Evolving Alkene Reductase Enantiocomplementarity Through Iterative Saturation Mutagenesis

MISSING IMAGE

Material Information

Title:
Evolving Alkene Reductase Enantiocomplementarity Through Iterative Saturation Mutagenesis
Physical Description:
1 online resource (188 p.)
Language:
english
Creator:
Walton, Adam Zane
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Chemistry
Committee Chair:
Stewart, Jon D
Committee Members:
Horenstein, Nicole A
Bruner, Steven Douglas
Aponick, Aaron Steven
Preston, James F

Subjects

Subjects / Keywords:
enantiocomplementarity -- mutagenesis -- oye
Chemistry -- Dissertations, Academic -- UF
Genre:
Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Old Yellow Enzyme (Saccharomyces pastorianus) was the first discovered flavo-protein and has served as a model for much of what is known about alkene reductase enzymes. These enzymes catalyze the net trans-addition of hydrogen to activated alkenes, namely, ketones, aldehydes, esters, and nitro alkenes. We introduce the use of two new compounds, 2-(hydroxymethyl) cyclohexenone and 2-(hydroxymethyl) cyclopentenone, to the list of known substrates. These compounds, along with hydroxymethyl methylacrylate, are reduced asymmetrically with a battery of Old Yellow Enzyme (OYE) homologs and OYE mutants with amino acid substitutions at the active site residue Trp116. The results yield a range of (R)- and (S)-reduction products. OYE 2.6 (Pichia stipitis) yields the best conversions and high stereoselectivity for the (S)-enantiomer of all three substrates. However, no variant gave a high yield and optical purity for the (R)-enantiomers. To obtain enantiocomplemantarity, we proposed using the strategy of Iterative Saturation Mutagenesis to evolve (R)-stereoselectivity for the three substrates of interest. In this approach, OYE 2.6 was subject to two rounds of saturation mutagenesis to achieve 89% enantiomeric excess of (R)-2-(hydroxymethyl) cyclopentanone and 90% enantiomeric excess of (R)-3-hydroxy-2-methylpropanoate (Roche ester). Also presented in this work, is a novel methodology for the generation of saturated mutant libraries. This approach includes assessment criteria for a pooled mutant plasmid library from sequencing chromatographic data. The result was an elimination of the requirement to sequence a large portion of a mutant library to assess overall library quality, dramatically streamlining the directed evolution cycle. The applicability of alkene reductases was also demonstrated by two novel preparative-scale enzymatic approaches to useful synthetic intermediates. The synthetic menthol intermediate (R)-citronellal was prepared with a volumetric productivity of 2.8 g/L/hr and an enantiomeric excess of 98%. The (S)-enantiomer of the chiral synthetic intermediate 2-methylpentanol was prepared at a volumetric productivity of 0.7 g/L/hr and an enantiomeric excess of 99%. The combination of this scale-up methodology along with the power of directed evolution vastly enhance the utility of alkene reductase enzymes for the production of fine chemicals and other synthetic intermediates.
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.
Statement of Responsibility:
by Adam Zane Walton.
Thesis:
Thesis (Ph.D.)--University of Florida, 2012.
Local:
Adviser: Stewart, Jon D.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-05-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0043949:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 EVOLVING ALKENE REDUCTASE ENANTIOCOMPLEMENTARITY THROUGH ITERATIVE SATURATION MUTAGENESIS By ADAM Z. WALTON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIRE MENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Adam Z. Walton

PAGE 3

3 In memory of 1 st Lieutenant Mark H. Dooley, 1978 2005

PAGE 4

4 ACKNOWLEDGMENTS Foremost ice, and for his unrelenting suppo rt over many years. His advice and guidance have been essential in the completion this study. At the risk of failing to do so in the following text I must thank and give credit upfront to Dr. Bradford (Brad) Sullivan for his many contributions. Specifically, it was his guidance and expertise that allowed me to complete the much of the scale up work described in Chapter 2. It was his idea to include methyl acrylate in our Baylis Hillman adduct panel and he conducted a vas t never achieved a working mutagenesis protocol. While we did much of the work and learning together more often than not he was the teacher and I the student. His con tributions are manifest in the pBS# plasmid nomenclature referenced throughout this document. For all of these things and countless others I say cheers. Specific t hanks is also due to Yuri Pompeu for solving the crystal structure of OYE 2.6 during the c ourse of this study and numerous discussi ons and keen observations on structural aspect s of the enzymes in this field. And finally, thanks to many unnamed Stewart group members over the past 10 years that have both influenced my education and provided pre cedence for and assistance with this research. I need to thank past leadership of the Department of Chemistry and Life Science, United States Military Academy, Brigadier General David Allbee (ret.), Colonel Patricia Dooley (ret.), and Lieutenant Colonel Br ad Dick (ret.) for their mentorship, guidance and faith in my abilities that gave me the opportunity to pursue this degree. Thanks is given to the current leadership of the Department, the U nited S tates M ilitary A cademy and the U S Army A dvanced C ivilia n S chooling program for financial support required to

PAGE 5

5 pursue this study. I am additionally grateful to the U.S. Army for seemingly endless opportunities for growth and development over the past 23 years. Finally, I would be remiss if I failed to acknowl edge the contributions of my family. I owe thanks to my beautiful wife, Jennifer, and our four sons for their encouragement and cheerleading during this work and my career. Their unconditional support, as always, is the source of my motivation for everyt hing that I do.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 10 LIST OF FIGURES ................................ ................................ ................................ ........ 11 ABSTRACT ................................ ................................ ................................ ................... 14 CHAPTER 1 OLD YELLOW ENZYME, THE ORIGINAL ALKENE REDUCTASE ....................... 16 Identification and Characterization ................................ ................................ .......... 16 Substrate Specifi city ................................ ................................ ............................... 18 Ketones and Aldehydes ................................ ................................ .................... 18 Esters ................................ ................................ ................................ ............... 21 Nitro Alkenes ................................ ................................ ................................ .... 22 Enzyme Structure ................................ ................................ ................................ ... 23 Histidine 191 and Asparagine 194 ................................ ................................ .... 24 Tyrosine 196 ................................ ................................ ................................ ..... 24 Threonine 37 ................................ ................................ ................................ .... 25 Tryptophan 116 ................................ ................................ ................................ 25 Experimental Strategy ................................ ................................ ............................. 26 Experimental Procedures ................................ ................................ ........................ 28 Preparation of Baylis Hillman Adducts ................................ ............................. 28 2 (Hydroxymethyl) cyclohex 2 enone ................................ ........................ 28 2 (Hydroxymethyl) c yclopent 2 enone ................................ ....................... 29 Methyl 2 (hydroxymethyl)acrylate ................................ .............................. 29 Screening Reactions ................................ ................................ ........................ 30 Larger Scale Preparation of ( S ) 4 ................................ ................................ .... 30 Larger Scale Preparation of ( S ) 5 and Assignment of Absolute Configuration ................................ ................................ ................................ 32 Results and Discussion ................................ ................................ ........................... 33 Alkene Reductase Homologs ................................ ................................ ........... 33 Tryptophan 116 Mutations ................................ ................................ ................ 34 Conclusions and Future Work ................................ ................................ ................. 35 2 PREPARATIVE SCALE ENZYMATIC ALKENE REDUCTIONS ............................. 51 Background ................................ ................................ ................................ ............. 51 ( R ) Citronellal ................................ ................................ ................................ ... 51 2 Methylpentanol ................................ ................................ .............................. 53 Experimental Strategy ................................ ................................ ............................. 53

PAGE 7

7 Preparation of Enantiomerically Pure ( R ) and ( S ) Citronellal Using OYE Homologs ................................ ................................ ................................ ...... 53 Preparation of ( S ) 2 Methylpentanol Using OYE2 and YahK ........................... 55 Experimental Procedures ................................ ................................ ........................ 55 GC/MS Analysis ................................ ................................ ............................... 55 PDMS Membrane Preparation ................................ ................................ ......... 56 Small scale Test Reactions ................................ ................................ .............. 57 CLEA Preparation ................................ ................................ ............................ 57 Acetylation of GST OYE 2.6. ................................ ................................ ............ 58 Preparation of ( R ) C itronellal ................................ ................................ ............ 58 Preparation of geranial ................................ ................................ ............... 58 Preparation of OYE 2.6 catalyst ................................ ................................ 59 Enzymatic reduction of geranial to ( R) citronellal ................................ ....... 60 Pre paration of ( S ) Citronellal ................................ ................................ ............ 61 Preparation of neral ................................ ................................ ................... 61 Preparation of NemA catalyst ................................ ................................ .... 62 Enzymatic reduction of neral to ( S) citronellal ................................ ............ 62 Preparation of ( S ) 2 Methylpentanol ................................ ................................ 63 Preparation of OYE2 and YahK catalyst ................................ .................... 63 Enzymatic reduction of 2 methyl 2 pentenal to ( S ) 2 methylpentanol ........ 64 Results and Discussion ................................ ................................ ........................... 65 ( R ) and ( S ) Citronellal ................................ ................................ ..................... 65 ( S ) 2 Methylpentanol ................................ ................................ ........................ 67 Conclusion ................................ ................................ ................................ .............. 68 3 PROTEIN ENGINEERING TO EVOLVE ENANTIOSELECTIVITY ......................... 76 Introduction ................................ ................................ ................................ ............. 76 Directed Evolution ................................ ................................ ................................ ... 76 Generating Diversity ................................ ................................ ......................... 77 Recombinative methods ................................ ................................ ............ 78 Non recombinative methods ................................ ................................ ...... 80 Semi rational methods ................................ ................................ ............... 82 The numbers problem ................................ ................................ ................ 85 Screening ................................ ................................ ................................ ......... 86 Comparing Directed Evolution Approaches ................................ ...................... 87 galactosidase ................................ ................................ .......................... 88 Epoxide hydrolase ................................ ................................ ..................... 90 Lipase ................................ ................................ ................................ ........ 95 ISM of an Alkene Reductase, YqjM ................................ ................................ .. 99 Conclusions ................................ ................................ ................................ .......... 101 4 CREATING SITE SPECIFIC SATURATED LIBRARIES ................................ ....... 109 Introduction ................................ ................................ ................................ ........... 109 Experimental Strategy ................................ ................................ ........................... 109 Site Selection ................................ ................................ ................................ 109

PAGE 8

8 Library Creation ................................ ................................ .............................. 111 Assessing Degeneracy ................................ ................................ ................... 112 Experimental Procedures ................................ ................................ ...................... 113 Construction of Pooled Degenerate Plasmid ................................ .................. 113 Degenerate Library Creation ................................ ................................ .......... 114 Results and Discussion ................................ ................................ ......................... 115 Developing a Mutagenesis Method ................................ ................................ 115 Transformation efficiency ................................ ................................ ......... 115 Primer concatemers ................................ ................................ ................. 117 Obtaining degeneracy ................................ ................................ .............. 118 Assessing Degeneracy ................................ ................................ ................... 120 Library Quality ................................ ................................ ................................ 121 Conclusions ................................ ................................ ................................ .......... 123 5 ITERATIVE SATURATION MUTAGEN ESIS: PUTTING IT ALL TOGETHER ...... 139 Introduction ................................ ................................ ................................ ........... 139 Experimental Strategy ................................ ................................ ........................... 139 Experimental Procedures ................................ ................................ ...................... 141 Library Master Plates ................................ ................................ ..................... 141 Auto Induction ................................ ................................ ................................ 141 Whole Cell Assays ................................ ................................ ......................... 142 Site directed Mutagenesis ................................ ................................ .............. 142 Results and Discussion ................................ ................................ ......................... 143 First Round Screening. ................................ ................................ ................... 143 T35X ................................ ................................ ................................ ........ 144 H188X and H191X ................................ ................................ ................... 144 Y193X ................................ ................................ ................................ ...... 145 I113X ................................ ................................ ................................ ....... 146 Summary of first round results ................................ ................................ 147 Second Round Screening ................................ ................................ ............... 147 Y78W ................................ ................................ ................................ ....... 148 Y78W, F247X ................................ ................................ ........................... 149 Purified enzyme screening results ................................ ........................... 150 Conclusions ................................ ................................ ................................ .......... 151 APPENDIX A DEGE NERATE LIBRARY SEQUENCING DATA ................................ ................. 167 B SEQUENCE DATA ................................ ................................ ............................... 176 C MUTAGENIC PRIMERS ................................ ................................ ....................... 179 D MUTAGENIC PLASMIDS ................................ ................................ ..................... 180 LIST OF REFERENCES ................................ ................................ ............................. 181

PAGE 9

9 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 188

PAGE 10

10 LIST OF TABLES Table page 2 1 Acetylation of OYE 2.6 ................................ ................................ ....................... 74 2 2 Productivity of alkene reductase biotransformations ................................ .......... 74 3 1 International Union of Biochemistry and Molecular Biology (IUBMB) nucleotide nomenclature ................................ ................................ .................. 105 3 2 Coding description of selected degenerate codons ................................ .......... 105 3 3 Overs ampling for 95% coverage as a function of degeneracy and number of simultaneously randomized sites. ................................ ................................ ..... 106 3 4 NNK degenerate codon usage. ................................ ................................ ........ 106 3 5 Summary of ANEH directed evolution efforts. ................................ .................. 108 4 1 OYE active site measurements. ................................ ................................ ....... 126 4 2 Summary of mutagenesis development efforts ................................ ................. 133 4 3 Amino acid bias. ................................ ................................ ............................... 135 4 4 Assessment of first round library degeneracy. ................................ .................. 135 5 1 Libraries screened in this study. ................................ ................................ ....... 156 5 2 Selected first round screening resu lts. ................................ .............................. 15 7 5 3 Summary second round screening results. ................................ ...................... 163 5 4 Summary purified enzyme screening results. ................................ ................... 165 C 1 List of mutagenic primers. ................................ ................................ ................ 179 D 1 Lis t of plasmids used in this study. ................................ ................................ ... 180

PAGE 11

11 LIST OF FIGURES Figure page 1 1 Warburg and Christian reaction system ................................ .............................. 37 1 2 OYE1 substrates. ................................ ................................ ............................... 38 1 3 OYE substrate binding modes ................................ ................................ ............ 45 1 4 OYE 1 active site crystal structure. ................................ ................................ ..... 46 1 5 Reaction scheme for the reduction of Baylis Hillman adducts ............................ 47 1 6 Baylis Hillman reaction mechanism ................................ ................................ .... 48 1 7 Results of OYE homolog screening with Baylis Hillman adducts ....................... 49 1 8 Results of W116X OYE1 screening with Baylis Hillman adducts ....................... 49 1 9 Results of W116X OYE 1 screening with methyl 2 (hydroxymethyl)acrylate ....... 50 2 1 Initial proposed scheme for the preparation of ( R ) citronellal ............................. 70 2 2 Proposed 2 compartment scheme for avoiding undesired side reactions .......... 71 2 3 General scheme for the preparation of ( S ) 2 Methylpentanol ............................. 72 2 4 Time course for OYE 2.6 reduction of citral under biphasic conditions ............... 72 2 5 Comparison of strategies for OYE 2.6 mediated reduction of geranial ............... 73 2 6 General scheme for the preparation of ( R) citronellal ................................ ......... 75 2 7 General scheme for the preparation of ( S) citronellal ................................ ......... 75 3 1 Numbers of publications dealing wi th in vitro directed evolution ....................... 103 3 2 The in vitro protein engineering process ................................ ........................... 104 3 3 Iterative Saturation Mutagenesis methodology for the evolution of increased stereoselectivity. ................................ ................................ ............................... 107 4 1 Summary of Chapter 4 experimental objectives and efforts ............................. 125 4 2 OYE 1 active site measurements. ................................ ................................ .... 126 4 3 OYE active site schematic. ................................ ................................ ............... 127

PAGE 12

12 4 4 Sequence alignment of OYE 1 and OYE 2.6. ................................ ................... 128 4 5 Overlay of OYE1 and OYE2.6 crystal structures ................................ .............. 129 4 6 OYE 2.6 active site schematic ................................ ................................ .......... 130 4 7 Plasmid map for pBS2 template. ................................ ................................ ...... 131 4 8 Sample calculation of Q pooled from a degenerate sequencing chromat ............. 132 4 9 Schematic of primer designs used ................................ ................................ .... 133 4 10 Pooled plasmid sequencing from Phusion SDM Kit experiment ..................... 134 4 11 Amino acid distribution for T35, F37, I113 and H188 libraries .......................... 136 4 12 Amino acid distribution for H191, Y193, F247 and N293 libraries .................... 137 4 13 Amino acid distribution for V294, F373 and Y374 libraries ............................... 138 5 1 Summary of Chapter 5 experimental objectives and efforts ............................. 153 5 2 Template for construction of Library Master Plates ................................ .......... 154 5 3 Growth apparatus for auto induct ion of E. coli cells ................................ .......... 155 5 4 First round screening results for T35X, F37X and I113X ................................ .. 158 5 5 First round screening results for H188X, H191X and Y193X ............................ 159 5 6 First round screening results for F247X, N293X and V294X ............................ 160 5 7 First round screening results for F373X and Y374X ................................ ......... 161 5 8 Selected substrate screening results for OYE 2.6 Tyr193 mutant s .................. 162 5 9 Second round screening results for Y78W I113X ................................ ............. 163 5 10 Results of Y78W I113X screening with 2 (hydroxymethyl)cyclopentenone ...... 164 5 11 Results of Y78W I113X screening with methyl 2 (hydroxymethyl)acrylate ....... 164 5 12 The role of Tyr78/Tyr82 in the active site ................................ ......................... 165 5 13 Final ISM pathways to observed best results ................................ ................... 166 A 1 Pooled and Observed Sequencing for T35X and F37X ................................ .... 167 A 2 Pooled and Observed Sequencing for I113X and H188X ................................ 168

PAGE 13

13 A 3 Pool ed and Observed Sequencing for H191X and Y193X ............................... 169 A 4 Pooled and Observed Sequencing for F247X and N293X ............................... 170 A 5 Pooled and Observed Sequencing for V294X and F373X ................................ 171 A 6 Pooled and Observed Sequencing for Y374X ................................ .................. 172 A 7 Pooled Sequencing for A68X, Y78X, and G292X ................................ ............. 173 A 8 Pooled Sequencing for Y78W and I113D Second Round Libraries .................. 174 A 9 Pooled Sequencing for I113W and V294P Second Round Libraries ................ 175 B 1 Sequence of pBS2 template plasmid. ................................ ............................... 176

PAGE 14

14 Abstr act o f Dissertation Pr esented to the Graduate School of the University o f Florida In Par tial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EVOLVING ALKENE REDUCTASE ENANTIOCOMPLEMENTARITY THROUGH ITERATIVE SATURATION MUTAGENESIS By Adam Z. Walton May 2012 Chair: Jon D. Stewart Major: Chemistry Old Yellow Enzyme ( Saccharomyces pastorianus ) was the first discovered flavo protein and has served as a model for much of what is known about alkene reductase enzymes. These enzymes catalyze the net trans addition of hydrogen to activated alkenes, namely, ketones, aldehyde s, esters, and nitro alke nes. We introduce the use of two new compounds, 2 (hydroxymethyl) cyclohexenone and 2 (hydroxymethyl) cyclopentenone, to the list of known substrates. These compounds along with hydroxymethyl meth yl acrylate, are reduced asymmetr ically with a battery of Old Yellow Enzyme ( OYE ) homologs and OYE mutants with amino acid substitutions at the active site residue Trp116. The results yield a range of ( R ) and ( S ) reduction products OYE 2.6 ( Pichia stipitis ) yields the best conversions and high stereoselectivity for the ( S ) enantiomer of all three substrates. However, no variant gave a hi gh yield and optical purity for the ( R ) enantiomers. To obtain enant iocomplemantarity, we propose d using the strategy of Iterative Saturation Mutagen esis to evolve ( R ) stereoselectivity for the three substrates of interes t. In this approach, OYE 2.6 was subject to two rounds of saturation mutagenesis to achieve 89 % enantiomeric excess of ( R ) 2 (hydroxymethyl)

PAGE 15

15 cyclopent anone and 90% enantiomeric excess of ( R ) 3 h ydroxy 2 methylpropanoate (Roche ester). Also presented in this work, is a novel methodology for the generation of saturated mutant libraries. This approach includes assessment criteria for a pooled mutant plasmid library from sequencing c hroma tograph ic data. The result wa s an elimination of the requirement to sequence a large portion of a mutant library to assess overall library quality dramatically streamlining the directed evolution cycle. The appli cability of alkene reductases was also dem ons trated by two novel preparative scale enzymatic approaches to useful synthetic intermediates. The synthetic menthol intermediate ( R ) c itronellal was pr epared with a volumetric productivity of 2.8 g/L/hr and an enantiomeric excess of 98%. The ( S ) enant iomer of the chiral synthetic intermediate 2 methylpen t anol was prepared at a volumetric productivity of 0.7 g/L/hr and an enantiomeric excess of 99%. The combination of this scale up methodology along with the power of directed evolution vastly enhance t he utility of alkene reductase enzymes for the production of fine chemicals and other synthetic intermediates.

PAGE 16

16 CHAPTER 1 OLD YELLOW ENZYME, T HE ORIGINAL ALKENE R EDUCTASE Identification and Characterization Old Yellow Enzyme was the first discovered flavo protein and has served as a model for much of what is known about this class of enzyme s 1 In recent years the catalytic activity of this enzyme and an entire subset of related homologs have appeared in the literature as valuable catalys ts with industrial applications 2 Our interest in this enzyme is to probe the protein sequence space in order to further our understanding of these enzymes. H opefully this will lead to the develop ment of biocatalysts that retain the unique activity o f the enzyme while expanding its substrate scope and stereoselectivity ; therefore in theory expanding its industrial applicabil ity as 3 Old Yellow Enzyme was first isolated from the brewer s bottom yeast Saccharomyces carlsbergensis (so named after the Danish brewery Carlsberg where it was first isolated and later reclassified as Saccharomyces pastorianu s ) b y Warburg and Christian in 1932 4 Th eir original aim was to elucidate the nature of biological oxidations in yeast. In the process of their work the y isolat ed a yellow protein, gel b e ferment that was pr esent in appreciable quantities in yeast lysate When investigating another enzyme that would later be identified as gluco se 6 pho sphate dehydrogenase, the yellow enzyme in the presence of a thermostable extract (NADP + /NADPH) and oxygen, was able to facilitate the oxidation of glucose 6 phosphate thus completing the respiratory chain ( Figure 1 1) The later ( 1938 ) discovery of a ne w yellow enzyme homolog in yeast by Haas led to the designation of the first reported yellow enzyme as referred to today. Later, in 195 6 Theorell and

PAGE 17

17 Akeson reported a rigorous purification technique leading to a much more homogenous enzyme further enabling detailed study of this enzyme 5 As a result, their work co nfirmed the enzyme to be a hom odimer with a 1:1 ratio of FMN to protein. In 1969 expanding on the w ork of Theorell, Matthews and Massey were able to improve on the purification of the Old Yellow E nzyme and set about its detailed characterization 6 The enzyme was found to catalyze the oxidation of NADPH to NADP + where the tw o electron transfer in the form of a hydride is transferred to the oxidized FMN cofactor. This now reduced flavin could then be reoxidized by molecular oxygen; however, the physiologic al substrate, or even putative substrates, at this point was not known. They reported that the enzy me displayed an absorption spectrum that had a maximum around 440nm that shifted to 600nm upon the addition of many phenolic compounds. This resulted in the visual color change from yellow to green and was attributed to the formation of a charge transfer complex 7 These phenol complexes were further characterized by Raman spectroscopy and found to involve interactions between the phenol (ate) ring and the i soalloxazine moiety of the bound flavi n 8 Mass e y exploited phenol binding to develop an affinity purification tech nique using a phenol functionalized resin for the sole chromatographic s tep 9 This further allowed a simpler and more effective isolation of the enzyme from its native host in greater yield. In 1991 the gene for Old Yellow Enzyme (now termed OYE 1) was cloned and overexpressed in E. coli 10 C ompari son of the overexpressed protein with non recombinant enzyme preparations led to the discovery that several OYE isozymes were present in brewers yeast 11 The cloned gene also led to a homogenous sour ce suitable for unambiguous characterization and structural determination.

PAGE 18

18 Substrate S pecificity To date the physiological electron acceptor(s) for Old Yellow Enzyme remain unknown. The generally accepted substrate for reducing the bound flavin mononucleotide cofactor is NADPH. It has been widely noted that molecular oxygen is capable of reoxidizing the FMN ; however the rate at which it does so is thought to be physiologically insignificant 6 Some of t he first substrates proposed for OYE1 were simple quinones such as th ose depicted in Figur e 1 2 a 11 Additionally cyclohex 2 ene 1 one has been proposed as a simp le model substrate for the reoxidization of FMN with significant activity in nearly all reported OYE homologs This has led to the widespread investigation of unsaturated carbonyl compounds in the characterization of OYE Ketones and A ldehydes O ne of the first and most extensive panels of substrates tested for OYE 1 was described by Vaz et al (Figure 1 2 b ) 12 It was observed that unsaturated aldehydes an d ketones, including cyclic ketones were among the most active substrates for OYE 1 Later while probing the kinetics of the oxid ative reaction, Kohli and Massey identified several additional cinnemaldehyde derivatives that displayed modest activity in the enzyme (Figure 1 2c) 13 The enzyme was found to have little to no activity for sterols. Notably the enzyme also showed no activity for vinyl acids, esters amides or nitriles in the study. Among the obse rvations in this work it was observed that increasing the size of alkyl substituents at the position dramatically reduced the rate of hydride transfer presumably due to steric hindrance within the active site. No characterization of the enantioselectivity of these re ductions was done in this work These workers also reported a novel dismutat ion reaction that occurs in the abs ence of the hydride donor NADPH. In this reaction OYE first oxidizes a molecule of the cyclohexenone substrate

PAGE 19

19 yielding reduced FMN and a cyclohexadienone. The FMNH 2 then transfers the electrons to a second molecule of the substrate producing a saturated cyclohexanone. The initial cyclohexadienone rapidly tautomerizes to the corresponding phenol, which tightly binds to the activ e site as previously described. The first systematic exploration of OYE 1 enantioselectivity was conducted in our group and reported in 2006 14 Based on results from 2 and 3 substituted cyc lohexen ones (Figure 1 2 d) it was observed that smaller alk yl substituents at either position yielded greater conversion s This observation is consistent with those made by Vaz et al Additionally all compounds tested yielded the same enantiomer indica ting a single binding mode for the h omologous substrates. This single binding mode, ( Figure 1 3) is one where the hydride adds to the carbon and a proton adds to the carbon, from below and above the plane of the page, respectively, as drawn in Figure 1 2 d. In 2007 Muller et al introduced several new compounds into the growing lis t of OYE substrates (Figure 1 2 d) 15 These included 2 methyl pentena l and both isomers of citral (neral and geranial) The enzyme was found to have a preference for the E isomer of citral (geranial) and yielded primarily the ( R ) enantiomer indicating binding analogous to that observed by Swiderska and Stewart. The Z isomer however gave much poorer conversion and requi red a flipped binding orientation to yield predominantly the same ( R ) enantiomer. Interestingly the observed stereochemistry of the reduced 2 methyl pentanal indicates the same flipped bindin g mode to yield the ( S ) enantiomer with a signific antly sma ller compound. I n the same study the significantly larger methyl cinnamaldehyde conforms to

PAGE 20

20 yield the ( R ) enantiomer in excess. In a separate work M uller et al further demonstrated the ability of OYE to also reduce an activated alkyne (Figure 1 2 d) 16 In this case 4 phenyl 3 butyne 2 one was reduced to ( E ) 4 phenyl 3 butene 2 one which is also a substrate previously identified by Vaz et al In this research group both ( R ) and ( S ) perillaldehyde 17 (Figure 1 2 d) and ( R ) and ( S ) carvone 18 (Figure 1 2 e) were added to the battery of substrates for OYE 1. While these compounds displayed varied bin ding modes when reduced with OYE homologs and mutants the wild type enzyme displayed consistent binding and high diastereomeric excess with these compounds that did not vary the configuration of the new substrate stereo center. In 2008, while comparing v arious OYE homologs, Hall et al reported several new compounds for OYE1 19 (Figure 1 2 e) Methyl substituted cyclopentenones, at both the 2 and 3 position s were found to be reasonable substrates for OYE. The 2 substituted compound yielded a racemic product mixture while the 3 substituted version gave predominantly ( S ) enantiomer product Additionally, 2 methyl maleimides were evaluated and found to give ( R ) enantiomer products supported by the observation that increasing the size of the N substituted alkyl group had little effect on the stereosele ctivity of the reaction. Following this work Stueckler et al demonstrated an efficient reduction of additional methyl cinnamaldehyde derivatives (Figure 1 2 f ) in the preparation o f fragrance compounds Lilial ( t butyl) and Helional (methylenedioxy) 20 They observed high enantiomeric excess in the presence of 20% TBME co solvent. The

PAGE 21

21 predominant ( S ) bulky nature of the trans phenyl and substituted phenyl rings. From the same group Winkler et al explored the suitability of 2 substituted alkoxy compounds (Figure 1 2 f) as novel OYE substrates 21 While overall conversion w as generally poor the 2 methoxy cyclohexenone substrate was the only compound that gave the ( R ) All other g where stereochemistry could be identified. Esters Hall et al ., using dimethyl esters of citraconic, itaconic, and mesaconic acid s (Figure 1 2 e), presented the first report of esters as substrate s for OYE 1 19 The except for the E isomer that p otential binding mode to avoid steric clash at the position T he same group also described the use of methyl 2 hydroxy methylacrylate ( Figure 1 2 e) as OYE 1 substrates to produce the synthetically useful compound 3 hy droxy 2 methylpropanoate 22 In all cases the observed yielding the ( R ) enantiomer presumably once again due to steric bulkiness. It appears as though considerable effort was wasted by systematically increasing the size of the ether functional group when available models were sufficiently useful to predict this outcome. A separate effort revealed the applicability of halo methylci nnamate derivatives ( Figure 1 2 f) for use with OYE 1 23 These compounds yield the ( S ) enantiomer hat is consistent with the cinnamaldehyde derivatives discussed above

PAGE 22

22 In pursuit of synthetically useful 2,4 dimethylbutyrolactone Korpak and Pietruszk used OYE 1 to reduce 2 methyl 4 oxopent 2 enoate ( Figure 1 2 f) as the first step in its preparation 24 Assuming that the keto carbonyl is the prefered electron wi thdrawing group the E isomer R ) enantiomer. Most recently Stueckler et al introduced d ehydroamino acid derivatives (Figure1 2 g) as suitable substrates for OYE1 25 In all cases the substrates theoretically bound as depicted in Figure1 2 g for OYE 1 yielding the ( S ) enantiomers. The diester E isomers each have two possible binding modes that will yield each enantiomer. Using the homolog OYE 3 ( from Saccharomyce s cerevisiae ) they were able to determine that inverted binding of the di ester compounds could be influenced by the size of the N acyl substitution on these compounds Nitro Alkenes In 2000 Meah and Massey first reported the reduction of unsaturated nit r o compounds (Figure 1 2 c) by OYE 1 26 As would be expected, the mechanisms of these r eductions were analogous to those previously described for unsaturated carbonyl compounds The major difference was that the nitronate ion f ormed after hydride transfer to the carbon bound to the enzyme weakly and therefor e c ould dissociate prior to enzyme mediated protonation. This presents potential problems in directing the stereoselectivity of the enzyme at the nitro bearing carbon. A year later Meah et al reported that OYE 1 reduced the nitrate esters nitroglycerine and propylene dinitrate (Figu re 1 2 c ). 27 T he mechanism involves a net two electron transfer to a nitrate ester nitrogen followe d by subsequent elimination of nitrite The enzyme preferred reaction with a terminal nitrate over secondary nitrate moieties No further reduction was observed with the propylene mono nitrate product of

PAGE 23

23 these reactions. Williams et al further demonstrated the ability of OYE 1 to reduce both trinitrotoluene (TNT) an d nitrobenzaldehyde (Figure 1 2 c) while investigating the ability of OYE homologs to degrade these and other compounds 28 This work further implied that hydride addition was directly to the nitro group and not to the ar omatic ring Our research group has previously investigated the asymmetric reduction of a series of nitro acrylates compound s (Figure 1 2 d) by OYE 1 29 In all cases the enzyme yielded a high excess of the ( R ) enantiomer. Deuterium labeling experiments also confirmed that hydri de was added to the carbon relative to the nitro group. Enzyme S tructure In 1994 Fox and Karplus reported the 2.0 resolution crystal structure of OYE1 30 The protein contained a single domain consisting of a parallel, eig ht stranded / barrel similar to that of triosephosphate isomerase (TIM) The largely hydrophobic active site is accessible to the solvent via a deep cleft. This cleft is, in part, obscured by a flexible loop that has implications in the dynamic bind ing of the NADPH substrate. The isoalloxazine ring of the FMN cofactor is postioned across the top of the barrel perpendicular to its axis. One crystal form contained the bound inhibitor p hydroxybenzaldehyde (PHB) bound directly above the FMN prosthetic group The aromatic rings of b oth ligands lay parallel to each other in close contact. The C2 of PHB lies directly above and 3.4 away from N5 of the flavin which is consistent with it s mimicking of C of a 2 cyclohexenone substrate. for formation of the charge transfer complex 8 Since the publication of this structure several efforts have uncovered the roles of key residues within the OYE 1 activ e site ( Figure 1 3 and 1 4)

PAGE 24

24 Histidine 191 and Asparagine 194 The phenolate oxygen of PHB in the Fox and Karplus structure has the appropriate distance and orientation to be hydrogen bonded with the side chains of His191 and Asn194. Brown et al construct ed H191N and H191N/N194H mutants of OYE 1 and characterized their kinetic properties 31 T he enzyme maintained activity with respect to the oxidation of NADPH in both mutan ts, presumably because of the additional contacts the cofactor makes with the enzyme upon binding. By contrast the mutants showed dramatically reduced activity towards the reduction of cyclohex enone and t he y also displayed reduced binding affinity for ph enol inhibitor s. These effects are likely due to altered alignment of the se compounds with in the active site. In order for catalysis to occur carbon of the substrate must be positioned precisely with respect to N5 of the flavin ring to facilitate efficient hydride transfer Hydrogen bonding to His191 and Asn194 stabilizes the enolate that is formed by the addition of the hydride. As the orientation of the hydrogen bond donors are altered the substrate is no longer optimally stabilized upon recei pt of the hydride thus reducing catalytic activity. Tyrosine 196 The same authors also probed the role of Tyr196 in the catalytic mechanism of OYE 1 13 Tyrosine 196 is positioned 3.5 above C3 of the PHB ligand opposite the flavin cofactor. This position is analogous to the carbon of 2 cyclohexenone. The acidity of the tyrosyl proton is potentially reduced from its normal pK a of 10.1 by the close proximity of Asn 251 (2.7 ). Thus Tyr196 is positioned to protonate and thereby yield a net trans addition of hydrogen t o the double bond. When Tyr196 was mutated to a phenylalanine the rate of 2 cyclohex enone reduction decreased by six orders of magnitude 13 unsaturated carbonyl compounds.

PAGE 25

25 Threonine 37 Threonine 37 is positioned in the OYE 1 active site near the bound FMN cofactor. The side chain hydroxyl group is within hydrogen bonding distance (2.8 ) of the C4 carbonyl oxygen of the i soalloxazine ring. By mutating the residue to an alanine Xu et al deduced that Thr 37 helpe d to stabilize negative charg e on the flavin ring system as a result of reduction by NADPH 32 The T37A mutant show ed an order of magnitude decrease in the rate of NADPH oxidation thus signifying a higher reduction potential for the FMN Conversely, the mutant displaye d a higher turnover (2 9 fold) for reducing all substrates investigated. This increase in rate was facilitated by the decreased stability of the reduced mutant enzyme. The overall turnover of the ping pong mechanism was still lower with the mutant enzyme given the more pronounced effect on the NADPH oxidation. Tryptophan 116 Given the observation that 2 cyclohexenones with large substituents at the 2 and 3 positions were reduced slowly (but with high ster e o selectivity ) 14 the Stewart group attempted to uncover the origins of these effects 18 From the crystal structure of OYE 1, the side chain of Trp116 appeared to sterically crow d t he active site, and computer modeling suggested that this hindrance would increase with increasing substituent size Site saturation mutagenesis at position 116 followed by screenin g for increased activity with 3 methyl cyclohexenone uncovered several var iants of interest. The W 116I and W116F mutants were purified and screened against a battery of potential substrates This investigation revealed an interesting result with respect to the reduction of ( R ) and ( S ) carvones (Figure 1 2 e). With ( R ) carvone both W116I and W116F as well as wild type OYE1 provided ( R ) selectivity for the newly created sp 3 center at C consistent

PAGE 26

26 substrate binding mode. ( S ) carvone gave the same result for W116F and the wild type enzyme On the other hand, the W116I mutant gave ( S ) selective reduction nding mode. It was postulated that this alternate substrate binding orientation is possible only when the bulky tryptophan residue is replaced with a n amino acid whose side chain is small enough to accommodate the isopropenyl group on C5 of the carvone su bstrate. A tryptophan is highly conserved among OYE homologs at position 116; however a BLAST search revealed a close rela tive of OYE 1 with an isoleucine at the analogous position (OYE 2.6 from Pichia stipitis 33 ). We cloned and expressed this protein to provide an additional alkene red uctase for our collection of biocatalysts. While its properties were similar to those of OYE 1 in some respects, it also had highly useful properties that were revealed in the course of the studies described here Experimental Strategy The previous sectio ns have describe d what is known about the structure, mechanism and substrate specificities of OYE 1. We had also collected a number of other OYE 1 homologs in our laboratory 18,34 as well as a complete set of amino acid replacements at position 116 in OYE 1 35 Because most of the work to date had focused on simple model substrates, we wanted to increase the utility of our enzyme library by applying it to more complex, and therefore more useful alkenes. We set three additional criteria in our search for new targets First they should be well behaved with respect to enantioselectivity and lack side reactions (i.e. racemization and ketone reduction ) Second, we wanted the reduction products to be useful intermediates for existing synthetic ro utes. Finally, the substrates needed to be commercially available or

PAGE 27

27 relatively simple to prepare in house. This search led us to Baylis Hillman adducts as alkene reductase substrates. We chose three representative examples ( Figure 1 5) ( S ) 2 ( hydroxym ethyl ) cyclohexanone [ 4 ] is a very useful chiral building block 36,37 It has been prepared by using L amino acid s to catalyze the asymmetric aldol condensation of formaldehyde with cyclohexanone. The compound has been used to prepare the antispasmodic drug rociverine 38 the cancer therapeutic pyranicin 39 the diabetes treatment penarolide sulfate A 1 40 the plant pheromones cladospolide A C 41 the C. elegans pheromone daumone 42 and the anti viral glycolipid macroviracin D 43 While the reported aldol route to ( S ) 4 was successful, we reasoned that this p rocess could be improved upon with respect to stereoselectivity and reaction time by using an appropriate enzymatic asymmetric alkene reduction. The required precursor alkene, 2 (hydroxymethyl) cyclohex 2 enon e had been prepared previously 44 by using a base catalyzed Baylis Hillman reaction to add an hydroxymethyl group to cyclohex enone in good yield ( Figure 1 6 ) W e also chose a biocatalytic route to ( S ) 2 ( hydroxymethyl ) cyclopentanone 45 [ 5 ] since this material has been used to prepare t he fragrance compound jasmine lactone 45 insect pheromones 46,47 and the signaling compound leukotrie ne B 5 48 Like the cyclohexene derivative, 2 (hydroxymethyl) cyclopent 2 enon e could be prepared in a similar manner using Baylis Hillman chemistry 49 Continuing our search we then focused on the Baylis Hillman theme ( R ) 3 h ydroxy 2 methylpropanoate [( R ) 6 ] ( also known as Roche e ster) is a common chiral precursor in the synthesis of vitamins, fragrance compounds, antibiotics, and natural products 22 The Baylis Hillman adduct precursor [ 3 ] to Roche ester could also be

PAGE 28

28 prepared by a similar approach 50 This compound and its derivatives had been previously tested as substrates for OYE 1 by Stueckler et al 22 ( Figure 1 2e) Our goal was to use enzyme rather than substrate engineering to identify an alkene reductase route to both enantiomers of the target compound [ 6 ]. The proposed enzymatic reduction s of the Baylis Hillman compounds is outli ned in Figure 1 5 All three could be prepared by simple r eactions from inexpensive, commercially available compounds. With the exception of [ 3 ] hydroxymethyl functionalized substrates represent a new class of en ones not yet explored with OYE or its homologs. Additionally, all three are useful intermediates in the synthesis of a diverse range of attractive synthetic targets. Experimental Procedures Preparation of Baylis Hillman Adducts Dr. Bradford Sullivan conducted a large portion of the synthesis of these compounds in our lab. This work is summarized as follows. 2 (Hydroxymethyl) cyclohex 2 enone A solution of aqueous 37% formaldehyde (12.5 mL, 154 mmol), 2 cyclohexen 1 one (5.00 g, 52.0 mmol), and 4 dimethylaminopyr idine (6.35 g, 52.0 mmol) in THF (15 mL) was stirred at room temperature for 12 h. The reaction mixture was acidified to ca pH 5 with 1 M HC l and the aqueous layer was extracted with CH 2 Cl 2 (5 15 mL). The combined organic layers were washed with H 2 O ( 25 mL), brine (25 mL), and dried over Na 2 SO 4 The crude material was subjected to flash column chromatography with a solvent gradient of hexanes and ethyl acetate (3 : 1, 1 : 1, then 1 : 3) to yield the title compound as a pale yellow oil (5.12 g, 78% yie ld ). R f 0.40 (1 : 1 hexanes ethyl acetate); 1 H NMR (300 MHz, CDCl 3 ) 6.93 (t, J = 4.1 Hz, 1H), 4.20 4.25 (m, 2H), 2.77

PAGE 29

29 (s, br OH), 2.35 2.46 (m, 4H), 1.95 2.05 (m, 2H) ppm; 13 C NMR (75 MHz, CDCl 3 ) 200.8, 147.2, 138.4, 62.0, 38.4, 25.8, 22.9 ppm. Spectral data were consistent with those reported previously 44 2 (Hydroxymethyl) cyclopent 2 enone To a stirred solution of aqueous 37% formaldehyde (1.0 mL, 12 mmol) and cyclopent 2 enone (700 mg, 8.52 mmol) in THF (1 mL) was added imidazole (29 mg, 0.42 mmol). The resulting mixture was stirred at room temperature for 17 days. The reaction mixture was acidified to ca pH 5 with 1 M HCl and the aqueous layer was extracted with CH 2 Cl 2 (5 5 mL). The combined organic layers were washed with H 2 O ( 10 m L) and brine ( 10 mL), then dried over Na 2 SO 4 The crude material was subjected to flash column chromatography with a solvent gradient of hexanes and ethyl acetate (1 : 1, 1 : 2, then 1 : 5) to yield the title compound as a white solid (219 mg, 22% yield ). R f 0.35 (1 : 3 hexanes ethyl acetate); m.p. 68 70 C ; 1 H NMR (300 MHz, CDCl 3 ) 7.51 7.57 (m, 1H), 4.33 4.41 (m, 2H), 2.61 2.69 (m, 2H), 2.53 (s, br OH), 2.43 2.49 (m, 2H) ppm; 13 C NMR (75 MHz, CDCl 3 ) Spectra l data were consistent with those reported by Kar and Argade 49 Methyl 2 (hydroxymethyl) acrylate To a stirred solution of 37% aqueous formaldehyde (1.45 mL, 17.9 mmol) and freshly distilled methyl acrylate (2.10 mL, 23.2 mmol) in a mixture of H 2 O and 1,4 dioxane (1:1, 50 mL) was added 1,4 diazabicyclo[2.2.2]octane (1.70 g, 15.1 mmol) at room temperature. The resu lting mixture was stirred for 72 h then poured into saturated aqueous NaCl (50 mL). Th is was extracted with Et 2 O (5 10 mL) and the combined organic layers were washed with water (20 mL), brine (20 mL), and dried over Na 2 SO 4 The crude material was sub jected to flash column chromatography with a

PAGE 30

30 solvent gradient of petroleum ether and ethyl acetate (7 : 1, 4 : 1, then 1 : 1) to yield the title compound as a clear and colorless oil (1.54 g, 57% yield ). R f 0.34 (1 : 1 petroleum ether ethyl acetate); 1 H N MR (300 MHz, CDCl 3 ) 6.18 6.20 (m, 1H), 5.78 5.80 (m, 1H), 4.24 4.27 (m, 2H), 3.72 (s, 3H), 3.12 (s, br OH) ppm; 13 C NMR (75 MHz, CDCl 3 ) 166.9, 139.5, 125.9, 62.4, 52.1 ppm Spectral data were consistent with those reported by Drewes et al 50 Screening R eactions Screening assays were carried out in 0.3 mL volumes of 50 mM KP i pH 7.5, supplemented with 200 mM glucose, 25 U/mL GDH102 (Biocatalytics), 0.3 mM NADP + 10 mM substrate and ~100 g of GST affinity purified alkene reductase enzyme. Reactions were allowed to proceed for 24 hours while shaken at room temperature followed by extraction with 0.5 mL of ethyl acetate and subsequent analysis by GC Chiral phase GC analyses were carried Dex 225 column (Supelco) with He as the carrier gas and FID. For analysis of 1 2 4 and 5 the temperature program involved 140C (10 min) followed by a 20C / min increase to 200C (3 min). Under these conditions, peaks eluted at 10.2 min (( S ) 4 ), 10.7 min (( R ) 4 ), and 13.1 min ( 1 ) or 10.3 min (( R ) 5 ), 11.4 min (( S ) 5 ), and 13.2 min ( 2 ) For analysis of 3 and 6 the temperature program involved 100C (12 min) followed by a 20C / min increase to 180C (5 min). Under these condit ions, peaks eluted at 10.7 min (( S ) 6 ), 11.3 min (( R ) 6 ), and 11.8 min ( 3 ). Larger Scale P reparation of ( S ) 4 E. coli BL21(DE3) cells overexpressing OYE 2.6 ( Pichia stipitis ) were grown on LB agar plates supplemented with 0.2 mg/mL ampicillin. A single colony was then used to prepare a 40 mL starter culture in LB media with ampicillin that was shaken overnight at

PAGE 31

31 37 C. This pre culture was used to inoculate 4 L of LB media supplemented with 80 mL of 20% glucose, 2 g/L ampicillin, and 0.5 mL AF 204 (Sig ma). The cells were grown in a New Brunswick Scientific M19 fermenter at 37 C, 700 rpm, and 4 L/min airflow for 2.5 hours until the OD 600 = 0.8. At this time the temperature was d ecreas ed to 30 C and protein expression was induced with the addition of 0. 48 mL of 840 mM iPTG. Growth was continued for an additional 4 hours until the cells reached an OD 600 = 4.98. Cells were then harvested by centrifugation at 5000 rpm to yield 28.15 g (wet cell weight) and were stored at 20 C overnight. After thawing in 30 mL of 0.1 M phosphate buffer (pH 7.5), the slurry was passed through a French Pressure cell twice at 15,000 psi in the presence of 10 M PMSF. Insoluble material was then removed by centrifugation at 15,000 rpm for 45 min to yield 45 mL of crude cell lysate. When assayed against 1 under the conditions described for screening reactions this crude contained 75 U/mL of activity and was used as is. The enzymatic reduction of 1 was conducted in a 500 mL three neck round bottom flask stirred continuously at 300 rpm with a magnetic bar. The reaction media consisting of 90 mL of 0.1 M phosphate buffer (pH 7.5) containing 8.0 g of solid dextrose monohydrate was degassed under reduced pressure for 1 hour prior to use. It was then transferred to the reaction ve ssel and kept under argon. To this 1 mg (100 U) of GDH 102 (Biocatalytics), 10 mg NADPH (12 mol), and 450 U of freshly assayed crude OYE 2.6 from the previous step was added. This mixture was allowed to equilibrate for 15 minutes at room temperature pri or to the addition of substrate. 0.63 g (5.0 mmol) of neat 1 was then added to the reaction vessel in a single bolus. Reaction pH was maintained at 7.5 by adding 1 M KOH via a pH stat. Base demand was monitored until it dropped noticeably after 1.5 hour s. GC MS

PAGE 32

32 confirmed that at 1.75 hours that the reaction was approximately 100% complete The mixture was then extracted with 100 mL of CH 2 Cl 2 The aqueous layer was extracted twice more with 50 mL of CH 2 Cl 2 and the organic layers were combined and filte red over Celite to remove insoluble material. The filtered organic solution was then washed with brine and dried over Na 2 SO 4 The solve nt was then removed under vacuum to yield 0.61 g (4.8 mmol) of ( S ) 4 The observed e.e. was 91 % and [ ] D 25 = +10.42 (c = 4.03 in CHCl 3 ). Lit. [ ] D 2 2 +11.4 ( c 1.0, CHCl 3 ) 37 Larger Scale Preparation o f ( S ) 5 and Assignment of Absolute Configuration E. coli BL21(DE3) cells overexpressing OYE2.6 ( Pichia stipitis ) were gro wn on LB agar plates supplemented with 0.2 mg/mL ampicillin. A single colony was then used to prepare a 15 mL culture in LB media with ampicillin that was shaken overnight at 37 C. This growth was used to inoculate 500mL of LB media supplemented with 5 m L of 40% glucose, 2 g/L ampicillin. The cells were grown at 37 C while shaking at 200 rpm in a baffled flask for 2 hours until the OD 600 = 1.0. The temperature was d ecrease d to 30 C and protein expression was induced with the addition of 60 L of 840 mM iPTG. Growth was continued for an additional 4 hours until the cells reached an OD 600 = 3.2. Cells were then harvested by centrifugation at 5000 rpm to yield 3.48 g (wet cell weight) They were then resuspended in 300 mL of 50 mM phosphate buffer (pH 7.5) containing 0.1 M glucose. To this slurry 300 mg of solid 2 was then added. GC MS confirmed that at 1.5 hours that essentially all the starting material had been consumed The mixture was then stirred with 500 mL CH 2 Cl 2 at room temperatur e overnight. The heavily emulsified organic layer was filtered over Celite to remove insoluble material. The filtered organic solution was then washed with brine and dried over Na 2 SO 4 The

PAGE 33

33 solvent was then removed under vacuum to yield 129 mg of 5 with an observed e.e. of 75 %. The absolute configuration of the OYE 2.6 reduction product was assigned after Baeyer Villiger oxidation to the corresponding lactone. To a stirred suspension of the reduction product from 5 (65 mg, 0.57 mmol) and NaHCO 3 (72 mg, 0 .85 mmol) in CH 2 CH 2 (1.0 mL) was added 3 chloroperoxybenzoic acid (77% purity 153 mg, 0.683 mmol) at 0 C. The reaction mixture was slowly warmed to room temperature over 3 h and then quenched by adding solid Na 2 S 2 O 3 (50 mg) and stirring for an additional 15 min. The reaction mixture was dried with Na 2 SO 4 and the solids were removed via filtration through a small plug of Celite with CH 2 Cl 2 The crude material was subjected to flash column chromatography with a solvent gradient of pentane and ethyl acetat e (2:1, 1:1, then 1:5) to yield 59 mg (80% yield) of ( S ) 6 (hydroxymethyl)tetrahydro 2 H pyran 2 one as a clear and colorless oil. R f 0.20 (1:3 hexanes ethyl acetate); 1 H NMR (300 MHz, CDCl 3 ) 4.41 (tdd, J = 11, 5.5, 3.3 Hz, 1H), 3.78 (dd, J = 12.3, 3.1 Hz, 1H), 3.66 (dd, J = 12.3, 5.4 Hz, 1H), 2.55 2.66 (m, 1H), 2.38 2.52 (m, 1H), 1.67 2.03 (m, 4H) ppm; [ ] D 23 +26.5 ( c 2.50, CHCl 3 ); Lit D 25 +33 ( c 1.3, CHCl 3 ) 46 Spectral data were consistent with those reported by Lees and Whitesides 51 Results and Discussion Alkene Reductase Homologs Our Baylis Hillman substrates were first screened against a collection of alkene reductase homologs previously prepared in our lab 35 See Figure 1 7 for complete results. Remarkably, P. stipitis OYE 2.6 was the only clone to displa y both high conversion and strong enantioselectivity for the ( S ) enantiomer of products 4 5 and 6 at 95%, 76% and 95% respectively It should also be noted that 5 appears to slowly

PAGE 34

34 epimerize under these rea c tion conditions osure to aqueous reaction conditions can yield an enantiomeric excess of 95% ( Chapter 5). Selecting the best enzyme to deliver the ( S ) enantiomer was quite simple ; however, a good solution to obtaining the ( R ) enantiomers was not quite as obvious. In the case of the ( R ) product, t he OYE homolog collection gave no result for 1 For substrate 2 similar enantiomeric excess was observed in several clones ; however conversion for these reactions was limited to 50%. Finally, for 3 OYE 1 and OYE 2 displayed l imited conversion but also showed a strong preference for the ( R ) product. These results are consistent with those reported by Stueckler et al except in the case of OYE3, which they reported as their best enzyme for the reduction of the methyl ester (37 % conversion, >99% ( R )) 22 In summary, OYE 1 was a good enantiocomplementary catalyst for OYE 2.6 except for its poor conversion. We therefore turned to our Trp116 replacement library of OYE 1 variants to identify a solution to this problem. Tr y p tophan 116 Mutations Given its success in previous studies 18 the Baylis Hillman ad ducts were screened against the set of mutants at position 116 35 See Figure 1 8 for a summary of results. Surprisingly all modifications with the exception of Arg, Cys, Thr, a nd Pro afforded both good conversion and a strong preference for the ( S ) enantiomer of 1 while none of the mutants indicated any preference for the opposite enantiomer. Similarly, the same mutants yielded a dramatic improvement in conversion for substrat e 2 with the same enantiopreference. Of note, o nly wild type OYE 1 retained any preference for the ( R ) product. Even more interesting were the results for the reduction of 3 It became apparent that modifications at position 116 could yield the full sp ectrum of conversion and

PAGE 35

35 stereoselectivity These are graphically depicted in Figure 1 9 In pursuit of the ( R ) product the most suitable candidates are Val, Tyr, and to a lesser extent Phe. These mutants all give greater than 70% e.e ., which is signif icant ly less than the 95% observed with the wild type enzyme The observed loss in e.e. for these specific mutants is offset to a certain extent by an increase from wild type conversion (19%) of 2 3 times (37 68%). Conclusions and Future Work At this poin t in our investigation P. stipitis OYE 2.6 remains the clear choice to obtaining the ( S ) product s for the reduction s of all three Baylis Hillman adducts. In the case of the ( R ) products we found no good solution leading to 4 a partial solution to achie ve compound 5 and multiple partial solutions to 6 Future efforts described in this work will advance on the goal of obtaining the ( R ) products for the reduction of compounds 1 3 Logical routes to this goal are 1) to continue searching for homologs tha t will yield the desired result or 2) pursue an approach using enzyme engineering via an appropriate mutagenesis strategy Considering the results depicted in Figures 1 7 and 1 8 the most attractive route is mutagenesis simply due to its ability to de liver a wide variety results. However, the generation of rationally designed mutants one at a time requires a significant amount of human effort and may be no easier than cloning ho m ologs in terms of that effort. Before embarking on our protein engineeri ng studies, we had to determine the best starting point for mutagenesis. That is to say is it better to attempt to evolve enantiocompementarity using a single template like OYE 2.6 where a solution to ( S ) enantioselectivity already exists or is it bett solution s manifest in our OYE 1 mutants? Complicating this is our desire to evolve a

PAGE 36

36 solution for all three Baylis Hillman adducts in the same effort rather than conducting three separate and distin ct directed evolution projects. A more streamlined approach toward enzyme engineering is certainly in order and will be discussed and exploited in future chapters.

PAGE 37

37 Figure 1 1. Warburg and Christian reaction system Glucose 6 phosphate is oxidized by glucose 6 p hosphate dehydrogenase ( zwischenferment ) in the presence of NADP + ( coferment ) Old Yellow Enzyme ( gelbe ferment ) serves to regenerate NADP + by facilitating a 2 electron transfer from NADPH to molecular oxygen. 1

PAGE 38

38 Figure 1 2a. OYE1 substrates. In Figures 1 2a through 2 g substrates are introduced in chronological order and are presented only if identified as a new substrate for OYE 1. Reference compounds (ref.) where available, are indicated for relative comparison of activity, turn over number, or conversion within a given study. By default, u orientation such that the hydride hydrogen adds to the position from below the plane of the page and th e proton hydrogen adds to the position from above the plane of the page to yield the stereoselectivity analogous to that observed by Swiderska and Stewart 14 Where absolute stereochemi stry was determined (Figures 1 2d through 2 g) substrates are drawn in a binding mode that corresponds to the major pro duct enantiomer. In these cases, dashed lines indicate

PAGE 39

39 Figure 1 2b. OYE1 substrates.

PAGE 40

40 Figure 1 2c. OYE1 substrates.

PAGE 41

41 Figure 1 2d. OYE1 substrates.

PAGE 42

42 Figure 1 2e. OYE1 substrates.

PAGE 43

43 Fig ure 1 2f. OYE1 substrates.

PAGE 44

44 Figure 1 2g. OYE1 substrates.

PAGE 45

45 Figure 1 3. OYE substrate binding modes In both binding modes, the model substrate (cyclohex enone ) docks in the active site above and parallel to the plane of the reduced isoalloxazine r ing system. The carbonyl oxygen forms hydrogen bonds with the side chains of residues Asn194 and His191. The hydride from N5 is transferred (red arrow) to the electron deficient carbon. Tyr196 functions as a general acid to protonate (blue arrow) the resulting enolate at the theoretically requires a shift in the angle of hydride and proton transfer and is sterically crowded by the presence of Trp116. The stereochemistry of each binding mode product is indicated wi th the protic hydrogen (blue) and the hydride hydrogen (red) below each scheme. For a prochiral substrate the two binding modes determine product stereochemistry.

PAGE 46

46 Figure 1 4. OYE 1 active site crystal structure. Active site view of OYE 1 complexed w ith p hydroxybenzaldehyde (PDB: 1OYB) as viewed from the entrance of the active site. Ligands are depicted in stick format with carbons in grey. Selected residue side chains are labeled and depicted with carbons in green. Key hydrogen bond interactions are indicated by dashed yellow lines.

PAGE 47

47 Figure 1 5. Reaction scheme for the reduction of Baylis Hillman adducts

PAGE 48

48 Figure 1 6 Baylis Hillman re a ction mechanism

PAGE 49

49 Figure 1 7 Results of OYE homolog s c reening with Baylis Hillman adducts Fig ure 1 8 Results of W116X OYE1 screening with Baylis Hillman adducts All ee% listed are for the (S) enantiomer except where indicated in bold

PAGE 50

50 Figure 1 9 Results of W116X OYE1 screening with m ethyl 2 (hydroxymethyl) acrylate Select data from Figure 1 8 with e.e.% on the vertical axis and total conversion on the horizontal axis.

PAGE 51

51 CHAPTER 2 PREPARATIVE SCALE ENZYMATIC ALKENE RED UCTIONS Background To validate the utili ty of alkene reductases for industrial applications we desired to find model chiral com pounds th at have current relevance and production at industrial scale. A literature review revealed two interesting target s where biocatalytic asymmetric reductions might be useful W e therefore attempted to develop scalable methodologies for the product ion of ( R ) citronellal and ( R ) 2 methylpentanol using alkene reductases in the key steps ( R ) C itronellal Menthol is a valuable commodity that is consumed in the production of food, pharmaceut icals, health care products, cosmetics and a variety of syntheti c applications 52 While menthol has several isomers, the most desirable form is ( ) menthol for these applications 53 and a pproximately 12,000 metric ton s of are consumed annually 54 Of this amount, nearly 20 % is produced by synthetic methods whereas the remainder is isolated from plant species ( Mentha piperita ). The most common method for the production of sy nthetic menthol is the Takasago process (Takasago International Corporation). In this approach, myrcene is converted to ( R ) citronellal u sing a chiral Rh BINAP catalyst 55 The aldehyde intermediate is then cyclized using a heteroge neous zinc bromide catalyst followed by hydrogenation to produce pure ( ) menthol Previous work has shown that ( R ) citronellal production by OYE homologs is possible ; however the reported volumetric productivity was generally low ( 1.5 g/L/h ) and reaction times longer than 1 hour were not investigated 15 Preliminary st udies by our group revealed that OYE 2.6 18 introduced in the previous chapter, yields ( R ) citronellal

PAGE 52

52 with high optical purity when provided with the unsaturated precursor, gerania l. Commercially, geranial ( E isomer) is available as a component of citral, a mixture of geometric alkene isomers, which includes ~40% neral ( Z isomer) as a contaminant. Unfortunately t he presence of this contaminant decreases both the rate of reaction and the optical purity of the final product when the results of geranial reduction by a series of OYE homol ogs was tested 15 We proposed a reaction scheme ( Figure 2 1 ) that commenced with the commercially available geraniol in geometrically pure form Dehydrogeneases present in whole ce lls and crude cell extracts would facilitate the ox idation of geraniol to geranial. OYE 2.6 would then be used to reduce geranial to ( R ) ci tronellal. In this scenario NADPH would be recycled in a closed loop process. Unfortunately attempts to use this scheme with whole cell s and crude cell extracts overexpressing alkene reductases and an alcohol dehydrogenase were plagued by two problems. Wolken et al demonstrated that isomerization can be facilitated by the presence of free amino acids in a buffer sys tem 56 Some isomerization of the unsaturated aldehyde was ob served under our reaction conditions, resulting in a mixture of enantiomers. W e also observed that the dehydrogenase present in our crude enzyme preparations had a tendency to reduce the desired product further to the unsaturated alcohol, citronellol. Co mbined, these two factors result ed in a complex mixture where after 6 hours a representative trial consisted of 7% geraniol starting material, 26% geranial, 19% isomerization product (neral), 48% over reduction product (( R ) and ( S ) citronellol) and no ne o f the desired product (( R ) citronellal). This revealed that significant optimization would be needed before this approach was viable

PAGE 53

53 2 Methylpentanol We also became interested in developing a biocatalytic route to ( R ) 2 methylpentanol (( R ) 2MP) 57 because of its utility in the synthesis of multiple phar maceutical compounds and in the production of liquid crystal technology. A prev iously described strategy was based on an evolved ket oreductase (KRED) that catalyzed the reduction of an aldehyde at the expense of the NADPH co factor. In this case, t he evolved KRED selectively reduced the ( R ) enantiomer of racemic 2 methylvaleraldehyd e. Unreacted ( S ) 2 methylvaleraldehyde was then removed by treatment with sodium bisulfite to form a water soluble adduct that allowed the desired alcohol to be easily recovered. A volumetric productivity of 4.7 g/L/hour was reported (2 L scale) We reas oned that an alkene reductase could provide a viable route to ( R ) 2MP and avoid the need for a kinetic resolution In our plan 2 methylpentanal (available as the pure trans isomer) would be reduced by an appropriate alkene reductase to yield enantiomerica lly pure methylvaleraldehyde ( Figure 2 3). R eduction to the desired alcohol would take place in situ by the alcohol dehydrogenase activity endogenous to our crude cell extracts containing the overexpressed alkene reductase. Since both steps involve NADPH a cofactor regeneration scheme would be required. Experimental Strategy Preparation of Enantiomerically Pure ( R ) and ( S ) Citronellal Using OYE Homologs Our i nitial strategy to avoid interference between the dehydrogenase and alkene reduction involved co mpartmentalizing the two reactions The idea was to generate the geranial in situ in one compartment (NADP + favored) but allow alkene reduction to

PAGE 54

54 proceed only in a physically separated compartment (NADPH favored) ( Figure 2 2 ). W e first worked to optimize OYE 2.6 mediated alkene reduction GST affinity purified OYE 2.6 and the c ommercially available geranial source, citral containing a mixture of isomers was used We discovered that the enzyme displays instability in the presence of the aldehyde substra tes and products To simplify the process further we prepared geometrically pure geranial for subsequent work. We initially attempted to compartmentalize the reaction system using a polydimethylsiloxane (PDMS) membrane as describe by Mwangi et al 58 This approach was complicated by the tendency for the PDMS material to absorb both starting material and product dramatically retar ding reaction progress Given the observation that the PDMS membrane appeared to function as a second phase we proceeded with investigation of solvent/aqueous biphasic reactions. Our initial aim was to use a biphasic reaction system as a means to limit OY E 2.6 exposure to substrate and product When this failed t o address the issue of enzyme stability we explored the use of cross linked enzyme aggregates ( CLEA s ) since these have been shown to improve enzyme stability in both single and multi enzyme syste ms 59 General protocols for the formation of enzyme aggregates require precipitation of the enzyme (or enzyme/protein mixture) using ammonium sulfate followed by cross linking of this material with glutaraldehyde or similar bi functionalized reag ent. Presumably, cross involving the exposed lysine residues displaying a free NH 2 group, on the protein surface and the cross linking agent. The resulting aggregate retains enzymatic activity in its solid form and due to its nature can be regenerated by simple separation and

PAGE 55

55 washing steps. Similarly and by the same logic, we pursued protection by acetylation of the soluble enzyme. The results of these isolation and protection efforts are describ ed below. When the above mentioned strategies proved unsatisfactory we adopted the most straightforward approach. This involved chemical oxidation of geraniol or nero l followed by alkene reductase mediated conversions to yield the desired saturated aldeh yde enantiomer s The alkene reductases were supplied as partially purified lysates from E. coli cells overexpressing OYE 2.6 ( P stipitis ) or NemA ( E. coli ) respectively. 60 We also found that reaction progress could be monitored by the use of a pH stat where base demand is tied stoichiometrcally to the mole s of gluconolactone formed by the NADPH recycling system ( Figure 2 2) Preparation of ( S ) 2 Methylpentanol Using OYE2 and YahK Process development in this case was more straightforward Commercially available 2 methyl 2 pentenal was sequentially reduced t o ( S ) m ethylvaleraldehyde by OYE 2 ( S. cerevisiae ) and then to ( S ) 2 m ethylpentanol by cloned alcohol dehydrogenase, E. coli YahK in a single pot reaction ( Figure 2 3 ) Glucose dehydrogenase (GDH) prepared from a donated plasmid pTgluDH3 (Biocatalyics) a nd processed as a crude cell lysate, wa s used to re generate reducing equivalents (NADPH) for the reaction. Experimental Procedures GC/MS A nalysis GC/MS analyses w ere performed on a Chirasil Dex CB column (25 m 0.25 mm) using a mass selective detector (EI 70 eV). The temperature program for the reduction of citral reductions involved 60C (2 min) followed by a 10C/ min increase to 180C (10

PAGE 56

56 min). Under these conditions, peaks eluted at 10.5 min (citronellal), 11.8 min (neral), 12.2 min (geranial) and 12.5 min (cintronellol). Chiral analyses for the reduction of citral were performed on a Dex 225 column (30 m 0.25 mm). The temperature program involved 95C (35 min) followed by a 5C/ min increase to 160C (2 min), then a 10C/ min increase to 200C (5 min). Under these conditions, peaks eluted at 26.6 min (( S ) citronellal), 27.0 min (( R ) c itronellal), 36.6 min (( S ) citronellol), 36.9 min (( R ) citronellol), 43.4 min (neral) and 46.2 min (geranial). Chiral analyses for the production of 2 methylpentanol were performed on a Chirasil Dex CB column (25 m 0.25 mm) using a mass selective detec tor (EI, 70 eV). The temperature program involved 65C (3 min) followed by a 1C/min increase to 75C (2 min), then a 5C/min increase to 120C (2 min), and completed with a 20C/min increase to 200C (5 min). Under these conditions, peaks eluted at 5.6 min (2 methylvaleraldehyde), 7.8 min (2 methyl 2 pentenal), 18.2 min (( R ) 2 methylpentanol) and 18.4 min (( S ) 2 methylpentanol). PDMS Membrane Preparation Polydimethylsiloxane (PDMS) reaction tubes were prepared using Sylgard 184 Silicone Elastomer Kit ( Dow Corning). The elastomer working solution (approximately Approximately half of the solution was applied by transfer pipet to a 0 .5 inch diameter brass rod while being turned on a hori zontal rotisserie at room temperature. After turning for 30 minutes the rotisserie was transferred to a 65C incubator and allowed to cure for one hour while turning. The single coat film was removed from the incubator and allowed to cool to room tempera ture before applying the remaining elastomer solution in the same

PAGE 57

57 manner. The rod was then turned overnight at room temperature followed by final curing at 65C for one hour. The two layer film was removed from the rod by placing the cooled rod in a grad uated cylinder containing hexane. The solvent swollen tube was then air dried to return to its original shape and trimmed to a desired length. The tubes were then sealed at one end by standing them on end in a thin layer fresh elastomer solution and cure d for two hours at 65C followed by trimming as necessary. Small scale Test Reactions Reaction conditions for preparative scale citronellal production were optimiz ed at an 8 mL working volume. These reactions were conducted in a 25 mL three neck round bot tom flask fitted with a pH probe and connected to an auto burette delivering 1M KOH to maintain a static pH of 7.5. Reactions were conducted in phosphate buffer (100 mM) supplemented with glucose (200 mM) and NADP + (0.2 mM), which were degassed and purged with argon prior to adding enzyme. When investigating biphasic reaction protocols an 8 mL volume of hexanes was introduced to the vessel after degassing. OYE 2.6 GST (1 mg/mL) and GDH 102 (0.25 mg/mL, Codexis) were then added to the stirred reaction medi a and allowed to equilibrate for 15 minutes before adding substrate (50% v/v in ethanol). Alternatively, the same enzyme loading was also delivered as a CLEA (see below). Reaction progress was monitored by base addition co nfirmed periodically by GC MS. C LEA P reparation Purified GST OYE 2.6 (17 mg, 0.50 mL) was mixed with purified GDH 102 (6 mg) and 0.50 mL of 100 mM KP i pH 7.5. A 0.50 mL aliquot was transferred to a microcentrifuge tube and 0.50 mL of saturated (NH 4 ) 2 SO 4 solution was added. The tube was rotated gently at 4C for 15 min, then 384 mg of solid (NH 4 ) 2 SO 4 was added the

PAGE 58

58 aqueous solution) was added to a final concentration of 75 mM and the tube was rotated gentl y at 4C for an additional 2 hr. The CLEA was collected by centrifugation and washed three times with cold 100 mM KP i pH 7.5. Acetylation of GST OYE 2.6. Four microcentrifuge tubes containing 4 mg of purified GST OYE 2.6 in a volum e of 0.50 mL were prepar ed (1 4). Sodium acetate (190 mg, 50% aqueous solution) was tubes were gently rotated at 4C for 1 hr, then samples 2 4 were dialyzed against 100 mM KPi, pH 7.5, 50 mM NaCl, 50% glycerol for 3 hr. Al 4 were transferred to microcentrifuge tubes and citronellal was added to a final at 4C. Each sample, along wit h a control that was not treated with citronellal was diluted 1 : 10 with 100 mM KP i pH 7.5 and assayed for protein concentration and catalytic activity. Activity was measured by incubating 2.5 mM 2 cyclohexenone (from a 1 M EtOH stock solution) with 100 mM KP i pH 7.5 that contained 0.2 mM NADPH and an appropriate quantity of protein in a total volume of 1.0 mL. The change in A 340 was measured at 25C. The small change in A 340 observed during the same time period in a control reaction lacking 2 cyclohexen one was subtracted from all of the determinations. Preparation of ( R ) C itronellal Preparation of g eranial A slurry of 5.25 g of geraniol (98%, Sigma) (33 .4 mmol), 10.85 g of activated m anganese (IV) oxide (Fluka) (125 mmol), and 30 mL of methylene chloride was mixed and stirred at room temperature. After 24 hours a 0.2 mL sample was diluted to 1.5 mL

PAGE 59

59 and passed through a silica plug to remove any solid oxidant. This sample was determined to be approximately 70% oxidized when analyzed by GC MS. An additi onal two equivalents of oxidant (63 mmol) was added and the reaction was allowed to proceed overnight. When analyzed at 46 hours in the same manner all geraniol had been oxidized with the resulting product being a 96% geranial and 4% neral mixture. The reaction slurry was then passed over a silica pad and the solvent was removed under vacuum to yield 4.72 g (89% yield) of bright yellow oil. This product was stored at 20 C under argon until used. Preparation of OYE 2.6 catalyst E. coli BL21(DE3) cells overexpressing OYE 2.6 ( P stipitis ) were grown on LB agar plates supplemented with 0.2 mg/mL ampicillin. A single colony was then used to prepare a 40 mL culture in LB media with ampicillin that was shaken overnight at 37 C. This pre culture was used t o inoculate 4 L of LB media supplemented with 80 mL of 20% glucose, 2 g/L ampicillin, and 0.5 mL antifoam AF 204 (Sigma). The cells were grown at 37 C, 700 rpm, and 4 L/min airflow for two hours until the OD 600 = 0.6. At this time the temperature was dec reased to 30 C and protein expression was induced with the addition of 0.48 mL of 840 mM iPTG. Growth was continued for an additional 4 hours until the cells reached an OD 600 = 4.15. Cells were then harvested by centrifugation at 5000 rpm to yield 28.18 g wet cell weight and were stored at 20 C until the next step. Cells from the previous growth were thawed at room temperature in 30 mL of 0.1M phosphate buffer (pH 7.5). The slurry was passed through a French Pressure cell twice at 15,000 psi in the pre sence of 10 M PMSF. Inso luble material was then

PAGE 60

60 removed by centrifugation at 15,000 rpm for 45 min to yield 44.5 mL of crude cell lysate. To this 15 mL of room temperature saturated (NH 4 ) 2 SO 4 was added slowly to bring the solution to 25% saturation and was then allowed to equilibrate for 10 minutes. Insoluble proteins were then removed by centrifugation at 15,000 rpm for 45 minutes to yield 55 mL of protein solution. To this 5.0 g of solid (NH 4 ) 2 SO 4 was slowly added while stirring and was allowed to equilibrate for 10 minutes. Insoluble proteins were separated by centrifugation and then redisolved in 25 mL of 0.1 M phosphate buffer and centrifuged once more to yield 29 mL of purified protein. When assayed this fraction contained 7.9 U/mL activity an d approximately 47% of the total activity in the original crude preparation. Aliquots of this preparation were stored at 20 C until needed. Enzymatic reduction of geranial to ( R) citronellal The enzymatic reduction of geranial was conducted in a 500 mL t hree neck round bottom flask stirred continuously at 300 rpm with a magnetic bar. The reaction media consisting of 85 mL of 0.1 M phosphate buffer (pH 7.5) containing 8.0 g of solid dextrose monohydrate was degassed for 1 hour prior to use. It was then t ransferred to the reaction vessel and kept under argon. To this 1 mg (100 U) of GDH 102 (Biocatalytics), 10 mg NADPH (12 mol), and 101 U of thawed and freshly assayed OYE 2.6 from the previous step was added. This mixture was allowed to equilibrate for 15 minutes at room temperature prior to the addition of substrate G eranial (2.38 g) dissolved in 2.5 mL of EtOH was then divided into three portions and added to the r eaction vessel upon initiation and then at 1.5 and 3 hours. Reaction pH was maintaine d at 7.5 by adding 1 M KOH via a pH stat. Base demand was monitored until it dropped to approximately maximum and 16.5 mL of base had been added by 5.5

PAGE 61

61 hours. After 5.75 hours, GC MS analysis confirmed that the reaction was 95% complete. T he mixture w as acidified with 1M HCl to pH 4 and extracted with 100 mL of CH 2 Cl 2 and allowed to stir overnight. The aqueous layer was extracted twice more with 50 mL of CH 2 Cl 2 and the organic layers were combined and filtered over Celite to remove insoluble material. The filtered organic layer was then washed three times with brine and dried over Na 2 SO 4 The resulting orange/brown solution was passed over a bed of silica resulting in a bright yellow solution. The solvent was then removed under vacuum to yield 2.12 g of yellow oil consisting of 91% citronellal when analyzed by GC MS. This product was then loaded onto 5 g of silica deactivated with 10% H 2 O and then onto a 60 g silica column equilibrated with hexanes. The product was eluted with 1:9 Et 2 O/hexanes. Fr actions containing citronellal as indicated by GC MS were pooled and the solvent was removed under vacuum. The final yield was 1.59 g (10.2 mmol) of a pale yellow liquid, 99% pure by GC MS, with an e.e. of 98%, [ ] D 24 = +18.22 (c = 7.30 in CHCl 3 ). Lit. [ ] D 24 = +16.2 (c = 1.0 0 in CHCl 3 ) 61 Preparation of ( S ) C itronellal Preparation of n eral A slurry of 21.0 g of nerol (97%, S igma) (1 36 mmol), 43.40 g of activated m anganese (IV) oxi de (Fluka) (500 mmol) and 150 mL of methylene chloride was prepared and stirred at room temperature. After 22 hours a 0.2 mL sample was diluted to 1.5 mL and passed through a silica plug to remove any solid oxidant. GC MS analysis showed that 10% of the initial starting material was still present. An additional 125 mmol of oxidant was added and the reaction was allowed to proceed for four more hours. After 27 hours the reaction slurry was passed over a silica pad and the solvent was removed under vacuum to yield 20.30 g (98.2% yield) of yellow oil. GC MS analysis showed the final

PAGE 62

62 product mixture contained 9 7% neral and 3% geranial This product was stored at 20 C under argon until used. Prepar ation of NemA catalyst E. coli BL21(DE3) cells overexpressing the OYE homolog NemA ( N e thylmaleimide reductase E. coli ) were grown on LB agar plates supplemented with 0.2 mg/mL ampicillin. A single colony was then used to prepare a 40 mL culture in LB me dia with ampicillin that was shaken overnight at 37 C. This growth was used to inoculate 4 L of LB media supplemented with 80 mL of 20% glucose, 2 g/L ampicillin, and 0.5 mL antifoam AF 204 (Sigma). The cells were grown at 37 C, 700 rpm, and 4 L/min airf low for 2.5 hours until the OD 600 = 0.8. At this time the temperature was decreased to 30 C and protein expression was induced with the addition of 0.48 mL of 840 mM iPTG. Growth was continued for an additional 4 hours until the cells reached an OD 600 = 4.98. Cells were then harvested by centrifugation at 5000 rpm to yield 28.15 g wet cell weight and were stored at 20 C overnight. After thawing at room temperature in 30 mL of 0.1M phosphate buffer (pH 7.5 ) the slurry was passed through a French Pressur e cell twice at 15,000 psi in the presence of 10 M PMSF. Insoluble material was then removed by centrifugation at 15,000 rpm for 45 min to yield 45 mL of crude cell lysate. When assayed this crude con tained 44 U/mL activity and was used as is. Enzymatic reduction of neral to ( S) citronellal The enzymatic reducti on of neral was conducted in a 500 mL three neck round bottom flask stirred continuously at 300 rpm with a magnetic bar. The reaction media consisting of 85 mL of 0.1 M phosphate buffer (pH 7.5) containing 8.0 g of solid

PAGE 63

63 dextrose monohydrate was degassed for 1 hour prior to use. It was then transferred to the reaction vessel and kept under argon. To this 1 mg (100 U) of GDH 102 (Biocatalytics), 10 mg NADPH (12 mol), and 484 U of freshly assayed crude NemA from the previous step was added. This mixture was allowed to equilibrate for 15 minutes at room temperature prior to the addition of substrate. 2.35 g of prepared neral was mixed with 2.5 mL of EtOH was then divided and added to the reaction vessel in equal portions upon initiation and then at 1.5 and 2.5 hours. Reaction pH was maintained at 7.5 by adding 1 M KOH via a pH stat. Base demand (1 M KOH) was monitored until it dropped noticeably after 13.5 mL of base had been added by 4 hours. GC MS confirmed that at 3.5 hours that the reaction was approximately 100% converted. The mixture was then acidified with 1M HCl to pH 4 and extracted with 100 mL of CH 2 Cl 2 and allowed to stir overnight. The aqueous layer was extracted twice more with 50 mL of CH 2 Cl 2 and the organic layers were combined and filtered over Celite to remove insoluble material. The filtered organic solution was then washed three times with brine and dried over Na 2 SO 4 The resulting pale yellow solution was passed over a bed of silica resulting in a pale yellow solution. The solvent was then removed under vacuum to yield 1.62 g (10.3 mmol) of pale yellow oil consisting of 98% citronellal when analyzed by GC MS. The observed e.e. was >99% and [ ] D 24 = 13.71 (c = 5.55 in CHCl 3 ). Lit. [ ] D 24 = 16.2 (c = 1.0 0 in CHCl 3 ) 61 Preparation of ( S ) 2 M ethylpentanol Preparation of OYE2 and YahK c atalyst E. coli BL21( DE3) cells overexpressing OYE 2 ( S. cerevisiae ), YahK ( E. coli ), and GDH (pTgluDH3, Biocatalyics) were grown separately on LB agar plates supplemented with 0.2 mg/mL ampicillin. The same procedure was used for all three overexpression

PAGE 64

64 strains. A single c olony was then used to prepare a 40 mL culture in LB media with ampicillin and this was shaken overnight at 37 C. The pre culture was used to inoculate 4 L of LB media supplemented with 80 mL of 20% glucose, 2 g/L ampicillin, and 0.5 mL antifoam AF 204 (S igma) in a New Brunswick Scientific M19 fermenter The cells were grown at 37 C, 700 rpm, and 4 L/min airflow for two hours until the OD 600 ~ 0.6. At this time the temperature was decreased to 30 C and protein expression was induced with the addition of 0.48 mL of 840 mM iPTG. Growth was continued for an additional 4 hours until the cells reached an OD 600 ~ 4. Cells were then harvested by centrifugation at 5000 rpm to yield approximately 7 g/L w.c.w. and were stored at 20 C until the next step. Cell p ellets were thawed in 0.1M phosphate buffer (pH 7.5). The slurries were passed through a French Pressure cell twice at 15,000 psi in the presence of 10 M PMSF. Insoluble material was them removed by centrifugation at 15,000 rpm for 45 min to yield crude cell lysates which were then divided into aliquots and either used immediately or stored at 20 C until needed. Prior to use, activity was estimated by measuring the decrease in A 340 at 25 C in pH 7.5 phosphate buffer (100 mM) containing NADPH (0.2 mM) a nd 10 mM of the appropriate substrate ( 2 methyl 2 pentenal 2 methylvaleraldehyde, or glucose). Enzymatic reduction of 2 methyl 2 pentenal to ( S ) 2 m ethylpentanol The enzymatic reduction of 2 methyl 2 pentenal was conducted in a 500 mL three neck round bot tom flask stirred continuously at 300 rpm with a magnetic bar. The reaction media consisting of 90 mL of 0.1 M phosphate buffer (pH 7.5) containing 8.0 g of solid dextrose monohydrate was degassed for 1 hour prior to use. It was then

PAGE 65

65 transferred to the r eaction vessel and kept under argon. To this 1 mL (300 U) of GDH lysate, 10 mg NADP + (13 mol), and 325 U of freshly assayed OYE 2 lysate from the previous step was added. This mixture was allowed to equilibrate for 30 minutes at room temperature prior to the addition of substrate. 2 methyl 2 pentenal (0.30 g, Sigma) was then added to the reaction vessel. Reaction pH was maintained at 7.5 by adding 1 M KOH via a pH stat. B ase demand indicated that approximately half of the substrate had been reduced a fter 1 hour At that time 1.5 mL YahK (122 U) was added to the reaction to initiate reduction to the final product. A second 0.30 g aliquot of neat 2 methyl 2 pentenal was added at 2 hours when base demand indicated that approximately 90% of the initial bolus had been consumed. Base demand was monitored until it dropped to approximately half the maximal value and a total of 15.5 mL of base had been added by 5.5 hours. The reaction was then acidified with 1M HCl to pH 4 and stirred overnight with 100 mL C H 2 Cl 2 The aqueous layer was extracted twice more with 100 mL of CH 2 Cl 2 and the organic layers were combined and filtered over Celite to remove insoluble material. The filtered organic layer was then washed once with 1M NaHCO 3 and twice with brine and dr ied over Na 2 SO 4 The solvent was then removed under vacuum to yield 0.37 g (59% yield) of a pale yellow oil consisting of ~98% 2 methylpentanol when analyzed by GC. Analysis of the final crude product indicates an e.e. of 99.6%, [ ] D 24 = 11.10 (c = 4.30 in CHCl 3 ). Lit. [ ] D 20 = 8.2 (c = 1.48 in CHCl 3 ) 62 Results and Discussion ( R ) and ( S ) Citronellal Attempts to use a biphasic system led to the conclusio n that substrate partitioning severely limi ted the overall reaction rates by effectively reducing the concentration of

PAGE 66

66 su bstrate available to the enzyme, which remained in the aqueous phase Figure 2 4 shows th e reaction time course for the reduction of citral (3 : 2, geranial : neral ) in a two ph ase system The data indicated that OYE 2.6 preferentially reduces geranial over neral. Because it was not clear whether the decrease in neral level wa s due to isomerization to geranial or by direct reduction of neral to citronellal geometrically pure g eranial was substituted for citral. Figure 2 5 shows the relative reaction progress of the biphasic reaction compared to a single phase aqueous reaction. Both r eactions appear to stop after approximately 12 hours Surprisingly, the single phase aqueous r eaction yielded approximately 2 fold higher product titer than the 2 phase strategy. We therefore carried out all subsequent studies under single phase conditions Given that the enzyme tended to inactivate after exposure to the aldehyde we pursued prepa ration of CLEAs. The assumption was that the immobilized aggregate would prove more stable in the presence of our aldehyde substrates. This was based on the fact that the cross linking agents were also aldehydes and might therefore pre condition the enzy me prior to substrate addition. We found that the reaction rate for aggregate enzymes was dramatically reduced for our system ( Figure 2 5 ) and displayed approximately a 4 fold decrease in activity from that of the free enzyme experiments in both single a nd bi phase approaches Acetylation was investigated as an approach to enhance enzyme stability without the decreased reaction rates associated with CLEA immobilization We used the acetylation protocol described by Riordan and Vallee 63 and discovered that the diminished activity for the enzyme yielded no apparent protection from aldehyde

PAGE 67

67 exposure Table 2 1 shows that after the acetylation protocol, both the un treated and acetylated enzyme ( NaOAc, AcO 2 ) preparations saw a ~16% reduction in specific activity a fter 24 hour incubation with citronellal Theref ore we concluded no benefit was obtained by acetylation and further attempts to functionalize the enzyme w ere abandoned. Based on these preliminary results, the final opt imized processes for producing ( R ) and ( S ) c itronellal utilized unmodified alkene reductases and aqueous reaction mixtures in a single compartment Our final reaction schemes (Figure 2 6 a nd 2 7) deviated from our original concept (Figure 2 2) of a compartmentalized reaction scenario. In this approach, alcohol oxidation and alkene reductions were separated both spatially and temporally. The productivity of these schemes are listed in the T able 2 2 and are consistent with general requirements for economically feasible industrial biocatalysis 3 Under our react ion conditions we observed very little geometric isomerization of th e substrate, geranial or neral. This eliminated the previously assumed requirement to produce geranial or neral in situ Ultimately our solution to the carbonyl over reduction problem w as to eliminate dehydrogenase activity by using simple ammonium sulfate fractionation of the crude cell lysate prior to alkene reduction While the volumetric pro ductivities listed in Table 2 2 certainly are capable of improvement through process design they are further limited by enzyme exposure to a total aldehyde concentration of ~ 200 mM. ( S ) 2 Methylpentanol Screening against our available alkene reductase library revealed that no enzyme could afford the ( R ) enantionmer of the desired target On the other hand, OYE 2 from Saccharomyces cerevisiae yield ed high enantiomeric excess for the ( S ) enantiomer To

PAGE 68

68 accomplish the second carbonyl reduction we introduced a crude preparation of a previously cloned alcohol dehydrogenase, YahK ( E. coli ). The YahK enzyme displayed high activity toward the primary aldehydes in this study and could be used to enhance the rate of secondary reduction as needed in the reaction scheme ( Figure 2 3 ). This process was complicated by two phenomena. First, the intermedi ate methylvaleraldhyde produced in the reaction demonstrated the tendency to racemi ze under reaction conditions. Secondly, the YahK dehydrogenase had a 3 : 1 preference for reducing the saturated aldehyde compared to the carbonyl of the unsaturated startin g material; however, at lower valeraldehyde concentrations the enzyme would alternatively reduce the pentenal starting material to produce quantities of the undesired allylic alcohol. As described in the experimental section, the addition of the YahK enzy me was delayed until approximately 50% of the initial substrate bolus had reacted. This served to avoid both intermediate reacemization and staring material reduction. As with the citronellal example, the total aldehyde exposure proved problematic and th u s limited s ubstrate additions to ~60 mM. The productivity of this reaction scheme was determined to be 0.67 g/L/h, significantly less than the approach published described by Gooding et al 57 but was capable of yielding pure product with high enantiomeric excess (Table 2 2 ). Conclusion Using simple reaction schemes we were able to demonstrate the applicability of alkene reductases in preparation of valuable synthetic intermediates. With the exception of the 2 methylpentanol target the reaction schemes developed in this study show sufficiently high volumetric productivity to be economically feasible 3 Aldehyde targets continue to be challenging substrates due to their apparent toxicity toward alkene

PAGE 69

69 reductase enzymes. Future applications of alkene re ductases in the production of enantiomerically pure primary aldehydes and alcohols could be greatly enhanced with the development of more substrate tolerant enzyme variants

PAGE 70

70 Figure 2 1. Initial proposed scheme for the preparation of ( R ) citronellal Ger aniol starting material is available from Sigma Aldrich at >97% purity and $ 0.06 per gram. Enzymes are used as crude cell lysates after protein expression in E. coli BL21 (DE3). Exogenous NADP + cofactor is added to a concentration of 0.2 mM. Dashed line circles indicate observed undesired side reactions.

PAGE 71

71 Figure 2 2. Proposed 2 compartment scheme for avoiding undesired side reactions NADPH oxidase from L. sanfranciscensis is used to regenerate oxidized cofactor without production of peroxide 64 All organic species are able to migrate across the indicated membrane (dashed line). In the reducing compartment, reduction of the unsaturated aldehyde is driven by the irreversible formation of gluconolactone.

PAGE 72

72 Figure 2 3. General scheme for the p reparation of ( S ) 2 Methylpentanol Figure 2 4. Time course for OYE 2.6 reduction of citral under biphasic conditions Quantities determined by GC are plotted as mmoles present (geranial, ; neral, ; citronellal, ) or mmoles consumed (glucose, ). The re action was carried out in a 1 : 1 mixture of hexanes and 100 mM KP i pH 7.5 using purified GST OYE 2.6 and GDH at room temperature.

PAGE 73

73 Figure 2 5. Comparison of strategies for OYE 2.6 mediated reduction of geranial Quantities determined by GC are plotted as mmoles of citronellal formed per unit of GST OYE 2.6 activity. The enzyme was employed as free protein in buffer ( ) or 1 : 1 mixture of hexanes and buffer ( ) or as a CLEA in buffer ( ) or a 1 : 1 mixture of hexanes and buffer ( ).

PAGE 74

74 Table 2 1 Acetylation of OYE 2.6 untreated NaOAc Ac 2 O NaOAc, Ac 2 O not incubated 1.75 0.17 1.43 0.08 0.12 0.0 1 1.02 0.0 4 incubated 1.47 0.09 1.46 0.14 0.26 0.02 0.84 0.04 (25mM citronellal) Specific activity of enzyme preparations are listed in moles of cyclohexene 2 one reduced per minute per mg of protein as measured by the decrease in the A 340 pea k of NADPH cofactor. The acetylation protocol used requires the addition acetic anhydride to the protein in the presence sodium acetate. Samples treated with only sodium acetate or acetic anhydride were run as controls. Following treatment, samples were incubated with 25 mM aldehyde product where indicated, f or 24 hours prior to specific activity assays. Table 2 2. Productivity of alkene reductase biotransformations Productivity Product e.e. % g product/L/h mol product/min/g d.c.w. ( R ) citronellal 98 2.8 17 ( S ) citronellal >99 4.1 25 ( S ) 2 methylpentanol >99 0.67 6 Dry cell weight (d.c.w.) of alkene reductase/dehydrogenase catalyst is calculated using 0.23 conversion factor from wet weight 65

PAGE 75

75 Figure 2 6. General scheme for the preparation of ( R) citronellal Figure 2 7. General sche me for the preparation of ( S) citronellal

PAGE 76

76 CHAPTER 3 PROTEIN ENGINEERING TO EVOLVE ENANTIOSEL ECTIVITY Introduction O ur interest in advanced protein engineering techniques is fundamentally a proof of principle endeavor While approaches to engineering va ry from rational to non rational we have observed the power and potential of a semi rational approach described in Chapter 1 and elsewhere. 18 The next logical step in complexity is that of directed evolution which by definition potentially requires successive combinations of beneficial mut ations When one considers the sheer magnitude of effort typically expended on these iterative cycles one can easily be dissuaded We did not attempt to to use dir ected evolution as a tool to explore structure function relationships in alkene reductases. Directed E volution Evol ution in nature has really one ultimate st r ategy: survival. However in directed or laboratory evolution of proteins the researcher controls the process and strategies leading to the desired outcome. In vitro protein engineering has become increasingly common over the past 20 years (Figure 3 1 ). Typically the objectives of these efforts are thermo stability or solvent tolerance, however, en antioselectivity or stereoselectivity have become increasingly common goals in recent years. Protein engineering in the laboratory requires a sound strategy for success. Manipulating proteins requires consideration of the following facts: 1 ) Protein seque nce space is large : with 20 amino acids there are 20 300 different ways to assemble a 300 residue enzyme Add to that the three dimensional nature of protein structure and the

PAGE 77

77 possible number of combinations approach the infinite. 2 ) Protein space is most ly empty of function. 3 ) The prospect of finding c ombinations of beneficial mutations are extremely rare. 66 The t wo basic approaches to protein eng ineering are rational design and d irected evolution ( Figure 3 2 ) In cases where much is known about the protein or enzyme structure and function rational design can yield significant results. 67 Rational design in this case is defined a s the introduction of a deliberate mutation for the purpose of changing the property of a given enzyme. In cases where little is known about the enz yme or how to evolve the desired property directed evolution is often the preferred solution. Directed e volution can be summarized as successive rounds of mutational pressure on a desired gene followed by subsequent screening that allows for selection of an improved gene product. The two most critical phases of a directed evolution effort are the generation of sufficient diversity so that a beneficial change can occur and a screening or selection strategy that uncovers these beneficial mutation s The following sections will outline existing approaches to mutagenesis and screening as applied to directed evol ution. Generating D iversity After having identified the target protein of interest the first step in directed evolution is creating the diversity N umerous approaches have been applied to this task 68 In general, they can be classified into three broad categories: recombi native (random) non reco mbinative (random) and semi rational approaches.

PAGE 78

78 Recombinative methods Recombinative approaches, often termed sexual evolution because of their attempt to mimic combination of two parent genes as in nat ure 66 involve various in vitro techniques based on the principle of the original method of DNA shuffling. DNA shuffling generate s large libraries of potentially beneficial mutations. The original app roach by Stemmer involved random fragmentation of a double stranded parent gene using DNaseI 69,70 followed by PCR amplification of resulting fragments both with and without primers for the gene. The result was a reassembled gene with a r elatively high frequency of point mutations due to various conditions of the PCR reaction. 71 When a single gene is subject to DNA shuffling the diversity generated is the result of these point mutations. It may also be desired to shuffle homologous genes in an attempt to combine properties of multip le parent genes. In this approach point mutation s ma y mask or complicate the effort, however, optimization to reduce these un intended mutations is possible. 71 The staggered extension process (S tEP) is an alternate method for random combinations of multiple closely related parent genes. 72 In this approach a single primer is introduced into a pool of homologous genes and subject to annealing and a significantly abbreviated extension step. The partially extended primer is then re annealed to one of the parent templates i n the pool and the process repeated. The result is a fully extended gene that is a random mixture of all the parents in the pool. Similar to these approaches is a technique called random chimeragenesis on transient templates (RACHITT). 73 In RACHITT a pool of single strand homologous genes is partially digeste d by DNaseI The fragments are then annealed to a homologous single strand template not included in the original digestion pool. Non

PAGE 79

79 Taq and Pfu DNA polymerases as well as gaps filled between the annealed fragments. The strand is ligated and then cloned into an appropriate vector. The crossover is the junction point in a reassembled sequence where a template switch takes place from one parent sequence to another 74 Crossovers are influenced by the homology and the size of the fragment annealing to the template The RACHITT strategy has the benefit of much higher number of crossovers per gene as compare d to previous methods as well as a significant increase in crossovers in regions with less than ten base pairs of sequence identity. The efficacy of these approaches has been furthered by the advent of methodology termed assembly of designed oligonucleoti des (ADO). 75,76 In these approaches synthetic oligonucleotides containing degeneracy at targeted positions are shuffled and reass embled using their inherent sequence overlap in a PCR based assembly process. This allows the researcher to insert a variety of point mutations in combination throughout the gene as well as predefine the number of crossover events in the mutagenic product The previously described approaches to recombinative mutations all require a certain degree of sequence homology between the members of the recombinative population theoretically limiting the scope of the directed evolution effort. Several methods have been developed to harness the utility of recombinative methods with a lack of sequence homology among the theoretical parents. Ostermeier et al developed several approaches based on an incremental truncation approach. 77 In these approaches two genes are designated for

PAGE 80

80 recombination at a single crossover point. One gene is digested by controlled end the other gene is digested in a similar manner from end and the reactions are quenched at set time intervals. The result of these digestions is gene fragments of all possible lengths for each gene. The resulting fragments are ligated and fusion produ cts corresponding to an approximate size are isolated by gel electrophoresis. While there is a chance that the fusion products are ligated out of frame and therefore useless the overall fusion construct was recombined at random from two potentially non ho mologous genes. This met hod and variations of it are given the title i ncremental truncation for the creation of hybrid enzymes (ITCHY) Traditional DNA shuffling of ITCHY libraries (termed SCRATCHY) can then be used to generate further diversity given tha t the ITCHY library now contains a degree of homology required. Sieber et al further developed an approach to fuse non homologous genes with a single crossover point while keeping the construct in frame 78 This approach is termed sequen ce homology independent protein recombination (SHIPREC) In this approach gene s are joined by a linker sequence containing useful restriction sites. This dimer is subject to fragmentation by DNaseI, treated to produce blunt ends and subsequently ligated t o form circular constructs Circular fragments that correspond to the size of a single gene are separated by gel electrophoresis. The linker is then removed by restriction digest revealing genes with crossover points distributed throughout the new gene c onstruct. Non recombinative methods One of the oldest and potential ly the most frequently used technique for diversity generation in directed evolution is error prone PCR ( epPCR ). This approach exploits

PAGE 81

81 the potential ly low fidelity of DNA polymerases und er certain conditions to insert random base substitutions in a given gene sequence during the PCR cycle. 79 The frequency of these mutations can be controlled by optimization of reaction conditions to yield anywhere from 1 20 nucleotide substitutions per 1 kb. 80 The most common methods for altering the fidelity of Taq DNA polymerase are the addition of Mn 2+ unbalanced nucleotide concentrations, and the use of nucleotide analogs. Alter natively a lower fidelity mutant polymerase may also be selected. 81 The method has the advantage of being able to insert mutations at nearly every po sition in the gene of interest. However, not all nucleotide substitutions lead to changes in amino acid incorporation. Base substitutions that occur in the third position often lead to no change in the amino acid at that position. Two sequential base su bstitutions would have a high probability of altering the amino acid but is statistically disfavored in the controlled epPCR reaction scenario. Statistical analysis of gene sequences indicates that only 45% of all nucleotide substitutions will yield an am ino acid substitution. Of those substitutions only 4 7 of the 20 possible amino acids will be represented at a given position with a disproportionate bias towards glycine or proline substitution 82 Another approach designed to overcome the inh erent biases of epPCR is the technique of Sequence Saturation Mutagenesis (SeSaM). Key to this approach is the construction of single strand sequences of the gene of interest of all lengths using a bead immobilization technique. The population of varied length nucleotides is then elongated with deoxyinosine using a terminal transferase and then fully elongated by an additional PCR reaction. The inosine containing genes are then subject to PCR where the inosine is replaced in a theoretically unbiased mann er with the four standard

PAGE 82

82 nucleotides. The result is a gene that contains mutations throughout its length that are independent of the mutational biases of DNA polymerase. Another random PCR based approach is the technique of random insertion/deletion (RI D). 83 With this meth od a cyclic single strand copy of a gene is subject to random cleavage by a Ce(IV) EDTA complex. Double strand anchor sequences containing the insertion sequence and unique restriction sites are then ligated to the randomly positioned ends of the cleaved gene. PCR fills in the new double stranded construct that is subsequently cleaved by restriction enzyme to remove the anchor sequence and the specified number of bases targeted for deletion. This blunt ended construct is ligated and digested again by uni que restriction enzyme to reveal a full length gene with the inserted sequence substituted randomly throughout the gene. To add diversity to the insertion library anchors constructed with degenerate bases can be used. Semi rational methods Semi rational approaches are so termed because they require some insight into the overall structure and function of the target enzyme. That is to say that we have some idea of where to look for improved function ; therefore that area of the protein space is targeted i n a random manner. The earliest example of this approach is combinatorial cassette mutagenesis (CCM). 84 In this method an oligonucleotide mimicking the target site on the gene is synthe sized with degenerate codons. The resulting degenerate cassette containing randomized nucleotides at desired positions is then ligate d i nto the gene to create a targeted mutant library for screening. A variation of this method has been described that reduces the redundancy of the genetic code by creating a cassette r and omized with a single codon for each of the 20 amino acids. 85 In

PAGE 83

83 this approach a set of 20 primers coding for each amino acid (MAX) is constructed for each position in a given cassette. These primers containing MAX randomization are hybridized with a synthetic template that is randomized with NNN at the same positions. The construct is then ligated, amplified by PCR, and restriction digested to isolate a cassette suitable for cloning. The resulting cassette therefore lacks the degeneracy of the genetic code theoretically simplifying subsequent sc reening efforts. One drawback to the use of cassette based strategies is the requirement for the parent gene to contain restriction sites suitable for liga tion of the synthetic cassette. A solution to this is to introduce degeneracy via mutagenic primers using a variation 86 of the s equence overlap extension (SOE) 87 method This approach requires the use of complementary mutagenic primers that are first used in separate reactions to amplify the forward and reverse halves of the gene when paired with appropriate non mutagenic primers. The two new overlapping halves are combined and extended via PCR to generat e a double stranded construct suitable for cloning. Another approach to saturation mutagenesis is the use of a whole plasmid strategy based on the popular QuikChange mutagenesis method 88 marketed by Stratagene. When applied with degenerate mutagenic primers the method was patented as Gene Site Saturation Mutagenesis ( GSSM ) 89 by Diversa Corporation and is also referenced in the literature as mutagenic plasmid amplification (MPA) 86 The replication of the whole plasmid using degenerate mutagenic primers eliminates the need for cloning as with the previous methods. The method also employs the use of template plasmid from a dam+ E. coli strain such that it is susceptible to digestion by

PAGE 84

84 DpnI endonuclease This therefore eliminates bias toward the wild type enzyme in the degenerate library. The introducti on of multiple mutagenic sites using variations of these whole plasmid approaches has also been described. The QuikChange Multi 90 kit as marketed by Stratagene requires the u se of a single degenerate mutagenic primer for each site of interest in the gene. All mutagenic primers are designed in the same direction. After annealing the plasmid is amplified and ligated in a single reaction using a polymerase and ligase blend. T he parent template is then digested to leave a single strand circular multi site mutated fragment that is then transfor med in to competent cells. Each cell then contains a single copy of the multi site mutant suitable for isolation and screening. An alte rnate approach to this has also been described 91,92 and recently optimized 93 based on the generation of mega primers. In these methods sense and anti sense primers located at distal sites and containing degenerate mutations are used to amplify a large fragment of a target plasmi d. In the later stages of the PCR process the large fragment functions as a mega primer for the amplification of the whole plasmid followed by subsequent DpnI digestion. Building on site saturation mutagenesis technique s Reetz and Carballeira describe d an approach that mimicked the directed evolution schemes involving multiple random mutagenesis cycles termed Iterative Saturation Mutagenesis (ISM). 94 Figure 3 3 outlines ISM me thodology. In the simplest version of their approach they use MPA to create degenerate mutagenic plasmids that are transformed into an expression strain and screened for desired activity. Positive results are then isolated, identified, and used as templat e for subsequent rounds of MPA at other beneficial sites. Key to their

PAGE 85

85 approach is the semi rational selection of sites subjected to saturation mutagenesis. When enhanced thermostability was the objective residues were selected based upon their crystall ographic B values. These values are a numeric reflect ion of smear ing of atomic electron densities with respect to their equilibrium positions as a result of thermal motion and positional disorde r. 94 Residues that display the highest degree of disorder are then subject to ISM. Conversely, when changes in substrate spe cificity or enantioselectivity we re the objective they propose d the use of a combinatorial active site saturation t est (CAST). 95 In this approach residues in the active site are grouped based on their side chain relative proximity to each other. Co randomization then in theory leads to synergistic effects not predictable by substitution of only a single amino acid side chain. When on e considers that when two sites are randomized simultaneously with NNN degeneracy the statistical oversampling required to observe all possible combinations exceeds 10 4 (for three sites this number approaches 10 6 see Table 3 3 ) it is clear that the CAST a pproach requires a degree of modification to be practical. The numbers problem To make the n umbers manageable Reetz proposed the use of restricted codon libraries. Using NNK degeneracy ( Table 3 1 for IUBMB abbreviations) this reduces the number of possib le codon s for 64 to 32 while still coding for all 20 amino acids. It also has the beneficial effect of reducing the nu mber of stop codons from three to one thereby reducing the theoretical number of inactive mutants in a given population. NNK codon usage is listed in Table 3 4 Further restricted codon usage can be used to target subsets of amino acids further reducing the total screening effort. Examples of a few of these are listed in Table 3 2 For example using an NDT degenerate codon will

PAGE 86

86 reduce th e number of codons to 12 each coding for a single amino acid. These 12 amino acids are a balanced mix of polar, nonpolar, aliphatic, aromatic, negative and positive charged residues that exclude structurally similar amino acid side chains. 96 Using restricted codon libraries also facilitates combinatorial degeneracy by reducing the total number of possible combination requiring screening. Specifically NNN degeneracy at two sites would create 4,096 possible variants (64 64 ) whereas with NNK the number is reduced to 1,024 possible combinations with a 67% reduction in the nu m ber of stop codons. Using Poisson statistics Reetz et al proposed the use of an oversampling factor, O f based on the perce nt probability of selecting all possible unbiased combinations of degenerate mutations. 96 In general terms, t he y determined that three fold oversa mpling was required to achieve 95% coverage of al l possible mutant combinations. These oversampling requirements are forecasted for a few degenerate profiles in Table 3 3 Using this as a guideline a researcher can tailor the library generation to fit the required screening effort. Screening The secon d phase of directed evolution is the s creening the generated diversity for beneficial mutations ( Figure 3 2 ). Screening large libraries is often the bottleneck of a directed evolution project. 97 Typically the reaction is chosen so that the enzyme yields a product that is both easily visible and non transient in nature. In some scenarios a pre screen can be adapted to eliminate non beneficial or deleterious mutations. The now reduced population can then be screened through more time consuming

PAGE 87

87 approaches. Examples of some of these screening methods are described in the following sections. The relative difficulty of a screening effort is more often than not a matter of perspective. That is to say if high throughput technology is available for a desired screen then the generation of an unusually large degenerate library may not seem like an insurmountable task. In our case we are interes ted in the stereo se lectivity of an enzyme where the product is most simply assayed by chiral gas chromatography. Depending on the product of interest the time to analyze 100 samples is measured in days (typically 1 3). Therefore, the construction of a 10,000 member library in a few days time could lead to a screening effort lasting a few months. Clearly the size of the library generated must be proportional to the capabilities of the screening system and the patience of the researcher. The diversity generation methods us ing the ISM approach described previously are the most straightforward way to generate manageable libraries for use in a directed evolution effort given our screening parameters. While it appears that this semi rational approach to directed evolution is t he most manageable its relative effectiveness compared to earlier traditional directed evolution methods needs to be evaluated. C omparing Directed Evolution A pproach es Given the myriad of diversity generation techniques used in directed evolution it wo uld be useful to compare successful approaches and the resulting outcomes. Many reviews have outlined directed evolution successes. 67,98 108 However, o ne significant problem arises in review of these work s is that typically only successful directed evolution efforts are published. Paramesvaran et al statistically compared the results of successful epPCR and site saturation approaches to determine the relative success of

PAGE 88

88 109 Their finding was that site site affecting substrate specificity were concerne d. But i t is very rare to find successful directed evolution efforts that use the same starting point enzyme, screening conditions and substrates required to truly evaluate each approach side by side Three particular examples serve to compare random dive rsity generation methods with semi rational methods in their efficacy. These examples include those of Escherichia coli galactosidase Aspergillus niger epoxide hydrolase and Pseudomonas aeruginosa lipase galactosidase galactosidase from Escheri chia coli (BGAL) is an enzyme that catalyzes the galactosides into their respective monosacchar ides. The enzyme is also known to cleave the glycosidic bond in 5 bromo 4 chloro indolyl D galactopyranoside (X G al) into galactose and the re spective indole that further dimerizes into an insoluble and intensely blue compound useful for visual screening. The enzyme is also known to catalyze the hydrolysis of glycosidic linkages in p nitrophenyl D galactopyranoside (PNPG ), and o nitrophenyl D galactopyranoside (ONPG ) The resulting nitrophenols are both soluble and yellow in color (420 nm) making them useful for measuring kinetic activity of galactosidase. Using these known substrate specificities of galactosidase Zhang et al set abo ut in vitro directed evolution of the enzyme using successive rounds of DNA shuffling 110 The specific aim of this work was to evolve fuco sidase activity in the native enzyme. They introduced low frequency point m utations into the galactosidase gene by DNA shuffling. 70 The resulting mutant plasmids were transformed into an

PAGE 89

89 expression strain and plated on selection media co ntaining 5 b romo 4 c hloro indolyl D fuc opyranoside (X Fuc ) The resulting library of ~10,000 transformants contained 2 5% that stained blue by visual inspection. 20 40 of the bluest colonies were selected and their DNA pooled for seven successive rounds of DNA s huffling b y the same technique. The most blue colony at the end of the seventh round was selected sequenced and kinetically characterized. When characterized against p nitrophenyl D fucopyranoside (PNPF) and o nitrophenyl D fucopyranoside (ONPF) the resulting mu tant showed a 300 fold increase in substrate specificity for PNPF and a 1000 fold increase in the specificity for ONPF over that of the wild type enzyme. The evolved gene contained 11 nucleotide substitutions that are further translated into 6 amino ac id changes in the peptide sequence. Three substitutions (P511S, Q573R, and N604S) were thought to directly influence the active site and a fourth (D908N) may potentially influence the active site. The remaining amino acid changes lie on the surface of th e protein and are thought to not b e catalytically significant. Nearly a decade later Parikh and Matsumura used a semi rational approach to evolving fucosidase activity. 111 In a direct comparison to the work described by Zhang et al they focused their attention of three residues (Asp201, His540, and Asn 604) known to coordinate a sodium ion in the enzyme active site This analysis i s similar to the CAST approach described earlier Each of these three residues was randomized with N NK degeneracy simultaneously. After transforming the multi site saturated plasmid into an expression strain the library was screened as previously described. Approximately 10,000 mutants (25% coverage) were screened, a number significantly smaller than t he 10 5 that required for 95% coverage. The eight transformants that

PAGE 90

90 displayed the highest fucosidase activity remain ed unchanged a t position 201. Remarkably they also contained the same mutations at t he other two positions (H540V and N604T). By compar ing the mutants containing each of the single substitutions with the evolved double mutant they were able to observe that the greatest contributor to fucosidase activity was the H540V mutation Most impressive in this effort was a 700,000 fold selectivity increase for PNPF compared to that of the wild type enzyme. This direct comparison of random and semi rational directed evolution approaches shows that the multi site saturation protocol is the clear winner with over 2000 times the enhancement in selectiv ity for D fucopyranoside s However, the semi rational approach required detailed information of the structure and mechanism of the galactosidase enzyme. It is important to note th e magnitudes of the screening effort in both approaches. The random app roach required visual screening of ~10,000 transformants for each of the seven rounds followed by detailed characterization of the rational approach required a single round of visual characterization followed by activity assays of the 30 best mutants. The top eight performers in this round were sequenced to reveal identical mutation profiles. Detailed characterization was done on the two single and the evolved double mutant to finalize the effort. Given the rel ative speed of the visual screen the major advantage with the semi rational approach is in the time required to conduct mutagenesis Combining the time advan tage with greater enhancement in substrate selectivity the semi rational site saturation approac h is clearly the superior technique in this direct comparison. Epoxide hydrolase Epoxide hydrolase from Aspergillus niger (ANEH) is an enzyme that catalyzes the hydrolysis of the model compound glycidyl phenyl ether to its respective diol with a

PAGE 91

91 slight pre ference for the ( S ) product (56% ee, 33% conversion in 45 minutes, E = 4.6). 112 Enantioselectivity, E has been described previously 113 as the ratio of specificity constants, V / K for each of the two competing enantiomers in a reaction and can be calculated as a function of conversion and enantiomeric excess. The presence of this epoxide can be quantified by simple assays One such assay is the formation of a blue dye when in the presence 4 p nitrobenzyl pyridine that can be quantified by absorbance at 560 nm. 114 Another is by simply using the inherent toxicity of the epoxide to select for bacterial transformants that display the abil ity to hydrolyze the epoxide to its less toxic diol product. 112 Using these properties Reetz et al set about in vitro directed evolution of the enzyme using a single round of epPCR. 112 Their aim was to increase the selectivity factor for the glycidyl phenyl ether substrate mutagenesis followed by s ubsequent screening. The mutagenic ANEH bearing plasmids were transformed into an E. coli host. Following expression 20,000 clones were incubated with racemic epoxide followed by staining with p nitrobenzyl pyridine The quick screen re vealed approxima tely 400 mutants with increased hydrolase activity. Those mutants were further subject to ESI MS to determine absolute stereoselectivity of the library population. 115 This assay requires the use of pseudo enantiomers in a mixture of 1:1 of (S) glycidyl phenyl ether and ( R ) D 5 glycidyl phenyl ether The ratio of the mixed diol product can then be determined by the 5 mass unit difference between the two possible products. Of the 10 most improved mutants the best was a triple muta nt ( A217V,

PAGE 92

92 K332E, A390E ) that displayed a more than 2 fold increase in enantioselectivity (74% ee, 39% conversion in 2 minutes, E = 10.8). Seeking to improve on these results Reetz et al pursued a new approach to directed evolution of ANEH. 116 This involved inspecting the ANEH active site and designating six regions using the CAST approach. Three 3 site libraries and three 2 site libraries were generated and designated A through F ( 193/1 95/196 (A), 215/217/219 (B), 329/330 (C), 349/350 (D), 317/318 (E), and 244/245/249 (F) ). B acterial growth was used as the initial quick screen followed by ESI MS as before. After screening libraries A, B and C they discovered that only library B contai ned mutants with improved enantioselectivity. Two of these mutants yielded an E = 14, three times greater than the wild type enzyme and slightly better than the epPCR directed evolution result ( E = 10.8). The library B variant was further randomized iter atively against the best variant of each subsequent round in sequence B D, F, E. The best resulting mutant from this series of iterative mutations contained nine total amino acid substitutions ( L215F, A217N, R219S, L249Y, T317W, T318V, M329P, L330Y, a nd C350V ) and gave an E = 115 10 ( 95% ee and 48% conversion in 1 hour). This result is 25 times greater than the wild type enzyme and a 10 fold improvement over the previous mutagenesis effort. This result is even more remarkable considering the total sc reening effort was identical in size (20,000 mutants screened) to the epPCR effort. This translates into less than 10% coverage for the three position libraries and less than 70% coverage for most two position libraries. However, the resulting quality of the targeted CAST libraries is significantly greater than the random approach library,

PAGE 93

93 which then translates into greater enantioselectivity improvements for the same screening effort. In a separate and later effort Reetz and Zheng revisited directed evo lution of ANEH. 117 The focus of this approach was to evolve an ANEH variant with improved expression, a problem that had plagued previous efforts to make ANEH a commercially viable catalyst. In this effort the ANEH ge ne was randomized by low frequency mutation epPCR and inserted in a plasmid connected by a linker region upstream from a gene coding for the galactosidase fragment. When transformed into E. coli DH5 and plated onto agar suppl emented with X Gal colonies stain ed blue in proportion to the relative expression of the fused fragment. Fourteen blue colonies were selected visually from a 15,00 0 member library. Expression was further quantified by whole cell ONPG assay s and the mutants were sequenced. The best performing mutant was determined to be a single substitution variant, P221S, and was found to have retained wild type hydrolase activit y ( E = 5). This mutant was further cloned into a pET22b vector under the control of the T7 promoter and co expressed with a chaperone bearing plasmid to yield a 50 fold increase in soluble protein expression. This plasmid construct was used as the starti ng point for site saturated mutagenesis study to evolve enantioselectivity. This second attempt at directed evolution of the ANEH enzyme was significant ly different than the first approach. This later analysis of the ANEH active site yielded four regions containing two positions each using the CAST approach (215/ 219 ( A ), 349/350 ( B), 317/318 ( C ), and 244/249 (D)). This selection contains six of the nine positions substituted in the previously evolved variant. To simplify screening effort, codons at

PAGE 94

94 these positions were randomized with NDT degeneracy which reduced the sampling effort for each library to 480 (five 96 well plates) in order to achieve 95% coverage. The P221S high expression variant was randomized iteratively against the best variant of each subsequent round in sequence D, A, C, B The best resulting mutant from this series of iterative mutations contained nine total amino acid substitutions ( F244C L249F L215F* T317F T318V* and L349V) (*Mutation found in previous ISM best variant.) and g ave an E = 160 (97 % ee and 45 % conversion in 1 hour). Interestingly they also observed a 13 mer peptide insert after the library C randomization that had no negative effect on enantioselectivity and was therefore left in place for the subsequent saturati on mutagenesis cycle. This insert is due to a double insert of the mutagenic primer used to construct t he C library. T he extent of this insertion within the library was not determined; however, it can be assumed that an insert of this size would typicall y have a significant negative impact on protein structure and folding, particularly if the primer insertion throws translation out of frame. Despite the relative success of this attempt the overall library quality is questionable in this case It is hig hly probable that successive primer sequence inserts in fact deactivated a large portion of the library. Given the limited screening employed (3,000 mutants) and the flawed library this m utagenesis effort was remarkably successful with respect to enantios electivity Comparing these three directed evolution experiments for ANEH gives great insight into directed evolution theory. Table 3 5 summarizes these efforts and results. I t can be seen that these methods can dramatically reduce screening effort while increasing coverage of all pos sible mutational combinations. Presupposing that the most appropriate sites are selected in a semi rational approach i t can be postulated that

PAGE 95

95 this increased coverage is directly related to the magnitude of the desired and im proved result. Lipase The lipase from Pseudomonas aeroginosa (PAL) is the most highly studied directed evolution target with respect to substrate acceptance and stereoselectivity. 118 PAL catalyzes the enantioselective hydrolysis of the model compound racemic p nitrophenyl 2 methyldecanoate to the corresponding acid with a slight preference for the ( S ) substrate ( 2 % ee, E = 1.1 ). Hydrolysis of the model compound can b e quickly screened by the formation of p nitrophenol (410 nm) Detailed study of the hydrolysis reaction can be monitored by traditional chiral GC methods after an efficient pre screening. In 1997 Reetz et al conducted the first reported directed evoluti on of an enzyme to enha nce its stereoselectivity 119 In this work the gene encoding PAL was subject to four successive rounds of low frequency mutational epPCR. In each round the best improvement in stereoselectivity was used as the template for the following round. The result was a substantial increase in the ( S ) selectivity of the enzyme (81% ee, 25% conversion, E = 11.3). The total screening effort for all four rounds was 7,600 mutants with only 29 displaying sufficient a ctivity to justify detailed characterization. In a continuation of this effort Liebeton et al constructed six sequential low error rate epPCR generations on the PAL enzyme 120 Each round involved the screening of 1000 7000 clones. After the sixt h generation they had successfully evolved from wild type ( E = 1.1) to a variant with an E = 13.5 containing nine amino acid substitutions ( S149G, S155L, V47G, F259L, L110R, Y8H, N21D, S158T, S248C) This result was very similar to the previous directed e volution effort ( E = 11.3). The selected best

PAGE 96

96 mutants from each of the first five rounds each harbored a single unique amino acid substitution. To explore further possibilities with their mutants they turned to saturation mutagenesis. NNN mutagenic prime rs were used to saturate each of the best variants from the first four generation s at position s corresponding to their respective amino acid substitutions ( in order 149, 155, 47, 259). 800 mutants were screened in each of the four degenerate libraries. I nterestingly, only one mutant from the s econd generation variant (S149G, S155L) showed any improvement in enantioselectivity. In this case the leucine to phenylalanine substitution at position 155 gave a small improvement fro m E = 4.4 to 5.7. Following this observation the third generation variant (S149G, S155L, V47G) was randomized at position 155 by the same approach. The best mutant isolated from this library contained, in fact, the S155F mutation and exhibited an E = 20.5. With the S155F mutation n ow considered critical to evolve the PAL enzyme the mutation was introduced into the wild type enzyme and the fourth and fifth generation variants by site directed mutagenesis. Each showed an increase in E but not greater than the value of 20.5. This b est variant (S149G, S155F V47G) was subject to yet another round of epPCR where yet another improvement was observed with two new unique substitutions at positions 55 and 164 which gave a further improvement to E = 25.3. The work described above was done without the a id of a PAL crystal structure ; thus epPCR served as a probe of enzyme protein sequence in search of influential positions. It also demonstrated that while epPCR may be efficient at finding those residues it may not be best suited for determ ining the optimal amino acid substitution as evident by the position 155 saturation mutagenesis.

PAGE 97

97 Reetz et al further pursued an improvement on the best variant from the Liebeton et al effort. 121 Thi s involved DNA shuffling of this variant with two others (E = 3 and E = 6.5) obtained from high error rate epPCR each containing three random amino acid substitutions. The resulting best mutant gave an E = 32 a new high, and contained all three mutation s from the Liebeton et al variant plus two new substitutions (S199G, T234S). Continuing further the Reetz group also pursued a strategy of combinatorial multiple cassette mutagenesis (CMCM) In this approach a 69 bp cassette with NNN degeneracy at po sitions co rresponding to residues 155 and 162 was shuffled with the wild type PAL enzyme The result was a new variant with an E = 34. When this cassette was shuffled with the two previous high error rate epPCR variants the result was a new enzyme with high enantioselectivity (95% ee, 24% conversion E = 51) containing six amino acid substitutions ( D20N, S53P, S155M, L162G, T180I, and T234S ). Notably this variant contained a methionine residue at position 155 and not the previously assumed critical phen ylalanine. The overall effort for this work involved the screening of approximately 40,000 mutants. The new mutagenesis strategy resulting from this work can be described in three steps as 1) using high error epPCR to probe the protein sequence, 2) usi ng CMCM to validate against the wild type enzyme, and 3) extending the CMCM modification to the early this approach yielded a 2 fold increase 120 over the single site saturated technique for fold improvement 119 over the probe alone.

PAGE 98

98 Similar to this work Zha et al set about evolving ( R ) enantioselectivity from the PAL enzyme. 122 In this approach the wild type enzyme was subject ed to three rounds of epPCR. This was followed by two successive rounds of DNA shuffling of t he epPCR variants then another round of epPCR. The result was a mutant with ( R ) enantioselectivity ( E = 30) similar to that of the ( S ) evolved variants. The mutant contained eleven amino acid substitutions and was the product of a net 45,000 mutant scre ening effort. Clearly the advent of an efficient UV vis pre screen is essential in these approaches to directed evolution. In the most recent and effective approach to evolving the PAL enzyme Reetz et al applied the previously discussed CAST technique in an ISM approach. 123 T hey opted to designate three libraries of two positions each ( A (Met16/Leu17), B (Leu159 /Leu162), and C (Leu231/Val232)) and introduced NNK randomization via degenerate mutagenic primers. This level of de generacy required screening 3000 mutants per library in order to obtain 95% coverage. After screening libraries A and C yielded no results while library B yielded several, the best being a single mutant (L162N) which gave an E = 8 for ( S ) p nitrophenyl 2 methyldecanoate. This mutant was used as the starting point for a second round of mutagenesis. In this round, the poor result s for the A and C libraries combined with a desire to reduce the screening effort prompted the use of DNT degeneracy for the tw o position randomization of libraries A and C. This amino acid subset was chosen because leucine is notably absent from the 11 possible amino acids ( Ala, Asn, Asp, Cys, Gly, Ile, Phe, Ser, Thr, Tyr, and Val ) and the screening effort could be reduced to 43 0 mutants per library to obtain 95% coverage. When randomized in this manner against the A

PAGE 99

99 grouping the best result obtained was an unprecedented value of E = 594 for the ( S ) substrate. This variant contained two new mutations (M16A, L17F). Given that randomization for the C grouping gave no improvements no further iterative cycles were performed. The total screening effort for this remarkable result was 10,000 mutants. The substrate scope of PAL has al so been explored in detail the CAST technique fo r site selection 95 and ISM 118 to achieve dramatic improvements for a wide range of p nitrophenyl esters. Based on all of these precedents, the ISM technique is the most straightfor ward and concise approach to altering the stereoselectivity of our target enzyme, OYE 2.6. The focus of our study involves the uncovering of an alternative, fli p ped binding mode of a pro chiral alkene in order to produce the opposite enantiomer. Preceden ce for using ISM for this purpose has been set with the alkene reductase Y qjM. ISM of an Alkene R eductase, Y qjM The alken e reductase and Old Yellow Enzyme homolog YqjM from Bacillus subtilis catalyzes the asymmetric reduction of prochiral unsaturated k etones in a reaction mechanism analogous to that of OYE 1 discussed in Chapter 1. The enzyme poorly reduces 3 methyl 2 cyclohexen 1 one (3% conversion) with a slight preference for the ( R ) product (79% ee). Bougioukou et al used the ISM approach to impr ove both activity (conversion) and steroselectivity of this 3 substituted model compound. 124 Given that the only feasible method to observe this reaction was through chiral GC analysis they established a modified screening protocol. M utants arrayed on a 96 well plate were pooled by column and screened for conversion by a short achiral GC method Pools that formed product quantities above the wild type threshold were subsequently deconvoluted to identify individual mutants of interest. Positive results at

PAGE 100

100 this stage were then further subject to chiral GC analysis to assess enantioselectivity. To further reduce screening effort individual amino acids were selected for randomization by the CAST approach. By not grouping CAST selections individual sites could be randomized and sampled to 95% confidence with only 94 mutants per site Using a crystal structure with a bound inhibitor (PDB 1Z 42) t wenty residues were selected with varying potentials to affect catalysis based on their distance from the carbon of the substrate analog. By randomizing each site individually th e first round screening effort ( not including the pooling stra tegy o r additional oversampling) was estimated at 2000 mutants or approximately 7 days of continuous non chiral GC operation for identification of active variants. Conversely construction of ten 2 site randomized libraries would theoretically increase the firs t r ound screening effort to 30,000 mutants in order to achieve similar coverage The assumption was then made that positive effects would be apparent at least to a small degree with single site randomization and that combinatorial effects could then be ex ploited in subsequent rounds of mutagenesis. After first round screening 35 variants were identified as having increased ( R ) or ( S ) enantiopreference and increased conversion. Each was given an impact score derived from the product of the % conversion an d % ee. In this manner each property was given equal weight in ranking the variants. Several variants were chosen as star t ing points for the second round of single site randomization (( R ) selective I69T and C26W; ( S ) selective C26D, C26G, T70H and A60C) Non systematic second round randomizations at these positions led to higher overall impact scores. The net result was several mutants displaying greater than 95% ee and greater than 50% conversion

PAGE 101

101 for both ( R ) and ( S ) reductions of the model compound ( ( R ) selective C26W( A104 Y and A104F), C26D( I69T ), C26D(A104Y, A104F, and A104W) ; ( S ) selective C26G (A60C, A60I and A60V ) Interestingly, the C26D mutant, that gave ( S ) selectivity in round one, yielded the best variants for ( R ) selectivity in round two. T his observation further emphasizes the importance of combinatorial effects and the non linearity of a directed evolution effort. The evolved variants were further screened against 3 substituted ethyl, isopropyl, n butyl, and methyl carboxylate derivatives of cyclohexenone as well as 3 substituted methyl and methyl carboxylate cyclopentenone. Several evolved variants displayed specificity for the additional substrates that was not present in the wild type enzyme. In this case the pool of evolved variants provided an effective source library for screening a family of related substrates. Conclusion s Our goal is to evolve OYE 2.6 to in duce the opposite binding mode (pro R ) for our Baylis Hillman derived substrates discussed in Chapter 1 Precedence for succ essful evolution of an Old Yellow Enzyme homolog is demonstrated by the efforts of Bougioukou et al Key to that work was the selection of a directed evolution strategy that could be conducted with an abbreviated screening effort. The use of ISM with res tricted codon usa ge enabled the construction of a focused high quality library that was manageable in size for a GC based assay system. Several examples show that combinatorial site saturated libraries as used in ISM methodology provide superior directed evolution results when compared side by side with more random mutagenesis approaches (e.g. DNA shuffling, epPCR). Key to using a site saturated approach such as ISM is the generation of high quality mutant libraries. In this case quality depends

PAGE 102

102 selectio n of the right sites and completeness of randomization while eliminating redundant and erroneous gene constructs. Next chapter describes our efforts for constructing mutant libraries for the Iterative Saturation Mutagenesis of OYE 2.6.

PAGE 103

103 Figure 3 1. N umbers of publications dealing with in vitro directed evolution The number of publications per year in the field of directed evolution (orange) and fraction of knowled ). The search term was used to retrieve the number of publications on directed evolution, and the additional search criteria ntioselectivity and stereoselectivity.

PAGE 104

104 Figure 3 2. The in vitro protein engineering process

PAGE 105

105 Table 3 1 International Union of Biochemistry and Molecular Biology (IUBMB) nucleotide nomenclature Letter code Name Bases Complement A Adenine A T C Cy tosine C G G Guanine G C T Thymine T A R puRine A G Y Y pYrimidine C T R S Strong (3 H bonds ) G C S W Weak (2 H bonds ) A T W K Keto T G M M aMino A C K B not A C G T V D not C A G T H H not G A C T D V not T A C G B N Unknown A C G T N Table 3 2 Coding description of selected degenerate codons Degeneracy # of c odons # of amino acids Distribution Properties NNN 64 20 all + 3 stop NNK 32 20 all + 1 stop NDT 12 12 CDFGHILNRSVY mixed DYK 12 8 AFILMSTV no charg ed RRK 8 7 DEGKNRS all polar Codon usage determined using CASTER v. 2.0 software available at for download at http://www.kofo.mpg.de/en/research/organic synthesis.

PAGE 106

106 Table 3 3 Oversampling for 95% coverage as a function of degeneracy and number of sim ultaneously randomized sites. # of codons NDT NNK NNN 1 34 94 190 2 430 3 066 12 269 3 5 175 98 163 78 5312 4 62 118 3 141 251 50 260 046 Oversampling calculated using CASTER v. 2.0 software available at for download at http://www.kofo.mpg.de/en/rese arch/organic synthesis. Table 3 4 NNK degenerate codon usage. Amino Acid Letter code Codons # Codons % Occurrence Alanine A GCG, GCT 2 6% Cysteine C TGT 1 3% Aspartate D GAT 1 3% Glutamate E GAG 1 3% Phenylalanine F TTT 1 3% Glycine G GG G, GGT 2 6% Histidine H CAT 1 3% Isoleucine I ATT 1 3% Lysine K AAG 1 3% Leucine L CTG, CTT, TTG 3 9% Methionine M ATG 1 3% Asparagine N AAT 1 3% Proline P CCG, CCT 2 6% Glutamine Q CAG 1 3% Arginine R AGG, CGG, CGT 3 9% Serine S AGT, T CG, TCT 3 9% Threonine T ACG, ACT 2 6% Valine V GTG, GTT 2 6% Tryptophan W TGG 1 3% Tyrosine Y TAT 1 3% Stop none TAG 1 3%

PAGE 107

107 Figure 3 3. Iterative Saturation Mutagenesis methodology for the evolution of increased stereoselectivity. In this ex ample l ibraries designated by a b c and d correspond to selected positions in the peptide sequence. In practice each library may consist of multiple sites randomized simultaneously by the CAST approach. Asterisk (*) indicates saturation with desired d egeneracy (i.e. NNK) Subscript numbers correspond specific amino acid substitutions (1 19 possible). The best mutant in each round serves as the starting template for the following round of mutations.

PAGE 108

108 Table 3 5 Summary of ANEH directed evolution ef forts. Methodology Screening effort Coverage Result ( E ) wild type 4.6 epPCR Single round 20,000 10.8 ISM NNK 20,000 5 70% 115 2 3 a.a./library ISM NDT 3,000 95% 160 2 a.a./library

PAGE 109

109 CHAPTER 4 CREATING SITE SPECIF IC SATURATED LIBRARIES Introduction This chapter describes the studies of how best to create degenerate libraries of OYE 2.6 variants used in directed evolution experiments The objectives and efforts described in this chapter are summarized in Fi gure 4 1 The screening of these libraries is a separate and distinct effort described in detail in the following chapter. Experimental Strategy Site S election At the onset of this work there was no solved crystal structure for the OYE 2.6 enzyme. We therefor e ch ose to establish candidate residues for OYE 2.6 CASTing by selecting active site residues in O YE 1 then identifying the OYE 2.6 counterparts by comparing amino acid sequences. Protein Data Bank accession number 1OYB is a 2.0 resolution crystal structur e for OYE1 with a bound inhibitor, p hydroxybenzaldehyde. 30 The phenol moiety in the bound ligand is oriented toward and most likely engaged in hydrogen bonding with the side chains of His191 and Asn194. This binding mod e is analogous to that of our carbony l containing substrates 14 For the purposes of CASTing, we a ssume that bound p hydroxybenzaldehyde is a good a pproximation of the size and orientation of our proposed substrates 1 2 and 3 Active site residues in OYE 1 were selecte d if their side chains met the following three criteria: L ocated above the plane of the catalytic FMN O riented toward the c enter of the active site pocket. C apable of interacting directly with the substrate by both position and distance

PAGE 110

110 Once r esidues were selected by visual inspection the distance s between the carbon of the residue side chain and the nearest carbon of the bound inhibitor were measured ( Figure 4 2 ) and recorded ( Table 4 1). Those r esidues with distances of 5 or less were determined to have a high probability of interacting with our substrate s while between 5 and 10 were designated as possibly interacting with our target substrates. Figure 4 3 shows a schematic array of these residues in the OYE 1 active site. These residues were matched in an alignment with OYE 2.6 ( Figure 4 4 ). From thi s alignment we determined that the best candidate residues for CASTing in OYE 2.6 were Thr35, Phe37, Ile113, His188, His191, Tyr193, Phe247, Asn293, Val294, Phe373, and Tyr 374. Residues Leu115 an d Gln248 were not selected based on their partial masking by adjacent active site residues. Figure 4 6 shows a schematic array of the OYE 2.6 active site residues. During the construction of these randomized libraries parallel efforts in our group eventually yielded crystal structures of OYE 2.6. 125 One structure contained a bound inhibitor, p chlorophenol, that binds above the FMN in an analogous manner to the phenolic liga nd in the 1OYB structure. Figure 4 5 depicts a ribbon diagram of OYE 1 (green) and OYE 2.6 (blue) overlaid with the FMN as the focal point of the alignment. From this depiction we can see that there is a large degree of structural similarity between the secondary and tertiary structures of the two enzymes. Additionally, the phenolic ligands bi nd with very close alignment. One region that did show significant deviation however, is the loop region centered on Pro295 of OYE 1. Sequence alignment revealed that OYE 2.6 lacks a corresponding proline residue and the loop

PAGE 111

111 itself is several amino acids shorter in length. Based on the se observations, Gly292 of OYE 2.6 was added to our list of CASTing libraries. Finally, w e also observed that OYE 2.6 Ala68 met t he criteria for selection. This residue had been overlooked in our original analysis based on OYE 1 because the corresponding residue in OYE 1 is a glycine that lacks a carbon and was interpreted as backbone in our earlier measurements. We had also overlooked Tyr78 in the OYE 2.6 active site due to its orientation away from the bound ligand in the OYE1 crystal structure This residue displays a significant shift in po sition toward the bound ligan d in OYE 2.6 Both of these residues were added to our list of libraries needed for complete coverage of the OYE 2.6 active site. Library C reation Our goal was to construct libraries using the protocol described by Bougioukou e t al. This approach was based on the QuikChange (Stratagene) protocol for site directed mutagenesis using degenerate primers. Modifications to this methodology are discussed in the results section of this chapter. The template DNA containing the OYE2. 6 GST fusion protein was constructed in our lab previously 18 and further mutated to remove a second NdeI restriction site that would interfere with future cloning steps. The resulting plasmid, pBS2 ( Figure 4 7 ), was further purified from a dam + strain so as to facilitate its digestion by DpnI endonuclease. Mutagenic primers were de signed with NNK degeneracy to provide the potential for 32 codons encoding all 20 amino acids. As previously noted, this library size required screening 94 samples to obtain a 95% probability of identifying at least one example of each codon. This modest effort fits within our constraints for minim ized downstream screening

PAGE 112

112 Assessing D egeneracy To the best of our knowledge there is no published method for determining the quality of a degenerate library from pooled sequencing data. This must be established prior to library screening to ensure that all possible clones actually are present within the collection. Kille developed an approach to estimate the relative amount of parent template in a pooled sample an d used it to determine the degree of oversampling required to achieve an adequate sample size. 126 To analyze the quality of our att empts in a quantitative manner we propose a method for calculating a value for the quality of de generacy, Q. This value i s determined by calculating the relative percentages of peak amplitudes for each base at each degenerate position from capillary sequencing data Then at each posit ion in the degenerate codon the absolute value of the deviation 50% for K) i for each base and i s then totaled at each position for all bases. The result of this treatment yields average of these values for all three positions in the codon can be used to calculate the quality of degeneracy for the codon, Q W hen calculated from fluorescence sequencing chrom atograph y (chromat) data from pooled plasmid DNA, the value is referred to as Q pooled which is an estimate of degeneracy. Figure 4 8 shows a sample calculation of Q pooled When determined from a sample population of individually sequenced mutants the va lue is referred to as Q codon and is an actual measure of degeneracy. The equations for determining Q are as follows: Q pooled = ( 0.4 Q N1 ) + ( 0.4 Q N 2 ) + (0.2 Q K ) where Q N = A C G T

PAGE 113

113 and Q K = G T At any N degenerate position for any base, B, B = 0.25 |0.25 X B | and at any K degenerate position for any base, B, B = 0.50 |0.50 X B | amplitudes at that position. This approach places a quantitative value on what was originally a qualitative observation when viewing sequencing chrom atograph data, i.e. the visual assessment (or not) in height. It assumes that all peaks and only those peaks expected (A, C, G, and T for N; G and T for K) are present in the sequencing chromat. A t tempts that display what a re obviously wild type or other disproportionate bias should be qualitatively assessed and not subject to this mathematical assessment Using the Q value facilitates discrimination between experiments that display varying degrees of degeneracy. Experimental Procedures Construction of Pooled Degenerate P lasmid Saturation mutagenesis libraries were prepared by a modification of the methods reported by Zheng et al. 127 Each Phusion 1 ng of pBS2 ( Figure 4 7 and Figure B 1 ) forward and ( Table C 1 ) and Phusion Hot Start II High Fidelity DNA Polymerase (1 U) was subjected to an initial denaturation step of 98C (30 s) followed by 25 cycles of 98C (10 s) followed by a range of extension temperatures from 62 to 72C (4 min) and completed by a final

PAGE 114

114 incubation at 72C (7 min). Amplicons were purified by DNA spin columns, digested with DpnI at 37C (10 U for 4 h followed by an add itional 10 U for 4 h) to remove adenine methylated template DNA and then purified by an additional DNA spin column. ) were used to transform Electo Ten Blue (Stratagene) electocompetent incubated for 1 h at 37C prior to selec tion on LB media plates supplemented with 200 mg/L ampicillin. After overnight incubation reactions yielding 300 colonies (3 plates combined) were pooled with th e aid of LB media and pelleted. Plasmid DNA was then purified (spin columns) and analyzed b y DNA sequencing to identify samples with the highest degree of pooled degeneracy at the desired position. Degenerate Library C rea tion After confirming adequate degeneracy of each pooled library 1 ng of each was used to transform electrocompetent E. coli BL21 Gold (DE3) (Stratagene, 40 ). SOC selection on LB medium supplemented with ampicillin After incubation at 37C for 16 hours individual colonies (95) were selected and used to inoculate wells conta ining 600 LB medium supplemented with ampicillin. The 96 th position in the plate was inoculated with a colony containing a plasmid encoding the wild type protein The resulting 2 mL deep well plate was then shaken (250 rpm) at 37C for 6 hours. Steri le glycerol was then added to bring each culture to a final concentration of 15% wt/vol. These plates were then replicated into multiple 150 volume copies and stored at 80 C until needed for screening ( Chapter 5) Eleven of these libraries were selec ted for

PAGE 115

115 whole plate sequencing. Copies of these plates were frozen at 20 C and sent for RCA protocol sequencing (UF ICBR) using appropriate primers. Results and Discussion Developing a Mutagenesis M ethod A significant amount of time an d ef fort was put in to developing an adequate saturation mutagenesis method. Our primary focus at the onset of this project was to develop a libr ary that was truly degenerate so that we had a 95% chance of obtaining all 32 possible codons in 95 randomly selected transformant s. Practically speaking an adequate library is one that yields all 20 amino acids. Eventually we discovered that with our self imposed sample size restrictions and the limits of PCR based methodologies our definition of adequate would have to be furthe r modified. Issues requiring attention during method development included : Transformation efficiency of PCR products Extraneous PCR derived errors Frequency of wild type appearance Distribution of degeneracy Table 4 2 and Figure 4 9 outline five distin ct approaches that we took in the search to resolve these issues. It is worth noting that methodology changes were attempted in a semi sequential manner with the goal of finding a workable method P arallel assessment and optimization of the methods was t herefore not performed and the discussion here does not serve to compare the methods to each other. Transformation efficiency One bottleneck shared by all library construction approaches is the transfor mation efficiency of mutagenized PCR products. We acc epted the requirement that we obtain 300+ colonies from our PCR product transformants as a sample large enough to assess

PAGE 116

116 degeneracy when pooled and sequenced. 124 D uring our investigations we frequently failed to achieve this number of transformants, which caused us to discard the attempt due to presumed inadequate degeneracy and actual insufficient quantity of DNA for sequencing. Several fact ors affect transformation efficiency. First is the concentration of mutagenic plasmid. Approaches described by Bougioukou 124 and Edelheit 128 are linear amp lification methods where the product of the preceding PCR cycle does not function as a template for subsequent cycles. This results in a non exponential amplification and provides lower DNA yields from the PCR process. We increased the volumes of the PCR reactions to 100 L in an effort to achieve a greater number of transformants with these methods. However, we ultimately moved on to the other methods listed in Table 4 2 which result in an exponential amplification, higher DNA yields and, in theory, mor e mutagenic transformants. The quality of the electrocompetent cells used to transform PCR products also played a significant role in yield. In general, we found that electrocompetent cells prepared fresh, not frozen, and used the same day for electropo ration resulted in the greatest transformation efficiencies. This observation held true for both E. coli DH 5 and JM109 cell lines. The most significant improvement in transformation efficiencies came with the use of E. coli ElectroTen Blue (Stratagene) cells. ElectroTen Blue cells were developed and are marketed as a high electroporation efficiency ( Hee ) phe notype. According to the manufacturer these cells better survive the electroporation event that in turn results in a larger mutant library. Using them afforded us up to two orders of magnitude greater

PAGE 117

117 transformants. The resulting 200 colonies per plate allowed us to achieve a larger sample size with fewer plates. Practically speaking we were able to reduce the number electroporations per library from two or three to one and the resulting number of plates pooled from six or nine to three. Primer conc atemers T he next issue was the elimination of PCR derived errors. Specifically these were manifest in the form of multiple repeating inserts of our mutagenic primers into the plasmid sequence. We originally set out to reproduce the mutagenesis method d escribed by Bougioukou. 124 After dealing with transformation efficiencies described in the previous section we obtained a pooled plasmid li brary for OYE 2.6 I113X that appeared significantly degenerate in the sequ encing chromatogram Using the previously described assessment the degeneracy estimate for that library scored a Q pooled = 0.84. The library was then transformed into an E. coli B L21 (DE3) expression strain and 95 random colonies along with one wild type control, were arrayed in a 96 well plate format and sequenced by RCA. While the quality of sequencing results was less than desired (66 of 96 wells sequenced) the results were su rprising. The degeneracy of this library was poor with only 15 of 20 amino acids present in the sample. The presence of wild type enzyme was slightly higher than expected in a truly degenerate population with 9% isoleucine present as opposed to 3% in a mo del population. At first glance we attributed the poor degeneracy to the smaller sequencing sample and the slightly higher parent template carry over. Upon closer inspection we observed that 23% of the sequenced population contained multiple repeating primer

PAGE 118

118 inserts. The number of inserts ranged from 2 up to 9 (!). More troubling was the fact that these multiple inserts were not apparent in DNA sequencing data from the pooled population To determine whether the problem was site specific or a more ge neral flaw in the methodology, we an alyzed two additional libraries. Pooled plasmid DNA sequencing data from both the F37X and V294X libraries revealed multiple primer inserts leading us to conclude that the problem was inherent in the PCR method. How mu ltiple primer s are inserted is unknown. Our initial approach to solve this problem was to adapt the Single Primer Reactions IN Parallel (SPRINP) protocol to our methodology. 128 In this approach the PCR reaction is divided in two and each portion is combined with either the forward or reverse mutagenic primers. During the PCR program only one strand is amplified in each reaction. Upon completion the two half reactions are combined and subject to a slow annealing step where, in theory, the two complementary degenerate pr oducts combine. The objective of this modification is to bypass the multiple in sertion mechanism. T his method did eliminate primer conc atemers; however, the resulting pooled libraries possessed poor degeneracy according to the pooled DNA sequencing data Moreover, this protocol yielded a high degree of parent template carry through W e therefore expanded our search for an adequate mutagenesis method. Obtaining degeneracy Since library construction was hindered by low DNA quantities obtained from linear a mplification based protocols, we examined the viability of other methods involving exponential amplification Three approaches were evaluated: Phusion Site directed Mutagenesis Kit protocol, Liu and Naismith 129 and Zheng et al 127 The results of these trials are summarized in Table 4 2

PAGE 119

119 The Phusion Site directed Mutagenesis Kit protocol is a significant deviation fro m previous attempts. In this strategy primers for both template strands are designed to be completely offset from one another but directly adjacent Melting temperature values were calculated according to the manufacture s protocol In addition, t he primers were phosphorylated at the Following the PCR amplification and DpnI digestion steps the mutagenized DNA was treated with ligase prior to transformation. The resulting circular DNA has a higher transformation efficiency that was also desirable in our case. Unfortunately, results using this approach failed to yield qualitatively accepta ble degeneracy. In general, pooled sequencing results appeared to codon (Figure 4 10 ) This was observed consistently even after a degree of annealing temperature optimization therefore the investigation wa s abandoned in search of a more appropriate strategy. The methods described by Liu and Naismith and Zheng et al use partially overlapping primers in which the degenerate codon is positioned in the center of the complementary segment. The methods vary in the length of both the overlapping and non overlapping segments. Ultimately the method described by Zheng et al gave us the best qualitative assessment of degeneracy, lacked an obvious favoring of parent template or other bias (i.e. GGG), and gave no in dication of primer concatemers (in either pooler or individual sequencing experiments). Additionally, their approach was designed and validated for use in the generation of site saturated libraries. Given this we continued to construct all first round li braries using this method, which is described in detail in the experimental section of this Chapter.

PAGE 120

120 It is interesting to note that when we conducted the Zheng et al phosphorylated primers followed by ligation, as done in the Phusion k it, the resulting pooled library overwhelmingly displayed a double primer insert. This implies that the duplex formed by the amplification product is further extended by the polymerase to form blunt ended product. Blunt ended product will transform at a much lower e fficiency than the sticky ended (and potentially circular) product we see in our normal protocol. Ligation, however, shows that a significant portion of the population is in fact blunt activit likely responsible for the PCR derived errors we have observed. Assessing D egeneracy In order to go forward into the scree ning phase of this project with a sample size of approximately 100 mutan ts we needed to ensure that the degree of degeneracy was as high as possible while the occurrence of wild type was minimal. The pooled plasmid sequencing data rarely indicated perfect degeneracy in our library plasmid mixtures. Degeneracy was therefore assessed at two stages: in the in itial population of pooled plasmids and at the individual clone level following transformation into the overexpression strain. After transforming the pooled degenerate plasmid library into our expression strain 95 randomly selected colonies were arrayed and cultured on a 96 well plate along with a wild type control. These bacterial cultures were sequenced in plate format using RCA methodology. From these sequencing results a value for Q codon can b e calculated based on the appearance of each base in the NNK codon using the same treatment previously described using chromatograph peak amplitudes We were able to use these values in combination to assess the quality of our degeneracy and the

PAGE 121

121 efficacy of the Q value. The Q value now serves as both an estima te and a threshold for the construction of a pooled degenerate library. Library Q uality We sequence d 95 mutants along with a wild type control from eleven first round libraries in order to determine the effectiveness of our mutagenesis methodology. The r esults of these sequencing attempts and the Q values associated with each library are summarized in Table 4 4 Among multiple attempts for each position library those with the highest Q pooled values were selected for transformation and sequencing of arra yed mutants. As we had hoped, there was a good correlation between Q pooled (pooled plasmids from the initial transformation) and Q codon f rom sequencing the resulting individual clones derived from the pooled plasmids Averaging the results of all librarie s yields a Q value of 0.71 for both Q pooled and Q codon However, for each library average deviation between Q pooled and Q codon is 0.07. On average libraries with a Q pooled = 0.71 0.07 yield ed 27.4 out of 32 possible codons and 17.5 out of 20 possible a mino acids. The highest number of codons observed across our sample was 31 and the lowest was 22. Only once did we obtain all 20 possible amino acids in a library O ur lowest level was 15. The actual amino acid distribution of these first eleven librar ies is depicted in Figures 4 1 1 thru 4 13 The actual amino acid saturation profile obtained generally correlates but is not an absolute function of the Q pooled values. With NNK degenerate codon usage we expect ed t hat amino acids with 1, 2 and 3 codons would represent 3%, 6%, and 9 %, respectivel y, of the population sampled. Table 4 3 lists the actual amino acid bias we observe using our methodology. The most

PAGE 122

122 over represented amino acids are Gly and Pro, each with only 2 codons but over 10% occurrence a cross our l ibraries. Since glycine uses G G G/T codons and p roline uses C C G/T it is likely that the overabundance of these amino acids is sequence specific and related to hydrogen bonding between base pairs. The most under represented amino acids are Cy s, Ile, and Phe all of which only have one codon in NNK degenerate libraries. There are very few published examples with which to compare our analysis of library quality. We o riginally carried out this analysis to assess the actual content of libraries c reated using our mutagenesis methodology; ultimately however we would also like to know whet h er these results are also typical of similar efforts. Unfortunately, few researchers have published the chromatographic sequence data required for comparison. Bougioukou et al published pie charts showing the base compositions of degenerate libraries similar to our presentation in Appendix A 124 From this we can qualitatively conclude that our degeneracy is similar, or in most cases, better than that obtained by their methodology We originally employed this methodology. A s described above, the degeneracy of pooled plasmid DNA appeared adequate ; unfortunately, sequencing individual library members revealed that a large fraction conatained primer concatemers. Based on our results, we suspect that some portion of the libraries reported earlier may also contain similar concatemer events that were undetected or underestimated during attempts at library validation. Zheng et al reported chromatograph ic sequencing data from their library methodology 127 for both the PCR product and DNA from transformed and pooled

PAGE 123

123 libraries. By our analysis their PCR product gives a Q pooled value of 0.78, which is acceptable by our analysis. However, th ey report pooled sequencing data with a Q pooled = 0.48 that is also visibly high in parent template sequence. They attribute this to inadequate DpnI digestion. This is a plausible explanation given the high transformation efficiency of the circular templ ate plasmid and the fact that they conducted only a single DpnI digestion as opposed to the two sequential digestions that we employed We also purified our PCR products prior to DpnI digestion so as to conduct a buffer exchange for the digestion as pres cribed by Zheng et al Sanchis et al performed similar analyses in developing an improved saturation mutagenesis method 93 however, they imply that they did not c onduct purification at this step. Even though they conducted direct comparisons to the Zheng et al approach and others 91,92 the data they report is significantly high in parent template. Q pooled for their data is consistently less than 0.50. While they do conduct two sequential DpnI digestions it appears that their digestions are poorly effective in removing parent template. Conclusions In summary we have successfully used OYE 1 as a model for identifying active site residues in OYE 2.6. After obtaining structural data for OYE 2.6 only minor changes were made to our original group of targeted residues. These site s were then further exploited using saturation mutagenesis. W e de veloped a successful approach for constructing site saturated mutagenesis libraries. For optimal success our approach requires degenerate primer design that facilitates exponential amplifica tion of template plasmid, robust elimination of template plasmid by DpnI digestion, and subsequent transformation at high efficiency. These

PAGE 124

124 measures ensure elimination of PCR derived errors and minimize the presence of parent template while maximizing deg eneracy of the target codon. Finally, w e have also developed a simple measure of degeneracy based on data provided from standard fluorescence based DNA sequencing data On average our methodology yields an estimate of degeneracy quality, Q pooled = 0.71 0.07. For a sample library of 95 members this translates into an average of 1 7.5 of 20 possible amino acids. This estimate serves as a guide in selecting libraries with sufficient degeneracy for further screening endeavors. In order to achieve all possi ble outcomes the sample library size would need to be increased to compensate for the carry over of wild type from the parent template and any inherent sequence specific bias as a result of the PCR methodology

PAGE 125

125 Figure 4 1. Summary of Chapter 4 experim ental objectives and efforts

PAGE 126

126 Table 4 1. OYE active site measurements. OYE1 Residue Ref. c arbon # OYE2.6 Residue N194 4.1 1 4.4 H191 Y196 6.7 1 7.0 Y193 G72* 6.4 2 4.8 A68 W116 7.0 2 6.8 I113 L118 8.9 2 9.1 L115 H191 7.3 2 7.5 H188 T37 4.0 3 3.9 T35 M39 9.4 3 9.7 F37 Y82 9.0 3 8.6 Y78 F374 9.1 4 8.9 F373 Y375 9.1 4 8.6 Y374 5 4.5 G292 N294** 9.4 5 4.5 N293 P295** 5.1 5 9.0 V294 F249 6.8 6 6.6 F247 N250 8.3 6 8.7 Q248 Distances are measured in angstroms from the carbon of the indicated residue to the nearest ring carbon of the bound ligand. Phenolic carbon is designated as 1 and proceeds clockwise as viewed from above the bound FMN cofactor. *measurement taken from carbon **region of poor sequence alignment w ith OYE2.6 Figure 4 2. OYE 1 active site measurements. This figure depicts the methodology used to conduct measurements listed in Table 4 1 using the crystal structure of OYE 1 (PDB: 1OYB).

PAGE 127

127 Figure 4 3. OYE active site schematic. This figure depic ts a 2 dimensional representation of the distribution of amino acid residues considered during this study. Despite their proximity to the modeled substrate, residues indicated in grey were not selected due to their obscuration by more closely positioned s elections.

PAGE 128

128 Figure 4 4. Sequence alignment of OYE 1 and OYE 2.6.

PAGE 129

129 Figure 4 5. Overlay of OYE1 and OYE2.6 crystal structures The ribbon diagram of OYE 1 (green, PDB: 1OYB) is shown aligned with that of OYE 2.6 ( blue, PDB: 4DF2). Bound ligands, p hy droxybenzaldehyde (OYE 1), p chlorophenol (OYE 2.6), and FMN are depicted in stick form with carbons in grey. The position of proline 295 (OYE 1), which has no analogous residue in OYE 2.6, is indicated with a black arrow.

PAGE 130

130 Figure 4 6. OYE 2.6 active s ite schematic This figure depicts a 2 dimensional representation of the distribution of amino acid residues considered during this study. Despite their proximity to the modeled substrate, residues indicated in grey were not selected due to their obscurati on by more closely positioned selections. A blue line connects the flexible loop segment (residues 292 294).

PAGE 131

131 Figure 4 7. Plasmid map for pBS2 template. Plasmid rendering done using PlasMapper v. 2.0 available at http://wisha rt.biology.ualberta.ca/PlasMapper/.

PAGE 132

132 Figure 4 8. Sample calculation of Q pooled from a degenerate sequencing chromat NNK degeneracy is depicted for a sample codon. A (green), C (blue), G (black), T (red), and average (dashed) peaks are depicted wh ere appropriate. Perfect degeneracy is defined as when all peaks overlap. Fractional peak values are indicated for each curve and the total value at any position sums to 1. Individual base peaks close r to the average yield a higher score for the base, t he position, and theoretically the codon as a whole.

PAGE 133

133 Table 4 2 Summary of mutagenesis development efforts Primers Reference (Method) overlap (5' end) non overlap (3' end) Q codon Notes Bougio u kou et al. 33 0.84 Primer concate mer ins erts (QuikChange ) 15 amino acids E delheit et al. 33 0.41 High parent template (SPRINP) Phusion 27 0.42 High parent template (SDM Kit) "GGG" rich Liu and Naismith 20 25 0.15 High parent te mplate Zheng et al. 25 15 0.69 17 amino acids Figure 4 9. Schematic of primer designs used

PAGE 134

134 Figure 4 10. Pooled plasmid sequencing from Phusion SDM Kit experiment Screen shot image of sequencing results from OYE 2.6 Ile113 NNK saturatio n mutagenesis using the Phusion SDM Kit. G is the predominant species at each position of the codon while the wild type codon (ATT) shows below. While small peaks indicate a degree of degeneracy at the first two positions in the codon they are only slig htly above background of the chromat. At our discretion this run was discarded after visual inspection and no Q value assessment was conducted.

PAGE 135

135 Table 4 3 Amino acid bias. Amino Acid # codons # observed % observed R 3 86 10.6% L 3 76 9.4% S 3 61 7.5% G 2 116 14.3% P 2 84 10.3% A 2 62 7.6% T 2 50 6.2% V 2 48 5.9% W 1 34 4.2% Q 1 31 3.8% E 1 22 2.7% D 1 19 2.3% H 1 19 2.3% M 1 18 2.2% Y 1 16 2.0% K 1 14 1.7% STOP 1 14 1.7% N 1 13 1.6% C 1 11 1.4% I 1 10 1.2% F 1 8 1.0% The number of each amino acid observed is the sum of occurrences in all plate sequenced libraries. Wild type codon appearance is not included in these totals. Amino acids encoded by NNK degeneracy with one, two, and three codons should theoretically provide 3%, 6% and 9% occurrence respectively. Table 4 4 Assessment of first round library degeneracy. Library # amino acids # codons # sequenced Q pool ed Q codon % wt T35X 18 30 79 0.66 0.71 0.05 20 F37X 20 30 74 0.80 0.83 0.03 3 I113X 17 26 84 0.69 0.76 0.07 17 H188X 19 31 92 0.72 0.79 0.07 15 H191X 16 28 91 0.84 0.78 0.06 5 Y193X 17 24 87 0.63 0.70 0.07 14 F247X 19 30 93 0.76 0.80 0.04 13 N293X 18 27 92 0.76 0.72 0.04 5 V294X 17 29 82 0.73 0.65 0.08 22 F373X 16 24 90 0.64 0.60 0.04 3 Y374X 15 22 87 0.62 0.44 0.18 31

PAGE 136

136 Figure 4 11. Amino acid distribution for T35, F37, I113 and H188 libraries Blue bars and numbers indicate the observed count of each amino acid. Grey bars depict the theoretical distribution based on sample size and NNK degeneracy.

PAGE 137

137 Figure 4 12. Amino acid distribution for H191, Y193, F247 and N293 libraries Blue bars and numbers indicate the observed count of each amino a cid. Grey bars depict the theoretical distribution based on sample size and NNK degeneracy.

PAGE 138

138 Figure 4 13. Amino acid distribution for V294, F373 and Y374 libraries Blue bars and numbers indicate the observed count of each amino acid. Grey bars depic t the theoretical distribution based on sample size and NNK degeneracy

PAGE 139

139 CHAPTER 5 ITERATIVE SATURATION MUTAGENESIS: PUTTING IT ALL TOGETHER Introduction In order to complete the evolutionary cycle the libraries created by saturation mutagenesis need ed t o be screened and assessed. The best mutants identified in first generation libraries become template DNA for a second round of saturation mutagenesis, hence the term Iterative Saturation Mutagenesis. This chapter discusses the screening, assessment and selection phases of the laboratory evolu tionary cycle as summarized in Figure 5 1. Experimental Strategy Rather than increase the degree of oversampling required to achieve saturation at a given library position we decided to a ccept the risk associated w ith keeping our screening libraries at 95 members. In short, we assumed that only 17 of 20 amino acids were likely present in any given saturated library. For libraries that display sufficiently interesting screening results we propose d to sequence the 95 member library to identify missing members. Amino acids not represented would then be individually constructed using site directed mutagenesis. In order to screen mutant libraries we decided to use a whole cell assay technique for the benefits of simp licity and time. Expression strains containing unique mutant plasmid s can be cultured in a 96 well plate format and induced. Rather than conducting cell lysis and providing exogenous co factor and the necessary recycling system as done previously 124 this approach allows us to keep the cells intact and use their intracellular NADPH pool and physiologic recycling mechanism Previous experien ce 130,131 led us to conclude that redox reactions involving esters and cyclic

PAGE 140

140 ketones are possible in the whole cell format. Some approaches to whole cell screening include the addition of small amounts of cofactor regeneration enzymes and additives to enhance the permeability of cell membranes. 126,132 Our methodology is more st raightforward and requires only the addition of glucose to the reaction buffer. Induction of mutant proteins is also further simplified using auto induction media as described by Studier. 133 Auto induction media is inoculated and allowed to grow to satura tion overnight. Care must be taken to provide adequate aeration in order to maximize expression in our host strain. Following induction the cell material is isolated and mixed with reaction buffer containing substrate and glucose and the reaction is all owed to proceed for six hours. Conveniently the whole cycle takes approximately 24 hours from inoculation to the start of our analysis phase by GC FID. GC analysis is conducted as described in Chapter 1 Experimental Procedures Total time for a single 96 well plate GC analysis is approximately 32 hours for substrates 1 and 2 and 38 hours for substrate 3 Induction and analysis cycles were stagger ed, where possible, to maximize throughput in our lab. Once first round screening was complete for all lib raries the best results were d library creation. Specifically, plasmid DNA from desirable mutants was isolated and subject to randomization at a site other than its original mutation that also gave an interestin g or desirable result. The resulting second round library was subject to screening as previously described. This process was to be repeated until our desired outcome was achieved. Given that OYE 2.6 reduces all three of our Baylis Hillman adducts efficie ntly and with a strong preference for the ( S ) enantiomer our primary consideration for a positive

PAGE 141

141 result was any mutant that could shift the reduction p roduct toward racemic or even preferentially to the ( R ) enantiomer. At the same time total conversion had to be taken into consideration. Results giving less than 10% conversion regardless of apparent ee%, were not considered for follow Experimental Procedures Degenerate libraries were constructed and arrayed as d escribed in Ch apter 4 into a 96 well format. These libraries were screened as constructed with the exception of the Library Master Plate (LMP) Library Master Plates First round libraries that were sequenced (Table 5 1) to determine the level of degenerac y actual ly present were arrayed into a consolidated format. These LMP s were constructed in accordance with the template depicted in Figure 5 2 The idea is that each codon (if available) is present only once. This eliminates duplicate screening reaction s and makes it possible for a single micro titer plate to contain three different saturated libraries. The resulting plates were then replicated and stored as previously described in Chapter 4 Experimental Procedures Auto I nduction Aliquots ( 20 L ) from each well of freshly prepare d library plates or stabs from library plates stored at 80 C were then used to inoculate a new plate containing 600 L ZYP 5052 auto inducing media (10 g/L tryptone, 5 g/L yeast extract, 1 mM MgSO 4 25 mM (NH 4 ) 2 SO 4 50 mM KH 2 P O 4 50 mM Na 2 HPO 4 5 g/L glycerol, 0.5 g/L anhydrous glucose, 2 g/ L lactose monohydrate) supplemented with 200 mg/L ampicillin. Plates were then mounted in a locally fabricated growth apparatus (Figure 5 3) designed to facilitate maximal oxygen transfer 134 and were shaken at 300 rpm and 37 C

PAGE 142

142 for 16 18 hours. To further ensure maximum oxygen transfer rate (OTR), cells were grown under these con ditions in square well plates 135 Following auto induction, t he cell pellets were harvested by centrifugat ion at 3000 rpm for 30 minutes and the media was removed by aspiration. Whole Cell A ssay s Cell pellets were resuspended in 300 L of 50 mM phosphate buffer, pH 7.0, augmented with 100 mM glucose and 10 mM substrate 1 2 or 3 The plates were shaken at 25 0 rpm and room temperature for 6 hours. 500 L of ethyl acetate was then added to each well and the plates were shaken as abov e for 30 minutes followed by a short centrifugation to facilitate separation of the aqueous and organic phases. The organic layer was then analyzed by chiral phase GC as described in Chapter 1 Site directed M utagenesis The I113D I113F, I113K, Y193C, Y1 93D, Y193K, Y78W/I113C, Y78W/I113W, and Y78W/I113Y mutants were prepared b y a modification of the method previously described in Chapter 4, Experimental Procedures, based on their absence from sequenced plate libraries Each PCR reaction (total volume 100 Phusion BS2 (2 Phusion Hot Start II High Fidelity DNA Polymerase (1 U) was subjected to an initial denaturation step of 98C (30 s) followed by 25 cycles of 98C (10 s) and 72C (4 min) followed by a final incubation at 72C (7 min). Amplicons were purified by DNA spin columns, digested with Dpn I at 37C (10 U for 4 h followed by an additional 10 U for 4 h) and then purified by an additional DNA spin column. E. coli JM109 (75

PAGE 143

143 selection on LB medium supplemented with ampicillin Plasmid DNA was pu rified (spin columns) from randomly chosen colonies and was subsequently analyzed by DNA sequencing to identify the desired OYE 2.6 variants. Once the desired mutations were identified the plasmid DNA was further transformed into E. coli BL21 Gold (DE3) f or the purpose of expression and screening. Results and Discussion Table 5 1 lists all libraries screened in this study. Each was screened against all three Baylis Hillman adducts discussed in Chapter 1. E leven completely sequenced first round libra ries were screened in LMP formats while the others were screened In the latter case, only clones yielding interesting results were sequenced. As previously discussed only the F37X library was complete, containing all 20 amino acid residues after deg enerate primer PCR. We decided to use manual methods to make the missing members of the I113X, Y193X, and Y78W/I113X libraries based on the results of screening ( vide infra ) First Round S creening. Figures 5 4 thru 5 7 graphically depict first round scree ning results from LMP libraries reported by exception only. In very general terms the total conversion for substrate 1 > substrate 2 > substrate 3 Based on these data we identifie d several mutations that influenced the stereoselectivity of OYE 2.6. These results are summarized in Table 5 2. The use of LMP s also provided us with useful information about the active site of OYE 2.6 by comparing those results to mutagenesis results o n analogous positions in the active site of OYE 1.

PAGE 144

144 T35X O YE 2.6 Thr35 is analogous by sequen ce alignment to Thr37 in OYE 1. In the case of OYE 2.6, o nly the T37S mutation retained activity similar to that of the wild type enzyme. This is logical given th e structural similarity of serine and threonine residues, both with a structurally similar hydroxyl group. P artial activity was also present when this position contained Cys, Met, and Ala. It is likely that Thr35 in OYE 2.6 also possibly helps to stabili ze charge on the reduced flavin as does the analogous Thr37 in OYE 1 described previously 32 Given the relatively poor results observed, we did not use this position in further rounds of mutagenesis. In the alkene reductase, YqjM Cys26 al igns with the a nalogous active site position. Bougioukou et al found that this enzyme tolerated several residues at this position (tryptophan, alanine, asparagine, aspartate, glycine, and valine) and these allowed for slight improvements in enzymatic act ivity and stereoselectivity when screened against the model substrate used in this study 124 When the glycine and aspartate mutations were co mbined with additional changes elsewhere in the protein, pronounced effects in stereoselectivity were observed. H188X and H191X Positions His188 and His191 are analogous by alignment to His191 a nd Asn194 in OYE 1. The side chains of these residues likely serve as hydrogen bond donors to the electronegative carbonyl oxygen that polarize the alkene and stabilize the enol ( ate ) intermediate s 31 Interestingly there appears to be no suitable substitute for His188 that retains enzyme activity. Partial conversion is recovered when His191 is substituted with asparagine This is logical given that OYE 1 has an asparagine at the same position ;

PAGE 145

145 however it is clear that OYE 2.6 has a distinct preference for histidine at that same position. A d etailed analysis of the role of these residues in the OYE homolog pentaerythritol tetranitrate (PETN ) reductase from Enterobacter cloacae PB2 has recently been conducted Initial results indic ated that, for a wide range of unsaturated carbonyl compounds, NADPH mediated alkene reduction was largely eliminated in nearly all members of site saturation mutant libraries at the analogous positions ( His181 and His184 ) 136 These results are consistent with our observations for OYE 2.6 A more focused study by the same group on the H181N and H184N mutants revealed that the enantioselectivity for the specific compound 2 phenyl 1 nitropropene (and its p chloro derivative) could be enhanced from ~50% ee (wild type enzyme) to ~90% ee (either H181N or H184N) 137 Unfortunately, a similar improvement in enantioselectivity was not observed in our case with OYE 2.6. However, the suitability of asparagine, at least at position 191 (OYE 2.6 mutant ) and positi on 194 (OYE 1 wild type) is consistent with our results. Further investigation of these positions in the E cloacae enzyme revealed that multiple substitutions at pos itions 181 (A, C, and N) and 184 (A, F, I, N, K, and R) were capable of similar enantiosel ectivity shifts with the model nitro compound 138 On th e other hand these results did not extend to ( R ) carvone, further supporting our observations and lending some hope that a broader substrate screen with our mutants may reveal alter ed enantiopreference s Y193X OYE 2.6 Tyr193 is analogous to Tyr196 in OYE 1. According to Kohli and Massey the role of Tyr at this position is to provide a proton to the stabilized enolate intermediate. 13 Without an appropriately positioned proton donor catalysis should not

PAGE 146

146 occur. In our investigation this would appear to be largely true with the exc eptions of Y193C and Y193T Whi le it seems plausible that cysteine could provide an acidic proton at that position it is difficult to imagine that threonine is sufficiently acidic to carry out the same task Given that these mutations provide near racem ic products it seems plausible that the anionic intermediate undergoes protonat ion by an alternate mechanism. Using the purified mutants, Y193C and Y193T, we conducted isolated enzyme screening of these mutants against a broader list of potential substrate s (Figure 5 8) In general terms, these results show dramatically reduced activity for these two mutants as compared to the wild type enzyme. When screened against 2 methyl substituted cyclic enones the stereoselectivity is still diminished but shows a s light preference for the ( R ) R ) and ( S ) carvone the enantiomeric excess of the products were reduced by 10 20% from that of the wild type enzyme. These results indicate that the formati on of racemic reduction products is not an absolute function of the mutant enzyme At this point it is not clear if the formation of racemic reduction products is due to altered substrate binding or an alternate protonation mechanism, either protein or so lvent in origin. I113X We expected that variations at Ile113 of OYE 2.6 would yield similar results as those observed for the analogous mutations at Trp116 of OYE 1 (results discussed in detail in Chapter 1) This was essentially true for 1 and 2 ; however 3 did not display a wide range of enantioselectiv ity as observed with OYE 1. I113W provided a shift to 11% ee for ( R ) 5 and I113D to 13% (S) 4 qualifying those positions s our second round investigatio ns.

PAGE 147

147 Summary of first round resul ts Table 5 2 summarizes the best results of our first round screening efforts Y78W experiments None of the mutants identified in this round of screening exhibited any propensity to alt er the enantioselectivity for 3 We chose Y78W, I113D, I113W, and V 29 4P as anchors for second round library construction. We deliberately excluded the Phe247 mutants listed in the table because of their relatively low impact on enantioselectivity. We al so chose to exclude the Tyr193 mutants given our estimation that the source of poor enantioselectivity is most likely an alternate protonation mechanism as opposed to an alternate binding mode that we believe is the key to stereoselectivity Second Round S creening Second round libraries were constructed with the best in the first round. Those mutations and their pooled degeneracy assessments are listed in Table 5 1. Table 5 3 summarizes our second round screening efforts In s everal cases the resulting libraries displayed a complete loss of activity. Others retained the stereoselectivity of the original parent and in some cases we saw a significant degradation in conversion. These libraries are indicated in Tabl e 5 3 as no change or low conversion respectively. In two libraries we saw results we classified as va ried. For the library anchored with I113W and randomized at Tyr78 varied results in enantioselectivity were observed for substrate 2 only. However, for all mem bers that displayed activity the results lay between racemic and high in ( S ) character. G iven its lack of improvement in ( R ) enantioselectivity this library was not sequenced

PAGE 148

148 Y78W The second round library for Y78W randomized at Ile113 yields significa nt and varied results for substrates 2 and 3 Figure 5 9 summarizes the data for this screening. Interestingly there is no general change in enantio selectivity when this library was assayed against substrate 1 although a wide variance in data is observ ed for 2 and 3 These results are dep icted graphically in Figures 5 10 and 5 11 These results are similar to the r esults observed for the OYE 1 W116X library against substrate 3 A range of both conversion and enantioselectivity over the full spectrum from ( R ) to ( S ) was observed for both enzymes However, when OYE 2.6 is altered with a Y78W mutation and position 113 is varied this same effect is seen for both 2 and 3 The bottom line is that changing Tyr78 to Trp makes OYE 2.6 behave more like OYE 1. We used the recently solved X ray crystal structure of OYE 2.6 to understand the results of mutagenesis. Overlaying the structure of OYE 1 W116I with substrate 2 bound in the active site in both productive and non productive binding modes (PDB 3RND) with the structure of wild type OYE 2.6 (PDB 4DF2 ) makes the role of Tyr78 evident. Figure 5 12 depicts the relative position of both tyrosine residues in the active site with respect to the bound ligand The distance from the hydroxymethyl oxygen to the tyr osyl oxygen in OYE 2.6 is 1.8 The distance between the same ligand and the analogous residue Tyr82 in OYE 1 W116I is 5.4 This three fold increase in distance is a function backbone positioning within the active site. It is a reasonable hypothesis that when Tyr78 is mutated to a tryptophan the lack of the po tential hydrogen bond with the hydroxymethyl moiety of the Baylis Hillman substrate weakens the affinity for the ( S ) binding mode. As a result changes in the nature of the residue at position

PAGE 149

149 113 have larger effects and are able to now provoke substrate binding in the alternate mode. Y78W, F247X At the same time, the apparent proximity of the hydroxyl moiety of Tyr78 to the modeled substrate (Figure 5 12) offers a n explanation as to OY E 2.6 wild type S ) hydroxymethyl products. Our results suggested that mutation of Tyr78 to Trp is a prerequisite to achieving altered binding of hydroxymethyl substrates. It is then a concern that other positions in the active site that may have influence on substrate binding analogous to that observed by variations at position 113 may have been overlooked during first round screening. To test this hypothesis, an additional library consisting of OYE 2.6 Y78W further randomized at position 247 was prepared. Phe247 is located in the active site in close proximity to where the substrate hydroxyl would be positioned in a pro ( R ) binding mode. In t heory, if the ( S ) stabilizing influence of Ty r 78 were removed by mutation to Trp then it may be able to stabilize the opposite binding mode with an appropriate side chain substitution Initial screening of this library against 1 and 3 revealed dramatically reduced conversion among most variants and mini mal shifts in enantiomeric excess where observed. However, when screened against 2 ten variants with greater than 90% conversion and enantiomeric excess approaching 90% were observed. Sequencing revealed that these mutants were F247H (89% ( R ) ), F247A (8 7% ( R ) ), and F247W ( 85% ( R ) ) These mutants have yet to be purified for further characterization. These observations support our hypothesis that stabilizing the alternate binding mode is possible when the primary mode is disfavored. The observation that three

PAGE 150

150 variants lead to increased enantiopreference mimics the r ange of possible solutions observed when variations are performed at position 113. The varied side chain properties of these three variants (His, Ala, and Trp) seem to imply that the stabilizi ng effect is more general than specific in nature. The single mutation (Y78W) parent enzyme in this screening only displays altered stereoselectivity for 2 Since this mutant offers no change in enantioselectivity for 1 and the enzyme is inactive with 3 it is not surprising that further mutations at position 247 have little effect on these substrates It follows that a more divergent approach to evolving enantiopreference may be required to further improve results for substrates 1 and 3 Purified enzym e screening results OYE 2.6 variants displaying the greatest degree of ( R ) stereoselectivity identified during the systematic investivgation of iterative mutagenesis were purified by GST affinity column and screened for activity as in Chapter 1. Screening resu lts are summarized in Table 5 4 U nder the same conditions wild type OYE 2.6 offers high ( S ) selectivity and full conversion for 10 mM 1 and 2 in less than one hour and within 3 hours for 3 None of the second generation mutants yield ed altered st ereo selectivity for 1 OYE 2.6 Y78W and Y78W/I113M provide for full reduction of 2 to 5 with an enantiomeric excess of 75% ( R ) within 6 or 8 hours respectively. In contrast, Y78W/I113C, Y78W/I113F, Y78W/I113L, and Y78W/I113V yield ed 90% ( R ) 6 to a limit of approximately 50% conversion within 24 hours. This is similar to the results obtained when screening OYE 1 W116V against 3 implying that the mutation of Tyr78 is an absolute requirement to make OYE 2.6 more like OYE 1 in active site topography.

PAGE 151

151 Con clusion s Foremost, we have both developed and demonstrated an approach to saturation mutagenesis that significantly reduces the library size and subsequent screening effort. These tools for directed evolution are therefore now more accessible to non speci alist high throughput laboratories like our own. Significant progress was made toward our goal of achieving enantiocomplementarity for the OYE 2.6 enzyme. For substrates 2 and 3 this result is essentially complete; however, additional work is needed to make these catalysts industrially viable. For reasons unknown, substrate 1 did not yield progress toward achieving the ( R ) enantiomer. In this case, it is plausible that the cyclohexenone ceeds through a dismutation mechanism resulting in enzyme inactivation with a bound phenol(ate) product. This would be especially pronounced if the mutant being evaluated had a reduced affinity for the NADPH substrate. We have demonstrated that an itera tive approach to saturation mutagenesis focused on the active site binding pocket residues in an alkene reductase is capable of altering the enantioselectivity of the parent enzyme ( Figure 5 13 ) In our case, this result was de pendent on the identities of three critical active site residues, Ile113, Tyr78 and Phe247 Previous experience with OYE 1 implicated position 113 as a focal point for altering the enantioselectivity of 2 substituted cyclic enones T his work led to identifying the role of Tyr78 in governing the enantioselectivity of the enzyme toward 2 hydroxymethyl substituted cyclic enones These results were achieved by systematic cross combination of our best results from first round screening. Conjecture led to the further investigation of Ph e247. Further combinations of variants at positions 113 and

PAGE 152

152 247 have a high probability of success in further increasing ( R ) stereoselectivity for hydroxymethyl substituted substrates. Fundamentally, we have demonstrated the ability to direct the binding of functionalized enones within an alkene reductase enzyme.

PAGE 153

153 Figure 5 1. Summary of Chapter 5 experimental objectives and efforts

PAGE 154

154 Figure 5 2. Template for construction of Library Master Plates

PAGE 155

155 Figure 5 3 Growth apparatus for auto inducti on of E. coli cells Layers are machined from 12.4 mm 16.7 mm aluminum sheets and are anodized black to finish In the assembled view (A) a 96 well microtiter p late (not shown) is mounted between the top assembly and the base plate by long screws (shown) This assembly can then be mounted to a shaker platform by screws ( not shown ) In the exploded view (B) a layer of g lass wool is inserted between the top plate and the red silicone rubber layer prior to assembly and autoclave sterilization. Drawings cou rtesy of Todd Prox, Department of Chemistry Machine Shop, University of Florida.

PAGE 156

156 Table 5 1 Libraries screened in this study. Library Q pooled Sequenced Complete T35X 0.66 Yes No F37X 0.80 Yes Yes A68X 0.62 No Unknown Y78X 0.75 No Unknown I113X 0.69 Yes Yes H188X 0.72 Yes No H191X 0.84 Yes No Y193X 0.63 Yes Yes F247X 0.76 Yes No G292X 0.72 No Unknown N293X 0.76 Yes No V294X 0.73 Yes No F373X 0.64 Yes No Y374X 0.62 Yes No Y78W, I113X 0.69 Yes Yes Y78W, F247X 0.63 No Unknown Y78W, V294X 0.7 7 No Unknown I113D, Y78X 0.87 No Unknown I113D, V294X 0.72 No Unknown I113W, Y78X 0.72 No Unknown I113W, V294X 0.76 No Unknown V294P, Y78X 0.90 No Unknown V294P, I113X 0.78 No Unknown *F37X was the only library to provide all 20 amino acids by NNK s aturation mutagenesis alone. All other libraries were completed by site directed mutagenesis where indicated.

PAGE 157

157 Table 5 2. Selected first round screening results. OYE 2.6 Substrate 1 Substrate 2 mutant % conv. % ee % conv. % ee Y78W 99 63 ( R ) I113W 20 11 ( R ) Y193C 49 10 ( R ) 26 2 ( R ) Y193T 15 12 ( R ) I113D 20 13 ( S ) V294P 11 61 ( S ) I113C 99 81 ( S ) 99 83 ( S ) F247Y 99 84 ( S ) F247W 99 86 ( S ) Substrate 3 did not display altered enantioselectivity in any of the first round library screen ings.

PAGE 158

158 Figure 5 4 First round screening results for T35X, F37X and I113X Shaded blocks indicate residues missing from the library. Bold text indicates deviation from wild type observations. N.d. = not determined. All e.e.% are ( S ) unless otherwise i ndicated.

PAGE 159

159 Figure 5 5 First round screening results for H188X, H191X and Y193X Shaded blocks indicate residues missing from the library. Bold text indicates deviation from wild type observations. N.d. = not determined. All e.e.% are ( S ) unless othe rwise indicated.

PAGE 160

160 Figure 5 6 First round screening results for F247X, N293X and V294X Shaded blocks indicate residues missing from the library. Bold text indicates deviation from wild type observations. N.d. = not determined. All e.e.% are ( S ) unle ss otherwise indicated.

PAGE 161

161 Figure 5 7 First round screening results for F373X and Y374X Shaded blocks indicate residues missing from the library. Bold text indicates deviation from wild type observations. N.d. = not determined. All e.e.% are ( S ) unle ss otherwise indicated.

PAGE 162

162 Figure 5 8 Selected substrate screening results for OYE 2.6 Tyr193 mutants Screening reactions were conducted under conditions described in Chapter 1. Substrate loading is 10 mM in all trials.

PAGE 163

163 Table 5 3 Summary second rou nd screening results. Anchor 2nd round Screened substrate Mutation % ee Substrate randomized 1 2 3 Y78W 63 ( R ) 2 I113X n.c. varied varied F247X n.c. varied low conv. V294X n.c. n.c. I113W 11 ( R ) 2 Y78X n.c. varied V294X low conv I113D 13 ( S ) 1 Y78X V294X V294P 61 ( S ) 1 Y78X low conv. I113X low conv. % ee listed is from first round screening for substrates indicated. Varied results from the Y78W an chor are depicted in Figures 5 9, 5 10 and 5 11 Varied results from the I113W anchor result in no improvements in % ee and were not sequenced. Dashes ( ) indicate complete loss of activity. n.c. = no change observed in enantioselectivit y across the screened library. low conv. = less than 5% co nversion in any screened well, insufficient to determine % ee *Results from first round screening. Figure 5 9 Second round screening results for Y78W I113X Shaded blocks indicate residues missing from the library. Bold text indicates deviation f rom wild type observations. N.d. = not determined. All e.e.% are ( S ) unless otherwise indicated.

PAGE 164

164 Figure 5 10 Results of Y78W I113X screening with 2 (hydroxymethyl)cyclopent enone Select data from Figure 5 9 with e.e.% on th e vertical axis and total conversion on the horizontal axis. Figure 5 11 Results of Y78W I113X screening with m ethyl 2 (hydroxymethyl) acrylate Select data from Figure 5 9 with e.e.% on the vertical axis and total conversion on the horizontal axis.

PAGE 165

165 Ta ble 5 4 Summary purified enzyme screening results. Enzyme Mutation(s) Substrate Time (h) % conv % ee OYE 2.6 none 1 <1 99 99 ( S ) OYE 2.6 none 2 <1 99 95 ( S ) OYE 2.6 none 3 3 99 99 ( S ) OYE 2.6 Y78W 2 6 99 75 ( R ) OYE 2.6 Y78W, I11 3M 2 8 99 75 ( R ) OYE 2.6 Y78W, I113C 3 24 57 90 ( R ) OYE 2.6 Y78W, I113F 3 24 48 90 ( R ) OYE 2.6 Y78W, I113L 3 24 64 85 ( R ) OYE 2.6 Y78W, I113V 3 24 47 87 ( R ) OYE 1 W116V 3 24 52 86 ( R ) Fig ure 5 12 The role of Tyr78/Tyr82 in the active site OYE 2.6 ( 4DF2 ) is shown in blue and OYE 1 W116I (3RND) is shown in green. Selected residue side chain carbons are shown in cyan Ligand carbons are shown in grey. Distance between hydroxymethy l and tyrosyl oxygens is 1.8 for OYE 2.6 and 5.4 for OYE1 W116I.

PAGE 166

166 Figure 5 13 Final ISM pathways to observed best results

PAGE 167

167 APPENDIX A DEGENERATE LIBRARY SEQUENCING DATA Figure A 1. Pooled and Observed Sequencing for T35X and F37X

PAGE 168

168 Figure A 2 Pooled and Observed Sequencing for I113X and H 188X

PAGE 169

169 Figure A 3 Pooled and Observed Sequencing for H191X and Y193X

PAGE 170

170 Figure A 4. Pooled and Observed Sequencing for F247X and N293X

PAGE 171

1 71 Figure A 5. Pooled and Observed Sequencing for V294X and F373X

PAGE 172

172 Figure A 6. Pooled and Observed Sequen cing for Y374X

PAGE 173

173 Figure A 7. Pooled Sequencing for A68X, Y78X, and G292X

PAGE 174

174 Figure A 8. Pooled Sequencing for Y78W and I113D Second Round Libraries

PAGE 175

175 Figure A 9. Pooled Sequencing for I113W and V294P Second Round Libraries

PAGE 176

176 APPENDIX B SEQUENCE DAT A TGGCGAATGGGACGCGCCCTGTAGCGGCGCATTAAGCGCGGCGGGTGTGGTGGTTACGCGCAGC GTGACCGCTACACTTGCCAGCGCCCTAGCGCCCGCTCCTTTCGCTTTCTTCCCTTCCTTTCTCG CCACGTTCGCCGGCTTTCCCCGTCAAGCTCTAAATCGGGGGCTCCCTTTAGGGTTCCGATTTAG TGCTTTACGGCACCTCGACCCCAAAAAACTTGATTAGGGTGATGGTTCACGTAGTGGGCCAT CG CCCTGATAGACGGTTTTTCGCCCTTTGACGTTGGAGTCCACGTTCTTTAATAGTGGACTCTTGT TCCAAACTGGAACAACACTCAACCCTATCTCGGTCTATTCTTTTGATTTATAAGGGATTTTGCC GATTTCGGCCTATTGGTTAAAAAATGAGCTGATTTAACAAAAATTTAACGCGAATTTTAACAAA ATATTAACGTTTACAATTTCAGGTGGCACTTTTCGGGGAAATGTGCGCGGAACCCCTATTTG TT TATTTTTCTAAATACATTCAAATATGTATCCGCTCATGAGACAATAACCCTGATAAATGCTTCA ATAATATTGAAAAAGGAAGAGTATGAGTATTCAACATTTCCGTGTCGCCCTTATTCCCTTTTTT GCGGCATTTTGCCTTCCTGTTTTTGCTCACCCAGAAACGCTGGTGAAAGTAAAAGATGCTGAAG ATCAGTTGGGTGCACGAGTGGGTTACATCGAACTGGATCTCAACAGCGGTAAGATCCTTGAG AG TTTTCGCCCCGAAGAACGTTTTCCAATGATGAGCACTTTTAAAGTTCTGCTATGTGGCGCGGTA TTATCCCGTATTGACGCCGGGCAAGAGCAACTCGGTCGCCGCATACACTATTCTCAGAATGACT TGGTTGAGTACTCACCAGTCACAGAAAAGCATCTTACGGATGGCATGACAGTAAGAGAATTATG CAGTGCTGCCATAACCATGAGTGATAACACTGCGGCCAACTTACTTCTGACAACGATCGGAG GA CCGAAGGAGCTAACCGCTTTTTTGCACAACATGGGGGATCATGTAACTCGCCTTGATCGTTGGG AACCGGAGCTGAATGAAGCCATACCAAACGACGAGCGTGACACCACGATGCCTGCAGCAATGGC AACAACGTTGCGCAAACTATTAACTGGCGAACTACTTACTCTAGCTTCCCGGCAACAATTAATA GACTGGATGGAGGCGGATAAAGTTGCAGGACCACTTCTGCGCTCGGCCCTTCCGGCTGGCTG GT TTATTGCTGATAAATCTGGAGCCGGTGAGCGTGGGTCTCGCGGTATCATTGCAGCACTGGGGCC AGATGGTAAGCCCTCCCGTATCGTAGTTATCTACACGACGGGGAGTCAGGCAACTATGGATGAA CGAAATAGACAGATCGCTGAGATAGGTGCCTCACTGATTAAGCATTGGTAACTGTCAGACCAAG TTTACTCATATATACTTTAGATTGATTTAAAACTTCATTTTTAATTTAAAAGGATCTAGGTG AA GATCCTTTTTGATAATCTCATGACCAAAATCCCTTAACGTGAGTTTTCGTTCCACTGAGCGTCA GACCCCGTAGAAAAGATCAAAGGATCTTCTTGAGATCCTTTTTTTCTGCGCGTAATCTGCTGCT TGCAAACAAAAAAACCACCGCTACCAGCGGTGGTTTGTTTGCCGGATCAAGAGCTACCAACTCT TTTTCCGAAGGTAACTGGCTTCAGCAGAGCGCAGATACCAAATACTGTCCTTCTAGTGTAGC CG TAGTTAGGCCACCACTTCAAGAACTCTGTAGCACCGCCTACATACCTCGCTCTGCTAATCCTGT TACCAGTGGCTGCTGCCAGTGGCGATAAGTCGTGTCTTACCGGGTTGGACTCAAGACGATAGTT ACCGGATAAGGCGCAGCGGTCGGGCTGAACGGGGGGTTCGTGCACACAGCCCAGCTTGGAGCGA ACGACCTACACCGAACTGAGATACCTACAGCGTGAGCTATGAGAAAGCGCCACGCTTCCCGA AG GGAGAAAGGCGGACAGGTATCCGGTAAGCGGCAGGGTCGGAACAGGAGAGCGCACGAGGGAGCT TCCAGGGGGAAACGCCTGGTATCTTTATAGTCCTGTCGGGTTTCGCCACCTCTGACTTGAGCGT CGATTTTTGTGATGCTCGTCAGGGGGGCGGAGCCTATGGAAAAACGCCAGCAACGCGGCCTTTT TACGGTTCCTGGCCTTTTGCTGGCCTTTTGCTCACATGTTCTTTCCTGCGTTATCCCCTGAT TC TGTGGATAACCGTATTACCGCCTTTGAGTGAGCTGATACCGCTCGCCGCAGCCGAACGACCGAG CGCAGCGAGTCAGTGAGCGAGGAAGCGGAAGAGCGCCTGATGCGGTATTTTCTCCTTACGCATC TGTGCGGTATTTCACACCGCATATATGGTGCACTCTCAGTACAATCTGCTCTGATGCCGCATAG TTAAGCCAGTATACACTCCGCTATCGCTACGTGACTGGGTCATGGCTGCGCCCCGACACCCG CC AACACCCGCTGACGCGCCCTGACGGGCTTGTCTGCTCCCGGCATCCGCTTACAGACAAGCTGTG Figure B 1. Sequence of pBS2 template plasmid.

PAGE 177

177 (continued) ACCGTCTCCGGGAGCTGCATGTGTCAGAGGTTTTCACCGTCATCACCGAAACGCGCGAGGCAGC TGCGGTAAAGCTCATCAGCGTGGTCGTGAAGCGATTCACAGATGTCTGCCTGTTCATCCGCGT C CAGCTCGTTGAGTTTCTCCAGAAGCGTTAATGTCTGGCTTCTGATAAAGCGGGCCATGTTAAGG GCGGTTTTTTCCTGTTTGGTCACTGATGCCTCCGTGTAAGGGGGATTTCTGTTCATGGGGGTAA TGATACCGATGAAACGAGAGAGGATGCTCACGATACGGGTTACTGATGATGAACATGCCCGGTT ACTGGAACGTTGTGAGGGTAAACAACTGGCGGTATGGATGCGGCGGGACCAGAGAAAAATCAC T CAGGGTCAATGCCAGCGCTTCGTTAATACAGATGTAGGTGTTCCACAGGGTAGCCAGCAGCATC CTGCGATGCAGATCCGGAACATAATGGTGCAGGGCGCTGACTTCCGCGTTTCCAGACTTTACGA AACACGGAAACCGAAGACCATTCATGTTGTTGCTCAGGTCGCAGACGTTTTGCAGCAGCAGTCG CTTCACGTTCGCTCGCGTATCGGTGATTCATTCTGCTAACCAGTAAGGCAACCCCGCCAGCCT A GCCGGGTCCTCAACGACAGGAGCACGATCATGCGCACCCGTGGGGCCGCCATGCCGGCGATAAT GGCCTGCTTCTCGCCGAAACGTTTGGTGGCGGGACCAGTGACGAAGGCTTGAGCGAGGGCGTGC AAGATTCCGAATACCGCAAGCGACAGGCCGATCATCGTCGCGCTCCAGCGAAAGCGGTCCTCGC CGAAAATGACCCAGAGCGCTGCCGGCACCTGTCCTACGAGTTGCATGATAAAGAAGACAGTCA T AAGTGCGGCGACGATAGTCATGCCCCGCGCCCACCGGAAGGAGCTGACTGGGTTGAAGGCTCTC AAGGGCATCGGTCGAGATCCCGGTGCCTAATGAGTGAGCTAACTTACATTAATTGCGTTGCGCT CACTGCCCGCTTTCCAGTCGGGAAACCTGTCGTGCCAGCTGCATTAATGAATCGGCCAACGCGC GGGGAGAGGCGGTTTGCGTATTGGGCGCCAGGGTGGTTTTTCTTTTCACCAGTGAGACGGGCA A CAGCTGATTGCCCTTCACCGCCTGGCCCTGAGAGAGTTGCAGCAAGCGGTCCACGCTGGTTTGC CCCAGCAGGCGAAAATCCTGTTTGATGGTGGTTAACGGCGGGATATAACATGAGCTGTCTTCGG TATCGTCGTATCCCACTACCGAGATATCCGCACCAACGCGCAGCCCGGACTCGGTAATGGCGCG CATTGCGCCCAGCGCCATCTGATCGTTGGCAACCAGCATCGCAGTGGGAACGATGCCCTCATT C AGCATTTGCATGGTTTGTTGAAAACCGGACATGGCACTCCAGTCGCCTTCCCGTTCCGCTATCG GCTGAATTTGATTGCGAGTGAGATATTTATGCCAGCCAGCCAGACGCAGACGCGCCGAGACAGA ACTTAATGGGCCCGCTAACAGCGCGATTTGCTGGTGACCCAATGCGACCAGATGCTCCACGCCC AGTCGCGTACCGTCTTCATGGGAGAAAATAATACTGTTGATGGGTGTCTGGTCAGAGACATCA A GAAATAACGCCGGAACATTAGTGCAGGCAGCTTCCACAGCAATGGCATCCTGGTCATCCAGCGG ATAGTTAATGATCAGCCCACTGACGCGTTGCGCGAGAAGATTGTGCACCGCCGCTTTACAGGCT TCGACGCCGCTTCGTTCTACCATCGACACCACCACGCTGGCACCCAGTTGATCGGCGCGAGATT TAATCGCCGCGACAATTTGCGACGGCGCGTGCAGGGCCAGACTGGAGGTGGCAACGCCAATCA G CAACGACTGTTTGCCCGCCAGTTGTTGTGCCACGCGGTTGGGAATGTAATTCAGCTCCGCCATC GCCGCTTCCACTTTTTCCCGCGTTTTCGCAGAAACGTGGCTGGCCTGGTTCACCACGCGGGAAA CGGTCTGATAAGAGACACCGGCATACTCTGCGACATCGTATAACGTTACTGGTTTCACATTCAC CACCCTGAATTGACTCTCTTCCGGGCGCTATCATGCCATACCGCGAAAGGTTTTGCGCCATTC G ATGGTGTCCGGGATCTCGACGCTCTCCCTTATGCGACTCCTGCATTAGGAAGCAGCCCAGTAGT AGGTTGAGGCCGTTGAGCACCGCCGCCGCAAGGAATGGTGCATGCAAGGAGATGGCGCCCAACA GTCCCCCGGCCACGGGGCCTGCCACCATACCCACGCCGAAACAAGCGCTCATGAGCCCGAAGTG GCGAGCCCGATCTTCCCCATCGGTGATGTCGGCGATATAGGCGCCAGCAACCGCACCTGTGGC G CCGGTGATGCCGGCCACGATGCGTCCGGCGTAGAGGATCGAGATCTCGATCCCGCGAAATTAAT ACGACTCACTATAGGGGAATTGTGAGCGGATAACAATTCCCCTCTAGAAATAATTTTGTTTAAC TTTAAGAAGGAGATATACATAATGACCAAGTTACCTATACTAGGTTATTGGAAAATTAAGGGCC TTGTGCAACCCACTCGACTTCTTTTGGAATATCTTGAAGAAAAATATGAAGAGCATTTGTATG A GCGCGATGAAGGTGATAAATGGCGAAACAAAAAGTTTGAATTGGGTTTGGAGTTTCCCAATCTT CCTTATTATATTGATGGTGATGTTAAATTAACACAGTCTATGGCCATCATACGTTATATAGCTG Figure B 1. Sequence of pBS2 template plasmid (continued).

PAGE 178

178 (continued) ACAAGCACAACATGTTGGGTGGTTGTCCAAAAGAGCGTGCAGAGATTTCAATGCTTGAAGGAGC GGTTTTGGATATTAGATACGGTGTTTCGAGAATTGCATATAGTAAAGACTTTGAAACTCTCAAA GTTGATTTTCTTAGCAAGCTACCTGAAATGCTGAAAATGTTCGAAGATCGTTTATGTCATAAAA CATATTTAAATGGTGATCATGTAACCCATCCTGACTTCATGTTGTATGACGCTCTTGATGTTGT TTTATACATGGACCCAATGTGCCTGGATGCGTTCCCAAAATTAGTTTGTTT TAAAAAACGTATT GAAGCTATCCCACAAATTGATAAGTACTTGAAATCCAGCAAGTATATAGCATGGCCTTTGCAGG GCTGGCAAGCCACGTTTGGTGGTGGCGACCATCCTCCAAAATCGGATCATCTGGTTCCGCGTCA TATGCCCATGTCTTCAGTCAAAATTTCTCCATTGAAGGATTCTGAAGCATTCCAGTCTATCAAA GTTGGTAACAACACTCTTCAAACCAAGATTGTCTATCCACCAACTACTAGA TTTAGAGCTTTAG AAGACCACACTCCTTCTGATTTGCAATTGCAGTACTATGGCGACAGATCCACTTTCCCAGGTAC TTTGCTTATCACTGAAGCTACTTTTGTCTCTCCTCAAGCCTCTGGTTATGAAGGTGCTGCTCCA GGTATTTGGACTGACAAGCACGCTAAAGCATGGAAGGTTATTACTGATAAAGTTCATGCCAACG GTTCTTTCGTTTCAACCCAGTTGATTTTTTTGGGAAGGGTTGCAGATCCAG CTGTTATGAAGAC CCGTGGGTTGAATCCAGTTTCTGCCTCTGCTACTTATGAAAGTGATGCCGCTAAAGAAGCTGCC GAAGCAGTTGGTAACCCTGTTAGAGCTTTGACTACCCAAGAAGTCAAGGATCTTGTTTACGAGG CTTACACCAACGCTGCTCAGAAGGCCATGGATGCTGGTTTCGACTATATTGAACTCCATGCTGC TCATGGCTACCTTTTAGATCAATTTTTGCAACCATGCACCAATCAAAGAAC TGATGAATACGGT GGATCCATTGAGAACAGAGCCAGGTTAATTCTTGAGTTGATTGACCATTTGTCTACCATTGTCG GTGCTGACAAGATTGGTATCAGAATCTCTCCATGGGCTACTTTCCAAAACATGAAGGCTCACAA GGACACTGTTCACCCATTGACTACTTTCTCTTACTTGGTCCACGAATTGCAACAGAGAGCTGAC AAGGGTCAAGGTATTGCCTACATTTCTGTCGTTGAGCCTCGTGTAAGTGGT AACGTCGACGTCT CTGAAGAAGACCAAGCTGGTGACAACGAATTTGTCTCCAAGATCTGGAAGGGTGTTATCTTGAA GGCAGGTAACTACTCCTACGATGCTCCAGAGTTCAAGACATTGAAGGAAGATATCGCTGACAAG CGTACATTAGTTGGCTTCTCCAGATACTTCACCTCGAATCCTAACTTGGTTTGGAAATTGCGTG ATGGAATTGACTTGGTGCCATACGACAGAAACACGTTCTACAGTGACAATA ACTATGGTTACAA TACCTTTTCTATGGATTCCGAAGAGGTTGATAAAGAATTAGAAATCAAGAGAGTTCCTTCGGCC ATTGAAGCTTTGTGATGCGGCCGCACTCGAGCACCACCACCACCACCACTGAGATCCGGCTGCT AACAAAGCCCGAAAGGAAGCTGAGTTGGCTGCTGCCACCGCTGAGCAATAACTAGCATAACCCC TTGGGGCCTCTAAACGGGTCTTGAGGGGTTTTTTGCTGAAAGGAGGAACTA TATCCGGAT Figure B 1. Sequence of pBS2 template plasmid (continued).

PAGE 179

179 APPENDIX C MUTAGENIC PRIMERS Table C 1 List of mutagenic primers. Mutation Sequence T35X Fwd 5' ATCCACCAACT NNK AGATTTAGAGCTTTAGAAGACCACAC 3' T35X Rev 5' GCTCTAAATCT MNN AGTTGGTGGATAG ACAATCTTGGTTT 3' F37X Fwd 5' CAACTACTAGA NNK AGAGCTTTAGAAGACCACACTCCTTC 3' F37X Rev 5' TCTAAAGCTCT MNN TCTAGTAGTTGGTGGATAGACAATCT 3' A68X Fwd 5' TTATCACTGAA NNK ACTTTTGTCTCTCCTCAAGCCTCTGG 3' A68X Rev 5' GAGACAAAAGT MNN TTCAGTGATAAGCAAAGTACCTGGGA 3' Y78X Fwd 5 AAGCCTCTGGT NNK GAAGGTGCTGCTCCAGGTATTTGGAC 3' Y78X Rev 5' GCAGCACCTTC MNN ACCAGAGGCTTGAGGAGAGACAAAAG 3' I113X Fwd 5' CAACCCAGTTG NNK TTTTTGGGAAGGGTTGCAGATCCAGC 3' I113X Rev 5' CTTCCCAAAAA MNN CAACTGGGTTGAAACGAAAGAACCG 3' I113N Fwd 5' CAACCCAGTTG AAT TTTTTGGGAA GGGTTGCAGATCCAGC 3' I113N Rev 5' CTTCCCAAAAA ATT CAACTGGGTTGAAACGAAAGAACCG 3' H188X Fwd 5' ATATTGAACTC NNK GCTGCTCATGGCTACCTTTTAGATCA 3' H188X Rev 5' CCATGAGCAGC MNN GAGTTCAATATAGTCGAAACCAGCAT 3' H191X Fwd 5' TCCATGCTGCT NNK GGCTACCTTTTAGATCAATTTTTGCA 3' H191 X Rev 5' AAAAGGTAGCC MNN AGCAGCATGGAGTTCAATATAGTCGA 3' Y193X Fwd 5' CTGCTCATGGC NNK CTTTTAGATCAATTTTTGCAACCATG 3' Y193X Rev 5' TGATCTAAAAG MNN GCCATGAGCAGCATGGAGTTCAATAT 3' Y193C Fwd 5' CTGCTCATGGC TGT CTTTTAGATCAATTTTTGCAACCATG 3' Y193C Rev 5' TGATCTAAAAG ACA G CCATGAGCAGCATGGAGTTCAATAT 3' Y193D Fwd 5' CTGCTCATGGC GAT CTTTTAGATCAATTTTTGCAACCATG 3' Y193D Rev 5' TGATCTAAAAG ATC GCCATGAGCAGCATGGAGTTCAATAT 3' Y193K Fwd 5' CTGCTCATGGC AAG CTTTTAGATCAATTTTTGCAACCATG 3' Y193K Rev 5' TGATCTAAAAG CTT GCCATGAGCAGCATGGAGTTCAATA T 3' F247X Fwd 5' CATGGGCTACT NNK CAAAACATGAAGGCTCACAAGGACAC 3' F247X Rev 5' TTCATGTTTTG MNN AGTAGCCCATGGAGAGATTCTGATAC 3' G292X Fwd 5' CTCGTGTAAGT NNK AACGTCGACGTCTCTGAAGAAGACCA 3' G292X Rev 5' ACGTCGACGTT MNN ACTTACACGAGGCTCAACGACAGAAA 3' N293X Fwd 5' GTGTA AGTGGT NNK GTCGACGTCTCTGAAGAAGACCAAGC 3' N293X Rev 5' GAGACGTCGAC MNN ACCACTTACACGAGGCTCAACGACAG 3' V294X Fwd 5' TAAGTGGTAAC NNK GACGTCTCTGAAGAAGACCAAGCTGG 3' V294X Rev 5' TCAGAGACGTC MNN GTTACCACTTACACGAGGCTCAACGA 3' F373X Fwd 5' ACAGAAACACG NNK TACAGTGACAATAAC TATGGTTACAA 3' F373X Rev 5' TTGTCACTGTA MNN CGTGTTTCTGTCGTATGGCACCAAGT 3' Y374X Fwd 5' GAAACACGTTC NNK AGTGACAATAACTATGGTTACAATAC 3' Y374X Rev 5' TTATTGTCACT MNN GAACGTGTTTCTGTCGTATGGCACCA 3'

PAGE 180

180 APPENDIX D MUTAGENIC PLASMIDS Table D 1 List of plasmids used i n this study. Plasmid Parent Mutation Description pDJB32 OYE2.6 GST fusion protein pBS2 pDJB32 Y368 silent deletion of NdeI restriction site pBS9 pBS2 I113D single mutation pBS10 pBS2 I113F single mutation pBS11 pBS2 I113K single mutation BTS unna med pAW3 I113C double mutation BTS unnamed pAW3 I113W double mutation BTS unnamed pAW3 I113Y double mutation pAW1 pBS2 I113W isolated from 1 st round screening pAW2 pAW7 I113N double mutation pAW3 pBS2 Y78W isolated from 1 st round screening pAW4 pBS2 I113E isolated from 1 st round screening pAW5 pBS2 V294P isolated from 1 st round screening pAW6 pBS2 I113N isolated from 1 st round screening pAW7 pBS2 F247Y isolated from 1 st round screening pAW8 pAW3 I113F isolated from 2 nd round screening pAW9 pAW3 I 113L isolated from 2 nd round screening pAW10 pAW3 I113M isolated from 2 nd round screening pAW11 pAW3 I113V isolated from 2 nd round screening pAW12 pBS2 Y193C single mutation pAW13 pBS2 Y193D single mutation pAW14 pBS2 Y193K single mutation pAW15 pBS2 Y193T isolated from 1 st round screening

PAGE 181

181 LIST OF REFERENCES (1) Williams, R.; Bruce, N. Microbiol. 2002 148 1607 1614. (2) Toogood, H. S.; Gardiner, J. M.; Scrutton, N. S. ChemCatChem 2010 2 892 9 14. (3) Hollmann, F.; Arends, I. W. C. E.; Holtmann, D. Green Chem. 2011 13 2285 2314. (4) Warburg, O.; Christian, W. Naturwissenschaften 1932 20 688 688. (5) Theorell, H.; Akeson, A. Arch. Biochem. Biophys. 1956 65 439 448. (6) Matthews, R.; Massey, V. J. Biol. Chem. 1969 244 1779 &. (7) Abramovitz, A.; Massey, V. J. Biol. Chem. 1976 251 5327 5336. (8) Nishina, Y.; Kitagawa, T.; Shiga, K.; Watari, H.; Yamano, T. J. Biochem. (9) Abramovitz, A.; Massey, V. J. Biol. Chem. 1976 251 5321 5326. (10) Saito, K.; Thiele, D.; Davio, M.; Lockridge, O.; Massey, V. J. Biol. Chem. 1991 266 20720 20724. (11) Stott, K.; Saito, K.; Thiele, D.; Massey, V. J. Biol. Chem. 1993 268 6097 6106. (12) Vaz, A.; Chakraborty, S.; Massey, V. Biochem. 1995 34 4246 4256 (13) Kohli, R.; Massey, V. J. Biol. Chem. 1998 273 32763 32770. (14) Swiderska, M. A.; Stewart, J. D. J. Mol. Catal. B Enzym. 2006 42 52 54. (15) Mueller, A.; Hauer, B.; Rosche, B. Biotechnol. Bioeng. 2007 98 22 29. (16) Mller, A.; Strmer, R.; Ha uer, B.; Rosche, B. Angew. Chem. Int. Edit. 2007 46 3316 3318. (17) Bougioukou, D. J.; Stewart, J. D. J. Am. Chem. Soc. 2008 130 7655 7658. (18) Padhi, S. K.; Bougioukou, D. J.; Stewart, J. D. J. Am. Chem. Soc. 2009 131 3271 3280. (19) Hall, M.; Stue ckler, C.; Hauer, B.; Stuermer, R.; Friedrich, T.; Breuer, M.; Kroutil, W.; Faber, K. Eur. J. Org. Chem. 2008 1511 1516. (20) Stueckler, C.; Mueller, N. J.; Winkler, C. K.; Glueck, S. M.; Gruber, K.; Steinkellner, G.; Faber, K. Dalton T. 2010 39 8472 84 76. (21) Winkler, C. K.; Stueckler, C.; Mueller, N. J.; Pressnitz, D.; Faber, K. Eur. J. Org. Chem. 2010 6354 6358.

PAGE 182

182 (22) Stueckler, C.; Winkler, C. K.; Bonnekessel, M.; Faber, K. Adv. Synth. Catal. 2010 352 2663 2666. (23) Brenna, E.; Gatti, F. G.; Manf redi, A.; Monti, D.; Parmeggiani, F. Eur. J. Org. Chem. 2011 4015 4022. (24) Korpak, M.; Pietruszk, J. Adv. Synth. Catal. 2011 353 1420 1424. (25) Stueckler, C.; Winkler, C. K.; Hall, M.; Hauer, B.; Bonnekessel, M.; Zangger, K.; Faber, K. Adv. Synth. Ca tal. 2011 353 1169 1173. (26) Meah, Y.; Massey, V. P. Natl. Acad. Sci. 2000 97 10733 10738. (27) Meah, Y.; Brown, B.; Chakraborty, S.; Massey, V. P. Natl. Acad. Sci. 2001 98 8560 8565. (28) Williams, R. E.; Rathbone, D. A.; Scrutton, N. S.; Bruce, N. C. Appl. Environ. Microbiol. 2004 70 3566 3574. (29) Swiderska, M. A.; Stewart, J. D. Org. Lett. 2006 8 6131 6133. (30) Fox, K.; Karplus, P. Structure 1994 2 1089 1105. (31) Brown, B.; Deng, Z.; Karplus, P. J. Biol. Chem. 1998 (32) Xu, D.; Kohli, R M.; Massey, V. P. Natl. Acad. Sci. 1999 96 3556. (33) Jeffries, T.; Grigoriev, I.; Grimwood, J.; Laplaza, J.; Aerts, A.; Salamov, A.; Schmutz, J.; Lindquist, E.; Dehal, P.; Shapiro, H. Nature Biotec hnol. 2007 25 319 326. (34) Bougioukou, D. Ph.D. The sis University of Florida, Gainesville, FL, 2006. (35) Walton, A. Z.; Conerly, W. C.; Pompeu, Y.; Sullivan, B.; Stewart, J. D. ACS Catal 2011 1 989 993. (36) Casas, J.; Sunden, H.; Cordova, A. Tetrahedron Lett 2004 45 6117 6119. (37) Chen, A.; Xu, J .; Chiang, W.; Chai, C. L. L. Tetrahedron 2010 66 1489 1495. (38) Di Bugno, C.; Colombani, S.; Dapporto, P.; Garzelli, G.; Giorgi, R.; Paoli, P.; Subissi, A.; Turbanti, L. Chirality 1997 9 713 721. (39) Crimmins, M. T.; Jacobs, D. L. Org. Lett. 2009 1 1 2695 2698. (40) Mohapatra, D. K.; Bhattasali, D.; Gurjar, M. K.; Khan, M. I.; Shashidhara, K. S. Eur. J. Org. Chem. 2008 2008 6213 6224. (41) Org. Lett. 2009 11 1107 1110.

PAGE 183

183 (42) Guo, H.; O'doherty, G. Org. Lett. 2005 7 39 21 3924. (43) Mlynarski, J.; Ruiz Caro, J.; Frstner, A. Chem Eur J 2004 10 2214 2222. (44) Rezgui, F.; Gaied, El, M. Tetrahedron Lett 1998 39 5965 5966. (45) Mase, N.; Inoue, A.; Nishio, M.; Takabe, K. Bioorg. Med. Chem. Lett. 2009 19 3955 3958. (46) Coutrot, P.; Grison, C.; Bomont, C. Tetrahedron Lett 1994 35 8381 8384. (47) Coutrot, P.; Grison, C.; Bmont, C. J Organometal. Chem. 1999 586 208 217. (48) Corey, E.; Pyne, S.; Su, W. Tetrahedron Lett. 1983 24 4883 4886. (49) Kar, A.; Argade N. P. Synthesis 2005 1234 1236. (50) Drewes, S.; Loizou, G.; Roos, G. Synthetic Commun 1987 17 291 298. (51) Lees, W. J.; Whitesides, G. M. J. Org. Chem. 1993 58 1887 1894. (52) Oertling, H.; Reckziegel, A.; Surburg, H.; Bertram, H. J. Chem Rev 2 007 107 2136 2164. (53) Nie, Y.; Niah, W.; Jaenicke, S.; Chuah, G. K. J Catal 2007 248 1 10. (54) Etzold, B.; Jess, A. Chem Eng Technol 2008 31 1282 1289. (55) Trasarti, A. F.; Marchi, A. J.; Apesteguia, C. R. J Catal 2007 247 155 165. (56) Wolken, W.; Have, Ten, R.; Van Der Werf, M. J Agr Food Chem 2000 48 5401 5405. (57) Gooding, O. W.; Voladri, R.; Bautista, A.; Hopkins, T.; Huisman, G.; Jenne, S.; Ma, S.; Mundorff, E. C.; Savile, M. M. Org Process Res Dev 2010 14 119 126. (58) M wangi, M. T.; Runge, M. B.; Hoak, K. M.; Schulz, M. D.; Bowden, N. B. Chem Eur J 2008 14 6780 6788. (59) Sheldon, R. A. Biochem Soc T 2007 35 1583 1587. (60) Bougioukou, D. J.; Walton, A. Z.; Stewart, J. D. Chem. Commun. 2010 46 8558 8560. (61) Yamamoto, T.; Ujihara, H.; Watanabe, S.; Harada, M.; Matsuda, H.; Hagiwara, T. Tetrahedron 2003 59 517 524. (62) Bergmann, J.; Lofstedt, C.; Ivanov, V.; Francke, W. Eur. J. Org. Chem. 2001 3175 3179.

PAGE 184

184 (63) Riordan, J.; Vallee, B. Meth. Enzymol. 1972 25 494 499. (64) Riebel, B.; Gibbs, P.; Wellborn, W.; Bommarius, A. Adv. Synth. Catal. 2002 344 1156 1168. (65) Bratbak, G.; Dundas, I. Appl. Environ. Microbiol. 1984 48 755. (66) Arnold, F. Acc. Chem Res 1998 31 125 131. (67) Shao, Z.; Arnold, F. C urr Opin Struc Biol 1996 6 513 518. (68) Neylon, C. Nucleic Acids Res 2004 32 1448 1459. (69) Stemmer, W. P. Natl. Acad. Sci. 1994 91 10747 10751. (70) Stemmer, W. Nature 1994 370 389 391. (71) Zhao, H.; Arnold, F. H. Nucleic Acids Res 1997 25 1307. (72) Zhao, H.; Giver, L.; Shao, Z.; Affholter, J. A.; Arnold, F. H. Nature Biotechnol. 1998 16 258 261. (73) Coco, W. M.; Levinson, W. E.; Crist, M. J.; Hektor, H. J.; Darzins, A.; Pienkos, P. T.; Squires, C. H.; Monticello, D. J. Nature Biotec hnol. 2001 19 354 359. (74) Moore, G.; Maranas, C.; Lutz, S.; Benkovic, S. P. Natl. Acad. Sci. 2001 98 3226 3231. (75) Zha, D.; Eipper, A.; Reetz, M. T. ChemBioChem 2003 4 34 39. (76) Ostermeier, M. Trends Biotechnol 2003 21 244 247. (77) Ostermei er, M.; Nixon, A.; Benkovic, S. Bioorg. Med. Chem. 1999 7 2139 2144. (78) Sieber, V.; Martinez, C.; Arnold, F. Nature Biotechnol. 2001 19 456 460. (79) Cadwell, R. C.; Joyce, G. F. Genome Res. 1992 2 28 33. (80) Cirino, P.; Mayer, K.; Umeno, D. Meth. Mol. Biol. 2003 231 3 10. (81) Wong, T.; Zhurina, D.; Schwaneberg, U. Comb Chem High T Scr 2006 9 271 288. (82) Wong, T. S.; Roccatano, D.; Zacharias, M.; Schwaneberg, U. J Mol Biol 2006 355 858 871. (83) Murakami, H.; Hohsaka, T.; Sisido, M. Nature Biotechnol. 2002 20 76 81. (84) Reidhaarolson, J.; Sauer, R. Science 1988 241 53 57.

PAGE 185

185 (85) Hughes, M. D.; Nagel, D. A.; Santos, A. F.; Sutherland, A. J.; Hine, A. V. J. Mol. Biol. 2003 331 973 979. (86) Georgescu, R.; Bandara, G.; Sun, L. Meth Mol. Biol. 2003 231 75 84. (87) Ho, S.; Hunt, H.; Horton, R.; Pullen, J.; Pease, L. Gene 1989 77 51 59. (88) Weiner, M. P.; Costa, G. L.; Schoettlin, W.; Cline, J.; Mathur, E.; Bauer, J. C. Gene 1994 151 119 123. (89) Kretz, K. A.; Richardson, T. H .; Gray, K. A.; Robertson, D. E.; Tan, X.; Short, J. M. Meth. Enzymol. 2004 388 3 11. (90) Hogrefe, H.; Cline, J.; Youngblood, G.; Allen, R. Biotechniques 2002 33 1158 (91) Kirsch, R. D.; Joly, E. Nucleic Acids Res 1998 26 1848 1850. (92) Tseng, W C.; Lin, J. W.; Wei, T. Y.; Fang, T. Y. Anal Biochem 2008 375 376 378. (93) Sanchis, J.; Fernandez, L.; Carballeira, J. D.; Drone, J.; Gumulya, Y.; Hoebenreich, H.; Kahakeaw, D.; Kille, S.; Lohmer, R.; Peyralans, J. J. P.; Podtetenieff, J.; Prasad, S .; Soni, P.; Taglieber, A.; Wu, S.; Zilly, F. E.; Reetz, M. T. Appl Microbiol Biot echnol. 2008 81 387 397. (94) Reetz, M.; Carballeira, J. Nature Protocols 2007 (95) Reetz, M. T.; Bocola, M.; Carballeira, J. D.; Zha, D.; Vogel, A. Angew. Chem. Int. Ed it. 2005 44 4192 4196. (96) Reetz, M. T.; Kahakeaw, D.; Lohmer, R. ChemBioChem 2008 9 1797 1804. (97) Reetz, M. Angew. Chem. Int. Edit. 2002 41 1335 1338. (98) Arnold, F. H. Nature 2001 409 253 257. (99) Bornscheuer, U.; Pohl, M. Curr Opin Chem Biol 2001 5 137 143. (100) Powell, K. A.; Ramer, S. W.; Del Cardayr, S. B.; Stemmer, W. P. C.; Tobin, M. B.; Longchamp, P. F.; Huisman, G. W. Angew. Chem. Int. Edit. 2001 40 3948 3959. (101) Zhao, H.; Chockalingam, K.; Chen, Z. Curr. Opin. Biotech. 2 002 13 104 110. (102) Turner, N. Trends Biotechnol 2003 21 474 478. (103) Cherry, J.; Fidantsef, A. Curr. Opin. Biotech. 2003 14 438 443. (104) Jaeger, K.; Eggert, T. Curr. Opin. Biotech. 2004 15 305 313.

PAGE 186

186 (105) Robertson, D.; Steer, B. Curr. Opin. Chem. Biol. 2004 8 141 149. (106) Lutz, S.; Patrick, W. Curr. Opin. Biotech. 2004 15 291 297. (107) Bommarius, A. S.; Blum, J. K.; Abrahamson, M. J. Curr. Opin. Chem. Biol. 2011 15 194 200. (108) Reetz, M. T. Angew. Chem. Int. Edit. 2011 50 138 17 4. (109) Paramesvaran, J.; Hibbert, E. G.; Russell, A. J.; Dalby, P. A. Protein Eng Des Sel 2009 22 401 411. (110) Zhang, J. H.; Dawes, G.; Stemmer, W. P. C. P. Natl. Acad. Sci. 1997 94 4504. (111) Parikh, M. R.; Matsumura, I. J. Mol. Biol. 2005 35 2 621 628. (112) Reetz, M. T.; Torre, C.; Eipper, A.; Lohmer, R.; Hermes, M.; Brunner, B.; Maichele, A.; Bocola, M.; Arand, M.; Cronin, A.; Genzel, Y.; Archelas, A.; Furstoss, R. Org. Lett. 2004 6 177 180. (113) Chen, C.; Fujimoto, Y.; Girdaukas, G.; Si h, C. J. Am. Chem. Soc. 1982 104 7294 7299. (114) Zocher, F.; Enzelberger, M.; Bornscheuer, U.; Hauer, B.; Schmid, R. Anal Chim Acta 1999 391 345 351. (115) Reetz, M.; Becker, M.; Klein, H.; Stockigt, D. Angew. Chem. Int. Edit. 1999 38 1758 1761. (116) Reetz, M. T.; Wang, L. W.; Bocola, M. Angew. Chem. Int. Edit. 2006 45 1236 1241. (117) Reetz, M. T.; Zheng, H. ChemBioChem 2011 12 1529 1535. (118) Prasad, S.; Bocola, M.; Reetz, M. T. ChemPhysC hem 2011 12 1550 1557. (119) Reetz, M.; Zonta, A.; Schimossek, K.; Liebeton, K.; Jaeger, K. Angew. Chem. Int. Edit. 1997 36 2830 2832. (120) Liebeton, K.; Zonta, A.; Schimossek, K.; Nardini, M.; Lang, D.; Dijkstra, B.; Reetz, M.; Jaeger, K. Chem Biol 2000 7 709 718. (121) Reetz, M.; Wilensek, S.; Zh a, D.; Jaeger, K. Angew. Chem. Int. Edit. 2001 40 3589 (122) Zha, D.; Wilensek, S.; Hermes, M.; Jaeger, K. E.; Reetz, M. T. Chem. Commun. 2001 2664 2665.

PAGE 187

187 (123) Reetz, M. T.; Prasad, S.; Carballeira, J. D.; Gumulya, Y.; Bocola, M. J. Am. Chem. Soc. 201 0 132 9144 9152. (124) Bougioukou, D. J.; Kille, S.; Taglieber, A.; Reetz, M. T. Adv. Synth. Catal. 2009 351 3287 3305. (125) Pompeu, Y.; Stewart, J. D. Unpublished Results (126) Kille, S. Ph .D Dissertation Max Planck Institut fur Kohlenforschung, Mu lheim an der Ruhr 2010 (127) Zheng, L.; Baumann, U.; Reymond, J. Nucleic Acids Res 2004 32 E115. (128) Edelheit, O.; Hanukoglu, A.; Hanukoglu, I. B MC Biotechnol. 2009 9 61. (129) Liu, H.; Naismith, J. BMC Biotechnol. 2008 8 91. (130) Walton, A. Z. ; Stewart, J. D. Biotechnol Progr 2002 18 262 268. (131) Walton, A. Z.; Stewart, J. D. Biotechnol Progr 2004 20 403 411. (132) Schwaneberg, U.; Otey, C.; Cirino, P. C.; Farinas, E.; Arnold, F. H. J Biomol Screen. 2001 6 111 117. (133) Studier, F. Protein Expres Purif 2005 41 207 234. (134) Duetz, W.; Ruedi, L.; Hermann, R.; O'connor, K.; Buchs, J.; Witholt, B. Appl. Environ. Microbiol. 2000 66 2641 2646. (135) Duetz, W.; Witholt, B. Biochem Eng J 2004 17 181 185. (136) Hulley, M. E.; Toogood, H. S.; Fryszkowska, A.; Mansell, D.; Stephens, G. M.; Gardiner, J. M.; Scrutton, N. S. ChemBioChem 2010 11 2433 2447. (137) Fryszkowska, A.; Toogood, H.; Sakuma, M.; Stephens, G. M.; Gardiner, J. M.; Scrutton, N. S. Catal Sci Technol 2011 1 948 957. (138) Toogood, H. S.; Fryszkowska, A.; Hulley, M.; Sakuma, M.; Mansell, D.; Stephens, G. M.; Gardiner, J. M.; Scrutton, N. S. ChemBioChem 2011 12 738 749.

PAGE 188

188 BIOGRAPHICAL SKETCH Lieutenant Colonel Adam Z. Walton graduated f rom the University of California, Davis in 1994 with a Bachelors of Science in Chemistry Upon graduation he was awarded a commission as a 2 nd Lieutenant in the US Army, branched Chemical Corps, via his completion of the Reserve Officers Training Course a t U.C. Davis. In 2001 he was selected to participate in the U S Army Advanced Civilian Schooling (ACS) program where he studied under Dr. Jon Stewart at the University of Florida and earned a Masters of Science in Chemistry. After leaving Florida he ser ved as an Instructor and Assistant Professor in the Department of Chemistry and Life Science at the United States Military Academy, West Point, New York. During his military career he has served in various positions in Louisiana, Alabama, Kentucky, Kansas and New York. Most recently he deplo yed to Afghanistan in support o f Operation Enduring Freedom with the historic 101 st Airborne Division. His military decorations include the Bronze Star Medal, the Meritorious Service Medal, and the Joint Meritorious Un it Award. In 2009 Lieutenant Colonel Walton was once again selected to participate in the U S Army ACS program to pursue a PhD in Chemistry at the University of Florida. Upon completion of these studies he will return to West Point and the Department of Chemistry and Life Science where he will continue to coach, teach, and mentor our