<%BANNER%>

I. NADPH Regeneration Sources in Non-Growing, Engineered Escherichia coli Cells II. Old and New Alkene Reductases as P...


PAGE 1

I. NADPH REGENERATION SOURCES IN NON-GROWING, ENGINEERED ESCHERICHIA COLI CELLS II. EVALUATION OF OLD AND NEW ALKENE REDUCTASES AS POTENTIAL BIOCATALYSTS By DESPINA BOUGIOUKOU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Despina Bougioukou

PAGE 3

To my loving parents: and

PAGE 4

iv TABLE OF CONTENTS page LIST OF TABLES............................................................................................................viii LIST OF FIGURES............................................................................................................x ABSTRACT.....................................................................................................................xvi CHAPTER 1 UNCOVERING THE NADPH REGENERATION PATHWAYS IN E. COLI : TOOLS AND METHODOLOGIES..............................................................................1 Introduction................................................................................................................1 The Past: Pentose Phosphate Pathway Partition......................................................4 Radiorespirometric Measurements.....................................................................8 Destructive Techniques.....................................................................................10 Measurements based on labeled nucleic acids.........................................10 Measurements based on labeled amino acids...........................................11 Regulation of Pentose Phosphate Pathway......................................................13 The Present: The Interplay of NADPH with the Metabolic Network.........................14 Theoretical Methods..........................................................................................15 In silico estimations for NADPH-related fluxes in E. coli ............................20 Computational methods: limitations...........................................................22 Experimental Methods......................................................................................23 NMR analysis for 13C labeled amino acids.................................................24 MS analysis for 13C-labeled amino acids...................................................27 Estimated fluxes from 13C labeling experiments........................................27 Other experimental methods: metabolic regulation analysis of NADPH related metabolic pathways at gene and protein expression level.........35 Experimental methods: limitations.............................................................37 Conclusion...............................................................................................................38 2 INVESTIGATION OF NAPDH REGENERATION SOURCES IN RESTING E. COLI CELLS ENGINEERED TO OVERPRODUCE A NAPDH-DEPENDENT KETOREDUCTASE.................................................................................................40 Introduction..............................................................................................................40 Examples of Enzymatic NADPH Regeneration Methods..................................43 Examples of Microbiological NADPH Regeneration Methods...........................45 Permeabilized E.coli cells..........................................................................45 Intact E.coli whole cells..............................................................................48 Economical Analysis.........................................................................................50

PAGE 5

v Manipulations of NADP(H /+) Levels in E. coli towards the Improvement of Biotechnological Processes...........................................................................52 Description of the System under Study....................................................................54 Experimental Strategy..............................................................................................56 Experimental Procedures.........................................................................................61 Materials............................................................................................................61 Bacterial Strains and Plasmids.........................................................................61 Cell Cultivations................................................................................................61 Analytical Methods............................................................................................62 In vitro Construction of a Linear Targeting Fragment by PCR: Synthesis of Targeting Marker with Long Flanking Homologous Regions (LFH)...............63 Gene Replacement...........................................................................................64 Detection and Analysis of Recombinants..........................................................64 Deletion of the Inserted Marker.........................................................................65 Preparation of [4S-2H] NADPH.........................................................................65 Preparation of [4R-2H] NADPH.........................................................................66 Preparation of D-glucose-1d.............................................................................66 Reduction of Ethyl-acetoacetate by Gre2p GST-fusion Reductase Using NADPDB or NADPDA......................................................................................67 Reduction of Ethyl-acetoacetate by Gre2p GST -fusion Reductase Using a mixture of NADPH and NADPDB...................................................................67 Standard Procedure for Ethyl-acetoacetate Biotransformations with Resting cells...............................................................................................................67 Results.....................................................................................................................68 Characterization of Mutated Strains..................................................................68 MS Data Analysis..............................................................................................76 Evaluation of the MS Data................................................................................85 Physiological conditions.............................................................................85 Calculation of the intracellular ratio NADPH/NADPD from mass data.......86 Productivity of Mutant Strains versus the Wild Type in a Shake Flask.............93 Discussion................................................................................................................95 Future Work.............................................................................................................99 3 E. COLI BL21 (DE3) yqhE STRAIN: A CLEANER HOST FOR OVERPRODUCTION OF NADPH-DEPENDENT KETOREDUCTASES...............101 Introduction............................................................................................................101 Carbonyl Reductases in the E. coli .................................................................102 Experimental Procedures.......................................................................................107 Results and Discussion..........................................................................................107 Characterization of yqhE Strain....................................................................107 Biotransformations..........................................................................................109 Conclusion.............................................................................................................112 4 BIOCATALYTIC REDUCTION OF FUNCTIONALIZED ALKENES: A SURVEY FOR EN-REDUCTASES........................................................................................113 Introduction............................................................................................................113 Biohydrogenation with Whole Cells........................................................................114 En-Reductases: Substrate Specificity....................................................................122 Old Yellow Enzyme Family.............................................................................123

PAGE 6

vi Old yellow enzyme from Saccharomyces carlsbergensis ........................123 Old yellow enzymes from yeasts..............................................................133 Old yellow enzymes from bacteria...........................................................136 Old yellow enzymes from plants..............................................................141 Enoate Reductases.........................................................................................145 Structure-Function Studies among Flavin-Containing En-Reductases...........149 Medium Chain Dehydrogenases Possessing En-Reductase Activity.............158 Short Chain Dehydrogenases Possessing En-Reductase Activity.................167 Unclassified En-Reductases...........................................................................168 Cofactor Specificity of En-Reductases...................................................................173 Concluding Remarks..............................................................................................175 5 IN VIVO RECOMBINATIONAL CLONING OF OLD AND NEW ENREDUCTASES: A COMPARATIVE SUBSTRATE SPECIFICITY STUDY............177 Introduction............................................................................................................177 Experimental Strategy............................................................................................179 Selection of Alkene Reductases Clones.........................................................179 Cloning and Expression of Alkene-Reductases..............................................183 Experimental Procedures.......................................................................................185 Purification of YNL134c...................................................................................185 Materials and supplies.............................................................................185 Enzyme activity assays............................................................................185 Cell growth for protein isolation................................................................186 Preparation of crude extract.....................................................................186 Streptomycin sulfate purification..............................................................187 Desalting by Sephadex G-25...................................................................187 DEAE CL-6B anion exchange chromatography.......................................187 HiLoad Q sepharose anion exchange chromatography...........................187 Dyematrex gel blue A affinity chromatography........................................188 Superdex 200 filtration chromatography..................................................188 HiLoad Q Sepharose anion exchange chromatography..........................189 Sephacryl S-300 Hi Prep filtration chromatography.................................189 Protein Sequencing..................................................................................189 Cloning and Purification of GST-Alkene Reductases......................................189 Biochemicals, plasmids and strains.........................................................189 Cloning of alkene reductases...................................................................190 Isolation of GST-fusion proteins......................................................................191 Regeneration of glutathione resin...................................................................192 Screening of Alkene Reductases....................................................................192 Preparation of 6-alkyl-2-cyclohexen-1-ones............................................192 Preparation of neral and geranial.............................................................193 Biotransformations...................................................................................194 Detection of reduced activated alkenes...................................................194 Results and Discussion..........................................................................................195 Isolation of YNL134c.......................................................................................195 Cloning and Expression of Alkene reductases................................................203 Construction of universal E.coli expression vector..........................................203 Construction of a yeast shuttle universal E.coli expression vector.................204 Cloning of GST-alkene reductases..........................................................205 Expression of GST-alkene reductases....................................................209

PAGE 7

vii Screening of Alkene Reductases....................................................................210 Concluding Remarks and Future Work..................................................................229 APPENDIX A ILLUSTRATION OF CONSTRUCTION OF GENE KNOCK-OUTS IN E.COLI ......232 B LIST OF STRAINS AND PLASMIDS.....................................................................234 C LIST OF PRIMERS................................................................................................235 LIST OF REFERENCES...............................................................................................241 BIOGRAPHICAL SKETCH............................................................................................263

PAGE 8

viii LIST OF TABLES Table page 1-1 Tabulation of glucose passage through the PP pathway.........................................7 1-2 Estimated fluxes though the PP pathway..............................................................16 2-1 Cost of nicotinamide cofactors and their content in E. coli. ...................................41 2-2 Cost analysis of different NADPH regeneration methods and tabulation of TTNs......................................................................................................................51 2-3 Cofactor levels of PHB producing strains and tabulation of TTN...........................54 2-4 Labeled sugars. Cost and possible application....................................................58 2-5 E. coli strains used in this study.............................................................................68 2-6 Colony PCR for pgi knock-out...............................................................................69 2-7 Colony PCR for zwf knock-out...............................................................................71 2-8 Colony PCR for pntAB knock-out..........................................................................73 2-9 Fractional abundance of NADPD...........................................................................91 3-1 Putative AKRs in E. coli .......................................................................................104 3-2 Colony PCR for yqhE knock-out..........................................................................108 3-3 -keto ester reductions by the starting and yqhE knockout strains.....................110 4-1 List of OYEs from yeasts and corresponding substrate specificity......................135 4-2 List of OYEs from bacteria and corresponding substrate specificity....................138 4-3 A comparative study among differ ent OYEs from yeast and bacteria.................141 4-4 List of OYEs from plants and corresponding substrate specificity.......................143 4-5 List of ERs and representative substrate specificity............................................149 4-6 En-Reductases in MDR superfamily....................................................................164 4-7 En-Reductases in SDR superfamily....................................................................168

PAGE 9

ix 4-8 List of unclassified en-reductases........................................................................171 4-9 Cofactor specificity of En-Reductases.................................................................174 5-1 List of alkene reductases cloned in this study.....................................................181 5-2 Percent identity in between OYEs cloned in this study........................................182 5-3 Starting materials for preparation of 6-substituted cyclohexenones....................193 5-4 Observed reduction of 1-octen-3-one or 2-cyclohexenone..................................197 5-5 Purification table for YNL134c.............................................................................199 5-6 Final plasmid constructs of akene reductases.....................................................208 5-7 Unsubstituted cyclo-alkenes................................................................................213 5-8 -substituted cyclo-alkenes.................................................................................213 5-9 -substituted cycloalkenes...................................................................................217 5-10 6-substituted cyclohexenones.............................................................................217 5-11 Perillaldehyde and derivatives.............................................................................219 5-12 N-substituted maleimides....................................................................................223 5-13 Straight chain alkenes.........................................................................................224 5-14 Substituted enals.................................................................................................226 5-15 Nitro-akenes........................................................................................................229 C-1 Oligonucleotides used to create pgi mutant.........................................................235 C-2 Oligonucleotides used to create zwf mutant........................................................236 C-3 Oligonucleotides used to create pntAB mutant....................................................237 C-4 Oligonucleotides used to create the library of En-Reductases............................238

PAGE 10

x LIST OF FIGURES Figure page 1-1 Overview of central metabolic pathways involved in NADPH formation in E. coli ...2 1-2 NADPH sources in E. coli metabolism.....................................................................3 1-3 The fate of G6P, labeled in different positions, via EMP (blue frame) and PP (green frame) pathways...........................................................................................5 1-4 The fate of labeled ACoA in TCA cycle. Enzymes are indicated by their 3-letter code.........................................................................................................................6 1-5 Tools and methods used in the last 25 years for elucidation of metabolic pathways................................................................................................................15 1-6 Basics of Flux Balance Analysis approach............................................................21 1-7 Synergy of experimental and computational methods towards the calculation of metabolic fluxes.....................................................................................................25 1-8 Basics of NMR analysis of biosynthetically directed fractional 13C labeling amino acids......................................................................................................................26 1-9 MS analysis: mass isotopomer distribution (MID) for the molecular ion. ..............28 1-10 Specific rates of NADPH production in glucose (C)and ammonia (N)-limited chemostat cultures of E. coli W3110.....................................................................31 1-11 The percent of NADPH produced by PP pathway, TCA cycle and PntAB transhydrogenase activity in E. coli during standard aerobic batch growth on glucose..................................................................................................................33 1-12 Glucose transport systems in E. coli ......................................................................34 1-13 Poly(3-hydroxybutyrate) synthesis: the acetoacetyl-CoA synthase requires NADPH..................................................................................................................36 2-1 Explanation of the relative prices among nicotinamide cofactors..........................41 2-2 NAD(P)H regeneration strategies and major disadvantages for each method......42 2-3 A substrate-coupled enzymatic regeneration method for NADPH.........................43

PAGE 11

xi 2-4 Glucose-6-phosphate dehydrogenase (G6PDH) as an auxiliary enzyme.............44 2-5 Glucose-6-phosphate dehygrogenase as an NAPDH regeneration catalyst.........44 2-6 Hydrogenase I from Pyrococcus furiosus for NADPH regeneration......................45 2-7 Example of two coupled E. coli strains in a two-phase system..............................46 2-8 Example of an E. coli cell overexpressing both GDH and a reductase.................46 2-9 Example of an E. coli cell overexpressing two auxiliary enzymes and a reductase...............................................................................................................47 2-10 Example of two E. coli strains coupled for regeneration of NADPH in one phase system...................................................................................................................48 2-11 Example of intact E. coli cells capable to regenerate NADPH for ethyl acetoacetate reduction by a ketoreductase...........................................................49 2-12 Example of intact E. coli cells overexpressing a G6PDH and a reductase............49 2-13 Cost of NADP+ as a function of TTN......................................................................50 2-14 Cartoon represents the system under study..........................................................57 2-15 Illustration of the experimental procedure..............................................................59 2-16 Results from colony PCR reactions probing the pgi locus before and after the disruption...............................................................................................................70 2-17 Results from colony PCR reactions probing the zwf locus before and after the disruption...............................................................................................................72 2-18 Results from colony PCR reactions probing the pntAB locus before and after the disruption.........................................................................................................74 2-19 Phenotype analysis of constructed mutants in M9 medium...................................75 2-20 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the wild-type strain (WT) using Glc-1D, under aerobic condition.................................76 2-21 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the wild-type strain (WT) using Glc-1D, under anaerobic condition.............................77 2-22 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the wild-type strain (WT) using Glc-1D, under aerobic and anaerobic conditions. .....78 2-23 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the wild-type strain (WT) using Glc-3D, under aerobic conditions...............................78 2-24 Fate of deuterium atoms of glucose at position 1, 3, 4 and 6 through glycolysis...79

PAGE 12

xii 2-25 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the wild-type strain (WT) using Glc-6,6 D2,..................................................................80 2-26 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by pgi strain using Glc-6,6 D2...........................................................................................81 2-27 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the pgi using Glc-1D..................................................................................................81 2-28 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by pgi strain using Glc-3D................................................................................................82 2-29 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by zwf strain using Glc-3D, or Glc-1D...............................................................................82 2-30 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by zwf strain using Glc-6,6 D2...........................................................................................83 2-31 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by icd strain using Glc-1D................................................................................................83 2-32 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by icd strain using Glc-6,6 D2...........................................................................................84 2-33 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by pntAB strain using Glc-1D, or Glc-6,6 D2..........................................................................84 2-34 Production of EHB by WT strain fed with glycerol under aerobic and anaerobic conditions...............................................................................................................85 2-35 Production of EHB by WT strain fed with glucose under aerobic and anaerobic conditions...............................................................................................................86 2-36 Equation for derivatization of the corrected A71/A72 from the observed ration.......87 2-37 Proposed correction for the NADPD fraction based on the corrected A71/A72 ratio and a kinetic isotope effect factor fKIE............................................................88 2-39 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by Gre2 in presence of NADPH and NADPD..........................................................................89 2-38 Preparation of stereospecific labeled NADPH and elucidation of Gre2 facial selectivity by H-NMR.............................................................................................90 2-40 Product formation of Gre2 overexpressed either in zwf or icd or pntAB strain......................................................................................................................94 2-41 Product formation of Gre2 overexpressed either in pgi or WT strain...................94 2-42 Proposed participation of NADPH regeneration sources in E. coli BL21(DE3)(pAA3) resting cells..............................................................................96

PAGE 13

xiii 2-43 Proposed participation of NADPH regeneration sources in E. coli BL21(DE3)(pAA3) mutant stains under non-growing conditions...........................97 3-1 Substrate specificity of dkgA................................................................................105 3-2 Stereoview of the ribbon structure of DkgA.........................................................105 3-3 Multiple alignment of eight AKRs from E .coli ......................................................106 3-4 Results from colony PCR reactions probing the yqhE locus before and after the disruption.............................................................................................................108 3-5 Growth curves for E. coli BL21(DE3) and yqhE strains.....................................110 3-6 The four possible stereoisomers of ethyl 2-fluoroacetoacetate reduction............111 3-7 Profile of ethyl 2-fluoroacetoacetate reduction from E. coli BL21(DE3) (WT) and yqhE (MT)strains...............................................................................................111 4-1 Preparation of a building block with two chiral centers from citral using bakers yeast as biocatalyst.............................................................................................115 4-2 Proposed hydrogenation of allylic alcohols by bakers yeast...............................115 4-3 Chemo-selective reduction of a trienal by bakers yeast......................................115 4-4 Preparation of goniothalamin, a putative anticancer compound..........................116 4-5 Preparation of raspberry ketone by bakers yeast reduction...............................117 4-6 Servis rules for A and B type reduction in bakers yeast.....................................117 4-7 The first reported enone reductase activity in bakers yeast................................118 4-8 Bakers yeast versus chemical mediated reduction of 4-oxo-isophorone............118 4-9 The six possible microbial transformations of zearalenone.................................119 4-10 Preparation of griseofulvin, an antifungal agent..................................................120 4-11 Biohydrogenation of abscisic acid, a plant stress hormone.................................120 4-12 General scheme of biohydrogenation with Clostridia species.............................121 4-13 A proposed classification of en-reductases.........................................................122 4-14 First generation of substrates and inhibitors for OYE1........................................125 4-15 Representation of catalytic cycle of Old Yellow Enzyme.....................................127 4-16 Second generation of substrates for OYE1.........................................................128

PAGE 14

xiv 4-17 Relative reactivity of substituted en-ones (als) based on Masseys studies........129 4-18 Stereoview of active site architecture of OYE1 with para-hydroxy-benzaldehyde (1OYB).................................................................................................................129 4-19 Stereochemical outcome of reduction of substituted cyclic enones by OYE1.....129 4-20 First generation of desaturation reactions by OYE1 in absence of NADPH........130 4-21 Proposed orientation of ODE above the si face of flavin.....................................131 4-22 Second generation of binding ligands for OYE....................................................131 4-23 Second generation of desaturation reactions by modified-OYE in absence of NADPH................................................................................................................132 4-24 Third and fourth generation of substrates for OYE1............................................133 4-25 Jasmonic acid biosynthesis.................................................................................142 4-26 Biohydrogenation of unsaturated carboxylic acids from three different types of enzymes..............................................................................................................146 4-27 Reductive part of the phenylalanine fermentation pathway in Clostridia sporogenes ..........................................................................................................147 4-28 Close up view of the catalytic microenvironment of OYE1 before (1OYA) and after (1OYB). .......................................................................................................150 4-29 Superposition of OYE1 and YqjM structure.........................................................153 4-30 Multiple alignment of twelve en-reductases belonging to OYE family.................154 4-31 Superposition of OYE1 and Leopr1 structure......................................................155 4-32 Close-up view of loop3 after superposition of five OYE crystal structures...........156 4-33 Birds view of the overall structure of the self-inhibited Leopr3 dimer. ...............157 4-34 Structural differences in between the flavor-bearing en-reductases....................158 4-35 The bifunctional action of AOR............................................................................159 4-36 Close up view of the hydrophobic pocket of AOR from guinea pig......................161 4-37 Proposed catalytic mechanism of the reductase activity of AOR from guinea pig........................................................................................................................161 4-38 Multiple alignment of the four purified AORs.......................................................162 4-39 Multiple alignment of zinc independent en-reductases belonging to MDR superfamily..........................................................................................................163

PAGE 15

xv 4-40 Superposition of two en-reductase structures belonging to the MDR superfamily..........................................................................................................167 4-41 Multiple alignment of the two known short chain en-reductases.........................169 4-42 Stereoview of the catalytic active site of porcine carbonyl reductase..................169 5-1 Different approaches towards the identification of novel biocatalysts..................178 5-2 Construction of the alkene reductase library.......................................................180 5-3 Protein gel after the Superdex 200 filtration column............................................198 5-4 Protein elution pattern from Mono Q column during the NaCl gradient...............201 5-5 Protein elution pattern from Sephacryl S-300 Hi Prep filtration column and protein gel for active fractions..............................................................................201 5-6 Sequence of YLN134c.........................................................................................202 5-7 Multiple alignment of YNL134c subfamily with FAQR amino acid sequence.......203 5-8 Construction of E. coli expression vector pDJB3.................................................204 5-9 Construction of the E. coli expression vector pDJB2...........................................205 5-10 Construction of pDJB5.........................................................................................206 5-11 Homologous recombinational cloning into universal vectors pDJB2 and pDJB3.207 5-12 SDS-PAGE analysis of total cell extr acts overexpressing alkene reductases.....210 5-13 SDS-PAGE analysis of alkene reductases after the affinity column....................210 5-14 Predicted stereochemical outcome for 2-methyl-2-cyclohexenone.....................215 5-15 Predicted stereochemical outcome for 3-methyl-2-cyclohexenone.....................215 5-16 Predicted stereochemical outcome for ( S )-perillaldehyde...................................221 5-17 Predicted stereochemical outcome for ( R )-perillaldehyde...................................221 5-18 An alternative, flipped geometry of ( S )-perillaldehyde in the catalytic center of OYE.....................................................................................................................222 5-19 Proposed mechanism for the Ltb4dh-mediated reduction of perillaldehyde........223 5-20 Proposed mechanism for the reduction of geranial or neral by OYEs.................228

PAGE 16

xvi Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy I. NADPH REGENERATION SOURCES IN NON-GROWING, ENGINEERED ESCHERICHIA COLI CELLS II. EVALUATION OF OLD AND NEW ALKENE REDUCTASES AS POTENTIAL BIOCATALYSTS By Despina Bougioukou December 2006 Chair: Jon D. Stewart Major Department: Chemistry In the first part of this dissertation, an effort was made to define the pathways by which the NADPH is regenerated in E. coli cells under non-growing conditions. We used the reduction of ethyl acetoacetate by a Saccharomyces cerevisiae reductase (Gre2) overexpressed in E. coli as a model system and specifically deuterated glucose (C1d1, C3d1 or C6d2) as isotopomer carbohydrate sources. Deuterium atoms originating at C1, C3 or C6 of glucose are first transferred to NADP+ and then appeared at the C3 atom of ethyl-3-hydroxybutyrate, allowing us to measure the isotopic composition of the intracellular NADPH pool by GC/MS. In addition, we perturbed the NADP(H/+) levels of the host cell by environmental and genetic manipulation means and evaluated the effects of those perturbations on the productivity of the system under study. Our data showed that the NADPH sources present in E. coli cells under growing conditions are also active in non-dividing cells and most likely are not the limited factor for the maximum productivity of this par ticular biotransformation.

PAGE 17

xvii In the second part, a set of 15 genes encoding known and putative alkene reductases were successfully cloned by a rapid recombination technique and overexpressed as GST-fused proteins in E. coli. Their ability to reduce ca. 40 simple enones and enals was evaluated. This study uncovered subtle differences in substrate specificity among the distinct en-reductase subfamilies, as well as the function of a previously uncharacterized in S. cerevisiae gene product.

PAGE 18

1 CHAPTER 1 UNCOVERING THE NADPH REGENERATION PATHWAYS IN ESCHERICHIA COLI: TOOLS AND METHODOLOGIES If youre converting carbohydrate into triglyceride, If you need pentose moieties to make nucleotide, Youll find that Embden-Meyerhof is not the game to play And youll do your biosynthesis the pentose phosphate way -The Pentose Phosphate Shunt Song, Author unknown1 Introduction The reduced form of nicotinamide adenine dinucleotide phosphate, NADPH, plays a central role in fueling biosynthetic reactions and maintaining the redox potential necessary to protect cells against oxidative stress.2 A careful inspection in the modelorganism database for Escherichia coli K-12 (EcoCyc, version 10.0)3 reveals five potential routes for NADPH formation. The exact location of each in the carbon central metabolism of E. coli is depicted in Figure 1-1. The pentose phosphate pathway (PP pathway) provides NADPH in two reacti ons catalyzed by glucose-6-phosphate dehydrogenase (zwf) and gluconate-6-phosphate dehydrogenase (gnd) respectively (Figure 1-2A); the tricarboxylic acid cycle (TCA) produces NADPH to a lesser extent by isocitrate dehydrogenase (icd, Figure 1-2B). In addition, an anaplerotic reaction catalyzed by the malate dehydrogenase (maeB, Figure 1-2C), as well as the membranebound pyridine nucleotide transhydrogenase (pntAB, Figure 1-2D), could serve as catalysts for the regeneration of NADPH. Even though the Embden-Meyerhof-Parnas (EMP) pathway is not directly involved in the NADPH synthesis, the branching points which connect it with the aforementioned pathways are of fundamental importance for control of NAPDH production. The efforts that have been made the last six decades

PAGE 19

2 towards the identification and estimation of the partition of each NADPH-producing step to the overall NADPH pool in E. coli will be described. Figure 1-1. Overview of central metabolic pathways involved in NADPH formation in E. coli Enzymes are indicated by their 3-letter code. Abbreviations: G6P, glucose-6-phosphate dehydrogenase; F6P, fructose-6-phosphate; 6PG, gluconate 6-phosphate; P5P, pentose 5-phosphate; E4P, erythrose 4phosphate; S7P, sedoheptulose 7-phosphate; T3P, triose 3-phosphate; 3PG, glyceraldehyde 3-phosphate; PGA, 3-phosphoglycerate; PEP, phosphoenolpyruvate; PYR, pyruvate; ACA: acetyl coenzyme A; OAA, oxaloacetate; ICT, isocitrate; AKG, -ketoglutarate; SUC, succinate; FUM, fumarate; MAL, malate; GOX, glyoxylate. Pgi, phosphogluconase isomerase; zwf, G6P dehydrogenase; gnd, 6PG, 6-phosphogluconate dehydrogenase; PntAB, membrane-bound transhydrogenase; sthA, soluble transhydrodenase; maeA, MAL dehydrogenase; icd, ICT dehydrogenase; PP pathway, pentose phosphate pathway; EMP, Embden-Meyerhof-Parnas pathway; ED pathway, Entner-Doudoroff pathway; TCA cycle, tricarboxylic acid cycle. (Adapted from references 60 and 50).

PAGE 20

3 Figure 1-2. NADPH sources in E. coli metabolism. A) both Zwf and Gnd are belonging to the PPP, B) the Icd is a part of TCA cycle, C) MaeB participates in an anaplerotic reaction, and D) the PntAB catalyzes the reversible transfer of a hydride ion equivalent between NAD(H) and NADP(H).

PAGE 21

4 The Past: Pentose Phosphate Pathway Partition The PP pathway, also known as the hexose monophosphate shunt (HMS), can be considered as a metabolic scrambler for the cells. It is an alternative pathway for glucose metabolism that shunts some G6P away from the EMP glycolytic pathway (oxidative part) and returns parts of the diverted hexose carbons to the latter pathway as F6P and 3PG (non-oxidative part, Figure 1-3). The extent of the contribution of the PP pathway to glycolysis has been a subject of numerous investigations, as this pathway, among other functions, provides ribose-5-phosphate (R5P) which is needed for nucleic acid synthesis and reduces NADP+ to NADPH, which is needed for amino acid and fatty acid synthesis as well as for the regeneration of glutathione, a key peptide in the antioxidant protection of the cells. In addition, recent studies4 have shown that pathways involved in redox-metabolism also play an important role in biosynthesis of metabolites of industrial importance. Indeed, the knowl edge of the PP pathway flux provides a basis for better understanding of the cell physiology and serves as a guide for directed improvements of performance of industrially exploited organisms.5 In classic isotope studies of metabolism, the fate of a label at a specific position of a substrate is traced by examining the incorporation of the labeled atom into the endproducts (e.g. CO2) or intermediates (e.g. NADPH) of the investigated pathways. Different pathways, even though they utilize the same labeled substrate, produce endproducts which are labeled at different positions and to different extents since distinct enzymes are participating in each of pathway. Thus, using the same substrate, in this particular case glucose, and labeling it at a different position (Figure 1-3 and Figure 1-4) each time, investigators estimated the percentage of glucose that is metabolized through the PP pathway. Their efforts are summarized in Table 1-1 and will be described in the following pages.

PAGE 22

5 Figure 1-3. The fate of G6P, labeled in different positions, via EMP (blue frame) and PP (green frame) pathways. Enzymes are indicated by their 3-letter code. Dashed arrows indicate that more than one step is involved. Hashed lines indicate the branching points in between the two pathways (Adapted from references 2, 3 and 68)

PAGE 23

6 Figure 1-4. The fate of labeled ACoA in TCA cycle. Enzymes are indicated by their 3-letter code. Dashed arrows and lines indi cate the fate of oxaloacetate acid (OAA) after the second and third turn of cycle. Hashed lines indicate the interconnection between the TCA cycle and the anaplelotic reaction catalyzed by maeA or maeB. Open cycles indicated partially labeled atoms (Adapted from references 3 and 24)

PAGE 24

7 Table 1-1. Tabulation of glucose passage through the PP pathway. Year (Ref.) E. coli strain Methods and Tools Physiological conditions PPPa (%) Growing cells 22-38 Resting cells <1 1950 (6) B 14C-1glucose, 14CO2 recovery measured Growing cells infected with T2 virus 6 1953 (7) B Activity of zwf and gnd in cell-free extracts Exponential aerobic growth on glucose 40 1958 (8) B C14-labeled glucose at 1 or 2 or 3 or 6 position, rediorespirometric measurements Resting cells in media optimum for growth 28 Rapid growth 30 1962 (9) W Glucose-1-O18 or glucose-6-O18 recovery of labeled CO2 Slow growth 10 Exponential growth in salts medium 25 Growth period in salts medium 16 Stationary phase in salts medium 10 During or after nitrogen deprivation 13-19 Minimal media with yeast extract or nutrient broth 13-14 Anaerobiosis 7 1967 (10) WT Glucose-1-O18, ratio of isotope recovered CO2 Anaerobiosis plus nitrate 13 K-10 Both PPP and EMP pathway active, ED pathway inactive 1967 (11) K-10 pgi Glucose-1-14C, gluconate-1-14C, radiolabeled alanine measurements Growing cells PPP major, ED pathway minor extent 1968 (12) B Glucose-1-O18, O18 in ribose and deoxyribose moieties measured by liquid chromatography MS Growing cells during early logarithmic phase 30 aPPP: Pentose Phosphate Pathway

PAGE 25

8 Table 1-1. Continued. Year (Ref.) E. coli strain Methods and Tools Physiological conditions PPP(%) 1969 (13) B [1-18O], [2-18O] and [6-18O] glucose and [2-18O] fructose O18 in ribose and deoxyribose moieties measured by liquid chromatography MS Growing cells in minimal media contained a limited amount of glucose 28-30 1972 (14) K-12 Enzymatic and end products determination Transition from aerobiosis to anaerobiosis 20-30 1974 (15) DF100 (K-12 derivative), DF102 (zwf 10fold-higher level mutant) C14-lebeled glucose at 1 or 6 position, rediorespirometric measurements, intracellular intermediates concentration measurements, in vivo and in vitro rates of glucose oxidation via PPP Growing cells during early logarithmic phase 0.4-fold higher rate of glucose oxidation via PPP in the mutant strain K10 Cells growing in glucose minimal medium aerobically 5-10 K10 Anaerobically < 5-10 1977 (16) K10 zwf [1-3H, U -14C]glucose, [3-3H, U -14C]glucose, [4-3H, U -14C]glucose, [6-3H, U -14C]glucose, amino acids from hydrolysates of the cellular protein were separated with an amino acid analyzer and their 3H/14C ratios calculated Aerobically or anearobically 0 1979 (17) DF567 (K-12 derivative) C14-lebeled glucose at 1 position, radiolebeled alanine measurements Cells growing aerobically in minimal medium 16 Radiorespirometric Measurements As the word implies, radiorespirometric methods measure the respiratory function of an organism by a combination of CO2 production and radioactive tracer techniques.8 In early attempts, the evolution of 14CO2 from E. coli cells growing under different conditions in presence of [1-14C]glucose was calculated18 based on the fact that the PP pathway will convert the C-1 carbon of glucose (colored in red in Figure 1-3) into CO2 much faster than the combination of EMP pathway and TCA cycle. However, the 14CO2 produced from the latter pathways (after 3 turns of TCA cycle, Figure 1-4) complicates

PAGE 26

9 the calculations, but a correction for the TCA cycle-associated contribution to the recovered 14CO2 was made by performing a parallel experiment8 in presence of [6-14C]glucose (14CO2 cannot be produced from the PP pathway starting from labeled glucose at position 6, colored in blue in Figure 1-3). To circumvent the two major problems inherent with 14C, namely: i) the possibility of conversion of glucose-6-14C to glucose-1-14C via the isomerization of the triose phosphate and ii) the fact that the final results depend on the relative degree of utilization of the 6 carbon atoms of glucose not only for CO2 but also for the synthesis of various cell components, Rittenberg and Ponticorvo9 developed a procedure based on C18O2 measurements. This approach is independent of the function of TCA cycle. When either glucose-1-18O or glucose-6-18O passes through the EMP pathway, the 18O is lost from the molecule at the formation of the PEP and oxidation of PYR yields nonlabeled CO2. On the other hand, when glucose-1-18O is oxidized to pentose via the PP pathway, the CO2 produced should contain 18O (colored in red in Figure 1-3). Consequently, the concentration of 18O in the CO2 will be proportional to the fraction of glucose-1-18O that enters the PP pathway. Regardless, of the spec ific labeled glucose used, the results led to estimations of between 10 and 40 % for the relative participation of the PP pathway into glucose utilization from the E.coli cells, with high dependence on the physiological conditions under which the experiments were carried out.10 The validity of the radiorespirometric measurements came under severe criticism by Katz and Wood,19, 20 who pointed out that theoretically, the PP pathway should be considered as a cycle and recycling of the re synthesized F6P is possible if the latter is isomerized back to G6P (hashed line in Figure 1-3). In such a case, the degree of recycling plays a vital role on fate of the labeled CO2, which is diluted by unlabeled CO2, and the PP pathway partition is underestimated. Nevertheless Katz and Rognstad,21 with the aid of a 7090 IBM computer and a mathematical model that they wrote in

PAGE 27

10 FOTRAN 2 language to fit the experimental data, showed that especially in E. coli cells, the impact of the carbon randomization during the concurrent operation of the PP and EMP pathway on radiorespirometric measurements is negligible. Destructive Techniques The attractiveness of the radiorespirometric method is based on its simplicity, since it is non-destructive and the only prerequisite is an apparatus to trap the released CO2, though only a single measurement can be made on each sample. On the other hand, due to many assumptions made in the calculations of the results, many investigators tried to verify those val ues by performing measurements on end-products other than CO2. Unfortunately, all these methods required the rupture of the cell membrane with unpredictable consequences on the measurements associated with it. Measurements based on labeled nucleic acids Caprioli and Rittenberg 12, 13, 22 examined the levels of radioactivity in the nucleic acids of the E. coli cells that had been grown on labeled glucose. The logic of these experiments lies on the fact that glucose metabolized exclusively by the PP pathway would produce a nucleoside labeled in the pentose moiety only if it is labeled as [6-18O] ([1-18O]glucose will give rise to unlabeled R5P and C18O2 while [2-18O]glucose would produce labeled in the C-1 oxygen R5P, but the 18O would be lost in the formation of the glycosidic linkage with base). The concentration of 18O in the C-5 atom of the pentose would be equal to that of C-6 oxygen atom of the original glucose. On the other hand, if glucose was metabolized exclusively via the EMP pathway, both [1-18O]glucose and [6-18O]glucose would give rise to [3-18O] 3PG and [2-18O]glucose would form [2-18O] 3PG. Through the reverse action of transketolase (tklAB in Figure 1-3), [5-18O] R5P would be produced from both [1-18O]glucose and [6-18O]glucose and [4-18O] R5P from the [2-18O]glucose. Hence, the growth of bacteria using [1-18O], [2-18O] and [6-18O]glucose as the sole carbon sources was terminated in early logarithmic phase, RNA and DNA were

PAGE 28

11 isolated, enzymatically degraded to the nucleoside level, purified further by liquid chromatography and the content of the O18 label in the pentose of nucleosides was determined directly by mass spectroscopy (M S). Two major conclusions emerged from this study: i) in line with the radiorespirometric measurements, 28-30 % of the pentose arose via the PP pathway, ii) the resynthesis of hexose phosphate is limited to about 3 %. Measurements based on labeled amino acids In an effort to investigate the function of the Entner-Doudoroff pathway in E. coli cells, Zablotny and Fraenkel11 isolated a pgi strain and fed both wild type and mutant with two different carbon sources, glucose-1-14C and gluconate-1-14C. Labeled glucose used by the EMP pathway would give rise to methyl-labeled PYR. Labeled gluconate used by the ED pathway could give rise to carboxyl-labeled PYR and should give unlabeled PYR when used by the PP pathway. Assuming that the alanine comes from the transamination of PYR, the authors isolat ed alanine from lysed cells and measured the radioactivity. Their results ruled out the function of the ED pathway as a third possible route in glucose utilization by E. coli cells. Later, Silva and Fraenkel17 estimated the PP partition at 16 % using the same method. In all previous studies, the amount of NADPH formed relative to the amount of consumed glucose can be calculated indirectly from the percentage of PP pathway since 1 mole of glucose yields 2 moles of NADPH when the pentose phosphate shunt is used exclusively from the cells. In 1977, however, Csonka and Fraenkel16raised a very interesting question: if a zwf E. coli strain, which cannot utilize glucose through the shunt, can grow easily on glucose23 how is it producing the 17 mmols of NADPH required from biosynthetic reactions to form 1 g of cells? In other words, if the PP pathway is not essential, how could NADPH be made in E. coli cells? To address this issue the investigators worked with E. coli strains that had mutations affected either

PAGE 29

12 directly or indirectly, putative NADPH-formi ng reactions. Specifically one of the following enzymes was deleted: phosphoglucose isomerase ( pgi ), glucose 6-phosphate dehydrogenase ( zwf ), isocitrate dehydrogenase ( icd ), gluconate 6-phosphate dehydrase ( edd ), malate dehydrogenase ( mdh ), a factor coupling electron transport to ATP formation ( uncB ) or a combination of them such as pgi edd, icd zwf and pgi uncB mutants. They grew the different strains aerobically or anaerobically in presence of glucose labeled uniformly with 14C and specifically with 3H in positions 1, 3, 4 or 6 and they isolated two pairs of amino acids: threonine and aspartate, or proline and glutamate from protein hydrolysates. This technique ensured that: i) the number of radioactive hydrogens derived from a particular position of glucose (1, 3, 4 or 6) can be normalized per carbon atom (according to 14C) in the compound isolated and ii) since the pairs of the amino acids selected differ by two NADPH-dependant reductions, the different 3H/14C ratios between them reflect labeling of the NADPH pool, even if there were also present radioactive hydrogens remaining carbon-bound from glucose.24 The most important observations from this thorough study are: Hydrogens from the 3 and 4 positions of glucose should only appear in endproducts from the NADPH-dependent action of the gnd and the NADH-dependent action of the 3GP dehydrogenase respectively. In the pgi mutant, under anaerobic conditions, the hydrogen at position 3 made no contribution to the reduction of the end-product isolated where the hydrogen at position 1 or 6 made a similar contributi on aerobically or anaerobically. Therefore, the PP pathway is not an obligatory source, at least anaerobically, and the hydrogens at position 1 or 6 of glucose must serve as reductive hydrides via other pathways. The Icd, a possible source for hydrides from the 6 position of glucose, is not essential even in cells also lacking the Zwf and the radioactivity from the hydrogens at position 1 and 6 is not decreased in a strain lacking the enzyme. Thus, a third source should be account for the labeled NADPH found. The NADH contribution to biosynthesis cannot be explained with a futile cycle proposed by Lowenstein25 which involves the ATP-dependent conversion of PYR

PAGE 30

13 to OAA, its NADH-dependant reduction to MAL (catalyzed by mdh) and the NADPH-dependant oxidative decarboxylation of MAL to PYR (catalyzed by meaB), since a mdh stain gave the same contribution of tritium from the C-4 of glucose as the wild type. The fact that the 4 hydrogen of glucose (NADH via the 3PG dehydrogenase) made a minor but measurable contribution to the biosynthetic reactions could be explained by the action of an energy -linked transhydrogenase since this contribution was abolished in the uncB mutant strain. Most of the drawbacks of the aforementioned study were pointed out from the authors. Namely: i) the possible isotope discrimination against the tritium atom26 should result in underestimation of the PP pathway; a factor of two is suggested as a correction (i.e., the PP pathway contributes to 20-40% of the total NADPH pool), ii) unfortunately strains missing a transhydrogenase (i.e., pntAB ) or the NAPDH-dependant malic enzyme were not available, at that moment, for the construction of all possible mutants, iii) the indirect measurements based on a uncB mutant were of limited value due to marginal growth characteristics of this particular strain. Nonetheless, the message from this work is clear: the PP pathway should not account for more than 50 % of the NADPH pool and it is not an obligatory route for NADPH formation in E. coli cells especially under anaerobic conditions or in zwf mutant. Regulation of Pentose Phosphate Pathway The results of isotopic experiments under various physiological conditions suggested that the PP pathway is regulated.10, 27 Stronger evidence for the regulation of the first step of the shunt came from Orthner and Pizers study.15 They calculated, using a mutant strain with 10-fold higher amount of Zwf compared to wild type, that the shunt was not used much faster in the mutant and that both Zwf and Gnd are not operating at their maximal velocity in the cell. Additionally, according to their data, glucose utilization by the shunt is not limited by the NADP+ or the NADP+/NADPH ratio since the concentration of NADPH in the mutant is not sufficient for the 90% inhibition required to justify the only 40% higher flux. A survey by Silva and Fraenkel17 for effectors that might

PAGE 31

14 control in vivo the Gnd reaction in E. coli was unsuccessful. In an elegant study,28 Wolf et al. showed the growth-rate-dependant alteration of zwf and gnd levels in E. coli They varied the specific growth rate, while using the same carbon source, by feeding the cells with different ratios of glucose/ -methylglucose (a competing with glucose for transport, but non-metabolizable, analog) and observed that the levels of the pertinent enzymes are proportional to growth rate. Interestingly, the regulatory site for the translational control of gnd lies within the protein-coding region of the regulated gene, as proved later by Baker and Wolf.29-31 The Present: The Interplay of NADPH with the Metabolic Network We may describe the first three decades of the NADPH-regeneration research as the reductionist era. At this point, there are well-documented speculations for all the possible sources of NADPH in E. coli metabolism. However, crucial questions for the physiology of the E. coli cell are still unanswered. Namely, are all the NADPH sources active under any physiological conditions? If so, to which extent does each one contribute to the overall NADPH pool and from which factors, if any, is the contribution regulated and how is it adjusted? The reconstruction of all the pertinent biochemical pathways with the help of new tools and methods (Figure 1-5) may give an answer to this problem.68 The results from these efforts are summarized in Table 1-2 with an emphasis on PP pathway flux, and they will be discussed in detail below.

PAGE 32

15 Figure 1-5. Tools and methods used in the last 25 years for elucidation of metabolic pathways.32, 33 Theoretical Methods There is no pure theoretical method for the calculation of a biochemical pathway flux (i.e., the rate at which input metabolites are processed to form output metabolites).68 In this context theoretical methods are considered all the approaches that are not using isotope tracers. Theoretical methods are coming in different flavors68, 69-70 but the keystone in all of them is the principle of mass conservation.

PAGE 33

16 Table 1-2. Estimated fluxes though the PP pathway. Year (Ref.) E. coli Strain Methods and Tools Physiological Conditions PPPa (%) Resting cells 1980 (34) K12 Glucose-1-13C, 1H correlation NMR Cells grown anaerobically in M9 medium 22 Aerobic growing cells without ATP maintenance requirements 90 1992 (35, 36) K12 FBAb Aerobic growing cells with ATP maintenance requirements 20-40 Aerated culture (in a fermenter) overexpressing P22 c2 repressor protein Air contact (in a conical flask) 1995 (37) W 3110 90% unlabeled glucose plus 10% [ U -13C]glucose, 2D [13C,1H]-COSY NMR of the purified and hydrolyzed P22 c2 repressor peptide Anaerobic (in a conical flask after purged the medium with argon) 20-30 1996 (38, 39) MG1655 Enzymatic analysis of endproducts, FBA Fed-batch cultivation of exponential grown VHbexpressing cells Higher PP partition compared to wild type 1996 (40) W 3110 90% unlabeled glucose plus 10%[ U -13C]glucose, 2D [13C,1H]-COSY NMR of the purified and hydrolyzed P22 c2 repressor Aerobic growing cells in minimal medium overexpressing P22 c2 repressor protein 15 JM101 20-30% 1997 (41) JM101 pykA pykF [1-14C]glucose, radiorespirometric method Exponentialphase growing cells in M9 medium Mainly through the PP pathway Slow growth 2 Log-growth phase 9 1998 (42) not reported [2-14C]glucose Stationary phase 16 a PPP: Pentose Phosphate Pathway. bFBA: Flux Balance Analysis.

PAGE 34

17 Table 1-2. Continued. Year (Ref.) E. coli Strain Methods and Tools Physiological Conditions PPP(%) Aerobic growth without recycling 46 Anaerobic growth without recycling 74 1999 (43) W 3110 90% and 92% unlabeled glucose plus 8 and 12% [ U -13C]glucose, 2D [13C,1H]-COSY NMR and GC/MS analysis of biomass samples With recycling 20-30 1999 (44) MG1655 90% unlabeled glucose plus 10% [ U -13C]glucose, 2D [13C,1H]-COSY NMR of hydrolyzed cell proteins, enzymatically analysis of excreted metabolites Fed-batch cultivation of cells carrying a high copy number plasmid under microaerobic conditions 32 Mid-exponential 29 Late exponential 25 1999 (45) B (ATCC 11303) 85-90% unlabeled glucose plus 15-10%[ U -13C] glucose, 2D [13C,1H]-COSY NMR of hydrolyzed cell proteins, metabolic flux ratio (METAFoR) analysis Stationary 24 2000 (46) MG1655 FBA was used with objective of maximizing growth and examination the change in the metabolic capabilities caused by gene deletions Simulated aerobic environment with glucose as carbon source 33 2000 (47, 48) K-12 FBA with objective to calculate the sensitivity of optimal cellular growth to altered flux levels of PP pathway genes Growing cells in glucose minimal medium 2-25% 2000 (49) MC 4100 Glycerol, acetate, glucose, DNA microarray technology for detection of differential transcription profiles of a total of 111 genes involved in central carbon metabolism Exponential growth phase in a minimal medium the PP genes were not regulated significantly with exception of gnd downregulated in glycerol and acetate MG1655 5-10 2001 (50) W 3110 pgi 0.45% (w/v) unlabeled glucose and 0.05% [ U -13C]glucose2D [13C,1H]COSY NMR of hydrolyzed cell proteins, metabolic flux ratio (METAFoR) analysis Mid-exponential growth phase in M9 medium 100

PAGE 35

18 Table 1-2. Continued. Year (Ref.) E. coli Strain Methods and Tools Physiological Conditions PPP (%) JM101 21-44 Climited), 1 (Nlimited) 2002 (51) JM101 ( pykAF) 10% [U-13C]glucose, 2DNMR, METAFoR analysis Glucoselimited chemostat or ammonialimited chemostat 1-28 (Climited) 31 (N-limited) 2002 (52, 53) MG1655 Glucose used as a carbon source, FBA and phenotype phase planes (PhPPs) analysis Growing cells under different oxygen uptake rates Flux decreases as oxygen uptake rate decreases JM101 23 2002 (54) JM101 PTSGlc[1-14C]glucose, [1-13C]glucose and [U-13C]glucose a combination of of genetic, biochemical and NMR techniques Growing cells during logarithmic phase in M9 medium 59 2003 (55) MG1655, W3110, JM101 20% or 100% [U-13C]glucose, GC-MS Midexponential growth-phase, batch cultures 4 (anaerobic)14-20 (fully aerobic) 2003 (56) K12 Glucose, enzyme activity assays and proteome analysis by 2-dimensional electrophoresis (2DE) Aerobic versus microanaerobic growth Zwf and gnd downregulated in microanaerobic growth 2003 (57, 58) JM109 pgi Semi-quantitative RT-PCR, intracellular NADPH measurements, extracellular metabolites analysis Batch cultivation in rich LB medium overproduction of NADPH, high flux through the PP 2003 (59) AB1159 Nicotinamide nucleotide enzyme levels, and cellular ratios of NADPH/ NADH measured after exposure of cells to H2O2 Growing cells in presence of H2O2 The activity of zwf was induced by 2.9 fold upon H2O2 challenge W3110 9 2 (C limited) 4 2 (N limited) W3110 pgi 47 9 (C limited) 37 9 (N limited) 2003 (60) W3110 zwf 90% unlabeled glucose plus 10% [ U -13C]glucose, extracellular metabolites analysis, enzyme activity assays, [13C, 1H]-COSY NMR analysis of amino acids, glycerol, glucose in the hydrolysates Aerobic chemostat cultivation in glucose or ammonialimited minimal medium 0 2 (C limited) 0 2 (N limited)

PAGE 36

19 Table 1-2. Continued. Year (Ref.) E. coli Strain Methods and Tools Physiological Conditions PPP(%) Aerobic chemostat cultivation in glucose minimal medium at dilution rate (D) 0.11 h-1 43 2003 (61) K12 7.5 % [ U -13C]glucose and 7.5 % [1-13C]glucose in presence of naturally labeled glucose, extracellular metabolites analysis, GC-MS analysis of biomass hydrolysates At dilution rate (D) 0.22 h -1 28 BW25113 20 2003 (62, 63) BW25113 gnd 10 % [ U -13C]glucose and 10 % [1-13C]glucose in presence of unlabeled glucose, enzyme activity assays, GC-MS and NMR analysis of biomass hydrolysates Aerobic chemostat cultivation in glucose minimal medium at dilution rate (D) 0.2 h-1 10 K12 34 2004 (64) K12 pykF 10 % [ U -13C]glucose and 10 % [1-13C]glucose in presence of unlabeled glucose, enzyme activity assays, GC-MS and NMR analysis of biomass hydrolysates Aerobic chemostat cultivation in glucose minimal medium at dilution rate (D) 0.2 h-1 79 BW25113 20 2004 (65) BW25113 zwf 10 % [ U -13C]glucose and 10 % [1-13C]glucose in presence of unlabeled glucose, enzyme activity assays, GC-MS and NMR analysis of biomass hydrolysates Aerobic chemostat cultivation in glucose minimal medium at dilution rate (D) 0 2004 (66) MG1655 20% or 100% [U-13C]glucose, GC-MS Aerobic batch cultures in M9 medium 17-23 Cells growing aerobically in glucose limited continuous culture at D= 0.044 h-1 39 % mol mol2005 (67) TG1 (a K-12 strain) FBA extracellular metabolites analysis, determination of ATP, ADP, AMP Cells growing anaerobically in glucose limited continuous culture at D= 0.44 h-1 73 % mol mol

PAGE 37

20 Similar to Kirchhoffs current law, (Figure 1-6) the fundamental equation for the flux balance analysis (FBA) states that a metabolite cannot be accumulated inside the metabolic network, therefore, over long period of time, the formation fluxes of a metabolite must be balanced by the degradation fluxes.72 For the FBA to be applicable:78 All possible biochemical pathways for the particular organism must be known. The stoichiometry of the biochemical reactions constituting the metabolic network should be well defined. The elemental composition of the biomass should be estimated. Physiological data such as nutrients uptake rates, metabolites secretion rates and biomass production rates should be calculated. The assumption of steady-state operation of the bioreaction network should be applied. For a network composed of n metabolites and m fluxes, this yields a system of n differential equations with m unknown fluxes which are time-independent when the network operates under steady-state conditions. Although, a number of constraints are applied to the specific system, based on experimental data, the number of unknown fluxes usually exceeds the number of metabolites and the solution is underdetermined. Thus, a multidimensional solution space satisfie s the constraints. To reach a single flux distribution further biochemical experiments (e.g. identification of active pathways under certain physiological conditions) must be performed and certain objectives need to be set (e.g. maximal cell growth or maximal production of a cofactor or metabolite). Then, an appropriate optimization algorithm may produce a unique solution. In silico estimations for NADPH-related fluxes in E. coli In one of the earliest applications of the FBA approach in E. coli Palsson and coworkers35, 36 calculated that the maximum yield of NADPH produced, per mole of glucose consumed, is 11 moles. When the optimal growth for E. coli in the presence of glucose, was set as objective, the PP pathw ay was found to be used almost exclusively from the bacterium.

PAGE 38

21 Figure 1-6. Basics of Flux Balance Analysis (FBA) approach (Adapted from references 72, 73, and 74)

PAGE 39

22 Realizing based on experimental data that this value is an overestimation they applied energy constraints to their model and showed t hat the PP pathway flux is very sensitive to the ATP maintenance requirements (e.g. maintenance of cellular osmolarity, transmembrane gradient, etc.). In an effort to evaluate the effects of gene knockouts in the E. coli metabolic network, Edwards and Palsson46 applied in silico gene deletions and calculated the resultant effects on growth flux. In line with the experimental data, they reported that the in silico zwf mutant had a growth rate of 99% of the wild type value with a concomitant increase of the TCA flux. In addition, the in silico pgi deletion strain had a decreased normalized growth yield under oxygen-limited conditions. The robustness (i.e. the ability of an organism to operate under a wide range of conditions)75 of E. coli was also examined. When the optimal flux for an essential for the PP pathway gene, tkt was decreased to 15% of that of the in silico wild type, the optimal growth flux did not change, but a utilizat ion of a transhydrogenase that converted NADH to NADPH was observed in addition to an increased TCA cycle flux. Also, under oxygen-limited conditions, the oxidative par t of PP pathway was not functional, the transhydrogenase reaction was again activa ted and the TCA cycle did not operate cyclically.52, 53 In a similar manner, but also measur ing experimentally the energetic state of the cells, Weber et al. demonstrated that the flux towards the PP pathway is increased at the cost of TCA cycle activity when growth rate increases in glucose-limited chemostat cultures.67 Computational methods: limitations Basically, the power of the computational approaches is based on the experimenters failure to estimate the in vivo kinetic parameters in a metabolic network. Under steady-state conditions the time-factor is cancelled out and the differential equation system is transformed to a set of linear equations that can be solved by matrix

PAGE 40

23 theory. In principle, this assumption holds true due to high turnover of the pools of metabolites, especially in continuous cultures and in exponential phase of a batch culture.76 However, several drawbacks are evident:77, 78 Metabolic flux analysis (MFA) fails to predict intracellular in parallel or cyclic metabolic pathways fluxes which are not coupled to measurable fluxes. MFA measures net fluxes. Therefore, it fails in bidirectional reaction steps such as the transaldolase and transketolase reactions in PP pathway.79 Assumptions made on the global currency metabolites ATP, NADH and NADPH are controversial. For example, in microorganisms bearing a transhydrogenase which can equilibrate both NADH and NADPH pools, the typical use of separate balances is an oversimplification. In addition, ATP-futile cycles cannot be estimated precisely.80 The accuracy of MFA on calculated fluxes is directly related to the experimental error of the measured extracellular rates. The postulation that a metabolism of a particular organism is designed by nature to follow a certain objective may or may not be true. Even thought the in silico predictions have been consistent with experimental data in an E. coli wild type strain81 they are questionable in a perturbed system, like the knock-out strains.82 Experimental Methods Nuclear Magnetic Resonance (NMR) spectr oscopy and mass spectroscopy (MS) are the dominant tools used during the last two decades for the elucidation of metabolic networks. Stable isotopes such as 18O, 2H, 13C had been used as tracers in some early studies in combination with MS, but radioactive tracers and the liquid scintillation counting technique was more approachable in the post Word War II era, for metabolic physiologists due to the high sensitivity of this method.83 The last two decades, the advent of high field NMR instruments and the coupling of MS with liquid (LC-MS) or gas chromatography (GC-MS) tipped the balance in favor of stable isotope tracers since problems due to detection limits are eliminated. Considering the fact that carbon is the central atom in biomass it is not a surprise that 13C is the predominant stable isotope for the labeling measurements in the latest studies especially when the 12C and 13C differ not only in mass but also in nuclear spin. Hence, their different physical properties can be detected by either of the abovementioned analytical instruments.

PAGE 41

24 Also, complementary information could be drawn by these two methods since both have the same starting point: Cells are growing in minimal medium containing a carbon labeling source, uniformly and/or 1-13C glucose is used almost exclusively (see Table 12). As the labeled substrate is traveling through the metabolic network, stable isotope labeled carbons are exchanged, diluted, recycled and finally accumulated into different metabolites in a very specific manner, which is dictated from the pathway that they followed.84 When a steady-state condition is established, biomass hydrolysates such as proteinogenic amino acids are subjected to NMR and/or GC analysis and the isotopic enrichment is measured. The isotopomer patterns (i.e. isomers differing only in isotope distribution) of the amino acids reflect the patterns of their related precursors which are usually key intermediate metabolites of the central carbon metabolism. It is not easy to detect these intermediates due to their low abundance in the cell, amino acids detection is relatively easy since approximately 55 % of biomass is proteins.85 Finally, what holds true for the theoretical methods is also true for the experimental studies: The complexity of the metabolic networks makes the mathematical and computational models absolutely necessary t ools for the experimentalists who want to fully exploit the meaning of their experimental data. Thus, in the last step of quantitative analysis, an isotopomer model of metabolism is integrated with an iterative fitting procedure of the measured data and the intracellular metabolic fluxes are derived (Figure 1-7). NMR analysis for 13C labeled amino acids Two types of information are usually derived from NMR 13C isotopomer analysis data.86 The first involves fractional enrichment measurements in which glucose labeled at a specific position (e.g. [1-13C] glucose) is used and the degree of 13C enrichment (i.e., the ratio of molecules bearing a 13C spin at a given position over the total number of molecules) is calculated.87 In the second and the most commonly employed method,

PAGE 42

25 uniformly labeled glucose, diluted with unlabeled glucose, is administrated into the cells and the conserved 13C-13C connectivities in the metabolic network are traced (Figure 18). This technique makes use of the fact that upon 90 % 12C dilution of 13C with random distribution of carbon atoms, the probability that any two adjacent carbon atoms are labeled in an individual molecule is approximately equal to the natural abundance of 13C (1.1 %). However, if the two adjacent carbon atoms in the final product derive from the same source molecule, the probability that both positions are labeled is about 10%. The extent to which a certain carbon atom has neighboring carbons that originate from the same source molecule of glucose is calculated from the quantitative analysis of the 13C fine structures. 88 Figure 1-7. Synergy of experimental and computational methods towards the calculation of metabolic fluxes (Adapted from reference 68).

PAGE 43

26 Figure 1-8. Basics of NMR analysis of biosynthetically directed fractional 13C labeling amino acids. (Adapted from references 87, 88) 2D heteronuclear single-quantum coherence [13C, 1H]-correlation NMR spectroscopy (2D [13C, 1H]-COSY) offers the highest sensitivity in deciphering the 13C-13C spin-spin scalar coupling fine structures. In addition, the two dimensional dispersion of 13C resonances enables the analysis of a mixture of amino acids without further purification. Subsequently, the relative intensities of the superimposed multiplets are

PAGE 44

27 determined and the metabolic flux ratios are estimated using a number of probabilistic equations and keeping in mind the biosynthetic reaction network of E. coli 89 MS analysis for 13C-labeled amino acids Compared to NMR instrumentation, MS offers higher sensitivity in a shorter time frame. Hence, a greater number of smaller volume samples can be analyzed, minimizing the otherwise high cost of labeled glucose and improving the accuracy of the calculated fluxes. The coupling of MS with gas chromatography (GC) enables the analysis of mixture of labeled amino acids eliminating laborious purification steps. However, in contrast to NMR, a derivatization step of the protic sites of the amino acids by silylation is required to increase the volatility of these polar compounds for GC separation before the MS analysis.91 Also, the MS data are difficult to interpret intuitively: for a molecule with three carbons, 23 positional isotopomers are expected, but a MS peak at M+2, for example, represents all isotopomers with two 13C nuclei, irrespective of their location in the carbon backbone (Figure 1-9). Consequently, in addition to analysis of mass isotopomer distribution (MID) for the molecular ion, corresponding fragments should be available for further interpretation. Taking into account that the production rate of each isotopomer depends on the probability of its formation, which is equal to the probability of formation of its precursor metabolite which, in turn, depends on real metabolic fluxes, correlation between isotopomer and MIDs is accomplished by using matrix calculus.90 Estimated fluxes from 13C labeling experiments In one of the earliest attempts to elucidate metabolic fluxes by NMR, Ogino et al. used 1H correlation NMR to follow glucose catabolism in anaerobic E. coli cells under various conditions.34 First the cells were growing in M9 minimal medium in the presence of glucose. Cells harvested at exponential or early stationary phase, were resuspended in the same buffer in a 5-mm NMR tube and incubated in a NMR spectrometer at 30 C,

PAGE 45

28 using glucose -1-13C as the sole carbon source. Spectral assignments for succinate, pyruvate, acetate, lactate and ethanol were made based on chemical shifts, spin coupling patterns and 13C-1H coupling constants. The 13C label is lost as CO2 via the PP pathway, whereas it is incorporated into the methyl carbon of pyruvate via the EMP pathway. Hence, it was estimated that in the case of resting cells or growing cells in M9 medium 22% of the glucose utilized by E. coli follows the PP pathway. Figure 1-9. MS analysis: mass isotopomer distribution (MID) for the molecular ion. (Adapted from reference 90) In 1995 Thomas Szyperski demonstrated the power of biosynthetically directed fractional 13C-labeling of proteinogenic amino acids in estimation of intracellular fluxes in E. coli.37 A K-12 derivative E. coli strain producing the P22 c2 repressor in high levels was used in this study. The overexpre ssed peptide was purified and subjected to hydrolysis. The labeled amino acids were analyzed as described before. Despite the examined physiological conditions from fully aerobic to strictly anaerobic, the flux through the PP pathway was set at 20-30 %. Interestingly, Schmidt et al. using the same set of experimental data, but applying a different mathematical model for the quantification of metabolic fluxes, arrived at twice higher numbers: 46 % under aerobic

PAGE 46

29 and 74 % under anaerobic conditions.43 The authors explained the deviation of their calculations by mentioning that the reversibility of reactions involved in PP pathway (e.g. pgi, tkt, tal) was not considered from the outset. Also, in their estimations the flux from ICT to AKG (the third source of NADPH) in TCA cycle was found to be 25 % under aerobic conditions and 2.2 % under anaerobic conditions. Trying to stimulate a situation met frequently in the biotechnological industry, Fiaux et al. performed a 13C NMR flux ratio analysis on a wild type E. coli MG1655 strain carrying a high copy number plasmid under microanaerobic conditions ([O2] < 0.02 mmol/L).39 The PP phosphate flux was 32%, whereas the flux from AKG to OAA was zero, showing that the TCA cycle is interrupted under anaerobiosis. In 1999 the term METAFoR (metabolic flux ratio) was introduced by Sauer et al. to describe the technique of imprinting the central carbon network history into cell proteins.45 In this study the interstrain differences among four strains (JM101, a K-12 strain, PB25, a pyruvate kinase-deficient JM101 strain, ATCC 11303, a wild type B strain, and KO20, an ethanol producing ATCC 11303 strain) were examined under different growth conditions. The major observations were: i) the impact of enzyme overexpression and gene disruption on central metabolism of E. coli was negligible, ii) of all central carbon fluxes, those in the TCA cycle were the most sensitive to env ironmental conditions and iii) surprisingly, the METAFoR patterns among E. coli strains with different genetic backgrounds were very similar in exponentially growing cells under aerobic conditions. In 2000, Dauner and Sauer introduced the GC-MS analysis of proteinogenic amino acids as a rapid method for the calculation of the isotopomer distributions.91 Three years later, Zhao and Shimizu combined the above technique with the theoretical approach of metabolic flux analysis, in order to estimate the net flux distribution in glucose metabolism in E.coli K-12 strain.61 The flux via the PP pathway was dependant on growth rate with higher values at higher dilution rates. This observation suggested that

PAGE 47

30 under increased growth rate, the flux through icd could not fulfill the NADPH demand for growth and as a consequence higher flux through the PP pathway is required. The old and popular method of creating mutants and reexamining the differences in fluxes in the genetically perturbed strains compared to the parent strain provide a new insight into metabolism of E. coli cell in recent years. Precise gene deletion in E.coli is a routine technique nowadays that the whole E. coli genome is annotated. Using METAFoR analysis by NMR, Canonaco et al. reinvestigated the impact of pgi deletion on glucose catabolism and confirmed the flux rerouting via the PP pathway.50 The observation of reduced specific growth and glucose-uptake rates due to Pgi-knockout was explored further by overexpressing the soluble transhydrogenase SthA in the mutant and the parent strain. While no difference was found in the wild type strain, the specific growth of the Pgi mutant increased by about 25%, but it could not be restored to 100% even when a 10-fold higher overexpression of SthA was attempted. This suggested that the kinetic limitation of glucose catabolism in the pgi strain is due to both the inability of the PP pathway to support higher fluxes and the limited capability of E.coli to reoxidize excess amounts of NADPH. Also, Fischer and Sauer observed a 30 % contribution of ED pathway for the pgi mu tant when they applied a GC-MS analysis to labeled proteinogenic amino acids from E. coli .55 The phenomenon could be explained by recalling that the ED pathway produces twice fewer moles of NADPH per mole of glucose compared to the PP pathway and thus it may serve as an alternative route for the reduction of the excess amounts of NADPH in pgi-knockout. Shimizus group combined the 13C-labeling data from NMR analysis with biomass composition and extracellular flux data in order to quantify intracellular fluxes in pgi and zwf knockouts.60 Interestingly, in an effort to minimize the overproduction of NADPH, the Pgi mutant bypassed the isocitrate dehydrogenase reaction by rerouting carbon flow through the glyoxylate shunt which is otherwise inactive in the wild type. Disruption of

PAGE 48

31 G6P dehydrogenase leads to activation of transhydrogenase activity towards the production of NADPH, especially in a ni trogen limited chemostat. Elevated maeB activity was also apparent in this mutant. The overall picture of regeneration rates of NADPH as they were estimated in this study is depicted in Figure 1-10. It is worth noting that the rates attributed to MaeB are based on the assumption that both NAD+ and NADP+ dependant malic enzymes (MaeA and MaeB) are equally active. Figure 1-10. Specific rates of NADPH production in glucose (C)and ammonia (N)limited chemostat cultures of E. coli W3110. The numbers in doughnuts are in mmol g-1 h-1 units. (Adapted from reference 60) A similar pattern with the one observed for the zwf mutant65 was also found when the 6-phosphogluconate dehydrogenase (gnd) deletion was studied by Jiao et al.62 Activation of the ED pathway, induction of malic enzyme and decreased flux through the PP pathway were the major features of this strain. The abovementioned work by Hua et al.60 introduced experimentally, for the first time since 1977,16 the transhydrogenases as players in NADPH production, at least in zwf mutant strain. With Enterobacteriaceae being the only known microbes containing both the soluble, SthA and the membrane bound, PntAB transhydrogenase92 the question of which contributes to NADPH pool remained open. Speculations for the physiological role of PntAB as NAPDH producer had existed since 1980 when Hanson and Rose carried out phenotypical studies on a zwf pnt double mutant and showed that

PAGE 49

32 it grew slower than the zwf mutant, as it would be expected when pntAB produces NADPH in absence of the PP pathway.93 The experimental answer for the physiological role of PntAB in the wild type E. coli strain came from Sauers group in 2004.66 First, in addition to the pgi strain, a new set of mutant strains such as pntAB udhA ( udhA pntAB ) and ( zwf edd eda ) were constructed. Second, the consumption rate of NADPH was estimated based on the growth rate of each mutant strain. Third, the specific rate of NADPH formation from PP phosphate pathway, from the icd and from the maeB was calculated based on METAFoR analysis of 13C-labeled amino acids by GC-MS and metabolic flux analysis. In the pgi strain, the specific rate of NADPH formation was two-fold higher than the rate of NADPH consumption. In pntAB ( udhA pntAB ) and ( zwf edd eda ) strains, the two rates were almost equal, but in udhA as well as in the wild-type strain, the rate of NADPH depletion surpassed by 20-40% the total rate of NADPH production from the three aforementioned source. These measurements led the authors to conclude that about 40% of the NADPH required to compensate for the imbalance between NADPH depletion and formation in the wild type strain must come from a transhydrogenase activity, especially from the PntAB transhydrogenase (Figure 1-11). Another interesting outcome from this study was the observation that the ( pgi udhA ) strain had lethal phenotype on glucose. Growth was recovered only upon plasmid expression of udhA. This result showed that the physiological role of SthA is to reoxidize the excess amounts of NADPH, restoring the imbalance of NADPH metabolism in the pgi strain. Besides the enzymes within the glucose-6-phosphate (G6P) node, enzymes related to phosphoenolpyruvate (PEP) node have also been examined for their ability to control the NADPH metabolism. One modification was the disruption of the pykA and pykF (Figure 1-12), which encode the two pyruvate kinases (PKs) in E. coli Radiorespirometric assays by Ponce et al.41 with resting cells of the mutant strain,

PAGE 50

33 lacking both PKs, metabolized glucose mainly through the PP pathway. A reexamination of the same mutant by Emmerling et al. with METAFoR analysis by NMR confirmed the observation only in ammonia-limited chemostat cultures.51 In glucoselimited cultures, similar or lower fluxes compared to parent strain were calculated for the PP pathway with concomitant higher fluxes for the TCA cycle, especially in the chemostat with the lower dilution rate. A single mutant pykF strain examined by Al Zaid Siddiquee et al.64 showed the same trend as the double mutant under non-growing conditions, i.e. elevated fluxes thought the PP pathway. In addition, activation of the malic enzyme with a flux about 10% towards the formation of PYR from MAL was calculated. Figure 1-11. The percent of NADPH produced by PP pathway, TCA cycle and PntAB transhydrogenase activity in E. coli during standard aerobic batch growth on glucose. (Adapted from reference 66).

PAGE 51

34 Figure 1-12. Glucose transport systems in E. coli The phosphoenolpyruvate: carbohydrate phosphotranferase system (PTS) and the galactose permaese (GalP) system. (Adapted from reference 94) Another modification around the PEP node was the disruption the of phosphotranferase transport system (PTS) which consumes at least 50% of PEP.94 Starting from a strain lacking PTS (PTS-), which has a limited ability to grow on glucose, a method based on continuous culture was applied to select for mutants (PTS-GLC+) with growth rates similar to wild type strain JM101 (a functional galP gene is required for this phenotype, Figure 1-12). METAFoR analysis of 13C-labeled amino acids by NMR revealed reduced carbon flux through the PP pathway and a 4-10% flux through the malic enzymes for the PTS-GLC+ mutant strain.54 Interestingly, in the PST-GLCmutant,

PAGE 52

35 in which glucose transport capacity is 10 times lower than the wild type JM101 strain, twice higher flux via the PP pathway was observed compared to parent strain. Other experimental methods: metabolic regulation analysis of NADPH related metabolic pathways at gene and protein expression level. Two dimensional electrophoresis (2DE) is a powerful technique for separation of complex protein mixtures and based on their different MWs and pIs. Ultimately, regulation of metabolic fluxes depends on enzyme activity which depends, to some extent, on protein expression levels. Thus, the different protein expression levels under different physiological conditions can be estimated and correlated to regulation of metabolic pathways95 by separating a protein mixture in the cell lysate by 2DE and quantifying each protein spot after electrophoresis. Utilizing this approach in combination with in vitro measurements of enzymes activities in E. coli Peng and Shimizu found that the two proteins involved in the PP pathway and related to NADPH production, zwf and gnd, are down-regulated during microanaerobic growth in glucose medium.56 This result was not a surprise because both enzymes were known to be subject of growth rate regulation, in fact the cells growth was quite slow under microanaerobic conditions. For isocitrate dehydrogenase, a source of NADPH in TCA cycle, the correlation of expression by activity was not linear since it is known that ICDH activity is regulated by reversible inactivation via phosphorylation. 96 Another technique for monitoring metabolic regulation at gene expression level is the reverse transcription polymerase reaction (RT-PCR) analysis. It is currently the preferred method for quantifying mRNA due to its higher sensitivity compared to other RNA analysis techniques such as Northern blot. In addition, it is relatively easy to run: isolated RNA is used as a template in order for a retroviral reverse transcriptase to create a complementary DNA (a cDNA) transcript. This product is subjected to a common PCR, where in the exponential phase, the signal is proportional to the amount

PAGE 53

36 of input RNA template or number of cycles. The measurements of relative changes in mRNA levels are reproducible as long as the PCR volumes are the same from tube to tube and the data are collected before the reaction reached a plateau. Finally, a standard curve is required to correlate the signal with the specific amount of mRNA. Kabir and Shimizu applied this approach, also known as semi-quantitative RT-PCR, to investigate patterns in a pgi knock-out E. coli strain expressing genes for poly(3hydoxybutyrate) (PHB) production (Figure 1-13).57, 58 Figure 1-13. Poly(3-hydroxybutyrate) PHB synthesis: the acetoacetyl-CoA synthase requires NADPH. Sauers group had shown that the deleterious effect on cell growth caused by the NADPH imbalance in the pgi E. coli mutant can be partially compensated by expressing an enzyme consuming NADPH such as SthA. A similar consequence may have the expression of the acetoacetyl-CoA synthase in Shimizus studies, which also requires NADPH. Indeed, expression of phbB in E. coli ( pgi ) increased the true cell mass by about 19%. Further measurements of NADPH concentrations confirmed the higher levels of NADPH in the mutant strain without phbB compared to the PHB producer

PAGE 54

37 strain. The transcript level analysis of 87 central metabolic pathway genes revealed upregulation of most of the genes in PP pathway except gnd The most interesting observation from this proteome analysis study was the absence of ptsG which encodes for a protein that is a component of PTS system, in 2DE gel for both pgi mutants even though the RT-PCR revealed the presence of ptsG transcript in both cases. This result led the authors to propose that the mutation of pgi affects ptsG expression after transcription initiation and an RNAaseE-mediated degradation of ptsG transcript97 might be another reason for lower glucose consumption rates in this mutant. Experimental methods: limitations Undoubtedly, the experimental techniques discussed in the second part of this chapter offer a more holistic picture of the NADPH regeneration sources in E. coli ; however, possible limitations and pitfalls of them must be encountered. Specifically for the 13C metabolic flux analysis: METAFoR analysis fails in the case that two alternative pathways operate according to the same carbon-carbon bond rearrangements. METAFoR analysis of 13C NMR and/or GC-MS data provides direct evidence for a particular flux or reaction in a metabolic network, but only ratios between fluxes around a metabolic node can be calculated. In order for the absolute intracellular fluxes to be estimated, 13C isotopic measurements must be applied, in parallel, with the theoretical approach of FBA, in which the 13C data are acting as additional constraints.51 As a consequence, prerequisites stated above for the FBA to be valid must be followed here as well. The merging of METAFoR analysis and FBA requires heavy mathematical and computational analysis98 which is usually performed by specially designed software with little interaction between the program and the experimentalist.99 Isotope abundances calculated from 13C data should be independent of possible limitations of the technique used for their derivatization such as 13C spin relaxation, strong spin-spin coupling effects, 13C isotope effects on carbon chemical shifts, isotope fractionation of the amino acids during the GC separation and nonlinearity of the MS detector response at different mass isotopomer ratios or different sample concentrations.78 In addition to a known metabolic network, 13C isotope effects on metabolic reactions must be negligible and metabolite channeling (i.e., when an intermediate

PAGE 55

38 metabolite is transferred from one enzyme to the other without previous equilibration with the rest of the molecules in the medium) should be absent.100 The accuracy of 13C data depends on the 13C/12C glucose with higher accuracy at a ratio of 50:50. The relative high cost of labeled sugars usually prohibits this type of experiments. The METAFoR analysis based on NMR data is quite laborious, time-consuming and expensive due to low sensitivity of the technique: about 300-400 ml of sample and 4-5 days are required before the calculation of the first set of data.101 In recent years, the problem was partially bypa ssed with the use of the GC-MS, but interpretation of the MS data is more complicated. The 13C metabolic flux analysis is currently restricted to two physiological conditions: continuous cultures and exponent ially growing cells, because isotopic equilibrium must be established before the first measurement. Last but not least this approach is not applicable in biological systems like resting cells performing biotransformations, due to insufficient de novo synthesis of biomass.87 2DE and RT-PCR, are still in their infancy. A common problem associated with 2DE is the identification of each spot. Amino acid sequencing and enzyme activities must be employed in parallel for the 2DE data to be justified. The RT-PCR is, in principle, a quick and extremely sensitive te chnique, but many parameters concerning the amount of PCR ingredients and the duration of PCR cycles must be verified and controlled carefully in advance. The design and quality of primers is also of paramount importance since genomic DNA contamination in a RNA preparation may be amplified too and produced false positives. Both techniques fail to detect post-translational regulation of enzymes and become extremely co stly if we consider all the additional control experiments that are required for their justification. Conclusion The aim of this chapter was to demonstrate the different tools and methods that researchers used the last six decades in order to answer two key questions regarding the NADPH metabolism in E. coli : Which are all the possible sources of NADPH in E. coli ? Most probably those five depicted in Figure 1-2. To which extent each source

PAGE 56

39 does contribute to the total NADPH pool? Unfortunately, as our discussion and a quick look at the last column of Table 1-1 and Table 1-2 revealed a global answer does not exist. In fact, the answer is another question: under which physiological conditions and based on which genetic background we are seeking for an answer? Different E. coli strains, in different environments, are us ing some or all of the abovementioned NADPH sources to a different degree in order to accommodate their needs. This statement may satisfy a physiologist who is trying to under stand how a biological system works. Definitely, it does not fulfill the ambitions of a metabolic engineer: Metabolic engineering is the improvement of cellular activities by manipulation of enzymatic, transport and regulatory functions of the cell with the use of recombinant DNA technology. [Toward a Science of Metabolic Engineering, James E. Bailey]102 In the next chapter, a couple of examples related to metabolic engineering of NADPH redox metabolism in E. coli will be introduced. A particular interesting case in redox biocatalysis, in which the NAPDH regeneration machinery of resting E. coli cells was used to produce a chiral compound in a gram-scale level will be discussed in detail.

PAGE 57

40 CHAPTER 2 INVESTIGATION OF NAPDH REGENERATION SOURCES IN RESTING ESCHERICHIA COLI CELLS ENGINEERED TO OVERPRODUCE A NAPDHDEPENDENT KETOREDUCTASE So far, the more valuable products, chiral building blocks for the synthesis of pharmaceuticals, flavors and agrochemicals, are potentially obtained with NADPHdependant enzymes. -Andreas Liese, Head of Institute of Technical Biocatalysis at TUHH103 Introduction NAPDH is an extremely high-priced hydride donor for an E. coli cell as well as for an organic chemist when stoichiometric amounts of it are required (Table 2-1 and Figure 2-1). Therefore, a reductive reaction would be economically feasible only upon regeneration of the core molecule at the expense of another cheaper reductive source. For example, the biosynthesis of one gram CDW (cell dry weight) of E.coli cells requires more than 15 mmols of NADPH; however, this amount of biomass barely contains 0.0014 mmols of the reduced cofactor.104 In other words, the same nicotinamide molecule has been recycled more than 10,000 times during this process, usually at the expense of a carbon source such as glucose. Accordingly, a plethora of ingenious NAD(P)H regeneration strategies have been employed by researchers who are explori ng the possibility to use a NAD(P)H-dependent reductase in their asymmetric synthesis. Modi fying slightly the original classification by Chenault and Whitesides,105 we should recognize that among the four general regeneration strategies (Figure 2-1): chemical, photochemical, electrochemical and biological, only the latter has unlimited compatibility with biological systems. In between the enzymatic and microbiological methods, the former has found a great deal of applications in laboratory as well as in industrial scale, especially when the regeneration

PAGE 58

41 of NADH is utilized.106 In contrast, the NADPH regeneration techniques are lagging behind. Table 2-1. Cost of nicotinamide cofactors and their content in E. coli. aBased on Sigma-Aldrich catalog (2006)107 bAdapted from reference 104 cAssuming that the mass of one cell is 3.0x10-13 g (DW).108 Figure 2-1. Explanation of the relative prices among nicotinamide cofactors.109 Step 1: Isolation of NAD+ from bakers yeast (1mmol/0.55kg CDW). Step 2: Reduction or phosphorylation of NAD+ leads to NADH or NADP+ respectively. Step 3: Reduction of NADP+ produces NADPH. Depending on the pH, NAPDH is 3 to 10 times less stable than NADH. Cofactor MW Costa ($/g) Specific cost ($/mmol of delivered [H]) Number of molecules per E. coli cellb Intracellular molarity (mM)b Cofactor content in E. coli ( mol/g CDW)c NADPH Na4 (~95%) 833.35 837 697 250,000 0.22 1.370 NADP+ Na ( 95%) 765.39 211 161 190,000 0.17 1.041 NADH Na2 (~95%) 709.4 70 50 440,000 0.90 2.411 NAD+ Na (~95%) 685.45 30 20 1,000,000 0.40 5.479

PAGE 59

42 Figure 2-2. NAD(P)H regeneration strategies and major disadvantages for each method. (Adapted from reference 106) The aim of this introduction is not to extensively review all the NAP(D)H regenerations methods, since this has been done recently.106, 110 In contrast, by selectively analyzing a few examples from the current literature, we will acknowledge the limitations of the enzymatic strategies in NAPDH regeneration, and we will point out the importance of the experimental work that follows.

PAGE 60

43 Examples of Enzymatic NADPH Regeneration Methods A substrate-coupled NADPH regeneration scheme, like the one depicted in Figure 2-3, is probably the most self-sufficient method in this category. In the presence of excess isopropanol, a recombinant alcohol dehydrogenase from Lactobacillus brevis (LBADH) reduced catalytic amounts of NAPD+ and simultaneously acted as an excellent regioand enantio-selective reductive catalyst on a number of 3,5-dioxocarboxylates.111 Unfortunately, the method is limited by the enzymes substrate specificity. Thus, when the opposite enantiomer was the desired product, the same researchers switched to one of the most well-established,112 enzyme-coupled, regeneration systems using glucose-6phoshate as a sacrificial cosubstrate (Figure 2-4).113 The subsequent hydrolysis of the corresponding lactone renders the reaction irreversible in this approach. Figure 2-3. A substrate-coupled enzymatic regeneration method for NADPH. Due to the relatively high cost of phosphorylated glucose, G6PDH has been replaced the recent years by glucose dehydrogenase (GDH), which accepts glucose as reductant (Figure 2-5).114 Unfortunately, the net benefit from this replacement, in terms of cost, is negligible due to a much higher price of GDH compared to that of G6PDH.

PAGE 61

44 Figure 2-4. Glucose-6-phosphate dehydr ogenase (G6PDH) as an auxiliary enzyme. Figure 2-5. Glucose-6-phosphate dehygrogenase as an NAPDH regeneration catalyst. Another concern about G(6P)DH regeneration systems is that both produce stoichiometric amounts of gluconic acid, which may impede the downstream purification process. To circumvent this issue, two routes have been explored so far: i) the use of a hydrogenase, which regenerates NADPH at the expense of the cheapest reducing agent, dihydrogen103 (Figure 2-6) and ii) the change of an enzymes cofactor diet from NAD+ to NADP+. The latter approach was first attempted for formate dehydrogenase (FDH), which releases CO2 as the only by-product.115 The first NADP+-dependant FDH from Pseudomonas sp. (mut-Pse FDH) is commercially available from Jlich Chiral Solution GmbH at a cost similar to GDH.116 In addition, a phosphite dehydrogenase

PAGE 62

45 (PTDH) has been converted to an NADP+ dependent enzyme with promising characteristics since phosphite is a very low cost reductant and the only by-product in this case is a phosphate ion.117 However, to the best of our knowledge, none of the three abovementioned systems have been applied to a larger, at least 1 mmol, reaction scale and thus we will not consider them further. Figure 2-6. Hydrogenase I from Pyrococcus furiosus for NADPH regeneration. Examples of Microbiological NADPH Regeneration Methods Under the umbrella of microbiological regeneration techniques, we may recognize two slightly different approaches, one in which the permeabilization of the cell membrane is intended and the other in which the E. coli cells are considered as absolutely selfcontained bioreactors. This feature implies that the cells are metabolically active and thus the integrity of their membrane is a prerequisite for a successful biotransformation. Permeabilized E.coli cells In one of the first examples of permeabilized E. coli cells, Kataoka et al.118 grew separately two JM109 E. coli strains, one overexpressing a GDH from Bacillus megaterium and another overproducing an aldehyde reductase from Sporobolomyces salmonicolon (Figure 2-7). The harvested cells were mixed together in different ratios and resuspended in potassium phosphate buffer containing catalytic amounts of NADP+.

PAGE 63

46 The presence of an organic layer of n-butyl-acetate rendered the cell membrane susceptible to the free transport of nicotinamides in between the two different strains. Figure 2-7. An example of two coupled E. coli strains in a two-phase system. A slightly modified approach will be the overproduction of both the auxiliary enzyme and a reductase in the same E. coli cell (Figure 2-8). Ema et al.119 applied this concept coupling the regeneration of NADPH by GDH with the reduction of 2, 4 octanedione by the Gre2 from Saccharomyces cerevisiae Even though a disruption of cell membrane or a pretreatment of the cells with organic solvent was not employed in this case, the authors reported that they froze the cells at -20C before their utilization. Hence, a semi-permeabilization of the cell membrane could be assumed due to this treatment. Interestingly, a 10% conversion was observed even with cells that overexpressed only Gre2. Figure 2-8. An example of an E. coli cell overexpressing both GDH and a reductase.

PAGE 64

47 A more complex scenario is depicted in Figure 2-9. In this case the authors tried to couple the most popular auxiliary enzyme for the regeneration of NADH (i.e. formate dehydrogenase (FDH) from Candida boidinii ) with the recombinant membrane-bound transhydrogenase from E. coli (PNTAB) which in turn shuttled the hydride from NADH to NADP+.120 The NADPH produced was used by an alcohol dehydrogenase (ADH from Lactobacillus kefir ) for the reduction of acetophenone. The whole process was carried out in the same E. coli cell, permeabilized with 1% toluene. Unfortunately, no application on a larger scale (i.e. > 1mmol) is reported for this system. Figure 2-9. An example of an E. coli cell overexpressing tw o auxiliary enzymes and a reductase. Surprisingly, Xu et al.121 reported recently that under optimal conditions, a two strain system similar to the one developed by Kataoka et al., is able to perform the same bioconversion without the addition of an organic solvent and/or extra NADP+. In this example, the overproduction of GDH in the first E. coli strain with the aid of high concentrations of 4-chloro-3-oxobutanoate caused an efflux of NADPH into the medium in the presence of glucose. Subsequently, the reduced nicotinamide was transferred into the second E. coli strain which is responsible for the transformation of the keto-ester to the corresponding chiral alcohol (Figure 2-10). A transmembrane transportation of

PAGE 65

48 NADPH was hypothesized and the percentage conversion found dependent on the initial concentration of the ketone with higher conversions at higher concentrations. It is worth noting that external addition of NADP+ did not improve the bioconversion in the two strain system but was found necessary w hen a one strain system, overexpressing both auxiliary and alcohol dehydrogenase, was utilized. Figure 2-10. An example of two E. coli strains coupled for regeneration of NADPH in one phase system. Intact E.coli whole cells In 2004, Walton and Stewart reported one of the most self-sufficient systems in this category.122 In this approach, the workers relied absolutely on the E. coli metabolism for the NADPH regeneration without the need to overexpress any aux iliary enzyme and/or add extra amounts of NADP(+ /H ) in the aqueous medium (Figure 2-11). Additionally, this example is one of the rare cases wher e the amount of sacrificial substrate (i.e. glucose) added is not in excess, as compared to the amount of product made. The overproduced reductase was Gre2 from bakers yeast.

PAGE 66

49 Figure 2-11. An example of intact E. coli cells capable to regenerate NADPH for ethyl acetoacetate reduction by a ketoreductase. A more sophisticated E. coli strain (SC16529) was developed by Hanson and coworkers123 to meet the needs for the production of a chiral chloro-alcohol (Figure 2-12) on a large scale. In this case, G6PDH from Saccharomyces cerevisiae was overexpressed with a ketoreductase from Hansenula polymorpha in the same E. coli cell. The authors relied on the E. coli s PTS activity for the phosphorylation of glucose. However, the addition of catalytic amounts of NADP+ was found necessary for the completion of this biotransformation. Figure 2-12. An example of intact E. coli cells overexpressing a G6PDH and a reductase.

PAGE 67

50 Economical Analysis How the different NADPH regeneration approaches could be evaluated from an economical perspective? A very informative parameter regarding this issue is the total turnover number (TTN), which is defined as the molar amount of product made by the end of the reaction per molar amount of cofactor used.124 This number embraces several factors such as the relative stability of the cofactor under the reaction conditions, the reaction time-frame and the catalysts turnover number.125 It is apparent from the Figure 2-13 that the higher the TTN, the more cost-efficient the bioprocess is. As proof, we present a cost analysis of the aforementioned examples and a tabulation of the corresponding TTN for each of them in Table 2-2. An estimation of intracellular NADP(+/H) content has to be made for those cases that whole cells were used. For this calculation we assumed 440,000 molecules of NADP(+/H) per E. coli cell. Figure 2-13. Cost of NADP+ as a function of TTN.

PAGE 68

51Table 2-2. Cost analysis of different NADPH regeneration methods and tabulation of TTNs. Year (ref.) Cofactor used (amount, cost) Sacrificial substrate (amount, cost) Auxiliary enzyme (amount, cost) Biomass in grams (cost)a Theoretical Yield in mmols (Total cost) Productivity (mmol h-1 g-1 CDW) Specific Cost ($/mol) TTN 1998 (118) NADP+ (0.019 mmol, $3.1) glucose (166 mmol, ngl) GDHb (261U, ngl) 3 ($4.5) 83.6c ($7.6) 4.2 100 3,980d,o 2000 (111) NADP+ (0.033 mmol, $5.3) acetone (66 mmol, ngl) recLBADH (600U, $39)m --9.9e ($44.3) --4,400 300f 2004 (113) NADP+ (0.17 mmol, ($27.3) G6Pg (4.4 mmol, $140) G6PDHg (2mg, $100) --2.8e ($267) --95,000 16.5f 2004 (122) --glucose ( 216 mmol, ngl) --17.5 ($26) 348j ($26) 7.0 70 27,840f,n 2005 (114) NADPH ( 0.05 mmol, $35) glucose (10 mmol, ngl) GDHh (50mg, $395) --4i ($410) --103,000 80f 2005 (123) NADP+ (0.15 mmol, $24) glucose ( 167 mmol, ngl) G6PDH (na, ngl) 75 ($112.5) 72.4l ($136.5) 0.13 1,890 337d,o 2006 (119) NADP+ (0.012 mmol, $ 1.9) glucose (8mmol, ngl) GDH (na, ngl) 2 ($3) 4.0k($ 4.9) 0.7 1,200 296d,o 2006 (121) --glucose (10mmol, ngl) GDH (27U, ngl) 0.22 ($0.2)5.7c($ 0.2) 7.8 30 35,625d,n a Assuming $1.5/g (wet weight). b In-house prepared. c Product recovery was not attempted. d Optimized conditions. e 77% recovered product. f not optimized conditions. g Based on Sigma-Aldrich catalog (2006).107 h Based on Biocatalytics Inc. catalog (2006).106 I 90% recovered product. j 76% purified product. k 71% recovered product. l 89% recovered product.m Based on J lich Chiral Solution GmbH catalog (2006).116 n For this calculation, the sum of NADP+ and NADPH pool is considered based on the values given in Table 1 and assuming that the mass of one cell is 1.0x10-12 g (wet weight).108 o Based on the NADP+ added and NADP(+/H) content of biomass used. Abbreviations: ngl, negligible; na, not applicable.

PAGE 69

52 The major conclusions from this analysis are: Regeneration methods based on pure auxilia ry enzymes are too expensive even on a gram-scale reaction. Only in-house prepared crude extracts of overexpressed proteins might accommodate the problem.127 A substrate-coupled regeneration method may be affordable if a TTN of 300 or higher is achieved and a high-value product is produced. The utilization of whole cells, intact or permeabilized, is the only foreseeable avenue for the large-scale production of fine chemicals, in which NADPH is mediated. This is particularly true for low-value chemicals. The misconception that an E. coli cell does not contain an efficient NADPHregeneration system is a myth,128 at least for NADPH regeneration rates in the order of 120 U/g (CDW). At the moment, several different methods that utilize E. coli whole cells overproducing oxidoreductases are under development. Therefore, caution must be paid when we are evaluating data such as the one presented in the first and the last row of Table 2-2 since isolated yields were not reported and the reaction scale in the last case was quite low (< 10 mmol). We should also be aware of a possible failure to repeat successful stories as the one presented in the forth row of Table 2-2 with substrates that may interfere with the intracellular metabolism of E. coli or they could not access the cells membrane. We also should not mislead the reader by making her or him believe that the TTN is the only effector on the final cost of a large-scale bioprocess. Useful reviews129-131 and discussions132, 133 exist in the literature about other factors that should be evaluated before a biocatalytic process could be considered successful, but they are beyond the scope of this introduction. Manipulations of NADP(H /+) Levels in E. coli towards the Improvement of Biotechnological Processes A number of researchers have studied the manipulation of the redox-metabolism in various microorganisms hoping to improve the biosynthesis of metabolites of industrial importance.134-137 However, similar efforts in E. coli are rare138, 139 In one of them, Flores et al.140 tried to determine if the overexpression of zwf in an E. coli JM101 strain could

PAGE 70

53 overcome the metabolic burden imposed on it by the presence of a high copy plasmid. Often, plasmid-DNA replication and synthesis of plasmid-encoded proteins reduce the growth rate of the producing culture. By increasing the expression of zwf and hence increasing the carbon flux through the PP pat hway, the authors were able to partially recover the growth capabilities of an E. coli strain harboring a multicopy plasmid responsible for the production of a recombinant TrpLE-proinsulin peptide. The growthrate values were three times higher upon overexpression of zwf, but still twice lower than the rates for the wild type. When the degree of zwf gene induction was altered from 3 to 100 times in terms of higher levels of zwf expression, it was found that an overexpression as little as three fold was adequate for the growth-rate of the strain overproducing the TrpLE-proinsulin to reach a plateau. Unfortunately, it is not clear from this study if NADPH or another metabolite from the oxidative part of PP pathway, which might be produced in excess, was responsible for the observed results. The second example is related to biosynthesis of the biodegradable polymer PHB from acetyl-CoA (Figure 1-13). The process requires one equivalent of NAPDH from each unit attached to the polymer. Lim et al.141 constructed plasmids containing zwf or gnd co-integrated with the phb operon and introduced them into E. coli The amount of PHB increased after expressing the genes, especially zwf However, this increment may be due to altered NADPH/NADP+ (i.e. six times higher compared to wild type), or redirection of the flux through the AcCoA, or the increased activity (i.e. 1.16-1.30 fold) of all the enzymes involved in PHB synthesis upon zwf overexpression. Again, which one was the crucial parameter was not clear from this study. In a very recent study Snchez et al.142 followed a different approach to tackle the same problem. They overexpressed the so luble pyridine transhydrogenase, UdhA, in order to increase the yield and productivity of PHB in E. coli The outcome from this manipulation is crystal clear: almost double yield of PHB obtained in the strain

PAGE 71

54 overproducing UdhA (Table 2-3). However, their discussion about the reason behind this outcome is controversial. The authors stated that by inducing a high level of UdhA, they increased the NADPH availability and thus the yield and productivity of PHB. They also commented that the drop of NADPH/NADP+ with the concomitant increment of NADH/NAD+ in the UdhA overexpressing strain suggests a depletion of NADPH pool and a higher NADH conversion rate matched with the PHB formation demands. Recalling, though, our discussion in the first chapter about the physiological role of UdhA (i.e., transfer of hydrides from NADPH to NAD+) we can also arrive at the same observations about the nicotinamide ratios in this example (Table 2-3). Nevertheless, this is the first example in the literature in which a NADPH manipulation in E. coli caused increased productivity of a NADPH-dependent compound. Also, it is worth noting the high TTN that this system reached. Analogous studies overexpressing the membranebound transhydrogenase, PNTAB, for the overproduction of xylitol in E. coli have not yet been fruitful.143 Table 2-3. Cofactor levels of PHB producing strains and tabulation of TTN. Strain Total yield (mmol g-1 DCW) Productivity (mmol h-1 g-1 DCW) NADPH (NADH) ( mol/gDWC) NAPD+ (NAD+) ( mol/gDWC) NADPH /NADP+ (NADH/NAD+) TTNC Controla 11.4 0.44 0.375 (0.6) 0.165 (2.6) 2.27 (0.23) 21,111 GJT001 22.8 0.88 0.250 (0.9) 0.185 (2.0) 1.35 (0.45) 52,413 a Expressing only phb operon. b Expressing both phb and udhA c Based on the calculated NADP(+/H) content. Description of the System under Study In all three aforementioned examples, the merits of the NADPH perturbations to the obtained results are not clear due to the fact that the provided carbon source contributes to bioprocess not only by pr oviding reducing equivalents but also by supplying carbon atoms to the final product with bits of carbon atoms. In contrast, a single-step biotransformation by a NADPH-dependant reductase like the one presented in Figure 2-11 has as a net gain only a pair of electrons and a proton, in the form of

PAGE 72

55 hydride. Thus, it exclusively depends on a metabolically active NADPH regeneration system, without causing a drain to other intracellular metabolites. In a preliminary attempt to shed light on the NADPH regeneration system within a resting E. coli cell during the course of a redox enzymatic reaction Walton and Stewart made the following observations: The profile of glucose consumption by the resting E. coli cells was the same in the presence and in the absence of the biotransformed starting material, in otherwise identical conditions. Both [NADH]/[NAD+] and [NADPH]/[NADP+] ratios remained constant at about 0.4 each, with an intracellular concentration for NADPH about 180 M. The molar ratio between product formation and glucose consumption was about 2.3. Consequently, they concluded that the relatively low level of [NADPH]/[NADP+] ratio, which is normally in the range of 0.7-1.3, may be due to additional needs of the overexpressed NADPH-dependent enzyme. In addition, they made an effort to identify the NADPH sources in this system by correlating the stoichiometry linking the product formation with glucose consumption in their system (i.e. 2.3 molar ratio) with the stoichiometry linking the product formation with glycerol consumption in a similar system developed by Woodley and coworkers (i.e. 0.96 molar ratio).144 Recognizing the PP pathway and the TCA cycle as the only NADPH sources in E. coli and setting the efficiency of the latter at 96% (each mole of glycerol can produce maximum 1 mole of NADPH), they suggested that the icd was responsible for around 83% [0.96 x 2.0)/2.3] of the NAPDH consumed by the ketone reduction. Therefore, less than 10% of glucose might be metabolized through the PP pathway in that system. The weak points of the abovementioned discussion are the following: The [NADPH]/[NADP+] ratio was 0.4 even at time, t=0 h, which is the point just before the beginning of the bioconversion. Therefore, this ratio simply may reflect the ratio of the nicotinamide levels at the end of the growing phase of cells and after their overnight storage at 4C.

PAGE 73

56 Even though there is a significant scattering in the data related to phosphorylated nicotinamide levels, it is obvious the trend of faster depletion of NADPH compared to NADP+ but this was not recognized by the authors. The observed drop in all nicotinamides levels could be easily explained in a system such as non-growing cells, in which de novo synthesis of those metabolites is excluded due to lack of a nitrogen source. Also, the integrity of the cell membrane is under question especially during the last 12 h of the bioprocess when high concentrations of organic material have been accumulated. Nevertheless, the estimated TTN for this system, based on the NADP(H/+) content provided by authors for it, is as high as 31,650 (assuming the volume of an E. coli cell = 1 x 10-15L).108 The data provided by Woodley et al. by no means could be correlated with the data obtained from Stewarts study because: i) the two E. coli strains were not isogenic (i.e. Woodleys host strain is isogenic to E. coli K-12 strain, Stewarts host strain is isogenic to E. coli B strain); ii) the calculated efficiency for the Krebs cycle is based on completely different carbon sources. Glycerol transport in E. coli is a facilitated diffusion process (i.e. requires no energy expenditure)68 whereas glucose transport process requires a high energy phosphate group from PEP; iii) when Woodley and coworkers examined their system further found that the amount of product formed within one hour, when glucose was used as a carbon source was half compared to what they obtained when glycerol was added to the medium.145 Unfortunately, the amount of consumed glucose was not reported. In addition, specific activity measurements with crude extracts revealed that the rate limiting step in Woodleys bioprocess is the transport of starting ketone through the cell membrane. To this end, the specific aims of the current work were threefold: To develop a method that will allow us to identify the NADPH regeneration sources in resting E. coli whole cells overexpressing a ke toreductase, a model system for bioreductions in lab-scale. To perturb the NADP(H/+) levels of the host cell by environmental and genetic manipulation means in order to test the capabilities of the methodology invented in the first step (Figure 2-14). To evaluate the effects of those perturbations on the productivity of the system under study. Experimental Strategy In lab-scale biotransformations, resting cells are usually the preferred method compared to batch and continuous culture methods. Non-growing cells are live cells that retain most of the enzymes activities of growing cells and are obtained from the culture medium at the time in which the activity of the overexpressed enzyme is at satisfied levels. Thus, the bioconversion is divided into two parts. In the first phase, cells are

PAGE 74

57 grown in rich medium (e.g. LB medium) while overproducing the desired biocatalyst. In the second phase, the concentrated cells are resuspended in a modified culture medium lacking nutrients for growth (e.g. M9 medium without NH4Cl) and the bioconversion takes place. This approach ensures cleaner reactions and easier isolation of products. In addition, cells from the first phase could be stored and used on demand at the desired amounts and under controlled conditions in the second phase. However, in such a system, the popular method of METAFoR analysis of 13C-labeled amino acids by NMR or GC-MS cannot be applied for the calculation of NADPH-related fluxes. An alternative choice would be a radiorespirometric method with all the disadvantages that we discussed in the first chapter. Figure 2-14. Cartoon represents the system under study. The perturbed points represented with X. Instead, we thought that a modified approach similar to one employed by Csonka and Fraenkel, which was described in detail in the first chapter, would be more

PAGE 75

58 informative. The major difference between the two systems is that in their system the labeled at a specific position of glucose atom (i.e. tritium) was transferred to NADP+ and then to certain amino acids, whereas in our case the labeled atom is transferred to ethylacetoacetate by the overexpressed NADPH-dependant keto-reductase, Gre2. An overview of the experimental strategy is depicted in Figure 2-15. The tritium was replaced with the stable isotope deuterium and thus GC-MS can be used to directly measure the deuterium content in the ethyl-3-hydroxybutyrate. The major advantage of this method is that is sensitive and not destructive. Thus, a small volume of sample is required and numerous of samples can be analyzed during the course of the same biotransformation. The labeled sugars that could possibly give useful information about the NADPH sources in this system are listed in Table 2-4. In this study only the glucose-1-d, -3-d, and -6,6 d2 were used. A detailed description about the fate of each of these isotopomers in the central carbon catabolism is given in Chapter 1 (see Figure 1-1,1-2,13 and1-4). In summary, glucose-1d and -3d could trace the NADPH sources in PP pathway and they could prove or eliminat e the existence of the ED pathway. The glucose -6,6 d2 could trace the activity of icd in the TCA cycle and perhaps the activity of meaB. The glucose-4d should give labeled NADH, which in turn should label NADP+ if a transhydrogenase activity is hypothesized. In principle, if glucose labeled in each position was used, almost all the produced NADPH should be labeled. Table 2-4. Labeled sugars. Cost and possible application. D-Glucose Cost ($/g)a Application -1-D 180 Zwf activity -3-D 570 Gnd activity -4-D 1,700 transhydrogenases activity -6,6 D2 75 icd activity -(1,2,3,4,5,6-D7) 570 Total activity (Control) a Based on Medical Isotopes Inc. catalog.146

PAGE 76

59 Figure 2-15. Illustration of the experimental procedure. Abbreviations: P, product. SM, starting material; IS, internal standard.

PAGE 77

60 The original bioreduction described by Walton and Stewart was carried out in a 3 L-scale containing 4 g/L glucose. Considering the high cost of the labeled sugars (Table 2-4), all biotransformations in this study were carried out in a 10 ml or 100 ml-scale under two distinct physiological conditions: i) under argon atmosphere (i.e. anaerobically) and ii) under air atmosphere (i.e. aerobically). In this context, the conditions under air atmosphere are named aerobic, but they may also be called microanaerobic due to relatively high density of cell cultures (O.D.600~8.0) used.147, 148 In addition to environmental perturbations, we also genetically perturbed the system at four positions. The first two were around the G6P node, disrupting either the pgi or the zwf gene. We expected that in the former mutant, the flux through the PP pathway would be maximized whereas in the latter case it would be diminished. The third one, an icd mutant, was at the isocitrate node, hence excluding the possibility of NADPH production from TCA cycle. The last disruption was the pntAB gene, which encodes the membrane-bound transhydrogenase responsible in the wild type strain for the production of NADPH from NADH. All mutant strains are isogenic to original strain BL21(DE3) used by Walton and Stewart for the ethyl-acetoacetate reduction. This host is necessary for the overproduction of Gre2, for which the expression is under the control of T7 promoter. We should make clear at this point that the measurements with labeled sugars and the different BL21(DE3) strains performed in this study allowed us to calculate the contribution of each NADPH source to the total NADPH pool present in the cell, in each strain. By no means is this method capable to measure net fluxes in the metabolic network. Finally, we examined the capabilities of the mutant strains toward the production of ethyl-3-hydroxybutyrate compared to that for the wild type in a shake flask.

PAGE 78

61 Experimental Procedures Materials D-Glucose-1-d (97% enrichment), D-Glucose-6,6-d2 (98% enrichment) 2-propanold8 and sodium deuteride (98% enrichment) obtained from Aldrich. D-Glucose-3-d (98% enrichment) was obtained from Medical Isotopes Inc. Glucose-6-phosphate dehydrogenase (from Bakers yeast Type XI) and alcohol dehydrogenase (from Thermoanaerobium brockii ) were purchased from Sigma. The purified GST-Gre2 fusion reductase from S. cerevisiae was prepared by Dr I. Kaluzna in our lab.149 Oligonucleotides were purchased from IDT (Coralville, IA) as deprotected and desalted. Long primers were IE-HPLC purified. Taq and Vent polymerases were obtained from New England Biolabs. In-house prepared Pfu was used instead of Vent in some cases. Trinder reagent, for glucose assay was purchased from Chemical Diagnostics. Bacterial Strains and Plasmids Bacterial strains and plasmids were obtained from the Coli Genetic Stock Center (CGSC) (Appendix B). The icd knockout strain (BL21(DE3) icd ) designated as MA1935 was kindly provided from Dr M. Aoshima.150 The pntAB knockout strain, (BL21(DE3) pntAB ) was made by Sylvie Boualavong in our lab. Plasmids were purified by density ultracentrifugation with CsCl in the presence of ethidium bromide. Genomic DNA from E. coli BL21(DE3) strain was purified as describe elsewhere.151 E. coli strain overexpressing S. cerevisiae Gre2p [BL21(DE3)(pAA3)] has been described elsewhere.152 All mutant strains were transformed with pAA3 via electroporation. Cell Cultivations LB medium used for cell growth contained Bacto-Tryptone (10 g/L), Bacto-Yeast Extract (5 g/L) and sodium chloride (10 g/L). Agar (15 g/L) was used to solidify media for plates. M9 nitrogen-free medium used for bioconversions under non-growing

PAGE 79

62 conditions contained disodium phosphate (12. 8 g/L), monopotassium phosphate (3 g/L), sodium chloride (0.5 g/L), magnesium sulfate (2 mM) calcium chloride (0.10 mM) and glucose (0.4%). The last three ingredients were added as sterile aqueous solutions (1 M, 0.10 M and 20%, respectively) after other materials had been autoclaved. M9 minimal medium used for phenotypic characterization of mutants was supplemented with ammonium choride (1 g/L) and glutamate (0.5 mM) in case of MA1935 strain. SOB and SOC media was prepared as described by Sambrook et al.153 Analytical Methods GC-MS analyses were conducted with a Hewlett-Packard Model 5890 GC/MS fitted with a 30 m DB-17 column using helium as carrier gas and a temperature gradient from 60 C to 180 C at 10 C/min with initial and final times 2 and 5 min, respectively. The MS was operating at electron impact ioniza tion mode. Biotransformation reaction mixtures were sampled for GC/MS analysis by mixing 50 l of the aqueous reaction mixture with 100 l of ethyl acetate containing 1.0 mM methyl benzoate (internal standard). After vortex mixing, the organic layer was separated and the aqueous layer was extracted a second time in the same manner. A sample of the combined organic layers was analyzed by GC/MS. A standard curve for ethyl-( S )-3-hydroxybutyrate linear up to 70 mM was constructed by preparing aqueous solutions containing varying concentrations of single analyte which was extracted and analyzed by GC/MS as described above. Glucose concentrations were determined using the Trinder reagent kit. Five l aliquots of water (blank), 5 mM glucose (standard) and the unknown sample, after centrifugation, were added to three separate 1 ml aliquots of reconstituted Trinder reagent. After gentle inversion, the tubes containing each were incubated at 37 C for five minutes. Immediately following incubation the absorbance (A) at 505 nm was

PAGE 80

63 measured. Unknown concentrations were determined based on the equation: [glucose] = 5 mmol/L x (Asample/Astandard) In vitro Construction of a Linear Targeting Fragment by PCR: Synthesis of Targeting Marker with Long Flanking Homologous Regions (LFH) The synthesis of the FRT-kan-FRT casse tte with long flanking homologous (LFH) regions of E. coli BL21 (DE3) DNA was carried out as described by Derbise et al.154 The procedure is based on the overlap extension PCR method and illustrated in Appendix A. In the first step two sets of primers were used to amplify the ~500-bp regions flanking the target gene of the E.coli BL21(DE3) chromosomal DNA. The chimeric primers used in each pair contained a 5 end homologous to one of the ends of the FRT-kan-FRT marker and a 3end that primed to the upstream or downstream region of the desired knockout fragment (step A). The sequence of primers was based on the known genomic DNA sequence of the E. coli K-12 MG1655 strain.155 The plasmid pKD13 was used as a template to amplify the kanamycin cassette with the primers set pKD13 F/R (step B). In the second step ~200 ng of upstream and downstream PCR megaprimers were mixed with ~1000 ng of the FRT-kan-FRT fragment, and with a second set of short primers annealing at the two ends of the targeting cassette. PCR cycling conditions were: 25-30 cycles of 94C for 1 min, 54C for 2 min and 72C for 3 min followed by 1 cycle of 72C for 10 min. In the first and second step Taq and Vent polymerases were mixed in ratio 3 to 1. The obtained products were concentrated by ethanol precipitation and purified by low melt agarose gel electrophoresis. The DpnI treatment of the PCR product from Step A before the gel-purification facilitates the degradation of the methylated plasmid template, which otherwise generates the selective genotype after the transformation with the targeting marker. The purified products were suspended in TE buffer (10 mM TrisHCl, pH 8.0, 1 mM EDTA, pH 8.0) and quantified by agarose gel electrophoresis.

PAGE 81

64 Gene Replacement E. coli BL21(DE3) cells were transformed, by electroporation, with the temperature sensitive replicon plasmid pKD46 containing the red recombinase genes under the control of the inducible arabinose promoter. BL21 (DE3) [pKD46] cells were grown at 30C overnight to saturation in LB medium with ampicillin (100 g/ml), diluted 100-fold in 50ml SOB (plus 100ug/ml ampicillin) and grown to OD600 of 0.1. L-Arabinose was added to 1 mM final concentration to induce the expression of the recombinase genes, and cells were grown to an OD600 ~ 0.5, before chilling into an ice slurry for 20 min. Cells were then centrifuged at 7000 rpm for 10min at 4C. The pellet was resuspended in icecold deionized water and centrifuged. Then the pellet was resuspended in ice-cold 10% glycerol and centrifuged again. The washing step with glycerol 10% was repeated twice, and the cell pellet was suspended in a final volume of 300-500 l ice-cold 10% glycerol. Electroporation was performed using ice-cold cuvettes and a Bio-Rad Gene Pulser set at 2.5kV, 25 uF with pulse controller of 200 Ohms. Transformations were usually performed: 40 l of induced cells, 40 l and 50 l of induced cells mixed with 1 l (30~70ng) or 2 l (60~140 ng) of linear marker, respectively. SOC medium (0.6 ml) was added after electroporation. The cells were transferred to 15 ml test tubes and incubated at 37C for 1-2h with shaking. The transformants were spread on LB plates containing 20-40 g/ml kanamycin. The plates were incubated at 37C to select for recombinants. Detection and Analysis of Recombinants Bacterial colony PCR was used to detect E. coli transformants having an altered structure caused by replacement of a gene. Each colony was touched with a sterile yellow tip and the cells were resuspended in 4-8 l of sterile water. The PCR reaction mixture for 2 l of the above suspension in 100 l of total volume was 1 l of Taq, 10 l 10x Pfu buffer (200 mM Tris-HCl pH 9.0, 0.2 mM MgSO4, 10 mg/ml BSA, 1 mM KCl, 1

PAGE 82

65 mM (NH4)2SO4, 1% Triton X-100) 0.2 mM dNTPs and 200 ng of each primer. Thermocycler parameters were the same as described before. Primers used for screening are listed in Appendix C. Products were analyzed on an agarose gel. Deletion of the Inserted Marker The E. coli BL21(DE3) clones resistant to both kanamycin, kanR, and ampicillin, ampR, were cured of the pKD46 plasmid by growth at 37C, a temperature not permissive for this replicon, and spread on LB, on LB plus amp (100 g/mL) and on LB plus kan (40 g/mL) in different dilutions to select for ampS colonies. KanR mutants were transformed with pCP20 that has a temperature-sensitive replication and thermal induction of FLP recombinase synthesis. Since the kan gene is flanked by the short direct repeats called Flp Recombinase Targets (FRT sites) the Flpase promotes recombination at those specific sites excising the kan cassette and leaving behind a short nucleotide sequence with one FRT site. Ampicillin-resistant transformants were selected after 1-day incubation at 30C, and were screened for kan deletion with colonyPCR as described above. Colonies having the deletion were grown non-selectively at 37C and then tested for loss of all antibiotic resistances. Preparation of [4S-2H] NADPH The preparation of ( S )NADPD was based on a method described by Viola et al.156 31mg NADP+ and 1.5 fold excess of D-glucose-1-d were dissolved in 4.3 ml phosphate buffer, 83mM, pH 8, containing DMSO at 40% (v/v) final concentration. 50 U of glucose6-phosphate dehydrogenase were added and the reaction mixture was stirred gently at RT. The progress of the reaction was followed by taking aliquots every hour and measuring the increment of the absorbance at 340nm. Maximum absorbance was achieved after 5h, conversion 95%. The reduction of NADP+ was almost complete because of the hydrolysis of gluconolactone to gluconate.

PAGE 83

66 The reduced nicotinamide was purified by a method described by Orr et al.157 with minor modifications. The G6PDH was separated from the solution by ultrafiltration (10.000 molecular weight cutoff filter). The NADPDB sample (5 x 1ml) was loaded on a FPLC Mono Q HR5/5 anion exchange column and was eluted by a 0-1 M ammonium bicarbonate gradient.158 Fractions showing a A260/A340 ratio between 2.3-2.7 were pooled (total volume: 15 ml) and lyophilized overnight affording 20 mg of white powdery solid containing 12.5 mg NADPD (based on A340). Final yield: 43%. Preparation of [4R-2H] NADPH The preparation of ( R )NADPD was based on a method described by Gready et al.159 33 mg NADP+,1.2 ml isopropanol-d8 and 4U of TBADH were added in 15ml of 25mM Tris buffer, pH 9 and the reaction mixture was stirred gently at 43 C. The thermostability of TBADH allowed the reaction to be carried out at this temperature, which also prevented the reverse reaction by evaporating the volatile product, acetone. The progress of the reaction was followed as before and the maximum conversion, 75%, was achieved after 3h. The reduced nicoti namide was purified as described above. Final yield: 39%. Preparation of D-glucose-1d A 5 ml aqueous solution of sodium borodeuteride (222 mg or 5.3 mmol) was added dropwise to a 10 ml aqueous solution of -gluconolactone ( 1.3 g or 7.3 mmol) at -4C-0 C at such a rate that all the reducing agent was added in 20 min.160 The pH of the reaction mixture was kept at 3-4 (for every ~10 drops of NABD4 solution, 4-5 drops of 1 M of H2SO4 solution were added). 10 min after the last addition of the reducing agent the reaction was quenched with 0.5 ml of 1 M H2SO4 and the solution was neutralized with 2 M NaOH. 50 ml of methanol was added and solvent was evaporated off under reduced pressure as methyl borate.161 The procedure was repeated four times until no boron was present in the solution based on the carmine reagent.162 The obtained syrup

PAGE 84

67 was lyophilized overnight, redissolved in 2 ml D2O and lyophilized again for 20 h affording 1.9 g of white powdery solid containing 1.3 g of glucose based on Trinder reagent assay. The labeled glucose was used without further purification. Reduction of Ethyl-acetoacetate by Gre2p GST-fusion Reductase Using NADPDB or NADPDA The stereospecificity of hydrogen transfer of NADPH catalyzed by Gre2p was determined by reacting 7mg (~1.6 mol) of NADPDB, with 3.2 mol ethyl-acetoacetate and ~450 g enzyme in total volume of 2.5 ml 100mM KPi pH 7. The reaction was monitored by following the decrement in A340. The oxidized co-factor was purified by FPLC as described before, lyophilized, resuspended in D2O and analyzed by 1H-NMR. The procedure was repeated for the NADPDA. The obtained ethyl-3-hydroxybutyrate was also analyzed by GC-MS. Reduction of Ethyl-acetoacetate by Gre2p GST -fusion Reductase Using a mixture of NADPH and NADPDB In a solution of 704 l 100mM KPi pH 7.0 were added 83 l of a stock solution of NADPDB (24.8 mM), 83 l of a stock solution of NADPH (21 mM) and 80 l of a stock solution of ethyl-acetoacetate (80 mM). The reaction was initiated by adding 50 l of a stock solution of Gre2 (~2.4 g/ml) and was monitored by following the absorbance at 340 nm and by extracting 50 l aliquots for GC-MS analysis. Standard Procedure for Ethyl-acetoacetate Biotransformations with Resting cells A single colony of the appropriate stain harboring the pAA3 was used to inoculate 10 ml of LB media supplemented with 40 g kanamycin and the culture was shaken overnight at 37C. Five ml of this solution was added to a 2 L baffled flask containing 500 ml of LB media, 4 g/L glucose, 40 g/ml kanamycin. The culture was incubated at 37C with shaking. Upon reaching an OD600~0.7, IPTG was added to a final concentration of 0.1 mM and the incubation was continued at 30 C for 6 to 8h. The

PAGE 85

68 harvested cells were resuspended in 5 ml nitrogen free medium and kept overnight or until used (maximum 2 days later) at 4 C. Bioconversions of ethyl-acetoacetate were carried out either in a 1 L flask or in 300 ml vessel or in 25 ml vessel containing 100 or 10 ml M9 medium without NH4Cl and cells prepared as described above at different optical densities. At the beginning, glucose and the organic compound were added to a concentration of 20mM and then appropriate amounts of both were added whenever was nec essary based on glucose consumption and product formation measurements The reacti on was incubated at 32C with shaking, when in flask, or by stirring, when in vessel. The vessels were equipped with a magnetic bar and an ace gas dispersion tube connected to an air or argon line. Most experiments regarding the icd and pntAB strains were performed by Sylvie Boualavong. Results Characterization of Mutated Strains All knockout mutants (Table 2-5) but icd were constructed by a marker-free deletion method163 of genes from start to stop codon as described in experimental section. Table 2-5. E. coli strains used in this study. Strain Pertinent genetic markers WT [BL21(DE3) (pAA3)] wild-type strain overexpressing Gre2 reductase pgi (DJB254) BL21(DE3) pgi (pAA3) (phosphoglucose isomerase deficient) zwf (DJB370) BL21(DE3) zwf (pAA3) (glucose-6-phosphate dehydrogenase deficient) icd (MA1935) BL21(DE3) icd (pAA3) (isocitrate dehydrogenase deficient) pntAB (SB2) BL21(DE3) pntAB (pAA3) (membrane-bound transhydrogenase deficient) Generally any number from one to 50 kanam ycin resistant colonies was obtained after the transformation of parent strain with the FRT-kan-FRT cassette. The recombination efficiency was from zero to 100 % based on about 13 colonies screened.

PAGE 86

69 In the former case the transformation was repeated with same or fresh-prepared linear DNA. Correct marker integration and excision was ve rified in all mutant stains by colony PCR. The PCR primers used were annealing either at the 5 and 3 of the target gene or 500 and 550 bp outside of targets gene coding region. Wild-type starting strain BL21(DE3) was used as a comparison. Results are summarized in Table 2-6, Table 2-7 and Table 2-8 and are depicted in Figure 2-16, Figure 2-17 and Figure 2-18. Table 2-6. Colony PCR for pgi knock-out. E. coli strain Primers Set Expected Size (kbp) Observed Size (kbp) BL21(DE3) pgi :: FRTkan -FRT pgiF, pgiR no band ~1.65, ~0.9, ~ 0.8 (major) BL21(DE3) pgi :: FRTkan -FRT Up500pgi, Down500pgi, 2.33 ~2.3 BL21(DE3) pgi :: FRTkan -FRT Up500pgi, k2 1.21 ~1.2 BL21(DE3) (pKD46) Up500pgi, k2 no band no band BL21(DE3) pgi :: FRTkan -FRT Down500pgi, k1 1.0 ~1.0 BL21(DE3) (pKD46) Down500pgi, k1 no band no band BL21(DE3) (pKD46) pgiF, pgiR 1.65 1.65 (major), 0.8 BL21(DE3) (pKD46) Up500pgi, Down500pgi 2.65 ~2.6 BL21(DE3) pgi :: FRTkan -FRT Up550pgi, Down550pgi, 2.43 ~2.4 BL21(DE3) (pKD46) Up550pgi, Down550pgi 2.75 ~2.7 BL21(DE3) pgi :: FRT pgiF, pgiR no band ~1.65, ~0.9, ~ 0.8 (major) BL21(DE3) pgi :: FRT Up500pgi, Down500pgi 1.1 ~1.0

PAGE 87

70 Figure 2-16. Results from colony PCR reactions probing the pgi locus before and after the disruption. Abbreviations: WT (wild type), MT (mutant strain); kanR, kanamycin resistant; kanS, kanamycin sensitive; M1and M2, DNA markers.

PAGE 88

71 In case of pgi deletion, a weak band with size similar for the structural gene was detected in kanamycin resistant colony along with multiple other weak bands when the pair of pgiF/R primers was used. However, we suspected that these low levels PCR products are due to aerosol contamination or mis-priming since colony PCR with outside primers confirmed the right construct. Table 2-7. Colony PCR for zwf knock-out E. coli strain Primers Set Expected Size (kbp) Observed Size (kbp) BL21(DE3) zwf :: FRT-kan-FRT zwfF, zwfR no band ~1.3, ~1.2, ~0.9 BL21(DE3) zwf :: FRT-kan-FRT Up550zwf, Down550 zwf 2.43 ~2.4 BL21(DE3) zwf :: FRT-kan-FRT Up570zwf, Down550zwf 2.45 ~2.4 BL21(DE3) (pKD46) Up570zwf, Down550zwf 2.62 ~2.5 BL21(DE3) zwf :: FRT-kan-FRT Down570zwf, Up550zwf 2.45 ~2.5 BL21(DE3) (pKD46) Down570zwf, Up550zwf 2.62 2.55 BL21(DE3) (pKD46) Up550zwf, k2 no band ~0.8 BL21(DE3) zwf :: FRT-kan-FRT Up550zwf, k2 1.23 ~1.2 (major), ~0.8 BL21(DE3) (pKD46) zwfF, zwfR 1.5 ~1.5 (major), ~0.9 BL21(DE3) (pKD46) Up550zwf, Down550zwf 2.6 ~2.5 BL21(DE3) zwf :: FRT-kan-FRT Down550zwf, k1 1.1 ~1.1 BL21(DE3) zwf :: FRT-kan-FRT zwf F, Down550zwf no band no band BL21(DE3) (pKD46) zwfF, Down550zwf 2.0 ~2.0 BL21(DE3) (pKD46) Down550zwf, k1 no band ~0.4 BL21(DE3) zwf :: FRT Up550zwf, Down550zwf 1.2 ~1.2

PAGE 89

72 Figure 2-17. Results from colony PCR reactions probing the zwf locus before and after the disruption. Abbreviations: WT (wild type), MT (mutant strain); kanR, kanamycin resistant; kanS, kanamycin sensitive. M1and M2, DNA markers.

PAGE 90

73 Table 2-8. Colony PCR for pntAB knock-out. E. coli strain Primers Set Expected Size (kbp) Observed Size (kbp) BL21(DE3) pntAB :: FRTkan -FRT Up510PntAF, Down510PntBR 2.35 ~2.3 BL21(DE3) pntAB :: FRTkan -FRT PntAF, PntBR no band no band BL21(DE3) pntAB :: FRTkan -FRT Up510 PntA, k2 1.23 ~1.2 BL21(DE3) (pKD46) Up510 PntA, k2 no band no band BL21(DE3) pntAB :: FRTkan -FRT Down510 PntBR, k11.03 ~1.1 BL21(DE3) (pKD46) Down510 PntBR, k1no band no band BL21(DE3) (pKD46) Up510PntAF, Down510PntBR 4.13 ~4.1 BL21(DE3) (pKD46) PntAF, PntAR 1.53 ~1.5 BL21(DE3) pntAB :: FRTkan -FRT PntAF, PntAR no band no band BL21(DE3) (pKD46) PntBF, PntBR 1.39 ~1.4 BL21(DE3) pntAB :: FRTkan -FRT PntBF, PntBR no band no band BL21(DE3) pntAB :: FRT Up510PntAF, Down510PntBR 1.11 ~1.0 The growth rate of mutant strains was compared with that of wild type in M9 minimal media. zwf strain grew almost as fast as the WT, at least for the first 8h, as it was reported before (Figure 2-19A).164 The pgi mutant showed limited growth and consumption of glucose (Figure 2-19A), as expected.165 In contrast to what has been observed for a K-derivatived E. coli strain166 lacking pntAB our pntAB mutant grew without any difference from the WT (Figure 2-19B). Finally, the icd mutant strain showed the reported elsewhere150 glutamate auxotrophy (Figure 2-19C). In this case, an additional portion of glutamate was added after 20h.

PAGE 91

74 Figure 2-18. Results from colony PCR reactions probing the pntAB locus before and after the disruption. Abbreviations: WT (wi ld type), MT (mutant strain); kanR, kanamycin resistant; kanS, kanamycin sensitive. M1 and M2, DNA markers.

PAGE 92

75 Figure 2-19. Phenotype analysis of constructed mutants in M9 medium. A. Growth curves for pgi and zwf strains. B. Growth curve for pntAB strain C. Growth curves for icd strain with (opened squares) or without (closed squares) glutamate.

PAGE 93

76 MS Data Analysis In this section, the raw MS data will be presented in a form of % fractional abundance of 72 (m/z) fragment (A72) of ethyl-hydroxybutyrate (EHB) obtained under different physiological conditions using di fferent glucose isotopomers for all the aforementioned strains. In the first experiment, the WT strain was fed with Glc-1D under aerobic conditions. A 20% incorporation of deuterium into product was observed. This experiment was repeated at least eight times (Figure 2-20) using different amount of cells, at two different scales and it was reproducible. In all cases, a plateau for the A72 (%) was observed after the first 2.5 h. We expected that pgi activity should give rise to labeled product whereas any other NADPH source should not contribute significantly to it. The second NADPD source in this case is the icd, albeit minor (i.e. pyruvate produced from Glc-1D is labeled at maximum 17%, Figure 1-3, Chapter 1). Figure 2-20. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the wild-type strain (WT) using Glc-1D, under aerobic condition.

PAGE 94

77 When the experiment with Glc-1D was repeated under anaerobic conditions for the WT, EHB was found labeled at a higher extent, up to 35% (Figure 2-21). However, a plateau value, if any, was not obtained at the first 6 h. Keeping in mind the equation presented in Figure 2-15 for the analysis of the data, a higher A72 (%) could be observed if a higher flux through the PP pathway occu rred (i.e higher [NADPD]) or if a major NADPH source was ceased (i.e. lower [NADPH]) under anaerobic conditions. Figure 2-21. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the wild-type strain (WT) using Glc-1D, under anaerobic condition. The distinct labeled pattern of EHB under aerobic or anaerobic conditions was confirmed with a third independent experiment in which the cells were fed with Glc-1D under aerobic conditions for the first 8h and then the air supply was halted (Figure 2-22). The possibility of a branch-point (i.e ED pathway) in between the pgi and the gnd was excluded by feeding the WT strain with Glc-3D (Figure 2-23). Almost the same degree of incorporated deuterium into the EHB (i.e., (18)%) was observed as when Glc-1D was used.

PAGE 95

78 Figure 2-22. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the wild-type strain (WT) using Glc-1D, under aerobic and anaerobic conditions. (Arrow indicates the time at which the air supply was turned off). Figure 2-23. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the wild-type strain (WT) using Glc-3D, under aerobic conditions. Glc-6,6D2 could trace a possible icd activity. Indeed, when the WT strain was fed with it, 11-13% of EHB was labeled under aerobic conditions. This is not a minute number, especially if we consider that less than a half of the pyruvate (i.e. maximum 33%) entering the TCA will be labeled (Figure 2-24).

PAGE 96

79 Figure 2-24. Fate of deuterium atoms of glucose at position 1, 3, 4 and 6 through glycolysis. Deuterium at position 1 should end up in PEP (25%), however only the position trans to carboxylic acid should be labeled, and PYR (only 16% labeled). Deuterium at position 3 should be found in water. Deuterium at position 4 should label NAD+. Deuterium atoms at position 6 should have the same fate as the atom at position 1, however PEP and PYR should bear twice more labeled atoms (50% and 33% respectively) (Adapted from references 176, 177) In contrast, under anaerobic condition the % A72 was diminished to 4%, a negligible number if we consider that even a reduction of EAA with NaBH4 is showing a 3.8% A72 due to the natural abundance of isotopes, as we will discuss in the next section. A very similar trend for the produced labeled EHB was detected when the WT strain was replaced by the pgi strain and fed with Glc-6,6D2 under either aerobic (i.e., 10-12%, Figure 2-26) or anaerobic conditions. The major difference compared to WT

PAGE 97

80 was observed when Glc-1D or Glc-3D was administrated to pgi strain under aerobic conditions. A higher amount of EHB (i.e., about 27% for Glc-1D and about 25% for Glc-3D) was found labeled (Figure 2-27 and Figure 2-28). These results could be easily explained since in this strain all labeled sugar should be catabolized through the PP pathway (i.e., higher [NADPD]). Figure 2-25. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the wild-type strain (WT) using Glc-6,6 D2, under aerobic (open squares and closed circles) and anaerobic conditions (open circles). (Arrow indicates the time at which the air supply was turned off) In contrast to what was observed for the pgi strain, the contribution of labeled Glc-1D or Glc-3D in EHB for zwf strain was negligible since the route through the PP pathway was disrupted (Figure 2-29). However, the contribution was maximized in this strain when Glc-6,6D2 was used under aerobic conditions. (Figure 2-30).

PAGE 98

81 Figure 2-26. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by pgi strain using Glc-6,6 D2, under aerobic (opened symbols) and anaerobic conditions (closed circles). Figure 2-27. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the pgi using Glc-1D, under aerobic (opened symbols) and anaerobic conditions (closed circles).

PAGE 99

82 Figure 2-28. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by pgi strain using Glc-3D. Figure 2-29. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by zwf strain using Glc-3D, or Glc-1D under anaerobic (opened circles) and aerobic conditions in 10 ml scale, OD600=7.0-9.0.

PAGE 100

83 Figure 2-30. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by zwf strain using Glc-6,6 D2, under aerobic (closed circles) and anaerobic conditions (opened circles) in 10 ml scale, OD600=8.0. The results for the obtained labeled EHB were reversed when an icd strain was used instead for a zwf strain. Maximum deuterium content in EHB was detected when Glc-1-D was used (Figure 2-31) and minimum when Glc-6,6-D2 was administrated (Figure 2-32). Figure 2-31. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by icd strain using Glc-1D, under aerobic (opened circles) and anaerobic conditions (closed triangles) in 10 ml scale, OD600=7.0-8.0.

PAGE 101

84 Figure 2-32. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by icd strain using Glc-6,6 D2 and anaerobic conditions in 10 ml scale, OD600=7.0-9.0. Finally, a very similar to pgi pattern was observed for the pntAB strain. (Figure 2-33). Figure 2-33. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by pntAB strain using Glc-1D, or Glc-6,6 D2 and anaerobic conditions in 10 ml scale, OD600=7.0-9.0.

PAGE 102

85 Evaluation of the MS Data The quality of the MS data and the physiological conditions under which they were collected were verified by number of control experiments. Physiological conditions E. coli K strain is unable to grow on glycerol under anaerobic conditions.168 In our experiments the conditions described as anaerobic were justified by the inability of the WT strain to produce EHB in presence of glycerol when the sparger fitted to the reaction vessel was connected to an argon line (Figure 2-34). In contrast, the same strain reduced the ethyl-acetoacetate when a stream of air passed through the reaction mixture in otherwise identical conditions. In presence of glucose, formation of product was observed when the biotransformation was carried out either anaerobically or aerobically (Figure 2-35). Figure 2-34. Production of EHB by WT strain fed with glycerol under aerobic and anaerobic conditions. OD600=8.5

PAGE 103

86 Figure 2-35. Production of EHB by WT strain fed with glucose under aerobic and anaerobic conditions. OD600=8.5-9.5. Calculation of the intracellular ratio NADPH/NADPD from mass data A typical fragmentation pattern of EHB is shown in Figure 2-15. Based on it, the ratio of unlabeled product ( PH)/ labeled product ( PD) could be calculated either from the pair of peaks A71, A72 or from A117, A118 if the naturally occurring elemental isotopes are also considered (i.e. the natural abundance of 13C is 1.1%). Even though both pairs gave similar results, we based our data analysis on the former pair due to its higher abundance. For the correction regarding the 72 peak, a sample of unlabeled EHB prepared from the reduction of ethyl-acetoacetate (EAA) with sodium borohydride was injected multiple times in the GS-MS and the % fractional abundance of A72 was found 3.8%. A similar correction was applied to peak 71. Apparently, when the EAA was reduced by sodium borodeuteride the % fractional abundance of A71 was 4.6%.

PAGE 104

87 Therefore the corrected ratio (A71/A72)COR could be derived from the observed ratio (A71/A72)OBS if the estimated error for each peak (A)ER is subtracted (Figure 2-36). Figure 2-36. Equation for derivatization of the corrected A71/A72 from the observed ration. A second concern was the possibility of an undesired isotope fractionation of the EHBs, labeled and unlabelled, during the GC separation. It is known that molecules bearing heavier atoms but otherwise identical are eluted first from the capillaly columns.169 To avoid bias due to this fractionation the full peak range of the EHB signal was integrated. The statement that the ratio PH/ PD, derived from the corrected ratio A71/A72, is at any time equal to intracellular NADPH/NADPD ratio is true only if the two reduced nicotinamide cofactors substituted with different isotopes (i.e NADPH and NADPD) reduce, via Gre2, the EAA, with the same rate. This may or may not be true for an enzyme-catalyzed reaction depending on whether or not the component step related with the cleavage of the bond containing the isotope is rate-determining for the overall reaction.170 In the latter scenario, the PH/ PD is proportional but not equal to NADPH/NADPD ratio and a correction including the kinetic isotope effect factor ( fKIE) must be employed (Figure 2-37).

PAGE 105

88 Figure 2-37. Proposed correction for the NADPD fraction based on the corrected A71/A72 ratio and a kinetic isotope effect factor fKIE. Unfortunately, the experimental determination of fKIE is not a facile task, especially if we consider that the intracellular concentration of substrate, coenzyme, enzyme, product and potential inhibitors are not well defined and the absolute value of fKIE may depend on any of these effectors. Nevertheless we attempted to verify if an in vitro isotopic discrimination is plausibly for Gre2. To do so, we first determined the strereospecificity of Gre2. Numerous of methods have been developed, in the past, for labeling each of the diastereomeric protons of NADPH at the 4-position.171 We prepared both NADPDB and NADPDA using a known B-specific enzyme (Figure 2-28A) and an A-specific enzyme (Figure 2-38B) correspondingly, as described in experimental section. Both nucleotides reacted separately with Gre2 and EAA ((Figure 2-38C). When NADPDB was the reduced cofactor, deuterium found incorporated into product, as confirmed by GC-MS, and the oxidized cofactor was unlabeled, as confirmed by H-NMR (Figure 2-38D1).172 The opposite pattern was observed when NADPDA replaced NADPDB (Figure 2-38D2).

PAGE 106

89 Therefore, the Gre2 is a B-specific enzyme as predicted by its sequence similarity to 3-hydroxysteroid dehydrogenases and to plant cinnamoylCoA reductases (all of which are B-specific enzymes).173 Finally, for an estimation of the apparent kinetic isotope effect (fKIE) by a competitive method, a mixture of NAPDH and NADPDB was allowed to react with the same substrate (i.e. EAA) in the presence of Gre2 simultaneously. The rationale behind this experiment is that at 100% conversion, the fractional abundance of peak 72 should reflect the initial fraction of labeled and unlabeled cofactors and should be constant during the time course of the reaction, in absence of an isotopic discrimination. Clearly, this was not the case at least under the conditions that this experiment was carried out (Figure 2-39). At a conversion lower that 50 % a lag-phase for the deuterated species was detected and we may assume final 72 71 initiail 72 71 KIEA A A A f or fKIE~1.3. Figure 2-39. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by Gre2 in presence of NADPH and NADPD.

PAGE 107

90 Figure 2-38. Preparation of stereospecific labeled NADPH and elucidation of Gre2 facial selectivity by H-NMR.

PAGE 108

91 Another question related to isotopic discrimination is if the extent by which its each particular pathway is utilized, is changing when the unlabeled sugar is replaced by the deuterated one. In such a case an underestimation of the partition from the step related to reduction of NADP+ is possible but a correction is not straightforward. Also, we should mention that the results presented above are the average incorporation during the time-frame reported for each biotransformation. Even though under aerobic conditions a plateau was achieved relatively early, this was not the case under anaerobic conditions. The lower rate of product formation as well as changes on relative fluxes, KIEs and physiological conditions (i.e., pH changes) may account for this phenomenon. Lastly, when labeled EHB was incubated only with cells and glucose, a 1.1 % loss of deuterium per hour was observed: however, this may not happen in presence of EAA and in any case since most data were collected during the first 5-6 hours the overall loss should not exceed the 5-6%. The best way to evaluate the consistency of the suggested corrections is by applying them to the initial MS data. The overall results are summarized in Table 2-9. Table 2-9. Fractional abundance of NADPD. Fractional abundance of NADPD (%), aerobically Fractional abundance of NADPD (%), anaerobically E. coli strains Glc-1-D Glc-3-D Glc-6,6-D2 Glc-1-D Glc-3-D Glc-6,6-D2 WT 21-23 20-22 10-12 30-40 ---a 0.1-0.5 pgi 28-34 27-30 9-12 37-41 --0.1-0.5 zwf 2-3 0.0-0.3 22-23 0.0-0.4 --3.0-4.0 icd 35-49 --1-2 25-27 --0.0-0.1 pntAB 29-30 --10-11 44-48 --0.0-0.8 aNot determined. Experiments under anaerobic conditions in pres ence of Glc-1D could serve as a guide for this evaluation. Under these conditions and especially in absence of a possible PntAB transhydrogenase activity, the theoretical maximum value for the fractional abundance of NADPD should be 50%, since there is no icd activity (< 1% deuterium content from Glc-6,6-D2) and the protium at position 3 of glucose should contribute the

PAGE 109

92 rest 50 % of the unlabeled product via Gnd. The experimental value, with corrections included, is 44-48%. Slightly lower numbers were estimated for the WT and pgi strains. The null incorporation for zwf strain is also expected in the absence of G6PDH activity in this mutant. The only inconsist ency is the somewhat lower value calculated for icd strain, even though this number is based only on one experiment. However, the same sugar under aerobic conditions labeled the product at 35-49% as if, in absence of icd, zwf and gnd are almost the exclusively contributors on NADP+ reduction. In contrast, in the WT strain under the same conditions, the two NADPH sources in PP pathway contribute at levels of ~21% each. The strong participation of icd in regeneration of NADPH is apparent for the WT. The ~11% labeled EHB when cells were fed with Glc-6,6-D2 suggests this trend. This number must be multiplied by an additional factor of 2 to 3 (Figure 2-24) in order for it to reflect the real contribution of TCA cycle in NADPH production in WT strain, under aerobic conditions. This participation is almost absent under anaerobic conditions in all strains but zwf It is unclear is this small, but measurable amount of deuterated product is due to maeB activation in this particular strain or due to experimental error. Also, one may ask if the small (2-3%) but measurable difference of deuterium content in EHB in between cells fed with Glc-1D and cells fed with Glc-3D is due to a small participation of ED pathway. This is unlikely at least for the WT strain. The most probable explanation is that the amount of labeled product in presence of Glc-1D reflects the sum of PP pathway activity and TCA cycle activity, whereas in presence of Glc-3D only the PP pathway activity is detected since deuterium at position 3 of glucose should be lost in water through glycolysis (Figure 2-24). Nevertheless, the contribution of deuterium from position 1 of glucose to the NADPD pool, due to icd activity only, is more than twice lower than the one from Glc-6,6-D2. In principle, we should expect exactly twice less contribution (i.e.

PAGE 110

93 two over one atoms labeled). For example, in zwf if Glc-6,6-D2 contributed to 22% the Glc-1-D should have participated at 11%, but only 2-3% was the experimental obtained labeled product. The hydrogen atom at position 1 may be prone to exchange with water more easily than the hydrogen atoms at position 6.167 Issues related with the stereospecifiity of TCA cycles enzymes may be another explanation.174 Productivity of Mutant Strains versus the Wild Type in a Shake Flask The zwf and icd strains produced EHB to lesser extents as compared to wild type (Figure 2-40). This was an expected outcome if we consider that the reaction is absolutely NADPH-dependant and both strains are missing two and one NADPHregeneration enzymes, respectively. Less problematic was the productivity of pntAB strain (Figure 2-40) indicating a minor contribution, if any, of this source to NADPH regeneration machinery. What about the pgi strain? All literature data suggested that the NADPH produced in this strain is in excess compared to its NADPH biosynthetic needs. Our experiments in a shake flask showed no significant difference from the WT strain, within the experimental error, in product rate formation at least for the first four hours. The observed retardation of product formation for the WT strain after the first six hours was due to faster pH drop in the flask containing this strain compared to the one containing cells of pgi strain. When the pH was adjusted manually in both flasks every three hours, the previously observed difference was diminished (data not shown). The same trend was observed when the rate of product formation of the two strains was tested in a 2-L fermenter under supervised conditions.175 The lower rate of glucose consumption observed for pgi may contribute to slower production of organic acids by this strain.

PAGE 111

94 Figure 2-40. Product formation of Gre2 overexpressed either in zwf or icd or pntAB strain. Data are normalized per biomass concentration expressed as optical density of culture medium at 600 nm. Figure 2-41. Product formation of Gre2 overexpressed either in pgi or WT strain. Data are collected from three independent experiments and normalized per biomass concentration expressed as optic al density of culture medium at 600 nm.

PAGE 112

95 Discussion By combing isotopic tracer experiments with the overexpression of a NADPHdependant reductase in a E. coli BL21(DE3) strain and using a very simple experimental setup, we successfully revealed at least three NADPH regeneration sources within this biocatalyst when glucose was used as a carbon source. The validity of our results in the wild-type strain was justified further by repeating the labeled studies in isogenic to WT strains overexpressing Gre2. Even though the presented methodology belongs to low resolution techniques for elucidation of metabolic fluxes, it still allows us to draw some conclusions about the relative participation of each NADPH source to ethyl-acetoacetate reduction in WT strain. Unfortunately, due to the additional, albeit not well defined, adjustments required for the derivatization of the icd contribution to NADPD pool from the experimental data (i.e. not one to one ratio of deuterium atoms between the donor and acceptor molecules and possible dilution of deuterium through glycolysis) this estimation is prone to higher error. Nonetheless, in order to ease our discussion in this section, we multiplied our numbers for the Glc-6,6D2 experiments by a factor of 3 assuming that trioses from the EMP pathway are accessing the TCA cycle only via pyruvate and that no scrambling from prot ium atoms occurred during this process. Clearly, both the PP pathway and the TC A cycle are competing equally in producing reductive equivalents for bioreduction (Figure 2-42) in this bioprocess. It would be beyond the resolution of this method to state which source dominates. Unfortunately, labeled experiments with Glc-4D, which may uncover a transhydrogenase role, were not performed in this study because of the high cost of this isotopomer sugar. However, we may be able to predict if a transhydrogenase activity is possible if we compare the results obtained from the WT strain with those from the pntAB strain.

PAGE 113

96 Figure 2-42. Proposed participation of NADPH regeneration sources in E. coli BL21(DE3)(pAA3) resting cells. In both Figure 2-42 and Figure 2-43, a transhydrogenase activity may fit under the label other sources. But, at least for the latter strain, a transhydrogenase activity towards the production of NADPH (i.e. pntAB) must be excluded since the structural gene is deleted. Hence, in WT strain, only half of the percentage denoted as other sources should be ascribed to possible pntAB action (i.e.10-20%). This hypothesis reveals that in pgi strain such an activity is also unlikely. A participation of pntAB similar to one observed for the wild type could be assumed in case of zwf mutant whereas more uncertain are the results for the icd strain. Ultimately, the question that is coming next is how our results could be correlated to the data obtained from similar studies for other systems. Before we attempted such a comparison, it is worth noting that the prediction made by Walton and Stewart about the importance of TCA cycles contribution in NADPH production, in our system, was on the right track. One reason for the possible overes timation of this contribution by a factor of

PAGE 114

97 2 compared to the data obtained in this study is that the authors almost ignored that oxidative catabolism does not necessarily imply complete oxidation of glucose.147 Figure 2-43. Proposed participation of NADPH regeneration sources in E. coli BL21(DE3)(pAA3) mutant stains under non-growing conditions. To compare our results with the already published data we will use two studies as references: i) the work reported by Csonka and Fraenkel176 because the methodology used in that work is more similar to our study (i.e. both are low resolution methods based on labeled hydrogen atoms attached to glucose molecule) and ii) the work done by Sauer et. al.177 because it is offering the most complete picture of the glucose catabolism in E. coli since is based on net flux measurements. In sharp contrast to what the data from the former study showed, our experiment al results implied a strong participation of Icd in the intracellular NADPH pool. Despite the controversial issue about the exact percent of contribution from this NADPH source, the facts that i) we observed higher deuterium content in product from the Glc-6,6D2 than from the Glc-1D, which the

PAGE 115

98 previous authors did not, ii) they observed in corporation of deuterium into amino acids from position 6 of the glucose even under anaerobic condition or even when they used a icd mutant and we did not, suggested, at least in our system, participation of icd to NADPH regeneration sources. Their suggestion for about 40% participation of PP pathway in NAPDH pool is inline with our results. Open is the possibility for a fourth regeneration source, most likely a transhydrogenase activity, in both studies. The most solid experimental evidences for the pntAB participation are provided by Sauers laboratory. We should mention, though, that even with their high resolution technique, the 35-45% estimation for the PntAB partition was based on an indirect calculation. They calculated the specific rate of NADPH production from the PP pathway and TCA cycle (i.e., 3.9 mmol h-1 g-1(DCW) and 2.3 mmol -1 g-1(DCW) respectively) from the metabolic flux analysis data. They then estimated the specific rate of NADPH consumption (i.e. 11 mmol h-1 g-1(DCW)) for biosynthetic needs of E. coli cell, using literature data and the calculated growth rate and since they found that the latter is higher than the former they proposed that the gap should be covered by the pntAB activity. Despite that, they also arrived to the same conclusion about the approximately 40% contribution from the PP pathway to NADPH pool as we did, whereas they estimated the merits for icd to 25% maximum. The last question, and most difficult to answer, is why our pgi strain did not show an improved productivity. We should not conceal the fact that the ultimate goal of this study was to improve the productivity of this NADPH-dependant bioprocess. But, in order to improve any process, we should be aware of the theoretical maximum that we are seeking to reach. Unfortunately, this is not a straightforward estimation due to the inherent complexity of our system. For example, what if the productivity of our system is limited by the rate of EAA transfer through the cell membrane? Or what if it is limited by

PAGE 116

99 the amount of the overproduced enzyme? Someone may suggest in vitro measurements for calculating the specific activity of Gre2, but again this kind of measurements may serve as a guide but they could exclude a higher or lower in vivo activity for the same biocatalyst. Therefore, if the productivity of our system is limited by a factor other than the NADPH availability within the cell, any manipulation regarding this issue will fail to improve the rate of produc t formation. To modify our original question we will ask if the apparent failure of the pgi strain towards the improvement of product rate formation is evidence that our systems maximum limit does not depend on NADPH regeneration rate. The answer is, absolutely not. Our study does not exclude the possibility of a back-flux from the NADPH to NAD+ via the soluble transhydrogenase UdhA as well as the co-existence in pgi mutant of a novel PEP-glyoxylate cycle, which does not produce NADPH and is redundant to UdhA activity.178 For example, from Sauers study we know that the production rate of NADPH in pgi stain is the same as the one in the WT, the difference is, the much lower specific rate consumption which in turn activates the back-flux activity via UdhA. A more clear solution to our problem it would be to use a double pgi udhA knockout strain Unfortunately, the Pgi-UdhA phenotype is lethal on glucose and it is unknown how this mutant grows on complex media. Future Work The ultimate goal of any bioprocess optimization is to increase the rate of product formation per liter of medium (volumetric productivity) and to improve the final product yield per liter (product titre) by using the minimum amount of biomass and carbon source. For a two step process, as the one that we described in this chapter, the expression levels of the NADPH regeneration sources are governed by the biosynthetic needs of the E. coli cell in the first phase (i.e. growth in complex media). Those levels

PAGE 117

100 may not be sufficient to meet the needs of an overproduced, NADPH-dependent biocatalyst in the second phase. Based on Walton and Stewart data, the productivity of this bioprocess was ~120 U/g (DCW) for resting cells, whereas the NADPH regeneration rate of aerobic growing E. coli cells in M9 medium is 180U/g (DCW).177 Therefore, we cannot exclude the possibility that in rich media, the cells are not adjusting their metabolism for the optimum NADPH regeneration rates observed in minimal, glucoselimited media. Controlled induction of zwf with concomitant deletion of competing for NADPH utilization enzymes may shed light on this problem. In addition, replacement of the NAD+ -dependant D-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) with a novel NADP+-GAPDH from Kluy v eromyces lactis 179, 180 and or overexpression of a transhydrogenase may add a NADPH regeneration source in the EMP pathway level and thus bypass the product formation dependence on TCA cycle which in turn depends on oxygen availability. In cases where glucose must be replaced by glycerol, expression of the dha operon (i.e., glycerol dehydratase ( dhaB ), 1,3-propanediol oxidoreductase ( dhaT ), glycerol dehydrogenase ( dhaD ) and dihydroxyacetone kinase ( dhaK )) from Klebsiella pneumoniae for dissimilation of glycerol under anaerobic conditions181, 182 could enhance the productivity of biocatalyst under oxygen-limited conditions. The above suggestions may be proved useful for easy substrates in other words for substrates that the cell membrane does not impose a barrier and the glucose catabolism is not poisoned by them. Unfortunately, those substrates are rather the exception than the rule in whole-cells biocatalysis. For those difficult substrates any combination of overexpressed reductase with an overexpressed NADPH-regeneration enzyme in the same or in separate, intact or permeabilized cells must be first employed in a small, exploratory scale before the progression to the final scale-up process.

PAGE 118

101 CHAPTER 3 E. COLI BL21 (DE3) yqhE STRAIN: A CLEANER HOST FOR OVERPRODUCTION OF NADPH-DEPENDENT KETOREDUCTASES Of the chemical aids in the living organism the ferments mostly referred to nowadays as enzymes are so pre-eminent that they may justifiably be claimed to be involved in most of the chemical transformations in the living cell. The examination of the synthetic glucosides has shown that the action of the enzymes depends to a large extent on the geometrical structure of the molecule to be attacked, that the two must match like lock and key. Consequently, with their aid, the organism is c apable of performing highly specific chemical transformations which can never be accomplished with the customary agents. To equal Nature here, the same means have to be applied, and I therefore foresee the day when physiological chemistry will not only make extensive use of the natural enzymes as agents, but when it will also prepare synthetic ferments for its purposes. -Syntheses in the purin and sugar group, Emil Fischer183 Introduction In 2000, seventy-six percent of the new drugs introduced in the market were single enantiomers, compared with 21% in 1991.184 The reasons for this include the established, by pharmacological research, differential, enantiomer-specific, relationship between the drug and the target biological macromolecule and the desire to minimize the cost and the wastage resulting from the production of unwanted enantiomers.185 The benefits derived from the asymmetric and catalytic operation of enzymes have been recognized today by both the academic and the industrial community.186-189 The most popular biocatalytic approach to asymmetric ketone reduction is the use of commercial bakers yeast.190 Whole cells of this microorganism provide a simple and inexpensive way to produce chiral alcohols si nce they possess a vast number of ketoreductases and they are able to regenerate simultaneously the required cofactors through the dissimilation of a cheap carbon source. Unfortunately, the reductive power of yeast, due to plurality of reductases in it, can also become a disadvantage in

PAGE 119

102 situations where two or more reductases with conflicting stereoselectivities are acting on the same pro-chiral ketone.191 The net result for this operation, which is not rare for this biocatalyst, is a product with a low enantiomeric purity. A number of empirical methods were developed in the past in an effort to improve the stereoselectivies of this microorganism.190, 191 Additionally, strain engineering strategies, including gene disruption and overexpression of the desired reductase in the same host, showed promising results but they did not completely overcome the problem.192, 193 An alternative strategy, initiated and completed by former students in our laboratory, gave an almost checkmate so lution to the aforementioned problem.194, 195 Briefly, a list of potential keto-reductases (ca. 50) was compiled by analyzing the complete genome sequence of S. cerevisiae196 and the genes encoding half of them were cloned, as GST-tagged, and overexpressed in the E. coli one by one. This methodology leads to: higher amounts of the desired biocatalyst, higher flexibility in choosing expression vectors, a simple and relatively fast enzyme-purification protocol, flexibility in using, at any time, the biocatalyst either as a single purified protein or as an overproduced enzyme in intact E. coli cells. The last point is of paramount importance fo r verifying the inherit stereospecificity of each biocatalyst since the ee values obtained when the purified enzyme is employed should match with the ones obtained when the same biocatalyst is used in whole-cells, in absence of competing E. coli reductases. Unfortunately, even this microorganism is not totally devoid of endogenous carbonyl-reductase activities. Carbonyl Reductases in the E. coli In 1996, two reports described the presence of carbonyl reductases in E. coli Miya et al.197 showed that whole cells of a number of Gram-negative bacteria, including E. coli reduced one specific -keto ester (ethyl 2-methylacetoacetate) to syn -(2 R ,3 S ) alcohol, whereas Misra et al.198 reported the purification of two enzymes, one NADH-

PAGE 120

103 dependent and another NADPH-dependent, from the E. coli capable to reduce methylglyoxal to acetol. Even though the molecular weight of both reductases was estimated by gel filtration on Sephacryl column (i.e. ~ 100.000 Da), the amino acid sequences were not determined and therefore the genes encoding them remained unknown. Yum et al.199 were the first who identified the yqhE and yafB genes as potentially encoding 2,5-diketo-D-gluconate reductases because of their high sequence similarity to other bacterial homologues with known function. They verified further their prediction by cloning and overexpressing the gene products in the same host. Although they showed that both enzymes catalyzed the reduction of 2,5KDG to 2-keto-L-gulonate, no other substrate was tested in that study. In our laboratory, the activity of keto-reductases in unmodified E.coli cells was recorded, for the first time, when two members of the aldo-keto reductase (AKR) superfamily (Gre3p and Gcy1p) were overexpessed in E. coli and the whole cells were used to reduce a panel of -keto esters.194 Subsequently, reductase activity against ethyl 2-methylacetoacetate was used to purify the enzyme responsible for this bioreduction from 100 g of E. coli BL21(DE3) cells.200 The N-terminal amino acid sequence analysis of the purified reductase matched with the protein encoded by the yqhE gene. A BLAST search of YqhE sequence against the EcoGene201 database revealed the presence of nine putative AKRs in E. coli (Table 3-1). Recently, an independent study by Ko et al.202 indicated that four of them ( yafB, yqhE yeaE and yghZ ), out of the five tested, are involved in the conversion of methylglyoxal (MG) to acetol in presence of NADPH. The authors suggested that MG may serve as endogenous substrate for these AKRs based on the fact that: i) 2,5-DKG does not appear to be an endogenous substrate of the E. coli cell ii) MG could be accumulated into the cells under physiological conditions with an uncontrolled carbohydrate metabolism and iii) the kcat/KM

PAGE 121

104 of YqhE for 2,5-DKG is lower than the one obtained for MG. However, strain lacking the yqhE did not exhibit the increased in vivo MG susceptibility that was observed for the other three knockout strains. With the dkgA reductase being the best characterized enzyme, with broad substrate specificity (Figure 3-1), among the AKRs in E. coli the question for their physiological role remains unanswered.203 However, multiple sequence alignment of all nine open reading frames (Figure 3-2) reveal ed the presence of the conserved among AKRs superfamily catalytic tetrad of amino acids (Tyr, His, Asp and Lys) in all but ydbC (i.e., His is replaced by Arg, not shown in Figure 3-2) which facilitate the transfer of a hydride ion from NADPH to the substrate wi th a concomitant protonation of the nascent alcohol. In addition to conserved catalyti c mechanism, the recently obtained crystal structure of DkgA204 showed the characteristic ( / )8 TIM barrel motif (Figure 3-3), common for all AKRs.205 Table 3-1. Putative AKRs in E. coli Primary gene name (alternative name)201 Function Identity (%) Crystal structure (PDB)206 Ref. dkgA ( yqhE ) 2,5-diketo-D-gluconate reductase A, beta-keto ester reductase; methylglyoxal reductase; NADPH-dependent 100 1MZR 199, 200, 202 dkgB ( yafB ) 2,5-diketo-D-gluconate reductase B, methylglyoxal reductase 38 --199 yeaE Methylglyoxal to acetol; NADPHdependent 28 --202 ydjG function unknown; NADHdependent; use methylglyoxal as substrate 25 --201 yajO 2-carboxybenzaldehyde reductase, function unknown 22 --201 ydhF Unknown 25 1OG6 201 tas ( ygdS ) Suppresses tyrosine requirement of tyrA14 O6 strain 31 1LQA 207 ydbC Unknown 23 --201 yghZ Methylglyoxal reductase 34 --208

PAGE 122

105 Because our laboratory had accumulated strong evidences for the participation of YqhE in the reduction of -keto esters, the aims of this work were: To knock out the yqhE gene in E. coli BL21(DE3) strain using a methodology which enables marker-free deletions. To assess the impact of the mutated strain on -keto esters reductions. Figure 3-1. Substrate specificity of dkgA. Relative activity among different substrates is indicated in parentheses (e.g. (++++), highest, (-) lowest). Figure 3-2. Stereoview of the ribbon structure of DkgA (chain A). This picture was rendered in 3D-Mol Viewer.

PAGE 123

106 Figure 3-3. Multiple alignment of eight AKRs from E .coli The alignment was generated using the TCoffee web server software.209

PAGE 124

107 Experimental Procedures Most of the experimental procedures were described in Chapter 2. Superbroth (SB) medium used for cell growth contained Bacto-Tryptone (32 g/L), Bacto-Yeast Extract (20 g/L), sodium chloride (5 g/L) and 5 ml 1M NaOH. The analytical separation of the diasteroisomers obtained from the bioreduction of ethyl 2-fluoroacetoacetate was conducted with a Hewlett-Packard Model 5890 GC fitted with a 30 m DB-17 column and a temperature gradient from 55 C to 180 C at 5 C/min with initial and final times 2 and 5 min, respectively. The synthesis of the FRT-kan-FRT cassette with short flanking homologous regions of E. coli BL21 (DE3) DNA was carried out in one-step PCR using long chimeric primers in which the 3 end primed to beginning or the end of the FRT-kan-FRT cassette and the 5 end (46nt) was homologous to upstream or downstream region of the yqhE (Appendix C). Results and Discussion Characterization of yqhE Strain Initially, we attempted to disrupt the yqhE in the E. coli BW25113 strain using the FRT-kan-FRT cassette flanking with short homologous regions (ca. 50 bp), as described by Datsenko and Wanner.210 In total, 50 kanR colonies from six independent transformations were screened. Unfortunately, none of them contained the desired replacement. Taken into account both the lower transformation efficiency of the E. coli BL21(DE3) strain, which was the desired host for the overexpression of the ketoreductases and the higher recombination efficiency observed by Court et al.211 when markers with longer homologous arms were used, we shifted to a marker cassette with longer (ca. 500 bp) homologous regions. One kanR colony was obtained after the

PAGE 125

108 transformation with the longer piece. Colony PCR confirmed the desired disruption (Figure 3-4 and Table 3-2). Figure 3-4. Results from colony PCR reactions probing the yqhE locus before and after the disruption. Abbreviations: WT (wild type), MT (mutant strain); kanR, kanamycin resistant; kanS kanamycin sensitive; M1and M2, DNA markers. Table 3-2. Colony PCR for yqhE knock-out. E. coli strain Primers Set Expected Size (kbp) Observed Size (kbp) BL21(DE3) yqhE :: FRTkan-FRT yqhEF, yqhER no band ~2.6, ~1.9, ~0.8, (all minor) BL21(DE3) yqhE :: FRTkan-FRT Up500yqhE500, Down500yqhE500, 2.3 ~2.3 BL21(DE3) yqhE :: FRTkan-FRT Up550yqhE, yqhER no band no band BL21(DE3) (pKD46) Up550yqhE, yqhER 1.38 ~1.4

PAGE 126

109 Table 3-2. Continued. E. coli strain Primers Set Expected Size (kbp) Observed Size (kbp) BL21(DE3) yqhE :: FRTkan-FRT Up500yqhE500, k2 1.18 ~1.2 BL21(DE3) (pKD46) yqhEF, yqhER 0.8 0.8 BL21(DE3) (pKD46) Up500yqhE, Down500yqhE 1.8 ~1.8 BL21(DE3) yqhE :: FRTkan-FRT Up550yqhE, Down550yqhE, 2.4 ~2.4 BL21(DE3) yqhE :: FRTkan-FRT Up550yqhE, k2 1.24 ~1.25 BL21(DE3) (pKD46) Up550yqhE, Down550yqhE 1.9 ~1.9 BL21(DE3) yqhE :: FRT Up550yqhE, Down550yqhE 1.2 ~1.2 Biotransformations To assess the impact of the single knockout E. coli strain on -keto ester reductions, we carried out three sets of bioconversions using known substrates for the DkgA reductase. Cultures of both strains were grown in 100 ml SB medium at 37 C under aerobic conditions in 1 L Erlenmeyer fl asks. There was essentially no difference in growth rates between the parent and the yqhE knockout strain (Figure 3-5). After 3.5 h of growth, substrate was added to a final concentration of 5 mM. Both ethyl-2methylacetoacetate and ethyl 2-allylacetoacetate were reduced completely by intact E. coli BL21(DE3) cells (WT) under growing conditions. By contrast, even extended reaction times (48 h) failed to result in formation of the corresponding, substituted at alpha position, -hydroxy esters when the mutated strain (MT) was employed instead. However, the reduction of the unsubstituted ethyl acetoacetate was only slightly affected by the presence or absence of the dkgA reductase (Table 3-3). Also, it is worth noting that when the bioconversion of ethyl-2-methyl acetoacetate was attempted with resting cells (OD600= 9.0) no conversion was observed, by either strain, even after 2 days incubation.

PAGE 127

110 Figure 3-5. Growth curves for E. coli BL21(DE3) and yqhE strains. Arrow indicates the time point at which -keto ester was added. Table 3-3. -keto ester reductions by the starting and yqhE knockout strains. Strain BL21(DE3) 100% reduction, 24h 100% reduction, 40h 96% reduction, 36h BL21(DE3) yqhE :: FRT-kan-FRT no reduction product detectable, 48h no reduction product detectable, 48h 72% reduction, 36h Another interesting group of chiral synthons212 was prepared recently in our laboratory by reducing stereoselectively mono-substituted, at alpha position, fluoro ketoesters by a set of purified keto-reductases from bakers yeast.213 Reduction of the simplest fluoro compound, ethyl 2-fluoroacetoacetate, could lead to four possible stereoisomers (Figure 3-6). Both E. coli stains, wild type and yqhE, reduce the abovementioned compound under non-growing conditions, albeit with different rates (Figure 3-7). The bioconversion by the parent stain was completed in 7 hours, whereas only about 50 % of the 10 mM

PAGE 128

111 starting material (SM) was converted to diastereisomeric procucts (PA, erythro) and (PB, threo) by the mutated strain at the same time. Figure 3-6. The four possible stereoisomers of ethyl 2-fluoroacetoacetate reduction. Figure 3-7. Profile of ethyl 2-fluoroacetoacetate reduction from E. coli BL21(DE3) (WT) and yqhE (MT)strains. The peak area of each eluted compound from the GC column was intergraded and normalized based on the peak area of the methyl-benzoate which was used as internal standard (IS). Interestingly, the rate formation of the one diastereisomer (PA) was unaffected by the absence of dkgA reductase, whereas the formation of the second one (PB) was drastically diminished in case of yqhE strain. The results clearly suggest the presence

PAGE 129

112 of at least one more active keto-reductase in E. coli with different stereospecificity from the dkgA. Conclusion The single knockout E. coli BL21(DE3) yqhE strain constructed in this work succeeded in eliminating the reduction background observed initially for a number alpha substituted -keto-esters. However, as new substrates are screened (i.e., fluorosubstituted keto-esters) and other type of reductases are overexpressed in this host (i.e., en-one(al) reductases) in our laboratory, we cannot exclude the possibility for a requirement of double or triple E. coli mutant in the future. Fortunately, the experimental method elected for the disruption of the YqhE gene is compatible with the deletion of additional genes using the yqhE as parent strain, in a time frame of one to three weeks.

PAGE 130

113 CHAPTER 4 BIOCATALYTIC REDUCTION OF FUNCTIONALIZED ALKENES: A SURVEY FOR ENREDUCTASES When in 1933 I went on a Rockefeller fellowship to Otto Warburgs institute in Berlin, Warburg and Christian had in the previous year produced a yellow-coloured preparation of an oxidation enzyme from yeast. It was here for the first time possible to localize the enzymatic effect to a definite atomic constellation: hydrogen freed from the substrate (hexose monophosphate) is, with the aid of a special enzyme system (TPN-Zwischenf erment) whose nature was elucidated somewhat later, placed on the nitrogen atoms of the flavin (1) and (10), giving rise to the colourless leucoflavin. This is reoxidized by oxygen, hydrogen peroxide being formed, and may afterwards be reduced again, and so forth. This cyclic process then continues until the entire amount of substrate has been deprived of two hydrogen atoms and been transformed into phosphogluconic acid; and a corresponding amount of hydrogen peroxide has been formed. At the end of the process the yellow enzyme is still there in unchanged form, and has thus apparently, as Berzelius expressed himself, aroused a chemical affinity through its mere presence. -The nature and mode of action of oxidation enzymes, Hugo Theorell214 Introduction Asymmetric hydrogenation of prochiral functionalized alkenes by chiral rhodium or ruthenium phosphines has resulted in an impressive number of enantioselective transformations during the last twenty years.215 A more critical evaluation reveals that the presence of certain functional groups (amide, carboxylate, alcohol) in proximity to the carbon-carbon double bond is a prerequisite for a high enantiomeric excess. However, aprotic oxygen functionalities such as ketones or esters have rarely been employed with success. 216-217 Moreover, the harsh reaction conditions that are often required, the complex preparation of the chiral reducing agents, as well as the demand for environmentally friendly chemical processe s are placing the organometallic methodology far behind from an ideal approach. Biohydrogenation on the other hand appears as an attractive alternative route.

PAGE 131

114 In the field of biotransformations, intact cells mediate the majority of the biocatalytic reductions of unsaturated compounds reported in the literature.220 There are two major reasons: i) the absolute requirement of reducing equivalents in the form of NAD(P)H, which can be easily and inexpensively regenerated in a cell, and ii) the lack of awareness, in many cases, with regards to particular enzyme catalyzes the saturation of the carbon-carbon double bond. The dominant term for describing the abovementioned enzymes, especially among organic chemists involved in biohydrogenations, is enoate-reductases.221, 222 We will not adopt this term here, but we will rather assign them the more generic term, enreductases. We will not also follow the common pattern observed in most reviews related to this subject220, 222-224 in which unsaturated compounds were grouped based on a common function group (e.g. enoates, enones, enals, nitro-alkenes etc.) or based on a common carbon skeleton (e.g. aliphatic, alicyclic etc.) and then biocatalysts, capable of reducing certain class were listed. Most often, these reductions utilized whole cells. While this type of classification beautifu lly demonstrates the enormous reductive capabilities of these bugs and the diversity of functional groups that they could accept,225, 226 it does not contribute to our understanding in terms of which specific enzyme is responsible for each particular bioreduction. This is the major scope of the present chapter. We will briefly introduce the biohydrogenations with whole cells in the first part whereas in the second part we will assign substrates to different en-reductase families. Structure-function implications will be also discussed. Biohydrogenation with Whole Cells Bakers yeast is more frequently used for the reduction of aliphatic unsaturated aldehydes and ketones leading to chiral synthons (Figure 4-1) that can be used to synthezise natural phytol,227 -tocopherol228, 229 and insect pheromones.230, 231

PAGE 132

115 Figure 4-1. Preparation of a building block with two chiral centers from citral using bakers yeast (BY) as biocatalyst. The reduction of unsaturated aldehydes is usually accompanied by the reduction of carbonyl group to the corresponding alcohols.232 By contrast, allylic alcohols must be first oxidized to corresponding aldehydes233-235 in order for the saturation of the C-C double bond to occur (Figure 4-2). However, the reduction of the double bond (db) is chemo-selective in the sense that only the db that is conjugated to the carbonyl group is reduced (Figure 4-3).Error! Bookmark not defined. Figure 4-2. Proposed hydrogenation of a llylic alcohols by bakers yeast. Figure 4-3. Chemo-selective reducti on of a trienal by bakers yeast. Bakers yeast cannot reduce unsaturated carboxylic acids unless a halogen is attached to carbon.236 The stereochemical outcome in this case depends on the initial

PAGE 133

116 diasteroisomer used (i.e. E or Z ).237 Unsaturated esters also cannot be reduced by BY with the exception of the unsaturated lactones (Figure 4-4)238 Figure 4-4. Preparation of goniothalam in, a putative anticancer compound. Hydrogenation of nitro-alkenes by bakers yeast afforded optically active nitro compounds239, 240 which are important intermediates in asymmetric synthesis. However, when a substituent other than hydrogen is present at position the obtained nitroalkanes are racemic.241 The unsaturated ketone in Figure 4-5 was used by Servi and coworkers in the production of raspberry ketone, a key component of the flavor of raspberry fruit,242 by natural means.243 This conversion also shed light on bakers yeast mediated reductions.244, 245 When the reaction was carried out in deuterated water, deuterium was found only at the -carbon whereas when specifically labeled cofactors NADPDA or NADDB were used the deuterium was attached to carbon. Deuterium was not incorporated into the product when NADPDB was used instead. Therefore hydride attack at the -carbon occurs by an A-type mechanism, for NADPH, with concomitant proton uptake from the solvent at the -position. Servi has also formulated some rules regarding the stereochemical outcomes of bakers yeast bioreductions with acyclic unsaturated compounds (Figure 4-6). 222, 226 He proposed that the stereochemistry of double bond reduction occurs with trans -hydrogen delivery when -substituted alkenes are employed (type A reaction) and the chirality of the obtained product is as depicted in Figure 4-6A. The opposite stereochemistry was proposed for -substituted alkenes (type B reaction). The question raised by Crout and coworkers246 is how we can explain those observations while taking into account the

PAGE 134

117 catalytic site of the enzyme responsible for the reaction? If we assume that the type A reaction was catalyzed by one enzyme, should we assume that the same enzyme also catalyzed B type reactions and that the compound adopted a different conformation (Figure 4-6B1)? Or we should assume that a second enzyme acted in syn mode (Figure 4-6B2) or even in trans mode but with opposite orientation from the one depicted in Figure 4-6A? Figure 4-5. Preparation of raspberry ketone by bakers yeast reduction. Figure 4-6. Servis rules for A and B type reduction in bakers yeast.

PAGE 135

118 The biohydrogenation of cyclopentenone and cyclohexenone derivatives has been reported to afford chiral building blocks useful for the syntheses of prostaglandins,247 carotenoids228 and other terpenoid compounds.248 In fact, the reduction of 3-methyl-2cyclohexenone, in the impressive scale of 30 g, by Saccharomyces cerevisiae was the first reported reduction of enone by Fischer and Weidemann, in 1935 (Figure 4-7).249 Figure 4-7. The first reported enone reductase activity in bakers yeast. Bakers yeast has also been used extensively for the production of levodione, a key intermediate in synthesis of carotenoids228 and flavor compounds.250 In contrast to routes using chemical means for this transfo rmation to control the chemo-selectivity of this reaction by utilizing different metals and varying solvents and reaction conditions251253 bakers yeast affords pure, crystalli ne levodione in 85% yield and excellent enantioselectivity (Figure 4-8).254 Figure 4-8. Bakers yeast versus chemic al mediated reduction of 4-oxo-isophorone.

PAGE 136

119 Unfortunately, moderate chemo-selectivity is a common problem when whole cells are utilized. One of the most impressive example is provided by zearalenone also known as RAL and F2 toxin. This mycotoxin is produced by some Fuasarium species when they grow on corn and other small grains in storage. It is a heat stable compound and causes infertility and other breeding problems, especially in swine. In an effort to understand better the metabolism of this compound in mammalian species, about 170 microorganisms were screened for their ability to produce different zearalenone analogues.255 Only Saccharomyces cerevisiae NRRL Y2034 was found to selectively reduce the carbon-carbon double bond without simultaneously acting on other functional groups of this molecule, as the rest of microorganisms did (Figure 4-9). Figure 4-9. The six possible microbial transformations of zearalenone. Other fungi have been also employed for the reduction of simple cyclic unsaturated compounds256-258 as well as for more exotic derivatives259, 260 (Figures 4-10 and 4-11). In addition, cyanobacteria and plant cells lines have shown notable, and in certain cases, complementary stereoselectivities as compared to bakers yeast261-264 but their applicability among organic chemists is narrow simply because complex and timeconsuming cultivation techniques are required.

PAGE 137

120 Figure 4-10. Preparation of griseofulvin, an antifungal agent. Figure 4-11. Biohydrogenation of abscisic acid, a plant stress hormone. We should recognize that bakers yeast may be the most frequently used microorganism for biohydrogenations but it is certainly not the most powerful one. The reductive capabilities of several Clostridia species are remarkable.265, 266Their ability to reduce unsaturated carboxylic acids, in a stereospecific and NAD(P)H-free manner, renders them a very attractive system (Figur e 4-12). It is not an overstatement to say that Helmut Simons laboratory has exami ned virtually every possible trisubstituted enoate as a substrate for these microorganisms.267, 268 In addition, numerous and sometimes innovative techniques for supplying cheap and by-product-free reductive equivalents have been combined with this system.269, 270 The simplest scheme comprises a cathode electrode, an electron mediator such as methyl-viologen, the unsaturated acid and the biocatalyst in a form of intact-cells, cell-extract or even purified

PAGE 138

121 enoate reductase. Unfortunately, most organic chemists do not enjoy working with electrochemical cells under strictly anaerobic conditions. The lack of oxygen is a prerequisite for these strict anaerobes to grow and retain their reductive ability. Therefore, none outside Simons laboratory have attempted to take advantage of their hidden reductive power, until now. Figure 4-12. General scheme of biohydrogenation with Clostridia species. If we compare the bakers yeast substrate specificity with that of Clostridia species it is clear that Clostridia in addition to reducing unsaturated carboxylic acids, also accepts as substrates most of the enones and enals known to react with bakers yeast. However, both microorganisms share a common disadvantage: they cannot tolerate double bulky substituents at position.

PAGE 139

122 En-Reductases: Substrate Specificity What is the color of an en-reductase? An answer to this seemingly trivial question could in fact facilitate many efforts to classify and therefore understand known and putative en-reductases. Based on our proposed categorization (Figure 4-13), an enreductase preparation may be yellow if it contains flavin and thus belongs either to the old yellow enzyme (OYE) family or to the enoate -reductase family (a distant relative to OYE family) or it may not possess any color if it belongs to medium or short chain dehydrogenase/reductase family. Figure 4-13. A proposed classification of en-reductases. Especially in the early days, most of these enzymes have been purified from their native hosts on the basis of their en-reductase activity. Only the last decade, and with the aid of bioinformatics has it become apparent that these enzymes may share common structural features and catalytic activities. We will build on these different families in order to describe the substrate specificity of en-reductases. We will also present cases in which a protein was purified because of its en-reductase activity but no ENREDUCTASES FLAVINCONTAINING FLAVIN-FREE UNCLASSIFIED Medium Chain Dehydrogenases/ Reductases Short Chain Dehydrogenases/ Reductases Old Yellow Enzyme Family Enoate Reductases

PAGE 140

123 additional effort was made for the determination of its amino-acid sequence. These orphan en-reductases may or may not belong to one of the abovementioned families and thus they will be discussed separately. Old Yellow Enzyme Family The isolation of the first biocatalyst capable of reducing activated carbon-carbon double bonds preceded271 even Fischers observation for bakers yeast reductive capabilities on the same functional group.249 Paradoxically, this particular molecular function of the so-called old yellow enzyme (OYE1) remained obscured to scientific community for about six decades since its first purification from brewers bottom yeast. Today, the old yellow enzyme family embrac es relatives found in numerous fungi, bacteria and plants.272 All members contain flavin mononucleotide (FMN) as a noncovalent prosthetic group, require NAD(P)H as cofactor and all are able, at least to some extent, of reducing functionalized alkenes. However, the cellular role of most of them is still unclear. Recent experimental evidences support defense against acrolein273 stress and protection of actin cytoskeleton in case of yeast OYEs.274 In the following discussion, we will present the substrate specificity of the OYE1 separately, whereas the rest of the purified OYEs are grouped based on their source of isolation. Old yellow enzyme from Saccharomyces carlsbergensis The detailed characterization of OYE1275 could be described as one mans accomplishment. We owe most of our knowledge today about the extraordinary and extremely divergent catalytic activity of t he OYEs from yeasts to Professor Vincent Masseys laboratory.276 -NADPH is considered to be the physiological reductant (2 electrons per flavin) for the OYE even thought the uncommon -anomer is actually slightly better reductant (Figure 4-14).277 Sodium dithionite can also be used, but in this case, the reduction

PAGE 141

124 proceeds via a formation of semiquinone inte rmediate (one electron reduced flavin), which disproportionates slowly into fully reduced and oxidized OYE. An electron mediator such as methyl viologen may accelerate the process. Only a few other reductants could be used, however, flavin r eoxidation can be achieved by a number of oxidants (Figure 4-14) with the molecular oxygen being an opportunistic one and yielding hydrogen peroxide and superoxide as products.278 The ability of the oxidized form of OYE1 to bind aromatic compounds, and especially phenols, gives rise to new, long wavelength, absorbance bands (green form of OYE) due to transfer of charge from the phenolate to flavin.279-281 This property was exploited for the purification of OYE1 by affinity chromatography on N-(4hydroxybenzoyl) aminohexyl agarose.282 Upon reduction with sodium dithionite, OYE1 has greatly diminished affinity for the phenol and hence is eluted from the column in an almost pure form. A second class of binding ligands is small molecules such as the monovalent anions: acetate, chloride and azide with the later binding tighter. (Figure 414). 277, 278 The last class of OYE1 inhibitors are pyridine nucleotide derivatives such as the acid hydrolysis products of NADPH. All three types of inhibitors compete with one other for OYE1 implicating a common binding site on the enzyme. In 1993, Stott et al.283 reported for the first time that 2-cyclohexenone could act as alternative electron acceptor for the reduced OYE1. Two observations stimulated Massey to test an enone as a substrate for OYE: i) the sequence similarity of OYE1 with a bile-acid-inducible protein in Eubacterium sp. and ii) the fact that all the reductive steps in the 7a-dehydroxylation of the bile acids in this microorganism involve the reduction of cyclohex-2-enone-like functional group. Indeed, in that work, Masseys spectrophotometric assays clearly showed that 2-cyclohexenone, similar to quinones, stopped the utilization of oxygen as electron acceptor during the reoxidation of OYE1.

PAGE 142

125 Figure 4-14. First generation of substrates and inhibitors for OYE1.

PAGE 143

126 The reduction, as it would be proved later, is stereospecific and involves a trans addition across the double bond with transfer of the proR -hydrogen of NADPH as a hydride [via the flavin N(5)] to the -carbon with a concomitant uptake of a solventderived proton at the -position (Figure 4-15).284-286 The overall process is showed to be concerted. Two years later, the same laboratory published an extensive study regarding the en-reductase potential of OYE1.287 Almost 50 compounds were tested spectrophotometrically for their ability to reoxidize OYE1 under anaerobic conditions (Figure 4-15). Several enones and enals were found to be good substrates for NADPHmediated reduction by OYE1, whereas unsaturat ed acids, esters, nitriles and amines did not react. The effect of substituents (i.e., substitution at or position and monoor disubstitution at position) on the rate of reduction was also explored and the general trend is depicted in Figure 4-17. Di-substitution at position, especially with substituents bigger than methyl, significantly reduced the rate of reaction of aliphatic compounds. According to the authors, the problem is not the binding of these compounds in the catalytic active site per se since comparable perturbation of the flavin spectrum was observed in each case, but in the oxidation rate of NADPH. In an independent study, Swiderska and Stewart used whole E. coli cells in which OYE1 was overexpressed and these authors showed a similar trend for 2-substituted cyclohexenones.288 Only the 2-methyl substituted compound was reduced completely, whereas bulkier alkyl groups (e.g. propyl, n -butyl) impeded the bioreduction. Even worse was the case when the methyl group at position 1 was replaced by an ethyl chain. Only 16% of the starting olefin (20 mg) was reduced even after prolonged incubation (i.e., 2 days). On the other hand, the recorded ee was higher than 90% in all cases. The stereochemical outcome of these biotransformations (Figure 4-19) is predictable

PAGE 144

127 and it could be derived from the crystal structure of the OYE1 with the para-hydroxy benzaldehyde (Figure 4-18).289, 290 Figure 4-15. Representation of catalytic cycle of Old Yellow Enzyme. The Stott et al. study also reported a new NADPH-independent dismutation reaction catalyzed by OYE1. In absence of any nicotinamide cofactor, 2-cyclohexenone was oxidized to phenol by reducing the OYE1 (Figure 4-20A, step 1) while in the second step, a second molecule of 2-cyclohexenone acted as an electron acceptor to reoxidize the OYE. The net result of the dismutation reaction is a 1:1 ratio phenol and cyclohexanone. Not all the substrates could undergo a dismutation reaction, however. For example 4,4 dimethyl-2-cyclohexenone is reduced in a NADPH-dependant manner by OYE1, but it could not reduce OYE1 in absence of cofactor. One explanation could be the unfavorable redox potential for such enones (i.e. a disubstituted at the same position cyclohexenone is not aromatizable after oxidation).

PAGE 145

128 Figure 4-16. Second generation of substrates for OYE1 (Numbers in parentheses indicates relative activity based on the measured turnover number for acrolein 170 min-1). Most reactive compounds are highlighted by red and less reactive by grey color. Compounds that did not react are not highlighted.

PAGE 146

129 Figure 4-17. Relative reactivity of substituted en-ones (als) based on Masseys studies. Figure 4-18. Stereoview of active site architecture of OYE1 with para-hydroxybenzaldehyde (1OYB). Picture was rendered in Pymol.291 Figure 4-19. Stereochemical outcome of reduction of substituted cyclic enones by OYE1.

PAGE 147

130 Using a stereospecifically labeled substrate, it was demonstrated that the second step of the dismutation reaction proceeds with a way similar to the one observed for the normal NADPH reduction whereas trans elimination was observed for the first step (Figure 4-20C) In order for the same substrate (e.g. ODE) to be able to reduce and then oxidize the flavin, a different accommodation above the si face must be assumed for the reductive ( Figure 4-21A) and the oxidative steps ( Figure 4-21B), respectively.292 Also, under aerobic conditions and for bulkier than 2-cyclohexenone substrates (e.g. ODE and ODEC), molecular oxygen may compete for reoxidation of OYE1 (Figure 4-20B and C). Most of steroids tested in that work, even though are bound strongly to OYE, did not undergo either an oxidation or an NADPH-dependent reduction by OYE1. The only exceptions were 1,4 adrostadiene,3,17-dione (Figure 4-22) and 19-nor-testosterone (Figure 4-22). Figure 4-20. First generation of desaturation reactions by OYE1 in absence of NADPH. (ODE, 3-oxodecalin-4-ene; ODEC 3-oxodecalin-4-ene-10-carboxaldehyde; HTN, 3-hydroxy-6,7,8,9-tetrahydronaphthalene)

PAGE 148

131 Figure 4-21. Proposed orientation of ODE above the si face of flavin during: the reductive (A) and the oxidative half (B) of the dismutation reaction. Figure 4-22. Second generation of binding ligands for OYE. Murthy et al. took advantage of the desaturase activity of the OYE1 to kinetically resolve recemic ketones and aldehydes (Figure 4-23).293 While this a nice example as a proof of principle, it is unlikely to find broader applications for three main reasons: i) the redox potential of flavin in OYE1 must become significantly less negative294-295 (this was accomplished by replacing the native FMN by a synthetic flavin-analog)296 ii) the yield of the reaction could not exceed the 50% and iii) the subsequent separation of the saturated chiral product from the corresponding enone may be difficult due to high similarity of both compounds.

PAGE 149

132 Figure 4-23. Second generation of desaturation reactions by modified-OYE in absence of NADPH. Nitro-alkenes could also be reduced easily by OYE1 (Figure 4-24A). Similar to a carbonyl group, a nitro moiety strongly polarizes a carbon carbon double bond, facilitating thus the hydride attack at the -carbon and enhancing the acidity of the proton at position.297 Lastly, OYE1 also shares a catalytic activity that has been mostly explored in OYEs from bacteria: reducing highly explosive organic nitrates such as GNT (Figure 424B). Massey and coworkers observed that the reductive cleavage of nitro-esters is also mediated by reduced free-flavin, but the ratio of the obtained products is different

PAGE 150

133 compared to that produced by the enzym atic reduction. OYE1 showed strong preference for the primary nitrates.298 Figure 4-24. Third and fourth generation of substrates for OYE1. Old yellow enzymes from yeasts Other OYEs from different yeast strains are listed in Table 4-1. OYE2 was detected in the bakers yeast genome after screening an S. cerevisiae DNA library using the OYE1 gene as a probe.283 OYE3 was discovered unexpectedly when a OYE2 yeast strain was found to retain OYE activity.299 Oye, initially named keto-isophorone (KIP) reductase, was first purified from its native host after a preliminary screening of different fungal species for the ability to reduce 4-oxoisophorone stereoselectively.300 Estrogen binding protein was initially purified from 200 g of C. albicans cells, the most common fungal pathogen for humans.391 Interestingly, the amino acid sequence showed no similarity to human estrogen receptor, as was initially suspected. Instead,

PAGE 151

134 the 46% amino acid identity with OYE2 forced the researchers to investigate a possible OYE activity. EBT1 not only reduces common OYE1 substrates and possesses desaturase activity but can also reduce 19-nor-testosterone in an NADPH-dependent manner whereas OYE1 can only aromatize the same compound.302 The methylotrpophic yeast Hansenula polymorpha overexpression of the HYE gene cluster confers resistance towards high concentrations of allyl alcohol in presence of alcohol oxidase (AO).303 The observation suggested that the AOs product (i.e. acrolein) is subsequently reduced by HYEs, t hereby diminishing thus its otherwise high toxicity. The purification of N-Ethylmaleimide Reductase (NEM) from Yarrowia lipolytica was the result of researchers failure to purify a similar protein from E. coli.304 In 1979, Mizugaki and co-workers reported the isolation of an enzymatic fraction able to reduce cis -2-alkenoyl-Coenzyme A to the corresponding saturated acyl-CoA derivatives in the presence of NADPH.305 During the course of inhibition studies on the same protein preparation, they found that including N-ethylmaleimide (a common binding reagent for thiol groups of cysteines) in the NADPH assay did not inhibit the reaction but instead accelerated NADPH reduction. The NEM reductase co-eluted in the same fraction with cis -2-enoyl-CoA reductase on Sephacryl S-200 Superfine gel-filtration column. These unsuccessful efforts to separate the two proteins led to a revised strategy of isolating a protein with similar catalytic activity from another host.306 Even though neither the amino acid sequence nor the UV-VIs spectrum of the protein was reported, this protein most likely belongs to OYE family. A BLAST search using OYE1 as a probe reveals six putative OYEs in Yarrowia lipolytica

PAGE 152

135 Table 4-1. List of OYEs from yeasts and corresponding substrate specificity. Year (Ref.) Organism Protein (Swiss-Prot accession no.) Gene Substrates Binding Ligands, inhibitors 1994 (301) Candida albicans Estrogen binding protein (P43084) EBP1 1993 (283) 1995 (299) Saccharomyces serevisiae Old yellow enzyme 2 Old yellow enzyme 3 (Q03558, P41816) OYE2 (YHR179W) OYE3 (YPL171C) 2002 (303) Hansenula polymorpha Hansenula yellow enzyme 1, 2, 3 (Q8J293, Q8J292, Q8J291) HYE1 HYE2 HYE3 2004 (309) Kluyveromyces lactis Kluveromyces yellow enzyme (P40952) KYE1 1997a (306) Yarrowia (Candida) lipolytica N-Ethylmaleimide Reductase of Y. lipolytica Unknown

PAGE 153

136 Table 4-1. Continued. Year (ref.) Organism Protein (Swiss-Prot accession no.) Gene Substrates 2002b (300) Candida macedoniensis ( Kluyveromyces marxianus) Old yellow enzyme (Q6I7B7) oye a Numbers in parentheses indica te relative activity based on the measured Vmax number for Nethylmaleimide 216 nmol min-1 mg-1. b Numbers in parentheses indicate relative activity based on spectrophotometric assays. In terms of biotechnological applications, the examples are limited. Oye, OYE2 and OYE3, have been cloned and overexpressed in E. coli and their activity towards the reduction of ketoisophore was evaluated. All of them produce (6 R )-levodione in excellent enantiomeric purity.307 Moreover, when both oye and glucose dehydrogenase were co-expressed in E. coli 100 times higher productivity was achieved, compared to the one reported with bakers yeast cells.308 In addition, when a second enzyme (a ketoreductase) was added to the reaction mixture after the completion of the first reduction step, the doubly chiral compound R -actinol was obtained in 94% ee Mergler et al. demonstrated a technique with which OYEs could be used as biological filters.309 KYE was displayed on the surface of an engineered Pichia pastoris cell. The novel strain was able to bind bisphenol A, a monomer used in industry for the production of polystyrene and polycarbonates and a known endocrine disruptor (i.e., it competes with estradiol for binding with estrogen receptors). Old yellow enzymes from bacteria In 1979, old yellow enzyme activity was detected in bacterium Gluconobacter suboxydans At that time the enone reductase activity of OYEs had not yet been recognized.310

PAGE 154

137 In 1994, Bruces laboratories painstaking efforts to elucidate the degradation pathway of morphine alkaloids in Pseudomonas species lead to the isolation of morphinone reductase from Pseudomonas putida M10.311, 312 The purified enzyme reduced the olefin bonds of both morphinone and codeinone. The products obtained, hydromorphone and hydrocodone, not easy to synthesize by chemical means and are important pharmaceutical drugs with useful analgesic and antitussive properties.313 Masseys 1993 report did not pass unnoticed in the Bruce lab, who demonstrated the ability of a bacterial old yellow enzyme (morB) to reduce 2-cycloxehenone in an NADHdependent manner for the fist time. Progesterone and cortisone bind to morphinone reductase but do not react. OYEs have been also isolated from other Pseudomonas species (Table 4-2) and other bacteria such as Agrobacterium radiobacter and Enterobacter cloacae species, albeit not for their action on alkaloids. Instead, it is their reductive power on organic nitrate esters that has been explored more while their enone reductase activity has been examined only superficially. It is beyond the scope of this introduction to discuss in detail the denitration activity of OYEs. Of particular importance is their ability to degrade the aromatic explosive 2,4,6-trinitrotoluene (TNT), persistent pollutant in military sites.314, 315 Since the first old yellow enzyme (pentaerythritol tetranitrate reductase) was expressed in transgenic tobacco plants and its bioremediation capabilities was demonstrated,316 the exact mechanism of this reaction is under vigorous investigation.317-323 The NAD(P)H dependent 2-cyclohexen-1-one reductase from phytopathogenic bacterium Pseudomonas syrigae pv. Glycinea was found to be preferentially present in soluble cellular protein fraction of this microorganism when the latter was cultivated at low temperatures (i.e 18 C). The protein spot was isolated from a two-dimensional electrophoresis gel and the N-terminal amino acid sequence showed similarity to

PAGE 155

138 morphinone reductase. Interestingly, the 3-methyl-2-cyclohexe-1-one significantly inhibit the activity of the purified as Ncr-His6 reductase while norsteroid compounds such progesterone and cortisone were not bound. Table 4-2. List of OYEs from bacteria and corresponding substrate specificity. Year (Ref.) Organism Protein (Swiss-Prot accession no.) Gene Substrates Inhibitors (Binding Ligands) 1979 (310) Gluconobacter suboxydans old yellow enzyme unknown O2 1994 (311) Pseudomonas putida M10 morphinone reductase (Q51990) morB 1999a (324) Pseudomonas putida II-B nitroester reductase (Q9R9V9) XenA 1999b (324) Pseudomonas fluorescens I-C nitroester reductase (Q9RPM1) XenB

PAGE 156

139 Table 4-2. Continued. Year (Ref.) Organism Protein (Swiss-Prot accession no.) Gene Substrates Inhibitors 2004 (325) Agrobacterium radiobacter glycerol trinitrate reductase (O31246) Ner 1996 (326) Enterobacter cloacae Pb2 Pentaerythritol tetranitrate reductase (P71278) Orn 2003c (327) Bacillus subtilis Escherichia coli JM109 YqjM (Q2B6D5 B14911_08447

PAGE 157

140 Table 4-2. Continued. Organism Protein (Swiss-Prot accession no.) Gene Substrates Inhibitor 1999 (329) Escherichia coli JM109 N-ethylmaleimide reductase (P77258) nemA 2005d (328) Shewanella oneidensis SYE1,SYE3, SYE4 AAN55488.1, AAN57126.1, AAN56390.1 a, b Numbers in parentheses indicate relative activity based on the measured specific activity (U/mg) for the enzyme with nitroglycer in (NG) as substrate (6.1 U/mg for XenA and 6.8 U/mg for XenB). c Numbers in parentheses indicate relative activity based on the measured kcat/KM number for the reaction N-ethylmaleimide with YqjM (8.07 M-1 sec1). d Numbers in parentheses indicate re lative activity based on the measured kcat/KM number for the reaction N-ethylmaleimide with SYE4 (5.96 M-1 sec-1) In 2003, the first gene encoding an OYE from a gram-positive bacterium ( Bacillus subtilis ) was cloned and the protein was overexpressed and isolated from E. coli in a particularly high yield (161 mg/L culture).327 In addition to some limited spectrophotometric studies regarding its substrate specificity, it was found that the presence of xenobiotics such as TNT or agents contributing to oxidative stress such as hydrogen peroxide led to rapid induction of YqjM in B. subtilis The observation further supports the notion that the physiological role of OYEs is for detoxification.

PAGE 158

141 The first comparative study among OYEs in the same bacterium was reported for Shewanella oneidensis .328 Out of the four putative OYEs in this species, three were successfully overexpressed in E. coli and purified in a GST-tagged and untagged form. Their relative activity towards four different substrates is summarized in the last entry of Table 4-2 for the untagged form. The corresponding GST-SYEs showed similar pattern but the kcat/ KM values were lower in most of the cases. A more comprehensive comparative study am ong OYEs from five different species had been earlier published by Bruce and coworkers.329 The results from this study are summarized in Table 4-3. The numbers are normalized based on the specific activity observed for N-ethylmaleimide reductase (NemA) from E. coli with glycerol trinitrate (GNT) as substrate (i.e. 37.9 U mg-1). Since the spectrophotometric assays were carried out under aerobic condition the background oxidase activity was determined and subtracted. The substrate preference for each enzyme is highlighted. Table 4-3. A comparative study among di fferent OYEs from yeast and bacteria. OYE from GNT PETN TNT 2-cyclo hexenone trans-2hexenal 1-nirtro cyclohe xene 2-nitro benzalde hyde Codein one E. coli 100% 84% 0% 12% 7% 22% 14% 0% E. cloacae (type strain) 94% 59% 1% 7.1% 5.8% 20% 14% 0.3% E. cloacae Pb2 52% 56% 4% 2% 4% 4% 5% 0.3% P. putida M10 -1% -1% 2.6%-1% -1% 1% 1% 36% S. carlsbergensis 1% -1% 8% 13% 11% 13% 8% -0.3% Old yellow enzymes from plants In the early 80s, the elucidation of octadecanoid biosynthesis in several plant tissues by Vick and Zimmerman led to the purification of the so-called 12oxophytodienoic acid reductase (OPR) from corn.330 OPR catalyzes the forth step in the jasmonic acid biosynthetic pathway starting from linolenic acid (Figure 4-25), reducing

PAGE 159

142 the endocyclic double bond of 12-oxo-10,15(Z)-phytodienoic acid (OPDA) to produce 3oxo-2-(2(Z)-pentenyl)-cyclopentane-1-octanoic acid (OPC-8:0).331 Figure 4-25. Jasmonic acid biosynthesis. Ten years later, Schaller and Weiler purified a second homologue from Corydalis sempervirens after examination of the OPR activity in five different plants.332 They wisely recognized that if the physiological role of OPR is the reduction of OPDA then their enzymatic preparation should preferably reduce the cis and not the trans diastereoisomer since the former and not the latter is found naturally from the allene oxidase/cyclase. Their GC/MS results were encouraging (6:1 preference for the cis

PAGE 160

143 isomer), but not conclusive. Unexpectedly, their additional experiments showed that this enzymatic preparation, known today as OPRI, as well as another homologue from Arabidopsis thaliana (OPR1) reduced the (9 R ,13 R )cis diastereoisomer and not (9 S ,13 S )cis diastereoisomer, which is however the natural product.333, 334 The validity of their initial hypothesis was reestablished in the subsequent publication in which a second isoenzyme (OPRII) was purified and showed a slight preference for the naturally occurred (9 S ,13 S )cis -OPDA diastereoisomer.335, 336 Interestingly the OYE from S. cerevisiae also showed preference for the same diastereoisomer. Today, the dominant trend regarding the substrate specificity of the OPRs is that they should be divided into two subgroups.337 Subgroup I consists of AtOPR1, AtOPR2, LeOPR1 (from tomato) and OsOPR1 (from rice) with uncertain biological function and subgroup II consists of AtOPR3 and LeOPR3 whose physiological activity is the reduction of the (9 S ,13 S )cis OPDA (Table 4-4). Table 4-4. List of OYEs from plants and corresponding substrate specificity. Year (Ref.) Organism Protein (Swiss-Prot accession no.) Gene Substrates Inhibitors 1986 (330) Zea mays (corn) 2-oxo-phytodienoic acid reductase not determined 1997 (332) Corydalis sempervirens OPRI not determined 1997 (333) Arabidopsis thaliana 12-oxophytodienoate reductase 1, (Q8LAH7) OPR1

PAGE 161

144 Table 4-4. Continued. Year (Ref.) Organism Protein (Swiss-Prot accession no.) Gene Substrate Inhibitors 2000 (336) Arabidopsis thaliana 12-oxophytodienoate reductase 2, (Q8GYB8) OPR2 2000 (336) Arabidopsis thaliana 12-oxophytodienoate reductase 3, (Q9FUP0) OPR3 2003 (337) Oryza sativa L. (rice) 12-oxophytodienoic acid reductase, (Q84QK0) opda 1999 (339) Lecopersicon esculentum cv. Castlemart II (tomato) 12-oxophytodienoate reductase 1, LeOPR1 (Q9XG54) OPR1

PAGE 162

145 Table 4-4. Continued. Year (Ref.) Organism Protein (Swiss-Prot accession no.) Gene Substrate 2003 (338) Pisum sativum (garden pea) PsOPR1-6 (Q76FS1, Q9AVK9, Q76FS0, Q76FR8, Q76FR9, Q76FR7) PsOPR1, OPDRA, PsOPR3-6 LeOPR1 is the only OYE in this category which is known to accept compounds other than 2-cyclohexenone and OPDA.339 In fact, the highest catalytic activity was observed for N-ethylmaleimide. It is interesting that activity was also recorded for maleic acid but not for fumaric acid. Matsui et al. classified the six OPR-like enzymes from pea into four groups based on their ability in r eduction of 2-cyclohexenone.338 They observed no catalytic activity for PsOPR5, little for PsOPR3, moderate activi ty for PsOPR1, 4 and 6 and highest activity for PsOPR2. To conclude, we should mention that almost all the experiments regarding the substrate specificity of OYEs have been performed only spectrophotometrically. In addition, despite the large number of isolated OYEs from bacteria and plants and the numerous biophysical studies of their structure and function, little is known about their substrate specificity. Enoate Reductases Enoate reductases (ERs) belong to a rare class of flavoenzymes containing both FMN and flavin adenine dinucleotide (FAD).340 In addition, four iron and four labile sulfur atoms are present per enzyme subunit. The mechanism of this class of enzymes is not as well studied as for the OYEs but EPR studies341 have shown that electrons from NADH flow via FAD and [4Fe-4S] cluster to FMN cofactor. The FMN domain is very

PAGE 163

146 similar to OYE structure and in that respect enoate reductases are considered distant relatives to OYEs. These multi-domain, gigantic (i.e.~ 940 kDa), enzymes have the unique ability to catalyze the reduction of non-activated 2-enoates, in contrast to what is observed for the enoyl-CoA and 2, 4-dienoyl-CoA reductases.342, 343 For the latter, conversion of the carboxyl group to the corresponding CoA ester is a prerequisite for the saturation of the carbon-carbon double bond to take place (Figure 4-26). 342, 344, 345 Figure 4-26. Biohydrogenation of unsaturated carboxylic acids from three different types of enzymes. In 1975, Simon and coworkers observed that the hydrogenation of ( E )-2-methylbutenoate by some Clostridia species could occur even when the acid was not converted to corresponding CoA ester.346 The key for this observation was the different stereochemical outcome of this reduction (2 R ) compared to what had been previously reported for the action of butyryl-CoA reductase on the corresponding CoA ester (2 S ).347 The enzyme responsible for the former transformation would be called 2-enoate reductase.

PAGE 164

147 Interestingly, in the first 2-enoate reductase preparation from Clostridium kluyveri the FMN cofactor was not found but catalytic activity was detected for the compounds listed in Table 4-5.353 A shorter enzyme purification protocol developed in the same laboratory allowed the isolation of the most studied enoate reductase from Clostridium tyrobutiricum .348 In that preparation, the 0.6-0.7 mol of FMN per mol of subunit found indicated labile binding of the FMN cofactor in these flavoproteins. In addition, the rapid deactivation of the catalyst in the presence of oxygen (1-2 min for the reduced form) renders its purification quite laborious since strictly anaerobic conditions are required in all steps. Attempts to express ER from Clostridium tyrobutiricum in E. coli have failed so far. However, a third ER from Clostridium thermoaceticum was successfully overexpressed in E. coli but only when the engineered strain grew under anaerobic conditions.341 Figure 4-27. Reductive part of the phenylalanine fermentation pathway in Clostridia sporogenes Glu, glutamate; 2-OG, 2-oxoglutarate (Adapted from ref. 350). The physiological role for the ER from Clostridium sporogenes has been elucidated by Simon and coworkers.349 It was found that certain Clostridia species ferment L-phenylalanine to phenylacetate and 3-phenylpropionate. Enoate reductase

PAGE 165

148 catalyzed the last step in the formation of the latter compound (Figure 4-27). However, the substrate specificity for this particular reductase is narrow and limited to cinnamic acid. In contrast, ERs from Clostridium tyrobutiricum and Clostridium kluyveri could accept a broad range of enoates. A complete list of substrates is not included in Table 4-5 because in most cases whole cells or crude extracts have been employed and the subject has been reviewed elsewhere.267 General rules derived from biohydrogenations with whole cells, mentioned in the first part, regarding the substrate specificity (no bulky disubstitution at the -position and different stereoisomers depending on the initial enoate E or Z stereochemistry) are also valid for the purified ERs. A complication, however, has been observed in the case of geraniate (Table 4-5). Hydrogenation of ( E ) and ( Z ) geraniate afforded 95% ee ( R )and 95% ee ( S )-citronellate respectively, when purified ER was employed. When whole cells were used instead, a lower ee value (60-85%) was observed for the Z geraniate presumably due to isomerization of the ( Z )-isomer to thermodynamically more stable ( E )geraniate by the cells.267 Enals could also be reduced easily by ERs. However, caution should be taken to remove simultaneously the obtained aldehyde product quickly from the reaction mixture, either by further enzymatic reduction to corresponding alcohol or by continuous extraction with an organic solvent, otherwise low ee values are observed.351 It has been shown that this is due to a desaturation reaction catalyzed by ERs when aldehydes, but not when carboxylates are used, in the presence of an electon acceptor such as oxygen. Surprisingly under those conditions, the half life of ER was more than 20 h. Lastly halogenated enoates undergo an elimination reaction by ERs affording saturated, halogen-free carboxylates with concomitant consumption of 2 equivalents of NADH.268

PAGE 166

149 Table 4-5. List of ERs and representative substrate specificity. Year (Ref.) Organism Protein(Swiss-Prot accession no.) Gene Substrates 1980 (352) Clostridium kluyveri 2enoate reductase 2-enoate reductase (O52933) enr 1979 (353) Clostridium tyrobutiricum ( Clostridium spec. La1) 2-enoate reductase (Q52922) enr 2000 (350) Clostridium sporogenes 2-enoate reductase unknown 2001 (344) Clostridium thermoaceticum 2-enoate reductase (O52935) enr Structure-Function Studies among Flavin-Containing En-Reductases The extensive biochemical and biophysical analyses of old yellow enzyme family members aim mostly to shed light on the catalytic microenvironment (Figure 4-28) of these enzymes as well as to pinpoint subtle structural differences among them.

PAGE 167

150 Figure 4-28. Close up view of the catalyti c microenvironment of OYE1 before (1OYA) and after (1OYB) the binding of para-hydroxy-benzaldehyde.290 Amino acid residues in distance equal of smaller of 4 around the binding ligand are shown. Amino acid residues for which conformational changes are imposed before and after the binding of the ligand (i.e Phe250, Phe296, and Tyr375) are colored in blue and magenta respectively. The figure was rendered in PyMOL. Tyrosine 196 was the first amino acid residue that was targeted by site-directed mutagenesis.284 The observation that the Y196F mutation did not alter the ability of S. carlsbergensis OYE1 to bind phenol analogs suggested that no significant structural changes occurred in the catalytic site upon this mutation. However, the nearly complete loss of catalysis for the oxidative-half reaction step, with 2-cyclohexenone as a substrate, was consistent with the role of Tyr196 as an active site general acid in the reduction of unsaturated ketones. By contrast, the observation that the rate of enzyme oxidation was not significantly altered w hen nitrocyclohexene was used as a substrate by the Y196F mutant suggested that hydride transfer and substrate protonation are decoupled events for this type of compound. Further investigation of the OYE1mediated reduction of nitro-alkenes revealed that the existence of Tyr196 may not be

PAGE 168

151 necessary for the reduction of nitrocyclohexene (i.e formation of nitronate intermediate) but it still accelerates the protonation of the rapidly released into the medium intermediate nitronate. In case of Y196F mutant, the reaction was found to be a slow non-enzymatic process.297 Interestingly, in the same study it was found that OYE1 catalyzed the non-redox deprotonation of nitrocyclohexane (pKa=8.65) forming a nitronate, the same intermediate formed as in the redox reaction, presumably via the Tyr196 side chain (pka=9.1). A tyrosine residue is present in an analogous position in all well-studied isoenzymes of OYE1, with the only exception being the morphinone reductase, in which Cys191 aligns with Tyr196. Surprisingly, mutagenesis studies (i.e. C191A in the case of MR and Y186F in the case of PETN reductase) argued against a similar role for either Cys191 or Tyr186 as an active general site acid in this OYE homolog.354 A water molecule has been suggested as the most likely proton donor in these enzymes.355 In addition, the protonation of the nitronate intermediate during the MR-mediated reduction of nitrocyclohexene is not an enzymatic reaction even though this nitroalkene is a good substrate for MR. In the crystal structure of OYE1 with para -hydroxybenzaldehyde, His191 and Asn194 are supposed to make hydrogen bonds with the hydroxy moiety of the ligand. Replacement of His191 with Asn reduced the ability of OYE1 to bind ligands or substrates whereas mutation of Asn194 to His resulted in a protein that did not bind FMN tightly. Surprisingly, a double O YE1 mutant H191N/N194H was reduced by NADPH much faster than the native enzyme.356 This result indicates that OYE is not optimized for the reductive half reaction alone. In effort to alter the redox potential of the FNM by protein engineering, Thr37 was mutated to Ala.357 This replacement lowered the redox potential of the bound FMN by 33mV a nd the rate of the oxidative half reaction

PAGE 169

152 was accelerated. However, the reductive-half reaction was slowed down and thus the overall catalytic turnover number was decreased. Mutagenesis of Trp102 (the corresponding Trp116 in Figure 4-28) to either Tyr or Phe in PETN reductase was designed to minimize steric clashes between the side chain of this amino acid and the C6 nitro group of picric acid and this resulted in tighter binding of the substrate by the mutant enzymes. The mutations altered also the chemical nature of products obtained upon the reduction of PETN reductase with picric acid.358 The X-ray crystal structure of B. subtilis YqjM has recently been solved.359 The enzyme is a homotetramer in contrast to other OYE homologues which are monomeric (OPR1, SEY1, MR) or dimeric (OYE1, PETN). In addition, even though the superposition of the two crystal structures of YqjM and OYE1 with bound para hydroxybenzaldehyde (p-HBA) places both flavin and ligand at almost the same position in both cases, only two amino acid residues (Tyr196 and His191, numbers are based on OYE1) are conserved (Figure 4-29). The most dramatic changes are the replacement of the amino acid residue Trp116 by Ala (not shown in Figure 4-29) and Thr37 by Cys26. Both Trp and Thr are well conserved aminoacids among other members of OYE family (Figure 4-30). The common replacement, Asn194 by His, found in other OYE members is also present in YqjM structure. The founding that an Arg residue from the C terminus of the second monomer participates in the catalytic environment of the first monomer may explain way a C terminus 6xHis-tagged YqjM is an almost completely inactive enzyme.360 Last but not least, the aldehyde moiety of pHBA is rotated 180 in YqjM structure in order to be placed in appropriate position for a hydrogen bond with Tyr28, since an analogous toTyr375 residue of OYE1 is absent in the YqjM structure.

PAGE 170

153 Figure 4-29. Superposition of OYE1 and YqjM structure. The flavin molecule as well as the p-HBA colored in yellow belongs to OYE1 structure. In blue are colored both flavin and p-HBA molecules belonging to YqjM structure. The Figure was rendered in PyMOL. Structure-function correlations in between OPR1 and OPR3 subfamilies are under vigorous investigations in order to explain their different substrate stereospecificities. When the X-ray structure of Leopr1 with bound 9 R ,13 R OPDA361 was compared to that of S. carlsbergensis OYE1, it became apparent that t he catalytically important amino acids are highly structurally conserved. It was proposed for the first time that amino acid residues in longer distance from the catalytic center should probably interact with the non-reactive part of the substrates and cause the observed altered stereoselectivities. Specifically, attention was paid to the two loops surrounding the entrance of the OYE catalytic cavity (Figure 4-31). Both loops showed pronounced flexibility. In fact, certain amino acid residues in loop6 are not well defined in the Leopr1 structure (i.e., black dots in Figure 4-31) as well as in the subsequently released X-ray structures of AtOPR1362 and AtOPR3.363

PAGE 171

154 Figure 4-30. Multiple alignment of twelve en-reductases belonging to OYE family. Black dots indicate amino acid residues present in the catalytic site of OYE1 and conserved among other OYE homologs. The replacement of T37 by C26 and W116 with A104 in case of YqjM as well as the replacement of Y196 by C191 in case of PETN reductase, discussed in the text, are highlighted in green boxes. The red box shows the replacement of N194 (OYE1) with His in some OYEs. The alignment was generated using the ClustalW web server software.

PAGE 172

155 Figure 4-31. Superposition of OYE1 and Leopr1 structure. A. Ribbon structure of OYE1 (blue) and Leop1 (purple). Arrows indicate loop3 and loop6 in Leopr1 and OYE1 structure respectively. B. Cartoon of superimposed structures. Arrow indicates the entrance of the catalytic center in both enzymes. Black dots indicate the disorder of residues 283-288 in loop6 of Leopr1 structure. The Figure was rendered in PyMOL. The accumulated X-ray structure data of OYEs from yeast, bacteria and plants clearly showed that the motif of loop3 is very similar in both bacterial and plant OYEs and consists of one parallel and one antiparallel beta strand (Figure 4-32). However, only the Leopr1 structure is the loop pointed toward the catalytic center. In the OYE1 structure, the loop3 is composed of two helices and two sheets and it does not capping the catalytic cavity. The crystallization of LeOPR3 uncovered an unprecedented property among flavoproteins.364 Unexpectedly, LeOPR3 crystallized as a self-inhibited dimer. The loop6 of the one monomer was protruded into the catalytic cavity of the second unit and vice versa (Figure 4-33). Mutagenesis studies revealed that replacement of Glu291, which is found at the tip of loop6, by Lys prevented the dimerization. The dimerization was also detected in solution and lead to slower catalytic turnover numbers. The cocrystallization of the dimer with two sulfate ions (i.e., red dots in Figure 4-33) at a hydrogen bond distance with Tyr364 suggested a possible regulation of catalytic activity

PAGE 173

156 in vivo by dimerization modulated by reversible phosphorylation. The observation that the Y364F mutant Leopr3 could not form a dimer supports this hypothesis. Figure 4-32. Close-up view of loop3 after superposition of five OYE crystal structures. Orange for OYE1 (1OYEA), black for Leopr1 (1ICP), green for AtOPR3 (1Q45), cyan for MR (1GWJ) and blue for SEY1 (2GOU).365 The Figure was rendered in PyMOL. Despite the accumulated knowledge regarding the structural details of OYE family members, the exact amino acid residues responsible for the differential stereoselectivities among OPRs have not yet identified. Further experimentation with compounds similar to OPDA as well as co mparative studies among OYEs from fungi, bacteria and plants may provide useful and complementary data to augment crystallographic studies. A preliminary, low resolution, crystal structure of the enoate reductase340 from Clostridium tyrobutyricum revealed that the overall fold resembles that of E. coli 2,4dienoyl-CoA Reductase (DCR).366 It seems that these two flavoproteins have evolved so that the reductive and oxidative half reaction take place in two distinct domains, whereas a 4Fe-4S cluster plays the role of an electron mediator across the two centers. In

PAGE 174

157 contrast, in the OYE family, both reduction and oxidation are carried out by the same FMN-bearing catalytic center (Figure 4-34). Figure 4-33. Birds view of the overall structure of the self-inhibited Leopr3 dimer. Arrows indicate the direction of loop6 in each monomer. Glu291 is shown in red (space filled model) from the left-hand unit is stacking above the flavin (colored in orange) of the right-hand unit blocking the catalytic center. Red dots highlight the two sulfate ions in the proximity of Tyr364, colored in blue, in the right-hand unit only. The Figure was rendered in PyMOL. Even though DCR and OYEs have been co-crystallized with their natural substrates or an analog, this is not the case for ER. In principle, DCR, OYEs, and ERs perform the same chemistry (reduction of a carbon-carbon double bond) in a similar TIM barrel domain; however, the substrate specificities are unique: enones/enals for OYE1, enoates and enones/enals for ER and CoA dienoyl esters for DCR) differs. Thus, determining which amino acids participate in substrate binding and catalysis for ER as well as additional experimentation with alternative substrates for OYEs and DCR will

PAGE 175

158 help us understand the catalytic features controlling the different substrate specificities among these flavoproteins. Such studies may also serve as a platform for subsequent protein engineering projects. Figure 4-34. Structural differences in between the flavor-bearing en-reductases. A. Cartoon of multi-domain structure of enoate reductase. B. Cartoon of OYE family single domain. Medium Chain Dehydrogenases Possessing En-Reductase Activity Alkenal/one oxidoreductase (AOR) compromises by itself a distinct family in the medium chain dehydrogenase/reductase (MDR) superfamily, called the LTD family367, 368 It is a key player in the inactivation of eico sanoids and can act either as an allylic alcohol dehydrogenase (leukotriene B4 12-hydroxydehydrogenase) or as an enone reductase (15-oxo-prostaglandin 13-reductase) on major endogenous lipids mediators (Figure 435) such as leukotriene B4 (LTB4), lipoxin A4 (LXA4) and prostaglandin E (PGE2).369

PAGE 176

159 Figure 4-35. The bifunctional action of AOR. AOR catalyzes the first step in degradation metabolism of the highly biologically active compounds LTB4, LXA4 and PGE2. AOR is found in a number of mammalian species and in various tissues. It was first isolated by Yokomizo et al. from the cytosolic fraction of porcine kidney as LTB4 12hydroxydehydrogenase (LTB4DH).370 It showed 3.5 times higher activity for 6trans -LTB4 compared to the one for LTB4, whereas its activity towards the 6trans -12epi -LTB4 was four times lower. The cDNA for the human enzyme was also cloned and overexpressed in E. coli from the same laboratory and showed 84.7% identity at the nucleotide level with that from pig.371 In an independent study, Tai and coworkers purified a 15-oxoprostaglandin reductase (PGR) from pig lung.372 Surprisingly, sequence homology analysis showed that the novel reductase differed from the LTB4DH only in one amino acid. Moreover, the reductase function of PGR on 15-oxo PGE2 was 300-fold higher compared to the dehydrogenase function on LTB4. Subsequent work uncovered guinea pig PGRs role on inactivation of lipoxin A4. The ternary complex structure of AOR for guinea pig with NADP+ and the -chain of 15-oxo-PGE2 (Figure 4-36) provided some insights for the possible mechanism of AOR as reductase (Figure 4-37).373 On the other hand the mechanistic action of the same enzyme as LTB4 dehydrogenase remains unclear.

PAGE 177

160 The significant sequence homology of AO R with the quinone reductase from E. coli374 forced Kensler and co-workers to postulate that the former enzyme possesses an additional role in chemoprotection by degrading toxic by-products of lipid peroxidation such as enones and enals.375, 376 They evaluated their hypothesis by measuring spectrophotometrically the catalytic activity of recombinant AOR from rat with a number of enones and enals. They found that enones are better substrates for AOR as compared to enals, especially when they bear a long aliphatic chain (ca. 8-10 carbon atoms, Table 4-6). Substitution at either or -carbons of enones or enals was not tolerated by the enzyme, nor could it reduce endocyclic double bonds. Interestingly, the sequiterpenes illudin S and illudin M, as well as semisynthetic derivatives such as acylfulvene, which all well-known alkylating agents and promising chemotherapeutic agents, are bioactivated by AOR.377, 378 In plants, the P1-crystallin (P1ZCr) quinone oxidoreductase from Arabidopsis thaliana ,379 an oxidative-stess induced enzym e, and the alkenal dehydrogenase from barley possess substrate specificities similar to AOR (Table 4-6).380 Due to this function, the alternative term NADPH: 2-alkenal/one hydrogenase (ALH) has been proposed to describe of this family of enzymes.381 In addition, the pulegone reductase from peppermint382 and the quinone oxidoreductase from strawberry383 were found to reduce the exocyclic double bond of (+)-pulegone and 4-hydroxy-5-methyl-2methylene-3(2H)-furanone (HMMF), respectively (Table 4-6). It is interesting to note that pulegone reductase lacks strict facial stereoselectivity in the reduction of pulegone and affords a mixture of (-)-methone and (+)-isomethone in a 55:45. The catalytic activity of both enzymes on other enones has not been explored so far.

PAGE 178

161 Figure 4-36. Close up view of the hydrophobic pocket of AOR from guinea pig (PDB, 1V3V). The chain of 15-oxo-PGE2 bearing the double bond is colored in green. NADP+ is colored in magenta. The three conserved Tyr residues (see Figure 4-38) are colored in blue. The Tyr residue 271, colored in purple, is replaced by Thr in rat and porcine AOR (See Figure 4-38). Figure 4-37. Proposed catalytic mechanism of the reductase activity of AOR from guinea pig. The enolate anion form of substrate is stabilized in the active site by making hydrogen bonds with the 2 hydroxyl group of nicotine amide ribose of NADPH and a water molecule (colored in blue). The proR hydrogen of NADPH as a hydride is attacking the electrophilic carbon C-13 of PGE2 (numbered as C4 in this figure). The Figure was rendered in PyMOL.

PAGE 179

162 Figure 4-38. Multiple alignment of the four purified AORs. The alignment was generated using the ClustalW web server software. Within a red frame the 3 Tyr residues, colored in blue in Figure 4-37. Wedges are showing the Ile and Tyr residues colored in purple in Figure 4-37.

PAGE 180

163 Figure 4-39. Multiple alignment of zinc independent en-reductase s belonging to MDR superfamily. First row the AOR from guinea pig. Last row the Quinone Oxidoreductase (QOR) from E. coli Alkenal hydrogenase from barley (ALH1), pulegenone reductase (PulR) from peppermint and P1-crystallin (P1 Cr) from Arabidopsis thaliana are supposed to belong to LTD family. Quinone reductase from strawberry (FaQR) and 2-haloacrylate reductase from Burkholderia sp. WS are supposed to belong to QOR family. The conserved NADPH binding moti f AXXGXXG is highlighted in green. In FaQR the first residue (Ala) in this motif is replaced by Gly. Wedges ar e indicating where residues (Cys, His and Cys respectively) for binding the catalytic zinc atom were supposed to be but they are absent in these two subfamilies of MDR superfamily. The alignment was generated using the MUSLE web server software.

PAGE 181

164 Lastly, an en-reductase from the soil bacterium Burkholderia sp. WS was inducibly synthesized when the cells were grown on 2-chloroacrylate.384 The purified protein catalyzed the reduction of chloroand bromo-acrylates to the corresponding ( S )-halo acids and thus it was named 2-haloacrylate reductase. It is unknown if the quinone oxirdoreductase from E. coli with which the 2-haloacrylate shares a significant sequence similarity (38.2% identity) also possesses en-reductase activity; however, the reductase from Burkholderia sp. did not act on quinones. In contrast to what is common for other oxidoreductases belonging to MDR superfamily, all the aforementioned en-reductases are metal independent and lack bound Zn+2. (Figure 4-39).345 The central cofactor-binding domain contains the characteristic Rossman fold. In the highly conver sed, in this fold, Gly-X-Gly-X-XGly (where X any amino acid) motif found in this fold, the first Gly has been replaced by an Ala in all but FaQR reductase, whereas a single residue insertion (i.e. Ala or Gly for FaQR) is apparent in all sequences (Figure 4-39).385 Thus, a nucleotide binding motif G/A-XX-G-XX-G is better description for this family of MDRs (Figure 4-40).386 Table 4-6. En-Reductases in MDR superfamily. Year (Ref) Organism Protein(Swiss-Prot accession no.) Gene Substrates Inhibitors 1993 (370) Sus scrofa (pig) NADP-dependent leukotriene B4 12hydroxydehydrogenase, 15oxoprostaglandin 13-reductase, (Q29073) LTB4DH 1996 (371) Homo sapiens (human) NADP-dependent leukotriene B4 12hydroxydehydrogenase, (Q14914) LTB4DH 2001 (387) Cavia porcellus (guinea pig) NADP-dependent leukotriene B4 12hydroxydehydrogenase, (Q9EQZ5) LTB4DH 2001 (375) Rattus novergicus (rat) Alkenal/one Oxidoreductase, (P97584) Ltb4dh

PAGE 182

165 Table 4-6. Continued. Year (Ref) Organism Protein (Swiss-Prot accession no.) Gene Substrates Inhibitors 2003 (382) Mentha x Piperita L. cv. Black Mitcham (+)-pulegenone reductase (Q6WAU0) AY300163

PAGE 183

166 Table 4-6. Continued. Year Organism Protein (Swiss-Prot accession no.) Gene Substrates 2000 (379, 381) Arabidopsis thaliana P1-crystallin (Q39172) P1 (At5g16970) 2005 (380) Hordeum vulgare cv. (barley), alkenal hydrogenase alkenal hydrogenase, ALH1 (Q2KM86) ALH 2006 (385) Fragaria x ananassa (strawberry) Fragaria x ananassa quinone oxidoreductase (Q84V25) QR 2005 (384) Burkholderia sp. WS 2-haloacrylate reductase (Q59I44) caa43

PAGE 184

167 Figure 4-40. Superposition of two en-reductase structures belonging to the MDR superfamily (green for LTB4DH from puinea pig; 1V3V, blue for 2-haloacrylate reductase;1WLY) with quinone oxidoreductase from E. coli (1QOR, red). Only the nucleotide binding motif is shown. Black loops, at the bottom, are part of the AAXGXXG motif, whereas the second inserted Ala residue is colored in magenta. The nicotinamide cofactor is shown only for QOR (red) and LTB4DH (green). The Figure was rendered in PyMOL. Short Chain Dehydrogenases Possessing En-Reductase Activity The Croteau laboratorys long efforts to understand monoterpenes metabolism in plants lead to the isolation of (-)-isopi peritone reductase (IspR) from peppermint382 (Table 4-7). A BLAST sequence analysis and identification of conserved motifs placed IspR in the short chain dehydrogenase/reduc tase superfamily (Figure 4-41). More interesting was the finding that the human carbonyl reductase (CR), a classical short chain dehydrogenase, showed 33% identity with the reductase from peppermint. Indeed, a year later the first experimental evidence for the dual action of CR on the same substrate (Table 4-7) would be published by Doorn et al.388 Previous work had demonstrated that this ubiquitous in nature enzyme catalyzes the NADPH-dependent reduction of many endogenous and xenobiotic carbonyl compounds and quinones.389 Its action on carbon-carbon double bonds has been explored so far only for 4-oxonon-2-

PAGE 185

168 enal (4ONE). It was found that the two carbonyl groups are prerequisite for the enone/al activity of human CR. If the one of the two carbonyl groups is reduced first, by the same enzyme, the obtained product is no longer a substrate for CR. Molecular modeling of 4ONE in porcine CR active site revealed no sp ecific or consistent contacts between the enzyme and substrate suggesting conformational freedom in the binding of 4ONE.388 Table 4-7. En-Reductases in SDR superfamily. Year Organism Protein Gene Substrates 2004 (388) Homo sapiens Human carbonyl reductase (P16152) CBR1 2003 (382) Mentha x Piperita L. cv. Black Mitcham, (-)-isopiperitenone reductase ( Q6WAU1) AY300162 CR and IspR possess classic characterist ics of the SDR superfamily NAD(P)H binding motif Gly-XXX-Gly-X-Gly (Figure 4-41).390, 391 On the other hand, in the catalytic motif Tyr-XXX-Lys, the Tyr residue has been replac ed by Glu in the case of IspR. A Tyr residue at this position has been shown to participate in catalysis by making a critical hydrogen bond with the carbonyl group (Figure 4-42).392 Unclassified En-Reductases The last decade, at least eight novel en-reductases have been isolated from different organisms unfortunately with no amino acid sequence information. Half of these enzymes have been purified from bakers yeast and their substrate specificities are listed in Table 4-8. Only the purification protocols for the two enone reductases EI and EII, have been reported in detail.393

PAGE 186

169 Figure 4-41. Multiple alignment of the two known short chain en-reductases (isopiperitenone reductase, IspR; and human carbonyl reductase, CR_Hs) with the porcine CR (CR_Ss) The conserved NAD(P)H binding motif GXXXGXG is highlighted in green. Within the red frame the catalytic motif YXXK. The alignment was generated using the Muscle web server software. Figure 4-42. Stereoview of the catalytic ac tive site of porcine carbonyl reductase (1CYD) with the nicotinamide cofactor in magenta and isopropanol in green. The key for the catalysis residues Tyr and Lys are colored in purple.

PAGE 187

170 Both the EI and EII reductases have the same substrate specificities but possess very distinct biophysical properties. EI was found to be an NADPH-dependent heterodimer which was decomposed under denaturing electrophoresis into two subunits of 34 kDa and 37 kDa. The second enzyme, EII, was also decomposed into two subunits of 56 kDa and 64 kDa but it was NADH-dependent. Both enzymes showed high enantioselectivity towards the reduction of cis and trans -phenyl-2-butenal, producing ( R )-2-phenylbutanal in >97% ee Interestingly the three en-reductases isolated from Nicotiana tabacum show divergent substrate specificities (Table 4-8).394 The 44 kDa reductase (p44) is only reduces -substituted cyclohexenones to corresponding ( R )-ketones, while the reduction of -alkylatedenones by the 90 kDa reductase (p90) affords ( S )-ketones. In contrast, the 74 kDa reductase (p74) was found to reduce only exocyclic double bonds. Unfortunately, even though detailed purification protocols have been reported for the abovementioned enzymes, their complete amino acid sequences have not been revealed. In case of the p74 reductase, the sequence of a small stretch of amino acids produced after digestion of p74 with lysylendopeptidase showed significant similarity with the isovaleryl-CoA dehydrogenase from potato that has desaturase activity and with other putative homologs in Arabidopsis thaliana .395 Lastly, the three en-reductases isolated from the protozoa in genus Euglena showed narrow substrate specificity with strongest preference for -substituted cyclohexenones.

PAGE 188

171 Table 4-8. List of unclassified en-reductases. Year (Ref.) Organism Protein Mr Substrates 1998 (393) Bakers yeast EI 34 kDa and 37 kDa 1998 (393) Bakers yeast EII 56kDa and 64 kDa

PAGE 189

172 Table 4-8. Continued. Year Organism Protein Mr Substrates 1998 (396) Bakers yeast Carbon-carbon double bond reductase unknown 2001 (397) Bakers yeast, Nitroalkene reductases Nitroalkene reductases: YNAR-I, YNAR-II, unknown 1996 (398) Nicotiana tabacum verbenone reductase (p90) 90 kDa (2 subunits)

PAGE 190

173 Table 4-8. Continued. Year Organism Protein Mr Substrates 1989 (399) Nicotiana tabacum carvone reductase (p44) 44 kDa (2 subunits) 2002 (395) Nicotiana tabacum pulegone reductase (p74) 74 kDa (2 subunits) 1998 (400) Euglena gracilis enone reductase 55 kDa 2000 (401) Astasia longa Reductase-I, Reductase-II 35 kDa, 36 kDa Cofactor Specificity of En-Reductases The cofactor specificity of the en-reductases presented in the previous sections is listed in Table 23. En-reductases belonging to the Old Yellow Enzyme family are A-type

PAGE 191

174 enzymes and possess more or less stronger preference for the NADPH instead of NADH cofactor with only exception being the morphinone reductase (MorR) which is a NADH-dependent enzyme. In contrast, enoate reductases are B-type, NADHdependent enzymes. For the non-flavin containing enzymes, experimental data exist only for the alkenal/one oxidoreductase, which is an A-type enzyme and the human carbonyl reductase, which is a B-type enzyme. Table 4-9. Cofactor specificity of En-Reductases. Entry Enzyme NADPH NADH Stereospe cificity 1 OYE1 Yes ( strongly preferred) Yes A (pro-R) 2 OYE2 Yes (preferred) Yes ---c 3 OYE3 Yes (preferred) Yes --4 EBP1 Yes nd --5 N-Ethylmaleimide Reductase of Y.lipolytica Yes nd A (pro-R) 6 oye Yes Yes --7 MorB No Yes --8 NemA Yes (slightly preferred) Yes --9 PB2 PETN reductase Yes ( strongly preferred) Yes A (pro-R) 10 Type strain PETN reductase Yes (slightly preferred) Yes --11 Glycerol trinitrate reductase --Yes --12 2-cyclohexen-1-one reductase Yes (preferred) Yes --13 YqjM Yes (slightly preferred) Yes --14 SYE1 Yes Yes (preferred) --15 SYE3 Yes (preferred)a Yes (preferred)b --16 SYE4 Yes (preferred) Yes --17 enoate reductase No Yes B (pro-S) 18 Alkenal/one oixidoreductase Yes No A (pro-R)

PAGE 192

175 Table 4-9. Continued. Entry Enzyme NADPH NADH Stereospe cificity 19 FaQR Yes Yes --20 P1-crystallin Yes ( strongly preferred) Yes --21 2-haloacrylate Reductase Yes No --22 (+)-pulegenone reductase Yes No --23 (-)-isopiperitenone reductase Yes (preferred) Yes --24 Carbonyl Reductase Yes (strongly preferred) Yes B (pro-S) 25 YNAR-I Yes (preferred) Yes --26 YNAR-II Yes (preferred) Yes --27 EI Yes No --28 EII No Yes --29 Enone reductase No Yes A (pro-R) 30 Pulegone reductase Yes (preferred) Yes --31 Verbenone reductase Yes (strongly preferred) Yes B (pro-S) 32 Reductase-I Yes Yes (preferred) A (pro-R) 33 Reductase-II Yes Yes B (pro-S) a Preferred with oxygen as substrate. b Preferred with NEM as substrate. c not determined. Concluding Remarks Where should we look for en-reductases useful in biocatalysis? In an attempt to answer the above question in this chapter, it became apparent that with the exception of enzymes belonging to enoate reductase family, no serious efforts have been made towards the utilization of en-reductases from other families in the field of biocatalysis. Unfortunately, functional heterologous ERs could not be easily produced in E. coli therefore problems related to overexpression and stabilization of these biocatalysts should be first solved before we envision their use in preparative scale biotransformations. On the other hand, mo re than 30 isoenzymes form the OYE family have been successfully overexpressed in E. coli but a systematic study regarding their substrate specificity is missing.

PAGE 193

176 Also, the characterization of flavin-free en-reductases is still in its infancy. A sequence similarity search in databases retu rns a large number of proteins in this category, albeit with unknown function. It woul d be interesting to learn if those enzymes could act as reductases, as well as if there are any specific features that discriminate them from the alcohol dehydrogenases and/or the quinone reductases. Also, the possibility that the orphan en-reductases may belong to any of the abovementioned families or they may uncover new en-reductase families remains open. To this end, we will present in the next chapter our efforts in answering the above questions as well as a comparative substrate specificity study among different enreductase subfamilies.

PAGE 194

177 CHAPTER 5 IN VIVO RECOMBINATIONAL CLONING OF OLD AND NEW EN-REDUCTASES: A COMPARATIVE SUBSTRATE SPECIFICITY STUDY By three methods we may learn wisdom: fi rst, by reflection, which is the noblest; second by imitation, which is the easiest; and third, by experience, which is the most bitter. -Confucius Quotes, Confucius402 Introduction Several studies have shown that the burning issue in the field of biocatalysis is the lack of a broad enzyme platform.403 Difficulties identifying the most appropriate biocatalyst for a particular biotransformation, in a short time-frame, usually render the bio-based technology a second choice in the manufacturing of pharmaceutical intermediates. The traditional approach to solving this problem is the perpetual screening of strain collections in hopes of identifying a suitable enzyme for the desired transformation. Unfortunately, this method is time consuming and unable to cover the estimated 99% of the unculturable microorganisms. Alternatively, different strategies (Figure 5-1) have been developed the recent years in an effort to create the desired diversity in vitro by means of directed evolution,404 or to gain access to the existed biodiversity. In the latter case, the metagenome approach and the sequence-based strategy have been recognized.405 While in the former, the entire genomic DNA from soil samples must be cloned and expressed in a heterologous host such as E. coli in the sequence-based approach, only a small number of protein sequences are required to act as baits for mining novel biocatalysts from the constantly increasing whole genome sequence databases.

PAGE 195

178 Figure 5-1. Different approaches towards the identification of novel biocatalysts. The concept of the sequence-based discovery for biocatalysis was beautifully demonstrated by Stewart and coworkers using the bakers yeast genome database for mining putative ketone reductases. Sequence si milarity searches using a small number of known reductases as probes revealed the presence of ca. 50 putative ketone reductases in the S. cerevisiae genome. Eighteen out of twenty two selected were successfully cloned and overproduced in E. coli as glutathione S-transferase fusion proteins A small scale screening of this collection towards various -keto esters not only uncovered biocatalysts with distinct and in some cases opposing stereopreferences but also provide the basis for the production of synthetically useful building blocks in gram scale quantities. The powerful features of this methodology were immediately Metagenome Approach Directed Evolution SequenceBased Approach Screening Strain Collections Novel Biocatalysts

PAGE 196

179 recognized by other research groups, which applied the same concept in order to build libraries of: i) ketoreductases from modular polyketide synthases,406 ii) epoxide hydrolases from bacteria407 and iii) Bayer-Villiger monooxygenases from Mycobacterium tuberculosis H37Rv.408 In this work, the abovementioned principle was applied in an expanded and improved form for the construction and evaluation of a collection of fifteen known and putative alkene reductases. Experimental Strategy Our experimental strategy is schematically depicted in Figure 5-2 and each step will be described more completely below. Selection of Alkene Reductases Clones Our first concern in selecting the alkene reductases genes was to include the highest possible diversity based on two constrai nts: i) the availability of the genomic DNA from which the desired gene will be amplified and ii) the available time for completion of the project. Our analysis of the different alkene reductases families, in the previous chapter, indicated that the Old Yellow Enzyme family is the wealthiest source of putative alkene reductases. We used the well-studied OYE1 from Saccharomyces carlsbergensis as a probe in order to select OYEs from other yeasts, as well as from bacteria and plants (Table 5-1). All of our plant targets have been overexpressed before in E. coli but their substrate specificities are the least studied. From bacteria, only NemA has been studied before and we felt that we should include this particular enzyme in our list since it is derived from a host that we routinely use to overexpress the rest of biocatalysts. Among the OYEs from fungi, the catalytic activities of OYEA and OYEB from Schizosaccharomyces pombe has not been tested before. Nonetheless the percent identity among the selected OYEs spans from 91 to 28 percent (Table 5-2).

PAGE 197

180 Figure 5-2. Construction of the alkene reductase library. A. Selection of target genes. B. In vivo recombinational cloning. C. Overexpression and purification of GST-En-Reductases.

PAGE 198

181 Table 5-1. List of alkene reductases cloned in this study. Entry Organism Gene (GI) Protein (Swiss Prot. Acc. #) Purified before? (mg/L)a 1 Saccharomyces carlsbergensis or Saccharomyces pastorianus OYE1 (4092) OYE1 (Q02899) Yes (32) 2 S. cerevisiae YHR179W (172029) OYE2 (Q03558) Yes (24) 3 S. cerevisiae YPL171C (460038) OYE3 (P41816) Yes (48) 4 Candida macedoniensis or Kluyveromyces marxianus or Candida kefyr oye (49387279) Oye (Q6I7B7) Yes (36) 5 Schizosaccharomyces pombe SPAC5H10.04 (854599) OYEA (Q09670) No (12) 6 S. pombe SPAC5H10.10 (854599) OYEB (Q09671) No (48) 7 E. coli BL21(DE3) nemA (1711243) NemA (P77258) Yes (44) 8 Pseudomonas Putida KT2440 PP2486 (24987239) ppOYE (Q88K07) No (48) 9 P. putida KT2440 nemA (24987239) ppNEMA (Q88I29) No (22) 10 Synechococcus elongatus PCC 7942 Synpcc7942_0473 (81167692) SeOYE (Q31R14) No (60) 11 Arabidopsis thaliana At1g76680 (2765082) OPR1 (Q8LAH7) Yes (32) 12 Arabidopsis thaliana At1g76690 (26450548) OPR2 (Q8GYB8) Yes (0) 13 Arabidopsis thaliana At2g06050 (5059114) OPR3 (Q9FUP0) Yes (34) 14 Lycopersicon esculentum (tomato) OPR1 (4894181) Leopr (Q9XG54) Yes (48) 15 Rattus norvegicus (rat) Ltb4dh (6012070) Ltb4dh (P97584) Yes (14) 16 S. cerevisiae YNL134C (45269933) YNL134c (P53912) Nob (30) a Amount of protein obtained in this study. b The protein has been cloned and overexpressed in E. coli and tested for its catalytic activity as a crude extract.

PAGE 199

182Table 5-2. Percent identity in between OYEs cloned in this study. aThe different OYE family subgroups: fungi, bacteria and plants ar e highlighted with yellow, blue and green respectively. Ind. (%)a OYE1 OYE2 OYE3 oye OYEA OYEB NemAppNe mA ppOYE SeOYE OPR1 OPR 2 OPR 3 Leopr OYE1 100 91 80 69 36 35 37 34 33 41 35 36 37 37 OYE2 91 100 81 71 37 36 37 35 35 40 31 36 38 38 OYE3 80 81 100 68 37 36 39 37 34 41 37 38 36 39 Oye 69 71 68 100 37 35 38 36 34 38 34 35 35 37 OYEA 36 37 37 37 100 57 32 32 30 36 35 37 35 38 OYEB 35 36 36 35 57 100 30 31 28 36 35 34 34 36 NemA 37 37 39 38 32 30 100 69 46 49 41 43 37 42 PpNEMA 34 35 37 36 32 31 69 100 44 50 39 41 38 42 PpOYE 33 35 34 34 30 28 46 44 100 47 38 38 35 39 SeOYE 41 40 41 38 36 36 49 50 47 100 45 49 44 49 OPR1 35 31 37 34 35 35 41 39 38 45 100 84 47 68 OPR2 36 36 38 35 37 34 43 41 38 49 84 100 51 72 OPR3 37 38 36 35 35 34 37 38 35 44 47 51 100 50 Leopr 37 38 39 37 38 36 42 42 39 49 68 72 50 100

PAGE 200

183 From the medium-chain reductase family, we selected the well-studied Ltb4dh from rat mainly because the cDNA was co mmercially available. In addition, its demonstrated broad substrate specificity, its enantioselectivity towards prochiral substrates had never been explored. Member s from the short chain reductase family were not included in this collection. The reported sensitivity of the enoate reductases to oxygen as well as their poor expression in E. coli discouraged our including this family in collection. We also made an effort to answer the following question: Besides the two known alkene reductases, namely OYE2 and OYE3 characterized in Masseys laboratory, do other genes in the bakers yeast genome encode proteins responsible for reducing activated alkenes? Based on our discussion in the last part of chapter 4, four out of eight reported orphan alkene reductases have been purified from this microorganism. We decided to repeat the purification protocol for the EI and EII enone reductases described by Wanner and Tressl in 1998, for following reasons: Both the EI and EII reductases showed broad substrate specificity, albeit similar to one another and to OYE1. this is the only report for which detail ed experimental protocol is given and the published SDS-PAGE analysis is clearly shown that these two enzymes have different molecular weights from the OYE2 or OYE3. it is intriguing that even though this report was published three years after Masseys characterization of OYE2 and two years after the release of the S. cerevisiae genome, the authors neither commented about the existence of the OYE2 and its catalytic activity towards enones nor explained why they did not determine the N-terminal amino acid sequence of their isolated proteins. Cloning and Expression of Alkene-Reductases The methodology for constructing of the yeast GST-reductase library by Kaluzna et al.409 was inspired by the seminal work of Phizicky and coworkers.410 In the latter, a genomic library of yeast strains expressing GST-ORF fusion proteins was constructed and used to link enzymatic activity to gene function. Their approach was rapid and sensitive mainly because:

PAGE 201

184 cloning and overexpression of the whole proteome were achieved in same host (i.e. yeast) and in a single step, taking advantage of the very efficient recombinational machinery of S. cerevisiae detection of 10-17 moles of product from a substrate with a single 32P-label was achieved with an enzyme concentration as little as 0.02 M side-reactions, by-products and catalysts inactivation or degradation were minimized due to the fact that contaminant proteins were eliminated during the protein affinity-chromatography purification step introduced before the screening of the yeast lysate. Since 1999 numerous improvements have been introduced to the original protocol;411, 412 however, whenever we attempt to impl ement biochemical and enzyme genomic approaches in the field of biocatalysis we should always keep in mind that: we are looking for enzymes capable of successfully transforming nonnatural substrates, which usually means lower turnover numbers and the requirement of higher amounts of protein, typically in a range of 0.5-5 M most of the compounds for which an organic chemist may show an interest do not possess a radio-labeled atom or a chromophore. As a consequence, higher amounts are required for their detecti on, typically in a range of 1-5x10-6 moles for 1 ml scale biotransformations. Driven by the need to produce sufficient amount s of yeast reductases, Stewart and coworkers had to switch to E. coli overproduction. Although the fusion protein library approach permits the rapid purification and screening of a number of biocatalysts, it suffers in terms of the time required to clone the target genes into the appropriate bacterial expression vector. In order to break the cloning bottleneck step, we introduced a one step recombination-mediated cloning method into our overall strategy. This modification allowed us, with appropriately designed primers, to PCR-amplify and fuse the target cassettes to the C-terminus of the GST-tag directly, thereby bypassing otherwise laborious sub-cloning steps. We evaluated two pathways: i) the wellestablished413 yeast in vivo cloning via homologous recombination (HR) ii) and the RecA mediated in vivo recombination cloning in E. coli which was introduced recently by Finley and co-workers.414 Even though a number of kits are commercially available for in vitro site specific HR cloning415 the in vivo HR strategy is much cheaper and convenient

PAGE 202

185 since no purchase of recombinase is required and any in-home modified vector could be used without the restrictions that the site-specific HR imposed. Experimental Procedures Purification of YNL134c Preliminary efforts in YNL134c isolation were performed in collaboration with Erin Hanlin in our laboratory and started from commercially available bakers yeast, as described by Tressl and Wanner. Our inability to reproduce the results obtained in the original published protocol lead us to the m odified purification protocol described below. All enzyme purification steps were carried out at 4 C. Phenylmethylsulphonylfluoride (PMSF) and 2-mercaptoethanol were added to the standard buffer immediately before use. SDS-PAGE analyses were carried out according to Laemmli416 and protein concentration was estimated by a Bradford assay using bovin serum albumin (BSA) as a protein standard.417 Materials and supplies The yeast strain DAY128 (oye2 oye3 ) used for protein isolation was generously provided by Dr. David Amberg. DEAE CL-6B was purchased from Sigma. Pre-packed HiLoad Q Sepharose, Mono Q anion exchange, Sephacryl S-300 Hi Prep 16/60 and fast desalting columns were all purchased from Pharmacia, Dyematrex Gel Blue A matrix was purchased from Millipore. Superdex 200 and Sephadex G-25 matrix were purchased from Pharmacia. Enzyme activity assays Enzyme activity was determined spectrophotom etrically at 25 C by following the decrease in absorbance of NAD(P)H at 340 nm. The standard assay mixture contained 0.2 mM NAD(P)H (10 L of 20 mM stock solution prepared immediately before use in 0.1 mM KPi, pH 7.0), 2.5 mM 1-octen-3-one (100 L of 25 mM stock solution made in acetate buffer, pH 4.8 containing 0.1 % Triton X-100) and the appropriate amount of the

PAGE 203

186 enzyme preparation in a final volume of 1 mL of acetate buffer pH 4.8. The above solution was passed through a 0.45m filter before being placed in a 1.00 cm cuvette. The background NAD(P)H oxidation was measured using an identical mixture in which the enone substrate was replaced by acetate buffer. One unit of enzyme activity was defined as the quantity sufficient to oxidize 1 mol of NAD(P)H per minute in the mixture described above. Unit per mL of enzyme preparation were calculated based on the following equation: Dilution enzyme V assay V *l 340 1000 dt dA mL Units where dA/dt the slope in AU/min, 340= 6270 mol-1 L cm-1, l=1 cm, Vassay = 1 mL, Venzyme = volume of enzyme added in mL, Dilution= dilution factor. Cell growth for protein isolation Day128 cells were pre-cultured in double strength, non-selective yeast growth medium (2 x YPD: 20 g Bacto-Yeast Extract, 40 g Bacto-Peptone, 40 g Dextrose per liter) overnight at 30 C and then diluted 1:100 into 500 mL of the same medium in eight 2 L baffled flasks. The cultures were grown at 30 C for 24 h and the cell harvested by centrifugation (5000x g for 10 min at 4 C) and stored at -20 C. Preparation of crude extract Cells (53 g wet mass) were thawed, washed twice with three volumes of deionized water, resuspended in 200 ml Tris-HCl pH 7.5 containing 0.1 mM phenylmethylsulphonylfluoride (PMSF) and 5 mM 2-mercaptoethanol (standard buffer) and disrupted in a French pressure cell at 16,000 psi. The process was repeated twice. Debris was removed by ultracentrifugation at 70000x g (30,000 rpm, 45 Ti rotor) for 30 min and the pellet was discarded.

PAGE 204

187 Streptomycin sulfate purification Streptomycin sulfate solution was added to the clarified cell extract (5.7 mL of 25% solution) dropwise over 20 min. The protein solution was stirred for additional 2.5 h and the pellet collected by ultracentrifugation at 70000x g for 20 min was discarded. The supernatant was concentrated overnight by ultrafiltration using an Amicon ultrafiltration cell fitted with a PM-10 membrane under nitrogen at 20 psi. Desalting by Sephadex G-25 The supernatant (~ 25 ml) was applied to a Sephadex G-25 column (40 cm x 3.5 cm) via a peristaltic pump at a rate of 1 ml/min equilibrated with standard buffer. The column was eluted with the same buffer at the same rate and 8 ml fractions were collected. All fractions bearing an absorbance at 280 nm >0.2 AU were pooled together and concentrated by ultrafiltration, as described above. DEAE CL-6B anion exchange chromatography The ultrafiltrate (~ 15 ml) was ultracentrifuged and applied to a DEAE CL-6B anion exchange column (18 cm x 3 cm) equilibrated with standard buffer. After the column was loaded and washed with 2 bed volumes of standard buffer, a NaCl gradient was applied from 0.0 to 0.5 M over 5 bed volumes. When the elution was completed, the column was further washed (1 bed volume) with the same buffer containing 0.5 M NaCl followed by 60 ml of standard buffer with 2 M NaCl. The proteins were eluted at a flow rate of 1 ml/min and fractions of 8 ml were collected. Fractions 31-38, having NADPH and NADH activity, were combined, concentrate d to a final volume of 35 ml and dialyzed overnight against 4 L ( 2 x 2L) of standard buffer. HiLoad Q sepharose anion exchange chromatography The desalted protein solution was centrifuged as before and loaded on a 10 cm x 1.6 cm Q Sepharose anion exchange column using a Pharmacia FPLC system. After the column was loaded and washed with 2 bed volumes of standard buffer, a NaCl

PAGE 205

188 gradient was applied from 0.0 to 0.5 M over five bed volumes. When the elution was completed, the column was further washed (1 bed volume) with the same buffer containing 0.5 M NaCl. The proteins were eluted at a flow rate of 1 ml/min and fractions of 5 ml were collected. Fractions 26-34, having NADPH and NADH activity, were combined, concentrated to a final volume of 12 ml and dialyzed overnight against 4 L (2 x 2L) of standard buffer. Dyematrex gel blue A affinity chromatography The desalted protein solution was loaded onto a Dyematrex gel blue A column (20 x 1.6 cm) at a flow rate of 0.3 ml/min. The elution was stopped for ~ 1.5 h in order to maximize interactions between the dye and the proteins. Then, the column was washed with 100 ml standard buffer until A280 of the eluant was ca. 0.2, and then it was eluted with 0.5 M NaCl in standard buffer at a flow rate of 0.5 ml/min. Fractions of 5 ml were collected. Fractions 6-12 (fraction FI), having NADPH and NADH activity, were combined, concentrated to a final volume of 10 ml and dialyzed overnight against 4 L ( 2 x 2L) of standard buffer. Fractions 22-28 (fraction FII), bearing absorbance at 280nm > 0.2 AU, were also collected, concentrated and desalted in the same manner. Both fractions, FI and FII, were further concentrated to a final volume of 1 ml. Superdex 200 filtration chromatography The fraction FI, from the previous step, was centrifuged and loaded onto a Superdex 200 filtration column ( 25 cm x 1.4 cm) at a flow rate of 0.5 ml/min equilibrated with 0.15 M NaCl in standard buffer. The proteins were eluted in the same buffer. Fractions of 2 ml were collected and NAD(P)H activity was found in fractions 5-8. The latter fractions were pooled together, dialyzed overnight against 2 L standard buffer, concentrated to final volume of 1 ml and desalted further using a Pharmacia fast desalting column. The desalted fractions were combined and concentrated to a final volume of 1 ml.

PAGE 206

189 HiLoad Q Sepharose anion exchange chromatography The desalted protein preparation from the previous step was loaded onto a Mono Q anion exchange column (5 cm x 0.5 cm) at a flow rate of 1 ml/ min. After the column was loaded and eluted with 10 ml of standard buffer, a 50 ml NaCl gradient was applied from 0.0 to 0.5 M NaCl in standard buffer. The column was further washed with 10 ml of 0.5 M NaCl in standard buffer. Fractions of 1 ml were collected and NAD(P)H activity was found in fractions 29-34. The latter fractions were combined and desalted as before. Sephacryl S-300 Hi Prep filtration chromatography The fraction from the previous step was concentrated to a final volume of 0.5 ml and was loaded onto a 60 cm x 1.6 cm Sephacryl S-300 Hi Prep filtration column at flow rate of 1 ml/min equilibrated with 0.15 M NaCl in standard buffer. The proteins were eluted in the same buffer. Fractions of 1 ml were collected and NAD(P)H activity was found in fractions 72-82. Fractions 74-78 were 10 fold concentrated individually with a centrifugal filter device (Microcon YM-3, 3,000 molecular weight cut off) before loading on a SDS-PAGE gel and submitted for LC-MS analysis. Protein Sequencing N-terminal sequencing and LC-MS analysis of the YNL134c were performed by the University of Florida-ICBR-Protein Core and sample preparation followed their recommended protocols. Protein digestion and LC-MS analysis was carried out by Dr. Scott McClung. Cloning and Purification of GST-Alkene Reductases Biochemicals, plasmids and strains Restriction endonuclucleases were purchased from New England Biolabs or Fermentas. T4 DNA ligase, Taq and Vent polymerases were obtained from New England Biolabs. Glutathione uniflow resin was purchased fron Clontech. cDNA from

PAGE 207

190 rat (normal tissue: liver) was obtained from BioChain. DNA from Synechococcus elongatus PCC 7942 and a cosmid containing the SeOYE gene were kindly provided by Dr Susan S. Golden. E. coli expression vector pET22b(+) was obtained from Novagen. pET-OYE was a gift from V. Masseys laboratory. Yeast shuttle vector pYEX4T-1-Rec. DOM. was a gift from Dr M. Martzen. Plasmids, pIK1(-) and pIK2 were constructed by Dr Iwona Kaluzna in our laboratory. Clones cTOF30K9, cLEY12J2 and cTOD19I19 for amplification of Leopr gene were obtained from sol genomics ne twork, Cornell University. Clones pda07378, pda10767 and pdx20368 for amplification of OPR1, OPR2 and OPR3 genes respectively, were purchased from RIKEN Tsukuba Institute. S. cerevisiae strain, BY4742 used for homologous recombination and DNA isolation was a gift from Dr. Thomas Lyons. Strains Kluyveromyces marxianus (ATCC 200965), Schizosaccharomyces pombe (ATTC 16979) and Pseudomonas pudida (ATCC2440) used from DNA isolation were purchased from American Type Culture Collection. E. coli KC8 competent cells, used for homologous recombination, were purchased from Clontech. Cloning of alkene reductases. Recombinant DNA techniques were carried out as described by Maniatis et al.418 DNA used as a template in PCR reactions, was isolated from different yeasts and bacteria as described elsewhere.419 Genes were PCR-amplified using a Perkin-Elmer Geneamp PCR system 2400. Clone cTOD wa s used for amplification of Leopr gene (restriction analysis showed that cLEY clone contains only a part of the Leopr gene and DNA sequencing of cTOF clone revealed a single nucleotide deletion in the Leopr gene). When an ampR plasmid was used as a template for PCR, linearization of the templateplasmid after PCR was found necessary for eliminating of false positives after the in vivo homologous recombination in E. coli KC8 strain. All PCR products and the linearized

PAGE 208

191 universal vectors pDJB2(3) were purified by agarose gel electrophoresis prior to recombinational cloning. As a rule, ca. 150 ng of purified PCR product and linearized vector were used for the homologous recombination. Routine E. coli transformations were performed by electroporation. Yeast transformations utilized the high e fficiency LiAc/SS-DNA/PEG protocol.420 E. coli KC8 chemically competent cells were transformed according to the manufactures instructions or they were prepared using a PEG-DMSO protocol.414 Yeast recombinants were plated on the appropriate selective media. Total yeast DNA was prepared and used to transform E. coli TOP10 cells in order to screen for recombinant molecules by restriction digestion analysis. E. coli KC8 recombinants were screened by colony-PCR. Purified plasmid DNA for sequencing was obtained by density gradient ultracentrifugation with CsCl in presence of ethidium bromide. Isolation of GST-fusion proteins An overnight culture of the appropriate ov erexpression strain grown in LB medium containing 100 g/ml amp.(40 g/mL kan was used for GST-OYE1) was diluted 1 : 100 into 500 mL of the same medium in a 2 L baffled flask. The culture was shaken at 37C until the optical density at 600 nm reached 0.5-1.0, then isopropylthio-D-galactoside (IPTG) was added to a final concentration of 100 M and the culture was shaken for an additional 6 hours at room temperature. The cells were collected by centrifugation, washed twice with cold sterile water and then resuspended in 30 mL loading buffer (50 mM Tris-Cl, 4 mM MgCl2, 1 mM DTT (added immediately before use), 1 mM PMSF (added immediately before use), 10% glycerol, pH 7.5). All purification steps were carried out at 4C. The cells were lysed by passage through a French pressure cell and debris was removed by centrifugation at 15,000 x g for 20 min at 4C. The supernatant was passed through a 0.45 m filter and loaded rapidly onto a column containing 10 mL of glutathione-uniflow resin. The resin was thoroughly mixed with the lysate and then

PAGE 209

192 allowed to settle in the column. The protein sample was continuously circulated through the column for 3-5 h via a peristaltic pump at a flow rate of ~0.5 ml/min. Then, the nonadsorbed lysate was drained and the resin was washed, without disturbing it, twice with 20ml of loading buffer. GST-fusion proteins were eluted with 40 ml of freshly prepared elution buffer (Loading buffer (39.6 ml), 0.40 ml NaOH (2 M) and solid reduced glutathione (0.31 g)). The eluant was concentrated and dialyzed overnight against 1 L of dialysis buffer (20 mM Tris-Cl, 4 mM MgCl2, 55 mM NaCl, 2 mM EDTA, 1 mM DTT, 50% glycerol, pH 7.5) prior to storage at -20C. Regeneration of glutathione resin The glutathione resin was regenerated by washing (3 x 200 mL) with regeneration buffer A (0.1 M Tris-Cl, 0.5 M NaCl, pH 8.5) and the same volume of regeneration buffer B (0.1 M Sodium acetate; 0.5 M NaCl, pH 4.5). After regeneration, the resin was equilibrated with loading buffer. Screening of Alkene Reductases Chemicals were purchased from Acros, Aldrich, Alpha Aesar or Lancaster and used without further purification. Preparation of 6-alkyl-2-cyclohexen-1-ones The methyl substituted cyclohexenones were prepared via tandem Michael addition-aldol condensation of the corresponding -ketoesters to conjugated enals421 (Table 5-3). To a stirred solution of -keto ester (7 mmol) and acrolein (7 mmol) in 7 ml of tBuOH was added a catalytic amount of t-BuOK at 4 C. The reaction mixture was stirred at that temperature for 30 min and then 1.75 mmol of t-BuOK were added and the mixture was held at for 20 h. After cooling to room temperature, the mixture was quenched with 10 ml of 1M HCl, diluted with a 1:1 mixture of ether and benzene (80 mL) washed with 1 M NaOH (20 mL x 3) and brine (20 mL x 2). The organic layer was dried

PAGE 210

193 over magnesium sulfate and concentrated under reduced pressure. The crude 6substituted 2-cyclohexenones were purified by flash column chromatography (SiO2, 1020 % ethyl acetate in hexanes). The yield did not exceed the 10-12% in any case. Table 5-3. Starting materials for preparation of 6-substituted cyclohexenones. -ketoesters enals Product Name 6-methyl-2cyclohexenone O 6-ethyl-2cyclohexenone 6-allyl-2cyclohexenone 6-methyl-2-cyclohexenone 1H NMR (300 MHz, CDCl3) : 1.14, (td, 3H J = 6.9), 1.69-1.80, (m, 1H), 2.03-2.10 (m, 1H), 2.38-2.40 (m, 3H), 5.98 (td, 1H, J = 10.5), 6.896.95 (m, 1H). 6-ethyl-2-cyclohexenone 1H NMR (300 MHz, CDCl3) : 0.92 (t, 3H J = 7.5), 1,371,47 (m, 1H), 1.72-1.88 (m, 2H), 2.06-2.24 (m, 2H), 2.36-2.38 (m, 2H), 5.98 (td, J = 10. 2, 1H), 6.86-6.92 (m, 1H) 6-allyl-2-cyclohexenone 1H NMR (300 MHz, CDCl3) :1.22-1.26 (m, 1H), 1.71-.175 (m, 1H), 2.05-2.14 (m, 2H), 2.36-2.38 (m, 2H), 2.58-2.65 (m, 1H), 5.07 (t, J = 8.4, 2H), 5.73-5.82 (m, 1H), 6.00 (dd, J = 9.9, 1H) 6.90-6.95 (m, 1H). Preparation of neral and geranial Nerol or geraniol was oxidized to corresponding neral and citral with iodosobenzene diacetate (IBD) and catalytic amount of TEMPO.422 To a stirred solution of allylic alcohol (3.2 mmol) in 3 ml CH3CN and 1 ml of 100 mM KPi, pH 7.0 was added 3.6 mmol IBD and a catalytic amount of TEMPO (0.3 mmol)

PAGE 211

194 at 0 C. The reaction mixture was stirred at that temperature until alcohol was no longer detectable by GC-MS analysis (in 1-2h). The reaction mixture was diluted with diethyl ether (30 ml) and extracted with saturated aqueous sodium thiosulfate ( 3 x 30 ml). The aqueous phase was separated and extracted once with 40 ml of diethyl ether. The combined organic layers were washed with saturated aqueous sodium hydrogen carbonate (50 ml) and then with brine (50 ml). The organic layer was dried over magnesium sulfate and concentrated under reduced pressure. The crude product was purified by column chromatography starting with hexanes and finishing with 5% diethyl ether in hexanes. The yield was 50 % in case of neral and 25% in case of geranial. Traces (2-3 %) of the opposite alkene isomer were detected in both cases by GC analysis. Neral 1H NMR (300 MHz, CDCl3) : 1.59 (s, 3H), 1.68 (s, 3H), 1.98 (d, 2H, J = 1.5 Hz), 2.20-2.27 (m, 1H), 2.58 (t, 2H, J = 6 Hz), 5.08-5.13 (m, 1H), 5.87 (d, J = 8.1 Hz, 1H), 9.90 (d, J = 8.1, 1H) Geranial 1H NMR (300 MHz, CDCl3) : 1.61 (s, 3H), 1.69 (s, 3H), 2.17 (d, 2H, J = 1.3 Hz), 2.22 (m, 3H), 5.06 (m, 1H), 5.88 (td, 1H, J = 8.1), 10.0 (d, J = 7.2, 1H) Biotransformations A typical 1ml-scale biotransformations contained NADP+ (0.20 moles), glucose6-phosphate (14 moles),and glucose-6-phosphate dehydrogenase (5 g) as an NADPH regeneration system, the appropriate amount of GST-En reductase (50-150 g) and 5 moles of functionalized alkene in 100 mM KPi, pH 7.0. Reaction mixtures were incubated at 30 C and sampled, by extraction with ethyl-acetate, for GC analysis, typically after 18-20 h. Detection of reduced activated alkenes Biotransformations were monitored by GC-MS (EI) using a DB-17 column (0.25 mm x 25 m, 0.25 m film thickness) and the following conditions: 60 C (2 min) to 180 C

PAGE 212

195 (5 min) at 10C/min. Chiral phase GC analyses were performed on a Chirasil-Dex CB column (0.25 mm x 25 m, 0.25 m film thickness) using one of the following conditions: i) 80 C (2 min) to 95 C ( 0.5 C/min, 1 min) followed by a 12C/min gradient to 180 C (7.00 min), ii) 60 C (2 min) to 115 C ( 0.5C/min, 5 min) followed by a 2.5C/min gradient to 180 C (5.00 min). Other analys es utilized a Beta DEX column (0.25 mm x 30 m, 0.25 m film thickness) using one of the followed programs: i) 95 C (35 min) to 160 C (5.0 C/min, 2 min) followed by a 10.0 C/min gradient to 200 C (5.00 min) ii) 95 C (35 min) to 160 (5.0 C/min, 2 min) followed by a 5.0 C/min gradient to 200 C (5.00 min). For the compounds, 2-cyclohexen-1-one, 2-cyclopenten-1-one, 2-methyl-2cyclohexen-1-one, carvone, 3-methyl-2-cyclohexen-1-one, 1-octen-3-one, trans-2decenal, trans, trans-2,4-decadienal, phorone, citral and 1-nitro-1-cyclohexene, the reduced products were compared with commercially available authentic standards. The reduced products of 6-substituted 2-cycl ohexenones had been previously characterized in our laboratory.423 The reduced products of trans-nitrostyrene and 2-nitro-1phenylpropene were prepared by reduction with sodium borohydride as described elsewhere.424 For the rest of the compounds tested in this study, the reduction of the double bond was inferred by GC-MS analysis. Results and Discussion Isolation of YNL134c In our preliminary attempts to isolate the EI and EII enone reductases, we used commercially available bakers yeast and we followed the purification protocol described by Wanner and Tressl with two major modifications: i) a strong anion exchange column (Q Sepharose Fast Flow) was used instead of the weak anion DEAE Sepharose CL-6B, which was employed in the original protocol, ii) a shorter gel filtration column (25 cm

PAGE 213

196 instead of 60 cm) was available for the last purification step. Our observations are summarized below: at the first and the most crucial step of the anion exchange chromatography we never succeeded in separating two fractions, EI at the beginning of the NaCl gradient and EII by the end of the gradient, one with NADPH activity and a second with NADH activity towards the reduction of 1-octen-3-one, as described in the published procedure. Even though the exact gradient of NaCl is not reported, we tried different gradients and in any case we always obtained one active fraction that accepted both NADPH and NADH. at the affinity chromatography step, ac tivity observed was mainly in the flow through and not in the protein fraction that it was bound to the dye. Even though it is not quite clear from the description gi ven by Wanner and Tressl in which fraction they observed activity, it is more likely that active fractions were detected during the NaCl elution. after the last purification step (i.e. the gel filtration column), the most intense band on the SDS-PAGE gel had a molecular weight at 40-45 kDa and a number of other bands were also visible (data not shown). when the purification procedure lasted more that 10 days, loss of activity was observed as we reached the final purification step. But in those cases that we completed the procedure and obtained active fractions, GC-MS analysis of reduced 1-octen-3-one revealed that the carbonyl moiety remained intact and only the double bond was reduced. To the above remarks we should add the following: OYE2 and OYE3 from bakers yeast, cloned and purified by us as GST-fusion proteins, were also found active towards the reduction of 1-octen-3-one under identical to the protein purification assay conditions, despite the nicotinamide cofactor used. Both OYE2 and OYE3 have a molecular weight of ~45kDa. To this end, we decided that it would be wiser to repeat the published purification protocol starting from a yeast strain missing the two genes encoding OYE2 and OYE3 (i.e., DAY1228). To confirm that the double knock-out strain still possesses enone reductase activity, we carried out whole-cell biotransformations using either 1-octen-3one or 2-cyclohexenone as a substrate. After in cubating yeast cells for one day at 30 C with the appropriate substrate, reactions were analyzed by GC-MS (Table 5-4). To our

PAGE 214

197 surprise, the oye strain was unable to reduce the 2-cyclohexenone, whereas it still transformed 1-octen-3-one to the corresponding saturated ketone. Table 5-4. Observed reduction of 1-octen-3-one or 2-cyclohexenone. Strain BYa Yes not determined DAY111b Yes Yes DAY128 Yes No EI enone reductasec Yes Yes EII enone reductasec Yes Yes a bakers yeast from grocery store. b background strain used by David Amberg for construction of DAY128 strain. c data reported by Wanner and Tresll. Only one explanation can accommodate both our data and Wanner and Tressls reported observations (last two entries in Table 5-4): the YHR179w (OYE2) and YPL171c (OYE3) are the only two genes in S. cerevisiae genome encode enone reductases able to reduce 2-cyclohexenone.The EI and EII protein samples prepared Wanner and Tressl must be contaminated by OYE2 and/or OYE3, even though these proteins are not visible in their SDS-PAGE gel. The possibility that an enone reductase different from OYE2 or OYE3, but also distinct from the one responsible for reducing the 1-octen-3-one in the DAY128 strain, was present in the yeast strain used by Wanner and Tressl but not expressed in DAY128 strain can be easily excluded since the parent DAY111 strain does reduce the 2-cyclohexenone. Based on these results, it was apparent to us that a third enone reductase (in addition to OYE2 and OYE3) is also responsible for the reduction of 1-octen-3-one. Convinced that a third enone reductase is present in DAY128 strain, we followed the purification procedure described in detail in the experimental section, which is still based on the reported protocol for the EI and EII enone reductases. In fact, in this case, we also prepared a weak anion exchange column using the matrix that Wanner and Tressl have been utilized before. However, neither this column nor the strong anion

PAGE 215

198 exchange column that we have used before afforded two different active fractions. In any case, only one NADHand NADPH-active frac tion was detected closer to the end of the NaCl gradient (data not shown). In addition, as we had observed before, the most active fractions after the affinity column were found in the flow through (fraction FI in Table 5-5) whereas the activity in the fractions bound to the blue dye was negligible (fraction FII in Table 5-5). After the Superdex 200 filtration column, the last step in Wanner and Tressls purification protocol, the picture of the SDS-PAGE gel was quite similar to the pattern we had observed in all our previous attempts: an intense protein band with molecular weight ~44kDa in presence of numerous other bands (Figure 5-3). In addition only 15-20-fold purification had been accomplished after this step compared to 165-257-fold purification accomplished by Wanner and Tressl. Figure 5-3. Protein gel after the Superdex 200 filtration column. Abbreviations: M :protein markers S: protein sample bearing NAD(P)H activity. We therefore desalted the active protein fraction and loaded it onto a third strong anion exchange column. To our satisfaction, this step afforded a 60-78-fold purification and two major bands in the SDS-PAGE (Figure 5-4). Unfortunately, when the protein sample was electroblotted onto a PVDF membrane, a number of additional protein bands became visible (data not shown).

PAGE 216

199Table 5-5. Purification table for YNL134c. Fraction Volume (mL) Protein (mg) Total Units (NADPH) Total Units (NADH) Specific Activity in mU/mg (NADPH) Specific Activity in mU/mg (NADH) Purification Factor (NADPH) Purification Factor (NADH) Crude extract 220 2860 ---a ----------Streptomycin sulfate precipitation 220 1626 31.9 31.1 19.6 19.1 ----Sephadex G-25 70 1050 18.2 16.2 17.3 15.5 1 1 DEAE CL-6B 62 371 20.5 16.5 55.2 44.2 3.2 2.8 HiLoad Q Sepharose 45 261 12.6 12.6 49 49 2.8 3.2 DyeMatrex Gel Blue A (Fraction FI) 50 54 4 8.8 74 163 4.3 10.5 DyeMatrex Gel Blue A (Fraction FII) 50 50 0.3 0.5 6 10 ----Superdex 200 20 20 5 6.5 252 324 14.6 20.9 MONO Q 5 5 2.5 2.9 1042 1217 60.2 78.5 Sephacryl S-300 0.7 0.17 0.08 0.05 470 294b ----a We were unable to measure the specific activity in the cr ude extract and the protein solution after the streptomycin precipit ation because of a strong precipitate formed in the enzyme assay solution even when it was passed through a 0.45 m filter. bSpecific activity is lowed in this step compared to the previous one ma inly because it was performed 2 weeks latter and enzyme activity was lost during the storage at 4 C.

PAGE 217

200 Nonetheless, an N-terminal analysis by Ed man degradation chemistry attempted for the most intense in Figure 5-4 protein band (~40kDa).No amino acid signal was obtained, suggesting that most likely, the N-terminus amino acid was blocked. In order to further enrich the protein sample for the enone reductase of interest, we loaded the mixture on a Sephacryl S-300 Hi Prep filtration column. Three peaks were detected during the protein elution (Figure 5-5), but only the last one (fractions 72-82) showed NAD(P)Hdependent activity towards 1-octen-3-one reduction as confirmed by both spectrophotometric assays and GS-MS analysis. Active fractions were individually concentrated and subjected to trichloroacetic acid (TCA) precipitation before being loaded onto an SDS-PAGE gel. Even though only the electrophoresis analysis of the fraction 76 is shown in Figure 5-5, all active fractions possessed the same composition with one major and one minor band. Both protein bands were excised from the gel, treated with a protease (trypsin) and subjected to LC-MS analysis, separately. The MASCOT search engine was used for correlation of the experimental MS data with proteins from primary sequences databases. In both cases, the major hit (score> 500) was the product of the YNL134c gene. Peptides found to match with the original sequence are highlighted in red and blue in Figure 5-6. Therefore, the minor band with a lower molecular weight is most likely derived from the major after proteolysis. Two additional hits were derived from this analysis: i) the cytoplasmic alanyl-tRNA synthetase (major band, score=185, 4 peptides matched the subject) ii) the phosphoribosylaminoimidazole-succinocarbox amide synthase (minor band, score=157, 5 peptides matched the subject). The results from the Mascot search engine strongly suggested that the gene encoding the third unknown enone reductase in DAY128 strain is YNL134c. Indeed, an independent study Yamamoto and Kimoto425 reached a similar conclusion by following a different approach. They purified and sequenced an enone reductase (MW ~ 42 kDa)

PAGE 218

201 from the fungus Kluyveromyces lactis using methyl vinyl ketone as a screening substrate. Then, based on a BLAST search and using the S. cerevisiae genome as a query, they identified three hypothetical proteins: YNN4 (YNL134c), YL460 (YLR460c), and YCZ2 (YCR102c) with high similarity (i.e ~68%) to the isolated one. They cloned and overexpressed the three putative reductases in E. coli and they verified their enone reductase activity using E. coli crude extracts based on the calculated specific activity for reducing methyl vinyl ketone. Higher activity was observed in the crude extract containing the overexpessesd YNL134c. Figure 5-4. Protein elution pattern from Mono Q column during the NaCl gradient (fractions 10-50) and protein gel for active fractions (29-34). Abbreviations: M: protein markers, S: protein sample bearing NAD(P)H activity. Figure 5-5. Protein elution pattern from Sephacryl S-300 Hi Prep filtration column and protein gel for active fractions (72-82). Abbreviations: M: protein markers, S: protein sample bearing NAD(P)H activity Red dots indicate the major (big dot) and the minor (small dot) protein bands submitted for LC-MS analysis.

PAGE 219

202 Figure 5-6. Sequence of YLN134c. Peptides found to match with the subject are highlighted either by red or blue color. It is probably not a surprise that the YLR460c was not detected in our sample since the Expasy database reveals that the later is present only at 799 molecules/ cell whereas the YNL134c is present at 3490 molecules/ yeast cell. The fact that OYE2 is at much higher level (15100 molecules/cell) virtually assures that it would be difficult to isolate pure YNL134c from a wild type yeast st rain, especially if we consider that both are NAD(P)H dependent enzymes with high activity for the screening substrate. Interestingly, during the purification of Gre2p, an NADPH-dependent methylglyoxal reductase, Chen et al. observed on a SDS-PAGE that the protein band corresponding to YNL134c was also enriched but it did not possess catalytic activity towards the screening substate.426 Other physical interactions studies revealed that YNL134c interacts with its homolog YCR102c,427 with the ribonucleotide-diphosphate reductase428 and with a domain of vacuolar ATPase.429 Nonetheless, YNL134c is listed as an uncharacterized ORF with unknown function in S. cerevisiae database. Our results for

PAGE 220

203 the substrate specificity of YNL134c will be discussed later in comparison with the other enone reductases. Amino acid sequence similarity search reveals a number of homologs in other fungi species with high identity to YNL134c. Most of them are annotated as putative zinc-binding dehydrogenases and/or NADPH quinone reductases. The most interesting finding is that YNL134c possesses a low but significant percent identity with FAQR, the enone oxidoreductase from strawberries described in the previous chapter. For comparison, a multiple sequence alignment of YNL134c, YCR102c, YLR460c and FAQR is depicted in Figure 5-7. We also noticed an unusual NAD(P)H binding motif in all three yeast sequences shown: instead of (G/A)-XX-G -XX-G the underlined glycine residue has been replaced by a threonine. Figure 5-7. Multiple alignment of YNL134c subfamily with FAQR amino acid sequence. The G-XX-T-XX-G is highlighted in green. Cloning and Expression of Alkene reductases Construction of universal E.coli expression vector The glutathione S-transferase gene was excised from pIK1 (-) vector as an AseI, NcoI fragment and cloned between the NdeI and NcoI sites of pET22b (-) to yield the expression vector pDJB3 (Figure 5-8).

PAGE 221

204 Figure 5-8. Construction of E. coli expression vector pDJB3. Construction of a yeast shuttle universal E.coli expression vector A fragment bearing the T7 terminator, the GST gene, the T7 promoter and the lacI gene (TT7-GST-PT7-lacI) was amplified from the pIK2 vector using appropriate primers that included sequences homologous to BstEII/HindIII-linearized pYEX4T-1-Rec. DOM. vector. The oligonucleotide primer for the 5 d (TTTCACACAGGAAAC AGCTATGACCATGATTACGCCTTTCAGCAAAAAACCCCTCAAGACC) included 36 residues homologous to Hind III end of the linearized pYEX4T-1-Rec. DOM. vector. The reverse oligonucleotide sequence 5 d(GTAATGTCTGCCCCTAAGAAGATCGTCCTTTTGCCAGCTTTCCAGTCG GGAAACCTGTCG) included 36 residues homologous to BstEII end of the linearized vector. Co-transformation of the PCR-amplified product flanked by the homologous regions with the BstEII/HindIII-linearized pYEX4T-1-Rec. DOM. vector into S. cerevisiae BY4742 strain or into E. coli KC8 (recA+) strain was successful and afforded the yeast shuttle E.coli expression vector pDJB2 (Figure 5-9). Both E.coli expression vectors (pDJB2 and pDJB3) possess features for propagation and selection in E.coli BL21 (DE3) strain and facilitate the cloning of the desired gene at the C-terminus edge of the GST with the same set of primers. It is noteworthy that the pDJB2 is a higher copy plasmid

PAGE 222

205 (pUC ori) compared to pDJB3 (pBR322 ori). Moreover, the pDJB2 bears features for propagation (2 micron origin) and selection (URA3 marker) in yeast. Cloning of GST-alkene reductases The gene encoding the OYE1 from S. carlsbergensis was excised from pET-OYE as an NdeI fragment and cloned at this site in pIK2 to afford pDJB5. (Figure 5-10) The appropriate orientation of the gene was confirmed by restriction enzyme digestion analysis. Figure 5-9. Construction of the E. coli expression vector pDJB2.

PAGE 223

206 Figure 5-10. Construction of pDJB5. The homologous recombination approach that we followed in order to clone the rest of alkene reductases is depicted in Figure 5-11. The pDJB3 vector was linearized using the HindIII and NcoI restriction enzymes. In case of pDJB2, the pair of HindIII and Ecl136II (M. SacI in Figure 5-11) was used instead, since both the NcoI and NdeI restriction sites are also present in the backbone of this plasmid. Co-transformation of the PCR-amplified product flanked by the homologous regions of pDJB2(3) in either yeast or E. coli KC8 strain afforded the final construct pDJBX. The overhand region between the Ecl136II and the NcoI was eliminated during the in vivo homologous recombination in yeast. The expression vector directed the synthesis of fused polypeptide in E. coli under control of the T7 promoter. The HR in yeast led typically to thousands of recombinants whereas after the HR in E. coli KC8 strain, we observed any number in between 1-100 of recombinants. We never failed in obtained recombinants in E. coli and four to six out of the six colonies typically screened, by colony PCR had the desired structure. The bacterial procedure is much faster: two to three days are required for yeast colonies to grow in minimal media compared to 16 h in for E. coli cells and no additional steps (i.e. plasmid rescue) other than a plasmid preparation for sequencing and transformation into E. coli BL21 (DE3) are required. Because of these advantages, the in vivo HR in E. coli was the method of choice for the remainder of the project.

PAGE 224

207 Figure 5-11. Homologous recombinational cloning into universal ve ctors pDJB2 and pDJB3. Only part of the final construct is s hown in this figure.

PAGE 225

Primers used for amplification of target genes are listed in Appendix C and the final plasmid constructs, in linear format, are listed in Table 5-6. Table 5-6. Final plasmid constructs of akene reductases. Plasmid Plasmid construct in lineal format pDJB8 pDJB9 pDJB11 pDJB13 pDJB15 pDJB17 pDJB19 pDJB21 pDJB22

PAGE 226

209 Table 5-6. Continued. Plasmid Plasmid constructs in lineal format pDJB23 pDJB24 pDJB25 pDJB26 pDJB27 pDJB29 Expression of GST-alkene reductases The expression and purification protocol for the GST fusion proteins is described in the experimental section and no additional effort was made to optimize these conditions for individual proteins. We failed to express only the OPR2 gene (Figure 5-12) and the lowest purification yield was obtained for OYEA. In two cases (OYE1 and SeOYE), a second protein band at the size of free glutathione transferase (26-28 kDa) is also copurified. The SDS-PAGE for the partially purified, GST-fusion alkene reductases is depicted in Figure 5-13. Sample used for preparation of this protein gel have been purified six to eighteen months before and kept at -20 C as 50 % glycerol solutions. Therefore a partial degradation in some cases could not be excluded and may explain

PAGE 227

210 the other lower molecular weight, proteins bands which are visible in some cases. We also noticed that the YNL134c was co-purified with another protein with higher molecular weight. Figure 5-12. SDS-PAGE analysis of total cell extracts overexpressing (8h postinduction) alkene reductases. Abbreviations: M, protein markers. Figure 5-13. SDS-PAGE analysis of alkene reductases after the affinity column. Screening of Alkene Reductases In principle, upon the completion of the cloning step, the obtained biocatalysts could be screened either as pure GST-fusion enzymes or as small scale (3-5 ml) wholecells biotransformations. Even though the latter approach is more convenient and faster, since the protein purification step could be omitted, side reactions from other enzymes

PAGE 228

211 present in E. coli or production of other metabolites408 may complicate the analysis of the data. In fact, it was the lack of stability of both substrate and enzyme in the crude yeast lysate that forced Phizicky and co-workers to include the affinity column chromatography step in their strategy before the screening. Nonetheless, we evaluated this approach for two substrates: the 2-cyclohexen-1-one and 2-phenyl-2-butenal. For the former substrate, even though reduction of the carbonyl group was not detected, reduction of double bond was detected even in an E. coli strain overexpressing an enzyme that in a pure form did not possess catalytic acti vity. When the unsaturated aldehyde was employed, reduction of the double bond as well as reduction of carbonyl moiety was observed, which in some cases led to formation of three products (data not shown). Those observations suggest that an alkene reductase as well as an aldehyde reductase present in E. coli may compete with some overexpressed alkene reductases for the same substrate, at least in these small scale reactions. Therefore, we switched to 1 ml screening reactions with partially purified GST-alkene reductases, and 5 mM of substrate as a more reliable method for evaluating new biocatalysts. In most cases, the biotransformations were carried out at 30 C with gentle mixing, extracted with ethyl acetate and analyzed periodically. Our results from ca. 600 small scale reactions are summarized in the following tables. Colored bars indicate conversion, usually after 18-24 h incubation. In all cases, the conversion is related to the reduction of the carbon-carbon double; reduction of carbonyl moiety was not observed in these reactions. Black and white bars indicate enantiomeric or diastereomeric excess values. In all bar graphs, the alkene reductases are presented with the same order (yellow bars for: OYE1, OYE2, OYE3, OYEA, and oye, blue bars for: NemA, PpNemA, PpOYE, SeOYE, green bars for: OPR1, OPR3 and Leopr, purple bar for Ltb4dh and orange bar for YNL134c). OYEB was not included since, with only one exception, did not s how any catalytic activity under the reaction

PAGE 229

212 conditions described in the experimental part with any of the substrates used in this study. The reduction of the 2-cyclohexen-1-one, a generic substrate for the OYE family, and 2-cyclopentenone from the OYEs presented in Table 5-7 clearly indicated that all the GST-tagged proteins are catalytically active. A small but measurable conversion was also detected for the same substrates with Ltb4dh, previously reported as nonreactive. Lower conversion was observed with OYEs from plants for 2-cyclopentenone compared to 2-cyclohexenone even though the former is the core skeleton of the OPDA, their known physiological substrate. This suggests that the longer alkyl substituents in OPDA enhance the catalysis. Mostly yeas t OYEs are capable of reducing the two lactones listed in Table 5-7 whereas bacteriaOYEs showed diminished activity, if any. In Table 5-8, most of cyclo-enones listed bear a methyl group at the -position. The stereochemical outcome of this biotransformation, for all the OYEs tested here, followed the mode described by Swiderska and St ewart for OYE1, with no exception. A crystal structure of OYE with 2-cyclohexenone exists only for pentaerythritol tetranitrate reductase (PETN). We built a methyl group at C-2 in order to visualize this transformation. In the proposed model (no energy minimization has been applied), depicted in Figure 5-14, protonation by the Tyr residue should afford the R-enantiomer in any of the relevant substrates listed in Table 5-8. In contrast, reduction of 3-methyl-2cyclohexenone should afford the S-enantiomer (Figure 5-15). Unfortunately, the constraints implied by the planar flavin structure, the strong requirement for a proper alignment of the carbon-carbon above and in clos e proximity with the N-5 atom of flavin and the orientation of the carbonyl moiety in such a way that hydrogen bonds could be made with the conserved Asp (or His) and His amino acid residues diminished our chances to observe diversity among the OYE family members for these type of compounds.

PAGE 230

213 Table 5-7.Unsubstituted cyclo-alkenes. Substrate (name) % Conversion O 2-Cyclohexen1-one 0 20 40 60 80 100 2-Cyclopenten1-one 0 20 40 60 80 100 5,6-Dihydro2H-pyran-2one 0 20 40 60 80 100 (2(5H)Furanone) 0 20 40 60 80 100O Y E1 O YE2 O Y E 3 OYEA Oye Nem A P pN emA Pp O YE SeOYE O P R 1 O P R 3 L eo pr L t b4dh Y N L134c Table 5-8. -substituted cyclo-alkenes. Substrate % Conversion 2-methyl-2cyclohexen-1one 0 20 40 60 80 100

PAGE 231

214 Table 5-8. Continued. Substrate % Conversion (R)-(+)Carvone 0 20 40 60 80 100 O (S)-(-)Carvone 0 20 40 60 80 100 O O oxoisophorone 0 20 40 60 80 100OYE1 OYE2 OYE3 OEYA Oye NemA ppNemA ppOYE SeOYE OPR1 OPR3 LeOPR Ltb4dh YNL134c 2-Methyl-2cyclopenten-1one 0 20 40 60 80 100 3-Methyl2(5H)furanone no conversion (R)-Pulegone no conversion

PAGE 232

215 Figure 5-14. Predicted stereochemical outcome for 2-methyl-2-cyclohexenone. Figure was rendered in Pymol. Figure 5-15. Predicted stereochemical outcome for 3-methyl-2-cyclohexenone. Figure was rendered in Pymol. Asymmetric transformations of -substituted cyclohexenones with whole cells of Synechococcus elongatus strain PCC 7942 were reported to afford the opposite enantiomer from the one we obtained with the purified OYE from the same microorganism.432 Most likely, another enone reductase, which does not belong to OYE family, was responsible for those biotr ansformations. Ltb4dh showed a small but measurable conversion with certain compounds but with the same stereoselectivity as the OYEs. The results reported for oxoisophorone, in Table 5-8, are after 3 h incubation

PAGE 233

216 since prolonged reaction-time lead to product racemization under the reaction conditions. The exocyclic double bond, in s-cis conformation to carbonyl group, in pulegone was not reduced by any of the alkene reductases tested. The dramatic decline in the catalytic activity of OYE1, when 3-methyl-2cyclohexenone was used as substrate, had been pinpointed by Massey. Unfortunately, this situation is the same or even worse is the case for all additional OYEs screened in this work (Table 5-9). Massey hypothesized that this compound somehow interfered with the NADPH reduction of the flavin. His studies on manipulating the redox potential of the flavin by site-directed mutagenesis (T27A, Thr residue is shown in blue in Figure 120) revealed that by lowering the redox potentia l, the rate of oxidative-half reaction of OYE1 with this substrate was increased by 3-fold (over the wild type rate, 0.062 s-1).The same fold increment was observed for 2-cyclohexenone (102 s-1 for the wild type) Considering also his measurements about the rate of reductive-half reaction for this mutant (0.43 s-1) we may predict that the T27A mutation should accelerate the overall reaction since for this particular substrate the oxidative-step is the rate-limited. However, even with this acceleration, the total turnover number will remain low compared to other substrates. Based on the above observations, Massey concluded that steric hindrance during the hydride transfer should also account for the overall phenomenon. We also explored substitution at the 6 position of cyclohexenones (Table 5-10). In this case, we are asking for a kinetic resolution of the racemic 6-substituted cyclohexenones and therefore in the best case the yield of reaction should not exceed 50%. Nonetheless, this is the first and the only example for which some diversity observed among the OYEs for cyclo-alkenes. In fact, the yeast-OYEs showed the opposite stereopreference from the one observed for bacteria and plants-OYEs.

PAGE 234

217 Unfortunately, the enantiomeric ratio (E) derived from the calculated ee of the product and the corresponding conversion was <10 in all cases examined. An E >15 is considered as respectable value for further applications. Table 5-9. -substituted cycloalkenes. Substrate % Conversion O 3-Methyl-2cyclohexen1-one 0 20 40 60 80 100O Y E 1 O Y E 2 O Y E 3 O E Y A o y e N e m A p p N e m A p p O Y E S e O Y E O P R 1 O P R 3 L e O P R L t b 4 d h Y N L 1 3 4 c Isophorone no conversion O O 5,6-dihydro4,6,6trimetylpyranone no conversion Table 5-10. 6-substituted cyclohexenones. Substrate % Conversiona,b O 6-methyl-2cyclohexen1-one -100 -80 -60 -40 -20 0 20 40 60 80 100OYE1 OYE2 OYE3 Oye NemA OPR1 OPR3

PAGE 235

218 Table 5-10. Continued. Substrate % Conversion O 6-ethyl-2cyclohexen1-one -100 -80 -60 -40 -20 0 20 40 60 80 100OYE1 OYE2 OYE3 Oye NemA OPR1 OPR3 6-allyl-2cyclohexen1-one -100 -80 -60 -40 -20 0 20 40 60 80 100OYE1 OYE2 OYE3 Oye NemA ppNemA OPR1 OPR3 a Black bars indicate the %ee of product for the (S) enantiomer in case of methyl and ethyl substituent and (R) enantiomer for the allylic substituent. The opposite is true for the white bars. bEnzymes not listed, were not screened. In the compounds listed in Table 5-11, the exocyclic carbonyl group is more flexible and can adopt an s-trans or an s-cis configuration. Even though an s-trans conformation is more likely preferred from OYEs, a favorable s-cis configuration in case of Ltb4dh could explain its high activity to wards those compounds. From the conversion values, observed in the first and second entry, we may conclude that enals are more reactive with OYEs in contrast to what it was observed for the Ltb4dh. In addition, an enhanced reactivity observed in all cases w hen the cyclohexene ring was substituted at 4 position, but only in case of (S)-perillaldehyde for the OYEs, in contrast both enantiomers were proved excellent substrates for Ltb4dh.

PAGE 236

219 Table 5-11. Perillaldehyde and derivatives. Substrate % Conversion O H 1-Cyclohexene1carboxaldehyde 0 20 40 60 80 100 1-Acetyl-1cyclohexene 0 20 40 60 80 100 (S)Perillaldehyde -100 -80 -60 -40 -20 0 20 40 60 80 100 (R)Perillaldehyde -100 -80 -60 -40 -20 0 20 40 60 80 100OYE1 OYE2 OYE3 OEYA oye NemA ppNemA ppOYE SeOYE OPR1 OPR3 LeOPR Ltb4dh YNL134c H O (-)-Myrtenal no conversion

PAGE 237

220 The two enantiomers of perillaldehyde allowed us to trace the stereopreference of the alkene reductases, since the obtained products are diastereoisomers. Bioreduction of perillaldehyde with bakers yeast afforded the trans saturated alcohol for the natural (S)-perillaldehyde in 78% de and the cis saturated alcohol for the synthetic (R)perillaldehyde in 60% de.431 We also observed the same pattern for all the OYE family members. The stereochemical outcome of this biotransformation can be also rationalized as before (Figure 5-16 and 5-17). Building the perillaldehyde structure on the core of cyclohexenone in the crystal structure of PETR (no energy minimization was applied) we can easily predict that the stereochemistry of the obtained products should be that reported by Serra and co workers. The major differences in our case are that only the carbon-carbon double bond is reduced and higher de values obtained for both trans and cis products. However, lower conversion and % de observed for the synthetic (R)-perillaldehyde compared to (S) natural compound in almost all cases. It also apparent from the graph that the lower conversion values are linked with lower de numbers. The loss of stereocontrol for the R enantiomer may be due to steric hindrance with the planar flavin cofactor and this may explain both the lower de and the lower conversion. The explanation provided by Serra for the lower de values obtained with bakers yeast was that more than one enone reductase compete for the same substrate. We cannot exclude this possibility but definitely this other reductase is not the YNL134c, which did not show any catalytic activity with this substrate under our screening conditions. A flipped geometry of (S) perillaldehyde in the catalytic center of the OYE as it is depicted in Figure 5-18 should lead to the cis diastereroisomer and therefore should be excluded at least for this enantiomer, but it may be an alternative explanation for the low de values observed for the (R) enantiomer.

PAGE 238

221 Figure 5-16. Predicted stereochemical outcome for (S)-perillaldehyde. Figure was rendered in Pymol. Figure 5-17. Predicted stereochemical outcome for (R)-perillaldehyde. Figure was rendered in Pymol.

PAGE 239

222 Figure 5-18. An alternative, flipped geometry of (S)-perillaldehyde in the catalytic center of OYE built based on the crystal structure of OYE1 with para-hydroxybenzaldehyde. Tyr375, colored in red, could serve as a hydrogen donor. Figure was rendered in Pymol. No energy minimization was applied. More interesting are the results obtained for perillaldehyde with Ltb4dh. Stereoselectivity studies with alkenal/ one oxidoreductase (AOR) and with most of medium-chain reductases have not been reported so far. Unfortunately, the crystal structure of AOR from rat has not been solved and therefore we will try to explain our data based on the structure of AOR from guinea pig with 15-oxo-prostaglandin E (Figure 5-19A). This structure suggested an s-cis conformation of the carbonyl group and a hydride attack from the re face of perillaldehyde. It is less clear, however, from where the -protonation occurred. The cis-1,2 dihydroperillaldehyde obtained, when the Rperillaldehyde was utilized, suggested a trans-addition across the double bond (Figure 519B). In contrast, for the same product, a syn addition must be hypothesized for the Senantiomer (Figure 5-19C), if we assume a similar conformation for both enantiomers in the catalytic center of AOR. The possibility that a different amino acid (or a water

PAGE 240

223 molecule) acts as a proton donor, depending on the substrate used, is open and in any case additional experiments are required for the elucidation of this mechanism. Figure 5-19. Proposed mechanism for the Ltb4dh-mediated reduction of perillaldehyde. A. Close-up view of Ltb4dh from guinea pig with 15-oxo-PGE2. B. Proposed hydrogen addition for the R-enantiomer. C. Proposed hydrogen addition for the S-enantiomer. The last group of cycloalkenes examined in the present study is listed in Table 512. All three maleimides were good substrates for the OYEs and even the Ltb4dh possesses some activity, especially with the two more hydrophobic compounds. Table 5-12. N-substituted maleimides. Substrate % Conversion maleimide 0 20 40 60 80 100 NPhenylmaleimide 0 20 40 60 80 100

PAGE 241

224 Table 5-12. Continued. Substrate % Conversion N O O 2-Methyl-Nphenylmaleimide 0 20 40 60 80 100OY E 1 OYE2 OYE3 OY E A Oye N emA PpNemA PpOYE S e OY E O P R1 OP R3 Le o p r Ltb4d h Y NL1 34 c YNL134c, unreactive with all cyclo-alkenes presented above, was able to reduce most of the acyclic alkenes listed in Table 5-13. Trans dienals were not good substrates for the OYEs from plants and their reactivity was modest towards trans-2-decenal. Methyl trans-4-oxo-2-pentenoate, and 1-octen-3-one were proved to be the two universal substrates for this alkene reductase library. In fact, the former was the only substrate for which a small conversion was detected (~12%) even with the OYEB. Interestingly, all reductases, except YNL134c showed high activity towards the dimethyl maleate. Lastly, both double bonds of phorone were reduced by Ltb4dh whereas no activity was observed from any of the O YEs. The conversion reported for the same compound for the YNL134c corresponds to a product in which only the one double bond was reduced. Table 5-13. Straight chain alkenes. Substrate % Conversion O 1-octen-3-one 0 20 40 60 80 100 trans-2-decenal 0 20 40 60 80 100

PAGE 242

225 Table 5-13. Continued. Substrate % Conversion trans,trans-2,4hexadienala 0 20 40 60 80 100 trans,trans-2,4Decadienala 0 20 40 60 80 100 methyl trans-4-oxo-2pentenoate 0 20 40 60 80 100 dimethyl maleate 0 20 40 60 80 100 phorone 0 20 40 60 80 100O Y E1 OY E2 OYE3 OYEA Oye Nem A PpNemA P p O YE S eO Y E OPR1 OPR3 Leopr L t b 4dh YN L 13 4 c aOnly the double bond conjugated to carbonyl group was reduced. The reduction of the two geometric isomers of citral, geranial (trans) and neral (cis), allowed us to examine the stereoand enantio-specificity of the reductases under study. Citral is a 60:40 mixture of trans and cis isomers, which, separately, are not commercially available.

PAGE 243

226 Table 5-14. Substituted enals. Substrate % Conversion citrala -100 -80 -60 -40 -20 0 20 40 60 80 100 geranialb -100 -80 -60 -40 -20 0 20 40 60 80 100OYE1 OYE2 OYE3 OEYA Oye NemA ppNemA ppOYE SeOYE OPR1 OPR3 LeOPR Ltb4dh YNL134c neralb -100 -80 -60 -40 -20 0 20 40 60 80 100OYE1 OYE2 OYE3 OEYA Oye NemA ppNemA ppOYE SeOYE OPR1 OPR3 LeOPR Ltb4dh YNL134c aBlack bars indicate % ee for (R) enantiomer and white bars % ee for (S) enantiomer bFor OYE2 and SeOYE twice more enzyme was used compared to what was used for citral. For ppOYE, OPR1, OPR3, LeOPR and Ltb4dh the amount of each was increased 3 fold compared to what was used for citral.

PAGE 244

227 Table 5-14. Continued. Substrate % Conversion 2-phenyl-2-butenalc 0 20 40 60 80 100 methylcinnamaldehyde -100 -80 -60 -40 -20 0 20 40 60 80 100OYE1 OYE2 OYE3 OEYA oye NemA ppNemA ppOYE SeOYE OPR1 OPR3 LeOPR Ltb4dh YNL134c cConversion values reported after 20 h incubation and % ee after 6h. Our results with the mixture of two isomers (first entry of Table 5-14), indicated a divergent behavior among the OYEs. To a fi rst approximation, yeast OYEs afforded (R)citronellal with good (OYE2) to moderate (OYE3) enantiomeric excess whereas OYEs from bacteria and plants afford the opposite enantiomer with a broad range (60-99%) of ee. The only exception was the ppNemA. If we applied the proposed trans addition of hydrogen across the double bond of citral, as we did before for the cyclohexenones it would become apparent (Figure 5-20) that the trans isomer should lead to (R)-citronellal and the cis isomer to (S)-citronellal. If we assume no isomerization of starting compounds and no preference from the OYEs fr om the one or the other isomer then we should observed a ~20% ee (R)-enantiomer in all cases. Clearly this is not the case. Our results with pure trans or cis enantiomer suggested that the OYEs from plants have

PAGE 245

228 a strong preference for neral whereas yeast-OYEs a strong preference for geranial. Bacterial-OYEs did not follow a consistent pattern: NemA followed the same patern as the OYEs from plants whereas ppNemA followed a similar to OYE3 patter. In all cases, an enhanced enantioselectivity was observed when the correct isomer was employed with the preferred enzyme. This is probably due to fact that when the isomer presented in the reaction mixture is not utilized i mmediately from the enzyme under study, alkene isomerization yielded the opposite, correct isomer which in turn was converted to the appropriate enantiomer. The amino acidmediated isomerization of pure neral or pure geranial to citral is a well-documented phenomenon432 and isomerization of the unreacted starting material was observed in our case too. Loss of stereocontrol, when the wrong isomer is reduced by OYEs, is also possible. With the existent experimental data and with the lack of a crystal structure of an OYE in presence of a straight-chain alkene it is hard to explain the observed distinct stereopreference among OYEs. Figure 5-20. Proposed mechanism for the reduction of geranial or neral by OYEs. A key substrate for the EI and EII enone reductases was the 2-phenyl-2-butenal, since both purified enzymes afforded the corresponding (R) 2-phenyl-butanal in 97% ee. In our hands, prolonged incubation of the above enal with all the enzymes tested led to a racemic product. We never succeeded in obtaining only one enantiomer, even at a lower pH (i.e. 4.8). Also, we did not observe any conversion with YNL234c. The observed enantioselectivity for -methyl-cinnamaldehyde was also low (Table 5-14).

PAGE 246

229 Nitro-alkenes were excellent substrates for OYEs as well as for Ltb4dh (Table 515). Unfortunately, a racemic mixture of 2-nitro-1-phenyl propane was obtained in both pH 7.0 and pH 4.8 for the last entry. Rapid release of the intermediate nitronate before protonation in the catalytic center, or enzyme-mediated racemization of the obtained chiral nitro-alkane, are two possible explanations that we did not investigate further in this work. Table 5-15. Nitro-akenes. Substrate % Conversion 1-Nitro-1cyclohexene 0 20 40 60 80 100 transnitrostyrene 0 20 40 60 80 100 2-Nitro-1phenylpropene 0 20 40 60 80 100OYE 1 O Y E2 O Y E3 OYE A Oye Ne m A P pN e mA PpOYE S eOY E O PR1 OP R 3 Leopr Ltb4 d h Y N L 13 4 c Concluding Remarks and Future Work The aim of this study was dual: production of previously uncharacterized alkene reductases and investigation of their substrate specificity. We successfully created a collection of sixteen clones out of which only one failed to overexpress in E. coli. One of these clones encodes a yeast enzyme with pr eviously uncharacterized function. Our efforts towards the purification of YNL134c from its native host suggested that OYE2 and

PAGE 247

230 OYE3 are the key players for the majority of biotransformation with enones and enals in bakers yeast since the YNL134c showed narrow substrate specificity. Therefore, whenever low stereoselectivities are observed with this microorganism, the most reasonable explanation is racemization of the obtained product or inherent incapability of the OYEs to produce a single enantiomer. The possibility of another alkene-reductase with opposite stereopreference is limited, if any. Our substrate specificity studies among OYEs uncovered: an extremely rigid action mode of all OYEs with or -substituted cyclohexenones probably due to the constraints that the planar of flavin imposes. a promising flexibility when straight chain alkenes were employed. The case of citral is important for two reasons. First and foremost, it underlines the necessity of more than one biocatalyst able to perform a similar transformation. For example, from this collection of OYEs, OYE2 appeared the more suitable biocatalyst for production of (R)-citronellal and NemA for the production of the (S) enantiomer. In principle, after this rapid screening, the next logical step is the scale-up of the process. Microbial transformations that have been explored so far suffer from low yields, mainly because of concomitant reduction of the carbonyl group.433 Gene deletions in E. coli can be easily constructed, as we already discussed in Chapter 3, and thus this bacterium is a promising workhorse for the production of both enantiomers of citronellal in a pure form. Secondly, citral may be proved useful probe for further structure-function studies among the old yellow enzyme family members. In fact, citral is the only natural product after the 12-oxophytodienoic acid for which divergent activity has been observed in this family. Considering the constantly increasi ng number of crystal structures from OYE members the last few years, a co-crystallization of any OYE with neral or geranial seems a feasible task and probably useful since none of the available structures contains a

PAGE 248

231 straight chain activated alkene. In parallel, compounds with similar to citral structures may help us to discover the limits of the observed diversity. In this work, for the first time, the stereoselectivity of Ltb4dh was examined. Even though the substrate specificity of this medi um chain reductase is narrower compared to that of OYEs, it is also less predictable. An important mechanistic detail, the proton donor during the reduction of the perillaldehy de, is missing and therefore experiments with labeled NADPH and site-directed mutagenesis based on the existed crystal structure may clarify this point. It will also be of great interest to explore whether the other AOR homologues possess the same substrate specificity. In conclusion, our decision to analyze genomes from fungi, bacteria, plants and mammals and not to limit ourselves to one species was proved very useful. For example, we could not have observed any dive rsity in this library if we limited to S. cerevisiaes genome. For the expansion of the current library the most challenging task is prediction of putative alkene reductases from the medium chain reductase family.

PAGE 249

232 APPENDIX A ILLUSTRATION OF CONSTRUCTION OF GENE KNOCK-OUTS IN E.COLI

PAGE 250

233

PAGE 251

234 APPENDIX B LIST OF STRAINS AND PLASMIDS Strain or Plasmid Description Source BW25113 (pKD46) lacIq rrnBT14 lacZWJ16 hsdR514 araBADAH33 rhaBADLD78 CGSC#: 7630 BW25141 (pKD13) lacIq rrnBT14 lacZWJ16 phoBR580 hsdR514 araBADAH33 rhaBADLD78 galU95 endABT333 uidA (MluI)::pir+ recA1 CGSC#: 7633 BT340 (pCP20) F( 80lacZ M15) endA1 recA1 supE44 thi-1 gyrA96hsdr17 relA1? (lacZYA-argF)U169 CGSC#: 7629 BL21(DE3) F-, ompT, hsdSB(rB -mB -), dcm+, gal, (DE3) tonA Invitrogen TOP10 F mcrA (mrr-hsdRMS-mcrBC) 80lacZ 15 lacX74 recA1 deoR araD139 (ara-leu)7697 galU galK rpsL (StrR) endA1 nupG Invitrogen KC8 hsdR, leuB600, trpC9830, pyrF::Tn5, hisB463, lac X74, strA, galU, K Clontech BY4742 MAT leu2, ura3, his3, lys2 Dr. Thomas Lyons DAY128 MATa ura3-52 leu2 1 trp1 63 his3 200 oye22::HIS3 oye32::TRP1 Dr. David Amberg

PAGE 252

235 APPENDIX C LIST OF PRIMERS Table C-1. Oligonucleotides used to create pgi mutant. Name (length) 5' to 3' primer sequence Application pKD13F (20nt) GTGTAGGCTGGAGCTGCTTCa pKD13 specific primer used to amplify FRT-kan-FRT cassette pKD13R (20nt) ATTCCGGGGATCCGTCGACC pKD13 specific primer used to amplify FRT-kan-FRT cassette pgiUpF500 (24nt) TTTTCAGCCTTGGCACAAGGGAAG E.coli specific primer used to amplify Upstream500 region pgiDownR500 (23nt) TTTCCCTTCATTGAATGAATGGAG E.coli specific primer used to amplify Downstream500 region pgiDownFpKD13R (45nt) GGTCGACGGATCCCCGGAAT TCATCGTCGATATGTAGGCCGGAT Long chimeric primer used to join FRT-kan-FRT fragment with Upstream 500 region pgiUpR-pKD13F (44nt) GAAGCAGCTCCAGCCTACACb TAGCAATACTCTTCTGATTTTGAG Long chimeric primer used to join FRT-kan-FRT fragment with Downstream 500 region PgiF (24nt) AACATCAATCCAACGCAGACCGCT E.coli specific primer used to detect pgi mutants PgiR (24nt) TTAACCGCGCCACGCTTTATAGCG E.coli specific primer used to detect pgi mutants pgi550F (24nt) CAGACATAACTACCTCGTGTCAGG E.coli specific primer used to detect pgi mutants pgi550R (24nt) AAGTCGCCGCAAGCGCAGATATGG E.coli specific primer used to detect pgi mutants k1 (20nt) CAGTCATAGCCGAATAGCCT kan specific primer used to detect the FRT-kan-FRT insertion k2 (20nt) CGGTGCCCTGAATGAACTGC kan specific primer used to detect the FRT-kan-FRT insertion a Bolding indicates regions of complementarity between oligonucleotides pairs. bItalics indicates reverse complementarity

PAGE 253

236 Table C-2. Oligonucleotides used to create zwf mutant. Name (length) 5' to 3' primer sequence Application pKD13F (20nt) GTGTAGGCTGGAGCTGCTTC pKD13 specific primer used to amplify FRT-kan-FRT cassette pKD13R (20nt) ATTCCGGGGATCCGTCGACC pKD13 specific primer used to amplify FRT-kan-FRT cassette zwfUp500F (24nt) ACAGAAACGATTCACCGTCGGTT C E. coli specific primer used to amplify Upstream500 region zwfDown500R (24nt) TGGATAGTGTTCATAAGGCTGGT G E. coli specific primer used to amplify Downstream500 region zwfUp570F (23nt) ATAGGGAGTGCCATTCGCCAGAC E. coli specific primer used to amplify Upstream570 region zwfDown570R (22nt) CATCACACATGCCCGGAA CACC E. coli specific primer used to amplify Upstream570 region zwfDownFpKD13R (45nt) GGTCGACGGATCCCCGGAATTAA TATCTGCGCTTATCCTTTATGG Long chimeric primer used to join FRT-kan-FRT fragment with Upstream 550 region zwfUpRpKD13F (45nt) GAAGCAGCTCCAGCCTACACGTC ATTCTCCTTAAGTTAA CTAACCC Long chimeric primer used to join FRT-kan-FRT fragment with Downstream 550 region zwfF (24nt) ATGCGGTAACGCAAACAGCCCAG E. coli specific primer used to detect zwf mutants zwfR (26nt) TTACTCAAACTCATTCAAGGAACG AC E. coli specific primer used to detect zwf mutants zwf550F (24nt) TGCCAGATGAAGTTTAAAATCAG G E. coli specific primer used to amplify Upstream550 region zwf550R (24nt) AACCTGACCAACCGCATTCGCTT C E. coli specific primer used to amplify Downstream550 region k1 (20nt) CAGTCATAGCCGAATAGCCT kan specific primer used to detect the FRT-kan-FRT insertion k2 (20nt) CGGTGCCCTGAATGAACTGC kan specific primer used to detect the FRT-kan-FRT insertion

PAGE 254

237 Table C-3. Oligonucleotides used to create pntAB mutant. Name (length) 5' to 3' primer sequence Application pKD13F (20nt) GTGTAGGCTGGAGCTGCTTC pKD13 specific primer used to amplify FRT-kan-FRT cassette pKD13R (20nt) ATTCCGGGGATCCGTCGACC pKD13 specific primer used to amplify FRT-kan-FRT cassette UpPntA510 (29nt) CGCCTTGCGCAAACCAGGTACT GGTATTG E. coli specific primer used to amplify Upstream510 region specific to E. coli BL21(DE3) Down510PntBR (26nt) GGCGATTCAACGTGCGATTGAC AGTG E. coli specific primer used to amplify Downstream510 region specific to E. coli BL21(DE3) PntAF (27nt) GCATACCAAGAGAACGGTTAACC AATG E. coli specific primer used to detect PntA region specific to E. coli BL21(DE3) PntAR (27nt) TTTGCGGAACATTTTCAGCATGC GCTG E. coli specific primer used to detect PntA region specific to E. coli BL21(DE3) PntBF (25nt) ATGTCTGGAGGATTAGTTACAGC TG E. coli specific primer used to detect PntB region specific to E. coli BL21(DE3) PntBR (24nt) TTACAGAGCTTTCAGGATTGCAT C E. coli specific primer used to detect PntB region specific to E. coli BL21(DE3) Up PntA R pKD13F (52nt) GAAGCAGCTCCAGCCTACACG ATATTCCCTTCCATCGGTTTTATT GATGATG Long chimeric primer used to join FRT-kan-FRT fragment with Upstream 510 region specific to E. coli BL21(DE3) Down pntB FpKD13 R (50nt) GGTCGACGGATCCCCGGAATC CCTGACGGCCTCTGCTGAGGC CGTCACTC Long chimeric primer used to join FRT-kan-FRT fragment with Downstream 510 region specific to E. coli BL21(DE3) k1 (20nt) CAGTCATAGCCGAATAGCCT kan specific primer used to detect the FRT-kan-FRT insertion k2 (20nt) CGGTGCCCTGAATGAACTGC kan specific primer used to detect the FRT-kan-FRT insertion

PAGE 255

238 Table C-4. Oligonucleotides used to create the library of En-Reductases Name (length) 5' to 3' primer sequence Application pYEX4T-pIK2 UP (61) TTTCACACAGGAAACAGCTATGACC ATGATTACGCCTTTCAGCAAAAAACC CCTCAAGACC pYEX4T specific primer used for construction of pDJB2 pYEX4T-pIK2 Down (60) GTAATGTCTGCCCCTAAGAAGATCG TCGTTTTGCCAGCTTTCCAGTCGGG AAACCTGTCG pYEX4T specific primer used for construction of pDJB2 pDJB2UpYPL171R (65) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATCAGTTCTTGTTC CAACCTAAATCTACTG pDJB2(3)-tagged Upstream primer used to amplify OYE3 pDJB2NcoIYPL171F (62) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGCCATTTGTAAA AGGTTTTGAGC pDJB2(3)-tagged Downstream primer used to amplify OYE3 pDJB2NcoIOYEAF (67) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGGCTGTTCAAA AGTTATATGAATCGCAG pDJB2(3)tagged Upstream primer used to amplify OYEA pDJB2UpOYEAR (58) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATCACTCAACCTTC TTGGTATCC pDJB2(3)-tagged Downstream primer used to amplify OYEA pDJB2NcoIOYEBF (74) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGAATGACCGAG GTGAATTGTTTAAACCCATCAAAG pDJB2(3)tagged Upstream primer used to amplify OYEB pDJB2NcoIOYEBF (64) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATCACTTTGATTGA AGGAACTCTTTGCTA pDJB2(3)-tagged Downstream primer used to amplify OYEB pDJB2-NemAF (71) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGTCATCTGAAAA ACTGTATTCCCCACTGAAAG pDJB2(3)tagged Upstream primer used to amplify NemA pDJB2-NemAR (63) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATTACAACGTCGG GTAATCGGTATAGCC pDJB2(3)-tagged Downstream primer used to amplify NemA pDJB2-OYE2F (69) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGCCATTTGTTAA GGACTTTAAGCCACAAGC pDJB2(3)tagged Upstream primer used to amplify OYE2 pDJB2-OYE2R (57) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATTAATTTTTGTCC CAACCGAG pDJB2(3)-tagged Downstream primer used to amplify OYE2 pDJB2-oyeF (65) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGTCGTACATGA ACTTTGACCCTAAGC pDJB2(3)tagged Upstream primer used to amplify oye pDJB2-oyeR (72) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATTAGTACTTCTTTT CCTCTTTCTTGTAACCCTTTGC pDJB2(3)-tagged Downstream primer used to amplify oye pDJB2-ltb4dhF (60) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGGTACAAGCTA AGACCTGGAC pDJB2(3)tagged Upstream primer used to amplify Ltb4dh pDJB2-ltb4dhR (63) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATCACGCTTTCACT ATAGTCTTCCCCAG pDJB2(3)-tagged Downstream primer used to amplify Ltb4dh

PAGE 256

239 Table C-4. Continued. pDJB2-OPR1F (65) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGGAAAACGGAG AAGCAAAACAGAGTG pDJB2(3)tagged Upstream primer used to amplify OPR1 pDJB2-OPRR (63) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATTAAGCTGTTGAT TCGAGGAAAGGGTA pDJB2(3)tagged Downstream primer used to amplify OPR1 pDJB2-OPR2F (60) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGGAAATGGTAA ACGCAGAAGC pDJB2(3)tagged Upstream primer used to amplify OPR2 pDJB2-OPR2R (61) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATTATGAAGGTATA ATGACTTAGAGT pDJB2(3)tagged Downstream primer used to amplify OPR2 pDJB2-OPR3F (64) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGACGGCGGCAC AAGGGAACTCTAAC pDJB2(3)tagged Upstream primer used to amplify OPR3 pDJB2-OPR3R (61) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATCAGAGGCGGGA AAAAGGAGCCAAG pDJB2(3)tagged Downstream primer used to amplify OPR3 pDJB2-PPOYEF (68) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGACACAATCAG AACTTTTCAAAACCACTC pDJB2(3)tagged Upstream primer used to amplify PpOYE pDJB2-PPOYER (64) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATTAGGATTTGTAA GTCGGGTAGTCGGTG pDJB2(3)tagged Downstream primer used to amplify PpOYE pDJB2PPNemAF (65) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGAAACTCTTGC AACCGCTGCAAATCG pDJB2(3)tagged Upstream primer used to amplify PpNemA pDJB2PPNemAR (63) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATCAAGCCTGCTTC AGGAACGGATAATC pDJB2(3)tagged Downstream primer used to amplify PpNemA pDJB2-seOYEF (63) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGTCCGAATCGC TCAAACTGCTGAC pDJB2(3)tagged Upstream primer used to amplify SeOYE pDJB2-seOYER (62) AGTGGTGGTGGTGGTGGTGCTCGA GTGCGGCCGCATTAGACAGATGCTG CTTCCAAACTGG pDJB2(3)tagged Downstream primer used to amplify SeOYE pDJB2-LeoprF (61) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGGAAAATAAAG TCGTTGAAGAG pDJB2(3)tagged Upstream primer used to amplify Leopr pDJB2-LeoprR (66) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATCATGTCATGGTT TCTAGAAATGGATAATC pDJB2(3)tagged Downstream primer used to amplify Leopr

PAGE 257

240 Table C-4. Continued. pDJB2-orf14F (61) TCCAAAATCGGATCATCTGGTTCCG CGTCATATGCCCATGTCCGCCTCGA TTCCAGAAACC pDJB2(3)tagged Upstream primer used to amplify YNL134c pDJB2-orf14R (64) CAGTGGTGGTGGTGGTGGTGCTCG AGTGCGGCCGCATTATTTCAAGACG GCAACCAACTTTTCG pDJB2(3)tagged Downstream primer used to amplify YNL134c lacI-START (23) GCGGGATCGAGATCTCGATCCTC pDJB2 specific primer for sequencing lacI region lacI-CONT (24) CTCGCAATCAAATTCAGCCGATAG pDJB2 specific primer for sequencing lacI region GST-END (25) ATGGACCCAATGTGC CTGGATG pDJB2(3) specific primer for sequencing Upstream region of enreductases genes

PAGE 258

241 LIST OF REFERENCES 1. Horecker, B. L. J. Biol. Chem. 2002, 277, 47965-47971. 2. Voet D.; Voet J.; Biochemistry 2nd ed.; Wiley, New York, 1995. 3. Encyclopedia of Eschericia coli K-12 Genes and Metabolism. Accessed on November 2006. http://ecocyc.org/ 4. Sanchez, A. M.; Bennett, G. N.; San, K.-Y. J. Biotechnol. 2006, 117, 395-405. 5. Portais, J.-C.; Delort, A.-M. FEMS Microbiol. Rev. 2002, 26, 375-402. 6. Cohen, S. S. Nature 1951, 168, 746-7. 7. Scott, D. B.; Cohen, S. S. Biochem. J. 1953, 55, 33-6. 8. Wang, C. H.; Stern, I.; Gilmour, C. M.; Klungsoyr, S.; Reed, D. J.; Bialy, J. J.; Christensen, B. E.; Cheldelin, V. H. J. Bacteriol. 1958, 76, 207-16. 9. Rittenberg, D.; Ponticorvo, L. J. Biol. Chem. 1962, 237, C2709-10. 10. Model, P.; Rittenberg, D. Biochemistry 1967, 6, 69-80. 11. Zablotny, R.; Fraenkel, D. G. J. Bacteriol. 1967, 93, 1579-81. 12. Caprioli, R.; Rittenberg, D. Proc. Natl. Acad. Sci. U S A. 1968, 60, 1379-82. 13. Caprioli, R.; Rittenberg, D. Biochemistry 1969, 8, 3375-84. 14. Thomas, A. D.; Doelle, H. W.; Westwood, A. W.; Gordon, G. L. J. Bacteriol. 1972, 112, 1099-105. 15. Orthner, C. L.; Pizer, L. I. J. Biol. Chem. 1974, 249, 3750-5. 16. Csonka, L. N.; Fraenkel, D. G. J. Biol. Chem. 1977, 252, 3382-91. 17. Orozco de Silva, A.; Fraenkel, D. G. J. Biol. Chem. 1979, 254, 10237-42. 18. Wang, C. H. Meth. Microbiol. 1972; 185-230. 19. Katz, J.; Wood, H. G. J. Biol. Chem. 1960, 235, 2165-77. 20. Katz, J.; Wood, H. G. J. Biol. Chem. 1963, 238, 517-23.

PAGE 259

242 21. Katz, J.; Rognstad, R. Biochemistry 1967, 6, 2227-47. 22. Caprioli, R.; Rittenberg, D. Proc. Natl. Acad. Sci. U S A. 1968, 61, 1422-7. 23. Fraenkel, D. G.; Vinopal, R. T. Annu. Rev. Microbiol. 1973, 27, 69-100. 24. Csonka, L. N. J. Biol. Chem. 1977, 252, 3392-8. 25. Lowenstein, J. M. J. Biol. Chem. 1961, 236, 1213-6. 26. Rose, I. A. J. Biol. Chem. 1961, 236, 603-9. 27. Fraenkel, D. G.; Levisohn, S. R. J. Bacteriol. 1967, 93, 1571-8. 28. Wolf, R. E., Jr.; Prather, D. M.; Shea, F. M. J. Bacteriol. 1979, 139, 1093-6. 29. Baker, H. V., Wolf, R. E., Jr. Proc. Natl. Acad. Sci. U S A. 1984, 81, 7669-73. 30. Rowley, D. L.; Fawcett, W. P.; Wolf, R. E., Jr. J. Bacteriol. 1992, 174, 623-6. 31. Pease, A. J.; Wolf, R. E., Jr. J. Bacteriol. 1994, 176, 115-22. 32. Sanford, K.; Soucaille, P.; Whited, G.; Chotani, G. Curr. Opin. Microbiol. 2002, 5, 318-322. 33. Sauer, U. Curr. Opin. Biotechnol. 2004, 15, 58-63. 34. Ogino, T.; Arata, Y.; Fujiwara, S. Biochemistry 1980, 19, 3684-91. 35. Varma, A.; Palsson, B. O. J. Theor. Biol. 1993, 165, 503-522. 36. Varma, A.; Palsson, B. O. J. Theor. Biol. 1993, 165, 477-502. 37. Szyperski, T. Eur. J. Biochem. 1995, 232, 433-48. 38. Tsai, P. S.; Hatzimanikatis, V.; Bailey, J. E. Biotechnol. Bioeng. 1996, 49, 139-50. 39. Frey, A. D.; Fiaux, J.; Szyperski, T.; Wuthrich, K.; Bailey, J. E.; Kallio, P. T. Appl. Environ. Microb. 2001, 67, 680-687. 40. Szyperski, T.; Bailey, J. E.; Wuethrich, K. Trends Biotechnol. 1996, 14, 453-459. 41. Ponce, E.; Martinez, A.; Bolivar, F.; Valle, F. Biotechnol. Bioeng. 1998, 58, 292-295. 42. Follstad, B. D.; Stephanopoulos, G. Eur. J. Biochem. 1998, 252, 360-371. 43. Schmidt, K.; Nielsen, J.; Villadsen, J. J. Biotechnol. 1999, 71, 175-189. 44. Fiaux, J.; Andersson, C. I. J.; Holmberg, N .; Buelow, L.; Kallio, P. T.; Szyperski, T.; Bailey, J. E.; Wuethrich, K. J. Am. Chem. Soc. 1999, 121, 1407-1408.

PAGE 260

243 45. Sauer, U.; Lasko, D. R.; Fiaux, J.; Hochu li, M.; Glaser, R.; Szyperski, T.; Wuthrich, K.; Bailey, J. E. J. Bacteriol. 1999, 181, 6679-6688. 46. Edwards, J. S.; Palsson, B. O. Proc. Natl. Acad. Sci. U S A. 2000, 97, 5528-5533. 47. Edwards, J. S.; Palsson, B. O. BMC bioinformatics 2000, 1, 1. 48. Edwards, J. S.; Palsson, B. O. Biotechnol. Prog. 2000, 16, 927-939. 49. Oh, M.-K.; Liao, J. C. Biotechnol. Prog. 2000, 16, 278-286. 50. Canonaco, F.; Hess, T. A.; Heri, S.; Wang, T.; Szyperski, T.; Sauer, U. FEMS Microbiol. Lett. 2001, 204, 247-252. 51. Emmerling, M.; Dauner, M.; Ponti, A.; Fiaux J.; Hochuli, M.; Szyperski, T.; Wuthrich, K.; Bailey, J. E.; Sauer, U. J. Bacteriol. 2002, 184, 152-164. 52. Edwards, J. S.; Ibarra, R. U.; Palsson, B. O. Nature Biotechnol. 2001, 19, 125-130. 53. Edwards, J. S.; Ramakrishna, R.; Palsson, B. O. Biotechnol. Bioeng. 2002, 77, 2736. 54. Flores, S.; Gosset, G.; Flores, N.; de Graaf, A. A.; Bolivar, F. Metab. Eng. 2002, 4, 124-137. 55. Fischer, E.; Sauer, U. Eur. J. Biochem. 2003, 270, 880-891. 56. Peng, L.; Shimizu, K. Appl. Microbiol. Biotechnol. 2003, 61, 163-178. 57. Kabir, M. M.; Shimizu, K. J. Biotechnol. 2003, 105, 11-31. 58. Kabir, M. M.; Shimizu, Appl. Microbiol. Biotechnol. 2003, 62, 244-255. 59. Brumaghim, J. L.; Li, Y.; Henle, E.; Linn, S. J. Biol. Chem. 2003, 278, 42495-42504. 60. Hua, Q.; Yang, C.; Baba, T.; Mori, H.; Shimizu, K. J. Bacteriol. 2003, 185, 70537067. 61. Zhao, J.; Shimizu, K. J. Biotechnol. 2003, 101, 101-117. 62. Jiao, Z.; Baba, T.; Mori, H.; Shimizu, K. FEMS Microbiol. Lett. 2003, 220, 295-301. 63. Jiao, Z.; Baba, T.; Mori, H.; Shimizu, K. Appl. Microbiol. Biotechnol. 2004, 64, 91-98. 64. Al Zaid Siddiquee, K.; Arauzo-Bravo, M. J.; Shimizu, K. Appl. Microbiol. Biotechnol. 2004, 63, 407-417. 65. Jiao, Z.; Baba, T.; Mori, H.; Shimizu, K. Metab. Eng. 2004, 6, 164-174. 66. Sauer, U.; Canonaco, F.; Heri, S.; Perrenoud, A.; Fischer, E. J. Biol. Chem. 2004, 279, 6613-6619.

PAGE 261

244 67. Weber, J.; Kayser, A.; Rinas, U. Microbiology, 2005, 151, 707-716. 68. Stephanopoulos G. N., Aristidou A., Nielsen N. Metabolic Engineering Principles and Methodologies; Academic Press: San Diego, 1988. 69. Pramanik, J.; Keasling, J. D. Biotechnol. Bioeng. 1998, 60, 230-8. 70. Schuster, S.; Dandekar, T.; Fell, D. A. Trends Biotechnol. 1999, 17, 53-60. 71. Mahadevan, R.; Edwards, J. S.; Doyle, F. J., III Biophys. J. 2002, 83, 1331-1340. 72. Lee, S. Y.; Papoutsakis, E. T. Metabolic Engineering; Dekker: New York, 1999. 73. Eden E. Accessed on November 2006. http://www.cs.technion.ac.il/~eraneden/ 74. Palsson B. Accessed on November 2006. http://gcrg.ucsd.edu/classes/2_10_fba.pdf 75. Kholodenko B. N., Westerhoff. H. V Metabolic Engineering in the Post Genomic Era; Horizon Bioscience: Norfolk, 2004. 76. Wiechert, W. Metab. Eng. 2001, 3, 195-206. 77. Wiechert, W.; Mollney, M.; Petersen, S.; de Graaf, A. A. Metab. Eng. 2001, 3, 265283. 78. Szyperski, T. Q. Rev. Biophys. 1998, 31, 41-106. 79. Follstad, B. D.; Stephanopoulos, G. Eur. J. Biochem. 1998, 252, 360-371. 80. van Winden, W.; Verheijen, P.; Heijnen, S. Metab. Eng. 2001, 3, 151-162. 81. Edwards, J. S.; Ibarra, R. U.; Palsson, B. O. Nature Biotechnol. 2001, 19, 125-30. 82. Segre, D.; Vitkup, D.; Church, G. M. Proc. Natl. Acad. Sci. U S A. 2002, 99, 1511215117. 83. Kelleher, J. K. Metab. Eng. 2001, 3, 100-110. 84. Boros L.G., Cascante M. and Lee W.-N.P., Eds. Metabolic profiling. Its role in biomarker discovery and gene function analysis Kruwer Academic Publishers, 2003;pp 141-169. 85. Neidhardt, F.C., Curtiss III, R., Ingraham, J.L., Lin, E.C.C., Brooks Low, K., Magasanik, B., Reznikoff, W.S., Riley, M ., Schaechter, M., and Umbarger, H.E., Eds. Escherichia coli and Salmonella typhimurium. ASM Press, Washington, D.C. 1996 86. Schmidt, K.; Nielsen, J.; Villadsen, J. J. Biotechnol. 1999, 71, 175-189. 87. Szyperski, T. Q. Rev. Biophys. 1998, 31, 41-106.

PAGE 262

245 88. Szyperski, T.; Bailey, J. E.; Wuethrich, K. Trends Biotechnol. 1996, 14, 453-459. 89. Szyperski, T.; Glaser, R. W.; Hochuli, M.; Fiaux, J.; Sauer, U.; Bailey, J. E.; Wuthrich, K. Metab. Eng. 1999, 1, 189-197. 90. Wittmann, C. Adv. Biochem. Eng. Biotechnol. 2002, 74, 39-64. 91. Dauner, M.; Sauer, U. Biotechnol. Prog. 2000, 16, 642-9. 92. Boonstra, B.; French, C. E.; Wainwright, I.; Bruce, N. C. J. Bacteriol. 1999, 181, 1030-1034. 93. Hanson, R. L.; Rose, C. J. Bacteriol. 1980, 141, 401-4. 94. Gosset, G. Microb. Cell Fact. 2005, 4, 14 95. Shimizu, K. Adv. Biochem. Eng. Biotechnol. 2004, 91, 1-49. 96. Cozzone, A. J.; El-Mansi, M. J. Mol. Microbiol. Biotechnol. 2005, 9, 132-146. 97. Kimata, K.; Tanaka, Y.; Inada, T.; Aiba, H. EMBO J. 2001, 20, 3587-3595. 98. van Winden, W.; Schipper, D.; Verheijen, P.; Heijnen, J. Metab. Eng. 2001, 3, 322343. 99. Zamboni, N.; Fischer, E.; Sauer, U. BMC Bioinformatics 2005, 6, 209 100. van Winden, W.; Verheijen, P.; Heijnen, S. Metab. Eng. 2001, 3, 151-162. 101. Wiechert, W.; Mollney, M.; Petersen, S.; de Graaf, A. A. Metab. Eng. 2001, 3, 265283. 102. Bailey, J. E. Science 1991, 252, 1668-75. 103. Mertens, R.; Greiner, L.; van den Ban, E. C. D.; Haaker, H. B. C. M.; Liese, A. J. Mol. Catal. B 2003, 24-25, 39-52. 104. Lundquist, R.; Olivera, B. M. J. Biol. Chem. 1971, 246, 1107-16. 105. Chenault, H. K.; Whitesides, G. M. Appl. Biochem. Biotech. 1987, 14, 147 106. Wichmann, R.; Vasic-Racki, D. Adv. Biochem. Eng.Biotech. 2005, 92, 225-260. 107. Sigma-Aldrich. Accessed on May 2006. www.sigmaaldrich.com 108. Institute of Biomolecular Design. Accessed on November 2006. http://redpoll.pharmacy.ualberta.ca/CCDB/cgi-bin/STAT_NEW.cgi 109. Leonida, M. D. Curr. Med. Chem. 2001, 8, 345-369. 110. Eckstein, M.; Daussmann, T.; Kragl, U. Biocatal. Biotransfor. 2004, 22, 89-96.

PAGE 263

246 111. Wolberg, M.; Hummel, W.; Wandrey, C.; Muller, M. Angew. Chem. Int. Ed. 2000, 39, 4306-4308. 112. Wong, C.-H.; Whitesides, G. M. J. Am. Chem. Soc. 1981, 103, 4890-9. 113. Wolberg, M.; Kaluzna, I. A.; Mueller, M.; Stewart, J. D. Tetrahedron: Asymmetry 2004, 15, 2825-2828. 114. Kalaitzakis, D.; Rozzell, J. D.; Kambourakis, S.; Smonou, I. Org. Lett. 2005, 7, 47994801. 115. Seelbach, K.; Riebel, B.; Hummel, W.; Kula, M.-R.; Tishkov, V. I.; Egorov, A. M.; Wandrey, C.; Kragl, U. Tetrahedron Lett. 1996, 37, 1377-80. 116. Julich Chiral Solutions. Accessed on May 2006. www.julich.com/public/default.cfm 117. Woodyer, R.; Zhao, H.; van der Donk, W. A. FEBS J. 2005, 272, 3816-3827. 118. Kataoka, M.; Rohani, L. P. S.; Wada, M.; Kita, K.; Yanase, H.; Urabe, I.; Shimizu, S. Biosci. Biotechnol. Biochem. 1998, 62, 167-169. 119. Ema, T.; Yagasaki, H.; Okita, N.; Takeda, M.; Sakai, T. Tetrahedron 2006, 62, 61436149. 120. Weckbecker, A.; Hummel, W. Biotechnol. Lett. 2004, 26, 1739-1744. 121. Xu, Z.; Liu, Y.; Fang, L.; Jiang, X.; Jing, K.; Cen, P. Appl. Microbiol. Biotechnol. 2006, 70, 40-46. 122. Walton, A. Z.; Stewart, J. D. Biotechnol. Prog. 2004, 20, 403-411. 123. Hanson, R. L.; Goldberg, S.; Goswami, A.; Tully, T. P.; Patel, R. N. Adv. Synth. Catal. 2005, 347, 1073-1080. 124. van der Donk, W. A.; Zhao, H. Curr. Opin. Biotechnol. 2003, 14, 421-426. 125. Zhao, H.; van der Donk, W. A. Curr. Opin. Biotechnol. 2003, 14, 583-589. 126. Biocatalytic, Inc. Accessed on May 2006. www.biocatalytics.com 127. Kataoka, M.; Rohani, L. P. S.; Yamamoto, K.; Wada, M.; Kawabata, H.; Kita, K.; Yanase, H.; Shimizu, S. Appl. Microbiol. Biotechnol. 1997, 48, 699-703. 128. van Beilen, J. B.; Duetz, W. A.; Schmid, A.; Witholt, B. Trends Biotechnol. 2003, 21, 170-177. 129. Duetz, W. A.; Van Beilen, J. B.; Witholt, B. Curr. Opin. Biotechnol. 2001, 12, 419425. 130. Straathof, A. J. J.; Panke, S.; Schmid, A. Curr. Opin. Biotechnol. 2002, 13, 548-556.

PAGE 264

247 131. Hummel, W. Trends Biotechnol. 1999, 17, 487-492. 132. J.D. Stewart, in Environmental Catalysis; Grassian V.H., Ed.; CRC Press, Boca Raton, FL, 2004; pp 649-665. 133. Held, M.; Schmid, A.; Van Beilen, J. B.; Witholt, B. Pure Appl. Chem. 2000, 72, 13371343. 134. Poulsen, B. R.; Nohr, J.; Douthwaite, S.; H ansen, L. V.; Iversen, J. J. L.; Visser, J.; Ruijter, G. J. G. FEBS J. 2005, 272, 1313-1325. 135. Moritz, B.; Striegel, K.; de Graaf, A. A.; Sahm, H. Metab. Eng. 2002, 4, 295-305. 136. Moreira dos Santos, M.; Raghevendran, V.; Koetter, P.; Olsson, L.; Nielsen, J. Metab. Eng. 2004, 6, 352-363. 137. Moreira dos Santos, M.; Thygesen, G.; Kotter, P.; Olsson, L.; Nielsen, J. FEMS Yeast Res. 2003, 4, 59-68. 138. Sanchez, A. M.; Bennett, G. N.; San, K.-Y. J. Biotechnol. 2005, 117, 395-405. 139. San, K.-Y.; Bennett, G. N.; Berrios-Rivera, S. J.; Vadali, R. V.; Yang, Y.-T.; Horton, E.; Rudolph, F. B.; Sariyar, B.; Blackwood, K. Metab. Eng. 2002, 4, 182-192. 140. Flores, S.; de Anda-Herrera, R.; Gosset, G.; Bolivar Francisco, G. Biotechnol. Bioeng. 2004, 87, 485-94. 141. Lim, S.-J.; Jung, Y.-M.; Shin, H.-D.; Lee, Y.-H. J. Biosc. Bioeng. 2002, 93, 543-549. 142. Sanchez, A. M.; Bennett, G. N.; San, K.-Y. J. Biotechnol. 2006, 117, 395-405. 143. Monroe M.;Pharkya P.;Chris, J.;Cirino P College of Engineering Research Symposium 2006. Accessed on November 2006. http://www.engr.psu.edu/symposium2006/papers/Session%202A%20%20Biological%20Systems/Monroe.pdf. 144. Doig, S. D.; Avenell, P. J.; Bird, P. A.; Gallati, P.; Lander, K. S.; Lye, G. J.; Wohlgemuth, R.; Woodley, J. M. Biotechnol. Prog. 2002, 18, 1039-1046. 145. Doig, S. D.; Simpson, H.; Alphand, V.; Furstoss, R.; Woodley, J. M. Enz. Microb. Tech. 2003, 32, 347-355. 146. Medical Isotopes, Inc. Accessed on November 2006. http://medicalisotopes.com/index.asp 147. Alexeeva, S.; Hellingwerf, K. J.; Teixeira de Mattos, M. J. J. Bacteriol. 2002, 184, 1402-1406. 148. Alexeeva, S.; de Kort, B.; Sawers, G.; Hellingwerf, K. J.; de Mattos, M. J. J. Bacteriol. 2000, 182, 4934-40.

PAGE 265

248 149. Kaluzna I., A.; Matsuda, T.; Sewell A., K.; Stewart J. D. J. Am. Chem. Soc. 2004, 126, 12827-32. 150. Aoshima, M.; Ishii, M.; Yamagishi, A.; Oshima, T.; Igarashi, Y. Biotechnol. Bioeng. 2003, 84, 732-737. 151. Birren B; Green E. D.; Klapholz S,; Myers R. M, Roskams J. Analyzing DNA: A Laboratory Manual : Genome Analysis, Cold Spring Harbor Laboratory Press,1997 152. Kaluzna, I.; Andrew, A. A.; Bonilla, M.; Martzen, M. R.; Stewart, J. D. J. Mol. Catal. B 2002, 17, 101-105. 153. Sambrook, J; Fritsch, E. F.; Maniatis, T. Molecular Cloning A Laboratory Manual; Cold Spring Harbor: Cold Spring Harbor, 1989 154. Derbise, A.; Lesic, B.; Dacheux, D.; Ghigo, J. M.; Carniel, E. FEMS Immunol. Med. Microbiol. 2003, 38, 113-116. 155. Colibri, Institute Pasteur. Accessed on November 2006. http://genolist.pasteur.fr/Colibri 156. Viola, R. E.; Cook, P. F.; Cleland, W. W. Anal. Biochem. 1979, 96, 334-40. 157. Orr, G. A.; Blanchard, J. S. Anal. Biochem. 1984, 142, 232-4. 158. Esaki, N.; Shimoi, H.; Nakajima, N .; Ohshima, T.; Tanaka, H.; Soda, K. J. Biol. Chem. 1989, 264, 9750-2. 159. Jeong, S.-S.; Gready, J. E. Anal. Biochem. 1994, 221, 273-7. 160. Wolfrom, M. L.; Thompson, A. Meth. Carbohydr. Chem. 1963, 2, 65-8. 161. Koch, H. J.; Stuart, R. S. Carbohydr. Res. 1978, 64, 127-34. 162. Hatcher, J. T.; Wilcox, L. V. Anal. Chem. 1950, 22, 567-9. 163. Datsenko, K. A.; Wanner, B. L Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 6640-6645. 164. Csonka, L. N.; Fraenkel, D. G. J. Biol. Chem. 1977, 252, 3382-91. 165. Sauer, U.; Canonaco, F.; Heri, S.; Perrenoud, A.; Fischer, E. J. Biol. Chem. 2004, 279, 6613-6619. 166. Canonaco, F.; Hess, T. A.; Heri, S.; Wang, T.; Szyperski, T.; Sauer, U. FEMS Microbiol. Lett. 2001, 204, 247-252. 167. Pionnier, S.; Robins, R. J.; Zhang, B.-L. J. Agr. Food Chem. 2003, 51, 2076-2082. 168. Aoshima, M.; Ishii, M.; Yamagishi, A.; Oshima, T.; Igarashi, Y. Biotechnol. Bioeng. 2003, 84, 732-737.

PAGE 266

249 169. Dauner, M.; Sauer, U. Biotechnol. Prog. 2000, 16, 642-649 170. Northrop, D. B. Ann. Rev. Biochem. 1981, 50, 103-31. 171. Arnold, L. J., Jr.; You, K.-S. Methods Enzymol. 1978, 54, 223-32. 172. Arnold, L. J., Jr.; You, K.-S.; Allison, W. S.; Kaplan, N. O. Biochemistry 1976, 15, 4844-9. 173. You, K.-S. CRC Crit. Rev. Biochem. 1985, 17, 313-451 174. Csonka, L. N.; Fraenkel, D. G. J. Biol. Chem. 1977, 252, 3382-91. 175. Parag Parekh Masters Thesis 2006, University of Florida 176. Csonka, L. N.; Fraenkel, D. G. J. Biol. Chem. 1977, 252, 3382-91. 177. Sauer, U.; Canonaco, F.; Heri, S.; Perrenoud, A.; Fischer, E. J. Biol. Chem. 2004, 279, 6613-6619. 178. Sauer, U.; Fischer, E. J. Biol. Chem. 2003, 278, 46446-46451. 179. Verho, R.; Londesborough, J.; Penttilae, M.; Richard, P. Appl. Environ. Microbiol. 2003, 69, 5892-5897. 180. Verho, R.; Richard, P.; Jonson, P. H.; Sundqvist, L.; Londesborough, J.; Penttilae, M. Biochemistry 2002, 41, 13833-13838. 181. Sprenger, G. A.; Hammer, B. A.; Johnson, E. A.; Lin, E. C. C. J. Gen. Microbiol. 1989, 135, 1255-62. 182. Tong, I. T.; Liao, H. H.; Cameron, D. C. Appl. Environ. Microbiol. 1991, 57, 3541-6. 183. Fisher, E. Syntheses in the Purine and Sugar Group. Accessed on November 2006. http://nobelprize.org/nobel_prizes/chemistry/laureates/1902/fischer-lecture.pdf 184. Filmore, D. Mod. Drug Discov. 2003, 6, 35-36, 38-39. 185. Federsel, H.-J. Nature Rev. Drug Discov. 2005, 4, 685-697. 186. Garcia-Urdiales, E.; Alfonso, I.; Gotor, V. Chem. Rev. 2005, 105, 313-354. 187. Schoemaker Hans, E.; Mink, D.; Wubbolts Marcel, G. Science 2003, 299, 1694-7. 188. Breuer, M.; Ditrich, K.; Habicher, T.; H auer, B.; Kesseler, M.; Stuermer, R.; Zelinski, T. Angew. Chem. Int. Ed. 2004, 43, 788-824. 189. Lorenz, P.; Eck, J. Nature Rev. Microbiol. 2005, 3, 510-516. 190. Johanson, T.; Katz, M.; Gorwa-Grauslund, M. F. FEMS Yeast Res. 2005, 5, 513-525.

PAGE 267

250 191. Stewart, J. D. Curr. Opin. Biotech. 2000, 11, 363-368. 192. Rodriguez, S.; Kayser, M.; Stewart, J. D. Org. Lett. 1999, 1, 1153-1155. 193. Rodriguez, S.; Kayser, M. M.; Stewart, J. D. J. Am. Chem. Soc. 2001, 123, 15471555. 194. Rodriguez, S.; Schroeder, K. T.; Kayser, M. M.; Stewart, J. D. J. Org. Chem. 2000, 65, 2586-2587. 195. Kaluzna, I. A.; Matsuda, T.; Sewell, A. K.; Stewart, J. D. J. Am. Chem. Soc. 2004, 126, 12827-12832. 196. Stewart, J.D., Rodriguez, S. Kayser, M.M.,Cloning, Structure and Activity of Ketone Reductases from Bakers Yeast. in Enzyme Technology for Pharmaceutical and Biotechnological Applications, H.A. Kirst, W.-K. Yeh and M.J. Zmijewski, Eds, Marcel Dekker: New York, 2001, pp 175-207. 197. Miya, H.; Kawada, M.; Sugiyama, Y. Biosci., Biotech., Biochem. 1996, 60, 95-8. 198. Misra, Kanika; Banerjee, Arun B.; Ray, Subhankar; Ray, Manju.Mol. Cell. Biochem. 1996, 156, 117-24. 199. Yum, D.-Y.; Lee, B.-Y.; Pan, J.-G. Appl. Environ. Microbiol. 1999, 65, 3341-3346. 200. Habrych, M.; Rodriguez, S.; Stewart, J. D. Biotechnol. Progr. 2002, 18, 257-261. 201. The EcoGene Database of Eschericia coli Sequence and Function Accessed on November 2006. http://www.ecogene.org 202. Ko, J.; Kim, I.; Yoo, S.; Min, B.; Kim, K.; Park, C. J. Bacteriol. 2005, 187, 5782-5789. 203. Bacterial Targets IGS-CNRS. Accessed on November 2006 http://igs-server.cnrsmrs.fr/Str_gen/ 204. Jeudy, S.; Monchois, V.; Maza, C.; Claverie, J.-M.; Abergel, C. Proteins: Structure, Function, and Bioinformatics 2005, 62, 302-307. 205. Ellis, E. M. FEMS Microbiol. Lett. 2002, 216, 123-131. 206. Protein Data Bank. Accessed on November 2006. http://www.rcsb.org/pdb/Welcome.do;js essionid=h+qtwqwGy+UQu0-VyP1ixw** 207. Obmolova, G.; Teplyakov, A.; Khil Pavel, P.; Howard Andrew, J.; Camerini-Otero, R. D.; Gilliland Gary, L. Proteins 2003, 53, 323-5. 208. Grant, A. W.; Steel, G.; Waugh, H.; Ellis, E. M. FEMS Microbiol. Lett. 2003, 218, 9399.

PAGE 268

251 209. Information Genomique et Stracturale, TCoffee. Accessed on November 2006. http://www.igs.cnrs-mrs.fr/Tcoffee/tcoffee_cgi/index.cgi 210. Datsenko, K. A.; Wanner, B. L Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 6640-6645 211. Yu, D.; Ellis, H. M.; Lee, E. C.; Jenkins, N. A.; Copeland, N. G.; Court, D. L. Proc. Natl. Acad. Sci. U.S.A 2000, 97, 5978-5983. 212. Kitazume, T.; Nakayama, Y. J. Org. Chem. 1986, 51, 2795-9. 213. Kaluzna, I. A.; Dascier, D.; Stewart, J. D. Unpublished data 214. Theorell, H. The nature and mode of action of oxidation enzymes. Accessed on November 2006. http://nobelprize.org/nobel_prizes/medicine/laureates/1955/theorelllecture.pdf. 215. Brown J. M. Hydrogenation of Functionalized Carbon-Carbon Double Bonds In Comprehensive Asymmetric Catalysis Vol I, Jacobsen E.N; Pfaltz, A. ; Yamamoto H., Eds. Speinger-Verlag, New York, 1999. 216. Dobbs, D. A.; Vanhessche, K. P. M.; Braz i, E.; Rautenstrauch, V.; Lenoir, J.; Genet, J.; Wiles, J.; Bergens, S. H. Angew. Chem. Int. Ed. Engl. 2000, 39, 1992-1995. 217. Ohta, T.; Miyake, T.; Seido, N.; Kumobayashi, H.; Takaya, H. J. Org. Chem. 1995, 60, 357-63. 218. Lipshutz, B. H.; Servesko, J. M. Angew. Chem. Int. Ed. Engl. 2003, 42, 4789-4792. 219. Lipshutz, Br. H.; Servesko, J. M.; Petersen, T. B.; Papa, P. P.; Lover, A. A. Org. Lett. 2004, 6, 1273-1275. 220. Matzinger P. K.; Leuenberger H. G. W. Enzyme-Catalyzed Hydrogenation, In Stereoselective Synthesis Vol 7, 1997, Houben J;Weyl T, Eds. Houben-weyl, 1997. 221. Faber K. Biotransformations in organic chemistry, 3rd ed.; Springer-Verlag, 1997. 222. Servi S. Stereoselective Synthesis of Chiral Compounds Using Whole-cell Biocatalysts In Stereoselective Biocatalysis Patel R. N. Ed.; Marcel Dekker, 1999. 223. Santaniello, E.; Manzocchi, A. Trends Org. Chem. 2000, 8, 21-34. 224. D'Arrigo, P.; Pedrocchi-Fantoni, G.; Servi, S. Adv. Appl. Microbiol. 1997, 44, 81-123. 225. Csuk, R.; Glaenzer, B. I. Chem. Rev. 1991, 91, 49-97. 226. Servi, S. Synthesis 1990, 1-25. 227. Gramatica, P.; Manitto, P.; Monti, D.; Speranza, G., Tetrahedron 1987, 43, 4481-6. 228. Leuenberger, H. G. W.; Boguth, W.; Barner, R.; Schmid, M.; Zell, Helv. Chim. Acta 1979, 62, 455-63.

PAGE 269

252 229. Fuganti, C.; Grasselli, P. Chem. Ind. 1977, 24, 983. 230. Fuganti, C.; Grasselli, P.; Servi, S. Perkin Trans. 1 1983, 241-4. 231. Gramatica, P.; Giardina, G.; Speranza, G.; Manitto, P. Chem. Lett. 1985, 9, 1395-8. 232. D'Arrigo, P.; Fuganti, C.; Fantoni, G. P.; Servi, S. Tetrahedron 1998, 54, 1501715026. 233. Gramatica, P.; Manitto, P.; Monti, D.; Speranza, G. Tetrahedron 1988, 44, 1299-304. 234. Ferraboschi, P.; Casati, S.; Santaniello, E. Tetrahedron: Asymmetry 1994, 5, 19-20. 235. Ferraboschi, P.; Grisenti, P.; Casati, R.; Fiecchi, A.; Santaniello, E. Perkin Trans. 1 1987, 1743-8. 236. Utaka, M.; Konishi, S.; Mizuoka, A.; Ohkubo, T.; Sakai, T.; Tsuboi, S.; Takeda, A. J. Org. Chem. 1989, 54, 4989-92. 237. Utaka, M.; Konishi, S.; Okubo, T.; Tsuboi, S.; Takeda, A. Tetrahedron Lett. 1987, 28, 1447-50. 238. Fuganti, C.; Pedrocchi-Fantoni, G.; Sarra, A.; Servi, S. Tetrahedron: Asymmetry 1994, 5, 1135-8. 239. Ohta, H.; Ozaki, K.; Tsuchihashi, G. Chem. Lett. 1987, 191-2. 240. Kawai, Y.; Inaba, Y.; Tokitoh, N. Tetrahedron: Asymmetry 2001, 12, 309-318. 241. McAnda, A. F.; Roberts, K. D.; Smallridge, A. J.; Ten, A.; Trewhella, M. A. Perkin Trans. 1 1998, 501-504. 242. Serra, S.; Fuganti, C.; Brenna, E. Trends Biotechnol. 2005, 23, 193-198. 243. Kosjek, B.; Stampfer, W.; Van Deursen, R.; Faber, K.; Kroutil, W. Tetrahedron 2003, 59, 9517-9521. 244. Fronza, G.; Fuganti, C.; Mendozza, M.; Ra llo, R. S.; Ottolina, G.; Joulain, D. Tetrahedron 1996, 52, 4041-52. 245. Fronza, G.; Fuganti, C.; Mendozza, M.; Rigoni, R.; Servi, S.; Zucchi, G. Pure Appl. Chem. 1996, 68, 2065-2071. 246. Koul, S.; Crout, D. H. G.; Errington, W.; Tax, J. Perkin Trans. 1 1995, 2969-88. 247. Jui J., Adv. Biochem. Eng. 1980, 17, 37 248. Gramatica, P.; Manitto, P.; Poli, L J. Org. Chem. 1985, 50, 4625-8 249. Fischer F.G.; Weidemann, O. Leibigs Ann. Chem. 1935, 520, 52-70

PAGE 270

253 250. Buque-Taboada, E. M.; Straathof, A. J. J.; Heijnen, J. J.; Van Der Wielen, L. A. M. Biotechnol. Bioeng. 2004, 86, 795-800. 251. Pillai, U. R.; Sahle-Demessie, E. Ind. Eng. Chem. Res. 2003, 42, 6688-6696. 252. Brunner, H.; Fisch, K. J. Organomet. Chem. 1993, 456, 71-5. 253. von Arx, M.; Mallat, T.; Baiker, A. J. Mol. Catal. Chem. 1999, 148, 275-283. 254. Cheetham, P. S. J. Adv. Biochem. Eng. Biotechnol. 2004, 86, 83-158. 255. El-Sharkawy, S. H.; Abul-Hajj, Y. J. J. Org. Chem. 1988, 53, 515-19. 256. Kergomard, A.; Renard, M. F.; Veschambre, H. J. Org. Chem. 1982, 47, 792-8. 257. Fauve, A.; Renard, M. F.; Veschambre, H. J. Org. Chem. 1987, 52, 4893-7. 258. Matsumoto, K.; Kawabata, Y.; Takahashi, J.; Fujita, Y.; Hatanaka, M. Chem. Lett. 1998, 283-284. 259. Sato, Y.; Oda, T.; Saito, H. Chem. Comm. 1977, 415-17. 260. Arnone, A.; Cardillo, R.; Nasini, G.; Vajna de Pava, O. Perkin Trans. 1 1990, 3061-3. 261. Hirata, T.; Takarada, A.; Matsushima, A.; Kondo, Y.; Hamada, H. Tetrahedron: Asymmetry 2004, 15, 15-16. 262. Shimoda, K.; Kubota, N.; Hamada, H.; Yamane, S.-y.; Hirata, T. Bull. Chem. Soc. Jpn. 2004, 77, 2269-2272. 263. Hamada, H.; Miyamoto, Y.; Ishihara, K.; Nakajima, N.; Hamada, H. Plant Biotechnol. 2002, 19, 203-205. 264. Shimoda, K.; Kubota, N.; Hamada, H.; Kaji, M.; Hirata, T. Tetrahedron: Asymmetry 2004, 15, 1677-1679. 265. Bader, J.; Guenther, H.; Rambeck, B.; Simon, H. Hoppe-Seyler's Zeitschrift fuer Physiologische Chemie 1978, 359, 19-27. 266. Simon, H.; Rambeck, B.; Hashimoto, H.; Guenther, H.; Nohynek, G.; Neumann, H. Angew. Chem. 1974, 86, 675. 267. Simon, H.; Bader, J.; Guent her, H.; Neumann, S.; Thanos, J. Angew. Chem. 1985, 97, 541-55. 268. Simon, H. Pure Apll. Chem. 1992, 64, 1181-1186 269. Thanos, I. C. G.; Simon, H. J. Biotechnol. 1987, 6, 13-29. 270. Thanos, I.; Bader, J.; Guenther, H.; Neumann, S.; Krauss, F.; Simon, H. Methods. Enzymol. 1987, 136, 302-17.

PAGE 271

254 271. Warburg, O. ; Christian, W. Naturwissenschaften 1932, 20, 688 272. Williams, R. E.; Bruce, N. C. Microbiology 2002, 148, 1607-1614. 273. Trotter, E. W.; Collinson, E. J.; Dawes, I. W.; Grant, C. M. Appl. Environ. Microbiol. 2006, 72, 4885-4892. 274. Haarer, B. K.; Amberg, D. C. Mol. Biol. Cell 2004, 15, 4522-4531. 275. Saito, K.; Thiele, D. J.; Davio, M.; Lockridge, O.; Massey, V. J. Biol. Chem. 1991, 266, 20720-4. 276. Massey, V. Biochem. Soc. Trans. 2000, 28, 283-296. 277. Massey, V.; Schopfer, L. M. J. Biol. Chem. 1986, 261, 1215-22. 278. Schopfer, L. M.; Massey, V. Old yellow enzyme. In A Study of Enzymes, Kuby, S. A., CRC Press, Cleveland, Ohio,1991. 279. Matthews, R. G.; Massey, V.; Sweeley, C. C. J. Biol. Chem. 1975, 250, 9294-8. 280. Abramovitz, A. S.; Massey, V. J. Biol. Chem. 1976, 251, 5321-6. 281. Abramovitz, A. S.; Massey, V. J. Biol. Chem. 1976, 251, 5327-36. 282. Matthews, R. G.; Massey, V. J. Biol. Chem. 1969, 244, 1779-86. 283. Stott, K.; Saito, K.; Thiele, D. J.; Massey, V. J. Biol. Chem. 1993, 268, 6097-106 284. Kohli, R. M.; Massey, V. J. Biol. Chem. 1998, 273, 32763-32770. 285. Xu, D.; Kohli, R. M.; Massey, V. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 3556-3561. 286. Brown, B. J.; Hyun, J.; Duvvuri, S.; Karplus, P. A.; Massey, V. J. Biol. Chem. 2002, 277, 2138-2145. 287. Vaz, A. D. N.; Chakraborty, S.; Massey, V. Biochemistry 1995, 34, 4246-56. 288. Swiderska, M. A.; Stewart, J. D. J. Mol. Catal. B 2006, 42, 52-54. 289. Karplus, P. A.; Fox, K. M.; Massey, V. FASEB J. 1995, 9, 1518-26. 290. Fox, K. M.; Karplus, P. A. Structure 1994, 2, 1089-105. 291. PyMOL, DeLano Scientific, LLC. Accessed on November 2006. www.pymol.org 292. Barna, T. M.; Khan, H.; Bruce, N. C.; Barsukov, I.; Scrutton, N. S.; Moody, P. C. E. J. Mol. Biol. 2001, 310, 433-447. 293. Murthy, Y. V. S. N.; Meah, Y.; Massey, V. J. Am. Chem. Soc. 1999, 121, 5344-5345.

PAGE 272

255 294. Murthy, Y. V. S. N.; Massey, V. J. Biol. Chem. 1998, 273, 8975-8982. 295. Stewart, R. C.; Massey, V. J. Biol. Chem. 1985, 260, 13639-47. 296. Murthy, Y. V. S. N.; Massey, V. Methods. Enzymol. 1997, 280, 436-460. 297. Meah, Y.; Massey, V. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 10733-10738. 298. Meah, Y.; Brown, B. J.; Chakraborty, S.; Massey, V. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 8560-8565. 299. Niino, Y. S.; Chakraborty, S.; Brown, B. J.; Massey, V. J. Biol. Chem. 1995, 270, 1983-91. 300. Kataoka, M.; Kotaka, A.; Hasegawa, A.; Wada, M.; Yoshizumi, A.; Nakamori, S.; Shimizu, S. Biosci. Biotechnol. Biochem. 2002, 66, 2651-2657. 301. Madani, N. D.; Malloy, P. J.; Rodriguez -Pombo, P.; Krishnan, A. V.; Feldman, D. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 922-6. 302. Buckman, J.; Miller, S. M. Biochemistry 1998, 37, 14326-14336. 303. Komduur, J. A.; Leao, A. N.; Monastyr ska, I.; Veenhuis, M.; Kiel, J. A. K. W. Curr. Genet. 2002, 41, 401-406. 304. Miura, K.; Tomioka, Y.; Suzuki, H.; Yonezawa, M.; Hishinuma, T.; Mizugaki, M. Biol. Pharmaceut. Bull. 1997, 20, 110-112. 305. Mizugaki, M.; Unuma, T.; Yamanaka, H. Chem. Pharmaceut. Bull. 1979, 27, 2334-7. 306. Mizugaki, M.; Nakazawa, M.; Yamamoto, H.; Yamanaka, H. J. Biochem. 1989, 105, 782-4. 307. Wada, M.; Yoshizumi, A.; Noda, Y.; Kataok a, M.; Shimizu, S.; Takagi, H.; Nakamori, S. Appl. Environ. Microbiol. 2003, 69, 933-937. 308. Kataoka, M.; Kotaka, A.; Thiwthong, R.; Wada, M.; Nakamori, S.; Shimizu, S. J. Biotechnol. 2004, 114, 1-9. 309. Mergler, M.; Wolf, K.; Zimmermann, M. Appl. Microbiol. Biotechnol. 2004, 63, 418421. 310. Adachi, O.; Matsushita, K.; Shinagawa, E.; Ameyama, M. J. Biochem. 1979, 86, 699709. 311. French, C. E.; Bruce, N. C. Biochem. J. 1994, 301, 97-103. 312. French, C. E.; Bruce, N. C. Biochem. J. 1995, 312, 671-8. 313. Boonstra, B.; Rathbone, D. A.; Bruce, N. C. Biomol. Eng. 2001, 18, 41-7.

PAGE 273

256 314. Ramos, J. L.; Gonzalez-Perez, M. M.; Caballero, A.; van Dillewijn, P. Curr. Opin. Biotechnol. 2005, 16, 275-281. 315. Stenuit, B.; Eyers, L.; El Fantroussi, S.; Agathos, S. N. Environ. Sci. Biotechnol. 2005, 4, 39-60. 316. French, C. E.; Rosser, S. J.; Davies, G. J.; Nicklin, S.; Bruce, N. C. Nature biotechnol. 1999, 17, 491-4. 317. French, C. E.; Nicklin, S.; Bruce, N. C. Appl. Environ. Microbiol. 1998, 64, 28642868. 318. Messiha H. L.; Bruce N. C.; Sattelle B. M.; Sutcliffe M. J.; Munro A., W.; Scrutton N., S. J. Biol. Chem. 2005, 280, 27103-10. 319. Moody, P. C. E.; Shikotra, N.; French, C. E.; Bruce, N. C.; Scrutton, N. S. Acta Crystallogr. Section D 1998, D54, 675-677. 320. Barna, T.; Messiha, H. L.; Petosa, C.; Bruce, N. C.; Scrutton, N. S.; Moody, P. C. E. J. Biol. Chem. 2002, 277, 30976-30983. 321. Blehert, D. S.; Knoke, K. L.; Fox, B. G.; Chambliss, G. H. J. Bacteriol. 1997, 179, 6912-6920. 322. Pak, J. W.; Knoke, K. L.; Noguera, D. R.; Fox, B. G.; Chambliss, G. H. Appl. Environ. Microbiol. 2000, 66, 4742-50. 323. Khan, H.; Harris, R. J.; Barna, T.; Craig, D. H.; Bruce, N. C.; Munro, A. W.; Moody, P. C. E.; Scrutton, N. S. J. Biol. Chem. 2002, 277, 21906-21912. 324. Blehert, D. S.; Fox, B. G.; Chambliss, G. H. J. Bacteriol. 1999, 181, 6254-6263. 325. Marshall, S. J.; Krause, D.; Blencowe, D. K.; White, G. F. J. Bacteriol. 2004, 186, 1802-1810. 326. French, C. E.; Nicklin, S.; Bruce, N. C. J. Bacteriol. 1996, 178, 6623-7. 327. Fitzpatrick, T. B.; Amrhein, N.; Macheroux, P. J. Biol. Chem. 2003, 278, 1989119897. 328. Brige, A.; Van den Hemel, D.; Carpentier, W.; De Smet, L.; Van Beeumen, J. J. Biochem. J. 2006, 394, 335-344. 329. Williams, R. E.; Rathbone, D. A.; Scrutton, N. S.; Bruce, N. C. Appl. Environ. Microbiol. 2004, 70, 3566-3574. 330. Vick, B. A.; Zimmerman, D. C. Plant Physiol. 1986, 80, 202-5. 331. Schaller, F. J. Exp. Bot. 2001, 52, 11-23. 332. Schaller, F.; Weiler, E. W. Eur. J. Biochem. 1997, 245, 294-299.

PAGE 274

257 333. Schaller, F.; Weiler, E. W. J. Biol. Chem. 1997, 272, 28066-28072. 334. Schaller, F.; Hennig, P.; Weiler, E. W. Plant Physiol. 1998, 118, 1345-1351. 335. Costa, C. L.; Arruda, P.; Benedetti, C. E. Plant Mol. Biol. 2000, 44, 61-71. 336. Schaller, F.; Biesgen, C.; Mussig, C.; Altmann, T.; Weiler, E. W. Planta 2000, 210, 979-984. 337. Sobajima, H.; Takeda, M.; Sugimori, M .; Kobashi, N.; Kiribuchi, K.; Cho, E.-M.; Akimoto, C.; Yamaguchi, T.; Minami, E.; Sh ibuya, N.; Schaller, F.; Weiler, E. W.; Yoshihara, T.; Nishida, H.; Nojiri, H.; Omori, T.; Nishiyama, M.; Yamane, H. Planta 2003, 216, 692-698. 338. Matsui, H.; Nakamura, G.; Ishiga, Y.; Toshima, H.; Inagaki, Y.; Toyoda, K.; Shiraishi, T.; Ichinose, Y. Mol. Genet. Genom. 2004, 271, 1-10. 339. Strassner, J.; Furholz, A.; Macheroux, P.; Amrhein, N.; Schaller, A. J. Biol. Chem. 1999, 274, 35067-35073. 340. Steinbacher, S.; Stumpf, M.; Weinkauf, S.; Rohdich, F.; Bacher, A.; Simon, H. Enoate Reductase in Flavins and Flavoproteins Berlin, 2002; p 941-949. 341. Caldeira, J.; Feicht, R.; White, H.; Teixeira M.; Moura, J. J. G.; Simon, H.; Moura, I. J. Biol. Chem. 1996, 271, 18743-18748. 342. He, X.-Y.; Yang, S.-Y.; Schulz, H. Eur. J. Biochem. 1997, 248, 516-520. 343. Fillgrove, K. L.; Anderson, V. E. Biochemistry 2001, 40, 12412-12421. 344. Rohdich, F.; Wiese, A.; Feicht, R.; Simon, H.; Bacher, A. J. Biol. Chem. 2001, 276, 5779-5787. 345. Liang, X.; Thorpe, C.; Schulz, H. Arch. Biochem. Biophys. 2000, 380, 373-379. 346. Hashimoto, H.; Guenther, H.; Simon, H. Hoppe-Seyler's Zeitschrift fuer Physiologische Chemie 1975, 356, 1195-1201. 347. La Roche, H. J.; Kellner, M.; Guenther, H.; Simon, H. Hoppe-Seyler's Zeitschrift fuer Physiologische Chemie 1971, 352, 399-402. 348. Kuno, S.; Bacher, A.; Simon, H. Biol. Chem. Hoppe-Seyler 1985, 366, 463-72. 349. Buehler, M.; Simon, H. Hoppe-Seyler's Zeitschrift fuer Physiologische Chemie 1982, 363, 609-25. 350. Dickert, S.; Pierik, A. J.; Linder, D.; Buckel, W. Eur. J. Biochem. 2000, 267, 3874-84. 351. Thanos, I.; Deffner, A.; Simon, H. Biol. Chem. Hoppe-Seyler 1988, 369, 451-60. 352. Bader, J.; Simon, H. Arch. Microbiol. 1980, 127, 279-87.

PAGE 275

258 353. Tischer, W.; Bader, J.; Simon, H. Eur. J. Biochem. 1979, 97, 103-12. 354. Messiha, H. L.; Bruce, N. C.; Sattelle, B. M.; Sutcliffe, M. J.; Munro, A. W.; Scrutton, N. S. J. Biol. Chem. 2005, 280, 27103-27110. 355. Khan, H.; Barna, T.; Bruce, N. C.; Munro, A. W.; Leys, D.; Scrutton, N. S. FEBS Journal 2005, 272, 4660-4671. 356. Brown, B. J.; Deng, Z.; Karplus, P. A.; Massey, V. J. Biol. Chem. 1998, 273, 3275332762. 357. Xu, D.; Kohli, R. M.; Massey, V. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 3556-3561. 358. Khan, H.; Barna, T.; Harris, R. J.; Bruce, N. C.; Barsukov, I.; Munro, A. W.; Moody, P. C. E.; Scrutton, N. S. J. Biol. Chem. 2004, 279, 30563-30572. 359. Kitzing, K.; Fitzpatrick, T. B.; Wilken, C.; Sawa, J.; Bourenkov, G. P.; Macheroux, P.; Clausen, T. J. Biol. Chem. 2005, 280, 27904-27913. 360. Fitzpatrick Teresa, B.; Auweter, S.; Kitz ing, K.; Clausen, T.; Amrhein, N.; Macheroux, P. Protein Expr. Purif. 2004, 36, 280-91. 361. Breithaupt, C.; Strassner, J.; Breitinger, U.; Huber, R.; Macheroux, P.; Schaller, A.; Clausen, T. Structure 2001, 9, 419-429. 362. Fox Brian, G.; Malone Thomas, E.; Johnson Kenneth, A.; Madson Stacey, E.; Aceti, D.; Bingman Craig, A.; Blommel Paul, G.; Buchan, B.; Burns, B.; Cao, J.; Cornilescu, C.; Doreleijers, J.; Ellefson, J.; Frederick, R.; Geetha, H.; Hruby, D.; Jeon Won, B.; Kimball, T.; Kunert, J.; Markley John, L.; Newman, C.; Olson, A.; Peterson Francis, C.; Phillips George, N., Jr.; Primm, J.; Ramirez, B.; Rosenberg Nathan, S.; Runnels, M.; Seder, K.; Shaw, J.; Smith David, W .; Sreenath, H.; Song, J.; Sussman Michael, R.; Thao, S.; Troestler, D.; Tyler, E.; Tyler, R.; Ulrich, E.; Vinarov, D.; Vojtik, F.; Volkman Brian, F.; Wesenberg, G.; Wrobel Russell, L.; Zhang, J.; Zhao, Q.; Zolnai, Z. Proteins 2005, 61, 206-8. 363. Malone, T. E.; Madson, S. E.; Wrobel, R. L.; Jeon, W. B.; Rosenberg, N. S.; Johnson, K. A.; Bingman, C. A.; Smith, D. W .; Phillips, G. N., Jr.; Markley, J. L.; Fox, B. G. Proteins 2004, 58, 243-245. 364. Breithaupt, C.; Kurzbauer, R.; Lilie H.; Schaller, A.; Strassner, J.; Huber, R.; Macheroux, P.; Clausen T. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, ;14337-14342 365. van den Hemel, D.; Brige, A.; Savvides, S. N.; Van Beeumen, J. J. Biol. Chem. 2006, 281, 28152-28161. 366. Hubbard, P. A.; Liang, X.; Schulz, H.; Kim, J.-J. P. J. Biol. Chem. 2003, 278, 3755337560. 367. Nordling, E.; Jornvall, H.; Persson, B. Eur. J. Biochem. 2002, 269, 4267-4276.

PAGE 276

259 368. Riveros-Rosas, H.; Julian-Sanchez, A.; Villalobos-Molina, R.; Pardo Juan, P.; Pina, E. Eur. J. Biochem. 2003, 270, 3309-34. 369. Hori, T.; Yokomizo, T.; Ago, H.; Sugahara, M.; Ueno, G.; Yamamoto, M.; Kumasaka, T.; Shimizu, T.; Miyano, M. J. Biol. Chem. 2004, 279, 22615-22623. 370. Yokomizo, T.; Izumi, T.; Takahashi, T.; Kasama, T.; Kobayashi, Y.; Sato, F.; Taketani, Y.; Shimizu, T. J. Biol. Chem. 1993, 268, 18128-35. 371. Yokomizo, T.; Ogawa, Y.; Uozumi, N.; Kume, K.; Izumi, T.; Shimizu, T. J. Biol. Chem. 1996, 271, 2844-50. 372. Ensor, C. M.; Zhang, H.; Tai, H.-H. Biochem. J. 1998, 330, 103-108. 373. Hori, T.; Yokomizo, T.; Ago, H.; Sugahara, M.; Ueno, G.; Yamamoto, M.; Kumasaka, T.; Shimizu, T.; Miyano, M. J. Biol. Chem. 2004, 279, 22615-22623. 374. Thorn, J. M.; Barton, J. D.; Dixon, N. E.; Ollis, D. L.; Edwards, K. J. J. Mol. Biol. 1995, 249, 785-99. 375. Dick, R. A.; Kwak, M.-K.; Sutter, T. R.; Kensler, T. W. J. Biol. Chem. 2001, 276, 40803-40810. 376. Dick R., A.; Kensler Thomas, W. J. Biol. Chem. 2004, 279, 17269-77. 377. Dick, R. A.; Yu, X.; Kensler, T. W. Clin. Cancer. Res. 2004, 10, 1492-1499. 378. Gong, J.; Neels, J. F.; Yu, X.; Kensler, T. W.; Peterson, L. A.; Sturla, S. J. J. Med. Chem. 2006, 49, 2593-2599. 379. Mano, J. i.; Babiychuk, E.; Belles-Boix, E.; Hiratake, J.; Kimura, A.; Inze, D.; Kushnir, S.; Asada, K. Eur. J. Biochem. 2000, 267, 3661-3671. 380. Hambraeus, G.; Nyberg, N. J. Agr. Food Chem. 2005, 53, 8714-8721. 381. Mano, J. i.; Torii, Y.; Hayashi, S.-i.; Takimoto, K.; Matsui, K.; Nakamura, K.; Inze, D.; Babiychuk, E.; Kushnir, S.; Asada, K. Plant Cell Physiol. 2002, 43, 1445-1455. 382. Ringer, K. L.; McConkey, M. E.; Davis, E. M.; Rushing, G. W.; Croteau, R. Arch. Biochem. Biophys. 2003, 418, 80-92. 383. Raab, T.; Lopez-Raez, J. A.; Klein, D.; Caballero, J. L.; Moyano, E.; Schwab, W.; Munoz-Blanco, J. Plant Cell 2006, 18, 1023-1037. 384. Kurata, A.; Kurihara, T.; Kamachi, H.; Esaki, N. J. Biol. Chem. 2005, 280, 2028620291. 385. Raab, T.; Lopez-Raez, J. A.; Klein, D.; Caballero, J. L.; Moyano, E.; Schwab, W.; Munoz-Blanco, J. Plant Cell 2006, 18, 1023-1037.

PAGE 277

260 386. Edwards, K. J.; Barton, J. D.; Rossjohn, J.; Thorn, J. M.; Taylor, G. L.; Ollis, D. L. Arch. Biochem Biophys. 1996, 328, 173-83. 387. Yamamoto, T.; Yokomizo, T.; Nakao, A.; Izumi, T.; Shimizu, T. Eur. J. Biochem. 2001, 268, 6105-6113. 388. Doorn, J. A.; Maser, E.; Blum, A.; Claffey, D. J.; Petersen, D. R. Biochemistry 2004, 43, 13106-13114. 389. Forrest, G. L.; Gonzalez, B. Chemico-Biological Interactions 2000, 129, 21-40. 390. Joernvall, H.; Persson, B.; Krook, M.; Atrian, S.; Gonzalez-Duarte, R.; Jeffery, J.; Ghosh, D. Biochemistry 1995, 34, 6003-13. 391. Tanaka, N.; Nonaka, T.; Nakamura, K. T.; Hara, A. Curr. Org. Chem. 2001, 5, 89111. 392. Ghosh, D.; Sawicki, M.; Pletnev, V.; Erm an, M.; Ohno, S.; Nakajin, S.; Duax, W. L. J. Biol. Chem. 2001, 276, 18457-18463. 393. Wanner, P.; Tressl, R. Eur. J. Biochem. 1998, 255, 271-278. 394. Hirata, T.; Shimoda, K.; Gondai, T. Chem. Lett. 2000, 850-851. 395. Shimoda, K.; Izumi, S.; Hirata, T. Bull. Chem. Soc. Jpn. 2002, 75, 813-816. 396. Kawai, Y.; Hayashi, M.; Inaba, Y.; Saitou, K.; Ohno, A. Tetrahedron Lett. 1998, 39, 5225-5228. 397. Kawai, Y.; Inaba, Y.; Hayashi, M.; Tokitoh, N. Tetrahedron Lett. 2001, 42, 33673368. 398. Shimoda, K.; Ito, D. I.; Izumi, S.; Hirata, T. Perkin Trans. 1 1996, 355-8. 399. Hirata, T.; Tang, Y.; Okano, K.; Suga, T. Phytochemistry 1989, 28, 3331-3. 400. Shimoda, K.; Ito, D. I.; Izumi, S.; Hirata, T. Perkin Trans. 1 1996, 355-8. 401. Shimoda, K.; Hirata, T. J. Mol. Catal. B 2000, 8, 255-264. 402. Zaadz, Inc. Accessed on November 2006. http://quotes.zaadz.com/Confucius?page=16 403. Leresche, J. E.; Meyer, H.-P. Org. Process Res. Dev. 2006, 10, 572-580. 404. Kaur, J.; Sharma, R. Crit. Rev. Biotech. 2006, 26, 165-199. 405. Bornscheuer, U. T. Adv. Biochem. Engin. Biotechn. 2005, 100, 181-203. 406. Bali, S.; O'Hare Helen, M.; Weissman Kira, J. Chembiochem 2006, 7, 478-84.

PAGE 278

261 407. van Loo, B.; Kingma, J.; Arand, M.; Wubbolts, M. G.; Janssen, D. B. Appl. Environ. Microbiol. 2006, 72, 2905-2917. 408. Bonsor, D.; Butz, S. F.; Solomons, J.; Grant, S.; Fairlamb, I. J. S.; Fogg, M. J.; Grogan, G. Org. Biomol. Chem. 2006, 4, 1252-1260. 409. Kaluzna, I. A.; Matsuda, T.; Sewell, A. K.; Stewart, J. D. J. Am. Chem. Soc. 2004, 126, 12827-12832. 410. Martzen, M. R.; McCraith, S. M.; Spinelli, S. L.; Torres, F. M.; Fields, S.; Grayhack, E. J.; Phizicky, E. M. Science 1999, 286, 1153-1155 411. Phizicky, E. M.; Grayhack, E. J. Crit. Rev. Biochem. Mol. Biol. 2006, 41, 315-327. 412. Kuznetsova, E.; Proudfoot, M.; Sanders, S. A.; Reinking, J.; Savchenko, A.; Arrowsmith, C. H.; Edwards, A. M.; Yakunin, A. F. FEMS Microbiol. Rev. 2005, 29, 263-279. 413. Hudson, J. R., Jr.; Dawson, E. P.; Rushing, K., L.; Jackson, C. H.; Lockshon, D.; Conover, D.; Lanciault, C.; Harris, J. R.; Simmons, S. J.; Rothstein, R.; Fields, S. Genome Res. 1997, 7, 1169-1173. 414. Parrish, J. R.; Limjindaporn, T.; Hines, J. A.; Liu, J.; Liu, G.; Finley, R. L., Jr. J. Proteome Res. 2004, 3, 582-586. 415. Lu, Q. Trends Biotech. 2005, 23, 199-207. 416. Lammli, U. K. Nature, 1970, 227, 680-685 417. Bradford M. M Anal. Biochem. 1976, 72, 248-254 418. Sambrook, J; Fritsch, E. F.; Maniatis, T. Molecular Cloning A Laboratory Manual; Cold Spring Harbor: Cold Spring Harbor, 1989 419. Birren B; Green E. D.; Klapholz S,; Myers R. M, Roskams J. Analyzing DNA: A Laboratory Manual : Genome Analysis, Cold Spring Harbor Laboratory Press,1997 420. Gietz, R.D. and R.A. Woods. Methods Enzymol. 2002 350: 87-96. 421. Chong, B.-D.; Ji, Y.-I.; Oh, S.-S.; Yang, J.-D.; Baik, W.; Koo, S. J. Org. Chem. 1997, 62, 9323-9325. 422. Piancatelli, G.; Leonelli, F.; Do, N.; Ragan, J. Org. Syntheses 2006, 83, 18-23. 423. Kyte, B. G.; Rouviere, P.; Cheng, Q.; Stewart, J. D. J. Org. Chem. 2004, 69, 12-17. 424. McAnda, A. F.; Roberts, K. D.; Smallridge, A. J.; Ten, A.; Trewhella, M. A. J. Chem. Soc. Perk. T 1 1998, 501-504. 425. Yamamoto, H.; Kimoto, K. ; (Daicel Chemical Industries, Ltd., Japan).: US-6780976

PAGE 279

262 426. Chen, C.-N.; Porubleva, L.; Shearer, G.; Svrakic, M.; Holden, L. G.; Dover, J. L.; Johnston, M.; Chitnis, P. R.; Kohl, D. H. Yeast 2003, 20, 545-554. 427. Krogan, N. J.; Cagney, G.; Yu, H.; Zhong, G.; Guo, X.; Ignatchenko, A.; Li, J.; Pu, S.; Datta, N.; Tikuisis, A. P.; Punna, T.; Peregrin-Alvarez, J. M.; Shales, M.; Zhang, X.; Davey, M.; Robinson, M. D.; Paccanaro, A.; Bray, J. E.; Sheung, A.; Beattie, B.; Richards, D. P.; Canadien, V.; Lalev, A.; M ena, F.; Wong, P.; Starostine, A.; Canete, M. M.; Vlasblom, J.; Wu, S.; Orsi, C.; Collin s, S. R.; Chandran, S.; Haw, R.; Rilstone, J. J.; Gandi, K.; Thompson, N. J.; Musso, G.; St. Onge, P.; Ghanny, S.; Lam, M. H. Y.; Butland, G.; Altaf-Ul, A. M.; Kanaya, S .; Shilatifard, A.; O'Shea, E.; Weissman, J. S.; Ingles, C. J.; Hughes, T. R.; Parkinson, J.; Gerstein, M.; Wodak, S. J.; Emili, A.; Greenblatt, J. F. Nature 2006, 440, 637-643. 428. Ho, Y.; Gruhler, A.; Heilbut, A.; Bader Ga ry, D.; Moore, L.; Adams, S.-L.; Millar, A.; Taylor, P.; Bennett, K.; Boutilier, K.; Yang, L.; Wolting, C.; Donaldson, I.; Schandorff, S.; Shewnarane, J.; Vo, M.; Taggart, J.; Goudreault, M.; Muskat, B.; Alfarano, C.; Dewar, D.; Lin, Z.; Michalickova, K.; Willems Andrew, R.; Sassi, H.; Nielsen Peter, A.; Rasmussen Karina, J.; Andersen J ens, R.; Johansen Lene, E.; Hansen Lykke, H.; Jespersen, H.; Podtelejnikov, A.; Nielsen, E.; Crawford, J.; Poulsen, V.; Sorensen Birgitte, D.; Matthiesen, J.; Hendrickson Ronald, C.; Gleeson, F.; Pawson, T.; Moran Michael, F.; Durocher, D.; Mann, M.; Hogue Ch ristopher, W. V.; Figeys, D.; Tyers, M. Nature 2002, 415, 180-3. 429. Ito, T.; Chiba, T.; Ozawa, R.; Yoshida, M.; Hattori, M.; Sakaki, Y Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 4569-4574. 430. Shimoda, K.; Kubota, N.; Hamada, H.; Yamane, S.-y.; Hirata, T. Bull. Chem. Soc. Japan 2004, 77, 2269-2272. 431. Fronza, G.; Fuganti, C.; Pinciroli, M.; Serra, S. Tetrahedron: Asymmetry 2004, 15, 3073-3077. 432. Wolken, W. A. M.; ten Have, R.; van der Werf, M. J. J. Agric. Food Chem. 2000, 48, 5401-5405. 433. Mueller, A.; Hauer, B.; Rosche, B. J. Mol. Catal. B 2006, 38, 126-130.

PAGE 280

263 BIOGRAPHICAL SKETCH Despina Bougioukou was born in Athens, Greece, on May 26th, 1974. She received a 1998 BS and a 2001 MS from the University of Crete, Greece. In 2002 she started her doctoral studies at University of Florida, working for Dr J. Stewart. Her research interests are in the field of biocatalysis. She expects to earn her PhD by the end of 2006.


Permanent Link: http://ufdc.ufl.edu/UFE0016880/00001

Material Information

Title: I. NADPH Regeneration Sources in Non-Growing, Engineered Escherichia coli Cells II. Old and New Alkene Reductases as Potential Biocatalysts
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0016880:00001

Permanent Link: http://ufdc.ufl.edu/UFE0016880/00001

Material Information

Title: I. NADPH Regeneration Sources in Non-Growing, Engineered Escherichia coli Cells II. Old and New Alkene Reductases as Potential Biocatalysts
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0016880:00001


This item has the following downloads:


Full Text











I. NADPH REGENERATION SOURCES IN NON-GROWING, ENGINEERED
ESCHERICHIA COLI CELLS
II. EVALUATION OF OLD AND NEWALKENE REDUCTASES AS POTENTIAL
BIOCATALYSTS













By

DESPINA BOUGIOUKOU


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

UNIVERSITY OF FLORIDA


2006


































Copyright 2006

by

Despina Bougioukou


































To my loving parents:'Eppa and Fidvvl.
















TABLE OF CONTENTS

page

L IS T O F T A B L E S ........................ ................. .......................................................... v iii

LIS T O F F IG U R E S .................................................. x

A B S T R A C T ........................ ............................................................................x v i

CHAPTER

1 UNCOVERING THE NADPH REGENERATION PATHWAYS IN E. COLI:
TOOLS AND METHODOLOGIES.................................................. 1

Introduction ........................................................... ..... ............ 1
The Past: Pentose Phosphate Pathway Partition .................................................. 4
Radiorespirometric Measurements .......... ............. .. ............ 8
Destructive Techniques......................................................... 10
Measurements based on labeled nucleic acids ....................................... 10
Measurements based on labeled amino acids................... 11
Regulation of Pentose Phosphate Pathway ............... ........... .............. 13
The Present: The Interplay of NADPH with the Metabolic Network ..................... 14
Theoretical M ethods................................. .......................15
In silico estimations for NADPH-related fluxes in E. coli ...........................20
Computational methods: limitations ....... ........ ..... ......... .. 22
Experimental Methods ........... .........................................23
NMR analysis for 13C labeled amino acids ................... ................. 24
MS analysis for 13C-labeled amino acids ................. .............. 27
Estimated fluxes from 13C labeling experiments ..............................27
Other experimental methods: metabolic regulation analysis of NADPH
related metabolic pathways at gene and protein expression level. ....... 35
Experimental methods: limitations ............................... ...... ......... 37
Conclusion .............. .. ................................... .......... ....... ......... 38

2 INVESTIGATION OF NAPDH REGENERATION SOURCES IN RESTING E.
COLI CELLS ENGINEERED TO OVERPRODUCE A NAPDH-DEPENDENT
KETOREDUCTASE ................................. ..... ........... 40

Introduction .............. .................. .... ................. ........40
Examples of Enzymatic NADPH Regeneration Methods ................................43
Examples of Microbiological NADPH Regeneration Methods.........................45
Permeabilized E.coli cells ............. ........................... 45
Intact E.coli whole cells ............. .............................48
Economical Analysis ................... ................. ...... ................. 50









Manipulations of NADP(H /) Levels in E. coli towards the Improvement of
Biotechnological Processes............. ....... .......................... ................... 52
Description of the System under Study............... ....... ... ..... ................. 54
Experimental Strategy ..................................... ............. ... .......... ...... .... .. 56
Experim ental P rocedures............................ ................................. ................. 61
Materials................................................ 61
Bacterial Strains and Plasm ids .............................................. ................. 61
C e ll C u ltiv a tio n s ................................................................... 6 1
A na lytica l M e th o d s .............. .. ... ....... ........................ .. .... ......... ............... 6 2
In vitro Construction of a Linear Targeting Fragment by PCR: Synthesis of
Targeting Marker with Long Flanking Homologous Regions (LFH)..............63
G ene R eplacem ent ............ ............. .. .... ......... ............ .. .............. .. 64
Detection and Analysis of Recombinants....... ........ .. .. .. .................64
Deletion of the Inserted Marker....................................................65
Preparation of [4S-2H] NADPH ................ .......................... .................65
Preparation of [4R-2H] NADPH ................ .......................... .................66
Preparation of D-glucose-ld ........................................ ................ ................. 66
Reduction of Ethyl-acetoacetate by Gre2p GST-fusion Reductase Using
N A D P D B or N A D P D A..................................................................................... 67
Reduction of Ethyl-acetoacetate by Gre2p GST -fusion Reductase Using a
m ixture of NADPH and NADPDB ............. ................... ............................ 67
Standard Procedure for Ethyl-acetoacetate Biotransformations with Resting
ce lls ............... ................................... ........................... 6 7
R results ................. ........... ......... ............. 68
Characterization of M utated Strains................................... .......................... 68
M S Data Analysis ........... ............................ ......... ... ...... .. .. .... ... 76
Evaluation of the MS Data .................. .................................85
Physiological conditions ................................ .................. ........... ...85
Calculation of the intracellular ratio NADPH/NADPD from mass data.......86
Productivity of Mutant Strains versus the Wild Type in a Shake Flask............. 93
D discussion ............... ......... ............................. ........................... 95
Future W ork .............. ..... ... ........ ..... .......... .... ..........99

3 E. COLI BL21 (DE3) AyqhE STRAIN: A 'CLEANER' HOST FOR
OVERPRODUCTION OF NADPH-DEPENDENT KETOREDUCTASES.............. 101

Introduction ............................................. .......... ................ 101
Carbonyl Reductases in the E. coli........ ......... ................................... 102
Experim ental Procedures.......................................................... .... ................. 107
Results and Discussion .................................................. .... .. ..... ..... ..... 107
Characterization of AyqhE Strain ...... ............ ....... ........ ................... 107
Biotransformations .......... ....................................... .. ... ..... ... ...... 109
Conclusion ..... .. .................................. .................. 112

4 BIOCATALYTIC REDUCTION OF FUNCTIONALIZED ALKENES: A SURVEY
FOR EN-REDUCTASES .............. ... ........ ... ........................ 113

Introduction .................................. ..... .................. ............ 113
Biohydrogenation with W hole Cells..................... ... ... ................. 114
En-Reductases: Substrate Specificity ................................. ......... ... ............. 122
Old Yellow Enzyme Family ...... ............... .............. 123


v









Old yellow enzyme from Saccharomyces carlsbergensis .............. 123
Old yellow enzymes from yeasts........... ...................................... .. ......... 133
Old yellow enzymes from bacteria ......... ...... .................... ........... 136
O ld yellow enzym es from plants ................. ................. .. ............... 141
Enoate Reductases ............................. ......... ...... ........... 145
Structure-Function Studies among Flavin-Containing En-Reductases .......... 149
Medium Chain Dehydrogenases Possessing En-Reductase Activity ............ 158
Short Chain Dehydrogenases Possessing En-Reductase Activity ................. 167
U classified En-Reductases ....................................................................... 168
Cofactor Specificity of En-Reductases... ............................... 173
C including Rem arks .......... ........... .......... .... ...................... ............. 175

5 IN VIVO RECOMBINATIONAL CLONING OF OLD AND NEW EN-
REDUCTASES: A COMPARATIVE SUBSTRATE SPECIFICITY STUDY ........... 177

Introduction ............................. .... ... ...... ...................... 177
Experim mental Strategy ........... ................... .. .... ............. ................. 179
Selection of Alkene Reductases Clones ........................................ ............... 179
Cloning and Expression of Alkene-Reductases ................. ........ ........ 183
Experimental Procedures ......... ... ........... ..... ................. 185
P purification of Y N L134c.............................................................................. 185
M materials and supplies ......... ...................................... .. .................. 185
Enzym e activity assays ................................................. ........................... 185
Cell growth for protein isolation............ ..................... 186
Preparation of crude extract........... ...................... 186
Streptomycin sulfate purification ............... ....................................... 187
Desalting by Sephadex G-25 ......... ............ ......... .... ........ 187
DEAE CL-6B anion exchange chromatography................ .................. 187
HiLoad Q sepharose anion exchange chromatography.......................... 187
Dyematrex gel blue A affinity chromatography ....................................... 188
Superdex 200 filtration chromatography ........................................ 188
HiLoad Q Sepharose anion exchange chromatography ..... .......... 189
Sephacryl S-300 Hi Prep filtration chromatography................................ 189
Protein Sequencing ................... .. .. ......... .. ........... ........... ... .. ......... .. 189
Cloning and Purification of GST-Alkene Reductases.................................... 189
Biochem icals, plasm ids and strains .................. ................................... 189
Cloning of alkene reductases.......... .. .. ... ............... 190
Isolation of GST-fusion proteins................. .................. .......... ....... 191
Regeneration of glutathione resin ............ ..... ................. 192
Screening of A lkene Reductases.......................................... .......................... 192
Preparation of 6-alkyl-2-cyclohexen-1-ones ................ ....... ......... 192
Preparation of neral and geranial..................................... ...................... 193
Biotransform nations .............. .. .. .. .. .. .............................. ............. 194
Detection of reduced activated alkenes .............................. .................... 194
R results and D discussion ......................................... ........ .. ...... ..... ........... 195
Isolation of YNL134c ....................................... .........1. 95
Cloning and Expression of Alkene reductases........... ................203
Construction of universal E.coli expression vector......................................... 203
Construction of a yeast shuttle universal E.coli expression vector ................ 204
Cloning of GST-alkene reductases ............. ................... 205
Expression of GST-alkene reductases ...................................209









Screening of Alkene Reductases ............................. ................................ 210
Concluding Remarks and Future W ork ........ .................................................229

APPENDIX

A ILLUSTRATION OF CONSTRUCTION OF GENE KNOCK-OUTS IN E.COLI......232

B LIST OF STRAINS AND PLASMIDS ............ .. ......................... 234

C LIST O F PRIM ERS .............. .. ..... .... .... ................. ................. ... 235

LIS T O F R E FE R E N C ES ......... ................. ........................................ ................. 241

BIO G RA PH ICA L S KETC H ............. .................................................... ................. 263
















LIST OF TABLES


Table page

1-1 Tabulation of glucose passage through the PP pathway.................................7

1-2 Estimated fluxes though the PP pathway. ................................. ....... .......... 16

2-1 Cost of nicotinamide cofactors and their content in E. coli. ...............................41

2-2 Cost analysis of different NADPH regeneration methods and tabulation of
T T N s ............................... ...................................................................................... 5 1

2-3 Cofactor levels of PHB producing strains and tabulation of TTN...........................54

2-4 Labeled sugars. Cost and possible application. .............. ................. 58

2-5 E. coli strains used in this study ........................................... ................68

2-6 Colony PCR for pgi knock-out. ..............................................................69

2-7 Colony PCR for zwf knock-out........ ................... ......... ............... ................. 71

2-8 Colony PCR for pntAB knock-out. ...........................................73

2-9 Fractional abundance of NADPD...................... ..........................91

3-1 Putative A KRs in E. coli. ............................ ............................... ................. 104

3-2 Colony PC R for yqhE knock-out .............................................. ........................... 108

3-3 P-keto ester reductions by the starting and yqhE knockout strains. .................... 110

4-1 List of OYEs from yeasts and corresponding substrate specificity. ....... ........ 135

4-2 List of OYEs from bacteria and corresponding substrate specificity .................. 138

4-3 A comparative study among different OYEs from yeast and bacteria. ................ 141

4-4 List of OYEs from plants and corresponding substrate specificity...................... 143

4-5 List of ERs and representative substrate specificity ........................................ 149

4-6 En-Reductases in MDR superfamily................................ 164

4-7 En-Reductases in SDR superfam ily. ............... ............................. ................. 168









4-8 List of unclassified en-reductases........................................ 171

4-9 Cofactor specificity of En-Reductases. ......... ..... ................. 174

5-1 List of alkene reductases cloned in this study. ........................ .................... 181

5-2 Percent identity in between OYEs cloned in this study ....................................... 182

5-3 Starting materials for preparation of 6-substituted cyclohexenones.................... 193

5-4 Observed reduction of 1-octen-3-one or 2-cyclohexenone................................ 197

5-5 Purification table for Y N L134c .............. .......... .. ...........................................199

5-6 Final plasmid constructs of akene reductases................. ................208

5-7 Unsubstituted cyclo-alkenes. ................. ...... ..................... 213

5-8 a-substituted cyclo-alkenes. ........ ............................................. .... ................... 213

5-9 P-substituted cycloalkenes.............................................. ................................ 2 17

5-10 6-substituted cyclohexenones. ........................................ ..............................217

5-11 Perillaldehyde and derivatives. ......................... .................... .......... 219

5-12 N-substituted m aleim ides. .... ........................................................................... 223

5-13 Straight chain alkenes. ......................................................................... 224

5-14 S substituted enals. ................ .......... .. .. .. .. .... ............... .... .......... 226

5-15 Nitro-akenes. ............. ................ ................... .......................229

C-1 Oligonucleotides used to create pgi mutant........................................................235

C-2 Oligonucleotides used to create zwf mutant. ............... ....................................236

C-3 Oligonucleotides used to create pntAB mutant....................................... .............237

C-4 Oligonucleotides used to create the library of En-Reductases................... 238
















LIST OF FIGURES


Figure page

1-1 Overview of central metabolic pathways involved in NADPH formation in E. coli...2

1-2 NADPH sources in E. coli metabolism.................... .............................................. 3

1-3 The fate of G6P, labeled in different positions, via EMP (blue frame) and PP
(green fram e) pathways............. ........................ ......... .. .. .. ...... ... 5

1-4 The fate of labeled ACoA in TCA cycle. Enzymes are indicated by their 3-letter
c o d e ............................................................................... 6

1-5 Tools and methods used in the last 25 years for elucidation of metabolic
pathw ays ........... ........................................................ ......... 15

1-6 Basics of Flux Balance Analysis approach ...... ..... .................... ..................21

1-7 Synergy of experimental and computational methods towards the calculation of
m etabolic fluxes. ........................ ..... .......... ................................... 25

1-8 Basics of NMR analysis of biosynthetically directed fractional 13C labeling amino
acids. ......................................... ....... ..... ...... .. ........... 26

1-9 MS analysis: mass isotopomer distribution (MID) for the molecular ion. ..........28

1-10 Specific rates of NADPH production in glucose (C)- and ammonia (N)-limited
chemostat cultures of E. coli W 3110 .. ... .... ......................... ............. .... 31

1-11 The percent of NADPH produced by PP pathway, TCA cycle and PntAB
transhydrogenase activity in E. coli during standard aerobic batch growth on
glucose. .............. .......... ... .... ....... ......... ......... .33

1-12 Glucose transport systems in E. coli.............. ............................ ................. 34

1-13 Poly(3-hydroxybutyrate) synthesis: the acetoacetyl-CoA synthase requires
NA DPH .............. ..... ..... .... ............ ........................... 36

2-1 Explanation of the relative prices among nicotinamide cofactors.......................41

2-2 NAD(P)H regeneration strategies and major disadvantages for each method......42

2-3 A substrate-coupled enzymatic regeneration method for NADPH.........................43









2-4 Glucose-6-phosphate dehydrogenase (G6PDH) as an auxiliary enzyme ............44

2-5 Glucose-6-phosphate dehygrogenase as an NAPDH regeneration catalyst.........44

2-6 Hydrogenase I from Pyrococcus furious for NADPH regeneration...................45

2-7 Example of two coupled E. coli strains in a two-phase system .............................. 46

2-8 Example of an E. coli cell overexpressing both GDH and a reductase. ................46

2-9 Example of an E. coli cell overexpressing two auxiliary enzymes and a
reductase. ......... .. ...... ............................. ................... ......... ........ 47

2-10 Example of two E. coli strains coupled for regeneration of NADPH in one phase
system ................. ................................................. ... ....... .. 48

2-11 Example of intact E. coli cells capable to regenerate NADPH for ethyl
acetoacetate reduction by a ketoreductase. ................................... .................. 49

2-12 Example of intact E. coli cells overexpressing a G6PDH and a reductase ...........49

2-13 Cost of NADP+ as a function of TTN ......... .... ... ...... ................. 50

2-14 Cartoon represents the system under study............. ......... ................. 57

2-15 Illustration of the experim ental procedure.................................. ..... ............. ....59

2-16 Results from colony PCR reactions probing the pgi locus before and after the
disruption. ............... ..... ... ........... ....... .......... .... ... .......... 70

2-17 Results from colony PCR reactions probing the zwf locus before and after the
disruption. ............... ..... ... ........... ....... .......... .... ... .......... 72

2-18 Results from colony PCR reactions probing the pntAB locus before and after
the disruption. .................. ................................... ....................... 74

2-19 Phenotype analysis of constructed mutants in M9 medium.................................75

2-20 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
wild-type strain (WT) using Glc-1D, under aerobic condition.................................76

2-21 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
wild-type strain (WT) using Glc-1D, under anaerobic condition .......................... 77

2-22 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
wild-type strain (WT) using Glc-1 D, under aerobic and anaerobic conditions. .....78

2-23 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
wild-type strain (WT) using Glc-3D, under aerobic conditions............ .............. 78

2-24 Fate of deuterium atoms of glucose at position 1, 3, 4 and 6 through glycolysis... 79









2-25 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
w ild-type strain (W T) using G lc-6,6 D2,. ...................................... ................... 80

2-26 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by Apgi
stra in using G lc-6 ,6 D 2 ......................... ... .... .... .... .... .... .... ..................... ..... 8 1

2-27 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
Apgi using Glc-1D. ............. ......... .... ........... ... ............... 81

2-28 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by Apgi
strain using Glc-3D. .......... ........................ ... .... ..................... 82

2-29 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by Azwf
strain using Glc-3D, or Glc-1 D............ ......... .... ......... .................. 82

2-30 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by Azwf
strain using Glc-6,6 D2 ............................... .... .... ................... .................. 83

2-31 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by Aicd
strain using G lc-1D .......... ...... ... ....... .. .. ............. .......... 83

2-32 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by Aicd
strain using Glc-6,6 D2 ............................... .... .... ................... .................. 84

2-33 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by ApntAB
strain using G lc-1 D, or G lc-6,6 D2................................................ ................... 84

2-34 Production of EHB by WT strain fed with glycerol under aerobic and anaerobic
c o n d itio n s .............................. .................................................... 8 5

2-35 Production of EHB by WT strain fed with glucose under aerobic and anaerobic
c o n d itio n s .............................. .................................................... 8 6

2-36 Equation for derivatization of the corrected A71/A72 from the observed ration....... 87

2-37 Proposed correction for the NADPD fraction based on the corrected A71/A72
ratio and a kinetic isotope effect factor fKIE .......................................................... 88

2-39 Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by Gre2 in
presence of NADPH and NADPD........ ... ...... .......... ................. 89

2-38 Preparation of stereospecific labeled NADPH and elucidation of Gre2 facial
se lectivity by H -N M R ................. ... ... ............ ... ... ... .. .. .. ........... .. ... .... 90

2-40 Product formation of Gre2 overexpressed either in Azwfor Aicd or ApntAB
s tra in ...................... .. .. ......... .. .. ............................................... 9 4

2-41 Product formation of Gre2 overexpressed either in Apgi orWT strain................... 94

2-42 Proposed participation of NADPH regeneration sources in E. coli
BL21(DE3)(pAA3) resting cells ............. .. .. ......... ....... .......... 96









2-43 Proposed participation of NADPH regeneration sources in E. coli
BL21(DE3)(pAA3) mutant stains under non-growing conditions...........................97

3-1 Substrate specificity of dkgA .............. ...... ... ................ ...... ........... 105

3-2 Stereoview of the ribbon structure of DkgA. ............... ........ .. ................ 105

3-3 Multiple alignment of eight AKRs from E .coli.................................................. 106

3-4 Results from colony PCR reactions probing the yqhE locus before and after the
disruption. ............ ....................................... ... .. ...... .............. 108

3-5 Growth curves for E. coli BL21(DE3) and AyqhE strains.................................. 110

3-6 The four possible stereoisomers of ethyl 2-fluoroacetoacetate reduction............ 111

3-7 Profile of ethyl 2-fluoroacetoacetate reduction from E. coli BL21(DE3) (VVT) and
AyqhE (MT)strains.. ......................................................... 111

4-1 Preparation of a building block with two chiral centers from citral using bakers'
yeast as biocatalyst. .............. ...... .... .... .... ........ .. .. .. ...... .......... .. 115

4-2 Proposed hydrogenation of allylic alcohols by bakers' yeast ............................ 115

4-3 Chemo-selective reduction of a trienal by bakers' yeast............... ............ 115

4-4 Preparation of goniothalamin, a putative anticancer compound........................ 116

4-5 Preparation of raspberry ketone by bakers' yeast reduction. ............................ 117

4-6 Servi's rules for A and B type reduction in bakers' yeast.................................. 117

4-7 The first reported enone reductase activity in bakers' yeast.................................... 118

4-8 Bakers' yeast versus chemical mediated reduction of 4-oxo-isophorone ........... 118

4-9 The six possible microbial transformations of zearalenone............................... 119

4-10 Preparation of griseofulvin, an antifungal agent. ............... ............................. 120

4-11 Biohydrogenation of abscisic acid, a plant stress hormone............................... 120

4-12 General scheme of biohydrogenation with Clostridia species. .......................... 121

4-13 A proposed classification of en-reductases. ........................... ................. 122

4-14 First generation of substrates and inhibitors for OYE1 ................ .................. 125

4-15 Representation of catalytic cycle of Old Yellow Enzyme.................................. 127

4-16 Second generation of substrates for OYE1. .................................................. 128









4-17 Relative reactivity of substituted en-ones (als) based on Massey's studies........ 129

4-18 Stereoview of active site architecture of OYE1 with para-hydroxy-benzaldehyde
(10 Y B) .................. ..................... ....................... ... .......... 129

4-19 Stereochemical outcome of reduction of substituted cyclic enones by OYE1..... 129

4-20 First generation of desaturation reactions by OYE1 in absence of NADPH........ 130

4-21 Proposed orientation of ODE above the si face of flavin. ................................. 131

4-22 Second generation of binding ligands for OYE............................. ................. 131

4-23 Second generation of desaturation reactions by modified-OYE in absence of
N A D P H ....................................................... ................. 132

4-24 Third and fourth generation of substrates for OYE1.............. .......... .......... 133

4-25 Jasmonic acid biosynthesis. ............................ .. .. .................... 142

4-26 Biohydrogenation of unsaturated carboxylic acids from three different types of
enzymes. ........ .................................. .................. 146

4-27 Reductive part of the phenylalanine fermentation pathway in Clostridia
sporogenes. ........................................................ ...................................... 147

4-28 Close up view of the catalytic microenvironment of OYE1 before (10YA) and
after (10YB). ....................... .. ..... ........................... .......... 150

4-29 Superposition of OYE1 and YqjM structure. ............................ ............... 153

4-30 Multiple alignment of twelve en-reductases belonging to OYE family. ................ 154

4-31 Superposition of OYE1 and Leoprl structure .............................. ....... ........... 155

4-32 Close-up view of loop3 after superposition of five OYE crystal structures........... 156

4-33 Bird's view of the overall structure of the self-inhibited Leopr3 dimer. ............... 157

4-34 Structural differences in between the flavor-bearing en-reductases.................. 158

4-35 The bifunctional action of AOR. ....... ........................... 159

4-36 Close up view of the hydrophobic pocket of AOR from guinea pig.................... 161

4-37 Proposed catalytic mechanism of the reductase activity of AOR from guinea
p ig ... ............... .................................. ......... ........... ....... 16 1

4-38 Multiple alignment of the four purified AORs. .......................... .................. 162

4-39 Multiple alignment of zinc independent en-reductases belonging to MDR
superfamily. ...... .................................... .................. 163









4-40 Superposition of two en-reductase structures belonging to the MDR
supe rfa m ily .. ....................................................................... 16 7

4-41 Multiple alignment of the two known short chain en-reductases ....................... 169

4-42 Stereoview of the catalytic active site of porcine carbonyl reductase.................. 169

5-1 Different approaches towards the identification of novel biocatalysts................. 178

5-2 Construction of the alkene reductase library. ..................... .......................... 180

5-3 Protein gel after the Superdex 200 filtration column .............. ........... ......... 198

5-4 Protein elution pattern from Mono Q column during the NaCI gradient...............201

5-5 Protein elution pattern from Sephacryl S-300 Hi Prep filtration column and
protein gel for active fractions ....................................................................... 201

5-6 Sequence of Y LN 134c ......... .................................................. ................. 202

5-7 Multiple alignment of YNL134c subfamily with FAQR amino acid sequence.......203

5-8 Construction of E. coli expression vector pDJB3 .................. .................... 204

5-9 Construction of the E. coli expression vector pDJB2...................................... 205

5-10 C construction of pDJB5. ............ ......... .. ........................ ..... ................. 206

5-11 Homologous recombinational cloning into universal vectors pDJB2 and pDJB3.207

5-12 SDS-PAGE analysis of total cell extracts overexpressing alkene reductases.....210

5-13 SDS-PAGE analysis of alkene reductases after the affinity column.................... 210

5-14 Predicted stereochemical outcome for 2-methyl-2-cyclohexenone .................215

5-15 Predicted stereochemical outcome for 3-methyl-2-cyclohexenone .................215

5-16 Predicted stereochemical outcome for (S)-perillaldehyde. ................................ 221

5-17 Predicted stereochemical outcome for (R)-perillaldehyde..................................221

5-18 An alternative, flipped geometry of (S)-perillaldehyde in the catalytic center of
O Y E ..................... .. .. ......... .. .. .................................................. 2 2 2

5-19 Proposed mechanism for the Ltb4dh-mediated reduction of perillaldehyde........ 223

5-20 Proposed mechanism for the reduction of geranial or neral by OYEs ................228
















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

I. NADPH REGENERATION SOURCES IN NON-GROWING, ENGINEERED
ESCHERICHIA COLI CELLS
II. EVALUATION OF OLD AND NEWALKENE REDUCTASES AS POTENTIAL
BIOCATALYSTS

By

Despina Bougioukou
December 2006

Chair: Jon D. Stewart
Major Department: Chemistry

In the first part of this dissertation, an effort was made to define the pathways by

which the NADPH is regenerated in E. coli cells under non-growing conditions. We used

the reduction of ethyl acetoacetate by a Saccharomyces cerevisiae reductase (Gre2)

overexpressed in E. coli as a model system and specifically deuterated glucose (Cl-dl,

C3-dl or C6-d2) as isotopomer carbohydrate sources. Deuterium atoms originating at C1,

C3 or C6 of glucose are first transferred to NADP+ and then appeared at the C3 atom of

ethyl-3-hydroxybutyrate, allowing us to measure the isotopic composition of the

intracellular NADPH pool by GC/MS. In addition, we perturbed the NADP(H/ ) levels of

the host cell by environmental and genetic manipulation means and evaluated the

effects of those perturbations on the productivity of the system under study. Our data

showed that the NADPH sources present in E. coli cells under growing conditions are

also active in non-dividing cells and most likely are not the limited factor for the

maximum productivity of this particular biotransformation.









In the second part, a set of 15 genes encoding known and putative alkene

reductases were successfully cloned by a rapid recombination technique and

overexpressed as GST-fused proteins in E. coli. Their ability to reduce ca. 40 simple

enones and enals was evaluated. This study uncovered subtle differences in substrate

specificity among the distinct en-reductase subfamilies, as well as the function of a

previously uncharacterized in S. cerevisiae gene product.















CHAPTER 1
UNCOVERING THE NADPH REGENERATION PATHWAYS IN ESCHERICHIA COLI:
TOOLS AND METHODOLOGIES

If you're converting carbohydrate into triglyceride,
If you need pentose moieties to make nucleotide,
You'll find that Embden-Meyerhof is not the game to play
And you'll do your biosynthesis the pentose phosphate way
-The Pentose Phosphate Shunt Song, Author unknown1

Introduction

The reduced form of nicotinamide adenine dinucleotide phosphate, NADPH, plays

a central role in fueling biosynthetic reactions and maintaining the redox potential

necessary to protect cells against oxidative stress.2 A careful inspection in the model-

organism database for Escherichia coli K-12 (EcoCyc, version 10.0)3 reveals five

potential routes for NADPH formation. The exact location of each in the carbon central

metabolism of E. coli is depicted in Figure 1-1. The pentose phosphate pathway (PP

pathway) provides NADPH in two reactions catalyzed by glucose-6-phosphate

dehydrogenase (zwf) and gluconate-6-phosphate dehydrogenase (gnd) respectively

(Figure 1-2A); the tricarboxylic acid cycle (TCA) produces NADPH to a lesser extent by

isocitrate dehydrogenase (icd, Figure 1-2B). In addition, an anaplerotic reaction

catalyzed by the malate dehydrogenase (maeB, Figure 1-2C), as well as the membrane-

bound pyridine nucleotide transhydrogenase (pntAB, Figure 1-2D), could serve as

catalysts for the regeneration of NADPH. Even though the Embden-Meyerhof-Parnas

(EMP) pathway is not directly involved in the NADPH synthesis, the branching points

which connect it with the aforementioned pathways are of fundamental importance for

control of NAPDH production. The efforts that have been made the last six decades









towards the identification and estimation of the partition of each NADPH-producing step

to the overall NADPH pool in E. coli will be described.


NADPH


Anaplerotic
~NADPH C02 O2
*L VkLoae


- p
S7P h
PYRW

y
H

NADPK NADH NADPH + NAD


Figure 1-1. Overview of central metabolic pathways involved in NADPH formation in E.
coli. Enzymes are indicated by their 3-letter code. Abbreviations: G6P,
glucose-6-phosphate dehydrogenase; F6P, fructose-6-phosphate; 6PG,
gluconate 6-phosphate; P5P, pentose 5-phosphate; E4P, erythrose 4-
phosphate; S7P, sedoheptulose 7-phosphate; T3P, triose 3-phosphate; 3PG,
glyceraldehyde 3-phosphate; PGA, 3-phosphoglycerate; PEP,
phosphoenolpyruvate; PYR, pyruvate; ACA: acetyl coenzyme A; OAA,
oxaloacetate; ICT, isocitrate; AKG, a-ketoglutarate; SUC, succinate; FUM,
fumarate; MAL, malate; GOX, glyoxylate. Pgi, phosphogluconase isomerase;
zwf, G6P dehydrogenase; gnd, 6PG, 6-phosphogluconate dehydrogenase;
PntAB, membrane-bound transhydrogenase; sthA, soluble
transhydrodenase; maeA, MAL dehydrogenase; icd, ICT dehydrogenase; PP
pathway, pentose phosphate pathway; EMP, Embden-Meyerhof-Parnas
pathway; ED pathway, Entner-Doudoroff pathway; TCA cycle, tricarboxylic
acid cycle. (Adapted from references 60 and 50).













A 9H
HO-P-0 OH

HO'" "OH
OH

i-D-glucose-6-phosphate


glucose 6-phosphate-1-
dehydrogenase: zwf


NADP+ NADPH


OH
HO-P-O 0

HO' 'OH
OH


D-glucono--lactone-6-phosphate


OH O0
HO 0
HO6-po HO'' '' OH
0 H" "OH

6-phospho-D-gluconate


O 0
OH V O

0' -OH
HO H O

isocitrate


6-phosphogluconate
dehydrogenase
idecartboylalu.ngp gnd


NADP4 NADPH


isocitrate dehydrogenase:
icd


NADP+ NADPH


HO-P-O0

HO OH
0

D-ribulose-5-phosphate


OH 0

0 -"OH
0
a-ketoglutarate


+ C02


+ CO2


O H OH
oA OH
0ala

malate


ralate dehydrogenase
Iowalcoacel le dlec:rt.oryla ilmg
maeB

NADP NADPH
NADP+ NADPH


D HO H NH2
N N

O=P-OH H
0
0=P-OH R Hn
0 = 0 N T HO 2
H / ONH2
HO '''OH
nicotinamide adenine
dinucleotide-reduced, (NADH)


pyridine nucleotide
transhydrogenase: pntAB


NADP+ NADPH


HO OH NH2



O=P-OH H
0 H
O=P-OH

-/ NH2
HO 'OH
nicotinamide adenine
dinucleotide-oxidized (NAD*)


Figure 1-2. NADPH sources in E. coli metabolism. A) both Zwf and Gnd are belonging
to the PPP, B) the Icd is a part of TCA cycle, C) MaeB participates in an
anaplerotic reaction, and D) the PntAB catalyzes the reversible transfer of a
hydride ion equivalent between NAD(H) and NADP(H).


OH

0
pyruvate


+ CO2









The Past: Pentose Phosphate Pathway Partition

The PP pathway, also known as the hexose monophosphate shunt (HMS), can be

considered as a metabolic scrambler for the cells. It is an alternative pathway for

glucose metabolism that shunts some G6P away from the EMP glycolytic pathway

oxidativee part) and returns parts of the diverted hexose carbons to the latter pathway as

F6P and 3PG (non-oxidative part, Figure 1-3). The extent of the contribution of the PP

pathway to glycolysis has been a subject of numerous investigations, as this pathway,

among other functions, provides ribose-5-phosphate (R5P) which is needed for nucleic

acid synthesis and reduces NADP+ to NADPH, which is needed for amino acid and fatty

acid synthesis as well as for the regeneration of glutathione, a key peptide in the

antioxidant protection of the cells. In addition, recent studies4 have shown that pathways

involved in redox-metabolism also play an important role in biosynthesis of metabolites

of industrial importance. Indeed, the knowledge of the PP pathway flux provides a basis

for better understanding of the cell physiology and serves as a guide for directed

improvements of performance of industrially exploited organisms.5

In classic isotope studies of metabolism, the fate of a label at a specific position of

a substrate is 'traced' by examining the incorporation of the labeled atom into the end-

products (e.g. CO2) or intermediates (e.g. NADPH) of the investigated pathways.

Different pathways, even though they utilize the same labeled substrate, produce end-

products which are labeled at different positions and to different extents since distinct

enzymes are participating in each of pathway. Thus, using the same substrate, in this

particular case glucose, and labeling it at a different position (Figure 1-3 and Figure 1-4)

each time, investigators estimated the percentage of glucose that is metabolized through

the PP pathway. Their efforts are summarized in Table 1-1 and will be described in the

following pages.
















OH
HO- -H 3 NAPD+ 3 NAPDH
4 }OH b
H -OH
zwf, pgl


O=P-OH
OH
G6P


OH



3 Oh OH
H- -OH


O =P-OH
OH


3 NAPD+ 3 NAPD


gnd 3 OJ


G6PG nrpe
H
0+OH
,=O


afkAB
baAB *


f

H


OH

H -OH



SF -OH
On
F6P

pfkAB
fbaAB

tPO1-,H

at0 DHAP


.0
3PG H --OH

OPOIH;

NADH --FgapA


HO, C,

H -OH
PGA -H

OPC,H;

ena


- -- H
H H O OH OH_._



H Ho
C -
O"PH ;O H H--OH

H pykAF aeEF 0
3 G IpdA 0-P-OH
o ., 2 OH
e1gapA F6P


S..
PEP n H
'Gti
. ', ano


H




H- -OH

op
O=P-OH
OH Ru5P

IrpiA,dlsl


H- -OH
O+OH
H- -OH


0=P-OH
OH
R5P

H


HO--H

OH
H- -OH



OP-OH
OH




-OH
H- -OH


O 0-OH


H

(5P =0
HO H
H- -OH


O=P-OH
,(a OH


Figure 1-3. The fate of G6P, labeled in different positions, via EMP (blue frame) and PP
(green frame) pathways. Enzymes are indicated by their 3-letter code.
Dashed arrows indicate that more than one step is involved. Hashed lines
indicate the branching points in between the two pathways (Adapted from
references 2, 3 and 68)


HO-


HO-

H

Pi A.u


GPO ,H.

H H P
HO 0


H n -A FI....


AC,-A F


i

















COOH
H 1s Itumr nt gln> H- -H
O HOOC OH
Sg 9gtA H0
DH
M I-S-CoA HO- "O
S O ACoA CIT

HO ..O O COOH



COOHHO
12 tHO- cO
2nd turn of cycle


NADH
NAD(P)+ NADIPPI
rlh "
S ,m'a aAB --
NAD Co, C


1/2
fumABC


COOH
H+OH
- HOOCH
acnAB

HO' "0
ICT

COOH

......) HOOC--H


HO"' '0


COOH
NADP' NADPH
S H H:H
icd
CO2
HO" '0
AKG


NADP* NADPH


CO2


C02 NAD*


3rd turn ot cycle >



H AO

PYR
PYR


O3
1/2 I -'OH

HOOC H
FUM


H COOH

HO 1/2 __ 1/2
I sdhABCD
O0


0 ,1OH



COOH


NAD
dA
BCD


C

1/2 '

HO-


COOH


HO- -H

HO "O



:H NAD*




NADH



OOH
-H

"O


HO- .O OH


HOH 4OH
O2H 02 OH
112 1/2


0

.----- "OH

HO Cf H
1/2


0,-OH OFH



HH OH HO'" O
1/2 1/2


H 0 OOH

HO, lI
0


Figure 1-4. The fate of labeled ACoA in TCA cycle. Enzymes are indicated by their 3-letter code. Dashed arrows and lines indicate
the fate of oxaloacetate acid (OAA) after the second and third turn of cycle. Hashed lines indicate the interconnection
between the TCA cycle and the anaplelotic reaction catalyzed by maeA or maeB. Open cycles indicated partially labeled
atoms (Adapted from references 3 and 24)


COOH

O 0
0" '-OH
112


HO, ,.O COOH
H H
1/2 H OH
HtOH T
COOH O^ 'OH
MAL


I


. . .









Table 1-1. Tabulation of glucose passage throu h the PP pathway.
Year (Ref.) E. coli Methods and Tools Physiological PPPa (%)
strain conditions
1950 (6) B 14C-1glucose, 14CO2 Growing cells 22-38
recovery measured Resting cells <1
Growing cells 6
infected with T2
virus
1953 (7) B Activity of zwf and gnd Exponential 40
in cell-free extracts aerobic growth
on glucose
1958 (8) B C14-labeled glucose at Resting cells in 28
1 or 2 or 3 or 6 media optimum
position, for growth
rediorespirometric
measurements
1962 (9) W Glucose-1-018 or Rapid growth 30
glucose-6-018 recovery Slow growth 10
of labeled CO2
1967 (10) WT Glucose-1-018, ratio of Exponential 25
isotope recovered CO2 growth in salts
medium
Growth period in 16
salts medium
Stationary phase 10
in salts medium
During or after 13-19
nitrogen
deprivation
Minimal media 13-14
with yeast
extract or
nutrient broth
Anaerobiosis 7
Anaerobiosis 13
plus nitrate
1967(11) K-10 Glucose--14C, Growing cells Both PPP and
gluconate-l-14C, radio- EMP pathway
labeled alanine active, ED
measurements pathway
inactive
K-10 PPP major,
Apgi ED pathway
minor extent
1968 (12) B Glucose-1-018, 018 in Growing cells 30
ribose and deoxyribose during early
moieties measured by logarithmic
liquid chromatography phase
MS
aPPP: Pentose Phosphate Pathway









Table 1-1. Continued.
Year (Ref.) E. coli Methods and Tools Physiological PPP(%)
strain conditions
1969 (13) B [1-18O], [2-18O] and [6-18O] Growing cells in 28-30
glucose and [2-180] minimal media
fructose 018 in ribose and contained a
deoxyribose moieties limited amount of
measured by liquid glucose
chromatography MS
1972 (14) K-12 Enzymatic and end Transition from 20-30
products determination aerobiosis to
anaerobiosis
1974 (15) DF100 C14-lebeled glucose at 1 or Growing cells 0.4-fold
(K-12 6 position, during early higher
derivative), rediorespirometric logarithmic rate of
DF102 measurements, phase glucose
(zwf 10- intracellular intermediates oxidation
fold-higher concentration via PPP in
level measurements, in vivo and the mutant
mutant) in vitro rates of glucose strain
oxidation via PPP
1977 (16) K10 [1-3H, U-14C]glucose, [3-3H, Cells growing in 5-10
U-14C]glucose, [4-3H, U- glucose minimal
14C]glucose, [6-3H, U- medium
14C]glucose, amino acids aerobically
-K10 from hydrolysates of the Any <5-
K10 cellular protein were Anaerobically < 5-10
cellular protein were
separated with an amino
K10 Azwf acid analyzer and their Aerobically or 0
3H/14C ratios calculated anearobically
1979 (17) DF567 C14-lebeled glucose at 1 Cells growing 16
(K-12 position, radiolebeled aerobically in
derivative) alanine measurements minimal medium

Radiorespirometric Measurements

As the word implies, radiorespirometric methods measure the respiratory function

of an organism by a combination of CO2 production and radioactive tracer techniques.8

In early attempts, the evolution of 14CO2 from E. coli cells growing under different

conditions in presence of [1-14C]glucose was calculated18 based on the fact that the PP

pathway will convert the C-1 carbon of glucose (colored in red in Figure 1-3) into CO2

much faster than the combination of EMP pathway and TCA cycle. However, the 14CO2

produced from the latter pathways (after 3 turns of TCA cycle, Figure 1-4) complicates









the calculations, but a correction for the TCA cycle-associated contribution to the

recovered 14CO2 was made by performing a parallel experiment8 in presence of [6-

14C]glucose (14CO2 cannot be produced from the PP pathway starting from labeled

glucose at position 6, colored in blue in Figure 1-3).

To circumvent the two major problems inherent with 14C, namely: i) the possibility

of conversion of glucose-6-14C to glucose-1-14C via the isomerization of the triose

phosphate and ii) the fact that the final results depend on the relative degree of

utilization of the 6 carbon atoms of glucose not only for CO2 but also for the synthesis of

various cell components, Rittenberg and Ponticorvo9 developed a procedure based on

C1802 measurements. This approach is independent of the function of TCA cycle. When

either glucose-l-180 or glucose-6-180 passes through the EMP pathway, the 180 is lost

from the molecule at the formation of the PEP and oxidation of PYR yields nonlabeled

CO2. On the other hand, when glucose-l-80 is oxidized to pentose via the PP pathway,

the CO2 produced should contain 180 (colored in red in Figure 1-3). Consequently, the

concentration of 180 in the CO2 will be proportional to the fraction of glucose-l-180 that

enters the PP pathway. Regardless, of the specific labeled glucose used, the results led

to estimations of between 10 and 40 % for the relative participation of the PP pathway

into glucose utilization from the E.coli cells, with high dependence on the physiological

conditions under which the experiments were carried out.10

The validity of the radiorespirometric measurements came under severe criticism

by Katz and Wood,19'20 who pointed out that theoretically, the PP pathway should be

considered as a cycle and recycling of the resynthesized F6P is possible if the latter is

isomerized back to G6P (hashed line in Figure 1-3). In such a case, the degree of

recycling plays a vital role on fate of the labeled CO2, which is diluted by unlabeled CO2,

and the PP pathway partition is underestimated. Nevertheless Katz and Rognstad,21

with the aid of a 7090 IBM computer and a mathematical model that they wrote in









FOTRAN 2 language to fit the experimental data, showed that especially in E. coli cells,

the impact of the carbon randomization during the concurrent operation of the PP and

EMP pathway on radiorespirometric measurements is negligible.

Destructive Techniques

The attractiveness of the radiorespirometric method is based on its simplicity,

since it is non-destructive and the only prerequisite is an apparatus to trap the released

CO2, though only a single measurement can be made on each sample. On the other

hand, due to many assumptions made in the calculations of the results, many

investigators tried to verify those values by performing measurements on end-products

other than C02. Unfortunately, all these methods required the rupture of the cell

membrane with unpredictable consequences on the measurements associated with it.

Measurements based on labeled nucleic acids

Caprioli and Rittenberg 12,13,22 examined the levels of radioactivity in the nucleic

acids of the E. coli cells that had been grown on labeled glucose. The logic of these

experiments lies on the fact that glucose metabolized exclusively by the PP pathway

would produce a nucleoside labeled in the pentose moiety only if it is labeled as [6-180]

([1-O80]glucose will give rise to unlabeled R5P and C1802 while [2-180]glucose would

produce labeled in the C-1 oxygen R5P, but the 180 would be lost in the formation of the

glycosidic linkage with base). The concentration of 180 in the C-5 atom of the pentose

would be equal to that of C-6 oxygen atom of the original glucose. On the other hand, if

glucose was metabolized exclusively via the EMP pathway, both [1-180]glucose and [6-

18O]glucose would give rise to [3-180] 3PG and [2-18O]glucose would form [2-180] 3PG.

Through the reverse action of transketolase (tklAB in Figure 1-3), [5-180] R5P would be

produced from both [1-180]glucose and [6-18O]glucose and [4-180] R5P from the [2-

180]glucose. Hence, the growth of bacteria using [1-80], [2-180] and [6-180]glucose as

the sole carbon sources was terminated in early logarithmic phase, RNA and DNA were









isolated, enzymatically degraded to the nucleoside level, purified further by liquid

chromatography and the content of the 018 label in the pentose of nucleosides was

determined directly by mass spectroscopy (MS). Two major conclusions emerged from

this study: i) in line with the radiorespirometric measurements, 28-30 % of the pentose

arose via the PP pathway, ii) the resynthesis of hexose phosphate is limited to about 3

%.

Measurements based on labeled amino acids

In an effort to investigate the function of the Entner-Doudoroff pathway in E. coli

cells, Zablotny and Fraenke11" isolated a Apgi strain and fed both wild type and mutant

with two different carbon sources, glucose-1-14C and gluconate-1-14C. Labeled glucose

used by the EMP pathway would give rise to methyl-labeled PYR. Labeled gluconate

used by the ED pathway could give rise to carboxyl-labeled PYR and should give

unlabeled PYR when used by the PP pathway. Assuming that the alanine comes from

the transamination of PYR, the authors isolated alanine from lysed cells and measured

the radioactivity. Their results ruled out the function of the ED pathway as a third

possible route in glucose utilization by E. coli cells. Later, Silva and Fraenkel17 estimated

the PP partition at 16 % using the same method.

In all previous studies, the amount of NADPH formed relative to the amount of

consumed glucose can be calculated indirectly from the percentage of PP pathway since

1 mole of glucose yields 2 moles of NADPH when the pentose phosphate shunt is used

exclusively from the cells. In 1977, however, Csonka and Fraenkel16raised a very

interesting question: if a AzwfE. coli strain, which cannot utilize glucose through the

shunt, can grow easily on glucose23 how is it producing the 17 mmols of NADPH

required from biosynthetic reactions to form 1 g of cells? In other words, if the PP

pathway is not essential, how could NADPH be made in E. coli cells? To address this

issue the investigators worked with E. coli strains that had mutations affected either









directly or indirectly, putative NADPH-forming reactions. Specifically one of the following

enzymes was deleted:

* phosphoglucose isomerase (Apgi),
* glucose 6-phosphate dehydrogenase (Azwf),
* isocitrate dehydrogenase (Aicd),
* gluconate 6-phosphate dehydrase (Aedd),
* malate dehydrogenase (Amdh),
* a factor coupling electron transport to ATP formation (AuncB) or a combination of
them such as ApgiAedd, AicdAzwf and ApgiAuncB mutants.

They grew the different strains aerobically or anaerobically in presence of glucose

labeled uniformly with 14C and specifically with 3H in positions 1, 3, 4 or 6 and they

isolated two pairs of amino acids: threonine and aspartate, or proline and glutamate from

protein hydrolysates. This technique ensured that: i) the number of radioactive

hydrogens derived from a particular position of glucose (1, 3, 4 or 6) can be normalized

per carbon atom (according to 14C) in the compound isolated and ii) since the pairs of

the amino acids selected differ by two NADPH-dependant reductions, the different

3H/14C ratios between them reflect labeling of the NADPH pool, even if there were also

present radioactive hydrogens remaining carbon-bound from glucose.24 The most

important observations from this thorough study are:

* Hydrogens from the 3 and 4 positions of glucose should only appear in end-
products from the NADPH-dependent action of the gnd and the NADH-dependent
action of the 3GP dehydrogenase respectively.

* In the Apgi mutant, under anaerobic conditions, the hydrogen at position 3 made
no contribution to the reduction of the end-product isolated where the hydrogen at
position 1 or 6 made a similar contribution aerobically or anaerobically. Therefore,
the PP pathway is not an obligatory source, at least anaerobically, and the
hydrogens at position 1 or 6 of glucose must serve as reductive hydrides via other
pathways.

* The Icd, a possible source for hydrides from the 6 position of glucose, is not
essential even in cells also lacking the Zwf and the radioactivity from the
hydrogens at position 1 and 6 is not decreased in a strain lacking the enzyme.
Thus, a third source should be account for the labeled NADPH found.

* The NADH contribution to biosynthesis cannot be explained with a futile cycle
proposed by Lowenstein25 which involves the ATP-dependent conversion of PYR









to OAA, its NADH-dependant reduction to MAL (catalyzed by mdh) and the
NADPH-dependant oxidative decarboxylation of MAL to PYR (catalyzed by meaB),
since a Amdh stain gave the same contribution of tritium from the C-4 of glucose
as the wild type.

* The fact that the 4 hydrogen of glucose (NADH via the 3PG dehydrogenase) made
a minor but measurable contribution to the biosynthetic reactions could be
explained by the action of an energy-linked transhydrogenase since this
contribution was abolished in the AuncB mutant strain.

Most of the drawbacks of the aforementioned study were pointed out from the

authors. Namely: i) the possible isotope discrimination against the tritium atom26 should

result in underestimation of the PP pathway; a factor of two is suggested as a correction

(i.e., the PP pathway contributes to 20-40% of the total NADPH pool), ii) unfortunately

strains missing a transhydrogenase (i.e., ApntAB) or the NAPDH-dependant malic

enzyme were not available, at that moment, for the construction of all possible mutants,

iii) the indirect measurements based on a AuncB mutant were of limited value due to

marginal growth characteristics of this particular strain. Nonetheless, the message from

this work is clear: the PP pathway should not account for more than 50 % of the NADPH

pool and it is not an obligatory route for NADPH formation in E. coli cells especially

under anaerobic conditions or in Azwf mutant.

Regulation of Pentose Phosphate Pathway

The results of isotopic experiments under various physiological conditions

suggested that the PP pathway is regulated.10'27 Stronger evidence for the regulation of

the first step of the shunt came from Orthner and Pizer's study.15 They calculated, using

a mutant strain with 10-fold higher amount of Zwf compared to wild type, that the shunt

was not used much faster in the mutant and that both Zwf and Gnd are not operating at

their maximal velocity in the cell. Additionally, according to their data, glucose utilization

by the shunt is not limited by the NADP+ or the NADP+/NADPH ratio since the

concentration of NADPH in the mutant is not sufficient for the 90% inhibition required to

justify the only 40% higher flux. A survey by Silva and Fraenkel17 for effectors that might









control in vivo the Gnd reaction in E. coil was unsuccessful. In an elegant study,28 Wolf

et al. showed the growth-rate-dependant alteration of zwf and gnd levels in E. coil. They

varied the specific growth rate, while using the same carbon source, by feeding the cells

with different ratios of glucose/ a-methylglucose (a competing with glucose for transport,

but non-metabolizable, analog) and observed that the levels of the pertinent enzymes

are proportional to growth rate. Interestingly, the regulatory site for the translational

control of gnd lies within the protein-coding region of the regulated gene, as proved later

by Baker and Wolf.29-31

The Present: The Interplay of NADPH with the Metabolic Network

We may describe the first three decades of the NADPH-regeneration research as

the reductionistt' era. At this point, there are well-documented speculations for all the

possible sources of NADPH in E. coil metabolism. However, crucial questions for the

physiology of the E. coil cell are still unanswered. Namely, are all the NADPH sources

active under any physiological conditions? If so, to which extent does each one

contribute to the overall NADPH pool and from which factors, if any, is the contribution

regulated and how is it adjusted? The reconstruction of all the pertinent biochemical

pathways with the help of new tools and methods (Figure 1-5) may give an answer to

this problem.68 The results from these efforts are summarized in Table 1-2 with an

emphasis on PP pathway flux, and they will be discussed in detail below.















































Figure 1-5. Tools and methods used in the last 25 years for elucidation of metabolic
pathways.32,33

Theoretical Methods

There is no 'pure' theoretical method for the calculation of a biochemical pathway

flux (i.e., the rate at which input metabolites are processed to form output metabolites).68

In this context theoretical methods are considered all the approaches that are not using

isotope tracers. Theoretical methods are coming in different 'flavors'68 69-70 but the

keystone in all of them is the principle of mass conservation.









Table 1-2. Estimated fluxes though the PP pathway.
Year (Ref.) E. coli Strain Methods and Tools Physiological PPPa (%)
Conditions
1980 (34) K12 Glucose-1-13C, 1H Resting cells 22
correlation NMR Cells grown
anaerobically in
M9 medium
1992 (35, K12 FBAb Aerobic growing 90
36) cells without ATP
maintenance
requirements
Aerobic growing 20-40
cells with ATP
maintenance
requirements
1995 (37) W3110 90% unlabeled glucose Aerated culture 20-30
plus 10% [U-13C]glucose, (in a fermenter)
2D [13C,1H]-COSY NMR of overexpressing
the purified and hydrolyzed P22 c2 repressor
P22 c2 repressor peptide protein
Air contact (in a
conical flask)
Anaerobic (in a
conical flask
after purged the
medium with
argon)
1996 (38, MG1655 Enzymatic analysis of end- Fed-batch Higher PP
39) products, FBA cultivation of partition
exponential compared
grown VHb- to wild
expressing cells type
1996 (40) W3110 90% unlabeled glucose Aerobic growing 15
plus 10%[U-13C]glucose, cells in minimal
2D [13C,1H]-COSY NMR of medium
the purified and hydrolyzed overexpressing
P22 c2 repressor P22 c2 repressor
protein
1997 (41) JM101 [-14C]glucose, Exponential- 20-30%
JM101 radiorespirometric method phase growing Mainly
ApykAApykF cells in M9 through
medium the PP
pathway
1998 (42) not reported [2-14C]glucose Slow growth 2
Log-growth 9
phase
Stationary phase 16
a PPP: Pentose Phosphate Pathway. bFBA: Flux Balance Analysis.









Table 1-2. Continued.
Year (Ref.) E. coli Methods and Tools Physiological PPP(%)
Strain Conditions
1999 (43) W3110 90% and 92% unlabeled Aerobic growth 46
glucose plus 8 and 12% without recycling
[U-13C]glucose, 2D Anaerobic growth 74
[13C,1H]-COSY NMR and without recycling
GC/MS analysis of With recycling 20-30
biomass samples
1999 (44) MG1655 90% unlabeled glucose Fed-batch 32
plus 10% [U-13C]glucose, cultivation of cells
2D [13C,1H]-COSY NMR of carrying a high
hydrolyzed cell proteins, copy number
enzymatically analysis of plasmid under
excreted metabolites microaerobic
conditions
1999 (45) B (ATCC 85-90% unlabeled glucose Mid-exponential 29
11303) plus 15-10%[U-13C] Late exponential 25
glucose, 2D [13C,1H]-COSY
NMR of hydrolyzed cell Stationary 24
proteins, metabolic flux
ratio (METAFoR) analysis
2000 (46) MG1655 FBA was used with Simulated aerobic 33
objective of maximizing environment with
growth and examination glucose as carbon
the change in the source
metabolic capabilities
caused by gene deletions
2000 (47, K-12 FBA with objective to Growing cells in 2-25%
48) calculate the sensitivity of glucose minimal
optimal cellular growth to medium
altered flux levels of PP
pathway genes
2000 (49) MC 4100 Glycerol, acetate, glucose, Exponential growth the PP
DNA microarray phase in a minimal genes were
technology for detection of medium not
differential transcription regulated
profiles of a total of 111 significantly
genes involved in central with
carbon metabolism exception of
gnd down-
regulated in
glycerol and
acetate
2001 (50) MG1655 0.45% (w/v) unlabeled Mid-exponential 5-10
W3110 glucose and 0.05% [U- growth phase in M9 100
Apgi 13C]glucose2D [13C,1H]- medium
COSY NMR of hydrolyzed
cell proteins, metabolic flux
ratio (METAFoR) analysis









Table 1-2. Continued.
Year (Ref.) E. coli Methods and Tools Physiological PPP (%)
Strain Conditions
2002 (51) JM101 10% [U-13C]glucose, 2D- Glucose- 21-44 C-
NMR, METAFoR analysis limited limited), 1 (N-
chemostat or limited)
JM101 ammonia- 1-28 (C-
(ApykAF) limited limited)
chemostat 31 (N-limited)
2002 (52, MG1655 Glucose used as a carbon Growing cells Flux
53) source, FBA and under different decreases as
phenotype phase planes oxygen uptake oxygen uptake
(PhPPs) analysis rates rate decreases
2002 (54) JM101 [1-4C]glucose, [1- Growing cells 23
13C]glucose and [U- during
JM101 13C]glucose, a logarithmic
S combination of of genetic, phase in M9
PTS- Glc
biochemical and NMR medium
techniques
2003 (55) MG1655, 20% or 100% [U- Mid- 4 (anaerobic)-
W3110, 13C]glucose, GC-MS exponential 14-20 (fully
JM101 growth-phase, aerobic)
batch cultures
2003 (56) K12 Glucose, enzyme activity Aerobic versus Zwf and gnd
assays and proteome microanaerobic down-
analysis by 2-dimensional growth regulated in
electrophoresis (2DE) micro-
anaerobic
growth
2003 (57, JM109 Semi-quantitative RT-PCR, Batch overproduction
58) Apgi intracellular NADPH cultivation in of NADPH,
measurements, rich LB high flux
extracellular metabolites medium through the PP
analysis
2003 (59) AB1159 Nicotinamide nucleotide Growing cells The activity of
enzyme levels, and cellular in presence of zwf was
ratios of NADPH/ NADH H202 induced by 2.9
measured after exposure fold upon H202
of cells to H202 challenge
2003 (60) W3110 90% unlabeled glucose Aerobic 9 2 (C limited)
plus 10% [U-13C]glucose, chemostat 4 2 (N limited)
W3110 extracellular metabolites cultivation in 47 9
Apgi analysis, enzyme activity glucose or (C limited)
assays, [13C, 1H]-COSY ammonia- 37 9
NMR analysis of amino limited minimal (N limited)
W3110 acids, glycerol, glucose in medium 0 2 (C limited)
Azwf the hydrolysates __0 2 (N limited)









Table 1-2. Continued.
Year (Ref.) E. coli Methods and Tools Physiological PPP(%)
Strain Conditions
2003 (61) K12 7.5 % [U-13C]glucose and Aerobic 43
7.5 % [1-13C]glucose in chemostat
presence of naturally cultivation in
labeled glucose, glucose minimal
extracellular metabolites medium at
analysis, GC-MS analysis dilution rate (D)
of biomass hydrolysates 0.11 h-'
At dilution rate 28
(D) 0.22 h -1
2003 (62, BW25113 10 % [U-13C]glucose and Aerobic 20
63) BW25113 10 % [-13C]glucose in chemostat 10
Agnd presence of unlabeled cultivation in
glucose, enzyme activity glucose minimal
assays, GC-MS and NMR medium at
analysis of biomass dilution rate (D)
hydrolysates 0.2 h-'
2004 (64) K12 10 % [U-13C]glucose and Aerobic 34
10 % [1-13C]glucose in chemostat
presence of unlabeled cultivation in
S glucose, enzyme activity glucose minimal
SApykF assays, GC-MS and NMR medium at
analysis of biomass dilution rate (D)
hydrolysates 0.2 h-'
2004 (65) BW25113 10 % [U-13C]glucose and Aerobic 20
BW25113 10 % [1-13C]glucose in chemostat 0
Azwf presence of unlabeled cultivation in
glucose, enzyme activity glucose minimal
assays, GC-MS and NMR medium at
analysis of biomass dilution rate (D)
hydrolysates
2004 (66) MG1655 20% or 100% [U- Aerobic batch 17-23
13C]glucose, GC-MS cultures in M9
medium
2005 (67) TG1 FBA, extracellular Cells growing 39 %
(a K-12 metabolites analysis, aerobically in mol mol-
strain) determination of ATP, glucose limited
ADP, AMP continuous
culture at D=
0.044 h-1
Cells growing 73 %
anaerobically in mol mol
glucose limited
continuous
culture at
D= 0.44 h-1









Similar to Kirchhoff's current law, (Figure 1-6) the fundamental equation for the flux

balance analysis (FBA) states that a metabolite cannot be accumulated inside the

metabolic network, therefore, over long period of time, the formation fluxes of a

metabolite must be balanced by the degradation fluxes.72 For the FBA to be applicable:78

* All possible biochemical pathways for the particular organism must be known.
* The stoichiometry of the biochemical reactions constituting the metabolic network
should be well defined.
* The elemental composition of the biomass should be estimated.
* Physiological data such as nutrients uptake rates, metabolites secretion rates and
biomass production rates should be calculated.
* The assumption of steady-state operation of the bioreaction network should be
applied.

For a network composed of n metabolites and m fluxes, this yields a system of n

differential equations with m unknown fluxes which are time-independent when the

network operates under steady-state conditions. Although, a number of constraints are

applied to the specific system, based on experimental data, the number of unknown

fluxes usually exceeds the number of metabolites and the solution is underdetermined.

Thus, a multidimensional solution space satisfies the constraints. To reach a single flux

distribution further biochemical experiments (e.g. identification of active pathways under

certain physiological conditions) must be performed and certain objectives need to be

set (e.g. maximal cell growth or maximal production of a cofactor or metabolite). Then,

an appropriate optimization algorithm may produce a unique solution.

In silico estimations for NADPH-related fluxes in E. coli

In one of the earliest applications of the FBA approach in E. coli, Palsson and

coworkers35,36 calculated that the maximum yield of NADPH produced, per mole of

glucose consumed, is 11 moles. When the optimal growth for E. coli, in the presence of

glucose, was set as objective, the PP pathway was found to be used almost exclusively

from the bacterium.











IS

IA--


KIRCHHOFF'S CURRENT LAW

-IA + Is + Ic + lo = 0


electric circuit


-metabolic network



metabolic network


external fluxes:
bI, b2, bs
S\ (measurable)

z_ \vi = v(k,. k2, ks...)
(unknown)


S. OR OR -- Sv
i .- di


Flux Balance Analysis Law


d=Sv O=S.v
dt


;:3 L.


steady state
assumption


-1 0 0 0 o
0


Problem: the number of reactions exceeds the number of metabolites,
S matrix has more columns than rowsl
Solution: find the physiologically feasible flux distributions.

HOW?


apply constrains
vi O,Vi
source: bri 0
sink: b2, b > 0


i set objectives
define objective factor (Z)
as a linear function of v
z =., cv
Cj 33


choose an objective factor
(i.e. cellular growth)

) optimize objectives (i.e. maximize cellular growth) under
certain constrains: use the simplex algorithm to solve the
linear programming (LP) problem.


Question: Is the optimal solution
biologically meaningful?

Answer: Do the experiments


3 metabolic flux v3




optimal
S solm ton
raslboile flux v2
xCvUXY2


x. metabolic flu


Figure 1-6. Basics of Flux Balance Analysis (FBA) approach (Adapted from references
72, 73, and 74)









Realizing based on experimental data that this value is an overestimation they applied

energy constraints to their model and showed that the PP pathway flux is very sensitive

to the ATP maintenance requirements (e.g. maintenance of cellular osmolarity,

transmembrane gradient, etc.).

In an effort to evaluate the effects of gene knockouts in the E. coil metabolic

network, Edwards and Palsson46 applied in silicon gene deletions and calculated the

resultant effects on growth flux. In line with the experimental data, they reported that the

in silicon zwf mutant had a growth rate of 99% of the 'wild type' value with a concomitant

increase of the TCA flux. In addition, the in silicon pgi deletion strain had a decreased

normalized growth yield under oxygen-limited conditions.

The robustness (i.e. the ability of an organism to operate under a wide range of

conditions)75 of E. coli was also examined. When the optimal flux for an essential for the

PP pathway gene, tkt, was decreased to 15% of that of the in silico wild type, the optimal

growth flux did not change, but a utilization of a transhydrogenase that converted NADH

to NADPH was observed in addition to an increased TCA cycle flux. Also, under

oxygen-limited conditions, the oxidative part of PP pathway was not functional, the

transhydrogenase reaction was again activated and the TCA cycle did not operate

cyclically.52'53 In a similar manner, but also measuring experimentally the energetic state

of the cells, Weber et al. demonstrated that the flux towards the PP pathway is increased

at the cost of TCA cycle activity when growth rate increases in glucose-limited

chemostat cultures.67

Computational methods: limitations

Basically, the power of the computational approaches is based on the

experimenters' failure to estimate the in vivo kinetic parameters in a metabolic network.

Under steady-state conditions the time-factor is cancelled out and the differential

equation system is transformed to a set of linear equations that can be solved by matrix









theory. In principle, this assumption holds true due to high turnover of the pools of

metabolites, especially in continuous cultures and in exponential phase of a batch

culture.76 However, several drawbacks are evident:77'78

* Metabolic flux analysis (MFA) fails to predict intracellular in parallel or cyclic
metabolic pathways fluxes which are not coupled to measurable fluxes.
* MFA measures net fluxes. Therefore, it fails in bidirectional reaction steps such as
the transaldolase and transketolase reactions in PP pathway.79
* Assumptions made on the global currency metabolites ATP, NADH and NADPH
are controversial. For example, in microorganisms bearing a transhydrogenase
which can equilibrate both NADH and NADPH pools, the typical use of separate
balances is an oversimplification. In addition, ATP-futile cycles cannot be
estimated precisely.80
* The accuracy of MFA on calculated fluxes is directly related to the experimental
error of the measured extracellular rates.
* The postulation that a metabolism of a particular organism is designed by nature to
follow a certain objective may or may not be true. Even thought the in silicon
predictions have been consistent with experimental data in an E. coil wild type
strain81 they are questionable in a perturbed system, like the knock-out strains.82

Experimental Methods

Nuclear Magnetic Resonance (NMR) spectroscopy and mass spectroscopy (MS)

are the dominant tools used during the last two decades for the elucidation of metabolic

networks. Stable isotopes such as 180, 2H, 13C had been used as tracers in some early

studies in combination with MS, but radioactive tracers and the liquid scintillation

counting technique was more approachable in the post Word War II era, for metabolic

physiologists due to the high sensitivity of this method.83 The last two decades, the

advent of high field NMR instruments and the coupling of MS with liquid (LC-MS) or gas

chromatography (GC-MS) tipped the balance in favor of stable isotope tracers since

problems due to detection limits are eliminated. Considering the fact that carbon is the

central atom in biomass it is not a surprise that 13C is the predominant stable isotope for

the labeling measurements in the latest studies especially when the 12C and 13C differ

not only in mass but also in nuclear spin. Hence, their different physical properties can

be detected by either of the abovementioned analytical instruments.









Also, complementary information could be drawn by these two methods since both

have the same starting point: Cells are growing in minimal medium containing a carbon

labeling source, uniformly and/or 1-13C glucose is used almost exclusively (see Table 1-

2). As the labeled substrate is traveling through the metabolic network, stable isotope

labeled carbons are exchanged, diluted, recycled and finally accumulated into different

metabolites in a very specific manner, which is dictated from the pathway that they

followed.84 When a steady-state condition is established, biomass hydrolysates such as

proteinogenic amino acids are subjected to NMR and/or GC analysis and the isotopic

enrichment is measured. The isotopomer patterns (i.e. isomers differing only in isotope

distribution) of the amino acids reflect the patterns of their related precursors which are

usually key intermediate metabolites of the central carbon metabolism. It is not easy to

detect these intermediates due to their low abundance in the cell, amino acids detection

is relatively easy since approximately 55 % of biomass is proteins.85

Finally, what holds true for the theoretical methods is also true for the experimental

studies: The complexity of the metabolic networks makes the mathematical and

computational models absolutely necessary tools for the experimentalists who want to

fully exploit the meaning of their experimental data. Thus, in the last step of quantitative

analysis, an isotopomer model of metabolism is integrated with an iterative fitting

procedure of the measured data and the intracellular metabolic fluxes are derived

(Figure 1-7).

NMR analysis for 13C labeled amino acids

Two types of information are usually derived from NMR 13C isotopomer analysis

data.86 The first involves fractional enrichment measurements in which glucose labeled

at a specific position (e.g. [1-13C] glucose) is used and the degree of 13C enrichment (i.e.,

the ratio of molecules bearing a 13C spin at a given position over the total number of

molecules) is calculated.87 In the second and the most commonly employed method,










uniformly labeled glucose, 'diluted' with unlabeled glucose, is administrated into the cells

and the conserved 13C13C connectivities in the metabolic network are traced (Figure 1-

8). This technique makes use of the fact that upon 90 % 12C dilution of 13C with random

distribution of carbon atoms, the probability that any two adjacent carbon atoms are

labeled in an individual molecule is approximately equal to the natural abundance of 13C

(1.1 %). However, if the two adjacent carbon atoms in the final product derive from the

same source molecule, the probability that both positions are labeled is about 10%. The

extent to which a certain carbon atom has neighboring carbons that originate from the

same source molecule of glucose is calculated from the quantitative analysis of the 13C

fine structures. 88




Biochemistry

] Predict:
SLabel enrichments
Formulate Isotopomer molecular
metabolic- Solve for metabolite weight distributions
isotopomer Isotopomer pools
balances Fine structure of NMR


Initial flux
guesses


Readjustfluxtes
I ---- /> [ estimates


Compare with
S experimental data
S Is the agreement
Satisfactory?






Flux estimates


Figure 1-7. Synergy of experimental and computational methods towards the calculation
of metabolic fluxes (Adapted from reference 68).















































Figure 1-8. Basics of NMR analysis of biosynthetically directed fractional 13C labeling
amino acids. (Adapted from references 87, 88)

2D heteronuclear single-quantum coherence [13C, 1H]-correlation NMR

spectroscopy (2D [13C, 1H]-COSY) offers the highest sensitivity in deciphering the 13C

13C spin-spin scalar coupling fine structures. In addition, the two dimensional dispersion

of 13C resonances enables the analysis of a mixture of amino acids without further

purification. Subsequently, the relative intensities of the superimposed multiplets are









determined and the metabolic flux ratios are estimated using a number of probabilistic

equations and keeping in mind the biosynthetic reaction network of E. coli. 89

MS analysis for 13C-labeled amino acids

Compared to NMR instrumentation, MS offers higher sensitivity in a shorter time

frame. Hence, a greater number of smaller volume samples can be analyzed,

minimizing the otherwise high cost of labeled glucose and improving the accuracy of the

calculated fluxes. The coupling of MS with gas chromatography (GC) enables the

analysis of mixture of labeled amino acids eliminating laborious purification steps.

However, in contrast to NMR, a derivatization step of the protic sites of the amino acids

by silylation is required to increase the volatility of these polar compounds for GC

separation before the MS analysis.91 Also, the MS data are difficult to interpret

intuitively: for a molecule with three carbons, 23 positional isotopomers are expected, but

a MS peak at M+2, for example, represents all isotopomers with two 13C nuclei,

irrespective of their location in the carbon backbone (Figure 1-9). Consequently, in

addition to analysis of mass isotopomer distribution (MID) for the molecular ion,

corresponding fragments should be available for further interpretation. Taking into

account that the production rate of each isotopomer depends on the probability of its

formation, which is equal to the probability of formation of its precursor metabolite which,

in turn, depends on real metabolic fluxes, correlation between isotopomer and MIDs is

accomplished by using matrix calculus.90

Estimated fluxes from 13C labeling experiments

In one of the earliest attempts to elucidate metabolic fluxes by NMR, Ogino et al.

used 1H correlation NMR to follow glucose catabolism in anaerobic E. coli cells under

various conditions.34 First the cells were growing in M9 minimal medium in the presence

of glucose. Cells harvested at exponential or early stationary phase, were resuspended

in the same buffer in a 5-mm NMR tube and incubated in a NMR spectrometer at 30 oC,










using glucose -1-13C as the sole carbon source. Spectral assignments for succinate,

pyruvate, acetate, lactate and ethanol were made based on chemical shifts, spin

coupling patterns and 13C-1H coupling constants. The 13C label is lost as CO2 via the PP

pathway, whereas it is incorporated into the methyl carbon of pyruvate via the EMP

pathway. Hence, it was estimated that in the case of resting cells or growing cells in M9

medium 22% of the glucose utilized by E. coli follows the PP pathway.


29 positional isotopomers


m,o MS analysis
a molecule
with 3 carbon
atoms A
b -




1 2 3
mass+2 Mass +


too. m"as+3


Figure 1-9. MS analysis: mass isotopomer distribution (MID) for the molecular ion.
(Adapted from reference 90)

In 1995 Thomas Szyperski demonstrated the power of biosynthetically directed

fractional 13C-labeling of proteinogenic amino acids in estimation of intracellular fluxes in

E. coli.37 A K-12 derivative E. coli strain producing the P22 c2 repressor in high levels

was used in this study. The overexpressed peptide was purified and subjected to

hydrolysis. The labeled amino acids were analyzed as described before. Despite the

examined physiological conditions from fully aerobic to strictly anaerobic, the flux

through the PP pathway was set at 20-30 %. Interestingly, Schmidt et al. using the

same set of experimental data, but applying a different mathematical model for the

quantification of metabolic fluxes, arrived at twice higher numbers: 46 % under aerobic









and 74 % under anaerobic conditions.43 The authors explained the deviation of their

calculations by mentioning that the reversibility of reactions involved in PP pathway (e.g.

pgi, tkt, tal) was not considered from the outset. Also, in their estimations the flux from

ICT to AKG (the third source of NADPH) in TCA cycle was found to be 25 % under

aerobic conditions and 2.2 % under anaerobic conditions.

Trying to stimulate a situation met frequently in the biotechnological industry, Fiaux

et al. performed a 13C NMR flux ratio analysis on a wild type E. coli MG1655 strain

carrying a high copy number plasmid under microanaerobic conditions ([02] < 0.02

mmol/L).39 The PP phosphate flux was 32%, whereas the flux from AKG to OAA was

zero, showing that the TCA cycle is interrupted under anaerobiosis. In 1999 the term

METAFoR (metabolic flux ratio) was introduced by Sauer et al. to describe the technique

of imprinting the central carbon network history into cell proteins.45 In this study the

interstrain differences among four strains (JM101, a K-12 strain, PB25, a pyruvate

kinase-deficient JM101 strain, ATCC 11303, a wild type B strain, and K020, an ethanol

producing ATCC 11303 strain) were examined under different growth conditions. The

major observations were: i) the impact of enzyme overexpression and gene disruption

on central metabolism of E. coli was negligible, ii) of all central carbon fluxes, those in

the TCA cycle were the most sensitive to environmental conditions and iii) surprisingly,

the METAFoR patterns among E. coli strains with different genetic backgrounds were

very similar in exponentially growing cells under aerobic conditions.

In 2000, Dauner and Sauer introduced the GC-MS analysis of proteinogenic amino

acids as a rapid method for the calculation of the isotopomer distributions.91 Three years

later, Zhao and Shimizu combined the above technique with the theoretical approach of

metabolic flux analysis, in order to estimate the net flux distribution in glucose

metabolism in E.coli K-12 strain.61 The flux via the PP pathway was dependant on

growth rate with higher values at higher dilution rates. This observation suggested that









under increased growth rate, the flux through icd could not fulfill the NADPH demand for

growth and as a consequence higher flux through the PP pathway is required.

The old and popular method of creating mutants and reexamining the differences

in fluxes in the genetically perturbed strains compared to the parent strain provide a new

insight into metabolism of E. coli cell in recent years. Precise gene deletion in E.coli is a

routine technique nowadays that the whole E. coli genome is annotated. Using

METAFoR analysis by NMR, Canonaco et al. reinvestigated the impact of pgi deletion

on glucose catabolism and confirmed the flux rerouting via the PP pathway.50 The

observation of reduced specific growth and glucose-uptake rates due to Pgi-knockout

was explored further by overexpressing the soluble transhydrogenase SthA in the

mutant and the parent strain. While no difference was found in the wild type strain, the

specific growth of the Pgi mutant increased by about 25%, but it could not be restored to

100% even when a 10-fold higher overexpression of SthA was attempted. This

suggested that the kinetic limitation of glucose catabolism in the Apgi strain is due to

both the inability of the PP pathway to support higher fluxes and the limited capability of

E.coli to reoxidize excess amounts of NADPH. Also, Fischer and Sauer observed a 30

% contribution of ED pathway for the pgi mutant when they applied a GC-MS analysis to

labeled proteinogenic amino acids from E. coli.55 The phenomenon could be explained

by recalling that the ED pathway produces twice fewer moles of NADPH per mole of

glucose compared to the PP pathway and thus it may serve as an alternative route for

the reduction of the excess amounts of NADPH in pgi-knockout.

Shimizu's group combined the 13C-labeling data from NMR analysis with biomass

composition and extracellular flux data in order to quantify intracellular fluxes in pgi and

zwf knockouts.60 Interestingly, in an effort to minimize the overproduction of NADPH, the

Pgi mutant bypassed the isocitrate dehydrogenase reaction by rerouting carbon flow

through the glyoxylate shunt which is otherwise inactive in the wild type. Disruption of









G6P dehydrogenase leads to activation of transhydrogenase activity towards the

production of NADPH, especially in a nitrogen limited chemostat. Elevated maeB

activity was also apparent in this mutant. The overall picture of regeneration rates of

NADPH as they were estimated in this study is depicted in Figure 1-10. It is worth noting

that the rates attributed to MaeB are based on the assumption that both NAD+ and

NADP+ dependant malic enzymes (MaeA and MaeB) are equally active.



APgW3110 pg, wf
C-limied C-imited I Cme




a icd
W3110 APi Azwf
N-limited N-imiled N-imited1 maeB

transhydrogenases


Figure 1-10. Specific rates of NADPH production in glucose (C)- and ammonia (N)-
limited chemostat cultures of E. coli W3110. The numbers in doughnuts are in
mmol g-1 h-1 units. (Adapted from reference 60)

A similar pattern with the one observed for the zwf mutant65 was also found when

the 6-phosphogluconate dehydrogenase (gnd) deletion was studied by Jiao et al.62

Activation of the ED pathway, induction of malic enzyme and decreased flux through the

PP pathway were the major features of this strain.

The abovementioned work by Hua et al.60 introduced experimentally, for the first

time since 1977,16 the transhydrogenases as players in NADPH production, at least in

Azwf mutant strain. With Enterobacteriaceae being the only known microbes containing

both the soluble, SthA and the membrane bound, PntAB transhydrogenase92 the

question of which contributes to NADPH pool remained open. Speculations for the

physiological role of PntAB as NAPDH producer had existed since 1980 when Hanson

and Rose carried out phenotypical studies on a zwfpnt double mutant and showed that









it grew slower than the zwf mutant, as it would be expected when pntAB produces

NADPH in absence of the PP pathway.93 The experimental answer for the physiological

role of PntAB in the wild type E. coli strain came from Sauer's group in 2004.66 First, in

addition to the Apgi strain, a new set of mutant strains such as ApntAB, AudhA, A(udhA

pntAB) and A(zwfedd eda) were constructed. Second, the consumption rate of NADPH

was estimated based on the growth rate of each mutant strain. Third, the specific rate of

NADPH formation from PP phosphate pathway, from the icd and from the maeB was

calculated based on METAFoR analysis of 13C-labeled amino acids by GC-MS and

metabolic flux analysis. In the Apgi strain, the specific rate of NADPH formation was

two-fold higher than the rate of NADPH consumption. In ApntAB, A(udhA pntAB) and

A(zwfedd eda) strains, the two rates were almost equal, but in AudhA as well as in the

wild-type strain, the rate of NADPH depletion surpassed by 20-40% the total rate of

NADPH production from the three aforementioned source. These measurements led the

authors to conclude that about 40% of the NADPH required to compensate for the

imbalance between NADPH depletion and formation in the wild type strain must come

from a transhydrogenase activity, especially from the PntAB transhydrogenase (Figure

1-11). Another interesting outcome from this study was the observation that the A(pgi

udhA) strain had lethal phenotype on glucose. Growth was recovered only upon plasmid

expression of udhA. This result showed that the physiological role of SthA is to reoxidize

the excess amounts of NADPH, restoring the imbalance of NADPH metabolism in the

Apgi strain.

Besides the enzymes within the glucose-6-phosphate (G6P) node, enzymes

related to phosphoenolpyruvate (PEP) node have also been examined for their ability to

control the NADPH metabolism. One modification was the disruption of the pykA and

pykF (Figure 1-12), which encode the two pyruvate kinases (PKs) in E. coli.

Radiorespirometric assays by Ponce et al.41 with resting cells of the mutant strain,









lacking both PKs, metabolized glucose mainly through the PP pathway. A

reexamination of the same mutant by Emmerling et al. with METAFoR analysis by NMR

confirmed the observation only in ammonia-limited chemostat cultures.51 In glucose-

limited cultures, similar or lower fluxes compared to parent strain were calculated for the

PP pathway with concomitant higher fluxes for the TCA cycle, especially in the

chemostat with the lower dilution rate. A single mutant ApykF strain examined by Al

Zaid Siddiquee et al.64 showed the same trend as the double mutant under non-growing

conditions, i.e. elevated fluxes thought the PP pathway. In addition, activation of the

malic enzyme with a flux about 10% towards the formation of PYR from MAL was

calculated.




20-25








35-45%
35-45


a zwf & gnd

I icd
pntAB


Figure 1-11. The percent of NADPH produced by PP pathway, TCA cycle and PntAB
transhydrogenase activity in E. coli during standard aerobic batch growth on
glucose. (Adapted from reference 66).







































Figure 1-12. Glucose transport systems in E. coli. The phosphoenolpyruvate:
carbohydrate phosphotranferase system (PTS) and the galactose permaese
(GalP) system. (Adapted from reference 94)

Another modification around the PEP node was the disruption the of

phosphotranferase transport system (PTS) which consumes at least 50% of PEP.94

Starting from a strain lacking PTS (PTS-), which has a limited ability to grow on glucose,

a method based on continuous culture was applied to select for mutants (PTS-GLC')

with growth rates similar to wild type strain JM101 (a functional galP gene is required for

this phenotype, Figure 1-12). METAFoR analysis of 13C-labeled amino acids by NMR

revealed reduced carbon flux through the PP pathway and a 4-10% flux through the

malic enzymes for the PTS-GLC mutant strain.54 Interestingly, in the PST-GLC- mutant,









in which glucose transport capacity is 10 times lower than the wild type JM101 strain,

twice higher flux via the PP pathway was observed compared to parent strain.

Other experimental methods: metabolic regulation analysis of NADPH related
metabolic pathways at gene and protein expression level.

Two dimensional electrophoresis (2DE) is a powerful technique for separation of

complex protein mixtures and based on their different MWs and pls. Ultimately,

regulation of metabolic fluxes depends on enzyme activity which depends, to some

extent, on protein expression levels. Thus, the different protein expression levels under

different physiological conditions can be estimated and correlated to regulation of

metabolic pathways95 by separating a protein mixture in the cell lysate by 2DE and

quantifying each protein spot after electrophoresis. Utilizing this approach in

combination with in vitro measurements of enzymes activities in E. coli, Peng and

Shimizu found that the two proteins involved in the PP pathway and related to NADPH

production, zwf and gnd, are down-regulated during microanaerobic growth in glucose

medium.56 This result was not a surprise because both enzymes were known to be

subject of growth rate regulation, in fact the cells growth was quite slow under

microanaerobic conditions. For isocitrate dehydrogenase, a source of NADPH in TCA

cycle, the correlation of expression by activity was not linear since it is known that ICDH

activity is regulated by reversible inactivation via phosphorylation. 96

Another technique for monitoring metabolic regulation at gene expression level is

the reverse transcription polymerase reaction (RT-PCR) analysis. It is currently the

preferred method for quantifying mRNA due to its higher sensitivity compared to other

RNA analysis techniques such as Northern blot. In addition, it is relatively easy to run:

isolated RNA is used as a template in order for a retroviral reverse transcriptase to

create a complementary DNA (a cDNA) transcript. This product is subjected to a

common PCR, where in the exponential phase, the signal is proportional to the amount









of input RNA template or number of cycles. The measurements of relative changes in

mRNA levels are reproducible as long as the PCR volumes are the same from tube to

tube and the data are collected before the reaction reached a plateau. Finally, a

standard curve is required to correlate the signal with the specific amount of mRNA.

Kabir and Shimizu applied this approach, also known as semi-quantitative RT-PCR, to

investigate patterns in a pgi knock-out E. coli strain expressing genes for poly(3-

hydoxybutyrate) (PHB) production (Figure 1-13).5758

Glycolysis


SNADPH
phbA I phbB
Acetyl-CoA Acetoacetyl-CoA 3-Hydrobutyryl-CoA

n NADP+ 4 phbC



TCA cycle H
-C-CH2-C
I II
CH3 O



Figure 1-13. Poly(3-hydroxybutyrate) PHB synthesis: the acetoacetyl-CoA synthase
requires NADPH.

Sauer's group had shown that the deleterious effect on cell growth caused by the

NADPH imbalance in the pgi E. coli mutant can be partially compensated by expressing

an enzyme consuming NADPH such as SthA. A similar consequence may have the

expression of the acetoacetyl-CoA synthase in Shimizu's studies, which also requires

NADPH. Indeed, expression of phbB in E. coli (Apgi) increased the true cell mass by

about 19%. Further measurements of NADPH concentrations confirmed the higher

levels of NADPH in the mutant strain without phbB compared to the PHB producer









strain. The transcript level analysis of 87 central metabolic pathway genes revealed

upregulation of most of the genes in PP pathway except gnd. The most interesting

observation from this proteome analysis study was the absence of ptsG, which encodes

for a protein that is a component of PTS system, in 2DE gel for both pgi mutants even

though the RT-PCR revealed the presence of ptsG transcript in both cases. This result

led the authors to propose that the mutation of pgi affects ptsG expression after

transcription initiation and an RNAaseE-mediated degradation of ptsG transcript97 might

be another reason for lower glucose consumption rates in this mutant.

Experimental methods: limitations

Undoubtedly, the experimental techniques discussed in the second part of this

chapter offer a more holistic picture of the NADPH regeneration sources in E. coli;

however, possible limitations and pitfalls of them must be encountered. Specifically for

the 13C metabolic flux analysis:

* METAFoR analysis fails in the case that two alternative pathways operate
according to the same carbon-carbon bond rearrangements.

* METAFoR analysis of 13C NMR and/or GC-MS data provides direct evidence for a
particular flux or reaction in a metabolic network, but only ratios between fluxes
around a metabolic node can be calculated. In order for the absolute intracellular
fluxes to be estimated, 13C isotopic measurements must be applied, in parallel,
with the theoretical approach of FBA, in which the 13C data are acting as additional
constraints.51 As a consequence, prerequisites stated above for the FBA to be
valid must be followed here as well.

* The merging of METAFoR analysis and FBA requires 'heavy' mathematical and
computational analysis98 which is usually performed by specially designed
software with little interaction between the program and the experimentalist.99

* Isotope abundances calculated from 13C data should be independent of possible
limitations of the technique used for their derivatization such as 13C spin relaxation,
strong spin-spin coupling effects, 13C isotope effects on carbon chemical shifts,
isotope fractionation of the amino acids during the GC separation and nonlinearity
of the MS detector response at different mass isotopomer ratios or different
sample concentrations.78

* In addition to a known metabolic network, 13C isotope effects on metabolic
reactions must be negligible and metabolite channeling (i.e., when an intermediate









metabolite is transferred from one enzyme to the other without previous
equilibration with the rest of the molecules in the medium) should be absent.100

* The accuracy of 13C data depends on the 13C/12C glucose with higher accuracy at
a ratio of 50:50. The relative high cost of labeled sugars usually prohibits this type
of experiments.

* The METAFoR analysis based on NMR data is quite laborious, time-consuming
and expensive due to low sensitivity of the technique: about 300-400 ml of sample
and 4-5 days are required before the calculation of the first set of data.101 In recent
years, the problem was partially bypassed with the use of the GC-MS, but
interpretation of the MS data is more complicated.

* The 13C metabolic flux analysis is currently restricted to two physiological
conditions: continuous cultures and exponentially growing cells, because isotopic
equilibrium must be established before the first measurement.

* Last but not least this approach is not applicable in biological systems like resting
cells performing biotransformations, due to insufficient de novo synthesis of
biomass.87

2DE and RT-PCR, are still in their infancy. A common problem associated with

2DE is the identification of each spot. Amino acid sequencing and enzyme activities

must be employed in parallel for the 2DE data to be justified. The RT-PCR is, in

principle, a quick and extremely sensitive technique, but many parameters concerning

the amount of PCR ingredients and the duration of PCR cycles must be verified and

controlled carefully in advance. The design and quality of primers is also of paramount

importance since genomic DNA contamination in a RNA preparation may be amplified

too and produced false positives. Both techniques fail to detect post-translational

regulation of enzymes and become extremely costly if we consider all the additional

control experiments that are required for their justification.

Conclusion

The aim of this chapter was to demonstrate the different tools and methods that

researchers used the last six decades in order to answer two key questions regarding

the NADPH metabolism in E. coli: Which are all the possible sources of NADPH in E.

coli? Most probably those five depicted in Figure 1-2. To which extent each source









does contribute to the total NADPH pool? Unfortunately, as our discussion and a quick

look at the last column of Table 1-1 and Table 1-2 revealed a global answer does not

exist. In fact, the answer is another question: under which physiological conditions and

based on which genetic background we are seeking for an answer? Different E. coli

strains, in different environments, are using some or all of the abovementioned NADPH

sources to a different degree in order to accommodate their needs. This statement may

satisfy a physiologist who is trying to understand how a biological system works.

Definitely, it does not fulfill the ambitions of a metabolic engineer:

Metabolic engineering is the improvement of cellular activities by manipulation of
enzymatic, transport and regulatory functions of the cell with the use of
recombinant DNA technology.
[Toward a Science of Metabolic Engineering, James E. Bailey]102

In the next chapter, a couple of examples related to metabolic engineering of

NADPH redox metabolism in E. coli will be introduced. A particular interesting case in

redox biocatalysis, in which the NAPDH regeneration machinery of resting E. coli cells

was used to produce a chiral compound in a gram-scale level will be discussed in detail.















CHAPTER 2
INVESTIGATION OF NAPDH REGENERATION SOURCES IN RESTING
ESCHERICHIA COLI CELLS ENGINEERED TO OVERPRODUCE A NAPDH-
DEPENDENT KETOREDUCTASE

So far, the more valuable products, chiral building blocks for the synthesis of
pharmaceuticals, flavors and agrochemicals, are potentially obtained with NADPH-
dependant enzymes.
-Andreas Liese, Head of Institute of Technical Biocatalysis at TUHH103

Introduction

NAPDH is an extremely 'high-priced' hydride donor for an E. coli cell as well as

for an organic chemist when stoichiometric amounts of it are required (Table 2-1 and

Figure 2-1). Therefore, a reductive reaction would be economically feasible only upon

regeneration of the core molecule at the expense of another cheaper reductive source.

For example, the biosynthesis of one gram CDW (cell dry weight) of E.coli cells requires

more than 15 mmols of NADPH; however, this amount of biomass barely contains

0.0014 mmols of the reduced cofactor.104 In other words, the same nicotinamide

molecule has been recycled more than 10,000 times during this process, usually at the

expense of a carbon source such as glucose.

Accordingly, a plethora of ingenious NAD(P)H regeneration strategies have been

employed by researchers who are exploring the possibility to use a NAD(P)H-dependent

reductase in their asymmetric synthesis. Modifying slightly the original classification by

Chenault and Whitesides,105 we should recognize that among the four general

regeneration strategies (Figure 2-1): chemical, photochemical, electrochemical and

biological, only the latter has unlimited compatibility with biological systems. In between

the enzymatic and microbiological methods, the former has found a great deal of

applications in laboratory as well as in industrial scale, especially when the regeneration









of NADH is utilized.106 In contrast, the NADPH regeneration techniques are lagging

behind.

Table 2-1. Cost of nicotinamide cofactors and their content in E. coli.
Cofactor MW Costa Specific cost Number of Intracellular Cofactor
($/g) ($/mmol of molecules molarity content in E.
delivered [H]) per E. coli (mM)b coli
cellb__ (pmol/g CDVV)
NADPH Na4 833.35 837 697 250,000 0.22 1.370
(-95%)_
NADP+ Na 765.39 211 161 190,000 0.17 1.041
(>95%)
NADH Na2 709.4 70 50 440,000 0.90 2.411
(-95%)___
NAD+ Na 685.45 30 20 1,000,000 0.40 5.479
(-95%)__
a- ,, "'10 bh c


Based on S-igma-Aldrich catalog (20065) AatdTor~re-0
that the mass of one cell is 3.0x1 0-13 g (DW).108


-Assuming


- $700

o
Io
4*-
0
E $160
C_


13

8 $50

$20


..*** IHI.

/ U

ATP .:
4:







) NADH

NAD*


steps of synthesis

Figure 2-1. Explanation of the relative prices among nicotinamide cofactors.109 Step 1:
Isolation of NAD+ from bakers' yeast (1mmol/0.55kg CDW). Step 2:
Reduction or phosphorylation of NAD+ leads to NADH or NADP+ respectively.
Step 3: Reduction of NADP+ produces NADPH. Depending on the pH,
NAPDH is 3 to 10 times less stable than NADH.











STRATEGIES FOR
NAD(P)H
REGENERATION


(Chemical ) Electrochemical


SPhotochemical )
limited stability and selectivity,
tow TTN, complex apparatus,
limited compatibility with
biological -
systems Microbi(


Mechani call or
cherricall, prereaate'l
h I.le cell S


Biolgical )


logical


Enzymatic


- II


I :Ll Itrl te -t


primitive state of development,
unknown stability, I
reduced enentiomeric
purity o product in Intact Whole-cells
some cases, tow volume
productivity, complicated product isolation


En yrn ?-
cou.._pl ed


high cost, low rates and low
volume productivity,
low TTN In one phase
system, complicated
product isolation in some
cases


Figure 2-2. NAD(P)H regeneration strategies and major disadvantages for each
method. (Adapted from reference 106)
The aim of this introduction is not to extensively review all the NAP(D)H

regenerations methods, since this has been done recently.106'110 In contrast, by

selectively analyzing a few examples from the current literature, we will acknowledge the

limitations of the enzymatic strategies in NAPDH regeneration, and we will point out the


importance of the experimental work that follows.









Examples of Enzymatic NADPH Regeneration Methods

A substrate-coupled NADPH regeneration scheme, like the one depicted in

Figure 2-3, is probably the most self-sufficient method in this category. In the presence

of excess isopropanol, a recombinant alcohol dehydrogenase from Lactobacillus brevis

(LBADH) reduced catalytic amounts of NAPD+ and simultaneously acted as an excellent

regio- and enantio-selective reductive catalyst on a number of 3,5-dioxocarboxylates.111

Unfortunately, the method is limited by the enzyme's substrate specificity. Thus, when

the opposite enantiomer was the desired product, the same researchers switched to one

of the most well-established,112 enzyme-coupled, regeneration systems using glucose-6-

phoshate as a sacrificial cosubstrate (Figure 2-4).113 The subsequent hydrolysis of the

corresponding lactone renders the reaction irreversible in this approach.

S OH
NADP OH 0 0

t-butyl (R)-5-hydroxy-3-oxohexanoate

LBADH LBADH


0 0 0 0
NADPH ot-Bu
t-butyl 3,5-dioxohexanoate

buffered medium containing sonicated E.coli cells & NADP*

Figure 2-3. A substrate-coupled enzymatic regeneration method for NADPH.

Due to the relatively high cost of phosphorylated glucose, G6PDH has been

replaced the recent years by glucose dehydrogenase (GDH), which accepts glucose as

reductant (Figure 2-5).114 Unfortunately, the net benefit from this replacement, in terms

of cost, is negligible due to a much higher price of GDH compared to that of G6PDH.












HO--OP OH NADP
HO' 'OH X

D-glucose-6-phosphate
G6PDH YGL157w





-ucono-6-atone-6-phosphat NADPH
Hbuffered medium containing NADP

buffered medium containing NADP


OH 0 0
Ch.. Ot-Bu

t-Butyl (R)-6-chloro-5-hydroxy-3-oxohexanoate






CI o o Ot-Bu
t-butyl 6-chloro-3,5-dioxohexanoate


Figure 2-4. Glucose-6-phosphate dehydrogenase (G6PDH) as an auxiliary enzyme.


HO. H NADP+

D-glucose GDH
GDH



H 0 4 \ NADPH
HO" 'OH
OH
D-glucono-S-Iactone


0 OH


4-hydroxy-3 methyl-2 pentanone




0 0


3 methyl-2,4 pentanedione


buffered medium containing NADPH


Figure 2-5. Glucose-6-phosphate dehygrogenase as an NAPDH regeneration catalyst.

Another concern about G(6P)DH regeneration systems is that both produce

stoichiometric amounts of gluconic acid, which may impede the downstream purification

process. To circumvent this issue, two routes have been explored so far: i) the use of a

hydrogenase, which regenerates NADPH at the expense of the cheapest reducing

agent, dihydrogen103 (Figure 2-6) and ii) the change of an enzyme's cofactor 'diet' from

NAD+ to NADP+. The latter approach was first attempted for format dehydrogenase

(FDH), which releases CO2 as the only by-product.115 The first NADP+-dependant FDH

from Pseudomonas sp. (mut-Pse FDH) is commercially available from Julich Chiral

Solution GmbH at a cost similar to GDH.116 In addition, a phosphite dehydrogenase









(PTDH) has been converted to an NADP+ dependent enzyme with promising

characteristics since phosphite is a very low cost reductant and the only by-product in

this case is a phosphate ion.117 However, to the best of our knowledge, none of the

three abovementioned systems have been applied to a larger, at least 1 mmol, reaction

scale and thus we will not consider them further.


H2 NADP+ CH


(1 R,2S)-phenyl-propane-1,2-diol
PF H2ase I ADH M



4H \ NADPH H3

(2S)-hyd roxy-1 -ph enyl-propa none
aqueous buffered medium on a YM10 membrane

Figure 2-6. Hydrogenase I from Pyrococcus furious for NADPH regeneration.

Examples of Microbiological NADPH Regeneration Methods

Under the umbrella of microbiological regeneration techniques, we may recognize

two slightly different approaches, one in which the permeabilization of the cell membrane

is intended and the other in which the E. coli cells are considered as absolutely self-

contained bioreactors. This feature implies that the cells are metabolically active and

thus the integrity of their membrane is a prerequisite for a successful biotransformation.

Permeabilized E.coli cells

In one of the first examples of permeabilized E. coli cells, Kataoka et al.118 grew

separately two JM109 E. coli strains, one overexpressing a GDH from Bacillus

megaterium and another overproducing an aldehyde reductase from Sporobolomyces

salmonicolon (Figure 2-7). The harvested cells were mixed together in different ratios

and resuspended in potassium phosphate buffer containing catalytic amounts of NADP+.










The presence of an organic layer of n-butyl-acetate rendered the cell membrane

susceptible to the free transport of nicotinamides in between the two different strains.


OH OH 0
.......... NADP' 4......... .... ciP o
HO OH
OH (R)-4-chloro-3-hydroxybutanoate
D-glucose GDH KAR


HO 00 OO
HO' OH ^--- NADPH."""... ... cifl o-
OH E. coli cell E. coli cell
D-glucono- -lactone 4-chloro-3-oxobutanoate

aqueous phase containing NADP organic phase

Figure 2-7. An example of two coupled E. coli strains in a two-phase system.

A slightly modified approach will be the overproduction of both the auxiliary

enzyme and a reductase in the same E. coli cell (Figure 2-8). Ema et al.119 applied this

concept coupling the regeneration of NADPH by GDH with the reduction of 2, 4

octanedione by the Gre2 from Saccharomyces cerevisiae. Even though a disruption of

cell membrane or a pretreatment of the cells with organic solvent was not employed in

this case, the authors reported that they froze the cells at -20C before their utilization.

Hence, a semi-permeabilization of the cell membrane could be assumed due to this

treatment. Interestingly, a 10% conversion was observed even with cells that

overexpressed only Gre2.


HO--0 0 .OH A OH 0
10 .......... NADP ..... ,1.
HO'"V 'OH
HH -l s GDH^ (S)-2-hydroxy-4-octanone
D-glucose GDH SCR



HO OH *"" NADPH .....
OH E. coli cell
D-glucono-5-lactone 2,4 octanedione
buffered medium containing NADP*


Figure 2-8. An example of an E. coli cell overexpressing both GDH and a reductase.









A more complex scenario is depicted in Figure 2-9. In this case the authors tried

to couple the most popular auxiliary enzyme for the regeneration of NADH (i.e. format

dehydrogenase (FDH) from Candida boidinii) with the recombinant membrane-bound

transhydrogenase from E. coli (PNTAB) which in turn shuttled the hydride from NADH to

NADP+.120 The NADPH produced was used by an alcohol dehydrogenase (ADH from

Lactobacillus kefir) for the reduction of acetophenone. The whole process was carried

out in the same E. coli cell, permeabilized with 1% toluene. Unfortunately, no application

on a larger scale (i.e. > 1mmol) is reported for this system.

o
OH
.... .
NAD+ + H+ .- .. """^ o^
format **(R)- phenylethanol

FDH / ADH


co, +" NADMNADP NADPH + H+
c02 .. NADH ...... N'
E. cocell PNTAB acetophenone


buffered aqueous medium containing NAD* & NADP*

Figure 2-9. An example of an E. coli cell overexpressing two auxiliary enzymes and a
reductase.

Surprisingly, Xu et al.121 reported recently that under optimal conditions, a two

strain system similar to the one developed by Kataoka et al., is able to perform the same

bioconversion without the addition of an organic solvent and/or extra NADP+. In this

example, the overproduction of GDH in the first E. coli strain with the aid of high

concentrations of 4-chloro-3-oxobutanoate caused an efflux of NADPH into the medium

in the presence of glucose. Subsequently, the reduced nicotinamide was transferred

into the second E. coli strain which is responsible for the transformation of the keto-ester

to the corresponding chiral alcohol (Figure 2-10). A transmembrane transportation of










NADPH was hypothesized and the percentage conversion found dependent on the initial

concentration of the ketone with higher conversions at higher concentrations. It is worth

noting that external addition of NADP+ did not improve the bioconversion in the two

strain system but was found necessary when a one strain system, overexpressing both

auxiliary and alcohol dehydrogenase, was utilized.


H" .......... \ ....NADP 4 ......... .... c A
HO' 'OH
-oH (S)4-chloro-3-hydroxybutanoate
D-glucose GDH CAR


Hn-O 00 .. 00....... 0
HO'" 'OH NADPH .......... b cI Jll k0
OH E. coli cell E. coli cell
D-glucono-8-lactone 4-chloro-3xobutanoate
aqueous buffered medium

Figure 2-10. An example of two E. coli strains coupled for regeneration of NADPH in
one phase system.

Intact E.coli whole cells

In 2004, Walton and Stewart reported one of the most self-sufficient systems in this

category.122 In this approach, the workers relied absolutely on the E. coli metabolism for

the NADPH regeneration without the need to overexpress any auxiliary enzyme and/or

add extra amounts of NADP(+/H ) in the aqueous medium (Figure 2-11). Additionally,

this example is one of the rare cases where the amount of sacrificial substrate (i.e.

glucose) added is not in excess, as compared to the amount of product made. The

overproduced reductase was Gre2 from bakers' yeast.












H
HO ........

HO'' 'OH
H j
D-glucose

O O


ethyl acetoacetate


CO2
&
organic acids


OH 0


Ethyl (S)-(+)-3-Hydroxybutyrate


aqueous buffered medium


Figure 2-11. An example of intact E. coli cells capable to regenerate NADPH for ethyl
acetoacetate reduction by a ketoreductase.

A more sophisticated E. coli strain (SC16529) was developed by Hanson and

coworkers123 to meet the needs for the production of a chiral chloro-alcohol (Figure 2-12)

on a large scale. In this case, G6PDH from Saccharomyces cerevisiae was

overexpressed with a ketoreductase from Hansenula polymorpha in the same E. coli

cell. The authors relied on the E. coli's PTS activity for the phosphorylation of glucose.

However, the addition of catalytic amounts of NADP+ was found necessary for the

completion of this biotransformation.


H
.PTS -PON-,
HO''''' 'OH

D-glucose G6PDH


organic acids
&
CO2


E. coli cell (SC16529)


OH
..... F Y^' -c





0

4H**** N
F-


buffered medium containing NADP+


Figure 2-12. An example of intact E. coli cells overexpressing a G6PDH and a
reductase.


',o






50


Economical Analysis

How the different NADPH regeneration approaches could be evaluated from an

economical perspective? A very informative parameter regarding this issue is the total

turnover number (TTN), which is defined as the molar amount of product made by the

end of the reaction per molar amount of cofactor used.124 This number embraces

several factors such as the relative stability of the cofactor under the reaction conditions,

the reaction time-frame and the catalysts turnover number.125 It is apparent from the

Figure 2-13 that the higher the TTN, the more cost-efficient the bioprocess is. As proof,

we present a cost analysis of the aforementioned examples and a tabulation of the

corresponding TTN for each of them in Table 2-2. An estimation of intracellular

NADP(+/H) content has to be made for those cases that whole cells were used. For this

calculation we assumed 440,000 molecules of NADP(+/H) per E. coli cell.


Cost of NADP* vs. TTN
1 assuming $161,0001/mo of cofactor)





z-





0 2 i. .
o






0 1 2 3 4 5 6

Iog(TTN)


Figure 2-13. Cost of NADP+ as a function of TTN.











Table 2-2. Cost analysis of different NADPH regeneration methods and tabulation of TTNs.
Year Cofactor used Sacrificial Auxiliary Biomass Theoretical Productivity Specific TTN
(ref.) (amount, cost) substrate enzyme in grams Yield in (mmol h-' g-1 Cost
(amount, cost) (amount, (cost)a mmols CDW) ($/mol)
cost) (Total cost)
1998 NADP+ (0.019 glucose GDHb 3 ($4.5) 83.6c ($7.6) 4.2 100 3,980do
(118) mmol, $3.1) (166 mmol, ngl) (261U, ngl)
2000 NADP+ (0.033 acetone (66 recLBADH -- 9.9e ($44.3) --- 4,400 300'
(111) mmol, $5.3) mmol, ngl) (600U, $39)m
2004 NADP+ (0.17 G6P9 G6PDH9 --- 2.8e ($267) --- 95,000 16.5'
(113) mmol, ($27.3) (4.4 mmol, $140) (2mg, $100)
2004 --- glucose (216 --- 17.5 ($26) 348 ($26) 7.0 70 27,840',n
(122) mmol, ngl)
2005 NADPH (0.05 glucose GDHh --- 4' ($410) --- 103,000 80'
(114) mmol, $35) (10 mmol, ngl) (50mg, $395)
2005 NADP+ (0.15 glucose ( 167 G6PDH 75 72.4 0.13 1,890 337d
(123) mmol, $24) mmol, ngl)na, ngl) ($112.5) ($136.5)
2006 NADP+ (0.012 glucose GDH 2 ($3) 4.0k($ 4.9) 0.7 1,200 296d,o
(119) mmol, $1.9) (8mmol, ngl) (na, ngl)
2006 -- glucose GDH 0.22 ($0.2) 5.7C($ 0.2) 7.8 30 35,625dn
(121) (10mmol, ngl) (27U, ngl)___
a Assuming $1.5/g (wet weight). b In-house prepared. c Product recovery was not attempted. d Optimized conditions. e 77% recovered
product. 'not optimized conditions. g Based on Sigma-Aldrich catalog (2006).107 h Based on Biocatalytics Inc. catalog (2006).106 '90%
recovered product. J76% purified product. k 71% recovered product.' 89% recovered product.m Based on JOlich Chiral Solution GmbH
catalog (2006).116 n For this calculation, the sum of NADP+ and NADPH pool is considered based on the values given in Table 1 and
assuming that the mass of one cell is 1.0x10-12 g (wet weight).108 0 Based on the NADP+ added and NADP(+/H) content of biomass
used. Abbreviations: ngl, negligible; na, not applicable.









The major conclusions from this analysis are:

* Regeneration methods based on pure auxiliary enzymes are too expensive even
on a gram-scale reaction. Only in-house prepared crude extracts of
overexpressed proteins might accommodate the problem.127
* A substrate-coupled regeneration method may be affordable if a TTN of 300 or
higher is achieved and a high-value product is produced.
* The utilization of whole cells, intact or permeabilized, is the only foreseeable
avenue for the large-scale production of fine chemicals, in which NADPH is
mediated. This is particularly true for low-value chemicals.
* The misconception that an E. coli cell does not contain an efficient NADPH-
regeneration system is a myth,128 at least for NADPH regeneration rates in the
order of 120 U/g (CDW).

At the moment, several different methods that utilize E. coli whole cells

overproducing oxidoreductases are under development. Therefore, caution must be

paid when we are evaluating data such as the one presented in the first and the last row

of Table 2-2 since isolated yields were not reported and the reaction scale in the last

case was quite low (< 10 mmol). We should also be aware of a possible failure to repeat

successful stories as the one presented in the forth row of Table 2-2 with substrates that

may interfere with the intracellular metabolism of E. coli or they could not access the

cell's membrane.

We also should not mislead the reader by making her or him believe that the TTN

is the only effector on the final cost of a large-scale bioprocess. Useful reviews129-131 and

discussions132, 133 exist in the literature about other factors that should be evaluated

before a biocatalytic process could be considered successful, but they are beyond the

scope of this introduction.

Manipulations of NADP(H /*) Levels in E. coli towards the Improvement of
Biotechnological Processes

A number of researchers have studied the manipulation of the redox-metabolism in

various microorganisms hoping to improve the biosynthesis of metabolites of industrial

importance.134-137 However, similar efforts in E. coli are rare138'139 In one of them, Flores

et al.140 tried to determine if the overexpression of zwfin an E. coli JM101 strain could









overcome the metabolic burden imposed on it by the presence of a high copy plasmid.

Often, plasmid-DNA replication and synthesis of plasmid-encoded proteins reduce the

growth rate of the producing culture. By increasing the expression of zwf and hence

increasing the carbon flux through the PP pathway, the authors were able to partially

recover the growth capabilities of an E. coli strain harboring a multicopy plasmid

responsible for the production of a recombinant TrpLE-proinsulin peptide. The growth-

rate values were three times higher upon overexpression of zwf, but still twice lower than

the rates for the wild type. When the degree of zwf gene induction was altered from 3 to

100 times in terms of higher levels of zwf expression, it was found that an

overexpression as little as three fold was adequate for the growth-rate of the strain

overproducing the TrpLE-proinsulin to reach a plateau. Unfortunately, it is not clear from

this study if NADPH or another metabolite from the oxidative part of PP pathway, which

might be produced in excess, was responsible for the observed results.

The second example is related to biosynthesis of the biodegradable polymer PHB

from acetyl-CoA (Figure 1-13). The process requires one equivalent of NAPDH from

each unit attached to the polymer. Lim et al.141 constructed plasmids containing zwf or

gnd co-integrated with the phb operon and introduced them into E. coli. The amount of

PHB increased after expressing the genes, especially zwf. However, this increment may

be due to altered NADPH/NADP+ (i.e. six times higher compared to wild type), or

redirection of the flux through the AcCoA, or the increased activity (i.e. 1.16-1.30 fold) of

all the enzymes involved in PHB synthesis upon zwf overexpression. Again, which one

was the crucial parameter was not clear from this study.

In a very recent study Sanchez et al.142 followed a different approach to tackle the

same problem. They overexpressed the soluble pyridine transhydrogenase, UdhA, in

order to increase the yield and productivity of PHB in E. coli. The outcome from this

manipulation is crystal clear: almost double yield of PHB obtained in the strain









overproducing UdhA (Table 2-3). However, their discussion about the reason behind

this outcome is controversial. The authors stated that by inducing a high level of UdhA,

they increased the NADPH availability and thus the yield and productivity of PHB. They

also commented that the drop of NADPH/NADP+ with the concomitant increment of

NADH/NAD+ in the UdhA overexpressing strain suggests a depletion of NADPH pool

and a higher NADH conversion rate matched with the PHB formation demands.

Recalling, though, our discussion in the first chapter about the physiological role of UdhA

(i.e., transfer of hydrides from NADPH to NAD ) we can also arrive at the same

observations about the nicotinamide ratios in this example (Table 2-3). Nevertheless,

this is the first example in the literature in which a NADPH manipulation in E. coli caused

increased productivity of a NADPH-dependent compound. Also, it is worth noting the

high TTN that this system reached. Analogous studies overexpressing the membrane-

bound transhydrogenase, PNTAB, for the overproduction of xylitol in E. coli have not yet

been fruitful.143

Table 2-3. Cofactor levels of PHB producing strains and tabulation of TTN.
Strain Total yield Productivity NADPH NAPD+ NADPH TTNc
(mmol (mmol h-1 (NADH) (NAD+) /NADP+
g-1 DCW) g-1 DCW) (pmol/gDWC) (pmol/gDWC) (NADH/NAD+)
Control 11.4 0.44 0.375 (0.6) 0.165 (2.6) 2.27 (0.23) 21,111
GJT001 22.8 0.88 0.250 (0.9) 0.185 (2.0) 1.35 (0.45) 52,413
a Expressing only phb operon. b Expressing both phb and udhA. c Based on the
calculated NADP(+/H) content.


Description of the System under Study

In all three aforementioned examples, the merits of the NADPH perturbations to

the obtained results are not clear due to the fact that the provided carbon source

contributes to bioprocess not only by providing reducing equivalents but also by

supplying carbon atoms to the final product with bits of carbon atoms. In contrast, a

single-step biotransformation by a NADPH-dependant reductase like the one presented

in Figure 2-11 has as a net gain only a pair of electrons and a proton, in the form of









hydride. Thus, it exclusively depends on a metabolically active NADPH regeneration

system, without causing a drain to other intracellular metabolites.

In a preliminary attempt to shed light on the NADPH regeneration system within a

resting E. coli cell during the course of a redox enzymatic reaction Walton and Stewart

made the following observations:

* The profile of glucose consumption by the resting E. coli cells was the same in the
presence and in the absence of the biotransformed starting material, in otherwise
identical conditions.
* Both [NADH]/[NAD+] and [NADPH]/[NADP+] ratios remained constant at about 0.4
each, with an intracellular concentration for NADPH about 180 pM.
* The molar ratio between product formation and glucose consumption was about
2.3.

Consequently, they concluded that the relatively low level of [NADPH]/[NADP+]

ratio, which is normally in the range of 0.7-1.3, may be due to additional needs of the

overexpressed NADPH-dependent enzyme. In addition, they made an effort to identify

the NADPH sources in this system by correlating the stoichiometry linking the product

formation with glucose consumption in their system (i.e. 2.3 molar ratio) with the

stoichiometry linking the product formation with glycerol consumption in a similar system

developed by Woodley and coworkers (i.e. 0.96 molar ratio).144 Recognizing the PP

pathway and the TCA cycle as the only NADPH sources in E. coli and setting the

efficiency of the latter at 96% (each mole of glycerol can produce maximum 1 mole of

NADPH), they suggested that the icd was responsible for around 83% [0.96 x 2.0)/2.3] of

the NAPDH consumed by the ketone reduction. Therefore, less than 10% of glucose

might be metabolized through the PP pathway in that system.

The 'weak' points of the abovementioned discussion are the following:

* The [NADPH]/[NADP+] ratio was 0.4 even at time, t=0 h, which is the point just
before the beginning of the bioconversion. Therefore, this ratio simply may reflect
the ratio of the nicotinamide levels at the end of the growing phase of cells and
after their overnight storage at 4C.









* Even though there is a significant scattering in the data related to phosphorylated
nicotinamide levels, it is obvious the trend of faster depletion of NADPH compared
to NADP+ but this was not recognized by the authors. The observed drop in all
nicotinamides levels could be easily explained in a system such as non-growing
cells, in which de novo synthesis of those metabolites is excluded due to lack of a
nitrogen source. Also, the integrity of the cell membrane is under question
especially during the last 12 h of the bioprocess when high concentrations of
organic material have been accumulated. Nevertheless, the estimated TTN for
this system, based on the NADP(H/ ) content provided by authors for it, is as high
as 31,650 (assuming the volume of an E. coli cell = 1 x 10-15L).108

* The data provided by Woodley et al. by no means could be correlated with the
data obtained from Stewart's study because: i) the two E. coli strains were not
isogenic (i.e. Woodley's host strain is isogenic to E. coli K-12 strain, Stewart's host
strain is isogenic to E. coli B strain); ii) the calculated efficiency for the Krebs cycle
is based on completely different carbon sources. Glycerol transport in E. coli is a
facilitated diffusion process (i.e. requires no energy expenditure)68 whereas
glucose transport process requires a high energy phosphate group from PEP; iii)
when Woodley and coworkers examined their system further found that the
amount of product formed within one hour, when glucose was used as a carbon
source was half compared to what they obtained when glycerol was added to the
medium.145 Unfortunately, the amount of consumed glucose was not reported. In
addition, specific activity measurements with crude extracts revealed that the rate
limiting step in Woodley's bioprocess is the transport of starting ketone through the
cell membrane.

To this end, the specific aims of the current work were threefold:

* To develop a method that will allow us to identify the NADPH regeneration sources
in resting E. coli whole cells overexpressing a ketoreductase, a model system for
bioreductions in lab-scale.

* To perturb the NADP(H/P) levels of the host cell by environmental and genetic
manipulation means in order to test the capabilities of the methodology invented in
the first step (Figure 2-14).

* To evaluate the effects of those perturbations on the productivity of the system
under study.

Experimental Strategy

In lab-scale biotransformations, resting cells are usually the preferred method

compared to batch and continuous culture methods. Non-growing cells are live cells that

retain most of the enzymes activities of growing cells and are obtained from the culture

medium at the time in which the activity of the overexpressed enzyme is at satisfied

levels. Thus, the bioconversion is divided into two parts. In the first phase, cells are










grown in rich medium (e.g. LB medium) while overproducing the desired biocatalyst. In

the second phase, the concentrated cells are resuspended in a modified culture medium

lacking nutrients for growth (e.g. M9 medium without NH4CI) and the bioconversion takes

place. This approach ensures 'cleaner' reactions and easier isolation of products. In

addition, cells from the first phase could be stored and used on demand at the desired

amounts and under controlled conditions in the second phase. However, in such a

system, the popular method of METAFoR analysis of 13C-labeled amino acids by NMR

or GC-MS cannot be applied for the calculation of NADPH-related fluxes. An alternative

choice would be a radiorespirometric method with all the disadvantages that we

discussed in the first chapter.







ADPH-




















represented with X.
6PGn r






o 0e ..................w t X

















Instead, we thought that a modified approach similar to one employed by Csonka

and Fraenkel, which was described in detail in the first chapter, would be more
..... ...........................
.........

::. ..... .......... ...
D ................ """
.........
.. ...............;1













and Fraenkel, which was described in detail in the first chapter, would be more









informative. The major difference between the two systems is that in their system the

labeled at a specific position of glucose atom (i.e. tritium) was transferred to NADP+ and

then to certain amino acids, whereas in our case the labeled atom is transferred to ethyl-

acetoacetate by the overexpressed NADPH-dependant keto-reductase, Gre2. An

overview of the experimental strategy is depicted in Figure 2-15. The tritium was

replaced with the stable isotope deuterium and thus GC-MS can be used to directly

measure the deuterium content in the ethyl-3-hydroxybutyrate. The major advantage of

this method is that is sensitive and not destructive. Thus, a small volume of sample is

required and numerous of samples can be analyzed during the course of the same

biotransformation.

The labeled sugars that could possibly give useful information about the NADPH

sources in this system are listed in Table 2-4. In this study only the glucose-1-d, -3-d,

and -6,6 d2 were used. A detailed description about the fate of each of these

isotopomers in the central carbon catabolism is given in Chapter 1 (see Figure 1-1,1-2,1-

3 and1-4). In summary, glucose-id and -3d could 'trace' the NADPH sources in PP

pathway and they could prove or eliminate the existence of the ED pathway. The

glucose -6,6 d2 could trace the activity of icd in the TCA cycle and perhaps the activity of

meaB. The glucose-4d should give labeled NADH, which in turn should label NADP+ if a

transhydrogenase activity is hypothesized. In principle, if glucose labeled in each

position was used, almost all the produced NADPH should be labeled.

Table 2-4. Labeled sugars. Cost and possible application.
D-Glucose Cost ($/g)a Application
-1-D 180 Zwf activity
-3-D 570 Gnd activity
-4-D 1,700 transhydrogenases activity
-6,6 D2 75 icd activity
-(1,2,3,4,5,6-D7) 570 Total activity (Control)
a Based on Medical Isotopes Inc. catalog.146











air or argon


Reaction Conditions:
10ml Minimal medium-
nitrogen source + cells
20mM labeled glucose
20mM ethyl-acetoacetate





water circulator
at 30-32 C


o0 0 G OH 0 OH 0
NADPH(D)
ethyl aceloacetate eth 3yl yaox>ybuaye




Sample preparation:
50ul of rxn mixture
extracted 2x with
EtOAc


GC-MS analysis
MS of P


Time (min)

MS-fragmentation


o..H "o.

H(D) "


S-e

H ;(D)
Uj^


C-I
CH3


-N


Analysis of data


A72 (A1+ A72) X 100 = % NADPD I (NAPDH + NADPD)


Figure 2-15. Illustration of the experimental procedure. Abbreviations: P, product. SM,
starting material; IS, internal standard.


o H 0









The original bioreduction described by Walton and Stewart was carried out in a 3

L-scale containing 4 g/L glucose. Considering the high cost of the labeled sugars (Table

2-4), all biotransformations in this study were carried out in a 10 ml or 100 ml-scale

under two distinct physiological conditions: i) under argon atmosphere (i.e.

anaerobically) and ii) under air atmosphere (i.e. aerobically). In this context, the

conditions under air atmosphere are named 'aerobic', but they may also be called

'microanaerobic' due to relatively high density of cell cultures (O.D.600~8.0) used.147' 148

In addition to environmental perturbations, we also genetically perturbed the

system at four positions. The first two were around the G6P node, disrupting either the

pgi or the zwfgene. We expected that in the former mutant, the flux through the PP

pathway would be maximized whereas in the latter case it would be diminished. The

third one, an icd mutant, was at the isocitrate node, hence excluding the possibility of

NADPH production from TCA cycle. The last disruption was the pntAB gene, which

encodes the membrane-bound transhydrogenase responsible in the wild type strain for

the production of NADPH from NADH. All mutant strains are isogenic to original strain

BL21(DE3) used by Walton and Stewart for the ethyl-acetoacetate reduction. This host

is necessary for the overproduction of Gre2, for which the expression is under the

control of T7 promoter.

We should make clear at this point that the measurements with labeled sugars and

the different BL21(DE3) strains performed in this study allowed us to calculate the

contribution of each NADPH source to the total NADPH pool present in the cell, in each

strain. By no means is this method capable to measure net fluxes in the metabolic

network.

Finally, we examined the capabilities of the mutant strains toward the production of

ethyl-3-hydroxybutyrate compared to that for the wild type in a shake flask.









Experimental Procedures

Materials

D-Glucose-1-d (97% enrichment), D-Glucose-6,6-d2 (98% enrichment) 2-propanol-

d8 and sodium deuteride (98% enrichment) obtained from Aldrich. D-Glucose-3-d (98%

enrichment) was obtained from Medical Isotopes Inc. Glucose-6-phosphate

dehydrogenase (from Bakers yeast Type XI) and alcohol dehydrogenase (from

Thermoanaerobium brockii) were purchased from Sigma. The purified GST-Gre2 fusion

reductase from S. cerevisiae was prepared by Dr I. Kaluzna in our lab.149

Oligonucleotides were purchased from IDT (Coralville, IA) as deprotected and

desalted. Long primers were IE-HPLC purified. Taq and Vent polymerases were

obtained from New England Biolabs. In-house prepared Pfu was used instead of Vent in

some cases.

Trinder reagent, for glucose assay was purchased from Chemical Diagnostics.

Bacterial Strains and Plasmids

Bacterial strains and plasmids were obtained from the Coli Genetic Stock Center

(CGSC) (Appendix B). The icd knockout strain (BL21(DE3)Aicd) designated as MA1935

was kindly provided from Dr M. Aoshima.150 The pntAB knockout strain,

(BL21(DE3)ApntAB) was made by Sylvie Boualavong in our lab. Plasmids were purified

by density ultracentrifugation with CsCI in the presence of ethidium bromide. Genomic

DNA from E. coli BL21(DE3) strain was purified as describe elsewhere.151 E. coli strain

overexpressing S. cerevisiae Gre2p [BL21(DE3)(pAA3)] has been described

elsewhere.152 All mutant strains were transformed with pAA3 via electroporation.

Cell Cultivations

LB medium used for cell growth contained Bacto-Tryptone (10 g/L), Bacto-Yeast

Extract (5 g/L) and sodium chloride (10 g/L). Agar (15 g/L) was used to solidify media

for plates. M9 nitrogen-free medium used for bioconversions under non-growing









conditions contained disodium phosphate (12.8 g/L), monopotassium phosphate (3 g/L),

sodium chloride (0.5 g/L), magnesium sulfate (2 mM) calcium chloride (0.10 mM) and

glucose (0.4%). The last three ingredients were added as sterile aqueous solutions (1

M, 0.10 M and 20%, respectively) after other materials had been autoclaved. M9

minimal medium used for phenotypic characterization of mutants was supplemented with

ammonium choride (1 g/L) and glutamate (0.5 mM) in case of MA1935 strain. SOB and

SOC media was prepared as described by Sambrook et al.153

Analytical Methods

GC-MS analyses were conducted with a Hewlett-Packard Model 5890 GC/MS

fitted with a 30 m DB-17 column using helium as carrier gas and a temperature gradient

from 600C to 1800C at 10C/min with initial and final times 2 and 5 min, respectively.

The MS was operating at electron impact ionization mode. Biotransformation reaction

mixtures were sampled for GC/MS analysis by mixing 50 pl of the aqueous reaction

mixture with 100 pl of ethyl acetate containing 1.0 mM methyl benzoate (internal

standard). After vortex mixing, the organic layer was separated and the aqueous layer

was extracted a second time in the same manner. A sample of the combined organic

layers was analyzed by GC/MS. A standard curve for ethyl-(S)-3-hydroxybutyrate linear

up to 70 mM was constructed by preparing aqueous solutions containing varying

concentrations of single analyte which was extracted and analyzed by GC/MS as

described above.

Glucose concentrations were determined using the Trinder reagent kit. Five pl

aliquots of water (blank), 5 mM glucose (standard) and the unknown sample, after

centrifugation, were added to three separate 1 ml aliquots of reconstituted Trinder

reagent. After gentle inversion, the tubes containing each were incubated at 370C for

five minutes. Immediately following incubation the absorbance (A) at 505 nm was









measured. Unknown concentrations were determined based on the equation: [glucose]

= 5 mmol/L x (Asample/Astandard)

In vitro Construction of a Linear Targeting Fragment by PCR: Synthesis of
Targeting Marker with Long Flanking Homologous Regions (LFH)

The synthesis of the FRT-kan-FRT cassette with long flanking homologous (LFH)

regions of E. coli BL21 (DE3) DNA was carried out as described by Derbise et al.154 The

procedure is based on the overlap extension PCR method and illustrated in Appendix A.

In the first step two sets of primers were used to amplify the -500-bp regions flanking the

target gene of the E.coli BL21(DE3) chromosomal DNA. The chimeric primers used in

each pair contained a 5' end homologous to one of the ends of the FRT-kan-FRT marker

and a 3'end that primed to the upstream or downstream region of the desired knockout

fragment (step A). The sequence of primers was based on the known genomic DNA

sequence of the E. coli K-12 MG1655 strain.155 The plasmid pKD13 was used as a

template to amplify the kanamycin cassette with the primers set pKD13 F/R (step B). In

the second step -200 ng of upstream and downstream PCR megaprimers were mixed

with -1000 ng of the FRT-kan-FRT fragment, and with a second set of short primers

annealing at the two ends of the targeting cassette. PCR cycling conditions were: 25-30

cycles of 94C for 1 min, 54C for 2 min and 72C for 3 min followed by 1 cycle of 72C

for 10 min. In the first and second step Taq and Vent polymerases were mixed in ratio 3

to 1.

The obtained products were concentrated by ethanol precipitation and purified by

low melt agarose gel electrophoresis. The Dpnl treatment of the PCR product from Step

A before the gel-purification facilitates the degradation of the methylated plasmid

template, which otherwise generates the selective genotype after the transformation with

the targeting marker. The purified products were suspended in TE buffer (10 mM Tris-

HCI, pH 8.0, 1 mM EDTA, pH 8.0) and quantified by agarose gel electrophoresis.









Gene Replacement

E. coli BL21(DE3) cells were transformed, by electroporation, with the temperature

sensitive replicon plasmid pKD46 containing the A red recombinase genes under the

control of the inducible arabinose promoter. BL21 (DE3) [pKD46] cells were grown at

30C overnight to saturation in LB medium with ampicillin (100 pg/ml), diluted 100-fold in

50ml SOB (plus 100ug/ml ampicillin) and grown to OD600 of 0.1. L-Arabinose was added

to 1 mM final concentration to induce the expression of the recombinase genes, and

cells were grown to an OD600~ 0.5, before chilling into an ice slurry for 20 min. Cells

were then centrifuged at 7000 rpm for 10min at 4C. The pellet was resuspended in ice-

cold deionized water and centrifuged. Then the pellet was resuspended in ice-cold 10%

glycerol and centrifuged again. The washing step with glycerol 10% was repeated twice,

and the cell pellet was suspended in a final volume of 300-500 pl ice-cold 10% glycerol.

Electroporation was performed using ice-cold cuvettes and a Bio-Rad Gene Pulser

set at 2.5kV, 25 uF with pulse controller of 200 Ohms. Transformations were usually

performed: 40 pl of induced cells, 40 pl and 50 pl of induced cells mixed with 1pl

(30~70ng) or 2pl (60-140 ng) of linear marker, respectively. SOC medium (0.6 ml) was

added after electroporation. The cells were transferred to 15 ml test tubes and

incubated at 37C for 1-2h with shaking. The transformants were spread on LB plates

containing 20-40 pg/ml kanamycin. The plates were incubated at 37C to select for

recombinants.

Detection and Analysis of Recombinants

Bacterial colony PCR was used to detect E. coli transformants having an altered

structure caused by replacement of a gene. Each colony was touched with a sterile

yellow tip and the cells were resuspended in 4-8 pl of sterile water. The PCR reaction

mixture for 2 pl of the above suspension in 100 pl of total volume was 1 pl of Taq, 10 pl

10x Pfu buffer (200 mM Tris-HCI pH 9.0, 0.2 mM MgSO4, 10 mg/ml BSA, 1 mM KCI, 1









mM (NH4)2S04, 1% Triton X-100) 0.2 mM dNTPs and 200 ng of each primer.

Thermocycler parameters were the same as described before. Primers used for

screening are listed in Appendix C. Products were analyzed on an agarose gel.

Deletion of the Inserted Marker

The E. coli BL21(DE3) clones resistant to both kanamycin, kanR, and ampicillin,

ampR, were cured of the pKD46 plasmid by growth at 37C, a temperature not

permissive for this replicon, and spread on LB, on LB plus amp (100 pg/mL) and on LB

plus kan (40 pg/mL) in different dilutions to select for amps colonies. KanR mutants were

transformed with pCP20 that has a temperature-sensitive replication and thermal

induction of FLP recombinase synthesis. Since the kan gene is flanked by the short

direct repeats called Flp Recombinase Targets (FRT sites) the Flpase promotes

recombination at those specific sites excising the kan cassette and leaving behind a

short nucleotide sequence with one FRT site. Ampicillin-resistant transformants were

selected after 1-day incubation at 30C, and were screened for kan deletion with colony-

PCR as described above. Colonies having the deletion were grown non-selectively at

37C and then tested for loss of all antibiotic resistances.

Preparation of [4S-2H] NADPH

The preparation of (S)NADPD was based on a method described by Viola et al.156

31mg NADP+ and 1.5 fold excess of D-glucose-1-d were dissolved in 4.3 ml phosphate

buffer, 83mM, pH 8, containing DMSO at 40% (v/v) final concentration. 50 U of glucose-

6-phosphate dehydrogenase were added and the reaction mixture was stirred gently at

RT. The progress of the reaction was followed by taking aliquots every hour and

measuring the increment of the absorbance at 340nm. Maximum absorbance was

achieved after 5h, conversion 95%. The reduction of NADP+ was almost complete

because of the hydrolysis of gluconolactone to gluconate.









The reduced nicotinamide was purified by a method described by Orr et al.157 with

minor modifications. The G6PDH was separated from the solution by ultrafiltration

(10.000 molecular weight cutoff filter). The NADPDB sample (5 x 1ml) was loaded on a

FPLC Mono Q HR5/5 anion exchange column and was eluted by a 0-1 M ammonium

bicarbonate gradient.158 Fractions showing a A260/A340 ratio between 2.3-2.7 were

pooled (total volume: 15 ml) and lyophilized overnight affording 20 mg of white powdery

solid containing 12.5 mg NADPD (based on A340). Final yield: 43%.

Preparation of [4R-2H] NADPH

The preparation of (R)NADPD was based on a method described by Gready et

al.159 33 mg NADP+,1.2 ml isopropanol-d8 and 4U of TBADH were added in 15ml of

25mM Tris buffer, pH 9 and the reaction mixture was stirred gently at 43 C. The

thermostability of TBADH allowed the reaction to be carried out at this temperature,

which also prevented the reverse reaction by evaporating the volatile product, acetone.

The progress of the reaction was followed as before and the maximum conversion, 75%,

was achieved after 3h. The reduced nicotinamide was purified as described above.

Final yield: 39%.

Preparation of D-glucose-ld

A 5 ml aqueous solution of sodium borodeuteride (222 mg or 5.3 mmol) was

added dropwise to a 10 ml aqueous solution of 5-gluconolactone ( 1.3 g or 7.3 mmol) at

-4C-0 C at such a rate that all the reducing agent was added in 20 min.160 The pH of

the reaction mixture was kept at 3-4 (for every -10 drops of NABD4 solution, 4-5 drops of

1 M of H2S04 solution were added). 10 min after the last addition of the reducing agent

the reaction was quenched with 0.5 ml of 1 M H2S04 and the solution was neutralized

with 2 M NaOH. 50 ml of methanol was added and solvent was evaporated off under

reduced pressure as methyl borate.161 The procedure was repeated four times until no

boron was present in the solution based on the carmine reagent.162 The obtained syrup









was lyophilized overnight, redissolved in 2 ml D20 and lyophilized again for 20 h

affording 1.9 g of white powdery solid containing 1.3 g of glucose based on Trinder

reagent assay. The labeled glucose was used without further purification.

Reduction of Ethyl-acetoacetate by Gre2p GST-fusion Reductase Using NADPDB
or NADPDA

The stereospecificity of hydrogen transfer of NADPH catalyzed by Gre2p was

determined by reacting 7mg (-1.6 pmol) of NADPDB, with 3.2 pmol ethyl-acetoacetate

and -450 pg enzyme in total volume of 2.5 ml 100mM KPi pH 7. The reaction was

monitored by following the decrement in A340. The oxidized co-factor was purified by

FPLC as described before, lyophilized, resuspended in D20 and analyzed by 1H-NMR.

The procedure was repeated for the NADPDA. The obtained ethyl-3-hydroxybutyrate

was also analyzed by GC-MS.

Reduction of Ethyl-acetoacetate by Gre2p GST -fusion Reductase Using a mixture
of NADPH and NADPDB

In a solution of 704 pl 100mM KPi pH 7.0 were added 83 pl of a stock solution of

NADPDB (24.8 mM), 83 pl of a stock solution of NADPH (21 mM) and 80 pl of a stock

solution of ethyl-acetoacetate (80 mM). The reaction was initiated by adding 50 pl of a

stock solution of Gre2 (-2.4 pg/ml) and was monitored by following the absorbance at

340 nm and by extracting 50 pl aliquots for GC-MS analysis.

Standard Procedure for Ethyl-acetoacetate Biotransformations with Resting cells

A single colony of the appropriate stain harboring the pAA3 was used to inoculate

10 ml of LB media supplemented with 40 pg kanamycin and the culture was shaken

overnight at 37C. Five ml of this solution was added to a 2 L baffled flask containing

500 ml of LB media, 4 g/L glucose, 40 pg/ml kanamycin. The culture was incubated at

37C with shaking. Upon reaching an OD600~0.7, IPTG was added to a final

concentration of 0.1 mM and the incubation was continued at 30 oC for 6 to 8h. The









harvested cells were resuspended in 5 ml nitrogen free medium and kept overnight or

until used (maximum 2 days later) at 4 oC.

Bioconversions of ethyl-acetoacetate were carried out either in a 1 L flask or in 300

ml vessel or in 25 ml vessel containing 100 or 10 ml M9 medium without NH4CI and cells

prepared as described above at different optical densities. At the beginning, glucose

and the organic compound were added to a concentration of 20mM and then appropriate

amounts of both were added whenever was necessary based on glucose consumption

and product formation measurements The reaction was incubated at 32C with shaking,

when in flask, or by stirring, when in vessel. The vessels were equipped with a magnetic

bar and an ace gas dispersion tube connected to an air or argon line. Most experiments

regarding the Aicd and ApntAB strains were performed by Sylvie Boualavong.

Results

Characterization of Mutated Strains

All knockout mutants (Table 2-5) but Aicd were constructed by a marker-free

deletion method163 of genes from start to stop codon as described in experimental

section.

Table 2-5. E. coli strains used in this study.
Strain Pertinent genetic markers
VVT [BL21(DE3) (pAA3)] wild-type strain overexpressing Gre2 reductase
Apgi (DJB254) BL21(DE3) Apgi (pAA3) (phosphoglucose isomerase deficient)
Azwf (DJB370) BL21(DE3) Azwf (pAA3) (glucose-6-phosphate
dehydrogenase deficient)
Aicd (MA1935) BL21(DE3) Aicd (pAA3) (isocitrate dehydrogenase deficient)
ApntAB (SB2) BL21(DE3) ApntAB (pAA3) (membrane-bound
transhydrogenase deficient)

Generally any number from one to 50 kanamycin resistant colonies was obtained

after the transformation of parent strain with the FRT-kan-FRT cassette. The

recombination efficiency was from zero to 100 % based on about 13 colonies screened.









In the former case the transformation was repeated with same or fresh-prepared linear

DNA.

Correct marker integration and excision was verified in all mutant stains by colony

PCR. The PCR primers used were annealing either at the 5' and 3' of the target gene or

500 and 550 bp outside of target's gene coding region. Wild-type starting strain

BL21(DE3) was used as a comparison. Results are summarized in Table 2-6, Table 2-7

and Table 2-8 and are depicted in Figure 2-16, Figure 2-17 and Figure 2-18.

Table 2-6. Colony PCR for pgi knock-out.
E. coli strain Primers Set Expected Size Observed Size
(kbp) (kbp)
BL21(DE3) Apgi:: pgiF, pgiR no band -1.65, -0.9, -
FRT-kan-FRT 0.8 (major)
BL21(DE3) Apgi:: Up500pgi, 2.33 -2.3
FRT-kan-FRT Down500pgi,
BL21(DE3) Apgi:: Up500pgi, k2 1.21 -1.2
FRT-kan-FRT
BL21(DE3) (pKD46) Up500pgi, k2 no band no band
BL21(DE3) Apgi:: Down500pgi, kl 1.0 -1.0
FRT-kan-FRT
BL21(DE3) (pKD46) Down500pgi, k1 no band no band
BL21(DE3) (pKD46) pgiF, pgiR 1.65 1.65 (major), 0.8
BL21(DE3) (pKD46) Up500pgi, 2.65 -2.6
Down500pgi
BL21(DE3) Apgi:: Up550pgi, 2.43 -2.4
FRT-kan-FRT Down550pgi,
BL21(DE3) (pKD46) Up550pgi, 2.75 -2.7
Down550pgi
BL21(DE3) Apgi:: pgiF, pgiR no band -1.65, -0.9, -
FRT 0.8 (major)
BL21(DE3) Apgi:: Up500pgi, 1.1 -1.0
FRT Down500pgi






























a Upstream pgi
C UP.00 FRT
UpSi, I kan 2 ki


y


FRT

Downstream pgi


Downstream pgi
FRT OownOO
DownM5


A Screening of kanR colonies for the disruption
of pgi, after transformation with FRT-kan-FRT,
with primers Up/Down500 M2 M1
MI M2





-47li


B Further screening of colony #10 (MT)
side by side with the wild type (WT)


pgiFIR Up5001k, Down5001k,


C Colony-PCR of MT before and after of the
FLP-mediated loss of the kanamycin marker


M1 M2 kans kanR kan R


Down500 Fipase
Downstream pgi Down550


Upstream pgi FRT
Apgi locus (-1220bp)


UplDown500 Up/Down550
Figure 2-16. Results from colony PCR reactions probing the pgi locus before and after the disruption. Abbreviations: VVT (wild type),
MT (mutant strain); kanR, kanamycin resistant; kans, kanamycin sensitive; Mland M2, DNA markers.


FRT kan
kan


Upstream pgi


S FRT-kan-FRT cassette


R
e WT locus (-2750bp)
d


Apgi:: FRT-kan-FRT locus (-2430bp)


M2 M1








C
I".-.
156o


Up500
Up550,
i1


M2 M1


-I --









In case of pgi deletion, a weak band with size similar for the structural gene was

detected in kanamycin resistant colony along with multiple other weak bands when the

pair of pgiF/R primers was used. However, we suspected that these low levels PCR

products are due to aerosol contamination or mis-priming since colony PCR with outside

primers confirmed the right construct.

Table 2-7. Colony PCR for zwf knock-out
E. coli strain Primers Set Expected Observed Size
Size (kbp) (kbp)
BL21(DE3) Azwf:: zwfF, zwfR no band -1.3, -1.2, -0.9
FRT-kan-FRT
BL21(DE3) Azwf:: Up550zwf, Down550 2.43 -2.4
FRT-kan-FRT zwf
BL21(DE3) Azwf:: Up570zwf, 2.45 -2.4
FRT-kan-FRT Down550zwf
BL21(DE3) Up570zwf, 2.62 -2.5
(pKD46) Down550zwf
BL21(DE3) Azwf:: Down570zwf, 2.45 -2.5
FRT-kan-FRT Up550zwf
BL21(DE3) Down570zwf, 2.62 2.55
(pKD46) Up550zwf
BL21(DE3) Up550zwf, k2 no band -0.8
(pKD46)
BL21(DE3) Azwf:: Up550zwf, k2 1.23 -1.2 (major), -0.8
FRT-kan-FRT
BL21(DE3) zwfF, zwfR 1.5 -1.5 (major), -0.9
(pKD46)
BL21(DE3) Up550zwf, 2.6 -2.5
(pKD46) Down550zwf
BL21(DE3) Azwf:: Down550zwf, k1 1.1 -1.1
FRT-kan-FRT
BL21(DE3) Azwf:: zwf F, Down550zwf no band no band
FRT-kan-FRT
BL21(DE3) zwfF, Down550zwf 2.0 -2.0
(pKD46)
BL21(DE3) Down550zwf, k1 no band -0.4
(pKD46)
BL21(DE3) Azwf:: Up550zwf, 1.2 -1.2
FRT Down550zwf















Upstream zwf
l


FRT kan
kan


\.


FRT


Downstream zwf
Downstream zwf


A Screening of kan" colonies for the disruption
of zwf, after transformation with FRT-kan-FRT,
with primers Up/Down500S


k I I


FRT-kan-FRT cassette


M2


B Further screening of colony #6 (MT)
side by side with the wild type (WT)


M2 M1



""t


e WT locus (-2610bp)
d


Downstrem zwf
FRT Dow.nzO
Down550


zwfFIR Up5501k, Down550/k,


Azwf :: FRT-kan-FRT locus (-2430bp)


Up500
Up550



Upstream zwf


DownS00
Flpase
Downstream zwf Down550


FRT 4


C Colony-PCR of MT before and after of the
FLP-mediated loss of the kanamycin marker
M2 M1 M1 M2 kans kanR







pDown50
UplDown55O


Figure 2-17. Results from colony PCR reactions probing the zwf locus before and after the disruption. Abbreviations: VVT (wild type),
MT (mutant strain); kanR, kanamycin resistant; kans, kanamycin sensitive. M1and M2, DNA markers.


0a Upstream zwf
c UpfD FRT
SUp550 kman 2 k

i uL
t
Y


Azwf locus (-1220bp)


- r- I --- u









Table 2-8. Colony PCR for pntAB knock-out.
E. coli strain Primers Set Expected Observed
Size (kbp) Size (kbp)
BL21(DE3) ApntAB :: Up510PntAF, 2.35 -2.3
FRT-kan-FRT Down510PntBR
BL21(DE3) ApntAB:: PntAF, PntBR no band no band
FRT-kan-FRT
BL21(DE3) ApntAB:: Up510 PntA, k2 1.23 -1.2
FRT-kan-FRT
BL21(DE3) (pKD46) Up510 PntA, k2 no band no band
BL21(DE3) ApntAB:: Down510 PntBR, k1 1.03 -1.1
FRT-kan-FRT
BL21(DE3) (pKD46) Down510 PntBR, k1 no band no band
BL21(DE3) (pKD46) Up510PntAF, 4.13 -4.1
Down510PntBR
BL21(DE3) (pKD46) PntAF, PntAR 1.53 -1.5
BL21(DE3) ApntAB:: PntAF, PntAR no band no band
FRT-kan-FRT
BL21(DE3) (pKD46) PntBF, PntBR 1.39 -1.4
BL21(DE3) ApntAB:: PntBF, PntBR no band no band
FRT-kan-FRT
BL21(DE3) ApntAB:: Up510PntAF, 1.11 -1.0
FRT Down510PntBR


The growth rate of mutant strains was compared with that of wild type in M9

minimal media. Azwfstrain grew almost as fast as the VVT, at least for the first 8h, as it

was reported before (Figure 2-19A).164 The Apgi mutant showed limited growth and

consumption of glucose (Figure 2-19A), as expected.165 In contrast to what has been

observed for a K-derivatived E. coli strain166 lacking pntAB, our ApntAB mutant grew

without any difference from the VVT (Figure 2-19B). Finally, the icd mutant strain showed

the reported elsewhere150 glutamate auxotrophy (Figure 2-19C). In this case, an

additional portion of glutamate was added after 20h.






























Downstream pntB


FRT
kan f
i rUpstream pntA
___Liw


WT locus
WT locus


A
M1










M1 M2


(-4060bp)


ApntAB :: FRT-kan-FRT locus (-2430bp)

Down510 Up510
FUpstream ppase
Upstream pntA


C
M1 M2









Wii


Downsteam pntB FRT

ApntAB locus (~1100bp)


FRT

Downstream pntB
I I


PntAFIR PntBFIR PntAFIBR Up510/k, Down510lOk



Colony-PCR of MT before and after of the
FLP-mediated loss of the kanamycin marker
M1 M2 kanR kans


I,53
lia
372



Ila
11


U pDown510


Figure 2-18. Results from colony PCR reactions probing the pntAB locus before and after the disruption. Abbreviations: VVT (wild
type), MT (mutant strain); kanR, kanamycin resistant; kans, kanamycin sensitive. M1 and M2, DNA markers.


Screening of kanR colonies for the disruption
of pntAB, after transformation with FRT-kan-FRT,
with primers Up/Down510








Further screening of colony # 4 (MT)
side by side with the wild type (WT)
M1 M2


FRT-kan-FRT cassette



Homologous recombination


pntBR pntAR F
pntB pntB F P pntA U m
I 1














2
1.8
1.6
1.4
I 1.2-

0 1
0.8
0.6
0.4
0.23


0 5 1 15 2


0 5 10 15 20
Time (h)






o



10




l O


o WT


WT
Azwf


2.2
2
1.8
1.6
8 1.4
0' 1.2
O 1
0.8
0.6


o WT
o ApntAB


IjI
------- -


n ----------


0.2


0 5 10 15 20 25 30 35 40 45 50
Time (h)


* Aicd

a Aicd + Glu

o WT


0 5 10


15 20 25 30 35 40 45

Time (h)


Figure 2-19. Phenotype analysis of constructed mutants in M9 medium. A. Growth curves for Apgi and Azwf strains. B. Growth
curve for ApntAB strain. C. Growth curves for Aicd strain with (opened squares) or without (closed squares) glutamate.


u



C
3

2.5

2

S1.5

1

0.5

0


------------------------------------W -


| I ,ll










MS Data Analysis

In this section, the raw MS data will be presented in a form of % fractional

abundance of 72 (m/z) fragment (A72) of ethyl-hydroxybutyrate (EHB) obtained under

different physiological conditions using different glucose isotopomers for all the

aforementioned strains.

In the first experiment, the WT strain was fed with Glc-1D under aerobic

conditions. A 202% incorporation of deuterium into product was observed. This

experiment was repeated at least eight times (Figure 2-20) using different amount of

cells, at two different scales and it was reproducible. In all cases, a plateau for the A72

(%) was observed after the first 2.5 h. We expected that pgi activity should give rise to

labeled product whereas any other NADPH source should not contribute significantly to

it. The second NADPD source in this case is the icd, albeit minor (i.e. pyruvate

produced from Glc-1D is labeled at maximum 17%, Figure 1-3, Chapter 1).


24

23 o = 6.2, 100 ml
SDo,= 7.5, 10 ml
22

0 07 OD,, = .0, 11.00 ml
. 21 -






9 x OD,= 213, 10 ml
18 --- ------------
0 OD^= 19.3, 100 ml
1 + ODa= 2.3, 100 ml

16
0 2 4 6 8 10 12 14 16 18 20 22 24 26

Time (h)

Figure 2-20. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
wild-type strain (WT) using Glc-1D, under aerobic condition.









When the experiment with Glc-1 D was repeated under anaerobic conditions for the

WT, EHB was found labeled at a higher extent, up to 35% (Figure 2-21). However, a

plateau value, if any, was not obtained at the first 6 h. Keeping in mind the equation

presented in Figure 2-15 for the analysis of the data, a higher A72 (%) could be observed

if a higher flux through the PP pathway occurred (i.e higher [NADPD]) or if a major

NADPH source was ceased (i.e. lower [NADPH]) under anaerobic conditions.


44
41
38
; 35 OD 0oo=10, 100ml
J: 000 0 6m
32 o x OD6o0=7.5, 10mi
29 A 00OD0=9, 100ml
^ 26 4
223
20
17
0 2 4 6 8 10 12 14 16 18 20
Time (h)

Figure 2-21. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
wild-type strain (WT) using Glc-1 D, under anaerobic condition.

The distinct labeled pattern of EHB under aerobic or anaerobic conditions was

confirmed with a third independent experiment in which the cells were fed with Glc-1D

under aerobic conditions for the first 8h and then the air supply was halted (Figure 2-22).

The possibility of a branch-point (i.e ED pathway) in between the pgi and the gnd

was excluded by feeding the WT strain with Glc-3D (Figure 2-23). Almost the same

degree of incorporated deuterium into the EHB (i.e., (182)%) was observed as when

Glc-1D was used.











35

S32

29

26

S23

20

17


10 15
Time (h)


20 25 30


Figure 2-22. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
wild-type strain (WT) using Glc-1D, under aerobic and anaerobic conditions.
(Arrow indicates the time at which the air supply was turned off).


U 0





o o
c li


OD 0o=7.5, 10ml

SODo=9, 10ml


0 1 2 3 4
Time (h)


5 6


Figure 2-23. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
wild-type strain (WT) using Glc-3D, under aerobic conditions.

Glc-6,6D2 could 'trace' a possible icd activity. Indeed, when the WT strain was fed

with it, 11-13% of EHB was labeled under aerobic conditions. This is not a minute

number, especially if we consider that less than a half of the pyruvate (i.e. maximum

33%) entering the TCA will be labeled (Figure 2-24).


*



S *e@.


* 00Do=2.3, 100ml


0 5


22

21

20

19

S18

17

16













H PRO-S
H-- -OH
HO 2
HO ---



- OP
H


H PRO-S
H- -OP

HO BH
H
-T-


G3P


PRO-S
H-0
=00


0 NA* O4/OP O. _0 OH
4 4 4 4 O -0
H --H H -5OH H OH H' 5 1 OP
H H OP OOH

PEP PYR
G3P

Figure 2-24. Fate of deuterium atoms of glucose at position 1, 3, 4 and 6 through
glycolysis. Deuterium at position 1 should end up in PEP (25%), however
only the position trans to carboxylic acid should be labeled, and PYR (only
16% labeled). Deuterium at position 3 should be found in water. Deuterium
at position 4 should label NAD Deuterium atoms at position 6 should have
the same fate as the atom at position 1, however PEP and PYR should bear
twice more labeled atoms (50% and 33% respectively) (Adapted from
references 176, 177)

In contrast, under anaerobic condition the % A72 was diminished to 4%, a

negligible number if we consider that even a reduction of EAA with NaBH4 is showing a

3.8% A72 due to the natural abundance of isotopes, as we will discuss in the next

section.

A very similar trend for the produced labeled EHB was detected when the VVWT

strain was replaced by the Apgi strain and fed with Glc-6,6D2 under either aerobic (i.e.,

10-12%, Figure 2-26) or anaerobic conditions. The major difference compared to VVT









was observed when Glc-1D or Glc-3D was administrated to Apgi strain under aerobic

conditions. A higher amount of EHB (i.e., about 273% for Glc-1 D and about 25% for

Glc-3D) was found labeled (Figure 2-27 and Figure 2-28). These results could be easily

explained since in this strain all labeled sugar should be catabolized through the PP

pathway (i.e., higher [NADPD]).


14
13 W
12
11
10
_ 10
Sg o OD0o= 10, 10ml
9

+ OD =7, 10ml
7 ~"
5M 60 D0=11oo= 10mi

4 0o
3
2
0 2 4 6 8 10 12 14 16 18 20 22 24 26
Time (h)

Figure 2-25. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
wild-type strain (VVT) using Glc-6,6 D2, under aerobic (open squares and
closed circles) and anaerobic conditions (open circles). (Arrow indicates the
time at which the air supply was turned off)

In contrast to what was observed for the Apgi strain, the contribution of labeled

Glc-1D or Glc-3D in EHB for Azwf strain was negligible since the route through the PP

pathway was disrupted (Figure 2-29). However, the contribution was maximized in this

strain when Glc-6,6D2 was used under aerobic conditions. (Figure 2-30).





















6

S4
Y,


0 1


0 0 0
00
o o


5 6 7


Time (h)

Figure 2-26. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by
Apgi strain using Glc-6,6 D2, under aerobic (opened symbols) and anaerobic
conditions (closed circles).


*3
0A






- ------------ --A ---------------
a o u



- ------l---I I II----


* OD00=5.7, 10ml
o OD60=7.6, 10ml


A OD00=7.8,
o OD6o0=8.0,


0 1


4 5 6
Time (h)


Figure 2-27. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by the
Apgi using Glc-1D, under aerobic (opened symbols) and anaerobic conditions
(closed circles).


* 0D600=8.2, lOml
* ODo=8.2, 1Oml



o OD6o=7.8, 10ml


100m/
10mi
















o OD6oo=8.0, 10ml


0 1 2 3 4
Time (h)


5 6 7


Figure 2-28. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by
Apgi strain using Glc-3D.

7 1


S A A A
A o0


0 1 2


3
Time


5 6 7


Figure 2-29. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by
Azwf strain using Glc-3D, or Glc-1D under anaerobic (opened circles) and
aerobic conditions in 10 ml scale, OD600=7.0-9.0.


()O


* Gic-ld, aerobically

o GIc-ld, anaerobically

* Gic-ld, aerobically

* GIc-3d, aerobically










24

21
-18
- 15

+1 2

9
-6

3
n


* 0






o 0


0 1 2 3 4 5 6 7
Time (h)

Figure 2-30. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by
Azwf strain using Glc-6,6 D2, under aerobic (closed circles) and anaerobic
conditions (opened circles) in 10 ml scale, OD600=8.0.

The results for the obtained labeled EHB were reversed when an Aicd strain was

used instead for a Azwf strain. Maximum deuterium content in EHB was detected when

Glc-1-D was used (Figure 2-31) and minimum when Glc-6,6-D2 was administrated

(Figure 2-32).


0
0







A A A
&-- a -I


A GIc-ld, anaerobically
o GIc-ld, aerobically

Glc-ld, aerobically

0 GIc- d, aerobically


0 1 2 3 4 5 6 7
Time (h)

Figure 2-31. Fractional abundance (%) of 72 m/z fragment (A72) of EHB produced by
Aicd strain using Glc-1D, under aerobic (opened circles) and anaerobic
conditions (closed triangles) in 10 ml scale, OD600=7.0-8.0.


* Glc-6,6 D2, aerobically


o Gic-6,6 D2, anaerobically


I W