.. i 1......“
........s....,..

533$.
hllt
. . .1
gr I
:hintfllkhr um... i 11%.“.
$1.33 ~23; a
>151; I. .15... 5:
.nfiiififl ‘Hlnmafi.
.J _ i .
1:46. 6.3.5”.— .5:
aJ-taztzm.‘ 3|)!
1 a

y 4 1. .... .
.15... 3:2. 11):... S u 13...:
-- I 5%.... Mics;
«a...» Xvi. .l\:?..t.tw .

g ‘ ‘.fl..I‘o .- V

‘ ;lt|tvsfl.:fl.§

:Ipl», I‘a.f..;.

.’:| !!I7§\.1.

I! .. .\

 

 

 

 

1 001/

 

LIBRARY
MlCt "gun State
Universlty

 

 

 

This is to certify that the

dissertation entitled

CHARACTERIZATION OF THE SUBSTRATE SPECIFICITY
OF 2,4-DICHLOROPHENOXYACETIC ACID/ALPHAexETO-
GLUTARATE-DEPENDENT DIOXYGENASE

presented by

JULIE CHRISTINE DUNNING HOTOPP

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in MIQRQBIQLQGY &
MOLECULAR GENETICS

6’

Major professor

Date ___0_4L1_21_2_Q.0.2_

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DatoDuo.p65—p. 15

 

CHARACTERIZATION OF THE SUBSTRATE SPECIFICITY OF 2,4-
DICHLOROPHENOXYACETIC ACID/a-KETOGLUTARATE-DEPENDENT
DIOXYGENASE
By

Julie Christine Dunning Hotopp

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Microbiology and Molecular Genetics

2002

ABSTRACT
CHARACTERIZATION OF THE SUBSTRATE SPECIFICITY OF 2,4-
DICHLOROPHENOXYACETIC ACID/a-KETOGLUTARATE-DEPENDENT
DIOXYGENASE
By

Julie Christine Dunning Hotopp

2,4-Dichlorophenoxyacetic acid (2,4-D)/Ot-ketoglutarate (OtKG) dioxygenase,
deA, couples the oxidative decarboxylation of aKG to the oxidation of the herbicide
2,4-D using a mononuclear non-heme Fe(II) active site. To better define the substrate
Specificity Of this enzyme, a variety of non-phenoxyacetate compounds were examined as
potential substrates. 2-Naphthoxyacetic acid was the best alternative substrate tested,
followed by benzofuran-Z-carboxylic acid, 2,4-dichlorocinnamic acid, 2-chlorocinnamic
acid, l-naphthoxyacetic acid, and 4-chlorocinnamic acid. deA appeared to oxidize the
Olefin bond of benzofiiran-Z-carboxylate and the cinnamic acids to form the
corresponding epoxides.

To facility further studies on substrate binding, a model of deA was constructed
and fluorescence techniques to measure binding affinities were developed. A structural
model Of deA was determined based on the coordinates of the ~30 % identical TauD.
The intrinsic fluorescence Of deA was shown to be quenched by 50-85% upon addition
Of Fe(II) or OtKG, allowing determination of binding affinities. Trp256 was initially
suspected to be the reporter since it lies within 5 A of the metal and OtKG binding sites;
however, W256L and W256F variants had only negligible differences from wild-type

deA in terms of relative fluorescence intensity and ligand-based fluorescence

quenching. Because Trpl 95 was predicted to be quite distant (~15 A) from the active site,
some combination ofTrp113 and Trp248 is likely to serve as the reporter that senses
metal and cofactor binding to deA.

Based on the deA structural model, potential 2,4-D and OLKG ligands were
identified and altered by site-directed mutagenesis. His2l4, Lys7l , Arg278, and
backbone amide of Ser117 were hypothesized to bind the 2,4-D carboxylate, and Lys95,
and possibly Lys71, were hypothesized to interact with the 2,4-D ether atom.
Additionally, Arg274 and Thr141 were suspected to be OLKG ligands. Consistent with
their important roles as substrate ligands, variants of Lys7l, Arg274, and Arg278 had
aKG Km and Kd values and a 2,4—D Km value that were increased 1000-, 1000-, and 100-
fold, respectively. Evidence supporting an interaction between Lys95 and the 2,4-D ether
atom included the K95L variant being significantly less prone to inactivation by
phenylpropiolic acid and having a 4-fold lowered 2,4-dichlorocinnarnic acid Km. Thr141
plays only a minor role, if any, in interacting with OtKG; the T141V had OtKG Km and K.
values and a 2,4-D Km value similar to wild-type protein.

Studies were also conducted in an attempt to examine the substrate specificity of
deA-like proteins in environmental and pathogenic organisms. The zfdA-like gene in
Bordetella pertussis was cloned, but soluble, properly folded protein could not be
Obtained to analyze substrate specificity. Environmental tfdA-like genes proved difficult
tO clone due to loss of the gene in these organisms over time. During the course Of these
studies, the ubiquity of tfd4-like genes was extended tO aerobic environmental isolates

from soils frozen in the Siberian permafrost for upwards of 10,000 years.

Cepyright by
JULIE CHRISTINE DUNNING HOTOPP
2002

ACKNOWLEDGMENTS

First, I would like to thank Bob Hausinger for being a wonderful mentor and
scientific guide. I appreciate his dedication to his graduate students, his ever Open door,
his encouragement, his guidance, and his confidence in me as a graduate student. I also
want to thank my committee members for their advice and support: John Breznak, Leslie
Kuhn, Pat Oriel, and Greg Zeikus. I owe special thanks to Leslie Kuhn for her assistance
in modeling the structure of deA. I also owe a great deal of thanks to past and present
members of the lab: Scott Mulrooney and Jason Kuchar for their scientific expertise and
comic relief, Deb Hogan for being my graduate student guide, Aileen Soriano for being
my pipetteman buddy, Gerry Colpas for always being ready to answer my questions,
Kristin Huizinga for doing many genomic preps during her rotation, Raghavakaimal
Padmakumar for getting me started in the lab, and Ruth Saari, Matt Ryle, and Tim
Henshaw for scientific discussions.

I am so greatly indebted to Kerry Hotopp that I am at a loss for words. He is my
life guide and has been of immense support and assistance in every aspect Of life. For
three years, he did this all from three hundred miles away, a testament of his commitment
to us. Additionally, he helped in the writing Of this thesis with patent research and
proofreading. I would also like to thank Sandi Clemens, Stephanie Eichorst, Jennifer
Gray, Kristin Huizinga, Monica Ponder, and Kristina Stredwick for all their support and

advice about science and life. I would also like to give honorable mention to EF and

BtVS. And lastly, but not least, I want to thank my parents and sister who gave me much
love and support and taught me the value of education. They have always appreciated and

encouraged me to be obstinate, to persevere, and to always reach for my dreams.

vi

TABLE OF CONTENTS

List of Tables ....................................................................................................................... x
List Of Figures .................................................................................................................... xi
List of Schemes ................................................................................................................. xii
List of Equations .............................................................................................................. xiii
List of Symbols and Abbreviations .................................................................................. xiv
CHAPTER 1 INTRODUCTION ....................................................................................... l
OtKG dioxygenase superfamily ................................................................................ 2

deA and 2,4-D degradation .................................................................................... 8

The History of 2,4-D .................................................................................... 9

Degradation of 2,4-D ................................................................................. l3

Prevalence of OtKG dioxygenases related tO deA .................................... 16

Critical Open questions and specific aims .............................................................. 17

CHAPTER 2 ALTERNATIVE SUBSTRATES OF 2,4-DICHLOROPHENOXY-

ACETIC ACID/Ot-KETOGLUTARATE DIOXYGENASE ................................. 18
Introduction...........................................L ................................................................ 19
Experimental .......................................................................................................... 21
Purification and assay of deA .................................................................. 21
Kinetics Of non-phenoxyacetic acid substrates .......................................... 21
Identification of the 2,4-dichlorocinnamic acid and benzofuran-Z-
carboxylic acid metabolites ............................................................ 23
Assessing non-phenoxyacetic acid substrates with whole cells ................ 24
Computing curve fits .................................................................................. 24
Results .................................................................................................................... 25
Identification of non-phenoxyacetic acid substrates ................................. ' .25
Kinetic analysis of non-phenoxyacetic acid substrates .............................. 26
Metabolite analysis .................................................................................... 27
Whole cell studies ...................................................................................... 30
Discussion .............................................................................................................. 33
Expansion of the substrate range of deA ................................................. 33
The ancestral role of deA? ....................................................................... 34
Biocatalysis ................................................................................................ 35
Acknowledgements ......................................... ' ....................................................... 37

CHAPTER 3 MODELING OF deA AND USE OF INSTRINSIC TRYPTOPHAN

FLUORESCENCE AS A PROBE OF METAL AND Ot-KETOGLUTARATE
BINDING ............................................................................................................... 38

vii

Introduction ................................................................................................... A ......... 39

Experimental .......................................................................................................... 41
Modeling methods ..................................................................................... 41
Protein purification and kinetics ................................................................ 43
Site-directed mutagenesis .......................................................................... 43
Fluorescence measurements ....................................................................... 43
Calculation of binding constants ................................................................ 45

Results and Discussion .......................................................................................... 46
Modeling Of the deA structure ................................................................. 46
Intrinsic tryptophan fluorescence as a probe of metal and OLKG binding

to deA ........................................................................................... 49
Identification of tryptophan residues acting as reporter groups ................. 51

CHAPTER 4 PROBING THE 2,4-DICHLOROPHENOXYACETATE/Ot-KETOGLUT-
ARATE DIOXYGENASE SUBSTRATE-BINDING BY SITE-DIRECTED

MUTAGENESIS AND MECHANISM-BASED INACTIVATION .................... 57
Introduction ................................................................................................................
Experimental .......................................................................................................... 6O
Site-directed mutagenesis .......................................................................... 60
Protein purification methods ...................................................................... 60
Kinetic analyses ......................................................................................... 60
Inactivation kinetics ................................................................................... 62
Large-scale inactivation Of deA by PPA followed by alkylation and
protease digestion ........................................................................... 63
Separation and characterization of deA peptides ..................................... 64
Results and Discussion .......................................................................................... 65
Site-directed mutagenesis Of prOposed 2,4-D ligands ................................ 65
Site-directed mutagenesis Of proposed OtKG ligands ................................ 70
Arg78 ......................................................................................................... 71
Site-directed mutagenesis of tyrosines near the active site ........................ 71
Inactivation of deA by PPA ..................................................................... 72
CHAPTER 5 PHYSIOLOGICAL AND GENETIC CHARACTERIZATION OF
ENVIRONMENTAL ISOLATES CONTAINING tfdA-like SEQUENCES ........ 81
Introduction ............................................................................................................ 83
Experimental .......................................................................................................... 85
Media and cultures ..................................................................................... 85
Amplification of the tfdA-like gene ........................................................... 85
Cinnamic acid and ferulic acid utilization studies ..................................... 85
CO-metabolism of 2,4-D ............................................................................ 85
Results and Discussion .......................................................................................... 86
Prevalence Of tfdA-like genes in Siberian permafrost samples .................. 86
Assessment of cinnamic acid utilization by environmental isolates .......... 88
Cloning Of a fldA-like gene from an environmental isolate ....................... 88
What happened to the tfdA-like genes in LTER freezer stocks? ............... 91
The future for the LTER isolates ............................................................... 93

viii

CHAPTER 6 CLONING OF A tfdA-like SEQUENCE FROM Bordetella pertussis AND
PRELIMINARY CHARACTERIZATION OF THE CORRESPONDING

PROTEIN ............................................................................................................... 94
Introduction ............................................................................................................ 95
Experimental .......................................................................................................... 97
Phylogenetic tree of deA relatives ........................................................... 97

Cloning ....................................................................................................... 97

Production and purification Of the B. pertussis deA-like protein ............ 98

Assessment of substrate Specificity ........................................................... 99

Molecular weight determination .............................................................. 100

UV/vis spectrOSCOpy ................................................................................ 100

Results and Discussion ........................................................................................ 101
CHAPTER 7 CONCLUSIONS AND FUTURE DIRECTIONS ................................... 104
APPENDIX A DEVELOPMENT OF METHODS TO EVOLVE A tfdA-like GENE ...124
Introduction .......................................................................................................... 125
Experimental ........................................................................................................ 126

Media and cultures ................................................................................... 126

Random mutagenesis ............................................................................... 126

Screening .................................................................................................. 126

Activity analysis ....................................................................................... 127

Plate matings ............................................................................................ 127

Results and Discussion ........................................................................................ 129

ix

Table 1.1.

Table 1.2.

Table 2.1.

Table 3.1.

Table 3.2.

Table 3.3.

Table 3.4.

Table 3.5.

Table 4.1.

Table 4.2.

Table 4.3.

Table 5.1.

Table 5.2.

Table 5.3.

Table 5.4.

Table A].

LIST OF TABLES

Phenoxyacetic acid substrate specificity ........................................................... 10
Ot-Ketoacid substrate specificity ....................................................................... 10
Kinetics of alternative substrates of deA ........................................................ 34
Forward mutagenic primer sequences .............................................................. 44
Summary Of the metal and OLKG binding to wild-type deA ........................... 52
Summary of metal binding to wild-type deA ................................................. 52
2,4-D kinetics of variant deA proteins ............................................................ 54
OtKG kinetics Of variant proteins ...................................................................... 54
Forward mutagenic primer sequences .............................................................. 61
2,4-D kinetics of variant proteins ..................................................................... 67
OtKG Kinetics of variant proteins ..................................................................... 69
Characteristics of permafrost isolates tested ..................................................... 87
Growth Of LTER isolates on media amended with cinnamic acid ................... 89
Growth of LTER isolates on media amended with ferulic acid ....................... 89
Cinnamic acid and ferulic acid degradation by LTER isolate .......................... 90
Characteristics Of selected mutagenized clones ............................................. 130

LIST OF FIGURES

Figure 1.1. 2-His-l-carboxylate facial triad ......................................................................... 3
Figure 1.2. Jellyroll architecture .......................................................................................... 7
Figure 1.3. Pounds of 2,4-D used in the twentieth century ................................................ 12
Figure 1.4. Pathway for the degradation of 2,4-D and organization of the Ifd genes in
Ralstonia eutropha JMP134 (pJ P4) ....................................................................... 15
Figure 2.1. Ascorbic acid dependence of 2,4-dichlorocinnamic acid utilization by deA
................................................................................................................................ 28
Figure 2.2. pH stability and nucleophilic reactivity of the deA reaction product derived
fiom 2,4-dichlorocinnamic acid ............................................................................. 31
Figure 2.3. Degradation of cinnamic acids by intact cells ................................................. 32

Figure 3.1. Secondary structure based sequence alignment of CA8], deA, and TauD...42

Figure 3.2. The active site of deA .................................................................................... 47
Figure 3.3. Ramachandran plot of deA ............................................................................ 48
Figure 3.4. Fluorescence spectra of deA .......................................................................... 50
Figure 3.5. Predicted positions Of tryptophan residues in deA ........................................ 55
Figure 4.1. The active site residues in deA ...................................................................... 66
Figure 4.2. deA inactivation kinetics in the presence of PPA ......................................... 74
Figure 4.3. PPA competitive inhibition kinetics for K95L deA ...................................... 78
Figure 4.4. Peptide purification and identification ............................................................ 80

Figure 6.1. Phylogenetic tree of characterized and hypothetical proteins homologous to
deA ....................................................................................................................... 96

xi

LIST OF SCHEMES

Scheme 1.1. Reaction types carried out by OtKG-dependent dioxygenases ........................ 5

Scheme 4.1. Theoretical mechanism Of deA inactivation by PPA .................................. 75

xii

LIST OF EQUATIONS

Equation 2.1. Equation for kinetics Of irreversible inactivation Of deA .......................... 23
Equation 3.1. Equation for determining K, from protein fluorescence changes ............... 45
Equation 4.1. Equation for kinetics Of irreversible inactivation Of deA .......................... 62
Equation 4.2. Equation for substrate inhibition ................................................................. 62
Equation 4.3. Equation for determining Kd from protein fluorescence changes ............... 62
Equation 4.4. Equation for mechanism-based inactivation ............................................... 63

xiii

2,4,5-T
2,4-D
ACCO
OtKG
AtsK
CarC
CAS
DAOCS
EDTA
EpoA
[ET]
HPLC
IPNS
AF

AF max
[I]
Mina“)
k(inact)max
KI

[L
LC-ESI
LTER
MALDI
ORF
PPA

Pt

Pch
SDS-PAGE
SyrB
TauD
TE
TFA
deA
Tris

Vi

YSD

LIST OF SYMBOLS AND ABBREVIATIONS

2,4,5-Trichlorophenoxyacetic acid
2,4-Dichlorophenoxyacetic acid
1-Amino-l-cyclopropanecarboxylic acid oxidase
Ot-Ketoglutarate

Alkylsulfatase

Carbapenem biosynthesis enzyme

Clavaminate synthase

Deacetoxycephalosporin C synthase
Ethylenediaminetetraacetic acid

Fosfomycin biosynthesis enzyme

Total concentration Of enzyme subunit

High pressure liquid chromatography
Isopenicillin N synthase

Change in fluorescence

Maximal fluorescence change

Inactivator concentration

Inactivation rate

Inactivation rate at saturating [I]
Inactivation/Inhibition constant

Total concentration Of ligand

Liquid chromatography electrospray ionization
Long term ecological research

Matrix-assisted laser desorption ionization
Open reading frame

Phenylpropiolic acid

Product accumulated at time t

Pyoverdine biosynthesis enzyme

Sodium dodecylsulfate polyacrylamide gel electrophoresis
Syringomycin biosynthesis enzyme 1
Taurine/OLKG dioxygenase

10 mM Tris (pH 7.7) and 1 mM EDTA
Trifluoroacetic acid
2,4-Dichlorophenoxyacetic acid/OLKG dioxygenase
Tris(hydroxymethyl)aminomethane

Initial velocity

Sulfonate/OLKG dioxygenase

xiv

CHAPTER 1

INTRODUCTION

In order to provide a framework for describing my studies of various family
members, this introductory chapter begins with a general description of the Ot-
ketoglutarate (OLKG) dioxygenase superfamily. Most of my efions have focused on 2,4-
dichlorophenoxyacetate/a-KG dioxygenase (deA), an enzyme involved in metabolism
of the herbicide 2,4-dichlorophenoxyacetic acid (2,4-D). I will introduce background
relevant to this herbicide, comment on its degradative pathway, and emphasize the role Of
deA. I then summarize what is known about close relatives of deA in environmental
isolates and pathogenic bacteria. Finally, I emphasize the Open questions that led to this

work.
OtKG DIOXYGENASE SUPERFAMILY

OLKG dioxygenases are an important and understudied superfamily of enzymes.
Members of the superfamily are widely distributed among eukaryotes and eubacteria and
catalyze a diverse array Of reactions (78). All members Of the aKG dioxygenase
superfamily chelate one ferrous ion per catalytic site in a 2-His-1-carboxylate facial triad
(42) with three waters bound at the other adjacent coordination sites (Figure 1.1). For
most family members, OLKG displaces two waters to coordinate the iron via its C-l
carboxylate and C-2 keto group (Figure 1.1). In contrast, 1-amino-1-
cyclopropanecarboxylic acid oxidase (ACCO) and isopenicillin N synthase (IPNS) do not
require an Ot-ketoacid, but are placed in “the superfamily based on sequence similarity (77,
80). When used as a cosubstrate, OLKG is oxidatively decarboxylated in a reaction that is
coupled to the oxidation of the primary substrate. The most common oxidation Of the

primary substrate is the hydroxylation of an unactivated carbon, but desaturations,

Fig. 1.1. 2-His-l-carboxylate facial triad. All members of the OtKG dioxygenase
superfamily chelate iron in a 2-His-1-carboxylate facial triad (top). For most family
members, OLKG binds the iron in a bidentate fashion by displacement of two water
molecules (bottom).

:17
N
O

’2
m
H.
m

0 H 20,00”
0,.

CD

\

-----.--|-r-J------..-
N
+

,/

tn
*0

OH

N

His

0
O
S‘
S
’4,
m
I—s
(I)

epoxidations, or ring formation and expansion reactions can also result (Scheme 1.1). The
OS carboxylate Of OLKG forms a salt bridge with a basic side chain, either lysine or
arginine, and in some cases is further stabilized by hydrogen bonding with a threonine or
serine. In combination, the iron and OtKG ligands form an HX(D/E)XnHX(6-.2)(R/K)
conserved sequence motif in these proteins. Crystal structures of IPNS (80), clavarninate
synthase 1 (CAS)1 (101), deacetoxycephalosporin C synthase (DAOCS) (109), proline
hydroxylase (16), anthocyanidin synthase (106), and taurine/OtKG dioxygenase (TauD)
(25) reveal that the conserved residues are located on eight B-strands in a conserved
jellyroll motif (Figure 1.2).

Historically, OLKG dioxygenases were slow to be identified since OtKG is not
routinely added to assay mixtures and the enzymes show little overall sequence identity.
Recently, however, there has been a rapid expansion of potential family members
identified through genome sequencing projects based on homology to characterized OLKG
dioxygenases. For example, analysis of the genome ofArabidopsis thaliana reveals over
sixty-four Open reading frames that show strong similarity to known OLKG dioxygenases;
less than 20% have known functions (78). Other proteins of known function, but
unknown biochemical mechanism, have been identified as OLKG dioxygenases by using
new search engines available to detect proteins with similar conserved motifs in
sequences with little overall homology (e.g., PSI-BLAST). PSI-BLAST searches to
convergence followed by alignments based on predicted secondary structure recently
revealed that the extracellular matrix protein leprecan, the disease-resistance-related

protein EGL-9, and the DNA-repair protein AlkB are members Of the OtKG dioxygenase

Scheme 1.1. Examples of the types of reactions carried out by OLKG dioxygenases.
Both hydroxylation and desaturation reactions are carried out by clavarninate synthase 1
(CAS) (9, 63). An epoxidase reaction leads to the formation of fosfomycin in a reaction
carried out by EpOA (104). Ring formation and ring expansion reactions are catalyzed by
IPNS (14) and deacetoxycephalosporin C synthase (DAOCS) (10, ll).

Scheme 1.1.

Hydroxylation

CASI

N NH Fe(II) : N OH
oi m I f \ o) m NH
I1:1] N 02 succinate HOOC E4
Ot-KG C02

 

H2

Desaturation
CASl

....O Fe(II) ..uO NH;
We: 7 \ = («Eff

 

 

 

COOH NH2 0; H20 COOH
aKG succinate
C02
Epoxidation
EpoA H H
H H .. -
> < Fean 2,7: ,9
H3C HO’ P: / \ OHO’
OH _
02 succinate
Ot-KG C02
Ring Formation
HOOC
WE SH HOOC H
N \
HZN O I/ :‘Ple Hzm S .3“
0 ea ) : O N‘fi
H COOH
HOOC? 02 21.120
Ring Expansion DAOCS

Fe(II)
7* HOOC
HOOC / \
N
2 S \x ‘ O
O a-KG succinate N /
N\8\ C02 O

O' OH
COOH CO

 

:1:

Figure 1.2. Jellyroll architecture. Conserved residues involved in binding ferrous ion
and OLKG are located on eight B-strands in a conserved jellyroll motif. Images in this
thesis/dissertation are presented in color.

 

superfamily (5). These approaches set the stage for biochemical analyses to define the
catalytic activity of these proteins. In a separate study, PSI-BLAST was used to divide
the OLKG dioxygenase superfamily into three groups each with their own more specific
conserved motif. The major difference between the groups is the size of the insert
between the carboxylic acid containing amino acid and the second histidine (n = 50-70,
Group I; n = 138-207, Group II; n = 72-101, Group III) (47). Work in this thesis focuses
on Group II members related tO Ralstom'a eutropha JMP134 (pJP4) deA, Escherichia
coli TauD, Streptomyces clavuligerus CASl and CAS2, Pseudomonas Sp. 7-
butyrobetaine hydroxylase, Caenorhabditis elegans y-butyrobetaine hydroxylase,
Pseudomonas putida alkylsulfatase (AtsK), Saccharomyces cerevisiae sulfonate/OLKG
dioxygenase (YSD), Erwinia carotovora carbapenem biosynthesis enzyme C (CarC),
Pseudomonas aeruginosa pyoverdine biosynthesis enzyme (Pch), and Pseudomonas
syringae syringomycin biosynthesis enzyme 1 (SyrB).
deA AND 2,4-D DEGRADATION
The Ralstonia eutropha JMP134 (pJP4) 2,4-D/OLKG dioxygenase (deA) is a
homodimer with a calculated monomer molecular weight of 32 KDa that hydroxylates
2,4-D while converting OLKG to succinate and carbon diOxide (29).
OVCOO' 0 COO OH COO 00‘:
+02+ ? ___> + (I:HO + 2 +C02
C1 C1 C1
’OOC

C1
C00'

Hydroxylation Of 2,4-D at the C-2 side chain position yields an unstable hemiacetal that
decomposes into glyoxylate and 2,4-dichlorophenol, which is further metabolized by the

cell (29, 86). Although deA prefers OtKG and 2,4-D, it can use a wide variety Of Ot-

ketoacid and phenoxyacetic acid substrates (Table 1.1 and 1.2). The key residues in deA
associated with the motif shown above are His] 14, Asp116, Thrl41, Hi5263, and
Arg278. Site-directed mutagenesis studies confirmed His] 14, Hi5263 and Asp116 as iron
ligands (47). Additionally, evidence was obtained that Hi5214 participates in substrate
binding (47). Below, I discuss the history Of 2,4-D, the pathway for its degradation, and
homologues of this protein or its gene.

The History of 2, 4-D. In 1880, Charles Darwin hypothesized that the tip of a plant
influences the bending of the plant toward light, or phototropism (4). Research continued
on phototropism through the early 1900’s, and in 1934, KOgl and Haagen-Smit from the
Netherlands discovered indoleacetic acid (auxin) to be the principle natural plant growth
hormone (4). Indoleacetic acid was unstable outside of plants, so researchers began
synthesizing indoleacetic acid homologues and studying their effects on plant growth
activity (4). In 1940, 2,4-D was first synthesized from monochloroacetic acid and 2,4-
dichlorophenol and was tested as a fimgicide (4). 2,4-D was reported in 1942 to be 300
times more potent than indolebutyric acid, the major plant growth regulator at that time
(4). The original 2,4-D patent was filed in February 1942 and defined plant growth
hormones as compounds “containing as an essential active ingredient a monocarboxylic
acid having one of its carbon atoms linked to a nuclear halogenated aromatic ring by
means of a polyvalent, strongly negative, non-metallic atom, namely, oxygen, sulfur and
nitrogen” (64). The patent did not mention the usage of the compounds as herbicides
(64). Originally, the work was not focused on the idea Of using synthetic plant growth
hormones as herbicides, but instead on using them at low concentration as rooting agents

(79).

Table 1.1. Phenoxyacetic acid substrate specificity (30).

 

 

 

 

 

Substrate K... (uM) km (min’l) [rm/1?...(minJ-mMT)
Phenoxyacetate 460 i 23 443 i 10 960
2-Chlorophenoxyacetate 110 :t 6.5 380 i 11 3,450
4-Chlorophenoxyacetate 117 $6.2 595 d: 15 5,090
2,3-Dichlorophenoxyacetate 102 i 8.5 288 i 12 2,820
2,4-D 17.5 i 1.0 529 :t 16 30,200
3,4-Dichlorophenoxyacetate 219 3: 7.6 307 i 5 1,400
2,4,S-Trichlorophenoxyacetate 59.6 i 3.4 96 fl: 3 1,610
4-Chloro-2-methylphenoxyacetate 89.0 :t 8.7 233 :1: 12 2,620
DL-2-Phenoxypropionate 1170 d: 120 5.1 :i: 0.2 4.4
2-(2,4-Dich1orophenoxy)propionate 191 i 6.7 61 i 3 320
3-Phenoxypropionate 12900 i 3200 3.2 i 0.4 0.25
Table 1.2. Ot-Ketoacid substrate specificity (30).

Substrate K m (uM) km (min‘l) [cm/Km (min'I-mM')
OtKG 3.20 i 0.54 643 :I: 44 200,000
Ot-Ketoadipate 20.6 i 1.1 290 at 5 14,100
Pyruvate 1020 i 86 58 d: 2 60
Ot-Ketobutyrate 464 d: 61 89 :t 6 190
Ot-Ketovalerate 607 i 47 404 :t14 660
Ot-Ketocaproate 583 i 50 . 158 :1: 7 270
Ot-Ketoisovalerate 745 i 36 16 i 1 20

 

10

During this time, researchers were also trying to identify selective herbicides for
broadleaf weeds in grasses and grains (4). In 1941, after Observing that certain plant
growth hormones were phytotoxic, Kraus proposed that synthetic plant growth hormones
might work as herbicides (4). In 1942, after Kraus and other scientists convinced the
Secretary of War, H.L. Stimson, of the potential dangers of these selective herbicides
when used as weapons, the United States Army set up Camp Detrick for the purpose of
studying chemicals for biological warfare (4). The scientists at Camp Detrick pursued
using herbicides to destroy crops and as forest defoliants to reduce cover of World War 11
Japanese defense positions (4). The research included the evaluation of over one
thousand compounds, cytotoxicity effects of specific compounds including 2,4-D,
demonstration Of the control of weeds, demonstration of the specificity of 2,4-D to kill
broad-leaf plants, and identification of the effects of applications of herbicides to soil and
water. This research was published by the scientists at the end of World War II (4).
Herbicides were not used in chemical or biological warfare until the introduction of
Agent Orange (a mixture of 2,4,5-triclorophenoxyacetic acid (2,4,5-T) and 2,4-D) as a
defoliant is the Vietnam War.

The patent for the usage of 2,4-D and other plant grth hormones as herbicides
was assigned to American Chemical Paint Company in May 1945 (49). In June 1945,
American Chemical Paint Company began marketing 2,4—D under the brand name
Weedone, and by 1946, 2,4-D was in high demand. It continues to be used on crops today
(Figure 1.3) under the brand names Hi-Dep®, Weedar® 64, Weed RHAP A-4D®. At
low concentrations, 2,4-D regulates plant cell development leading to bending and

swelling, proliferation and overgrowth of leaves and stems, control Of root growth, and

11

Figure 1.3. Pounds of 2,4-D used in the twentieth century.

Usage of 2,4-D in the United States

 

 

 

 

6O
50.

6‘

<

S

5 40*

‘U

0

m

3

=3 30.

v.

N

b-

O

a 20.

fl

=

O

A.
10-
0. .
1900 1920 1940 1960 1980 2000

Years

12

the initiation Of new roots on Stems and leaves (64). At higher concentrations, 2,4-D acts
as a selective, post-emergence herbicide against broad-leaf plants (49). Because of these
properties, 2,4-D is used today on major field crops, rangeland, pastureland, alfalfa for
forage, turf grass, small grain production, weed control, and orchard, vineyard, soft fruit,
and nut crops (4). Approved uses of 2,4-D have not been scientifically documented to
cause any acute or chronic human health risks (4). In fact, a 2,4-D researcher revealed
having eaten 0.5 g of 2,4-D per day for three weeks with no ill effects (4). Due to its
many uses and low toxicity in humans and animals, 2,4-D is the most widely used
herbicide in the world today and the third most widely used in the United States (4).

Degradation of 2, 4-D. Early reports suggested that 2,4-D was rapidly
decomposed in soil (7). By the 1960’s organisms had been isolated that degrade 2,4-D by
a pathway involving 2,4-dichlorophenol (12, 92). Since then organisms from numerous
genera have been found to degrade 2,4-D including Acinetobacter, Arthrobacter,
Aspergillus, Bordetella, Bradyrhizobia, Burkholderia, Corynebacterium, Pseudomonas,
Ralstonia, Rhodoferax, Stenotrophomonas, Sphingomonas, Variovorax, and
Xanthobacter (20, 21, 26, 27, 38, 50-52, 55, 58, 65, 67, 96, 107).

The individual genes involved in 2,4-D metabolism (the tfd genes) and the
plasmids that typically carry them are often mobile. For example, the 2,4-D catabolic
genes can be transferred in the environment between different genera of soil bacteria (31,
66, 88, 99). Transposons carrying the {fd genes have been identified (13). Some plasmids
carrying the tfd genes can be transferred by conjugation (19, 72) and can be integrated

and excised from the chromosome (53, 96).

13

Numerous pathways exist for 2,4-D degradation ranging from various oxidative
schemes to completely anaerobic metabolism (40). The best-studied pathway is the one
encoded by the tfd genes on the pJP4 plasmid from R. eutropha JMP134 (74). The pJP4
plasmid is a 51 MDa, Inc P1 plasmid. It carries all the genes necessary for converting
2,4-D to chloromaleylacetate (Figure 1.4). Enzymes from chromosomally encoded genes
carry out the subsequent breakdown of chloromaleylacetate. The plasmid pJP4 also
confers resistance to merbromin, phenylmercury acetate, and mercuric ions (21), and it
confers the ability to degrade 3-chlorobenzoate (3 7), but not phenoxyacetate (21). Since
its discovery and characterization, the pJP4-encoded catabolic system has been used as a
model for the evolution and dispersal in the environment of catabolic pathways to
degrade xenobiotics.

Three lines of evidence suggest that separate sets of the individual tfd genes were
recruited onto the plasmid backbone to generate a single plasmid capable of conferring to
the organism the ability to degrade 2,4-D. First, tfdA and the IdeCDEF gene cluster are
individually regulated on pJP4 (61) (Figure 1.4). Second, a comparison of 2,4-D-
degradative plasmids containing homologous tfd genes Shows that the genes are uniquely
organized on unrelated plasmids (66). Finally, phylogenetic trees of the 168 rDNA, tfdA,
ide, and tde produce discrepancies (71 , 102). In other words, the tfd genes have not
evolved together in one organism or on one plasmid backbone. Additionally, tdeCDE
Operons may have predated the application of 2,4-D, since 2,4-dich1orophenol is
produced in the environment from phenol by the metabolic action of Penicillium Sp. (39).
Large-scale use of 2,4-D may have increased the prevalence of these genes rather than

selecting for a novel function.

14

Figure 1.4. Pathway for the degradation of 2,4-D and organization of the tfd genes in
Ratstonia cutropha JMP134 (pJP4).

Cl

I
Irmn.

OH
Cl
-OOC
c) / a TidEl
‘ // Cl
-00C '00C /

-00C
00-
\

deC.
—->

 

OH
CI

 

TrdBl
—->

 

OH
Cl

deA
—->

CI
-OOC
deFn y deF“
<— <——
-OOC

Cl
-OOC

 

,0

OOC
A.

15

Homologues of all the genes except tfdA have been identified that encode
enzymes with substrates similar to those in the pJP4-encoded pathway providing hints of
their possible recruitment origin. The IdeDE genes (encoding for dichlorocatechol
degradation) are closely related to the cchBD genes encoding for chlorocatechol
degradation (35, 36). Likewise, tde (encoding for 2,4-dichlorOphenol hydroxylase) is
homologous to a similar gene, pheA, which encodes a phenol hydroxylase (73).

Although no homologue of tfdA has been found to participate in decomposition of
non-chlorinated aromatic substrates, some evidence suggests that one may exist. Bacterial
isolates from pristine areas throughout the world have been shown to degrade 2,4-D,
albeit at a much slower rate than an organism carrying the canonical pJP4 plasmid (32).
One particular isolate from a pristine area in northern Saskatchewan, Canada, carries a
tfdA gene homologous to the one on pJP4 that can insert itself onto the pJP4 backbone
(55). Furthermore, bacterial isolates that possess the tfdA gene, but do not degrade 2,4-D,
have been isolated from agricultural plots that were not treated with 2,4-D (46). Lastly,
plasmid capture experiments with a recipient plasmid containing tdeCDEF, but not tfdA,
led to the identification of a plasmid containing only {721.4 (100).

Prevalence of deA homologues. Apparent homologues of tfdA are present in
approximately 30% of bacterial isolates obtained from agricultural plots. The role of such
tfdA-like genes is unknown, as these soil microorganisms do not degrade 2,4-D (46).
Similar sequences have been identified in several microorganisms including

Mycobacterium tuberculosis and Bordetella pertussis. These pathogenic bacteria would

16

not be expected to have been exposed to 2,4-D; thus, the tfdA-like genes in these ,
microbes likely function in a distinct role.
CRITICAL OPEN QUESTIONS AND SPECIFIC AIMS

The overriding question addressed by this dissertation is: what governs substrate
specificity in deA and related enzymes? In order to answer this question, experiments
were undertaken to examine the breadth of compounds that the pJP4-encoded deA can
use as a substrate (Chapter 2), to model the structure Of deA and develop fluorescence
methods to assess metal and OtKG binding (Chapter 3), and to test the model by site-
directed mutagenesis and analysis of substrate-based inactivation of deA by
phenylpropiolic acid (Chapter 4). In addition, attempts were made to compare the
substrate specificity of deA to the substrate Specificity of environmental or pathogen
deA-like proteins mentioned above. This involved characterizing environmental isolates
containing tfdA-like genes (Chapter 5), cloning and performing a preliminary
characterization of the protein encoded by a Bordetella pertussis q'dA-like gene (Chapter
6), and developing methods to evolve a tfdA-like gene into one encoding deA activity

(Appendix A).

17

CHAPTER 2
ALTERNATIVE SUBSTRATES OF 2,4-DICHLOROPHENOXYACETIC
ACID/Ot-KETOGLUTARATE DIOXYGENASE

This chapter is based on the published article of the same title: Dunning Hotopp, J .C. and
RP. Hausinger (2001) “Alternative substrates of 2,4-dichlorophenoxyacetate/a-
ketoglutarate dioxygenase.” J. Molec. Catalysis B: Enzymatic 15:155-62.

18

2,4-Dichlorophenoxyacetic acid (2,4-D), a broad leaf herbicide, is mineralized by
a wide variety of environmental isolates (40) including the best-studied example,
Ralstonia eutropha JMP134 (pJ P4) (formerly Alcaligenes eutrophus). The tfdABCDEF
genes, carried on the pJP4 plasmid in this microorganism (21), encode all of the enzymes
necessary for the degradation of 2,4-D to chloromaleylacetate. Chromosomally borne
genes encode enzymes that carry out the subsequent metabolism of chloromaleylacetate.
Since its discovery and characterization, pJP4 has been used to study the evolution and
environmental dispersal of the genes encoding a catabolic pathway for xenobiotic
degradation (40).

deA catalyzes the first step in this 2,4-D degradation pathway. This ferrous ion
and tat-ketoglutarate (OLKG) dependent dioxygenase hydroxylates 2,4-D while converting
OLKG to succinate plus C02 (29).

0 COO‘ O OH DOC
\/ coo‘ COO‘ 2
+ I + '1' C02
.4. 02 + ___.’. CHO
C1 C1 C1 Cl
'OOC

COO'

Hydroxylation of 2,4-D at the C-2 position yields an unstable hemiacetal that
decomposes into 2,4-dichlorophenol and glyoxylate (29). deA is related in sequence

(~3 0% amino acid identity) to a group of sulfonate degrading OtKG dioxygenases (24, 45)
and an alkyl sulfate ester degrading OtKG dioxygenase (54). In addition to these protein
homologues, nucleotide sequences closely homologous to tfdA are present in
approximately 30% of bacterial isolates obtained from agricultural plots. The role of the

tfdA-like genes containing these sequences is unknown since these soil microorganisms

do not degrade 2,4-D (46).

19

The current study explores alternative substrates for deA in order to better define
the substrate profile and to explore potential biotechnological uses for the enzyme. In
particular, we demonstrate that purified enzyme or whole cells can be used to produce
synthetically valuable epoxides (17, 34) of chlorinated cinnamic acids and benzofuran-Z-
carboxylic acid. Furthermore, since the 2,4-D-degrading enzyme may retain residual
activities for the natural substrate, this work may provide clues about the role of deA-

like proteins in nature.

20

EXPERIMENTAL

Purification and assay of deA. Escherichia coli DHSOt (pUS31 l) carries IfdA on
a pUC19-derived plasmid (29). These cells (4 L) were grown at 30 °C in LB medium for
16-20 hr, harvested by centrifugation, and suspended in 30 ml of TE buffer (10 mM Tris,
1 mM EDTA, pH 7.7) containing 1 mM phenylmethylsulfonyl fluoride and 10 ug/mL
leupeptin. Crude cell extracts were obtained by passing the cells twice through a pre-
cooled French pressure cell (16,000 psi) and clarifying the debris by centrifugation
(100,000 x g for 45 min) at 4 °C. After Splitting the sample into two pools, deA was

enriched by chromatography at 4 °C on a column of DEAE-Sepharose (2.5 x 19 cm)
using TE buffer and a 400 mL linear gradient to 200 mM NaCl. The enzyme eluted at
approximately 100'mM NaCl. deA-containing fractions were dialyzed in TE buffer and
chromatographed on a Mono Q column (HRlO/ l 0) (Pharmacia) at room temperature. The
enzyme eluted at about 40 mM NaCl when using a 100 mL linear gradient to 200 mM
NaCl in TE buffer. When necessary, deA was further purified by phenyl-Sepharose
chromatography. The sample was adjusted to 1 M in ammonium sulfate, applied to a
column (HRlO/ 10) equilibrated with TE buffer containing the same concentration of this
salt, and chromatographed with a 100 mL linear gradient fi'om l M to 0 M ammonium
sulfate. The enzyme eluted from this resin at approximately 350 mM salt. deA was
routinely assayed with 2,4-D as the substrate by using the previously described 4-
aminoantipyrene spectrophotometric assay (30). Protein concentrations were determined
by using a commercial protein assay (BioRad) with bovine serum albumin as a standard.
Kinetics ofnon-phenoxyacetic acid substrates. The activity Of deA toward non-

phenoxyacetic acid potential substrates was determined at 30 °C in 10 mM imidazole

21

buffer (pH 6.8) containing 1 mM OtKG, 50 UM (NI-I4)2Fe(SO4)2 , and 200 M ascorbic
acid by using one or more of four different assays. For all methods, stock solutions of
(NH4)2Fe(SO4)2 and ascorbic acid were made fresh prior to each set of experiments.

A YSI model 5300 biological oxygen monitor was used to determine rates of
oxygen consumption. The probe was equilibrated using air-saturated MilliQ water, and
conuOl runs included samples lacking in ascorbic acid, iron, enzyme, aKG, and substrate.

Samples analyzed by HPLC (Hewlett Packard 1050) were resolved by using a
Merck Lichrosorb RP-18 column. The peak areas or peak heights for samples absorbing
at 230, 254, and/or 280 nm were determined by using the Hewlett Packard ChemStation
software. Reaction mixtures containing 2,4-dichlorocinnamic acid were quenched with
NaOH (0.1 M final concentration), neutralized with HCl (0.1 N final concentration), and
analyzed by utilizing a mixture of 65:35:01 methanol:water:phosphoric acid (effluent A).
A standard curve for the reaction product was prepared by complete conversion of
selected concentrations of substrate. The area under the product peak was assumed to
correspond to the original concentration of substrate. Reaction mixtures that contained
naphthoxyacetic acid were quenched with EDTA (5 mM final concentration) and
analyzed by using a 50:50:01 mixture of methanol:water:phosphoric acid (effluent B) as
the mobile phase. Standard curves for the degradation of 1-naphthoxyacetic acid and 2-
naphthoxyacetic acid were prepared by using l-naphthol and 2-naphthol, respectively.
All other potential substrates tested were quenched with NaOH (0.1 M final
concentration) and neutralized with HCl (0.1 N final concentration). Standard curves for
the reaction products were created by correlating the loss of substrates with gain of

products at various concentrations of substrate. Effluent B was used as the mobile phase,

22

and the kinetics were determined in all cases by analyzing the gain of product; except, the
4-chlorocinnamic acid reaction kinetics were estimated from the loss of substrate.

Two spectrophotometric assays also were utilized. Analysis of 1-naphthoxyacetic
acid was carried out by using the 4-arninoantipyrene assay with l-naphthol as the
standard. For 2,4-dichlorocinnamic acid, an alternative spectrophotometric assay made
use of the 250 nm absorption associated with the conjugated system (8250 = 7800 M’lcm'
1). The loss of absorbance at this wavelength was monitored in studies designed to assess
the ascorbic acid dependence of the reaction.

As previously documented (30, 85), deA activity decreased over time by both
irreversible and ascorbate-reversible inactivation. To analyze the kinetics of non-
phenoxyacetic acid substrates, progress curves in the presence of 200 trM ascorbate were
analyzed by fitting the data to the following equation:

Pt = V, (1-e"‘“"‘°‘)‘)ktinact)" - Eq. 2.1
where P, is the accumulated product at time t, Vi is the initial velocity, and k(inact) is the
inactivation rate constant (85).

Identification of the 2, 4-dichlorocinnamic acid and benzofitran-Z-carboxylic acid
metabolites. Samples for NMR analysis were generated at 30 °C in D20 containing 40
mM phosphate buffer (pH 6.8), 200 UM OtKG, 100 uM substrate, 50 uM (NH4)2Fe(SO4)2,
50 uM ascorbic acid, and 1.6 uM deA. The metabolites generated from 2,4-
dichlorocinnamic acid and benzofuran-Z-carboxylic acid were analyzed by lH-NMR
using a Varian VXR 500 MHz NMR spectrometer.

The 2,4-dichlorocinnamic acid metabolite stability was assessed with a sample

prepared by enzymatic conversion of 10 umol of 2,4-dichlorocinnamic acid. The reaction

23

mixture was applied to a Pharmacia PEP-RPC HR10/10 column, washed with 0.1%
trifluoroacetic acid, and eluted by using a 40:60:0.1 mixture of
methanol:water:trifluoroacetic acid while monitoring the absorbance at 254 nm. Aliquots
of the sample were incubated for varying time periods using specified conditions, and the
concentrations of the remaining metabolite and the degradation products were measured
by HPLC analysis with effluent A.

Assessing non-phenoxyacetic acid substrates with whole cells. R. eutropha
JMP134 (pJP4) carries the pJP4 plasmid containing tfdABCDEF (21). R. eutropha
JMP228 (pBHSOlaE) is a derivative of R. eutropha JMP134 (pJP4) where the tfdA gene
has been interrupted by transposon mutagenesis (100). R. eutropha JMP228 is the strain
lacking the plasmid. These cells were grown to late-exponential phase at 30 °C in MMO
minimal medium (91) amended with the indicated carbon source. Cells were centrifuged
and resuspended (A260 = 1.5) in fresh MMO with no carbon source. The cell suspensions
were aerated on a stir plate. Carbon sources were added and samples were removed at
various times, centrifuged, diluted, and transferred to HPLC vials. Isocratic HPLC with
effluent A was used to analyze the samples for loss of the substrate peak and gain of the
product peak.

Computing curve fits. KaleidaGraph for Windows by Abelbeck Software was

used for computing all curve fits.

24

RESULTS

Identification of non-phenoxyacetic acid substrates. Prior studies have evaluated
the ability of ptuified deA to degrade various phenoxyacetic acids, thiophenoxyacetic
acids, and phenoxypropionic acids (30, 85, 86). The substrate associated with the highest
catalytic efficiency was identified as the xenobiotic compound, 2,4—D. Here, we
investigated a range of naturally occurring, non-phenoxyacid, aromatic compounds as
substrates of the enzyme, along with some related synthetic derivatives of these
compounds. In particular, cinnamic acids, auxin-like compounds, and naphthoxyacetic
acids were tested by using oxygen electrode, HPLC, and spectrOphotometric methods.

The oxygen electrode assay showed clearly enhanced levels of oxygen
consumption over background for deA (1.25 uM) assay mixtures containing 2-chloro, 4-
chloro, and 2,4-dichlorocinnamic acids, 1- and 2-naphthoxyacetic acids, and benzofuran-
2-carboxylic acid (each at 300 uM). When each enzyme-containing sample was adjusted
to 200 uM 2,4-D, oxygen consumption immediately increased to match that observed for
samples containing only 2,4-D. Thus, none of these compounds exhibited significant
inhibition or inactivation of 2,4-D hydroxylation activity by deA. For other compounds,
oxygen consumption rates were close to background levels. This assay did not directly
measure substrate conversion (e.g., some substrates may have uncoupled oxygen
consumption from substrate oxidation), and spurious results were Observed for at least
one sample: 3,4-dihydroxycinnamic acid. This compound formed a blue color when
mixed with ferrous ion (consistent with metal chelation by the catechol group) and
consumed oxygen in the absence of enzyme when ascorbic acid was present, presumably

similar to the non-enzymatic oxygen consumption observed for EDTA (not shown) (28).

25

Given these concerns, the potential substrates were quantitatively studied by using HPLC
and spectrophotometric methods.

HPLC or (for l-naphthoxyacetic acid) spectrophotometric methods provided
similar results for conversion of the potential substrates. Under conditions equivalent to
those used with the oxygen electrode, 2-chlorO-, 4-chloro-, and 2,4-dichlorocinnamic
acids, 1- and 2-naphthoxyacetic acids, and benzofuran-2-carboxylic acid were converted
to products. In addition, small and variable amounts of 3,4-dihydroxycinnamic acid
(caffeic acid), 3,5-dimethoxy-4-hydroxycinnamic acid (Sinapinic acid), 4-hydroxy-3-
methoxycinnamic acid (ferulic acid), 2,4-dimethoxycinnamic acid, 3,5-
dimethoxycinnarnic acid, unsubstituted cinnamic acid, 4-methoxycinnamic acid, and 3-
methoxycinnarnic acid were converted to products. NO indication of substrate loss or
product formation was observed for 2-hydroxycinnamic acid, 3-hydroxycinnamic acid, 4-
hydroxycinnamic acid (coumaric acid), or 2-methoxycinnamic acid. Similarly,
chromone-2-carboxylic acid, indole-2-carboxylic acid, indole-3-carboxylic acid, indole-
3-acetic acid (auxin), indole-3-acrylic acid, hippuric acid, and phenylpropionic acid were
not substrates of deA.

Kinetic analysis of non-phenoxyacetic acid substrates. Detailed studies of the
enzyme kinetics were carried out with the test compounds that were convincingly shown
to be substrates. Characterization of the kinetic parameters was complicated due to an
irreversible inactivation of the ferrous containing enzyme that occurs during exposure to
oxygen (85). An average k(inact) = 0.49 i: 0.26 min'1 was determined for all substrates,
regardless of the substrate concentration. Initial rates were obtained by fitting the

progress curves for substrate conversion, as described in the Materials and Methods. The

26

enzyme kinetics were further complicated by the dependence of the reaction on ascorbic
acid. As illustrated in Figure 2.1, the concentration of this reductant had a large effect on
the initial rate of 2,4-dichlorocinnamic acid utilization. Similar results were previously
reported for conversion Of the poor substrate, thiOphenoxyacetic acid (85). To overcome
this requirement, 200 uM ascorbic acid was used for all kinetics studies.

The results of the kinetic investigations are provided in Table 2.1. For
comparison, data were obtained for deA metabolism of 2,4-D. The values of Km, kw,
and [rm/Km (the catalytic efficiency) associated with 2,4-D were Similar to, and more
accurate than, previous reports (17.5 i 1.0 uM, 529 i 16 min", and 30,200 rnin'l uM'l,
respectively) (3 0). Each of the non-phenoxyacetic acid substrates was associated with a
higher Km, a lower kw, and a lower catalytic efficiency than was observed for 2,4-D.

Metabolite Analysis. lH-NMR analysis of the sample mixture after complete
transformation of 2,4-dichlorocinnamic acid by deA revealed (a) the disappearance of
two doublets at 5 6.4 ppm and 8 7.6 ppm associated with protons bound to carbon atoms
participating in the Olefin bond and (b) the appearance of two new doublets at 6 3.5 ppm
and 5 4.2 ppm (data not shown). The epoxide protons of phenylglycidic acid (the epoxide
of cinnamic acid) exhibit identical resonances (18); thus, we conclude that deA oxidized
the side chain double bond of 2,4-dichlorocinnamic acid to produce 2,4-
dichlorophenylglycidic acid. In the case of benzofuran-Z-carboxylic acid, NMR evidence
revealed an analogous change in chemical shift of the proton on carbon-3 from 8 7.2 ppm
to 5 3.6 ppm (data not Shown). This result was consistent with similar formation of an

epoxide during substrate conversion.

27

Figure 2.1. Ascorbic acid dependence of 2,4-dichlorocinnamic acid utilization. by
deA. The effect of ascorbic acid concentration on the initial rate Of 2,4-
dichlorocinnamic acid loss was monitored by using a continuous spectrophotometric
assay. The standard assay buffer contained 391 nM deA dimer and 100 uM 2,4-
dichlorocinnamic acid.

 

 

 

l6
14
12
'7: 10
.§ 8
.3 6
«a 4
2

0 1 l

 

 

20 40 60 80 100
[Ascorbate], uM

28

Table 2.1. Kinetics of alternative substrates of deA.

 

 

Substrate K... (0M) k... (mini) kart/Km (min'i-mM’)
2,4-D 20.0 i 5.4 1020 1|: 90 51,000
2-Naphthoxyacetic acid 134 i 32 263 i 5 1960

2,4-Dichlorocinnamic acid 190 :1: 56 52.5 :t 5.1 276
Benzofuran-Z-carboxylic acid 254 i 120 105 i 16 413

2-Chlorocinnamic acid 264 :l: 52 22.1 :h 1.3 83.9
l-Naphthoxyacetic acid 622 :t 310 15.7 i 2.9 25.2
4-Chlorocinnamic acid >900 >38

 

29

To further test whether the product of the 2,4—dichlorocinnamic acid is 2,4-
dichlorophenylglycidic acid, we examined its pH stability and nucleophilic reactivity.
Acidic conditions (pH 3) led to the rapid decomposition of the sample as two degradation
products were formed (Figure 2.2, panel A). In contrast, the reaction product was stable
for at least 8 days at pH 13 (data not shown). Inclusion of the strong nucleophile
hydroxylarnine at low pH led to the formation of two new degradation products (Figure
2.2, panel B). The low pH reactivity and distinct product profile generated in the presence
of hydroxylarnine are consistent with an epoxide being the initial product of 2,4-
dichlorocinnamic acid metabolism by deA.

Whole cell studies. The potential of using intact bacterial cells to metabolize 2,4-
dichlorocinnamic acid was examined using R. eutropha JMP134 (pJP4), R. eutropha
JMPZZ8 (pBHSOlaE), and R. eutropha JMP228 (Figure 2.3). The first strain encodes the
entire 2,4-D pathway, whereas the second and third cultures lack tfdA and the pJP4
plasmid, respectively. Degradation of 2,4-dichlorocinnamic acid was observed only in the
first isolate, consistent with a requirement for tfdA (panel A). Notably, product
accumulation mirrored substrate disappearance in agreement with the inability of this
strain to grow on 2,4-dichlorocinnarnic acid as a sole carbon source. In contrast to the
2,4-dichlorocinnamic acid results, all three strains decomposed unsubstituted cinnamic
acid (panel B). These results demonstrate that a cinnamic acid degrading system exists in

the cells and does not involve tfdA or other loci on the pJP4 plasmid.

Figure 2.2. pH stability and nucleophilic reactivity of the deA reaction product
derived from 2,4-dichlorocinnamic acid. The peak areas at 230 nm were determined for
the 2,4-dichlorocinnamic acid reaction product (0) and degradation products associated
with retention times of 2.5 min (O), 4 min (I), 6 min (A), and 7 min (X) of samples
incubated at pH 3 (A) and pH 4 with 2 mM hydroxylamine (B).

 

4 I I I

 

Relative HPLC Peak Area

 

 

 

      

 

 

4
0 50 . 100 .150 200
T1me,mm

31

Figure 2.3. Degradation of cinnamic acids by intact cells. The time dependence was
examined for decomposition of 2,4-dichlorocinnamic acid (panel A) and cinnamic acid
(panel B) by R eutropha JMP134 (pJP4) (a), R eutropha JMP228 (I), and R eutropha
JMP228 (pBHSOlaE) (o) by using HPLC methods and analysis of the resulting
absorbances at 230 nm. In addition, the time dependence was determined for production
of the metabolite derived from 2,4-dichlorocinnamic acid by R. eutropha JMP134 (pJP4)
(A) in panel A.

14
12
10

 

 

 

 

HPLC Peak Height

 

 

 

o 50 100 150 200
Time, min

32

DISCUSSION

Expansion of the known substrate range of T fdA. We demonstrated that several
substituted cinnamic acids are utilized (albeit poorly) as substrates by deA. The best of
this group of substrates is 2,4-dichlorocinnamic acid, whereas the 2- or 4-monochloro
derivatives are used with less efficiency and the unsubstituted cinnarrric acid is
metabolized even more poorly. This pattern of substrate preference parallels the situation
for phenoxyacetic acids, where the 2,4-dichloro species is preferred over the
monochlorinated derivatives, which are more readily metabolized than the unsubstituted
species (30). Based on NMR evidence and stability/reactivity studies of the 2,4-
dichlorocinnamic acid metabolite, we propose that the deA-generated products arising
from the cinnamic acids are the Side chain epoxides. Thus, deA is not limited to
inserting oxygen into unactivated C-H bonds—it also can catalyze the easier oxidation of
a C-C double bond.

Because 2,4-D cherrrically mimics auxin in its action as an herbicide (33), we
tested whether the converse situation may occur; i.e., can auxin or auxin-like compounds
chemically mimic 2,4-D and serve as a substrate for deA? We showed that the natural
plant hormone, indole-3-acetic acid, is not a substrate of the enzyme. Furthermore, we
demonstrated that indole-3-carboxylate, indole-3-acrylate, and indole-2-carboxylate also
are not transformed by the enzyme. Interestingly, however, benzofuran-Z-carboxylate
(the analogue of indole-2-carboxylate in which the indole N is replaced by 0) does serve
as a substrate of deA. Perhaps related to this reaction, benzofuran-2-carboxylate mimics
phenoxyacetic acid with the side chain linked to the aromatic ring via a methenyl carbon.

NMR evidence suggests that the double bond of this substrate is converted to an epoxide

33

by deA. The ability of deA to selectively catalyze the transformation of benzofuran-Z-
carboxylate while not oxidizing indole-2-carboxylate may relate to the differences in
resonance energies between the pyrrole (22 kcal/mol) and furan (16 kcal/mol) rings (23).
Thus, oxygen insertion into the double bond of benzofuran-Z-carboxylate is more
favorable than for the indole-2-carboxylate. Another compound related in structure to
both benzofuran-Z-carboxylate and phenoxyacetic acid is chromone-2-carboxylate, which
is not a substrate of deA.

1- and 2-Naphthoxyacetic acids, a third class of test compounds, were Shown to
be substrates of the enzyme. These compounds closely resemble phenoxyacetic acid, but
contain a larger aromatic ring. The fact that deA is capable of transforming these
compounds demonstrates that the active site is sufficiently large to allow entry of these
species. Of interest, 2-naphthoxyacetic acid is a synthetic auxin predecessor to 2,4-D that
was used as a rooting agent. It was never used widely as an herbicide since less 2,4-D
could be applied to achieve the same benefits (79).

The ancestral role of T fdA? Soon after 2,4-D was introduced into the environment
in the early 19405, reports suggested that 2,4-D was rapidly decomposed in soil (e. g.,
(7)). To account for decomposition of this xenobiotic compound, we earlier proposed the
existence of an ancestral gene encoding a degradative enzyme with greatest specificity
toward a natural product, but capable of utilizing 2,4-D with low efficiency (40).
According to this scenario, mutations arose over time to enhance the specificity of the
ancestral enzyme toward 2,4-D and the new gene recombined with genes for 2,4-
dichlorophenol degradation to create a 2,4-D degradation pathway. In support of this

hypothesis, approximately a third of soil bacterial isolates possess a tfdA-like gene, but do

34

not measurably degrade 2,4-D (46). Thus, environmental isolates possess tfdA-like genes
that appear to play another, yet unidentified, role besides 2,4-D catabolisrn. Based on the
studies reported, we propose that the ancestral role of deA may have involved
degradation of ciruramic acids. Ring-substituted cinnamic acids are widely distributed in
nature and are present in sizable amounts. For example, 4-hydroxycinnamic acid is found
at a concentration of 8.7 mg carbon/ g carbon in the top meter of peat soil (56). Plants
synthesize large amounts of cinnamic acids by phenylalanine deamination, and they
catalyze a variety of ring substitution, condensation, degradation, reduction, or
conjugation reactions to produce a host of cinnamic acid-related compounds including
flavonoids, hydroxybenzoates, and lignin (93). Degradation of substituted cinnamic acids
occurs by several pathways (for a review, see (82)), but it would not be surprising to learn
that another pathway remains to be described. Future studies will test the cinnamic acid
degrading abilities of the enzyme products from tfdA -like genes in environmental
isolates.

Biocatalysis. Because of the wide range of enantiospecific chemical reactions in
which epoxides can participate, epoxides serve as intermediates in a variety of medically
important chemical syntheses (6). Our demonstration that purified deA enzyme and
whole cells containing this enzyme can convert chlorinated cinnamic acids to the
corresponding epoxides expands the repertoire of available synthetic building blocks. In
particular, phenylglycidic acid, the epoxide of unsubstituted cinnamic acid, has been used
as an intermediate in the production of taxol (17) and (25,3S)-diltiazem (34). The
bioconversion of the chlorinated cinnamic acids enables the synthesis of several drug

analogues containing novel functional groups. Additionally, the use of deA may allow

35

for the production of enantiopure epoxides. deA has been shown previously to
hydroxylate only (S)-dichloroprop, a phenoxypropionate herbicide (86). Additionally, a
fellow member of the superfamily of Ot-KG dependent dioxygenases, EpoA, catalyzes the
enantiomer specific production of fosfomycin ((—)-cis-1,2-epoxypropylphosphonic acid)

from cis-propenylphosphonic acid in Penicillium decumbens (104).

36

ACKNOWLEDGEMENTS
I would like to thank Raghavakaimal Padmakumar for initiating several of these studies

and performing NMR analysis of the benzofuran-Z-carboxylate product.

37

CHAPTER 3

MODELING OF deA AND USE OF INSTRINSIC TRYPTOPHAN

FLUORESCENCE AS A PROBE OF METAL AND Ot-KETOGLUTARATE
BINDING

This chapter includes my contributions to two submitted papers.

The modeling of deA is presented in: Elkins, J. M., Ryle, M. J., Clifton, I. J ., Dunning
Hotopp, J. C., Lloyd, J. S., Burzlaff, N. 1., Baldwin, J. E., Hausinger, R. P., and P. L.
Roach (2002) “X-Ray crystal Structure of Escherichia coli taurine/a-ketoglutarate
dioxygenase complexed to ferrous iron and substrates,” Biochemistry, in press.

The intrinsic tryptophan fluorescence is presented in: Dunning Hotopp, J. C., Auchtung,
T. A., Hogan, D. A., and R. P. Hausinger (2002) “Intrinsic tryptophan fluorescence as a
probe of metal and Ot-ketoglutarate binding to deA, a mononuclear non-heme iron
dioxygenase,” Journal of Inorganic Biochemistry, in press. My contributions to this paper
include pmification and characterization Of the tryptophan mutants and measruing the
dissociation constants for the various metals tested. They do not include designing or
synthesizing the plasmids containing the tryptophan mutants.

38

Ot-Ketoglutarate (OtKG) dioxygenases comprise an important yet understudied
superfamily of enzymes. Members of the superfamily are widely distributed among
eukaryotes and eubacteria where they catalyze a diverse array of reactions (78). OtKG
dioxygenases chelate one ferrous ion and bind one OtKG per catalytic site. Substrate and
oxygen then bind to the enzyme and the oxidative decarboxylation of OtKG is coupled to
the oxidation of the primary substrate. The most common oxidation of the primary
substrate involves hydroxylation of an unactivated carbon, but desaturations,
epoxidations, or ring formation and expansion reactions can also result (Scheme 1.1).

The Studies described here focus on 2,4-dichlorophenoxyacetic acid (2,4-D)/OLKG
dioxygenase (deA), which hydroxylates the herbicide 2,4-D producing a hemi-acetal
that spontaneously decomposes to 2,4-dichlorophenol.

OVCOO- 0 COO' OH COO' DOC

r1 53 DE 2

c1 c1 Cl CI COO
'OOC

A model of deA was develOped based on the crystal structure of two homologues,
taurine/OLKG dioxygenase (TauD) (25) and clavarninate synthase 1 (CA8) (63). The latter
proteins are 30 % and 11.5 % identical in sequence, respectively, to the target protein.
deA is particularly well characterized by biochemical and biophysical methods (e. g.,
(30, 43, 47, 62, 85, 86, 105)) allowing for substantiation of the model. In addition, a
simple method was developed to rapidly determine metal and OtKG binding affinities SO
as to circumvent concerns about enzyme metallocenter oxidation. The intrinsic

tryptophan fluorescence associated with deA was Shown to be quenched by these

components, providing a useful method to measure their binding affinities. Site-directed

39

mutagenesis methods were used to demonstrate that the tryptophan residue predicted to
lie closest to the metal- and OtKG-binding sites in a model generated of deA does not

serve as the fluorescent reporter.

4O

EXPERIMENTAL

Modeling Methods. The PHD predicted deA secondary structure elements (83)
were aligned with the crystallographically determined secondary structures of TauD
(PDB ID. lGQW) (25) and CASl (PDB ID. lDSl) (109) (Figure 3.1). The deA
sequence was modeled onto the TauD chain A and CASl crystal structures ignoring iron,
OLKG, and primary substrates using Modeller4 (87). deA residues for which there were
not coordinates of an equivalent residue in TauD were masked. From the initial model,
the loop refinement method was used to correct an Ot-helix distortion near residues 34-36
and to remove six disallowed phi/psi angles. This method was also used in efforts to
remodel the loop formed by residues 86-110 which contained a knot. Numerous possible
models were generated for this region, including the additional criterion that residues 96-
103 form an Ot-helix as suggested by PHD secondary structure prediction. No model was
convincingly superior to the others and, since this region of deA contains a sixteen
amino acid insertion when compared to TauD, residues 86-111 were left out of the
model. Insertion of OLKG and iron into the deA model was based on alignment of the 2-
His-l-carboxylate facial triads of deA and TauD, with direct substitution of the OLKG
and iron positions. The residues interacting with the C-5 carboxylate of OLKG (Thr14l
and Arg274) were adjusted to optimize hydrogen bonding and ionic interactions, based
on the CASl and TauD coordinates. 2,4-D was manually positioned into the modeled
deA active site using InsightII (Molecular Simulations Inc.). As an initial constraint, the
0. carbon of the 2,4-D Side chain and the carbon adjacent to the sulfur Of taurine (i.e., the
sites of hydroxylation) were superimposed. Since deA selectively oxidizes the pro-R

hydrogen of 2,4-D (86), this hydrogen was positioned to face the iron active site. The

41

Figure 3.1. Secondary structure based sequence alignment of CASl, deA, and
TauD. An initial aligmnent was produced based on a 3D alignment of CA8] and TauD.
deA was then aligned based on a PSI-BLAST to convergence and its predicted
secondary structure elements to the existing alignment. The amino acids corresponding to
those in TauD secondary structure elements and the homologous residues in CA8] and
deA are colored red for helix, green for B-strands of the jellyroll motif and blue for other
B-Strands. Conserved Fe(II)-binding ligands are marked with black arrows.

deA --MSVVANPLHPLFAAGVEDIDLREALG -------------------------------------
TauD MSERLSITPLGPYIGAQISGADLTRPLS -------------------------------------
CASl -------------- MTSVDC -------- TAYGPELRALAARLPRTPRADLYAFLDAAHTAAASLP
deA STEVREIERLMDEK--—SVLVFRGQPL ----------------------- SQDQQIAFARNFGPL
TauD DNQFEQLYHAVLRH--—QVVFLRDQAI ----------------------- TPQQQRALAQRFGEL

CASI GALATALDTFNAEGSEDGHLLLRGLPVEADADLPTTPSSTPAPEDRSLLTMEAMLGLVGRRLGLH

deA EGGFIKVNQRPSRF ------- KYAELADISNVSLDGKVAQRDAREVVGNEANQLWHSDSSFQQPA
TauD HIHPVYPHAEGVD -------- EIIVLDTHNDNPPDND ---------------- NWHTDVTFIETP
CASl TG ----------- YRELRSGTVYHDVYP --------- SPGAHHLSSETSETLLEERTENAYHRLQ
deA ARYSMLSAVVVPPS----GGDTEFCDMRAAYDALPRDLQSELEGLRAEHYALNS----RFLLGDT
TauD PAGAILAAKELPST----GGDTLWTSGIAAYEALSVPFRQLLSGLRAEHDFRKSFPEYKYRKTEE

CASl PNYVMLACSRA---DHERTAATLVASVRKALPLLDERTRARLLDRRMPCCV --------------

deA D---YSEAQRNAMPPVNWPLVRTHAGSG---RKFLFIGAHASH--VEGLPVAEGRMLLAELLEHA
TauD EHQRWREAVAKNP-PLLHPVVRTHPVSG---KQALEVNEGFTTRIVDVS-EKESEALLSFLFAHI

CASl -DVAFRGGVDDPGAIAQVKPLYG ----- DADDPFLGYDRELLA——~--PEDPADKEAVAALSKAL
deA TQREFVYRHRWNVGDLVMWDNRCVLHRGRRYDISA --------- RRELRRATTLDDAVV ------
TauD TKPEFQVRWRWQPNDIAIWDNRVTQHYANADYLPQ --------- RRIMHRATILGDKPFYRAG--
CAST D—-EVTEAVYLEPGDLLIVDNFRTTiART ------ PFSPRWDGKDRWLHRVYIRTDR--NGQLSG
deA --------------
TauD --------------

CASl GERAGDVVAFTPRG

42

substrate carboxylate was rotated to allow favorable interactions with Arg278, Lys7],
Hile4 and the backbone amide of Ser117.

Protein purification and kinetics. Wild-type and variant forms of deA were
purified from Escherichia coli DHSOL as previously described (Chapter 2). Protein
concentrations were assessed using a commercial protein assay with standard curves
determined with bovine serum albumin. Activity measurements were made using the 4-
aminoantipyrene assay (30). Curve fits were computed using KaleidaGraph for Windows
by Abelbeck Software.

Site-directed mutagenesis. The amino acid residue numbering scheme used here is
based on that derived from the gene sequence rather than that of the purified protein, and
differs by the addition of one residue compared to that used previously (47, 62). Trp113,
Trp195, Trp248, and Trp256 deA variants were created using the Stratagene
QuikChange Mutagenesis System, the forward mutagenic primers listed in Table 3.1,
their complements, and pUS3ll (30) as the starting plasmid. This pUCl9 derivative
contains the Ralstonia eutropha JMP134 (pJP4) tfdA gene (94).

Fluorescence Measurements. The intrinsic tryptophan fluorescence of 0.25 uM
deA dimer in 25 mM MOPS, pH 6.75, was examined at room temperature (20 °C) by
using a Hitachi F-4500 spectrofluorometer. The excitation wavelength typically was
maintained at 290 nm (5 nm slit width) while monitoring emission from 300 to 360 nm (5
nm slit width). OtKG was prepared in the same MOPS buffer as the protein, whereas
metals were freshly dissolved in water. Cations were supplied as Fe(NI-I4)2(SO4), ZnSO4,

CuSO4, C0304, MnSO4, MgSO4, CaClz, LiCl, LiZSO4, NaCl, N32804, KCl, and K2804.

43

Table 3.1. Forward mutagenic primer sequences.

 

 

Mutant Forward Primer

WI 13?7 5'-GAA CCA OCT CTT CCA CAO CGA CAG C-3'

W113L 5'-GAA CCA OCT CTT GCA CAO COA CAO C-3'

W195F 5'-CGC COO TCA ACT TOC CGC TOO TTC G-3'

W195L 5'-CGC COO TCA ACT TCC COC TOO TTC G-3'

W248F 5'-CGT OTA CCG OCA TCO CTT CAA COT OOO AGA TCT GG-3'
W248L 5'-CCG GCA TCO CTT GAA COT OOO AGA "TC-3'

W256F 5'-GAT CTO OTO ATO TTC GAC AAC COC TGC-3'

W256L 5'-GAT CTO OTO ATO TTO OAC AAC CGC TGC-3'

 

Calculation of binding constants. Protein fluorescence changes accompanying the
binding of metal or OtKG were analyzed by curve fitting (KaleidaGraph) according to

equation 3.1.

 

AF = Arm, - Kd +1117]+[Er1-J<K;[;I§TI+IETI>-4ILTIIETI

 

Eq. 3.1

In this equation, the Observed change in fluorescence (AF) is related to the maximal
fluorescence change (AF max) extrapolated to infinite concentration of titrant. Based on the

known total concentrations of enzyme subunit ([ET]) and ligand ([Lfl), the Kd and AFM

were calculated.

45

RESULTS AND DISCUSSION

Modeling of the deA structure. The three-dirnensional structure of deA was
modeled and substrates were positioned at its active site, as illustrated in Figure 3.2 (PDB
ID. lGQX). Analysis of the deA model by Procheck gave an acceptable overall average
G factor of —O.22 and no residues in disallowed regions (Figure 3.3). The OtKG chelates
the metal center and is predicted to form a salt bridge to Arg274 and a hydrogen bond to
Thr14l. Interactions with the 2,4-D carboxylic acid are proposed to involve Arg278,
Lys7], Hile4 and the backbone amide of Ser117. Two of these interactions, Arg270 and
the N-H of Va1102, correspond to the taurine sulfonate-binding residues of TauD. Gly67
in deA replaces the third sulfonate-binding residue, His70, whereas deA residues
Hile4 and Lys7] do not have Similar counterparts in TauD. In addition to the potential
interaction with the carboxylate, deA Lys7] could reasonably interact with the ether
oxygen of 2,4-D. Additional interactions with the substrate are certain to involve residues
86-111 which could not be convincingly modeled. However, some potential models
suggested that Lys95' could orient to hydrogen bond with the ether oxygen of 2,4-D.

deA is particularly well characterized by biochemical and biOphysical methods
(e.g., (30, 43, 47, 62, 85, 86, 105)) allowing for substantiation of the model. The deA
model is consistent with previous observations and predictions made about deA. For
example, Arg78 is surface exposed in the model and accounts for the reported protease
sensitivity of this site (30). In addition, the model supports previous predictions based on
site-directed mutagenesis and Spectroscopic analyses that His] 14, Asp116, and Hi5263
bind iron (47). The model vindicates a prior suggestion that Arg274 forms a salt bridge

with the C-5 carboxylate of OtKG (47). Thrl4l is positioned to hydrogen bond this

46

Figure 3.2. The active site of deA. The crystal structures of TauD and CASl were used
to model the homologous protein deA. The residues believed to be involved in substrate
binding, and for which variants were made in these studies, are illustrated.

§ 2,4-D

Aspll6

Hi8214
m

k 1—

Hi3263

Figure 3.3. Ramachandran plot of the deA model. A Ramachandran plot of the deA
model was generated using Procheck. Of the 258 residues presented in the model, no
residues were found in disallowed regions and only 1.8% of residues were found in
generously allowed regions. Of the remaining 98.2%, 19.2% were found in additional
allowed regions and 79.0% were found in the most favored regions.

180

 

135

90

 

45

Psi (degrees)
0

-l35

 

 

-180 .7 ., ‘ I. ‘ 1' ‘ . «a .-..
-180 -135 -45 0 135 180

Phi (degrees)

48

moiety of OLKG, consistent with conservation of a hydroxylated amino acid at that
position in Group II dioxygenases. The model is in agreement with previous results
suggesting that His2l4 functions in 2,4-D binding or catalysis (47). Specifically, Hi5214
interaction with a carboxylate oxygen of 2,4-D accounts for the lO-fold increase in Km of
2,4-D in a H214A mutant. Although substitution of Hi3217 by Ala was also shown to
lead to a 2.5-fold increase in Km of 2,4-D (47), this residue is more distant fiom the active
site and is predicted not to directly bind substrate or participate in catalysis. The Observed
effects may alternatively arise from reduced substrate access to the active Site or to
altered positioning of a 2,4-D ligand. Finally, evidence for an interaction with the
substrate ether atom, presumed to involve Lys71, but possibly also Lys95, is provided by
studies involving the substrate 2,4-dichlorocinnamic acid (Chapter 2). This compound
closely resembles 2,4-D in structure, but lacks the ether bond leading to a much higher
Km.

Intrinsic Ttyptophan Fluorescence as a Probe of Metal and aKG Binding to
deA. The tryptophan fluorescence spectrum of deA, a representative OtKG-dependent
dioxygenase, is quenched by addition of Fe(II), OtKG, or both compounds (Figure 3.4).
The presence Of saturating concentrations (200 uM) of either F e(II) or OtKG leads to a 50-
85% reduction in fluorescence compared to that of deA apoprotein. When both
components are added, the fluorescence is further diminished to relative intensities that
are ~10% of the initial value. Additions of 200 uM Fe(II) and/or OtKG to 5 pM
tryptophan in water do not cause decreases in fluorescence. The binding of the 2,4-D to
deA was not examined by this approach because this substrate absorbs appreciably at

the excitation wavelength, thus interfering with the analysis. The fluorescence changes in

49

Figure 3.4. Fluorescence spectra of deA. The fluorescence Spectra of deA were
obtained in MOPS buffer alone (solid line), or in MOPS buffer containing 200 uM Fe(II)
or 200 pM OtKG (dashed lines), and with both compounds (dotted line).

 

p—I
O
0

1—1

00
O
r‘r-r-‘fr I ‘T—Y"T“T“T T’

[IIJLIIIII

      
 

l l

 

 

Relative Fluorescence
O\
O

 

 

300 310 320 330 340 350 360
Wavelength

50

deA likely arise from changes in the environment of one or more of the four tryptophan
residues in deA. .
Fluorescence titration studies were used to directly determine the Kd for Fe(II) and
OtKG binding to deA (Table 3.2). The 3.35 1.1M Kd of OtKG calculated by this approach
closely matches the reported OtKG Km values of 3.2 pM and 4.9 pM based on kinetic
studies (30, 47). Similarly, the 7.45 uM Kd of F e(II) determined here approximates the
concentration (~2.0 uM) shown to give rise to half-maximal activity (47). The presence
of 200 uM OtKG decreased the apparent affinity for Fe(II), perhaps due to partial
chelation of the metal ion by the Ot-keto acid. In contrast, the presence of 200 uM Fe(II)
had essentially no effect on the apparent affinity for OLKG (Table 3.2). The dications of
Mg, Ca, Co, Cu, Mn, and Zn bound to deA with Kd values Similar to that observed for
Fe(II) (Table 3.3), whereas monovalent cations had negligible effect. The Similarities in
Kd values for the dications do not correspond to their ability to inhibit deA; e.g., Cu
exhibits competitive inhibition with K; ~0.2 uM (43) while Mg does not inhibit.
Identification oftryptophan residues acting as reporter groups. Site-directed
mutagenesis studies were carried out to create deA variants lacking each of the four Trp
residues (Trp113, Trp195, Trp248, and Trp256) in an attempt to identify the aromatic
group(s) acting as the fluorescent reporter group. Unfortunately, cultures expressing
recombinant genes encoding the W248F, W248L, W195F, and W195L variants of deA
produced insoluble protein for all cell growth conditions examined. This problem also
had been encountered in prior deA mutagenesis studies (47). The W113F and W113L

variants also could not be Studied because these proteins were either not synthesized or

51

Table 3.2. Summary of the metal and aKG binding to wild-type deA

 

 

Sample Kd of Fe(II) (uM) Kd of OLKG (pM)

Tram 7.45 i 0.61 3.35 i 0.35
deA + 200 1.1M Fe(TI) -- 4.78 i 0.69
deA + 200 11M OtKG 36.3 i- 19.1 --

 

Table 3.3. Summary of metal binding to wild-type deA.

 

 

Metal Kd of Metal (uM)
Te" 7.45 :l: 0.61

Cu.2+ 5.0 a 0.4

Zn2+ 8.4 :1: 2.9

C62+ 3.0 a 0.7

Mn2+ 6.8 3: 0.7

M + 6.2 i 1.3
5 Ca I 3.8 i 0.5

 

52

were very rapidly degraded in the cell. Perhaps related to this result, deA is known to
catalyze enzyme self-hydroxylation during extended incubation with F e(II), OtKG, and
oxygen resulting in Trpl 13 conversion to hydroxytryptophan (62). This reaction was
hypothesized to serve a protective role for the enzyme; i.e., hydroxylation of this Side
chain may prevent the occurrence of more damaging oxidative reactions such as those
resulting in cleavage of the peptide backbone. The lack of this protective side chain in the
W113F and W113L variants may lead to their rapid oxidative degradation in the
cytoplasm. The deA fluorescence spectrum that is quenched by Fe(II) and OtKG is not
associated with hydroxytryptophan based on the distinct properties (e.g., excitation
beyond 300 nm) of this modified aromatic group (22). Because of these complications,
only W256F and W256L deA variants were purified and characterized. The kinetic
properties of these variants closely resembled those of the wild-type enzyme (Table 3.4).
The model of the deA structure based on that of TauD predicts that Trp256 lies
only 5 A from the Fe(II) Site and 4.5 A from the C-1 carboxylate of OLKG (Figure 3.5).
Nevertheless, this tryptophan is n_ot the reporter group studied above as shown by the
fluorescence properties of the W256F and W25 6L deA variants. The relative intensity
(data not shown) and the Fe(II)- and OtKG-dependent changes in fluorescence intensity of
these mutant proteins closely paralleled those of the wild-type enzyme (data not shown
and Table 3.4, respectively), and these mutants resemble wild-type protein in their kinetic
properties (Table 3.4-3.5). The deA structural model predicts that Trp195 lies on the
surface of the protein nearly 15 A from, and on the Opposite face to, the active Site.
Crystallographic Studies with other OtKG-dependent dioxygenases reveal the absence of

significant protein conformational changes upon binding Fe(II) or OtKG (101, 106).

53

 

Table 3.4. 2,4-D kinetics of variant deA proteins.

 

 

 

 

 

Range kw Km ken/Km

Mutant (11M ) (min’l) (11M ) (min'l mM'l)
Wt' 5-1000 1020 i 90 20.0 i 5.4 51,000
W256F 10-1000 352 i 10 17.6 i 1.6 20,000
W256L 10-1000 403 i 26 13.0 at 3.7 31,000
aChapter 2
Table 3.5. OtKG kinetics of variant proteins.

Range kw Km ken/Km Kd(uM)
Mutant (pM) (min’l) (pM) (min'l mM")
Wt 1.5-10 643 i 442‘ 3.2 i 0.621 201,000r 3.4 i 0.4
W256F 5-1,000 362 :1: 32 <10 >36200 2.3 i 0.3
W256L 5-1,000 333 :1: 38 <10 >33300 2.9 :1: 0.3

 

a'Fukumori et a1. (30).

54

Figure 3.5. Predicted positions of tryptophan residues in deA. Of the four Trp
residues in deA, three are close to the active site and could reasonably sense changes
induced by F e(II) and OtKG binding: Trp113 is adjacent to a metal ligand, Trp256 lies
close to both the metal and cofactor, and Trp248 is close to the OtKG binding site. In
wntrast, Trpl 95 is on the surface and distant from the active site.

Asp116

.0
.O
O

 
   
  

His] 14

 

Trp248
His263

   

55

Thus, we conclude that Trp195 is unlikely to sense changes induced by Fe(II) or OtKG
binding. The remaining two tryptophan residues remain as reasonable options for the
fluorescence reporter group(s) based on the structural model. Trp248 is only 4.1 A from
the C-5 carboxylate of OtKG while also forming a hydrogen bond with Thrl41 that
potentially binds the a-keto acid, but this residue is more distant (9.7 A) to the metal site.
Although adjacent to a metal liganding residue, the aromatic ring OfTrp113 is
approximately 10 A from both the F e(lI) and OtKG binding sites. We conclude that the
Fe(II)- and OtKG-dependent changes in fluorescence are likely due directly to binding of
molecules to the active Site or to a resultant conformational change which then leads to an
altered environment that influences the fluorescence of either or both Trp113 and Trp248.
These residues are partially conserved among the OtKG-dependent dioxygenases; thus,
this technique may be applicable to estimating metal and/or OtKG binding affinities of

other family members.

56

CHAPTER 4

PROBIN G THE SUBSTRATE-BINDING OF 2,4-

DICHLOROPHENOXYACETATE/Ot-KETOGLUTARATE DIOXYGENASE BY
SITE-DIRECTED MUTAGENESIS AND MECHANISM-BASED
INACTIVATION

57

 

Ot-Ketoglutarate (OtKG) dioxygenases are mononuclear non-heme Fe(II) enzymes
that couple the oxidative decarboxylation of OLKG to oxidation of their primary
substrates. Depending on the specific enzyme, the primary substrate may be a protein
side chain, a cellular metabolite, or a compound taken up from the environment (78).
Most commonly, substrate oxidation involves hydroxylation at an unactivated carbon
atom, but desaturations, epoxidations, or ring formation or expansion reactions can also
result. Based on crystal structures and site-directed mutagenesis studies, the metal is
bound by two histidines and a carboxylic acid-containing side chain (16, 25, 47, 80, 101,
106, 109). The iron is also coordinate by the OLKG C-l carboxylate and C-2 keto group,
while the C-5 carboxylate forms a salt bridge with an arginine or lysine Side chain. A
conserved jellyroll architecture serves as the platform for these residues (16, 25, 80, 101,
106, 109), which form a conserved HX(D/E)X,.HX(-10)R motif (47).

2,4-Dichlorophenoxyacetic acid (2,4-D)/OLKG dioxygenase (deA) is a
representative family member that hydroxylates the herbicide 2,4-D producing a herrri-

acetal that spontaneously decomposes to 2,4-dichlorophenol and glyoxylate.

O
OVCOO- COO- OH C00
C00-
CHO
Cl C' or Cl
-OOC -00C

Previous site-directed mutagenesis studies revealed the roles Oins114, Asp116, and
Hi5263 in binding iron and suggested that Hile4 functions in 2,4-D binding or catalysis.
Recently, a model of the three-dimensional structure of deA (Fig. 4.1) was developed
based on the crystal Structure of two homologues, taurine/Ot-ketoglutarate dioxygenase

(TauD) (25) and Clavaminate synthase 1 (CAS) (63). The deA model was consistent

58

 

with prior mutagenesis studies and identified a set of residues likely to be involved in 2,4-
D binding and OLKG binding.

The studies described here utilize site-directed mutagenesis approaches to test the
importance of the proposed substrate binding residues of deA by characterizing selected
mutant proteins. Additional deA mutants were characterized in order to assess the
effects of removing a protease-sensitive site and to explore the importance of tyrosine
residues located near the metal center. Furthermore, results demonstrating that
phenylpropiolic acid (PPA) is a mechanism-based, irreversible inactivator of deA are

presented and related to what is known about the structure of deA.

59

EXPERIMENTAL

Site-directed mutagenesis. The amino acid residue numbering scheme used here is
based on that derived from the gene sequence rather than that of the purified protein, and
differs by the addition of one residue compared to that used previously (47, 62). deA
variants were created using the Stratagene QuikChange Mutagenesis System using the
primers listed in Table 4.1 and pUS311 (30) as the starting plasmid. This pUCl9
derivative contains the Ralstonia eutropha JMP134 (pJP4) tfdA gene (94).

Protein purification methods. Wild-type and all variant forms of deA except
Y126F protein were purified from Escherichia coli DHSOt as previously described
(Chapter 2). Protein concentrations were assessed by using the BioRad assay and bovine
serum albumin as the standard. Y126F deA was temperature sensitive so the protein was
purified at 4 °C with alterations to the last two chromatographic steps. Specifically, the
active fractions eluting off the DEAE-Sepharose column were pooled, adjusted to 1 M
ammonium sulfate, loaded onto a phenyl-Sepharose column (2.5 cm x 19 cm)
equilibrated with TE buffer (10 mM Tris, 1 mM EDTA, pH 7.7) containing the same
concentration Of this salt, and chromatographed with a 400 mL linear gradient from 1 to 0
M ammonium sulfate. The active fractions were dialyzed to remove the ammonium
sulfate and then chromatographed on a Q-Sepharose column (2.5 cm x 19 cm) using TE
buffer and a 400 mL linear gradient to 200 mM NaCl in TE.

Kinetic analyses. Typical activity measurements used the 4-aminoantipyrene
assay (30) for detecting 2,4-dichlorOphenol. Oxygen consumption or 2,4-
dichlorocinnamic acid oxidation were monitored using previously described oxygen

electrode or HPLC methods, respectively (Chapter 2). As previously documented (85),

60

 

Table 4.1. Forward mutagenic primer sequences

 

Mutant
K71Q
K71L
R78Q
Y81F
K95L
K95Q
T141V
Y169F
Y244F
R274Q
R274L
R278Q
R278L

Forward Primer

5'-GGC GGT TTC ATC CAG GTC AAT CAA AG-3'
5'-GGC GGT TTC ATC TTG GTC AAT CAA AG-3'
5'-GTC AAT CAA AGA CCT TCG CAA TTC AAG TAC GCG GAG TTO-3'
5'-CGA GAT TCA AGT TCG CGG AGT TGG-3'
5'-CAG TCT CGA CGG CCT GGT CGC GCA AC-3'
5'-CAG TCT CGA CGG CCA GGT CGC GCA AC-3'
5'-CGG GCG GCG ACG TCG AGT TCT GCG AC-3'
5'-GCC GAG CAC TTC GCA CTG AAC TCC C-3'
5'-GAA TTC GTG TTC CGG CAT CGC TG-3'

5'-CAT CTC GGC CAG GCA AGA GCT GCG CCG G-3'
5'-CTC GGC CAG GCT TGA GCT GCG CC-3'

5'-GTG AGC TGC GCC AGG CGA CCA CCC-3'
5'-GTG AGC TGC GCC TCG CGA CCA CCC TG-3'

61

deA activity decreases over time by a combination of irreversible inactivation and
ascorbate-reversible inactivation. To analyze the kinetics of designated variant proteins
whose activity decreased over the assay time course, progress curves were analyzed by
fitting the data to the Eq. 4.1 where P, is the accumulated product at time t, V; is the initial
velocity, and k(inact) is the inactivation rate constant (85).

Pt = n(1-e'“‘““°°‘)k(inact)" Eq. 4.1
In cases where high concentrations of substrate decreased the kw, the data were analyzed
by fitting to Eq. 4.2

16.0.15]

[S]2
K+[S]+ K

k:

 

Eq. 4.2

 

The [(4 values associated with binding Of F e(II) and aKG were measured by using
intrinsic tryptophan fluorescence and fitting to Eq. 4.3 as previously described (Chapter

3).

 

* K4 +[LT]+[ET]_\[(K¢I +[LT]+[ET])-4ILT][ET]
max ZlEr]

 

Eq. 4.3

In this equation, the observed changes in fluorescence (AF) are related to the maximal
fluorescence change (AF max) extrapolated to infinite concentration of titrant. Based on the

known total concentrations of enzyme subunit ([ET]) and ligand ([Lfl), the Kd and AFM
were calculated. Curve fits were computed using KaleidaGraph for Windows by
Abelbeck Software.

Inactivation kinetics. The inactivation of deA by PPA was studied in 10 mM

irrridazole, 1 mM OtKG, 5 uM (NH4)2Fe(SO4)2 , and 20 uM ascorbic acid containing

varying concentrations of the inactivator at 30 °C. Stock solutions of all reagents were

62

 

made fresh prior to each set of experiments. Aliquots were withdrawn at various time
intervals, diluted 100-fold into fresh assay buffer containing 1 mM 2,4-D, and residual
activities were determined as in the standard 4-aminoantipyrene assay procedure. These
data were plotted according to log(residual activity) versus time in order to Obtain the

inactivation rates, k(inact), and further analyzed by fitting to the following equation:

k(inact)“, * [I]
K, + [I]

 

k(inact) = Eq. 4.4

where [I] is the inactivator concentration, k(inact)max is the inactivation rate at saturating
[I], and K; is the [I] that gives rise to half maximal rates of inactivation.

Large-scale inactivation of T fdA by PPA followed by alkylation and protease
digestion. deA (1 -5 mg) was inactivated by PPA in 10 mM imidazole buffer (pH 6.8)
containing 1 mM OtKG, 1 mM PPA, 5 11M (NH4)2Fe(SO4)2 , and 20 uM ascorbic acid at
30 °C . Inactivation was complete by 2 min as confirmed by using the 4-aminoantipyrene
assay. PPA-treated protein was dialyzed against water, lyophilized, and redissolved in 6
M guanidine hydrochloride, 100 mM Tris (pH 8), and 10 mM EDTA . Using darkened
vials, the sample was degassed, and dithiothreitol was added at ten-fold excess over the
cysteines. After 30 min at 50 °C, iodoacetate was added in two-fold excess to
dithiothreitol. The alkylation reaction was quenched with B-mercaptoethanol and the
sample was dialyzed into 50 mM ammonium bicarbonate (pH 7.8). Trypsin (1 % wt/wt)
was added twice with incubation at 37 °C for 1 h after each addition. Digestion was
confirmed by SDS-PAGE and matrix-assisted laser desorption ionization (MALDI) mass
spectrometry (Michigan State University Mass Spectrometry Facility). The MALDI mass

Spectra were measured at an accelerating potential of +25kV using a Voyager STR time

63

of flight mass Spectrometer. Spectra represent the accumulation of 50-100 laser shots
fired at the sample in the matrix Ot-cyano-4-hydroxycinnamic acid.

Separation and characterization of deA peptides. The u'yptic digests of control
and modified proteins were chromatographed on a pep-RPC (HR10/10) (Pharmacia)
column by using a linear 200 mL gradient from 0.1 % trifluoroacetic acid (TFA) in water
to methanol containing 0.1% TFA while monitoring the absorption at 280 nm. All
peptides containing an aromatic moiety were characterized by MALDI mass

Spectrometry (Michigan State University Mass Spectrometry Facility).

64

RESULTS AND DISCUSSION

Site-directed mutagenesis of proposed 2,4-D ligands. Based on deA structural
modeling studies, interactions with the 2,4-D carboxylic acid were suggested to involve
Arg278, Lys7l, Hi3214 and the backbone amide of Serl 17 (Figure 4.1) (Chapter 3). The
model also indicated that a loop comprising residues 86-111 made up one face of the 2,4-
D binding pocket, however a consensus structure for this region could not be Obtained
and these twenty-five residues were omitted from the final model. Notably, several
possible structures for this loop placed Lys95 where it could assist in binding the 2,4-D
via the carboxylate or the ether moiety. Site-directed mutagenic evidence for the
participation of Hi5214 in binding 2,4-D was described previously (47) and supports at
least that portion of the model. In an extension of this approach, variants of deA with
substitutions of Arg278, Lys7l, and Lys95 were constructed and the proteins
characterized

The R278Q, K71L, K71Q, K95L, and K95Q deA variants were purified and
characterized to assess the roles of the residues in catalysis and 2,4-D binding. Iron
binding to these proteins was not affected based on Kd determinations using fluorescence
techniques (data not shown). Kinetic analyses revealed only modest changes in km“, with
an approximately two-fold increase in the rate for the K71Q mutant and four- to five-fold
decreases for the K71L and R278Q variants (Table 4.2). Tumover-dependent inactivation
rates were unaffected (data not shown). The most striking changes observed for these
proteins involved Km, ranging from a 2-fold increase in the case of K95Q to a 200-fold
increase for R278Q (Table 4.2). These results are consistent with a role for Arg278,

Lys7] , and perhaps Lys95 in substrate binding. To further examine the roles of these

65

Fig. 4.1. The active site residues in deA. The crystal structures of TauD and CASl
were used to model the homologous protein deA. The residues believed to be involved
in substrate binding, and for which variants were made in these studies, are illustrated.

Lys7]
§ 2,4-D
Arg278
Aspl l6

 

111le4
‘. \

Thrl4l :

Hi3263

 

Table 4.2. 2,4-D kinetics of variant proteins.a

 

 

294’D kcat Km [‘ch Ksi
Mutant Range (min'l) (11M) (min'l mM'l) (mM)
—Wt° 51000 1020 i 90 20.0 i 5.4 51,000 --
K71Ld" 10010000 256 a 48 370 :t 170 692 16 a 11
K71Qd" loo-10,000 2093 a 150 1,100 a 300 1,900 --
R78Q 125-5,000 452 3: 21 1,500 a 200 301 --
Y81F 51000 346 a 12 33.2 a 3.3 10,400 -—
K95L“ 25-1000 893 a 41 78.5 :1: 12 11,400 --
K95Q 10-1000 733 a 22 39.5 a 5.2 18,600 --
Y126F 5-1000 576i26 31.93: 5.1 18,100 --
T141v 51000 707 a 19 21.6 a 2.2 32,700 --
Y169F 51000 444 a 10 42.9 a 5.1 10,300 --
Y244F 51000 512 a 32 30.8 a 4.0 16,600 --
R274Q°’f 250-10,000 208 3: 16 1600 i 30 131 39 a 18
R274I. 250- 10,000 296 a: 176 2500 i 1300 118 9.9 a 7.4
R278Q“ 25010000 256 a 17 4200 i 200 160 11 a 8

 

“The kinetics were linear and determined at 1 mM OtKG except where noted.
bCalculated inhibition constant for substrate inhibition when fit to Eq. 4.2.

cChapter 2.
dDetermined at 10 mM aKG.

eReaction progress over time was non-linear and analyzed using Eq. 4.1.

fDetermined at 2 mM OtKG.

67

residues, the variant proteins were incubated with the alternative substrate 2,4-
dichlorocinnamic acid. This compound resembles 2,4-D, but possesses a carbon-carbon

double bond in the side chain rather than an ether oxygen.

0
\ C00- C00- coo-
COO-
+ 02 + ——> + CO2 +
Cl C1 C1
-OOC

Cl

Wild-type deA converts 2,4-dichlorocinnamic acid to the epoxide, 2,4-
dichlorophenylglycidic acid, with a Km of 190 :t 56 uM and a km of 52.5 i 5.1 min'1
(Chapter 2). No detectable epoxide synthase activity was observed for the K71L, K71Q,
and R278Q variants of deA in the presence of 1 mM 2,4-dichlorocinnamic acid. These
results are interpreted to be due to a large increase in the Km for this substrate similar to
the Situation for 2,4-D. Higher concentrations of 2,4-dichlorocinnamic acid could not be
tested due to its limited solubility in water. 2,4-Dichlorocinnamic acid was oxidized by
the K95Q variant with an estimated K... > 500 uM. Furthermore, K95L deA had an
approximate four-fold decrease in the Km and similar km for 2,4-dichlorocinnamic acid
(Km = 56.5 :1: 14.0 and kw = 41.3 i 3.2) when compared to wild-type enzyme. These
results imply that changing the basic lysine to the more hydrophobic leucine enhances the
substrate preference for the more aliphatic side chain of 2,4-dichlorocinnamic acid. Such
an interpretation is consistent with Lys95 having a role in binding the ether of 2,4-D.
Lysines and arginines are expected to have multiple hydrogen-bonding partners
because of their large positively charged side chains. Substitution of these residues by
leucine or glutamine could reasonably have longer-range effects. One such effect

Observed here is the significantly increased OtKG Km and Kd Of these mutants (Table 4.3),

68

Table 4.3. OtKG kinetics of variant proteins.a

 

 

Mutant Range km, Km km/Km Kd [(31——
(11M) (“fin") (11M ) Wm" “M" (11M) (HIM)

Wt Look up 643 1 44C 3.2 i 0.6c 201,000c 3.4 :1: 0.4 -
K71Ld’° 2500-106 112 3 10 11,700 3 3300 9.6 2100 3 400 --
K71Qd" 2500—106 >1900 >20,000 -- 4800 3 300 --
R78Q 54,000 403 3 45 <40 >10,100 9.9 3 0.7 .-
Y81F 51,000 285 3 21 <10 >28,500 8.9 3 0.7 --
K95Ld 54,000 461 3 51 <35 >13,200 10.0 3 1.1 -—
K95Q 31,000 301 3 5 <10 >20,100 14.9 3 2.6 --
Y126F 51,000 435 3 3 <20 >21,800 14.8 3 1.0 --
T141V 54,000 560 3 48 <20 >28,000 10.9 3 1.2 --
Y169F 5-1.000 454 3 14 <10 >45,000 11.1 3 2.2 --
Y244F 5,1000 512 3 32 <10 >51,200 16.6 3 3.7 --
R274Q'~"f 500-106 263 3 72 8200 3 3900 32 2300 3 200 44 3 24
R274L 250.106 271 3 68 4800 3 2700 56 9000 3 600 116 3 92
R278Q“ 500-10‘5 224 3 12 4100 3 500 53 1350 3 50 133 3 24

 

2‘The kinetics were linear and determined at 1 mM 2,4-D except where noted.

bCalculated inhibition constant for substrate inhibition when fit to Eq. 4.2.

cFulcumori et a1. (30).

6Reaction progress over time was non-linear and analyzed by using Eq. 4.1.

cDetermined at 10 mM 2,4-D.

{Determined at 2 mM 2,4-D.

69

possibly resulting fi'om repositioning of various active site residues, shifting of the
protein backbone, and/or disruption of secondary structure elements.

Site-directed mutagenesis of proposed aKG ligands. Many of the OLKG-
dependent dioxygenases have a conserved arginine and threonine/serine motif, the
residues of which are hypothesized to interact with the C-5 carboxylate of OLKG (16, 25,
47, 80, 101, 106, 109). In deA, this motif is comprised of Arg274 and Thr14l (47) and
is in agreement with the computer-generated structural model (Chapter 3) (Figure 4.1).
To test whether these residues are critical to ctKG binding, additional mutagenesis studies
were carried out. The T141V, R274L, and R274Q variants of deA were purified and
characterized (Table 4.1 and 4.2). None of the variants exhibited significant differences
in iron Kd (not shown) and the substitutions had only modest effects on kw. T141V also
had no significant change in its OtKG Km, OIKG Kd, or 2,4-D Km, suggesting that it has at
most a minor role in binding to OtKG and may merely be conserved for structural reasons.
In contrast, R274L and R274Q were found to have an ~1000-fold increase in the ctKG
Km and K, (Table 4.3), confirming the critical role for Arg274 at the active Site.
Complimentary to the situation with variants of 2,4-D binding residues described above,
the mutations involving the OtKG binding residue also affected the binding of 2,4-D with
a 100-fold increase in the 2,4-D Km (Table 4.2). This is somewhat surprising since
conversion of the corresponding arginine in deacetoxycephalosporin C synthase
(DAOCS, and OLKG dioxygenase) to a glutamine was not reported to affect the Km Of the
primary substrate (41). The loss of activity arising from conversion of the corresponding
arginine to glutamine in DAOCS and phytanoyl-COA hydroxylase (another family

member) can be reversed by using longer chain length Ot-ketoacids, a phenomenon

70

termed co-substrate rescue (41, 69). In particular, Ot-ketocaproate and Ot-ketoisovalerate

restored activity to greater than 100% Of the OtKG-dependent wild-type activity in the
R25 8Q mutant of DAOCS. Co-substrate rescue was not observed with the R274Q variant
of deA, which had Km values of 1.7 i 0.5 mM and 0.70 :t 0.09 mM and kw values of
365 i 31 min'l and 22.4 :t 2.8 min'1 for Ot-ketocaproate and Ot-ketoisovalerate,
respectively.

Arg78. In addition to mutating codons encoding the residues hypothesized from
the model to be involved in substrate binding, Arg78 was converted to a glutamine in
order to eliminate this protease-sensitive site in deA (30). NO significant proteolysis
product was Observed for purified R78Q. Removal of this protease-sensitive site might be
beneficial in future studies, such as crystallography. Although the deA model suggests
that Arg78 is not located near the active site, the R78Q variant exhibited nearly a 100-
fold increase in the 2,4-D Km. This variant was unaffected in the iron K, and, unlike
variants of the active site 2,4-D-binding residues, did not have an increase in the OtKG K;
or OLKG Km. This lack of effect on OtKG binding is consistent with the residue not being
in the active site. A hypothetical role of Arg78 may be to assist in 2,4-D entry into the
active site. This role has also been hypothesized for His217, since the alanine variant
(HZl7A) of this residue has an increased 2,4-D Km but is not in the active site (47).
Possibly related to these findings, the crystal structure of the OtKG dioxygenase
anthocyanidin synthase reveals a second binding Site for the primary substrate that is
proposed to help bring substrate to the active site (106).

Site-directed mutagenesis oftyrosines near the active site. Tyrosine mutants of

deA were of interest because in TauD (a 30% identical protein) a tyrosine radical was

71

detected and mentioned as a possible catalytic intermediate (unpublished observations by
M. J. Ryle, A. Liu, R. B. Muthukumaran, B. S. Phinney, R. Y. N. Ho, J. McCracken, L.
Que, Jr., and R. P. Hausinger). In order to assess the importance of the deA tyrosines,
the four tyrosines closest to the active site (Tyr8l, Tyr126, Tyr169 and Y244) were
altered to phenylalanines and characterized. All four variants were only slightly affected
when compared to wild-type enzyme in terms of iron K4, OLKG Kd, OLKG Km, and 2,4-D
Km suggesting that these tryosines are not involved in catalysis or substrate binding. The
Y81F, Y168F, and Y244F proteins were purified as normal, but Y126F deA bad the
unique characteristic of being thermally instable. This lability was not evident during the
typical time course of the assay, but was evident in the initial purification steps at room
temperature. This lack of temperature sensitivity during catalysis suggests that the
presence of one or more of the substrates stabilizes the protein during activity
measurements.

Inactivation of T fdA by PPA. In addition to examining substrate binding by site-
directed mutagenesis, inactivation of deA by PPA, an acetylenic analogue of 2,4-D, was
examined. PPA appeared to strongly inhibit or inactivate the enzyme based on the
absence of oxygen consumption after adding 2,4-D to protein previously incubated with
iron, OtKG, ascorbate, and this compound. Kinetic analysis of enzyme inactivation
revealed a first-order loss of activity (Figure 4.2, Panel A) that was dependent on the
concentration Of PPA and required OtKG, and ferrous ions. Consistent with binding of the
inactivator to the active site, high concentrations of 2,4-D protected the enzyme against
inactivation. Saturation kinetics were Observed for the loss of activity, allowing the

calculation of K = 38.1 i 6.0 nM and k(inact)max = 2.3 :1: 0.1 (mm!) (Figure 4.2, Panel

72

B). Inactivation of deA by PPA was not reversible by dialysis against 10 mM imidazole
(pH 6.8) for 72 h, consistent with it being an irreversible, mechanism-based inactivator.

A potential mechanism for deA inactivation by PPA is illustrated in Scheme 4.1.
This mechanism is based on that reported for thymine hydroxylase (an OtKG
dioxygenase) when treated with an acetylenic analogue of thymine (60, 97, 98). Thus,
initial oxidation of PPA by deA likely produces a highly reactive oxirane intermediate
that rearranges to form either of two carbene intermediates. Carbenes are known to be
able to react with amino acid Side chains. Thus, a covalent linkage may form between the
inactivator and an active site residue. Alternatively, a carbene may react with solvent to
spare the enzyme and release an aromatic product. To explore this model further, two
approaches were used. The PPA-dependent inactivation kinetics of the K95L variant
were analyzed and peptide studies were canied out on inactivated deA.

Modeling and site-directed mutagenesis studies suggest that Lys95 might be
positioned near the 2,4-D ether oxygen and thus may lie close to the suspected oxirane
intermediate formed from PPA acid. In order to examine the possible role of Lys95 in
deA inactivation by this compound, the kinetics of inactivation by PPA were explored
with the K95L variant. At 1 mM PPA, no inactivation of the K95L variant was observed
for up to four rrrin, while wild-type deA under comparable conditions was completely
inactivated within one minute. The K71L, K71Q, and R278Q variants were not examined
by this approach since they have Significantly increased 2,4-D Km values, and would
likely bind the inactivator poorly. As a control, the Y81F variant was examined and
Shown to be inactivated at the same rate as wild-type enzyme. The lack of inactivation in

the K95L mutant was not due to a lack of binding since PPA is a competitive inhibitor

73

Fig. 4.2. deA inactivation kinetics in the presence of PPA. Panel A shows the-first-
order loss of activity over time for 2 11M (0), 4 11M (X), 6 11M (O), 8 11M (A), and 10 11M
(V) PPA. Panel B is the plot of the individual inactivation rates versus the concentration
of PPA fit to Eq. 4.4.

 

 

 

 

 

 

 

 

 

 

 

0.8 —
.Q‘
,2 0.6 -
*5
35 04 _
on .
.9.
0.2 -
0 ' i
0 l 2 3 4 5
Time (min)
I
. 3
_/'\
is f
E
v
O
0
<6
.5
V
.32
0 l I L l I
0 200 400 600 800 1000 1200

[Phenylpropiolic acid] (11M)

74

Scheme 4.1. Theoretical mechanism of deA inactivation by PPA.

 

 

 

 

COO- " O ‘
/ _
7? COO-
Fe(IV)=O Fe(ll) L _
O
.. O ..
+
000- COO-
1. 3
H20
(and K95L deA variant)
Covalently modified, An aromatic compound
inactivated deA + active deA

75

 

 

with respect to 2,4-D with a Ki 2 60 11M (Figure 4.3, Panels C and D). A reasonable
explanation for the Observed lack of inactivation of deA by PPA is that any carbene
intermediate that is formed reacts with water. For example, replacement of Lys95 by a
leucine may result in great solvent access to the active site. If PPA is turned over by the
enzyme, the reaction rate was sufficiently low to be indiscernible from background based
on oxygen consumption.

In order to further characterize how wild-type deA is inactivated by PPA, treated
and untreated samples were digested with trypsin and compared by using MALDI mass
spectrometry and reverse-phase chromatography. No differences were observed between
the direct analyses of digestions of modified and unmodified proteins, even though 92 %
of the peptides with molecular weights between 1100 and 2600 Da were able to be
identified. The peptide containing Lys95 was the one peptide in this Size range that could
not be identified in either digest presumably due to mass spectrometric suppression. In an
extension of these studies, tryptic digests of modified and unmodified protein were
chromatographed on a reverse phase column while monitoring the absorbance at 280 nm
(allowing detection of peptides containing Trp, Tyr, and possibly the PPA aromatic ring).
When the chromatography profiles were compared, one novel peak was observed in the
digest of modified protein without a loss of intensity in any other peak (Figure 4.4, Panel
A top). Each of the peaks was collected from the chromatography of the modified protein
and identified by MALDI mass Spectrometry (Figure 4.4, Panel B). All four Trp and all
seven Tyr in deA were accounted for in these peptides. Notably, the peptide containing
Lys95 gave a very weak MALDI signal, consistent with the suppression observed with

the whole digest. Since this peptide was clearly present in peak five, Lys95 can be

76

Fig. 4.3. PPA competitive inhibition kinetics for K95L deA. Panel A and Panel B
depict progress curves of variant enzyme incubated with 2.5 11M and 150 pM PPA,
respectively, with varying 2,4-D concentrations [16 11M (O), 40 uM (O), 80 uM (I), 160
uM (X), 400 11M (+), 800 11M (A), 1600 11M (A), and 4000 11M (7)]. The initial reaction
velocities of these and additional progress curve [1.25 11M (0), 2.5 uM (I), 50 11M (0),
and 150 11M (X) PPA] were plotted versus the concentration Of 2,4-D (Panel C). The K"n
wad deduced by replotting the apparent Km/Vmax as a function of PPA concentration
(Panel D).

77

 

l 000

 

 

800

600
4-D

 

3

400
M 2

 

 

0

 

 

20

 

 

 

 

 

 

 

 

50 100 150 200

1.1M PPA

-50

 

m

0 .

 

2 I l.
AwEEE>oEE
been... 3335

8:50:53 mmEmoE

A83...» 3333.0.

Fig. 4.3. PPA competitive inhibition kinetics for K95L deA.

78

excluded as the site of modification. Indeed, no tyrosine or tryptophan containing peptide
(those absorbing at 280 nm) appears to be modified since there were no differences in the
intensity or retention time of these peaks in the chromatograms. The novel A230-
absorbing peak that likely represents a covalent adduct of PPA and a deA peptide could
not be identified by MALDI mass Spectrometry. This result may arise from suppression
effects or because the species has a mass (< 700 Da) that is obscured by the matrix
features. In order to identify this peak, amino acid analysis and fast atom bombardment
mass spectrometry should be performed to determine the amino acid composition and
mass of this peptide, respectively. Nevertheless, the above results are consistent with
PPA being a mechanism-based inactivator of deA that most likely becomes covalently

bound to an active Site peptide.

79

Fig. 4.4. Peptide purification and identification. Tryptic digests of PPA treated (lower
trace) and untreated (upper trace) deA were chromatographed on a reverse phase
column while monitoring at 280 nm (Panel A). Each of the peaks was collected fiom the
chromatography of the modified protein and identified by MALDI mass spectrometry
(Panel B). An asterisk indicates a peak Observed only in the treated sample.

A

 

 

 

 

 

 

 

 

 

W

 

 

 

 

 

 

 

W
l 2 3 4 * 5 6 7 8 9 10

Peak Number Molecular Weight Sequence B
1 1061.3 AEHYALNSR

881.2 RYDISAR and/or YDISARR
2 714.0 EFVYR
3 877.2 AAYDALPR
4 1 5 1 5.1 FLLGDTDYSEAQR
5 1595.1 YAELADISNVSLDGK
6 1394.4 NAMPPVNWPLVR
7 2589.8 EVVGNFANQLWHSDSSFQQPAAR
8 2420.2 YSMLSAVWPPSGGDTEFCDMR
9 1505.2 WNVGDLVMWDNR
10 2234.0 Unidentified

1 525. l Unidentified

 

80

ACKNOWLEDGEMENTS
I thank Rhonda Husain and Beverly Chamberlain for discussion about interpreting mass

spectrometric data and Jon Elkins for the preparation of Figure 1.

81

CHAPTER 5

PHYSIOLOGICAL AND GENETIC CHARACTERIZATION OF
ENVIRONMENTAL ISOLATES CONTAINING tfdA-like SEQUENCES

82

In prior studies from this laboratory, bacteria isolated from agricultural soils
(specifically a culture collection derived from the MSU-Kellogg Biological Research
Station long-term ecological research (LTER) plots) were examined for the presence of
tfdA-like genes and their ability to degrade 2,4-D (46). PCR amplification of an internal
fragment Of ddA revealed that approximately 37% of these isolates contain a q’dA-like
gene, although none of the isolates degraded or incorporated 2,4-D (46). In this chapter, I
report three sets of studies to better define the role of tfdA-like genes in environmental
isolates.

I initiated a series of experiments to differentiate between two hypotheses that are
compatible with the widespread prevalence of tfdA -like genes in 2,4-D non-degrading
isolates. In the first hypothesis, tfdA-like genes encoded enzymes that degraded 2,4-D (or
more generally, phenoxyacetic acid), but in the absence of 2,4-D exposure, the regulatory
or downstream genes were lost and tfdA became a cryptic gene. Such a sequence of
events would have required very rapid dissemination of the gene followed by a rapid loss
of function since 2,4-D was patented and first applied to the environment in the 1940s.
An alternative, and more tenable, hypothesis is that the tfdA-like genes present in these
bacteria function in a distinct and highly ubiquitous role in soil having nothing to do with
phenoxyacetate decomposition. To address these possibilities, I examined thirty-three
Siberian permafrost cultures using the same procedure of PCR amplification of genomic
DNA with tfdA-specific primers. These isolates had been frozen in the Siberian
permafrost from present time to over a thousand years. In those cells believed to have
been frozen over a thousand years, 2,4-D exposure is highly unlikely and detection of a

tfdA -like gene would be indicative of an alternative function for tfdA.

83

In a second set of studies, I examined whether the tfdA-like gene of environmental
isolates produced enzymes that might function to degrade cinnamic acid or fenrlic acid.
Cinnamic acid and its derivatives are abundant in nature, structurally resemble 2,4-D, and
were found to be substrates of deA (see Chapter 2). Therefore, the degradation of
cinnamic acids may explain the abundance of the tfdA-like gene in the environmental
isolates.

In an additional study, I undertook efforts to clone a tfdA-like gene from an LTER

isolate in order to characterize the substrate utilization profile of the recombinant protein.

84

EXPERIMENTAL

Media and cultures. Escherichia coli DHSOt (pUS311) carrying tfdA on a pUCl9-
derived plasmid (29) was grown in LB medium. R. eutropha JMP134 carries the pJP4
plasmid containing tfdABCDEF (21). R. eutropha JMP228 (pBH501aE) is a derivative of
R. eutropha M1 34 (pJP4) where the tfdA gene has been interrupted by transposon
mutagenesis (100). These cells were grown in MO minimal medium (91) amended
with 2,4-D or succinate. The permafrost isolates were supplied by Dr. James Tiedje and
were routinely cultured on 20% tryptic soy broth. The LTER isolates are a collection of
environmental isolates that have been phylogenetically characterized and examined for
the presence of the tfdA gene (46). They were cultured on MMO amended with 2,4-D or
succinate, 20% TSB, or R2 (0.5 g/L yeast extract, 0.5 g proteose peptone 3, 0.5 g/L
casarnino acids, 0.5 g/L dextrose, 0.5 g/L starch, 0.3 g/L sodium pyruvate, 0.3 g/L
potassium phosphate, 0.05 g/L magnesitun sulfate).

Amplification of the tfdA-like gene. Genomic DNA was isolated from overnight
cultures (8). A 360-bp internal tfdA fragment was amplified from 10-100 ng of genomic
DNA using previously described primers (46).

Cinnamic acid and ferulic acid utilization studies. The isolates were grown by
Shaking at 30 °C in tubes Of R2 amended with 0.5 mg/mL trans-cinnamic acid or ferulic
acid. After 2 days, the tubes were centrifuged and the culture supematants were analyzed
by HPLC (Hewlett Packard 1050). Usage of cinnamic acid or ferulic acid was determined
by analyzing the peak height at 280 nm after isocratic chromatography on a Merck

Lichrosorb RP-18 column with 50:50:0.1 mixture of methanol:water:phosphoric acid.

85

RESULTS AND DISCUSSION

Prevalence oftfdA-like genes in Siberian permafrost samples. It has been
hypothesized that the 2,4-D biodegradation pathway arose rapidly upon application of
2,4-D in the 1940s and continues to arise rapidly in pristine soils treated with 2,4-D (32).
This situation requires that a form of each of the tfd genes be present in soils prior to
herbicide application. Additionally, the encoded proteins must either use or rapidly
evolve to use the xenobiotic or corresponding pathway intermediate. For the genes
downstream of tfdA in the pJP4 plasmid, homologues have been identified that work on
more “natural” substrates (35, 36, 73). In contrast, a homologue of tfdA functional toward
a naturally occurring chemical has remained elusive. To further examine the evolution
and dissemination of tfdA, I carried out a set of studies with Siberian soil isolates. I
examined the presence of tfdA-like sequences in Siberian permafrost cultures using the
previously described method of PCR amplification of genomic DNA. Of thirty-three
permafrost culture stocks from which genomic DNA was purified, eight genomic preps
(or 24%) tested positive for the presence of a tfdA -like gene (Table 5.1). Since some of
these isolates have been frozen in the Siberian permafrost for over a thousand years prior
to culturing, 2,4-D exposure with these isolates is highly unlikely. Theoretically, the
permafrost samples have been isolated away from even incidental 2,4-D exposure; yet,
the tfdA gene was prevalent in isolates from these soils at levels Similar to those found in
current agricultural soils. Detection of tfdA-like genes in isolates frozen away more than a

hundred years ago is indicative of an alternative function for tfdA.

86

Table 5.1. Characteristics of the Permafrost Isolates Tested.

 

 

Strain Estimated Age Putative Identification tfdA-like geneI
23-9 F Iavobacterium Sp. (76) + and -

26-2 -

33-1 1000K (76) Arthrobacter Sp. (76) -

33-9 +

215-1 20-30K (76)
215-2 20-30K (76) -
215-4 20-30K (76) 3-
215-5 20-30K (76) 3
215-14 20-30K (76) -
215-15 20-3OK (76) -

215-30 -
309-5 20-30K (76) + and -
309-16 20-30K (76) -
312-12 -
342-5 -

392-1 Modern (76) ‘ -
392-7 Modern (76) -
1892 -
241 1 -
5138-1 Exiguobacterium sp. (57) +
641 1 -
ED 23 Acinetobacter Sp. - -
ED 28 + and -
ED 37 -
EDM 6-4 + and -

 

 

' A positive Sign indicates the presence of an amplification product derived from a tfdA-like gene.
Duplicate freezer stocks existed when stocks were made at different times or when a spontaneous mutation
yields a new colony phenotype (a color change, a change in antibiotic susceptibility, etc). In the case of
ED28, 23-9, 309-5, and EDM6-4, one culture tested positive while the other tested negative. This
phenomenon may be similar to the results seen with the LTER isolates where the gene seems to disappear
from the freezer stocks over time. In other duplicates tested (ED 37 and EDM 33-1), neither genomic prep

contained the tfdA-like gene.

87

Assessment ofcinnamic acid utilization by environmental isolates. Substituted
cinnamic acids are prevalent in nature, resemble 2,4-D in structure and were recently
found to be substrates of deA (Chapter 2). Thus, it is plausible that a tfdA-like gene
exists in nature where it participates in degradation of a non-chlorinated compound like
cinnamic acid. Incorporation of a few mutations may have converted the gene product to
one that degrades the chlorinated xenobiotic 2,4-D. For example, an E. coli homologue of
an atrazine-degrading enzyme specific for the natural compound melamine differs from
the herbicide-degrading protein by only 9 of 475 amino acids (89). Similarly, directed
evolution experiments using DNA shuffling or random mutagenesis have provided other
examples of how a few changes in the amino acid sequence can result in dramatic
changes in substrate specificity (e.g. (108)).

I grew the LTER isolates in R2 amended with cinnarrric acid or ferulic acid in
order to assess their ability to degrade these natural products. After analyzing media
supematants by HPLC to examine the loss of cinnamic or ferulic acid (Tables 5.2-5.4), I
concluded that the presence of a tfdA -like gene in an LTER isolate was not predictive of
its ability to degrade cinnamic acid or ferulic acid.

Cloning of a tfdA-like gene fiom an environmental isolate. In an effort to
determine the functions of the tfdA-like genes in the environmental isolates, I isolated
genomic DNA from four isolates (LTER 3, 8, 28, and 40) so as to clone the regions
flanking the tfdA-like gene. Unfortunately, in the three years that have elapsed since the
LTER isolates were initially examined, the freezer stocks seem to have lost the gene.
Using genomic DNA preparations prepared identically to those described by Hogan et a1.

(46), I obtained DNA from which the 168 rDNA could be amplified, but

88

 

Table 5.2. Growth of LTER isolates on media amended with cinnamic acid.

 

 

 

 

Isolates containing a Isolates lacking a Total
tfdA-like gene tfdA-like gene
No cinnamic acid
degradation 10 24 34
Cinnamic acid
degraded 17 23 40
“”31 27 47 74

 

 

 

 

x2=1.35; p > 0.20 and the null hypothesis cannot be rejected

Table 5.3. Growth of LTER isolates on media amended with ferulic acid.

 

 

 

 

 

Isolates containing a Isolates lacking a Total
tfdA-like gene tfdA-like gene
No ferulic acid
de 1 tion 16 37 53
F erulic acid
degraded 11 10 21
ma] 27 46 74

 

 

 

 

x2=3.25; p > 0.20 and the null hypothesis cannot be rejected

89

 

 

Table 5.4. Cinnamic acid and ferulic acid degradation by LTER isolate.

 

 

 

 

 

 

 

 

 

LTER Degrades Degrades LTER Degrades Degrades
Isolate cinnamic acid ferulic acid Isolate cinnamic acid ferulic acid
1 + + 44 + +
2 - + 47 - -
3 + + 48 + -
4 - - 52 - -
5 + - 54 - +
6 - - 57 - -
8 + + 59 + -
9 - - 61 + -
10 + + 62 + -
11 + - 63 - -
12 + - 64 - -
13 - - 65 - +
14 + + 66 - -
15 + + 67 + +
17 - - 68 - +
18 - - 69 + -
20 + + 70 + -
21 + + 71 + -
22 - - 72 - -
24 - - 73 - +
25 + + 74 - -
26 - - 75 - -
27 + - 77 + -
28 + + 78 - -
29 + + 79 - -
30 - - 81 - -
31 + - 82 ND ND
32 + - 86 - -
33 + + 88 + -
34 + - 89 + -
36 - - 91 - -
37 + - 92 + -
38 - - 94 — -
39 - - 95 - -
40 + - 97 + -
41 + + 98 + -
42 + + 100 + -
43 + -

 

 

 

Isolates numbered in bold contain a tfdA-like gene (46).

90

the tfdA-like gene could not. I varied the growth condition (e.g., using a variety of-media
such as TSB, MMO, and R2) in an attempt to enhance the persistence of tfdA, but I
obtained no amplification products with the tfdA specific primers. In addition, I examined
LTER 40 genomic DNA that was still available from 1997 and quantified the abundance
of tfdA by a PCR titration. I Observed a visible band on an ethidium bromide stained gel
after PCR amplification of pUS3 11 (containing tfdA) DNA at 20 fg/uL but not at 10

fg/ 11L. This corresponded to 4 x 104 plasmids/11L. Similarly, I found that the lowest level
of amplification of the LTER 40 genomic DNA required 1 ng/pL or 1 x 106 Mb/pL
genomic DNA. Assuming 810 Mb genome, this corresponds to 1 x 105 genomes/1.1L.
Therefore, I conclude there is approximately 1 tfdA-like gene per 25 genome equivalents
in this previously prepared DNA, and less in freshly prepared samples.

What happened to the U'dA -like genes in the LTER freezer stocks? The internal
fragment of tfdA is no longer amplifiable from cultures grown from the LTER freezer
stocks. This lack of reproducibility may be due to contamination of the original genomic
preparations, to loss of the tfdA-like gene from the freezer Stocks over time, or to loss of
the cells containing the tfdA-like gene more rapidly in the freezer than those lacking the
tfdA-like gene. Several lines of evidence suggest that the original amplification of the
JdA fragment was not due to Contamination. PCR amplification was confirmed by at least
one additional amplification reaction (46). If contamination occurred, one might expect
contamination in batches of genomic preparations or that the contaminants would be all
the same. Neither the amplification nor the hybridization profile of the tfdA fragments
could be correlated to certain groups of genomic preparations and amplification reactions

(44). Furthermore, the sequences of the fragments amplified were different. It was shown

91

that the PCR fragments exhibited differential hybridization to the canonical R eutropha
JMP134 (pJP4) tfdA gene. Approximately half of the genes hybridized to the R eutropha
JMP134 (pJP4) tfdA gene with high stringency. It was confirmed that these were highly
homologous to the R. eutropha J MP134 (pJP4) tfdA gene by sequencing the PCR
products of a select few (46). LTER 40 which hydribidized at high stringency to pJP4
was shown by sequencing to be identical to the pJP4 tfdA (46). Those that did not
hybridize at high stringency to the pJP4 tfdA were either novel sequences (LTER 5 and
75), or were homologous to the tfdA encoded on the chromosome of Burkholderia
cepacia RASC (LTER 1, 8, and. 11) (46). Additionally, genorrric DNA from several
isolates was routinely obtained and amplified up to a year later (44). The results of this
current study suggest that the tfdA-like gene may have been present in only 4% of the
genomes in this DNA.

There are several examples of unstable DNA elements that might account for this
loss. Acinetobacter strain Adpl loses a lOO-kb fragment of its chromosome that contains
genes for the degradation of ferulic acid when the isolate is not grown on this carbon
source (90). More commonly, plasmids and transposable elements can be lost upon
cultrning when selection is lacking. The genes required for 2,4-D degradation are often
associated with plasmids and transposable elements (13, 75); however, such genes are not
expected to be lost in freezer stocks. Perhaps related to the Observed instability, loss of
cell viability on antibiotic containing agar plates has been noted when freezer stocks of
cells containing pET vectors are made with a concentration of glycerol greater than 10%
(95). The phenomenon has not been studied in detail, but perhaps higher concentrations

of glycerol inhibit the ability of the cells to grow under selection due to loss of the

92

plasmid between freezing and plating. The current LTER freezer stocks contain 15-20%
glycerol (44); thus, a Similar uncharacterized phenomenon may have occurred. The
presence of the tfdA-like gene at low levels in the 1997 genomic preparations may be due
to the freezer stocks having been subcultured three times (44). The repeated culturing of
Ralstonia sp. strain TFD41, a 2,4-D degrading organism, is known to cause alterations in
the genomic DNA (70).

The future for the LT ER isolates. The lack of stability of the tfdA gene confounds
the interpretation of results with the isolates. I suggest that any LTER studies subsequent
to the 1997 paper are suspect. The tfdA-like gene could be cloned from the Older genomic
preps using a BAC library approach, but the loss of the tfdA-like gene in the freezer
stocks makes it impossible to go back to the isolate and do physiological studies to
confirm the function of the gene. For these reasons, further studies with these

environmental isolates were abandoned.

93

CHAPTER 6
CLONING OF A tfdA-like SEQUENCE FROM Bordetella pertussis AND

PRELIIVIINARY CHARACTERIZATION OF THE CORRESPONDING
PROTEIN

94

p)

 

B. pertussis is a gram negative, obligate aerobic, coccobacillus, and the causative
agent of pertussis (whooping cough). The genomes of Bordetella pertussis, Bordetella
parapertussis, and Bordetella bronchiseptica have recently been sequenced and shown to
contain two, two, and three open reading frames, respectively, related to tfdA (Figure 6.1)
(1-3). The tfdA-like sequence at ~1.75 Mb on the B. pertussis chromosome encodes the
most closely related protein to deA of any microorganism not known to metabolize 2,4-
D. Although free-living B. pertussis is not thought to be present in nature (its only known
reservoir is humans), other Bordetella strains have been found in the environment.
Furthermore, a Bordetella strain isolated from enrichment cultures inoculated with
activated sewage Sludge has been shown to degrade 2,4—D (38). Like many characterized
2,4-D-degrading strains (i.e., Ralstonia eutropha JMP134 and Burkholderia cepacia
RASC), B. pertussis is a B-proteobacterium. I hypothesize that the B. pertussis tfdA -like
gene is a close relative of tfdA ’s ancestor. Characterization of the B. pertussis protein may
elucidate the function of proteins encoded by tfdA-like sequences in the environment. In
order to determine the function of the B. pertussis tfdA -like gene, I cloned the Open
reading frame and studied the corresponding recombinant protein and a recombinant

maltose binding protein (MBP) fusion of the protein.

95

 

Figure 6.1. Phylogenetic tree of characterized and hypothetical proteins homologous
to deA. Cloned genes with corresponding characterized recombinant proteins are in

bold.

_| Burkholderia mallei ORF —
Burkholder pseudomallei ORF
—— Mycobacterium smegmatis ORF

Yersinia pestis ORF
TauD Escherichia coli Sulfonate
Pseudomonas aeruginosa ORF Dioxygenases

Mycobacterium avium ORF
YDO Saccharomyces cerevisiae
Burkholderia pseudomallei ORF
‘ _l Burkholderia mallei ORF
Burkholderia pseudomallei ORF

 

AtsK Pseudomonas putida S-313
__E Pseudomonas aeruginosa ORF Sulfate ES“? Cleaving
Burkholderia pseudomallei ORF Dioxygenases
Burkholderia mallei ORF _

_—lBurkholderia pseudomallei ORF
Bordetella bronchiseptica ORF
Bordetella pertussis ORF

Bordetella parapertussis ORF
Pseudomonas putida ORF
Pseudomonas putida ORF
Pseudomonas syringae pv tomato ORF
Pseudomonas aeruginosa ORF

. Burkholderia pseudomallei ORF

__l Bordetella bronchiseptica ORF

 

 

Bordetella pertussis ORF
Bordetella parapertussis ORF
~—Pseudomonas aeruginosa ORF
Mycobacterium avium ORF
Mycobacterium smegmatis ORF
Pseudomonas syringae pv tomato ORF
deA Burkholderia cepacia RASC
deA on pEMT8 plasmid 2,4-D dioxygenases
deA Ralstonia eutropha JMP134 (pJP4)
Bordetella pertussis ORF
Bradyrhizobia sp. ORF
Mycobacterium avium ORF

 

 

.———4 o a a o
100 substrtutrons per 100 ammo acrds

96

EXPERIMENTAL

Phylogenetic tree of deA relatives. Protein sequences more than 30%
homologous to the entire Open reading frame of deA were identified in GenBank and the
genome sequencing projects using tblastn. A rooted phylogram of these protein
sequences was developed using the GrowTree program in the GCG Wisconsin Package
with a blosurn62 sequence comparision matrix, the Jukes-Cantor distance correction
method, and the neighbor joining tree construction method.

Cloning. B. pertussis genomic DNA was graciously supplied by Trevor Stenson
of Alison Weiss’s lab at the University of Cincinnati. The BP1665 Open reading flame
(corresponding to nucleotides 1751989-1752825 on the B. pertussis genome) was PCR
amplified from the genomic DNA using the forward primer 5’-
GCTCTAGAATGACCATCACCATTACC-3' that contains an Xbal cleavage site 5'-
extension and the reverse primer 5'-CATCAAGCTTCAGACCGGTICCAGC-3' that
contains a HindIII cleavage Site 3'-extension and cloned into pMAL-c2 for expression in
Escherichia coli DHSOt. PCR was carried out in 20 mM Tris-HCl (pH 8.4), 50 mM KCl,
2.5 mM MgC12, 1.25 mM each dNTP (Gibco), 1 11M each primer, and 0.05 units/1.1L
Gibco T aq DNA polymerase. Initial denaturation occurred at 94 0C for 5 min, followed
by elongation for 35 cycles of denaturation for 1 min at 95 °C, annealing at 55 °C for 45
sec, and elongation at 72 °C for 45 see, with a final 10 min elongation step at 72 °C. The
resulting PCR product was cleaved with HindIII and XbaI and inserted into the pMAL-c2
cut with the same enzymes to create pMAL-BP1665. The resulting plasmid was
sequenced with the New England BioLabS Ml3/pUC and malE sequencing primers to

ensure the presence of the appropriate insert.

97

The B. pertussis tfdA-like gene was subcloned into pET24a(+) (for production of
the native protein under regulation of the T7 promoter) and into pET28a(+) (to produce a
His-tagged protein under regulation of the T7 promoter) and expressed in Escherichia
coli C41 (68). The B. pertussis tfdA-like gene was PCR amplified from 2 pg/uL pMAL-
BP1665 plasmid using as the forward primer 5'-
AGGATATACATATGACCATCACCATTACC-3' that contains a 5’-extension and
mutation creating a NdeI cleavage site and New England BioLabs M13/pUC as the
reverse primer. The PCR product was cleaved with NdeI and HindIII and inserted into
pET28a(+) and pET24a(+) cut with the same enzymes to produce pET28-BP1665 and
pET24-BP1 665 respectively.

Production and purification of the B. pertussis T fdA-like protein. Escherichia coli
DHSOt (pMAL-BP1665) was grown from a 1% inoculation of an overnight starter culture
at 30 °C in LB medium containing 100 ug/mL ampicillin. At an A600 of ~0.4, the cultures
were induced for two h with a final concentration of 0.4 mM IPTG, harvested by
centrifirgation, and suspended in 30 ml of column buffer (20 mM Tris, 1 mM EDTA, 200
mM NaCl, pH 7.8) containing 1 mM phenylmethylsulfonyl fluoride and 10 ug/mL
leupeptin. Crude cell extracts were obtained by passing the cells twice through a pre-
cooled French pressure cell (16,000 psi) and clarifying the debris by centrifugation
(100,000 x g for 45 min) at 4 °C. After diluting the cell extracts to 150 mL with column
buffer, the cell extracts were applied at 4 °C to a column containing New England
Biolabs amylose resin (2.5 x 6 cm). The column was washed with an additional 150 mL
of column buffer and the protein was removed from the column with column buffer

amended with 10 mM maltose. The fractions were examined by 12% SDS-PAGE gel and

98

 

 

pooled accordingly. The resulting MBP-BP1665 was dialyzed against 10 mM imidazole
to remove the EDTA. The protein concentration was determined by using a commercial
protein assay (BioRad) with bovine serum albumin as a standard.

E. coli C41 (DE3) (pET28-BP1665) and E. coli C41 (DE3) (pET24-BP1665)
were grown from 1% inocula of overnight starter cultures at 30 °C in LB medium
containing 100 ng/mL ampicillin. At A600 of ~2, the cultures were induced for 2 h with a
final concentration of 0.4 mM IPTG, harvested by centrifugation, and suspended in 30 ml
TE buffer (20 mM Tris, 1 mM EDTA pH 7.8) containing 1 mM phenylmethylsulfonyl
fluoride and 10 ug/mL leupeptin. Crude cell extracts were obtained in the same manner
as already described. Native length BP1665 protein was enriched by chromatography at 4
°C on a column of DEAE-Sepharose (2.5 x 19 cm) using TE buffer and a 400 mL linear
gradient to 200 mM NaCl. The enzyme eluted at approximately 100 mM NaCl. deA-
containing fractions were dialyzed in TE buffer and chromatographed on a Mono Q
column (HRlO/ 10) (Pharmacia) at room temperature. The enzyme eluted at about 20 mM
NaCl when using a 100 mL linear gradient to 100 mM NaCl in TE buffer.

Assessment of substrate specificity. The activity of the MBP-BP1665 and native
length BP1665 toward potential substrates was determined at 30 °C in 10 mM imidazole
buffer (pH 6.8) containing 1 mM Ot-ketoglutarate (OLKG), 50 1.1M (N114)2Fe(SO4)2 , and
200 11M ascorbic acid by using one of two different assays. For all methods, stock
solutions of (NH4)2Fe(SO4)2 and ascorbic acid were made fresh prior to each set of
experiments. The capability to degrade phenoxyacetate and 2,4-D was tested by using the
previously described 4-aminoantipyrene spectrophotometric assay (30). The release of

sulfite from the sulfonates taurine, isethionate, cysteate, and taurocholate was examined

99

by using the previously described Ellman’s reagent assay (24). The activity toward other
potential substrates was determined by monitoring the rates of oxygen consumption at 30
°C with a YSI model 5300 biological oxygen monitor as previously described (Chapter
2).

Molecular weight determination. The molecular weights Of the MBP-BP1665 and
native length BP1665 were estimated by gel filtration on a Pharmacia HR 10/30 Superose
6 column in the above column buffer. Calibration was carried out by using gel filtration
standards (BioRad) containing thyroglobulin (670 kDa), bovine y-globulin (158 kDa),
chicken ovalbumin (45 kDa), equine myoglobin (17 kDa) and vitamin B-12 (1.35 kDa).

UV/vis spectroscopy. Detection of a charge-transfer transition characteristic of

OtKG dioXygenases was determined by the method of Ryle et a1. (84). The UV/vis

spectrum was taken of a degassed solution of 190-250 11M protein and 1 mM ctKG under
nitrogen. Degassed ferrous ammonium sulfate under nitrogen was added to a

concentration equal to the protein concentration.

100

RESULTS AND DISCUSSION

Genome sequencing has revealed a preponderance of uncharacterized OtKG
dioxygenases. For example, the genome ofArabidopsis thaliana contains over sixty-four
open reading frames that Show strong similarity to known OtKG dioxygenases and less
than 20% have known functions (78). To roughly assess the prevalence of OtKG
dioxygenases related to deA, BLAST was used to identify proteins in the genome
sequencing projects with >3 0% identity to the entirety of deA. Although the genome
sequencing projects searched covered bacteria from most of the phylogenetic branches of
the eubacteria, homologues of deA were only found in the Mycobacteria in the Gram
positives, the Bordetella and Burkholderia in the B-proteobacteria, and the Pseudomonas
and Enterobacteriaceae in the y-proteobacteria.

Upon examination of a phylogenetic tree made from these protein sequences
(Figure 6.1), only eleven of the thirty uncharacterized open reading frames (ORFS) could
be grouped with characterized proteins. All three known 2,4-D/OLKG dioxygenases
grouped together. Two uncharacterized ORFS formed a clade with AtsK, a Pseudomonas
putida S-313 sulfate-ester cleaving/aKG dioxygenase (103). Nine uncharacterized ORFS
grouped with the sulfonate/orKG dioxygenases of Escherichia coli (24) and
Saccharomyces cerevisiae (45). Interestingly, Burkholderia pseudomallei and
Burkholderia mallei have three and two ORFS respectively clustering with the
sulfonate/OtKG dioxygenases suggesting that these organisms widely utilize sulfonates
for growth.

Numerous ORFS seem to be strictly conserved among phylogenetically related

organisms. Three ORFS are similar when compared between the species in B. mallei and

101

B. pseudomallei. Likewise, B. pertussis, B. parapertussis, and B. bronchiseptica contain
two ORFS that are very similar in these genera. Intriguingly, B. pertussis has an open
reading frame (BP1665) that is distinct from the open reading flames of the other
Bordetella and that is most closely related to the 2,4-D/OLKG dioxygenases. In order to
determine if this Open reading frame produces an enzyme that could use 2,4-D and to
explore the ancestral substrate specificity of deA, this open reading flame was cloned
for the production of recombinant protein.

MBP-BP1665 was produced from the pMAL-BP1665 plasmid and purified by
using the amylose resin. No activity was detected for this protein when assayed with
phenoxyacetates, sulfonates, benzofuran-Z-carboxylic acid, chromone-Z-carboxylic acid,
cinnamate, hippurate, indole-3-acetic acid, indole-3-acrylic acid, indole-2-carboxylic
acid, indole-3-carboxylic acid, indole-3-pyruvate, 2-naphthoxyacetic acid, PPA, butyrate,
sodium dodecylsulfate, Z-hydroxycinnamate, 3-hydroxycinnamate, 4-hydroxycinnamate,
2-methoxycinnamate, 3-methoxycinnamate, 4-methoxycinnamate, ferulate, 3,5-
dimethoxycirmarnate, 4-aminoethylphosphonic acid, thymine, isethionate, DL-alanine, L-
cysteine, L-aspartate, L-glutamate, L-phenylalanine, L-glycine, L-histidine, L-isoleucine,
L-lysine. L-leucine, L-methionine, L-asparagine, L-proline, L-glutarrrine, L-serine, L-
threonine, L-valine, L-tryptophan, L-tyrosine, or phosphonoforrnate.

Gel filtration chromatography followed by electTOphoresis of selected flactions on
an SDS-PAGE gel indicated that MBP-BP1665 was mainly present as high molecular
weight multimers (>4,000 kDa) that eluted in the column void volume. TauD and deA,
other closely related OLKG dioxygenases, are dimers (24, 30), and it is presumed that the

BP1665 Should be a dimer as well. I conclude that the high molecular weight multimers

102

are non-native aggregates, perhaps analogous to a previously described state of two
inactive mutants of deA (47). In those cases, the non-flrsion H245A and H262A mutants
of deA formed inclusion bodies while the corresponding maltose-binding protein
fusions were soluble but formed high molecular weight multimers (47).

Additionally, MBP-BP1665 did not develop a charge-transfer transition
characteristic of OtKG dioxygenases. In the absence of oxygen and the presence of OtKG
and ferrous ions, deA and TauD have a characteristic metal-to-ligand charge-transfer
transition at 500-600 nm arising flom the bidentate binding of OtKG to iron (43, 84). Lack
Of this charge-transfer transition is consistent with mis—folded protein that cannot bind
substrate. I conclude that overproduced MBP-BP1665 did not fold properly in E. coli.

The B. pertussis tfdA-like gene was flirther subcloned into pET28 and pET24. The
His-tagged fusion of the BP1665 protein expressed from pET28 was insoluble and
formed inclusion bodies. The non-fusion version of BP1665 expressed from pET24 was
mostly insoluble, but the small proportion of soluble protein was able to be purified flom
the cell lysates. NO activity was detected for this protein when assayed with
phenoxyacetates or sulfonates. Gel filtration chromatography followed by electrophoresis
of selected fractions on an SDS-PAGE gel indicated that the protein was mainly present
as monomers (27 kDa). Similarly, this protein did not develop the metal-to-ligand charge-
transfer transition at 500-600 nm. I conclude that the overproduced non-fusion BP1665

protein also does not fold properly in E. coli.

103

 

CHAPTER 7

CONCLUSIONS AND FUTURE DIRECTIONS

104

 

The goal of this work was to determine what governs substrate specificity in deA
and related enzymes. 2,4-D is the preferred substrate for deA followed by 2-
naphthoxyacetic acid, benzoflrran-2-carboxylic acid, 2,4-dichlorocinnamic acid, 2-
chlorocinnamic acid, l-naphthoxyacetic acid, and 4-chlorocinnamic acid. A generic
structure of a substrate for deA might be an aromatic ring with a three atom side chain
in an extended conformation and ending in a carboxylate. Compounds with a terminal
sulfonate or phosphonate were not found to be substrates.

The latter characteristic can be understood upon examination of the deA
structural model. In the model of deA, the two carboxylate oxygens are coordinated to
Arg278, Lys95, His214, and the backbone amide of Ser117. Kinetic analyses of variants
of these residues were consistent with these roles. In contrast, the ~30% identical TauD
selects for a sulfonate-containing substrate through coordination of the three sulfonate
oxygens to arginine, histidine, and a backbone amide. Additional insights into the
specificity of deA for binding 2,4-D were obtained by Site-directed mutagenesis of
Lys95 and inactivation of the enzyme by PPA. The results are consistent with Lys95
being hydrogen bonded to the ether of 2,4-D, partly explaining the preference of deA for
2,4-D over 2,4-dichlorocinnamic acid. Residues involved in interacting with the ring of
2,4-D or with the ring chlorines were not identified and could be the focus of further
research. In the deA structural model, no amino acid side chains interact with the ring or
these chlorines, but residues participating in such interactions may be present in a protein
loOp that was omitted from the structure. In order to further explore the role of these
omitted residues, alanine-scanning mutagenesis could be carried out and/or a crystal

structure obtained. deA samples were provided to the crystallographer who deduced the

105

TauD crystal structure, however a crystal structure of deA is unavailable because thus
far the crystals have been amorphous. At least some of the crystallography problems are
likely due to a proteolysis site in deA that now has been eliminated in the R78Q variant,
so further efforts with this variant should be continued. Although analysis of substrate
interactions may be complicated in this variant because of its high 2,4-D Km, it would be
worthwhile especially since Arg78 is not expected to be an actual 2,4-D ligand.

Based on the deA structural model, Arg274 and Thrl4l were identified as
possible OtKG ligands. Site-directed mutagenesis data are consistent with Thr141 having
only a minor role, if any, and Arg274 having a major role in binding aKG. Importantly,
the activity of the R274Q variant did not return to wild-type levels when using other Ot-
keto acids (termed co-substrate rescue), highlighting the differences among the members
of the OLKG dioxygenase superfamily. These. differences between deA and other OLKG
dioxygenase family members may be important when trying to understand the trade-Offs
made when a protein evolves a new substrate specificity.

The intrinsic tryptOphan fluorescence of the protein was used to develop a binding
assay for OtKG and Fe(II). In addition, each of the four tryptophan residues were altered
in an effort to identify the fluorescent reporter in deA. Even though Trp256 was the
most likely candidate due to its close proximity to the active site, it was excluded as the
reporter group. The reporter is expected to be some combination ofTrp113 and Trp248.
Further work Should be carried out to determine the role OfTrp113 in the protein. This
tryptophan has been identified as becoming hydroxylated when protein containing OtKG,
but not substrate, is incubated in the presence of oxygen for thirty minutes. Additionally,

in the studies described here, variants OfTrp113 could not be expressed suggesting an

106

 

important role for this tryptophan. The hydroxylation of Trpl 13 has been hypothesized
to serve a protective role for the enzyme by preventing the occurrence of more damaging
oxidative reactions such as those resulting in cleavage of the peptide backbone. The lack
of this protective side chain in the W113F and W113L variants may lead to their rapid
oxidative degradation in the cytoplasm. Alternatively, in vivo hydroxylation ofTrp113
may signal a lack of substrate and lead to degradation of the protein. Further work on
Trp113 could involve fluorescence techniques Since the characteristics of
hydroxytryptophan differ flom that of tryptophan. Additionally, examination of the
W113F and W113L variants in a different expression system may allow for expression of
the variants.

In order to examine the evolution of substrate specificity of a tfdA gene product, I
sought to clone a tfdA-like gene from environmental organisms isolated from modern
agricultural soil. Unfortunately, tfdA -like genes proved to be unstable in the available soil
isolates and this approach was abandoned. With projects like the meta-genome project
(81), where DNA from soil is being cloned and sequenced, the sequence of an
environmental #dA-like gene is likely to be determined. This information will facilitate
the cloning of the gene and characterization of the encoded protein. Although my
preliminary efforts to evolve deA-like activity were unsuccessflrl, I did develop plate
assay methods to detect formation of a protein that preferably hydroxylates 2,4-D. As an
alternative direction to the evolution work, I showed that tfdA-like genes are present in
permafrost soil samples with an estimated age greater than 10,000 years. These results
suggest that tfdA-like genes have not appeared in response to 2,4-D exposure; rather, they

are prevalent in the environment and of ancient ancestry.

107

 

In an effort to further evaluate the evolution of substrate specificity of a tfdA gene
product, the Bordetella pertussis tfdA-like gene was cloned and overexpressed, and the
protein characterized. Unfortunately, enzyme activity was not detected for any potential
substrate examined. The B. pertussis tfdA -1ike gene may be of flnther interest to study
since it is not found in the initial shotgun sequencing of Bordetella parapertussis or
Bordetella bronchiseptica. If these results hold up, this might suggest that the target gene
encodes one of several proteins that differentiates Bordetella pertussis from the other A
Bordetella sp. Further information about the function of this ORF may be provided in i
rnicroarray work when B. pertussis arrays become available. If this protein is involved in

pathogenesis, its function should be flirther explored after expression in alternative

 

systems such as yeast or a high GC organism.

When I began this research, a homologue to deA had just been identified,
nothing was known about the Specific role of any amino acid in deA, and little was
known about the substrate specificity of deA or the residues involved in governing that
substrate specificity. Four years later, numerous alternative substrates of deA have been
identified that may be of potential biosynthetic importance (Chapter 2), a model for the
structure of deA has been Obtained based on the crystal structure coordinates of a ~30%
identical homologue (Chapter 3), and this model has been used to identify potential
substrate ligands which were confirmed by Site-directed mutagenesis techniques (Chapter

4).

108

REFERENCES

109

 

 

b)

10.

The Bordetella pertussis genome sequence data was produced by the Pathogen
Sequencing Unit at the Sanger Centre, and can be accessed from
http:l/www.sanger.ac.uk/Projects/B_pertussis.

The Bordetella parapertussis genome sequence data was produced by the
Pathogen Sequencing Unit at the Sanger Centre.

The Bordetella bronchiseptica genome sequence data was produced by the
Pathogen Sequencing Unit at the Sanger Centre.

Biologic and economic benefits from the use of phenoxy herbicides in the United
States. NAPIAP l-PA-96. U. S. Department of Agriculture.

Aravind, L., and E. V. Koonin. 2001. The DNA-repair protein AlkB, EGL-9, and
leprecan define new familes of 2-Oxoglutarate- and iron-dependent dioxygenases.
Genome Biol. 2(3):research0007.1-0007.8.

Archelas, A., and R. Furstoss. 1997. Synthesis of enantiopure epoxides through
biocatalytic approaches. Annu. Rev. Microbiol. 51:491-525.

Audus, L. J. 1952. The decomposition of 2,4-dichlorophenoxyacetic acid and 2-
methyl-4-chlorophenoxyacetic acid in the soil. J. Sci. Food Agric. 3:268.
Ausubel, F ., R. Brent, R. Kingston, D. Moore, J. Seidman, I. Smith, and K. Struhl.
1987. Current protocols in molecular biology. John Wiley and Sons, New York.
Baggaley, K. H., A. G. Brown, and C. J. Schofield. 1997. Chemistry and
biosynthesis of clavulanic acid and other clavarns. Nat. Prod. Rep. l4(4):309-333.
Baldwin, J. E. 1990. The biosynthesis of penicillins and cephalosporins. J.

Heterocycl. Chem 27:71-78.

110

 

ll.

13.

14.

15.

16.

17.

18.

Baldwin, J. E., and E. Abraham. 1988. The biosynthesis of penicillins and.
cephalosporins. Nat. Prod. Rep. 5(2): 129-145.

Bell, G. R. 1957. Some morphological and biochemical characteristics of a soil
bacterium which decomposes 2,4-dichlorophenoxyacetic acid. Can. J. Microbiol.
3 :82 1-840.

Bulinski, D. A., and C. H. Nakatsu. 1998. Involvement of a transposon in the
dissemination of 2,4-D catabolic genes. 8th International Symposium on
Microbial Ecology, 9-14 August, Halifax, N.S.:115.

Burzlaff, N. 1., P. J. Rutledge, I. J. Clifton, C. M. H. Hensgens, M. Pickford, R.
M. Adlington, P. L. Roach, and J. E. Baldwin. 1999. The reaction cycle of
isopenicillin N synthase observed by X-ray diffraction. Nature 401(6754):721-
724.

Cadwell, R. C., and G. F. Joyce. 1992. Randomization Of genes by PCR
mutagenesis. PCR Methods and Applications 2:28-33.

Clifton, I. J., L. C. Hsueh, J. E. Baldwin, K. Harlos, and C. J. Schofield. 2001.
Structure of proline 3-hydroxylase - Evolution of the family of 2-Oxoglutarate
dependent oxygenases. Eur J Biochem 268(24):6625-6636.

Denis, J.-N., A. E. Greene, A. A. Serra, and M.-J. Luche. 1986. An efficient,
enantioselective synthesis of the taxol side chain. J. Org. Chem. l986(51):46-50.
DiGennaro, P., A. Colmegna, E. Galli, G. Sello, F. Pelizzoni, and G. Bestetti.
1999. A new biocatalyst for production of optically pure aryl epoxides by styrene
monooxygenase from Pseudomonasfluorescens ST. Appl. Environ. Microbiol.

65(6):2794—2797.

111

 

 

19.

20.

21.

22.

23.

24.

25.

DiGiovanni, G. D., J. W. Neilson, I. L. Pepper, and N. A. Sinclair. 1996. Gene
transfer ofAlcaIigenes eutrophus JMP134 plasmid pJP4 to indigenous soil
recipients. Appl. Environ. Microbiol. 62(7):2521-2526.

Ditzelmuller, G., M. Loidl, and F. Streichsbier. 1989. Isolation and
characterization of a 2,4-dichlorophenoxyacetic acid-degrading soil bacterium.
Appl. Microbiol. Biotechnol. 31:93-96.

Don, R. H., and J. M. Pemberton. 1981. Properties of six pesticide degradation
plasmids isolated flom Alcaligenes paradoxus and Alcaligenes eutrophus. J.
Bacteriol. 145(2):681-686.

Efiink, M. R. 1997. Fluorescence methods for studying equilibrium
macromolecule-ligand interactions, p. 221-257. In L. Brand and M. L. Johnson
(ed.), Methods in Enzymology: Fluorescence spectroscopy, vol. 278. Academic
Press, New York.

Ege, S. N. 1994. Organic chemistry: Structure and reactivity, 3rd ed. D. C. Heath
and Company, Toronto.

Eichhom, E., J. R. van der Ploeg, M. A. Kertesz, and T. Leisinger. 1997.
Characterization of Ot-ketoglutarate-dependent taurine dioxygenase from
Escherichia coli. J. Biol. Chem. 272(37):23031-23036.

Elkins, J. M., M. J. Ryle, I. J. Clifton, J. C. Dunning Hotopp, J. S. Lloyd, N. I.
Burzlaff, J. E. Baldwin, R. P. Hausinger, and P. L. Roach. 2002. X-Ray crystal
structure of Escherichia coli taurine/Ot-ketoglutarate dioxygenase complexed to

ferrous iron and substrates. Biochemistry 41(16):5185-5192.

112

~1- .IF

 

 

26. -

27.

28.

29.

30.

31.

32.

33.

34.

Evans, W. C., B. S. W. Smith, H. N. Fernley, and J. I. Davies. 1971. Bacterial
metabolism of 2,4-dichlorophenoxyacetate. Biochem. J. 122:543-551.

Faulkner, J. K., and D. Woodcock. 1964. Metabolism of 2,4-
dichlorophenoxyacetic acid ('2,4-D') by Aspergillus niger van Tiegh.
Nature(4947):865.

Feig, A. L., and S. J. Lippard. 1994. Reactions of non-heme iron(II) centers with
dioxygen in biology and chemistry. Chem. Rev. 94:759-805.

F ukumori, F ., and R. P. Hausinger. 1993. Alcaligenes eutrophus JMP134 "2,4-
dichlorophenoxyacetate monooxygenase" is an or-ketoglutarate—dependent
dioxygenase. J. Bacteriol. 175(7):2083-2086.

Fukumori, F., and R. P. Hausinger. 1993. Pmification and characterization of 2,4-
dichlorophenoxyacetate/a-ketoglutarate dioxygenase. J. Biol. Chem.
268(32):24311-24317.

Fulthorpe, R. R., C. McGowan, O. V. Maltseva, W. E. Holben, and J. M. Tiedje.
1995. 2,4-Dichlorophenoxyacetic acid-degrading bacteria contain mosaics of
catabolic genes. Appl. Environ. Microbiol. 61(9):3274-3281.

Fulthorpe, R. R., A. N. Rhodes, and J. M. Tiedje. 1996. Pristine soils mineralize
3-chlorobenzoate and 2,4-dichlorophenoxyacetate via different microbial
populations. Appl. Environ. Microbiol. 62(4):] 159-1166.

Garraway, J. L., and R. L. Wain. 1976. The design of auxin-type herbicides. Drug
Design 72115-164.

Gentile, A., and C. Giordano. 1992. The enzymatic preparation of (2R, 3S) phenyl

glycidic acid Esters. J. Org. Chem. 57:6635-6637.

113

 

 

35.

36.

37.

38.

39.

40.

41.

Ghosal, D., and l.-S. You. 1988. Nucleotide homology and organization of
chlorocatechol genes of plasmids pJP4 and pAC27. Mol. Gen. Genet. 211:113-
120.

Ghosal, D., and I.-S. You. 1989. Operon structure and nucleotide homology of the
chlorocatechol oxidation genes of plasmids pJP4 and pAC27. Gene 83:225-232.
Ghosal, D., I.-S. You, D. K. Chatterjee, and A. M. Chakrabarty. 1985. Genes
specifying degradation of 3-chlorobenzoic acid in plasmids pAC27 and pJP4.
Proc. Natl. Acad. Sci., USA 82:1638-1642.

Greer, L., J. Robinson, and D. Shelton. 1992. Kinetic comparison of seven strains
of 2,4-dichlorophenoxyacetic acid-degrading bacteria. Appl. Environ. Microbiol.
58:1027-1030.

Gribble, G. W. 1992. Naturally occurring organohalogen compounds-A survey.
J. Nat. Prod. 55(10):]353-1395.

Hausinger, R. P., F. Fukumori, D. A. Hogan, T. M. Sassanella, Y. Kamagata, H.
Takami, and R. E. Saari. 1997. Biochemistry of 2,4-dichlorophenoxyacetic acid
(2,4-D) degradation: Evolutionary implications, p. 35-51. In K. Horikoshi, M.
Fukuda, and T. Kudo (ed.), Microbial Diversity and Genetics of Biodegradation.
Karger, New York.

Hee, H.-J., M. D. Lloyd, 1. J. Clifton, K. Harlos, A. Dubus, J. E. Baldwin, J .-M.
Frere, and C. J. Schofield. 2001. Alteration of the co-substrate selectivity of
deacetoxycephalosporin G synthase--The role of arginine 258. J. Biol. Chem.

276:18290—18295.

114

 

42.

43.

45.

46.

47.

48.

49.

50.

Hegg, E. L., and L. Que. 1997. The 2-his-1-carboxylate facial triad: an emerging
structural motif in mononuclear non-heme iron (II) enzymes. Eur. J. Biochem.
250:625-629.

Hegg, E. L., A. K. Whiting, R. E. Saari, J. McCracken, R. P. Hausinger, and L.
Que, Jr. 1999. Herbicide-degrading Ot-keto acid-dependent enzyme deA: metal
coordination environment and mechanistic insights. Biochemistry 38: 16714-
16726.

Hogan, D. A. 2001. Personal communication.

Hogan, D. A., T. A. Auchtung, and R. P. Hausinger. 1999. Cloning and
characterization of a sulfonate/ct-ketoglutarate dioxygenase fl'om Saccharomyces
cerevisiae. J. Bacteriol. 181 :5876-5879.

Hogan, D. A., D. H. Buckley, C. H. Nakatsu, T. M. Schmidt, and R. P. Hausinger.
1997. Distribution of the tfdA gene in soil bacteria that do not degrade 2,4-
dichlorophenoxyacetic acid (2,4-D). Microb. Ecol. 34:90-96.

Hogan, D. A., S. Smith, E. Saari, J. McCracken, and R P. Hausinger. 2000. Site-
directed mutagenesis of 2,4-dichlorophenoxyacetic acid/Ot-ketoglutarate
dioxygenase. J. Biol. Chem. 275(17):12400-12409.

Itoh, K. 2001. Personal communication.

Jones, F. D. May 4, 1945. U. S. patent 2,390,941.

Ka, J. O., W. E. Holben, and J. M. Tiedje. 1994. Analysis of competition in soil
among 2,4-dichlorophenoxyacetic acid-degrading bacteria. Appl. Environ.

Microbiol. 60:1121-1128.

115

 

51.

52.

53.

54.

55.

56.

57.

58.

Ka, J. 0., W. E. Holben, and J. M. Tiedje. 1994. Genetic and phenotypic diversity
of 2,4-dichlorophenoxyacetic acid (2,4-D)-degrading bacteria isolated from 2,4-
D-treated field soils. Appl. Environ. Microbiol. 60:] 106-1 1 15.

Ka, J. 0., W. E. Holben, and J. M. Tiedje. 1994. Use of gene probes to aid in
recovery and identification of functionally dominant 2,4-dichlorophenoxyacetic
acid-degrading populations in soil. Appl. Environ. Microbiol. 60:1116-1120.
Ka, J. 0., and J. M. Tiedje. 1994. Integration and excision of a 2,4-
dichlorophenoxyacetic acid-degradative plasmid in Alcaligenes paradoxus and
evidence Of its natural intergeneric transfer. J. Bacteriol. 176(17):5284-5289.
Kahnert, A., and M. A. Kertesz. 2000. Characterization of a sulfur-regulated
oxygenative alkylsulfatatse from Pseudomonas putida S-313. J. Biol. Chem.
275(41):31661-31667.

Kamagata, Y., R. R. Fulthorpe, K. Tamura, H. Takami, L. J. Fomey, and J. M.
Tiedje. 1997. Pristine environments harbor a new group of Oligotrophic 2,4-
dichlrophenoxyacetic acid-degrading bacteria. Appl. Environ. Microbiol.
63(6):2266-2272.

Katase, T., and R. Kondo. 1989. Vertical profiles Of trans- and cis-4-
hydroxycinnamic acids and other phenolic acids in Horonobe peat soils, Japan.
Soil Science 148(4):258-264.

Kathariou, S. 2001. Personal communication.

Kilpi, S. 1980. Degradation of some phenoxyacid herbicides by mixed cultures of
bacteria isolated from soil treated with 2-(2-methyl-4-chloro)phenoxypropionic

acid. Microb. Ecol 6:261-270.

116

 

59.

60.

61.

62.

63.

64.

65.

66.

King, R. J ., K. A. Short, and R J. Seidler. 1991. Assay for detection and .
enumeration of genetically engineered microorganisms which is based on the
activity of a deregulated 2,4-dichlorophenoxyacetate monooxygenase. Appl.
Environ. Microbiol. 27(6): 1790-1792.

Lai, M.-t., W. Wu, and J. Stubbe. 1995. Characterization of a novel, stable
norcaradiene adduct resulting from the inactivation of thymine hydroxylase y 5-
ethynyluracil. J. Am. Chem. Soc. 117:5023-5030.

Leveau, J. H. J., A. J. B. Zehnder, and J. R van der Meer. 1998. The tde gene
product facilitates uptake of 2,4-dichlorophenoxyacetate by Ralstonia eutropha
JMP134(pJP4). J. Bacteriol. 180(8):2237-2243.

Liu, A., R- Y. N. Ho, and L. Que, Jr. 2001. Alternative reactivity of an or-
ketoglutarate-dependent iron(II) oxygenase: enzyme self-hydroxylation. J. Am.
Chem. Soc. 123:5126-5127.

Lloyd, M. D., K. D. Merritt, V. Lee, T. J. Sewell, B. Wha-Son, J. E. Baldwin, C.
J. Schofleld, S. W. Elson, K. H. Baggaley, and N. H. Nicholson. 1999. Product-
substrate engineering by bacteria: Studies on Clavaminate synthase, 3 trifunctional
dioxygenase. Tetrahedron 55(33): 10201-10220.

Lontz, J. F. February 20, 1942. U. S. patent 2,322,761.

Loos, M. A., R. N. Roberts, and M. Alexander. 1967. Phenols as intermediates in
the decomposition of phenoxyacetates by an Arthrobacter species. Can. J.
Microbiol. 13:679-690.

Mac, A. A., R. O. Marits, N. R. Ausmees, V. M. KOiv, and A. L. Heinaru. 1993.

Characterization of a new 2,4-dichlorophenoxyacetic acid degrading plasmid

117

 

 

67.

68.

69.

70.

71.

72.

pEST4011: Physical map and localization of catabolic genes. J. Gen. Microbiol.
139:3165-3170.

Mai, P., O. S. Jacobsen, and J. Aamand. 2001. Mineralization and co-metabolic
degradation of phenoxyalkanoic acid herbicides by a pure bacterial culture
isolated from an aquifer. Appl. Microbiol. Biotechnol. 56:486-490.

Miroux, B., and J. E. Walker. 1996. Over-production Of proteins in Escherichia
coli: mutant hosts that allow synthesis of some membrane proteins and globular
proteins at high levels. J. Mol. Biol. 260:289-298.

Mukherji, M., W. Chien, N. J. Kershaw, I. J. Clifton, C. J. Schofield, A. S.
Wierzbicki, and M. D. Lloyd. 2001. Structure-function analysis of phytanoyl-COA
2-hydroxylase mutations causing Refsum's disease. Hum Mol Genet 10(1 8): 1971-
1982.

Nakatsu, C. H., R. Korona, R. E. Lenski, F. J. de Bruijn, T. L. Marsh, and L. J.
Fomey. 1998. Parallel and divergent genotypic evolution in experimental
populations of Ralstonia sp. J. Bacteriol. 180(17):4325-4331.

Nakatsu, C. H., R. Korona, R. E. Lenski, F. J. deBruijn, T. L. Marsh, and L. J.
Fomey. 1998. Parallel and divergent genotypic evolution in experimental
populations of Ralstonia sp. J .Bacteriol. 180(17):4325-4331.

Nielson, J. W., K. L. Josephson, I. L. Pepper, R. P. Arnold, G. D. DiGiovanni,
and N. A. Sinclair. 1994. Frequency of horizontal gene transfer of a large

catabolic plasmid (pJP4) in soil. Appl. Environ. Microbiol. 60(11):4053-4058.

118

“flit!

 

73.

74.

75.

76.

77.

78.

79.

80.

81.

Nurk, A., L. Kasak, and M. Kivisaar. 1991. Sequence of the gene (pheA)
encoding phenol monooxygeanse from Pseudomonas Sp. ESP 1001: expression in
Escherichia coli and Pseudomonas putida. Gene 102:13-18.

Pemberton, J. M., B. Corney, and R. H. Don. 1979. Evolution and spread of
pesticide degrading ability among soil micro-organisms, p. 287-299. In K. N.
Timrrris and A. Puhler (ed.), Plasmids of Medical, Environmental, and
Commercial Importance. Elsevier/North-Holland Biochemical Press, Amsterdam.
Pemberton, J. M., and P. R. Fisher. 1977. 2,4-D plasmids and persistence. Nature
268:732-733.

Ponder, M. 2001. Personal communication.

Prescott, A. G. 1993. A dilemma of dioxygenases (or where biochemistry and
molecular-biology fail to meet). J. Exp. Bot. 44:849.

Prescott, A. G., and M. D. Lloyd. 2000. The iron (II) and 2-oxoacid-dependent
dioxygenases and their role in metabolism. Nat. Prod. Rep. 17:367-3 83.

Que Hee, S. S., and R. G. Sutherland. 1981. The phenoxyalkanoic herbicides:
volume I chemistry, analysis, and environmental pollution. CRC Press, Inc., Boca
Raton, Florida.

Roach, P. L., I. J. Clifton, V. Fulop, K. Harlos, G. J. Barton, J. Hajdu, I.
Andersson, C. J. Schofield, and J. E. Baldwin. 1995. Crystal structure of
isopenicillin N synthase is the first from a new structural family of enzymes.
Nature 37 5:700-704.

Rondon, M. R., P. R. August, A. D. Betterrnann, S. F. Brady, T. H. Grossman, M.

R. Liles, K. A. Loiacono, B. A. Lynch, 1. A. MacNeil, C. Minor, M. Gilrnan, C. L.

119

'-'. ‘4 "ii—cg

 

 

82.

83.

84.

85.

86.

87.

88.

Tiong, M. S. Osbume, J. Clardy, J. Handelsman, and R. M. Goodman. 2000.
Cloning the soil metagenome: a strategy for accessing the genetic and functional
diversity of uncultured microorganisms. Appl. Environ. Microbiol. 66(6):2541-
2547.

Rosazza, J. R, Z. Huang, L. Dostal, T. Volrn, and B. Rousseau. 1995. Review:
biocatalytic transformations of fenrlic acid: An abundant aromatic natural
product. J. Ind. Microbiol. 15(6):457-471.

Rost, B., and C. Sander. 1993. Prediction of protein secondary structure at better
than 70-percent accuracy. Journal of Molecular Biology 232(2):584-599.

Ryle, M. J ., R. Padmakumar, and R. P. Hausinger. 1999. Stopped-flow kinetic
analysis of Escherichia coli taurine/Ot-ketoglutarate dioxygenase: interactions
with (It-ketoglutarate, taurine, and oxygen. Biochemistry 38:15278-15286.

Saari, R. E., and R. P. Hausinger. 1998. Ascorbic acid-dependent turnover and
reactivation of 2,4-dichlorophenoxyacetic acid/Ot-ketoglutarate dioxygenase using
thiophenoxyacetic acid. Biochemistry 37:3035-3042.

Saari, R. E., D. A. Hogan, and R. P. Hausinger. 1999. Stereospecific degradation
of the phenoxypropionate herbicide dichlorprop. J. Molec. Catalysis B: Enzymatic
62421-428.

Sali, A., and T. L. Blundell. 1993. Comparative protein modelling by satisfaction
of Spatial restraints. J. Mol. Biol. 234:779-815.

Sandmann, E. R. I., and M. A. Loos. 1988. Aromatic metabolism by a 2,4-D

degrading Arthrobacter sp. Can. J. Microbiol. 34: 125-130.

120

“1‘1 .313: .9.-

 

 

89.

90.

91.

92.

93.

94.

95.

96.

Seffernick, J. L., M. L. deSouza, M. J. Sadowsky, and L. P. Wackett. 2001.
Melamine deaminase and atrazine chlorohydrolase: 98 percent identical but
functionally different. J. Bacteriol. 183(8):2405-2410.

Segura, A., P. V. Bunz, D. A. D'Argenio, and L. N. Omston. 1999. Genetic
analysis of a chromosomal region containing vanA and vanB, genes required for
conversion of either ferulate or vanillate to protocatechuate in Acinetobacter. J.
Bacteriol. 181(11):3494-3504.

Stanier, R, N. Palleroni, and M., Doudoroff. 1966. The aerobic Pseudomonads: A
taxonomic study. J. Gen. Microbiol. 43: 159-271.

Steenson, T. 1., and N. Walker. 1957. The pathway of breakdown of 2,4-dichloro-
and 4-chloro-2-methyl-phenoxyacetic acid by bacteria. J. Gen. Microbiol. 16: 146-
155.

Strack, D. 1997. Phenolic metabolism, p. 387. In P. M. Dey and J. B. Harbone
(ed.), Plant Biochemistry. Academic Press, San Diego.

Streber, W. R, K. N. Tirnrnis, and M. H. Zenk. 1987. Analysis, cloning, and high-
level expression of 2,4-dichlorophenoxyacetate monoxygenase gene tfdA of
Alcaligense eutrophus JMP134. J. Bacteriol. 169:2950-2955.

Studier, F. W., A. H. Rosenberg, J. J. Dunn, and J. W. Dubendorff. 1990. Meth.
Enzymol. 185:60-89.

Suwa, Y., A. D. Wright, F. Fukimori, K. A. Nummy, R P. Hausinger, W. E.
Holben, and L. J. Fomey. 1996. Characterization of a chromosomally encoded
2,4-dichlorophenoxyacetic acid/Ot-ketoglutarate dioxygenase from Burkholderia

Sp. strain RASC. Appl. Environ. Microbiol. 62(7):2464-2469.

121

 

97.

98.

99.

100.

101.

102.

103.

Thomburg, L. D., and J. Stubbe. 1993. Mechanism-based inactivation of thymine
hydroxylase, an Ot-ketoglutarate-dependent dioxygenase, by S-ethynyluracil.
Biochemistry 32:14034-14042.

Thomburg, L. D., and J. Stubbe. 1989. Mechanism-based inhibition of thymine
hydroxylase. J. Am. Chem. Soc. 111:7632-7633.

Top, E. M., W. E. Holben, and L. J. F orney. 1995. Characterization Of diverse
2,4-dichlorophenoxyacetic acid-degradative plasmids isolated flom soil by F
complementation. Appl. Environ. Microbiol. 61(5):11691-1698. ,i
Top, E. M., O. V. Maltseva, and L. J. Fomey. 1996. Capture of a catabolic

plasmid that encodes only 2,4-dichlorophenoxyacetic acid: Ot-ketoglutaric acid '1

 

dioxygenase (T fdA) by genetic complementation. Appl. Environ. Microbiol.
62(7):2470-2476.

Valegard, K., A. Terwisscha van Scheltinga, C., M. D. Lloyd, T. Hara, S.
Ramaswarny, A. Perrakis, A. Thompson, H.-J. Lee, J. E. Baldwin, C. J. Schofield,
J. Hajdu, and I. Andersson. 1998. Structure of a cephalosporin synthase. Nature
394:805-809.

Vallaeys, T., L. Courde, C. McGowan, A. D. Wright, and R. R. Fulthorpe. 1999.
Phylogenetic analyses indicate independent recruitment of diverse gene cassettes
during assemblage of the 2,4-D catabolic pathway. FEMS Microbiol. Ecol.
28(1999):373-382.

Verrneij, P., C. Wietek, and M. Kertesz. 2000. Characterization of a sulfur-
regulated oxygenative alkylsulfatase from Pseudomonas purtida S-313. J. Biol.

Chem. 275(41):31661-31667.

122

 

104.

105.

106.

107.

108.

109.

Watanabe, M., N. Sumida, S. Murakanri, H. Anzai, C. J. Thompson, Y. Tateno,
and T. Murakarrri. 1999. A phosphonate-induced gene which promotes
Penicillium-mediated bioconversion of cis-Propenylphosphonic acid to
fosfomycin. Appl. Environ. Microbiol. 65(3):1036-1044.

Whiting, A. K., R. E. Saari, M. A. Fredrick, J. McCracken, R. P. Hausinger, and

L. Que. 1997. Metal coordination environment of a Cu(II)-substituted Ot-keto

acid-dependent dioxygenase that degrades the herbicide 2,4-D. J. Am. Chem. Soc.

119(14):34l3-3414.

Wilmouth, R, J. Tumbull, R. Welford, I. J. Clifton, A. G. Prescott, and C. J.
Schofield. 2002. Structure and mechanism of anthocyanidin synthase from
Arabidopsis thaliana. Structure 10(1):93-103.

Xia, X.-S., S. Aathithan, K. Oswiecirnska, A. R. W. Smith, and I. J. Bruce. 1998.
A novel plasmid pIJBl possessing a putative 2,4-dichlorophenoxyacetate,
degradative transposon Tn5530 in Burkholderia cepacia strain 2a. Plasmid
39:154-159.

Yano, T., S. One, and H. Kagarniyama. 1998. Directed evolution of an aspartate
arrrinotransferase with new substrate specificities. Proc. Natl. Acad. Sci. USA
95:5511-5515.

Zhang, Z., J. Ren, D. K. Stammers, J. E. Baldwin, K. Harlos, and C. J. Schofield.
2000. Structural origins of the selectivity of the trifunctional oxygenase

clavarninic acid synthase. Nature Struct. Biol. 7(2): 127-133.

123

13'

 

 

 

APPENDIX A

DEVELOPMENT OF METHODS TO EVOLVE A mill-like GENE

124

 

 

The tfd genes appear to have been recruited from multiple metabolic pathways
onto a single plasmid backbone to generate a plasmid capable of conferring to the
organism the ability to degrade 2,4-D (40). Homologues of the tdeCDE genes are known
to encode proteins that utilize naturally occurring non-chlorinated substrates related to the
2,4-D degradative intermediates (35, 36, 73), but the substrate of the environmental
homologues of deA have not been identified. Based on PCR amplification studies, y'dA-
like sequences are widespread in the environment but the organisms containing these
putative genes generally do not degrade 2,4-D (46). F mthermore, these tfdA-like 1

sequences have been present in the environment for over a thousand years based on my

 

analysis of Siberian permafrost soil isolates (Chapter 5). I hypothesize that tfdA-like E
genes in nature work on a substrate Similar in structure to 2,4-D and that, through a small
number of mutations, the gene was able to mutate into a tfdA gene encoding the 2,4-D
degrading enzyme. I suggest that this process Should be able to be reproduced artificially
by directed evolution. For example, by artificially introducing small numbers of point
mutations into a tfdA -like gene and screening the constructs for increased activity with
2,4-D it may be possible to obtain reasonable deA activity after several rounds.

To attempt to mimic the natural evolution of deA, I used error prone PCR to
introduce small numbers of point mutations into Bordetella pertussis and Bradyrhizobia
Sp. g'dA-like genes. The B. pertussis and Bradyrhizobia Sp. genes were chosen as
candidates for directed evolution because the encoded proteins are closely related to

deA, yet do not prefer 2,4-D as a substrate.

125

 

EXPERIMENTAL

Media and cultures. Escherichia coli DHSO. (pMAL-tfdA) carrying tfdA on the
pMAL plasmid (47) was used for experimental positive controls. E. coli DHSa (pMAL-
brad) containing a Bradyrhizobia sp. tfdA -like gene, graciously supplied by Kazuhito Ito,
and E. coli DHSa (pMAL-BP1665), containing a Bordetella pertussis tfdA-like sequence
previously described (Chapter 6), were used as sources of plasmid for the directed
evolution studies. The proteins encoded for by the plasmids had an N-terminal maltose
binding protein fusion. All cultures were grown in LB or on LB containing 1.5% agarose.

Random mutagenesis. Plasmid DNA was isolated fiom overnight cultures using

Promega Wizard preps. The tfdA-like genes were mutagenized by amplification of the

 

corresponding pMAL plasmid by the method of Cadwell and Joyce (15) with the New
England Biolabs malE and M13 primers in 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 7
mM MgC12, 0.5 mM MnClz, 5 mM dCTP, 5 mM dTTP, 1 mM dATP and 1 mL dGTP, 1
uM each primer, and 0.05 units/1.1L Gibco T aq DNA polymerase. Non-mutagenic
amplification was carried out with the same primers in 20 mM Tris-HCl (pH 8.4), 50 mM
KCl, 2.5 mM MgC12, 1.25 mM each dNTP (Gibco), 1 pM each primer, and 0.05 units/uL
Gibco Taq DNA. Initial denaturation occurred at 94 °C for 5 min, followed by elongation
for 35 cycles of denaturation for 1 min at 95 °C, annealing at 55 °C for 45 sec, and
elongation at 72 °C for 45 see, with a final 10 min elongation step at 72 °C.

Screening. The resulting PCR products were cleaved with HindIII and Xbal, and
inserted into pMAL-c2 cut with the same enzymes to create a libraryof mutagenized
plasmids. This library of mutagenized plasmids was transformed into Gibco DH5a MAX

Emciency cells so as to yield approximately 1000 colony forming units on a 150 mm plate.

126

 

The colonies were transferred to sterile nitrocellulose filters, and the filters were soaked
in 50 mM Tris buffer (pH 7) with the colony side up for 30 min. (It should be noted that
the filters were not soaked in IPTG to induce the cells; instead, the plate assay relied on
the leakiness of the tac promoter when cells are grown on LB not amended with glucose.)
The filters were transferred to a shallow tray containing 10 mM imidazole, 1 mM aKG, 1
mM 2,4-D, 10 uM ferrous ammonium sulfate, and 10 uM ascorbate and soaked for 10
min. 4-Aminoantipyrene was added to 0.02%, pH 10 buffer to 0.1X, NaOH to adjust the
pH to 9.0, and 0.08% potassium ferricyanide (59). The color was developed for 5 min.
Areas identified as having more activity than non-mutagenized colonies were marked on
the original plates. The original plates were re-grown and colonies from that region of the
plate were streaked for isolation and re-screened by the filter assay to identify the exact
colony with enhanced activity.

Activity analysis. Cell lysates were prepared as previously described (Chapter 2)
for isolates identified as having enhanced activity as well as for E. coli (pMAL-brad) and
E. coli (pMAL-tfdA) as controls. Cell lysates were prepared from both static cultures and
well-aerated cultures. Purified proteins were obtained by passing the cell lysates over an
amylose column as previously described (47). Cell lysates and purified proteins were
assayed by using the 4-aminoantipyrene assay (59) and proteins concentrations were
determined by using a commercial protein assay (BioRad) with bovine serum albumin as
a standard.

Plate matings. Overnight cultures of E. coli DH50L (pACYC) in LB amended with

34 ug/mL chlorarnphenicol and R. eutropha JMP228 (pBH501aE) in LB amended with

50 pg/mL kanamycin and 100 ug/mL rifampin were diluted 1000-fold into 10% LB. LB

127

 

 

plates were prepared with only 1 mL of the R. eutropha JMP228 (pBH501aE) dilution, 1
mL of the E. coli DHSa (pACYC) dilution, or 1 mL of the R eutropha JMP228
(pBH501aE) combined with 1 mL of the E. coli DHSa (pACYC) dilution. After
incubating overnight at 30 °C, the plates were scraped into 5 mL sterile PBS, serially
diluted and plated on LB amended with 34 ug/mL chlorarnphenicol, 50 ug/mL

kanamycin, and 100 ug/mL rifampin. The colonies were counted after overnight

incubation at 30 0C.

128

 

 

RESULTS AND DISCUSSION

The B. pertussis and Bradyrhizobia sp. tfdA-like genes were chosen as candidates
for directed evolution because the encoded proteins are closely related to deA (Figure
6.1), yet they do not prefer 2,4-D as a substrate. The maltose binding protein fusions of
the B. pertussis deA-like protein had no detectable activity with 2,4-D and the
Bradyrhizobia sp. deA-like protein degraded 0.03 umol of 2,4-D min'l mg’1 (48). This
value corresponds to 0.4% of the activity of the maltose binding protein fusion of wild-
type deA (30). Consistent with these results with purified proteins, cells producing the
B. pertussis protein did not turn red, those producing the Bradyrhizobia protein turned
pink, and those producing the R. eutropha deA protein turned dark red.

Once I confirmed that the colorimetric plate assay could be used to screen for
colonies with increased activity towards 2,4-D as substrate, the B. pertussis and
Bradyrhizobia sp. tfdA -1ike genes were mutagenized by PCR mutagenesis. I utilized a
method involving the addition of manganese and an altered dNTP ratio since this
procedure was determined to maximize the mutation rate without bias for certain types of
mutations (15). An initial round of mutagenesis and cloning was carried out with the B.
pertussis tfdA-like gene, the Bradyrhizobia #dA -like gene, and the VdA gene. From each
set, six colonies were picked and grown in LB. Fourteen of the eighteen cultures
contained plasmids of the expected size for vector with insert (Table A. 1). Of the other
four, one plasmid was larger than expected, one was smaller than expected, and two did
not yield plasmid. Of the fourteen cultures containing appropriately sized plasmids, seven

produced appropriately sized proteins, five produced truncated proteins, and two

129

.._ rm...“ Wit-imam

 

 

Table A.1 Characteristics of selected mutagenized clones.

 

 

 

 

 

Expected Insert Clone Expression Insert
Number (assessed on SDS-PAGE (assessed on agarose gel)
gel)
pMAL-brad 1 Full length Yes
(mutagenized) 2 Full length Yes
3 Full length No plasmid
4 No Smaller than expected
5 Truncated Yes
6 Full length Yes
pMAL-BP1665 1 Full length Yes
(mutagenized) 2 Almost firll length Yes
3 Almost full length Yes
4 No No plasmid
5 Truncated Yes
6 No Yes
7 Full length Yes
pMAL-tfdA 1 Full length Bigger than expected
(mutagenized) 2 Full length Yes
3 Almost full length Yes
4 No Yes
5 Full length Yes
pMAL-tfdA Full length Yes

 

130

 

 

did not produce any deA-like protein. Although there seems to be a disproportionate
number of truncations the mutagenesis was deemed successful and a larger screening was
undertaken.

Six plates, two for the mutagenized Bradyrhizobia gene, two for the mutagenized
Bordetella gene, and two for the mutagenized tfdA, were prepared with approximately
one thousand colonies each, replica filtered, and assayed. The plates containing the
mutagenized B. pertussis gene yielded no pink or red colonies. Approximately seven
percent of the colonies containing mutagenized ddA turned red. The plates containing
colonies with the mutagenized Bradyrhizobia gene contained twenty areas that were red
and numerous more areas that were pink. The twenty red areas were picked, streaked, and
reanalyzed by the assay. Of these, four were deemed to be active. A colony from each of
these was further analyzed (1-4-6-4, 1-4-7-4, 1-4-8-2, and 1-4-9-2).

Shaken and static cultures of 1-4-6-4, 1-4-7—4, 1-4-8-2, 1-4-9-2, E. coli (pMAL-
brad), and E. coli (pMAL-tfdA) were grown, cell lysates were prepared, and the proteins
were purified by amylose column chromatography. No activity was detected in the
purified proteins from the mutagenized clones or the non-mutagenized E. coli (pMAL-
brad) cultures. In contrast, the control protein was active as expected (47).

As an alternative to the phenol staining assay approach, I explored using R.
eutropha JMP228 (pBH501aE) for screening the evolved proteins. R eutropha JMPZZS
(pBH501aE) is a derivative of R. eutropha J MP1 34 (pJP4) that contains the tdeCDEF
genes, but lacks tfdA. I showed that pACYCl84 is compatible with the plasmid already in
the host and found that both plasmids could be maintained. Initially studies with R.

eutropha JMP228 (pBH501aE) and E. coli (pACYCl 84) confirmed that pACYCl 84 was

 

 

mobilized into R. eutropha JMP228 (pBH501aE) during plate matings with a yield of
about 6700 transforrnants per plate mating. Six attempts were carried out using four
different locations for inserting the Bradyrhizobia sp. q'dA -like gene into pACYC184;
none were successful. As a control, the same attempts were made with tfdA, but these
were also unsuccessful. The final attempt included containing different antibiotics on
each piece of DNA being ligated and screening with both antibiotics. No appropriately
sized transformants could be obtained in any of the cases. I presume that the genes are
not stable in the plasmid.

Kazuhito Ito has cloned the Bradyrhizobia sp. tfdA-like gene into a different

plasmid and a construct of R. eutropha JMP228 (pBH501aE) containing this plasmid has

 

been made and is currently being characterized (48). Future studies could be carried out
in that construct if the plasmid and gene are stable and if protein isolated from those cells

dimerizes correctly.

 

132

llllllllllllllllll '