (6M HCl) and extracted into diethyl ether (1 mL). After treatment of the extract with excess diazomethane, the amino acid derivatives were isola ted as a mixture of ethyl and methyl esters and analyzed by GC/EIMS, as before. Diagnostic frag ment ions from the ethyl ester derivative of the biosynthetic -2 were analyzed to determine the regiochemistry of t he deuterium atoms. 5.2.12. Assessing the Stereospecificity of the Hydrogen Reb ound at C Catalyzed by Cc TAM (2 S)-[3,3- 2H2]- 1 (5 mM) was incubated with Cc TAM (0.8 mg) at 31 °C in 50 mM phosphate buffer (40 mL, pH 8.5) for 36 h. The incubation mix ture was lyophilized, and the remaining residue was dissolved in methanol (7 mL) for 2H-NMR analysis or CD 3OD (7 mL) for 1H-NMR 123 analysis. As a control sample, authentic unlabeled ( R)- 2 (2 mM) in 50 mM phosphate buffer (40 mL, pH 8.5) was lyophilized, and the remaining resi due was dissolved in CD 3OD (7 mL) for 1H- NMR analysis. The magnitude of the coupling constan ts of the ABX spin system observed in the 1H NMR for authentic unlabeled ( R)- 2 in CD 3OD was used to assign the chemical shifts of the protons at the prochiral center of 2. This data was then used to assign the absolute stereochemistry at C of the biosynthetic [ 2H]- 2. 5.2.13. Assessing the Intramolecular Proton Transfer Step o f the Cc TAM Reaction (2 S)-[3,3- 2H2]- 1 (1 mM) was incubated with Cc TAM (0.1 mg) at 31 °C in 50 mM phosphate buffer (10 mL, pH 8.5). Aliquots (1 mL) were withdr awn from the reaction mixture at 10 min, then at 2, 4, 7, 12, 22, 33 and 45 h. The samples w ere derivatized in situ with ethyl chloroformate (50 µL, 0.5 mmol) and pyridine (50 µL, 0.6 mmol) to their 4'- O,N-di((ethoxycarbonyl)) derivatives, acidified to pH 2 (6M HCl), and extrac ted into diethyl ether (1 mL). After treatment of the ethanolic extract with diazomethane, the ami no acid derivatives were isolated as a mixture of ethyl and methyl esters. During the derivatizati on, 4'-hydroxycinnamate was converted to its 4'- O- (ethoxycarbonyl)-( E)-coumaric acid as an ethyl and methyl ester mixtur e. The methyl esters of the -tyrosine and ( E)-coumaric acid derivatives and the ethyl ester of the -tyrosine derivative were analyzed by GC/EIMS (see Figure A 4 4-46) for representative diagnostic fragment ions observed for the derivatized biosynth etic 2 and 4'-hydroxycinnamate in this time course study). GC/EIMS was used in selected-ion mod e to calculate the ion abundance ratio of [2H1](M) + ( m/z 354) and [ 2H2](M) + ( m/z 355) for the derivatives of the biosynthetic deute rium labeled 2. This ratio informed on the D H exchange during the reaction. The ion abundance o f 124 [2H2](M) + ( m/z 355) was corrected by subtracting the abundance of the naturally occurring 13 C- isotopomer (i.e., the [ 2H1](M + 1) + ( m/z 355) of the molecular ion [ 2H1](M) + ( m/z 354)) of the singly deuterium-labeled derivative. 5.2.14. Assessing the Effect of pH on Cc TAM Stereoselectivity (S)- 1 (1 mM) was incubated with Cc TAM (0.1 mg) separately at pHs 7, 8, and 9 (6 mL of each) in 50 mM phosphate buffer at 31 °C. Aliquots (1 mL) were withdrawn at 1, 2.5, 5, 11, and 25 h. The amino acids were derivatized in situ with ( S)-2-methylbutyric anhydride to form the 4'- O,3- N-di(( S)-2-methylbutanoyl)- 2 and methyl esterified with diazomethane and analyz ed by GC/EIMS. The sum of ion abundances for fragment ion s m/z 278 [M Œ methyl butyl] + (90% of base peak) and m/z 194 [M Œ (2 × methyl butyl) + H] + (base peak) for each enantiomer was compared to calculate the product ratio of ( R)- and ( S)- 2 in the sample. 5.2.15. Synthesis of ( R)-2 Methyl Ester In brief, to ( R)- 2 (18.1 mg, 0.1 mmol) dissolved in methanol (1 mL, 2 5 mmol) was added trimethylsilyl chloride (25 L, 0.2 mmol). 123 The suspension was stirred for 16 h and the methanol was evaporated in vacuo to obtain the (19.5 mg, quantitative yield). 1H NMR (500 MHz, CD 3OD) d: 7.29 (d, 8.5 Hz, 2 H), 6.85 (d, 8.5 Hz, 2 H), 4.6 3 (dd, J = 6.3, 7.8 Hz, 1 H), 3.69 (s, 3 H), 3.09 (dd, J = 7.8, 17.0 Hz, 1 H), 2.98 (dd, J = 6.3, 17.0 Hz, 1 H). 125 5.3. Results 5.3.1. Cc TAM Activity and Stereochemistry Cc TAM was expressed from the pET-28a(+) vector in Escherichia coli (BL21), and then Cc TAM and ( S)- 1 were incubated. The amino acids and hydroxycinnama te were derivatized and analyzed by GC/EIMS to show that product 2 (90 mol %) and by-product 4'-hydroxycinnamate (10 mol %) were formed (Figure A 40 and Figure A 41 ). The biosynthetic -tyrosine was also derivatized with a chiral auxiliary that indicated a mixture containing ( R)- 2 and ( S)- 2 at an 85:15 ratio (See Figure A 38 and Figure A 39). This produ ct distribution was consistent with that shown in an earlier study. 46 5.3.2. Assignment of the Prochiral Hydrogens of ( R)-2 by 1H-NMR The 1H-NMR of authentic unlabeled ( R)- 2 showed signals at 4.41 (dd, J = 4.4, 10.1 Hz, C -H), 2.73 (dd, J = 10.1, 16.6 Hz, C -H), 2.61 (dd, J = 4.4, 16.7 Hz, C -H) in CD 3OD solvent (Figure 5.1). According to the Karplus equation for 1H NMR, 124 3J coupling constants of vicinal protons are largest when the dihedral torsion angle ( ) is constrained at 0° (eclipsed conformation) or 180° (anti conformation). Smaller 3J coupling constants are observed when approaches 90°. Thus, the magnitudes of the 3J couplings ( Jeq-eq and Jeq-ax ( = 60°) between 3-4 Hz, and Jax-ax ( = 180°) between 8-13 Hz) of the conformationally-r estricted vicinal axial (ax) and equatorial 85 protons of cyclohexane 125 , 126 were used as a guide to assign the 1H NMR signals of ( R)- 2. The geminal C -H signals (designated as A and B spins) of the unl abeled ( R)- 2 126 spin-coupled with C -H (designated as the X spin) were identified by th eir distinct J-values [AX (3JAX = 10.1 Hz, 180°) and BX ( 3JBX = 4.4 Hz, 60°)] (Figure 5.4). Figure 5.1 . Partial 1H-NMR profile of unlabeled 2 and the 3J coupling constants for the ABX spin system of ( R)-2. 127 (2 S,3 R)-[2,3- 2H2]- 2 Biosynthetic Product a (2 S)-[3,3- 2H2]- 1 Substrate b (R)- 2 c Figure 5.2. a) 2H-NMR (after solvent exchange into CH 3OH); the relative area of the peaks at 4.41 and 2.61 are shown and b) 1H-NMR (after solvent exchange into CD 3OD) of a mixture 128 Figure 5.2 (cont'd). containing the remaining substrate (2 S)-[3,3- 2H2]-1 and the biosynthetic (2 S,3 R)-[2,3- 2H2]-2 after a Cc TAM-catalyzed reaction.c) 1H-NMR (in CD 3OD) of authentic ( R)-2. The signals for the prochiral protons of authentic ( R)-2 are aligned (boxes) with signals for the deuter ium labeled product in the biosynthetic sample. 5.3.3. Using NMR to Assess the Mechanism of the Hydrogen T ransfer at C in the Cc TAM Reaction The 2H-NMR chemical shifts of the propanoid side chain d euteriums at C and C of [ 2H2]-2 biosynthesized from (2 S)-[3,3- 2H2]- 1 by Cc TAM were at 2.61 (relative peak area = 0.7) and 4.41 (relative peak area = 1.0), respectively (Figu re 5.2a). The deuterium signals were compared to the 1H-NMR chemical shifts of the prochiral hydrogens of authentic ( R)- 2 (Figure 5.2a and Figure 5.2c) to assign the absolute stereochemistry of the [ 2H]-labeled biosynthetic sample. The 1H-NMR signal at 2.73 (singlet) produced by the biosynthetic [ 2H]-labeled ( R)- 2 (Figure 5.2b) coincided with the chemical shift of the pro -(2 R) proton of authentic ( R)- 2 (Figure 5.2c). Two nearby doublets were also observed at 2.75 (d, J = 16.8 Hz, 1 H) and 2.63 (d, J = 16.8 Hz, 1 H) with a peak area that was ~30% of the singlet a t 2.73 (Figure 5.2b). These resonances corresponded to a [ 2H]-labeled isotopomer of 2 containing two geminal hydrogens at C , creating an AB-coupled spin system. 129 5.3.4. Assessing the Mode of the Amino Group Attachment at C by Cc TAM. 5.3.4.1. Synthesis of (2S,3S)-[2,3- 2H2]- and (2S,3R)-[3- 2H]-(S)-1 The mode of the amino group transfer to C during the isomerization reaction catalyzed by Cc TAM was assessed. (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)-[3- 2H]- 1 were synthesized by stereospecific reduction of unlabeled and [3- 2H]-labeled ( Z)-2-benzamido-3-(4'- hydroxyphenyl)acrylic acid using a chiral [Rh(( R)-Prophos)(NBD)]ClO4 catalyst with deuterium or hydrogen gas, according to an earlier procedure. 117 The [ 2H]-labeled isotopomers of 1 were treated with a chiral auxiliary ( S)-2-methylbutyric anhydride and titrated with diazo methane to make the 4'- O,3- N-di(( S)-2-methylbutanoyl)- -tyrosine methyl ester derivatives. The retention times and mass spectrometry fragmentation of the sy nthesized [ 2H]-labeled ( S)- 1 derivatives observed by GC/EIMS analysis were identical to thos e of an identically derivatized sample of authentic ( S)- 1. 5.3.5. Analysis of 2 Made by Cc TAM Catalysis from (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)-[3- 2H]- 1 Cc TAM was incubated separately with (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)-[3- 2H]- 1. Afterwards, the reaction was basified, and the amino acids were derivatized to their 4'- O,3- N-di(ethoxycarbonyl) derivatives, acidified, extracted from the aqueous reaction buffer, and reacted with diazomethane to make the methyl esters. Intere stingly, during the basification step of the derivatization procedure, ethanol by-product from t he excess ethyl chloroformate caused partial ethyl esterification of the amino acids (10 Œ 30 mo l% compared to the methyl esters). 130 Table 5.1. EI-MS Fragmentation of 4'- O,3- N- Di((ethoxycarbonyl)) Ethyl Ester Derivatives of Labeled and [ 2H]-Labeled Biosynthesi zed Isotopomers of 2 m/z 353 Isotopomers of Substrate 1 Fragment ion F1: Cleavage at bond a Fragment ion F2: Cleavage at bond b Fragment ion F3: Cleavage at bonds b and c; H-transfer g to O a (2 S)-unlabeled m/z 280 m/z 266 m/z 194 b(2 S,3 S)-[2,3- 2H2] m/z 282 ( m/z 281 D H exchange) m/z 266 m/z 194 c (2 S,3 R)-[3- 2H] m/z 281 m/z 267 m/z 195 d(2 S)-[3,3- 2H2] m/z 282 ( m/z 281 D H exchange) m/z 267 m/z 195 gD-transfer was negligible compared to H-transfer ba sed on analysis of mass spectrometry fragments 131 The mass spectrum of the methyl ester derivatives of the -isomers yielded ions of diagnostic labeled fragments that were perturbed by overlappin g satellite fragment ions. Therefore, the ethyl esters of the -isomers were separated by GC, and the fates of the deuteriums on the -amino acids, originating from the substrate, were evaluat ed by the identity of diagnostic fragment ions made in the mass spectrometer (Figure A 41, Figure A 42, and Figure A 43 for fragment ions). The molecular ion [M+] ( m/z 353) of the 4'- O,3- N- di(ethoxycarbonyl) ethyl ester derivative of unlabeled ( R)- 2 fragmented into diagnostic ions F1A , F2A , and F3A (m/z 280, 266, and 194, respectively) (Table 1A). The corresponding fragmen t ions of the derivatized -amino acid biosynthesized by Cc TAM from (2 S,3 S)-[2,3- 2H2]- 1 were F1B , F2B , and F3B ( m/z 282, 266, and 194, respectively, Table 1b). Fragment ions F2B ( m/z 266) and F3B ( m/z 194) are identical to those of the unlabeled derivatized product (Table 1 a), indicating that deuterium was not at C of the biosynthetic 2. Further, the fragment ion F1B ( m/z 282, Table 1b) corresponding to the intact phenylpropanoid was shifted by two mass units above the similar fragment ion of the un labeled isotopomer. These data indicate that both deuterium s are at C of the biosynthetic product derived from (2 S,3 S)-[2,3- 2H2]- 1. Notably, the mass spectrum of the biosynthetic pr oduct 2 made from (2 S,3 S)-[2,3- 2H2]- 1 showed a molecular ion ([ 2H2](M)+, m/z 355) and ion m/z 354 ([ 2H1](M)+), one mass unit lower (Table 1b), indicating 36% deuterium was lost from the product compared to the substrate. This deuterium-t o-hydrogen (D H) exchange during the Cc TAM isomerization of (2 S,3 S)-[2,3- 2H2]- 1 was also supported by the ratio of diagnostic fragment ion F1B ( m/z 282) (Table 1b) and its [ F1B Œ 1] partner ( m/z 281), showing 32% deuterium loss. Conversely, when (2 S,3 R)-[3- 2H]- 1 was used as the substrate with Cc TAM, the resulting derivatized [ 2H]-labeled 2 yielded fragment ion F1C ( m/z 281) that was shifted one 132 mass unit over its unlabeled counterpart (Table 1c) , indicating retention of one deuterium in the propanoid side chain. Fragment ions F2C ( m/z 267) and F3C ( m/z 195) (Table 1c) were also one mass unit higher than the corresponding fragments d erived from the unlabeled product. The latter clarified that the deuterium at C of the substrate remained at this position after t he Cc TAM isomerization reaction to 2. 5.3.6. Assessing the D H Exchange Rate during Cc TAM Catalysis (2 S)-[3,3- 2H2]- 1 was incubated with Cc TAM in a time course experiment to assess the exten t of the D H exchange during the isomerization (Table 1d). The -amino acids isolated at designated time intervals were derivatized to their 4'- O,3- N- di(ethoxycarbonyl) ethyl esters and analyzed by GC/EIMS, as before. Tracking the molecu lar ion [ 2H2](M)+ ( m/z 355) and its [2H1](M)+ partner ( m/z 354) showed that the deuterium enrichment decrease d from 85% at 10 min to 60% after 12 h, under steady state reaction conditions. After 45 h, the reaction reached equilibrium (70% conversion of substrate to product s 4'-hydroxycinnamate and 2, and the D H exchange stabilized at 30% deuterium retention (see Table 1d, Figure A 43, Figure A 44, Figure A 45). Moreover, the relative abundances of fragment ions F2D ( m/z 267) and F3D ( m/z 195) (Table 1d) (both one mass unit above their unlabele d counterparts) from the derivatized biosynthetic product 2, at each time point, confirmed that a deuterium re mained at C . Thus, one of the geminal deuteriums of (2 S)-[3,3- 2H2]- 1 was 100% retained at C of the product, while the migratory deuterium was partially exchanged with hy drogen. This deuterium loss during the reaction was supported by the parallel decrease in the ratio between ion F1D ( m/z 282) (a fragment containing two deuteriums, one at C (exchangeable during catalysis), the other at C 133 (100% retained)) and ion [ F1D Œ 1 ] ( m/z 281) (a fragment containing one deuterium at C ) in the same spectrum (Table 1d and see Figure A 43 as refe rence). 020 40 60 80 100 050010001500200025003000 mol % Time (min) Figure 5.3. Plotted are the D ®H exchange ( ×) and (2 S)-[3,3- 2H2]- 1 ( ), [2,3- 2H2]- 2 ( ), and [3- 2H]-4'-hydroxycinnamic acid ( ) (as mol %) during the Cc TAM conversion of (2 S)-[3,3- 2H2]- 1 to labeled 2. 5.3.7. Re-evaluation of the Stereoisomeric Product Distrib ution Catalyzed by Cc TAM A previous study on Cc TAM reported that the enzyme stereoselectivity was 69% ee for ( R)- 2 at pH 8.8. 46 The earlier study however also showed that ( S)- 2 was produced at <5 mol% for the duration of the steady state (0 to 4 h) of the reac tion containing Cc TAM (10-50 µg), but increased significantly as the reaction entered equ ilibrium. 46 This interesting perturbation in the production of ( R)- and ( S)- 2 at equilibrium prompted us to re-evaluate this rea ction phenomenon. 134 Here, GC/EIMS was used to separate the diastereoiso meric 4'- O,3 -N-di(( S)-2-methylbutanoyl)- -tyrosine methyl ester derivatives of enantiomers of 2 produced by Cc TAM at steady state. Even after 1% conversion of the substrate, after 5 min, Cc TAM (50 µg) was already producing an enantiomeric mixture containing 80% ee of ( R)- 2 from 1 mM of substrate. The mixture approached 85:15 R:S (i.e., 70% ee of ( R)- 2) at a steady state rate just before reaching equilibrium at 120 min (Figure 5.5). The steady sta te production of ( S)- 2 (from 5 mol% to 15 mol%) over 2 h starkly contrasted the <5 mol% produ ction of ( S)- 2 prior to reaching equilibrium over 4 h, as reported earlier. 46 5.3.8. pH Effect on the Stereoselectivity of the Reaction Catalyzed by Cc TAM The Cc TAM reaction was incubated with ( S)- 1 separately at different pH values. After the conversion of ( S)- 1 to 2 reached 13, 35, and 56% at each pH, the amount of the ( S)-isomer (the minor antipode) relative to the total amount of 2 was calculated. The rate of the reaction was similar at pHs 8 and 9, yet half as fast at pH 7. A fter the conversion of substrate to product reached ~13% in each assay, the samples incubated at pHs 7, 8, and 9 contained 10.8%, 13.0%, and 16.7% ( S)- 2, respectively, relative to the ( R)-antipode (Figure 5.6). At higher percentages of conversion, the proportion of ( S)- 2 continued to increase, and the trend of increasing production of ( S)- 2 with higher pHs was also maintained (Figure 5.6). 135 5.4. Discussion 5.4.1. Retention of Configuration at the Migration Termini during the Cc TAM Reaction. 5.4.1.1. Amino Group Migration The stereochemistry of the major biosynthetic produ ct 2 made by Cc TAM was confirmed herein as (3 R); this was consistent with the assignment made in an earlier report. 46 The synthesis of stereospecifically deuterium labeled isotopomers (2 S,3 S)-[2,3- 2H2]- and (2 S,3 R)-[3- 2H]- 1 made it possible to assess the mode of attachment o f the amino group at C during the isomerization reaction catalyzed by Cc TAM. GC/EIMS profiles of derivatized isotopomers of 2 showed diagnostic fragment ions that informed on th e location of the deuteriums. Fragment ions F2B and F3B (Table 2b) revealed that the C deuterium of the (2 S,3 S)-[2,3- 2H2]- 1 substrate was replaced by the amino group en route to 2 (Table 1). Fragment ion F1B revealed that the deuterium migrated reciprocally to C (Table 2b), thus confirming that Cc TAM migrated the pro -( 3S ) hydrogen from C to C . Further, fragments ions F2C and F3C of the derivatized [ 2H]-labeled product (Table 2c) showed complementary that the C deuterium of the (2 S,3 R)-[3- 2H]-substrate was retained at its original position. Coupled with the known ( R)- 2 stereochemistry made in the aminomutase reaction, the deuterium remaining at C confirmed that Cc TAM uses a retention-of-configuration mechanism at the amino migration terminus. 136 5.4.1.2. Hydrogen Migration Stereochemistry The transient C -deuterium of (2 S,3 S)-[2,3- 2H2]- 1 migrates from C to C as the amino group moves reciprocally from C to C . The stereochemical mode of attachment of the deut erium to C was assessed by 1H- and 2H-NMR analyses of the zwitterion of the product dis solved in methanol. The magnitude of the difference between t he AX and BX coupling constants ( 3J 6 Hz) of ( R)- 2 was measure and compared to the experimentally cal culated 3J (~7 Hz) for the vicinal protons of the cyclohexane structure. 125 , 126 These similar 3J magnitudes calculated for these two structures suggested that the propionate hydrogens of ( R)- 2 are conformationally restricted as they are in cyclohexane. It can be im agined that a monodentate interaction between the ammonium cation and carboxylate anion of the ( R)- 2 zwitterion forms a pseudo-staggered six-membered ring to account for the magnitude of t he observed 3J (Figure 5.4). Figure 5.4. Intramolecular salt-bridge between the ammonium io n and carboxylate group of (R)- 2 in methanol. Dihedral angles ( 1 and 2) between H A and H X and between H B and H X, respectively, in the pseudo six-membered ring forme d by 2 are shown in Newman Projection. 137 To support the proposed restricted rotamer conforma tion, the carboxylate group of ( R)-2 was methyl esterified to disrupt the ionic ammonium/car boxylate ionic interaction. In acyclic systems, the conformational preference of hydrogens in an ABX spin system is commonly lower than in a restricted conformer. As an example, 3J(AX) (7.8 Hz) and 3J(BX) (6.3 Hz) coupling constants ( 3J(AX)Œ(BX) = 1.5 Hz) of the ( R)-2 methyl ester were nearly identical in CD 3OD. This observation suggested that H A and H B in the ( R)-2 methyl ester was conformationally more equivalent with respect to H X due to greater rotational freedom about the C -C bond. This likely resulted from removal of the intramolecular ionic interaction of underivatized ( R)- 2 (see Figure 5.4). The proposed six-membered ring conformation of ( R)- 2 places the side chain hydrogens of the amino acid in defined axial and equatorial posi tions by restricted rotation about the C -C bond (Figure 5.4). In addition, the carboxylate and aromatic ring of ( R)- 2 are considered to be positioned anti , according to a described lowest energy rotamer of 1.122 Therefore, based on the 1H-NMR chemical shifts and coupling constants, H A was definitively assigned as pro -(2 R) and HB as pro -(2 S) (see Figure 5.1). These chemical shift assignment s were used as reference in 1H- and 2H-NMR analyses of the deuterium-labeled biosyntheti c samples. Cc TAM shuttled one deuterium of (2 S)-[3,3- 2H2]- 1 from C and attached it to C with retention-of-configuration, placing the deuterium in the formal pro -(2 S) position of ( R)- 2. Previous stereochemical studies showed that a relat ed MIO-dependent phenylalanine aminomutase from Taxus plants ( Tc PAM) exchanges the position of the NH 2 and hydrogen migration partners of ( S)- -phenylalanine with retention of configuration at t he terminal 138 carbons. 5 Thus, the Cc TAM and Tc PAM reactions are likely mechanistically similar, 66 where both remove NH 2 and H from their substrates to form an acrylate in termediate. Apparently, these enzymes can rotate their intermediates 180° about t he C 1-C and C -C ipso bonds prior to the rebound of NH 2 and H to retain the stereochemistry in the corresp onding -products. 38 010 20 30 40 50 50 60 70 80 90 100 050100150200250300350 mol % of 2(R)- 2 (%) Time (min) Figure 5.5 . Analysis of the diastereomeric mixture of product s catalyzed by Cc TAM. Plotted are mol % of ( R)- 2 ( ) and ( S)- 2 ( ) relative to amount of ( S)- 1 added. The amount of ( R)- 2 (as %) () relative to the total amount of ( R)- and ( S)- 2 made at steady state. (Average of duplicate assays is plotted). It is important to note that ( R)- 2 produced by Cc TAM decreased from 90% to 80% (while ( S)- 2 increased accordingly) during the steady state pha se of the reaction from 5 to 120 min (Figure 5.5). The isotope enrichment in fragment ion cluste rs of the biosynthetic [ 2H]- 2 (Table 1) in the mass spectrometer did not indicate that the ( S)-isomer was labeled regioselectively different fro m (R)- 2. In addition, 2H-NMR analysis of the labeled mixture of 2 indicated isochronous signals for 139 the deuteriums at C of both enantiomers of 2 present in the reaction mixture. Therefore, the biosynthetic [ 2H]-( S)- 2 was labeled as the enantiomer of ( R)- -isomer. In contrast to the retention of configuration mechanism to access ( R)- 2, Cc TAM must use an inversion-of-configuration process to obtain the ( S)-isomer (Scheme 5.3) 0510 15 20 25 30 12.835.455.6 (S)- 2 (%) % Conversion pH 7 pH 8 pH 9 Figure 5.6. ( S)- 2 (as % of total 2) measured after substrate ( S)- 1 was depleted by 13%, 35%, and 56% at pHs 7, 8, and 9 while incubated with Cc TAM. 5.4.2. Hydrogen Exchange during Migration (2 S)-[3,3- 2H2]- 1 was incubated with Cc TAM in a time course study to track the D H exchange rate of the migratory pro- (3S ) deuterium during the isomerization reaction. The deuterium enrichment in the [ 2H]- 2 decreased from 85% at 5 min to 30% after 10 h as t he reaction reached equilibrium. This level of D H exchange was also observed for the ( R)- -phenylalanine product made in a similar reaction ca talyzed by the plant Tc PAM with (2 S)-[3,3- 140 2H2]- -phenylalanine as the substrate. 66 However, no deuterium exchange was observed in the reaction catalyzed by an isozyme from the bacteria Pantoea agglomerans ( Pa PAM) with (2 S)- [3,3- 2H2]- -phenylalanine substrate. 61 Structural data showed that the Pa PAM active site is slightly smaller and more ordered than Tc PAM. 38 , 114 Thus, bulk water likely can access the active site of Tc PAM better than that of Pa PAM. While the Cc TAM structure is not yet solved, it can be imagined that its active site, like that of Tc PAM, allows access by bulk water to account for the observed proton exchange during catalysis. The D H exchange rate progressively increases as the reac tion progresses under steady state conditions. A monoprotic Tyr52 residue, also report ed for other MIO aminomutases (and located on a flexible loop structure) 38 , 69 , 114 presumably, serves as the general base in the Cc TAM reaction. In this case, after deuterium abstraction from the (2 S)-[3,3- 2H2]- 1, the D H exchange on the catalytic Tyr52 should theoretically remain constant if all of the conditions stayed the same for the duration of the reaction; however, thi s is not the case. One notion to explain the observed increase in deuterium washout uses ureases as a model. Ureases are known to change the pH of the enzyme-microenvironment upon release of ammonia from urea. 127 , 128 Likewise, each of the known MIO-dependent arylalanine aminomu tases catalyzes an ammonia lyase reaction that releases ammonia and produces an aryl acrylate as a by-product (see Figure A 45). 38 , 46 , 61 , 115 Thus, the microenvironment of Cc TAM, containing increasing amounts of ammonia, might increase the pH and influence the proton exch ange rates during catalysis. It is conceivable that as the microenvironment ultimately reaches equ ilibrium with bulk water, the D H exchange becomes zero order as the reaction reaches equilibrium (see Figure A 45). To assess 141 this, Cc TAM was separated from the (2 S)-[3,3- 2H2]- 1 substrate and biosynthesized products after a 24-h time course study. The isolated enzyme was r e-incubated with more (2 S)-[3,3- 2H2]- 1, and the change in the deuterium washout rate was the sa me as observed before (see Figure A 45). Thus, the factors affecting the deuterium exchange were reversible, and the microenvironment was reset to its initial conditions. 5.4.3. Diastereomeric Product Ratio Catalyzed by Cc TAM Thus far, the MIO-dependent aminomutases can be cat egorized according to their stereoselectivity. Sg TAM and Pa PAM catalysis stereoselectively produce ( S)- -arylalanines. 6, 61 It can be imagined that these aminomutases follow a similar stereochemical course during removal and rebounding of the transient hydrogen an d NH 2 groups. Scheme 5.3: Inversion and Retention of Configuratio n Pathways Catalyzed by Cc TAM a (R)- 1 (S)- 1 (S)- 1 (S)- 1 a(Path a) Retention and ( Path b) inversion-of-configuration at C and C after exchange and reattachment of the pro -(3 S) proton (H ƒ) and the NH 2 of the ( S)- 1 substrate to make ( R)- 2 and (S)- 2, respectively by Cc TAM catalysis. 142 a) Pa PAM b) Sg TAM c) Tc PAM d) Cc TAM Figure 5.7. Comparison of the Tc PAM, Pa PAM, and Sg TAM active site structures co- crystallized with phenylpropanoid adducts or comple xes and the Cc TAM active site with 4'- hydroxycinnamate was modeled on the Tc PAM crystal structure (PDB# 3NZ4). The orientation of phenylpropanoid (center of each diagram) relativ e to the Arg residue (at right of each diagram) is shown. Also shown are key non-catalytic residues involved in binding and positioning the substrate; the catalytic tyrosine r esidue is above the plane in each drawing and is not shown. 143 The crystal structures of a phenylalanine aminomuta se ( Pa PAM) (Figure 5.7a) and a tyrosine aminomutase ( Sg TAM) (Figure 5.7b) show that the trajectories of th e phenylalanine and tyrosine substrates, respectively, are nearly identical. How ever, these aminomutases use distinct enzyme/substrate interactions to orient their subst rates. Sg TAM apparently uses residues His93 and Tyr415 to form a hydrogen bond network with the 4'-OH of the tyrosine substrate and place the carboxylate in a monodentate salt bridge intera ction with Arg311. 129 By contrast, these H-bond interactions are absent in Pa PAM, which has hydrophobic residues Val108 and Phe4 28 positioned analogously to His93 and Tyr415, respect ively, of Sg TAM. Instead, Pa PAM uses the steric bulk of Phe455 to force the phenylalanine su bstrate into a monodentate salt bridge interaction with Arg323. This steric interaction al igns phenylalanine in Pa PAM at a similar orientation as the enzyme/substrate pairing in Sg TAM. It was postulated that after elimination of the hydrogen and NH 2 from the Pa PAM substrate, the distinct angle somehow prevents rotation of the cinnamate intermediate. In this way, the hyd rogen and NH 2 can rebound to the same face of the intermediate from which they were removed. T he similar substrate docking conformations and trajectories of Pa PAM and Sg TAM likely, in part, define their identical (3 S)-product stereoselectivity. By analogy, Cc TAM (described herein) and Tc PAM from Taxus plants both make -amino acid product with (3 R)-stereochemistry; 5, 45 , 46 thus, these aminomutases likely progress through similar stereochemical profiles. E arlier structural analyses showed that Tc PAM has Asn458 positioned in contrast to the sterically larger Phe455 of isozyme Pa PAM. The smaller Asn of Tc PAM enables the carboxylate of the transient cinnam ate intermediate to engage in a bidentate salt bridge with Arg325 (Figure 5.7c ). In addition, the aryl portion of cinnamate sits in hydrophobic pocket that is, in part, compri sed of residues Leu108, and Ile431, which also help aligns the reaction intermediate in Tc PAM ~ 15° different from that of phenylalanine abou t 144 the C -axis in the Pa PAM 114 (Figure 5.7a). The altered angle presumably allows the Tc PAM intermediate to rotate 180° and retain the configur ation at the chiral and prochiral centers of the (R)- -phenylalanine product. Cc TAM probably also proceeds through an identical pro cess to obtain ( R)- 2. The Cc TAM structure has not yet been solved. However, its reaction stereochemistry (retention of configuration at both C and C in the major reaction product) is identical to tha t of Tc PAM. This suggested that the two enzymes have commo n active site architecture and likely position their reaction intermediates similarly. Th us, Cc TAM was modeled on Tc PAM (PDB# 3NZ4) (30% sequence similarity). Placing the 4'-hyd roxycinnamate in an orientation identical to that of cinnamate in Tc PAM shows the hydroxyaryl portion of the former int eracts with His81 and Tyr403 residues of Cc TAM (Figure 5.7d). The plausible hydrogen bonding between Cc TAM and the 4'-hydroxyaryl of the substrate likely help s align the substrate with Arg298 to form a bidentate salt bridge interaction (Figure 5.7d), si milar to the interaction in Tc PAM. As mentioned, a significant difference between the aminomutase isozymes is the TAMs use His and Tyr residues in the aromatic pocket to help bind and orient the tyrosine substrate through hydrogen bonding. The PAMs, instead, use hydrophobi c residues in the aromatic pocket and likely, other not yet fully understood factors to d irect the substrate binding orientation and stereoselectivity. It is feasible that hydrogen bon ding with the phenol of the substrate, in part, governs the stereoselectivity of the TAMs and enabl es the minor antipodal product to form through a new reaction route. As mentioned previous ly, each MIO-dependent arylalanine aminomutases catalyzes an ammonia lyase reaction, r eleasing ammonia and an arylacrylate as by-products (see Figure A 45). 38, 46, 61, 115 We posited earlier that the observed D H exchange for 145 the Cc TAM reaction was likely influenced by the increased pH of the microenvironment caused by ammonia-release (Figure A 45). Variation in bulk and local pH is known generally to disturb hydrogen bonding networks within an enzyme active s ite. 130 -132 In the TAM enzymes, the perturbation of the local pH could also affect the H-bonding network within the active site. Alteration of the hydrogen bonding network in the 4 '-hydroxyaryl binding pocket of the TAM enzymes could potentially change the enantioselecti vity, since this region likely contributes significantly to the substrate docking conformation . Therefore, in this study, the buffered pH of a Cc TAM reaction was adjusted to 9 and below to assess whether this would change the % ee of the product. A lower % conversion of ( S)- 1 to 2 correlated with a higher % ee of ( R)- 2 for each pH tested. Also, at higher pHs, the relative proportion of ( S)- 2 made by Cc TAM increased compared to ( R)- 2. Cc TAM consistently produced more ( S)- 2 enantiomer at higher pHs, regardless of the amount of ( S)- 1 converted to 2 (Figure 5.6). As we proposed earlier, ( S)- 2 is made via a conformer of a 4'-hydroxycinnamate intermediate that receives the NH 2 group and hydrogen (dwelling temporarily on the enzyme) on the same face they we re removed from ( S)- 1. ( R)- 2 is provided via a rotamer of the initially formed conformer of 4'-hydroxycinnamate. Therefore, since the production of ( S)- 2 increases with higher pH, these conditions can be imagined to restrict the rotation of the 4'-hydroxycinnamate intermediate wi thin the Cc TAM active site by possibly altering the substrate binding angle or by strength ening the hydrogen bond network with the 4'- OH of the substrate. Increases in pH are known to p artition the phenol groups of active site tyrosines to the phenolates and thereby shorten hyd rogen bond distances interacting with these oxyanions. 131 However, the mechanism through which the pH exactl y affects the product stereochemistry of Cc TAM remains unclear and will require further inquir y. 146 5.5. Co nclusion In summary, the synthesis of stereospecifically [ 2H]-labeled 1 enabled us to provide evidence to support that Cc TAM catalyzes its isomerization reaction along a pa th similar to that of Tc PAM from a plant. That is, both likely overcome a torsi onal barrier to access two pivotal rotameric intermediates. Cc TAM converts one rotamer to ( R)- 2 as the major product through retention of configuration. This pathway competes with a route i nvolving a second rotamer, producing ( S)- 2 as the minor enantiomer through inversion of config uration. Tc PAM uses only one rotamer to advance to a single enantiomer, ( R)- -phenylalanine. Thus, this study potentially starts to shed light o n how Cc TAM catalysis uses the 4'-hydroxyl group to orient the substrate in the active site. I n addition, this investigation provides a basis to probe the driving force that rotates the branch poi nt intermediate in Cc TAM to produce two enantiomers. 147 APPENDIX 148 050 100 150 200 03006009001200 1/ vo (min/nmol) [( E)-4'-Chlorocinnamate] ( mM) 0612 18 24 0150300450600 1/ vo (min/nmol) [( E)-4'-Methylcinnamate] ( mM) 0246803006009001200 1/ vo (min/nmol) [( E)-4'-Fluorocinnamate] ( mM) 0123403006009001200 1/ vo (min/nmol) [( E)-Cinnamate] ( mM) Figure A 1. The linear relationship between the reciprocal of t he steady-state rate (1/ vo) and the concentrations of competitive inhibitors ([I] o in eq. 4 above): Top Left) trans -4'-Chloro- (9), Top Right) trans -4'-Methyl- (10), Middle Left) trans -4'-Fluorocinnamate (11), Middle Right) trans -Cinnamate (12). 149 Figure A1 (cont'd) . 00.5 11.5 205001000150020002500 1/ vo (min/nmol) [( E)-2'-Thienylacrylate] ( mM) 00.5 11.5 205001000150020002500 1/ vo (min/nmol) [( E)-3'-Methylcinnamate] ( mM) The linear relationship between the reciprocal of t he steady-state rate (1/ vo) and the concentrations of competitive inhibitors ([I] o in eq. 4 above): Bottom Left) trans -2'- Thienylacrylate ( 13 ), and Bottom Right) trans -3'-Methylcinnamate ( 14 ) (each at 50, 100, 150, 300, 500, 1000, 2000 M) are shown. The inhibitors were co-incubated with Tc PAM and ( S)- styryl- -alanine ( 6), in triplicate. The production of (2 E,4 E)-Styrylacrylate ( 7) was monitored. The standard deviation for each calculated slope ( KI) was approximately ± 6% of the mean. 150 0246810 00.020.040.060.080.10.12 1/ vo (min .nmol -1 )1/[4'-Methyl- a-phenylalanine] ( mM-1 ) Figure A 2. Double Reciprocal Plots of the Rate of Conversion o f 8b/ to 8b/ and Concentration of Substrate ( S)-4'-Methyl- -phenylalanine ( 8b/ ) in Tc PAM Reactions Containing (2 E, 4E)- Styrylacrylate ( 7) at 0 ( ), 50 ( ), 100 ( ), 200 ( ) M to Assess the KI of 7. 151 a b Figure A 3. a) 1H-NMR spectrum ((500 MHz, CDCl 3) d: 7.68 (s, 1 H), 7.32 - 7.17 (m, 4 H), 2.91 (t, J = 8.4 Hz, 2 H), 2.64 (dt, J = 1.5, 8.5 Hz, 2 H))of and the b) 13 C-NMR spectrum ((126 MHz, CDCl 3) d: 172.2, 138.7, 137.2, 132.3, 129.9, 128.8, 128.4, 1 27.7, 126.8, 27.5, 21.9) of 16 -Acr 152 MŁ+ = 377 m/z Diagnostic Fragments a b Time (min) 12.813.013.213.413.613.814.0 020 40 60 80 100 13.72 13.75 Relative Abundance 50100150200250300350 400 020 40 60 80 100 377 318 240 180 149 109 125 83 Relative Abundance c d Time (min) 13.413.513.613.713.813.914.0 020 40 60 80 100 13.83 13.82 Relative Abundance m/z 50100150200250300350400 020 40 60 80 100 377 346 304 196 164 139 122 Relative Abundance Figure A 4 . Overlay of gas chromatography profiles of; a) N-[(1' S)-camphanoyl] methyl esters 153 Figure A 4 (cont'd). of (2 S)- -4'-fluorophenylalanine (13.72 min) derived from in termolecular amino group transfer by Tc PAM catalysis (solid line) and of the N-[(1' S)-camphanoyl] methyl ester of (2 S)- -4'- fluorophenylalanine (13.75 min) (dashed line); b) N-[(1' S)-camphanoyl] methyl ester of (3 R)- -4'-fluorophenylalanine (13.82 min) derived from int ermolecular amino group transfer catalysis of Tc PAM (solid line) and of the N-[(1' S)-camphanoyl] methyl ester of (3 R)- -4'- fluorophenylalanine(13.83 min) (dashed line). The m ass spectrometry profile of c) N-[(1' S)- of (2 S)- -4'-fluorophenylalanine (13.72 min), and d) N-[(1' S)-camphanoyl] methyl ester of (3 R)- -4'-fluorophenylalanine (13.82 min) and derived from intermolecular amino group transfer catalysis of Tc PAM. 154 a 149 m/z 109 m/z 318 m/z b 164 m/z 346 m/z 304 m/z Figure A 5 . GC EI/MS fragments of a) N-[(1' S)-camphanoyl] methyl ester of (2 S)- -4'- fluorophenylalanine and b) N-[(1' S)-camphanoyl] methyl ester of (3 R)- -4'- fluorophenylalanine 155 4'-Hydroxycinnamic acid MŁ+ = m/z 250 m/z 178 m/z 147 Cleavage at bond a; H-transfer to 4'- O Cleavage at bonds a, b; H-transfer to 4'- O m/z 50100150200250 Relative Abundance 020 40 60 80 100 250 178 147 219 206 119 161 91 Figure A 6. EI-MS fragmentation of 4'- O-(ethoxycarbonyl) methyl ester derivative of authentic 4'-hydroxycinnamic acid. Diagnostic fragm ent ions are m/z 178 and 147. 156 ˘ ˘ ˇ ˇ ˆ ˆ ˙ ˙ ˝ ˝ ˛˚˜ !"#$%˚!&'()) * +++++#,,""-˜&.#,"-˜&.˜'˚#/0$ ˜'˚#/0$ Figure A 7 . a) 1H NMR of 4'- O-(ethoxycarbonyl) methyl ester of 4'-hydroxycinnami c acid 1H NMR (500 MHz, CDCl 3) d: 7.68 (d, J = 16.1 Hz, 1 H), 7.55 (d, J = 8.3 Hz, 2 H), 7.22 (d, J = 8.3 Hz, 1 H), 6.41 (d, J = 16.1 Hz, 1 H), 6.41 (d, J = 16.1 Hz, 1 H), 4.34 (q, J = 7.3 Hz, 2 H), 3.82 (s, 3 H), 1.41 (t, J = 7.3 Hz, 3 H) 157 Figure A 8 . a) 13 C NMR of 4'- O-(ethoxycarbonyl) methyl ester of 4'-hydroxycinnami c acid 13 C NMR (126 MHz, CDCl 3) d: 167.2, 153.2, 152.4, 143.6, 132.2, 129.2, 121.6, 1 18.1, 65.1, 51.7, 14.2 158 (S)- 1 MŁ+ = m/z 339 m/z 250 m/z 178 m/z 107 Elimination of N-(ethoxycarbonyl) and C -H Elimination of N-(ethoxycarbonyl) and C -H; Cleavage at bond a; H- transfer to 4'- O Cleavage at bonds a and b; H-transfer to 4'- O m/z 50100150200250300 Relative Abundance 020 40 60 80 100 250 178 107 280 206 135 147 Figure A 9 . EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) methyl ester derivatives of unlabeled ( S)-1. Diagnostic fragment ions are m/z 250, 178, and 107. 159 1˘ˇˆ˙˝˛˚˜ !"#$%˚!&'()) * 23 +++++++#4, ," -˜&.˚!-˜˚!.",#,,#-˜.˚!& Figure A 10 . a) 1H NMR of 4'- O,2- N-di(ethoxycarbonyl) methyl ester of -tyrosine. 1H NMR (500 MHz, CDCl3) = 7.21 - 6.98 (m, 4 H), 5.10 (d, J = 7.3 Hz, 1 H), 4.67 - 4.55 (m, 1 H), 4.28 (q, J = 7.1 Hz, 2 H), 4.08 (q, J = 6.7 Hz, 2 H), 3.69 (s, 3 H), 3.10 (dd, J = 6.1, 14.0 Hz, 1 H), 3.05 (dd, J = 6.1, 14.0 Hz, 1 H), 1.36 (t, J = 7.0 Hz, 3 H), 1.20 (t, J = 6.7 Hz, 3 H). 160 ˛˚˜ !"#$%˚!&'()) * ˝˙ˆˇ ˘15 ˝˝ ˝ˆ4˝ˇ ˝˙ ++23 +++++˝˙ˆˇ˘˘1 5˝˝ ˝˙ ˝ˇ ˝ˆ Figure A 11 . a) 13 C NMR of 4'- O,2- N-di(ethoxycarbonyl) methyl ester of -tyrosine. 13 C NMR (126 MHz, CDCl3) = 171.9, 155.8, 153.5, 150.2, 133.6, 130.3, 121.0, 64.8, 61.2, 54.5, 52.3, 37.6, 14.4, 14.1 161 ˘ ˘ ˇ ˇ ˆ ˆ ˙ ˙ ˝ ˝ ˛˚˜ !"#$%˚!&'()) * +++++23 ++#,"-˜&.˚!67#,#,"-&.˜˚67!%!$!"0/ 89˜#:˜ ;#'˜9 Figure A 12. 1H NMR of 4'- O,2- N-di(ethoxycarbonyl) ethyl ester of -tyrosine. 1H NMR (500 MHz, CDCl3) = 7.17 (d, J = 8.6 Hz, 2 H), 7.14 - 7.11 (m, J = 8.6 Hz, 2 H), 5.14 (d, J = 7.8 Hz, 1 H), 4.63 (dd, J = 5.6, 13.4 Hz, 1 H), 4.33 (q, J = 7.1 Hz, 2 H), 4.17 (q, J = 7.2 Hz, 2 H), 4.12 (q, J = 7.1 Hz, 2 H), 3.11 (d, J = 5.1 Hz, 2 H), 1.40 (t, J = 7.1 Hz, 3 H), 1.25 (t, J = 7.3 Hz, 3 H), 1.24 (t, J = 7.2 Hz, 3 H) 162 Figure A 13. 13 C NMR of 4'- O,2- N-di(ethoxycarbonyl) ethyl ester of -tyrosine. 13 C NMR (126 MHz, CHCl 3) d: 171.7, 156.1, 153.8, 150.4, 133.9, 130.6, 121.3, 6 5.1, 61.8, 61.4, 54.9, 38.0, 14.7, 14.4, 14.3 163 9.09 min [M +Ł ] = 339 m/z 9.79 min [M +Ł ] = 353 m/z OONH OOOOOjedfkigaabbc OONH OOOOOjfdhligaabbke Figure A 14 a) GC profile of derivatized -tyrosine (with ethylchloroformate and CH 2N2 164 Figure A 14 (cont'd) . b c b) MS profiles of peak at 9.09 min retention time a nd c) of peak at 9.79 min retention time. 165 (R)- 2 MŁ+ = m/z 353 m/z 50100150200250300350 Realative Abundance 020 40 60 80 100 353 280 266 206 308 236 222 194 178 150 120 Figure A 15 . EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester derivatives of authentic ( R)-2. Diagnostic fragment ions are m/z 280, 266, and 194. 166 ˛˚˜ !"#$%˚!&'()) * +++++32 ++#,"-˜&.˚!67;#'˜9 #,#,"-˜&.˚!67 !67.&˜ Figure A 16 . a) 1H NMR of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester of -tyrosine. 1H NMR (500 MHz, CDCl 3) d: 7.34 (d, J = 8.5 Hz, 2 H), 7.16 (d, J = 8.8 Hz, 2 H), 4.33 (q, J = 7.3 Hz, 2 H), 4.12 (q, J = 7.1 Hz, 2 H), 4.09 (q, J = 7.3 Hz, 2 H), 2.92 - 2.78 (m, 2 H), 1.40 (t, J = 7.3 Hz, 3 H), 1.24 (t, J = 7.2 Hz, 3 H), 1.19 (t, J = 7.3 Hz, 3 H) 167 Figure A 17 . a) 13 C NMR of 4'- O, 3- N-di(ethoxycarbonyl) ethyl ester of -tyrosine. 13 C NMR (126 MHz, CDCl 3) d: 170.7, 155.8, 153.5, 150.3, 138.8, 127.4, 121.3, 6 4.8, 61.0, 60.7, 51.0, 40.6, 14.5, 14.1, 14.0 168 Figure A 18 . 1H NMR of 4'- O,3- N-di(ethoxycarbonyl) methyl ester of -tyrosine . 1H NMR (500 MHz, CDCl 3) = 7.29 (d, J = 8.5 Hz, 2 H), 7.12 (d, J = 8.5 Hz, 1 H), 5.66 (br. s., 1 H), 5.17 - 5.07 (m, 1 H), 4.28 (q, J = 6.9 Hz, 2 H), 4.08 (q, J = 7.3 Hz, 2 H), 3.60 (s, 3 H), 2.90 - 2.73 (m, 2 H), 1.36 (t, J = 7.3 Hz, 3 H), 1.20 (t, J = 7.0 Hz, 3 H) 169 Figure A 19. 13 C NMR of 4'- O,3- N-di(ethoxycarbonyl) methyl ester of -tyrosine. 13 C NMR (126 MHz, CD 3COCD 3) d: 170.7, 155.9, 153.4, 150.5, 140.3, 128.4, 121.7, 6 5.0, 64.4, 60.0, 51.0, 40.2, 14.1, 13.6 170 a Figure A 20 . Crystal structure of a) ( Z)-2-benzamido-[3- 2H]-3-(4'-hydroxyphenyl) acrylic acid and Color coding: Carbon (black), Hydrogen (Cyan), Deuterium (green) Oxygen (red), and Nitrogen (blue), 171 Figure A 20 (cont'd). b Crystal structure of b) ( Z)-2-benzamido-3-(4'-hydroxyphenyl) acrylic acid. Co lor coding: Carbon (black), Hydrogen (Cyan), Deuterium (green) Oxygen (red), and Nitrogen (blue), 172 Figure A 21 . 1H NMR spectrum of ( Z)-2-Benzamido-[3- 2H]-3-(4'-hydroxyphenyl)acrylic Acid. 1H NMR (500 MHz, DMSO-d 6) d: 12.49 (br. s., 1 H), 9.91 (s, 1 H), 9.77 (s, 1 H), 7.99 (d, J = 7.1 Hz, 2 H), 7.59 (t, J = 7.6 Hz, 1 H), 7.56 - 7.49 (m, 4 H), 6.77 (d, J = 8.3 Hz, 2 H) 173 Figure A 22 . 13 C NMR spectrum of ( Z)-2-Benzamido-[3- 2H]-3-(4'-hydroxyphenyl)acrylic Acid. 13 C NMR (126 MHz, DMSO-d 6) d: 166.6, 165.9, 158.8, 133.7, 133.4, 131.9, 131.8, 1 28.5, 128.3, 127.7, 127.5, 124.6, 123.9, 115.5, 115.4 174 Figure A 23 . 1H NMR spectrum of ( Z)-2-Benzamido-3-(4'-hydroxyphenyl)acrylic Acid. 1H NMR (500 MHz, DMSO-d 6) d: 12.49 (br. s., 1 H), 9.91 (br. s., 1 H), 9.82 - 9. 72 (m, 1 H), 7.99 (d, J = 7.3 Hz, 2 H), 7.59 (t, J = 7.1 Hz, 1 H), 7.56 - 7.47 (m, 4 H), 7.41 (s, 1 H ), 6.77 (d, J = 8.3 Hz, 2 H) 175 Figure A 24 . 13 C NMR spectrum of ( Z)-2-Benzamido-3-(4'-hydroxyphenyl)acrylic Acid. 13 C NMR (126 MHz, DMSO-d 6) d: 166.6, 165.9, 158.8, 134.0, 133.8, 131.9, 131.8, 1 28.5, 128.4, 127.7, 127.5, 124.6, 124.0, 115.5, 115.4 176 ˘ ˘ ˇ ˇ ˆ ˆ ˙ ˙ ˝ ˝ ˛˚˜ !"#$%˚!&'()) * "˝˝ ,ˇ˝ˆ #˙ <˚ ˛$+ ˇ"#,˛2˛$ ˆ˙ ˙ ˝ ˇ˝ ˇ˝ ˝ ˝ ˝ˇ Figure A 25 . 1H NMR spectrum of [Rh(NBD) 2]ClO 4. 1H NMR (500 MHz, CDCl 3) d: 5.20 (q, J = 2.2 Hz, 8 H), 4.15 - 4.11 (m, 4 H), 1.51 (t, J = 1.6 Hz, 4 H) 177 Figure A 26 . 1H NMR spectrum of [Rh(NBD)( R)-Prophos]ClO 4Ł0.5 CH 2Cl 2. Trace amounts of THF, H 2O, and silicon grease are present as impurities. 1H NMR (500 MHz, CDCl 3) = 7.79 - 7.73 (m, 2 H), 7.72 - 7.67 (m, 3 H), 7.65 - 7.61 (m , 2 H), 7.61 - 7.58 (m, 6 H), 7.58 - 7.55 (m, 6 H), 7.47 - 7.41 (m, 2 H), 7.35 - 7.29 (m, 2 H), 5.4 2 (br. s., 2 H), 5.31 (s, 1 H), 4.87 (br. s., 1 H), 4.28 (br. s., 1 H), 4.16 (br. s., 1 H), 2.71 - 2.59 (m, 2 H), 2.04 (td, J = 7.4, 12.5 Hz, 1 H), 1.84 - 1.76 (m, 2 H), 1.21 (dd, J = 6.5, 12.5 Hz, 3 H) 178 Figure A 27 . 13 C NMR spectrum of [Rh(NBD)( R)-Prophos]ClO 4Ł0.5 CH 2Cl 2. Trace amounts of THF, is present as impurities. 13 C NMR (126 MHz, DMSO-d 6) = 143.1, 135.1, 135.1, 134.8, 134.5, 134.1, 133.9, 132.8, 132.7, 132.0, 13 1.9, 131.7, 131.1, 131.0, 130.9, 130.1, 129.3, 129.3, 128.9, 128.8, 128.8, 128.6, 128.5, 128.2, 12 7.8, 125.9, 125.5, 63.4, 48.2, 34.2, 33.0, 14.7 179 Table A 1. EI-MS Fragment Ions of 4'- O,2- N- Di(ethoxycarbonyl) Methyl Ester Derivatives of Authentic and Synthesi zed Isotopomers of ( S)-1. m/z 339 Isotopomers of Substrate 1 Fragment ion F4: Cleavage at Bond e Ion abundance ratio a [F4] +:[F4 + Œ 1] +:[F4 Œ 2] + D atom % 100:0:0 0 a (2 S)-unlabeled m/z 280 280:279:278 93:5:2 98 b (2 S,3 S)-[2,3- 2H2] m/z 282 282:281:280 98:2:0 99 c (2 S,3 R)-[3- 2H] m/z 281 281:280:279 93:5:2 98 d (2 S)-[3,3- 2H2] m/z 282 282:281:280 aRatio of ion abundances of m/z [ F4 ]+:[ F4 - 1] +:[ F4 - 2] + for entries b, c, and d (where [F4 ]+ = [M + - 59] is calculated to determine deuterium enrichm ent). 180 (2 S,3 S)-[2,3- 2H2]- 1 MŁ+ = m/z 341 m/z 252 (251) m/z 180 (179) m/z 108 Elimination of N-(ethoxycarbonyl) and C -H Elimination of N-(ethoxycarbonyl) and C -H; Cleavage at bond a; H- transfer to 4'- O Cleavage at bonds a and b; H-transfer to 4'- O m/z 50100150200250300 Relative Abundance 020 40 60 80 100 282 252 180 149 136 108 208 Figure A 28 . EI-MS fragmentation of 4'- O,2- N-di(ethoxycarbonyl) methyl ester derivatives of (2 S,3 S)-[2,3- 2H2]-1. Diagnostic fragment ions are m/z 252, 180, and 108. 181 (2 S,3 R)-[3- 2H]- 1 MŁ+ = m/z 340 m/z 251 (250) m/z 179 (178) m/z 108 Elimination of N-(ethoxycarbonyl) and C -H Elimination of N-(ethoxycarbonyl) and C -H; Cleavage at bond a; H- transfer to 4'- O Cleavage at bonds a and b; H-transfer to 4'- O m/z 50100150200250300 Relative Abundance 020 40 60 80 100 281 251 207 179 148 136 108 Figure A 29 . EI-MS fragmentation of 4'- O,2- N-di(ethoxycarbonyl) methyl ester derivatives of (2 S,3 R)-[3- 2H]-1. Diagnostic fragment ions are m/z 251, 179, an d 108. 182 (2 S)-[3,3- 2H2]- 1 MŁ+ = m/z 341 m/z 251 m/z 179 m/z 109 Elimination of N-(ethoxycarbonyl) and C -H Elimination of N-(ethoxycarbonyl) and C -H; Cleavage at bond a; H- transfer to 4'- O Cleavage at bonds a and b; H-transfer to 4'- O m/z 50100150200250300 Relative Abundance 020 40 60 80 100 282 251 207 179 109 148 137 Figure A 30 . EI-MS fragmentation of 4'- O,2- N-di(ethoxycarbonyl) methyl ester derivatives of (2 S)-[3,3- 2H2]-1. Diagnostic fragment ions are m/z 251, 179, and 109. 183 Partial 1H-NMR Spectra a) b) c) ! !" ˘ ˇ # ! " ˘ ˜$%&'()%*+,$- " "ˇ " " ! ˘ ˘# ˘# ˘ˇ Figure A 31 . Partial NMR spectra of isotopomers of ( S)-1. 184 Figure A 31 (cont'd). Partial 2H-NMR Spectra d) e) f) Partial NMR spectra of isotopomers of ( S)-1. 185 Figure A 32 . 1H NMR spectrum of -tyrosine. 1H NMR (500 MHz, D 2O) d: 7.22 (d, J = 8.5 Hz, 2 H), 6.92 (d, J = 8.5 Hz, 2 H), 3.96 (dd, J = 5.1, 7.8 Hz, 1 H), 3.23 (dd, J = 5.0, 14.8 Hz, 1 H), 3.08 (dd, J = 7.8, 14.6 Hz, 1 H) 186 a ++32 ˆ2+ =2##,,=" ˛˚˜ !"#$%˚!&'()) * #, " b Figure A 33 . a) 1H NMR spectrum of (2 S,3 S)-[2,3- 2H2]- -tyrosine. 1H NMR (500 MHz, D2O) d: 7.22 (d, J = 8.5 Hz, 2 H), 6.92 (d, J = 8.5 Hz, 2 H), 3.08 (s, 1 H). b) 2H NMR spectrum of (2 S,3 S)-[2,3- 2H2]- -tyrosine. 2H NMR (77 MHz, H 2O) d: 3.97 (bs, 1 2H) 3.22 (bs, 1 2H). 187 a b Figure A 34 . a) 1H NMR spectrum of (2 S,3 R)-[3- 2H]- -tyrosine. 1H NMR (500 MHz, D 2O) d: 7.22 (d, J = 8.5 Hz, 2 H), 6.92 (d, J = 8.5 Hz, 2 H), 3.97 (d, J = 5.0 Hz, 1 H), 3.22 (d, J = 5.0 Hz, 1 H) b) 2H NMR spectrum of (2 S,3 R)-[3- 2H]- -tyrosine. 2H NMR (77 MHz, H 2O) d: 3.08 (bs, 1 2H). 188 ˛˚˜ !"#$%˚!&'()) * 2+ ++32 ˆ#,"-˜22##,,2"-˜ Figure A 35 . 1H NMR spectrum of -tyrosine. 1H NMR (500 MHz, CD 3OD) d: 7.25 (d, J = 8.8 Hz, 2 H), 6.81 (d, J = 8.8 Hz, 2 H), 4.41 (dd, J = 4.2, 10.0 Hz, 1 H), 2.73 (dd, J = 10.0, 16.6 Hz, 1 H), 2.61 (dd, J = 4.2, 16.6 Hz, 1 H) 189 a ˘ ˘ ˇ ˇ ˆ ˆ ˙ ˛˚˜ !"#$%˚!&'()) * 2+ ++32 ˆ=(2* 2##,,="#,"-(-* 8$>"˜90$ b Figure A 36 . a) 1H NMR and b) 2H NMR spectra of the reaction mixture of biosynthes ized products catalyzed by Cc TAM from (2 S)-[3,3- 2H2]- -tyrosine. 190 Figure A 37 . 1H NMR spectrum of -tyrosine methyl ester. 1H NMR (500 MHz, CD 3OD) d: 7.29 (td, J = 2.2, 8.5 Hz, 2 H), 6.85 (td, J = 2.2, 8.5 Hz, 2 H), 4.63 (dd, J = 6.3, 7.8 Hz, 1 H), 3.09 (dd, J = 7.8, 16.6 Hz, 1 H), 2.98 (dd, J = 6.3, 16.7 Hz, 1 H) 191 (R)- 2 MŁ+ = m/z 363 m/z 278 m/z 194 Cleavage at bond a Cleavage at bonds a, b; H-transfer to 4'- O a) Retention Time (min) 1718192021 Relative Abundance 014 29 43 57 71 86 100 18.68 19.06 (S)- 2(R)- 2 Figure A 38 . GC trace of 4'- O,3- N-di(( S)-2-methylbutanoyl) methyl ester derivatives of authentic ( S)-2 (18.68 min) and ( R)-2 (19.06 min) (a). EI-MS fragmentation of 4'- O,3- N-di(( S)-2-methylbutanoyl) methyl ester derivatives of aut hentic ( R)-2. Diagnostic fragment ions are m/z 278 and 194. 192 Figure A 38 (cont'd). b) m/z 50100150200250300350 Relative Abundance 020 40 60 80 100 278 194 363 332 206 179 136 122 57 (b). EI-MS fragmentation of 4'- O,3- N-di(( S)-2-methylbutanoyl) methyl ester derivative of authentic ( S)-2 was virtually identical. 193 (S)- 1 MŁ+ = m/z 363 m/z 278 m/z 194 Cleavage at bond a Cleavage at bonds a, b; H-transfer to 4'- O a) Retention Time (min) 1718192021 Relative Abundance 025 50 75 100 18.68 19.06 (S)- 2(R)- 2 Figure A 39 . GC trace of 4'- O,3- N-di(( S)-2-methylbutanoyl) methyl ester derivatives of ( S)-2 (18.68 min) and ( R)-2 (19.06 min) (a) and EI-MS fragmentation of 4'- O,3- N-di(( S)-2- methylbutanoyl) methyl ester derivatives of ( R)-2 biosynthesized by Cc TAM from unlabeled (S)-1. 194 (S)- 1 MŁ+ = m/z 250 m/z 178 m/z 147 Cleavage at bond a; H-transfer to 4'- O Cleavage at bonds a, b; H-transfer to 4'- O m/z 50100150200250 Relative Abundance 020 40 60 80 100 250 178 147 219 206 119 161 91 Figure A 40 . EI-MS fragmentation of 4'- O-(ethoxycarbonyl) methyl ester derivative of 4'- hydroxycinnamic acid biosynthesized by Cc TAM from unlabeled ( S)-1. 195 Figure A 40 (cont'd). b) m/z 50100150200250300350 Relative Abundance 020 40 60 80 100 363 332 278 206 179 194 136 122 57 Diagnostic fragment ions are m/z 178 and 147. (b). Diagnostic fragment ions are m/z 278 and 194). EI-MS fragmentation of 4'- O,3- N-di(( S)-2-methylbutanoyl) methyl ester derivatives of (S)-2 biosynthesized by Cc TAM from unlabeled ( S)-1 was virtually identical. 196 (S)- -Tyrosine (( S)- 1) MŁ+ = m/z 353 m/z 50100150200250300350 Relative Abundance 020 40 60 80 100 353 F1 280 F2 266 F3 194 208 120 178 150 308 236 222 Figure A 41 . EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester derivatives of 2 biosynthesized by Cc TAM from unlabeled ( S)-1. Diagnostic fragment ions F1, F2, and F3 (m/z 280, 266, and 194) are highlighted. 197 (2 S,3 S)-[2,3- 2H2]- 1 M Ł+ = m/z 355 m/z 50100150200250300350 Relative Abundance 020 40 60 80 100 355 F1 282 F2 266 F3 194 281 310 238 222 180 150 122 210 Figure A 42 . EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester derivatives of isotopomers of 2 biosynthesized by Cc TAM from (2 S,3 S)-[2,3- 2H2]-1. Diagnostic fragment ions F1, F2, and F3 (m/z 282, 266, and 194) are hig hlighted. 198 (2 S,3 R)-[3- 2H]- 1 MŁ+ = m/z 354 m/z 50100150200250300350 Relative Abundance 020 40 60 80 100 354 F1 281 F2 267 F3 195 309 237 223 209 151 178 123 Figure A 43 . EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester derivatives of isotopomers of 2 biosynthesized by Cc TAM from (2 S,3 R)-[3- 2H]-1. Diagnostic fragment ions F1, F2, and F3 (m/z 281, 267, and 195) are highligh ted. 199 (2 S)-[3,3- 2H2]- 1 M Ł+ = m/z 355 m/z 50100150200250300350 Relative Abundance 020 40 60 80 100 355 F1 282 281 F2 267 F3 195 210 238 223 179 151 123 310 Figure A 44 . EI-MS fragmentation of 4'- O,3- N-di(ethoxycarbonyl) ethyl ester derivatives of isotopomers of 2 biosynthesized by Cc TAM from (2 S)-[3,3- 2H2]-1. Diagnostic fragment ions F1, F2, and F3 (m/z 282, 267, and 195) are highligh ted. 200 (2 S)-[3,3- 2H2]- 1 MŁ+ = m/z 251 m/z 179 m/z 148 Cleavage at bond a; H-transfer to 4'- O Cleavage at bonds a, b; H-transfer to 4'- O m/z 50100150200250 Relative Abundance 020 40 60 80 100 148 179 251 220 207 120 162 92 Figure A 45 . EI-MS fragmentation of 4'- O-(ethoxycarbonyl) methyl ester derivative of biosynthesized [3- 2H]-4'-hydroxycinnamic acid biosynthesized by Cc TAM from (2 S)-[3,3- 2H2]-1. Diagnostic fragment ions are m/z 179 and 148. 201 REFERENCES 202 REFERENCES 1. Garcion, E.; Lamprecht, A.; Heurtault, B.; Paill ard, A.; Aubert-Pouessel, A.; Denizot, B.; Menei, P.; Benoit, J. P., A new generation of antic ancer, drug-loaded, colloidal vectors reverses multidrug resistance in glioma and reduces tumor progression in rats. Mol. Cancer Ther. 2006, 5, (7), 1710-1722. 2. von Nussbaum, F.; Spiteller, P.; Rüth, M.; Stegl ich, W.; Wanner, G.; Gamblin, B.; Stievano, L.; Wagner, F. E., An iron(III)Œcatechol complex as a mushroom pigment. Angew. Chem., Int. Ed. Engl. 1998, 37, (23), 3292-3295. 3. Crews, P.; Manes, L. V.; Boehler, M., Jasplakino lide, a cyclodepsipeptide from the marine sponge, Jaspis sp . Tetrahedron Lett. 1986, 27, (25), 2797-2800. 4. Christenson, S. D.; Liu, W.; Toney, M. D.; Shen, B., A novel 4-methylideneimidazole-5- one containing tyrosine aminomutase in enediyne ant itumor antibiotic C-1027 biosynthesis. J. Am. Chem. Soc. 2003, 125, (20), 6062-6063. 5. Rachid, S.; Krug, D.; Weissman, K. J.; Müller, R ., Biosynthesis of ( R)- b-tyrosine and its incorporation into the highly cytotoxic chondramide s produced by Chondromyces crocatus . J. Biol. Chem. 2007, 282, (30), 21810-21817. 6. Magarvey, N. A.; Fortin, P. D.; Thomas, P. M.; K elleher, N. L.; Walsh, C. T., Gatekeeping versus promiscuity in the early stages of the andrimid biosynthetic assembly line. ACS Chem. Biol. 2008, 3, (9), 542-554. 7. Benaglia, M.; Cinquini, M.; Cozzi, F., The S-thioester enolate/imine iondensation: a shortcut to -lactams. Eur. J. Org. Chem. 2000, 2000, (4), 563-572. 8. Fueloep, F.; Martinek, T. A.; Toth, G. K., Appli cation of alicyclic b-amino acids in peptide chemistry. Chem. Soc. Rev. 2006, 35, (4), 323-334. 9. Disney, M. D.; Hook, D. F.; Namoto, K.; Seeberge r, P. H.; Seebach, D., N-Linked glycosylated -peptides are resistant to degradation by glycoamid ase A. Chem. Biodiversity 2005, 2, (12), 1624-1634. 10. Wu, Y.-D.; Han, W.; Wang, D.-P.; Gao, Y.; Zhao, Y.-L., Theoretical analysis of secondary structures of -peptides. Acc. Chem. Res. 2008, 41, (10), 1418-1427. 11. Seebach, D.; Beck, A. K.; Bierbaum, D. J., The world of - and ?-peptides comprised of homologated proteinogenic amino acids and other com ponents. Chem. Biodiversity 2004, 1, (8), 1111-1239. 203 12. Duursma, A.; Minnaard, A. J.; Feringa, B. L., H ighly enantioselective conjugate addition of dialkylzinc reagents to acyclic nitroalkenes: A catalytic route to 2-amino acids, aldehydes, and alcohols. J. Am. Chem. Soc. 2003, 125, (13), 3700-3701. 13. Peña, D.; Minnaard, A. J.; de Vries, J. G.; Fer inga, B. L., Highly enantioselective rhodium-catalyzed hydrogenation of -dehydroamino acid derivatives using monodentate phosphoramidites. J. Am. Chem. Soc. 2002, 124, (49), 14552-14553. 14. Abele, S.; Seebach, D., Preparation of achiral and of enantiopure geminally disubstituted -amino acids for -peptide synthesis. Eur. J. Org. Chem. 2000, 2000, (1), 1-15. 15. Liu, M.; Sibi, M. P., Recent advances in the st ereoselective synthesis of -amino acids. Tetrahedron 2002, 58, (40), 7991-8035. 16. Cardillo, G.; Tomasini, C., Asymmetric synthesi s of b-amino acids and a-substituted b-amino acids. Chem. Soc. Rev. 1996, 25, (2), 117-128. 17. Podlech, J.; Seebach, D., The ArndtŒEistert rea ction in peptide chemistry: A facile access to homopeptides. Angew. Chem., Int. Ed. Engl. 1995, 34, (4), 471-472. 18. Jefford, C. W.; McNulty, J.; Lu, Z.-H.; Wang, J . B., The enantioselective synthesis of -amino acids, their -hydroxy derivatives, and the N-terminal components of bestatin and microginin. Helv. Chim. Acta 1996, 79, (4), 1203-1216. 19. Lakner, F. J.; Chu, K. S.; Negrete, G. R.; Kono pelski, J. P., Enantiomerically pure b-amino acids from 2- tert -butyl-1-carbomethoxy-2,3-dihydro-4(1 H)-pyrimidinone: ( R)-3- amino-3-( p-methoxyphenyl)propionic acid. Org. Synth. 1996, 73, 201-209. 20. Sibi, M. P.; Deshpande, P. K., A new methodolog y for the synthesis of b-amino acids. J. Chem. Soc., Perkin Trans. 1 2000, 0, (9), 1461-1466. 21. Fitzi, R.; Seebach, D., Resolution and use in -amino acid synthesis of imidazolidinone glycine derivatives. Tetrahedron 1988, 44, (17), 5277-5292. 22. d'Angelo, J.; Desmaële, D.; Dumas, F.; Guingant , A., The asymmetric Michael addition reactions using chiral imines. Tetrahedron: Asymmetry 1992, 3, (4), 459-505. 23. Zhu, G.; Chen, Z.; Zhang, X., Highly efficient asymmetric synthesis of -amino acid derivatives via rhodium-catalyzed hydrogenation of -(acylamino)acrylates. J. Org. Chem. 1999, 64, (18), 6907-6910. 24. Cimarelli, C.; Palmieri, G., Stereoselective re duction of enantiopure -enamino esters by hydride: A convenient synthesis of both enantiopure -amino esters. J. Org. Chem. 1996, 61, (16), 5557-5563. 25. Li, G.; Chang, H.-T.; Sharpless, K. B., Catalyt ic asymmetric aminohydroxylation (AA) of olefins. Angew. Chem., Int. Ed. Engl. 1996, 35, (4), 451-454. 204 26. Palomo, C.; Aizpurua, J. M.; Ganboa, I.; Oiarbi de, M., -Lactams as versatile intermediates in - and -amino acid synthesis. Synlett 2001, 2001, (12), 1813-1826. 27. Weiner, B.; Poelarends, G. J.; Janssen, D. B.; Feringa, B. L., Biocatalytic enantioselective synthesis of N-substituted aspartic acids by aspartate ammonia ly ase. Chem. Eur. J. 2008, 14, (32), 10094-10100. 28. Kim, J.; Kyung, D.; Yun, H.; Cho, B.-K.; Seo, J .-H.; Cha, M.; Kim, B.-G., Cloning and characterization of a novel -transaminase from Mesorhizobium sp . strain LUK: A new biocatalyst for the synthesis of enantiomerically p ure -amino acids. Appl. Environ. Microbiol. 2007, 73, (6), 1772-1782. 29. Wohlgemuth, R., Asymmetric biocatalysis with mi crobial enzymes and cells. Curr. Opin. Chem. Biol. 2010, 13, (3), 283-292. 30. Turner, N. J., Deracemisation methods. Curr. Opin. Chem. Biol. 2010, 14, (2), 115-121. 31. van Rantwijk, F.; Sheldon, R. A., Enantioselect ive acylation of chiral amines catalysed by serine hydrolases. Tetrahedron 2004, 60, (3), 501-519. 32. Gotor-Fernández, V.; Busto, E.; Gotor, V., Candida antarctica lipase B: An ideal biocatalyst for the preparation of nitrogenated org anic compounds. Adv. Synth. Catal. 2006, 348, (7-8), 797-812. 33. Altenbuchner, J.; Siemann-Herzberg, M.; Syldatk , C., Hydantoinases and related enzymes as biocatalysts for the synthesis of unnatu ral chiral amino acids. Curr. Opin. Biotechnol. 2001, 12, (6), 559-563. 34. Köhler, V.; Bailey, K. R.; Znabet, A.; Raftery, J.; Helliwell, M.; Turner, N. J., Enantioselective biocatalytic oxidative desymmetriz ation of substituted pyrrolidines. Angew. Chem., Int. Ed. Engl. 2010, 49, (12), 2182-2184. 35. Wanninayake, U.; DePorre, Y.; Ondari, M.; Walke r, K. D., ( S)-Styryl- a-alanine used to probe the intermolecular mechanism of an intramolec ular MIO-aminomutase. Biochemistry 2011, 50, (46), 10082-10090. 36. Koszelewski, D.; Tauber, K.; Faber, K.; Kroutil , W., @-Transaminases for the synthesis of non-racemic -chiral primary amines. Trends Biotechnol. 2010, 28, (6), 324-332. 37. Walker, K. D.; Floss, H. G., Detection of a phe nylalanine aminomutase in cell-free extracts of Taxus brevifolia and preliminary characterization of its reaction. J. Am. Chem. Soc. 1998, 120, (21), 5333-5334. 38. Feng, L.; Wanninayake, U.; Strom, S.; Geiger, J .; Walker, K. D., Mechanistic, mutational, and structural evaluation of a Taxus phenylalanine aminomutase. Biochemistry 2011, 50, (14), 2919-2930. 205 39. Klettke, K. L.; Sanyal, S.; Mutatu, W.; Walker, K. D., b-Styryl- and b-aryl- b-alanine products of phenylalanine aminomutase catalysis. J. Am. Chem. Soc. 2007, 129, (22), 6988-6989. 40. Szymanski, W.; Wu, B.; Weiner, B.; de Wildeman, S.; Feringa, B. L.; Janssen, D. B., Phenylalanine aminomutase-catalyzed addition of amm onia to substituted cinnamic acids: A route to enantiopure a- and b-amino acids. J. Org. Chem. 2009, 74, (23), 9152- 9157. 41. Wu, B.; Szyma Aski, W.; Wybenga, G. G.; Heberling, M. M.; Bartsch, S.; de Wildeman, S.; Poelarends, G. J.; Feringa, B. L.; Dijkstra, B. W.; Janssen, D. B., Mechanism-inspired engineering of phenylalanine aminomutase for enhanc ed -regioselective asymmetric amination of cinnamates. Angew. Chem., Int. Ed. Engl 2012, 51, (2), 482-486. 42. Ratnayake, N. D.; Walker, K. D., Arch. Biochem. Biophys 2013, In preparation. 43. Gloge, A.; Zon, J.; Kovari, A.; Poppe, L.; Rete y, J., Phenylalanine ammonia-lyase: The use of its broad substrate specificity for mechanis tic investigations and biocatalysis - synthesis of L-arylalanines. Chem. Eur. J. 2000, 6, (18), 3386-3390. 44. Takaç, S.; Akay, B.; Özdamar, T. H., Bioconvers ion of trans -cinnamic acid to L- phenylalanine by L-phenylalanine ammonia-lyase of Rhodotorula glutinis : Parameters and kinetics. Enzyme Microb. Technol. 1995, 17, (5), 445-452. 45. Walker, K. D.; Klettke, K.; Akiyama, T.; Crotea u, R., Cloning, heterologous expression, and characterization of a phenylalanine aminomutase involved in Taxol biosynthesis. J. Biol. Chem. 2004, 279, (52), 53947-53954. 46. Krug, D.; Müller, R., Discovery of additional m embers of the tyrosine aminomutase enzyme family and the mutational analysis of CmdF. ChemBioChem 2009, 10, (4), 741- 750. 47. MacDonald, M. J.; D™Cunha, G. B., A modern view of phenylalanine ammonia lyase. Biochem. Cell Biol. 2007, 85, (3), 273-282. 48. Kyndt, J. A.; Meyer, T. E.; Cusanovich, M. A.; Van Beeumen, J. J., Characterization of a bacterial tyrosine ammonia lyase, a biosynthetic en zyme for the photoactive yellow protein. FEBS Lett. 2002, 512, 240-244. 49. Baedeker, M.; Schulz, G. E., Structures of two histidine ammonia-lyase modifications and implications for the catalytic mechanism. Eur. J. Biochem. 2002, 269, (6), 1790- 1797. 50. Schwede, T. F.; Rétey, J.; Schulz, G. E., Cryst al structure of histidine ammonia lyase revealing a novel polypeptide modification as the c atalytic electrophile. Biochemistry 1999, 38, (17), 5355-5361. 206 51. Wickner, R. B., Dehydroalanine in histidine amm onia lyase. J. Biol. Chem. 1969, 244, (23), 6550-6552. 52. Röther, D.; Poppe, L.; Viergutz, S.; Langer, B. ; Retey, J., Characterization of the active site of histidine ammonia-lyase from Pseudomonas putida . Eur. J. Biochem. 2001, 268, (23), 6011-6019. 53. Tsien, R. Y., The green fluorescent protein. Annu. Rev. Biochem. 1998, 67, (1), 509-544. 54. Cooke, H. A.; Bruner, S. D., Probing the active site of MIO-dependent aminomutases, key catalysts in the biosynthesis of b-amino acids incorporated in secondary metabolites. Biopolymers 2010, 93, (9), 802-810. 55. Hanson, K. R.; Havir, E. A., L-Phenylalanine am monia-lyase: Evidence that the prosthetic group contains a dehydroalanyl residue a nd mechanism of action. Arch. Biochem. Biophys. 1970, 141, (1), 1-17. 56. Calabrese, J. C.; Jordan, D. B.; Boodhoo, A.; S ariaslani, S.; Vannelli, T., Crystal structure of phenylalanine ammonia lyase: Multiple helix dipo les implicated in catalysis. Biochemistry 2004, 43, (36), 11403-11416. 57. Christianson, C. V.; Montavon, T. J.; Festin, G . M.; Cooke, H. A.; Shen, B.; Bruner, S. D., The mechanism of MIO-based aminomutases in b-amino acid biosynthesis. J. Am. Chem. Soc. 2007, 129, (51), 15744-15755. 58. Schuster, B.; Rétey, J., The mechanism of actio n of phenylalanine ammonia lyase: The role of prosthetic dehydroalanine. Proc. Natl. Acad. Sci. U. S. A. 1995, 92, (18), 8433- 8437. 59. Poppe, L.; Rétey, J., FriedelŒCrafts-type mecha nism for the enzymatic elimination of ammonia from histidine and phenylalanine. Angew. Chem., Int. Ed. Engl. 2005, 44, (24), 3668-3688. 60. Bartsch, S.; Bornscheuer, U. T., A single resid ue influences the reaction mechanism of ammonia lyases and mutases. Angew. Chem., Int. Ed. Engl. 2009, 48, (18), 3362-3365. 61. Ratnayake, N. D.; Wanninayake, U.; Geiger, J. H .; Walker, K. D., Stereochemistry and mechanism of a microbial phenylalanine aminomutase. J. Am. Chem. Soc. 2011, 133, (22), 8531-8533. 62. Christenson, S. D.; Wu, W.; Spies, M. A.; Shen, B.; Toney, M. D., Kinetic analysis of the 4-methylideneimidazole-5-one-containing tyrosine am inomutase in enediyne antitumor antibiotic C-1027 biosynthesis. Biochemistry 2003, 42, (43), 12708-12718. 63. Montavon, T. J.; Christianson, C. V.; Festin, G . M.; Shen, B.; Bruner, S. D., Design and characterization of mechanism-based inhibitors for the tyrosine aminomutase Sg TAM. Bioorg. Med. Chem. Lett. 2008, 18, (10), 3099-3102. 207 64. Strom, S.; Wanninayake, U.; Ratnayake, N. D.; W alker, K. D.; Geiger, J. H., Insights into the mechanistic pathway of the Pantoea agglomerans phenylalanine aminomutase. Angew. Chem., Int. Ed. Engl. 2012, 51, 11-17. 65. Wanninayake, U.; Walker, K. D., Assessing the d eamination rate of a covalent aminomutase adduct by burst phase analysis. Biochemistry 2012, 51, (26), 5226-5228. 66. Mutatu, W.; Klettke, K. L.; Foster, C.; Walker, K. D., Unusual mechanism for an aminomutase rearrangement: Retention of configurati on at the migration termini. Biochemistry 2007, 46, (34), 9785-9794. 67. Wanninayake, U.; Walker, K. D., A bacterial tyr osine aminomutase proceeds through retention or inversion of stereochemistry to cataly ze its isomerization reaction. J. Am. Chem. Soc. 2013, 135, (30), 11193-11204. 68. Chesters, C.; Wilding, M.; Goodall, M.; Micklef ield, J., Thermal bifunctionality of bacterial phenylalanine aminomutase and ammonia lya se enzymes. Angew. Chem., Int. Ed. Engl. 2012, 51, (18), 4344-8. 69. Christianson, C. V.; Montavon, T. J.; Van Lanen , S. G.; Shen, B.; Bruner, S. D., The structure of L-tyrosine 2,3-aminomutase from the C- 1027 enediyne antitumor antibiotic biosynthetic pathway. Biochemistry 2007, 46, (24), 7205-7214. 70. Louie, G. V.; Bowman, M. E.; Moffitt, M. C.; Ba iga, T. J.; Moore, B. S.; Noel, J. P., Structural determinants and modulation of substrate specificity in phenylalanine-tyrosine ammonia-lyases. Chem. Biol. 2006, 13, (12), 1327-1338. 71. Ritter, H.; Schulz, G. E., Structural basis for the entrance into the phenylpropanoid metabolism catalyzed by phenylalanine ammonia-lyase . Plant Cell 2004, 16, (12), 3426- 3436. 72. Wu, B.; Szymanski, W.; Wietzes, P.; de, W. S.; Poelarends, G. J.; Feringa, B. L.; Janssen, D. B., Enzymatic synthesis of enantiopure a- and b-amino acids by phenylalanine aminomutase-catalysed amination of cinnamic acid de rivatives. ChemBioChem 2009, 10, (2), 338-344. 73. Ege, M.; Wanner, K. T., Synthesis of b-amino acids based on oxidative cleavage of dihydropyridone derivatives. Org. Lett. 2004, 6, (20), 3553-3556. 74. Cox, B. M.; Bilsborrow, J. B.; Walker, K. D., E nhanced conversion of racemic a-arylalanines to ( R)- b-arylalanines by coupled racemase/aminomutase catal ysis. J. Org. Chem. 2009, 74, (18), 6953-6959. 75. O'Maille, P. E.; Tsai, M. D.; Greenhagen, B. T. ; Chappell, J.; Noel, J. P., Gene library synthesis by structure-based combinatorial protein engineering. In Protein Engineering , Elsevier Academic Press Inc: San Diego, 2004; Vol. 388, pp 75-91. 208 76. Dunkley, K. D.; Callaway, T. R.; Chalova, V. I. ; Anderson, R. C.; Kundinger, M. M.; Dunkley, C. S.; Nisbet, D. J.; Ricke, S. C., Growth and genetic responses of Salmonella typhimurium to, pH-shifts in an anaerobic continuous culture. Anaerobe 2008, 14, (1), 35-42. 77. Turner, N. J., Ammonia lyases and aminomutases as biocatalysts for the synthesis of a-amino and b-amino acids. Curr. Opin. Chem. Biol. 2011, 15, (2), 234-240. 78. Moffitt, M. C.; Louie, G. V.; Bowman, M. E.; Pe nce, J.; Noel, J. P.; Moore, B. S., Discovery of two cyanobacterial phenylalanine ammon ia lyases: Kinetic and structural characterization. Biochemistry 2007, 46, (4), 1004-1012. 79. Poppe, L.; Pilbak, S.; Paizs, C.; Retey, J., Me chanistic aspects and biocatalytic implications of the MIO-containing ammonia-lyase/am inomutase family. Stud. Univ. Babes-Bolyai, Chem. 2008, 53, (2), 15-19. 80. Sawada, S.; Kumagai, H.; Yamada, H.; Hill, R. K .; Mugibayashi, Y.; Ogata, K., Stereochemistry of ammonia elimination from L-tyros ine with L-phenylalanine ammonia-lyase. Biochim. Biophys. Acta, Enzymol. 1973, 315, (1), 204-7. 81. Hanson, K. R.; Havir, E. A., L-Phenylalanine am monia-lyase. Evidence that the prosthetic group contains a dehydroalanyl residue a nd mechanism of action. Arch. Biochem. Biophys. 1970, 141, (1), 1-17. 82. Ellis, B. E.; Zenk, M. H.; Kirby, G. W.; Michae l, J.; Floss, H. G., Steric course of the tyrosine ammonia-lyase reaction. Phytochemistry 1973, 12, (5), 1057-8. 83. Thompson, J. D.; Gibson, T. J.; Higgins, D. G., Multiple sequence alignment using ClustalW and ClustalX. In Curr Protoc Bioinformatics , 2008/09/17 ed.; 2002; Vol. Chapter 2, p Unit 2 3. 84. Kvicala, J.; Hrabal, R.; Czernek, J.; Bartosova , I.; Paleta, O.; Pelter, A., Low-temperature F19 NMR spectroscopy of 1-fluoro-1-lithioethenes - sta bility, shifts and unexpected coupling constants. J. Fluorine Chem. 2002, 113, (2), 211-218. 85. Sayers, E. W.; Barrett, T.; Benson, D. A.; Brya nt, S. H.; Canese, K.; Chetvernin, V.; Church, D. M.; DiCuccio, M.; Edgar, R.; Federhen, S .; Feolo, M.; Geer, L. Y.; Helmberg, W.; Kapustin, Y.; Landsman, D.; Lipman, D. J.; Madd en, T. L.; Maglott, D. R.; Miller, V.; Mizrachi, I.; Ostell, J.; Pruitt, K. D.; Schule r, G. D.; Sequeira, E.; Sherry, S. T.; Shumway, M.; Sirotkin, K.; Souvorov, A.; Starchenko , G.; Tatusova, T. A.; Wagner, L.; Yaschenko, E.; Ye, J., Database resources of the na tional center for biotechnology information. Nucleic Acids Res. 2009, 37, (suppl 1), D5-D15. 86. Xiang, L.; Moore, B. S., Inactivation, compleme ntation, and heterologous expression of EncP, a novel bacterial phenylalanine ammonia-lyase gene. J. Biol. Chem. 2002, 277, (36), 32505-32509. 209 87. Xiang, L.; Moore, B. S., Biochemical characteri zation of a Prokaryotic phenylalanine ammonia lyase. J. Bacteriol. 2005, 187, (12), 4286-4289. 88. Grecu, T., Masters of philosophy thesis, University of Manches ter (UK) 2010 . 89. Oclarit, J. M.; Okada, H.; Ohta, S.; Kaminura, K.; Yamaoka, Y.; Iizuka, T.; Miyashiro, S.; Ikegami, S., Anti- Bacillus substance in the marine sponge, Hyatella species, produced by an associated Vibrio species Bacterium . Microbios 1994, 78, (314), 7-16. 90. Koksal, M.; Jin, Y.; Coates, R. M.; Croteau, R. ; Christianson, D. W., Axadiene synthase structure and evolution of modular architecture in terpene biosynthesis. Nature 2011, 469, (7328), 116-120. 91. Jennewein, S.; Wildung, M. R.; Chau, M.; Walker , K.; Croteau, R., Random sequencing of an induced Taxus cell cDNA library for identification of clones invo lved in taxol biosynthesis. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, (24), 9149-9154. 92. Creighton, T. E., Proteins. Structure and molecular principles . 2nd ed.; Freeman: NewYork, 1984; p 515 pp. 93. Wu, B.; Szymanski, W.; Wijma, H. J.; Crismaru, C. G.; de Wildeman, S.; Poelarends, G. J.; Feringa, B. L.; Janssen, D. B., Engineering of an enantioselective tyrosine aminomutase by mutation of a single active site res idue in phenylalanine aminomutase. Chem. Commun. (Cambridge, U. K.) 2010, 46, (43), 8157-8159. 94. Takaç, S.; Akay, B.; Bzdamar, T. H., Bioconversion of trans -cinnamic acid to L- phenylalanine by L-phenylalanine ammonia-lyase of Rhodotorula glutinis : Parameters and kinetics. Enzyme Microb. Technol. 1995, 17, (5), 445-452. 95. Fersht, A., Enzyme structure and mechanism. W.H Freeman and Company 1985, second ed., (Chapter 4 ), 128-147. 96. Lipshutz, B. H.; Ghorai, S.; Boıkovi C, . V., Tandem olefin metathesis-elimination reactions. A new route to doubly unsaturated carbon yl derivatives. Tetrahedron 2008, 64, (29), 6949-6954. 97. Gutfreund, H.; Sturtevant, J. M., Mechanism of the reaction of chymotrypsin with p-nitrophenyl acetate. Biochem. J. 1956, 63, (4), 656-661. 98. Ahn, H. J.; Kim, Y. S.; Kim, J.-U.; Han, S. M.; Shin, J. W.; Yang, H. O., Mechanism of taxol-induced apoptosis in human Skov3 ovarian carc inoma cells. J. Cell. Biochem. 2004, 91, (5), 1043-1052. 99. Belyaev, N. A.; Kolesanova, E. F.; Kelesheva, L . F.; Rotanova, T. V.; Panchenko, L. F., Effect of the aminopeptidase inhibitor bestatin on rat-brain enkephalin levels. Bull. Exp. Biol. Med. 1990, 110, (11), 1483-1485. 210 100. Prabhakaran, P. C.; Woo, N. T.; Yorgey, P. S.; Gould, S. J., Biosynthesis of blasticidin S from L- a-arginine. Stereochemistry in the arginine-2,3-amin omutase reaction. J. Am. Chem. Soc. 1988, 110, (17), 5785-5791. 101. Liu, W.; Christenson, S. D.; Standage, S.; She n, B., Biosynthesis of the enediyne antitumor antibiotic C-1027. Science 2002, 297, (5584), 1170-1173. 102. Yin, X. H.; O'Hare, T.; Gould, S. J.; Zabriski e, T. M., Identification and cloning of genes encoding viomycin biosynthesis from Streptomyces vinaceus and evidence for involvement of a rare oxygenase. Gene 2003, 312, 215-224. 103. Weaver, D. F.; Tan, C. Y. K.; Kim, S. T.; Kong , X.; Wei, L.; Carran, J. R. Antiepileptogenic agents. WO 2002073208, March 13, 2002. 104. Georg, G. I., The Organic Chemistry of b-Lactams . Wiley-VCH, New York, N. Y.: 1993; p 381 pp. 105. Gademann, K.; Kimmerlin, T.; Hoyer, D.; Seebac h, D., Peptide folding induces high and selective affinity of a linear and small b-peptide to the human somatostatin receptor 4. J. Med. Chem. 2001, 44, (15), 2460-2468. 106. Sonti, R.; Gopi, H. N.; Muddegowda, U.; Ragoth ama, S.; Balaram, P., A Designed Three-Stranded -Sheet in an / Hybrid Peptide. Chem. Eur. J. 2013, 19, (19), 5955-65. 107. Murray, J. K.; Farooqi, B.; Sadowsky, J. D.; S calf, M.; Freund, W. A.; Smith, L. M.; Chen, J.; Gellman, S. H., Efficient synthesis of a b-peptide combinatorial library with microwave irradiation. J. Am. Chem. Soc. 2005, 127, (38), 13271-13280. 108. Ibrahem, I.; Rios, R.; Vesely, J.; Zhao, G.-L. ; Cordova, A., Catalytic enantioselective 5- hydroxyisoxazolidine synthesis: an asymmetric entry to -amino acids. Synthesis 2008 , (Copyright (C) 2013 American Chemical Society (ACS) . All Rights Reserved.), 1153- 1157. 109. Davies, S. G.; Mulvaney, A. W.; Russell, A. J. ; Smith, A. D., Parallel synthesis of homochiral -amino acids. Tetrahedron: Asymmetry 2007, 18, (13), 1554-1566. 110. Tan, C. Y. K.; Weaver, D. F., A one-pot synthe sis of 3-amino-3-arylpropionic acids. Tetrahedron 2002, 58, (37), 7449-7461. 111. Christianson, C. V.; Montavon, T. J.; Festin, G. M.; Cooke, H. A.; Shen, B.; Bruner, S. D., The mechanism of MIO-based aminomutases in b-amino acid biosynthesis. J. Am. Chem. Soc. 2007, 129, (51), 15744-15745. 112. Asano, Y.; Kato, Y.; Levy, C.; Baker, P.; Rice , D., Structure and function of amino acid ammonia-lyases. Biocatal. Biotransform. 2004, 22, (2), 131-138. 113. Rettig, M.; Sigrist, A.; Rétey, J., Mimicking the reaction of phenylalanine ammonia lyase by a synthetic model. Helv. Chim. Acta 2000, 83, (9), 2246-2265. 211 114. Strom, S.; Wanninayake, U.; Ratnayake, N. D.; Walker, K. D.; Geiger, J. H., Insights into the Mechanistic Pathway of the Pantoea agglomerans Phenylalanine Aminomutase. Angew. Chem. Int. Ed. 2012, 51, 11-17. 115. Christenson, S. D.; Wu, W.; Spies, M. A.; Shen , B.; Toney, M. D., Kinetic Analysis of the 4-Methylideneimidazole-5-one-Containing Tyrosin e Aminomutase in Enediyne Antitumor Antibiotic C-1027 Biosynthesisâ• Biochemistry 2003, 42, (43), 12708-12718. 116. Sasse, F.; Kunze, B.; Gronewold, T. M. A.; Rei chenbach, H., The chondramides: Cytostatic agents from Myxobacteria acting on the actin cytoskeleton. Journal of the National Cancer Institute 1998, 90, (20), 1559-1563. 117. Fryzuk, M. D.; Bosnich, B., Asymmetric synthes is. An asymmetric homogeneous hydrogenation catalyst which breeds its own chirali ty. J. Am. Chem. Soc. 1978, 100, (17), 5491-5494. 118. Fryzuk, M. D.; Bosnich, B., Asymmetric synthes is. Production of optically active amino acids by catalytic hydrogenation. J. Am. Chem. Soc. 1977, 99, (19), 6262-6267. 119. Cleary, T.; Brice, J.; Kennedy, N.; Chavez, F. , One-pot process to Z-a-benzoylamino- acrylic acid methyl esters via potassium phosphate- catalyzed Erlenmeyer reaction. Tetrahedron Lett. 2010, 51, (4), 625-628. 120. Kirby, G. W.; Michael, J., Stereoselective b-labelling of aromatic amino-acids with deuterium and tritium. J. Chem. Soc., Perkin Trans. 1 1973, 0, (0), 115-120. 121. Jin, M.; Fischbach, M. A.; Clardy, J., A biosy nthetic gene cluster for the acetyl-CoA carboxylase inhibitor andrimid. J. Am. Chem. Soc. 2006, 128, (33), 10660-10661. 122. Oba, M.; Ueno, R.; Fukuoka, M.; Kainosho, M.; Nishiyama, K., Synthesis of L-threo- and L-erythro-[1-13C, 2,3-2H2]amino acids: novel pr obes for conformational analysis of peptide side chains. J. Chem. Soc., Perkin Trans. 1 1995, 0, (12), 1603-1609. 123. Li, J.; Sha, Y., A convenient synthesis of ami no acid methyl esters. Molecules 2008, 13, (5), 1111-9. 124. Lambert, J. B.; Shurvell, H. F.; Lightner, D.; Cooks, R. G., Organic Structural Spectroscopy . 1st ed.; Prentice Hall: New Jersey, 1998; p 568. 125. Jensen, F. R.; Bushweller, C. H., Separation o f conformers: Axial and equatorial isomers of chlorocyclohexane and trideuteriomethoxycyclohex ane. J. Am. Chem. Soc. 1969, 91, (12), 3223-3225. 126. Garbisch, E. W.; Griffith, M. G., Proton coupl ings in cyclohexane. J. Am. Chem. Soc. 1968, 90, (23), 6543-6544. 212 127. Kumar, S.; Dwevedi, A.; Kayastha, A. M., Immob ilization of soybean (Glycine max) urease on alginate and chitosan beads showing impro ved stability: Analytical applications. J. Mol. Catal. B: Enzym. 2009, 58, (1-4), 138-145. 128. Stingl, K.; Uhlemann, E. M.; Schmid, R.; Alten dorf, K.; Bakker, E. P., Energetics of Helicobacter pylori and its implications for the me chanism of urease-dependent acid tolerance at pH 1. J. Bacteriol. 2002, 184, (11), 3053-3060. 129. Cooke, H. A.; Bruner, S. D., Probing the activ e site of MIO-dependent aminomutases, key catalysts in the biosynthesis of -amino acids incorporated in secondary metabolites. Biopolymers 2010, 93, (9), 802-810. 130. Katz, B. A.; Elrod, K.; Verner, E.; Mackman, R . L.; Luong, C.; Shrader, W. D.; Sendzik, M.; Spencer, J. R.; Sprengeler, P. A.; Kolesnikov, A.; Tai, V. W. F.; Hui, H. C.; Guy Breitenbucher, J.; Allen, D.; Janc, J. W., Elaborat e Manifold of Short Hydrogen Bond Arrays Mediating Binding of Active Site-directed Se rine Protease Inhibitors. J. Mol. Biol. 2003, 329, (1), 93-120. 131. Hanoian, P.; Sigala, P. A.; Herschlag, D.; Ham mes-Schiffer, S., Hydrogen Bonding in the Active Site of Ketosteroid Isomerase: Electronic In ductive Effects and Hydrogen Bond Coupling. Biochemistry 2010, 49, (48), 10339-10348. 132. Lyubimov, A. Y.; Lario, P. I.; Moustafa, I.; V rielink, A., Atomic resolution crystallography reveals how changes in pH shape the protein microenvironment. Nat. Chem. Biol. 2006, 2, (5), 259-264.