Citation
Application of Native and Mutant Threonine Aldolases to Chemical Synthesis

Material Information

Title:
Application of Native and Mutant Threonine Aldolases to Chemical Synthesis
Creator:
Beaudoin, Sarah Franz
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (282 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Chemistry
Committee Chair:
STEWART,JON DALE
Committee Co-Chair:
BRUNER,STEVEN DOUGLAS
Committee Members:
HORENSTEIN,NICOLE ALANA
CASTELLANO,RONALD K
DING,YOUSONG

Subjects

Subjects / Keywords:
aldol -- aldolase -- biocatalysis -- biocatalyst -- biochemistry -- chemistry -- crystallography -- engineering -- enzymes -- mutagenesis -- protein -- threonine
Chemistry -- Dissertations, Academic -- UF
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Chemistry thesis, Ph.D.

Notes

Abstract:
beta-Hydroxy-alpha-amino acids are an important class of natural products that have found many uses in pharmaceuticals. Although several "chemical" strategies have been described for these products, we chose a biocatalytic approach since it could allow for a one step synthesis of these amino acids. Threonine aldolases (TAs) catalyze carbon-carbon bond formation between glycine and an aldehyde, producing beta-hydroxy-alpha-amino acids. Four L-TAs were cloned and overexpressed in Escherichia coli (Aeromonas jandaei L-allo-TA, E. coli L-TA, Thermotoga maritima L-allo-TA, and Pseudomonas putida L-TA). A Design of Experiments strategy was used to identify optimal reaction conditions for each enzyme. These conditions were used to characterize the substrate and stereoselectivity of each TA toward a panel of aldehyde acceptors. In general, the A. jandaei L-allo-TA performed best, and six representative examples were scaled up 10-fold in order to develop downstream steps for product isolation. The use of glycine oxidase to degrade the residual starting material simplified this process greatly. We also solved the x-ray crystal structure of L-TA from P. putida using molecular replacement. Like the other TAs, P. putida L-TA was a homotetramer with the active site composed of a "catalytic dimer". This structure and others were used to design site directed mutagenesis studies. Three A. jandaei L-allo-TA site saturation mutagenesis libraries (His 85, Tyr 89, and His 128) were screened against a variety of aldehyde acceptors to probe for enhanced enzyme activity and improved diastereoselectivity. Histidine 85 was found to be unequivocally essential for enzyme catalysis as any change at this position left the enzyme inactive. On the other hand, a few changes at positions 128 and 89 increased diastereoselectivity for a select group of aldehyde acceptors and in some cases, reversed diastereoselectivity completely. Finally, we showed that wild type A. jandaei L-allo-TA accepted fluorinated acetones, the first example of a non-aldehyde acceptor for TAs. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2017.
Local:
Adviser: STEWART,JON DALE.
Local:
Co-adviser: BRUNER,STEVEN DOUGLAS.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2018-06-30
Statement of Responsibility:
by Sarah Franz Beaudoin.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Embargo Date:
6/30/2018
Classification:
LD1780 2017 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

APPLICATION OF NATIVE AND MUTANT THREONINE ALDOLASES TO CHEMICAL SYNTHESIS By SARAH FRANZ BEAUDOIN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FO R THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2017

PAGE 2

2017 Sarah Franz Beaudoin

PAGE 3

To my family

PAGE 4

4 ACKNOWLEDGMENTS First of all, I would like to acknowledge and express my ultimate gratitude to my advisor and mentor Dr. Jon Stewart for this support and guidance over the past four and a half years. It was a pleasure to work with him and h is inspiration and compassionate teaching style has been invaluable to me I would also like to thank Dr. Br ent Feske for his leadership and encouragement through my first years of research as an undergraduate and leading up to my decision to attend graduate school I thank Dr. Matthew Burg for his patience and instruction towards solving the crystal structure of threonine aldolase I am very thankful for current and past group members of the Stewart group, in particular Michael Hanna for working alongside me during the mutagenesis studies as my undergraduate assistant. Dr. Robert Powell III, Dr. Louis Mouterd e, Hyunjun Choe Richard Watkins, Thinh Nguyen, Kevin Fisher, Cristina Iturrey and others have been invaluable coworkers and friends. T heir support, combined expertise and friendship through the years have made my time here at the University of Florida no t only knowledgeable but also enjoyable. I am also thankful for my committee members Dr. Steven Bruner, Dr. Nicole Horenstein, Dr. Ronald Castellano and Dr. Yousong Ding for their time, advice and suggestions And lastly, I would like to thank my family a nd friends for their love and support throughout my entire degree. Especially my husband Steven for his love and pronounced encouragement to pursue my dreams.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 14 A BSTRACT ................................ ................................ ................................ ................... 18 CHAPTER 1 HYDROXY AMINO ACIDS ................................ ................................ ............... 20 Introduction ................................ ................................ ................................ ............. 20 Organic Synthesis Strategies for Synthes Hydroxy amino Acids .............. 21 By Addition of Imides to Aldehydes ................................ ................................ .. 21 Aziridine Ring Opening ................................ ................................ .................... 23 Chiral Glycine Enolate ................................ ................................ ...................... 25 aza Claisen Rearrangements of Allylic Acetimidates ................................ ....... 27 Dynamic Kinetic Resolution via Ruthenium Catalyzed Dehydrogenation of N substituted Amino keto Esters ................................ ............................ 29 Potential Targets ................................ ................................ ................................ ..... 31 Droxidopa ................................ ................... 31 LPC 058 A Potent Gram Negative Bacteria Inhibitor ................................ .... 34 Rhizobitoxine A PLP Dependent Enzyme Inhibitor ................................ ....... 35 Sphingofungins Antifungal Agents ................................ ................................ 37 2 THREONINE ALDOLASE AS BIOCATALYSTS ................................ ..................... 59 Introduction ................................ ................................ ................................ ............. 59 The Mechanism of Threonine Aldolase ................................ ............................ 61 Threonine Aldolases Utilized for Chemical Synthesis ................................ ...... 61 Substrate Selectivity of Threonine Aldolase ................................ ............................ 62 Glycine/Alkyl Aldehydes ................................ ................................ ................... 63 Glycine/Aryl Aldehydes ................................ ................................ ..................... 77 Other Amino Acid Donors (D Ala, D Ser, and D Cys) ................................ ...... 79 Structures of Threonine Aldolases ................................ ................................ .......... 86 T. maritima L allo Threonine Aldolase ................................ .............................. 86 E. coli L Threonine Aldolase ................................ ................................ ............. 87 A. jandaei L allo Threonine Aldolase ................................ ................................ 89 A. xylosoxidans D Threonine Aldolase ................................ ............................. 91 Other Threonine Aldolase Structures ................................ ............................... 93 Protein Engineering Studies of Threonine Aldolases ................................ .............. 93

PAGE 6

6 Improving Catalytic Activity ................................ ................................ ............... 94 Improving Thermostability ................................ ................................ ................ 95 Improving Stereoselectivity ................................ ................................ ............... 97 Introducing and Optimizing Threonine Aldolase Activity into a Novel Scaffold ................................ ................................ ................................ ......... 99 Conclusions and Future Work ................................ ................................ ............... 100 3 SUBSTRATE PROFILING OF THREE THEONINE ALDOLASES ....................... 104 Introduc tion ................................ ................................ ................................ ........... 104 Results and Discussion ................................ ................................ ......................... 106 Gene Cloning and Protein Overexpression ................................ .................... 106 Derivatization of Amino Acids for Analysis ................................ ..................... 107 Optimization of Reaction Conditions ................................ .............................. 108 Screening of Aldehyde Acc eptors ................................ ................................ .. 113 Screening of Amino Donors ................................ ................................ ............ 115 Optimizing Isolation and Purification Procedure ................................ ............. 115 Preparative Conversions ................................ ................................ ................ 117 Assignment of Relative Configurations ................................ ........................... 119 Monitoring Transaldimination of Amino Donors ................................ .............. 123 Deprotonation of Amino Donors ................................ ..................... 124 Investigation into the Thermodynamic Reversibility of Aldol Prod ucts ............ 125 Conclusions ................................ ................................ ................................ .......... 126 Experimental Procedures ................................ ................................ ...................... 126 4 STRUCTURE DETERMINATION AND SUBSTRATE PROFILING OF P. putida L THREONINE ALDOLASE ................................ ................................ .................. 142 Introduction ................................ ................................ ................................ ........... 142 Results and Discussion ................................ ................................ ......................... 143 Gene Cloning and Protein Overexpression ................................ .................... 143 Purification of P. putida L Threonine Aldolase and Screening of Protein Crystals ................................ ................................ ................................ ....... 144 Overall Structure of the L Threonine Aldolase from P. putida ........................ 145 The Active Site of Threonine Aldolase ................................ ........................... 147 Structural Comparison of L Threonine Aldolases ................................ ........... 148 Optimization of Reaction Conditions ................................ .............................. 149 Screening of Aldehyde Acceptors ................................ ................................ .. 152 Conclusion ................................ ................................ ................................ ............ 153 Experimental Procedures ................................ ................................ ...................... 154 5 SITE SATURATION MUTAGENESIS OF A. jandaei L allo THREONINE ALDOLASE ................................ ................................ ................................ ........... 170 Introduction ................................ ................................ ................................ ........... 170 Results and Discussion ................................ ................................ ......................... 171 Site Saturation Mutagenesis Library Construction and Overexpression ......... 171

PAGE 7

7 Histidine 85 Library ................................ ................................ ......................... 172 Histidine 128 Library ................................ ................................ ....................... 174 Tyrosine 89 Library ................................ ................................ ......................... 180 Conclusion ................................ ................................ ................................ ............ 186 Experimental Procedures ................................ ................................ ...................... 186 6 BIOCATALYTIC SYNTHESIS OF TERTIARY METHYL FLUORINATED HYDROXY AMINO ACIDS BY THREONINE ALDOLASE ................................ 196 Introduction ................................ ................................ ................................ ........... 196 Results and Discussion ................................ ................................ ......................... 196 Screening of Fluorinated Acetones ................................ ................................ 196 Preparative Conversions ................................ ................................ ................ 199 Conclusion ................................ ................................ ................................ ............ 200 Experimental Procedures ................................ ................................ ...................... 200 7 CONCLUSIONS AND FUTURE WORK ................................ ............................... 207 Wild Type L Threonine Aldolase ................................ ................................ ........... 207 Mutant L Threonine Aldolase ................................ ................................ ................ 207 Histidine 128 Mutants ................................ ................................ ..................... 208 Tyrosine 89 Mutants ................................ ................................ ....................... 208 Structure of L Threonine Aldolases ................................ ................................ ...... 209 Future Work ................................ ................................ ................................ .......... 210 Double Mutations ................................ ................................ ........................... 210 Additional Mutations ................................ ................................ ....................... 210 Structure of Mutant L Threonine Aldo lase ................................ ...................... 211 APPENDIX ................................ ................................ ................................ .................. 212 LIST OF REFERENCES ................................ ................................ ............................. 265 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 282

PAGE 8

8 LIST OF TABLES Table page 2 1 Synthetic applications of threonine aldolases using glycine as the donor .......... 64 2 2 Synthetic applications of threonine aldolases using alanine as the donor .......... 80 2 3 Synthetic applications of threonine aldolases using serine as the donor ............ 83 2 4 Synthetic applications of threonine aldolases using cysteine as the donor ........ 85 2 5 Kinetic parameters for L allo threonine aldolase from A. jandaei and its mutants ................................ ................................ ................................ ............... 90 3 1 Substrate specificity of L TA catalyzed aldol reactions ................................ ..... 109 3 2 Preparative scale reacti on results ................................ ................................ .... 118 3 3 Expected values for diagnostic coupling constants in pure staggered conformers of the two diastereomers ................................ ............................... 120 3 4 Diagnostic coupling constants measured in the two diastereomers of M and m ................................ ................................ ................................ ...................... 121 4 1 Data collection and refinement statistics ................................ .......................... 146 4 2 Substrate specificity of L TA catalyzed aldol reactions ................................ ..... 150 5 1 Screening of histidine 85 mutants against aldehyde acceptors ........................ 173 5 2 Initial screening of histidine 128 mutants against aldehyde acceptors ............. 175 5 3 Extended screening of histidine 128 mutants against aliphatic aldehydes ....... 177 5 4 Extended screening of histidine 128 mutants against aromatic aldehydes ....... 179 5 5 I nitial screening of tyrosine 89 mutants against aldehyde acceptors ................ 181 5 6 Extended screening of tyrosine 89 mutants against aliphatic aldehydes .......... 183 5 7 Extended screening of tyrosine 89 mutants ag ainst aromatic aldehydes ......... 185 5 8 Forward and reverse primer sets for site saturation mutagenesis .................... 188 6 1 L TA catalyzed aldol co ndensation of fluorinated acetones ............................. 197 6 2 Forward and reverse primer sets for mutagenesis ................................ ........... 2 02

PAGE 9

9 Table page A 1 Assignment of 1 H and 13 C chemical shifts in diastereomers M and m in D 2 O (or methanol ................................ ................................ ............ 212 A 2 R S sign resulting from disubstitution with MPA ................................ ........... 213 A 3 Chemical shifts and R S sign in double MPA derivatives of the n butyl and cyclohexyl compounds ................................ ................................ ..................... 213 A 4 Chemical shifts in the methyl ester of 3 phenyl 2 amino 3 hydr oxypropanoic acids and its amides with MPA. ................................ ................................ ........ 214

PAGE 10

10 LIST OF FIGURES Figure page 1 1 hydroxy amino a cid s as precursors to or stand alone pharmaceuticals ................................ ................................ ................................ 39 1 2 Addition of ethyl isothiocyanatoacetate to benzaldehyde ................................ ... 39 1 3 Addition of oxazolidinone to benzaldehyde ................................ ........................ 39 1 4 Addition of oxazolidinone to 3 chloro 4 p henoxybenzaldehyde to yield the protected amino acid, a precursor to vancomycin ................................ .............. 40 1 5 Addition of oxazolidinone to an aldehyde by N N dioxide nickel(II) catalyst complex ................................ ................................ ................................ .............. 40 1 6 First documented example of an acid catalyzed aziridine ring opening .............. 41 1 7 Generalized summary of aziridine ring openings ................................ ................ 41 1 8 Optimized synthesis of ( ) chloramphenicol by aziridine ring opening strate gies ................................ ................................ ................................ ............ 42 1 9 hydroxy amino acid ethyl esters by a 1,3 dipolar cycloaddition with an aziridine and benzaldehyde ................................ .............. 42 1 10 One pot synthesis of trans oxazolidinones by an asymmetric aziridine intermediate ................................ ................................ ................................ ........ 42 1 11 hydroxy amino acid synthesis ................................ ................................ ................................ ............ 43 1 12 syn hydroxy amino acids by oxazolidinone chiral glycine enolates ................................ ................................ 43 1 13 The natural products cy closporine and echinocandin D ................................ ..... 44 1 14 Caddick et al. route to syn hydroxy amino acids by imidizolidinone chiral glycine enolates ................................ ................................ ................................ .. 45 1 15 anti hydroxy amino acids by chiral glycine enolates ... 45 1 16 First proposed mechanism for metal catalyzed aza Claisen rearrangements .... 46 1 17 Synthesis of hydroxy amino acids by means of aza Claisen rearrangement ................................ ................................ ................................ .... 46

PAGE 11

11 Figure page 1 18 The synthesis of (2 R ,3 S ) 2 amino 3,4 dihydro xybutyric acid by aza Claisen rearrangement ................................ ................................ ................................ .... 47 1 19 Dynamic kinetic resolution ................................ ................................ .................. 47 1 20 Dynamic kinetic resolution towards the sy nthesis of important intermediates .... 48 1 21 (2 S ,3 R ) and (2 R ,3 R ) p chloro 3 hydroxytyrosines in vancomycin ...................... 48 1 22 A hydroxy amino intermediate in ( ) balanol ................................ ................ 49 1 23 Dynamic kinetic resolution towards anti hydroxy amino acids ..................... 49 1 24 Pr oposed mechanism for dy namic kinetic resolution for anti and syn hydroxy amino acids ................................ ................................ ....................... 50 1 25 Coupling therapy: selectively targeting dopamine synthesis by methyltyrosine inhibitor ................................ ................................ ....................... 52 1 27 Biocatalytic synthesis of L threo dihydroxyphenylserine by L threonine aldolase ................................ ................................ ................................ .............. 54 1 28 Structure comparison of the natural s ubstrate of LpxC versus the inhibitor ........ 54 1 29 Initial Route to LPC 058 ................................ ................................ ..................... 55 1 30 Final route in the synthesis of LPC 058 ................................ .............................. 56 1 31 Proposed rhizobitoxine biosynthesis pathway ................................ .................... 57 1 32 Sphingofungins A F ................................ ................................ ......................... 57 1 33 Sphingofungin synthesis strategies ................................ ................................ .... 58 1 34 The first step in the biosynthesis of sphingolipids, serine palmitoyltransferase .. 58 2 1 Aldol condensation with threonine aldolase ................................ ...................... 102 2 2 Mechanism of threonine aldolase ................................ ................................ ..... 102 2 3 T hreonine aldolase kinetic pathway ................................ ................................ .. 103 2 4 Aldol products from a mutant Ala racemase ................................ ..................... 103 3 1 Threonine aldolase reactions ................................ ................................ ............ 139 3 2 Glycine oxidase reaction ................................ ................................ .................. 139

PAGE 12

12 Figure page 3 3 Derivatization of amino acids ................................ ................................ ............ 140 3 4 Screening of amino donors ................................ ................................ ............... 140 3 5 Preparative reactions ................................ ................................ ........................ 140 3 6 Transaldimination of amino donors ................................ ................................ ... 141 3 7 Thermodynamic reversibility of L threonine ................................ ...................... 141 4 1 Ribbon representation of P. putida L threonine aldolase ................................ .. 160 4 2 Overview of the active site of P. putida L threonine aldolase ........................... 162 4 3 Overview of the active site pocket of chain A from P. putida L threonine aldolase showin g key interactions between residues in the active site and the pyridoxal phosphate cofactor ................................ ................................ ............ 163 4 4 ESPript sequence alignment of four threonine aldolases ................................ 165 4 5 Detailed view of the active site pocket of chain A from P. putida L threonine aldolase showing key interactions between conserved residues in the active site ................................ ................................ ................................ .................... 166 4 6 Structural comparison of four L threonine aldolases ................................ ........ 168 5 1 Initial screening of H128 mutants ................................ ................................ ..... 192 5 2 Initial screening of Y89 mutants. ................................ ................................ ....... 194 6 1 methylfluoro threonine analogues by threonine aldolase ................. 206 A 1 Mass spectrum of (4 S ,5 R ) 2 amino 3,4,5,6 tetrahydroxyhexanoic acid ........... 214 A 2 NMR spectra of (2 S ,3 S ) 2 amino 3 hydroxyheptanoic acid ............................. 216 A 3 NMR spectra of (2 S ,3 R ) 2 amino 3 hydroxycyclohexanepropanoic acid ......... 223 A 4 N MR spectra of (2 S ,3 S ) 3 hydroxyphenylalanine ................................ ............. 227 A 5 NMR spectra of 2 amino 3 hydroxy 4 pyridinepropanoic acid .......................... 232 A 6 NMR spectra of 3 hydroxy 2 methoxy phenylalanine ................................ ....... 239 A 7 NMR spectra of 2 amino 3 (2 chloro 3 pyridine) 3 hydroxypropanoic acid ...... 243 A 8 1 H NMR for thermodynamic reversibility of L allo threonine aldolase ............... 248

PAGE 13

13 Figure page A 9 MS for thermodynamic reversibility of L allo threonine aldolase ...................... 253 A 10 Mass spectrum of 2 amino 4,4 difluoro 3 hydroxy 3 methylbutanoic acid ....... 257 A 11 Proton and fluorine NMR spectra of 2 amino 4,4 d ifluoro 3 hydroxy 3 methylbutanoic acid synthesized by P. putida L TA ................................ ......... 259 A 12 Proton and fluorine NMR spectra of 2 amino 4,4 difluoro 3 hydroxy 3 methylbutanoic acid synthesized by A. jandaei L allo TA ................................ 261 A 13 Proton and fluorine NMR spectra of 2 amino 4 fluoro 3 hydroxy 3 methylbutanoic acid synthesized by A. jandaei L allo TA. ................................ 263

PAGE 14

14 LIST OF ABBR EVIATIONS 1 H NMR P roton nuclear magnetic resonance A 500 Absorbance at 500 nm acac Acetylacetonate AcOH Acetic acid Ala Alanine Arg Arginine Asp Aspartate Asn Asparagine BINAP 2,2' B is(diphenylphosphino) 1,1' binaphthyl bipy Bipyridine BuOH Butano l CoA Coenzyme A Cys Cysteine d e Diastereomeric e xcess D 2 O Deuterium Oxide DCC N N Dicyclohexylcarbodiimide DHAP Dihydroxyacetone phosphate DKR Dynamic kinetic resolution DMSO Dimethyl sulfoxide DNA Deoxyribonucleic a cid DOE Design of e xperime nts DOPS Dihydroxyphenylserine e.e. Enantiomeric excess Et 2 O Diethyl ether

PAGE 15

15 Et 3 N Triethylamine e TA L Threonine aldolase from E. coli FAA Trifluoroacetylacetone GC/MS Gas chromatographer/mass spectrometer Glu Glutamate Gln Glutamine Gly Glycine H 2 O Water Hg Mercury Hht 3 Hydroxyhomotyrosine His Histidine Hmp 3 Hydroxy 4 methylproline HPLC High pressure liquid chromatography HRP Horseradish peroxidase Ile Isoleucine i Pr 2 EtN Hu n i N N diisopropy l ethylamine IPTG Isopropyl D thiogalactopuranoside KDA Potassium diisopropylamide L Liter LATA L allo Threonine aldolase from A. jandaei LB Luria b roth Leu Leucine LpxC UDP 3 O ( R 3 hydroxymyristoyl) N acetylglucosamine deacetylase L TA L Threonine aldolase LTAPP L Threonine aldolase from P. putida

PAGE 16

16 Lys Lysine MeBMT 3 Hydroxy 4 methyl 2 (methylamino) 6 octenoic acid MeOH Methanol Met Methionine Mg(ClO 4 ) 2 Magnesium p erchlorate MPA Methoxyphenylacetic acid MS Mass spectrometry MSTFA N Methyl N (trimethylsilyl ) trifluoroacetamide NA Noradrenaline NMR Nuclear magnetic resonance nm Nanometers NOH N eurogenic O rthostatic H ypotension NTBB Sodium tetraborate O.D. 600 Optical density at 600 nm OPA/NAC o rtho Phthalaldehyde N acetylcysteine PCR Polymerase chain r eaction Pd Palladium Phe Phenylalanine PLP phospate ppm Parts per million Pro Proline Psi Pounds per square inch pybox Pyridine bis(oxazoline) Ru Ruth en ium Ser Serine

PAGE 17

17 SPT Serine palmitoyltransferase TA Threonine aldolase t BuOMe tert Butyl methyl ether THF Tetrahydrofuran Thr Threonine TLC Thin layer chromatography Trp Tryptophan t TA L allo Threonine aldolase from T. maritima Tyr Tyrosine UDP Uridine diphosphate UV Vis Ultaviolet v isible spectroscopy Val Valine vvm Volume o f Air Under Standard Conditions per Volume of Liquid per Minute wcw Wet cell weight

PAGE 18

18 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosop hy APPLICATION OF NATIVE AND MUTANT THREONINE ALDOLASES TO CHEMICAL SYNTHESIS By Sarah Franz Beaudoin December 2017 Chair: Jon D. Stewart Major: Chemistry Hydroxy amino acids are an important class of natural products that have described for these products, we chose a biocatalytic approach since it could allow for a one step synthesis of these amino acids. Threonine aldolases (TAs) catalyze carbon hydroxy amino acids. Four L TA s were cloned and overexpressed in Escherichia coli ( Aeromonas jandaei L al lo TA E. coli L TA, Thermotoga maritima L allo TA, and P seudomonas putida L TA ). A D esign of Experiments strategy was used to identify optimal reaction conditions for each enzyme. These conditions were used to characterize the substrate and stereoselect ivity of each TA toward a panel of aldehyde acceptors. In general, the A. jandaei L allo TA performed best, and six representative examples were scaled up 10 fold in order to develop downstream steps for product isolation The use of glycine oxidase to d egrade the residual starting material simplified this process greatly. We also solved the x ray crystal structure of L TA from P. putida using molecular replacement. Like the other TAs, P. putida L TA was a homotetramer with the active

PAGE 19

19 site composed of a directed mutagenesis studies. Three A jandaei L allo TA site saturation mutagenesis libraries (His 85, Tyr 89, and His 128) were screened against a variety of aldehyde acceptors to pro be for enhance d enzyme activity and improved diastereoselectivity. Histidine 85 was found to be unequivocally essential for enzyme catalysis as any change at this position left the enzyme inactive. On the other hand, a few changes at positions 128 and 89 increased diastereoselectivity for a select group of aldehyde acceptors and in some cases, reversed diastereoselectivity completely. Finall y we showed that wild type A. jandaei L allo TA accepted fluorinated acetones, the first example of a non aldehyde acceptor for TAs.

PAGE 20

20 CHAPTER 1 HYDROXY AMINO ACIDS Introduction Hydroxy amino acids comprise an important class of natural products that include the natural amino acids threonine and serine as well as valuable precursors to a wide range of antibioti cs such as vancomycin, 1 polyoxin A, 2 and myriocin 3 They are also found in antifungal agents including sphingofungins, 4, 5 rhizobitoxine 6 and pharmaceuticals such as d roxidopa 7, 8 and cyclosporine 9 Droxidopa (L threo 3,4 dihydroxyphenylserine or L threo DOPS ) is the most notable member of this family rosses the blood brain barrier and is metabolize d to L noradreline which is responsible for increasing 10 Additionally, Rhizobitoxine is an effective inhibitor of PLP dependent enzymes that hydroxy amino acid 4 hydroxy L threonine as a precursor. 6 hydroxy amino acids are useful in the food industry. D Glucosaminic acid, or 2 amino 2 deoxy D gluconic acid, is a natural product from Aeromonas oxydans that acts as an artificial sweetener. 11 More recently, D g lucosaminic acid has been found useful as a building block for several glycosidase inhibitors, such as (2 S ,4 S ,5 R ) 4,5,6 t rihydroxynorleucine 12, 13 Figure 1 1 highlight s a few examples of the hydroxy amino a cid m otif (pink). Since the 1980s, significant efforts have been devoted to producing the se optically pure hydroxy amino acids on large sc ale s using chemical approaches These include but are not limited to chiral glycine enolate, 14 Sharpless dihydroxylation, epoxidation, and aminohydroxyla tion 15 17 aza Claisen rearrangements of allylic

PAGE 21

21 acetimidates 18 addition of imides to aldehydes, 19, 20 organocatalytic asymmetric aldol amino aldehydes, 21 gly cine Schiff base, 22 dynamic kinetic resolution via ruthenium catalyzed dehydrogenation of N substituted amino keto esters, 23 30 aminohydroxylation of olefins, 31 aziridine ring opening, 32 ammonium ylides, 33 chiral ammonium salts, 34, 35 and numerous others. This chapter describes a subset of the chemical synthes is approaches described above as well as a few of the pharmaceutical targets that are of key hydroxy amino acids by threonine aldolases. Organic Synthesis Strategies for Synthesizing Hydro xy amino Acids By Addition of Imides to Aldehydes In 2002, Willis and coworkers 19 devised a strategy to produce synthetically useful protected aryl variants of h ydroxy amino a cids by a direct addition of an isothiocyanate substituted ester to a variety of aryl aldehyd es using an achiral catalyst. Th is approach was inspired by the popular chiral glycine enolate strategy, but allowed the use of a catalyst for both enolate formation and addition to the aldehyde. In order for coordination alcohol g roup to coordinate to a metal catalyst, they used an isothiocyanate substituted ester. A combination of Lewis acids (metal triflates ) and weak amin e bases ( such as triethylamine) were investigated to find the optimal combination The best mixture 10 mol % magnesium perchlorate, 20 mol% triethylamine (Et 3 N), and 10 mol% bipyridine (bipy), generated a 94% yield of ethyl isothiocyanatoacetate with benzaldehyde in 24 hours (Figu re 1 2). Willis and coworkers investigated other aryl aldehydes using the same ca talyst system and found it useful for a variety of aromatic aldehydes although a 60 : 40 ( syn : anti ) mixture of diastereomers was obtained for most products. 19

PAGE 22

22 Subsequently, Willis et al. attempted to improve the catalyst system into a more highly enantioselective variant 20 They exchang ed the bipyridine ligand with an as sortment of enantiomerical ly pure bidentate ligands; however low enantioselectivities were obtained and they re focused their efforts on other aspect s of the system. They used oxazolidinone as the glycine equivalent and reaction c onditions for this system were optimized as before. The best conditions were 10 mol% magnesium perchlorate, 20 mol% Hn i N N Diisopropyethylamine or i Pr 2 EtN), and 11 mol% pyridine bis(oxazoline) (pybox). Reactions utilizing these conditions with benzaldehyde as the acce ptor resulted in an 86% yield and a n 85 : 15 ( syn : anti ) mixture of diastereomers (Figure 1 3). Willis and coworkers also tested the enantioselective catalyst system with other aryl aldehydes. Although most of the products retained the high diastereosel ectivity (even up to 91 : 9), ortho substituted ph enyl rings were not tolerated well and resulted in a 50 : 50 ratio of diastereomers 20 One key advantage of this strategy i s that all components are commercially available and reaction conditions are simple. Its relevance was demonstrated by Willis succes sful s ynthesi s of a protected amino acid with the same substitution pattern as the functionalized tyrosine residue found in vancomycin in a 78% yield and 94% d.e (Figure 1 4a). Seidel and coworkers 36 extended the W illis approach by substituting additional imide starting materials and catalysts. The dimethyl analog proved to be best along with 1 (4 nitrophenyl) 3 ((1 R ,2 R ) 2 (pyrrolidin 1 yl)cyclohexyl)thiourea as the catalyst. This combination yielded the protected syn amino acid in outstanding yields (>98%) and high selectivity (93 : 7) using benzaldehyde as the electrophile. Further optimization of

PAGE 23

23 these conditions allowed the catalyst loading to be reduced from 20 to 5 mol% and the equivalents of aldehyde from 2 to 1.2. A variety of aldehydes were tested with the new enantioselective catalyst. Most a romatic aldehydes provided the same high yield s and stereo selectivit ies as benzaldehyde ; however, aliphatic aldehydes were less reactive and gave slightly lower ste reo selectivit ies 36 The only disadvantage of the improved catalyst was that it was not commercially available Recently, Feng and coworkers 37 sought a catalyst that could achieve these h igh yields and selectivities with better functional group tolerance They employ ed N N dioxide metal complexes since they have proven to be useful chiral auxiliaries 38 41 and catalysts in enantioselective transfo rmations. 42 45 In this study, seven different N N dioxide ligands were coordinated to various metal salts for the aldol condensation between benzaldehyde and oxazolidinone. The optimum catalyst, a n N N nickel(II) catalyst complex, was subjected to optimization using temperature and solvent studies, followed by an extensive screening of thirty aldehydes under the optimal conditions (Figure 1 5). Most aldehydes gave >90% yield and diastereoselectivities > 95%. The few exceptions were the troublesome aliphatic aldehydes identified previously by Li et al. 36 Nonetheless, the N N dioxide metal catalyst designed by Chen e t al provided better results even for these dif ficult cases The system developed by Feng and coworkers is currently the best approach in this area, producing relatively high yields (up to 98%) and outstanding diastereoselectivity (up to >99 : 1) for a broad range of substrates. Aziridine Ring Opening Aziridines, highly strained three membered heterocycles, are extremely susceptible to ring opening reactions. For this reason they are incredibly useful

PAGE 24

24 intermediates for numerous products 46 50 T he earliest ex ample was reported in 1895 when Gabriel and Stelzner documented the first acid catalyzed isomerization of an aziridine (Figure 1 6). 51 Over the years chemists have extended the initial findings of Gabriel and Stelzner. Figure 1 7 represents a generalized summary of aziridine ring openings. Since the discovery of asymmetric aziridine ring openings, much emphasis has be en placed on synthesizing these starting materials with different substitut ed patterns including tri substitut ion s. 49, 50, 52, 53 Ring opening of aziridines to yield hydroxy amino acids was first demonstrated in 2001 by Wulff and Loncaric in their synthesis of ( ) chloramphenicol, an antibacterial agent. 32 The optimized 4 step synthesis begins with commercially available p nitrobenzaldehyde. The key aziridine ring opening step is carried out using VAPOL as the catalyst and refluxing 1,2 dichloroethane to give the hydroxyl acetamide as a single diastereomer and an 80% yield. ( ) Chloramphenicol was obtained at >99% enantiomeric excess with a 74% yield by a simple reduction with sodium borohydride. This work is summarized i n F igure 1 8. 32 Recently, groups led by Somfai in Sweden 54 and Maruoka in Japan 53 devised similar str ategies for synthesizing hydroxy amino acid derivatives by aziridine ring rearrangements to oxazolidines and oxazolidinones, respectively. This idea is reminiscent of the Willis strategy described above, which u sed oxazolidinone as the glycine equivalent for the addition of imides to aldehydes. 20 Danielsson et al. combined a 1,3 dipolar cycloaddition with an aziridine and an aldehyde followed by simple hydrolysis to execute their plan ( F igure 1 9 ) The only downside to this strategy was the

PAGE 25

25 moderate diastereo selectivities which depend ed on the starting aldehyde M ost react ions favored the erythro isomer with >50% diastereomeric excess. 54 Hashimoto et al. devised a n elegant hydroxy amino acids 55 by a ring rearrangement of trisubstituted aziridines to afford trans oxazolidinones. 53 The rearrangement was mediated by trifilic acid at 40 o C in only 30 min. Since the reaction conditions for both the asymmetric aziridination and subsequent ring rearrangement were so similar, a one pot method tha t yielded the oxazolidinone and provide d hydroxy amino acid derivatives. 53, 55 This wo rk is summarized in F igure 1 10. Chiral Glycine Enolate The first account of a chiral glycine enolate used for the pro hydroxy amino acids was published in 1981 by Mukaiyama. 56 Both the magnesium counter ion requir ed in t he second step as well as the strong base that allow ed the nucleophilic addition to the aldehyde were optimized. Iodide and potass ium diisopropylamide (KDA) respectively, proved best at these roles (Figure 1 11). 56 Mukaiyama and coworkers focused their efforts into using this strategy for the synthesis of other natural products. 57 59 Shortly after, Evans and Weber designed their chiral glycine enolate using substituents originally developed for aldol condensations 60, 61 A n oxazolidinone was chosen as the chiral auxiliary for stereochemical control and an isothiocyanate as the activated group (Fig ure 1 12). 14, 62 Their focus, as with many others 63 65 was to synthesize of the distinctive am ino acid (MeBMT) in cyclosporine, an immunosuppressant (Fig ure 1 13). The previously reported synthesis of MeBMT by Wenger required 24 steps ; 9 the Evans strategy reduced it to six steps. This publication

PAGE 26

26 prompted later efforts to use hydroxy amino acids and many other natural products amino acids 66 68 and chiral dienophiles 69 Evans and Weber continued their work and reported the total synthesis of hydroxy amino acid derivatives fo und within the natural product ( 3 hydroxyhomotyrosine Hht and 3 hydroxy 4 methylproline Hmp) (Fig ure 1 13). They followed a similar plan as described in F igure 1 12 where each amino ac id was synthesized in four steps from an (isothiocyanoacetyl)oxazolidinone. They completed the synthesis with an overall yield of 50%. 70 Caddick Parr, and Pritchard enhanced by utilizing an imidizolidinone derivative as the chiral glycine enolate versus the standard oxazolidinone. The simple experimental conditions and high optical purity made their method attractive for p reparing syn hydroxy amino acids. 68, 71 The chiral auxiliary imidizolidinone, ha d been used previously by this group in the d ynamic k inetic r esolution method 72, 73 Their synthetic approach with chiral gly cine enolates is summarized in F igu re 1 14. The diastereoselectivity for their synthesis ultimately depended on the R group of the aldehyde. In the case shown in F igure 1 14, using benzaldehyde as the starting material gave an overall yield of 65% and a diastereomeric ratio of 97 : 3. 68 All the methods established by Evans and Weber 60 which were continued by Caddick, Parr, and Pritchard 68 were very effective for obtaining syn hydroxy amino acids A n effective method for the anti product remained elusive Iwanowicz et al. addressed this issue by developing a novel oxazolidine based chiral glycine equivalent. They v aried steric bulk and found that large substituents were required not only at the 2

PAGE 27

27 and 4 positions of the oxazolidine but also on the carboxylic acid. Figure 1 15 summarizes their work for the synthesis of anti hydroxy amino acids. 74 aza Claisen R earrangements of A llylic A cetimidates The use of metal catalyzed aza Claisen rearrangements of allylic trichloroacetimid ates to synthesize the corresponding trichloroacetamide s was first reported by Overman in 1974 using Hg(II) salts. 75 He found that a wide variety of allylic trichloroacetimidates underwent a thermal [3,3] sigmatropic rearrangement by refluxing m xylene ; however when 0.10 equivalent s of mercuric trifluoroacetate or mercuric nitrate were present the re action occurred almost instantaneous ly at room temperature Stronger Lewis acids such as silver fluoroborate and aluminum chloride etherate failed to catalyze the rearrangement. The Overman group proposed a mechanism for the rearrangement, shown in Figur e 1 16. 75 77 synthesize trichloroacetamide derivatives, it was also important to note that this rearrangement yielded chiral amines. This new approach to asymmetric amines was ext ended by the discovery t hat Pd(II) salts were more effective at catalyzing the [3,3] sigmatropic rearrangement of allylic esters. 78 80 Overman and Knoll reported that 4 mol% of PdCI 2 (MeCN) could catalyze the rearr angement of a variety of allylic acetates at room temperature 80 Both Hg(II) and Pd(II) salts were limited to allylic esters (unsubstituted at the C 2 position) ; however Pd(II) was the preferred catalyst because the allylic rearrangements were faster and required lower cataly st loadings 79 It was not aza Claisen rearrangements were used in hydroxy amino acids by the Sutherland group 81 E allylic trichloroacet imidate s were synthesized in six steps using commercially available

PAGE 28

28 hydroxy acids. Pd(II) catalyze d the [3,3] sigmatropic rearrangement with high yields and diastereo selectivities Their method was applied to five different starting materials to define the scope of the reaction (Figure 1 17) 81 It is important to note that sterically demanding substituents ( i.e. R = isopropyl) can reduce some Pd(II) to Pd(0) which catalyzed the undesir able [1,3] rearrangement. This reduction was likely caused by elimination which became competitive due to the slow rearrangement mediated by Pd(II) This effect could be eliminated by addi ng p b enzoquinone 81, 82 This Pd(0) catalyzed rearrangement was also observed by Ov erman, Ikariya and others. 78, 81, 83 85 The Overman, Sutherland and Jirgensons group s continued their work on aza Claisen rearrangements for several years searching for better reaction conditions allowed by other ca talysts such as Pt(II) and Au(I) 86 89 and solvents such as THF and toluene 86 I n most cases, the original conditions of PdCl 2 (MeCN) 2 in THF proved superior although Pt(II) and Au(I) were also effect ive at catalyzing the aza Claisen rearrangement. The S utherland group attempted to maximize yields for the ir synthesis of the piperdine alkaloid (+) conhydrine and its pyrolidine analogue, which both contain the 1,2 aminoalcohol functional group 90 They likewise target ed the synthesis of (2 R ,3 S ) 2 amino 3,4 dihydroxybutyric acid using this strategy (Figure 1 18). The rearrangement proceeded with a 68% yield and a 4 : 1 ratio of diastereomers using the Pd(II) catalyst. 91 The aza Claisen rearrangement has also been used in the synthesis of other important natural products including (+) monanchorin, 92 clavaminol A, C and H, 93 7 deoxypancratistatin analogues 94 and s phingosine 95

PAGE 29

29 In 2003, the Overman group synthesized and tested f ive chiral palladacyclic catalysts for their ability to catalyze aza Claisen rearrangement s of trifluoroacetimidates They found the asymmetric catalyst di chlorobis[( 5 ( S ) (p R ) 2 isopropyl) oxazolinylcyclopentadienyl,1 C N )) ( 4 tetraphenylcyclobutadiene)cobalt]dipalladium ( COP Cl ) catalyzed the arrangement with high yields and enantioselectivit ies even for sterically hindered trifluoroacetimidates 89, 96, 97 T he Sutherland group used this catalyst in the synthesis of syn 1,2 aminoalcohols and found a diastereomeric ratio of 52 : 1 rather than the anti isomer normally displayed by the Pd(II) catalyst 98 Dynamic Kinetic Resolution via Ruthenium Catalyzed Dehydrogenation of N substituted Amino keto Esters The use of a d ynamic kinetic resolution (DKR) for the hydroxy amino acids was first published in 1989 by both Noyori 28 and Gent. 99 Their goal was to synthesize this product with high diastere oselectiviti es as most previously reported selectivities using kinetic resolution did not exceed 50%. 100 The idea was to racemiz e the starting material and employ an enantioselective hydrogenation catalyst that only accepted one of the starting material antipodes and also selectively form ed only one diastereomer out of the four possible reducti on products Noyor i et al previously published a BINAP Ru catalyst that gave high enantio carbon position, but the catalyst afforded a racemic mixture carbon 101 By employing other BINAP Ru catalysts along with ster ically hindered functional groups allow ed for selectivity between the syn and anti transition states. Gent et al. used a different asymmetric catalyst, CHIRAPHOS Ru, in the synthesis of L and D threonine (>99% d.e.). 29 Subsequently the G ent group synthesiz ed a wide range of chiral Ru(II) catalysts and investigat ed their utility in DKR of additional starting materials that

PAGE 30

30 contained other functional chloro methyl hydroxy moieties 102 106 The general method is illustrated in F igure 1 1 9 Since the initial publication s by Noyori and Gen t many other groups have devised different asymme tric Ru catalysts for a variety of starting materials including but not limited to N amino keto esters (examples are summarized in reviews on DKR 105 110 ). DKR has been used over the years in the synthesis of hydroxy amino intermediates to many important compounds, including vancomycin 111 s tatin and its analogues, 112 b iphenomycin A 113 ( ) balano l ( a protein kinase C inhibitor ) 114 carbapenems 115 (2 S ,3 R ) 3 hydroxylysine 116 therapy drug, droxidopa 117 g ymnangiamide 118 and many others. Two of th ese examples are summarized in F igure 1 20 The synthesis by DKR of one of hydroxy amino acids in vancomycin was described by Gent and coworkers in 1996 111 Vancomycin contains both the (2 S 3 R ) and (2 R 3 R ) p c hloro 3 h ydroxytyrosines linked with a central p hydroxyphenylglycine (Figure 1 21 ) The DKR method was used for the key step in the synthe sis of the syn amino acid and involved an asymmetric Ru catalyst, RuB r 2 [( R ) MeOB IPHEP ], which resulted in both high yields (>99%) and diaste reoselectiviti es (>95%) (Figure 1 20 a). 111 Gent a DKR strategy towards the amino acid precursor to ( ) b alanol using the same catalyst (Figure 1 20 b and Figure 1 22 ). Over the course of a decade much effort was devoted to synthesizing a nd screening a collection of a s y m metric Ru catalysts (most involving ( R ) BINAP or ( R ) MeOB IPHEP as the chiral ligand) for the DKR of numerous s tarting materials. In 2001, Mohar et al. sought to enhance the catalyst system by employing electron withdrawing

PAGE 31

31 fluorosulfonyl groups on the diamine ligand. 119 The more important challenge was to find a better catalyst for anti product formation as a t th e time, only a few had been reported. 105, 108 In 2004, the Gen t group disclosed a Ru mediated hydrogenation with high anti diastereoselectivity (>90%) for most starting materials using Ru(II) catalysts with the SYNPHOS ligand (Figure 1 23). 120, 121 This significantly improved the toolbox for DKR and allow ed it to be use d in the synthesis of precursors such as ( + ) lactacystin 120 (Figure 1 2 3 ), symbioramide sulfobacin A 122 and others. In 2008, Makino et al. recognized that the anti isomer of these product s can also be formed by using [RuCl 2 (( S ) BINAP)] as the asymmetric cat alyst 123 They predict ed that when the amine is unprotected the reaction proceeds by a five membered cyclic transition state involving chelation between t he amine and the keto enol functional group formed by tautomerization This yields the anti product 123 (Figure 1 2 4 a). The original hydrogenation re action contains a protected amino group and involves a six membered cyclic transition state formed by chelation between the two carbonyl groups of the starting material This rearrangement yields the syn product 124 (Figure 1 2 4 b). Potential Targets Droxidopa Dihydroxyphenylserine (DOPS) was first reported as a component of a coupling therapy with the tyrosine hydroxylase inhibitor met hyltyrosine T he de carboxylation of threo or erythro DOPS by DOPA decarboxylase in mammalian cells produce s noradrenaline (NA). Methyltyrosine inhibits tyrosine hydroxylase in the brain and this blocks the synthesis of both NA and dopamine without affecting serotonin lev els 125 The coupling therapy allowed dopamine to be selectively targeted while NA was maintained at normal levels (Figure 1 25 ) 126, 127 By administering racemic erythro DOPS to mice

PAGE 32

32 that were previously depleted of NA a buildup of the unnatural ( S ) NA was found without the natural ( R ) isomer. 128 Alternatively, L threo DOPS wa s decarboxylated to the natural form of NA in the brain and hearts of rats, while the D threo isomer wa s recalcitrant to decarboxyla ti on On the other hand, in the presence of D threo DOPS or erythro DOPS, decarboxylase activity was inhibited by half. 129 131 NA is a molecule that produces analgesia by inhibiting the dorsal horn neurons that res pond to noxious inputs. 132 NA alone cannot pass the blood brain barrier; for this reason, L threo DOPS was clinically introduced as a precursor to noradrenaline by the direct conversion of L threo DOPS to L NA in vivo by an aromatic L amino acid dec arboxylase because of its ability to cross the blood brain barrier 133 L threo DOPS has been investigated for its ability to treat neurogenic ortho static hypotension (NOH) in amyloidotic polyneuropathy, pure autonomic failure and other pathologies 134 141 Patients that suffer from NOH related dise ases are unable to maintain healthy blood pressure while in the standing position. In a study presented by Carvalho et al. bedridden patients suffering from type I familial amyloidotic polyneuropathy who were administered L threo DOPS saw improvement aft er 3 5 days in th eir ability to walk freely and this effect continued throughout the six month treatment. 135 In a double b lind, placebo controlled crossover study administered by Kaufmann et al. patients suffering from NOH experienced increased blood pressure and less lightheadedness while standing. 10 Recently, droxidopa was examined for its ability to treat other diseases such as 142, 143 After one month of treatment, transgenic mice exhibited

PAGE 33

33 improved learning in the Morris water maze te st. Enhancement in learning was revealed to have a direct correlation between the NA levels in the brain and latency times in the water maze test. 142, 144 There has been additional interest in testing ability to treat a variety of other diseases or injuries including hemodynamic and renal alterations of liver cirrhosis in portal hypertensive rats, 145, 146 spinal cord injury, 147 down syndrome, 143 and more. In most studies droxidopa had no significant side effects and was tolerated by most patients even after a 12 month study. 137, 140, 141, 148 In 2007, Sudalai an d coworkers developed a chemical synthesis of L threo DOPS by sodium periodate mediated asymmetric bromohydroxylation with high regio and diastereoselectivity. 149 This methodology was based on the chiral glycine enolate discuss ed previously and eliminated the need for heavy metals and molecular bromi n e/ N halo succinimides as the halogen sources used by Herbert et al. and Hegeds et al in 2001 and 1975, respectively 150, 151 The Sudalai preparation of d roxidopa used carboxamide as the starting material and yielded L threo DOPS at an overall yield of 51 % and >99% diastereomeric excess (Figure 1 2 6 a ). Very recently, Guan et al. proposed a route towards the chemical synthesis of L threo DOP S by initially using a rhodium catalyzed asymmetric hydrogenation of an enamide which resulted in high yields and enantioselectivities (Figure 1 2 6 b). They used this method on a variety of relating starting materials and achi eved high syn selectivity wit h all substituents (>95%). 152 In a ddition, Wang and coworkers used the DKR method first established by Noyori and Gent in 1989 28, 99 and a Ru ( II ) catalyst to prepare d roxidopa in gram quantities Using different Ru(II) catalyst /base systems, they

PAGE 34

34 devel oped a n efficient way to make the syn diastereomer ( >99% ) of d roxidopa (Figure 1 2 6 c) 117 Baik et al. synthesized L threo DOPS by utilizing the l ow s pecific ity L threonine aldolase from S. coelicolor A3(2) in one enzymatic step from simple starti ng materials (Figure 1 27 ). 8 To obtain a more thermostable biocatalyst, error prone PCR was employed O ne mutant (H177Y ) retained 5 9 % activity after heat treatment at 60 C Kinetic studies revealed that this mutation improved thermostability of the enzyme without affecting its catalytic ability to synthesize L threo DOPS. 8 Since the d.e. of the wild type enzyme was only 14%, Gwon et al. preformed additional round s of error prone PCR to improve diastereoselecti vity, yielding six mutants. The best mutant ( A48T ) increased the d.e. by 3.3 fold Unfortunately, even this improvement represented on ly a 43% d.e. still too low for a useful L threo DOPS synthe tic process. T hey t herefore combined several mutations usi ng site directed mutagenesis to create three double variants All double muta n ts retained L threo DOPS synthesis activity and also showed dramatically increased d.e. values compared to the wild type enzyme. The best double muta n t ( Y34C/A28T ) was used to synthesize L threo DOPS in a batch process, yielding 3.7 mg/mL. Although mutagenesis was successful at improving the aldol condensation diastereoselectivity more development would be required before this biocatalytic method could be employed on a large scale. LPC 058 A Potent Gram Negative Bacteria Inhibitor Zhou, Toone, and coworkers recently discovered a difluoromethyl allo threonyl hydroxamate LPC 058, that displayed broad spectrum antibiotic activity toward Gram negative bacteria. Steady state ki netic studies revealed that LPC 058 exhibited up to a 44 fold greater potency over other inhibitors of the same class for a wide variety of

PAGE 35

35 Gram negative pathogens. 153, 154 LPC 058 inhibits UDP 3 O ( R 3 hydroxymyr istoyl) N acetylglucosamine de acetylase (LpxC), which catalyzes the first irreversible step in lipid A biosynthesis Lipid A is the outermost monolayer of the outer membrane of Gram negative bacteria and is comprised of lipopolysaccharides These are ess ential to protect against antibiotics and detergents. 155, 156 I nhibitors that target LpxC bind competitively with the natural substrate (Figure 1 28 ). The long acyl chain of the substrate (and the extended aryl n etwork of the inhibitor) is encapsulated by of the enzyme 156 The initial 10 step synthesis of LPC 058 was re ported by Zhou, Toone and coworkers. This route began from a difluoro threonine methyl ester an d a diacetylene carboxylic acid. These were later coupled to yield the final product (Figure 1 29) 153 The same authors continued their work by optimizing reaction conditions. On a large scale, their initial approach was expensive due to the chiral sulfoxide employed to synthesize the difluoro threonine methyl ester starting material It also required tedious chromatogra phic separation of the diastereomers. They improved the synthesis of both starting materials and also shortened the route by two steps ( F igure 1 3 0 ) 153, 157 Rhizobitoxine A PLP D ependent Enzyme Inhibitor Rhiz obitoxine was first discovered in the 1960s by Owens et al. as a phytotoxin that inhibited greening of new leaf tissue and was produced by certain strains of soy beans such as Rhizvobium japonticum 158, 159 S hortly after its discovery it was also found to inhibit many other pathways such as a PLP dependent enzyme in the methionine biosynthesis by Salmonella typhimurium cystathionase, 160, 161 and ethylene biosynthesis in sorghum seedlings by hinderin g the conversion of methionine

PAGE 36

36 to ethylene. 159 Interest in the absolute configuration 162 and total synthesis 163 of this amino acid peaked when its potential as a herbicide was investigated. 164 During the course of two decades, the effects of rhizobitoxine was investigated in many different plants, such as pears and avocado. 165, 166 Although a total synthesis was established, Mitchell et al. discovered that rhizobitoxine along with L threo hydroxythreonine (a biosynthetic intermediate for this phytotoxin) could be isolated from Pseudomonas andropogonis 167, 168 Shortly after, they established a proposed biosynthetic pathway by using 14 C and 13 C labeled materials. In this pathway 14 C aspartate and 14 C homoserine were tested by feeding experiments and both of t hese precursors led to incorporat ion of the 14 C l abel into rhizobitoxine It was suggested that hydroxythreonine was the common product of both (Figure 1 3 1 ). Although the route from hydroxythreonine to rhizobitoxine is not known, it has been hypothesized that serinol, which is found to accumulate in s oy bean nodules, could directly couple to hydroxythreonine to yield 3 hydroxydihydrorhizobitoxine and upon dehydration would result in rhizobitoxine. Although 3 hydroxydihydrorhizobitoxine has not been found in cell extracts, it may be present in low qua ntities or it may dehydrate immediately to its final form without being released from the cell. 6, 169, 170 Over the past two decades, research on rhizobitoxine has slowed, mainly focusing on its biosynthesis. Some efforts were made to find a positive role for rhizobitoxine. Duodu et al. found that some plant species, like Vigna radiata (mungbean) are susceptible to high levels of ethylene. The use of B. elkanii rhizobitoxine mutant strains was found to jump start aborted nodules at all stages of development. 171

PAGE 37

37 S phingofungins Antifungal Agents The search for new and safe antifungal medications increased greatly in the early 1980s when there was a rise in pa tients suffering from AIDS related fungal infections. 172 A new family of antifungal agents fumifungin, were first discovered in 1987 from A. Fumigatus Fresenius 1863 This led to the discovery of a s imilar class of antifungal agents, sphingofungins 173 which were isolat ed from a different strain of A. Fumigatus (ATCC 20857) (sphingofungins A D) and P Variotii (ATCC 74079) (sphingofungins E and F) by a Merck group in 1992. 174 176 These natural products were identified for thei r antibiotic effects and by their ability to inhibit serine palmitoyltransferase (SPT). 177 S phingofungins A, B, C, and D were the first to be investigated (Figure 1 3 2 ). Sphingofungins A C and E F were found to be potent antifungal agents against C. neoformans an d showed selective activity towards Candida species and others. Sphingofungin D was notably less potent than its counterparts. 176, 177 Since their discovery in 1992, there has been much effort in the total synthesi s of sphingofungins 5 The first 10 step synthesis was published in 1994 by Mori who used N acetyl D mannosamine, 1 heptyne, and ( R ) epoxyoctane as the starting materials 178, 179 Shortly thereafter, Kobayashi et al. developed a 7 step synthetic route utilizin g simple achiral compounds produced by catalytic asymmetric aldol reactions and chiral heterocycles as key intermediates. 180, 181 The most prominent and widely used total synthesis was established by Trost et al. in 2001 by asymmetric allylic alkylation of a gem diacetate pall adium catalyst as the key step and methyl substituted azlactone as starting materials. 182, 183 These total syntheses are summarized in Figure 1 3 3 Sphingofungins act as antifungal agents against a variety of fungal species by inhibiting SPT an enzyme essential in the biosynt hesis of sphingolipids. 176, 177 This

PAGE 38

38 PLP dependent enzyme catalyzes the condensation a fatty acyl CoA (in this case palmitoyl CoA) with serine to biosynthesize 3 ketodihydrosphingosine (Figure 1 3 4 ). Sphingofungi ns are known to inhibit this enzyme alone but not others in the pathway. Zweerink et al. discovered that when S. cerevisiae was grown in the presence of both the inhibitor and the other downstream intermediates in the pathway, sphingolipids were still su ccessfully biosynthesized. 177

PAGE 39

39 Figure 1 1. hydroxy amino a cid s as precursors to or stand alone pharmaceuticals Figure 1 2. Addition of ethyl isothiocyanatoacetate to benzaldehyd e Figure 1 3. Addition of oxazolidinone to benzaldehyde

PAGE 40

40 Figure 1 4. Addition of oxazolidinone to 3 chloro 4 phenoxybenzaldehyde to yield ( a ) the protected amino acid, a precursor to ( b ) vancomycin Figure 1 5. Addition of o xazolidinone to an aldehyde by N N dioxide nickel(II) catalyst complex

PAGE 41

41 Figure 1 6. First documented example of an acid catalyzed aziridine ring open ing Figure 1 7 Generalized summary of aziridine ring openings

PAGE 42

42 Figure 1 8. Optimized synthesis of ( ) chloramphenicol by aziridine ring opening strategies Figure 1 hydroxy amino acid ethyl esters by a 1,3 dipolar cycloaddition with an aziridine and benzaldehyde Figure 1 10. One pot synthesis of trans oxazolidinones by an asymmetric aziridine intermediat e

PAGE 43

43 Figure 1 hydroxy amino acid synthesis Figure 1 syn hydroxy amino acids by oxazolidinon e chiral glycine enolates

PAGE 44

44 Figure 1 13. The natural products ( a ) cyclosporine and ( b ) echinocandin D

PAGE 45

45 Figure 1 14. Caddick et al. route to syn hydroxy amino acids by imidizolidinone ch iral glycine enolates Figure 1 anti hydroxy amino acids by chiral glycine enolates

PAGE 46

46 Figure 1 16. First proposed mechanism for metal catalyzed aza Claisen r earrangements Figure 1 17. Synthesis of hydroxy amino acids by means of aza Claisen rearrangement

PAGE 47

47 Figure 1 18. The synthesis of (2 R ,3 S ) 2 amino 3,4 dihydroxybutyric acid by aza Clais en rearrangement Figure 1 1 9 Dynamic kinetic resolution

PAGE 48

48 Figure 1 20 Dynamic kinetic resolution towards the synthesis of important intermediates Figur e 1 21 (2 S 3 R ) and (2 R 3 R ) p ara c hloro 3 h ydroxytyrosines in vancomycin

PAGE 49

49 Figure 1 22 A hydroxy amino intermediate in ( ) balanol Figure 1 2 3 D ynamic kinetic resolution towards anti hydroxy amino acids

PAGE 50

50 ( a ) Figure 1 24. Proposed mechanism for dynamic kinetic resolu tion for ( a ) anti and ( b ) syn hydroxy amino acids

PAGE 51

51 ( b ) Figure 1 2 4 Continued

PAGE 52

52 Figure 1 2 5 Coupling therapy: selectively targeting dopamine synthesis by methyltyrosine inhibit or

PAGE 53

53 Figure 1 2 6 Organic synthesis strategies to L threo dihydroxyphenylserine

PAGE 54

54 Figure 1 27 Biocatalytic synthesis of L threo dihydroxyphenylserine by L threonine aldolase Figure 1 28 Structure comparison of the natural substrate of LpxC versus the inhibitor

PAGE 55

55 Figure 1 29 Initial Route to LPC 058

PAGE 56

56 Figure 1 30 Final route in the synthes is of LPC 058

PAGE 57

57 Figure 1 3 1 Proposed rhizobitoxine biosynthesis pathway Figure 1 3 2 Sphingofungins A F

PAGE 58

58 Figure 1 3 3 Sphingofungin synthesis strategie s Figure 1 3 4 The first step in the biosynthesis of sphingolipids, serine palmitoyltransferase

PAGE 59

59 CHAPTER 2 i THREONINE ALDOLASE AS BIOCATALYSTS Introduction Aldol condensations are one of the most common ways that nature accomplishes carbon carbon bond formation and/or cleavage. The reaction is widely applicable since many common metabolites especially carbohydrates contain aldehyde or ketone moieties and this allows them to function as either enol(ate) nucleophiles or e lectrophiles. Because the same molecule can serve as both donor and acceptor, the scope of accessible products is very large. The aldol addition product is usually favored, although the exact equilibrium position is dictated by the relative thermodynamic stabilities of the reactants and products as well as their concentrations. A variety of aldolases have evolved to facilitate these conversions in both primary and secondary metabolic pathways. Native aldolases generally tolerate little or no variation i n the enol(ate) partner and this is often used to classify these enzymes. By contrast, aldolases usually accept a wide variety of aldehyde electrophiles, which forms the foundation of t heir synthetic versatility. Interest in using aldolases for non native substrates grew rapidly in the 1980s after a seminal publication by Wong et al 185 Carbohydrates and their derivatives were logical targets for the first generation syntheses involving aldolases because the reactions closely mimicked their normal m etabolic roles and substrate a cceptance was simplified. In this regard, dihydroxyacetone phosphate (DHAP) dependent aldolases i A review of threonine aldolases was published in 2014 by Sarah E. Franz and Jon D. Stewart with the Advances in Applied Microbiology Reference: 184. Franz, S. E.; Stewart, J. D., Chapter Three Threonine Aldolases. In Advances in Applied Microbiology Sima Sariaslani and Geoffrey Michael, G., Ed. Academic Press: 2014; Vol. Volume 88, pp 57 101.

PAGE 60

6 0 found particular utility and many ingenious applications were developed using a diverse range of aldehyde acceptors. The subsequ ent identification of four stereocomplementary DHAP aldolases that provided each of the four possible diastereomeric aldol addition products allowed this technology to mature into a well accepted synthetic methodology 186, 187 The major drawback of DHAP dependent aldolases is their near complete specificity for DHAP. This narrows the scope of accessible products to those c ontaining this substructure (or those derivable by subsequent transformations of the DHAP moiety) and motivated a search for aldolases that accept other enol(ate) donors. These efforts yielded pyruvate dependent aldolases and 2 keto 3 deoxygluconate aldol ase, N acetylneuraminic acid (NeuAc) lyase, and 2 keto 3 deoxy 6 phosphogluconate aldolase, all of which have b een applied to organic synthesis (examples are summa rized in Brovetto et al 2011 and Clapes et al 2010). 186, 187 To date, all known PLP dependent aldolases utilize amino acids as their native substrates (typically glycine, serine, or threonine). In particular, threonine aldolases (TAs) have emerged as useful enzymes for organic synthesis since the aldol reaction creates two n ew, adjacent stereocenters (Figure 2 1 ) These enzymes have been divided into four classes based on their stereochemical preferences: high specificity L and D threonine aldolases and low specificity L and D threonine aldolases. The high specificity enzymes can be further subdivided into the threonine and allo threonine subtypes. It should be noted are in fact highly selective for a particular term arises because they yield a mixed population of carbon configurations.

PAGE 61

61 The possibility of forming only one enantiomer (out of the four potential products) starting from simple, inexpensive building block s has motivated most of the efforts in this research area. This chapter briefly summarizes the range of substrates and products that have been employed with TA s and then describes our structural knowledge of these enzymes and the efforts to use this infor mation to increase their substrate range and stereoselectivities. The Mechanism of Threonine Aldolase All of the aforementioned aldolases follow chemical mechanisms that involve either a metal ion stabilized enol(ate) or a synthetically equivalent enamine intermediate (utilizing an active phosphate (PLP) dependent aldolases follow a fundamentally different pathway ( Figure 2 2). These enzymes first establish a S amino group and PLP (referred to as an external aldimine). The cationic pyridinium ring facilitates deprotonation on the carbon to the amine by an enzyme general base, yielding a highly resonance stabilized anion. This nucleophile adds to the aldehyde acceptor, thereby forming the C C bond between the aldol product and the PLP cofactor The catalytic cycle is completed by an exchange (transaldimination) that transfers the cofactor from the product back to the active site side chain (referred to as an internal aldimine). Threonine Aldolases Utilized for Chemical Synthesis A handful of TA s have dominated the published synthetic applications, particularly the L TA s from Aeromonas jandaei 188, 189 Candida humicola 190 Pseudomonas putida 191, 192 Streptomyces coelicolor 193 Escherichia coli 194 197

PAGE 62

62 Aeromonas veronii 198 Shewanella loihica 198 and Raoultella ornithinolytica 198 along with D TA s produced by Alcaligenes xylosoxidans 7, 191, 192, 199, 200 Pseudomonas sp. 188 Xanthomonas oryzae 196 Arthrobacter sp. 200, 201 Pseudomonas aeruginosa 198 and Pseudomonas protegens 198 In addition to these bona fide TA s, the Hilvert group has developed a mutant alanine racemase from Geobacillus stearothermophilus that endows the variant with D TA activity 202, 203 This alanine racemase is evolutionarily related to D TA s, and the mutation removed one of the two acid base groups re qu ired for alanine epimerization. All of these workhorse enzymes have been cloned and overexpressed in E coli at high levels, which simplifies their use in chemical synthesis. The Griengl group recently created and surveyed a larger collection of these e nzymes in hopes of uncovering examples with higher diastereoselectiviti es 204 Whether C overcome by testing additional wild type isolates or by applying protein engineering technologies to existing TAs awaits experimental testing. Substrate Selectivity of Threonine Aldolase In 2014, we published a book chapter that contained a complete list of synthetic applications of TAs that were published by early 2014 (Tables 2 1 2 4). 184 The examples have been grouped first by the amino acid donor nucleophile (glycine, alanine, serine, or cysteine; Tables 2 1 2 4, respectively). Within each table, examples are ordered by increasing size and structural complexity of the aldehyde acceptor (alkyl aldehydes followed by aryl aldehydes). In most cases, aldol products were not isolated from the reaction mixtures and only fractional conversions based on chromatographic analysis (typically HPLC) are available. Stereochemical purities were

PAGE 63

63 almost universally assessed by chromatographic separations. Finally, it should be noted that large excesses of the amino acid were typically employed to drive reactions toward the desired aldol product. For these reasons, the fractional conversion achieved in a given example should only be taken as a rough guide with regard to estimating th e synthetic feasibility of a preparative scale reaction. Glycine/Alkyl Aldehydes Simple alkyl aldehydes can be converted to L anti product s with high diastereoselectiviti es by the E coli L TA (Table 2 1, entries 4, 11, and 16). 196 The active site of this enzyme can accommodate relatively large n alkyl aldehydes, although the anti diastereoselectivity and fractional conversion decline as aldehyde size increas es (Table 2 1, entries 24 and 30). To date, a D TA with comparable levels of diastereoselectivity has not been identified, although the A xylosoxidans enzyme can show good syn selectivity in favorable cases ( e.g ., Table 2 1, entry 19). TAs generally tole rate relatively bulky and highly substituted aldehyde acceptors. Halo alkoxy and amino moieties are acceptable, even when the latter are derivatized by large protecting groups such as benzyl and Cbz ( Table 2 1, entries 40 59). The main drawback is that diastereomeric mixtures are usually obtained and the pr eferences for C stereochemistry are relatively modest. Aldehyde acceptors with substituents even relatively large ones are also tolerated by TA s ( Table 2 1, entries 60 78). The general conclusion is that nearly all alkyl aldehydes can serve as accep tors for glycine; however, it is likely that a mixture of diastereomers will be obtained at the carbon. Preexisting chiral centers in the aldehyde have modest impa cts on diastereoselectivity ( Table 2 1, entries 57 59 and 75 79).

PAGE 64

64 Table 2 1. Synthetic applications of threonine aldolases using glycine as the donor Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e Entry References C. humicola L TA 30 40% N/A 1 190 P. putida L TA n.d. 4% ( anti ) 2 192 E. coli L TA 40% (24 h) 82% ( anti ) 3 196 E. coli L TA 35% (3 h) 99% ( anti ) 4 196 A. xylosoxidans D TA n.d. 2% ( anti ) 5 192 X. oryzae D TA 60% (3 h) 6% ( syn ) 6 196 X. oryzae D TA 50% (24 h) 6% ( syn ) 7 196 C. humicola L TA 30% n.d. 8 190 P. putida L TA 71% 28% ( syn ) 9 192 E. coli L TA 18% (24 h) 97% ( anti ) 10 196 E. coli L TA 15% (3 h) 99% ( anti ) 11 196 A. xylosoxidans D TA 52% 9% ( syn ) 12 192 X. oryzae D TA 37% 3.3% ( syn ) 13 196 P. putida L TA 55% 42% ( syn ) 14 192 E. coli L TA 10% (3 h) 76% ( anti ) 15 196 E coli L TA 1% (24 h) 99% ( anti ) 16 196 Y265A Ala racemase 0% n.d. 17 202 X. o ryzae D TA 49% 86% ( syn ) 18 196 A. xylosoxidans D TA 24% >95% ( syn ) 19 192

PAGE 65

65 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry References P putida L TA 94% 10% ( syn ) 20 192 A xylos oxidans D TA 65% 54% ( syn ) 21 192 P putida L TA 92% 31% ( syn ) 22 192 E coli L TA 16% (24 h) 26% ( anti ) 23 196 E c oli L TA 7% (3 h) 84% ( anti ) 24 196 X oryzae D TA 31% (3 h) 22% ( syn ) 25 196 X oryzae D TA 23% (24 h) 28% ( syn ) 26 196 A xylos oxidans D TA 42 % (30% DMSO) 68% ( syn ) 27 192 A xylos oxidans D TA 33% (no cosolvent) 73% ( syn ) 28 192 P. putida L TA 25% 9% ( syn ) 29 192 E. coli L TA 2% 44% ( anti ) 30 196 X. oryzae D TA 3% 38% ( syn ) 31 196 A. xylos oxidans D TA 12% 55% ( syn ) 32 192 P putida L TA 29% 23% ( anti ) 33 192 A x ylos oxidans D TA <1% n. d. 34 192

PAGE 66

66 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry References P putida L TA 33% 15% ( anti ) 35 192 A x ylos oxidans D TA <1% n.d. 36 192 P putida L TA 11% 15% ( anti ) 37 192 A xylos oxidans D TA <1% n.d. 38 192 C h umi cola L TA 30% n.d. 39 190 P. putida L TA 50% 93% ( syn ) 40 192 A. xylos oxidans D TA 30% 97% ( syn ) 41 192 C humi cola L TA <5% n.d. 42 190 P putida L TA 65% 40% ( syn ) 43 192 A x ylos oxidans D TA 26% 73% ( syn ) 44 192

PAGE 67

67 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver si on d.e. Entry Refer ences P putida L TA 20% 73% ( syn ) 45 192 A x ylos oxidans D TA 6% 82% ( syn ) 46 192 C humi cola L TA >75% n.d. 47 190 E coli L TA 13% 0% 48 197 E coli L TA 30% 20% ( syn ) 49 194 C humi cola L TA >75% 84% ( anti ) 50 190 E coli L TA 36% 88% ( anti ) 51 196 X. oryzae D TA 80% 40% ( syn ) 52 196 C humi cola L TA 45 75% 84% ( anti ) 53 190 C humi cola L TA 45 75% 84% ( anti ) 54 190 E coli L TA 18% 40% ( syn ) 55 194 C humi cola L TA 30% n.d. 56 190

PAGE 68

68 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry Referen ces E coli L TA 129% 41% ( syn ) 57 194 E coli L TA 54% 64% ( syn ) 58 194 E coli L TA 40% 68% ( syn ) 59 194 C humi cola L TA 10 30% n.d. 60 190 E coli L TA 10% (30% DMSO) 44% ( anti ) 61 196 E coli L TA 5% (no co solvent) 66% ( anti ) 62 196 X oryzae D TA 16% 74% ( syn ) 63 196 C humi cola L TA 53% 6% ( anti ) 64 190 E coli L TA 10% 88% ( anti ) 65 196 X oryzae D TA 45% (3 h) 29% ( syn ) 66 196 X oryzae D TA 35% (25 min) 64% ( syn ) 67 196 E coli L TA 11% 0% 68 194 C humi cola L TA >75% n.d. 69 190

PAGE 69

69 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry References C humicola L TA 10% n.d. 70 190 C humicola L TA 10 30% n.d. 71 190 C humicola L TA 10% n.d. 72 190 E coli L TA 67% 0% 73 197 E coli L TA 34% n.d. 74 197 E coli L T A 35% 92% ( anti ) 75 196 X oryzae D TA 73% 40% ( syn ) 76 196 E coli L TA 70% 0% 77 196 X oryzae D TA 84% 76% ( syn ) 78 196

PAGE 70

70 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. En try References C humicola L TA 30% n.d. 79 190 T maritim a L allo TA 25% (5 min) 76% ( syn ) 80 204 P putida L TA 80% (30 min) 21% ( syn ) 81 192 P aerugi n osa L TA 80% (30 min) 21% ( syn ) 82 204 A jandaei L allo TA 30% (5 min) 27% ( syn ) 83 204 P putida L TA 40% (1 min) >30% ( syn ) 84 192 P aerugin osa L TA 10% (1 min) >30% ( syn ) 85 204 E coli L TA 9% (24 h) 60% ( syn ) 86 196 E coli L TA 3% (3 h) 71% ( syn ) 87 196 A jandaei L allo TA <20% (<1 min) anti 88 204 S cerevisiae L low TA 60% (5 h) 22% ( anti ) 89 204 S cerevisiae L low TA 4% (1 m in) 40% ( anti ) 90 204 C humicola L TA 45% 40% ( anti ) 91 190 B bronchi septica L low TA 10% (5 min) 70% ( anti ) 92 204 S pomperoyl D low TA 80% (5 days) 21% ( syn ) 93 204 X oryzae D TA 10% (24 h) 73% ( syn ) 94 196

PAGE 71

71 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry References X oryzae D TA 11% (3 h) 74% ( syn ) 95 196 Y265A Ala racemase 17% (24 h) 76% ( syn ) 96 202 Y265A Ala racemase 10% (3 h ) 97% ( syn ) 97 202 A xylosoxidans D TA 79% 98% ( syn ) 98 192 P putida L TA 68% 35% ( syn ) 99 192 A xyloso xidans D TA 68% 95% ( syn ) 100 192 P. putida L TA 90% 52% ( syn ) 101 192 A. xylos oxidans D TA 27% 67% ( syn ) 102 192 P. putida L TA 79% 34% ( syn ) 103 192 A. xylos oxidans D TA 6% 35% ( syn ) 104 192 Y265A Ala racemas e <1% n.d. 105 202 P. putida L TA 99% 32% ( syn ) 106 192 E. coli L TA 93% (24 h) 42% ( anti ) 107 196

PAGE 72

72 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry References E. coli L TA 46% (3 h) 68% ( a nti ) 108 196 X. oryzae D TA 89% (3 h) 44% ( syn ) 109 196 A. xylos oxidans D TA 18% 65% ( syn ) 110 192 Y265A Ala racemase 1% >97% ( syn ) 111 202 P. putida L TA 6 4% 27% ( syn ) 112 192 A. xylos oxidans D TA 54% 81% ( syn ) 113 192 P putida L TA 69% 30% ( syn ) 114 192 A xy losoxidans D TA 60% 85% ( syn ) 115 192 P putida L TA 63% 55% ( syn ) 116 192

PAGE 73

73 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry Ref erences A xylosoxidans D TA 43% 71% ( syn ) 117 192 E coli L TA 43% 46% ( syn ) 118 196 P putida L TA 56% 51% ( syn ) 119 192 X oryzae D TA 53% (24 h) 46% ( syn ) 120 196 X oryzae D TA 54% (3 h) 48% ( syn ) 121 196 Y265A Ala racemase 3% 70% ( syn ) 122 202 A xylosoxidans D TA 76% 86% ( syn ) 123 192 P putida L TA 74% 21% ( syn ) 124 192 A xylosoxidans D TA 90% 80% ( syn ) 1 25 192 Y265A Ala racemase 55% (24 h) 85% ( syn ) 126 202 Y265A Ala racemase 20% (3 h) 93% ( syn ) 127 202 E coli L TA 17% 20% ( anti ) 128 196 X oryzae D TA 25% 14% ( anti ) 129 196

PAGE 74

74 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry Referen ces P putida L TA 51 % 29% ( syn ) 130 192 A xylosoxidans D TA 42% 91% ( syn ) 131 192 P putida L TA 57% 17% ( syn ) 132 192 A xyl osoxidans D TA 26% 86% ( syn ) 133 192 P putida L TA 47% 14% ( syn ) 134 192 A xylosoxidans D TA 12% 74% ( syn ) 135 192 P putida L TA 11% 36% ( syn ) 136 192 C humicola L TA 30% 40% ( anti ) 137 190 A xylos oxidans D TA 15% 70% ( syn ) 138 192

PAGE 75

75 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Con version d.e. Entry Referen ces P putida L TA 79% 24% ( syn ) 139 192 E coli L TA 53% (24 h) 6% ( anti ) 140 196 E coli L TA 35% (3 h) 28% ( anti ) 141 196 X oryzae D TA 88% (24 h) 10% ( syn ) 142 196 X o ryzae D TA 72% (3 h) 16% ( syn ) 143 196 Y265A Ala racemase 36% (24 h) 40% ( syn ) 144 202 A xylos oxida ns D TA 31% 75% ( syn ) 145 192 P putida L TA 68% 53% ( syn ) 146 192 A xylos oxidans D TA 63% 99% ( syn ) 147 192 P putida L TA 92% 24% ( syn ) 148 192 A xylos oxidans D TA 53% >90% ( syn ) 149 192

PAGE 76

76 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry Referen ces P putida L TA <1% n.d. 150 192 S coelicolor L TA n.d. 14% ( syn ) 151 193 V81I/R241C/ Y306C S coelicolo r L TA n.d. 21% ( syn ) 152 193 R241C/A287V S coelicolor L TA n.d. 21% ( syn ) 153 193 Y306C S coelicolor L TA n.d. 26% ( syn ) 154 193 Y36C/Y306C/ R316C S coelicolor L TA n.d. 28% ( syn ) 155 193 Y39C/Y306C S coelicolor L TA n.d. 38% ( syn ) 156 193 Y39C/T306C/ A48T S coelicolor L TA n. d. (fourfold of wt) 43% ( syn ) 157 193 E coli L TA 71% 60% ( syn ) 158 195 A xylosoxidans D TA <1% n.d. 159 192 Y265A Ala racemase 5% 70% ( syn ) 160 202 P putida L TA 15% 16% ( syn ) 161 192

PAGE 77

77 Table 2 1. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry Referen ces A xylos oxidans D TA 16% 46% ( syn ) 162 192 E coli L TA 21% (24 h) 40% ( anti ) 163 196 E coli L TA 19% (3 h) 50% ( anti ) 164 196 X oryzae D TA 49% (24 h) 30% ( syn ) 165 196 X oryzae D TA 42% (3 h) 52% ( syn ) 166 196 E coli L TA 40% 32% ( syn ) 167 196 X oryzae D TA 60% 22% ( syn ) 168 196 A xylos o xidans D TA 70% 99% ( syn ) 169 200 Arthro b acter sp. D TA 70% 99% ( syn ) 170 200 Note: Alkyl aldehydes are shown first, in the approximate order of increasing size and structural complexity. Aryl a ldehydes follow, also in the order of increasing size and structural complexity. When a given reaction has been carried out by more than one threonine aldolase, entries are arranged in the order of increasing anti selectivity, followed by increasing syn s electivity for both L and D selective aldolases. n.d., not determined or not reported. d. e diastereomeric excess = % major diastereomer % minor diastereomer. Glycine/Aryl Aldehydes Because it is a critical component of some semisynthetic lactam antibiotics, phenylserine has been a key target for TAs ( Table 2 1 entries 80 98). A number of L

PAGE 78

78 TA s yield the desired target from benzaldehyde and glycine, although the diastereoselectivity i s incomplete. By contrast, A xylosoxidans D TA pr oduces the D syn product with both high conversion and excellent diastereoselectivity ( Table 2 1, entry 98). This is one of the more successful uses of TA s in preparative synthesis. Interestingly, the engineered alanine racemase also gives very good D sy n diastereoselectivity in this reaction, although this is tempered by poor fractional conversion ( Table 2 1, entries 96 and 97). A variety of monosubstituted benzaldehydes have been tested as glycine acceptors by a variety of TAs ( Table 2 1 entries 99 149). In many cases, the A xylosoxidans D TA affords good diastereoselectivity as does the a la nine racemase point mutant. The P putida L TA also accepts a wide variety of monosubstituted benzaldehydes; however, the diastereoselectivities are generally poor to moderate. The Yamada group published an early study focused on 4 methylthiophenylserine involving the same substrate pair as entries 146 and 147, but employing a low specificity D TA from Arthrobacter sp. DK 38 201 Unfortunately, the diastereomeric purity of t he product was not reported. Next to phenylserine itself, 3,4 dihydroxyphenylserine has been the most popular application of TAs in preparative synthesis since this compound has been used to treat Table 2 1 entries 151 158). Earlie r synthetic routes involved difficult separations of diastereomeric mixtures; the possibility that an enzyme could directly yield only the desired material was a major impetus in exploring TAs as an alternative. Interestingly, some further derivatization of the hydroxyl groups of 3,4 dihydroxybenzaldehyde can be tolerated, for example, 3,4 methylenedioxy or 3,4

PAGE 79

79 dimethoxy ( Table 2 1 entries 160 162), but monomethoxy analogs were not accepted by the enzyme. 191, 192 In 2015, the Goldberg group from Bristol Myers Squibb used the D TAs from A. xylosoxidans and Arthrobacter sp. DK 38 to synthesize (2 R ,3 S ) 2 amino 3 hydroxy 3 (pyridin 4 yl)propanoic acid, a precursor to a developmental drug candidate The reactions wer e carried out at 4 C with a 10 : 1 molar ratio of glycine : aldehy de. The hydroxy amino acid precipitated from the solution and was isolated by simple filtration with 9 9 % d.e and 70% overall conversion (Table 2 1, entry 169 and 170) 200 Th e use of precipitation to drive the aldol product formation was a simple and elegant solution to the otherwise unfavorable equilibrium problem that also avoids the erosion of diastereoselectivity after extended reaction times. Ot her A mino A cid D onors (D Ala, D Ser, and D Cys) After screening a variety of native TAs, the Griengl group identified two enzymes that accepted more complex amino acids in addition to glycine. 188 Despit e forming only L threonine and analogs when glycine was the donor, the A jandaei enzyme could also accept D Ala as a substrate. More surprisingly, L Ala was not accepted. A panel of representative aldehydes was tested as partners for D Ala, and the resu lts paralleled those observed for Gly with the same aldehydes ( Table 2 2 ). Despite the relatively poor diastereoselectivities (caused by mixtures at the C chiral center), these results are significant since a quaternary, non racemizable center is created at the carbon with very high enantioselectivities. This remains a difficult challenge in organic synthesis. In addition to the A jandaei enzyme, a D TA from Pseudomonas sp. also utilized D Ala and a variety of aldehyde acceptors, albeit with modest d iastereoselectivities.

PAGE 80

80 The same pair of aldolases described in the preceding text also accepted D Ser and D Cys as nucleophiles in aldol reactions ( Tables 2 3 and 2 4, respectively). 188, 198 While neither fracti onal conversion nor diastereoselectivities were high, the ability to exercise high control over a quaternary carbon center is an equally impressive achievement. Very recently, ii Fesko et al. searched for more TAs that might have broader amino donor specificity by using the sequences of A. jandaei L allo TA and Pseudomonas sp D TA to search sequence databases f or related proteins Ten TAs were found with sequence similarities of 55 85 % compared to the genes. Among these, only five show ed aldol condensation with D Ala and D Ser (Tables 2 2 and 2 3, respectively) 198 These included A. veronii L TA, S. loihica L TA, R. ornithinolytica L TA, P. aeruginosa D TA and P. protegens D TA. Among the mo st impressive examples were provided by the D TAs with the m nitrobenzaldehyde and D Ala (Table 2 2, entries 33 and 34) This gave moderate conversions (>45%) but high diastereoselectivities (>80% for the syn isomer). Table 2 2. Synthetic applications of threonine aldolases using alanine as the donor Aldehyde Acceptor Product Enzyme Yield or Conver s ion d.e. Entry References A jandaei L allo TA 20% 46% ( anti ) 1 188 Pseudomonas sp. D TA 54% 42% ( syn ) 2 188 ii Stewart, but included here to complete the story of amino donor selectivity of threonine aldolase.

PAGE 81

81 Table 2 2. Continued Aldehyde Acceptor Product Enzyme Yield or C onver s ion d.e. Entry References A jandaei L allo TA 6% 26% ( anti ) 3 188 A. veronii L TA 45% 21% ( syn ) 4 198 S. loihica L TA 6% 11% ( anti ) 5 198 R. ornithinolytica L TA 8% 5% ( anti ) 6 198 Pseudom onas sp. D TA 32% 66% ( syn ) 7 188 P. aerug inosa D TA 29% 60% ( syn ) 8 198 P. protegens D TA 21% 17% ( syn ) 9 198 A. jandaei L allo TA 58% 8% ( anti ) 10 188 A. veronii L TA 48% 32% ( anti ) 11 198 S. Ioihica L TA 4% 18% ( anti ) 12 198 R. ornithinolytica L TA 18% 21% ( anti ) 13 198 Pseudomonas sp. D TA 84% 33% ( syn ) 14 188 P. aeruginosa D TA 55% 13% ( syn ) 15 198 P. protegens D TA 30% 16% ( syn ) 16 198 A. jandaei L allo TA 35% 6% ( anti ) 17 188 A. veronii L TA 30% 0 18 198 S. Ioihica L TA 10% 74% ( syn ) 19 198 R. ornith inolytica L TA 21% 0 20 198 Pseudom onas sp. D TA 11% 65% ( syn ) 21 188

PAGE 82

82 Table 2 2. Continued Aldehyde Acceptor Product Enzyme Yield or Conver s ion d.e. Entry References P. aeruginosa D TA 11% 65% ( syn ) 22 198 P. protegens D TA 9% 55% ( syn ) 23 198 A. jandaei L allo TA 24% 35% ( syn ) 24 188 Pseudomonas sp. D TA 21% 95% ( syn ) 25 188 A. janda ei L allo TA 60% 7% ( anti ) 26 188 A. veronii L TA 92% 4% ( syn ) 27 198 S. Ioihica L TA 23% 16% (a nti ) 28 198 R. orni th inolytica L TA 64% 40% ( syn ) 29 198 Y265A Ala racemase 12% (24 h) 65% ( syn ) 30 202 Pseudomonas sp. D TA 36% 76% ( syn ) 31 188 Y265A Ala racemase 6% (24 h) 80% ( syn ) 32 202 P. aeruginosa D TA 69% 80% ( syn ) 33 198 P. protegens D TA 45% 92% ( syn ) 34 198 A. veronii L TA 97% 20% ( syn ) 35 198 S. Ioihica L TA 18% 73% ( syn ) 36 198 R. ornith inolytica L TA 52% 12% ( syn ) 37 198

PAGE 83

83 Table 2 2. Continued Aldehyde Acceptor Product Enzyme Yield or Conver s ion d.e. Entry Referen ces P. aerug inosa D TA 56% 88% ( syn ) 38 198 P. prote gens D TA 38% 80% ( syn ) 39 198 A veronii L TA 29% 28% ( anti ) 40 198 S. Ioihica L TA 10% 36% (a nti ) 41 198 R. ornith inolytica L TA 27% 31% ( anti ) 42 198 P. aerug inosa D TA 27% 17% ( syn ) 43 198 P. prote gens D TA 16% 49% ( syn ) 44 198 Note: Alkyl aldehydes are shown first, in the approximate order of increasing size and structural complexity. Aryl aldehydes follow, also in the order of increasing size and structural complexity. When a given reaction has been carried out by more than one threonine aldolase, entries are arranged in the order of increasing anti selectivity, followed by increasing syn selectivity for both L and D selective aldolases. n.d., not determined or not reported d e diastereomeric excess = % major diastereomer % minor diastereomer. Table 2 3. Synthetic applications of threonine aldolases using serine as the donor Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry References A jandaei L allo TA 6% 65% ( anti ) 1 188 Pseudom onas sp. D TA 23% 11% ( anti ) 2 188 A. jandaei L allo TA 30% 45% ( anti ) 3 188 A. veronii L TA 23% 41% ( anti ) 4 198

PAGE 84

84 Table 2 3. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry References S. Ioihica L TA <1% 95% ( anti ) 5 198 R. ornith inolytica L TA 6% 53% ( anti ) 6 198 Pseudom onas sp. D TA 43% 24% ( anti ) 7 188 P. aeruginosa D TA 11% 2% ( anti ) 8 198 P. protegens D TA 8% 9% ( anti ) 9 198 A. jandaei L allo TA 10% 40% ( anti ) 10 188 A. veronii L TA 15% 55% ( syn ) 11 198 S. Ioihica L TA 4% 62% ( syn ) 12 198 R. ornithinolytica L TA 6% 54% ( syn ) 13 198 Pseudomonas sp. D TA <1% n.d. 14 188 P. aeruginosa D TA n.d. n.d. 15 198 P. protegens D TA n.d. n.d. 16 198 A. jandaei L allo TA 15% 65% ( anti ) 17 188 A. vero nii L TA 39% 22% ( anti ) 18 198 S. Ioihica L TA 2% 0 19 198 R. ornith inolytica L TA 8% 0 20 198 Pseudom onas sp. D TA <5% 23% ( anti ) 21 188 P. aerug inosa D TA 4% 23% ( anti ) 22 198 P. protegens D TA 1% 4% ( syn ) 23 198

PAGE 85

85 Table 2 3. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry Referen ces A. veronii L TA 25% 44% ( anti ) 24 198 S. Ioihica L TA 5% 80% ( anti ) 25 198 R. ornith inolytica L TA 4% 95% ( anti ) 26 198 P. aerug inosa D TA 5% 79% ( anti ) 27 198 P. protegens D TA n.d. n.d. 28 198 A. veronii L TA 17% 60% ( anti ) 29 198 S. Ioihica L TA n.d. n.d. 30 198 R. ornith inolytica L TA 4% 95% ( anti ) 31 198 P. aerug inosa D TA 6% 35% ( anti ) 32 198 P. prote gens D TA n.d. n.d. 33 198 Note: Alkyl aldehydes are shown first, in the approximate order of increasing size and structural complexity. Aryl aldehydes follow, also in the order of increasing size and structural complexity. When a given reaction has been carried out by more than one threonine aldolase, entries are arranged i n the order of increasing anti selectivity, followed by increasing syn selectivity for both L and D selective aldolases. n.d., not determined or not reported. d e diastereomeric excess = % major diastereomer % minor diastereomer. Table 2 4. Synthe tic applications of threonine aldolases using cysteine as the donor Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry References A jandaei L allo TA 27% 18% ( anti ) 1 188

PAGE 86

86 Table 2 4. Continued Aldehyde Acceptor Product Enzyme Yield or Conver sion d.e. Entry References Pseudomona s sp. D TA 33% 20% ( anti ) 2 188 A jandaei L allo TA 30% 12% ( anti ) 3 188 Pseudomonas sp. D TA 39% 6% ( anti ) 4 188 Note: Alkyl aldehydes are shown first, in the approximate order of increasing size and structural complexity. Ary l aldehydes follow, also in the order of increasing size and structural complexity. When a given reaction has been carried out by more than one threonine aldolase, entries are arranged in the order of increasing anti selectivity, followed by increasing sy n selectivity for both L and D selective aldolases. n.d., not determined or not reported. d e diastereomeric excess = % major diastereomer % minor diastereomer. Structure s of Threonine Aldolase s In addition to the extensive catalytic characterizati on studies carried out with TAs, our understanding of their structures has also increased in recent years. Several representative crystal structures are known, although some key enzymes used widely for synthesis remain unsolved. Interestingly, L and D T As are structurally and evolutionarily distinct, the former belonging to the aspartate aminotransferase family and the latter to the alanine racemase family. Currently, much more is known about the structures of L TA s than their D counterparts. T. maritim a L allo Threonine Aldolase The first crystal structure of a TA was reported by Kielkopf and Burley in 2001 and 2002 (PDB codes 1JG8 and 1M6S ). 205 This is a low specificity enzyme produced

PAGE 87

87 by a thermophilic organism. Its amino acid sequence is similar to thos e of several L TAs that have been applied to chemical synthesis (including those from A jandaei E coli P aeruginosa and S cerevisiae ). In addition to the native enzyme, cocrystallized forms with either Gly (PDB code 1LW5) or L allo Thr (PDB code 1L W4 ) were also solved. External aldimine formation with either amino acid did not change the overall protein structure apart from a small rotation of the PLP ring. Of the four active sites in the homotetrameric enzyme, two were fully occupied with substra te/product, a third was partially occupied and the fourth showed only the resting form (the internal aldimine, in diversity of active site occupancies reflects kinetic cooper ativity or was an artifact of the crystallization process remains unknown. The hydroxyl group of bound L allo Thr interacted with both an active site h is tidine side chain (residue 83) and an ordered water molecule that was in turn hydrogen bonded to the phosphate of PLP. This suggests that the side chain of His 83 might act as a general acid base group in the catalytic mechanism. E. coli L Threonine Aldolase Safo, Contestabile, and coworkers solved several X ray crystal structure s of the low specificity L TA from E coli 206 The overall structure of the unliganded enzyme (PDB code 4LNJ) was similar to that of the T maritim a L allo TA described above 205 One key difference is that the active sites of the homotetrameric E coli enzyme are composed of residues that converge from three subunits. iii In common with other iii Active sites in most homologues are formed at the dimer interface between two subunits.

PAGE 88

88 enzymes in this family, each monomer consists of two domains with the PLP cofactor located at their interface and bou 197. Cocrystallizing E coli L TA with L Ser yielded a mixture of glycine bound to PLP with a fraction in which the cofac tor was covalently bound to Lys 197 (PDB code 4LNM ). 206 External aldimine formation did not change the overall protein structure, although the PLP ring rotated slightly within the active site, a phenomenon observed previously in other PLP d ependent enzymes. When E coli L TA was cocrystallized with L Th r (PDB code 4LNL ), one active site contained the glycine/PLP covalent complex and the other contained a mixed population of this species along with a mixture of PLP covalently bound to L Thr and L allo Thr. Importantly, the side chain hydroxyl groups of b oth amino acids were located in the same position; their respective methyl groups occupied different locations. This has important ramifications for the catalytic mechanism since the aldol/ r etro aldol reaction involves direct acid base interactions with t his functional group. It also suggests ways to modify the active site to enhance carbon stereoselectivity. The hydroxyl group mak es hydrogen bonds with both His 83 and His 126, suggesting that these might play a role as general acid/base catalysts. S urprisin gly, the mutation of either His 83 or His 126 alone was tolerated by the enzyme, although some variants at these positions were unstable and precipitated after extended storage. M oreover, replacing His 126 with Asn or Phe actually increased k cat va lues by up to three fold, although compensating changes in K M blunted the impact somewhat. These results argue against a direct acid base role by either His side chain, which was further supported by the retention of some catalytic activity in a double Hi s 83 / His 126 variant

PAGE 89

89 At this time, the identity of the group responsible for hydroxyl p rotonation/ deprotonation remains a mystery. Two repl acements were also made for Phe 87, which had been proposed to be a key determinant of substrate specificity. Th e properties of both the Ala and Asp mutants were not in accord with predictions, leaving this as another open issue. The authors also pointed out that the active site is much larger than needed to accommodate the relatively small substrate acetaldehyde. Furthermore, L allo Thr is not a recognized metabolite in E coli It is therefore possible that the true physiological role of this enzyme despite threonine may actually involve different (and possibly multiple) substrates i n the native host rather than Thr and/or allo Thr. A. jandaei L allo Threonine Aldolase iv In 2014, Tanokura, Shimizu, Kawabata, and coworkers solved the crystal structure of both the wild type enzyme and the H128Y/S292R variant of the L allo TA from A. jan daei DK 39 (PDB codes 3WGB and 3WGC, respectively). 207 The overall structure of this aldolase and its mutant variant are similar to the other two structures that were highlighted earlier 205, 206 The TA family has been reported as homotetramers and funct ion as catalytic dimers by forming interfaces between the two large domains. The most important and conserved active site residues in the TA family are Ser 8, His 85, Arg 171, Lys 199, and Arg 313. The amino acid re sidues Arg 171, Arg 313, and Ser 8 anch or the substrate/product to the active site by interactions between the aldimine and the amino acid side chains. The side chain of Lys 199 forms the Schiff base with iv Not included Stewart, but included here to complete the story of the structure of threonine aldolase.

PAGE 90

90 PLP and any mutations at this site render the enzyme inactive. 189 Finally, His 85 is responsible for the regulation of the degree of stereospecificity between the L and L allo stereomers and pi stacks with the pyridinium ring of PLP to help stabilize the active site 205 207 The noteworthy differences between the wild type enzyme an d the H128Y/S292R double mutan t occurred between Ala 123 and Pro 131. When His 128 was mutated to Tyr, the residue moved 4.2 out of the active site 207 The electron density map for the resi due at position 292 in both the wild type and mutant structures was poor and therefore could not confirm how this position differed. The L allo TA from A. jandaei prefers the substrate L allo Thr to that of L Thr with k cat / K M values of 26.9 mM 1 s 1 and 0.0 0203 mM 1 s 1 respectively (Table 2 5) The double mutation H128Y/S292R presented a three fold and 322 fold increase in k cat / K M towards L allo Thr and L Thr, respectively ( compared to the wild type enzyme ) Additionally, t he single mutations at these two positions showed similar k cat / K M towards L allo Thr, however the H128Y mutant exhibited a higher increase in k cat / K M toward L Thr than the S292R mutant, supporting the notion that the mutation of His 128 to Tyr was likely the site of improved activity. T he enhanced activity for the S292R mutant could not be explained due lack of structural information at the distorted loop where the position resides. Table 2 5. Kinetic p arameters for L allo t hreonine a ldolase from A. jandaei and its m utants Enzyme L allo Thr L Thr k cat (s 1) K M (mM) k cat / K M (mM 1 s 1 ) k cat (s 1) K M (mM) k cat / K M (mM 1 s 1 ) Wild Type 13.8 0.513 26.9 0.641 31.6 0.00203 H128Y/S292R 22.0 0.386 83.3 2.14 0 3.27 0.654 00 H128Y 16.5 0.402 41.0 1.81 0 4.48 0.404 00 S292R 18.3 0.444 41.2 0.913 32 .7 0.0279 0 Note Reference: Qin et al. 2014

PAGE 91

91 The side chain of His 128 (along wit h that of His 85) was found to be in hydrogen bonding distance to recognize the hydroxyl group on L allo Thr and explain s why this enzyme prefers L allo Thr as the native subst rate. Surprisingly Qin et al. discovered that when this residue was mutated to a Tyr, the residue moved 4.2 outwards from the active site to form a new hydrogen bond with Val 31, expanding the active site and explain ing why this mutation has broad subs trate stereoselectivity. To confirm this, they carried out site saturation mutagenesis of His 128 and dis covered that only a few amino acid substitutions ( i.e. Tyr, Phe, and Met) resulted in an improved activity towards L allo Thr 207 On the other hand, replacement by Tyr, Phe, Leu, Ser, Met, Tr p, Ile, Lys, and Val resulted in an increased activity towards L Thr as compared to the wild type enzyme. Most hydrophobic substitutions showed an increased activity towards L Thr due to the hydrophobic side chains and m ethyl group of L Thr. Overall, the mutant that showed highest activity towards both L allo Thr and L Thr was H128Y. A xylosoxidans D Threonine Aldolase v Very recently, Gruber, Schrmann, and coworkers crystallized the very first D TA from A. xylosoxidans (PDB code 4V15). 208 In order to improve crystal quality, they methylated the D TA by the selective lysine methylation method established by Rayment in 1997 209 that successfully enhanced the quality of the crystals from 3.5 to well under 2 resolution Interestingly, there was no indication of methylate d lysines in the electron density showing some process during the methylation Based on homology models, D TAs belong to the v Franz and Jon D. Stewart, but included here to complete the story of the structure of threonine aldolase.

PAGE 92

92 alanine racemase family (type III fold) 210 which differ s from that of L TAs, indicating a different evolutionary origin. Unlike L TAs, the presence of divalent ions (particularly manganese) for aldolase activity in D TAs is crucial. 7, 211 In the structure of A. xylosoxidans D TA, a Mn 2+ ion was modeled into the ion binding site located relatively close to PLP. The octahedral coordination sphere of the Mn 2+ io n was coordinated by two amino acid residues, His 347 a nd Asp 349 and four water molecules. The Mn 2+ ion binding site was located 5 from the aldehyde group of PLP and nearly coplanar with the pyridinium ring of the cofactor. The role of the Mn 2+ ion was determined by modeling (2 R ,3 S ) phenylserine as an ex hydroxyl group of the substrate with a Mn O distance of 2.3 , replacing a water molecule that is typically coordinated to the metal in the active site. The overall structure of A. xylosoxidans D TA is completely different from that of L TAs ; 205 207 however t he proposed mechanism for D TAs derived from this crystal structure is similar to that of L TAs. The only noteworthy difference between the mechan isms is the use of the Mn 2+ ion by D replacement for the second His residue in L TAs hydroxyl group of the substrate. Although the overall structures are unalike, D TAs and L TAs are classified as enantio complementarity enzymes which by definition are enzymes that fold differently, but whose active sites are mirror images. 212 In this case, they both include the active site His that deprotonates the hydroxyl group of the substrate, but the His residue in D TA is positioned at the re face of PLP and His residue in L TA is positioned at the si face.

PAGE 93

93 Other Threonine Aldolase Structures In addition to the more complete studies descri bed in the preceding text, several other TA crystal structures have been solved as part of structural genomics projects with little or no additional data available. One example from Leishmania major was reported in 2011 (PDB code 1SVV). A second example was described in 2010, a TA from Listeria monocytogenes EGD E (PDB code 3PJ0). No published synthetic applications of these enzymes have appeared in the literature as of late 201 7 Very recently, Hirato et al. successfully crystallized the D TA from Chla mydomonas reinhardtii at a 1.85 resolution and they are currently solving the structure. 213 The final structural entry, which lacks a formal literature citation, is the phenylserine aldolase from P. putida (PDB code 1V72). Protein Engineering Studies of Threonine Aldolases Recent years have witnessed signific ant improvements in protein engineering methodologies, high throughput screening, and selections along with structure determination using X ray crystallography. These developments have made it possible for even smaller academic laboratories to undertake p rotein mutagenesis projects aimed at improving the performance of biocatalysts in synthetic applications. Surprisingly, only a relatively few examples where these techniques have been applied to TAs have been published as of late 201 7 and these are summar ized in the succeeding text. The selection and/or screening methodology dictates the number of variants that can be screened, and this is nearly always the limiting factor in protein engineering studies. In cases where L Thr is the desired product, enabli ng growth of a Thr auxotroph in minimal medium can be used to select the desired variants. Even when

PAGE 94

94 substrates beyond L easily be applied to libraries containing up to 10 10 variants. v i The danger is that the best mutant for a novel conversion may have lost the ability to accept glycine and/or also more challenging to devise growth based high thro ughput assays that directly interrogate stereoselectivity (both enantioselectivity and diastereoselectivity) without resorting to GC or HPLC analyses. In such cases, library sizes are practically limited to < 1000 members unless pooling/deconvolution strat egies are employed. 214 Improving Catalytic Activity Lee and coworkers devised a growth based selection for TAs with greater catalytic efficiencies based on the observation that aldehydes such as acetaldehydes depress the growth rate of E coli 215 In the synthetic direction, TAs consume aldehydes; by depleting the local medium of the toxic substrate, transformed E coli cells grow at correspondingly faster rates (positive selection). In principle, this is a generally applicable strategy that should allow the most active TA variants to predominate. The authors also considered a n egative selection based on an analogous strategy (using the acetaldehyde from Thr degradation to inhibit cell growth). Unfortunately, this proved impossible to implement in practice since the levels of acetaldehyde never rose to toxic levels. The positi ve selection strategy was applied to P aeruginosa L TA with the goal of increasing its catalytic efficiency. 215 Error prone PCR yielded a library of ca. 20,000 colonies that were grown in the presence of 20 mM acetaldehyde under conditions vi In these cases, the library size is limited primarily by transformation efficiency.

PAGE 95

95 where the cloned TA variants were overexpressed. The initial selection provided ten hits, and plasmid DNA was isolated from each. After retransformation, five of the ten plasmids retained the ability to confer high level acetaldehyde resistance and these were further characterized. The best variant exhibited a two fold improvement in catalytic activity as compared to the wild type. Despite the relatively modest impact on catalytic activity, the selection method devised in this study may be useful in other protein engineering studies. In cases where expanded substrate range is desired, supplementing the growth medium with both acetaldehyde and the amino donor might allow one to identify the desired variants. vii This se lection scheme should also be applicable to other aldehyde acceptors, although the precise concentrations needed for cell toxicity will need to be established empirically for each substrate. Improving Thermostability As described previously, the stereosele ctive synthesis of 3,4 dihydroxy phenylserine has been an important synthetic application for TAs. Baik and coworkers cloned and expressed an L TA from S. coelicolor A3(2) as the basis for their strategy. 8 While the wild type enzyme provided an acceptable reaction rate and stereoselectivity, its longevity under process conditions was too low for practical use. Error prone PCR was therefor e used to introduce random changes throughout the entire length of the protein. Approximately, 15,000 clones were individually screened for the ability to degrade L Thr after a 65 C heat treatment step (using a colorimetric assay for the vii In practice, it may also be necessary to reduce glycine l evels to favor reaction with the amino donor of interest. This can be accomplished by employing a stringent glycine auxotroph and supplementing the growth medium with limiting concentrations of glycine. Hilvert and coworkers have developed one example of a glycine auxotrophic host strain that might be useful in this strategy 216. Giger, L.; Toscano, M. D.; Bouzon, M.; Marliere, P.; Hilvert, D., Tetrahedron 2012, 68 (37), 7549 7557.

PAGE 96

96 acetaldehyde by product). This search yielded eight variants that appeared to be more thermostable than the wild type enzyme; four were chosen for additional studies. All four had only single amino acid changes and all four changes wer e unique. The best mutant (H177Y ) retained 86% of the original activity after 20 min at 60 C ; under similar conditions, the wild type retained only 11%. Importantly, greater thermostability was not achieved at the expense of catalytic activity, and the H177Y variant had essentially the s ame steady state kinetic values as the parent protein. In whole cell format, the improved variant performed at the same level for 20 successive batch reactions and provided a final product concentration of 4 g/ L Because the targeted level of enzyme impro vement was reached after one generation of mutagenesis and selection, the beneficial mutations were not examined combinatorically nor were additional random mutations added to the best first generation variants. In the absence of data, one must speculate as to whether the observed thermostabilities are the best that can be reached. The major drawback to the final process is that the molar yield of the final product was only 0.7%, based on the aldehyde added (glycine was present in vast excess). Wieteska a nd coworkers used site directed mutagenesis at the interchain interface of T. maritima L allo TA viii to improve its thermostability. These efforts were based on the crystal structure reported by Kielkopf et al 205, 217 They targeted sites that allowed for additional salt bridges or new intrachain disulfide bonds. Among the ten predicted thermostable muta nt s, only two actually showed increased thermo tolerance viii Stewart, but included here to co mplete the story of the improving thermostability of threonine aldolase.

PAGE 97

97 ( P56C and A21C ). The P56C mutation increased stability by 10 15% on average with no loss in activity. 217 Improving Stereo selectivity Enzymes are by nature homochiral catalysts, and the ability to direct reactions into single product enantiomers is one of their most important attributes. As noted previously, TAs create two adjacent stereocenters, and as a general rule, these enzymes are carbon. By contrast, they often have relatively carbon. Reversibility also plays a role in governing stereoselectivity. For any chiral cente r, a racemic mixture is always the thermodynamic minimum. When more than one stereocenter is present (as is the case for TA s), a diastereomeric mixture of enantiomers nearly always occurs at equilibrium. While the precise composition depends on the actua l product structure, it is rare that a single diastereomer predominates at equilibrium for most synthetically interesting targets. For these reasons, it is almost always essential that preparative reactions be carried out under kinetically controlled cond itions and the reverse (retro aldol) reaction should be avoided as much as possible. The main drawback is that the yield of the desired product is almost always low when reactions are limited to far from equilibrium conditions. Several years ago, the Grien gl group carefully analyzed the properties of r epresentative TA s 204 This study included members of all four available classes: high sp ecificity L Thr L allo Thr and D Thr types and a low specificity L Thr type. The formation of phenylserine from glycine and benzaldehyde was chosen as the model system. As expected, all four TA types yielded the same thermodynamic mixture of products ( 60 : 40 syn : anti ) after extended reaction times. Of the four classes, only the

PAGE 98

98 high specificity D TA provided high diastereoselectivity under kinetic conditions in the early phase of the reaction; the rest gave mixtures of products from the start. Int erestingly, the high specificity D TA reached the equilibrium product mixture only after five days (compared to 5 h for the other three aldolase types). This observation is critical since it demonstrates conclusively that high diastereoselectivity combin ed with good product yield is possible using TAs showed conclusively that rapid carbon epimerization is not an intrinsic flaw of TAs This implies that the problem can be removed from other TAs using the appropriate pr otein engineering. Griengl and coworkers subsequently carried out an NMR study to understand why some TAs catalyzed rapid product epimerization at the carbon while others did not. 204 13 C Labeled syn product was mixed in a 60 : 40 ratio with unlabeled anti product in the presence of glycine, benzaldehyde, and enzyme that matched their equilibrium values. Based on the known chemical mech anism for TAs, the relevant species can be deduced ( Figure 2 3 ). Both carbon epimerization and the back reaction to free glycine proceed via the cofactor stabilized anion that results from retro aldol cleavage of the syn product external aldimine. The NMR study yielded the relative rates of 13 C label transfer from the syn product to the anti product and to free glycine, which can also be described as the partition ratio for the cofactor stabilized anion. In comparing data from the four TAs, three yielded partition ratios ranging from 0.5 to 20 (epimerization/back reaction ). As expected, these values correlated with the initial diastereoselectivities of the reactions. Because it was not possible to measure microscopic rate constants under the experimental conditions, the relative net rate constants could not be further

PAGE 99

99 de composed into the individual contributions from the multiple steps that occur in each branch. It might be possible to measure the relative contributions of proton transfer steps and aldehyde binding/release by incorporating additional isotopic labels. Th is would provide extremely valuable guidance for future protein engineering studies by focusing the improvements on the most relevant step(s) in the reaction pathway. Introducing and Optimizing Threonine Aldolase Activity into a Novel Scaffold In 2003, the Hilvert group reported that a single amino acid substitution was sufficient to convert a PLP dependent alanine racemase from G stearothermophilus into a TA. 218 The specific mutation (Y 265A) was designed to allow a histidine side chain (His 83) to act as an acid base group for oxyanion protonation/deprotonation in the aldol reaction. In addition, the smaller Ala s ide chain created additional active site volume to allow larger substrates to bind. While the catalytic activity was relatively modest when assessed against the standard glycine/benzaldehyde benchmark reaction, the mutant was more than five orders of magnitude more efficient than the starting racemase. The mutant enzyme was also highly selective for the D configuration at the ca rbon. The preference, if any for carbon stereochemistry, was not reported. In a follow up study, the ability of the Y265A mutant to accept methyl substrates was assessed. 219 This is an important application since aldol condensations that yield quaternary centers remain particularly challenging. Steady state kinetic constants were measured for three substrates ( Figure 2 4 ). Interestingly, the steady state kinetic values for hydrogen a nd methyl (2 R ,3 S ) diastereomers were very similar, implying that steric bulk at the carbon was well tolerated. This may be a consequence of the evolutionary heritage of the enzyme, which originally bound Ala.

PAGE 100

100 The D anti analog had a ca. ten fold lowe r k cat value, but K M was also decreased by a similar extent so that the k cat / K M ratio was nearly the same. Given the large number of microscopic rate constants in the catalytic cycle of TAs, it is very difficult to determine which individual step(s) was i mpacted and a deeper understanding will require future pre steady state kinetic investigations. To support future protein engineering studies, the Hilvert group recently developed a new growth based selection system for TAs. 216 Rather than target Thr, their strategy uses an engineered E coli strain with four simultaneously inactivated in minimal medium when supplemented with the amino acid or when retro aldol activity by a cloned TA yields glycine. The advantage of this selection is that the substrate of interest can be directly interrogated. This includes selection for stereoselectivity if a diastereomerically pure Thr analog added t o the growth medium. limitation is that glycine must be the amino acid partner in the aldol/retro aldol reaction. The utility of this screen was demonstrated by creating a library of simultaneous random amino acid replacements at four p ositions in a previously uncharacterized L TA from Caulobacter crescentus CB15. The data revealed that only one of the four amino acids (His 91) was absolutely essential for catalytic activity. Conclusions and Future Work TAs have been clearly established as useful catalysts for asymmetric organic synthesis. Their ability to control the stereochemistry at the carbon is excellent and they accept a diverse array of acceptor aldehydes. On the other hand, these enzymes have several drawbacks that must be o vercome before they can be employed routinely. The lack of stereochemical control at the carbon is a significant problem that detracts

PAGE 101

101 C epimerization from retro aldol clea vage. Enhancing these properties is an obvious target for protein engineering efforts. The establishment of several selection/screening methodologies for TAs should simplify these studies. The other majo r drawback of TA s is that high substrate concentrati ons are usually needed to drive the conversion to products (and avoid equilibrating conditions that erode diastereomeric purities). While glycine is inexpensive, many aldehydes of synthetic interest do not share this trait and this practically limits the range of usable substrates. One possibility is to employ coupled enzyme systems that further convert the aldol product in an effectively irreversible reaction, for example, by lipase mediated acylation or redox conversions. Similar strategies have proven quite useful in transaminations and may provide inspiration to this field as well 220

PAGE 102

102 Figu re 2 1. Aldol condensation with threonine aldolase Figure 2 2. Mechanism of threonine aldolase

PAGE 103

103 Figure 2 3. Threonine aldolase kinetic pathway Figure 2 4. Aldol products from a mutant Ala racemase

PAGE 104

104 CHAPTER 3 SUBSTRATE PROFILING OF THREE THEONINE ALDOLASES Introduction Hydroxy amino acids comprise an important class of natural products that often serve as building blocks for antibiotics such as vancomycin 1 polyoxin A 2 and rhizobitoxine 6 Several pharmaceuticals also incorporate these building blocks, most notably d roxidopa (L threo 3,4 dihydroxyphenylserine, L DOPS), an important therapy for P brain barrier is the key to its effectiveness 7, 8 D Glucosaminic acid (2 amino 2 deoxy D gluconic acid), produced by Aeromonas oxydans can act as an artificial sweetener 221 More recently, D g lucosaminic acid has been found useful as a building block for several glycosidase inhibitors, such as (2 S ,4 S ,5 R ) 4,5,6 trihydroxynorleucine. 12, 13 Many strategies hydroxy amino acids in optically pure form have been devised. Chemical approaches include a chiral glycine enolate 14 Sharpless dihydroxylation, epoxid ation, aminohydroxylation 15 17 and aza Claisen rearrangements of allylic acetimidates 18 among others. The target compounds can also be prepared by enzymatic aldol additions of glycine to aldehydes using threonine aldolases (TAs ) (Figure 3 1). 184, 222, 223 These PLP dependent enzymes feature an active site lysine 7 subsequently acts as a general base during the catalytic cycle. In general, TAs are nearly specific for glycine as the nucleophile, although some exceptions have been noted, e.g. L Ala, L Ser and L Cys. 184, 188, 189, 198, 222, 223 By contrast, TAs tolerate a wide variety of ald ehyde acceptors, ranging from long chain aliphatic aldehydes to

PAGE 105

105 substituted benzaldehydes and heteroaromatic aldehydes. 184, 222, 223 It is this latter property that has generated interest in using TAs for preparat ive synthesis. While a large number of aldehydes have been tested as substrates for TAs, much substrate spectrum remains to be explored. One goal of this study was to better define the substrate range and stereoselectivity of three key family members, Aer omonas jandaei L allo threonine aldolase (L allo TA), Escherichia coli L threonine aldolase (L TA) and Thermotoga maritima L allo threonine aldolase (L allo TA). A second important goal was to develop general methods for isolating the aldol products from these reactions, which would make TAs much more useful for preparative synthesis. In a few favorable cases, the hydroxy amino acid precipitates in nearly pure form from the reaction mixture ; 224 however, this is not generally the case and instea d, a highly polar product must be isolated from an aqueous mileu that also contains buffers, salts, proteins and (usually) a large excess of glycine. Reactions catalyzed by TAs are readily reversible and a large excess of glycine is often used to favor the desired aldol conversion. Unfortunately, this seriously amino acids with similar ionic properties are hydroxy amino acid product). We s ought to overcome this problem by selective degradation of glycine after the completion of the aldol reaction. To the best of our knowledge, this strategy has not been explored previously. Glycine oxidase is an FAD dependent enzyme that oxidatively deami nates glycine to yield glyoxalate (Figure 3 2). Enzymes of this family are involved in thiamin pyrophosphate biosynthesis and are also important in degrading the herbicide glyphosate. 225, 226 For this study, we c hose glycine oxidase

PAGE 106

106 from Bacillus subtilis since it has been cloned, overexpressed in E. coli and screened against a wide range of amino acids. 227, 228 B. subtilis glycine oxidase accepts only Gly and small D amino acids. Its substrate selectivity is therefore orthogonal to the product spectrum of threonine aldolases, which w ould allow it to be added directly to a TA reaction mixture with no possibility of degrading the desired aldol product. The reversibility of TA catalyzed reactions has serious, negative consequences carbon configuration is essentially complete, carbon. This loss of stereochemical integrity is a complex interplay of enzyme, substrate and reaction conditions. For this reason, we chose several representative TAs and two time points in order to identify combinations that afforded both high product concentration s and high diastereoselectivities. Results and Discussion Gene C loning and P rotein O verexpression Genes encoding the three required TAs were obtained by colony PCR 229 ( E. coli L TA and T. maritima L allo TA) or by chemical gene synthesis ( A. jandaei L allo TA). Each was individually ligated into pET 15b and the resu lting plasmids were used to transform the E. coli overexpression strain BL21 Gold(DE3). All TAs were efficiently overproduced by the recombinant strains, which allowed crude lysates to be employed for screening reactions. Control reactions were carried o ut for each aldehyde substrate with a crude extract from an un transformed E. coli strain; none gave significant aldol product.

PAGE 107

107 Derivatization of Amino Acid s for Analysis In order to analyze the overall conversion and diastereomeric excess of the natural p roducts as they are being synthesized we worked through a series of analytical techniques for derivatizing amino acids for GC/MS or HPLC detection The challenge was that w e not only needed to detect the presence of the products but also separate the di astereomers. We first employed a technique used by Khuhawar et al. using trifluoroacetylacetone ( Figure 3 3a) ; however there was no detection of L Thr and it s diastereomer L allo Thr using this strategy. 230 Our next approach was to deploy the standard OPA/NAC method ( Figure 3 3b) established by Nimura and Kinoshita in 1986 and used in most TA papers. 231 233 Although derivatization was a success, Thr and Gly derivatives were close in retention time to each other. Since Gly is used in a five times molar excess in the reactions, separation would have been an issue. The bigger reason for finding another derivatization method was that the der iv atized amino acids were analyzed by HPLC and since we planned to screen over a dozen substrates, the lack of an auto sampler would have made analysis tedious. Therefore, we sought for a derivatization method that allowed for GC/MS analysis since an autosampler was readily available in our lab N Methyl N (trimethylsilyl) trifluoroacetamide (MSTFA) de rivatization was first 234 by Yoon ( Figure 3 3c) MSTFA h as been used in the silyla tion of numerous compounds such as phenolic s, sterols, and sugars. 234 236 The original protocol allowed for three evaporation events to ensure samples were completely free from water including the remov al of the original solvent system addition and removal of 80% MeOH, and lastly addition and removal of methylene chloride. This monotonous

PAGE 108

108 evaporation protocol was explored and optimization revealed that evaporation under speed vacuum was only required one time to remove the origina l solvent system. The derivatization conditions required 70 L of MSTFA and 40 L of dried pyridine for every 5 20 nmol of amino acid at 37 C and vigorous shaking (250 rpm) for 30 minutes However, s ince MSTFA is somewhat expensive and pyridine is toxic we fine tuned the quantities of both derivatization reagents. Optimization revealed that only 50 L of MSTFA and 1 L of dried pyridine were required for derivatization. The derivatization techniques explored are summarized in F igu re 3 3. Additionally, it was possible to detect un derivatized Gly and amino acids products by t hin layer chromatography (TLC) with a ninhydrin s tain 197 Ninhydrin reacts with primary amine s so this was an obvious method for the detection of amino acids This method was only used as an initial detection of product formation and subsequent MSTFA derivatization was required for conversi on and selectivity values. Optimization of R eaction C onditions Prior to carrying out extensive aldehyde screening studies, we used the DOE methodology to identify optimal conditions for each of the three overexpressed TAs. Based on previous literature, th e starting point was 25 C, pH 8 and a 5 fold molar excess of glycine versus the aldehyde. Variations of each of these three parameters were investiga ted: reaction pH (5, 7, 8, 9.7 and 12), reaction temperature (4 C 18 C 25 C 37 C and 42 C ) and th e glycine : aldehyde ratio (2, 4, 8, and 10). A total of 92 individual reactions were carried out for each enzyme substrate pair and both relative conversion and product diastereomeric excess values were determined for each. Aldehydes chosen to optimiz e each enzyme gave measurable, but incomplete

PAGE 109

109 conversions under standard conditions; this allowed both improved and detrimental changes to the reaction conditions to be identified. For A. jandaei L allo TA, reactions involved glycine and aldehyde 10 (Table 3 1). The starting conditions gave a relative conversion i of 0.11 using a 10 : 1 ratio of glycine : aldehyde. Conversion increased significantly as pH, temperature and substrate molar ratio were increased. Solving for the optimal conditions (pH 9.7, 37 C and a glycine : aldehyde ratio of 10) gave a relative conversion of 0.68, more than a 6 fold improvement. In addition, an 87% d e was obtained, underscoring the power of this process improvement strategy. The outcomes of reactions catalyzed by E. col i L TA and T. maritima L allo TA were also improved by the DOE strategy, albeit more modestly. These studies involved aldehydes 6 and 10 respectively (Table 3 1). The E. coli enzyme performed best under the same conditions as its A. jandaei counterpart while the T. maritima enzyme required slightly higher pH value of 12. Table 3 1. Substrate specificity of L TA catalyzed aldol reactions Entry Aldehyde Product Enzyme a Reaction Time (hr) Conver sion b (%) d.e. a (%) 1 1 A 4 20 22 28 99 99 B 4 20 <5 c 6 c 99 C 4 20 <5 c 13 -c 99 i Relative conversions for these studies were defined as the peak area ratio of MSTFA derivatized product/internal standard using GC/MS.

PAGE 110

110 Table 3 1. Continued Entry Aldehyde Product Enzyme a Reaction Time (hr) Conver sion b (%) d.e. (%) 2 2 A 4 20 10 23 99 ( anti ) 96 ( anti ) B 4 20 <5 c <5 c -c -c C 4 20 <5 c 16 -c 33 ( anti ) 3 3 A 4 20 2 9 32 19 ( syn ) 23 ( syn ) B 4 20 10 17 61 ( anti ) 51 ( anti ) C 4 20 20 35 12 ( syn ) 37 ( syn ) 4 4 A 4 20 52 71 n.d. d n.d. d B 4 20 -<5 c --C 4 20 ----5 5 -A 4 20 ----B 4 20 ----C 4 20 ----6 6 A 4 20 13 18 28 17 B 4 20 <5 c <5 c -c -c C 4 20 <5 c 7 -c 17 7 7 A 4 20 6 12 47 21 B 4 20 <5 c <5 c -c -c C 4 20 <5 c 6 -c 9

PAGE 111

111 Table 3 1. Continued Entry Aldehyde Product Enzyme a Reaction Time (hr) Conver sion b (%) d.e. (%) 8 8 -A 4 20 ----B 4 20 ----C 4 20 ----9 9 A 4 20 11 19 31 ( anti ) 26 ( anti ) B 4 20 <5 c <5 c -c -c C 4 20 <5 c 14 -c 12 ( anti ) 10 10 A 4 20 30 63 73 ( syn ) 30 ( anti ) B 4 20 -<5 c --c C 4 2 0 31 47 20 ( syn ) 18 ( syn ) 11 11 A 4 20 43 70 11 18 B 4 20 <5 c <5 c -c -c C 4 20 <5 c 21 -c 4 12 12 A 4 20 37 66 49 ( syn ) 41 ( syn ) B 4 20 <5 c 25 -c 34 ( anti ) C 4 20 <5 c 14 -c 6 ( syn ) 13 13 -A 4 20 ----B 4 20 ----C 4 20 ----

PAGE 112

112 Table 3 1. Continued Entry Aldehyde Product Enzyme a Reaction Time (hr) Conver sion b (%) d.e. (%) 14 14 A 4 20 29 58 99 ( syn ) 24 ( syn ) B 4 20 -<5 c --c C 4 20 10 35 56 ( syn ) 12 ( syn ) 15 15 A 4 20 5 8 27 ( syn ) 20 ( syn ) B 4 20 -<5 c --c C 4 20 <5 c <5 c -c -c 16 16 A 4 20 73 80 n.d. d n.d. d B 4 20 <5 c 8 -c n.d. d C 4 20 <5 c 7 -c,d n.d. d 17 17 A 4 20 47 68 n.d. d n.d. d B 4 20 <5 c 8 -c n.d. d C 4 20 <5 c 11 -c n.d. d 18 18 A 4 20 8 13 n.d. d n.d. d B 4 20 <5 c 7 -c n.d. d C 4 20 <5 c 6 -c n.d. d Note: Reaction mixtures contained 500 mM glycine, 100 mM aldehyde, 0.05 mM PLP, 20% ethanol, and Cl, pH 9.7 (Enzymes A, B) or 50 mM CAPS, pH 12 (Enzyme C). Reactions were incubated at 37 C. a Enzyme A: A. jandaei L allo TA; Enzyme B: E. coli L TA; Enzyme C: T maritima L allo TA. b Conversion and diastereomeric excess values were determined by GC MS after MSTFA derivatization. Reactions with 1 e omeric purity are provided since the small peak sizes preclude accurate integration. c Accurate values for conversion and diastereomeric excess could not be determined due to the very low peak areas. d It was not possible to determine the diastereoselectivity of this reaction by chiral phase GC since temperat ures required for elution were greater than the maximum column temperature.

PAGE 113

113 Screening of A ldehyde A cceptors Once optimized reaction conditions had been identified, the substrate ranges of the three selected L TAs were investigated using various aldehyde a cceptors (Table 3 1). Aldehydes 1 and 9 are well known substrates for TA catalyzed reactions; they were included here to allow comparisons to previous studies. Aldehydes 2 and 4 6 have not previously been tested as TA substrates. 184 Cyclohexanecarboxaldehyde 3 is a very sterically demanding acceptor, and previous studies had failed to identify a suitable TA for this substrate. 237 Hetereocyclic aldehyde 10 was used very successfully with a D TA by Go ldberg et al ; 224 here, our goal was to identify a stereocomplementary L TA. Aldehydes, 11 15 have not been previously studied as TA substrates to the best of our knowledge. The o rtho halogenated analogs 16 18 previously explored for other TA s by Steinreiber et al were included for comparison. 191 All screening react ions used a 5 : 1 molar ratio of glycine : aldehyde. Under these conditions, conversions were expected to be modest; the goa l was to identify useful enzyme/ substrate combinations, rather than to form high product titers in this phase of the study. All thr ee L TAs were found to accept a wide range of new aldehydes, although the overall performance of the A. jandaei L allo TA was nearly always superior. As expected, diastereoselectivities at 4 hr were greater than those at 20 hr; however, very low conversio ns at 4 hr often thwarted accurate stereochemical measurements. The smallest substrate (propionaldehyde 1 ) gave only a single diastereomer regardless of reaction time ( Table 3 1, entry 1). This is particularly interesting since the active site binding po cket can accommodate much larger aldehydes in multiple orientations. That a small aldehyde seemingly occupies only one orientation with very high preference testifies to the complexity of TA substrate binding interactions. Valeral dehyde 2 also

PAGE 114

114 gave high diastereoselectivity with A. jandaei L allo TA as well as with the T. maritima enzyme ( Table 3 1, entry 2). Bulky aldehyde 3 was accepted by all three enzymes examined. While the diastereoselectivities were modest, it was possible to find enzymes that fa vored both the syn and anti diastereomers ( Table 3 1, entry 3). Neither D threose nor D erythrose were accepted significantly by any of the three enzymes ( Table 3 1, entries 4 and 5). ii Furan thiophene and pyrrole 2 carboxaldehydes presented an intere sting contrast ( Table 3 1, entries 6 8). While the first two aldehydes gave some conversion, the last was not a substrate for any of the three TAs. The conversions are correlated with the carbonyl IR stretching frequencies, which give an approximate me asure of electrophilicity iii We also tested a variety of aromatic aldehydes as possible acceptors for TAs ( Table 3 1, entries 9 18). Our data for benzaldehyde itself confirmed the modest anti selectivity that had been observed previously. Pyridines 1 0 and 11 can be considered formal benzaldehyde analogs. While both were also accepted by two of the TAs, the syn diastereomers were favored in each case ( Table 3 1, entries 10 and 11). The presence of an o chloro substituent allowed the aldehyde to be ac cepted by all three TAs, with one favoring the syn product, one the anti product and the third with essentially no diastereoselectivity ( Table 3 1, entry 12). The intriguing stereochemical results from aldehyde 12 prompted us to investigate a series of o rt ho substituted benzaldehydes ( Table 3 1, entries 13 18). All but o salicylaldehyde 13 afforded aldol products with varying degrees of ii The aldol product of D erythrose was confir med by MS (Figure A 1). iii Furfural, 1700 cm 1, thiophene 2 carboxaldehyde, 1680 cm 1, pyrrole 2 carboxaldehyde, 1650 cm 1

PAGE 115

115 diastereoselectivity. A. jandaei L allo TA was particularly useful for these substrates, although o iodobenzaldehyde 1 8 was a poor substrate, even for this enzyme. Given the ability of the TAs to accept relatively bulky o rtho substituents ( Table 3 1, entries 14 18), the inability to form aldol products from 13 was puzzling. We tested the ability to 13 to bind to the ac tive site of A. jandaei L allo TA by adding equal concentrations of 9 and 13 As expected, only the aldol product from 9 was formed. In addition, the product concentration was approximately half that from an analogous reaction lacking 13 That 13 can ac t as a competitive inhibitor supports the notion that can bind to the active site, but once bound, cannot undergo nucleophilic addition. Screening of Amino Donors Although these enzymes were found to accept a wide variety of aldehyde acceptors their high selectivity for glycine is a disadvantage (Figure 3 4 ). The amino donors 21 23 (Figure 3 4 a) showed no conversion to the aldol product, even though as previously stated, some L TAs show conversion with these amino donors. 188, 189, 198, 222, 223 Additionally, o ther unconventional amino donors were also investigated such as amino donors 24 and 25 but unfortunately these substrates showed no conversion to the aldol product ( Figure 3 4 b ). Optimizing Isolation and Purification Procedure The screening results suggested that TAs wer e powerful tools in the synthesis of hydroxy amino acids, however, an effective method for isolation and separation from starting materials, particularly Gly, has not been described S ince the reaction require d five equivalents of Gly to afford suitable yields, the purification process of separating large amounts of Gly from the aldol product wa s very difficult Only a few methods were found in the literature for hydroxy amino acids from its Gly

PAGE 116

116 starting material including MeOH precipitations, 192, 196 ion exchange chromatography, 196, 197, 238 activated carbon, 238 and silica gel chromatography. 192, 197 The first attempt at solving this problem began with the l arge scale reaction with Gly and aldehyde 10 The unreacted aldehyde was extracted with ether and the aqueous phase lyophilized to afford a white residue. MeOH precipitate d most of the Gly, enzyme, and phosphate buffer salts. The mixture was filtered an d MeOH evaporated under reduced pressure for 1 H NMR analysis. It was found that d euterated MeOH cannot be used in this approach, as MeOH overlaps with the chemical shift of Gly, however deuterated water (D 2 O) was the obvious choice as it dissolves all ami no acids. In the first isolation attempt, a large amount of Gly was present in our crude product, showing that further purification was required. Additional MeOH precipitation steps were employed in an attempt to allow for easy isolation of the products w ithout using column chromatography ; however traces of Gly were always found in the 1 H NMR spectrum. We t herefore explored chromatographic strategies to separate Gly from the aldol products. After an initial MeOH precipitation step, the crude product was applied to an ion exchang e resin to allow for separation. Despite testing different elution solvents glycine could not be separated from the product with this strategy. Silica gel chromatography was the second strategy investigated for purifying the amin o acid products The appropriate solvent system was established as 50% EtOH and 5% AcOH in water through TLC with a ninhydrin stain. Although this method was slightly better at separating Gly from the desired aldol products, only around 20% of the aldol products were isolated With the failure of these two purification strategies, we

PAGE 117

117 explored a biocatalytic alternative to remove the excess Gly prior to downstream purification steps. One obvious approach to removing excess glycine was to exploit glycine o xidase, a n FAD dependent enzyme that oxidatively deaminated Gly to glyoxylate ( Figure 3 2 ). The glycine oxidase gene from B subtilis was chemically synthesized by Genscript, cloned into a pET 15b vector, and overexpressed in E. coli The advantage of th is enzyme was that it was extremely selective for Gly and small D amino acids Administration of any larger amino acid, including threonine and the TA catalyzed aldol products, resulted in no conversion to the corresponding glyoxylate analogue. The crude extract containing B. subtilis glycine oxidase was purified by binding its hexahistidine tag to a HiTrap Chelating HP column. After elution, the p urified enzyme was concentrated to 5 mg /mL and activity was measured by monitoring the formation of H 2 O 2 by UV Vis spectroscopy at 500 nm by a coupled HRP assay. 228 Purified g lycine oxidase (2 mg) was added to the TA reaction mixture after a single MeOH precipitation step to remove most of the glycine, followed by ion e xchange chromatography to separate glyoxylate from the aldol product (Figure 3 5). By utilizing this biocatalyst, problems of isolating and purifying hydroxy amino acids w ere solved. Preparative C onversions Taken together, our screening results suppor t the current view that TA catalyzed reactions have broad substrate tolerance, but suffer from a loss of diastereoselectivity as the reaction progresses. Employing a large molar excess of glycine is a common approach to maximizing conversion over short re action times. When only small scale screening reactions are considered, this is not a problem; however, if preparative

PAGE 118

118 reactions are contemplated, the hydroxy amino acid purified. Because these downstream operations have generally received less attention, we chose six examples from our screening efforts for preparative reactions (Table 3 2). Glycine oxidase was used to simplify downstream processing by removing excess glycine from the reaction mixture (Figure 3 5 ). Table 3 2. Preparative scale reaction results Entry Aldehyde Major Product Isolated Yield (%) M ole F raction anti : syn a 1 2 16 0.60 : 0.40 2 3 50 0.14 : 0.86 3 9 32 0.33 : 0.66 4 14 22 0.10 : 0.90 5 10 28 0.33 : 0.66 6 12 50 0.33 : 0.66 Note: Reactions were catalyzed by A. jandaei L allo threonine aldolase at pH 9.7. Excess glycine was r emoved by glycine oxidase prior to aldol product isolation (Figure 2 3). See Figures A 2 7 for NMR spectra. a The diastereomeric composition was determined by NMR analysis ( anti erythro ; syn = threo ). Because A. jandaei L allo TA provided the best conversions and diastereoselectivities, all preparative scale reactions employed this enzyme. Screening reaction conditions were scaled up ten fold. After the 4 hr, any unreacted a ldehyde was

PAGE 119

119 removed by extracting with diethyl ether, then the water was removed by lyophilization. The solid was extracted with two portions of methanol, which left most unreacted glycine and buffer salts undissolved. After evaporating the methanol, the residue was dissolved in buffer at pH 8.0 and glycine oxidase was added to decompose any remaining glycine. An anion exchange resin eluted with 0.5% aqueous acetic acid was used for final purification prior to final lyophilization the afforded the aldol products. While some yields were disappointing, e.g. entries 1 and 4, others were more synthetically relevant. The low yields were mainly due to poor conversions by the enzymes, rather than to losses during isolation. Assignment of R elative C onfiguration s iv The diastereomeric compositions of aldol products isolated from all six preparative scale reactions were elucidated by the J analysis method of Matsumori 239 and the results are shown in Table 3 2 Elucidation of the relative stereochemistry of two chiral carbons separated by a single bond imp lies simultaneous elucidation of the rotamer equilibrium. The expected coupling constants in the three staggered rotamers of the two diastereomers are given in Table 3 3. As is often the case, two of the rotamers here Ia and IIa display the same patt ern of coupling constants. The values for large and small coupling constants were taken from the values reported by Matsumori 239 for the case where the two carbons carry oxygen atoms: 3 J H H : large 7 10 Hz, small 0 4 Hz; 3 J C H : large 5 7 Hz, small 1 3 Hz; 2 J C H large 4 to 6 Hz, small 0 2 Hz. iv All relative configurations were assigned by Ion Ghiviriga at the University of Florida by NMR analysis.

PAGE 120

120 Table 3 3. Expected values for diagnostic coupling constants in pure staggered conf ormers of the two diastereomers (2 S ,3 S ) + (2 R ,3 R ) anti (erythro) Ia Ib Ic 3 J H2 H3 Hz lg sm sm 3 J H2 C4 Hz sm sm lg 3 J H3 C1 Hz sm lg sm 2 J H2 C3 Hz lg sm lg 2 J H3 C2 Hz lg lg sm (2 S ,3 R ) + (2 R ,3 S ) syn (threo) IIa IIb IIc 3 J H2 H3 Hz lg sm sm 3 J H2 C4 Hz sm sm lg 3 J H3 C1 Hz sm sm lg 2 J H2 C3 Hz lg sm lg 2 J H3 C2 Hz lg sm lg The H H coupling constants were measured in the proton spectrum. The H C coupling constants were measured in the f1 dimension of EXSIDE spectra. 240 The values are given in Table 3 4. The 1 H and 13 C chemical shifts assignments, on which the elucidation of the relative stereochemistry relies, were based on the cross peaks seen in the gHMBC spectra. v v For Tables and 2D NMR, see supporting information section, Table A 1 and Figures A 2 7.

PAGE 121

121 Tab le 3 4. Diagnostic coupling constants measured in the two diastereomers of M and m Rotamer, x Ic + Ib, 0.60 IIb + IIa, 0.40 IIb, 0.86 0.14 IIb + IIa, 0.66 Ic + Ia, 0.33 3 J H2 H3 Hz 3.7 sm 4.5 sm m 3.3 sm 3.8 sm 4.3 sm m 4.2 sm 3 J H2 C4 Hz 3.6 m 1.1 sm 0.6 sm nm 0.7 sm 4.8 m lg 3 J H3 C1 Hz 2.7 m sm 1.4 sm 1.2 sm nm 1.3 sm 1.5 sm 2 J H2 C3 Hz 4.5 lg <0.4 sm sm nm 2.6 m 5.7 lg 2 J H3 C2 Hz 3.1 m lg 2.5 m sm sm nm 0.9 sm <0.2 sm Rotamer, x IIb, 0.90 0.10 IIb +IIa, 0.66 Ic + Ib 0.33 IIb, 0.66 Ib, 0.33 3 J H2 H3 Hz 2.6 sm 2.9 sm 4.1 sm m 3.4 sm 4 .0 sm 2.9 sm 3 J H2 C4 Hz 0.5 sm nm 0.3 sm 2.8 m sm sm sm 3 J H3 C1 Hz 0.9 sm nm 0.6 sm 2.6 m sm sm 5.7 lg 2 J H2 C3 Hz 2.4 sm m nm 2.2 m 3.4 m lg 1.2 1.2 2 J H3 C2 Hz 1.0 sm nm 1.2 sm 2.4 m sm 4.3 lg Note: Some values are missing for the minor diastereom ers where they were in low concentration; in these cases (products derived from aldehydes 3 and 14 diastereomer. The elucidation of the stereochemistry of the products derived from aldehyde 2 illustrates the me thodology. For the minor product, the coupling constants are all small, as in IIb Larger values, i.e m sm, for 3 J H2 H3 and 2 J H3 C2 indicate that IIa is also present. In the major diastereomer, a large 2 J H2 C3 would point towards Ia Ic IIa and / or I Ic A small 3 J H2 H3 is consistent with Ic and/ or IIc but not with Ia and IIa Finally, a small 3 J H3 C1 is consistent with Ic Medium values for 3 J H2 C4 3 J H3 C1 and 2 J H3 C2 demonstrate that the rotamer of I in which these coupling constants are of oppo site magnitude, i.e. some 1b is also present. Both the anti (major) and syn diastereomers prefer the conformations in which the OH and the NH 3 + groups are gauche which can be explained by hydrogen bonding or electrostatic interactions.

PAGE 122

122 For the aldol pro duct derived from aldehyde 3 the coupling constants indicate that the major species corresponds to rotamer IIb The same gauche conformational preference is observed for the hydroxyl and ammonium moieties but conformers in which the bulkier cyclohexyl g roup is gauche to the carboxyl are absent. The high diastereomeric ratio precludes measuring all of the coupling constants for the minor product, but the H2 H3 coupling points towards Ib or Ic The steric requirements of an aromatic moiety are less than o f a cyclohexyl, and although the preferred conformations are still IIb for the anti and Ic for the syn diastereomers some other rotamers are also present. The absolute stereochemistry of these 3 hydroxy 2 amino acids was determined by derivatization with methoxyphenylacetic acid (MPA). 241 D 2 O or methanol d4 which dissolve the 2 hydroxy 1 amino acid, are not appropriate solvents for the esterificat ion, which u ses DCC as a hydrization reagent. The methyl ester of the n butyl amino acid was soluble in chloroform d, while for the other 2 hydroxy 1 amino acids pyridine d5 was used for derivatization. All the MPA esters were prepared in the NMR tube, and characterized in the reaction mixture. 242 The absolute stereochemistry of both chiral carbons C2 and C3 was determined for the n butyl compound by double deri vatization, and was found to be 2 S, 3 S in the major and 2 S 3 R in the minor, in agreement with the relative stereochemistry found by J analysis. For the methyl ester of the cyclohexyl compound, double derivatization proved the major to be 2 S 3 R an d the minor 2 R 3 R again in agreement with the relative stereochemistry previously determined. The presence of the methoxy protons in position 1 of the ester allowed the elucidation of the stereochemistry of the phenyl compound by single derivatization to the amide. C2 was

PAGE 123

123 found S in the major and R in the minor; using the knowledge of relative stereochemistry, the major was assigned as 2 S 3 R and the minor as 2 R 3 R vi Monitoring Transaldimination of Amino Donors Since the initial screening of non glycine amino donors w as unsuccessful, the focus was directed to the first step in catalysis: transaldimination of the amino donors. An external aldimine has an absorbance of 422 nm, however an internal aldimine has an absorbance of 388 nm. 243 By utilizing UV Vis spectroscopy, the formation of the external aldimine could be monitored by an increase in the absorbance at 422 nm. A simple kinetics experiment was run over ten minutes on the L allo TA from A. jandaei taking absorbance readings every 30 seconds at 422 nm. The results for these amino donors are shown in F igure 3 6 Gly wa s administered as the positive control as it is the native substrate for this enzyme. We found that the negative control ( addition of PLP alone ) gave virtually no increased absorbance at 422 nm over the ten min. A mino donors 21 and 22 (F igure 3 6, purple and red) showed optimistic results for the formation of the external aldimine. It is possible that the enzyme can form the external aldimine, bu t not undergo deprotonation carbon Amino donors 24 and 25 (F igure 3 6, red and green) gave similar values to the PLP negative control, indicating no external aldimine formation. Amino donors 24 and 25 were probed further to see if the enzyme, L allo TA from A. jandaei al lowed for any binding to the active site by adding equal amounts of Gly and 24 (or Gly and 25 ) with aldehyde 9 As expected, only the aldol product from Gly was formed as seen in the initial screening results However the product vi For Tables with chemical shifts, see supporting information section, Tables A 2 4.

PAGE 124

124 concentration was redu ced compared to the analogous reaction lacking 24 or 25 This supports the perception that these amino donors 24 and 25 can bind to the active site, but once bound, cannot undergo transaldimination or aldol condensation with an aldehyde Deprotonation of Amino Donors The next step was to probe for deprotonation on carbon of the amino donor the next step in the cataly t i c cycle of TA By applying deuterium oxide (D 2 O) and mass spectrometry (MS) one can determine whether TA s can catalyze deprotonation on carbons a djacent to amines. The reaction was monitored by GC/MS to detect the possible increased molecular weight due to deuterium incorporation from the solvent As the natural substrate, Gly was used as a positive contro l to test this method and to monitor the deprotonation of the enzyme. Gly was found to have complete deuterium exchange within 1.5 hours. However, analysis of non glycine amino donors indicated no deuterium exchange of the proton after 24 or even 48 hours. Although this experimen t lead to the result that TA s do not deprotonate non glycine amino donors, it did give some insight into the catalysis by these enzymes T he positive control was carried out in both the absence and presence of an aldehyde and in the presence of aldehyde 13 a comp etitive inhibitor that binds to the active site but is unable to undergo aldol addition. The idea was to determine whether an aldehyde must be present for glycine deprotonation to occur. It was found that an aldehyde wa s not required for deprotonation of Gly. Secondly, these L TAs are known to only make the L isomer of these aldol products Taken together, the assignment configuration of the preparative scaled reactions confirmed this fact and that this experiment revealed that

PAGE 125

125 only one proton was excha nged with deuterium over time, c onfirmed the fact that these enzymes only make the L isomer. And lastly, we have developed an efficient method for monitoring the exchange of the proton with deuterium Investigation into the Thermodynamic Reversibility of Aldol Products One of th e major limitations of TAs is the ir reversibility that leads to loss of diastereoselectivity This loss of d.e. has been seen in the literature 184, 192, 196, 204 and in our results (Table 3 1 ). T he thermodynamic reversibility was monitored by proton nuclear magnetic resonance ( 1 H NMR) and MS experiments using deuterated acetaldehyde ( Figure 3 7 ). The reaction wa s monitored by 1 H NMR at several time points, probing the reversibility by observing the transformation of the from a doublet to a singlet. After 24 hrs, we saw approximately a 2 : 1 ( singlet : doublet ) ratio in the 1 H NMR spectrum, indicating some reversibility occur r ed overnight (Figure A 8 g ) It was also possible to monitor the reversibility b y MS and the addition of plus four in Thr The reaction was monitored at numerous time points and after 24 hrs the ratio between L allo Thr and L allo T hr d4 was approximately 1 : 1 (Figure A 9 f ). Diastereomeric excess was also monitored in this experiment, however no loss i n diastereoselectivity was found This was expected as these are the natural substrates for this enzyme. By utilizing both of these techniques, it was confirmed that thermodynamic reversibility does occur especially over a long period of time For the natural substrates, no loss in d.e. was found, however the screening results indicate with larger aldehyde acceptors, reversibility wa s damaging to the d .e. of these products

PAGE 126

126 Conclusions In summary, L TAs from A. jand aei E. coli and T. maritima were tested on an assortment of aldehyde acceptors using glycine as the amino donor A. jandaei L allo TA proved superior in all cases. The conversions using aldehyde acceptors could be scaled up moderately and the aldol pro ducts isolated from the reaction mixtures using glycine oxidase to degrade excess glycine. Establishing these downstream processing steps should increase the practical impact of TA catalyzed reactions in asymmetric synthesis. Additionally, these L TAs we re screened against a ha ndful of unnatural amino donors, nevertheless the results were inferior To investigate this further, the transaldimination of amino donors was monitored by UV Vis spectroscopy and revealed that both L Ala and L Ser undergo transal dimi nation with the active site PLP; however aminomethylphosphonate and aminomethanesulfonate do not form the external aldimine Sequentially, t he deprotonation was monitored by MS and deuterium oxide ; however non glycine amino donors exhibited no deu terium exchange. Experimental Procedures General LB medium contained 10 g/L Bacto Tryptone, 5 g/L Bacto Yeast Extract and 10 g/L NaCl; 15 g/L agar was added for plates. PCR amplifications were performed with Phusion Hot Start II DNA polymerase using the protocols. Electroporation was carried out with a BioRad GenePulser apparatus using 0.2 cm cuvettes. Promega Wizard kits and CsCl buoyant density ultracentrifugation 244 were used for small and large scale plasmid purification s, respectively. Fluorescent Sanger DNA sequencing was performed by the University of Florida ICBR. GC/MS eV. The temperature program involved an initial hold at 95 C fo r 5 min, an initial

PAGE 127

127 increase of 5 C /min to 138 C followed by an increase of 10 C /min to 180 C then a final increase of 2 C /min to 200 C and a hold at that temperature for 10 min. Cloning of A. jandaei L allo TA The gene encoding L allo TA from A. jandaei (accession number D87890) was synthesized by GenScript and ligated into a pUC 57 with flanking Nde I and Xho I restriction sites at the 5' and 3' ends, respectively. The TA gene was excised by digesting with these restriction enzymes and ligated w ith Nde I, Xho I cut pET 15b (Novagen). After transformation into E. coli ElectroTen Blue, plasmid DNA from a randomly chosen colony was isolated, restriction mapped and then sequenced to verify the desired structure. The resulting plasmid (designated pSF3 ) was used to transform E. coli BL21 Gold(DE3) strain for protein overexpression. Cloning of E. coli L TA. The gene encoding L TA (accession number NC000913) was amplified from E. coli ElectroTen Blue by colony PCR 229 using 5' ATAAGGACAT CATATG ATTGATTTAC 3' and 5' ACGTC T GGATCC TTAACGCG 3' as forward and reverse prim ers, respectively. These primers also introduced flanking Nde I and Bam HI restriction sites (underlined). After purification, the PCR product was digested sequentially with Nde I and Bam HI, then ligated with Nde I, Bam HI digested pET 15b. After transformat ion into E. coli ElectroTen Blue, plasmid DNA from a randomly chosen colony was isolated, restriction mapped and then sequenced to verify the desired structure. The resulting plasmid (designated pSF4) was used to transform E. coli BL21 Gold(DE3) strain fo r protein overexpression. Cloning of T. martima L allo TA. The L TA gene from T. maritima (accession number AE000512) was PCR amplified from T. maritima genomic DNA using 5' CGTGTGGGAGGTGAC CATATG ATCGATC TCAGGTCCGACACC 3' and 5'

PAGE 128

128 GAAATTTTTT GGA TCC TCAGGAGAA TTTTCTGAAGAGTTTTTCGAAGA T 3' as forward and reverse primers, respectively. The gene was inserted into pET 15b in the same manner as E. coli L TA, yielding pSF5, which was used to transform E. coli BL21 Gold(DE3) for protein overexpression. Cloning of B. subtilis g lycine o xidase. The complete coding sequence for B. subtilis glycine oxidase (accession number NC000964) was synthesized by GenScript and ligated into pUC57 with flanking Nde I and Bam HI restriction sites at the 5' and 3' ends, respectively Sile nt mutations were introduced to the coding region to remove internal Nde I and Bam HI sites that occur in the native sequence. The gene was subcloned as an Nde I, Bam HI fragment between these sites into pET 15b. After tran s forming E. coli ElectroTen Blue, pl asmid DNA was isolated from a randomly chosen colony, restriction mapped and sequenced to verify that the desired plasmid had been prepared (designated pSF9). This was used to transform E. coli BL21 Gold(DE3) for protein overexpression. Protein o verexpres sion. A single colony of the appropriate strain was used to overnight at 37 C a 40 mL portion of the preculture was added to 4 L of LB medium ampicillin, 80 mL of 20% glucose, and 1.5 mL of Sigma Antifoam 204 in a New Brunswick M19 fermenter. The culture was grown at 37 C with stirring at 400 rpm and an air flow of 1 vvm until the O.D. 600 reached 0.5 0.6. Protein overexpression was induced by adding 10 mL of 0.16 M IPTG (to yield a final concentration of 0.4 mM) and adjusting the temperature to 30 C and shaking. After 3 hr, the cells were har vested by centrifuging at 6,300 g for 15 min at 4 C resuspended

PAGE 129

129 in 50 mM KP i pH 8.0 (1 mL buf fer per gram wcw), then lysed by a French pressure cell at 17,000 psi. Insoluble debris was pelleted by centrifuging at 39,000 g for 1 hr at 4 C and the yellow supernatant was used for TA catalyzed reactions. Glycerol was added to a final concentr ation of 20% and the protein was stored in aliquots at 80 C Affinity p urification of B. subtilis g lycine o xidase A crude extract containing glycine oxidase was prepared as described above, then the sample was applied to a 5 mL HiTrap Chelating HP colu mn (GE Healthcare Life Sciences) that had been equilibrated with binding buffer (20 M NaP i 500 mM NaCl, 20 mM imidazole, pH 7.4). After washing with 50 mL of binding buffer, the desired protein was eluted by elution buffer (20 mM NaPi, 500 mM NaCl, 500 m M imidazole, pH 7.4). A flow rate of 2 mL/min was employed throughout. The eluate was concentrated by ultrafiltration (Amicon Ultra), then diluted with 50 mM KPi, pH 8.0 and re concentrated. This was repeated two more times. The final glycine oxidase s ample was diluted with the same buffer to 5 mg/mL, then glycerol was added to a final concentration of 10% and the protein was stored in aliquots at 80 C Enzyme a ssays for t hreonine a ldolase. The activity of L TAs was measured by mixing 0.1 m solution in 50 mM KPi, pH 8.0 (total volume of 1 mL). The mixture was gently rotated at overnight. Samples were derivatized with MSTFA and analyzed by GC/MS. The temperature program involved an initial hold at 95 C for 5 min, an initial increase of 5 C /min to 1 20 C followed by an increase of 2 C /min to 1 38 C then a final increase of 10 C /min to 20 0 C and a hold at that temperature for 5 min

PAGE 130

130 Enzyme assay for g lycine o xidase. Glycine oxidase activity was measured by monitoring the formation of H 2 O 2 by UV Vis spectroscopy at 500 nm using a coupled HRP assay 228 Reaction mixtures containing 10 mM glycine and 0.25 mg glycine aminoantipyrine, 2 incubating at 37 C for 1 0 min, the A 500 value was used to calculate units glycine oxidase activity. 228 Enzyme k inetics. Kinetic experiments were executed at 422 nm for 10 minutes at room temperature monitoring the absor bance every 30 seconds The reaction contained 10 mol of amino donor, 0.1 mol of PLP, and 10 L of enzyme lysate in b uffer (total volume of 1 mL). After the addition of enzyme, t he cuvette was inverted three times and the absorbance was immediately mon itored for 10 minutes. The experiment was performed in triplicate and an average was taken for the final results. Derivatiz ation of a mino a cids by FAA d erivatization. Amino acid standards were dissolved in 100 L of buffer pH 8.0, following an addition of 0.1 M ammonium acetate pH 7.0 (200 L) and 2% trifluoroacetylacetone (FAA) (200 L). After heating for 95 C for 20 min, the solvent system acetonitrile : water : MeOH : pyridine (21 : 21 : 4 : 4) (200 L) was added to the reaction mixture. The deriva tized amino acid was extracted by chloroform (500 L) and analyzed by GC/MS. The temperature program involved an initial hold at 95 C for 2 min, a temperature increase of 3 C /min to 200 C and a hold at that temperature for 5 min. Derivatization of ami no acids by OPA/NAC d erivatization. Amino acid standards (10 L) were diluted with 0.2 M NTBB buffer pH 10.5 (500 L). The

PAGE 131

131 OPA/NAC derivatizing agent (200 M OPA/600 L NAC in MeOH) (10 L) was added to the amino acid solution and inverted four times to mix. After incubating for 15 min at room temperature, the derivatized amino aci ds were analyzed by HPLC. A C 18 column was used as the stationary phase. Potassium phosphate (50 mM) and acetonitrile were used as the mobile phase systems A and B, respectiv ely. The solvents were filtered through a 0.22 membrane filter. The flow rate was 1 mL/min and the eluents were monitored at 250 nm. The injection volumes were 20 L. The gradient was as follows: The i nitial mobile phase was solvent A (100%) for 1 m L/min for 10 min, the mobile phase was switched to solvent B (100%) for 3 min, and the final mobile phase was reverted back to solvent A (100%) for 15 min. Derivatization of amino acids by MSTFA d erivatization. The original procedure required reaction mix tures (or standards of 5 nmol of amino acid) to be dried completely under reduced pressure for 30 min by a Savant SpeedVac SVC100, then the residue was taken up by 80% MeOH (50 L). The sample was then dried for a second time under reduced pressure for 30 min, following the addition of methylene chloride (40 L). The sample was dried for a third and final time under reduced pressure for 10 min. The dried sample was taken up in the derivatization solvents: dried shaking at 37 C for 30 min, the mixtures were analyzed by GC/MS. Derivatization of amino acids by MSTFA d erivatization ( o ptimized). The optimized procedure aimed to limit the amount of drying steps and derivatization reagents. R eaction mixtures (or stan dards of 5 nmol of amino acid) were dried completely under reduced pressure for 30 min by a Savant SpeedVac SVC100 The

PAGE 132

132 dried sample was taken up dry 37 C for 30 min, the mixtures were analyzed by GC/MS. For L Thr and L allo T hr standards and enzyme assay, t he temperature program involved an initial hold at 95 C for 5 min, an initial increase of 5 C/min to 1 20 C followed by an increase of 2 C /min to 1 38 C, then a final increase of 10 C /min to 200 C and a hold at that temperature for 5 min For screening reactions, t he temperature program involved an initial hold at 95 C for 5 min, an init ial increase of 5 C /min to 138 C followed by an increase of 10 C /min to 180 C then a final increase of 2 C /min to 200 C and a hold at that temperature for 10 min. Detection of a mino a cids by TLC with a n inhydrin s tain. TLC silica gel plates (Merck 60 F254) were used to visualize Gly and amino acid products at 254 nm and by treatment with a 2% ninhydrin reagent (in methanol). The TLC solvent system contained Bu OH : H 2 O : AcOH (4 : 2 : 1). After separation by TLC, the silica plate was sprayed with 2% ninhydrin solution to visualize primary amines. Deprotonation of a mino d onors without aldehyde 13 R eactions con tained 7.5 mol of amino donor 750 nmol of PLP and 2.5 volume of 1 mL) made with D 2 O These were gently rotated overnight at room temp erature and samp led after 0.5 hr, 1 hr, 1.5 hr, 3 hr, 5 hr and overnight for MSTFA derivatization and GC/MS analysis. Deprotonation of a mino d onors with aldehyde 13 Reactions contained 3.75 mol of ortho hydroxybenzaldehyde ( 13 ) 7.5 mol of amino donor, 750 nmol of PLP and 2.5 made with D 2 O

PAGE 133

133 These were gently rotated overnight at room temp erature and sampled after 0.5 hr, 1 hr, 1.5 hr, 3 hr, 5 hr and overnight for MSTFA derivatization and GC/MS analysis. Pro bing t hermodynami c r eversibility of a ldol r eaction s by 1 H NMR Six reactions were performed to complete this experiment. Positive control reactions contained (1) 0.1 mmol of Gly, 0.1 mmol of acetaldehyde 10 mol of PLP and 10 L of enzyme lysate in buff er (total volume of 0.5 mL) made with D 2 O, (2) 0.1 mmol of Gly, 0.1 mmol of acetaldehyde d4 10 mol of PLP and 10 L of enzyme lysate in buffer (total volume of 0.5 mL) made with D 2 O, and (3) 50 mol of L allo Thr 0.5 mol of PLP and 10 L of enzyme lysa te in buffer (total volume of 0.5 mL) made with D 2 O Negative control reactions contained (4) 50 mol of L allo Thr and 0.1 mmol of acetaldehyde d4 in buffer (total volume of 0.5 mL) made with D 2 O and (5) 50 mol of L allo Thr, 0.1 mmol of acetaldehyde 0 .5 mol of PLP and 10 L of enzyme lysate in buffer (total volume of 0.5 mL) made with D 2 O. And lastly, the reversibility reaction contained (6) 50 mol of L allo Thr, 0.1 mmol of acetaldehyde d4 0.5 mol of PLP and 10 L of enzyme lysate in buffer (tota l volume of 0.5 mL) made with D 2 O. These were allowed to react overnight at room temp erature and sampled after 0 hr, 1 hr, 3 hr, 6 hr, and overnight for 1 H NMR analysis Probing t hermodynami c reversibility of aldol reaction s by GC/MS Six reactions wer e performed to complete this experiment. Positive control reactions contained (1) 0.2 mmol of Gly, 0.1 mmol of acetaldehyde, 20 mol of PLP and 10 L of enzyme lysate in buffer (total volume of 1 mL) made with D 2 O, (2) 0.2 mmol of Gly, 0.1 mmol of acetald ehyde d4, 20 mol of PLP and 10 L of enzyme lysate in buffer (total volume of 1 mL) made with D 2 O, and (3) 0.1 mmol of L allo Thr, 10 mol of PLP and 10

PAGE 134

134 L of enzyme lysate in buffer (total volume of 1 mL) made with D 2 O. Negative control reactions contai ned (4) 0.1 mmol of L allo Thr and 0.1 mmol of acetaldehyde d4 in buffer (total volume of 1 mL) made with D 2 O and (5) 0.1 mmol of L allo Thr, 0.1 mmol of acetaldehyde, 10 mol of PLP and 10 L of enzyme lysate in buffer (total volume of 1 mL) made with D 2 O And lastly, the reversibility reaction contained (6) 0.1 mmol of L allo Thr, 0.1 mmol of acetaldehyde d4, 10 mol of PLP and 10 L of enzyme lysate in buffer (total volume of 1 mL) made with D 2 O. These were gently rotated overnight at room temp erature and sampled after 0.5 hr, 1 hr, 1.5 hr, 3 hr, 5 hr and overnight for MSTFA derivatization and GC/MS analysis. General p rocedure for s creening a ldehyde a cceptors. Reactions contained 0.1 mmol of aldehyde, 0.5 mmol of glycine, 10 nmol of in buffer (total volume of 1 mL). These were gently rotated overnight at room temp erature and sampled after 4 hr and overnight for MSTFA derivatization and GC/MS analysis. General p rocedure for s creening a mino d onors Reac tions contained 0.1 mmol of aldehyde (acetaldehyde, hexanal, or benzaldehyde) 0.5 mmol of amino donor were gently rotated overnight at room tempe rature and sampled after 4 hr and overnight for MSTFA derivatization and GC/MS analysis. Design of Experiments To optimize reaction conditions, all three TAs underwent a series of experiments varying temperature, pH, and glycine concentration. All reactions contained 0.1 mmol o f aldehyde, 0.2 1.0 mmol of glycine, 10 nmol of PLP

PAGE 135

135 The four different glycine concentrations were tested at five different temperatures (4 C 18 C 25 C 37 C and 42 C ) and four different pH values (6, 8, 9.7, and 12). taken after 4 hr and overnight. Samples were derivatized with MSTFA and analyzed by GC/MS. Optimization o f Purification Procedure F irst purification a ttempt : After the TA catalyzed reaction was completed the unreacted aldehyde was removed by extraction with Et 2 O (2 30 mL) and the aqueous layer was lyophilized. The solid residue was mixed thoroughly wit h 30 mL of MeOH and the solution was filtered and evaporated under reduced pressure. 1 H NMR analysis revealed the presence of a large amount of glycine. S econd a ttempt : After the TA catalyzed reaction was completed the unreacted aldehyde was removed by extraction with Et 2 O (2 30 mL) and the aqueous layer was lyophilized. The solid residue was mixed thoroughly with 30 mL of MeOH and the solution was filtered and evaporated under reduced pressure. This procedure was repeated to remove additional unrea cted glycine. 1 H NMR analysis revealed trace amounts of glycine. Third a ttempt: After the TA catalyzed reaction was completed the unreacted aldehyde was removed by extraction with Et 2 O (2 30 mL) and the aqueous layer was lyophilized. The solid residu e was mixed thoroughly with 30 mL of MeOH and the solution was filtered and evaporated under reduced pressure. This procedure was repeated to remove additional unreacted glycine T he crude product was dissolved in water and applied to an 11 1.5 cm DOWE X, 1 2 (HO form) column. The column

PAGE 136

136 was washed with 100 mL of deionized water, then the desired product was eluted by washing with 50 mL of 20 % acetic acid. The solvent was removed using a SpeedVac to afford the final product. 1 H NMR analysis reveale d no separation of glycine was achieved. Separately, different concentrations (10%, 5%, 2%, and 1%) of eluent solvent (acetic acid) were used in an attempt for a greater separation of glycine from the products, but only slight separation occurred with 1% AcOH. F ourth a ttempt: After the TA catalyzed reaction was completed the unreacted aldehyde was removed by extraction with Et 2 O (2 30 mL) and the aqueous layer was lyophilized. The solid residue was mixed thoroughly with 30 mL of MeOH and the solution was filtered and evaporated under reduced pressure. This procedure was repeated to remove additional unreacted glycine T he crude product was dissolved in 50% EtOH and 5% AcOH in water and applied to a 15 2 cm silica gel column. The column was washed with 50% EtOH and 5% AcOH in water The solvent was found by attempting several solvent systems by TLC and ninhydrin stain, including such solvents systems as 100% EtOH, 80% EtOH, and 50% EtOH. The solvent that allowed for best separation was 50% EtOH an d 5% AcOH in water ( rf values of 0.3 and 0.7 for glycine and products, respectively). The fractions were monitored by TLC and ninhydrin stain. The solvent was removed using a SpeedVac to afford the final product. 1 H NMR analysis revealed a slight separa tion of products from glycine. Separately, a 20 x 2 cm silica gel column was used in the attempt to allow for great separation, but this was also unsuccessful. The f ifth and f inal a ttempt is summarized in the following section.

PAGE 137

137 Preparative s h ydroxy a mino a cids using A. jandaei L allo TA Reaction mixtures contained 2.0 mmol of aldehyde, 10 mmol glycine, 100 nmol After gently rotating at 37 C f or 4 hr, the unreacted aldehyde was removed by extraction with Et 2 O (2 30 mL) and the aqueous layer was lyophilized. The solid residue was mixed thoroughly with 30 mL of MeOH and the solution was filtered and evaporated under reduced pressure. This pro cedure was repeated to remove additional unreacted glycine. The crude product was resuspended in 10 mL of 50 mM KP i pH 8.0, then 1 mg of purified glycine oxidase was added and the mixture was gently rotated at 37 C for 8 hr. An additional of 1 mg porti on of purified glycine oxidase was then added and incubation at 37 C was continued for an additional 20 hr. The reaction mixture was lyophilized. The residue was stirred with MeOH, leaving most phosphate undissolved. After evaporating the solvent, the crude product was dissolved in water and applied to an 11 1.5 cm DOWEX, 1 2 (HO form) column. The column was washed with 100 mL of deionized water, then the desired product was eluted by washing with 50 mL of 0.5% acetic acid. The solvent was remove d using a SpeedVac to afford the final product. (2 S ,3 S ) 2 amino 3 hydroxyheptanoic acid. White solid 53 mg, 16% yield, 22% d.e., m.p. 226 C (lit. m.p. 223 224 C) 245 1 H NMR (300 MHz, D 2 3.65 3.84 (1H, dd), 1.3 1.45 (2H, m), 0.85 (3H, m) ppm. 13 C NMR (300 MHz, D 2 171.47, 69.39, 59.34, 32.76 30.5, 27.39, 21.54, 13.07 ppm (Figure A 2 ). (2 S ,3 R ) 2 amino 3 hy droxycyclohexanepropanoic acid. White solid 185 mg, 50% yield, 99% d.e., m.p. 210 C (lit. m.p. 216 217 C) 245 1 H NMR (300 MHz, D 2 O)

PAGE 138

138 3.89 4.00 (1H, dd), 3.58 3.61 (1H, dd), 1.00 1.93 (11H, m) ppm. 13 C NMR (300 MHz, D 2 (Figure A 3 ) (2 S ,3 S ) 3 hydroxyphenylalanine. White solid 115 mg, 32% yield, 86% d.e., m.p. 188 190 C (lit. m.p. 189 191 C) 246 1 H NMR (300 MHz, D 2 5.29 5.27 (1H anti dd), 3.90 3.89 (1H, dd) ppm. 13 C NMR (300 MHz, D 2 O) 171.79, 139.00, 128.89, 128.75, 128.52, 126.25, 125.82, 71.18, 60.70 ppm (Figure A 4 ) 2 amino 3 hy droxy 4 pyridinepropanoic acid. Orange solid 101 mg, 28% y ield, 40% d.e., m.p. 180 C. 1 H NMR (300 MHz, D 2 5.70 5.69 (0.70H syn d), 5.58 5.57 (0.30H anti d), 3.97 4.17 (1H, dd) ppm. 13 C NMR (300 MHz, D 2 (F igure A 5 ) 3 hydroxy 2 methoxy phenylalanine White solid 91 mg, 22% yield, 89% d.e, m.p. 170 C. 1 H NMR (300 MHz, D 2 7.38 (2H, m), 7.04 7.01 (2H, m), 5.57 5.56 (1H syn dd), 4.41 4.40 (1H, dd), 3.81 (3H, s) ppm (Figure A 6 ) 2 amino 3 (2 chloro 3 pyr idine) 3 hydroxypropanoic acid. Orange solid 215 mg, 50% yield, 56% d.e, m.p. 179 180 C. 1 H NMR (300 MHz, D 2 7.93 (2H, m), 7.44 7.36 (1H, m), 5.53 5.52 (0.65H syn d), 5.35 5.34 (0.30H anti d), 4.09 4.07 (1H, dd) ppm (Figure A 7 )

PAGE 139

139 Figure 3 1. T hreonine aldolase reactions Figure 3 2. Glycine oxidase reaction

PAGE 140

140 Figure 3 3. Derivatization of amino acids Figure 3 4. Screening of amino donors Figure 3 5 Preparative reactions

PAGE 141

141 Figure 3 6. Transaldimination of amino donors Figure 3 7. Thermodynamic reversibility of L threonine 0.7 0.75 0.8 0.85 0.9 0.95 1 0 100 200 300 400 500 600 700 Absorbance Time (s) Transaldimination of Amino Donors Glycine Aminomethylphosphonate Aminomethylsulfonate L-Alanine L-Serine (-) Control

PAGE 142

142 CHAPTER 4 STRUCTURE DETERMINATION AND SUBSTRATE PROFILING OF P. putida L THREONINE ALDOLASE Introduction L Threonine aldolases (L TAs) catalyze the retro aldol cleavage of threonine to glycine and acetaldehyde. Although named for the retro aldol reaction, they also catalyze the formation of threonine by carbon position of the donor amino acid and the carbonyl carbon of the acceptor aldehyde. Many L TAs have been characterized and screened against an assortment of aldehydes, ranging fr om long chain aliphatic aldehydes to an array of aromatic aldehydes with altered electronic effects 184, 222, 223 TAs var y in their stereo selectivit ies positions of the aldol products and have been named L threonine aldolase, L allo threonine aldolase, low specificity L threonine aldolase/L allo threonin e aldolase and D threonine/D allo threonine aldolases The ability of these enzymes to control two adjacent chiral centers along with C C bond forma tion gives them potential for biotechnology. In recent years, these enzymes have been applied to the synthes is of important pharmaceuticals including L threo 3,4 disease therapy 238 and (2 R ,3 S ) 2 amino 3 hydroxy 3 (pyridin 4 yl) propanoic acid, a precursor to the drug development candidate (2 R ,3 S ) 2 amino 3 hydroxy 3 (pyridin 4 yl ) 1 (pyrrolidin 1 yl)propan 1 one. 200 Currently, three L TAs, from T hermogota maritima 205 A eromonas jandaei 207 and E coli 206 have known crystal structures. This collection of data has given us a detailed The two most importan t and conserved amino acid residues wi thin the active site are Lys 199 and His 85. The Lys forms the Schiff base with PLP and any mutations at this site render the enzyme inactive. 189

PAGE 143

143 The His is r esponsible for regulating of the degree of stereo selectivity between the L and L allo stereomers and may also be t he catalytic base for the retro aldol cleavage of threonine 205 207 P seudomonas putida L TA has be en studied for its broad s ubstrate range in producing hydroxy amino acids 184, 191, 192, 247 Although other L TAs have shown a high tolerance in the synthesis of phenylserine derivatives and long chain amino acids 184 the diastereoselectivity of P. putida L TA was found to be rather poor with the remarkable exception of 4 fluorothreonine (93% d e ). 191 On the other hand in most cases, conversions were generally high for this family of enzymes Additionally, temperature, pH, and co solvent addition studies were executed by Steinreiber et al. who found that the most favorable conditions were pH 8.0 and room temperature. Addition of co solvents did not increase reacti vity of the enzyme 192 Here, we report the crystal structure determination and substrate profiling results for L TA from P. putida (LTAPP). The enzyme was screened against a variety of aldehyde acceptors. In compa rison to other enzymes in the same class, diastereom er ic purify was found to be quite low (Table 3 1). We hoped to use t he crystal structure to understand the poor stereoselectivity found during screening studies Results and Discussion Gene C loning and P rotein O verexpression The gene encoding the L TA was obtained by colony PCR 229 from a P. putida strain purchased from Carolina Biological Company. After amplification, the L TA gene was ligated into pET 15b and the resulting plasmid was used to transform the E. coli overexpression strain BL21 Gold(DE3). The L TA wa s efficiently overproduced by the recombinant strain, which allowed crude lysates to be employed for screening reactions.

PAGE 144

144 Control reactions were carried out for each aldehyde substrate with a crude extract from an un transformed E. coli strain; none gave significant aldol product. Purification of P. putida L Threonine Aldolase and Screening of Protein Crystals After isolating the crude lysate, the L TA was applied to a HiTrap Chelating HP column to separate the hexahistidine tagged L TA by affinity chromat ography. The eluate was concentrated by ultrafiltration and t he purified protein was applied to a for further purification. The N terminal hexahistidine tag was cleaved by using thrombin from bovine plasma. The histidine tag was sep arated from the desired protein by applying it to the same HiTrap Chelating HP column. The desired protein was concentrated by ultrafiltration and reconstituted in 50 mM KP i pH 8.0 for crystal screening Hampton Research 96 well screening kits were use d for the initial screening of P. putida L TA. All initial screenings used the sitting drop vapour diffusion method at room temperature with 10 and 15 mg/mL protein concentrations. After 72 hours, c rystals were identified in the C 6 well of the Salt Rx H T kit. The precipitant solution contained 3.5 M sodium formate, 0.1 M sodium acetate trihydrate pH 4.6. Crystallization scale ups using both hanging drop and sitting drop vapour diffusion methods and careful observation revealed that crystals formed in l ess than 24 hours. Optimized crystal conditions were sought in order to slow crystallization. A 24 well large sc reening plate was set up with varied concentrations of both sodium formate (0.7 6.3 M) and sodium acetate (0.1 1 M) buffers This screen ing revealed that crystals could not form in sodium formate and sodium acetate concentrations greater than 4.9 and 0.3 M, respectively Some conditions grew crystals in 24 hours, however a

PAGE 145

145 few wells presented crystal form ation in 120 hour s. T w o of these crystals were determined at 2.80 and 2.82 resolutions, respectively. The previous screening u sed 15 mg/mL of purified L TA; the next attempt involved decreasing the protein concentration added to each drop (5 10 mg/mL) with varied concentrations of bot h sodium formate (2.1 4.9 M) and sodium acetate (0.1 0.2 M) buffers. As before, m ost of the conditions resulted in crystals after 24 hours ; however careful observation revealed that after 240 hours few rod shaped crystals had formed. The final cryst allization condition was a hanging drop vapour diffusion method with a precipitant solution of 2.1 M sodium formate, 0.1 M sodium acetate trihydrate pH 4.6 at room temperature with 5 mg/mL of L TA (1 : 1, precipitant solution : protein). The crystals up t o this poi n t had been a tetragonal bipyramid shape, therefore the new rod shape d crystal could have indicated a different packing of the crystal lattice that might have signified a better resolution. Two of these crystals were used to collect a full data set: one un soaked and one soaked with the natural substrate L Thr. These two L TA structures were determined at 2.27 and 2.75 resolution, respectively. Overall Structure of the L Threonine Aldolase from P. putida The crystal structure of L TA from P. p utida (LTAPP) was determined at 2.27 resolution and was a homotetramer. This is consistent with other crystallographic studies of L TAs from T. maritima 205 E. coli 206 and A. jandaei 207 The struc ture was first determined by molecular replacement with the E. coli L TA ( PDB code 4LNJ). The data collection and refinement stat istics are summarized in T able 4 1.

PAGE 146

146 T able 4 1. Data collection and refinement statistics LTAPP a LTAPP a Resolution range 26.71 2.275 (2.356 2.275) CC(work) 0.957 (0.855) Space group C 1 2 1 CC(free) 0.960 (0.663) Unit cell 198.167 186.96 53.269 90 98.924 90 Number of non hydrogen atoms 11258 Total reflections 585838 (53082) M acromolecules 10667 Unique reflections 87029 (8273) L igands 84 Multiplicity 6.7 (6.4) S olvent 507 Completeness (%) 99.42 (95.01) Protein residues 1375 Mean I/sigma(I) 14.50 (2.65) RMS(bonds) 0 .014 Wilson B factor 41.45 RMS(angles) 1.30 R merge 0.08797 (0.6277) Ramachandran favored (%) 94.13 R meas 0.09536 (0.682) Ramachandran allowed (%) 5.06 R pim 0.03654 (0.2635) Ramachandran outliers (%) 0.81 CC1/2 0.998 (0.879) Rotamer outli ers (%) 1.91 CC* 0.999 (0.967) Clashscore 10.79 Reflections used in refinement 87014 (8270) Average B factor 54.48 Reflections used for R free 1999 (190) macromolecules 54.67 R work 0.1814 (0.2745) ligands 53.89 R free 0.2215 (0.3123) solvent 50 .52 a Statistics for the highest resolution shell are shown in parentheses.

PAGE 147

147 In the structure of LTAPP (Figure 4 1a), two of the four chains (A and D or B and helices and sheets, shown in F igure 4 1b. The overall structure of a single monomer equates well to the structure of E. coli L TA (Figure 4 sheets. The two differences are hig hlighte d in the zoomed view in F igure 4 1d,e. The purple loop (Figure 4 sheet in E. coli L TA 206 sheet in T. maritima L allo TA 205 and a loop in A. jandaei L allo TA 207 ( Figure 4 1f,g ). helices on the surface of the protein is approx imately 13 amino acid residues long By comparing it s position and length with other L TAs shown in green in F igure 4 1e g we can determine that it is very flexible as it is located on the surface of the protein and contains several more amino acid residues (4 7) than other L TAs. Since LTAPP has a slightly longer loop and is perceived as flexible, it could have distorted the folding of th helix, displaying the multiple conformations shown by the electron density at this position (not shown). The Active Site of Threonine Aldolase The binding pocket of the TA family is composed of positively charged residues in order to stabiliz e the negatively charged PLP and substrate/product. The electrostatic sur face potential is displayed in F igure 4 2a to show the binding pocket of one monomer. The active site of this family is comprised of five conserved residues: Ser 10, His 89, Arg 177 Lys 207, and Arg 321 (Figure 4 2b). The two most important and conserved amino acid residues within the active site are Lys 207 and His 89. Lys207 is essential for catalytic activity since it forms a Schiff base with the PLP cofactor. M utations at t his position leave the enzyme inactive 189 His 89 is responsible

PAGE 148

148 for regulati ng of the degree of dia stereospecificity between the L and L allo stereo iso mers 205 207 A more detailed illustration of the active site pocket is represented in F igure 4 3a. The unbound PLP has hydrogen bonding interactions with Gly 64, Thr 65, and Arg 177 to stabilize the negatively charge d phosphate and hydroxide groups along wit h Asp 174 to help stabilize the pyridinium ring of PLP. In addition to the hydrogen bonding interactions, His 89 donates pi stacking interactions with the pyridine ring to help position the cofactor in the active site. Although we were unsuccessful at soa king our crystals with the native substrate, there is a clear binding pocket for the subst rate shown as a gray circle in F igure 4 3a. As described in previous crystallographic studies, 205 207 the carboxylic group of the PLP Gly interacts with the side chains of Ser 10, Arg 177, and Arg 321. When the aldol condensation occurs, His 89 and His 133 hydrogen bond to the hydroxyl group of threonine. These interactions are summarized in F igure 4 3b d. Structural Compari son of L Threonine Aldolases As mentioned previously, crystal structures of L TAs from T. maritima 205 E. coli 206 and A. jandaei 207 have been solved. Sequence alignment (Figure 4 4) of the three enzymes with LTAPP resulted in 43 completely conserved residues (shown in red). Among these conserved residues 16 are located within the active site, 9 of which are highli ghted in F igure 4 3a (gray). The remaining 7 conserved residues include Asp 11, Tyr 35, Asn 68, Glu 94, Gly 175, A la 176, and Asn 290 (F igure 4 5 ) Only Tyr 35 is sufficiently close to interact directly with the substrate/product (Figure 4 5a). All other amino acid residues hydrogen bond to one of the conserved active site residues. The Asp 11 hydrogen

PAGE 149

149 bonds to N Arg 177, Asn 290 with O Asp 11, Asn 68 with O Asp 174, and Glu 94 with N His 89 ( Figure 4 5b c). The overall secondary structures of the fo ur L TAs were very comparable at all angles (Figure 4 6a). The only major differences was in surface loops and the loops in F igure 4 1d e. In order to see key differences in the active site residues, all four structures were aligned and active site resi d ues highlighted. As shown in F igure 4 6b, the cofactor and conserved residues occupy the same space and interact with the same amino acid residues. The only residue within the active site that differs in LTAPP from the other L TAs crystallized thus far w as the 93 position In LTAPP this position was occupied by Asp; however Tyr was found at the analogous position in T. maritima L allo TA and E. coli L TA and Phe occurred at this location in A. jandaei L allo TA. 2 05 207 This particular position is help ful to define the hydrophobic pocket where the methyl group of L Thr binds for the native reaction Site saturat ion mutagenesis at this position might determine whether this site plays an important role in the poor diastereoselectivity of this enzyme and this will be in the subject of the following chapter. Optimization of Reaction Conditions Prior to carrying out extensive aldehyde screening studies, we used the same DOE methodology employed in the previous chapter to identify optimal conditions for the overexpressed L TA from P. putida A total of 92 individual reactions with varied temperature (4 C, 18 C, 25 C, 37 C and 42 C) glycine : aldehyde ratio (2, 4, 8, and 10) and pH (5, 7, 8, 9.7 and 12) were carri ed out for the enzyme substrate pair and both relative conversion and product diastereomeric excess values were determined for each. Based on previous literature, the starting point was 25 C, pH 8 and a 5 fold

PAGE 150

150 molar excess of glycine versus the aldehyd e. The starting conditions gave a relative conversion i of 0. 51 using a 4 : 1 ratio of glycine : aldehyde. Aldehyde 10 (Table 4 2 ) was chosen to optimize the enzyme as it gave measurable, but incomplete conversions under standard conditions; this allowed both improved and detrimental changes to the reaction conditions to be identified. The only condition that affected the relative conversio n with aldehyde 10 was the glycine to aldehyde ratio Conversion increased s light ly as the substrate molar ratio w a s increased; s olving for the optimal conditions (pH 8 25 C and a glycine : aldehyde ratio of 10) and presenting a relative conversion of 0. 89 less than a 2 fold improvement. Although the outcome of this DOE strategy was more modest than others, it did provide that the increased ratio of glycine versus aldehyde increased the overall conversion. Table 4 2 Substrate specificity of L TA catalyzed aldol reactions Entry Aldehyde Product Reaction Time (hr) Conversion a (%) d.e. a (%) 1 1 4 20 59 78 99 99 2 2 4 20 69 72 26 13 3 3 4 20 35 48 37 25 4 4 4 20 35 50 n.d. c n.d. c i Relative conversions for these studies were defined as the peak area ratio of MSTFA derivatized pr oduct/internal standard using GC/MS.

PAGE 151

151 Table 4 2. Continued Entry Aldehyde Product Reaction Time (hr) Conversion a (%) d.e. a (%) 5 5 -4 20 ----6 6 4 20 24 36 33 28 7 7 4 20 8 9 20 12 8 8 -4 20 ----9 9 4 20 22 24 16 14 10 10 4 20 5 47 42 13 11 11 4 20 60 58 10 13 12 12 4 20 6 32 48 45 13 13 -4 20 ----

PAGE 152

152 Table 4 2. Continued Entry Aldehyde Product Reaction Time (hr) Conversion a (%) d.e. a (%) 14 14 4 20 24 64 n.d. c n.d. c 15 15 4 20 <5 b 11 -b n.d. c 16 16 4 20 54 75 n.d. c n.d. c 17 17 4 20 14 30 n.d. c n.d. c 18 18 4 20 <5 b <5 b -b -b Note: Reaction mixtures contained 500 mM glycine, 100 mM aldehyde, 0.05 mM PLP, 20% ethanol, and i pH 8. Reactions were incubated at 25 C. a Conversion and diastereomeric excess values were determined by GC / MS after MSTFA deri vatization. Reactions with 1 e omeric purity are provided since the small peak sizes preclude accurate integration. b Accurate values for conversion and diastereomeric excess could not be determ ined due to the very low peak areas. c It was not possible to determine the diastereoselectivity of th is reaction by chiral phase GC since temperatures required for elution were greater than the maximum column temperature. Screening of Aldehyde Acceptors Once optimized reaction conditions had been identified, the substrate range of the P. putida L TA w as investigated using numerous aldehyde acceptors (Table 4 2 ) The aldehydes were selected so that relative conversions and diastereoselectivi t i es could be compared with the other L TAs from the previous chapter (Table 3 1). Overall, the diastereoselectivity of this enzyme was poor, which agrees with literature reports 191

PAGE 153

153 The only aldehyde that displayed diastereoselectivi t i es greater than 5 0% was 1 The aldol condensation of this aldehyde with glycine gave a two fold increase in relative conversion compared to the L TAs from A. jandaei E. coli and T. maritima while retaining the 99% d.e Although conversions of other aldehyde acceptors w ith this L TA were improved from the other L TAs, the diastereoselectivity was always inadequate Conclusion The structure of P. putida L TA was successfully crystallized and data collected at a resolution of 2. 27 . Molecular replacement and refinement s trategies were used to build the final model of the structure and it was determined to be a homotetramer, comparable to other L TAs in its class. The active site lysine was determined to be Lys 207 and other highly conserved amino acid residues included: Ser 10, His 89, His 133, Arg 177, and Arg 321 C omparison with the other structures revealed two main differences. First, the loop located near the active site was slightly longer than the others (4 7 amino acid residues) which distort s the folding of helix Second the amino acid residue at position 93 differed In LTAPP, it was an Asp residue ; however in other L TAs, the residue at the analogous position was a large aromatic residue ( Tyr or Phe ) We also extensively screened the e nzyme w ith a variety of aldehyde acceptors. This study revealed that the enzyme has a broad substrate tolerance although diastereoselecivity was poor in most cases. This deficiency could be due to the extended loop or to the acidic residue at position 9 3 or to other, more subtle factors Mutagenesis studies of this position are described in the following chapter.

PAGE 154

154 Experimental Procedures General. LB medium contained 10 g/L Bacto Tryptone, 5 g/L Bacto Yeast Extract and 10 g/L NaCl; 15 g/L agar was added for plates. PCR amplifications were protocols. Electroporation was carried out with a BioRad GenePulser apparatus using 0.2 cm cuvettes. Promega Wizard kits and CsCl buoyant de nsity ultracentrifugation were used for small and large scale plasmid purifications, respectively. Fluorescent Sanger DNA sequencing was performed by the University of Florida ICBR. Crystallography kits for 96 well screening were purchased from Hampton Research. EI at 70 eV. The temperature program involved an initial hold at 95 C for 5 min, an initial rate of 5 C/min to 138 C, followed by a rate of 10 C/min to 180 C, and a final rate of 2 C/min to 200 C, held at that temperature for 10 min. Plasmid c onstruction. The L TA gene from P. putida (accession number AP013070) was isolated and amplified from a P. putida strain purchased f rom Carolina Biological Company by co lony PCR 229 CGTTCACAGGACCGT CATATG ACA GATAAGAGCCAACAA TTCGCC CTGGCTTGCCGGCGATTGG GGATCC TCAGGCGGT GATGATGCTGCGGATA respectively. These primers also introduced flanking Nde I and Bam HI restriction sites (underlined). After purification, the PCR product was dige sted sequentially with Nde I and Bam HI, then ligated with Nde I, Bam HI digested pET 15b. After transformation into E. coli ElectroTen Blue, plasmid DNA from a randomly chosen colony was isolated, restriction mapped and then sequenced to verify the desired s tructure. The resulting

PAGE 155

155 plasmid (designated pSF6) was used to transform E. coli BL21 Gold(DE3) strain for protein overexpression. Protein p urification. A single colony of E. coli BL21 Gold(DE3) containing pSF6 was used to inoculate 50 mL of LB medium sup plemented with 100 g/mL ampicillin. After shaking at 37 C until becoming turbid, a 40 mL portion of the preculture was added to 4 L of LB medium supplemented with 100 g/mL ampicillin, 80 mL of 20% glucose, and 1.5 mL of antifoam 204 in a New Brunswick M19 fermenter. The culture was grown at 37 C with stirring at 400 rpm and an air flow of 4 vvm until the O.D. 600 reached 0.5 0.6. Protein overexpression was induced by adding of 10 mL of 0.16 M IPTG (to yield a final concentration of 0.4 mM) and adjus ting the temperature to 30 C and shaking. After 3 hours, the cells were harvested by centrifugation at 6,300 g for 15 min at 4 C, resuspended in 50 mM KP i pH 8.0 ( 1 mL/g ) then lysed by a French pressure cell at 17,000 psi. Insoluble debris was pell eted by centrifugation at 39,000 g for 1 hr at 4 C and a portion of the yellow supernatant was used for TA catalyzed reactions. The remaining crude lysate was applied to a 5 mL HiTrap Chelating HP column (GE Healthcare Life Sciences) that had been equi librated with binding buffer (0.02 M NaH 2 PO 4 0.5 M NaCl, 20 mM imidazole, pH 7.4). After washing with 50 mL of binding buffer, the desired protein was eluted by elution buffer (0.02 M NaH 2 PO 4 0.5 M NaCl, 0.5 M imidazole, pH 7.4). A flow rate of 2 mL/mi n was employed throughout. The eluate was concentrated by ultrafiltration (Amicon Ultra) to yield a final concentration of 25 mg/mL. column (Pharmacia Biotech) that had been equilibrated with 50 mM Tris HCl, 50 mM NaCl, pH 7.4. The desired protein was eluted with the same buffer. A flow rate of

PAGE 156

156 0.3 mL/min was employed throughout. The eluate was concentrated by ultrafiltration (Amicon Ultra) to yield a final concentration of 25 mg/mL. The N termin al histidine tag was cleaved using thrombin from bovine plasma. The histid in e tag was separated from the desired protein by applying it to the same treatment described above using the 5 mL HiTrap Chelating HP column. The desired protein was eluted with b inding buffer and eluate was concentrated by ultrafiltration (Amicon Ultra), then diluted with 50 mM KP i pH 8.0 and re concentrated. This was repeated 2 more times. The final L TA sample was concentrated to 15 mg/mL and used for immediately for crystal screening. Crystal screenin g The first round of crystallization screening of the L TA from P. putida used two commercial screening kits from Hampton Research ( PEG Rx HT (HR2 086) and Crystal Screen Cryo HT (HR2 133) ) These kits were used with the s itting drop vapo r diffusion meth od at room temperature with 10 15 mg/mL protein concentrations ( added at a ratio of 1 : 1, precipitant solution : protein solution) This screening resulted in one only condition that formed quasi crystals. The second ro und used three commercial screening kits from Hampton Research ( Salt Rx HT (HR2 136), Index HT (HR2 134) and Additive Screen HT (HR2 138 ) ). The additive screen kit used the precipitant solution that formed quasi crystals fr om the first round of scre ening ( 0.075 M HEPES pH 7.5, 0.6 M NaH 2 PO 4 1H 2 O, 0.6 M KH 2 PO 4 25% glycerol ) T he screening was accomplished with the sitting drop vapor diffusion method at room temperature with 10 and 15 mg/mL protein concentrations (added at a ratio of 1 : 1, precipita nt solution : protein solution) Within 72 hours, t his screening revealed four conditions that formed quasi crystals and one that formed ordered crystals in well C 6 of the Salt Rx HT screening kit The precipitant

PAGE 157

157 solution consisted of 3.5 M sodium form ate, 0.1 M sodium acetate trihydrate pH 4.6. A large scale screening with careful observation using these same conditions revealed the protein had actually crystallized within 24 hours. Attempts to optimize crystal growth involved varying concentrations of both salts ( 0.7 M 6.3 M sodium formate and 0.1 1 M sod ium acetate trihydrate pH 4.6) and monitor ing crystal formation daily. The protein concentration used was 15 mg/mL and it was added at different ratios with the precipitant solution (1 : 4, 2 : 4, and 3 : 4) via both the hanging drop and sitting drop vapor diffusion method at room temperature. Within 24 hours, some conditions resulted in crystallized protein with a tetragonal bipyramid al shape ; however a few wells showed crystal formation in 1 20 hours with a precipitant solution consisting of 2.1 M (or 2.8 M) sodium formate, 0.1 M sodium acetate trihydrate pH 4.6. In the final optimiz ation screening study the concentrations of both salts (2.1 M 4.9 M sodium formate and 0.1 0.2 M sodium ace tate trihydrate pH 4.6) were varied and the protein concentration decreased. Crystal formation was monitored daily. The protein concentration used in this screening was 5 10 mg/mL and it was added at different ratios with the precipitant solution (1 : 4, 2 : 4, and 3 : 4) via the hanging drop vapour diffusion method at room temperature. As before, most of the conditions resulted in crystals after 24 hours, but careful observation revealed that after 240 hours a few rod shaped crystals had formed with a precipitant solution consisting of 2.1 M sodium formate, 0.1 M sodium acetate trihydrate pH 4.6. Two of these crystals were used to collect a complete data set for X ray diffraction analysis : one un soaked and one soaked with the natural substrate L Thr (5 mM) for 1 hour at room temperature.

PAGE 158

158 D ata collection The crystals were flash cooled in liquid nitrogen using 20% (v/v) glycerol as a cryoprotectant. X ray diffraction data were collected at the Advanced Photon Source (Argonne National Laboratory) o n beamline 21 ID G Reflection data were processed using XDS. 248 The structure of LTAPP was solved by the molecular replacement method using AUTOMR in the PHENIX suite 249 using the s tructure of L TA from E. coli (Accession number NC000913; PDB entry 4LNJ ) 206 Manual model building and refinement were performed with Coot 250 and phenix.refine 251 respectively. The structure was refined at 2.27 resolution with an R factor and an R free of 0.1799 and 0.2171, respectively. Amino a cid d erivatization. Reaction mixtures were dried completely under reduced pressure for 30 min by a Savant SpeedVac SVC100, the n the residue was After shaking at 37 C for 30 min, the mixtures were analyzed by GC/MS. General p rocedure for s creening a ldehyde a cceptors Reactions contained 0.1 mmol of aldehyde, 0.5 mmol of glycine, 10 in 50 mM KP i pH 8 buffer (total volume of 1 mL). These were gently rotated overnight at room temp erature and sampled after 4 hr and overnight for MSTFA derivatization and GC/MS analysis. General p rocedure for s cree ning a mino d onors Reactions contained 0.1 mmol of aldehyde (acetaldehyde, hexanal, or benzaldehyde) 0.5 mmol of amino donor 50 mM KP i pH 8 buffer (total volume of 1 mL). These were gently rotated overnight at room tempe rature and sampled after 4 hr and overnight for MSTFA derivatization and GC/MS analysis.

PAGE 159

159 Design of e xperiments To optimize reaction conditions, the L TA was used in a series of experiments with varied temperature, pH, and glycine ratio All reactions contained 0.1 mmol of aldehyde, 0.2 enzyme lysate in buffer plus 2% (v/v) ethanol in a total volume of 1 mL. The four different glycine concentrations were tested at five different temperatures (4 C 18 C 25 C 37 C and 42 C ) and four different pH values (6, 8, 9.7, and 12). The mixtures and overnight. Samples were derivatized with MSTFA and analyzed by GC/MS.

PAGE 160

160 ( a ) F igure 4 1. Ribbon representation of P. putida L threonine aldolase ( a ) The homotetramer of LTAPP. The four monomers are colored teal (chain A), green (chain B), orange (chain C), and magenta (chain D). ( b ) Chain C of LTAPP sho wing PLP binding in yellow. ( c ) Monomer of E. coli L TA (salmon) showing PLP binding in yellow for a comparison of structure with LTAPP. ( d g ) Zoomed view of N and C terminus highlighting main differences in structure. The purple loop indicates the e sheet in E. coli L TA (salmon) helix in T. maritima L allo TA (yellow) and the longer loop is shown in green. Part of the green loop wa s deleted in the A. jandaei L allo TA (brown) structure.

PAGE 161

161 ( b ) ( c ) ( d ) ( e ) ( f ) ( g ) Figure 4 1. Continued

PAGE 162

162 ( a ) ( b ) Figure 4 2. Overview of the active site of P. putida L threonine aldolase ( a ) Electrostatic surface potential displayed in blue for positive residues, red for negative, yello w for sulfur containing residues, and cyan for neutral residues. PLP is shown as sticks in active site hole. ( b ) Active site with conserved residues highlighted. PLP (yellow) is shown in predicted position.

PAGE 163

163 ( a ) ( b ) Figure 4 3. Overview of the active site pocket of chain A from P. putida L threonine aldolase showing key interactions between residues in the active site and the pyridoxal phosphate cofactor Red dashed lines show hydrogen bonding and green dashed lines show pi bonding interactions between His 89 and PLP. ( a ) PyMOL representation. Gray circle denotes the space in the active site that the substrate/product binds. ( b ) Chemical structure representation of internal aldimine. ( c ) Chemical structure represe ntation of Gly PLP. ( d ) Chemical structure representation of Thr PLP.

PAGE 164

164 ( c ) ( d ) Figure 4 3. Continued

PAGE 165

165 Figure 4 4. ESPript sequence alignment of four threonine aldolases ii ii Reference: 252. Robert, X.; Gouet, P., Nucleic Acids Research 2014, 42 (W1), W320 W324.

PAGE 166

166 ( a ) Figur e 4 5. D e tailed view of the active site pocket of chain A from P. putida L threonine aldolase showing key interactions between conserved residues in the active site Red dashed lines show hydrogen bonding between residues. Magenta residues highlight the conserved active site residues found after sequence alignment. ( a ) Tyr 35 lies in close range to the binding pocket. The gray circle represents the space in the active site in which the substrate/product likely binds. ( b ) Hydrogen bonding between Asn 6 8 Asp 174 and Glu 94 His 89. ( c ) Hydrogen bonding between Asn 290 Asp 11 Arg 177.

PAGE 167

167 ( b ) ( c ) Figure 4 5. Continued

PAGE 168

168 ( a ) Figure 4 6. Structu ral comparison of four L threonine aldolases T. maritima L allo TA (teal), Aeromonas jandaei L allo TA (pink), and Escherichia coli L TA (green) and the studied enzyme, P. putida L TA (orange). ( a ) Secondary structure alignments rotated at 90 to show structual similarit i es. ( b ) Active site alignment with key residues highlighted. PLP shown in yellow

PAGE 169

169 ( b ) Figure 4 6. Continued

PAGE 170

170 CHAPTER 5 SITE SATURATION MUTAGENESIS OF A. jandaei L allo THREONINE ALDOLASE Introduction Threonine aldolases (TAs) are PLP dependent enzymes that have proven useful hydroxy amino acids. 184, 223 They catalyze the C C bond formation between g lycine and an aldehyde acceptor, resul ting in two adjacent stereochemistry, but typically show much lower diastereoselectivity carbon particularly with longer reaction times or when larger aldehyde acceptors are used This loss of diastereo selectivity must be addressed before TAs can be considered synthetically useful Site saturation mutagenesis (SSM) has been widely employed to improve a variety of enzyme characteristics such as substrate range, stereoselectivity, and thermostability. 253 258 Mutations in and around the active site have usually proven to be the most influential for substrate specificity and enantioselectivity 259 262 A pplying SSM to the active site residues of a TA was an obvious strategy to tackle the diastereoselectivity problem TAs have been studied extensively for their substrate tolerance on a broad range of aldehyde acceptors, 184 but few mutagenesis studies have been performed with these enzymes to this date. These include the work done by Baik and coworkers on the enzymatic synthesis of droxidopa 8, 193, 238 They used error prone PCR on the L TA from S. coelicolor A3(2) and found that the triple mutation Y39C/T309C/A48T offered a four fold increase in overall conversion and three fold increase in diastereoselectivity (14% to 43% d.e. for t he syn isomer ) compared to the wild type enzyme. 193

PAGE 171

171 The crystal structure of A. jandaei L allo TA was solved by Qin et al. in 201 4 207 These authors also carried out random mutagenesis and SSM of His 128. Random mutagenesis and screening of approximately 3 000 colonies revealed only one variant that showed significant activity towa rds L allo and L Thr ( H128Y/S292R ) The double mutant showed a three fold and 322 fold increased k cat /K M towards the two substrates respectively E nzyme kinetic studies of the two single muta nt s revealed that H128Y was likely the position of increased selectivity towards L Thr. Therefore, Qin et al. preformed SSM on His 128 and found increased relative activity towards L Thr with the Phe, Trp, Tyr, Ile, Leu, Met, and Ser variants. Interestingly, H128Y proved superior. 207 Although these mutants were useful in the retro aldol cleavage of L Thr, Qin et al. did not further screen for aldol condensation to L Thr. Here, we report three SSM libraries of A. jandaei L allo TA at active site residues His 85, His 128, and Tyr 89. Each library was initially screened against three representative a ldehyde acceptors and interesting variants were further screened for altered substrate tolerance. We also hoped to solve the d iastereoselectivity problem by this approach Results and Discussion S ite Saturation Mutagenesis Library Construction and Overexpression The g ene encoding wild type L allo TA from A. jandaei w as originally obtained by chemical gene synthesis from GenScript. The native gene was ligated into pE T 15b and the resulting plasmid w as used to transform the E. coli overexpression strain BL21 Gold(DE3). Individual primer sets were deployed in 19 individual PCR amplifications to obtain a collection of plasmid s with each of the 19 possible variants. After co nfirming

PAGE 172

172 each mutation by Sangar sequencing, the plasmids were used to t ransform E. coli BL21 Gold(D E3) to isolat e each muta n t Histidine 85 Library Three positions within the active site were chosen to und ergo SSM on the L allo TA from A. jandaei : His 85, His 128, and Tyr 89. The crystal structure of this TA was solved by Qin et al. 207 in 2014 These data along with the structures from T. maritima 206 E. coli 206 and P. putida (Chapter 4) we re used to choose these three positions His tidine 85 was selected since it is responsible for regulating of the degree of stereo selectivit y between the L and L allo configurations by hydrogen bonding to the hydroxyl group of the product. Along with his tidine 85, histidine 128 also hydrogen bonds to the hydroxyl group of the products Qin et al. preformed SSM at this position and found enhanced activity for the retro aldol cleavage of L Thr ; however, these variants were never tested for their aldol cond ensation activity Histidine 128 was selected for this reason. And lastly, tyrosine 89 was chosen for its location in the hydrophobic pocket where the methyl group of L allo Thr resides. This residue is not conserved in the L TA family and therefore bel ieved to be an interesting site for SSM. 205 207 All three SSM libraries were screened against three aldehyde acceptors ( 3 9 and 1 1 ) These aldehydes (one aliphatic aldehyde and two aromatic aldehydes) were chose n to minimize screening efforts while still obtaining a broad picture of the substrate selectivity of the mutants. Unfortunately, all mutations at the 85 position rendered the enzyme completely inactive. This could be due to His 85 acting as the catalyti c base. Originally, water was predicted to be the catalytic base 205 207 for the retro aldol cleavage of L allo Thr; that is why we considered this site safe for mutagenesis. His 85 could

PAGE 173

173 also be important for its pi stacking interactions with the pyridinium ring of the PLP cofactor. It is possible tha t this His residue is required to stabiliz e and position the cofactor with in the active site See T able 5 1 for full screening results. Table 5 1. Screening of his tidine 85 mutants against aldehyde acceptors Entry Aldehyde Product Variant Conversion a (%) d.e. a (%) 1 3 H85G --H85A <1 b -b H85V <1 b -b H85I <1 b -b H8 5L <1 b -b H85P <1 b -b H85M <1 b -b H85F <1 b -b H85Y <1 b -b H85W --H85S <1 b -b H85C <1 b -b H85T <1 b -b H85N <1 b -b H85Q <1 b -b H85R <1 b -b H85K <1 b -b H85D --H85E --W.T. 25 2 6 2 9 H85G 2 -b H85A 3 -b H85V 5 18 H85I 5 25 H85L 1 -b H85P 2 -b H85M 2 -b H85F 6 9 H85Y 1 -b H85W 1 -b H85 S 2 -b H85C 1 -b H85T 6 18 H85N 3 -b H85Q 2 -b H85R 4 -b H85K 4 -b H85D 2 -b H85E 1 -b W.T. 19 26

PAGE 174

174 Table 5 1. Continued Entry Aldehyde Product Variant Conversion a (%) d.e. a (%) 3 11 H85G --H85A <1 b -b H85V <1 b -b H85I 1 -b H85L --H85P --H85M --H85F 3 -b H85Y <1 b -b H85W --H85S <1 b -b H85C -H85T 1 -b H85N <1 b -b H85Q <1 b -b H85R 6 99 H85K 3 -b H85D 3 -b H85E --W.T. 70 18 Note: Reaction mixtures contained 500 mM glycine, 100 mM aldehyde, 0.05 mM PLP, 20% ethanol, and in 1 mL of 50 mM Tris Base, pH 9.7. Reactions were incubated at 37 C. a Conversion and diastereomeric excess values were determined by GC/MS after MSTFA derivatization. r e omeric purity are provided since the small peak sizes preclude accurate integration. b Accurate values for conversion and diastereomeric excess could not be determined due to the very low peak areas. Histidine 128 Library The second His that occupies th e active site of t hese TAs is located at position 128 in the L allo TA from A. jandaei Along with His 85, t his His residue also hydrogen bonds to the hydroxyl group of the products. di Salvo et al. observed that this His was quite flexible, presenting t wo different conform ations in the electron density map of the crystal structure. 206 Qin et al. took this observation and mutated His 128 to Tyr and found that the Tyr side chain moved 4.2 out of the active site This group carried out SSM on thi s position, but only monitored the retro aldol cleavage of L allo Thr. They found that the enzyme tolerated substitution s by many amino acids, but the Tyr mutant gave the best results an 8.4 fold increase in selectivity towards L Thr They did not

PAGE 175

175 test aldol condensation activity. 207 We t herefore hypothesized that m utagenesis might create sufficient room within the active site to accommodate larger, more functionalized acceptor aldehydes and/or potentia lly different amino donors The initial screening results of the His 128 mutants revealed increased diastereoselectivity with aldehyde 3 On the other hand, m utations at this position n either increased con version nor diastereoselectivit i es for the aromati c aldehydes 9 and 1 1 These results are displayed in T able 5 2 Amino acid residues with increased diastereoselectivity were primarily hydrophobic amino acids like Ala, Ile, and Phe. Table 5 2. Initial screening of histidine 128 mutants against aldehyde acceptors Entry Aldehyde Product Variant Conversion a (%) d.e. a (%) 1 3 H128G <5 b -b H128A 12 60 H128V 37 11 H128I 18 99 H128L <5 b -b H128P <5 b -b H 128M <5 b -b H128F 5 99 H128Y 19 99 H128W --H128S --H128C --H128T <5 b -b H128N 31 99 H128Q <5 b -b H128R --H128K --H128D <5 b -b H128E <5 b -b W.T. 25 26 2 9 H128G <5 b -b H128A 26 16 H128V 12 12 H128I <5 b -b H128L <5 b -b H128P 17 14 H128M <5 b -b H128F 5 16 H128Y 16 4 H128W <5 b -b

PAGE 176

176 Table 5 2. Continued Entry A ldehyde Product Variant Conversion a (%) d.e. a (%) 2 9 H128S <5 b -b H128C <5 b -b H128T <5 b -b H128N 16 13 H128Q <5 b -b H128R <5 b -b H128K <5 b -b H128D <5 b -b H128E <5 b -b W.T. 19 26 3 1 1 H128G <5 b -b H128A 43 10 H128V 34 9 H128I <5 b -b H128L --H128P 34 8 H128M <5 b -b H 128F <5 b -b H128Y 24 17 H128W --H128S --H128C --H128T <5 b -b H128N 7 39 H128Q <5 b -b H128R <5 b -b H128K --H128D <5 b -b H128E <5 b -b W.T. 70 18 Note: Reaction mixtures contained 500 mM g lycine, 100 mM aldehyde, 0.05 mM PLP, 20% ethanol, and C. a Conversion and diastereomeric excess values were determined by GC/MS after MSTFA derivatization. Reacti ons with 1 e omeric purity are provided since the small peak sizes preclude accurate integration. b Accurate values for conversion and diastereomeric excess could not be determined due to the ver y low peak areas. More extensive screening of the Ile, Phe, Tyr, Asn, and other variants was carried out on a variety of aliphatic aldehydes i.e straight ch ain, cyclic, branched, and unsaturated aldehydes and a few aromatic aldehydes (Table s 5 3 and 5 4 ) The Ile mutant showed >90% d.e. across all aliphatic aldehydes screened ; however conversions were uniformly low (<20%). Neither increas ing the e nzyme concentration nor adjusting temperature or pH improved these conversions Substituting pur i fied

PAGE 177

177 eny zme was also unsuccessful The Asn mutant gave comparable diastereoselectivit i es (>85%, with the exception of aldehyde 38 ) along with reasonable conversions Applying DOE strategies to this mutant increased conversion onl y slight ly, at 25 C and pH 8 Ta ble 5 3. Extended screening of histidine 128 mutants against aliphatic aldehydes Entry Aldehyde Product Variant Conversion a (%) d.e. a (%) 1 1 H128V 5 99 H128I 2 -b H128T 3 -b H128P 5 99 H128F 3 -b H128Y 5 99 H128N 6 99 W.T. 9 99 2 26 H128I 3 -b H128F 3 -b H128Y 11 99 H128N 10 99 W.T. 27 99 3 27 H128I <1 b -b H128F --H128Y <1 b -b H128N <1 b -b W.T. 6 99 4 28 H128I --H128F 3 -b H128Y 4 -b H128N <1 b -b W.T. 3 -b 5 29 H128I 4 -b H128F 4 -b H128Y 24 99 H128N 16 96 W.T. 42 99 6 30 H128I 3 -b H128F 3 -b H128Y 14 99 H128N 17 96 W.T. 79 97 7 31 H128I 7 99 H128F 11 99 H12 8Y 12 99 H128N 31 99 W.T. 27 58

PAGE 178

178 Table 5 3. Continued Entry Aldehyde Product Variant Conversion a (%) d.e. a (%) 8 32 H128I 3 -b H128F 4 -b H128Y 3 -b H128N 3 -b W.T. 3 -b 9 33 H128I --H128F --H128Y --H128N --W.T. --10 34 H128I 1 -b H128F 3 22 H128Y 25 58 H128N 16 31 W.T. 64 58 11 2 H128A 14 99 H128V 43 99 H128I 9 99 H128T 30 99 H128P 28 99 H128F 5 99 H128Y 24 99 H128N 28 99 W.T. 80 99 12 3 5 H128I 19 93 H128F 16 57 H128Y 91 70 H128N 50 82 W.T. 92 83 13 3 6 H128I --H128F --H128Y <1 b -b H128N <1 b -b W.T. 4 12 14 3 7 H128I 5 -c H128F 10 -c H128Y --H128N 15 -c W.T. --15 20 H128I 5 99 H128F 5 99 H128Y 31 63 H128N 20 94 W.T. 34 41 16 38 H128I <1 b -b H128F --H128Y <1 b -b H128N 5 16 W.T. 3 -b

PAGE 179

179 Table 5 3. Continued Entry Aldehyde Product Variant Conversion a (%) d.e. a (%) 17 39 H128I --H128F 3 -b H128Y --H128N <1 b -b W.T. 2 -b Note: Reaction mixtures contained 500 mM glycine, 100 mM aldehyde, 0.05 mM PLP, 20% ethanol, and 0 mM Tris Base, pH 9.7. Reactions were incubated at 37 C. a Conversion and diastereomeric excess values were determined by GC/MS after MSTFA derivatization. e omeric pur ity are provided since the small peak sizes preclude accurate integration. b Accurate values for conversion and diastereomeric excess could not be determined due to the very low peak areas. c Accurate values for diastereomeric excess could not be determine d due to the original aldehyde acceptor being a racemic mixture and four possible products were made. Table 5 4. Extended screening of histidine 128 mutants against aromatic aldehydes Entry Aldehyde Product Variant Conversion a (%) d.e. a (%) 1 6 H128A 17 13 H128V 15 11 H128T 4 -b H128P 11 8 H128Y 16 5 W.T. 11 20 2 7 H128A 7 2 H128V 7 22 H128T 4 -b H128P 11 16 H128Y 7 38 W.T. 12 21 3 10 H128A 43 36 H128V 44 17 H128T <1 b -b H128P 22 87 H128Y 36 18 W.T. 63 30 4 12 H128A 35 40 H128V 41 33 H128T 5 62 H128P 29 30 H128Y 23 20 W.T. 66 41 Note: Reaction mixtures contained 500 mM glycine, 100 mM aldehyde, 0.0 5 mM PLP, 20% ethanol, and C. a Conversion and diastereomeric excess values were determined by GC/MS after MSTFA derivatization. Reactions with 1 5% conversions ar e omeric purity are provided since the small peak sizes preclude accurate integration. b Accurate values for conversion and diastereomeric excess could not be determined due to the very low peak areas.

PAGE 180

180 T yrosine 89 Library We also carried out SSM on the non conserved amino acid at position 89, which is most commonly a Tyr in L TAs, 206, 207 but can be Phe 205 or Asp (Chapter 4) This residue is located in the hydrophobic pocket occupied by the methyl group of L allo Thr when it is bound as an external aldimine with PLP. This residue i s thought to control the preference for this isomer. 205 207 Since th e residue at position 89 is not conserved, we hypothesized that a mutation at this position may allow for larger aldehyde acceptors to bind The initial screening of these mutation s offered some unexp ected results. Figure 5 2 represents the amino acid mutations that off er conversions >5%. The Gly, Glu and Trp substitutions offered increased diastereoselectivit ies with aldehyde 3 as the substrate. Interestingly the steric bulk of the amino acid side chain did correlate with increased diastereoselectivity since both Gly and Trp showed the same effect When all 19 variants were tested with benzaldehyde 9 we found increased diastereoselec tivity with the Pro and Glu mutants With respect to the 3 pyri dinecarboxaldehyde 1 1 diastereoselectivity was increased only slightly by Asn and Gln substitution s By contrast the Pro mutant showed complete ly reverse d diastereoselectivity. T he Ala, Gln and Lys variants also showed reversed diastereoselectivity T he Y89P mutants increased the wild type diastereoselectivity ( to 81% d.e. ) with benzaldehyde 9 but gave almost completely reversed selectivity for 3 pyridinecarboxaldehyde 1 1 (to 91% d.e.) despite the very similar sizes of the aldehydes Table 5 5 summar izes these results for the initial screening.

PAGE 181

181 Table 5 5. Initial screening of tyrosine 89 mutants against aldehyde acceptors Entry Aldehyde Product Variant Conversion a (%) d.e. a (%) 1 3 Y89G 12 99 Y89A 6 6 Y89V <5 b -b Y89I <5 b -b Y89L 7 65 Y89P <5 b -b Y89M <5 b -b Y89H 5 34 Y89F 5 99 Y89W 6 99 Y89S <5 b -b Y89C 6 13 Y89T <5 b -b Y89N <5 b -b Y89Q <5 b -b Y89R <5 b -b Y89K <5 b -b Y89D 15 26 Y89E 23 88 W.T. 25 26 2 9 Y89G 10 15 Y89A 12 50 Y89V <5 b -b Y89I <5 b -b Y89L 10 21 Y89P 5 81 Y89M 9 17 Y89H 8 32 Y89F 10 22 Y89W 9 33 Y89S 6 26 Y89C 12 40 Y89T 8 18 Y89N 12 49 Y89Q 11 58 Y89R <5 b -b Y89K 12 36 Y89D 16 48 Y89E 7 85 W.T. 19 26 3 11 Y89G 18 14 c Y89A 28 52 c Y89V --Y89I <5 b -b Y89L 35 13 c Y89P 7 91 c Y89M 19 12 c Y89H 12 18 c Y89F 17 16 c Y89W 27 11 c

PAGE 182

182 Table 5 5. Continued Entry Aldehyde Product Variant Conversion a (%) d.e. a (%) 3 11 Y89S 16 15 c Y89C 26 36 c Y89T 25 7 c Y89N 31 70 Y89Q 24 54 c Y89R <5 b -b Y89K 25 44 c Y89D 47 31 Y89E 38 41 W .T. 70 18 Note: Reaction mixtures contained 500 mM glycine, 100 mM aldehyde, 0.05 mM PLP, 20% ethanol, and a Conversion and diastereomeric excess values were de termined by GC/MS after MSTFA derivatization. Reactions with 1 e omeric purity are provided since the small peak sizes preclude accurate integration. b Accurate values for conversion and diaster eomeric excess could not be determined due to the very low peak areas. c Reverse diastereoselec tivity was observed. More e xten sive screening was performed on interesting variants identified by the initial characterization The Gly, Trp, Asp and Glu varia nts were tested against a variety of a liphatic aldehyde acceptors (Table 5 6) In most cases, diastereoselectivity was increased compared to the wild type enzyme by at least one of the four mutations. Non cyclic, branched aldehyde acceptors ( 28 31 and 3 4 ) revealed reverse d diastereoselectivity with all four mutations. Pivaldehyde 28 was particularly interesting as the overall conversion was increased seven fold by the Y89D mutation which also completely revers ed the diastereoselectivity (99% d.e.). Is obutryaldehyde 31 presented similar results : a 2.7 fold increase in conversion (73% ) as well as 99% d.e. for the opposite diastereomer Surprisingly cycl ic aliphatic aldehydes did not show reversed diastereoselectivity.

PAGE 183

183 Table 5 6. Extended screening of tyrosine 89 mutants against aliphatic aldehydes Entry Aldehyde Product Variant Conversion a (%) d.e. a (%) 1 1 Y89G 3 -b Y89W 4 -b Y89D 8 78 Y89E 4 -b W.T. 9 99 2 26 Y89G 8 99 Y89W 11 99 Y89D 33 99 Y89E 15 99 W.T. 27 99 3 27 Y89G 1 -b Y89W 2 -b Y89D <1 -b Y89E <1 -b W.T. 6 99 4 28 Y89G 9 99 d Y89W 11 99 d Y89D 22 99 d Y89E 13 99 d W.T. 3 -b 5 29 Y89G 8 99 Y89W 11 99 Y89D 26 99 Y89E 10 99 W.T. 42 99 6 30 Y89G 14 99 Y89W 17 72 Y89D 70 58 Y8 9E 75 66 W.T. 79 97 7 31 Y89G 9 92 d Y89W 11 94 d Y89D 73 99 d Y89E 22 97 d W.T. 27 58 8 32 Y89G 2 -b Y89W 4 -b Y89D 16 -c Y89E 3 -b W.T. 3 -b 9 33 Y89G --Y89W --Y89D --Y89E --W.T. --10 34 Y89G <1 -b Y89W 1 -b Y89D 8 74 d Y89E <1 -b W.T. 64 58

PAGE 184

184 Table 5 6. Continued Entry Aldehyde Product Variant Conversion a (%) d.e. a (%) 11 2 Y89G 6 93 Y89W 12 99 Y89D 40 99 Y89E 9 99 W.T. 80 99 12 35 Y89G 12 89 Y89W 14 16 Y89D 23 57 Y89E 20 80 W.T. 92 83 13 36 Y89G --Y89W <1 -b Y89D <1 -b Y89E <1 -b W.T. 4 -b 14 37 Y89G <1 -b Y89W 2 -b Y89D 2 -b Y89E 1 -b W.T. --15 20 Y89G <1 -b Y89W <1 -b Y89D <1 -b Y89E --W.T. 34 41 16 38 Y89G --Y89W --Y89D --Y89E --W.T. 3 -b 17 39 Y89G -Y89W 3 -b Y89D 6 -c Y89E 2 -b W.T. 2 -b Note: Reaction mixtures contained 500 mM glycine, 100 mM aldehyde, 0.05 mM PLP, 20% ethanol, and 37 C. a Conversion and diastereomeric excess values were determined by GC/MS after MSTFA derivatization. Reactions with 1 e omeric purity are provided since the small peak sizes preclude accura te integration. b Accurate values for conversion and diastereomeric excess could not be determined due to the very low peak areas. c Accurate values for diastereomeric excess could not be determined due to the original aldehyde acceptor being a racemic mix ture and four possible products were made. d Reverse diastereoselec tivity was observed. The final screening efforts for this library were carried out to investigate the reversed diastereoselectivity observed for aldehyde 11 The Ala, Leu, Pro and Glu

PAGE 185

185 mut ants were tested with a few aromatic aldehyde acceptors (Table 5 7). The Pro substitution (Y89P) was by far the most interesting mutation as it single handedly triggered elevated and reversed diastereoselectivity (91% d.e.) for aldehyde 1 0 (Table 5 7, ent ry 3) and 11 (Table 5 5, entry 3) O n the other hand it increased diastereoselectivity (81% d.e.) for the original diastereomer with benzaldehyde 9 as the substrate (Table 5 5, entry 2) Table 5 7. Extended screening of tyrosine 89 mutants against aro matic aldehydes Entry Aldehyde Product Enzyme Conversion a (%) d.e. a (%) 1 6 Y89A 11 92 Y89L 5 4 Y89P 9 30 Y89E --W.T. 11 20 2 7 Y89A 3 -b Y89L 3 -b Y89P <1 -b Y89E --W.T. 12 21 3 10 Y89A 46 31 c Y89L --Y89P 29 57 c Y89E 1 -b W.T. 63 30 4 12 Y89A 7 3 Y89L 9 24 Y89P 6 12 Y89E 12 26 W.T. 66 41 Note: Reaction mixtures contained 500 mM glycine, 100 mM aldeh yde, 0.05 mM PLP, 20% ethanol, and a Conversion and diastereomeric excess values were determined by GC/MS after MSTFA derivatization. Reactions with 1 5% conv e omeric purity are provided since the small peak sizes preclude accurate integration. b Accurate values for conversion and diastereomeric excess could not be determined due to the very low peak areas. c Reverse diastereoselec tivity was observed. In the previous chapter, it was observed that the L TA from P. putida contained an Asp residue at position 93, analogous to the Tyr 89 in A. jandaei We predicted that this position may have been the culprit for the observed poor diatereoselectivites of this

PAGE 186

186 enzyme. In the case of the A. jandaei L allo TA T yr 89 SSM library, extended screening revealed that the Y89D mutation actually gave some of the best diatereoselectivites and overall conversions. Therefore, the poor diastereoselectivity observed for P. putida L TA must be due to additional factors beyond the residue at position 89. Conclusion Three SSM libraries were constructed and screened against a wide range of aldehyde acceptors. The complete collectio n of mutants were tested with three representative aldehydes ( 3 9 and 11 ) Interesting variants were then tested against a broader range of aldehydes to uncover their effects on substrate tolerance and diastereoselectivity Mutations at His 85 left the enzyme completely inactive indicating that this amino acid was absolutely crucial, perhaps because of cofactor stabilization or because it acts as the catalytic base. Substitutions for His 128 increased diastereoselectivity ( Asn or Ile ) particularly for branched aldehyde acceptors. Reverse d diastereoselectivity was observed for branched, aliphatic aldehyde acceptors and pyridinium aldehyde acceptors when T yr 89 was mutated. The Y 89P variant was among the most interesting mutation in this series : this s ingle change increased diastereoselectivity for aldehyde 9 while reversing diastereoselectivity for aldehyde 11 In conclusion, w e have benefitted from SSM strategies and solved the diastereoselectivity problem for at least some of the substrates. Experim ental Procedures General LB medium contained 10 g/L Bacto Tryptone, 5 g/L Bacto Yeast Extract and 10 g/L NaCl; 15 g/L agar was added for plates. PCR amplifications were protocol s. Electroporation was carried out with a BioRad GenePulser apparatus using

PAGE 187

187 0.2 cm cuvettes. Promega Wizard kits and CsCl buoyant density ultracentrifugation 244 were used for small and large scale plasmid purifications, respectively. Fluor escent Sanger DNA sequencing was performed by the University of Florida ICBR. GC/MS eV. The temperature program involved an initial hold at 95 C for 5 min, an initial inc rease of 5 C/min to 138 C followed by an increase of 10 C/min to 180 C, then a final increase of 2 C/min to 200 C and a hold at that temperature for 10 min. Plasmid c onstruction The gene encoding L allo TA from A. jandaei (accession number D87890) was synthesized by GenScript and ligated into a pUC57 with flanking Nde I and Xho I restriction sites at the 5' and 3' ends, respectively. The TA gene was excised by digesting with these restriction enzymes and ligated with Nde I, Xho I cut pET15b (Novagen) After transformation into E. coli ElectroTen Blue, plasmid DNA from a randomly chosen colony was isolated, restriction mapped and then sequenced to verify the desired structure. The resulting plasmid (designated pSF3) was used to transform E. coli BL21 Gold(DE3) strain for protein overexpression. Site s aturation m utagenesis. Each mutation was accomplished by PCR with individual primer sets using pSF3 as the template DNA and the corresponding forward and reverse primers for each mutation (Table 5 8) T hese primer sets introduced a specific mutation at a specific site using the most frequent codon usage for E. coli (underlined in Table 5 8 ). After purification, the PCR product was digested with Dpn I to cut the template DNA (pSF3) After transformation into E. coli ElectroTen Blue, plasmid DNA from a randomly chosen colony was isolated and sequenced to verify the desired

PAGE 188

188 mutation The resulting plasmid with mutation was used to transform E. coli BL21 Gold(DE3) strain for protein overexpression. Table 5 8. Forward and reverse primer sets for site saturation mutagenesis Mutation of pSF3 Forward Primer Reverse Primer H85G GCAGCC GGC ATCTAT CGCTATGAGGCG ATAGAT GCC GGCTGC CGAGCC H85A GCAGCC GCG ATCTAT CGCTATGAGGCGC AT AGAT CGC GGCTGC CGAGC H85V GCAGCC GTG ATCTAT CGCTATGAGGCGCAG ATAGAT CAC GGCTGC CGAGCC H85I GCAGCC ATT ATCTAT CGCTATGAGGCGCAGG ATAGAT AAT GGCTGC CGAGCCC H85L GCAGCC CTG ATCTAT CGCTATGAGGCGCAG AT AGAT CAG GGCTGC CGAGCC H85P GCAGCC CCG ATCTAT CGCTATGAGGCGCAG ATAGAT CGG GGCTGC CGAGC H85M GCAGCC ATG ATCTAT CGCTATGAGGCGCAG ATAGAT CAT GGCTGC CGAGCCC H85F GCAGCC TTT ATCTAT CGCTATGAGGCGCA G G ATA GAT AAA GGCTGC CGAGCCCAG H85 Y GCAGCC TAT ATCTAT CGCTATGAGGCGCAGG ATAGAT ATA GGCTGC CGAGCCCAG H85W GCAGCC TGG ATCTAT CGCTATGAGGCGC ATAGAT CCA GGCTGC CGAGCC H85S GCAGCC AGC ATCTAT CGCTATGAGGCGCA ATAGAT GCT GGCTGC CGAGCCCA H85C GCAGCC TGC ATCTAT CGCTATGAGGCGC ATAGAT GCA GGCTGC CGAGCCCA H85T GCAGCC ACC ATCTAT CGCTATGAGGCGCA ATAGAT GGT GGCTGC CGAGCCCA H85N GCAGCC AAC ATCTAT CGCTATGAGGCGCAG ATAGAT GTT GGCTGC CGAGCCCA H85Q GCAGCC CAG ATCTAT CGCTATGAGGCGCAG ATAGAT CTG GGCTGC CGAGCC H85R GCAGCC CGT ATCTAT CGCTATGAGGCGCAGG ATAGAT ACG GGCTGC CGAGCC H85K GCAGCC AAA ATCTAT CGCTATGAGGCGCAG 5 ATAGAT TTT GGCTGC CGAGCCC H85D GCAGCC GAT ATCTAT CGCTATGAGGCGCAGG ATAGAT ATC GGCTGC CGAGCCC H85E GCAGCC GAA ATCTAT CGCTATGAGGCGCAG ATAGAT TTC GGCTGC CGAGCC Y89G TATCGC GGC GAGGCG CAGGGTT CG CCTC GCC GCGATA GATGTGGGCT Y89A TATCGC GCG GAGGCG CAGGGTT CGCCTC CGC GCGATA GATGTGGGCTGC Y89V TATCGC GTG GAGGCG CAGGGTTCT CGCCTC CAC GCGATA GATGTGGGC Y89I TATCGC ATT GAGGCG CAGGGTTCTG CGCCTC AAT GCGATA GATGTGGGCTG

PAGE 189

189 Table 5 8. Continued Mutation of pSF3 Forward Primer Reverse Primer Y89L TATCGC CTG GAGGCG CAGGGTTCTG CGCCTC CAG GCGATA GATGTGGGCTGC Y89P TATCGC CCG GAGGCG CAGGGTT CGCCTC CGG GCGATA GATGT GGGCT Y89M TATCGC ATG GAGGCG CAGGGTTCTG CGCCTC CAT GCGATA GATGTGGGCTGC Y89H TATCGC CAT GAGGCG CAGGGTTCTG CGCCTC ATG GCGATA GATGTGGGCTGC Y89F TATCGC TTT GAGGCG CAGGGTTCTG CGCCTC AAA GCGATA GATGT GGGCTGCC Y89W TATCGC TGG GAGGCG CAGGGTTCTG CGCCTC CCA GCGATA GATGTGGGCTGC Y89S TATCGC AGC GAGGCG CAGGGTTCT CGCCTC GCT GCGATA GATGTGGGCTGC Y89C TATCGC TGC GAGGCG CAGGGTTCT CGCCTC GCA GCGATA GATG TGGGCTGC Y89T TATCGC ACC GAGGCG CAGGGTTCT CGCCTC GGT GCGATA GATGTGGGCTGC Y89N TATCGC AAC GAGGCG CAGGGTTCTG CGCCTC GTT GCGATA GATGTGGGCTGC Y89Q TATCGC CAG GAGGCG CAGGGTTCTG CGCCTC CTG GCGATA GAT GTGGGCTGC Y89R TATCGC CGT GAGGCG CAGGGTTCT CGCCTC ACG GCGATA GATGTGGGCTG Y89K TATCGC AAA GAGGCG CAGGGTTCTGCT CGCCTC TTT GCGATA GATGTGGGCTGC Y89D TATCGC GAT GAGGCG CAGGGTTCTG CGCCTC ATC GCGATA G ATGTGGGCTGC Y89E TATCGC GAA GAGGCG CAGGGTTCTG CGCCTC TTC GCGATA GATGTGGGCTG H128G GCCCCTGACGATGTC GGC TTTACCCCGACTCGC GCGAGTCGGGGTAAA GCC GACATCGTCAGGGG H128A GCCCCTGACGATGTC GCG TTTACCCCGACTCGC GCGAGTCGGGGTAAA CGC GACATCGTCAGGGG H128V GCCCCTGACGATGTC GTG TTTACCCCGACTCGCCTC GCGAGTCGGGGTAAA CAC GACATCGTCAGGGGC H128I GCCCCTGACGATGTC ATT TTTACCCCGACTCGCCTCGT GCGAGTCGGGGTAAA AAT GACATCGTCAGGGGCGATG H128L GCCCCTGACGATGTC CTG TTTACCCCGACTCGCCT GCGAGTCGGGGTAAA CAG GACATCGTCAGGGGC H128P GCCCCTGACGATGTC CCG TTTACCCCGACTCGCCT GCGAGTCGGGGTAAA CGG GACATCGTCAGGGG H128M GCCCCTGACGATGTC ATG TTTACCCCGACTCGC CTCG GCGAGTCGGGGTAAA CAT GACATCGTCAGGGGCGAT H128F GCCCCTGACGATGTC TTT TTTACCCCGACTCGCCTCG GCGAGTCGGGGTAAA AAA GACATCGTCAGGGGCGATG H128Y GCCCCTGACGATGTC TAT TTT ACCCCGACTCGCCTCGTCT GCGAGTCGGGGTAAA ATA GACATCGTCAGGGGCGATG H128W GCCCCTGACGATGTC TGG TTTACCCCGACTCGCC 5 GCGAGTCGGGGTAAA CCA GACATCGTCAGGGGC H128S GCCCCTGACGATGTC AGC TTTACCCCGACTCGCCTC GCGAGTCGGGGTAAA GCT GACATCGTCAGGGGC

PAGE 190

190 Table 5 8. Continued Mu tation of pSF3 Forward Primer Reverse Primer H128C GCCCCTGACGATGTC TGC TTTACCCCGACTCGC 5 GCGAGTCGGGGTAAA GCA GACATCGTCAGGGGC H128T GCCCCTGACGATGTC ACC TTTACCCCGACTCGCC GCGAGTCGGGGTAAA GGT GACATCGTCAGGGGC H128 N GCCCCTGACGATGTC AAC TTTACCCCGACTCGCCTC GCGAGTCGGGGTAAA GTT GACATCGTCAGGGGCG H128Q GCCCCTGACGATGTC CAG TTTACCCCGACTCGCCTC GCGAGTCGGGGTAAA CTG GACATCGTCAGGGGC H128R GCCCCTGACGATGTC CGT TTTACCCCGACTCGCCTCT 5 GCGAGTCGGGGTAAA ACG GACATCGTCAGGGGC H128K GCCCCTGACGATGTC AAA TTTACCCCGACTCGCCTCT GCGAGTCGGGGTAAA TTT GACATCGTCAGGGGC H128D GCCCCTGACGATGTC GAT TTTACCCCGACTCGCCTCTC 5 GCGAGTCGGGGTAAA ATC GACATCGTCAGGG GCG H128E GCCCCTGACGATGTC GAA TTTACCCCGACTCGCCTC 5 GCGAGTCGGGGTAAA TTC GACATCGTCAGGGGC Note: Mutation is underlined in each primer. Protein o verexpression. A single colony of the appropriate strain was used to inoculate 25 mL o overnight at 37 C, a 10 mL portion of the preculture was added to 1 L of LB medium C with shaking at 250 rpm until the O.D. 600 reached 0.5 0.6. Protein overexpression was induced by adding 0.4 mL of 0.16 M IPTG (to yield a final concentration of 0.4 mM). After 3 hr, the cells were har vested by centrifuging at 6,300 g for 15 min at 4 C, resuspended in 50 mM KP i pH 8.0 (1 mL buffer per gram wcw), then lysed by a French pressure cell at 17,000 psi. Insoluble debris was pelleted by centrifuging at 39,000 g for 1 hr at 4 C and the yellow supernatant was used for TA catalyzed reactions. Glycerol was added to a final concentration of 20% and the protein was stored in aliquots at 80 C Amino a cid d erivatization. Reaction mixtures were dried completely under reduced pressure for 30 min by a Savant SpeedVac SVC100, then the residue was After shaking at 37 C for 30 min, the mixtures were analyzed by GC/MS.

PAGE 191

191 Enzyme a ssays The activity of each L TA mutant was measured by mixing 0.1 mM KPi, p H 8.0 (total volume of 1 mL). The mixture was gently rotated at room 20 hr Samples were derivatized with MSTFA and analyzed by GC/MS. The temperature program involved an initial hold at 95 C for 5 min, an initial increase of 5 C / min to 1 20 C followed by an increase of 2 C / min to 1 38 C then a final increase of 10 C / min to 200 C and a hold at that temperature for 5 min General p rocedure for s creening a ldehyde a cceptors. Reactions contained 0 .1 mmol of aldehyde, 0.5 mmol of glycine, 10 nmol of PLP and 20 in 50 mM Tris base, pH 9.7 (total volume of 1 mL). These were gently rotated overnight at 37 C and sampled after 20 hr for MSTFA derivatization and GC/MS analysis.

PAGE 192

192 ( a ) ( b ) Figure 5 1. Initial screening of H1 28 mutants against the three aldehydes: ( a ) cyclohexanecarboxaldehyde, ( b ) benzaldehyde, and ( c ) 3 pyridinecarboxaldehyde. Each mutation was represented by its one letter abbreviation. W.T. stands for the wild type enzyme for comparison. All 20 amino ac ids were put through this initial screening, however if a conversion of <5% was observed, it was excluded from the graph. The magnitude of the each chart represents the % conversion with major (blue) and minor (orange) diastereomers noted. Reaction condi tions are as follows: Reaction mixtures contained 500 mM glycine, 100 mM aldehyde, 0.05 mM PLP, 20% ethanol, Tris Base, pH 9.7 Reactions were incubated at 37 C. 0 5 10 15 20 25 30 35 40 A V I F Y N W.T. H128 Mutants Against Cyclohexanecarboxaldehyde Major Diastereomer Minor Diastereomer 0 5 10 15 20 25 30 A V P F Y N W.T. H128 Mutants Against Benzaldehyde Major Diastereomer Minor Diastereomer

PAGE 193

193 ( c ) Figure 5 1 Continued 0 10 20 30 40 50 60 70 80 A V P Y N W.T. H128 Mutants Against 3 Pyridinecarboxaldehyde Major Diastereomer Minor Diastereomer

PAGE 194

194 ( a ) Figure 5 2. Initial screening of Y89 mutants against the three aldehydes: ( a ) cyclohexanecarboxaldehyde, ( b ) benzaldehyde, and ( c ) 3 pyridinecarboxaldehyde. Each mutation was represented by its one letter abbreviation. W.T. stands for the wild type en zyme for comparison. All 20 amino acids were put through this initial screening, however if a conversion of <5% was observed, it was excluded from the graph. The magnitude of the each chart represents the % conversion with major (blue) and minor (ora nge) diastereomers noted. In some cases diastereoselectivity was reversed, indicated the original diastereomer in blue. Reaction conditions are as follows: Reaction mixtures contained 500 mM glycine, 100 mM aldehyde, of enzyme lysate in 1 mL of 50 mM Tris Base, pH 9.7 Reactions were incubated at 37 C. 0 5 10 15 20 25 30 G A L F H W C D E W.T. Y89 Mutants Against Cyclohexanecarboxaldehyde Major Diasteromer Minor Diastereomer

PAGE 195

195 ( b ) ( c ) Figure 5 2. Continued 0 2 4 6 8 10 12 14 16 18 20 G A L P M F H W S C T N Q K D E W.T. Y89 Mutants Against Benzaldehyde Major Diastereomer Minor Diastereomer 0 10 20 30 40 50 60 70 80 G A L P M F H W S C T N Q K D E W.T. Y89 Mutants Against 3 Pyridinecarboxaldehyde "Major" Diastereomer "Minor" Diasteromer

PAGE 196

196 CHAPTER 6 BIOCATALYTIC SYNTHESIS OF TERTIARY METHYL FLUORINAT ED HYDROXY AMINO ACIDS BY THREONINE ALDOLASE Introduct ion Hydroxy amino acids have been valuable precursors for wide variety of pharmaceuticals. 263 266 Zhou, Toone and coworkers discovered LPC 058, ( difluoromethyl allo threonyl hydroxamate ) a broad spectrum antib iotic. 153, 157 LPC 058 inhibits the enzyme UDP 3 O (R 3 hydroxymyristoyl) N acetylglucosamine deacetylase (LpxC), which catalyzes the irreversible first step in the biosynthesis of lipid A. 155, 156 Zhou, Toone, and coworkers initially reported a 10 step total synthesis 153 which was later shortened to 8 steps 157 by modifying the synthesis of the difluoromethylthreonine methyl ester. Threonine aldolases (TAs) are pyridoxal phosphate (PLP) dependent enzymes that yield hydroxy amino acids by a one step aldol condensation between glycine and an aldehyde acceptor. TAs have been shown to accept a wide variety of aldehyde partners ranging from long chain aliphatic chains to complex aromatic side chains 184 The Griengl group also identified a handful of TAs that accepted more complex amino acids such as D Ala D Ser and D Cys increasing the utility of these en zymes. 184, 188, 198 However, to the best of our knowledge, no TA has catalyzed an aldol reaction between an amino acids and a ketone acceptor. Here, we report the first example of such a reaction (Figure 6 1) Re sults and Discussion Screening of Fluorinated Acetones Gene cloning, protein overexpression, and optimized reaction conditions were completed as previously described (Chapter s 3 5 ). The following L TA s w ere

PAGE 197

197 investigated for their tolerance toward four f luorinated acetones 40 43 : L allo TA from A. jandaei H128N L allo TA from A. jandaei Y89D L allo TA from A. jandaei Y89P L allo TA from A. jandaei and L TA from P. putida (Table 6 1). As expected, no aldol condensation was observed when the enzyme s were tested with glycine and trifluoro acetone 4 3 as >99% of this ketone is in the hydrate form. 267 While ketones 4 0 41 and 42 all gave measurable aldol products yields were below 10 %. Interestingly for ketone 40 both that Y89D and Y89P variants of A. jandaei L allo TA favored the opposite diastereomer versus that afforded by the wild type enzyme (Table 6 1, entry 1). The symmetry of ketone 41 i s advantageous since o nly one chiral center was formed upon aldol condensation by an L TA. T hese enzymes generally exert virtually complete control over the configuration at the carbon all e.e. values were >99%. S creening reactions used a ten fold molar excess of glycine versus the ketone acceptor to afford maximum conversions, although the standard five fold molar excess gave similar results Thus while the overall conver sio ns were poor, our goal in this phase was to determine whether these five enzymes could tolerate a ketone acceptor Table 6 1. L TA c atalyzed a ldol c ondensation of f luorinated a cetones Entry Fluorinated Acetone Product Enzyme Reaction Time (hr) Conve rsion a (%) d.e. a (%) 1 40 A. jandaei L allo TA 4 20 3 4 -b -b H128N A. jandaei L allo TA 4 20 2 4 -b -b Y89D A. jandaei L allo TA 4 20 <1 <1 -b,d -b,d Y89P A. jandaei L allo TA 4 20 <1 <1 -b,d -b,d P. putida L TA 4 20 3 3 -b -b

PAGE 198

198 Table 6 1. Continued Entry Fluorinated Acetone Product Enzyme Reaction Time (hr) Conversion a (%) d.e. a (%) 2 41 A. jandaei L allo TA 4 20 5 7 99 c 99 c H128N A. jandaei L allo TA 4 20 5 6 99 c 99 c Y89D A. jandaei L allo TA 4 20 <1 1 -b -b Y89P A. jandaei L allo TA 4 20 2 2 -b -b P. putida L TA 4 20 4 6 -b 99 c 3 42 A. jandaei L allo TA 4 20 8 10 30 23 H128N A. jandaei L allo TA 4 20 5 5 32 1 Y89D A. jandaei L allo TA 4 20 <1 <1 -b -b Y89P A. jandaei L allo TA 4 20 1 2 -b -b P. putida L TA 4 20 6 10 36 2 4 43 -A. jandaei L allo TA 4 2 0 ----H128N A. jandaei L allo TA 4 20 ----Y89D A. jandaei L allo TA 4 20 ----Y89P A. jandaei L allo TA 4 20 ----P. putida L TA 4 20 ----Note: Reaction mixtures contained 1 M of enzyme lysate in 1 mL of buffer. Reactions with A. jandaei L allo TA and its mutants were in 50 mM Tris base, pH 9.7 and were incubated at 37 C. Reactions with P. putida L TA were in 50 mM KP i pH 8 and w ere incubated at room temperature a Conversion and diastereomeric excess values were determined by GC/MS after MSTFA derivatization. Reactions with 1 provided since the small peak sizes preclude accurate integration. b Accurate values for conversion and diastereomeric excess could not be determined due to the very low peak areas. c This value was enantiomeric excess as a result of the symmetrical ketone. d Reve rse diastereoselec tivity was observed.

PAGE 199

199 Preparative Conversions Since the overall conversion s w ere very low for all of these ketone substrates, we scale d up the reactions t o provide sufficient amount s of products to allow 1 H NMR analysis All pr eparative conversions were accomplished with wild type enzymes, as they provided best overall conversions. Screening reaction conditions were scaled up ten fold. After the 20 hr, any unreacted aldehyde was removed by extracting with Et 2 O then the water was remove d by lyophilization. The solid was extracted with two portions of methanol, which left most unreacted glycine and buffer salts undissolved. After evaporating the methanol, the residue was dissolved in buffer at pH 8.0 and glycine oxidase was added to dec ompose any remaining glycine. An anion exchange resin eluted with 0.5% aqueous acetic acid was used for final purification prior to final lyophilization the afforded the aldol products. The low yields were mainly due to poor conversions by the enzymes, rather than to losses during isolation. While yields were disappointing enough material was isolated for MS and NMR analysis (Figure s A 10 and A 11 13 ) We first confirmed product formation MS analysis of the product from ketone 42 i The theoretical [M H] = 168.0478 was observed as 168.0473, confirming the aldol condensation with ketone 42 (Figure A 10 ). Both 1 H and 19 F NMR also confirmed the aldol product (Figure A 11 and 12 ) resulting in the first known example of threonine aldolases in the synthe trisubstituted amino acids. Both ketones 40 and 41 were also scaled up ten fold; however yie lds for ketone 41 were too low for NMR analysis. Both 1 H and 19 F NMR confirmed the aldol product from ketone 40 (Figure A 13). i Nominal mass spectrometry was determined by the UF Mass Spectrometry Lab.

PAGE 200

200 Conclusion In summary, fo ur fluorinated acetones were tested as substrates for the L TA from P. putida and A. jandaei and a few mutations of A. jandaei Three of the four ketones were tolerated (albeit poorly). The product from ketone s 40 and 42 was isolated on a sufficient scal e to allow MS and NMR analysis which confirmed its structure. This represents the first example of a ketone acceptor for a threonine aldolase. Experimental Procedures General. LB medium contained 10 g/L Bacto Tryptone, 5 g/L Bacto Yeast Extract and 10 g /L NaCl; 15 g/L agar was added for plates. PCR amplifications were protocols. Electroporation was carried out with a BioRad GenePulser apparatus using 0.2 cm cuvettes. Promega W izard kits and CsCl buoyant density ultracentrifugation 244 were used for small and large scale plasmid purifications, respectively. Fluorescent Sanger DNA sequencing was performed by the University of Florida ICBR. GC/MS analysis employed a column and ionization by EI at 70 eV. The temperature program involved an initial hold at 95 C for 5 min, an initial increase of 5 C/min to 138 C followed by an increase of 10 C/min to 180 C, then a final increase of 2 C /min to 200 C and a hold at that temperature for 10 min. Cloning of A. jandaei L allo TA. The gene encoding L allo TA from A. jandaei (accession number D87890) was synthesized by GenScript and ligated into a pUC57 with flanking Nde I and Xho I restriction sites at the 5' and 3' ends, respectively. The TA gene was excised by digesting with these restriction enzymes and ligated with Nde I, Xho I cut pET15b (Novagen). After transformation into E. coli ElectroTen Blue, plasmid

PAGE 201

201 DNA from a randomly chosen colony was isolated, restriction mapped and then sequenced to verify the desired structure. The resulting plasmid (designated pSF3) was used to transform E. coli BL21 Gold(DE3) strain for protein overexpression. Cloning of P. putida L TA The L TA gene from P. putida (accession number AP013070) was isolated and amplified from a P. putida strain purchased f rom Carolina Biological Company by colony PCR 229 CGTTCACAGGACCGT CATATG ACA GATAAGAGCCAACAATTCGCC CTGGCTTGCCGGCGATTGG GGATCC TCAGGCGGT GATGATGCTGCGGATA respectivel y. These primers also introduced flanking Nde I and Bam HI restriction sites (underlined). After purification, the PCR product was digested sequentially with Nde I and Bam HI, then ligated with Nde I, Bam HI digested pET 15b. After transformation into E. coli ElectroTen Blue, plasmid DNA from a randomly chosen colony was isolated, restriction mapped and then sequenced to verify the desired structure. The resulting plasmid (designated pSF6) was used to transform E. coli BL21 Gold(DE3) strain for protein overex pression. Cloning of B. subtilis g lycine oxidase. The complete coding sequence for B. subtilis glycine oxidase (accession number NC000964) was synthesized by GenScript and ligated into pUC57 with flanking Nde I and Bam HI restriction sites at the 5' and 3' ends, respectively Silent mutations were introduced to the coding region to remove internal Nde I and Bam HI sites that occur in the native sequence. The gene was subcloned as an Nde I, Bam HI fragment between these sites into pET 15b. After tran s forming E. coli ElectroTen Blue, plasmid DNA was isolated from a randomly chosen colony, restriction mapped and sequenced to verify that the desired plasmid had

PAGE 202

202 been prepared (designated pSF9). This was used to transform E. coli BL21 Gold(DE3) for protein overexpres sion. Mutagenesis of A. jandaei L allo TA Each mutation was accomplished by PCR with individual primer sets using pSF3 as the template DNA and the corresponding forward and reverse primers for each mutation These primer sets introduced a specific mutat ion at a specific site, using the most frequent codon usage for E. coli (underlined in Table 6 2). After purification, the PCR product was digested with Dpn I to cut the template DNA (pSF3) After transformation into E. coli ElectroTen Blue, plasmid DNA f rom a randomly chosen colony was isolated and sequenced to verify the desired mutation The resulting plasmid with mutation was used to transform E. coli BL21 Gold(DE3) strain for protein overexpression. Table 6 2. Forward and reverse primer sets for mut agenesis Mutation of pSF3 Forward Primer Reverse Primer Y89P TATCGC CCG GAGGCG CAGGGTT CGCCTC CGG GCGATA GATGTGGGCT Y89D TATCGC GAT GAGGCG CAGGGTTCTG CGCCTC ATC GCGATA GATGTGGGCTGC H128N GCCCCTGACGATGTC AAC TTTACCCCGACTCGCCTC GCGAGTCGGGGTAAA GTT GACATCGTCAGGGGCG Note: Mutation is underlined in each primer. Protein overexpression. A single colony of the appropriate strain was used to picillin. After shaking overnight at 37 C, a 40 mL portion of the preculture was added to 4 L of LB medium Antifoam 204 in a New Brunswick M19 fermenter. The culture was g rown at 37 C with stirring at 400 rpm and an air flow of 1 vvm until the O.D.600 reached 0.5 0.6. Protein overexpression was induced by adding 10 mL of 0.16 M IPTG (to yield a final concentration of 0.4 mM) and adjusting the temperature to 30 C and sh aking. After 3

PAGE 203

203 hr, the cells were harvested by centrifuging at 6,300 g for 15 min at 4 C, resuspended in 50 mM KP i pH 8.0 (1 mL buffer per gram wcw), then lysed by a French pressure cell at 17,000 psi. Insoluble debris was pelleted by centrifuging a t 39,000 g for 1 hr at 4 C and the yellow supernatant was used for TA catalyzed reactions. Affinity p urification of B. subtilis g lycine o xidase A crude extract containing glucose oxidase was prepared as described above, then the sample was applied to a 5 mL HiTrap Chelating HP column (GE Healthcare Life Sciences) that had been equilibrated with binding buffer (20 M NaP i 500 mM NaCl, 20 mM imidazole, pH 7.4). After washing with 50 mL of binding buffer, the desired protein was eluted by elution buffer (20 mM NaPi, 500 mM NaCl, 500 mM imidazole, pH 7.4). A flow rate of 2 mL/min was employed throughout. The eluate was concentrated by ultrafiltration (Amicon Ultra), then diluted with 50 mM KPi, pH 8.0 and re concentrated. This was repeated two more time s. The final glycine oxidase sample was diluted with the same buffer to 5 mg / mL, then glycerol was added to a final concentration of 10% and the protein was stored in aliquots at 80 C Enzyme a ssays for t hreonine a ldolase. The activity of L TAs was m easured solution in 50 mM KPi, pH 8.0 (total volume of 1 mL). The mixture was gently rotated at and overnight. Samples were derivatized with MSTFA and analyzed by GC/MS. The temperature program involved an initial hold at 95 C for 5 min, an initial increase of 5 C/min to 1 20 C followed by an increase of 2 C/min to 1 38 C, then a final incr ease of 10 C/min to 200 C and a hold at that temperature for 5 min

PAGE 204

204 Enzyme assay for glycine oxidase. Glycine oxidase activity was measured by monitoring the formation of H 2 O 2 by UV Vis spectroscopy at 500 nm using a coupled HRP assay 228 Reaction mixtures containing 10 mM glycine and 0.25 mg glycine aminoantipyrine, 2 in cubating at 37 C for 10 min, the A 500 value was used to calculate units glycine oxidase activity. 228 Derivatization of amino acids by MSTFA. Reaction mixtures were dried completely under reduced pressure for 30 min by a Savant SpeedVac SVC100, then the residue was taken u C for 30 min, the mixtures were analyzed by GC/MS. General procedure for screening fluorinated acetone acceptors. Reactions contained 0.1 mmol of fluorinated acetone, 1 mmol of glycine, 10 nmol of PLP and 10 overnight at room temp erature and sampled after 4 hr and overnight for MSTFA derivatization and GC/MS analysis. Preparative s h ydroxy a mino a cids Reaction mixtures lysate in 20 mL 50 mM Tris, pH 9.7 and 2% (v/v) ethanol. After gently rotating at 37 C (if using A. jandaei L allo TA) or 25 C (if using P. putida L TA) for 20 hr, the unreacted aldehyde was removed by extraction with Et 2 O (2 30 mL) and the aqueous layer was lyophilized. The solid residue was mixed thoroughly with 30 mL of MeOH and the solution was filtered and evaporated under reduced pressure. Th is procedure was

PAGE 205

205 repeated to remove additional unreacted glycine. The crude product was resuspended in 10 mL of 50 mM KP i pH 8.0, then 1 mg of purified glycine oxidase was added and the mixture was gently rotated at 37 C for 8 hr. An additional of 1 mg portion of purified glycine oxidase was then added and incubation at 37 C was continued for an additional 20 hr. The reaction mixture was lyophilized. The residue was stirred with MeOH, leaving most phosphate undissolved. After evaporating the solvent the crude product was dissolved in water and applied to an 11 1.5 cm DOWEX, 1 2 (HO form) column. The column was washed with 100 mL of deionized water, then the desired product was eluted by washing with 50 mL of 0.5% acetic acid. The solvent was removed using a SpeedVac to afford the final product. 2 A mino 4,4 difluoro 3 hydroxy 3 methylbutanoic acid ( synthesized by P. putida L TA ) White solid 30 mg, 8.8 % yield, 0 % d.e 1 H NMR (300 MHz, D 2 6.34 5.79 ( 1 H, q ), 3.88 3.78 (1H, d ), 1.48 1 .27 ( 3H, d) ppm 1 9 F NMR (300 MHz, D 2 O) 129.49 132.24 ( d q ), 132.39 136.37 ( d q ) ppm (Figure A 11 ). 2 A mino 4,4 difluoro 3 hydroxy 3 methylbutanoic acid (synthesized by A. jandaei L allo TA). White solid 12.9 mg, 7.2 % yield, 37 % d.e 1 H NMR (30 0 MHz, D 2 6.34 5.79 ( 1 H, q ), 3.88 3.78 (1H, d ), 1.48 1.28 ( 3H, d) ppm. 1 9 F NMR (300 MHz, D 2 129.50 132.22 ( dq ), 132.38 136.37 ( dq ) ppm (Figure A 1 2 ). 2 Amino 4 fluoro 3 hydroxy 3 methylbutanoic acid (synthesized by A. jandaei L allo TA). White solid 13 mg, 8 % yield, 6 % d.e 1 H NMR (300 MHz, D 2 4.63 4.33 ( 2 H, m ), 3.86 3.72 (1H, d ), 1.40 1.16 ( 3H, d) ppm. 1 9 F NMR (300 MHz, D 2 3.83 3.53 ( t ), 2.49 2.15 ( t ) ppm (Figure A 13 ).

PAGE 206

206 Figure 6 methylfluoro threonine analogues by threonine aldolase

PAGE 207

207 CHAPTER 7 CONCLUSIONS AND FUTURE WORK Wild Type L Threonine Aldolase The L TAs from A. jandaei E. coli T. maritima and P. putida were success fully isolated and screened against a variety of aldehyde and ketone acceptors using glycine as the amino donor Among these L TAs, the L allo TA from A. jandaei proved superior hydroxy amino acids with moderate conversions and diaster eoselectivities Since a simple i solation and purification protocol for the products had not been established prior to our work one was devised and s ix aldehyde acceptors were scaled up moderately T he aldol products were isolated from the reaction mixt ures using glycine oxidase to simplify downstream processing by removing excess glycine from the reaction mixture Mutant L Threonine Aldolase One of the major limitations of these enzymes was the reversibility of the reactions and the unfavorable thermody namic equilibrium Because of this reversibility, diastereoselectivity degrade d over time resulting in a mixture of diastereomers. It had been shown previously that shortened reaction times would prevent this erosion of diastereoselectivity; however, in most cases fractional conversion had to be significantly sacrificed. In order for TAs to be synthetically relevant in an industrial setting, this issue had to be address ed. T herefore, site saturation mutagenesis was carried out on the L allo TA from A jandaei at three active site residues, His 128, His 85, and Tyr 89. Each mutant was initially screened using glycine and three aldehyde acceptors (one aliphatic aldehyde and two aromatic aldehydes) in order to minimize screening efforts while still sket ching the broad picture of the substrate selectivity of the

PAGE 208

208 mutants. His 128 and Tyr 89 mutants gave increased diastereoselectivity for some aldehyde acceptors; however, mutations at the 85 position rendered the enzyme completely inactive. In summary, we can definitively say that by targeting active site residues and using SSM strategies the diastereoselectivity problem for at least some of the substrates was solved and specific results are summarized in the following paragraphs. Histidine 128 Mutants H is tidine 128 was selected for mutagenesis because it hydrogen bonds with the hydroxyl group of L allo Thr. Qin et al. revealed some m utations at this position increased activity (8.4 fold) for the retro aldol cleavage of L Thr ; however, the aldol condensa tion of glycine with aldehyde acceptors was never tested with these variants. 207 In this study, the His 128 mutants that displayed the best overall results in the initial screening were the Ile and Asn var iants. More extensive screening of these variants was carried out on a variety of aliphatic aldehydes, i.e straight chain, cyclic, branched, and unsaturated aldehydes and a few aromatic aldehydes. The Ile mutant showed >90% d.e. across all aliphatic ald ehydes screened although conversions were uniformly low (<20%). On the other hand, the Asn mutant gave comparable diastereoselectivities (>85%) along with reasonable conversions for some substrates Tyrosine 89 Mutants Tyr osine 89 was targeted because it is located in the hydrophobic pocket where the methyl group of L allo Thr rests when it is bound as an external aldimine with PLP. The initial screening of these mutations revealed not only increased diastereoselectivity with some variants (Gly, Pro, Asp Glu, and Trp), but also reversed diastereoselectivity (Ala, Pro, Gln and Lys). The best variants in this library w ere the Pro (Y89P) and Asp

PAGE 209

209 (Y89D) mutants The Y89P mutant increased the wild type diastereoselectivity (to 81% d.e.) with benzaldehyde 9 but gave almost complete reversed selectivity for 3 pyridinecarboxaldehyde 1 1 (to 91% d.e.) despite the very similar sizes of the aldehydes Y89D revealed reversed diastereoselectivity for n on cycl ic, branched aldehyde acceptors. For example, the overa ll conversion of p ivaldehy de 28 was increased seven fold and completely reversed the diastereoselectivity (99% d.e.) with this single mutation Isobutryaldehyde 31 presented similar results: a 2.7 fold increase in conversion (73%) as well as 99% d.e. for the opposite diastereomer. Surprisingly, cyclic aliphatic aldehydes did not show reversed diastereoselectivity. In conclusion, the diastereoselectivity problem for at least some of the substrates was solved. Structure of L Threonine Aldolases The L TA fr om P putida was studied for its broad s ubstrate range in producing hydroxy amino acids Although other L TAs displayed tolerance in the synthesis of a variety amino acids the diastereoselectivity of P. putida L TA was found to be rather poor compare d to the other L TAs ; however in most cases, conversions were generally high for this of enzyme We hoped to identify the problem by determining the crystal structure of this enzyme. After careful optimization of crystallogenesis conditions, the structu re of P. putida L TA was determined at a 2.27 resolution. The active site lysine was determined to be Lys 207 and other highly conserved amino acid residues included Ser 10, His 89, His 133, Arg 177, and Arg 321 A c omparison with the other L TA struc tures revealed two main differences. First, the loop located near the active site was slightly longer in P. putida L TA than in the others ( by approximately 4 7 amino acid s ) This was thought to distort the folding helix as the elect ron density was poor. Second, the amino acid

PAGE 210

210 residue at position 93 differed from those of other L TA s In P. putida L TA, it was Asp; in other L TAs, the residue at the analogous position was a large aromatic residue (Tyr or Phe). The poor diastereosel ectivity of P. putida L TA might therefore be due to the extended loop the acidic residue at position 93, or to other, more subtle factors Interestingly, our m utagenesis studies of the equivalent position (93) in A. jandaei L allo TA revealed that an As p at this position actually increased diastereoselectivity in that enzyme It is therefore likely that the poor diastereoselectivity observed for P. putida L TA was due to additional factors beyond the nature of the residue at position 93. Future Work Dou ble Mutations The double mutations H128N/Y89P and H128N/Y89D of L allo TA from A. jandaei should be isolated, overexpressed in E. coli and screened against aldehyde and ketone acceptors. Since both single mutations had an augmented effect on diastereosele ctivity, the double mutants might provide even higher stereoselectivities If the screening yields desirable outcomes scale d up reactions should be preformed followed by isolation of aldol products. Additional Mutations One goal of this project was to in crease the diastereoselectivity of A. jandaei L allo TA by manipulating the residues that interact with the hydroxyl and side chain of the product The next step in this project would be to mutate the amino acid residues that interact with the carboxylate group of the starting material and product to allow for different amino donors with interesting functional groups ( i.e. sulfonate and phosphonate). The active site of L TAs contain five conserved residues. These include

PAGE 211

211 Ser 10, Arg 177, and Arg 321 whi ch hydrogen bond with the carboxyl group of Gly and L Thr The most probable solution to help stabilize the extra charge from a phosphonate or sulfonate amino donor would be to mutate the Ser 10 to an Arg or Lys variant. This would produce a more positiv ely charged environment that might allow for electrostatic interactions between the negatively charge phosphonate/sulfonate and the guanidinium group of Arg or the protonated amine of Lys. Structure of Mutant L Threonine Aldolase H128N, Y89P and Y89D varia nts of A. jandaei L allo TA have proven useful at solving the diastereoselectivity problem for some aldehyde acceptors and in some cases, revers ing the diastereoselectivity completely C rystallograph ic studies on these mutations should be carried out to help understand how these changes affected the active site Since Qin et al. solved the crystal structure of both the wild type and the H128Y/S292R variant, these efforts should be fairly straight forward M olecular replacement with the native L allo TA should allow for simple structure determination.

PAGE 212

212 APPENDIX SUPPORTING INFORMATION Table A 1. Assignment of 1 H and 13 C chemical shifts in diastereomers M and m in D 2 O (or methanol Compound Position M m M m M m 1 C 171.6 172.8 172.1 170.7 171.8 170.9 2 C 57.4 57.1 56.0 56.9 60.7 6 0.3 2 H 3.75 3.55 3.65 3.72 3.84 4.00 3 C 69.4 69.6 3.6 74.5 71.2 71.1 3 H 4.00 3.97 3.78 3.57 5.21 5.26 4 C 30.5 32.8 40.1 39.7 139.0 136.8 4 H 1.4 0 1.49 1.47 nm 5 C 27.4 27.1 nm nm 125.8 126.3 5 H 1.37, 1.25 1.35, 1.25 nm nm 7.37 7.30 6 C 21.6 21.6 nm nm 129.0 129.0 6 H 1.25 1.25 nm nm 7.37 7.35 7 C 13.1 13.2 nm nm 128.6 128.6 7 H 0.78 0.79 nm nm 7.30 7.31 Compound Position M m M m M m 1 C 169.0 nm 171.2 169.9 171.3 170.0 2 C 56.7 nm 60.0 59.6 57.8 56.9 2 H 4.25 4.01 3.88 4.08 4.01 4.02 3 C 65.8 nm 70.0 69.9 67.6 68.2 3 H 5.61 5.44 5.29 5.33 5.47 5.29 4 C 126.9 nm 153.2 153.6 134.1 134.3 5 C 156.1 nm 122.6 122.6 148.0 147.1 5 H nm nm 7.60 7.60 6 C 110.2 nm 146.7 145.4 6 H 7.03 6.98 8.50 8.49 7 C 129.4 nm 148.8 148.6 7 H 7.36 7.31 8.20 8.17 8 C 120.5 nm 123.8 123.3 8 H 7.04 6.99 7.39 7.33 9 C 126.6 nm 138.0 138.6 9 H 7.59 7.52 8.00 7.90 10 C 54.5 n m 10 H 3.90 3.85

PAGE 213

213 Table A 2. R S sign resulting from disubstitution with MPA 2 amino 3 hydroxy acid 2 R 3 R 2 R 3 S 2 S 3 R 2 S 3 S H2 0 0+ 0 0+ H3 0 0 +0 +0 R + -++ + Table A 3. Chemic al shifts and R S sign in double MPA derivatives of the n butyl and cyclohexyl compounds R1 CD R (ppm) H3 (ppm) H2 (ppm) n Butyl (OH) (M) R,R 0.68 5.14 4.81 anti S,S 0.80 5.02 4.73 2 S, 3 S RS 0.18 0.12 0.08 n Bu tyl (m) R,R 0.86 5.40 4.71 syn S,S 0.56 5.32 4.77 2 S, 3 R RS 0.30 0.08 0.06 Cyclohexyl (OMe) (M) R,R 1.56 5.23 4.86 syn S,S 1.10 5.11 4.90 2 S, 3 R RS 0.46 0.12 0.04 Cyclohexyl (m) R,R 1.63 4.97 4.89 anti S,S 1.50 5.11 4.87 2 R, 3 R RS 0.13 0. 14 0.02 a a Of the two possible configurations with RS at H2<0, the one which allows for a RS at R>0 was chosen, because of the absolute values of RS

PAGE 214

214 Table A 4. Chemical shifts in the methyl ester of 3 phenyl 2 amino 3 hydroxypropanoic aci ds and its amides with MPA. R MPA H5 (ppm) MeO1 (ppm) Ph (M) N/A 7.78 3.60 syn R 7.78 3.54 2 S 3 S S 7.70 3.65 Ph (m) N/A 7.91 3.67 anti R 7.79 3.63 2 R 3 S S nm nm a a For the minor, the shielding of H5 in the R MPA was compared to the one in t he amine. ( a ) Figure A 1. Mass spectrum of (4 S ,5 R ) 2 amino 3,4,5,6 tetrahydroxyhexanoic acid

PAGE 215

215 ( b ) Figure A 1. Continued

PAGE 216

216 ( a ) Figure A 2 NMR spectra of (2 S ,3 S ) 2 amino 3 hydroxyheptanoic acid

PAGE 217

217 ( b ) Figure A 2 Continued

PAGE 218

218 ( c ) Figure A 2 C ontinued

PAGE 219

219 ( d ) Figure A 2 Continued

PAGE 220

220 ( e ) Figure A 2 Continued

PAGE 221

221 ( f ) Figure A 2 Continued

PAGE 222

222 ( g ) Figure A 2 Continued

PAGE 223

223 ( a ) Figure A 3 NMR spectra of (2 S ,3 R ) 2 amino 3 hydroxycyclohexanepropanoic acid

PAGE 224

224 ( b ) Figure A 3 Continued

PAGE 225

225 ( c ) Figure A 3 Continued

PAGE 226

226 ( d ) Figure A 3 Continued

PAGE 227

227 ( a ) Figure A 4 NMR spectra of (2 S ,3 S ) 3 hydroxyphenylalanine

PAGE 228

228 ( b ) Figure A 4 Continued

PAGE 229

229 ( c ) Figure A 4 Continued

PAGE 230

230 ( d ) Figure A 4 Continued

PAGE 231

231 ( e ) Figure A 4 Continued

PAGE 232

232 ( a ) Figure A 5 NMR spect ra of 2 amino 3 hy droxy 4 pyridinepropanoic acid

PAGE 233

233 ( b ) Figure A 5 Continued

PAGE 234

234 ( c ) Figure A 5 Continued

PAGE 235

235 ( d ) Figure A 5 Continued

PAGE 236

236 ( e ) Figure A 5 Continued

PAGE 237

237 ( f ) Figure A 5 Continued

PAGE 238

238 ( g ) Figure A 5 Continued

PAGE 239

239 ( a ) Figure A 6 NMR spectr a of 3 hydroxy 2 methoxy phenylalanine

PAGE 240

240 ( b ) Figure A 6 Continued

PAGE 241

241 ( c ) Figure A 6 Continued

PAGE 242

242 ( d ) Figure A 6 Continued

PAGE 243

243 ( a ) Figure A 7 NMR spectra of 2 amino 3 (2 chloro 3 pyr idine) 3 hydroxypropanoic acid

PAGE 244

2 44 ( b ) Figure A 7 Continued

PAGE 245

245 ( c ) Figure A 7 Continued

PAGE 246

246 ( d ) Figure A 7 Continued

PAGE 247

247 ( e ) Figure A 7 Continued

PAGE 248

248 ( a ) Figure A 8. 1 H NMR for thermodynamic reversibility of L allo threonine aldolase. Negative control at ( a ) zero hours and ( b ) 24 hours L allo Thr and acetaldehyde d4 i n the presence of L allo TA at ( c ) zero hours, ( d ) 1 hour, ( e ) 3 hours, ( f ) 6 hours and ( g ) 24 hours

PAGE 249

249 ( b ) ( c ) Figure A 8. Continued

PAGE 250

250 ( d ) ( e ) Figure A 8. Continued

PAGE 251

251 ( f ) Figure A 8. Continued

PAGE 252

252 ( g ) Figure A 8. Continued Approx. 1 : 2

PAGE 253

253 ( a ) Figure A 9. MS for thermodynamic reversibility of L allo threonine aldolase. Negative control at ( a ) 1 hour and ( b ) 24 hours. L allo Thr and acetaldehyde d4 in the presence of L allo TA at ( c ) 1 hour, ( d ) 3 hour s ( e ) 5 hours and ( f ) 24 hours.

PAGE 254

254 ( b ) ( c ) Figure A 9. Continued

PAGE 255

255 ( d ) ( e ) Figure A 9. Continued

PAGE 256

256 ( f ) Figure A 9. Continued Approx. 1 : 1

PAGE 257

257 ( a ) Figure A 10 Mass spectrum of 2 amino 4,4 difluoro 3 hydroxy 3 methylbutanoic acid

PAGE 258

258 ( b ) Figure A 10 Continued

PAGE 259

259 ( a ) Figure A 11 Proton and fluorine NMR spectra o f 2 amino 4,4 difluoro 3 hydroxy 3 methylbutanoic acid synthesized by P. putida L TA

PAGE 260

260 ( b ) Figure A 11 Continued

PAGE 261

261 ( a ) Figure A 1 2 Proton and fluorine NMR spectra of 2 amino 4,4 difluoro 3 hydroxy 3 methylbutanoic acid synthesized by A. jandaei L allo TA

PAGE 262

262 ( b ) Figure A 1 2 Continued

PAGE 263

263 ( a ) Figure A 1 3 Proton and fluorine NMR spectra of 2 a mino 4 fluoro 3 hydroxy 3 methylbutanoic acid synthesized by A. jandaei L allo TA.

PAGE 264

264 ( b ) Figure A 1 3 Continued

PAGE 265

265 LIST OF REFERENCES 1. Yim, G.; Thaker, M. N.; Koteva, K.; Wright, G., Journal of Antibiotics 2014, 67 (1), 31 41. 2. Suzuki, S.; Isono, K.; Nagatsu, J.; Mizutani, T.; Kawashim.Y; Mizuno, T., Journal of Antibiotics 1965, 18 (3), 131. 3. Miyake, Y.; Kozutsumi, Y.; Naka mura, S.; Fujita, T.; Kawasaki, T., Biochemical and Biophysical Research Communications 1995, 211 (2), 396 403. 4. Kang, S. H.; Kang, S. Y.; Lee, H. S.; Buglass, A. J., Chemical Reviews 2005, 105 (12), 4537 4558. 5. Ohfune, Y.; Shinada, T., European Jour nal of Organic Chemistry 2005, 2005 (24), 5127 5143. 6. Pirrung, M. C.; Nunn, D. S.; McPhail, A. T.; Mitchell, R. E., Bioorganic and Medicinal Chemistry Letters 1993, 3 (10), 2095 2098. 7. Liu, J. Q.; Odani, M.; Yasuoka, T.; Dairi, T.; Itoh, N.; Kataoka, M.; Shimizu, S.; Yamada, H., Applied Microbiology and Biotechnology 2000, 54 (1), 44 51. 8. Baik, S. H.; Yoshioka, H.; Yukawa, H.; Harayama, S., Journal of Microbiology and Biotechnology 2007, 17 (5), 721 727. 9. Wenger, R. M., Helvetica Chimica Acta 19 83, 66 (7), 2308 2321. 10. Kaufmann, H.; Saadia, D.; Voustianiouk, A.; Goldstein, D. S.; Holmes, C.; Yahr, M. D.; Nardin, R.; Freeman, R., Circulation 2003, 108 (6), 724 728. 11. Mitsubishi; Co., G. C. Glucosaminic Acid Production by Aeromonas. 1982. 12 Gu, W. X.; Xia, W. S., Journal of Carbohydrate Chemistry 2006, 25 (4), 297 301. 13. Varela, O.; Nin, A. P.; Delederkremer, R. M., Tetrahedron Letters 1994, 35 (50), 9359 9362. 14. Evans, D. A.; Weber, A. E., Journal of the American Chemical Society 198 6, 108 (21), 6757 6761. 15. Alonso, M.; Riera, A., Tetrahedron Asymmetry 2005, 16 (23), 3908 3912. 16. Nagamitsu, T.; Sunazuka, T.; Tanaka, H.; Omura, S.; Sprengeler, P. A.; Smith, A. B., Journal of the American Chemical Society 1996, 118 (15), 3584 3590

PAGE 266

266 17. Kim, I. H.; Kirk, K. L., Tetrahedron Letters 2001, 42 (48), 8401 8403. 18. Fanning, K. N.; Jamieson, A. G.; Sutherland, A., Organic and Biomolecular Chemistry 2005, 3 (20), 3749 3756. 19. Willis, M. C.; Piccio, V. J. D., Synlett 2002, 2002 (10), 1 625 1628. 20. Willis, M. C.; Cutting, G. A.; Piccio, V. J. D.; Durbin, M. J.; John, M. P., Angewandte Chemie International Edition 2005, 44 (10), 1543 1545. 21. Thayumanavan, R.; Tanaka, F.; Barbas, C. F., Organic Letters 2004, 6 (20), 3541 3544. 22. Be lokon, Y. N.; Bulychev, A. G.; Vitt, S. V.; Struchkov, Y. T.; Batsanov, A. S.; Timofeeva, T. V.; Tsyryapkin, V. A.; Ryzhov, M. G.; Lysova, L. A., Journal of the American Chemical Society 1985, 107 (14), 4252 4259. 23. Genet, J. P.; Pinel, C.; Mallart, S.; Juge, S.; Thorimbert, S.; Laffitte, J. A., Tetrahedron Asymmetry 1991, 2 (7), 555 567. 24. Mordant, C.; Dunkelmann, P.; Ratovelomanana Vidal, V.; Genet, J. P., European Journal of Organic Chemistry 2004, (14), 3017 3026. 25. Makino, K.; Goto, T.; Hirok i, Y.; Hamada, Y., Angewandte Chemie International Edition 2004, 43 (7), 882 884. 26. Makino, K.; Hiroki, Y.; Hamada, Y., Journal of the American Chemical Society 2005, 127 (16), 5784 5785. 27. Makino, K.; Fujii, T.; Hamada, Y., Tetrahedron Asymmetry 200 6, 17 (4), 481 485. 28. Noyori, R.; Ikeda, T.; Ohkuma, T.; Widhalm, M.; Kitamura, M.; Takaya, H.; Akutagawa, S.; Sayo, N.; Saito, T.; Taketomi, T.; Kumobayashi, H., Journal of the American Chemical Society 1989, 111 (25), 9134 9135. 29. Gent, J. P.; Pin el, C.; Mallart, S.; Juge, S.; Thorimbert, S.; Laffitte, J. A., Tetrahedron Asymmetry 1991, 2 (7), 555 567. 30. Ratovelomanana Vidal, V.; Gent, J. P., Canadian Journal of Chemistry 2000, 78 846 851. 31. Li, G. G.; Chang, H. T.; Sharpless, K. B., Ange wandte Chemie International Edition in English 1996, 35 (4), 451 454. 32. Loncaric, C.; Wulff, W. D., Organic Letters 2001, 3 (23), 3675 3678.

PAGE 267

267 33. Wang, Y. H.; Chen, Z. Y.; Mi, A. Q.; Hu, W. H., Chemical Communications 2004, (21), 2486 2487. 34. Mettath S.; Srikanth, G. S. C.; Dangerfield, B. S.; Castle, S. L., The Journal of Organic Chemistry 2004, 69 (19), 6489 6492. 35. Ooi, T.; Taniguchi, M.; Doda, K.; Maruoka, K., Advanced Synthesis and Catalysis 2004, 346 (9 10), 1073 1076. 36. Li, L.; Klauber, E. G.; Seidel, D., Journal of the American Chemical Society 2008, 130 (37), 12248 37. Chen, X. H.; Zhu, Y.; Qiao, Z.; Xie, M. S.; Lin, L. L.; Liu, X. H.; Feng, X. M., Chemistry a European Journal 2010, 16 (33), 10124 10129. 38. Chelucci, G.; Murineddu, G.; Pinna, G. A., Tetrahedron: Asymmetry 2004, 15 (9), 1373 1389. 39. European Journal of Organic Chemistry 2007, 2007 (1), 29 36. 40. Yu, Z.; Liu, X.; Dong, Z.; Xie, M.; Feng, X., Angewandte Chemie 2008, 120 (7), 1328 1331. 41. Shang, D.; Xin, J.; Liu, Y.; Zhou, X.; Liu, X.; Feng, X., The Journal of Organic Chemistry 2008, 73 (2), 630 637. 42. Itoh, K.; Kanemasa, S., Journal of the American Chemical Society 2002, 124 (45), 13394 13395. 43. Itoh, K.; Hasegawa, M.; Tanaka, J .; Kanemasa, S., Organic Letters 2005, 7 (6), 979 981. 44. Evans, D. A.; Seidel, D., Journal of the American Chemical Society 2005, 127 (28), 9958 9959. 45. Evans, D. A.; Thomson, R. J.; Franco, F., Journal of the American Chemical Society 2005, 127 (31) 10816 10817. 46. Heine, H. W., Angewandte Chemie International Edition in English 1962, 1 (10), 528 532. 47. Padwa, A.; Woolhouse, A. D., 5.04 Aziridines, Azirines and Fused ring Derivatives A2 Katritzky, Alan R. In Comprehensive Heterocyclic Chemi stry Rees, C. W., Ed. Pergamon: Oxford, 1984; pp 47 93.

PAGE 268

268 48. Tanner, D., Angewandte Chemie International Edition in English 1994, 33 (6), 599 619. 49. Pellissier, H., Tetrahedron 2010, 66 (8), 1509 1555. 50. Zhang, Y.; Lu, Z.; Wulff, W. D., Synlett 2009, 2009 (17), 2715 2739. 51. Gabriel, S.; Stelzner, R., Berichte der Deutschen Chemischen Gesellschaft 1895, 28 52. Huang, L.; Wulff, W. D., Journal of the American Chemical Society 2011, 133 (23), 8892 8895. 53. Hashimoto, T.; Nakatsu, H.; Yamamoto, K.; Watanabe, S.; Maruoka, K., Chemistry an Asian Journal 2011, 6 (2), 607 613. 54. Danielsson, J.; Toom, L.; Somfai, P., European Journal of Organic Chemistry 2011, 2011 (3), 607 613. 55. Cativiela, C.; Daz de Villegas, M. D., Tetrahedron: Asymmetry 2007, 18 (5), 569 623. 56. Nakatsuka, T.; Miwa, T.; Mukaiyama, T., Chemistry Letters 1981, 10 (2), 279 282. 57. Mukaiyama, T.; Miwa, T.; Nakatsuka, T., Chemistry Letters 1982, (2), 145 148. 58. Mukaiyama, T.; Hayashi, H.; Miwa, T.; Narasaka, K., Chemistry L etters 1982, (10), 1637 1640. 59. Mukaiyama, T., Pure and Applied Chemistry 1982, 54 (12), 2455 2464. 60. Evans, D. A.; Bartroli, J.; Shih, T. L., Journal of the American Chemical Society 1981, 103 (8), 2127 2129. 61. Volkmann, R. A.; Davis, J. T.; Mel tz, C. N., Journal of the American Chemical Society 1983, 105 (18), 5946 5948. 62. Rich, D. H.; Sun, C. Q.; Guillaume, D.; Dunlap, B.; Evans, D. A.; Weber, A. E., Journal of Medicinal Chemistry 1989, 32 (8), 1982 1987. 63. Tung, R. D.; Rich, D. H., Tetra hedron Letters 1987, 28 (11), 1139 1142. 64. Schmidt, U.; Siegel, W., Tetrahedron Letters 1987, 28 (25), 2849 2852.

PAGE 269

269 65. Aebi, J. D.; Dhaon, M. K.; Rich, D. H., Journal of Organic Chemistry 1987, 52 (13), 2881 2886. 66. El Achqar, A.; Boumzebra, M.; Roum estant, M. L.; Viallefont, P., Tetrahedron 1988, 44 (17), 5319 5332. 67. Evans, D. A.; Britton, T. C.; Ellman, J. A.; Dorow, R. L., Journal of the American Chemical Society 1990, 112 (10), 4011 4030. 68. Caddick, S.; Parr, N. J.; Pritchard, M. C., Tetrah edron 2001, 57 (30), 6615 6626. 69. Evans, D. A.; Chapman, K. T.; Bisaha, J., Journal of the American Chemical Society 1988, 110 (4), 1238 1256. 70. Evans, D. A.; Weber, A. E., Journal of the American Chemical Society 1987, 109 (23), 7151 7157. 71. Cadd ick, S.; Parr, N. J.; Pritchard, M. C., Tetrahedron Letters 2000, 41 (31), 5963 5966. 72. Caddick, S.; Jenkins, K., Chemical Society Reviews 1996, 25 (6), 447 73. Caddick, S.; Jenkins, K.; Treweeke, N.; Candeias, S. X.; Afonso, C. A. M., Tetrahedron Let ters 1998, 39 (15), 2203 2206. 74. Iwanowicz, E. J.; Blomgren, P.; Cheng, P. T. W.; Smith, K.; Lau, W. F.; Pan, Y. Y.; Gu, H. H.; Malley, M. F.; Gougoutas, J. Z., Synlett 1998, 1998 (06), 664 666. 75. Overman, L. E., Journal of the American Chemical Soci ety 1974, 96 (2), 597 599. 76. Overman, L. E., Journal of the American Chemical Society 1976, 98 (10), 2901 2910. 77. Overman, L. E., Accounts of Chemical Research 1980, 13 (7), 218 224. 78. Ikariya, T.; Ishikawa, Y.; Hirai, K.; Yoshikawa, S., Chemistry Letters 1982, (11), 1815 1818. 79. Overman, L. E., Angewandte Chemie International Edition in English 1984, 23 (8), 579 586. 80. Overman, L. E.; Knoll, F. M., Tetrahedron Letters 1979, (4), 321 324. 81. Fanning, K. N.; Jamieson, A. G.; Sutherland, A. Organic and Biomolecular Chemistry 2005, 3 (20), 3749 3756.

PAGE 270

270 82. Zhang, Y. F.; Farrants, H.; Li, X. C., Chemistry an Asian Journal 2014, 9 (7), 1752 1764. 83. Schenck, T. G.; Bosnich, B., Journal of the American Chemical Society 1985, 107 (7), 2058 2066. 84. Calter, M.; Hollis, T. K.; Overman, L. E.; Ziller, J.; Zipp, G. G., Journal of Organic Chemistry 1997, 62 (5), 1449 1456. 85. Hollis, T. K.; Overman, L. E., Tetrahedron Letters 1997, 38 (51), 8837 8840. 86. Jamieson, A. G.; Sutherland, A., Organic and Biomolecular Chemistry 2006, 4 (15), 2932 2937. 87. Jaunzeme, I.; Jirgensons, A., Synlett 2005, (19), 2984 2986. 88. Jaunzeme, I.; Jirgensons, A., Tetrahedron 2008, 64 (24), 5794 5799. 89. Anderson, C. E.; Overman, L. E., Journal of the American Ch emical Society 2003, 125 (41), 12412 12413. 90. Jamieson, A. G.; Sutherland, A., Organic Letters 2007, 9 (8), 1609 1611. 91. Swift, M. D.; Sutherland, A., Tetrahedron Letters 2007, 48 (22), 3771 3773. 92. Zaed, A. M.; Sutherland, A., Organic and Biomole cular Chemistry 2010, 8 (19), 4394 4399. 93. Zaed, A. M.; Sutherland, A., Organic and Biomolecular Chemistry 2011, 9 (23), 8030 8037. 94. Ahmad, S.; Swift, M. D.; Farrugia, L. J.; Senn, H. M.; Sutherland, A., Organic and Biomolecular Chemistry 2012, 10 ( 19), 3937 3945. 95. Calder, E. D. D.; Zaed, A. M.; Sutherland, A., Journal of Organic Chemistry 2013, 78 (14), 7223 7233. 96. Overman, L. E.; Owen, C. E.; Pavan, M. M.; Richards, C. J., Organic Letters 2003, 5 (11), 1809 1812. 97. Watson, M. P.; Overman L. E.; Bergman, R. G., Journal of the American Chemical Society 2007, 129 (16), 5031 5044. 98. Swift, M. D.; Sutherland, A., Tetrahedron 2008, 64 (40), 9521 9527.

PAGE 271

271 99. Gent, J. P.; Mallart, S.; Jug, S. Improvement in the Manufacture of Chiral Catalyst s Based on Ruthenium and Phosphocomplexes. 1989. 100. Kagan, H. B.; Fiaud, J. C., Kinetic Resolution. In Topics in Stereochemistry Eliel, E. L., Ed. John Wiley and Sons, Inc.: 1988; Vol. 18, p 249. 101. Noyori, R.; Ohkuma, T.; Kitamura, M.; Takaya, H.; Sayo, N.; Kumobayashi, H.; Akutagawa, S., Journal of the American Chemical Society 1987, 109 (19), 5856 5858. 102. Gent, J. P.; Pinel, C.; Ratovelomananavidal, V.; Mallart, S.; Pfister, X.; Deandrade, M. C. C.; Laffitte, J. A., Tetrahedron Asymmetry 1994 5 (4), 665 674. 103. Gent, J. P.; Pinel, C.; Ratovelomananavidal, V.; Mallart, S.; Pfister, X.; Bischoff, L.; Deandrade, M. C. C.; Darses, S.; Galopin, C.; Laffitte, J. A., Tetrahedron Asymmetry 1994, 5 (4), 675 690. 104. Gent, J. P.; Ratovelomananav idal, V.; Deandrade, M. C. C.; Pfister, X.; Guerreiro, P.; Lenoir, J. Y., Tetrahedron Letters 1995, 36 (27), 4801 4804. 105. Pellissier, H., Tetrahedron 2003, 59 (42), 8291 8327. 106. Pellissier, H., Tetrahedron 2008, 64 (8), 1563 1601. 107. Noyori, R.; Tokunaga, M.; Kitamura, M., Bulletin of the Chemical Society of Japan 1995, 68 (1), 36 55. 108. Ward, R. S., Tetrahedron: Asymmetry 1995, 6 (7), 1475 1490. 109. Pellissier, H., Advanced Synthesis and Catalysis 2011, 353 (5), 659 676. 110. Li, P.; Hu, X. Q.; Dong, X. Q.; Zhang, X. M., Molecules 2016, 21 (10). 111. Girard, A.; Greck, C.; Ferroud, D.; Gent, J. P., Tetrahedron Letters 1996, 37 (44), 7967 7970. 112. Nishi, T.; Kitamura, M.; Ohkuma, T.; Noyori, R., Tetrahedron Letters 1988, 29 (48), 632 7 6330. 113. Schmidt, U.; Meyer, R.; Leitenberger, V.; Griesser, H.; Lieberknecht, A., Synthesis Stuttgart 1992, (10), 1025 1030. 114. Phansavath, P.; de Paule, S. D.; Ratovelomanana Vidal, V.; Genet, J. P., European Journal of Organic Chemistry 2000, (23), 3903 3907.

PAGE 272

272 115. Mashima, K.; Matsumura, Y.; Kusano, K.; Kumobayashi, H.; Sayo, N.; Hori, Y.; Ishizaki, T.; Akutagawa, S.; Takaya, H., Journal of the Chemical Society Chemical Communications 1991, (9), 609 610. 116. Coulon, E.; Cristina, M.; de Andr ade, C.; Ratovelomanana Vidal, V.; Gent, J. P., Tetrahedron Letters 1998, 39 (36), 6467 6470. 117. Sun, G. D.; Zhou, Z. H.; Luo, Z. H.; Wang, H. L.; Chen, L.; Xu, Y. B.; Li, S.; Jian, W. L.; Zeng, J. B.; Hu, B. Q.; Han, X. D.; Lin, Y. C.; Wang, Z. Q., Or ganic Letters 2017, 19 (16), 4339 4342. 118. Tone, H.; Buchotte, M.; Mordant, C.; Guittet, E.; Ayad, T.; Ratovelomanana Vidal, V., Organic Letters 2009, 11 (9), 1995 1997. 119. Mohar, B.; Valleix, A.; Desmurs, J. R.; Felemez, M.; Wagner, A.; Mioskowski, C., Chemical Communications 2001, (24), 2572 2573. 120. Mordant, C.; Dunkelmann, P.; Ratovelomanana Vidal, V.; Genet, J. P., Chemical Communications 2004, (11), 1296 1297. 121. de Paule, S. D.; Jeulin, S.; Ratovelomanana Vidal, V.; Genet, J. P.; Champi on, N.; Dellis, P., European Journal of Organic Chemistry 2003, (10), 1931 1941. 122. Labeeuw, O.; Phansavath, P.; Genet, J. P., Tetrahedron Asymmetry 2004, 15 (12), 1899 1908. 123. Makino, K.; Goto, T.; Hiroki, Y.; Hamada, Y., Tetrahedron: Asymmetry 20 08, 19 (24), 2816 2828. 124. Mordant, C.; Reymond, S.; Ratovelomanana Vidal, V.; Genet, J. P., Tetrahedron 2004, 60 (43), 9715 9723. 125. Spector, S.; Sjoerdsm, A.; Udenfrie, S., Journal of Pharmacology and Experimental Therapeutics 1965, 147 (1), 86 1 26. C reveling, C. R.; Daly, J.; Tokuyama, T.; Witkop, B., Biochemical Pharmacology 1968, 17 (1), 65. 127. Bartholini, G.; Constantinidis, J.; Tissot, R.; Pletscher, A., Biochemical Pharmacology 1971, 20 (6), 1243 1247. 128. Porter, C. C.; Torchiana, M. L .; Stone, C. A., Life Sciences 1972, 11 (16, Part 1), 787 795. 129. Bartholini, J.; Constantinidis, J.; Puig, M.; Tissot, R.; Pletscher, A., Journal of Pharmacology and Experimental Therapeutics 1975, 193 (2), 523 532.

PAGE 273

273 130. Inagaki, C.; Fujiwara, H.; Tana ka, C., The Japanese Journal of Pharmacology 1976, 26 (3), 380 382. 131. Chiyoko, I.; Chikako, T., Biochemical Pharmacology 1978, 27 (7), 1081 1086. 132. Belcher, G.; Ryall, R. W.; Schaffner, R., Brain Research 1978, 151 (2), 307 321. 133. Kawabata, A.; Kasamatsu, K.; Umeda, N.; Takagi, H., British Journal of Pharmacology 1994, 111 (2), 503 508. 134. Narabayashi, H.; Nakanishi, T., Clinical Evaluation 1987, 15 423 457. 135. Carvalho, M. J.; van den Meiracker, A. H.; Boomsma, F.; Man in 't Veld, A. J.; Freitas, J.; Costa, O.; Falcao de Freitas, A., Journal of the Autonomic Nervous System 1997, 62 (1 2), 63 71. 136. Mathias, C. J.; Senard, J. M.; Braune, S.; Watson, L.; Aragishi, A.; Keeling, J. E. A.; Taylor, M. D., Clinical Autonomic Research 2001, 11 (4), 235 242. 137. Kaufmann, H., Clinical Autonomic Research 2008, 18 19 24. 138. Knosalla, M.; Weis, J.; Isenmann, S.; Haensch, C. A., Klinische Neurophysiologie 2011, 42 (2), 103 109. 139. Perez Lloret, S.; Rey, M. V.; Traon, A. P. L.; Rascol, O., E xpert Opinion on Emerging Drugs 2013, 18 (1), 39 53. 140. Biaggioni, I.; Freeman, R.; Mathias, C. J.; Low, P.; Hewitt, L. A.; Kaufmann, H.; Behalf Droxidopa, I., Hypertension 2015, 65 (1), 101 141. Isaacson, S.; Shill, H. A.; Vernino, S.; Ziemann, A.; R owse, G. J., Journal of Parkinsons Disease 2016, 6 (4), 751 759. 142. Kalinin, S.; Polak, P. E.; Lin, S. X.; Sakharkar, A. J.; Pandey, S. C.; Feinstein, D. L., Neurobiology of Aging 2012, 33 (8), 1651 1663. 143. Phillips, C.; Fahimi, A.; Das, D.; Mojabi, F. S.; Ponnusamy, R.; Salehi, A., Current Alzheimer Research 2015, 13 (1), 68 83. 144. Heneka, M. T.; Nadrigny, F.; Regen, T.; Martinez Hernandez, A.; Dumitrescu Ozimek, L.; Terwel, D.; Jardanhazi Kurutz, D.; Walter, J.; Kirchhoff, F.; Hanisch, U. K.; Ku mmer, M. P., Proc Natl Acad Sci U S A 2010, 107 (13), 6058 63. 145. Coll, M.; Rodriguez, S.; Raurell, I.; Ezkurdia, N.; Brull, A.; Augustin, S.; Guardia, J.; Esteban, R.; Martell, M.; Genesca, J., Hepatology 2012, 56 (5), 1849 1860.

PAGE 274

274 146. Rodriguez, S.; Ra urell, I.; Ezkurdia, N.; Augustin, S.; Esteban, R.; Genesca, J.; Martell, M., Liver International 2015, 35 (2), 326 334. 147. Wecht, J. M.; Rosado Rivera, D.; Weir, J. P.; Ivan, A.; Yen, C.; Bauman, W. A., Archives of Physical Medicine and Rehabilitation 2013, 94 (10), 2006 2012. 148. Mathias, C. J., Clinical Autonomic Research 2008, 18 25 29. 149. George, S.; Narina, S. V.; Sudalai, A., Tetrahedron Letters 2007, 48 (8), 1375 1378. 150. Herbert, B.; Kim, I. H.; Kirk, K. L., The Journal of Organic Chemi stry 2001, 66 (14), 4892 4897. 151. Hegeds, B.; Krass, A. F.; Noack, K.; Zeller, P., Helvetica Chimica Acta 1975, 58 (1), 147 162. 152. Guan, Y. Q.; Gao, M.; Deng, X.; Lv, H.; Zhang, X. M., Chemical Communications 2017, 53 (58), 8136 8139. 153. Lee, C J.; Liang, X.; Wu, Q.; Najeeb, J.; Zhao, J.; Gopalaswamy, R.; Titecat, M.; Sebbane, F.; Lemaitre, N.; Toone, E. J.; Zhou, P., 2016, 7 10638. 154. Titecat, M.; Liang, X. F.; Lee, C. J.; Charlet, A.; Hocquet, D.; Lambert, T.; Pages, J. M.; Courcol, R.; S ebbane, F.; Toone, E. J.; Zhou, P.; Lemaitre, N., Journal of Antimicrobial Chemotherapy 2016, 71 (10), 2874 2882. 155. Raetz, C. R.; Whitfield, C., Annu Rev Biochem 2002, 71 635 700. 156. Pei Zhou and Adam, W. B., Current Pharmaceutical Biotechnology 20 08, 9 (1), 9 15. 157. Liang, X.; Gopalaswamy, R.; Navas, F.; Toone, E. J.; Zhou, P., The Journal of Organic Chemistry 2016, 81 (10), 4393 4398. 158. Owens, L. D.; Wright, D. A., Plant Physiology 1965, 40 (5), 927 159. Owens, L. D.; Lieberman, M.; Kunis hi, A., Plant Physiology 1970, 46 32 160. Giovanelli, J.; Owens, L. D.; Mudd, S. H., Biochimica Et Biophysica Acta 1971, 227 (3), 671 161. Owens, L. D.; Guggenheim, S.; Hilton, J. L., Biochimica Et Biophysica Acta 1968, 158 (2), 219

PAGE 275

275 162. Keith, D. D .; Debernardo, S.; Weigele, M., Tetrahedron 1975, 31 (20), 2629 2632. 163. Keith, D. D.; Tortora, J. A.; Ineichen, K.; Leimgruber, W., Tetrahedron 1975, 31 (20), 2633 2636. 164. Owens, L. D., Weed Science 1973, 21 (1), 63 66. 165. Wang, C. Y.; Mellenthi n, W. M., Plant Physiology 1977, 59 (4), 546 549. 166. Baker, J. E.; Anderson, J. D.; Lieberman, M.; Apelbaum, A., Plant Physiology 1979, 63 (5), 91 91. 167. Mitchell, R. E.; Frey, E. J.; Benn, M. H., Phytochemistry 1986, 25 (12), 2711 2715. 168. Mitche ll, R. E.; Frey, E. J., Physiological and Molecular Plant Pathology 1988, 32 (2), 335 341. 169. Mitchell, R. E., Phytochemistry 1989, 28 (6), 1617 1620. 170. Mitchell, R. E.; Coddington, J. M., Phytochemistry 1991, 30 (6), 1809 1814. 171. Duodu, S.; Bhu vaneswari, T. V.; Stokkermans, T. J. W.; Peters, N. K., Molecular Plant Microbe Interactions 1999, 12 (12), 1082 1089. 172. Mills, J.; Masur, H., Scientific American 1990, 263 (2), 50 57. 173. Mukhopadhyay, R.; Roy, K.; Coutinho, L.; Rupp, R. H.; Ganguli B. N.; Fehlhaber, H. W., The Journal of Antibiotics 1987, 40 (7), 1050 1052. 174. Vanmiddlesworth, F.; Dufresne, C.; Wincott, F. E.; Mosley, R. T.; Wilson, K. E., Tetrahedron Letters 1992, 33 (3), 297 300. 175. Vanmiddlesworth, F.; Giacobbe, R. A.; L opez, M.; Garrity, G.; Bland, J. A.; Bartizal, K.; Fromtling, R. A.; Polishook, J.; Zweerink, M.; Edison, A. M.; Rozdilsky, W.; Wilson, K. E.; Monaghan, R. L., Journal of Antibiotics 1992, 45 (6), 861 867. 176. Horn, W. S.; Smith, J. L.; Bills, G. F.; Rag hoobar, S. L.; Helms, G. L.; Kurtz, M. B.; Marrinan, J. A.; Frommer, B. R.; Thornton, R. A.; Mandala, S. M., J Antibiot (Tokyo) 1992, 45 (10), 1692 6. 177. Zweerink, M. M.; Edison, A. M.; Wells, G. B.; Pinto, W.; Lester, R. L., Journal of Biological Chemi stry 1992, 267 (35), 25032 25038.

PAGE 276

276 178. Mori, K.; Otaka, K., Tetrahedron Letters 1994, 35 (49), 9207 9210. 179. Otaka, K.; Mori, K., European Journal of Organic Chemistry 1999, (8), 1795 1802. 180. Kobayashi, S.; Furuta, T.; Hayashi, T.; Nishijima, M.; Hanada, K., Journal of the American Chemical Society 1998, 120 (5), 908 919. 181. Kobayashi, S.; Furuta, T., Tetrahedron 1998, 54 (35), 10275 10294. 182. Trost, B. M.; Lee, C. B., Journal of the American Chemical Society 2001, 123 (49), 12191 12201. 183 Trost, B. M., Chemical and Pharmaceutical Bulletin 2002, 50 (1), 1 14. 184. Franz, S. E.; Stewart, J. D., Chapter Three Threonine Aldolases. In Advances in Applied Microbiology Sima Sariaslani and Geoffrey Michael, G., Ed. Academic Press: 2014; Vol. Volume 88, pp 57 101. 185. Wong, C. H.; Whitesides, G. M., Journal of Organic Chemistry 1983, 48 (19), 3199 3205. 186. Brovetto, M.; Gamenara, D.; Mendez, P. S.; Seoane, G. A., Chemical Reviews 2011, 111 (7), 4346 4403. 187. Clapes, P.; Fessner, W. D.; Sprenger, G. A.; Samland, A. K., Current Opinion in Chemical Biology 2010, 14 (2), 154 167. 188. Fesko, K.; Uhl, M.; Steinreiber, J.; Gruber, K.; Griengl, H., Angewandte Chemie International Edition 2010, 49 (1), 121 124. 189. Liu, J. Q.; Dairi, T.; Kata oka, M.; Shimizu, S.; Yamada, H., Journal of Bacteriology 1997, 179 (11), 3555 3560. 190. Vassilev, V. P.; Uchiyama, T.; Kajimoto, T.; Wong, C. H., Tetrahedron Letters 1995, 36 (23), 4081 4084. 191. Steinreiber, J.; Fesko, K.; Mayer, C.; Reisinger, C.; S churmann, M.; Griengl, H., Tetrahedron 2007, 63 (34), 8088 8093. 192. Steinreiber, J.; Fesko, K.; Reisinger, C.; Schrmann, M.; van Assema, F.; Wolberg, M.; Mink, D.; Griengl, H., Tetrahedron 2007, 63 (4), 918 926. 193. Gwon, H. J.; Baik, S. H., Biotechn ology Letters 2010, 32 (1), 143 149.

PAGE 277

277 194. Gutierrez, M. L.; Garrabou, X.; Agosta, E.; Servi, S.; Parella, T.; Joglar, J.; Clapes, P., Chemistry a European Journal 2008, 14 (15), 4647 4656. 195. Gwon, H. J.; Yoshioka, H.; Song, N. E.; Kim, J. H.; Song, Y. R.; Jeong, D. Y.; Baik, S. H., Preparative Biochemistry and Biotechnology 2012, 42 (2), 143 154. 196. Kimura, T.; Vassilev, V. P.; Shen, G. J.; Wong, C. H., Journal of the American Chemical Society 1997, 119 (49), 11734 11742. 197. Sagui, F.; Conti, P.; Roda, G.; Contestabile, R.; Riva, S., Tetrahedron 2008, 64 (22), 5079 5084. 198. Fesko, K.; Strohmeier, G. A.; Breinbauer, R., Applied Microbiology and Biotechnology 2015, 99 (22), 9651 9661. 199. Liu, J. Q.; Dairi, T.; Itoh, N.; Kataoka, M.; Shimizu, S .; Yamada, H., Journal of Molecular Catalysis B Enzymatic 2000, 10 (1 3), 107 115. 200. Goldberg, S. L.; Goswami, A.; Guo, Z. W.; Chan, Y.; Lo, E. T.; Lee, A.; Truc, V. C.; Natalie, K. J.; Hang, C.; Rossano, L. T.; Schmidt, M. A., Organic Process Research and Development 2015, 19 (9), 1308 1316. 201. Liu, J. Q.; Odani, M.; Dairi, T.; Itoh, N.; Shimizu, S.; Yamada, H., Applied Microbiology and Biotechnology 1999, 51 (5), 586 591. 202. Fesko, K.; Giger, L.; Hilvert, D., Bioorganic and Medicinal Chemistry L etters 2008, 18 (22), 5987 5990. 203. Toscano, M. D.; Muller, M. M.; Hilvert, D., Angewandte Chemie International Edition 2007, 46 (24), 4468 4470. 204. Fesko, K.; Reisinger, C.; Steinreiber, J.; Weber, H.; Schurmann, M.; Griengl, H., Journal of Molecula r Catalysis B Enzymatic 2008, 52 53 19 26. 205. Kielkopf, C. L.; Burley, S. K., Biochemistry 2002, 41 (39), 11711 11720. 206. di Salvo, M. L.; Remesh, S. G.; Vivoli, M.; Ghatge, M. S.; Paiardini, A.; D'Aguanno, S.; Safo, M. K.; Contestabile, R., Febs Jo urnal 2014, 281 (1), 129 145. 207. Qin, H. M.; Imai, F. L.; Miyakawa, T.; Kataoka, M.; Kitamura, N.; Urano, N.; Mori, K.; Kawabata, H.; Okai, M.; Ohtsuka, J.; Hou, F.; Nagata, K.; Shimizu, S.; Tanokura, M., Acta Crystallographica Section D Biological Crys tallography 2014, 70 1695 1703.

PAGE 278

278 208. Uhl, M. K.; Oberdorfer, G.; Steinkellner, G.; Riegler Berket, L.; Mink, D.; van Assema, F.; Schurmann, M.; Gruber, K., Plos One 2015, 10 (4). 209. Rayment, I., Macromolecular Crystallography, Pt A 1997, 276 171 179. 210. Paiardini, A.; Contestabile, R.; D'Aguanno, S.; Pascarella, S.; Bossa, F., Biochimica et Biophysica Acta (BBA) Proteins and Proteomics 2003, 1647 (1 2), 214 219. 211. Liu, J. Q.; Dairi, T.; Itoh, N.; Kataoka, M.; Shimizu, S.; Yamada, H., Journal of Biological Chemistry 1998, 273 (27), 16678 16685. 212. Mugford, P. F.; Wagner, U. G.; Jiang, Y.; Faber, K.; Kazlauskas, R. J., Angewandte Chemie International Edition 2008, 47 (46), 8782 8793. 213. Hirato, Y.; Goto, M.; Tokuhisa, M.; Tanigawa, M.; Nis himura, K., Acta Crystallographica Section F Structural Biology Communications 2017, 73 86 89. 214. Bougioukou, D. J.; Kille, S.; Taglieber, A.; Reetz, M. T., Advanced Synthesis and Catalysis 2009, 351 (18), 3287 3305. 215. Lee, S. J.; Kang, H. Y.; Lee, Y., Journal of Molecular Catalysis B Enzymatic 2003, 26 (3 6), 265 272. 216. Giger, L.; Toscano, M. D.; Bouzon, M.; Marliere, P.; Hilvert, D., Tetrahedron 2012, 68 (37), 7549 7557. 217. Wieteska, L.; Ionov, M.; Szemraj, J.; Feller, C.; Kolinski, A.; Gro nt, D., Journal of Biotechnology 2015, 199 69 76. 218. Seebeck, F. P.; Hilvert, D., Journal of the American Chemical Society 2003, 125 (34), 10158 10159. 219. Seebeck, F. P.; Guainazzi, A.; Amoreira, C.; Baldridge, K. K.; Hilvert, D., Angewandte Chemie International Edition 2006, 45 (41), 6824 6826. 220. Tufvesson, P.; Lima Ramos, J.; Jensen, J. S.; Al Haque, N.; Neto, W.; Woodley, J. M., Biotechnology and Bioengineering 2011, 108 (7), 1479 1493. 221. Tominaga, M.; Nagashima, M.; Taniguchi, I., Electro chem. Commun. 2007, 9 911 914. 222. Dckers, N.; Baer, K.; Simon, S.; Groger, H.; Hummel, W., Applied Microbiology and Biotechnology 2010, 88 (2), 409 424. 223. Fesko, K., Applied Microbiology and Biotechnology 2016, 100 (6), 2579 2590.

PAGE 279

279 224. Goldberg, S L.; Goswami, A.; Guo, Z. W.; Chan, Y.; Lo, E. T.; Lee, A.; Truc, V. C.; Natalie, K. J.; Hang, C.; Rossano, L. T.; Schmidt, M. A., Organic Process Research and Development 2015, 19 (9), 1308 1316. 225. Yao, P.; Lin, Y. J.; Wu, G. B.; Lu, Y. L.; Zhan, T.; Kumar, A.; Zhang, L. L.; Liu, Z. D., International Journal of Biological Macromolecules 2015, 79 965 970. 226. Zhang, K.; Guo, Y. M.; Yao, P.; Lin, Y. J.; Kumar, A.; Liu, Z. D.; Wu, G. B.; Zhang, L. L., Enzyme and Microbial Technology 2016, 85 12 18. 227. Nishiya, Y.; Imanaka, T., Febs Letters 1998, 438 (3), 263 266. 228. Job, V.; Marcone, G. L.; Pilone, M. S.; Pollegioni, L., Journal of Biological Chemistry 2002, 277 (9), 6985 6993. 229. Packeiser, H.; Lim, C.; Balagurunathan, B.; Wu, J.; Zhao, H., Applied Biochemistry and Biotechnology 2013, 169 (2), 695 700. 230. Khuhawar, M. Y.; Majidano, S. A., Chromatographia 2011, 73 (7 8), 701 708. 231. Zuman, P., Chemical Reviews 2004, 104 (7), 3217 3238. 232. Vasanits, A.; Kutlan, D.; Sass, P.; Molnar Per l, I., Journal of Chromatography A 2000, 870 (1 2), 271 287. 233. Nimura, N.; Kinoshita, T., Journal of Chromatography 1986, 352 169 177. 234. Donike, M., Journal of Chromatography 1969, 42 103. 235. Yoon, H. R., Archives of Pharmacal Research 2013, 3 6 (3), 366 373. 236. Schummer, C.; Delhomme, O.; Appenzeller, B. M. R.; Wennig, R.; Millet, M., Talanta 2009, 77 (4), 1473 1482. 237. Steinreiber, J.; Schuermann, M.; van Assema, F.; Wolberg, M.; Fesko, K.; Reisinger, C.; Mink, D.; Griengl, H., Advanced Synthesis and Catalysis 2007, 349 (8 9), 1379 1386. 238. Gwon, H. J.; Yoshioka, H.; Song, N. E.; Kim, J. H.; Song, Y. R.; Jeong, D. Y.; Baik, S. H., Preparative Biochemistry and Biotechnology 2012, 42 (2), 143 154. 239. Matsumori, N.; Kaneno, D.; Murata, M.; Nakamura, H.; Tachibana, K., Journal of Organic Chemistry 1999, 64 (3), 866 876. 240. Krishnamurthy, V. V., Journal of Magnetic Resonance Series A 1996, 121 (1), 33 41.

PAGE 280

280 241. Seco, J. M.; Quio, E.; Riguera, R., Chemical Reviews 2012, 112 4603 4641. 242. Ghiviriga, I., Journal of Organic Chemistry 2012, 77 (8), 3978 3985. 243. Schirch, D.; Dellefratte, S.; Iurescia, S.; Angelaccio, S.; Contestabile, R.; Bossa, F.; Schirch, V., Journal of Biological Chemistry 1993, 268 (31), 23132 23138. 244. Sambr ook, J.; Fritsch, E. F.; Maniatis, T., Molecular Cloning. A Laboratory Manual. Second ed.; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, New York, 1989. 245. Blaskovch, M. A.; Evindar, G.; Rose, N. G. W.; Wilkinson, S.; Luo, Y.; Lajoie, G. A., J. Org. Chem. 1998, 63 3631 3646. 246. Looker, J. H.; Thatcher, D. N., J. Org. Chem. 1957, 22 1233 1237. 247. Steinreiber, J.; Schuermann, M.; van Assema, F.; Wolberg, M.; Fesko, K.; Reisinger, C.; Mink, D.; Griengl, H., Advanced Synthesis and Catalysi s 2007, 349 (8 9), 1379 1386. 248. Winn, M. D.; Ballard, C. C.; Cowtan, K. D.; Dodson, E. J.; Emsley, P.; Evans, P. R.; Keegan, R. M.; Krissinel, E. B.; Leslie, A. G. W.; McCoy, A.; McNicholas, S. J.; Murshudov, G. N.; Pannu, N. S.; Potterton, E. A.; Powe ll, H. R.; Read, R. J.; Vagin, A.; Wilson, K. S., Acta Crystallographica Section D Biological Crystallography 2011, 67 235 242. 249. Adams, P. D.; Afonine, P. V.; Bunkoczi, G.; Chen, V. B.; Davis, I. W.; Echols, N.; Headd, J. J.; Hung, L. W.; Kapral, G. J.; Grosse Kunstleve, R. W.; McCoy, A. J.; Moriarty, N. W.; Oeffner, R.; Read, R. J.; Richardson, D. C.; Richardson, J. S.; Terwilliger, T. C.; Zwart, P. H., Acta Crystallographica Section D Biological Crystallography 2010, 66 213 221. 250. Emsley, P.; L ohkamp, B.; Scott, W. G.; Cowtan, K., Acta Crystallographica Section D Biological Crystallography 2010, 66 486 501. 251. Afonine, P. V.; Grosse Kunstleve, R. W.; Echols, N.; Headd, J. J.; Moriarty, N. W.; Mustyakimov, M.; Terwilliger, T. C.; Urzhumtsev, A.; Zwart, P. H.; Adams, P. D., Acta Crystallographica Section D Biological Crystallography 2012, 68 352 367. 252. Robert, X.; Gouet, P., Nucleic Acids Research 2014, 42 (1), 320 324. 253. Reetz, M. T., Journal of Organic Chemistry 2009, 74 (16), 5767 5 778. 254. Reetz, M. T.; Krebs, G. P. L., Comptes Rendus Chimie 2011, 14 (9), 811 818.

PAGE 281

281 255. Patterson Orazem, A.; Sullivan, B.; Stewart, J. D., Bioorganic and Medicinal Chemistry 2014, 22 (20), 5628 5632. 256. Liu, Z. Q.; Zhang, X. H.; Xue, Y. P.; Xu, M.; Zheng, Y. G., Journal of Agricultural and Food Chemistry 2014, 62 (20), 4685 4694. 257. Lutz, S., Current Opinion in Biotechnology 2010, 21 (6), 734 743. 258. Sullivan, B.; Walton, A. Z.; Stewart, J. D., Enzyme and Microbial Technology 2013, 53 (1), 70 77. 259. Chica, R. A.; Doucet, N.; Pelletier, J. N., Current Opinion in Biotechnology 2005, 16 (4), 378 384. 260. Morley, K. L.; Kazlauskas, R. J., Trends in Biotechnology 2005, 23 (5), 231 237. 261. Ozaki, S. I.; Demontellano, P. R. O., Journal of the American Chemical Society 1994, 116 (10), 4487 4488. 262. Reetz, M. T.; Wilensek, S.; Zha, D. X.; Jaeger, K. E., Angewandte Chemie International Edition 2001, 40 (19), 3589 263. Breuer, M.; Ditrich, K.; Habicher, T.; Hauer, B.; Keeler, M.; Strmer, R.; Zelinski, T., Angewandte Chemie International Edition 2004, 43 (7), 788 824. 264. Panke, S.; Wubbolts, M., Current Opinion in Chemical Biology 2005, 9 (2), 188 194. 265. Sun, H.; Zhang, H.; Ang, E. L.; Zhao, H., Bioorganic and Medicinal Chemistry 2017 266. Patel, R. N., Bioorganic and Medicinal Chemistry 2017 267. Guthrie, J. P., Canadian Journal of Chemistry 1975, 53 (6), 898 906.

PAGE 282

282 BIOGRAPHICAL SKETCH Sarah Franz Beaudoin was born in Dunedin, Florida in 1989 Shortly thereafter her famil y moved to Savannah, Georgia where s he attended Sava nnah Christian Prepatory School. After completing her coursework in 2008, she entered Armstrong Atlantic State University in Savannah, Georgia. She began undergraduate level research in the lab of Dr. B rent Feske in the summer of 2010. She published a first author paper in Synthesis lactones. While attending Armstrong Atlantic State University Sarah received several prestigious rewards including the Robert Kolodney Scholarship and the Department of Chemistry and Physics biochemistry award. She received her Bachelor of Science in chemistry from Armstrong Atlantic State University in December 2012. Sarah followed her passion for biocatalysis and attended the University of Flori da in the summer 2013. She began her graduate studies in the lab of Dr. Jon Stewart where she launched her project on threonine aldolases. In 2014, she published a book Chapter 3: Advances in Applied Microbiology While attending the University of Florida, she received the Sarah received her Doctor of Philosophy in chemistry from the University of Florida in December 2017.