[.1 a. a. $1.53.;
a

a...» Faun... in}. .
:5». A .
at}. . .

L

1.1

$.31,

bum:

. .2
1 I ‘

Q: n
S.“ I: MMWH‘WAM‘KKMIMM
Sun.

I i.
«TREE . .
.3: .. I
In“: E
7...;

CL

 

 

I
n .

c

 

hlu

n

 

I‘vhltc r I. >D p :00: K . : IA .3 a v 1017.1
- A111.» I |u . A v ”Ir. II 1|
ll |'l I‘I

 

.. . in? g ,i ‘ . L .. ; . . o: , m.
.1.c.5!.l-... .. : i. 1.... -. .3
n x . D II ‘ ‘ r k .‘ In .1] \\1 . olo‘olh‘ 1|,y iv.- 4'1“1 r:
i -I 3: It!!!

THESIS

Illlllllllllllllll ! '

293 01565 93641

                                                                           

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the

thesis entitled

Optimization of Biocatalytic 3-Dehydroshikimic Acid
Production from D—Glucose in Escherichia Coli

presented by

Mark R. Mikola

has been accepted towards fulfillment
of the requirements for

M.S. Chemical Engineering

degree in

 

 

KW WW

Major professor

Date i'lZl M7

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date duo.

DATE DUE ' DATE DUE DATE DUE

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I. An Affirmative Action/Equal Opportunity Institution
Wyn-9.1

 

OPTIMIZATION OF BIOCATALYTIC 3-DEHYDROSHIKIMIC ACID PRODUCTION
FROM D-GLUCOSE IN ESCHERICHIA COLI
By
Mark Raymond Mikola

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

MASTER OF SCIENCE

Department of Chemical Engineering
1996

ABSTRACT

OPTIMIZATION OF BIOCATALYTIC 3-DEHYDROSHIKINIIC ACID PRODUCTION
FROM D-GLUCOSE IN ESCHERICHIA COLI

By

Mark Raymond Mikola

Biocatalysis is an environmentally friendly alternative to petrochemical processing
for production of small molecules. The biocatalytic process objective is efficient and high-
yielding conversion of starting material to product. Success of the process depends on
manipulation of metabolic fluxes through desired pathways by genetic manipulaan and
process engineering.

A strain of Escherichia coli was genetically engineered to overproduce
dehydroshikimic acid (DHS). This intermediate of the common aromatic amino acid
pathway is of commercial interest, because it can readily converted into a variety of
valuable products. This E. coli strain features a feedback-insensitive DAHP synthase
(aroF‘h') that increases carbon flux into the common pathway. A fed-batch fermentation
process was developed using this strain resulting in a DHS titer of 40 g L" in 48 h. A
novel method was developed to obtain the aroF" where ultraviolet light mutagenesis was
followed by phenotype selection utilizing chemotaxis in a diffusion gradient chamber
(DGC).

To my fiancée Anna, and my family for their love and support

ACKNOWLEDGMENTS .

First of all I would like to express my gratitude to Dr. R. Mark Worden and Dr.
John W. Frost for their encouragement and support. I hope the collaboration between the
two research groups will successfully continue.

I wish to thankfufly acknowledge Dr. Karen Draths and Kai Li, each for their
contributions to the research. I want to also thank Mark Widman for his help with the
DOC. I am very grateful to all the members of both Dr. Worden’s and Dr. Frost’s research
group, for their advice, support, and friendship. I wish everyone the best in their
endeavors.

Lastly, I wish to thank to thank my family and especially Anna. My parents have
given me encouragement and support at the most needed times. I wish to thank Anna

because of her help and great patience enabling me to concentrate on my graduate career.

iv

TABLE OF CONTENTS

LIST OF TABLES ......................................................................... vii

LIST OF FIGURES ...................................................................... viii

LIST OF ABBREVIATIONS ............................................................. x

CHAPTER 1

INTRODUCTION .......................................................................... l
Biocatalyst .......................................................................... 2
Common Pathway of Aromatic Amino Acid Biosynthesis .................... 2
Common Pathway Metabolite Uses .............................................. 5
Central Metabolism ................................................................ 6
Selection of DAHP Synthase Isozyme for Overexpression ................... 7
Feedback-Insensitive AroF ....................................................... 8
Increasing DHS Production ....................................................... 9
Metabolic Control Analysis and Metabolic Engineering ...................... 11
Plasmid Maintenance ............................................................. 12
Fermentor .......................................................................... 14
Research Objectives ............................................................... 15
LIST OF REFERENCES ........................................................ 16

CHAPTER 2 .

TYROSINE FEEDBACK INSENSITIVE DAHP SYNTHASE ..................... 21
Background ........................................................................ 21
Molecular Biology Background ................................................. 23
Development of the Strain for Mutagenesis .................................... 23
UV Mutagenesis ................................................................... 29
Development and Use of the Diffusion Gradient Chamber Method ......... 30
DGC Chemotaxis Experiments .................................................. 34
Generation and Isolation of Fee'dback-Insensitive Mutant ................... 36
Confirmation ofFeedback-Insensitivity .............. 38
Isolation of the Feedback-Insensitive aroF ..................................... 41
Sequencing of the Feedback-InsenSitive aroF ................................. 42
Feedback Insensitive AroF Specific Activity .................................. 43
Discussion and Conclusion ...................................................... 53
LIST OF REFERENCES ........................................................ 56

CHAPTER 3

FERMENTOR PRODUCTION OF DHS ............................................... 59
Fermentor Advantages ............................................................ 59
Biosynthetic Pathways for DHS Production ................................... 60
Optimization Strategy for DHS Production .................................... 63

V

Development of a Microbial Catalyst ........................................... 64

Development of a New Host ..................................................... 64
B. Braun Fermentor .............................................................. 65
Optimized Fermentor Conditions ................................................ 66
NMR Assays ...................................................................... 71
DAHP Synthase Assays .......................................................... 72
Confirmation of Plasmid Maintenance .......................................... 79
Fermentations ...................................................................... 79
Scaled-Up Fermentations ........................................................ 88
Discussion and Conclusions ..................................................... 95
LIST OF REFERENCES ........................................................ 99
CHAPTER 4 .
EXPERIMENTAL ....................................................................... 101
General Methods ................................................................ 101
Spectroscopic Measurements ......................................... 101
Bacterial Strains ........................................................ 101
Storage of Bacterial Stains ............................................ 101
Culture Medium ........................................................ 102
1H NMR Analysis of Culture Supernatant .......................... 102
Restriction Enzyme Digest of DNA .................................. 103
Agarose Gel Electrophoresis .......................................... 103
Preparation and Transformation of Competent Cells .............. 103
Enzyme Assay .......................................................... 104
Chapter 2 Experimental ......................................................... 106
Insertion of a Synthetic Cassette into serA of AB3248 ............ 106
UV Mutagenesis ........................................................ 107
Diffusion Gradient Chamber (DGC) Setup ......................... 107
Recovery of Cells ...................................................... 110
Culture Conditions for
DAHP Synthase Assay of DGC Mutant .................... 111
Isolation of the Feedback-Insemitive aroF gene ................. 111
Isolation of the Feedback-Insensitive aroF gene ................... 111
Generation of Truncated aroF‘”r ...................................... 1 12
Replacement of the aroF“ Native Promoter
with the tac Promoter ......................................... l 12
Shake Flask Culture Conditions for
aroFb‘ Specific activity Evaluations ......................... 112
Chapter 3 Experimental ......................................................... 114
B. Braun Fermentor ................................................... 114
Fermentor Medium. .................................................... 1 15
Fermentor Inoculum for Evaluation of
KL3/pKL4.32B and KL3/pKL4.3BB ...................... 115
Optimized Fermentor Inoculum ...................................... 1 16
Fermentor Conditions ................................................. 1 16
Fermentor Setup ........................................................ 117
Fermentor Samples .................................................... 1 19
LIST OF REFERENCES ...................................................... 120

LIST OF TABLES

Table 1 - PCR Primers .................................................................... 27
Table 2 - UV Radiation Survival ......................................................... 30
Table 3 - Fermentor Medium Components .............................................. 67
Table 4 - Trace Mineral Supplements .................................................... 68
Table 5 - Phase 2 DO. PID Controller Parameters .................................... 70
Table 6 - Starter Culture for Evaluation of the Truncated/Nontruncated aroF ...... 80
Table 7 - Summary of Fermentations .................................................... 94

LIST OF FIGURES

 

Figure 1 - The Common Pathway ......................................................... 3
Figure 2 - Alternate Molecule Production From The Common Pathway ............. 5
Figure 3 - Plasmid Map of pKAD76A..‘ ............. . ................................... 25
Figure 4 - Synthetic Cassette ............................................................. 26
Figure 5 - Plasmid Map of pKADlO.156A ............................................. 27
Figure 6 - Plasmid Map of pKADlO.186A ............................................. 28
Figure 7 - Diagram of a Diffusion Gradient Chamber ................................. 32
Figure 8a - Chemotaxis of AB2.24 ...................................................... 36
Figure 8b - Image Analysis of Centerline ............................................... 36
Figure 9a - DGC Prior to Mutagenesis (3 days) ........................................ 37
Figure 9b - DGC 7 Days after Mutagenesis ............................................. 37
Figure 9c - DGC 9 Days after Mutagenesis ............................................. 37
Figure 10 - Mutant Strain Tyrosine Insensitivity Assay ............................... 39
Figure 11 - Percent Relative Activity .................................................... 40
Figure 12 - Phenylalanine and Tryptophan Sensitivity Test .......................... 41
Figure 13 - Plasmid Map of Cloned Mutants ........................................... 42
Figure 14 - Map of Feedback Insensitive aroF ......................................... 43
Figure 15 - Cloned AroF Specific Activity .............................................. 44
Figure 16 - Map of aroF tyr Boxes ...................................................... 45
Figure 17 - Plasmid Map of pCL2-13A-trunc .......................................... 46
Figure 18 - Nontruncated and Truncated AroF Specific Activity ..................... 46
Figure 19 - Minimal Medium Nontruncated and Truncated AroF activity ........... 47
Figure 20 - Plasmid Map of pKL4.32B and pKL4.33B .......................... ,. . . .48
Figure 21 - Truncated/Nontruncated AroF Activity and DHS Production .......... 49
Figure 22 - Plasmid Maps of pKL4.66A and pIGA.66B ............................. 50
Figure 23 - Specific Activity and DHS Production of pKL4.66A/B ................. 51
Figure 24 - Plasmid Map of pKL4.79B ................................................. 52
Figure 25 - AroF Specific Activity and DHS Production of pKL4.79B ............ 52
Figure 26 - Biosynthetic Pathways for DHS Production .............................. 61
Figure 27 - Flux Distribution for Maximum DHS Yield .............................. 63
Figure 28 - Braun Fermentor ............................................................. 65
Figure 29 -’H NMR of 3-Dehydroshkirnic Acid (DHS) ............................... 73
Figure 30 -‘H NMR of Glucose ......................................................... 74
Figure 31 - 1H NMR of Gallic Acid ...................................................... 75
Figure 32 - 1H NMR of Acetic Acid ..................................................... 76
Figure 33 - 'H NMR of Formic Acid .................................................... 77
Figure 34 - Typical 1H NMR for 36 hour Fermentation ............................... 78
Figure 35 - Fermentation Evaluating KL3/pKL4.32B and KL3/pKL4.33B ........ 81
Figure 36 - Nonoptimized Fermentation ................................................ 82
Figure 37 - KL3/pKL4.66A Fermentation .............................................. 84
Figure 38 - Fermentations of KL3/pKL4.66A and KL3/pKL4.66B ................ 85
Figure 39 - Fermentations of KL3/pK1A.33B ......................................... 86
Figure 40 - Fermentation of KL3/pKL4.79B 1.0 g IPTG ............................ 89

Figure 41 - Fermentations of KL3/pKL4.79B, IPTG Titration of aroF ............. 90
Figure 42 - DHS Production as a Function of AroF Activity ......................... 91
Figure 43 - 50 L Scale Fermentation ..................................................... 93

Cm
DAHP
DHQ
DHS

PEP
PCR

8pc
TSP

LIST OF ABBREVIATIONS

ampicillin

chloramphenicol
3-deoxy-D-arabino-heptulosonic acid 7 -phosphate
3-dehydroquinate

3-dehydroshikimic acid

D—erythrose 4—phosphate

hour

isopropyl BoD-thiogalactopyranoside
kanamycin

kilobasc

Luria broth

m-fluorotyrosine

minute

nuclear magnetic resonance
phosphoenolpyruvate

polymerase chain reaction

Ribosome Binding Site

spectinomycin

sodium 3-(trimethy1silyl)propionate-2,2,3,3,-d,

CHAPTER]

INTRODUCTION

The development and use of microorganisms for biocatalytic production of small
molecules is a growing technology as indicated by the increasing demand for bioreactors
and projected biotechnology sales forecasts.‘ A factor in this increase is the use of
biological production for aromatic compounds and their chemical precursors as an
alternative to traditional synthetic routes based on petroleum feedstocks.2 Biocatalysis
utilizes benign, inexpensive, and renewable starting materials such as D-glucose. In
comparison, traditional chemical synthesis utilizes petroleum-derived starting materials
such as benzene and other hydrocarbons that are in many cases toxic. Chemical synthesis
often has the disadvantage of producing a mixture of products and byproducts that require
separation. Many of the environmentally harmful aspects of using petroleum feedstocks
are a result of the chemistry required to add oxygen atoms to the starting materials and
intermediates. Petroleum typically has a low ratio of oxygen to carbon, whereas plant-
derived feedstocks, such as starch and cellulose, have a high oxygen to carbon ratio.
Petroleum is a nonrenewable resource that has been plagued by inadvertent releases into the
environment. Therefore, the advantages of biocatalytic production are the benefits of
exploiting starting materials such as D-glucose and the initially high oxygen to carbon ratio
of the feedstocks. However, efficient and high-yielding conversion of starting material to
product requires optimization of the microbial catalyst by genetic modification and

optimization of the fermentation conditions.

2
W

A plethora of engineered microorganism have been exploited for biocatalytic
synthesis of aromatic amino acids. A partial listing of these microbes include
Corynebacterium glutamivum,3 Brevibacterium lactofermentum,4 and Escherichia coli.5
The organisms utilized in this study are E. coli K-12 derivatives. E. coli are facultative,
gram-negative, rod shaped, prokaryotic bacteria of the family Enterobacteriaceae. There
are many reasons for choosing E. coli. It is a well studied organism. The K-12 strain was
given preferential treatment by the National Institutes of Health guidelines for work with
recombinant organisms because the safety of K-12 strains were more actively investigated.‘S
E. coli have high growth rates, rapidly metabolize substrates, require inexpensive medium
components, are physically rugged, and survive in a wide variety of environmental
conditions. Also, an advantage to using E. coli for the present study was the necessary
phenotype to produce 3-dehydroshikimic acid (DHS, the desired fermentation product) had
already been created.7 Two derivatives of K-12, AB2834 and AB3248, were previously
obtained by the Frost group from the Yale Genetic Stock Center. AB2834 had the desired
phenotype for DHS production, and AB3248 had the phenotype necessary to be used as
the host to study a key enzyme in the metabolic pathway.8 I

C E l E E . E . i . I E' l .

The elucidation in the early 19603 of the common aromatic amino acid pathway
utilized bacterial auxotrophs of E. coli and Kleblsiella areogenes.9 The pathway consists of
seven enzymatically catalyzed reactions that convert D—erythrose 4-phosphate (E4P) and
phosphoenolpyruvate (PEP) to chorismic acid (Figure 1). The pathway is named for the
fact that it is the source for the common precursor of aromatic amino acids and related
secondary metabolites. This pathway is also present in other bacteria, plants, fungi, and
molds.10 PEP is produced through glycolysis, and E4P is produced from the non-
oxidative branch of the pentose phosphate pathway.

0H 2‘00.“ HO . 002” HO I 002“
'0” PEP A ' '
o . —* V OH-BTPiO
-—-’
0H

.-. Hzoaao OH
HO OH H pi Hzoam 6H
D-glucose H8 0 DAHP DHQ
E4P 0
H2

002“ 0021-1 NADP NADPH COZH

L-phenylalanine

L-tyrosine *_‘

L-Wptophan o JLcozn HaHo" OH #0 - OH
5H 5H
chorismate shikimic OHacid DHS

(A) DAHP synthase (aroF, aroG, aroH); (B) DHQ synthase (aroB); (C) DHQ dehydratase
(aroD); (D) shikimate dehydrogenase (aroE).

Figure 1 - The Common Pathway.

In E. coli, the first committed step in the common pathway is the condensation
between E4P and PEP to form 3-deoxy-D-arabino-heptulosonate 7 -phosphate (DAHP) and
inorganic phosphate. This reaction is catalyzed by three DAHP synthase isozymes that are
encoded by aroF, aroG, and araH, whose products are sensitive to feedback inhibition“ by
L-tyrosine, L-phenylalanine, and L-tryptophan, respectively (typically, names of genes are
italicized, and the first letter of protein names are capitalized). At the repressor-mediated
transcriptional level, aroF and aroG are repressed by the TyrR protein and aroH is
repressed by the TrpR protein.‘2

The tyrosine-sensitive isozyme of DAHP synthase (AroF) was used in the
constructs for DHS production, and the choice for the use of this isozyme is detailed later
in this chapter (p 7). It is an iron-containing dimer with subunit molecular weight of about
40,000.13 When E. coli are grown in minimal salts medium supplemented with all the

amino acids except the aromatic amino acids, AroF enzyme comprises the major isozyme."

The reaction mechanism is ordered and sequential with PEP being the first substrate to
bind. The KIn value (the substrate concentration at which the reaction rate is half its
maximal value) for PEP has been reported to be 5.8 M for the purified enzyme.” It is
assumed that PEP-enzyme complex is the native form of the enzyme since the intracellular
concentration of PEP never falls below 88 uM, and enzyme activity is lost and
unrecoverable during purification if the enzyme buffer solution does not contain PEP.

The second enzyme in this pathway, dehydroquinate synthase, is encoded by the
aroB locus. Dehydroquinate (DHQ) and inorganic phosphate are formed from DAHP in
this NAB-requiring reaction. AroB is a cobalt-requiring, monomeric protein with a
molecular weight of 44,000.‘6 Next DHQ dehydratase, encoded by aroD, catalyzes the
dehydration of DHQ to form DHS. This enzyme does not require a metal as do the
previous two enzymes. There are conflicting data concerning the molecular weight and
structure of AroD. The native enzyme has been reported to be both a dimer and a tetrarner
with a molecular weights, respectively, of 40,000 to 60,000.17 Both AroB and AroD are
constitutively expressed and are not sensitive to feedback inhibition by any of the aromatic
amino acids or chorsimate.18 It is noteworthy that DHS contains the first double bond of
what becomes the aromatic ring in phenylalanine, tyrosine, and tryptophan. The last
enzyme of importance for DHS production in the common pathway is aroB-encoded
shikimate dehydrogenase. This enzyme catalyzes the reduction of DHS to shikimic acid
and requires NADPH. Shikimate dehydrogenase is a monomeric protein with a molecular
weight of 32,000.” The host strain utilized in this study, AB2834, lacks shikimate
dehydrogenase activity and therefore cannot further metabolize DHS.20

Since the strain is unable to further metabolize DHS, it is unable to produce
chorismic acid, the last metabolite in the common pathway. Chorismic acid is the precursor
for six end products. Three terminal pathways lead to the aromatic amino acids
phenylalanine, tyrosine, and tryptophan. The three other end products include folic acid,

ubiquinone, and enterochelin, which are involved in coenzyme biosynthesis, electron
transport, and iron uptake, respectively.”
W

The common pathway provides a metabolic route to not only aromatic amino acids
and essential metabolites, but also to other pathway intermediates that are potentially
valuable starting materials for chemical synthesis. Through metabolic engineering, key
intermediates can be overexpressed, and their increased availability may further the use and
the demand for these unique molecules as starting materials for chemical synthesis.
Diversion of carbon flow from the common pathway to products not normally synthesized
by E. coli is possible by the addition of foreign genes to a genetically modified host strain.
Examples of novel syntheses using heterologous microbes include the production of
protocatechuic acid,22 catechol,23 and adipic acid,24 from glucose via the common pathway

intermediate DHS (Figure 2). Also, DHS has several possible industrial uses for abiotic

OH 002H 002H
0 "OH
OH —" 0 OH A B ”0 OH
; ; ——> HO OH ——> OH
HO OH OH OH '
D-glucose DHS gallic acid pyrogallol
C
002H Hozc
D Qua E
<ECJH ———>- OH -——>
OH COzH
PCA catechol adipic acid

(A) chemical oxidation; (B) enzymatic decarboxylation; (C) DHS dehydratase; (D)
enzymatic decarboxylation; (E) (i) catechol 1,2-dioxygenase, (ii) chemical hydrogenation.

Figure 2 - Alternate Molecule Production from The Common Pathway.

6

synthesis of aromatic compounds such as gallic acid and pyrogallol.” Thus, research to
increase DHS production in the fermentor can serve as the archetype for other common
pathway metabolites products such as shikimic acid and other products that can be
produced by the addition of foreign genes.

CentralMetalmh’sm

Glycolysis and the pentose phosphate pathway are two central routes of
carbohydrate metabolism.26 The pentose phosphate pathway connects glycolysis with
several biosynthetic pathways. It consists of an oxidative and non-oxidative branch. The
oxidative branch produces ribose S-phosphate and carbon dioxide from glucose 6-
phosphate, with concomitant production of NADPH. The non—oxidative branch coverts
glycolytic intermediates such as fructose 6-phosphate and glyceraldehyde 3-phosphate into
a variety of C-3 through C-7 monophosphate sugars, including E4P. Two enzymes of the
non-oxidative branch, transketolase and transaldolase, work in concert to produce E4P.
These two enzymes are essential to the formation of ribose 5-phosphate, sedoheptulose 7-
phosphate and E4P, which are required, respectively, for the biosynthesis of nucleotides,
lipopolysaccharide, and aromatic amino acids and vitamins.

The other substrate for DAHP synthase, PEP, is an intermdiate in the glycolytic
pathway. The production of PEP via glucose metabolism leads to a compeduon between
the PEP-requiring biosynthetic steps and the PEP-dependent phosphorotransferase system
(PTS) for glucose uptake.27 PEP is the phosphate donor used by PTS in the uptake and
phosphorylation of glucose. In the glycolytic pathway one mole of glucose produces two
moles of PEP. During glucose transport by the PTS one mole of PEP is converted into one
mole of pyruvate. Therefore, glucose transport results in only one mole of PEP per mole
of glucose consumed being available for biosynthesis. Wild-type E. coli apparently do not
recycle pyruvate back to PEP. Pyruvate formed from the PTS-rnediated glucose uptake is
further metabolized in the TCA cycle and used in the production of organic acids, carbon

dioxide, or cell mass.

7
, E : .

It was determined that the first step in the common pathway was limiting the carbon
flow into the pathway and its overexpression was necessary to increase the pathway flux.28
Since there are three isozymes of DAHP synthase present in wild-type E. coli, a selection
of which isozyme to use had to be made. The most important criteria for this selection
were optimal enzymatic activity and the effect enzymatic and transcriptional regulation.

The AroH isozyme was ruled out as the choice for overexpression. This isozyme
normally accounts for only a small portion of the DAHP synthase activity under a variety of
growth conditions.” In cells extracts, when fully derepressed, the activity of AroH is 10%
of the AroG activity and 20% of the AroF activity.30 It would be reasonable to assume the
lower activity of AroH is indicative of a weak native promoter,31 and significant
overexpression of AroH is not possible.

There are many similarities between AroG and AroF. Both enzymes contain one
mole of iron per mole of enzyme. There is approximately 50% identity in the nucleotide
sequence of AroG and AroF.32 Also aroG and aroF are transcriptionally repressed by the
TyrR protein.33 In addition, feedback insensitive mutants have been generated for both
AroF" and AroG.” ‘ A

AroG is a tetramer with a subunit molecular weight of 35,000. The AroF isozyme
is a dimer with a subunit molecular weight of 40,000. During growth in minimal medium,
without any amino acid supplementation, AroG comprises 80% or more of the total DAHP
synthase activity. Similarly, when E. coli are grown in iron-starved conditions AroG
comprises the majority of the DAHP synthase activity.36 Although, when E. coli are
grown on minimal salts medium supplemented with all amino acids except aromatic amino
acids, AroF accounts for most of the DAHP synthase activity. The AroG isozyme is 50%
inhibited by 13 11M phenylalanine, and AroF is 50% inhibited by 20 M tyrosine. During
the stationary phase of E. coli growth, AroF is significantly more sensitive to proteolysis

relative to AroG.37

8

There are certain advantages to each of the isozymes. The isozyme, AroG or
AroF, that comprises most of the DAHP synthase activity varies with growth conditions.
Since iron is added to the medium, the higher specific activity of AroG during iron
starvation is not significant. The higher specific activity of AroF under growth conditions
where aromatic amino acid concentrations are limiting is relevant since these are the
conditions anticipated for biocatalytic synthesis of DHS. Although, the main reason for
use of a feedback-insensitive AroF was because it has been successfully used in high titer
fermentations of L-phenylalanine and L-tryptophan.38 Several different mutations in AroF
confer insensitivity to feedback inhibition, and it was believed this would increase the
likelihood that a feedback—insensitive AroF could be generated in this lab. The feedback-
insensitive AroF was needed to increase the carbon flow into the pathway when E. coli is
grown in medium supplemented with tyrosine.

E I] l -I . . E E

Biocatalytic synthesis of various metabolites in high yields often requires alteration
of the cell’s natural regulatory mechanisms to increase carbon flow into selected
biosynthetic pathways.” Alterations can be performed genetically by mutation of genes,
changing the promoter, increasing the number of gene copies in the cell, or by overcoming
transcriptional repression. Regulatory mechanisms governing carbon flow into the
common pathway include both feedback inhibition of translated enzymes and
transcriptional repression.40

Feedback inhibition involves reduction of activity of key biosynthetic enzymes by
either pathway intermediates or products. Typically, the first enzyme in the common
biosynthetic pathways and the first enzyme in branch pathways are subject to regulation by
feedback inhibition. In the presence of excess inhibitory product, feedback inhibition
reduces the affected enzyme’s activity, which results in reduction of product biosynthesis.
Feedback inhibition has been demonstrated to be the dominant regulatory mechanism

controlling carbon flow into the pathway of aromatic amino acid biosynthesis in E. coli.“

9

Repression of transcription occurs when a repressor protein binds to a conserved
DNA sequence upstream from the structural gene. In the presence of the repressor,
transcription of the gene by RNA polymerase is blocked. Often the ability of a repressor
molecule to bind to DNA is dictated by binding of an effector molecule to the repressor.
Either biosynthetic end products or biosynthetic intermediates can be effector molecules.
An example of such regulation occurs in the aroF tyrA transcriptional unit. The structural
gene for aroF is preceded by three separate conserved sequences called ryrR boxes.42 A
repressor protein, TyrR, when activated by binding to the effector tyrosine, binds to the
tyrR box such that transcription cannot occur. In the presence of tyrosine, transcriptional
repression can lead to a 20-fold reduction in DAHP synthase activity relative to the activity
in the absence of tyrosine.‘3

Since the key enzyme AroF is regulated by both feedback inhibition and
transcriptional repression, increasing the in vivo activity of this isozyme of DAHP synthase
was considered essential. The host strain used for producing DHS is an aromatic amino
acid auxotroph that requires the addition of tyrosine, phenylalanine, and tryptophan, as
well as the aromatic vitamins for growth in minimal medium. A feedback-insensitive AroF
was thought to be needed to avoid feedback inhibition caused by in vivo, steady-state
concentration of tyrosine. The intracellular tyrosine concentration of AB2834 may be high
compared to the wild-type E. coli intracellular tyrosine concentration (since wild-type E.
coli would be grown in medium that is not supplemented with tyrosine) if the rate of active
transport of tyrosine into the cell is greater than the rate of tyrosine consumption by
incorporation into proteins. In addition, the pseudo steady state concentration of tyrosine
may be high even if the transport and initialization rates of tyrosine are the same. The
nature of AroF inhibition is noncompetitive with respect to E4P and competitive with
respect to PEP.“ The enzyme is 50% inhibited by 20 1.1M tyrosine.“

10

I . 12 HE E l .

The first step in the metabolic optimization of the common pathway was to increase
DAHP synthase levels.46 This was achieved by insertion of the locus that encoded for
either aroG or aroF onto a multicopy plasmid. Plasmid-based expression of DAHP
synthase increased the flux of carbon through the cormnon pathway, resulting in a situation
where subsequent enzymes became limiting."7

The next metabolic improvements were made using a strain that produced
phenylalanine. In this system it was determined that 3-dehydroquinate (DHQ) synthase
activity was insufficient to catalyze the conversion of DAHP into DHQ at a adequately rapid
rate to avoid DAHP accumulation (termed rate-limiting). An amplification of aroB-encoded
DHQ synthase expression as part of a synthetic cassette provided sufficient amplification to
remove the rate-limiting character of DHQ synthase.“8 Based on these findings, the
minimum requirement to have. increased carbon flow through the common pathway to
produce DHS is the overexpression both aroF and aroB.

Next, E4P in vivo availability was increased using the transketolase (tktA) locus
inserted into a multicopy plasmid. The plasmid also contained the locus for aroF. The
plasmid was transformed into a host that produced DAHP. The DAHP titer and yield
increased two-fold.49

A theoretical analysis of the pathways involved indicated that the yield was then
limited by PEP availability.’0 One option to increase PEP availability was to recycle
pyruvate generated through the PTS back to PEP, by using the gene for
phosphoenolpyruvate synthase (pps). This would increase the theoretical yield by a factor
of two (from 43 mol % to 86 mol %), since two moles of PEP produced from one mole of
glucose consumed would be available for biosynthesis. The gene was localized on a
plasmid and transformed into a strain that accumulates DAHP. The result was an increase

in the DAHP titer and an increase in the yield to near the new theoretical maximum.’ ’

11

An attempt to increase in vivo PEP availability was examined by the localizauon of
phosphoenolpyruvate carboxykinase (pck) on a plasmid. This enzyme converts
oxaloaeetate from the TCA cycle into PEP. Liao er. al. reported that the overexpression of
pck adversely affected the growth.’2 The explanation given was pck overexpression altered
the concentration of PEP-related metabolites which were previously unknown to be
involved in global regulation. There was no mention of DAHP yield or titer, presumably
because the growth was so poor that DAHP was not produced. This genetic modificauon
illustrates that metabolic engineering techniques sometimes do not work because of
unknown cellular processes.

Still another route to increase yield is to use xylose (a pentose) as a carbon source.
Xylose does not utilize the PTS to phosphorylate the sugar.’3 Instead, its metabolism
involves transport through the membrane, isomerization to xylulose, and ATP-dependent
phosphorylation of xylulose. Most strains of E. coli grow on xylose, but a mutation is
necessary for strain K-12 to grow on this carbohydrate." A theoretical analysis of the
pathways involved indicated that the yield should increase to 71 mol % since PEP is not
consumed in the phosphorylation of xylose. This experiment was conducted using a strain
that produced DAHP. The yield of DAHP from xylose was near the theoretical maximum
and the DAHP titer was higher when compared to glucose as the carbon source.”

Most recently, the impact of amplified expression of transaldolase (talB) has been
used to increase E4P in vivo availability. The talB locus was localized onto a multicopy
plasmid that also contained amF and aroB. The resulting plasmid was transformed into a
host that produced DHS. The construct resulted in only slightly higher DHS titer compared

to the control construct without talB.’6

 

The methodology of metabolic control analysis has emerged since the early 1970s.
It aims to characterize the sensitivity of metabolic responses to changes in enzymatic
activities or parameters without the use of full mathematical models.’7 Metabolic control

12

analysis takes advantage of the relationship of characteristic parameters at steady-state.
Valuable information can be gained by analysis with respect to identifying rate-limiting
enzymes by using control analysis. Knowledge of which enzymes are rate-limiting is
useful because it can identify enzymatic steps the exert the most control on the flux through
the pathway. This methodology has been used and further developed by Liao et. al. for
analysis the production of DAHP.’8 This analysis has shown, in the production of DAHP,
that after increasing E4P availability, PEP becomes limiting. Therefore, PEP availability
would be the next target for improving DAHP production, illustrating the usefulness of
control analysis.

Metabolic engineering as stated by Stephanopoulos et al is, “not another form of
classical manipulation of intermediary metabolism. It is, rather, the purposeful design of
metabolic networks”?9 This methodology has been applied by Stephanopoulos to the
production of the aspartate family of amino acids in Corynebacteriwn lactofennentum. The
analysis identifies which branch points in the biosynthetic pathway for the amino acid
production exert the most control. The objective of metabolic engineering is to identify
rate-limiting enzymes for amplification that would most efficiently increase carbon flux
through desired pathways. Flux control coefficients and elasticity coefficients are two tools
used to implement the metabolic engineering objective. Flux control coefficients indicate
how much an enzyme restricts the pathway as a whole. Elasticity coefficients indicate the
sensitivity of single reaction to changes in metabolite concentrations. Similarly to the
metabolic control analysis, metabolic engineering can provide the information necessary for
rational design of a microbial catalyst. Both methodologies offer means for analysis of
limiting steps in biosynthesis, and the results can be used to determine the next logical step
for improvement.

El . II I .
Significant plasmid loss during fermentation can result in low productivity and

indicate the need for a better plasmid maintenance system. Many studies have investigated

13

plasmid loss and its cause. There are many environmental factors that influence plasmid
maintenance, ranging from the level of dissolved oxygen to amino acid supplementation.‘50
Plasmid maintenance has been a concern in the industrial use of recombinant
microorganisms. Previous production of DHS in a fermentor indicated that after 48 h of
fermentation, 50-90% of the cells were lacking plasmids. In actuality, the loss of plasmids
by an individual host is a rare event.61 However, cells lacking plasmids no longer have the
metabolic burden caused by overproduction of the plasmid-encoded enzymes, and
therefore have a higher growth rate and soon out grow the plasmid bearing cells. This was
addressed by using a nutritional requirement to maintain the plasmid instead of antibiotic
resistance. There have been several examples of nutritional requirements used for plasmid

L62

maintenance. One such method was employed by Porter er a The ssb gene, whose
product is responsible for DNA replication, was deleted from the E. coli chromosome. The
locus for ssb was placed onto a plasmid. Therefore, a cell lacking ssh-encoding plasmid
would be unable to further replicate. Plasmid stability was achieved under nonselective
culture conditions.

Another method for plasmid maintenance relies on the postsegregational killing of
plasmid-free cells.‘53 The plasmid carries for the parB locus which encodes two genes,
Hok and Sok. The Hok gene (host killing) encodes for a small polypeptide that results in
rapid death of the host. The Sok gene (suppression of killing) product inhibits the
translation of the Hok mRNA. The Hok mRNA is much more stable than the Sok mRNA.
In plasmid-carrying cells Sok RNA prevents the synthesis of the Hok ' protein and the cell
remains viable. In a plasmid-free cell, the Sok RNA quickly decays thereby allowing
translation of the Hok mRNA accessible for translation. The Hok protein is synthesized,
and cell death occurs. Other methods rely on a similar strategy, wherein the plasmid free
cells no longer have the ability to prevent cell death due to lack of a protein. The use of the
parB locus may not be the best choice for plasmid maintenance since it involves more

cloning and will increase the metabolic burden on the host cell.

14

The use of antibiotic resistance for plasmid maintenance in itself is problematic for
many reasons. Antibiotics such as ampicillin are degraded by B-lactamase enzymes which
are exported from the cell.64 This enables cells lacking B—lactamase to grow because the
enzyme conferring resistance has been transported into the culture broth. It is likely that
during 48 h fermentations, all of the ampicillin initially present in the growth medium has
been hydrolyzed.65 Another problem with the use of some antibiotics is their influence on
cellular metabolism. Chloramphenicol is one such example.“ The mechanism of
chloramphenicol deactivation entails acetylation of chloramphenicol using acetyl-CoA as the
donor. Since the acetyl donor group is from acetyl-CoA this reaction may have the
unwanted effect of depleting the acetyl-CDA pool. Lastly, it is undesirable to have an
antibiotic resistant organism used for industrial processes due to possible release into the
environment.

Eermentor

Significant work has been performed towards the optimization of E. coli
fermentations to produce phenylalanine and tryptophan.‘57 This previous work served as a
useful starting point for optimizing DHS production.

At the laboratory scale, fermentors provide a significant advantage over shake
flasks. The fermentor can control many variables on-line, such as temperature, pH,
dissolved oxygen, and substrate addition. Through computer usage, it is also possible to
have a variety of complex control algorithms calculated from on-line data. Another
advantage is the greater aeration capacity provided by the air sparger and the impeller in the
fermentor, which is necessary for aerobic growth of high density E. coli cultures"8 Much
higher cell densities and growth rates are typically achieved due to the increased aeration in
a stirred-tank fermentor than in a shake flask, which results in a higher concentration of
biocatalyst.

There are also many operating advantages in utilizing a fermentor. Dissolved

oxygen can be controlled by the impeller speed and or air flow rate in the Braun fermentor.

15

A better degree of mixing is achieved by the impeller and baffles. In high density culture,
foaming is a significant problem, but it can be controlled by a foam sensor and computer-
controlled additions of antifoam. Alteration of medium temperature can be conveniently
accomplished in a fermentor, which allows for temperature-dependent expression systems
to be used. Samples can be removed under sterile conditions using the harvest pipe. On-
line monitoring and electronic storage allows for data recording and retrieval of essential
parameters.
E I Q] . 'v

This research was a collaborative effort between Drs. John W. Frost and R. Mark
Worden to develop a microbial catalysts and a fermentation process for industrial
production of DHS. The first objective was to develop a feedback-insensitive AroF. The
second chapter of this thesis will discuss the use of a novel technique to generate and select
for a feedback-insensitive aroF mutant in a diffusion gradient chamber. The feedback-
insensitive aroF obtained was then used in constructing strains relevant to industrial
synthesis of hydroaromatic and aromatic compounds. Also investigated was the effect of
removing part of the ryrR boxes upstream of the feedback-insensitive aroF in an effort to
remove transcription regulation by repression.

The third chapter of this thesis and the second objective was the optimization of
DHS production in a fermentor through both genetic and reaction engineering. The
parameters optimized were DHS titer and yield. Optimizatr' ' 'on studies investigated the
effects of several variables, including fermentor medium composition, starter culture
medium, dissolved oxygen setpoint, pH setpoint, and glucose feeding strategies.
Genetically modified constructs were evaluated for DHS production with the use of the
feedback-insensitive aroF. The effect of DAHP synthase activity on DHS production was
studied utilizing the tac promoter and repressor lacl' to manipulate the DAHP synthase
activity by transcription control. The alterations of aroF orientation with respect to a

promoter and the number of aroF copies on the plasmid were also investigated.

M-FUJN

10

11

12

13

16

Vieth, Wolf R. Biapracess Engineering; John Wiley & Sons: New York, 1994, p
52, 147.

Draths, K. M.; Frost J. W. J. Am. Chem. Soc. 1995, 117, p 2395.
Ikeda, M.; Ozaki A.; Katsumata, R. App. Microbial. Biotechnol. 1993, 39, p 318.
Chic, Y. J .; Tribe, D. E. Biatechnal. Lett. 1982, 4, p 223.

Ito, H.; Sato, K.; Matsui, K., Enei, H.; Hirose, Y. App. Micravial. Biotechnol.
1990, 33, p 190.

Swartz, J. R. In Escherichia coli and Salmonella typhimurium, Neidhart F. C.,
Ed.; American Society for Microbiology: Washington , DC, 1995; Vol. 2, p 1696

Pittard, J; Wallace B. J. J. Bacterial. 1966, 4, p 1494.
Pittard, J; Wallace B. J. J. Bacteriol. 1966, 4, p 1494.

(a) Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C. ,
Ed.; American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 368 (b)
Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983,
p 301. (c) Cambell, M.; Sainsbury, M.; Scarle, P.; Synthesis 1993, p 179.

(a) Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C. ,
Ed.; American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 368.
(b) Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Hegmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983,
p 1. ‘

(a) Dewick, P. M.; Natural Product Report 1992, p 149. (b) Herrmann, K. M.; In
Amino Acids: Biosynthesis and Genetic Regulation; Herrmann, K. M.; Somerville,
R. L. Eds.; Addison-Wesley: Reading, MA, 1983, p 305.

Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Harorznann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983,
p .

Pittard, J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 371.

14

15

16

17

18

19

20

21

22
23
24

25
26

27

17

Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983,
p 304.

Pittard, J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 371 . (b)
Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983,
p 307.

Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 372. (b)
Frost J. W., Binder J. L.; Kadonaga, J.T.; Knowles, R. J. Biochemistry, 1984,
23, p 4470.

Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 372. (b)
Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983,
p 313.

Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 371.

Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 373.

Pittard, J; Wallace B. J. J. Bacterial. 1966, 4, p 1494.

Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983,
p 304.

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

(a) Draths, K. M.; Frost J. W. J. Am. Chem. Soc. 1994, 114, p 9725. (b) Draths
K. M.; Frost, J. W. In ACS Synpasium Series 577: Benign by Design; Anastas, P.
T. Farris, C.A., Eds.; American Chemical Society: Washington, DC, 1994; p
32.(c) Omston, L. N .; Neidle, E. L. In The biology of Acinetabacter; Towner, K.
J., Bergogne-Berezin, E., Fewson, C. A., Eds.; plenum: New York, 1991; p
201. (d) Neidle, E. L.; Ornston, L. N. J. Bacterial. 1986, 168, p 815.

Dell, K. A.; PhD. Thesis, Purdue University, 1989

Franenkel, D. G. In Escherichia coli and Salmonella typhimurium, Neidhart F. C. ,
Ed.; American Society for Microbiology: Washington , DC, 1987; Vol. 1, p 142

(a) Vieth, Wolf R. Biapracess Engineering; John Wiley & Sons: New York, 1994,

p 166. (b) Postma, P. W. In Escherichia coli and Salmonella typhimurium,

Neidhart l1:2.7C., Ed.; American Society for Microbiology: Washington, DC, 1987;
o . p .

28
29

30
31

32

33

34

35

36

37

38

39
40

41

42

18

Frost, J. W.; Knowles, J. R. Biochemistry 1984, 23, p 4465.

Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 371.

Tribe, D. E.; Camakaris, H.; Pittard, J. J. Bacterial. 1976, 127, p 1085.

Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 371.

Shultz, J.; Hermodson, M. A.; Garner, C. C.; Herrmann, K. M. J. Biol. Chem.
1984. 259. p 9655.

Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 371.

(a) Konstantinov, K.; Nishio, N.; Seki, T.; Yoshida, T. J. Ferment. Bioeng.
1991, 71, p 350. (b) Weaver, L. M.; Herrmann K. M.; J. Bacterial. 1990, 172, p
6581. (c) Sugimoto, S.; Yabuta, M.; Kato, N.; Seki, T.; Yoshiomi, T.; Taguchi,
H. J. Biatech. 1987, 5, p 237.

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,
p3956

Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987 ; Vol. 1, p 371.

(a) Gottesman, 8.; Methods in Enzymalagy 1990, 185, p 119. (b) Pittard, A. J. In
Escherichia coli and Salmonella typhimurium, Neidhart F. C.,Ed.; American
Society for Microbiology: Washington, DC, 1987; Vol. 1, p 371. (c),Tribe, D. E.;
Pittard, J.; Appl. Environ. Microbial. 1979, 38, p 181.

(a) Konstantinov, K.; Nishio, N.; Seki, T.; Yoshida, T. J. Ferment. Bioeng.
1991, 71 , p 350. (b) Weaver, L. M.; Herrmann K. M.; J. Bacterial. 1990, 172, p
6581. (c) Sugimoto, S.; Yabuta, M.; Kate, N.; Seki, T.; Yoshiomi, T.; Taguchi,
H. J. Biatech. 1987, 5, p 237.

Stephanopoulos, G.; Vallino, J. J. Science 1991, 252, p 1675.

(a) Lewin, B.; Genes V. 1994, p 414. (b) Herrmann, K. M.; In Amino Acids:
Biosynthesis and Genetic Regulation; Herrmann, K. M.; Somerville, R. L. Eds.;
Addison-Wesley: Reading, MA, 1983, p 309.

(a) Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983,
p 310. (b) Ogino T.; Garner, S. J.; Sverlow, G. G.; Turpen, T. H.; Grill, L. K.
Biotechnology 1982, 79, p 5828.

(a) Pittard, J.; Davidson, B. E. Mal. Microbial. 1991, 5, p 1585. (b) Herrmann,
K.; Garner, C.;J. Biol. Chem. 1985, 260, p 3820. (C) Herrman, K.; Schultz J.;
Hermodson, M.; Garner C.; J. Biol. Chem. 1984, 259, p 9655.

43

45

46
47
48
49

50

51

52

53

54

55
56
57
58

59

19

Cui, J.; Sommerville, R. L. J. Bacterial. 1993, 175, p 303.

Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983;

p 307.

Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983;
p 307.

Frost, J. W.; Knowles, J. R. Biochemistry 1984, 23, p 4465.
Snell K. D.; Draths, K. M.; Frost J. Am. Chem. Soc. 1996, 118, p 5605.
Snell K. D.; Draths, K. M.; Frost J. Am. Chem. Soc. 1996, 118, p 5605.

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,
p 3956

(a) Liao, J .; Patnaik R.; Spitzer, R. G. Biotechnol. Bioeng. 1995, 46, p 361. (b)
Pratnaik, R.; Liao, J. Appl. Environ. Microbial. 1994, 60, p 3903.

(a) Liao, J.; Patnaik R.; Spitzer, R. G. Biotechnol. Bioeng. 1995, 46, p 361. (b)
Chao, Y.; Pratnail, R.; Roof, W.; Young, R.; Liao J. J. Bacterial. 1994, 175, p
9639.

Liao, J. C.; Hou, S. Y.; Chao, Y. P. Biatechnal. and Bioeng. 1996, 52, p 129.
Liao, J.; Patnaik R.; Spitzer, R. G. Biotechnol. Bioeng. 1995,46, p 361. (b) Lin,
E. C. C. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology. Washington, DC, 1987; Vol. 1, p 255.

Lin, E. C. C. In Escherichia coli and Salmonella typhimurium, Neidhart F. C.,
Ed.; American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 255.

Liao, J.; Patnaik R.; Spitzer, R. G. Biotechnol. Bioeng. 1995, 46, p 361.
Farabaugh, M. Unpublished data.

Liao, J. C.; and Delgado, J. Biatechnal. Prag. 1993, 9, p 221.

(a) Laio, J. C.; and Delgado, J. Biatechnal. Prag. 1993, 9, p 221. (b) Liao J. C.;
Delgado, J. Biochem. J. 1992, 285, p 965. (c) Liao, J. C.; and Delgado, J.
Biatechnal. Prag. 1991, 7, p 15.

33mpsgg, T. W,; Colon, G. E.; Stephanopoulos, G. Biochem. Soc. Trans. 1995,
p l.

(a) Shu, J.; Shuler, M. L.; Biatechnaol. and Bioeng. 1992, 40, p 1197. (b)
Haung, J.; Dhulster, P., Barbotin, J.; Thomas, D. J. Chem. Tech. Biotechnol.
1989, 45, p 259. (c) Marin-Iniesta, F.; Nasri, M.; Dhulster, P; Barbotin, J.;
Thomas, D. App. Microbial. Biotechnol. 1988, 28, p 455.

61

62

63

65

66

67

68

20

(a) Margaritis, A.; Bassi, A. S. Improving Product Expessian and Recovery 1990,
p 317. (b) Ensley, B. D. CRC Critical Reviews in Biotechnology 1986, 4, p 283.

(a) Porter, R. D.; Black, S.; Pannuri, S.; Carlson, A. Bio/1‘ echnal. 1990, 8, p 47.
(b) Nakayama, K. Kelly, S. M.; Curtis III, R. Bia/I'echnal. 1988, 6, p 693 (c)
Mortensen, U.; Nilsson, J. Prac. 4th European Cong. Biotechnol, Neijssel, O.
M.; van der Meer, R. R.; Luybem, A. M.Eds.; 1987, 3, p 303.

(a) Kulakauskas, S.; Lubys, A.; Whrlich, D. J. Bacterial. 1995, 177, p 3451. (b)
Margaritis, A.; Bassi, A. S. Improving Product Expession and Recovery 1990, p
317. (c) Porter, R. D.; Black, S.; Pannuri, S.; Carlson, A. Biafl’echnal. 1990, 8,
p 47. (d) Gerdes, K. Bia/I‘echnal. 1988, 6, p 1402. (e) Mortensen, U.; Nilsson,
J. Prac. 4th European Cong. Biotechnol, Neijssel, O. M.; van der Meer, R. R.;
Luybem, A. M.Eds.; 1987, 3, p 303(1) Meacock, R; Cohen, S. Cell, 1980, 20,
p 529.

Ward, J. B. In Antibiotic Inhibitors of Bacterial Cell Wall Biosynthesis, Tipper, D.
J. Ed.; Pergamon Press, New York, 1987; p 23.

(a) Swartz J. R., In Escherichia coli and Salmonella typhimurium, Neidhart F. C.,
Ed.; American Society for Microbiology: Washington, DC, 1995; Vol. 2, p 1693.
(b) Patnaik, R.; Liao, J. C.; Appl. Environ. Microbial. 1995, 60, p 3903. (c)
Jung, G.; Denefle, P.; Becquart, J .; Mayaux, J. F. Ann. Inst. Pasteur Microbial.
1988, 139, p 129.

(a) Swartz J. R., In Escherichia coli and Salmonella typhimurium, Neidhart F. C.,
Ed.; American Society for Microbiology: Washington, DC, 1995; Vol. 2, p 1693.

(a) Konstantinov K.; Nishio, N.; Seki, T.; Yoshida, T. J. Ferm. Bioeng. 1991,
71 , p 350. (b) Terasawa, M.; Fukushima, M.; Yukawa Prac. Biachem Inc. 1990,
25, p 172. (c) Sugimoto, S.; Yabuta, M.; Karo, N.; Seki, T.; Yoshida, T.;
Taguchi, H. J. Biatech. 1987, 5, p 237. (d)Tribe, D.; Pittard J. Appl. Env.
Microbiol. 1979, 38, p 181.

Swartz, J. R. In Escherichia coli and Salmonella typhimurium, Neidhart F. C.,
Ed.; American Society for Microbiology: Washington, DC, 1995; Vol. 2, p 1693.

CHAPTER 2

TYROSINE FEEDBACK-IN SEN SITIVE DAHP SYNTHASE

Background

Biocatalytic synthesis of various metabolites often requires alteration of a microbe’s
natural regulatory mechanism so that increased carbon can be directed into selected
biosynthetic pathways.‘ Alterations can be performed genetically by mutation of genes,
changing the promoter, increasing the number of gene copies in the cell, or by overcoming
transcriptional repression. Regulatory mechanisms governing carbon flow in the common
pathway include both feedback inhibition of enzymes and repressor-mediated
transcriptional repression. 2

In Escherichia coli the first committed step in the common aromatic biosynthetic
pathway is the condensation between erythrose 4—phosphate _ (E4P) and
phosphoenolpyruvate (PEP) to form 3-deoxy-D-arabina-heptulosonate 7-phosphate
(DAHP). This reaction is catalyzed by three DAHP synthase isozymes encoded by araF,
aroG, and wall, which are feedback inhibited by L-tyrosine, L-phenylalanine, and L-
tr'yptophan, respectively.3 At the repressor-mediated transcriptional level, araF and aroG
are repressed by the TyrR protein, and mail is repressed by the TrpR protein.‘ Feedback
inhibition has been demonstrated to be the dominant regulatory mechanism in controlling
carbon flow into the common pathway of aromatic amino acid biosynthesis in E. cali.’

Repressor—mediated transcriptional repression is a second mode of control for araF
expression at the gene level. In the case of araF, the transcriptional regulation is a result of

the araF tyrA transcriptional unit. The structural gene for araF is preceded by three

21

22

separate conserved sequences called tyrR boxes.6 A repressor protein, TyrR, when
activated by binding to the effector tyrosine, binds to the tyrR box such that transcription
cannot occur. In the presence of tyrosine, transcriptional repression can lead to a 20-fold
reduction in DAHP synthase activity relative to the activity in the absence of tyrosine.7

Since the key enzyme AroF is regulated by both feedback inhibition and
transcriptional repression, increasing the in vivo activity of this isozyme of DAHP synthase
was essential. The host used for producing dehydroshikimate (DHS) is an aromatic amino
acid auxotroph that requires the addition of tyrosine, phenylalanine, and tryptophan, as
well as aromatic vitamins for growth in minimal medium. The aromatic vitamins p-
aminobenzoic acid, 2,3-dihydroxybenzoic acid, and p-hydroxybenzoic acid are
biosynthetic precursors to folic acid, enterochelin, and ubiquinone, respectively. A
feedback-insensitive AroF was thought to be needed to avoid feedback inhibition caused by
in vivo, steady-state concentration of tyrosine. Another motivation for creating the
feedback-insensitive AroF is the intracellular tyrosine concentration could be higher than
the extracellular concentration. The intracellular tyrosine concentration of ABZ834 may be
high compared to the wild-type E. coli intracellular tyrosine concentration (since wild-type
E. coli would be grown in medium that is not supplemented with tyrosine) if the rate of
active transport of tyrosine into the cell is greater than the rate of tyrosine consumption by
incorporation into proteins. In addition, the pseudo steady state concentration of tyrosine
may be high even if the transport and initialization rates of tyrosine are the same.

The tyrosine-sensitive form of DAHP synthase is an iron-containing dimer with
subunit molecular weight of about 40,000.8 Under minimal growth conditions that are
supplemented with all amino acids except the aromatic amino acids, AroF comprises the
major isozyme. The reaction mechanism is ordered and sequential with PEP being the first
substrate to bind. The KIn value for PEP has been reported to be 5.8 11M for the purified
enzyme.9 It is assumed that PEP-enzyme complex is the native form of the enzyme since

the intracellular concentration of PEP never falls below 88 11M, and enzyme activity is lost

23

and unrecoverable during purification if the enzyme buffer solution does not contain PEP.
The enzyme is 50% inhibited by 20 M tyrosine.lo The enzyme is also inhibited by the
nonmetabolized analog m-fluorotyrosine (m-FT).” The nature of the feedback inhibition is
noncompetitive with respect to E4P and competitive with respect to PEP.

To obtain a feedback-insensitive AroF, a novel method was deve10ped where
ultaviolet light mutagenesis was followed by phenotype selection utilizing chemotaxis in the
diffusion gradient chamber (DGC). Chemotaxis is the ability of an organism to move in
response to a gradient of a chemical species.” A chemical species that attracts an organism
is referred to as a chemoattractant. The separation and selection of the desired mutant was
aided by the exploitation of chemotaxis.

W

In this chapter, several molecular biology techniques are described, and the
following terminology is defined to help clarify these techniques. An open reading frame
(OPR) contains a series of DNA base triplets coding for amino acids without any
termination codons, and the sequence is potentially translatable into protein. The OPR does
not contain the genetic sequence for a promoter. Polymerase chain reaction (PCR) was
used to isolate the feedback-insensitive araF. It is a technique in which cycles of
denaturation, annealing with primer, and extension with DNA polymerase, are used to
amplify the number of copies of a target DNA sequence by >106 timesl3 . The ribosome-
binding site is the DNA sequence at which initiation of transcription occurs. It is a short
sequence of bases that precedes the actual coding region.

12 l E l S . E I I .

The E. coli strain mutagenized contained only the AroF DAHP synthase isozyme.
The activity of the other two DAHP synthase isozymes were not present. It was developed
from host, AB3248, that lacks all three DAHP synthase isozyme activities,” and is also
auxotrophic for the amino acids histidine (H), isoleucine (I), proline (P), arginine (R), and
valine (V). Minimal salts medium must be supplemented with these amino acids in order

24

for AB3248 to grow. To obtain the correct phenotype, a copy of araF was added either by
insertion of a multicopy plasmid or insertion into the genome. Insertion into the genome
was the method chosen for several reasons. This suategy eliminates the need to add an
antibiotic for selective pressure to maintain a plasmid. One problem with using araF on a
multicopy plasmid stems from overexpression of the AroF protein. With use of plasmid-
encoded araF, this enzyme may constitute 10-15% of the total soluble protein of the
organism.15 The selection might not work well under these conditions since a significant
amount of DAHP synthase activity may remain in the presence of m-F'I‘. The genomic
insertion would provide only one copy of araF. This would reduce the possibility that the
enzyme concentration may become too high relative to the inhibitor concentration. A
further advantage of the genomic versus plasmid localization of araF arises from the
presence of only one copy of the mutated gene requiring isolation. Use of a plasmid-
bearing araF would complicate the isolation of the mutant gene since multiple copies of
mutated and unmutated genes may be present. The insertion of araF into the genome of
AB3248 was conducted by Dr. Karen Draths.

Genomic insertion of genetic sequences into E. coli is possible by several methods.
Successful insertions have been performed using transposons," circular plasmids,'7 and
linear DNA fragments18 as a introduction vehicle. The disadvantage of many of these
methods is the required use of special strains with specific mutations in order to achieve
insertion. Site specific insertion independent of the genotype of the recipient strain was
desired for our purpose.19

Usually plasmids are maintained as extrachromosomal, circularized DNA in E. coli,
but occasionally recombination events occur such that plasmid DNA is integrated into the
host cell’s genome. Exploitation of this rare event provides a method for simple site-
specific insertion of a gene flanked by sequences homologous to the desired insertion
location in the genome. Plasmid-bearing cells are usually differentiated from cells without

plasmid by selection due to antibiotic resistance conferred by the plasmid. The difficulty in

25

differentiation between the cells with integrated plasmid DNA and replicating plasmids is
that both are resistant to the antibiotic. The use of a non-replicating plasmid allows for the
exclusive selection of integrated plasmid DNA since cells will possess antibiotic resistance
only if the plasmid resides in the genome.

A plasmid in which the replication mechanism can be turned on and off is desirable
for genomic insertions. Under conditions where the replication of the plasmid is inactive,
genomic insertions can be accomplished, whereas normal cloning and preparation of the
plasmid can be performed with conditions conductive to active replication. Plasmids
possessing a temperature-sensitive replicon [rep (ts)] are capable of being manipulated in
this fashion. Plasmid pKAD76A20 (Figure 3) contains such a temperature sensitive
pSC 101 replicon, a chloramphenicol acetyltransferase (cm) gene conferring resistance to
chloramphenicol, and a copy of the serA locus, which has a suitable cloning site (EcaRI).
The serA gene on the plasmid was necessary for homologous recombination into the
genomic serA. The serA gene converts 3-phosphoglycerate into 3-
phosphohydroxypyruvate for serine biosynthesis. The plasmid is able to replicate normally
when the host cell is grown at 30 °C, but it is unable to replicate when the host cell is
cultured at 44 °C. Thus genetic manipulation of the plasmids is carried out at 30 °C while
cells that had the integration of the plasmid into the genome can be selected at 44 °C.

 

pKAD76A
7.4 kb EcoRI
serA I
O 4—
rep (ts) Cm

Figure 3 - Plasmid Map of pKAD76A.

26

Synthetic Cassette

Ec RI BamI-II EcoRI
o araF Kan

 

 

Figure 4 - Synthetic Cassette.

A synthetic cassette was developed to aid in the selection of the integration of the
plasmid. An araF gene fragment with EcaRl and BamHI ends was produced from the
polymerase chain reaction (PCR), using primers JWF-22 and JWF-83 (Table l), which
have EcoRI and BamI-II ends, respectively. The araF fragment was ligated into a linearized
plasmid (pKAD62A, a kanamycin resistance-bearing plasmid) at a cloning site next to the
kanamycin (kan) resistance gene. This product was then digested with EcoRI to produce a
cassette of the araF and kanamycin (kan) resistance genes (Figure 4). The fragment was
then cloned into pSU18 (ampicillin and chloramphenicol resistance-bearing plasmid)
creating the plasmid pKD10.156A (Figure 5). The correct ligation was confirmed by
plasmid DNA digestion and plating onto minimal medium (M63) plates with kanamycin.
The cells would not grow on the M63/Kan plates unless both the kan resistance and a
functional araF was contained on the plasmid. Next, the planned insertion of the cassette
into the genomic copy of serA in strain AB3248 necessitated construction of the plasmid
containing the synthetic cassette flanked by the serA locus.

The synthetic cassette was isolated from pKDlO.156A as an EcaRI fragment and
was cloned into the EcoRI site of serA in pKAD76A. The resulting 9.8 kb plasmid,

27
Table l - PCR Primers.

Primer Name Sequence
Eco RI
JWF-l9 5' WAAGCCACGCGAGCCGT 3'

EcaR I
JWF-22 5' WMAGGGAGTGTAAATTTAC 3'

JWF-79 5' GCI'I'I'TCCATTGAGCCTGCA 3'

JWF-80 5' AACGATCCCCATATGGATGG 3'
JWF-8 1 5' GCAGTCTGGCAACAGCAATT 3'
JWF-82 5' AAGAT'TATCGCCGTCAGCCT 3'

Barn HI
JWF-83 5' GCGGATCCTCTTAAGCCACGCQAQQCGT 3'

Eco RI
JWF—94 5' GQAAILCTGTACGAAATATGGAT'I‘GAA 3'

Eco RI
JWF-97 5' GQAAILCTTAAGCCACGCCCGT 3'

Eco RI
JWF-103 5' WATGCAAAAAGACGCGCT GA 3'

 

 

 

 

pKDlO. 156A
4.7 kb
EcoRI BamHI EcoRl
—>
Cm

Figure 5 - Plasmid Map of pKDloJStiA.

28

pKD10.186A (Figure 6), contained the synthetic cassette flanked by portions of the serA
gene in a host vector containing a temperature-sensitive replicon. The plasmid was
transformed into the host JC158, a serine auxotroph lacking serA activity. After
transformation and plating, the colonies were then replicate plated onto M63/Cm plates with
and without serine supplementation. The desired insertion inactivated the serA locus on the

host cell’s genome resulting in an inability of the cells to grow on minimal plates lacking

 

 

 

serine supplementation.
pKDlO. 186A
9. 8 kb
EcoRI BamHI EcoRI
serA I araF l kan I serA
O 4——
reP (t8) Cm

Figure 6 - Plasmid Map of pKD10.186A.

The synthetic cassette was inserted into the genome of AB3248 using homologous
recombination into the genomic locus of serA. Competent AB3248 cells were transformed
with pKD10.186A, and integration of the plasmid into the genome was selected for at 44
°C on LB plates containing chloramphenicol and kanamycin. The importance of the
chloramphenicol resistance marker is two-fold. First, it confirms plasmid integration into
the genome, and later chloramphenicol sensitivity was used to confu'm that the plasmid had
been excised from the genome. Only those colonies containing integrated plasmid DNA are
able to grow in the presence of the antibiotics since the temperature-sensitive replicon in
pKD10.186A does not allow replication at 44 °C. Eleven cointegrates were isolated in this

manner after a series of transformations. Removal of the plasmid from the genome was

29

performed by growing cointegrates at 30 °C in LB medium without antibiotics. Two more
cycles of growth were carried out at 30 °C by diluting cultures (l:20,000) into fresh LB
medium without antibiotics. Growth of the cointegrate strain at a temperature permissive to
plasmid replication (30 °C) creates an unstable environment for the integrated plasmid and a
second spontaneous recombinational event occurs allowing the excision of the plasmid
from the genome. The removal occurs such that the plasmid is either excised with its
original synthetic cassette insert or with an intact serA sequence on the plasmid indicating
that the cassette remained in the genome. Subsequent growth in liquid culture at 44 °C
resulted in the loss of the excised plasmids from the progeny.

Colonies were finally selected for kanamycin resistance and chloramphenicol
sensitivity at 44 °C to identify cells that retained the kan resistance and excised the cm
plasmid marker from the genome. Fifteen such colonies originating from the same
cointegrate exhibited the correct markers from which two were identified and characterized
further. These two strains were designated AB2.23 and AB2.24. Plasmid DNA
preparations confirmed that no plasmid DNA remained in the strains.

The strains were replicate plated on M63 plates with and without serine to verify
that genomic insertion of the synthetic cassette disrupted the serA gene. The colonies were
unable to grow without supplemented serine, signifying that site-specific insertion into
serA had occurred.

W515

The conditions for mutagenesis were adapted from the method by Miller.21 The
desired survival rate was 0.1 to 0.01 % to obtain efficient mutagenesis. Conditions giving
the desired survival rate were determined by manipulating the distance of the culture from
the UV source and the UV radiation exposure time. This was accomplished by growing
A8224, followed by dilution and plating onto M63 plates lacking aromatic amino acid
supplementation. The lack of aromatic amino acid supplementation was selected for cells

that have a functional araF. The plates were then exposed to the UV radiation (Sylvania

30

germicidal 8W lamp) for a duration of 0 (control), 5, 10, 15, 30, 60, 80, or 100 seconds
immediately after plating. Two UV source distances were investigated: 18 and 24 inches.
The optimal time, at a distance of 24 inches, was 10 seconds which resulted in a 0.6%

survival rate (Table 2).

Table 2 - UV Radiation Survival.

UV Exposure Time (8) Number of Colonies Survial Rate (%)

 

0 169 100.0

5 75 44.4
10 l 0.6
15 3 1.8
30 2 1.2
60 l 0.6
80 0 0.0
100 0 0.0

 

In the DGC, the cells grew and migrated (by chemotaxis) through the agarose gel
rather than only on the surface. As a confirmation that UV radiation would penetrate
through the agarose gel, the UV absorbance of the gel was determined. The gel did not
absorb significantly in the wavelength range of 200 to 800 nm.

 

Experiments using replicate plating to select for feedback-insensitive mutants after
chemical mutagenesis resulted in false positives (growth in the presence of m-FT without a

feedback-insensitive AroF). The selection was for growth on minimal plates in the

31

presence of m—FT. The false positives were believed to have resulted from enough
overexpression of the feedback-sensitive AroF to allow for growth even in the presence of
moderate levels of m—FT. The false positives greatly increased effort required to find the
desired mutant.

One advantage of the DGC method is it allows for simultaneous selection and
mutagenesis in the chamber, resulting in a significant reduction in effort to isolate the
desired mutant. The DGC has other advantages over plate selection. The replicate plating
assay compares growth of a single colony to growth of a single colony on a plate
supplemented with m-FI‘. It is difficult to know in advance the optimal concentration of m-
FT to use for selection using the plating technique. The growth of the sensitive AroF
strains on the m-FT plates raised questions about the strength of the selection pressure
provided by m-FT. The DGC allows for a range of m-FT tyrosine concentration to be
applied simultaneously in the DGC. A second possible reason for the false positive
phenomenon on plates is the fact that some cells uptake the m-FI‘ leaving a zone lacking 111-
FT thereby allowing the m-FT-sensitive cells to grow. This would be avoided in the DGC
since the m-FT gradient is continuously replenished throughout the experiment.

Another advantage is that differentiation between phenotypically (or possibly)
genetically different mutants is possible in the DGC, based on their different sensitivities to
m-FT. One mutation may allow for partial insensitivity to the inhibitor. Presumably the
partially insensitive mutant would only be able to grow up to a certain distance towards the
source of the m-FI‘. Beyond that point, the m-FI‘ concentration would be to high for
growth. A more desirable mutant that was completely insensitive to inhibition, however,
could grow into the region of highest m—Fl‘. The DGC may be able to select among many
different mutations for the one with the highest m-FT insensitivity.

The DGC (Figure 7) is part of the diffusion gradient system commercially produced
by Koh Development (Ann Arbor, Michigan). The chamber consists of an arena and the

32

 

 

 

 

 

 

 

 

 

 

 

TOP VIEW
l l I Reservoir
I L____-___-______E I j
“1 i" — -
g :
: i l—
-__"_:L.. mitt-=1
I. I
i <—— 5 cm —-> E :1
1 I
[__E :h___ -
. :r_________ -___-__.: ' / Gasket
I I I I I Semi-permeable
\ membrane
Inlet Outlet
SIDE VIEW .
L1d (removable)

 

l I / Gas Port

0 .=1

 

 

l
a

 

 

 

 

Figure 7 - Diagram of a Diffusion Gradient Chamber.

33

recesses that form the cavity of the four reservoirs. There is an opening between each
recess and the arena. The reservoirs contain stainless steel inlet and outlet ports (1.0 and
2.5 mm diameters, respectively). The outlet port is higher than the inlet port to allow the
reservoir to fill with liquid and release gas bubbles. The purpose of the larger outlet port
diameter is to prevent back pressure of the outflowing liquid. The membrane allows for
diffusion of small molecules from the reservoir into the gel, which is contained in the
arena, and prevents organisms from moving from the gel into the reservoirs. The arena
and the reservoirs are both machined from polycarbonate (PC). For reservoirs in use
during the experiments, a 0.05 11M pore size PC filter membrane (Koh Development) and a
gasket (on both sides of the membrane) were placed between each opening into the arena
and the reservoir. The reservoirs not in use were sealed off from the arena with non-
permeable silastic sheeting (Dow Corning Inc., Midland, Michigan). The reservoirs are
secured to the DGC with thumb screws. The volume of each reservoir is 3 mL. The total
volume of the arena, excluding the reservoirs, is 50 mL (5 cm x 5 cm x 2 cm). The lid and
bottom plates are clear PC and are fastened by thumb screws. The entire system is
sterilized by autoclaving prior to use.

The DGC was placed on a transilluminator box (TB) which can accommodate up to
three DGCs. The TB contained two 30 cm fluorescent light fixtures (single 8W, cool white
bulb). The inner walls of the TB were white and a piece of black felt was placed on the
bottom and the sides of the TB to provide contrasts for the pictures. This design provided
cool, diffuse, even illumination from beneath the DGC, which was essential for
visualization of microbial growth patterns in the chamber. The light was turned off when
not in use as to prevent heating of the DGC. The TB also had a bracket for mounting a
camera above the DGC to record growth patterns photographically. The camera used was a
PULNiX TM-7CN CCD-camera (Sunnyvale, CA). The pictures were taken using the
programs Photofinish (Zsoft, Marietta, GA) and AutoCap. AutoCap written in the Warden

34

lab is the program responsible for recording the images of the DGC. Photofinish is a
commercially available program to view and manipulate pictures.

To generate the gradient, solutes contained in Erlenmeyer flasks were continuously
pumped through the reservoirs of the DGC. The flow rate of 2.5 mL h" was controlled
with a dual channel peristaltic pump (LAB Bromma Microplex). An effluent chamber
mounted on a stand next to the TB served three functions. Its height relative to the height
of the DGC reservoir outlets regulated the back pressure in the chamber reservoirs. This
was critical since excessive back pressure causes flooding of the gel due to bulk flow of
liquid through the membrane. Insufficient back pressure causes shrinkage of the gel due to
siphoning of the liquid. The effluent chamber also consolidated all of the reservoir
outflows into one large waste flask as well as serving as a sterile break in the liquid flow.

DECQ 'E .

Chemotactic experiments in the DGC were carried out similarly to the experiments
described in Emerson et al.22 Because strain AB3248 had undergone several rounds of
mutagenesis, we needed to prove that neither the Chemotactic response nor motility were
destroyed. A minimal salt medium (M63) was used for all DGC experiments. This
medium was supplemented with 5 mM glycerol, and 40 mg L'1 of the amino acids histidine
(H), isoleucine (I), proline (P), arginine (R), valine (V), and serine (S). Serine was
supplemented in order to complement the serine auxotrophy that was created through the
homologous recombination. The glycerol served as the carbon source for growth, but it is
not a chemoattractant for E. coli.23 The arena medium, which was stabilized with 0.15%
agarose, was supplemented as described above and initially did not contain glucose. The
low percentage of agarose provided enough strength for a stable gel matrix, but did not
prevent movement of the cells through the gel. The sink reservoir (800 mL) contained
supplemented M63 medium only. The source reservoir (800 mL) contained the
supplemented M63 medium and was additionally supplemented with 5 mM glucose as a

chemoattractant." The sink and source reservoirs were on opposite sides of the DGC.

35

This arrangement created a l-dimensional gradient that spanned 0 to 5 mM glucose from
sink to source.

The first step was inoculation of the starter culture (5 mL of LB medium) with a
single colony of AB2.24. The culture was then grown overnight at 37 °C with shaking in
a water bath. Next, 1 mL of the starter culture was added to 100 mL of supplemented M63
medium containing 10 mM glycerol in a 250 mL Erlenmeyer flask. This culture was
grown (37 °C and 250 rpm) to stationary phase (24 h, 0D,“, 3). Then four 1 mL aliquots
were concentrated (4x) by microcentrifugation, combined and centrifuged again to give a
final concentration of 16x. The center point of the DGC was inoculated with 15 11]. of the
16x concentrated culture using a micropipette to disperse the cells evenly throughout the
depth of the agarose as the pipette was withdrawn from the gel. The DGC and TB were set
up in an approximately 30 °C warm room.

The flow of liquid through the reservoirs was started 6 h before inoculation to
initiate the glucose gradient. Photographs were taken every 1 h for the 72 h length of the
experiment. The photograph shown in Figure 8a (48 h) illustrates the chemotatic response
of AB2.24 towards glucose. A clear bias of the E. coli growth towards the higher glucose
concentration confirmed that AB2.24 is chemotatic towards glucose. Figure 8b is a
computer image analysis of the relative light intensity (grey scale) of the Figure 8a
centerline (vertical). In this figure, 0 mm is the glucose source reservoir (5 mM glucose)
and the top of Figure 8a. The 50 mm position corresponds to the sink reservoir (0 mM
glucose). Inoculation was at about position 23 mm. Significantly more cells moved
towards. the highest glucose concentration at 0 mm, as indicated by the light intensity.
There was little chemotaxis towards the low concentration of glucose at the 50 mm
position. This pattern is indicative of chemotaxis since the cell concentration is biased

towards the glucose.

36

 

 

 

 

 

250
200 -
D
E 150~
m
>5
5 100 --
50—-
0 5 s
O \O N 1‘ M O\ In '— VD
e—t ‘— N N m V V
Position(mm)
a b

Figure 8 - a) Chemotaxis of AB2.24, Glucose Gradient Source at Top (i);
b) Image Analysis of Centerline.

E . ll 1' [E 1] 1-1 ..V II

After the chemotatic response towards glucose was established, the DGC was used
for mutagenesis and selection of a mutant containing feedback-insensitive AroF. The
purpose of proving chemotaxis was to utilize chemotaxis to isolate feedback-insensitive
mutants. Chemotaxis was used to draw the cells into higher glucose (and m-FT)
concentrations, where only mutants with a feedback-insensitive AroF could survive.

The experimental setup was the same as for the previous DGC experiment, with
one alteration. The glucose source reservoir now was also supplemented with 125 11M m-
FI‘. This created an additional gradient, that spanned 0 to 125 11M m-FI‘ from sink to
source.

The cells were allowed to grow in the DGC for 3 days before mutagenesis was
performed. This allowed for growth of a large initial pool of cells and establishment of the
glucose and m-FI‘ gradient. At the top of the picture (Figure 9a, 9b, and 9c) is the glucose

source reservoir and therefore the highest glucose and m—FT concentrations in the DGC.

37

The m-Fl‘ clearly inhibited the cell growth prior to mutagenesis (Figure 9a). After the 3
days of initial growth the DGC lid was removed and the cells were exposed to UV radiation
using the optimum conditions determined from the kill-curve experiments. The lid was
quickly replaced, and, the cells were allowed to continue growth until the putative
feedback-insensitive mutants had grown into the regions of high m—FI‘ concentrations
(Figures 9b and 9c). In Figure 9b, 7 days after mutagenesis, several “nodes” were
observed growing into higher m—FI‘ concentrations. The most predominant growth was
from the central region and less distinctly is the growth on the left side. Also, there is a
more diffuse growth on the right side growing into higher m—FT concentration. In Figure
10c, 9 days after mutagenesis (the conclusion of the experiment) the growths extended into

still higher m—FT concentrations and had become more distinct. The difference in the

contrasts of the pictures is an artifact of the camera.

   

Figure 9 - a) DGC Prior to Mutagenesis (3 days); b) DGC 7 Days after
Mutagenesis; c) DGC 9 Days after Mutagenesis.

38

After the completion of the DGC experiment, cells samples were taken from the
three regions showing growth in high m-FI‘ concentrations by streaking onto LB/Kan
plates and incubated at 37 °C for 10 b. Two methods of sampling were explored. In the
first method a sterile wooden applicator was stuck into the gel and used to streak out
colonies. In the second method, a micropipette was used to remove a plug of gel to streak
the plates. Two plates were streaked, one by each method of sampling, for each of the
three regions described above. This resulted in six plates of possible mutants.

After the growth on the plates, single colonies were selected and replicate plated
onto M63/HIPVRS plates with 0, 30, and 150 11M m—FT and a LB/Kan master plate. The
plating was performed in the listed order to assure that an adequate number of cells were
introduced onto each plate. Since, if there were not a sufficient amount of cells on the
applicator to plate onto all four plates, there would be no growth on the last plate (LB/Kan).
Thus, the LB/kan plate was to assure that the lack of growth on any minimal plate was due
to inhibition and was not a result of poor replicate plating. Colonies that grew well on
either the 30 or 150 11M m-FI‘, as compared to the 0 11M m-FI‘ plate, were tested for
tyrosine insensitivity. Four colonies were tested for insensitivity and two were insensitive.

To assay for DAHP synthase specific activity colonies were taken from the LB/Kan
master plates. The colony was removed with an applicator and first used to make a master
plate of the colonies tested, then used to inoculate the starter culture, 50 mL LB/Kan in a
250 mL Erlenmeyer flask (37 °C and 250 rpm). After 10 h of growth, 5 mL of the starter
culture was added to the growth flask consisting of 500 mL LB/Kan in a 2 L Erlenmeyer
flask (37 °C, 250 rpm, 12 h). To harvest the culture for the enzyme assay, centrifugation
of the culture (4000 rpm, 5 min 4 °C) was followed by resuspension of the cell pellet in the
resuspension buffer. The cells were disrupted by two passes through a French Pressure
cell (SML Aminco) at 11,000 psi. Cellular debris was removed from the lysate by
centrifugation (20,000 rpm, 20 min, 4 °C). Protein was quantified using the Bradford

39

dye-binding procedure.25 A standard curve depicting the absorbance at 595 nm versus
protein concentration was prepared using bovine serum albumin. The protein assay
solution (5x concentration) was purchased from Bio—Rad. The specific activity of araF
was quantified by the DAHP synthase/thiobarbituric acid (TBA) assay.26 One unit of
DAHP synthase activity was defined as 1 11le of DAHP formed per minute.

To confirm insensitivity to tyrosine, the assay was performed first without tyrosine
in the assay buffer to obtain the control specific activity. Next, the buffer was
supplemented with tyrosine to a final concentration of 125 11M and the assay performed
with another aliquot. From the test of four possible mutants, two were completely
insensitive (AC1-l7 and AC2-13), one was sensitive (AR2-20), and one was sensitive
(ALI-48) but had much higher specific activity relative to the other three tested mutants
(Figure 10). The two insensitive mutants were both from the central “node”, while the

 

0.18
0.16
0.14
0.12

0.1
0.08

 

 

 

 

 

 

 

Specific Activity (units/mg)

 

 

 

 

 

 

ACl-l7 AC2-13 AR2-20
‘ Strain ‘

The assay buffer containing: (1) No tyrosine; (2) 125 11M tyrosine.

Figure 10 - Mutant Strain Tyrosine Insensitivity Assay.

40

sensitive strain (AR2-20) was from the diffuse growth on the right side. The mutant with
higher activity was from the left “node” of the DGC. Mutant araF (isolated from AC2-13)
and wild-type araF relative specific activities in the presence of tyrosine were compared.
The mutant araF was not inhibited by tyrosine concentrations as high as 330 11M (Figure
11), and may have even been somewhat stimulated by tyrosine, which was also reported

for a feedback-insensitive AroF by Herrmann.27

 

IInsensitive AroF
uSensitive AroF

Percent Relative Activity
8

 

 

 

 

Tyrosine (11M)

Figure 11 - Percent Relativity Activity.

Next, the strains were tested for inhibition by phenylalanine and tryptophan to
ensure that the new mutation did not confer sensitivity to the other aromatic amino acids.
The concentrations of phenylalanine and tryptophan used to detect inhibition were
determined from the concentrations necessary to inhibit AroG and AroH, respectively. In
cell free extract, AroG is 60% inhibited by 10 11M phenylalanine and AroH is 20%
inhibited by 10 11M tryptophan.28 A concentration of 100 11M of phenylalanine or

tryptophan was used to ensure the detection of any sensitivity. The assay for sensitivity

41
was performed as before. A control assay was performed without added phenylalanine or
tryptophan, and then the assay was performed in the presence of 100 11M phenylalanine or
tryptophan. Neither AC1-17 nor AC2-13 was sensitive to phenylalanine or tryptophan
(Figure 12).

 

 

 

 

 

 

E?
E
S
v II
.E‘
.g I2
It; 133
E
i I4
tn

 

 

 

ACl-l7 AC2-l3

The assay buffer containing: (1) No phenylalanine; (2) 100 11M phenylalanine;
(3) No tryptophan; (4) 100 11M tryptophan.

Figure 12 - Phenylalanine and Tryptophan Sensitivity Test.

I l . E l E I] l I . 'v E

The feedback-insensitive araF (aroFb') was then cloned from the genome. The
cloning and isolation of araF" were conducted by Kai Li (PhD. student) and Dr. Karen
Draths. The first step was the isolation of genomic DNA. The method used was modified
from Silhavy.29 Three genomic isolations were performed. One colony of AC 1-17 and
two colonies of AC2-13, designated AC2-13A and AC2-13B, were used. The second step

was the amplification of araF" from the genomic DNA.

42

The PCR amplification unexpectedly had problems for unknown reasons. After
several attempts, the araF“ gene was amplified from the genome using the primers JWF-19
and JWF-22 (Table 1) which have EcaRI ends. The PCR products were then ligated into
the low copy number plasmid pCL1920 (spectinomycin (Spc) resistance-bearing plasmid)
for sequencing. The three resulting plasmids, one from each genonric DNA isolation, were

designated pCL1-17-1,pCL2-13A-1, and pCL2-13B-l (Figure 13).

pCLl-l7-l pCL2-13A-1 pCL2-13B-l

5.5 kb EcoRI EcoRl
araF

 

 

 

 

Figure 13 - Plasmid Map of Cloned Mutants.

S . E l E 1! l -I . . E

The polyethylene glycol precipitation method was used by Dr. Karen Draths to
isolate the DNA fragment for sequencing.” The primer used for sequencing were JWF-
22, JWF-79, JWF-80, JWF-81, and JWF-82 (Table 1). The primers and template were
sent to the Molecular Sequencing Facility at Michigan State University for sequencing. The
sequence of the araF" was found to be different from the tyrosine sensitive araF by a
cytosine to a thymine base change. This base change causes a proline (residue 148) to
leucine substitution conferring insensitivity. This is the same mutation that was obtained
by Herrmann.31 This base change introduced a Bng restriction site into the gene (Figure
14). Digestion of the three plasmids containing araF“ with both EcoRI and Bng

confirmed the presence of the new restriction site.

43

Feedback Insensitive araF
B 111
0.4 kb i 0.8 kb

 

‘J

Figure 14 - Map of Feedback Insensitive araF.

As stated by Herrmann,32 this particular amino acid residue change is noteworthy.
In tyrosine-sensitive AroF residue 148 is proline. The corresponding residue in AroG and
AroH is methionine. Leucine and methionine are similar amino acids in respect to
hydrophobicity and their effect on protein secondary structure. Residue 148 of AroF is in a
region of little sequence homology to AroH. This region of AroH has been identified as a
major part of the allosteric binding site. Changes of AroH amino acid residues flanking
residue 148 have conferred insensitivity to tryptophan. The corresponding region in AroF
appears to be involved in the allosteric binding site also. The variation of the sequence in
this region between AroF and AroH are critical to the different allosteric binding pockets of
the enzymes. It appears that AroH and AroF have similar domains that are critical to
allosteric binding sites.

ElllIHEESfi!”

The three plasmids containing feedback-insensitive amF were each transformed
into competent A33248 in order to assay the specific activity of araF". A33248 does not
possess any native DAHP synthase activity. The strains were grown as previously
described for specific activity assays, except the medium contained spectinomycin (50 mg
L") for plasmid selection pressure. The activities are illustrated in Figure 15. The results
show that PCR did not alter feedback insensitivity.

0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05

Specific Activity (units/mg)

 

pCL2-13A—1 pCL2-13B-l pCL2-17-1
Strain

The assay buffer containing: (1) No tyrosine; (2) 125 [1M tyrosine.

Figure 15 - Cloned AroF Specific Activity.

The next genetic manipulation was the removal of tyrR boxes in an attempt to
eliminate or reduce transcriptional repression by tyrosine (Figure 16). A similar approach
was used in the production of phenylalanine, whereby the entire native promoter was
removed and replmd with a temperature-sensitive promoter.33 For our purpose, all of
tyrR Box 3 and half of tyrR Box 2 were eliminated by PCR amplification of araF“ from
pCL2-13A using the primers JWF-94 and JWF-22 (Table 1). The PCR product was
cloned into pCLl920 generating the new plasmid pCL2-l3A-trunc (Figure 17), which was
transformed into competent AB3248.

The two constructs were assayed for DAHP synthase activity using the previously
described conditions. The results are illustrated in Figure 18. Activity of the truncated

araF"I was lower. The lower activity of the truncated araF was unexpected. It was

45

thought that the specific activity of the truncated araF would either be the same or better

then the nontruncated araF because only the regulatory sequence of the gene was changed.

araF

 

‘J

/-f‘tyrRBoxl

tyrR Box 3
tyrR Box 2

Figure 16 - Map of araF tyrR Boxes.

The AroF specific activity was then assessed during shake flask experiments under
conditions used to accumulate DHS. This was done to better simulate the actual growth
conditions used to produce DHS. The experiment was performed for both AB3248/pCL2-
13A (as a control) and AB3248/pCL2-13A-trunc. A single colony was used to inoculate
50 mL of LB in a 250 mL Erlenmeyer flask. After growth for 10 h (37 °C and 250 rpm),
5 mL aliquots of this starter culture were used to inoculate, in triplicate, 500 mL of
supplemented M9 in a 2 L flask. The carbon source was 10 g L" (56 mM) glucose, and
supplemented with was 40 mg L" phenylalanine, tyrosine, and tryptophan each, and 10 mg
L" pcaminobenzoic acid, p-hydroxybenzoic acid, and 2,3-dihydroxybenzoic acid. These
accumulation cultures were grown for 36 h (37 °C and 250 rpm). At 12, 24, and 36 h, one
culture of each construct was harvested for the enzyme assay. The results are summarized
in Figure 19. Under these conditions the specific activity of the truncated araF“ is higher.
Since the trend is the opposite to that shown in Figure 18, it appears that the AroF specific
activity is dependent on the growth conditions. The truncated and nontruncated AroF

46

pCL2-13A-1-trunc

5.4 kb ECORI ECOR]
I araF

 

 

 

 

Figure 17 - Plasmid Map of pCL2-l3A-trunc.

0.35

 

 

 

 

.o
m

.o

N

M
I

 

p
N

 

I1
:12

 

FD

H

U!
I

.0
fl
1

 

 

Specific Activity (units/mg)

 

 

 

 

 

 

 

 

.0—

 

pCL2-l3A-1 pCL2-13A-trunc
Strain

The assay buffer containing: (1) No tyrosine; (2) 125 [J.M tyrosine.

Figure 18 - Nontruncated and Truncated AroF Specific Activity.

47

0.45 ...__
0.4
0.35
0.3‘

0.25 IPCL2-13A
0.2 DpCL2-13A-trunc

 

 

 

 

 

 

 

 

 

 

 

0.15

0.1 :
O- i , 4 .
24 36

12

 

 

 

 

 

 

Specific Activity (units/mg)

 

 

 

Time (b)
Figure 19 - Minimal Medium Nontruncated and Truncated AroF Activity.

specific activities both decreased with time. This trend has been reported before for the
araF isozyme of DAHP synthase34 and it has been suggested” to be the result of selected
proteolysis of the araF isozyme during stationary phase.

The plasmid vector was subsequently changed, so that the araF“ could be used in a
host/plasmid system and grown under conditions more similar to the fermentor conditions.
This would also allow for monitoring of DHS production to investigate the relationship
between AroF specific activity and DHS production. DHS was quantified by ‘H NMR
analysis which is described in Chapter 3 (p 71). Both the araF" and truncated araF" were
cloned into the plasmid vector pSU l 8. The truncated araF"r was only obtained in the
opposite direction of the lac promoter that is on the pSU18 plasmid vector, and this new
plasmid was designated pKL4. 19B. The nontruncated araF" was obtained in both
orientations with respect to the lac promoter yielding plasmids pKL4.20A and pKL4.20B .
Subsequently the serA locus was cloned into pKL4. 19B and pKL4.20B creating

48

pKL4.33B and pKL4.32B, respectively (Figure 20). The serA was added to complement
the lack of serA activity in the host. Kai Li used the same homologous recombination
method described previously (p 23) to create a new host to use for DHS accumulation. The
original strain was AB2834, and the new host was designated KL3. AB2834 and KL3
lack shikimate dehydrogenase (aroB) activity and therefore produce DHS. KL3 also carries
an additional copy of aroB (encoded dehydroquinate synthase) inserted into serA locus of
the host. This allowed for enough amplification of aroB to ameliorate the rate-limiting
character of DHQ synthase. A nutritional requirement was used, by localization of aroB
onto the plasmid, to maintain the plasmid instead of an antibiotic resistance. The method of
growth was similar to the previous experiment. An important difference was that the starter
culture medium was the same as the accumulation medium (supplemented M9). This
required the starter culture to be grown for 24 h to reach the same optical density (OD)

obtained when LB medium was used as the starter culture. The two resulting accumulation

 

 

 

 

 

pKL4.32B
5.4 kb
P
1864 araF truncated serA
pg —
——>
Cm
pKL4.33B
5.45 kb
Plac araF serA
K _
-—->
Cm

Figure 20 - Plasmid Map of pKL4.32B and pKL4.33B.

49

cultures were assayed for DHS production and AroF activity. The results are shown in
Figure 21. The truncated araF“ construct also was evaluated in the fermentor. These
studies are summarized in Chapter 3 (p 79). At this point in the development of the
microbial catalyst the truncated form of araF“t was abandoned because of low specific
activity and negligible DHS production in both shake flask and fermentor experiments.

 

0.6 6

 

~ pKL4.32B
Specific
Activity

= pKL4.33B
Specific
Activity

+ KL4.32B
HS

Production

 

Specific Activity (unit/mg)

—a—pKL4.33B
DHS
Production

 

 

 

 

 

 

24 36
Time (h)

Figure 21 - Truncated/Nontruncated AroF Activity and DHS Production

To increase AroF activity, two more genetic approaches were tried. First, a second
copy of araF" was inserted into pKL4.33B. The second araF" fragment had been PCR
amplified with xba2 ends that were suitable for ligation into pKL4.33B. The two new
plasmids were generated because of the two possible orientations of the insert and were
designated pKL4.66A and pKL4.66B (Figure 22). This approach was used because
Michael Farabaugh, of the Frost group, had previously seen a significant increase in AroF

xbaZ xbaZ

 

 

 

 

”U
fr
‘5

3.
I3.
5.
‘L

 

 

 

 

 

——-b
Cm
pKL4.66B
6.7 kb
xba2 xba2
Pm; araF serA araF
r: '>

—->
Cm

Figure 22 - Plasmid Maps of pKL4.66A and pKL4.66B.

activity when his constructs had two copies of are)" on a plasmid (unpublished data). The
growth and enzyme assay conditions were the same as for the last experiment. The enzyme
activities and DHS production for the two constructs are shown in Figure 23 (the
supernatant was only analyzed at 24 and 36 h). The difference in DHS production
indicates that there may be an effect due to the genes’ orientations, although the effect might
not be the same in the fermentor due to the differences in growth conditions.

The second approach was the complete removal of the native promoter and
ribosome binding site (RBS). The inducible tac promoter"6 was used instead used to
control araF expression. The repressor protein, encoded by lacI‘, binds to the tac promoter
region preventing transcription of the gene. The inducer isopropyl B-D-

thiogalactopyranoside (IPTG) binds to the product of lacI‘, inactivating the repressor.

51

 

 

 

 

 

 

 

 

 

 

1'2 ~— - KL4.66A
" 12 pecific
A 1 -~ Activity
{’50 q- 10
.2 o s -_ =pKL4.66B
5, ' -. 3 5 Specific
,9 E Actrvrty
.5 0.6 T' _’ 6 m
2 I +pKL4.66A
C
° 0.4 -. DHS .
E. / ~" 4 Production
m 0.2 ~- .. 2 —u—pKL4.66B
DHS
0 . : 0 Production
24 36
Tune (b)

Figure 23 - Specific Activity and DHS Production of pKL4.66A/B.

Only the open reading frame of araFhr was PCR amplified using primers JWF-103 and
JWF-97 (Table 1). The serA locus and newly generated araF“ fragment were cloned into
pBR322, which contains the rac promoter, tac repressor lacI‘I, and ampicillin resistance
gene, generating pKL4.79B (Figure 24). The araF fragment was cloned such that the tar:
promoter and RBS of pBR322 were in the conect location and orientation to express the
gene. Three cultures were grown in triplicate using the same conditions and host (KL3) as
previously used. Seven hours after inoculating the accumulation cultures, IPTG was
added. To the first three flasks no IPTG was added, to the second set of three flasks IPT G
was added to a final concentration of 0.05 mM, and to the remaining flasks IPTG was
added to a final concentration of 0.05 mM. One flask of each of the three culture types (0,
0.05, and 0.5 mM IPTG) was harvested for the enzyme assay at 10, 21, and 44 h, and the
supernatant was analyzed for DHS. The results are summarized in Figure 25. The
induction of the expression of araF appears to be time-dependent since the highest specific

52

 

 

pKL4.79B
8.3 kb
Pix amF serA
# b
lac! q AP

Figure 24 - Plasmid Map of pKL4.79B.

1.4 5

 

-0.0 mM IPT' G
Specific
1.2 4. Activity
__ 4 _0.05 mM IPTG
Specific
Activity
A :05 mM IPTG
E Smific
.5 ; Acuvnty
m
G

1r
:5
M

1 l
U) U)
M

+0.0 mM IPTG
DHS
.5 Production

+0.05 mM IPT G
DHS

Production
+0.5 mM IPTG

DHS
Production

 

Specific Activity (units/mg)

 

 

 

 

 

Time (b)

Figure 25 - AroF Specific Activity and DHS Production of pKL4.79B.

53

activity occurred 14 h after the addition of IPTG. The 0.05 mM IPTG culture had the
highest specific activity. Again, the characteristic loss of specific activity was seen over the
48 h duration of the experiments. No DHS was produced in the 0 mM IPTG flasks, which
suggests that the expression of araF was tightly controlled. The 0.05 and 0.5 mM IPTG
flasks produced about the same amount of DHS indicating that the difference in specific
activity under these conditions had little effect on DHS production.

11° . l E l . n

The DGC experiment and related strain development resulted in a feedback-
insensitive DAHP synthase (aroFm‘). The mutation that conferred insensitivity was one
base pair change that resulted in a proline (residue 148) to leucine substitution. The
sequence of the araF" is different from the tyrosine-sensitive araF by the a cytosine to a
thymine base change. The single base change also introduces a 83111 restriction site into
the aroF“’r gene which, upon digestion, yields 0.4 and 0.9 kb fragments. The
characterization of araF”I demonstrated that this isozyme was insensitive to feedback
inhibition by tyrosine up to a concentration of 330 11M tyrosine. A slight stimulatory effect
in the presence of tyrosine was observed for the araF“.

The novel, in-situ mutagenesis and selection approach demonstrated in the DGC is
generic and can be applied to many other biological systems. The DGC method is a new
approach to develop and isolate mutants. It is based on microbial chemotaxis and selection
with respect to an inhibitory compound.

As stated by Herrmann, this particular amino acid residue change in araF“ is
noteworthy. In the tyrosine-sensitive AroF, residue 148 is proline, while the
corresponding residue in AroG and AroH is methionine. Leucine and methionine are
similar amino acids in respect to hydrophobicity and their effect on protein secondary
structure. Residue 148 of AroF is in a region of little sequence homology to AroH. This
region of AroH has been identified as an allosteric binding site. Changes of AroH amino
acid residues flanking residue 148 have conferred insensitivity to tryptophan. The

54

corresponding region in AroF also appears to be involved in allosteric binding resulting in
feedback inhibition.

Subsequent experiments evaluated the specific activity of araF" under several
different growth conditions. All growth conditions resulted in the decay of AroF activity
over 48 h, and this has been previously reported for the native araF isozyme of DAHP
synthase.37 This trend has been suggested to be the result of selected proteolysis of AroF
during stationary phase.38 The relative magnitude of the specific activity varied with the
culture conditions, but the decay of activity was not avoided. To overcome this decay of
activity, a tac promoter was used. At certain concentrations of the inducer IPTG, the AroF
absolute specific activity was comparatively higher than previous experiments that did not
use the tac promoter. Nevertheless, AroF specific activity still declined over time. It also
did not translate into better DHS production.

The specific activity of araF" varied greatly according to the growth conditions.
This effect was very profound when comparing truncated and nontruncated araF" specific
activity. When grown in LB, the nontruncated araF“ had higher specific activity relative to
the truncated araF". The lower activity of the truncated araF was unexpected since it was
thought that the specific activity of the truncated araF would either be the same or better
then the nontruncated araF under these growth conditions. It was thought that removing
the tyrR boxes of araF" would not adversely affect expression since only the genetic
sequence responsible for repression was removed. This relationship was reversed when
the hosts were grown exclusively in minimal medium. The nontruncated araF" specific
activity decreased by more than an order of magnitude when grown entirely in minimal
medium as compared to using LB as the starter culture medium. Another reversal of
relative specific activities occurred after the host was changed from AB3248 to KL3 and the
vector was changed fiom pCLl920 to pSU18. This change also necessitated insertion of a

serA locus into the plasmid. Eventually, the truncated version was abandoned because

55

constructs that contained the truncated version did not produce DHS in the fermentor
(Chapter 3, p 80), and had low activity in the shake flask.

DHS production was quantified to conelate the AroF activity with increased carbon
flow directed into the common pathway of aromatic amino acid biosynthesis. Since the
truncated araF did not produce DHS, it appears that the promoter may have been affected
when removing the tyrR boxes. The second tyrR box is near the ribosome binding site,
and this may be responsible for the lack of DHS production.

The data obtained in shake flasks on the correlation between DHS production and
DAHP synthase activity is of limited use. For consistent results DHS-synthesizing
constructs need to be grown in a fermentor. These results did indicate that the level of
DHS production varied with culture conditions. Since commercial fermentations are
performed in stirred-tank fermentors, subsequent experiments were carried out in a 1 L,
computer controlled stirred-tank fermentor. Chapter 3 describes studies of the medium
components, fermentor conditions, and araF”I activity of cultures grown in the fermentor.
The studies provide insight into the relationship between DAHP synthase activity and DHS

production under conditions which are more industrially relevant.

10

ll
12

56

Stephanopoulos, G.; Vallino, J. J. Science 1991, 252, p 1675.

(a) Lewin, B.; Genes V. 1994, p 414. (b) Herrmann, K. M.; In Amino Acids:
Biosynthesis and Genetic Regulation; Herrmann, K. M.; Somerville, R. L. Eds.;
Addison-Wesley: Reading, MA, 1983; p 309.

(a) Dewick, P. M.; Natural Product Report 1992, p 149. (b) Herrmann, K. M.; In
Amino Acids: Biosynthesis and Genetic Regulation; Herrmann, K. M.; Somerville,
R. L. Eds.; Addison-Wesley: Reading, MA, 1983; p 305.

Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983;
p 304.

Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983;
p 310.

(a) Pittard, J.; Davidson, B. E. Mol. Microbiol. 1991, 5, p 1585. (b) Herrmann,
K.; Garner, C.;J. Biol. Chem. 1985, 260, p 3820. (C) Herrman, K.; Schultz J .;
Hermodson, M.; Garner C.; J. Biol. Chem. 1984, 259, p 9655.

Cui, J.; Sommerville, R. L. J. Bacteriol. 1993, 175, p 303.

Pittard, J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C.,Ed.;
American Society for Microbiology: Washington, DC, 1995; Vol. 1, p 371.

Pittard, J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1995; Vol. 2, p . (b)
Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
ngamann, K M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983;
p 7. '

Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Hag-31m, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983;
p .

Weaver, L. M.; Herrmann, K. M.; J. Bacteriol. 1990, 172, p 6581.

(a) Widman, M.T.; Emerson D.; Chiu, C. C.; Worden R. M. Biotech. and Bioeng.
1996, in press. (b) Macnab R. M.; In Escherichia coli and Salmonella
typhimurium, Neidhart F. C., Ed.; American Society for Microbiology:
Washington, DC, 1987; Vol. 1, p 732.

13
14
15

16

17

18

19
20

21

22

23

24

25
26
27
28

57

(a) Lewin, B.; Genes V. 1994, p 1248.
Pittard, J; Wallace B. J. J. Bacteriol. 1966, 4, p 1494.

Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983;
p 307.

(a) Kulakauskas, S.; Wikstrom, P. M.; Berg, D. E. J. Bacterial. 1991, 173, p
2633. (b) Herrero, M.; De Lorenzo, V.; Tirnmis, K. N. J. Bacterial. 1990, 172, p
6557. (c) Ginter, N. Gene 1983, 21, p 133.

(a) Dell, K. A. Biacatlaytic Conversion of D-Glucase into Aromatics 1993. (b)
Hamilton, C. M.; Aldea, M.; Washbum, B. K.; Babitzke, P.; Kushner, S. R. J.
Bacteriol. 1989, 171, p 4617. (c) Penfold, R. J.; Pemberton J. M. Gene 1992,
118, p 145. (d) Gil, D.; Bouche, J. P. Gene, 1991, 105, p 17. (e) Recorbet, G.;
Robert, C.; Givaudan, A.; Kulda, B.; Normad, P.; Faurie, G. Appl. Environ.
Microbial. 1993, 59, p 1361. (f) Oden, K. L.; deVeaux, L. C.; Vibat, C. R. T.;
Cronan Jr., J. B.; Gennis, R. B. Gene 1990, 96, p 29. (g) Resnik, B.; LaPorte,
D. C. Gene 1991, 107, p 107.

Shevell, D. B.; Abou-Zamzam, A. M.; Demple, B.; Walker, G. C. J Bacteriol
1988, 170, p 3294.

Dell, K. A.; PhD. Thesis, Purdue University, 1989

Hamilton, C. M.; Aldea, M.; Washbum, B. K.; Babitzke, P.; Kushner, S. R. J.
Bacterial. 1989, I 71, p 4617.

Miller, J. H. Experiments in Molecular Genetics; Cold Spring Harbor Laboratory:
Plainview, NY, 1972.

Emerson, D.; Worden R. M.; Breznak, J. A. Appl. Environ. Microbiol. 1994, 60,
p 1269.

(a) Seymour, F. W. K.; Doetsch, R. N.; J. Gen. Microbiol. 1973, 78, p 287. (b)
Mesibov, R., Ordal, G. W.; Adler, J. J. Gen. Physiol. 1973, 62, p 203. (c) Adler
J. Science, 1969, 166, p 1588.

(a) Seymour, F. W. K.; Doetsch, R. N.; J. Gen. Microbial. 1973, 78, p 287. (b)
Mesibov, R., Ordal, G. W.; Adler, J. J. Gen. Physiol. 1973, 62, p 203. (c) Adler
J. Science, 1969, 166, p 1588.

Brasford, M. M.; Anal. Biochem. 1976, 251, p 248.

Schoner, R.; Herrmann, K. M. J. Biol. Chem. 1976, 251, p 5440.

Weaver, L. M.; Herrmann K. M.; J. Bacterial. 1990, I 72, p 6581.

Gibson, F.; Pittard, J. Curr. Topics Cell. Reg. 1970, 2, p 36.

29

30

31
32
33

34

35
36
37

38

58

Silhavy, T. J.; Berman, M. L.; Enquist, L. W. Experiments with Gene Fusions;
Cold Spring Harbor Laboratory: Plainview, NY, 1984.

Sambrook, J; Fritsch, E. F.; Maniatis, T. Molecular Cloning: A laboratory
Manual; Cold Spring Harbor Laboratory: Plainview, NY, 1984.

Weaver, L. M.; Herrmann K. M.; J. Bacteriol. 1990, 172, p 6581.
Weaver, L. M.; Herrmann K. M.; J. Bacterial. 1990, 172, p 6581.

Sugimoto, S.; Yabuta, M.; Kato, N.; Seki, T.; Yoshiomi, T.; Taguchi, H. J.
Biatech. 1987, 5, p 237.

(a) Gottesman, 8.; Methods in Enzymalagy 1990, 185, p 119. (b) Pittard, A. J. In
Escherichia coli and Salmonella typhimurium, Neidhart F. C.,Ed.; American
Society for Microbiology: Washington, DC, 1987; Vol. 1, p 371. (c) Tribe, D. B.;
Pittard, J.; Appl. Environ. Microbial. 1979, 38, p 181.

Gottesman, S. methods in Enzmalogy 1990, 185, p 119.

Amann, E.; Ochs, B.; Abel, K. J. Gene 1988, p 301.

(a) Gottesman, 8.; Methods in Enzymalagy 1990, 185, p 119. (b) Pittard, A. J. In
Escherichia coli and Salmonella typhimurium, Neidhart F. C.,Ed.; American
Society for Microbiology: Washington, DC, 1987; Vol. 1, p 371. (c) Tribe, D. B.;
Pittard, J.; Appl. Environ. Microbial. 1979, 38, p 181.

Gottesman, S. methods in Enzmalagy 1990, 185, p 119.

CHAPTER 3

FERMENTOR PRODUCTION OF DHS

EennentQLAdxantaees

Significant work has been performed towards the optimization of fermentations to
produce phenylalanine and tryptophan in E. coli .1 This previous work would serve as a
starting point for optimization of DHS production in the fermentor. At the laboratory scale,
fermentors provide a significant advantage over shake flasks. The fermentor can control
and manipulate many variables on-line, such as temperature which allows for temperature-
dependent expression systems to be used, pH, dissolved oxygen, and addition of the
carbon source. Through computer usage, it is also possible to have a variety of complex
control algorithms calculated from on-line data. One significant advantage is the greater
aeration provided by the air sparger and the impeller, which is necessary for aerobic growth
of high density E. coli cultures.2 Due to the increased aeration, much higher cell densities
and growth rates are typically achieved in the fermentor than shake flasks.

There are also many operating advantages in utilizing a fermentor. Dissolved
' oxygen can be controlled by the impeller speed and or air flow rate in the Braun fermentor.
A better degree of mixing is achieved by the impeller and baffles. In high density culture,
foaming is a significant problem, but it can be controlled by a foam sensor and computer-
controlled additions of antifoam. Samples can be removed under sterile conditions using
the harvest pipe. On-line monitoring and electronic storage allows for data recording and

retrieval of essential parameters.

59

60

The topic of this chapter is the optimization of DHS production in a fermentor
through both genetic and reaction engineering. The first parameter optimized DHS titer,
which resulted in the highest yield. The independent variables included the components
and concentration of the starter culture and fermentor medium, dissolved oxygen, and pH.
Genetically modified constructs were evaluated to obtain information relating the genetic
changes of the microbial catalyst to DHS production. The feedback-insensitive araF
generated using the diffusion gradient chamber was used and tested in fermentations.
Investigations included the relationship between DAHP synthase activity and DHS
production using the tac promoter and repressor lac]? to manipulate the DAHP synthase
activity by transcriptional control. DAHP synthase activity was also investigated by
exploitation of the gene orientation with respect to a promotm and the number of gene
copies on the plasmid.

Pathways required for biosynthetic production of DHS include, glycolysis, the
pentose phosphate pathway, and the common pathway for aromatic amino acid
biosynthesis (Figure 26). The TCA cycle is also involved because of its importance as a
major sink for the pyruvate that is produced by the phosphorotransferase system (PTS).
Glycolysis produces phosphoenolpyruvate (PEP) which is one of the two substrates
required for DAHP (3-deoxy-D—arabino-heptulosonate 7-phosphate) synthase. The pentose
phosphate pathway generates the second precursor erythrose 4-phosphate (E4P), which is
the second substrate required by DAHP synthase. DAHP synthase is the first enzyme of
the common pathway of aromatic amino acid biosynthesis.

The production of PEP via glucose metabolism leads to a competition between the
PEP-requiring biosynthetic steps and the PEP—dependent PTS for glucose uptake (Chapter
1, p 6).3 The utilization of the PTS is an important aspect in the metabolic model of DHS

production.

61

PYR: I

 

PEP—fl

 

 

 

 

 

 

 

on i
-\ car \
Ribulose—SP
1 . , .
m, xsr asp
AT?
FormicAcid
eeco2 ‘2 AD? o 57?
1.6FBP
DHAP —. GAP w. F6P J
mummi
H
mm WW"
. rep
CoA
DAHP
m l
t }/ NADH
>\ _ /m.
2coz
+2NADH

+GTP

Figure 26 - Biosynthetic Pathways for DHS Production.

The pentose phosphate pathway connects glycolysis with several biosynthetic
pathways. It consists of an oxidative and non-oxidative branch. Two enzymes of the non-
oxidative branch, transketolase and transaldolase, work in concert to produce E4P.

The first step committed in the common pathway is the condensation between E4P
and PEP to form DAHP. This reaction is catalyzed by three DAHP synthase isozymes that
are encoded by araF, aroG, and araH, which are sensitive to feedback inhibition by L-
tyrosine, L-phenylalanine, and L—tryptophan, respectively.‘ At the repressor-controlled
level transcriptional level, araF and aroG are repressed by the TyrR protein and araH is
repressed by the TrpR protein.’ The second enzyme of the common pathway is DHQ
synthase, which is encoded by the aroB locus. DHQ and inorganic phosphate are formed
from DAHP in this NAD-requiring reaction catalyzed by AroB. The third enzyme of the
pathway is DHQ dehydratase, encoded by araD. It catalyzes the dehydration of DHQ

62

which results in DHS formation. Both AroB and AroD enzymes are constitutively
expressed and are not sensitive to feedback inhibition mediated by any of the aromatic
amino acids or intermediate metabolites of the common pathway.6

The last common pathway enzyme that is important to DHS production is shikimate
dehydrogenase (aroE). This enzyme catalyzes the reduction of DHS to shikimic acid and
requires NADPH. E. coli strain ABZ834 and its derivative KL3 lack shikimate
dehydrogenase activity and therefore cannot further metabolize DHS.7

During fermentation, there are often unwanted byproducts generated such as acetic
and formic acid. Typically these result from either glucose overfeeding or anaerobic
growth.8 Acetic acid is inhibitory to growth. The presence of acetic acid will reduce the
fermentation productivity, so avoidance of its accumulation is desirable. The production of
either acid also constitutes an undesirable loss of carbon from the desired pathways. If the
aeration of the culture is insufficient acetic acid and sometimes formic acid are formed.
Formic acid production is indicative of severely anaerobic conditions, since the enzyme
pyruvate-formate-lyase catalyzes the production of formic acid is irreversible and rapidly
inactivated in the presence of oxygen.9 Acetate can also be produced because of
excessively high glucose concentration. If glucose is overfed, the TCA cycle and other
biosynthetic pathways can no longer consume all the pyruvate that is being produced from
glycolysis, and the extra pyruvate is converted into acetate and exported from the cell.10
Knowledge of conditions promoting acetic acid and formic acid formation allows optimal
fermentor conditions to be chosen to avoid their production.

In Figure 27 (only the pathways necessary for DHS production are shown) the
numbers next to the reaction arrows represent the maximum theoretical flux distribution, as
calculated by metabolic control analysis, for DHS production.11 From these calculations,
24 moles of DHS are produced from 56 moles of glucose consumed. Therefore, the

theoretical maximum molar yield of DHS from glucose is 43%.

Glucose

Ribulose-SP W

8/ \8

8
561
8 XSP RSP

   

 

 

Figure 27 - Flux Distribution for Maximum DHS Yield.

Q°"S EEHSEI'
To exploit the advantages of biocatalysis for large scale production a strategy must

be developed incorporating both industrial requirements and knowledge of the effect of
environmental conditions on DHS production. The goal is the efficient and high-yielding
conversion of D-glucose to DHS. It required genetic optimization of both microbial catalyst
by genetic modification and the optimization of fermentation conditions.

To optimize the microbial catalyst, there are various important aspects for DHS
production. An important strategy on the optimization is identification of the rate-limiting

enzymes and enzymatic regulatory mechanisms. Both central and peripheral metabolism

54

must be understood in its relation to DHS production, such as the PTS.12 Some of the
genetic modifications must be evaluated in a fermentor since the physiology of E. coli
varies greatly under differing growth conditions. Fermentor conditions and microbial
constructs were simultaneously evaluated and optimized. The fermentor conditions
investigated for optimization included, the starter culture medium, fermentor medium
components, glucose feeding rate, amino acid supplementation, and aeration.
mm

The necessary metabolic alterations necessary for DHS production were detailed in
Chapter 1 (p 9). To review the key points of the first chapter, both araF and aroB need to
be overexpressed. The overexpression of araF is accomplished by insertion of the araF
locus on a multicopy plasmid. The aroB locus was inserted into the genomic copy of serA
in the strain AB2834, creating KL3. The insertion of aroB into the genome served an
additional purpose besides overexpression, it was also used for plasmid maintenance,
described below.

W

A nutritional requirement for plasmid maintenance was used in the development of a
new host strain. The starting host strain was AB2834 which produces. DHS. The
methodology used for the generation of AB2.24 for mutagenesis was applied to the
generation of the new host for DHS production. This time the homologous recombination
into the genomic copy or serA was an aroB cassette. The inactivation of serA generated a
new strain designated KL3. For plasmid maintenance, all plasmids transformed into KL3
had the serA locus. KL3 containing serA-encoding plasmids were cultured in minimal
medium lacking supplementation. When KL3 was cultured in rich medium such as Luria
Broth (LB), chloramphenicol or ampicillin was added to the medium for plasmid selection

pressure.

65

Water
The fermentor utilized was a B. Braun Biotech Biostat‘ MD (Figure 28). The

fermentor was equipped with a platinum thermister temperature probe, a polarizable
dissolved oxygen probe, a pH probe, and an antifoam probe. It was equipped with three
six-bladed disk impellers. The temperature was maintained at the setpoint by the circulation
of either cooled or heated water through the jacket of the vessel. The vessel was sterilized
at 121 °C for 25 min with the medium inside. The fermentor vessel was connected to a
distributed control unit (DCU) for control. The DCU unit was interfaced to a Compaq
personal computer running the Braun fermentor control software MFCS. The fermentor
could be controlled by either the DCU or the NIFCS software, but only when using the
software would the data be recorded.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ill-V

 

Figure 28 - Braun Fermentor.

66

The temperature, pH, and glucose feeding were controlled with a PID algorithm.
Only the glucose feeding control constants were changed from the preset values from
Braun for fermentations. The acid and base were added with peristaltic pumps that were
part of the DCU assembly. The glucose feed was supplied by an external pump connected
to the DCU.

Samples were removed from the vessel via the harvest pipe. There were 2 outlets
for the exhaust. The main outlet was equipped with a condenser to minimize evaporative
loss, and the second outlet was an emergency exhaust if the main exhaust became clogged.
The exhausts were temporarily closed to create pressure inside the vessel, and the harvest
pipe was opened to allow the sample to flow out. When the air flow into the vessel was
high (greater than 1 L min") the back pressure was enough to push the sample out the
harvest pipe even without the exhausts closed off. The pH probe was calibrated before
sterilization, but the oxygen probe was calibrated after sterilization. The DC. probe was
calibrated to 0 and 100% saturation with nitrogen and air, respectively.

0 . . 15 C l' .

The inoculum for the fermentor was 100 mL of LB supplemented with glucose to
the final concentration of 20 g L". The antibiotic chloramphenicol or ampicillin was added
for plasmid maintenance. Since LB is a rich medium, the serine auxotrophy of KL3 would
be ineffective as a plasmid maintenance strategy. The 100 mL inoculum was grown at 37
°C for approximately 14 h in a rotary shaker at 250 rpm.

To 850 mL of distilled and deionized water the optimized fermentor mdium
components listed in Table 3 were added (Fe(IlT) ammonium citrate concentration is not
listed because the iron content varies from 28-32%). The pH was adjusted to 7 by the
addition of KOH. The Fe(III) ammonium citrate was added to supply the Fe(III) for AroF.
The AroF isozyme requires for activity one mole of iron per mole of enzyme. The citric
acid was included to solubilize the Fe(III). The dibasic potassium phosphate and
concentrated sulfuric acid were added as the phosphate and sulfur sources. The medium

67

was added to the fermentor vessel, which was then sterilized by autoclaving at 121 °C for
25 rrrin. After autoclaving, the additional sterile trace minerals and nutrients, described

below, were added.

Table 3 - Fermentor Medium Components.

 

Component Grams Concentration (mM)
Kzl'IPO4 7.5 43

Citric Acid 2 .0 10

Fe(IlI) Ammonium Citrate 0.30

Concentrated H280. 1.2 (mL) 22

 

Since KL3 is unable to further metabolize DHS, it is unable to produce chorismic
acid, the last metabolite in the common pathway. Chorismic acid is the precursor for six
end products. Three terminal pathways lead to the aromatic arrrino acids phenylalanine,
tyrosine, and tryptophan. The three other end products are essential aromatics including
folic acid, ubiquinone, and enterochelin, which are involved in coenzyme biosynthesis,
electron transport, and iron uptake, respectively.13 To overcome the aromatic amino acid
auxotrophy of KL3, the growth medium was supplemented with phenylalanine (0.7 g L"),
tyrosine (0.7 g L"), and tryptophan (0.35 g L"). Based upon the average E. coli
requirements for aromatic amino acids,” about 25 g dry weight of cells L" should be
produced given this supplementation. The concentration of tyrosine at 0.7 g L" is above its
solubility limit. Therefore, the amino acids were added as a powder just before inoculation
since it was not possible to make a concentrated stock solution. The essential metabolites

were supplemented with biosynthetic precursors (aromatic vitamins). The aromatic

68

vitamins are p-aminobenzoic acid, 2,3-dihydroxybenzoic acid, and p—hydroxybenzoic acid
and supplements for the essential aromatic metabolites are, folic acid, ubiquinone, and
enterochelin, respectively. The vitamins were each supplemented from a concentrated
stock solution to a final concentration of 10 mg L". A concentrated stock trace minerals
solution was added; the constituents and their final concentrations are described in Table 4.
The trace mineral and aromatic vitamin stock solution were sterile filtered through a 0.2 um

pore membrane.

Table 4 - Trace Mineral Supplements

 

Component mg/L Concentration (11M)
(NHA)6M07024(H20)7 3.7 3

H3B 03 24.7 400
MDC]:(H20)4 15.8 80
ZnSO.(HzO)7 2.88 10
CuSO.(HzO)5 . 2.49 10

 

A stock solution of glucose (18 g per 50 mL) was autoclaved separately. After
autoclaving, the stock glucose solution was added to the fermentor medium. This addition
adjusted the final concentration, including the 2 g of glucose which was added from the
inoculum, to 20 g L". For the constructs that had only one copy of araF with the native
promoter, the initial glucose concentration was 8 g L".

The stock solution of MgSO4 (1.0 M) was autoclaved separately. To the fermentor
medium, 2 mL of the stock MgSO4 solution was added (final concentration of 2.0 mM).
Finally, the 100 mL inoculum was added to the fermentor to give a final volume of about 1

69

L. The pH was controlled with the addition of 2 N HCl and 28% NH4OH. Sigma 204
antifoam was added manually when needed.

The choice for the dissolved oxygen (D.O.) setpoint was somewhat arbitrary.
Konstantinov et. al. reported that phenylalanine production in E. coli was independent of
the DO. level in the range of 0-40%.15 Their choice of a 20% DO. setpoint was made out
of convenience and simplicity. Also, at this D.O. level the growth rate is near its
maximum.“5

The fermentations were divided into two phases that had different control strategies.
In the first phase, the DO. was controlled by impeller speed and airflow rate. The
experiment was begun with a DO. value of about 100% saturation following the DD.
probe calibration. When then D.O. concentration fell to 20% the response was activated.
The controller operated by increasing the impeller speed and then the airflow rate. The
initial impeller speed was 50 rpm, and this value increased during the fermentation to a
maximum vale of 900 rpm. After the maximum rpm was reached, the airflow rate was
increased to maintain the 20% DO. The minimum and maximum airflow rates were 0.06
and 3.0 L min", respectively. The first phase lasted for 10 to 18 h depending on the
construct used. A

After the glucose in the first phase was consumed, the second phase of fermentation
was started. At this time, the impeller was usually at the maximum rpm (900), and there
was a rapid increase in the airflow rate to maintain the DO. The airflow rate would then
reach 3.0 L min'1 for a short time before rapidly decreasing which corresponded to
depletion of the glucose. At this point, the use of the impeller rate and airflow rate to
control D.O. concentration were stopped. Two different methods were used to start the
second phase. For both methods the impeller rate was set to 900 rpm. Preferably the
maximum airflow rate (3.0 L min'l) had been reached and had not begun declining,
allowing the flow rate to be set to this value. For some of the first fermentations using this

method, the maximum airflow rate point was missed, and it had begun to decline. So, the

70

airflow rate was set to the current value at the time of the phase change, usually about 1.0 L
min". Then the flow rate was increased in increments of about 0.25 - 0.5 L min1 to reach
a final flow rate of 2.5 - 3.0 L min;1 within about 12 - 15 h after the start of phase two.

The DC. concentration was controlled in the second phase by the glucose-feeding
rate. The glucose feedstock had a concentration of either 400 or 600 g L". Glucose was
fed when the DO. rose above the 20% setpoint, indicating glucose limitation by decreased
respiratory activity. The PID controller parameters (T able 5) were found empirically. A
dynamic simulation of the process was used to obtain a to obtain a first approximation of
the controller parameters. Then based on experience with the fermentation, the parameters

were significantly modified.

Table 5 - Phase 2 D.O. PID Controller Parameters.

Control Parameter Value

”CD 0
It 999.9 8
Xe 950.0%

 

The equation for the output of the controller is given by Equation 1.17 The
controller response as a function of time is C(t). The process gain (Kc) is defined as ratio
of the change in output (D.O. level in our case) to the change of input (glucose feeding),
Equation 2. Proportional control creates a controller response that is proportional to the
error. The error as a function of time (£(t)) is defined as the difference between the
output (DO) and the setpoint (20%), Equation 2. The integral time constant (1'!) defines
the time needed by the controller to repeat the initial proportional control response to a

71

continuing deviation error. The derivative time constant (To) and the derivative term
anticipates what the error will be in the immediate future. The proportional band (Xp) is
inversely related to the process gain, Kc (Equation 3). It characterizes the range over
which the error must change in order to drive the controller’s response over its full range.
The larger the proportional band, the lower the sensitivity of the controller’ s response to the
deviation error will be. This means that the larger the magnitude of the proportional band,

the slower glucose feed rate is for a given deviation from the DO. setpoint. The

Ker d8
t=Ke t+— etdt+Ker— 1
c() an “(,0 Ddt <>

_ AOutput
Alnput

Kc (2)

100

X”: K.

 

(3)

derivative control feature was turned off ( To: 0). The integral control parameter 1'! was
set to the maximum value allowed by the computer field width, 999.95 (Equation 1). The
control action due to integral control is inversely proportional to the parameter value;
therefore the integral control action was set to a minimum. These control values typically
worked well as demonstrated by little to no accumulation of glucose. The setpoint for the
culture temperature was 37 °C. The pH was maintained at 7.0 :t 0.05 pH units.
W
DHS, glucose, gallic acid, acetic acid and formic acid were assayed by ‘H NMR

(nuclear magnetic resonance) that were recorded on a Varian Gemini-300 spectrometer at

72

300 MHz. An aliquot (5-6 mL) of the culture was withdrawn, and cells were removed by
centrifugation. A portion (0.25-3.0 mL) of the culture supernatant was concentrated to
dryness under reduced pressure, concentrated to dryness two additional times from D20 (1
mL), and then dissolved in D20 (1 mL) containing a known concentration of TSP (8 0.00).
Concentrations of metabolites in the supernatant were determined by comparison of
integrals corresponding to each metabolite with the integral corresponding to TSP in the 1H
NMR spectra. DHS resonances are at 8 6.4, 4.3, 4.0, 3.1, and 2.7 (citric acid also has
overlapping resonances at 8 2.7). The resonances used for quantification are either 8 3.1
(doublet of doublets) or 8 4.2 (doublet), and each corresponds to a single proton (Figure
29). The resonance at 8 3.1 is preferred because the proton that corresponds to 8 4.2 is
slightly exchangeable with deuterium. The resonance at 8 4.2 is used for quantification
only when a significant amount of glucose is present which begins to overlap with the DHS
peak at 8 3.2. Glucose is quantified by summing of the integrals of the resonances at 8 4.6
and 5.2, which corresponds to the total concentration of a and B enantiomers in solution
(Figure 30). Gallic acid is quantified by the singlet resonance at 8 7.1 (Figure 31). This ‘
resonance corresponds to the protons of the aromatic ring. Acetic acid is quantified by the
8 2.1 singlet corresponding to 3 protons (Figure 32). Lastly, if formic acid is present, it is
quantified by the singlet resonance 8 8.5 corresponding to one proton (Figure 33). A
typical optimized fermentation spectrum at 36 h is shown in Figure 34. Glucose can also
be assayed using the Glucose Trinder Assay marketed by Sigma (St. Louis, Missouri).
The supernatant was typically quantified every 6 h after the start of the fermentation.
WW

DAHP synthase activity was measured according to the procedure described by
Schoner.18 Fermentor samples for the assay were taken at 12, 24, 36, and 48 h. The
volume of culture taken was 40 mL at the 12 h time point and 25 mL was taken for each
point after that. The lysate was diluted with the appropriate buffer for the assay. The assay

"' O
b

Ania—v Eu< omfiifimeuchaoA—é he :22 E. . an 0.52%

.Q: .u .36 8 .2 can a: .vu .36 8 8: ”Evacuate—Nae c8 moo-8:80”—

 

73

 

 

 

 

 

28 25238528-»

:0
:0 v- o

o N

P
Imoo

 

 

74

 

I

.382“. .e as: a. . en 8.5....

d: 2. 552.55 a «a Ea Lon—353 a 3. 8 .1 ”353.3556 .8 885.80%

omooawé

:Q.

:0 o:

00
IO

 

 

 

F D h b . I

 

 

 

b
b
I

75

.23. ease .e :22 m. . S 9...»...—

Amm .m ._ s 8 .m ”cognates—G .5.“ 35.88%

{am . ~ n 4 m o m... a

....... Nurhpnp-npp»pppnrhbhrihhbbbpbnpb-pinnpnbnrp ppnbnhbbth-n-ppppb

w - a?

 

 

28 2:8

IO
10 CI

F
zuoo

 

 

 

76

.23. 2.84.. .e :22 .m. . an 9...»...—

Amm .m ._ .N 8 .3 eeuaomweaec 8.. guacamom

 

 

 

Eon ozoom

mzooaoz
P

 

77

.32 9.5.3..— ue fizz I. . mm 0.5»...—

.E=€oE 388:3 5 d: .m .n.w 3 .: conga—manna c8 cone—88¢

Inc a .
. r. .P- P

rat
a
v-o
If!
u
bk
~o
.

h F p D

4 . 4 .. a-

 

 

 

 

 

Bow 2.52

 

Iwooz
F

 

 

78

.‘c o a u a v n o
p D b D b D

.eeeseeaeea 2.2. en .3. as: a. .825 . an as»;

Sofia ea 388< e8 .22.. 855 £8

. . a a

Id. . 3 33 at Tali .-

 

 

 

 

 

 

 

 

 

 

 

 

79

time points were taken every 15 or 30 s for a total of 6 data points. The concentration of
DAHP produced was determined by thiobarbituric acid visualization of the periodate
cleavage products as described by Gollub.’9 The absorbance of the cleavage products was

measured at 549 nm for quantification.

 

The effectiveness of the nutritional basis for plasmid maintenance was examined.
After each fermentation, an aliquot of the culture broth was diluted and plated. The plates
were LB, LB/Cm or LB/Amp depending on the plasmid host vector, and M9/glucose. The
growth on LB indicated how many viable cells were present in the culture, while the
LB/antibiotic plate indicated the number of cells that maintained the plasmid as verified by
antibiotic resistance. The ratio of the number of colonies on the LB/antibiotic plate and LB
plate gave an estimate of plasmid retention. The M9/glucose plate was used to confirm that
the strain was still an aromatic amino acid auxotrophy and had not reverted to prototrophy.
There should not be any growth on this plate because it lacked the proper supplementation
to complement the aromatic amino acid auxotrophy.

A second confirmation of the plasmid stability was also used. The plasmid DNA
was isolated and digested with proper endonucleases. The digested DNA was then
separated by size using electrophoresis on a 0.7% agarose gel. The bands were stained
with ethidium bromide to visualize the DNA fiagments with a UV lamp. This confirmed
not only the presence of the plasmid, but also the genetic composition of the plasmid.

W

The constructs used to compare the nontruncated and tnmcated araF",
KL3/pKL4.33B and KL3/pKL4.32B (Chapter 2, p 48) were further investigated in the
fermentor. The objective of the two fermentations was to determine whether the truncated
araF“ gave higher titers or yields than the nontruncated araF". The conditions were
slightly different from the optimized conditions listed above since the strain development

and optimization of the fermentor conditions occurred simultaneously. The composition of

80

the starter culture is listed Table 6. The starter culture volume was 80 mL, but the other
growth conditions were the same as for the optimized fermentations. The DC. was
maintained at 5% saturation. This D.O. setpoint was used at this time because some
previous fermentations produced higher titers at lower D.O. This trend was not observed
for subsequent fermentations. The fermentor medium composition was the same as before
although the initial concentration of glucose in the medium was 5 g L". After inoculation (6
h) glucose was fed at approximately 1 g h" for 42 h. The DC. control was that of phase
one of the optimized conditions. The impeller speed and airflow rate were manipulated to

maintain the DC. at 5%.

Table 6 - Starter Culture for Evaluation of Truncated/Nontruncated araF.

 

Component gL Concentration (mM)
K2I‘IPO4 4.8 28

KH2P04 12.2 90

Yeast Extract 2.50

(N PL);SO. 2.5 20
Glucose 1 .2 6.6

MgSO. 0.5 4. 1

 

The first fermentation was performed for KL3/pKL4.32B (truncated araF“). N o
DHS accumulated and, only acetic acid formation was observed. Fifty mL aliquots of the
fermentor culture were taken at 24 h and 48 h for the DAHP synthase specific activity
assays. The growth rate and DAHP synthase specific activity were low compared to the
next fermentation (Figure 35). The experiment was repeated for KL3/pKL4.33B

81

(nontruncated araF“). There was a small amount of DHS (l g L") produced in this
fermentation, and the DAHP synthase specific activity was higher at 24 h than the DAHP
synthase specific activity at 24 h for the truncated araF". Based on the lack of any DHS
accumulation by 10.3 containing the truncated araF” plasmid, this construct was
abandoned.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.5
- Specific
33 2 Activity
25, pKL4.32B
’§ fl = Specific
V 1 5 Actrvrty
g: - pKL4.33B
'3 D : 7 ' +g Cells Dry Wt.
5 £3 pKL4.32B
:5 as 1
cg. an / _u—g Cells Dry Wt.
2 ' |.//
0 : - , I:—
0 12 24 48
Time (h)

Figure 35 - Fermentation Evaluating KL3/pKL4.32B and KL3/pKL4.33B.

For a point of reference, a fermentation performed using unoptimized conditions is
illustrated in Figure 36. The construct used for this fermentation was KL3/pKL4.66A
which is a double araF-bearing plasmid (Chapter 2, p 50) The starter culture was grown
in 100 mL LB/Cm for 14 h (37 °C and 250 rpm). After 14 h the cells were harvested (4 °C
and 3000 x g, for 5 min) and resuspended in 100 mL of the fermentor medium. Amino
acids used as supplements included phenylalanine (1.0 g L"), tyrosine (1.0 g L"), and
tryptophan (0.5 g L"). The initial glucose concentration was 8 g L". The dissolved

82

oxygen was maintained at 5% saturation. The fermentation control was carried out in two
phases as for the optimum conditions with the exception that the maximum impeller speed
was 400 rpm. The cell growth was better than most previous fermentations, but the cell
concentration was still about one third of the optimized fermentation’s cell concentration.
The DHS production was an order of magnitude less then for the optimized case. The

molar yield of DHS was 15%. The acetic production was kept to a minimum. Typically,

 

9 4
8 4 IDHS
7 3
g 6 3
aAcetic
g 5 2 Acid
g. 4 2
(’3 3 1
an 2 1 “Cells
Drth.
1 0
0 0
0 4 8 12 17.5 24 31 37 42 48
Time(h)

Figure 36 - Nonoptimized Fermentation.

the acetic acid was consumed during phase two due to the strict control of glucose feeding
as is shown for this fermentation. This fermentation was included to demonstrate the
differences between optimized and non-optimized fermentations. A key point learned was
at 5% DC. the cell concentration was much lower than the maximum cell concentration as
calculated by the aromatic amino acid supplementation (25 g dry wt. cells L").

All subsequent fermentations cited were conducted using the optimized conditions.

The two double araF constructs (pKL4.66A and pKL4.66B, Chapter 2 p 50) were tested

83

in the fermentor. Two effects were investigated in these fermentations. The effect of
adding a second araF gene to the construct was explored to increase the DAHP synthase
specific activity. Secondly, since more than one orientation was obtained when ligating the
second araF, any effect due to the different orientation between the two construct was
explored. The first fermentation (Fermentation 960037) using the optimized conditions
was with I0.3/pKL4.66A (double araF construct, Figure 22, p 50). The fermentation was
run for 60 h to explore the DHS production over a longer time period (Figure 37). During
this fermentation, the airflow rate was incrementally increased in phase two. The DHS titer
was 19 g L" and overall molar yield from glucose was 24%.

The next three fermentation described are discussed as a group. One fermentation
was with I0.3/pI0.4.66A and the other two were run with KL3/pKL4.66B (Figure 22, p
50), designated I0.3/pl0.4.66B(1) and I0.3/p10.4.66B(2) (Figure 38). Fermentation
960037 was not included in the comparison because some of the operation was different
from the other fermentations (i.e. the airflow rate at the beginning of phase two was not
3.0 L min" ), and it ran for much longer. The DHS titers were 40, 37, and 34 g L", and
the molar yields were 21%, 20%, and 17% for pKL4.66A, pKL4.66B(l), and
pKL4.66B(2), respectively. Unlike the shake-flask experiments, the DHS production was
nearly the same for the two constructs, and there was not a difference between the two
constructs with respect to specific activity in the fermentor.

The single araF construct, KL3/pKL4.33B, was grown twice in the fermentor
(Figure 39). The DHS titers were 21 and 25 g L". The molar yields were 18% and 10%.
A fermentation was attempted using 20 g L" glucose, but after 24 h only acetic acid was
produced. The conditions had to be slightly modified from the previous fermentations.
The initial concentration of glucose in the fermentor was 8 g L" instead of 20 g L". For the
20 g L" glucose run, the growth rate and glucose consumption were slower initially.

Since AroF and the PTS compete for PEP, the PEP availability may be the cause
for the this run’s lack of DHS production. It has been assumed that the PEP-enzyme

84

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m
E +g Cells
. Dry Wt.
S
3
E I DHS
c3
60
Time (h)

30
g 25
<
.0
3 20
El IAcetic
a Acrd
g 15
3‘
If, ' uSpecific
2 10 AWVHY
E 5
‘2
§ 4——~I-

0

 

 

 

 

 

 

N N M V W '0
Time (h)

Figure 37 - KL3/pKL4.66A Fermentation.

 

—o—g Cells Dry Wt.
pKL4.66A

Cells Dry Wt.
pKL4.66B( l)

+g Cells Dry Wt.
pKL4.66B(2)

+DHs pKL4.66A

g Cells Dry WtJL, DHS (g/L)

—o—DHS
pI0.4.66B( 1)

 

 

 

_|—DHS
pKL4.66B(2)

 

 

IpKIA.66A
(10 x)

 

—""'"l

 

npKL4.66B(l)
(100 x)

 

 

 

 

 

 

_ IpKL4.66B(2)
(10 x)

 

 

 

Specific Activity (units/mg)

 

 

 

 

6 12 18 24 30 36 42 48
Time (h)

30

 

 

20 _ “10.4.6“

 

 

15 apl0.4.66B(1)

 

 

Acetic Acid (mM)
1
I

IpKL4.66B(2)

 

 

 

 

 

 

 

 

42 48

 

Figure 38 - Fermentations of KL3/pKL4.66A and pKL4.66B.

g Cells Dry WtJL, DHS (g/L)

Specific Activity (units/mg) Acetic Acid (mM)

86

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

35
30 -. ' - —-0—-6vCellsDry
' t.
pKL4.33B( l)
25 —~
0“ \ 4‘
I \ or, \ J 6 —o—g Cells Dry
20 -» Wt.
I pKL4.33B(2)
- . ' I
15 ' / - -A- -DHS
f pKL4.333( 1)
l
10 _. I- r
l I
I —-o—DHS
5 _ _ pKL4.33B(2)
0 1 i L 1
0 10 20 30 40 50
Time (h)
30
a Specific
Activity
25 ' pKL4.3BB(1)
20 ISpecific
Actrvrty
pKL4.33B(2)
15 J
I Acetic Acid
pKL4.33B(1)
10 —
5 _ nAcetic Acid
pKL4.33B(2)
0 _
12 18 24 30 36 42 48
Time (b)

Figure 39 - Fermentations of KL3/pKL4.33B.

87

complex is the native form of araF since the K... for PEP is 5.8 M, and the intracellular
PEP concentration never falls below 88 rrM.20 With the double araF construct, more PEP
is channeled into the biosynthesis of DHS initially because there is more of the enzyme to
bind the PEP, resulting in a reduction of glucose uptake. The construct with a single cOpy
of araF would siphon less PEP into the synthesis of DHS, thereby increasing the PEP
availability for the PTS. Since initially more glucose is transported into the cell because of
high PEP availability, the natural regulatory mechanism of catabolite repression would be
significant. A kinetic rate expression for the PTS was derived by Liao et. al.,“ and it
predicts the rate of glucose transport is affected by the PEP/pyruvate ratio. Increases in the
ratio raise the apparent maximum reaction velocity of the PTS (V...) and reduce the
apparent K... for glucose. The kinetic rate expression demonstrates that there is a link
between the PEP availability and glucose transport. The two single araF fermentations had
lower DHS titers and lower DAHP synthase activities as compared to the double aroF
constructs. The DHS titers of the single araF construct decreased almost 40% compared to
the double araF constructs. It appears that the double araF under these conditions is a
better construct.

A strain used the tac promoter to manipulate araF expression was then tested. The
gene that encoded for the tac repressor lac]q on a plasmid allowed for manipulation of araF
expression into which a copy of araF was cloned. Expression could be controlled by
addition of the inducer isopropyl B—D—thiogalactopyranoside (IPTG). This compound binds
to the repressor and changes its conformation so that the repressor can no longer bind to the
ribosome binding site to stop transcription. The plasmid pKL4.79B was described in
Chapter 2 (Figure 24 ,p 52). The first experiment was to add IPTG after the initial lag
period, about 4 h after inoculation. The cell concentration was at least an order of
magnitude greater in the fermentor than in shake flasks, so it was not clear how to correctly
scale the IPTG. Addition of l g of IPTG was arbitrarily chosen for a first fermentation.
The results are summarized in Figure 40. The initial specific activity was high but dropped

88

to about the activity of the double aroF constructs. DHS titer, 17 g L", was lower than
either the double or single araF constructs. The molar yield was 29%.

For the next five fermentations, IPTG was added throughout the fermentation to
attempt to maintain a stable level of DAHP synthase activity over the 48 h fermentation.
The subsequent schedule for adding the IPTG was one addition at the time of the phase
change and then one addition every 6 h. The last addition was after the start of stationary
phase. The IPTG was typically added at about 8, 12, 18, 24, 30, 36, and 42 h after
inoculation. The amount of IPTG added to the culture, was either 0, 0.32, 1.6, 8.0, or 40
mg per addition. For each fermentation the same amount of IPTG was added for each
addition (i.e. one fermentor run had seven additions of 0.32 mg IPTG). The total amount
added was 0, 2.24, 11.2, 56, or 280 mg IPTG. While this method produced a range of
DAHP synthase specific activities, a constant specific activity throughout the fermentation
was never achieved (Figure 41). The DHS molar yields were 30%, 42%, 52%, 30%, and
26%, respectively. The most striking result from this set of fermentations was that the
highest specific activity did not produce the highest DHS titer. Apparently, there is an
optimal specific activity for DHS production that does not correspond to the maximum
specific activity (Figure 42). The highest DHS titer and best yield were for the 1.6 mg
IPTG addition fermentation. Also, unlike the shake flask experiment with this construct,
expression of araF under these conditions cannot be completely stopped in the absence of
IPTG. This finding further supports the idea that the physiology of the organism varies
greatly with growth conditions.

Sealedzllafiennematinns

The fermentation conditions was scaled up to a 50 liter working volume fermentor
at Biosys Inc. (Columbia, Maryland). This fermentor was also a B. Braun Biostat". The
physical design of the 50 L fermentor was very similar to the IL fermentor. The same
probes (temperature, pH, and DO.) and impeller design were used, and the acid and base

89

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

30
g +gCells
D Drth.
5
3 u DHS
5
63
00
Time (h)
14

E 12 '1

E

2

.55 10

< 1 DAcetic

E) 8... Acrd

3 6m —1 ISpecific

g 4-—

<

0

SE 2-—

m

0 . a i 1 L;
4 7 12 24 30 36 48
Time (b)

Figure 40 - Fermentation of KL3/pKL4.79B 1.0 g IPTG.

Specific Activity (units/mg) DHS (3M 8 Cells Dry WtJL

AceticAcid(mM)

Figure 41 - Fermentations of KL3/pKL4.79B, IPTG Titration of araF.

 

_._40m
IPT'

+8.0 mg
lP'rG

—.—1.6 mg
1PTG

—0—-0.32 mg
IPTG

 

 

10

‘ ' + 0.0 mg
20 30 40 50 IPTG

Tune (11)

 

 

 

° in" "('3‘

+8.0 mg
IPTG

+1.6 mg
IPTG

_._0.32 mg
IPTG

 

 

-—0.0 mg
IPTG

I40 mg IPTG

l8.0 mg IPTG

I 1.6 mg IPTG
(10 x)

.032 mg IPTG
(10 x)

l0.0 mg IPTG

Time (h) (100 x)

I40 mg IPTG

l8.0 mg IPTG

al.6mgIP'I‘G

.032 mg [PIC

18 24 30 36 42 48
Time (11)

no.0 mg IPTG

91

 

 

 

 

 

0 5 10 15 20 25 30 35 40
[lP'I‘G] (ms/addition)

Figure 42 - DHS Production as a Function of AroF Activity.

were added through peristaltic pumps that were a part of the fermentor unit. Again, the
glucose was fed by an external pump but a second D.O. probe was used to control the
glucose feeding in phase two since the first D.O. probe could not be connected to a
controller for glucose feeding. This fermentor did not have the capability to control D.O.
using first the impeller and then the airflow rate. Instead, the air flow rate was set to 40 L
min" throughout the fermentation, and the impeller speed was used to control DO. The
medium composition was the same as for the 1 L fermentations. After about 8 h, the
culture stopped growing and the dissolved oxygen quickly rose. Upon addition of
NH.OH, the DD. quickly dropped indicating an increased respiratory demand. The
difference between this fermentation and the previous 1 L fermentations was at the IL scale
the pH dropped upon inoculation, and NI-LOH was added to increase the pH to near
neutrality. It had been noted that he pH did not drop after inoculation of the 50 L culture
and therefore no NH,OH was added. These results suggest that nitrogen was limiting

92

growth. The first fermentation produced a significant amount of acetate (146 mM) and only

about 6 g L" of DHS.
For the second fermentation, the pH of the medium was adjusted with NH,OH

instead of KOH to supply more nitrogen to the culture at the beginning of the fermentation.
This fermentation accumulated 43 g L‘1 of DHS (Figure 43), and the molar yield was 39%.
The acetic acid production was very different from the 1 L scale. Acetic acid was produced
throughout the fermentation, whereas the 1 L fermentations typically only produced acetic
acid at beginning and end. The acetic acid production throughout the fermentation may be
because that the glucose feeding control parameters were arbitrarily set, and may have
cause over feeding of the culture or oxygen-liming conditions. A summary of all
fermentation in this chapter is given in Table 7.
E. . 1 C 1 .

The double araF construct produced the highest DHS titer (40 g L") obtained to
date. None of the fermentations using the nuuitional requirement for plasmid maintenance
showed significant plasmid loss or instability. Through the manipulation of araF activity,
by addition of IPTG, it was indicated that an intermediate level of DAHP synthase activity
gave the highest DHS titer. The fermentations have reached 25% DHS yield from glucose.
There was a general correlation between DHS titer cell concentrations (Figures 37, 38, 39,
40, 41, and 43). This correlation has also been reported for tryptophan production.22

Using the data from the highest DHS titer fermentation (KL3/pKL4.66A) the yield
coefficients, productivity, and average oxygen masss transfer rate were calculated. The
coefficient for the production of DHS from glucose was 0.16 g DHS per gram glucose (Y m
= 0.16), and the yield coefficient of DHS based on oxygen was 0.053 (YW = 0.053). The
productivity was 0.83 g DHS L" h". An electron and carbon balance were also calculated
for this fermentation to provide insight into the loss of energy and carbon. The electron

balance showed that 66% of the electrons available from the glucose were not accounted for

i

g Cells Dry WtJL, DHS (g/L),
Acetic Acid (mM)
8 a

10 --

3

2.5
37)
E
B
E

g 1.5
<
2

’g 1
I")

0.5

0

93

 

 

4... g Cells
Dry Wt.

+DHS

+Acctic

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12

Figure 43 - 50 L Scale Fermentation.

94

 

 

or: we 3. 2 R a mere ease: as manaemfi
OE we de M: an .e 98:: .0888: 0:: mmESMEmAM
or: me e: S mm e use .228... .5... managed.
or: w... o... 2 3 e mere cease: es amalgama—
oE. we 9. : S e mere 3.28.: ea 8:40:36.
00:0 @0203 0?: wag m _ S 3. RE: :08an 08 mafiaémd—
202:8 “00:35:: C em 5 RE: 03:0: m8.3v_&mig
202:8 3385:: 8 mm 5 RE: 03:2. acqaémig
202:8 00383:: Na ca 5 RE: 03:0: <3.§&md~
accede Basic 3 a .e use 228 3.8.30.5:—
m:c:3:oo 33:33:: m _ fin .e he: 03:0: <cc.a&mi§
252:8 8:58 2 2 e nee ewe... managed.
23:38 3.38:9: 3 K a RE: 0&5: mmm.3v_:\mx§
2. RE: @8355

.23 :ommtwafioo vé _ .e keg: 3.00555: mannamxmig

5 he: 32.0555:
5; 53:38.00 6 o 5 he: 303055 mafigémd

8:08:80 Aeev 20; .332 A63 :05. $5 83:08:35 3.303% $8.:ch

95

in the products of cell mass and DHS, and from the carbon balance calculation, 131 g
glucose (56%) was not accounted for in the products. From the carbon balance it was also
calculated that 3.7 moles of carbon dioxide (C02) was produced per mole of glucose
consumed. The theoretical minimum production of CO2 is 3 moles CO2 per mole of
glucose, indicating the carbon balance does not violate the theoretical yield. The time
averaged oxygen mass transfer rate (at 20% DD.) was calculated to be 514 h“ (K,a = 514
h'1 ), which provides a minimum oxygen transfer rate that needs to be maintained during
phase 2 of the fermentation.

Due to the large number of variables involved many of the parameters were not
fully investigated. The concentrations of the medium components have not been optimized.
Although DHS production medium components are all necessary, such as the Fe(III)
ammonium citrate, the exact concentrations may not be critical.

The fermentation scaled well to 50 L. However, this fermentation demonstrates
that several issues still need to be addressed. The air flow rate used the l L scale of 3 L
min'1 was very high. This airflow rate corresponds to 3 volumes of air per volume of
culture per minute (3 vvm). The air flow rate used for the 50 L run was only 40 L min", or
about 1 vvm. The new air flow rate did not alter the DHS titer. The choice for the DD.
setpoint was somewhat arbitrary. The choice was based upon a phenylalanine production
study by Konstantinov et. al.,23 in which phenylalanine production was found to be
independent of the D.O. level in the range of 040%.“ Therefore their criteria for the 20%
DO. setpoint was convenience and simplicity. Research has indicated this D.O. level is
not always practical. To maintain 20% DC. at the end of a fermentation, the airflow rate
was 3 L min", causing liquid to be carried out of fermentor and thereby clogging the
hydrophobic exhaust-air filters. Also, this flow rate (3 vvm) is too high for industrial scale
use. It will be important in the future to investigate the DHS production at lower D.O.

values.

96

The glucose-feeding strategy for phase two was chosen for several reasons.
Previous fermentation experience indicated that at high glucose concentrations very little, if
any, DHS is produced. Also the feeding strategy needs prevent the glucose addition rate
from exceeding the respiratory capacity of the cells for a given rate of oxygen transfer into
the medium. The method chosen was to keep the glucose concentration low enough to be
rate-limiting and to feed glucose when the DD. rose above the setpoint value. This
approach prevented accumulation of glucose to levels that would lead to metabolic
overflow25 or surpass the respiratory capacity and result in acetate production.

The inoculum size also deserves more investigation. The 10% (100 mL to a final
volume of l L) value used in this study is larger than is typically used in industrial
fermentations (2-5% inoculum). A smaller inoculum would greatly affect the costs
associated with the inoculum, since a series of fermentors are used to grow the starter
cultures for large-scale fermentations.

Significant plasmid loss using the nutritional requirement for maintenance was
never detected by either the plating on antibiotic-containing medium relative to controls
lacking antibiotics or by restriction enzyme analysis. This results demonstrates that the
nutritional plasmid-maintenance strategy worked very well.

The use of a nutritional requirement for plasmid maintenance that would not be
supplemented by a rich medium, such as LB, would be desirable for industrial-scale DHS
production. In an industrial fermentation the aromatic amino acids would most likely be
supplemented by a complex medium component (such as yeast extract). The complex
medium component would also supplement serine, rendering the plasmid selection pressure
based on serA complementation with the host KL3 ineffective. Currently, the aromatic
amino acids are added as powders. The cost of pure aromatic amino acids is high
compared to a complex medium component such as yeast extract. In addition, there has
been the concern about the nonsterile powder additions of the aromatic amino acids. At the

bench, scale this has not been a problem, but in an industrial setting, there would be cause

97

for concern. There have been other nutritional requirements, besides amino acids, used for
plasmid maintenance.26 One such method was employed by Porter et al.” The ssb gene
whose product is responsible from DNA replication was deleted for the E. coli
chromosome. The ssb locus was then placed onto a plasmid. Therefore any plasmid
lacking cell would be unable to further replicate and produce more plasmid-lacking cells.
Plasmid stability was achieved under nonselective culture conditions. An approach similar
to this maybe more appropriate for the future development of the microbial catalyst. It is
unlikely that undefined supplementation would complement the deficiency caused by the
ssb chromosomal deletion.

The glucose feeding rate had a profound effect on DHS production. When the
glucose concentration was high in the culture medium, acetic acid was formed. In the
second fermentation at the 50 L scale, a significant amount of DHS was produced despite
the high concentration of acetic acid present. There was no accumulation of glucose in the
NMR spectra from this fermentation, so the acetic acid formed was probably a result of
oxygen limitation. Formic acid was also produced in this fermentation which is only
formed in anaerobic conditions. It appears that the fermentation was anaerobic for a period
of time. However, there may be another cause to the low DHS production in fermentations
with high glucose concentration other than the production of inhibitory acetic acid. The
effect of the PTS on DHS production was discussed earlier. At very low glucose
concentrations a second high affinity glucose transport system is utilized by E. coli.28 The
significance of this system is that glucose is phosphorylated by ATP and not by PEP. This
transport system utilizes the galactose transport system and involves complex changes in
the cell membrane. The optimum medium glucose concentration for the high affinity
transport system is about 111M.” This level is well below the NMR detection limits of
about 100 M. At micromolar glucose concentrations, the E. coli growth rate was
comparable to growth rates at millimolar glucose concentrations. Also the ability to uptake
glucose through the high affinity system increased for at least 50 generations for cultures

 

98

grown in a chemostat. So, the E. coli can also mutationally adapt to the low glucose
concentration environment. It is possible that the previous fermentations primarily have
been using the high-affinity transport system. Use of the high-affmity glucose uptake
system may be the ideal solution to the problem of in vivo PEP limitations. The cell uses
its own regulatory mechanism to make more PEP available, and this method would
circumvent problems associated with manipulation of the cell’s central metabolism. In
addition, micromolar levels of glucose induce the transport systems of other sugars
opening the possibility to utilize multiple substrates during the fermentation.

The result that the optimum DAHP synthase activity for DHS production is not the
maximum specific activity raises the question of which enzyme becomes rate-limiting. As
stated in Chapter 1 (p 11), the hypothesis of metabolic engineering is identification and
amplification of rate-limiting enzymes to most efficiently increase carbon throughput in a
desired pathway. A next logical step might be to increase the in vivo availability of either
PEP or E4P. Some possible methods to increase the availability of PEP and E4P have
been tested in shake flask experiments and were discussed in Chapter 1 (p 9). A word of
caution is necessary because increasing the PEP and E4P have unseen results. In Chapter
1 (p 10) overexpression of an enzyme whose product was PEP was found to adversely
alter the central metabolism of the cell. Also, PEP is involved in more than one key step
(both the PTS and DAHP synthase reactions) for the production of DHS, and therefore any

changes involving PEP production may have unknown consequences.

 

10

11
12
13

99

(a) Konstantinov K.; Nishio, N.; Seki, T.; Yoshida, T. J. Ferm. Bioeng. 1991,
71, p 350. (b) Terasawa, M.; Fukushima, M.; Yukawa Proc. Biochem Inc. 1990,
25, p 172. (c) Sugimoto, S.; Yabuta, M.; Kato, N.; Seki, T.; Yoshida, T.;
Taguchi, H. J. Biotech. 1987, 5, p 237. (d)Tribe, D.; Pittard J. Appl. Env.
Microbiol. 1979, 38, p 181.

Swartz, J. R. In Escherichia coli and Salmonella typhimurium, Neidhart F. C.,
Ed.; American Society for Microbiology: Washington, DC, 1995; Vol. 2, p 1693.

(a) Vieth, Wolf R. Biapracess Engineering; John Wiley & Sons: New York, 1994,
p 166. (b) Postma, P. W. In Escherichia coli and Salmonella typhimurium,
Neidhart F. C., Ed.; American Society for Microbiology: Washington, DC, 1987;
Vol. 1 p 127.

(a) Dewick, P. M.; Natural Product Report 1992, p 149. (b) Herrmann, K. M.; In
Amino Acids: Biosynthesis and Genetic Regulation; Herrmann, K. M.; Somerville,
R. L. Eds.; Addison-Wesley: Reading, MA, 1983, p 305.

Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;
Herrmann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983,
p 304.

Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 371.

Pittard, J; Wallace B. J. J. Bacteriol. 1966, 4, p 1494.

Konstantinov, K; Kishimoto, M.; Seki, T.; Yoshida, T. Biotechnol. and Bioeng.
1990, 36, p 750.

Knappe, J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 151.

(a) Chao, Y. P.; Liao, J. C. App. Env. Microbiol. 1993, 59, p 4261. (b) Nan, K.;
Lim, H. C.; Hong, J. Biotechnol. and Bioeng. 1992, 39, p 663.

Patnaik, R.; Spitzer, R. G.; Liao, J. C. Biotechnol. and Bioeng. 1995, 46, p 361.
Kramer, R. J. Biotechnol. 1996, 45, p l
Herrmann, K. M.; In Amino Acids: Biosynthesis and Genetic Regulation;

Hgarnann, K. M.; Somerville, R. L. Eds.; Addison-Wesley: Reading, MA, 1983,
p .

14

15

16

17

18
19
20

21
22
23

24

25

26

27
28

29

100

Neidhardt, F. C. In Escherichia coli and Salmonella typhimurium, Neidhart F. C.,
Ed.; American Society for Microbiology: Washington, DC, 1987 ; Vol. 1, p 3.

Konstantinov, K. B.; Nishio, N .; Seki, T. Yoshida, T. J. Ferment. Bioeng. 1991,
71, p 350.

Huang, J.; Dhulster, P.; Barbotin, J.; Thomas, D. J. Chem. Tech. Biotech. 1989,
45, p 259.

Stephanopoulos, G. Chemical Process Control; Prentice Hall, Englewood Cliffs,
New Jersey, 1984, p 248.

Schoner, R.; Herrmann, K. M. J. Biol. Chem. 1976, 251, 5440.
Gollub, E.; Zalkin, H.; Sprinson, D.B. Methods Enzymol. 1971, 17A, 394.

Pittard, A. J. In Escherichia coli and Salmonella typhimurium, Neidhart F. C., Ed.;
American Society for Microbiology: Washington, DC, 1987; Vol. 1, p 371.

Liao, J. C.; Hou, S. Y.; Chao, Y. P. Biotechnol. Bioeng. 1996, 52, p 129.
Tribe, D. B.; Pittard, J. Appl. Environ. Microbiol. 1979, 36, p 181.

Konstantinov, K. B.; Nishio, N.; Seki, T. Yoshida, T. J. Ferment. Bioeng. 1991,
71, p 350.

Konstantinov, K. B.; Nishio, N.; Seki, T. Yoshida, T. J. Ferment. Bioeng. 1991,
71, p 350.

(a) Chao, Y. P.; Liao, J. C. App. Env. Microbiol. 1993, 59, p 4261. (b) Nan, K.;
Lim, H. C.; Hong, J. Biotechnol. and Bioeng. 1992, 39, p 663. '

(c) Porter, R. D.; Black, S.; Pannuri, S.; Carlson, A. Bia/I‘echnal. 1990, 8, p 47.
(b) Nakayama, K. Kelly, S. M.; Curtis III, R. Bia/I‘echnol. 1988, 6, p 693 (c)
Mortensen, U.; Nilsson, J. Prac. 4th European Cong. Biotechnol., Neijssel, O.
M.; van der Meer, R. R.; Luybem, A. M.Eds.; 1987, 3, p 303.

Porter, R. D.; Black, S.; Pannuri, S.; Carlson, A. Biofl'echnal. 1990, 8, p 47.
(a)Notley, L.; Ferenci, T. J. Bacterial. 1996, 178, p 1465. (b) Alexander, D. M.;
Damerau, K; St. John, A. C.; Curr. Microbiol. 1993, 27, p 335. (c) Death, A.;
Ferenci, T.; Res Microbial. 1993, 144, p 529.

Ferenci, T. FEMS Microbiol. Rev. 1996, 18, p 301.

 

Chapter 4

EXPERIMENTAL

W

Spectroscopic Measurements
‘H NMR spectra were recorded on a Varian Gemini-300 spectrometer at 300 MHz.
Chemical shifts were reported in parts per million (ppm) downfield from sodium 3-
(trimethylsilyl)propionate-2,2,3,3,-d. (TSP 6 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 R100A chart recorder. Additional
UV spectra were measured on a Hewlett Packard 8452A Diode Array Spectrophotometer

equipped with HP 89532A UV-Visible Operating Software.

Bacterial Strains
E. coli AB2834 [tsx-352 supE42 A“ araE353 malA352 (35)] and AB3248 [F'
araF363 araG365 araH367 praA2 argE3 ilv-7 his-4 lac gal-2 tsx-358 thi 0.)] were
obtained from the E. coli Genetic Stock Center at Yale University.‘ E. coli KL3
(serA::aroB) derived from AB2834 and AB2.24 derived from AB3248 (serAzzaroB/kan)
were generated for this study.

Storage of Bacterial Strains
All bacterial strains were stored at -78 °C in glycerol. Glycerol samples were

prepared by adding 0.75 mL of an overnight culture to a sterile vial containing 0.25 mL of

101

 

102

sterile 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 medium2 contained
(per liter) tryptone (10g), yeast extract (5g), and N aCl (10g). M9 medium2 contained (per
liter) Na2HPO4 (6 g), KHZPO, (3 g), NILCl (1 g), NaCl (0.5 g), MgSO, (0.12 g), and 1)-
glucose (10 g). M63 medium2 contained (per liter) KHzPO4 (13.6g), (NH,),SO4 (2g),
FeSO,(H,O)7 (0.5 mg), MgSO, (0.12 g), and either glycerol (0.46 g) or the combination of
glycerol (1.1 g) and D-glucose (0.90 g). The M63 medium was adjusted to 7.0 pH with
KOH before autoclaving. Isopropyl B—D-thiogalactopyranoside (IPTG) was added in
varying concentration to the culture medium for strains possessing plasmids derived from
pSU18 and pSU19. Ampicillin (50 mg L"), chloramphenicol (20 mg L"), kanamycin (50
mg L"), and spectinomycin (50 mg L") were added to appropriate cultures. Stock
solutions of antibiotics were prepared in water with the exception of chloramphenicol (95%
ethanol). Solutions of inorganic salts, magnesium salts, and carbon sources were
autoclaved separately and then combined Antibiotics and IPTG were sterilized through a
0.2 pm membranes prior to addition to the culture medium. Solid medium was prepared by -
addition of 1.5% (w/v) Difco agar to LB and 1.5% (w/v) agarose (Sigma, type II: medium
EEO) to M9 or M63 medium.

1H NMR Analysis of Culture Supernatant
A 5-6 mL aliquot of the fermentor culture or a 10 mL aliquot of shake flask culture
was removed at a specific time point, and cells were removed by centrifugation (3000, 5
min). A portion (0.25-3.0 mL) of the culture supernatant was concentrated to dryness
under reduced pressure, concentrated to dryness two additional times from D20, and then
dissolved in D20 (99.9 atom %D) containing a known concentration of TSP (5 0.00).

 

103

Concentrations of metabolites in the supernatant were determined by comparison of
integrals corresponding to each metabolite with the integral corresponding to TSP in the 1H
NMR spectrum.

Restriction Enzyme Digest of DNA
Resuiction enzyme digests were performed using buffer solutions supplied by BRL
or New England Biolabs. A typical digest contained approximately 1 ug of DNA in 8 uL
TE (10 mM TrisHCl 1 mM EDTA, pH 8.0), 1 [1L of restriction enzyme buffer (10X
concentration), and 1 [IL of restriction enzyme (10 units). The reaction was incubated at 37
°C for 1 to 2 h. The digest was quenched by addition of 1.1 pL of Endostop solution (10x

concentration) followed by agarose gel electrophoresis.

Agarose Gel Electrophoresis
Agarose gels were run in TAE buffer containing 40 mM Tris-acetate and 2 mM
EDTA (pH 8.0) The concenuation of the gel was 0.7% agarose (w/v) in TAB buffer.
Ethidium bromide (0.5 ug mL") was included in the agarose to visualize the DNA fragment
with a UV lamp. The size of the DNA was determined by using two sets of DNA
standards: 71. DNA digested with HindIII or A DNA digested with EcoRI and HindIII.

Preparation and Transformation of Competent Cells

Competent cells were prepared using a procedure modified from Sambrook et al.3
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 growth (judged
from the absorbance at 600 nm reaching 0.6). The culture was transferred to a large
centrifuge bottle that had been previously sterilized with 10% (v/v) bleach solution and
rinsed with sterile water. The cells were collected by centrifugation (3000 x g, 5 min, 4

 

104

°C) and the culture medium discarded. At this point all further manipulations were
performed on ice. The cell pellet was washed with 100 mL of cold 0.9% (w/v) NaCl and
then resuspended in 50 mL of cold 100 mM CaCl,. The suspension was placed on ice for
at least 30 min and then centrifuged (3000 x g, 5 min, 4 °C). The cell pellet was
resuspended in 4 mL of cold 100 mM CaCl2 containing 15% glycerol (v/v). Aliquots (0.25
mL) were dispensed into sterile 1.5 mL microcentrifuge tubes and immediately frozen in
liquid nitrogen. The competent cells were stored at -78 °C with no significant decrease in
u'ansformation efficiency over a period of several months.

Frozen competent cells were thawed on ice for 5 min before transformation. A
small aliquot (1 to 10 pl.) of plasmid DNA was added to an aliquot of thawed competent
cells (0.10 mL). The solution was gently mixed and placed on ice for 30 min. The cells
were then beat shocked at 42 °C for 2 min and placed on ice (2 min). LB (1 mL, no
antibiotics) was added to the cells, and the sample was incubated at 37 °C (no agitation) for
1 h. Cells were collected in a microcentrifuge (30 s), resuspended in a small volume of LB
(0.1 mL), and then spread onto LB plates that contained the appropriate antibiotics. A
sample of competent cells with no plasmid 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. Strains were
retransfonned every 4 weeks to ensure homogeneity, since transformed colonies on plates

older than 4 weeks can give variable results.

Enzyme Assay
Centrifugation of the appropriate culture (3000 x g, 5 min, 4 °C) was followed by

resuspension of the cell pellet in the resuspension buffer. The resupension buffer was
comprised of 50 mM potassium phosphate (pH 6.5) that contained 10 mM PEP and 0.05
mM CoCl,. The volume of resuspension buffer (mL) was twice the wet weight (g) of the
cells. The cells were disrupted by two passes through a French Pressure cell (SML

 

105

Aminco) at 11000 psi. Cellular debris was removed from the lysate by centrifugation
(48,000 x g, 20 min, 4 °C). Protein was quantified using the Bradford dye-binding
procedure.‘ A standard curve was prepared using Bovine serum albumin. The protein
assay solution (5x concentration) was purchased from Bio-Rad.

DAHP Synthase activity was measured according to the procedure described by
Schoner.’ The cells were disrupted as previously described. The lysate was diluted in
potassium phosphate (50 mM), PEP (0.5 mM) and 1,3-propanediol (250 mM), pH 7.0.
The lysate dilution buffer, when assaying for aromatic amino acid sensitivity, also
contained either 250 M tyrosine, 200 M phenylalanine, or 200 M tryptophan. A dilute
solution of E4P was concentrated to 12 mM by rotary evaporation and neutralized with 10
N KOH. A solution was the prepared that contained E4P (6 mM), PEP (12 mM),
ovalbumin (1 mg mL"), and potassium phosphate (25 mM), pH 7.0. An aliquot (0.5 mL)
of the diluted lysate was combined with 1 mL of the E4P/PEP solution after 5 min of
incubation at 37 °C separately. After these solutions were mixed (time = 0), the reaction
was incubated at 37 °C for 5 min. Aliquots (0.15 mL) were removed, during the 5 min, at
time intervals and quenched by addition to 0.1 mL of 10% trichloroacetic acid (w/v).
Precipitated protein was removed by rrricrocentrifugation (2 min), and DAHP in each
sample was quantified by the thiobarbituric acid assay.

The concentration of DAHP was determined by the thiobarbituric acid visualization
of the periodate cleavage products.‘5 An aliquot (0.1 mL) of solution was incubated at 37
°C for 5 min. The periodate oxidation was quenched by adding 0.5 mL of a solution
containing 0.8 M Na,SO4 in 0.1 M H2804 (pH 7.0). The mixture was incubated at 100 °C
for 15 nrin and the pink chromophore was extracted into 4 mL of distilled cyclohexanone.
The organic layer was separated by centrifugation (2000 x g, 15 min, room temperature)
The absorbance of the organic layer was measured at 549 nm (e = 680000 L mol" cm").
One unit of DAHP synthase activity was defined as 1 pmol of DAHP formed per min.

 

106

W

Insertion of a Synthetic Cassette into serA of AB3248

Homologous recombination of the synthetic cassette at serA was performed with
modified literature procedures.7 Competent AB3248 cells were transformed with the
temperature sensitive plasmid containing the cassette for insertion. Cells were heat shocked
at 42 °C (2 min) followed by the addition of LB (1 mL) without antibiotics, and the cells
were incubated at 44 °C for 1 h. Transformation mixtures were plated onto LB plates
containing chloramphenicol and kanamycin and incubated at 44 °C until colonies appeared
(24 h). Colonies with the ability to grow in the presence of chloramphenicol and
kanamycin at 44 °C contained plasmid DNA inserted into the genome. After several
transformations, 11 colonies were isolated.

Cointegrates were inoculated into test tubes containing LB without antibiotics (5
mL) and grown at 30 °C with agitation for 12 h to allow excision of plasmid DNA from the
genome. Cultures were diluted (1:20,000) into fresh LB and incubated at 30 °C with
agitation for 12 h. One additional cycle of growth at 30 °C was canied out to ensure the
cells had excised plasmid DNA from the genome. Cultures were diluted (1': 20,000) into
fresh LB and incubated at 44 °C for 12 h to allow plasmid loss. Serial dilutions of each
culture were spread on LB plates containing kanamycin and incubated overnight at 44 °C.
Cells with the ability to grow on kanamycin at 44 °C still retained the kanamycin marker in
the genome.

Single colonies were replicate plated on LB plates containing chloramphenicol and
LB plates containing kanamycin and were scored for colonies that were kanamycin-
resistant and chloramphenicol sensitive at 44 °C. Kanamycin resistance signifies retention
of the kanamycin marker in the genome, whereas chloramphenicol-sensitivity signifies the
loss of plasmid DNA from the genome. Digestion of the cells and isolation of plasmid

DNA yielded no plasmid DNA, verifying the loss of free replicated plasmids for AB2.23

107

and AB2.24. To verify the cassette insertion inactivated serA, AB2.23 and AB2.24 were
replicated plated on minimal plates with and without serine. The strains were unable to

grow without serine supplementation.

UV Mutagenesis

The conditions for mutagenesis were determined through a modified method of
Miller.8 A single colony of AB2.24 was used to inoculate a 5 mL culture (LB, 50 mg L").
The culture was grown overnight (37 °C, 12 h, agitation) followed by dilution with M63
salts and spreading onto M63 medium plates supplemented with HIPRVS. The lack of
aromatic amino acid supplementation selects for cells that still have a functional araF. The
plates were exposed immediately after plating to the UV radiation for a duration of 0
(control), 5, 10, 15, 30, 60, 80, or 100 s at a distance of 18 or 24 inches from the UV
source. The UV Lamp (Sylvania Germicidal 5 W) was warmed up 0.5 h prior to the
mutagensis. The plates were incubated at 37 °C for 24 h, and the viable colonies counted.
The optimal time, at a distance of 24 inches, was 10 s. As a confirmation that UV radiation
would penetrate through the agarose gel, the absorbance of the gel was examined. The
DGC medium was prepared as describe below and an aliquot was added to a UV cuvette.
After solidification, the UV absorbance was scanned through the wavelength range of 200
to 800 nm.

Diffusion Gradient Chamber (DGC) Setup

The DGC is part of the diffusion gradient system commercially produced by Koh
Development (Ann Arbor, Michigan). For reservoirs in use during the experiments, a 0.05
urn pore size PC filter membrane (Koh Development) and a gasket (on both sides of the
membrane) were placed between each opening into the arena and the reservoir. The
reservoirs not in use were sealed off from the arena with non-permeable silastic sheeting
(Dow Corning Inc., Midland, Michigan). The reservoirs are secured to the DGC with

thumb screws. The lid and bottom plates are fastened by thumb screws.

108

The DGC was placed on a transilluminator box (TB) which can accommodate up to
three DGCs. The TB contained two 30 cm fluorescent light fixtures (single 8W, cool white
bulb). The insides of the TB were white and a piece of black felt was placed on the bottom
and the sides of the TB to provide contrasts for the pictures. The light was turned off when
not in use to prevent heating of the DGC. The TB also had a bracket for mounting a
camera above the DGC to record growth patterns photographically. The camera used was a
PULNiX TM-7CN CCD—camera. The pictures were taken using the programs Photofinish
and AutoCap.

To generate the gradient in the DGC arena, solutes contained in Erlenmeyer flasks
were continuously fed through the reservoirs of the DGC via tubing. The flow rate of 2.5
mL h" was controlled with a dual channel peristaltic pump (LAB Bromma Microplex). An
effluent chamber was mounted on a stand next to the TB served three functions. The
effluent chamber essential because it permitted control of back pressure in the chamber
reservoirs by varying the inlet height into the effluent chamber relative to the height of the
DGC reservoir outlets. This was critical since excessive back pressure may cause flooding
of the gel due to bulk flow of liquid through the membrane, or insufficient back pressure
may cause shrinkage of the gel due to siphoning of the liquid. The effluent chamber also
consolidated all of the reservoir outflows into one large waste flask as well as serving as a
sterile break in the liquid flow.

A minimal salt medium (M63) was used for the DGC experiments, and all solutions
were prepared with distilled, deionized water. The glycerol stock solution (92 g L"),
glucose stock solution (181 g L"), and MgSO4 stock solution (121 g L") were made to a 1
M concentration and autoclaved separately. The amino acid stock solution contained 20 mg
mL" of histidine (H), isoleucine (I), proline (P), arginine (R), valine (V), and serine (S)
each. The stock amino acid solution was sterilized through a 0.2 pm membrane prior to
use. The m-Fl‘ tyrosine was added as a powder to the medium. To a liter of M63 medium
5 mL of glycerol stock solution (5 mM glycerol), 1 mL of MgSO4 stock solution ( 1 mM

109

MgSO,), and 2 mL of the stock amino acid solution (40 mg L" of each amino acid) was
added (supplemented M63).

Chemotactic experiments in the DGC were carried out similarly to the experiments
described by Emerson et al.9 The arena medium (supplemented M63) was stabilized with
0.15% agarose and initially contained neither glucose nor m-fluorotyrosine (m-Fl‘). The
sink reservoir (800 mL) contained supplemented M63 medium only. The source reservoir
(800 mL) contained the supplemented M63 medium and was additionally supplemented to
the final concentration of either 5 mM glucose or 5 mM glucose plus 125 M m—FI‘. Sink
and source reservoirs were placed opposing each other. This arrangement created a 1
dimensional gradient that spanned 0 to 5 mM glucose, and 0 to 125 M m-FI‘ if present,
from sink to source reservoir.

The DGC, effluent chamber, and empty flasks were autoclaved as one unit. The
source and sink flasks containing medium were autoclaved separately, since the pressure of
the liquid and the vapor in the flask during autoclaving would disturb the reservoirs gasket
seals. After sterilization, the DGC was placed in a sterile laminar flow hood, the empty
flasks were exchanged for the full flasks, the supplements were added. Since the m-FI‘
was added as a powder, it was stirred overnight to dissolve after addition to the medium.
While still in the laminar flow hood, the arena is filled with 40 mL of the sterile molten
0.15% agarose stabilized medium cooled to approximately 50 °C. This creates a layer (1.6
cm thick) that covers the openings between the arena and each reservoir.

The starter culture (5 mL LB, 50 mg L" kanamycin) was inoculated with a single
colony of AB2.24 and was grown overnight (37 °C with shaking in a water bath). A
portion of the starter culture (1 mL) was added to 100 mL of supplemented M63 medium
containing 10 mM glycerol and 5 mM glucose in a 250 mL Erlenmeyer flask. This culture
was grown (37 °C, 250 rpm) to stationary phase (24 h). Four 1 mL aliquots were
concentrated (4x) by microcentrifugation, combined, centrifuged and combined again to
give a final concentration of 16x. The center point was inoculated with 15 ttL of the 16x

110

concentrated culture using a micropipette to disperse the cells evenly throughout the depth
of the agarose as the pipette was withdrawn from the gel. The DGC and TB were set up in
an approximately 30 °C warm room.

The flow of liquid through the reservoirs was started 6 h prior to the inoculation to
begin the establishment of the gradients. For the experiment to prove chemotaxis of
AB2.24, the photographs were taken every 1 h for the 72 h length of the experiment. For
the mutagenesis experiment, the DGC lid was removed 3 days after inoculation, and the
cells were exposed to UV radiation for 10 s at a distance of 24 inches from the source, and
the lid was quickly replaced. The cells in the DGC were allowed to continue growth until
the mutants had separated from the regions of inhibited growth (15 d). The mutagenesis
experiment pictures were taken as just prior to mutagenesis and at several time points to

document the cells growth patterns.

Recovery of Desired Cells

The cells were recovered by streaking onto LB/kan plates and grown at 37 °C for
10 h. Two methods of sampling from the 3 desired locations of the gel were used. The
first method of sampling was a sterile applicator stuck into the gel and used to streak out
colonies. The alternative method used a micropipette for removing a plug of gel for
streaking. Two plates were streaked out for each region of interest, one by each method of
sampling. This resulted in 6 plates of possible mutants.

After the growth on the plates, single colonies were selected and replicate plated
onto M63/HIPRVS plates with 0, 30, and 150 M m-Fl‘ and a LB/kan master plate. The
plating was performed in the listed order to assure that an adequate number of cells were
introduced onto each plate. Too few plated cells would be reflected by an absence of
growth on the last LB/kan plate. The LB/kan plate was a check that the lack of growth on
the minimal plates was due to inhibition and was not a result of poor replicate plating.

111

Colonies that grew well on either the 30 or 150 M m-FT, as compared to the 0 uM m—FI‘

plate, were tested for tyrosine insensitivity. Four colonies were tested for insensitivity.

Culture Conditions for DAHP Synthase Assay of DGC Mutants

The appropriate colonies for the assays were taken from the LB/kan master plates.
Each colony was removed with an applicator, used to make a master plate of the colonies
tested, and then used to inoculate the starter culture (50 mL LB, 50 mg L" kanamycin in a
250 mL Erlenmeyer flask). After 10 h of growth (37 °C, 250 rpm), 5 mL of the starter
culture was added to the growth flask (500 mL LB, 50 mg L" kanamycin, in a 2 L
Erlenmeyer flask) and incubated for 10 h (37 °C, 250 rpm). The cells were harvested and
the DAHP synthase/TBA assay was performed as described in the general methods section.

Isolation of the Feedback-Insensitive araF gene

The cloning and isolation of the feedback-insensitive araF (aroF‘h') was conducted
by Kai Li and Dr. Karen Draths. The first step was the isolation of genomic DNA. The
method used was from Silhavy.10 Three genomic isolations were performed. One colony
of AC1-l7 and two colonies of AC2~13, designated AC2-13A and AC2-13B, were used.
The second step was the amplification of araF“ from the genomic DNA.

The araF“ gene was amplified from the genome using the primers JWF- l9 and
JWF-22 (Table 1, Chapter 2) which have EcaRI ends. The PCR products were then
ligated into the low copy number plasmid pCLl920 for sequencing. The three resulting
plasmids were designated pCL1-17-1, pCL2-13A-1, and pCL2-13B-1.

Sequencing of Feedback-Insensitive araF gene
The polyethylene glycol precipitation method of isolation of the DNA fragment for
sequencing was used. The experiment was performed by Dr. Karen Draths.” The primer
used for sequencing were JWF-22, JWF-79, JWF-80, JWF-81, and JWF-82 (Table 1,

112

Chapter 2). The primers and template were sent to the Molecular Sequencing Facility at
Michigan State University for sequencing. The mutation introduced a Bng restriction site
into the gene, as indicated by the sequencing information. Digests of the araF"t with BglII

confirmed the presence of the new restriction site.

Generation of Truncated araF"

The truncated araF" gene was created by eliminating tyrR box 3 and half of tyrR
box 2. This was accomplished by PCR amplification of araF" from pCL2-13A using the
primers JWF-94 and JWF-22 (Table 1, Chapter 2). The PCR product was ligated into
pCLl920 generating the new plasmid pCL2-13A-trunc.

Replacement of the araF“r Native Promoter with the tac Promoter
The araF" open reading frame was PCR amplified with JWF-103 and JWF-97
(Table 1, Chapter 2). The serA locus and the new generated arolmt fragment were cloned
into pBR322, generating the plasmid pKL4.79B, which contains the tac promoter. The
araF" fragment was cloned such that the tac promoter and RBS of pBR322 was in the

correct location and orientation to express the gene.

Shake Flask Culture Conditions for araF“r Specific Activity Evaluations

Three plasmids, pCLl-l7-l, pCL2-13A—l, and pCL2-13B-1 were transformed into
competent AB3248 in order to assay the specific activity of araF“. A single colony of
AB3248/pCL1-17-1, AB3248pCL2~l3A-l, and AB3248pCL2-13B-1 each was used to
inoculate a starter culture and (50 mL LB, 50 mg L" Spectinomycin, in a 250 mL
Erlenmeyer flask) grown for about 10 h (37 °C, 250 rpm). After 10 h, 5 mL of each
starter culture was used to inoculate the growth flask (500 mL LB, 50 mg L"
spectinomycin, in 2 L Erlenmeyer flask) and grown for 10 h (37 °C, 250 rpm). The exact
experimental conditions were repeated for AB3248/pCL2-13A (as a control) and

113

AB3248/pCL2-13A-trunc. The assay for specific activities was performed as described on
the general methods section of this chapter.

Both AB3248/pCL2-13A (as a control) and AB3248/pCL2-13A-trunc were grown
using culture conditions for DHS accumulation. A single colony of each strain was used to
inoculate a starter culture (50 mL LB, 50 mg L" spectinomycin, in a 250 mL Erlenmeyer
flask) and grown for 10 h (37°C, 250 rpm). At 10 h 5 mL of the starter culture was used
to inoculate, in triplicate, growth flasks for each construct (500 mL M9, 40 mg L" aromatic
amino acids, 10 mg L" vitamins, and 50 mg L" spectinomycin, in a 250 mL Erlenmeyer
flask). The growth culture was grown for 36 h (37 °C, 250 rpm). One culture of each
construct was harvested for the enzyme assay at 12, 24, and 36 h.

A single colony of the constructs KL3/pKL4.32B (truncated araF“) and
KL3/pKL4.33B (nontruncated araF“) was used to inoculate starter cultures (50 mL M9,
40 mg L" aromatic amino acids, 10 mg L" vitamins, and 20 mg L" chloramphenicol, in a
250 mL Erlenmeyer flask) which were grown for 24 h (37 °C and 250 rpm). After 10 h, 5
mL of the starter culture was used to inoculate triplicate growth flasks (500 mL M9, 40 mg
L" aromatic amino acids, 10 mg L" vitamins, and 50 mg L" chloramphenicol, in a 250 mL
Erlenmeyer flask) for each construct. The growth culture was grown for 36 h (37 °C, 250
rpm). After 12, 24, and 36 h, one culture of each construct was harvested for the enzyme
assay and quantification of DHS.

For the double araF“’r constructs (KL3/pKL4.66A and KL3/pKL4.66B), the
grth conditions were exactly the same as for the previous experiment. The cultures were
harvested at 12, 24, and 36 h for the enzyme assay, but the supernatant was only analyzed
for DHS at 24 and 36 h.

For KL3/pKL4.79B three cultures were grown, in triplicate, (9 cultures total) using
the just previously described conditions, except 7 h after the inoculum was added to the
growth cultures, IPTG was added. To the first three flasks no IPTG was added, to the
second set of three flasks IPTG was added to a final concentration of 0.05 mM, and to the

114

remaining flasks [PT G was added to a final concentration of 0.5 mM. One flask of each of
the three culture types (0, 0.05, and 0.5 mM IPTG) was harvested for the enzyme assay at
10, 21, and 44 h, and the supematants were analyzed for DHS.

W

B. Braun Fermentor

The fermentor utilized was a B. Braun Biotech Biostat" MD. The fermentor was
equipped with a platinum thermister temperature probe, a polarizable dissolved oxygen
probe, a pH probe, and an antifoam probe. It was equipped with three six-bladed disk
impellers. The temperature was maintained at the setpoint by the circulation of either cooled
or heated water through the jacket of the vessel. The fermentor vessel was interfaced to a
DCU unit for control. The DCU unit was connected to a Compaq personal computer
running the Braun fermentor control software MFCS. The fermentor could be controlled
be either the DCU or the MFCS software, but usually the fermentor was controlled by the
MFCS software because only it could record and store the data.

The temperature, pH, and glucose feeding were controlled with a PID controller.
Only the glucose feeding control constants were changed for fermentations. The acid and
base were added with peristaltic pumps that were part of the DCU unit. The glucose feed
was supplied by an external pump connected to the DCU.

Samples were removed from the culture via the harvest pipe. There were 2 outlets
for the air exhaust. The main outlet was equipped with a condenser to minimize
evaporative loss of the culture and the second outlet was an emergency exhaust if the main
exhaust became clogged. The exhausts were temporarily closed to create pressure inside
the vessel, and the harvest pipe was opened to allow the sample to flow out. When the air
flow into the vessel was high (greater than 1 L min '1) the back pressure created by aeration

was enough to push the sample out the harvest pipe without the exhausts closed off.

 

115

Fermentor Medium

The optimized fermentation medium contained (per liter) KzHPO, (7.5 g), Fe(III)
ammonium citrate (0.3 g), citric acid (2.0 g), concentrated H280, (1.22 mL), and MgSO4
(0.24 g). The medium was prepared by adding the salts (except MgSO4) to 850 mL of
distilled and deionized water. The reason for using only 850 mL was to have a l L final
volume after all supplements and inoculum were added. The pH was adjusted to 7 by the
addition of KOH, except the 50 L fermentation was adjusted with NH,OH. The MgSO,
was autoclaved separately from the fermentor medium.

Aromatic amino acids were supplemented as a dry powder by the addition of
phenylalanine (0.7 g L"), tyrosine(0.7 g L"), and tryptophan (0.7 g L"). The aromatic
vitamins p—aminobenzoic acid, 2,3-dihydroxybenzoic acid, and p-hydroxybenzoic acid,
were prepared as a stock solution (10 mg mL"). The stock trace mineral solution
contained (per liter) (NIL),Mo,Ou(I-I,O)., (3.7 g), H,BO3 (24.7 g), MnC12(H,O), (15.8 g),
ZnSO,(HzO)7 (2.88 g), and CuSo,(H20), (2.49 g). The trace mineral and aromatic vitamin
stock solution were sterilized by filtration through a 0.2 um pore membrane prior to
addition

A stock solution of glucose was prepared and autoclaved separately for each
fermentation. Either 18 or 7.2 g of glucose was added to 50 mL of water and autoclaved.
This was to add to the fermentor to adjust the final concentration of glucose in the
fermentor at the start of the experiment to either 20 or 8 g L", including the glucose that
was added from the inoculum.

Fermentor Inoculum for Evaluation of KL3/pKL4.32B and KL3/pKL4.33B

The starter culture for the evaluation was (per liter) KzHPO4 (4.8 g), KHZPO4 (12.2
g), yeast extract (2.5 g), (NH,),SO, (2.5 g), MgSO4 (0.5 g), and glucose (1.2 g). A single
colony was used to inoculate the starter culture (80 mL medium, 50 mg L"
chloramphenicol, in a 250 mL Erlenmeyer flask). The start culture was grown for 12 h (37

116

°C , 250 rpm) and harvested (3000 x g, 5 min, 4 °C ). The cell pellet was resupsended in
the optimized fermentor medium and added to the fermentor vessel. The only other
difference in culture conditions from the optimized conditions was the DO. setpoint at 5%

instead of 20% during the fermentation.

Optimized Fermentor Inoculum
A single colony was used to inoculate the starter culture (100 mL LB in a 500 mL
Erlenmeyer flask) supplemented with glucose to the final concentration of either 8 g L" or
20 g L". The antibiotic chloramphenicol (20 mg L" ) or ampicillin (50 mg L") was added
for selection pressure. The inoculum was grown at 37 °C for approximately 14 h at 250

rpm.

Fermentor Conditions

The fermentation conditions were divided into two phases. In the first phase, the
DO. was controlled by impeller speed and airflow rate. Dissolved oxygen started at
approximately 100% saturation. The liquid phase dissolved oxygen concentration was
maintained at 20% air saturation by first the impeller speed, and then the airflow rate. The
impeller speed was set to a minimum value of 50 rpm, and a maximum value of 900 rpm.
After the controller reached the maximum rpm, the airflow rate was increased to maintain
the 20% dissolved oxygen. The airflow rate minimum and maximum was 0.06 and 3.0 L
min", respectively. The first phase lasted for 10 to 18 h depending on the construct.

After the glucose in the first phase was consumed, the second phase of fermentation
was started. At this time, the impeller was usually at the maximum rpm (900), and there
was a rapid increase in the airflow rate to maintain the DO. The airflow rate would then
reach 3.0 L min:1 for a short time before rapidly decreasing which corresponded to
consumption of the glucose. At the point of the phase change, the control of the impeller

and airflow rate was stopped. Two different methods, were used to start the second phase.

117

For both methods the impeller was set to 900 rpm. Preferably, the point of maximum
airflow rate (3.0 L min") was reached and had not begun declining, allowing the flow rate
to be set to this value. For some of the first fermentations the second method was used.
When the point of maximum airflow rate was missed and had begun to decline, the airflow
rate was set to the current value at the time of the phase change (usually about 1.0 L min").
Then the flow rate was increased in increments of about 0.25 - 0.5 L min'1 to reach a final
flow rate of 2.5 - 3.0 L min '1 within about 12 - 15 h after the start of phase two.

The DC. was controlled, in the second phase, by glucose feeding. Glucose was
fed with either a 400 or 600 g L" solution. The glucose was fed on demand based upon the
increase of the DC. from the 20% saturation set point. The PID controller parameters
were found empirically. The parameter were set to 0.0 (off) for the derivative control (ID),
999.9 s (minimum control action) for the integral control (1,), and 950.0% for the
proportional band (X,). A simulation of the process was performed to obtain a first
approximation of the controller parameters, but when this was done, very different
conditions were employed for the fermentation (i.e. glucose feed rate and final cell
density). As a result, the parameters were significantly modified. The setpoint for the
culture temperature was 37 °C. The pH was maintained at 7.0 :1: 0.05 pH units. The pH
was controlled with the addition of 2 N HCl and 28% NH40H. Sigma 204 antifoam was
added manually when needed.

Fermentor Setup
The fermentor vessel was prepared for autoclaving by first adding medium (850
mL) to the fermentor vessel. Sparger, impeller motor connection, and all probe
connections were protected during autoclaving by covering in aluminum foil. The
harvest pipe and the gas exhausts were closed off by clamps. One unused port was left
loose when autoclaved. This was to prevent pressure build-up in the fermentor during

the depressurize cycle of autoclaving and was critical. An additional 1 L of the

 

118

fermentor medium was autoclaved to replace any volume loss in the fermentor due to
autoclaving.

The fermentor was autoclaved for 25 min at 121 °C, at least 12 h before the
start of the fermentation. The fermentor vessel was allowed to cool in the autoclave for
about 20 min after the end of the autoclaving cycle. This was to prevent the medium
from bailing out of the fermentor vessel.

The D.O. probe was connected to the DCU after autoclaving. The DC. must
be polarized before use and was connected about 12 h before the start of the
fermentation. All other connections (probes, water, and airline) were made about 1 h
before the start of the fermentation. The pH probe was calibrated with buffer solution
of pH 4.01 and 7.0 before autoclaving. The DC. probe was calibrated about 1 h
before the start of the fermentation. Before D. 0. probe calibration, the fermentor
vessel was warmed to the operating temperature of 37 °C and the impeller speed was
set to 100 rpm. The probe was calibrated with N2, for the 0% dissolved oxygen, and
then was calibrated with air, for the 100% dissolved oxygen. The flow rate for both
calibrations was 3 L min".

Just prior to the start of the fermentation, the supplementation was added. To
the medium, 2 mL of stock MgSO4, 1 mL of stock vitamin solution, the 50 mL glucose
solution, 1 mL of trace mineral solution, and the aromatic amino acids were added
through an open port. Lastly the inoculum was added and the data collection by the
MFCS software was begun.

Fermentor Samples
Culture broth samples were withdrawn from the vessel every 6 - 8 h. When
only the DHS concentration was to be determined, a 5 mL aliquot was taken. The cells
were removed by microcentrifugation, and the supernatant was analyzed for DHS. For

DHS concentration and enzyme assay sample, 50 mL was withdrawn for the first

119

sample (12 h) and subsequent sample were 25 mL (24, 36, and 48 h). The cells were
pelleted as described in the enzyme assay section in this chapter, and the supernatant
was analyzed for DHS. For all samples the first 5 mL of sample was discarded due to
the dead volume present in the harvest pipe.

\lO‘Ut-h

10

11

120

Pittard, J.; Wallace, B. J. J. Bacterial. 1966, 91, p 1494.

Miller, J. H. Experiments in Molecular Genetics; Cold Harbor Spring Laboratory;
Planview, NY, 1072.

Sarnbrook, J .; Frotsch, E. F.; Maniatis, T. Molecular Cloning: A Laboratory
Manual; Cold Spring Harbor Laboratory: Plainview, NY, 1990.

Bradford, M. M. Ana;. Biochem. 1976, 72, p 248.

Schoner, R.; herrmann, K. M. J. Biol. Chem. 1976, 251, p 5440.

Gullub, E. Zalkin, H.; Sprinson, D. B. Methods Enzymol. 1971, 17A, p 349.

(a) Ohta, K.; Beall, D. S.; Mejia, J. P.; Shanmugam, K.T.; Ingram, L. 0. Appl.
Environ. Microbiol. 1991, 57, p 893.(b) Hamilton, C. M.; Aldea, M.; Washbum,
B. K.; Babitke, P.; Kushner, S. R. J. Bacteriol. 1989, 171, p 4617.

Miller, J. H. Experiments in Molecular Genetics; Cold Spring Harbor Laboratory:
Plainview, NY, 1972.

Emerson, D.; Worden R. M.; Breznak, J. A. Appl. Environ. Microbiol. 1994, 60,
p 1269.

Silhavy, T. J.; Berman, M. L.; Enquist, L. W. Experiments with Gene Fusions;
Cold Spring Harbor Laboratory: Plainview, NY, 1984.

Sarnbrook, J; Fritsch, E. F.; Maniatis, T. Molecular Cloning: A Laboratory
Manual; Cold Spring Harbor Laboratory: Plainview, NY, 1984.