LIBRARY
Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

Syntheses and Downstream Purification of 1,2,4—Butanetriol

presented by

Man Kit Lau

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Chemistry

 

 

 

/ Major Professor’s Signature

7/2 3/ C» 1
Date

 

MSU is an affinnative-action, equal—opportunity employer

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE ' DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProj/Aoc8Pres/CIRCIDateDue.indd

 

SYNTHESES AND DOWNSTREAM PURIFICATION OF 1,2,4—BUTANETRIOL
By

Man Kit Lau

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

2007

ABSTRACT
SYNTHESES AND DOWNSTREAM PURIFICATION OF 1,2,4-BUTANETRIOL
By

Man Kit Lau

An extensive application of the thermally stable, high energetic material 1,2,4-
butanetriol trinitrate, is hindered by the lack of an economic route to synthesize its
precursor, 1,2,4-butanetriol. Our goal is to design an efficient synthesis of 1,2,4-
butanetriol using low-cost, renewable and abundantly available starting material. The
application of cutting-edge molecular biology offers opportunities to create 1,2,4-
butanetn'ol biocatalytic routes that can ultimately be scaled-up to commercially viable
processes. 2-Deoxyribose-5-phosphate aldolase condenses acetaldehyde and
glycoaldehyde into 3,4-dihydroxy-D-butanal. The success in coupling this enzymatic
reaction with a Ru/C catalyzed hydrogenation yielded D-l,2,4-butanetriol in 30% yield.
This chemoenzymatic synthesis represents the shortest route to synthesize a single 1,2,4-
butanetriol enantiomer under relatively mild reaction conditions using inexpensive
starting materials.

As an alternative to enzymatic processes, a large portion of this thesis was
devoted to explore and improve fermentation processes using E. coli microbial catalysts
that in a single step converts renewable carbohydrates into 1,2,4-butanetn'ol. To explore
novel microbial alternatives, a biosynthesis was created to synthesize 1,2,4-butanetriol in
a two-step conversion starting from erythritol. However, the identified erythritol

dehydratase activity was not sufficient enough to elaborate this biosynthesis into a

microbial process. Earlier research efforts in the Frost group created a single E. coli
microbe WNl3/pWN7.126B that synthesized D-l,2,4—butanetriol from D-xylose at a
concentration of 10.5 g/L when cultivating under an improved fermentation conditions.
Attempts to identify a better 3-deoxy-glycero-pentulosonate decarboxylase did not lead to
significantly improved alternatives. Approaches to improve P. putida mdlC gene
encoding benzoylformate decarboxylase expression resulted in metabolic burden. E. coli
ath encoding alcohol dehydrogenase was identified to be involved in the conversion of
3,4-dihydroxy-D-butanal into D-1,2,4-butanetriol. E. coli microbe KIT10/pWN7.126B
synthesized only 6.5 g/L D-1,2,4-butanetriol. A second generation D-1,2,4-butanetriol
biocatalyst KIT18/pWN7.126B was created by overexpressing ath chromosomally with
two Z-keto acid dehydrogenase (YiaE and chW) knockouts. More recently, cultivation
of E. coli KIT18/pWN7.126B yielded D-1,2,4-butanetriol at a concentration of 18 g/L in
50% (mol/mol) yield from D-xylose.

A downstream purification method for D-1,2,4-butanetriol was developed using
Karr reciprocating column. Using 2—butanol as solvent, D-1,2,4-butanctriol was
successfully recovered in 74% from the fermentation broth. Salt-free D-l,2,4-butanetriol
with >99% purity was obtained by Karr column extraction and negative adsorption using

Dowex 1X8 (OH') followed by Dowex 50 (H+) ion-exchange resins.

Copyright by
hmmnkfitLau

2007

To my parents

For their love and support

ACKNOWLEDGMENTS

This dissertation would not have been possible without the help of so many people.
I would first like to express my warmest appreciation to Professor John W. Frost. His
enthusiasm, integral view of science, and constructive criticism have all contributed a
great deal to the success of my research. I would also like to thank the members of my
graduate committee, Prof. Chris C. K. Chang, Prof. Babak Borhan, and Prof. Kevin D.
Walker for their interest and intellectual input during the preparation of this thesis.

I am deeply indebted to Dr. Karen M. Draths for her guidance and invaluable
inputs to my research. I wish to thank Carolyn Wemple for resolving most of the
bureaucratic issues I have encountered throughout my studious years at MSU. There are
no better colleagues and fi'iends than Justas Jancauskas and Dr. Jihane Achkar, whom I
must thank for their devoted friendship, encouragement and support. Brad Cox who
struggled with my writing style certainly helped to put this thesis together. My gratitude
is given to other Frost group members, past and present, including Dr. Vu Bui, Dr.
Stephen Cheley, Dr. Sigal Lechno-Yossef, Dr. Wensheng Li, Dr. Jiantao Guo, Dr.
Mapitso Molefe, Dr. Wei Niu, Dr. Ningqing Ran, Dr. Heather Stueben, Dr. Dongming Xie,
Dr. Jinsong Yang Dr. Jian Yi and Xiaofei Jia for providing a pleasant research
environment. I am also grateful to my dear friends Cangel, Fei, Harrison, Sewa, Sing,
Kostas, Thalia, Nineta, Chrysoula and Jennie for their friendship and help. The jokes and
time we shared will never be forgotten. Finally, the love and support I have been given by
my parents Joseph and Doris has been overwhelming. Even during the toughest time, I
have always felt secure knowing they were there for me. Without all of you I would never

have found my way.

vi

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. xii
LIST OF FIGURES ........................................................................................................... xiv
LIST OF ABBREVIATIONS ............................................................................................ xix
CHAPTER ONE ................................................................................................................... 1
Introduction ........................................................................................................... 1
Metabolic pathway engineering for microbial catalysts ................................... 4

Shikimic acid and quinic acid ....................................................................... 4

Indigo dye ...................................................................................................... 7

l ,3-Propanediol ............................................................................................. 8

The significance of 1,2,4-butanetn'ol and 1,2,4-butanetriol trinitrate ............... 9

Chemical syntheses of 1,2,4-butanetriol ......................................................... 12

Creation of a 1,2,4-butanetriol biosynthetic pathway ..................................... 14

A single microbe for D-l,2,4-butanetn'ol synthesis .......................................... 16

Important chemicals derived from the microbial 1,2,4-butanetriol synthesis. 18
3,4-Dihydroxy-D-butanoic acid ................................................................... 18

3 ,4-Dihydroxy-threo-L-norvaline ............................................................... 1 9

REFERENCE ...................................................................................................................... 20
CHAPTER TWO ................................................................................................................ 23
Chemo-Enzymatic Synthesis of D-1,2,4-Butanetriol Using 2-Deoxyribose-5-
Phosphate Aldolase ............................................................................................. 23
Background ..................................................................................................... 23

Synthesis of 3,4-dihydroxy-D-butanal using 2-deoxyribose-5-phosphate

aldolase ........................................................................................................... 30

Protein purification of 2-deoxyribose-5-phosphate aldolase ...................... 30

Chemical synthesis of 3,4-dihydroxy-D-butanal ......................................... 33

Enzymatic synthesis of 3,4-dihydroxy-D-butanal ....................................... 34

Catalytic hydrogenation of racemic 3,4-dihydroxybutanal ............................. 36
Chemo-enzymatic synthesis of D-1,2,4-butanetriol ........................................ 38

Discussion and future work ............................................................................. 40
REFERENCE ...................................................................................................................... 44
CHAPTER THREE ............................................................................................................ 46
Improving Microbial Synthesis of 1,2,4-Butanetriol .......................................... 46
Overview ......................................................................................................... 46

Artificial biosynthesis of 1,2,4-butanetriol from erythritol ............................. 48
Background ................................................................................................. 48

Identification of erythritol dehydratase candidates ..................................... 52

Plasmid Constructions ................................................................................. 55

vii

Enzyme assays for in vitro erythritol dehydratase activity ......................... 63

1,2,4-Butanetn'ol biosynthesis using erythritol as substrate ........................ 65

In search of 3-deoxy-glycero-pentulosonate decarboxylase genes for the
microbial synthesis of 1,2,4-butanetriol from pentoses .................................. 72
Background ................................................................................................. 72
Cloning and Screening for 3-deoxy-glycer0-pentulosonate decarboxylase
candidates .................................................................................................... 76
Site-directed Mutagenesis of benzoylformate and pyruvate decarboxylase
gene ............................................................................................................. 85
Enzyme activities for 3-deoxy-glycero-pentulosonate decarboxylase
candidates .................................................................................................... 88
Codon usage ................................................................................................ 92
Amino acid and nucleotide sequence identity ............................................. 92
Identification of novel benzoylformate decarboxylases ............................. 93
Directed evolution of 3-deoxy-glycero-pentulosonate decarboxylase for
microbial 1,2,4-butanetriol synthesis .............................................................. 95
Background ................................................................................................. 95
Subcloning 3-deoxy-glycer0-pentulosonate decarboxylase encoded genes98
Chimera library construction for ssDNA family shuffling ....................... 102
Chimera library construction for dsDNA family shuffling ....................... 103
Chimera library screening for 3-deoxy-glycero-pentulosonate
decarboxylase ............................................................................................ 106
Bacteriophage T7 promoter and E. coli chaperones for deC expression....112
Background ............................................................................................... 1 12
Expression of plasmid-home mdlC gene under phage T7 promoter ......... 113
E. coli chaperonins GroES-GroEL for deC gene expression .................. 120
Discovery of alcohol dehydrogenases for D-1,2,4-butanetriol biosynthesis. 127
Background ............................................................................................... 127
Cloning and screening for D-l ,2,4-butanetriol dehydrogenase candidates
................................................................................................................... 127
Chromosomal ath gene inactivation ...................................................... 134
Plasmid-based expression of ath gene ................................................... 137
Chromosomal expression of E. coli ath ................................................. 147
Construction of the second generation D-1,2,4-butanetriol synthesizing E.
col i microbe ............................................................................................... 1 56
Future work ................................................................................................... 158
Directed Evolution of 3-deoxy-D-glycero-pentulosonate decarboxylase .158
Chemo-microbial synthesis of (S)-3-hydroxy-y-butyrolactone ................. 161
Enzyme activity studies for E. coli ath expressing microbes ................ 164
REFERENCE .................................................................................................................... 167
CHAPTER FOUR ............................................................................................................. 173
Downstream Purification of 1,2,4-Butanetriol ................................................. 173
Background ................................................................................................... 1 73
Solvent candidates screening using continuous extraction method .............. 179

viii

Residual protein removal by membrane ultra-filtration ............................ 179

Liquid-liquid extraction using a single solvent system ............................. 181
Liquid-liquid extraction using a solvent-blend system ............................. 184
Development of a scalable extraction process using a bench-top Karr column
....................................................................................................................... 1 87
Bench-top Karr column unit setup ............................................................ 187
Distribution coefficient for solvent candidates ......................................... 189
Start-up conditions for Karr column extractions ....................................... 190
Preliminary extraction attempts using Karr column ................................. 192
Further optimization of Karr column extraction ....................................... 194
Final step toward purified D-1,2,4-butanetriol .......................................... 196
Alternative method for D-1,2,4-butanetriol purification ........................... 199
Future plan .................................................................................................... 201
REFERENCE .................................................................................................................... 203
CHAPTER FIVE .............................................................................................................. 204
Experimental ..................................................................................................... 204
General chemistry ......................................................................................... 204
Reagents and solvents ................................................................................... 204
Chromatography ........................................................................................... 205
SpectrOSCOpic measurements ........................................................................ 206
Microbial strains and plasmids ..................................................................... 206
Storage of microbial strains and plasmids .................................................... 209
Culture medium ............................................................................................ 210
General fed-batch fermentation conditions ................................................... 212
Analysis of fermentation broth ..................................................................... 213
Genetic manipulations .................................................................................. 214
Large scale purification of plasmid DNA ..................................................... 216
Small scale purification of plasmid DNA ..................................................... 217
Restriction enzyme digestion of DNA .......................................................... 218
Determination of DNA concentration ........................................................... 219
Agarose gel electrophoresis .......................................................................... 219
Isolation of DNA from agarose ..................................................................... 219
Treatment of DNA with Klenow fragment ................................................... 220
Treatment of vector DNA with calf intestinal alkaline phosphatase ............ 221
Ligation of DNA ........................................................................................... 221
Preparation and transformation of competent E. coli cells ........................... 222
Purification of E. coli genomic DNA ........................................................... 224
P1 phage-mediated transduction ................................................................... 226
Enzyme assays .............................................................................................. 228
2-Deoxyribose-5-phosphate aldolase assay .................................................. 228
Glycerol and diol dehydratase assay ............................................................. 228
2-Keto acid decarboxylase assay .................................................................. 229
Alcohol dehydrogenase assay ....................................................................... 230
Protein SDS-PAGE analysis ......................................................................... 231

ix

Chapter 2 ........................................................................................................... 232

Purification of 2-deoxyribose-5-phosphate aldolase ..................................... 232
Catalytic hydrogenation of 3,4-dihydroxy-D-butanal ................................... 234
Chapter 3 ........................................................................................................... 235
Artificial biosynthesis of 1,2,4-butanetriol from erythritol ........................... 235
Plasmid pML5.176 .................................................................................... 235
Plasmid pML5.179 .................................................................................... 235
Plasmid pML5.200 .................................................................................... 235
Plasmid pML5.226 .................................................................................... 236
Plasmid pWN5.22OA ................................................................................ 236
Plasmid pWN5.260A ................................................................................ 236
Plasmid pWN5.258A ................................................................................ 237
Anaerobic cultivation of Klebsiella and Lactobacillus strains ................. 237
In vitro dehydratase reaction of erythritol ................................................. 238
In search of 3-deoxy-glycero-pentulosonate decarboxylase genes for the
microbial sysnthesis of 1,2,4-butanetn'ol from pentoses .................................. 238
Plasmid pML2.118 .................................................................................... 238
Plasmid pML2.123 .................................................................................... 238
Plasmid pML2.162 .................................................................................... 239
Plasmid pML2.208 .................................................................................... 239
Plasmid pML2.214 .................................................................................... 239
Plasmid pML2.256 .................................................................................... 240
Plasmid pML2.286 .................................................................................... 240
Plasmid pML3.04O .................................................................................... 240
Plasmid pML3.054 .................................................................................... 241
Plasmid pML3.062 .................................................................................... 242
Plasmid pML3.084 .................................................................................... 242
Plasmid pML3.086 .................................................................................... 243
Plasmid pML3.104 .................................................................................... 243
Directed evolution of 3-deoxy-glycero-pentulosonate decarboxylase for
microbial 1,2,4-butanetriol synthesis ............................................................ 244
Plasmid pML3.200 .................................................................................... 244
Plasmid pML3.224 .................................................................................... 244
Plasmid pML3.225 .................................................................................... 245
Plasmid pML3.226 .................................................................................... 245
Chimera library generation for ssDNA family shuffling .......................... 245
Chimera library generation for dsDNA family shuffling .......................... 247
Library screening for 3-deoxy-glycer0-pentulosonate decarboxylase ...... 248
Bacteriophage T7 promoter and E. coli chaperones for deC expression....250
E. coli KIT2 ............................................................................................... 250
E. coli KIT15 ............................................................................................. 251
Plasmid pML7. 128 .................................................................................... 251
Plasmid pML7.135 .................................................................................... 251
Plasmid pML7.166 .................................................................................... 252
Plasmid pML7.175 .................................................................................... 252
Plasmid pML7.18O .................................................................................... 252

Plasmid pML7.202 .................................................................................... 253
Discovery of alcohol dehydrogenases for D-1,2,4-butanetriol biosynthesis.253

E. coli KITl ............................................................................................... 253
E. coli KIT3 ............................................................................................... 254
E. coli KIT4 ............................................................................................... 254
E. coli KITS ............................................................................................... 255
E. coli KIT6 ............................................................................................... 256
E. coli KIT7 ............................................................................................... 256
E. coli KIT8 ............................................................................................... 256
E. coli KIT9 ............................................................................................... 257
E. coli KITIO ............................................................................................. 257
E. coli KIT16 ............................................................................................. 258
E. coli KIT17 ............................................................................................. 258
E. coli KIT18 ............................................................................................. 259
Plasmid pML5.179 .................................................................................... 259
Plasmid pML6.09O .................................................................................... 259
Plasmid pML6. 128 .................................................................................... 260
Plasmid pML6.133 .................................................................................... 260
Plasmid pML6.135 .................................................................................... 260
Plasmid pML6.166 .................................................................................... 260
Plasmid pML6.168 .................................................................................... 261
Plasmid pML6. 185 .................................................................................... 261
Plasmid pML6.195 .................................................................................... 261
Plasmid pML6.255 .................................................................................... 262
Plasmid pML6.259 .................................................................................... 262
Plasmid pML6.261 .................................................................................... 262
Plasmid pML6.263 .................................................................................... 263
Plasmid pML6.272 .................................................................................... 263
Plasmid pML7.042 .................................................................................... 263
REFERENCE .................................................................................................................... 265

xi

Table 1.
Table 2.

Table 3.

Table 4.
Table 5.
Table 6.
Table 7.
Table 8.
Table 9.
Table 10.
Table 11.
Table 12.
Table 13.
Table 14.
Table 15.
Table 16.
Table 17.
Table 18.
Table 19.
Table 20.
Table 21.

Table 22.

LIST OF TABLES

Purification table of 2-deoxyribose-5-phosphate aldolase ................................... 32
Purification table of 2-deoxyribose-5-phosphate aldolase ................................... 33
Influence of 2-deoxyribose-5-phosphate aldolase (DERA) purity on 3,4—
dihydroxy-D-butanal yield. .................................................................................. 35
Catalytic hydrogenation of authentic 3,4-dihydroxybutanal. .............................. 38
Glycerol and diol dehydratase candidates ........................................................... 54
In vitro dehydratase activities. ............................................................................. 64
1,3-Pr0panediol oxidoreductase activities. .......................................................... 68
In vitro enzymatic reactions using erythritol as substrate .................................... 69
3-Deoxy-glycero-pentulosonate decarboxylase candidates. ................................ 85
Forward primers used to generate mutations. .................................................... 88
Activity table for wild-type pJF118EH clones. ................................................. 90
Activity table for benzoylformate and pyruvate decarboxylase mutants. ......... 9O
Amino acid composition of the gene candidates. .............................................. 91
Rare codon analysis of the gene candidates ....................................................... 91
Nucleotide and amino acid (in parentheses) sequence identity. ........................ 93
Data reproducibility using in viva GC screening assay. .................................. 108
Test for false-negative results. ......................................................................... 108
Test for mutants with growth issues. ............................................................... 108
Shake flask experiments for improved decarboxylase activities. .................... 109
E. coli alcohol dehydrogenase candidates. ...................................................... 134
E. coli strains for chromosomal ath inactivation study. ............................... 135
Other alcohol dehydrogenase candidates ......................................................... 137

xii

Table 23

Table 24.

Table 25.

Table 26.
Table 27.
Table 28.
Table 29.
Table 30.
Table 31.
Table 32.
Table 33.
Table 34.
Table 35.
Table 36.
Table 37.

Table 38.

E. coli strains for chromosomal ath expression study. ................................. 149
Creation of the second generation D-1,2,4-butanetriol synthesizing E. coli....158

Preliminary results on the cultivation of E. coli KIT18/pWN7.126B under

fermentor-controlled conditions ...................................................................... 158
Oxidative decarboxylation of 3-deoxy-D-glycero-pentulosonic acid. ............. 163
E. coli strains with ath gene expression. ...................................................... 166
Methods for residual protein removal from fermentation broth. ..................... 181
Single solvent candidates. ................................................................................ 182
Summary of single solvent screening experiments .......................................... 183
Liquid-liquid extraction using 2-butanone/ethyl acetate solvent-blend. ......... 185
Liquid-liquid extraction using ethanol/ethyl acetate solvent blend. ................ 186
Distribution coefficients for various solvent candidates .................................. 189
Preliminary Karr column extraction studies. ................................................... 194
Further optimization for Karr column extraction. ........................................... 195
Dowex 1X8 (0H) for extraction-free purification. ......................................... 201
D-l,2,4-Butanetriol recovery from Dowex 1X8 (OH'). ................................... 201
Microbial strains and plasmids. ....................................................................... 206

xiii

Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.
Figure 9.
Figure 10.

Figure 11.

Figure 12.
Figure 13.
Figure 14.
Figure 15.

Figure 16.

Figure 17.
Figure 18.

Figure 19.

Figure 20.

Figure 21.

LIST OF FIGURES

The shikimate pathway and biosynthesis of quinic acid. ..................................... 5
Synthesis of hydroquinone from glucose. ............................................................ 7
Chemical and microbial syntheses of indigo. ....................................................... 8
Schematic of the reaction network from D-glucose to 1,3-propanediol. ............ 10
Industrial synthesis of 1,2,4-butanetriol trinitrate. ............................................. 11
Syntheses of 1,2,4-butanetriol. ........................................................................... 12
Routes to 1,2,4-butanetriol previously developed in the Frost lab. .................... 14
Microbial syntheses of D- and L-l,2,4-butanetriol. ............................................ 15
Native E. coli D-xylonic acid catabolic pathway ................................................ 17
Microbial synthesis of D-l ,2,4-butanetriol using WN13/pWN7.126B. ........... 17
Drugs synthesized from 3,4-dihydroxy-D-butanoic acid and 4,5-dihydroxy-
threo-L-norvaline. ............................................................................................ l 9
In vivo reaction catalyzed by 2-deoxyribose-5-phosphate aldolase. ................ 23
2-Deoxyribose-5-phosphate aldolase catalyzed tandem aldol reaction ............ 24
Retro-synthesis of atorvastatin (Lipitor). ......................................................... 25
Catalytic hydrogenation of malic acid. ............................................................. 27
Proposed chemo-enzymatic reaction using 2-deoxyribose-5-phosphate
aldolase. ........................................................................................................... 28
Industrial syntheses of acetaldehyde and glycolaldehyde. ............................... 29
Construction of plasmid pVH17 ....................................................................... 31
SDS-PAGE of 2-deoxyribose-5-phosphate aldolase purified from E. coli
DH50I. .............................................................................................................. 32
Chemical synthesis of racemic 3,4-dihydroxybutanal ...................................... 34
NaBH4 reaction for 3,4-dihydroxy-D-butanal quantification ........................... 34

xiv

Figure 22.

Figure 23.

Figure 24.
Figure 25 .
Figure 26.
Figure 27.
Figure 28.
Figure 29.
Figure 30.
Figure 31.
Figure 32.
Figure 33.
Figure 34.
Figure 35.
Figure 36.
Figure 37.
Figure 38.
Figure 39.

Figure 40.

Figure 41.
Figure 42.
Figure 43.

Figure 44.

Reaction parameters for the chemo-enzymatic synthesis of

D—1,2,4-butanetriol. .......................................................................................... 3 9
Proposed artificial biosynthetic pathway for 1,2,4-butanetriol from

erythritol. ......................................................................................................... 46
Anaerobic utilization of glycerol. ..................................................................... 49
Anaerobic utilization of 1,2-propanediol. ........................................................ 50
Plasmid maps of pCS120, pF L1 and pXY39. .................................................. 56
Construction of plasmid pML5.176. ................................................................ 58
Constuction of plasmid pML5.200. .................................................................. 59
Construction of plasmid pWN5.220A. ............................................................. 60
Construcion of plasmid pWN5.260A. .............................................................. 61
Construction of plasmid pWN5.258A. ............................................................. 62
In vitro dehydratase assays using different substrates. ..................................... 64
Construction of plasmid pML5.179. ................................................................ 66
Construction of plasmid pML5.226. ................................................................ 67
GC chromatogram for reaction 1 (control) ....................................................... 69
GC chromatograms for reaction 2. ................................................................... 70
GC chromato grams for reaction 4. ................................................................... 71
Restriction enzyme map of plasmids ................................................................ 72
Mandelate pathway in P. putida. ...................................................................... 74
Reaction mechanism for the benzoylformate decarboxylase-catalyzed
decarboxylation of benzoylformate. ................................................................ 74
Construction of plasmid pML2.118. ................................................................ 77
Construction of plasmid pML2.123. ................................................................ 78
Construction of plasmid pML2.162. ................................................................ 79
Construction of plasmid pML3.040. ................................................................ 80

XV

 

Figure 45.
Figure 46.
Figure 47.
Figure 48.
Figure 49.
Figure 50.
Figure 51.
Figure 52.
Figure 53.
Figure 54.
Figure 55.

Figure 56.

Figure 57.

Figure 58.

Figure 59.
Figure 60.
Figure 61.

Figure 62.

Figure 63.
Figure 64.
Figure 65.
Figure 66.

Figure 67.

Construction of plasmid pML2.208. ................................................................ 81
Construction of plasmid pML2.214. ................................................................ 82
Construction of plasmid pML2.256. ................................................................ 83
Construction of plasmid pML2.286. ................................................................ 84
QuikChange Site directed mutagenesis. ........................................................... 87
Coupled enzyme assay for decarboxylase activities. ....................................... 89
Multiple sequence alignment using ClustalW algorithm. ................................ 94
Schematic diagram for single-stranded DNA family shuffling ........................ 97
Construction of plasmid pML3.224. ................................................................ 99
Construction of plasmid pML3.225. .............................................................. 100
Construction of plasmid pML5.226. .............................................................. 101
Sequencing results of 20 ramdomly chosen clones from ssDNA family
shuffling ......................................................................................................... 103
DNA agarose gels for dsDNA family shuffling. ............................................ 104
Sequencing results of 22 ramdomly chosen clones fi'om dsDNA family
shuffling ......................................................................................................... 105
Sequence alignment of selected mutants. ....................................................... 111
Construction of plasmid pML7.128. .............................................................. 114
Construction of plasmid pML7. 135. .............................................................. 115
Biosynthesis of 3-deoxy-D-glycero-pentulosonic acid under fermentor-
controlled conditions. .................................................................................... 1 19
Construction of plasmid pML7.166. .............................................................. 121
Construction of plasmid pML7.175. .............................................................. 122
Construction of plasmid pML7.180. .............................................................. 123
Construction of plasmid pML7.202. .............................................................. 124

Cultivation of chaperonin groEL-groES constructs under fermentor-controlled
conditions. ..................................................................................................... 126

xvi

 

Figure 68. Construction of plasmid pML6.166. .............................................................. 129
Figure 69. Construction of plasmid pML6.168. .............................................................. 130
Figure 70. Construction of plasmid pML6.259. .............................................................. 131
Figure 71. Construction of plasmid pML6.261. .............................................................. 132
Figure 72. Construction of plasmid pML2.263. .............................................................. 133
Figure 73. Cultivation of E. coli strains with ath gene inactivation under fermentor-
controlled conditions. .................................................................................... 136
Figure 74. Construction of plasmid pML5.179. .............................................................. 138
Figure 75. Construction of plasmid pML6.185. ............................................................... 140
Figure 76. Construction of plasmid pML6.195. .............................................................. 141
Figure 77. Construction of plasmid pML6.133. .............................................................. 142
Figure 78. Construction of plasmid pML6.090. .............................................................. 143
Figure 79. Construction of plasmid pML6.128. .............................................................. 144
Figure 80. Construction of plasmid pML6.135. .............................................................. 145
Figure 81. Cultivation of E. coli constructs expressing plasmid-bome alcohol
dehydrogenases under ferrnentor-controlled conditions. .............................. 146
Figure 82. Strategy to integrate xdh-ath-Pm gene cassette into E. coli W3110 ........... 149
Figure 83. Construction of plasmid pML6.255. .............................................................. 150
Figure 84. Construction of plasmid pML6.272. .............................................................. 151
Figure 85. Construction of plasmid pML7.042. .............................................................. 153
Figure 86. Cultivation of E. coli constructs having a chromosomal Pmc-ath gene
insertion under fermentor-controlled conditions ........................................... 154
Figure 87. Cultivation of E. coli constructs having a chromosomal T5 promoter insertion
under fermentor-controlled conditions .......................................................... 155
Figure 88. Proposed chemo-microbial synthesis of (S)-3-hydroxy-y-butyrolactone. ...... 162
Figure 89. Enzyme activity time-course for microbial syntheses of D-1,2,4-butanetriol

under fermentor-controlled conditions .......................................................... 166

xvii

Figure 90. Single-stage liquid-liquid extractors ............................................................... 176
Figure 91. Agitated counter-current extraction columns. ................................................ 178
Figure 92. Glass continuous liquid-liquid extractor (Sigma-Aldrich). ............................ 179
Figure 93. D-1,2,4-Butanetriol purified by 50% 2-butanone in ethyl acetate extraction. 185
Figure 94. D-1,2,4-Butanetriol purified by 10% ethanol in ethyl acetate extraction. ...... 186
Figure 95. Benchtop Karr column in action ..................................................................... 187
Figure 96. Kremser equation ............................................................................................ 197
Figure 97. Kremser type plot. .......................................................................................... 197
Figure 98. GC traces for Karr column extraction. ........................................................... 198
Figure 99. Large scale purification of E. coli WN13/pWN7.126B fermentation broth. .199
Figure 100. Cyclic acetal formation of 1,2,4-butanetriol ................................................. 202

xviii

 

ADH
Ap
ATP
bp
BSTFA
BT
CIAP
Cm
DEAE
DERA
DHBA
DGP
DHN
DNA
D.O.
DPA
Flp
FRT
G3P
GAP

GC

LIST OF ABBREVIATIONS

alcohol dehydrogenase

ampicillin

adenosine triphosphate

base pair
N,0-bis(trimethylsilyl)trifluoroacetamide
1 ,2,4-butaneuiol

calf intestinal alkaline phosphatase
chloramphenicol

diethylarninoethyl
2-deoxyribose-5-phosphate aldolase
3,4-dihydroxybutanoic acid
3-deoxy-glycero-pentulosonic acid
4,5-dihydroxy-thre0-norvaline
3-deoxyribonucleic acid

dissolved oxygen
3-deoxy-glycero-pentanoic acid
flipase

flipase recognition target sequence
sn-glycerol 3-phosphate
D-glyceraldehyde 3-phosphate

gas chromatography

xix

GRAS

His
HPLC

IPTG

kg

LB

mg

mL

pL
mM
min
NAD
NADH
NADP
NADPH
NMR
OD
ORF
PEG

PEP

generally regard as safe

hour

L-histidine

high pressure liquid chromatography

isopropyl B-D-thiogalactopyranoside

kanamycin

kilogram

Luria-Bertani broth

molar

milligram

milliliter

microliter

millimolar

minute

nicotinamide adenine dinucleotide, oxidized form
nicotinamide adenine dinucleotide, reduced form
nicotinamide adenine dinucleotide phosphate, oxidized form
nicotinamide adenine dinucleotide phosphate, reduced form
nuclear magnetic resonance

optical density

open reading frame

polyethylene glycol

phosphoenolpyruvate

XX

PCR

ppm

psi

PTS

rpm

11

SDS
SDS-PAGE
TCA cycle
TSP

UV

polymerase chain reaction

parts per million

pounds per square inch

phosphoenolpyruvate:carbohydrate phosphotransferase systems
revolutions per minute

room temperature

sodium dodecyl sulfate

SDS polyacrylamide gel electrophoresis

tricarboxylic acid cycle

sodium 3-(trimethylsilyl)propionic-2,2,3,3-d4

ultraviolet

xxi

CHAPTER ONE

Introduction

Industrial fermentation processes have existed for thousands of years to meet
various human needs. Traditionally, most were accidentally discovered by mankind and
incorporated into daily life to produce foods and beverages such as bread, wine, beer,
vinegar and cheese.1 The idea behind this technology was not clearly understood until the
nineteenth century when a French chemist named Louis Pasteur first connected
microorganisms to fermentation. His pioneering work that included the discovery of
yeast, lactic acid bacillus, development of ‘germ theory’, and the invention of the
Pasteurization process laid the foundation for modern biotechnology.2 Modern industrial
biocatalysis began when the first large scale fermentations were developed for the
manufacturing of solvents, organic acids, vitamins etc. In the 1940’s, antibiotic
fermentations came into existence. Sir Alexander Fleming discovered microbe-
synthesized penicillin, and its scale-up fermentation process was developed jointly by
Merck & Co. and Princeton University.3 In 1972, the first recombinant DNA plasmid was
made in the laboratories of Stanford University and University of California at San
Francisco.3 In 1975, first DNA sequencing technique was established.4 In 1988, Kary
Mullis published the first paper on the application of polymerase chain reaction (PCR).5
Since then, molecular biology has revolutionized traditional microbiology and established

a new global biotechnology industry.

A prominent feature by which enzymes are favorably distinguished from chemical
catalysts is their specificity. While enzymes are used to increase the regioselectivity of a
reaction, their greatest advantage lies in the ability to differentiate between enantiomeric
substrates, resulting in enantiomerically pure products. In addition, the high specificity of
enzymes do not require highly purified substrate stream. Another advantage of
biocatalysis over a chemical route is its mild, near ambient reaction conditions of
temperature and pH. The preference of an aqueous reaction medium is often regarded as
an important benefit to develop an environmentally benign manufacturing process. The
possibility of using simple and inexpensive carbohydrates such as D-glucose and D-xylose
as starting material makes this biocatalytic approach economically attractive and
environmentally acceptable.

In terms of applications, enzymes have long been known to compete well with
chemical methods for resolution. Ibuprofen, phenylethylamine and acrylamide are
commonly cited in literature as examples that are synthesized using enzyme-based
processes.6 With the advances in technology, such as enzyme immobilization in amino
acids manufacturing, 7 ’ 8 new methods for accessing biodiversity to identify novel
enzymes}9 and more recently, by directed evolution strategies to optimize enzymes for

existing and new applications,10

new cutting-edge applications continue to be discovered.
For instance, a fed-batch process of L-DOPA using Erwinia herbicola has been developed
(Ajinomoto Co. Ltd.) to give a final product titer of 110 g/L.ll This single step reaction
utilizes catechol, pyruvic acid and ammonia as starting materials to afford L-DOPA,

producing water as the only byproduct in a short production cycle of 3 days. This process

provides a relatively cheaper route to L-DOPA, which is a metabolic precursor of

dopamine, an important drug for the treatment of Parkinson’s disease. The other example
is the microbial synthesis of Shikimic acid developed in the Frost group.12 Shikimic acid
is the starting material for the manufacture of the anti-influenza drug Tamiflu.
Traditionally, Shikimic acid was extracted from star anise seeds. The Frost group has
optimized an alternative process using a microbe to synthesize 60 g/L concentration of
Shikimic acid from glucose.12 These examples show that biocatalysis has been evolving
into a highly promising area of research, and the new generation of bioprocesses is
becoming more competitive with conventional routes.

This dissertation presents research results for the syntheses of D-1,2,4-butanetriol
using different methodologies. In chapter 2 of this thesis, a biocatalytic alternative to the
industrially employed chemical synthesis of 1,2,4-butanetriol is described. A chemo-
enzymatic approach utilized E. coli deoC encoding deoxyribose-S-phosphate aldolase
(DERA) to condense acetaldehyde and glycoaldehyde into 3,4-dihydroxy-D-butanal ‘3
followed by catalytic hydrogenation to form D-1,2,4-butanetriol. In chapter 3,
development of a novel biosynthetic route to produce 1,2,4-butanetriol from erythritol will
be discussed. Different approaches to improve the microbial synthesis of D-1,2,4-
butanetriol from D-xylose using genetic engineered E. coli WN13/pWN7.l26Bl4 will be
presented as well. In contrast to the chemo-enzymatic approach, whole cell microbial
synthesis exploits the metabolism of the microbe to supply necessary co-substrates and to
regenerate cofactors. Although genetic manipulation of the microbe is necessary, at the
end of the process, the organism produced can self-regenerate. In chapter 4, a downstream
purification of D-l,2,4-butanetriol will be presented. Recovery of polyols from the

complex and dilute fermentation broth is a challenging task due to relatively high boiling

points and inherently high hydrophilicity of related compounds. An efficient purification
protocol is therefore critical for the development of a commercially viable bioconversion
process for 1,2,4-butanetriol synthesis. To this end, a liquid-liquid extraction

methodology was developed using a bench-top Karr reciprocating plate extraction column.

Metabolic pathway engineering for microbial catalysts

By manipulating metabolic pathways in microorganisms, a wide range of natural
compounds, such as amino acids, organic acids, and nucleotides, can be produced.
Essential for the successful production of chemicals, however, will be the integration of
metabolic engineering with biochemical engineering. Successful industrially viable
bioprocess development will achieve improvements in several aspects such as higher
catalyst turnover, more efficient channeling of carbon and most importantly, higher

product yield. Three independent cases will be discussed as examples.

Shikimic acid and quinic acid

Shikimic acid is the starting material for the synthesis of the neuraminidase
inhibitor Tarniflu, which is an anti-influenza drug. Shikimic acid has traditionally been
extracted from star anise seeds. More recently, fermentor-controlled microbial synthesis
provides an alternative avenue to this molecule. Currently, these two routes supply the
global demand for Shikimic acid. The host strain used for microbial synthesis of Shikimic
acid is a genetically engineered E. coli that lacks shikimate kinase activity (Figure 1).”

The first reported shikimate-producing E. coli synthesized 27 g/L Shikimic acid in a 1 L

high density cell culture. After further genetic manipulation of the shikimate pathway and
optimization of fermentor-controlled culture conditions, the current process can produce

Shikimic acid at a concentration of 62 g/L in yield from glucose of 33% (mol/mol).

 

 

 

OH O
H203P°MH
OH
HO HQ HQ
D-erythrose 4-phosphate ’- COZH ‘ COZH ‘ COZH
AroF , o _Ar_os.. Ares. .
grog i OH 0 i OH H0“ , OH
OPOaHz ° H203PO OH 0H 0H
002H 3-deoxy-o-arabino- 3—dehydroquinic quinic
phosphoenolpyruvic acid heptulosonic acid acrd acud
7-phosphate
AroD
COzH C02H COzH
(it E1 £1
‘— ._—-_
Hzoapo“ i OH AWL no“ i OH 0 i OH
OH OH OH
Sh'k'mate 3‘Ph°sphate shikirnic 3—dehydroshikimic
acid acid
lAroA O O\/
COZH COZH multi-steps chemical
JL AroC ,iL synthesrs : O“ ""NH2
H203Po“ , o COzH i o COzH “NY
OH OH O
5—enolpyruvylshikimate chorismic acid '
3-phosphate TamIflu

Figure 1. The shikimate pathway and biosynthesis of quinic acid. Enzymes: DAHP
synthase (AroF, AroG, AroH); 3-dehydroquinate synthase (AroB); 3-dehydroquinate
dehydratase (AroD); shikimate dehydrogenase (AroE); shikimate kinase (AroK, AroL); 5-
enolpyruvoylshikimate 3-phosphate synthase (AroA); chorismate synthase (AroC).

Quinic acid is another shikimate pathway metabolite that is a useful natural
product. Quinic acid has been identified to be biologically active and can be isolated from
the bark of Uncaria tomentosa, which is better known as cat’s claw.15 Cat’s claw extract
is widely used in South America as a historic medicinal treatment for gastric, arthritis,
cancer and inflammatory conditions. Microbial synthesis of quinic acid using E. coli has
been accomplished by exploiting the catalytic promiscuity of shikimate dehydrogenase,
which in addition to catalyzing the reduction of 3-dehydroshikimate, can also catalyze the
reduction of 3-dehydroquinate to quinic acid. High-titer, high-yielding microbial
syntheses of quinic acid have subsequently been established based on overexpression of
shikimate dehydrogenase in 3-dehydroquinate-synthesizing E. coli strains and have led to
concentration of 54 g/L of quinic acid synthesized in 20% (mol/mol) yield from glucose.16
Quinic acid is also a key intermediate in the synthesis of hydroquinone from glucose.16
As a pseudocommodity chemical mainly used for photographic developing, the
approximate 4.5 X 107 kg/yr production of hydroquinone is derived from aniline, phenol,
and p-diisopropylbenzene, which are all manufactured from benzene.17 Quinic acid in
partially purified fermentation broth was readily converted into hydroquinone employing
HOCl or Ag3PO4/K28203 (Figure 2). More recently, quinic acid has been identified to
have intriguing activities associated with stimulation of the immune system. In rats

treated with the anticancer agent doxorubicin, cat’s claw extract has been observed to

accelerate the recovery of white blood cells.15

OH

HQ. 002H 0
0 ..0H
. OH Ho“ i OH Ho“ OH

HO OH OH OH
D-gluoose quinic acid 3R,5R-III hYdI’OXY
cyclohexanone

\ /
OH
O —*
—a>
OH
benzene hydroquinone

Figure 2. Synthesis of hydroquinone from glucose. (a) microbial synthesis; (b)
Ag3PO4/K28203; (c) HOCl; ((1) heat.

Indigo dye

Another success of pathway engineering is found in the microbial synthesis of
indigo dye. An estimated world market of 16,000 tons per year for indigo is dominated by
its use to dye denim. In the chemical process, phenylglycine reacts with KOH-NaOH at
900 °C with NaNHz to produce indoxyl, which subsequently oxidizes in air to form indigo
(Figure 3).18 The microbial synthesis of indigo from D-glucose was developed by
Genencor using metabolic engineering to incorporate a single P. putida gene encoding
naphthalene dioxygenase (NDO) into E. coli. This heterologously expressed protein
modified the tryptophan pathway in E. coli to cause high-level production of indole.
Indole was converted by NDO into indoxyl, which spontaneously oxidized to form indigo
(Figure 3).18 As another example of microbial synthesis of chemicals, genetically
engineered E. coli has been successfully used to synthesized 1,3-propanediol that is not

naturally produced in E. coli.19

indolxyl H
phenylglyci ne (keto-form) O —— N O
N

H o
COZCHa

L-t to han '"dOXY'
ryp p indole (enoI-form)

Figure 3. Chemical and microbial syntheses of indigo. (a) NaOH/KOH/NaNHz, 900

°C; (b) spontaneous oxidation; (c) E. coli native tryptophanase; (d) P. putida naphthalene
dioxygenase (N DO); (e) spontaneous oxidation.

1, 3 -Propanediol

1,3-propanediol is an important monomer in the production of poly(propylene
terephthalate) (PPT) polymer and fibers,20 which has long been known to have more
preferable properties over poly(ethylene terephthalate) (PET), and poly(butylenes
terephthalate) (PBT).21 Ethylene glycol and 1,4-butanediol, which are used to make PET
and PBT, however, are much more expensive than 1,3-propanediol, which is the
intermediate for PPT. Traditionally, 1,3-propanediol is produced either from the reductive
hydration of acrolein (Degussa-DuPont process), or through reductive carbonylation of
ethylene oxide (Shell process).

In the late 19903, microbial synthesis of 1,3-propanediol jointly developed by
DuPont and Genencor using renewable carbohydrates as the starting material came onto
the scene.‘9 The process originally patented by DuPont in 1996 consisted of a two-stage
fermentation, the first in S. cerevisiae leading to the synthesis of glycerol from glucose,

and the second in K. pneumoniae to convert the glycerol into 1,3-propanediol.22 The yeast

8

produces glycerol from the glycolytic intermediate dihydroxyacetone-3-phosphate using
glycerol-3-phosphate dehydrogenase (GPDl) and glycerol-3-phosphate phosphatase
(GPP2). The natural pathway for the production of 1,3-propanediol from glycerol requires
two enzymes, glycerol dehydratase (DhaB) and 1,3-propanediol dehydrogenase (DhaT).
A single microbe synthesis of 1,3-propanediol was reported in 1998 based on the
expression of GPDl, GPP2, DhaB and DhaT in a single E. coli microbe. E. coli cannot
convert dihydroxyacetone-3-phosphate into glycerol or synthesize 1,3—propanediol from
glycerol (Figure 4). E. coli provides several advantages over other organisms. It has been
an intensively studied organism, an extensive range of methodologies is available for its
genetic manipulation, and it has been used for some time for large scale production under
fermentor-controlled conditions. In addition, since E. coli does not naturally produce
glycerol or 1,3-propanediol, there is no natural regulation to overcome. Currently,
microbial synthesis of 1,3-propanediol from glucose using a single microbe achieves a

titer of 129 g/L in 34% yield (g of 1,3-propanediol/g of glucose consumed).23

The significance of 1,2,4-butanetriol and 1,2,4-butanetriol trinitrate

Mixed acid nitration of racemic 1,2,4-butanetriol produces racemic 1,2,4-
butanetriol trinitrate, which is used as an energetic plasticizer in propellant and explosive
formulations (Figure 5). Although 1,2,4-butanetn'ol trinitrate and nitroglycerin are used in
similar applications, 1,2,4-butanetriol trinitrate is less volatile, more thermally stable, and
has a lower shock sensitivity than nitroglycerin. Substitution of 1,2,4-butanet1ioltrinitrate
in plasticizer applications would therefore reduce hazards associated both with the

propellant manufacturing process and use of the final product.

OH

 

anaerobic TCA cycle

0 . \OH
i OH
HO OH .
/ D-gluoose \
O a O
H203PO/\;/U\H HOdK/OPOIBHZ
OH dihydroxyacetone
D-glyceralde hyde phosphate
3-phosphate l b
l ..
HO\/'\/ 0P03H2
glycerol

:QIIIIIIIIIIIIIIIIIOIIIIIIIIIIIIIIIIIIII‘: 3-phosphate
/u\cozH 0
g pyruvic acid g OH
5 5 HOW OH
g 5 glycerol
: COzH 0H 2
E succinic acid ethanol E HO\/l\¢o
3 5 3-hydroxy-
5 + : propionaldehyde
: OH 0 E
: _ e
5 ACOzH 2k0H l
E D-lactic acid acetic acid 5 [.10on
E g 1,3-propanediol

Figure 4. Schematic of the reaction network from D-glucose to 1,3-propanediol. (a)
triose phosphate isomerase (TIM); (b) S. cerevisiae glycerol dehydrogenase (GPDl); (c) S.
cerevisiae glycerol 3-phosphate phosphatase (GPP2); (d) K. pneumoniae glycerol
dehydratase (DhaB); (e) K. pneumoniae 1,3-propanediol dehydrogenase (DhaT).

10

OH O OH b ONO:
a
H3CONOCH3 ——> HONOH ——> OzNONONOz
O
dimethyl malate 1,2.4—butanetriol 1,2,4—butanetriol trinitrate

(3T) (BTTN)

Figure 5. Industrial synthesis of 1,2,4-butanetriol trinitrate. (a) NaBH4, ROH; (b)
H2804, HNO3.

The advantages of using 1,2,4-butanetriol trinitrate are unfortunately offset by its
expense, which largely reflects the expense of 1,2,4-butanetriol starting material.
Approximately 5,000 lbs of 1,2,4-butanetriol priced at $35/lb are used in the manufacture
of 10,000 lbs/yr of BTTN for the US Department of Defense.24 Lowering the price of
1,2,4-butanetriol to $20/lb could see demand grow for 1,2,4-butanetriol to 70,000 lb/yr
due to the substitution of nitroglycerin with 1,2,4-butanetriol trinitrate.24b

Beyond its use as an energetic material, 1,2,4-butanet1iol could have significant
impact in other aspects. Polyurethane foams made with 1,2,4-butanetriol have intriguing
elastic properties. These foams have the same compression-bending characteristics as

natural rubber. 25

D-l,2,4-Butanetriol trinitrate provides intriguing opportunities as an
alternative to nitroglycerin for use as a vasodilator in the treatment of angina.26 Beyond
its advantages in manufacture and formulation over nitroglycerin, 1,2,4-butanetriol
trinitrate is degraded more slowly by organic nitrate reductase than nitroglycerin.26 This
has the potential to translate into longer lasting plasma levels of D- and L-1,2,4-butanetriol
trinitrate relative to nitroglycerin. Using only a single enantiomer of 1,2,4-butanetriol
trinitrate could conceivably minimize the number of metabolites generated from the action

of organic nitrate reductase and correspondingly minimize the chances of encountering

adverse side effects.

11

Chemical syntheses of 1,2,4-butanetriol

1,2,4-Butanetriol has been commercially manufactured by the reduction of
dimethyl malate using sodium borohydride in a mixture of a C2-6-alcohols and
tetrahydrofuran (Figure 6).27 Such a stoichiometric reduction is expensive. In addition,
each ton of 1,2,4-butanetriol produced results in the generation of 2-5 tons of borate salts,

which constitute an environmental impediment to large scale manufacture.27

i O
ADH —> l)\/OH HON 4.9— HON

 

 

 

 

lcidol
aIIyIanohoI 9y \ y 0 3-butene-1-ol
OH \)O\H/\ O
a HO f
H3C02C\/\CO20H—> 0” ‘ HOAOH
3 1,2,4-butanetriol
dimethyl malate
e
c
H
>=O+H:H b HO:OH’9"HONOH
H
formaldehyde acetylene 2-butyne-1,4-diol 2-butene—1,4-diol

Figure 6. Syntheses of 1,2,4-butanetriol. (a) NaBH4, THF, MeOH; (b) CuCz, 50 atm;
(c) i) HgSO4, H2804, ii) 2CuO'CrzO7, H2; (d) Lindler’s catalyst, H2; (e) H202, H2WO4; (f)
I’d/C, H2; (g) H202, H2W04; (h) H2304; (0 H202; (l) i) C02(C0)8, (C0, H2), ii) LiA1H4.

A variety of other alternative routes to 1,2,4-butanetriol have also been explored
(Figure 6). Reaction of acetylene and formaldehyde catalyzed by CuCz to form 2-butyn-
1,4-diol provides two avenues to 1,2,4-butanetriol. Oxymercuration of 2-butyn-1,4-diol
and subsequent reductive demercuration gives 1,4-dihydroxy-2-butanone, which is

hydrogenated to afford 1,2,4-butanetriol. Alternatively, 2-butyn-1,4-diol can be converted

l2

to 2-butene-1,4-diol. Oxidation to the corresponding epoxide followed by hydrogenation
yields 1,3,4-butanetriol. A low-yielding route to 1,2,4-butanetriol consists of the reaction
of glycidol with CO and H2 followed by reduction of the 3-hydroxy-y-butyrolactone
intermediate. 3-Buten-1-ol, which is derived from acetaldehyde, can be oxidized to an
epoxide followed by hydrolysis to obtain 1,2,4-butanetriol. The most direct chemical
route to 1,2,4-butanetriol entails the low-yielding reaction of allyl alcohol with
formaldehyde.

In the Frost lab, metal-mediated catalytic hydrogenation of malate was examined
as an alternative to sodium borohydride and the salt streams attendant with this
stoichiometric reduction. Ruthenium catalyzed hydrogenation of malic acid in water
afforded 1,2,4-butanetriol in 74% yield. However, elevated H2 pressure (5,000 psi) and
elevated reaction temperature (135 °C) were required.28 Competing deoxygenation and
cracking reactions generated byproducts that were difficult to separate from 1,2,4-

butanetriol.28

A chemo-enzymatic synthesis of 1,2,4-butanetriol using L-ascorbic acid as
starting material was also investigated (Figure 7). Oxidation of L-ascorbic acid using
H202 gave L-threonate, which was dehydrated using dihydroxy-acid dehydratase.
Cyclization of the resulting 4-hydroxy-2-ketobutyric acid during extraction from water
afforded 2-hydroxy-2-buten-4-olide. High-pressure, high temperature catalytic
hydrogenation of 2-hydroxy-2-buten-4-olide gave 1,2,4-butanetriol in 53% overall yield.29

Given the shortcomings of the chemical and chemo-enzymatic routes, a microbial

synthesis of 1,2,4-butanetriol as pursued.

13

0H 0
HO

OH a
O r ‘d \ OH
ma lC act HO
OH
OH OH O 1 2 4obuta ‘
_ e . , netnol
HOJH. O o _b.. HO\/'\/CO2Na _C_. HO\/\",CO2Na 4. go /
HO OH 0“ 44‘ d0 OH
L-asoorbic acid sodium L-threonate y roxy- 2-h drox -2-
2-ketobutyrate butdh-4—dlide

Figure 7. Routes to 1,2,4-butanetriol previously developed in the Frost lab. (a) 5%
Ru on C at 1.3% Ru/malate, 340 atrn H2, 135 °C, 10 h, 74% yield; (b) Na2C03, 30% H202;
(c) E. coli JWFl/p0N1.118B cell extract, pH 8.2, 37 °C, N2; ((1) HCl, pH 1.5; (e) 1 mol%
Ru on C, 170 atrn H2, 125 °C, 53% overall yield from L-ascorbic acid.

Creation of a 1,2,4-butanetriol biosynthetic pathway

Microbial synthsis of 1,2,4-butanetriol poses a unique challenge, largely due to the
fact that 1,2,4-butanetriol is not a natural product. Since no biosynthetic routes apparently
exist in nature for D-1,2,4-butanetriol or L-1,2,4-butanetriol, a biosynthetic pathway had to
be created de novo. The first route developed in the Frost group was a two-microbe
synthesis of D-l,2,4-butanetriol from D-xylose (Figure 8). A wild-type microbe
Pseudomonas fi'agi ATCC4973 was used to oxidize D-xylose, which afforded 77 g/L
concentrations of D-xylonic acid in 70% yield.28 An E. coli construct then catalyzed the
conversion of D-xylonic acid into D-1,2,4-butanet1iol. The Frost lab discovered that native
E. coli expresses D-xylonate dehydratase (thG and YagF) and is able to use D-xylonic
acid as a sole source of carbon to support growth. Therefore, transport of D-xylonic acid
and its conversion to 3-deoxy-D-glycero-pentulosonic acid (Figure 8) relied solely on the
native E. coli activity. Conversion of D-xylonic acid to 1.6 g/L of D-l,2,4-butanetriol in

25% yield required only the expression of plasmid-localized Pseudomonas putida

14

ATCC 1263 3 mdl C -encoded benzoylformate decarboxylase in E. col i

DHSor/pWN6.168A.28

E. coli WN13/pWN7.1268

 

 

 

I t
OH OH 0H 0H 0H 0 0H 0 0H
HOMH a HOMOH b HOMOH c Ho\/'\)’\H d HO\/'\/\OH
0H 0 OH O O .
D-xylose D-xylonic acid 3'deOXY'D‘9/YCGIO' 3,4—dIhydroxy- 0-1 ,2,4-butanetrio|
L j L pentulosonlc acrd D-butanal +
P. fragi ATCC4973 E. coli DH5a/pWNG.186A
on OH oH OH OH 0 9H 0 OH
HOMH e HOMOH f, HOMOH g HOMH h, Ho\/'\/\OH
0H 0 OH 0 0 .
L-arabinose L-arabinonic acid MWXY'L‘QIYCGCO' 3,4-dlhydroxy- L—1,2,4-butanetriol
[ + l pentulosonic acrd L-butanal +

 

 

P. fragi ATCC4973 E. coli BL21(DE3)/p\M\16.222A

Figure 8. Microbial syntheses of D- and L—1,2,4-butanetriol. (a) D-xylonate
dehydrogenase; (b) D-xylonate dehydratase; (c) benzoylformate decarboxylase; ((1) alcohol
dehydrogenase; (e) L-arabinonate dehydrogenase; (f) L-arabinonate dehydratase; (g)
benzoylformate decarboxylase; (h) alcohol dehydrogenase.

A parallel biosynthetic pathway was also created for L-l,2,4-butanetriol. P. fi'agi
ATCC4973 was again recruited for the oxidation of L-arabinose, which afforded 55 g/L L-
arabinonic acid in 54% yield. The construction of the second E. coli catalyst, however, is
more complicated than the former case since E. coli is unable to catabolize D-arabinonic
acid. To circumvent this issue, a genomic library of P. fi'agi ATCC4973 was constructed
in E. coli followed by selection for E. coli strains that acquired the ability to use L-
arabinonic acid as a sole carbon source for growth and metabolism. This led to isolation
of the gene designated aadh encoding L-arabinonate dehydratase. The aatp gene encoding

arabinonate transport protein was also discovered to be contiguous to the aadh gene.

15

Conversion of L-arabinonic acid to 2.4 g/L of L-l,2,4-butanetriol in 19% yield was
accomplished using E. coli BL21(DE3)/pWN6.222A, which expressed the plasmid
localized aatp-encoded L-arabinonate transport protein, aadh-encoded L-arabinonate

dehydratase, and mdlC-encoded benzoylformate decarboxylase (Figure 8).28

A single microbe for D-l,2,4-butanetriol synthesis

Synthesis of D-1,2,4-butanetriol from D-xylose using a single microbe was then
pursued. The major challenges to the direct, microbe-catalyzed conversion of D-xylose
into D-1,2,4-butanetriol included the lack of a gene encoding D-xylonate dehydrogenase
and competing catabolism of intermediate 3-deoxy-D-glycero-pentulosonic acid.

A partial amino acid sequence of D-xylonic acid dehydratase was isolated from P.
fragi ATCC4973 by protein purification, by trypsin digestion, and N-terminal sequence
analysis of the peptide fragments. A Burkholderia fungorum LB400 protein was
identified in the ERGO bacterial genome database to encode D-xylonate dehydratase
activity using this information.14 Because genes encoding catabolic enzymes for the same
carbon source tend to form gene clusters in chromosomes, open reading frames adjacent to
this B. fungorum D-xylonate dehydratase were examined leading to identification of the D-
xylonate dehydrogenase in B. fimgorum. A similar approach also led to the discovery of
another D-xylonate dehydrogenase in Caulobacter crescentus CB15.14

E. coli has a different D-xylonic acid catabolism than P. fragi and elucidation of
this pathway turned out to be critical for improving the yields and concentrations of
biosynthesized D-l ,2,4-butanetriol. Two sets of catabolic enzymes encoded by the yjh and

yag gene clusters were discovered to be responsible for E. coli utilization of D-xylonic

16

acid as a sole source of carbon for growth (Figure 9).14 YagE and thH were further

identified to catalyze the cleavage of 3-deoxy-glycero-D-pentulosonic acid to form pyruvic
acid and glycolaldehyde.14

0

)Kn’OH —> TCA cycle
0
OH OH OH O pyruvic acid
HOMOH aI HO\/'\/“\n,0H b. + H0\/\OH
OH O o O / ethylene glycol
D- | - 'd 3-deoxy-D-glycero-
xy omc acr pentulosonic acid HO\)L H O
glycolaldehyde glycolate
HO\/U\OH ' catabolism

glycolate

Figure 9. Native E. coli D-xylonic acid catabolic pathway. (a) E. coli D-xylonate
dehydratase (thG, YagF); (b) 3-deoxy-D-glycero-pentulosonate aldolase (YagE, thH)

OH OH OH 0

HO OH

OH 0 HO OH
HOM/OH o
3-deoxy-D—glycero— 3.4-dihydroxy-D—
OH pentanoic acid butanoic acid
D-xylulose (5 g/L) (3 g/L)
el 91 1
OH OH OH OH OH O OH O OH
HOMH a HOMOH b, HO\/'\/“\",OH c HOMH d HO\/'\/\OH
OH O OH O 0 . -
pentulosonic 801d D-butanal (5 9/L)
y “W” \f
0H NH2 0 o
HOMOH MOH + HO\/[L
H
0 o
4,5-dihydroxy-threo— pyruvic acid glycolaldehyde
L-norvaline
(4 Q/L)

Figure 10. Microbial synthesis of D-l,2,4-butanetriol using WNl3/pWN7.126B. (a)
C. crescentus D-xylose dehydrogenase (xdh); (b) E. coli D-xylose dehydratase (yth and
yagF); (c) P. putida benzoylformate decarboxylase (mdlC); (d) E. coli alcohol
dehydrogenase(s); (e) inactivated E. coli D-xylose isomerase (xylA); (f) inactivated E. coli
3-deoxy-D-glycero-pentulosonate aldolase (yagE and yth); (g) E. coli dehydrogenase(s);
(h) E. coli transaminase; (i) E. coli dehydrogenase(s).

17

A single microbe, E. coli WN13/pWN7.126B, was subsequently constructed to
express all four enzyme activities (Figure 10) required for the synthesis of D-1,2,4-
butanetriol from D-xylose. The D-xylose dehydrogenase gene (xdh) from C. crescentus
was integrated to the xylAB site in the E. coli WN13 chromosome to convert D-xylose into
D-xylonic acid. Native E. coli D-xylonate dehydratases (YagF and thG) utilize D-
xylonic acid to produce 3-deoxy-D-glycero-pentulosonic acid. Knockouts of the 2-keto
acid aldolase genes (yagE and yth) eliminated catabolic competition for 3-deoxy-D-
glycero-pentulosonate. Plasmid pWN7.126B contains the benzoylformate decarboxylase
gene (mdlC) from P. putida. Expression of this gene from a tac promoter produced the
enzyme activity for conversion of 3-deoxy-D-glycero-pentulosonic acid into 3,4-
dihydroxy-D-butanal. As the final step in the biosynthesis, the conversion of 3,4-
dihydroxy-D-butanal into D-1,2,4-butanetriol relied solely on unidentified native E. coli
dehydrogenase activity. This E. coli microbe synthesized 6 g/L D-l,2,4-butanetriol in
28% yield (mol/mol) under fermentor-controlled conditions together with 3-deoxy-D-
glycero-pentulosonic acid, 3-deoxy-D-glycero-pentanoic acid, 4,5-dihydroxy-thre0-L-

norvaline and 3,4—dihydroxy‘D-butanoic acid as byproducts (Figure 10).

Important chemicals derived from the microbial 1,2,4-butanetriol synthesis

3, 4-Dihydr0xy-D—butan0ic acid

3,4-dihydroxy-D-butanoic acid lactonizes to form (S)-3-hydroxy-y-butyrolactone
(Figure 11), which is an important synthetic precursor. This compound is used in the

commercial manufacture of Astra Zeneca’s cholesterol-lowering drug Crestor® (Figure

18

11). One of the two patented routes to the cholesterol-lowering drug Zetia® (Figure 11)
also employs (S)-3-hydroxy-y-butyrolactone as a chiral synthon. Merck and Schering-

Plough jointly market Zetia®, which has a unique mode of action for lowering cholesterol.

3, 4 -Dihydroxy-thre0-L-n0rval ine

3,4-dihydroxy-threo-L-norevaline has been used as a precursor for the synthesis of
a clavam antibiotic, Clavalanine.30 This antibiotic was isolated in 1983 from Streptomyces
clavuligerus,31 which is known to be a prolific B-lactam antibiotic producer. Clavalanine
is unique for being an antimetabolite of O-succinylhomoserine which results in disruption
of methionine biosynthesis.3 1° This stands in marked contrast to disruption of
peptidoglycan biosynthesis, which characterizes the mode of action of other B-lactam

antibiotics.

OH 0 . . HO
HOMOH lactonlzatlon mo

3,4—dihydroxy-D— (S)—3-hydroxy.y.
butanoic acid butyrol actone

 

Crestor

OH NH2

HH NH

O
4,5-dihydroxy-threo- Clavalanine
L-norvaline

Figure 11. Drugs synthesized from 3,4-dihydroxy-D—butanoic acid and 4,5-
dihydroxy-threo-L-norvaline.

l9

REFERENCE

' Pannuri, S.; DiSanto, R.; Kamat, S. In Kirk-0thmer Encyclopedia of Chemical
Technology Online; Biocatalysis, 2003, Wiley.

2 Junker, B. In Kirk-0thmer Encyclopedia of Chemical Technology Online; Fermentation,
2004, Wiley.

3 Adrio, J. L.; Demain, A. L. In Microbial Enzymes and Biotransformations; Barrede, J. L.
Ed.; Humana Press: Totowa, NJ, 2005; pl.

4 Cohen, 5.; Chang, A. c. Y.; Boyer, H. W.; Helling, R. B. Proc. Natl. Acad. Sci. USA.
1973,70,3240.

5 Saiki, R. K.; Scharf, 3.; Faloona, F.; Mullis, K. 13.; Horn, G. r; Erlich, H. A.; Arnheim,
N. Science 1985, 4732, 1350.

6 Vasic-Racki, D. In Industrial Biotransformation; Liese, A.; Seelbach, K.; Wandrey, C.
Eds.; Wiley-VCH: Weinheim, 2006; pl.

7 Bomscheuer, U. T.; Buchholz, K. Eng. Life Sci. 2005, 5, 309.

8 Buchholz, K.; Poulsen, P. B. In Applied Biocatalysis; Straathof, A. J. J .; Adlercreutz, P.
Eds.; Harwood Academic Publisher: Amsterdam, 2000.

9 Robertson, D. E.; Chaplin, J. A.; DeSantis, G.; Podar, M.; Madden, M.; Chi, E.;
Richardson, T.; Milan, A.; Miller, M.; Weiner, D. P.; Wong, K.; McQuaid, J .; Farwell, B.;
Preston, L. A.; Tan, X.; Snead, M. A.; Keller, M.; Mathur, E.; Kretz, P. L.; Burk, M. J.;
Short, J. M. Appl. Environ. Microbiol. 2004, 70, 2429.

'0 (a) Arnold, F. H.; Georgiou, G. In Methods in Molecular Biology, Vol. 230; Humana
Press: Totowa, NJ, 2003. (b) Arnold, F. H.; Georgiou, G. In Methods in Molecular
Biology, Vol. 231; Humana Press: Totowa, NJ, 2003. (c) Brakmann, S.; Johnsson, K.
Directed Molecular Evolution of Enzymes; Wiley-VCH: Weinheim, 2002. (d) Brakmann,
S.; Schwienhorst, A. Evolutionary Methods in Biotechnology; Wiley-VCH: Weinheim,
2004.

H Ghisalba, 0. In New Trends in Synthetic Medicinal Chemistry; Gualtieri, F. Ed.; Wiley-
VCH: Weinheim, 2000; p175.

12 (a) Draths, K. M.; Knop, D. R.; Frost, J. W. J. Am. Chem. Soc. 1999, 121, 1603. (b)
Frost, J. W.; Frost, K. M.; Knop, D. R. WO 2000044923 A1 20000803. (c) Knop, D. R.;

20

Draths, K. M.; Chandran, S. S.; Barker, J. L.; von Daeniken, R.; Weber, W.; Frost, J. W. J.
Am. Chem. Soc. 2001, 123, 10173. (d) Chandran, S. S.; Yi, J.; Draths, K. M.; von
Daeniken, R.; Weber, W.; Frost, J. W. Biotech. Prog. 2003, 19, 808.

13 (a) Barbas, C. F .; Wang, Y.-F; Wong, C.-H. J. Am. Chem. Soc. 1990, 112, 2013. (b)
Chen, L.; Dumas, D. P.; Wong, C.-H. J. Am. Chem. Soc. 1992, 114, 741.

14 Niu, W.; Frost, J. W. unpublished results.

15 Sheng, Y.; Akesson, C.; Holmgren, K.; Bryngelsson, C.; Giamapa, V.; Pero, R. W. J.
Ethnopharmacol. 2005, 96, 577.

'6 Ran, N.; Knop, D. R.; Draths, K. M.; Frost, J. w. J. Am. Chem. Soc. 2001, 123, 10927.

17 Krumenacker, L.; Constantini, M.; Potal, P.; Sentenac, J. In Kirk-0thmer Encyclopedia
of Chemical Technology Online; Hydroquinone, Resorcinol, and Catechol, 1995, Wiley.

'8 (a) Ferreira, E. S. B.; Hulme, A. N.; McNab, H.; Quye, A. Chem. Soc. Rev. 2004, 33,
329. (b) Bommarius, A. S.; Riebel, B. R. Biocatalysis; Wiley-VCH: Weinheim, 2004.

‘9 Gatenby, A. A.; Haynie, s. L.; Nagarajan wo 9321339, 1998.

20 Polk, M. B.; Vigo, T. L.; Turbak, A. F. In Kirk-0thmer Encyclopedia of Chemical
Technology Online; High Performance Fibers, 2004, Wiley.

2' Hansen, S.; Atwood, K. B. In Kirk-0thmer Encyclopedia of Chemical Technology
Online; Polyester Fibers, 2005, Wiley.

22 Haynie, s. L.; Wagner, L. w. wo 35799, 1996.

23 Emptage, M.; Haylie, s. L.; Laffend, L. A.; Pucci, J. P.; Whited, G. M. US 7,067,300,
2006.

24 (a) Charles Cappuccino, Operations Manager, Copperhead Chemical Company Inc. (b)
Charles Painter, ManTec Center, Naval Surface Warfare Center, Indian Head Division.

25 Gmitter, G. T. US 3,149,083, 1964. (b) Wilson, C. L.; Riedeman, W. L. DE 1047420,
1958.

26 Needleman, P. Annu. Rev. Pharmacol. Toxicol. 1976, 16, 81.

21

27 (a) Monteith, M. J .; Schofield, D.; Bailey, M. W0 98/08793, 1998. (b) ikai, K.;
Amagasaki-shi, H. EP 1,061,060 A1, 1999.

28 (a) Niu, W.; Molefe, M.; Frost, J. w. 1 Am. Chem. Soc. 2003, 125, 12998. (b) Frost, J.
w. wo 2005/068642, 2005.

29 Molefe, M. PhD Thesis, Michigan State University, 2005.

30 Bernardo, S. D.; Tengi, J. P; Sasso, G. J .; Weigele, M. J. Org. Chem. 1985, 50, 3457.

31 (a) Higgens, C. E.; Kastner, R. E. Int. J. Syst. Bacteriol. 1971, 21, 326. (b) Evans, R.
H.; Ax, H.; Jacoby, A.; Williams, T. H.; Jenkins, E.; Scannell, J. P. J. Antibiotics 1983, 36,

213. (c) Pruess, D. L.; Kellet, M. J. Antibiotics 1983, 36, 208. (d) Muller, J.-C.; Toome,
V.; Pruess, D. L.; Blount, J. F.; Weigele, M. J. Antibiotics 1983, 36, 217.

22

CHAPTER TWO

Chemo-Enzymatic Synthesis of D-l,2,4-Butanetriol
Using 2-Deoxyribose-S-Phosphate Aldolase

Background

Aldolases are emerging into effective and economically viable tools for the
industrial syntheses of chiral molecules. They catalyze enantioselective carbon-carbon
bond formations to generate up to two chiral centers under mild reaction conditions.
Among the aldolases proven to be useful for large-scale synthesis is the deoC gene
encoding 2-deoxyribose-5-phosphate aldolase (DERA, EC 4.1.2.4).1 In E. coli, this
aldolase is associated with thymidine phosphorylase (DeoA), purine nucleoside
phosphorylase (DeoD), and phosphopentomutase (DeoB) in the deo operon.2 2-Deoxy-D-
ribose-S-phosphate aldolase is a 28 kDa enzyme that reversibly catalyzes the retro-aldol
reaction of 2-deoxy-D-ribose-5-phosphate, a degradation product of deoxyribonucleic acid
to form acetaldehyde and D-glyceraldehyde-3-phosphate (Figure 12).3 Interestingly, the
chemical equilibrium lies towards 2-deoxy-D-ribose-5-phosphate, with an equilibrium

constant of 4.2 x 103 M".4

 

O O O OH
DERA -
HJK + HJYOPOf’ ..— ‘ HWOPoai"
OH OH
acetaldehyde D-glyoeraldehyde— D-2-deoxyribose—
3-phosphate 5-phosphate

Figure 12. In vivo reaction catalyzed by 2-deoxyribose-5-phosphate aldolase.

23

Mechanistically, 2-deoxyribose-5-phosphate aldolase belongs to the type I class of
aldolases with the enzymatic reaction proceeding through a Schiff base intermediate
between the donor substrate and an active site lysine residue.5 It is the only aldolase
known to date which accepts two aldehydes in the condensation reaction. Other aldolases
require a ketone and an aldehyde. Several literature reports show that 2-deoxyribose-5-
phosphate aldolase has high substrate specificity towards using acetaldehyde as the donor
substrate.6 Its application in organic synthesis has been emerging since the first report of
an unprecedented one-pot tandem aldol reaction to synthesize six-membered lactol
derivatives, in which two equivalents of acetaldehyde were added in sequence to a variety
of two-carbon aldehyde acceptors (Figure 13).7 Since 2-deoxyribose-5-phosphate aldolase
catalyzes reactions in both directions, the intermediate 4-carbon adduct is reversibly
formed under the reaction conditions. The second condensation took place between this

adduct and a second equivalent of acetaldehyde to form a more stable lactol.

 

 

 

 

 

o DERA OH 0 DERA OH OH 0 R 0 0“

DERA = 2-deoxyribose-5-phosphate aldolase
R = H, Cl, OCH3, N3

Figure 13. 2-Deoxyribose-5-phosphate aldolase catalyzed tandem aldol reaction.

The industrial interest in 2-deoxyribose-5-phosphate aldolase lies in the structural
similarity of its enzymatic products to the lactone moiety of the cholesterol lowering 3-
hydroxy-3-methylglutaryl (HMG)-COA CoA reductase inhibitors, which are collectively

known as statins. Unlike the earlier generation of statins that were mostly fungal

24

metabolites or semi-synthetic molecules, Pfizer’s Lipitor (atorvastatin) (Figure 14) and
AstraZeneca’s Crestor (rosuvastatin) (Figure 11) are completely synthetically made. A
(3R,5S)-dihydroxyhexanoate moiety that is commonly shared by these two drugs turns out
to be essential for its activity. Chemical construction of the two chiral centers involved in
this side chain proved to be a challenging task. For instance, the (3R,5S)-
dihydroxyhexanoate in atorvastatin accounts for about 25% of the compound’s total
molecular weight. In addition, a synthesis of this intermediate with greater than 99.5%
enantiomeric and 99% distereomeric excess is required to meet the standard for
pharmaceutical use. To circumvent these issues associated with the traditional chemical
synthesis, DSM8 and the enzyme developer Diversa9 have independently developed two
improved 2-deoxyribose-5-phosphate aldolase variants for statin intermediate synthesis.
These improved bioprocesses allows the intermediate (3R,5S)-6-chloro-2,4,6-
trideoxyhexapyranoside to be synthesized in a large quantity and compete with other

traditional routes to supply the 220 tons per year market demand worldwide.

 

(3R.5S)-6-chloro-2,4.6- chloro-
tn‘deoxyhexapyranoside acetaldehyde

atorvastatin (Lipitor)

acetaldehyde

Figure 14. Retro-synthesis of atorvastatin (Lipitor).

25

The Frost group developed several approaches to the environmental benign
synthesis of 1,2,4-butanetriol as discussed previously in Chapter 1. Among them is the
direct catalytic hydrogenation of D- and L-malic acid. Hydrogenation of malic acid in
aqueous solution using 5% ruthenium on carbon as catalyst was investigated in detail.m’“
Carboxylates with electron withdrawing groups attached to the adjacent carbon atom are
known to be activated towards metal-mediated reduction.12 However, the presence of the
second unactivated carboxylate in malic acid is far less reactive and its hydrogenation
requires elevated H2 pressure and temperature. Under an H2 pressure of 1,000 psi and
temperature of 135 °C, malic acid was converted into 1,2,4-butanetriol in 25% yield using
5% ruthenium on carbon at 1.3% ruthenium to malic acid ratio. Exclusive reduction of the
activated carboxylate on malic acid led to the formation of 3,4-dihydroxybutanoic acid
and 3-hydroxybutyrolactone as the side product in a combined yield of 60%.11 Further
optimization of the hydrogenation process resulted in complete reduction of malic acid but
also formed other side products. Under an elevated H2 pressure of 5,000 psi and 135 °C,
1,2,4-butanetriol was synthesized in a maximum yield of 74% with other reaction
parameters unchanged.11 The elevated H2 pressure and temperature led to the formation of
various polyol byproducts due to carbon-carbon and carbon-oxygen bond cleavage
reactions (Figure 15). Due to the structural similarity of 1,2,4-butanetriol and these
byproducts, downstream purification becomes challenging. In addition, the expense of an
industrial scale hydrogenation at 5,000 psi H2 pressure was projected to be quite high.
Although malic acid can be obtained using microbial synthesis with a renewable

feedstock, its catalytic hydrogenation to synthesize 1,2,4-butanetriol is problematic.

26

OH OH

/\/\/OH

1,2,4-butanetn'ol, 74% 1,2-propanediol, 11% 1,4-butanediol, 8%

OH 0
a
”ONon —+ OH O
o
Ho/\/OH \/S\,OH All/E0
malicacid H0

ethylene glycol, 3% 1,2-butanediol, 3% 3-hydroxy-
butyrolactone, 1%

Figure 15. Catalytic hydrogenation of malic acid. (a) 5% Ru on C at 1.3% Ru/malate,
5,000 psi H2, 135 °C, 10 h, 74% yield.

An alternate route to 1,2,4-butanetriol has also been examined in the Frost group.
2-Hydroxy-buten—4-olide (Figure 7) was targeted as the substrate for catalytic
hydrogenation to afford 1,2,4-butanetriol based on its single carboxylate possessing an
activating hydroxyl group attached to the adjacent carbon atom. Milder reaction
conditions for the hydrogenation of butenolide relative to those required for the
hydrogenation of malic acid were of interest to avoid the formation of polyol byproducts.
1,2,4-butanetriol was synthesized from 2-hydroxy-2-buten-4-olide under H2 pressure of
2,500 psi at 125 °C using 1.0 mol% ruthenium catalyst (5% ruthenium on carbon) in a
yield of 96% without observable byproduct formation.ll Synthesis of 2-hydroxy-2-buten-
4-olide started from L-ascorbic acid and proceeded through intermediacy of L-threonate
and 4-hydroxy-2-ketobutyrate using a chemo-enzymatic route in a yield of 55% (Figure
7). Compared with the catalytic hydrogenation of malic acid, the hydrogenation of 2—
hydroxy-2-buten—4-olide afforded a significant improvement in both catalytic
hydrogenation reaction yield and product purity. However, the overall yield for the
chemo-enzymatic process to 1,2,4-butanetriol from L-ascorbic acid was only 53% (Figure
Figure 7). Furthermore, replacing an industrially well-established one-step sodium

borohydride reduction by this four-step process is not economically feasible.

27

Accordingly, a two-step chemo-enzymatic synthesis of D-1,2,4-butanetriol has
been elaborated. It involves an enzymatic asymmetric aldol reaction followed by catalytic
hydrogenation (Figure 16). The first step in our synthetic scheme is an enantioselective
aldol condensation reaction using acetaldehyde and glycolaldehyde. Beyond the
traditional demand of enantioselective control, the proposed reaction also requires high
regioselectivity preference. The acetaldehyde molecule participates strictly as a donor
substrate in the aldol reaction, while glycolaldehyde acts as receptor to form the
corresponding 3,4-dihydroxy-D-butanal as the sole product and thus avoided the non-
enzymatic mixed-aldol products formation. 0n the other hand, the reaction environment
has to be controlled to avoid the retro-aldol reaction to take place. The success of
developing this chemo-enzymatic process relied on a strategy using 2-deoxyribose-5-
phosphate aldolase, the only enzyme that accepts two aldehydes (acetaldehyde and
glycolaldehyde) as substrates to produce 3,4-dihydroxy-D-butanal. The resulting
intermediate 3,4-dihydroxy-D-butanal was anticipated to undergo facile catalytic
hydrogenation relative to the carboxylates employed in previous studies and yield D-1,2,4-

butanetriol as the target product.

0 o DERA 0 9H H2/cat. OH
H’”\ + H’U\/OH ‘— H’I'J\/\/OH HONOH
acetaldehyde glycolaldehyde 0-3.4—dihydroxybutanal D-1,2,4—butanetriol

Figure 16. Proposed chemo-enzymatic reaction using 2-deoxyribose-5-phosphate
aldolase.

28

Another important advantage of this proposed synthesis lies on its simple and
inexpensive starting materials. Acetaldehyde and glycolaldehyde are readily available in
large quantities. Acetaldehyde is manufactured by liquid phase Wacker oxidation of
ethylene catalyzed by palladium(II) chloride and copper(II) chloride. This is a process
first developed in 1957 — 1959 by Wacker-Chemie and Farbwerke Hoechst (Figure 17).13
Glycoaldehyde is produced by rhodium metal catalyzed reaction between syn gas with
forrnaldehydxe at 30 MPa. Formaldehyde is industrially made by a silver catalyzed

oxidation of methanol at 600 — 650 °C (Figure 17).14

o
H2C=CH2 —a——> HJJ\

 

ethylene acetaldehyde
b O c O
CH30H JL —-> Jk/OH
H H H
methanol formaldehyde glycolaldehyde

Figure 17. Industrial syntheses of acetaldehyde and glycolaldehyde. (a) PdCl2, CuC12,
02. (b) Ag, 02, 1 atm, 600 — 650 °C. (C) CO, H2, Rh, 30 MPa, 150 °C.

In this chapter, the 2-deoxyribose-5-phosphate aldolase catalyzed aldol reaction
and the catalytic hydrogenation will first be presented independently. This allows us to
determine all necessary reaction parameters. In the last part of this chapter, the two-step
synthetic process was scaled up. Downstream purification using short-path distillation
and Kugelrohr distillation will be discussed. Both yields determined by gas

chromatography and isolated yields will be reported.

29

Synthesis of 3,4-dihydroxy-D-butanal using 2-deoxyribose-5-phosphate aldolase

Protein purification of 2-deoxyribose-5 -phosphate aldolase

Over-expressed E. coli 2-deoxyribose-5-phosphate aldolase was obtained by
culturing the strain E. coli DH5o/pVH17 (ATCC86963) (Figure 18).15 A 4 L culture
generally provides 80,000 units of crude 2-deoxyribose-5-phosphate aldolase lysate
activity. The protein quality is of paramount importance to the isolated yield of D-1,2,4-
butanetriol, as was unambiguously proven in the course of our study. To obtain large
quantities of pure 2-deoxyribose—5-phosphate aldolase for the reaction trials, classical
enzyme purification was performed. This technique includes protein precipitation, ion-
exchange chromatography using an open-column, and FPLC. During the first attempt, a
literature purification protocol was followed with minor modifications.6 E. coli
DH50L/pVH17 crude lysate was subjected to protein precipitation using 1% streptomycin
sulfate. The resulting supernatant was fractionated using ammonium sulfate (40 —- 60%).
Subsequently the protein pellet was resuspended into buffer and loaded on a Mono-Q
10/ 10 anion exchange column controlled by an Amersham Bioscience AKTA Purifier.
Near-homogeneous 2-deoxyribose-5-phosphate aldolase was obtained and analyzed by
SDS-PAGE protein electrophoresis (Figure 19, left, lane 2 and 3). Purified protein
migrated as a major band on the gel with a molecular weight of approximately 28 kDa.
During the purification, 2-deoxyribose-5-phosphate aldolase activity was monitored by a
coupled enzyme assay using glycerol 3-phosphate dehydrogenase and triose phosphate
isomerase.6 The activity was determined by the depletion of NADH as indicated by

measuring optical density at 340 nm (Table 1).

3O

A ddeo CABD DNA fragment

1) E0051 digest
2) Aval digest

EcoRl Aval

l 1.7 kb
[2" ”l

_> 3.»
.'<-.."A‘,,‘ I, 1,;
L. » . .v . '
- -~~~"* ' ’ (1600
24.21;; .L H/

1) EcoFil digest
i 2) Aval digest i
3) CIAP treatment

 

 

 

, /,. (fix
”2'" deoC 9% x ddeo CABD DNA fragment
l‘l pVH10 ti 1) 15le digest

\X I? EcoFtl EcoFll

x .8?"
(6:3. APR ”/3".
.k‘l:;-7‘ .4 4'1"

c.2247 . 424a?” 0-5 kb
1) 15le digest
2) CIAP treatment

Figure 18. Construction of plasmid pVHl7.15

31

 

 

 

 

 

"" 200 kDa \_ I
, fi,//116koa -. - .
4 ’/ 97 kDa ’ w* , ._ “a
W ——————- 45 kDa /' ‘
L ‘1 -.- — /_ ’ -‘ w . 5. _' .
u an m ‘ —_ 29 kDa w ‘

 

 

 

 

 

Figure 19. SDS-PAGE of 2-deoxyribose-S-phosphate aldolase purified from E. coli
DHSa. (left) using literature protocol; (right) using simplified protocol. M, Molecular
size marker; lane 1, after (N H4)2S04 fractionation; lane 2, fractions collected after Mono-

Q purification; lane 3, Mono-Q eluent after dialysis; lane 4, after Q-Sepharose F ast-Flow;
lane 5, after Mono-Q.

Table 1. Purification table of 2-deoxyribose-5-phosphate aldolase.

 

 

step volume [protein] pingin 22:33: abciwtaiiy recovery purification
(mL) (mg/mL) ("19) (Wins) (U) M) (fold)
crude 193 29.7 5732 10.1 57,900 100 1
strep. sulb 206 36.9 7601 8 60,800 100 0.8
(NHr)2$Orc 70 48.1 3367 14 47,100 81 1.4
dialysis 40 33.1 1324 11.2 14,800 26 1.1
Mono-Q 40 6.7 268 38 10,200 17 3.8

 

acne unit (U) of 2-deoxyribose-5-phosphate aldolase corresponds to the formation of 1 pmol of
cheraldehyde-3-phosphate per min at 25 °C.
streptomycin sulfate precipitation (strep. sul.)
0ammonium sulfate precipitation ((NH4)2SO.,)
As indicated in Table 1, the two protein precipitation steps resulted in 1.4 fold
purification. Surprisingly, about 32,300 units total enzyme activity was lost after

dialyzing the protein solution against buffer, presumably due to the instability of 2-

deoxyribose-S-phosphate aldolase in high salt solution. FPLC protein purification using

32

Mono-Q 10/10 anion exchange column turned out to be successful and 3.8 fold
purification was obtained in this single step. Further investigation resulted in a simplified
purification protocol involving two column chromatography steps using anion exchange
resin. An open column containing 300 mL Q-Sepharose Fast-F low anion exchange resin
was used to purify 3,000 mg crude protein in a single passage. The protein solution was
eluted with a gradient up to 1 M NaCl in buffer resulting in a 2.5-fold purification and a
77% recovery of enzyme activity. Near-homogeneous 2-deoxyribose-5-phosphate
aldolase was then obtained by chromatography using a Mono-Q column. These protein

samples were analyzed by SDS-PAGE protein electrophoresis (Figure 19, right, lane 4-5).

Table 2. Purification table of 2-deoxyribose-5-phosphate aldolase.

 

 

step volume [protein] pg: n 2233;: a331,, recovery purification
(mL) (mg/mL) (m9) (U/mQ) (U) (%) (fold)
crude 68 31.8 2160 17.6 38,000 100 1
Q-Sep.b 29 23.1 670 43.9 29,400 77 2.5
Mono-Q 9 32.7 294 52.4 15,400 40 3

 

8one unit (U) of 2-deoxyribose-5-phosphate aldolase corresponds to the formation of 1 pmol of
cheraldehyde-3-phosphate per min at 25 °C.
Q-Sepharose Fast-Flow anion exchange column (Q-Sep.)

Chemical synthesis of 3, 4-dihydroxy-D-butanal

3,4-dihydroxy-D-butanal is not a commercially available compound. To provide
enough authentic material for the titled chemo-enzymatic process, racemic 3,4-dihydroxy-
D-butanal was synthesized from racemic 1,2,4-butanetr'iol.l6 The 1,2-diol moiety was first
selectively protected with acetone as the corresponding acetal in the presence of p-
toluenesulfonic acid. The remaining terminal hydroxyl group was then oxidized to the

aldehyde using pyridinium chlorochromate. These two intermediates were easily purified

33

by using short-path distillation under reduced pressure. Target racemic 3,4-
dihydroxybutanal was obtained in 50% overall yield upon the deprotection of the acetal.
The final product was characterized by 1H and 13 C NMR spectroscopy and was identical
to literature values.

to ,, C}

0H 0
a
HO\/K/\OH O\)\/\OH M0

0H

0

Figure 20. Chemical synthesis of racemic 3,4-dihydroxybutanal. (a) acetone, p-
toluenesulfonic acid, 85%; (b) pyridium chlorochromate, CH2C12, 60%; (c) Dowex
5)(H+), H20, quantitative.

Enzymatic synthesis of 3, 4-dihydroxy-D-butanal

Small scale enzymatic reactions were carried out using literature conditions for

7 In order to

tandem aldol reactions to synthesize six-membered lactol derivatives.
establish the relationship between the enzyme purity and the yield of 3,4-dihydroxy-D-
butanal, this unstable aldehyde intermediate was quantitatively converted into the

corresponding D-1,2,4-butanetriol by stoichiometric sodium borohydride reduction.

0 OH 8 OH
HMOH ——> HO\/'\/\OH

D-3,4-dihydroxybutanal D-1,2,4-butanetriol

Figure 21. NaBH4 reaction for 3,4—dihydroxy-D-butanal. (a) NaBH4, H20.

34

Upon sodium borohydride reduction, the reaction mixture was passed through a
small pad of Dowex 50(H+) resin. Boric acid was removed by repeated azeotropic
distillation with methanol. This protocol was used to determine the influence of enzyme
purity on the yield of 3,4-dihydroxy-D-butanal as summarized in Table 3. Interestingly,
using the same amount but higher purity of 2-deoxyribose-5-phosphate aldolase resulted
in a significant improvement in the final yield of 3,4-dihydroxy-D-butanal. Using 1,000
activity units of aldolase purified by protein precipitation increased the yield of 3,4-
dihydroxy-D-butanal by 9% as supposed to reaction catalyzed by crude cell lysate (Table
3, entries 1 and 2). The same amount of aldolase purified by anion exchange columns
generally doubled the product obtained, up to a maximum yield of 50% (Table 3, entries 3
and 4). Furthermore, the amount of 3,4-dihydroxy-D-butanal obtained using Q-Sepharose
purified 2-deoxyribose-5-phosphate aldolase (43.9 U/mg) was comparable to the same
reaction using enzyme purified by Mono-Q equipped F PLC (52.4 U/mg) (Table 3, entries
3 and 4). This suggested that passing the crude lysate through Q-Sepharose resin offers a

one-step method to obtain 2-deoxyribose-5-phosphate aldolase for synthetic purposes.

Table 3. Influence of 2-deoxyribose-5-phosphate aldolase (DERA) purity on 3,4-
dihydroxy-D-butanal yield.

 

 

specific activitya total units reaction time product yield
entry DERA batch
(U/mg) (U) (daYS) (%)
1 crude 12.0 1,000 3 14
2 protein pptb 14.0 1.000 3 22
3 Q-Sep.c 43.9 1,000 3 45
4 Mono-Q 52.4 1 .000 3 50

 

aOne unit (U) of 2-deoxyribose-5-phosphate aldolase corresponds to the formation of 1 umol of
gcheraldehyde—3-phosphate per min at 25 °C.

Protein precipitation steps using streptomycin sulfate followed by ammonium sulfate (protein ppt.)
cQ-Sepharose Fast-Flow anion exchange column (Q-Sep.)

35

Notably, the lactol 2,4-dideoxy-D-hexapyranoside, which resulted from the
condensation of one molecule of glycolaldehyde and two molecules of acetaldehyde
catalyzed by 2-deoxyribose-5-phosphate aldolase was also formed.7 However, this double
aldol product was only observed as a minor byproduct in contrast to other a-substituted

aldehyde receptors.

Catalytic hydrogenation of racemic 3,4-dihydroxybutanal

Having succeeded in the enzymatic-catalyzed condensation, focus was changed to
the catalytic hydrogenation step in the proposed chemo-enzymatic synthesis of D-l,2,4-
butanetriol. Based on the previous successful use of commercially available 5 wt%
ruthenium on carbon in the hydrogenation of D- and L-malic acid in aqueous medium, the
same catalyst was employed to hydrogenate chemically synthesized 3,4-dihydroxybutanal
(Figure 16).m’” Variation of pressure, temperature, reaction time, concentration of
substrate, rate of stirring and catalyst loading were explored. Representative reaction
trials are presented in Table 4. A typical hydrogenation was conducted by dissolving
authentic racemic 3,4-dihydroxybutanal in distilled, deionized water (100 mL) in a glass
liner. The catalyst was suspended in the 3,4-dihydroxybutanal solution. The liner was
inserted into a 500 mL Parr 4575 stainless steel high temperature-high pressure reactor,
and the vessel was sealed. The temperature and stir rate were controlled by a Parr 4842
temperature controller. Hydrogen was bubbled through the reaction mixture for 10 — 15
min to remove air while stirring at 100 rpm. The vessel was then charged with hydrogen
gas to a pressure below the desired value. After heating the reaction to desired

temperature, the hydrogen pressure was adjusted to the desired pressure. The reaction was

36

stirred at 200 rpm for 1 — 10 h at a constant temperature. When the reaction was
completed, the vessel was cooled to room temperature and the pressure was released.
After removal of the catalyst by filtration, the reaction mixture was concentrated to
dryness under vacuum to afford a colorless oil, which was derivatized with N,0-bis-
(trimethylsilyl)-trifluoroacetamide (BSTF A) and analyzed by gas chromatography.

In the case of catalytic hydrogenation of D- and L-malic acid, elevated temperature
and pressure at 135 °C and 5,000 psi is required.”’” Competing deoxygenation and
cracking reactions generate byproducts that are very difficult to separate from 1,2,4-
butanetriol. In the chemo-enzymatic synthesis, a carboxylate functionality is substituted
by a more reactive aldehyde and thus milder reaction conditions could be employed. In
Table 4, catalytic hydrogenation of a 250 mM aqueous solution of 3,4-dihydroxybutanal at
400 psi H2, a 700 rpm stir rate, and a reaction temperature of 50 °C for 5 h resulted in a
23% yield of D-1,2,4-butanetriol (entry 1). Increasing the catalyst loading had a
pronounced impact on the yield of 1,2,4-butanetriol even under reduced temperature and
pressure. The yield of 1,2,4-butanetriol increased to 68% using 0.5 mol% of the same
ruthenium on carbon catalyst at room temperature (entry 2). Lowering the pressure from
400 psi to 100 psi did not change the yield of product (entry 3). Increasing the catalyst
loading to 1 mol% led to another increase in product yield (entry 4). Attempts to increase
the catalyst loading beyond 1 mol% did not translate to further improvement in 3,4-
dihydroxybutanal reduction yields (result not shown). In addition, reducing the stir rate to
400 rpm resulted in no difference in 3,4-dihydroxybutanal reduction yields (entry 5).
However, a pronounced decrease in product yield was noticed with doubled substrate

concentration (entry 6). At last, both reaction time and temperature were found to be the

37

keys to complete conversion. A reaction with product yield exceeding 90% was achieved
simply by doubling the reaction time (entry 7 vs. entry 6). Increasing reaction temperature
led to a 10% increase in the final yield of 1,2,4-butanetriol (entry 10 vs. 8). Catalytic
hydrogenation of a 250 mM aqueous solution of 3,4-dihydroxybutanal at 200 psi H2, a 700
rpm stir rate, and a reaction temperature of 30 °C for 5 h resulted in a quantitative yield of

1 ,2,4-butanetriol.

Table 4. Catalytic hydrogenation of authentic 3,4-dihydroxybutanal.

 

 

Ru on C substrate temp pressure stir rate time GC yield
entry (mol%) (mM) (°C) (psi) (rpm) 00 (%)
1 0.1 250 50 400 700 5 23
2 0.5 250 25 400 700 5 68
3 0.5 250 25 100 700 5 71
4 1 250 25 100 700 5 82
5 1 250 25 100 400 5 80
6 1 500 25 100 400 5 62
7 1 500 25 100 400 10 91
8 1 250 25 200 700 5 93
9 1 250 20 200 700 5 81
10 1 250 30 200 700 5 quant.

 

Chemo-enzymatic synthesis of D-1,2,4-butanetriol

Having deve10ped the steps separately, the chemo-enzymatic synthesis was applied
to the preparation of D-1,2,4-butanetriol using acetaldehyde and glycolaldehyde as starting
material. As shown in Figure 22, the conversion began with the condensation of
acetaldehyde and glycolaldehyde using 2-deoxyribose-5-phosphate aldolase. The

intermediate, 3,4-dihydroxy-D-butanal was allowed to pass through a pad of Dowex 50

38

(H) The reaction mixture was then concentrated to about 150 mL in volume and
subjected to the high pressure-high temperature parr reactor for catalytic hydrogenation.
The optimized hydrogenation conditions using authentic racemic 3,4-dihydroxybutanal
was applied in this step. The yield of product D-1,2,4-butanetriol in the final reaction

mixture was found to be 30% as quantified by using 1H NMR and gas chromatography.

0 + o a 0 OH b OH
HJK,0H H/LK ——-—* HMOH ———> HONOH
glycolaldehyde acetaldehyde D-3,4—dihydroxybutanal D-1,2,4-butanetriol
Reaction condition: a) 50 mM triethanolamine b) 1 "10' °/0 5 Wl- % RU 0" C

1 mM EDTA 30 °c, 200 psi

300 mM acetaldehyde 5 hrs 700 rpm

100 mM lycolaldehyde '

20000 U ERA 150 mL total volume

400 mL total volume

Figure 22. Reaction parameters for the chemo-enzymatic synthesis of D-1,2,4-
butanetriol.

To determine the enantiomeric purity of chemo-enzymatically synthesized D-1,2,4-
butanetriol, product sample in the reaction mixture was partially purified and derivatized
using Moshers reagent, (S)-(+)-01-methoxy-or-(trifluoromethyl)phenylacetyl chloride. 17
Analysis of esterified 1,2,4-butanetriol using an HPLC Chiralpak AD column (Daicel
Chemical) revealed that the ee% of chemo-enzymatically synthesized D-1,2,4-butanetriol
was >99%.

Purification of product D-1,2,4-butanetriol, however, was not straightforward.
Attempts to purify crude D-1,2,4-butanetriol in the reaction mixture using short path

distillation in vacuo was unsuccessful. Distillation using a Kugerohr apparatus resulted in

39

a 17% yield of pure D-1,2,4-butanetriol as characterized by both 1H NMR and gas

chromatography.

Discussion and future work

The commercially employed route for the manufacture of 1,2,4-butanetriol relies
on a stoichiometric sodium borohydride reduction of dimethyl malate.18 The major issues
associated with this current synthesis include the disposal of byproduct borate salts and the
substantial cost of using a stoichiometric reducing agent sodium borohydride. Although
its derivative 1,2,4-butanetriol trinitrate offers superior physical properties over
nitroglycerin as an energetic plasticizer, these factors significantly limit the application of
1,2,4-butanetriol trinitrate for military and civilian purposes. To this end, novel
alternatives involve catalytic hydrogenation of malic acid using heterogeneous ruthenium

10’” chemo-enzymatic synthesis using L-ascorbic acid11 and the whole-

on carbon catalyst,
cell microbial synthesis using a created biosynthetic pathwaym’16 had been developed in
the Frost group to synthesize precursor 1,2,4—butanetriol. Catalytic malic acid
hydrogenation does not generate a salt stream as a reaction byproduct. However, the cost
of the ruthenium metal catalyst and the capital investment for constructing a high
pressure-high temperature reactor for large-scale 1,2,4-butanetriol production are
problematic. In addition, the formation of a number of polyol byproducts complicates the

downstream processing of the target moleculelo’ll

The four-step chemo-enzymatic
approach using L-ascorbic acid afforded 1,2,4-butanetriol in 50% overall yield.11

Although competing deoxygenation and cracking reactions are eliminated, a long

synthesis and loss of half of the starting material makes this process less attractive. The

40

microbial synthesis methodology offers an alternative that ultimately turned out to be the
best approach. However, at the time when the research was conducted, this complicated
two microbe, four enzyme synthesis suffered from low yields, low product concentrations
and byproduct formation.10 This prompted elaboration of a chemo-enzymatic synthesis
using 2-deoxyribose-5-phosphate aldolase.

Chemo-enzymatic synthesis of D-1,2,4-butanetriol presents several interesting
features that have not been explored by other approaches. This two-step route has the
advantage of relying on an enzyme that is produced in great abundance by a commercially
available strain and requires the catalytic hydrogenation of an aldehyde as opposed to a
carboxylate. Intermediate 3,4-dihydroxybutanal underwent catalytic hydrogenation over
ruthenium on carbon at a temperature of 30 °C and H2 pressure of 200 psi, which sharply
contrasts with the conditions (135 °C, and 5,000 psi H2) that were required for catalytic
malic acid hydrogenation. In addition, starting materials for this synthesis are achiral
molecules acetaldehyde and glycolaldehyde. These small molecular building blocks are
commercially available.

1,2,4-Butanetriol trinitrate that has been used as an energetic material has been a
racemic mixture, which reflects the use of methyl D-, L-malate as the starting material for
commercial synthesis of precursor racemic 1,2,4-butanetriol. The difference between the
melting points of the pure enantiomers relative to the melting point for the racemic
mixture of enantiomers is an important consideration in the utilization of energetic
materials. Moreover, 1,2,4-butanetriol trinitrate was proven to be an alternative to
nitroglycerin for use as a vasodilator in the treatment of angina.19 Using a pure chiral

chemical could conceivably minimize the chances of encountering adverse side effects. In

41

the 1,2,4-butanetriol syntheses employing D-, L-malic acid or L-ascorbic acid as starting
materials, racemization was observed during the problematic reduction of carboxylates

under elevated temperature and pressure.H

Fortunately, synthesizing a single D-1,2,4-
butanetriol using this chemo-enzymatic synthesis was found to be possible due to the
milder catalytic hydrogenation reaction condition involved. Unlike the microbial
approach where the stereochemistry is defined by the starting D-xylonic acid or L-
arabinonic acid, 2-deoxyribose—5-phosphate aldolase generates the product chirality using
an enantioselective aldol condensation.

Despite all these advantages of utilizing this chemo-enzymative synthesis using 2-
deoxyribose-S-phosphate aldolase, the enzymatic condensation of glycolaldehyde and
acetaldehyde and the catalytic hydrogenation only led to a 30% overall yield of D-l,2,4-
butanetriol. The major concern lies in the enzymatic aldol condensation reaction using 2-
deoxyribose-S-phosphate aldolase. First, the catalyst loading was 20,000 units per gram
of D-1,2,4-butanetriol synthesized, which is considered to be very high for an enzyme-
catalyzed reaction. Use of such a high concentration of enzyme would make the process
prohibitively expensive, in addition to making isolation of product from the reaction
mixture difficult. Second, the reaction time was on the order of several days, therefore
leading to a considerable increase in anticipated production costs. These issues can be
addressed by developing an improved 2-deoxyribose-5-phosphate aldolase. As a case
study, there are generally two different approaches persued in the industrial community to
improve this enzyme toward the synthesis of atorvastatin. DSM engineered 2-

deoxyribose-S-phosphate aldolase for higher affinity and resistance toward the substrate

chloroacetaldehyde, which is the key building block for atorvastatin side chain synthesis.8

42

They employed error-prone PCR and DNA shuffling techniques to generate the aldolase
mutant library. Subsequent screening led to the isolation of a ten-fold improved variant
for the reaction of interest. In 2002, Diversa also discovered a 2-deoxyribose-5-phosphate
aldolase alternative from an environmental genomic DNA library collected from a variety
of habitats.9 By using a combination of enzyme selection and screening, the newly
discovered enzyme resulted in a production rate of 30.6 g/L/h. Catalyst loading was also
improved by a 10-fold decrease while maintaining high enantio- and diastereoselectivity.
These examples suggest that it may be possible to increase the affinity and resistance of 2-
deoxyribose-S-phosphate aldolase towards glycolaldehyde. At the time when our chemo-
enzymatic synthesis of D-1,2,4-butanetriol was completed, our research focus turned to the
development of microbial syntheses using pentoses as starting materials. We believed that
a synthesis employing a single microbe would ultimately represent a general, practical,

and economical route to 1,2,4-butanetriol enantiomers.

43

REFERENCE

‘ Valentin-Hansen, P.; Boetius, F .; Hammer-Jespersen, K.; Svendsen, 1. Eur. J. Biochem.
1982, 125, 561.

2 Blank, J.; Hoffe, P. Mol. Gen. Genet. 1972, 116,291.
3 Racker, E. J. Biol Chem. 1952, 196, 347.
4 Pricer, W. E.; Horecker, B. L. J. Biol. Chem. 1960, 5, 1292.

5 (a) Hoffee, P. Arch. Biochem. Biophys. 1968, 126, 795. (b) Horecker, B. L.; Pontremoli,
S.; Ricci, 0; Cheng, T. Proc. Natl. Acad. Sci. USA 1961, 47, 1942.

6 (a) Barbas, C. F. 111; Wang, Y.-F.; Wong, C.-H. J. Am. Chem. Soc. 1990, 112, 2013. (b)
Chen, L.; Dumas, P. D.; Wong, C.-H. J. Am. Chem. Soc. 1992, 114, 741.

7 (a) Gijsen, H. J. M.; Wong, C.-I-I. J. Am. Chem. Soc. 1994, 116, 8422. (b) Wong, C.-H.;
Garcia-Junceda, E.; Chen, L.; Blanco, 0.; Gijsen, H. J. M.; Steensma, D. H. J Am. Chem.
Soc. 1995, 117, 3333.

8 Jennewein, S.; Schurmann, M.; Wolberg, M.; Hilker, I.; Luiten, R.; Wubbolts, M.; Mink,
D. Biotechnol. J. 2006, 1, 537.

9 (a) Liu, J .; Hsu, C.-C.; Wong, C.-H. Tetrahedron Lett. 2004, 45, 2439. (b) Greenberg,
W. A.; Varvak, A.; Hanson, S. R.; Wong, K.; Huang, H.; Chen, P.; Burk, M. J. Proc. Natl.
Acad. Sci. USA. 2004, 101, 5788.

‘0 Niu, W.; Molefe, M. N.; Frost, J. w. J. Am. Chem. Soc. 2003, 125, 12998.

'1 Molefe, M. N. PhD Thesis, Michigan State University, 2005.

'2 (a) Folkers, K.; Adkins, H. J. Am. Chem. Soc. 1932, 54, 1145. (b) Trenner, N. R.;
Bacher, F. A. J. Am. Chem. Soc. 1949, 71, 2352. (c) Zhang, 2.; Jackson, J. E.; Miller, D.
J. Appl. Cat. A 2001, 219, 89.

13 Hagemeyer, H. J. In Kirk-0thmer Encyclipedia of Chemical Technology; Wiley: 2002,
DOI: 10.1002/047123 8961 .0103052008010705.a01 .pub2.

'4 Gerberich, H. R.; Seaman, G. C.; Hoechst-Celanese Corporation In Kirk-0thmer

Encyclopedia of Chemical Technology; Wiley: 1994, DOI:
10.1002/0471238961.0615181307051802.a01

44

'5 (a) Valentin-Hansen, P.; Aiba, H.; Schumperli, D. EMBO J. 1982, I, 317. (b)
Jorgensen, P.; Collins, J.; Valentin-Hansen, P. Mol. Gen. Genet. 1977, 155, 93.

‘6 Niu, W. PhD Thesis, Michigan State University, 2004.
17 Dale, J. A.; Mosher, H. S. J. Am. Chem. Soc. 1973, 95, 512.

‘8 (a) Monteith, M. J.; Schoefield, D.; Bailey, M. PCT Int. Appl. wo 98/08793, 1998. (b)
Ikai, K.; Mikarni, M.; Furukawa, Y.; Ho, S. PCT Int. Appl. W0 99/44976, 1999.

‘9 Needleman, P. Annu. Rev. Pharmacol. Toxicol. 1976, 16, 81.

45

CHAPTER THREE

Improving Microbial Synthesis of 1,2,4-Butanetriol

Overview

The objective of this chapter is to develop a novel and improve an existing
microbial synthesis for 1,2,4-butanetriol as a precursor for the production of 1,2,4-
butanetriol trinitrate. A newly proposed biosynthesis of 1,2,4-butanetriol using erythritol
as the starting material under an anaerobic environment was explored (Figure 23). Efforts
were also diverted to improve the existing aerobic, pentose-based microbial process to D-

1,2,4-butanetriol using D-xylose as starting material (Figure 10).

0H0 0H0

HOMH HOMH
3,4—dihydroxy- 3,4-dihydroxy-
D-butanal L-butanal

/ N V \

OH OH OH
HO\/K/\OH HO\/'\i/\OH HO\/\/\OH
OH
D-1 2.4-butanetriol meso-erythritol L-1,2,4—butanetriol
N la V
O
HO\/u\/\OH

1,4~dihydroxy-
2-butanone

Figure 23. Proposed artificial biosynthetic pathway for 1,2,4-butanetriol from
erythritol. (a) diol or glycerol dehydratase. (b) alcohol dehydrogenase

46

Biosyntheses of D-1,2,4-butanetriol and L-butanetriol has been accomplished by
creating biosynthetic pathways not found in nature. A two-step enzymatic conversion of
erythritol into 1,2,4-butanetriol was also thought to be possible. For example, an active
erythritol dehydratase could convert erythritol into 3,4-dihydroxybutanal or 1,4-
dihydroxy-2-butanone (Figure 23). Alcohol dehydrogenase activity could then reduce D-
or L-3,4-dihydroxybutanal (Figure 23) or 1,4-dihydroxy-2-butanone (Figure 23) into 1,2,4-
butanetriol. Although the desired erythritol dehydratase activity was identified, the
activity was too low to be practically exploited. Instead of elaborating this newly explored
route into a microbial process, the research focus was changed to improving the pentose-
based pathway previously developed in the Frost group using D-xylose. In the second part
of this chapter, efforts toward isolating a more active 3-deoxy-glycero-pentulosonate
decarboxylase will be discussed. Potential wild-type decarboxylases were identified in
silico using bioinformatics tools and decarboxylase mutants were generated by site
directed mutagenesis and directed evolution methodologies. UV/vis spectrophotometry,
gas chromatography and high-throughput 96-well plate assays were developed to evaluate
these enzyme candidates. Finally, E. coli WNl3/pWN7.l26B, which synthesizes D-1,2,4-
butanetriol from D-xylose was subjected to further metabolic engineering to improve the
concentrations and yields of D-1,2,4-butanetriol. Differential expression of P. putida 3-
deoxy-glycero-pentulosonate decarboxylase (MdlC), E. coli molecular chaperonines
(GroES and GroEL), D-1,2,4-butanetriol dehydrogenase (Ath) and keto-acid
dehydrogenases (YiaE and chW) were studied. The resulting second generation
microbial catalyst of D-1,2,4-butanetriol was evaluated under fermentor-controlled

conditions.

47

Artificial biosynthesis of 1,2,4-butanetriol from erythritol

Background

The first step in the proposed bioconversion of erythritol to D-, and L-1,2,4-
butanetriol involves the dehydration of erythritol to D- and L-3,4-dihydroxybutanal (Figure
23). It was believed that either a glycerol dehydratase or diol dehydratase would be able
to catalyze this dehydration reaction. Glycerol dehydratase plays a critical role in the
ability of microbes to ferment glycerol to 1,3-propanediol while diol dehydratase catalyzes
a similar dehydration essential to the microbial fermentation of 1,2-propanediol.l The
final step of the erythritol-based route requires the reduction of D- and L-3,4-
dihydroxybutanal to the corresponding D-, and L-1,2,4-butanetriol, respectively (Figure
23).

Erythritol is commercially used as a sweetening agent in foods and is derived from
glucose via a microbial fermentation route.2 Cerester, which has been the major producer
of erythritol, was acquired by Cargill in 2002.3 The starch fraction obtained during corn
wet-milling is an abundant source of inexpensive glucose. Titers as high as 130 g/L and
yields of 46% have been reported for the microbial synthesis of erythritol from glucose.4
Considerable progress had been made in detailing the steps involved in microbial
synthesis of erythritol from glucose.5 This progress suggests, over the long term, that it
may be possible to construct a microbe capable of synthesizing 1,2,4—butanetriol directly
from glucose with erythritol as an in vivo biosynthetic intermediate.

Glycerol dehydratase (EC 4.2.1.30) and diol dehydratase (EC 4.2.1.28) can
catalyze the conversion of glycerol (Figure 24), 1,2-propanediol (Figure 25) and 1,2-

ethanediol to the corresponding aldehydesé’7 These enzymatic reactions proceeds through

48

a mechanism involving coenzyme 812 as an essential cofactor. Coenzyme 812 contains a
cobalt-carbon bond that is stable in water, although easily undergoes a homolytic cleavage
reaction when a suitable substrate is bound to the enzyme. This homolytic bond cleavage

initiates the catalytic cycle of all coenzyme Br 2 dependent enzymatic processes.8'9

OH
Home
H20 glycerol
a
d
HO 0 0
W NADH HOdK/OH
3-h drox - .
propioiialdeliyde dIhydroxyacetone
ATP
e D K
ADP
O
HO OH
\/\/ HO\)K/OP032‘
1 .3-propanediol dihydroxyacetone-
phosphate
0 I
OH
HO\/\/OP032'
glyceraldehyde-
3-phosphate

lll

acetate, ethanol. lactate,

forrnate, succinate, H2, C02

Figure 24. Anaerobic utilization of glycerol. (a) glycerol dehydrogenase; (b)
dihydroxyacetone kinase; (c) glyceraldehyde-3-phosphate dehydrogenase; (d) glycerol
dehydratase; (e) 1,3-propanediol dehydrogenase.

49

OH
A/OH

1,2-propanediol

j a
N0
propionaldehyde

2e' 2e'
)4 N
CH /\n,COA
O

propi onyl-CoA
Pi

/\/
propanol

C

/\n,opoaz'
o

propi onyl-
phosphate

ADP
14
ATP
/\ 002'
propi onate

Figure 25. Anaerobic utilization of 1,2-propanediol. (a) diol dehydratase; (b)
propionaldehyde dehydrogenase; (c) phosphotransacylase; (d) propionate kinase; (e) 1-
propanol dehydrogenase.

Glycerol and diol dehydratases are involved in the utilization of small molecule
under anaerobic conditions (Figure 24 and 3). These enzymes catalyze molecular
rearrangements that generate an aldehyde, which can be reduced during glycerol
fermentation. 0n the other hand, the aldehyde generated can also dismutate to oxidized
and reduced compounds in 1,2-propanediol fermentation. The anaerobic degradation of
glycerol is initiated by two enzymes (Figure 24). Glycerol dehydrogenase forms
dihydroxyacetone and glycerol dehydratase produces 3-hydroxypropionaldehyde which is

further reduced to l,3-propanediol.6"0"1 1,2-Propanediol fermentation shares a very

50

similar mechanism as ethanolarnine ammonia-lyase, which utilizes coenzyme B12 (Figure
25). Diol dehydratase first converts 1,2-propanediol into propionaldehydelz-‘B’l4 This
aldehyde dismutates into propanol and propionyl-CoA. These reactions allow the
generation of ATP during the propionate kinase catalyzed reaction (Figure 25).”’ '5
Glycerol and its closely related diol dehydratases were extensively studied. It had been
shown that these enzymes have similar molecular masses and substrate spectra.6’l6"7’18

Glycerol and diol dehydratases are known to undergo irreversible mechanism-
based inactivation by glycerol during catalysis.7’ '9 Inactivation by glycerol involves
irreversible cleavage of the cobalt-carbon bond of coenzyme B12, forming 5’-

193’ 20 Irreversible inactivation is

deoxyadenosine and an alkylcobalamin—like species.
brought about by tight binding of the modified coenzyme. Such suicide inactivation
seems enigmatic, since glycerol is essential for glycerol metabolism. However, it had
been shown that glycerol-inactivated dehydratases can be reactivated by an exchange of
the modified coenzyme for an intact coenzyme B12 in the presence of ATP, Mg2+ and a
rescue protein.21 The rescue protein is made up of two subunits and is located in the same
gene cluster as its corresponding dehydratase.

The proposed alternate biosynthesis of 1,2,4-butanetriol utillizes only two enzyme-
catalyzed steps, which appears to make this erythritol-based synthesis a simpler route
relative to the pentose-based syntheses previously developed in the Frost group. This
presumption is reinforced with the successful identification of dehydrogenases capable of

reducing D- and L-3,4-dihydroxybutanal. However, the stereoselectivity as well as the

regioselectivity in the enzyme-catalyzed erythritol dehydration reactions cannot be

51

predicted. This contrasts with the predictable, stereospecific synthesis of D-1,2,4-
butanetriol from D-xylose and L-l,2,4-butanet1iol from L-arabinose.
Dehydratase-catalyzed formation of an aldehyde from erythritol could lead to
either D- or L-3,4-dihydroxybutanal. The enantioselectivity of this desymmetrization
cannot be predicted (Figure 23). Furthermore, the regioselectivity of the dehydratase-
catalyzed reaction cannot be predicted. It is possible that 1,4-dihydroxy-2-butanone may
be produced from erythritol by the dehydratase. Formation of 1,4-dihydroxy-2-butanone
would wipe out the prochirality of erythritol. Whether pure enantiomers or enantiomeric
mixture of 1,2,4-butanetriol were synthesized would be dictated by the enantionselectivity
of the alcohol dehydrogenase catalyzing the reduction of 1,4-dihydroxy-2-butanone.
Irrespective of 3,4-dihydroxybutanal or 1,4-dihydroxy-2-butaneone intermediacy,
synthesis of mixtures of 1,2,4-butanetriol enantiomers is an acceptable outcome. This
follows from the current exclusive use of racemic 1,2,4-butanetriol trinitrate in high

energy material applications.

Identification of erythritol dehydratase candidates

A wide range of microorganisms are able to ferment glycerol and 1,2-propanediol.
To investigate whether these naturally occurring diol and glycerol dehydratase converts a
structurally similar substrate erythritol into its corresponding aldehyde or ketone, enzyme
candidates from seven bacteria were identified and are shown in Table 5. Fermentation of
glycerol by Citrobacter fieundii and Klebsiella pneumoniae had been studied
exclusively. 22 In the absence of an external oxidant, glycerol is fermented by a

dismutation process involving two pathways, leading to glycerol oxidation, the other being

52

responsible for oxidation of NADH (Figure 24). All four key enzymes and the
corresponding genes have been characterized in C. freundii and K.

6’11’23’24’25’26 In C. freundii, the structural genes of the glycerol dehydratase

pneumoniae.
(dhaBCE) are part of the dha regulon along with the genes encoding three other key
enzymes (dhaD, dhaK, dhaT) of the pathway. The glycerol dehydratase in K. pneumoniae
is comprised of three different subunits (gldABC) and catalyzes the same anaerobic
conversion of glycerol. The pddABC encoding diol dehydratase in K. pneumoniae has

17,28

also been identified and characterized. 27 Klebsiella oxytoca and Salmonella

. . 29
typhzmurzum

were reported to produce a similar diol dehydratase when grown
anaerobically in 1,2-propanediol or glycerol. Kledsiella oxytoca pddABC genes were
further characterized by cloning and expressing in E. coli.16 Interestingly, although the
pduCDE genes of S. typhimurium and the pddABC genes of K. oxytoca are over 90%
identical in nucleotide sequence, S. typhimurium cannot use 1,2-pr0panediol nor glycerol
as a sole carbon source to grow under anaerobic conditions.29 The glycerol fermentation
pathway of Clostridium pasteurianum is slightly different. In addition to the products
mentioned above, it synthesizes butyric acid, butanol and ethanol.”3 1 The dhaBCE gene
cluster encoding glycerol dehydrates in CI. pasteurianum were identified and had a high
amino acid sequence identity to those chararterized in K. pneumoniae and C. freundii.
The last candidates were Lactobacillus brevis strains LB18 and 19. Although genetic
information were not available for these strains, they were the only known strains that
transform meso-2,3-butanediol into 2-butanol. They are of particular interest due to the

structural similarlity between meso-2,3-butanediol and erythritol.32

53

Table 5. Glycerol and diol dehydratase candidates

 

 

entry strain gene locus plasmid

1 Citrobacterfreundii DSM30040 dhaBCE” pcs12033

2 Klebsiella pneumoniae ATCC25955 gIdBC pM L 5' 1 7 6
9’dABCb pML5.200
dhaT° (Ptac) pML5.179
dhaTc (pm pWN5.220A
gldABC-dhaT PML5-225
p d d A BCd_ dh aT pWN5.260A

3 Klebsiella oxytoca ATCCB724 pddABCd-ddrABf-dhaT pWN5.258A

4 Klebsiella ox ytoca ATCC43165

5 Salmonella typhimurium 'l‘l'10324 pduCDEa pXY3929

e Clostridium pasteurianum DSM525 dhaBCEa pm34

7 Lactobacillus brevis LB18

8 Lactobacillus brevis LB19

 

a'bglycerol dehydratase

d'ediol dehydratase
c1,3-propanediol oxidoreductase
rescue protein

ATCC43165 were obtained from the American Type Culture Collection. L. brevis LB18
(CNRZ734) and LB19 (CNRZ735) were obtained from the Centre National de Recherches
Zootechniques in France.
screened in vivo by culturing anaerobically in medium containing erythritol as substrate.

The resulting culture medium was subject to sodium borohydride reduction followed by

BSTFA derivatization and analyzed by gas chromatography.

production was observed.

54

Bacterial strains K. pneumoniae ATCC25955, K. oxytoca ATCC8724 and

As a preliminary study, these five wild-type strains were

No 1,2,4-butanetriol

Plasmid Constructions

A total of six plasmids were employed in the in vitro enzyme screening assays
using glycerol, 1,2-propanediol and erythritol as substrates. The genes encoding
coenzyme B12 dependent glycerol or diol dehydratase in C. freudii DSM30040, Cl.
Pasteurianum DSM525, K. pneumoniae ATCC25955, K. oxytoca ATCC8724 and S.
typhimurium TT10324 were expressed in E. coli. These plasmids were either obtained
from individual laboratories or prepared in-house (Table 5). A 2.7 kb PCR product
containing the dhaBCE gene fragment was amplified using C. freundii DSM3004O
genomic DNA as a template. The plasmid pCS120 was constructed by cloning this PCR
product into pBluescript SK+ (Stratagene).33 The dhaBCE gene cluster was transcripted in
the same direction as the lac promoter on the pBluescript SK+ cloning vector.
Recombinant cosmid pFLl contains a 13.5 kb insert of Cl. Pasteurianum genomic DNA
and harbors the genes encoding the reductive branch of glycerol breakdown. 3‘4 The
presence of dhaBCE genes encoding glycerol dehydratase was confirmed by DNA
sequencing. Plasmid pXY39 was constructed by cloning S. typhimurium pduCDE into
vector placIQPO-Bglll.29 This vector expresses the LaclQ protein and transcription of the
cloned dehydratase genes will therefore be induced by the addition of IPTG to the

medium. These three plasmids were obtained as gifts from other research laboratories.

55

 

, e
\
, x, .\
./ «'4‘-
at,
/ V'
r I ’V/).
,1 I V3.)
/ I
I.
ll ' f
A
I r
! I I
7 ,'
.1 7‘ 1’1

’4
.1}
.\ ./
‘\ \. //
.‘ \x a //' .'
\\ ‘ '/( H,
x‘ ‘\“‘ ‘ r," ‘)’
x‘ .1" ".414. ’w’
CJ\ ,w'” _..-
"f ._..

 

Figure 26. Plasmid maps of pCSlZO, pFLl and pXY39.

56

To construct a plasmid with a K. pneumoniae gldABC, a 1 kb DNA fragment
encoding gldBC was amplified from K. pneumoniae ATCC25955 genomic DNA and
subsequently digested with Sail and HindIII. Ligation of this fragment into the
SalI/HindIII digested pJF118EH resulted in pML5.176 (Figure 27). Another 1 kb
fragment encoding glycerol dehydratase subunit GldA was amplified using the same
genomic DNA followed by restriction enzyme digestion using EcoRI and Soil. Plasmid
pML5.200 was obtained by ligating this fragment into EcoRI/Sall digested pML5.176
(Figure 28). The resulting gldABC gene cluster was transcribed in the same direction as
the toe promoter located on pJF118EH. Transcription of the glycerol dehydratase was
initiated by the pJF118EH-localized tac promoter, which was under the control of LaclQ
repressor protein encoded by the laclQ gene.

The plasmid that expresses K. pneumoniae pddABC encoding diol dehydratase
and dhaT encoding 1,3-propanediol oxidoreductase was derived from cloning vector pT7-
7. This particular 1,3-propanediol oxidoreductase is known to be active toward the
conversion of 3,4-dihydroxybutanal into 1,2,4-butanetriol.3S A 1.2 kb fragment containing
the K. pneumoniae ATCC25955 dhaT gene was excised from plasmid pWN5.02236 using
BamHI. This fragment was ligated into the BamHI site of pT7-7 to afford plasmid
pWN5.220A (Figure 29). A 2.9 kb DNA fragment containing the pddABC gene was
amplified from the genomic DNA of the same K. pneumoniae strain and subsequently
digested with EcoRI. Ligation of this fragment into the EcoRI site of pWNS .220 yielded
plasmid pWN5.260A (Figure 30). On plasmid pWN5.260A, the pddABC and dhaT genes

are transcribed in the same direction as the phage T7 promoter on pT7-7.

57

 

PCR gldBC from K. pneumoniae
AT0025955 genomic DNA

1) Sall digest
2) Hindlll digest

 

 

 

Sall Hindlll
[ 1 kb \l
L-
gldBC
1) Sail digest
i 2) Hindlll digest l
3) CIAP treatment

 

,. v ' . ”IF-"~15,”
if pML5.1 76

‘11
6.3 kb ll
7::
’[if‘l
[-111
7’ "1
//
/'
,/’;./
I .< .r/
“#32:”: "

 

Figure 27. Construction of plasmid pML5.176.

58

PCR gldA from K. pneumoniae
ATCC25955 genomic DNA

1) EcoFil digest
2) Sail digest

 

 

 

 

EcoFil 1.7 kb Sail
gldA L

2) Sell digest

1) EcoRl digest
i 3) CIAP treatment i

   

51‘ \ \ [30,0 //

.5 I. /'
. g. I . .
“d"""§.~.' ’Jfl’"%/
«4W4?

Figure 28. Constuction of plasmid pML5.200.

59

PCR dhanrom K. pneumoniae
ATCC25955 genomic DNA

1) BamHl digest

BamHl

7' BamH' 1.2 kb
/ H
dhaT

9. .-//
e2 go

1) BamHl digest
2) CIAP treatment

 

   

’9” PW dhaT

./'
,/

if:

{If pWN5.220A

II 3.7 ,b

 

Figure 29. Construction of plasmid pWN5.220A.

60

PCR pddABC from K. pneumoniae
ATCC25955 genomic DNA

11) EcoRl digest

 

 

 

 

Echl 29 kb EcoRl
pddABC ‘
1) EcoRl digest
2) CIAP treatment
y P77 1:1,“,
"3.1x.
ApR it

‘1 ddABC
pWN5.260A ‘3 p

 

2" 1‘.
1:35 _,‘ 2.,
'rfj,‘ ll .I'
‘5}, III. I
' . I /

1' /‘)/
/ T '
/””'./

Figure 30. Construcion of plasmid pWN5.260A.

61

    
  

'3‘"

«:99

p

  

PCR pddABC- ddrAB from K. oxytoca
dhaT ATCCB724 genomic DNA
/

pWN5.220A l1) EcoRl digest
3.7 kb

 

. EcoFil
2:;

I 5.4 kb
._ ' {no.3}. . .t..= .. s 3 ~‘. . '
\~'\_\K\ _ /,‘rff;:"/V
“'«Z‘r—rLZ-Ii”

 

 

p.

pddABC-ddrAB
1) EcoRl digest
2) CIAP treatment

um."

M'L-"v
I" ',-’t
/f'/ dhaT A
I/"l/ I").
I'll

‘ \
\
4 (334]
wt%»)?
9wa

\ RA§X

ll pWN5.258A

p 9.1 kb

“:1 Ap
pddABC 1\

21’
>
CD
‘17
~\.

D
“as.
N;“ x“

‘~..
N.

\

/
7
‘ .

7*."
(x
\

Figure 31. Construction of plasmid pWN5.258A.

62

The final plasmid pWN5.258A harboring the K. oxytoca pddABC encoding diol
dehydratase, ddrAB encoding reactivating protein, and K. pneumoniae dhaT encoding 1,3-
propanediol oxidoreductase was derived from the same cloning vector pT7-7. Due to
substrate similarity, erythritol might inactivate dehydratases under the same glycerol
inactivation mechanism. This glycerol inactivated dehydratase is known to be reactivated
in the presence of ATP, Mg2+ and the reactivating protein. It was interesting whether diol
dehydratase could be revived in the case when erythritol inactivation exists. A 5.35 kb
fragment containing pddABC and ddrAB genes was amplified from the K. oxytoca
ATCC8724 genomic DNA and subsequently digested with EcoRI. Ligation of this
fragment into the EcoRI site of pWN5.220 led to the formation of plasmid pWN5.258A.
Genes pddABC, ddrAB and dhaT in plasmid pWN5.258A were transcribed in the same

orientation as the phage T7 promoter.

Enzyme assays for in vitro erythritol dehydratase activity

In vitro dehydratase activity was assayed according to the widely applied literature
procedure by Abeles (Figure 32).37 A typical assay mixture was composed of the
substrate of interest, yeast alcohol dehydrogenase, NADH and coenzyme B12 in a
phosphate buffer at pH 8 in a cuvette. A unit of activity equals 1 umol per min of NADH
oxidized at 37 °C. The heterologously expressed dehydratases were first examined by
assaying their activity using erythritol as substrate. The possible oxidation reaction in the
coupled enzyme assay was followed spectrophotometrically by monitored formation of
NAD at 340 nm. Enzyme assays of these dehydratases were carried out in crude cell

lysate without further purification. The catalytic activity of each enzyme towards its

63

native substrates glycerol and 1,2—propanediol was also determined. Catalytic activity
towards erythritol was detected for K pneumoniae GldABC in crude lysate (Table 6).

H20 NADH NAD

HO\/OKH/OH ‘4’ ”ONO LA

a b HO\/\/OH
9'YC8F0l 3-hydroxypropanal 1 ,3—propanediol

OH H20 NADH NAD
a

b
1 ,2-p ropanediol 1 -propana| 1 -propanol

on
H c
O o

4 'n -
H20 3' figgfig‘fxy NADH NAD

OH
HOVIY‘ ‘1’ or M Hoe/C‘DE/xon

OH OH a b
0 1 2 4-butanetriol
erythritol Ho\/U\/\OH ’ '
1.4-dihydroxy-
2-butanone

Figure 32. In vitro dehydratase assays using different substrates. (a) glycerol or diol
dehydratase; (b) yeast alcohol dehydrogenase

Table 6. In vitro dehydratase activities.

 

 

 

specific activity (U/mg)
construct gene(s) _ _
native substrate erythritol

DH5a/pCS120 c. freundii dhaBCE no activity“ no activity

Cl. Pasteurianum 13.5 kb genomic a . _
JM109/pF L1 . . 0.25 no actrvrty

fragment containing dhaBCE
oH5o/pXY39 s. typhimurium pduCDE 034" no activity
W31 10/pML5.200 K. pneumoniae gldABC 0.61 a 0.001
BL21(DE3)/pWN5.260A K. pneumoniae pddABC-dhaT 1.32b no activity
BL21(DE3)/pWN5.258A K. oxytoca pddABC-ddrAB-dhaT 1.45” no activity

 

aglycerol as substrate
b1,2-propanediol as substrate

64

1,2, 4-Butanetriol biosynthesis using erythritol as substrate

The successful identification of erythritol dehydratase activity using K.
pneumoniae GldABC makes erythritol a possible starting material for the biosynthesis of
1,2,4-butanetriol. To further establish this biosynthetic route, it was necessary to identify
all the enzymes responsible to the biosynthetic conversion. Our goal was to construct a
single plasmid capable of synthesizing all the enzymes necessary in the pathway,
ultimately leading to a microbe that synthesizes 1,2,4-butanetriol from erythritol. To this
end, two plasmids were constructed harboring the K. pneumoniae dhaT gene encoding
1,3-propanediol oxidoreductase. A 1.2 kb fragment containing the dhaT gene was
amplified from the K. pneumoniae genomic DNA and subsequently digested with HindIlI.
This fragment was then ligated into the HindIII site of pJF118EH to yield plasmid
pML5.179 (Figure 33). To obtain a construct that harbors both gldABC and dhaT genes,
the HindIII digested dhaT fragment was ligated into pML5.200, which had been
linearlized by HindIII to afford plasmid pML5.226 (Figure 34). Transcription of gldABC
and dhaT was initiated by the pJF118EH-localized tac promoter. The 1,3-propanediol
oxidoreductase activity was determined by a reverse assay proposed by Johnson and Lin.38
The assay mixture contained the substrate of interest and NAD in a bicarbonate buffer at
pH 9. A unit of activity is given as umol per min at 25 °C. In vitro oxidoreductase
activities of these two plasmids expressed in wild-type W3110 are shown in Table 7.
Activities were observed using both 1,3-propanediol and 1,2,4-butanetriol as substrates,
indicating that the oxidoreductase activities were expressed heterologously in both
constructs. Our control experiment with construct E. coli W3110/pML5.200 showed no

detectable activity.

65

BamHl
Smal Sall
EcoRl Hindlll

Ptac

PCR dhanrom K. pneumoniae
ATCC25955 genomic DNA

 
 
 

.2
£531

 

f

l1) Hindlll digest
iaeP pJF118EH i;
5.3 kb

Hindlll Hindlll

.’ .
."rt
, ,i', 3’ o
'.. ‘- 574'
‘1‘ x... 1/1/
'\‘ “\ / -. I
~ \_ /
\ x
-. x _.- I
3“, x“ I 1

- - .. dhal
‘3. ‘ . a I'
~ .3. .
\xy‘ngflifiy _ -’ i ’ "’/
a. .
J

) Hindlll digest
) CIAP treatment

 

Figure 33. Construction of plasmid pML5.179.

66

Hindlll

,J‘xle

34:... PCR dhanrom K. pneumoniae
1 ldBC \

3.1 g ApR

.’/

ATCC25955 genomic DNA

3
9L
\J

h

1) Hindlll digest
1 pM L5 200 ‘ _

Hindlll Hindlll

a l 1.3 kb I
-. .//~'/
P186 ageg‘\ Iacp ' J7,t/./

I“; ’i’i’ yo , / dha T
\Itfitzrv‘firrt'v'rgf/f

1) Hindlll digest
2) CIAP treatment

d \
’../Al": I; t f , XIX: ‘
I ' T‘s-2.5“

" l

,- ”at.
/ / ‘ " ‘3
gldBC I" '4’ \\$:‘1
7' 1r, .
I/ 7' 1')!

hi pML5.226 ii
1 g 9.3 kb

[act0 I?

Figure 34. Construction of plasmid pML5.226.

67

Table 7. 1,3-Propanediol oxidoreductase activities.

 

 

 

specific activity (Ulmg)
construct gene(s) 1,3-propanediol 1,2,4-butanetriol
W3110/pML5.179 K. pneumonia dhaT 4.3 0.01
W3110/pML5.200 K. pneumoniae gldABC no activity no activity
W3110/pML5.226 K. pneumoniae gldABC-dhaT 1.1 0.003 '

 

Previous experiments suggested that K. pneumoniae gldABC encoding glycerol
dehydratase, while coupled with dhaT encoding 1,3-propanediol oxidoreductase, converts
erythritol to 1,2,4-butanetriol. As a second line of evidence, in vitro enzymatic reactions
using erythritol as substrate was setup to verify this finding. Clarified E. coli
W3110/pML5.179, W3110/pML5.200 and W3110/pML5.226 cell-free lysate were
prepared. A total of four reactions were set up under different conditions (Table 8). A
typical enzymatic reaction mixture contained 100 mM erythritol and 0.13 mM coenzyme
B12 in 40 mM phosphate buffer at pH 8 and was incubated overnight with the cell-free
lysate of interest. The first two reactions employed lysate solution obtained from E. coli.
W3110/pML5.200 as a catalyst. The resulting enzyme reaction mixture was directly
subject to sodium borohydride reduction, BSTFA derivatization and analyzed by gas
chromatography for the presence of 1,2,4-butanetriol. 1,2,4-butanetriol was detected in
the gas chromatograrn of reaction 2 with a total yield of 0.004% based on the amount of
erythritol in the reaction mixture. No 1,2,4-butanetriol was detected in the control
reaction 1 without supplemented coenzyme B12. Reaction 3 employed W3110/pML5.226
lysate as a catalyst while reaction 4 was incubated in W3110/pML5.179 and
W3110/pML5.200 lysate mixture in a 1:1 ratio. Both reactions were derivatized using
BSTFA and analyzed by gas chromatography directly without prior treatment. No 1,2,4-

butanetriol was detected in reaction 3. However, to our surprise, 1,2,4-butanetriol

68

formation was observed in reaction 4, which might indicate the lack of either dehydratase
or oxidoreducatase expression in reaction 3. These results provide a proof of concept in
using erythritol as substrate for 1,2,4-butanetriol biosynthesis. The discovery that the K
pneumoniae gldABC gene cluster encodes erythritol dehydratase activity will serve as
parent genes for directed enzyme evolution experiments designed to improve this

dehydratase activity.

Table 8. In vitro enzymatic reactions using erythritol as substrate.

 

reaction lysate reaction conditionsa 1,2,4-butanetriol

 

enzyme reaction (no B12 added) was not detected
1 W3110/pML5.200 _ _
subject to NaBHr reduction

enzyme reaction (with B12) was subject detected
2 W3110/pML5.200 ,
to NaBH4 reduction

3 W3110/pML5.226 enzyme reaction with NADH (0.58 mM) not detected
W3110/pML5.200 and

4 enzyme reaction with NADH (0.58 mM) detected
W3110/pML5.179 (1:1)

 

aAll enzymatic reactions were performed in potassium phosphate buffer (40 mM pH 8),
coenzyme B12 (0.13 mM) and erythritol (100 mM) at rt.

 

 

 

 

 

 

 

 

 

 

 

 

 

4 5 6 7 8 min

Figure 35. GC chromatogram for reaction 1 (control).

69

1 ,2,4-butanetriol

 

 

 

 

 

 

 

 

 

4 5 6 7 8 min

i 124-butanetriol Vt " \J

| /
WLW WW Pr“

r

4 5 6 7 8 min

 

 

 

 

 

 

 

 

Figure 36. GC chromatograms for reaction 2. (upper) reaction 2; (lower) reaction 2
and authentic 1,2,4-butanetriol co-injection.

7O

1 2.4-butanetriol

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4 5 6 7 8 min

1 2.4-butanetriol

«IL not. / In

I 7 fi f v I v v v v I r v v ‘V I v v v f 1

4 5 6 7 8 min

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 37. GC chromatograms for reaction 4. (upper) reaction 4; (lower) reaction 4
and authentic 1,2,4-butanetriol co-injection.

71

In search of 3-deoxy-glycero-pentulosonate decarboxylase genes for the microbial

synthesis of 1,2,4-butanetriol from pentoses

Background

Earlier research efforts in the Frost group elucidated two parallel biocatalytic
pathways that start from D-xylose and L-arabinose to give D- and L-1,2,4-butanetriol,
respectively.39 Each biocatalytic system involves two microbes that synthesize all four
enzymes required for the conversion. Pseudomonas fragi ATCC4973 oxidizes the
pentoses into the corresponding carboxylic acids. Recombinant E. coli.
DH501/pWN6.186A and BL21(DE3)/pWN6.222A catalyze the conversion of D-xylonic

acid to D-1,2,4-butanetriol and L-arabinonic acid to L-1,2,4-butanetriol, respectively.

pWN6. 186A (8.1kb)

 

 

E S S E E
I (l ) ( I ) l l
< —> <— —> _->
(APR) KmR (ApR) lac/Q Pee deC

pWN6.222A (10.1 kb)

 

(H) (S) (S) (39) (39) B B (H) E E (H)
l m i l l l l l l l I
— <—— —> ———> —> ——> <——— —> ————>
(ApR) KmR ApR Pracaatp PT; aadh lac]Q Prac deC

Figure 38. Restriction enzyme map of plasmids. Sites are abbreviated as follow: E =
EcoRI; S = ScaI; H = Hindlll; Bg = BglII; B = BamHl.

Although using whole-cell recombinant microbes to produce 1,2,4-butanetriol has
proven to be the most efficient approach so far, E. coli constructs DH501/pWN6.186A and
BL21(DE3)/pWN6.222A produced 1,2,4-butanetriol enantiomers in unsatisfactory yields

and concentrations. The problem is believed to lie in the decarboxylation step.35’36 To

72

tackle this issue, gene candidates encoding decarboxylase activity towards 3-deoxy-
glycero-pentulosonic acid were identified using bioinformatics tools. Cloning and enzyme
activity assays of the resulting recombinant proteins followed.

Pseudomonas putida ATCC12633 benzoylformate decarboxylase (EC 4.1.1.7) is a
thiarnin—diphosphate dependent enzyme that catalyzes the non-oxidative decarboxylation
of benzoylformate to form benzaldehyde and carbon dioxide. The genes for
benzoylformate decarboxylase, (S)-mandelate dehydrogenase and mandelate racemase are
arranged in an operon (mdlCBA), while the genes for the benzaldehyde dehydrogenases
are independently transcribed (mdlDE). The enzymes in the mandelate pathway allow
both enantiomers of mandelic acid to be utilized as a sole carbon source in
pseudomonads.40 (R)-Mandelic acid is first converted into (S)-mandelic acid, which is
then oxidized to its keto-acid. Benzoylformic acid is decarboxylated to form
benzaldehyde and finally degraded into benzoic acid (Figure 39). Benzoic acid is
subsequently metabolized by the B-ketoadipate pathway and citric acid cycle. The
mechanism of the deC gene encoding benzoylformate decarboxylase catalyzed reaction
has been studied in detail (Figure 40) and the crystal structure of this enzyme is
available.41 In the initial step, the imino group of thiamin-diphosphate abstracts a proton
from the thiazolium ring, resulting in the formation of an ylide. The ylide attacks the
carbonyl of the substrate to form the tetrahedral intermediate 2-a-mandelyl complex of
thiamin-diphosphate. Decarboxylation of this complex results in the enamine
intermediate. Protonation of the enamine provides another tetrahedral 2-hydroxybenzyl-
thiarnin complex. Finally, benzaldehyde is eliminated from the enzyme active site (Figure

40).

73

01°” . or“ . 04°

.II ——+ 'I’ —->

H O“ 011* cor-
‘..

B-ketoadipate Q40 d ©_/<O

pathway 0' H

Figure 39. Mandelate pathway in P. putida. (a) mandelate racemase (mdlA); 0)) (S)-
mandelate dehydrogenase (mdlB); (c) benzoylformate decarboxylase (mdlC); ((1) NAD
and NADP benzaldehyde dehydrogenase (mdlD, mdlE)

H3C>=<R' H30 R'
H30 R' R‘gsrs 9N: \— s: We
(an / H \ R (5.4
R 9‘5 ThDP w
0 'o
H)i\© benzoylformate o\'
benzaldehyde H C R' 81H
3
('9 W.—
R ‘ N \ S 1
O .
0‘) 002

82H HC R
H3C R' 3 '
7=< (9H
R'g')‘ S R'N‘ S
W 3 Ho,.
Ho HC R' O J
I 3 _
e f 1—1— (9

B 3' R ‘ N S 2-a-mandelyl-
2-a-hydroxybenzyl-ThDP | ThDP
Ho’ \E j

enamine

Figure 40. Reaction mechanism for the benzoylformate decarboxylase-catalyzed
decarboxylation of benzoylformate.42 R = 4’amino-2-methyl-5-pyrimidyl; R’ = B-
hydroxyethyldiphosphate.

74

Historically, the first attribute established about an enzyme is the reaction that it
catalyzed. A particular enzyme is either cloned in another microorganism or
overexpressed in its native host. Crude extract is then purified and subjected to activity
assay using the substrate of interest. The purified enzyme can be used to deduce the
amino acid sequence, and subsequently the gene sequence is obtained. The discovery of
3-deoxy-glycer0-pentulosonate decarboxylase activity for 1,2,4-butanetriol biosynthesis
using P. putida mdlC encoding benzoylformate decarboxylase represents an example.
This enzyme was obtained after screening for decarboxylase activities through more than
10 independent decarboxylase candidates.39 Since P. putida benzoylformate
decarboxylase is well studied, the gene sequence was therefore obtained simply by
accessing a gene database.40’41 In recent years, a different path has been established which
follows almost the reverse direction of the conventional route. By now moving from
sequence to activity, there is a higher chance of success with fewer resources, including
less time investment for more hits. Gene sequences, which are often publicly available,
can be accessed by computers. Bioinformatics, which is a discipline combining biology
and information technology, is the essential tool to process these enormous amounts of
information. Gene-mining among the available genome sequences and comparing gene
sequence similarity can be used to identify suitable or interesting open reading frames for
different functions.43

In general, web-based bioinformatics tools were used using links provided through
web-sites such as National Center for Biotechnology Information (NCBI),44 European
Bioinformatics Institute (EBI) 45 and the ERGO Bioinformatics Suite provided by

Integrated Genomics, Inc.46 Using BLAST, one can search either for similar proteins

75

within a database or for new genes similar to the protein sequence at hand.47 The results
are given in a graphic display with the scores in different colors, which is then deciphered
into a hit list. Multiple sequence alignment between sequences of interest can be done
using ClustalW.48 It is particularly useful to compare whole protein/DNA sequences or
domains. Even in the area of structural information, long regarded as one with access for
experts only, there are handy molecular visualization softwares like PyMOL available to
help.49 PyMOL reads in molecular coordinates of atoms in the protein structure from a
.pdb file and these files can be found at the Protein Data Bank.50 With the ability to
visulize an enzyme’s three-dimensional structure, rational protein design provides an

alternative avenue toward an enzyme mutant with improved properties.

Cloning and Screening for 3-deoxy-glycero-pentulosonate decarboxylase candidates

Eight genes bearing high sequence identity with P. putida mdlC gene were selected
using sequence alignment algorithms such as BLAST (Table 9). The strains or genomic
DNA were obtained mainly from ATCC and revived according to the enclosed
instructions. Genomic DNA was extracted from each of the candidates using the standard
cetyltrimethylammonium bromide (CTAB) protocol. PCR reactions using PfizTurbo DNA
polymerase (Stratagene) were employed to amplify the target decarboxylase genes and the
plasmid pJF118EH was used as the cloning vector. Both the vector and the insert were
digested with the suitable restriction enzymes and ligated. The resulting recombinant
plasmids were transformed into E. coli BL21 expression host strain and grown to evaluate

enzyme activities (Figure 41-26).

76

PCR 1.7 kb gene candidate from
P. fluorescens Pf-5
ATCC-BAA477 genomic DNA

in EcoRI and

 

BamHl digests
EcoRl BamHl
L __1 7 kb :4
Pf candidate
1)EcoR|and
l BamHl digests l
2) CIAP treatment

 

Figure 41. Construction of plasmid pML2.118.

77

BamHl
Smal Sail .
EcoR l Hlnd I | I

PCR 1.7 kb gene candidate
from P. aeruginosa

   

ATCC47085 genomic DNA
1) EcoFll and
; BamHl digests
IacP pJF118EH Z
5-3 kb 52? EcoRl BamHl
’ l ,-.__,--1_-z-kb. ..J
\ //// Pa candidate
1) EcoRI and
l BamHl digests l
2) CIAP treatment

 

Figure 42. Construction of plasmid pML2.123.

78

  
   

Hindlll

PCR 1.7 kb gene candidate from
B. fungorum LB400 genomic DNA

1 1) Smal digest

Smal Smal
[ 1.7 kb NJ

I;

Bf candidate

 

tv
I

I

[I
.
. /
.‘v .
x, 4 m. my
"x" ‘- " 5 67'
\. ‘ . W V' l
‘x‘ V. ..'.;.:,jf-.'1 ’b’
- W

1) Smal digest
2) CIAP treatment

Bf candidate :3»EZ3:\

// \fx
,x/ ApR ‘téx
,4 ’f a

   

.4 ’2
[*2/2/
,4 4;;-
k ¢,W_ .‘4 a

Figure 43. Construction of plasmid pML2.162.

79

BamHl

Smal Sail
EcoFll Hind l I l

PCR 1.7 kb gene candidate
from A. ferrooxidans

 

 

 

ATC023270 genomic DNA
1) EcoRI and
BamHI digests
IacP pJF118EH
5.3 kb EcoRl BamHl
1.7 kb
I a
., Af candidate
W71) EcoRI and
l BamHl digests l
2) CIAP treatment

 

Figure 44. Construction of plasmid pML3.040.

8O

BamHl

 

 

Smal Sail .
EcoRI Hindlll
Ptac
1'

a leap pJF118EH El
f" 5'3 kb {.15
figl‘ If?“
\x, /////
V531 '
«(Vi)»_ "’_',:p’,/
~.yc:::.. 5 WW-’ a: Z:/

‘ ‘W 44“.“ ~"‘

1) EcoRI and
i BamHl digests
2) CIAP treatment

PCR 1.7 kb gene candidate
from S. coelicolor
ATCC-BAA471 genomic DNA

1) EcoRI and
BamHl digests

EcoRl BamHl
L 1.7 kb A/J

I

So candidate

1

 

Figure 45. Construction of plasmid pML2.208.

81

BamHl

Smal Sail
EcoFll Hindlll

PCR 1.7 kb gene candidate
from N. aromaticivorans
ATCC700278 genomic DNA

 

 

 

 

 

  

1) EcoRI and
” BamHl digests
IacP pJF118EH $5;
5.3 kb 5, [€le BamHl
1.7 kb ’\/l
3.3. / Na candidate
1) EcoRI and
l BamHl digests l
2) CIAP treatment

 

x'. \‘
‘t-x
\" 73 p
'25 [ac
‘QE‘Q¢»,,_ / I};

. -rl ." ’-
.‘. . by. -.‘..--. A!
‘Ué',’./v,".""r‘r, 7.".’/.'. 23";
fizz/.13}: -z’i-V

Figure 46. Construction of plasmid pML2.214.

82

BamHl

   

 

 

 

Smal San .
EcoRI leldlll
Ptac PCR 1.7 kb gene candidate
_’ ¥‘ from R. palustris
ATCC-BAA98 genomic DNA
\ 1) [5le and
f3, BamHl digests
IacP pJF118EH ‘3
5.3 kb 3:} EcoRl BamHl
\ [ii 1.7 kb [J
“a; 11%/fly Hp candid ate
1) EcoRI and
l BamHl digests l
2) CIAP treatment

 

Figure 47. Construction of plasmid pML2.256.

83

BamHl
Smal Sail
ECOFII Hindlll

PCR 1.7 kb gene candidate
from M. smegmatis
ATCC700084 genomic DNA

1) EcoRI and

 

 

 

 

 

h BamHl digests
1&le pJF118EH $1

5.3 kb )3 EcoRl BamHl

1.7 kb
// Ms candidate
WM?) EcoRI and
l BamHl digests l
2) CIAP treatment

 

': ‘15.
Va; 9" '.f.l.-'.r'
“<74, fen x», . «w r'. 7;,“
4‘“ . . _ .' .‘ .,
m

4"

Figure 48. Construction of plasmid pML2.286.

84

Table 9. 3-Deoxy-gb’cero-pentulosonate decarboxylase candidates.

 

 

plasmid strain Identity to P. putida deC (%)
pWN5.238 Pseudomonas putida ATCC12633 100
pML2.118 Pseudomonas fluorescens ATCC-BAA477 60
pML2.123 Pseudomonas aeruginosa ATCC47085 59
pML2.162 Burkholderia fungorum LB400 61
pML3.04O AcidithiobaciI/us ferrooxidans ATCCZ327O 37
pML2.208 Streptomyces coelicolor ATCC-BAA471 49
pML2.214 Novosphingobium aromaticivorans ATCC700278 37
pML2.256 Rhodopseudomonas palustn’s ATCC-BAA98 40
pM L2.286 Mycobacterium smegmatis ATCC700084 42

 

Site-directed Mutagenesis of benzoylformate and pyruvate decarboxylase gene

Benzoylformate decarboxylase from P. putida and pyruvate decarboxylase from Z
mobilis catalyze the decarboxylation of aromatic benzoylforrnic acid and pyruvic acid,
respectively. In fact, benzoylformate decarboxylase poorly accepts D- and L-3-deoxy-
glycero-pentulosonic acid as substrate. In a 2004 report, Kenyon and McLeish identified
some of the amino acid residues potentially involved in substrate binding and thus in the

specificity of these enzymes. 5'

Among various mutants involved in their study,
benzoylformate decarboxylase mutant A4601 (BFD A4601) and pyruvate decarboxylase
mutant I472A (PDC I472A) are of particular interest and was generated. Both mutants
were successful at increasing the catalytic activities in using C4 — C6 keto-acids as
substrates by manipulating a single amino acid residue inside the active site. In addition
to the two mutants mentioned above, pyruvate decarboxylase amino acid residue I472
were exchanged for glycine, serine and threonine in attempts to either open up the binding

pocket (PDC 14720) or to increase the interactions between the protein and the hydroxyl

groups on the 3-deoxy-glycero-pentulosonic acid substrate (PDC 14728 and I472T) (Table

85

10). These mutants were prepared using PfuTurbo DNA polymerase and the QuikChange
site-directed mutagenesis kit (Stratagene, LaJolla, CA) (Figure 49).52 Wild-type P. putida
mdlC encoding benzoylformate decarboxylase and Z. mobilis pdc encoding pyruvate
decarboxylase were available in the Frost group strain collection as plasmids pWN5.23 8A
and pL0127653, respectively. These plasmids also served as templates for site-directed
mutagenesis experiments without further genetic modifications. The forward primers used
for the mutagenesis are shown in Table 10.

The mutated codons are underlined, and the lowercase letters indicate a base
change from wild-type. The QuikChange methodology (Figure 49) was performed using
PfuTurbo DNA polymerase and a temperature cycler. The oligonucleotide primers
containing the desired mutation, each complementary to opposite strands of the double-
stranded DNA (dsDNA) plasmid template harboring the decarboxylase gene were
extended during temperature cycling by the DNA polymerase. Incorporation of the
oligonucleotide primers generated a mutated plasmid containing staggered nicks.
Following the temperature cycling, the reaction mixture was treated with DpnI, which
digested the parental methylated DNA template. The nicked plasmid DNA containing the
desired mutation was then transformed into E. coli BL21. The transformants were plated
on LB containing ampicillin and incubated at 37 °C overnight. Plasmid was purified from

two single colonies and the mutation was verified by sequencing.

86

A plasmid template
'x' represents the

STEP 1
desired mutation.

Plasmid Preparation

STEP 2

Temperature Cycling oligonucleotide primers

containing the desired mutation.

. ' A' represents the

STEP 3 . .
The PCR reaction IS performed
in a thennocycler.

Digestion

 

' ) The template is digested
with Dpnl.

STEP 4 .
Transformatlon . After transformation. the nicks
in the mutated plasmid is repaired.

Figure 49. QuikChange Site directed mutagenesis.52

87

Table 10. Forward primers used to generate mutations.

 

plasmid mutation fonlvard primers

 

benzoylformate decarboxylase
pWN5.238A wta

pML3.054 A460! 5’-ATGAACAACGGCACgTACGGTatcTTGCGATGG
pyruvate decarboxylase

pL0I27e53 wt"

pM L3.062 I472A 5’-GATCAATAACTATGGTTACACQQQCGAAGTTATGATCCATGATG
pM L3.084 |472$ 5’-TCAATAACTATGGTTACACCAQQGAAGTTATGATCCATGATG
pM L3.086 I47ZG 5’-GATCAATAACTATGGTTACACngQGAAG'lTATGATCCATGATG
pML3. 1 04 I472T 5’-TCAATAACTATGGTI'ACACCA_CQGAAGTTATGATCCATGATG

 

awild-type P. putida mdlC encoding benzoylformate decarboxylase
bwiId-type Z. mobilis pdc encoding pyruvate decarboxylase

Enzyme activities for 3 -deoxy—glycero-pentulosonate decarboxylase candidates

The catalytic activity of the eight wild-type candidates generated using
bioinformatics and the five benzoylformate and pyruvate decarboxylase site-directed
mutants were first examined by coupling the potential decarboxylation reaction with
equine liver alcohol dehydrogenase-catalyzed reduction of 3,4-dihydroxybutanal (Figure
50). Racemic 3-deoxy-glycero-pentulosonic acid was synthesized using a method

described earlier.35

The reactions were followed by monitoring the consumption of
NADH at 340 nm using a UV-vis spectrophotometer. Enzyme assays were carried out in
the cell lysate of E. coli strains that expressed the enzyme candidates (Table 11). As a
second line of evidence, each 3-deoxy-glycero-pentulosonate decarboxylase candidates
was incubated with a reaction mixture of racemic 3-deoxy-glycero-pentulosonic acid,
thiamin diphosphate, NADH and equine liver alcohol dehydrogenase for 24 h. After

removal of the protein, the reactions were subject to BSTFA derivatization and analyzed

by gas chromatography (Table 11). Based on the results of these two assays, eight novel

88

benzoylformate decarboxylases were identified with lower activity than the P. putida
mdlC encoding gene product. Among these decarboxylases, four of them showed
activities toward the non-native substrate 3-deoxy-glycero-pentulosonic acid. It is likely
that BL21/pML2.118, BL21/pML2.123, BL21/pML2.162 and BL21/pML3.040 contain 3-
deoxy-glycero-pentulosonate decarboxylase activities. In contrast, single amino acid
mutant of benzoylformate decarboxylase BL21/pML3.054 (BFD A4601) showed more
than a 25-fold activity decrease in using 3-deoxy-glycero-pentulosonic acid as substrate
(Table 12). The ability of this enzyme to use benzoylformic acid as substrate also dropped
compared to what is reported in the literature.51 All four pyruvate decarboxylase mutants,
on the other hand, were able to synthesize a small amount of 1,2,4-butanetriol, as
indicated by the in vitro enzymatic reaction experiments (Table 12). Although 3-deoxy-

glycero-pentulosonate decarboxylase activities were successfully generated in these

mutants, their activities were low when comparing to the P. putida deC encoding

 

 

 

 

decarboxylase.
NADH NAD
benzoylformate
O _ decarboxylase O \ /
: : OH
@002 'T PP ©/N\ H horse liver @
alcohol dehydrogenase
benzoylformate benzyl alcohol
NADH NAD
pentulosonate
OH O decarboxylase OH O \ /: OH
HOMCOZ- “I’PP , HO horse liver ”ONO...

o. L-deoxy-glycero-
pentulosonate

alcohol dehydrogenase
1 2.4-butanetriol

‘TPP = thiamine pyrophosphate

Figure 50. Coupled enzyme assay for decarboxylase activities.

89

Table 11. Activity table for wild-type pJF118EH clones.

 

 

 

specific activity (U/mg) GC yield of ET”
construct bacterial strain benzoylformate pentulosonatea (%)
DH50deN5.238A P. putida 25.5 0.01 5.3
BL21/pML2.118 P. fluorescens 0.79 0.003 0.5
BL21/pML2.123 P. aeruginosa 0.26 <0.001 0.6
BL21/pML2.162 B. fungorum 0.51 0.009 0.15
BL21/pML3.040 A. ferroxidans 0.01 0.03 0.5
BL21/pML2.208 S. coelicolor 0.03 0.002 <0.01
BL21/pML2.214 N. aromaticivorans 0.003 <0.001 <0.01
BL21/pML2.256 R. palustris 0.005 0.008 <0.01
BL21/pML2.286 M. smegmatis 0.006 0.001 0

 

a3-deoxy-g/ycero-pentulosonic acid (pentulosonate)
b1,2,4-butanetriol (BT)

Table 12. Activity table for benzoylformate and pyruvate decarboxylase mutants.

 

 

 

specific activity (U/mgl GC yield of BT"
construct mutation benzoylformate 2-oxo-hexanoate pentulosonatea (%)
pWN5.238A wtc 26.2 0.2 0.01 5.3
pML3.054 A460l 0.5 0.3 <0.01 0.2
pL01276 wtd 0.03 0.08 <0.01 0
pML3.062 I472A 0.3 10.3 <0.01 0.03
pML3.084 14723 <0.01 0.13 <0.01 0.08
pML3.086 I472G 0.01 1.26 <0.01 0.04
pML3.104 I472T <0.01 0.14 <0.01 0.04

 

a'3-deoxy-glycero-pentulosonic acid (pentulosonate)
b1,2,4-butanetriol (BT)
cwild type P. putida deC encoding benzoylformate decarboxylase

dwild type 2. mobilis pdc encoding pyruvate decarboxylase

9O

Table 13. Amino acid composition of the gene candidates.

 

 

 

 

 

 

. Arg Leu Pro
bacteria occur.a % occur.a "/0 occur."3 % occur.a %
P. putida 23 4.4 50 9.5 25 4.7 40 7.6
P. fluorescens 29 5.5 54 10.2 22 4.2 47 8.9
P. aeruginosa 32 6.1 58 11.0 21 4.0 49 9.3
B. fungorum 29 5.5 51 9.7 28 5.3 41 7.8
A. ferroxidans 23 4.3 54 10.1 37 6.9 42 7.9
S. coelicolor 38 8.0 50 10.5 10 2.1 43 9.1
N. aromaticivorans 35 6.7 50 9.6 19 3.6 39 7.5
R. palustris 29 6.1 47 9.8 16 3.3 47 9.8
M. smegmatis 34 5.4 53 8.5 22 3.5 57 9.1
anumber of occurance (occur)

Table 14. Rare codon analysis of the gene candidates.

, GC content a c d
bacteria (%) Arg Leu lle Pro
P. putida 62.4 9 1 0 4
P. fluorescens 61.2 0 0 0 16
P. aeruginosa 67.1 1 0 0 1 3
B. fungorum 62.3 3 2 0
A. ferroxidans 58.3 6 3 9
S. coelicolor 72.3 0 0 0 19
N. aromaticivorans 62.5 3 12 0
R. palustn’s 64.7 1 0 0 7
M. smegmatis 67.4 9 2 2 23

 

aAGG, AGA, CGA; DCTA; cATA; dccc

91

Codon usage

In a nutshell, four novel 3-deoxy-glycero-pentulosonate decarboxylases were
successfully identified in the previous experiment. Insight into these genes was gained
using bioinformatics tools to evaluate codon usage (Table 13 — 10), compare sequence
identities (Table 15) and search for conserved domains (Figure 51). Analysis of crude
cell-free lysate by SDS-PAGE suggested that protein expression levels of the recombinant
plasmids were low. Codon bias in E. coli is believed to be one of the factors leading to
low heterologous gene expression. Codons that were suspected to be associated with
translational inefficiency in E. coli. were examined (Table 13 — 10). The differences in
codon usage between E. coli and the bacterial strains in our study can account for the
observed low protein expression level. For example, most gene candidates encode a large
amount of proline often found at turns of folded proteins, thus errors in translation of the

associated codons can result in protein misfolding and inclusion bodies.

Amino acid and nucleotide sequence identity

In light of our wild-type decarboxylase activities, use of sequence alignment tools
is inadequate in searching for an active 3-deoxy-glycero-pentulosonate decarboxylase to
replace the P. putida benzoylformate decarboxylase. Advanced protein engineering
methods can provide an answer to those limitations. For example, gene family shuffling54
during which chimeric progeny genes are generated to maximize sequence diversity by
recombining a set of naturally occurring homologous genes is a powerful method. The
resulting shuffled library can be screened for activity towards a particular substrate.
However, such a method requires above 60% nucleotide sequence identities among

shuffled genes. Nucleotide and amino acid sequence identities among the active

92

candidates point toward P. fluorescens, P. aeruginosa and B. fungorum as suitable for us

in gene shuffling experiments (Table 15).

Table 15. Nucleotide and amino acid (in parentheses) sequence identity.

 

P. putida P. fluorescens P. aeruginosa B. fungorum A. ferroxidans

 

P. putida 64% (62%) 66% (63%) 66% (64%) 30% (39%)
P. fluorescens 64% (62%) 81% (82%) 70% (67%) 33% (40%)
P. aeruginosa 66% (63%) 81% (82%) 72% (71%) 33% (40%)
B. fungorum 66% (64%) 70% (67%) 72% (71%) 31% (40%)

A. ferroxidans 30% (39%) 33% (40%) 33% (40%) 31% (40%)

 

Identification of novel benzoylformate decarboxylases

It is noteworthy that the P. putida mdlC gene product is the only benzoylformate
decarboxylase known to date. In the course of our investigation of 3-deoxy-glycero-
pentulosonate decarboxylases with high identity with deC, we identified three new
benzoylformate decarboxylases in the cloned P. fluorescens, P. aeruginosa, B. fungorum
gene products (Table 11). This is all the more interesting as recent publications identified
several important amino acid residues within the active site of P. putida BFD that bind to
the substrate and co-factor thiamine pyrophosphate by a high resolution X-ray crystal
structure.41 Figure 3 shows the multiple sequence alignments of P. putida and the other
four active candidates. P. fluorescens, P. aeruginosa, B. fungorum genes show a perfect
alignment of important residues, which suggests that these genes are likely to be encoding

thiamine pyrophosphate dependent benzoylformate decarboxylases.

93

Pf -MKTVHSASYDILRQQGLTTVFGNPGSNELPFLKGFPEDFRYILGLHEGAVVGMADGFAL 59

Pa eMKTVHSASYEILRRHGLTTVFGNPGSNBLPFLKDFPEDFRYILGLHEGAVVGMADGFAL 59
Bf -MKTVQHAAYEILRRHGLTTIFGNPGSNELPFLKHFPSDFRYILGLHEGVVTGMADGYAQ 59
Pp -MASVHGTTYELLRRQGIDTVFGNPGSNELPFLKDFPEDFRYILALQEACVVGIADGYAQ 59
Af MHVSVKTATFKLLRQLGMTWIVGNPGSTEETFLQDLPEDFTYIQALHEGSVVAIADGLAQ 60
Pf ASGQPAFVNLHAAAGTGNGMGALTNAWYSHSPLVITAGQQVRSMIGVEAMLANVDAPQLP 119
Pa ASGRPAFVNLHAAAGTGNGMGALTNAWYSHSPLVITAGQQVRSMIGVEAMLANVDAGQLP 119
Bf ATGNPAFVNLHSAAGTGNAMGALANAWNSHTPLVVTSGQQVRSTIGMEPLLANVDAVNLP 119
Pp ASRKPAFINLHSAAGTGNAMGALSNAWNSHSPLIVTAGQQTRAMIGVEALLTNVDAANLP 119
Af GLRQAVLVNVRTGVGLGNAMGAILTAYQNKTPLIITSGQQTRDMLLFEPLLTNVNAVTMP 120
120 to 239
Pf CPFPTRHPSFRGVLPAAIAGISRCLADHDLILVVGAPVFRYHQFAPGDYLPAGTELLHIT 299
Pa CPFPTRHACFRGVLPAAIAGISRLLDGHDLILVVGAPVFRYHQFAPGDYLPAGAELVQVT 299
Bf CSFPTTHACFRGVLPAGIASISRLLDGHDLILVVGAPVFRYHQYEPGALLPAGAELVSIT 299
Pp CPFPTRHPCFRGLMPAGIAAISQLLEGHDVVLVIGAPVFRYHQYDPGQYLKPGTRLISVT 299
Af AIFPCSNSLYAGILPAAIAPLADCLRGHDLAIVVGAPVFRYYPYVAGHYLPEGTSLLQIT 297
300 to 418
Pf E----RRVIGVIGDGSANYGITALWTAAQYQIPVVFIILKNGTYGALRWFAGVLQVSDAP 474
Pa R—~--RQVIGIIGDGSANYGITALWSAAQYRVPAVFIILKNGTYGALRWFAGVLEVPDAP 474
Bf G-e--RRVIGIIGDGSANYGITALWTAAQYSIPTIFIIMKNGTYGALRWFAGVLGVEDVP 474
Pp E----RQVIAVIGDGSANYSISALWTAAQYNIPTIFVIMNNGTYGALRWFAGVLEAENVP 474
Af ISGRHRPVIVFVGDGAFNYAPQCIYTGVRHHTHVIFVVLQNHEYAILKEFAIEEKLKNIP 477
Pf GLDVPGLDFCAIGRGYGVHSVQANTREAFAQALSEALAGDRPVLIEVPTLTIEP---- 528
Pa GLDVPGLDFCAIARGYGVEALHAATREELEGALKHALAADRPVLIEVPTQTIEP---- 528
Bf GLYVPGIDFCALARGYGVEALHADSGASLTVALERALSSSRPTLIEVETLA ------- 525
Pp GLDVPGIDFRALAKGYGVQALKADNLEQLKGSLQEALSAKGPVLIEVSTVSPVK---- 528
Af GLELPGIAIADIGSAYGAHATVAITAIQLEAAFREAMGYQGVSVLQVPISSDLHPLID 535

Figure 51. Multiple sequence alignment using ClustalW algorithm. Important amino
acid residue are highlighted (Pp = P. putida, Pf = P. fluorescens, Pa = P. aeruginosa, Bf =
B. fungorum, Af = A. ferrooxidans).

94

Directed evolution of 3-deoxy-glycero-pentulosonate decarboxylase for microbial

1,2,4-butanetriol synthesis

Background

Created biosyntheses of 1,2,4-butanetriol from pentoses involves four enzymatic
conversions and were designed based on Pseudomonas fiagi D-xylose, L-arabinose
catabolism and E. coli D-xylonic acid catabolism. The enzymatic decarboxylation of 3-
deoxy-glycero-pentulosonic acid to 3,4-dihydroxybutanal using Pseudomonas putida
mdlC gene-encoded benzoylformate decarboxylase is known to play a critical role in the
final 1,2,4-butanetriol yield and concentration. However, this decarboxylase showed
unsatisfactory catalytic activity towards the non-native substrate 3-deoxy-glycero-
pentulosonic acid. To this end, searching for a more active 3-deoxy-glycero-
pentulosonate decarboxylase has been a central focus of efforts to improve the pentose-
based microbial synthesis of 1,2,4-butanetriol. A variety of strategies had been employed
targeting this goal. In the earlier section, bioinformatics searches on available wild-type
bacterial genomes yielded eight decarboxylase candidates. Among them, Pseudomonas
fluorescens, Pseudomonas aeruginosa and Burkholderia fimgorum were confirmed to
encode 3-deoxy-glycero-pentulosonate decarboxylase activity. The enzyme activity of
these novel decarboxylases was found to be lower than the deC gene encoding
benzoylformate decarboxylase from P. putida. However, these genes appeared to be
shared a high degree of nucleotide sequence identity (Table 15). An alternative strategy to
take advantage of this fact is to evolve a better decarboxylase for microbial 1,2,4-

butanetriol synthesis using DNA family gene shuffling.

95

Since its first application about a decade ago, directed evolution has rapidly
become a powerful tool used in protein engineering research.55 This methodology allows
significant improvement of enzyme performance without prior knowledge of its structure
or mechanism. In addition, by studying the mutants obtained from directed evolution
experiments, one can gain a better understanding toward fundamental questions related to
protein folding, substrate binding, and to identify important amino acid residues of an
enzyme. 56 A previous attempt to evolve a better 3-deoxy-glycero-pentulosonate
decarboxylase in the Frost group using error-prone PCR successfully isolated an enzyme
mutant with a two-fold activity improvement.35 It is generally believed that directed
evolution of a single parent gene would be limited by inadequate sequence diversity.54 To
circumvent this issue, directed evolution employing single-stranded DNA (ssDNA)
(Figure 52) and double-stranded DNA (dsDNA) family shuffling have been employed
using P. putida deC and the newly identified 3-deoxy-glycero-pentulosonate
decarboxylase genes from P. fluorescens, P. aeruginosa and B. fimgorum. DNA family
shuffling represents an alternative or additional approach for generating genetic diversity
based on the mixing and concatenation of genes from a number of parent sequences with
high nucleotide sequence identity.54 As compared to random mutagenesis, DNA shuffling
is advantageous in concentrating beneficial mutations that have risen independently and
subsequently translating into preferred enzyme mutant properties. These four parent
genes were randomly fragmented and recombined to generate a chimera library. A 96-
well format high-throughput screening method to assay in vivo activity of decarboxylase
chimeras using intact E. coli microbes was also developed. In the end, 4,000 independent

mutant clones were screened.

96

 

 

 

 

 

 

 

 

 

 

 

 

3= r®'

; <9

- ®
®——l—_: ®——:
®=i ®—————:
_1 ®——i

 

¢ Individual lambda exonuclease digestion

 

+++

 

 

 

 

 

 

 

 

1 Mix in equal ratio

 

+

 

 

DNase digestion

$ Fragments recombination

 

 

l Amplification

 

 

 

 

 

 

Figure 52. Schematic diagram for single-stranded DNA family shuffling.

97

Subcloning 3-deoxy-glycero-pentulosonate decarboxylase encoded genes

The EcoRI and BamHI restriction sites were eliminated from the B. fungorum
decarboxylase gene through a QuikChange site-directed mutagenesis kit used to facilitate
the chimera library. This double mutant was prepared using PfuTurbo DNA polymerase
with the following oligonucleotide primers: 5’-CCAACGC-
GTGGAACTCGCATACGCCG-3’ and 5’-GCGGCGGCGGGCATCCAGCTCGCG-3'.
The EcoRI and BamHI sites are italicized and the mutations are in boldface. The
mutations were designed such that the encoded amino acids remain unchanged. The
plasmid that harbors this B. fungorum mutated gene was designated pML3.200.

The decarboxylase genes from P. fluorescens, P. aeruginosa and B. fungorum
were amplified by PCR from plasmids pML2.118, pML2.123 and pML3.200,
respectively. Only the ORFs were amplified from these new plasmids and they shared the
same 20 nucleotides at both ends of the genes. The inclusion of a HindIII restriction
enzyme site on the sense primer and an EcoRI restriction enzyme site on the anti-sense
primer allowed digestion of the PCR product with both restriction enzymes and ligation of
the digestion product into vector pJF118HE. These gene fragments were subcloned into
pJF118HE to yield plasmids pML3.224 (P. fluorescens, Figure 53), pML3.225 (P.
aeruginosa, Figure 54) and pML3.226 (B. fungorum, Figure 55). As a consequence, the
inserted DNA could only be located into the vector in a defined orientation. This
directional cloning strategy ensured that the insert gene would be expressed off a tac
promoter located upstream from the Hindlll recognition site on pJF118HE. Due to the
expression of a plasmid-encoded Lac repressor protein, the tac promoter initiated

transcription of mutant genes in different clones could be regulated by IPTG addition.

98

, r». [‘1‘
' 1." “New?“

     

BamHl

Hinduisa" smalfcofil

 

1) PCR gene ORF
2) Hindlll and
EcoRI digests

EcoRl

..\\

 

l
5.3 kb ,2; Hindlll
| 1.7 kb >|

/‘ y
L
j.

" Pf candidate

...s¥‘§§§§\&\mx

v
\\.
\

EcoRl digests

WI?) Hindlll and ¢
2) CIAP treatment

. ,x '5'“; ”h,
Pf candidate we» ”3 ~43,
//'/ A\nf’i\
/ /‘ p \\
\\ tx
\( ‘ 1'

It
7.0 kb )t
Ptac‘ ,}

 

Figure 53. Construction of plasmid pML3.224.

99

"L

BamHl

Hindlllsan sma’EcoRI

 

l '.. h
«'1‘ r "
gqrf.x\‘,~.'~’
<

, mg.»-
uu, , '..‘.;.s‘»*

1) PCR gene ORF
2) Hindlll and

ECOR' digests
Hind||| ECORI
I17 kb; ’-
,//;,,

,/ Pa candidate
1) Hindlll and
l EcoFil digests

2) CIAP treatment ¢

g
.b
'o
c.
'n
-L
A
on
I
ITI

5.3 kb

3.
‘g..x~" . V
‘Q§\ ‘t\

0 \
‘<5:_-i;7'.~

 

- ,»”""t<.7f-
Pa candidate/.337,»m»- “-

K‘?‘\\
/ \ X
/ x
/ Ap\RA “x.
x' /"
,I

/ ”3..

‘ ’ pML3.225 1

7.0 kb 5??

Ptact {,1

 

Figure 54. Construction of plasmid pML3.225.

100

J‘ . -— .,
. 1",»:q/_:-__..<c72;7’
Bf candldate/j:::l,,/ «71:2: 1,1”
/// APR “"3
/ \m
BamHl / 1
San Smal / pML3.200

l' ‘1
Hindlll EcoRl P 7.0 kb )1
\{ ’ac

”y

 

Ptac

 

‘ , 1 I _
.43 .
[’3 ”1.1. ‘r,/
""127 P'”; 'x'-
//>“ .49”

    
 

, :21 1) PCR gene ORF
"if ‘69 2) Hindlll and

IacP pJF118HE EcoRl digests

’_’ .xt‘.‘ .\\\ \._ 1 ~
y“ .ua— N3. ~
‘ . t ‘3 “.‘\‘\‘\~“.\\
~‘~‘\‘\'I“\A‘Clh-’

 

 

 
 

, 5-3 kb Hindlll EcoRl
‘\ ['5/ l I“
‘..‘~.“\\ ’77 L’-
h \- /./‘ 0
"fix M Bf candidate
«er 4&2 '

EcoFil digests

1 1) Hindlll and 1
2) CIAP treatment

 

Figure 55. Construction of plasmid pML5.226.

101

Chimera library construction for ssDNA family shuffling

The decarboxylase parent genes were amplified from plasmids pML3.224,
pML3.225 and pML3.226, respectively using suitable phosphorylated primers (Figure 52).
5 pg of each purified PCR product was subject to lambda exonuclease digestion to yield
the desired upper or lower ssDNA. 0.5 pg of each purified ssDNA was then mixed and
digested with DNase I for 30 sec. The 500 - 1,000 bps DNA fragments were gel-purified
and recovered using spin columns. The gel purified fragments were subsequently
subjected to primerless DNA fragment recombination using a thermocycler. This
reassembly reaction mixture served as a DNA template for the subsequent normal PCR
reaction. As a final step, the full length PCR product was purified, digested by Hindlll
and EcoRI and then cloned into pJF118HE to generate the chimera library. This chimera
library was transformed into E. coli W3110 for sequence analysis.

Twenty clones were randomly chosen from a 96 well plate and submitted for high-
throughput sequence analysis at the Genomics Technology Support Facility at Michigan
State University, using the following primers: 5’-CGGCTCGTATAATGTGTGGA-3’, 5’-
TCCATTCCCTACGACGACTG-3’, 5’-GATGCGCTGGTGGGAGACAT-3’, and 5’-
TCGGCCAACTACGGTATCAC-3’. These primers were designed such that the
sequencing results would cover the entire gene. Comparison of the nucleotide sequence of
20 shuffled pentulosonate decarboxylase genes showed the unsuccessful recombination
among P. fluorescens, P. aeruginosa and B. fimgorum genes (Figure 56). These chimeras
had an average number of crossover equal to one, where B. fungorum gene fragments
rarely ended up into the hybrid genes. In addition, the chimera library was heavily biased

towards P. fluorescens at the 5’-end, and P. aeruginosa at the 3’-end. There were several

102

possible explanation for these results. Family shuffling with large parent gene sizes (>1.5
kb) with low nucleotide homology (< 70%) was generally agreed upon to be problematic.
The use of large ssDNA fragments (500 — 1000 bps) in the recombination step was
certainly another factor. Attempts to generate the chimera library using smaller fragments

was unsuccessful, leading us to explore dsDNA family shuffling as an alternative.

 

._.l.._‘.l .g. ‘..:7 ..- .- 6",- n 5': ‘r_‘;.;— _..':“ ‘_t‘_,. .;._’.. ‘..3- tax, 1
v 7" "VZ -. ‘ - _V ‘.. - 'r~‘- ..
9338,35: JFL'Z-s «is»: '-‘5:.‘:v”-1'.!"‘fi“ “Em?
_ Eat-a: mm. assumes:
1. | I " "j "n'.’." . .. i "U“, I n.7,," L , c, H
- - --~.---s J3. .- --..-— --. 1:; . L» 4,“

 

 

Pa: _ Bf; _

Figure 56. Sequencing results of 20 ramdomly chosen clones from ssDNA family
shuffling. Pf = P. fluorescens; Pa = P. aeruginosa; Bf = B. fungorum

Chimera library construction for dsDNA family shuffling

dsDNA family shuffling is highly similar to the ssDNA family shuffling method,
however, the use of phosphorylated primers and lambda exonuclease digestion are not

required. Joern, Arnold and coworkers reported the protocol to generate a chimera library

103

for dsDNA family shuffling. 57 A 2.1 kb gene was successfully recombined with an
average crossover of 3.7. Given their success, a chimera library of decarboxylase genes
from P. putida, P. fluorescens, P. aeruginosa and B. fungorum was generated.

The four parent genes were amplified using a pair of non-phosphorylated primers.
Purified PCR products (1 ug) from each microorganism was mixed and digested with
DNase I for 4 min. The 100 — 300 bps DNA fragments were gel-purified and recovered
using spin columns. These fragments were subsequently recombined in a therrnocycler
(Figure 57, left). This reassembly reaction mixture (5 uL) was employed as a DNA
template for the subsequent PCR reaction (Figure 57, right). The full length 1.6 kb PCR
product was purified, digested by Hindlll and EcoRI, and ligated into plasmid pJF118HE.
This chimera library was then transformed into E. coli W3110 for sequence analysis and

in vivo enzyme activity screening.

123456789101112

s..-

M:fififi3wfififi“‘

   

Figure 57. DNA agarose gels for dsDNA family shuffling. (left) primerless
recombination reaction, lane 1 and 2: DNA size markers, lane 3: recombination reaction
sample; (right) normal PCR for full length gene, lane 1 and 2: DNA size marker, lanes 3 —
7: PCR reaction at different annealing temp. 51.2 °C, 53.6 °C, 56.4 °C, 58.2 °C, 60.0 °C,
respectively, lanes 8 — 12: replicates of lane 3 — 7.

22 clones were randomly chosen from a 96-well plate and submitted for high-

throughput sequence analysis. The results are summarized in Figure 58. In these 22

104

sequenced clones, the highest crossover obtained was 7 with an average of 3.4. The
diversity of a chimera library depends on the number of crossovers, and more crossovers
are expected when participating genes share a high level of sequence identity. This
explains the gene fragment distribution shown in Figure 58. P. fluorescens and P.
aeruginosa genes share up to 82% sequence identity, the fraction of crossover occurring
between them were expected to be more frequent. On the other hand, the P. putida gene
only shares 64% identity to the rest, leading to the fact that the fraction of crossovers was

relatively low.

 

 

 

Figure 58. Sequencing results of 22 ramdomly chosen clones from dsDNA family
shuffling. Pp = P. putida; Pf = P. fluorescens; Pa = P. aeruginosa; Bf = B. fungorum.

105

Chimera library screening for 3 ~deoxy—glycero-pentulosonate decarboxylase

Designing a reliable and efficient methodology for the identification of active
decarboxylase mutants possessing improved catalytic activity towards 3-deoxy-glycero-
pentulosonic acid is of premier importance in this study. Decarboxylase activities of the
mutants were screened in vivo by cultivating E. coli W3110 transformants of the chimera
library in minimal salts medium containing D-xylonic acid. D-Xylonic acid was
synthesized according to literature procedure58 and served both as a carbon source for
growth as well as substrate for the potential decarboxylase. The decarboxylase activity
was measured by the amount of D-l,2,4-butanetriol accumulated in the culture medium
based upon a response factor using gas chromatography.

Screening of the 3-deoxy-glycero-pentulosonate decarboxylase chimera library
was carried out in a 96-well format. Single colonies from the chimera library were
inoculated and cultured in a 96-well grth block. After 36 h of incubation at 37 °C, a
glycerol freeze plate was made and stored at -80 °C.. Simultaneously, a fresh grth block
was inoculated using a 10% seed culture from the first block and incubated for another 24
h. Protein expression was induced with IPTG, and the cells were harvested by
centrifugation after 12 h. 100 uL of clarified culture medium was sampled from each well
and evaporated to dryness in a CentriVap (Labconco) under reduced pressure. The residue
was derivatized with a reagent mixture containing bis(trimethylsilyl)trifluoroacetamide
(BSTF A). 1,2,4-Butanetriol was quantified by gas chromatography with dodecane as an
internal standard.

Chimera library screening of 3,960 independent mutants was carried out in a 96-

well format using the method previously described. In general, around 90% of the

106

mutants were found to be inactive, while the rest of the mutants carried genes with 3-
deoxy-glycero-pentulosonate decarboxylase activity lower than our benchmark,
benzoylformate decarboxylase from P. putida. In order to determine whether the failure
to isolate better mutants was due to possible experimental errors from our screening
method, a secondary screening using the same methodology was carried out and the
results were compared to the primary screening. Five active mutants were picked
arbitrarily based on the results obtained in the primary screening and were screened again
(Table 16). The activity fold between the two screenings indicated a maximum error of
i0.3, so therefore it was unlikely that potent mutants would have been missed due to poor
assay design. Since over 90% of the mutants were assayed as inactive clones, testing for
the possibility of false-negative results was also of particular importance. Seven mutants
were randomly screened accordingly (Table 17). This experiment showed zero activity
for all mutants toward D-1,2,4-butanetriol synthesis, which was consistent to the results
obtained in the primary screening.

Consistency of cell growth inside a 96-well block was found to be problematic,
which was believed to be the consequence of inadequate aeration inside then incubated
shaker. This became more obvious during screening experiments involving a prolonged
incubation period of six days. Putting spacers between blocks and leaving space between
stacks of blocks were measures taken to facilitate aeration. At the end of the screening,
eight mutants showed relatively high D-1,2,4-butanetriol formation with low cell density
and were subject to a second screen. All eight candidates showed normal cell growth in
this second screening experiment and the results are shown in Table 18. Except 21G5, all

other mutants showed lower activity relative to P. putida MdlC.

107

Table 16. Data reproducibility using in viva GC screening assay.

 

 

 

cell density activity folda
entry mutant ODsoo fireen 2"“ screen
1 16010 1.67 0.9 0.9
2 18A5 2.67 0.4 0.5
3 19F3 3.39 0.9 0.7
4 2802 2.64 0.9 0.6
5 2906 2.17 0.9 1.1

 

“activity fold = mol of 1,2,4-butanetriol produced with mutant/mol of 1,2,4-butanetriol produced with
W31 10/pWN5.238A (control)

Table 17. Test for false-negative results.

 

 

 

cell density activity folda
entry mutant ODsoo 1§i screen 2"?I screen
1 29A3 2.53 0 0
2 29E6 2.41 0 0
3 30A2 3.05 0 0
4 30E9 1.70 0 0
5 32A2 1.49 0 0
6 3235 2.75 0 0
7 32E12 3.07 0 0

 

“activity fold = mol of 1,2,4-butanetriol produced with mutant/mol of 1,2,4-butanetriol produced with
W31 lO/pWN5.238A (control)

Table 18. Test for mutants with growth issues.

 

 

 

cell density activity folda
entry mutant ODsoo fi screen 2"“ screen
1 2165 2.76 3.5 1.3
2 22G2 2.50 1.7 1.0
3 26H5 3.04 2.8 0.5
4 27A9 2.18 1.5 0.4
5 2709 2.16 2.0 0.8
6 2837 2.63 1.1 0.9
7 29H11 2.06 2.2 0.6
8 31G8 2.40 1.5 0.4

 

“activity fold = mol of 1,2,4—butanetriol produced with mutant/mol of 1,2,4-butanetriol produced with
W31 lO/pWN5.238A (control)

108

Mutants 29D6 and 21G5 showed activity fold increases of 1.1 and 1.3,
respectively. They were subject to a third round of screening for D-1,2,4-butanetriol
production in baffled flasks. Two culture tubes of 5 mL M9/D-xylonate/Ap cultures were
inoculated with a single colony of each mutant and allowed to grow overnight at 37 0C.
They were used as seed culture to inoculate 100 mL M9/xylonite/Ap medium in 500 mL
Erlenmeyer flasks. The experiment was conducted at 30 °C with agitation at 250 rpm.
After 24 h, 1 mL of culture was sampled from each flask. Clarified culture medium was
subsequently evaporated to dryness, subject to BSTFA derivatization and finally assayed
by gas chromatography. Control constructs E. coli W3110/pWN5.23 8A,
W3110/pML3.224, W3110/pML3.225 and W3110/pML3.226 showed reproducible
activities (Table 19). The activity fold for mutants 29D6 and 21G5 were found to be 0.4
and 0.8, respectively. To understand the structure of these chimera decarboxylase genes,

all the mutant candidates that entered the secondary screening were sequenced.

Table 19. Shake flask experiments for improved decarboxylase activities.

 

 

cell density . _ a
entry construct 00600 actiwty fold
1 W3110/pWN5.238A (Pp) 4.2 1.0
2 W3110/pML3.224 (Pf) 4.0 0.4
3 W3110/pML3.225 (Pa) 4.1 0.5
4 W3110/pML3.226 (Bf) 4.0 0.2
5 W3110/21G5 4.0 0.8
6 W3110/29D6 4.1 0.4

 

“activity fold = mol of 1,2,4-butanetriol produced with mutant/mol of 1,2,4-butanetriol
produced with W31 10/pWN5.23 8A (control)

109

Sequence alignment of mutants from the secondary screening is shown in Figure
59. The average numbers of crossovers for active and inactive clones are 4.5 and 6.2,
respectively. Gene fragments from P. putida were not incorporated in any of these
mutants, presmnably due to the relatively low nucleotide identity (around 64%) to the
other candidates. B. fungorum fragments were found in most inactive mutants, suggesting
that these fragments might be responsible to the high percentage of inactive clones
generated (>90%). Mutants 29D6 and 21G5 were found to have the same amount of
crossover (equals 2) and share a similar gene recombination pattern. Both of them contain
a small P. fluorescens gene fragment in the middle, flanked by P. aeruginosa fragments.
Interestingly, the absence of P. putida gene fragments in mutant 21G5 suggested its
decarboxylase activity was solely inherited from P. fluorescens and P. aeruginosa genes,
which in fact resulted in a 1.6-fold activity improvement relative to either parent genes.
Nonetheless, this improvement only generated another decarboxylase with activity
comparable to the benchmark P. putida MdlC.

In summary, P. putida, P. fluorescens, P. aeruginosa and B. fungorum were
identified to bear pentulosonate decarboxylase encoded genes that were suitable for DNA
family gene shuffling. A DNA chimera library of these decarboxylase genes were
successfully generated with high numbers of crossover (average crossover = 4.5) using
double-stranded DNA family shuffling techniques. Library screening of 3,960 mutants
were conducted. Identification of the mutant 21G5 with at least 1.6 fold activity
improvement compared to its parent genes was achieved. However, the activity of mutant

21G5 decarboxylase was similar to the P. putida MdlC. Nevertheless, the success in

110

generating a diverse mutant library and the valuable experience gained from this study

allows us to revisit the task in the future once high-throughput robotics become available.

 

 

 

 

 

 

 

 

*- ps= van am. urea-.1 nmxwrnrim .3 7 ----—
2837 ENE-Fm...“ .-........ iii-i .......-........ ”ra—
. “‘ xll J U 7’ H ‘ ‘ ’17- ‘ Jun- I ‘ i A.

 

 

 

 

 

 

 

 

 

Inactive

 

pp- 95- pa-— 3?—

Figure 59. Sequence alignment of selected mutants. Pp = P. putida; Pf = P.
fluorescens; Pa = P. aeruginosa; Bf = B. fimgorum.

111

Bacteriophage T7 promoter and E. coli chaperones for deC expression

Background

F ermentor-controlled synthesis of D-1,2,4-butanetriol using a single recombinant
E. coli W3110 derivative WN13/pWN7.126B was scaled up to 100 L at MBI
International.59 A 500 g quantity of D-1,2,4—butanetriol was successfully isolated as a
clear oily liquid using this process. Byproduct formation during D-1,2,4—butanetn'ol
biosynthesis provided a guide for what aspects of this microbial catalyst needed to be
improved. As shown in Figure 10, although 6 g/L in 28% yield (mol/mol) of D-l,2,4-
butanetriol was synthesized at l L working volumes by E. coli WN13/pWN7.126B under
fermentor-controlled conditions, substantial concentrations of other byproducts are also
synthesized. These byproducts included 3-deoxy-D-glycero-pentanoic acid (5 g/L), 4,5-
dihydroxy-threo-L-norvaline (4 g/L), 3-deoxy-D-glycero-pentulosonic acid (1 g/L), and
3,4-dihydroxy-D-butanoic acid (3 g/L). One of these byproducts (3-deoxy-D-glycero-
pentulosonic acid) was likely the precursor to two of the other products (3-deoxy-D-
glycero-pentanoic acid and 4,5-dihydroxy-threo-L-norvaline).36 Reduction of byproduct
formation would not only increase the yield and concentration of D-1,2,4-butanetriol
production, but would also substantially lower the cost currently associated with
downstream purification of product from the fermentation broth.

In the previous section, a large effort towards the discovery of novel 3-deoxy-
glycero-pentulosonate decarboxylases was discussed. By increasing the rate of conversion
of 3-deoxy-D-glycero-pentulosonic acid into 3,4-dihydroxy-D-butanal, three of the four

byproducts observed during the synthesis of D-l,2,4-butanetriol from D-xylose would

112

likely be eliminated. While this approach would undoubtedly be effective, alternative
approaches will be discussed in this section.

The plasmid-borne P. putida mdlC gene encoding benzoylformate decarboxylase
in our benchmark E. coli microbe WN13/pWN7.126B is transcribed based on a tac
promoter system. Another commonly used system for achieving high level of protein
production depends on the extremely selective nature of the bacteriophage T7 RNA
polymerase for specific promoters.”61 To investigate the effect on using a strong T7
promoter, the P. putida mdlC gene was cloned into pET28c(+). The resulting plasmid was
introduced into a DE3 lysogen of E. coli WN13 to generate E. coli KIT15/pML7.135.
However, culturing this E. coli microbe resulted in reduced D-1,2,4-butanetriol
production. Co-expression of E. coli chaperonins GroEL-GroES in this E. coli strain was
subsequently pursued in an attempt to restore proper protein folding and assembly of this
heterologously expressed P. putida benzoylformate decarboxylase. E. coli strains

WN13/pWN7.180 and WN13/pWN7.202 were generated and evaluated.

Expression of plasmid-borne deC gene under phage T7 promoter

Often the first step to improve plasmid-localized enzyme expression in E. coli is
the use of a strong and controllable expression vector system. Protein expression vectors
that utilize the bacteriophage T7 polymerase/promoter system are capable of very high
levels of protein production. In the pET vector system (Novagen), the gene encoding T7
RNA polymerase is usually supplied by the host bacterial cell in the form of a A lysogen

that expresses the polymerase gene under control of the IacUVS promoter.60

113

’I~ . 4
/ ‘

/.-/fndlC
88" l" i” "ii“.

Hindlll EcoRl ’ "l
Xhol amHl 7.0 kb 1;

 

 

I .J
!.
."
l I
I I
I} ,i
('.i/
,1 1
.277"
9‘2’1'771‘ A //
x" ‘ 14“”- fink .1 xv"
r/' 1". 733%,, /»,"’V’
I. '/l > I
1' r." ,' , I
’3‘ P77 ”6?. WW “.1
. a, 22-.---.
'x’/u
V}
3175
$2.1
‘:

pET28c(+) H 1) EcoRl digest
5-4kb Iale i”
I

7 EcoRl 1 6 kb EcoRl
1”? L . 1d
//./

/ /
_ ,1

ftp '1 , , deC
j 1) EcoRl digest
)

 

 

I
1'

2 CIAP treatment

 

Figure 60. Construction of plasmid pML7.128.

114

45.12”,” ~ plasmid thC1.558

~/ / ‘\,."\,

’x’,” R mdlC'Qx
57 Km \ -
xii \‘x 1) Smal digest

:7 pML7.128 ll

 

1.7 kb
‘K \\ (Aq v , . . . .... . . . a. . v, . .. -. r... . I
‘4

 

serA

 

1) Smal digest
2) CIAP treatment

 

Figure 61. Construction of plasmid pML7.135.

115

The target gene, P. putida mdlC gene was encoded on plasmid pET28c(+) so that
transcription could be driven by the T7 RNA polymerase. A 1.6 kb fragment encoding the
P. putida mdlC gene was excised from plasmid pWN5.23 8A by digestion with EcoRI and
ligated to the 5.4 kb plasmid pET28c(+), which had been previously treated with EcoRI to
afford a 7.0 kb plasmid pML7.128 (Figure 60). A 1.6 kb serA locus was excised from
plasmid pRC1.55B by digestion with Smal and ligated to the Smal digested plasmid
pML7.128. The ligation mixture was transformed into E. coli WN13. Transforrnants
carrying the serA insert were selected on M9 medium plates. In the resulting 8.7 kb
plasmid pML7.135, the serA and deC genes are convergently transcribed (Figure 61).
The DE3 lysogen E. coli KIT2 was prepared using the DE3 lysogenation kit of Novagen
according to the manufacturer’s protocol.62 E. coli host strain KITIS was constructed and
retransformed with pML7.135. The resulting E. coli KIT15/pML7.l35 was evaluated
under fermentor-controlled conditions.

Fed-batch fermentations were performed in a 2.0 L working capacity Biostat MD
B. Braun fermentor equipped with a DCU-3 system and a Dell Optiplex Gs+ 5166M
personal computer utilizing B. Braun MF CS/W in sofiware (v.2.0) for data acquisition and
automatic process monitoring. Temperature, pH, and glucose feeding were controlled by
PID control loops. Temperature was maintained at 36 °C by temperature adjusted water
flow through a jacket surrounding the fermentor vessel. The pH was maintained at 7.0 by
the addition of concentrated ammonium hydroxide or 2 N H2804. Dissolved oxygen
(D.O.) was measured using a Mettler-Toledo 12 mm 0; sensor fitted with an Ingold A-

type 02 permeable membrane. D.O. level was maintained at 10% of air satuation.

116

Antifoam (Sigma 204) was added manually as needed. Fed-batch fermentations were run
in duplicate, and reported results represent an average of the two runs.

Inoculants were started by introduction of a single colony picked from an M9 agar
plate containing glucose into 5 mL of M9 glucose medium. Cultures were grown at 37 °C
with agitation at 250 rpm until they were turbid and subsequently transferred to 100 mL of
fresh M9 glucose medium. After culturing at 37 °C and 250 rpm for an additional 10 h,
the inoculants (OD600 = 3.0-4.0) were transferred into the fermentation vessel and the
batch fermentation was initiated (t = 0 h). The initial glucose concentration in the
fermentation medium ranged from 24 to 28 g/L according to the growth requirement of
different constructs.

Cultivation under fermentor-controlled conditions was divided into three stages
with each stage corresponding to a different method for controlling D.O. In the first stage,
the airflow was kept constant at 0.06 L/L/min and the impeller speed was increased from
50 rpm to 950 rpm to maintain the DO. at 10%. Once the impeller speed reached its
preset maximum of 950 rpm, the mass flow controller started to maintain the DO. by
increasing the airflow from 0.06 L/L/min to 1.0 L/L/min. Depending on the construct
under examination, the two stages took a total of 12 — 15 h for completion. At constant
impeller speed and constant airflow rate, the DO. level was maintained at 10% for the
remainder of the fermentation by use of an 02 sensor to control glycerol feeding. At the
beginning of the third stage, the DO concentration fell below 10% due to the residual
glucose initially added to the medium. This lasted from approximately 10 — 30 min before
glucose feeding (650 g/L) started. Fermentation cultures typically entered the stationary

phase between 24 — 30 h after inoculation of the fermentor culture medium. IPTG stock

117

solution was added to the culture medium at 18 h to a final concentration of 0.5 mM. The
solution of D-xylose was added to the fermentation medium at 24 h, 30 h, 36 h, and 42 h.

Fermentations were run for 48 h. Samples (5 — 10 mL) of the fermentation culture
were removed every 6 h after the fermentor was inoculated. Cell densities were
determined by dilution of fermentation broth with water (1:100) followed by measurement
of OD600. Dry cell weight of E. coli cells (g/L) was calculated using a conversion
coefficient of 0.43 g/L/ODéoo. The remaining fermentation culture was centrifuged to
obtain cell-free broth and analyzed by 1H NMR and gas chromatography. The cell pellets
were used for enzyme assays.

Compared with our benchmark E. coli WN13/pWN7.126B, which produced D-
1,2,4-butanetriol at a concentration of 10.5 g/L at 50% yield after modification of various
culturing parameters, KIT15/pML7.135 produced only 1.7 g/L of D-1,2,4-butanetriol in
8% yield under the same culture conditions. Interestingly, this construct yielded 23 g/L 3-
deoxy-D—glycero-pentulosonic acid, which accounted to 80% of the total D-xylose in the
system. Later on, E. coli KIT15/pRC1.55B was constructed. Plasmid pRC1.55B carries
the serA gene locus in order to complement the serA mutation of the E. coli KIT15. The
absence of deC gene encoding benzoylformate decarboxylase in this construct was
expected to give a similar yield and titer of 3-deoxy-D-glycero-pentulosonic acid as
observed in E. coli KIT15/pML7.135. The cultivation of this new construct, however,
only produced 11.8 g/L 3-deoxy-D-glycero-pentulosonic acid at 40% yield. The fact that
nearly half of the D-xylose was recovered at the end of the fermentation indicated that the
T7 promoter is not suitable for producing functional decarboxylase. The translation of

genes from GC rich organisms is inefficient in E. coli.

118

 

 

 

 

 

 

 

 

 

 

 

 

 

160 70

140 l (a)

120 A
s i s,
E 100 " :1
V .c
C en
.9 80 - '-
15 S
E 60 4 0 =
s .0. is
8 '0

20 - o
0 i
12 18

160 70

140 4 (b) r 60
£120 " __ 50 g,
E 100 ‘ . v
E’ . 0 t40 E
g 80 . O 2’
g ' - 30 E
§ 60 " i
c 40 ~ ' ' 2° a
8 u

20 ~ - 10

O
0 l I I r l' o
12 18 24 3O 36 42 48
time (h)

160 70

140 - (c) . 60
’2‘ 120 . _ 50 g
z - 40 E
.g 80 ~ .93
g , - 30 Q
: 60 " . ' . o =
8 - - 20 8
C 40 - Z‘
8 'o

20 - - 10

O
0 i . . - P 0
12 18 24 30 36 42 48
time (h)

Figure 62. Biosynthesis of 3-deoxy-D-giycero-pentulosonie acid under fermentor-
controlled conditions. (a) benchmark construct E. coli WN13/pWN7.126B; (b) E. coli
KIT15/ pRC1.55B; (c) E. coli KIT15/pML7.135; - D-l,2,4-butanetriol; l: 3,4-
dihydroxy-D-butanoic acid; -3-deoxy-D-glycero-pentanoic acid; -4,5-dihydroxy-
threo—L-norvaline; 12:1 3-deoxy-D-glycero-pentulosonic acid; E D-xylonic acid.

119

E. coli chaperonins GroES-GroEL for deC gene expression

E. coli is commonly used as a host for heterologous protein expression. However,
there are issues associated with this widely used system. For instances, inclusion bodies
and protease degradation are often encountered by researchers as a consequence of
improper protein folding. This phenomenon was observed in our previous attempt to
express P. putida deC under a strong T7 promoter. Molecular chaperones have been
demonstrated to be involved in the protein folding process in vivo, 63 and the
overexpression of these proteins proved to enhance active protein production.64 The E.
coli chaperonin GroEL first binds ATP and GroES co-chaperonin to create a large binding
cavity. The resulting complex then encapsulates partially folded or misfolded proteins and
unfolds it. As a final step in this GroEL-GroES mode of action, the partially unfolded
protein will undergo conformational changes yielding a hydrophobic core that compels the
misfolded protein to refold and obtain its correct structure.65 In a recent report, Jaeger and
coworkers utilized a commercially available chaperone plasmid system (Takara Bio USA)
with success.66 P. putida MdlC protein was co-expressed with chaperone proteins in E.
coli and increased in vitro enzyme activity was reported.66

To construct a plasmid harboring the Pm-groEL-groES gene locus, a 3.4 kb
fragment carrying the groEL-groES chaperones and the trigger factor (TF) genes was
amplified from the plasmid pG-Tf2 (Takara Bio USA). The PCR product was digested
with Hindlll and ligated with Hindlll digested pJF118EH to afford the 8.8 kb plasmid
pML7.166. This subcloning step replaced the tetracycline induced promoter Pztl in the
original plasmid with a tac promoter, thus avoided the use of light sensitive tetracycline

for induction during fermentation.

120

 

”332.23% PCR 3.3 kb gene containing groES, groEL
/ and the trigger factor from pG-T12

l1) Hindlll digest

    

Hindlll 3.3 kb findl“
groESL-tig
1) Hindlll digest
2) CIAP treatment

 

g pML7.166
8.8 kb

\r’

R IaCP , ”I

Ptac ///

 

 

Figure 63. Construction of plasmid pML7.166.

121

PCR 1.5 kb gene containing the tee promoter,

’/"/.,"1/7

 

groES and groEL from pML7.166
l4/mdl \
1) Smal digest
' pWN5.238A it
tpta ii

   

1) Smal digest
2) CIAP treatment

groESL 5 APR K,
/ ‘5),
e pML7.175 V:
g 8.5 kb 121
ii} if }
Ptaag Iale 1’?
, .\de0 ///
"‘ 32;“ if???”
Ptac

Figure 64. Construction of plasmid pML7.175.

122

__ APR 297’s
- groESL plasmid pRC1.55B

5
_§ pML7.175
l 8'5 kb 11) Smal digest

i

‘l

‘Ptac :9»!
‘ ~ 5m" 1.6 kb Sma'
' ' ' “ ' "'5'" zit? :~~"‘-"'.’;'r j:3§:';:';::';f]

   

serA

1) Seal digest
2) CIAP treatment

’

’1. " . ~‘A ': .
” | -‘. ‘1‘
\‘ a

/ ' .- \‘\?:“~.‘
'/ serA xx,
7'. ‘ .‘Lx

pML7.180
10.1 kb )

Figure 65. Construction of plasmid pML7.180.

123

x
. u
\\‘ \“

PCR 1.5 kb gene containing the tac promoter,
groES and groEL from pML7.166

'1’ s
\
.‘ \ x
1’ \ .
\. x,
\ .
\ .

{wiserA pML7.135 PW}? 1) Smal digest
\ 8.7 kb ,2,
’30P Xi?" Smal Smal
Ya: . /'/:,' 1 .5 kb
2:35 ,//

715‘: 'r- ” l -
c. ,.,~ .-' ,r
‘... 1‘3"». -’ t’ I ES‘

2) Klenow

1) Bglll digest
i 3) CIAP treatment L

”wt“,
/- \ ‘

, ‘..-4e ..

. ’- ‘.. -,
rr" ‘1 ‘s ‘

.4
r .-
y"

_, if;
/
i/ /
,Il/ser A

deC\\~ Pr7

ii pML7.202
,1 10.2 kb

J

Figure 66. Construction of plasmid pML7.202.

124

The 1.5 kb P,ac-groEL-groES locus was amplified by PCR from the plasmid
pML7.166. The PCR product was digested with Smal and ligated to Smal digested
pWN5.238A to yield plasmid pML7.175. A 1.6 kb serA locus was excised from plasmid
pRC1.55B by digestion with Smal and ligated to the Seal digested plasmid pML7.175 to
afford the 10.1 kb plasmid pML7.180. A similar strategy was used to insert the chaperone
locus into a plasmid expressing deC under a phage T7 promoter. In this case, plasmid
pML7.135 was used as the parent plasmid for cloning. The same 1.5 kb fragment carrying
the Pmc-groEL-groES fragment obtained from PCR was digested with Smal. This
fragment was subsequently ligated to pML7.135, which had been treated sequentially with
BglII, Klenow and CIAP, to afford upon ligation the 10.2 kb plasmid pML7.202.
Plasmids pML7.180 and pML7.202 were then transformed into host strains E. coli strains
KIT2 and KIT15, respectively, and evaluated under the same fermentation controlled
condition as discussed earlier. However, similar to the case of E. coli KIT15/pML7.l35,
all D-xylose was consumed and these E. coli microbes synthesized 21 g/L and 20 g/L 3-
deoxy-D-glycero-pentulosonic acid as product, respectively.

In a nutshell, neither using a strong phage T7 promoter or co-expressing E. coli
chaperonins GroES and GroEL proteins was a successful approach in yielding a better D-
1,2,4-butanetriol producing construct. Cultivation of these constructs under fermentor-
controlled conditions resulted in a significant decrease in D-l,2,4-butanetriol production.
The expression of deC under the T7 promoter vector system resulted in insufficient 3-
deoxy-glycero-pentulosonate decarboxylase activity. Metabolic burden associated with
using a strong promoter had led to a significant drOp in cell mass compared to the

microbial synthesis using our benchmark microbe WN13/pWN7.126B. Nevertheless, E.

125

coli microbes KIT15/pML7.135, KIT2/pML7.180 and KIT15/pML7.202 were able to

synthesize 3-deoxy-D-glycero-pentulosonic acid as the major product (Figure 62).

160
140 - (a)

 

o: m 8 r3
0 O o O
0

dry cell weight (glL)

concentration (mM)
-§
0

204 o

 

 

 

 

12 18
time(h)

 

160 70
140 . (b)
120 -
100 -
80 ~
60 4
40 ~ 0
20 ~

- 60
- 50
~40
- 30

dry cell weight (gIL)

L20

concentration (mM)

~10

 

 

 

 

Figure 67. Cultivation of chaperonin groEL-groES constructs under fermentor-
controlled conditions. (a) E. coli KIT2/pML7.180; (b) E. coli KIT15/ ML7.202; -
D—1,2,4-butanetriol; CZ] 3,4-dihydroxy-D-butanoic acid; 3-deoxy-D-glycero-
pentanoic acid; 4,5-dihydroxy-threo-L-norvaline; [2:1 3-deoxy-D-glycero-
pentulosonic acid; E D-xylonic acid.

126

Discovery of alcohol dehydrogenases for D-1,2,4-butanetriol biosynthesis

Background

The remaining byproduct formed during microbial D-1,2,4-butanetriol synthesis
was 3,4-dihydroxy-D-butanoic acid, which was derived from intermediate 3,4-dihydroxy-
D-butanal. Elimination of the formation of this byproduct was pursued by identification of
the enzyme and encoding gene or genes responsible for the conversion of 3,4-dihydroxy-
D-butanal into D-1,2,4-butanetriol. Overexpression of the identified dehydrogenase was
anticipated to reduce or eliminate formation of byproduct 3,4-dihydroxy-D-butanoic acid.
Screening of NAD- and NADP- dependent 1,2,4-butanetriol dehydrogenases in E. coli
was therefore pursued. E. coli ath, adhE and yiaY encoding substrate-nonspecific
alcohol dehydrogenases were found active towards using 3,4-dihydroxy-D-butanal as
substrate. Among these dehydrogenases, the importance of native E. coli Ath in the
1,2,4-butanetriol biosynthesis was demonstrated using a chromosomal ath inactivation
of E. coli WN13/pWN7.126B. Overexpression of ath gene in an E. coli D-1,2,4-
butanetriol producing host strain with additional 2-ketoacid dehydrogenase gene
inactivations led to a second generation microbe E. coli KIT4/pWN7.126B with improved
D-1,2,4-butanetriol production and reduced organic acid byproduct formation. Recently,
varying culture parameters for this new construct resulted in the synthesis of 18 g/L D-

l,2,4-butanetriol in 51% yield.

Cloning and screening for D-1,2, 4-butanetriol dehydrogenase candidates

To characterize the last biosynthetic step in WN13/pWN7.126B that converts 3,4-

dihydroxy-D-butanal into D-1,2,4-butantriol, five E. coli NAD- or NADP- dependent

127

enzymes that are commonly annotated as alcohol dehydrogenases or putative
dehydrogenase were screened. These enzymes are designated as Ath, AdhE, YhdH,
YiaY and YdjO in the ERGO bacterial genome database. AdhE is an iron-dependent bi-
functional enzyme with both dehydrogenase and pyruvate formate lyase activities. YiaY
is believed to be another iron-dependent dehydrogenase and it shares 63% amino acid
sequence identity to Zymomonas mobilis ath encoding alcohol dehydrogenase. YdjO is
annotated as an alcohol dehydrogenase in ERGO with no significant similarity to any
other known enzymes. Among these candidates, zinc-dependent Ath and YhdH alcohol
dehydrogenases shared over 80% nucleotide sequence identity with Z. mobilis adhA and
were of particular interest due to their ability to utilize propanal as substrate. These gene
candidates were independently amplified by PCR using Platinum Taq DNA polymerase
(Invitrogen) and cloned into pKK223-3 (ath, Figure 68 and adhE, Figure 69) and
pJF118EH (yhdH, yiaY and ya'jO, Figure 70 — 50) for protein expression. The resulting
recombinant plasmids were transformed into an E. coli DH50l expression host. The cell-
free extract of these constructs was used for in vitro enzymatic assay studies. The possible
reduction reaction was assayed by monitoring the depletion of NADH at 340 nm using
acetaldehyde and racemic 3,4-dihydroxybutanal as substrates. No measurable enzyme
activities were observed with YhdH and YdjO using both acetaldehyde and racemic 3,4-
dihydroxybutanal as substrates (Table 20). Ath, AdhE and YiaY, however, successfully
catalyzed the conversion of these substrates to form ethanol and 1,2,4-butanetriol,
respectively (Table 20). These results suggested Ath, AdhE and YiaY indeed had 1,2,4-
butanetriol dehydrogenase activities. Among them, enzyme Ath showed 20 times higher

activity over the other dehydrogenases and was subjected to further study.

128

PCR 1.0 kb 21th gene from
E. coli W3110 genomic DNA

«(if ,
'V

 

, 1) BamHl digest
‘23, 2) Klenow

ii. 4.6 kb f}

3 M 1.0 kb

   

1) 15le digest
L 2) Klenow j
3) CIAP treatment

 

Figure 68. Construction of plasmid pML6.166.

129

PCR 2.7 kb adhE gene from
E. coli W3110 genomic DNA

1) BamHl digest
2) Klenow

 

if pKK223-3 a
it, 4.6 kb g

   
       
  
 

 

‘Jfigfiifiiafigflgfl‘ifiifim '3 .

Y I. M 4‘5:
We.

     

.1 I ‘ _
'\. , . ‘. 2
\‘ . a. f .
K .' .5 , '2.
\u,‘;?"’7 n-,._,,p f:,(r,gl’
. ‘ . ‘. . .2'. ' 'P
4. K #44 .,, M

l) 15le digest
j 2) Klenow j
3) CIAP treatment

 

V)
pML6.168 )3,
7.3 kb ii
C?» m. w.“ 4.2"”

Figure 69. Construction of plasmid pML6.168.

130

PCR 1.0 kb yth gene from
E. coli W3110 genomic DNA

   

I],
/ 1) BamHl digest
,r/ 2) Klenow

6% 5.3 kb l};

£431} "1' .'

{g} I/
"T 1) EcoFll digest
j 2) Klenow L
3) CIAP treatment

 

Figure 70. Construction of plasmid pML6.259.

131

BamHl

Smal Sail '
EcoRl Hindlll
Ptac _ PCR 1.2 kb yiaY gene from

E. coli W3110 genomic DNA

 

1) BamHl digest
2) Klenow

5.3 kb

   

//",

«4’36"

‘ “~. M"::;;’
«.3?» as” ‘..-W

1), EcoRl digest
j 2) Klenow j
3) CIAP treatment

 

Figure 71. Construction of plasmid pML6.26l.

132

PCR 0.8 kb ydjO gene from
E. coli W3110 genomic DNA

1) BamHl digest

       

2) Klenow
0.8 kb
sataewmasa
_ MO
1) EcoRl digest
j 2) Klenow j
3) CIAP treatment

 

Figure 72. Construction of plasmid pML2.263.

133

Table 20. E. coli alcohol dehydrogenase candidates.

 

size crude lysate activity (U/mg)

 

 

entry gene (kb) °°"s"“°t acetaldehyde 3,4—dihydroxybutanal
1 ath 1.0 DH5a/pML6.166 9.6 0.8
2 adhE 2.7 DH5a/pML6.168 0.06 0.02
3 yhdH 1.0 DH5a/pML6.259 o o
4 yiaY 1.2 DH5a/pML6.261 0.13 0.03
5 ydjO 0.8 DH5a/pML6.263 o o

 

Chromosomal ath gene inactivation

To further examine whether Ath was responsible for in vivo conversion of 3,4-
dihydroxy-D-butanal into D-1,2,4-butanetriol, the E. coli strain WN13/pWN7.126B with a
chromosomal ath gene inactivation was prepared. A typical deletion was introduced
into E. coli using the Datsenko-Wanner methodology.67 PCR primers for the ath gene
inactivation were designed based on 40 nt homologous sequences for this gene and 20 nt
for the priming site on the template plasmid pKD3. Plasmid pKD3 contains a
chloramphenicol resistant gene flanked by Flp recognition target (FRT). E. coli wild-type
W3110 carrying plasmid pKD46 encoding 7t. Red recombinase was made competent and
subsequently transformed with the F RT-flanked chloramphenicol linear PCR gene
fragment. The transformants were selected on an LB solid agar plate containing
chloramphenicol. Genomic DNA was purified from three single colonies that appeared on
this plate using the CTAB method as described in the earlier chapter. Disruption of
chromosomal ath was confirmed by PCR using a pair of primers flanking the ath
locus. The resulting E. coli strain W31 10athzzFRT-CmR-F RT was designated E. coli

KIT8.

134

This mutation was then transferred to the E. coli WN13/pWN7.126B by P1 phage-
mediated transduction into E. coli KIT2 using KIT8 as a donor strain. The resulting strain
E. coli KIT9 was selected on an LB plate containing chloramphenicol. Subsequent
removal of the drug marker was achieved by thermal induction of Flp recombinase in E.
coli KIT9/pCP20.68 The resulting marker-free strain was designated E. coli KIT10. Both
E. coli strains KIT9 and KIT10 were transformed with the deC carrying plasmid
pWN7.126B and their catalytic efficiencies were evaluated under fermentor-controlled

conditions.

Table 21. E. coli strains for chromosomal ath inactivation study.

 

 

strain genotype

WN7 W31 10 yth::FRTyagE::FRTserA

WN13 W31 10 yth::FRTyagE::FRTserAxyIAB::xdh-FRT—CmR-FRT

KIT2 W31 10yth: : FRTyagE::FRTserAxyIAB:: xdh-FRT

KIT8 W31 1 Oath. : F RT-CmR-FRT

KIT9 W3110yth::FRTyagE::FRTserAxyIAB:: xdh-F RT ath.:FRT-CmR-FRT
KIT10 W31 10 ythzzFRTyagE::FRTserAxyIAB:: xdh-FRT ath.:FRT

 

E. coli KIT10/pWN7.126B produced only 6.5 g/L D-1,2,4-butanetriol in 31% yield
under fermentor-controlled conditions. D-3,4-Dihydroxybutanoic acid accumulated to 5.3
g/L in the culture medium, which accounted for a 15% increase in concentration compared
to its parent microbe E. coli WN13/pWN7.126B. This result suggests that E. coli Ath is
indeed an in vivo D-1,2,4-butanetriol dehydrogenase. Since KIT10/pWN7.126B was still
able to synthesize D-l,2,4-butanetriol, it therefore becomes tempting to speculate that one
or more other native E. coli dehydrogenases may reduce 3,4-dihydroxy-D-butanal to D-

1,2,4-butanetriol in vivo.

135

concentration (mM)

concentration (mM)

concentration (mM)

160

 

d —L d
N # O) on O N h
C O O O O O O
.l l l l l

(a)

70
i- 60

 

 

O

160

12

 

 

140 l
120 J
100 J
80 .
so ~
40 -
20 ~

 

(b)

 

 

12

time (h)

 

160
140 .
120 <
100 -
80 «
60 «
40 J
20 «

 

(C)

 

 

 

 

dry cell weight (gIL) dry cell weight (g/L)

dry cell weight (g/L)

 

Figure 73. Cultivation of E. coli strains with ath gene inactivation under
fermentor-controlled conditions. (a) E. coli WN13/pWN7.126B; (b) E. coli KIT9/

pWN7.126B; (c) E. coli KIT10/pWN7.126B; - D-l,2,4-butanetriol; [:21 3,4-
dihydroxy-D-butanoic acid; 3-deoxy-D-glycero-pentanoic acid; 4,5-
dihydroxy-threo-L-norvaline; 3-deoxy-D-glycero-pentulosonic acid; E D-

xylonic acid.

136

Plasmid-based expression of ath gene

Further understanding of the dehydrogenation reaction in the D-1,2,4-butanetriol
biosynthetic pathway offered us an opportunity to improve the performance of our
benchmark E. coli WN13/pWN7.126B. Three additional alcohol dehydrogenase
candidates identified in other microorganisms were screened in an attempt to obtain a
better wild-type D-1,2,4-butanetriol dehydrogenase. Z. mobilis AdhA and Ath shares
80% and 63% amino acid identity to E. coli Ath and YiaY, respectively. Klebsiella
pneumoniae DhaT uses 1,3-propanediol as its native substrate and is of particular interest
because of the structural similarity between 1,3-propanediol and 1,2,4-butanetn'ol. Z
mobilis alcohol dehydrogenase candidates adhA and ath were localized in plasmids
pLOIl35 and pLOIZ95, respectively. K. pneumoniae dhaT was amplified from K.
pneumoniae ATCC25955 genomic DNA and subsequently cloned into pJF118EH to yield
pML7.179. These plasmid-localized genes were expressed in E. coli DHSa and the
enzyme activities were determined using the same spectrophotometn'c assay as described
earlier (Table 22). Crude lysate of AdhA and DhaT showed activities of 0.06 and 0.003
U/mg, respectively, using 3,4-dihydroxybutanal as substrate. Ath showed no detectable

activity. The enzyme activities toward their native substrates were also determined.

Table 22. Other alcohol dehydrogenase candidates

 

 

 

size crude lysate activity (U/_mg)
entry gene (kb) construct native substrate 3,4—dihydroxybutanal
1 adhA 1.2 DH5a/pLOI135 73" 0.06
2 ath 1-8 DH5or/pLOI295 508 <0.001
3 dhaT 1.2 DH5a/pML5.179 4.3" 0.003

 

“Ethanol was used as substrate, NAD was used as cofactor.
b1,3-propanediol was used as substrate, NAD was used as cofactor.

137

 

iii
1:51 IacIQ pJF118EH E}
iii 5.3 kb

\ ’s
I“ ’...

- ,-.-/v , ,1 2..)
“w a??? .‘rv ”MHz-r".- ,4"
~q.- . » _. . a;
use. .MJ

j 1) Hindlll digest

PCR 1.2 kb dhaT gene from
K. pneumoniae genomic DNA

1) Hindlll digest

Hindlll 1.2 kb Hindlll

Ethernet?

 
   
 

   
    

 

2) CIAP treatment ¢

 

Figure 74. Construction of plasmid pML5.179.

138

E. coli ath, Z. mobilis adhA and K. pneumoniae dhaT were subsequently cloned
into plasmid pWN7.126B. The effect of expressing these alcohol dehydrogenases during
the biosynthesis of D-1,2,4-butanetriol were examined under fermentor-controlled
conditions. A 1.3 kb fragment carrying the tac promoter region and the ath gene was
excised from plasmid pML6.166 with BamHI and ligated to BamHl digested pWN5.23 8A
to afford plasmid pML6.185. A 1.6 kb serA locus excised from plasmid pRC1.55B with
Smal was ligated with ScaI treated pML6.185 to afford plasmid pML6.195. On the other
hand, plasmid pWN6.096B containing Z mobilis adhA and deC gene fragments was
obtained from the Frost group strain collection. The same serA locus was ligated to ScaI
digested pWN6.096B to yield plasmid pML6.133. Finally, the K. pneumoniae dhaT gene
was excised from plasmid pML5.179 by treating with BamHI and Klenow. This gene
fragment was subsequently ligated to pKK223-3, which was sequentially treated with
EcoRI and Klenow to yield plasmid pML6.090. The Pm-dhaT locus was excised with
BamHI and ligated to BamHl treated pWN5.238A to afford plasmid pML6.128. Inserting
the Smal digested serA gene into the Seal site of this plasmid yielded plasmid pML6.135.
The resulting mdlC genes and dehydrogenase candidates in these three plasmids were
transcribed in the opposite direction under a separate tac promoter. Transcription of both
genes was controlled by the LaclQ repressor protein encoded in the lac]Q gene and
initiated by addition of IPTG. With transformation of these plasmids into E. coli host
WN13, constructs WN13/pML6.195, WN13/pML6.133 and WN13/pML6.135 were
obtained and evaluated in fermentors using the same cultivation conditions as described

earlier.

139

 

I e
' I
I
III. 3
'v . I
I, /’
rs. ,. x
,r” ‘ u . 7r
- , , (“g/Ill? /
I' " "U 4. 3.5,,
I II 2 I" .r
‘I i '2’ .1 1" .
I /. ‘II I'-
I / .Il . ’.
. /
/ 'W 4' ’1
deC _
I '1
. /

 

  

pr5 238A 1) BamHl digest

 
  

 
  

/’ BamHl 1 .5 kb 33ml.“

" ' ' \.~-'v'.-
i ' J 6% a 151'? a": aadv'iflvfinf“ , V.

1) BamHl digest
2) CIAP treatment

(.1
v

fi.“\‘
I s l 5\
. )3")

 

deC i l 8.5 kb

1‘ I, [(1
\ I r‘
\
1 .~ ./
‘\\ \\ . 1'
s ‘\ .
l .\
N‘ 1

Figure 75. Construction of plasmid pML6.185.

140

plasmid pRC1.558

 

f l" pML6.185 1.1. .
l, lmdlc 8.5 kb H l" 3m" “'99?“

\

J

1.6 kb sma'
' ' .1453,,j;;'.,.~.;:;'jg141;];,..U'.;TZ:T;..;;LI]

 

«2.1.2..- serA

1) Seal digest
2) CIAP treatment

 

iii
1 1 10.1 kb ‘1

 

Figure 76. Construction of plasmid pML6.195.

141

Pt ‘Lad'hA \Ewfl
8% APR i. plasmid pRC1.558
.51 ‘1:

" pWN7.0968 1
lmdle 8.4 kb til

-\

l1) Smal digest

’ «.—
’.r’
a

// Smal 1. 6 k6 Smal
4L . ., I

\ lac!Q

Ptac W 6.. ,,,/

~ _.. w serA

¢ 1) Seal digest 6

 

2) CIAP treatment

,7? ‘
fill/l ‘51
deC 1 i pML6.133 ‘ng
l i 10.2 RD N

 

Figure 77. Construction of plasmid pML6.133.

142

Isma' PSflHindlll

 

1:412», '4 ‘
'. Iva, '1: ,1, -‘ I
I/ )7. - tram.»- ' "
““4444 4.32,. .. “I

1) BamHl digest
; f 2) Klenow
1,

 

2) Klenow

6 1) EeoRl digest 6
3) CIAP treatment

 

Figure 78. Construction of plasmid pML6.090.

143

BamHl

 

x ’l.— 1"
a

;...
/ r
/;-’deC
i" ,

.w 5* 1,,
. Hex Jr.) .
‘- rm“... ..-.v' '.-

 

pWN5.238A
iptae 7'0 kb ll
h’ BamHl 1.5 kb BamHl

   

/_ /
/
A]
1‘ /
/’- 1",
I ,
-4. ",u’

1) BamHl digest
2) CIAP treatment

 

if pML6.128 ii
i

' I :ff.
deC (‘6 8.5 kb {6.
‘ i. r
l“ l.“ ,7]
431/
63;?
«/'/

Figure 79. Construction of plasmid pML6.128.

144

 

plasmid pRC1.558
ML6.128 \f‘i .
[‘ imdlC p 8.5 kb M 1) Smal digest

   

1.6 kb Smal

 

 

‘77 7 serA
1) Seal digest
2) CIAP treatment

 

Figure 80. Construction of plasmid pML6.135.

145

160 70

 

 

 

 

 

 

 
   

 

 

 

 

 

 

 

 

140- (a) . .60
120 ,4
g T . O . 750%
§6100 . I:
S 80 . r 40 £6
8 L30 g
‘E 60 ~ . ‘§
c 40 . - 20 g;
8 .
204 o '10
0 F 0
12 18
160 70
140 . (b) ~ 60
" 120 - 6 _ 50 a~
2 5.
£1004 . I:
c - 40 g;
:g 80 . . '2
g '30-
c 60" §
E44 ' ”"5
O x
20« ' -10
0 1 I I '0
12 18 24 30 36 42 48
finned»
160 70
140 ~ .
120 (C) 60
A ‘ 502
£1004 ' g3
g 80" o .9,
'5 ° .¥
15,60- . a
c 40 . 2‘
8 "O
20 - ' E
o I I U ml
12 18 24 30
thne(h)

Figure 81. Cultivation of E. coli constructs expressing plasmid-borne alcohol
dehydrogenases under fermentor-controlled conditions. (a) E. coli WN13/pML6.195;
(b) E. coli WN13/pML6.133; (c) E. coli WN13/pML6.135; - D-1,2,4-butanetriol;
[:1 3,4-dihydroxy-D-butanoic acid; 3-deoxy-D-glycero-pentanoic acid; 4,5-
dihydroxy-threo-L-norvaline; 1:21 3-deoxy-D-glycero-pentulosonic acid; E D-
xylonic acid.

146

As shown in Figure 81, E. coli microbes WN13/pML6.195, WN13/pML6.133 and
WN13/pML6.135 synthesized D-1,2,4-butanetriol at a concentration of 4.6 g/L (22%
yield, mol/mol), 2.5 g/L (12% yield, mol/mol) and 2.9 g/L (14% yield, mol/mol),
respectively. Overexpressing plasmid-localized ath in these constructs led to poor cell
growth and decreased benzoylformate decarboxylase activities as indicated by the
accumulation of 3-deoxy-D—glycero-pentulosonic acid. In addition, a high glutamic acid
concentration was observed in lH NMR spectra of clarified culture medium at different
time intervals. Glutamic acid accumulation has long been known to associate with
hyperosmotically stressed E. coli and Streptomyces typhimurium.69 According to the
literature, it is advantageous in terms of stability, genetic regulation, and metabolic burden
to modulate expression of relevant genes on the chromosome rather than relying on
overexpressing these genes on a multi-copy plasmid.7o To strive for a balance between
getting higher enzyme activity and avoiding stress to cells, chromosomal insertion of an
alcohol dehydrogenase on the chromosome was pursued. Only the E. coli ath gene was
employed in this study to avoid potential complication associated with heterologously

expressing Z mobilis adhA and K. pneumoniae dhaT in E. coli.

Chromosomal expression of E. coli ath

Two strategies were pursued to increase expression of Ath activity in E. coli
chromosomally. E. coli KIT3 was derived from E. coli WN7 by replacing the genomic
copy of the xylele gene cluster with an xdh(Caulobacter crescentus)-ath-P,ac-FRT-
CmR-FRT gene cassette. The xylA gene encodes D-xylose isomerase, while the xle gene
encodes D-xylulose kinase. These are the two enzymes essential for E. coli catabolism of

D-xylose. C. crescentus xdh gene encoding D-xylose dehydrogenase was identified in a

147

previous study to convert D-xylose into D-xylonic acid.36 This gene was successfully
incorporated in our benchmark construct E. coli WN13/pWN7.126B to synthesize D-1,2,4-
butanetriol from D-xylose. The proposed chromosomal modification in E. coli KIT3
therefore abolished its ability to utilize D-xylose as a sole carbon source for growth. As a
second consequence, E. coli KIT3 could express a D-xylose dehydrogenase activity under
the control of the xylA promoter, while the alcohol dehydrogenase encoding gene ath
was expressed under an independent tac promoter supplied in the cassette. The xdh and
(1th genes would transcribe in opposite directions.

A typical chromosomal gene insertion in E. coli was achieved using a modified
Datsenko-Wanner methodology (Figure 82). First, the Pam-ath gene locus was
amplified from plasmid pML6.166 using Platinum T aq DNA polymerase (Invitrogen) and
subsequently cloned into the AflIII site of plasmid pKD3 to yield the plasmid pML6.255
(Figure 83). The template plasmid containing both ath and xdh genes was obtained by
cloning the xdh gene into the SphI site of pML6.255. The resulting plasmid was
designated pML6.272 (Figure 84). PCR primers for the gene insertion experiment were
then designed based on 40 nt homologous sequence for this gene and 20 nt for the priming
site on the template plasmid pML6.272. E. coli wild-type W3110 carrying plasmid
pKD46 encoding the A. Red recombinase was made competent and transformed with this
gene cassette. The transformants were selected on an LB solid agar plate containing
chloramphenicol. Chromosomal gene insertion was verified by PCR using a pair of
primers flanking the xylele gene locus. The resulting E. coli strain was designated E.
coli KIT1. The mutation in E. coli KIT1 was then transferred to E. coli WN7 by P1

phage-mediated transduction. The resulting E. coli KIT3 was selected on an LB plate

148

containing chloramphenicol. Subsequent removal of the drug marker was achieved by
thermal induction of Flp recombinase in KIT3/pCP20 to yield E. coli strain KIT4. Both
KIT3 and KIT4 were transformed with the plasmid pWN7.126B carrying the deC gene

and evaluated under fermentor-controlled conditions (Table 23 and Figure 86).

PCR product of pML6.272 Ptac

_——_‘:L-——_—-——

xdh ath FRT CmR FRT

ny/A E. coli W31 1 0 chromosome

‘[:'_————

xyIA xyIB

homologous recombination

nylA E. coli KIT1 chromosome Ptac
xdh ath FRT CmR FRT

Figure 82. Strategy to integrate xdh-ath-Pm gene cassette into E. coli W3110.

Table 23. E. coli strains for chromosomal ath expression study.

 

 

strain genotype

WN7 W31 10 yth::FRTyagE::F RTserA

WN1 3 W31 1 Oyth: : FRTyagE: : F RTserAxyIAB: :xdh—FRT-CmR-FRT

KIT1 W311OxyIAB::xdh-ath-PtachRT-CmR-FRT

KIT3 W31 10yth: : FRTyagE: : FRTserAxy/AB: :xdh-ath-Ptac- F RT-CmR-FRT
KIT4 W31 1 0yth: : FRTyagE: : FRTserAxy/AB: :xdh-ath-Ptac-F RT

KIT5 W3110P75- FRT-CmR-FRT

KIT6 W31 10yth: : FRTyagE::FRTserAxyIAB: :xdh-FRT PT5- F RT-CmR-FRT
KIT7 W31 10yth::FRTyagE::FRTserAxylAB::xdh-FRT P-r5-FRT

 

149

1: PCR 1.3 kb fragment containing
#1.; $1,...” the tac promoter and ath
gene from plasmid pML6.166

     

1) Afllll digest
Afllll 1.3 kb Afllll
1) Afllll digest
2) CIAP treatment

4.1 kb

 

 

Figure 83. Construction of plasmid pML6.255.

150

  
 

plasmid pWN9.068

  
 

“"3: ‘.
-"~‘;-=
‘ o’, A
"9:.
D '35?
a
3'13.
'.13.
I f
I

pML6.255 l1) Sphl digest

 

xdh

1) Sphl digest
2) CIAP treatment

 

Figure 84. Construction of plasmid pML6.272.

151

The second E. coli strain was derived from E. coli WN13. Similarly, the modified
Datsenko-Wanner methodology was employed to replace the native ath promoter with
the gene locus containing the phage T5 promoter region, the lacO gene and the ribosome
binding site obtained from plasmid pQE30. This strategy was particular appealing
because only a minimal chromosomal modification was required to improve native ath
expression. A gene cassette PT5-FRT-CmR-FRT was constructed by cloning the P75 locus
into the NdeI site of plasmid pKD3 to yield plasmid pML7.042 (Figure 85). The resulting
chloramphenicol resistant gene is transcribed in an opposite direction as the phage T5
promoter. This plasmid pML7.042 also delivers a ‘universal’ phage T5 cassette for other
native gene expression experiments along the E. coli genome.

To replace the native E. coli ath promoter with this phage T5 cassette along with
a chloramphenicol marker, the phage T5 cassette was amplified by PCR and subsequently
transformed into electrocompetent E. coli W3110. The transformants were subject to
selection on LB plates containing chloramphenicol. The resulting strain was verified by
PCR and designated as E. coli KITS. Strain KIT6 was prepared from E. coli WN13 by P1
phage-mediated transduction using KITS as the donor strain. This E. coli KIT6 was
selected on the same LB/Cm plate and its genotype was verified by PCR. Using the same
recombination method, thermal induction of Flp recombinase of KIT6/pCP20 excised the
chloramphenicol marker and afforded strain KIT7. E. coli KIT7 contains only a 0.2 kb
genomic insert. Both E. coli KIT6 and KIT7 were transformed with plasmid pWN7.126B

and evaluated under fermentor-controlled conditions (Table 23 and Figure 87).

152

PCR 0.2 kb fragment containing the
phage T5 promoter from plasmid pML6.166

1) Ndel digest
pKDS

2.8 kb

 

Ndel Ndel
III:/ I

43:?” Ndel 0.2 kb

.r
,' . I: r
'..' I.. _q '.J’
V". .1 HIV” . r"

.. -2, P T5

¢ 1) Ndel digest ¢

 

"x
“A

2) CIAP treatment

1’7”- ,I/HV»

, ’x'-1,1"
1,1;Ifllwgy/Z‘Augg‘” b .

If, f I»‘

for}, ‘4“451'2».

-K/ I '
Aura"

1.; 'lv/

 
    

CmR

 

pML7.042 :3
3.0 kb .523
P T5 ’5’]
M
37.7,» _" Marx/é:

Figure 85. Construction of plasmid pML7.042.

153

160 70

 

 

 

 

 

 

 

 

 

“0 J (a) ° -60
0

g 120 "‘ . . _ 50 g
e 100 . g
E’ .r:
g 80 .. 0 .§’
g 60 O =
§ 8
g 40 4 o _g‘

20 «

o T ' I \ I
12 18 24 30 36 42 48
time (h)

160 70

140 4 (b) - 50

120 1 . A
A - 50 d
2
g 100 - f."
5 80 fi’
'13 1 . 8
E 60 q =
8 E
§ 40 - 1,

20 1 o

o
12 18

 

Figure 86. Cultivation of E. coli constructs having a chromosomal Pm-ath gene
insertion under fermentor-controlled conditions. (a) E. coli KIT3/pWN7.126B; (b) E.
coli KIT4/pWN7.126B; - D-l,2,4-butanetriol; 1:] 3,4-dihydroxy-D-butanoic
acid; 3-deoxy-D-glycero-pentanoic acid; 4,5-dihydroxy-threo-L-norvaline;
[:21 3—deoxy-D-glycero-pentulosonic acid; E D-xylonic acid.

154

160 70
140 1 (a) .. 60
120 1
100 1 °
80 1
60 1
40 1 0
20 1

 

dry cell weight (gIL)

concentration (mM)
0

 

 

 

 

time (h)

160 70

 

140 1 (b) . 60
120 1
100 1
80 1
60 1
40 1
20 1 0

concentration (mM)
dry cell weight (g/L)

 

 

 

 

1218 24 3O 36 42 48
time(h)

Figure 87. Cultivation of E. coli constructs having a chromosomal T5 promoter
insertion under fermentor-controlled conditions. (a) E. coli KIT6/pWN7.126B; (b) E.
coli KIT7/pWN7.126B; - D-l,2,4-butanetriol; 12:1 3,4-dihydroxy-D-butanoic
acid; 3~deoxy-D-glycero-pentanoic acid; 4,5-dihydroxy-threo-L-norvaline;
CZ] 3-deoxy-D-glycero-pentulosonic acid; E D-xylonic acid.

155

It is known in the literature that the insertion of an antibiotic gene on the
chromosome might exert a polarity effect on downstream genes, ultimately leading to

growth issues.“70

Excising the chloramphenicol resistance gene in E. coli D-1,2,4-
butanetriol synthesizing constructs was of importance as observed in the cultivation
experiments shown in Figure 86 and 65. E. coli microbes KIT3/pWN7.126B and
KIT6/pWN7.126B synthesized D-1,2,4-butanetriol at a concentration of 10.0 g/L (47%
yield, mol/mol) and 9.8 g/L (46% yield, mol/mol), respectively. A 15% increase of
product concentration was observed in the cultivations using the antibiotic marker-free
derivatives of these constructs. E. coli KIT4/pWN7.126B and E. coli KIT7/pWN7.126B
synthesized D-1,2,4-butanetriol at a concentration of 11.5 g/L (55% yield, mol/mol) and
11.0 g/L (52% yield, mol/mol), respectively. E. coli KIT4/pWN7.126B was able to
synthesize D-1,2,4-butanetriol at a concentration and yield 10% higher than the benchmark
E. coli WN13/pWN7.126B. The success in using chromosomal ath expression for the

microbial D-1,2,4-butanetriol synthesis led to the construction of a second generation E.

coli microbial catalyst.

Construction of the second generation D-1,2, 4—butanetriol synthesizing E. coli microbe

The new generation of D—1,2,4—butanetriol synthesizing E. coli strain contains all
the genetic modification developed in E. coli KIT4/pWN7.126B. In addition,
chromosomal inactivations of yiaE and ych genes were introduced. In an earlier study,
E. coli yiaE and ych genes encoding enzymes were identified to catalyze the reduction
of 3-deoxy-D-glycero-pentulosonic acid to form 3-deoxy-D-glycero-pentanoic acid using

NADPH as a cofactor.36 Introducing these two inactivations into the second generation

156

microbe would likely reduce the production of 3-deoxy—D-glycero-pentanoic acid,
ultimately increasing the yield for the desired product D-1,2,4-butanetn'ol.

Gene yiaE was inactivated by replacing it with a chloramphenicol resistant gene
flanked with Flp recognition target sequence (FRT) amplified by PCR from plasmid
pKD3. E. coli KIT16 containing this yiaE gene inactivation was prepared directly from E.
coli KIT2 using the same homologous recombination method as described earlier. Gene
)2ch was replaced by a kanamycin resistant gene that was obtained using a different
template plasmid pKD4. The introduction of this kanamycin gene locus into E. coli
KIT16 yielded a double knockout strain KIT17. Thermal induction of the F lp
recombinase of KIT17/pCP20 excised the chloramphenicol and kanamycin marker in a
single step and afforded strain KIT18. Both E. coli strains KIT17 and KIT18 were then
transformed with deC containing plasmid pWN7.126B and evaluated (Table 24).
Preliminary fermentation results were shown in Table 25. E. coli KIT17/pWN7.126B
synthesized only 5.5 g/L (26% yield, mol/mol) D-1,2,4-butanetriol, presumably due to the
polarity effect. KIT18/pWN7.126B yielded 11.2 g/L (53% yield, mol/mol) D-1,2,4-
butanetriol, which is consistant with the earlier result using KIT4/pWN7.126B as catalyst
(Table 24, line 3). Most importantly, a decrease in the concentration of 3,4-dihydroxy-D-
butanoic acid and 3-deoxy-D-glycero-pentanoic acid was observed for this fermentation,
indicating that the (1th genomic expression and the 2-keto-acid dehydrogenase
inactivations were fimctional. Attempts to vary the amount of substrate D-xylose yielded
an increase in D-l,2,4-butanetriol production (Table 25, line 4 —- 6). Using 50 g/L D-
xylose as substrate, E. coli KIT18/pWN7.126B was able to synthesize 18.0 g/L D-1,2,4-

butanetriol in 54% yield.

157

Table 24. Creation of the second generation D-l,2,4-butanetriol synthesizing E. coli.

 

 

strain genotype

K'T16 W31 1 0 yth: : FRT yagE : : FRTserAX yIAB: :xdh-ath-Ptac-FRTyiaE: : F RT-CmR-F RT

KIT1 7 W31 1 0yth: : FRTyagE: : FRTserAxyIAB: :xdh-ath-Ptac-FRTyiaE: : FRT-CmR-F RT
ych: : FRT-KmR-FRT

KIT13 W31 1 0 yth: : FRTyagE: : F RTserAxyIAB: :xdh-ath-Ptac-FRTyiaE: : F RTych: : F RT

 

Table 25. Preliminary results on the cultivation of E. coli KIT18/pWN7.126B under
fermentor-controlled conditions.

 

 

 

xylose time titer (1L) BT yield

construct (g) (h) BTa DHBA” DPAC DHNd DGP" (%)
WN13/pWN7.1268 30 48 10.2 4.6 5.1 3.8 o 48
KIT17/pWN7.1268 30 48 5.5 3.8 2.1 4.4 12.6 26
KIT18/pWN7.1263 30 48 11.2 3.9 2.9 5.3 o 53
KIT18/pWN7.126B 50 48 16.5 4.9 5.4 6 3 47
KIT18/pWN7.126B 50 54 18.0 5.2 5.5 5.9 o 54
KIT18/pWN7.126B 70 60 18.3 4.7 4.7 8.5 15.3 37

 

aD-1,2,4-butanetriol (BT)
b3,4-dihydroxy-D-butanoic acid (DHBA)
c3-deoxy-D-glycero-pentanoic acid (DPA)
d4,5-dihydroxy-threo-L-norvaline (DHN)
°3-deoxy-D-glycero-pentuIosonic acid (DGP)

Future work

Directed Evolution of 3 -deoxy-D-glycero-pentulosonate decarboxylase

A variety of strategies was pursued to identify a 3-deoxy-D-glycero-pentulosonate
decarboxylase with improved activity. Using the bioinformatics approach, four novel 3-
deoxy-glycero-pentulosonate decarboxylases were successfully identified. Unfortunately,

these decarboxylases were less active than the P. putida MdlC currently employed in the

158

D-1,2,4-butanetriol synthesizing construct. Attempts to evolve an active decarboxylase
using family gene shuffling were unsuccessful due to insufficient screening capability and
sequence identity. Recently, a robotic screening workstation was purchased and is
available for access. This robotic screening workstation features a Genetic QPix2XT
colony picker, a Tecan Freedom Evo 200 workstation and a Molecular Devices
SpectraMax Plus 384 microplate reader. Mutant libraries will be generated by error-prone
PCR. Mutants with improved activity will then be subject to DNA shuffling for further
improvement. Library screening of deC mutants will be carried out in 96-well format.
Purified DNA obtained from error-prone PCR7| will be transformed into electrocompetant
E. coli DHSa and plated on a QTray (Genetix) containing LB/Ap solid medium. This step
will be optimized such that independent colonies appearing on each tray will be
recognized by the CCD camera of the Genetix QPix2XT colony picker. Single colonies on
the tray will then be screened using the following protocol recently revised.

A sterilized deep-well block (Genetix) with each well containing 1 mL of LB/Ap
liquid medium will be inoculated with single colonies from the mutant library using the
96-pin head of the colony picker. Simultaneously, a master plate for these mutants will be
prepared by inoculating the material left on the 96-pin head into the wells of a microplate

containing 100 uL LB-freeze medium in each well. Wells A1 and B1 of the deep-well

block will be inoculated with DHSor/pKK223-3 and DH50t/pML7.118 (P. putida mdlC
gene cloned into pKK223-3) as controls, while well H12 will be left as blank. Therefore,
a total of 93 independent mutants will be assayed in a typical screening cycle. The deep-

well block will be covered by the Qiagen Airpore Sheet and incubated in the shaker at 37

159

°C for 12 h with agitation at 250 rpm, while the master plate will be incubated overnight
in an oven at 37 °C.

After the incubation, glycerol will be added to the wells and the master plate will
be stored at -80 °C. The deep-well block will be removed from the shaker and the cells
will be pelleted by centrifugation at 4,000 rpm for 5 min. With the exception of the
centrifugation steps, all subsequent microplates handling and liquid pipetting steps will be
automated by the Tecan Freedom Evo 200 Workstation coupled with the Evoware Plus
v1.4 software. The supernatant will be discarded and the cell pellets at the bottom of the
block will be resuspended in 100 uL BugBuster solution (Novagen), transferred to a 96-
well microplate (N unc), and incubated in a microplate hotel at room temperature for 0.5 h
without agitation. The cell lysate will then be centrifuged at 4,000 rpm for 20 min to
remove cell debris. Protein concentration will be estimated by the measurement of optical
density at 280 nm using the Molecular Devices SpectraMax Plus 384 microplate reader
that is integrated to the Tecan workstation. This microplate will then be stored in the
cooling hotel at 4 °C prior to use.

The enzymatic assay reaction for decarboxylases will be setup as described earlier
followed by a 6 h incubation period at room temperature. The reactions will be quenched
by the addition of 30 uL of 10% trichloroacetic acid solution followed by protein removal
using centrifugation. 50 uL of Schiff’s reagent72 will be added to each well and incubated
at room temperature for another 30 min. Schifi’s reagent has been demonstrated to react
selectively with 3,4-dihydroxybutanal to give a red dye. The formation of this red dye in

microplate wells will be recorded by measuring the absorbance at 550 nm. The higher

160

absorbance readings at 550 nm will be interpreted as indicative of a mutation leading to a
more reactive decarboxylase.

Enzymes corresponding to these putatively more reactive decarboxylases will be
assayed a second time at higher dilution using Schifi’s reagent. The more active
decarboxylase mutants will be assayed for a third time using an in vivo assay described
earlier in this chapter. E. coli W3110 transformed with plasmid-localized mutant genes
will be cultivated using D-xylonic acid as sole carbon source for growth. The resulting
culture medium will be derivatized with BSTFA and the D-l,2,4-butanetriol will be
quantified with our highly sensitive gas chromatography interfaced with a flame-

ionization detector.

C hemo-microbial synthesis of (S)-3 -hydroxy- 7-butyrolactone

As discussed in chapter one, (S)-3-hydroxy-y-butyrolactone is an important
synthetic intermediate for a variety of chiral compounds. For instance, the use of this
precursor in the manufactures of Crestor and Zetia attracts enormous commercial interest.
There are various strategies in the literature toward the synthesis of this molecule. (S)-3-
hydroxy-y-butyrolactone can be synthesized from the selective reduction of L-malic acid
ester using a dimethyl sulfide complex of borane and a catalytic amount of sodium

73

borohydride. Methods of enzymatic or catalytic reduction of B-ketoester are known,

however, with low enantioselectivity. 74 Many other methods were derived using
carbohydrates such as cellobiose,75 maltose,761actose,77 starch78 or cellulose78 as starting
materials. However, issues associated with these methods such as low yielding and low
stereoselective reactions, and byproducts formation of glycolic acid, isosaccharinic acid,

formic acid, ketones and diketones increase the product cost and complicate the

161

downstream purification process.”78 The recently discovered microbial synthesis of 3-
deoxy-D-glycero-pentulosonic acid using E. coli microbe KIT15/pML7.135 from D-xylose
provides a commercially viable alternative toward the synthesis of (S)-3-hydroxy-y-
butyrolactone. A simple oxidative dacarboxylation of 3-deoxy-D-glycero-pentulosonic
acid into 3,4-dihydroxy-D-butanoic acid followed by an acid-catalyzed lactonization to
afford this important chiral synthon was envisioned. This stereospecific synthesis

guarantees the enantio-purity of product (S)-3-hydroxy-y-butyrolactone (Figure 88).

OH O HO

OH O
a b
0 0“ o
3-deoxy- D-glycero- 3,4-dihydroxy-D- (S)-3- hyd roxy-
pentulosonic acid butanoic acid y—buty rolactone

Figure 88. Proposed chemo-microbial synthesis of (S)-3-hydroxy-y-butyrolactone.
(a) oxidative decarboxylation; (b) lactonization.

To facilitate this study, a simple purification method of 3-deoxy-D-glycero-
pentulosonic acid was developed from the fermentation broth of KIT15/pML7.l35. The
fermentation broth was first passed through an anion-exchange column packed with
Dowex 1X8 (Cl' form) and eluted with 100 mM HCl. The HCl eluent was then evaporated
to dryness and the residue washed with methanol. Evaporation of the methanol layer
yielded a colorless oily product 3-deoxy-D-glycero-pentulosonic acid with > 99% purity as
indicated by lH and 13 C NMR.

Oxidative decarboxylation of purified 3-deoxy-D-glycero-pentulosonic acid in
water under reflux yielded trace amount of 3,4-dihydroxy-D—butanoic acid (Table 26, entry

1). Attempts to decarboxylate 3-deoxy-D-glycero-pentulosonic acid under nitrogen

162

atmosphere was unsuccessful (Table 26, entry 2). Addition of 3% H202 to the reaction
mixture yielded 3,4-dihydroxy-D-butanoic acid in 90% under room temperature conditions
(Table 26, entry 3). Quantitative conversion was obtained with the addition of 5 mol%
FeSO4 as a catalyst. This synthesis was further improved using 1.5% H202 and 1% FeSO4
while the same conversion was maintained (Table 26, entry 4 — 6). To take a step further,
quantitative oxidative decarboxylation of 3-deoxy-D-glycero-pentulosonic acid in
fermentation broth KIT15/pML7.l35 without prior purification was achieved even
without extra FeSO4 addition as catalyst (Table 26, entry 7 — 8). The fact that this reaction
took only 15 min to complete certainly made it further appealing and commercially viable.
Research is now focused on developing a reactive extraction method for (S)-3-hydroxy-y-
butyrolactone. In an ideal case, both acid-catalyzed lactonization and purification will be

achieved in a single extraction step in Karr reciprocating column using ethyl acetate.

Table 26. Oxidative decarboxylation of 3-deoxy-D-glycero-pentulosonic acid.

 

 

_ . temp time H202 FeSO4 conversion
entry starting ketoacrd (°C) (min) (%) (mol%) (%) note

1 1g purified 100 30 0 0 trace

2 1g purified 100 30 0 0 trace purged with N2
3 1g purified rt 15 3 0 90

4 1g purified rt 15 3 5 quant.

5 1g purified rt 15 1.5 5 quant.

6 1g purified rt 15 1.5 1 quant.

7 49 in brotha rt 15 1.5 1 quant.

8 49 in brotha rt 15 1.5 0 quant.

 

aFermentation broth contained 3-deoxy-D-glycero—pentulosonic acid at a concentration of 21 g/L.
Cells and protein were separated by ultra-filtration prior to use.

163

Enzyme activity studies for E. coli ath expressing microbes

The studies on alcohol dehydrogenases is important to fulfill our understanding to
otherwise unknown native E. coli enzyme conversions of 3,4-dihydroxy-D-butanal into D-
1,2,4-butanetriol. Among the five E. coli alcohol dehydrogenases identified using
bioinformatics, only three showed in vitro enzyme activity using racemic 3,4-dihydroxy-
D-butanal as substrate. The benzoylformate decarboxylase activity time-course study
indicated that both E. coli WN13/pWN7.126B and its derivative KIT10/pWN7.126B with
chromosomal ath inactivation appeared to be similar during fermentation, and achieved
the highest value of 65 and 57 U/mg, respectively, at 48 h (Figure 89a). On the other
hand, no measurable D-1,2,4-butanetriol dehydrogenase activity was observed for
KIT10/pWN7.126B during the fermentation (Figure 89b). Since these two constructs
expressed about the same level of benzoylformate decarboxylase activity, activity of Ath
became the only possible factor that could contribute to the difference in D-1,2,4-
butanetriol production.

Construction of E. coli microbes by co-expressing 1,2,4-butanetriol dehydrogenase
candidates E. coli ath, Z. mobilis adhA and K. pneumoniae dhaT genes together with P.
putida mdlC gene in a single multi-copy plasmid had led to poor cell growth and reduced
decarboxylase activity (Figure 89a). Research was then focused on E. coli ath to avoid
potential complications due to heterologous gene expression of dhaA and dhaT in E. coli.
Our efforts toward studying ath expression for D-l,2,4-butanetriol synthesis are
summarized in Table 27. In the case of E. coli WN13/pML6.195 (E. coli ath on a
plasmid), only 32 U/mg benzoylformate decarboxylase activity was expressed at the end

of the fermentation, which was half of the activity observed for other constructs with a

164

single mdlC gene localized plasmid (Figure 89a). On the other hand, modulated
expression of ath on the E. coli chromosome afforded E. coli KIT4/pWN7.126B and E.
coli KIT7/pWN7.126B. These microbes successfully produced 11.5 g/L and 11 g/L D-
1,2,4-butanetriol, which was an average of 12% improvement in concentration over the
benchmark E. coli WN13/pWN7.126B. The benzoylformate decarboxylase activity of
these two constructs was restored as indicated by the enzyme activity time-course study
(Figure 89a). In addition, while maintaining the same level of decarboxylase activity,
ath gene encoding dehydrogenase expression was also achieved in these constructs. In
the case of E. coli KIT4/pWN7.126B, Ath activity attained a maximum of 0.05 U/mg at
48 h, which corresponded to a five-fold increase comparing to the wild-type E. coli
(Figure 89b). These enzyme activity studies provide a second line of evidence that E. coli
ath is indeed the dehydrogenase that participates in the D-1,2,4-butanetriol biosynthesis.
Due to the fact that inactivating the ath gene did not shut down D-l,2,4-butanetriol
biosynthetic pathway, it becomes tempting to speculate that one or more other native E.
coli dehydrogenases might be acting as in vivo D-1,2,4-butanetriol dehydrogenase. Last
but not least, expressing ath chromosomally under a tac or phage T5 promoter only
afforded slight improvement in D-1,2,4-butanetriol production, which provides an
additional support that MdlC catalyzed 3-deoxy-D-glycero-pentulosonate decarboxylation
is indeed the limiting step in the pathway. Although preliminary cultivation of the second
generation microbe E. coli KIT18/pWN7.126B successfully achieved D-1,2,4-butanetriol
synthesis at a concentration of 18 g/L, it is believed that an active 3-deoxy-D-glycero-
pentulosonate decarboxylase will determine the ultimate success of this pentose-based

pathway.

165

 

 

 

 

 

 

 

100 0.20
(a) (b)
75 1 0.15 1
a A
s E
3 50 1 a 0.10
E 5‘
>
g '8
25 ~ " 0.05
0 J—- . . 0.00 -
0 12 24 36 48 0 12 24 36 48
time (h) - tlmelh)

Figure 89. Enzyme activity time-course for microbial syntheses of D-1,2,4-
butanetriol under fermentor-controlled conditions. (a) benzoylformate decarboxylase
activity. (b) 1,2,4-butanetriol dehydrogenase activity. E. coli WN13/pWN7.126B
(square); E. coli WN13/pML6.195 (diamond); E. coli KIT10/pWN7.126B (triangle); E.
coli KIT4/pWN7.126B (asterisk); E. coli KIT7/pWN7.126B (circle). One unit of activity
is defined as 1 mol of product formation per minute.

Table 27. E. coli strains with ath gene expression.

 

 

 

 

genomic ath titer, g/L (yield, %) activity, U/mg

construct R

Ptac P15 knockout Cm BT BA MdlC Ath
WN13/pWN7.1268 «J 10.5(50) 4.5(19) 67 0.006
WN13/pML6.195 1! 4.6(22) 4.2(17) 38 0.08
KIT2/pWN7.126B 9.1(43) 4.5(19) 77 0.004
KIT3/pWN7.126B 1] 1! 10.0(47) 4.4(18) 35 0.02
KIT4/pWN7.126B \/ 11.5(55) 4.5(19) 64 0.02
KIT6/pWN7.1268 1! \/ 9.8(46) 4.5(19)
KIT7/pWN7. 1268 ‘1 11.0(52) 4.7(20) 51 0.02
KIT9/pWN7.126B \/ \/ 8.2(39) 5.0(21)
KIT10/pWN7.1263 1! 6.5(31) 5.3(22) 57 0

 

CmR, chloramphenicol gene on the host chromosome; BT, D-1,2,4-butanetriol; BA, 3.4-dihydroxy-
D—butanoic acid; MdlC. P. putida benzoylformate decarboxylase; Ath, E. coli alcohol
dehydrogenase.

166

REFERENCE

‘ Daniel, R.; Bobik, T. A.; Gottschalk, G. FEMS Microbiol. Rev. 1999,22, 553.

2 Lawson, M. E. In Kirk-0thmer Encyclopedia of Chemical Technology; Sugar Alcohols,
1991, Wiley.

3 http://wwwcargill.com/today/releases/02_O6tl6ificonf.htm.

4 (a) Lin, S.-J.; Wen, C.__J.; Liau, J .-C.; Chu, W. S. Process Biochem. 2001, 36, 1249. (b)
Kasumi, T. Jap. Agricul. Res. Quart. 1995, 29, 49.

5 Veiga—da-Cunha, M.; Santos, H.; Van Schafiingen, E. J. Bacteriol. 1993, 175, 3941. (b)
Veiga-da-Currha, M.; Firrne, F.; San Romao, M. V.; Santos, H. Appl. Environ. Microbiol.
1992, 58, 2271.

6 Seyfried, M.; Daniel, R.; Gottschalk, G. J. Bacteriol. 1996, 1 78, 5793.

7 Toraya, T.; Shirakashi, T.; Kosuga, T.; Fukui, S. Biochem. Biophys. Res. Commun. 1976,
69, 475.

8 Abeles, R. H.; Dolphin, D. Acc. Chem. Res. 1976, 9, 114.

9 Toraya, T. In Vitamin 812 and BIZ-Proteins; Krautler, B.; Arigoni, D.; Golding, B. T.
Eds.; Wiley-VCH: Weinheim, 1998; p303.

'0 Boenigk, R.; Bowien, S.; Gottschalk, G. Appl. Microbiol. Biotechnol. 1993, 38, 453.
" Daniel, R.; Stuertz, K.; Gottschalk, G. J. Bacteriol. 1995, 177, 4392.

'2 Roth, J. R.; Lawrence, J. G.; Bobik, T. A. Annu. Rev. Microbiol. 1996,50, 137.

‘3 Toraya, T.; Honda, 3.; Fukui, s. J. Bacteriol. 1979, 139, 39.

'4 Badia, J.; Ros, J.; Aguilar, J. J. Bacteriol. 1985, 161, 435.

‘5 Scarlett, F. A.; Turner, J. M. J. Gen. Microbiol. 1976, 95, 173.

'6 Tobimatsu, T.; Hara, T.; Sakaguchi, M.; Kishimoto, Y.; Wada, Y.; Isoda, M.; Sakai, T.;
Toraya, T. J. Biol. Chem. 1995, 270, 7142.

167

‘7 Poznanskaja, A. A.; Tanizawa, K.; Soda, K.; Toraya, T.; Fukui, S. Arch. Biochem.
Biophys. 1979, 194, 379.

'8 Tobimatsu, T.; Azuma, M.; Matsubara, H.; Takatori, H.; Niida, T. Nishimoto, K.; Satoh,
H.; Hayashi, R.; Toraya, T. J. Biol. Chem. 1996, 271, 22352.

‘9 (a) Bachovchin, w. W.; Moore, K. W.; Richards, J. H. Biochemistry, 1977, I6, 1082.
(b) Schneider, 2.; Pawelkiewicz, J. Acta Biochim. Pol. 1966, I3, 311. (c) Poznanskaya, A.
A.; Yakusheva, M. I.; Yakovlev, V. A. Biochim. Biophys. Acta. 1977, 484, 236.

20 Toraya, T.; Fukui, S. In Big; Dolphin, D. Ed.; John Wiley & Sons, Inc.: New York,
1982;p233.

21 (a) Honda, 8.; Toraya, T.; Fukui, S. J. Bacteriol. 1980, 143, 1458. (b) Ushio, K.;
Honda, 8.; Toraya, T.; F ukui, S. J. Nutr. Sci. Vitaminal. 1982, 28, 225.

22 Homann, T.; Tag, C.; Biebl, H.; Deckwer, W.-D.; Schink, B. Appl. Microbial.
Biotechnol. 1990, 33, 121.

23 Daniel, R.; Gottschalk, G. FEMS Microbial. Lett. 1992, 100, 281.
24 Tong, I.-T.; Liao, H. H.; Cameron, D. C. Appl. Environ. Microbial. 1991, 57, 3541.
25 Daniel, R.; Boenigk, R.; Gottschalk, G. J. Bacterial. 1995, I 77, 2151.

26 Sprenger, G. A.; Hammer, B. A.; Johnson, E. A.; Lin, E. C. C. J. Gen. Microbial. 1989,
135, 1255.

27 Tobimatsu, T.; Azuma, M.; Hayashi, S.; Nishimoto, K.; Toraya, T. Biosci. Biotechnol.
Biochem. 1998, 62, 1774.

28(a) Toraya, T.; Uesaka, M.; Kondo, M.; Fukui, S. Biochem. Biophys. Res. Commun.
1973, 52, 350. (b) Toraya, T.; Uesaka, M.; F ukui, S. Biochemistry 1974, 13, 3895.

29 Bobik, T. A.; Xu, Y.; Jeter, R. M.; Otto, K. E.; Roth, J. R. J. Bacterial. 1997, I 79, 6633.
3" Dabrock, 13.; Bahl, 11.; Gottschalk, G. Appl. Environ. Microbiol. 1992,58, 1233.

3' Heindrickx, M.; de Vos, R; Vancanneyt, M.; de Ley, J. Appl. Microbial. Biotechnol.
1991, 34, 637.

168

32 (a) Speranza, G.; Manitto, P.; Fontana, G.; Monti, D.; Galli, A. Tetrahedron Lett. 1996,
37, 4247. (b) Speranza, G.; Corti, S.; Fontana, G.; Manitto, P. J. Agric. Food Chem. 1997,
45, 3476. (c) Manitto, P.; Speranza, G.; Fontana, G.; Galli, A. Helv. Chim. Acta 1998. 81,
2005. (d) Speranza, G.; Morelli, C. F.; Orlandi, M.; Scarpellina, M.; Manitto, P. Helv.
Chim. Acta 2001, 84, 335.

33 Seifert, C.; Bowien, s.; Gottschalk, G.; Daniel, R. Eur. J. Biochem. 2001, 268, 2369.

34 Luers, F.; Seyfried, M.; Daniel, R.; Gottschalk, G. FEMS Microbiol. Lett. 1997, I54,
337.

35 Niu, w. PhD Thesis, Michigan State University, 2004.
36 Niu W.; Frost, J. W. unpublished results.

37 Toraya, T.; Kazutoshi, U.; Fukui, S.; Hogenkamp, H. P. C. J. Biol. Chem. 1977, 252,
963.

38 Johnson, E. A.; Lin, B. c. C. J. Bacterial. 1987, 169, 2050.

39 (a) Niu, W.; Molefe, M. N.; Frost, J. W. J. Am. Chem. Soc. 2003, 125, 12998. (b) Frost,
J. W. WO 2005/068642, 2005.

40 (a) Hegeman, G. D. J. Bacterial. 1966, 91, 1140. (b) Hegeman, G. D. J. Bacterial.
1966, 91, 1155. (c) Hegeman, G. D. J. Bacteriol. 1966, 91, 1161. (d) Fewson, C. A.
FEMS Microbiol. Rev. 1988, 54, 85.

41 (a) Hasson, M. S.; Muscate, A.; Henehan, G. T.; Guidinger, P. F.; Petsko, G. A.; Ringe,
D. Kenyon, G. L. Biochemistry 1998 37, 9918. 0)) Klunger, R. Pure Appl. Chem. 1997,
69, 1957. (c) Weiss, P. M.; Garcia, G. A.; Kenyon, G. L.; Cleland, W. W.; Cook, P. F.
Biochemistry 1998, 27, 2197. (d) Reynolds, L. J .; Garcia, G. A.; Kozarich, J. W.; Kenyon,
G. L. Biochemistry 1998, 27, 5530.

42 Polovnikova, E. S.; Mcleish, M. J.; Sergienko, E. A.; Burgner, J. T.; Anderson, N. L.;
Bera, A. K.; Jordan, F.; Kenyon, G. L.; Hasson, M. S. Biochemistry 2003, 42, 1820.

43 (a) Bommarius, A. S.; Riebel, B. R. Biacatalysis; Wiley-VCH: Weinheim, 2004. (b)
Krawetz, S. A.; Womble, D. D. Introduction to Bioinformatics: A Theoretical and
Practical Approach; Humana Press: Totowa, NJ, 2003.

44 http://www.ncbi.nlm.nih.gov/

169

45 http://www.ebi.ac.uk/

46 http://wit.integratedgenomics.com/GOLD/

47 http://wwwncbi.nlm.nih.gov/BLAST/

48 http://www.ebi.ac.uk/clustalw/
49 http://pvmol.sourceforgenet/
50 http://www.rcsb.org/pdb/home/home.do

5' (a) Siegert, P.; Pohl, M.; Kneen, M. M.; Pogozheva, I. D.; Kenyon, G. L.; McLeish, M.
J. In Thiamine: Catalytic mechanisms in normal and disease states; Jordan, F.; Patel, M.
S. Eds; Marcel Dekker, Inc.: NY, 2004; p275. (b) Pohl, M.; Siegert, P.; Mesch, K.;
Bruhn, H.; Grotzinger, J. Eur. J. Biochem. 1998, 25 7, 538.

52 http://www.stratagenecom/manuals/2OOS 1 8.pdf

53 Conway, T.; Osman, Y. A.; Konnan, J. I.; Hoffmann, E. M.; Ingram, L. O. J. Bacteriol.
1987, 169, 949.

54 Arnold, F. H.; Georgiou, G. Directed Evolution Library Creation: Methods and
Protocols; Humana Press: Totowa, NJ, 2003. (b) Arnold, F. H.; Georgiou, G. Directed
Enzyme Evolution: Screening and Selection Methods; Humans Press: Totowa, NJ, 2003.

55 (a) Chen, K.; Arnold, F. H. Proc. Natl. Acad. Sci. USA 1993, 93, 5618. (b) Powell, K.
A.; Ramer, S. W.; Cardayre, S. B.; Stemmer, W. P. C.; Tobin, M. B.; Longchamp, P. F.;
Huisman, G. W. Angew. Chem. Int. Ed. 2002, 40, 3948. (c) Schmidt-Dannert, C.
Biochemistry 2001, 40, 13125. (d) Tao, H.; Cornish, V. W. Curr. Opin. Chem. Biol. 2002,
6, 858.

56 Arnold, F. H.; Wintrode, P. L.; Miyazaki, K.; Gershenson, A. Trends Biochem. Sci.
2001, 26, 100.

57 Joem, J. M.; Meinhold, P.; Arnold, F. H. .1. M01. Biol. 2002, 316, 643.
58 Burchert, J .; Viikari, L.; Linko, M.; Markkanen, P. Biotechnol. Lett. 1986, 8, 541.

59 Frost, J. W.; Niu, W.; Lau, M. K.; Dodds, D.; Saffron, C. unpublished results.

170

60 (a) Studier, F. W.; Moffatt, B. A. J. Mol. Biol. 1986, 189, 113. (b) Studuer, F. W.;
Rosenberg, A. H.; Dunn, J. J .; Dubendorff, J. W. Methods Enzymal. 1990, 185, 60.

6‘ Tabor, S.; Richardson, c. C. Proc. Natl. Acad. Sci. USA 1985,82, 1074.
62 http://www.emdbiosciences.com/docs/docs/PROT/TB031 .pdf

63 (a) Thomas, J. G.; Ayling, A.; Baneyx, F. Appl. Biochem. Biotechnol. 1997, 66, 197.
(b) Wall, J. G.; Pluchthun, A. Curr. Opin. Biotechnol. 1995, 6, 507.

64 (a) Nichihara, K.; Kanemori, M.; Kitagawa, M.; Yanagi, H.; Yura, T. Appl. Environ.
Microbial. 1998, 64, 1694. (b) Nishihara, K.; Kanemori, M.; Yanagi, H.; Yura, T. Appl.
Environ. Microbiol. 2000, 66, 884.

65 (a) Horovitz, A. Curr. Opin. Struct. Biol. 1998, 8,93. 05) Sigler, P. B.; Xu, z. H.; Rye,
H. S.; Burston, S. G.; Fenton, W. A.; Horwich, A. L. Annu. Rev. Biochem. 1998, 67, 581.
(c) Thirumalai, D.; Lorimer, G. H. Annu. Rev. Biophys. Biamal. Struct. 2001, 30, 245.

66 Henning, H.; Leggewie, C.; Pohl, M.; Muller, M.; Eggert, T.; Jaeger, K.-E. Appl.
Environ. Microbiol. 2006, 72, 7510.

67 Datsenko, K. A.; Wanner, B. L. Proc. Natl. Acad sci. USA 2000, 97, 6640.

68 Cherepanov, P. P.; Wackemagel, W. Gene 1995, 158, 9.

69 (a) McLaggan, D.; Naprstek, J .; Buurrnan, E.; Epstein, W. J. Biol. Chem. 1994, 269,
1991. (b) Yan, D.; Ikeda, T. P.; Shauger, A. E.; Kustu, S. Proc. Natl. Acad. Sci. USA
1996, 93, 6527.

70 (a) Ohta, K.; Beall, D. S.; Mejia, J. P.; Shanmugam, K. T.; Ingram, L. 0. Appl. Environ.
Microbial. 1991, 5 7, 893. (b) Mantinez, A.; York, S. W.; Yomano, L. P.; Pineda, V. L.;
Davis, F. C.; Shelton, J. C.; Ingram, L. O. Biotechnol. Prog. 1999, 15, 891.

7‘ Cadwell, R. C.; Joyce, G. F. PCR Meth. Appl. 1992, 2, 28.

72 Kieman, J. A. Histological and histachemical methods: Theory and practice 3rd ed.;
Butterworth Heinemann: Oxford, UK, 1999.

73 Saito, S.; Hasegawa, T.; Inaba, M.; Nishida, R.; Fujii, T.; Nomizu, S.; Moriwake, T.
Chem. Lett. 1984, 1389.

171

74 (a) Taber, D. F.; Raman, K. J. Am. Chem. Soc. 1983, 105, 5935. (b) Nakamura, K.;
Kawai, Y.; Ohno, A. Tetrahedron Lett. 1990, 31, 267.

75 Rowell, R. M.; Somers, P. J.; Barker, S. A.; Stacey, M. Carbohydrate Res. 1969, 11, 17.
76 Green, J. w. J. Am. Chem. Soc. 1956, 78, 1894.
77 Jacks, T. B. we 9804543, 1998.

78 (a) Whistler, R. L.; Schweiger, R. J. Am. Chem. Soc. 1959, 81 , 3136. (b) Arts, S. J. H.
F. F.; Mombarg, E. J. M.; van Bekkum, H.; Sheldon, R. A. Synthesis 1997, 597.

172

CHAPTER FOUR

Downstream Purification of 1,2,4-Butanetriol

Background

The microbial synthesis of 1,2,4-butanetriol from renewable resources is an
attractive process. However, the recovery of this high boiling hydrophilic compound from
a fermentation broth is a challenging task. This is also considered critical for the
development of a commercially viable process. Conventionally, vacuum distillation is
often used in processes for the purification of other polyhydroxy compounds such as
glyceroll and 1,3-propanediol.2 To distill a dilute aqueous solution requires a substantial
energy input that ads to the cost of the final product. Earlier attempts in the Frost group to
distill the 1,2,4-butanetriol containing reaction mixture from catalytic malic acid
hydrogenation 3 and chemo-enzymatic synthesis using 2-deoxyribose-5-phosphate
aldolase4 under vacuum resulted in poor product recovery. Undesired polymer and
decomposition products were recovered at the end of the process. Thus, an alternative
method for the purification of 1,2,4-butanetriol is needed. Replacement of distillation by
energy efficient solvent extraction is believed to be an alternative to reduce the cost of the
product separation significantly. It is also suitable for large scale operation. More
importantly, solvent extraction does not require excessive heat input to vaporize the
compound, therefore issues associated with polymerization and decomposition are thus
eliminated. To this end, research involved a systematic and comprehensive examination
of the liquid-liquid extraction of 1,2,4—butanetriol from fermentor broth. Potential solvent

candidates were screened using a glass continuous liquid-liquid extractor apparatus. A

173

scalable downstream process was then designed using a bench scale Karr reciprocating
column (Koch Modular Process Systems, KMPS).

Setting up an efficient liquid-liquid extraction process is often neglected because it
only contributes to a minor part of the overall process scheme. However, significant
reduction of process cost can be achieved by fine-tuning an extraction system. Liquid-
liquid extraction is a mass transfer operation in which a liquid solution (the feed) contacts
with an immiscible or nearly immiscible liquid (the solvent) that exhibits preferential
affinity towards the target compound (the solute) in the feed.5 Feed and solvent streams
resolve after the contact of these two phases. The extract is a solvent rich solution that
contains the desired solute, while the raffinate refers to the solute lean feed solution after
the extraction.

To design an efficient liquid-liquid extraction process, solvent selection, choice of
extraction equipment and operating conditions represent different parameters that require
careful consideration and optimization. Among these, a suitable solvent choice is of
primary importance. This solvent must possess a relative low boiling point, therefore
separates easily from the extracted components. Little or no miscibility with the feed
solution is preferred. Corrosiveness, toxicity, flammability and stability are certainly
other concerns. The solvent must also be cost effective, and readily available for a
commercial operation. Additional consideration such as solvent recovery and solvent
release are gaining more attention due to stricter regulations on solvent release and to
lower the overall process cost. An ideal single solvent for an extraction system can
usually be theoretically judged, but the performance of solvent blends is difficult to

predict solely by threoretical means. Therefore, actual experimentation still occupies a

174

crucial role to make more accurate decisions. Once the solvent candidates are identified,
measuring distribution coefficients (m) for the solute of interest and selectivity tests are
the next tasks to be accomplished. The distribution coefficient is defined as the ratio of
the concentration of solute in extract phase to the concentration in the raffinate.6 On the
other hand, selectivity concerns the ability of the solvent to pick up that particular solute
versus other components in the feed. The solvent of choice has a high distribution
coefficient and a good selectivity towards the solute.

The most commonly used single-stage extracting device in the laboratory is a
separatory funnel. The industrial equivalence of it is a mixer-settler that was developed in
the past among pharmaceutical and agrochemical industries (Figure 90a).‘5 These extractor
tanks are used today for washing and neutralization processes that require only a few
stages. Extractions are usually operated in a cross-current mode and are particularly
flexible and economical for batch operations. When this extraction tank is in operation,
solvent is first added to the feed and mechanically mixed. This mixture will then be
allowed to settle in a different chamber where it resolves into two layers. Single stage
extraction can be easily achieved using this apparatus. If more than one theoretical stage
is required, multiple tanks will be setup in series. In this case, a mixer-settler may not be a
desirable apparatus. Using the wrong apparatus will lead to an unnecessary waste of
solvent and low extraction yields. Another similar industrial scale extractor is a
centrifugal extractor (Figure 90b). It features a high-speed centrifuge that offers the
advantage of low residence time. Centrifugal extractors are ideal for systems in which the
density difference between the feed and solvent is very small. A mechanical device is

used to agitate the mixture by increasing the interfacial area and decreasing mass transfer

175

resistance. In most cases, the number of stages in a centrifugal extractor is limited to one.
In contrast to the mixer-settler, centrifugal extractors are usually associated with high

setup and maintenance cost.

extract

      
 
 

\
\

:\\

solvent

    
 
   

i/ 5;“ l,
4% >9; ,
”' dis ersion feed 1%? e" ‘—
agitator p extract [gig/9’7)?
J . 7g’r'./'l‘{'/>/‘..‘ 7."; o

\ i. ,
.\_.
)5.

‘\ - i". . - .‘ 0'

feed solvent

‘
h “I .
v .. .. 1.
\l' l
. .,
\ .. s. _ u
1 . v ' a» -‘ _ c: ‘ v "4
se er . . -‘ _ _I; _ 7'
-5.-.an ‘ - _ , r", ”‘7 :7 _ .
‘ \‘ 1 ' . 9,. - j ' ._._.
~ 7 ' — _ ‘I”’; 1“, '
. I 1‘ "._~ .‘I'
'1 . , ‘
l .
I . c
, I
- 1
I

-4 gal) raffinate

5‘

 

raffinate

 

Figure 90. Single-stage liquid-liquid extractors. (a) a mixer-settler.7 (b) a centrifugal
extractor (Baker-Perkins, Saginaw, MI).

Counter-current column extractors are probably the most popular in the chemical
industry. They can be static or agitated. A static column provides a moderate number of
stages with relatively low operating and maintenance cost. The packed colrunn, tray
column and spray column have different materials arranged in the column and these
materials are used to increase the interfacial area in which the two phases contact.
However, scaling up a laboratory-scale static column process into an industrial operation
is not simple, and less efficient than the mixer-settler process. Agitated columns offer
advantages such as reasonable capacity, low operating cost, and high efficiency. An
agitated column typically requires less than a third of the number of theoretical stages as

compared to a static column. These columns are therefore ideal for any system that a high

176

number of theoretical stages are required. An associated reduction of solvent usage is also
beneficial to the process economics. Agitation in these columns are supplied either by a
rotating disc, which usually. runs at much higher speeds than a turbine type impeller
(Scheibel column, Figure 918), or with a reciprocating pump, in which a rapid
reciprocating motion of baffled and perforated plates are used to create a dispersion (Karr
column, Figure 91b). Among these two, Karr columns have an important advantage over
Scheibel columns to handle difficult extraction systems that tend to emulsify or flood
easily. In the Karr column, a rapid reciprocating motion with small amplitude is created
by a pump equipped in the column to superimpose the normal counter-current flow of the
liquid phases due to the solvent/feed density difference. The pulsation of the perforated
plates causes the light liquid to be dispersed into the heavy phase on the upward stroke
and heavy liquid phase to jet into the light phase on the downward stroke. It generates a
certain number of theoretical stages required to transfer solute from one phase to the other.
By increasing the surface area and turbulence, a more efficient extraction process will be

achieved.

177

   
   

. Variable Speed
(3) =1 / Drive
Light
Heavy Phase Out
Phase In w”, . . .2 ‘5‘

 
  
 

V N x.

4“" Rotating

,1." Shaft \n,
1’5 Horizontal Vessel \
Outer Baffle Walls \ 1‘

    
  

 

 
    

E‘s, Turbine 'Horizontal /
1..“ Impeller InnerBaffle'g

 

.1

 
   

‘1' .l
4.. '54.,“ ”fr
interface Interface 1* » co»-
1.. (Control

 
 

Heavy
Phase Out

 
 

D 1
(b) Adstnbly "’

   
 
  
   

......... . .. Tle Rods
. , 4 & Spacers

Plate

/ Baffle Plate

II ‘
Interface
4,, Control

 
 

Heavy
Phase Out

Figure 91. Agitated counter-current extraction columns.8 (a) Scheibel column. (b)
Karr column.

178

Solvent candidates screening using continuous extraction method

Residual protein removal by membrane ultra-filtration

 

 

 

EtOAc—-'

 

 

 

 

lennentation broth-->

 

 

 

g=_J

Figure 92. Glass continuous liquid-liquid extractor (Sigma-Aldrich).

Fermentation broth used in this study was obtained by the cultivation of E. coli
microbe WN13/pWN7.126B. Cells were removed by centrifugation at 14,000 rpm for 10
min. Extraction experiments were carried out in a glass continuous liquid-liquid
extraction apparatus obtained from Sigma-Aldrich. A typical extraction started by filling
the extraction tower with 100 mL of fermentation broth. The solvent of interest was then
poured into the same tower until it reached the side-arm. A 250 mL round-bottomed flask
filled with 100 mL of the same solvent was attached to the other end of the side-arm.
Finally, a condenser was fitted to the top of the extraction tower. The extraction began by

refluxing the solvent inside the round-bottomed flask. The extractant evaporated,

179

condensed and directed back into the dispersion created inside the extraction tower by
vigorous magnetic stirring. Due to the density difference, D-1,2,4-butanetriol-rich
extractant was eventually separated from the feed and collected in the round-bottom flask.
At the end of the extraction, D-l,2,4-butanetriol was accumulated with extractant inside
the round-bottom flask. Subsequent removal of solvent was achieved by rotary
evaporation and subjected to 1H NMR and gas chromatographic (GC) analyses.

Direct D-l,2,4-butanetriol extraction from cell-free fermentation broth using ethyl
acetate as solvent was unsuccessful due to protein precipitation (Table 28, entry 1).
Methods were developed to remove residual soluble proteins in the fermentation broth.
Earlier research in the Frost group clarified broth by adjusting the pH of the cell-free broth
to 2.5 using concentrated H2804. This method was used to treat D-l,2,4-butanetriol broth
and the protein precipitate was pelleted using centrifugation. Subsequent extraction of this
acidified broth led to decomposition of D-1,2,4-butanetriol (Table 28, entry 2). In another
attempt, this acidified broth was neutralized using 10 N KOH and the extraction started
once again. In this case, D-1,2,4-butanetriol was recovered in the solvent without
decomposition (Table 28, entry 3). As an alternative approach, charcoal treatment was
found to be an attractive way to remove protein by adsorption (Table 28, entry 4). It
eliminated the extra salt stream created during acidification and neutralization of the
fermentation broth. As an additional advantage, charcoal addition was also found useful
in reducing the product color and unpleasant odor. The optimal way to remove protein
was achieved by membrame ultra-filtration. Fitting an Amicon concentrator with a 10,000

MWCO Miomax membrane (Millipore) yielded clarified broth without extra process cost

180

due to acid/base or charcoal addition (Table 28, entry 5). Either charcoal or membrane-

treated broth yielded D-1,2,4-butanetriol upon ethyl acetate extraction.

Table 28. Methods for residual protein removal from fermentation broth.

 

 

entry conditions recovery
1 direct extraction (pH 7) failed
2 protein was precipitated at pH 2.5 (H2804) BT decomposed

extracted for 48 h, replaced w/fresh EtOAc every 12 h
no organics left in the raffinate

3 protein was precipitated at pH 2.5 (H2804) BT recovered
pH adjusted to 6.5 w/ KOH
extracted for 24 h, NMR samples were taken every 12 h

4 10 g/L charcoal was added to the cell-free broth BT recovered
filtered through Celite
extracted for 24 h, NMR samples were taken every 12 h

5 cell-free broth was subjected to ultra-filtration BT recovered
extracted for 24 h, NMR samples were taken every 12 h

 

Liquid-liquid extraction using a single solvent system

A list of common industrial solvents was obtained from the Kirk-Othmer
Encyclopedia of Chemical Technology.9 Potential solvent candidates for the D-1,2,4-
butanetriol extraction were identified based on several considerations. Solvent boiling
point lower than 120 °C was preferred to avoid potential D-l,2,4-butanetriol
decomposition. Water solubility was certainly another important factor in liquid-liquid
extraction. Organic solvents that are miscible with water could not be used. A difference
in the density of saturated liquid phases should be as large as possible for physical
separation of the two phases. As the solvent must be recovered for reuse, the formation of

an azeotrope with the extracted solute is not desirable. The solvent should also be

181

inexpensive and non-toxic if possible. Lastly, the solvent should be chemically stable and

inert to other components in the system.

Table 29. Single solvent candidates.

 

 

entry solvent boiling point (°C) water solubility (wt%)
1 ethyl acetate 77 3
2 1-butanol 118 9
3 2-butanol 98 12.5
4 3-pentanone 101.5 5
5 2-butanone 80 25
6 nitromethane 101 10

 

Single solvent candidates were screened and are summarized in Table 30. The
extraction experiments were performed using the same protocol as described previously.
The 2-butanol extraction experiment was sampled every hour, while the rest of the solvent
candidates were sampled every 12 h. Extractions were performed for a maximum of 24 h
when needed. Both GC based and isolated recovery were calculated using these samples.
The total recovery was obtained by adding up the recoveries at different time-points. The
purity was based on the gas chromatogram of the isolated D-1,2,4-butanetriol. At this
early stage of our study, final D-1,2,4-butanetriol quality was of primarily importance.
The rate of recovery came second and was tuned at a later stage of this work.
Experimental results revealed that ethyl acetate (Table 30, entry 1) and 2-butanone (Table
30, entry 5) possessed excellent selectivity toward D-1,2,4-butanetriol, however, the
extractions were performed at a relatively low rate. On the other hand, extractions with 1-
butanol (Table 30, entry 2) or 2-butanol (Table 30, entry 3) as solvent achieved at least
ten-fold improvement in the rate of recovery. As a drawback, impurities were also

extracted into these two solvents. Extraction using 3-pentanone (Table 30, entry 4) was

182

unsuccessful due to poor recovery (40% GC based). Attempts to use nitromethane as a
solvent, which has a higher density than water also failed. A dispersion between the feed
and the solvent was not formed during extraction with nitromethane. During the study,
inconsistency between the GC-based and isolated total recovery was observed, indicating
the accumulation of NMR and GC inactive impurities (ie. poly-glycolslo) in the product
after prolonged heating. In addition, attempts to scale up the process using a 1 L glass
liquid-liquid extractor experienced a significant decrease in the rate of recovery. A dark
yellowish oil was obtained even using ethyl acetate as solvent. These observations
indicate that continuous extraction using a single solvent system is not a good method. To
address this issue, other extraction alternatives using a solvent—blend system and counter-

current extraction column were explored.

Table 30. Summary of single solvent screening experiments.

 

time BT extracteda (g) BT recovery (%) total recovery (%) purity”

 

 

entry “Went (h) GC isolated GC isolated GC isolated (%)

1 ethyl acetate 12 0.24 0.22 44 40 64 51 >99
24 0.1 1 0.06 20 1 1

2 1-butanol 12 0.53 0.92 96 167 96 167 >90

3 2-butanol 1 0.29 0.36 52 65 99 1 13 >90
2 0.26 0.26 47 48

4 3-pentanone 12 0.14 0.22 25 40 40 67 >99
24 0.08 0.15 15 27

5 2-butanone 12 0.41 0.58 75 105 75 105 >99

 

aThe charcoal clarified fermentation broth (100 mL) contained 0.55 g D—1,2,4-butanetriol.
bPurity was calculated based on the gas chromatogram of isolated D-1,2,4-butanetriol.

183

Liquid-liquid extraction using a solvent-blend system

Given that ethyl acetate provided the best selectivity of D-l,2,4-butanetriol as
discovered in the previous section, ethyl acetate and 2-butanone mixture were studied as
the first solvent-blend system. Three experiments were setup to evaluate different
concentration of 2-butanone (0%, 25% and 50%) in ethyl acetate. The extractions were
carried out for 24 hours using the same protocol. A sample was taken every 12 h and
analyzed by 1H, '3 C NMR and GC to quantify D-1,2,4-butanetriol and to determine the
product purity. All solvent system were successfully used to extract D-1,2,4-butanetriol in
99% purity (Figure 93). Increase in product recovery was observed upon increasing 2-
butanone concentration as shown in Table 31.

Ethyl acetate extraction with ethanol as a co-extractant in the charcoal clarified
fermentation broth was also studied. Different amount of ethanol (2%, 5%, 10%) was
added into the clarified broth before extraction started. The extraction experiments were
carried out for 24 h and was sampled every 12 h. Solvent was evaporated by rotary
evaporation under reduced pressure and the product was analyzed by 1H, 13C NMR and
GC. Similar to the 2-butanone/ethyl acetate case, EtOAc/EtOH extracted D-1,2,4-
butanetriol with over 99% purity (Figure 94). In addition, 87% isolated recovery was
successfully obtained by extracting clarified broth with 10% ethanol (Table 32). Although
ethyl acetate/ethanol extraction was proven to be a better approach relative to other tested
solvent systems, recycling two solvents was not economically viable for an industrial
process. Furthermore, scaling up a continuous liquid-liquid extraction process into
production scale is problematic. Nevertheless, solvent candidates screening using the

glass continuous extractor provided fundamental knowledge towards extracting D-1,2,4-

184

butanetriol from the fermentation broth, leading to the development of a scalable

extraction process using a Karr column.

 

180 160

1/

 

 

140

dodecane, 4.5 min

 

120 100 80 60

D-1,2,4-butanetriol, 6.5 min

/

 

 

4O 20 0

 

 

Figure 93. D-l,2,4-Butanetriol purified by 50% 2-butanone in ethyl acetate
extraction. (upper) 1H NMR spectrum, in ppm; (middle) 13 C NMR spectrum, in ppm;

(lower) GC chromatogram.

Table 31. Liquid-liquid extraction using 2-butanone/ethyl acetate solvent-blend.

 

2-butanone time

BT extracteda (%) BT recovery (%) total recovery (%) purity”

 

 

entry In EtO Ac (%) (h) GC isolated GC isolated GC isolated (%)
12 0.083 0.07 32 27

6 0 50 45 >99
24 0.047 0.047 18 18
12 0.091 0.1 36 39

7 25 58 66 >99
24 0.057 0.07 22 27
12 0.12 0.13 46 50

8 50 67 69 >99
24 0.055 0.05 21 19

 

aThe charcoal clarified fermentation broth (100 mL) contained 0.26 g D-1,2,4-butanetriol.
b Purity was calculated based on the gas chromatogram of isolated D-1,2,4-butanetriol.

185

 

 

 

‘_ . 4:. 'L‘.-——. 2
a.

 

180 160 140 120 100 80 60 40 20 0
D-1,2,4-butanetriol, 6.5 min

dodecane, 4.5 min
if/ it.

Figure 94. D-l,2,4-Butanetriol purified by 10% ethanol in ethyl acetate extraction.
(upper) 1H NMR spectrum, in ppm; (middle) 13 C NMR spectrum, in ppm; (lower) GC
chromatogram.

 

 

 

 

 

Table 32. Liquid-liquid extraction using ethanol/ethyl acetate solvent blend.

 

Ethanol in time BT extracteda(g) BT recovery (%) total recovery (%) purityb

 

 

6‘er broth (%) (h) GC isolated GC isolated GC isolated (%)
12 0.09 0.09 41 41
9 0 55 59 >99
24 0.031 0.04 14 18
12 0.106 0.1 48 46
10 2 63 55 >99
24 0.033 0.02 15 9
12 0.104 0.14 47 64
1 1 5 70 82 >99
24 0.054 0.04 24 18
12 0.1 1 0.14 50 64
12 10 80 87 >99
24 0.064 0.05 30 23

 

aThe charcoal clarified fermentation broth (100 mL) contained 0.22 g D-1,2,4-butanetriol.
bPurity was calculated based on the gas chromatogram of isolated D-1,2,4-butanetriol.

186

Development of a scalable extraction process using a bench-top Karr column

Bench-tap Karr column unit setup

A Karr reciprocating column (Koch Modular Process System, KMPS) is a counter-
current column extractor that is designed to handle difficult extraction systems having a
low interfacial tension and tends to emulsify. The bench-top Karr column unit is widely
used for feasibility testing during early stages of process development. By using this
apparatus, overall process feasibility, economics and different solvents will be evaluated.
In most applications, data generated in this bench-top unit will form the basis for a larger

scale pilot test program or even the design of a production scale column.

 

Figure 95. Benchtop Karr column in action.

187

The bench-top Karr column was set up in a laboratory fumehood. Three Fluid
Metering QD-Pump drive modules fitted with RH pump heads were used to control liquid
flows within this extraction system. This pump was setup to deliver a liquid flow from 8
mL/min to 86.25 mL/min to the system. Precise control over the flow could be adjusted
by a control ring graduated in 450 divisions. Two among these variable speed pumps
were used for charging feed and solvent, and a special Teflon coated Tygon tubing was
employed to deliver solvent to the column. The agitator assembly was powered by an air
driven motor to drive the agitation. A needle valve flow meter was provided to regulate
the gas flow, therefore providing an avenue to vary the agitation speed of the plate stack
assembly. In addition, the agitation speed (stroke per minute, spm) was monitored by a
handheld touchless digital tachometer. A Teflon and a stainless steel plate stack assembly
built on stainless steel shafts were supplied with the system. They were interchangeable to
meet different solvent characteristics. The third variable speed pump was responsible for
controlling an interface in the lower disengaging chamber and discharging the raffinate
out of the column. An optional control over very small raffrnate discharge rate could also
be accomplished manually using a metal clip. To maintain precise input/output rate and
material balance, a stop watch and a graduated cylinder were also used during the
experiment. The total volume of this column is 220 mL, and the operating throughput is
expected to be in the range between 15 — 120 mL/min. Manufacturer suggested agitation
speed is between 100 -— 400 spm. While systems with a large density difference between
the phases or high interfacial tension will be expected to operate in the high end of this

range, systems with opposite properties will be operated in the low end of this range.

188

Distribution coefficient for solvent candidates

Transition of the earlier knowledge obtained using a continuous liquid-liquid
extractor to a scalable process using a counter-current Karr extraction column can be
facilitated by understanding distribution coefficients between solvent candidates and D-
1,2,4-butanetriol. The distribution coefficient (m) is defined using the following formula:
m = y;l / x3, where a is the solute, ya is the concentration of component a in the solvent and
xa is the concentration of component a in the raffinate. The distribution coefficient for our
solvent candidates was calculated from the liquid-liquid equilibrium data set, which had
been generated experimentally using known concentration of authentic 1,2,4-butanetriol in
water and the solvent of interest. A typical liquid-liquid equilibrium data set was obtained
by vigorously mixing 1,2,4-butanetriol solution (50 mL) and solvent (50 mL) in a
separatory funnel and resolved back into two phases. The solvent was collected and the
concentration of 124-butanetriol in solvent and raffmate were determined using GC. The
raffinate was extracted again twice more and the distribution coefficient calculated from
each extraction was averaged at the end of the experiment. This averaged number
represents the distribution coefficient of a particular solvent. This method was applied on

eight solvent candidates and their distribution coefficients are summarized in Table 33.

Table 33. Distribution coefficients for various solvent candidates.

 

 

 

 

entry solvent m entry solvent m
1 2-butanol 0.24 5 1-pentanol 0.06
2 1-butanol 0.14 6 2-pentanol 0.06
3 iso—butanol 0.10 7 2-butanone 0.03
4 isoamyl alcohol 0.07 8 ethyl acetate 0.004

 

 

Abbreviation: distribution coefficient (m)

189

According to earlier experiments, ethyl acetate provided the best selectivity over
the others toward extracting D-1,2,4-butanetriol from the fermentation broth. However, it
is not suitable to be used in an industrial column extraction process due to its small
distribution coefficient towards D-l,2,4-butanetriol (Table 33, entry 8). Although 2-
butanol and l-butanol were less selective toward D-l,2,4—butanetriol that yielded products
with only 90% purity, they possess distribution coefficients that are two orders of
magnitudes greater than that using ethyl acetate as solvent (Table 33, entries 1 and 2).
Attempts to explore solvent alternatives yielded an extra five solvent candidates as iso-
butanol, isoamyl alcohol, l-pentanol, 2-pentanol and 2-butanone. From a distribution
coefficient standpoint, however, none of them shows a better potential over 2-butanol
(Table 33, entries 3 — 7). To this end, 2-butanol was the only solvent employed in the

following Karr column extraction experiments.

Start-up canditiansfar Karr column extractions

There are several conditions associated with the use of the bench-top Karr column.
Careful optimization of all parameters will not only maximize the performance of this
laboratory scale column but also provide a solid foundation to perform tests in a medium
scale pilot size column that is available on a rental basis at the KMPS facility located in
Houston, Texas. The pilot test is important to determine other system limitations before
significantly increasing the capacity by building an operating column.

The initial conditions for the Karr column extraction are different for each specific
case and were determined for the D-l,2,4-butanetriol extraction based on the operating

manual and consultation with Mr. Donald J. Glatz at KMPS. Prior to start-up, it was

190

important to determine the continuous phase of the system. Generally speaking, the
continuous phase is the higher flowing phase and will be filled in the column before the
extraction begins. In contrast, the lower flowing phase will be dispersed. Conventionally,
nearly all old cases involving fermentation broth column extraction, the broth (the feed) is
used as the continuous phase, while the solvent is the disperse phase.11 The material used
in making the plate stack poses another option based on the ‘wetting’ characteristics of the
system. If coalescing of the dispersed phase occurs on the plates and shaft spacers, then
either the phase should be reversed or the plate stack should be replaced with the alternate
plate stack made from a different material. In general, stainless steel works best for
aqueous phase continuous extractions, while teflon will be used for organic-based phases.
In our initial plan, since fermentor broth was the continuous phase, a stainless steel shaft
was used. The initial agitation speed is usually defined to void flooding, which is a
phenomenon where the upward or downward flow of the disperse phase ceases to form a
second interface in the column. This value was set to 200 rpm for the D-1,2,4-butanetriol
extraction. By knowing that 2-butanol has a low distribution coefficient at 0.24 for D-
1,2,4-butanetriol extraction, it was therefore reasonable to adjust the starting total
throughput to the low end, which was 15 mL/min. To select a reasonable solvent to feed
ratio is of almost importance and this requires knowledge of the solvents distribution
coefficients. It is not difficult to understand that a high recovery can be obtained with a
high solvent to feed ratio. However, using a large amount of solvent will certainly
increase the cost of the process. These two factors compete against each other and a
formula for the extraction factor (B) exists to make a reasonable judgment on the initial

amount of solvent: E = m X (vol. of solvent/vol. of feed) _>_ 1.'1 For instance, in the case

191

of 2-butanol extraction, the minimum solvent to feed ratio would be 4:1. After operating
the Karr column at these initial conditions, evaluations were made for the extraction
performance. Related parameters were changed accordingly to maximize the efficiency of

the Karr column extraction.

Preliminary extraction attempts using Karr column

A typical column extraction started by filling up the column with the continuous
phase, either with clarified fermentation broth or fresh solvent. Once the column was full
of liquid, the agitator was turned on at a slow rate (50 rpm). The pumps for the feed and
solvent were then turned on with the flow rate being set as required to establish an
interface. A stable interface was achieved by controlling the bottom raffmate take-off
rate. The agitation was then brought up slowly to the initial set point. During the first run
after a shutdown period, the column was allowed to operate for five column turnovers
before sampling. A turnover is defined as the column volume divided by the combined
flow rate. Solvent and raffinate (5 mL) were obtained for each set of parameters and
analyzed by ‘H, 13 C NMR and GC. All the NMR spectra were obtained on a 500 MHz
spectrometer for better resolution. The results were used to calculate the recovery and
purity. For each additional run on the same day, samples were collected after three
column turnovers.

In the first five entries, the solvent to feed ratio was set to 3:1 to allow precise
broth flow rate and maintained a low throughput at 15 mL/min (feed flow rate below 4 mL
was not accurately controlled by the pump). Among these experiments, attempts to

extract cell-free broth as a continouous phase failed shortly after five minutes due to

192

severe protein precipitation that resulted in column blockage (Table 34, entry 1). Further
attempts using different clarified broths were not successful due to entrainment (a portion
of feed was carried by the solvent and collected at the wrong end), leading to the change
of the continuous phase to solvent (Table 34, entry 2 and 3). In entry 4, it was discovered
that using solvent as the continuous phase provided a more defined interface and resulted
in stable extraction. Unfortunately, acidified broth yielded D-1,2,4-butanetriol with
serious contamination, as indicated earlier during the continuous extraction studies. Karr
column extraction of neutralized broth under the same conditions led to the isolation of
1,2,4-butanetriol with 86% purity (Table 34, entry 5). Unfortunately, the recovery was
only 15%. Increasing the agitation to 260 spm was unsuccessful and resulted in
entrainment (Table 34, entry 6). Doubling the total throughput to 30 mL/min did not
make an impact on the recovery (Table 34, entry 7), however, reducing the feed flow rate
by half increased the recovery to 22% (Table 34, entry 8). At this point, it was interesting
to investigate the recovery with even smaller feed flow rates. To this end, a Q pump
(Fluid Metering, Inc.) was setup to provide a finely controlled feed delivery to the column.
By lowering the total throughput back to 15 mL/min with a broth flow rate at 2.5 mL/min,
a 43% recovery was successfully achieved using the Karr column (Table 34, entry 9).

More importantly, entries 6 — 8 yielded D-l ,2,4-butanetriol with purity around 90%.

193

Table 34. Preliminary Karr column extraction studies.

 

 

 

entry brotha ”3::ng S/Fb 83mg)" totaflllow Edgtbmubndlilhnt wig/X?” p52?
1 cell-freed broth 3:1 200 15 4 11 column blocked
2 acid" broth 3:1 200 15 4 11 entrainment

3 acid/basef broth 3: 1 200 1 5 4 1 1 entrainment

4 acid° solvent 3:1 200 15 4 11 poor BT quality
5 acid/basef solvent 3:1 200 15 4 11 15 . 86
6 acid/basef solvent 3:1 260 1 5 4 1 1 entrainment

7 acid/basef solvent 3:1 200 30 8 22 14 87
8 acid/casef solvent 7:1 200 32 4 28 22 92
9 acid/basef solvent 4:1 200 12.5 2.5 10 43 92

 

aFor entries 1 — 3, the titer of D-1,2,4-butanetriol is 11.0 glL; for 4 — 8, 11.7 g/L; for 8, 10.6 glL.
b Solvent to feed ratio (S/F)

cStroke per minute (spm), stainless steel plate stack was used in all experiments.
dFermentation broth was cell-free, without underwent protein precipitation steps.

eBroth residual protein was precipitated at pH 2.5.

fAcidified broth was adjusted back to pH 7 after protein precipitation.

Further optimization of Karr column extraction

A new set of experiments was designed to optimize the Karr column extraction in
an attempt to increase the recovery. Given the earlier experience, several conditions were
adjusted. First, protein-free clarified fermentation broth was obtained from membrane
ultrafiltration as discussed earlier. This method avoided the introduction of salt streams
and reduced the cost by eliminating unnecessary use of acid and base. Second, coalescing
of the dispersed phase (feed) was observed in the previous experiments. Interfacial area
was therefore decreased as expected. To address this issue, the stainless steel plate stack
was replaced by the one made of Teflon. Third, broth delivery was tightly controlled by
the pump that was made to handle low flow rate (< 10 mL). More accurate and

reproducible extraction data were thus expected.

194

Attempts to extract D-l,2,4-butanetriol at a solvent to feed ratio at 4:1, where the
flow rate for broth and solvent were at 2 mL/min and 8 mL/min, respectively, recovered
the target product in 70% (Table 35, entry 1). Given this success, the total throughput
were increased and two extractions were run at 22 and 30 mL/min without significant
decrease of recovery (Table 35, entry 2 and 3). Increasing the throughput beyond this
point was not encouraging. A 54% recovery of D-1,2,4-butanetriol was obtained with a
total throughput at 40 mL/min (Table 35, entry 4), and 50 mL/min throughput ended up in
entrainment (Table 35, entry 5). Further attempts to increase the recovery by increasing
solvent to feed ratio to 8:1 and 12:1 yielded little success. D-1,2,4-Butanetriol was
recovered in 72% and 78%, respectively (Table 35, entry 6 and 7). The exhaustive use of
solvent in these cases would increase the cost of the process and therefore was not
preferred. From these experiments, it is clear that obvious advantages exists for filtering
off protein by ultrafiltration. Using the Teflon plate stack for agitation certainly created a

better dispersion to facilitate improved extraction performance.

Table 35. Further optimization for Karr column extraction.

 

 

 

2:2, <> e22,
1 ultrafil. solvent 4:1 200 10 2 8 70 2.5
2 ultrafil. solvent 4:1 200 22 4.5 18 74 2.5
3 ultrafil. solvent 4: 1 200 30 6 24 64 1 .5
4 ultrafil. solvent 4: 1 200 40 8 32 54 1 .3
5 ultrafil. solvent 4:1 200 50 10 40 entrainment
6 ultrafil. solvent 8: 1 260 56 6 50 72 1 .2
7 ultrafil. solvent 12:1 200 78 6 72 78 <1

 

aThe titer of D-1,2,4-butanetriol is 11.9 g/L. Residual protein was membrane filtered.
bSolvent to feed ratio (S/F)

cStroke per minute (spm), Teflon plate stack was used in all experiments.
dTheoretical stage (TS), estimated by using the Klemser type plot.

195

The operating efficiency of Karr column extractions can also be evaluated by using

the Kremser equation.7’”

The Kremser equation can be used to calculate the number of
theoretical stages (N s) and the equation is defined as follows (Figure 96), where X}: is the
mass concentration of solute in the feed, XN is the mass concentration of solute in the
raffinate, Y5 is the mass concentration of solute in the solvent, m is the distribution
coefficient, S/F is the mass ratio of solvent rate to feed rate. Finally, E is the extraction
factor that was discussed earlier. As an alternative, the number of theoretical stages can
be easily determined using the Kremser type plot (Figure 97).11 Theoretical stages for the
Karr column extraction experiments were determined and shown in Table 35. According
to all the evidence obtained, entry 2 represents the most efficient Karr colmnn extraction

process by having a 22 mL/min throughput, a theoretical stage of 2.5 and a recovery at

74%. These conditions will be used for future process scale-up.

Final step toward purified D-1,2,4-butanetrial

The Karr column was successfully employed to yield D-1,2,4-butanetriol in 90%
purity using 2-butanol as solvent. Under optimized conditions, recovery of this product
achieved up to 74% at a solvent to feed ratio of 4:1. The GC chromatographic traces for
the extraction process are shown in Figure 98 using dodecane as an internal standard.
Negative adsorption using a small plug of Dowex 1X8 (OH') resin was employed in an

attempt to obtain D-l,2,4-butanetriol with higher purity.

196

Figure 96. Kremser equation.

 

 

 

 

 

 

 

 

 

 

Y
xF-J
m 1 1
In —— 1-— +—
me E E
in
NS:
lnE
1° 1 = 0.3
0.4
__ 0.1
U
Q:
I?
re 0.04
5
X
2
5
P
.§
U
2 0.01
5
LI.-
0.004
0.001
0.0005
1 2 3 .1 '3 6 731015 2G
Number of ideal Stages

 

 

Figure 97. Kremser type plot.ll

197

 

 

(a)

D-1,2,4-butanetriol, 6.5 min _

dodecane, 4.5 min

/

.11 IL A

 

 

(b)

 

 

 

(C)

 

 

 

 

 

Figure 98. CC traces for Karr column extraction. (a) clarified broth; (b) 2-butanol
extract; (c) raffinate.

2-Butanol extract eluted from the Karr column containing D-1,2,4-butanetriol (500
mL) was evaporated to dryness under reduced pressure. The residue was redissolved into
10 mL H20 and filtered through a plug of anion-exchange Dowex 1X8 (OH') resin (30
mL). The resin was washed with two column volumes of water and all the flowthrough
was allowed to filter through another cation-exchange Dowex 50 (H) column (30 mL).
The second column contained a cation-exchange resin that was expected to remove
residual metal ion contamination in the D-1,2,4-butanetriol. The filtrate was evaporated to
dryness under reduced pressure to yield >99% pure salt-free D-l,2,4-butanetriol.

Combining the Karr column extraction and the resin filtration methodologies, 7 g D-1,2,4-

198

butanetriol was successfully purified from a 1 L fermentation broth of E. coli

WN13/pWN7.126B with >99% purity (Figure 99).

 

 

 

 

 

 

 

l t
(a)
l j
1' 110111 ._ -_ 2,212-1 L ,2 l_
6 5 4 3 2 1 0
(b) ,
"1‘36"““120'"in“ no ' 1‘60"“ so” a” "a” W 200
(C) D-1,2,4-butanetriol, 6.5 min
/
dodecane, 4.5 min
4/ 11 _

 

 

 

 

 

Figure 99. Large scale purification of E. coli WN13/pWN7.126B fermentation broth.
(a) 1H NMR spectrum, in ppm; (b) '3 C NMR spectrum, in ppm; (0) GC chromatogram.

Alternative methadfar D-1,2,4-butanetriol purification

Given success in using Dowex 1X8 (OH‘) anion-exchange resin as a second
purification step to obtain 99% pure D-l,2,4-butanetriol, it became interesting to
investigate whether Karr column extraction was indeed necessary for our purpose. To this
end, D-1,2,4-butanetriol was purified directly from membrane clarified fermentation broth
using Dowex 1><8 (OH') resin followed by a Dowex 50 (H') column. Although the
amount of cation-exchange resin was proven to be similar to the earlier process, it turned

out that the amount of the anion-exchange resin played an important role in the final

199

quality of product in this extraction-free purification. Therefore experiments were setup to
optimize this parameter and are shown in Table 36. D-1,2,4-Butanetriol with >99% purity
was obtained from 1 L fermentation broth using 2 L of Dowex 1X8 (OH') resin. It was an
eight-fold increase in the amount of resin involved comparing to the earlier process. In
addition, since more resin was involved in the purification, more water was used to wash
off D-1,2,4-butanetriol from the resin, leading to a requirement of 4 L of water to obtain a
recovery of 96%. Evaporation of this volume of water was deemed problematic (Table
37). Finally, regenerating the Dowex 1X8 (OH') resin after each purification was also
found to be problematic. Ten column volumes of 5 N NaOH was used to wash off the
bound material, including organic acids, salts and most importantly, some unidentified
colored impurities in the fermentor broth. Exhaustive resin regeneration is known to
shorten the resin lifetime, ultimately leading to an increase of product cost. MBI
International purified 500 g D-1,2,4-butanetriol with >99% purity using this extraction-
free method and a cost study was also prepared.'2 It costs $164 to purify 1 kg D-l,2,4-
butanetriol from the fermentation broth, while $48 was the cost of sodium hydroxide
involved in the resin regeneration. Apparently Karr column extraction using 2-butanol did
not only selectively extract D-1,2,4-butanetriol over other impurities, most of the salt and
coloring material were also separated from the product and discarded in the raffinate.
Although extra equipment cost associated with the construction of a production scale Karr
column will be involved when the process is commercialized, the solvent involved can be
easily recycled by distillation. As a long term consideration, the process involving the
Karr column extraction will lower the production cost by avoiding exhaustive use of

expensive Dowex 1 resin and associated NaOH.

200

Table 36. Dowex 1X8 (0H) for extraction-free purification.

 

 

resin (L resin per L broth) purity (%)
0.5 95
1 96
1.5 97
2 99

 

Table 37. D-l,2,4-Butanetriol recovery from Dowex 1X8 (OH').

 

 

Water (CV) recovery (%)
0.5 2
1 76
1 .5 93
2 96

 

Abbreviation: column volume (CV)

Future plan

Compared to the conventional purification method for polyols using distillation,
Karr column extraction was proven to be useful in D-1,2,4-butanetriol purification to
obtain product in a 74% recovery and 99% purity . Using liquid-liquid extraction would
significantly reduce the cost associated in the product separation. In addition, residual salt
and impurities would separate from the product during extraction. However, the
distribution coefficient of D-1,2,4-butanetriol into extraction solvent appears to be too
small to make the process efficient enough. A similar extraction approach using the Karr
column might serve as a solution to tackle this issue. D-1,2,4-Butanetriol is not easily
extracted into organic solvent due to the extensive hydrogen bonds formed between the
molecule and water. A possible avenue toward easier recovery of 1,2,4-butanetriol would

therefore be to react them with a suitable aldehyde to form the corresponding acetal

201

(Figure 100). Simultaneously, this aldehyde will also serve as solvent for the liquid-liquid
extraction process. By lowering the polarity of the compound, this D-1,2,4-butanetriol
derivative will therefore possess a larger distribution coefficient and readily extract into
the aldehyde solvent. This acetal cyclization only selectively occurs with compounds
having hydroxyl groups attached, therefore additional purification will be expected. In the
end, the aldehyde will be recycled by distillation. This reactive extraction methodology
had been successfully applied toward the recovery of 1,3-propanediol from a dilute
stream.l3 With the similarity of D—l,2,4-butanetriol and 1,3-propanediol, it is believed that
reactive extraction will apply to D-1,2,4-butanetriol. Solvent screening using different
aldehydes involving acetaldehyde, isobutyraldehyde and benzaldehyde is currently

underway.

R
OH + 0 11+ 7‘0
R = CH3-
Ph-
(CH3)2CH-

Figure 100. Cyclic acetal formation of 1,2,4-butanetriol.

202

REFERENCE

1 Brockmann, R.; Jeromin, L.; Johannisbauer, W.; Meyer, H.; Michel, O.; Plachenka, J.
US 4,655,879, 1987.

2 (a) Ames, T. T. US 6,361,983, 2002. (b) Haas, T.; Wiegand, N.; Amtz, D. US
5,334,778, 1994.

3 Molefe, M. N. PhD Thesis, Michigan State University, 2005.
4 See Chapter 2.

5 (8) Robbins, L. A. Chem. Eng. Prag. 1980, 76, 58. (b) http://wwwcheresourcescom
/extraction.html

6 Belter, P. A.; Cussler, E. L.; Hu, W.-S. Bioseparations: Downstream Processing for
Biotechnology; Wiley-VCH: Weinheim, 1988.

7 Gu, T. In Handbook of Biaseparation; Ahuja, A. Ed; Academic Press: NY, 2000; p329.
8 http://www.liquid-extraction.com/default.htm

9 Sullivan, D. A.; Shell Chemical Company In Kirk-Othmer Encyclopedia of Chemical
Technology; Wiley: 1997, DOI: 10.1002/0471238961.1915122219211212.a01.

'0 Dhale, A. D.; Myrant, L. K.; Chopade, S. R; Jackson, J. E.; Miller, D. J. Chem. Eng.
Sci. 2004, 59, 2881.

H Glatz, D. J. personal communications.
12 Saffron, C.; Dodds, D.; Lau, M. K.; Frost, J. W. unpublished results.
‘3 Broekhuis, R. R.; Lynn, 8.; King, C. J. Ind. Eng. Chem. Res. 1994, 33, 3230. (b)

Malinowski, J. J. Biotechnol. Prag. 2000, 16, 76. (c) Hao, J .; Liu, H.; Liu, D. Ind. Eng.
Chem. Res. 2005, 44, 4380.

203

CHAPTER FIVE

Experimental

General chemistry

All reactions sensitive to air and moisture were carried out in oven and/or flame
dried glassware under positive argon pressure. Air or moisture sensitive reagents and
solvents were transferred to reaction flasks fitted with rubber septa via syringes or
cannula. Solvents were removed using either a Biichi rotary evaporator at water aspirator

pressure or under high vacuum.

Reagents and solvents

CH2C12 and benzene were distilled from calcium hydride under argon. CH3OH

was distilled from sodium metal under argon and stored over Linde 4 A molecular sieves
under argon. THF and diethyl ether were distilled under nitrogen from sodium
benzophenone ketyl. DMF, DMSO, hexanes and acetone were dried over activated Linde
4 A molecular sieves under nitrogen. Water was glass distilled and deionized. All
reagents and solvents were used as available from commercial sources or purified
according to published procedures. Organic solutions of products were dried over

anhydrous MgSO4.

204

Chromatography

Gas chromatography was performed on an Agilent 6890N equipped with an HP-5
capillary column (30 m X 0.25 mm X 0.25 micron). Temperature programming began
with an initial temperature of 120 °C for 3 min. The temperature was increased to 210 °C
at a rate of 15 °C/min, and held at the final temperature for l min. The split injector was
maintained at a temperature of 300 °C and the F ID detector was kept at 350 °C. Samples
analyzed by gas chromatography were derivatized using bis(trirnethylsilyl)trifluoro-
acetamide and quantified relative to an internal standard of dodecane against a calibration
curve.

HPLC analysis was performed on an Agilent 1100 HPLC installed with
ChemStation acquisition software (Rev. A.08.03). Columns used in the enantiomeric
purity analysis of 1,2,4-butanetriol include Chiralpak AD column (Daicel Chemical, 4.6
mm X 250 mm) and Chirapak ADH column (Daicel Chemical, 4.6 mm X 250 mm).
Protein purification utilized the same HPLC system equipped with a Pharrnacia Resource
Q column (6.4 mm X 30 mm, 1 mL). Solvents were routinely filtered through 0.45-um
membranes (Gelman Science) prior to use.

Dowex 50WX8-200 (11+) and Dowex l><8-400 (CT) were purchased from Sigma-
Aldrich. Previously used Dowex 50 (H+) was cleaned by treatment with bromine. An
aqueous suspension of resin was adjusted to pH 14 by addition of solid KOH. Bromine
was added to the solution until the suspension turned a golden yellow color. Additional
bromine was added (1-2 mL) to obtain a saturated solution. The mixture stood at room
temperature overnight, and the Dowex 50 resin was collected by filtration and washed

exhaustively with water followed by 6 N HCl. Dowex 50 (H+) was stored at 4 °C. AG-

205

1X8 (acetate form and chloride form) and hydroxyapatite Bio-Gel HTP gel were
purchased from Bio-Rad. Phenylsepharose was purchased from Pharmacia.
Diethylaminoethyl cellulose (DEAE) was purchased from Whatrnan. Ni-NTA resin was

purchased from Qiagen.

Spectroscopic measurements

lH NMR and 13C NMR spectra were recorded on either a Varian VX-300 or a
Varian VXR-500 FT-NMR spectrometer. Chemical shifts for 1H NMR and '3 C NMR
spectra were reported in parts per million (ppm) relative to sodium 3-
(trimethylsilyl)propionate-2, 2, 3, 3-d4 (TSP, 6 = 0.0 ppm) with D20 as the solvent. UV and
visible measurements were recorded on a Perkin-Ehner Lambda 3b UV—vis
spectrophotometer or on a Hewlett Packard 8452A Diode Array Spectrophotometer
equipped with HP 89532A UV-Visible Operating Software. Measurements of multiple
samples in microplates were carried out using a Benchmark rrricroplate reader (Bio-Rad
Laboratories) equipped with Microplate Manager 111 Macintosh Data Analysis and

Kinetics Software.

Microbial strains and plasmids

Plasmid constructions were carried out in E. coli DH50t and XLl-Blue.

Table 38. Microbial strains and plasmids.

 

strain relavant characteristics source

 

DHSOL F’ d801acZAM15 A(lacZYA-argF)U169 deoR Invitrogen

recAI endAI hstI 7(rk', mf) phaA I supE44
thi-I gyrA96 relAI
BL21(DE3) E. coli B F dcm ompT hst(r3- m3-) gal A (DE3) Novagen

 

206

Table 38. (continue)

 

strain

relavant characteristics

 

source

XLl-Blue recAI endAI gyrA96 thi-I hstI 7 supE44 relAl Stratagene
lac [F' proAB lachZAM15 TnI 0 (Tet)].

W31 10 wild-type K12 CGSC

Acidithiabacillus wild-type ATCC

ferroxidans

Burkholderia wild-type ARS

fungorum LB400

Caulobacter wild-type ATCC

crescentus CB 1 5

Klebsiella wild-type ATCC

oxytoca

Klebsiella wild-type ATCC

pneumoniae

Lactobacillus wild-type CNRZ

brevis LB18

Lactobacillus wild-type CNRZ

brevis LBl9

Mycobacterium wild-type ATCC

smegmatis

Novosphingobium wild-type ATCC

aromaticivorans

Pseudomonas wild-type ATCC

fragi

Pseudomonas wild-type ATCC

F luorescens PfS

Pseudomonas putida wild-type ATCC

Rhodopseudomonas wild-type ATCC

palustris

WN7 W31lOserijhH::FRTyagE::FRT ref 1

WN13 WN7xylAB::xdh-FRT-CmR-FRT ref 1

KIT1 W31 leylAB : :xdh-ath-Pyac-FRT-CmR-F RT Chapter 3

KIT2 WN7xylAB: :xdh-FRT Chapter 3

KIT3 WN7xylAB::xdh-ath-Pm-FRT-CmR—FRT Chapter 3

KIT4 WN7xylAB : :xdh-ath-Pmc-F RT Chapter 3

KIT5 W3110P75-FRT-CmR-FRT Chapter 3

KIT6 KIT2Pr5-FRT-CmR-FRT Chapter 3

KIT7 KIT2PT5-FRT Chapter 3

KIT8 W3110athzzFRT-CmR-FRT Chapter 3

KIT9 KlT2ath::FRT-CmR-FRT Chapter 3

KIT10 KIT2ath::FRT Chapter 3

KIT16 KIT4yiaEzzFRT-CmR-FRT Chapter 3

KIT17 KIT16ycd W: :FRT-KmR-FRT Chapter 3

KIT18 KIT4yiaE ::FRTych::F RT Clgiter 3

 

207

Table 38. (continue)

 

 

plasmid relavant characteristics source
pVH17 ApR, deoC in pBR322 ATCC, ref2
pCS 120 ApR, 1),... dhaBCE from C. freundii ref 3
pFLl ApR, dhaBCE from C. pasteurianum ref 4
pXY39 ApR, lale, Pm pduCDE from S. typhimurium ref 5
pUC18 ApR, Ptac ref 6
pLOIl 35 ApR, Z. mobilis adhA in pUC18 ref 7
pL01276 ApR, z. mobilis pdc in pUC18 ref 29
pLOIZ95 ApR, Z. mobilis ath in pUC18 ref 8
pKD3 ApR, F RT-flanked CmR ref 35
pKD4 ApR, FRT-flanked KmR ref 35
pKD46 ApR, araC, PmBy, ,6, exa, ts-pAlOl replicon CGSC35
pCP20 ApR, CmR, Flp+, 1. c1857+ ref 34
pG-Tf2 CmR, Pztl groESL-tig Takara Bio
pRC1.55B serA in pSU18 lab

pT7-7 ApR, P77 lab9
pKK223-3 ApR, Pm ref 10
pJF118EH ApR, lac1Q in pKK223-3 ref 11
pET28c(+) KmR, lale, P77 Novagen
pQE3O ApR, lacO, PT; Qiagen
pWN5.220A ApR, dhaT in pT7-7 labl
pWN5.260A ApR, pddABC in pWN5.220A labl
pWN5.258A ApR, pddABC-ddrAB in pWN5.220A rahl
pWN5.23 8A ApR, laCIQ, deC in pJF118EH lab‘
pWN6.186A KmR, lacIQ, Pm mdlC labl
pWN6.222A KmR, Pm aatp, P77 aadh in pWN6.186 lab‘
pWN7.096B ApR, lale, Pm mdlC, 1).... adhA lab'
pWN7.126B serA in pWN5.238A rahl
pML2.118 P. fluorescens decarboxylase in pJF118EH Chapter 3
pML2.123 P. aeruginosa decarboxylase in pJF118EH Chapter 3
pML2.162 B. fungorum decarboxylase in pJF118EH Chapter 3
pML2.208 S. coelicolor decarboxylase in pJF118EH Chapter 3
pML2.214 N. aromaticivorans decarboxylase in pJF118EH Chapter 3
pML2.256 R. palustris decarboxylase in pJF118EH Chapter 3
pML2.286 M. smegmatis decarboxylase in pJF118EH Chapter 3
pML3.04O A. ferrooxidans decarboxylase in pJF118EH Chapter 3
pML3.054 P. putida mdlC mutant A4601 in pJF118EH Chapter 3
pML3.062 Z. mobilis pdc mutant I472A in pUC18 Chapter 3
pML3.084 Z mobilis pdc mutant 14728 in pUC18 Chapter 3
pML3.086 Z mobilis pdc mutant I472G in pUC18 Chapter 3
pML3.lO4 Z. mobilis pdc mutant I472T in pUC18 Chapter 3
pML3.17l Bf gene with EcoRI site deletion in pJF118EH Chapter 3
pML3.200 Bf gene with BamHI site deletion in pML3.200 Chapter 3
pML3.224 P. fluorescens decarboxylase in pJF 1 18HE Clmter 3

 

208

Table 38. (continue)

 

 

plasmid relavant characteristics source
pML3.225 P. aeruginosa decarboxylase in pJ F1 18HE Chapter 3
pML3.226 B. fungorum decarboxylase mutant in pJF118HE Chapter 3
pML5.176 ApR, gldBC in pJF118EH Chapter 3
pML5. 179 ApR, dhaT in pJF118EH Chapter 3
pML5.200 ApR, gldA in pML5.176 Chapter 3
pML5.226 ApR, dhaT in pML5.200 Chapter 3
pML6.090 ApR, dhaT in pKK223-3 Chapter 3
pML6.128 ApR, Pm-dhaT in pWN5.238 Chapter 3
pML6.133 serA in pWN7.096B Chapter 3
pML6.135 serA in pML6.128 Chapter 3
pML6.166 ApR, ath in pKK223-3 Chapter 3
pML6.168 ApR, adhE in pKK223-3 Chapter 3
pML6.185 ApR, Pm-ath in pWN5.238A Chapter 3
pML6.195 serA in pML6.185 Chapter 3
pML6.255 ApR, CmR, PmC-ath in pKD3 Chapter 3
pML6.259 ApR, yth in pJF118EH Chapter 3
pML6.261 ApR, yiaY in pJFl 18EH Chapter 3
pML6.263 ApR, who in pJF118EH Chapter 3
pML6.272 ApR, CmR, xdh in pML6.255 Chapter 3
pML7.042 ApR, CmR, P15 in pKD3 Chapter 3
pML7.128 KmR, deC in pET28c(+) Chapter 3
pML7.135 serA in pML7.128 Chapter 3
pML7.166 ApR, graESL-fig in pJF118EH Chapter 3
pML7.175 ApR, graESL in pWN5.23 8A Chapter 3
pML7.180 serA in pML7.175 Chapter 3
pML7.202 graESL in pML7.135 Chapter 3

 

Storage of microbial strains and plasmids

All bacterial strains were stored at -78 °C in glycerol. Plasmids were transformed
into DHSCL for long-term storage. Glycerol samples were prepared by adding 0.75 mL of
an overnight culture to a sterile vial containing 0.25 mL of 80% (v/v) glycerol. The

solution was mixed, left at room temperature for 2 h, and then stored at ~78 °C.

209

Culture medium

Bacto tryptone, Bacto yeast extract, nutrient broth, casamino acids, agar, and
MacConkey agar base were purchased from Difco. Casein hydrolysate was obtained from
Sigma. Nutrient agar was purchased from Oxoid.

All solutions were prepared in distilled, deionized water. LB medium12 (1 L)
contained Bacto tryptone (10 g), Bacto yeast extract (5 g), and NaCl (10 g). LB-glucose
medium contained glucose (10 g), MgSO4 (0.12 g), and thiamine hydrochloride (0.001 g)
in 1 L of LB medium. LB-freeze buffer contained K2HPO4 (6.3 g), KHzPOt; (1.8 g),
MgSOa (1.0 g), (NH4)2804 (0.9 g), sodium citrate dihydrate (0.5 g) and glycerol (44 mL)
in 1 L of LB medium. M9 salts12 (l L) contained NazHPOa (6 g), KH2PO4 (3 g), NH4C1
(l g), and NaCl (0.5 g). M9 minimal medium12 contained D-glucose (10 g), MgSO.; (0.12
g), and thiamine hydrochloride (0.001 g) in 1 L of M9 salts. Other M9 media contained
carbon sources (10 g) including D-lactose, L-arabinose, glycerol, D-xylonate, or L-
arbinonate in place of D-glucose in M9 minimal medium. For example, M9 glycerol
medium contained glycerol (10 g), MgSO4 (0.12 g), and thiamine hydrochloride (0.001 g)
in 1 L of M9 salts. M9 L—arabinose medium also contained 0.4% (w/v) casamino acids,
which were added into the M9 salts solution before sterilization in an autoclave. M9
medium (1 L) was supplemented where appropriate with L-phenylalanine (0.040 g), L-
tyrosine (0.040 g), L—tryptophan (0.040 g), p-hydroxybenzoic acid (0.010 g), potassium p-
aminobenzoate (0.010 g), and 2,3-dihydroxybenzoic acid (0.010 g). L-Serine was added
to a final concentration of 40 mg/L where indicated. Antibiotics were added where
appropriate to the following final concentrations: ampicillin (Ap), 50 pg/mL;

chloramphenicol (Cm), 20 pg/mL; kanamycin (Kan), 50 pg/mL; tetracycline (Tc),

210

12.511g/mL. Stock solutions of antibiotics were prepared in water with the exceptions of
chloramphenicol which was prepared in 95% ethanol and tetracycline which was prepared
in 50% aqueous ethanol. Aqueous stock solutions of isopropyl-B-D-thiogalactopyranoside
(IPTG) were prepared at various concentrations. Solutions of LB medium, M9 inorganic
salts, MgSO4, D-glucose, D-xylose, glycerol and erythritol were autoclaved individually
and then mixed. Solutions of potassium D-xylonate, thiamine hydrochloride, antibiotics,
and IPTG were sterilized through 0.22-11m membranes. Other solid media were prepared
by addition of Difco agar to a final concentration of 1.5% (w/v) to the liquid medium.

Pseudomonas fragi and Pseudomonas putida were grown either in nutrient broth
or on solid nutient agar plates. Nutrient broth and nutrient agar were prepared according
to procedures recommended by the manufactures. Klebsiella pneumoniae was cultured in
liquid or on solid ATCC medium 561. ATCC medium 561 (800 mL) contained yeast
extract (1 g), casein hydrolysate (1.4 g), K2HP04 (0.6 g), Mg804 (0.5 g), K2804 (1 g), and
glycerol (20 mL). Solid ATCC medium 561 was prepared by addition of Difco agar to a
final concentration of 1.5% (w/v) to the liquid medium.

The standard fermentation medium (1 L) utilized in Chapter 3 contained K2HP04
(7.5 g), ammonium iron (III) citrate (0.3 g), citric acid monohydrate (2.1 g), and
concentrated H2804 (1.2 mL). Fermentation medium was adjusted to pH 7.0 by addition
of concentrated NH40H before autoclaving. The following supplements were added
immediately prior to initiation of the fermentation: D-glucose, Mg804 (0.24 g), potassium
and trace minerals including (NH4)6(M07024)-4H20 (0.0037 g), ZnSO4-7H2O (0.0029 g),
H3BO3 (0.0247 g), Cu804-5H20 (0.0025 g), and MnCl2-4H2O (0.0158 g). IPTG stock

solution was added as necessary to the indicated final concentration. Glucose feed

211

solution and Mg804 (1 M) solution were autoclaved separately. Glucose feed solution
(650 g/L) was prepared by combining 300 g of glucose and 280 mL of H20. Solutions of
trace minerals and IPTG were sterilized through 0.22-pm membranes. Antifoarn (Sigma

204) was added to the fermentation broth as needed.

General fed-batch fermentation conditions

Fermentations employed a 2.0 L working capacity B. Braun M2 culture vessel.
Utilities were supplied by a B. Braun Biostat MD controlled by a DCU-3. Data
acquisition utilized a Dell Optiplex Gs+ 5166M personal computer (PC) equipped with B.
Braun MFCS/W in software (v1.1) or a Dell Optiplex GX200 personal computer (PC)
equipped with B. Braun MFCSfW in software (v2.0). Temperature, pH, and carbon Source

feeding were controlled with PID control loops. pH was maintained at 7.0 by addition of

concentrated NH40H or 2 N H2804. Dissolved oxygen (D.O.) was measured using a
Mettler-Toledo 12 mm sterilizable 02 sensor fitted with an Ingold A-type O2 permeable
membrane. Inoculants were started by introduction of a single colony picked from an agar
plate into 5 mL of M9 medium. Cultures were grown at 37 °C with agitation at 250 rpm

until they were turbid and subsequently transferred to 100 mL of M9 medium. Cultures
were grown at 37 °C and 250 rpm for an additional 10 h. The inoculant (OD600 = 1.0-3.0)

was then transferred into the fermentation vessel and. the batch fermentation was initiated
(t = 0 h).
Three staged methods were used to maintain D.O. concentrations at desired air

saturation during the fermentations. With the airflow at an initial setting of 0.06 L/L/min,

212

the DO. concentration was maintained by increasing the impeller speed from its initial set
point of 50 rpm to its preset maximum rate. With the impeller speed constant, the mass
flow controller then maintained the DO. concentration by increasing the airflow rate from
0.06 L/L/min to a preset maximum of 1.0 L/L/min. At constant impeller speed and
constant airflow rate, the DO. concentration was finally maintained at the desired air
saturation for the remainder of the fermentation by oxygen sensor-controlled carbon
source feeding. At the beginning of this stage, the DO. concentration fell below the
desired air saturation due to residual initial carbon source in the medium. This lasted for
approximately 10 min to 30 min before carbon source feeding commenced. The carbon

source feed PID control parameters were set to 0.0 s (off) for the derivative control (TD)

and 999.9 5 (minimum control action) for the integral control (1}). Xp was set to 950% to

achieve a Kc of 0. 1.

Analysis of fermentation broth

Samples (5-10 mL) of fermentation broth were removed at the indicated timed

intervals. Cell densities were determined by dilution of fermentation broth with water

(1:100) followed by measurement of absorption at 600 nm (OD600). Dry cell weight of E.

coli cells (g/L) was calculated using a conversion coefficient of 0.43 g/L/OD600. The

remaining fermentation broth was centrifuged to obtain cell-free broth.
Glucose concentrations in cell-free broth were measured using the Glucose
Diagnostic Kit purchased from Sigma. For the biosynthesis of D-1,2,4-butanetriol,

organic acid byproducts such as D-xylonic acid, 3-deoxy-D-glycera-pentulosonic acid, 3-

213

ddeoxy-D-glycera-pentanoic acid, and 4,5-dihydroxy-threa-L-norvaline concentrations in
the cell-free broth were quantified by 1H NMR. A portion (0.5-2.0 mL) of the cell-free
broth was concentrated to dryness under reduced pressure, concentrated to dryness one
additional time from D20, and then redissolved in D20 containing a known concentration
of the sodium salt of 3-(trimethylsilyl)propionic-2, 2, 3, 3-d4 acid (TSP, Lancaster Synthesis
Inc.). Concentrations were determined by comparison of integrals corresponding to each
compound with the integral corresponding to TSP (8 = 0.00 ppm) and were converted by
response factors determined using authentic materials. Compounds were quantified using
the following resonances: D-xylonic acid (8 4.08, d, 1 H); 3-deoxy-D-glycera-pentulosonic
acid (8 4.58, m, 1 H); 3-deoxy-D-glycera-pentonic acid (8 1.94, ddd, 1 H); 4,5-dihydroxy-
threa-L-norvaline (8 4.01 , dd, 1 H). A standard concentration curve was determined for
metabolites using solutions of authentic samples. The concentration of D-l ,2,4-butanetriol
and 3,4-dihydroxy-D-butanoic acid in cell-free broth was quantified by GC analysis. A
portion of the fermentation broth (0.5-1.0 mL) was concentrated to dryness under reduced
pressure, and the residue was redissolved in pyridine (0.9 mL). To this pyridine solution,
dodecane (0.1 mL) and bis(trimethylsilyl)trifluoroacetamide (BSTFA, 2 mL, 7.53 mmol)
were sequentially added. Silyation of D-1,2,4-butanetriol and 3,4-dihydroxy-D-butanoic
acid were carried out at room temperature with stirring for 10 h. Samples were then

analyzed using gas chromatography.

Genetic manipulations

Recombinant DNA manipulations generally followed methods described by

1.13

Sarnbrook et a Restriction enzymes were purchased from Invitrogen or New England

214

Biolabs. T4 DNA ligase was obtained from Invitrogen. F ast-LinkTM DNA Ligation Kit
was obtained from Epicentre. Zymoclean Gel DNA Recovery Kit and DNA Clean &
Concentrator Kit was obtained from Zymo Research Company. Maxi and Midi Plasmid
Purification Kits were obtained from Qiagen. Calf intestinal alkaline phosphatase was
obtained from Boehringer Mannheim. Agarose (electrophoresis grade) was obtained from
Invitrogen. Phenol was prepared by addition of 0.1 % (w/v) 8-hydroxyquinoline to
distilled, liquefied phenol. Extraction with an equal volume of 1 M Tris-HCl (pH 8.0) two
times was followed by extraction with 0.1 M Tris-HCl (pH 8.0) until the pH of the
aqueous layer was greater than 7.6. Phenol was stored at 4 °C under an equal volume of
0.1 M Tris-HCl (pH 8.0). SEVAG was a mixture of chloroform and isoamyl alcohol

(24:1, v/v). TE buffer contained 10 mM Tris-HCl (pH 8.0) and 1 mM Na2EDTA (pH
8.0). TAE buffer contained 40 mM Tris-acetate (pH 8.0) and 2 mM NazEDTA. Endostop
solution (10x concentration) contained 50% glycerol (v/v), 0.1 M NazEDTA, pH 7 .5, 1%
sodium dodecyl sulfate (SDS) (w/v), 0.1% bromophenol blue (w/v), and 0.1% xylene
cyanole FF (w/v) and was stored at 4 °C. Prior to use, 0.12 mL of DNase-free RNase was

added to 1 mL of 10X Endostop solution. DNase-free RNase (10 mg mL'l) was prepared
by dissolving RNase in 10 mM Tris-H01 (pH 7.5) and 15 mM NaCl. DNase activity was
inactivated by heating the solution at 100 °C for 15 min. Aliquots were stored at -20 °C.
PCR amplifications were carried out as described by Sambrook et al.13 A standard
reaction (0.1 mL) contained 10 mM KCl, 20 mM Tris-HCl (pH 8.8), 10 mM (NH4)2804, 2
mM MgSO.;, 0.1% Triton X-100, dATP (0.2 mM), dCTP (0.2 mM), dGTP (0.2 mM),
dTTP (0.2 mM), template DNA, 0.5 ,uM of each primer, and 2 units of Taq polymerase.

Template concentration varied from 0.02 ,ug to 1.0 ,ug.

215

Large scale purification of plasmid DNA

Plasmid DNA was purified on a large scale using a modified alkaline lysis method
described by Sambrook et al.13 In a 2 L Erlenmeyer flask, 500 mL of LB containing the
appropriate antibiotics was inoculated from a single colony, and the culture was incubated
in a gyratory shaker at 37 °C for 14 h with agitation at 250 rpm. Cells were harvested by
centrifugation (4,000g, 5 min, 4 °C) and then resuspended in 10 mL of cold GETL
solution (50 mM glucose, 20 mM Tris-HCl (pH 8.0), 10 mM NazEDTA, pH 8.0) into
which lysozyme (5 mg/mL) had been added immediately before use. The suspension was
stored at room temperature for 5 min. Addition of 20 mL of 1% sodium dodecyl sulfate
(w/v) in 0.2 N NaOH was followed by gentle mixing and storage on ice for 15 min.
Fifteen milliliters of an ice-cold solution containing 3 M KOAc (prepared by combining
60 mL of 5 M potassium acetate, 11.5 mL of glacial acetic acid, and 28.5 mL of H20) was
added. Vigorous shaking resulted in formation of a white precipitate. After the
suspension was stored on ice for 10 min, the cellular debris was removed by centrifugation
(48,000g, 20 min, 4 °C). The supernatant was transferred to two clean centrifiige bottles
and isopropanol (0.6 volumes) was added to precipitate the DNA. After the samples were
left at room temperature for 15 min, the DNA was recovered by centrifugation (20,000g,
20 min, 4 °C). The DNA pellet was then rinsed with 70% ethanol and dried.

Further purification of the DNA sample involved precipitation with polyethylene
glycol (PEG). The isolated DNA was dissolved in TE (3 mL) and transferred to a Corex
tube. Cold 5 M LiCl (3 mL) was added and the solution was gently mixed. The sample

was then centrifuged (12,000g, 10 min, 4 °C) to remove high molecular weight RNA. The

216

clear supernatant was transferred to a clean Corex tube and isopropanol (6 mL) was added
followed by gentle mixing. The precipitated DNA was collected by centrifugation
(12,000g, 10 min, 4 °C). The DNA was then rinsed with 70% ethanol and dried. After re-
dissolving the DNA in 0.5 mL of TE containing 20 ,ug/mL of RNase, the solution was
transferred to a 1.5 mL microcentrifuge tube and stored at room temperature for 30 min.
DNA was precipitated from solution upon addition of 500 ,uL of 1.6 M NaCl containing
13% PEG-8000 (w/v) (Sigma). The solution was mixed and centrifuged (microcentrifuge,
10 min, 4 °C) to recover the precipitated DNA. The supernatant was removed, and the
DNA was then re-dissolved in 400 ,uL of TE. The sample was extracted sequentially with
phenol (400 ,uL), phenol and SEVAG (400 ,uL each), and finally SEVAG (400 ,uL).
Ammonium acetate (10 M, 100 ,uL) was added to the aqueous DNA solution. After
thorough mixing, 95% ethanol (1 mL) was added to precipitate the DNA. The sample was
left at room temperature for 5 min and then centrifuged (microcentrifuge, 5 min, 4 °C).
The DNA was rinsed with 70% ethanol, dried, and then redissolved in 200-500 ,uL of TE.
Alternatively, DNA was purified using a Qiagen Maxi Kit or Midi Kit as described

by the manufacturer. The purity of DNA isolated by these kits was adequate for DNA

sequencing.

Small scale purification of plasmid DNA

An overnight culture (5 mL) of the plasmid-containing strain was grown in LB
containing the appropriate antibiotics. Cells from 3 mL of the culture were collected in a
1.5 mL microcentrifuge tube by centrifugation. The resulting cell pellet was liquefied by

vortexing (30 sec) and then resuspended in 0.1 mL of cold GETL solution into which
217

lysozyme (5 mg/mL) had been added immediately before use. The solution was stored on
ice for 10 min. Addition of 0.2 mL of 1% sodium dodecyl sulfate (w/v) in 0.2 N NaOH
was followed by gentle mixing and storage on ice for 5-10 min. To the sample was added
0.15 mL of cold KOAC solution. The solution was shaken vigorously and stored on ice
for 5 min before centrifugation (15 min, 4 °C). The supernatant was transferred to another
microcentrifuge tube and extracted with equal volumes of phenol and SEVAG (0.2 mL).
The aqueous phase (approximately 0.5 mL) was transferred to a fresh microfuge tube, and
DNA was precipitated by the addition of 95% ethanol (1 mL). The sample was left at
room temperature for 5 min before centrifugation (15 min, room temperature) to collect
the DNA. The DNA pellet was rinsed with 70% ethanol, dried, and redissolved in 50 to
100 ,uL TE. DNA isolated using this method was used for restriction enzyme analysis,
although the concentration of DNA could not be accurately determined by spectroscopic

methods.

Restriction enzyme digestion of DNA

Restriction enzyme digests were performed in buffers provided by Invitrogen or
New England Biolabs. A typical restriction enzyme digest contained 0.8 pg of DNA in 8
11L of TE, 2 uL of restriction enzyme buffer (10x concentration), 1 11L of bovine serum
albumin (0.1 mg/mL), 1 uL of restriction enzyme and 8 11L TE. Reactions were incubated
at 37 °C for 1 h, terminated by addition of 2.2 11L of 10X Endostop solution and analyzed

by agarose gel electrophoresis. When DNA was required for cloning experiments, the

digest was terminated by addition of 1 1.1L of 0.5 M NazEDTA (pH 8.0) or by heating at

218

70 °C for 15 min followed by extraction of the DNA using Zymoclean gel DNA recovery

kit.

Determination of DNA concentration

The concentration of DNA in the sample was determined as follows. An aliquot
(10 pL) of DNA was diluted to 1 mL in TE and the absorbance at 260 nm was measured
relative to the absorbance of TE. The DNA concentration was calculated based on the fact

that the absorbance at 260 nm of 50 ,ug/mL of double stranded DNA is 1.0.

Agarose gel electrophoresis

Agarose gel typically contained 0.7% agarose (w/v) in TAE buffer. Ethidium
bromide (0.5 ,ug/ml) was added to the agarose to allow visualization of DNA fragments
under a UV lamp. Agarose gel was run in TAE buffer. The size of the DNA fragments
were determined using two sets of DNA molecular weight standards: 21 DNA digested
with Hindlll (23.1-kb, 9.4-kb, 6.6-kb, 4.4-kb, 2.3-kb, 2.0-kb and 0.6-kb) and 2. DNA
digested with EcoRI and Hindlll (21.2-kb, 5.1-kb, 5.0-kb, 4.3-kb, 3.5—kb, 2.0-kb, 1.9-kb,

1.6-kb, 1.4-kb, 0.9-kb, 0.8-kb and 0.6-kb).

Isolation of DNA from agarose

The band of agarose containing DNA of interest was excised from the gel while
visualized with long wavelength UV light. Two methods were used for isolating DNA

from agarose gels. The first method used Zymoclean gel DNA recovery kit to isolate

219

DNA from the agarose gel according to the procedure provided by Zymo Research.
Alternatively, the agarose gel containing DNA was chopped thoroughly with a razor and
then transferred to a 0.5 mL microfuge tube packed tightly with glass wool and having an
18 gauge hole at the bottom. The tube was centrifuged for 5 min using a Beckman
microfuge to extrude the DNA solution from the agarose into a second 1.5 mL microfuge
tube. The DNA was precipitated using 3 M NaOAc and 95% ethanol as described

previously and subsequently redissolved in TE.

Treatment of DNA with Klenow fragment

DNA fragments with recessed 3’ termini were modified to DNA fragments with
blunt ends by treatment with the Klenow fragment of E. coli DNA polymerase 1. After
restriction digestion (20 ,uL) of the DNA (0.8-2 ,ug) was complete, a solution (1 ,uL)
containing each of the four dNTPs was added to a final concentration of 1 mM for each
dNTP. Addition of 1-2 units of Klenow fragment was followed by incubation at room
temperature for 20-30 min. Since the Klenow fragment works well in the common
restriction enzyme buffers, there was generally no need to purify the DNA after restriction
digestion and prior to filling recessed 3’ termini. Klenow reactions were quenched by
extraction with equal volumes of phenol and SEVAG. DNA was recovered using the
Zymoclean gel DNA recovery kit or by precipitation as described previously and

subsequently dissolved in TE.

220

Treatment of vector DNA with calf intestinal alkaline phosphatase

Following restriction enzyme digestion, plasmid vectors were dephosphorylated to
prevent self-ligation. Digested vector DNA was dissolved in TE (88 ,uL). To this sample
was added 10 ,uL of dephosphorylation buffer (10x concentration) and 2 ,uL of calf
intestinal alkaline phosphatase (2 units). The reaction was incubated at 37 °C for 1-2 h.
The phosphatase was inactivated by the addition of 1 ,uL of 0.5 M Na2EDTA (pH 8.0)
followed by heat treatment (70 °C, 15 min). The sample was extracted with phenol and
SEVAG (100 ,uL each) to remove the protein, and the DNA was purified as previously

described and subsequently dissolved in TE.

Ligation of DNA
Molar ratios of insert to vector were typically maintained at 3 to l for DNA
ligations. A typical reaction contained 0.1 ug of vector DNA and 0.05 to 2.0 ug of insert
DNA in a total volume of 7 11L. To this was added 2 11L of T4 ligation buffer (5x
concentrations) and 1 11L of T4 DNA ligase (2 units). The reaction was incubated at 16 °C
for at least 4 h and then used to transform competent cells.
Alternatively, A F ast-LinkTM DNA Ligation Kit (Epicentre, Madison, WI) was
used for ligation of insert DNA with cohesive or blunt ends into predigested vectors with

compatible ends according to the protocol provided by Epicentre.

221

Preparation and transformation of competent E. coli cells

Competent cells were prepared according to a procedure modified from Sambrook
et a1.13 LB medium (5 mL) containing antibiotics where appropriate, was inoculated with
a single colony from 8 LB plate containing antibiotics where appropriate. The culture was
grown at 37 °C with shaking at 250 rpm for 10-12 h. An aliquot (1 mL) from the culture
(5 mL) was used to inoculate LB (100 mL) containing the appropriate antibiotics. The
culture was grown at 37 °C with shaking at 250 rpm in a NBS series 25 incubator shaker
until the optical density at 600 nm was between 0.4 and 0.6. The culture was transferred
to a centrifuge bottle that had been sterilized with a 25 % (v/v) bleach solution and rinsed
four times with sterile, deionized water. The cells were harvested by centrifugation (4000
g, 5 min, 4 °C) and the culture medium was decanted. All subsequent manipulations were
carried out on ice. The harvested cells were resuspended in ice-cold 0.9 % NaCl (100
mL), and the cells were collected by centrifugation (4,000 g, 5 min, 4 °C). The 0.9 %
NaCl solution was decanted, the cells were resuspended in ice-cold 100 mM CaCl2 (50
mL) and stored on ice for 30 min. After centrifugation (4,000 g, 5 min, 4 °C), the cells
were resuspended in 4 mL of ice-cold 100 mM CaCl2 containing 15% glycerol (v/v).
Aliquots (0.25 mL) of competent cells were added to 1.5 mL microfuge tubes,
immediately frozen in liquid nitrogen, and stored at -78 °C.

Frozen competent cells were thawed on ice for 5 min before transformation. A
small aliquot (1 to 10 pL) of plasmid DNA or a ligation reaction was added to the thawed
competent cells (0.1 mL). The solution was gently mixed by tapping and stored on ice for
30 min. The cells were then beat shocked at 42 °C for 30 seconds and returned to ice

briefly (1 min). LB (0.5 mL, no antibiotics) was added to the cells, and the sample was

222

incubated at 37 °C (no agitation) for l h. Cells were collected by centrifugation (30 s) in a
microcentrifuge. If the transformation was to be plated onto LB plates, 0.5 mL of the
culture supernatant was removed, and the cells were resuspended in the remaining 0.1 mL
of LB and subsequently spread onto plates containing the appropriate antibiotics. If the
transformation was to be plated onto minimal medium plates, the cells were washed twice
with a solution of M9 salts (0.5 mL). After resuspension in a fresh aliquot of M9 salts (0.1
mL), the cells were spread onto a plate. An aliquot of competent cells with no DNA
added was also carried through the transformation protocol as a control. These cells were
used to check the viability of the competent cells and to verify the absence of growth on
selective medium.

Transformations were also performed by electroporation using electrocompetent
cells. An aliquot (1 mL) from an overnight culture (5 mL) was used to inoculate 500 mL
of 2xYT containing the appropriate antibiotics. The cells were cultured at 37 °C with
shaking at 250 rpm. Once an absorbance of 0.6-0.8 at 600 nm was observed, the cells
were kept on ice for 10 min and harvested (3,000g, 5 min, 4 °C). The cells were gently
washed three times with sterile, cold water (450 mL once and 250 mL twice) and then
resuspended in 100 mL sterile, ice-cold aqueous 10% glycerol (v/v). After centrifugation
(3,000g, 5 min, 4 °C), the cells were resuspended in 1.5 mL sterile ice-cold aqueous 10%
glycerol (v/v). Aliquots (0.1 mL) of electrocompetent cells were dispensed into 1.5 mL
microfuge tubes, and immediately frozen in liquid nitrogen and stored at -—78 °C.

The electroporation was performed in Bio-Rad Gene Pulser cuvettes with an
electrode gap of 0.2 cm. The cuvettes were chilled on ice for 5 min prior to use.

Electrocompetent cells were thawed in ice for 5 min, and 40 uL of thawed cells was added

223

to the chilled cuvette. To this was added 1-10 11L of plasmid DNA (1 ug mL'l), and the
mixture was gently shaken. The Bio-Rad Gene Pulser was set at 2.5 kV, 25 pF and 200 Q.
The outside surface of the cuvette was wiped clean and it was placed in the sample
chamber. A single pulse was applied, the cuvette was removed, and 1 mL of freshly
prepared SOC was added into it. The contents of the cuvette were transferred to a 15 mL
sterile centrifugation tube. The cells were incubated at 37 °C for 1 h with shaking at 250
rpm. The transformed cells were plated in the same manner as in the transformation with

chemically competent cells.

Purification of E. coli genomic DNA

Genomic DNA was purified using a method modified from Silhavy et al.14 A
single colony of an E. coli strain was inoculated into 100 mL of TB medium (500 mL
Erlenmeyer flask). The cells were cultured in a gyratory shaker (37 °C, 250 rpm) for
12 h. Centrifugation (4,000g, 5 min, 4 °C) of the culture was followed by resuspension of
the cell pellet in 5 mL of buffer (50 mM Tris-HCl, 50 mM Na2EDTA, pH 8.0) and storage
at -20 °C for 20 min to freeze the suspension. To the frozen cells was added 0.5 mL of
0.25 M Tris-HCl (pH 8.0) that contained 5 mg of lysozyme. The suspension was thawed
at room temperature in a water bath with gentle mixing and then stored on ice for 45 min.
The sample was then transferred to a Corex tube. After addition of 1 mL of STEP
solution (25 mM Tris-HCl (pH 7.4), 200 mM Na2EDTA (pH 8.0), 0.5% SDS (w/v), and
proteinase K (1 mg mL"), prepared just before use), the mixture was incubated at 50 °C
for at least 1 h with gentle, periodic mixing. The solution was then divided into two Corex

tubes, and the contents of each tube were extracted with phenol (4 mL). The organic and

224

aqueous layers were separated by centrifugation (1,000g, 15 min, room temperature), and
the aqueous layer was transferred to a fresh Corex tube. All transfers of the aqueous layer
were carried out using wide bore pipette tips to minimize shearing of the genomic DNA.
The contents of each tube were extracted again with a mixture of phenol (3 mL) and
SEVAG (3 mL). Extractions with phenol/SEVAG were repeated (approximately 6 times)
until the aqueous layer was clear.

Genomic DNA was precipitated by addition of 0.1 volume of 3 M NaOAc (pH
5 .2) followed by gentle mixing and addition of 2 volumes of 95% ethanol. Threads of
DNA were spooled onto a sealed Pasteur pipette and transferred to a Corex tube that
contained 5 mL of 50 mM Tris-HCl (pH 7.5), 1 mM Na2EDTA (pH 8.0), and 1 mg of
RNase. The mixture was stored at 4 °C overnight to allow the DNA to dissolve
completely. The solution was then extracted with SEVAG (5 mL) and centrifuged (1,000
g, 15 min, room temperature). The aqueous layer was transferred to a fresh Corex tube
and the genomic DNA was precipitated as described above. The threads of DNA were
spooled onto a Pasteur pipette and redissolved in 2 mL of 50 mM Tris-HCI (pH 7 .5) and 1
mM Na2EDTA (pH 8.0). Genomic DNA was stored at 4 °C.

Alternatively, genomic DNA was purified using a method described by Pitcher et
al.'5 A single colony of an E. coli strain was inoculated into 20 mL of LB medium. The
cells were culture in a gyratory shaker (37 °C, 250 rpm) overnight, and were subsequently
harvested by centrifugation (1,000g, 15 min, room temperature). The cell pellet obtained
was resuspended into 100 ,uL of TE and incubated at 37°C for 30 min. Cells were lysed
with 0.5 mL 5 M guanidine thiocyanate, 100 mM EDTA and 0.5% (v/v) sarkosyl (GES

reagent), which was prepared as follows. Guanidine thiocyanate (60 g), 0.5 M Na2EDTA,

225

pH 8.0 (20 mL), and deionized water (20 mL) were heated at 65°C with mixing until
dissolved. After cooling, 5 mL of 10% v/v sarkosyl were added, the solution was made up
to 100 mL with deionized water, filtered through a 0.22-pm membrane and stored at room
temperature.

The cell suspension was vortexed briefly and checked for lysis (clear solution)
after 5-10 min. The lysate was cooled on ice, and a cold solution of ammonium acetate
(7.5 M, 0.25 mL) was added with mixing on ice for 10 rrrin. To this sample, 0.5 mL
SEVAG was added and mixed thoroughly. After centrifugation in a 1.5 mL Eppendorf
tube (25,000g, 10 min, room temperature), the supernatant was transferred to an
Eppendorf tube and 0.54 volumes of cold 2-propanol was added. The tubes were inverted
for 1 min to mix the solutions and the fibrous DNA precipitate was collected by
centrifugation (6,500g, 20 sec, room temperature). DNA pellet was washed five times
with 70% ethanol and dried in air at room temperature for 20 min. Finally, the genomic

DNA was redissolved in 100 pL TE.

Pl phage-mediated transduction

Transduction with P1 phage was carried out using a method modified from
Miller.l6 P1 phage lysate was prepared by propagation of phage in the donor strain using
the following procedure. Serial dilutions of P1 phage stock (0.1 mL, 10'l to 10's) in LB
were prepared in sterile test tubes (13 x 100 mm). An aliquot (0.1 mL, approximately 5 x
108 cells) of an overnight culture of the donor strain was added to each tube. Sterile,
molten soft agar (45 °C) was added to each tube. The contents of each tube were mixed

and poured immediately onto a pre-warmed (37 °C) L plate and swirled gently to achieve

226

uniform coverage of the plate. After the agar had solidified, the plates were incubated at
37 °C until confluent lysis had occurred (approximately 8 h). Because the multiplicity of
infection is critical to phage generation, confluent lysis occurred on only one or two of the
plates. To the plates displaying confluent lysis, L-broth (4 mL) was added, and the plates
were then stored overnight at 4 °C to allow the phage particles to diffuse into the broth.
The L-broth was collected from the plate and vortexed with several milliliters of CHCl3.
The solution was centrifuged (2,000g, 5 nrin, room temperature) to separate the layers.
Aqueous phage lysate was stored in 1.5 mL microfuge tubes over several drops of CHCl3
at 4 °C.

Infection of the recipient strain with phage lysate proceeded as follows. An
overnight culture (2 mL) of the recipient strain was centrifuged (microfuge, 30 sec, 4 °C)
and the growth medium was discarded. The cells were resuspended in 1 mL sterile
solution containing of 5 mM CaCl2 and 100 mM MgSO.;, and shaken (250 rpm) at 37 °C
for 15 min to promote aeration of the cells. In the meantime, 0.1 mL serial dilutions (100
to 103) of phage lysate in LB were prepared in sterile microfuge tubes. An aliquot (0.1
mL) of aerated recipient cells was added to each of the phage dilutions, the samples were
gently mixed. and then incubated at 37 °C for 20 min without shaking. Sodium citrate (1
M, 0.2 mL) was added to each sample, and the cells were harvested (microfuge, 30 3,
room temperature) and resuspended in 0.2 mL of LB containing 100 mM sodium citrate.
After incubation at 30 °C for 30 min, cells were again harvested (microfuge, 30 sec, room
temperature), resuspended in 0.1 mL of growth medium, and plated out onto appropriate

agar plates.

227

Enzyme assays

After collected and resuspended in the proper resuspension buffer, the cells were
disrupted by two passages through a French pressure cell (SLM Aminco) at 16,000 psi.
Cellular debris was removed from the lysate by centrifugation (48,000g, 20 min, 4 °C).
Protein was quantified using the Bradford dye-binding procedure.17 A standard curve was
prepared using bovine serum albumin. Protein assay solution was purchased from Bio-

Rad and used as described by the manufacture.

2-Deoxyribose-5-phosphate aldolase assay

2-Deoxyribose 5-phosphate aldolase activity was assayed as described by Wong,
C.-H. in literature.18 2-Deoxyribose-5-phosphate (1 mM), 0.28 mM NADH, and a mixture
of glycerophosphate dehydrogenase and triose phosphate isomerase was incubated in 50
mM Tris, 0.1 mM EDTA, pH 7.6 at 25 °C. The assay was initiated by the addition of 2-
deoxyribose 5-phosphate aldolase, and the decrease in the absorbance at 340 nm was

monitored. The extinction coefficient for NADH was taken as 6.22 x 103 M’lcm".

Glycerol and diol dehydratase assay

Glycerol and dial dehydratase was assayed according to the procedure described
by Abeles.19 A coupled enzyme reaction in which the aldehyde formed by the dehydratase
reaction was reduced to the corresponding alcohol. The assay mixture was composed of
an appropriate amount of dehydratase, 0.2 M substrate, 12 U yeast alcohol dehydrogenase

(Sigma-Aldrich), 0.2 mM NADH, 0.04 M potassium phosphate buffer (pH 8.0), and 10

228

11M coenzyme B12 in a total volume of 1.0 mL. The reaction was carried out at 37 °C in a
cuvette with a 1.0 cm light path and started by adding coenzyme to the reaction mixture
which was preincubated at 37 °C for 5 min. One unit is defined as the amount of enzyme
activity catalyzing the formation of 1 umol of aldehyde per min under the standard

conditions.

2-Keto acid decarboxylase assay

The specific activities of 2-keto acid decarboxylases including pyruvate
decarboxylase and benzoylformate decarboxylase were determined by coupling the
decarboxylation reaction with aldehyde-dependent oxidation of NADH by equine liver
alcohol dehydrogenase. Resuspension buffer contained sodium phosphate (50 mM, pH
6.5) and MgCl2 (10 mM). The enzyme assay solution (1 mL) contained sodium phosphate
(50 mM, pH 6.5), MgCl2 (10 mM), thiamine pyrophosphate (0.15 mM), NADH (0.2 mM),

and an appropriate amount of cell lysatezo’2|

When pyruvate decarboxylase was assayed
for specific activity on pyruvate, equine liver alcohol dehydrogenase (0.05 U) and
pyruvate (5 mM) were also included in the assay.20 When benzoylformate decarboxylase
was assayed for specific activity on benzoylformate, equine liver alcohol dehydrogenase
(0.05 U) and benzoylformate (2 mM) were also included in the assay.21 When 2-keto
decarboxylases were assayed for specific activity on 3-deoxy-D,L-glycera-pentulosonic
acid, equine liver alcohol dehydrogenase (1 U) and 3-deoxy-D,L-glycera-pentulosonate

(50 mM) were also included in the assays. Enzyme activity was measured

spectrophotometrically by monitoring the oxidation of NADH at 340 nm. One unit of 2-

229

keto acid decarboxylase was defined as the conversion of l umol of NADH (e = 6,220 M-

1 -l -
cm ) per mm at room temperature.

Alcohol dehydrogenase assay

The enzyme activities of alcohol dehydrogenases including adhA- and ath-
encoded alcohol dehydrogenases 22 of Zymamanas mobilis and 1,3-propanediol
oxidoreductase23 were measured in the oxidative direction by monitoring the formation of
NADH. Resuspension buffer for adhA- and ath-encoded alcohol dehydrogenase
contained potassium phosphate (30 mM, pH 6.5), sodium ascorbate (10 mM), and
Fe(NH4)2(804)2 (0.5 mM).22 Enzyme assays (1 mL) of AdhA and Ath on the native
substrate contained Tris-HCl (30 mM, pH 8.5), NAD (1 mM), appropriate amount of cell
lysate, and ethanol (1 mM).22 E. coli alcohol dehydrogenase candidates were assayed
under the same conditions as AdhA and Ath. Cell resuspension buffer for 1,3-
propanediol oxidoreductase contained potassium carbonate (100 mM, pH 9.0) and
ammonium sulfate (35 mM).23 When the specific activity of 1,3-propanediol
oxidoreductase on native substrate was measured, the enzyme assay solution (1 mL)
contained potassium carbonate (100 mM, pH 9.0), NAD (1 mM), appropriate amount of
cell lysate, and 1,3-propanediol (0.1 mM).23 When the above alcohol dehydrogenases
were assayed for specific activity on nonnative substrate, 1,2,4-butanetriol (50 mM) was
included in the assay in place of the native substrate. Enzyme activity was measured

spectrophotometrically by monitoring the formation of NADH at 340 nm. One unit of

230

alcohol dehydrogenase was defined as the formation of l umol of NADH (c = 6,220 M'1

.1 .
cm ) per mm at room temperature.

Protein SDS-PAGE analysis

Protein SDS-PAGE analysis followed the procedure described by Harris. 24
Preparation of a 10% separating gel started from mixing 3.33 mL of 30% (w/v) aqueous
acrylamide stock solution containing N,N’-methylene-bisacrylamide (0.8% (w/v)), 2.5 mL
of 1.5 M Tris-HCl (pH 8.8), and 4 mL of distilled deionized water. After degassing the
solution using a water aspirator for 30 rrrin, 0.1 mL of 10% (w/v) aqueous ammonium
persulfate solution, 0.1 mL 10% (w/v) aqueous SDS solution, and 0.005 mL of N, N, N’,
N’-tetramethylethylenediamine (TEMED) were added. The solution was mixed
thoroughly and poured into a 0.1 cm-width gel cassette to about 1.5 cm below the top of
the gel cassette. t-Amyl alcohol was overlaid on top of the solution and the gel was
allowed to polymerize for l h at rt. The stacking gel was prepared by mixing 1.7 mL 30%
acrylamide stock solution containing N,N’-methylene-bisacrylamide (0.8% (w/v)), 2.5 mL
Tris-HCI solution (0.5 M, pH 6.8), and 5.55 mL of distilled deionized water. After
degassing for 30 min, 0.1 mL of 10% ammonium persulfate, 0.1 mL 10% SDS, and 0.01
mL of TEMED was added, and the solution was mixed thoroughly. t-Arnyl alcohol was
removed from the top of the gel cassette, which was subsequently rinsed with water and
wiped dry. After insertion of the comb, the gel cassette was filled with stacking gel
solution, and the stacking gel was allowed to polymerize for 1 h at rt. After removal of
the comb, the gel cassette was installed into the electrophoresis apparatus. The electrode

chamber was then filled with electrophoresis buffer containing glycine (192 mM), Tris

231

base (25 mM), and 0.1% SDS (w/v). Following dilution with Laemmli sample buffer (10
uL, Sigma 8-3401) consisting of 4% SDS, 20% glycerol, 10% 2-mercaptoethanol, 0.004%
bromophenol blue, and Tris-HCI (125 mM, pH 6.8), each protein sample (10 pL) was
heated at 100 °C for 10 min. Samples and markers (MW-SDS-200, Sigma) were then
loaded into the sample wells and the gel was run under constant current at 30 mA until the
blue tracking dye (bromophenol blue) reached the interface of stacking gel and separating
gel. The protein gel was then run at a higher current (50 mA). When the blue tracking
dye reaches the bottom of the gel, electrophoresis was terminated. The protein gel was
subsequently removed from the cassette and submerged in 10% (w/v) aqueous
trichloroacetic acid solution with constant shaking for 30 min. The protein gel was then
transferred into a solution containing 0.1% (w/v) Comassie Brilliant Blue R, 45% (v/v)
MeOH, 10% (v/v) HOAC in H20 and stained with constant shaking for 4 h. Destaining of
the protein gel was carried out in a solution containing 45% (v/v) MeOH, 10% (v/v)
HOAC in H2O for 2-3 h. For long—term storage, SDS-PAGE gels were sealed in plastic

bags containing 10% glycerol.

Chapter 2

Purification of 2-deoxyribose-5-phosphate aldolase.

E. coli strain DH50L/pVH17 was obtained from the American Type Culture
Collection (ATCC 86963). Cells were grown in rich medium (1 L) containing 30 g
tryptone, 20 g yeast extract, and 10 g 4-morpholine-propanesulfonic acid at pH 7.

Cultures were initiated by inoculating a single colony into 100 mL of rich medium

232

containing 50 ug/mL ampicillin. Inoculants were grown at 37 °C with agitation at 250
rpm for 12 h, and a 10 mL aliquot was subsequently transferred to 1 L of rich growth
medium. A total of 4-1 L cultures of inoculated rich media were then cultured for 12 h at

37 °C. Following centrifugation at 4,000g for 5 min, approximately 50 g of wet cells were

obtained. All subsequent manipulations were performed at 4 °C. Cells were resuspended
in 100 mL 100 mM Tris, 2 mM EDTA, pH 7.6 and lysed by two passages through a
French press (1220 atm). Cellular debris was removed by centrifugation at 20,000g for 30
min to afford cell-free lysate containing 38,000 units of aldolase activity. The crude lysate
was applied to Q-Sepharose Fast-Flow anion exchange resin (300 mL), which had been
equilibrated with 100 mM Tris and 2 mM EDTA, pH 7.6. The column was eluted with a
linear gradient (500 mL + 500 mL, 0—1 M) of NaCl in the same buffer. The fractions were
assayed as described previously”25 Active fractions were pooled and concentrated to

afford a total of 29,400 units of aldolase activity with a specific activity of 43.9 U/mg.

Enzymatic synthesis of 3,4-dihydroxy-D-butanal

The synthesis of 3,4-dihydroxy-D-butanal was modified from a literature
procedure.25 2-Deoxyribose-5-phosphate aldolase (20,000 U) was added to a 400 mL
solution of 50 mM triethanolamine buffer (pH 7.3) containing glycoaldehyde dimer (2.25
g, 18.7 mmol), acetaldehyde (4.94 g, 112.2 mmol), and EDTA (0.15 g, 0.4 mmol). The
resulting solution was stirred in the dark for 72 h under Ar. The reaction was quenched by
addition of 1 L of acetone and then cooled to 0 °C for l h. Precipitated protein was
removed by centrifugation. After removal of the solvent under reduced pressure, the

residue was passed through Dowex 50 (H+) to remove triethanolamine and then

233

concentrated to a solution of 150 mL. This crude 3,4-dihydroxy-D-butanal was
characterized by 1H and '3 C NMR spectroscopy26 as the major product and subsequent

hydrogenation was carried out without further purification.

Catalytic hydrogenation of 3,4-dihydroxy-D-butanal

The solution obtained from the previous step was placed in a glass reaction vessel
along with 5 wt % Ru on C (0.76 g, 1.0 mol %). The glass reaction vessel was inserted
into the 500 mL Parr 4575 stainless steel high temperature-high pressure reactor and the
vessel sealed. The temperature and stirring rate were controlled by a Parr 4842
temperature controller. Hydrogen was bubbled through the reaction mixture for 10-15
min to remove air while stirring at 100 rpm. The vessel was then charged with 13.6 atm
H2. After heating the reaction to 30 °C, it was allowed to stir for 5 h at 30 °C. After
removal of the catalyst by filtration through Celite, the reaction solution was concentrated
to a yellow syrup, and a portion of the residue was derivatized by
bis(trimethylsilyl)trifluoroacetamide followed by GC analysis. D—l,2,4-Butanetriol was
obtained as a yellow oil (0.34 g, 17%) following Kugelrohr distillation (165 °C/ 0.2 mm

Hg) of the syrup. The 1H NMR and 13 C NMR were identical to authentic sample.

234

Chapter 3

Artificial biosynthesis of 1,2,4-butanetriol from erythritol

Plasmid pML5. I 76

The gldBC gene was amplified from K. pneumoniae ATCC25955 genomic DNA
using 5’-GTAGTGGTCGACGTGCAACAGACAACCCAAAT as the forward primer, 5’-
CTATACAAGCTTGCTGACCTCCGCTTAGCTTC as the reverse primer (SalI and
Hindlll restriction sites are underlined), and PfuTurbo as the DNA polymerase. The
resulting 1 kb was cloned into pJF118EH to afford pML5.176. Transcription of gldBC is

in the same orientation as the Plac promoter.

Plasmid pML5. 1 79

The dhaT gene was amplified by PCR from K. pneumoniae ATCC25955 genomic
DNA using 5’-CGGCAGAAGCTTGAGAAGGTATATTATGAGCT-3’ as the forward
primer and 5’-GTTAGCAAGCTTCCTCGTTAACACTCAGAATG-3’ as the reverse
primer (Hindlll restriction sites are underlined), and PfuTurbo as the DNA polymerase.
The resulting 1.2 kb fragment was cloned into pJF118EH to afford pML5.179.

Transcription of dhaT is in the same orientation as the Pm promoter.

Plasmid pML5. 200

The gldA gene (glycerol dehydratase 0t subunit) was amplified by PCR from K.
pneumoniae ATCC25955 genomic DNA using 5’-CGTCGCGAATTCCCI I I IGAGC-
CGATGAACAA as the forward primer, 5’-GCTAACGTCGACACCGCCTTATTCA-

ATGGTGT as the reverse primer (EcoRI and SalI restriction sites are underlined), and

235

PfuTurbo as the DNA polymerase. The resulting 1.7 kb fragment was digested and
ligated with pML5.176 to produce pML5.200. The gldA gene is transcribed in the same

direction as gldBC.

Plasmid pML5. 226

The dhaT gene was amplified by PCR from pWN5.022A using 5’-CGGCAG-
AAGCTTGAGAAGGTATATTATGAGCT-3’ as the forward primer and 5’-
GTTAGCAAGCTTCCTCGTTAACACTCAGAATG-3’ as the reverse primer (Hindlll
restriction sites are underlined), and PfuTurbo as the DNA polymerase. The resulting 1.2
kb fragment was cloned into pML5.200 to afford pML5.226. The dhaT gene cluster is

transcribed in the same direction as gldABC.

Plasmid pWN5. 220A

The dhaT gene was amplified by PCR from K. pneumoniae ATCC25955 genomic
DNA using 5’-ACGGATCCGCGAGAAGGTATATTATGAGC as the forward primer,
5’-TCGGATCCCCTCGTTAACACTCAGAATGC as the reverse primer (BamHl
restriction sites are underlined), and PfuTurbo as DNA polymerase. The resulting 1.2 kb
fragment was digested and ligated with pT7-7 to produce pWN5.220A. Transcription of

dhaT is in the same orientation as the phage T7 promoter.

Plasmid pWN5.260A

The pddABC gene cluster was amplified by PCR from K. pneumoniae
ATCC25955 genomic DNA using 5’-CCGAATTCACAI I IAATATCCGCGAGGC as

the forward primer, 5’-CGGAATTCGCTTACCCCTGTTAATCGTC as the reverse

primer (EcoRI restriction sites are underlined), and PfuTurbo as DNA polymerase. The

236

resulting 3.0 kb fragment was digested and ligated with pWN5.220A to produce

pWN5.260A. The pddABC gene cluster is transcribed in the same direction as dhaT.

Plasmid p WN5. 258A

The pddABC-ddrAB gene cluster was amplified by PCR from K. oxytoca
ATCC8724 genomic DNA using 5’-CCGAATTCCGTCCTACATCTGATACCCA as the
forward primer, 5’-CGGAATTCTGTTTCAACCAGTCCCAGTG as the reverse primer
(EcoRI restriction sites are underlined), and PfilTurbo as DNA polymerase. The resulting
3.0 kb fragment was digested and ligated with pWN5.220A to produce pWN5.258A. The

pddABC-ddrAB gene cluster is transcribed in the same direction as dhaT.

Anaerobic cultivation of Klebsiella and Lactobacillus strains

Anaerobic culturing of Klebsiella pneumoniae, Klebsiella oxytoca and
Lactobacillus brevis strains followed the literature protocol.27 All solutions were prepared
in distilled, deionized water. Culture medium contained KH2P04 (5.4 g), (NH4)2SO4 (1.2
g), MgSO4-7H20 (0.4 g), tryptone (2.0 g), yeast extract (2.0 g) and carbon source (9 g) in
l L H20. The buffering reagent, grth factor and carbon sources were prepared and
autoclaved separately. Prior to autoclaving, solutions were saturated with Nz/COz (95:5,
v/v) gas by bubbling through a boiled solution. The process continued until the solution
was cooled down to rt. Serum bottles containing treated solution were sealed with a butyl
rubber stopper and aluminum cap. The cell culture was initiated by incubation of a single
colony of bacterial strain into nutrient broth (5 mL) and cultured at 37 °C with agitation.
A 5 mL aliquot of overnight culture was transferred into the serum bottle using a syringe.

The inoculant was cultured at 37 °C without agitation. Four different carbon sources were

237

fed to the cultures. These include glycerol, erythritol, glycerol/erythritol (1:1, wt/wt), and
glycerol/erythritol (1 :8, wt/wt). Cell free culture medium of each strain was analyzed by

GC.

In vitro dehydratase reaction of erythritol

In vitro enzymatic reactions contained potassium phosphate (33 mM, pH 8.6), KCl
(50 mM), vitamin BI; (15 uM), erythritol (0.2 M), and an appropriate amount of cell
lysate. A 5 mL reaction mixture was setup and stirred for 24 h at rt. Formation of product

1,2,4-butanetriol was monitored by GC.

In search of 3-deoxy-glycero-pentulosonate decarboxylase genes for the microbial

sysnthesis of 1,2,4-butanetriol from pentoses

Plasmid pML2. 1 18

The potential decarboxylase gene was amplified from Pseudomonas fluorescens
Pf—5 ATCC BAA477 genomic DNA using 5’-CGGAATTCGGCGGCTGCAAGAGT-
AACTC as the forward primer, 5’-TCGGATCCGCGTCATCTATCACAGCAGT as the
reverse primer (EcoRI and BamHI restriction sites are underlined), and PfuTurbo as the
DNA polymerase. The resulting 1.7 kb gene was cloned into pJF118EH to afford

pML2.118. Transcription of this ORF is in the same orientation as the Plac promoter.

Plasmid pML2. 123

The potential decarboxylase gene was amplified from Pseudomonas aeruginosa
ATCC47085 genomic DNA using 5’-CGGAATTCGTAGCGCAAATCCAGGAAAC as

the forward primer, 5’-ATGGATCCTCAGGGTTCGATGGI I IGCG as the reverse
238

primer (EcoRI and BamHI restriction sites are underlined), and PfuTurbo as the DNA
polymerase. The resulting 1.7 kb gene was cloned into pJF118EH to afford pML2.123.

Transcription of this ORF is in the same orientation as the Plac promoter.

Plasmid pML2. 1 62

The potential decarboxylase gene was amplified from Burkholderia fungorum
LB400 genomic DNA using 5’-GCGCCCCCGGGTCATTTGATAGATATTAAGG as
the forward primer, 5’-TTCCCCCGGGTAAGTGGCTTGCTCATGGCT as the reverse
primer (Smal restriction sites are underlined), and PfiiTurbo as the DNA polymerase. The
resulting 1.7 kb gene was cloned into pJF118EH to afford pML2.162. Transcription of

this ORF is in the same orientation as the Pzac promoter.

Plasmid pML2.208

The potential decarboxylase gene was amplified from Streptomyces coelicolor
ATCC BAA471 genomic DNA using 5 ’-CGGAATTCTCACTTCGAAAGGAACGTCC
as the forward primer, 5’-TAGGATCCTGATCTTCGGTGTGGTGTCG as the reverse
primer (EcoRI and BamHI restriction sites are underlined), and PfiiTurbo as the DNA
polymerase. The resulting 1.7 kb gene was cloned into pJF118EH to afford pML2.208.

Transcription of this ORF is in the same orientation as the Plac promoter.

Plasmid pML2. 2 1 4

The potential decarboxylase gene was amplified from Novosphingobium
aromaticivorans ATCC700278 genomic DNA using 5’-CGGAATTCTGCGGCAGTTC-

ATCCTCGGT as the forward primer, 5’—ATGGATCCAAGGGCGGTCGTCGATAAGG

as the reverse primer (EcoRI and BamHI restriction sites are underlined), and PfuTurbo as

239

the DNA polymerase. The resulting 1.7 kb gene was cloned into pJF118EH to afford

pML2.214. Transcription of this ORF is in the same orientation as the P1“ promoter.

Plasmid pML2. 25 6

The potential decarboxylase gene was amplified from Rhodopseudomonas
palastris ATCC BAA98 genomic DNA using 5’—ATGAATTCAGGAAGCTTCGGCGG-
TGGG as the forward primer, 5’-TCGGATCCGCTGGACAACTGATGACGCA as the
reverse primer (EcoRI and BamHI restriction sites are underlined), and PfiiTurbo as the
DNA polymerase. The resulting 1.7 kb gene was cloned into pJF118EH to afford

pML2.25 6. Transcription of this ORF is in the same orientation as the P106 promoter.

Plasmid pML2. 286

The potential decarboxylase gene was amplified from Mycobacterium smegmatis
ATCC700084 genomic DNA using 5’-ATGAATTCACTGGCAGGACGGCGTGAGC as
the forward primer, 5’-GCGGATCCCATCGTTCGGTCTCCTGTTT as the reverse
primer (EcoRl and BamHI restriction sites are underlined), and PfuTurbo as the DNA
polymerase. The resulting 1.7 kb gene was cloned into pJF118EH to afford pML2.286.

Transcription of this ORF is in the same orientation as the P1“ promoter.

Plasmid pML3. 040

The potential decarboxylase gene was amplified from Acidithiobacillus
ferrooxidans ATCC BAA477 genomic DNA using 5’-CGGAATTCGTTGCI I IGGGG-
GGTTGACA as the forward primer, 5’-TCGGATCCTATGCCCAACCATTCCATTC as

the reverse primer (EcoRI and BamHI restriction sites are underlined), and PfirTurbo as

240

the DNA polymerase. The resulting 1.7 kb gene was cloned into pJF118EH to afford

pML3.040. Transcription of this ORF is in the same orientation as the Plac promoter.

Plasmid pML3. 05 4

All point mutations were introduced into the plasmid with a QuikChange site
directed mutagenesis kit (Stratagene). The general protocol was described in the
manufacturer’s manual. Oligonucleotide primers carrying the mutation BFD A4601 and
flanked by complementary sequence of the plasmid template pWN5.238A were designed
as follows: 5’-ATGAACAACGGCA_CgTACGGTATCTTGCGATGG and 5’-
CCATCGCAAGATA-CCGTA_cG_TGCCGTTGTTCAT (The mutated codons are
underlined, and the lowercase letters indicate a base change from wild-type).28 The PCR
reaction mixture containing the primers (125 ng each), 1 uL supplied dNTP mix, 10 ng
plasmid pWN5.238A, 5 uL reaction buffer and 2.5 U PfizTurbo DNA polymerase was
made up to 50 uL using membrane filtered deionized H20. Following temperature
cycling, the reaction mixture containing the 1.7 kb nicked plasmid was treated with DpnI,
subsequently the mixture was transformed into XLl-Blue supercompetent cells. The
transformants were plated on LB solid medium containing ampicillin. Plasmid was
purified from a single colony and the mutation was confirmed by sequencing using
primers 5’-ATAACGGTTCTGGC-AAATATT, 5’-GCCGGCTTTCATTAACCTGCAT,
5’-GATCCTCAGTCCCACCAC-CTTT, 5’-CGAAGGTCACGATGTGGTTTTG, and 5’-
GAGAATGCGATTTACCTG-AACG. The resulting plasmid was designated as

pML3.054.

241

Plasmid pML3. 062

Oligonucleotide primers carrying the mutation PDC I472A and flanked by
complementary sequence of the plasmid template pLOI27629 were designed as follows:
5’-GATCAATAACTATGG'I'TACACQgQCGAAGTTATGATCCATGATG and 5’-
CATC—ATGGATCATAACT‘TCGgfiGTGTAACCATAGTTATTGATC (The mutated
codons are underlined, and the lowercase letters indicate a base change from wild-type).
The point mutation was generated as described earlier and confirmed by sequencing
analysis using primers 5’-CAAGGAAACAGCTATGACC, 5’-ACAACCTCGTCCT-
TCTTGAC, 5’-GAAGAAGCCGGTTTATCTC, 5’-ACCGCGTT-CTGTCGTCGTTA,
and 5’-CAGGAAGTCGCTCAATGGT. The resulting plasmid was designated as

pML3.062.

Plasmid pML3. 084

Oligonucleotide primers carrying the mutation PDC I4728 and flanked by
complementary sequence of the plasmid template pLOI27629 were designed as follows:
5’-TCAATAACTATGGTTACACCAgQGAAGTTATGATCCATGATG and 5’-
CATCATGGATCATAACTTC_Gc_TGGTGTAACCATAGTTATTGA (The mutated
codons are underlined, and the lowercase letters indicate a base change from wild-type).
The point mutation was generated as described earlier and confirmed by sequencing
analysis using the same primer set as for generating plasmid pML3.062. The resulting

plasmid was designated as pML3.084.

242

Plasmid pML3. 086

Oligonucleotide primers carrying the mutation PDC 14720 and flanked by
complementary sequence of the plasmid template pLOI27629 were designed as follows:
5’-GATCAATAACTATGGTTACACngQGAAGTTATGATCCATGATG and 5’—
CATCATGGATCATAACTTCQ_§9_GGTGTAACCATAGTTATTGATC (The mutated
codons are underlined, and the lowercase letters indicate a base change from wild-type).
The point mutation was generated as described earlier and confirmed by sequencing
analysis using the same primer set as for generating plasmid pML3.062. The resulting

plasmid was designated as pML3.086.

Plasmid pML3 . 1 04

Oligonucleotide primers carrying the mutation PDC I472T and flanked by
complementary sequence of the plasmid template pLOI27629 were designed as follows:
5’-TCAATAACTATGGTTACACCAQQGAAGTTATGATCCATGATG and 5’-
CATCATGGATCATAACTTCQgIGGTGTAACCATAGTTATTGA (The mutated
codons are underlined, and the lowercase letters indicate a base change from wild-type).
The point mutation was generated as described earlier and confirmed by sequencing
analysis using the same primer set as for generating plasmid pML3.062. The resulting

plasmid was designated as pML3.104.

243

Directed evolution of 3-deoxy-glycero-pentulosonate decarboxylase for microbial

1,2,4-butanetriol synthesis

Plasmid pML3. 200

Oligonucleotide primers carrying the deletion of EcoRI site and flanked by
complementary sequence of the plasmid template pML2.162 were designed as follows:
5’-CCAACGCGTGGA_A_CTCGCATACGCCG and 5’-CGGCGTATGCGAgT_TCCACG-
CGTTGG (The mutated codons are underlined, and the lowercase letters indicate a base
change from wild-type). The resulting plasmid was designated as pML3.171. To delete
the BamHI site, another set of primers were designed as follow: 5’-
GCGGCGGCGGGCATCCAGCTCGCG and 5’-CGCGAGCTGGATGCCCGCCGCC-
GC. The resulting plasmid was designated pML3.200. These mutations were verified by
PCR analysis using primers 5’-ATAACGGTTCTGGCAAATATT and 5’-

GTGGGACCACCGCGCTACTGC.

Plasmid pML3. 224

The P. fluorescens parent gene candidate was amplified from plasmid pML2.118
using 5’-TACGCTCAAGCTTGCCTCCATGAAAACCGTCCA as the forward primer,
5’-GCGTCTGAATTCTCTCTCATCCGCCAAAACAG as the reverse primer (Hindlll
and EcoRI restriction sites are underlined), and PfiiTurbo as the DNA polymerase. The
resulting 1.7 kb gene was cloned into pJF 1 18HE to afford pML3.224. Transcription of

this ORF is in the same orientation as the P100 promoter.

244

Plasmid pML3.225

The P. aeruginosa parent gene candidate was amplified from plasmid pML2.123
using 5’-TACGCTCAAGCTTGCCTCCATGAAAACCGTCCA as the forward primer,
5’-GCGTCTGAATTCTCTCTCATCCGCCAAAACAG as the reverse primer (Hindlll
and EcoRI restriction sites are underlined), and PfuTurbo as the DNA polymerase. The
resulting 1.7 kb gene was cloned into pJF118HE to afford pML3.225. Transcription of

this ORF is in the same orientation as the Plac promoter.

Plasmid pML3. 226

The B. fungorum parent gene candidate was amplified from plasmid pML3.200
using 5’-TACGCTCAAGCTTGCCTCCATGAAAACCGTCCA as the forward primer,
5’-GCGTCTGAATTCTCTCTCATCCGCCAAAACAG as the reverse primer (Hindlll
and EcoRI restriction sites are underlined), and PfitTurbo as the DNA polymerase. The
resulting 1.7 kb gene was cloned into pJF118HE to afford pML3.226. Transcription of

this ORF is in the same orientation as the Plac promoter.

Chimera library generation for ssDNA family shuflling

The original DNA shuffling was invented by Stemmer.30 ssDNA shuffling is a
similar DNA chimera library generation technique developed by Zhao and Arnold.31 The
decarboxylase genes of P. fluorescens, P. aeruginosa and B. fungorum were amplified
from plasmids pML3.224, pML3.225 and pML3.226, respectively, using the following
primers: 5’-TACGCTCAAGCTTGCCTCCATGAAAACCGTCCA and 5’-GCGTCT-
GAATTCTCTCTCATCCGCCAAAACAG. The Hindlll and EcoRI sites are underlined.

Phosphorylated forward or reverse primer was used to yield the desired single DNA

245

strand. A 100 uL reaction mixture contained: 10 uL of 10X Pfiz buffer, 10 uL of 2 mM
dNTP mix, 50 pmol of each primer, 8 uL of dimethylsulfoxide (DMSO), 50 ng of
template plasmid and 2.5 U of PfiiTurbo DNA polymerase. Amplification of DNA was

carried out for 35 cycles under the following conditions: 94 °C for l min, 58 °C for l min,

and 72 °C for 2.5 min.

5 pg of each purified decarboxylase gene candidates were diluted to 45 uL with
sterile water and 5 uL of 10>< lambda exonuclease buffer was added. Each mixture was
digested with 10 U lambda exonuclease at 37 °C for 90 min. The resulting ssDNA was
then purified by agarose gel electrophoresis. 0.5 ug of each purified ssDNA was mixed
and diluted to 42.5 uL, treated with 5 uL of 0.5 M Tris-HCl (pH 7.5) and 2.5 uL of 0.2 M
MnClz, and digested with 0.02 U DNase I at 15 °C for 30 sec. The reaction was
terminated by the addition of endostop and was purified by low melting agarose gel
electrophoresis (1%). 500 -— 1,000 bps DNA fragments were recovered using a Zymoclean
gel extraction kit.

Fragments were reassembled in a 20 pl. reaction mixture containing 5 uL of
purified fragment DNA, 2 uL of 10X Pfu buffer, 1 uL DMSO, 2 uL dNTP mix (2 mM
each) and 1 U PfuTurbo DNA polymerase. Cycling protocol: 96 °C for 90 seconds; 50
cycles of (94 °C for 1 min; 55 °C for 1 min; 72 °C for 2.5 min); 72 °C for 7 min and 4 °C
thereafter. This step was monitored by running a small aliquot on an agarose gel until a
smear extending through 1.6 kb was seen.

To amplify full-length 1.6 kb genes, the reassembly reaction served as the DNA
template and was diluted 2-fold. The same 50 uL PCR mixture used to acquire DNA for

lambda exonuclease digestion was then setup. The PCR program is as follow: 4 min at

246

96 °C followed by 35 cycles of 1 min at 94 °C, 1 min at 58 °C, 2.5 min at 72 °C and finally
7 min at 72 °C. The final full-length PCR product was purified by agarose gel
electrophoresis. The purified genes were digested by Hindlll and EcoRI and subsequently

cloned into pJF118HE for in vivo screening.

Chimera library generation for dsDNA family shuflling

Chimera library for dsDNA family shuffling was generated according to a similar
procedure described by Joern and Arnold. 32 P. fluorescens, P. aeruginosa and B.
fungorum genes were amplified using non-phosphorylated primers: 5 ’-TACGCTCAAG-
CTTGCCTCCATGAAAACCGTCCA and 5’-GCGTCTGAATTCTCTCTCATCCG-
CCAAAACAG from pML3.224, pML3.225, and pML3.226, respectively. A 100 uL
reaction mixture contained: 10 uL of 10X Pfu buffer, 10 uL of 2 mM dNTP mix, 50 pmol
of each primer, 8 uL of dimethylsulfoxide (DMSO), 50 ng of template plasmid and 2.5 U
of PfuTurbo polymerase was setup. Amplification of DNA was carried out for 35 cycles
under the following conditions: 94 °C for 1 min, 58 °C for 1 min, and 72 °C for 2.5 min.
The P. putida gene was obtained by restriction enzyme digestion from pWN5.23 8A using
enzymes EcoRI and BamHI. All four gene candidates were purified by agarose gel
electrophoresis prior to use.

After purification and quantification, equal amounts of parent DNA (as determined
spectrOSCOpically by UV-vis absorbance, it = 260 nm) were mixed and subjected to DNase
I digestion. A 50 uL digestion contained 4 ug of parent DNA mix, 5 uL of 0.5 M Tris-

HCl (pH 7.4), 2.5 pL of 0.2 M MnClz, and 0.02 U of DNase I. Afier 4 min digestion at 15

247

°C, 5 1.1L of endostop was added. Using Zymoclean gel extraction kit, fragments from
100 — 300 kb were extracted.

Fragments were reassembled in a 50 uL reaction containing all the purified DNA
fragments, 10 uL of 2 mM dNTP mix, 5 uL of Pfu buffer, 2 uL DMSO and 2.5 U of
PfuTurbo DNA polymerase. Cycling was according to the following protocol: 96 °C, 90
sec; 35 cycles of (94 °C, 30 s; 65 °C, 90 s; 62 °C, 90 s; 59 °C, 90 s; 56 °C, 90 s; 53 °C, 90
s; 50 °C, 90 s; 47 °C, 90 s; 44 °C, 90 s; 41 °C, 90 s; 72 °C, 4 min); 72 °C, 7 min; 4 °C
thereafter. 5 uL of the reaction mixture was analysed on the gel and a smear of
reassembled DNA that extended above the parent gene size (1.6 kb) was seen.

PCR was carried out for the reassembly mixture using primers 5’-TACGC-
TCAAGCTTGCCTCCATGAAAACCGTCCA, and 5’-GCGTCTGAATTCTCTCTC-
ATCCGCCAAAACAG under the same conditions used to acquire DNA for
fragmentation. The DNA template used in the reaction, however, was diluted 80-fold.
Cycling was done using the following program: 94 °C for 1 min, 58 °C for 1 min, and 72
°C for 4 min. The final full-length PCR product was purified by agarose gel
electrophoresis. The purified genes were digested by Hindlll and EcoRI and subsequently

cloned into pJF l 18HE for in vivo screening.

Library screening for 3 ~deoxy-glycero-pentulosonate decarboxylase

Library screening was carried out in a 96-well format. Purified genes obtained
from the family shuffling were transformed into electrocompetant E. coli W3110 and
plated on M9/xylonate/Ap plates. This step was optimized such that the amount of

colonies appeared on each plate was around 200. Single colonies were then used for

248

screening using the protocol as follows. The daily throughput of this assay is 200
colonies.

A fresh growth block (Qiagen) with each well containing 1 mL of M9/xylonate/Ap
medium was inoculated with single colonies from the chimera library. The first column of
each block was inoculated with the controls. Wells A1 and H1 were blanks. Wells Bl —
G1 were inoculated with E. coli W3110/pWN5.238A, W3110/pWN5.238A* (this plasmid
harbors an mdlC mutant isolated earlier with 2-fold activity improvement 33 ),
W3110/pML3.224, W3110/pML3.225, W3110/pML3.226, and W3110/pWN5.238A in
order. The growth block was then covered by Qiagen Airpore Sheet and incubated in the
shaker at 37 °C for 36 hours.

The growth block was removed from the shaker and 150 uL was withdrawn from
each well and poured into an autoclaved 96-well plate (Nunc Inc.) containing 50 uL of
80% sterilized glycerol. The glycerol freeze plate was then covered with aluminum tape
and stored at —80 °C. From the same growth block, another 10 uL was drawn from each
well and used to inoculate another fresh block. The new block was then covered with
Qiagen Airpore Sheet and incubated at 37 °C for another 24 hours.

The growth block was removed from the shaker and 100 uL aliquot was poured
into a 96-well optical plate and the OD600 was measured. At this stage, the OD600 should
be around 0.6 and IPTG was added to a final concentration of 0.5 mM. The induced
culture block was then covered with a fresh piece of Airpore Sheet and was incubated at
37 °C for 48 hours.

The growth block was removed from the shaker and the final OD600 was recorded.

The cell was then centrifuged down at 400 rpm for 10 min. 100 uL supernatant from each

249

well was pipetted to a new block (F isherbrand, polypropylene) and this block was placed
in the CentriVap (Labconco). The supernatant was evaporated to dryness within 2 hours
by setting the core temperature to 70 °C. The residue in each well was then resuspended
into 10 uL H20 and submitted to GC derivatization using bis(trimethylsilyl)-
trifluoroacetamide (BSTFA). The BSTFA solution mix was prepared by mixing 50 mL
pyridine, 50 mL BSTFA and 500 uL dodecane in situ. The reaction began by pouring 1
mL of this solution mix into each well of the block. The block was sealed tight with a
polypropylene lid (Fisherbrand) and allowed to react overnight with agitation at 250 rpm
in a shaker.

Finally, the solution from the block was pipetted into GC sample vials for analysis.
The GC method used was developed to be isothermal with a 7 min method runtime.
Therefore, the time required for analyzing 200 samples is 24 hours. In a typical GC
chromatogram, the BT comes out at 2.86 min while the internal standard, dodecane comes
out at 2.34 min. Enzymes corresponding to these putatively more reactive
decarboxyalases were assayed for a second time using the same assay. Confirmed
decarboxylase mutants were assayed for a third time by culturing in 100 mL

M9/xylonate/Ap medium.

Bacteriophage T7 promoter and E. coli chaperones for deC expression

E. coli KIT2

The genomic antibiotic marker on E. coli WN13 was excised by the method

described by Cherepanov et al.34 E. coli WN13 was transformed with plasmid pCP20, and

250

ampicillin resistant transformants were selected at 30 °C, after which a few were colony-
purified once non selectively at 43 °C . They were then tested for the loss of all antibiotic
resistances. Disruption of the genomic chloramphenicol marker was confirmed by PCR
analysis using the following primers: 5’-TACGACATCATCCATCACCCGCGGCA-
TTAC and 5’-CAGAAGTGGC-TGATAGAGGCGACGGAACGT. This W3110ijhH-

AyagEserAxyIABzzxdh-F RT strain was designated as E. coli KIT10.

E. coli K1T15

E. coli KIT2 was grown in LB supplemented with 0.2% maltose, 10 mM MgSO4 at
37 °C in the shaker to an OD600 = 0.5. The DE3 lysogen of E. coli KIT2 was then
prepared using the DE3 lysogenation kit of Novagen according to the manufacturer’s

protocol. E. coli KIT15 was the result and this new strain was used as host for phage T7

expression vectors.

Plasmid pML 7. 128

A 1.6 kb fragment encoding the P. putida benzolformate decarboxylase gene deC
was excised from plasmid pWN5.238A by digestion with EcoRI and ligated to the 5.4 kb
plasmid pET28c(+), which had been previously treated with EcoRI, followed by CIAP, to

afford a 7.0 kb plasmid pML7.128.

Plasmid pML 7. 135

A 1.6 kb serA locus was excised from plasmid pRCl.55B by digestion with Smal
and ligated to the 6.5 kb plasmid pML7.128, which had been previously treated with Smal

and CIAP. The ligation mixture was transformed into E. coli WN13. Transfonnants

251

carrying the serA insert were selected on M9 medium plates. In the resulting 8.7 kb

plasmid pML7.135, the serA and deC genes are convergently transcribed.

Plasmid pML 7.1 66

To construct a plasmid harboring the Pm-groEL-groES gene fragment, a 3.4 kb
fragment carrying the groEL-groES chaperones and the trigger factor (TF) genes without
the tetracycline induced promoter Pztl sequence was amplified by PCR using 5’-CGG-
AAGCTTATGAATATTCGTCCATTGCA as the forward primer and 5’-ATAAGCTTA-
CGGGCCTTTGTGCGAATTTA as the reverse primer (HindIII restriction sites are
underlined) from the plasmid pG-Tf2 (Takara Bio Inc.). The PCR product was digested
with Hindlll and ligated with Hindlll digested pJF118HE to afford the 8.8 kb plasmid

pML7.166.

Plasmid pML 7.1 75

The 1.5 kb PmC-groEL-groES locus was amplified by PCR using 5’-
ACCCCGGGCGTAAATCACTGCATAATTCG as the forward primer and 5’-
ACCCCGGGTACCTCAAAAAATCACAGTGG as the reverse primer (Smal restriction
sites are underlined) from plasmid pML7.166. The PCR product was digested with Smal
and ligated to the 7.0 kb plasmid pWN5.238A, which had been sequentially treated with

Smal and CIAP, to afford the 8.5kb plasmid pML7.175.

Plasmid pML 7. 1 80

A 1.6 kb serA locus was excised from plasmid pRCl.55B by digestion with Smal
and ligated to the 8.5 kb plasmid pML7.128, which had been previously treated with Seal

and CIAP. The ligation mixture was transformed into E. coli WN13. Transformants

252

carrying the serA insert were selected on M9 medium plates. In the resulting 10.1 kb
plasmid pML7.135, the serA and P,ac-mdlC-P,ac-groEL-groES are convergently

transcribed.

Plasmid pML 7.202

A 1.5 kb fragment carrying the Pmc-groEL-groES region was amplified by PCR
using the same method as described earlier. The Smal digested PCR product was ligated
into pML7.135, which had been sequentially treated with BglII, Klenow and CIAP, to

afford the 10.2 kb plasmid pML7.202.

Discovery of alcohol dehydrogenases for D-1,2,4-butanetriol biosynthesis

E. coli K1 T1

The chromosomal gene insertion in E. coli was generated by modifying a method
described by Datsenko et al.35 To replace native E. coli xylAB with the xdh-ath-Pm
along with a chloramphenicol selectable marker in the chromosome, the PCR-mediated 7&—
Red recombination methodology was employed. A linear DNA cassette including these
genes was generated by PCR from pML6.272 using Platinum T aq High Fidelity DNA
polymerase (Invitrogen), with 5’-CACCCGCGGCATTACCTGATTATGGAGTTCA-
_A_TA_TGTCCTCAGCCATCTATCCC as the forward primer and 5 ’-
CAGAAGTTGCTGATAGAGGCGACGGAACGTTTCTCATATGAATATCCTCCTTA-
GT as the reverse primer. The sequences underlined represent the homology arms, while
the remainders are the priming sequences for hybridization to complementary nucleotide

sequences on the template plasmid pML6.272. The 3 kb PCR product was agarose gel-

253

purified and transformed into electrocompetent E. coli W3110/pKD46. SOC medium was
added after electroporation and incubated at 37 °C for 1 h. The cells were spread on LB
plates containing 25 ug/mL chloramphenicol and incubated overnight at 37 °C to select
for antibiotic transformants. The k—Red recombinase expression plasmid pKD46 has a
temperature sensitive replication origin and can be eliminated at 37 °C. Strain KIT1
(W3 1 10xylAB::xdh-ath-P,ac-F RT-CmR-F RT) was therefore constructed. The gene
insertion was verified by PCR analysis using primers 5’-TACGACATCATCC-

ATCACCCGCGGCATTAC and 5’-CAGAAGTGGCTGATAGAGGCGACGGAACGT.

E. coli KIT3

Pl phage mediated transduction was used to transfer the xylAB::xdh-ath-P,ac-
FRT-CmR-FRT mutation into E. coli WN7. E. coli strain KIT1 was used as the donor
strain while strain WN7 served as the recipient strain. The transformants were selected on
LB/Cm plate to produce another new strain KIT3 (W31lOijhHAyagEserAxyIABzzxdh-
ath-Pm-FRT-CmR-FRT). The gene insertion was verified by PCR analysis using
primers 5’-TACGACATCATCCATCACCCGCGGCATTAC and 5’-CAGAAGTGGC-

TGATAGAGGCGACGGAACGT.

E. coliK1T4

The genomic antibiotic marker on E. coli KIT3 was excised by the method
described by Cherepanov et al.34 E. coli strain KIT3 was electroporated with plasmid
pCP20 containing Flp recombinase gene, and grown on LB plate containing 100 ug/mL

ampicillin at 30 0C overnight. The Flp recombinase excises the chloramphenicol gene

flanked by FRT sites from the chromosome. After growth at 42 °C for one day in LB to

254

cure KIT3/pCP20 that has a temperature sensitive replication origin, ampicillin and
chloramphenicol sensitivity were tested to verify loss of pCP20 and the antibiotic marker.
The loss of the antibiotic gene marker was verified by PCR analysis using primers 5’—
TACGACATCATCCATCACCCGCGGCATTAC and 5’-CAGAAGTGGCTGATAGAG-
GCGACGGAACGT. The resulting strain was designated as E. coli KIT4

(W3 1 lOijhHAyagEserAxylAB: :xdh-ath-Pmc-FRT).

E. coli KIT5

The chromosomal gene insertion in E. coli was generated by the same protocol
used to generate strain KIT1. A 1.2 kb fragment carrying the PT5-FRT-CmR-FRT
sequence was amplified by PCR using Platinum T aq High Fidelity DNA polymerase, with
5 ’ -TCTCCCTTTTTCAGATCTCATCCATACTGGGTAGTGGCGAATAAACGTCTTG-
AGCGATTGTGTAG as the forward primer and 5’-AACTGCAGCCTTCATAGTTC-
CTCCTTTTCGGATGATGTTCTGCATAGTTAATTTCTCCTCTTTAATG as the
reverse primer. The sequences underlined represent the homology arms, while the
remainders are the priming sequences for hybridization to complementary sequences on
the template plasmid pML7.042. The PCR product was agarose gel purified and was
transformed into electrocompetent E. coli W3110/pKD46. The transformants were
selected on LB/Cm plate and verified by PCR analysis using primers 5’-DAGGG-
ATCCATGCAGAACATCATCCGAAAA and 5’-AGGGATCCTTAGTGACGGAAAT-

CAATCAC. E. coli strain KIT5 (W3110Pr5-FRT-CmR-FRT) was thus constructed.

255

F

E. coli K1 T6

E. coli train KIT6 (W31 10ijhHAyagEserAxylAB::xdh-FRT PT5—FRT-CmR-FRT)
was made from E. coli strain WN13 by P1 phage mediated transduction using KIT5 as the
donor strain. The newly constructed strain KIT6 was selected on an LB/Cm plate. The
inserted PT5-FRT-CmR-FRT gene fragment was verified by PCR analysis using primers
5’-DAGGGATCCATGCAGAACATCATCCGAAAA and 5’-AGGGATCCTTAGTGA-

CGGAAATCAATCAC.

E. coli KIT7

The genomic chloramphenicol marker was excised by the same protocol for
generating E. coli KIT4. Thermal induction of the Flp recombinase in KIT6/pCP20
excised the chloramphenicol marker and afforded strain KIT7 (W3110ijhHAyagEserA-
xylABzzxdh-F RT PT5-FRT). The loss of the antibiotic marker gene was verified by PCR
analysis using primers 5’-DAGGGATCCATGCAGAACATCATCCGAAAA and 5’-

AGGGATCCTTAGTGACGGAAATCAATCAC.

E. coli KIT8

The genomic ath inactivation was generated by the methods described by
Datsenko et a1.35 Linear DNA with an FRT-flanked chloramphenicol gene was amplified
by PCR from pKD3 using Platinum Taq DNA polymerase, with 5’-
ACTGGGTAGTGGCGAATAAATCTCATTTGCCTCACCTGCTGTGTAGGCTGGA-
GCTGCTTC as the forward primer, and 5’-GATGCTGCGACCCGAACATGGCAGTC-
GCAGCAAAGGCCTCCATATGAATATCCTCCTTAG as the reverse primer. The

sequence underlined represent the homology arms, while the remainders are the priming

256

sequences for hybridization to complementary sequences on the template plasmid pKD3.
The 1 kb PCR product was agarose gel purified and subsequently transformed into
electrocompetent E. coli W3110/pKD46. The transformants were selected on an LB/Cm
plate. Disruption of ath was confirmed by PCR analysis using the following primers:
5’-DAGGGATCCATGCAGAACATCATCCGAAAA and 5’-AGGGATCCTTAGTGA-
CGGAAATCAATCAC. The W3110athzzFRT-CmR-FRT strain was designated as E.

coli KIT8.

E. coli KIT9

P1 phage mediated transduction was employed to transfer the athzzFRT-CmR-
FRT mutation into E. coli KIT2 to produce KIT9 (W3110ijhHAyagEserAxylAB::xdh-
FRTathzzFRT-CmR-FRT). Disruption of ath in E. coli KIT9 was confirmed by PCR
analysis using the following primers: 5’-DAGGGATCCATGCAGAACATCATCCGAA-

AA and 5’-AGGGATCCTTAGTGACGGAAATCAATCAC.

E. coli KIT10

The genomic chloramphenicol marker was excised by the same protocol for
generating E. coli KIT4. Thermal induction of the F 1p recombinase in KIT9/pCP20
excised the chloramphenicol marker and afforded strain KIT7. The loss of the antibiotic
marker gene was verified by PCR analysis using primers 5’-
DAGGGATCCATGCAGAACATCATCCGAAAA and 5’-AGGGATCCTTAGTGA-
CGGAAATCAATCAC. This W3110ijhHAyagEserAxyIAB::xdh-FRTath::FRT strain

was designated as KIT10.

257

E. coliK1T16

The E. coli genomic yiaE inactivation was generated by the same protocol for
generating E. coli KIT8. The F RT-flanked chloramphenicol gene was amplified from
pKD3 using 5’-CGCCCGAAAAA'ITCTGCTGATGCACACGCCAACCGTATTAT-
GTGTAGGCTGGAGCTGCTTCG as the forward primer, and 5’-AGATCCAATATGC-
GGTACTGCGACGACGTTGGCCATTGAGCATATGAATATCCTCCTTAGT as the
reverse primer. The sequence underlined represent the homology arms, while the
remainders are the priming sequences for hybridization to complementary sequences on
the template plasmid pKD3. The 1 kb PCR product was agarose gel purified and
subsequently transformed into electrocompetent E. coli W3110/pKD46. The
transformants were selected on an LB/Cm plate. Disruption of yiaE was confirmed by
PCR analysis using the following primers: 5’-CGGAATTCGAATGGAGAGAAGCA-
TGAAG and 5’-CGGAATTCGATTATCGGGCTTTACTCTA. This E. coli
W31 lOijhHAyagE—serAxyIAB: :xdh-ath—Pm-FRTyiaE: :FRT-CmR-FRT was designated

as KIT16.

E. coli KIT17

Pl phage mediated transduction was employed to transfer the ych::FRT-KmR-
FRT mutation into E. coli KIT16 to produce KIT17 (W3110ijhHAyagEserAxylAB::xdh-
ath-Pm—FRTyiaE::FRT-CmR-FRTych::FRT-KmR-FRT). The phage (W3110-
ychzzFRT-KmR-F RT) was obtained in the Frost group strain collection. Disruption of
ych in E. coli KIT17 was confirmed by PCR analysis using the following primers: 5’-
CGGAATTCAGTATGGATATCATCTTTT and 5’-CGGAATTCCTGATGCTTTAT-

TAGTAG.

258

E. coliK1T18

The genomic chloramphenicol and kanamycin markers were excised by the same
protocol for generating E. coli KIT4. Thermal induction of Flp recombinase of
KIT17/pCP20 excised both antibiotic markers. The loss of antibiotic marker gene was
verified by PCR analysis using 5’-CGGAATTCGAATGGAGAGAAGCATGAAG, 5’-
CGGAATTCGATTATCGGGCTTTACTCTA, 5’-CGGAATTCAGTATGGATATCAT-
CTTTT and 5’-CGGAATTCCTGATGCTTTATTAGTAG This W3110ijhHAyagE-

serAxylAB::xdh-ath-PmC-FRTyiaE::FRTych::FRT strain was designated as KIT18.

Plasmid pML5. 1 79

The dhaT gene was amplified by PCR from K. pneumoniae genomic DNA using
5’CGGCAGAAGCTTGAGAAGGTATATTATGAGCT as the forward primer and 5’-
GTTAGCAAGCTTCCTCGTTAACACTCAGAATG as the reverse primer (Hindlll
restriction sites are underlined), and PfilTurbo as the DNA polymerase. The resulting 1.2

kb fragment was cloned into pJF118EH to afford a 6.5 kb plasmid pML5.179.

Plasmid pML 6. 090

A 1.2 kb dhaT gene was excised from plasmid pML5.179 by digestion with
BamHI. The resulting gene fragment was treated with Klenow, gel purified and ligated to

EcoRI-digested, Klenow treated and dephosphorylated pKK223-3 to generate the plasmid

pML6.090.

259

Plasmid pML6. 128

A 1.5 kb fragment carrying the tac promoter region and the dhaT gene was excised
from plasmid pML6.090 by digestion with BamHI and ligated to pWN5.238A, which had

been sequentially treated with BamHI and CIAP, to afford a 8.5 kb plasmid pML6.128.

Plasmid pML6. I 3 3

A 1.6 kb serA locus was excised from plasmid pRCl.55B by digestion with Smal
and ligated to the 8.4 kb plasmid pWN7.096B, which had been previously treated with
Scal and CIAP. The ligation mixture was transformed into electrocompetent E. coli
WN13. Transformants carrying the serA insert were selected on M9/glucose medium

plates. This generated the 10.2 kb plasmid pML6.133.

Plasmid pML6. 135

A 1.6 kb serA locus was excised from plasmid pRCl .55B by digestion with Smal
and ligated to the 8.4 kb plasmid pML6.128, which had been previously treated with Seal
and CIAP. The ligation mixture was transformed into electrocompetent E. coli WN13.
Transformants carrying the serA insert were selected on M9/ glucose medium plates. This

generated the 10.2 kb plasmid pML6.135.

Plasmid pML6. 1 66

E. coli ath gene was amplified using 5’-DAGGGATCCATGCAGAACATCAT-
CCGAAAA as the forward primer, 5’-AGGGATCCTTAGTGACGGAAATCAATCAC
as the reverse primer (BamHl restriction sites are underlined) using PfuTurbo as the DNA

polymerase. The resulting 1 kb PCR products was digested with BamHI, Klenow, gel

260

purified and ligated to EcoRI-digested, Klenow treated and dephosphorylated pKK223-3

to generate the plasmid pML6.166.

Plasmid pML6. I 68

E. coli adhE gene was amplified using 5’-ATGGATCCATGGCTGTTACTAA-
TGTCGC as the forward primer, 5’-CGGGATCCTTAAGCGGATTTTTTCGCTT as the
reverse primer (BamHl restriction sites are underlined) using PfuTurbo as the DNA
polymerase. The resulting 2.7 kb PCR products were digested with BamHI, Klenow, gel
purified and ligated to EcoRI-digested, Klenow treated and dephosphorylated pKK223-3

to generate plasmids pML6.168.

Plasmid pML6. 1 85

A 1.5 kb fragment carrying the toe promoter region and the ath gene was excised
from plasmid pML6.166 by digestion with BamHI and ligated to pWN5.238A, which had

been sequentially treated with BamHI and CIAP, to afford a 8.5 kb plasmid pML6.185.

Plasmid pML6. 1 95

A 1.6 kb serA locus was excised from plasmid pRCl.55B by digestion with Smal
and ligated to the 8.4 kb plasmid pML6.185, which had been previously treated with ScaI
and CIAP. The ligation mixture was transformed into electrocompetent E. coli WN13.
Transformants carrying the serA insert were selected on M9/ glucose medium plates. This

generated the 10.1 kb plasmid pML6.195.

261

Plasmid pML6. 255

A 1.3 kb fragment carrying the toe promoter region and the ath gene was
amplified by PCR using 5’-CGACGCGTCCCGTTCTGGATAATGT’ITT as the forward
primer and 5’-CTACGCGTATCCGCCAAAACAGAAGCTT as the reverse primer (AflIII
restriction sites are underlined) from plasmid pML6.166 using Platinum T aq High Fidelity
DNA polymerase. The PCR product was digested with AflIII and ligated to a 2.8 kb
plasmid pKD3, which had been sequentially treated with AflIII and CIAP, to afford the 4

kb plasmid pML6.255.

Plasmid pML6. 25 9

A 1.0 kb fragment encoding the yhdH gene was amplified by PCR from E. coli
W3110 genomic DNA using 5’-AAGGATCCCCACTTATGCAGGCGTTACT as the
forward primer and 5’-AGGGATCCGTTAGTTAACCTTCACCAGC as the reverse
primer (BamHl restriction sites as underlined) using Platinum Taq High Fidelity DNA
polymerase (Invitrogen). The resulting gene was cloned into BamHl digested pJF118EH
to afford pML6.259. Transcription of yhdH gene is in the same orientation as the Plac

promoter.

Plasmid pML6.261

A 1.2 kb fragment encoding the yiaY gene was amplified by PCR from E. coli
W3110 genomic DNA using 5’-AGGGATCCATGGCAGCTTCAACGTTCTT as the
forward primer and 5’-AGGGATCCTGATGATTACCTCGCTGCAT as the reverse

primer (BamHl restriction sites as underlined) using Platinum T aq High Fidelity DNA

262

polymerase. The resulting gene was cloned into BamHl digested pJF118EH to afford

pML6.261.

Plasmid pML6.263

A 0.8 kb fragment encoding the ydjO gene was amplified by PCR from E. coli
W3110 genomic DNA using 5’-ATGGATCCATGAGGCGGGAGATGTTTTC as the
forward primer and 5’-ATGGATCCTTCTTATGCTGGCAACGGTC as the reverse
primer (BamHl restriction sites as underlined) using Platinum Taq High Fidelity DNA
polymerase. The resulting gene was cloned into BamHl digested pJF118EH to afford

pML6.263.

Plasmid pML6. 2 72

A 0.8 kb fragment encoding the C. crescentus xylose dehydrogenase gene xdh was
excised from plasmid pWN9.068 by digestion with Sphl and ligated to the 4 kb plasmid
pML6.255, which had been previously treated with Sphl and CIAP. In the resulting 4.8 kb
plasmid pML6.272, the ath gene is transcribed by the tac promoter, while the xdh gene

was cloned alone without any promoter.

Plasmid pML 7.042

A fragment containing a phage T5 promoter, lacO and ribosome binding site was
amplified by PCR from plasmid pQE30 using 5’-GGGAATTCCATATGCGTCTTCAC-
CTCGAGAAATC as the forward primer, 5’-GGGAATTCCATATGGTGATGCG-
ATCCTCTCATAG as the reverse primer (Ndel restriction sites are underlined) using

Platinum Taq High Fidelity DNA polymerase. The resulting 0.2 kb PCR product was

263

digested with Ndel, gel purified and ligated with Ndel digested and dephosphorylated

pKD3 vector, generating the 3.0 kb plasmid pML7.042.

264

REFERENCE

1 Niu, W.; Frost, J. W. unpublished results.

2 (a) Valentin-Hansen, P.; Aiba, H.; Schumperli, D. EMBO J. 1982, 1, 317. (b) Jorgensen,
P.; Collins, J .; Valentin-Hansen, P. Mol. Gen. Genet. 1977, 155, 93.

3 Seifert, c; Bowien, s.; Gottschalk, G.; Daniel, R. Eur. J. Biochem. 2001, 268, 2369.

4 Luers, F.; Seyfried, M.; Daniel, R.; Gottschalk, G. FEMS Microbiol. Lett. 1997, 154,
337.

5 Bobik, T. A.; Xu, Y.; Jeter, R. M.; Otto, K. E.; Roth, J. R. J. Bacteriol. 1997, I 79, 6633.
6 Yanisch-Perron, C.; Vieira, J.; Messing, J. Gene 1985, 33, 103.

7 Conway, T.; Sewell, G. W.; Osman, Y. A.; Ingram, L. O. J. Bacteriol. 1987, 169, 2591.

8 Keshav, K. F .; Yomano, L. P.; An, H. J .; Ingram, L. O. J Bacteriol. 1990, 172, 2491.

9 Tabor, S.; Richardson, C. C. Proc. Natl. Acad. Sci. USA, 1987, 84, 4767.

'0 Brosius, J .; Holy, A. Proc. Natl. Acad. Sci. USA, 1984, 81 , 6929.

H F firste, J. P.; Pansegrau, W.; Frank, R.; Blocker, H.; Scholz, P.; Bagdasarian, M.; Lanka,
E. Molecular cloning of the plasmid Rp4 primase region in a multi-host-range Pm

expression vector. Gene 1986, 48, 119-131.

12Miller, J. H. Experiments in Molecular Genetics; Cold Spring Harbor Laboratory:
Plainview, NY, 1972.

13Sambrook, J.; Russell, D. W. Molecular Cloning: A Laboratory Manual; Cold Spring
Harbor Laboratory Press: Cold Spring Harbor, NY, 2001.

l4Silhavy, T. J.; Berman, M. L.; Enquist, L. W. Experiments with Gene Fusions; Cold
Spring Harbor Laboratory: Plainview, NY, 1984.

‘5 Pitcher, D. G.; Saunders, N. A.; Owen, R. J. Rapid extraction of bacterial genomic DNA
with guanidium thiocyanate. Lett. Appl. Microbiol. 1989, 8, 151-156.

'6 Miller, J. H. A Short Course in Bacterial Genetics; Cold Spring Harbor Laboratory:
Plainview, NY, 1992.

265

'7 Bradford, M. M. A rapid and sensitive method for the quantitation of microgram
quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 1976,
72, 248-254.

‘8 (a) Barbas 111, C. F.; Wang, Y.-F.; Wong, on J. Am. Chem. Soc. 1990, 112, 2013. (b)
Chen, L.; Dumas, D. P.; Wong, C.-H. J. Am. Chem. Soc. 1992, 114, 741.

'9 Toraya, T.; Krodel, E.; Mildvan, A. S.; Abeles, R. H. Biochemistry 1979, 18, 417.

20 (a) Raj, K. C.; Ingram, L. O.; Maupin-Furlow, J. A. Arch. Microbiol. 2001, 1 76, 443.
(b) Raj, K. C.; Talarico, L. A.; Ingram, L. O.; Maupin-Furlow, J. A. Appl. Environ.
Microbiol. 2002, 68, 2869. (c) Conway, T.; Osman, Y. A.; Konnan, J. 1.; Hoffmann, E.
M.; Ingram, L. O. J. Bacteriol. 1987, 169, 949.

2' Tsou, A. Y.; Ransom, S. C.; Gerlt, J. A. Buechter, D. D.; Babbitt, P. C.; Kenyon, G. L.
Biochemistry 1990, 29, 9856.

22 Keshav, K. F.; Yomano, L. 1).; An, H. J.; Ingram, L. o. J. Bacteriol. 1990, 172, 2491.

23 (a) Tong, I. T.; Liao, H. H.; Cameron, D. C. Appl. Environ. Microbiol. 1991, 5 7, 3541.
(b) Johnson, E. A.; Lin, B. C. C. J. Bacteriol. 1987, 169, 2050.

24 Harris, E. L. V.; Angal, S. In Protein Purification Methods: A Practical Approach;
Oxford University Express: Oxford, New York, Tokyo, 1989.

25 Gijsen, H. J. M.; Wong, C.-H. J. Am. Chem. Soc. 1994, 116, 8422.
2" Synder, J.; Serianni, A. s. Carbohydr. Res. 1991, 210,21.

27 Demain, A. L.; Davies, J. E. Manual of Industrial Microbiology and Biotechnology;
ASM Press: NY, 1986.

28 Siegert, P.; Pohl, M.; Kneen, M. M.; Pogozheva, I. D.; Kenyon, G. L.; McLeish, M. J.
In Thiamine: Catalytic mechanisms in normal and disease states; Jordan, F.; Patel, M. S.
Eds; Marcel Dekker, Inc.: NY, 2004; p275.

29 Conway, T.; Osman, Y. A.; Konnan, J. I.; Hoffmann, E. M.; Ingram, L. O. J. Bacteriol.
1987, 169, 949.

30(a) Stemmer, W. P. C. DNA shuffling by random fragmentation and reassembly: in vitro
recombination for molecular evolution. Proc. Natl. Acad. Sci. USA 1994, 91 , 10747-
10751. (b) Stemmer, W. P. C. Rapid evolution of a protein in vitro by DNA shuffling.

266

Nature 1994, 370, 389-391.

31(a) Zhao, H.; Arnold, F. H. Optimization of shuffling for high fidelity recombination.
Nucleic Acids Res. 1997, 25, 1307-1308. (b) Arnold, F. H.; Georgiou, G. Directed
Evolution Library Creation: Methods and Protocols; Humana Press: Totowa, NJ, 2003.
32Joem, J. M.; Meinhold, P.; Arnold, F. H. J. Mol. Biol. 2002, 316, 643.

33 Niu, W. PhD Thesis, Michigan State University, 2004.

34 Cherepanov, P. P.; Wackemagel, W. Gene 1995, 158, 9.

35 Datsenko, K. A.; Wanner, B. L. Proc. Natl. Acad. Sci. USA 2000, 97, 6640.

267

3 03062 44 6

3 129