:3.
VI ... . , .
I a“ . . 5 r
I .5...

arm

T
:v
i .

O l

A. .. i... , man. ..
. 2.3.33 .3
fit

: .(aoV: 9.3.1:

a
i = t. 5". .
. .2 i?! 9.:
‘ a

21... _
{Jinnah
in...
.. herb
.. x2.
zishtfiin
5,

z
. s3. . ‘
.yrs.....s....£.. i... u I33.
;.n.u.......... 2.3
.1”:
Examlgt

8:1.

1...: r

. I $7....
1):?85!‘

to. v1.
:ztzuu .

\u...lt..in1u\
‘ 3 \\n 15.3....
. .\ ‘liveia ?

 

:3L...

 

"vilihuv.

 

 

mm «willwill”llulllwurlllml
Michigan State
Universlty

 

 

This is to certify that the

dissertation entitled

MICROBIAL SYNTHESES OF VALUE-ADDED
CHEMICALS FROM D-GLUCOSE

presented by

Kai Li

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in Chemistry

QLWt

Major professor

 

53/23/79

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

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.

 

DA E DATE DUE DATE DUE

 

E

\

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11/00 C'JCIRC/DaIODUOJfiS-p, 1‘

 

 

MICROBIAL SYNTHESES OF VALUE-ADDED CHEMICALS FROM D-GLUCOSE
By

Kai Li

A DISSERTATION
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry
1999

ABSTRACT

MICROBIAL SYNTHESES OF VALUE-ADDED
CHEMICALS FROM D-GLUCOSE

By
Kai Li

Microbial syntheses of value-added chemicals from renewable resources are
becoming important alternatives for current petroleum-based chemical syntheses in
industry. Environmental considerations and the non-renewable character of petroleum
necessitate the development of microbial biocatalysis through a combination of metabolic
engineering and fermentation technologies. The primary tasks are to develop novel
environmentally-benign routes to target molecules, to increase the titer and yield of the
synthesized product, and to reduce production costs through alternative feedstocks. In this
thesis, the synthesis of vanillin from D-glucose,I the fed-batch fermentor synthesis of 3-
dehydroshikimic acid (DHS) using recombinant Escherichia coli.2 and the selection of
optimal carbon sources3v4 illustrate these three challenging aspects of microbial
biocatalysis.

Vanillin is the major flavor component of vanilla extract. Synthesis of vanillin from
glucose has been achieved using a microbe~cata1yzed conversion of glucose into vanillic
acid followed by an enzyme-catalyzed reduction of the vanillic acid to afford vanillin. A
genetically engineered E. coli synthesized 5.0 g/L of vanillic acid from glucose under fed-
batch fermentor conditions. .Aryl-aldehyde dehydrogenase purified from Neurospora
crassa was used. to reduce vaniliic acid to vanillin in 66% isolated yield.

DHS is a hydroaromatic intermediate in the common pathway of aromatic amino
acid biosynthesis. In addition to being a potent antioxidant, it is the most advanced
intermediate shared by aromatic amino acid biosynthesis and biocatalytic synthesis of adipic
acid, catechol, vanillin. and gallic acid. Microbial synthesis of DHS, like other

intermediates in the common pathway, has previously been examined only under shake

flask conditions. A series of E. coli biocatalysts have been constructed to evaluate the
synthesis of DHS under fed-batch fermentor conditions. The limiting factors for achieving
high DHS yields and titers were identified and alleviated. DHS titers of 69 g/L were
synthesized in 30% yield.

Carbon source selection is one of the most important considerations confronting
rnicrobe—catalyzed chemical synthesis. The phosphotransferase uptake system imposes an
upper limit on the yields and concentrations of chemicals which can be synthesized when
microbes such as E. coli employ glucose as a carbon source. The concentration and yield
of microbially synthesized DHS has been examined as a function of using D-xylose, L-
arabinose, D-glucose, or a mixture consisting of a 3/3/2 molar ratio of
glucose/xylose/arabinose as carbon sources under fed—batch fermentor conditions. The
sugar mixture is a direct simulation of corn fiber hydrolysate. In addition, butane-derived
succinic acid has been evaluated as a glucose adjunct for the synthesis of DHS to determine
how glucose might be improved upon as a source of carbon. Catabolite repression of the
uptake of both pentoses and succinic acid in the presence of glucose were successfully
circumvented under fed-batch fermentor conditions. Our discovery of higher—titer, higher-
yielding syntheses of DHS using these alternative feedstocks carries significant
ramifications relevant to the employment of corn fiber and the recruitment of butane

feedstocks in microbial synthesis of value-added chemicals.

References:
1. Li, K.; Frost, J. W. J. Am. Chem. Soc. 1998, 120, 10545.

2. Li, K.; Mikola, M. R.; Draths, K. M.; Worden, R. M.; Frost, J. W. Biotechnol.
Bioeng. 1999, 64, 61.

3. Li, K.; Frost, J. W. Biotechnol. Prog. 1999, In press.
4. Li, K.; Frost, J. W. J. Am. Chem. Soc. 1999, In press.

Copyright by
Kai Li
1 999

To my parents and my wife

For their love and support

ACKNOWLEDGMENTS

First of all, I would like to express my great gratitude to Professor John W. Frost
for his patience, encouragement and guidance throughout the course of my PhD. studies.
His enthusiasm, even in the face of difficult problems, together with his aggressive
commitment to "getting the job done" has impressed me (forever) the way cutting-edge
science must be done. In addition, I want to thank the members of my graduate committee,
Professor Chi-Kwong Chang, Professor Kris A. Berglund, and Professor Gary J.
Blanchard for their intellectual input during my preparation of this thesis. I am grateful to
Professor R. Mark Worden for his help on my research as well.

I am greatly indebted in Dr. Karen Draths for her patient instruction in experimental
techniques and intelligent input to my research. As a close mentor of mine, her
encouragement and friendship helped me make it through my tough graduate studies.
Meanwhile I am appreciative of current and former group members: Dr. Jean-Luc
Montchamp, Dr. Jirong Peng, Dr. David Spear, Dr. Marie Migaud, Dr. John Arthur, Dr.
Amy Dean, Dr. Feng Tian, Michael Farabough, Spiros Kambourakis, Mark Mikola, Sunil
Chandran, Jessica Barker, Dave Knop, Chad Hansen, Padmesh Venkitasubramanian, Jian
Yi, Wei Niu, Ningqing Ran, Jiantao Guo, and Satyamaheshwar Peddibhotla for their
assistance and friendship.

The love and support from my parents, my sister and brother-in-law were sincerely
appreciated. Finally, I would like to thank my wife, Hongyan Liang, for her love and
understanding. Without them, my adventure in science would not have been so exciting

and rewarding.

vi

TABLE OF CONTENTS

LIST OF TABLES ................................................................................ x
LIST OF FIGURES .............................................................................. xii
LIST OF ABBREVIATIONS .................................................................... xvi
CHAPTER 1
INTRODUCTION ................................................................................ 1
The Shikimate Pathway .................................................................. 2
Microbial Syntheses of Value-Added Chemicals Utilizing the Shikirnate
Pathway ................................................................................... 5
Metabolic Engineering of E. coli to Increase Product Titers and Yields ........... 14
CHAPTER 2
SYNTHESIS OF VAN IL] IN FROM GLUCOSE ............................................ 22
Background ............................................................................... 22
Microbial Synthesis of Vanillic Acid from Glucose .................................. 27
A.The Construction of the Host Strain ............ . ........................... 27
B. Plasmid Constructions ...................................................... 33
C. Fed-Batch Fermentor Synthesis of Vanillic Acid ........................ 40
Enzymatic Reduction of Vanillic Acid to Vanillin .................................... 45
A. Purification of Aryl-Aldehyde Dehydrogenase from Neurospora
crassa ............................................................................. 45
B. Reduction of Vanillic Acid .................................................. 48
Discussion ................................................................................. 66
Future Work .............................................................................. 67
CHAPTER 3
FED-BATCH FERMENTOR SYNTHESIS OF 3-DEHYDROSHIKIMIC ACID
USING RECOMBINANT ESCHERICHIA COLI ........................................... 70
Introduction ............................................................................... 70
Biocatalysts and Fed—Batch Fermentor Conditions for DHS Synthesis ........... 73
A. Shared Genomic and Plasmid Elements .................................. 73
B. Fed-Batch Fermentor Condition ........................................... 77
Carbon Flow Directed into the Common Pathway as a Function of DAHP
synthase Expression ..................................................................... 80
Overexpression of Transketolase under Fed- Batch Fermentor Conditions ........ 92
Discussion ................................................................................ 101
A. Comparison of Titers and Yields ................................. . ........ 101
B. DHQ and Gallic Acid Formation ........................................... 104
C. B4P Limitation versus PEP Limitation .................................... 106

vii

CHAPTER 4
CARBON SOURCE SELECTION: IS D-GLUCOSE THE BEST CARBON

SOURCE FOR MICROBIAL SYNTHESIS? ................................................. 115
Introduction ............................................................................... 1 15
Fermentation Conditions ................................................................ 117
A Comparative Analysis of D-Xylose, L-Arabinose, and D-Glucose Carbon
Sources for Microbial Synthesis of DHS .............................................. 118
A. Background .................................................................. 1 18
B. Biocatalyst .................................................................... 120
C. Glucose Fermentation Versus Xylose or Arabinose
Fermentation. .................................................................... 121
D. Glucose, Xylose and Arabinose as a Mixed Carbon Source ........... 124
E. The Impact of Transketolase ................................................ 129
F. Discussion ................................................................... 132

Utilizing Succinic Acid as a Glucose Adjunct 1n Fed- Batch Fermentation. Is

Butane a Feedstock Option 1n Microbial- -Catalyzed Synthesis? ..................... 135
A. Introduction ................ . ................................................ 135
B. Biocatalyst Construction .................................................... 137
C. Channeling Carbon from Succinic Acid into DHS Synthesis .......... 142
D. Replacing Succinic Acid with Other TCA Cycle Intermediates ........ 146
E. Discussion .................................................................... 148

CHAPTER 5

EXPERIMENTAL ................................................................................ 150
General Methods ...................................... . .................................. 150

Chromatography ......... . ....................................................... 150
Spectroscopic Measurements .................................................. 150
Bacterial Strains ............................................. . ................... 150
Storage of Bacterial Strains and Plasmids .................................... 151
Culture Medium .................................................... . ............ 151
Conditions for Shake-Flask Cultivation ...................................... 152
General Fed-Batch Fermentor Conditions .................................... 153
1H NMR Analysis of Culture Supernatant ....... . ........................... 154
Genetic Manipulations .......................................................... 155
General ................................................................. 155
Large Scale Purification of Plasmid DNA .......................... 155
Small Scale Purification of Plasmid DNA ............ . ............. 157
Restriction Enzyme Digestion of DNA ............................. 158
Agarose Gel Electrophoresis ......................................... 158
Isolation of DNA from Agarose .................................... 159
Treatment of Vector DNA with Calf Intestinal Alkaline
Phosphatase ............................................................ 159
Treatment of DNA with Klenow fragment ......................... 159
Ligation of DNA ....................................................... 160
Preparation and Transformation of Competent Cells .............. 160
Purification of Genomic DNA ....................................... 161
Enzyme Assay ................................................................... 162
DAHP Synthase ....................................................... 163
DHQ Synthase ......................................................... 164
DHS Dehydratase ..................................................... 164
Catechol-O-Methyltransferase ....................................... 165
Aryl-Aldehyde Dehydrogenase ..................................... 166

viii

PEP Carboxykinase ................................................... 166

Chapter 2 .................................................................................. 167
Purification of Neurospora crassa Aryl-aldehyde Dehydrogenase ........ 167
Buffer ................................................................... 167
Neurospora crassa Cultivation ....................................... 167
Purification of Aryl-Aldehyde Dehydrogenase .................... 168

Strain Constructions ............................................................ 168
KL7 ..................................................................... 168
Plasmid pKL3.272A .................................................. 170
Plasmid pKL5.26A ................................................... 170
Plasmid pKL5.97A ................................................... 170
Biocatalytic Synthesis of Vanillic Acid ....................................... 171
Reduction of Vanillic Acid to Vanillin ........................................ 171
Chapter 3 .................................................................................. 173
Strains Constructions ........................................................... 173
KL3 ..................................................................... 173

Strain A3224 ......................................................... 173

Strain AC2- 13A ....................................................... 174
Plasmid pKL4.2OB ................................................... 174
Plasmid pKL4.33B ................................................... 175
Plasmid pKL4.66A ................................................... 175
Plasmid pKL4.l30B .................................................. 175
Plasmid pKL4.71A ................................................... 175
Plasmid pKL4.79B .................................................. 176
Plasmid pKL4. 124A .................................................. 176
Plasmid pKDl 1.291A ................................................ 176
Plasmid pKLS. 17A ................................................... 176
Fed-Batch Fermentation ........................................................ 177
Chapter 4 .................................................................................. 178
Strain Constructions ............................................................ 178
Plasmid pKL2.222 .................................................... 178
Plasmid pKL6. 198A .................................................. 178
Plasmid pKL6.218A .................................................. 178
Fermentation Conditions ....................................................... 179

Conditions for the Study of "A Comparative Analysis of D-Xylose,
L-Arabinose, and D-Glucose Carbon Sources for Microbial
Synthesis of DHS" .................................................... 179
Conditions for the Study of "Utilizing Succinic Acid as a Glucose
Adjunct in Fed-Batch Fermentation: Is Butane a Feedstock Option
in Microbial-Catalyzed Synthesis?" ................................. 180

BIBLIOGRAPHY ................................................................................. 182

11

1.1

LIST OF TABLES

Table 1. Confirmation of AroB and AroZ activities in pKL4.237A ........................ 30

Table 2. Plate selection for characterization of genomic insertion strain KL7 ............. 31

Table 3. Products formed after 48 h under fed- batch fermentor conditions as a
function of catechol- -0-methyltransferase activity and L-methionine

supplementation ................................................................................... 40
Table 4. Purification of aryl-aldehyde dehydrogenase from N. crassa ..................... 46
Table 5. The process of reducing vanillic acid to vanillin .................................... 47
Table 6. Plate selection for characterization of genome insertion strain KL3 .............. 76
Table 7. Product titers and yield after 48 h fermentation for KL3/pKL4.33B ............ 82

Table 8a. DAHP synthase activities when aroFFBR expression is controlled by
Ptac. ................................................................................................. 83

Table 8b. Product titers and yields when aroFFBR expression is controlled by Pmc ...... 84

Table 9. Product titers and yield synthesized by KL3/pKL4.66A after 48 h .............. 90
Table 10. Product titers and yield synthesized by KL3/pKDl 1.291A after 48 h ......... 90
Table 11a. DAHP synthase activities in TktA-overexpression constructs ................. 96
Table 11b. Product titers and yield in TktA-overexpression constructs .................... 97
Table 12. KL3/pKL4.79B synthesis of DHS with different carbon source ............... 124
Table 13. KL3/pKL4.124A synthesis of DHS with different carbon source ............. 130

Table 14. Consumption of the individual components of the mixed carbon source ...... 131

Table 15. Product titers and yields after 48 h under fed-batch fermentor condition
as a function of plasmids, carbon sources and PEP carboxykinase activities .............. 143

Table 16. KL3/pKL6.218A synthesis of DHS with different TCA cycle intermediate
supplementation ................................................................................... 147

LIST OF FIGURES

Figure l. The common pathway of aromatic amino acid biosynthesis ..................... 3

Figure 2. Chemicals synthesized from tryptophan, tyrosine, and phenylalanine ......... 5

Figure 3. Comparison of microbial and chemical synthesis of p-aminobenzoic acid
(PABA) ............................................................................................. 6

Figure 4. Comparison of p-hydroxybenzoic acid (PI-TB) synthesis from glucose
versus potassuim phenoxide ..................................................................... 7

Figure 5. Comparison of benzoquinone and hydroquinone synthesis from glucose
and benzene ........................................................................................ 9

Figure 6. Value added-chemicals synthesized from glucose via DHS intermediacy ...... 10

Figure 7. Synthesis of gallic acid from glucose ............................................... 11
Figure 8. Comparison of microbial and chemical synthesis of catechol ................... 12
Figure 9. Comparison of microbial and chemical synthesis of adipic acid ............... 13

Figure 10. Theoretical flux distribution for directing carbon into common pathway

with glucose as the carbon source ............................................................... 16
Figure 11. Reactions catalyzed by transketolase and transaldolase ......................... 18
Figure 12. Substrates and products associated with PEP synthase (Pps) activity ........ 19
Figure 13. The structure of vanillin ............................................................ 22
Figure 14. Biosynthesis of natural vanillin .................................................... 23
Figure 15. Chemical synthesis of vanillin ..................................................... 24
Figure 16. Biosynthesis of vanillin from glucose ............................................ 25

xii

F.qv'
.‘.-
O

EV”
Flku
'

FhVU
”Lu
V

FVT‘
8.9

F1T1
I H

(yr

'I'J
IV

I ‘I"

Figure 17. Preparation of Plasmid pKL4.237A .............................................. 29

Figure 18. Preparation of Plasmid pKL4.276B .............................................. 32
Figure 19. Preparation of Plasmid pKL4.20B ................................................ 35
Figure 20. Preparation of Plasmid pKL4.33B .............. . ................................. 36
Figure 21. Preparation of Plasmid pKL3.272A .............................................. 37
Figure 22. Preparation of Plasmid pKL5.26A ................................................ 38
Figure 23. Preparation of Plasmid pKL5.97A ................................................ 39

Figure 24. Fermentation culture of KL7/pKL5.26A without methionine
supplementation ................................................................................... 41

Figure 25. Fermentation culture of KL7/pKL5.97A without methionine
supplementation .. ................................................................................. 42

Figure 26. Fermentation cultures of (a) KL7/pKL5.26A and (b) KL7/pKL5.97A

supplemented with methionine ................................................................... 44
Figure 27. Mechanism of aryl-aldehyde dehydrogenase reduction of aromatic acid ..... 45
Figure 28. Control of 1H NMR of vanillic acid ........ . ...................................... 50
Figure 29. Control of 1H NMR of isovanillic acid ........ . .................................. 52
Figure 30. Control of 1H NMR of protocatechuic acid ...................................... 54
Figure 31. Control of 1H NMR of 3-dehydroshikimic acid ................................. 56
Figure 32. Control of 1H NMR of vanillin... .. .............................................. 58

xiii

Figure 33. Control of 1H NMR of isovanillin ................................................ 60

Figure 34. 1H NMR after EtOAc extraction of the KL3/pKL5.97A fermentation
broth (with methionine supplementation) ....................................................... 62

Figure 35. 1H NMR after CH2C12 extraction of the reduction reaction .................... 64

Figure 36. DHS as the most advanced intermediate shared by aromatic amino acid

biosynthesis and biocatalytic synthesis of adipic acid and catechol ......................... 72
Figure 37. Preparation of Plasmid pKL3.82A ................................................ 74
Figure 38. Serine biosynthesis .................................................................. 76
Figure 39. B. Braun fermentor ................................................................. 78

Figure 40. KL3/pKL4.33B synthesis of DHS under fed-batch fermentor

conditions ........ . ................................................................................. 81
Figure 41. Preparation of Plasmid pKL4.71A ................................................ 85
Figure 42. Preparation of Plasmid pKL4.79B ................................................ 86
Figure 43. Preparation of Plasmid pKL4.66A ................................................ 87
Figure 44. Preparation of Plasmid pKDl 1.291A ............................................. 88
Figure 45. Preparation of Plasmid pKL4.124A .............................................. 93
Figure 46. Preparation of Plasmid pKL4.13OB .............................................. 94
Figure 47. Preparation of Plasmid pKL5. 17A ................................................ 95
Figure 48. KL3/pKL4.66A fermentation ...................................................... 99
Figure 49. KL3/pKL4.13OB fermentation ......... . .......................................... 100

xiv

«k.»

I t

~uh.

PM.“

-\

.0?“

u

V.hb
PT.

. u
s\h»

r1

Figure 50. Control of 1H NMR of gallic acid ................................................. 109

Figure 51. Control of 1H NMR of 3-dehydroquinic acid .................................... 111
Figure 52. 1H NMR of a typical DHS fermentation after 48 h .............................. 113
Figure 53. Synthesis of DHS by KL3/pKL4.79B ........................................... 123

Figure 54. KL3/pKL4.79B cultivation prior to initiation of oxygen sensor-
controlled addition of the glucose : xylose : arabinose mixture .............................. 126

Figure 55. KL3/pKL4.79B in fed-batch fermentation using a mixture of glucose,
xylose, and arabinose at a ratio of 3 : 3 : 2 ..................................................... 127

Figure 56. KL3/pKL4.124A in fed-batch fermentation using a mixture of glucose,

xylose, and arabinose at a ratio of 3 r 3 : 2 ............ . ........................................ 128
Figure 57. Manufacture of succinic acid from butane ........................................ 135
Figure 58. Aromatic amino acid biosynthesis and TCA cycle... ............................ 136
Figure 59. Preparation of Plasmid pKL2.222 ................................................ 139
Figure 60. Preparation of Plasmid pKL6.198A .............................................. 140
Figure 61. Preparation of Plasmid pKL6.218A ..... .. ........................................ 141
Figure 62. Products and cells produced during fed-batch fermentation .................... 145

Figure 63. KL3/pKL6.218A cultivation prior to initiation of oxygen sensor-
controlled addition of the glucose/succinic acid mixture ...................................... 146

XV

2, 3-DHB
DHQ
DHS
L—DOPA

E4P

LIST OF ABBREVIATIONS

4-amino-4-deoxychorismic acid
adenosine monophosphate

adenosine triphosphate

arnpicillin

base pair

chorismic acid

cyclic adenosine monophosphate
complimentary DNA

flavin adenine dinucleotide, reduced form
calf intestinal alkaline phosphatase
chloramphenicol
catechol-O-methyltransferase
3-deoxy-D-arabin0-heptulosonic acid
3-deoxy-D-arabino-heptulosonic acid 7-phosphate
digital control unit

diethylaminoethyl
2,3-dihydroxybenzoic acid
3-dehydroquinic acid
3-dehydroshikimic acid
L-3,4-dihydroxyphenylalanine
dissolved oxygen

dithiothreitol

D-erythrose 4—phosphate

xvi

EPSP

FB S
GA

HPLC

Kan
kb

MOPS
mRN A
NADP
NADPH

ORF
PABA
PCA
PEG
PEP
PHB

S-enolpyruvoylshikimate 3-phosphate
feedback resistant

feedback sensitive

gallic acid

hour

high pressure liquid chromatography
isopropyl B-D-thiogalactopyranoside
kanamycin

kilobase

Michaelis constant

molar

minute

milliliter

millimolar

4-morpholinepropanesulfonic acid
messenger RNA

nicotinarnide adenine dinucleotide phosphate, oxidized form
nicotinamide adenine dinucleotide phosphate, reduced form
nuclear magnetic resonance spectroscopy
optical density

open reading frame

p-aminobenzoic acid

protocatechuic acid

polyethylene glycol

phosphoenolpyruvic acid

p-liydroxybenzoic acid

xvii

PID
Pck
PCR
Phe
PMSF

Pps
PTS
QA

rpm
SA
SAM
SDS
S3P

TBA
Tc
TCA
Trp
TSP

UDP

Vmax

proportional-integral-derivative
PEP carboxykinase
polymerase chain reaction
L-phenylalanine
phenylmethylsufonyl fluride
PEP carboxylase

PEP synthase
phosphotransferase system
quinic acid

ribosome binding site
rotations per minute
shikimic acid
S-adenosylmethionine
sodium dodecyl sulfate
shikimate 3-phosphate
spectinomycin

thiobarbituric acid
tetracycline

tn'carboxylic acid
L-tryptophan

sodium 3-(trimethylsilyl)propionate-2,2,3,3-d4
L-tyrosine

uridine diphosphate
ultraviolet

maximal velocity

xviii

CHAPTER 1

INTRODUCTION

Chemistry is moving into an era in which renewable resources and starting
materials such as D-glucose, D-xylose, and L-arabinose will likely be prominent features of
industrial chemical manufacture. Current chemical manufacture typically relies on abiotic,
chemical catalysts and on starting materials derived from petroleum. As a nonrenewable
natural resource, petroleum has several negative environmental and geopolitical problems
associated with its use. For example, aromatics are currently derived predominantly from
the benzene, toluene, xylene (BTX) fraction of petroleum refining. With annual production
levels at 5.4 x 109 kg in the United States, benzene is the most important component of the
BTX fraction.1 Benzene is a potent carcinogen,2 and it is also included by the
Environmental Protection Agency on the list of chemicals covered by the Chemical
Manufacturing Rule that requires drastic reductions in emissions of hazardous organic air
pollutants.3

Beyond the health costs associated with benzene, the costs of deriving this starting
material from nonrenewable fossil fuel feedstocks are important to consider. Oil spills and
land reclamation along with geopolitical complications substantially add to the true cost of
aromatic manufacture from fossil fuel-derived BTX. By contrast, plant-derived starch,
cellulose, and hemicellulose are abundant, renewable sources of glucose, xylose, and
arabinose. The comparatively low temperatures, near-atmospheric pressures, and use of
water as reaction solvent, which characterize microbial biocatalysis, are environmentally
benign. In addition, there is a net consumption of carbon dioxide from the environment
when glucose, xylose, and arabinose are employed as starting materials for chemical
synthesis. Add to this the avoidance of toxic intermediates, reagents, and byproducts, and

the ability to synthesize a chemical by microbial biocatalysis from renewable feedstocks and

nontoxic starting materials presents itself as an appealing alternative to traditional chemical
manufacture.4

In Chapter 2 of this thesis, a novel route to synthesize vanillin biocatalytically is
described, which serves as a good example for the comparison of biocatalysis with
traditional chemical manufacture. The main portion of the route was developed through
metabolic engineering of the common pathway of aromatic amino acid biosynthesis (the
shikimate pathway) in Escherichia coli. Chapter 3 and 4 of this thesis focus on increasing
the carbon flow directed into the shikimate pathway to improve titers and yields of
biosynthesized chemicals. 3-Dehydroshikimic acid (DHS) was used as a model compound
because it is the most advanced intermediate shared by both biosynthesis of aromatic amino

acids and biocatalytic syntheses of a variety of value-added chemicals.

Wm

The shikimate pathway, also referred to as the common pathway of aromatic amino
acid biosynthesis, has been the subject of considerable study due to its role in the
transformation of simple carbohydrate precursors into aromatics in plants, bacteria, fungi,
and molds.5 It consists of seven enzyme-catalyzed reactions converting
phosphoenolpyruvic acid (PEP) and erythrose 4-phosphate (E4P) into chorismic acid
(Figure 1). Three terminal pathways then lead from chorismic acid to L-tryptophan, L-
tyrosine, and L-phenylalanine. In addition, biosynthetic pathways leading to ubiquinone,
folic acid and enterochelin also branch away from the common pathway at chorismic acid
(Figure 1).5 Folic acid-derived coenzymes are frequently involved in the biosynthetic
transfer of one carbon fragments; ubiquinones are involved in electron transport; and

enterochelin is an iron chelator responsible for iron uptake in numerous microorganisms.

"h. .
JV :g-EC

foléc ac

 

 

 

 

H203PO

/ COgH H3PO4 HO__. COZH H3PO4 HO__. COZH H20
PEP 4;, )j _j_, {j _2,
OH O aroF - 0H 3mg 0 - 0“ aroD
H203POMH aroG H203P0 6H 0H
6H 8’0” DAHP DHQ
E4P
C02H NADPH NADP COgH ATP ADP COgH PEP H3PO4
0 : 0” aroE H0". : OH aroK H203PO'" : 0“ aroA
OH OH aroL OH
DHS shikimic acid SSP
002H H3P04 COgH 002 H
i) 4G <1“
—-> L-t to ban
H203 PO"' o’chozH aroC oJLcozH ”p p
EPSP chrrismic acid anthranilic acid
COzH C02 gH HO... COZH
ubiquinone ‘— G: —> L-phenylalanine
folic acid OH O ’ L-tyrosine
X 2.3-dihydroxy- OH
X = OH PHB benzoic acid prephenic acid
X=NH2 PABA ii

enterochelin

Figure l. The common pathway of aromatic amino acid
biosynthesis. Genetic loci are as follows: aroF aroG aroH, DAHP
synthase; aroB, DHQ synthase; aroD, DHQ dehydratase; aroE, shikimate
dehydrogenase; aroL aroK, shikimate kinase; aroA, EPSP synthase; aroC,
chorismate synthase.

EX

t.)
u-y—J

f.

511

:0:

n .
I

be;

Individual pathway enzymes have received attention due to the novel mechanisms
employed during turnover of substrate into product and as targets for inhibition. The
existence of the shikimate pathway in plants and bacteria but not in humans provides an
appealing mode of action for enzyme-targeted herbicides and antibiotics.

The substrates and products of the seven enzymatic reactions that convert PEP and
E4P into chorismic acid (Figure 1) were identified by the early 19605 from studies of
bacterial auxotrophs of Escherichia coli and Klebsiella aerogenes. The first committed step
of aromatic amino acid biosynthesis involves the condensation of PEP and E4P to form 3-
deoxy-D-arabino-heptulosonic acid 7-phosphate (DAHP) and is catalyzed by DAHP
synthase.5 Three isozymes of DAHP synthase exist in E. coli, each of which is sensitive
to feedback inhibition by one of the three aromatic amino acids. The genes aroF, aroG and
aroH encode for tyrosine-sensitive, phenylalanine-sensitive, and tryptophan-sensitive
isozymes of DAHP synthase, respectively. DAHP is converted into 3-dehydroquinic acid
(DHQ) by DHQ synthase which is encoded by aroB6 in a complex reaction where active
site NAD is used as a catalytic active site residue.6 A syn elimination of water from DHQ
affords 3-dehydroshikimic acid (DHS).7 This reaction is catalyzed by DHQ dehydratase,
which is encoded by aroD. Reduction of DHS to shikimic acid in the presence of NADPH
is catalyzed by aroE-encoded shikimate dehydrogenase.8 Shikimic acid is further
converted to shikimate 3-phosphate by phosphoryl group transfer from ATP. This reaction
is catalyzed by two isozymes of shikimate kinase encoded by the loci ar0L9 and aroK'.10 5-
Enolpyruvoylshikimate 3-phosphate (EPSP) synthase, encoded by aroA,ll catalyzes the
reversible condensation of shikimic 3-phosphate and PEP. Product EPSP is formed along
with inorganic phosphate. The last enzyme of the common pathway is chorismate
synthase.12 Encoded by aroC, it catalyzes the elimination of inorganic phosphate from

EPSP to afford chorismic acid.

u- H..- 11.1‘ ‘ 0 V1- wan). ‘ 'al 'lizii- h 1,_I_1L- ‘ ,'- hwa

Aromatic amino acids such as phenylalanine and tryptophan figure prominently
among the chemicals being microbially manufactured from glucose.13 A variety of other
chemicals are then synthesized from these aromatic amino acids. For example,
phenylalanine can be enzymatically or chemically converted to aspartame (Figure 2),14
which has the largest sales volume of all food additives.15 Introduction of naphthalene
dioxygenase into a tryptophan-synthesizing microbe that also expresses tryptophanase
results in biocatalytic synthesis of indigo (Figure 2),16 the vat dye that gives blue jeans
their faded-blue coloration. Indigo is the highest volume dye produced worldwide.
Tyrosine can be converted into melanin (Figure 2),17 the mammalian pigmentation that is

responsible for hair coloring as well as protection from solar irradiation.

Trp —__’_,.

Tyr ——"’

 

 

 

melanin

CH3020 o

KIH
3
Phe —’, HJKC

CO;
aspartame

Figure 2. Chemicals synthesized from tryptophan, tyrosine,
and phenylalanine.

Some common pathway intermediates are important value-added chemicals. For
example, shikimic acid is a valuable chiral synthon used in the synthesis of neuraminidase
inhibitors effective in the treatment of influenza infection.18 Because the current isolation
of shikimic acid from the fruit of Illicium plants19 precludes its use in kilogram-level
synthesis, an E. coli biocatalyst has been constructed which synthesizes 50 g/L of shikimic
acid from glucose under fed-batch fermentor conditions. This shikimate-synthesizing
biocatalyst resulted from disruption of the genomic aroL and aroK loci in E. coli and
overexpression of feedback-insensitive, aroFFBR—encoded DAHP synthase to channel more

carbon into the common pathway.20

 

 

C02H PabA C02H
o .IIOH see Fig1 o PabB o
——->’ .U. ——’ JL _
s OH 5 O COZH a 0 CO2
D-glucose chorismic acid ADC
CH3 CH3 COZH N“2
<3 HNOa, sto4 HNO3 Fe, m PABA
9 ———>
30-45 °C
toluene p-nitrotoluene p-nitrobenzoic acid

Figure 3. Comparison of microbial and chemical synthesis of p-
aminobenzoic acid (PABA).

Intermediates in the common pathway of aromatic amino acid biosynthesis can also
be transformed into valuable chemicals either through chemical conversion or
bioconversion. For instance, p-aminobenzoic acid (PABA) is biosynthesized by initial
conversion of chorismic acid to 4-amino-4-deoxychorismic acid (ADC) followed by
elimination of pyruvic acid (Figure 3). ADC synthase consists of two subunits encoded by
the pabA and pabB loci. ADC lyase is encoded by pabC. Each of the E. coli genes

associated with the conversion of chorismic acid into PABA has been cloned and

an;

I03

‘.
in

q

‘illl

 

sequenced.21 Although its biological role is as an intermediate in the biosynthesis of folic
acid, PABA's chemical use is as an ingredient in UV-blocking formulations, and it can be
esterified to form the local anesthetic known as benzocaine.22

PABA is currently industrially synthesized from toluene (Figure 3). The first step
entails reaction of toluene with a 20/60 (v/v) of nitric acid/sulfuric acid at 30—45 °C.22 p-
Nitrotoluene is then separated from the other nitrotoluene isomers by sequential distillation
and crystallization. Oxidation of the methyl group of p-nitrotoluene with nitric or chromic
acids is followed by reduction to afford PABA. Almost every step of this manufacturing
route has to contend with a health/safety hazard. These range from strong acid solutions to

manipulation of toxic, flammable toluene and highly toxic p-nitrotoluene.23

 

OH COZH COZH
0H see Fig 1 see Fig 1
0 II 0
(:1 —*’ "a —>’ (1 JL
.-. OH HO ; 0H ; O COZH
Ho OH OH OH
D-glucose shikimic aicd chorismic acid
H\ ;/UbiC
002K COZH
20 atm CO2 H+
180-250 °C
OK OH OH
potassium potassium PHB
phenoxide p-hydroxybenzoate

Figure 4. Comparison of p-hydroxybenzoic acid (PHB)
synthesis from glucose versus potassuim phenoxide.

Chorismic acid can also be converted directly into p-hydroxybenzoic acid in a
reaction catalyzed by ubiC—encoded24 chorismate-pyruvate lyase (Figure 4). An E. coli
biocatalyst has been constructed consisting of a genome-localized aroAaroLaroBaroCkanR
cassette and plasmid-localized aroFFBR, tktA, and ubiC. The genes encoding other

chorismate-utilizing enzymes are disrupted to prevent biocatalytic conversion of chorismic

acid into anthranilic acid and prephenic acid (Figure 1). The biocatalyst can synthesize 12
g/L of p-hydroxybenzoic acid from glucose under fed-batch fermentor conditions.25 p-
Hydroxybenzoic acid can also be synthesized by means of chemical dehydration of
shikimic acid. This reaction is catalyzed by 1 M sulfuric acid in refluxing acetic acid
(Figure 4).25 p-Hydroxybenzoic acid is a component of liquid crystal polymers such as
Xydar,26 which have attracted considerable attention because of their use in high-
performance applications. Esters of p-hydroxybenzoic acid are also widely used as food
preservatives.”

p-Hydroxybenzoic acid is currently manufactured by Kolbe-Schmitt reaction of
dried potassium phenoxide with 20 atm dry carbon dioxide at 180-250 °C (Figure 4).27
Product p-hydroxybenzoate potassium salt is converted to its free acid upon addition of
mineral acid. Besides the required temperatures and pressures, p-hydroxybenzoic acid
manufacture has to contend with handling of phenol which is listed as a highly toxic,
corrosive chemical.23

Quinic acid is another useful molecule that can be microbially synthesized through
the utilization of the shikimate pathway (Figure 5).28 Quinic acid is a widely used chiral
starting material in multi-step chemical synthesis.29 It is currently obtained by an
expensive isolation from plant sources. An E. coli biocatalyst has been constructed via
mutational inactivation of aroD-encoded DHQ dehydratase, overexpression of aroFFBR—
encoded feedback-insensitive DAHP synthase and overexpression of aroE-encoded
shikimate dehydrogenase. The biocatalyst can synthesize up to 80 g/L of quinic acid.3O
Quinic acid can be converted into either benzoquinone or hydroquinone through simple
chemical oxidation (Figure 5).28 Hydroquinone's selective reduction of photoactivated
silver ion is the basis for this organic's widespread use in photography. Benzoquinone is

an important chemical precursor in the manufacture of various chemicals.31

 

NH2
H2304

benzene aniline

00°

 

 

MnOz, H+ benzoquinone
OH OH OH

 

; OH HO" HO" \
HO OH
OH .QH . MnO2
D-glucose DHQ qumIc ac1d
y hydroquinone
/\\ NaOH
O —->
benzene p-diisopropyl
benzene

Figure 5. Comparison of benzoquinone and hydroquinone synthesis from
glucose and benzene.

Hydroquinone is produced globally at volumes of 4.5 - 5.0 x 107 kg/year.32 The
dominant route to manufacture hydroquinone is through hydroperoxidative synthesis.32
The p-diisopropylbenzene is synthesized by zeolite-catalyzed Friedel-Crafts reaction of
benzene or cumene with propylene or isopropanol. Air oxidation of the p-
diisopropylbenzene proceeds at 90 - 100 °C in an aqueous NaOH solution also containing
organic bases along with cobalt or copper salts. Hydroperoxycarbinol and dicarbinol are
produced along with the dihydroperoxide during air oxidation. Treatment with acid and
H202 converts the hydroperoxycarbinol and dicarbonyl to the dihydroperoxide which is
cleaved to form acetone and hydroquinone. During the acidic cleavage, explosive organic
peroxide can form which presents a safety hazard.32 Benzoquinone is primarily
manufactured from oxidation of aniline (Figure 5). The oxidant employed is Mn02 in
aqueous H2804 (Figure 5). Benzoquinone can also be reduced by Fe or hydrogenated to

afford hydroquinone. Although accounting for approximately 10% of hydroquinone

ilgUr
illltrn

production,” this manufacturing route generates large quantities of MnSO4, (NH4)2SO4,

and iron oxide salts.”

o H r; 51
W «MN; 0 0 NHCH3
u n +0

 

   

0
nylon 6,6 bendiocarb
HOZC
+ H N N NH
NH2 8 HOQOH 2 . \‘r’ 2
CO; COZH OH IN
adipic acid pyrogallol
HO T T CH30 OCH3
OH OCH3
L-DOPA COZH trimethoprim
HO‘ : HO‘ : “OH
OH OH
catechol gallic acid
0 H COZH O O\/\
CH30 O’KD‘OH HO OH
OH OH OH
vanillin propyl gallate

 

OH

= OH
HO OH
D-glucose

Figure 6. Value added-chemicals synthesized from glucose via DHS
intermediacy.

10

3-Dehydroshikimic acid (DHS) is another important common pathway intermediate.
It can be converted both biocatalytically and chemically into a variety of industrially useful
molecules (Figure 6).33 DHS itself is a potent antioxidant.“ Chemical oxidation of DHS
can afford gallic acid.34 DHS can also be converted to gallic acid biocatalytically via
protocatechuic acid (PCA) intermediary (Figure 7). DHS dehydratase, which is encoded
by the aroZ locus isolated from Klebsiella pneumoniae, converts DHS to PCA.
Hydroxylation of PCA catalyzed by a mutant isozyme of p-hydroxybenzoate hydroxylase
isolated from Pseudomonasfluorescens then affords gallic acid. An E. coli biocatalyst has
been constructed which synthesizes approximately 15 g/L of gallic acid from glucose under
fed-batch fermentation conditions.35 Derivatives of gallic acid include: propyl gallate, an
important food-grade antioxidant; trimethoprim, an antibacterial agent; and pyrogallol, a
product of chemical or enzymatic decarboxylation of gallic acid (Figure 6). Pyrogallol is
used in photographic developing solutions and is a key building block in the manufacture

of the insecticide bendiocarb (Figure 6).36 Gallic acid is currently isolated from gall nuts

 

   
   

and tara powder.
OH COzH COZH
O ,.OH see Fig1 Aroz
-T>_, ——>
; OH O ; OH HO
HO OH OH OH
D-glucose DHS PCA
@244sz A/PobA
COZH
HO OH
OH
gallic acid

 

Figure 7. Synthesis of gallic acid from glucose.

11

Catechol is also microbially synthesized from DHS (Figure 8).37 A genetically
modified E. coli strain mediates the dehydration of DHS to form protocatechuic acid
(Figure 8), which then undergoes an enzyme-catalyzed, non-oxidative decarboxylation to
catechol. As in biocatalytic synthesis of gallic acid, dehydration of DHS to protocatechuic
acid is catalyzed by aroZ-encoded DHS dehydratase. Conversion of protocatechuic acid
into catechol is catalde by aroY-encoded PCA decarboxylase, which was isolated from
K. pneumoniae. Although DHS and protocatechuic acid are intermediates in the

conversion of glucose into catechol, only catechol accumulates in the culture supernatant.

acetone
hydroquinone

©7Y© JIOOQ

 

 

benzene cumene phenol Q
HO
OH
OH 002 H C02” AroY catechol
flOH—JZ—p see Fig 10?: AroZ :H
= OH ——>HO
HO OH
D—glucose DHS PCA

a = propylene, solid H3PO4 catalyst, ZOO-260°C, 400-600 psi
b = 02, 80-130°C then 302, 60-100°C
C = 700/0 H202, EDTA, F92+ Ol' 002+, 70'80 0C

Figure 8. Comparison of microbial and chemical synthesis of catechol.

Global, non-captive production of catechol is approximately 25,000 tons per
year.38 Estimating total world production is difficult since catechol is often part of captive
markets where it is produced and then converted into higher, value-added products before
ever seeing the open market as catechol. Chemical products derived from catechol include

pharmaceuticals (L-DOPA, adrenaline, papaverine), flavors (vanillin, eugenol, isoeugenol),

12

‘Q

1")“
.1:

agrochemicals (carbofuran, propoxur), and polymerization inhibitors and antioxidants (4-
tert-butylcatechol, veratrol).39 Most catechol production begins with Friedel-Crafts
alkylation of benzene to afford cumene (Figure 8). Subsequent Hock-type, air oxidation of
the cumene leads to formation of acetone and phenol. The phenol is then oxidized to a
mixture of catechol and hydroquinone using 70% hydrogen peroxide either in the presence
of transition metal catalysts or in formic acid solution where performic acid is the actual

oxidant. Catechol and hydroquinone are separated by distillation (Figure 8).39

©—»0 6 <3\

 

  
    

 

benzene cyclohexaneb cyclo- cycIo-\ H020
hexanol hexanone 8
0H . H020 / C02H
(OOH see Fig 8 Q CatA 8 H2, Pt adipic acid
; OH HO I
HO OH OH 002“
D-glucose catechol cis,cis-muconic acid

a = Nl-A1203, H2, 370-800 psi, 150-250 °C
b = C0, 02, 120-140 psi, 150-160 00
C = CU, NH4VO3, 60%) HNOs, 60'8000

Figure 9. Comparison of microbial and chemical synthesis of
adipic acid.

Inclusion of a catA gene encoding catechol 1,2-dioxygenase in the catechol-
producing microbe converts catechol into cis,cis-muconic acid (Figure 9).“0 The catA gene
was isolated from Acinetobacter calcoaceticus.41 Catalytic hydrogenation at 50 psi and
room temperature using 10% Pt on carbon converts the cis,cis-muconic acid into adipic
acid.40 Primarily used in production of nylon-6,6, the annual global demand for adipic

acid exceeds 1.9 x 10 9 kg.42 Benzene is the principal starting material from which adipic

13

acid is currently synthesized. Hydrogenation of benzene to produce cyclohexane is
followed by air oxidation to yield a mixture of cyclohexenol and cylohexanone. Nitric acid
oxidation then yields adipic acid and nitrous oxide,“3 which is involved in ozone depletion
and global warming.44 Adipic acid production may account for some 10% of the annual
increase in atmospheric nitrous oxide levels.43

Construction of microbial catalysts typically involves disruption and/or
overexpression of genes in a microbial host such as E. coli. Overexpressed genes are often
not native to the host microbe and have been recruited from other microbes. This has been
illustrated in the biocatalytic syntheses of Trp, Phe, PABA, p-hydroxybenzoic acid,
shikimic acid, quinic acid, gallic acid, catechol, and adipic acid. In this thesis, similar
strategies were used where K. pneumoniae and rat genes were recruited for synthesis of
vanillic acid from glucose. A second enzymatic step is then used to reduce the vanillic acid
to vanillin. The enzyme employed was aryl-aldehyde dehydrogenase purified from

Neurospora crassa.45

U'.'.h‘ . I'I'vru‘o I01 [1-‘I.' rib-.1. 'e
To compete with chemical synthesis, microbial synthesis must produce value-added
chemicals in a high yield (percent conversion) and high titer (product concentration) while
taking advantage of abundant, inexpensive feedstocks. Metabolic engineering of
biocatalytic syntheses utilizing the shikimate pathway has focused on the elimination of
feedback inhibition and repression of gene expression, overexpression of enzymes leading
to the desired product, and elimination of enzymes and pathways resulting in depletion of

the desired product. For example, in commercial tryptophan-producing strains, the ma

l4

gene encoding the tryptophan-degrading enzyme tryptophanase is deleted, and the genes in
the terminal pathway are all overexpressed.46

Alteration of central metabolism in E. coli to channel increased carbon flow into the
common pathway has recently been the focus of considerable research activity.47 In
addition to overexpressing feedback-insensitive DAHP synthase, strategies have been
evaluated to increase E4P and PEP availability. The rate of aromatic amino acid
biosynthesis is controlled by modulation of the catalytic activity of DAHP synthase. There
are three isozymes of DAHP synthase expressed in E. coli, each of which is sensitive to
feedback inhibition by one of the three aromatic amino acids. The genes aroF, aroG and
aroH encode tyrosine-sensitive, phenylalanine-sensitive, and tryptophan-sensitive DAHP
synthase isozymes, respectively. These isozymes' in vivo activities are dictated by their
expression levels, feedback inhibition by aromatic amino acids, and the availability of their
E4P and PEP substrates.

One option to manipulate AroF activity is to introduce a mutation into the locus that
encodes the aporepressor (TyrR) for aroF transcription. In lieu of TyrR, transcription of
aroF is derepressed and the number of molecules of AroF increases.48 Another way to
increase the transcription of aroF is to replace the native promoter (Pamp) with a strong
promoter such as a tac promoter (Ptac). RNA polymerase binding to the strong tac
promoter is not influenced by the TyrR repressor protein, resulting in more transcription.
Expression of AroF can also be increased by increasing the number of aroF genes in a
given microbe. This is often accomplished by extrachromosomally localizing aroF on a
multi-copy plasmid. Such a strategy increases the number of aroF genes beyond the
number of available TyrR repressor protein molecules thereby overriding transcriptional

regulation of aroF.

15

Glucose

 

 

 

  
  

 

Pva¢7 T PEP—\
7
G6P \
Ribulose-SP
7 1/ \1
F6P X5P RSP
1 Tkt
6
1 6FDP GAP S7P
V \2 1 .
6
101
3
DAHP synthase

DAHP

1

common pathway

Figure 10. Theoretical flux distribution for directing carbon
into common pathway with glucose as the carbon source. The
numbers are the relative fluxes needed to convert 7 mol of glucose into
DAHP. G6P, glucose 6—phosphate; F6P, fructose 6-phosphate; 1,6FDP,
1,6-fructose diphosphate; DHAP, dihydroxyacetone phosphate; GAP,
glyceraldehyde 3-phosphate; RSP: ribose 5-phosphate; XSP, xylulose 5-
phosphate; S7P, sedoheptulose 7-phosphate; PYR, pyruvate.

Amplified expression of DAHP synthase does not necessarily mean that the in vivo
activity of DAHP synthase has been increased, because of the prominent regulatory role
played by feedback inhibition.49 Several alleles that encode feedback-resistant DAHP
synthase have been obtained by mutation of the aroF50, aroG47a, and aroH5l loci.
Extensive mutation is not required to make the encoded DAHP synthase feedback-
insensitive. For instance, a single amino acid change can render AroF catalytic activity
insensitive to the concentration of Tyr.50 Insensitivity to feedback inhibition by aromatic

amino acids increases the in vivo catalytic activity of each molecule of DAHP synthase.

16

Ultimately, the catalytic activity of DAHP synthase increases to a point where
further amplification of even feedback-resistant DAHP synthase does not lead to improved
synthesis of either aromatic amino acids or precursors to these end products. Historically,
attention was focused on the in vivo availability of phosphoenolpyruvic acid (PEP) at this
juncture. PEP is a high energy metabolite which provides 3 of the 7 carbons of DAHP
(Figure 10). A number of different cellular processes and enzymes compete with DAHP
synthase for PEP including pyruvate kinase, PEP carboxylase, and the PEP-dependent
carbohydrate:phosphotransferase (PT S) system for uptake of glucose and structurally
related sugars. Efforts to improve intracellular PEP availability began with mutational
inactivation of PEP carboxylase47°~ 52 and pyruvate kinase.53 More recent efforts have
focused on circumventing PTS-mediated uptake of glucose in microbes such as E. coli.
Discovery that E4P availability is a critical limiting factor in aromatic amino acid
biosynthesis separated these two periods of research focusing on intracellular PEP
availability.47a

Mutational inactivation of PEP carboxylase and pyruvate kinase did not lead to
improvements in aromatic amino acid biosynthesis that were significant or practical. For
example, mutational inactivation of ppc‘ encoded PEP carboxylase results in a slow-
growing E. coli ppc‘ strain that requires succinic acid supplementation.“c Although E.
coli ppc' produces 10-fold higher phenylalanine titers relative to E. coli ppc+, these
phenylalanine titers are 10-fold lower than the amount of acetic acid which is produced.
Also, the 1.5 g/L titers of phenylalanine produced by E. coli ppc‘ are minuscule relative to
the 46 g/L of phenylalanine which can be produced by E. coli.54

In 1990, Frost and coworkers published the first work indicating that E4P
availability was also an important factor limiting in vivo DAHP synthase activity (Figure
10).55 As an aldose phosphate which can not exist in solution in a cyclic form, E4P is
prone to dimerization, trimerization and polymerization. Dissociation of these E4P forms

back to monomeric E4P is quite slow. This is probably the underlying reason why nature

17

r)_

‘f'
CL
5

ilgu

Ola;
ilI, ,
101ch

El .1111 '

“Nuk-

closely matches the rate of E4P synthesis with the rate of E4P utilization, thereby
maintaining low, steady-state concentrations of E4P. Low steady-state concentrations of

E4P likely limit this substrate's availability thereby limiting in vivo activity of DAHP

 

 

 

 

synthase.
H203PO OH O H203PO O H203PO OH H203PO OH OH
= + ; H & = H + ;
OH OH OH OH OH O OH O
o-fructose o-glyceraldehyde o-xylulose
6-phosphate 3-phosphate 5-phosphate
H203PO OH O H203PO OH O H203P0 OH H203PO OH OH OH
+ _a_s H '
:. :. .: H ‘__ : + :. :.
OH OH OH OH OH OH O OH OH O
o-fructose o-ribose o-sedoheptulose
6-phosphate 5-phosphate 7-phosphate
H203PO OH OH OH H203PO O b H203PO OH H203PO OH O
' + '—‘ H +
.1 : : H ‘— ; :-
OH OH O OH OH O OH OH OH
o-sedoheptulose o-glyceraldehyde erythrose o-fructose
- - D- -
7 phosphate 3 phosphate 4-phosphate 6 phosphate
(E4P)

Figure 11. Reactions catalyzed by transketolase (a) and transaldolase (b).

Inspection of the enzymes which catalyze reactions where E4P is either a substrate
or a product led to the pentose phosphate pathway and the enzyme transketolase (Figure
11). Amplified expression of transketolase increases the levels of E4P available to the cell
for channeling into the common pathway.55 Two of the three intracellular E4P-generating
reactions are catalyzed by transketolase while the third reaction is catalyzed by transaldolase
(Figure 11). Transketolase also serves to generate the substrate D-sedoheptulose-7-
phosphate for the E4P-generating reaction catalyzed by transaldolase. Thus, transketolase
plays a major role in increasing the levels of E4P available to the cell for aromatic

production. DAHP synthase, which catalyzes the first irreversible reaction of the common

18

tam

flO\

ppsax
syntht
loans:
03ers:
dating
expect:
[hf C0}

upon ;

lofialz

 

 

 

pathway, commits elevated levels of carbon generated by transketolase to aromatic amino
acid biosynthesis. Work combining expression of feedback-insensitive aroG with
amplified expression of tktA has revealed that increased DAHP synthase catalytic activity in
tandem with increased transketolase catalytic activity leads to a twofold increase in carbon
flow directed into the common pathway above that achieved with amplified expression of

DAHP alone.47a

PEP
0 synthase

0
OiJLOH + ATP + H20 ——>‘_ H203P0 OH + AMP + Pi
pyruvic acid PEP

Figure 12. Substrates and products associated with PEP
synthase (Pps) activity.

Liao and coworkers subsequently examined the impact of amplified expression of
pps-encoded PEP synthase on aromatic synthesis in E. coli aroB (Figure 10).47dv ‘3 PEP
synthase catalyzes the reaction of pyruvate with adenosine triphosphate (ATP) resulting in
formation of PEP, adenosine monophosphate (AMP) and inorganic phosphate (Figure 12).
Overexpression of PEP synthase was designed to recycle the pyruvate generated from PEP
during PTS-mediated glucose uptake back to PEP (Figure 10). Even though this might be
expected to alleviate intracellular PEP limitations, no increase in carbon flow directed into
the common pathway of aromatic amino acid biosynthesis was detected in E. coli aroB
upon amplified expression of pps and aroGFBR. However, expression (plasmid
localization) of tktA along with pps and aroGFBR resulted in a twofold improvement in
carbon flow as measured by the yield of DAH synthesized by E. coli (”08.47de These
results suggest that E4P is the first limiting metabolite in aromatic amino acid biosynthesis.
Once this limitation is removed, aromatic amino acid biosynthesis improves when PEP

availability increases as is the case with PEP synthase-catalyzed recycling of pyruvic acid.

19

Amplified expression of transaldolase also relieves E4P limitation in the presence of
amplified PEP synthase, but no further improvements in aromatic amino acid biosynthesis
are observed relative to when transketolase is overexpressed.56 While recycling of
pyruvate back to PEP using overexpressed PEP synthase (in the presence of overexpressed
transketolase and DAHP synthase) demonstrates that improved aromatic biosynthesis can
be realized with improved PEP availability, amplified expression of PEP synthase might
not be a long term solution. PEP synthase is a heavily regulated enzyme and the
consequences of its overexpression on the metabolic viability of the microbial biocatalyst
are unknown. Unfortunately, all experiments to date with amplified pps have been
performed under shake-flask conditions. Corresponding data on the impact of
overexpressed pps and tktA on aromatic biosynthesis under fed-batch fermentor conditions
has never made its way into the literature. Indeed, prior to this dissertation, all research
examining the impact of transketolase overexpression on carbon flow directed into the
common pathway was performed under shake-flask and not fed-batch fermentor
conditions.

Under typical shake-flask conditions, cells were first grown to stationary phase in
rich medium using antibiotics as the selection pressure to maintain plasmids. These cells
were then harvested and resuspended in minimal medium. Glucose in the minimal medium
is then converted to the desired molecule with limited or no cell growth. Although this type
of cell cultivation has been widely employed in evaluating pathway engineering, there are
several inherent problems associated with such experimental procedures. For examples,
the use of rich medium and antibiotics can dramatically increase the cost if such culturing
methods were scaled up. Centrifugation and resuspension of cells also complicate large-
scale manufacture. Under shake-flask conditions, because cells are grown in rich medium
prior to resuspension in minimal medium, calculated yields do not really reflect the percent
conversion of carbohydrate into product. DAH yields exceeding the theoretical maximum

yield were repeatedly reported when aroGFBR, tktA and pps were overexpressed in an E.

20

coli aroB cultured under shake-flask conditions.47a’ d» e In addition, under shake-flask
conditions, cultures begin in glucose-rich environment and end in a glucose—deficient
environment, and both oxygenation levels and pH are difficult to control.

In this thesis, fed-batch fermentor conditions were employed for DHS synthesis.
One of the advantages of using fed-batch fermentor conditions is that it increases cell
density at least lS-fold relative to the shake-flask cultivation. In addition, fed-batch
fermentor conditions overcome most of disadvantages of shake-flask conditions. For
examples, only minimal medium is used in the process, and the growth of the cells and the
production of DHS occur "simultaneously" under fed-batch fermentor conditiOns. A
strategy of using nutritional pressure to maintain the plasmid has also been developed. As
a result, antibiotics are no longer used in the fermentor. In addition, under fed-batch
fermentor conditions, variables such as temperature , pH, dissolved oxygen levels and the

steady-state concentration of the glucose can be easily controlled and manipulated on line.

21

CHAPTER 2

SYNTHESIS OF VANILLIN FROM GLUCOSE

Mame

Vanillin is the common name for 3-methoxy-4-hydroxybenzaldehyde (Figure 13).
It is one of the most important aromatic compounds in the production of flavors for foods
and fragrances for perfumes. As a food flavor, vanillin is second only to aspartame and
well ahead of both citric acid and monosodium glutamate in terms of annual sales.15 About
12 x 106 kg of vanillin are globally manufactured annually.57 Besides its popularity in the
food and perfume industries, vanillin has many other useful applications. For examples,
vanillin is used as a palatabilty enhancer to make animal feed more appetizing by flavor—
masking minerals with off-taste.58 Vanillin is also used in the pharmaceutical industry.
Since 1970, the use of vanillin as a chemical intermediate in the production of
pharmaceuticals has surpassed the quantities used for flavoring purposes.58 The single
largest use of vanillin is as a starting material for the manufacture of an antihypertensive
drug having the chemical name of Methyldopa or L-3-(3,4-dihydroxyphenyl)-2-

methylalanine.58 Vanillin is also a potent antioxidant and an antimicrobial agent.58’ 59

O H

CH30
OH

vanillin

Figure 13. The structure of vanillin.

As the major flavor and aroma component in vanilla extract, the concentration of
vanillin dictates the quality of vanilla extract.60 Free, unconjugated vanillin does not exist

in green vanilla beans. Natural vanillin is produced from glucovanillin when the beans of

the orchid Vanilla planifolia are submitted to a multi-step curing process.61 Free,
unconjugated vanillin is formed during the curing process by B—glucosidase cleavage from
the precursor glucovanillin which is biosynthesized from coniferol (Figure 14).61 The
labor intensity associated with vine cultivation, bean harvesting, and hand pollination of
flowers contributes to the difficulty in employing V. planifolia as a commercial source of
vanillin.51»62 In addition. the long and expensive curing process to produce vanilla extract
from vanilla beans only yields a product that has an inconsistent quality. As a result,
natural vanillin can supply only 0.2% of the demand for vanillin flavoring and the

consumption of naturally-occurring vanilla has gradually given way to synthetic vanillin.58

CHaOESLa “CHaofigb bCH305 —C+CH30:§:

0- glucose O-glucose
coniferol coniferin glucovanillin vaniolliHn

Figure 14. Biosynthesis of natural vanillin. (a) UDP—
glucose:coniferyl alcohol glucosyltransferase; (b) unidentified enzymes; (c)
B-glucosidase.

Synthetic vanillin can be produced either from the waste sulfite lye obtained from
wood pulping operations or from guaiacol and glyoxylic acid (Figure 15).58 To derive
vanillin from the lignin content of sulfite waste, lignin is treated with sodium hydroxide or
with calcium hydroxide solution under oxidative conditions. The reaction is generally run
at 160-175 °C and 150-160 psi.58 When the reaction is completed, the solid wastes are
removed. Vanillin is extracted from the acidified solution with butanol or benzene, and
then reextracted with sodium hydrogen sulfite solution. Reacidification with sulfuric acid
followed by vacuum distillation yields technical-grade vanillin, which must be
recrystallized several times to obtain food-grade vanillin. Although this process was

widely used in the 1970s, it has faced serious problems since then. The availability of

23

sulfite waste is reduced because newer processes for making paper paste yields less liquor.
The process's environmental impact is also problematic because over 160 ton of caustic

waste are produced from every ton of vanillin manufactured.58

HO O 02H
HJOKgOI-I
(CH30)2302
——->
HO H3CO NaOH __*H 3HCO
HO

HO HO

4- -hydroxy-3-methoxy

-phenyl glyoxylic acid vanillin

catechol guaiacol mandelic acid

Figure 15. Chemical synthesis of vanillin.

The other synthetic route to produce vanillin begins with the condensation of
guaiacol with glyoxylic acid (Figure 15). Air oxidation of the resulting mandelic acid
affords phenylglyoxylic acid. Crude vanillin is obtained by acidification and
decarboxylation of 4-hydroxyl-3-methoxyphenyl glyoxylic acid solution. Commercial
grades are obtained by vacuum distillation and subsequent recrystallization.58 Although
this process is currently the major source for synthetic vanillin, it has several inherent
problems. Guaiacol has historically been obtained from condensation of dimethyl sulfate
with catechol. Dimethyl sulfate is classified as a highly toxic, cancer-suspect agent.
Catechol is listed as being toxic and corrosive while guaiacol is a toxic irritant. Catechol is
currently manufactured from benzene (Figure 8) which is carcinogenic and derived from
nonrenewable petroleum.63

The limited availability of vanilla beans has led to extensive research into the use of
microorganisms to synthesize vanillin from ferulic acid. Such vanillin can be labeled as a
natural or nature-equivalent flavoring.64 Fungi, soil bacteria and plant cells are the main
organisms that have been examined. A two-step process for synthesis of vanillin has been
developed using Aspergillus niger to produce vanillic acid from ferulic acid followed by

employment of Pycnoporus cinnabrinus to reduce vanillic acid to vanillin. However, the

24

(I

r.)

h

)-

"I

 

90m ;

I».

‘1 l; :5";

 

process was plagued by low overall yield. The process also produced many impurities
because of the predominance of the vanillic acid oxidative system and the degradation route
of vanillin in the filamentous fungi. Under optimal conditions, the A. niger and P.
cinnabrinus route produced only 560 mg/L vanillin.65 Several soil bacteria have also been
carefully selected or mutagenized for their metabolic capacity to produce vanillin from
ferulic acid. A selected Streptomyce strain can produce 650 mg/L to 1 g/L vanillin after at
least three days' cultivation.“ 66 As another alternative, root cells from V. planifolia have
been cultured to produce vanillin from ferulic acid. However, this method is handicapped
by the fact that maintenance of plant cell cultures is rather expensive and time-consuming.
The lack of an abundant, inexpensive supply of ferulic acid with its toxicity are key
impediments to all biocatalytic approaches using ferulic acid as a starting material for

vanillin synthesis. 67

 

H203PO
/ 02H HO_. C02H HO COzH
D-glucose _., PEP AroFFBR O AroB CA
CH 0 - OH OH
OH R=PO3H2; DAHP
E4P R=H; DAH
O H
C302H 02H CO2H aryl-aldehyde
AroD gong AroZ dehydrogenase,
OH_’ HOCH3 OCH3
HO HSAM SAH ATP AMP, PP. 0”
DHS vanillic vanillin
acid NADPH NADP+

Figure 16. Biosynthesis of vanillin from glucose.

In this Chapter, a synthesis of vanillin from glucose will be elaborated. Glucose is
converted into vanillic acid by a recombinant Escherichia coli biocatalyst under fed-batch

fermentor conditions (Figure 16). Reduction of vanillic acid to vanillin is catalyzed by aryl-

25

aldehyde dehydrogenase isolated from Neurospora crassa (Figure 16). This synthesis
qualifies both as a route to natural vanillin and as a first step towards large-scale,

environmentally-benign manufacture of vanillin using biocatalysis.

26

A. The construction of the host strain

My.

AB2834, an E. coli aroE strain,68 was chosen as the parental strain for constructing
a vanillate-producing biocatalyst. The lack of aroE-encoded shikimate dehydrogenase
results in the synthesis of DHS when carbon flow is directed into the common pathway.
An aroBaroZ cassette was synthesized and site-specifically inserted into the serA locus of
E. coli AB2834 aroE via homologous recombination to generate E. coli KL7. Site-specific
homologous recombination followed from serA sequences which flanked the cassette. The
serA locus encodes D-3-phosphoglycerate dehydrogenase which is the first of the three
enzymes responsible for L-serine biosynthesis. Insertion of aroB into the genome
increases the activity of DHQ synthase, which is the only rate-limiting enzyme in the
portion of the common pathway of aromatic amino acid biosynthesis required for vanillin
synthesis.69 DHS dehydratase, encoded by aroZ, catalyzes the conversion of DHS into
PCA. Incorporation of aroZ into the genome was designed to economize on the number of
genes which needed to be plasmid-localized. Since serA-encoded 3-phosphoglycerate
dehydrogenase is an enzyme necessary for biosynthesis of L—serine, microbial growth in
minimal salts medium lacking L-serine supplementation is only possible if plasmids
containing serA inserts are stablely maintained. This strategy employing nutritional
pressure for plasmid maintenance obviates any need for use of antibiotics and plasmid-

localized antibiotic resistance for plasmid maintenance.

W
The aroB gene was obtained as a 1.6 kb fragment following digestion of pJB14

with EcoRI. Plasmid pJB14 is a 6.5 kb pKK223-3 derivative.70 The 1.6-kb fragment

included the native promoter and an inverted repeat downstream of the aroB gene capable

27

of forming a stem-loop structure characteristic of a rho-independent terminator. The aroZ
gene was obtained as a 2.2-kb fragment through PCR amplification from pSU1-28.
Plasmid pSU1-28 is a 5.8-kb plasmid which is a derivative of pSU1971 and contains a 3.5-
kb aroZ fragment isolated from the genome of K. pneumoniae. The primers used for PCR
were: 5'-CGGGATCCGCGCATACACATGC and 5'-
CGGGATCCGGGTACAGAGGGTGTTGT. Ligation of the 2.2-kb PCR product into
the BamHI site of pSUl8 afforded pSK4.99A. The 2.2-kb aroZ fragment includes its
native promoter and is transcribed in the same orientation relative to the lac promoter in
pSK4.99A. The 1.6-kb aroB fragment was blunt ended through Klenow fragment
treatment and subsequently inserted into the SmaI site of pSK4.99A to afford pKL4.237A
(Figure 17). The aroB gene is transcribed in the same orientation as aroZ in plasmid
pKL4.237A.

The purpose of employing pSK4.99A instead of pSUl-28 for pKL4.237A
construction was to use a shorter aroZ fragment for the cassette. Surprisingly, the BamHI
site between the aroB and aroZ disappeared in pKL4.237A. However, the loci of aroB and
aroZ were intact as were confirmed first by multiple restriction enzyme digestions. In
addition, pKL4.237A was transformed into competent AB2834/pMF62A to afford
AB2834/pMF62A/pKL4.237A. Plasmid pMF62A is a pBR322 derivative that bears a
copy of aroFFBS encoding feedback-sensitive DAHP synthase and a copy of tin/1.543 In
shake-flask conditions, construct AB2834/pMF62A/pKL4.237A converted 56 mM glucose
into 27.1 mM PCA, which was the sole product. As a control, construct
AB2834/pMF62A/pSK4.99A was also tested under the same conditions. It afforded a
mixture of 6.5 mM DAH(P) and 19.6 mM PCA (Table 1). Clearly, pKL4.237A
overexpressed aroB compared to the parental plasmid pSK4.99A, while both plasmids
conferred the active aroZ gene. The 3.9-kb aroBaroZ cassette was then isolated from
pKL4.237A with the inclusion of BamHI recognition sequences at the 5’- and 3’- ends

using PCR amplification. The following primers were used for the PCR: 5‘-

28

Smal BamHI Hindlll

 

 

pSU1-28
i PCR
BamHI BamHI
2.2 kb
aroz
BamHI digest 1) BamHI digest
2) CIAP treatment
T4 Ligase
Smal BamHI
pJB14
EcoRl digest
pSK4.99A
4'5 kb EcoFil EooFII
I 1.6 kb I
aroB
1) Smal digest
2) CIAP treatmentl Hindlll l Klenow treatment

(Smal) (BamHI)

 

Hindlll

Figure 17. Preparation of Plasmid pKL4.237A

29

CGGGATCCATCGGAGCTGATGTGGG and 5'-
CGGGATCCGGGTACAGAGGGTGTTGT.

Table 1. Confirmation of AroB and AroZ activities in pKL4.237A.

 

 

 

 

 

Construct DAH(P) DHS PCA
AB2834/pMF62A/pSK4.99A 6.5 mM 0 19.6 mM
A82834/pMF62A/pKL4.237A O 0 27.1 mM

11 ' ' ' ar r asse ' 28 4 cre

Although plasmids are maintained in E. coli as extrachromosomal, circularized
DNA, recombination events occasionally occur such that plasmid DNA is integrated into
the chromosome of the host cell. Exploitation of this rare event provides a method for
simple, site-specific insertion of a gene by flanking the gene with sequences homologous to
the genomic region targeted for disruption. Identification of cells with integrated plasmid
DNA is difficult since both freely replicating and integrated plasmids express resistance to
drugs encoded by plasmid markers. By using a conditional non-replicative plasmids,
selection of integrated plasmid DNA becomes possible since cells express drug resistance
only if the plasmid resides in the genome.

Normal cloning and preparation of plasmids containing conditional origins of
replication can be performed under permissive conditions whereas genomic insertions are
accomplished under nonperrnissive conditions where the replication machinery is inactive.
Plasmids possessing a temperature-sensitive replicon are capable of being manipulated in
this fashion. Plasmid pMAK705 contains a temperature-sensitive pSClOl replicon, a
chloramphenicol resistance marker, and a convenient multiple cloning site.72 The plasmid
is able to replicate normally when the host cell is grown at 30 °C but is unable to replicate

when the host cell is cultured at 44 °C. Thus genetic manipulations of the plasmid are

30

carried out at 30 °C while the integration of the plasmid into the genome can be selected for
at 44 °C.

The planned insertion of the synthetic cassette into the genomic serA locus in strain
AB2834 necessitated construction of a plasmid containing the synthetic cassette flanked by
serA DNA. The gene for serA was isolated from plasmid pD2625, which was digested
with EcoRV and DraI to liberate a 1.9-kb serA fragment. Plasmid pMAK705 was digested
with BamHI and treated with Klenow fragment. Subsequent ligation of the serA fragment
to pMAK705 afforded pLZl.68A.73 Since the original BamHI site in pMAK705 was
destroyed due to the Klenow treatment, the BamHI site inside the serA gene becomes a
unique restriction site, which facilitates the cloning of the synthetic cassette. Insertion of
the cassette into the BamHI site of serA in pLZl.68A yielded pKL4.276B (Figure 18).
Both aroB and aroZ are transcribed in the opposite orientation relative to the lac promoter.
DHSOI/pKL4.276B was tested for AroB and AroZ activities before plasmid pKL4.276B
was used for homologous recombination. Compared to DHSOI, DHSa/pKL4.276B
expressed 2.5-fold higher DHQ synthase activity. DHSor/pKL4.276B also afforded a 0.01

U/mg specific activity of DHS dehydratase while no DHS dehydratase activity was detected
in the E. coli DHSa.

Table 2. Plate selection for characterization of genomic insertion strain.

 

 

M9/glucose/

 

 

 

. M9/glucose/
StraIn Tyr IPhe IS A Tyr/Pne/SA/ LB/Cm LB
senne
A32834 + + _ +
KL7 — + — +
SA: shikimic acid

31

Sall BamHI EcoRl

 
   
 
     

p02625

Plac
EcoRV/ Dral digest

pMAK705

 

 
 
   
 

 

   

  
  
   

  

EcoRV Dral rep (ts)
I 1.9 kb 4 5-4 "b
‘ ~ 011
serA V
1) BamHI digest
2) Klenow treatment
3) CIAP treatment
T4 Ligase
(BamHI) EcoFil
rep (ts)
serA KL4.237A
pLZ1.68A p
7.4kb 0m
/ l PCR
BamHI BamHI BamHI
| ac 1.6 kb 2.2 kb
aroB amZ
(BamHI) Sall
1) BamHI digest BamHI digest
2) CIAP treatment

(BamHI) EcoFil

 

Figure 18. Preparation of Plasmid pKL4.276B.

32

Conditions for homologous recombination were based on those previously
describedfi9b The competent host strain was transformed with pKL4.276B. Following
heat-shock treatment, cells were incubated in LB at 44 °C for l h and subsequently plated
on LB plates containing Cm. Plates were incubated at 44 °C for approximately 20 h before
colonies appeared. The resulting cointegrates were cultured in 5 mL of LB containing no
antibiotics and were grown at 30 °C for 12 h to allow excision of the plasmid from the
genome. Cultures were diluted (1:20000) in LB without antibiotics, and two more cycles
of growth at 30 °C for 12 h were carried out to enrich cultures for more rapidly growing
cells that had lost the temperature-sensitive replicon. Growth cultures were then diluted
(1:20000) into LB and grown at 44 °C for 12 h to promote plasmid loss from the cells.
Serial dilutions of each culture were spread onto LB plates and incubated at 30 °C
overnight. The resulting colonies were screened on multiple plates to select the desired
recombinants. E. coli KL7 was isolated based on the following growth characteristics:
growth on M9 containing L-tyrosine, L-phenylalanine, shikimic acid and serine; no
growth on M9 containing L-tyrosine, L-phenylalanine, shikimic acid; growth on LB; and

no growth on LB containing Cm (Table 2).

B. Plasmid Constructions

mam

Three genes needed to be localized in a plasmid. These included aroFFBR, serA,
and COM T. The aroFFBR insert encodes a 3-deoxy-D-arabino-heptulosonic acid 7-
phosphate synthase isozyme insensitive to feedback inhibition for increasing carbon flow
into the common pathway. Plasmid-localized serA was necessary for complementation of
the serA mutation in the KL7 genome. This provided the nutritional pressure for stable
maintenance of plasmids carrying serA by KL7 when cultured in minimal salts medium
lacking serine supplementation. Expression of the COM T gene encoding catechol-0-

methyltransferase (COMT) was required for conversion of PCA into vanillic acid. The

33

activities of catechol-O-methyltransferase were primarily found in plant and mammalian
tissue.74 The cDNA encoding COMT from rat liver had been cloned and expressed in E.
coli. 75 An E. coli promoter and ribosome binding site sequences were added in front of the
open reading frame (ORF) of the COMT cDNA to ensure transcription and translation in E.
coli. Because catechol-O-methyltransferase has a turnover number of only 24 per min,76

the strong Pm promoter was used to achieve adequate activity levels.

W

A 1.3-kb aroFFBR fragment encoding a feedback-insensitive DAHP synthase AroF
was obtained through UV mutagenesis. Ligation of the aroFFBR fragment into the EcoRI
site of pSU18 led to pKL4.20B. The aroFFBR gene was transcribed in the opposite
orientation relative to the lac promoter in pKL4.2OB (Figure 19). Plasmid pKL4.20B was
further digested with Smal. Ligation of a 1.9-kb DraI/EcoRV fragment encoding serA
obtained from pD2625 into the Smal site of the pKL4.20B afforded pKL4.33B (Figure
20). The serA gene was transcribed in the same direction as aroFFBR.

Plasmid pKK223-370 is a cloning vector with a pMBl origin of replicon. It
contains a strong tac promoter and a ribosome binding site (RBS) which can be utilized to
express gene inserts. Plasmid pCRX-275 carries the open reading frame (ORF) of the
COMT gene. Digestion of pRCX2 with EcoRI/HindIII was followed by isolation of a 0.7-
kb fragment encoding the ORF of COMT. Plasmid pKK223-3 was also digested with
EcoRI/HindIII to yield a 5.3-kb fragment. Ligation of these two fragments resulted in
pKL3.272A, in which the COMT was behind the tac promoter and an E. coli RBS (Figure
21).

Digestion of pKL4.33B with SphI and Hindlll afforded a 5.5-kb fragment, while
the same digestion of pKL3.272A yielded a 1.1-kb PMCCOMT fragment. Ligation of these
two purified fragments resulted in pKL5.26A (Figure 22). Plasmid pKL5.26A was
digested with HindIII/Xbal and the 1.1 kb PMCCOMT fragment was isolated and treated

34

with Klenow fragment. Plasmid pKL5.26A was also digested with Hindlll and treated
with Klenow fragment. Subsequent ligation of the PmcCOM T to pKL5.26A afforded
pKL5.97A. The two PmCOMT gene are transcribed in the same orientation (Figure 23).

EcoRl Smal Hindlll

 

 

EcoRl ECORI
l 1.3 kb
aroFFER
EcoRl digest 1) EcoFil digest
2) CIAP treatment

T4 Ligase

EcoFll

    
     
    

l”lac

aroFFER
pKL4.208
3.5 kb

\Q
EcoRl
Smal

Hindlll

Figure 19. Preparation of plasmid pKL4.20B.

35

p02625

l EcoRV/ Dral digest

EcoRV Dral
L 1.9 kb _|

l )

serA

    
 

pKL4.208
3.5 kb

Hindlll

1) Smal digest
2) CIAP treatment

T4 Ligase

   
  
    

pKL4.338
(Cm 5.5 kb

Hindlll

Sphl (Smal)

Figure 20. Preparation of plasmid pKL4.33B.

EcoRl Smal Hindlll

    
     
 

Sphl

Ptac
pRCX2

AP
pKK223-3

EcoRl / Hindlll digest
4.6 kb

 
 
   
      

  

EcoRl Hindlll
I 0.7 kb 'I
COMT
1) EooRI I Hindlll digest
2) CIAP treatment
T4 Ligase
EcoFil
Sphl
Ptac
COMT
Hindlll

pKL3.272A
5.3 kb

V

Figure 21. Preparation of plasmid pKL3.272A.

37

a;

  

pKL3.272A

   
   

jam / Hindlll digest (Cm pKL4.338

5.5 kb

1.1 kb
Sphl EcoRl Hindlll
Hindlll

Pmcow 39'“ (Small
1) Sphl / Hindlll digest
2) CIAP treatment

T4 Ligase
pKL5.26A
Hindlll

 

EcoRl Sphl

Figure 22. Preparation of Plasmid pKL5.26A.

38

pKL5.26A

¢ Xbal / Hindlll digest

Xbal 1.1 kb lflndfll
PtacCOMT

lKlenow treatment

“findfln

GflndHD

      
 
      

pKL5.26A
6.6 kb

serA

lfindHl
COMT Ptac

EcoRl Sphl

2) Klenow treatment

i 1) Hindlll digest
3) CIAP treatment

T4 Ligase

EcoRl

 

Figure 23. Preparation of Plasmid pKL5.97A.

39

C. F ed-Batch Fermentor Synthesis of Vanillic Acid

KL7/pKL5.26A were cultured for 48 h under fed-batch fermentor conditions at 37
°C. pH 7.0, and dissolved oxygen at 20% of air saturation.33 Cells reached stationary
phase at approximately 24 h after inoculation. Extracellular accumulation (Figure 24) of
vanillic, isovanillic, protocatechuic, and 3-dehydroshikimic acids began in mid log phase of
microbial growth. Upon cessation of the fermentation at 48 h, the production of vanillic
acid, isovanillic acid, PCA and DHS were 2.5 g/L, 0.4 g/L, 9.7 g/L and 0.9 g/L,
respectively. The 6-fold molar excess of vanillic acid synthesized relative to isovanillic acid
(Table 3) is consistent with the reported selectivity of catechol-O-methyltransferase towards
meta-hydroxyl group methylation. DHS usually constituted 6 mol% of the total product
mixture indicating that the rates for its biosynthesis and dehydration were nearly equal.
However, the molar dominance of protocatechuic acid relative to vanillic acid pointed to
inadequate catechol-O-methyltransferase activity. In addition, little vanillic acid was
produced during the stationary phase (Figure 24). A 25 mL of aliquot of fermentation
broth was taken at 12 h and 36 h for determination of catechol-O-methyl transferase
activity. COMT activity was apparently stable with a specific activity of 0.0064 U/mg and
a specific activity of 0.0056 U/mg at 36 h.
Table 3. Products formed after 48 h under fed-batch fermentor conditions

as a function of catechol-0-methyltransferase activity and L-methionine
supplementation.

 

 

 

COMT L-methionine vanillic isovanillic PCA DHS

construct . . , _ b acid acid
specrfic activnya addItIon (g/L) (g/L) (g/L) (g/L)
KL7/pKL5.26A 6.0 x 10-3 — 2.5 0.4 9.7 0.9
KL7/pKL5.26A 5.5 x 10'3 + 4.9 1.3 7.1 1.0
KL7/pKL5.97A 12.0 x 10-3 - 3.0 0.6 12.9 1.0
KL7/pKL5.97A 103 x 10-3 + 5.0 1.2 10.5 1.8

 

 

a specific activity: umol/min/mg. b 0.4 g/L added every 6 h beginning at 12 h.

40

To improve COMT activity, plasmid pKL5.97A was constructed, which contains
two copies of PmcCOMT as opposed to the single copy of PmcCOMT in pKL5.26A. The
specific activity of catechol-O-methyl transferase of 0.012 U/mg in KL7/pKL5.97A was
twice the levels of the transferase activity in KL7/pKL5.26A (Table 3). Unfortunately,
increased catechol-0-methyltransferase specific activity (Table 3) in KL7/pKL5.97A
relative to KL7/pKL5.26A had little impact on the concentrations (Table 3, Figure 25) of
synthesized vanillic acid. As with KL7/pKL5.26A, vanillic acid production was minimal
during the stationary phase of KL7/pKL5.97A fermentation (Figure 25). This led to

consideration of S-adenosylmethionine (SAM) availability as a possible factor limiting

COMT activity.

 

\l
0
GD
0

0'! O)
O O
1 l
l l

h
C
1
I

 

 

0)

C
l
1

u
J

 

N
o
l
1

 

isovanillate, PCA, DHS (mM)
.1.
0

cells (g/L), vanIIlate (mM)

 

.a
o
n
r
J
1
—L
o

 

 

 

 

 

i

 

 

o
L
t
o

0 912182430364248
time (h)
Figure 24. Fermentation culture of KL7/pKL5.26A without
methionine supplementation. —o— vanillic acid; misovanillic acid;

I: protocatechuic acid (PCA); — 3-dehydroshikimic acid (DHS);
-0— cell mass.

Catechol-O-methyltransferase catalyzes the methyl transfer reaction from SAM to

PCA to form vanillic acid and isovanillic acid (Figure 16). SAM is a central metabolite of

41

E. coli. Synthesized from L-methionine and ATP by SAM synthetase, SAM is the major

methyl donor in metabolism.77

 

 

 

 

 

 

 

 

 

 

80 60
A70.-
3 — n "502
v60» 3
a)
I - 400
050-- — P g
<‘ E
240- g
2'30- ’3’
g a
.E - V
$20 %
910- °

0d

 

 

 

 

0 91218243036 4248
time (h)

Figure 25. Fermentation culture of KL7/pKL5.97A without
methionine supplementation. —o— vanillic acid; misovanillic acid;
:2: protocatechuic acid (PCA); — 3-dehydroshikimic acid (DHS);
—0- cell mass.

Because supplementation with L-methionine had been reported to increase the
intracellular SAM levels,77v 73 L-methionine was added to both KL7/pKL5.26A and
KL7/pKL5.97A ferrnentations at timed intervals. In both cases, increased vanillic acid
concentrations were synthesized as a result of L-methionine supplementation. With 0.4 g/L
L-methionine addition every 6 h beginning at 12 h, KL7/pKL5.26A and KL3/pKL5.97A
synthesized 4.9 g/L and 5.0 g/L of vanillic acid, respectively, after 48 h fermentationCTable
3, Figure 26a, b). Although PCA is still a major product in both cases, the titers of vanillic
acid with L-methionine addition are almost twice as high as those of without L-methionine
addition. The ratios of PCA/vanillic acid also decreased significantly while the specific

COMT activities remained the same (Table 3). Of equal importance, vanillic acid synthesis

42

was observed during stationary phase as well as during exponential phase in both cases
(Figure 26a, 26b). Clearly, methionine addition has an important impact on the conversion
of PCA to vanillic acid. However, further increases in the amount of L-methionine added

did not increase the concentrations or yields of synthesized vanillic acid.

43

 

40--

 

 

 

 

 

30-; FT“

 

N
o
L
I
I
fl
'0
o

 

IsovanIllate, PCA, DHS (mM)
8
o

cells (g/L), vanIllate (mM)

A
O
l
I

1

fi
A
O

 

 

 

 

 

 

 

 

o
I
I

o

o 9 12 18 24 so 36 42 48
thw m)

 

\l
O
O}
O

O)
O
I
I

J

C”
O
I
I

 

.h

0

4
I

 

 

 

(A)

o
I
I

 

 

I
I
N
C

 

 

N
o
I
I

 

isovanillate, PCA, DHS (mM)
8
0

cells (g/L), vanIIIate (mM)

 

..L
O
I
I
I
I
.a
o

 

 

 

 

 

o
I
I

C

0 912182430364248
time (h)

Figure 26. Fermentation cultures of (a) KL7/pKL5.26A and (b)
KL7/pKL5.97A supplemented with methionine. + vanillic acid;
Inmisovanillic acid; I=Iprotocatechuic acid (PCA); — 3-
dehydroshikimic acid (DHS); -0- cell mass.

A. Purification of Aryl-Aldehyde Dehydrogenase from Neurospora crassa

WW

Microbial reductions of aromatic carboxylic acids, usually to their corresponding
alcohols, have been observed with whole cell biotransformations by a number of
microorganisms, including Actinomyces, Aspergillus niger, Neurospora crassa , Nocardia
sp. and Corynespora melonis.79 Among them, Neurospora crassa was the first to be
studied.80 In the early 1970s, Gross and Zenk discovered that Neurospora crassa reduces
aryl carboxylic acids to aryl alcohols via an intermediate aryl aldehyde. The enzyme that
catalyzes the reduction of the carboxylic acid to aldehyde was identified and named aryl-
aldehyde oxidoreductase or aryl-aldehyde dehydrogenase. Mycelia of the fungus
Neurospora crassa has to be grown in culture medium supplemented with salicylate in
order to form the enzyme. The aryl-aldehyde dehydrogenase can reduce a variety of
aromatic acids including benzoic acid, salicylic acid, p-hydroxybenzoic acid and vanillic
acid. The enzyme was purified and identified as a monomeric protein with an apparent
molecular mass of 120,000 Da. The reduction required ATP, Mg2+ and NADPH as
cofactors. Further work showed that this enzyme catalyzed the initial reaction between an
aromatic acid and ATP to form an acyl-AMP intermediate which was subsequently reduced

by NADPH to form the aldehyde (Figure 27).80

M924-

 

 

aromatic acid + ATP + enzyme enzyme-acyl-AMP + PPi

R-SH
enzyme - acyI-AMP + NADPH + H+ = aromatic aldehyde + AMP+NADP+

Figure 27. Mechanism of aryl-aldehyde dehydrogenase
reduction of aromatic acids.

45

Neurospora crassa SY 7A was obtained from the American Type Culture Collection
(ATCC 24740). It was grown on solid growth medium at 24 °C for 7 days and a mixture
of mycelium and spores was obtained. After suspension in sterilized water and filtration of
the mixture through sterilized glass wool, a spore suspension was obtained. The
concentration of the spores in the suspension was estimated by measuring the absorbance at
650 nm ( 1 A650 unit = 5.0 x 106 spores/mL).81 Spores were then inoculated in liquid
growth medium ( 2.5 x 10‘5 spores per liter medium) supplemented with 1.6 g/L sodium
salicylate. The fungus starts to produce salicylic alcohol (saligenin) after a lag phase of
about 50 h. Maximum aryl-aldehyde dehydrogenase activity was observed at 60 h when
approximately 20% of salicylate were reduced to saligenin. The mycelium was harvested
by filtration at this time and frozen at -20 °C. It is critical to use relatively fresh spores
(stored at 4 °C less than two weeks) for the inoculation in order to obtain the optimal aryl-

aldehyde dehydrogenase activity at 60 h.

Table 4. Purification of aryl-aldehyde dehydrogenase from N. crassa.

 

 

 

Maggie" Wants (fig/mfg 3?ng new
crude Iysate 3600 \ \ \ \
DEAE 809 53 0.072 1 100%
FledA 1 O7 55 0.52 7 96%

 

 

1 unit = 1 pmol NADPH oxidized /min.

'0 - h dro enase

Besides aryl-aldehyde dehydrogenase, a native enzyme called aryl-alcohol
dehydrogenase also exists in Neurospora crassa. This aryl-alcohol dehydrogenase can
reduce the aromatic aldehyde to the corresponding alcohol. Specifically, it can reduce

vanillin to vanillyl alcohol.80 Therefore. the primary purpose of the protein purification

46

was to purify away the unwanted aryl-alcohol dehydrogenase. In addition, other NADPH-
consuming enzymes in Neurospora crassa needed to be purified away from aryl aldehyde
dehydrogenase since these oxidoreductases can compete with the aryl-aldehyde

dehydrogenase for cofactor NADPH.

Table 5. The process of reducing vanillic acid to vanillin.

 

 

 

termenation EtOAc CHZCIZ
broth extraction reprecipitation reduction extraction
DHS (g) 0.17 o o o o
PCA (g) 1.04 0.93 0.14 0.095 0
lsovanilllc acid (9) 0.12 0.098 0.062 0.007 0
vanillic acid (9) 0.50 0.44 0.38 0.03 0
ggzfigdeecg). — — — 0.043 0
isovanillin (g) — — — 0.045 0.03
vanillin (g) — — — 0.32 0.30
ylelda (mol/mol) 100% 88% 76% 71 % 66%

 

 

a vanillic acid or vanillin recovery yield.

The purification was performed using a DEAE column followed by a RedA
column. Aryl—aldehyde dehydrogenase activities were monitored by assaying for the
reduction of benzoic acid using NADPH and ATP as cofactors. Benzoic acid was used for
the assay as opposed to vanillic acid because benzoic acid is a better substrate for the aryl-
aldehyde dehydrogenase. The specific activity of aryl-aldehyde dehydrogenase could not
be determined in crude mycelium extract because of contaminating dehydrogenase
activities. Aryl-aldehyde dehydrogenase purified through the two columns (Table 4) was
not homogenous. However, contaminating aryl-alcohol dehydrogenase was no longer
detectable. In addition, without addition of benzoic acid, NADPH was not consumed.

This indicated that other NADPH-competing enzymes had also been purified away from

47

aryl-aldehyde dehydrogenase. Yields and specific activities for each step of the purification
of aryl-aldehyde dehydrogenase are summarized in Table 4. This partially purified enzyme

was used for the in vitro reduction of vanillic acid to vanillin.

B. Reduction of Vanillic Acid

KL7/pKL5.97A was cultured under fed-batch fermentor conditions with optimal L-
methionine supplementation for 48 h. A portion (100 mL) of the fermentation broth was
then acidified to pH 3.0 and the precipitated proteins were removed by centrifugation.
Extraction of the protein-free broth with EtOAc separated DHS from vanillic,
protocatechuic, and isovanillic acids which were in the organic layer after the extraction.
Removal of the EtOAc afforded a solid mixture of vanillic acid, PCA and isovanillic acid
with 88% (mol/mol) yield recovery. The extraction did not change the ratio of PCA and
vanillic acid. A subsequent reprecipitation step increased the vanillic acid/protocatechuic
acid ratio from 1:2 to 2.5:1 (mol/mol). This reprecipitation was accomplished by taking
advantage of the solubility differences between PCA and vanillic acid. The solid mixture
after the extraction and solvent removal was dissolved in water adjusted to pH 7.5 by
NaOH addition. Subsequent dropwise addition of concentrated sulfuric acid acidified the
solution to pH 1.8 and resulted in precipitation of a solid which was filtered and dried. The
recovery of vanillic acid for this step was 86% (mol/mol). The resulting aromatic mixture
was incubated with glucose 6-phosphate dehydrogenase (to recycle NADP+) and aryl-
aldehyde dehydrogenase at 30 °C and pH 8.0 using 0.07 equiv. of NADP+ and 2 equiv. of
ATP relative to vanillic acid. Reduction of vanillic acid to vanillin proceeded in 92%
(mol/mol) yield in 7 h. Reduction of PCA was slower with a 33% (monol) yield of
protocatechualdehyde after 7 h. Vanillin was extracted from the enzymatic reduction with
CH2C12 leaving protocatechualdehyde and protocatechuic acid in the aqueous phase.
Removal of the solvent afforded vanillin as a solid with isovanillin present at 10 mol% as

the only contaminant (Figure 35). Extraction of the fermentor broth, selective precipitation

48

to remove excess protocatechuic acid, aryl-aldehyde dehydrogenase reduction, and the final
CH2C12 extraction led to a 66% overall yield (mol/mol) for conversion of vanillic acid into

vanillin (Table 5).

49

Figure 28. Control 1H NMR of vanillic acid. Resonances: 8 7.55 (d, 1H), 6.95
(d, 1H), 7.47 (dd, 1H), 3.92 (S, 3H).

50

o Eu m m w

P—nbnh-phi-P—anPF—bbb—hbpbhnh-g—phnhnnppnm

m

m

u

m

..-.~.».»—.-p-_PpP»—.....—pppphppp_.~L~—.p-b.PPP.

D

 

g

 

 

+14

?

 

J

51

Figure 29. Control 1H NMR of isovanillic acid. Resonances: 5 7.41 (d, 1H),
7.06 (d, 1H), 7.49 (dd, 1H), 3.91 (s, 3H).

52

o

P_.-..-pn-—-Pnp-.p.b—---b-.~pppn—-.p

I!“

r if

 

m

n

v

 

II

111

 

(11

1‘ IJJII‘I

{‘11

I;

d

‘1

1

 

j

m

m

5

III?) I'PII'I

F

—.-.-.pp-P.-p.—P.Pb-..nP-pb.--pp—.pn.—p-upppn-P_

 

1.11

 

‘4 I]

{I

53

Figm
7.45 I

 

 

Figure 30. Control 1H NMR of protocatechuic acid (PCA). Resonances: 8
7.45 (d, 1H), 6.95 (d, 1H), 7.40 (dd, 1H).

54

o 1am” m

L—-.._-_-—.».__np.-_pbp--Pp

 

m w

“hbnb—b-bb—Phbhnbbbb

r

14

m
_

i

 

m

.--b~.-_.-»»--p

1111

u

bu.Pp—-.~

P-~P---_

 

55

Figu
6 6.4

 

 

Figure 31. Control 1H NMR of 3-dehydroshikimic acid (DHS). Resonances:
5 6.42 (d, 1H), 4.28 (d, 1H), 4.00 (ddd, 1H), 3.07 (dd, 1H), 2.66 (ddd, 1H).

56

tau

nnph

.lJ

bubbpunbpw

7i

 

1111‘

N

:p._:.P_.£:::

1-

m

+

::CL~PP.:.:.-__....::._:..:.:

4.14

 

 

 

m

o

 

3"}

 

‘

m

..:_.:._:_:t:

o
_

 

la

 

44

57

Fig

 

Figure 32. Control 1H NMR of vanillin. Resonances: 8 7.39 (s, 1H), 6.98 (d,
1H), 7.46 (d, 1H), 3.88 (8, 3H), 9.63, (s, 1H).

58

In: .

1rLl _

bl

P

 

 

 

 

 

 

 

 

 

j

 

59

FigUI
Id. 11

 

 

Figure 33. Control 1H NMR of isovanillin. Resonances: 8 7.38 (s, 1H), 7.18
(d, 1H), 7.55 (d, 1H), 3.95 (8, 3H), 9.70 (8, 1H).

60

 

 

 

 

 

 

 

61

Figu
fern
resor
resor.

6 7.4

 

 

Figure 34. 1H NMR after EtOAc extraction of the KL3/pKL5.97A
fermentaiton broth (with methionine supplementation). Vanillic acid

resonances: 8 7.55 (d, 1H), 6.95 (d, 1H), 7.47 (dd, 1H), 3.92 (s, 3H). Isovanillic acid
resonances: 8 7.41 (d, 1H), 7.06 (d, 1H), 7.49 (dd, 1H), 3.91 (s, 3H). PCA resonances:
8 7.45 (d, 1H), 6.95 (d, 1H), 7.40 (dd, 1H).

62

 

 

 

 

 

 

 

 

63

FigI
\"an
111 I.
9.70

 

Figure 35. 1H NMR after CHzClz extraction of the reduction reaction.
Vanillin resonances: 8 7.39 (s, 1H), 6.98 (d, 1H), 7.46 (d, 1H), 3.88 (s, 3H), 9.63 (s,

1H). Isovanillin resonances: 8 7.38 (s, 1H), 7.18 (d, 1H), 7.55 (d, 1H), 3.95 (s, 3H),
9.70 (s, 1H).

can c

P F b

L

P

 

 

 

 

 

 

 

 

 

 

 

65

ran
rout
“1th
and

lSlhc

pros]
Phen
dehy
Corr

prec;

 

hon:
USEC
ads:
Hm:
abur

PTO_?I

 

Discussion

Biocatalytic synthesis of vanillin from glucose has a number of advantages relative
to other biocatalytic vanillin syntheses. Coniferol, formed during phenylpropanoid
biosynthesis, is converted into coniferin by a glucosyltransferase in Vanilla planifolia.61
Coniferin is then transformed into glucovanillin which is finally hydrolyzed to vanillin by a
B—glucosidase. Synthesis of vanillin via 3-dehydroshikimic, protocatechuic, and vanillic
acids circumvents phenylpropanoid biosynthesis and glucosylation/deglucosylation
reactions. This substantially reduces the number of enzymes required to synthesize
vanillin. Ferulic acid can also be microbially converted into vanillin. However, these
routes produce vanillin titers below 1 g/L. Cultured plant tissue in medium supplemented
with ferulic acid has also been used to produce vanillin although this conversion is slow
and scale-up is relatively difficult. A problem with all vanillin syntheses from ferulic acid
is the absence of low-cost, commercial production of this phenylpropanoid.

Biocatalytic synthesis of vanillin from glucose, although a longer term commercial
prospect, also has advantages relative to synthetic vanillin manufacture (Scheme 1).
Phenol and guaiacol are toxic and are derived from carcinogenic benzene. Nontoxic 3-
dehydroshikimic, protocatechuic, and vanillic acids are derived from innocuous glucose.
Corrosive H202 used for oxidation of phenol into catechol requires special handling
precautions. By contrast, biocatalytically synthesized vanillin derives its oxygen atoms
from the oxygen atoms of glucose. Dimethyl sulfate, a carcinogen, has historically been
used to methylate catechol. Protocatechuic acid methylation employs S -
adenosylmethionine generated and consumed intracellularly. Finally, synthetic vanillin
manufacture is based on use of nonrenewable petroleum whereas glucose is derived from
abundant, renewable starch. This difference in feedstock utilization is important given

projected fierce international competition as global petroleum production diminishes.82

66

redu
rani

rani

afier
eren
liar
than

Ianfl

enzyI

 

Future Wurk

Use of an intact microbe to reduce vanillic acid will be essential for future large-
scale vanillin synthesis. Vanillic acid synthesized by one microbe from glucose could be
reduced to vanillin by a second, different microbe. For example, E. coli can be used for
vanillic acid synthesis and Neurospora crassa can be used to convert vanillic acid into
vanillin. In active, growing Neurospora crassa cultures, it was observed that addition of
vanillic acid (10 mM) resulted in formation of vanillyl alcohol in the culture supernatant
after 60 h. All of the added vanillic acid and the initially formed vanillyl alcohol were
eventually metabolized after cultivation for 5 days. This observation suggests that intact
Neurospora crassa could be employed to reduce vanillic acid to vanillin. The primary
challenge is to mutagenically inactivate the aryl-alcohol dehydrogenase, which converts
vanillin into its alcohol form. Mutagenizing genes encoding other vanillin-utilizing
enzymes may also be necessary.

Conversion of glucose into vanillin using a single vanillate-synthesizing microbe
expressing aryl-aldehyde dehydrogenase may also be possible. For example, the
Neurospora crassa gene encoding aryl-aldehyde dehydrogenase can be cloned, sequenced,
and expressed in the vanillate-synthesizing E. coli biocatalyst. The clone of the gene
encoding aryl-aldehyde dehydrogenase can be isolated either from a Neurospora crassa
genomic or cDNA library. Direct isolation and expression of a gene from a N. crassa
genomic DNA library in E. coli has been successful in the past.83 However, this strategy
is risky. Fungal promoters are not recognized in bacteria and introns, if present, cannot be
excised from the primary transcript in E. coli, which results in expression of truncated
protein.84

To avoid problems with promoters and introns, a Neurospora crassa cDNA library
can be constructed and screened with synthetic oligonucleotide probes designed from the
determined amino acid sequence of aryl-aldehyde dehydrogenase. The aryl-aldehyde

dehydrogenase will first need to be purified to homogeneity followed by determination of

67

its :
nuc
be 1
P5P
tern

p05

5an

librz

 

 

its amino acid sequence. Because an oligonucleotide probe containing fifteen to twenty
nucleotides is sufficient to screen a library, only a small portion of the total protein needs to
be sequenced.“ The purified protein needs to be digested with trypsin and the resulting
peptides separated. Once several peptides have been partially sequenced from their N-
terminus, the six or seven amino acid stretch that can be encoded by the smallest number of
possible DNA sequences will be determined and the corresponding degenerate probe
synthesized. Radiolabelling of the probe can be carried out by standard procedure.86

N. crassa poly(A+) mRNA will need to be isolated in order to construct the cDNA
library. Two strategies can be used to reduce the size of the cDNA library needed for
screening the desired aryl-aldehyde dehydrogenase gene. The aryl-aldehyde
dehydrogenase is an induced enzyme, which is expressed upon addition of sodium
salicylate. Subtractive hybridization can be used to differentiate the mRN A of induced gene
from the large mRNA pool.87 In addition, the aryl-aldehyde dehydrogenase has a
molecular weight of 120 kDa, which is higher than the average Neurospora crassa protein.
Its mRNA must therefore be larger in size than the. bulk of the mRNA in Neurospora
crassa. Size-based fractionation of the subtracted mRNA for aryl-aldehyde dehydrogenase
will further reduce the size of the mRNA pool.86 Subsequent synthesis of the first strand
of cDNA with reverse transcriptase will be followed by replacement synthesis of the
second strand of cDNA using RNAase H and the Klenow fragment from E. coli. After
addition of synthetic linkers, the cDNA will be ligated into the bacteriophage XZAPII
vector. The ligation mixture will be packaged in vitro and plated on XLl-Blue E. coli to
obtain plaques.88 The recombinant 7t virions present in plaques on a lawn of E. coli will be
transferred to a nylon membrane. Hybridization to the radiolablled probes and
autoradiography will be carried out until the positive clone is located. The identified gene
can then be sequenced and plasmid-localized to provide a promoter and ribosome binding

site for expression in E. coli.

68

metl
is It
met}
bios
cere
mull
solu
with
feed
of C

IUI:

 

Improving protocatechuic acid methylation will also be essential. The lack of
significantly improved protocatechuic acid methylation with increased catechol-0-
methyltransferase activity and the improvement in methylation observed with L-methionine
supplementation suggest that cosubstrate S-adenosylmethionine availability and/or feedback
inhibition39 may be limiting in vivo methyltransferase activity. Overexpression of E. coli
SAM-synthase (MetK) along with use of L-methionine supplementation did not increase the
ratio of vanillic acid/PCA.90 This suggests that increasing SAM availability is a rather
complicate issue. For example, supplementation with methionine represses MetK and most
of the methionine biosynthetic enzymes.91

Ethionine-resistant Saccharomyces cerevisiae mutants have been reported to
accumulate significant amounts of SAM92 while ethionine-resistant E. coli overproduces
methionine.93 Additional research has disclosed that a mutation in the metJ locus in E. coli
is responsible for the ethionine resistance.94 Met] protein is a global regulator in
methionine biosynthesis. Mutation of metJ results in derepression of methionine
biosynthesis to afford increased in vivo synthesis of methionine (in E. coli) or SAM (in S.
cerevisiae). Mutations in the met] locus can be easily identified by the resistance these
mutations impart towards ethionine. Creation of a KL7 metJ mutant could be a possible
solution for increasing SAM availability. Alternatively, a metJ mutation could be combined
with MetK overexpression. In the meantime, DNA shuffling can be used to obtain a
feedback-insensitive COMT and/or to improve the properties of COMT. The gene product
of COMT has a Km of 136 1.1M and a Vmax of 4.6 umol L'l min1 for vanillic acid95 and a

turnover number of 24 per min.

69

CHAPTER 3

FED-BATCH FERMENT OR SYNTHESIS OF 3-DEHYDROSHIKIMIC ACID USING
RECOMBINANT ESCHERICHIA COLI

IntreducuLQn

Microbe-catalyzed conversion of D-glucose into bioproducts such as L-
phenylalanine and L-tryptophan are intimately related to microbe-catalyzed syntheses of
industrial chemicals such as adipic acid and catechol from D-glucose (Figure 36).96 For
example, the carbon flow which is directed into the common pathway of aromatic amino
acid biosynthesis is of central importance to product yields and titers in each of these
syntheses. Carbon flow directed into this common pathway can be conveniently measured
by the accumulation of 3-dehydroshikimic acid (DHS) in the culture supematants of
microbial mutants lacking shikimate dehydrogenase activity (Figure 36). DHS is of
particular importance since it is the most advanced common pathway intermediate shared by
aromatic amino acid biosynthesis and biocatalytic syntheses of adipic acid, catechol,
vanillin and gallic acid.37’ 40. 45’ 35 DHS synthesis is also important in its own right given
the potent antioxidant activity which has been reported for this hydroaromatic}14

As the first enzyme in the common pathway of aromatic amino acid biosynthesis,
DAHP synthase activity dictates the amount of cellular carbon directed into DHS synthesis.
Historically, transcriptional repression and feedback inhibition of DAHP synthase by
aromatic amino acids have been viewed as the regulatory mechanisms that control the in
vivo catalytic activity of this enzyme. More recently, the intracellular concentrations of
substrates phosphoenolpyruvate (PEP) and D-erythrose 4-phosphate (E4P) have come
under scrutiny as critical determinants of in vivo DAHP synthase activity. Intracellular E4P
availability was been reported to be markedly influenced by expression levels of the

enzyme transketolase.47a' 553 Overexpression of PEP synthase recycles pyruvic acid back

70

[0

U3

bloc
com
and
con
0056
glue

The

 

 

to PEP resulting in increased carbon flow directed into the common pathway as long as
transketolase is overexpressed.47

Biocatalytic syntheses of common pathway intermediates have been examined
under shake-flask conditions. Such culturing conditions are problematic. Cells are in a
glucose-rich environment at the beginning of shake-flask cultures but experience glucose-
deficient conditions by the end of shake-flask cultures. Oxygenation levels and pH are also
difficult to control under standard shake-flask conditions.47 Cultivation of E. coli under
fed-batch fermentor conditions offers a number of important advantages. Variables such as
temperature, pH, dissolved oxygen levels and the steady-state concentration of glucose can
be controlled. Greater aeration is provided by the fermentor stir sparger and the impeller.
This increased aeration results in much higher cell densities and growth rates for E. coli
cultured under fed-batch fermentor conditions relative to shake-flask conditions.97

In this chapter, the synthesis of DHS is examined by a series of recombinant E. coli
biocatalysts under fed-batch fermentor conditions where oxygenation levels and pH are
controlled and where glucose availability is maintained at a constant, limiting level. Titers
and yields for synthesized DHS are compared with previously reported syntheses of
common pathway intermediates in shake flasks. Transketolase overexpression was
observed to have a pronounced impact on the yields and titers of DHS synthesized from
glucose under fed-batch fermentor conditions even in lieu of PEP synthase overexpression.

Theoretical maximum yields for product DHS and cell mass formation are also discussed.

71

COZH

 

 

Won-ck
CO2H PEP L-phenylalanine
% OH O L-tyrosine
H 0
H020 ’- 3P04/K5/il-1LH L-tryptophan
adipic acid E4P
TH AroF T
2- AroG T
PVC AroH
COZH HO ., COzH C02H
' 23 II
| ; OH ; O COZH
HOZC H203PO OH OH
cis,cis-rnuconic DAHP chorismic acid
aCId
T CatA lAroB T AroC
Ho, COZH CO?”
HO’Q OQOH H203PO" ; OJiCOz"l
OH OH OH
catechol DHO EPSP
TAroY lAroD T No“
AroZ
HO ‘ H203PO" : OH
OH OH
PCA 33"
\ / \ AroL
\ / NOE AroK
COZH COZH
I: (1
HO OH HO' ; OH
OH OH
gallic acid shikimic acid

Figure 36. DHS as the most advanced intermediate shared by.
aromatic amino acid biosynthesis and biocatalytic synthesis of
adipic acid and catechol.

72

danc
inserti
aroF’r
mean
absent
conveI
supern
mmnm
V'lldml]
Rubs
damn

dihydrc

from in
encode,
into prc
DAHP

u"IllCh a

-301“. -._I,-_ ‘9-3- _I,_'III‘I0 91001 0-. ._ i 's

A. Shared Genomic and Plasmid Elements

All of the DHS-synthesizing biocatalysts shared several genetic and recombinant
elements including a mutation in the genomic aroE locus, a second copy of the aroB gene
inserted into the genomic serA locus, plasmid-localized serA and plasmid-localized
aroFFBR. E. coli AB2834, an aroE mutant lacking shikimate dehydrogenase activity, was
the ancestral strain used to construct the KL3 host used in all of the DHS syntheses. The
absence of catalytically-active, aroE-encoded shikimate dehydrogenase, which catalyzed the
conversion of DHS into shikimic acid, resulted in the accumulation of DHS in the culture
supematants of E. coli KL3. Growth of E. coli AB2834 requires supplementation with
aromatic amino acids for protein biosynthesis along with supplementation with aromatic
vitamins for biosynthesis of folic acid, coenzyme Q, and enterochelin. Aromatic amino
acids supplements include L-phenylalanine, L-tyrosine, and L-tryptophan while aromatic
vitamin supplements consisted of p-aminobenzoic acid, p-hydroxybenzoic acid, and 2,3-
dihydroxybenzoic acid.

Because of the increased carbon flow directed into the common pathway resulting
from increased in vivo activity of DAHP synthase, wild-type expression levels of aroB-
encoded DHQ synthase is inadequate in E. coli AB2834 for conversion of substrate DAHP
into product DHQ at a rate which is sufficiently rapid to avoid substrate accumulation.
DAHP undergoes dephosphorylation to 3-deoxy-D-arabino-heptulosonic acid (DAH)
which accumulates in the culture supernatant resulting in reductions in the titer, yield, and
purity of synthesized DHS. An approximately 2-fold increase in DHQ synthase activity,
which can be accomplished by introduction of a second copy of aroB into the genome of E.

coli AB2834, is required to eliminate DAH accumulation.69

Sphl

EcoRl

 
   
   
    

rep (ts)

EcoRl \

 

 

 

 
     
   

serA pKAD76A
EcoRI
pJB14
8 hi
p Hindlll
EcoRi partial digest $500M digest
EcoRl Sphl EcoFll E oFll Hind ll Sphl EcoFll EooFli EooHl
1 JV 1 1.6 kb
serA a... ._. v serA aroB
rep (ts) Cm
T4 Ligase
Sphl EcoRl
EcoFll
EcoRi
Cm
aroB ‘ pKL3.82A

9.0 kb

EcoFll

serA .

Figure 37. Preparation of Plasmid pKL3.82A.

74

The genomic modification in E. coli KL3 responsible for increased DHQ synthase
expression resulted from site-specific insertion of a cassette consisting of aroB with
flanking serA nucleotide sequences into the serA locus of E. coli AB2834. This was
achieved through homologous recombination. Localization of the serA gene in
temperature-sensitive plasmid pMAK705 followed by insertion of the aroB into an EcoRI
site internal to the serA directed recombination of the aroB into the serA locus of the
genome (Figure 37). Digestion of pKAD63 with Sphl liberated a 1.9-kb serA fragment,
which was subsequently inserted into the Sphl site of pMAK705 to afford pKAD76A. The
aroB gene was obtained as a 1.6-kb fragment following digestion of pJB14 with EcoRI.
Insertion of the aroB fragment into the EcoRI site of serA was complicated by two
additional EcoRI sites in pKAD76A. Following EcoRI partial digestion of pKAD76A, the
resulting DNA fragments were resolved on an agar gel and the 7.4-kb fragment
corresponding to the linearized plasmid was isolated. Ligation of the linearized plasmid to
the 1.6-kb EcoRI fragment of aroB afforded pKL3.82A (Figure 37). Temperature-
sensitive plasmid pKL3.82A was used for homologous recombination. Conditions for
homologous recombination98 were identical to those described in Chapter 2 of this thesis.
KL3 was isolated based on the following growth characteristics: growth on M9 containing
L-tyrosine, L-phenylalanine, shikimic acid and serine; no growth on M9 containing L-
tyrosine, L-phenylalanine, shikimic acid; growth on LB; and no growth on LB containing
Cm (Table 6).

Disruption of the genomic serA locus was also the basis for plasmid maintenance.
Gene serA encodes D-3-phosphoglycerate dehydrogenase, which is the first of the three
enzymes responsible for serine biosynthesis99 (Figure 38). Since KL3 is a serine
auxotroph, growth in medium lacking L-serine supplementation required expression of
serA localized in all plasmids carried by KL3 biocatalysts. Using this nutritional pressure
for plasmid maintenance avoided the necessity of using antibiotics in fermentation medium.

However, all of the DHS-synthesizing plasmids carried markers encoding antibiotics

75

because rich medium with added antibiotic was used as the inoculum for all the

fermentations discussed in this Chapter.

Table 6. Plate selection for characterization of genome insertion strain.

 

 

 

 

 

M9/glucose/
Strain figfigfgz’ Tyr/Phe/SA/ LB/Cm LB
y senne
A32834 + + — +
KL3 — + _ +
SA: shikimic acid
'03PO NCOO H O NCOO
OH NH;
3-phosphoglycerate serine

NAD+ Pl
serA serB
NADH Glutamate a-Ketoglutarate H20

- COO' - ‘

03PO/\n’ \A oaPo’chzo
O serC NHa

3-phosphohydroxypyruvate 3-phosphoserine

Figure 38. Serine biosynthesis. Genetic loci are as follows: serA, 3-

phosphoglycerate dehydrogenase; serC, 3-phosphoserine aminotransferase;

serB, 3-phosphoserine phosphatase.

In E. coli, the most important regulation of DAHP synthase is feedback inhibition
of the enzymes by aromatic amino acids.100 All DHS-synthesizing constructs therefore
employed a mutant isozyme of DAHP synthase, designated as aroFFBR, which was
insensitive to feedback inhibition by aromatic amino acids. This mutant isozyme was

obtained by photochemical mutagenesis of an E. coli AB2.24, which expressed only the

genome-encoded, tyrosine-sensitive isozyme (AroF) of DAHP synthase. The aroFFBR

76

gene was isolated from a mutant selected as a result of its more rapid growth in a diffusion
gradient chamber against an increasing concentration of m-flurotyrosine.101 Sequencing of
the isolated aroFFBR gene revealed a Pro-148 to Leu-148 point mutation which
corresponds to a previously reported AroF mutant isozyme insensitive to feedback

inhibition by L-tyrosine.102

B. Fed-batch fermentor conditions

Fed-batch fermentation was performed in a 2.0 L capacity Biostat MD B-Braun
fermentor connected to a DCU system and a Compaq computer equipped with B-Braun
MFCS software for data acquisition and automatic process monitoring (Figure 39). The
temperature, pH and glucose feeding were controlled with a PID controller. The
temperature was maintained at 37 °C, and pH was maintained at 7.0 by addition of
concentrated NH4OH or 2 N H2804. Dissolved oxygen was maintained at 20% air
saturation throughout the fermentation process using a Braun polarographic probe.
Antifoam (Sigma 204) was added manually as needed.

Inoculants were grown in 100 mL LB medium (enriched with 2 g glucose)
containing the appropriate antibiotic for 12 h to 14 h at 37 °C with agitation at 250 rpm.
The inoculants were then transferred to the fermentor. The initial glucose concentration in
the fermentation varied from 18 g/L to 23 g/L according to the growth rates of different
constructs. Three different methods were used to maintain dissolved oxygen (D.O.) level
at 20% air saturation during the course of each fed-batch fermentor run. After inoculation
of the fermentor solution containing inorganic salts, aromatic amino acids, aromatic
vitamins, and a quantity of glucose, D.O. was maintained by increasing the impeller
stirring rate until a preset maximum value (940 rpm) was reached. Approximately 10 h
were required before the impeller reached its maximum stirring rate. The mass flow
controller then maintained D.O. levels at 20% saturation at the constant impeller stirring rate

by increasing the airflow rate until a preset maximum value (3.0 LIL/min) was reached.

77

Approximately 2 h were needed for the airflow to increase to its maximum rate. At a
constant impeller stirring rate (940 rpm) and constant airflow (3 LIL/min), D.O. levels were
then maintained at 20% saturation by oxygen-sensor-controlled glucose feeding for the rest
of the fermentation. At the beginning of this stage, dissolved oxygen levels fell below 20%
saturation due to residual glucose in the medium. This lasted for approximately 1 h before

all residual glucose was consumed and the glucose feeding started.

 

 

 

 

 

 

 

 

 

 

 

 

 

Ti”?
Midi

 

 

 

 

 

w ' V

 

Figure 39. B. Braun Fermentor.

The concentration of the glucose added to fermentor runs was 60% (w/v). The PID
control parameters were set to 0.0 (off) for the derivative control (To). 999.9 5 (minimum
control action) for the integral control (ti), and 950.0% for the proportional band (X9).

Oxygen sensor control of glucose usually became too difficult to maintain at about 48 h into

78

a fermentation run. Loss of oxygen sensor control was characterized by unregulated
addition of glucose to the fermentor culture.

Fermentor runs typically entered logarithmic growth 6 h after inoculation. After
approximately 24 h, fermentor cultures moved from a logarithmic to a stationary growth
phase. Microbial cell density normally reached a maximum of 25-30 g/L dry cell weight.
Over the course of the fermentations, the culture solution turned progressively darker. By
the end of all of the fermentor runs, the culture solutions were always a deep black color.
Acetic acid accumulation was observed at 6 h and at the conclusion of fermentor runs
which corresponds to early logarithmic and late stationary microbial growth phases,
respectively. Concentrations of acetic acid declined or were absent for most of the
logarithmic and early stationary microbial growth phases. Maximum productivity in DHS
synthesis generally started at 12 h and continued until 48 h. DHS synthesis typically did

not continue beyond 48 h.

79

D r s i

The key determinant of carbon flow directed into a biosynthetic pathway is often the
in vivo activity of the first enzyme in the pathway. For the common pathway of aromatic
amino acid biosynthesis, the in vivo activity of DAHP synthase is dictated by feedback
inhibition, transcription repression, and the availability of substrates E4P and PEP. E. coli
uses three different isozymes of DAHP synthase encoded by aroF, aroG, and aroH which
are feedback-inhibited, respectively, by L—tyrosine, L-phenylalanine, and L-tryptophan.
Feedback inhibition was circumvented in the DHS synthesizing strains by use of aroFFBR
which is obtained via photochemical mutagenesis of aroF. Choice of aroFFBR, as opposed
to use of feedback-insensitive mutants of other DAHP synthases, followed from previous
employment of aroFFBR to achieve the highest titers thus far reported for microbial
synthesis of L-phenylalanine under fed-batch fermentor conditions.54

The first DHS-synthesizing biocatalyst was created by transforming plasmid
pKL4.33B into the KL3 host. Plasmid pKL4.33B includes a copy of aroFFBR under the
transcription control of its native promoter and a copy of serA in a pSU18 vector (Figure
19, 20).71 This vector is has a p15A origin of replicon with a copy number of
approximately 12 per cell. Vector pSU18 also contains a lac promoter and a genetic marker
encoding Cm resistance. Plasmid pKL4.33B was created as follows: After UV
mutagenesis and chemotactic selection using a Diffusion Gradient Chamber, the aroFFBR
locus was amplified by PCR with its native promoter to yield a 1.3-kb fragment. Inclusion
of EcoRI recognition sequences at the 5’- and 3’- ends of the aroFFBR fragment facilitated
its insertion into the EcoRI site of pCL1920 to afford pCL2-13A. Plasmid pCL2-13A was
digested with EcoRI and the 1.3-kb aroFFBR fragment was isolated and ligated into the
EcoRI site of pSU18 to create pKL4.20B. The aroFFBR gene is transcribed in the opposite
orientation relative to the lac promoter in pKL4.20B (Figure 19). Ligation of a 1.9-kb

Dral/EcoRV fragment encoding serA obtained from pD2625 into the Smal site of the

80

pKL4.20B afforded pKL4.33B. The serA gene is transcribed in the same orientation as

 

     
   

 

 

 

 

aroFFBR (Figure 20).
3 4° —D—cells (g/L)
g 35-~ +DHS (g/L)
g 30 -
D
- 25 -
:i
E 20 1
a 15 -
'O
3 1O 1
e .
8 5
o i 4 : i i
o 12 13 24 so 36 42 48
time (hour)

Figure 40. KL3/pKL4.33B synthesis of DHS under fed-batch

fermentor conditions.

KL3/pKL4.33B was examined under fed-batch fermentor conditions. The
fermentation reached stationary phase at around 24 h. DHS was synthesized from 12 h to
36 h after which DHS production leveled off (Figure 40). After 48 h, KL3/pKL4.33B
produced 20.3 g/L DHS in a 17% (mol of DHS synthesized /mol of glucose consumed)
yield. DHQ (1.0 g/L) and gallic acid (0.5 g/L) also accumulated in the fermentation
medium. A total yield, which includes all of the accumulated common pathway
intermediates and their derivatives, was used to gauge carbon flow directed into the
common pathway. The total yield of KL3/pKL4.33B fermentation is 18% (mol/mol)
including DHS, DHQ and gallic acid (Table 7). DAHP synthase specific activity was also
monitored at four different time points during the 48 h fermentation (Table 7). It started at
0.118 U/mg at 12 h and increased to 0.228 U/mg at 24 h. At 36 h, DAHP synthase
specific activity decreased significantly to 0.014 U/mg and 0.018 U/mg at 48 h (Table 7).

The coincidence between the low DAHP synthase activity and the leveling off of DHS

81

synthesis after 36 h indicated that DAHP synthase activity might be the limiting factor in
achieving higher DHS titers and yields.

Table 7. Product titers and yield after 48 h fermentation for
KL3/pKL4.33B

 

 

I DH S] DHS DHQ GA T0131 DAHP synthase
yield [ l l 1 yield specific activity

(g/L) (mol/moi) (g/L) (g/L) (monol) 12h 24h 36h 48h

 

20.3 17% 1.0 0.5 18% 0.118 0.228 0.014 0.018

 

 

One approach taken to increase the levels of DAHP synthase expression entailed
replacement of the native promoter of aroFFBR with a strong Pmc promoter. Use of a
stronger promoter and the attendant tighter binding of RNA polymerase leads to increased
transcription. However, DAHP synthase overexpression can reach a level where cell
growth and metabolism are compromised. Inclusion of lale in the same plasmid was
designed to control this trade-off. The lacIQ gene product is the Lac repressor protein,
which binds to the lac operator DNA sequence. Since the lac operator is included in the
Pm promoter region, Lac repressor encoded by plasmid-localized lacIQ will repress the
transcription of the gene controlled by the Pm promoter. However, binding of lactose to
the Lac repressor can cause a conformational change of the repressor so that the repressor
no longer binds to the operator. Transcription of the gene under the control of Pmc is thus
derepressed. Clearly, the advantage of this Ptao/lacIQ system is the ability to modulate the
concentration of the inducible gene products in the cell by varying the lactose inducer
concentration. The frequently used inducer is the nonhydrolyzable analog of lactose,
isopropyl B-D—thiogalactopyranoside (IPTG). AroF specific activity was thus controlled
by the amount and frequency of IPTG added to the fermentor medium.

Plasmid pKL4.79B was created for the inducible, controlled expression of

FBR

aroF . First, the open reading frame (ORF) of the aroFFBR was amplified from

82

pKL4.33B using following primers: 5’- GGAATTCATGCAAAAAGACGCGCTGA and
5’- GGAATTCTTAAGCCACGCGAGCCGT. Localization of the resulting 1.1-kb
fragment in pJF118EH afforded pKL4.71A (Figure 41). Cloning vector pJF118EH
contains a tac promoter and lacIQ gene.103 With the copy number of about 15 per cell, it
also contains a pMBI origin of replicon and an Ap resistant marker. In pKL4.71A, the
ORF of aroFFBR is transcribed in the same orientation as the tac promoter to ensure the
proper transcription and translation of aroFFBR (Figure 41). Further ligation of a 1.9-kb
Dral/EcoRV serA fragment obtained from pD2625 into the Smal site of the pKL4.71A
afforded pKL4.79B. The serA gene in pKL4.79B is transcribed in the opposite

orientation relative to the arol"1r BR gene (Figure 42).

Table 8a. DAHP synthase activities (umol/min/mg) when
aroFFBR expression is controlled by Pm.

 

DAHP synthase
Entry 'PTG specific activity
“91““ 12h 24h 36h 48h
a1 b0.0 0.009 0.011 0.012 0.007
’2 ”0.32 0.11 0.098 0.13 0.048
as b1.6 0.10 0.33 0.13 0.12
84 b4.6 0.24 0.30 0.27 0.42
as b13.0 4.0 2.8 1.2 1.1
8‘6 t’40.0 1.5 1.5 0.91 1.2
a7 °1000 3.0 2.2 0.29 0.21
a KL3/pKL4.79B

9 Amount of IPTG (mg) added at 12, 18 24, 30, 36, and 42 h.
0 Amount of IPTG (mg) added at 4 h.

Optimization of IPTG addition of KL3/pKL4.79B fermentation led to several
trends. During addition of IPTG at timed intervals, 3 maximum specific activity of DAHP
was achieved after which an additional increase in IPTG concentration resulted in a decline

in DAHP synthase specific activity (entry 5. Table 8a). A high specific activity for DAHP

83

synthase could also be achieved by a single addition of a relatively large quantity of IPTG

early in the fermentor run (entry 7, Table 8a).

Table 8b. Product titers and yields when aroFFBR expression
is controlled by P1“.

DHS Total

 

Entry (g/L) (mol/moi) (g/L) (g/L) (mol/moi)
a1 24.3 13% 1.4 0.6 14%
a2 29.1 16% 2.6 1.5 18%
as 34.0 21% 2.0 1.4 23%
a4 52.0 20% 6.3 3.7 24%
as 26.8 12% 1.8 1.7 14%
a6 22.0 11% 1.0 1.4 12%
a7 17.4 12% 0.1 0.2 13%

a KL3/pKL4.79B

AroF is precedented104 to be labile to protease activity during the stationary phase
of E. coli growth. However, DAHP synthase specific activities were stable over the course
of most of the fed-batch fermentations where IPTG was added at timed intervals (entries 2-
4 and 6, Table 8a). This sharply contrasts with the decline in DAHP synthase activity
observed overtime with KL3/pKL4.33B where aroFFBR expression was under its native
promoter (Table 7). These differences may suggest that a Ptac but not a Pam]: promoter
allows for aroFFBR transcription during stationary phase. Genes expressed from a Plac
promoter and a Ptrp promoter continue to be transcribed during stationary phase in E.
coli.105 There does appear to be a ceiling above which AroF specific activities can not be
maintained even when aroFFBR is under Pm promoter control. This is evident in the
pronounced decline in enzyme activity observed for the two IPT G addition regimes
resulting in the highest DAHP synthase activities (entries 5 and 7, Table 8a) measured at 12

h into the fermentation runs.

84

EcoRl Smal Hindlll

 

 

 

 

 

pKL4.33B
pJF118EH
l PCR .. 5.3 kb
EcoFil EcoRl
1.1 kb
aroFFBR (ORF)
. 1) EcoRl digest

lEG‘ORI dIgest i 2) CIAP treatment

T4 Ligase

 

EcoFll
Smal
Hindlll

pKL4.71A
6.4 kb

 

Figure 41. Preparation of Plasmid pKL4.71A.

85

EcoFll
Smal
Hindlll

p02625 - pKL4.71A
6.4 kb

i EcoRV / Dral digest

 

 

 

EcoRV Dral
| 1.9 kb J
serA '
1) Smal digest
2) CIAP treatment

T4 Ligase

      
 
  

 
  
 

8.3 kb
‘—Ap/ (Smal)

Hindlll

Figure 42. Preparation of Plasmid pKL4.79B.

86

   
  
    

1°“ pKL4.33B
pKL4.338 5.5 kb (Smal)
i PCR

Xbal Xbal -

1.3 kb Hindlll

Xbal (Smal)
emf-FER
. 1) Xbal digest
leal d'995t l 2) CIAP treatment

T4 Ligase

 

Figure 43. Preparation of Plasmid pKL4.66A.

87

   
  
    

   
  

pMFaaA 0n pKL4.ssB
5-5 kb (Smal)
1 PCR
0.1 kb
Xbal Xbal Hindlll
Laud Xbal (Smal)
ParoF
. 1) Xbaii digest
leal digest i 2) CIAP treatment
T4 Ligase
EcoRl
EcoRl
{Cm pKDI1.291A 13””
Hindlll
Xbal
Xbal (Smal)

Figure 44. Preparation of Plasmid pKDll.29lA.

88

Incremental increases in added IPTG concentration led to corresponding
improvements in DHS titers and yields (Table 8a). The highest DHS titer and yield was
achieved when 4.8 mg of IPTG was added at 12, 18, 24, 36, and 42 h. KL3/pKL4.79B
synthesized 52.0 g/L DHS in a 20% (mol/mol) yield (entry 4, Table 8a and 8b). DHQ (6.3
g/L) and gallic acid (3.7 g/L) were also synthesized leading to a total yield of hydroaromatic
and aromatic products of 24% (mol/mol). Beyond this optimal IPT G addition, addition of
higher IPTG concentration resulted in a pronounced decrease in DHS titers and yields
(entries 5 and 6, Table 8a and 8b). This cut-off point appears to correspond to a ceiling in
DAHP synthase activity above which further increases in DAHP synthase activity levels
may diminish the biocatalyst's ability to synthesize DHS.

As shown in Table 8a and Table 8b, higher than optimal DAHP synthase
expression levels results in a significant decrease in DHS titers and yields. This is
probably due to the competition of PEP availability between DAHP synthase and
phosphotransferase (PTS) mediated glucose uptake. Too much DAHP synthase drives
PEP into the common pathway, thereby leading to a diminish rate of glucose uptake by the
cell. This is consistent with the observation that cells grow slower when IPTG additions
exceed 4.8 mg per time point. The increased metabolic burden imposed on the cell's
metabolism by excessive enzyme overexpression may also be responsible for the decline in
DHS synthesis observed when DAHP synthase expression exceeds an optimal level.

Because of the anticipated need to use strong promoters such as Pm for amplified
expression of enzymes other than DAHP synthase, options were explored for exploiting
the native promoter of aroFFBR for amplified expression of this gene. In E. coli, the tyrR
gene product represses transcription of aroF. Upon plasmid localization, the increased
copies of aroFFBR with its unmodified native promoter should titrate away the cellular
supply of TyrR which binds to the operator region of aroFFBR. Some percentage of the
aroFFBR promoters thus evade TyrR binding and are derepressed. This was the reason for

comparing localization of one aroFFBR locus per plasmid, two aroFFBR loci per plasmid,

89

and one aroFFBR locus accompanied by one Pam): promoter per plasmid. The strategy
where only I’m-0F, the promoter region of aroFFBR, was inserted into the multicopy plasmid
along with a single copy of aroFFBR was designed to achieve derepression of aroFFBR
while reducing the amount of DNA contained in the multicopy plasmid.

Table 9. Product titers and yield synthesized by KL3/pKL4.66A after
48 h

 

 

Total DAHP th
[DHS] yield [DHQ] [GA] yield speciiicfaratitfitsye
(g/L) (monoi) (g/L) (g/L) (monol) 12h 24h 36h 48h

 

38.5 16% 2.2 3.2 18% 0.41 0.14 0.077 0.11

 

 

Table 10. Product titers and yield synthesized by KL3/pKDll.291A
after 48 h

 

 

Total DAHP synthase
[DHS] yield [DHQ] [GA] yield specific activity
(g/L) (mol/moi) (g/L) (g/L) (mol/mol) 12h 24h 36h 43h

 

41.2 18% 2.9 2.6 21% 0.059 0.06 0.072 0.054

 

 

Plasmid pKL4.66A was created to contain two copies of aroFFBR. The aroFFBR
locus was amplified by PCR from pKL4.20B with Xbal ends. Localization of the 1.3-kb
aroFFBR into the Xbal site of pKL4.33B resulted in pKL4.66A. Transcription of the
aroFFBR locus in the Xbal site of pKL4.66A proceeds in the opposite orientation of serA
(Figure 43). Plasmid pKDl 1.291A was created to contain a copy of aroFFBR and a copy
of Pump. This 5.6-kb plasmid was constructed by inserting a fragment encoding the
promoter region of aroF into the Xbal site of pKL4.33B. Pam]: was amplified from

pMF63A using following primers: 5’- GCT CT AGAGAA'I'I‘CAAAGGGAGTGTA and 5’-

90

GCTCTAGACCTCAGCGAGGATGACGT. Transcription from Pam; is in the same
orientation as the serA gene (Figure 44).

Both KL3/pKL4.66A and KL3/pKDll.29lA were examined using the same
fermentation conditions. KL3/pKL4.66A expressed significantly higher DAHP synthase
specific activities relative to KL3/pKL4.33B where only a single copy of aroFFBR is
plasmid-localized (Table 7, 9). DAHP synthase specific activity was lower for
KL3/pKDll.291A compared to that for KL3/pKL4.66A (Table 10). However, this
biocatalyst's DAHP synthase specific activities remained reasonably constant over time
(Table 10). Both KL3/pKL4.66A and KL3/pKDl 1.291A produces much higher DHS
titers and similar DHS yield relative to KL3/pKL4.33B. KL3/pKL4.66A synthesized 38.5
g/L of DHS in 16% (monol) yield while KL3/pKD1 1.291A synthesized 41.2 g/L of
DHS in 18% (monol) yield. Taking into consideration the synthesized DHQ and gallic
acid, total yields of hydroaromatic and aromatic products were 18% (monol) and 21%

(monol) for KL3/pKL4.66A and KL3/pKDl 1.291A, respectively (Table 9, 10).

91

OV‘,‘.-L‘_lut .‘olo.-‘ U‘I' 'n-Ba I ‘lll‘l'. 0.13:0

As described in the previous section, several different strategies were explored for
circumventing transcriptional repression including placing plasmid-localized aroFFBR under
Pm control , employing two plasmid-localized copies of aroFFBR, and insertion of Pump in
the same plasmid containing aroFFBR. However, higher DAHP synthase expression levels
did not necessarily translate into higher DHS yields or titers as evidenced by the
relationship between DAHP synthase specific activity and DHS synthesized when plasmid-
localized aroFFBR was under Pm control (Table 8a, b). Even though DAHP synthase
specific activity and DHS synthesis initially increased with increased IPTG concentration, a
point was reached where further increases in IPTG concentrations and associated
improvements in DAHP synthase activity did not result in increased DHS titers or yields.
Indeed, the IPTG addition regimens leading to the three highest DAHP synthase specific
activities resulted in DHS titers which were no more than 52% (monol) of the DHS titers
observed for the best DHS synthesis when aroFFBR was under Pm control (Table 8a, b).

E4P availability is another factor likely limiting in vivo DAHP synthase activity.
Using shake-flask experiments, Draths et al. discovered that the E4P availability was an
important limiting factor in aromatic amino acid biosynthesis.“a Steady-state
concentrations of E4P are low even in the absence of amplified DAHP synthase activity.106
E4P is prone to dimerization, trimerization, and polymerization.107 Dissociation of these
E4P forms back to monomeric E4P is quite slow. By closely matching the rate of E4P
synthesis with the rate of E4P utilization, the resulting low, steady-state concentration of

E4P may be N ature's in vivo strategy favoring the monomeric form of E4P.

92

     
   

    
     
 
    
   
 

EcoRl
(Smal)
pM F51 A
pKL4.793
l BamHI digest 33 kb 5904
BamHI BamHI A p
(Smal)
Hindlll
tktA

2) Klenow treatment

lKlenow treatment
3) CIAP treatment

¢ 1) Hindlll digest

T4 Ligase

serA

pKL4.124A

10.5 kb
(Smal)

(Hindlll)

w

(Hindlll)

Figure 45. Preparation of Plasmid pKL4.124A.

93

pMF51A

 

¢ BamHI digest
BamHI BamHI
2.2 kb
tktA

2) Klenow treatment

lKlenow treatment
3) CIAP treatment

i 1) Hindlll digest

T4 Ligase

EcoRl

(Hindlll) pKL4.1308 serA

 

Xbal

Figure 46. Preparation of Plasmid pKL4.130B.

94

aroFFER

EcoFtt

    

 

 

 

pKD11.291A 15”"
Sphl Xbal (Smal)
1) Sphl/Neal digest 1) Sphl/Neal digest
2) get purification 2) gel purification
Ncol SP” Ncol Sphl
L—fl I We fi‘ri
a“ MA 3: tilroFFBR serA Pam).-

i i

(Hindlll)

 

(Hindlll) Sphl

Figure 47. Preparation of Plasmid pKL5.17A.

95

Amplified expression of either transketolase or transaldolase108 has been
demonstrated to increase the in vivo availability of E4P in E. coli cultured under shake-
flask conditions. T ransketolase and transaldolase are enzymes in the nonoxidative pentose
phosphate pathway which facilitate the interconversion of C-7, C-6, C-5, and C-4 aldoses
and ketoses. In order to test whether E4P is a limiting factor at optimized DAHP synthase
expression levels under fed-batch fermentor conditions, the tktA gene was inserted in the
plasmid pKL4.79B to overexpress transketolase. Plasmid pKL4.124A was created for this
purpose. This 10.5-kb plasmid was constructed by inserting the tktA fragment from
pMF51A into pKL4.79B. Following digestion of pMF51A with BamHI, the 2.2-kb tktA
fragment was blunt ended using Klenow fragment. Plasmid pKL4.79B was digested with
Hindlll and also treated with Klenow fragment. Subsequent ligation of the tktA fragment
with linearized pKL4.79B afforded pKL4.124A. The tktA gene is transcribed in the same
orientation as the are BR gene (Figure 45).

Table 11a. DAHP synthase specific activities (umol/min/mg) in
TktA-overexpression constructs

DAHP synthase
specific activity
1 2h 24h 36h 48h

Construct

KL3/pKL4.124A8 0.27 0.12 0.092 0.062
KL3/pKL4.1308 0.23 0.11 0.08 0.075
KL3/pKL5.17A 0.037 0.031 0.033 0.025

a4.8 mg IPTG added at 12, 18, 24, 30, 36. 42h.
KL3/pKL4.124A synthesis of DHS was examined under fermentation conditions
where 4.8 mg IPTG were added at 12, 18, 24, 36, and 42 h. DAHP synthase activities

were also monitored during the fermentation run and were observed to decrease relative to

those measured for the non-tktA KL3/pKL4.79B (Table 8a, Table 11a). This was

96

probably due to the metabolic burden imposed by the expression of the additional gene
(tktA) localized in pKL4.124A. However, the titer and yield of DHS increased
significantly. KL3/pKL4.124A synthesized 66.3 g/L of DHS in 28% (monol) yield.
DHQ (6.0 g/L) and gallic acid (6.0 g/L) were also produced. The total yield of DHS,
DHQ, and gallic acid synthesized from glucose was 33% (monol) (Table 11b).

Table 11b. Product titers and yield in TktA-overexpression constructs

DHS Total
[DHS] yield [DHQ] [GA] yield

 

canatmat (g/L) (monol) (g/L) (g/L) (mol/moi)
KL3/pKL4.124A3 66.3 23% 6.0 6.0 33%
KL3/pKL4.1308 69.0 30% 6.3 6.6 36%
KL3/pKL5.17A 53.1 24% 3.6 4.6 27%

a 4.8 mg IPTG added at 12, 18, 24, 30, 36, 42h.

The impact on synthesized DHS by amplified transketolase was also examined
when aroFFBR expression was under the transcription control of its native promoter.
Insertion of tktA gene into pKL4.66A and pKDll.29lA resulted in pKL4.130B and
pKL5.17A, respectively. Plasmid pMF51A was digested with BamHI and the resulting
2.2-kb tktA fragment was treated with Klenow fragment. Plasmid pKL4.66A was
digested with HindIII and treated with Klenow fragment. Subsequent ligation of the tktA
fragment with linearized pKL4.66A afforded pKL4. 130B. The tktA gene is transcribed in
the same orientation as the serA gene (Figure 46). The 7.8-kb pKL5.17A plasmid was
created by replacing the 1.0-kb NcoI/SphI fragment of pKDll.291A with a 3.2-kb
NcoI/Sphl fragment from pKL4. 130B that included tktA. Digestion of pKDl 1.291A with
NcoI and Sphl afforded a 4.6-kb fragment while similar digestion of pKL4.130B yielded a

97

3.2-kb DNA fragment. Ligation of these two purified fragments resulted in pKL5.17A
(Figure 47).

When two copies of plasmid-localized aroFFBR system were utilized for
overexpressing transketolase, DHS yields and titers went from a titer of 39 g/L and 16%
yield to a titer of 69 g/L and 30% yield upon plasmid-localization of tktA (KL3/pKL4.66A
versus KL3/pKL4.130B). A DHS titer of 41 g/L synthesized in 18% (mol/mol) yield
improved to a titer of 58 g/L synthesized in 24% (mol/mol) yield upon plasmid localization
of tktA along with aroFFBR and Pam]: (KL3/pKDl 1.291A versus KL3/pKL5.17A). The
highest concentrations of synthesized gallic acid (6.6 g/L) and DHQ (6.6 g/L) were again
observed for the biocatalyst that also synthesized the highest concentration of DHS (Table
11b).

Figure 48 and Figure 49 compare the profiles of the fermentations for
KL3/pKL4.66A and KL3/pKL4.13OB. Both fermentations yield similar cell mass.
However, compared to non-transketolase overexpressing construct KL3/pKL4.66A, the
transketolase overexpressing KL3/pKL4.130B produced DHS, DHQ and gallic acid at a
much higher rate. With transketolase overexpression, more acetic acid accumulated in both

the log phase and the late stationary phase of the fermentation.

98

 

-D—Cells (g/L)

30
l+oHs ( /L)

70‘
60-
50‘

j

 

I

 

    

 

Cells (9 dry wt/L), DHS (g/L)
.h
C

 

 

0 6 12 18 24 30 36 42 48

 

 

time (hour)
b
7T
I DHQ (g/L)
6 . a Gallic Acid (g/L
5 DAcetate (g/L)
4 -

Concentration (g/L)

 

 

 

 

0 612182430364248
time (hour)

Figure 48. KL3/pKL4.66A (a) DHS production and cell growth. (b)
DHQ, gallic acid, and acetic acid production.

99

 

8° —D—Ce|ls (g/L)
7o-r +DHS(/L) l
60--
50--
40--
30-- I
20--
10"?
0 .. . i : : i i
0 612182430364248
time (hour)

Cells (9 dry wt/L), DHS (g/L)

 

 

 

 

  

I DHQ (g/L)

          
 

:IGaIlic Acid (g/L)

I

D Acetate (g/L)

  
       

 

 

 

 
 
    

Concentration (g/L)

    
  

IIIIIIIIIIIIIIIIIIIIIIIIIIII
Q

  
   

 

lIIllIIlllllIIIlllllllllllllllllllllllll
-\

 
   
  

 

0 612182430364248
time (hour)

Figure 49. KL3/pKL4.130B (a) DHS production and cell growth.
(b) DHQ, gallic acid, and acetic acid production.

Discussing

A. Comparison of Titers and Yields

Deterrrrining109 the theoretical maximum yield for biocatalytic synthesis of DHS
begins with balancing (Eq. (1)) the PEP and E4P inputs with DHS product and
byproducts. The PEP and E4P inputs are then equated to the amount of D-glucose which is
required to form theses substrates (Eq. (2a)). Because E. coli relies on the carbohydrate
phosphotransferase (PTS) system for glucose uptake, a pyruvic acid term is included in
Eq. (2a) to reflect the conversion of one molecule of PEP into pyruvic acid for each
molecule of glucose transported into the cytoplasm as glucose 6-phosphate. Eq. (2a)
reflects the apparent absence of pyruvic acid recycling back to PEP which would be
catalyzed by the enzyme PEP synthase. PT S-generated pyruvic acid is considered to be the
primary carbon source for the anabolism and catabolism required for generation of E. coli

cellular biomass.

(1) PEP + E4P _, 2 H3PO4 + H20 + DHS

(2a) x glucose —> PEP + E4P + x pyruvic acid

(2b) x6(C) —> 3(C)+4(C)+x3(0)

s
) .. _J t
V l. w“? I! “I" y, 5‘ j“

'r

A coefficient is determined (Eq. (2b)) to balance the number of carbon atoms in the
glucose starting material with the total number of carbon atoms formed in PEP, E4P, and
pyruvic acid. The determined coefficient leads to a maximum theoretical yield of 43% (mol
DHS/mo] glucose) for synthesis of DHS from glucose. KL3/pKL4.130B, the construct
producing the highest yields and titers of DHS when aroFFBR was expressed from its
native Pam]: promoter, synthesized DHS at 70% of the theoretical maximum yield. DHS
was synthesized at 66% of theoretical maximum yield by KL3/pKL4.124A, which

produced the highest titer and yield of DHS when two copies of aroFFBR were expressed

101

from their Ptac promoters. In considering the performance of KL3/pKL4.130B and
KL3/pKL4.124A, synthesis of DHQ and gallic acid must also be taken into consideration.
DHQ is a metabolic precursor to DHS and gallic acid is a metabolite derived from DHS.
Including DHQ and gallic acid synthesis, KL3/pKL4.130B and KL3/pKL4.124A
channeled a total of 83% and 77%, respectively, of the theoretical maximum amount of

carbon which can be directed into aromatic amino acid biosynthesis.

(3) 2 pyruvic acid + 2 ATP + 2 NADH —> glucose
(4) glucose + 0.87 NADH + 14.8 ATP —-—-> 6 C-mole cells

(5) 2 pyruvic acid + 16.8 ATP + 2.87 NADH ——> 6 C-mole cells

A similar balancing approach was used to estimate the theoretical maximum E. coli
cell yield from the pyruvic acid, ATP, and electron equivalents generated by this pathway.
For every seven glucose molecules entering the pathway, 7 pyruvic acid, 10 reduced
N ADH, and 4 ATP molecules are produced. The stoichiometries for production of pyruvic
acid from glucose via glycolysis (Eq. (3)) and cell mass production from glucose110 (Eq.
(4)) can be summed to estimate the stoichiometry to cell mass production from pyruvic acid
(Eq. (5)). A fraction of the pyruvic acid produced must be converted to acetyl CoA and
then catabolizcd to C02 in the Krebs cycle to provide energy and electrons needed for cell
synthesis. Each FADH2 produced in the Krebs cycle was assumed to yield two ATP in the
electron transport chain, and each NADH was assumed to yield three ATP. Balanced
equations for pyruvic acid, cells, FADHz, NADH, and ATP were written, based on Eq.
(5) and the known stoichiometry of pyruvic acid oxidation via the Krebs cycle. The
pseudo-steady-state assumption was invoked for non-secreted intermediates.‘ 10 Solving
the balances simultaneously resulted in a cell yield of 15 C-mols of cells per seven moles of
glucose consumed. This number corresponds to 28% of the pyruvic acid derived from the

PTS being catabolizcd, and 72% being used for cell synthesis. The mass of one C-mol of

102

E. coli cells is 24.6 g, based on a measurement of the cell's elemental composition. Using
this value, the theoretical maximum cell yield on a mass basis is 0.29 g cell/g glucose.
Typically, the cell concentrations obtained in these fermentations were about 25 g/L, and
the glucose consumption levels were about 250 g/L. Thus, the experimental cell yields
were about 0.10 g cells/g glucose, which is well below the theoretical maximum value.
Microbial synthesis of DHS leading to a titer of 5.2 g/L and a yield of 54%
(mollmol) has previously been realized under shake-flask conditions.111 These
experiments employed overexpressed native aroF and tktA. Accumulation in the culture
supernatant of DAH, which results from dephosphorylation of DAHP, has been more
frequently employed as the measure of carbon flow directed into the common pathway of
aromatic amino acid biosynthesis. DAH has been synthesized under shake flask conditions
at a titer of 6.2 g/L and a yield of 63% (mollmol) by an E. coli construct overexpressing the
isozyme of DAHP synthase encoded by aroGFBR which is insensitive to feedback
inhibition by L-phenylalanine.47¢ e Overexpression of transketolase and PEP synthase in
addition to overexpression of aroGFBR led to a titer of 12.5 g/L of DAH synthesized in a
yield of 94% (mol/mol).47dt '3 All of these reported yields of DHS and DAH are
substantially in excess of the theoretical maximum yield of 43% (mollmol) and 86%
(mollmol) for DAH and DHS synthesis in the absence and presence, respectively, of PEP
synthase-catalyzed recycling of pyruvate back to PEP. Such yields reflect the shake flask
conditions employed for DHS and DAH syntheses where the E. coli construct is initially
grown in rich medium, harvested, and then resuspended in minimal salts medium
containing glucose where synthesis of DHS proceeds. Using fed-batch fermentor
conditions similar to those employed in this Chapter for DHS synthesis, an E. coli
construct overexpressing aroFFBR synthesized 46.8 g/L of L-phenylalanine in 20%

(mollmol) yield.54

103

B. DHQ and Gallic Acid Formation

Substantial concentrations of 3-dehydroquinic acid (DHQ) and gallic acid
accumulated during DHS synthesis. Various common pathway metabolites have been
reported to accumulate in the culture supematants of E. coli constructs in which carbon
flow directed into the common pathway of aromatic amino acid biosynthesis has been
substantially increased.69 These metabolites result from rate-limiting common pathway
enzymes which can not catalyze the conversion of substrate into product at a sufficiently
rapid rate to avoid substrate accumulation and subsequent export of these metabolites into
the culture supernatant. However, DHQ accumulation has not previously been reported
for E. coli biocatalyst despite the detailed scrutiny this microbe has received for rate-
limiting, common pathway enzymes.69 Gallic acid formation during DHS synthesis is also
surprising. Beyond its occurrence in plants112 and reported biosynthesis in the filamentous
fungus Phycomyces blakesleeanus,“3 gallic acid synthesis has not previously been
observed in E. coli or any other bacterium. For constructs KL3/pKL4.130B and
KL3/pKL4.124A, which synthesized DHS in the highest yields and titers, substantial
quantities of DHQ (6-7 g/L) and gallic acid (6-7 g/L) were produced.

Metabolite accumulation has typically been attributed to impeding, rate-limiting
enzymes when the common pathway is operating in the forward direction towards
biosynthesis of aromatic amino acids. For DHQ accumulation, it is possible that the aroD
encoded, DHQ dehydratase is rate-limiting under the fed-batch fermentor conditions.
However, an alternative explanation may be that synthesized DHS accumulating in the
fermentor culture medium is transported back into the cytoplasm where DHQ dehydratase
catalyzes the conversion of the DHS back into DHQ. DHS uptake by E. call may be
mediated by protein(s) encoded by the shiA locus114 which catalyze uptake of the
structurally similar shikimic acid. In addition, DHQ dehydratase is known to catalyze DHS
hydration.115 Further conversion of DHQ is unlikely since the next common pathway

enzyme, DHQ synthase, catalyzes an irreversible reaction which precludes conversion of

104

DHQ back into DAHP. Equilibration of DHS with DHQ is supported by the fact that molar
ratio of [DHS]/[DHQ] synthesized by DHS producer is similar in magnitude to the known
equilibrium constant (Keq = 15) for the reversible DHQ dehydration/DHS hydration
catalyzed by DHQ dehydratase. If DHQ accumulation is due to rate-limiting DHQ
dehydratase, expression of the aroD locus needs to be amplified. Increasing DHQ
dehydratase activity, however, would not reduce DHQ formation if this byproduct is the
result of microbial equilibration of initially synthesized DHS. Eliminating transport of
DHS from the culture medium into the cytoplasm would likely be a more productive
strategy. By cloning another copy of aroD in the plasmid and testing the new plasmid
under the same conditions, the reason of DHQ accumulation would be clarified according
to the changed (or unchanged) ratio of [DHS]/[DHQ].

Several different mechanisms can be offered for formation of gallic acid. Abiotic,
inorganic phosphate-catalyzed conversion of DHS into gallic acid is precedented.32
Alternatively, DHS could be enzymatically converted into gallic acid by DHS dehydratase-
catalyzed conversion of DHS into protocatechuic acid (PCA) followed by hydroxylation of
the PCA to form gallic acid. Oxidoreductase-catalyzed dehydrogenation of either the G4
or C-5 hydroxyl group of DHS could also lead to gallic acid formation since subsequent
aromatization would be expected to be spontaneous and rapid. Despite repeated attempts,
no conversion of DHS into gallic acid was observed in cell-free extracts to which a variety
of different cofactors were added. This lack of assayable enzyme activity could be
consistent with either an abiotic route for gallic acid formation or indicative of an enzymatic
process where the responsible enzymes do not survive cell lysis. Abiotic, inorganic
phosphate-catalyzed conversion of DHS into gallic acid is always accompanied by PCA
formation. The absence of PCA when gallic acid is formed during microbial synthesis of
DHS is thus not consistent with an abiotic conversion. Furthermore, much higher
concentrations of inorganic phosphate are required to catalyze abiotic conversion of DHS

into gallic acid than were employed during the fed-batch fermentor syntheses of DHS.34

105

The absence of PCA formation appears to be more consistent with oxidoreductase
oxidation of the C-4 or C-5 alcohol of DHS rather than dehydration of DHS and
hydroxylation of the intermediate PCA. However, the possibility remains that PCA might

be present in undetected low, steady-state concentrations.

C. E4P Limitation versus PEP Limitation

E4P and PEP availabilities are two possible limiting factors for directing more
carbon into the common pathway after DAHP synthase has been overexpressed.
Historically, examination of small molecule limitations to aromatic amino acid biosynthesis
has focused on PEP availability.116 Competition between PTS-mediate glucose uptake and
DAHP synthase for PEP is an obvious problem during biocatalytic synthesis of aromatics
and hydroaromatics. Other enzymes which employ PEP as substrate such as PEP
carboxylase, pyruvate kinase, PEP carboxykinase, and 3-deoxy-D-manno-octulosonate 8-
phosphate synthase add to this intracellular competitive fray. Although various strategies
have been examined for increasing in vivo PEP availability, amplified expression of PEP
synthase is particularly effective. PEP synthase catalyzes the conversion of pyruvic acid
into PEP at the expense of both of ATP's high-energy phosphodiesters (Figure 12).
Overexpressed PEP synthase enables PI‘S generated pyruvate to be recycled back into PEP
thereby ameliorating the competition between glucose uptake and aromatic amino acid
biosynthesis for intracellular PEP concentrations.

The impact of both E4P and PEP availability on the in vivo activity of DAHP
synthase has been examined by Liao and coworkers under shake-flask conditions.47dt ‘3
This was accomplished by using amplified expression of tktA-encoded transketolase and
pps-encoded PEP synthase activities. These experiments demonstrated that, when
aroGFBR, pps and tktA were all overexpressed, yields close to or exceeding the theoretical
maximum value were achieved for a common pathway intermediate used to gauge carbon

flow directed into the common pathway. Overexpression of feedback-insensitive aroGFBR

106

and pps-encoded PEP synthase in lieu of tktA overexpression did not increase the flow of
carbon directed into the common pathway of aromatic amino acid biosynthesis relative to
overexpression of just aroGF BR. Only a modest increase in common pathway carbon flow
was detected when aroGFBR and tktA were overexpressed in lieu of amplified pps
expression. These observations suggested that even in a biocatalyst environment
possessing ample PEP concentrations and overexpressed, feedback-insensitive DAHP
synthase, E4P availability was a critical factor which limited aromatic amino acid
biosynthesis.

DHS synthesis under fed-batch fermentor conditions clearly demonstrates the
pronounced effect of tktA overexpression even in the absence of PEP synthase
overexpression. This contrasts with the report47d» e by Liao that carbon flow directed into
the common pathway was only modestly increased by transketolase overexpression in the
absence of amplified PEP synthase expression. The impact of tktA overexpression on
DHS synthesis is an important observation since overexpression of PEP synthase inhibits
growth of E. coli. In actively growing E. coli cultures, this growth inhibition negates the
positive impact PEP synthase has on titers and yields of synthesized common pathway
intermediate.47dv ‘3

An important difference between this study and the previous Liao study may be
D.O. control of glucose addition during fed-batch fermentor synthesis of DHS relative to
the shake flask culturing conditions employed by Liao. The steady-state concentration of
glucose was maintained at approximately 200 11M during most of the fed-batch fermentor
synthesis of DHS as a result of employing DC. to control glucose concentration. By
contrast, glucose concentrations are not maintained at a constant level during shake-flask
cultures. Such cultures begin in a glucose-rich environment and end in a glucose-deficient
environment. D.O. levels are also maintained at a constant value over the entire course of
the fed-batch fermentor runs, whereas D.O. levels are uncontrolled in shake-flask cultures.

Titers and yields of DHS synthesized in this Chapter may therefore highlight a fundamental

107

difference in the in vivo availability of PEP in E. coli cultured under fed-batch fermentor

conditions relative to culturing of E. coli under shake-flask conditions.

108

Figure 50. Control 1H NMR of gallic acid. Resonances: 8 7.06 (s, 2H).

109

in: u N m

uiP—pnphpp-ipp—.-p_.bbp—npbn—pb-p—r.

Jet

 

P-Ean..—.n-b-..._LL~b—..Epr-P-pbnhpb-bPrppb

I -Lfit

 

 

110

Figure 51. Control 1H NMR of 3-dehydroquinic acid (DHQ). Resonances: 8
2.35 (d, 1H), 2.39 (d, 1H), 2.60 (dd, 1H), 3.21 (d, 1H), 3.95 (ddd, 1H), 4.36 (d, 1H).

111

Ean

 

 

 

 

____...__

 

 

112

Figure 52. 1H NMR of a typical DHS fermentation after 48 h. ( This

specific NMR was obtained using KL3/pKL5.17A). DHS Resonances: 8 6.42
(d, 1H), 4.28 (d, 1H), 4.00 (ddd, 1H), 3.07 (dd, 1H), 2.66 (ddd, 1H).

113

can

 

 

 

 

 

 

 

 

 

 

114

CHAPTER 4

CARBON SOURCE SELECTION: IS D-GLUCOSE THE BEST CARBON SOURCE
FOR MICROBIAL SYNTHESIS?

Wu

Carbon source selection is one of the most important considerations confronting
microbe-catalyzed chemical synthesis. Although D-glucose derived from corn starch has
historically been used for microbial synthesis of chemicals under fed-batch fermentor
conditions, complete dependence on glucose may not be advisable. Microbes such as
Escherichia coli and Bacillus subtilis employ a phosphoenolpyruvate (PEP)-expending
phosphotransferase system (PTS) for glucose transport. When glucose serves as the sole
carbon source for synthesis of chemicals via the common pathway of aromatic amino acid
biosynthesis, limitations imposed on PEP availability reduce titer and yield of synthesized
product.47dt '3

Corn fiber, an inexpensive byproduct of wet milling, constitutes 10% of the dry
weight of processed corn.117 Fiber is separated from corn, mixed (typically) with steeping
liquor, dried, and sold as gluten feed.118 This use of corn fiber constitutes a waste of a
valuable synthetic starting material. Corn fiber is a polymer composed of D-glucose, D-
xylose, and L-arabinose monomers which is end-capped with phenylpropanoids such as
ferulic acid. Besides its price and availability, the corn fiber's D-xylose and L-arabinose
content are particularly attractive in terms of serving as carbon sources for microbial
synthesis. D-Xylose and L-arabinose are transported into E. coli by ATP-based
permeases.“9 Therefore, they should not suffer, in theory, from the pronounced yield
limitation associated with PT S-driven glucose utilization. Compared to microbe-catalyzed

synthesis of chemicals from glucose, microbe-catalyzed syntheses of chemicals using

115

xylose and particularly arabinose as the carbon source have received comparatively little
attention.

Succinic acid is manufactured via succinic anhydride and maleic anhydride from
butane whose principal source is petroleum refining.120 Succinic acid availability benefits
from the wide availability of butane and the simple process to convert butane into succinic
acid. In the biological world, succinic acid is an intermediate in the tricarboxylic (TCA)
cycle . The TCA cycle is the common mode of oxidative metabolism in both eukaryotes
and prokaryotes. In E. coli, the TCA cycle is connected with the common pathway of
aromatic amino acid biosynthesis through reactions catalyzed by PEP carboxylase
(Ppc)116C and PEP carboxykinase (Pck).121 While Ppc catalyzes an irreversible
condensation of PEP with bicarbonate to form oxaloacetate, Pck catalyzes conversion of
PEP into oxaloacetate in a reversible reaction.

This Chapter will explore the potential use of both corn fiber hydrolysate and a
mixture of glucose and succinic acid as synthetic starting materials in the microbial
synthesis of value-added chemicals. The focus is again on synthesis of DHS using E. coli
as the microbe. In the first part of this Chapter, xylose and arabinose are evaluated as
carbon sources individually and as a 3/3/2 molar mixture of glucose/xylose/arabinose. The
3/3/2 molar mixture of glucose/xylose/arabinose simulates the carbohydrate composition of
corn fiber hydrolysate. In the second part, a 1/1 molar ratio of glucose/succinic acid was
used as a carbon source in conjunction with overexpression of Pck. For both pentose and
glucose/succinic acid mixtures, the DHS yields and/or titers are significantly increased

relative to the data obtained using glucose as a sole carbon source.

116

E °EI"S

Fed-batch cultures were performed in a 2.0 L capacity Biostat MD B-Braun
fermentor connected to a DCU system and a Dell Optiplex Gs+ 5166M personal computer
equipped with B-Braun MFCS/win software for data acquisition and automatic process
monitoring. The temperature, pH and substrate feeding were controlled with a PID
controller. The temperature was maintained at 36 °C and pH 7.0 was maintained by
addition of concentrated NI-hOH or 2 N H2S04. Dissolved oxygen was measured using a
Braun polarographic probe. Antifoam (Sigma 204) was added manually as needed.

The major differences in the fermentation conditions used in the study of this
Chapter relative to those of Chapter 3 are inoculum and the preset maximum airflow rate.
As in the research summarized in Chapter 3, a minimal medium containing inorganic salts,
aromatic amino acids, aromatic vitamins, and a quantity of carbon source was used for the
fermentation medium. Previously, rich medium (LB medium) supplemented with
antibiotics was used as the inoculum. In this Chapter, minimal medium without antibiotic
supplementation was used as the inoculum. Inoculants were typically started by
introduction of a single colony into 5 mL of M9 medium supplemented with L-
phenylalanine, L-tyrosine, L-tryptophan, p-aminobenzoic acid, 2,3-dihydroxybenzoic acid,
and p-hydroxybenzoic acid. The culture was grown at 37 °C with agitation at 250 rpm for
18 h and subsequently transferred to 100 mL of M9 medium. After growth at 37 °C and
250 rpm for an additional 12 h, the inoculant was ready for transferring into the
fermentation vessel. The fermentation process was similar as the one reported in Chapter

3. However, the maximum air flow was decreased from 3.0 LIL/min to 1.0 LIL/min.

117

i ._l!': IUV' 5 1- i. .'419.- Tisfi' . . 1.1. .G_ 0_' 1 .011

c or 'rbi n si fD

A. Background

D-Glucose is transported into the E. coli cytoplasm by the phosphoenolpyruvate
(PEP) dependent phosphotransferase system (PTS), while xylose and arabinose are
transported by ATP-based permease systems. These different transport mechanisms are
reflected in the different theoretical maximum yield for biocatalytic synthesis of DHS from
glucose versus xylose or arabinose.

Determining the theoretical maximum yield begins with balancing the PEP and E4P
inputs with DHS product and byproducts (Eq. (1)). The PEP and E4P are then equated to
the amount of D-glucose or D-xylose or L-arabinose which is required to form these
substrates (Eq. (2a), (3a), and (4a)). A pyruvate acid term is included in Eq. (2a) to reflect
the conversion of one molecule of glucose transported into the cytoplasm as glucose 6-
phosphate. A coefficient is determined (Eq. (2b)) to balance the number of carbon atoms in
the glucose starting material with the total number of carbon atoms formed in PEP, E4P
and pyruvic acid. The determined coefficient leads to a maximum theoretical yield of 43%
(mol of DHS/mol of glucose). Similar calculations (Eq. (3b) and Eq (4b)) result in a
maximum theoretical yield of 71% (mol of DHS/mol of xylose or arabinose) for both
xylose and arabinose. In the case of glucose, PTS-generated pyruvic acid is considered to
be the carbon source for the anabolism and catabolism required for generation of E. coli
cellular biomass.33 However, when xylose or arabinose was used as the sole carbon
source, some of the carbon sources must be diverted to biomass production. As a result,
the calculateed yield of 71% (mollmol) must be considered to be the upper limit for the

theoretical maximum yield for both xylose and arabinose.

118

(1) p51: 4. E4P _> 2 H3PO4 + H20 + DHS

(2a) x glucose ——> PEP + E4P + x pyruvic acid

(2b) x6(C) -> 3(C)+4(C)+x 3(0)

(33) x xylose —> PEP+E4P
(3b) x5(C) —> 3(C)+4(C)

(43) x arabinose—-> PEP + E4P

(4b) x5(C) —’ 3(C)+4(C)

As previously mentioned, corn fiber represents a cheap, renewable resource that is
abundantly available from the corn wet milling industry.122 A number of pretreatment
processes have been developed in order to utilize the corn fiber as a low-cost feedstock for
production of fuel ethanol. All processes yield a mixture of glucose, xylose and arabinose,
although the ratio of the components varies depending on the treatment condition. For
example, when the corn fiber was treated with dilute sulfuric acid at pH 1.0 and 140 °C for
20 min, the hydrolysate obtained had a glucose, xylose and arabinose molar ratio of 3.3 :
3.6 : 2;123 When corn fiber was pretreated with dilute sulfuric acid at pH 2.0 and 160 °C
for 20 min, neutralized and then hydrolyzed with an enzyme mixture, the resulting sugar
mixture had a glucose, xylose and arabinose molar ratio of 5.1 : 3.5 : 2.123
Saccharification of alkaline hydrogen peroxide (AHP) pretreated corn fiber using crude
enzyme from Aureobasidium sp. NRRL Y-2311-1 afforded a mixture of glucose, xylose
and arabinose at a molar ratio of 6.6 : 2.2 : 2.124 It has also been reported that no
difference is observed between using corn fiber hydrolysate and mixtures of pure
carbohydrates that simulate the carbohydrate composition of hydrolysate for production of
ethanol through fermentation.122 In this section, the performance of individual

carbohydrates including glucose, xylose and arabinose will be compared in terms of

119

synthesized product titers and yields. The utilization of a 3 : 3 : 2 molar ratio of
glucose/xylose/arabinose to simulate the composition of the carbohydrate mixture resulting

from mild acid hydrolysate of corn fiber125 will also be reported.

B. Biocatalyst

KL3/pKL4.33B was the first biocatalyst tested using xylose as a sole carbon
source. However, this biocatalyst grew slowly on xylose. The starter inoculum (5 mL)
had to be grown for 24 h at 37 °C and subsequently transferred into 100 mL of culture for
an additional 24 h growth at 37 °C in order to reach a cell density appropriate for
inoculating a fermentation run. The fermentation also had a long initial lag phase and did
not reach the third stage and stationary phase until after 31 h and 48 h, respectively. By
contrast, KL3/pKL4.33B cultured on glucose required only 18 h and 12 h, respectively,
for growth of the 5 mL and 100 mL starter cultures. KL3/pKL4.33B cultured on glucose
under fed-batch fermentor conditions reached the third stage and stationary phase at 13 h
and 24 h, respectively. Therefore, KL3/pKL4.33B was not used for this study due to the
significant growth rate difference when xylose versus glucose was employed as a carbon
source.

DAHP synthase expression levels are known to affect E. coli growth rates. A
slower growth rate is typically observed with higher DAHP synthase overexpression
levels. Plasmid pKL4.79B has a PmcaroFFBR/lacIQ system which allows DAHP synthase
overexpression to be controlled by IPTG addition. Without IPTG addition, DAHP
synthase (AroF) expression is minimal due to the repression of transcription by the Lac
repressor. KL3/pKL4.79B was thus expected to show faster growth without IPT G
addition relative to KL3/pKL4.33B in which AroF is expressed through its native
promoter. For KL3/pKL4.79B cultured on xylose or glucose in the absence of IPTG, only
18 h and 12 h, respectively, were required for the 5 mL and 100 mL starter cultures. IPTG

was not added until the beginning of the third stage of the fermentation. The xylose

120

fermentation reached the third stage and stationary phase at about 14 h and 30 h,
respectively. For comparison, the third stage and stationary phase were reached at
approximately 13 h and 24 h, respectively, when glucose was the carbon source.
Therefore, KL3/pKL4.79B was used as the model biocatalyst for the research summarized
in this section. In all fermentations, 4.8 mg IPTG were added at the beginning of stage 3,

at 18 h, and after every subsequent 6 h time interval until completion of the fermentor run.

C. Glucose Fermentation Versus Xylose or Arabinose Fermentation.

As a control, KL3/pKL4.79B was first examined under the new fed-batch
fermentor conditions using glucose as a sole carbon source (Figure 53a). The fermentation
was again divided into three stages according to the three different methods used to control
dissolved oxygen (D.O.) levels at 20% air saturation. After inoculation, D.O. levels were
maintained by increasing the impeller stirring rate until a preset maximum value was
reached. Approximately 11 hours was required before the impeller reach its maximum
stirring rate (940 rpm). The mass flow controller then maintained D.O. levels at 20% air
saturation at the constant impeller stirring rate (940 rpm) by increasing the airflow until a
preset maximum value (1.0 LIL/min) was reached. Approximately 2.5 h was needed for
the airflow to increase to its maximum rate. At a constant impeller stirring rate (940 rpm)
and constant airflow rate (1.0 LIL/min), D.O. levels were then maintained at 20%
saturation by oxygen sensor-controlled glucose feeding for the rest of the fermentation.
The initial glucose concentration in the fermentation culture medium was 18 g/L. The
concentration of the glucose solution added under DC. control was 60% (w/v).

When KL3/pKL4.79B was cultivated under the same conditions using xylose or
arabinose as a sole carbon source (Figure 53b, 0), fermentations behaved the same during
the first two stages as when glucose was used as the carbon source. The initial xylose and
arabinose concentrations in the fermentation culture medium was 23 g/L. Approximately

14 h was needed for completion of the first two stages of DC. control. However, after the

residual xylose or arabinose was completely consumed at the beginning of the third stage of
the fermentation, the DD. increased to 40% and could not subsequently be reduced back to
20% even with aggressive addition of pentose. Addition of pentose was consequently set
to a constant rate (10 mL/h) in order to feed xylose (65% w/v) or arabinose (61% w/v) in
excess of what was consumed in the fermentation to avoid cell starvation. As a result, the
xylose or arabinose slowly accumulated in the fermentation medium. The accumulation
continued until 36 h when the pump controlling pentose addition was shut off. Microbial
metabolism was then maintained on the excess of xylose or arabinose present in the culture
medium until the end of the fermentor run at 48 h (Figure 53 b, c). During the entire third
stage, the DO. level slowly increased from about 40% up to about 60% air saturation.

All fermentor runs entered logarithmic growth 6 h after inoculation (Figure 53).
Glucose fermentation reached stationary phase at approximately 24 h while xylose and
arabinose fermentation reached stationary phase at about 30 h. Glucose fermentation also
afforded a slightly higher cell mass than xylose and arabinose fermentation (Figure 53).
Despite the accumulation of sugars in the xylose and arabinose fermentation, no acetic acid
was observed during most of the logarithmic phase and the entire stationary phase. No
acetic acid was observed during the same period of time in the glucose fermentation.

Glucose fermentation of KL3/pKL4.79B afforded 36.4 g/L of DHS in a yield of
22% (Table 12). DHQ (3.6 g/L) and gallic acid (2.6 g/L) were also formed. Use of xylose
as a carbon source for culturing KL3/pKL4.79B afforded 41.7 g/L of DHS in a 32% yield
while use of arabinose as a carbon source afforded 41.1 g/L of DHS synthesized in a 35%
yield (Table 12). Along with DHQ and gallic acid, use of both xylose and arabinose as a
carbon source produced significant amounts of DAH (9.7 g/L and 9.5 g/L for xylose and
arabinose fermentation, respectively) (Table 12). This is different from glucose

fermentation, where no detectable DAH was formed throughout the fermentation run.

122

 

 

 

 

 

 

 

20 50
,1: 18
8 16 «40%
§ 14 03'
82:12 q‘30 g;
13:10 J a 0’
3" 8 '20 31"
. 6 a)“
:CI:3 4 --10 E
0 2
0 '0
0 12 18 24 30 36 42 48
Time (h)

b
‘O
'3 2
.g 8
82 is
2 EB
D or“
C5 : :l:
I : O
D Z

0 12 18 24 30 36 42 48

Time (h)

c
E g
t“ 8
.2 .
5 8A
‘3’ Ed
2 32
o to
d w‘
I I
o o

 

 

01218 24 30 36 42 48
Time (h)

Figure 53. Synthesis of DHS by KL3/pKL4.79B. (a) D-glucose
(b) D-xylose (c) L-arabinose: —I:l— cell dry wt; —¢- DHS; + glucose

or xylose or arabinose; EIID DHQ; = DAH; — gallic acid;

123

Along with the increased DHS, DHQ, and gallic acid production, the DAH accumulation
results in a higher total yield of hydroaromatics and aromatics synthesized from xylose and
arabinose relative to use of glucose as the carbon source (Table 12). A total yield of 44%
(mollmol) for xylose and 47% (mollmol) for arabinose was achieved relative to 26%
(mollmol) for use of glucose as the carbon source.

Table 12. KL3/pKL4.79B synthesis of DHS using different
carbon sources.

 

PVOdUCtS glucose xylose arabinose mix°
oHsa 36.4 41.7 41.1 53.0
DHS yieldb 22% 32% 35% 36%
DAHa 0 9.7 9.5 7.3
0Hoa 3.6 4.3 3.0 7.3
(iAa 2.6 4.0 4.7 2.3
total yieldb 26% 44% 47% 45%
theoretical yieldb 43% 71% 71% 60%

3 g/L; b mol/mol; c 3/3/2 molar ratio of
glucose/xylose/arabinose.

D. Glucose, Xylose and Arabinose as a Mixed Carbon Source

Culturing E. coli KL3/pKL4.79B under fed-batch fermentor conditions at 36 °C
and pH 7.0 was also examined when a 3/3/2 molar mixture of glucose/xylose/arabinose
was used as the carbon source. With the carbohydrate mixture as the initial substrate in the
first two stages of the fermentation, the cells exhibits diauxic grth (Figure 54). Glucose
is the first carbohydrate consumed due to the inhibition of both arabinose and xylose

transport by catabolite repression. Arabinose consumption did not begin until the cells

124

consumed all of the available glucose. Unexpectedly, xylose utilization did not start until
arabinose was exhausted, which suggests that arabinose may repress xylose uptake.

At 10 h, when the cells exhausted glucose, the fermentation typically was at the
beginning of second stage with the air flow at approximately 0.15 LIL/min. Because the
cells must divert their energies from growth to “retool” for the new carbon supply, an air
flow rate of 0.15 LIL/min at the maximum stir rate (940 rpm) was adequate for maintenance
of the DO. level at 20% air saturation. At the transition between glucose and arabinose
consumption, the airflow rate decreased sharply to its minimum value (0.06 LIL/min)
followed by a decrease in the stirring rate to around 850 rpm. Upon initiation of arabinose
consumption, the DO. level was again maintained at 20% air saturation first by increasing
the impeller rate and then by increasing the airflow rate. About 1 h was needed before all
the arabinose was consumed. A sudden decrease in airflow rate followed by a drop in the
stirring rate to about 800 rpm was again observed at the transition between arabinose and
xylose consumption. The beginning of xylose consumption corresponded with an increase
of the stirring rate followed by an increase in the air flow rate. About 3 h was required for
the cells to consume all the remaining xylose when the fermentation was in its second stage
with the air flow rate at about 0.3 LIL/min. Approximately 2 mL of the carbohydrate
mixture was added to the fermentation to drive the airflow rate up to its preset maximum
instantly due to the presence of the glucose. Only several minutes was required to consume
all of the added carbohydrate mixture.

D.O. sensor-controlled addition of the carbohydrate mixture began immediately
after the aliquot of the carbohydrate mixture was consumed. The D.O. level was
maintained at 20% air saturation throughout the fermentation. The initial carbon source
concentrations in the fermentation culture medium were 7.6/6.3/4.1 g/L for
glucose/xylose/arabinose mixture. The concentrations (w/v) of the individual
carbohydrates in the carbohydrate mixture were 26.3%, 21.9%, and 14.4% for glucose,

xylose, and arabinose, respectively.

125

 

-O—glucose (g/L)
-D—xylose (g/L)
—A-arabinose (g/L)

 

 

Sugar cocentrations (g/L)

 

 

 

 

0 6 7 3 91011121314
Time(h)

Figure S4. KL3/pKL4.79B cultivation prior to initiation of
oxygen sensor-controlled addition of the glucose : xylose :
arabinose mixture.

The steady-state concentration of glucose was maintained below 0.5 mM (not
detectable by 300 MHz NMR) while D.O. levels were used to control addition of the
carbohydrate mixture. This glucose concentration was sufficiently low that catabolite
repression of xylose and arabinose utilization was avoided.126 As a result, no
accumulation of pentoses was observed until 42 h when about 6 g of arabinose
accumulated in the culture (Figure 55). Fermentations were typically stopped at 42 h
because running beyond 42 h resulted in both xylose and arabinose accumulation (Figure
55).

Cultivation of KL3/pKL4.79B on the mixed carbohydrates as a carbon source was
similar to cultivation on glucose in terms of time required to reach the stationary phase and
the final cell mass which was produced (Figure 52a, Figure 55a). At no time during
cultivation on the glucose/xylose/arabinose mixture was acetic acid accumulation detected.
The concentration of DHS synthesized by KL3/pKL4.79B when cultivated on the mixed

carbohydrate as a carbon source was substantially in excess of the concentration of DHS

126

synthesized when any of the individual carbohydrates were employed as the carbon source.
A DHS titer of 53 g/L was synthesized in 36% yield (mollmol). Taking into consideration
(Table 12) the concentrations of DHQ (7.8 g/L), DAH ( 7.3 g/L) and gallic acid (2.8 g/L)

synthesized, the total yield for the synthesis of hydroaromatics and aromatics was 45%

 

     
 
  

 

 

 

 

 

 

(mollmol).
a
7

35 =glucose (g/L) O
m 30.. t=lxylose (g/L) "60 E:
S —arabinose (g/L) U
E- 25 " —D—cell dry wt (9) “50 72'”
8: 20" +0143 (g/L) «40 E;
C D) .‘d
8 v 15 -- " 3° 3 g
e 33
a, 10+ “20 "’
a = I

5 + [410 0

o '1 E '0

36 42

 

   

 

 

 

DHQ, DAH, gallic acid (g/L)
O)

     

g; Illlllllllllllllllllllr

w lllllllllllllll

0 12 18 24
Time (h)

Figure 55. Cultivation of KL3/pKL4.79B under fed-batch
fermentor conditions using a mixture of glucose, xylose, and
arabinose at a molar ratio of 3 : 3 : 2. (a) DHS production, cell
growth and carbon source concentrations (b) DHQ, DAH, and gallic acid
concentrations.

127

35 70

 

   
  
   
 
   
    

 

 

 

 

 

 

 

=glucose (g/L) .

m 30 -- =xylose (g/L) --60 2'
g L —arabinose (g/L) 4L '0
“g 25i -D—cell dry wt (9) 50 g”
E: 201- +DHS (cm 440 =5}:
0 \ \
§3 15 4. 4e 30 j 3
g 101' «20 (5,?
(1:) = I

5 - 5 -- 10 D

0 - '3 - 0

 

 

12- noHo (glL)
10‘ DDAH (g/L)
I aliic acid IL

   

        

 
   

 

 

 

DHQ, DAH, gallic acid (g/L)
a:

     

llllllllllllllllllll
llllllllllllllllllllllllllllll

illllllllll

, =1
0 12 18 24 30
Time (h)

w illlllllllllllllllllllllll

Figure 56. Cultivation of KL3/pKL4.124A under fed-batch
fermentor conditions using a mixture of glucose, xylose, and
arabinose at a molar ratio of 3 : 3 : 2. (a) DHS production , cell
growth and carbon source concentrations (b) DHQ, DAH, and gallic acid
concentratio

ns.

128

E. The Impact of Transketolase

KL3/pKL4.124A, which was obtained by cloning tktA gene into pKL4.79B
(Figure 45), was also examined using individual carbohydrate as well as the carbohydrate
mixture as carbon sources in order to compare the effect of transketolase overexpression.
Arabinose was not tested as a carbon source given the similar results when KL3/pKL4.79B
was cultured on arabinose or xylose. Despite the report that overexpression of
transketolase only modestly increased the carbon flow directed into the common pathway
of aromatic amino acid biosynthesis when glucose was employed as the carbon source
under shake-flask conditions,45dt '3 Chapter 3 clearly demonstrated that transketolase can
significantly improve the flow of carbon directed into the common pathway under fed-batch
fermentor conditions when glucose was employed as the carbon source.

KL3/pKL4.124A was examined under the conditions described above. IPTG (4.8
mg) was added at the beginning of stage 3, at 18 h, and after each subsequent 6 h time
interval to realize optimal DAHP synthase expression levels. Results are listed in Table 13.
Using glucose as a carbon source, KL3/pKL4.124A produced DHS at a titer of 46.0 g/L
with a yield of 28% (mollmol), which sharply contrasts with a titer of 36.4 g/L and a yield
of 22% (mollmol) that was observed when KL3/pKL4.79B was cultured under the same
culture conditions (Table 12, 13). Although both DHQ and gallic acid production
increased, DAH was not detected in the fermentation supernatant when glucose was used
as the carbon source. Fed-batch fermentor cultivation using xylose as the carbon source
did not show a significant change upon overexpression of transketolase. KL3/pKL4.124A
produced DHS, DHQ, and gallic acid at similar levels relative to KL3/pKL4.79B (Table
12, 13). The production of DAH was slightly increased which led to a small increase in
total yield. The yield of DHS synthesis remained unchanged. For the fermentation
utilizing a 3 : 3 : 2 molar ratio of glucose/xylose/arabinose, overexpression of transketolase
showed a significant impact on both titer and yield. A DHS titer of 64.0 g/L was

synthesized by KL3/pKL4.124A after 42 h of cultivation. The yield of DHS was 41%

129

(mollmol) relative to 36% (mollmol) in KL3/pKL4.79B. Combination of the DHQ (9.2
g/L), DAH (10.7 g/L), gallic acid (3.1 g/L) and DHS synthesis results in a total yield of
synthesized products of 55% (mollmol) (Table 13).

Table 13. KL3/pKL4.124A synthesis of DHS with different
carbon source.

 

products glucose xylose mix°
DHsa 46.0 43.1 64.0
DHS yieldb 23% 33% 41 %
DAHa 0 12.3 10.7
ol-loa 4.5 4.3 9.2

GA3 3.3 3.3 3.1

total yieldb 33% 47% 55%
theoretical yieldb 43% 71 % 53%

3 g/L; b mol/mol; C 3/3/2 molar ratio of
glucose/xylose/arabinose.

When glucose and xylose were used as carbon sources, KL3/pKL4.124A showed
the same sugar accumulation of carbohydrates as did KL3/pKL4.79B. KL3/pKL4.124A
did not accumulate any glucose when glucose was the sole carbon source. However, when
cultivated in xylose, KL3/pKL4.124A accumulated xylose up to approximately 40 g/L at
36 h. When the carbohydrate mixture was used as a carbon source, a noteworthy
difference between KL3/pKL4.124A and KL3/pKL4.79B was the accumulation of both
xylose and arabinose by transketolase-overexpressing KL3/pKL4.124A beginning at 24 b
(Figure 56) while the fermentation was under D.O. sensor control. By 42 h, 18.6 g/L of
xylose and 15.4 g/L of arabinose had accumulated in the fermentation medium. Table 14
compares the total amount of individual carbohydrates consumed by KL3/pKL4.79B and

KL3/pKL4.124A when the carbohydrate mixture was the source of carbon.

130

KL3/pKL4.79B and KL3/pKL4.124A consumed approximately the same amount of
xylose and arabinose. On the other hand, overexpression of transketolase in
KL3/pKL4.124A significantly increased the amount of glucose consumed. Transketolase
overexpression apparently increased the glucose utilization rate but did not increase the

utilization rate of xylose and arabinose.

Table 14. Consumption of the individual components of
the mixed carbon source.

KL3/pKL4.79B KL3/pKL4.124A

 

glucose

consumed 82:5 9 103-0 9

coxrilaiaisiried 64:9 9 53.6 9

56:23:89: 37:0 9 37-2 9
23:; 3.7 : 3.5 : 2 4.8 : 3.4 : 2

3‘ glucose/xylose/arabinose

F. Discussion

In considering the performance of DHS-producers under different conditions,
synthesis of DAH, DHQ and gallic acid must also be taken into consideration. Both DAH
and DHQ are metabolic precursors to DHS and gallic acid is a metabolite derived from
DHS. The theoretical maximum yield of DHS, DAH, DHQ and gallic acid is 43%
(mollmol) for glucose and 71% (mollmol) for xylose. KL3/pKL4.79B directed a total of
63% and 62% of the theoretical maximum amount of carbon which can be channeled into
aromatic amino acid biosynthesis when glucose and xylose, respectively, were used as

carbon sources. For KL3/pKL4.124A, the total yield corresponds to 76% and 66% of the

131

theoretical maximum yield when glucose and xylose, respectively, were used as carbon
sources.

Determining the theoretical maximum yield for biocatalytic synthesis of DHS from
the carbohydrate mixture takes into account the consumption ratio of each individual sugar.
For KL3/pKL4.79B, the glucose, xylose and arabinose was consumed at a ratio of 3.7 :
3.5 : 2. Calculation according to Eq. (5a) leads to a theoretical maximum yield of 60%
(mol/mol). For KL3/pKL4.124A, since the actual sugar consumption is 4.8 : 3.4 : 2 for
glucose, xylose and arabinose, respectively, a similar calculation results in a theoretical
yield of 58% (mol/mol) (Eq. (5b)). Thus, KL3/pKL4.79B channeled 75% of the
theoretical maximum amount of carbon which can be directed into aromatic amino acid
biosynthesis. With the help of transketolase, KL3/pKL4.124A channeled 94% of the

theoretical maximum amount of carbon which can be directed into aromatic amino acid

 

 

 

 

biosynthesis.
3.7 3.5 2
(5a) x 43%». —— x 71% + x 71% = 60%
3.7+3.5+2 3.7+3.5+2 3.7+3.5+2
4.8 3.4 2
(5b) X 43% + —_ X 7170 + X 710/0 = 58%,
4.8+3.4+2 4.8+3.4+2 4.8+3.4+2

KL3/pKL4.124A, cultivated on glucose, consumed approximately 240 g of glucose
which generated 1.3 mol of pyruvic acid for biomass synthesis. When the carbohydrate
mixture was the carbon source, KL3/pKL4.124A consumed only 108 g of glucose (T able
14) which translates into 0.6 mol of pyruvic acid. Since both fermentations produced
roughly the same amount of cell mass, the efficiency for conversion of pyruvic acid to
biomass when using the carbohydrate mixture as the carbon source is more than two times

higher than when glucose was the sole carbon source.

132

In the Chapter 3, it was demonstrated that introduction of a second aroB copy into
the genome of E. coli constructs expressing elevated levels of DAHP synthase activity
eliminated accumulation of DAH in the culture supernatant due to the attendant increase in
DHQ synthase activity. However, use of xylose, arabinose, or the
glucose/xylose/arabinose mixture as a carbon source for KL3/pKL4.79B or
KL3/pKL4.124A cultivation results in DAH becoming a major byproduct. This
accumulation of DAH is an indication of the basic success in exploiting pentoses to alleviate
the competition between DAHP synthase and PTS-mediated glucose transport for in vivo
PEP supplies. Since xylose and arabinose do not require PEP expenditure for their
transport into the cytoplasm, aromatic amino acid biosynthesis is not limited by the
availability of PEP. An increase in the rate of aromatic amino acid biosynthesis is reflected
in the accumulation of DAH. Further amplification in the expression of aroB could reduce
or even eliminate DAH accumulation.

KL3/pKL4.124A afforded the highest DHS titer and yield when cultivated on a
mixture of glucose, xylose and arabinose. Nevertheless, the accumulation of xylose and
arabinose in the medium represents a waste of material and may also affect the purification
of the target molecule. One strategy to explore for eliminating pentose accumulation might
entail changing the ratio of glucose/xylose/arabinose added to the fermentor medium to
reflect the 4.8/3.4/2 molar ratio of glucose/xylose/arabinose consumed by
KL3/pKL4.124A during DHS synthesis. Depending on the different method employed in
the pretreatment of corn fiber, the ratio of these three carbohydrates in the corn fiber
hydrolysate varies.122v123t124 A number of pretreatment methods can yield mixtures with
higher glucose contents. For example, the enzymatic hydrolysate of ammonia fiber
explosion (AFEX) pretreated corn fiber gives a mixture of glucose, xylose and arabinose at
a molar ratio of 12.6 : 1.6 : 2.122 Whatever the sugar ratio is in the corn fiber hydrolysate,
the pure sugar can always be added to the corn fiber hydrolysate to make the ratio more

suitable for the fermentation. An alternative strategy to avoid the accumulation of xylose

133

and arabinose is to shut the feeding pump off at 42 h. The fermentation will end upon

exhaustion of the xylose and arabinose in the fermentation culture medium.

134

.I.‘ - '1' ‘.L: -O‘WJ ' Chi-.1 ‘m‘nl-II.

' Mi i - a zed n esis?

A. Introduction

PEP limitation is an inherent problem associated with the synthesis of industrial
chemicals from glucose using recombinant E. coli. A severe price is exacted in terms of
reduced titers of synthesized products because of PT S-mediated glucose transport. As
reported in the pervious section, PEP limitation can be resolved by switching the carbon
source from D-glucose to D-xylose and L-arabinose, which are transported into the E. coli
cytoplasm by ATP-driven permeases. Alternatively, a strategy can be explored to allow E.
coli to continue to use PEP-expending, PTS-mediated glucose transport while
supplementing the glucose-containing medium with metabolic precursors to PEP. The
primary focus of the supplementation strategy explored in this section is the TCA cycle
intermediate succinic acid. Succinic acid is readily available. It is manufactured via
hydrogenation of maleic anhydride and hydrolysis of the resulting succinic anhydride
(Figure 57). Maleic anhydride is produced by reaction of butane (derived from liquefied
petroleum gas) with oxygen using a vanadium phosphorus oxide catalyst. Succinic acid
can also be biosynthetically derived from renewable carbon sources such as glucose.127

DHS synthesis was again used as the model system for this study.

(VO) P O _ Ni, H H O
N _02_2,2 m __2_, [j _2_, Hozc’VcozH
2 O O O O O O
butane maleic anhydride succinic anhydride succinic acid

Figure 57. Manufacture of succinic acid from butane.

135

COASH C02 + NADH

o
SucAB
C°2*NA°H HOZCJK/‘COZH Hozc’Vco‘acaA 60P,H3Po,

 

 

 

N AD 2-ketoglutan'c acid succinyl CoA
OH led SucCD GTP
HO C O H O H
2 50214 2 HO zc’VC 2
o, L-isocitric acid succinic acid
FAD
H20
TCA Cycle SdhABCD FADH2
H \ O H
020’ng1 2 H 02 C V9?)
cls-aconitic acid tumanc ac1d
Wen PumA//\ H20
H20
HOZCYCOz HOZC CO?”
0 C O H
Mdh L-malic acid
H020‘ $00214 NAD
O
COASH Cd oxaloacetic acid NADH
CH3C-SCOA ATP
Pckt
ADP + 002
\ 02H
'30st
9“ PEP HO 02H
0 "'OH —) + AfOFFBR o
—> —’
: 0H TktA OH 0 : OH
HO OH H203PONLH OH
D-glucose OH Fl = P03H2 DAHP
E4P Fl = H DAH
AroB
CO?” HO COzH
AroE AroD b‘
HO" OH O = OH
H OH
shikimic acid \ DHO
L-phenylalanlne COZH vanillin
L-tyrosine catechol
L-t t han adi ic acid
ryp op HO OH 9
OH
gallic acid

Figure 58. Aromatic amino acid biosynthesis and the TCA cycle.
transketolase; AroFFBR: feedback-insensitive 3-deoxy-D-arabino-heptulosonate 7-
phosphate synthase; AroB: 3-dehydroquinate synthase; AroD: 3-dehydroquinate
dehydratase; AroE: shikimate dehydrogenase; Pck: PEP carboxykinase; GltA: citrate
synthase; Acn: aconitase; ch: isocitrate dehydrogenase; SucAB: 2-ketoglutrarate
dehydrogenase; SucCD: succinyl-CoA synthetase; SdhABCD: succinate dehydrogenase;

FumA: fumarase; Mdh: malate dehydrogenase.

136

B. Biocatalyst Construction

E. coli can employ dicarboxylic acid intermediates of the TCA cycle such as
succinic, fumaric, and L-malic acids as sole sources of carbon for growth. The transport
system consists of an inner membrane component encoded by dctA and dctB and a binding
protein component encoded by cbt. Upon transport into the cytoplasm, these
dicarboxylates can be converted into oxaloacetate, which leads to an increase in turnover of
the TCA cycle. Oxaloacetate is an anaplerotic metabolite whose concentration is maintained
at a low, constant value. If growth on dicarboxylates such as succinic acid leads to an
increase in oxaloacetate concentrations, various enzymes can catalyze reactions with TCA
cycle intermediates withdrawing carbon flow from the TCA cycle. The reaction between
oxaloacetate and ATP catalyzed by pck-encoded PEP carboxykinase, which results in
formation of PEP, ADP, and C02, was of particular interest to our research. With pck
overexpression, the increased carbon flow moving into the TCA cycle due to
supplementation of glucose with succinic acid will result in increased synthesis of PEP,
which was designed to translate into increased DHS synthesis (Figure 58).

Plasmid pKL4.79B was used as a parental plasmid for cloning of the pck gene
(Figure 60). The choice of pKL4.79B for the study was to obtain data which could be
compared with data collected in pentose studies, which employed this same plasmid. As a
source of the pck gene, plasmid pHG26128 was fully digested with EcoRI and partially
digested with EcoRV to isolate a 2.6-kb fragment containing the pck gene. Subsequent
ligation of the pck fragment into the EcoRI/Smal digested pSU18 afforded pKL2.222
(Figure 59). Following digestion of pKL2.222 with EcoRI/BamHI, the 2.6-kb pck
fragment was modified to blunt ends using Klenow fragment. Plasmid pKL4.79B was
digested with SalI and treated with Klenow fragment. Subsequent ligation of the pck
fragment into pKL4.79B afforded pKL6.198A. The pck gene is transcribed in the same
orientation as the aroFFBR gene (Figure 60). Plasmid pMF51A was digested with BamHI

and the resulting 2.2-kb tktA fragment was treated with Klenow fragment. Plasmid

137

pKL6.198A was digested with HindIII and treated with Klenow fragment. Subsequent
ligation of the tktA fragment to pKL6.198A afforded pKL6.218A. The tktA gene is

transcribed in the same orientation as the aroFFBR gene (Figure 61).

138

EcoRV

EcoRV
EcoRl Smal BamHI Hindlll

 

EcoRl

1) EcoFll digest
2) EcoRV partial digest
3) gel purification

 

 
        

EcoRl EcoRV EcoRV
2.6 kb
pck
1) EcoFll/Smai digest
2) CIAP treatment
T4 Ligase
EcoRl
Plac
pck
pKL2.222
10’“ 4.9 kb
(Smal)
BamHI

Hindlll

Figure 59. Preparation of Plasmid pKL2.222.

139

 

   

 

      
        

EcoFtl

(Smal)
pKL2.222
i EcoRl/Baml-tl digest
serA
ECORI BamHI
2.6 kb
pck
(Smal)

 

Hindlll

1) Sall digest
2) Klenow treatment

lKlenow treatment
3) CIAP treatment

T4 Ligase

 

serA

pKL6.198A

10.9 kb (Smal)

(Sall)

w

Hindlll

(Sall)

Figure 60. Preparation of Plasmid pKL6.198A.

140

pMF51 A
t BamHI digest
BamHI BamHI
2.2 kb
tktA

lKlenow treatment

(Hindlll)

 

serA

      
 
   
    

pKL6.198A
10.9 kb (Smal)

(Sall)

W

Hindlll (Salt)

1) Hindlll digest
2) Klenow treatment
3) CIAP treatment

T4 Ligase

(Smal)

(kt/I (Hindlll)

Figure 61. Preparation of Plasmid pKL6.218A.

141

serA

      
       

 

pMF51 A
pKL6.198A
. 10.9 kb
BamHI digest (Smal)
(Sall)
A
BamHI BamHI ‘\P
I 2.2 kb I
tktA -
Hindlll (Sall)
1) Hindlll digest
lKlenow treatment 2) Klenow treatment
3) CIAP treatment
T4 Ligase
EcoRl
“namFFBR EcoFll
”fig" Pta H.“ (Smal)
.. lacl°
serA
(Smal)
pKL6.218A (San)
13.1 kb
‘9
(Sall)
(Hindlll) WA (Hindlll)

Figure 61. Preparation of Plasmid pKL6.218A.

141

 

cuhi
baud
soun
ackl

enzy
and;
requi

PEIK

36GC1
synth
dunn
fénnt
ofcm
gfluco
the fit
€21er
CeHs

Vflue
[Hahn
SUCci

Carbq

 

Prese
The c

glUCt‘i

 

C. Channeling Carbon from Succinic Acid into DHS Synthesis

The basic design for use of succinic acid as a glucose supplement entailed
cultivation of a DHS-synthesizing E. coli construct, which overexpresses pck, under fed-
batch fermentor conditions where equimolar of succinic acid and glucose were used as the
source of carbon. Glucose could be viewed as a specialized source of E4P while succinic
acid was the specialized source of PEP. Conversion of glucose into E4P requires fewer
enzymes than conversion of succinic acid into E4P, which would require both TCA cycle
and gluconeogenesis enzymes. On the other hand, processing of succinic acid into PEP
requires fewer enzymatic steps than processing of glucose through glycolysis to obtain
PEP.

All fermentations were cultured for 48 h under fed-batch fermentor conditions at
36°C, pH 7.0, and dissolved oxygen maintained at 20% air saturation. Induction of DAHP
synthase overexpression was accomplished by incremental addition of IPT G (4.8 mg)
during the fermentation run. The initial carbon sources and their concentrations in the
fermentation culture medium were 18 g/L when glucose was employed as the sole source
of carbon, and 9.0 g/L of glucose and 5.9 g/L of succinic acid when the 1 :1 molar ratio of
glucose/succinic acid was used as the carbon source. Diauxic growth was observed during
the first two stages of DC. control when the glucose/succinic acid mixture was used as the
carbon source. This likely reflects catabolite repression of succinic acid transport.129 After
cells exhausted the glucose supply at 11-12 h, the airflow dropped sharply to its minimal
value. Succinic acid then became the only carbon source and the DO. level was again
maintained at 20% air saturation by increasing the air flow rate. Cells consumed all the
succinic acid in the culture medium in about 1 h. At this juncture, 2 mL of the mixed
carbon source was added to the fermentation to allow the airflow rate to increase to its
preset maximum at which time D.O. sensor-controlled addition of the carbon source began.
The concentration for added glucose solution was 60% (w/v). The concentration of added

glucose/succinic acid mixture was 31.3/20.5% (w/v).

142

 

ffabl
plaSI

g/L
KL3
and y
15).
“any
that
Succ

PEP

 

 

Table 15. Product titers and yields synthesized by KL3 as a function of
plasmid and carbon source.

 

 

pKL4.79B pKL4.124A pKL6.198A pKL6.218A

 

 

 

 

MA - -l- - +
POK - "' + +
glucose
[DHS], g/L 36 46 ND. 44
DHS yielda 22% 28% ND. 28%
total yieldb 26% 33% ND. 32%
activity6 0.04 0.04 N.D. 0.48
glucose/succinated
[DHS], g/L ND. 46 34 57
DHS yielda ND. 24% 21 % 29%
total yieldb ND. 29% 25% 37%
activityc Mb. 0.04 0.77 0.44

 

 

a mol/mol; 9 mol of (DHS + DHQ + DAH + gallic acid)/ mol of carbon sources
consumed; C umol/min/mg (Since stable Pck activities were observed through

the fermentation run, the number reported is the average of 18 h and 36 h.); 9
molar ratio: 1 : 1.

With glucose serving as a sole carbon source, KL3/pKL4.79B synthesized 36.4
g/L of DHS in a 22% (mol/mol) yield (Table 15). As demonstrated previously,
KL3/pKL4.124A, which overexpresses transketolase, significantly increased the DHS titer
and yield with the synthesis of 46.0 g/L DHS in 28% (mol/mol) yield from glucose (Table
15). For comparison, use of a 1/1 molar mixture of glucose/succinic acid in conjunction
with overexpression of Pck in KL3/pKL6.198A resulted in a DHS titer and yield similar to
that observed for KL3/pKL4.79B cultured on glucose (Table 15). Since growth on
succinic acid increases the carbon flow moving into the TCA cycle, the in vivo synthesis of
PEP is believed to be increased when the Pck is also amplified. The difference between
titers and yields of DHS synthesized by KL3/pKL4.124A from glucose and

143

avo
con
and
Ole;
acid

and

 

KL3/pKL6.198A from glucose/succinic acid mixture is consistent with E4P, rather than
PEP, being the first limiting metabolite in aromatic amino acid biosynthesis when DAHP
synthase is overexpressed. After E4P limitation was alleviated by overexpressing
transketolase, overexpression of Pck coupled with the succinic supplementation further
increased the titer of DHS. KL3/pKL6.218A synthesized 57.0 g/L DHS in 29% yield
from the glucose/succinic mixture (Table 15). However, neither KL3/pKL6.218A using
glucose only or KL3/pKL4.124A using glucose/succinic acid mixture was able to achieve
the same increase, indicating that the in vivo PEP availability can only be increased through
a combination of Pck overexpression and succinic acid supplementation (Table 15).
Substantial concentrations of 3-dehydroquinic acid (DHQ) and gallic acid were also
synthesized in all cases. DAH accumulation was only observed in KL3/pKL6.218A
fermentation when the glucose/succinic acid mixture was used. The observation of DAH in
the culture indicates again that two copies of aroB in the KL3 genome are insufficient to
avoid substrate accumulation. This further demonstrates that the carbon flow into the
common pathway was significantly increased by the combination of Pck overexpression
and succinic acid supplementation after E4P limitation was relieved with transketolase
overexpression. While KL3/pKL4.124A synthesized 4.5 g/L of DHQ and 3.8 g/L of gallic
acid using glucose as a sole carbon source, KL3/pKL6.218A synthesized 7.1 g/L of DHQ
and 4.4 g/L of gallic acid in addition to 4.2 g/L of DAH when the glucose/succinic acid

mixture was used as a carbon source (Figure 62).

144

Fi

ml

 

Pil

at

 

 

 

 

 

 

   
   

 

     

12 60
3101' «50 3
Z , ‘ T9
3 tie , --4o c0
3 E
g 6'”- I I I “-30 A
I- - d

I _ 03
3 4* é "2° 3

’ E E E E
c:3 2-- 5 a a "10 0
D : : :

0‘ : : : -0

36 42 48
b

12 60
3’10" «50 3
7.; :9
$8" ”40 a)
o I
E _ 5 a
$6" 5 E ‘30 A
I: I g d
< .i. , E s . 9
0.4 5 5 E 5 2° 9
O : : -'- : '03
I . : : = : O
02 E E E 5 F10

0" = 2 E E .0

 

 

 

 

    

01218 24 30 36 42 48
time (h)

Figure 62. Products and cells produced during fed-batch fermentation.

—D—cell dry wt; -A-DHS; EIIIIDHQ; =2: DAH; — gallic acid. (a)
KL3/pKL4.124A cultured on glucose (b) KL3/pKL6.218A cultured on a 1/1 molar
mixture of glucose/succinic acid.

Succinic acid transport is precedented to be subject to catabolite repression in the
presence of glucose.129 In cultures where glucose concentrations are relatively high, cyclic

adenosine monophosphate (CAMP) levels are low, resulting a repression of succinic acid

145

 

upt:

C00

616

8y]

 

uptake. In our study, diauxic growth has been observed prior to use of the DD. sensor to
control addition of the glucose/succinic acid mixture. Before the initial glucose is
exhausted, the uptake of succinic acid is minimal although the detection of trace furmaric
acid suggests there is some level of transport (Figure 63). When addition of the
glucose/succinic acid mixtures subsequently is controlled by a dissolved oxygen sensor,
the steady-state glucose concentration is maintained at approximately 200 M. Under these
glucose-limited conditions, an elevated CAMP level has been reported.”0 This accounts
for the absence of succinic acid accumulation during D.O. sensor controlled addition of the

glucose/succinic acid mixture.

 

 

n-L

O-‘MOD-hU'IONCDCDo

-D—glucose (g/L)
+succinic acid (g/L)
+fumaric acid (g/L)

 
    

 

carbon source cocentration
( 9 / L )

 

 

 

O 7 8 91011121314
Time (h)

Figure 63. KL3/pKL6.218A cultivation prior to initiation of
oxygen sensor-controlled addition of the glucose/succinic acid
mixture.

D. Replacing succinic acid with other TCA cycle intermediates
To further demonstrate that supplementation with TCA cycle dicarboxylates can
elevate carbon flow directed into the TCA cycle, which ultimately results in increased PEP

synthesis, 1/ 1 molar mixtures of glucose/L—malic acid and glucose/2—ketoglutaric acid were

146

 

als<
fen
sou
sup
we:
ferr
sou
init:
The
gluc

Tat
fun

 

and

5.4

 

also examined using KL3/pKL6.218A as the catalyst. The glucose/L-malic acid
fermentation was run in similar fashion to the glucose/succinic acid fermentation. Carbon
source concentrations were 9.0 g/L of glucose and 6.7 g/L of L-malic acid for malic acid
supplementation. For 2-ketoglutaric acid supplementation, it was observed that DO. levels
were difficult to control at 20% air saturation in the beginning of the third stage of
fermentation when a glucose/Z-ketoglutaric acid mixture was used as an initial carbon
source. Therefore, 18 g/L of glucose was used as the initial carbon source prior to
initiation of DO. sensor-controlled addition of the glucose/Z-ketoglutaric acid mixture.
The concentration of glucose/L-malic acid was 32%/24% (w/v), and the concentration of
glucose/Z-ketoglutaric acid was 31%/26% (w/v).

Table 16. Product titers and yields synthesized by KL3/pKL6.218A as a
function of glucose supplement.

 

 

glucose/ glucose/ glucose/
glucose succinate malate 2-ketoglutarate

[DHS]a 44 57 62 58

DHS yield” 28% 29% 30% 29%
[DAHP o 4.2 5.4 4.8'
[DHQ]£1 4.5 7.1 9.1 7.9
[gallic acidlfil 3.8 4.4 4.6 4.3
total yield0 32% 37% 38% 37%

 

 

3 g/L; b mol/mol; c mol of (DHS + DHQ + DAH + gallic acid)/ mol of
carbon sources consumed.

Similar increases in titers of DHS and co-products were observed with malic acid
and 2-ketoglutaric acid supplementation. KL3/pKL6.218A synthesized 61.9 g/L of DHS,
5.4 g/L of DAH, 9.1 g/L of DHQ and 4.6 g/L of gallic acid from the glucose/L-malic acid
mixture. The DHS yield was 30% (mol/mol) and the total yield of hydroaromatic and

147

aron
4.8 1
acid

aron

witi
the j
equ.
sub:
con
dio:
(Ihr
die:
and
PET
wet
in 1
m0
DH

Site

are

be
Wh

CO]

aromatic products was 38% (mol/mol). KL3/pKL6.218A synthesized 57.6 g/L of DHS,
4.8 g/L of DAH, 7.9 g/L of DHQ, 4.3 g/L of gallic acid from the glucose/Z-ketoglutaric
acid mixture. The DHS yield was 29% (mol/mol) and the total yield of hydroaromatic and

aromatic products was 37% (mol/mol) (Table 16).

E. Discussion

Determining the theoretical yield for DHS synthesis when glucose is supplemented
with succinic acid begins with balancing (Eq. (1)) the substrate and cosubstrate inputs with
the products and byproducts. The PEP and E4P input required for DHS synthesis is then
equated to the amount of D-glucose and succinic acid that is required to form these
substrates (Eq. (2a)). A pyruvic acid term is then added (Eq. (23)) to take into
consideration the operation of the PEP phosphotransferase system in E. coli. A carbon
dioxide term is also added because the conversion of one molecule of succinic acid
(through oxaloacetic acid) to one molecule of PEP produces one molecule of carbon
dioxide. Coefficients are determined to balance the number of carbon atoms in the glucose
and succinic acid input with the total number of carbon atoms constituted by PEP, E4P,
pyruvic acid and carbon dioxide formation. Since equal moles of glucose and succinic acid
were consumed, the same coefficient for the glucose term and the succinic acid term is used
in the carbon balance (Eq. (3b)) for overall conversion (Eq. (3a)). A requirement of 1.17
molecule of glucose and 1.17 molecule of succinic acid for synthesis of each molecule of
DHS leads to a maximum theoretical yield of 43% [mol of DHS/(mol of glucose + mol of
succinic acid)] for synthesis of DHS from glucose and succinic acid.

The basic strategy to direct carbon from TCA cycle intermediate succinic acid into
aromatic amino acid biosynthesis pathway to alleviate the PEP limitation has been proven to
be feasible. DHS titers were improved but DHS yields did not improved significantly
when both the moles of glucose and the moles of added succinic acid are taken into

consideration. KL3/pKL4.124A synthesized DHS and associated byproducts from

148

 

gnu
assc

yiel

cata
conn
conc
nuo

DA}
pnci
adde
Unan

gase

 

fins:

for r

 

glucose at 77% of the theoretical maximum yield. KL3/pKL6.218A synthesis of DHS and
associated byproducts from glucose and succinic acid was 86% of the theoretical maximum

yield.

(1) PEP + E4P -—> DHS

(23) xglucose + xsuccinic acid —-> PEP + E4P + xpyruvic acid + x002
(2b) x6(C) + x4(C) -—> 3(0) + 4(0) 2. x3(C) + x 1(0)

(3a) 1.17 glucose + 1.17succinic acid —->DHS + 3.51 pyruvic acid + 1-17002
(3b) 11.70 —> 7(0) + 351(0) + 1.17(0)

In relation to the aforementioned strategies for improving product titers in microbe-
catalyzed synthesis, use of succinic acid as a glucose adjunct is distinguished by the
combination of genetic manipulation and carbon source modification. Under fermentor
conditions, normal growth rates of the E. coli constructs are observed, and carbon flow
into aromatic amino acid biosynthesis is increased as reflected by DHS titers as well as by
DAH and DHQ titers. Pure succinic acid is also commercially available. Although current
pricing of succinic acid restricts its use to microbe-catalyzed synthesis of higher, value-
added chemicals, economies of scale remain to be fully exploited in succinic acid
manufacture. In light of the abundant availability of butane from liquefied petroleum
gases,131 use of butane-derived succinic acid as a glucose supplement demonstrates how a
fossil fuel feedstock can be combined with a renewable feedstock to create a carbon source

for microbe-catalyzed chemical synthesis, which is superior to glucose.

149

 

VHS
Innn

purc

MHI

 

CHAPTER 5

EXPERIMENTAL

rMes

Chromatography
HPLC analyses employed a Rainin HPLC interfaced with a Rainin Dynamax UV-
Vis detector (Model UV-l). A C18 column (5 pm, Rainin Microsrobe-MVTM, 4.6 x 250
mm) was used for all HPLC separations. Diethylaminoethyl cellulose (DE52) was

purchased from Whatman. Dye Matrex Red A was obtained from Amicon.

Spectroscopic Measurements

1H NMR spectra were recorded on a Varian VX-300 FT-NMR spectrometer (300
MHz). Chemical shifts were reported in parts per million (ppm) downfield from internal
sodium 3-(trimethylsilyl)propionate-2,2,3,3-d4 (TSP, 8 = 0.00) with D20 as solvent. TSP
was purchased from Lancaster. UV and visible measurements were recorded on a Perkin-
Elmer Lambda 3b UV-vis spectrophotometer connected to a RIOOA chart recorder or on a
Hewlett Packard 8452A Diode Array Spectrophotometer equipped with HP 89532A UV-
Visible Operating Software.

Bacterial Strains
E. coli DHSOL [F ' endAI hstI 7( r“ Km+ K) supE44 thi-1 recAI gyrA relAI
¢8OIacZAM15 A(lacZYA-argF)u169] and RB791 (W3110 100L819) were obtained
previously by this laboratory.”2 AB283468 [tsx-352 supE42 11' aroE353 malA352 (1')]
was obtained from the E. coli Genetic Stock Center at Yale University. AB3248133 [F '
aroF 363 aroG365 ar0H367 pr0A2 argE3 ilv-7 his-4 lac gal-2 tsx-358 thi (11)] was obtained

150

 

frox
frm
the

fror

into
an <

solu

sup}

 

 

from the laboratory of Professor C. Yanofsky. Neurospora crassa SY 7A was obtained
from the American Type Culture Collection (ATCC 24740). Plasmid pCRX-275 carrying
the open reading frame of catechol-O-methyltransferase cDNA from rat liver was obtained

from Orion Corporation.

Storage of Bacterial Strains and Plasmids
All bacterial strains were stored at -78 °C in glycerol. Plasmids were transformed
into DHSa 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.

Culture Medium

All solutions were prepared in distilled, deionized water. LB medium134 (1 L)
contained Bacto tryptone (10 g), Bacto yeast extract (5 g), and NaCl (10 g). TB medium
contained (1 L) tryptone (10 g) and NaCl (5 g). After autoclaving and directly before use,
MgSO4 (10 mL of 1 M stock per L) was added to the TB medium. M9 salts (l L)
contained NazHPO4 (6 g), KH2P04 (3 g), NH4C1 (l g), and NaCl (0.5 g). M9 medium
contained D-glucose (10 g), MgSO4 (0.12 g), and thiamine (0.001 g) in 1 L of M9 salts.
M63 medium (1 L) contained KH2P04 (13.6 g), (NH4)2SO4 (2 g), FeSO4-7H20 (0.0005
g), D-glucose (2 g), MgSO4 (0.12 g), and thiamine (0.001 g). The pH of M63 inorganic
salts was adjusted to 7.0 with 1 N KOH before autoclaving. Solutions of inorganic salts,
magnesium salts. and carbon sources were autoclaved separately and then mixed.
Fermentation medium (1 L) contained KzHPO4 (7.5 g), ammonium iron(III) citrate (0.3 g),
citric acid monohydrate (2.1 g), L-phenylalanine (0.7 g), L-tyrosine (0.7 g), and L-
tryptophan (0.35 g), and concentrated H2S04 (1.2 mL). The culture medium was adjusted
to pH 7.0 by addition of concentrated NH4OH before autoclaving. The following

supplements were added immediately prior to initiation of the fermentation: glucose (18 g,

151

 

auto
(122
appr
50 u

its/I)
chkn

p—n
:3
LI:

imfle

Ined

519(
then

Cuhi
antii
gr0\(
Shak
due(
appr
(37

 

20 g or 23 g), MgSO4 (0.24 g), aromatic vitamins p-aminobenzoic acid (0.01 g), 2,3-
dihydroxybenzoic acid (0.01 g), and p-hydroxybenzoic acid (0.01 g), and trace minerals
(N H4)6(Mo7024)-4H20 (0.0037 g), ZnSO4-7H20 (0.0029 g), H3BO3 (0.0247 g),
CuSO4-5H20 (0.0025 g), and MnC12-4H20 (0.0158 g). D-Glucose and MgSO4 were
autoclaved separately while aromatic vitamins and trace minerals were sterilized through
0.22-um membranes prior to addition to the medium. Antibiotics were added where
appropriate to the following final concentrations: chloramphenicol, 20 ug/mL; arnpicillin,
50 ug/mL; kanamycin, 50 ug/mL; tetracycline, 12.5 ug/mL; and spectinomycin, 50
ug/mL. Stock solution of antibiotics were prepared in water with the exception of
chloramphenicol which was prepared in 95% ethanol and tetracycline which was prepared
in 50% aqueous ethanol. L-Phenylalanine, L-tyrosine, shikimic acid, L-histidine, L-
isoleucine, L-valine, L-proline, L-arginine, and L-serine were added to M9 medium or M63
medium where indicated to a final concentration of 0.04 g/L. Antibiotics, thiamine and
amino acid supplements were sterilized through 0.22-um membranes prior to addition to
M9 or M63 medium. Solid medium was prepared by addition of 1.5% (w/v) Difco agar to
the medium.

Conditions for Shake-Flask Cultivation

For analysis of product accumulation in shake flasks, bacterial strains were
cultivated as follows. One liter of LB (4 L Erlenmeyer flask) containing the appropriate
antibiotics and if necessary, IPTG, was inoculated with 10 mL of an overnight culture
grown in LB with the appropriate antibiotics. Cultures were grown at 37 °C in a gyratory
shaker at 250 rpm for 10 h. Cells were collected by centrifugation (4000 x g, 5 min) and
directly resuspended in l L of M9 medium (4 L Erlenmeyer flask) containing the
appropriate antibiotics and if necessary, IPTG. Cultures were then returned to the shaker

(37 °C, 250 rpm). Samples were taken at timed intervals for product analysis.

152

 

conn
soft\
MFC
PID
main
was

Anni

in th
intro
cont:
at 25
repoi
of 11
21min

Was 1
1110C!)
meth

the i
leV’e]

940

 

General Fed-Batch Fermentor Conditions

Fed-batch cultures were grown in a 2.0 L capacity Biostat MD B-Braun fermentor
connected to a DCU system and a Compaq computer equipped with B-Braun MFCS
software or a Dell Optiplex Gs+ 5166M personal computer equipped with B-Braun
MFCS/win software. The temperature, pH and substrate feeding were controlled with a
PID controller. The temperature was maintained at 36 °C or 37 °C as specified. pH was
maintained at 7.0 by addition of concentrated NILOH or 2 N H2504. Dissolved oxygen
was measured using a Braun polarographic probe and was set at 20% air saturation.
Antifoam (Sigma 204) was pumped in manually as needed.

Two types of inoculants were prepared for the fermentation experiments described
in this thesis. For the data reported in Chapter 2 and 3, inoculants were started by
introduction of a single colony into 100 mL LB medium (enriched with 2 g glucose)
containing the appropriate antibiotic. After grth for 12 h to 14 h at 37 °C with agitation
at 250 rpm, the inoculant was ready for transfer into the fermentation vessel. For the data
reported in Chapter 4, inoculants were started by introduction of a single colony into 5 mL
of M9 medium supplemented with L-phenylalanine, L-tyrosine, L-tryptophan, p-
aminobenzoic acid, 2,3-dihydroxybenzoic acid, and p-hydroxybenzoic acid. The culture
was grown at 37 °C with agitation at 250 rpm for 18 h and subsequently transferred to 100
mL of M9 medium. After growth at 37 °C and 250 rpm for an additional 12 h, the
inoculant was ready for transfer into the fermentation vessel.

The typical fermentation can be divided into three stages according to three different
methods used to maintain dissolved oxygen (D.O.) at the set point of 20% air saturation.
The dissolved oxygen concentration was first maintained by gradually increasing the
impeller speed from a minimum value of 50 rpm to a maximum setting of 940 rpm. After
the impeller reached 940 rpm, the mass flow controller governing airflow maintained D.O.
levels at 20% saturation by increasing the airflow rate while the impeller was maintained at

940 rpm. Airflow ranged from 0.06 LIL/min to a preset maximum value. Two preset

153

 

re:
3P
st:

99

prc

takc
For
and
SUpe
one
come
(Con-
tnte§l

PP“?

 

 

maximum values for air flow rate were used in this thesis. For data reported in Chapter 2
and Chapter 4, the value was 1.0 LIL/min. For date reported in Chapter 3, the value was
3.0 LIL/min. Approximately 1-2 h were needed for the airflow to increase to its maximum
rate. Once the impeller speed and airflow rate reached their pre-set maximum values, D.O.
levels were maintained at 20% saturation by oxygen sensor-controlled substrate feeding.
At the beginning of this stage, dissolved oxygen levels fell below 20% saturation due to
residual substrate in the medium. This period lasted anywhere from several minutes to
approximately 1 h before all residual substrate was consumed and the substrate feeding was
started. The PID control parameters were set to 0.0 (off) for the derivative control (171)),
999.9 s (minimum control action) for the integral control (11), and 950.0% for the

proportional band (Xp). Typical fermentations lasted for 48 h.

1H NMR Analysis of Culture Supernatant

For strains being evaluated in shake flasks, samples (20 mL) of the culture were
taken at timed intervals, and the cells were removed by centrifugation (4000 x g, 5 min).
For strains being evaluated in a fermentor, samples (5 mL) were taken at timed intervals
and the cells were removed using a Beckman rrricrofuge. A portion (0.5-2.0 mL) of the
supernatant 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).
Concentrations of metabolites in the supernatant were determined by comparison of
integrals corresponding to each metabolite with the integral corresponding to TSP (8 = 0.00
ppm) in the 1H NMR.

154

 

Sanfl
Biol
phos
\vas
hydr
'Tns-
the;
equa
isoar
Piagf
01 h
[flue

RNA
and

lSrr

 

 

Genetic Manipulations
Citrate]

Recombinant DNA manipulations generally followed methods described in
Sambrook et al.37 Restriction enzymes were purchased from Gibco BRL or New England
Biolabs. T4 DNA ligase was obtained from Gibco BRL. Calf intestinal alkaline
phosphatase was obtained from Boehringer Mannheim. Agarose (electrophoresis grade)
was obtained from Gibco BRL. 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
N azEDTA (pH 8.0). Endostop solution (10X concentration) contained 50% glycerol (v/v),
0.1 M N azEDTA, 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 DNAase-free RNAase was added to 1 mL of 10X Endostop solution. DNAase-free
RNAase (10 mg mL'l) was prepared by dissolving RN Aase in 10 mM Tris-Cl (pH 7.5)
and 15 mM NaCl. DNAase 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.88 Each reaction (0.1 mL) contained 10 mM KC], 20 mM Tris-Cl (pH 8.8),
10 mM (NH4)2SO4, 2 mM MgSO4, 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 Vent polymerase. Initial template concentrations varied from 0.02 ug to 1.0 pg.

Pl ' NA
Plasmid DNA was purified on a large scale using a modified alkaline lysis method
described by Sambrook et 01.86 In a 2 L Erlenmeyer flask, LB (500 mL) containing the

155

appropriate antibiotics was inoculated from a single colony, and the culture was incubated
in a gyratory shaker (250 rpm) for 14 h at 37 °C. Cells were harvested by centrifugation
(4000 x g, 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'l) 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 (48000 x g, 20
min, 4 °C). The supernatant was transferred to two clean centrifuge 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 (20000 x g, 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 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 (12000 x g, 10 min, 4 °C) to remove high molecular weight RNA. The 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 (12000 x
g, 10 min, 4 °C). The DNA was then rinsed with 70% ethanol and dried. After
redissolving the DNA in 0.5 mL of TE containing 20 ug/mL of RNAase, the solution was
transferred to a 1.5 mL microcentrifuge tube and stored at room temperature for 30 min.
500 uL of 1.6 M NaCl containing 13% PEG-8000 (w/v) (Sigma) was added to this
sample. 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

156

redissolved 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 11L) 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 11L of TE.

The concentration of DNA in the sample was determined as follows. An aliquot

(10 11L) of the 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 a 50 ug mL'1 of plasmid DNA is 1.0.

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
lysozyme (5 mg mL'l) 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

157

50 -100 1.1L TE. DNA isolated from this method was used for restriction enzyme analysis,

and the concentration was not determined by spectroscopic methods.

e . . . s . N

Restriction enzyme digests were performed using buffer solutions supplied by BRL
or New England Biolabs. A typical digest contained approximately 0.8 ug of DNA in 8 11L
TE, 2 uL of restriction enzyme buffer (10X concentration), 1 1.1L of restriction enzyme, and
TE to a final volume of 20 1.1L. Reactions were incubated at 37 °C for l h. Digests were
terminated by addition of 2.2 “L of Endostop solution (10X concentration) and
subsequently analyzed by agarose gel electrophoresis. When DNA was required for
subsequent cloning, restriction digests were terminated by addition of l uL of 0.5 M
NazEDTA (pH 8.0) followed by extraction of the DNA with equal volumes of phenol and
SEVAG and precipitation of the DNA. DNA was precipitated by addition of 0.1 volume of
3 M NaOAc (pH 5.2) followed by thorough mixing and addition of 3 volumes of 95%
ethanol. Samples were stored for at least 2 h at -78 °C. Precipitated DNA was recovered
by centrifugation (15 min, 4 °C). To the DNA pellet was added 70% ethanol (100 uL),
and the sample was centrifuged again (15 rrrin, 4 °C). DNA was dried and redissolved in

TE.

e l ‘ c resi

Agarose gels were run in TAE buffer containing 40 mM Tris-acetate and 2 mM
EDTA (pH 8.0). Gels typically contained 0.7% agarose (w/v) in TAE buffer. Higher
concentrations of agarose (1%) were used to resolve DNA fragments smaller than 1 kb.

Ethidium bromide (0.5 ug mL"1) was added to the agarose to allow visualization of DNA

fragments over a UV lamp. The size of the DNA fragments were determined by using two
sets of DNA standards: )1. DNA digested with Hindlll (23.1-kb, 9.4-kb, 6.6-kb, 4.4-kb,

158

2.3-kb, 2.0-kb, and 0.6-kb) and 71. DNA digested with EcoRI and HindIII (21.2-kb, 5.1-
kb, 5.0-kb. 4.3-kb, 3.5-Rb, 2.0-kb, 1.9-kb, 1.6-kb, 1.4-kb, 0.9-kb, 0.8-kb, and 0.6-kb).

I l ' o N s

The band of agarose containing DNA of interest was excised from the gel and
chopped thoroughly with a razor in a plastic weighing tray. The agarose was then
transferred to a spin column consisting of a 500 uL microfuge tube packed tightly with
glass wool and with an 18 gauge hole in its bottom. The spin column was then centrifuged
for 5 min using a Beckman microfuge to separate the DNA solution from the agarose. The
DNA—containing aqueous phase was collected during centrifugation in a 1.5 mL microfuge
tube. The DNA was precipitated with 3 M NaOAc and 95% ethanol as previously
described and redissolved in TE.

win-n o, - 1 DN‘ W1 ._ In - .1. .i ‘_'l,0 11..-..--

Plasmid vectors digested with a single restriction enzyme were dephosphorylated to
prevent self-ligation. Digested vector DNA was dissolved in TE (88 11L). To this sample
was added 10 11L of dephosphorylation buffer (10X concentration) and 2 uL of calf
intestinal alkaline phosphatase (2 units). The reaction was incubated at 37 °C for 1 h. The
phosphatase was inactivated by addition of 1 11L of 0.5 M NazEDTA (pH 8.0) followed by
heat treatment (65 °C, 20 min). The sample was extracted with phenol and SEVAG (100

uL each) to remove the protein, and the DNA was precipitated as previously described and

redissolved in TE.

WWW
DNA fragment with recessed 3' termini was modified to blunt-ended fragment by

treatment with the Klenow fragment of E. coli DNA polymerase I. After the DNA (0.8-2

ug) restriction digestion was completed in a 20 uL reaction, a solution (1 11L) containing

159

each of the desired dNTPs was added to a final concentration of 1 mM. Addition of 1-2
units of the Klenow fragment was followed by incubation of the mixture at room
temperature for 20-30 min. Since the Klenow fragment works well in the common buffers
used for restriction digestion of DNA, there was 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 by DNA

precipitation.

I . . [EN 1

DNA ligations were designed so that the molar ratio of insert to vector was 3 to l.
A typical ligation reaction contained 0.03 to 0.1 ug of vector and 0.05 to 0.2 pg of insert in
a total volume of 7 uL. To this sample was added 2 11L of ligation buffer (5X

concentration) and l 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.

 

Competent cells were prepared using a procedure modified from Sambrook et al.83
An aliquot ('1 mL) from an overnight culture (5 mL) was used to inoculate 100 mL of LB
(500 mL Erlenmeyer flask) containing the appropriate antibiotics. The cells were cultured
in a gyratory shaker ( 37 °C, 250 rpm) until they reached the mid-log phase of growth
(judged from the absorbance at 600 nm reaching 0.4-0.6). The culture was poured into a
large centrifuge bottle that had been previously sterilized with bleach and rinsed with sterile
water. The cells were collected by centrifugation (4000 x g, 5 min, 4 °C) and the culture
medium was discarded. All manipulations were carried out on ice during the remaining
portion of the procedure. The cell pellet was washed with 100 mL of cold 0.9% NaCl
(w/v) and then resuspended in 50 mL of cold 100 mM CaClz. The suspension was stored

on ice for a minimum of 30 min and then centrifuged (4000 x g, 5 min, 4 °C). The cell

160

pellet was resuspended in 4 mL of cold 100 mM CaC12 containing 15% glycerol (v/v).
Aliquots (0.25 mL) were dispensed into 1.5 mL microcentrifuge tubes and immediately
frozen in liquid nitrogen. Competent cells were stored at -78 °C with no significant
decrease in transformation efficiency over a period of six months.

Frozen competent cells were thawed on ice for 5 min before transformation. A
small aliquot (1 to 10 uL) of plasmid DNA or a ligation reaction was added to the thawed
competent cells (0.1 mL). The solution was gently mixed and stored on ice for 30 min.
The cells were then heat shocked at 42 °C for 2 min and placed on ice briefly (l min). LB
(0.5 mL, no antibiotics) was added to the cells, and the sample was incubated at 37 °C (no
agitation) for l h. Cells were collected in a microcentrifuge (30 s). If the transformation
was to be plated onto LB plates, the cells were resuspended in a small volume of LB
medium (0.1 mL), and then spread onto plates containing the appropriate antibiotics. If the
transformation was to be plated onto minimal medium plates, the cells was washed once
with the same minimal medium. After resuspension in fresh minimal medium (0.1 mL),
the cells was spread onto the plates. A sample of competent cells with no DNA added was
also carried through the transformation procedure 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.

E 'E . E E . E11 !
Genomic DNA was purified using a modified method described by Silhavy.135 A

single colony of the 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 (4000 x g, 5 min, 4 °C) of the culture was followed by resuspension of the
cell pellet in 5 mL of buffer [50 mM Tris-HCl (pH 8.0), 50 mM EDTA (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

161

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 EDTA (pH 8.0), 0.5% SDS (w/v), and
proteinase K (1 mg mL'l, Sigma), 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 aqueous layers were separated by centrifugation (1000 x g, 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), 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 EDTA (pH 8.0), and 1 mg of RNAase. 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 ( 1000 x 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-HCl (pH 7.5) and 1 mM EDTA (pH 8.0).

Genomic DNA was stored at 4 °C.

Enzyme Assays
After collected and resuspended in proper resuspension buffer, the cells were
disrupted by two passages through a French pressure cell (SLM Aminco) at 16000 psi.

Cellular debris was removed from the lysate by centrifugation (48000 x g, 20 min, 4 °C).

162

 

Prc
pf€

Bic

10

m)
cor
dif‘
sol
pot
dill
ren

Pre-

qu;

Na
M

col
(pl
the

 

fig;
ab.K

 

Protein was quantified using the Bradford dye-binding procedure. A standard curve was
prepared using bovine serum albumin. The protein assay solution was purchased from

Bio-Rad.

WW

DAHP synthase was assayed according to the procedure described by Schoner.136
Harvested cells were resuspended in 50 mM potassium phosphate (pH 6.5) that contained
10 mM PEP and 0.05 mM CoClz. The cells were disrupted using a French press as
described above. Cellular lysate was diluted in potassium phosphate (50 mM), PEP (0.5
mM), and 1,3-propanediol (250 mM), pH 7.0. A dilute solution of E4P was first
concentrated to 12 mM by rotary evaporation and neutralized with 5 N KOH. Two
different solutions were prepared and incubated separately at 37 °C for 5 min. The first
solution (1 mL) contained E4P (6 mM), PEP (12 mM), ovalbumin (1 mg/mL), and
potassium phosphate (25 mM), pH 7.0. The second solution (0.5 mL) consisted of the
diluted lysate. After the two solutions were mixed (time = 0), aliquots (0.15 mL) were
removed at timed intervals and quenched with 0.1 mL of 10% trichloroacetic acid (w/v).
Precipitated protein was removed by centrifugation, and the DAHP in each sample was
quantified using thiobarbituric acid assay137 as described below.

An aliquot (0.1 mL) of DAHP containing sample was reacted with 0.1 mL of 0.2 M
NaIO4 in 8.2 M H3PO4 at 37 °C for 5 min. The reaction was quenched by addition of 0.8
M NaAst in 0.5 M Na2804 and 0.1 M H2SO4 (0.5 mL) and vortexed until a dark brown
color disappeared. Upon addition of 3 mL of 0.04 M thiobarbituric acid in 0.5 M NaZSO4
(pH 7), the sample was heated at 100 °C for 15 min. Samples were cooled (2 min), and
the pink chromophore was then extracted into distilled cyclohexanone (4 mL). The
aqueous and organic layers were separated by centrifugation (2000 x g, 15 min). The

absorbance of the organic layer was recorded at 549 nm (8 = 68000 L mol'l cm'l). One

163

UH

at

 

Sol

f0;

‘3’ 3

[er

PTE

“’a
Dr

\

D

aci
Hc
80}

"0

 

unit of DAHP synthase activity was defined as the formation of 1 umol of DAHP per min

at 37 °C.

DHstnthase
DHQ synthase activity was determined by measuring the disappearance of DAHP

over time using the thiobarbituric acid.137 Harvested cells were resuspended in Buffer A
(100 mL) consisting of 50 mM potassium 3-(N-morpholino)propanesulfonate (MOPS), pH
7.5, and 0.25 mM CoClz. After centrifugation, the cells were resuspended in Buffer A (2
mL g'l cells) and disrupted using a French press as previously described. An assay
solution was prepared containing MOPS (50 mM), pH 7.5, NAD+ (0.25 mM), CoC12
(0.25 mM), and DAHP (0.2 mM). The assay solution (1.5 mL) was preincubated at 37 °C
for 10 min. An aliquot (0.2 mL) of appropriately diluted enzyme (diluted with buffer A)
was added (t = 0) and the sample was incubated at 37 °C. Aliquots (0.2 mL) were
removed at timed intervals (15 s) and quenched with 10% trichloroacetic acid (0.1 mL).
Precipitated protein was removed by centrifugation using a Beckman microfuge and DAHP
was quantified using the thiobarbituric acid (TBA) assay as described above. One unit of

DHQ synthase activity is defined as the loss of one umol of DAHP per min at 37 °C.

W

DHS dehydratase activity was assayed by measuring formation of protocatechuic
acid at 290 nm. Harvested cells were resuspended in a solution containing 100 mM Tris-
HCl and 2.5 mM MgC12, pH 7.5 and disrupted as previously described. The assay
solution contained 0.1 M Tris-HCl (pH 7.5), 25 mM MgC12, and 1 mM DHS in a total
volume of 1 mL. The reaction was initialized upon the addition of the enzyme. The

absorbance at 290 nm was monitored for 5 min at room temperature. The specific activity

was expressed as umol of protocatechuic acid generated per min per mg protein at room

164

[CT

 

temperature. The molar extinction coefficient for protocatechuic acid at 290 nm is 3890 L

mol'1 cm'l.138

mm

A modification of the method of Reenila was used to assay catechol-O-methyl
transferase activity.139 Buffer A containing 10 mM NaHzPO4, pH 7.4 and 0.5 mM
dithiothreitol was used for washing and resuspending the cells and diluting the cellular
lysate. The cells were washed twice and then resuspended in buffer A. The cells were
disrupted as previously described. Two solutions were prepared separately and incubated
at 37 °C for 3 min. The first solution (4 mL) contained NaHzPO4 buffer (125 mM) pH
7.4, Mng (6.25 mM), S-adenosyl-L—methionine (0.75 mM), and protocatechuic acid (0.5
mM). The second solution (1 mL) contained the diluted lysate. The two solutions were
combined, vortexed briefly and incubated at 37 °C (time = 0). Aliquots (0.5 mL) were
removed at timed intervals (1 min) and quenched with 40 uL of ice-cold 4 M HC103.
Precipitated protein was removed by centrifugation using a Beckman microfuge. The
resulting supernatant was either analyzed immediately by HPLC as described below or
stored at -78 °C for up to two days.

The protocatechuic acid, vanillic acid and isovanillic acid components in the
supernatant were quantitated by HPLC. Individual component was separated by isocratic
elution (17:2:1 HzO/CH3/CH3COZH v/v) on a C18 column. The absorbance of the eluent at
250 nm was monitored. Samples were quantitated by comparison of the peak area
corresponding to each component with a standard curve. One unit of catechol-0-

methyltransferase activity is defined as the formation of l umol of the sum of vanillic acid

and isovanillic acid per min at 37 °C.

165

 

I'D

tl

1)
it
1.
tr

(1.

ti.

Andzaldsbldflshxdrmm

Aryl-aldehyde dehydrogenase assay solution (1 mL) contained Tris-HCl (100 mM),
pH 8.0, MgC12 (10 mM), dithiothreitol (20 mM), NADPH (0.15 mM), ATP (20 mM), and
benzoic acid (4 mM), and was pre-incubated at 30 °C. After addition of solution containing
aryl-aldehyde dehydrogenase, reduction of benzoic acid was monitored by following the

loss of absorbance at 340 nm. One unit of aryl-aldehyde dehydrogenase activity is defined
as the loss of 1 umol of NADPH (e = 6220 L mol'l cm'l) per min at 30°C.

W

PEP carboxykinase was assayed by modification of a procedure previously
described by Goldie.128 Harvested cells were resuspended in 100 mM Tris-HCl (pH 7.5)
that contained 1 mM EDTA and 0.1 mM DTT. The cells were disrupted as previously
described. A buffer solution at pH 7.5 was prepared containing Tris-HCl (100 mM), PEP
(10 mM), NaHCO3 (50 mM), ADP (4 mM), MgC12 (80 mM). Lysate (0.1 mL) and the
buffer solution (16 mL) were incubated separately at 31°C for 3 min. After addition of the
lysate solution to the buffer solution (time = 0), aliquots (2 mL) were removed at timed
intervals (1 min) and quenched with 0.75 mL absolute ethanol. Following the addition of
100 11L of freshly prepared Fast Violet B Salt (2% w/v in H20), the mixture was allowed
to stand at 31°C for 10 min followed by addition of 250 uL of 1 M HCl to stop the color
development. After a 1:5 dilution of each sample with H20, the absorbance was recorded
at 520 nm. Oxaloacetic acid formation was determined by comparison to a standard curve
prepared using oxaloacetic acid purchased from Sigma. One unit of activity is defined as

the formation of 1 mol of oxaloacetic acid per min in 31°C.

166

Chapter 2

Purification of Neurospora crassa AryI-Aldehyde Dehydrogenase

Buffers

Buffers used for purification of aryl-aldehyde dehydrogenase included buffer A:
Tris-HCl (100 mM) and L-cystine (10 mM), pH 7.6; buffer B: Tris-HCl (50 mM), EDTA
(1 mM), DTT (1 mM), and PMSF (0.4 mM), pH 7.6; buffer C: Tris-HCl (50 mM), EDTA
(1 mM), DTT (1 mM), and PMSF (0.4 mM), and KCl (400 mM), pH 7.6; buffer D: Tris-
HCl (20 mM), EDTA (0.4 mM), DTT (0.4 mM), and PMSF (0.15 mM), pH 7.5; buffer
B: Tris-HCl (20 mM), EDTA (0.4 mM), DTT (0.4 mM), PMSF (0.15 mM), and KCl (2.5
M), pH 7.5.

M r l . .

Growth medium for cultivation of N. crassa was prepared in distilled, deionized
water. Solid growth medium (1 L) contained sucrose (20 g), sodium citrate dihydrate
(2.5g), KHZPO4 (5.0 g), NH4NO3 (2.0 g), CaClz-ZHZO (0.1 g), MgSO4 (0.1 g), biotin
(5.0 pg), and trace elements including citric acid monohydrate (5.0 mg), ZnSO4-7HZO
(5.0 mg), Fe(NH4)2(SO4)2-6HZO (1.0 mg), CuSO4-5HZO (0.25 mg), MnSO4-H20
(0.05mg), H3BO3 (0.05 mg), NazMoO4-2HZO (0.05 mg). Difco agar was added to a final
concentration of 2% (w/v). The liquid growth medium (1 L) differed from solid growth
medium only in the addition of Difco yeast extract (2.0 g) and sodium salicylate (1.6 g).
N. crassa SY 7A was grown on solid growth medium in a petri dish (dimension: 100 mm x
80 mm) at 24 °C for 7 days and a mixture of mycelium and spores was obtained, which
was subsequently suspended in 20 mL of sterile water and filtered through glass wool that
had been previously sterilized by autoclaving. The resulting spore suspension can be

stored at 4 °C for up to two weeks. The concentration of spores in the suspension was

estimated by measuring the absorbance at 650 nm (A650 = 1.0 = 5.0 x 106 spores/mL).

167

 

Spores were inoculated into 2 L liquid growth medium in a 4 L Erlenmeyer flask to give a
final concentration of 2.5 x 106 spores/L. After culturing at room temperature on a gyratory
shaker at 200 rpm for 60 h, the mycelium was harvested by filtration through Whatman
filter paper on a Buchner funnel. The mycelium can be stored at -20 °C for at least 6

months with minimal loss of the aryl-aldehyde dehydrogenase activity.

1- n

All protein purification manipulations were carried out at 4 °C. Frozen mycelium
(400 g, wet weight) was thawed in 900 mL buffer A, and disrupted in a Waring blender for
20 min at 4 °C. To prevent overheating the lysate, the blending time was divided into 10
two-minute sessions with three-minute cooling time between two consecutive sessions.
The debris was removed by centrifugation at 18000 x g for 30 min followed by
concentration of the supernatant to 200 mL through ultrafiltration (PM-10 Diaflo
membranes from Amicon). After dialysis against buffer B (3x), the mycelium extract was
applied to a DEAE column (5 x 23 cm) equilibrated with buffer B. The column was
washed with 500 mL of buffer B followed by elution with a linear gradient (1.5 L + 1.5 L,
buffer B/buffer C). Fractions containing aryl-aldehyde dehydrogenase were combined and
concentrated to 30 mL. After dialysis against buffer D (3x), the protein was loaded onto a
RedA column (2.5 x 8 cm) equilibrated with buffer D. The column was washed with 200
mL buffer C and eluted with a linear gradient (150 mL + 150 mL, buffer D/buffer E).
Fractions containing aryl-aldehyde dehydrogenase were concentrated, frozen in liquid

nitrogen, and stored at -78 °C.

Strain Constructions

ISLZ

E. coli KL7 was prepared by homologous recombination of an aroBaroZ cassette

into the serA locus69 (serA::aroBaroZ) of AB2834.68 Localization of the serA gene in

168

pMAK70572 followed by insertion of the aroBaroZ cassette into a BamHI site internal to
serA directed recombination of the aroBaroZ cassette into the serA locus of the genome.
Plasmid pMAK705 contains a temperature-sensitive pSClOl replicon. Since derivatives of
pMAK705 replicate at 30 °C but are unstable at 44 °C, isolation of all pMAK705
derivatives required culturing of cells at 30 °C.

Digestion of pD262569b with EcoRV and Dral liberated a 1.9-kb serA fragment.
Plasmid pMAK705 was digested with BamHI and treated with Klenow fragment.
Subsequent ligation of serA fragment to pMAK705 afforded pLZl.68A.73 The aroBaroZ
cassette was created as follows: The aroB gene was obtained as a 1.7-kb fragment
following digestion of pJB14 with EcoRI. After treatment with the Klenow fragment, the
aroB fragment was inserted into the Smal site of pSK4.99A35 which was constructed by
ligation of 2.1-kb aroZ fragment into BamHI site of pSU18 to afford pKL4.237A. The
3.9-kb aroBaroZ cassette was amplified from pKL4.237A with the inclusion of BamHI
recognition sequences at the 5’- and 3’- ends. Insertion of the cassette into the BamHI site
of serA in pLZl.68A yielded pKL4.276B. Both aroB and aroZ are transcribed in the
opposite orientation relative to the lac promoter.

Conditions for homologous recombination were based on those previously
described.69b Competent AB2834 was transformed with pKL4.276B. Following heat—
shock treatment, cells were incubated in LB at 44 °C for 1 h and subsequently plated onto
LB plates containing Cm. Plates were incubated at 44 °C for approximately 20 h before
colonies appeared. The resulting cointegrates were inoculated into 5 mL of LB containing
no antibiotics, and the cells were grown at 30 °C for 12 h to allow excision of the plasmid
from the genome. Cultures were diluted (1:20000) in LB without antibiotics, and two
more cycles of growth at 30 °C for 12 h were carried out to enrich cultures for more rapidly
growing cells that had lost the temperature-sensitive replicon. Cultures were then diluted
(1:20000) into LB and grown at 44 °C for 12 h to promote plasmid loss from the cells.

Serial dilutions of each culture were spread onto LB plates and incubated at 30 °C

169

overnight. The resulting colonies were screened on multiple plates to select the recombined
ones. E. coli KL7 was isolated based on the following growth characteristics: growth on
M9 containing L-tyrosine, L-phenylalanine, shikimic acid and serine; no growth on M9
containing L-tyrosine, L-phenylalanine, shikimic acid; growth on LB; and no growth on

LB containing Cm.

ElaamiinKLLZZZA

Plasmid pCRX-275 carries the open reading frame (ORF) of the COM T. Digestion
of pCRX-2 with EcoRI/HindIII was followed by isolation of a 0.7-kb fragment encoding
the ORF of COMT. Plasmid pKK223-37O was also digested with EcoRI/HindIII to yield
a 5.3-kb fragment. Ligation of these two fragments resulted in pKL3.272A in which the

COM T gene was transcribed from a tac promoter.

W

This 6.6-kb plasmid was constructed by inserting the PmCOM T fragment from
pKL3.272A into pKL4.33B. Digestion of pKL4.33B with Sphl and HindIH afforded a
5.5-kb fragment while similar digestion of pKL3.272A yielded a 1.1-kb PmcCOM T

fragment. Ligation of these two purified fragments resulted in pKL5.26A.

W

Plasmid pKL5.26A was digested with Hindlll/Xbal and the 1.1-kb PmCOM T
fragment was isolated and treated with Klenow fragment. Plasmid pKL5.26A was also
digested with Hindlll and treated with Klenow fragment. Subsequent ligation of the
PmCOMT to pKL5.26A afforded pKL5.97A. The two PWCOMT genes are transcribed in

the same orientation in pKL5.97A.

170

Biocatalytic Synthesis of Vanillic Acid

Fermentation process was performed as described previously. The initial glucose
concentration in the fermentation medium was 20 g/L. The glucose feed was 60% (w/v).
A solution (500 mL) of glucose feed was prepared by autoclaving a mixture of 300 g of
glucose and 300 mL of water. L-Methionine, when employed, was added as an autoclaved
solution (40 g/L) in timed intervals (6 h) starting at 12 h after initiation of a fermentor run.

Fermentation samples (6 mL) were removed at 6 h intervals. A portion (1 mL) was
used to measure the cell density. The remaining 5 mL of each fermentation sample was
centrifuged and the DHS, protocatechuic acid, vanillic acid and isovanillic acid components
in the supernatant were quantitated by HPLC. Individual components were separated by
isocratic elution (17:2:1 HZO/CH3/CH3COZH v/v) on a C18 column. The absorbance of the
eluent at 250 nm was monitored. Samples were quantitated by comparison of the peak area
corresponding to each component with a standard curve. Separate aliquots (25 mL) of
fermentation broth were taken at 12 h and 36 h for assay of catechol-O-methyl transferase
activity. After the fermentation was complete, the entire broth was centrifuged at 16000 x g

for 10 min, and the resulting supernatant was stored at 4 °C.

Reduction of Vanillic Acid to Vanillin

Fermentation broth (100 mL) containing DHS, protocatechuic acid, vanillic acid,
and isovanillic acid was acidified to pH 3.1 by addition of concentrated HCl and the
resulting precipitated protein was removed by centrifugation at 16000 x g for 10 min. After
three extractions of the resulting clear brown supernatant with EtOAc (100 mL), the organic
extracts were combined and the solvent was removed under reduced pressure. The
resulting solid containing protocatechuic acid, vanillic acid and isovanillic acid was
dissolved in 12 mL of water, and adjusted to pH 7.5 by addition of NaOH (10 N).
Subsequent dropwise addition of concentrated H2804 to pH 1.8 resulted in precipitation of

a solid which was filtered and dried. The collected precipitate containing protocatechuic

171

acid, vanillic acid and isovanillic acid was dissolved in a solution (100 mL) containing Tris-
HCl (200 mM), pH 8.0, MgC12 (100 mM), DTT (10 mM), ATP (60 mM), NADP+ (2
mM), glucose 6-phosphate (60 mM), 2,000 units of glucose 6-phosphate dehydrogenase
and 200 units of the partially purified aryl-aldehyde dehydrogenase. The reaction was
incubated at 30 °C and monitored by HPLC. After 7 h, 92% of the starting vanillic acid
and 34% of the protocatechuic acid had been reduced. The reaction mixture was then
extracted with 100 mL CHzClz (3x). The combined organic extracts were washed one time
with an equal volume of water. Concentration of the organic layers afforded a yellow
powder consisting of vanillin (0.30 g) and isovanillin (0.03 g). 1H NMR (D20) for
vanillin were: 8 3.88 (s, 3H), 6.98 (d, J = 8 Hz, 1H), 7.39 (s, 1H), 7.46 (d, J = 8 Hz,
1H), 9.63 (s, 1H); 1H NMR (D20) for isovanillin were: 5 3.95 (s, 3H), 7.18 (d, J = 8
Hz, 1H), 7.38 (s, 1H), 7.55 (d, J = 8 Hz, 1H), 9.70 (s, 1H).

172

Sham}

Strain Constructions

mum

E. coli KL3 was prepared from AB283468 by homologous recombination of the
aroB gene into the serA locus69 (serA::aroB). Localization of the serA gene in
pMAK70573 followed by insertion of the aroB into an EcoRI site internal to the serA
directed recombination of the aroB into the serA locus of the genome. Digestion of
pKAD6314o with Sphl liberated a 1.9-kb serA fragment, which was subsequently inserted
into the Sphl site of pMAK705 to afford pKAD76A.”9 The aroB gene was obtained as a
1.7-kb fragment following digestion of pJB14139 with EcoRI. Insertion of the aroB
fragment into the EcoRI site of serA was complicated by two additional EcoRI sites in
pKAD76A. Following EcoRI partial digestion of pKAD76A. the resulting DNA fragments
were resolved on an agarose gel and the 7.4-kb fragment corresponding to the linearized
plasmid was isolated. Ligation of the linearized plasmid to the 1.7-kb EcoRI fragment of
aroB afforded pKL3.82A. Following homologous recombination of the serA::aroB locus
of pKL3.82A into AB2834, E. coli KL3 was isolated based on the following growth
characteristics: growth on M9 containing L-tyrosine, L-phenylalanine, shikimic acid and
serine; no growth on M9 containing L-tyrosine, L-phenylalanine, shikimic acid; growth on

LB; and no growth on LB containing Cm.

W

E. coli AB2.24 was constructed to facilitate mutagenesis of aroF to a feedback
insensitive isozyme of DAHP synthase (aroFFBR). AB2.24 was derived from AB3248133
by homologous recombination of an aroFkanR locus into the serA gene. The aroF locus
was obtained as a 1.3-kb EcoRI/BamHI fragment following PCR amplification from

pKD136,l 11 while a 1.2-kb BamHI/EcoRI kanR fragment was obtained from pKAD62A.69

173

Ligation of aroF and kanR to pSU18 which had been linearized by EcoRI afforded
pKDlO.156. Digestion of pKDlO.156 with EcoRI yielded the 2.5-kb aroFkanR fragment
which was subsequently localized in the EcoRI site internal to the serA gene in pKAD76A
to afford pKDlO.186A. Following homologous recombination of the serA::aroFkanR
locus of pKDlO.186A into A83248, E. coli AB2.24 was isolated based on the following
growth characteristics: no growth on M63 containing histidine, isoleucine, valine, proline,
and arginine; no growth on M63 containing histidine, isoleucine, valine, proline, arginine,
and shikimic acid; growth on M63 containing histidine, isoleucine, valine, proline,

arginine, and serine; growth on LB containing Kan; and no growth on LB containing Cm.

W

AB2.24 was subjected to in situ UV mutagenesis, and chemotactic selection was
performed in a Diffusion Gradient Chamber101 to select mutant strains bearing the
aroFFBR. Strain AC2-13A was isolated using the process previously reported.97 In vitro
assay of DAHP synthase in the presence of 125 M tyrosine confirmed the aroF locus in
AC2-13A was feedback insensitive. The aroFfBR locus was amplified from AC2-13A
with its native promoter to yield a 1.3-kb fragment. Inclusion of EcoRI recognition
sequences at the 5’- and 3’- ends of the aroFFBR fragment facilitated its insertion into the

EcoRI site of pCL1920 to afford pCL2-13A. Plasmid pCL2-13A was used to sequence the

aroFFBR locus.

W
Plasmid pCL2-13A was digested with EcoRI and the 1.3-kb aroFFBR fragment
was isolated and ligated into the EcoRI site of pSU18 to create pKL4.20B. The aroFFBR

gene is transcribed in the opposite orientation relative to the lac promoter in pKIA.20B.

174

W
This 5.5-kb plasmid was created by ligation of a 1.9-kb Dral/EcoRV fragment
encoding serA obtained from pD262569b into the Smal site of the pKL4.20B. The serA

gene is transcribed in the same orientation as aroFFBR.

W

The aroFFBR locus was amplified by PCR from pKL4.20B with Xbal ends.
Localization of the 1.3-kl) amFFBR into the Xbal site of pKL4.33B resulted in pKL4.66A.
Transcription of the aroFFBR locus in the Xbal site of pKL4.66A proceeds in the opposite

orientation of serA.

W

This 8.9-kb plasmid was created by inserting the tktA fragment from pMF51A56
into pKL4.66A. Plasmid pMF51A was digested with BamHI and the resulting 2.2—kb tktA
fragment was treated with Klenow fragment. Plasmid pKL4.66A was digested with
HindIII and treated with Klenow fragment. Subsequent ligation of the tktA fragment to
pKL4.66A afforded pKL4.130B. The tktA gene is transcribed in the same orientation as

the serA gene.

Plasmid QM] 1A

This 6.4-kb plasmid was constructed by ligating the open reading frame (ORF) of
aroFFBR into the EcoRI site of pJF118EH.103 The aroFFBR ORF was amplified from
pKL4.33B using following primers: 5’- GGAATTCATGCAAAAAGACGCGCTGA and
5’- GGAATTCTTAAGCCACGCGAGCCGT. Localization of the resulting 1.1-kb
fragment into pJF118EH afforded pKL4.71A. The ORF of aroFFBR is transcribed in the

same orientation as the tac promoter in pKL4.71A.

175

 

 

W
This 8.3-kb plasmid was created by ligation of a 1.9-kb Dral/EcoRV serA fragment
obtained from pD2625 into the Smal site of the pKL4.71A. The serA gene is transcribed in

the opposite orientation relative to the aroFFBR gene.

W

This 10.5-kb plasmid was constructed by inserting the tktA fragment from
pMF51A into pKL4.79B. Following digestion of pMF51A with BamHI, the 2.2-kb tktA
fragment was treated with Klenow fragment. Plasmid pKL4.79B was digested with
HindHI and treated with Klenow fragment. Subsequent ligation of the tktA fragment to
pKL4.79B afforded pKL4.124A. The tktA gene is transcribed in the same orientation as

the aroFFBR gene.

W

This 5.6-kb plasmid was constructed by inserting a fragment encoding the promoter
region of aroF into the Xbal site of pKL4.33B. Pm; was amplified from pMF63A using
following primers: 5’- GCTCTAGAGAATTCAAAGGGAGTGTA and 5’-
GCTCTAGACCTCAGCGAGGATGACGT. Transcription from Pam; is in the same

orientation as the serA gene.

WA

This 7 .8-kb plasmid was created by replacing the 1.0-kb NcoI/Sphl fragment of
pKDl 1.291A with a 3.2-kb NcoI/SphI fragment from pKL4.130B that included tktA.
Digestion of pKDl 1.291A with N001 and Sphl afforded a 4.6-kb fragment while similar
digestion of pKL4.130B yielded a 3.2-kb DNA fragment. Ligation of these two purified
fragments resulted in pKL5.17A.

176

Fed-Batch Fermentation

Fermentation process was performed as described previously. The initial glucose
concentration in the fermentation was either 18 g/L (for KL3/pKL4.79B, KL3/pKL4.33B,
and KL3/pKD1 1.29lA) or 23 g/L (for KL3/pKL4.124A, KL3/pKL4.66A,
KL3/pKL4.13OB, and KL3/pKL5.17A). The glucose feed concentration was 60% (w/v)
which was prepared as described previously. As needed, IPT G was added every 6 h from
12 h into the fermentation until the end of the fermentation

Sample (5 mL) of fermentation broth were taken at timed intervals. A portion (1
mL) was used to determine the cell density by measuring the absorption at 600 nm
(013600). The remaining 4 mL of each fermentation broth sample was centrifuged using a
Beckman microfuge. A portion (0.5-3.0 mL) of the culture supernatant was used for 1H
NMR analysis. The OD600 was converted to the cell dry weight by using a conversion
coefficient determined as follows: Fermentation broth with a known OD600 was
centrifuged, and all the cells were collected. After washing three times with fresh M9 salts
(400 mL), the cells were transferred to a container and dried in a 100 °C oven until the
weight was constant. The dry cell weight was determined and a conversion coefficient (dry

cell weight / OD600) of 0.43 was obtained using an average value of three experiments.

177

cicada

Strain Constructions
W
Plasmid pHG26128 was fully digested with EcoRI and partially digested with
EcoRV to isolate the 2.6-kb pck fragment. Subsequent ligation of the pck fragment into the
EcoRI/Smal digested pSU18 afforded pKL2.222.

mm

This 10.9-kb plasmid was constructed by inserting the pck fragment from
pKL2.222 into pKL4.79B. Following digestion of pKL2.222 with EcoRI/BamHI, the
2.6-kb pck fragment was treated with Klenow fragment. Plasmid pKL4.79B was digested
with Sall and treated with Klenow fragment. Subsequent ligation of the pck fragment to
pKL4.79B afforded pKL6.198A. The pck gene is transcribed in the same orientation as

the aroFFBR gene.

Blasmidnxrdziaa

This 13.1-kb pKL6.218A was created by inserting the tktA fragment from
pMF51A into pKL6.198A. Plasmid pMF51A was digested with BamHI and the resulting
2.2-kb tktA fragment was treated with Klenow fragment. Plasmid pKL6.198A was
digested with Hindlll and treated with Klenow fragment. Subsequent ligation of the tktA
fragment to pKL6.198A afforded pKL6.218A. The tktA gene is transcribed in thesame

orientation as the aroFFBR gene.

178

Fermentation Conditions
011,11 0 1' -1 o r ; ,‘_l.01__o._V‘ ‘ .. i o ,D-..,10' -‘_-.oos art I-
1 ' ' si "

The fermentation was performed as described previously. The inoculum carbon
sources were either 4 g/L when glucose, xylose, and arabinose were used individually or a
combination of 1.67 g/L of glucose, 1.4 g/L of xylose, and 0.92 glL of arabinose for sugar
mixture fermentations. Initial sugar concentrations in the fermentor culture medium were
18 g/L for glucose, 23 g/L for xylose or arabinose, and 7.6/6.3/4.1 g/L for the
glucose/xylose/arabinose mixture.

The feeding method differed depending on the carbon substrate. For both glucose
and the sugar mixture fermentations, at a constant impeller speed and a constant airflow,
D.O. was maintained at 20% air saturation by oxygen sensor-controlled substrate feed.
The concentration of the substrate feed for glucose fermentation was 60% (w/v) as was
prepared as described previously. The feed concentration for sugar mixture fermentations
was 26.3% (w/v) glucose, 21.9% (w/v) xylose, and 14.4% (w/v) arabinose. A typical
sugar mixture feeding solution (480 mL) was prepared by autoclaving a mixture of glucose
(126 g), xylose (105 g), arabinose (69 g) in water (300 mL). When xylose or arabinose
were used as carbon sources, the D. 0. could not be controlled at 20% air saturation in the
third stage of the fermentation. In these cases, the carbon substrate was added at a constant
rate of 10.0 mL/h after the initial substrate was exhausted. A typical xylose feed solution
(460 mL) was prepared by autoclaving a mixture of xylose (300 g) and water (280 mL)
(65% w/v). A typical arabinose feed solution (490 mL) was prepared by autoclaving a
mixture of arabinose (300 g) and water (300 mL) (61% w/v). IPTG (4.8 mg) was added
to all the fermentations at the beginning of the third stage, and then every 6 h started from

18 h until the end of the fermentation.

179

or..-” o .r' - 1.1 o "z _ 'L A ' a, . , {"11 '1 _e-0_=.. 1
Fe ,n- t .1” l ;_ .t- . 1.17.01.011411u,,ul.,-a-.z" _ t- is?"

The fermentation was performed as described previously. The carbon source for
the inoculum was 4 g/L of glucose in all cases. Initial carbon sources and their
concentrations in the fermentation culture medium were 9.0 g/L of glucose and 5.9 g/L of
succinic acid for glucose/succinic acid mixture, and 9.0 g/L glucose and 6.7 g/L of L—malic
acid for glucose/L-malic acid mixture. D.O. could be maintained at 20% air saturation
during the third stage of fermentation when 2-ketoglutaric acid was used as the glucose
adjunct only if glucose (at an initial concentration of 18 g/L) was used as the sole carbon
source for the first two stages.

The concentration of glucose/succinic acid feed mixture was 31.3/20.5% (w/v). A
typical glucose/succinic acid feed solution (480 mL) was prepared as follows: An
ammonium succinate solution was prepared by dissolving succinic acid (98.3 g) in H20
(130 mL) with the addition of concentrated NH40H (52 mL). A glucose solution was
prepared by mixing glucose (150 mL) with H20 (140 mL). The two solutions were
autoclaved separately and mixed after both were cooled to room temperature. The
concentration of glucose/L-malic acid feed mixture was 31.9/23.8% (w/v). A typical
glucose/L-malic acid feed solution (470 mL) was prepared as follows: A L-malic acid
solution was prepared by dissolving L-malic acid (111.7 g) in H20 (200 mL). A glucose
solution was prepared by mixing glucose (150 g) with H20 (100 mL). The two solutions
were autoclaved separately and mixed after both were cooled to room temperature. The
concentration of added glucose/2-ketoglutaric acid mixture was 31.3/25.4% (w/v). A
typical glucose/2-ketoglutaric acid feed solution (480 mL) was prepared as follows: A 2-
ketoglutaric acid solution was prepared by dissolving 2-ketoglutaric acid ( 121.8 g) in H20
(160 mL). A glucose solution was prepared by mixing glucose (150 g) with H20 (140

mL). The two solutions was autoclaved separately and mixed after both were cooled to

180

 

 

room temperature. IPTG (4.8 mg) was added for all the fermentations at the beginning of

the third stage, at 18 h and every 6 h increment after 18 h until the end of the fermentation.

181

BIBLIOGRAPHY

1 Reisch, M. S. Chem. Eng. News 1992, 70 (15), 16.

2 Lewis, R. J. Carcinogenically Active Chemicals, Van Nostrand Reinhold, New York,
199l,p.68.

3 Ember, L. Chem. Eng. News 1994, 72, 4 (March 7).

4 Frost, J. W.; Lievense, J. New J. Chem. 1994, 18, 341.

5. (a)Haslam, E. The Shikimic Pathway; Wiley: New York, 1974. (b) Bentley, R. Crit.
Rev. Biochem. Mol. Biol. 1990, 25, 307. (c) Herrmann, K. M. In Amino Acids:
Biosynthesis and Genetic Regulation; Herrrnann, K. M., Somerville, R. L., Ed.; Addison-
Wesley: Reading, 1983: p. 301.

6 Carpenter, E. P.; Hawkins, A. R.; Frost, J. W.; Brown, K. A. Nature 1998, 394, 299.
7 (a) Haslem, E.; Smith, B. W.; Turner, M. J. J. Chem. Soc., Perkins Trans. 1 1995,
52. (b) Duncan, K.; Chaudhuri, S.; Campbell, M. S.; Coggins, J. R. Biochem. J. 1986,
238, 475.

3 (a) Anton, I. A.; Cogins, J. R. Biochem. J. 1988, 249, 319. (b) Chaudhuri, S.;
Coggins, J. R. Biochem. J. 1985, 226, 217.

9 (a) DeFeyter, R. C.; Pittard, J. J. Bacterial. 1986, I65, 226. (b) DeFeyter, R. C.;
Pittard, J. J. Bacteriol. 1986, 165, 231. (c). DeFeyter, R. C.; Pittard, J. J. Bacteriol.
1986, 165, 233.

1° Laner-Olesen, A.; Marinus, M. G. J. Bacterial. 1992, 174, 525.

11 Duncan, K.; Coggins, J. R. Biochem. J. 1986. 234, 49. (b) Duncan, K.; Lewendon,
A.; Coggins. J. R. FEBS Lett. 1984, 165, 121.

12 White, P. J.; Millar, G.; Coggins. J. R. Biochem. J. 1988, 251, 313.

‘3 Hodgson J. Biotechnology 1994, 12, 152.

‘4 Oyama, K; Irina, S; Hagi, N. Methods Enzymal. 1987, 136. 503.

‘5 Thayer, A. M. Chem. Eng. News 1991, 69, 9.

‘6 Murdock, D. Ensley, B. D.; Serdar, C.; Thalen, M. Biotechnology 1993, 11, 381.

17 della-Cioppa. G.; Garger, S. J.; Sverlow, G. G.; Turpen, T.H.: Grill, L. K.
Biotechnology 1990, 8, 634.

13 Kim, C. U.; Lew, W.; Williams, M. A.; Liu, H.; Zhang. L.; Swaminathan, S.;
Bischofberger, N.; Chen. M. S.; Mendel, D. B.; Tai, C. Y.; Laver, W. G.; Stevens, R. C.
J. Am. Chem. Soc. 1997, 119, 681.

19 Haslam, E. Shikimic Acid: Metabolism and Metabolites; Wiley & Sons: New York,
1993, p. 40.

2° Draths, K. M.; Knop, D. R.; Frost, J. W. J. Am. Chem. Soc. 1999, 121, 1603.

21 (a) Goncharoff, P.; Nichols, B. P.; J. Bacterial. 1984, 159, 57. (b) Green, J. M.;
Merkel, W. K.; Nichols, B. P. J. Bacterial. 1992, 174, 5317. (0) Kaplan, J. B.; Nicols,
B. P.; J. Mal. Biol. 1983, I68, 451.

22 Szmant, H. H. In Organic Building Blocks of the Chemical Industry; New York:
Wiley, 1989, p. 465.

23. Lenga, R. E.; Votoupal, K. L. The Sigma-Aldrich Libray of Regulatory and Safety
Data; Sigma-Aldrich: Milwaukee, WI, 1993.

24 Siebert, M.; Berchthold, A.; Melzer, M.; May, U. Berger, U. FEBS Lett. 1992, 307,
347.

25 Barker, J .; Draths, K. D.; Frost, J. W. Unpublished result.
26 Kirsch, M. A.; Williams, D. J. CHEMTECH 1994, 24, 40.

27 Szmant, H. H. In Organic Building Blocks of the Chemical Industry; New York:
Wiley, 1989, p. 467.

28 Draths, K. M.; Ward, T. L.; Frost, J. W. J. Am. Chem. Soc. 1992, 114, 9725.

29 (a) Hanessian, S.; Beaulieu, P.; Dube, D. Tetrahedron lett. 1986, 27, 2332. (b)
Flack, J. R.; Yadagiri, P. J. Org. Chem. 1989, 54, 5851. (c) Molin, H. ; Pring, B. G.
Tetrahedron lett. 1985, 26, 677. (d) Montchamp, J .-L.; Piehler, L. T.; Frost, J. W. J.
Am. Chem. Soc. 1992, 114, 4453. (e) Widlanski, T.; Bender, S. L.; Knowles, J. R. J.
Am. Chem. Soc. 1989, II 1, 2299.

30 Knop. D. R.; Draths, K. M.; Frost, J. W. Unpublished result.

31 Szmant, H. H. In Organic Building Blocks of the Chemical Industry; New York:
Wiley, 1989, p. 512.

32 Krumenacker, L.; Constantini, M.; Pontal, P.; Sentrnac. J. In Kirt-Othmer
encyclopedia of chemical technology, 4th ed. Kroschwitz, J. I., Howe-Grant, M., Eds.;
Wiley, New York, 1995, vol 13, p 996.

33 Li, K.; Mikola, M. R.; Draths, K. M.; Worden, R. M.; Frost, J. W. Biotechnol.
Bioeng. 1999, 64, 61.

34 Richman, J. E.; Chang. Y.-C.; Kambourakis, S.l; Draths, K. M.; Almy, E.; Kristi, D.
S.; Strasburg, G. M.; Frost, J. W. J. Am. Chem. Soc. 1996, 118, 11587.

35 Kambourakis, 8.; Frost, J. W. Unpublished result.

36 Szmant, H. H. In Organic Building Blocks of the Chemical Industry; New York:
Wiley, 1989, p. 519.

183

37 Draths, K. M.; Frost, J. W. J. Am. Chem. Soc. 1995, 117, 2395.

33 Krumenacker, L.; Costantini, M.; Pontal, P.; Sentenac, J. In Kirt-Othmer encyclopedia
of chemical technology, Kroschwitz, J. 1., Howe-Grant, M., Eds.; Wiley, New York,
1995, vol 13, p 996.

39 (a) Franck, H.-G.; Stadelhofer, J. W. Industrial Aromatic Chemistry; Springer-Verlag:
New York, 1988; p. 183-190. (b) Varagnat, J. In Kirt-Othmer encyc10pedia of chemical
technology, 3rd ed.; Grayson, M., Ed.; Wiley: New York, 1981; vol. 13, p. 39-69. (0)
Szmant, H. H. In Organic Building Blocks of the Chemical Industry; New York: Wiley,
1989, p. 512-519.

40 Draths, K. M.; Frost, J. W. J. Am. Chem. Soc. 1994, 116, 399.

41 Neidle, E. L.; Omston, L. N. J. Bacteriol. 1986, I68, 815.

42 Davis, D.D.; Kemp, D. R. In Kirt-Othmer encyclopedia of chemical technology, 4th
ed.; Kroschwitz, J. I., Howe-Grant, M., Eds.; Wiley, New York, 1991, vol 1, p 466-
493.

43 Thiemens, M. H.: Trogler, W. C. Science 1991, 251, 932.

44 Dickinson, R. E.; Cicerone, R. J . Nature 1986, 319, 109.

45 Li, K; Frost, J. W. J. Am. Chem. Soc. 1998, 120, 10545.

46 (a) Aiba, S.; Tsunekawa, H.; Imanaka. T. Appl. Env. Microbial. 1982, 43, 289. (b)
Tsunekawa, H.; Imanaka, T.; Aiba, S. J. Gen. Microbial. 1980, 118, 253.

47 (a) Draths, K. M, Pompliano, D. L.; Conley, D. L; Frost, J. W.; Berry, A.;
Disbrow, G. L.; Staversky, R. J.; Lievense, J. C. J. Am. Chem. Soc. 1992,114, 3956.
(b) Gubler, M, Jetten, M., Lee, S. H., Sinskey, A. J. Appl. Env. Microbial. 1994,60,
2494. (0) Miller, J. E.; Backman, K.C.; O'Conner, M. J .; Hatch, R. T]. Ind. Microbial.
1987, 2, 143. (d) Patnaik, R.; Liao, J. C. Appl. Environ. Microbiol. 1994, 60, 3903.
(e) Patnaik, R.; Spitzer, R. G.: Liao, J. C. Biotechnol. Bioeng. 1995, 46, 361.

43 Pitard, A. J. In Escherichia coli and Salmonella typhimurium; Neidhardt, F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p. 368-394.

49 Takashi, O.; Garner, C.; Markley, J. L. ; Herrrnann, K. M. Proc. Natl. Acad. Sci.
USA 1982. 79, 5828.

50 Weaver, L. M.; Herrrnann, K. M. J. Bacterial. 1990, 172, 6581.

51 Ray, J. M.; Yanofsky, C.; Bauerle, R. J. Bacterial. 1988, 170, 5500.

52 Backman, K.C. U S. Patent 5,169,768, 1992.

53 Mori. M.; Yokota, A. ; Sugitomo, S.; Kawamura, K. Patent JP 62,205,782, 1987.

184

54 (a) Konstantinov, K. B.; Nishio, N.; Yoshida, T. Ferment. Bioeng. 1990, 70, 253.
(b) Konstantinov, K. B.; Nishio, N.; Seki, T.; Yoshida, T. Ferment. Bioeng. 1991, 71,
350.

55 (a) Draths, K. M.; Frost, J. W. J. Am. Chem. Soc. 1990, 112, 1657. (b) Frost, J.
W. US. Patent 5,168,056, 1992.

56 (a) Farabaugh, M. MS. Thesis, Michigan State University, April 1996. (b) Lu, J .-L.;
Liao, J. C. Biotechnol. 1997, 53, 132.

57 Clark, G. S. Perfum. Flavor. 1990, 15, 45.

53 (a) Esposito, L.; Formanek, K.; Kientz, G.; Mauger, F.; Maureaux, V.; Robert, G.;
Truchet, F. In Kirk-0thmer Encyclopedia of Chemical Technology; Fourth Ed.,
Kroschwitz, J. I.; Howe-Grant, M., Ed.; Wiley: New York, 1997; Vol. 24, p 812. (b)
Van Ness, J. H. In Kirk-Othmer Encyclopedia of Chemical Technology, 3rd Edition;
Wiley: New York, 1983, Volum 23, p. 704.

59 Burri, J.; Graf, M.; Lambelet, P.; Loliger, J. J Sci. Food Agric. 1989, 48, 49.
‘50 Guarino, P. A.; Brown, S. M. J. Assoc. Ofi. Anal. Chem. 1985, 68, 1198.

51 Ranadive, A. S. In Spices, Herbs, and Edible Fungi; Charalambous, G., Ed.; Elsevier:
Amsterdam, 1994; p. 517.

62 Westcott, R. J.; Cheetham, P. S. J.; Barraclough, A. J. phytachemistry 1994, 35,
135.

63 (a) Lewis, Sr., R. J. Hazardous Chemicals Desk Reference, Third Edition; Van
Nostrand Reinhold: New York, 1993. (b) Lenga, R. E.; Votoupal, K. L. The Sigma-
Aldrich Libray of Regulatory and Safety Data; Sigma-Aldrich: Milwaukee, WI, 1993.

54 Benz, 1.; Muheim, A. Spec. Publ. R. Soc. Chem. 1996, 197, 111.

65 (a) Falconnier, B.; Lapierre, C.; Lesage-Meessen, L.; Yonnet, G.; Brunerie, P.;
Colonna Ceccaldi, B.; Corrieu, G.; Asther, M. J. Biotechnol. 1994, 37, 123. (b) Lesage-
Meessen, L.; Delattre, M.; Haon, M.; Thibault, J. F.; Colonna Ceccaldi, B.; Brunerie, P.;
Asther, M. J. Biotechnol. 1996, 50, 107. (c) Lesage-Meessen, L.; Haon, M.; Delattre,
M.; Thibault, J. F.; Colonna Ceccaldi, B.; Asther, M. Appl. Microbial. Biotechnol. 1997,
47, 393.

65 (a) Audras, B.; More, J. PCT Int. Appl. WO 9634971 A1, Nov. 7, 1996. (b)
Takasago Co Ltd, Japanese patent application 35338, Sept. 7, 1993.

67 Frenkel-Havkin, D.; Dorn, R. ACS Symp. Ser. 1997, 660 (Spices), 29.
63 Pittard, J.; Wallace, B. J. J. Bacterial. 1966, 91, 1494.

69 (a) Snell, K. D.; Frost, J. W. J. Am. Chem. Soc. 1993, 115, 11581. (b) Snell, K.
D.; Draths, K. M.; Frost, J. W. J. Am. Chem. Soc. 1996, 118, 5605.

70 Brosius, J.; Holy, A. Pmc. Natl. Acad. Sci. USA 1984, 81, 6929.

185

71 Bartolome, B.; Jubete, Y.; Martines, E.; de la Cruz, F. Gene, 1991, 102, 75.

72 Hamilton, C. M.; Aldea, M.; Washbum, B. K.; Babitzke, P.; Kushner, S. R. J.
Bacteriol. 1989, 171, 4617.

73 Zhu, L.; Frost, J. W. Unpublished result.
74 Guldberg, H. C.; Marsden, C. A. Pharmacol. Rev. 1975, 2 7, 135.

75 (a) Salminen, M.; Lundstrom, K.; Tilgmann, C.; Savolainen, R.; Kalkkinen, N.;
Ulmanen, I. Gene 1990, 93, 241. (b) Lundstrom, K.; Tilgmann, C.; Peranen, J.;
Kalkkinen, N.; Ulmanen, I. Biochim. Biophys. Acta 1992, 1129, 149.

76 Vidgren, J.; Svensson, A.; Liljas, A. Nature, 1994, 368, 354.

77 (a) Markham, G. D.; DeParasis, J.; Gatmaitan, J. J. Bio. Chem. 1984, 259, 14505.
(b) Markham, G. D.; Hafner, E. W.; Tabor, C.W., Tabor, H. J. Bio. Chem. 1980, 255,
9082. (c) Satishchandran, C.; Taylor, J. C.; Markham, G. D. Mal. Microbia. 1993, 9,
835. (d) Newman, E. B.; Budman, L. 1.; Chan, E. C.; Greene, R. C.; Lin, R. T.;
Woldringh, C. L.; D'Ari, R. J. Bacteriol. 1998, 180, 3614.

73 Greene, R. In Escherichia coli and Salmonella typhimurium; 2nd Ed., Neidhardt, F.
0, Ed.; American Society for Microbiology: Washington, DC, 1996; Vol. 1, p.542-560.

79 Li, T.; Rosazza, J. P. N.; J. Bacterial. 1997, 179, 3482.

30 (a) Gross, G. G.; Bolkart, K. H.; Zenk, M. H. Biochem. Biophys. Res. Commun.
1968, 32, 173. (b) Gross, G. G.; Zenk, M. H. Eur. J. Biochem. 1969, 8, 413. (0)
Gross, G. G. Eur. J. Biochem. 1972, 31, 585. (d) Zenk, M. H.; Gross, G. G. Recent
Adv. Phytachem. 1972, 4, 87.

31 Kirk, T. K.; Schultz, E.; Connors, W. J.; Lorenz, L. F.; Zeikus, J. G. Arch.
Microbiol. 1978, 117, 277.

82 Campbell, C. J.; Laherrere, J. H. Sci. Am. 1998, 278(3), 78.

83 (a) Vapnek, D.; Hautala, J. A.; Jacobson, J. A.; Giles, N. H.; Kushner, S. R. Prac.
Natl. Acad. Sci. USA 1977, 74, 3508. (b) Schweizer, M.; Case, M. BR; Dykstra, C.
C.; Giles, N. H.; Kushner, S. R. Gene 1981, I4, 23. (c) Buxton, F. P.; Radford, A.;
Mol. Gen. Genet. 1983, 190, 403.

84 Turner, G. In Applied Malecilar Genetics of Fungi; Peberdy, J. F.; Caten, C. E.;
Ogden, J. E.; Bennett, J. W., Ed.; Cambridge Unversity Press: Cambridge, 1991,
Chapter 2, p. 31.

85 Lodish, H.; Baltimore, D.; Berk, A.; Zipursky, S. L.; Matsudaira, P.; Darnell, J.
Molecular Cell Biology Freeman:New York, 1995.

36 Sambrook, J .; Fritsch, E. F.; Maniatis, T. Molecular Cloning: A laboratory Manual;
2nd ed.; Cold Spring Harbor Laboratory: Plainview, NY, 1990.

37 Sagerstroem, C. G.; Sun, B. 1.; Sive, H, L. Annu. Rev. Biochem. 1997, 66, 751.

186

33 Short, J. M.; Femandezx, J. M.; Sarge, J. A.; Huse, W. D. Nucleic Acids Res.
1988, 16, 7583.

39 Coward, J. K.; Slisz, E. P.; Wu, F. Y.-H. Biochemistry 1973, 12, 2291.

90 Li, K.; Frost, J. W. Unpublished result.

91 Holloway, C. T.; Greene, R. C.; Su, C.-H. J. Bacterial. 1970, 104, 734.

92 (a) Heiland, P. G; Hill, F. F. Process Biochem. 1993, 28, 171. (b) Shiomi, N. ;
Fukuda, H.; Murata, K.; Kimura, A. Appl. Micro. Biotech. 1995, 42, 730. (c) Shionri,
N. et.al. Biotech. Bioeng. 1990, 35, 1120.

93 (a) Barra, J. L.; Mautino, M. R.; Rosa, A. L. Genetics, 1996, 144, 1455. (b)
Lawrece, D. A.; Smith, D. A.; RowBury, R. J. Genetics, 1968, 58, 473. (c) Mautino,
M. R.; Barra, J. L.; Rosa, A. L. Genetics, 1996, 142, 789.

94 Su, C.-H.; Greene, R. C. Prac. Natl. Acad.Sci. USA 1971, 68, 367.

95 Reenila, I.; Tuomainen, P.; Tilgmann, C. Mannisto, P. T. Pharmacol. Toxicol. 1995,
77, 414.

96 Frost, J. W., Draths, K. M. Annu Rev. Microbiol. 1995, 49, 557.

97 Mikola, M. R. M.S. thesis, 1996, Michigan State University.

93 (a) Hamilton, C. M.; Aldea, M.; Washburn, B. K. Babitzke, P.; Kushner S. R. J.
Bacterial. 1989, 171, 4617. (b) Ohta, K.; Beall, D. S.; Mejia, J. P.; Shanmugarn, K. T.;
Ingram, L. 0. Appl. Environ. Microbiol. 1991, 57, 893.

99 Tobey, K. L.; Grant, G. A. J. Biol. Chem. 1986, 261, 12179.

‘00 Ogino, T.; garner, C.; Markley, J. L.; Herrrnann, K. M. Prac. Natl. Acad. Sci. USA
1982, 79, 5828.

101 Mikola, M. R., Widman, M. T., Worden, R. M. Appl. Biochem. Biotechnol. 1997,
70, 905.

102 Weaver, L. M.; Herrrnann, K. M. J. Bacterial. 1990, 172, 6581.

103 Fiirste, J. P.; Pansegrau, W.; Frank, R.; Blocker, H.; Scholz, P.; Bagdasarian, M.;
Lanka, E.; Gene 1986, 48, 119.

104 Tribe, D. E.; Pittard, J. Appl. Environ. Microbiol. 1979, 38, 181.

105 Tanaka, K.; Takayanagi, Y.; Fujita, N. Ishihama, A.; Takahashi, H. Prac. Natl.
Acad. Sci. USA 1993, 90. 3511.

106 (a) Paoletti, P.; Williams, J. F.; Horecker, B. L. Anal. Biochem. 1979, 95, 250. (b)

William, J. F.; blackmore, P. F.; Duke, C. C.; Macleod, J. F. Int. J. Biochem. 1980, 12,
339.

187

107 (a) Blackmore, P. F.; William, J. F.; MacLeod, J. F. FEBS Lett. 1976, 64, 222. (b)
Dike, C. C.; Macleod, J. F.; Williams, J. F. Carbohydr. Res. 1981, 95, 1.

103 Lu, J.-L, Liao, J. C. Biotechnol. Bioeng. 1997, 53, 132.

109 Forberg, C.; Eliaeson, T.; Haggstr'om, L. J. Biotechnol. 1988, 7, 319.
”0 Papoutsakis, E. T. Biotechnol. Bioeng. 1984, 26, 174.

”1 Draths, K. M.; Frost, J. W. J. Am. Chem. Soc. 1990, 112, 9630.

“2 Gross, G. G. In Plant polyphenols, Hemingway, R. W.; Lakes, P. E.; Ed. New
York: Plenum, 1992, p. 43-60.

”3 Rivero, F.; Cerda-Olmedo, E. Mycalagia, 1994, 86, 781.
114 Brown, K. D.; Doy, C. H. Biochim. Biophys. Acta 1976, 42, 550.

“5 (a) Davis, B. D.; Givarg, C.; Mitsuhashi, S. Methods Enzymal. 1955, 2, 300. (b)
Mitsuhashi, S.; Davis, B. D. Biochim. Biophys. Acta. 1954, 15, 268.

”6 (a) Chen, R. Z.; Yap, W. M. G. J.; Postma, P. W. Bailey, J. E. Biotechnol. Bioeng.
1997, 56, 583. (b) Flores, N.; Xiao, J.; Berry, A.; Bolivar, F.; Valle, F. Nat.
Biotechnol. 1996, 14, 620. (0) Miller, J. E.; Backman, K. C.; O'Cornor, M. J .; Hatch,
R. T. J. Ind. Microbiol. 1987, 2, 143.

“7 Bothast, R. J .; Grohmann, K. Saccharification of Com Fiber by Combined Treatment
with Dilute Sulphuric Acid and Enzymes. Process Biochem. 1997, 32, 405-415.

113 May. J. B. In Corn: Chemistry and Technology; Watson, S. A.; Ramstad, P.E., Ed.
American Association of Cereal Chemists: St. Paul, MN, 1987; p.377-397.

“9 Park, C.; Song, S. Organization and Regulation of the D-Xylose Operons in
Escherichia coli K-12: Xle Acts as a Transcriptional Activator. J. Bacterial. 1997 , 179,
7025-7032.

120 (a) Fumagalli, C.; Spa, L. In Kirk-Othmer Encyclopedia of Chemical Technology, 4th
ed.; Kroschwitz, J. 1., Howe-Grant, M., Eds.; Wiley: New York, 1997; Vol. 22, p. 1074.
(b) Felthouse, T. R.; Burnett. J. C.; Mitchell, S. F.; Mummey, M. J. In Kirk-Othmer
Encyclopedia of Chemical Technology, 4th ed.; Kroschwitz, J. I., Howe-Grant, M., Eds.;
Wiley: New York, 1997; Vol. 15, p. 893. (c) Szmant, H. H. Organic Building Blocks of
the Chemical Industry; Wiley: New York 1989; p. 363.

121 Medina, V.; Pontarollo, R.; Glaeske, D.; Tabel, H.; Goldie, H. J. Bacterial. 1990,
I 72, 7151.

122 Moniruzzaman, M.; Dien, B. S.; Ferret, B.; Hespell, R. B.; Dale, B. E.; Ingram, L.
O.; Bothast, R. J. Biotech. Letts. 1996, 18, 985-990.

123 Bothast, R. J .; Grohmann, K. Process Biochem. 1997, 32, 405-415.
124 Gupta, S. C.; Leathers. T. D. Appl. Bioch. Biotech. 1996, 59, 337-347.

188

125 Carlson, T. In Prelininary Program Proceedings, Corn Utilization Conference V, June
8-10, 1994; National Corn growers Association: St. Louis, MO.

126 Lendenmann, U.; Egli, T. Microbiology 1995, I41, 71.

127 (a) Millard, C. S.; Chao, Y-P; Liao, J. C.; Donnelly, M. 1. Appl. Environ. Microbia.
1996, 62, 1808. (b) Stols, L.; Donnelly, M. 1. Appl. Environ. Microbia. 1997, 63,
2695.

128 Goldie, H.; Medina, V. Mol. Gen. Genet. 1990, 220, 191.

129 (a) Kay, W. W.; Kornberg, H. L. Eur. J. Biochem. 1971, I8, 274. (b) Lo, T. C.
Y.; Rayman, K., Sanwal, B. D. J. Biol. Chem. 1972, 247, 6323.

130 Ferenci, T. FEMS Microbiol. Rev. 1996, 18, 301.
131 True, W. R. Oil Gas J. 1998, 96(20), 49.

132 Current Protocols in Molecular Biology; Ausubel, F. M.; Brent, R.; Kingston, R. E.;
Moore, R. E.; Seidman, J. G.; Srrrith, J. S.; Struhl, K. Eds.; Wiley: New York, 1987.

133 Zurawski, G.; Gunsalus, R. P.; Brown, K. D.; Yanofsky, C. J. Mal. Biol. 1981,
145, 47.

134 Miller, J. H. Experiments in Molecular Genetics; Cold Spring Harbor Laboratory:
Plainview, NY, 1972.

135 Silhavy, T. J.; Berman, M. L.; Enquist, L. W. Experiments with Gene Fusions; Cold
Spring Harbor Laboratory: Plainview, NY, 1984.

136 Schoner, R.; Herrrnann, K. M. J. Biol. Chem. 1976, 251 , 5440.

137 Gollub, E.; Zalkin, H.; Sprinson, D. B. Meth. Enzymal. 1971, 17A, 349.

133 Stroman, P.; Reinert, W. R.; Giles, N. H. J. Biol. Chem. 1978, 253, 4593.
139 Reenila, I.; Tuomainen, P.; Mannisté, P. T. J. Chromatogr. B 1995, 663, 414.
140 Dell, K.A. Ph.D. Thesis, Purdue University, December, 1993.

189

MICHIGAN smt: UNIV. LIBRARIES
llllmllllll”Will“lllllllllllllllllllll
31293021061860