5.2..“

s). 5...:

.1 sh u.-

Hm.

may}: ,

{as »

u .... :w. .
gwfiusfl

.
.

Lu... .1

.....u.wm.u..v...lum.
.. .r. u.

v‘.
$1.35..“
. Iv.‘

-L 1‘ \l bl . ‘l
w. .u..ta....,.wucflt. Err mm:
19. ustvhm... new: . h".

ammu V2.4
A r ’1‘ I—VI ‘ l
.n “Ivan... «Nd. . 4

”5.1.1.3?

 

 

‘r ~: mg LIBRARY
Z Michigan State
2% f University

This is to certify that the
dissertation entitled

KINETIC/SPECTROSCOPIC INVESTIGATION OF TauD
INTERACTIONS WITH INHIBITORS AND
ISOLATION/CHARACTERIZATION OF THE PUTATIVE
Fen/a-KETOGLUTARATE DEPENDENT DIOXYGENASE
AND FLAVIN-DEPENDENT DEHYDROGENASE CsiD
AND YgaF.

presented by
Efthalia Kalliri

has been accepted towards fulfillment
of the requirements for the

Ph. D. degree in Chemistry

 

 

1W armed

Major Professor’s Sigufiure'

Ida/07

Date

MSU is an affirmative-action, equal-opportunity employer

4. —.-o-a--—O-l-O-I-n-I-l-I-I-h-O-l-l-t-l-v-l-I-b-n-I-A-o-l-

-._.-n--n--o---.—-- A

-.-.-—-—----.--a-u-n-c-o-—.— _

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K lProilAccapres/CIRCIDaleDue indd

KINETIC/SPECTROSCOPIC INVESTIGATION OF TauD INTERACTIONS WITH
INHIBITORS AND ISOLATION/CHARACTERIZATION OF THE PUTATIVE Fen/a-
KETOGLUTARATE DEPENDENT DIOXYGENASE AND FLAVIN-DEPENDENT
DEHYDROGENASE CsiD AND YgaF.

By

Efthalia Kalliri

A DISSERTATION

Submitted to
Michigan State University
In partial fulfillment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry

2008

ABSTRACT

KINETIC/SPECTROSCOPIC INVESTIGATION OF TauD INTERACTIONS

WITH INHIBITORS AND ISOLATION/CHARACTERIZATIONS OF THE PUTATIVE

F eu/a-KETOGLUTARIC DEPENDENT DIOXYGENASE AND FLAVIN-

DEPENDENT DEHYDROGENASE CsiD AND YgF.
By
Efthalia Kalliri

TauD, an Escherichia coli Fen/a-ketoglutarate (aKG)-dependent dioxygenase,
catalyzes the hydroxylation of aminoethanesulfonate (taurine) to produce an intermediate
that decomposes to provide sulfite as a cellular sulfur source. Complexes of TauD with
Co", Ni", and N-oxalylglycine (NOG) (known inhibitors of this class of enzymes) were
kinetically and spectroscopically investigated. The metal ions were shown to be slow-
binding competitive (with Fe") inhibitors of the enzyme, and substitution of C01' for Fell
was found to produce a chromophore that can serve as a diagnostic marker for aKG-
dependent dioxygenases. Kinetic studies revealed that NOG is a weak competitive (with
aKG) inhibitor of TauD. Despite the close structural similarity of NOG to aKG, the TauD-
Fen-NOG complex was incapable of activating oxygen or catalyzing NOG
decarboxylation.

CsiD is encoded by a gene (csiD) located upstream of the y-aminobutyric acid
(GABA) operon (gabDTP-csiR) in E. coli and is induced by carbon starvation. The crystal

structure of CsiD revealed it to be a putative Fell/aKG-dependent dioxygenase.

Spectroscopic studies of purified CsiD were used to prove that Fe" and aKG bind as a

chelate at its active site. An uncoupled reaction was observed in which aKG and oxygen
were consumed and succinate was produced in the absence of any primary substrate.
Comparisons between the E. coli wild type and a csiD-knockout strain were undertaken in
order to examine the function of CsiD. Numerous compounds were tested as potential
substrates of the putative dioxygenase by monitoring oxygen consumption with an oxygen
electrode assay, but the function of this protein remains unknown.

The ygaF gene of E. coli is located immediately downstream of csiD, with which it
is coregulated, and just upstream of the gabDTP-csiR operon. On the basis of sequence
comparisons, YgaF had been proposed to be a putative FAD-dependent oxidoreductase
with an unknown function. Cloning and overexpression of ygaF was used to produce the
YgaF protein for characterization. YgaF was shown to possess non-covalently bound
FMN, which undergoes a single-step, two-electron redox process (E° ~20 mV) during
photoreduction or titration with dithionite. Rapid reoxidation of reduced YgaF by oxygen
and formation of a characteristic flavin-sulfite complex provided evidence that it is an
oxidase. The function of YgaF was investigated via metabolic studies of E. coli wild-type
and ygaF-knockout cells together with examination of various compounds as substrates of

YgaF. The results identified YgaF as an L-2-hydroxyglutarate oxidase.

To my dearest father, mother, sister
&

Beloved fiance’ Kostas

iv

ACKNOWLEDGMENTS

I would like to thank many people for their help during my PhD. Firstly; I am
really grateful] to my advisor, Prof. Robert Hausinger, who accepted me in his group in
the middle of my PhD. Dr. Hausinger has provided a valuable help through the course of
my work with his continuous guidance, support and countless discussions on my projects.
Also, I would like to thank my former advisor, Prof. Joan Broderick, who had been an
excellent advisor because of the motivation that she inspired me the first years of my PhD.

Research is a team work and in order to be successful several people get involved.
Therefore, I would like to thank: Prof. Daniel Jones assisted with GC-MS studies, Prof.
David Ballou for the stopped flow studies on YgaF, Dr. Nicolai Burzlaff for providing
NOG, Dr. Christopher Lima for providing the pSMT3 contained the csiDl gene. Dr Tina
Muller involved in the CsiD project, Dr. Piotr Grzyska involved in spectroscopic studies
on the TauD protein, and Dr. Meng Li assisted with EPR spectroscopy.

I also thank the many graduate and undergraduate students I have worked with in
Prof. Hausinger’s group: Meng, Rachel, Jana, Soledad, Tina, Scott, Kim, Piotr, Andrea,
Bruce, Aaron, Kimberly and Melody. In addition I would like to thank the graduate
students in Prof. Broderick’s lab: Sophia, Meng, Magda, Egis, Jeff, Cliff, Mbako, Dan,
Sujuan, Yi, Washington and Jim. The existence of many good friends and helpful
labmates (in both labs) had enormous positive effect in all these five and a half years.

Being in a foreign country, friends substitute your family. I was very lucky to meet
Greek and international people as well, with whom I shared many happy moments:
Nineta, Chryssoula, Justas, Kit, Kyoungsoo, Maria and Marina. Especially I would like to

thank Nineta, my dearest friend, who was next to me any time I needed her.

Finally, I would like to thank my family. My father, mother, and sister for their
love, support and understanding, and my fiance Kostas with whom we did this journey

together.

vi

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................... ix
LIST OF FIGURES .................................................................................................. x
LIST OF SCHEMES ............................................................................................. xiii
ABBREVIATIONS ............................................................................................... xiv
CHAPTER 1; INTRODUCTION ............................................................................. 1
a-ketoglutarate dependent dioxygenases .................................................................. 2
Reactions of aKG-dependent dioxygenases and related enzymes .................... 2
General characteristics and mechanistic studies of the aKG-dependent
dioxygenases ...................................................................................................... 7
Flavoenzymes ......................................................................................................... 15
Structures of flavins ........................................................................................ 15
Diversity of flavoenzymes .............................................................................. 15
F lavin redox chemistry ................................................................................... 17
Classification of flavoenzymes ....................................................................... 25
Thesis outline .......................................................................................................... 36
REFERENCES ....................................................................................................... 39
CHAPTER 2; Kinetic and Spectroscopic Investigation of Co", Ni", and N-
Oxalylglycine Inhibition of the F en/a-Ketoglutarate Dioxygenase, TauD ............ 47
ABSTRACT .................................................................................................... 48
INTRODUCTION .......................................................................................... 49
EXPERIMENTAL PROCEDURES ............................................................... 52
RESULTS AND DISCUSSION ..................................................................... 55
CONCLUSIONS ............................................................................................. 63
REFERENCES ............................................................................................... 64
CHAPTER 3; Isolation and characterization of CsiD ............................................ 68
ABSTRACT .................................................................................................... 69
INTRODUCTION .......................................................................................... 70
EXPERIMENTAL PROCEDURES ............................................................... 74
RESULTS ....................................................................................................... 84
CONCLUSIONS ........................................................................................... 101
REFERENCES ............................................................................................. 106
CHAPTER 4; Cloning of the ygaF gene, and purification of the encoded protein.
............................................................................................................................... 109
ABSTRACT .................................................................................................. 110
INTRODUCTION ........................................................................................ 1 1 1
EXPERIMENTAL PROCEDURES ............................................................. 113

vii

RESULTS ..................................................................................................... 123
CONCLUSIONS ........................................................................................... 148
REFERENCES ............................................................................................. 153

viii

LIST OF TABLES

Table 1. List of the facial triad of several aKG-Ds ............................................................. 10
Table 2. List of amino acids and other compounds detected by amino acid analysis ....... 81

Table 3. List of the compounds tested as potential substrates of YgaF and the methods that
were used. ......................................................................................................................... 138

ix

LIST OF FIGURES

Figure 1. General reaction of the a-ketoglutarate-dependent dioxygenases (aKG-Ds). ...... 3
Figure 2. Prolyl 4-hydroxylase (P4H) reaction. .................................................................... 4
Figure 3. Reaction of AlkB. .................................................................................................. 5
Figure 4. CAS is involved in three out of four steps of clavulanic acid biosynthesis. ......... 5
Figure 5. Reaction of TauD ................................................................................................... 6
Figure 6. Reaction of HPPD. ................................................................................................ 7
Figure 7. Structure of the jelly-roll motif .............................................................................. 8
Figure 8. General mechanism of dKG-dependent dioxygenases. ......................................... 9
Figure 9. The structures of flavins. ..................................................................................... 16
Figure 10. Redox states of flavin. ....................................................................................... 17
Figure 11. Absorption spectra of the redox states of glucose oxidase. ............................... 18
Figure 12. Reductive and oxidative half reactions of flavoproteins. .................................. 19
Figure 13. Covalent protein-flavin bonds observed in flavoenzymes. ............................... 22
Figure 14. Photoreduction of AidB. .................................................................................... 24
Figure 15. Possible pathways of reaction of oxygen with reduced flavin. ......................... 26
Figure 16. The flavin-sulfite complex is a covalent N(5) adduct. ...................................... 27
Figure 17. Binding of sulfite to glycine oxidase from Bacillus subtilis. ............................ 27
Figure 18. Structure of the active site of sevelar flavoproteins .......................................... 29
Figure 19. Mechanism of external flavoprotein monooxygenases. .................................... 32
Figure 20. Reactions catalyzed by flavin monooxygenases. .............................................. 33
Figure 21. Titration of OYE with p-chlorophenol. ............................................................. 36
Figure 22. Inhibition of TauD by Co” and Ni”. .................................................................. 56

Figure 23. Time-dependent inhibition of TauD by Co” and Ni". ....................................... 58

Figure 24. Electronic spectra of Con-substituted, NOG-Fe“, and orKG-Fen forms of TauD
............................................................................................................................................. 60

Figure 25. Location of csiD and ygaF and regulation of the csiD-ygaF-gabDTP gene
cluster. ................................................................................................................................. 71

Figure 26. The metabolic pathways for catabolism of GABA, putrescine, agmatine,
arginine and omithine in E. coli, highlighting the interrnediacy of GABA. ....................... 71

Figure 27. A) Quaternary structure of CsiD. B) A subunit of CsiD. .................................. 73

Figure 28. SDS-PAGE analysis of the expression and purification of CsiD. and CsiDz. ..84

Figure 29. Determination of the native size of the CsiD; (red) and CsiD; (blue). ............. 85
Figure 30. Absorption spectra of CsiDz .............................................................................. 87
Figure 31. EPR spectra of CsiD./Fe(II)/NO (black) and CsiD./Fe(II)/oKG/NO (red) ....... 88
Figure 32. Oxygen consumption of the uncoupled reaction of CsiD .................................. 89

Figure 33. The aKG consumption (A) and succinate production (B) curves for the

uncoupled reaction of CsiD as followed by HPLC. ............................................................ 91
Figure 34. Proton NMR spectra of the uncoupled reaction and a blank experiment. ......... 92
Figure 35. Photos of Biolog plates. ..................................................................................... 96

Figure 36. Comparison of the amino acid analysis of the BW25113 (black) and csiD-KO
(red) strains. ........................................................................................................................ 98

Figure 37. Pull down assays with Hisb-tagged CsiDz mixed with cell extracts of non—

tagged YgaF. ..................................................................................................................... 100
Figure 38. SDS-PAGE analysis of the expression conditions tested for ygaF. ................ 124
Figure 39. . Purification of YgaF as monitored by SDS-PAGE analysis. ........................ 125
Figure 40. Determination of the native size of YgaF by sephacryl 300 chromatographzyg
........................................................................................................................................... 1

Figure 41. FMN is not covalently bound on YgaF. .......................................................... 128
Figure 42. Photoreduction of YgaF .................................................................................. 129

xi

Figure 43. Reductive titration of YgaF by dithionite. ....................................................... 130

Figure 44. A) Titration of YgaF with sulfite. B) Absorbance change at 450 nm versus the
concentration of free sulfite. ............................................................................................. 131

Figure 45. Analysis of the YgaF flavin reduction potential using MB as the redox dye.. 132

Figure 46. Analysis of the YgaF flavin reduction potential using PMS as the redox dye.
........................................................................................................................................... 133

Figure 47. Comparison of the growth of BW25113 (WT) and ygaF-KO strains with
different C- and N-sources. ............................................................................................... 135

Figure 48. Comparison of the amino acid analysis of the BW25113 (black) and ygaF-KO
(red) strains ....................................................................................................................... 137

Figure 49. Titration of anaerobic YgaF with L-2-hydroxyglutarate. ............................... 141

Figure 50. Determination of the Km and Vmax of the L-2-hydroxyglutarate oxidation from a
graph of the initial velocity (vi) versus its concentration. ................................................. 142

Figure 51. Identification of the product of the enzymatic reaction of L-2-hydroxyglutarate.

........................................................................................................................................... 144
Figure 52. Timecourse of the aKG production for the reaction of YgaF with the isomers of
2-hydroxyglutarate. ........................................................................................................... 146
Figure 53. Titration of anaerobic YgaF with D-3-phosphoglycerate ................................ 147

Images in this dissertation are presented in color.

xii

LIST OF SCHEMES

Scheme 1. Equilibrium of different reduction states of flavin. ............................... 19

Scheme 2. Mechanism of flavoprotein photoreduction with free flavin as catalyst.

................................................................................................................................. 24
Scheme 3. Oxidation of the two-electron reduced flavin by molecular oxygen. ....25
Scheme 4. Catalytic mechanism of TauD. .............................................................. 51
Scheme 5. Slow-binding inhibition kinetics. .......................................................... 58
Scheme 6. Binding mode of aKG and postulated modes for binding NOG to iron
metal ion .................................................................................................................. 62
Scheme 7. Synthesis of S-S-aminovalerate from L-omithine. .............................. 122

Scheme 8. Structural similarity of 2-hydroxyglutarate and 3-phosphoglycerate..147

Scheme 9. The catalytic YgaF reaction with L-2-hydroxyglutarate. .................... 151

xiii

aKG

aKG-Ds

CarC

CAS

DAOC S

DEA/NO

DTT

EDTA

EPR

EXAFS

FAD

FMN

HEPES

HPPD

KO

LMCT

MLCT

NADH

NADPH

NIR MCD

NMR

ABBREVIATIONS

a-Ketoglutarate
aKG-dependent dioxygenases

Carbapenem synthase
Clavaminate synthase
Deacetoxycephalosporin C synthase
Diethylammonium (Z)-1-(N,N-diethylamino)diazen-1-ium
-1,2-diolate

Dithiothreitol

Ethylenediamine tetraacetic acid

Electron paramagnetic resonance

Extended X-ray Absorption Fine Structure
Flavin adenine dinucleotide

F lavin mononucleotide
N-(2-Hydroxyethyl)piperazine-N'-(2-ethanesulfonic acid)
4-Hydroxyphenylpyruvate dioxygenase
Knock-out

Ligand-to-metal charge-transfer
Metal-to-ligand charge-transfer
Nicotinamide adenine dinucleotide
Nicotinamide adenine dinucleotide phosphate
Near-infrared magnetic circular dichroism

Nuclear magnetic resonance

xiv

NO

NOG

OPDA

PCD

PCA

P3H

P4H

RR

SDS-PAGE

TauD

deA

Nitric acid

N—oxalylglycine

o-Phenylenediamine

Protocatechuic 3,4-dioxygenase

Protocatechuic acid

Proline 3-hydroxylase

Prolyl 4-hydroxylases

Resonance Raman

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis
Taurine/aKG dioxygenase

2,4-D dioxygenase

XV

CHAPTER 1

INTRODUCTION

a-ketoglutarate dependent dioxygenases

Dioxygen (02) is one of the major air components (20.95%), and it is necessary for
aerobic respiration in animals. The reduction of oxygen is very strictly controlled by
nature. Despite the fact that the reactions of oxygen with organic molecules are
thermodynamically favored (exothermic), these reactions are relatively slow. The kinetic
barrier has its origin in the fact that oxygen’s ground state has a triplet spin, while organic
biological molecules have singlet ground states, and consequently interactions between
them are spin forbidden.I Nature overcomes this barrier by using molecules capable of
one-electron chemistry that can reduce dioxygen by exciting it to one of its singlet states.
Enzymes that use 02 to oxidize their substrates utilize specific cofactors that promote the
spin inversion. Such cofactors include organic molecules that can stabilize a radical (e.g.,
the isoalloxazine ring of a flavin), metal ions that are redox active (e.g. iron), and
combinations of the two (e. g., the porphyrin ring of a heme). Non-heme iron dioxygenases
including a-ketoglutarate (aKG)-dioxygenases are included in the hybrid cofactor class
because they employ a combination of a metal ion (Fe(II)) and an organic molecule (aKG)

to perform the 0; spin inversion.

Reactions of aKG-dependent dioxygenases and related enzymes

aKG-dependent dioxygenases (aKG-Ds) consist of an enzyme family covering a
wide range of reactions of biological significance. All orKG-Ds have an iron atom at their
active sites and usually they require three substrates: aKG, molecular oxygen and a
primary substrate, which varies depending on the role of the particular enzyme. The

general reaction catalyzed by these enzymes is the hydroxylation of their primary
2

substrate (shown in Figure l). Enzymes in this family are called dioxygenases because
both oxygen atoms of molecular dioxygen are inserted into the substrates; one into the
cosubtrate aKG leading to carbon dioxide (C02) and succinate, and the other into the
primary substrate.

orKG-Ds carry out a large number of chemical reactions including hydroxylation
(aliphatic and aromatic), hydroxylation-e1imination, carboxyl formation (from aldehyde or
ketone), epoxidation, desaturation, ring expansion/closure, epimerization, and
halogenation (that was discovered very recently).2‘3 The proteins of this family have
highly diverse biological roles too, including protein side-chain modifications, repair of
alkylated DNA/RN A, biosynthesis of antibiotics and plant products, lipid metabolism, and
biodegradation of a wide variety of small molecules (reviewed by Hausinger).3 A brief
introduction of the reactions catalyzed by several important representatives of the aKG-Ds

is given below, followed by a discussion of the general catalytic cycle.

0

aKG-D
RH * HOOCMCOOH + o

 

2 R(O)H + Hooc/\/COOH + 002

Figure 1. General reaction of the a-ketoglutarate-dependent dioxygenases (aKG-Ds).

Prolyl 4 -hydr0xylase

Prolyl 4-hydroxylase (P4H) is included in the side-chain modification subgroup,
and it is historically significant because it was the first identified aKG-D in 1966.4 The
reaction catalyzed by P4H is the hydroxylation of the proline residue, as shown in Figure

2. P4H activity has been found in mammals (where it plays a key step in collagen

biosynthesis and serves a critical role in hypoxic signalling) and in plants (where it

modifies proline-rich glycoproteins that are components of the cell wall.5’6

 

o\>_R 0%,R
02 (:02 OH

OIKG succinate

Figure 2. Prolyl 4-hydroxylase (P4H) reaction.

AlkB

Alkylating agents can produce lethal mutagenic lesions by methylating DNA.
AlkB is an Escherichia coli demethylase that repairs 1-methyladenine and 3-
methylcytosine lesions in DNA and RNA. Its activity remained unknown for more than 20
years from the time alkB was first identified as a gene involved in repair of alkylation
damage.7 Finally, in 2002 the enzyme was shown to hydroxylate the methyl group by
aKG—D chemistry, resulting in a spontaneous departure of formaldehyde and recovery of

adenine and cytosine, respectively (Figure 3). 89*”)

C lavaminate synthase

Clavaminate synthase (CAS) is involved in the biosynthesis of the antibiotic
clavulanic acid. CAS catalyzes three steps in the biosynthesis.ll The first step is the
hydroxylation of deoxyguanidinoproclavaminic acid (B-lactam) yielding

guanidinoproclavamic acid. The next step is a hydrolysis of the guanidine group catalyzed

by an amidinohydrolase. CAS then catalyzes two steps, cyclization of proclavaminic acid

and desaturation, producing finally clavulanic acid (Figure 4).

 

 

 

 

A)
NH2 NH2
N \Kl/ AlkB <: NADH N NH2
</ I / ; : IKN
,N N) A DNAN N//I ‘\ <’N lN/J
DNA 02 coz HCHO /
“KG succinate DNA
B)
NH2 NH2
\fi/ Ame fi/\OH ””2
'N’J m 7 N) N I U)"
l 02 CO ' N
2 HCHO |
DNA (IKG succinate DNA DNA

Figure 3. Reaction of AlkB having A) l-methylated adenine and B) 3-methylated cytosine
as a substrate.

0

O
N [If 0
COOH COOH Cf
£2 HO? . N
proclavamlnate COOH
CAS amidinohydrolase; HO?

NH
HN=<$ /\\b; HN=<
NH2 NH2

GKG succinate

 

 

0

Ci 0
N
HO COOH CAS ‘ 0% CA S Ofio
/r‘\b COOH ' COOH
C 2 02

NH2 OLKG succinate NH2 aKG succinate

 

 

Figure 4. CAS is involved in three out of four steps of clavulanic acid biosynthesis.

5

T aurine dioxygenase

aKG-Ds are also involved in biodegradation of the molecules supplying with
sulfur, phosphorous or carbon when the cell is undergoing starvation for the particular
element. In E. coli under sulfur starvation, the operon tauABCD is activated and provides
the organism with proteins that are involved in taurine (2-aminoethanesulfonate) uptake
(TauA, TauB, and TauC) and degradation (TauD).'2‘l3 TauD is an aKG-D that
hydroxylates the C-1 position of taurine, creating an unstable intermediate which
decomposes to aminoacetaldehyde and sulfite, as shown in Figure 5.14 TauD is one of the
very well studied enzymes of this superfarnily. Some experiments that I performed with

this particular enzyme are reported in Chapter 2.

 

 

- TauD _
N803 - SO - /O -
HZN A I HzN/Y 3 HzN/V * ”303

aKG succinate

Figure 5. Reaction of TauD.

4-hydr0xyphenylpyruvate dioxygenase

Several enzymes are included in the aKG-D superfamily because they exhibit
similarities in sequence or catalytic mechanism even though they may not require aKG. 4-
Hydroxyphenylpyruvate dioxygenase (HPPD) exhibits chemical parallels, but is not
related in sequence to aKG-Ds. It catalyzes the second step of the tyrosine catabolism
pathway.l5 HPPD doesn’t utilize aKG as a cosubstrate, and its only substrate is an a-keto

acid (4-hydroxyphenylpyruvate) which is cleaved (Figure 6).'6 Actually the mechanism

involves decarboxylation, acetyl group migration and oxygenation (with both oxygen

atoms getting incorporated into the aromatic product) steps.

 

HOOC
0 OH
HPPD COOH
OH 02 002 OH

Figure 6. Reaction of HPPD.

General characteristics and mechanistic studies of the aKG-dependent dioxygenases

3D structure

Crystallization of aKG-D proteins started only about 10 years ago, and proceeded
so fast that there are more than 50 entries of crystal structures in the Protein Data Bank.
These enzymes have been reported to exist in various oligomeric forms. For instance,
CAS is a monomer;‘7 whereas DAOCS (deacetoxycephalosporin C synthase), which is
involved in the cephalosporin synthesis from penicillin, has been found in both

8 TauD crystallized as a dimer;‘9 and AtsK

monomeric and trimeric states.l
(alkylsulfatase), a close relative of TauD that produces sulfate from alkylsulfates, is a
tetramer.20 The highest oligomeric form found in the aKG-D family is hexameric, reported
for CarC (carbapenem synthase), an enzyme involved in carbazole degradation.”

The monomers of all known crystal structures of this family contain the “jelly-roll”

motif. This fold consists of two four-stranded antiparallel B-sheets that are twisted

clockwise (Figure 7).22 The structure of the jelly roll motif is supported by a-helices that

are conserved and pack alongside the major B-sheet. HPPD, which is a relative of this

enzyme family, doesn’t exhibit this motif in its structure. '6

5 4
s 3
7 2
a 1
c N

 

Figure 7. Structure of the jelly-roll motif.23

General reaction mechanism

The catalytic cycle (Figure 8) of the aKG-Ds starts with the metal ion being
hexacoordinated with three amino acid residues (Asp/Glu and two His) and three water
molecules in an octahedral geometry. The usual order for addition of substrates starts with
binding of the cosubstrate uKG, resulting in bidentate chelation of the Fe(II), displacing
two water molecules (intermediate 8).”25 Subsequently, addition of the primary substrate
leads to the removal of the last water molecule from the coordination environment of the
metal ion (intermediate C), creating a place for oxygen binding. The primary substrate
coordinates in the active site, in close vicinity to F e(II), but it doesn’t interact directly with
it.l9 Oxygen binding to the Fe(II) site has been proposed to lead to Fe(III)-superoxo or
Fe(IV)-peroxo complexes (intermediates D),3 which finally collapse to a ferry] 0x0
complex (intermediate E). a species which is to date the only identified and characterized

iron—oxygen intermediate.”27 Finally, the products, oxygenated primary substrate,

8

succinate, and carbon dioxide, are released and three water molecules bind to the Fe(II),
returning the enzyme to its starting state, ready for a new cycle. Each step of the catalytic

cycle is analyzed below.

 

 

B C
H20 R-H
Glu/Asngu /O\ /R R‘-H 2 H O GIu/Asp,’ (.0 R
2 H20 Fe” 2 n"! \
/ 2 1 Fe.‘
His i ‘0 His/ i \‘O
aK HIS - 0 His - 0
H2
Glu/Aspo,‘ : H20
A "Fe", 02
His/,1 \HQO _ —1
HIS R'-H
._ D R'-H
O 2
GIU/Aspz’a'. ."10 R GIu/Asz Q 0 R
Fen: —— aw
co His .1 ‘o - 1 ~.
2 HIS _ 0 Hrs ,
succinate HIS — O
R'(O)H
__ __l
3H20 I
R" R.“
- R 0’2 R
Gnu/Asp... TH .0 G'UIAsp’a 55 x9
F ”Fem” .. 0 ¢——— Flelil O E
' His “ : :
HIS His! O=C=O His 0 C 0

Figure 8. General mechanism of aKG-dependent dioxygenases. Description of the steps is
reported in the text.

Resting state
The metal ion is involved only in the mechanism of the enzyme, and has nothing to
do with the structure of the proteins, since apoenzymes of several aKG-Ds have been

crystallized and showed that they adopt similar overall conformations with the

holoenzymes.2 The Fe(II) in the holoenzyme in the absence of any substrate (resting state)
appears in an octahedral geometry with three water molecules and three amino acid
residues as ligands. The three amino acid residues that bind Fe(II) in the active site
constitute “the facial triad”, which includes two histidines and one carboxylic acid (Asp or
Glu) found in a HXD/EXnH motif.28 Examples of amino acids constituting the facial triad
in different enzymes are given in Table l. HPPD shows deviations from this behavior,
having two histidines (Hisl6l and HisZ40) and a glutamic acid (Glu322) binding the
Fe(II) in a HX~65HX~50E motif that rather resembles an extradiol dioxygenase.29

The crystal structures of the resting state of P3H (proline 3-hydroxylase) and
DAOCS have been solved.30’3 1 The near-infrared magnetic circular dichroism (N IR MCD)
spectrum of CAS in the resting state exhibited two d-d transitions at 917 and 1085 nm

associated with to the distorted octahedral center.32

 

 

 

 

 

 

 

Protein Organism Metal coordinating residues
TauD Escherichia coli His99, AsplOl , His255
CAS Streptomyces clavuligerus Hisl44, Glul46, His279
DAOCS Streptomyces clavuligerus Hisl83, Aspl 85, Hi3243
CarC Erwinia carotovora Hilel, Asp103, His251
AtsK Pseudomonas putida HileS, Asp110, Hi5264

 

 

 

 

Table 1. List of the facial triad of several chG-Ds.'9"7*3"2“20

aKG binding to the active site
Displacement of two water molecules permits binding of aKG in a bidentate

fashion to the ferrous active-site ion via the Cl carboxylate and the C2 keto group as

10

shown in Figure 8. This complex structure was first proposed based on spectroscopic
experiments on CAS,32 and later proven from various aKG-Ds crystal structures (e.g. CAS
and AtsK).”‘20 All proteins of this family in the presence of aKG and in the absence of
oxygen, exhibit a characteristic chromophore at around 500 nm with slight variations of
the maxima and intensities (e.g., 530 nm for TauD, 500 nm for CAS).24‘33 This absorption
was assigned by Solomon and coworkers to metal-to-ligand charge-transfer (MLCT)
transitions due to excitation of electrons from the d (yz, xz-yz, 22) orbitals of the iron atom
to the 7t* orbitals of (JtKG.33 Resonance Raman (RR) spectroscopy of TauD yielded
vibrations at 460 cm'I and 1700 cm", which were assigned to metal-to-ligand stretching
and C = O vibrations respectively.34 The aKG cosubstrate can bind to the Fe(II) in two
different ways, “in line” and “off line”, depending on the coordination position of the C1
carboxylate group of dKG. When the Cl carboxylate group of aKG is trans to the
proximal His, this is called “in-line” coordination and oxygen can bind and directly attack
the substrate. Examples of enzymes that belong in this category are CAS and TauD.'7"9
On the other hand, in the “off-line” coordination, the C1 carboxylate group of aKG is
trans to the distal histidine residue and the active site must undergo reorganization for
bound 0; to react with the substrate. Representatives of this category are the enzymes

CarC and DAOCS.2"35

Binding of the primary substrate
The primary substrate binds in close proximity to, but not in direct contact with,
the metal ion, forcing the last water molecule ligand to leave. The departure of the solvent

results in a change from six-coordinate octahedral geometry to five-coordinate square

11

pyramidal geometry of the iron complex. Several X-ray structures of enzymes with both
substrates bound (under anaerobic conditions) have been published.3 The change in the
iron coordination can be monitored spectroscopically. For example, the addition of taurine
to TauD alters the characteristic MLCT chromophore so that it is blue shifted from 530
nm to 520 nm.24 RR spectra of TauD also demonstrate a shift, from features at 460 and
1700 cm'1 to features at 470 and 1688 cm], upon taurine binding to the active site with a

consequent change in the iron complex geometry.34

Oxidation

Dioxygen binding to the iron complex is a critical step in the oxidation of the
cosubstrate and primary substrate of all the aKG-Ds. Delocalization of electron density
from the cosubstrate to the iron ion facilitates the single-electron reduction of the
molecular oxygen that is necessary for its spin inversion. 02 takes the place of a water
molecule in the primary coordination sphere of the ferrous ion. Observation of the
accumulation of three spectroscopically distinct oxygen intermediates in HPPD proves the
existence of multiple oxygen complexes during a single turnover.36 Nevertheless, the only
proven iron-oxygen intermediate is the ferryl-oxo (Fe(IV)=O'2) complex (intermediate E
in Figure 8), which has been observed in TauD and P4H catalytic cycle.26’37’38 The
particular intermediate of TauD was observed by stopped-flow UV-visible
spectrophotometry to absorb at 318 nm.37 Rapid freeze-quench techniques combined with
Mossbauer spectroscopy (on a TauD sample) were used to trap and characterize the TauD
intermediate as being paramagnetic with an integer-spin ground-state S22. Subsequent

cryoreduction of the complex formed a high-spin Fe(III) species, confirming the Fe(IV)

12

oxidation state of the reported intermediate.37 The existence of the ferryl-oxo species was
also confirmed by Proshlykov et a1. using RR and cryogenic continuous flow methods.26
The Fe(IV)=O'2 was directly identified by the isotope-sensitive vibration vFe-so in the 800
cm'1 region, which shifted from 821 cm'1 for '60 to 787 cm'1 for 180. Moreover, EXAF S
spectroscopy was used to determine that the distance between the iron and the oxygen of
the TauD oxo-intermediate was 1.62 A, similar to the corresponding distance of model
complexes.39 This intermediate abstracts a hydrogen atom from substrate to form a carbon
centered radical and Fe(III)-OH, followed by hydroxyl radical rebound to generate the
product.

HPPD reacts with molecular oxygen 3,600 times faster in the presence of its
primary substrate and such reactivities with 02 are similarly enhanced in other (zKG-Ds;40
nevertheless, it has been shown in vitro that the absence of primary substrate leads to
uncoupled reactions and results in enzyme inactivation and self-modification.4"42’43 In the
general uncoupled reaction one atom of dioxygen incorporates into the cosubstrate as the
metal site oxidizes and the inactive enzyme is reactivated by ascorbate. An alternative type
of uncoupled reaction is found in TauD which generates a 550 nm chromophore; in this
case an aKG-dependent process leads to hydroxylation of a side chain, which by RR
spectroscopy was identified as an Fe(III)-catecholate species. The self-hydroxylated Tyr73
residue, or dihydroxyphenylalanine, forms a complex with the oxidized, active-site iron
ion to generate the distinctive ligand-to-metal charge-transfer feature.41 A similar example
was reported for the dichlorophenoxyacetate/a-ketoglutarate dioxygenase (deA) which
formed a hydroxylated Trp in the absence of its primary substrate.42 The blue ferric

complex that is formed under these conditions is catalytically inactive. Reductants like

13

ascorbate can return the iron to its ferrous state, but it cannot reverse the side chain
modifications; in some cases this treatment restores a portion of the catalytic activity of

the dioxygenase.44

Products formation and release

Formation of the oxygenated primary substrate (R(O)H) accompanies the oxidative
decarboxylation of aKG, and release of the products is one of the last steps of the catalytic
cycle. An interesting experiment was performed by Yu et al. in order to observe the
generation of the products of AlkB.45 Anaerobic crystals of holo-AlkB complexed with
aKG and l-methyl adenine nucleotide (dT-( 1-me-dA)-dT) were exposed to molecular
oxygen. As a result, aKG converted to succinate and demethylation of the nucleotide was
observed. Concerning product release, there is debate whether CO2 and succinate
dissociate during or after the turnover is complete.2 In any case, C02 is the first product to

dissociate, followed by succinate.

l4

Flavoenzymes

Structures of flavins

More than a century ago, Blyth reported the existence of a yellow pigment in the
content of whey, which he named lactochrome.46 Over the years many other yellow
pigments were observed in nature, including hepatoflavin and verdoflavin that were
named on the basis of their biological sources or their physical properties. Subsequently,
these compounds were all shown to be riboflavins. Richard Kuhn and Paul Karren
independently determined the structure of flavin and proved it by chemical synthesis, as
reviewed'by Vincent Massey.58 The structure of riboflavin and its natural variants are
shown in Figure 9. Riboflavin got its name from its structural components 7,8-
dimethylisoalloxazine (which was named flavin from the latin “flavous” that means
yellow) and the reduced form of D-ribose. Lumichrome and lumiflavin are riboflavin
derivatives after photolysis at pH < 7 and pH > 9 respectively. The riboflavin moiety in
some cases is attached to adenosine diphosphate forming FAD (flavin adenine
dinucleotide), and in other circumstances is bound to phosphate forming FMN (flavin

mononucleotide).

Diversity of flavoenzymes

The biologically active forms of riboflavin, FMN and FAD, have been found to be
cofactors for many enzymes. It has been estimated that genes encoding flavin—containing
proteins comprise of 1-3 % of the prokaryotic and eukaryotic DNA.47 The high abundance
of flavoproteins in cells, together with the importance of their biological roles, has led to

extensive studies of these enzymes. They are involved in an amazingly wide range of

15

reactionsfmfl50 It has been shown that flavoenzymes are involved in typical
dehydrogenation reactionsSI (using a large variety of metabolites as substrates), electron
transfer from and to redox centers,50 light emission,52 activation of dioxygen,
photochemistry53 signal transduction related to programmed cell death54 and DNA
damage55 and repair.53 Their substrates can be alcohols, aldehydes, ketones, acids, amino

acids, amines, dithiols, hydroxy acids, and other compounds.56

 

 

 

 

 

 

 

Lumiflavin
\ OH OH
O O O
Lumichrome H2C/‘z _ \p/ \p\/
E / //
g 0/ \OH 0 OH
OH
1’
H C N N O
3 9\ 10 /1
I8 2
7 4a 3
6 5/ NH
H3C / N
0
L Riboflavin
I FMN
FAD

Figure 9. The structures of flavins.

l6

Flavin redox chemistry

Redox states of flavin.

Flavoenzymes are involved in a plethora of reactions because of the chemical
versality of the alloxazine ring. The isoalloxazine moiety is the redox active part of the
flavin molecule, whereas the side-chain, including the adenine moiety, functions mainly to
anchor the coenzyme to the active site. Flavins can exist in the oxidized, one-electron-
reduced (semiquinone) and two-electron (fully) reduced state, as shown in Figure 10.
Every state can occur in cataionic, neutral, or anionic forms.S7 Based on the pKa, six out of

the nine forms are observed under physiological conditions.

 

 

CATIONIC NEUTRAL ANIONIC
R H R """"""""""""""""""""""""""" R ---------- 1
HC N N o =iHc N N o pKa=1o HC N N 05
OXIDIZED 2 1):“ Y ”‘3 0; 2 DE ’ ‘F 2 I: ’ Y :
H30 N’ NH H+ EH30 N’ N” H“ H3C N’ '9 :
o 5 o o

/

 

 

 

 

 

 

E R E
R H i F3 - :
HZCJCKN I NYC pKa=2.3E HZCUN /N\'¢O pKa=3.3 HZCUNfiYOE
SEMI UlNONE . 5 . \ NH :
Q H3C +NIn/NH H+ : H3C uLrNH H+ H3C u 0
H o 5 o - 5
R H
(FULLY) H2O N Nq¢0 pKa<o=
READUCED m I NH .
H3C *N H+ i
H2 0 :

 

PHYSIOLOGICALLY POSSIBLE

Figure 10. Redox states of flavin.

An advantage of studying flavoenzymes is that the investigator can monitor the
redox-dependent reactions they catalyze using a spectrophotometer. An example is shown
in Figure 11.58 The oxidized form of the flavin has yellow color and a distinct spectrum
with dual maxima at ~380 nm and ~450 nm (the wavelength varies slightly in each
protein). For an anionic semiquinone the intensity of the 380 nm feature increases, while
the 450 nm peak decreases. The neutral anionic form of the cofactor has an absorbance

transition around 600 nm. The fully reduced form of flavin is colorless.

 

 

    
   

 

A Anionic
E ,3 Semiquinone
f 16-
': -_ . .
E '4 /’\. Oxrdized
v
E 12“ _ \
m /
'5 10* ‘
t i _. ""5
8 B ‘ -'; .
2 5i 5 :1 1, Neutral Semiquinone
:3 a... f :1:. ‘t‘ 6.55"," N
O 4' ,’ ‘ :5"
.E 2 ’ t
12 " Reduced \~-...
LLI .

aoo

Wavelenght (nm)

Figure 11. Absorption spectra of the redox states of glucose oxidase.58

Many flavoenzymes catalyze reactions through a single two-electron flavin-

reduction step and do not form any semiquinone intermediate. On the other hand, other

18

flavin-containing enzymes catalyze mechanisms involving two, one-electron steps for
flavin reduction with the intermediacy of a semiquinone. The reduced proteins are
reoxidized by reactions that again utilize either one, two-electron step, or two, one-

electron steps (only the overall processes are depicted in Figure 12.

Oxidized
RH2 F lavin XHz
Reduced
R F lavrn X

(Flavin-H2)
Figure 12. Reductive and oxidative half reactions of flavoproteins.

In solution, a mixture of oxidized and reduced free flavin is observed in
equilibrium, where an amount of flavin semiquinone is formed, as shown in Scheme 1. At
physiological pH (pH 7) the radical species accounts for only 5% of the total, in a mixture
containing equal amounts of oxidized and reduced states.58 When flavin is bound in the

active site of a protein, this equilibrium can be altered dramatically resulting in a broad

range of stabilization of the semiquinone species.

Flox + Flred = 2 PET

Scheme 1. Equilibrium of different reduction states of flavin.

19

Redox potential of flavoproteins.

Free flavin has a redox potential of —210 mV at pH 7;48 however, when interacting
with residues at the active center of flavoenzymes, this value can change dramatically. The
lowest flavoprotein redox potential ever reported is -495 mV for the semiquinone/reduced
couple of the flavodoxin from Azotobacter vinelandii,” and the highest potential recorded
is +80 mV for the semiquinone/reduced couple of thiamin dehydrogenase.60 The wide
range of potentials reflects the strong modulation exerted by the protein environment on
the cofactor redox properties. Strongly positive charges in the proximity of the flavin
cause an increase in the observed redox potential,49 as confirmed by crystallographic
studies.61 By contrast, negative charges near the coenzyme lead to decreases of the redox
potential.

Hydrogen bonds between the flavin and protein residues can tune the redox
potential as well. The N1-C2=0 portion of the isoalloxazine (shown in Figure 9) has been
shown to be in contact with positive charges associated with Lys or Arg side chains, as
well as with the a helix N-terminus.49 This interaction is responsible for the stabilization
of the reduced form of the flavin; therefore it increases the redox potential. Another
important locus for hydrogen bonds is the N5 of the isoalloxazine. The N5 position
becomes protonated upon oxidation of a substrate; therefore the particular hydrogen bond
disfavors the reaction from occurring. Higher redox potentials have been found for
enzymes which lack such a hydrogen bond with the N5 of the flavin (e.g. Flex/FIN: +55
mV for vanilly-alcohol oxidase, Flox/Flscmiq: -33 mV and Flsemiq/Flred: -17 mV for glycolate

oxidase)” 62

2O

The flavin molecule typically is strongly, but non-covalently, bound to
flavoenzymes. In addition, some enzymes bind the coenzyme using a covalent bond, as
shown in Figure 13.63 The covalent bond between the enzyme and the cofactor reduces the
oxidative power of the flavin, as proven by mutagenesis experiments where the covalent

bond was elimimated.64

21

R
/ I
His(Nl)-8a-f|avin N1—H2C N
N3\ |8 9\ 10
7
6 5/
H3C / N
R
His(N3)-8a-flavin / |
Ng—H2C N
N1§/ l8 9\ 10
7
6 5/
H3C / N
R
Cys(S)-8a-flavin I
”MS—H c N
2 8 9\ 10
l?
6 5/
H3C / N
i
Tyr(0)-8a-flavin O‘—H2C N
l8 9\ 10
7
6 5/
H3C / N
T
H3C N N o
CYS(S)-6-flavin '8 9\ 10 / 1%
7 3
H3C / N
S
M

Figure 13. Covalent protein-flavin bonds observed in flavoenzymes.

22

4
O

Photoreduction.

Photoreduction is another common property of flavins. The flavin takes two
electrons from a variety of electron donors to form its fully reduced state.46 As for
chemical reductants, reduction can occur either in a single, two-electron step or with the
intermediacy of a partially reduced species (semiquinone). Photoreduction of
flavoenzymes often is more difficult than for the free coenzyme; however, Vincent
Massey and co-workers demonstrated that photoreduction of several flavoenzymes is
facilitated by having free flavin as a catalyst in the reaction solution.65 It was envisioned
that catalytic amounts of free flavin can trigger the light-mediated reduction of
flavoenzymes, as is shown in Scheme 2. EDTA (ethylenediamine tetraacetic acid) is the
most common photoreductant in such studies, and is added to the mixture of flavoenzyme
and free flavin. Free flavin (F l) is photoreduced by light irradiation (reaction 1), then
reacts with the flavoenzyme (E-Fl). This reduction can go through one two-electron step
(no intermediates generated) or via two one-electron reactions (semiquinone intermediate
is formed; reactions 3a and 3b). The most effective, and therefore the most commonly
utilized, free flavin used in this experiment is 5-deazaflavin, which possesses a nitrogen
atom at the position typically occupied by C5.66

As expected, the photoreduction of a flavoenzyme can be monitored
spectroscopically. The absorption at ~450 nm of the oxidized protein decreases as the
reduction proceeds. There are many examples where semiquinones (indicated by the
increase in absorbance at ~370 nm for the anionic species) can be observed, such as the
one shown in Figure 14.67 In other cases, neutral semiquinones are formed or direct

reduction to the fully reduced flavin is observed.

23

EDTA

Fl ——>
OX hV

FlredHZ

a) no intermediate FI,,.,dH2 + E-Fl0X —» Flox + E'F'rede
formed

b) intermediate FlredH2+ E-Flox —> “H + E‘FlredH
formed

FlredHZ + E'FlredH _’ FIH. I E'FlredHZ

2FIH’ —— FIox + F'redHZ

reaction 1

reaction 2

reaction 3a

reactioan

reaction 3c

Scheme 2. Mechanism of flavoprotein photoreduction with free flavin as catalyst.

0.14
0.12
0.10
0.08
0.06 1

 

 

Absorbance

0.02 i

A

 

0.00 - r - . .
300 350 400
Wavelength (nm)

 

'7

450

Figure 14. Photoreduction of AidB.67

24

560

Classification of flavoenzymes

The abundance of flavoenzymes in nature and their participation in a wide variety
of roles is reflected in their classification, which is based on the substrates recognized and
the reactions catalyzed. The classic categorization has separated these enzymes into three
groups: oxidases, monooxygenases, and dehydrogenases. The large number of
dehydrogenases identified with diverse properties has forced the scientific society to split
that category into subclasses. Therefore, by taking into account the type of reactions
catalyzed, the reactivity with oxygen and the nature of additional redox centers in the
protein, flavoenzymes are now constituted into five groups oxidases, monooxygenases,

electron transferases, flavoproteins with auxiliary redox centers, and enzymes with

. 5
unknown function:0

Oxidases.
Oxidases are flavoproteins which, when in their fully reduced form, react very
rapidly with molecular oxygen to yield the oxidized form of the enzyme and hydrogen

peroxide (Scheme 3).

F'redHZ + 02 __T FIOX + H202

Scheme 3. Oxidation of the two-electron reduced flavin by molecular oxygen.

The high redox potential of the O2/H202 couple (+270 mV at pH 7) drives the
reaction of Scheme 3 to the products, but this is a hindered process because it is spin

forbidden.68 Oxygen is in the triplet state and the reduced-form flavin is in the singlet

25

state. In order to proceed, an electron transfer from the flavin to the oxygen takes place,
yielding a caged radical pair.69 The rate of the reaction of molecular oxygen with
flavoenzymes varies widely, depending on the accessibility and environment of the flavin.
In particular, some oxidases react rapidly with oxygen (e.g., 1.5 x 106 M'l s'1 and 8.5 x 104
are the rate constant of glucose oxidase and glycolate oxidase respectively).7°’7' For free
flavins the step after the formation of the caged radical pair is the appearance of a flavin
hydroperoxide intermediate, which is heterolytically cleaved to H202 and oxidized flavin
(Flavin 7). For the oxidases, there is no detectable flavin hydroperoxide intermediate. This
can mean: 1) the intermediate is formed and decays really fast, or 2) instead of the
generation of that intermediate, a second electron is transferred from the flavin radical to

the caged superoxide to directly form H202 (Figure 15).

 

. H
H30 N NYC
caged radical pair H3C H
0.8-
H30:[£:[:|O NF: 02 “BC :N N NE:
H3O H3C H02- kHOz
02-0 R
H3Cji>N [N 0
H30 NIf...
0

Figure 15. Possible pathways of reaction of oxygen with reduced flavin.

Another reaction characteristic of this group of enzymes is the formation of a
stable flavin N(5) adduct with sulfite (Figure 16).72 This complex doesn’t absorb in the
visible range, therefore titration of sulfite in an oxidase solution exhibits a gradual

26

decrease of the flavin absorbance. An example is shown in Figure 17. The dissociation
constant Kd of the flavin-sulfite complex for oxidases is in the micromolar range. It has

been found that the Kd of sulfite can be correlated with the oxidation reduction potential of

the flavoprotein.73

|

Figure 16. The flavin-sulfite complex is a covalent N(5) adduct.

 

      
 
 

 

‘05 A 1 L 1.5 A z
ISuIl'ite](mM)

Absorbance
P
e
UI

  

 

 

 

A
0
300 400 500 600 700

Figure 17. Binding of sulfite to glycine oxidase from Bacillus subtilis.74

27

A question that still needs to be answered is whether there are any structural rules
that can give hints about the reactivity of flavoenzymes with oxygen. In other words: can
we predict if a flavoprotein is an oxidase by looking its 3D structure? According to
Mattevi, the answer to this question is no.75 Two proteins that are very similar structurally
are glycolate oxidase and flavocytochrome b2 (Figure 18). Despite their highly conserved
active centers, they exhibit very different oxygen reactivities. Flavocytochrome b2 has a
rate constant for reactions with oxygen of close to zero, whereas glycolate oxidase reacts

' s". The only differences are a

with a second-order rate constant of 8.5 x 104 M'
replacement of a Trp to Leu and an orientation change of a peptide (Alal98) in the
flavocytochrome b2 that are in contact with the N4-C4a locus of the flavin. From the

above observation it can be concluded that changes in the steric constraints and polarity of

the area around N4-C4a can greatly modulate oxygen activity.

28

 

Figure 18. Structure of the active site of several flavoproteins. A) glycolate oxidase, B)
flavocytochrome b2, C) D-amino acid oxidase, and D) glucose oxidase.75

Another factor that has been examined is the ability of oxygen to reach the flavin

in the active site of a protein. Although it had been thought that the flavins of the oxidases

29

might be more accessible than those in other flavoenzymes, flavocytochrome b2 and
selected other proteins have accessible flavin which doesn’t react with oxygen, and D-
amino acid oxidase and other representatives react with oxygen even though they have a
shielded coenzyme (Figure 18). Thus, flavin accessibility is not an obvious determining
factor for defining the protein reactivity with oxygen. Similarly, whether the flavin is
planar or distorted does not determine the reactivity of oxidases. For example, the flavin
of D-amino acid oxidase has a planar conformation, whereas that of glucose oxidase is
distorted (Figure 18). Therefore, there is no clear connection between the structure of a

flavoprotein and its reactivity with oxygen.

F lavoprotein monooxygenases.

As for the case in oxidases, the reduced state of monooxygenases reacts with
oxygen to form a caged radical pair (Figure 19). The difference between an oxidase and a
monooxygenase is that the first uses molecular oxygen as an electron acceptor, whereas
the second inserts an oxygen atom into the substrate.50 Such enzymes that use NADH or
NADPH for their reductive step are called external flavoprotein monooxygenases.76
Alternatively, the internal flavoprotein monooxygenases utilize a non-nicotinamide
primary substrate to reduce the flavin. An example of this group is the lactate
monooxygenases, where lactate is oxidized to pyruvate by reducing the flavin of the
enzyme. The reduced enzyme then reacts with oxygen, and the C4a—hydroperoxy
intermediate reacts with pyruvate to form acetate and carbon dioxide. The resulting C4a-

hydroxyflavin completes a catalytic cycle by retuming to its oxidized form by

dehydration. In the absence of substrate the hydroperoxy intermediate decays to the

30

oxidized form of protein and hydrogen peroxide, giving the same product as if it were an
oxidase. The reactions that are catalyzed by flavin monooxygenases are very diverse, as
shown in Figure 20.76

These enzymes stabilize neither the cationic nor the neutral semiquinone
intermediate. Also, in contrast to oxidases, they don’t react with sulfite to form the flavin
N(5)-sulfite adduct.68 Depending on the type of oxygenation, the intermediate C(4a)-
hydroperoxyflavin acts as nucleophile (e.g., in Baeyer-Villiger oxidation) or electrophile

(e.g., hydroxylation).68

31

No
If

NH

 

IZ Z-IJ

MCKIN NYO NADH/NADPH H3011
H3C H3C

oxidized flavin fully reduced flavin

O

caged radical pair

J

R
H3Cj::[
l 1::[NLT/NH OC(4a)-—hydroxyflavin
H C N
3 H (I) O
HO

RH
R(O)HQ/ \/ H202
R (absence of substrate)
H c N N 0
H30 N 3
H o
HO

oxidized flavin

 

1!

H20

Figure 19. Mechanism of external flavoprotein monooxygenases.

32

 

 

 

. . I
hydroxylation ___. > amine \ _ _ +_ -
> OH oxidation /N R3 R2 [ii 0
R2 R2 R2 R3
BaeyerVilliger R1 R‘ R‘ R1
oxidation F O ——> O): o epoxrdation \ ___§ :1: 0
R2 2 2
R
R1 1 . R R
. l selemde 1 1
phosphite ester \ _ _ +_ - . . \ \ __
oxidation )3 R3 R2 ll” 0 oxrdation /Se ———> §e—O
R2 R3 R2 R2
/OH OH R1 R1
organoboron R1“B R1—O-B/ sulfoxidation \s ——> \s=o
oxrdatlon \ \ / /
OH OH R2 R2

Figure 20. Reactions catalyzed by flavin monooxygenases.

Electron transferases.

Flavoproteins that react with one-electron acceptors or donors and produce 02' on
reaction with molecular oxygen are called electron transferases. In contrast to oxidases
and monooxygenases, the transfereses react sluggishly with oxygen. Also, they don’t form
the flavin N(5)-sulfite adduct that is characteristic of the oxidase reaction. These proteins
stabilize the blue neutral semiquinone intermediate. Another characteristic of these
proteins is that most of the flavin molecule is buried in the active site, and only the
dimethyl benzene ring is freely accessible to solvent.

There are two groups of electron transferases.50 The first includes proteins such as

flavodoxin and ferredoxin-NADP reductase. This group catalyses electron transfer

33

between two redox proteins as part of photosynthetic, nitrogen- or sulfate-reducing,
hydrogen-evolving, or other electron systems.

Acyl-coenzyme A (CoA) dehydrogenases, which constitute the second group of
transferases, catalyze electron transfer reactions that feed electrons into the respiratory
chain. Acyl-CoA dehydrogenases are reduced by two electrons as they convert substrate to
trans-enoyl-COA, containing a carbon-carbon double bond.77 The reduced acyl-CoA
dehydrogenase transfers the 2 6', one by one, to another flavoprotein, the electron-
transfem'ng flavoprotein.78 An unusual characteristic of the enzyme is that in the absence
of the enoyl-CoA product, the blue neutral semiquinone is stabilized, whereas the red

cationic semiquinone is stabilized in the presence of the product.50

F lavoproteins with auxiliary redox centers.

The flavoprotein disulfide oxidoreductase has a disulfide as the additional redox
center in close juxtaposition to the FAD.50 NADH or NADPH interacts with the flavin and
transfers two electrons that reduce a redox active disulfide. The reduced cysteines then
interact with the second substrate which is usually a dithiol, such as glutathione in
glutathione reductase, thioredoxin in thioredoxin reductase, or trypanothione in
trypanothione reductase. In some cases, the second substrate is not a disulfide; e.g.,
mercuric ion reductase, NADH peroxidase, NADH oxidase, and 2-ketopropyl-coenzyme
M carboxylase/oxidoreductase catalyze reduction of Hg”, hydrogen peroxide, and
molecular oxygen, and the reductive carboxylation of 2-ketopropyl-coenzyme M,

respectively.

34

A second subcategory of flavoenzymes with auxiliary redox centers includes the

flavocytochromes, which are heme-containing flavoproteins.50

A very well known
example of this class is yeast lactate dehydrogenase, which oxidizes lactate to pyruvate
while the flavin (FMN) is reduced by two electrons. The FMNHZ transfers the reducing
equivalents one by one to the cytochrome b, which is reoxidized by its external electron
acceptor.79

The third subset is the non-cytochrome metal-containing group. A representative
example is xanthine oxidase, which contains a molybdopterin cofactor and an iron-sulfiir
cluster ([FezszD in addition to the FAD coenzyme. As M0 is reduced (Mo‘i6 goes to
M0”), the substrate xanthine is oxidized to urate.80 The molybdenum cofactor is then re-

oxidized by sequential one-electron transfer steps to the FAD through the iron-sulfur

clusters. Finally, the FADH; reduces NAD+ in order to complete the catalytic cycle.

F lavoenzymes of unknown function.

A very interesting case of a flavoenzyme with unknown firnction is the old yellow
enzyme (OYE), which surprisingly was the first flavoprotein discovered (1932).81 It was
characterized “old” because in 1938 a second “new” yellow protein was isolated.82 This
enzyme together with many other uncharacterized enzymes that have sequence similarity
with it comprise a separate family because of their distinct physicochemical characteristics
together with uncertainty about their physiological role.

83 It is

OYE is rapidly reduced by NADPH and can be reoxidized by oxygen.
believed that NADPH is the physiological reductant; however, the reoxidation with

molecular oxygen is slow, and is lower than that obtained with substrates such as

35

quinones.82 This enzyme has a variety of substrates, all of them phenolic compounds.
Additionally, a very unique behavior of OYE is the development of an intense long

absorption spectrum upon binding of the phenolic substrate (Figure 21).83

 

Absorbance

 

 

 

 

 

Wavelength (nm)

Figure 21. Titration of OYE with p-chlorophenol.83

Thesis outline

The following chapters describe three projects that I carried out in the Hausinger
laboratory, and do not include additional studies performed while in the laboratory of Joan
Broderick.

Chapter 2 details my kinetic and spectroscopic investigation of TauD interactions

with Co(II), Ni(ll) and N-oxalylglycine (NOG) inhibitors. These metal ions were shown to

36

cause slow-binding, competitive inhibition of the enzyme. When Fe(II) was replaced by
Ni(II), no observable chromophore was produced. On the other hand, when Co(II) was
used, a chromophore was obtained, which could be used as a diagnostic marker for aKG-
dependent dioxygenases. Kinetic studies on NOG revealed that, despite its close structural
similarity to aKG, it is a weak competitive inhibitor of TauD. The NOG-bound state of the
enzyme was shown to be incapable of oxygen catalyzed oxidative decarboxylation.

Chapter 3 describes my efforts toward purification and characterization of an
Escherichia coli aKG-dependent dioxygenase, CsiD, of unknown function. This protein,
encoded by the csiD gene and activated by carbon starvation, has been crystallized;
however, its ability to bind and utilize aKG had not been shown. Anaerobic absorption
and EPR spectra of the holoenzyme in the presence of the cosubstrate proved that aKG
binds in a bidentate manner to Fe(II) at the active site. Furthermore, the uncoupled
reaction of CsiD was probed, using various methods. Finally, the function of the
dioxygenase was investigated with a direct method (oxygen consumtion), together with
metabolic studies that compared the Escherichia coli wild type with the CsiD knock out
strains utilizing several probes.

The final chapter describes my examination of YgaF, which is encoded by a gene
that is adjacent to csiD. YgaF had been proposed to be an FAD-oxidoreductase, but its
role was unknown. Here, the first cloning and overexpression of the enzyme is reported.
General characterization of YgaF includes photoreduction, titration with sulfite, gradual
reduction with dithionite, determination of its redox potential and demonstration that its
flavin is non-covalently bound. Oxygen was shown to reoxidize the flavin rapidly,

demonstrating that the protein is an oxidase. Various compounds were tested as substrates

37

of YgaF and studies that compared the metabolism of E. coli wild type and YgaF knock
out cells were undertaken as well. The only significant activity of YgaF was observed with

L-2-hydroxyglutarate.

38

REFERENCES

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

2 Purpero, V., and Moran, G. (2007) The diverse and pervasive chemistries of the u—keto
acid dependent enzymes. J. Biol. Inorg. Chem. 12, 587-601.

3 Hausinger, R. P. (2004) Fe(II)-a-ketoglutarate-dependent hydroxylases and related
enzymes. Crit. Rev. Biol. Mol. Biol. 39, 21-68.

4 Hutton, J. J ., Jr., Trappel, A. L., and Udenfriend, S. (1966) Requirement for a-
ketoglutarate, ferrous ion and ascorbate for collagen proline hydroxylase. Biochem.
Biophys. Res. Commun. 24, 179—184.

5 Kivirikko, KL, and Pihlajaniemi, T. (1998) Collagen hydroxylases and the protein
disulfide isomerase subunit of prolyl 4-hydroxylase. Adv. Enzymol. Rel. Areas Mol. Biol.
72, 325—398.

6 Hieta, R., and Myllyhatju, J. (2002) Cloning and characterization of a low molecular
weight prolyl 4-hydroxylase from Arabidopsis thaliana. Effective hydroxylation of
proline-rich, collagen-like, and hypoxia-inducible transcription factor a-like peptides. J.
Biol. Chem. 277, 23965—23971.

7 Kataoka, H., Yamamoto,Y., and Sekiguchi, M. (1983) A new gene (alkB) of Escherichia
coli that controls sensitivity to methyl methane sulfonate. J. Bacteriol. 153, 1301—1307.

8 Chen, B. J ., Carroll, P., and Samson, L. (1994) The Escherichia coli AlkB protein
protects human cells against alkylation-induced toxicity. J. Bacteriol. I 76, 6255—6261.

9 Dinglay, S., Trewick, S. C., Lindahl, T., and Sedgwick, B. (2000) Defective processing
of methylated single-stranded DNA by E. coli alkB mutants. Genes Develop. 14, 2097-
2105.

1° Aas, P. A., Otterlei, M., Falnes, P. 0., Vagbe, C. 13., Skorpen, F., Akbari, M.,
Sundheim, 0., Bjoras, M., Slupphaug, G., Seeberg, E., and Kurkan, U. E. (2003) Human
and bacterial oxidative demethylases repair alkylation damage in both RNA and DNA.
Nature 421, 859—863.

in Lloyd, M. D., Merrit, K. D., Lee, V., Sewell, T. J ., Wha-Son, 3., Baldwin, J. E.,
Schofield, C. J ., Elson, S. W., Baggaley, K. H., and Nicholson, N. H. (1999) Product-

39

substrate engineering by bacteria: studies on Clavaminate synthase, a trifunctional
dioxygenase. Tetrahedron 55, 10201-10220.

'2 Van Der Ploeg, J. R., Weiss, M. A., Saller, E., Nashimoto, H., Saito, N., Kertesz, M. A.,
and Leisinger, T. (1996) Identification of sulfate starvation-regulated genes in Escherichia

coli: a gene cluster involved in the utilization of taurine as a sulfur source. J. Bacteriol.
I 78, 543 8-5446.

‘3 Fukumori, F., and Hausinger, R. P. (1993) Alcaligenes eutrophus JMP134 "2,4-
dichlorophenoxyacetate monooxygenase" is an a-ketoglutarate-dependent dioxygenase. J.
Bacteriol. 1 75, 2083-2086.

‘4 Eichhom, E., van der Ploeg, J. R., Kertesz, M. A., and Leisinger, T. (1997)
Characterization of a-ketoglutarate-dependent taurine dioxygenase fi'om Escherichia coli.
J. Biol. Chem. 272, 23031-23036.

15 Gissen, P., Preece, M. A., Willshaw, H. A., and McKieman, P. J. (2003) Ophthalmic
follow-up of patients with tyrosinaemia type I on NTBC. J. Inherit. Metab. Dis. 26, 13-6.

'6 Moran, G. R. (2005) 4-Hydroxyphenylpyruvate dioxygenase. Arch. Biochem. Biophys.
433, 117-128.

'7 Zhang, Z., Ren, J ., Starnmers, D. K., Baldwin, J. E., Harlos, K., and Schofield, C. J.
(2000) Structural origins of the selectivity of the trifunctional oxygenase clavaminic acid
synthase. Nat. Struct. Biol. 7, 127-133.

‘8 Lloyd, M. 1)., Lee, H.-J., Harlos, K., Zhang, Z.-H., Baldwin, J. 13., Schofield, c. J.,
Charnock, J. M., Garner, C. D., Hara, T., Terwisscha Van Scheltinga, A. C., Valegard, K.,
Viklund, J. A. C., Hajdu, J ., Andersson, 1., Danielsson, A., and Bhikhabhai, R. (1999)
Studies on the active site of deacetoxycephalosporin C synthase. J. Mol. Biol. 287, 943-
960.

‘9 Elkins, J. M., Ryle, M. J., Clifton, I. J., Dunning Hotopp, J. 0., Lloyd, J. 3., Burzlaff, N.
I., Baldwin, J. E., Hausinger, R. P., and Roach, P. L. (2002) X-ray crystal structure of
Escherichia coli taurine/a-ketoglutarate dioxygenase complexed to ferrous iron and
substrates. Biochemistry 41, 5185-5 192.

2° Muller, I., Kahnert, A., Pape, T., Sheldrick, G. M., Meyer-Klaucke, W., Dierks, T.,
Kertesz, M., and Uson, I. (2004) Crystal structure of the alkylsulfatase AtsK: insights into
the catalytic mechanism of the Fe(II) a-ketoglutarate-dependent dioxygenase superfamily.
Biochemistry 43, 3075-3088.

40

2' Clifton, I. J ., Doan, L. X., Sleeman, M. C., Topf, M., Suzuki, H., Wilmouth, R. C., and
Schofield, C. J. (2003) Crystal structure of carbapenem synthase (CarC). J. Biol. Chem.
278, 20843-20850.

22 Stirk, H. J ., Woolfson, D. N., Hutchinson, E. G., and Thornton, J. M. (1992) Depicting
topology and handedness in jellyroll structures. F EBS Lett. 308, 1-3.

23 Picture taken from: http://courses.cm.utexas.edu/jrobertus/ch339k/overheads-
l/JellvRollBarreljpg

 

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

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

26 Proshlyakov, D. A., Henshaw, T. F., Monterosso, G. R., Ryle, M. J ., and Hausinger, R.
P. (2004) Direct detection of oxygen intermediates in the non-heme Fe enzyme taurine/a-
ketoglutarate dioxygenase. J. Am. Chem. Soc. 126, 1022-1023.

27 Riggs-Gelasco, P. J ., Price, J. C., Guyer, R. B., Brehm, J. H., Barr, E. W., Bollinger, J.
M., Jr., and Krebs, C. (2004) EXAFS spectroscopic evidence for an F e=O unit in the
Fe(IV) intermediate observed during oxygen activation by taurinezalpha-ketoglutarate
dioxygenase. J. Am. Chem. Soc. 126, 8108-9.

28 Clifion, I. J ., McDonough, M. A., Ehrismann, D., Kershaw, N. J ., Granatino, N., and
Schofield, C. J. (2006) Structural studies on 2-oxoglutarate oxygenases and related
double-stranded B-helix fold protein. J. Inorg. Biochem. 100, 644-669.

29 Serre, L., Sailland, A., Sy, D., Boudec, P., Rolland, A., Pebay-Peyroula, E., and Cohen-
Addad, C. (1999) Crystal structure of Pseudomonasfluorescens 4-hydroxyphenylpyruvate

dioxygenase: an enzyme involved in the tyrosine degradation pathway. Structure 7, 977-
988.

30 Clifton, I. J ., Hsueh, L.-C., Baldwin, J. E., Harlos, K., and Schofield, C. J. (2001)

Structure of proline 3-hydroxylase. Evolution of the family of 2-oxoglutarate dependent
dioxygenases. Eur. J. Biochem. 268, 6625-6636.

41

3' Valegard, K., Terwisscha van Scheltinga, A. C., Lloyd, M. D., Hara, T., Ramaswamy,
S., Perrakis, A., Thompson, A., Lee, H. J ., Baldwin, J. E., Schofield, C. J ., Hajdu, J ., and
Andersson, I. (1998) Structure of a cephalosporin synthase. Nature 394, 805-809.

32 Pavel, E. G., Zhou, J ., Busby, R. W., Gunsior, M., Townsend, C. A., and Solomon, E. I.
(1998) Circular dichroism and magnetic circular dichroism spectroscopic studies of the
non-heme ferrous active site in Clavaminate synthase and its interaction with a-
ketoglutarate cosubstrate. J. Am. Chem. Soc. 120, 743-753.

33 Pavel, E. G., Zhou, J ., Busby, R. W., Gunsior, M., Townsend, C. A., and Solomon, E. I.
(1998) Circular dichroism and magnetic circular dichroism spectroscopic studies of the
non-heme ferrous active site in Clavaminate synthase and its interaction with a-
ketoglutarate cosubstrate. J Am. Chem. Soc. 120, 743-753.

34 Ho, R. Y. N., Mehn, M. P., Hegg, E. L., Liu, A., Ryle, M. A., Hausinger, R. P., and
Que, L., Jr. (2001) Resonance Raman studies of the iron(Il)-u-keto acid chromophore in
model and enzyme complexes. J. Am. Chem. Soc. 123, 5022-5029.

35 Lee, H. J., Lloyde, M. 1)., Clifton, 1. J., Baldwin, J. 13., and Schofield, C. J. (2001)
Kinetic and crystallographic studies on deacetoxycephalosporin C synthase (DAOCS). J.
Mol. Biol. 308, 937-948.

36 Winters, K. J ., Purpero, V. M., Kavana, M., and Moran, G. R. (2005) Accumulation of
multiple intermediates in the catalytic cycle of (4-hydroxypheny1)pyruvate dioxygenase
from Streptomyces avermitilis. Biochemistry 44, 7189-7199.

37 Price, J. C., Barr, E. W., Tirupati, B., Bollinger, J. M., Jr., and Krebs, C. (2003) The first
direct characterization of a high-valent iron intermediate in the reaction of an a-
ketoglutarate-dependent dioxygenase: a high-spin Fe(IV) complex in taurine/a-
ketoglutarate dioxygenase (TauD) from Escherichia coli. Biochemistry 42, 7497-7508.

3“ Hoffart, L. M., Barr, E. W., Guyer, R. 13., Bollinger, J. M., Jr., and Krebs, C. (2006)
Direct spectroscopic detection of a C-H-cleaving high-spin Fe(IV) complex in a prolyl-4-
hydroxylase. Proc. Natl. Acad. Sci. USA 103, 14738-14743.

39 Riggs-Gelasco, P. J ., Price, J. C., Guyer, R. B., Brehm, J. H., Barr, E. W., Bollinger, J.
M., Jr., and Krebs, C. (2004) EXAFS spectroscopic evidence for an F e=O unit in the

Fe(IV) intermediate observed during oxygen activation by taurine:alpha-ketoglutarate
dioxygenase. J. Am. Chem. Soc. 126, 8108-8109.

42

4° Winters, K. J ., Purpero, V. M., Kavana, M., Nelson, T., and Moran, G. R. (2003) (4-
Hydroxyphenyl)pyruvate dioxygenase from Streptomyces avermitilis: the basis for
ordered substrate addition. Biochemistry 42, 2072-2080.

4‘ Koehntop, K. D., Marimanikkuppam, S., Ryle, M. J ., Hausinger, R. P., and Que, L.
(2006) Self-hydroxylation of taurine/a-ketoglutarate dioxygenase: evidence for more than
one oxygen activation mechanism. .1. Biol. Inorg. Chem. 11, 63-72.

42 Liu, A., Ho, R. Y. N., Que, L., Jr., Ryle, M. J., Phinney, B. s., and Hausinger, R. P.
(2001) Alternative reactivity of an a-ketoglutarate-dependent iron(II) oxygenase: enzyme
self-hydroxylation. .1. Am. Chem. Soc. 123, 5126-5127.

43 Bradley, F. C., Lindstedt, s., Lipscomb, J. D., Que, L., Jr., Roe, A. L., and Rundgren,
M. (1986) 4-Hydroxyphenylpyruvate dioxygenase is an iron-tyrosinate protein. J. Biol.
Chem. 261, 11693-11696.

44 Myllyla, R., Majamaa, K., Gfinzler, V., Hanauske-Abel, H. M., and Kivirikko, K. I.
(1984) Ascorbate is consumed stoichiometrically in the uncoupled reactions catalyzed by
propyl 4-hydroxylase and lysyl hydroxylase. J. Biol. Chem. 259, 5403-5405.

45 Yu, B., Edstrom, W. C., Hamuro, Y., Weber, P. C., Gibney, B. R., and Hunt, J. F.
(2006) Crystal structures of catalytic complexes of the oxidative DNA/RNA repair
enzyme AlkB. Nature 439, 879-884.

46 Chemistry and Biochemistry of Flavoenzymes. Volume I. by Franz Miiller (1991)
published by CRC Inc. Florida.

47 De Colibus, L., and Mattevi, A. (2006) New frontiers in structural flavoenzymology.
Curr. Opin. Struct. Biol. 16, 722-728.

43 Ghishla, and S., and Massey, V. (1989) Mechanisms of flavoprotein-catalyzed
reactions. Eur. J. Biochem. 181,1-17.

49 Fraaije, M., and Mattevi, A. (2000) Flavoenzymes: diverse catalysts with recurrent
features. Trends BiochemSci. 25, 126-132.

50 Massey V. (1995) Introduction: Flavoprotein structure and mechanism. FASEB .1. 9,
473-475.

5' Fitzpatrick, P. F. (2001) Substrate dehydrogenation by fiavoproteins. Acc. Chem. Res.
34, 299-307.

43

52 Becvar, J. E., and Hastings, J. W. (1975) Bacterial luciferase requires one reduced flavin
for light emission. Proc. Natl. Acad. Sci. USA 72, 3374-3376.

53 Komori, H., Masui, R., Kuramitsu, S., Yokoyama, S., Shibata, T., Inoue, Y., and Miki,
K. (2001) Crystal structure of thermostable DNA photolyase: pyrimidine-dimer
recognition chemistry. Proc. Natl. Acad. Sci. USA 98, 13560-13565.

54 Susin, S. A., Lorenzo, H. K., Zamzani, N., Marzo, 1., Snow, B. E., Brothers, G. M.,
Mangion, J ., Jacotot, E., Constantini, P., and Loeffler, M. (1999) Molecular
characterization of mitochondrial apoptosis-inducing factor. Nature 397, 441-445.

55 Woodmansee, A. N., and Imlay, J. A. (2002) Reduced flavin promotes oxidative DNA
damage in non-respiring Escherichia coli by delivering electrons to intracellular free iron.
J. Biol. Chem. 277, 34055-34066.

56 Walsh, C. (1980) Flavin coenzymes: at the crossroads of biological redox chemistry.
Acc. Chem. Res. 13, 148-155.

57 Miura, R. (2001) Versatility and specificity in flavoenzymes: control mechanisms of
flavin reactivity. Chem. Rec. 1, 183-194.

58 Massey, V. (2000) The chemical and biological versatility of riboflavin. Biochem. Soc.
Trans. 28, 283-296.

59 Barman, B. G., and Tollin, G. (1972) Flavin-protein interactions in flavoenzymes.

Thermodynamics and kinetics of reduction of Azotobacter flavodoxin. Biochemistry 11,
4755-4759.

60 Gomez-Moreno, C., Choy, M., and Edmondson, D. E. (1979) Purification and
properties of the bacterial flavoprotein: thiamin dehydrogenase. J. Biol. Chem. 254, 7630-
7635.

6‘ Burnett, R. M., Darling, G. D., Kendall, D. S., LeQuesne, M. S., Mayhew, S. G., Smith,
W. W., and Ludwig, M., L. (1974) The structure of the oxidized form of Clostridial

flavodoxin at 1.9-A resolution. Description of the flavin mononucleotide binding site. J.
Biol. Chem. 249, 4383-4392.

62 Pace, C., and Stankovich, M. (1986) Oxidation-reduction of glycolate oxidase.
Biochemistry 25 , 2516-2522.

63 Decker, K. F. (1993) Biosynthesis and function of enzymes with covalently bound
flavin. Ann. Rev. Nutr. 13,17-41.

44

64 Fraaije, M. W., van de Hawvel, R. H. H., van Berkel, W. J. H., and Mattevi, A. (1999)
Covalent flavinylation is essential for efficient redox catalysis in vanillyl-alcohol oxidase.
J. Biol. Chem. 274, 35514-35520.

65 Massey, V., Stankovich, M., and Hemmerich, P. (1978) Light mediated reduction of
flavoproteins with flavin as catalysts. Biochemistry 1 7, 1-8.

66 Massey, V., and Hemmerich, P. (1978) Photoreduction of flavoproteins and other
biological compounds catalyzed by deazaflavins. Biochemistry 1 7, 9-17.

67 Rohankhedkar, M. S., Mulrooney, S. B., Wedemeyer, W. J ., and Hausinger, R. P.
(2006) The AidB component of the Escherichia coli adaptive response to alkylating
agents is a flavin-containing, DNA-binding protein. J. Bacteriol. 188, 223-230.

68 Massey, V. ( 1994) Activation of the molecular oxygen by flavins and flavoproteins. J.
Biol. Chem. 269, 22459-22462.

69 Kemal, C., Chan, T. W., and Bruice, T. C. (1977) Reaction of 302 with dihydroflavins.
1. N3,5-dimethyl-1,5-dihydrolumiflavin and 1,5-dihydroisoalloxazines. J. Amer. Chem.
Soc. 99, 7272-7286.

70 Roth, J. P., and Klinman, J. P. (2003) Catalysis of electron transfer during activation of
02 by the flavoprotein glucose oxidase. Proc. Natl. Acad. Sci. USA. 100, 62-67.

7‘ Macheroux, P., Massey, V., and Thiele D. J. (1991) Expression of spinach glycolate
oxidase in Saccharomyces cerevisiae: purification and characterization. Biochemistry 30,
4612-4619.

72 Massey, V., Muller, F., Feldberg, R., Schuman, M., Sullivan, P. A., Howell, L. G.,
Mayhew, S. G., Matthews, R. G., and F oust, G. P. (1969) The reactivity of flavoproteins
with sulfite. J. Biol. Chem. 244, 3999-4006.

73 Muller, F., and Massey, V. (1969) Flavin-sulfite complexes and their structures. J. Biol.
Chem. 244, 4007-4016.

74 Job, V., Marcone, G. L., Pilone, M. S., and Pollegioni, L. (2002) Glycine oxidase from
Bacillus subtilis. J. Biol. Chem. 277, 6985-6993.

75 Mattevi, A., (2006) To be or not to be an oxidase: challenging the oxygen reactivity of
flavoenzymes. Trends Biochem. Sci. 31, 276-283.

45

76 Van Berkel, W. J. H., Kamerbeek, N. M., and Fraaije, M. W. (2006) Flavoprotein
monooxygenases, a diverse class of oxidative biocatalysts. J. Biotechnol. 124, 670-689.

77 Ghisla, S., Thorpe, C., and Massey, V. (1984) Mechanistic studies with general acyl-
CoA dehydrogenase and butyryl-CoA dehydrogenase: evidence for the transfer of the [3-
hydrogen to the flavin N(5)-position as a hydride. Biochemistry 23, 3154-3161.

78 Thorpe, C., and Kim, J. J. P. (1995) Structure and mechanism of action of the acyl-CoA
dehydrogenases. FASEB J. 9, 718-725.

79 Rao, K. S., and Lederer, F. (1998) About the pKa of the active-site histidine in
flavocytochromes b2 (yeast L-lactate dehydrogenase). Protein Sci. 7, 1531-1537.

80 Hille, R. (1996) The mononuclear molybdenum enzymes. Chem. Rev. 96, 2757-2816.

8' Warburg, O. and Christian, W. (1932) Bin zweites sauerstoffiibertragendes Ferment und
sein Absorptionspektrum. Naturwissenschaften 20, 688.

82 Williams, R. E., and Bruce, N. C. (2002) ‘New uses of an old enzyme’ — the old yellow
enzyme family of flavoenzymes. Microbiology 148, 1607-1614.

83 Karplus, P. A., Fox, K. M., and Massey, V. (1995) Structure-function relations for old
yellow proteins. FASEB J. 9, 1518-1526.

46

CHAPTER 2

Kinetic and Spectroscopic Investigation of Co", Ni", and N-

Oxalylglycine Inhibition of the Fen/a-Ketoglutarate Dioxygenase, TauD

Piotr Grzyska carried out analysis of the electronic spectra of Con-substituted

forms of TauD.

47

ABSTRACT

Co", Ni", and N-oxalylglycine (NOG) are well-known inhibitors of Fell/0t-
ketoglutarate (aKG)-dependent hydroxylases, but few studies describe their kinetics and
no spectroscopic investigations have been reported. Using taurine/aKG dioxygenase
(TauD) as a paradigm for this enzyme family, time-dependent inhibition assays showed
that Co" and Ni" follow slow-binding inhibition kinetics. Whereas Ni'l-substituted TauD
was non-chromophoric, spectroscopic studies of the Con-substituted enzyme revealed a 6-
coordinate site (protein alone or with aKG) that became S-coordinate upon taurine
addition. The C011 spectrum was not perturbed by a series of anions or oxidants,
suggesting the Co" is inaccessible and could be used to stabilize the protein. NOG
competed weakly (Ki ~290 uM) with aKG for binding to TauD, with the increased
electron density of NOG yielding electronic transitions for NOG-Feu-TauD and taurine-
NOG-Fen-TauD at 380 nm (s380 90-105 M" x s"). The spectra of the NOG-bound TauD
species did not change significantly upon oxygen exposure, arguing against the formation

of an oxygen-bound state mimicking an early intermediate in catalysis.

48

INTRODUCTION

Fen/a-ketoglutarate (aKG)-dependent dioxygenases couple the oxidative
decarboxylation of aKG to the oxidation of their primary substrates using a mononuclear,
non-heme iron metallocenter.l These enzymes catalyze a wide variety of crucial chemical
transformations including the repair of alkylation damage in DNA or RNA,2’3 sensing of
hypoxia,4 modification of structural proteins,5 synthesis of various metabolites ranging
from antibiotics to plant hormones,6‘7‘8 and degradation of compounds such as herbicides
and phytanic acids.9’l0 The focus of this study is the archetype member of this enzyme
family, Escherichia coli taurine/aKG dioxygenase (TauD) that metabolizes aminoethane-
sulfonate (Figure 5) to produce sulfite as a cellular sulfur source. ”

Recent crystallographic and spectroscopic studies of TauD have confirmed many
aspects of the general enzyme mechanism of Fell/aKG-dependent dioxygenases (Scheme
4) that was first proposed over two decades ago.'2 Structural studies reveal that the Fe"
center is ligated by three amino acid side chains on the same face of the metal: His99,
Asp101, and Hi5255.l3‘l4 In the absence of substrate, three water molecules complete the
six-coordinate environment (A). Two waters are displaced upon binding of ocKG (shown
as RCOCOO'), which coordinates the Fe" center through its C-l carboxylate and C-2
carbonyl moieties (B) producing a diagnostic metal-to-ligand charge-transfer transition
with a kmax at 530 nm and 8530 of 140-240 M’1 X cm".'5‘l6 Taurine (illustrated by R'-H)
binds near the active site and promotes dissociation of the remaining water ligand to shift
the 7mm to 520 nm with 8520 = 180-270 M'l >< cm",'5‘l6 leaving the Fe" five-coordinate and

Ill

primed to react with oxygen (C). Binding of oxygen produces a yet uncharacterized Fe -

superoxo or Few-peroxo species (D) that attacks the ocKG carbonyl group, leading to

49

decomposition of aKG and heterolytic O-O bond cleavage. The resulting Few-0x0 species
(E) inserts oxygen into the target C-H bond of the substrate by hydrogen atom transfer and
oxygen rebound, as found in heme-type oxygenases, to restore the Fe'1 state of the
enzyme. The Few-0x0 species has been identified on the basis of stopped-flow UV/visible
spectroscopy, freeze-quench Mtissbauer analyses, EPR spectroscopy of cryoreduced
sample, cryogenic continuous-flow resonance Raman studies, and X-ray absorption
Spectroscopy.l6,l7,18,l9,20,21

Here, we examine the kinetics and spectroscopy of TauD interaction with three
well-known inhibitors of this class of enzyme. CoII and Ni" inhibit prolyl and asparaginyl
hydroxylases that target the hypoxia inducible factor (HIF), involved in oxygen sensing,22
so the metal ions lead to cellular gene expression changes mimicking those observed
during hypoxia?"24 In addition, Co" inhibits several other enzyme family members
including TauDll and l-aminocyclopropanecarboxylate oxidase, for which the Co-bound
crystal structure is known.25 Despite the importance of these inhibitory metal ions to the

filnctioning of Fell/aKG-dependent dioxygenases, few kinetic characterization studies and

no spectroscopic investigations have been reported. Similarly, the CLKG analogue N-
oxalylglycine (NOG) is an established inhibitor of procollagen prolyl 4-hydroxylase2‘5’27
as well as HIP-specific prolyl and asparaginyl hydroxylases.28‘29‘30 In the asparaginyl
hydroxylase known as factor inhibiting HIF (F 1H), the structure of the NOG-bound
enzyme was determined and NOG was found to coordinate Fe in the same manner as the
aKG shown in B of Scheme 4.3 I Furthermore, inhibition of the iron-mediated degradation

of the iron regulatory protein 2 by dimethyl-NOG (which is hydrolyzed to NOG by

esterases within the cell) was cited as evidence for the participation of an Fell/aKG-

50

dependent dioxygenase in this pathway.32‘33 As with the inhibitory metals, few NOG-
related studies address the kinetics of inhibition and none examine the spectroscopic
properties of the inhibited enzyme. We’ve chosen to study these aspects of Co", Ni", and
NOG interaction with TauD as a paradigm for related systems. In addition, we sought to

obtain new insights into the structures and properties of the early intermediates in TauD

 

 

 

catalysis.
R'-H
R'-H
01712
p,. x0 R A813 4"0\ R
B F1301) Fe(II) C
Hls . O HlsH o O
HIS - 0 H20
_ O2
Rcocoo
2H20 \
-- 9 R'-H =
WOHO RH 120 R D Asp. 92 R
Aspm .~~‘ 2 Aspa, ,- , i,’ \
A Hts’FilH)‘H20 Fe(lll) \ Femool
HIS His/ 1‘0 His/l ‘0 o
R'-OH His- 0 HIS -
RCOO‘ e —
R-i: - R
3H20 A 9,0 C02
SP”;Fe(IV) 0
His ,l
HIS
E

Scheme 4. Catalytic mechanism of T auD.

51

EXPERIMENTAL PROCEDURES

Purification of T auD apoprotein.

Cultures of E. coli BL21(DE3) with pME4l4l,” containing tauD under the
control of the T7 RNA polymerase promoter, were grown in TB medium (1 L) at 37 °C
with stirring at 200 rpm. When cultures reached an A600 of ~0.4, tauD expression was
induced by the addition of isopropyl B-D-thiogalactopyranoside to a final concentration of
0.1 mM. Cells were harvested after ~12 h by centrifugation for 10 min at 10,000 g and 4
°C. The cell pellet was resuspended in 20 mL of 20 mM Tris with 1 mM EDTA (TE)
buffer (pH 8.0). The sample was stored at -80 0C until disrupting the cells by using a
French pressure apparatus. The lysate was clarified by centrifugation for 40 min at
150,000 g and applied to a DEAE-Sepharose column (5.0 cm X 30 cm, Amersham
Biosciences). The column was rinsed with 2 column volumes of TE buffer (pH 8.0) and
eluted by using a linear 1500 ml gradient from 70 to 230 mM NaCl in the same buffer at a
flow rate of 7 ml/min. Fractions containing TauD were concentrated in an Amicon stirred
cell concentrator with a 30 kDa cutoff membrane. The concentrated eluant was loaded
onto a high-perforrnance phenyl-Sepharose column (2.5 cm X 30 cm, Amersham
Biosciences) equilibrated with TE buffer (pH 8.0) containing 500 mM (NH4)2$O4.
Proteins were eluted with a 1200 mL linear gradient from 500 to 0 mM (NH4)2SO4 in the
same buffer at a flow rate of 5 ml/min. The TauD-containing fractions were concentrated
and extensively dialyzed against 25 mM Tris buffer (pH 8.0) at 4°C. Purified TauD
apoprotein exhibited a single 32.2 kDa band when examined by denaturing
polyacrylamide gel electrophoresis. Protein concentrations were estimated by using 8230

46,400 M'l X cm]. The dialyzed apoprotein was stored frozen at -80 °C.

52

Enzyme assays.

TauD activity was measured by using Ellman's reagent to quantify sulfite, as
previously described.Il One unit (U) of enzyme activity is defined as the amount of
enzyme that releases 1 umol of sulfite per min at 30 °C in assay buffer containing 25 mM
Tris (pH 8.0), 50 or 100 M Fe", 50 or 100 M ascorbate, 100 or 500 M OLKG, and 1

mM taurine. The TauD used in these studies had a specific activity ranging from 3 to 6.4

U (mg of protein)".

Characterization of inhibition kinetics.

Steady-state inhibition assays were carried out during 5 min incubations using the
typical assay conditions (sometimes with one component varied in concentration) and
amended with the indicated concentrations of inhibitor. The assays were initiated either by
adding enzyme or by adding Fe" to the assay buffers. In addition, the time dependence of
inhibition was assessed in some cases by removing timed aliquots to an EDTA quench

solution.

Electronic spectroscopy.

Spectra were recorded at room temperature on a Shimadzu UV-2401 UV/visible
spectrophotometer. All stock solutions for anaerobic binding studies were prepared inside
serum vials sealed with butyl rubber stoppers and purged of oxygen by several rounds of
vacuum degassing and flushing with argon using a vacuum manifold. Stock solutions of

orKG (50 or 100 mM), taurine (50 or 100 mM), and NOG (kindly provided by Dr. Nicolai

Burzlaff) (100 mM) were prepared in 25 mM Tris buffer (pH 8.0). Ferrous ammonium

53

sulfate stock solutions (25 mM) containing 5 mM ascorbate were prepared by several
rounds of degassing and argon flushing of the solids, followed by addition of the desired
volume of H20. CoClz and MG; stock solutions (25 mM) were prepared by several
rounds of degassing and flushing with argon inside a sealed serum vial. TauD apoprotein
(0.25 or 0.5 mM subunit in 25 mM Tris buffer, pH 8.0) was placed into a 1 cm path
length, 1 mL quartz cuvette fitted with a stopper and purged with argon. Other
components were added by using gastight syringes (Hamilton) that had been flushed with
anaerobic buffer. Selected samples were mixed with an equal volume of buffer sparged
with 100 % Oz, and spectral changes were monitored over time. The effect of added H202

was examined in one sample. All spectra were corrected to account for sample dilutions.

54

RESULTS AND DISCUSSION

Kinetics of Inhibition of TauD by C o” and NiII

The effects of Co" and Ni" on TauD activity were examined by using two methods
to initiate the assays. First, TauD apoprotein was added to standard assay mixes containing
FeII and varied concentrations of the inhibitory metal ions, incubated for 5 min, and the
total sulfite product measured (Figure 22, solid lines). The data were fit to equation 1
(where [M] is the inhibitory metal ion concentration) to determine the ICso (inhibitor
concentration resulting in 50% inhibition, with the approximate range of this value shown
in parentheses). The ICSO of Co" was 41 11M (30-70 uM) while that of Ni'1 was 32 11M (20-
40 uM). When the TauD inhibition assays were repeated using Fell addition to initiate the
reaction (Fig. 1, dashed lines), the observed IC50 values were 1.9 uM (l-3.5 HM) and 0.71
11M (O.60-1.0 uM) for Co" and Ni", respectively. The large differences in ICso values
observed when using the distinct methods to initiate the reaction suggest that FeIl does not
readily displace the inhibitory metal ions previously bound to the protein. These
differences also highlight the fact that steady-state assays that assume rapid equilibrium
kinetics are inadequate for defining the true kinetic inhibition mechanism of metal ions.
Despite this caveat, the ICSO values obtained for TauD were compared to those reported
for three human HIF-specific prolyl 4-hydroxylase isozymes (38 :i: 8, 100 i 15, and 9 i 4
pM for Co"; 130 :t 76, >1000, and 120 i 49 11M for Ni”) and a collagen-specific prolyl 4-
hydroxylase (14 i 3 uM for Coll and 37 i 11 uM for Ni”) along with the K, values
estimated for FIH inhibition (1.0 a 0.4 M for Co" and 4 a 1 M for Ni") 22. Curiously,

C0" and Ni" inhibition of the HIF prolyl 4-hydroxylases was incomplete with up to 50 %

55

activity remaining at 0.5 mM Co” and up to 55 % activity remaining at 1 mM Ni”. We
attribute the incomplete inhibition of the prolyl 4-hydroxylases to their purification as

partial holoproteins, compared to TauD that was purified as the apoprotein.

% activity remaining = 100 — 100[M]/(IC50 + [M]) Equation 1

 

% Activity
8

8

8

 

 

 

 

 

 

01 1 10 100

Cobalt (11M) Nickel (uM)

Figure 22. Inhibition of TauD by Coll and Ni". The concentrations of sulfite produced
during 5 min incubations were used to assess the percent activity of TauD in standard
assay conditions containing the indicated concentrations of Col1 (left panel) and Nill (right
panel). The solid circles represent data associated with assays initiated with TauD
apoprotein, whereas the open circles represent data for assays initiated by FeII addition to
samples exposed to inhibitory metal ions for 2 min. The data were fit to equation 1 to
calculate IC50 values.

To better define the kinetic mechanism of inhibition by metal ions, the time-
dependence of Co" and Ni" inhibition were determined (Figure 23A and Figure 23B). In
the absence of inhibitory metal ion, sulfite production began immediately (i.e., TauD
apoprotein binds F e” rapidly) and increased steadily. When various concentrations of Co"
or NilI were added at 23 sec into the assays, the rates of sulfite production were observed

to decrease over time. The apparent first-order rate constants of enzyme inactivation were

56

calculated for each assay according to equation 2 (where P, is the amount of product at
time t, v, is the initial rate, and kmaa is the apparent inactivation rate constant), and the
values replotted as a function of inhibitory metal ion concentrations (Figure 23C and
Figure 23D). The apparent kmact values were observed to saturate at high concentrations of
metal ions, consistent with slow-binding inhibition kinetics (Scheme 5).34 The initial
dissociation constant K, (kn/kl) was estimated to be 600 i 180 uM for Coll and 166 i 65

uM for Ni", and k3 (equivalent to (kmact)max) was estimated as 0.044 at 0.008 s'1 for Co”

and 0.078 :- 0.0012 s" for Ni".

P, = v, (1-exp(-kmactt))/kmam Equation 2

57

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A) 30 E '3 B) 35 . i
F 1 3° 3 ‘3
. 1 > ‘
b 7 h 7
A e 4 25 L .1
E : 1 g I 1
1 v n b . 4
v m ,._ —-< 1 h 1
“a? : : : 20 r 1
t . 0 t 1
a: > ‘ t ' 4
3 T 7 ’5 15 '— -‘
(I) . . m , .
‘3‘ : I I 3
10 ~ “ 2' 10 i ‘
; v 3 t 1
i- ‘ i A 5 E I
‘ '1
i . i i
0 1.4.1....l1...1.LA.1L...1.A. OLA“.1_1‘_1‘A‘AL1L1‘11‘JA1__A‘1
o 100 200 300 0 so 100 150 200 250 300
tlme (sec) time (sec)
C 0.0% >1 7‘7! rT v v v I 1 v Y Y I v Y rt I v v v '1 rY v v '1 v 1 v I v Y Y I4 D hry-y '1 r v Y I I v V V V T ‘ I I I 1‘
) i ) : . I
r ‘l 7 4
0.006 _- ° 1
0.02 - d . ‘
F A . 4
' . . 4
FA 1 A .— . .1
'8 0015 ' i ‘8 I . :
I — d r T
3.1 _ . g; 0.004 E i
Q ‘ Q .
8 ,. , . l
A“ A“ h. ‘
a; 001 — . ‘ *3 ' ‘
5 i ° V 0.002 ”- -‘
’ : 0 i
0.005 - o 4 I 1
lo 0 1
o l- 4 L. l l ‘ ‘ A ‘ L ‘ A l l ‘ LAALLL ‘ ‘ l ‘ A | t l L ‘ A A l A ‘ A ‘4 o A ALIA A A A I I A A A I A A A A I A A A A l A A A A l A A A A I A A A A
o 200 400 000 80C 0 20° 40° 60° 80°
[CO] (0M) 1“" 0’”)

Figure 23. Time-dependent inhibition of TauD by Con and Ni". TauD apoprotein was
added to standard assay conditions and varied concentrations of Co'1 or NiII (including 0
uM, o; 50 1.1M, I, 100 uM, O; 300 uM, V, and 700 uM, A) were added after 23 sec
(equivalent to the zero time point in the subset of studies illustrated in panels A and C).
Subsequent productions of sulfite were analyzed according to equation 2 to determine the
effects of inhibitory metal concentrations on kinaa, the apparent rates of inactivation, as
illustrated in panels B and D.

 

 

 

 

E+l - 7 El - 7 El*
[(1 k-Z

Scheme 5. Slow-binding inhibition kinetics.

58

Spectroscopy of C0I I and Nill interaction with TauD

The addition of Co” to an anaerobic sample of TauD (Figure 24A) resulted in
formation of a broad peak between 450 and 600 nm, with Ame at 530 nm and 8530 of 70 M'
I X cm'I (calculated on the basis of the difference spectrum of Co'l-TauD minus TauD).
The addition of CLKG had little effect on the spectrum, whereas the further addition of
taurine resulted in features at 565, 552, and 500 nm with extinction coefficients of 204,
200, and 127 M" X cm’l (calculated on the basis of the difference spectrum for taurine-
aKG-Con-TauD minus TauD). The magnitude of the Con extinction coefficient has been
shown empirically to correlate with the coordination number in Coll proteins: six-
coordinate sites have extinction coefficients about 50 M'I x cm", five-coordinate sites
have values between 50 and 300 M" x cm", and four-coordinate sites possess a coefficient

of more than 300 M'1 x cm".35 We conclude that the C0" sites in Con-TauD and aKG-

Con-TauD are six-coordinate, whereas that in taurine-ocKG-Con-TauD is five-coordinate.
Thus, the binding of substrate most likely leads to dissociation of a water molecule in this
metal-substituted protein just as in the active, Fell-containing enzyme.

We examined the accessibility of TauD-bound C0 to exogenous ligands
using both taurine-aKG-Con-TauD and orKG-Cou-TauD. No spectral changes were
detected for either species upon addition of CN', OCN', SCN', or ClO' at 4 mM
concentrations, or in the presence of 200 mM NaCl. Although the metal coordination
environment of taurine-aKG-Con-TauD appears to closely resemble that of taurine-aKG-

Fen-TauD, the spectrum of the Con-containing species remained unchanged when exposed

to oxygen. Furthermore, despite precedence for the transformation of Col1 species to Com-

59

36.37

OOH model compounds, no spectral perturbations were observed when H202 (up to

750 uM) was added to taurine-orKG-Co'l-TauD.
In contrast to the situation for Co”, the addition of Nill did not affect the

spectra of TauD, protein plus orKG, protein with both substrates, or for these components

plus oxygen (data not shown).

 

 

 

 

 

 

 

 

A) B)
250‘ A 120:
f f“; E 100 4 ,‘Jr‘.
1100-: . v '
c 11 ,9 IN»: . L
g, 150 “'7’ .3.“ g
2 ‘Kwe 2
g I”: ... a g
(5 504 0.5. ~ 1 (U
B , K“ 3
otoovtso'soo 580'600‘67507700‘7750 0 I 450 550 650 F7507 600
e (M‘1 cm") A€(M‘1cm'1)
C) A
E
C
V
.C
'8)
C
9. .
:5
° 150* 500 ' an shale
Wavelength (nm)
A: (1111" cm
1)

Figure 24. Electronic spectra of Coll-substituted, NOG-Fe", and aKG-Fe" forms of
TauD. (A) An anaerobic solution of TauD apoprotein (550 uM subunit, circles) was
adjusted to contain near stoichiometric amounts of Con (inverted triangles), 2 mM aKG
(squares), and 2 mM taurine (triangles) in 25 mM Tris buffer, pH 8.0. (B) Analogous
spectra were collected for 250 11M TauD while substituting F e“ for ColI and 10 mM NOG
for aKG, with difference spectra shown for the taurine-NOG-Fe'l-TauD minus Fen-TauD
(dashed line) and the NOG-Fen-TauD minus Fell-TauD samples (solid line). (C)
Analogous spectra were collected using Fe" and aKG, again showing difference spectra

60

for the taurine-aKG-Fen-TauD minus F eH-TauD (dashed line) and the aKG-Fe'l-TauD
minus Fell-TauD (solid line) samples.

Kinetics of Inhibition of T auD by NOG

Steady-state kinetic inhibition studies (data not shown) were used to establish that
NOG competed with CLKG for binding to TauD, with a K, of 290 i 90 1.1M. N-oxalyglycine
should be able to form all of the ionic interactions of aKG and it contains an oxamic acid
moiety as a potentially better bidentate ligand for the enzyme-bound Fe", so we were
surprised by the weak inhibition of TauD by this compound. For comparison, NOG was
shown to be a competitive inhibitor of collagen prolyl 4-hydroxylase and FIH with
markedly different K, values of 1.9 to 7.0 uM and 1.2 i 0.3 mM.2("30 The reasons for this
wide range of K, values for these enzymes remain unclear, but interactions with the protein
side chains probably play a role in determining this value. Studies with the structurally
related compound N-oxalyl-D-phenylalanine provided a Ki of 83 i 18 [AM for F IH,l4-fold

lower than that of orKG.30

Spectroscopy of NOG interaction with T auD

The spectra of NOG-Feu-TauD and taurine-NOG-Fe'l-TauD (Figure 24B) differ
significantly from the previously described spectra of orKG-Fe'l-TauD and taurine-ocKG-
Fen-TauD (shown for comparison in Figure 24C). We interpret this result in terms of the
excess electron density of NOG compared to orKG (note the trianionic form of NOG
versus the dianionic aKG, Scheme 6) which causes a shift of the metal-to-ligand charge-
transfer transition to higher energy. While aKG-Fen-TauD is known to react with oxygen

resulting in aKG decomposition, protein self hydroxylation, and the generation of a 550

61

nm chromophore,38 the spectra of NOG-FeH-TauD and taurine-NOG-FeH-TauD did not
change at a significant rate upon exposure to oxygen. The reduced 02 reactivity for these
species also contrasts with that observed for the product complex (succinate-FeH-TauD)
that generates a 720 nm chromophore when exposed to 02.39 The aKG- and succinate-
derived ligand-to-metal charge-transfer transitions arise from chelation of Fem by the
catecholate produced by hydroxylation of Tyr73, with or without bound bicarbonate
ligand.39 We conclude that hydroxylation of Tyr73 is greatly reduced using NOG-Fe"-

TauD and that NOG does not undergo oxidative decarboxylation.

— 0 _ O\H/i ' of
0 o
a-Ketoglutarate N-Oxalylglycine deprotonated N-oxaiylglycine

Scheme 6. Binding mode of aKG and postulated modes for binding NOG to iron
metal ion.

62

CONCLUSIONS

Steady-state kinetic approaches are inappropriate for determining the kinetics of
inhibition of Fell/aKG-dependent hydroxylases by metal ions; rather, slow-binding kinetic
inhibition methods must be utilized. The chromophore generated upon binding of both
substrates to Con-substituted protein might serve as a useful diagnostic marker for this
enzyme family. While the resulting Co" center is most likely 5-coordinate in TauD, it does
not react with added ligands including oxidants. The low reactivity of the Co"- and Ni"-
substituted enzymes could be exploited during purification and crystallization efforts to
stabilize related proteins against self-hydroxylation reactions. Despite the close structural
similarity of NOG to ocKG, this is a weak competitive inhibitor of TauD and the K ranges
widely for different enzymes. The NOG-bound state of the enzyme does not react with

oxygen to catalyze oxidative decarboxylation.

63

REFERENCES

1 Hausinger, R. P., (2004) Fe(II)/a-ketoglutarate-dependent hydroxylases and related
enzymes. Crit. Rev. Biochem. Mol. Biol. 39, 21-68.

2 Trewick, S. C., Henshaw, T. F., Hausinger, R. P., Lindahl, T., and Sedgwick, B. (2002)
Oxidative demethylation by Escherichia coli AlkB directly reverts DNA base damage.
Nature 419, 174-178.

3 Aas, P. A., Otterlei, M., Falnes, P. O., Vagbe, C. B., Skorpen, F., Akbari, M., Sundheim,
O., Bjoras, M., Slupphaug, G., Seeberg, E., and Krokan, H. E. (2003) Human and bacterial

oxidative demethylases repair alkylation damage in both RNA and DNA. Nature 421 ,
859-863.

4 Schofield, C. J. and Ratcliffe, P. J. (2004) Oxygen sensing by hydroxylases. Nat. Rev.
Molec. Cell Biol. 5, 343-354.

5 Kivirikko, K. I., and Pihlajaniemi, T. (1998) Collagen hydroxylases and the protein
disulfide isomerase subunit of prolyl 4-hydroxylase. Adv. Enzymol. Rel. Areas Mol. Biol.
72, 325-398.

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

7 Dixon, R. A., and Steele, C. L. (1999) Flavonoids and isoflavonoids - a gold mine for
metabolic engineering. Trends Plant Sci. 4, 394-400.

8 Hedden, P. (1999) Recent advances in gibberellin biosynthesis. J. Exper. Bot. 50, 553-
563.

9 F ukumori, F ., and Hausinger, R. P. (1993) Purification and characterization of 2,4-
dichlorophenoxyacetate/a-ketoglutarate dioxygenase. J. Biol. Chem. 268, 24311-24317.

'0 Mukhelji, M., Schofield, C. J ., Wierzbicki, A. S., Jansen, G. A., Wanders, R. J. A., and
Lloyd, M. D. (2003) The chemical biology of branched-chain lipid metabolism. Prog.
Lipid Res. 42, 359-376.

” Eichhom, E., van der Ploeg, J. R., Kertesz, M. A., and Leisinger, T. (1997)

Characterization of a-ketoglutarate-dependent taurine dioxygenase from Escherichia coli.
J. Biol. Chem. 272, 23031-23036.

64

'2 Hanauske-Abel, H. M., and Giinzler, V. (1982) A stereochemical concept for the
catalytic mechanism of prolylhydroxylase. Applicability to classification and design of
inhibitors. J. Theor. Biol. 94, 421-455.

'3 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 Roach, P. L. (2002) X-ray crystal structure of
Escherichia coli taurine/a-ketoglutarate dioxygenase complexed to ferrous iron and
substrates. Biochemistry 41, 5 185-5192.

'4 O'Brien, J. R., Schuller, 1). J., Yang, v. s., Dillard, B. 1)., and Lanzilotta, w. N. (2003)
Substrate-induced conformational changes in Escherichia coli taurine/oc-ketoglutarate
dioxygenase and insight into the oligomeric structure. Biochemistry 42, 5547-5554.

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

16 Grzyska, P. K., Ryle, M. J., Monterosso, G. R., Liu, J ., Ballou, D. P., and Hausinger, R.
P. (2005) Steady-state and transient kinetic analyses of taurine/a-ketoglutarate
dioxygenase: effects of oxygen concentration, alternative sulfonates, and active site
variants on the Fe(IV) intermediate. Biochemistry 44, 3845-3 855.

'7 Price, J. C., Barr, E. W., Glass, T. E., Krebs, C., and Bollinger, J. M., Jr. (2003)
Evidence for hydrogen abstraction from C1 of taurine by the high-spin Fe(IV)

intermediate detected during oxygen activation by taurine:a-ketoglutarate dioxygenase
(TauD). J. Am. Chem. Soc. 125, 13008-13009.

‘8 Price, J. C., Barr, E. W., Tirupati, B., Bollinger, J. M., Jr., and Krebs, C. (2003) The first
direct characterization of a high-valent iron intermediate in the reaction of an 01-

ketoglutarate-dependent dioxygenase: a high-spin Fe(IV) complex in taurine/01-
ketoglutarate dioxygenase (TauD) from Escherichia coli. Biochemistry 42, 7497-7508.

'9 Price, J. C., Barr, E. W., Hoffart, L. M., Krebs, C., and Bollinger, J. M., Jr. (2005)
Kinetic dissection of the catalytic mechanism of taurine:a-ketoglutarate dioxygenase
(TauD) from Escherichia coli. Biochemistry 44, 8138-8147.

20 Proshlyakov, D. A., Henshaw, T. F., Monterosso, G. R., Ryle, M. J ., and Hausinger, R.
P. (2004) Direct detection of oxygen intermediates in the non-heme Fe enzyme taurine/01-

ketoglutarate dioxygenase. .1. Am. Chem. Soc. 126, 1022-1023.

2' Riggs-Gelasco, P. J ., Price, J. C., Guyer, R. B., Brehm, J. H., Barr, E. W., Bollinger, J.
M., Jr., and Krebs, C. (2004) EXAF S spectroscopic evidence for an Fe=O unit in the

65

Fe(IV) intermediate observed during oxygen activation by taurine:a-ketoglutarate
dioxygenase. J. Am. Chem. Soc. 126, 8108-8109.

22 Hirsila, M., Koivunen, P., Xu, L., Seeley, T., Kivirikko, K. 1., and Myllyhalju, J. (2005)
Effect of desferrioxamine and metals on the hydroxylases in the oxygen sensing pathway.
FASEB J. 19, 1308-1310.

23 Goldberg, M. A., Dunning, S. P., and Bunn, H. F. (1988) Regulation of the
erythropoietin gene: evidence that the oxygen sensor is a heme protein. Science 242,
1412-1415.

24 Vengellur, A., Phillips, J. M., Hogenesch, J. B., and LaPres, J. J. (2005) Gene
expression profiling of hypoxia signaling in human hepatocellular carcinoma cells.
Physiol. Genomics 22, 308-318.

25 Zhang, 2., Ren, J.-S., Clifton, 1. J., and Schofield, C. J. (2004) Crystal structure and
mechanistic implications of 1-aminocyclopropane-1-carboxylic acid oxidase--the
ethylene-forming enzyme. Chem. Biol. 11, 1383-1394.

26 Baader, E., Tschank, G., Baringhaus, K. H., Burghard, H., and Gunzler, v. (1994)
Inhibition of prolyl 4-hydroxylase by oxalyl amino acid derivatives in vitro, in isolated
microsomes and in embryonic chicken tissues. Biochem. J. 300, 525-530.

27 Cunliffe, C. J ., Franklin, T. J ., Hales, N. J ., and Hill, G. B. (1992) Novel inhibitors of
prolyl 4-hydroxylase. 3. Inhibition by the substrate analogue N-oxaloglycine and its
derivatives. .1. Med. Chem. 35, 2652-2658.

28 Chan, D. A., Sutphin, P. D., Denko, N. C., and Giaccia, A. J. (2002) Role of prolyl
hydroxylation in oncogenically stabilized hypoxia-inducible factor-101. J. Biol. Chem. 277,
40112-40117.

29 Wojtaszek, P., Smith, C. G., and Bolwell, G. P. (1999) Ultrastructural localization and
further biochemical characterization of prolyl 4-hydroxylase from Phaseolus vulgaris:
comparative analysis. Int. J. Cell Biol. 31 , 463-477.

30 McDonough, M. A., McNeil], L. A., Tilliet, M., Papmichaél, C. A., Chen, Q.-Y.,
Banelji, B., Hewitson, K. S., and Schofield, C. J. (2005) Selective inhibition of factor
inhibiting hypoxia-inducible factor. J. Am. Chem. Soc. 127, 7680-7681.

3‘ Elkins, J. M., Hewitson, K. S., McNeill, L. A., Seibel, J. F., Schlemminger, 1., Pugh, C.
W., Ratcliffe, P. J ., and Schofield, C. J. (2003) Structure of factor-inhibiting hypoxia-

66

inducible factor (HIF) reveals mechanism of oxidative modification of HIF-1 01. J. Biol.
Chem. 278, 1802-1806.

32 Wang, J ., Chen, G., Muckenthaler, M., Galy, B., Hentze, M. W., and Pantopoulos, K.
(2004) Iron-mediated degradation of IRP2, an unexpected pathway involving a 2-
oxoglutarate-dependent oxygenase activity. Molec. Cell. Biol. 24, 954-965.

3’ Hanson, E. S., Rawlins, M. L., and Leibold, E. A. (2003) Oxygen and iron regulation of
iron regulatory protein 2. J. Biol. Chem. 278, 40337-40342.

34 Morrison, J. F., and Walsh, C. T. (1988) The behavior and significance of slow-binding
enzyme inhibitors. Adv. Enzymol. Rel. Areas Mol. Biol. 61, 201-301.

35 Bertini, I., and Luchinat, C. High spin cobalt(II) as a probe for the investigation of
metalloproteins. In Advances in Inorganic Biochemistry, Eichhom, G. L.; Marzilli, L. G.,
Eds. Elsevier Science Publishing Company: New York, NY, 1984; Vol. 6, pp 71-111.

36 Rajani, C., Kincaid, J. R., and Petering, D. H. (2004) Resonance Raman studies of
HOD-Co(III)bleomycin and Co(III)bleomycin: identification of two important vibrational
modes, nu(Co-OOH) and nu(O-OH). J. Am. Chem. Soc. 126, 3829-3836.

37 Chaves, F. A., and Mascharak, P. K. (1999) Co(III)-alkylperoxo complexes: syntheses,
structure-reactivity correlations, and use in the oxidation of hydrocarbons. Acc. Chem.
Res. 33, 539-545.

38 Ryle, M. J ., Liu, A., Muthukumaran, R. 8., Ho, R. Y. N., Koehnt0p, K. D., McCracken,
J ., Que, L., Jr., and Hausinger, R. P. (2003) Oz- and oc-ketoglutarate-dependent tyrosyl

radical formation in TauD, an a-keto acid-dependent non-heme iron dioxygenase.
Biochemistry 42, 1854-1862.

39 Ryle, M. J ., Koehntop, K. D., Liu, A., Que, L., Jr., and Hausinger, R. P. (2003)

Interconversion of two oxidized forms of TauD, a non-heme iron hydroxylase: evidence
for bicarbonate binding. Proc. Natl. Acad. Sci. USA 100, 3790-3795.

67

w

Isolation and characterization of CsiD

Dr. Tina Miiller participated in the cloning, the Biolog plates studies, the GC-MS
analysis, and testing of some molecules as CsiD substrates. Meng Li collected the EPR

spectra. Dr. Jones helped in the GC-MS analysis.

68

ABSTRACT

CsiD, an Escherichia coli protein of unknown function, is a putative 01-
ketoglutarate (aKG)-dependent dioxygenase whose crystal structure has been reported.
The csiD gene is located upstream of the gabDTP-csiR operon and regulated by garbon
starvation induction, perhaps hinting at its physiological role. After cloning and
overexpression of the gene, the CsiD protein was purified and shown to be a tetramer, in
agreement with the literature. Binding of the aKG cosubtrate to the active site of CsiD was
probed by EPR and spectrophotometric methods. The enzyme was found to catalyze an
uncoupled reaction in which oxygen and uKG were consumed in the absence of primary
substrate by an ascorbate-dependent process. Many compounds were tested as potential
substrates of the putative dioxygenase by testing for enhanced oxygen consumption using
an oxygen electrode assay, but without any positive result. Additionally, a series of
comparisons between the E. coli wild type and csiD-knockout strains was undertaken,
including use of Biolog plates, GC-MS metabolic studies and amino acid analysis;

however, the observed differences did not reveal the role of this protein.

69

INTRODUCTION

The Escherichia coli csiD and ygaF genes encode proteins of unknown function
and are positioned adjacent to the gabDTP-csiR operon (Figure 25), leading to the
suggestion that CsiD may participate in the metabolic pathway for y-aminobutyric acid
(GABA) catabolism.7 The gab cluster includes gabP encoding a transport carrier protein
responsible for the uptake of GABA, gabD encoding a succinate semialdehyde
dehydrogenase, and gabT encoding a GABA transaminase.l Succinate, the final product of
the GABA catabolism, is an intermediate of the Krebs cycle. Proposed precursors to
GABA include omithine, putrescine, arginine, and agmatine (Figure 26);2 however, a
recent publication showed that GABA is not typically an intermediate in the catabolism of
arginine or omithine.3 Rather the astCADBE operon is necessary for the catabolism of
arginine to glutamate and AstC is involved in the catabolism of omithine. Nevertheless,
the catabolism of GABA can reasonably contribute to the metabolism of putrescine and
agmatine.3‘4 The ygaF gene encodes a protein whose role has been investigated and is
discussed in the next chapter of my thesis.

The csiD-ygaF-gabDTP-csiR gene cluster constitutes a complex operon which is
regulated by three promoters (Figure 25).5 The first promoter, csiDp, is activated by CsiR
(encoded downstream of the gene cluster) either upon carbon starvation or at stationary
phase and affects the expression of all five genes. The 0’ factor (a subunit of RNA
polymerase) is responsible for the initiation of transcription in bacteria under various
stress conditions in E. coli (e.g. starvation, pH drop, high or low temperature). Gene
activation is triggered by (SS accumulation in cells under specific stress conditions. It was

found that the csiD gene is induced exclusively by carbon starvation and requires a

70

cAMP-CRP activator.6 The other two promoters, gaprr and gaprz, are located before

the gabDTP-csiR operon and are triggered by multiple stress induction and upon nitrogen

starvation, respectively.

 

 

 

 

 

Figure 25. Location of csiD and ygaF and regulation of the csiD-ygaF-gabDTP—csiR
gene cluster. The gene encoding YgaF is positioned downstream of and co-regulated with
csiD. Although located immediately upstream of the gabDTP-csiR operon, they are not
involved in GABA metabolism. The downstream gene csiR encodes a regulatory protein

that binds at csiDp.

 

 

 

 

 

NH2 9H0 COOH
(CH2)3 astC (QHzlz putA (CH2)2
H N-CH H N-CH ”—’ '
speC 2 . 2 . m HzN‘QH
co COOH EKG GLT COOH FAD FAD2 COOH
2“—" ornithine glutamic
semialdehyde glutamate
putrescine
NH2 pat NH per NH gabT 9H0 gabD COOH
we; (3:214 (CH2): ”—fi (CH2)3 ($332}; (3H2):
2 aKG GLT CH9 NAD NADH coon aKG GLT .. NADP NADPH OOH
speB y-amino ' GAB A succrnrc succinate
NH2 butyraldehyde semialdehyde
C=NH
NH
(CH2), 0% NH2 coorr
NH2 \SpeA SRNH astA 6518 8316 8310 astE (9H2)2
agmatine . A T T /‘\ 7T 3 HzN-CH
H N -3CH succrnleoA CoA C02 aKG GLT NAD NADH succrnate
2 ' + 2NH3 glutamate
COOH
arginine

Figure 26. The metabolic pathways for catabolism of GABA, putrescine, agmatine,
arginine and omithine in E. coli, highlighting the intermediacy of GABA.4

71

The acronym csi stands for garbon starvation induced. Although CsiD is a protein
of unknown function, its crystal structure was solved in 2002 and revealed it to be a
member of the or-ketoglutarate (aKG)-dependent dioxygenases.7‘8 The resolved crystal
structure had 2 A resolution and behaved as a homotetramer (Figure 27). The authors used
gel filtration chromatography of purified CsiD to show that it is a tetramer in the solution,
confirming that the structure represents the natural state of the protein. The quaternary
structure is formed through interactions of helices 011-3 from one promoter with 06-7 of
the adjacent subunit via hydrophobic, hydrogen bonding, and salt interactions. Eight [3
sheets ([33, BS,B6,B7,BS,BI4, [315 and 016) form a distorted jelly roll motif characteristic of
this family of enzymes, (Figure 27B).9 The Fe(II) in the active site is coordinated by
Hisl60, Hi5292 and Asp162. The metal has octahedral geometry with water molecules
occupying the remaining positions. Surprisingly, aKG did not bind to the active site when
included it in the crystallization conditions. The authors did note some electron density at
the putative active site; however, it was present at low occupancy and could not be

identified.

72

 

Figure 27. A) Quaternary structure of CsiD. B) A subunit of CsiD.l

73

EXPERIMENTAL PROCEDURES

Cloning and overexpression osziD.

One clone containing csiD was described previously,7 and a second one was
constructed here. The first, csiD], encoded CsiD with a HisG-tag followed by a SUMO-
fusion tag at the N-terminus, with the first fifteen amino acids missing from the protein
and was provided by Dr. Christopher Lima. The second construct, csiDz, encoded full-
length CsiD with an N-terminal His6-tag. Expression studies with each system are
described separately.

Plasmid pSMT3 (derived from pET28b plasmid) contained csiD] between the
BamHI and HindIII restriction sites. The pSMT3 vector was transformed into DHSa
maximum efficiency competent cells (Invitrogen), and the amplified plasmid was
transformed into E. coli C41(DE3) cells. The transformants were plated onto Luria Broth
(LB) agar containing 50 ug/ml of kanamycin. A single colony was used to inoculate 50 ml
of LB medium (Difco) containing 50 ug/ml kanamycin, and cells were grown at 37 °C for
16 h. A portion (15 m1) from the overnight growth was inoculated into 1 L of Terrific
Broth (Fisher Biotech) containing 50 ug/ml kanamycin at 25 °C and grown to an optical
density of 0.5 at 600 nm. B—D-isopropyl-thiogalactopyranoside (IPTG) was added (0.1
mM final concentration) and the growth continued at 25 °C for 16 h. Cells were harvested
by centrifugation at 7,500 x g for 10 min at 4 0C. These Optimum expression conditions,
chosen after experimentations with various growth temperatures, IPTG concentrations and
cell types, yielded high levels of soluble CsiD protein.

The second construct was created by amplifying csiD from E. coli MG1655 DNA

using primers 5’-CATAAGAGGATCGCTTCATATGAATGCACTGACCG-3’ and 5’-

74

AAGCTI‘TTACTGATGCGTCTGGTA-3’ that introduced NdeI and HindIII sites at the 5’
and 3’ end of the gene, respectively. Ligation of the endonuclease treated PCR product
into identically digested pET42b (Novagen) provided in an in-frame region encoding an
N-terminal His6 tag. The ligated plasmid, pET42b-csiD, was transformed into DH50r
competent cells, and the amplified plasmid was transformed into E. coli BL21(DE3) cells
that were plated on LB agar containing 50 ug/ml of kanamycin. A colony was used to
inoculate a culture of 50 ml LB medium containing 50 1.1ng of kanamycin and this was
grown overnight at 37 °C. An aliquot (15 ml) was used to inoculate 1 L of TB containing
50 ug/ml kanamycin, and this was grown at 37 0C until reaching an O.D.600 of 0.2. This
culture was moved to 25 °C and shaken vigorously until reaching an O.D.6oo of 0.5 at
which point the cells were induced with 0.1 mM IPTG and the grth continued overnight
at 25 °C. The cells were harvested by centrifugation at 7,500 x g for 10 min at 4 °C. These
optimized expression conditions were chosen after experimentations with various growth

temperatures, IPTG concentrations, and transformation-competent cell types.

Purification of CsiD proteins

Both CsiD forms were purified by using the same procedure. Approximately 3 g of
harvested cells were suspended in 10 ml of lysis buffer containing 20 mM imidazole, 300
mM NaCl, 50 mM NazHPO4, and 1 mM DTT, pH 8. The suspension was sonicated in a
Sonifier 450 (Branson) for 1 min at 30 W output power, and 50 % duration of the pulse
per sec. The sonication was repeated four more times with 1 min intervals in ice. The
broken cells were centrifuged at 100,000 g for 1 h at 4 °C. The soluble cell extracts (9 ml)

were loaded onto an 8 ml Ni-Sepharose 6-Fast Flow column (GE Healthcare), which had

75

been equilibrated with buffer A (20 mM imidazole, 300 mM NaCl, 50 mM NazHP04 pH
8). The column was washed with buffer B (100 mM imidazole, 300 mM NaCl, 50 mM
NazHPO4, pH 8) in order to remove any protein weakly bound to the resin. CsiD was
eluted with buffer C (500 mM imidazole, 300 mM NaCl, 50 mM NazHPO4, pH 8).
Exchange of the buffer was performed by dialysis against 25 mM Tris, 1 mM DTT, pH 8,
at 4 °C. In some cases (where specifically mentioned) the CsiD solution was incubated

with 1 mM EDTA prior to dialysis against the Tris buffer.

Protein analytical methods

The protein concentration was determined as described by Bradford, with bovine
serum albumin used as the standard.10 Sodium dodecyl sulfate polyacrylamide gel
electrophoresis (SDS-PAGE) (with stacking and running gels containing 5 % and 12 %
acrylamide) was used for measurement of protein overexpression and assessment of
protein purity.ll The standard protein mixture used for comparison included
phosphorylase b (M, 97,400), bovine serum albumin (M, 66,200), ovalbumin (M, 45,000),
carbonic anhydrase (M, 31,000), trypsin inhibitor (M, 21,500) and lysozyme (M, 14,400)
(Bio-Rad Laboratories). The native sizes of purified CsiD] and CsiDz were estimated by
gel filtration chromatography using a 10 um Protein-pak Diol(OH) column (Waters),
which had been equilibrated with 100 mM Tris, 300 mM NaCl, pH 7.5, with a flow rate
was 1 ml/min. The calibration protein mixture contained thyroglobulin (M, 670,000), 7-
globulin (M, 158,000), ovalbumin (M, 44,000), myoglobin (M, 17,000) and vitamin 13;;

(M, 1,350) (Bio-Rad), with added bovine serum albumin (M, 66,200).

76

Binding of aKG to F e(II)-CsiD monitored by absorption spectroscopy

Anaerobic stock solutions were prepared in sealed serum vials by alternative
cycles of vacuum and argon. A stock solution of ctKG (150 mM) was prepared in 25 mM
Tris, pH 8.0 buffer. Mixtures of ferrous ammonium sulfate (25 mM) and ascorbic acid (5
mM) were prepared by several rounds of degassing and argon flushing of the solids,
followed by addition of the desired volume of water. The procedure followed was the
same as that reported for TauD.‘2 CsiDz apoprotein (370 uM subunit in 25 mM Tris
buffer, pH 8.0) was placed into a 1 cm path length, 300 pl quartz cuvette fitted with a
stopper and was made anaerobic by 10 rounds of degassing and flushing with argon.
During the whole process the protein was kept in an ice bath to avoid precipitation.
Absorption spectra were taken at 25 °C on a Shimadzu UV-2401 UVNisible

spectrophotometer using the protein buffer as the baseline spectrum.

Binding of aKG to Fe(II)-CsiD monitored by electron paramagnetic resonance (EPR)
analysis

A solution of Diethylammonium (Z)-1-(N,N-diethylamino)diazen-1-ium-1,2-
diolate (DEA/NO) (Cayman Chemical) was added to 115 11M anaerobic protein (already
containing 2 mM of aKG and/or 115 uM Fe(II)) to reach 550 11M final concentration in
300 pl quartz cuvettes fitted with rubber stoppers. Electronic spectra (300-800 nm) were
recorded at 2 min intervals for ~ 40 min at 25 °C in order to demonstrate that the complex
CsiD/Fe(II)/NO or CsiD/F e(II)/01KG/NO was formed.

The anaerobic solutions were transferred to degassed EPR tubes and frozen.

Continuous wave X band EPR spectra were recorded at 4 K on a Bruker ESP300E

77

spectrometer equipped with an Oxford liquid He cryostat by using 100 kHz modulation,

1.99 mW microwave power, and 9.99 GHz microwave frequency.

Oxygen consumption

Oxygen consumption assays were carried out using a Clark-type oxygen electrode
and air-saturated buffer (50 mM imidazole pH 6.75) at 25 °C. In all cases, 2 11M YgaF was
mixed with 50 BM Fe(II), 400 uM ascorbate, 1 mM aKG, and 2 mM of the potential

substrates in a total volume of 5 m1.

Organic acids HPLC column

Reaction mixtures (containing enzyme and reactants in 1 ml of 25 mM imidazole
buffer at pH 6.75) were incubated for various times at 30 oC, and quenched by adding 5 pl
of 6 M sulfuric acid to the 300 pl aliquots. After centrifugation at 10,000 g for 5 min the
supernatant was centrifuged in a spin column (Amicon ultrafree-MC from Millipore) at
10,000 g for 1 min, and 200 111 was injected onto the organic acids HPLC column,
equilibrated with 13 mM sulfuric acid. The refractive index was monitored and the

integrated peak intensities were compared to those of authentic standards.

Nuclear magnetic resonance WMR) spectroscopy

NMR spectroscopy was used for identification of the product of aKG metabolism.
The reaction mixture (10 ml) of 2 uM CsiDz, 50 uM Fe(II), 400 uM ascorbate and 1 mM
aKG in 25 mM imidazole buffer (pH 6.75) was incubated for 30 min at 30 °C. The

reaction was quenched with 160 pl formic acid and the sample was centrifuged at 10,000

78

g for 5 min. The supernatant was centrifuged in a spin column (Amicon ultrafree-MC
from Millipore) at 10,000 g for 1 min. The solvent was evaporated under reduced pressure
(vacuum pump), the remaining solid was dissolved in deuterated water (D20) and the 300
MHz proton NMR spectrum was obtained. NMR spectra of control samples of orKG,
succinate, and a mixture of Fe(II), ascorbate and oKG in 25 mM imidazole buffer (pH

6.75), were obtained with the same way.

Gas chromatography-mass spectrometry (GC-MS)

GC-MS was used to detect differences in the metabolites of the wild type strain (E.
coli BW25113) and the csiD-deletion strain (csiD-KO). Colonies fi'om BW25311 and
csiD-KO strains were inoculated into 10 ml M9 medium (with either D-Ala as a C-source
and NH3 as the N-source, or succinate as a C-source and L-Trp as the N-source) for 24 h
at 37 °C. The reason for using these sources is because the two strains had the biggest
growth differences when grown in D-Ala and L-Trp as the C- and N-source, respectively.
The culture of the knock out strain in addition contained kanamycin (50 ug/ml). Portions
(0.5 ml) from the starter cultures were inoculated into 50 ml M9 medium (lacking
kanamycin to avoid metabolite differences due to the antibiotic) and grown at 37 °C until
they reached the stationary state. The cultures were mixed with equal volume of a 32.5 %
MeOH in M9 salts solution (pre-chilled to —25 °C), and centrifuged for 7 min at 5800 g.
The supematants (Angsm and Amp-04(0) were separated from the pellets (Ban/253” and
Beg-0.1m) and frozen at —80 °C. The pellets were transferred into a pro-chilled with liquid
nitrogen mortar, and homogenized with a pestle. The resulting white powder of broken

cells was dissolved in 1.4 ml of methanol-chloroforrn-water solution in a ratio of 10:32].

79

An internal standard of ribitol (50 pl of 2 mg/ml) was added to the solutions which were
incubated at 25 °C for 15 min. Water (1.4 ml) was added, and after vigorous shaking the
samples were centrifugated at 2300 g for 5 min. Aliquots of 1 ml of the aqueous phase
were pipetted into glass vials (Aligent) and dried in a speed vacuum overnight. The
samples were derivatized by mixing a portion (80 pl) with 80 pl of 20 mg/ml
methoxyamine hydrochloride dissolved in pyridine solution and incubating at 30 °C for 90
min. Then, 80 p1 of N-methyl-N-(trimethylsilyl)trifluoroacetamide (MSTFA) (Aldrich)
was mixed with the samples and incubated at 37 °C for 30 min. The GC-MS spectra were

obtained by the MS facility at MSU.

Pull down assays to test for interactions between CsiD and YgaF

A pull down assay was performed using purified CsiD; (containing a Hisé-tag) and
cell extracts of untagged YgaF. Purified CsiDz, as eluted from the Ni-NTA column was
desalted using an Econo-pac 10 DG disposable chromatography column (BioRad) to
change the buffer to 20 mM imidazole and either a) 300 mM NaCl, b) 100 mM NaCl, or
c) 0 mM NaCl. CsiD; ( 3.2 mg in 2 ml) was incubated with cell extracts of untagged YgaF
(0.2 ml) and 0.8 ml of buffer for 16 h at 4 °C. Subsequently, the mixture was loaded onto a

Ni-Sepharose 6-fast flow column and the procedure for CsiD purification was followed.

Biolog plate growth assays
Four Biolog phenotype bioarray plates (Biolog) of 96 wells each with different
carbon, nitrogen, phosphorous and sulfur sources were used to compare the growths of the

E. coli BW25113 and csiD-K0 strains. Plates denoted PM-l and PM-2 provide various

80

carbon sources, PM-3 included selected nitrogen compounds and PM-4 has an array of
phosphorous and sulfur sources. The two strains were streaked out on BUG+G plates
(included in the kit) and incubated at 37 OC. Colonies were suspended in 40 ml inoculation
fluid, and their turbidities were adjusted (by dilution or by suspension of more colonies) to
match the turbidity standard as measured by the absorbance at 600 nm. Suspensions from
both strains were provided succinate as the carbon source for the N, S, and P assays. Cells
(100 pl) were added to each well and the plates were incubated at 37 °C for 30 h (PM-1
and PM-2) or 48 h (PM3 and PM4). The growth was monitored by the development of a

purple color associated with tetrazolium dye reduction by the respiring bacteria.

Amino acid analysis

E. coli BW25113 and csiD-K0 strains were grown in 3 ml LB that contained 50 pg/ml
kanamycin for 16 h at 37 °C. The cultures (0.5 m1) inoculated into 50 ml of minimal
medium (8.8 ml 5X M9 salt, 8.8 ml of 1 M MgSO4, 4.4 pl of 1 M CaClz, 1 ml 50X
succinate ferric citrate and 40 ml water) and incubated at 37 0C to stationary phase. Cells
were harvested by centrifugation at 7,500 g for 10 min, re-suspended in 2 ml of 20 mM
Tris-HCI, pH 8 buffer, and sonicated. Soluble cell extracts were collected by
ultracentrifugation at 100,000 g for 1 h, and filtered using a Millipore centrifuge filters
(cat. Number 4104), at 2000 g for 15 min. The filtered samples of the BW25113 and
CsiD-KO strains were subjected to amino acid analysis by the Macromolecular Structure
Facility of MSU. The amino acids and other compounds that can be detected by this

method are listed in Table 2.

81

Table 2. List of amino acids and other compounds detected by amino acid analysis.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 . Phosphoserine 24. Leucine

2. Taurine 25. Tyrosine

3. Phosphoethanolamine 26. Cystathionine

4. Aspartic acid 27. Phenylalanine

5. Methionine sulfoxide 28. B-Alanine

6. Threonine 29. B-Amino-isobutyric acid
7. Serine 30. y-Amino butyric acid
8. Asparagine 3 1 . Homocysteine

9. Glutamic acid 32. Ethanolamine

10. Glutarnine 33. Tryptophan

l 1. Sarcosine 34. Ammonia

12. a-Aminoadipic acid 35. S-Hydroxylysine

1 3. Glycine 36. Aminoethylcysteine
l4. Alanine 37. Omithine

15. Citrulline 38. S-Adenosyl Homocysteine
16. a-Amino-n-butyric acid 39. Lysine

1 7. Valine 40. 1 -Methyl histidine

1 8. Methionine 41 . Histidine

19. Allo-isoleucine 42. 3-Methyl histidine
20. Hydroxy-proline 43. Anserine

21 . Carnosine 44. Cysteine

 

 

 

 

82

 

 

22. Arginine 45. Isoleucine

 

 

23. Proline

 

 

 

 

Table 2 (continued).

Cloning of YgaF

The ygaF gene was amplified by PCR from E. coli M01655 using Pfu polymerase
and primers 5’-CAAAGGAATTGAGCATATGTATGATTTTG-3 ’ and 5’-
GCTACATCCTGTTTTCAAAAGCTTTTATTGATTAAATGCGGC GTG-3 ’ , which
introduced NdeI and HindIII cutting sites into the 5’ and the 3’ gene ends, respectively.
Ligation of the 1,269 base pair PCR product into endonuclease-treated pET42b plasmid
allowed production of a non-tagged version of the YgaF protein. The pET-42b-ygaF
plasmid was transformed into E. coli DHSa (maximum efficiency) competent cells, and
the amplified plasmid was transformed into BL21(DE3) that was plated onto LB agar
containing 50 pg/ml of kanamycin. A single colony was used to inoculate 50 ml of LB
medium containing 50 pg/ml kanamycin, and the culture was grown overnight at 37 °C. A
portion (15 ml) from the overnight growth was used to inoculate 1 L of TB containing 50
pg/ml kanamycin, and this culture was incubated at 25 °C for 24 h. The cells were
harvested by centrifugation at 7,500 X g for 10 min, and 6 g of cells were resuspended in
lysis buffer (30 mM tricine, 5 mM EDTA, 20 % glycerol, 50 pM FAD, 0.5 mM PMSF,
pH 8.0). Cells were broken by sonication (Branson sonifier) and centrifuged at 27,000 X g

for 1 hat4 °C.

83

 

RESULTS
Purification of CsiD; and CsiD;

The form of CsiD whose crystal structure is known,6 here indicated as CsiD], has
an N-terrninal Hiss-SUMO fusion tag and the first 15 amino acids of the protein sequence
are missing (in order to enhance expression and solubility). The CsiD form generated by
an alternate approach, the CsiDz protein, includes the complete annotated coding region
(amino acids 1-325) and an N-terminal H156 tag. Both proteins were purified to
homogeneity from soluble cell extracts by Ni-NTA affinity column chromatography
(Figure 28). The proteins migrated as expected from the gene sequences (49.17 kDa and
39.52 kDa, respectively). DTT (1 mM) was included in all purification steps, as well as
during storage, because it enhanced protein stability. The usual yield from 1 L culture was

30-40 mg for CsiD, and 200-250 mg for CsiDz.

A) B) C)

 

234567

   

Figure 28. SDS-PAGE analysis of the expression and purification of CsiD, and CsiDz.
Denatured samples were analyzed on a 12% acrylamide gel. A) Expression of CsiD]: lane

84

1, cell extracts; lane 2, pellet; and lane 3, standard. B) Purification of CsiDI: lane 1, cell
extracts; lane 2, flow through of the Ni-Sepharose 6 fast flow column; lane 3, wash of the
Ni resin with buffer containing 100 mM imidazole; lane 4, CsiDt elution fiom Ni resin
with buffer containing 500 mM imidazole; lane 5, dialyzed CsiD, against 25 mM Tris, 1
mM DTT at pH 8; lane 6, concentrated CsiDr; and lane 7, standard. C) Expression and
purification of CsiD2: lane 1, pellet; lane 2, standard; lane 3, cell extracts; lane 4, flow
through of the Ni-Sepharose 6 fast flow column; lane 5, wash of the Ni resin with 100 mM
imid; lane 6, CsiD: elution from Ni column with 500 mM imidazole; and lane 7, CsiD:
after dialysis against 25 mM Tris, 1 mM DTT at pH 8.

Native size of CsiD.

An estimation of the native size of both purified CsiD proteins was provided by gel
filtration chromatography. When compared to results derived for standard proteins (Figure
29), the elution volume of CsiDr and CsiD: (6.98 ml and 7.36 ml, respectively) indicated a
tetrameric organization for both proteins (3 .9-mer and 4.6-mer, respectively). These

results were consistent with the X-ray analysis reported in the literature.7

3)
-l
74

5-1

4
55‘4-
3 3.

1
2-
‘

 

 

I ' I ' 1

3 ‘ é ' 9 12 15
elution volume (ml)

Figure 29. Determination of the native size of the CsiDr (red) and CsiD; (blue). The
standard curve is shown in black.

85

Binding of aKG to F e(II)-CsiD monitored by absorption and EPR spectroscopy.

A diagnostic property of the non-heme Fe(II)/aKG-dependent dioxygenase family
is the formation of a chromophore observed when anaerobic apoprotein is incubated with
Fe(II) and with the cosubstrate, 01KG.l4 This chromophore has been attributed to metal-to-
ligand charge-transfer (MLCT) transitions. '3 CsiD; was treated with EDTA to remove any
Fe(III) or Ni(II) that might be bound to the purified protein. After dialysis against EDTA-
free, 25 mM Tris, 1 mM DTT buffer, the absorption spectra of anaerobic CsiD; was
obtained. An Fe(II)/ascorbate mixture was added, followed by addition of aKG to yield
the spectra of Figure 30A. As shown by the difference spectrum in Figure 30B, the
chromophore has Amax at 510 nm with (5.0 of 158 M'l-cm". The result is similar to other
proteins of this family like TauD (Am, at 530 nm and £530 in the range of 140-250 M" -cm'
1),”l5 and XanA (Amax at 506 nm and £506 of 145 M‘l-cm") (Meng Li, unpublished
observations).

The EPR spectra of the CsiD/Fe(II)/N0 and CsiD/Fe(II)/01KG/NO complexes were
obtained in order to monitor the changes of the spectra upon aKG binding (Figure 31). N0
is an oxygen analogue that converts the nonparamagnetic high spin Fe(II) (S=2) into a
paramagnetically detectable species, {FeNO}7 (S=3/2).l6 The EPR samples were prepared
in anaerobic cuvettes (as described in the experimental methods section) and used
DEA/N0 to provide N0. N0 binding to the Fe(II) complexes was confirmed
spectroscopically by the absorbance increase at 443 nm (data not shown)” The

CsiD/Fe(II)/NO EPR sample had an axial feature at 3.97 g at 1722 G, but when aKG was

added the shape of the EPR signal in region near g i became more rhombic as indicated by

86

the shoulder at 1682 G. This provides added evidence for the interaction of the cosubstrate

with the Fe at the active site of the enzyme.

 

 

 

 

1500.
A 10004
‘7
E
0
t
‘r'
a
n, 500-
0 1 I 1 f 1
400 500 600 700
wavelength (nm)
160-
A 120‘
‘7
E
o i
I
‘7 30..
2
v
8
4o-
0 f I f F ' l
400 500 600 700
wavelength (nm)

Figure 30. Absorption spectra of CsiDz. (A) Spectra of anaerobic CsiD/Fe(II)/ascorbate
(black) and CsiD/Fe(m/ascorbate/GKG (red). CsiD; (370 pM of subunit) was mixed with
ferrous ammonium sulfate/ascorbate mixture (740 pM/l48 pM) and 1.5 mM otKG in 25

87

mM Tris buffer, pH 8.0, containing 1 mM DTT. (B) Difference spectra shown for the
CsiD/Fe(m/ascorbate/aKG minus CsiD/Fe(m/ascorbate.

 

 

 

 

8000
6000
4000
g 2000
g 01": 2a.: ‘ -4. 7 W+__.L
-2000 f
.4000
1400 T600 1m

 

G

Figure 31. EPR spectra of CsiD;/Fe(ll)/NO (black) and Cami/Fe(m/aKG/NO (red).
Protein 115 pM was adjusted to contain 110 pM ferrous ammonium sulfate and examined

in the absence (black) or in the presence (red) of 2 mM orKG. In both samples 550 pM
DEA/N0 was added ~ 40 nrin prior to freezing. The buffer was 25 mM Tris, pH 8.

Identification of an uncoupled reaction of CsiD.

Several F e(II)/o.KG dioxygenases are known to catalyze uncoupled reactions in
vitro in which dKG and 02 are consumed in the absence of any transformation of the
primary substrate. Such aberrant chemistry often is stimulated by the presence of
inhibitors or poor substrates.18 These reactions typically lead to enzyme inactivation and in
some cases result in self-modification (e. g. deA, AlkB, 'I‘auD).19’2°'21 The ability of CsiD
to catalyze an uncoupled reaction was examined by oxygen and aKG consumption and

succinate production.

88

CsiD; and CsiDz exhibited similar oxygen consumption behavior as assessed by
using an oxygen electrode (Figure 32). The initial rate of oxygen consumption by purified
CsiD (i.e., that occurred in the first 30 sec) is 700-900 (nmoles/min)/mg protein,
depending on the purification batch and on the number of days CsiD was stored at 4 °C.
The reduction in rate at longer times was shown to be due to depletion of ascorbate during

the experiment.

240-

220-

 

 

E

v 200 -
c
o
m

5‘ 1
O

180 4

«1

160 -l

I ' I ' I * I
200 250 300 350
time (sec)

Figure 32. Oxygen consumption of the uncoupled reaction of CsiD. aKG (1 mM) was
added to buffer (25 mM imidazole, pH 6.75) containing ascorbate (200 pM) and Fe(II)
(100 pM) at 170 sec, and the reaction was initiated with addition of 2 pM CsiD at 208 sec.

To determine whether the oxygen concentration was tied to aKG consumption and
succinate production, the concentrations of these compounds were monitored using the
organic acids HPLC column. A reaction mixture of CsiDr, Fe(II), ascorbate, and aKG was
incubated at 30 °C and aliquots were taken at several time points (Figure 33). The initial

rates of aKG consumption (738 nmol/min/mg protein) and succinate production (612

89

nmol/min/mg protein) agree very well with the rate of 02 consumption. The aKG
consumption and succinate production curves mirror each other, as shown in Figure 33.

The presence of ascorbate was shown to be required for oxygen consrunption, aKG
consumption, and succinate production (data not shown). The uncoupled oxidation of
aKG has been linked to a requirement of ascorbate in several enzymes (e.g. prolyl-4-
hydroxylase).22 This dependence has been attributed to the enzyme reacting with aKG to
form the iron-0x0 state which decays to the inactive Fe(III) state. Ascorbate is suggested
to reduce the metal to restore the active Fe(II) species, capable of further uncouple
chemistry. Consistent with this hypothesis, ascorbate is consumed stoichiometrically in
these uncoupled reactions.

To further confirm the production of succinate, the reaction products were
examined by proton NMR spectroscopy (Figure 34). In a control experiment Fe(II) (50
pM), ascorbate (400 pM) and aKG (1 mM) were mixed in 25 mM imidazole buffer (pH
6.75), without CsiD present, and incubated at 30 °C for 30 min. Formic acid was added
and water was evaporated under reduced pressure. The NMR spectrum of this mixture
included the resonances expected of aKG at 2.27 and 2.80 ppm, ascorbate at 3.62 and
3.90 ppm, and imidazole at 7.26 and 8.36 ppm in Figure 34. The same mixture in the
presence of CsiD2 (2 pM) was subjected to the same conditions. A proton NMR spectrum
of the crude reaction mixture in D20 for the uncoupled reaction showed the presence of

features derived from succinate at 2.30 ppm as the only species produced.

90

2500000 ]

I
2000000 -1

.1
I

1500000 -l
1000000 -«

500000 —

area of the aKG peak (uV‘sec)

 

 

1000000 -
aooooo -
600000 -l
400000 4

200000 -

.1.

 

area of succinate peak (uV‘sec)

I ' I ' I ' I

' 1
40 60 80 100 120

 

20
time (min)
I
I
I
I
r 1 r ' T r r f r T r fl
0 20 40 60 80 100 120
time (min)

Figure 33. The aKG consumption (A) and succinate production (B) curves for the
uncoupled reaction of CsiD as followed by HPLC. CsiD. (2.5 pM) was mixed with

’ Fe(II) (50 pM), ascorbate (0.4 mM) and aKG (1 mM) in 25 mM imidazole buffer at pH
6.75, and incubated at 30 °C. Aliquots of 300 pl were collected from the reaction mixture
at 5, 20, 40, 60 and 120 min. A control experiment lacking CsiD, was incubated for 120
min, and interpreted as the reaction at 0 min. The separation and detection of orKG and
succinate was done by HPLC equipped with a refractive index detector.

91

Figure 34. Proton NMR spectra of the uncoupled reaction and a blank experiment.
A) Blank experiment: ascorbate mixed with aKG in imidazole buffer in the absence of
CsiD. B) Reaction: ascorbate mixed with aKG in imidazole buffer in the presence of
CsiD. C) Blow up of the interesting area of spectrum B.

 

 

 

 

A) l .. ‘
i l
. . ‘
.-
3 .. ‘._ _-_. _, 0.)”
.' T _ 7‘ 0.317
t
a! —-1 I
l
i l
\‘\
g _:i ;:L_-.”-:.: r-» 1.2”
:3 a i ”if
, l
‘ l
. l
- 4
l
at
1
U! 5 \
.4 4"" 4‘: — — -_ —‘ WW-—
(L.
f...
l
. e Y
1.1
u .
o ."
L. 4 b
u l
A... l
H l
«'o-i > Er“ -—-——-—---a.ns
o ‘ [
.l
2‘ r L _”4r~‘3.392
3 L 1 '1“ "“M-aau
” I
.1
.1
P _J
1
1i
'6 l
'8 ..l

 

92

,i- .¥_._.kn_—-._._-.—.. ..,. _

Figure 34 (continued).

93

«wu-

\.

-»_J_-3°3°‘

 

.—
. .
a
3 «a 1 ,
s 2 1 7
. C C n. a
1. . L J W .
. 4 w 3 C 5 n
I. u . o I.-
..n.,_. \_ 3 C
. \.. 7 o
H. I.
1 . C \u v
o n 2‘
, . . . n
o .
a a
. x .
. m
_ .
,3
,. c . ,
s. . ,
.3 A
. M 7.
a, i
I
(
. x/ ‘ e .. . ,u 7 r.
9). i r l
W. .. \(l .I \l I) ‘4 ‘1. r . . If . / \1
tr) ( ,)|lt t/l\r\(1i!4|.\l (l( v f . t). lift-I) llltlll, {((l \l‘rLl/Ilillttlll l lIllr ‘10.le link. .llllt'l‘ll'lltlllilllllllg-ill. g It’ll) lll\ll'
« _ t I a l I J l l 4' a.“ 4 l litid, 1,...1IWI.4I «I'll l J1 g‘ d I? Jlll lllllfl.‘

..o u.¢ w.a u... w.» u.o ”.0 u.m N.‘ n.» 0‘.

 

H.uu ”.00

Figure 34 (continued).

C)

94

Biolog plates analysis.

A method for studying differences in metabolism between two bacterial strains is
Phenotype Microarrays (PM) analysis.23 Many phenotypes can be tested simultaneously in
a very simple, efficient, standardized format. PMs are sets of 96-well microtiter plates.
Each well contains a different dried cell culture medium that is designed to test a unique
phenotype or cell function. A colorimetric method is used to detect respiring bacteria. If
the cells can grow on the specific culture medium they respire normally to give a dark
purple color. If the growth is weak, the respiration is slow and only light purple color is
observed. If the phenotype and growth are negative, the well remains colorless. The plates
provide different carbon (PM-l and PM-Z plates), nitrogen (PM-3 plate), and phosphorous
and sulfur sources (both in PM-4 plate).

The two strains that we compared with this method were BW25113 wild-type cells
and the isogenic csiD-KO mutant. The knock out mutant was generated by insertion of a
kanamycin resistance gene into the reading frames of the csiD gene. The PMs were
inoculated with each strain and incubated at 37 °C for 30 h (C-sources) and 48 h (all
others), respectively. For the first 30 h, pictures of the plates were taken every 6 h and
then, for the N, S and P plate a last picture at 48 h. In several cases the CsiD-KO strain
grew slower but finally reached the same intensity as BW25113. The most interesting
differences between the strains involved the following C-sources: D-alanine, acetic acid,
L-glutamine, m-tartaric acid, B-methyl-D-glucoside, N-acetyl-B-D-mannosamine and
glucuronamide (Figure 35A). N-sources of particular interest were: L-arginine, L-aspartic
acid, L-tryptophan, D-valine, L-homoserine, agmatine, adenine, D,L-a-amino-N-butyric

acid, D,L-a-amino-valeric acid, 6-amino-N-valeric acid and alanine-leucine (Figure 35B).

95

For the S-sources conclusions couldn’t made because growth was observed even in the
negative control. All the interesting molecules were tested as substrates of CsiD using

oxygen electrode (vide infra).

 

 

 

 

Wl' CsiD-KO
, _ . , p \ .. D-Alanine
{f 9 Q; . . . . ‘3' . ".l\. g. g‘ ‘ ‘. ‘l
toooooooowe .:~ ,
C-sourse o; o o o ,. :- o o o ,o ::- o " ' Amt'c acud
0‘ v : 2 O O I >' 3 i‘ :0 ‘ W ' ?. L-Glutamme
~ , f l' r “ -
9, ’ ' ' " ° 3’ ,9 ' f ;. B—methyl-D-gluoosude
°__ :: f : . ,z .. ,: _ .- N-acetyI-B-D-mannosamine
:. . ‘ . .,../. .. .-, ' ' glucuronamide
’ L-Arginine
, _ EfL-Aspartic acid
‘ 0 00.02.09! ‘45" -. .
.1. - . 7. .. ‘3 a; (g 0 O ' ‘ ' O V E; O K. c- . B‘LWFtOpha"
' ’ .: .; .' ‘0 e »' . ' ' . - ame .
N-sourse - 1 .. ' .. z : . \ ’ : _ _[:l—.—l;-EI:L-homosenne
. j j .. . - m ..... . ,1-“ v. ,Qgg‘naigge
' 'f 9 * t " ' ‘ “'Lil' ‘ ' ' ‘* ' >D.L—a-amino-N-butyric acid

 

..2‘.......o;..’o:- 0.3;. cove. _ _
o‘o‘b‘oo'oo‘o’o’o’r'o 6660‘ n 'o'o’r 0' 93m'"°'N“'a'°"°

. Q-amino-N-valeric acid
alanme-leucme

Figure 35. Photos of Biolog plates. Plate with C-sources (PM-1) of BW25113 and CsiD-
KO of 30 h of incubation at 37 °C. Plate with N—sources (PM-3) of BW251 13 and CsiD-
KO of 48 h of incubation at 37 °C.

GC-MS analysis.

A second method comparing the metabolites of two bacterial strains involves GC-
MS analysis. BW25113 and csiD-KO strains were grown identically up to the stationary
phase and then the metabolites were extracted, derivatized and finally measured using a
GC-MS instrument. Initially, the growths were centrifuged and both supematants and cells
were saved. The supernatant solutions have all the compounds that were exported from the

cells (Amt/253” and AcsiD-KO), whereas the pellets have all the molecules that remain within

96

(BM/25311 and Bea-04(0). From the analysis having D-Ala as the C-source the samples
B3w253n and 8,3li0 exhibited differences in two unidentified molecules with molecular
weights after derivatization of 373 and 290 respectively and almost identical mass spectra;
whereas the interesting compounds from the analysis of the A3w253” and Amp-K0 samples
were 3-carboxy-6-ethoxy-4-hydroxy-1,S-naphthyridine, N-methylaminopropionic acid,
hexanoic acid, aspartic acid and acetic acid. For the cells that grew in L-Trp as the only
N-source the samples Ban/253” and Begum exhibited differences in (l-ethyl-2-
methylpropyl)methylamine and 3,5-dichloro-4-methoxy-2,6—dimethyl-pyridine; whereas
Aszsm and Away) showed differences in 4,6-dimethyl-1-oxa-4,6-diazacyclooctane-5-
thione, benzenebutanoic acid, silanol, pyrroindole, glycine, dihydrocodeine, D-threo-2,5-
hexodiulose, malic acid, 3-methylvaleric acid, pentanedioic acid, erythrose and

benzenepropanoic acid.

Amino acid analysis.

Another attempt to identify the biological role of CsiD involved comparison of the
cellular amino acids produced by the BW25113 and csiD-KO strains (Figure 36). Several
amino acids were observed to differ in concentration in the two strains, with the biggest
difference noted for glutamic acid where the CsiD-KO strain possessed half the level of
glutamic acid compared to the BW25113 strain (Figure 36). This result could imply that
CsiD is involved in a pathway that enhances the production of this particular amino acid
or that it reduces its rate of consumption. The two strains exhibit differences in other
amino acids too (e.g. Ala, Gly, a-aminoadipic acid), although those changes were

considerably smaller, and within the range of experimental error.

97

0.50 -

  
   

 

  

 

a: 0.45;
37; 30.40;
5.; £035?
£54 gt:
E". 3 0.203
3 3‘, gons‘
gzj 20.10
31 3 0.05

0 0.00

gaggaéssssaéessgsiga észgggfiggsgagsisisas

Figure 36. Comparison of the amino acid analysis of the BW25113 (black) and csiD-
KO (red) strains. Acronyms except of natural amino acids that have been used are: Pser
(phosphoserine), Tau (taurine), PEA (phosphoethanolamine), MetSO4 (methione
sulfoxide), Sar (sarcosine), a-AAA (a-aminoadipic acid), Cit (citruline), a-ABA (a-amino-
n-butyric acid), Allo-Ile (allo-isoleucine), Cysthi (cystathionine), b-Aiba (B-amino-
isobutyric acid), Hcys (homocysteine), EOHNH2 (ethanolamine), Hylys (6-
hydroxylysine), AEC (aminoethylcysteine), Om (omithine), S-AHCys (S-adenosyl
homocysteine), 1-MeHis(1-methyl histidine) and 3-MeHis (3-methyl histidine).

Investigation of potential substrates using an oxygen electrode.

A search for the primary substrate of CsiD was performed by using an oxygen
electrode. Compounds tested included those proposed to be involved in GABA
metabolism (Figure 26, putrescine, GABA, agmatine, L-arginine, L-glutamate) along with
all other native amino acids. All tested substrates consumed the same amount of oxygen as
observed for the uncoupled reaction except L-tryptophan, which acted as an inhibitor by
diminishing the uncoupled reaction after its addition to the CsiD mixture. Other molecules
(chosen on the basis of their reactivity of the Biolog studies, their use as substrates by
other distantly related enzymes, or to test specific hypotheses of CsiD function) were

tested by using the same assay and included: D-alanine, D-alanine-D-alanine, D-valine, L-

98

homoserine, L-phenyl propionic acid, thymine, adenine, ribose, sucrose, glucose, 5-
aminovaleric acid, y-butyrobetaine, 5-amino-N-valeric acid, D,L-a-amino-N-butyric acid,
glucuronamide, m-tartaric acid, N—acetyl-B-D-mannosamine, D,L-2-aminobutyric acid and

kynurenine. None of these substrates stimulated oxygen consumption.

Pull down assays.

In the absence of any evidence to directly identify the primary substrate of CsiD, I
tested whether this protein might form a complex with the product of the downstream
gene ygaF. As discussed in Chapter 4, YgaF, is an FMN-containing L-2-hydroxyglutarate
oxidase. To test for interaction between CsiD and YgaF, purified CsiD; was mixed with
cell extracts containing YgaF, incubated overnight at 4 oC, and loaded onto a Ni-NTA
column. If the two enzymes interacted strongly, both proteins would bind to the resin
because of the Hisb-tag on CsiD. In the absence of interaction, YgaF wouldn’t bind to the
affinity column. The experiment was carried out using three different salt concentrations
in the buffer: 20 mM imidazole containing 300, 100 and 0 mM NaCl. In no case did the
enzyme co-elute from the column; (Figure 37) rather, YgaF consistently failed to bind to
the column (eluting in the flow through fraction). In contrast, CsiD eluted with the 500
mM imidazole buffer in the first experiments, and in the wash buffer in the third
experiment. A blank experiment using only the cell extracts of YgaF proved that this

protein doesn’t bind to the column, as expected.

99

1234567891011 12131415161718 1920212223242526

     

hen

 

Figure 37. Pull down assays with Hing-tagged CsiD; mixed with cell extracts of non-
tagged YgaF. Blank experiment lanes 1-7: purified CsiDz. 1; standard, 2; cell extracts of
non-tagged YgaF, 3; cell extracts of YgaF with buffer instead of CsiD, 4; flow through, 5;
wash, 6; elution, 7. Pull down assay with 300 mM NaCl: cell extracts of YgaF with CsiD,
8; flow through, 9; wash, 10; elution, 11. Pull down assay with 100 mM NaCl: purified
CsiDz, 12; cell extracts of YgaF, 13; cell extracts of YgaF with CsiD, 14; flow through,
15; wash, 16; elution, 17; standard, 18. Pull down assay with 0 mM NaCl: purified CsiDz,
19; cell extracts of YgaF, 20; standard, 21; cell extract of YgaF with CsiD, 22; flow
through, 23; wash, 24; elution with buffer that doesn’t contain salt, 25; elution with buffer
that contains 300 mM NaCl, 26. Wash buffer: 100 mM imidazole with A) 300 mM, B)
100 mM and C) 0 mM NaCl at pH 8. Elution buffer: 500 mM imidazole with A) 300 mM,
B) 100 mM and C) 0 mM NaCl at pH 8.

100

CONCLUSIONS

The E. coli csiD gene is located in the csiD-ygaF-gabDTP-csiR operon, controlled
by the csiDp promoter, activated exclusively under carbon starvation and encodes a
protein, CsiD, that has been crystallized and shown to belong to the non-heme Fe(II)
dioxygenase family. In this study the purification and general characteristics of CsiD have
been described.

Purification and native size. Two forms of CsiD were isolated. The CsiD. protein
is the one that was crystallized and has an N-terminal Hisé-SUMO-tag, but is missing the
first fifteen amino acids. The CsiD; protein is the full-length protein with an N-terminal
Hisb-tag. Both were purified very effectively in one step by using affinity column
chromatography. Their stabilities were enhanced in the presence of the DTT and they were
able to be stored at 4 0C for several weeks. Estimation of the native size of the CsiD forms
by gel filtration chromatography revealed that both are tetramers, agreeing with the
crystallographic result. The presence of the tags didn’t affect the quaternary structure of
the protein because the N-terminis is not involved in the interactions between the
monomers.

CsiD-Fe(ID-aKG complex. The crystals of CsiD were grown under anaerobic
conditions in the presence of 20 mM aKG. Nevertheless, diffraction analysis didn’t
indicate any complexation of the aKG with Fe(II) at the active site. To investigate whether
aKG can bind to enzyme-bound Fe(II), two approaches, absorption spectroscopy and EPR
spectroscopy, were used. All the aKG-dependent dioxygenases develop a characteristic
chromophore in the visible area of the spectrum upon binding Fe(II) and aKG due to

metal-to-ligand charge-transfer transitions.24 The spectrum of anaerobic CsiD/Fe(II)

101

mixed with aKG developed the characteristic chromophore (hm = 510 nm, (510 = 158 M'
1cm"). The binding of aKG to the CsiD active site was also confirmed by EPR studies, in
which superimposition of the EPR spectra of CsiD/Fe(II) with and without aKG indicated
that the two spectra are not identical. The EPR spectrum of CsiD/Fe(II)/0tKG exhibits a
more rhombic signal compared to the one where aKG was absent. Both experiments prove
that aKG binds to the active site of CsiD. Concerning the failure of Chance et al. to
observe such a complex in the crystal structure of CsiD,7 I hypothesize that the complex
was destabilized by the conditions used to grow the crystals.

Uncoupled reaction. Another characteristic property of aKG-dependent
dioxygenases is the ability to oxidize the cosubstrate in the absence of the primary
substrate, a situation known as an uncoupled reaction. In air-saturated buffer, CsiD
consumed oxygen and in the presence of Fe(II), ascorbate and aKG, as expected. The
presence of ascorbate seemed to be completely necessary. This phenomenon can be
explained as an aberrant short cutting of the normal catalytic cycle of the dioxygenase. In
the absence of primary substrate, oxidative decarboxylation of aKG slowly forms an
Fe(IV)-0x0 intermediate which decays to yield an Fe(IlI) species. Ascorbate acts as a
reductant, returning the iron from the Fe(III) to its active Fe(II) state, making the protein
ready for another cycle. The consumption of aKG during the uncoupled reaction was
confirmed by using an organic acids column to separate products while monitoring the
refractive index. The appearance of succinate (the product of the oxidation of oKG)
coincided with the loss of aKG during the course of the reaction. The production of

succinate was also demonstrated by NMR.

102

Searching for the CsiD substrate. Four methods were used to investigate potential
substrates for CsiD. First, metabolic analysis used GC-MS methods and revealed several
compounds that differed in abundance in BW25113 and csiD-KO growths both outside
and inside the cells. Further studies on the molecules that looked promising showed that
none of them served as a substrate for CsiD. Biolog plate analysis was another method
that gave us the opportunity to compare the growths of E. coli BW25113 and CsiD-KO
having many different carbon, nitrogen, phosphorous and sulfur sources. Several subtle
changes were noted, but follow up studies failed to identify any primary substrate from
this approach. It is possible that E. coli possesses more than one pathway to metabolize the
specific compound of interest. If this is the case for CsiD, then the csiD-KO strain would
grow in exactly the same way, or maybe a little bit slower than the WT strain. A third
approach to test for the primary substrate of CsiD involved direct assays for oxygen
consumption. Again, the approach was fruitless as none of the compounds served as a
substrate of CsiD. A forth method examined the role of CsiD by amino acid analysis,
where we specifically investigated differences in the amino acids levels of BW25113 and
csiD-KO cells. The most pronounced difference between the two strains was found in the
amount of L-glutamic acid, but it is unclear how this result relates to the presence or
absence of CsiD.

A very recent review by Shelley D. Copley discusses reasons to explain why some
metabolic enzymes in E. coli are essential for growth on glucose as a sole carbon source,
whereas others are not. 25 Surprisingly, 80 out of 227 genes encoding enzymes in central
metabolic pathways are not essential for growth on glucose. This discrepancy can be

explained by the existence of: i) alternative pathways, ii) the availability of isozymes, iii)

103

enzymes with broad-specificity, or iv) redundant multifunctional enzymes. In some cases,
interconnecting metabolic pathways allow the organism to bypass a defective metabolic
step, however it is possible the cell will not survive due to the toxicity which can be
caused by the accumulation of the substrate of the missing enzyme. With regard to the
Krebs cycle, often considered an essential series of reactions for obtaining energy during
growth on glucose several steps are in fact nonessential because the intermediates can be
produced by alternative pathways. Many nonessential metabolic genes encode enzymes
for which isozymes are encoded by other genes and catalyze the same reaction using
different reactants. Enzymes with broad substrate specificities often can compensate for a
missing enzyme. For example, the transaminase class enzymes convert a wide range of 2-
keto acids to amino acids. There are also occurrences where enzymes have redundant
catalytic activities. A representative example is the conversion of shikimate to shikimate
3-phosphate from both shikimate kinases AroK and AroL, with differing Km for shikimate,
equal to 200 and 20 uM respectively. The fact that a significant number of genes (~ 35 %)
encoding enzymes involved in metabolism of E. coli is not essential for the growth of the
cell on glucose, raises the possibility that csiD might be included in this class.

Pull-down assay. A final effort to investigate the role of CsiD explored whether it
formed a complex with the product of ygaF , a gene that is adjacent to csiD and encodes a
protein not previously characterized. The possibility of interaction between CsiD and
YgaF was tested by pull-down experiments using Hiss-tagged CsiD and cell extracts
containing overexpressed YgaF. No interactions between these proteins were detected.

In summary, I demonstrated that aKG binds to the active site of CsiD, something

that was not obvious in the crystal structure, and conclude that the protein is an (1K0-

104

dependent dioxygenase. Moreover, the uncoupled reaction, an in vitro reaction performed
by several members from this family, was observed by several methods and provides
further evidence for this classification. I tried to identify the biological role of CsiD by
several approaches, but was not successful. Finally, I used pull-down assays with cell
extracts containing overexpressed YgaF to show that no interactions are formed between

these proteins.

105

REFERENCES

l Niegemann, E., Schulz, A., and Bartsch, K. (1993) Molecular organization of the
Escherichia coli gab cluster: nucleotide sequence of the structural genes gabD and gabP
and expression of the GABA permease gene. Arch. Microbiol. 160, 454—460.

2 Shaibe, E., Metzer, E., and Halpem, Y. S. (1985) Metabolic pathway for the utilization
of L-arginine, L-ornithine, agmatine, and putrescine as nitrogen sources in Escherichia
coli K-12. JBacteriol. 163, 933-937.

3 Schneider, 13., L., Kiupakis, A., K., Reitzer, L., J. (1998) Arginine catabolism and the
arginine succinyltransferase pathway in Escherichia coli. J. Bacteriol. 180, 4278-4286.

4 Schneider, B. L., Ruback, S., Kiupakis, A. K., Kasbarian, H., Pybus, C., Reitzer, L.
(2002) The Escherichia coli gabDTPC operon: specific y-aminobutyrate catabolism and
nonspecific induction. J. Bacteriol. 184, 6976-6986.

5 Metzner, M., Germer, J ., and Hengge, R. (2004) Multiple stress signal integration in the
regulation of the complex (rs-dependent csiD-ygaF-gabDTP operon in Escherichia coli.
Molec. Microbiol. 51, 799-81 1.

6 Marschall, C., Labrousse, V., Kreimer, M., Weichart, D., Kolb, A., and Hengge-Aronis,
R. (1998) Molecular analysis of the regulation of csiD, a carbon starvation-inducible gene

in Escherichia coli that is exclusively dependent on as and requires activation by cAMP-
CRP. J. Molec. Biol. 276, 339—353.

7 Chance, M. R., Bresnick, A. R., Burley, S. K., Jiang, J. S., Lima, C. D., Sali, A., Almo,
S. C., Bonanno, J. B., Buglino, J. A., Boulton, S., Chen, H., Eswar, N., He, G., Huang, K,
Ilyin, V., McMahan, L., Pieper, U., Ray, 8., Vidal, M., and Wang, L. K. (2002) Structural
genomics: a pipeline for providing structures for the biologist. Protein Sci. 11, 723-738.

8 Hausinger, R. P. (2004) Fe(II)la-ketoglutarate-dependent hydroxylases and related
enzymes. Crit. Rev. Biochem. Mol. Biol. 39, 21—68.

9 Schofield, C. J ., and Zhang, Z. (1999) Structural and mechanistic studies on 2-

oxoglutarate-dependent oxygenases and related enzymes. Curr. Opin. Struct. Biol.
9, 722-731.

'0 Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram

quantities of protein utilizing the principle of protein-dye binding. Anal] Biochem. 72,
248-254.

106

” Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head
of bacteriophage T4. Nature (London) 22 7, 680-685.

‘2 Kalliri, E., Grzyska, P. K., and Hausinger, R. P. (2005) Kinetic and spectroscopic
investigation of Co", Ni", and N-oxalylglycine inhibition of the Fell/a-ketoglutarate
dioxygenase, TauD. Biochem. Biophys. Res. Commun. 338, 191—197.

'3 Purpero, V., and Moran, G. R. (2007) The diverse and pervasive chemistries of the a-
keto acid dependent enzymes. J. Biol. Inorg. Chem. 12, 587-601.

'4 Ryle, M. J ., Padmakumar, R., and Hausinger, R. P. (1999) Stopped-flow kinetic analysis
of Escherichia coli taurine/a-ketoglutarate dioxygenase: interactions with o-ketoglutarate,

taurine, and oxygen. Biochemistry 38, 15278-15286.

'5 Grzyska, P. K., Miller, T. A., Campbell, M. G., and Hausinger, R. P. (2007) Metal
ligand substitution and evidence for quinone formation in taurine/a-ketoglutarate
dioxygenase. J. Inorg. Biochem. 101, 797-808.

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

'7 Enemark, J. H., and F eltharn, R. D. (1974) Principles of structure, bonding, and
reactivity for metal nitrosyl complexes. Coord. Chem. Rev. 13, 339.

‘8 Welford, R. W. D., Schlemminger, I., McNeill, L. A., Hewitson, K. S., and Schofield,
C. J. (2003) The selectivity and inhibition of AlkB. J. Biol. Chem. 278, 10157-10161.

‘9 Bradley, F. C., Lindstedt, S., Lipscomb, J. D., Que, L., Jr., Roe, A. L., and Rundgren,
M. (1986) 4-Hydroxyphenylpyruvate dioxygenase is an iron-tyrosinate protein. J. Biol.
Chem. 261, 11693-11696.

20 Liu, A., Ho, R. Y. N., Que, L., Jr., Ryle, M. J ., Phinney, B. S., and Hausinger, R. P.
(2001) Alternative reactivity of an a-ketoglutarate-dependent iron(II) oxygenase: enzyme
self-hydroxylation. J. Amer. Chem. Soc. 123, 5126-5127.

2' Ryle, M. J ., Liu, A., Muthukumaran, R. B., Ho, R. Y. N., Koehntop, K. D., McCracken,
J ., Que, L., Jr., and Hausinger, R. P. (2003) Oz— and a-ketoglutarate-dependent tyrosyl
radical formation in TauD, an a-keto acid-dependent non-heme iron dioxygenase.
Biochemistry 42, 1854-1862.

107

22 Myllyléi, R., Majamaa, K., Gt'lnzler, V., Hanauske-Abel, H. M., and Kivirikko, K. I.
(1984) Ascorbate is consumed stoichiometrically in the uncoupled reactions catalyzed by
propyl 4-hydroxylase and lysyl hydroxylase. J. Biol. Chem. 259, 5403-5405.

23 Loh, K. D., Gyaneshwar, P., Papadimitriou, E. M., Fong, R., Kim, K. S., Parales, R.,
Zhou, Z., Inwood, W., and Kustu, S. (2006) A previously undescribed pathway for
pyrimidine catabolism. Proc. Nalt. Acad. Sci. USA 103, 5114-5119.

24 Pavel, E. G., Zhou, J ., Busby, R. W., Gunsior, M., Townsend, C. A., and Solomon, E. I.
(1998) Circular dichroism and magnetic circular dichroism spectroscopic studies of the
non-heme ferrous active site in Clavaminate synthase and its interaction with a-
ketoglutarate cosubstrate. J. Amer. Chem. Soc. 120, 743-753.

25 Kim, J ., and Copley, S. D. (2007) Why metabolic enzymes are essential or nonessential
for growth of Escherichia coli K12 on glucose. Biochemistry ASAP.

108

CHAPTER 4

Cloning of the ygaF gene, and purification of the encoded protein.

109

ABSTRACT

The ygaF gene, encoding a protein of previously unknown function in Escherichia
coli, was cloned and overexpressed, and the YgaF protein was shown to possess non-
covalently bound FMN. Reaction with dithionite and photoreduction both led to the two-
electron reduction of the flavin. Reduced reacted rapidly with oxygen and the oxidized
protein was bleached by reaction with sulfite; both of these results are consistent with
YgaF being an oxidase. The wild-type strain Escherichia coli BW25113 was compared
with the ygaF-KO strain, lacking an effective ygaF, in with growth studies, Biolog plate
assays, and amino acid analyses, but these experiments did not identify the primary
substrate. Numerous potential substrates with various functional groups were tested in
several types of assays, and YgaF was shown to exhibit L-2-hydroxyglutarate oxidase
activity. The product of the oxidation of L-2-hydroxyglutarate was proven to be a-
ketoglutaric acid by HPLC and colorimetric methods. The inability of anaerobic, reduced
enzyme to reverse the reaction by reducing the product a-ketoglutaric acid is explained by
the very high reduction potential (+19 i 8 mV) of this enzyme. Also, it was shown that

the D isomer of 2-hydroxyglutarate is not a substrate of YgaF.

110

INTRODUCTION

The ygaF gene of Escherichia coli is located immediately downstream of csiD (see
Chapter 3), encoding a crystallographically-characterized protein of unknown function,1
and just upstream of the gabDTP-csiR operon,2 encoding succinic semialdehyde
dehydrogenase, y-aminobutyric acid (GABA) transaminase, and a GABA-specific
permease (Figure 25, Chapter 3). The set of five genes is co-regulated by CsiR, cAMP-
CRP, and as acting at csiDp during carbon starvation and at stationary phase.3 Expression
of gabDTP-csiR is separately controlled by 05 binding to gaprl, which triggered by
multiple stress induction, and by Nae/cs7O interaction with gaprz in response to nitrogen
starvation.2 YgaF is not obviously involved in GABA metabolism since it does not affect
the catabolism of GABA or any other nitrogen source (agmatine, arginine and ornithine),4
and its role is unknown.

On the basis of its amino acid sequence, YgaF is likely to be a flavoenzyme. It has
been estimated that 1-3 % of identified proteins in prokaryotic and eukaryotic cells contain
flavin,5 and these abundant enzymes catalyze a wide range of reactions with a diverse set
of substrates including alcohols, aldehydes, ketones, amines, dithiols, amino acids, and
hydroxy acids.6 Most of these enzymes transition between the fully oxidized and two-
electron reduced forms of their cofactor, but in some cases the one-electron reduced
semiquinone species is stabilized. Reoxidation of the reduced flavin coenzyme can take
place via several processes including by reaction with oxygen, as in the case of flavin
oxidases. The flavin generally is tightly bound to these enzymes, and in selected examples
the coenzyme is covalently attached to the protein. Sequence comparisons of YgaF reveal

this 422-amino acid E. coli protein to be homologous to human mitochondrial L-2-

111

hydroxyglutarate dehydrogenase (41% identity over 398 residues),7 Helicobacter pylori
malatezquinone oxidoreductase (24% identity over 421 residues),8 Bacillus creatinase
sarcosine oxidase (23% identity over 255 residues),9 human mitochondrial
dimethylglycine dehydrogenase (24% identity over 227 residues),'0 Bacillus subtilis
glycine oxidase (25% identity over 146 residues),ll human peroxisomal L-pipecolic acid
oxidase (21% identity over 219 residues),12 and many other flavoenzymes. This list
includes both dehydrogenases and oxidases, with some representatives having covalently

bound flavin,12"3‘ ‘4

while others do not. In this chapter I describe the cloning and
overexpression of ygaF, the purification and characterization of YgaF, and the
demonstration that it is an oxidase that possesses non-covalently bound FMN. In an effort
to define the function of YgaF, I compare wild-type and ygaF mutant (BW25113 and
ygaF-KO) strains using growth studies, Biolog plate assays and amino acid analysis.

Finally, I show that YgaF is an L-2-hydroxyglutarate oxidase and discuss on the potential

relevance of this activity to E. coli.

112

EXPERIMENTAL PROCEDURES

Materials

D-3-phosphoglycerate disodium salt, O-phospho-L-serine, O-phospho-L-
threonine, D-2-hydroxygluratate disodium salt, butyraldehyde, (:t)-citramalic acid, L-
mandelic acid, D-mandelic acid, L-malic acid, DL-malic acid, L-glutamic acid, D-
glutamic acid, L-glutamine, L-omithine-hydroxide, D-omithine-hydroxide, and L-
threonine were obtained from Sigma. Sarcosine, 2-hydroxycaproic acid, and y-
aminovaleric acid were obtained from Aldrich. DL-lactic, a-ketoglutaric acid disodium
salt, agmatine and D-valine were obtained from Fluka. y-aminobutyric acid was purchased
from General Mills Inc. L-2-hydroxyglutarate zinc salt was obtained from City Chemicals

LLC.

Cloning and expression of ygaF

The gene encoding YgaF was amplified by PCR using E. coli MG1655 DNA as
template, Pfu polymerase, and primers (5’-CAA AGG AAT TGA GCA TAT GTA TGA
TTT TG-3’ and 5’-GCT ACA TCC TGT TTT CAA AAG CTT TTG ATT AAA TGC
GGC GTG-3’) that introduce Ndel and HindIII sites into the 5’ and the 3’ ends of the
gene, respectively. Ligation of the 1,269-hp PCR product into pET42b (Novagen)
provided in an in-frame region encoding a C-terrninus His6 tag. The ligated pET42b-ygaF
plasmid was transformed into E. coli Max Efficiency DH50l competent cells (Invitrogen),
and the amplified plasmid was transformed into E. coli C41(DE3)l5 and BL21(DE3) cells.

The transformants were plated onto Luria Broth (LB) agar containing 50 ug/ml of

kanamycin. Optimized expression conditions were chosen after experimentations with

113

various growth IPTG concentrations (0, 0.1 and 1 mM), and transformation-competent
cell types (C41(DE3) and BL21(DE3)). In detail, a single colony was used to inoculate 3
ml of LB medium with 50 ug/ml kanamycin, and the culture was grown overnight at 37
°C. A portion (1.2 ml) from the overnight growth was used to inoculate 12 ml of Terrific
Broth (Fisher Biotech) containing 50 ug/ml kanamycin, and this was incubated at 30 °C to
an optical density of 0.7 at 600 nm. IPTG was added (0, 0.1 and 1 mM final concentration)
and the growth continued at 30 °C overnight. The cells were harvested by centrifugation at
10,000 X g for 10 min and resuspended in 1 m1 of lysis buffer (30 mM imidazole, 300 mM
NaCl, 50 mM NazHPO4, 20 % glycerol, 50 uM FAD, 0.5 mM PMSF, pH 7.2). Cells were
broken by sonication (Branson sonifier, 5 reputations of 1 min each at 30 W output power,
and 50 % duration) and centrifuged at 10,000 x g for 20 min at 4 °C and the expression
was examined by SDS-PAGE.16 The chosen optimized growth conditions were followed
to express YgaF in large scale. A single colony of C41(DE3) cells was used to inoculate
50 ml of LB medium with 50 ug/ml kanamycin, and the culture was grown overnight at 37
°C. A portion (15 ml) from the overnight growth was used to inoculate l L of Tenific
Broth (Fisher Biotech) containing 50 ug/ml kanamycin, and this was incubated at 30 °C to
an optical density of 0.7 at 600 nm. IPTG was added (1 mM final concentration) and the
growth continued at 30 °C overnight. The cells were harvested by centrifugation at 7,500
X g for 10 min and resuspended in 30 ml of lysis buffer (30 mM imidazole, 300 mM NaCl,
50 mM NazHPO4, 20 % glycerol, 50 uM FAD, 0.5 mM PMSF, pH 7.2). Cells were

broken by sonication and centrifuged at 100,000 X g for 1 h at 4 °C.

114

Purification of His-tagged Y gaF

The cell extracts were loaded onto a Ni-NTA-Sepharose 6-Fast Flow column (2.5
cm diameter by 2 cm, GE Healthcare), which had been equilibrated with buffer A (30 mM
imidazole, 300 mM NaCl, 50 mM NazHPO4, 20 % glycerol, pH 7.2). The column was
washed with buffer B (100 mM imidazole, 300 mM NaCl, 50 mM NazHPO4, 20 %
glycerol, pH 7.2) in order to remove any weakly bound protein from the resin, and YgaF
was eluted with buffer C (500 mM imidazole, 300 mM NaCl, 50 mM NazHPO4, 20 %
glycerol, pH 7.2). The YgaF was immediately exchanged into buffer D (25 mM HEPES,
100 mM NaCl, 5 mM EDTA, 1 mM dithiothreitol (DTT), 20 % glycerol, pH 8.2) by using
a Superdex G-25 column that had been equilibrated with this buffer. The fractions were
analyzed by (SDS)-PAGE 12% polyacrylamide gel. Molecular weight markers included
phosphorylase b (M, 97,400), bovine serum albumin (M, 66,200), ovalbumin (M, 45,000),
carbonic anhydrase (M, 31,000), trypsin inhibitor (Mr 21,500) and lysozyme (M, 14,400)

(Bio-Rad Laboratories).

Native size of Y gaF

The native size of purified YgaF was investigated by gel filtration chromatography
using a Superdex 75 (1.5 cm diameter by 66 cm GE Healthcare) and a Sephacryl 300 (1.5
cm diameter by 66 cm, GE Healthcare) column. The first column was equilibrated with 25
mM HEPES buffer containing 100 mM NaCl, at pH 8.2. The second column was
equilibrated with 25 mM HEPES buffer containing 100 mM or 300 mM NaCl, at pH 8.2.

In all cases the flow rate was 1 ml/min. The calibration protein mixture contained

115

thyroglobulin (M, 670,000), y-globulin (M, 158,000), ovalbumin (M, 44,000), myoglobin

(M, 17,000) and vitamin 13.2 (M, 1,3 50) (Bio-Rad).

Spectroscopic studies

Absorption spectra were obtained at room temperature using a Shimadzu UV-2401
UV/visible spectrophotometer. For analyses requiring the absence of oxygen, anaerobic
cuvettes (1 ml or a 200 pl, Helma) were used. YgaF was made anaerobic by 10 cycles of
vacuum-argon while keeping the protein on ice to avoid precipitation and evaporation. In
some cases, trace amounts of oxygen were eliminated from the buffers by including 0.04
units/ml protocatechuic 3,4-dioxygenase (PCD) and its substrate 80 uM protocatechuic
acid (PCA) (both obtained from Sigma-Aldrich). Additives were prepared in sealed serum
vials and similarly treated by alternate cycles of vacuum and argon flushing. Solutions
were added using gas-tight syringes.

To assess whether the flavin chromophore was covalently attached to the enzyme,
YgaF (600 pl of 23 pM enzyme) was treated for 1 h on ice in the dark with 10 %
trichloroacetic acid.17 After centrifugation at 10,000 g for 10 min to separate the protein
pellet from the solution, the spectrum of the protein free sample was compared to that of
the original enzyme solution. Exactly the same procedure was followed for detaching the
flavin from YgaF for the cofactor identification experiment. The TCA-treated supernatant
after centrifugation was filtrated (5 kDa cut off, Amicon ultra filters) and the MALDI
spectra were obtained by the MS facility at MSU.

The effect of dithionite on the YgaF spectrum was examined. Anaerobic YgaF (10

pM) in 25 mM HEPES buffer with 100 mM NaCl, 1 mM DTT, 20 % glycerol at pH 8.2

116

was incubated with PCD and PCA in a 200 pl cuvette for 10 min at 4 °C to ensure
removal of all oxygen, and titrated with an anaerobic solution of sodium dithionite (2
mM). The reaction was monitored spectroscopically, and the fully reduced sample was
mixed with an equal volume of buffer equilibrated with 100 % oxygen. The concentration
of dithionite was calibrated by titration of a solution of authentic FAD (Sigma).

Photoreduction of YgaF was analyzed. Anaerobic enzyme (40 pM) was incubated
with PCD, PCA, and 5-deazaflavin (4 uM) on ice for 10 min, adjusted to contain 30 mM
EDTA, and exposed to light (using a 50 W halogen spotlight at a distance of 6 inches) at 4
°C. Spectra were monitored over time.

The effect of sulfite addition on the spectrum of YgaF was investigated. YgaF was
exchanged into 25 mM HEPES, 100 mM NaCl, 5 mM EDTA, 1 mM DTT, 20 % glycerol,
pH 7.0. Sodium sulfite (Baker Analyzed Reagents) was dissolved in the same buffer and
used to titrate the enzyme (15 pM), allowing the mixtures to equilibrate for 2 min before
monitoring their spectra. The dissociation constant Kd of sulfite was calculated by using
equation 3.

AA = (AAmax o[L]) / (K, + [L]) (eq. 3)

In this equation, AA is the observed change in absorbance at 450 nm, AA,“ax is the
maximum change in absorbance, and [sulfite] is the concentration of free ligand (in this
case sulfite).

Potential substrates were examined spectroscopically for their capacity to reduce
or oxidize the flavin of YgaF. Putative substrates were prepared as stock solutions (5-20
mM) in 25 mM HEPES, 100 mM NaCl, 5 mM EDTA, 1 mM DTT, 20 % glycerol, pH 8.2,

and made anaerobic by vacuum/argon cycling. To test whether these chemicals were

117

capable of reducing the enzyme flavin, anaerobic YgaF in the same buffer was titrated
with the test chemicals while monitoring the spectral changes. To test whether the
compounds would oxidize the bound F MN, the anaerobic enzyme was first treated with

sufficient dithionite to reduce the flavin and then titrated with the potential oxidants.

Determination of the Y gaF F MN redox potential

The reduction potential of the FMN cofactor of YgaF was determined
spectroscopically by reductive titration of a mixture of the protein and each of several
redox dyes. The most useful indicators were methylene blue (MB) and phenazine
methosulfate (PMS) with reduction potentials of -5 mV and +65 mV, respectively, at pH
7.5.‘8’19 YgaF (10 pM) and the selected redox dye (10 pM) were placed into an anaerobic
cuvette along with PCD in (25 mM HEPES buffer with 100 mM NaCl, 1 mM DTT, 20 %
glycerol at pH 7.5). The mixture was degassed and PCA was added to eliminate any traces
of oxygen. The sample was titrated with freshly prepared dithionite (calibrated by titration
of free F MN) while monitoring the changes in the spectrum after allowing time to achieve
redox equilibration. The reduction potential of the system (E) was calculated by using the
Nernst equation (Eq. 4) where Em is the midpoint potential of the redox dye, R is the gas
constant (8.314 J mol'1 K'l), T is the temperature in K, n is the number of electrons
transferred and F is the Faraday constant (96.5 J (mV)I mol"). By plotting log([YgaFox] /

[YgaF,cd]) versus E, the Em of the YgaF flavin was calculated.

2.3 RT [dYeoxl
log

 

 

E=Em+
(Eq- 4)

118

Growth studies

E. coli BW25113 and ygaF-KO are wild-type and ygaF deletion strains obtained
from the Keio collection.20 The cells were plated on LB agar containing 50 ug/ml
kanamycin and incubated overnight at 37 °C. Colonies from the plates were transferred
(using autoclaved cotton-tip applicators) into 15 ml Corning tubes that contained 4 ml of
M9 minimal medium that was missing the C-source. The specific C-sources of interest
were added to the cells (0.4%), and a portion (1 ml) was used to monitor the initial O.D. at
600 nm. The remaining samples (3 ml) were incubated at 37 °C for several days and the
final O.D. at 600 nm was monitored. The same procedure was followed for investigating
various N-sources (0.4%) using M9 minimal medium lacking ammonium chloride, but

containing succinate (as the C-source).

Biolog plates growth assays
Growths of the BW25113 and ygaF-KO strains were compared for different
carbon, nitrogen, phosphorous and sulfur sources by using Biolog plate assays. The

procedure followed that of Chapter 3.

Amino acid analysis

The abundance of amino acids and cellular amines was estimated for E. coli
BW25113 and ygaF-KO strains by amino acid analysis carried out by the Macromolecular
Structure Facility of MSU. The preparation of these samples is the same as that described

in Chapter 3.

119

Ortho-Phenylenediamine (OPDA) assay

A series of 2-hydroxyacid compounds were tested as substrates of YgaF by using
OPDA. This reagent is known to react with many or-ketoacids,2"22 the potential products
of the reaction. In addition, this reagent was used to determine the kinetic parameters of L-
2-hydroxyglutarate oxidation in comparison with authentic a-ketoglutarate. For the latter
experiments, a freshly prepared stock solution containing 10 mg OPDA in 10 ml of 1 M
phosphoric acid was adjusted to pH 2 and 25 ul B-mercaptoethanol was added. Samples
(7.76 ml) of 1 uM YgaF (in 25 mM HEPES, pH 7.0, buffer containing 100 mM NaCl, 5
mM EDTA, 1 mM DTT, and 20 % glycerol) were mixed with L-2-hydroxyglutarate (0-
400 pM) and incubated at room temperature. At selected time points, aliquots (0.97 ml)
were transferred to glass tubes containing OPDA (280 pl of 8.9 mM) which quenched the
reactions. To monitor the amount of product generated, samples were boiled 3 min,

cooled, and the absorbance at 340 nm determined.

Oxygen electrode assay

To assay any potential oxidase activity of YgaF, the enzyme was mixed with
various compounds while monitoring 0; consumption with a Clark-type oxygen electrode
(YSI Incorporated). The experiments were carried out at room temperature using air-
saturated buffer (25 mM HEPES, 100 mM NaCl, 5 mM EDTA, 1 mM DTT, 20 %

glycerol, pH 8.2), 2 uM YgaF, and 1 mM of the potential substrate in a 5 ml solution.

120

Organic acids HPLC analysis

Reaction mixtures (1 ml) containing 1 pM enzyme and 500 pM L-2-
hydroxyglutarate in 25 mM HEPES buffer (pH 8.2), 100 mM NaCl, 5 mM EDTA, 20 %
glycerol, and 2 mM DTT were incubated at room temperature for 4.5 h. Aliquots (300 pl)
were quenched with 5 pl of 6 M sulfuric acid, centrifuged at 10,000 g for 5 min, and 250
pl portions of each supernatant was filtered (Amicon ultrafree-MC from Millipore) at
10,000 g for 1 min. Samples (200 pl) was analyzed by using Waters Breeze HPLC system
equipped with an organic acids column (7.8 mm diameter by 300 mm, Bio-Rad) that had
been equilibrated with 13 mM sulfuric acid with detection by reductive index. The
concentrations of L-2-hydroxyglutarate (eluting at 23.0 min) and a-ketoglutarate (20.1

min) were determined by comparison to standards.

Synthesis of R and S-5-aminovalerate

For the synthesis of S-5-amino-2-hydroxy valerate, L-omithine was the starting
material (Scheme 7).23 L-omithine - HCl (1 g, 6 mmol) was dissolved in 3 ml water,
loaded onto an Amberlite IRA-120 (H+) column (8 ml of resin), washed with 30 ml of
water, and eluted with 40 ml of 3% ammonium hydroxide. The solution was lyophilized
and the powder was dissolved in 5 ml water. After cooling in ice (~ 0 °C), the solution
was mixed with 1.75 ml concentrated sulfuric acid, followed by dropwise addition of
sodium nitrite (1.35 g dissolved in 5 ml water) over a period of 1 h. The solution was
allowed to stand as the temperature was raised from 0 °C to 25 °C and incubated at room
temperature for 50 h. The unreacted nitrite was quenched with urea, which was added

until the solution gave a negative KI—starch test. The 1H NMR of sample dissolved in D20

121

revealed two multiplets at 1.5-1.8 ppm, a feature at 2.76 ppm, and a triplet at 3.85 ppm
similar to the previously reported NMR spectrum.23 For synthesis of R-S-amino-Z-
hydroxyvalerate, the same procedure was followed, but using D-omithine as the starting

material.
NH2

OH
2 eq NaN02
NH2 : NH2

mix in ice, then r.t. for ~ 50 h

 

Scheme 7. Synthesis of S-S-aminovalerate from L-ornithine.

122

RESULTS

Cloning and expression of ygaF

The E. coli gene ygaF was successfully ligated into pET-42b plasmid and
transformed into BL21(DE3) and C41(DE3) competent cells, as described in the
experimental methods. The optimized expression conditions of the ygaF gene were chosen
after comparison among growths in both cell types treated with several IPTG
concentrations (0, 0.1 and 1 mM final concentration). The gene of interest was expressed
in most of the conditions tested and appeared in both soluble and insoluble fractions
(Figure 38). The conditions that were chosen as the best were use of C41(DE3) cells and 1

mM IPTG, where the expressed level of soluble YgaF was satisfactory.

123

 

Figure 38. SDS-PAGE analysis of the expression conditions tested for ygaF.
Denatured samples were analyzed on a 12% acrylamide gel. Expression of ygaF in
BL21(DE3) cells: lanes la and l b, 0 mM IPTG; lanes 3a and 3b, 0.1 mM IPTG; and lanes
4a and 4b, 1 mM IPTG. Expression of ygaF in C41(DE3) cells: lanes 5a and 5b, 0 mM
IPTG; lanes 6a and 6b, 0.1 mM IPTG; and lanes 7a and 7b, 1 mM IPTG. Standard: lane 2.
The pellets are symbolized with ‘a’ and the cell extracts with ‘b’.

Purification of YgaF

YgaF containing a C-terrninal Hi36-tag was purified by Ni-NTA-Sepharose 6-Fast
Flow chromatography from soluble extracts of E. coli C41 (DE3) containing pET42b-
ygaF. Immediately after elution, the protein was exchanged into imidazole-free buffer to
enhance its stability. Glycerol, EDTA, and DTT further stabilized the protein (data not
shown). SDS-PAGE revealed the protein to be ~95 % homogeneous with a molecular
mass in agreement with that predicted from the sequence (47.64 kDa), as shown in Figure

39. A l L culture typically provided 55-65 mg of YgaF after the Ni-Sepharose column or

124

30-40 mg after Sephadex G-25 chromatography. Purified YgaF was stored in buffer D (25
mM HEPES, 100 mM NaCl, 5 mM EDTA, 1 mM DTT, 20 % glycerol, pH 8.2) at 4 °C,

conditions where activity was stable for several weeks.

 

Figure 39. . Purification of YgaF as monitored by SDS-PAGE analysis. Denatured
samples were analyzed on a 12% acrylamide gel. Lane 1, standards; lane 2, cell extracts;
lane 3, flow through of the Ni-NTA-Sepharose 6-fast flow column; lane 4, wash of the Ni
resin with buffer containing 100 mM imidazole; lane 5, YgaF elution from Ni resin with
buffer containing 500 mM imidazole; lane 6, YgaF elution from Superdex G-25.
(Molecular weight markers: phosphorylase b (M, 97.4 kDa), bovine serum albumin (M,
66.2 kDa), ovalbumin (M, 45.0 kDa), carbonic anhydrase (M, 31.0 kDa), trypsin inhibitor
(M, 21.5 kDa) and lysozyme (M, 14.4 kDa)).

Native size of His6-tagged YgaF

Two gel filtration columns were used for testing the polymerization state of YgaF.
The protein eluted from a Superdex 75 column and a Sephacryl 300 column primarily as
an aggregate. The results for the latter column are illustrated (Figure 40A and B). SDS-
PAGE revealed that YgaF eluted both at 44.16 min (the point where the protein
absorbance is greater at 280 nm) and in all subsequent fractions (Figure 40C). The same

results were noted when the salt concentration was increased from 100 mM to 300 mM

125

NaCl. In conclusion, most YgaF appeares to be highly aggregated and a portion of YgaF

appears to interact with the gel exclusion resins.

 

 

. . . v . . . ,
40 50 00 70 00 00 100 110 120
volume (ml)

 

 

swanwsooomsawmmmmmm
Timolmlmms]

Figure 40. Determination of the native size of YgaF by sephacryl 300
chromatography. A) A standard curve that correlates the elution time with the log(MW)
of the standard proteins (I)and the estimated position of YgaF (o). B) Elution of YgaF.
Time in minutes is equal to elution volume in ml. C) 12 % SDS-PAGE: lane 1, fraction
12; lane 2, fraction 13; lane 3, fraction 14; lane 4, fraction 15; lane 5, fraction 16; lane 6,
fraction 22; lane 7, fraction 24; lane 8, fraction 26; lane 9, fraction 28; lane 10, fraction 31;
lane 11, fraction 32; lane 12, fraction 33; lane 13, fraction 34; lane 14, standard; lane 15,
and fraction 35.

126

Testing whether the flavin binds covalently to YgaF and identification of the flavin
cofactor by MALDI analysis.

Purified YgaF is yellow in color, and its spectrum (maxima at 378 and 450 nm,
with a shoulder at 476 nm) is consistent with that of a flavoprotein. To distinguish whether
the flavin is covalently attached to YgaF, as is reported for several sequence-related
enzymes,'2’l3’l4 the protein was precipitated with trichloroacetic acid (TCA). The resulting
pellet was white whereas the supernatant remained yellow. The spectrum of the
supernatant was similar with that of the as-purified YgaF (Figure 41), demonstrating that
the cofactor is not covalently attached to YgaF. Direct confirmation of the presence of
FMN and not FAD, was obtained by MALDI-MS analysis, which revealed a feature at m/z
of 458.1 (data not shown) (the calculated values for FMN and FAD are 456.1 and 785.1

respectivelly).

127

0.4

Absorbance

0.2

 

 

wavelength (nm)

Figure 41. FMN is not covalently bound to YgaF. Black line: 23 pM untreated YgaF
with TCA, red lane: supernatant alter the centrifugation of YgaF that had been treated
with TCA. Baseline is the protein buffer.

Reduction and oxidation of YgaF

It is known that EDTA can be used as a source of redth in the presence of
small amounts of fiee flavins (like S-deazaflavin) and light to catalyze the photoreduction
of flavoenzymes.”25 Photoreduction of YgaF for 72 min in the presence EDTA and 5-
deazaflavin converted the protein to its reduced form, pointing out that YgaF undergoes a
two-electron reduction leading directly to the firlly reduced cofactor, with no semiquinone

intermediate observable (Figure 42).

128

 

 

 

300 400 500 600
Wavelength (nm)

Figure 42. Photoreduction of YgaF. Anaerobic YgaF (40 pM) in the presence of EDTA
and 5-deazaflavin (upper scan) was photoreduced at 4 °C and spectra were obtained after
7, 14, 36, 54, 72, 92 and 120 min.

Similarly, chemical reduction of YgaF with dithionite led to a smooth transition to
the two-electron reduced species (Figure 43). No anionic or neutral semiquinone species
was observed during the titration. Approximately 1.6 equivalents of dithionite were
required to reduce YgaF (Figure 43), consistent with the presence of some remaining
oxygen in the sample. The addition of an equal volume of buffer equilibrated with 100 %
oxygen to dithionite-reduced enzyme led to the immediate reoxidation of the FMN to half
(due to dilution) the starting spectral intensity. The rapid reoxidation of the flavin with

oxygenated buffer is evidence that YgaF is an oxidase.

129

A) B)

 

 

 

 

 

 

 

 

0.15
I 0.12
E 0.1 O
E O
010 g 0.00 o
‘— O
l i” ’
g 8 0.04 O .
005 2 DIR
O . , a , . .
O 0.4 0.6 1.2 1.6 2
equivalent: of dlhlonlte
am
am no 500 am 700

Figure 43. Reductive titration of YgaF by dithionite. A) Anaerobic YgaF (10 pM)
(upper scan) was treated with sequential additions of dithionite until the flavin was
completely reduced. The fully reduced sample was mixed with an equal volume of
oxygenated buffer, resulting in immediate reoxidation of the flavin (dashed line). B)
Absorbance of the flavin at 450 1101 versus the equivalents of dithionite added.

FIavin-sulfite adduct

Many flavin-containing oxidases (but not other flavoproteins) are bleached by
formation of a complex between sulfite and FMN.26 If a similar reaction takes place with
YgaF that would provide additional evidence that it is an oxidase. The addition of sulfite
to YgaF resulted in loss of the flavin absorbance (Figure 44), providing a sulfite
dissociation constant (Kd) of 102 i 7 pM at pH 7. A correlation has been noted between
the measured Kd of sulfite and the redox potentials of several flavoenzymes;27
extrapolation of those data allowed us to roughly estimate the redox potential of the YgaF

flavin as approximately —25 mV at pH 7.

130

A) B)
0.30 0.12 -

0.09 —.

Absorbance
O
a
A Abs at 450 nm
9
8

.0
O
U
r

 

300 460 500 96o 0 307' 400 e60 000
wavelength (nm) [sulfite] (pM)

Figure 44. A) Titration of YgaF with sulfite. YgaF (15 pM) was titrated with 25, 50, 75,
100, 125, 150, 200, 250, 500 and 700 pM sulfite in 25 mM HEPES buffer, pH 7.0,
containing 100 mM NaCl, 5 mM EDTA, 1 mM DTT, and 20 % glycerol. B) Absorbance
change at 450 nm versus the concentration of free sulfite.

Determination of the redox potential of the flavin

To more directly determine the reduction potential of the YgaF-bound FMN,
reductive titrations were carried out in the presence of redox dyes. To illustrate, YgaF was
mixed with MB (Em= -5 mV at pH 7.5), made anaerobic, and titrated with increasing
levels of dithionite (Figure 45A). YgaF was preferentially reduced over MB, leading to the
conclusion that the redox potential of the flavin should be higher than the MB potential.
By monitoring the concentration of oxidized MB (Am, = 666 nm with 3666 = 35,440 M ‘
cm’l, while its reduced form has essentially no absorbance at this wavelength),28 the
system reduction potential (E) after each addition could be determined by using the Nernst
equation. A comparison of E versus the log of the ratio of the oxidized to reduced YgaF

was used to determine the Em of +265 :h 5 mV for YgaF (Figure 45B).

131

A)

 

 

0.45
so
0 0 30
o ' 50
s E
E '“ 40 y . 22.313x+ 26.490
2 0.15 R2 - 090$
30 . . , , .
0.4 0.6 0.8 1 1.2
om log(FMNoxIFMNred)
$0 400 500 000 7W 000

 

 

 

wavelength (nm)

Figure 45. Analysis of the YgaF flavin reduction potential using MB as the redox dye.
A) An anaerobic mixture of YgaF and MB (10 pM each) was titrated with sodium
dithionite while monitoring the absorption spectrum. The changes in the relative
concentrations of the reduced and oxidized forms of MB were used to deduce the system
reduction potential (E) for each condition. B) Correlation of the measured E with the
log(FMNox/FMNM) of YgaF.

A second redox dye used to determine the redox potential of the YgaF-bound FMN
was phenazine methosulfonate (PMS), which has a midpoint potential at pH 7.5 of +65
mV. Upon reduction, PMS exhibits a sharp absorption feature at 388 nm (33in; = 21,390 M
' cm'l). The PMS absorbancies at 388 nm and 450 nm overlap with the YgaF
absorbancies, so it was necessary to calculate the fractions of PMS and YgaF that were
oxidized and reduced by taking into account the observed changes for PMS and YgaF
reduced separately. PMS was preferentially reduced over YgaF when treated together with
dithionite, indicating that the redox potential of the YgaF flavin was smaller than +65 mV
(Figure 46A). The estimated Em value for YgaF was +11 i 5 mV by comparison of E of
the system versus the log of the ratio of the oxidized to reduced YgaF (Figure 46B). Thus,
by taking into account both experiments (MB and PMS), we can conclude that the redox

potential of the flavin bound to YgaF is 19 i 8 mV.

132

A) B)

 

 

 

 

 

 

0.4
80
0.3
3 5- 60
a E 40
g 0.2 m 20 y = 74.3x+ 10.735
U)
2 0 l T ‘li l 1
°-1 0.2 0.4 0.6 0.8 1
.. I09([FMN]0x/[FMN]red)
"360 400 500 000 760

 

wavelength (nm)

Figure 46. Analysis of the YgaF flavin reduction potential using PMS as the redox
dye. A) An anaerobic mixture of YgaF and PMS (10 pM each) was titrated with sodium
dithionite while monitoring the absorption spectrum. The changes in the relative
concentrations of the reduced and oxidized forms of PMS were used to deduce the system
reduction potential (E) for each condition. B) Correlation of the measured E with the
log(FMNu/FMNM) of YgaF.

Growth studies

An attempt to identify the biological role of YgaF involved comparison of the
growths of the BW25113 and ygaF-KO strains while they grew on particular carbon and
nitrogen sources. The sources that were investigated included several L- and D-amino
acids, and molecules that could potentially be involved in GABA metabolism. The cells
were grown to stationary phase and their absorbances at 600 nm were compared. As
shown in Figure 47 A, both strains grew with succinate (positive control) and with L-Pro,
but none of the other C-sources led to growth, under these conditions. On the other hand,
when the same molecules were tested as N-sources (with NH4C1 as the positive control),
growth was observed in both strains for several conditions (Figure 47B). The most

interesting difference observed between the BW25113 and knock-out strains was using L-

133

Glu as an N-source, where there was less growth after 67 h of incubation for the mutant
strain, which disappeared after 90 h. L-Glu along with other interesting molecules were

tested as substrates of purified YgaF, as shown below.

Biolog plates analysis

The Biolog plate analysis (described in Chapter 3) was used to compare the
metabolism of BW25113 and ygaF-KO strains. The ygaF-KO mutant, like the csiD-KO
mutant, was generated by insertion of a kanamycin resistance gene into the reading frame
of the ygaF gene. Many carbon (PMl and PM2), nitrogen (PM3), phosphorous and sulfur
sources (PM4) were tested. C-sources that resulted in less extensive growth for the knock
out mutant compared to the BW25113 strain included: L-Asn, L-Glu, M-tartaric, pyruvic
acid, L-sorbose, S-keto-D-gluconic acid. The most interesting differences between the
strains involved the following N-sources: L-Glu, L-Thr, D-Asn, L-homoserine, agmatine,
N-acetyl-D-mannosamine, guanosine, and D,L-a-aminocaprylic acid. Moreover, several
N-sources that caused a slower growth in the ygaF-KO cells, which finally reached similar
growth level with the BW25113 cells were: L-Cys, L-Lys, L-Pro, L-Ser, L-ornithine and
S-amino-valeric acid. The P-sources that should some differences were: D-3-
phosphoglyceric acid, O-phospho-L-serine, and O-phospho-L-threonine. For S-sources,

conclusions could not be made because the negative control of both strains grew.

134

 

 

 

 

 

 

 

 

 

 

as 7.

_L_
5.25.1
2 1

 

RE: 88 oocafiomn< AS: 33 oocanaoon<

 

was—

 

Figure 47. Comparison of the growth of BW25113 (WT) and ygaF-KO strains with
different C- and N-sourees. A and B) Absorbance of strains grown on several C-sources
at 600 nm at 0 h (blue) and after 116 h (red). C and D) Absorbance of strains grown on

several N-sources at 600 nm at 67 h (blue) and after 90 h (red).

mp;

135

 

 

C)

 

Abs (600 nm)

WT DOS
KO pos
WT neg
K0 neg
WT L—Pro
K0 L-Pro
WT D-Pro
KO L-Val
WT D-Val
KO D-Val
WT agm
KO agm
WT sper
K0 L-Gln
WT GABA
KO GABA

 

K0 D-Pro
WT L-Val
KO sperm
WT L-Gln

 

 

 

 

 

 

 

D)
2
AL:
1.
E 1.4
o 1.2
O 1
80.3
3 0.6
0.4
< 0.2
0
1’0 2 ..
§§§§$$3$5¢$§§§§§§E
sageaseatsrigigoii
E¥E¥E¥Eg 9 ”56'
E 2
Figure 47 (continued).
Amino acid analysis

As described in detail in Chapter 3, amino acid analysis enables the comparison of
the cellular levels of free amino acids among different strains. This experiment was the
third attempt for identification of the physiological substrate of YgaF. The most
pronounced differences in abundance between the two strains are present in glutamic acid
and a-aminoadipic acid, as shown in Figure 48. Both amino acids were found to be in

136

smaller abundance in the ygaF-KO mutant compared to BW25113 cells. The fact that
similar behavior was reported for the csiD-KO mutant (see Chapter 3), can be interpreted
as indicating that the encoded proteins CsiD and YgaF might be closely correlated in a yet

unknown pathway.

0.5

Q ‘1
I

,0 .0

U A

Estimated abundance (nmoles)
A
I
F’
—h

Estimated concentration (nmoles)
o
'n

 

_A
l

 

Figure 48. Comparison of the amino acid analysis of the BW25113 (black) and ygaF-
KO (red) strains. Acronyms other than the of natural amino acids are: Pser
(phosphoserine), Tau (taurine), PEA (phosphoethanolamine), MetSO4 (methione
sulfoxide), Sar (sarcosine), a-AAA (a—aminoadipic acid), Cit (citruline), a-ABA (tr-amino-
n-butyric acid), Allo-Ile (allo-isoleucine), Cysthi (cystathionine), b-Aiba (IS-amino-
isobutyric acid), Hcys (homocysteine), EOPWHZ (ethanolamine), Hylys (5-
hydroxylysine), AEC (aminoethylcysteine), Om (omithine), S-AHCys (S-adenosyl
homocysteine), 1-MeHis (l-methyl histidine) and 3-MeHis (3-methyl histidine).

Investigations of potential substrates of YgaF

Several compounds were tested as potential substrates of YgaF by assaying for
their ability to (a) reduce FMN or oxidize FMNHz in the anaerobic enzyme, (b) stimulate
oxygen consumption, or (c) react with OPDA, a reagent for detecting a-ketoacids (Table
3). Significantly, neither the reduced nor the oxidized forms of NAD+ or NADP+ affected

137

the flavin spectrum; thus, YgaF is not a nicotinarnide-dependent enzyme. No
spectroscopic changes or 02 consumption activity was detected when YgaF was incubated
with GABA or several compounds that could plausibly be used in GABA production
(agmatine, putrescine, glutamic acid, glutamine, and the R- and S- isomers of 5-amino-2-
hydroxyvaleric acid). Similarly, no activity was detected for selected methylated
compounds (dimethylglycine, sarcosine), representative aldehyde (butyraldehyde) and 3-
hydroxyacid (3-hydroxybutyric acid) compounds. Furthermore most 2-hydroxyacids were
ineffective as substrates, including L- or DL-malic acid, DL-lactic acid, L- or D-mandelic
acid, and 2-hydroxycaproic acid. In contrast, as described below, robust activity was

detected in the case of L-2-hydroxyglutaric acid.

Table 3. List of the compounds tested as potential substrates of YgaF and the
methods that were used. The four experimental methods included oxygen consumption,
UV-vis absorption spectra of potential substrate added to the oxidized and to the reduced
form of YgaF, and the OPDA assay. The method(s) used for each hypothetical substrate is
indicated by “J”.

 

 

 

 

 

 

 

 

 

 

 

Substrate 02- Anaerobic Anaerobic OPDA
consump YgaF + YgaF + assay
tion substrate dithionite +
substrate
NADH "
NADPH "
D-Ala " "
Putrescine " "
DL-diarninopimelic acid "

Malic acid "
Butyraldehyde "
Sarcosine ./
n-butyl-CoA "

 

 

 

 

 

 

138

 

DL-2-OH-Glu

 

L-2-OH-Glu

 

D-2-OH-Glu

 

2-OH-caproic acid

'\

 

(+/-) citramalic acid

 

L-mandelic

\\\\\\\

 

D-mandelic

 

DL-lactic

 

DL-malic

\\\\\

 

L-malic

 

Dimethylglycine

 

3-OH-butyric acid

 

S-5-amino-2-OH-valeric acid

 

R-5-amino-2-OH-valeric acid

\'\'\\'\\

 

R/S-S-amino-2-OH-valeric acid

 

2-propanol

 

aKG

 

L-Glu

 

D-Glu

\\'\\

 

L-Arg

 

L-Gln

 

Agmatine

 

D(-)-3-phosphoglycerate

 

O-phospho-L-Ser

 

O-phospho-L-Thr

\'\\\

 

GABA

 

4-amino benzoic acid

 

Butyric acid

 

 

3-chloropropionic acid

 

 

 

\\'\'\

 

 

Table 3 (continued).

139

 

L-2-hvdroxvglutarate is a substrate for YgaF :

The behavior of L-2-hydroxyglutarate as a substrate of YgaF was examined in
greater detail. As illustrated in Figure 49, the addition of increasing concentrations of L-2-
hydroxyglutarate to an anaerobic solution of YgaF resulted in successive reduction of the
FMN (the feature at 410 nm was a contaminant in this particular preparation), with fully
reduced sample requiring about 1.5 equivalents of substrate. Analysis of the concentration
dependent changes in the difference spectra by use of equation 3 yielded an L-2-hydroxy-
glutarate Kd of 20 d: 4 pM. Furthermore L-2-hydroxyglutarate was shown to be a substrate
of YgaF according to both the oxygen electrode, and OPDA assays. The latter method was
used at several substrate concentrations to produce an initial velocity (v,) versus L-2-
hydroxyglutarate concentration graph, providing a Km of 95 i 26 pM, Vmax of 113 i 14

nmoles min'l (mg of protein)'l, and a turnover number, km, of 0.08 s'1 (Figure 50).

140

 

 

 

 

A) B)
0.004
0.9
0.55.
E 050- -
Q 0.04 3 0,45.
5 3 0.40-1 I
3 0.354 .
0.3-
0.304
I
0.25- I I
0.9 . i i “-29 i . . i i 1 i u .
am 400 500 600 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Wavelength (nm) nmoles L-2-hydroxyglutarata I nmoles YgaF
C) to-
E .
g 0.8-
E .
3 0.6-
.: .
g 0.4
E 0.2-
=8
E

 

2'0 4'0 6'0 81) 100 120 140
[L-2-hydroxyglutarateh] (pM)

Figure 49. Titration of anaerobic YgaF with 1.4-hydroxyglutarate. A) Anaerobic
YgaF (47 pM in 25 mM HEPES, pH 8.2, containing 100 mM NaCl, 5 mM EDTA, 1 mM
DTT, and 20 % glycerol) was adjusted to contain 19, 38, 57, 95, 133 and 171 pM L-2-
hydroxyglutarate. B) Change in absorbance at 450 nm as a function of the concentration of
free substrate in the solution. C) Fractional change in absorbance at 450 nm as a function
of the concentration of free L-2-hydroxyglutarate.

141

100-

v,- (nmoles/min*mg)

 

 

 

0 t 50 1100 700 ' 200 ' 250 ' 300 V 300
[L-2-hydr0xyglutarate] (pM)

Figure 50. Determination of the KIn and V,,,,.,, of the L-2-hydroxyglutarate oxidation
from a graph of the initial velocity (vi) versus its concentration.

The product of the reaction of YgaF with L-2-hydroxyglutarate was expected to be
a-ketoglutarate on the basis of the OPDA reactivity and the oxidase activity with this
substrate. The production of 330 pM a-ketoglutarate from 380 pM substrate was
confirmed by HPLC (details in the experimental methods). As shown in Figure 51A, in
the first control experiment (where YgaF is absent) L-2-hydroxy glutarate elutes after 23
min, and there is no observable peak at an elution time of 20.1 min, where aKG elutes. On
the other hand, when the enzyme is present and incubated at room temperature for the
same amount of time as the control experiment (4.5 h), the peak that corresponded to the
substrate disappeared and a new peak appeared where aKG elutes (Figure 51C). The
amount of aKG produced from the enzymatic reaction (330 pM) was calculated from its
standard curve. Although a-ketoglutarate is the product of the enzymatic reaction, the

addition of this oxoacid to anaerobic YgaF (with its FMN reduced by dithionite) did not

142

result in flavin oxidation; these results indicate the reaction is essentially irreversible (data
not shown), in agreement with the high reduction potential of the flavin.

In order to make sure that the Hisé-tag did not have a major interference with the
activity of YgaF, cell extracts containing the over-expressed His6-tagged protein were
compared on the basis of activity to extracts containing overexpressed un-tagged YgaF.
The amounts of aKG produced in both cell extracts in a period of 20 min at room
temperature, were very similar. This experiment demonstrates the His6-tag has little effect
on the activity of YgaF, though it is an estimation taking into account that the cell extracts

don’t contain exactly the same amount of YgaF.

143

 

A)

 

 

 

 

 

3)

 

 

 

C)

 

ittitiiii

 

 

El
8
auto

 

 

 

Figure 51. Identification of the product of the enzymatic reaction of L-2-
hydroxyglutarate. Organic acid HPLC analysis for A) 380 pM L-2-hydroxyglutarate, B)
1 pM YgaF and C) reaction of 1 pM YgaF with 380 pM L-2-hydroxyglutarate at room
temperature for 4.5 h. The bufi’er for all three HPLC samples was : 25 mM HEPES, 100
mM NaCl, 5 mM EDTA, 20 % glycerol, 2 mM DTT, pH 8.2. The peak at ~26.8 min
elution time is glycerol.

The D-2-hvdroxvglutarate isomer is not a substrate of YgaF:

In contrast to the results with L-2-hydroxyglutarate, no flavin reduction was
detected using D-2-hydroxyglutarate. Furthermore, only very low levels of activity were
detected using the OPDA procedure. As illustrated in Figure 52, the low level reactivity of
D-2-hydroxyglutarate was transient and was repeated when another aliquot of the
substrate was added, consistent with the enzyme remaining active and suggesting that D-
2-hydroxyglutarate contains a low concentration of the L isomer. As further confirmation
that the enzyme retains activity after incubation with the D isomer, high level of activity
was observed when L-2-hydroxyglutarate was added to the reaction mixture (afier the
system had reached the first platue). The aKG production under these conditions was
identical with that where the D-isomer was absent, indicating that the D-isomer doesn’t

significantly inhibit YgaF.

145

 

 

l
100- 0
80-
E .0. .
8
a 40‘
o
20- °
. O A A
AA
0.1 1 ' r V I T I V I T I T T V ‘7 ‘ j ‘
O 5 10 15 20 25 30 35 40 45

flme (min)

Figure 52. Timecourse of the aKG production for the reaction of YgaF with the
isomers of 2-hydroxyglutarate. YgaF (1 pM) was incubated with D-2-hydroxyglutarate
(300 pM) in 25 mM HEPES buffer containing 100 mM NaCl, 1 mM DTT, and 20 %
glycerol at pH 8.2 at 25 °C. Aliquots of 970 pl were collected fi'om the reaction mixture at
1, 3, 5, 10 and 25 min and treated with OPDA (black squares). The remaining reaction
mixture was separated into two equal portions (7.76 ml). D-2-hydroxyglutarate (300 pM)
was added to one (blue triangles); and L-2-hydroxyglutarate (300 pM) was added to the
other (red cycles). Aliquots of 970 pl were collected from both reactions at 1, 3, 5, 10, 20
and 40 min and treated with OPDA

D-3- h h i r su f Y

As discussed above, the ygaF-KO mutant grew slower but finally reached the same
growth point with the BW25113 strain on D-3-phosphoglycerate as a P-source. 3-
Phosphoglycerate is a close structural mimic of 2-hydroxyglutarate (Scheme 8); thus, we
tested this common cellular intermediate as a substrate of YgaF. As shown in Figure 53A,
90, 270 and 540 pM of 3-phosphoglycerate led to only sligh reduction of 18 pM YgaF,
whereas 900 pM of this particular substrate was able to reduce the enzyme at a slow rate.

Therefore, D-3 -phospho-D-glycerate, acts as a poor substrate of the enzyme. By contrast,

I46

two other phosphorylated compounds, O-phospho-L-serine and O-phospho-L-threonine

exhibited no activity.
OH 0 | l O
HO—P
OWOH \O/\HJ\OH
HO
OH OH
2-hydr0xyglutarate 3-phosphoglycerate

Scheme 8. Structural similarity of 2-hydroxyglutarate and 3-phosphoglycerate.

 

 

 

 

 

 

 

A) B)
0.30
0.24 025.
0.18 °'" . .
° 0104
3 0.12 ° .
< ° ° ° 1
0.06 015-1
0 . . . , . . . .
0.10-
0 20 4o 60 80 100 120 14c
tlma(nin) 000-
300 40) 50.0 000 no

Figure 53. Titration of anaerobic YgaF with D-3-phosphoglycerate. A) Change in
absorbance at 450 nm as a function of time. Anaerobic YgaF (18 pM in 25 mM HEPES,
pH 8.2, containing 100 mM NaCl, 5 mM EDTA, 1 mM DTT, and 20 % glycerol) was
adjusted to contain 90 (blue squares), 270 (pink squares), 540 (yellow triangles) and 900
pM (cyan cycles) D-3-phosphoglycerate. B) Anaerobic YgaF incubated with 900 pM of
D-3-phosphoglycerate for 10, 20, 35, 50, 65, 80 and 93 min.

147

CONCLUSIONS

The E. coli ygaF gene is located in the csiD-ygaF—gabDTP-csiR operon, controlled
by the csiDp promoter and activated exclusively under carbon starvation together with the
csiD gene. It encodes a protein, YgaF, which has been considered a putative
dehydrogenase with unknown function. In this study I cloned, expressed, purified and
characterized the general properties of YgaF.

Cloning, expression and purification. In this work, the E. coli ygaF gene was
cloned and expressed with success for the first time. The encoded protein, C-terminus-
Hisé-tagged YgaF, was easily purified in one step by using a Ni-NTA affinity column. The
color of the purified protein was yellow, as it was expected for a flavin-containing protein.
The protein was more than 95 % homogeneous, according to the SDS-PAGE, and
obtained with an overall yield of 30-40 mg per L growth. YgaF was stable in 25 mM
HEPES buffer that contained 100 mM NaCl, 20 % glycerol, 5 mM EDTA and 1 mM DTT
at pH 8 for several weeks at 4 °C. The attempt to estimate the native size of the YgaF by
gel filtration chromatography was inconclusive, due to unidentified interactions between
the protein and the resin, but the major portion of the protein was a large aggregate.

YgaF is a FMN bound oxidase. The absorbance spectrum of the purified enzyme
has the characteristic shape of a bound flavin, with dual maxima at 378 nm and 450 nm.
The flavin of YgaF was proven to be non-covalently bound FMN. Photoreduction of YgaF
occurred in the presence of EDTA and S-deazaflavin with no interrnediated semiquinone
observed. This experiment indicates that the YgaF goes through two-electron reduction in
a single step. Similar results were noted when the flavin was reduced chemically by

dithionite. Immediate reoxidation of chemically-reduced YgaF by addition of an equal

148

volume of oxygenated buffer implies that YgaF is an oxidase. Support for this assignment
was obtained by a characteristic reaction catalyzed only by oxidases involving the
formation of a flavin-sulfite adduct through a covalent bond at the N(5) position of the
isoalloxazine ring of the flavin. The dissociation constant (Kd = 102 pM) of the FMN-
sulfite complex is in the micromolar range, as commonly observed for flavin-containing
oxidases,29 and was used to approximate the reduction potential of the protein (-24 mV at
pH 7). This value is close to the reduction potential measured spectroscopically when
using the MB and the PMS as redox dyes (+19 mV at pH 7.5), and suggests that Em
increases with increasing pH.

Searching for the YgaF substrate. Several methods were used to identify the
primary substrate of YgaF. Growths of E. coli BW25113 and ygaF-KO strains revealed
that the mutant grows more slowly than the wild type cells using L-Glu as the N-source.
Further studies indicated that L-Glu is not a substrate of YgaF. Biolog plate analysis,
enabled the comparison of growths of E. coli BW25113 and ygaF-KO for a large variety
of different carbon, nitrogen, phosphorous and sulfur sources. This study identified several
potentially interesting compounds, but none were subsequently shown to be substrates of
YgaF. As mentioned in the discussion part of Chapter 3 for the csiD gene, it cannot be
excluded that E. coli possesses more than one pathway to metabolize the specific
compound of interest, so that the analyses of the growth of the ygaF-KO and BW25113
strains may not be definitive.30 A third approach to compare the BW25113 and the ygaF-
KO strains was by amino acid analysis. This method specifically detects the level of
amino acids and specific other compounds in the cells. The mutant cells contained less L-

Glu, a result that might be connected with the slower growth rates on this N-source for the

149

knock-out mutant compared to the WT. Other experiments attempted to directly identify
the substrates of YgaF using purified YgaF and four detection methods. These were
assaying for the ability to reduce FMN or to oxidize FMNHz under anaerobic conditions,
to consume oxygen, or to react with OPDA, which is a reagent that can react with or-
ketoacids. The molecules investigated included compounds involved in GABA
catabolism, along with alcohols, aldehydes, primary and secondary amines, acids, amino
acids, and a and B-hydroxy acids. Most molecules failed to react with YgaF when
examined by any method. The only compounds that showed activity were L-2-
hydroxyglutarate and the D-3-phosphoglycerate, with the latter one showing that it was
poor substrate for YgaF.

L-2-hydroxyglutarate is a substrate of YgaF. YgaF exhibits robust activity toward
L-2-hydroxyglutarate (Scheme 9), while negligible activity was detected with the D
isomer. The use of L-2-hydroxyglutarate as a substrate was not surprising given the close
sequence similarity (41 % identity) between this E. coli protein and human FAD-
dependent L-2-hydroxyglutarate dehydrogenase. Furthermore, the finding that YgaF is an
oxidase rather than a dehydrogenase is consistent with the ability of the protein to generate
a complex with sulfite, as discussed above. It is instructive to compare YgaF to the
mammalian enzyme L-2-hydroxyglutarate dehydrogenase to which it is related. The Km of
YgaF for L-2-hydroxyglutarate (95 pM) is lower than those reported for L-2-
hydroxyglutarate dehydrogenases from human or rat (800 pM and 150 pM,
respectively).7’31 The dissociation constant K, of the YgaF-L-Z-hydroxyglutarate was
found to be 20 i 4 pM. The expected product of this enzymatic reaction (aKG) was

shown to be produced by using both the OPDA assay (because OPDA reacts with 2-keto

150

acids) and by the organic acids HPLC column, where the product of the reaction had the
same elution time with aKG. Importantly, the reaction seems to be irreversible. The
relatively high reduction potential of this flavoprotein explains the apparent irreversibility
of the reaction (i.e., the reduction of a-ketoglutarate by anaerobic, reduced YgaF) when

monitored spectroscopically. Also, the Hisb-tag seems not to affect the activity of YgaF.

 

OH YgaF O
HOOCMCOOH o HoocMCOOH
2
L-2-hydroxyglutaric acid a-ketoglutaric acid

Scheme 9. The catalytic YgaF reaction with L-2-hydroxyglutarate.

The potential relevance of the L-2-hydroxyglutarate activity of YgaF to E. coli. In
mammalian cells, L-2-hydroxyglutarate has been suggested to arise from the non-specific
reduction of a-ketoglutarate by L-malate dehydrogenase. Thus, the physiological role of
the human enzyme is proposed to be a metabolite repair enzyme to prevent accumulation
of this compound in tissues.32 Mutation of the gene encoding human L-2-hydroxyglutarate
dehydrogenase (in particular, mutations associated with K81E and E176D variants or
deletion of exon 9) leads to such accumulation,“ a disease state known as L-2-
hydroxyglutarate aciduria and characterized by ataxia, mental deficiency with subcortical
leukoencephathology, and cerebellar atrophy.7 L-malate dehydrogenase may play a similar
role in E. coli, but in addition SerA is known to reduce a-ketoglutarate to form both D-
and L-2-hydroxyglutarate.33 The normal reaction catalyzed by SerA is the oxidation of 3-
phospho-D-glycerate (a poor substrate of YgaF) to form 3-phospho-hydroxypyruvate,

using NAD+ as cofactor. We propose that E. coli uses YgaF to oxidize L-2-
151

hydroxyglutarate (and 3-phospho-L-glycerate), where these compounds arise by such

aberrant side reactions.

152

REFERENCES

' Chance, M. R., Bresnick, A. R., Burley, S. K., Jiang, J. S., Lima, C. D., Sali, A., Almo,
S. C., Bonanno, J. B., Buglino, J. A., Boulton, S., Chen, H., Eswar, N., He, G., Huang, K,
Ilyin, V., McMahan, L., Pieper, U., Ray, 8., Vidal, M., and Wang, L. K. (2002) Structural
genomics: a pipeline for providing structures for the biologist. Protein Sci. 11, 723-738.

2 Metzner, M., Gerrner, J ., and Hengge, R. (2004) Multiple stress signal integration in the
regulation of the complex (rs-dependent csiD-ygaF -gabDT P operon in Escherichia coli.
Molec. Microbiol. 51, 799-811.

3 Marshchall, C., Labrousse, V., Kreimer, M., Weichart, D., Kolb, A., and Hengge-Aronis
R. (1998) Molecular analysis of the regulation of csiD, a carbon starvation-inducible gene

in Escherichia coli that is exclusively dependent on sigmaS and requires activation by
cAMP-CRP. J. Mol. Biol. 276, 339-353.

4 Schneider, B. L., Ruback, S., Kiupakis, A. K., Kasbarian, H., Pybus, C., Reitzer, L.
(2002) The Escherichia coli gabDTPC operon: specific y-aminobutyrate catabolism and
nonspecific induction. J. Bacterial. 184, 6976-6986.

5 De Colibus, L., and Mattevi, A. (2006) New frontiers in structural flavoenzymology.
Curr. Opin. Struct. Biol. 16, 722-728.

6 Walsh, C. T. (1979) Enzymatic reaction mechanisms. W. H. Freeman and Company, San
Francisco.

7 Rzem, R., Veiga-da-Cunha, M., Noél, G., Goffette, S., Nassogne, M.-C., Tabarki, B.,
Schtiller, C., Marquardt, T., Vikkula, M., and Van Schafiingen, E. (2004) A gene
encoding a putative FAD-dependent L-2-hydroxyglutarate dehydrogenase is mutated in L-
2-hydroxyglutaric aciduria. Proc. Natl. Acad. Sci. USA 101, 16849-16854.

8 Tomb, J. F., White, 0., Kerlavage, A. R., Clayton, R. A., Sutton, G. G., F leischmann, R.
D., Ketchurn, K. A., Klenk, H. P., Gill, 8., Daugherty, B. A., Nelson, K., Quackenbush,

J ., Zhou, L., Kirkness, E. F., Peterson, 8., Loftus, B., Richardson, D., Dodson, R., Khalak,
H. G., Glodek, A., McKenney, K., F itzergerald, L. M., Lee, N., Adams, M. D., Hickey, E.
K., Berg, D. E., Gocayne, J. D., Utterback, T. R., Peterson, J. D., Kelley, J. M., Cotton, M.
D., Weidman, J. M., Fujii, C., Bowman, C., Watthey, L., Wallin, B, Hayes, W. S.,
Borodovsky, M., Karp, P. D., Smith, H. 0., Fraser, C. M., and Venter, J. C. (1997) The
complete genome of the gastric pathogen Helicobacter pylori. Nature 388, 539-547.

153

9 Suzuki, K., Sagai, H., Imamura, S., and Sugiyama M. (1994) Cloning, sequencing,
overexpression in Escherichia coli of a sarcosine oxidase-encoding gene linked to the
Bacillus creatinase gene. J. Ferm. Bioeng. 77, 231-234.

'0 Binzak, B. A., Vockley, J. G., Jenkins, R. 13., and Vockley, J. (2000) Structure and
analysis of the human dimethylglycine dehydrogenase gene. Molec. Genet. Metab.
69,181-187.

'1 Kunst, F., Ogasawara, N., Moszer, I., Albertini, A. M., Alloni, G., Azevedo, V.,
Bertero, M. G., Bessieres, P., Bolotin, A., Borchert, S., Borriss, R., Boursier, L., Brans,
A., Braun, M., Brignell, S. C., Bron, S., Brouillet, S., Bruschi, C. V., Caldwell, B.,
Capuano, V., Carter, N. M., Choi, S.-K., Codani, J .-J ., Connerton, I. F., Cummings, N. J .,
Daniel, R. A., Denizot, F., Devine, K. M., Diisterhbft, A., Ehrlich, S. D., Emmerson, P.
T., Entian, K. D., Errington, J ., Fabret, C., Ferrari, E., Foulger, D., Fritz, C., Fujita, M.,
Fujita, Y., Fuma, S., Galizzi, A., Galleron, N., Ghim, S.-Y., Glaser, P., Goffeau, A.,
Golightly, E. J ., Grandi, G., Guiseppi, G., Guy, B. J ., Haga, K., Haiech, J ., Harwood, C.
R., Henaut, A., Hilbert, H., Holsappel, S., Hosono, S., Hullo, M.-F., Itaya, M., Jones, L.,
Joris, B., Kararnata, D., Kasahara, Y., Klaerr-Blanchard, M., Klein, C., Kobayashi, Y.,
Koetter, P., Koningstein, G., Krogh, S., Kumano, M., Kurita, K., Lapidus, A., Lardinois,
S., Lauber, J ., Lazarevic, V., Lee, S.-M., Levine, A., Liu, H., Masuda, S., Mauél, C.,
Medigue, C., Medina, N., Mellado, R. P., Mizuno, M., Moesti, D., Nakai, S., Noback, M.,
Noone, D., O'Reilly, M., Ogawa, K., Ogiwara, A., Oudega, B., Park, S.-H., Parro, V.,
Pohl, T. M., Portetelle, D., Porwollik, S., Prescott, A. M., Presecan, E., Pujic, P., and
Pumelle, B. (1997) The complete genome sequence of the Gram-positive bacterium
Bacillus subtilis. Nature 390, 249-256.

'2 Dodt, G., Kim, D. G., Reimann, S. A., Reuber, B. E., McCabe, K., Gould, S. J ., and
Mihalik S. J. (2000) L-Pipecolic acid oxidase, a human enzyme essential for the

degradation of L-pipecolic acid, is most similar to the monomeric sarcosine oxidases.
Biochem. J. 345, 487-494.

‘3 Hassan-Abdallah, A., Zhao, G., and Jams M. S. (2006) Role of the covalent flavin
linkage in monomeric sarcosine oxidase. Biochemistry 45, 9454-9462.

'4 Leys, D., Basran, J ., and Scutton, N. S. (2003) Channeling and formation of 'active'
formaldehyde in dimethylglycine oxidase. EMBO J. 22, 4038-4048.

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

154

'6 Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head
of bacteriophage T4. Nature (London) 22 7, 680-685.

‘7 Mihalik, S. J ., McGuinness, M., and Watkins, P. A. (1991) Purification and
characterization of peroxisomal L-pipecolic acid oxidase from monkey liver. J. Biol.
Chem. 266, 4822-4830.

'8 Pollegioni, L., Porrini, D., Molla, G., and Pilone, M. S. (2000) Redox potentials and
their pH dependence of D-amino acid oxidase of Rhodotorula gracilis and Trigonopsis
variabilis. Eur. J. Biochem. 267, 6624-6632.

‘9 Prince, R. C., Linkletter, S. J. G., and Dutton, L. (1981) The thermodynamic properties
of some commonly used oxidation-reduction mediators, inhibitors and dyes, as determined
by polarography. Biochim. Biophys. Acta 635, 132-148.

20 http://ecoli.naist.jp/gb6/ResourceS/deletion/deletion.html.

2‘ Hayashi, T., Tsuchiya, H., Todoriki, H., and Naruse, H. (1982) High-performance liquid
chromatographic determination of alpha-keto acids in human urine and plasma. Anal.
Biochem. 122, 173-179.

22 Romanov, V., Merski, M. T., and Hausinger, R. P. (1999) Assays for allantoinase. Anal.
Biochem. 268, 49-53.

23 Sefler, A. M., Kozlowski, M. C., Guo, T., and Bartlett, P. A. (1997) Design, Synthesis,
and Evolution of a Depsipeptide Mimic of Tendamistat. J. Org. Chem. 62, 93-102.

24 Massey, V., Stankovich, M., and Hemmerich, P. (1978) Light-mediated reduction of
flavoproteins with flavins as catalysts. Biochemistry 17, 1-8.

25 Massey, V., and Hemmerich, P. (1978) Photoreduction of flavoproteins and other
biological compounds catalyzed by deazaflavins. Biochemistry 17, 9-17.

2" Massey, V., Muller, F., Feldberg, R., Schuman, M., Sullivan, P. A., Howell, L. G.,
Mayhew, S. G., Matthews, R. G., and Foust, G. P. (1969) The reactivity of flavoproteins

with sulfite. Possible relevance to the problem of oxygen reactivity. J. Biol. Chem. 244,
3999-4006.

27 Mtiller, F., and Massey, V. (1969) Flavin-sulfite complexes and their structures. J. Biol.
Chem. 244, 4007-4016.

155

28 Pande, S., Ghosh, S. K., Nath, S., Praharaj, S., Jana, S., Panigrahi, S., Basu, S., and Pal,
T. (2006) Reduction of methylene blue by thiocyanate: kinetic and thermodynamic
aspects. J. Colloid Interface Sci. 299, 421-427.

29 Job, V., Marcone, G. L., Pilone, M. S., and Pollegioni, L. (2002) Glycine oxidase from
Bacillus subtilis. J. Biol. Chem. 2 77, 6985-6993.

30 Kim, J ., and Copley, S. D. (2007) Why metabolic enzymes are essential or nonessential
for growth of Escherichia coli K12 on glucose. Biochemistry ASAP.

3 I Rzem, R., Van Schaftingen, E., and Veiga-da-Cunha, M. (2006) The gene mutated in L-
2-hydroxyglutaric aciduria encodes L-2-hydroxyglutarate dehydrogenase. Biochimie 88,
113-116.

32 Rzem, R., Vincent, M.-F., Van Schaftingen, E., and Veiga-da-Cunha M. (2007) L-2-
Hydroxyglutaric aciduria, a defect of metabolite repair. J. Inherit. Metab. Dis. ,in press.

33 Zhao, G., and Winkler, M. E. (1996) A novel a-ketoglutarate reductase activity of the

serA-encoded 3-phosphog1ycerate dehydrogenase of Escherichia coli K-12 and its
possible implications for human 2-hydroxylglutaric aciduria. J. Bacteriol. 178, 232-239.

156

M

Illllilllllnlglilll

       

lllll