.
,1 0..

.y .. .{1 (11,51.

63 . -
x. swung p.§..rnn._v.:u.u...

{darn

 

W5...»

.m
5 .. an S,

. v

i n,
0‘ .2 In???

4.3
toxins?!
,a‘J. , v. It!“ I
.sptrnl:
. V 5313. ‘
2617:! '

k.
:33?! .
Q...».\. \
15“..
e... 3.

. 3.3.:
.\. t

A‘ .Q."

|.Ilv&
2 “ .. ‘11)}:

My:

(,9
’_)r if

Mullah

This is to certify that the
dissertation entitled

STUDIES OF TWO IRON-CONTAINING ENZYMES FROM
ESCHERICHIA COLI: PYRUVATE FORMATE-LYASE
ACTIVATING ENZYME AND ALKB

presented by

TIMOTHY FLEMING HENSHAW

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Chemistry

 

 

aefifim LL ,_ .

U ”Majfirofesso s I nature
2% 5::

 

Date

MSU is an Aflinnative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/Dateouepss-pJS

 

STUDIES OF TWO IRON-CONTAINING ENZYMES FROM ESCHERICHIA COLI:
PYRUVATE FORMATE-LYASE ACTIVATING ENZYME AND ALKB

By

Timothy Fleming Henshaw

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Chemistry

2004

ABSTRACT
STUDIES OF TWO IRON-CONTAINING ENZYMES FROM ESCHERICHIA COLI:
PFL-AE and ALKB
By

Timothy Fleming Henshaw

Pyruvate forrnate-lyase activating enzyme (PFL-AE) is an iron-sulfur protein that
catalyzes the generation of an essential radical at glycine-734 of pyruvate formate-lyase
(PFL). This radical-generating reaction is dependent on S-adenosylmethionine (AdoMet)
and is thought to involve a 5’-deoxyadenosyl radical that is generated by the reductive
cleavage of AdoMet. PFL—AE was purified under anaerobic conditions and the iron-
sulfur cluster was characterized by a combination of EPR, Mossbauer, and UV-visible
spectroscopies. As isolated, PFL-AE was found to contain a variety of iron-sulfur
clusters, including [4Fe—4S]2+, [2Fe-28]2+, and linear and cuboidal [3Fe-4S]+ clusters,
with the cuboidal [3Fe-4S]Jr being the dominant form of the cluster. These preparations
were found to exhibit high specific activity in the absence of added iron. Under reducing
conditions, these cluster types were converted to a mixture of [4Fe-4S]2+ and [4Fe-4S]1+
clusters.

In order to investigate the possibility of a unique iron site in the cluster of PF L-
AE, a dual-isotope approach (56Fe/57Fe) was used to selectively label the [4Fe-4S]2+
cluster. Mossabuer spectroscopy of these samples demonstrated that the 57Fe was
incorporated into the unique site. In the presense of AdoMet, a large increase in the

isomer shift was observed (0.43 mm/s to 0.73 mm/s) indicating that AdoMet binds

directly to this site in the cluster. In addition, detailed characterization of a single
turnover of PFL-AE shows that the [4Fe-4S]l+ cluster provided the electron that is
required for the reductive cleavage of AdoMet.

AlkB from Escherichia coli is a component of the “Ada response,” which is a
coordinated defense against the toxic and mutagenic effects of environmental and
endogenous alkylating agents. Despite its identification in 1983, the function of AlkB
had not been determined. Theoretical fold prediction had suggested that AlkB might be a
member of the iron and a-ketoglutarate dependent dioxygenase superfamily. This
possibility was tested, and the results in this dissertation Show that AlkB functions by
coupling the oxidative decarboxylation of a-ketoglutarate to the oxidation of the aberrant
methyl group, resulting in release of formaldehyde and restoration of the original base.

In addition, in the absence of primary substrate, AlkB was found to catalyze the oxidation

of a tryptophan residue located close to the metal center.

Dedicated to the memory of
Judith Fleming Henshaw

April 17, 1945 — June 30, 2003

iv

Table of Contents

List of Figures ................................................................................................................... vii
List of Schemes .................................................................................................................. ix
List of Tables ...................................................................................................................... x
List of Abbreviations ......................................................................................................... xi
INTRODUCTION .............................................................................................................. l
The Fe/S AdoMet enzymes ............................................................................................. 2
Pyruvate forrnate-lyase ............................................................................................... 2
Pyruvate forrnate-lyase activating enzyme ................................................................. 3
Ribonucleotide reductase activating enzyme .............................................................. 7
Lysine aminomutase ................................................................................................... 8
Sulfur insertion: BioB and LipA .............................................................................. 10
Spore photoproduct lyase .......................................................................................... 11
DNA repair and Fe(II)/a-ketoglutarate dioxygenases .................................................. 12
The Adaptive response to alkylating agents ............................................................. 13
Fe(II)/a-ketoglutatarate Dioxygenases ...................................................................... 16
Amino Acid Hydroxylases ........................................................................................ l7
Fe(II)/a-KG Dioxygenase Mechanism ..................................................................... 18
Non-heme Models 20
Identification of an F e(IV)=O Intermediate .............................................................. 20
AlkB as an Fe(II)/a-ketoglutarate Dioxygenase ....................................................... 22
References ..................................................................................................................... 23
CHARACTERIZATION OF THE IRON-SULF UR CENTER OF PYRUVATE
FORMATE-LYASE ACTIVATING ENZYME .............................................................. 29
11.1 Introduction ............................................................................................................ 30
11.2 Experimental Methods ........................................................................................... 32
Grth and expression of PFL ................................................................................. 32
Subcloning, growth, and expression of PFL-AE ...................................................... 32
Purification of pyruvate formate-lyase-activating enzyme (PFL-AB) ...................... 33
Activity, protein and iron assays ............................................................................... 34
EPR and Mossbauer Spectroscopy ........................................................................... 35
[1.3 Results .................................................................................................................... 36
Enzyme Activity ....................................................................................................... 36
Characterization of purified PFL-AE ........................................................................ 36
Characterization of the dithionite-reduced enzyme .................................................. 45
11.4 Discussion .............................................................................................................. 51
References ..................................................................................................................... 53
IDENTIFICATION OF A UNIQUE IRON SITE ............................................................ 55
111.] Introduction ........................................................................................................... 56
1112 Experimental Methods .......................................................................................... 57

Preparation of Specifically Labeled Samples ........................................................... 58

111.3 Results and Discussion ......................................................................................... 58
References ..................................................................................................................... 63
THE ACTIVE FORM OF THE CLUSTER: SINGLE TURNOVER EXPERIMENTS 64
NJ Introduction .......................................................................................................... 65
IV.2 Methods ............................................................................................................... 67
Preparation of single turnover samples ..................................................................... 67
EPR Spectroscopy ..................................................................................................... 68
IV.2 Results and Discussion ........................................................................................ 68
OXIDATIVE DNA REPAIR ............................................................................................ 74
V.1 Introduction ............................................................................................................ 75
V2 Experimental Methods ........................................................................................... 76
Expression and purification of AlkB ........................................................................ 76
Radioactive substrates and assays ............................................................................. 76
HPLC assays ............................................................................................................. 77
High-level methylation of an oligionucleotide ......................................................... 77
UV-visible spectroscopy ........................................................................................... 78

Gas chromatography/mass spectrometry .................................................................. 78
Oxygen consumption ................................................................................................ 78

V3 Results and Discussion ........................................................................................... 79
Release of ethanol-soluble radioactivity ................................................................... 79
UV-visible spectroscopy ........................................................................................... 79
Identification of l-meA and 3-meC as AlkB substrates ........................................... 83
Formaldehyde release and oxygen consumption ...................................................... 84
References ..................................................................................................................... 88
ABBERANT ACTIVITY OF THE DNA REPAIR ENZYME ALKB ............................ 9O
V1.1 Introduction ........................................................................................................... 91
V1.2 Experimental ......................................................................................................... 92
Sample Preparation ................................................................................................... 92
UV-visible (UV-vis) spectroscopy ........................................................................... 92
Mass Spectrometry (MS) .......................................................................................... 93
Structural Modeling .................................................................................................. 94
V1.3 Results and Discussion ......................................................................................... 95
Chromophore Formation ........................................................................................... 95
Mass spectrometric analysis ..................................................................................... 96
Protein Modeling ...................................................................................................... 99
Conclusions ............................................................................................................. 101
References ................................................................................................................... 103
CONCLUSIONS AND FUTURE DIRECTIONS .......................................................... 105
References ................................................................................................................... 1 10

vi

List of Figures

Figure 11.1 UV—visible absorption spectrum of pyruvate formate-lyase activating enzyme
as isolated. ................................................................................................................. 37

Figure 11.2 X-band EPR spectra of pyruvate formate-lyase-activating enzyme as a

function of temperature. ............................................................................................ 38
Figure 11.3 Mossbauer spectra of anaerobically purified native PFL-AB. ....................... 40

Figure 11.4 Mdssbauer spectrum of dithionite-reduced native PFL-AE recorded at 4.2 K

in a magnetic field of 0.05 T applied parallel to the y-beam ..................................... 46
Figure 11.5 Mbssbauer spectra of dithionite-reduced PFL-AE ......................................... 47

Figure 11.6 EPR spectrum of dithionite-reduced PF L-AE showing a weak signal
originating from the [4Fe-4S]Jr cluster. ..................................................................... 48
Figure 111.1 Mossbauer spectra of 56Fe PFL-AE reconstituted with 5 7F e and DTT in the
absence (A) and presence (B) of AdoMet ................................................................. 59
Figure IV.1 X-band EPR spectra of photoreduced PFL-AB before and after addition of
PFL. ........................................................................................................................... 69
Figure IV.2 Spin quantitation of the EPR spectra shown in Figure 1V.1A ([4Fe-4S]l+) and
1B (gly') as a function of illumination time. ............................................................ 70
Figure V.1 Release of ethanol-soluble material from methylated poly(dA) by AlkB in the
presence of Fe(II) and a-ketoglutarate. ..................................................................... 80
Figure V.2 Repair of l-methyladenine (l-meA) and 3-methylcytosine (3-meC) by AlkB.

................................................................................................................................... 81

vii

Figure V.3 Production of formaldehyde and consumption of Oz by AlkB ...................... 85

Figure V1.1 UV-visible spectroscopic analysis of the reaction of orKG-Fe(II)AlkB with

oxygen. ...................................................................................................................... 97
Figure V1.2 LC/MS/MS identification of hydroxy-Trp 178 in AlkB. .............................. 98
Figure V1.3 Structure of AlkB determined by protein modeling. ................................... 100

viii

List of Schemes

Scheme 1.1 Reaction catalyzed by PFL-AE and ARR-AE ................................................. 3
Scheme 1.2 Activation of PFL ............................................................................................. 4
Scheme 1.3 Formation of an adenosyl adduct ..................................................................... 4
Scheme 1.4 Structural comparison of AdoCbl and AdoMet ............................................... 5
Scheme 1.5 Reaction catalyzed by lysine aminomutase ..................................................... 9
Scheme 1.6 Reaction catalyzed by BioB ........................................................................... 11
Scheme 1.7 Proposed mechanism of spore photoproduct repair by SPL .......................... 12
Scheme 1.8 Sites of Methylation in DNA ......................................................................... 13
Scheme 1.9 1-methyladenosine and 3-methylcytosine ...................................................... 16
Scheme 1.10 Other Fe(II)/a-KG catalyzed reactions ....................................................... 18
Scheme 1.11 Fe(II)/a-KG Dioxygenase Mechanism ....................................................... 20

Scheme 1.12 Proposed Fe(II)/a-KG dependent AlkB mechanism: “oxidative DNA repair”

................................................................................................................................... 22
Scheme IV.1 Proposed redox role for the iron-sulfur cluster of PFL-AB ....................... 66
Scheme IV.2 Proposed Mechanism of PFL-AB .............................................................. 73
Scheme V.1 Oxidative DNA repair mechanism catalyzed by AlkB ................................ 87

ix

List of Tables

Temperature-Dependent Mdssbauer Parameters of the_Fe-S Clusters in Native PFL-AE _

.................................................................................................................................. .43

a-KG
BioB

1 -meA
3-meC
5’-dAdo
AdoCbl
AdoMet
CAS
CoA
DAOCS
DMS
DTT
EDTA
ENDOR
EPR

F eS
GC/MS
IPTG
LAM
LB
LMCT
MLCT
MMS
MNNG
NMNU
PFL
PFL-AB
PF PH
RN R
SDS-PAGE
TauD
deA
Tris
UV-vis

List of Abbreviations

a—ketoglutarate

biotin synthase

1 -methyladenine

3-methylcytosine

5’-deoxyadenosine

adenosylcobalamin
S-adenosylmethionine

clavaminate synthase

coenzyme-A

deacetoxycephalosporin C synthase
dimethylsulfate

dithiothreitol

ethylenediaminetetraacetic acid
electron-nuclear double resonance
electron paramagnetic resonance
iron-sulfur cluster

gas chromatography/mass spectrometry
isopropyl-B-D-thiogalactopyranoside
lysine aminomutase

Lauria Bertini media

ligand to metal charge transfer transition
metal to ligand charge transfer transition
methyl methanesulphonate
N—methyl-N-nitro-N-nitroguanidine
N-methyl-N-nitrosourea

pyruvate formate-lyase

pyruvate formate-lyase activating enzyme
pentafluorophenylhydrazine
ribonucleotide reductase

sodium dodecylsulfate polyacrylamine gel electrophoresis
taurine/OLKG dioxygenase
2,4-dichlorophenoxyacetic acid/aKG dioxygenase
tris(hydroxymethyl)aminomethane
ultraviolet-visible

xi

CHAPTER I

INTRODUCTION

Many crucial biological processes require steps that involve single electron chemistry.
This dissertation investigates two enzymes from Escherichia coli that catalyze reactions
utilizing single electron chemistry. The first, pyruvate formate-lyase activating enzyme
(PFL-AE, encoded by pflA) catalyzes the reductive cleavage of S-adenosylmethionine
(AdoMet) to generate a carbon radical intermediate. The second, the gene product of
alkB, is an iron and u-ketoglutarate dependent dioxygenase that initiates the repair of

alkylated DNA by a hydrogen atom abstraction step.

The Fe/S AdoMet enzymes

The classical role of AdoMet in biological systems is that of a methylating agent,
for example in the methylation of cytosine bases - a crucial mechanism of epigenetic
regulation (1). In recent years, there has been an explosion of interest in AdoMet
dependent enzymes that do not involve methylation (2). Rather, they utilize AdoMet and
an iron-sulfur (FeS) cluster to generate substrate and enzyme-based radicals. The
reactions accomplished by Fe-S/AdoMet enzymes include rearrangements, sulfur
insertion, DNA repair, and glycyl radical generation. I will describe the chemistry of
several of these enzymes, focusing on exciting recent developments in the generation of

5’-deoxyadenosyl radicals and the fascinating ways in which they are employed.

Pyruvate formate-lyase

Pyruvate formate-lyase (PF L) catalyzes the first committed step in the anaerobic

glucose catabolism by Escherichia coli; i.e. the reversible conversion of pyruvate and

2

coenzyme-A (CoA) to formate and acetyl-CoA . As isolated from aerobically growing
cells, PFL is inactive and contains no metals or organic cofactors but the enzyme may be
activated under strict anaerobic conditions. This activation is catalyzed by another
enzyme, pyruvate formate-lyase activating enzyme, and involves the introduction of a
glycyl radical at Gly-734 (3-6)(Scheme 1). This radical is absolutely required for PFL
activity and accounts for the extreme oxygen lability of the enzyme. Upon exposure to
oxygen, the radical is irreversibly quenched and the peptide chain is cleaved at the radical

site, leading to the formation of 3 kDa and 82 kDa protein fragments (4).

 

i i H .

‘ N. . N. ,

R an/ R PFL-AE R flJTII/ R
o 0

Scheme 1.1 Reaction catalyzed by PFL-AE and ARR-AE

The glycyl radical does not directly participate in catalysis. Rather, the radical is
transferred to one of two active site cysteine residue which attacks the carbonyl carbon of

pyruvate, displacing formate and generating an acyl-enzyme intermediate.

Pyruvate formate-lyase activating enzyme

The post-translational introduction of the glycyl radical into PFL requires PFL-
AE, AdoMet, and a source of electrons (Scheme 1.2) (7). AdoMet is reductively cleaved

to yield methionine and 5’deoxyadenosine (S’dAdo) (8).

PFL, PFL.

 

 

 

 

GlyH G'Y '
5'-deo adenosine
S-adenosyI-L-methionine > xy+
methionine
f fl F 1
PF L-AE PFL-AE
[4Fe-4S]1+ [4Fe-4312“
L J L J

 

 

 

 

 

 

Scheme 1.2 Activation of PFL

1n important early experiments, Knappe and coworkers labeled Gly-734 with
deuterium and showed that the label was incorporated into 5’-deoxyadenosine, thus
implicating a 5’-deoxyadenosyl radical as the hydrogen atom abstracting species (9).
More recently, the same laboratory used a PFL peptide analog that contained
dehydroalanine in place of Gly-734 as a substrate for PFL-AE (10). Using mass
spectrometry, the authors characterized a 5’-deoxyadenosyl adduct of the peptide where
the adenosyl moiety was linked to the B-carbon of the modified amino acid, thereby
providing further evidence for the involvement of a 5’deoxyadenosyl radical in the

activation of PFL (10). (Scheme 1.3)

 

O A
CH2 AdoMet methionine HZC/w
H O PFL-AE H 0

Scheme 1.3 Formation of an adenosyl adduct

Deoxyadenosyl radicals had long been proposed as intermediates in
adenosylcobalamin-dependent reactions, including rearrangements and nucleoside
reduction. Enzymes generate 5’-deoxyadenosyl radicals from AdoCbl by destabilizing
the weak carbon-cobalt bond and promoting its homolysis. AdoMet is a much simpler
molecule with no metal cofactors (Scheme 1.4) and it was unclear how a 5’-
deoxyadenosyl radical might be generated, prompting Perry Frey to ask the question, “Is
adenosylmethionine a poor man’s adenosylcobalamin?” (11) A significant portion of this
dissertation addresses that question. It is interesting to note that Professor Frey has
changed his rhetoric regarding the radical generating role of AdoMet. Ten years after his
previous query, he recently asked the question “S-adenosylmethionine: a wolf in sheep’s
clothing or a rich man’s adenosylcobalamin” (2) as a reflection of the rapid pace of

progress in this field.

OI-DH

 

Hci‘

Scheme 1.4 Structural comparison of AdoCbl and AdoMet

Early evidence suggested that an organic cofactor (7) or mononuclear metal site

(12) plays a role in PFL-AE catalyzed radical generation. The correct line of
5

investigation was found when an iron-sulfur cluster was identified in the enzyme (13).
Broderick and coworkers used a combination of resonance Raman, ultraviolet-visible,
and variable-temperature magnetic circular dichroism spectroscopies to identify a
mixture of [4Fe-4S]2+ and [2Fe-28]2+ clusters in the as-isolated enzyme. Reduction with
dithionite yielded entirely cuboidal [4Fe-4S] clusters, with the more reduced and electron
paramagnetic resonance active 1+ cluster oxidation state reached only in the presence of
AdoMet (13).

Knappe and coworkers also identified an F eS cluster in PFL-AB that had been
artificially reconstituted with iron and sulfide (14). They observed an axial EPR signal
consistent with the presence of an [4Fe-4S]l+ cluster that underwent a shifi to a rhombic
Signal upon AdoMet binding. Significantly, using site-directed mutagenesis, they
identified three cysteine residues that were critical for cluster reconstitution and
enzymatic activity.

The identification of an FeS cluster and the suggestion that AdoMet interacts with
the cluster led to the proposal that the cluster could serve a redox role in the AE reaction
(13). The cleavage of AdoMet to methionine and 5’-deoxyadenosine is a two-electron
reduction. A hydrogen atom is abstracted from Gly-734, accounting for one electron. A
reasonable proposal is that the cluster serves as the source of the additional electron, as
shown in Scheme 1.2. One aim of this dissertation is to test this mechanism and
determine the active form of the cluster.

Another interesting question is the identity of the fourth ligand to the cluster. The
C-X3-C-X2-C motif is conserved in all FeS/AdoMet enzymes (15). An interesting

proposal is that AdoMet itself is the fourth ligand in the active form of the enzyme. The

observations that AdoMet seemed to alter the reduction potential or the anisotropy and
the cluster electronic environment suggest an interaction between these species, but they
do not necessarily indicate a direct interaction. A precise determination of the
cluster/AdoMet interaction would greatly enhance our understanding of the mechanism

of glycyl radical generation.

Ribonucleotide reductase activating enzyme

The reduction of ribonucleotides to deoxyribonucleotides is the rate-determining
step in DNA synthesis (16). All three classes of ribonucleotide reductases utilize a
radical mechanism in the reduction of the sugar, but vary in the mechanism by which the
radical is generated. The class I enzymes use a diiron center and dioxygen as a radical
generating mechanism. Class II enzymes use AdoCbl. Class III reductases function
under anaerobic conditions and utilize AdoMet to generate a stable glycyl radical in a
manner directly analogous to the PFL system (Scheme 1.1) (17-19). A small subunit, also
termed the activating enzyme or [3 subunit, generates the radical that resides on Gly-681
of the larger or a subunit. The glycyl radical reacts with the active site cysteine that in
turn abstracts a hydrogen atom from substrate to initiate reduction of the sugar at the 5’-
carbon.

As in the case of the PFL system, AdoMet and a source of electrons is required to
generate the glycyl radical on the a subunit (17). Like PF L-AE, the 0 subunit also
contains a [4Fe-4S] cluster that can access both the 2+ and 1+ oxidation states (20, 21).
Thus, it has been hypothesized that the [4Fe-4S]1+ cluster provides the electron needed

for the reductive cleavage of AdoMet, allowing subsequent activation of the

7

ribonucleotide reductase. Fontecave and coworkers correlated the production of
methionine with the oxidation of the [4Fe-4S]1+ cluster and determined similar first order
rate constants for both processes (22). They initially determined a stoichiometry of two
methionines produced for each cluster oxidized, a ratio that is inconsistent with the
reaction scheme. More recent work by the same lab has resolved the stoichiometry with
regard to the activation, but they also observed reductive AdoMet cleavage catalyzed by
the B subunit in the absence of the a subunit (23). This side reaction produced two
methionines for each cluster that was oxidized. An unprecedented mechanism was
proposed that invokes a “hyper-reduced” [3Fe-4S]2' cluster which could donate two
electrons to reduce two equivalents of AdoMet, thus accounting for the observed
methionine production. Experimental evidence to support this novel redox state has not

been reported.

Lysine aminomutase

Lysine 2,3 -aminomutase (LAM) catalyzes the reversible conversion of L-lysine
and L-B-lysine, a reaction that is highly reminiscent of adenosylcobalamin dependent
rearrangements (2)(Scheme 1.5), but LAM activity is strictly dependent on AdoMet, not
AdoCbl (24). When [adenosyl-5’-3H]AdoMet is used, the radioactive label is
incorporated into both L-lysine and L-B-lysine (25). This result shows that AdoMet
mediates hydrogen exchange in a manner directly analogous to the AdoCbl systems and

strongly suggests that a 5’-deoxyadenosyl radical participates in this process.

H guns»: *H (\H
+H3NWCOO- = ”WA/53V“)-
H . LAM

H +H3N H

Scheme 1.5 Reaction catalyzed by lysine aminomutase

Like other family members, LAM may accommodate a variety of cluster types.
Of the various clusters and redox states that have been identified, Frey and coworkers
have shown that the [4F e-4S] H cluster participates in the catalytic cycle (26, 27). Recent
creative work has provided the first direct spectroscopic evidence for a 5’-deoxyadenosyl
radical in the AdoMet dependent systems. AnhydroAdoMet was used to stabilize and
allow characteriation of a 5’ally1ic radical in the LAM reaction (28, 29).

Recent work also has shed light on the interaction between AdoMet and the
cluster in LAM (30). Frey and coworkers used selenoAdoMet (which contains selenium
instead of sulfur) and were able to detect a Fe-Se interaction at 2.7 A using X-ray
absorption spectroscopy. This scattering was only observable in samples that contained
the substrate analog trans-4,5-dehydrolysine, which traps the enzymatic reaction at an
intermediate afier cleavage of AdoMet. The new FT peak is then a reflection of
methionine, a cleavage product of AdoMet, bound to the cluster (30). In addition to the
information gained from X-ray spectroscopy, electron-nuclear double resonance
(ENDOR) spectroscopy has been used to probe the interaction between AdoMet and the
cluster (31). AdoMet was specifically labeled with 17O at the cabroxylate group and 151N
at the amino group. The results show that AdoMet chelates an iron of the cluster through
the amino acid groups to form a five-membered ring. (Similar experiments have been
done with the iron-sulfur cluster of PFL-AB and will be introduced in the final chapter

(32, 33).)

Sulfur insertion: BioB and LipA

BioB catalyzes the insertion of a sulfur atom across two unactivated C-H bonds of
dethiobiotin in the final biosynthetic step of biotin, an essential vitamin that is only
produced by plants and certain microorganisms (for reviews see (2, 34)) (Scheme 1.6).
Like other family members, BioB is able to accommodate a wide variety of cluster types,
however BioB binds two distinct clusters simultaneously (35). Jarrett and coworkers
used specific 57Fe labeling and Mossbauer spectroscopy to identify two cluster types in
BioB: a [4Fe-4S]2+ and a [2Fe-2S]2+ cluster (36). In addition, they were able to show
that the [2Fe-28] cluster was degraded during turnover while the [4Fe-4S] cluster
remained intact (37). From these results they postulate that the [2Fe-ZS] cluster serves as
the sulfur source for the reaction, while the [4Fe-4S] cluster catalyzed the reductive
cleavage of AdoMet to generate 5’-deoxyadenosyl radicals that activate the C-H bonds of
dethiobiotin. The Fontecave group has detected a PLP-dependent cysteine desulfurase
activity catalyzed by BioB and has shown that this activity is dependent on two (non-
cluster ligand) cysteine residues (38). They postulate that the cysteine-derived sulfide
forms a persulfide at the required cysteines and is then incorporated into dethiobiotin. It
is possible that the results of both groups are correct and self-consistent. If the [2Fe-ZS]
cluster is degraded during turnover to provide the sulfur source, a regeneration
mechanism (perhaps in the form of a cysteine desulfurase activity) would allow for

multiple turnovers, a feat that has not yet been detected in vitro.

10

HN NH 0

Hm BioB ”w
H30 0-
Scheme 1.6 Reaction catalyzed by BioB

Recent exciting results include the determination of the crystal structures of BioB
(39) and HemN (40), another FeS/AdoMet enzyme. These will be discussed in the

context of the results presented in this dissertation in Chapters IV and VII.

Spore photoproduct lyase

In addition to the reactions already discussed, F e/ S/AdoMet enzymes also
participate in DNA repair. Spores of Bacillus subtilis form specific thiamine dimers upon
exposure to UV radiation, which are repaired by the spore photoproduct lyase (SPL) in a
reaction that is dependent on AdoMet (41-43). (Scheme 1.7). A [4Fe-4S] cluster has been
identified in SPL and it has been reported that AdoMet undergoes reductive cleavage to
methionine and 5’-deoxyadenosine (44). In contrast, later experiments have
demonstrated that AdoMet functions as a cofactor rather than a cosubstrate in SPL (45).
Check and Broderick generated photoproduct-containing DNA that was labeled with
tritium at either the C—6 or methyl positions. When substrate was labeled at 06,
significant radioactivity was incorporated into AdoMet, but not when substrate was
labeled at the methyl group (45). The results support the mechanism first suggested by
Mehl and Begley (46) as shown in Scheme 1.7. AdoMet is reduced by the Fe/S cluster to

generate a 5’dAdo radical which abstracts a hydrogen atom from C-6 of the photoproduct

11

dimer. The substrate radical then undergoes B-scission and the abstracts a hydrogen atom
from 5’dAdo to regenerate the S’dAdo radical which recombined with methionine to

regenerate AdoMet.

.-.P.'§‘3°*..

 

 

 

 

Scheme 1.7 Proposed mechanism of spore photoproduct repair by SPL

Reflecting the diversity of chemistry that has already been demonstrated, Sophia
et al. used PSI-BLAST and other bioinforrnatic tools to identify over 500 unique
sequences that are thought to encode FeS/AdoMet enzymes (15). Some of the enzymes
they identified had been characterized by biochemical methods, but most had not yet

been identified as members of the FeS/AdoMet family.

‘ DNA repair and Fe(II)/a-ketoglutarate dioxygenases
DNA is constantly under attack. The molecule that carries our genetic

information can be damaged or modified by oxidation, radiation, or toxins, as well as by

12

some endogenous biomolecules (47). To counter this constant threat to the security of
the genetic information, elaborate and diverse mechanisms have evolved to reverse the
DNA damage. Three general strategies for DNA repair are base excision, nucleotide
excision, or direct repair. In base excision repair, the glycosidic bond is hydrolyzed by a
glycolyase, removing the nucleoside base and resulting in an abasic site. Specialized
endonucleases then must remove the remaining sugar phosphate and a DNA polymerase
fills in the resulting gap. In nucleotide excision repair, a variable DNA segment that

contains the lesion is removed and replaced.

The Adaptive response to alkylating agents

Alkylating agents are a broad class of toxins that have been shown to target DNA
and other biomolecules. These methylating agents have been broadly separated into SNI
and the SN2 categories, based on the mechanism of nucleophilic substitution reaction they
employ to alkylate the DNA base. As shown in Scheme 1.8, a plethora of sites in the four

bases are subject to damage from of methylating agents (48).

l I (I’m

NH2\ rig]

</ \iN.......... ...... ...HN \ N\ / NHN/ \ N\
N \\ >—_‘N dR dR

dR/ OIIIIHHHIIIIIHII HZN dRI (T)

Scheme 1.8 Sites of Methylation in DNA

Escherichia coli has developed an elegant response to the stress caused by exposure to
alkylating agents that is described as both the “adaptive response to alkylating agents”
and the “Ada response,” and includes four genes: ada, alkB, alkA, and aidB. The
proteins encoded by these genes serve differing roles that all seem to protect the cell from
acute exposure to alkylating agents.

Ada is a fascinating trifunctional protein. It activates the transcription of the
adaptive response and also repairs two types alkylated DNA lesions. The C-terminal
domain is responsible for one type of DNA repair activity. Alkyl groups from the
mutagenic lesions 6-methylguanine and 4-methylthymine are directly transferred to a
specific nucleophilic cysteine residue (Cys-321) (49). After completing one DNA repair
turnover the C-terminal domain is rendered inactive and de novo protein synthesis is
required to provide additional resistance. The N-terminal domain of Ada similarly
exhibits methyltransferase actitivity, but also serves as the transcriptional regulator of the
adaptive response. A methyl group from the phosphotriester backbone is transferred to
Cys-3 8, made possible by the coordination of a Zn2+ ion which enhances the
nucleophilicity of the cysteine thiol (50, 51). Upon methylation of the N-terminal
domain, the binding of Ada to the promoter regions of the ada-alkB operon and the alkA
and aidB genes is enhanced, thus promoting expression of the adaptive response.

AlkA was quickly identified as a 3-methyladenosine glycosylase (52, 53). The
function of AidB remains a mystery, but it may function by metabolizing SNl alkylating
agents, decomposing modified nucleotide precursors, or reversing damage to DNA .
These hypotheses are currently being tested in our lab. It is interesting that in addition to

these activities encoded to genes in the Ada response, E. coli has alkylation defense genes

14

that are constitutively expressed, suggesting that the cells are exposed to alkylating
agents in both persistent and acute mechanisms. For example, ogt encodes a 06-
methylguanine methyltransferase (54) and tag encodes a 3-methyladenosine glycosoylase
(55).

The protein encoded by alkB has been an enigma for two decades. In 1983 alkB
mutants were isolated that were defective in their ability to reactivate methyl
methanesulfonate-treated lambda phage (56). The gene was cloned (57) and the protein
was purified without clarification of its role (58). AlkB was shown to protect cells from
the toxic effects of the 81.12 alkylating agents dimethylsulfate (DMS) and methyliodide
(Mel), but is less effective in protection against the SN] reagents N-methyl-N-nitro-N-
nitroguanidine (MNN G) and N-methyl-N-nitrosourea (NMNU). AlkB mutants are not
defective in repairing 3-methyladenine, DNA strand breaks, abasic sites, and other
lesions that may arise at abasic sites such as DNA—protein cross-links and DNA
interstrand cross-links (59). Expression of the E. coli gene in human cells afforded
Similar resistance to alkylating agents, suggesting that AlkB functioned in the absence of
other factors (60). A major piece of the AlkB puzzle was put in place when Sedgewick
and coworkers found that alkB mutants were defective in reactivating single-stranded
lambda phage that had been treated with SN2 alkylating agents (61). There was very little
defect in the mutants’ ability to reactivate double stranded phage that had been treated in
the same manner, and the effect of SN2 methylating agents was much greater than that of
SNl agents. These results strongly suggest that AlkB repairs DNA lesions created in
single stranded DNA. This may be because AlkB shows preference toward single-

stranded substrate or it may repair lesions that are formed more readily in single- rather

15

than double-stranded DNA. Two prospective lesions that are formed primarily in single-
stranded DNA were identified as potential substrates: l-methyladenosine and 3-
methylcytosine (Scheme 1.9) (48, 61). The nitrogens that are methylated in these lesions

are protected by Watson-Crick base pairing in double-stranded DNA.

NH2 NH2

, ,CH

«"fl CH3 [1 3

l‘ ~“ 'I‘ 0
dR dR

Scheme 1.9 l-methyladenosine and 3-methylcytosine

Recent in silico work also has contributed to our understanding of AlkB. Aravinid and
Koonin used structural prediction tools to show that that AlkB likely adopts a fold that is

characteristic of a-ketoglutarate and iron dependent dioxygenases (62).

F e(II)/a-ketoglutatarate Dioxygenases

The enzymes that make up the iron and a—ketoglutarate (Fe(II)/aKG) dependent
dioxygenase superfamily catalyze a wide variety of oxidations. Each enzyme couples the
reduction of dioxygen to the oxidative decarboxylation of a-ketoglutarate to generate
high-valent iron-0x0 intermediates that are responsible for the oxidation of the primary
substrate. The following paragraphs briefly summarize the critical features of this

enzyme family before returning to AlkB.

16

Amino Acid Hydroxylases

Prolyl 4-hydroxylase, a representative Fe(II)/a-KG dioxygenase, catalyzes the
hydroxylation of proline within many structural and transcriptional proteins. In type I
collagen, these hydroxylated proline residues stabilize the triple helical structure of the
protein, accounting for its structural elasticity (63). Although studied for decades,
interest in this sidechain hydroxylase has intensified with the discovery that the hypoxia
response is regulated by an Fe(II)/a-KG dioxygenase. The hypoxia inducible factor is a
dimeric (HIFa and HIFB) transcription factor that is constitutively expressed and
activates the transcription of a variety of genes (reviewed in (64)). Under normal oxygen
conditions, Pro-564 of HIFa is hydroxylated by a prolyl 4-hydroxylase, targeting the
transcription factor for ubiquitinylation and degradation, thus suppressing transcription
(65-68). In an additional mechanism, hydroxylation of Asn 803 of HIFu prevents its
interaction with p300, a coactivator that is required for HIP-dependent transcriptional
activation. The enzyme responsible for this hydroxylation has been identified as an a-
KG dioxygenase (69). The medical significance in these systems is noteworthy, as
hypoxia plays a significant role in a variety of disease processes, including tumor growth,
diabetes, and other ischemic diseases (64).

Although the work presented in Chapters V and V1 concerns an a-KG dependent
hydroxylase, other Fe(II)/a-KG family members catalyze remarkably varied chemistry.
For example, clavaminate synthase (CAS) catalyzes three oxidative steps in the synthesis
of clavulanic acid: a hydroxylation, an oxidative ring closure, and a desaturation, all of
which consume 02 and a-KG (Scheme 1.10) (70). Deacetoxycephalosporin C synthase

(DAOCS) catalyzes both a ring expansion and a hydroxylation (Scheme 1.10) (71-73).

17

Thymine hydroxylase catalyzes three successive oxidations of the methyl group of

thymine, from the alcohol to the aldehyde to the acid (Scheme 1.10) (74).

aKG succinate

R EH 0, co2 R OH 51H
0 WN NH2 3.4 0 Wu NH2

COOH H CA3 COOH

aKG succinate aKG succinate

I] OH 0 co2 o O
N 2 l ' o2 co2
o YK/‘NHZE i o N ”bx—4 01 N' \ NH
COOH CAS COOH C A S COOH
R—N S \ aKG succinate H aKG succinate /N
If 02 002 R’ oz co; R
0 NRA :4 N / g 4 o N / OH
COOH vocs 0 COOH DAOCS COOH

O KG ' I O O O
CHO. SUCCInae
w 3 02 002 HN I CHZOH HN)‘]/CHO HNJlj/coori

O H Thymine H 0 N
hydroxylase H

Scheme 1.10 Other Fe(II)/a-KG catalyzed reactions

F e(II)/a-KG Dioxygenase Mechanism

An abundance of spectroscopic, structural and kinetic studies have dramatically

improved our understanding of these enzymes in recent years (for reviews see (75, 76))

and have unveiled a number of reoccurring themes. From structural and spectroscopic

studies, a clear picture of the first few steps of the catalytic cycle is clear (Scheme 1.11).

In the resting state of the enzyme, the Fe(II) is coordinated by two histidines and one

acidic residue, and the remaining three sites in the distorted octahedral environment are

18

occupied by water. a-Ketoglutarate chelates the iron through the C-1 carboxylate and C-
2 ketone oxygens, displacing two waters, while the water in the sixth coordination site
remains bound. The primary substrate binds to the enzyme, but not directly to the iron,
causing a shift in the coordination environment as water disassociates, thus opening a
coordination site for oxygen binding and activation. This communication via the protein
matrix serves a critical regulatory role, since initiation of oxidative chemistry in the
absence of primary substrate is destructive to the enzyme (77, 78) (Chapter VI).
Significant strides have been made towards understanding the later steps of the cycle,
although some steps remain uncertain. Oxygen binds to the iron, most likely at the open
coordination site, and may be reduced to either the superoxo or peroxo state. The oxygen
then attacks the carbonyl carbon of a-KG to yield an alkylperoxo intermediate which
decomposes to release C02 and generate an Fe(IV)—oxo intermediate. This high-valent
iron species abstracts a hydrogen atom from the substrate, resulting in an Fe(III)-
hydroxide and a substrate radical. Recombination of the substrate radical with the
hydroxide restores the Fe(II) state of the enzyme and results in the hydroxylated

substrate.

19

“3‘0

0H2 a-KG 0H2 substrate

Aspmm'l .\\OH2 Asphhhl .nO O Aspu ("I \O O

./Fe(”Q /Fe(||) / """ Fe”

Hls | OH; His | ‘0 His/ | \0’
His 2H20 His

H 0 His
COOH 2 COOH
02
succinate, product 1
OH COJ
ASP’hn I"). .... “0 ASP’nu-geflY‘lg
"'Fe / ' "" . /
His/l \o HIS | ‘o)\\\
”‘5 COOH “'3 COOH

Scheme [.11 Fe(II)/a-KG Dioxygenase Mechanism

Non-heme Models

The inorganic community has provided model systems that mimic the
coordination environment and reactivity of Fe(II)/a-KG dioxygenases. Specifically, Que
and coworkers have demonstrated ligand hydroxylation in an Fe(II)/a-keto acid complex
upon exposure to dioxygen, an activity that directly mimics the u-KG hydroxylases (79).
An exciting recent result is the crystallographic characterization of an Fe(IV)-oxo
complex in a non-heme mononuclear environment (80). This proves that the long-
postulated high valent intermediate is a viable possibility and provides an important

spectroscopic point of reference for comparison to enzyme systems.

Identification of an Fe (IV) =0 Intermediate

A likely Fe(IV) intermediate has also been identified in the cycle catalyzed by

TauD. Bollinger, Krebs and coworkers used freeze-quench techniques to trap

20

intermediates and characterized them by Mossbauer and EPR spectroscopies (81). The
Mossbauer parameters were consistent with a formal Fe(IV) oxidation state. The
intermediate exhibited an integer spin state with S 2 2 and cryoreduction by gamma
irradiation led to an EPR signal typical of a high-spin Fe(III) S = 5/2 ground state. Taken
together, these data provide strong evidence for the participation of an Fe(IV)
intermediate. In an additional report, the same researchers determined that the rate of
decay of this intermediate exhibits a 37-fold decrease when deuterated taurine is used
(82). This substantial kinetic isotope effect indicates that the Fe(IV) intermediate is
responsible for direct hydrogen atom abstraction from the primary substrate.

In a crucial extension of this work, our laboratory has used isotope difference
time-resolved electronic absorption and resonance Raman (rR) spectroscopy to probe the
nature of this intermediate (83). The experimental apparatus used a continuous flow
technique that allows rR excitement at a selected delay time. When the chromophore
associated with the Fe(IV) intermediate is excited, the dominant oxygen dependent mode
is observed at 821 cm'1 and shifts to 787 cm'l when 1802 is used in place of 1602. The
position and isotope shift of this mode compares favorably with the well characterized
Fe(IV)-oxo intermediate in heme oxygenases. We conclude that this mode most likely
arises from the symmetric stretch of a formally F e(IV)-oxo intermediate which abstracts a
hydrogen atom from taurine. Additional modes were identified at 555 cm'l and 859 cm'l
which were not assigned, but could arise from either a Fe(II)-02 or Fe(III)-superoxo,
intermediates that are proposed to precede formation of the F e(IV)-oxo. Experiments to

discriminate among these possibilities are currently in progress.

21

AlkB as an F e(II)/a-ketoglutarate Dioxygenase

The observation that AlkB repairs methylated lesions in DNA and the possibility
that it may adopt the fold of an Fe(II)/a-KG dioxygenase led us to hypothesize that it may
repair methylated DNA by hydroxylating the aberrant methyl group (Scheme 1. 12). This
hydroxymethyl intermediate would spontaneously decompose to yield formaldehyde and
the unmodified base in an oxidative N-demethylation mechanism. If correct, this
mechanism would explain the experimental observations that had been obtained over two
decades. Such a mechanism has precedent in heme systems and would account for
previous unsuccessful attempts to identify an enzymatic activity for AlkB, as iron and u-
ketoglutarate are not common additions to enzyme activity screens. In fact, such a
mechanism would be anathema to the DNA repair community, as one most ofien thinks
of “oxidative DNA damage,” not “oxidative DNA repair.” Chapter V presents

experiments designed to test this mechanism.

NH2 NH2 0 o NH2
, H2 H
/N \+N ’ CH3 </N I \+N [N \ N
< | . A < |
J aKG succmate N N N //I
'1' N oz co I N

2 dR l
dR \ / dR
NH2 “”2 \ NH2

Vi ' Us I J:
N 0 'fl 0 N o
dR

5R dR

Scheme 1.12 Proposed Fe(II)/a-KG dependent AlkB mechanism: “oxidative DNA
repair”

22

10.

11.

12.

13.

14.

References

Bird, A., Genes Dev, 2002, 16, 6—21.
Frey, P. A. and Magnusson, O. T., Chemical Reviews, 2003, 103, 2129-2148.

Knappe, 1., Schacht, 1., Mockel, W., Hopner, T., Vetter, H., 1r., and Edenharder,
R., Eur J Biochem, 1969, 11, 316-27.

Wagner, A. F. V., Frey, M., Neugebauer, F. A., Schafer, W., and Knappe, 1.,
Proceedings of the National Academy of Sciences of the United States of America,
1992, 89, 996-1000.

Knappe, 1., Neugebauer, F. A., Blaschkowski, H. P., and Ganzler, M.,
Proceedings of the National Academy of Sciences of the United States of America-
Biological Sciences, 1984, 81, 1332-1335.

Unkrig, V., Neugebauer, F. A., and Knappe, 1., European Journal of
Biochemistry, 1989, 184, 723-728.

Conradt, H., Hohmannberger, M., Hohmann, H. P., Blaschkowski, H. P., and
Knappe, 1., Archives of Biochemistry and Biophysics, 1984, 228, 133-142.

Knappe, 1. and Schmitt, T., Biochemical and Biophysical Research
Communications, 1976, 71, 1110-1117.

Frey, M., Rothe, M., Wagner, A. F. V., and Knappe, 1., Journal of Biological
Chemistry, 1994, 269, 12432-12437.

Wagner, A. F. V., Demand, 1., Schilling, G., Pils, T., and Knappe, 1., Biochemical
and Biophysical Research Communications, 1999, 254, 306-310.

Frey, P. A., Faseb Journal, 1993, 7, 662-670.

Wong, K. K., Murray, B. W., Lewisch, S. A., Baxter, M. K., Ridky, T. W.,
Ulissidemario, L., and Kozarich, 1. W., Biochemistry, 1993, 32, 14102-14110.

Broderick, 1. B., Duderstadt, R. E., Fernandez, D. C., Wojtuszewski, K.,
Henshaw, T. F., and Johnson, M. K., Journal of the American Chemical Society,
1997, 1 19, 73 96-73 97.

Kulzer, R., Pils, T., Kappl, R., Huttermann, 1., and Knappe, 1., Journal of
Biological Chemistry, 1998, 273, 4897-4903.

23

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

Sofia, H. 1., Chen, G., Hetzler, B. G., Reyes-Spindola, 1. F., and Miller, N. E.,
Nucleic Acids Res, 2001, 29, 1097-106.

Stubbe, 1. and van der Donk, W. A., Chemical Reviews, 1998, 98, 705-762.

Harder, 1., Eliasson, R., Pontis, E., Ballinger, M. D., and Reichard, P., Journal of
Biological Chemistry, 1992, 267, 25548-25552.

Bianchi, V., Eliasson, R., Fontecave, M., Mulliez, E., Hoover, D. M., Matthews,
R. G., and Reichard, P., Biochemical and Biophysical Research Communications,
1993, 197, 792-797.

Sun, X., Eliasson, R., Pontis, E., Andersson, 1., Buist, G., Sjoberg, B.-M., and
Reichard, P., J. Biol. Chem, 1995, 270, 2443-2446.

Mulliez, E., Fontecave, M., Gaillard, 1., and Reichard, B, Journal of Biological
Chemistry, 1993, 268, 2296-2299.

Ollagnier, S., Mulliez, E., Gaillard, J., Eliasson, R., Fontecave, M., and Reichard,
P., Journal of Biological Chemistry, 1996, 271, 9410-9416.

Ollagnier, S., Mulliez, E., Schmidt, P. P., Eliasson, R., Gaillard, 1., Deronzier, C.,
Bergman, T., Graslund, A., Reichard, P., and Fontecave, M., Journal of
Biological Chemistry, 1997, 272, 24216-24223.

Padovani, D., Thomas, F., Trautwein, A. X., Mulliez, E., and Fontecave, M.,
Biochemistry, 2001, 40, 6713-6719.

Chirpich, T. P., Zappia, V., Costilow, R. N., and Barker, H. A., J. Biol. Chem,
1970, 245, 1778-1789.

Moss, M. and Frey, P. A., Journal of Biological Chemistry, 1987, 262, 14859-
14862.

Petrovich, R. M., Ruzicka, F. 1., Reed, G. H., and Frey, P. A., Biochemistry, 1992,
31,10774-10781.

Lieder, K. W., Booker, S., Ruzicka, F. 1., Beinert, H., Reed, G. H., and Frey, P.
A., Biochemistry, 1998, 37, 2578-2585.

Magnusson, O. T., Reed, G. H., and Frey, P. A., Journal of the American
Chemical Society, 1999, 121, 9764-9765.

Magnusson, O. T., Reed, G. H., and Frey, P. A., Biochemistry, 2001, 40, 7773-
7782.

Cosper, N. 1., Booker, S. J ., Ruzicka, F., Frey, P. A., and Scott, R. A.,
Biochemistry, 2000, 39, 15668-15673.

24

31.

32.

33.

34.

35.
36.

37.
38.

39.

40.

41.

42.

43.

44.

45.
46.
47.

Chen, D. W., Walsby, C., Hoffman, B. M., and Frey, P. A., Journal of the
American Chemical Society, 2003, 125, 11788-11789.

Walsby, C. 1., Hong, W., Broderick, W. E., Check, 1., Ortillo, D., Broderick, 1. B.,
and Hoffman, B. M., Journal of the American Chemical Society, 2002, 124, 3143-
3151.

Walsby, C. 1., Ortillo, D., Broderick, W. E., Broderick, J. B., and Hoffinan, B. M.,
Journal of the American Chemical Society, 2002, 124, 11270-11271.

Fontecave, M., Ollagnier-de-Choudens, S., and Mulliez, E., Chem Rev, 2003, 103,
2149-66.

Ugulava, N. B., Gibney, B. R., and Jarrett, J. T., Biochemistry, 2001, 40, 8343-51.

Ugulava, N. B., Surerus, K. K., and Jarrett, J. T., J Am Chem Soc, 2002, 124,
9050-1.

Ugulava, N. B., Sacanell, C. 1., and Jarrett, 1. T., Biochemistry, 2001, 40, 8352-8.

Ollagnier-De-Choudens, S., Mulliez, E., Hewitson, K. S., and Fontecave, M.,
Biochemistry, 2002, 41, 9145-52.

Berkovitch, F., Nicolet, Y., Wan, J. T., Jarrett, J. T., and Drennan, C. L., Science,
2004, 303, 76-9.

Layer, G., Moser, 1., Heinz, D. W., Jahn, D., and Schubert, W. D., Embo J, 2003,
22, 6214-24.

Varghese, A. 1., Biochemical and Biophysical Research Communications, 1970,
38, 484-&.

Munakata, N. and Rupert, C. 8., Molecular & General Genetics, 1974, 130, 239-
250.

Rebeil, R., Sun, Y. B., Chooback, L., Pedraza-Reyes, M., Kinsland, C., Begley, T.
P., and Nicholson, W. L., Journal of Bacteriology, 1998, 180, 4879-4885.

Rebeil, R. and Nicholson, W. L., Proceedings of the National Academy of
Sciences of the United States of A merica, 2001, 98, 9038-9043.

Check, 1. and Broderick, 1. B., J Am Chem Soc, 2002, 124, 2860-1.
Mehl, R. A. and Begley, T. P., Organic Letters, 1999, I, 1065-1066.
Wood RD, M. M., Sgouros J, Lindahl T, Science, 2001, 291, 1284-1289.

25

48.

49.

50.
51.

52.

53.

54.
55.
56.

57.

58.

59.

60.

61.

62.
63.

64.

65.

Singer, B. and D., G., Molecular Biology of Mutagens and Carcinogens:
Reactions of Directly Acting Agents with Nucleic Acids. 1983, New York: Plenum
Press.

Demple, B., Sedgwick, B., Robins, P., Totty, N., Waterfield, M. D., and Lindahl,
T., Proc Natl Acad Sci U S A, 1985, 82, 2688-92.

He, C., Wei, H., and Verdine, G. L., J Am Chem Soc, 2003, 125, 1450-1.

Sedgwick, B., Robins, P., Totty, N., and Lindahl, T., J Biol Chem, 1988, 263,
4430-3.

Nakabeppu, Y., Kondo, H., and Sekiguchi, M., J Biol Chem, 1984, 259, 13723-9.

Nakabeppu, Y., Miyata, T., Kondo, H., Iwanaga, S., and Sekiguchi, M., J Biol
Chem, 1984, 259, 13730-6.

Margison, G. P., Cooper, D. P., and Potter, P. M., Mutat Res, 1990, 233, 15-21.
Steinum, A. L. and Seeberg, E., Nucleic Acids Res, 1986, 14, 3763-72.

Kataoka, H., Yamamoto, Y., and Sekiguchi, M., Journal of Bacteriology, 1983,
153,1301-1307.

Kataoka, H. and Sekiguchi, M., Molecular and General Genetics, 1985, I98, 263-
269.

Kondo, H., Nakabeppu, Y., Kataoka, 11., Kuhara, S., Kawabata, S., and
Sekiguchi, M., Journal of Biological Chemistry, 1986, 261, 15772-15777.

Dinglay, S., Gold, 3., and Sedgwick, B., Mutation Research-DNA Repair, 1998,
407, 109-116.

Chen, B. R. 1., Carroll, P., and Samson, L., Journal of Bacteriology, 1994, 1 76,
6255-6261.

Dinglay, S., Trewick, S. C., Lindahl, T., and Sedgwick, 3., Genes &
Development, 2000, 14, 2097-2105.

Aravind, L. and Koonin, E. V., Genome Biol, 2001, 2, RESEARCH0007.

Kivirikko, K. I. and Pihlajaniemi, T., Adv Enzymol Relat Areas Mol Biol, 1998,
72, 325-98.

Hewitson, K. S., McNeil], L. A., and Schofield, C. 1., Current Pharmaceutical
Design, 2004, 10, 821-833.

Bruick, R. K. and McKnight, S. L., Science, 2001, 294, 1337-1340.

26

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.
77.

78.

Epstein, A. C. R., Gleadle, J. M., McNeill, L. A., Hewitson, K. S., O'Rourke, J.,
Mole, D. R., Mukherji, M., Metzen, E., Wilson, M. 1., Dhanda, A., Tian, Y. M.,
Masson, N., Hamilton, D. L., Jaakkola, P., Barstead, R., Hodgkin, J., Maxwell, P.
H., Pugh, C. W., Schofield, C. J., and Ratcliffe, P. J., Cell, 2001, 107, 43-54.

Ivan, M., Kondo, K., Yang, H. F ., Kim, W., Valiando, 1., Ohh, M., Salic, A.,
Asara, 1. M., Lane, W. S., and Kaelin, W. G., Science, 2001, 292, 464-468.

Jaakkola, P., Mole, D. R., Tian, Y. M., Wilson, M. 1., Gielbert, J., Gaskell, S. 1.,
von Kriegsheim, A., Hebestreit, H. F., Mukherji, M., Schofield, C. J., Maxwell, P.
H., Pugh, C. W., and Ratcliffe, P. 1., Science, 2001, 292, 468-472.

Hewitson, K. S., McNeill, L. A., Riordan, M. V., Tian, Y. M., Bullock, A. N.,
Welford, R. W., Elkins, 1. M., Oldham, N. 1., Bhattacharya, S., Gleadle, 1. M.,
Ratcliffe, P. J., Pugh, C. W., and Schofield, C. 1., Journal of Biological
Chemistry, 2002, 277, 26351-26355.

Lloyd, M. D., Merritt, K. D., Lee, V., Sewell, T. J., Wha-Son, B., Baldwin, 1. E.,
Schofield, C. J., Elson, S. W., Baggaley, K. H., and Nicholson, N. H.,
Tetrahedron, 1999, 55, 10201-10220.

Dotzlaf, J. E. and Yeh, W.-K., Journal of Bacteriology, 1987, I 69, 1611-1618.

Baldwin, 1. E., Adlington, R. M., Coates, J. B., Crabbe, M. J. C., Crouch, N. P.,
Keeping, J. W., Knight, G. C., Schofield, C. J., Ting, H. H., Vallejo, C. A.,
Thomiley, M., and Abraham, E. P., Biochemical Journal, 1987, 245, 831-841.

Samson, S. M., Belagaje, R., Blankenship, D. T., Chapman, 1. L., Perry, D.,
Skatrud, P. L., Vanfrank, R. M., Abraham, E. P., Baldwin, 1. E., Queener, S. W.,
and Ingolia, T. D., Nature, 1985, 318, 191-194.

Thomburg, L. D., Lai, M. T., Wishnok, J. S., and Stubbe, 1., Biochemistry, 1993,
32, 14023-33.

Solomon, E. 1., Brunold, T. C., Davis, M. 1., Kemsley, J. N., Lee, S. K., Lehnert,
N., Neese, F., Skulan, A. J., Yang, Y. S., and Zhou, 1., Chemical Reviews, 2000,
100, 235-349.

Hausinger, R. P., Crit Rev Biochem Mol Biol, 2004, 39, 21-68.

Ryle, M. J., Liu, A., Muthukumaran, R. 8., Ho, R. Y., Koehntop, K. D.,
McCracken, 1., Que, L., 1r., and Hausinger, R. P., Biochemistry, 2003, 42, 1854-
62.

Liu, A., Ho, R. Y., Que, L., Jr., Ryle, M. J., Phinney, B. S., and Hausinger, R. P.,
J Am Chem Soc, 2001, 123, 5126-7.

27

79.

80.

81.

82.

83.

Mehn, M. P., Fujisawa, K., Hegg, E. L., and Que, L., Jr., J Am Chem Soc, 2003,
125, 7828-42.

Rohde, J. U., In, J. H., Lim, M. H., Brennessel, W. W., Bukowski, M. R., Stubna,
A., Munck, E., Nam, W., and Que, L., Science, 2003, 299, 1037-1039.

Price, 1. C., Barr, E. W., Tirupati, B., Bollinger, J. M., Jr., and Krebs, C.,
Biochemistry, 2003, 42, 7497-508.

Price, 1. C., Barr, E. W., Glass, T. E., Krebs, C., and Bollinger, J. M., Jr., J Am
Chem Soc, 2003, 125, 13008-9.

Proshlyakov, D. A., Henshaw, T. F., Monterosso, G. R., Ryle, M. J., and
Hausinger, R. P., J Am Chem Soc, 2004, 126, 1022-3.

28

CHAPTER II

CHARACTERIZATION OF THE IRON-SULFUR CENTER OF

PYRUVATE FORMATE-LYASE ACTIVATING ENZYME

The results presented in this chapter have been published in Broderick, J. B., Henshaw, T.
F., Cheek, J ., Wojtuszewski, K., Smith, S. R., Trojan, M. R., McGhan, R. M., Kopf, A.,
Kibbey, M., and Broderick, W. E., Biochemical and Biophysical Research
Communications, 2000, 269, 451-456 and Krebs, C., Henshaw, T. F., Check, 1., Huynh,
B. H., and Broderick, J. B., Journal of the American Chemical Society, 2000, 122, 12497-
12506.

My contribution to this work did not include acquisition or analysis of Mossbauer spectra

29

11.1 Introduction

Pyruvate forrnate-lyase is a pivotal enzyme in the anaerobic glucose metabolism
of Escherichia coli and other facultative anaerobes. It catalyzes the reversible non-
oxidative cleavage of pyruvate to formate and acetyl-CoA, the determinate step in
anaerobic glucose metabolism. PF L is a member of a class of enzymes that contain a
catalytically essential glycyl radical which is generated in a post-translational
modification; in the case of PFL this modification is catalyzed by pyruvate forrnate-lyase
activating enzyme (1, 2). This modification is dependent on S-adenosylmethionine,
which is reductively cleaved to methionine and 5’-deoxyadenosine (3). When the glycine
residue is labeled with deuterium, the label is incorporated into 5’-deoxyadenosine,
implicating a 5’-deoxyadenosyl radical as the direct hydrogen-abstracting species (4). A
major interest in the Broderick lab is the mechanism by which this putative radical is
generated.

The first detailed studies of PF L-AE suggested the presence of an organic
cofactor, as indicated by a broad absorbance in the visible region, but further
characterization was hampered by the low availability of the enzyme (1). These
problems were overcome with the cloning (5) and overexpression of the gene which
yielded large quantities of enzyme for detailed study (6). Unfortunately, most of this
protein was found in insoluble inclusion bodies, and the researchers resorted to purifying
the enzyme under denaturing conditions followed by refolding and reconstitution with a
variety of metal ions. The activity was found to dependent on Fe(II), which could be

incorporated into the enzyme in an approximate 1:1 ratio, and was inhibited by other

30

thiophilic metals, suggesting that the metal was coordinated by cysteine residues. The
activity exhibited by the refolded protein was much lower than the native enzyme
purified previously and showed no appreciable visible chromophore, observations that
indicate a fundamental difference between the native and refolded protein (1, 6).

The first work to identify an iron-sulfur cluster in PFL-AB was done by the
Broderick lab at Amherst College in collaboration with Professor Michael Johnson at the
University of Georgia (7). The evidence that PFL-AB contained an iron-sulfur cluster
was obtained using enzyme that had been purified using nitrogen-filled glove bags and
argon-purged buffers. Resonance Raman spectroscopy indicated the presence of a
mixture of [4Fe-4S]2+ and [2Fe-2S]2+ clusters. Then the sample was reduced with excess
dithionite, only the [4F e-4S]2+ clusters remained, as indicated by resonance Raman and
the lack of an EPR signal. When the enzyme was reduced in the presence of AdoMet, a
fast-relaxing EPR signal was observed that was attributed to a [4Fe-4S] 1+ cluster which
was not observed in the absence of substrate. While the method of purification used in
these studies allowed for the isolation of enzyme containing intact clusters, they could
hardly be considered strictly anaerobic conditions. In addition, the plasmid used to
overexpress pflA (the gene that encodes PFL-AB) used a heat-inducible promoter,
requiring growth at 42 °C. This elevated temperature could contribute to degradation of
the cluster.

The work described in this chapter presents a detailed characterization of the iron-
sulfur cluster of PFL-AB, a necessary prerequisite for determining the role, if any, this
cluster plays in the radical-generating reaction. The work in this chapter also represents a

major step forward in the evolution of our methods of purifying and handling the

31

enzyme. Two improvements in particular have facilitated the characterization of the
enzyme. The first is the subcloning of pflA into a pET-based vector that allows IPTG-
induced expression of the gene. The second is the much more rigorous anaerobic

conditions used in the purification and manipulation of the enzyme using anaerobic

chambers.

11.2 Experimental Methods

Growth and expression of PF L

pKK-PFL (gift from John Kozarich) was used to transform BL21(DE3)pLysS. A
single colony of transformed cells was used to inoculate 5 to 50 mL of LB media
containing 50 mg/mL ampicillin (LB/Amp). This culture was grown to saturation and
then used to inoculate LB/Amp in 2.8 L Fembach flasks or in a 10-L bench-top fermentor
(New Brunswick). The cultures were grown at 37°C with vigorous shaking (flasks) or
continuous air purge and vigorous agitation (fermentor) to early log phase (OD600 ~ 0.6-
0.8), and then induced by addition of isopropyl-B-D-thiogalactopyranoside (IPTG) to 1
mM. The cultures were grown for 2 more h before harvesting by centrifugation (8000

rpm, Sorvall GS3 rotor). The supernatant was decanted and the cells stored at -80°.

Subcloning, growth, and expression of PF L-AE

The pflA gene, which encodes PFL-AE, was digested from pMG-AE (gift from
John Kozarich) using NdeI/HindIII and ligated to pCAL-n-EK that had been cut with the
same enzymes. The resultant vector (pCalnAE) was used to transform BL21(DE3)pLysS.
A single colony of the resulting overexpressing strain was used to inoculate 50 mL of LB

32

containing 50 ug/mL ampicillin (LB/Amp). This culture was grown to saturation at 37
°C and then used to inoculate 10 L of LB/Amp. The 10 L culture was grown at 37 °C in
a bench-top fermentor (New Brunswick) with a continuous air purge and vigorous
agitation. When the culture reached an optical density of 0.6 to 0.8, isopropyl-B-D-
thiogalactopyranoside (IPTG) was added to 1 mM final concentration, and the medium
was supplemented with 150 mg F e(N H4)2(SO4)2 per liter of culture. The culture was
allowed to grow for an additional 2 h, at which time the temperature was reduced to 4 °C
and the culture was purged with nitrogen. The culture was incubated for 14—24 h at 4 °C
under nitrogen before harvesting under anaerobic conditions. The harvested cells were
stored under nitrogen at -80 °C until used for purification.

In order to prepare PF L-AE for Mossbauer studies, the
pCalnAE/BL21(DE3 )pLysS strain described above was grown in a defined MOPS
medium (8) that was modified to include 0.5 mM CaClz, 1% (w/v) casamino acids, and
0.001% each of pyroxidine, riboflavin, niacinamide, folic acid, cyanocobalamin,
pantothenic acid, and thiocitic acid. 57Fe was added to a final concentration of 20 uM
from a concentrated stock that was prepared by dissolving the metal in a minimal volume

of 2:3:1 H20:HCI (concentrated):HNO3 (concentrated).

Purification of pyruvate formate-lyase-activating enzyme (PFL-A E)

PFL-AB was purified from E. coli BL21(DE3)pLysS transformed with pCAL-n-
AE3, prepared as described above. All steps in the purification were performed in a
single day under strictly anaerobic conditions in a Coy anaerobic chamber (Coy
Laboratories, Grass Lake, M1) at ambient temperature except where noted. Solutions and

buffers used in the purification were thoroughly degassed prior to bringing them in to the

33

Coy chamber. Approximately 4 to 6 g of cell paste was suspended in 5 to 10 mL of
enzymatic lysis buffer containing 50 mM Tris—sulfate, pH 7.5, 200 mM NaCl, 1% Triton
X-100, 5% glycerol, IOmM MgClz, lmM DTT, 40 mM dithionite, 1 mM PMSF, 8 mg
lysozyme, and trace amounts (approximately 0.1 mg each) of DNase I and RNase A.
This suspension was incubated at ambient temperature for 1 h, and then centrifuged at
15,000 rpm (SS34) for 15 min at 4°C. The extract was decanted and used directly in
purification. Up to 10 mL of the crude extract was loaded onto a Sephacryl S-200 HR
column (5 x 60 cm) equilibrated with 50 mM Hepes, 200 mM NaCl, pH 7.2. The protein
was eluted with this same buffer at 3 mL/min. AE eluted from the column in a relatively
sharp peak at approximately 750 mL after injection. The fractions were analyzed by
SDS-PAGE, and those determined to be >95% pure were pooled and concentrated using
an Amicon concentrator with W] 0 filter membranes. If additional purification was
required, the protein was chromatographed on a Superdex 75 column (1.6 x 60 cm) using
the same buffers. The concentrated, purified protein was flash-frozen and stored in O-

ring-sealed tubes at -80 °C.

Activity, protein and iron assays

PF L-AE activity was determined by the direct detection of the PFL glycyl radical
by EPR spectroscopy. A 1 ml reaction mix contained 0.1 M Tris-HCl, pH 7.6, 0.1 M
KCl, 10 mM DTT, 10 mM oxamate (allosteric effector), 10 mg/ml PFL, 200 M 5-
deazariboflavin, 0.2 mM AdoMet and catalytic amounts of PFL-AB. This mix was
prepared in an Mbraun glove box from anaerobic stocks. The mix was divided into EPR
tubes and illuminated with a 500 W halogen lamp for varying times. The samples were

maintained at ambient temperature (20-25 °C) by immersion in a water bath, which was

34

cooled with ice as necessary. The amount of glycyl radical generated was determined by
EPR spectroscopy.

Protein concentrations were determined using the method of Bradford using a kit
purchased from Bio-Rad. For PFL-AB, these results were calibrated by amino acid
hydrolysis of the purified enzyme done at the MCB core facility at the University of
Massachusetts.

Iron concentrations were determined according to the method of Beinert (9).

EPR and Mossbauer Spectroscopy

Mossbauer spectroscopy was done in collaboration with Professor Boi Hanh
(Vincent) Huynh and Dr. Carsten Krebs at Emory University.

First derivative EPR spectra were recorded on a Bruker ER—200D-SRC or
ESP300E spectrometer equipped with a continuous-flow helium cryostat. Details of
conditions are given in figure legends. Mossbauer spectra were recorded in either a
weak-field spectrometer equipped with a Janis 8DT variable-temperature cryostat or a
strong-field spectrometer fumished with a Janis CNDT/ SC SuperVaritemp cryostat
encasing an 8-T superconducting magnet. Both spectrometers operate in a constant
acceleration mode in a transmission geometry. The zero velocity of the spectra refers to

the centroid of a room-temperature spectrum of a metallic iron foil.

35

11.3 Results

Enzyme Activity

The preparations of PF L-AE prepared as described above exhibit specific
activities that are higher than those previously published, presumably as a result of the
direct detection of the glycyl radical by EPR spectroscopy rather than the coupled assay
used previously (1, 6, 10). A preparation AE from cells grown in enriched media exhibits
a specific activity of 95 U/mg and AE purified from cells grown in media exhibits a
specific activity of 48 U/mg. As discussed below, these specific activities correlate

directly with iron content.

Characterization of purified PF L-AE

Purified PFL-AE contains iron (2.8 :t 0.3 mol/mol PFL-AB). The UV-visible
spectrum of purified PFL-AE is indicative of the presence of an iron sulfur cluster
(Figure 11.1), with a maximum at 412 nm (e = 3.0 mM"em°‘), and shoulders at 320 nm (e
= 4.4 mM'em"), 455 nm (a = 2.6 mM“em"), and 550 nm (e = 0.94 mM"em"). The
energies and extinction coefficients of these absorption bands are similar to those
observed for the cuboidal [3Fe—4S] form of aconitase (11). PFL-AE exhibits a strong,
nearly isotropic electron paramagnetic resonance (EPR) signal, which is centered at g =
2.02 and observable only below approximately 30 K (Figure 11.2). The g value and low
anisotropy of this fast-relaxing signal are consistent with its assignment to a [3Fe—4S]’r
cluster. The signal, quantified by using a Cu(11)edta standard and the method outlined by
Aasa and Vanngard (12), accounts for 62% of the total iron in the sample, with the

remainder of the iron being present in an EPR-silent form.

36

2.01

 

1.5-

1.0..

Absorbance

0.5—7

 

 

.0.0 I --

l

400 600
Wavelength

800

Figure 11.1 UV—visible absorption spectrum of pyruvate formate-lyase activating
enzyme as isolated.

The protein was 0.34 mM in 50 mM Hepes/200 mM NaCl, pH 7.2, and the spectra were
recorded in a 1 mm pathlength cuvette under anaerobic conditions at 4°C. Reproduced
with permission from (13).

37

 

J 12K

 

 

 

 

 

 

 

 

 

 

 

 

é‘
“"3 l
3 l
.5
2‘. l ”‘
in 20K
AL— 30K
._ -39}-
I .. .59K
1‘ T l i

Figure 11.2 X-band EPR spectra of pyruvate formate-lyase-activating enzyme as a
function of temperature.

PFL-AE is 1.26 mM in 50 mM Hepes/200 mM NaCl, pH 7.2. Conditions of
measurement, T as indicated, microwave power, 0.02 mW, microwave frequency, 9.4792

GHz, modulation amplitude, 10.084, and receiver gain, 2 x 104, 5 scans accumulated for
each spectrum. Reproduced with permission from (13).

38

Preparations of the 57Fe-enriched PF L-AE were analyzed for iron and protein
content. The results Show an iron to protein monomer ratio of 1.3 :l: 0.1. This value is
lower than the 2.8 Fe/protein monomer for the enzyme purified from bacteria grown in
enriched medium with naturally available Fe. Since the 57Fe-enriched PFL-AE was
isolated from bacteria grown in defined medium, this observed difference in Fe content
may suggest inefficient incorporation of the Fe cluster under suboptimal growth
conditions. Two observations, however, indicate that the only difference between 57Fe-
enriched and unenriched PFL-AE is the absolute quantity of iron incorporated, not the
form in which it is incorporated. First, the EPR and UV-visible spectroscopic properties
of 57Fe-enriched PFL-AE are essentially identical to those of PFL-AE isolated from
bacteria grown with naturally available Fe. Second, the specific activity of the 57Fe-
enriched PFL-AE containing 1.3 Fe per protein monomer (48 U/mg) is approximately
half that for unenriched PFL-AE containing 2.65 Fe per protein monomer (95 U/mg).
The specific activities reported here are higher than any previously published for PFL-
AE, a difference that is likely a result of the direct detection of the glycyl radical rather
that the coupled assay previously used (1, 6).

Figure 11.3 shows a Mossbauer spectrum of the as-isolated PFL-AE recorded at
4.2 K in a parallel applied field of 50 mT. At least three spectral components are
discernible from this spectrum. The major component, which accounts for 66% of the Fe
absorption, displays magnetic hyperfine interactions with absorption extending from -2 to
3 mm/s (solid line plotted above the data). This component is composed of three equal-
intensity subspectral components and exhibits field-orientation and field-strength

dependence consistent with an S = 1/2 electronic system. At temperatures

39

A b A 1 .. I . .i
$1.0L * it: '5' ‘:
“320i l bWW'O
I-I - a
g -. — :5 - I ‘1 I I r1.0
< W 1‘ ,N ,

o
c
Ԥ
5
i
i
l
‘i
l
g
'5
..‘
l
I
ll
I
ii
I
11L
“9
co

 

 

 

 

 

 

 

 

 

 

0.0 ,5 t J;
1.0 ‘ RWLOD
. , , .
:3 l M“ N,“ ‘1.0
4 ; it
. '5; -2.0
.31-41514 3' -8‘-41014‘81
Velocity (mm/s) Velocity (mm/ S)

Figure [1.3 Miissbauer spectra of anaerobically purified native PFL-AE.

The spectra were recorded at 4.2 K in a parallel field of 0.05 T (A and C), 170 K in
zero field (B), 4.2 K in a parallel field of 4 T (D), or 8 T (E). For the spectra shown in
C, D, and E, contributions from the [4Fe-4S]2+ (8%) and [2Fe-2S]2+ (12%) clusters
have been removed. The solid line overlaid with the experimental spectrum in A is
the sum of the theoretical simulations of the cuboidal [3Fe-4S]+ (soild line above the
experimental spectrum, 66%), [2Fe-2S]2+ (dashed line, 12%), and [4Fe-4$]2+
(dotted line, 8%) clusters. The solid line in B is the sum of the quadrupole doublets
arising from the [4Fe-4S]2+ cluster (dotted line above the experimental spectrum)
and from the cuboidal [3Fe-48]+ and [2Fe-2S]2+ (dashed line) clusters. The solid
lines shown in C, D, and E are theoretical simulations of the linear [3Fe-4S]+ cluster
using the parameters given in (11). The arrows in D indicate the positions at which
the outermost lines of the three Fe sites of the linear [3Fe-4S]+ cluster overlap at 4 T
(see text). Reproduced with permission from (14). Copyright 2000 American
Chemical Society.

40

above 77 K, this magnetic component collapses into a quadrupole doublet (Figure 1138)
with parameters (5 = 0.23 mm/s and AEQ = 0.58 mm/s at 170 K) typical of tetrahedral
sulfur-coordinated high-spin Fe(III). All these features are characteristic of the all-ferric
cuboidal [3Fe-4S]+ cluster. This magnetic component is thus assigned to such a cluster.
This assignment is consistent with the EPR finding that the majority of the Fe-S cluster in
the as-isolated PF L-AE is in an S = 1/2 [3Fe-4S]+ state.

A second spectral component accounting for 12% of the total Fe absorption is a
sharp quadrupole doublet observed at 4.2 K (dashed line plotted above the data in Figure
4A) with parameters (listed in Table 1) characteristic of [2Fe-2S]2+ clusters (15, 16).
Spectra recorded at strong applied field indicate that this component originates from a
diamagnetic species. On the basis of this observed diamagnetism and the characteristic
parameters, this component is assigned to a [2Fe-2S]2+ cluster. Removal of the
contributions of the [3 Fe-4S]+ and [2Fe-28]2+ clusters from the raw data reveals a third
component identical to that of the [4Fe-4S]2+ cluster observed in the dithionite-reduced
PFL-AB (discussed below and dotted line plotted above the data in Figure II.3A). The
high-energy line of this doublet can be seen as a shoulder at ~l mm/s in the raw data.
This component is estimated to account for ~8% of the total Fe absorption. A sum of the
three components (the [3Fe-4S]+, [2F e-2S]2+, and [4Fe-48]2+) yields the solid line plotted
over the data in Figure II.3A. Although the agreement between the experiment and the
simulated spectrum is reasonable, approximately 14% of the Fe absorption remains not
accounted for. Spectra recorded at strong applied field reveal the presence of a fourth

component.

41

The complexity of the above-discussed low-temperature spectrum is greatly
reduced when data are collected at high temperatures. This is due to the fact that, at high
temperature, the [3 Fe-4S]+ cluster exhibits a single quadrupole doublet, and this doublet
is similar to that of a [2Fe-2S]2+ cluster since both clusters are composed of only
Fe(III)S4 units. As expected, the 170 K spectrum of the as-isolated PFL-AB (Figure
11.3B) shows predominantly a quadrupole doublet arising from both the [3F e-4S]+ and
[2F e-2S]2+ clusters (dashed line in Figure 11.38). This central doublet accounts for 80%
of the total Fe absorption, consistent with the sum of the amounts of these two types of
clusters determined at 4.2 K. The presence of a small quantity of the [4Fe-4S]2+ cluster is
supported by the appearance of a shoulder at the high-energy line of the central doublet.
To illustrate this point, we plot in Figure 1133 a simulated spectrum (dotted line in
Figure 1133) of the [4Fe-4S]2+ cluster using the parameters determined from the
dithionite-reduced sample (Table 11.1). The plotted spectrum is normalized to 8% of the
total Fe absorption. Addition of the simulated central quadrupole doublet with the
simulated [4F e-4S]2+ doublet yields the solid line overlaid with the experimental data.
Again, reasonable agreement between experiment and simulation is observed.

The fourth spectral component is readily observed in a spectrum recorded at 4.2 K
in a parallel field of 4 T (Figure 11.3D). This component is seen as a magnetic spectrum
with its outermost lines appearing at approximately -5 and 5 mm/s (indicated by arrows).
In a much weaker (0.05 T) or stronger (8 T) applied field (Figure 11.3, spectra C and B,
respectively), however, this magnetic component becomes broad and is difficult to detect.
This field-strength dependence and the observed magnetic splitting are consistent with

those of a linear [3F e-4S]+ cluster (11, 17). In the following, we use the results obtained

42

Table 11.1 Temperature-Dependent Miissbauer Parameters of the

Fe-S Clusters in Native PFL-AE

 

 

T(K)
4.2 113 170
[4Fe-4S]2+site1 8(mm/s) 045(2) 044(2) 044(2)
AEQ(mm/s) 1.15(4) l.l3(4) 1.10(4)

1) 0.3
[4Fe-4812+ site2 5(mm/s) 045(2) 044(2) 042(2)
AEQ (mm/s) 1.00 (4) 0.83 (4) 080(4)

11 0.7
[2Fe-2S]2+ 8 (mm/s) 029(2) 0.26 (2) 0.23 (2)
AEQ(mm/s) 058(4) 058(4) 058(4)

0 0

 

43

for the linear [3Fe-4S]+ cluster in aconitase (l 1) as an example to illustrate this point and
to compare with the PFL-AE spectra. For a linear [3Fe-4S]+ cluster, the three S = 5/2
Fe(III) sites are spin-coupled to form an S = 5/2 ground state. The spin of the middle Fe
site (site 3) is antiparallel to those of the two terminal Fe sites (sites 1 and 2).
Consequently, the internal field of site 3 is parallel to the applied field, while those of the
other two sites oppose the applied field. For the linear [3Fe-4S]+ cluster in aconitase, (11)
the observed internal fields for sites 1, 2, and 3 are -33.6, -31.9, and 24.2 T, respectively.
The signs indicate the direction of the internal field in relation to that of the applied field.
At 0.05 T, since all three internal fields are different in magnitude, the spectra arising
from the three Fe sites show different magnetic splittings. Superposition of these three
spectra results in a relatively broad spectrum (solid line overlaid with the experimental
spectrum in Figure II.3C). At 4 T, to a good approximation, the magnitudes of the
effective fields at sites 1 and 2 are reduced by 4 T to 29.6 and 27.9 T, respectively, while
that of site 3 is increased to 28.2 T. These values are quite similar, and thus, all three sites
show similar magnetic spectra, superposition of which results in a "single" magnetic
spectrum with its intensity tripled (solid line in Figure II.3D). At 8 T, the magnitudes of
the effective fields at sites 1, 2, and 3 are again different, and a broad magnetic spectrum
is again observed (solid line in Figure II.3E). On the basis of the above analysis, it is
concluded that the as-isolated PFL-AE contains linear [3Fe-4S]+ clusters, accounting for
approximately 10% of the total iron absorption.

In summary, analysis of the Mossbauer data of the as-isolated PFL-AB shows the
presence of mixtures of Fe-S clusters. A majority of the clusters are present in the

cuboidal [3F e-4S]+ state (~0.29 cluster/protein monomer), while minor portions are in the

44

[2Fe-2S]2+ (~0.08 cluster/monomer), [4F e-4S]2+ (~0.03 cluster/monomer), and linear

[3Fe-4S]Jr (~0.04 cluster/monomer) states.

Characterization of the dithionite-reduced enzyme

The 4.2 K Mossbauer spectrum of a dithionite-reduced PF L-AE sample (Figure
11.4), recorded in a magnetic field of 50 mT oriented parallel to the ‘i-beam, displays an
intense central quadrupole doublet (marked by a bracket), a broad absorption peak at ~2.8
mm/s, and weak absorptions between -2 and +2.5 mm/s. These features are associated
with three spectral components and thus represent three different Fe species. The broad
absorption peak at ~2.8 mm/s is the high-energy line of a quadrupole doublet (shown as a
solid line in Figure 11.4), of which the parameters (see legend of Figure 11.4) and shape
are indicative of adventitiously bound high-spin Fe(II). Removal of the contribution of
the adventitious Fe(II) (25% of the total Fe absorption) from the raw data reveals the
spectrum arising from the Fe-S clusters (Figure 11.5A). The central quadrupole doublet,
accounting for 66% of the total Fe absorption, is best simulated with two overlapping
quadrupole doublets (dashed line in Figure 11.5A). The parameters used in the simulation
are listed in Table 11.1 and are typical for [4F e-4S]2+ clusters (18, 19). Without
exception, [4Fe-4S]2+ clusters have an S = 0 ground state, resulting from
antiferromagnetic coupling of the two valence-delocalized Fe(II)Fe(III) units. To
examine the spin state of the species associated with the central doublet, a Spectrum of
the dithionite-reduced PFL-AB was recorded in a strong external field of 8 T at 4.2 K
(Figure 11.5C). The solid line plotted over the data in Figure 2C is a simulation using the
parameters obtained for the central doublet from the weak-field spectrum and assuming

diamagnetism. The agreement between the simulation and the central portion of

45

 

0.0 *-

 

.. I ..
6‘- 1.0 . 'l I. '
é _ I . ' 1 .-
..8 , '
g 20 I I '
§ ' - "
.D I "
< __ -
I. I.
3.0- ' I ..
I
)- cur

 

 

 

Velocity (mm/s)

Figure 11.4 Miissbauer spectrum of dithionite-reduced native PFL-AE recorded at
4.2 K in a magnetic field of 0.05 T applied parallel to the y-beam.

The bracket indicates the positions of the quadrupole doublet arising from the [4Fe-4S]2+
cluster. The solid line is the theoretical simulation of the adventitiously bound Fe(II)
assuming two quadrupole doublets with (5(1) = 1.28 mm/s, AEQ(1) = 3.07 mm/s, 5(2) =
1.10 mm/s, and AEQ = 2.42 mm/s. Doublet 1 contributes 17% of the total Fe absorption,
and doublet 2 contributes 8%. Reproduced with permission from (14). Copyright 2000
American Chemical Society.

46

 

 

 

 

 

 

 

 

 

Figure 11.5 Mt'issbauer spectra of dithionite-reduced PFL-AE

Spectra were recorded at 4.2 K in a parallel field of 0.05 T (A), 170 K in the absence of
applied field (B), and 4.2 K in a parallel field of 8 T (C). The contributions from the
adventitiously bound Fe(II) have been removed for the spectra shown in A and B. The
soild lines in A and B are summations of theoretical simulations of the [4Fe-4S]2+
(dashed lines, 66% of total Fe absorption) and the [4Fe-4S]+ (dotted lines, 12%) clusters.
The solid line in C is the theoretical simulation of the [4F e-4S]2+ cluster, assuming
diamagnetism. Reproduced with permission from (14). Copyright 2000 American
Chemical Society.

47

 

 

 

 

3.0 3.2 34 3.6 3.8 4.0

L

 

Magnetic Field (T)

Figure 11.6 EPR spectrum of dithionite-reduced PFL-AE showing a weak signal
originating from the [4Fe-4S]’r cluster.

The spectrum was recorded with the following instrumental settings: temperature, 12 K;
microwave power, 2 mW; microwave frequency, 9.65 GHz; modulation amplitude, 1
mT; modulation frequency, 100 kHz. Reproduced with permission from (14). Copyright
2000 American Chemical Society.

48

the experimental spectrum confirms that the central doublet is arising from a diamagnetic
species. Thus, this component is assigned to a [4Fe-4S]2+ cluster.

The component having absorption extending from -2 to 2.5 mm/s, which accounts
for 12% of the total Fe absorption, is originating from a paramagnetic species and can be
attributed to a [4Fe-4S]+ cluster with an S = 1/2 ground state. Miissbauer spectra of such
clusters consist of two equal intensity subspectra representing a valence-delocalized
Fe(II)Fe(III) pair and a diferrous pair (15, 18). Because of the weak intensity and
because the number of variables required for spectral simulation of a [4Fe-4S]+ cluster is
large, we did not attempt to determine the specific parameters for the [4Fe-4S]+ cluster in
the dithionite-reduced PFL-AE by fitting this paramagnetic component. Instead, to show
that this paramagnetic component is consistent with that of a [4Fe-4S]+ cluster, we have
used the parameters obtained for the [4Fe-4S]+ cluster of Bacillus stearothermophilus
ferredoxin (18) to simulate a spectrum for comparison with this component. The
simulation is plotted in Figure 11.5A as a dotted line. It can be seen that the magnetic
splitting and shape of the paramagnetic component in PFL-AB agree well with the
simulation. Addition of the simulated [4Fe-4S]+ spectrum (12%) with the least-squares fit
spectrum for the [4Fe-4S]2+ cluster (66%) generates the solid line overlaid with the
experimental spectrum shown in Figure 11.5A. The agreement observed between the
theory and the experimental spectrum supports the above assignment.

At high temperatures the electronic relaxation of a [4Fe-4S]+ cluster is fast in
comparison with the Larmor frequency of the 57Fe nucleus, resulting in cancellation of
the internal field and collapse of the low-temperature magnetic spectrum into two

quadrupole doublets associated with the Fe(II)F e(111) and the diferrous pairs. (15, 18)

49

Figure 11.5B shows a spectrum of the dithionite-reduced PF L-AE recorded at 170 K in
the absence of an applied field (the contribution of the adventitiously bound F e(11) has
been removed from the spectrum shown). As expected, the paramagnetic component
observed at 4.2 K has collapsed into quadrupole doublets at this temperature. A small
shoulder at ~1.3 mm/s is observed. A comparison with the simulated spectrum of the B.
stearothermophilus [4Fe-4S]+ cluster (18) at this temperature (dotted line in Figure 11.5B)
indicates that this shoulder is at the position of the high-energy line of the quadrupole
doublet of the diferrous pair. The intensity of this shoulder is also consistent with the
percent absorption (12% of total Fe absorption) determined for the [4Fe-4S]+ cluster from
the low-temperature data. To demonstrate that the high-temperature spectrum is in
agreement with the assignment made from analyzing the low-temperature data, we have
again added the simulated central doublet (long-dashed line in Figure 1153 for the [4Fe-
4S]2+ cluster) and the simulated B. stearothermophilus [4Fe-4S]+ spectrum (dotted line)
according to the intensities determined from the low-temperature spectra (66% for the
[4Fe-4S]2+ and 12% for the [4Fe-4S]+) and plotted the sum (solid line) over the
experimental spectrum. The agreement between the simulation and experiment indicates
that both the high-temperature and low-temperature Mossbauer data are consistent with
the conclusion that, except for the adventitiously bound Fe(II), only [4F e-4S] clusters are
present in the dithionite-reduced PF L-AE. The percent absorption of the [4Fe-4S] cluster
determined from the Miissbauer measurement, together with the iron and protein
determinations, yields a stoichiometry of ~0.25 [4Fe-4S] cluster per protein monomer, of
which 85% are in the diamagnetic [4F e-4S]2+ state and the remainder are in the

paramagnetic [4F e-4S]+ state.

50

The presence of a small quantity of [4Fe-4S]+ cluster in the dithionite-reduced
PF L-AE is also supported by EPR measurements. An X-band EPR spectrum of a
dithionite-reduced PFL-AE prepared in parallel with the Mbssbauer sample displays an
axial S = I/2 signal with g values at 2.01 and 1.94 (Figure 11.6). This signal is very similar
to that reported for the dithionite-reduced reconstituted PFL-AE (10). Double integration
of this signal yields a spin quantitation of 0.03-0.04 spin/protein monomer, a value that is

consistent with the Mdssbauer finding of ~0.04 [4F e-4S]+ cluster per protein monomer.

[1.4 Discussion

The results presented in this chapter show that native PFL-AE may be isolated
under anaerobic conditions and may be isolated with nearly a firll complement of iron-
sulfur cluster. In addition these preparations exhibit high specific activities in the
absence of Fe(II) in the assay mix. Previous studies on PFL—AB included 0.2 mM Fe(II)
in the assay and this added iron was thought to be an absolute requirement for activity (1,
6). In addition, the specific activities of different enzyme preparations correlate directly
with the iron content of the preparation. For example, the enzyme that was isolated for
Mossbauer studies showed an iron content of 1.3 F e/monomer and a specific activity of
48 U/mg. A preparation isolated from cells grown in enriched media had an iron content
of 2.65 Fe/monomer and a specific activity of 95 U/mg. The direct correlation between
the iron content, in the form of an FeS cluster, indicate that this cluster is required for
enzyme activity.

The detailed spectroscopic studies presented in this chapter show that AB has the

ability to coordinate a plethora of cluster types, including the [2Fe-28]2+, the [4Fe-4S]2+

51

and the cuboidal [3Fe-4S]+ with the latter being the predominant cluster type. In
addition, a small amount of linear [3 Fe-4S]+ cluster was identified by Mossbauer
spectroscopy. Under reducing conditions, these cluster types are converted to a mixture
of [4Fe-4S]2+/1+ clusters. This finding is consistent with the expectation that the [3Fe-4S]
cluster represents an inactive form of the enzyme and that the [4Fe-4S] clusters are
generated under the reducing conditions of the activity assay. The results presented in
Chapter IV explicitly demonstrate that the [4Fe-4S]1+ is the catalytically relevant form of
the cluster.

The ready accessibility of the [3Fe-4S] and [4Fe-4S] cluster types invites a
comparison with aconitase and suggests that there may be a unique iron Site in PFL-AE.
In addition, aconitase is the only other enzyme system in which a linear [3Fe-4S] cluster
has been identified (11). A CX3CX2C cluster binding motif has been identified in the
F eS/AdoMet enzyme family (20), but the identity of the fourth ligand to the cluster has
not been identified. The results presented in Chapter III demonstrate that there is a

unique iron site in PFL-AE and that AdoMet is a ligand to this site.

52

10.

11.

12.

13.

14.

References

Conradt, H., Hohmannberger, M., Hohmann, H. P., Blaschkowski, H. P., and
Knappe, 1., Archives of Biochemistry and Biophysics, 1984, 228, 133-142.

Knappe, J., Neugebauer, F. A., Blaschkowski, H. P., and Ganzler, M.,
Proceedings of the National Academy of Sciences of the United States of America-
Biological Sciences, 1984, 81, 1332-1335.

Knappe, J. and Schmitt, T., Biochemical and Biophysical Research
Communications, 1976, 71, 1110-1117.

Frey, M., Rothe, M., Wagner, A. F. V., and Knappe, 1., Journal of Biological
Chemistry, 1994, 269, 12432-12437.

Rodel, W., Plaga, W., Frank, R., and Knappe, 1., European Journal of
Biochemistry, 1988, 1 77, 153-158.

Wong, K. K., Murray, B. W., Lewisch, S. A., Baxter, M. K., Ridky, T. W.,
Ulissidemario, L., and Kozarich, J. W., Biochemistry, 1993, 32, 14102-14110.

Broderick, J. B., Duderstadt, R. E., Fernandez, D. C., Wojtuszewski, K.,
Henshaw, T. F., and Johnson, M. K., Journal of the American Chemical Society,
1997, 119, 7396-7397.

Garboczi, D., Hullihen, J ., and Pedersen, P., J. Biol. Chem, 1988, 263, 15694-
15698.

Beinert, H., Methods Enzymol, 1978, 54, 435-45.

Kulzer, R., Pils, T., Kappl, R., Huttermann, J., and Knappe, 1., Journal of
Biological Chemistry, 1998, 273, 4897-4903.

Kennedy, M., Kent, T., Emptage, M., Merkle, H., Beinert, H., and Munck, E., J.
Biol. Chem, 1984, 259, 14463-14471.

Aasa, R. and Vanngard, T., Journal of Magnetic Resonance, 1975, 19, 308-315.

Broderick, J. B., Henshaw, T. F., Check, 1., Wojtuszewski, K., Smith, S. R.,
Trojan, M. R., McGhan, R. M., Kopf, A., Kibbey, M., and Broderick, W. E.,
Biochemical and Biophysical Research Communications, 2000, 269, 451-456.

Krebs, C., Henshaw, T. F., Cheek, J., Huynh, B. H., and Broderick, J. B., Journal
of the American Chemical Society, 2000, 122, 12497-12506.

53

15.

16.

17.

18.

19.

20.

Trautwein, A. X., Bill, E., Bominaar, E. L., and Winkler, H., Structure and
Bonding, 1991, 78, 1-95.

Ferreira, G. C., Franco, R., Lloyd, S. G., Pereira, A. S., Moura, 1., Moura, 1.1. G.,
and Huynh, B. H., Journal of Biological Chemistry, 1994, 269, 7062-7065.

Girerd, 1. 1., Papaefthymiou, G. C., Watson, A. D., Gamp, E., Hagen, K. S.,
Edelstein, N., Frankel, R. B., and Holm, R. H., Journal of the American Chemical
Society, 1984, 106, 5941-5947.

Middleton, P., Dickson, D. P. E., Johnson, C. E., and Rush, 1. D., European
Journal of Biochemistry, 1978, 88, 135-141.

Middleton, P., Dickson, D. P. E., Johnson, C. E., and Rush, 1. D., European
Journal of Biochemistry, 1980, 104, 289-296.

Sofia, H. J., Chen, G., Hetzler, B. G., Reyes-Spindola, J. F., and Miller, N. E.,
Nucleic Acids Res, 2001, 29, 1097-106.

54

CHAPTER III

IDENTIFICATION OF A UNIQUE IRON SITE

The results presented in this chapter have been published in: Krebs, C., Broderick, W. E.,
Henshaw, T. F ., Broderick, J. B., and Huynh, B. H., Journal of the American Chemical
Society, 2002, 124, 912-913.

55

111.1 Introduction

Pyruvate forrnate-lyase activating enzyme is a member of a growing class of
enzymes that use AdoMet and iron-sulfur clusters in radical generating roles. These
enzymes have been shown to participate in the generation of stable protein radicals,
sulfur insertion, DNA repair, and tRNA modification. PFL-AB catalyzes the reductive
cleavage of AdoMet to methionine and a putative 5’-deoxyadenosyl radical which
abstracts a hydrogen atom from Gly-734 of PFL to generate a stable glycyl radical.

A conserved feature of the “Radical SAM” (1) or “AdoMet radical” (2) enzymes
is a -C-X3-C-X2-C- cluster-binding motif. However, [4F e-4S] clusters are typically
coordinated to the protein by four ligands, and other conserved cluster-binding residues
have not been identified. In addition, [3Fe-4S] clusters have been identified in several of
the FeS/AdoMet enzymes (3-7), suggesting that there is a unique iron site in these
systems, one that is not ligated by a cysteine residue. AdoMet itself has several
functional groups that could bind the cluster and is an appealing ligand candidate from a
mechanistic perspective. It could ligate the cluster through the ribose hydroxyl groups,
the amino and carboxylate moieties, or the sulfonium center. Direct binding of AdoMet
to the cluster would position it for reductive cleavage and generation of the 5’-dAdo
radical. All of these observations suggest that AdoMet binds directly to the cluster, but
other experiments are necessary to define the nature of the interaction.

Aconitase catalyzes the interconversion citrate and isocitrate in the Krebs cycle
and is an iron-sulfur cluster protein that contains a [3Fe-4S] cluster in the inactive form
and a [4Fe-4S] cluster in the active form. The [4Fe-4S] cluster is ligated by three

cysteines and water in the absence of substrate. Méssbauer spectroscopy is a powerful

56

technique in the elucidation of the roles of iron sulfur clusters. It allows the investigation
of the oxidation state and coordination environment of all forms of iron in the sample.
Beinert, Munck and coworkers took advantage of the isotope selective nature of
Mtissbauer in characterizing the iron-sulfur cluster in aconitase and its role as a Lewis
acid in catalysis (8). When isolated under aerobic conditions, aconitase contains an
inactive [3F e-4S] 1+ cluster and may be activated with iron and reductant to generate an
active [4Fe-4S]2+ cluster. During this activation, the iron is incorporated into a unique
iron site, F ea. By activating aconitase with 57Fe, the researchers obtained clusters with a
powerful spectroscopic probe at the unique iron and showed that the Mtissbauer and
parameters were dramatically altered when bound to the substrate, isocitrate (9).

In the experiments presented in this chapter, a similar dual-isotope approach was
used in combination with Mossbauer spectroscopy to show the existence of a unique iron
site in PF L-AE and provide evidence that AdoMet directly binds to this site in the [4Fe-

4S]2+ cluster.

111.2 Experimental Methods

PFL-AE was expressed and purified as described in Chapter 11 with the inclusion of 1
mM DTT in all purification buffers. In addition, the ratio of absorbance of each PFL-
AE-containing fraction was measured at 280 and 426 nm. Fractions with the highest
A426/A230 ratios (an indirect measure of iron content) that did not show a blue shift of the

A230 peak (an indication of nucleic acid contamination) were pooled and concentrated.

57

Preparation of Specifically Labeled Samples

The [3Fe-4S]+ cluster state was generated by exposing PF L-AE (isolated in the
[4Fe-4S]2+ state) to air for 30 minutes on ice followed by gel filtration through a 5 ml
HiTrap desalting column (Pharrnacia) with 50 mM Tris, 200 mM NaCl, pH 8.5 as eluant.
The red protein was collected and the amount of [3Fe-4S]+ was quantified by EPR. The
iron content was determined by the method of Fish (10) and the protein concentration
determined as described in Chapter II. Reconstitution with 5 7F e was accomplished by
addition of 1.4 equiv (relative to [3Fe-4S]+) of S7FeSOs to the [3Fe-48]+, under an inert
atmosphere, followed by addition of DTT (10 mM). The 57FeSO4 solution was prepared

by dissolving the metal in dilute H2804 followed by neutralization with hydroxide.

111.3 Results and Discussion

The nature of the 57Fe that was incorporated into the [4Fe-4S] cluster was
investigated by Mossbauer spectroscopy. A 4.2 K Mtissbauer spectrum (Figure 111.1A,
hashed marks) of the reconstituted enzyme shows that 70% of the added ”Fe appears as a
quadrupole doublet typical of a [4F e-4S]2+ in PFL-AE, as determined in Chapter 11 (solid
line, Figure III.1A). We estimate the amount of 57Fe incorporated in [4Fe-4S] clusters
(365 pM) is comparable to the amount of [3Fe-4S]+ (3 80 1.1M, from EPR) present prior to
addition of 57Fe and DTT. Taken together, these data strongly suggest that under the
reconstitution conditions the [3Fe-4S]+ cluster is converted into a [4Fe-4S]2+ cluster by
selective incorporation of the 57Fe into the fourth, unique Fe site, a process also observed
in aconitase (9). Addition of AdoMet (10 equiv) to the reconstituted enzyme yields the

Mossbauer spectrum shown in Figure 111.1B. The quadrupole doublet arising from the

58

 

 

 

 

 

' ' F 1 V I V I F
.
’/
’f
'r'
ODr- -I
3 ‘4 7°“
8 " B -' .
5 ln- - PK .' -
'. r '- -
o - - .' “It 4
a
i: I 5;
on- -
0.5- C -
5: 5,
00- -
D
M-
I l l I I 1 l l I
4 -2 o 2 4
VELOCITYOII‘n/I)

Figure 111.1 Mossbauer spectra of 56Fe PFL-AE reconstituted with 57Fe and DTT in
the absence (A) and presence (B) of AdoMet.

The data (hashed marks) were recorded at 4.2 K in a magnetic field of 50 mT applied
parallel to the yrays. The solid line in A is the experimental spectrum of [4F e-4S]2+
clusters in PFL-AE normalized to 70% of the total Fe absorption of A. The solid line in B
is the spectrum of a control sample containing only the reconstitution ingredients and
AdoMet but without PFL—AE and is normalized to 15% of the total Fe absorption of B. A
difference spectrum of B minus A is shown in C. Spectrum D is a difference spectrum of
the spectra of samples A and B recorded in a parallel field of 8 T. The solid lines in C and
D are difference spectra of theoretical simulations of the unique Fe site with and without
AdoMet using the parameters given in the text and assuming diamagnetism. Reproduced
with permission from (1 1). Copyright 2002 American Chemical Society.

59

unique Fe site has reduced in intensity while a new quadrupole doublet appears,
indicating that a portion of the unique Fe site has converted into a new Fe species. This
conversion is best illustrated by a difference spectrum between A and B to eliminate
contributions from Fe species that are common to both samples (Figure 111.1C, B minus
A, hashed marks). The amount of the unique Fe site that converts to the new Fe species
appears as a doublet pointing upward and accounts for 32% of the total 57Fe in the
sample, while the new Fe species accounting for the same portion appears as a doublet
pointing downward. In the absence of AdoMet, as expected, the quadrupole splitting
(AEQ = 1.12 mm/s) and isomer shift (8 = 0.42 m/s) are typical for Fe sites in a [4Fe-
4S]2+ cluster. In the presence of AdoMet, however, the parameters (AEQ = 1.15 mm/S, 8
= 0.72 mm/s) are distinct, with the unusually large isomer shift signaling an increase of
coordination number and/or binding of more ionic ligands, suggesting coordination of the
substrate AdoMet to the unique Fe site. The solid line shown in Figure 111.1C is a
theoretical difference spectrum of the two doublets mentioned above.

To Show that the doublets in A and B originate from Fe associated with PF L-AE
and not from Fe in solution, we obtained the Mossbauer spectrum of a sample containing
5 7F e(11), DTT, and AdoMet but without PFL-AB. The spectrum, shown as a solid line in
Figure 111.1B and normalized to 15% of the Fe absorption, exhibits a broad quadrupole
doublet with parameters (AEQ = 3.38 mm/s, 8 = 0.73 mm/s) typical of a tetracoordinate
Fe(II)S4 complex. The quadrupole doublets assigned to iron-sulfur clusters of PFL-AE
are not observed in the control sample.

To demonstrate that both quadrupole doublets are due to an Fe site that is part of a

diamagnetic [4F e-4S]2+ cluster, we have recorded the spectra of the reconstituted PFL-

60

AE with and without AdoMet in a parallel applied field of 8 T. A difference spectrtun of
the two 8-T spectra is shown in Figure 1D. The solid line overlaid with the experimental
data is a theoretical difference spectrum using the parameters obtained for the two
doublets mentioned above and assuming diamagnetism. The excellent agreement between
theory and experiment establishes unambiguously that the two doublets are indeed arising
from an iron site associated with a diamagnetic system. This observation also indicates
that upon binding of the substrate, the unique Fe site remains exchange coupled to the
other 3 Fe atoms in the cluster.

Analogous experiments were also done using 57Fe-containing PF L-AE that had
been reconstituted with natural abundance iron. The Méssbauer spectra of these samples
were similar to those of the specifically labeled samples, but were insensitive to the
presence of AdoMet.

While the existence of a unique iron site in several Fe/S AdoMet enzymes has
long been postulated, the results presented in this chapter provide the first direct evidence
for such a site and suggest that AdoMet binds to this iron in a catalytically relevant
manner. The observation that AdoMet binds before reduction to the catalytically active
[4Fe-4S]1+ suggests that the bound form of the oxidized cluster represents an ES complex
that is a prerequisite for radical generation. The strict conservation of the C-XXX-C-XX-
C motif also suggests that such a site-differentiated cluster may be a common feature of
the Fe/S-AdoMet enzymes. These three cysteine residues have been repeatedly
implicated as cluster ligands, but the fourth ligand had not been identified in any AdoMet
radical system. The results presented here show that AdoMet itself is a ligand to the

cluster of PFL-AE.

61

It has been proposed that the sulfonium center of AdoMet interacts with an iron
site in the reaction catalyzed by LAM (12). Using seleno-substituted AdoMet (Se-
AdoMet), a Se X-ray absorption study detected 2.7 A interaction between Se-Met (the
cleavage product of Se-AdoMet). It is important to note that no such interaction was
detected between AdoMet and the cluster prior to cleavage. The large Mdssbauer shifi
observed here disfavors coordination of the sulfonium center to the cluster, however
AdoMet has several other functionalities that could coordinate the cluster, specifically the
amino and carboxylate groups of the methionine moiety, and the ribose hydroxyls. The
change in isomer shift observed upon addition of AdoMet is insufficient to differentiate

among these possibilities.

62

10.

ll.

12.

References

Sofia, H. J., Chen, G., Hetzler, B. G., Reyes-Spindola, J. F., and Miller, N. E.,
Nucleic Acids Res, 2001, 29, 1097-106.

Jarrett, J. T., Curr Opin Chem Biol, 2003, 7, 174-82.

Petrovich, R. M., Ruzicka, F. 1., Reed, G. H., and Frey, P. A., Biochemistry, 1992,
31,10774-10781.

Miller, 1. R., Busby, R. W., Jordan, S. W., Cheek, J., Henshaw, T. F., Ashley, G.
W., Broderick, J. B., Cronan, 1. E., and Marletta, M. A., Biochemistry, 2000, 39,
15166-15178.

Krebs, C., Henshaw, T. F ., Cheek, J., Huynh, B. H., and Broderick, J. B., J Am
Chem Soc, 2000, 122, 12497-12506.

Pierre], F ., Bjork, G. R., Fontecave, M., and Atta, M., J Biol Chem, 2002, 2 77,
13367-70.

Mulliez, E., F ontecave, M., Gaillard, J., and Reichard, P., J Biol Chem, 1993, 268,
2296-9.

Kent, T. A., Dreyer, 1. L., Kennedy, M. C., Huynh, B. H., Emptage, M. H.,
Beinert, H., and Munck, B, Proc Natl Acad Sci U S A, 1982, 79, 1096-100.

Emptage, M. 11., Kent, T. A., Kennedy, M. C., Beinert, H., and Munck, B, Proc
Natl Acad Sci U S A, 1983, 80, 4674-4678.

Fish, W. W., Methods in Enzymology, 1988, 158, 357-364.

Krebs, C., Broderick, W. E., Henshaw, T. F ., Broderick, J. B., and Huynh, B. H.,
J Am Chem Soc, 2002, 124, 912-913.

Cosper, N. J., Booker, S. 1., Ruzicka, F ., Frey, P. A., and Scott, R. A.,
Biochemistry, 2000, 39, 15668-15673.

63

CHAPTER IV

THE ACTIVE FORM OF THE CLUSTER: SINGLE TURNOVER

EXPERIMENTS

The results presented in this chapter have been published in Henshaw, T. F .; Check, 1.;

Broderick, J. B. J. Am. Chem. Soc. 2000, 122, 8331

64

1V.1 Introduction

The work in Chapter 11 detailed the ability of PF L-AE to coordinate a plethora of
iron-sulfur clusters and the work presented in Chapter 111 demonstrated that AdoMet
binds to a unique iron site in the [4Fe-4S] cluster. The work in this chapter determines
which cluster type is relevant in catalysis and identifies a redox role for the cluster. The
reductive cleavage of AdoMet represents a two-electron reduction, one of which is
supplied as a hydrogen atom that is abstracted from Gly—734 of PFL to generate the
glycyl radical. The source of the second electron has been difficult to conclusively
identify. Since iron-sulfur clusters are well know to play redox roles, it is reasonable to
propose that the cluster provides the additional electron. A primary goal of my graduate
research was to determine whether this is the case by allowing PF L-AE to complete a
well-characterized single turnover (1). This includes determination of the oxidation state
of the cluster before turnover, confirmation of the production of a stoichiometric amount
of product (the glycyl radical of PFL), and characterization of the oxidation state after
turnover. A reasonable proposal is that the [4Fe-4S]1+ cluster provides the reducing
equivalents for the generation of the putative 5’deoxyadenosyl radical and is oxidized to
the [4Fe-4S]2+ state during turnover (Scheme IV.1). At the time these results were
published, there were few reports that implicated the [4Fe-4S]1+ cluster in FeS/AdoMet
catalysis. Fontecave and coworkers had published a report that the rate of methionine
production was comparable to the rate of oxidation of the [4Fe-48] '1 cluster of the AR
activating enzyme, however, the stoichiometry they observed was inconsistent with the

reaction scheme (2). They observed three methionines produced for each cluster

65

oxidized or glycyl radical produced, calling their results into question. In addition, they
did not report any data that characterized the oxidation state of the cluster afier turnover
(2). Work with LAM showed that reduction to the [4Fe-4S]1+ oxidation state is required
for activity (3), but AdoMet is used as a cofactor rather than a substrate in the LAM

reaction and so stoichiometric cluster oxidation could not be demonstrated.

OOCWS\ CH3

 

+
'OOC S +H3N
W6 A H
+H3N O o: H A
O
OH OH
OH OH
S—Fe’X X 2+
Cys‘Ser/ CySI'I‘EE‘l7S
8 Fe
Cys \Cys

 

 

 

 

Scheme 1V.l Proposed redox role for the iron-sulfur cluster of PFL-AE

There were significant challenges to rigorously testing the hypothesis shown in
Scheme 1V.1. Even in the presence of a large excess of dithionite, EPR quantitation of
the signal from the reduced cluster accounted for only a fraction of the iron in the sample,
indicating that most of the cluster was not in the [4Fe-4S]1+ oxidation state. In addition,
excess dithionite could rereduce the cluster after turnover, making characterization of the
cluster after a single turnover impossible. Various attempts to remove excess reductant
resulted in loss of the [4Fe-4S]'+ EPR signal. Ti(111) was used as a reductant, and while
preliminary Mdssbauer results showed that the yield of [4Fe—4S]1+ increased, the S = ‘/2

EPR signal from excess reductant made characterization by EPR difficult.
66

The solution to these problems was twofold. First, the inclusion of 1 mM DTT in
all purification buffers allowed for the isolation of PFL-AB that contained primarily [4Fe-
4S]2+ clusters rather than [3F e-4S]+ clusters. The second was the use of 5-
deazariboflavin and light as reductant. I found that this method allowed for the
quantitative reduction to the [4Fe-4S] H oxidation state. In addition, since photoreduction
may be used to reduce the cluster to the [4F e-4S] '1, removal of excess reductant becomes

a trivial matter, as the samples may simply be stored in the dark.

IV.2 Methods

PFL-AE was expressed and purified as described in Chapter 11 with modifications. DTT
was included at a concentration of 1 mM in all purification buffers. In addition, cells
used for purification were grown in the modified MOPS medium containing natural
abundance iron. PFL was purified as described. Protein and iron assays were performed

as described in Chapter 11.

Preparation of single turnover samples

Samples were prepared in an anaerobic chamber at 0 °C using 200 uM PFL-AE
(2.65 F e/monomer) in 50 mM Hepes, 50 mM Tris (pH 7.4). 5-Deazariboflavin was added
in the dark to a final concentration of 100 uM. The samples were illuminated by a 500 W
halogen lamp for various times, followed by addition of AdoMet to 2 mM. In the dark,
each sample was split into two, and to one an equal volume of PFL solution (200 uM
PFL, 20 mM oxamate (allosteric effector), 100 uM 5-deazariboflavin) was added.

Samples were stored frozen in the dark until EPR spectra were recorded.

67

EPR Spectroscopy

EPR first-derivative spectra were obtained at X-band on a Bruker ESP300E
spectrometer equipped with a liquid He cryostat and a temperature controller from
Oxford Instruments. The sample spin concentrations were estimated by comparing the
double integral of the sample spectrum to that of a 1.04 mM K2(S03)2NO solution (for
the radical signals) or a 0.1 mM Cu(II) 1 mM EDTA solution (for the cluster signals)

recorded under identical conditions.

IV.2 Results and Discussion

As indicated by Scheme 1V.1, single turnover conditions for PFL-AB can be
achieved by limiting the amount of reductant. The clean conversion to [4Fe-4S]1+
provided by photoreduction has allowed us to carry out single turnover experiments for
glycyl radical production, since removing illumination eliminates the exogenous
reductant. PFL-AE was photoreduced for various times, after which a lO-fold excess of
AdoMet was added and the sample was wrapped in aluminum foil to prevent further
reduction. Each sample was then split into two halves and equimolar PFL was added to
one-half in the dark. EPR spectra were recorded to detect formation of [4Fe-4S]1+ and
glycyl radical in these samples. Figure 1A shows, from bottom to top, the 12 K EPR
spectra of PFL-AE photoreduced in the presence of 5-deazariboflavin for 0, 1, 2, 5, 10,

and 30 min. Quantitation of these EPR signals results in 0, 2.8(i0.5), 17(i2), 28(2t3),

68

 

 

 

 

444 :

 

 

 

 

 

 

 

 

1 W
0 .t..*v‘.fi _ r—‘w‘; T ‘f‘w ““
l I I I l I '— I
3400 3600 3300 3400

Field (gauss)

Figure 1V.l X-band EPR spectra of photoreduced PFL-AE before and after
addition of PFL.

Panel A: EPR spectra recorded after photoreduction of PFL-AE in the presence of 5-
deazariboflavin for the times indicated. Panel B: EPR spectra of the photoreduced PFL-
AE samples after addition of PFL. Conditions of measurement: T = 12 (A) or 60 K (B);
microwave power, 2 mW (A) or 20 uW (B); microwave frequency, 9.48 GHz;
modulation amplitude, 10.084 (A) or 5.054 G (B); single scan. Protein concentrations are
200 (A) or 100 uM (B). Reproduced with permission from (1). Copyright 2000
American Chemical Society.

 

 

 

 

 

 

 

 

 

60"
.. 1- 1' +
.3 40- t
(D ' +
23‘ 20— ‘1' 1+
q +[4Fe-4S]
+ I glyc
+
O I ' I ' I '
0 20 40 60

lllum. Time (min)

Figure IV.2 Spin quantitation of the EPR spectra shown in Figure IV.1A ([4Fe-
4S]“) and 1B (gly-) as a function of illumination time.

Included are additional data points for 20 min (51 d: 5 uM [4Fe-4S]'+ and 48 :I: 5 uM
gly-) and 60 min (56 a 6 [4Fe-4S]'+ and 57 a 6 uM gly°) illumination. Reproduced with
permission from (1). Copyright 2000 American Chemical Society.

70

41(i4), and 54(15) uM spins, respectively. The nearly axial EPR signals shown in
Figure 1V. 1A are characteristic of a [4Fe-4S]1+ cluster, and are essentially identical to the
EPR signal previously reported for dithionite-reduced PFL-AB in the presence of
AdoMet (4). After 60 min of illumination, 85% of the cluster in PFL-AE is in the
reduced [4Fe-4S] '1 state. Saturation of cluster reduction is indicated by the illumination
time course shown in Figure 2.

EPR spectra (60 K) for samples with PFL added are shown in Figure 1V.lB.
Increasing amounts of a multiplet EPR signal characteristic of the PFL glycyl radical are
observed with increasing time. Spin quantitation of the glycyl radical EPR signals at each
time point show a 1:1 correspondence between the amount of glycyl radical observed and
the amount of [4Fe-4S] 1+ cluster present prior to addition of PFL, as shown in Figure
IV.2. The glycyl radical spin quantitations are 3.6(i0.5), 16(i2), 28(i3), 36(:l:4), and
52(15) uM for 1, 2, 5, 10, and 30 min illumination, respectively.

The glycyl radical spectra were recorded at 60 K, a temperature at which the
[4Fe-48]'+ signal of PFL-AB is not observable. However, EPR spectra recorded at 12 K
for the same samples also showed no [4Fe-4S]1+ signal. This observation demonstrates
that the [4F e-4S]l+ cluster has been converted to an EPR-silent state upon addition of
PFL to the [4F e-4S]‘+/PFL-AE/AdoMet and subsequent generation of the glycyl radical.
Cleavage of AdoMet is stoichiometric with PF L glycyl radical generation, and requires a
source of electrons. Our results strongly suggest that the required electrons come from
the [4Fe-4S] H cluster, thereby converting it to an EPR-silent [4Fe-4S]2+ cluster. This

conclusion is supported by the observation of a UV-vis spectrum which is typical of a

71

[4Fe-4S]2+ cluster. In addition, the increase in 8400 upon addition of PFL is consistent
with oxidation of the [4Fe-4S]1+ to a [4Fe-4S]2+ cluster. Furthermore, re-illumination of
samples containing the EPR-silent cluster can regenerate the [4F e-4S] 1+ EPR signal (data
not Shown), indicating an ability to cycle readily between the [4Fe-4S]1+ and [4Fe-48]2+
states.

The results presented in this chapter demonstrate that the [4Fe-4S]1+ cluster of
PF L-AE generated in the presence of AdoMet is sufficient to generate the glycyl radical
of PFL and is oxidized during turnover to the [4F e-4S]2+ oxidation state. The results
presented in Chapter III demonstrate that AdoMet binds directly to the cluster. As such,
we are now prepared to propose a mechanism of radical generation that is consistent with
the results presented in this dissertation (Scheme IV.2). In this mechanism AdoMet
chelates the unique iron site of the [4F e-4S]2+ cluster, accounting for the observed
changes in the Mossbauer parameters of the unique site and positioning the sulfonium
group in close proximity to one of the bridging sulfides. This binding mode was
demonstrated by later ENDOR experiments (5, 6). Following reduction to the [4F e-4S]1+
state, the sulfide-sulfonium interaction facilitates an inner-sphere electron transfer
promoting homolytic cleavage of the sulfonium-carbon bond. This generates the 5’-
deoxyadenosyl radical, which abstracts a hydrogen atom from gly-734 of PFL to generate

the glycyl radical.

72

 

Fe\ls

 

s/ Ife\s. ’

LS I>i|=e”\

\ S, Fe\

I

”2
‘\\ N

 

 

2+

2+

PFL-gly-
+ dAdo

PFL-gly

Scheme IV.2 Proposed Mechanism of PFL-AE

73

 

 

 

 

1+

2+

+ dAdo-

CHAPTER V

OXIDITAVE DNA REPAIR

The results presented in this chapter have been published in Trewick, S. C., Henshaw, T.
F., Hausinger, R. P., Lindahl, T., and Sedgwick, B., Nature, 2002, 419, 174-178.

My contribution to this work did not include radioactivity or HPLC-based assays.

74

V.l Introduction

DNA alkylating agents occur endogenously, are present in the environment and are used
in chemotherapy (1, 2). Escherichia coli exposed to these compounds respond by
inducing the expression of four genes, ada, alkA, aidB and alkB. Ada is an 06-
methylguanine-DNA methyltransferase and regulates this adaptive response. AlkA is a
3-methyladenine-DNA glycosylase, and AidB is proposed to destroy certain alkylating
agents (3). AlkB is conserved from bacteria to mammals, but its role has not been
resolved despite the early isolation of an E. coli alkB mutant (4). Expression of E. coli
alkB confers alkylation resistance to human cells (5), and conversely, a human
homologue has been reported to convey methyl methanesulphonate (MMS) resistance to
the E. coli mutant (6). AlkB processes the cytotoxic DNA damage generated in single-
stranded DNA by SN2 methylating agents, such as MMS, dimethylsulphate (DMS) and
methyl iodide (7). l-Methyladenine (lme-A) and 3-methylcytosine (3-meC) are
predominant forms of base damage in single-stranded DNA because the sites of
methylation are normally protected by base pairing (8, 9). These lesions are cytotoxic
because they stall DNA replication, and are not removed by known DNA repair
pathways. 1-Methyladenine and 3-methylcytosine in DNA were previously proposed as
candidate substrates of AlkB, nevertheless, many attempts to develop assays for this
enzyme were unsuccessful (10). Theoretical sequence profile and fold recognition
searches suggested that AlkB an a-ketoglutarate- and Fe(II)-dependent dioxygenase (11).
The results in this chapter demonstrate DNA repair activity for AlkB using DNA
substrates containing l-methyladenine or 3-methylcytosine, a-ketoglutarate as a co-

substrate, and Fe(II) as a cofactor.

75

V.2 Experimental Methods

Expression and purification of A lkB

AlkB with an N-terrninal was purified as previously described (10) with the
inclusion of 1 mM EDTA in purification buffers. EDTA was removed following
purification by extensive dialysis against 50 mM HEPES, 100 mM NaCl, 5 % glycerol

(w/v) pH 8.

Radioactive substrates and assays

A total of 0.6 mg poly(dA) (average length 310 residues; Amersham Biosciences)
was treated with 1 mCi [14C]methyl iodide (58 mCi mmol") (Amersham Biosciences) in
0.8 ml of 10 mM sodium cacodylate, pH 7, at 30 °C for 6 h. The methylated polymer was
precipitated in ethanol, dissolved in 10 mM Tris-HCl, pH 8, and had a specific activity of
1,860 counts per min (c.p.m.) per ug. To prepare a double-stranded substrate, [14C]-
labelled methylated poly(dA) was annealed to poly(dT) at 20 °C. A total of 1.2 mg
poly(dC) (average length 370 residues) in 1.3 ml of 50 mM HEPES-KOH, pH 8, was
similarly treated with [”C]methyl iodide to yield a specific activity of 700 c.p.m. pg". A
plasmid coding for Flag-tagged AlkB was constructed by digestion of pBAR54 (10) with
NdeI and NcoI to remove DNA coding for the His tag. This DNA fragment was replaced
by a double-stranded oligonucleotide coding for the Flag tag. The Flag-tagged protein
was purified by immnunoaffinity chromatography (Sigma). Purified AlkB was incubated
with the [14C]methyl iodide-treated substrates in 50 mM HEPES-KOH, pH 8, 75 uM

76

Fe(NH4)2(SOa)2 °6H20, 1 mM a-ketoglutarate, 2 mM ascorbate, and 50 ug ml”I BSA for
15 min at 37 °C. The reaction was stopped by adding 11 mM EDTA, and the substrate
was precipitated in ethanol in the presence of carrier calf thymus DNA. Ethanol-soluble

radioactive material was monitored by scintillation counting.

HPLC assays

Methylated adenine residues were released from [HQ-methylated poly(dA) by
hydrolysis in 0.1 M HCl at 95°C for 1 h, and 3-meC from methylated poly(dC) by
treatment with 90% formic acid at 180°C for 20 min. Methylated bases were analyzed by
HPLC on a Whatman Partisil 10 cation exchange column in 0.1 M ammonium formate,
pH 3.6. A gradient of MeOH from 20 to 40% was applied to separate the methylated

adenine derivatives, and a gradient from 5 to 40% MeOH was used to analyze 3-meC.

High-level methylation of an oligionucleotide

A 41-mer oligonucleotide, TT'I'I‘TT(AT'IT1TT)5, was treated with 50 mM DMS
in 75 mM sodium cacodylate, pH 7.4, at 30°C, 2 times for 2 h and then 4 times for 1 h.
Between each treatment the DMS was removed by centrifirgation through a G25
Sephadex mini-column equilibrated in the same buffer. The level of methylation of
adenine residues was examined by acid hydrolysis, HPLC and A260 measurements. The
A260 ratio of adenine to l-meA was 1.04 which was determined by monitoring known

amounts of these purines in the same conditions.

77

U V-visible spectroscopy

Samples containing AlkB (0.35 mM) and a-KG (1.2 mM) in 50 mM HEPES, 100
mM NaCl, 5 % glycerol (w/v) pH 8 were sealed with a rubber septum in a 1 cm
pathlength quartz cuvette. The enzyme was made anaerobic by several rounds of
evacuating and refilling the cuvette with argon. A background spectrum was recorded
and ferrous ammonium sulfate was added (using a gas-tight Hamilton syringe) to a final
concentration of 0.3 mM from an anaerobic stock and the UV-vis spectrum was again

recorded.

Gas chromatography/mass spectrometry

Oligo(dA) (25-mer) was treated with 500 mM MMS at 30°C for 30 min and,
after removal of the MMS, annealed to Oligo(dT). AlkB (38 pM) was incubated with this
substrate (50 uM) in 50 mM HEPES-KOH, pH 8, containing 1 mM orKG, 75 uM
Fe(NHa)2(S04)2'6H20, and 2 mM ascorbate in a final volume of 200 pl. After 10 min at
30°C, 200 pl 15 mM PFPH in 1.2 M phosphoric acid was added, and incubation
continued at 50°C for 2 h. The reaction mixtures were extracted with 100 pl of 85:15

hexanezdichloromethane, and 20 ul was analyzed by GC/MS as described (12).

Oxygen consumption

50 mM HEPES, pH 8, with 1 mM aKG, 75 uM Fe(NHa)2(SO4)2'6H20, 100 uM
ascorbate, and 50 uM oligonucleotide substrate, prepared as described for GC/MS, were

mixed for 3 min in a YSI 5 300 oxygen electrode at 30°C to allow equilibration with

78

atmospheric oxygen. The electrode was precalibrated with air saturated water (236 uM

02). AlkB was added through a gas tight syringe to a final concentration of 9 uM.

V.3 Results and Discussion

Release of ethanol-soluble radioactivity

A substrate containing l4C-labelled methylated adenine residues was prepared by
treating poly(dA) with [1°C]methyl iodide. 1n the presence of a-ketoglutarate and Fe(II),
purified His-tagged AlkB protein released ethanol-soluble radioactive material from this
methylated substrate (Figure V. 1 A). Similar activity was also observed with Flag-tagged
AlkB protein (data not shown). The activity was dependent on ot-ketoglutarate, inhibited
by EDTA and stimulated by ascorbate (Figure V.lB). Inhibition by EDTA was
overcome by adding an excess of Fe(NH4)2(SO4)206H20, thus demonstrating a
requirement for F e(II). At lOO-fold higher AlkB concentrations, the effect of ascorbate
was reduced. To determine whether AlkB could also act on double-stranded DNA, the
[”C]-labelled methylated poly(dA) was annealed to poly(dT). AlkB was approximately
threefold more active on the double- compared with the single-stranded substrate (Figure

v.1C).

U V-visible spectroscopy

Figure V. 1 D shows the visible difference spectrum of an anaerobic AlkB sample
in the presence of F e(II) and a-KG which exhibits a weak chromophore at 500 nm and is
attributed to a metal to ligand charge transfer transition that is characteristic of the

Fe(II)/0KG dioxygenases (13-15). This chromophore was not observed in the absence of

79

8

 
 
 
  
  
 

 

 

 

 

 

 

 

 

I o
g soot g poly(dA)-pawn
a. .9150.
§ § 100.
'3 1004 '3 50]
g and»
I.“
0t. . f . . - - .
0 50 100 150 200 O 50 100 150
Mallard) AIkBamol)
I! ActMtyM)
Complete 100
NoAIkB <05
Nan-ketoglutarate <05
Noaaoorbate <0.5
+ 0.2 mM EDTA 1.1
+ 02 mM EDTA + Fe“ 0.7 mM 99
d 0.08‘I
0.06-
0.024

 

 

300 400 500 000 Too
Wavelengthmm)

Figure V.1 Release of ethanol-soluble material from methylated poly(dA) by AlkB

in the presence of Fe(II) and a-ketoglutarate.

a, [14C]Methyl iodide-treated poly(dA) (1,200 c.p.m.) was incubated with AlkB in the
complete reaction mixture, and the release of radioactive material was monitored. b,
Requirements for AlkB activity (2 pmol AlkB). c, Comparison of AlkB activity on [14C]-
labelled methylated poly(dA) (1,200 c.p.m.) (open circles) and afier annealing to
poly(dT) (filled circles), assayed at 20 °C for 30 min to maintain stable duplexes. d,
Difference absorption Spectrum of anaerobic AlkB (0.35 mM) in the presence of u-
ketoglutarate (1 mM) and Fe(II) (0.3 mM) minus the spectrum without Fe(II).

80

Figure V.2 Repair of l-methyladenine (l-meA) and 3-methylcytosine (3-meC) by
AlkB.

a, b, [”C]Methyl iodide-treated poly(dA) (a) or poly(dC) (b) were incubated without
AlkB (triangles) or with 2.5 pmol AlkB (squares). The [14C]-labelled methylated bases
remaining in the substrates were analysed by HPLC and scintillation counting. 3-meA
and 7-meA, 3- and 7-methlyadenine, respectively. c, Direct reversion of l-methyladenine
to adenine by AlkB. A thymine-rich oligodeoxynucleotide containing five adenine
residues was heavily methylated and then incubated without (top) or with (bottom)

920 pmol AlkB at 37 °C for 30 min. The bases present in the oligonucleotide were
analysed by HPLC and A260 measurements. The early eluting peaks were Oligo(dT)
fragments.

81

S

J

8

1

Methylated m (c.p.m.) I

1 -maA

+NoAIkB
-{}-+Alk8

 

 

 

 

 

l

 

 

   

Lurll'l'l'L'!

 

 

 

 

 

 

 

 

 

 

 

 

o I 1 r I
0 25 50 75 100
Fraction number
0
0.020 No AlkB
0.010
«a 0.012
0.008 "m“
Adenine
0.004
7-meA _
0 2 3 meA
0 5 10 15 20 25 30
Time (mln)
0.020 Adenlne + AlkB
0.016
3 0.012
0.008
01134 1-maA
T-NIBA 3411“
0
0 5 10 15 20 25 30
Tlme (mini

82

a-KG or iron and provides strong evidence that AlkB coordinates Fe(II) in a 2-His-1-Asp

facial triad (16) and that Ill-KG chelates the iron.

Identification of I-meA and 3-meC as AlkB substrates

To examine which methylated adenines were processed by AlkB, the poly(dA)
substrate was incubated with AlkB, acid hydrolysed, and analysed by high-performance
liquid chromatography (HPLC). In the poly(dA) polymer treated with [”C]methyl
iodide, l-methyladenine was the principal methylated base, 3-methyladenine was also
abundant, but 7-methyladenine was a minor product. After incubation with AlkB, the
levels of l-methyladenine were reduced whereas the amounts of 3-methyladenine and 7-
methyladenine remained unchanged (Figure V.2A). Thus, AlkB specifically catalyzed
removal of l-methyladenine from methylated poly(dA). From the data in Figure V.lA,
0.1 pmol AlkB protein removed 1.7 pmol l-methyladenine from methylated poly(dA) in
15 min, indicating that AlkB acts enzymatically and is not consumed during the reaction.
This distinguishes its mode of action from O°-methylguanine-DNA methyltransferase.

A second modification that is formed to a greater extent in single- compared with
double-stranded DNA is 3-methylcytosine. This lesion was also considered as a
candidate substrate of AlkB. AlkB protein released radioactive material from poly(dC)
treated with [”C]methyl iodide (data not shown). HPLC analysis showed that 3-
methylcytosine was the only detectable modified base in this substrate, and that it
disappeared on incubation with AlkB (Figure V.2B). We have therefore identified two
substrates of AlkB, l-methyladenine and 3-methylcytosine, that are both generated in

single-stranded DNA on treatment with 8N2 methylating agents.

83

It is proposed that AlkB repairs l-meA and 3-meC in DNA by oxidative
demethylation. In such a mechanism, the lesions would be reverted to adenine and
cytosine, formaldehyde would be generated and 02 consumed. To demonstrate that AlkB
directly reverts l-meA in DNA to adenine residues, a non-radioactive substrate was
prepared in which 76% of the adenine residues were methylated to form l-meA. This
was achieved by repeated treatments with DMS of an oligonucleotide containing adenine
residues interspersed between inefficiently methylated thymine residues. Only 4% of the
adenines were recovered as 3-meA, and 2% were 7-meA (Figure V.2C). The low amount
of 3-meA might be caused by instability of the glycosyl bond and loss of this
modification during the extensive DMS treatments. The heavily methylated substrate
was incubated with AlkB in the optimized assay conditions, the DNA hydrolyzed and
individual bases quantified by HPLC and A260 measurements. In the presence of AlkB, a
decrease in the amount of l-meA correlated with a stoichiometric increase in the amount
of adenine recovered (Figure V.2C). It is concluded conclude that AlkB converts l-meA

directly to adenine in DNA.

Formaldehyde release and oxygen consumption

To determine whether formaldehyde was a reaction product, AlkB was incubated
with MMS-treated poly(dA) annealed to poly(dT), the products were derivatised with
pentafluorophenylhydrazine (PFPH) and analysed by gas chromatography (GC). One
derivatised product, as indicated by the arrow in Figure V.3A, arose only when the DNA
substrate was methylated and F e(11) and orKG were present (Figure V.3B), and had a
mass spectrum (MS) identical to that previously described for the HCHO-PFPH adduct

(12)(Figure V.3C). The release of HCHO was also monitored by a coupled
84

 

 

 

 

 

 

 

182
8 l g
g g 210
a § ‘
3
4 5 6 7 8 9 ' 80 160 4' 240
Tune (min) m/z
b d 0 Non-methylated

WY
:[1096 02

1 min
I--I

 

Abundance

 

4 5 6 7 8 9
Tlmelmin)

 

Figure V.3 Production of formaldehyde and consumption of Oz by AlkB.

Methylated Oligo(dA)-Oligo(dT) was incubated with AlkB. The products were derivatized
with PFPH and analysed by gas chromatography mass spectrometry. Gas
chromatography traces are shown of derivatized products generated when the DNA was
methylated (a) or not methylated (b). c, Mass spectrum of the product indicated by the
arrow in a. (1, Consumption of 02 during incubation of AlkB with either methylated or
non-methylated Oligo(dA)-Oligo(dT) was determined by using an oxygen electrode. AlkB
(9 uM) was added at the point indicated by the arrow.

85

spectrophotometric assay using formaldehyde dehydrogenase (17), and again was
detected only in complete assay conditions with the heavily methylated DNA substrate
(tumover number 1.5 3", data not shown). A yield of 140 uM HCHO correlated
stoichiometrically with the amount of l-meA in the DNA substrate (7-10% of the
adenines), the amount of Oz consumed (Figure V.3D) and the succinate generated (data
not shown). This stoichiometric relationship further verifies the proposal that AlkB is an
aKG-dependent dioxygenase. In the absence of methylated DNA, an observed slow
consumption of oxygen (Figure V.3D) and orKG was consistent with a partial uncoupling
of alKG decomposition and hydroxylation of the methylated DNA bases. Such
uncoupling is a well-known property of this family of enzymes (18, 19).

Oxidative demethylation is an unprecedented mechanism of DNA repair. The
proposed reaction mechanism by which AlkB repairs 1-meA and 3-meC in DNA is
shown in Scheme V. 1. Due to the stability of the N-C bond in l-meA and 3-meC,
demethylation by hydrolysis would be energetically unfavorable; consequently oxidative
demethylation by reactive iron-oxygen species is required. Direct reversal of this base
damage to unsubstituted parent residues would be a highly accurate form of DNA repair
and agrees with in viva observations that repair by AlkB is non-mutagenic (10). E. coli
alkB mutants are more sensitive to alkylating agents during active growth than in
stationary phase probably because l-meA and 3-meC are produced in single-stranded
regions of DNA in replication forks and transcription bubbles (10). DNA unfolds only
. transiently during replication and transcription, so it is beneficial that AlkB repairs its

substrates not only in single-strands but also, and even more efficiently, after DNA

86

reannealing (Figure V. l ). Interestingly, Caulabacter crescentus alkB expression is cell

cycle regulated with a pattern similar to activities required for DNA replication (20).

NH2 NH2 CH OH MHZ
, H ’ 2
Q otKG succinate N N/) < 21
N N o CO I N N

dR :2 :v 2 GR dR

 

NH NH
2 CH \ ,CHZOH HCHO 2
\N’ 3 N \

I vs ' Us I I;

N o '1 0 N o

dR dR dR

Scheme V.1 Oxidative DNA repair mechanism catalyzed by AlkB

We have extended the family of alKG-Fe(11) dependent dioxygenases to include
AlkB, as recently proposed on theoretical grounds (11). These enzymes catalyse a
variety of reactions including hydroxylations, desaturations and oxidative ring closures,
and account for oxidation of proline in collagen, steps in the biosynthesis of several
antibiotics and cellular metabolites, as well as biodegradation of selected compounds
(21). No other members of this family are presently known to act on DNA, but it is
notable that a fungal enzyme, thymine-7-hydroxylase, oxidises the methyl group of free

thymine (22, 23).

87

10.

ll.

12.

13.

14.

15.

16.
17.

References

Rebeck, G. W. and Samson, L., Journal of Bacteriology, 1991, 1 73, 2068-2076.

Vaughan, P., Sedgwick, B., Hall, J ., Gannon, J ., and Lindahl, T., Carcinogenesis,
1991, 12, 263-268.

Lindahl, T., Sedgwick, B., Sekiguchi, M., and Nakabeppu, Y., Annu Rev
Biochem, 1988, 57, 133-57.

Kataoka, H., Yamamoto, Y., and Sekiguchi, M., Journal of Bacteriology, 1983,
153, 1301-1307.

Chen, B. R. J ., Carroll, P., and Samson, L., Journal of Bacteriology, 1994, 1 76,
6255-6261.

Wei, Y. P., Carter, K. C., Wang, R. P., and Shell, B. K., Nucleic Acids Research,
1996, 24, 931-937.

Dinglay, 8., Gold, B., and Sedgwick, B., Mutation Research-DNA Repair, 1998,
407, 109-116.

Singer, B. and D., 6., Molecular Biology of Mutagens and Carcinogens:
Reactions of Directly Acting Agents with Nucleic Acids. 1983, New York: Plenum
Press.

Bodell, W. J. and Singer, B., Biochemistry, 1979, 18, 2860-2863.

Dinglay, S., Trewick, S. C., Lindahl, T., and Sedgwick, B., Genes &
Development, 2000, 14, 2097-2105.

Aravind, L. and Koonin, E. V., Genome Biol, 2001, 2, RESEARCHOOO7.

Heck, H. D., White, B. L., and Casanovaschmitz, M., Biomedical Mass
Spectrometry, 1982, 9, 347-353.

Ryle, M. J ., Padmakumar, R., and Hausinger, R. P., Biochemistry, 1999, 38,
15278-86.

Hegg, E. R., Ho, R. Y. N., and Que, L., Jr., Journal of the American Chemical
Society, 1999, 121, 1972-1973.

Pavel, E. 6., Zhou, J ., Busby, R. W., Gunsior, M., Townsend, C. A., and
Solomon, E. 1., Journal of the American Chemical Society, 1998, 120, 743-753.

Hegg, E. L. and Que, L., Jr., Eur J Biochem, 1997, 250, 625-9.

Lizcano, J. M., Unzeta, M., and Tipton, K. F., Anal Biochem, 2000, 286, 75-9.
88

18.

19.

20.
21.
22.
23.

Myllyla, R., Majamaa, K., Gunzler, V., Hanauske-Abel, H. M., and Kivirikko, K.
I., J Biol Chem, 1984, 259, 5403-5.

de Jong, L., Albracht, S. P., and Kemp, A., Biochim Biophys Acta, 1982, 704,
326-32.

Colombi, D. and Gomes, S. L., Journal of Bacteriology, 1997, 1 79, 3139-3145.
Hausinger, R. P., Crit Rev Biochem Mol Biol, 2004, 39, 21-68.
Thomburg, L. D. and Stubbe, J ., Biochemistry, 1993, 32, 14034-42.

Thomburg, L. D., Lai, M. T., Wishnok, J. S., and Stubbe, J ., Biochemistry, 1993,
32, 14023-33.

89

CHAPTER VI

ABBERANT ACTIVITY OF THE DNA REPAIR ENZYME ALKB

The results presented in this chapter have been published in Henshaw, T. F., Feig, M.,
and Hausinger, R. P., J Inorg Biochem, 2004, 98, 856-61.

90

V1.1 Introduction

AlkB and its human homologues belong to a family of enzymes known as the
iron(II)- and alpha-ketoglutarate (ocKG)-dependent dioxygenases (1). Like other family
members, they activate dioxygen at a redox-active mononuclear center and abstract a
hydrogen atom from a chemically non-reactive C-H bond to initiate transformations that
can include hydroxylations, ring closures, desaturations, and other reactions. An
additional key feature of most of these enzymes is their ability to couple the oxidative
decarboxylation of aKG to oxygen reduction; thus generating a putative Fe(IV)-oxo
intermediate. Evidence for the production of an F e(IV) intermediate was obtained in
another family member, taurine/aKG dioxygenase (TauD) (2-4), and model studies have
demonstrated that Fe(IV)-oxo species can be generated at a mononuclear site (5). Also
consistent with the intermediacy of an Fe(IV)-oxo species is the finding of enzyme self-
hydroxylation reactions in other family members (6-8). For example, in the presence of
aKG, but absence of taurine, TauD oxidizes Tyr 73 located near the active site to form a
tyrosyl radical that converts to a catechol (8). Coordination of the catecholate side chain
to the oxidized metal site creates greenish-brown (7tmax 550 run; 8550 ~700 M'lcm'l) or
green (km, 700 nm; 8700 ~3 80 M'lcm'l) ligand-to-metal charge-transfer (LMCT)
chromophores, depending on whether bicarbonate is present (6). Similarly, in the
absence of its primary substrate 2,4-dichlorophenoxyacetic acid (2,4-D), but presence of
ocKG, the 2,4-D/aKG dioxygenase deA hydroxylates Trp 113, a residue located
adjacent to a metal ligand of this protein (7). Coordination of the hydroxytryptophan

(OH-Trp) to the oxidized metallocenter yields a blue chromophore (km, 580 nm; 8530

91

~1000 M'lcm'l). These self-hydroxylation reactions are likely to arise from side
reactions of the highly activated F e(IV)-oxo species.

In this chapter, I provide evidence that AlkB also catalyzes self-oxidation of an
amino acid side chain when incubated with Fe(II), aKG, and oxygen in the absence of its
primary substrate, methylated DNA/RNA. This hydroxylated residue coordinates the
oxidized metal site to produce a blue chromophore, analogous to that observed in deA.
On the basis of mass spectrometric evidence, we identify the modification as
hydroxylated Trp 178. Modeling of the AlkB structure provided evidence that Trp 178 is
appropriately positioned near the metallocenter to undergo the observed oxidation

chemistry.

V1.2 Experimental

Sample Preparation

His-tagged AlkB was overproduced in E. coli BL21(DE3) [pBAR54] and purified
as described in Chapter VI. Significantly, cell disruption was carried out in buffers
containing ethylenediaminetetraacetic acid (EDTA) to chelate the Fe(II) that is weakly
bound to the enzyme and minimize any oxidative damage to the protein sample during

purification.

UV-visible (U V-vis) spectroscopy

A cuvette containing 500 pl of AlkB (240 uM protein in 50 mM Hepes buffer, pH
8.0, with 300 mM NaCl) was made anaerobic by several rounds of vacuum and argon on

an anaerobic Schlenk line. Ferrous ammonium sulfate was added from an anaerobic

92

stock (10 mM) to a final concentration of 200 uM. aKG was added to a concentration of
1 mM. The enzyme was oxidized at room temperature (~23° C) by the addition of 300
uL of water that had been saturated with 02 at atmospheric pressure. The headspace of
the cuvette was replaced with pure 02. UV-vis spectra were recorded at times indicated

in the figure using a Beckman DU 7500 diode array spectrophotometer.

Mass Spectrometry (MS)

MS samples were run by Brett Phinney of the MSU Mass Spectrometry Facility.

AlkB samples were extensively dialyzed, diluted to a concentration of l nmole in
50 ul of 100 mM ammonium bicarbonate buffer (pH 8.0), digested overnight with
sequencing-grade modified trypsin (Promega) (1:40 molar ratio) at 37°C, and analyzed
by nanoscale liquid chromatography mass spectrometry/mass spectrometry (LC/MS/MS).
500 frnole aliquots were applied to a picofrit column (75 um id. x 5 cm) packed with 3
pm Magic C-18 material (Micron Bioresources). Mobile phase A consisted of 0.1%
formic acid while mobile phase B consisted of 0. 1 % formic acid in 90/10
acetonitrile/water. Peptides were eluted during a 25 min gradient extending from 2% B
to 60% B. The picofrit column made electrical contact through a pre-column ZDV
titanium union, terminating in a 8 pm tip i.d. outlet spray needle. Mass spectra were
acquired by data dependent analysis in a Micromass Q-TOF Ultima mass spectrometer,
where peptide ions detected in the MS survey scans triggered MS/MS fragmentation for
obtaining product ion spectra. In order to identify peptides in the digest, the Mascot

search engine was used to search the AlkB sequence.

93

Structural Modeling

Structural modeling was done by Michael Feig of the MSU Physics Department.

The structure for E. coli AlkB was predicted based on fold recognition in a
functionally related oxidoreductase (deacetoxycephalosporin C synthase from
Streptomyces clavuligerus) where the structure is available from crystallography (PDB
codes: lDCS and 1RXG). This structure was identified as the best template with the
Bioinfo Meta Server (http://bioinfo.pl/l\/Ieta) (9). The alignment of secondary structure
elements, as generated with the ORFeus method (10), was used to generate a structure
scaffold by comparative modeling, where side chains in the template structure are
replaced with the corresponding amino acids in E. coli AlkB. Connecting loops and other
missing fragments were then added with the MMTSB Tool Set (11) by sampling
conformational space with a low-resolution lattice model (12) and selecting the most
favorable conformation based on clustering and an all-atom energy scoring function (13).
The position of the aKG ligand and Fe(II) were copied from the 1RXG crystal structure
where the ligands were present. The complete structure including the ligand and iron was
then minimized for 1000 steps with the molecular mechanics program CHARMM (14)

under harmonic restraints to keep backbone and CB atoms near their initial positions.

The ligand and iron were fixed during the minimization.

94

 

 

V1.3 Results and Discussion l.

Chromophore Formation

The addition of ocKG and Fe(II) to AlkB under anaerobic conditions generates a
weak metal-to-ligand charge-transfer (MLCT) absorption at 450-500 nm (Chapter V and
(15)). When mixed with oxygen-saturated water, the pink color associated with this
species disappeared and was replaced by a more intense blue chromophore (3590 960 M'
1cm'1 on the basis of iron concentration) with an absorption centered at 595 nm (Figure
V1.1). The slow color transformation occurred in two phases, with the relatively rapid
disappearance of the orKG/Fe(II) chromophore (note the decrease in absorption at ~500
nm) preceding the increase of the 595 nm absorption. Full development of the blue
chromophore required more than 10 min and followed an approximate first-order process
(0.36 i 0.05 min'1 when measured starting at 2 min). Incubation of the sample under
vacuum and exchange with an anaerobic atmosphere did not affect the blue color.
Addition of dithionite to reduce the sample led to protein precipitation. In contrast to the
situation for orKG/Fe(II)AlkB, no color was generated when Fe(II)AlkB lacking a-keto
acid was mixed with oxygen-saturated water. The behavior observed here for AlkB and
the resulting 595 nm spectrum are very similar to what was previously reported for deA,

where the chromophore was attributed to the development of a LMCT transition

involving OH-Trp coordinated to Fe(III) (7).

95

 

Mass spectrometric analysis i -

A hydroxylated Trp side chain of AlkB was identified by nanoscale capillary
LC/MS/MS analysis. In this analysis, AlkB samples were first extensively dialyzed,
diluted into ammonium bicarbonate (a volatile buffer), and digested with trypsin before
injection. This method allows for very accurate and sensitive determinations of post-
translationally modified amino acids within a protein. Using this approach, clear
evidence was obtained for hydroxylation of a Trp side chain at position 178 (with

numbering on the basis of the translated gene sequence). Product ion spectra revealed

96

 

0.14
- “ o

0.00 ~ 9

0.06 ~

0.04 - g

0.02 ~ 0

o I I I l
0 5 10 15 20 25

Absorbance 590 nm

 

 

 

Time (min)

 

Absorbance

 

 

 

 

300 400 500 600 700 800

Figure V1.1 UV-visible spectroscopic analysis of the reaction of aKG-FeaDAlkB
with oxygen.

The Fe(II)- and orKG-bound form of AlkB was mixed with 0.2-saturated water and
spectra were recorded as a function of time. Panel A shows the loss of the MLCT
absorption near 500 nm associated with the Fe(II)/aKG center and the development of an
intense visible chromophore centered at 595 nm. The kinetics of the spectral changes at

590 nm are illustrated in Panel B.

97

 

- Y(5)

 

 

 

 

 

 

 

 

. §
:21 5???
> +'
g i' 3;:
' a i "5% g
[lllllnllll I. II II. I l
' 200 000 1000 '

 

Figure V1.2 LC/MS/MS identification of hydroxy-Trp 178 in AlkB.

Tryptic peptides derived from oxygen-exposed aKG-Fe(II)AlkB were resolved by HPLC
and analyzed by tandem MS/MS. The figure depicts the observed secondary MS product
ions (mono- and di-protonated) derived from fragmentation of a parent ion (mass of
1780.91) associated with a single tryptic peptide. B ions are amino-terminal fragments
and Y ions include the carboxyl terminal fragments. The parent ion and product ion
masses match the sequence LLLEHGDVVVWGGESR (residues 168-183 of AlkB) with
a mass addition of 16 Da starting on the B11 and Y6 product ions. This 16 Da mass
addition was consistent with the presence of hydroxy-Trp at residue 178.

98

that all fragments containing Trp 178 were shifted by 16 Da and fragments not containing
Trp 178 were unshified (Figure V1.2). Surprisingly, similar analyses revealed the
presence of the same modification in a tryptic peptide of the control sample. This result
is consistent with at least a portion of the overproduced AlkB undergoing modification
during cell growth. We interpret the finding of Trp 178 hydroxylation as evidence that an
Fe(IV)-oxo, or similarly reactive species, is generated in the reaction of aKG-Fe(II)AlkB
with 02 in the absence of methylated DNA substrate. Oxidation of the nearby Trp
initiates the eventual formation of the hydroxy-Trp feature observed in the mass spectral

analysis.

Protein Modeling

An overview of the predicted structure for E. coli AlkB is shown in Figure V1.3A.
It features the same double-stranded beta-helix and jelly-roll topology as the
deacetoxycephalosporin C synthase templates used to build the model. Residues 1-15 are
not shown since the prediction of the amino terminus is considered less reliable for the
lack of a suitable template. The closeup View shown in Figure VI.3B focuses on the
immediate environment of the Fe(II) binding site. As expected, His 131, Asp 133, and
His 187 are in direct interaction with the Fe(II). The model also shows Trp 178 located
above His 187 and in the immediate vicinity. In fact, the distance between the Trp
carbon at the 5 position, where hydroxylation is most likely, and the Fe(II) ion is only 4.7
A in the predicted model after force-field based minimization. This compares favorably
with the geometry in TauD where the distance between F e(II) to Tyr 73, the tyrosine that

becomes hydroxylated, is 6.5 A (8).

99

 

 

Figure V1.3 Structure of AlkB determined by protein modeling.

Panel A depicts the predicted fold of the monomer and reveals a B-jellyroll structure as
found in other family members. Residues 1-15 are not included since their positions
could not be reliably predicted. Panel B focuses on the region near the metallocenter and
shows the three amino acid side chains that serve as ligands, the aKG, and the nearby
Trp 178 that undergoes hydroxylation.

100

 

Conclusions

When exposed to oxygen, aKG-Fe(II)AlkB forms a blue chromophore that
closely resembles the spectrum of oxygen-exposed orKG-Fe(II)deA. On the basis of
resonance Raman and mass spectrometric evidence, the deA chromophore previously
was assigned to a LMCT band associated with OH-Trp coordination to F e(III) (7).
LC/MS/MS evidence confirms that OH-Trp also is produced in AlkB and we propose
that the spectrum similarly derives from a OH-Trp/Fe(III) LMCT transition. Whereas the
modified Trp in deA (Trp 113) immediately precedes a metal ligand (His 114), the AlkB
modification (OH-Trp 178) is more distant from the closest ligand (His 187) in the
sequence. Structure prediction efforts provide a robust three-dimensional model of AlkB
that is consistent with Trp 178 being close to the metallocenter and capable of becoming
hydroxylated.

Of interest, OH-Trp 17 8 also was identified in a control AlkB sample. We
attribute this finding as indicating that self-hydroxylation can occur within the cell, where
aKG, F e(II), and oxygen are present, at least in the case of an overexpressing strain.
Because oxygen exposure of Fe(II)AlkB does not yield a detectable blue chromophore,
either the amount of OH-Trp is quite small in the sample or the chromophore only can be
generated from unmodified AlkB, i.e., OH-Trp does not coordinate F e(III) when
previously modified protein is amended with F e(II) and the metal is oxidized. These
results suggest that anaerobic growth conditions could be used to help insure the
production of fully active, unmodified protein.

The phenomenon of enzyme self-hydroxylation appears to be widespread in non-

heme Fe(II) oxygenases. In addition to the hydroxylation of a Trp side chain noted above

101

for AlkB and previously described for deA (7), a tyrosine residue is hydroxylated to
form a catecholate derivative in TauD (6, 8) and phenylalanine hydroxylation has been
noted in both tyrosine hydroxylase (16, 17) and the Y325F variant of phenylalanine
hydroxylase (18). Furthermore, both tyrosine and phenylalanine hydroxylation have been
observed in various mutants of the R2 subunit of ribonucleotide reductase, a dinuclear
iron protein (19-22). The blue oxidized form of 4-hydroxyphenylpyruvate dioxygenase
(23), assigned to a species possessing tyrosinate-Fe(III) LMCT transitions (24), might
derive from analogous hydroxylation of an active site phenylalanine (as proposed in (25))
on the basis of the enzyme crystal structure which shows the absence of tyrosines, but
presence of phenylalanines, in the vicinity of the metallocenter (26). Biomimetic
complexes designed to mimic the reactivity of aKG-dependent dioxygenases also
undergo self-hydroxylation chemistry involving a ligand phenyl group (27, 28). In
summary, the hydroxylation chemistry of several enzymes and model compounds is
consistent with the generation of a highly reactive intermediate that oxidizes an aromatic
substituent when the primary substrate is absent. It remains unclear whether the aberrant
enzymatic reactions of AlkB and these other systems may benefit the proteins by
protecting them from more harmful oxidative reactions such as backbone cleavage

reactions (as observed in aminocyclopropane carboxylate oxidase (29)).

102

10.

11.

12.

13.
14.

15.

16.

References

Prescott, A. G. and Lloyd, M. D., Natural Products Reports, 2000, 1 7, 367-383.

Price, J. C., Barr, E. W., Glass, T. B., Krebs, C., and Bollinger, J. M., Jr., J Am
Chem Soc, 2003, 125, 13008-9.

Price, J. C., Barr, E. W., Tirupati, B., Bollinger, J. M., Jr., and Krebs, C.,
Biochemistry, 2003, 42, 7497-508.

Proshlyakov, D. A., Henshaw, T. F., Monterosso, G. R., Ryle, M. J ., and
Hausinger, R. P., J Am Chem Soc, 2004, 126, 1022-3.

Rohde, J. U., In, J. H., Lim, M. H., Brennessel, W. W., Bukowski, M. R., Stubna,
A., Munck, E., Nam, W., and Que, L., Science, 2003, 299, 1037-1039.

Ryle, M. J ., Koehntop, K. D., Liu, A., Que, L., Jr., and Hausinger, R. P., Proc
Natl Acad Sci U S A, 2003, 100, 3790-5.

Liu, A., Ho, R. Y., Que, L., Jr., Ryle, M. J ., Phinney, B. S., and Hausinger, R. P.,
J Am Chem Soc, 2001, 123, 5126-7.

Ryle, M. J ., Liu, A., Muthukumaran, R. B., Ho, R. Y., Koehntop, K. D.,
McCracken, J ., Que, L., Jr., and Hausinger, R. P., Biochemistry, 2003, 42, 1854-
62.

Ginalski, K., Elofsson, A., Fisher, D., and Rychlewski, L., Bioinformatics, 2003,
19,1015-1018.

Ginalski, K., Pas, J ., Wyrwicz, L. S., von Grotthuss, M., Bujnicki, J. M., and
Rychlewski, L., Nucleic Acids Research, 2003, 31 , 3804-3 807.

Feig, M., Karanicolas, J ., and Brooks, C. L., 3rd, J Mol Graph Model, 2004, 22,
377-95.

Skolnick, J ., Kolinski, A., and Ortiz, A. R., Journal of Molecular Biology, 1997,
265, 217-241.

Feig, M. and Brooks, C. L., 3rd, Proteins, 2002, 49, 232-45.

Brooks, B. R., Bruccoleri, R. E., Olafson, B. D., States, D. J ., Swaminathan, S.,
and Karplus, M., Journal of Computational Chemistry, 1983, 4, 187-217.

Trewick, S. C., Henshaw, T. F., Hausinger, R. P., Lindahl, T., and Sedgwick, B.,
Nature, 2002, 419, 174-178.

Ellis, H. R., Daubner, S. C., McCulloch, R. I., and Fitzpatrick, P. F .,
Biochemistry, 1999, 38, 10909-14.

103

17.

18.

19.

20.

21.

22.

23.
24.

25.
26.

27.

28.

29.

Goodwill, K. B., Sabatier, C., and Stevens, R. C., Biochemistry, 1998, 37, 13437- '.
45.

Kinzie, S. D., Thevis, M., Ngo, K., Whitelegge, J ., Loo, J. A., and Abu-Omar, M.
M., J Am Chem Soc, 2003, 125, 4710-1.

Baldwin, J ., Voegtli, W. C., Khidekel, N., Moenne-Loccoz, P., Krebs, C., Pereira,
A. S., Ley, B. A., Huynh, B. H., Loehr, T. M., Riggs-Gelasco, P. J ., Rosenzweig,
A. C., and Bollinger, J. M., Jr., J Am Chem Soc, 2001, 123, 7017-30.

Ormo, M., deMare, F ., Regnstrom, K., Aberg, A., Sahlin, M., Ling, J ., Loehr, T.
M., Sanders-Loehr, J ., and Sjoberg, B. M., J Biol Chem, 1992, 267, 8711-4.

Ormo, M., Regnstrom, K., Wang, Z., Que, L., Jr., Sahlin, M., and Sjoberg, B. M.,
J Biol Chem, 1995, 270, 6570-6.

Logan, D. T., deMare, F., Persson, B. O., Slaby, A., Sjoberg, B. M., and
Nordlund, P., Biochemistry, 1998, 37, 10798-807.

Lindstedt, S. and Rundgren, M., J Biol Chem, 1982, 25 7, 11922-31.

Bradley, F. C., Lindstedt, S., Lipscomb, J. D., Que, L., Jr., Roe, A. L., and
Rundgren, M., J Biol Chem, 1986, 261 , 11693-6.

Ryle, M. J. and Hausinger, R. P., Curr Opin Chem Biol, 2002, 6, 193-201.

Serre, L., Sailland, A., Sy, D., Boudec, P., Rolland, A., Pebay-Peyroula, E., and
Cohen-Addad, C., Structure Fold Des, 1999, 7, 977-88.

Hegg, E. R., Ho, R. Y. N., and Que, L., Jr., Journal of the American Chemical
Society, 1999, 121, 1972-1973.

Mehn, M. P., Fujisawa, K., Hegg, E. L., and Que, L., Jr., J Am Chem Soc, 2003,
125, 7828-42.

Zhang, Z., Barlow, J. N., Baldwin, J. B., and Schofield, C. J ., Biochemistry, 1997,
36, 15999-6007.

104

CHAPTER VII

CONCLUSIONS AND FUTURE DIRECTIONS

105

 

The results presented in this dissertation represent a significant advance in our 1
understanding of the FeS/AdoMet enzymes. When I first started work on AE, the
existence of an iron sulfur cluster was a hypothesis. Now the role of the cluster in radical
generation has been identified and the interaction with AdoMet established. Chapter 11
demonstrated the exceptional flexibility of the cluster-binding site of PF L-AE to
accommodate a wide variety of cluster types. The work presented in Chapter IV

+
”’2 redox

differentiated among the various types and demonstrated that the [4Fe-4S]
couple is catalytically relevant.

The work presented in Chapter III demonstrated the existence of a unique iron site
in the iron-sulfur cluster of PFL-AB and the binding of AdoMet to this site, however the
results were unable to differentiate between various potential binding modes. The
sulfonium center was deemed unlikely to be a ligand, since the large change in isomer
shift observed was more consistent with coordination of the amino, carboxylate, or
hydroxyl groups. Since these results were reported, further work with PFL-AB and other
systems has clarified our understanding of the interaction between AdoMet and the iron-
sulfur cluster. ENDOR studies of PFL-AB using specifically labeled AdoMet have
provided strong evidence that AdoMet indeed chelates the unique iron site of PFL-AB
through the amino and carboxylate groups, perhaps positioning the sulfonium group for
an inner-sphere electron transfer or attack by one of the bridging sulfides (1, 2). Using
170 and 15N labeled AdoMet the Broderick and Hoffman groups clearly showed that both

the amino and carboxylate groups coordinate an iron of the cluster. This coordination

mode is consistent with the data presented in Chapter III of this dissertation.

106

Mfissbauer studies of BioB show that the isomer shift of one of the mixed-valence
pairs of the [4Fe-4S]2+ cluster increases from 0.47 to 0.64 mm/s in the presence of
AdoMet, a similar shift to that observed in PFL-AB . In addition, the symmetric
breathing mode of the cluster shifts from 338 to 342 cm”1 upon addition of AdoMet, a
shift that is consistent with the ligation of an O/N ligand to the cluster (3). Significantly,
the crystal structures of two FeS/AdoMet enzymes have been reported. The structure of
BioB shows AdoMet bound to a unique site in the [4Fe-4S] cluster through the amino
and carboxylate groups (4), in a manner similar to that of PFL-AB based on the available
spectroscopic data (1, 2, 5). The crystal structure of HemN, another FeS/AdoMet
enzyme, also shows AdoMet chelated to a unique iron site of the [4Fe-48] cluster with
the sulfonium group approximately equidistant from the nearest iron and bridging sulfide
(6). In the absence of any structural information on PF L-AE, it might be possible to build
a useful model using the structures of BioB and HemN as templates. An interesting
aspect of these structures is that they are both built around a TIM barrel fold, perhaps the
most common fold known. The authors of the BioB structure paper suggest that this
would be a favorable architecture for an enzyme with a small-molecule substrate but
raises questions about how other FeS/AdoMet enzymes could gain access to large-
molecule substrates, such as other proteins. The answer might come from observing the
HemN structure, which is built around an incomplete TIM barrel (6). Perhaps the
substrate completes the TIM barrel, providing a basis for protein docking in the enzyme-
substrate complex. The crystal structure of PFL is known (7-9), and with a model of

PFL-AB, it might be possible to dock the structures such that PFL completes an 0113

107

barrel in PFL-AE and Gly-734 is positioned close to the 5’ carbon of the adenosyl moiety
of AdoMet.

The chelation of a unique iron site of the [4Fe-4S] cluster by the amino and
carboxylate groups of AdoMet will likely be a common theme in this group of enzymes.
The exact mechanism by which this complex generates a 5’dAdo radical remains
uncertain. Although an inner sphere electron transfer may generate the adenosyl radical,
it is also possible that there are cluster bound AdoMet derived intermediates. The PFL-
AE system represents an excellent opportunity to investigate this question using pre-
steady state methods, such as freeze-quench and stopped-flow spectroscopy. To my
knowledge, the single turnover conditions described in Chapter IV are more precisely
defined than those generated for other F eS/AdoMet enzymes.

The function and mechanism of AlkB in protecting cells from the toxic effects of
methylating agents has been an open question for decades. The results presented in
Chapter V establish that AlkB functions as an F e(II)/a-KG dioxygenase by a novel
mechanism of oxidative DNA repair. It is assumed that the enzyme utilizes the same
mechanism as other F e(II)/a-KG dioxygenases in generating an F e(IV)=O intermediate
(10-12), however, there are several interesting questions that are unique to AlkB. The
first is how does it bind DNA. The model presented in Chapter VI could be used to
identify potential DNA-interacting residues and mutagenesis targets. It is also unknown
how AlkB recognizes the l-meA and 3-meC substrates. Our lab, in collaboration with
Professor John McCracken, is currently investigating this question by EPR, ESEEM, and
ENDOR spectroscopy using NO-bound AlkB. Using deuterium-labeled methylated

DNA, we hope to obtain information on the distance and orientation of the methyl group

108

,_ .

 

with respect to the metal center. The crystal structure of AlkB bound to substrates would 3'

answer many of these questions and is being actively pursued in multiple crystallography

labs.

A human homolog of AlkB was identified in 1996 (hABHl for human AlkB
homolog) and was reported to complement an E. coli alkB mutant, but more recent
studies have been unable to reproduce these findings and have not identified an
enzymatic activity for this homolog (13, 14). As such, the function of hABHl remains a
mystery. Two recently identified human homologs of AlkB (hABH2 and 3) have been
shown to demethylate 1-meA and 3-meC (13, 14). In addition to their activities with
methylated DNA, AlkB and hABH3 have been shown to repair l-meA and 3-meC in
RNA (14). In order to ascertain the role of these different AlkB homologs, interference
RNA knockdown cell lines and knockout mice are being generated (15). In addition to
the AlkB homologs that have been characterized by biochemical means, bioinformatic
work has identified an additional five putative homologs in the human genome (16).
These additional homologs should be investigated by biochemical and genetic

approaches.

109

10.

11.

12.

13.

14.

15.

References

Walsby, C. J ., Hong, W., Broderick, W. B., Cheek, J ., Ortillo, D., Broderick, J. B.,
and Hoffman, B. M., Journal of the American Chemical Society, 2002, 124, 3143-
3151.

Walsby, C. J., Ortillo, D., Broderick, W. B., Broderick, J. B., and Hoffman, B. M.,
Journal of the American Chemical Society, 2002, 124, 11270-11271.

Cosper, M. M., Jameson, G. N. L., Hernandez, H. L., Krebs, C., Huynh, B. H.,
and Johnson, M. K., Biochemistry, 2004, 43, 2007-2021.

Berkovitch, F., Nicolet, Y., Wan, J. T., Jarrett, J. T., and Drennan, C. L., Science,
2004, 303, 76-9.

Krebs, C., Broderick, W. E., Henshaw, T. F., Broderick, J. B., and Huynh, B. H.,
Journal of the American Chemical Society, 2002, 124, 912-913.

Layer, G., Moser, J ., Heinz, D. W., Jahn, D., and Schubert, W. D., Embo J, 2003,
22, 6214-24.

Leppanen, V. M., Merckel, M. C., Ollis, D. L., Wong, K. K., Kozarich, J. W., and
Goldman, A., Structure Fold Des, 1999, 7, 733-44.

Becker, A. and Kabsch, W., J Biol Chem, 2002, 277, 40036-42.

Becker, A., Fritz-Wolf, K., Kabsch, W., Knappe, J ., Schultz, S., and Volker
Wagner, A. F., Nat Struct Biol, 1999, 6, 969-75.

Price, J. C., Barr, E. W., Glass, T. B., Krebs, C., and Bollinger, J. M., Jr., J Am
Chem Soc, 2003, 125, 13008-9.

Price, J. C., Barr, E. W., Tirupati, B., Bollinger, J. M., Jr., and Krebs, C.,
Biochemistry, 2003, 42, 7497-508.

Proshlyakov, D. A., Henshaw, T. F ., Monterosso, G. R., Ryle, M. J ., and
Hausinger, R. P., J Am Chem Soc, 2004, 126, 1022-3.

Duncan, T., Trewick, S. C., Koivisto, P., Bates, P. A., Lindahl, T., and Sedgwick,
B., Proc Natl Acad Sci U S A, 2002, 99, 16660-5.

Aas, P. A., Otterlei, M., Falnes, P. O., Vagbo, C. B., Skorpen, F., Akbari, M.,
Sundheim, O., Bjoras, M., Slupphaug, G., Seeberg, B., and Krokan, H. B., Nature,
2003, 421, 859-63.

Samson, L., personal communication.

110

16. Kurowski, M. A., Bhagwat, A. S., Papaj, G., and Bujnicki, J. M., BMC Genomics,
2003, 4, 48.

111

 

       

M ARS

llllllllllfllllflfllfllflllllflllfllll

1