sizy‘
. .3

- :2 .l. 1
.1. ts!

\

. l5
. . . . 1.) n :3 .
an. 1. 3L ... .. tn...
4 s .1. a. 3D '12
. , .175. :3.
. . 2.3333.
Anna“...

 

-

$1.63!. to
t .2.) K

 

A‘
x a 1):: .
.Ia

 

x.

:37:

t... . s

. J. n.
2..
z...

.9. I!
3.3“... (it: ..
..w$....!.5.£3..5..

 

?
a:
1‘ :1... s!. .

 

 

 

 

 

 

 

 

.. ‘ : Q... :. ... auxin... u.
invigfi ‘éaéfiamfiyifif
. , a . . . e gap 33:: V . ‘
”fl.c .. a. :m. t. 2,1. . . .nm er???- , : ‘ . . .
inmuffirflfigfiw2 . : : .

   

 
 
 
 

 

  
  
  
 
  
 
  
 
 
  
 
 
 
  
   
  
     

 

 

LIBRARY
Michigan State.
University '

Tl 's is to certify that the
thesis entitled
INTERACTION 0F S—ADENOSYLMETHIONINE WITH ‘

THE IRON-SULFUR CLUSTER OF I
PYRUVATE FORMATE-LYASE ACTIVATING ENZYME

presented by

Wei Hong

has been accepted towards fulfillment ‘
of the requirements for J

”.5. degree in Chemistry l

Major professor

'6’?) 077
l , ,

(/ZLKL/ flag/é L#(/
Date g/Zs/OI '

0-7639 MS U i: an Affirmative Action/Equal Opportunity lruu‘mtiou

 

 

.3.
Jim

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

 

N 5220112».

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c/CIRC/DatoDuepfi5-Q1 5

 

 

 

 

 

INTERACTION OF S—ADENOSYLMETHIONINE WITH
THE IRON-SULFUR CLUSTER OF
PYRUVATE FORMATE-LYASE ACTIVATING ENZYME

By

Wei Hong

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Chemistry

2001

INTERACTION OF S—ADENOSYLMETHIONINE WITH
THE IRON-SULFUR CLUSTER OF
PYRUVATE FORMATE-LYASE ACTIVATIN G ENZYME

By

Wei Hong

AN ABSTRACT OF A THESIS
Submitted to
Michigan State University .
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Chemistry

2001

Professor Joan B. Broderick

 

 

' ml-

ABSTRACT

INTERACTION OF S-ADENOSYLMETHIONINE WITH
THE IRON-SULFU R CLUSTER OF
PYRUVATE FORMATE—LYASE ACTIVATIN G ENZYME

By

Wei Hong

Pyruvate formate-lyase activating enzyme (PFL-AB) is a member of a growing
family of enzymes that utilize [4Fe-4S] clusters and S-adenosylmethionine (AdoMet) to
generate catalytically essential radicals. PFL-AB generates a glycyl radical on pyruvate
formate-lyase (PFL), and converts AdoMet stoichiometrically to methionine and S’-
deoxyadenosine. An AdoMet-derived adenosyl radical has been implicated as the
intermediate responsible for abstraction of the pro-S hydrogen atom of PFL Gly734. In
order to probe the mechanism by which the Fe-S cluster interacts with AdoMet to
generate the adenosyl radical intermediate, the methyl group of AdoMet was labeled with
2H and 13C, and the sulfonium sulfur was replaced by selenium. Investigations of the
interaction between AdoMet and the Fe-S cluster of PFL-AB were carried out with the
isotopically labeled AdoMet and Se-AdoMet, using a variety of spectroscopic methods.
A close association of the PFL-AB [4Fe-4S] cluster with AdoMet has been demonstrated
for the first time by using 2H and 13C ENDOR spectroscopy. Our results suggest that

the Fe-S cluster of PFL-AB is directly involved in generating the putative adenosyl

radical from AdoMet.

 

 

 

 

 

 

ACKNOWLEDGEMENTS

First of all, I would like to thank my advisor, Prof. Joan B. Broderick for guiding
me through my graduate study. She has shown a great deal of support, encouragement,
understanding and patience for the past two years that I worked in the lab. I am grateful
to Dr. William E. Broderick, who has put so many great ideas in my work, and provided
suggestions in my thesis writing.

I want to thank Jennifer and Tim for being so helpful all the time. I gained so
much knowledge, as well as hands-on experience in our group’s research from them.
Also, Jennifer ran all the EPR samples for me. Mbako, Danilo, Silvana, and all the
people who used to be in this group, have been great friends to me. I have enjoyed being
a part of this group.

I appreciate the tremendous contributions that Dr. Charles Walsby, Prof. Brian
Hoffman, and all the collaborators have given to the work I did in the lab and to our
research as a whole.

Lastly, I am truly indebted to my parents’ endless love and support, and all the

things they have done for me during the month they spent with me here.

 

 

 

 

 

TABLE OF CONTENTS

LIST OF SCHEMES ............................................................................... vi
LIST OF TABLES ................................................................................. vii
LIST OF FIGURES ................................................................................ viii
LIST OF ABBREVIATIONS ....................................................................... x
CHAPTER I

INTRODUCTION .................................................................................. 1
1.1 Pyruvate formate-lyase ...................................................................... l
1.2 Pyruvate formate-lyase activating enzyme ............................................... 3
1.3 Adenosylmethimine-dependent iron-sulfur enzymes ................................... 6

1.4 S-adenosylmethionine and the Fe-S cluster of PFL-AB in radical generation ...... 10

CHAPTER II

OVEREXPRESSION AND CHARACTERIZATION OF
PYRUVATE FORMATE-LYASE AND

PYRUVATE FORMATE-LYASE ACTIVATING ENZYME ........................... 14
II. 1 Introduction ................................................................................. 14
11.2 Experimental methods ..................................................................... 17
11.3 Results and Discussion .................................................................... 24

iv

 

 

 

 

 

 

 

CHAPTER HI

SYNTHESES AND CHARACTERIZATION OF ISOTOPICALLY LABELED
S-ADENOSYL-L—METHIONINES AND

Se-ADENOSYL-L-SELENOMETHIONINE ................................................ 33
111.1 Introduction ................................................................................. 33
HI.2 Experimental methods ..................................................................... 37
111.3 Results and Discussion .................................................................... 41
CHAPTER IV

INVESTIGATING THE INTERACTION OF ADOMET WITH PFL-AE

USING EPR AND EN DOR SPECTROSCOPY ............................................ 51
IV] Introduction ................................................................................. 51
IV.2 Experimental methods ..................................................................... 55
IV.3 Results and Discussion .................................................................... 58
CHAPTER V

Se-ADENOSYL-L-SELENOMETHIONINE AND ITS INTERACTION WITH
PYRUVATE FORMATE-LYASE ACTIVATING ENZYME ........................... 75
V.1 Introduction ................................................................................. 75
V2 Experimental methods ..................................................................... 81
V3 Results and Discussion .................................................................... 84
CHAPTER VI

CONCLUSIONS .................................................................................. 94
REFERENCES ............................................. . ...................................... 96

 

 

 

 

 

Scheme 1.1

Scheme 1.2

Scheme [.3

Scheme [.4

Scheme [.5

Scheme [.6

Scheme [.7

Scheme L8

Scheme 1.9

Scheme IH.1

Scheme III.2

Scheme HI.3

Scheme V.1

 

_—.-—. _Iq I ,7 nui—

LIST OF SCHEMES

PFL catalyzed interconversion of pyruvate and CoA to

acetyl-CoA and formate ........................................................... 1
Structure of the glycyl radical in PFL ............................................ l
The activation / deactivation of PFL by generation or quenching of

the glycyl radical on PFL ......................................................... 3
AdoMet-dependent PFL glycyl radical generation by PFL-AE ............. 4
Reaction catalyzed by benzylsuccinate synthase activating enzyme ......... 8

The interconversion between L-lysine and L-fl-lysine
catalyzed by lysine 2,3-aminomutase ............................................ 9

Biosynthesis of biotin from dethiobiotin catalyzed by biotin synthase ...... 9

Reaction catalyzed by lipoate synthase (LipA) ................................ 10
General mechanism for radical generation in PFL-AB ....................... 12
Reductive cleavage of AdoMet by PFL-AE in PFL activation ............. 34
AdoMet and the Fe-S cluster of PFL-AE are proposed to be

in close proximity during radical generation .................................. 35
Enzymatic synthesis of S—adenosyl-L-methionine ............................. 36
Selenoamino acids found in plants ............................................. 77

vi

 

 

w

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES

Table H.l Ratio of absorbances at 426 nm and 280 nm of
PFL-AE fractions of? the 2nd run of the gel filtration column .............. 28

Table H.2 Summary of activity assay of PFL-AE ......................................... 30

Table IV.l T values derived from the dipolar interpretation of

the 2H ENDOR data .............................................................. 73

vii

Figure 11.1

Figure H.2

Figure 11.3

Figure H.4
Figure 11.5

Figure 11.6

Figure 11.7

Figure 11.8
Figure HI.1

Figure HI.2

Figure 111.3

Figure H14

Figure IH.5

Figure [11.6

Figure IV.1

 

LIST OF FIGURES

Chromatograms of purification of PFL ........................................ 25

SDS-PAGE analysis of PFL fractions ofl‘
the Accell Plus QMA anion-exchange column ................................ 26

SDS-PAGE analysis of PFL fractions off

the Phenyl-Sepharose hydrophobic column ................................... 26
SDS-PAGE analysis of overexpression of PFL-AB in E. coli cells ........ 27
Purification of PFL-AE by gel-filtration chromatography ................... 28

SDS-PAGE analysis of PFL-AE fractions ofi‘ the 2nd run of
the gel filtration column ......................................................... 29

X-band EPR spectrum of PFL-AE as-isolated

in the presence of 1 mM DTT ................................................... 30
Activity assay of PFL-AE ........................................................ 31
SDS-PAGE analysis of AdoMet synthetase ................................... 41

SDS-PAGE analysis of AdoMet synthetase fractions off
the Phenyl-Sepharose column ................................................... 42

Chromatogram of isolation of CDg-AdoMet by
SOURCE 158 cation exchange chromatography ............................. 45

Chromatogram of isolation of 13CHg-AdoMet by
SOURCE 15S cation exchange chromatography ............................. 46

Chromatogram of isolation of Se-AdoMet by
SOURCE 15$ cation exchange chromatography ............................. 47

‘H NMR (300 MHz) of AdoMet, CDg-AdoMet,
'3CH3-AdoMet, and Se-AdoMet ................................................ 48

X—band EPR spectrum of photoreduced PFL-AE
in the presence of CDg-AdoMet ................................................ 60

viii

 

 

 

 

 

Figure IV.2

Figure IV.3

Figure 1V.4

Figure 1V.5

Figure IV.6

Figure V.l

Figure V.2

Figure v.3

Figure V.4

Figure V.5

 

 

X-band EPR spectrum of photoreduced PFL-AB
in the presence of l3CH3--AdoMet .............................................. 60

35 GHz Mims pulsed-ENDOR spectra of PFL-AE with
methyl-D3 AdoMet ............................................................... 68

35 GHz Mims pulsed-ENDOR spectra of PFL-AB with
methyl-13C AdoMet .............................................................. 70

13C field dependence with simulation of line positions and
relative intensities ................................................................. 71

Structural model of AdoMet bound to
the [4Fe-48] cluster of PFL—AE ................................................ 74

Activity assay of PFL-AB in the presence of Se-AdoMet ................... 84

X-band EPR spectra of photoreduced PFL-AB (A) in the presence of
Se-AdoMet and (B) in the presence of AdoMet .............................. 86

Se K-edge X-ray absorption spectra (A) and Fourier transforms (B) of
Se-Met and Se-AdoMet ......................................................... 88

Se K-edge X-ray absorption spectra (A) and Fourier transforms (B) of
PFL-AB [4Fe-48]2+ incubated with Se-Met / 5’-dAdo (EPOSA),

Se—Met / 5’-dAdo / PFL (EPRPA and EPRPB), Se-AdoMet (EPOEB),
and Se-AdoMet / PFL (EPOQA) ............................................... 90

Se K-edge X-ray absorption spectra (A) and Fourier transforms (B) of
PFL-AB [4F e-4S]1+ in the presence of Se—AdoMet .......................... 9]

ix

 

 

 

LIST OF ABBREVIATIONS

5’-dAdo .......................................................................... 5’-deoxyadenosine
AdoMet .................................................................. S—adenosyl-L-methionine
aRNR ........................................................ anaerobic ribonucleotide reductase
aRNR-AB .............................. anaerobic ribonucleotide reductase activating enzyme
ATP ......................................................................... adenosine triphosphate
,B-ME .............................................................................. ,B-mecaptoethanol
CD3~AdoMet .................................................... adenosyl-methyl-Dg-methionine
‘3 CHg-AdoMet .................................................. adenosyl-methyl-l3C-methionine
CoA ...................................................................................... coenzyme A
DTT ..................................................................................... dithiothreitol
E. coli ............................................................................... Escherichia coli
ENDOR ....................................................... electron nuclear double resonance
EPR .............................................................. electron paramagnetic resonance
ESEEM ................................................. electron spin echo envelope modulation
Hepes ................................. N—(2-hydroxyethyl)piperazine—N’ -2—ethanesulfonic acid
IPTG ........................................................ isopropyl-fl-D-thiogalactopyranoside
LAM ....................................................................... lysine 2,3-aminomutase
LB ...................................................................................... Luria-Bertani
NMR .................................................................... nuclear magnetic resonance
PFL .......................................................................... pyruvate formate-lyase

 

   

 

 

 

 

PFL-AB .............................................. pyruvate formate-lyase activating enzyme

PMSF ............................................................... phenylmethylsulfonyl fluoride
SDS-PAGE ..................... sodium dodecyl sulfate-polyacrylamide gel electrophoresis
Se-AdoMet ..................................................... Se-adenosyl-L-selenomethionine
Se—Met ............................................................................. selenomethionine
Tris ............................................................ tris(hydroxymethyl)aminomethane
XAS ............................................................... X-ray absorption spectroscopy

 

   

 

 

 

CHAPTER I

INTRODUCTION
1.1 Pyruvate formate-lyase
Pyruvate formate-lyase (acetyl-CoAzformate C-acetyltransferase, EC 2.3.1.54;
PFL) catalyzes the first committed step in anaerobic glucose metabolism in Escherichia
coli (E. coli) cells, the conversion of pyruvate and coenzyme A (CoA) to formate and

acetyl-CoA (Scheme 1. 1) (1-3). Knappe et al. identified PFL as the first enzyme known

0

0 PFL 0
H3C + HSCoA —> /fl\ + A
H o- COAS

CH3
0

Scheme 1.1 PFL catalyzed interconversion of pyruvate and CoA to

acetyl-CoA and formate

to contain a stable glycyl radical. The catalytically essential radical of PFL was shown to
be located on the a—carbon of glycine 734 through isotopic labeling, EPR spectroscopy,
and analysis of the products of oxygenolytic cleavage of the activated PFL (4). The

glycyl radical (Scheme 1.2) is stable in the absence of dioxygen, and its stability can be

H II
N . C
\(i/ \C/ \n/
0 IL

Scheme 1.2 Structure of the glycyl radical in PFL

 

 

 

 

_’....

 

attributed to the summation of the effects of resonance electron withdrawal by the glycyl-
carbonyl group and resonance electron donation by the adjacent amide nitrogen through
its lone electron pair. Gly734 does not directly participate in catalysis, but rather serves
as the source of an unpaired electron that can be relayed to the active site in the form of a
cysteinyl-thiyl radical, which is directly involved in homolytic cleavage of the pyruvate
C-C bond (5, 6). Based on steric relationships among the amino acid residues, the
Gly734 radical was proposed to abstract a hydrogen atom from Cys4l9, and the Cys4l9
thiyl radical in turn relays the radical center to Cys418, which is in the pyruvate binding
site (7).

PFL from E. coli is a homodimeric protein of 170 kDa. The enzyme can exist in
an active form (PFLa), which contains one glycyl radical per dimer, and an inactive form
(PFLi) in which the glycyl radical is reduced (1). The catalytically essential glycyl
radical of PFL is post-translationally generated under anaerobic conditions by the
stereospecific abstraction of the pro-S hydrogen atom of the Gly734 methylene group (8).
The interconversion of PFL between its inactive and active forms requires the pyruvate
formate-lyase activating enzyme (PFL-AB), S—adenosyl-L-methionine (AdoMet),
fiavodoxin, and flavodoxin reductase (Scheme 1.3) (1). In addition, a “deactivase”, a
regulatory component of the PFL system, reduces the glycyl radical of PFL to glycine,

preventing its destruction by 02 (9, 10).

 

 

 

5’-deoxyadenosine
+
methionine

ll
PFL, PFL-AB

[flavodoxin reductasa

NADPH

N AD+

 

 

PFLi PFL-AB
NADH

G'YH [4Fe—4S]l+

 

S-adenosyl-L-methionine

Scheme 1.3 The activation / deactivation of PFL by generation or quenching of the

glycyl radical on PFL.

1.2 Pyruvate formate-lyase activating enzyme

Pyruvate formate-lyase activating enzyme (PFL-AB) generates the glycyl radical
at G734 on PFL. PFL—AE requires S-adenosylmethionine (AdoMet) as a co-substrate in
the radical generation, subsequently cleaving it to methionine and 5’-deoxyadenosine (5’-
dAdo) (Scheme 1.4) (2). The hydrogen atom abstracted from Gly734 of PFL has been
shown to be incorporated into the 5’-dAdo product by isotopic labeling, suggesting that
the role of AdoMet is to produce an adenosyl radical intermediate as the immediate
hydrogen atom abstractor. By using an octapeptide analogous to the Gly734 site of PFL,
but with a dehydroalanyl residue in the glycyl position, Wagner et al. have trapped the

adenosyl radical intermediate by C-adenosylation of the dehydroalanyl residue (11).

 

 

 

11!?"

 

H H E i
’4, PFL-Gly PFL-Glyo A
51" rt PFL-AE LLLH 0 Jr",

1" e

Scheme 1.4 AdoMet-dependent PFL glycyl radical generation by PFL-AB

PFL-AB was initially isolated aerobically from non-overexpressing E. coli cells
and found to be a 28 kDa monomer with a broad absorbance from 310 to 550 nm
suggesting the presence of a covalently bound cofactor (1). The catalytic activity of PFL—
AE was shown to be completely dependent on the presence of exogenous iron in the
assay, which was the first indication of the role of iron in glycyl radical generation. The
first over-expression of PFL-AE in E. coli was reported by Kozarich et al. (12), but
solubility problems made it necessary to purify and then refold the denatured protein.
The refolded protein was able to bind stoichiometric quantities of Fe as well as other
divalent metals such as Cu(II) and Co(II), but it was found to have enzymatic activity
only in the presence of Fe(11) and DTT. A cysteinal coordination sphere for the iron
center of PFL-AB was suggested because thiophilic metals such as Cu(H), Zn(II), Hg(II)
and Cd(II) were found to be inhibitors of PFL-AE activity.

The first isolation of PFL-AE in its native state under strictly anaerobic conditions
and identification of the presence of an iron-sulfur cluster in PFL-AE were achieved by

Broderick et al. ( 13, 14). PFL-AE with an intact Fe-S cluster has been purified and a

 

 

 

 

 

 

 

 

high Specific activity (95 U/mg in the absence of added iron) was obtained for PFL-AE

containing 2.65 Fe per protein. When purified under anaerobic conditions in the absence
of DTT, PFL-AE was shown to contain primarily a [3Fe-4S]+ cluster by a combination of
UV-visible, EPR and resonance Raman spectroscopic methods (14). A complete
description of the states of the cluster present in the enzyme was provided by Mossbauer
spectroscopy (15). The major component was confirmed to be the cuboidal [3Fe-4S]+
cluster, accounting for 66 % of the total iron. Minor contributions from [2Fe-28]2+ (12 %
of the total Fe), [4Fe-4S]2+ (8 % of the total Fe) and linear [3Fe-48]+ (~10 % of the total
Fe) were also found in the Mossbauer spectrum of as isolated PFL-AE. When PFL-AE
was anaerobically reduced with dithionite, complete conversion of all cluster types to
[4Fe-48]2+’+ clusters was observed by Mossbauer spectroscopy.

The above work has clearly demonstrated that the Fe-S cluster of PFL-AB is
required for enzymatic activity, and that no ferrous iron is required for activity when an
intact cluster is present. Among the mixture of Fe-S clusters present in PFL-AB, the
[4Fe-48] cluster was considered to be the catalytically relevant cluster, as PFL-AE
activity is observed in vitro only under anaerobic reducing conditions (1). A definite
assignment of the [4Fe-4S]1+ of PFL-AE as the catalytically active cluster was achieved
in a “single turnover” experiment (16). Deazariboflavin-mediated photoreduction
afforded quantitative reduction of PFL-AE to the [4Fe-48]1+ state. After the excess
reductant was removed by placing the sample in the dark, either AdoMet alone, or
AdoMet plus PFL (equimolar to PFL-AB) was added to the reduced PFL-AB. Spin
quantitation of the resulting EPR spectra, taken as a function of illumination time, show a

1:1 correspondence between the amount of PFL glycyl radical generated and the amount

 

 

 

011‘“:e-4S]1+ cluster present in the PFL-AB prior to addition of PFL. The [4Fe-4S]1+ was
converted to an EPR silent state concomitant with glycyl radical formation, and
preliminary data suggest that the final cluster state is [4Fe-48]2+. This is the first direct
quantitative spectrosc0pic evidence that the [4Fe-4S]l+ of PFL—AB is the catalytically
relevant cluster, and this cluster provides the electron necessary for AdoMet-dependent
glycyl radical generation.

Site-directed mutagenesis studies have identified Cy829, Cys33, Cys36 as cluster
ligands in PFL-AB ( 17). In general, a similar cluster-binding motif comprised of only
three cysteines (CXXXCXXC) is common to all of the Fe-S/AdoMet-dependent enzymes
for which the gene sequence is known (18-23). Considering that the [4Fe-4S] cluster of
PFL-AE is the catalytically relevant cluster, the fourth ligand to the [4Fe-48] cluster is
presumably non-cysteine, therefore, resulting in a unique iron site, which may be

important in binding / interacting with AdoMet.

1.3 Adenosylmethionine—dependent iron-sulfur enzymes

Iron-sulfur clusters are widespread in biological systems and participate in a
broad range of functions (24-28), including electron transport, mediating redox catalysis,
and non-redox catalysis. Aconitase is one of the most thoroughly studied enzymes that
mediate non-redox catalysis, in which a [4Fe-4S]2+ cluster serves as a Lewis acid to
catalyze the interconversion between citrate and isocitrate (29). Fe-S clusters also play
roles in regulation of gene expression, for example in the iron-responsive protein (IRP)
(30), fumarate-nitrate reduction protein (FNR) (31), and SoxR (32) responding to

changes in levels of iron, oxygen, and superoxide respectively.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A new role for Fe-S clusters has emerged in recent years as a number of enzymes

have been identified that utilize Fe-S clusters and AdoMet to initiate radical catalysis.
This Fe-S cluster-mediated radical catalysis includes the generation of catalytically
essential glycyl radicals (4, 13, 33, 34), the generation of substrate radical intermediates
(35), cofactor biosynthesis (36-39), and repair of DNA damage (40). Amazingly, a novel
protein superfamily of more than 600 related enzymes that involve AdoMet-derived
radical biochemistry has very recently been discovered by iterative profile searches and
analyzed with bioinformatics and information visualization methods (41). This radical
SAM superfamily, as named by Sofia et al., has highly diverse functions in biochemical
pathways and reflects an ancient conserved mechanistic approach to difficult chemistries.
The following is a brief introduction of some defining members of the superfamily,
besides PFL-AE, that are being investigated in various laboratories.
The reduction of ribonucleotides into deoxyrebonucleotides is catalyzed by
ribonucleotide reductases (RNRs) in order to supply DNA precursors to cells (42-45).
All known RNRs fall into four distinct classes according to their metal cofactor, but all
use a protein radical to activate the ribonucleotide substrate (45). Class III RNRs are
utilized under anaerobic conditions by certain microbes, and are similar to PFL / PFL-AB
in that they have a glycyl radical and are AdoMet dependent. The anaerobic RNR

(aRNR) from E. coli was originally purified in an inactive dimeric a; form. In its active

form, a glycyl radical is generated by an activating enzyme, the fl protein, which contains

an Fe-S cluster (46-49). Evidence has suggested that the [4Fe-4S]1+ cluster of aRNR-AB

is catalytically relevant (34).

 

 

The activating enzyme of benzylsuccinate synthase is another probable member

of the Fe-S / AdoMet enzyme groups, which was found to have a high degree of
sequence homology with both PFL-AE and aRNR-AB, especially in the region of the
cluster-binding site (50). Benzylsuccinate synthase catalyzes the first step in anaerobic
degradation of toluene, the addition of the toluene methyl carbon to the double bond of
fumarate to form benzylsuccinate (Scheme 1.5) through a glycyl radical mechanism (51).
The sequence of benzylsuccinate synthase is significantly similar to that of PFL and

aRNR, with a highly conserved region around the glycyl radical site and an active-site

 

 

cysteine.
HOOC
H 00C COOH
+ \L :
COOH
Toluene Fumarate Benzylsuccinate

Scheme 1.5 Reaction catalyzed by benzylsuccinate synthase activating enzyme

Lysine 2,3-aminomutase (LAM) from Clostridiumsubtenninale 8B4 catalyzes the
interconversion between L-lysine and L-fl-lysine (Scheme 1.6). The reaction catalyzed by
LAM is analogous to those that are catalyzed by adenosylcobalamin-dependent
aminomutases, as well as other adenosylcobalamin isomerases. However, LAM
reactivity does not depend on adenosylcobalamin, but instead requires AdoMet, a Fe-S

cluster, and pyridoxal 5’-phosphate (PLP) (52-54). The [4Fe-4S]1+ cluster generated in

 

 

 

L-lysine L-fl—lysine

Scheme 1.6 The interconversion between L-lysine and L-fl-lysine

catalyzed by lysine 2,3-aminomutase

the presence of AdoMet has been demonstrated to be the catalytically active species of
the enzyme that is directly involved in the generation of a substrate radical intermediate
(35). By using the cofactor analog S-3’,4’-anhydroadenosyl-L-methionine, Magnusson et
al. have recently provided the first spectroscopic evidence for an analog of the adenosyl
radical intermediate in the Fe-S / AdoMet enzymes (55, 56).

The biosynthesis of biotin from dethiobiotin (Scheme 1.7) is catalyzed by the
enzyme biotin synthase (BioB, the bioB gene product) with the absolute requirement of
AdoMet as a co-substrate and the presence of an Fe-S cluster (37, 57). Isotopic labeling
studies have demonstrated that the Fe-S cluster is actually the sulfur donor (37), and an

AdoMet-derived deoxyadenosyl radical is involved in the cleavage of the OH bonds of

 

dethiobiotin (58).
1 ° °
HN NH HNJKNH __’ HNJkNH
HSMCOOH MCOOH WCOOH
sx S
Dethiobiotin Thiol intermediate Biotin

Scheme 1.7 Biosynthesis of biotin from dethiobiotin catalyzed by biotin synthase

The lipA gene product lipoate synthase (LipA) has been found to be involved in
the biosynthesis of lipoate (59-61) by catalyzing the insertion of either one (59) or both
(21, 61) of the sulfur atoms into the backbone of octanoic acid to produce lipoic acid
(Scheme 1.8). LipA has also been shown to contain an Fe-S cluster(38), and the [4Fe-

4S]1+ cluster form of LipA appears to be the active enzyme state (62).

S—S

OH
—-> -—->
W OH
O

O
Octanoic Acid Lipoic acid

Scheme 1.8 Reaction catalyzed by lipoate synthase (LipA)

Spore photoproduct lyase, an enzyme responsible for repairing DNA damage, has
recently been found to contain an Fe-S cluster (40). This enzyme has substantial
sequence homology to the Fe-S cluster binding regions of both aRNR-AB and PFL-AB
from Ecoli (22, 63), and its in vitro activity is dependent upon reducing conditions and

the addition of AdoMet as cofactor (40).

1.4 S—adenosylmethionine and the Fe-S cluster of PFL-AE in radical generation
Despite the differences in cluster properties for the AdoMet-dependent Fe-S
enzymes, evidence has pointed to a common mechanism involving an adenosyl radical
intermediate. The only other known source for 5’-deoxyadenosyl radical intermediates in
biology is adenosylcobalamin, an essential cofactor for numerous radical-mediated
rearrangements as well as ribonucleotide reduction in certain organisms (45, 64, 65). In
adenosylcobalamin, the adenosyl radical intermediate is reversibly generated by

homolytic cleavage of the weak Co—C bond. How the same radical intermediate can be

10

 

 

 

 

 

 

 

 

generated from the much simpler cofactor, AdoMet, through an interaction with an Fe-S

cluster is an intriguing mechanistic question. It has been shown for LAM (35), PFL-AB
(l6), and aRNR-AB (34) that the [4Fe-4S]'+ is the cluster that interacts with AdoMet to
initiate radical chemistry, and one-electron oxidation of the cluster appears to accompany
radical generation (16, 35). Based on this information, a general mechanistic framework
for the adenosyl radical generation in PFL-AE, as well as these related enzymes is
pr0posed (Scheme 1.9). In this scheme, the unidentified ligand to the labile iron site is
denoted as “X”. This ligand is likely to be an exogenous ligand in the resting enzyme or
in the environment, such as H20, but may be displaced by AdoMet in its catalytically
relevant state (Step 1). The presence of AdoMet has been shown to dramatically affect
the EPR signal of the PFL-AB [4Fe-4S]1+ (15, 16), suggesting the possibility of a direct
interaction between AdoMet and Fe-S cluster of PFL-AB, though the nature of the
enzyme-AdoMet complex is not clear. This direct interaction could occur through either
the iron or the sulfides of the cluster, and could involve the sulfonium, 5’-carbon,
hydroxyls, or amino and carboxylate functionalities of AdoMet. In the central and most
intriguing mechanistic step, the reduced cluster donates an electron to convert AdoMet to
methionine and 5’-deoxyadenosyl radical intermediate (Step 2). Label-transfer from the
glycine residue of PFL to 5’-dAdo product has provided indirect evidence for an
adenosyl radical intermediate (8), but direct spectroscopic evidence for the existence of
the adenosyl radical intermediate is not available due to the instability of this radical
intermediate. An allylic analog of the 5’-deoxyadenosyl radical has been detected in the
LAM catalyzed reaction (55, 56). The adenosyl radical intermediate then abstracts a

hydrogen atom from PFL Gly734 to generate 5’-dAdo and the glycyl radical (Step 3).

ll

 

 

 

 

 

 

 

 

 

 

PFL-AB consumes one AdoMet per turnover, and therefore, after release of methionine

of 5’-dAdo, an external reductant is necessary to reduce the [4Fe-4S]2+ to the [4Fe-4S]1+
cluster, enabling further turnover. In vivo, the source of reductant is flavodoxin/

flavodoxin reductase (l).

H . ' H
-0 + H ooc\'/\/ ’3 H
OC\{/\/ d NH+ H3C p A
+H3N CH3 4' 3
- 1

 

3 OH OH OH OH
2+

Wei—“T“

CY’L'Eef’

 

 

 

 

 

 

 

 

 

 

 

Scheme 1.9 General mechanism for radical generation in PFL-AE

12

 

The key question regarding the mechanism by which the Fe-S cluster interacts

with adenosylmethionine to generate an adenosyl radical intermediate remains to be
answered. As the sulfur-5’-deoxyadenosyl carbon bond is reductively cleaved during the
radical generation, the sulfonium center of AdoMet is likely to be positioned close to the
Fe-S cluster. In the work described here, the putative interaction between AdoMet and
the Fe-S cluster was probed. The sulfonium center of AdoMet was synthetically
modified, and used along with a variety of spectroscopic methods to investigate the
interaction between AdoMet and the Fe-S cluster of PFL-AB. Specifically, the methyl
group of AdoMet was labeled with 2H and '3 C, and the sulfur was replaced by selenium
to take advantage of the natural abundance of 77Se (7.50 %). All of these nuclei have
non-zero nuclear spins, and therefore, are effective probes in spectroscopic studies, such
as electron paramagnetic resonance (EPR) and electron nuclear double resonance
(ENDOR) spectroscopy, to demonstrate the possible interaction between AdoMet and the

Fe-S cluster of PFL-AE.

l3

 

 

 

 

 

 

 

 

 

 

 

 

 

CHAPTER II

OVEREXPRESSION AND CHARACTERIZATION
OF PYRUVATE FORMATE—LYASE AND
PYRUVATE FORMATE-LYASE ACTIVATING ENZYME

H.l Introduction

Pyruvate formate-lyase (PFL) plays a central role in anaerobic glucose
fermentation, catalyzing the reversible conversion of pyruvate and coenzyme A (CoA) to
acetyl-CoA and formate (Scheme 1.1). The catalytic activity of PFL is regulated at both
the transcriptional and post-translational levels (66, 67). Under aerobic conditions, PFL is
expressed constitutively but is inactive; conversion of the inactive form to the active form
occurs post-translationally upon a shift to anaerobic conditions, and is catalyzed by the
pyruvate formate-lyase activating enzyme (PFL-AE) (Scheme 1.4), which requires S-
adenosylmethionine and flavodoxin as cosubstrates. Pyruvate or oxamate is also required
as an allosteric effector for PFL activation (66-68). The inactive form of PFL has an (12
oligomeric structure (2x85 kDa) and contains no cofactor (2). The active form of PFL
contains a stable glycyl radical located at residue 734 (4). During the PFL activation
reaction, the glycyl radical is formed by abstraction of the pro-S hydrogen atom of G734
of PFL, and AdoMet is converted stoichiometrically to methionine and 5’-
deoxyadenosine (5’-dAdo) (2, 8).

PFL-AB was first isolated by Knappe et al. from non-overexpressing E. coli cells
(2). The E. coli cells were grown anaerobically, but the enzyme was purified aerobically

and shown to be a 28 kDa monomer. A broad absorbance from 310 to 550 nm suggested

14

 

 

the presence of a covalently bound chromaphore. The enzymatic activity was shown to

be completely dependent on the presence of exogenous iron in the assay.

The first over-expression of PFL-AB in E. coli was reported by Wong et al. ( 12).
The majority of the overexpressed protein, however, was found in the form of insoluble
inclusion bodies. PFL-AE was purified from these inclusion bodies by denaturation in
guanidine-HCI, followed by gel-filtration chromatography and refolding under anaerobic
conditions. Fe(II) was not required for refolding but could be included with a
stoichiometry of approximately 1:1, and was again an absolute requirement for activation
of PFL. The apo-enzyme could also be reconstituted with various divalent metals,
although none of these metal-substituted proteins showed activity in the absence of added
iron. Thiophilic metals such as Cu(H), Zn(II), Hg(II) and Cd(II) were found to be
inhibitors of PFL-AE activity, which was consistent with the proposal that cysteines are
the metal ligands.

The ability to isolate large quantities of native PFL-AB, without resorting to
denaturation and artificial reconstitution, is critical to understanding the nature of the iron
center in PFL-AE and its role in radical generation. The first isolation of PFL-AB in its
native state under strictly anaerobic conditions and identification of the presence of an
iron-sulfur cluster in PFL-AB were achieved by Broderick et al. (13, 14). When purified
under anaerobic conditions in the absence of DTT, PFL-AE was shown to contain
primarily a [3Fe-4S]+ cluster by a combination of UV-visible, EPR and resonance Raman
spectroscopic methods (14). A complete description of the states of the cluster present in
the enzyme was provided by Messbauer spectroscopy (15), a method that can detect and

quantify all iron species in the samples. Detailed analysis of the Mbssbauer data

15

d "V. ‘— (Ear—m

 

 

 

 

 

indicates a mixture of Fe-S clusters with the cuboidal [3Fe-4S]+ cluster as the primary

cluster form, accounting for 66 % of the total iron. Also present are [2Fe-28]2+ (12 % of
the total Fe), [4Fe-4S]2+ (8 % of the total Fe) and linear [3Fe-4S]+ (~10 % of the total Fe).
The isolated native enzyme has nearly a full complement of Fe-S cluster with
approximately 3 irons and 3 sulfides per protein molecule. It exhibits a high specific
activity (95 U/mg) in the absence of added iron, while the apo-enzyme exhibits no such
activity, indicating the cluster present in native enzyme is essential and sufficient for
enzymatic activity.

In the dithionite-reduced form of PFL-AE, all cluster types were converted into
the [4Fe-48] form (15). By analogy to aconitase (69), the [4Fe-48] cluster was expected
to be the catalytically relevant cluster, which is generated under the reducing conditions
present in the activity assay. Recently, Henshaw et al. have demonstrated that the [4Fe—
4S]1+ is the catalytically active cluster of PFL-AB and that it donates the electron required
for reductive cleavage of AdoMet ( 16). The purification of PFL-AB containing primarily
[3Fe-4S]+ clusters implies a labile fourth iron site, and is consistent with the observation
that only three cysteines have been implicated in cluster coordination (17). These three
cysteines exist in a CX3CX2C motif that is common to all of the AdoMet dependent Fe-S
enzymes for which a sequence has been determined (18—23). The identity of the fourth
ligand to the [4Fe-48] cluster‘in PFL-AB and these related enzymes is currently
unknown. By analogy to aconitase, in which substrate coordinates to the unique iron site
(70-72), and solvent binds in the absence of substrate or product (70), it is reasonable to
propose that exogenous ligands such as water or substrate may bind to the unique labile

iron site in PFL-AB.

l6

 

 

 

 

 

 

Currently, PFL-AB is purified under anaerobic reducing conditions in the
presence of DTT to yield essentially EPR-silent clusters, presumably in the [4Fe-4S]2+
state, which can be readily reduced to [4Fe-4S]1+ - the cluster that is responsible for

providing the electron necessary for AdoMet dependent glycyl radical generation on PFL.

11.2 Experimental methods
11.2.1 Materials

The plasmids pMG-AE and pKK-PFL were obtained as a generous gift from John
Kozarich (Merck). The Escherichia coli BL21(DE3)pLysS strain and pCAL-n-EK
expression vector were obtained from Stratagene. 5-Deazariboflavin was previously
synthesized in our laboratory according to published procedures (73-75), and
characterized using N MR and TLC. All other chemicals were obtained commercially

and used as received.

11.2.2 Growth and expression of PFL

pKK-PFL was used to transform BL21(DE3)pLysS. A single colony of
transformed cells was used to inoculate 50 mL LB media containing 50 pg/mL ampicillin
(LB/Amp). This culture was grown for 16 h to saturation and then used to inoculate
LB/Amp in a 9 L bench-top fermentor (New Brunswick). The 9 L culture was grown at

37 °C with continuous air purge and vigorous agitation to early log phase (OD600 ~ 0.6-

0.8), and then induced by addition of isopropyl-fl-D-thiogalactopyranoside (IPTG) to

17

 

 

 

 

 

 

 

 

 

 

I,

 

 

[\/

 

 

 

 

1 mM. The culture was grown for 2 more hours before harvesting by centrifugation at

10,816 x g (8,000 rpm, Sorvall GS3 rotor). The supernatant was decanted and the cells

stored at -80 °C.

11.2.3 Purification of PFL

PFL was purified from the BL21(DE3)pLysS/pKK-PFL cells. Cell paste
(typically approximately 10 g) was suspended in enzymatic lysis buffer (5 mL per gram
of cell paste) containing 20 mM Hepes, pH 7.2, l % (w/v) Triton X-100, 5 % (w/v)
glycerol, 10 mM MgC12, 8 mg lysozyme, 1 mM PMSF, and trace amounts
(approximately 0.1 mg each) RNase A and DNase I. The suspension was agitated and
then incubated at ambient temperature for 1 hour. The lysed cells were centrifuged at
26,892 x g (15,000 rpm, S834) for 20 min at 4°C. The crude extract (typically
approximately 50 mL) was decanted and loaded onto an Accell Plus QMA Anion
Exchange column (Quaternary Methylamine, 300 A, Waters Corp., 5 x 30 cm)
equilibrated with Buffer A (20 mM Hepes, pH 7.2, 1 mM DTT). The column was
washed with 300 mL of the same buffer prior to running a gradient from Buffer A to
Buffer B (20 mM Hepes, pH 7.2, 500 mM NaCl, 1 mM DTT) over 900 mL. PFL eluted
at approximately 240 mM NaCl. Fractions containing 2 75 % pure PFL (as judged by
SDS-PAGE on a 5 % - 14 % Tris-HCl gel) were combined, flash frozen and stored at
-80 °C. Another 50 mL of crude extract was run through the same procedure, and the
2 75 % pure fractions from both runs were combined, dialyzed against Buffer C (40 mM
Hepes, pH 7.2, l M ammonium sulfate, 1 mM DTT), and centrifuged to remove

precipitated protein. The supernatant was loaded onto a Phenyl-Sepharose column

18

 

 

 

 

 

 

 

(Pharmacia 16/10) equilibrated with Buffer C. The column was washed with 50 mL of

Buffer C prior to running a gradient from Buffer C to Buffer A over 50 mL, followed by
a wash with 50 mL of Buffer A. PFL was eluted through the last half of the gradient.
Fractions containing 2 95 % pure PFL (as judged by SDS-PAGE) were combined,

dialyzed against Buffer A, concentrated, flash-frozen and stored at —80 °C.

11.2.4 Growth and expression of PFL-AE

A single colony of transformed BL21(DE3)pLysS/pCAL—n-AE3 was used to
inoculate 50 mL LB/Amp. This culture was grown for 16 h to saturation at 37 °C and
then used to inoculate 9 L of defined MOPS medium based on one previously described
(76). The media was modified to include (per 9 L) 75.6 g MOPS, 99 g casarrrino acid,
7.2 g tricine, 26.3 g NaCl, 14.4 g KOH, 4.6 g NH4C1, 200 mL of 20 % glucose, 20 mL of
“0” solution, 20 mL of l M KH2P04, 10 mL of 276 mM K2804, and 50 mL of 0.1 M
CaClz. The “0” solution consists of 0.1 g FeC12-4H20 in 10 mL concentrated HCl, 1 mL
of “T” solution which includes 8 mL concentrated HCl, 18.4 mg of CaC12-2H20, 64 mg
H3B03, 40 mg MnClz-4H20, 18 mg CoC12-6H20, 4 mg CuClz-2H20, 340 mg ZnClz, 605
mg Na2M004-2H20, and 2.68 g MgC12-6H20. Ampicillin (9 mL of 50 mg/mL) and 10
mg each of riboflavin, thioctic acid, vitamin B 12, niacinamide, pantothenic acid,
piridoxine and folic acid were added to the 9 L culture right before the inoculation. The
9 L culture was grown at 37 °C in a bench-top fermentor (New Brunswick) with a
continuous air purge and vigorous agitation to early log phase (OD600 ~0.5), and then
induced by addition of IPT G to 1 mM. At this time the medium was supplemented with

750 mg Fe(N1-14)2(SOa)2. The culture was allowed to grow for an additional 2 h, purging

19

with air before cooling down and purging with nitrogen for 20 min. The medium was
then supplemented with another 750 mg Fe(NHa)2(SO.t)2, When the temperature was

reduced to 20 °C, the culture was incubated for 14-17 hours at 4 °C under argon. The
cells were harvested by centrifugation at 10,816 x g (8,000 rpm, GS3) under anaerobic

conditions. The harvested cells were stored under nitrogen at —80°C until used for

purification.

11.2.5 Purification of PFL-AE

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 or purged with nitrogen prior to
bringing them into the Coy chamber. Cell paste was suspended in enzymatic lysis buffer
(2 mL per gram of cells) containing 50 mM Tris-sulfate, pH 7.5, 200 mM NaCl, 1 %
Triton X-100, 5 % glycerol, 10 mM MgC12, 1 mM DTT, 8 mg lysozyme, 1 mM PMSF
and trace amounts (approximately 0.1 mg each) of RNase A and DNase I. The
suspension was agitated with a 10 mL syringe and then incubated at ambient temperature
for 1 h. The suspension was centrifuged at 38,724 x g (18,000 rpm, 8834) for 30 min at
4 °C. The extract was decanted and used directly in purification. Up to 30 mL of the
crude extract was loaded onto a Sephacryl S-200 HR column (5 x 60 cm) equilibrated
with 50 mM Hepes, pH 7.5, 200 mM NaCl, 1 mM DTT. The protein was eluted with this

same buffer at 3 mUmin. PFL—AE eluted from the column in a relatively sharp peak at

20

 

 

 

 

 

a‘PIJI'OXimately 680 mL after injection. The fractions that had fairly dark red/brown color

were pooled and concentrated down to less than 10rnL using an Arnicon concentrator
with YM10 filter membranes. Another aliquot of crude extract was run through the same
procedure. Fractions pooled from both runs were combined, concentrated to less than 20
mL, and re-run on the same column as above. The final fractions were checked by the
ratio of UV absorbance at 426 nm and 280 nm. Fractions that have the highest ratio
(greater than 0.15, but fractions with a blue-shifted 280 nm peak, indicating nucleic acid
contaminants, were avoided) were pooled, concentrated, flash-frozen, and stored at

—80 °C.

11.2.6 Protein assays

Routine determinations of protein concentrations were done by the method of
Bradford (77), using a kit purchased from Bio-Rad, and bovine serum albumin as a
standard. Calibration of the results from the Bradford assays of PFL-AE was obtained by
amino acid hydrolysis of the purified enzymes, performed at the MCB Core Facility,
University of Massachusetts, Amherst. Actual protein concentrations could then be

determined by applying a correction factor of 0.65 to the Bradford assays.

11.2.7 Iron assays

Iron assays were carried out by using the method of Beinert (78).

11.2.8 Sulfide assays
Sulfide assays were carried out with a modification of the method of Beinert (14,

79). The use of siliconized Eppendorf tubes was found to yield more reproducible

21

 

 

 

 

 

results, perhaps due to the minimized head space for loss of sulfide as H2S. The tubes

were kept tightly capped except when adding reagents. Rather than using stir bars, the

tubes were closed and vortexed when mixing was called for. The procedure used was as
follows. The sample volumes were brought to 100 ,uL with MQ H2O, pH 8.1. One at a

time, each tube was opened, 300 ,uL l % ZnOAc and 15 ,uL 12 % N aOH were added
simultaneously, the tube was closed tightly and vortexed. When all tubes had been
treated in this way, they were allowed to sit for 12 to 15 h before addition (again, one
tube processed at a time) of 75 ,uL DMPD (0.1 % in 5 M HCl) and 2 ,uL Fer (23 mM in
1.2 M HCl). Na2S-9H2O was used as standard and prepared as follow: a small-to-
medium chunk of Na2S-9H2O was rinsed with MQ H2O, pH 8.1, dried by gently patting
with Kim-wipe, weighed to 4 decimal places, and dissolved in 100 mL deoxygenated

0.1 M NaOH solution to make a standard solution of concentration between 1.2 to

2.7 mM. The standard solution was sealed with septum, purged thoroughly with

nitrogen, and was good for at least a month.

11.2.9 Activity assay of PFL-AE

The activity of PFL-AB was assayed using a direct enzyme assay in which the
amount of glycyl radical on PFL generated by PFL-AB as a function of time was
quantified by EPR spectroscopy. The PFL-AB reaction mix contained in a final volume
of 1 mL: 0.1 M Tris-HCl, pH 7.6, 0.1 M KC], 10 mM DTT, 10 mM oxamate (allosteric
effector), 10.01 mg PFL, 200 ,uM 5-deazariboflavin, 0.2 mM AdoMet, and 0.186 #M
PFL-AB. This mix was made in an anaerobic chamber by combining reagents from

anoxic stock solutions in the order listed to the final concentrations indicated. The 1 mL

22

 

 

 

 

mix was split into 5 EPR tubes and the reaction was initiated by illumination of the
samples with a 500 W halogen lamp. The samples were situated at a distance of 5 cm
from the lamp and maintained at ambient temperature (20-26 °C) during illumination by
immersion in a water bath to which ice was added as needed. After specified periods of
illumination, typically 5, 10, 20, 30, and 40 min, the samples were flash-frozen to stop
the reaction. The amount of glycyl radical generated was determined by EPR
spectroscopy. One unit of PFL-AB activity was defined as lnmol of glycyl radical
generated per min, and the specific activity of PFL-AB was defined as the number of

units per mg of PFL-AE.

11.2.10 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. Spectra were recorded at 12 K for [3Fe-48]” and [4Fe-4S]l+, and at
60 K to detect glycyl radical. Spin quantifications were done as described previously
(80). The double integrals of the EPR signals were evaluated by using a computer on-
line with the spectrometer. Spin concentrations in the protein samples were determined
by calibrating double integrals of the EPR spectra recorded under nonsaturating
conditions (i) with a standard sample of 0.1 mM Cu(II) and 1 mM EDTA solution for the
cluster signals, or (ii) with a 1.04 mM K2(803)2NO solution for the glycyl radical signals.
The concentration of the K2(SO3)2NO standard was determined using the optical

extinction coefficient (81).

23

 

 

 

 

“-3 Results and Discussion

11.3.1 Expression and purification of PFL

PFL was purified from BL21(DE3)pLysS cells harboring the pKK-PFL plasmid.
The cells were lysed by an enzymatic procedure. Two portions of partially purified PFL
from the first ion-exchange column (Chromatogram see Fig. 11.1; SDS-PAGE gel see
Figure 11.2) were combined, dialyzed, and run on an hydrophobic column to yield 2 95%
pure PFL (Chromatogram see Fig. 11.2; SDS-PAGE gel see Figure 11.3). Yield was

approximately 50 mg of purified PFL per liter of bacterial culture.

24

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.5 ——1
2 100% BufferB 7 films,“
3 1.5 4 ‘ ——Wabs
; ? —Conductiv'ly
D 1 " ‘ Gradient Pimp
0.5 -
o 1 1 1 l 1 l l J
0 200 400 600 800 1000 1200 1400
Volume (mL)
3' 2.5 r... — - »— ~ 1
2 140m$lcm 100% Buffer 13
i
g 1.5 ——Wabs
; -———Conductivity
D 1 ‘ Gradient Rum
0.5 ;
o i Y 1 I l T ‘l’ ‘
0 20 4o 60 80 100 120 140
l Volume (mL)
Figure H.l Chromatograms of purification of PFL. A. Elution profile for the Accell

Plus QMA Anion Exchange column (Quaternary Methylamine, 300 A, Waters Corp, 5 x
30 cm), gradient from Buffer A (20 mM Hepes, pH 7.2, 1 mM DTT) to Buffer B (20 mM
Hepes, pH 7.2, 500 mM NaCl, 1 mM DTT) over 900 mL, flow rate 5 mL/min. PFL
eluted at approximately 240 mM NaCl, and was dialyzed against Buffer C (40 mM
Hepes, pH 7 .2, 1 M ammonium sulfate, lmM DTT). B. Elution profile for the Phenyl-

Sepharose column (Pharmacia 16/10), gradient from Bufi‘er C to Buffer A over 50 mL,
followed by a wash with 50 mL of Buffer A, flow rate 1 mL/min. PFL was eluted
through last half of the gradient. (Images in this thesis are presented in color.)

25

w

123 4 5 67 8 9101112131415

W

.f E, _,

fioo— ...

115 —-
97.4 —- - ,__

66 '— ~-
45 '— a.»

 

31—”.

215—»
145-:

Figure 11.2 SDS-PAGE analysis of PFL fractions off the Accell Plus QMA anion-

exchange column. Lane 1, molecular marker (kDa); lane 2-15, fractions 19-45 (every

other one). Fractions 23-43 were pooled.

6 7891011

12345

 

Figure 11.3 SDS-PAGE analysis of PFL fractions off the Phenyl-Sepharose
hydrophobic column. Lane 1-11, fractions 7-27 (every other one). Fractions 11-21 were

pooled.

26

~fi

v

 

 

 

“-32 Expression and purification of PFL—AE

PFL-AE was purified from BL21(DE3)pLysS cells harboring the pCAL-n-AE3
expression vector. Overexpression of the protein in the pre—induced and post-induced
cells were checked by SDS-PAGE on a 12 % Tris-HCl gel (Figure 11.4). The cells were

1 2 3
P!

97.4 Fl »

66 g j

45 y

31 . ‘fi .
215.

Figure 11.4 SDS-PAGE analysis of overexpression of PFL-AE in E. coli cells. Lane 1,
molecular marker (kDa); lane 2, pre-induced cells; lane 3, post-induced cells.

lysed by an enzymatic procedure, and PFL-AE was purified from this crude extract by
two passages through a preparative gel filtration column (Figure 11.5). The enzyme
eluted as a reddish-brown peak, and pure fractions from the final run of the column were
identified by the highest ratio of absorbance at 426 nm and 280 nm. Both [3Fe—4S]'+ and
[4Fe-4$]2+, the major cluster forms found in purified PFL-AB, have significant
absorbance at 426 um. All proteins have maximal absorbance at 280 nm. The high iron
content in the enzyme, thus, is indicated by the high ratio of the absorbance values (Table
H. 1). Fraction purity was confirmed by SDS-PAGE (Figure H6), and the purest fractions
were combined, concentrated, and stored under nitrogen in small aliquots at —80 °C.

Yield was typically 40-50 mg of purified PFL-AE per liter of bacterial culture.

27

1m; "

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.6

0.5 -+
E 0.4 ~
2
O
3 0.3 -
Q)
m -
'8 0'2 Run 1
>
3 0.1 +

Run 2
o J ”I I I T r fl
(5 200 400 soo 800 1000 1200
-0 1
Volume (mL)
Figure 11.5 Purification of PFL-AB by gel-filtration chromatography (Sephacryl S-

200 HR column, 5 x 60 cm). The protein was eluted with 50 mM Hepes, pH 7.5, 200 mM
NaCl, 1 mM DTT at 3 mL/min. PFL-AE eluted from the column in a relatively sharp
peak at approximately 680 mL after injection. The chromatograrn labeled “ Run 2”
shows the final purification of the pooled fi'actions from two first runs on the column.

 

 

 

 

 

 

Fraction Ratio 426 nm / 280 nm Abs at 426 nm Abs at 280 nm
17 0.085 0.022 0.255
18 0.128 0.046 0.359
19 0.154 0.067 0.435
21 0.184 0.069 0.378
23 0.213 0.046 0.216
25 0.153 0.019 0.125
27 0.101 0.009 0.091
28 0.095 0.008 0.080

 

Table 11.1

2nd run of the gel filtration column.

Ratio of absorbances at 426 nm and 280 nm of PF L-AE fi'actions off the

 

 

 

 

 

12345678

97.4 Not
56 w

45 ...
31 "w , -'

215 i
145 4

Figure 11.6 SDS-PAGE analysis of PFL-AE fractions off the 2nd run of the gel
filtration column. Lane 1, molecular marker (kDa); lane 2-8, fractions 17, 18, 19, 20, 22,

24, 26. Fractions 19-25 were pooled.

11.3.3 Characterization of purified PFL-AE

PFL-AB purified under anaerobic reducing conditions in the presence of lmM
DTT was essentially EPR-silent, presumably in the [4Fe-4S]2+ state, containing
approximately 2 % [3Fe-4S]1+ as indicated by EPR spectroscopy (Figure H7). The

protein was found to contain 2.5-3.8 mol Fe / mol protein and a stoichiometric amount of

acid-labile sulfide.

29

   

 

 

 

 

 

 

 

 

 

 

 

EPR Intensity

 

 

 

 

3000 3200 3400 3600 3800
Field (Gauss)

 

 

 

Figure 11.7 X-band EPR spectrum of PFL-AE as-isolated in the presence of 1 mM
DTT. Protein concentration, 400 ,uM; [3Fe-4S]1+ concentration, 8.3 ,uM, g _L = 2.017,

g,, = 2.038. Conditions of measurement, T = 12 K; microwave power, 2 mW; microwave

frequency, 9.49 GHz; modulation amplitude, 9.57 G; modulation frequency, 100 kHz,

single scan.

Using purified PFL and PFL-AB, a specific activity for PFL-AE of 109 units / mg
was achieved using the direct activity assay (Figure 11.8), and the specific activity per

holoenzyrne was calculated to be 115 U/mg based on 3.8 mol Fe/mol protein (Table 11.2).

Rate UM glcyl radical] min) ‘ 0.5697
Units (nmol PFL/ min) 0.5697
Specific Activity (U/ mg PFL-AB) * _ ’ 109 _
Number of Fe in PFL-AE (mol Fe/ mol PFL-AE) 3.8

Specific Activity per holoenzyme (U/ mg) 115 _

Table 11.2 Summary of activity assay of PFL-AE

30

 

—'

 

 

 

 

 

 

 

 

. 40 min
30 min

. 20 min
, 10 min

EPR Intensity of
glycyl radical signals

 

' 5min

 

 

 

3000 3200 3400 3600 3800
Field (Gauss)

 

 

 

 

 

  

 

 

 

25
20 —
2" 15 -
a
3 10 -
5 j y= 0.5697x
R2 = 0.9876
0 r t t i
0 10 20 30 40 so

 

 

Time (min)

 

Figure 11.8 Activity assay of PFL-AB. The assay mix (1 mL) contained 10.01 mg/mL
(58.9 mM) PFL, 0.1 M Tris-HCl, pH 7.6, 0.1 M KCl, 10 mM DTT, 10 mM oxamate, 0.2
mM deaza, 0.2 mM AdoMet, and 0.186 M PFL-AB (3.8 mol Fe/ mol PFL-AB, 4.0 mol
S/ mol PFL-AB). The assay mix was illuminated for 5, 10, 20, 30, and 40 rrrin, and the
amount of glycyl radical generated was determined by EPR spectroscopy. The glycyl
radical is indicated by a doublet signal at g = 2.007. Conditions of measurement, T = 60
K; microwave power, 20 pW; microwave frequency, 9.47 GHz; modulation amplitude,

5.04 G; modulation frequency, 100 kHz. The spectra shown are single scans.

31

 

 

The optimization of the growth and purification conditions for PFL-AE has been a

long and evolving process in our laboratory. Inducing at an earlier log phase (OD600 ~
0.5 vs. ~0.8) and supplementing iron to the medium after making the culture anaerobic
contributed to better overexpression and iron inclusion of PFL-AB. The component of
holoenzyme in the isolated native enzyme that contains full complement of iron was also
improved in the purification procedure. The enzyme was purified by passing through the
gel filtration column twice under strictly anaerobic conditions in the presence of 1 mM
DTT, and pooling fractions based not only on SDS-PAGE analysis but also on the UV
absorbance ratio at 426 nm and 280 nm. In summary, under current growth and
purification conditions, approximately 400 mg of purified PFL-AB was obtained from a
9 L bacteria culture; this represents a 2-3-fold improvement over earlier procedures. The
PFL-AB contains 2.5 - 3.8 mol Fe/ mol protein and a stoichiometric amount of acid-
labile sulfide. The F-S clusters in the isolated protein are essentially EPR-silent,
presumably in the [4Fe-4S]2+ form, with about 2 % of [3Fe—4S]1+ cluster. A high specific
activity of 109 U/ mg enzyme was achieved for PFL-AB containing 3.8 mol Fe/ mol

protein.

32

 

CHAPTER III

SYNTHESES AND CHARACTERIZATION OF
ISOTOPICALLY LABELED S—ADENOSYL-L—METHIONINES
AND Se-ADENOSYL-L-SELENOMETHIONINE

111.1 Introduction

S-adenosyl-L-methionine (AdoMet) was first discovered and identified in 1953 by
Cantoni (82), and a great deal of research has been directed towards its synthesis and
biological functions since then. AdoMet plays various roles in biological systems. It is
the primary methyl donor in the methylation of DNA, RNA, proteins, lipids, vitamin B12
and many other substrates, by AdoMet-dependent methyltransferases (83). Alternatively,
AdoMet can undergo decarboxylation to act as three-carbon donor. The three carbon
arninopropyl group is incorporated into the polyamines, spermine and sperrnidine (84),
and is also a precursor of the plant hormone, ethylene (85). As a five-carbon donor, the
adenosyl moiety of AdoMet is the precursor of the cyclopentenediol of the tRN A wobble
base queuosine (86). AdoMet also has clinical applications as a therapeutic drug in the
treatment of liver disease, as a potential cancer chemo-preventive agent (87), and as an
antidepressant (88-90). Another important function of AdoMet resides in the newly
emerging group of AdoMet dependent iron-sulfur enzymes, where AdoMet participates
in radical generation, with evidence suggesting a common mechanism involving an
intermediate 5’-deoxyadenosyl radical (4, 34, 35, 55). A superfamily of more than 600
related enzymes that involve radical-based biochemistry has very recently been identified

and named the radical SAM superfamily (41).

33

 

 

 

 

 

 

 

 

 

During the PFL activation reaction, PFL-AE requires AdoMet as a cosubstrate,
cleaving it to methionine and 5’-deoxyadenosine (2). Knappe et al. have shown that the
hydrogen atom abstracted from Gly734 of PFL is incorporated into the 5’-dAdo product
(Scheme H1. 1), indicating that the role of AdoMet is to produce an adenosyl radical

H
1: Ad

    

HO OH

H

PFL-Glyo «1111/le

 

Scheme 111.1 Reductive cleavage of AdoMet by PFL-AE in PFL activation

intermediate as the immediate hydrogen atom abstractor (8). Recently, Henshaw et al.
have demonstrated that the [4Fe-4S]1+ is the catalytically active cluster of PFL-AE and
that it donates the electron required for reductive cleavage of AdoMet (16). The presence
of AdoMet has also been shown to dramatically affect the EPR signal of the [4Fe-4S]l+
of PFL-AE (15, 16), suggesting the possibility of a direct interaction between AdoMet
and Fe-S cluster of PFL-AE. The mechanistic question of how the Fe-S cluster of PFL-
AE interacts with AdoMet to generate the putative adenosyl radical, however, still
remains unclear. As the S-C bond is reductively cleaved during the radical generation,
the sulfonium center of AdoMet is likely to be positioned close to the Fe—S cluster

(Scheme HI.2). Using isotopically labeled AdoMets and AdoMet analogs in which

34

 

 

 

 

 

 

 

NHz H H3N coo N/ N
N/ N K I /> . +.
K I /> N CH2 ”“3
N N + H 0 AA
§\0H3 _ a 3 \ 000'
6 OH H
HO H F S—Fe
T\Fe 'e\s S\'\Fe—-' \s
Fe- —8 / \-_ \
s/ \s \Fe 3 S_F
\S S
Me 4511. [4Fe-4$]2*

Scheme 111.2 AdoMet and the Fe-S cluster of PFL-AE are proposed to be in close

proximity during radical generation.

atoms in the sulfonium center are substituted with nuclei that have non-zero nuclear
spins, hyperfine interactions with the Fe-S cluster of PFL-AB can be displayed if AdoMet
coordinates or is in close proximity to the cluster. In order to probe hyperfine
interactions between AdoMet and the cluster, the methyl group of AdoMet was labeled
with 2H and '3 C, and the sulfur was replaced by selenium to take advantage of the natural
abundance of 77Se (7.50 %). These isotopically labeled AdoMets and AdoMet analog are
effective probes in spectroscopic studies to demonstrate the possible interaction between
AdoMet and the Fe-S cluster of PFL-AE.

AdoMet contains a chiral sulfonium center and thus exists in two diastereo-
isomeric forms, (-)- and (+)-AdoMet. (-)-AdoMet is the only form enzymatically
synthesized (91), and the only methyl donor in vivo (92). Chemical coupling reactions
between adenosyl homocysteine and methyl group donating agents can produce AdoMet,
however the diastereoselectivity is poor (91). The formation of (-)-AdoMet is catalyzed

by S-Adenosylmethionine synthetase (ATPzL-methionine S-adenosyltransferase

35

 

 

EC2.S.1.6; AdoMet synthetase) (Scheme H13). AdoMet synthetase from E.coli has a
molecular weight of 180 kDa, and the native enzyme is a tetramer of 4 identical subunits

(93).

2+ .K 3H3 (IN I \Tfl
'o c 3 Mg , K+ up 5; , N
2 \ + AT + N

Scheme 111.3 Enzymatic synthesis of S-adenosyl-L-methionine

Kinetic experiments have demonstrated that the enzymatic reaction proceeds sequentially
in 4 steps: (1) random addition of methionine and MgATP; (ii) formation of AdoMet and
tripolyphosphate; (iii) oriented cleavage of tripolyphosphate to yield orthophosphate (Pi)
and pyrophosphate (PPi); (iv) product release with orthophosphate and pyrophosphate
dissociation before AdoMet (93). Two divalent metal ions, such as Mg”, bind to the
enzyme active site in the presence of ATP, and both are required for activity (82, 94, 95).
A single divalent cation, such as K“, binds in the presence of AdoMet, and stimulates the
rate of AdoMet formation up to 100-fold (93).

Prior to a report by Park et al. (96), preparative scale synthesis of AdoMet was
infeasible, due to product inhibition at substrate concentrations >1 mM, which may result
from the formation of inactive AdoMet-bound-enzyme complexes (93-95, 97-99). The
inhibition problem was successfully overcome in incubations using 10 mM substrate in

the presence of various additives, including p—toluenesulfonate (p-TsONa), and more

efficiently, a high concentration of ,B-mercaptoethanol (fl-ME), acetonitrile or urea (96).

36

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

In the work described here, 8 % ,B—ME was used as additive to overcome product
inhibition. Pyrophosphate also acts as an inhibitor of AdoMet synthetase (93). Though it
is a weaker inhibitor than AdoMet, it is still strong enough to show an inhibitory effect in
reactions containing > 1 mM substrate. Therefore, a small amount of pyrophosphatase
was added to the reaction mixture to hydrolyze pyrophosphate to orthophosphate, a

weaker non-competitive inhibitor of AdoMet synthetase (93).

111.2 Experimental methods
111.2.1 Materials

AdoMet synthetase overproducing strain DM22 (pK8) was a general gift from Dr.
George D. Markham (Fox Chase Cancer Center). Methyl-Dy L-methionine and methyl-
l3C-L-methionine were purchased from Isotech Inc. Seleno-L-methionine, ATP, and
inorganic pyrophosphatase were purchased from Sigma. These and all other chemicals

were of the highest purity obtainable from commercial sources.

111.2.2 Growth and purification of AdoMet synthetase
Expression of E. coli AdoMet synthetase

AdoMet synthetase overproducing strain DM22 (pK8) was stored in 50 %
glycerol at —80 °C. A single colony of transformed cells was used to inoculate 50 mL LB
media containing 30 yg/mL oxytetracycline (LB/Tet). This culture was grown for 12 -
14 h to saturation and then used to inoculate 700 mL LB/Tet in each of 5 2800 mL
Fembach culture flasks. The culture was grown at 37 °C with vigorous shaking for 12 -
14 h before harvesting by centrifugation at 10,816 x g (8,000 rpm, GS3 rotor). The

supernatant was decanted and the cells stored at -80 °C.

37

 

 

Preparation of cell lysate for AdoMet synthesis

Cell paste was suspended in 100mM Tris-HCl, pH 8.0 containing 1 mM EDTA (3
— 3.5 mL/g cell). Lysozyme was added to 50 ,ug/mL and the suspension was incubated at
room temperature for 30 min. Phenylmethylsulfonyl fluoride (PMSF) was added to a
final concentration of 0.1 mM. Cells were lysed by sonication (10 cycles of l min
sonication with 1 min rest in between) in an ice bath. The suspension was centrifuged at
26,892 x g (15,000 rpm, 8834) for 20 min. The supernatant was decanted and the cells
stored at —80 °C until needed. Protein concentrations were determined by the method of

Bradford (77), using a kit purchased from Bio-Rad.

Purification of AdoMet synthetase

Preparation of cell lysate: Cell paste was suspended in 100 mM Tris-HCl, pH 8.0
containing 10 % glycerol, 1 mM EDTA, 0.1 % fl-ME (4 mL/g cell). PMSF was added to
a final concentration of 0.1 mM, followed by lysozyme (50 ,ug/mL) and trace amount of
DNase and RNase (about 0.1 mg each). The lysate was incubated in ice for 1-1.5 hours,
and then centrifuged at 10,816 x g (8,000 rpm, G83) for 20 min. The supernatant was
decanted and used for the next step.

Ammonium sulfate fractionation: For each 100 mL of the supernatant solution,
22 g of ammonium sulfate were added. After stirring at 4 °C for 20 min the suspension
was centrifuged. The supernatant was decanted and for each 100 mL of the supernatant,
12 g of ammonium sulfate were added. After stirring at 4 °C for 20 min the precipitated

protein was collected by centrifugation.

38

 

 

Phenyl-Sepharose chromatography: The protein was dissolved in 10 mM Tris-

HCl, pH 8.0, containing 10 % glycerol, 1 mM EDTA, 0.1 % ,B-ME, and 0.75 M
ammonium sulfate. The protein solution was then loaded onto a Phenyl-Sepharose
column (Pharmacia 16/10), which had been equilibrated with the above buffer. The
column was washed with 2 volumes of the above buffer prior to running a reverse linear
gradient of 0.75 M to 0 M ammonium sulfate in 200 mL 10 mM Tris-HCl, pH 8.0
containing 10 % glycerol, 1 mM EDTA, 0.1 % fi-ME. Fractions containing AdoMet
synthetase, which were eluted at the middle of the gradient (as judged by SDS-PAGE on
a 12 % Tris—HCI gel), were pooled and concentrated in an Amicon concentrator with

YM30 filter membranes. The purified protein was flash-frozen and stored at —80 °C.

111.2.3 Synthesis and purification of isotopically labeled AdoMets and Se-AdoMet

AdoMet synthesis reactions (10 mL) were carried out at room temperature with
moderate stirring in 100 mM Tris-HCl, pH 8.0 containing 50 mM KCl, 26 mM MgCl2,
13 mM ATP, 1 mM EDTA, 8 % ,B-ME, 10 mM isotopically labeled methionine or
seleno-L- methionine, a small amount of inorganic pyrophosphatase (about 0.25 U) and
1 mL AdoMet synthetase lysate (approximately 13 mg of total protein) . All reagents
were added in the order as listed. The reaction was monitored by thin-layer
chromatography (TLC) on silica plates developed in butanol /acetic acid /water (4:1:1).
The reaction was terminated by addition of 1 mL of l M HCl and precipitated protein
was removed by centrifugation at 26,892X g (15,000 rpm, SS34) for 20 min at 4 °C. The
supernatant was decanted and split in half. Half of the supernatant was loaded onto a
SOURCE lSS cation exchange column (Pharmacia, 8 mL), which had been charged with
1 M HCl and equilibrated with MQ H2O. The column was run with a linear gradient of

39

 

0‘1 M HCl, and AdoMet eluted from 38 - 56 % of the gradient as a distinct peak. The

other half of the supernatant was run through the same procedure. Fractions containing
products were pooled, lyophilized, and stored at —80 °C until needed.
Adenosyl-methyl-D3-methionine (CD3-AdoMet) was synthesized overnight as
described above with 10 mM methyl-D3-L-methionine.
Adenosyl-methyl—'3C-methionine (”CHyAdoMet) was synthesized overnight as
described above with 10 mM methyl-'3C-L-methionine.
Se-adenosyl-L-selenomethionine (Se-AdoMet) was synthesized either at room

temperature for 2-3 hours or at 4 °C for 6-7 hours with 10 mM seleno-L-methionine.

III.2.4 NMR spectroscopy
lH, 13C and 2H NMR spectra were recorded at room temperature on a Varian

Inova-300 or a VXR—300 spectrometer (300.11, 75.43 and 46.04 MHz respectively).

III.2.5 Mass spectroscopy

Mass spectral data were obtained at the Michigan State University Mass
Spectrometry Facility, which is supported, in part, by a grant (DRR-00480) from the
Biotechnology Research Technology Program, National Center for Research, National

Institutes of Health.

40

 

111.3 Results and Discussion

111.3.1 Growth and purification of AdoMet synthetase
AdoMet synthetase was purified from DM22(pK8) overproducing strain.
Overexpression of the protein in the E. coli cells was checked by SDS-PAGE on a 12 %
Tris-HCl gel (Figure 111.1). Typical yields were 5-8.5 g of cells per liter of bacterial
culture. The cell lysate for AdoMet synthesis was prepared by a conjunction of
enzymatic and physical lysis procedures. The purity of the crude extract was checked by
SDS-PAGE on a 12 % Tris-HCI gel (Figure 111.1). The crude extract was concentrated to
approximately 13 mg/mL for AdoMet synthesis.

123456
A

97.4 1:13» B. I 2
66 w p 974“-
»‘t s‘ 45 .. t...
31 ... ‘-'
31 --~
215 .. q 215.
145 ,
3 -- ---~ ...... 145v

Figure 111.] SDS-PAGE analysis of AdoMet synthetase. A. Overexpression of
AdoMet synthetase in E. coli cells(no induction is required). Lane 1, molecular marker
(kDa); lane 2-6, duplicate samples. B. AdoMet synthetase crude extract. Lane 1,

molecular marker (kDa); lane 2, crude extract.

AdoMet synthetase was further purified from the crude extract by ammonium

sulfate fractionation and Phenyl-Sepharose chromatography. Purified protein was > 99 %

41

pure as judged by SDS-PAGE on a 12 % gel (Figure HI.2). Yield was 16.7 mg of

purified protein from 23 g cells.

97.‘ ......
66 'I"
45 W.““-_-~~H““
31 ......
215 a---
145 ’—

Figure [11.2 SDS-PAGE analysis of AdoMet synthetase fractions off the Phenyl-
Sepharose column. Lane 1, molecular marker (kDa); lane 2-10, fractions 3240.

Fractions 33-38 were pooled. ‘
III.3.3 Syntheses and purification of isotopically labeled AdoMets and Se-AdoMet
Intitial AdoMet synthesis reactions were carried out with purified AdoMet
synthetase, using natural abundence methonine and ATP as substrates, but
pyrophosphatase was not added in the reaction mixture. The reaction was run for 5 h at
ambient temperature according to the literature procedure(96), and the reaction progress
was monitored by TLC on cellulose developed in butanol /acetic acid /water (25:4: 10)
(92). The Rf value of ATP is 0 under these condition, and due to its high polarity,
AdoMet also has a very small Rf value, reportedly being 0.04 by descending TLC
developed in butanol /acetic acid /water (5:4: 1) (82). The reaction did not proceed to the
formation of AdoMet as indicated by TLC. It was then found, in some cases, that the
crude extract of AdoMet synthetase worked better than the purified enzyme, as a low Rf
value spot was observed only on the TLC of the reaction system containing crude extract

of AdoMet synthetase. Though the reason for the crude extract of AdoMet synthetase

42

performing better than the purified enzyme is not clear, it was speculated that the
pyrophosphate inhibition was partially overcome by the endogenous pyrophosphatase
present in the cell lysate. Some other unknown component in the cell lysate might as
well be beneficial to the performance of the enzyme. Therefore, crude extract of AdoMet
synthetase was used for the reaction. Not only does it save the labor and time to purify
the enzyme, which gives a rather low yield of purified enzyme, but it works very well in
terms of yield and purity of the AdoMets synthesized. In addition to using crude extract,
a small amount of inorganic pyrophosphatase was introduced into the reaction system to
facilitate the hydrolyzation of pyrophosphate to orthophosphate, which is a weaker non-
competitive inhibitor of AdoMet synthetase.

The TLC for monitoring the reaction progress was improved by using butanol

/acetic acid /water (4:4: 1) as the mobile phase and silica plates as the stationary phase to
get a better development of AdoMet and a clearer observation of the spots under
ultraviolet light. The Rf value of AdoMet, both isotopically labeled and Se substituted
AdoMet, was 0.12. Unreacted ATP didn’t develop on the plates (Rf: 0), and several
faint spots were also observed from the reaction mixture, which were presumably
contaminants and decomposed products. Longer reaction time (overnight instead of 5 h)
was also found to help allow the reaction to go to completion. The isolation of AdoMet
from the reaction mixture by chromatography was achieved on a Source SIS (Pharmacia)
cation exchange column, after various other cation exchange columns, including Dowex-
50W, Amberlite XE-64 and Mono S (HR 5/5, Pharmacia), had been tried without

SUCCESS.

43

Taking all the previous experience together, the AdoMet synthesis reaction is
currently carried out as described in the Experimental Methods. Isolated AdoMets were
lyophilized to yield crystalline colorless solids, presumably in the chloride salt form. The
purified AdoMet was analyzed by 1H NMR to check its diastereoisomeric purity, as the

methyl group resonances of (-)- and (+)- forms are well resolved (A6 ~ 0.4 ppm). Mole
ratios of approximately 94:6 for (-)- / (+)-AdoMet were obtained based on integrations of
the two peaks in 1H NMR spectra. Since (-)-AdoMet is the only form enzymetically
synthesized, the contamination of (+)-AdoMet in the purified product results from the
method of isolation and the temperature at which the reaction was done, which could
cause epimerization of the natural product. This was previously demonstrated by the
observation of only (-)- isomer in the HPLC analysis of the enzymatic reaction mixture
before isolation (91). A virtually epimerically pure (-)-AdoMet was obtained in small
quantities via rapid enzymatic synthesis followed by HPLC purification.

CD3-AdoMet was synthesized overnight under normal conditions as for natural
AdoMet, and isolated by chromatography on a SOURCE 15S cation exchange column
(Figure 111.3). Yield was 38.5 mg (87.9 %) for a 10 mL of 10 mM reaction.
1H NMR (D20): 8 2.18-2.39 (m, HB & HB’), 3.32-3.42 (m, Hy& Hy’), 3.52-3.62 (m,
Hut), 3.76-3.98(m, H5’, H5” & H4’), 4.38-4.73 (m, H3’ & H2’), 6.01 (d, J=3.91Hz, Hl’),
8.32 (5, H2 & H8) ppm. The 2H labeling was confirmed by the absence of a methyl peak

in the 1H NMR (Figure 111.6), as well as the presence of a methyl peak in the 2H NMR.

 

 

 

 

 

 

 

 

 

 

 

 

Y 2.5
100% B
2 - 380mS/ctm
CDa'AdOMet
m 1.5 ~ -———-Wabs
.9
g h \ —-———-Conductiviy
D 1 “ 4+ Gradient amp
0.5 k /
O I L11! ... F 1 I - 1
O 20 40 60 80 100 120
Time (min)

 

 

 

Figure [11.3 Chromatogram of isolation of CDg-AdoMet by SOURCE 15S cation
exchange chromatography (Pharmacia, 8 mL), gradient from 0-1 M HCl, flow rate
1 mL/min. CDg-AdoMet eluted from 38 -— 56 % of the gradient as a distinct peak.

l3CHg-AdoMet was also synthesized overnight under normal conditions as for
natural AdoMet, and isolated by chromatography on 3 SOURCE 158 cation exchange
column (Figure 111.4). Yield was 37.5 mg (86.0 %) for a 10 mL of 10 mM reaction.
1H NMR (D20): 5 2.20-2.35 (m, HB & HB’), 2.62 & 3.11 (d, J=l46.49, (-)-”CH3), 2.59
& 3.08 (d, J=146.49, (+)-‘3CH3), 3.37-3.43 (m, Hy & Hy’), 3.44-3.58 (m, Ha), 3.77—4.01
(m, H5’, H5” & H4’), 4.38-4.71 (m, H3’ & 1-12’), 6.01 (d, J=3.91, Hl’), 8.32 (5, H2 &
H8) ppm. The 13 C labeling was confirmed by the distinct large splitting of the methyl

peak in the 1H NMR Gigure 111.6), as well as the presence of a 13 C peak in the 13C NMR.

45

 

 

‘ 2.5T——— »» 7 7* ,., 7” ,_,. a.

l3CHz-AdoMet

\

        

 

o I I l l 1 Ti
0 20 40 60 8O 1 00 1 20

Time (min)

 

 

 

 

Figure III.4 Chromatogram of isolation of 13CH-j—AdoMet by SOURCE 158 cation
exchange chromatography (Pharmacia, 8 mL), gradient from 0-1 M HCl, flow rate 1
mL/min. l3CI~I;;-AdoMet eluted from 38 — 56 % of the gradient as a distinct peak.

In contrast to the isotopically labeled methionines, which behave identically to the
natural substrate methionine, selenomethionine has much different chemical properties
than its sulfiir counterpart. Selenomethionine reacts more rapidly than methionine in the
synthesis reaction, as has been reported for AdoMet synthetase from both yeast and E.
coli (93, 100, 101). The higher reactivity of selenomethionine can be explained as due to
the better nucleophilicity of selenium relative to sulfiir (102, 103). In addition, more
decomposed products and contaminants were found in the reaction mixture as more spots
were observed on TLC plates, and the contaminant peaks were of higher intensity in the
Chromatogram (Figure 111.5). Therefore, the Se—AdoMet synthesis reaction was carried
out either at room temperature for 2-3 h or at 4 °C for 6-7 h. Yields were 10.2 mg
(42.3 %) for a 5 mL of 10 mM reaction run at room temperature and 14.0 mg (58.1 %)

for a 5 mL of 10 mM reaction run at 4 °C respectively.

46

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.5
2 -
l Se-AdoMet
31.5 4 \ —---UVabs
a —-—Oonductivity
3 1 . Gradientmnp
0.5 "’
0 . T I
O 20 40 60 80 100 120
Time (min)

 

 

 

Figure III.5 Chromatogram of isolation of Se-AdoMet by SOURCE 15S cation
exchange chromatography (Pharmacia, 8 mL), gradient from 0-1 M HCl, flow rate 1
mL/min. Se-AdoMet eluted from 38 — 56 % of the gradient as a distinct peak. This

reaction was run at room temperature for 2 h.

1H NMR (D20): 5 2.21-2.33 (m, HB & HB’), 2.72 (s, (-)-CH3), 2.68 (s, (+)-CH3), 3.24-
3.34(m, Hy & Hy’), 3.40-3.50 (m, Ha), 3.69-3.92 (m, H5’, H5” & H4’), 4.37-4.73 (m,
H3’ & HZ’), 6.01 (d, J=3.91, Hl’), 8.32 (d, J=1.5, H2 & H8) ppm. The methyl peak of
Se-AdoMet shified upfield relative to the one in the natural abundence AdoMet, which is
consistent with the higher electron density of Se relative to S (Figure 111.6).

MS (NBA, +FAB)I III/Z 447 (C15H2305NGSC).

47

 

 

 

Lrlr‘vr9\alli"f¥'_!l5f,|f\-A'Il \g-gunrnunt .

 

v
...
r
.
A.

3203.. we 839$ merges: 28 8385852 onus—20m ”:85

3203-8 Ea .azogamu: 6202-60 3283‘ co 3% 83 «22 m3 3: 2:5

 

 

 

 

 

 

 

 

can
A: ON Qm oé 9m o6 ON 9w ad 3
a fit a 320v< ,
(a mrqjmamj ) WI Kl $20u<AQU
2. A
1 j HDEOU<IMIUM~
fi fi) 820v<uom
O
«I z .0
K
2/ 7

 

£93

«:2

48

In conclusion, a fairly simple procedure for enzymatic synthesis of AdoMet on a
preparative scale has been established. AdoMet synthetase was overexpressed and grown
in E. coli cells. The enzyme was lysed by a combination of enzymatic and physical
procedures, and the crude extract was used for AdoMet synthesis reaction without further
purification. Product inhibition was overcome by addition of 8 % of ,B-ME in the
reaction mixture. A 10 mM reaction can afford over 80 % yield of AdoMet. AdoMet
isolated by the method described here has high diastereoisomeric purity, and appears to
be more stable than AdoMet iodide salt from commercial sources. AdoMet from Sigma
shows more than one spot on the TLC plate, while purified homemade AdoMet shows
only one spot. In a NMR time course study, the NMR spectrum of a homemade natural
AdoMet sample, which was kept in D20 at room temperature for 6 days, was essentially
the same as the one that was taken right after this AdoMet was made, except for a small
decrease of the mole ratio of the (-)- and (+)- forms. The enhanced stability of
homemade AdoMet is probably due to the increased purity relative to AdoMet form
Sigma. In addition, the high acidity of the lyophilized product, which is eluted by HCl of
around 0.5 M, may help to stabilize AdoMet.

This method can be easily applied to isotopically labeled substrates, as has been
shown on methyl-D3-methionine, methyl-”C-methionine, and 1SN-methionine, which has
also been synthesized in our laboratory. With some modifications of the reaction
conditions according to the properties of the substrate analogs being used, AdoMet
analogs can also be made. In the case here, selenomethionine is a better substrate to
AdoMet synthetase in terms of reaction rate, but the product is more prone to degrade

during the synthesis, and the reaction time was thus much shortened to enhance the yield.

49

 

 

 

Combined with other synthetic techniques, more isotopically labeled AdoMets and
AdoMet analogs will be synthesized from a variety of isotopically labeled methionine
and ATP, or methionine and ATP analogs. These AdoMet analogs will be used as probes
to study the interaction of AdoMet with the Fe-S cluster of PFL-AB from different

perspectives.

50

 

 

 

 

 

 

 

 

CHAPTER IV

INVESTIGATING THE INTERACTION OF
ADOMET WITH PFL-AE
USING EPR AND ENDOR SPECTROSCOPY

IV.1 Introduction

PFL catalyzes the conversion of pyruvate and coenzyme A (CoA) to formate and
acetyl-CoA (1, 3, 104, 105). The active form of PFL contains a stable glycyl radical
located at residue 734 (4). This catalytically essential protein radical is generated post-
translationally by PFL-AB, under anaerobic conditions in a reaction that is strictly
dependent on the presence of AdoMet as a cosubstrate (66-68). During the PFL
activation reaction, the glycyl radical is formed by abstraction of the pro-S hydrogen
atom of Gly734 of PFL, and AdoMet is converted stoichiometrically to methionine and
5’-deoxyadenosine (Scheme 1.4) (2, 8). Isotope-labeling experiments have established
that the hydrogen atom abstracted from Gly734 is incorporated into the methyl group of
the 5’-dAdo product, suggesting that a 5’-deoxyadenosyl radical is the species that
abstracts a hydrogen atom from Gly734 (8).

PFL-AE is a representative of a rapidly growing group of enzymes that utilize
iron-sulfur clusters and AdoMet as required cofactors in radical generation (106). The
functions of these AdoMet-dependent iron-sulfur enzymes are diverse, including the
generation of stable protein radicals (4, 33, 34), as well as substrate radical intermediates
(35). However, a common key mechanism is proposed to be involved in the chemistry
catalyzed by these enzymes: the generation of an intermediate 5’-deoxyadenosyl radical
that initiates catalysis by hydrogen atom abstraction (4, 34, 35, 55).

51

 

 

 

 

   

 

 

A variety of iron-sulfur clusters, including [2Fe-2S], [3Fe—4S] and [4Fe-4S], have
been identified in these AdoMet-dependent enzymes (13, 14, 17, 34-39, 107). Mdssbauer
data have provided the first clear description of all the forms of the cluster present in
anaerobically isolated PFL-AE and in the dithionite-reduced form of the enzyme (15).
The anaerobically isolated enzyme in the absence of DTT contains a mixture of Fe-S
clusters with the cuboidal [3Fe-4S]'+ clusters as the primary cluster form, accounting for
66 % of the total iron. Other forms present include [2Fe-2S]2+ (12 % of total Fe), [4Fe-
4S]2+ (8% of total iron), and also a linear [3Fe-4S]1+ (~10 % of total iron). Reduction of
the as-isolated enzyme by dithionite converts all cluster types into the [4Fe-4S] form with
a majority of 2+ (66 % of total iron) and a small amount of 1+ (12 % of total iron)
oxidation states. The quantitation of the paramagnetic [4Fe-4S]1+ was also confirmed by
parallel EPR studies. Under current reducing purification conditions in the presence of 1
mM DTT, PFL-AE as-isolated is essentially EPR silent, presumably in [4Fe-4S]2+ state,
with a very small amount of [3Fe-4S]1+ accounting for approximately 2 % of total iron.

The observation that only cuboidal [4Fe-48] clusters are present upon reduction
of the enzyme suggests that the [4Fe-4S] cluster is the catalytically relevant cluster,
especially considering that PFL-AE activity is observed in vitro only under anaerobic
reducing conditions (1). It has been difficult, however, to identify unequivocally the
catalytically relevant cluster by which the 5’-dAdo radical intermediate is generated. A
[4Fe—4S]1+ has been shown to be the catalytically active for aRNR-AB (34, 108) and
LAM (35). A definite assignment of the [4Fe-4S]1+ of PFL-AE as the catalytically active
cluster was achieved in an elegantly designed single turnover experiment (16). In the

above work, deazariboflavin-mediated photoreduction of PFL-AB followed by addition

52

 

 

of AdoMet has been shown to afford primarily [4Fe-48] ”. Saturation of cluster

reduction is reached in 1 hour, at which time 85 % of the cluster in PFL-AB is reduced to
[4Fe-4S]” state. By removing illumination, the exogenous reductant is eliminated and
single turnover conditions can be achieved. Equimolar PFL is mixed with the [4Fe-
4S]”/AdoMet after removing illumination in a time-dependent manner. 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. No glycyl radical was observed when only
the [4Fe-4S]2+ cluster was present prior to the addition of PFL. This is the first direct
quantitative spectroscopic evidence that the [4Fe-4S]l+ of PFL-AB is the catalytically
relevant cluster, and demonstrates that this cluster provides the electron necessary for
AdoMet-dependent glycyl radical generation.

However, a detailed mechanism of this unprecedented reaction still remains
elusive. A central mechanistic question is whether the Fe-S cluster plays a simple
electron-transfer role in the reductive cleavage of AdoMet, or is more directly involved in
the radical generation reaction. In the absence of AdoMet, the [4Fe—4S]1+ of PFL—AB
gives rise to a rhombic EPR signal. However, in the presence of AdoMet, the observed
EPR signal is nearly axial. The observed effect of AdoMet on the EPR signal of the
catalytically relevant [4Fe-4S]1+ could reflect either a close association between AdoMet
and the cluster in the enzyme-substrate (ES) complex or allosteric perturbations of the
cluster.

In the work described here, the interaction between AdoMet and the [4Fe-4S]

cluster was probed by using electron nuclear double resonance (ENDOR) spectroscopy.

53

 

 

 

 

 

 

The methyl group on the sulfonium center of AdoMet has been chosen as the initial target

of study, because of its proximity to the S-C bond of AdoMet, where the cleavage takes
place during radical generation. The methyl group was labeled with 2H and 13C
respectively, both of which have non-zero nuclear spin and can display hyperfine
interactions if the atoms coordinate or are in proximity to the paramagnetic Fe-S cluster.
Electron-nuclear hyperfine interactions, which are too small to be resolved within the
natural width of the EPR line, are generally amenable to study by ENDOR spectroscopy.
To characterize the binding of AdoMet to PFL-AE, initial 35 GHz pulsed 2H and ”c

EN DOR studies were performed with labeled AdoMet. Binding of AdoMet to the cluster
in the equilibrium geometry associated with the paramagnetic 1+ state of the cluster
(denoted [1+/AdoMet]) was examined, as well as binding to the diamagnetic 2+ state
(denoted [2+/AdoMet]). In order to probe the binding of AdoMet to the diamagnetic
[4Fe-4S]2+ state, as-isolated PFL-AB samples were frozen in the presence of AdoMet and
then yirradiated at 77 K to produce the 1+ valency cluster trapped in the geometry of the
2+ state (denoted [2+/AdoMet],ed). In this cryoreduction technique, yirradiation of
solutions frozen at cryogenic temperatures (~77 K) generates mobile electrons from the
matrix that can reduce redox-active sites producing an EPR—visible reduced enzyme form
retaining the structure of the parent oxidized species. This technique has been applied to
a wide variety of metalloproteins, such as heme proteins (109, 110), diiron oxo proteins

(111—114), and iron—sulfur proteins (115-117).

54

 

IV.2 Experimental methods

IV.2.1 Synthesis and purification of CD3-AdoMet and l3CH3-AdoMet

As described in Chapter 111, Section 111.2.3.

IV .2.2 [1+ / AdoMet] sample preparation

The as-isolated PFL-AB, in 50 mM Hepes, 200 mM NaCl, 1 mM DTT, pH 7.5,
was exchanged with 50 mM Tris-HCl, 200 mM NaCl, 1 mM DTT, pH 8.5 buffer, and
concentrated in an Amicon pressure cell with an YM-lO membrane in the Coy Chamber.
The concentration of the protein was determined by the method of Bradford (77).

The following procedures were carried out anaerobically in an inert atmosphere
glove box (Mbraun) with 02 level 5 3 ppm. All buffers, solvents and bath water were
deoxygenated using a vacuum / inert gas manifold before being taken into the glove box.
Solid chemicals were pumped in as solids. Ice was pre-chilled with liquid nitrogen
before pumping into the glove box.

PFL-AE was mixed with 20-25 % (v/v) glycerol in an eppendorf to a total volume
of 200 [1L for EPR samples, or 300 ,uL for parallel EPR and ENDOR samples, and pre-
reduced by addition of 1 equivalent of sodium dithionite (freshly made 50 mM solution in
50 mM Tris buffer, pH 8.5). 5-deazariboflavin (10 mM stock solution in DMSO) was

added in the dark to a final concentration of 100 ,uM for 400 M or less concentrated

PFL-AB, or 200 M for PFL-AB of over 400 M, and mixed well. The sample was then

transferred to an EPR tube. The EPR tube was capped and inserted in a beaker tightly

55

packed with ice-water. The sample was illuminated on ice using a 500 W halogen lamp
for 1 hour with the lamp situated about 5 cm from the beaker. After illumination, the
protein sample was mixed with 2 equivalents of AdoMet on ice by adding the protein to
pre-measured AdoMet or isotopically labeled AdoMet solution (10 mM or 30 mM
solution in 50 mM Tris buffer, pH 8.5). The sample was transferred back to the EPR
tube, or split to EPR and ENDOR tubes, and flash-frozen in liquid nitrogen in the glove

box.

IV.2.3 [2+ / AdoMet] sample preparation

As-isolated PFL-AB containing 0.13 % of [3Fe-48]1+ (0.73 M spin for 560 MM
protein, as determined by EPR) was exchanged with 50 mM Tris-HCI, 200 mM NaCl, 1
mM DTT, pH 8.5 buffer, and concentrated as described in IV.2.2. In the inert
atmosphere glove box, 574 #M PFL-AE, 20 % (v/v) glycerol and 2 equivalents of CD3-
AdoMet or l3CH3-AdOMCI were mixed on ice to make a final volume of 60 #1.. The
sample was loaded into an ENDOR tube, and flash-frozen in liquid nitrogen in the glove

box.

IV.2.5 [2+/AdoMet]md sample preparation

[2+/AdoMet]md sample preparation was done in collaboration with Professor
Brian M. Hoffman, and Dr. Charles Walsby, Department of Chemistry, Northwestern
University.

The [2+ / AdoMet] sample prepared as described above was y-irradiated at 77 K

to produce the 1+ valency cluster trapped in the geometry of the 2+ state.

56

____.—.

 

 

 

IV .2.6 EPR spectroscopy

As described in Chapter II, Section 11.2. 10.

IV.2.7 ENDOR spectroscopy

ENDOR spectroscopy was done in collaboration with Professor Brian M.
Hoffman, and Dr. Charles Walsby, Department of Chemistry, Northwestern University.

35 GHz pulsed ENDOR spectra were recorded on a spectrometer equipped with a
helium immersion dewar and all the measurements were carried out at approximately 2
K. Pulsed ENDOR measurements employed the three-pulse Mims EN DOR sequence
(rt/2-t-1t/2-T-rt/2-‘t-echo) and its four-pulse ’Remims’ variant sequences, where the RF
was applied during the interval T. For a nucleus with hyperfine coupling A, the Mims
techniques have a response R which depends on the product, At, according to the
equation,

R ~ [1 - cos(21t At ] (Equation 0)

This function has zeroes, corresponding to rninima in the ENDOR response (hyperfine
“suppression holes”), at At = n; n = 0, 1, , and maxima at A1: = (2n+l)/2; n = 0, 1,...
The applied radio-frequency was broadened in bandwidth by mixing with a 60 KHz noise
to improve the signal-to-noise ratio.

For a nucleus (N), of spin 1 = 1/2 interacting with a paramagnetic center, the first
order ENDOR resonance condition for a single molecular orientation can be written as
follows:

V: = IVN i A/2I (Equation 1)

57

 

 

 

 

 

where vN is the Larmor frequency and A is the orientation dependent hyperfine constant

of the coupled nucleus. The equation predicts a doublet with separation A and centered at
VN when VN >A/2, as is true here for both 1H and 13C nuclei. Similarly, for a nucleus (N)
of spin 1:1, when we >A/2, the first order ENDOR resonance condition can be written as
follows:

vi = IVN :t A/2+P/2(2m1-1)l (Equation 2)
where P is the quadrupolar splitting term. This is the case with 2H nucleus in the present
study. The full hyperfine tensor of a coupled nucleus, both principal values and
orientation parameters (Euler angles) with respect to the g-tensor frame, is obtained by
simulating the 2-D pattern of orientation-dependent ENDOR spectra recorded across the
EPR envelope using the procedures and program described elsewhere. The computer
simulation and analysis of the frozen-solution ENDOR spectra used in the present work

are described in the same references.

IV.3 Results and Discussion
IV.3.1 Generation of the PFL-AB [4Fe-4S] 1* / isotopically labeled AdoMet complexes
by photoreduction
In order for ENDOR spectroscopy to clearly observe possible hyperfine
interactions between the isotopes and the [4Fe-4S]1+ cluster of PFL-AB, it is necessary to
generate large quantities of PFL-AB [4Fe-4S]1+ in the presence of isotopically labeled
AdoMet. Photoreduced 5-deazariboflavin acts as an electron donor, which conveniently
replaces the physiological flavodoxin system. Tris buffer is a suitable photo-substrate to

generate the effective 5-deazariboflavin radical. Efficient photoreduction of highly

58

..y‘ J—‘TIO .

 

 

concentrated PFL-AE samples was not as easily achieved as for fairly dilute protein

samples. Photoreduction conditions were optimized, including the amount of 5-
deazariboflavin and other additives present, the time period of illumination, and the
distance of the light source from the samples. The optimal pH for photoreduction was
found to be 8.5. Addition of 20-25 % of glycerol to concentrated PFL-AB samples and
pre-reduction with 1 equivalent of dithionite also afforded more efficient reduction. In
addition, the concentration of 5-deazariboflavin present affects the efficiency of
reduction, as a final concentration of 100 M deazariboflavin works better for 400 ,uM or
less concentrated PFL-AE, and 200 ,uM deazariboflavin for PFL-AE of over 400 M.
The illumination time was kept at 1 hour, and ice was continuously added to the ice-water
bath to keep the samples cool during illumination.

The X-band EPR spectra of the PFL-AB [4Fe-4S] H in the presence of CD3-
AdoMet and '3CH3—AdoMet are essentially indistinguishable from the [4Fe-4S] 1+/
AdoMet spectra. Photoreduced PFL-AB in the presence of CD3-AdoMet generated
38.0 % of [4Fe—4S]”, based on 324 ,uM spin for 910,uM protein with 3.75 mol Fe/ mol
PFL-AE (Figure IV.1). Photoreduced PFL-AB in the presence of ”CHg-AdoMet
generated 68.9 % of [4Fe—4S]”, based on 404 W spin for 777 ,uM protein with 3.02

mol Fe/ mol PFL-AE) (Figure IV.2).

59

 

 

 

 

A
W v '

 

EPR Intensity

 

 

 

 

I I

3000 3200 3400 3600 3800 4000
Field (Ga us)

 

 

 

Figure IV.1 X-band EPR spectrum of PFL-AE [4Fe-4S]1+ in the presence of

CD3-AdoMet. The sample contained 910 ,uM PFL-AE, 1 equivalent of Na2S204, 200 M
5-deazariboflavin, and 2 equivalents of CD3-AdoMet; 38.0 % reduction (based on 324

,uM spin for 910,11M protein with 3.75 mol Fe/ mol PFL-AB); g = 2.005, 1.949, 1.883.
Conditions of measurement: T = 12 K; microwave power, 20 ,uW; microwave
frequency, 9.47 GHz; modulation amplitude, 10.053 G; single scan.

 

 

EPR Intensity

 

 

 

3000 3200 3400 3600 3800 4000
Field (Gauss)

 

 

 

Figure IV.2 X-band EPR spectrum of PFL-AB [4F e-4S]1+ in the presence of
13'CH3-AdoMet. The sample contained 77 7 ,uM PFL-AB, 1 equivalent of Na28204,

200 pM 5-deazan‘boflavin, and 2 equivalents of 13CHg-AdoMet; 68.9 °/o reduction (based
on 404 ,uM spin for 777pM protein with 3.02 mol Fe/ mol PFL-AB); g = 2.011, 1.948,
1.880. Conditions of measurement: T = 12 K; microwave power, 20 ,uW; microwave
frequency, 9.49 GHz; modulation amplitude, 9.571 G; single scan.

60

 

 

IV .3.2 ENDOR spectroscopic evidence for a close interaction between AdoMet and the
[4Fe-4S] cluster of PFL-AE
2H and ”C ENDOR
The 2H ENDOR spectra of [1+/CD3—AdoMet] and [2+/CD3-AdoMet]md a=igure

IV.3), and the corresponding l3C spectrum of [1+/13CH3-AdoMet] and [2+/”CH3-
AdoMet],ed (Figure IV.4), immediately demonstrate that the cofactor sits close to the
cluster in both the 1+ and 2+ states. The 2H spectra of [1+/CD3—AdoMet] taken at g; and
g" (Figure IV.3) show a well-defined deuteron pattern that overlays a less-intense signal
seen in the unlabelled [1+/AdoMet] sample (see supplementary material); this latter is
assigned to double-quantum transitions from weakly-coupled 14N, as seen in similar Fe-S
clusters. In the 2H spectrum collected at g, the breadth of the signal is 1.1 MHz which,
when corrected for unresolved 2H quadrupole splitting of ~0.1 MHz (based on the known
value for quadrupole coupling for CD3) corresponds to a 1H coupling of ~ 67 MHz. For
comparison, this coupling is ~1/2 that of water bound to a low-spin heme and greater than
those to the protons of histidine bound to Fe. Such an interaction could not arise from an
AdoMet bound at a remote site and influencing the cluster EPR spectrum by an allosteric
process; at minimum it requires the AdoMet lie adjacent to the cluster. The
[2+/AdoMet]md state has an EPR signal identical to that of the [1+/AdoMet] state,
indicating that AdoMet binds with the same geometry to both the 1+ and 2+ clusters.
This inference is confirmed by the 2H EN DOR spectra of [2+/AdoMet]md as illustrated in

Figure IV.3: the spectra of [1+/AdoMet] and [2+/AdoMet],ed are indistinguishable.

61

 

 

 

l3C ENDOR spectra collected from [1+/13CH3-AdoMet] and [2+/”CH3-
AdoMet]md (Figure IV.4) lead to identical conclusions. The two labeled samples exhibit
identical hyperfine-split doublets centered at the 13C Larmor frequency and arising from
coupling to 13 C of the labeled AdoMet lying adjacent to the cluster. The spectra also
show ENDOR signals from natural abundance 57Fe. These are assigned to the v.
transition of a single iron with a coupling of A(57Fe) = 26 MHz, which is similar to that
of the labile Fe site in aconitase ES (A(57Fe) = 29 MHz). The existence of this signal is
useful because it allows us to compare spectra from natural—abundance and l3C-enriched
samples, by scaling to the signal from the 57Fe. A further use of this ability to scale

spectra is described below.

Field dependence of 13C ENDOR

Details of the AdoMet binding have been obtained by analysis of the 2D patterns
of 35 GHz Mims-pulsed l3C ENDOR spectra from [1+/'3CH3-AdoMet] and 2H spectra
from [1+/CD3-AdoMet], each collected across the entire EPR envelope. Figure IV.5
presents the 2D set of Mims pulsed ‘3 C ENDOR spectra; at each field the spectrum is
normalized to the 2-pulse electron-spin echo (ESE) signal intensity. We note that with the
value, 1: = 600 ns employed in the Mims ENDOR pulse sequence, the n = 1 'Mims
suppression holes' fall sufficiently far from the '3 C Larmor frequency (holes at Av(l3 C) =
$0.833 MHz) as not to significantly distort the observed l3C pattern. However the n = 0
'hole' centered at the 13 C Larmor frequency (Av('3C) = 0 MHz) regardless of 1:, has a
major impact on the observed signals and as a result it is essential to incorporate Mims

suppression into the simulations. This procedure is not completely understood because

62

 

 

the ’simple’ application of the Mims suppression formula can be compromised by spin-

diffusion effects. However, by the use of the Mims formula in the simulation program
and by use of carefully normalized spectra, it is possible to analyze the experiments
persuasively. In particular, by reproducing the relative peak intensities from field to field
in the 2D pattern, as well as the frequencies, helps to characterize the dipolar part of the
hyperfine interaction; this is particularly important at fields and for orientations where the
hyperfine coupling approaches zero and may change sign. As shown in Figure 1V.5,
systematic efforts at simulations yield excellent modeling of the peak positions and
intensities, in this case generated with a 13 C hyperfine tensor with principal values,
A(‘3C) = {-0.6(1), +0.4(l.5), -0.5(1)] MHz (NOTE: The relative signs are fixed by
experiment. The absolute sign is not, and is chosen to make the dipolar interaction
positive. This is consistent with Equation 3 for K positive.) and Euler angles relative to
the g-tensor frame of 0 = 30, w = 30. We note that although A. and A3 do not appear to
differ, within error, the errors are correlated. To reproduce the curvature of the outer
edges of the 2D plot, A1 must be ~0.l MHz larger than A3, and thus A(13C) is constrained
to be non-axial.

The 13 C tensor can be decomposed into the sum of an isotropic part, aiso(l3 C) =
-O.23MHz, and two, mutually perpendicular, purely-dipolar tensors, using A(l3C) = {-06,
+0.4, -05] as an example, T(‘3C) = {-0.33, +0.66, 0.33] = [-Tc, 2 TC, -TC] MHz, and
(-)t(l3C) = (+)[-0.03, -0.03, +0.06] = [-t, -t, 2t] MHz. The former we assign to the direct
through-space dipolar interaction between the 13 C and spin of the cluster; the latter we

assign to the ’local’ interaction with the spin on the l3C itself whose presence is disclosed

63

 

 

 

 

.- 9: _""' _
§ '

£1- 6‘},

 

 

by the isotropic term. The hyperfine tensor, thus, can be decomposed into the ’non-local’

dipolar term and a Iocal’ term that arises from spin density on the 13C, A(13C )loc =
~[Iaiso('3C )11 + t(13C )]. The presence of spin-density at the methyl group of AdoMet, as
manifest in this local term, requires that AdoMet lie in contact with the cluster, weakly
interacting with it through an incipient bond/antibond.

The through-space coupling tensor T, contains information about the geometry of
AdoMet binding. The dipole interaction between nucleus j of AdoMet, here the 13 C of the
labeled methyl, to a cluster can be written as

H”, = 52 Tj-i,
where: 1

5}. ~ 8.5.3.5. (Equation 3)

4
nzgefiegjflnz r3 1] ~ 3 Kt

i=1 0 r",- n nj
Here K, is the spin-projection coefficient for Fe,- of the spin-coupled S = 1/2 cluster, the
Fej‘NUCj distance is r,-,-, and It,- is the dimensionless through-space electron-nuclear dipolar
matrix. Extensive calculations indicate that only the dipole coupling to the nearest iron
ion, k, need be considered to a first approximation, as indicated in Equation 3, and hence
the dipole coupling to nucleus j is characterized by the parameter, T,- EggjnKk/rjk3.

The modeling process and determination of r64, itself for the '3 C -methyl, requires
knowledge of K". Mossbauer spectroscopic studies of [l+(AdoMet)] do not yet provide
the K,- for the four Fe ions, and therefore we consider the known spin coefficients of the
1+ cluster of substrate-bound aconitase (ES), which also has a labile, non-cysteine cluster
ligand; these are IKI = 0.86, 1.57, 1.60 and 1.78 (118). When interpreting the 13C dipolar

interaction constant, Tc = 0.33 MHz, in this way, the point-dipole-calculated distance

between the methyl ‘3 C and Pet of the cluster then would take one of four approximate

64

 

 

distances, depending on which value of K is associated with Pet. The range of possible
values for n.1, given the uncertainties in the simulations and K then falls in the range 3.5 -

5.6 A.

Field dependence of 2H ENDOR

The 2H and '3 C ENDOR data has been tested for self-consistency by analysis of
the 2D set of 2H Mims pulsed ENDOR spectra collected for [1+/CD3-AdoMet]. Figure
IV.3 includes the spectra at both g1 and g"; the intervening spectra (not included) show

that the breadth of the pattern decreases monotonically from g; to g... The deuteron

spectrum is not highly resolved but the hyperfine splitting at g r is roughly twice that of
the splitting at g... For a CD3 moiety, the maximum 2H quadrupole splitting is only 0.1
MHz and a 2H spectrum as broad as that shown in Figure IV.3 is dominated by the
hyperfine interaction to the three deuterons (j = 1-3). The spectra are compatible with a
surprisingly simple model in which the AdoMet is bound alongside the cluster and the
outer features of the CD3 ENDOR response are governed by the through-space dipolar
interaction between the closest methyl deuteron and the spin density on a single Fe ion of
the cluster.

Modeling the outer features of the field-dependent pattern by the interaction of
one deuteron with a single iron Fen gives the dipole—coupling parameter TD 5 gflgflnKn
/rD_k3 = 0.60 MHz, along with the orientation of the Fek—D; vector relative to the g-tensor
frame. Contributions from the more remote deuterons were calculated under constraints
of the tetrahedral geometry around the methyl carbon and some simple assumptions as to

the orientation of the AdoMet. As shown by the calculated spectra in Figure IV.3, the

65

 

data can be well represented by the dipolar spin Hamiltonians of Equation 3. The point-

dipole-calculated distance between the closest methyl deuteron and Fe), of the cluster then
would take one of four approximate distances, depending on which value of K is
associated with Pet. These are: ~3.0, 3.6, 3.7 and 3.8 A. Included in the simulations of
the perpendicular and parallel spectra in Figure IV.3 are the predicted signals for the
other, more distant, deuterons of the labeled methyl group. The distance of these
deuterons is dependent on the orientation of the methyl group relative to the cluster and
on the value of K used in the calculation; for the simulations in Figure IV.3 a

representative value of K = 1.6 (rperD = 3.7 A) was selected.

The above analysis is not unique, however, because the behavior of the D1 pattern
cannot be modeled at small couplings due to uncertainties about the contributions from
the D2, D3 signals, as well as because of the n = 0 'Mims hole'. Thus, the data do not rule
out an alternate model in which an isotropic coupling makes a significant contribution to
the 2H observed pattern. Simulations of the 2H data with a significant isotropic
component are not unique, but a representative simulation that describes the outer
spectral features utilized a tensor, A = [0.3,0.3, l .0] MHz, corresponding to an isotropic
coupling a = 0.53 MHz and an anisotropic contribution of the form [-0.23,-0.23,0.47]
MHz. In this model, the lack of resolution in the 2H spectra again is attributed to the
presence of signals from the other two deuterons, which although chemically equivalent,
would have different isotropic couplings. The isotropic coupling likely varies as cos2 0,

where Bis the dihedral angle between the it orbital on S that requires spin from the

cluster and the C—D bonds. For example, if one 2H has a maximum coupling (6: 0), the

66

 

 

 

an a" -

.q

 

other two will have four-fold smaller values (6: i21r/3, c0326: 1A), thus contributing

intensity near the center of the spectrum as required by experiment.

67

 

 

 

 

 

 

 

 

.— 1 .va

'1'»

_)‘a

 

 

 

-F'; .q

. w“

 

 
    

"l1".‘l§?:rf

    
  

I
N;

‘-',‘ ,-rrtenrvsu.iqgrt0_u rut-utqwssflenfntf‘. 1.

 

 

 

Figure IV.3 35 GHz Mims pulsed-ENDOR spectra of PFL-AB with methyl-D3
AdoMet. a) and d) photoreduced sample, b) cryoreduced sample. The spectra at gt lave
been scaled to the height of the natural abundance ”Fe peaks visible to higher fiequemy
of the 13C signals. Conditions: T = 2 K, M = 34.8 GHz, MW pulse lengths = 80 n8,

1 = 456 ns, RF pulse length = 60 ps, repetition rate = 30 Hz. Each spectrum consists of
256 points with each point an average of 240-300 transients. c) and e) Simulations with
axial, completely anisotropic A tensors. Closest 2H: T = 0.6 MHz (corresponding R - 3.1
AforK= 1.0), ot=p= 30°, y=0°', 3P=0.l MHz ct= B=y=0°. Moredistantmethyl
deuterons: R = 4.7 A, K = 1.0 (corresponding T= 0.12 MHz) angles tetrahedtuly

disposed, quadrupole as above.

 

 

 

 

 

 

 

fiesta:
Si? :5.

 

.
J-"

0.5

 

0
v-v(2H) (MHz)

I

-0.5

 

 

rte: 1

'7‘

AJ

(1) photoreduced PFL-AB [4Fe-48]1+/ CD3-AdoMet at g“

a) photoreduced PFL-AB [4Fe-4S]”/ CD3-AdoMet at gi
e) simulations of c)

b) cryoreduced PFL-AB [4Fe-48]2+/ CD3-AdoMet at g

c) simulations of a) and b)

Figure IV.3

69

 

 

 

 

   

 

13C

 

A
57Fe
A
a)
b)
C)
d)
-1 0 l
v - v('3C)

Figure [VA 35 GHz Mims pulsed-ENDOR spectra of PFL-AE with
methyl-”C AdoMet.

a) photoreduced PFL-AE [4Fe-4S]'+/ '3CH3-AdoMet at gl
b) cryoirradiated PFL-AB [4Fe-48]”/ 13CH3-AdoMet at g..
c) photoreduced PFL-AB [4Fe-48]”/ AdoMet (natural abundance '3 C) scaled to 57Fe
d) photoreduced PFL-AE Hrs-43117 '3CH3-AdoMet at g.l

Experimental conditions as for Figure IV.3, except that T = 600 nS and number of
transients = 600.

70

 

 

 

 

 

 

 

 

 

 

 

 

I I I I I

-l O 1
v - v(l3C) (MHz)

Figure IV.5 l3C field dependence with simulation of line positions and relative
intensities. Field dependence data (Full line) conditions as for Figure IV .4. Simulation
(dashed line) parameters: A = [-0.6(i0.1), 0.4(i0. 15), -0.5(i0.1)] MHz, or: B = 30°, 7':
0°; EPR linewidth = 170 MHz, ENDOR linewidth = 0.2 MHz; T = 600 ns, 100% Mims
suppressron.

71

 

 

 

 

 

Summary

The observation of substantial 2H and 13C hyperfine couplings from labeled
AdoMet bound to the 1+ cluster of PFL-AE clearly demonstrates that AdoMet binds
adjacent to the [4Fe-48] cluster. The observation of identical spectra from the 1+ and 2+
cryo-reduced enzyme further shows that the position of AdoMet relative to the Fe-S
cluster in PFL-AB is identical in the oxidized and reduced states.

Most intriguingly, the existence of an isotropic contribution to the 13C tensor
requires that there be overlap between orbitals on the cluster and on AdoMet, namely that
there are incipient covalent interactions between AdoMet and the cluster. The most
plausible interpretation is that this interaction is of a dative character, between a
negatively-charged sulfide of the cluster and the positively-charged sulfur of AdoMet.

As shown in Table IV.l, the values of r derived from the dipolar interpretation of
the 2H data produce distances to the closest iron ion which are consistent with the C-H
single bond separation, with the methyl protons closer than the methyl carbon as required
sterically. Any interpretation of the 2H 2D pattern which includes a significant isotropic
component produces a smaller T and thus predicts a larger value of r(2H-Fe,,). Given that
the 2D pattern always demands a significant dipolar interaction, as evidenced by the 2-
fold increase in pattern width between g.. and g1 and that r(2H-Fe,,) must not be large
enough to place the closest deuterons on the 'other' side of the methyl carbon. We are
able to select the purely-dipole model as the better approximation. The distances
predicted from T of the model with substantial isotropic 2H coupling are close to those for
r('3C-Fek); for example from the fitting quoted above for the covalent model, choosing

the K values in descending order produces distances of 4.5, 4.4, 4.3 and 3.5 A all of

72

 

 

 

"f 96“". ‘:0-_'> - ._ - r . .

 

 

 

 

 

which are within the range of the distances quoted for r(13 C-Fek). We do not, however,

believe that the data rules out all isotropic contributions.

 

 

 

 

 

K r(2H-Fe,,) - 10% uncertainty in T r('3C-Fe,,) - range in Tfrom simulation
1.78 3.8 30.1 5010.6
1.60 3.7 x 0.1 4.9 :t: 0.6
1.57 3.6 1 0.1 4.9 :1: 0.6
0.86 3.0 t 0.1 4.0 i 0.6

 

 

 

 

 

Table IV.1 r values derived from the dipolar interpretation of the 2H ENDOR data

73

 

 

 

 

The result of these investigations is a structural model of AdoMet bound to both
1+ and 2+ states of PFL-AB in which a cluster sulfide and the AdoMet sulfur are
adjacent, with weak dative orbital overlap, while the AdoMet methyl-group carbon lies

~4—5 A from an Fe ion. This model is illustrated by the structure presented in Figure

 

 

 

 

IV.6.
NH2 H H3N COO
N
N/ l >
0 513-0? "'
H
S Fe
OH OH |\Fe |\S
s/ I
|
Fe— --—8
/ \Q \
S a re\
8
[4Fe-48]

Figure IV.6 Structural model of AdoMet bound to the [4Fe-4S] cluster of PFL—AE

74

 

 

CHAPTER V

Se-ADENOSYL-L—SELENOMETHIONINE AND
ITS INTERACTION WITH
PYRUVATE FORMATE-LYASE ACTIVATING ENZYME

V.1 Introduction

Selenium is in a bridging position in Group VIA, lying between two nonmetals,
oxygen and sulfur, and the increasingly metallic tellurium and polonium. In its period it
lies between the metalloid arsenic and the halogen bromine. Selenium has six naturally
occurring stable isotopes, 74Se (0.87%), 76Se (9.02%), 77Se (7.58%), 78Se (23.52%),
80Se (49.82%), 82Se (9.19%) and several radioactive isotopes, the most important of
which is 75 Se, a yemitter(t1/2 = 120.4 days), which is widely used as a tracer in
biological studies and in certain radiologic diagnostic procedures. Of the naturally
occurring isotopes only 77Se has a nuclear spin (1 = 1/2) and is an increasingly useful
probe in some spectroscopic studies (119). Selenium is widely distributed in nature (120-
122). It ranks seventieth in order of abundance of the elements and comprises
approximately 7 x 10'5 weight per cent of the earth’s crust. Selenium has various
applications in industry, such as in electronics and metallergy; in the glass and ceramics
industry; in rubber, pigment, and explosive manufacture, and as a catalyst or a constituent
in some pharmaceuticals (119).

In terms of chemical properties, selenium resembles sulfur more closely than
any other element. Both of them occur in the same valence states, -2, 0, +2, +4, and +6,
and have similar covalent radii and eletronegativities (123). However, the two elements

display noteworthy differences in biological systems. Selenium is a paradoxical element

75

 

 

 

 

in that it is highly toxic and yet is an essential micronutrient for mammals, birds, several
bacteria and probably for fish and many other animal species. “Alkali disease” and
“blind stagger”, diseases of livestock known for some time, are now attributed to
selenium poisoning, resulting from the high toxicity of “selenium accumulator plants” to
most animals (121). On the other hand, selenium has been demonstrated to be essential
in view of the numerous specific and well-characterized occurrences of selenium in
biomolecules (124, 125), and a number of selenium deficiency diseases found in animals
as well as human beings (121, 126). Despite its toxicity, selenium compounds have been
used as complexing agents to treat heavy metal poisoning, arising from such metals as
silver, cadmium, mercury and lead (121, 126). The detoxification is thought to occur
because selenium, having a greater affinity than sulfur for heavy metals, can strip the
heavy metals from their in vivo ligands. In addition, the difference in retention and
distribution of sulfur and selenium compounds in the body may facilitate detoxification.
Selenium in plants and animals is found primarily in the form of selenoamino

acids (119). Shown in Scheme V.1 are the major selenoamino acids that have been
isolated from plants and chemically synthesized. Selenium is also known to be a normal
component of several enzymes, proteins and some arninoacyl transfer nucleic acids
(tRNAs) (119). Interestingly, most of the selenoamino acids detected in animal and
bacterial proteins are selenocysteine, with a number of other selenoamino acids reported
(127), including selenomethionine (125). Forrnate dehydrogenase from E. coli, glycine
reductase from amino acid-fermenting clostridia, hydrogenases from Methanococcus
vannielii and from Desulfovibrio baculatus, and mammalian glutathione peroxidase are

some examples of the enzymes that have been demonstrated to contain selenium in the

76

 

_:

49.9””.

r

 

 

 

 

 

 

 

 

 

 

 

 

 

int
Se—CHz—CH-COOH NH2
HSe—CHZ—CH—COOH

Se—CHZ—(lIH—COOH
NH2

Selenocystine SCICHOCYStCine

N112 NH2
HOOC—HC-CHf—Se—CHZ— CHZ—CH—COOH
Selenocystathionine

NH2

Se—CHz—CHZ—CH-COOH NHZ

HSe—CHz—CHz—CH—COOH
Se—CHz—CHz—CIZH—COOH
NH2

Selenohomocystine Selenohomocysteine
NHZ
H3c—Se—CH2— CHz—CH-COOH

Selenomethionine

Scheme v.1 Selenoamino acids found in plants

form of selenocysteine (125). Four genes, whose products are required for the incorpora-
tion of selenocysteine into proteins, have been identified in E. coli. These findings truly
establish selenocysteine as the twenty-first naturally occurring amino acid (125).
Following the rationale that selenium substitutes for sulfur in particular functions
in various biomolecules, selenium has been artificially incorporated into specific sites of

proteins in order to modify their properties and ultimately to understand the targeted sites

77

 

1
1

 

 

 

of the proteins. Selenocoenzyme A and selenobiotin have been synthesized and their
activities have been studied to compare with their native counterparts (128-130). Many
modifications have also been targeted at cysteine or methionine residues (131-137) and
inorganic sulfur atoms of iron-sulfur proteins. Substitution of selenium in the active site
of ferredoxins has led to interesting findings (Reviewed in Ref.138).

There are also a number of enzymes, which under normal conditions catalyze
reactions that result in the transformation of a particular sulfur compound, which have
also been found to undergo the same reaction with the analogous selenium compound
(139- 141). Selenium-containing biomolecules can thus frequently be prepared in vitro by
enzymatic methods. Some notable enzymes that catalyze reactions in which selenium-
containing substrates have been shown to react in place of the corresponding sulfur-
containing substrates are tRNA sulfur transferase and cysteinyl-tRNA synthetase
(selenocysteine) (140, 142, 143), ATP sulfurylase (adenosine-5’-selenophosphate) (144),
AdoMet synthetase (selenomethionine) (100), and AdoMet-dependent methylases and
AdoMet decarboxylases (Se-adenosylmethionine) (100, 131, 145, 146).

Selenium compounds are usually more reactive than their sulfur analogues
presumably because of the slightly greater polarity and lower bond strength of the C-Se
O'bond as well as other 0' bonds such as the N-Se and O-Se bonds (119). Selenides are
also more nucleophilic than analogous sulfur molecules (102, 103). For example, seleno-
methionine has been shown to be more reactive than methionine (147), and is much less
stable to acid hydrolysis. Selenomethionine completely reacts with cyanogens bromide
in 0.1 M HCl to produce homoserine in 15 min while methionine requires 24 h for the

same reaction. Selenomethionine also reacts more rapidly than methionine in certain

78

 

7‘

 

 

   

synthesis reactions, as has been reported for AdoMet synthetase from both yeast and E.
coli (93, 100, 101).

As discussed previously, the chemical properties of selenium are sufficiently
similar to those of sulfur to allow replacement of the latter by the former in many
biomolecules. However, the two elements differ significantly from each other by some
of their physical characteristics, and selenium has proved to be a more efficient probe
than sulfur for several spectroscopic methods (119).

The two chalcogens differ considerably in their atomic mass number A, Se (A =
79) being more than twice as heavy as S (A = 32). The atomic mass of Se qualifies it as
an absorbing atom for X-ray absorption spectroscopy (XAS), which sometimes is diffcult
to perform with sulfur. For example, S-XAS is not a good method for detecting changes
in ligation of an Fe-S cluster, because of the large number of Fe-S bonds present in a
typical Fe-S cluster, and since XAS “sees” an average coordination environment for all
molecules of a given element in the sample. Measurements on Fe-Se synthetic
compounds have demonstrated the feasibility of Se-XAS (148), and it has been used to
show that the coordination sphere of Se in a hydrogenase from D. baculatus includes
nickel and carbon (149). Recently, Frey et al. have successfully demonstrated a direct
interaction of selenomethionine with the Fe-S cluster of lysine-2,3-aminomutase, a
member of the AdoMet dependent Fe-S enzymes, by using Se K-edge XAS in
combination with selenium derivatives of methionine and AdoMet (150).

Another advantage of selenium is the presence of an isotope (77Se) with nuclear
spin 1 = 1/2. When Se atoms are involved in or near a paramagnetic center, the presence

of 77Se may result in EPR spectra displaying hyperfine structures that can provide

79

 

 

“I

 

 

 

detailed information on the environment of the paramagnetic center. This property has
been implemented for several Fe-S proteins, where Se was artificially incorporated into
the Fe-S clusters or the bacteria which overexpress the proteins were grown in selenium
enriched media (138, 151). Moreover, electron nuclear double resonance (ENDOR) and
electron spin echo envelope modulation (ESEEM) also have potential in revealing
hyperfine interactions of 77Se with a paramagnetic center. ENDOR and ESEEM are two
related but complementary techniques that derive from the basic EPR experiment.
Broadly speaking, ENDOR is used for investigation of strongly coupled “closer” nuclei
having coupling constant A typically in the range of 2 — 40 MHz, while ESEEM is used
to study “ more distant” weakly coupled nuclei, having A less than about 10 MHz,
depending on the nucleus (152). In addition, infrared and resonance Raman
spectroscopy, circular dichroism, X-ray diffraction and 77Se NMR also have applications
in studies of Se-containing biomolecules (119).

In view of the wide existence of Se in biological systems, its bio-compatibility to
several enzymes that catalyze reactions involving sulfur compounds, and its application
in certain spectroscopic methods, Se—AdoMet has been proposed to be an effective probe
in studying the interaction of the sulfonium center of AdoMet with the Fe-S cluster of
PFL-AB. Se-AdoMet, the substrate analog, and selenomethionine (Se-Met), the cleaved
product analog, along with the [4Fe-4S] clusters of PFL-AB at different oxidation states
were used to mimic the enzyme-substrate (ES) and enzyme-product (EP) complexes, and
to follow the course of the cleavage reaction by a variety of spectrosc0pic methods,

including EPR, XAS, ENDOR and ESEEM.

80

 

 

 

 

 

 

V .2 Experimental methods

V.2.1 Synthesis and purification of Se-AdoMet

As described in Chapter 111, Section 111.2.3.

V.2.2 Activity assay of PFL-AB with Se-AdoMet

As described in Chapter H, Section 11.2.9.

The PFL-AB reaction mix contained in a final volume of 1 mL: 0.1 M Tris-HCl,
pH 7.6, 0.1 M KCl, 10 mM DTT, 10 mM oxamate, 9.84 mg PFL, 200 W deaza-

riboflavin, 0.2 mM Se-AdoMet, and 0.186 M PFL-AB.

V.2.3 [1+ / Se-AdoMet] sample preparation

EPR, ENDOR and XAS samples were prepared as described in Chapter IV,
Section IV.2.2, except using Se-AdoMet as substrate. [1+/Se-AdoMet] samples for EPR
and ENDOR spectroscopy contained 2 equivalents (relative to protein) of Se-AdoMet.

[1+/Se-AdoMet] samples for Se K-edge XAS contained a final volume of 160 ,uL:
typically approximately 1 mM PFL-AB, 25 % (v/v) glycerol, 1 equivalent of sodium
dithionite, 200 #M deaza-riboflavin and no more than 0.95 equivalents (relative to
protein) of Se-AdoMet. A high content of glycerol present in the samples was to prevent
formation of ice crystallites, which could result in failure of data collection. A slight
excess of protein (relative to seleno-compound) was present in the samples to insure that
all metal in the sample was protein-bound, since XAS “sees” an average structure of all
selenium sites, including any excess selenium that is coordinated to solvent. The XAS

samples were either prepared in parallel with EPR samples, with aliquots transferred to

81

 

 

 

< 11.1111;

.-1 “qu

tailli‘fi’ 7‘
. risen}.
4, . .

mum 112:?

. -
. §
. __ .:__
I"!HG5[§‘¢"'
. ' r F. a
-1?! P".
' .

    

 

 

both an EPR tube and an XAS cuvet, or the samples were prepared from a previous EPR

sample that was thawed and transferred to an XAS cuvet.

V.2.4 [2+ / Se—AdoMet] XAS sample preparation
As described in Chapter IV, Section IV .2.3.
The [2+/Se-AdoMet] mix contained in a final volume of 160 #1.: typically

approximately 1 mM PFL-AE, 25 % glycerol and 0.95 equivalents of Se-AdoMet.

V.2.5 [2+ / Se-Met / 5’-dAdo] XAS sample preparation

The [2+ / Se-Met / 5’-dAdo] mix contained in a final volume of 160 ,1th typically
approximately 1 mM PFL-AB, 25 % glycerol, 0.95 equivalents of Se-Met, and
4 mM 5’-deoxyadenosine. The sample was mixed well, loaded into an XAS cuvet, and

flash-frozen in liquid nitrogen in the glove box.

V.2.6 [2+ / Se-Met / 5’-dAdo / PFL] XAS sample preparation

The [2+ / Se-Met / 5’-dAdo / PFL] mix contained in a final volume of 160 uL:
typically approximately 0.4 mM PFL-AB, 25 % glycerol, 0.95 equivalents of Se-Met,
4 mM 5’-deoxyadenosine, and 1.1 equivalents of PFL. The sample was mixed well,

loaded into an XAS cuvet, and flash-frozen in liquid nitrogen in the glove box.

V.2.7 [2+ / Se-AdoMet / PFL] sample preparation
[2+ / Se-AdoMet / PFL] samples were prepared under single turnover conditions

(16). The “PFL mix” contained approximately 0.9 mM PFL, 25 % glycerol, 250 mM

82

 

ro'JY'pzl

I . -
LI. u +-

5. 11314.31

   

 

 

dithionite, 10 mM DTT, and 200 [AM 5-deazariboflavin, and was pre-reduced by
illumination on ice using a 500 W halogen lamp for 30 min. The “PFL—AB mix” was
prepared as described in V.2.3, and an aliquot of the mix was saved to determine the
amount of [4Fe-4S]1+ by EPR spectroscopy. The “PFL mix” was then mixed with
equimolar “PFL-AB mix”. The final protein concentration was thus diluted by a factor of
2, and was approximately 0.4 mM. The sample was split into an EPR tube and an XAS
cuvet, and flash-frozen in liquid nitrogen in the glove box. The amount of glycyl radical

generated was determined by EPR spectroscopy.

V.2.8 EPR spectroscopy

As described in Chapter H, Section 11.2. 10.

V29 ENDOR spectroscopy
ENDOR spectroscopy was done in collaboration with Professor Brian M.

Hoffman, and Dr. Charles Walsby, Department of Chemistry, Northwestern University.

V2. 10 Se K-edge X-ray absorption spectroscopy
XAS was done by collaboration with Professor Robert R. Scott, and Nathaniel
Cosper, Department of Chemistry, University of Georgia. XAS data were collected at

Stanford Synchrotron Radiation Laboratory (SSRL), beamline 7-3.

83

 

    
 
 
 
 
 
 
  
  
  

1110M
izuuum
‘. . '90-. it]

'1) mm

.v‘.Ul1)
f
. .-

. r4201.)

:11' 11.111118

 

V3 Results and Discussion
V31 PFL-AB activity with Se—AdoMet
The activity of PFL-AB in the presence of Se—AdoMet was assayed using a direct

enzyme assay in which the amount of glycyl radical on PFL generated by PFL-AE as a
function of time was quantified by EPR spectroscopy (Figure V. 1). The specific activity
of PFL-AB in the presence of Se-AdoMet was estimated to be at least 300 U/mg. This
activity is only a lower limit because the time points taken were beyond the linear region
of activity. The specific activity determined for PFL-AB with Se-AdoMet is

approximately 3 times higher than that with regular AdoMet, which was found to be

95 U/mg.

 

 

25

20~

15~

uM 91y.

10~

 

 

o I I I

0 10 20 30 40
Time (min)

 

 

 

 

Figure V.l Activity assay of PFL-AB in the presence of Se—AdoMet. The assay mix
(1 mL) contained 9.84 mg/ mL (57.9 mM) PFL, 0.1 M Tris-HCI, pH 7.6, 0.1 M KCl, 10
mM DTT, 10 mM oxamate, 0.2 mM 5-deazariboflavin, 0.2 mM Se-AdoMet, and 0.186
,uM PFL-AB (3.8 mol Fe/ mol PFL-AB, 4.0 mol S/ mol PFL-AB). The assay mix was
illuminated for 5, 10, 20, 30, and 40 min, and the amount of glycyl radical generated was

determined by X-band EPR spectroscopy.

84

 

___..

..4. "5* .‘m V-

 

 

V .3 .2. Generation of [4Fe-4S] H of PFL-AB in the presence of Se-AdoMet

Deazariboflavin-mediated photoreduction of PFL-AE followed by addition of Se-
AdoMet afforded a high percentage reduction of the [4Fe-4S] cluster of PFL-AB. In
initial experiments, precipitation of protein upon addition of Se-AdoMet solution was
observed, which was thought to result from residue acid from the Se-AdoMet
purification. The problem was solved by dissolving Se-AdoMet in 50 mM Tris buffer,
pH 8.5.

Photoreduction of PFL-AE followed by addition of Se-AdoMet generated a [4Fe-
4S]” EPR signal accounting for 77.5 % of the iron, based on 512 W spin for 775 #M
protein with 3.41 mol Fe/ mol PFL-AE. A [1+/AdoMet] sample prepared at the same
time generated 63.0 % [4Fe-4S]'+, based on 416 W spin for 775 M protein with 3.41
mol Fe/ mol PFL-AB. (Figure V2). The EPR spectrum of [1+/ Se-AdoMet] was similar
to that of [1+/ AdoMet], suggesting that the complexes formed with AdoMet and Se-
AdoMet were similar. However, the Se-AdoMet signal was considerably broader than
that of the AdoMet. In order to investigate whether the effects of Se-AdoMet on the EPR
signal were due to hyperfine interactions of the 7.5 % abundant 77Se, 77Se-ESEEM and
77Se-ENDOR experiments have been carried out with the [1+/ Se-AdoMet] samples. To
date, however, 77Se-ESEEM and 77Se-ENDOR spectroscopy have not shown evidence for

hyperfine interactions between 77Se and the Fe-S cluster of PFL-AB.

85

 

 

 

 

 

 

 

 

 

EPR Intensity
w

 

 

 

I I 1 I

3000 3200 3400 3600 3800 4000
Field (Ge use)

 

 

 

Figure V.2 X—band EPR spectra of photoreduced PFL-AB (A) in the presence of
Se-AdoMet; the sample contained 775 ,uM PFL-AB, 1 equivalent of Na28204, 200 M
5-deazariboflavin, and 2 equivalents of Se-AdoMet; 77.5 % reduction (based on 512 ,uM
spin for 77 5 M protein with 3.41 mol Fe/ mol PFL-AE); and (B) in the presence of
AdoMet; the sample contained 775 ,uM PFL-AB, 1 equivalent of Na28204, 200 ,uM
S-deazariboflavin, and 2 equivalents of AdoMet; 63.0 % reduction (based on 416 [AM
spin for 7 75 M protein with 3.41 mol Fe/ mol PFL-AB). Conditions of measurement:

T = 12 K; microwave power, 20 ,uW; microwave frequency, 9.48 GHz; modulation

amplitude, 9.57 G; single scan.

86

 

~73." M'J

 

 

 

V .3 .3 XAS investigation of Se-Fe interaction between the [4Fe-4S] cluster of PFL-AB

and Se-AdoMet

Se K-edge XAS spectra of Se-Met and Se-AdoMet displays the expected change
in edge position and a significant change in edge shape (Figure V3). The shift in
absorption edge position between Se-Met (12659.6 eV) and Se-AdoMet (12661.2 eV) is
indicative of a change in the Se oxidation state of the sample (153). Fourier transform
(FT) spectra are characterized by the reduction in FT peak intensity for Se-Met, which is
consistent with the coordination number for Se-Met being lower than that for Se-
AdoMet. The above features make Se edges a diagnostic fingerprint for distinguishing
between samples that resemble these two compounds.

Moreover, a Se-Fe interaction was observed in lysine-2,3-aminomutase after the
enzyme turned over (150). The Se-Fe interaction was characterized by a distinct peak at

ca. 2.7 A in the FT spectra. This result provides a precedence of observation of Se-Fe

interaction in Fe-S enzymes by XAS.

87

 

 

 

 

 

2.5 , .

 

 

A —— ESAMA. Se- SAM (04/99)
— EMETA Se MET (04/99)

,3 2.0 — /\ -
tn
5, .
g 1.5 — -
E 1.6 —
E
o 0.5 — ..
Z

 

 

0.0 M

" l l 1 i l l

1 2650 1 2660 1 2670 1 2680
Energy (eV)

 

 

 

1 2 _ 1 l 1 I l 1 __
' B ———- ESAMA. Se-SAM (04/99)
-—-- EMETA. Se-MET (04/99)

 

 

 

 

R' (A)

Figure V.3 Se K-edge X—ray absorption spectra (A) and Fourier transforms (B) of

Se-AdoMet (ESAMA) and Se-Met (EMETA).

88

 

 
 

 

1 ll. 1'1 il I i {fill-t...t...-. m1 i . I

“ I
I ' I
I
I

A series of enzyme-substrate (ES) and enzyme-product (EP) complexes were
prepared with Se-AdoMet, Se-Met, and the [4Fe-4S] clusters of PFL-AB in different
oxidation states, and the course of the AdoMet cleavage reaction was followed by Se K-
edge XAS.

Incubating [4Fe-4S]2+ of PFL-AB with Se-Met, Se-Met / 5’-dAdo, and Se-Met/
5’-dAdo / PFL yielded Se-edge and FT spectra that were similar to that of Se-Met alone.
The spectrum of [4Fe-4$]2+ of PFL-AB with Se-AdoMet was similar to that of Se-
AdoMet alone (Figure V.4). The observation that the selenium compounds maintained
their initial structures was expected, since the [4Fe-4S]2+ cluster is not the catalytically
relevant cluster, and therefore should not react with AdoMet. In addition, no Se-Fe
interaction was observed in the FT spectra of any of the above samples, unlike what was
observed in LAM incubated with Se-Met, 5’-dAdo and its substrates (150). The [2+/
Se-AdoMet / PFL] sample was prepared under single turnover conditions. The “PFL-AB

mix” was photoreduced to generate 319 fllVI [4Fe-4S]1+ (out of 927 M protein) after

addition of Se-AdoMet, and 11.2 M glycyl radical was generated upon mixing
equimolar of the pre-reduced “PFL-AB mix” and the “PFL mix”. The glycyl radical
accounted for 2.8 % of PFL, as the final protein concentration was 395 pM. Although
little glycyl radical was produced, it seems that the PFL-AE turned over the Se-AdoMet,
since the XAS edge of the [2+ / Se-AdoMet / PFL] sample was more reminiscent of that
of Se-Met (Figure V.4). However, the edge of this species was slightly more intense and
shifted to higher energy relative to that of Se-Met, indicating that some Se-AdoMet
remained. The low percentage of enzymatic turnover could be attributed to reaction of

the PFL-AB [4Fe-4S] 1+ with Se-AdoMet prior to mixing with PFL. No Se-Fe

89

 

‘—"“

 

 

 

12650

2.5
E‘ 2.0
(D
5
g 1.5
1::
a 1.0
‘s"
3 0.5
2
00
1.0
0.8
o
'0
,2- 0.6
r:
8
2 0.4
1:
0.2
0.0
0
FigureV.4

 

interaction was observed in this sample, indicating that Se-Met does not coordinate to the

cluster afier turnover, in contrast to the results for LAM (150).

 

 

      

,1" — PFL-AE [4Fe-4S]2+ISeMoMet/PFL. EPOQA (0112001)

1‘ If"
I

—— PFL-AE [4Fe4S]2+ISeMet/dAdo. EPOSA (01/2001) _
— PFL-AE [4Fe—4$]2+lSeMet/PFUdAdo, EPRPA (0112001)
—-—— PFL-AE [4Fe-4S]2+ISeSAM, EPOEB (05/2001)
----- PFL-AE [4Fe-4$]2+ISeMet/PFUdAdo. EPRPB (05/2001)

1 . 1 I -

L

 

 

12660

12670 12680
Energy (eV)

 

 

 

I A l I I
.... PFL-AE [4Fe-48124-ISeAdoMeUPFL. EPOQA (0112001)
...... PFL-AE [4FMS]2+IScMetfdMO. EPOSA (01/2001)
__ PFL-AE [4Fe-48]2+ISeMeUPFUdAdO. EPRPA (0112001)-

 

 

 

 

Se K-edge X-ray absorption spectra (A) and Fourier transforms (B) of

PFL-AB [41:e-4sl2+ incubated with Se—Met / 5’-dAdo (EPOSA), Se-Met / 5’-dAdo / PFL

(EPRPA and EPRPB), Se-AdoMet (EPOEB), and Se—AdoMet / PFL (EPOQA).

90

 

   

   

 

Interestingly, the edge for the three samples of PFL-AB [4Fe-4S]l+ in the presence

of Se-AdoMet but in the absence of PFL indicated varying levels of turnover

 

 

 

 

 

 

 

 

 

(Figure V.5).
2-5 l r I 1 r
A
«3‘ 2.0 - -
c
a: ,r
E 1.5 - )- \. a“ _.
_ / . “x.
‘o / 7"“ _.
.5? 1.0 e x xi)
m
E 0.5 1— .;z — PFL-AE [4Fe-48]1+ISeSAM. EPREA(O512001)-
g .' —— PFL-AE [4Fe-4S]1+ISeSAM. EPREB (0512001)
_ , - — PFL-AE[4Fe—4S]1+ISeSAM. EPREC(0512001)
0.0 ... "" 1 l J 1 1 4
12650 12655 12660 12665 12670 12675 12680
Energy (eV)
I I I V l r
1.2 " B —— PFL-AE[4Fe-48]1+ISeSAM, EPREA(05/2001) "
—— PFL-AE [4Fe-48]1+lSeSAM, EPREB (05/2001)
1.0 — — PFL-AE [4Fe-48]1+/SeSAM, EPREC (0512001) -
3
,3 0.8 - -
c:
g 0.6 h i -
2
E 0.4 - A -
”'2 " « '   & sh... "
OOfiA 4‘ I W‘ ' I ' .. *‘vm
0 1 2 3 4 5 6 7

R'(A)

Figure V.5 Se K-edge X-ray absorption spectra (A) and Fourier transforms (B) of
PFL-AB [4Fe.43]“ in the presence of Se—AdoMet. Sample EPREA: 910 M PFL-AB,
590 pM Se-AdoMet, and 538 M [41=e-4s11+ generated; EPREB: 600 pM PFL-AB, 575
pM Se-AdoMet; EPREC: 1098 M PFL-AE, 633 uM Se—AdoMet, and 214 M [4Fe-
4S]1+ generated.

91

 

 

 

 

. I‘
, _ ‘I I

The difference in the amounts of Se-AdoMet being cleaved in the three samples
may result from the fast degradation of Se-AdoMet that took place between the time of
sample preparation and freezing, which varied from one preparation to another due to the
difficulties in loading the XAS cuvettes. Regardless of the amount of Se—AdoMet being
cleaved in each sample, the observation of the cleavage of Se-AdoMet in the absence of
PFL suggests that PFL might not be required to initiate formation of the adenosyl radical.
Thus, the cleavage of Se—AdoMet may be less strictly controlled than in LAM, where no
cleavage of Se-AdoMet was observed unless the substrate analog trans-4,5—dehydrolysine
was present (150).

None of the PFL-AB samples in the presence of Se-AdoMet displayed a Se-Fe
interaction in Se K-edge XAS, in contrast to the observations in LAM. LAM is distinct
in the class of AdoMet-dependent Fe-S enzymes in that the cleavage of AdoMet is freely
reversible. Once the adenosyl radical is generated in LAM, it becomes catalytic, and the
enzyme can complete multiple turnovers with only one equivalent of AdoMet. Frey et al.
postulated that the interaction of methionine with the Fe-S cluster might aid not only in
cleaving the cofactor, but also in maintaining methionine in the active site of the enzyme
for the back reaction to regenerate AdoMet, which in turn makes it easier to trap the
“enzyme-product” complex and observe the Se—Fe interaction (150). However, in PFL-
AE, adenosyl radical generation is required for each turnover, which means one mole of
AdoMet is expended per mole of glycyl radical. Thus, the AdoMet binding pocket in
PFL-AB would have to be emptied after the adenosyl radical is generated, requiring that
any Se-Met and Fe-S cluster interaction be short lived. In addition to the fundamental

nature of the radical generation in PFL-AE, simply the heterogeneity of the species

92

 

 

 

resulting from the different levels of Se-AdoMet cleavage makes it beyond the capability

of the current XAS experiments to observe any putative Se-Fe interaction.

93

 

 

 

 

 

 

 

 

 

 

 

 

CHAPTER VI

CONCLUSIONS

Since the first isolation of PFL-AE in its native state under strictly anaerobic
conditions and identification of the presence of an iron—sulfur cluster in PFL-AB (13, 14),
PFL—AE growth and purification conditions have been modified, and the yield of purified
protein, as well as the content of intact Fe—S clusters in the as-isolated protein, have been
significantly improved. Under current growth conditions, approximately 400 mg of
purified PFL-AE was obtained from a 9 L bacteria culture; this represents a 2-3-fold
improvement over earlier procedures. In the purification procedures, two passages
through the gel filtration column under strictly anaerobic conditions in the presence of 1
mM DTT afforded a higher content of intact Fe-S clusters in the as-isolated PFL-AE.
The as-isolated PFL-AB was found to contain 2.5 — 3.8 mol Fe/ mol protein and a
stoichiometric amount of acid-labile sulfide. The F-S clusters in the as-isolated protein
are essentially EPR-silent, presumably in the [4Fe-4S]2+ form, with about 2 % of [3Fe-
4S]'+ cluster. A high specific activity of 109 U/ mg enzyme was achieved for PFL-AB
containing 3.8 mol Fe/ mol protein.

A fairly simple procedure for enzymatic synthesis of S-adenosylmethionine on a
preparative scale has been established. AdoMet is synthesized using a crude extract of E.
coli cells overexpressing AdoMet synthetase. Product inhibition was overcome by
addition of 8 % fl-mercaptoethanol. The purified AdoMet has high diastereoisomeric
purity, and appears to be more stable than AdoMet iodide salt from commercial sources.

Methyl-Dg-AdoMet, methyl-”C—AdoMet, and Se-AdoMet have been synthesized using

94

 

 

 

 

the method described here, with some modifications of the reaction conditions according

to the properties of the substrates.

Using isotopically labeled AdoMet and Se-AdoMet, along with 2H and 13C
ENDOR spectroscopy, a close association of PFL-AB [4Fe-4S] cluster with AdoMet has
been demonstrated for the first time. These results suggest that the Fe-S cluster of PFL-
AE is directly involved in generating the putative adenosyl radical from AdoMet. The
2H and 13C ENDOR studies have demonstrated that the position of AdoMet relative to
the Fe—S cluster of PFL-AE is identical in the oxidized and reduced states, and the
sulfonium methyl group lies closely to the Fe—S cluster (Figure IV.6).

EPR spectrum of PFL-AB [4Fe—4S]1+ / Se-AdoMet complex displays some
different features from the one of PFL-AB [4Fe-4S]'+ / natural AdoMet complex,
suggesting an effect of selenium on the interaction between AdoMet and the [4Fe-4S]1+
cluster of PFL-AB. However, spectroscopic evidence for the Se-Fe interaction has not
yet been observed by using Se XAS, and ENDOR spectra of PFL-AB [4Fe-4S]H / Se-

AdoMet complex awaits further analysis.

95

 

 

 

 

REFERENCES

Conradt, H., Hohmann-Berger, M., Hohmann, H. P., Blaschkowski, H. P., and
Knappe, J. (1984) Archives of Biochemistry and Biophysics 228, 133—142.

Knappe, J ., Neugebauer, F. A., Blaschkowski, H. P., and Ganzler, M. (1984)
Proceedings of the National Academy of Science U.S.A. 81, 1332-1335.

Knappe, J ., Elbert, S., Frey, M., and Wagner, A. F. V. (1993) Biochem. Soc.
Trans. 21, 731-734.

Wagner, A. F. V., Frey, M., Neugebauer, F. A., Schafer, W., and Knappe, J.
(1992) Proc. Natl. Acad. Sci. U.S.A. 89, 996-1000.

Reddy, S. G., Wong, K. K., Parast, C. V., Peisach, J ., Magliozzo, R. S., and
Kozarich, J. W. (1998) Biochemistry 37, 558-563.

Parast, C. V., Wong, K. K., Lewisch, S. A., and Kozarich, J. W. (1995)
Biochemistry 34, 2393-2399.

Becker, A., Fritz-Wolf, K., Kabsch, W., Knappe, J ., Schultz, S., and Wagner, A.
F. V. (1999) Nature Structural Biology 6, 969-975.

Frey, M., Rothe, M., Wagner, A. F. V., and Knappe, J. (1994) The Journal of
Biological Chemistry 269, 12432—12437.

Kessler, D., Leibrecht, I., and Knappe, J. (1991) FEBS Letters 281, 59-63.

Kessler, D., Herth, W., and Knappe, J. (1992) The Journal of Biological
Chemistry 267, 18073-18079.

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

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

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

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.
(2000) Biochemical and Biophysical Research Communications 269, 451-456.

96

 

 

 

l5.

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

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

Henshaw, T. F ., Cheek, J ., and Broderick, J. B. (2000) Journal of The American
Chemical Society 122, 8331-8332.

Kfilzer, R., Pils, T., Kappl, R., Hfittermann, J ., and Knappe, J. (1998) The Journal
of Biological Chemistry 273, 4897-4903.

Rddel, W., Plaga, W., Frank, R., and Knappe, J. (1988) Eur. J. Biochem 177, 153-
158.

Sun, X., Eliasson, R., Pontis, E., Andersson, J ., Buist, G., Sjoeberg, R.-M., and
Reichard, P. (1995) The Journal of Biological Chemistry 270, 2443-2446.

Otsuka, A. J ., Buoncristiani, M. R., Howard, P. K., Flamm, J ., Johnson, C.,
Yamamoto, R., Uchida, K., Cook, C., Ruppert, J ., and Matsuzaki, J. (1988) The
Journal of Biological Chemistry 263, 19577-19585.

Hayden, M. A., Huang, 1., Bussiere, D. E., and Ashley, G. W. (1992) The Journal
of Biological Chemistry 267, 9512-9515.

Fajardo—Cavazos, P., Salazar, C., and Nicholson, W. L. (1993) Journal of
Bacteriology 175, 1735-1744.

Ruzicka, F. J ., Lieder, K. W., and Frey, P. A. (2000) Journal of Bacteriology 184,
469-476.

Beinert, H. (2000) Journal of Biological Inorganic Chemistry 5, 2-15.
Beinert, H., Holm, R. H., and Munck, E. (1997) Science 277, 653-659.
Johnson, M. K. (1994) Encyclopedia of Inorganic Chemistry, 1896-1915.
Howard, J. B., and Rees, D. C. (1996) Chemical Reviews 96, 2965-2982.
Flint, D. H., and Allen, R. M. (1996) Chemical Reviews 96, 2315-2334.

Beinert, H., Kennedy, M. C., and Stout, C. D. (1996) Chemical Reviews 96, 2335-
2374.

Harrison, P. M., and Arosio, P. (1996) Biochim Biophys Acta 1275, 161-203.

Khoroshilova, N ., Popescu, C., Mfinck, B., Beinert, H., and Kiley, P. J. (1997)
Proc. Natl. acad. Sci. U.S.A. 94, 6087-6092.

97

 

 

 

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

Hidalgo, B., Ding, H. G., and Demple, B. (1997) Trends Biochem. Sci. 22, 207-
210.

Leuthner, B., Leutwein, C., Schulz, H., Horth, P., Haehnel, W., Schlitz, B.,
Schagger, H., and Heider, J. (1998) Molecular Microbiology 28, 615-628.

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

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

Duin, E. C., Lafferty, M. B., Crouse, B. R., Allen, R. M., Sanyal, 1., Flint, D. H.,
and Johnson, M. K. (1997) Biochemistry 36, 11811.

Sanyal, 1., Cohen, 6., and Flint, D. H. (1994) Biochemistry 33, 3625.

Busby, R. W., Schelvis, J. P. M., Yu, D. S., Babcock, G. T., and Marletta, M. A.
(1999) Journal of The American Chemical Society 121, 4706.

Ollagnier—de Choudens, 8., and Fontecave, M. (1999) F EBS Letters 453, 25.

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

Sofia, H. J ., Chen, G., Hetzler, B. G., Reyes-Spindola, J. F., and Miller, N. E.
(2001) Nucleic Acids Research 29, 1097-1106.

Reichard, P. (1988) Annual Review of Biochemistry 57, 349-374.

Reichard, P. (1993) Science 260, 1773-1777.

Reichard, P. (1997) Trends in Biochemical Science 22, 81-85.

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

Eliasson, R., Pontis, B., Fontecave, M., Gerez, C., Harder, J ., Jcmvall, H., Krook,
M., and Reichard, P. (1992) The Journal of Biological Chemistry 267, 25541-
25547.

Bianchi, V., Eliasson, R., Fontecave, M., Mulliez, B., Hoover, D. M., Matthews,
R. G., and Reichard, P. (1995) Biochemical and Biophysical Research
Communications 197, 792-797.

98

 

 

 

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

61.

62.

63.

 

Mulliez, E., Fontecave, M., Gaillard, J ., and Reichard, P. (1993) The Journal of
Biological Chemistry 268, 2296-2299.

Tamarit, J ., Mulliez, E., Meier, C., Trautwein, A., and Fontecave, M. (1999) The
Journal of Biological Chemistry 274, 31291-31296.

Coschigano, P. W., Wehrman, T. S., and Young, L. Y. (1998) Appl. Environ.
Microbiology 64, 1650-1656.

Bell, H. R., and Spommann, A. M. (1997) Journal of Bacteriology 179, 670-676.

Chirpic, T. P., Zappia, V., Costilow, R. N., and Barker, H. (1970) The Journal of
Biological Chemistry 245, 1778-1789.

Song, K. B., and Frey, P. A. (1991) The Journal of Biological Chemistry 266,
7651-7655.

Petrovich, R. M., Ruzicka, F. J ., Reed, G. H., and Frey, P. A. (1991) The Journal
of Biological Chemistry 266, 7656-7660.

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

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

Ifuku, 0., Kishimoto, J ., Haze, S., Yanagi, M., and Fukushima, S. (1992) Biosci.
Biotechnol. Biochem. 56, 1780-1785.

Escalettes, F., Florentin, D., Bui, B. T. S., Lesage, D., and Marquet, A. (1999)
Journal of The American Chemical Society 121, 3571-3578.

Reed, K. J ., and Cronan, J. E. J. (1993) Journal of Bacteriology 175, 1325-1336.

Vanden Boom, T. J., Reed, K. E., and Cronan, J. E. J. (1991) Journal of
Bacteriology 173, 6411-6420.

Hayden, M. A., Huang, 1. Y., Iliopoulos, G., Orozco, M., and Ashley, G. W.
(1993) Biochemistry 32, 3778-3782.

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

Nicholson, W. L., Chooback, L., and Fajardo-Cavazos, P. (1997) Mol. Gen. Genet
255, 587-594.

99

 

 

Jam's- . 4. '- - 3m

 

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

 

Frey, P. A. (1990) Chemical Reviews 90, 1343-1357.

Frey, P. A. (2001) Annual Review of Biochemistry 70, 12 l- 148.

Knappe, J ., Schacht, J ., W., M., Hoepner, T., Vetter, H., Jr., and Edenharder, R.
(1969) European Journal of Biochemistry 11, 316-327.

Knappe, J ., Blaschkowski, H. P., Groebner, P., and Schmitt, T. (1974) European
Journal of Biochemistry 50, 253-263.

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

Kennedy, M. C., Emptage, M. H., Dreyer, J .-L., and Beinert, H. (1983) The
Journal of Biological Chemistry 258, 11098-1 1105.

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

Werst, M. M., Kennedy, M. C., Houseman, A. L. P., Beinert, H., and Hoffman, B.
M. (1990) Biochemistry 29, 10526-10532.

Werst, M. M., Kennedy, M. C., Houseman, A. L. P., Beinert, H., and Hoffman, B.
M. (1990) Biochemistry 29, 10533-10540.

J anda, M., and Hemmerich, P. (1976) Angewandte Chemie International Edition
English 15, 7.

Ashton, W. T., Brown, R. D., and Tolman, R. L. (1978) Journal of Heterocyclic
Chemistry 15, 489-491.

Smit, P. (1986) Recyl. Trav. Chim. Pays-Bays. 105, 538-544.

Garboczi, D. N., Hullihen, J. H., and Pedersen, P. L. (1988) The Journal of
Biological Chemistry 263, 15694—15698.

Bradford, M. (1976) Analytical Biochemistry 72, 248.
Beinert, H. (1978) Methods in Enzymology 54, 435-445.
Beinert, H. (1983) Analytical Chemistry 131, 373—378.
Aasa, R., and Vanngard, T. (1975) J. Magn. Reson. 19, 308.

Murib, J. H., and Riter, D. M. (1952) Journal of The American Chemical Society
74, 339.

 

 

 

 

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

 

Cantoni, G. L. (1953) The Journal of Biological Chemistry 204, 403-416.
Tabor, C. W., and Tabor, H. (1984) Advances in Enzymology 56, 251-282.
Tabor, C. W., and Tabor, H. (1985) Microbiology Reviews 49, 81.

Satoh, S., and Yang, S. F. (1989) Plant Physiology 91, 1036-1039.

Slany, R. K., 8631, M., Crain, P. F., and Kersten, H. (1993) Biochemistry 32,
7811-7817.

Pascale, R. M., Marras, V., Simile, M. M., Daino, L., Pinna, G., Bennati, S.,
Carta, M., Seddaiu, M. A., Massarelli, G., and Feo, F. (1992) Cancer Research
52, 4979-4986.

Kagan, B. L., Sultzer, D. L., Rosenlicht, N., and Gemer, R. H. (1990) American
Journal of Psychiatry 147, 591-595.

Bell, K. M., Plon, L., Bunney, J. W. E., and Potkin, S. G. (1988) American
Journal of Psychiatry 145, 1 1 10-1114.

Metz, J. (1993) Nutrition Reviews 51, 12-15.

Matos, J. R., Raushel, F. M., and Wong, C.-H. (1987) Biotechnology and Applied
Biochemistry 9, 39-52.

Schalk-Hihi, C., and Markham, G. D. (1999) Biochemistry 38, 2542-2550.

Markham, G. D., Hafner, E. W., Tabor, C. W., and Tabor, H. (1980) The Journal
of Biological Chemistry 255, 9082-9092.

Markham, G. D. (1981) The Journal of Biological Chemistry 256, 1903-1909.

Zhang, C., Markham, G. D., and LoBrutto, R. (1993) Biochemistry 32, 9866-
9873.

Park, J ., Tai, J ., Roessner, C. A., and Scott, A. I. (1996) Bioorganic & Medicinal
Chemistry 4, 2179-2185.

Markham, G. D. (1986) The Journal of Biological Chemistry 261, 1507-1509.

Markham, G. D., and Leyh, T. S. (1987) Journal of The American Chemical
Society 109, 600-602.

Gilliland, G. L., Markham, G. D., and Davies, D. R. (1983) The Journal of
Biological Chemistry 258, 6963-6964.

101

 

 

 

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

Mudd, S. H., and Cantoni, G. L. (1957) Nature 180, 1052.
Greene, R. C. (1969) Biochemistry 8, 2255-2265.
Barth, H., and Gosselck, J. (1961) Z. Naturforsch. 166, 280.

Pearson, R. G., Sobel, H., and Songstad, J. (1968) Journal of The American
Chemical Society 90, 319.

Knappe, J ., and Sawers, G. (1990) FEMS Microbiology Reviews 75, 383-398.
Wong, K. K., and Kozarich, J. W. (1994) Met. Ions Biol. Syst. 30, 279-313.

Cheek, J ., and Broderick, J. B. (2001) Journal of Biological Inorganic Chemistry
6, 209-226.

Ollagnier, S., Meier, C., Mulliez, E., Gaillard, J ., Schuenemann, V., Trautwein,
A., Mattioli, T., Lutz, M., and Fontecave, M. (1999) Journal of The American
Chemical Society 121, 6344.

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

Davydov, R. M., Yoshida, T., Ikeda-Saito, M., and Hoffman, B. M. (1999)
Journal of The American Chemical Society 121, 10656-10657.

Davydov, R., MacDonald, I. D. G., Makris, T. M., Sligar, S. G., and Hoffman, B.
M. (1999) Journal of The American Chemical Society 121, 10654-10655.

Davydov, R., Valentine, A. M., Komar-Panicucci, S., Hoffman, B. M., and
Lippard, S. (1999) Biochemistry 38, 4188-4197.

Davydov, R., Sahlin, M., Kuprin, S., Graslund, A., and Ehrenberg, A. (1996)
Biochemistry 35, 5571-5576.

Davydov, A., Davydov, R., Graslund, A., Lipscomb, J ., and Anderson, K. K.
(1997) The Journal of Biological Chemistry 272, 7022-7026.

Davydov, R., Kuprin, S., Graslund, A., and Ehrenberg, A. (1994) Journal of The
American Chemical Society 116.

Dikanov, S. A., Davydov, R. M., Xun, L., and Bowman, M. K. (1996) Journal of
Magnetic Resonance 112, 289-294.

Telser, J ., Davydov, R., Kim, C. H., Adams, M. W. W., and Hoffman, B. M.
(1999) Inorganic Chemistry 38, 3550-3553.

102

 

 

 

117.

118.

119.

120.

121.

122.

123.

124.

125.

126.

127.

128.

129.

130.

131.

132.

133.

Yoo, S., J ., Angove, H. C., Burgess, B. K., Mfinck, B., and Peterson, J. (1998)
Journal of The American Chemical Society 120, 9704-9705.

Telser, J., Huang, H., Lee, H. I., Adams, M. W. W., and Hoffman, B. M. (1998)
Journal of The American Chemical Society 120, 861-870.

Odom, J. D. (1983) Structure and Bonding 54, 1-26.

Bisbjerg, B. (1972) Studies on Selenium in Plants and Soils, Danish Atomic
Energy Commission.

Rosenfeld, I., and Beath, O. A. (1964) Selenium Geobotany, Biochemistry,
Toxicity and Nutrition, Academic Press, New York.

Committee on Medical and Biologic Effects of Environmental Pollutants:
Selenium ( 1976) National Academy of Sciences, Washington, D. C.

Pauling, L. (1960) The Nature of Chemical Bond, Chaps. 3 and 7, Cornell
University Press, Ithaca, N. Y.

Spallholz, J. E., Martin, J. L., and Ganther, H. E. (1981) Selenium in Biology and
Medicine.

Stadtman, T. C. (1990) Annual Review of Biochemistry 59, l 1 1.

Diplock, A. T. (1981) Philosophical Transactions of the Royal Society of
Londond Series B, Biological Sciences 294, 105-117.

Sunde, R. A. (1990) Annual Review of Nutrition 10, 451.

Gunther, W. H. H., and Mautner, H. G. (1965) Journal of The American Chemical
Society 87, 2708.

Piffeateau, A., Gaudry, M., and Marquet, A. (1976) Biochemical and Biophysical
Research Communications 73, 773.

Marquet, A. (1977) Pure and Application Chemistry 49, 183.
Wu, M., and Wachsman, J. T. (1971) Journal of Bacteriology 105, 1222-1223.

Wu, Z.-P., and Hilvert, D. (1989) Journal of The American Chemical Society 111,
4513-4514.

Walter, R., and Chan, W. Y. (1967) Journal of The American Chemical Society
89, 3892-3898.

103

 

 

 

134.

135.

136.

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.

149.

150.

 

Hartrodt, B., Neubert, K., Bierwolf, B., Blech, W., and Jakubke, H.-D. (1980)
Tetrahedron Letters 21.

Oikawa, T., Esaki, N., Tanaka, H., and Soda, K. (1991) Proceedings of the
National Academy of Science U.S.A. 88, 3057-3059.

Frank, P., Licht, A., Tullius, T. D., and Hodgson, K. O. (1985) The Journal of
Biological Chemistry 260, 5518-5520.

Yang, W., Hendrickson, W. A., Kalman, E. T., and Crouch, R. T. (1990) The
Journal of Biological Chemistry 265, 13553-13559.

Meyer, J ., Moulis, J .-M., Gaillard, J ., and Lutz, M. (1992) Advances in Inorganic
Chemistry 38, 73-115.

Stadtman, T. C. (1974) Science 183, 915.
Stadtman, T. C. (1979) Advances in Enzymology 48, l. '

Klayman, D. L., and Gunther, W. H. H. e. (1973) Orangic Selenium Compounds:
Their Chemistry and Biology, Wiley, New York, Chap. XHIE.

Lipsett, M. N. (1965) The Journal of Biological Chemistry 240, 3975.

Abrell, J. W., Kaufman, E. F., and Lipsett, M. N. (1971) The Journal of
Biological Chemistry 246, 294.

Dilworth, G. L., and Bandurski, R. S. (1977) Biochemical Journal 163, 521.

Bremer, J ., and Natori, Y. (1960) Biochimica et Biophysica Acta. International
Journal of Biochemistry and Biophysics 44, 367-370.

Pegg, A. E. (1969) Biochimica et Biophysica Acta. International Journal of
Biochemistry and Biophysics 177, 361-364.

Shepherd, L., and Huber, R. E. Canadian Journal of Biochemistry 47, 877-881.

Weatherill, T. D., Rauchfuss, T. B., and Scott, R. A. (1986) Inorganic Chemistry
25, 1466-1472.

Eidsness, M. K., Scott, R. A., Prickril, B. C., derVartanian, D. V., LeGall, J ., .
Moura, I., Moura, J. J. G., and Peck, H. D., Jr. (1989) Proceedings of the National

Academy of Science U.S.A 86, 147.

Cosper, N. J ., Booker, S. J ., Ruzicka, F., Frey, P. A., and Scott, R. A. (2000)
Journal of The American Chemical Society 39, 15668-15673.

104

 

 

 

 

 

 

 

151. He, S. H.,Teixeira,M., LeGa11,J., Patil, D. S., Moura, I., Moura, J. J. G.,
DerVartanian, D. V., Huynh, B. H., and Peck, H. D., Jr. (1989) The Journal of
Biological Chemistry 264, 2678.

152. Chasteen, N. D., and Snetsinger, P. A. (2000) Physical Methods in Bioinorganic
Chemistry, Spectroscopy and Magnetism, University Science Books.

153. Pickering, 1. J ., George, G. N., Fleet-Stalder, V. V., Chasteen, T. G., and Prince,
R. C. (1999) Journal of Biological Inorganic Chemistry 4, 791-794.

105