..;.,.
Via; . 1
Ann“. .uufirfi...

faru

fink 55".“.
n . .
, .4”. 11 .
l. ‘9!
4....
I I a

tilt!“
Juummmmmm

. 1 .3
Igfltflfl‘.
7.. t. “Ina!“

5,.: .
Sinfii.

5

1..
.11
Lt}:

IQ. rt.
53!“
.4 ta

. .z
1.0.1.2111":

:50»! a.

an..huunuur..hwunu .
¢.Juman“uu.‘:. ‘

I. 0.1.1.
Ila:

 

.I. 114111.:

9051..

«1‘ I"

in

fabric?!

x 13:5. 4?
in“ 3!”.

. 1 .Ixttl‘.‘ ..

. Didi“?!

318%.? .ihmu
in!

.1...lil....r€..n,(.h~4. .2?!
.33 , x... .11
1.7 53.1.8 3.
.2 £93.... (Inn...
1...)... #3.".
{arkflfildfi tux}!
«Hiya-V...»

..:

“Fifi

no

 

LIBRARY
Michigan State
, “yniversity

 

 

 

This is to certify that the
dissertation entitled

PYRUVATE FORMATE LYASE AND PYRUVATE FORMATE
LYASE ACTIVATING ENZYME: SPECTROSCOPIC
CHARACTERISTICS, INTERACTION AND MECHANISM

presented by

YI PENG

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Chemistry

 

 
       
 

   
 

,) ///‘V / V '\
'ajor P ofes r’s Sign‘at re

AM +5 297 2&95/

Date

MSU is an affinnative-acfion, equal-opportunity employer

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K lProflAccAPres/ClRC/DateDue,indd

PYRUVATE FORMATE LYASE AND PYRUVATE FORMATE LYASE
ACTIVATING ENZYME: SPECTROSCOPIC CHARACTERISTICS,
INTERACTION AND MECHANISM
By

Yi Peng

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

2008

ABSTRACT

PYRUVATE FORMATE LYASE AND PYRUVATE F ORMATE LYASE
ACTIVATING ENZYME: SPECTROSCOPIC CHARACTERISTICS, INTERACTION
AND MECHANISM
By

Yi Peng

Pyruvate formate-lyase (PF L) reversibly converts pyruvate and coenzyme A
(CoA) into formate and acetyl-CoA, a key step in anaerobic metabolism for facultative
anaerobes. A glycyl radical on Gly-734 in PFL, generated by the pyruvate formate-lyase
activating enzyme (PFL-AB), initiates the reaction. PFL-AB utilizes S-adenosylmthionine
(AdoMet or SAM), which binds to a unique site in the Fe4S4 cluster of PFL-AB, to
produce a putative deoxyadenosyl radical intermediate. It is this intermediate that
abstracts the hydrogen from PFL GIy-734 to generate the glycyl radical. Sulfur K-edge
X-ray absorption spectroscopy (XAS) has been used to probe the interaction between
AdoMet and the cluster. The spectra show a sulfur K-edge absorption at 2472.5 eV and a
pre-edge feature at 2470.3 cV. Upon AdoMet binding, the intensity of the pre-edge
absorption increases, indicating change in Fe-S bond covalency. DFT calculations
illustrate a 37% increase in covalency of the Fe-S bonds, in which 12% is attributed to
FC'H3Ssulfidc and 25% to Fe-‘Sthioim. The optimized model shows that the AdoMet 5’-C-S
0'" bond has backbonding interactions with filled Fe-S bonds, which shifts the charge
density from the cluster to AdoMet. The effect causes enhancement of overall cluster

covalency, and induces the S Is to 5’-C-S transition in the pre-edge feature.

To investigate the possible exogenous ligand binding to the unique Fe in the
absence of AdoMet, nuclear resonance vibrational spectroscopy (NRVS) was used. The
spectra exhibit cubane cluster twist bands below 100 cm", S-Fe-S’ bending modes
around 160 cm", and Fe-S stretching modes above 200 cm". In the presence of AdoMet,
the intensity of the Fe-ligand + Fe—tS stretching mode at 432 cm'l increases, which might
provide clues as to the identity of the exogenous ligand.

PFL-AB has no direct access to Gly—734, which is buried in the core of PFL,
however direct interaction between them is necessary to generate the glycyl radical. We
have obtained evidence from EPR spin quantification, H/D exchange of Gly-734 and PFL
activity assays that Gly-734 is more exposed and unstable in the presence of PFL-AB,
supporting the notion that the PF L/PFL-AE interaction involves Significant
rearrangement of the glycine loop. Circular dichroism (CD) spectra of activated PFL
Show that PFL-AB causes a conformational change in PFL. All evidence supports the
proposal that the glycine loop undergoes conformational change to allow Gly—734 to bind
at the PFL-AB active site during activation. Investigation of the PFL structure suggests
that the C-terminal region, residues 710 — 759, might be the structure interacting with
PFL-AE and pulling Gly-734 out.

It has been proposed that a thiyl radical on Cys-418 or Cys-419 might initiate PFL
catalysis. EPR spectra Show that activated C418A-PFL in D20 has H/D exchange on the
glycyl radical, indicating Cys-419 is not acetylated by pyruvate. The result implies that
CyS-419 does not attack pyruvate directly, and that Cys-418 might be the residue to

initiate the reaction.

I dedicate this work to my lovely and pretty wife Haolan Dong for her thorough support
in my life no matter what it is, and putting up during all the hard time. I could never have
made it this far without you and expect much more in the rest of our life.

iv

ACKNOWLEDGEMENTS

I would like to express sincere appreciation to my advisor Dr. Joan Broderick,
who has provided me with academic guidance through my whole graduate program, and
encouraged me to pursue research goals. Thank you for everything and wish you the best
in your academic and personal lives.

I would also like to thank my graduate committee members, Dr. David Weliky,
Dr. Jack Watson and Dr. Gary Blanchard who helped me set up the laser for the Raman
spectrometer and gave me the suggestion in resonance Raman spectroscopy. I thank Dr.
Gavin Reid who did the mass spectrometry for the product of PFL.

I would like to thank Dr. Mbako Nnyepi for teaching me how to conduct PF L
activity assay. I wish I could see you again sometime in Botswana. I also appreciate Dr.
Jian Yang showing me how to use the EPR machine. My thanks go to Dr. Danilo Ortillo
and Dr. Jeff Buis for their help in using the lab equipments. I acknowledge Dr. Will
Broderick from whom I learned how to purify PF L-AE.

I appreciate our collaborators, Prof. Ed Solomon and Dr. Abhishek Dey for their
work in the XAS project, Dr. Tim Sage and Weiqiao Zeng for their work in the NRVS
project, and Prof. Vincent Huynh and Dr. Ricardo Garcia for his help of the Mdssbauer
spectra of the NRVS samples.

I would also like to thank my lab mates, Dr. Meng Li for her friendship; Dr.Susan
Veneziano for explaining to me so many American slangs; Dr. Tilak Chandra’s
incredible Indian foods; and all other members in our lab, Magdalena Makowska-Gryska,
Ziyang Su, Efihalia Kalliri, Kaitlin Duschene, Sunshine Silver, Rachel Undelhoven and

Eric Sheperd. It has been always a lot of fun working with you.

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. xi

LIST OF SCHEMES ............................................................................... xii
LIST OF FIGURES ................................................................................ xiii
LIST OF ABBREVIATIONS ..................................................................... xv
CHAPTER I

INTRODUCTION ..................................................................................... 1

1.1 PF L is a key enzyme in anaerobic metabolism for facultative anaerobes ................. 1
1.2 PF L is a free radical enzyme .................................................................... 2
1.3 Structure of PFL .................................................................................. 4
1.4 Generation of the glycyl radical in PFL requires an activating enzyme .................. 5

1.5 PFL-AE is a radical SAM superfamily enzyme .............................................. 7
1.6 AdoMet binds to the [4Fe-48] cluster through the unique Fe Site ........................ 13
References ............................................................................................ 16
CHAPTER II

OVEREXPRESSION AND PURIFICATION OF PYRUVATE FORMATE-LYASE
AND PYRUVATE F ORMATE-LYASE ACTIVATIN G ENZYME ........................ 21
11.1 Introduction ..................................................................................... 21

11.2 Materials and methods ........................................................................ 26
11.2.1 Chemicals, bacterial cells and plasmids .................................................... 26
11.2.2 Overexpression of PFL ...................................................................... 26
11.2.3 Purification of PFL ............................................................................ 27
11.2.4 Overexpression of PF L-AE ................................................................. 29
11.2.5 Purification of PF L-AE ..................................................................... 30
11.2.6 Determination of protein concentration .................................................... 31
11.2.7 Determination of iron content in PFL-AB ................................................ 31

11.3 Results and discussion ......................................................................... 32
11.3.1 Overexpression and purification of PFL ..................................................... 32
11.3.2 Overexpression and purification of PF L-AE ............................................... 34
11.4 Conclusions ..................................................................................... 36
References ............................................................................................ 38
CHAPTER 111

SULFUR LIGAND K-EDGE X-RAY ABSORPTION SPECTROSCOPIC STUDY OF
PYRUVATE F ORMATE-LYASE ACTIVATIN G ENZYME ............................... 40
111.1 Introduction .................................................................................... 40
111.2 Materials and methods ......................................................................... 47
1112.1 Chemicals and materials ................................................................... 47

vi

111.2.2 Overexpression and purification of PFL-AE ............................................ 47

111.2.3 Synthesis of S-adenosylmethionine ...................................................... 47
111.2.4 Sample preparation for sulfur XAS ...................................................... 48
111.2.5 PFL-AE photo-reduction .................................................................. 49
111.2.6 UV-Vis spectroscopy ....................................................................... 49
[11.2.7 EPR spectroscopy ............................................................................ 50
111.2.8 Sulfur K-edge XAS ......................................................................... 50
111.3 Results and discussion ......................................................................... 52
111.3.1 Overexpression and purification of PFL-AB ............................................ 52
111.3.2 Synthesis of S-adenosylmethionine ...................................................... 53
111.3.3 UV-Vis spectra of PFL-AB purified in ascorbic acid and DTT ....................... 54
111.3.4 EPR spectra of photo-reduced PFL-AB ................................................... 57
111.3.5 Sulfur K-edge XAS of PF L-AE ........................................................... 63
111.4 Conclusions .................................................................................... 66
References ............................................................................................ 69
CHAPTER IV

NUCLEAR RESONANCE VIBRATIONAL SPECTROSCOPIC STUDY OF
PYRUVATE FORMATE-LYASE ACTIVATING ENZYME ............................... 71
NJ Introduction .................................................................................... 71
1V.2 Materials and methods ......................................................................... 79
1V.2.l Chemicals and materials .................................................................... 79
1V.2.2 57Fe labeled PFL-AE preparation ........................................................... 8O
1V.2.3 NRVS sample preparation ................................................................. 81
1V.2.4 Mossbauer spectroscopy ................................................................... 81
1V.2.5 Resonance Raman spectroscopy .......................................................... 81
1V.2.6 NRVS spectroscopy ........................................................................ 82
IV.3 Results and discussion ......................................................................... 82
IV.3.1 Mossbauer spectra of PF L-AE ............................................................ 82
IV.3.2 Resonance Raman spectra of PFL-AE .................................................... 84
IV.3.3 NRVS spectra of PFL-AE .................................................................. 86
1V.4 Conclusions .................................................................................... 90
References ............................................................................................ 92
CHAPTER V

STRUCTURE CHANGE OF PYRUVATE F ORMATE-LYASE DURING
ACTIVATION ....................................................................................... 94
V.1 Introduction ..................................................................................... 94
V2 Materials and methods ......................................................................... 95
V21 Chemicals and materials ...................................................................... 95
V22 PF L activation ................................................................................. 95
V23 EPR spectroscopy .............................................................................. 96
V24 Hydrogen-deuterium exchange of activated PFL in D20 .............................. 96
V25 PFL activity assay ............................................................................ 97
V26 CD sample preparation ...................................................................... 98
V3 Results and discussion .......................................................................... 99

vii

V.3.1 Spin quantification of glycyl radical in activated PFL .................................. 99

V.3.2 Hydrogen-deuterium exchange of activated PFL in D20 ............................. 107
V.3.3 PFL activity assay ........................................................................... 110
V.3.4 CD spectra of PFL activated with different amounts of PFL-AE .................... 111
V.3.5 PFL structure analysis ..................................................................... 113
VA Conclusions .................................................................................... 116
References .......................................................................................... 1 19
CHAPTER V1

INVESTIGATIONS OF THE PYRUVATE F ORMATE-LYASE CATALYTIC
MECHANISM ....................................................................................... 121
V1.1 Introduction ................................................................................... 121
V1.2 Materials and methods ...................................................................... 124
V1.2.1 Chemicals and materials ................................................................... 124
V1.2.2 Synthesis of sodium 3,3,3-d3-pyruvate .................................................. 124
V1.2.3 PFL activity assay .......................................................................... 124
V1.2.4 PFL enzyme reaction ...................................................................... 125
V1.2.5 Mass spectrometry of acetyl-CoA from PF L enzyme reaction ...................... 125
V1.2.6 H/D exchange in mutants after reaction with pyruvate ................................ 126
V1.2.7 EPR spectroscopy .......................................................................... 126
V1.3 Results and discussion ....................................................................... 127
V1.3.1 PFL activity assay of sodium pyruvate and sodium 3,3,3-d3-pyruvate ............. 127
V1.3.2 Mass spectra of 2’,2’,2’-d3-acetyl-CoA ................................................ 128
V1.3.3 Activity assay of C418A-PFL and C419A-PF L ....................................... 130
V1.3.4 H/D exchange of PF L and PFL mutants after reaction with pyruvate .............. 131
V1.4 Conclusions ................................................................................... 133
References .......................................................................................... 136
CHAPTER VII

GENERAL CONCLUSIONS AND FUTURE DIRECTIONS .............................. 138
V11.1 The interaction between AdoMet and the iron-sulfur cluster in PFL-AE ........... 138
V11.2 The vibrations of the iron-sulfur cluster in PFL-AB ................................... 140
VII.3 PFL structure change during activation .................................................. 142
V11.4 The first step in PFL catalysis ............................................................. 142
VII.5 Overview of PFL-AE and PFL catalytic mechanism .................................. 144
VII.6 Future work on sulfur XAS of reduced PF L-AE ........................................ 144
v11.7 Future work on NRVS studies of PFL-AB in H2130 .................................. 144
V11.8 Future work on the PFL and PFL-AE interaction ...................................... 145
VII.9 Future work on the investigation of the first step in PFL catalysis .................. 145
References .......................................................................................... 146

viii

LIST OF TABLES
Table 111.1 DFT calculations of covalencies of the [4Fe-4S]2+ cluster in PF L-AE ....... 65
Table 1V.1 NRVS normal modes assignments of PFL-AB and PFL-AE with AdoMet...88

Table V.1 Spin quantification of glycyl radical in 5.0 uM PF L activated with different

amount of PFL-AB ................................................................................. 102
Table V.2 Spin quantification of glycyl radical in 50 uM PF L activated with 0.5 and 50
uM PFL-AB for different time ................................................................... 105
Table V.3 Spin quantification of glycyl radicals in 50 uM PFL illuminated for 2 hours
with different amounts of AdoMet ............................................................... 106
Table V.4 Activity of 100 M of PF L activated with 100, 10 and 1 M PFL-AE .......... 111
Table V1.1 Turnover rates of sodium pyruvate and sodium 3,3,3-d3-pyruvate ........... 128
Table V1.2 Activity assays of wild-type PF L, C418A-PFL and C419A-PFL ............ 131

ix

LIST OF SCHEMES
Scheme 1.1 Ping-pong mechanism of PFL catalysis ............................................. 1
Scheme 1.2 Acetylated PFL radical reacting with hypophosphite ............................. 3

Scheme 1.3 Addition of deoxyadenosyl to the AAla as a result of AdoMet cleavage ....... 6

Scheme 1.4 AdoMet cleavage by radical SAM superfamily enzymes ........................ 8
Scheme 1.5 Iron-sulfur clusters conversion .............. . ........................................ 10
Scheme 1.6 AdoMet cleaved by [4Fe—48] H ...................................................... 11
Scheme 111.1 AdoMet synthesis catalyzed by SAM synthetase ................................ 48
Scheme 1V.l AdoMet binding to the unique iron ............................................... 72
Scheme V1.1 PF L mechanism proposed by Kozarich’s and Knappe’s group ............. 122
Scheme V1.2 PF L mechanism proposed by Broderick’s group .............................. 123

LIST OF FIGURES

Figure 1.1 Conserved Fe-S cluster binding motif in radical SAM superfamily ............ 12
Figure 1.2 Isocitrate binds to the unique Fe in aconitase ....................................... 13
Figure 1.3 AdoMet binds to the [4F e-4S] cluster through the unique Fe .................... 14
Figure 11.1 Protein expression controlled by Lac operator ..................................... 22
Figure 11.2 Structures of allolactose and IPTG .................................................. 23
Figure 11.3 SDS-PAGE of PFL overexpression ................................................ 32
Figure 11.4 SDS-PAGE of PFL purified by anion exchange chromatography .............. 33

Figure 11.5 SDS-PAGE of PFL purified by hydrophobic interaction chromatography. . .34

Figure 11.6 SDS-PAGE of overexpression and purification of PF L-AE ..................... 35
Figure 11.7 EPR spectrum of [3Fe-4S]1+ in PFL-AB ........................................... 36
Figure 111.1 Cleavage of AdoMet ................................................................. 42
Figure 111.2 Ligand pre K-edge transition ....................................................... 44
Figure 111.3 Energy separation of an electron in a magnetic field ........................... 46
Figure 111.4 Electron delocalization in a Fezsz subunit ........................................ 47
Figure 111.5 Experimental setup of X-ray absorption spectrometer .......................... 51

Figure 111.6 SDS-PAGE of PF L-AE purified in the presence of 1 mM ascorbic acid. ...53
Figure 111.7 Chromatography of AdoMet purified by cation exchange column... .........54
Figure 111.8 UV-Vis spectra of PFL-AB purified in DTT and ascorbic acid.................55
Figure 111.9 UV-Vis difference spectra of PF L-AE with and without exogenous thiol...56
Figure 111.10 EPR spectra of reduced PFL-AB and reduced PFL-AE with AdoMet. . ....58
Figure 111.11 EPR spectra of ascorbic acid purified PF L-AE with exogenous thiol ...... 59

Figure 111.12 EPR spectra of DTT purified PFL-AE with exogenous thiol ................ 60

xi

Figure 111.13 EPR spectra of reduced PF L-AE with exogenous thiol and AdoMet ....... 62

Figure 111.14 Sulfur K-edge XAS spectra of PFL-AE and PF L-AE with AdoMet ........ 64

Figure 111.15 DFT optimized structure of F e4S4 cluster bound with AdoMet .............. 66
Figure 1V.1 NRVS photon absorption in 57Fe .................................................... 74
Figure 1V.2 NRVS experimental set-up .......................................................... 75
Figure IV.3 Schematic diagrams of Raman scattering ......................................... 79

Figure 1V.4 Mossbauer spetra of unique site 57Fe labeled PF L-AE with AdoMet. . . . . ....84
Figure 1V.5 Resonance Raman spectra of PF L-AE and PFL-AB with AdoMet... . . . . .....85
Figure 1V.6 NRVS spectra of 5 7F e reconstituted PF L-AE and PFL-AE with AdoMet. ..87
Figure 1V.7 NRVS spectrum of all sites 57Fe labeled PFL-AE ............................... 89
Figure V.l PF L activity assay ..................................................................... 98
Figure V.2 Spin quantification of activated PF L with different amounts of PFL-AB. . . 101
Figure V.3 Spin quantification of activated PF L illuminated for different time .......... 104
Figure V.4 Spin quantification of PFL activated with different amounts of AdoMet. .. 106
Figure V.5 EPR spectra of activated PFL in H20 and D20 ................................. 107

Figure V.6 EPR spectra of activated PFL with different amounts of PFL-AB in D20..109

Figure V.7 CD spectra of activated PF L with different amounts of PFL-AE ............. 112
Figure V.8 Structure of PFL active site ......................................................... 114
Figure V.9 Sequence conservation of C-terminal in PFLs ................................... 114
Figure V.10 Crystal structure of PFL ........................................................... 116
Figure V1.1 Mass spectra of 2’,2’,2’-d3-acetyl-CoA ......................................... 129
Figure V1.2 MS/MS of 2’,2’,2’-d3-acetyl-CoA ................................................ 130

Figure V1.3 EPR spectra of wild-type PFL, C41 8A-PF L and C419A-PF L in D20. . ....133

xii

LIST OF ABREVIATIONS

AdoMet ................................................................... S-adenosyl-L—methionine
aRN R ........................................................... anaerobic ribonucleotide reductase
ATP ........................................................................... adenosine triphosphate
BioB .................................................................................... biotin synthase
B-ME .............................................................................. B—mercaptoethanol
BSA ........................................................................... bovine serum albumin
CD ................... . ............................................................... circular dichroism
CoA ....................................................................................... coenzyme A
DFT ........................................................................... density function theory
DTT ....................................................................................... dithiothreitol
EDTA ........................................... ethylenedinitriolotetra-acetic acid disodiurn salt
ENDOR ......................................................... electron nuclear double resonance
EPR ............................................................... electron paramagnetic resonance
EXF AS .................................................. extended X-ray absorption fine structure
F PLC ........................................................... fast protein liquid chromatography
HemN ....................................... oxygen independent coprophyringogen-III-oxidase
E. coli ................................................................................. Escherichia coli
IPTG .......................................................... isopropyl-fi-D-thioglactopyranoside
LAM ........................................................................ lysine 2,3-aminomutase
LB ........................................................................................ Luria Bertani
MOPS ...................................................... 3-(N-morphlino)propanesu1fonic acid
LMCT ............................................................... ligand to metal charge transfer

xiii

NAD+ ....................................... nicotinamide adenine dinucleotide (oxidized form)

NADH ....................................... nicotinamide adenine dinucleotide (reduced form)
NMR .................................................................... nuclear magnetic resonance
NRVS ................................................ nuclear resonance vibrational spectroscopy
PF L ........................................................................... pyruvate formate—lyase
PFL-AE ................................................ pyruvate formate-lyase activating enzyme
PMSF ............................................................... phenylmethyl sulfonyl
fluoride

RNA .................................................................................. ribonucleic acid
SAM ....................................................................... S-adenosyl-L-methionine
SDS-PAGE ....................... sodium dodecyl sulfate-polyacrylamide gel electrophoresis
SPL ........................................................................ spore photoproduct lyase
XAS ................................................................. X-ray absorption spectroscopy

xiv

CHAPTER 1

INTRODUCTION

1.1 PFL is a key enzyme in anaerobic metabolism for facultative anaerobes
Glycolysis, a series of reactions converting glucose into pyruvate, is an almost
universal route for organisms to generate metabolic energy. Under aerobic conditions,
glycolysis is followed by the citric acid cycle, in which pyruvate is completely oxidized
to CO2 and H20, and reducing equivalents that feed into the electron transport chain are
generated. Most of the ATPs required in the metabolism are generated during the process.
To enter the citric acid cycle, pyruvate and coenzyme A (CoA) are irreversibly converted
into acetyl CoA and C02, catalyzed by pyruvate dehydrogenase complex. Facultative
anaerobes such as E. coli, have an alternative reaction under anaerobic conditions in
which pyruvate formate-lyase (PFL) reversibly converts pyruvate and CoA into formate
and acetyl-CoA; this represents the first and key step in anaerobic metabolism. ' PFL
exhibits two-step ping-pong kinetics with an acetylated enzyme intermediate (Scheme

1.1).

0
ll
CH3COC00' + OE-SHv—OE-S-CCH3 + HCOO'

 

0 1.)
||
gas-con, + CoA ‘ ~ oE-SH + CoA-CCH3

 

 

Scheme 1.1 Schematic representation of ping-pong mechanism of PFL catalysis.

1.2 PF L is a free radical enzyme

It has been demonstrated that PF L can exist in either inactive or active forms; "2
the later is generated under anaerobic conditions and contain a stable protein-centered
free radical. 3 Electron paramagnetic resonance (EPR) spectra revealed a hyperfine
coupling between the electronic spin and the nuclear spin on 13 C in partially l3C-labeled
PFL; together with the lack of coupling to 33 S and other nuclei, this indicated the radical
was carbon centered. 4 In 1992, Knappe and coworkers determined that the radical was
on a glycine residue based on the hyperfine interaction between the unpaired electron and
l3C-labeled glycine. 5 The radical form of PFL is extremely sensitive to 02. 3 In the
presence of 02, activated PFL is cleaved into two fragments, a larger fragment (82 kDa)
and a smaller one (3 kDa). 5 Sequencing of the smaller fragment illustrated that the
enzyme was cleaved at position Gly-734, which was, thus, assigned as the radical site. 5

Incubation of PFL with 2-‘4C-pyruvate results in labeling of the protein to
generation of the acetyl-enzyme intermediate. 6 To identify the position of the 14C label,
the acetyl-PFL was cleaved by CNBr to generate several fragments. Sequencing of the
peptide with radioactive label revealed that the l4C-acetyl moiety was located on Cys-419.
6 When the l4C-acetylated PFL was treated with hypophosphite (a formate analogue
known as an inactivator of PFL 3) to mimic the reverse reaction, the 14C label was found
on Cys-418. 6 The structure of the hypophosphite treated intermediate was determined as
l-hydroxyethylphosphonate with a thioester linkage to Cys-418, using 31P-NMR
spectroscopy and chemical characterization (Scheme 1.2). 6 These results suggested that
both Cys-418 and Cys-419 were involved in PF L radical mechanism. 6 To further

investigate this proposal, two PFL mutants, C419S and C418S, were made by Knappe

and coworkers. 7 Although EPR spectra showed that both mutants could be activated to
contain the glycyl radical as with wild-type PF L, both mutants had no activity, suggesting
that these two cysteines are critical for PFL catalysis. 7

o O
l l (Cy s-4 1 8)

—0'——> 13—— s— P—O'

—O

 

{II—“U

CH3— C_H

0H

Scheme 1.2: Schematic representation of acetylated PFL radical reacting with
hypophosphite.

The glycyl radical gives rise to a doublet EPR signal (g = 2.0037, A = 1.5 mT).
The coupling is due to the interaction between the unpaired electron and the nuclear spin
of the IH on the a-carbon of the glycyl residue. The doublet signal changes to a singlet
when PFL is mixed with D20, 3‘ 4 suggesting that the a-hydrogen on Gly-734 exchanges
with deuterium from the solvent. Kozarich and coworkers showed that C419S-PF L
retained its doublet EPR signal in D20, 8 indicating that Cys-419 was critical for the
exchange. Together, the results suggested that the glycyl radical abstracts a hydrogen
atom from Cys-419, thereby transferring the radical to Cys-419 to form a thiyl radical; in
D20 the thiol of Cys-419 would be deuterated, and so the glycyl radical would abstract
deuterium from Cys-419. Multiple H/D atom abstraction between Gly-734 and Cys-419

would result in essentially full deuteration. Activated C4188-PFL, on the other hand,

undergoes the doublet to singlet EPR signal change, indicating that Cys-418 is not
required for the solvent hydrogen exchange on the glycyl radical. 8

The results just described led to a proposal in which an intermediate thiyl radical
on either Cys-4l9 or Cys-4l 8 attacks pyruvate as a key step in PF L catalysis. 9'” A thiyl
radical is also known to participate in the mechanism of E. coli ribonucleotide reductase
(RNR) catalysis. '2 Abstraction of a 3’-hydrogen from the nucleotide by a cyeteine thiyl
radical is believed to initiate the reduction of the ribonucleotide. Although the thiyl
radical has yet to be observed in PFL, a thiyl radical coupled to cob(11)alarnin in
ribonucleoside triphosphate reductase (RTN) from Lactobacills leichmannii was trapped
by rapid freeze quench techniques and detected by EPR. '3 The detection of a thiyl radical

in PFL would provide strong evidence for the proposed radical mechanism.

1.3 Structure of PFL

PFL is a 170 kDa homodimer, containing 759 amino-acid residues in each subunit
with no metals or bound cofactors. 14‘ 15 The crystal structure of the non-radical form of
PFL and its complex with the substrate analog oxamate were firstly reported by Knappe
and coworkers. ” The overall structure of the monomer is a 10—stranded B/a barrel. Gly-
734 and Cys-419/Cys-41 8 are located on two short loops, pointing toward each other and
approximately 4A apart. The structure of PF L in complex with both substrates (pyruvate
and CoA) bound was solved by Becker and Kabsh. '6 In this structure, pyruvate was
clearly located in the active site, whereas the adenine moiety of CoA was bound at a

distance of 15 A from the active site with its thiol moiety not far away from the pyruvate

binding pocket. This is consistent with the ping-pong mechanism in which CoA binds to

the active site only after pyruvate is cleaved.

1.4 Generation of the glycyl radical in PFL requires an activating enzyme

PFL purified aerobically is in an inactive form. 1,2, '3 To convert PFL to an active
form containing the glycyl radical, the pyruvate formate-lyase activating enzyme (PF L-
AB) is required. 17’ ‘8 During PF L activation, S-adenos'ylmethionine (AdoMet or SAM) is
stoichiometrically converted to methionine and 5’-deoxyadenosine. When 2H-glycine
labeled PF L was activated with PF L-AE and AdoMet, a 2H was observed in the product
5’-deoxyadenosine via mass spectrometry and NMR. 19 It was proposed that PFL-AE
reductively cleaves AdoMet to form the deoxyadenosyl radical, which abstracts the
hydrogen from Gly-734 to activate PF L. '9 To further investigate this proposal, an
octapeptide mimic of the short loop containing Gly-734, was used to trigger the AdoMet
cleavage in PFL-AE. 20 A dehydroalanyl residue (AAla) took the place of Gly-734 in this
peptide. After the reaction, it was discovered by MS and 2D NMR that the
deoxyadenosyl moiety was connected to the C=C double bond of AAla (Scheme 1.3). 20
The mechanism of this reaction was proposed to involve the deoxyadenosyl radical

attacking the C=C double bond.

 

 

 

 

Suc-Arg-Val-Pro-AA1a-Tyr-Ala-Val-Arg-NH2 CH O
2
CH2 /
H2C
- + '
$554\ 1% PFL AE e if; 0H OH
H N
H
0
H 0
'00C S H
+
A
/ Methionine
1H3N H3C O
AdoMet
0H 0H

Scheme 1.3: Schematic representation of the addition of deoxyadenosyl to the AAla as a
result of AdoMet cleavage. 20

DNA sequencing showed that the E. coli PFL-AE was a 28 kDa enzyme
containing 245 amino-acid residues. 14 The PFL and PF L-AE genes are located adjacent
to each other on the E. coli chromosomal map. Overexpression of PFL-AE was initially
conducted by Kozarich and coworkers. 21 The aerobically purified enzyme was barely
soluble in H20, and thus was purified after denaturation. The protein, reconstituted with
Fe2+ and DTT after it was refolded, contained one mole Fe per mole enzyme. Broderick
and coworkers subsequently demonstrated that overexpressed and anaerobically purified
PFL-AE contained an iron-sulfur cluster. 22 The average number of Fe in one mole of
PFL-AB was 1.5, and the enzyme also had much higher activity than the form
characterized by Kozarich and coworkers. 2“ 22 UV-Vis, EPR and resonance Raman (RR)

spectroscopies showed that the enzyme contained [2Fe-ZS]2+ and [4Fe-4S]2+ clusters. 22

Hence, PFL-AE was identified as an iron-sulfur cluster protein, and a F e484 cluster was
proposed to be required for activity. Anaerobically reconstituted protein also showed
F e484 cluster signals in UV-Vis and EPR. 23

A direct measurement of the generation of the PFL glycyl radical by the [4F e-
4S]'+ cluster of PFL-AE was carried out by EPR spin quantification of the cluster and the
radical. 24 PFL-AE was photoreduced by 5-deazariboflavin, and AdoMet was added after
photoreduction. The quantity of the [4F e-4S]'+ increased with illumination time up to 30
minutes. PFL was then added to each sample, and EPR showed the presence of glycyl
radicals in amounts directly corresponding to the amount of [4F e-48]'+ originally present.
There was no [4Fe-4S]1+ cluster signal in PFL-AE after the glycyl radicals were
generated, suggesting a stoichiometric turnover of PFL by PFL-AE in a controlled

reducing environment.

1.5 PFL-AE is a radical SAM superfamily enzyme

AdoMet, a widely used substrate in biological systems, is involved in many
important biochemical reactions. 25 In various types of organisms, AdoMet is considered
as the major methyl donor in methylation reaction, 26 including the insertion of methyl
groups into DNA, hormones, neurotransmitters, phospholipids and the cobalamin
cofactors. AdoMet is also a methylene source for cyclopropane fatty acid formation in

27

eubacteria and some eukaryotic organisms, as well as for the post transcriptional

modification of tRNAs. 28 Moreover, this compound can donate its amino group. 29‘ 30

The discovery that AdoMet was essential for the activity of three different
enzymes, lysine 2,3-aminomutase (LAM), PF L-AE and anaerobic ribonucleotide
reductase (aRN R), all of which had free radical mechanisms, led to the idea that AdoMet
could also be a radical initiator. “'34 It is now believed that these enzymes uses reductive
cleavage of AdoMet to produce a 5’-deoxyadenosyl radical as a key intermediate.
Enzymes that utilize AdoMet (also called SAM) to generate the 5’-dAdo radical
intermediate (Scheme 1.4) comprise the radical SAM superfamily. 35 Although there has
not yet been any direct observation of this extremely reactive radical intermediate, a
radical on an allylic analog, S-3’,4’-anhydroadenosyl-methionine, has been detected by

EPR at 77 K in the enzymatic reaction of LAM. 36

 

. 0
H3 / o , H20 0
CH3 Radrcal SAM Enzymes ’
(6')
OH OH OH OH
+
AdoMet . .
methromne

Scheme 1.4: AdoMet is cleaved by radical SAM superfamily enzymes to generate a
deoxyadenosyl radical and a methionine.

The radical SAM superfamily is thought to be composed of more than 600
different enzymes with different functionalities. Some well-studied samples include the

37, 38

sulfur insertion enzymes biotin synthase (BioB) and lipoate synthase (LipA); 39 the

mutase lysine 2,3-aminomutase (LAM); 40

the radical activation enzymes pyruvate
formate-lyase activating enzyme (PF L-AE), 32 anaerobic ribonucleotide reductase
activating enzyme (aRNR-AB) 4' and benzylsuccinate synthase activating enzyme; 42 the
DNA repair enzyme spore photoproduct lyase (SPL); 43 and the biological cofactor
biosynthesis enzymes oxygen-independent coproporphyrinogen-III oxidase (HemN) 44
and MoaA 45 (BioB and LipA are also considered in this category).

Early work on these enzymes suggested the involvement of Fe-S clusters were
observed under different conditions, however the exact form was unclear since numerous
different cluster types ([2Fe-28]2+, [3Fe-4S]l+ and [4Fe-4S]2+) were observed under
different conditions. 22‘ 46'” Only LAM was unequivocally identified as an Fe4S4 cluster
enzyme. 50‘ 5' Iron-sulfur clusters, in synthesized complexes or in proteins, have long
been known to be remarkably versatile. 52 Distinct forms are interchangeable under
different conditions (Scheme 1.5). In the radical SAM proteins PFL-AE and BioB, EPR,
RR, UV and Mossbauer spectroscopic techniques demonstrated that the [2F e-ZS]2+ and
[3Fe-4S]l+ can be converted to [4Fe-4S]2+ under reducing conditions. 49‘ 53’ 54 In earlier
PFL-AE preparations, 66% of total irons were in the [3Fe-4S]'+ state, 12% were in the
pre-2512+ state, 8% were in the [41:e-4s12+ state and 10% in the linear [3Fe-4S]'+ state.
53 Upon dithionite reduction, 88% of total iron was in Fe4S4 clusters. 53 Further reduction

could convert [4F e-4S]2+ into [4F e-4S]1+, the active form to cleave AdoMet (Scheme 1.6).

24‘ 55 Because the cluster interconversions occurred without added iron, these studies also

suggested the conversions were a result of “cannibalization”, which meant the extra irons
and sulfurs needed to assemble the [4Fe-4S] cluster were taken from other clusters with
lower numbers of iron. More strictly anaerobic purification of PF L-AE in the presence of

1 mM DTT now results in almost exclusively [4Fe-4S]2+ clusters. 56

—2+ \

. S /
/F Fe\ 2 — s i I
8 OX S Fe

 

 

 

 

 

 

 

 

 

\\%‘Dl “— /. .4
0

O a ‘ 0+
N; A
2+

 

 

 

 

 

 

 

 

 

 

Scheme 1.5: Schematic representation of the conversions among different F e-S clusters.

'ooc
*H3N

 

Scheme 1.6: Schematic representation of AdoMet cleavage by [4Fe-4S]'+.

 

 

 

 

 

 

 

 

 

 

2+

)>

OH OH

 

 

>

OH OH

11

 

 

+ Methionine

 

OH OH

The conversions among iron-sulfur clusters in the radical SAM superfamily,
especially the one between the [3Fe-4S]1+ and [4Fe-4S]2+, suggested that there was at
least one labile Fe in the 4F e cluster. The proposed labile site was consistent with the
conserved three cysteines motif (Fig. 1.1) shown to be required for activity. 23 In PF L-AE,
Cys-29, Cys-33 and Cys-36 were proposed to bind 3 of the 4 irons in the [4Fe-4S]2+
cluster, which left one Fe not coordinated and thus more labile. The special CX3CX2C
motif is highly conserved in radical SAM superfamily enzymes (Fig 1.1), and thought to

be crucial for the catalytic mechanism of the cluster. 31‘33‘ 35

PFL-AB 24 ITF F QGCLMRCLYCHNRD
SPL 86 IPFATGCMGHCHYCYLQT
aRNR-AB 20 VLFVTGCLHKCEGCYNRS
BioB 47 SIKTGACPQDCKYCPQTS
LipA 48 MILGAICTRRCPFCDVAH
LAM 132LL1TDMCSMYCRHCTRRR
HemN 53 YFHIPFCQSMCLYCGCSI
MoaA 15 IAVTPECNLDCFFCHMEF

 

 

 

 

 

 

Figure 1.1: Conserved Fe-S cluster binding motif in radical SAM superfamily

12

1.6 AdoMet binds to the [4Fe-4S] cluster through the unique Fe site

The implication of the three-cysteine binding motif is that the unbound Fe could
be the coordination site of AdoMet as seen in the substrate binding site in aconitase. 57
Aconitase is a F e484 cluster enzyme, although not a radical SAM superfamily member,
catalyzing the reversible conversion of citrate to isocitrate. 58 Three cysteines Cys-358,
Cys-421 and Cys-424 provide the ligation of the cluster to the enzyme 59. Spectroscopic
studies and the crystal structure of aconitase show that isocitrate, the substrate, binds to

the cluster by coordination to the unique Fe site (Fig 1.2), 58 ' 6°

 

 

 

S-Cys
C00'
.000 H20 0 S/ 9/: S-Cys
/ ‘~ Sl——'Fe
\C/C / S / \

\Fe
H2 ‘ / \ S-Cys

O/C\o H20

Figure 1.2: Isocitrate binds to the unique Fe of the [4F e-4S] cluster in aconitase.

To investigate the possibility of AdoMet coordination to the cluster, 57Fe
reconstituted PFL-AE was studied by Mossbauer spectroscopy. 6' The 5 7F e was
incorporated into the unique site by adding 1 equivalent of 57Fe to PFL-AB containing
[3F e-4S]1+ under reducing conditions. In the presence of AdoMet, a significant change of
the spectral parameters of the unique site was observed, suggesting an increase in
coordination number of the unique Fe; this in turn suggested that AdoMet binds to the

cluster. Further studies of PFL-AB were carried out by electron nuclear double resonance

13

(ENDOR) spectroscopy. 56’ 62‘ 63 The carboxylate group in AdoMet was labeled with 17O
and 13C, and the or-amino group was labeled with 15N. The cluster in PFL-AE was
photoreduced to [4Fe-48]”, and the interaction between the cluster and AdoMet was
monitored. The hyperfine couplings of 17O and 13 C to the cluster were almost the same as
that observed for the carboxylate group of isocitrate coordinating to the unique Fe in
aconitase. 56 The hyperfine interaction of N to the Fe was comparable to the interaction
between the histidine N and the ferrous Fe in Rieske protein with a [2Fe-2S] cluster. 64
These results suggested that AdoMet was binding to the unique Fe via the carboxylate

and amino groups (Fig 1.3).

 

 

 

 

C , S’— Fe
I "‘Fe/ s/ \
S—Cys
CH‘N
/ H2
c/
H2 CH3
OH OH

Figure 1.3: AdoMet binds to the [4Fe-4S] cluster through the unique Fe.

Similar ENDOR results were observed in AdoMet binding to LAM. 65 RR, EPR

and Mossbauer spectroscopies were used to investigate the coordination between AdoMet

14

and the cluster in reconstituted BioB. 66 In the presence of AdoMet, the symmetric
stretching mode of the [4F e-4S]2+ cluster shifted from 338 to 342 cm", which was out of
the range (333 — 339 cm’l) for a Fe4S4 cluster coordinated by 4 exogenous cysteinyl
ligands. The Mossbauer data also indicated an increase of coordination number and/or a
noncysteinyl ligand binding to a unique Fe. The theoretical simulation of an AdoMet
interaction, similar to that in PFL-AE and LAM, fit well with the experimental data. 66
The results from different proteins suggested the binding mode between AdoMet and the
unique Fe site of the F e484 cluster is common in radical SAM superfamily members.
Recently published crystal structures of BioB (3.4 A), LAM (2.1 A), HemN (2.1
A), and MoaA (2.2 A) confirmed this type of AdoMet coordination in the superfamily. 45’
67'” All the enzymes have (IS/or) barrel or TIM barrel structures, with a (15/003 barrel for
BioB, and a % TIM barrel or (B/a)6 barrel for the other three proteins. This type of
structure has been proposed as a common feature in the radical SAM enzymes. 70 The
distances from the AdoMet carboxylate O to the unique Fe are 2.0, 2.3 and 2.5 A in
HemN, BioB and MoaA respectively. The distances between the amino N and the unique
Fe are 2.3, 2.4 and 2.6 A in the three enzymes. The distance between the Se and the
unique Fe is 3.2 A in the LAM structure crystallized with Se-AdoMet. Our crystal
structure study of PFL-AE (manuscript in preparation) shows distances from the
carboxylate O, the amino N, and the sulfonium S to the unique Fe similar to the published
structures. The noncysteine bound Fe might be the position to lock AdoMet binding and
donate one electron from the cluster to the substrate to catalyze the deoxyadenosyl radical

generation.

15

References

1. Knappe, J.; Blaschkowski, H. P.; Groebner, P.; Schrnitt, T., European Journal of
Biochemistry 1974, 50, (1), 253-63.

2. Knappe, 1.; Schacht, 1.; Moeckel, W.; Hoepner, T.; Vetter, H., Jr.; Edenharder, R.,
European Journal of Biochemistry 1969, ll, (2), 316-27.

3. Knappe, J.; Neugebauer, F. A.; Blaschkowski, H. P.; Gaenzler, M., Proceedings
of the National Academy of Sciences of the United States of America 1984, 81, (5), 1332-
5.

4. Unkrig, V.; Neugebauer, F. A.; Knappe, 1., European Journal of Biochemistry
1989, 184, (3), 723-8.

5. Wagner, A. F. V.; Frey, M.; Neugebauer, F. A.; Schaefer, W.; Knappe, 1.,
Proceedings of the National Academy of Sciences of the United States of America 1992,
89, (3), 996-1000.

6. Plaga, W.; Frank, R.; Knappe, 1., European Journal of Biochemistry 1988, 178,
(2), 445-50.

7. Knappe, 1.; Elbert, S.; Frey, M.; Wagner, A. F. V., Biochemical Society
Transactions 1993, 21, (3), 731-4.

8. Parast, C. V.; Wong, K. K.; Lewisch, S. A.; Kozarich, J. W.; Peisach, J.;
Magliozzo, R. 8., Biochemistry 1995, 34, (8), 2393-9.

9. Ulissi-DeMario, L.; Brush, E. 1.; Kozarich, 1. W., Journal of the American
Chemical Society 1991, 1 13, (11), 4341-2.

10. Brush, E. 1.; Kozarich, 1. W., Enzymes (3rd Ed.) 1992, 20, 317-403.

11. Becker, A.; Fritz-Wolf, K.; Kabsch, W.; Knappe, 1.; Schultz, S.; Wagner, A. F. V.,
Nature Structural Biology 1999, 6, (10), 969-975.

12. Mao, S. S.; Yu, G. X.; Chalfoun, D.; Stubbe, 1., Biochemistry 1992, 31, (40),
9752-9.

13. Licht, S.; Gerfen, G. J.; Stubbe, 1., Science 1996, 271, (5248), 477-81.

14. Roedel, W.; Plaga, W.; Frank, R.; Knappe, 1., European Journal of Biochemistry
1988, 177, (1), 153-8.

16

15. Conradt, H.; Hohmann-Berger, M.; Hohmann, H. P.; Blaschkowski, H. P.;
Knappe, 1., Archives of Biochemistry and Biophysics 1984, 228, (1), 133-42.

16. Becker, A.; Kabsch, W., Journal of Biological Chemistry 2002, 277, (42), 40036-
40042.

17. Knappe, J.; Schmitt, T., Biochemical and Biophysical Research Communications
1976, 71, (4), 1110-17.

18. Blaschkowski, H. P.; Neuer, G.; Ludwig-Festl, M.; Knappe, 1., European Journal
of Biochemistry 1982, 123, (3), 563-9.

19. Frey, M.; Rothe, M.; Wagner, A. F. V.; Knappe, 1., Journal of Biological
Chemistry 1994, 269, (17), 12432-7.

20. Wagner, A. F. V.; Demand, 1.; Schilling, G.; Pils, T.; Knappe, 1., Biochemical and
Biophysical Research Communications 1999, 254, (2), 306-310.

21. Wong, K. K.; Murray, B. W.; Lewisch, S. A.; Baxter, M. K.; Ridky, T. W.; Ulissi-
DeMario, L.; Kozarich, 1. W., Biochemistry 1993, 32, (51), 14102-10.

22. Broderick, 1. B.; Duderstadt, R. E.; Fernandez, D. C.; Wojtuszewski, K.; Henshaw,
T. F.; Johnson, M. K., Journal of the American Chemical Society 1997, 119, (31), 7396-
7397.

23. Kulzer, R.; Pils, T.; Kappl, R.; Huttermann, 1.; Knappe, 1., Journal of Biological
Chemistry 1998, 273, (9), 4897-4903.

24. Henshaw, T. F.; Cheek, 1.; Broderick, 1. B., Journal of the American Chemical
Society 2000, 122, (34), 8331-8332.

25. Cantoni, G. L., Annu Rev Biochem 1975, 44, 435-51.

26. Chiang, P. K.; Gordon, R. K.; Ta], 1.; Zeng, G. C .; Doctor, B. P.; Pardhasaradhi,
K.; McCann, P. P., Faseb J 1996, 10, (4), 471-80.

27. Grogan, D. W.; Cronan, 1. E., Jr., Microbiol Mol Biol Rev 1997, 61, (4), 429—41.

28. Bjork, G. R.; Ericson, 1. U.; Gustafsson, C. E.; Hagervall, T. G.; Jonsson, Y. H.;
Wikstrom, P. M., Annu Rev Biochem 1987, 56, 263-87.

29. Stoner, G. L.; Eisenberg, M. A., J Biol Chem 1975, 250, (11), 4037-43.

30. Stoner, G. L.; Eisenberg, M. A., J Biol Chem 1975, 250, (l 1), 4029-36.

17

31. Frey, P. A.; Magnusson, 0. T., Chemical Reviews (Washington, DC, United
States) 2003, 103, (6), 2129-2148.

32. Jarrett, J. T., Current Opinion in Chemical Biology 2003, 7, (2), 174-182.

33. Check, 1.; Broderick, 1. B., Journal of biological inorganic chemistry .' JBIC : a
publication of the Society of Biological Inorganic Chemistry 2001, 6, (3), 209-26.

34. F ontecave, M.; Mulliez, E.; Ollagnier-de-Choudens, S., Current Opinion in
Chemical Biology 2001, 5, (5), 506-512.

35. Sofia, H. 1.; Chen, G.; Hetzler, B. G.; Reyes-Spindola, 1. F.; Miller, N. E., Nucleic
Acids Research 2001, 29, (5), 1097-1106.

36. Magnusson, 0. T.; Reed, G. H.; Frey, P. A., Biochemistry 2001, 40, (26), 7773-82.

37. Lotierzo, M.; Bui, B. T. S.; Florentin, D.; Escalettes, F.; Marquet, A., Biochemical
Society Transactions 2005, 33, (4), 820-823.

38. Jarrett, 1. T., Arch Biochem Biophys 2005, 433, (1), 312-21.
39. Ollagnier-de Choudens, S.; Fontecave, M., FEBS Lett 1999, 453, (1-2), 25-8.

40. Frey, P. A., Book of A bstracts, 216th ACS National Meeting, Boston, August 23-
27 1998, B10L-OO7.

41. Tamarit, 1.; Gerez, C.; Meier, C.; Mulliez, E.; Trautwein, A.; F ontecave, M.,
Journal of Biological Chemistry 2000, 275, (21), 15669-15675.

42. Leuthner, B.; Leutwein, C.; Schulz, H.; Horth, P.; Haehnel, W.; Schiltz, E.;
Schagger, H.; Heider, 1., Mol Microbiol 1998, 28, (3), 615-28.

43. Cheek, 1.; Broderick, J. B., J Am Chem Soc 2002, 124, (12), 2860-1.

44. Layer, G.; Verfurth, K.; Mahlitz, E.;1ahn, D., Journal of Biological Chemistry
2002, 277, (37), 34136-34142.

45. Haenzelrnann, P.; Schindelin, H., Proceedings of the National Academy of
Sciences of the United States of America 2004, 101, (35), 12870-12875.

46. Broderick, J. B.; Henshaw, T. F .; Cheek, 1.; Wojtuszewski, K.; Smith, S. R.;
Trojan, M. R.; McGhan, R. M.; Kopf, A.; Kibbey, M.; Broderick, W. E., Biochemical
and Biophysical Research Communications 2000, 269, (2), 451-456.

47. Ollagnier, S.; Mulliez, E.; Gaillard, 1.; Eliasson, R.; Fontecave, M.; Reichard, P.,
J Biol Chem 1996, 271, (16), 9410-6.

18

48. Miller, J. R.; Busby, R. W.; Jordan, S. W.; Cheek, J.; Henshaw, T. F .; Ashley, G.
W.; Broderick, 1. B.; Cronan, 1. E., Jr.; Marletta, M. A., Biochemistry 2000, 39, (49),
15166-78.

49. Sanyal, 1.; Cohen, G.; Flint, D. H., Biochemistry 1994, 33, (12), 3625-31.

50. Song, K. B.; Frey, P. A., J Biol Chem 1991, 266, (12), 7651-5.

51. Petrovich, R. M.; Ruzicka, F. 1.; Reed, G. H.; Frey, P. A., J Biol Chem 1991, 266,
(12), 7656-60.

52. Beinert, H.; Holm, R. H.; Munck, E., Science (Washington, D. C.) 1997, 277,
(5326), 653-659.

53. Krebs, C.; Henshaw, T. P.; Check, 1.; Huynh, B. H.; Broderick, J. B., Journal of
the American Chemical Society 2000, 122, (50), 12497-12506.

54. Duin, E. C.; Lafferty, M. E.; Crouse, B. R.; Allen, R. M.; Sanyal, 1.; Flint, D. H.;
Johnson, M. K., Biochemistry 1997, 36, (39), 11811-11820.

55. Wu, W.; Booker, S.; Lieder, K. W.; Bandarian, V.; Reed, G. H.; Frey, P. A.,
Biochemistry 2000, 39, (31), 9561-70.

56. Walsby, C. 1.; Hong, W.; Broderick, W. E.; Cheek, 1.; Ortillo, D.; Broderick, 1. B.;
Hoffman, B. M., Journal of the American Chemical Society 2002, 124, (12), 3143-3151.

57. Werst, M. M.; Kennedy, M. C.; Houseman, A. L.; Beinert, H.; Hoffman, B. M.,
Biochemistry 1990, 29, (46), 10533-40.

58. Beinert, H.; Kennedy, M. C.; Stout, C. D., Chemical Reviews (Washington, D. C.)
1996, 96, (7), 2335-2373.

59. Robbins, A. H.; Stout, C. D., Proc Natl Acad Sci U S A 1989, 86, (10), 3639-43.

60. Werst, M. M.; Kennedy, M. C.; Beinert, H.; Hoffman, B. M., Biochemistry 1990,
29, (46), 10526-32.

61. Krebs, C.; Broderick, W. E.; Henshaw, T. F.; Broderick, J. B.; Huynh, B. H.,
Journal of the American Chemical Society 2002, 124, (6), 912-913.

62. Walsby, C. J.; Ortillo, D.; Broderick, W. E.; Broderick, 1. B.; Hoffman, B. M.,
Journal of the American Chemical Society 2002, 124, (38), 1 1270-1 1271 .

63. Walsby, C. 1.; Ortillo, D.; Yang, 1.; Nnyepi, M. R.; Broderick, W. E.; Hoffman, B.
M.; Broderick, J. 3., Inorganic Chemistry 2005, 44, (4), 727-741.

19

64. Gurbiel, R. J.; Doan, P. E.; Gassner, G. T.; Macke, T. 1.; Case, D. A.; Ohnishi, T.;
Fee, 1A.; Ballou, D. P.; Hoffman, B. M., Biochemistry 1996, 35, (24), 7834-45.

65. Chen, D.; Walsby, C.; Hoffman, B. M.; Frey, P. A., Journal of the American
Chemical Society 2003, 125, (39), 11788-11789.

66. Cosper, M. M.; Jameson, G. N. L.; Davydov, R.; Eidsness, M. K.; Hoffman, B.
M.; Huynh, B. H.; Johnson, M. K., Journal of the American Chemical Society 2002, 124,
(47), 14006-14007.

67. Berkovitch, F .; Nicolet, Y.; Wan, 1. T.; Jarrett, 1. T.; Drennan, C. L., Science
(Washington, DC, United States) 2004, 303, (5654), 76-80.

68. Lepore, B. W.; Ruzicka, F. J.; Frey, P. A.; Ringe, D., Proceedings of the National
Academy of Sciences of the United States of America 2005, 102, (39), 13819-13824.

69. Layer, G.; Moser, J.; Heinz, D. W.; Jahn, D.; Schubert, W.-D., EMBO Journal
2003, 22, (23), 6214-6224.

70. Nicolet, Y.; Drennan, C. L., Nucleic Acids Res 2004, 32, (13), 4015-25.

20

CHAPTER II

OVEREXPRESSION AND PURIFICATION OF PYRUVATE FORMATE-LYASE
AND PYRUVATE FORMATE-LYASE ACTIVATING ENZYME

11. 1 Introduction

Purifying proteins from organisms or tissues was once a very difficult and time-
consuming process. It took large quantities of cells or tissues to obtain a relatively small
amount of protein, since the protein of interest usually comprised less than 1% of the total
protein content in the raw material. 1 Numerous steps were required to get enough
purified protein for research. During the process, much of the target protein was lost or
degraded. It was extremely difficult to study human proteins as sufficient quantity of
human tissues or organs was always a big problem.

In the 1950’s and 1960’s, protein overexpression systems, utilizing plasmid
derived cloning vectors in bacteria, were discovered and developed. 2 In vitro studies on
proteins have been remarkably advanced by the resulting large-scale production. In order
to do protein overexpression, the gene of the target protein is inserted into a plasmid (a
circular DNA vector), containing antibiotic resistance, promoters, and regulatory
elements. A common system utilizes the Lac operator, promoter, and, which natively
regulates lactose metabolism in E. coli and some other bacteria, to control the target
protein overexpression (Fig 11.1). The target gene can is inserted behind the LAC
operation promoter, and in the absence of lactose, the Lac repressor, binds to the operator
and suppresses the transcription. When cells are grown in the presence of lactose,

allolactose (a metabolite of lactose) binds to the LAC repressor, changing its

21

conformation, and thus lowering its affinity for the lac operator. The target gene is then

transcribed and the protein can be produced at high levels. 3

lac I Promoter Operator target protein gene Terminator

»_M

lacl repressor

| I no protein expression
» _

lactose unavailable

RNA polymerase

 
   

high level of expression

allolactose or analogues available

Figure 11.1: Schematic diagram of protein expression controlled by lac promoter, operator
and repressor. Little target protein is produced in the absence of lactose. Overexpression
occurs when allolactose or analogues bind to the lac] repressor.

In practice, the allolactose analogue, isopropyl B-D-thiogalacto-pyranoside
(IPTG) is commonly used to control protein overexpression, because it can bind to the
lactose repressor like allolactose, but cannot be metabolized by E. coli (Fig 11.2).

Therefore, its concentration is constant during cell growth, and the rate and level of

protein production is sustained during protein overexpression. An additional advantage is

22

that IPTG is transported into the cell efficiently, independent of the lactose transportation

 

 

 

 

 

 

 

protein.
CH2 OH
O
H 0
\CH2
OH H
H 1
OH
H OH
H I H
Allola ctose H OH
CH2OH CH3
OH H
H I H
H OH
IPTG

Figure 11.2: Structures of allolactose and IPTG. IPTG can bind to the lac repressor, but
cannot be metabolized by the bacteria.

Once a protein is overexpressed, it generally must be isolated from all other
cellular proteins. The soluble extract from the cells is a complicated mixture of hundreds
of soluble proteins, although the overexpressed protein is ideally present at a much higher

quantity than the others. Column chromatography is the most common method used to

23

purify the protein of interest from the extract. The instrumentation is usually comprised
of pumps, an inj ecting/loading valve, a column, a spectrometer (generally UV-Vis) as the
detector, a fraction collector and a computer as the control center. Gel filtration, ion
exchange and affinity columns are widely used to purify based on the protein properties.

The gel filtration method was first developed by Porath and Flodin. 4 The so-
called gel is composed of cross-linked dextran beads, which consist of an open three-
dimensional molecular network. The bead form allows easy column packing and flow
optimization. The pores in the beads have various sizes. Small molecules can penetrate
all the pores, whereas large molecules, like proteins, can only pass through the pores
bigger than their molecular size. The accessibility of the pores to different sizes of
molecules ranges from O — 100%, with the coefficient Km, defining the proportion of the
pores that can be penetrated by a specific molecule. One expression of the coefficient is
that K“ = (V e — Vo)/(Vt — V0), where V, is the total column volume, V0 (the void volume)
is the volume outside of the bead, and V, is the elution volume. 5 Bigger molecules
occupy fewer pores, and therefore have smaller Kav values. That means large molecules,
with less elution volume, will pass through a certain column with packed beads faster
than small molecules. Thus the components in a mixture are separated by molecular size
in gel filtration.

Ion exchange columns are another widely used method for protein purification.
The net charges on proteins interact with the functional group embedded on the beads. 6
There are generally two types of ion exchangers, anion exchangers and cation exchangers.

Anion exchangers have a positively charged matrix, interacting with the negative charges

24

on the protein surface. Cation exchangers, oppositely, have a negatively charged matrix
to attract protein’s positive charges.

Diethylaminoethyl (DEAE) and quaternary amino ethyl (QAE) substituents,
connected to hydroxyl groups on the matrix, are commonly used anion exchangers. The
positive charge is on the nitrogen atom of the exchangers. There are no dissociable
protons for the quaternary amine, but in DEAE the proton on the nitrogen can leave,
which makes DEAE partially positive charged at neutral pH. During the purification, it
can act as a strong buffer, aiding the separation of proteins.

There are three classes of cation exchangers, the “weak” substituents
carboxymethyl groups, the “strong” substituents sulfonate groups, and phosphate. The
“weak” adsorbents are widely used above pH 4.5. For pH below 4, the “strong”
substituents are required, as even at low pH they will not be protonated . Phosphate
groups not only adsorb positive charges on the protein, but also interact with proteins that
bind to phosphorylated substrates such as nucleic acids. 7

The ion substituents on ion exchangers are neutralized by counterions. For
instance, Cl' is a common counterion for anion exchangers and H+ is a common
counterion for cation exchangers. The charges on proteins can replace the counterions on
the substituents, and therefore the proteins are adsorbed on the matrix under appropriate
conditions. To elute the adsorbed protein, salt gradients are usually applied. The salts can
weaken the interaction between the protein and the adsorbents, which releases the protein
into the buffer solution. Proteins with stronger adsorption are kept on the matrix longer
and eluted out later than the protein with weaker adsorption, and thus proteins are

separated by their interaction with the ion exchangers.

25

Hydrophobic interactions are also used in protein purification. 8 1n aqueous
solutions, the hydrophobic amino acid residues on protein’s surface are attracted to the
hydrophobic adsorbents on the matrix. The extent of binding of a specific protein to the
matrix is dependent on the protein, the type and density of the substituents, and the buffer
conditions. The hydrophobic adsorbents are usually short aliphatic chains (C4 to C10) or a
benzyl group connected to the matrix by hydroxyl groups. High salt concentrations in the
buffer promote the hydrophobic interaction. The separation in hydrophobic columns is

usually poorer than in ion exchange columns due to more variations in this interaction.

11.2 Materials and methods
11.2.1 Chemicals, bacterial cells and plasmids

All chemicals were commercially purchased and of the highest quality except as
noted otherwise. E. coli BL21(DE3)plysS competent cells were purchased from
NovagenTM. The plasmids pKK/pfl and pMG-AE 9 were generously donated by Dr. John
Kozarich. The pKK/pfl was amplified in our lab as previously described, '0 and the AE
gene from pMG-AE was engineered into the pCAL-n—EK expression vector (Stratagene)
to produce pCAL-n-AE, which was amplified and stored as a stock solution. '0 Both of
the pKK and pCAL-n-EK vectors carry the lac operator and ampicillin (AMP) resistance.

SDS-PAGE gels were purchased from Bio-Rad Scientific.
11.2.2 Overexpression of PF L

The plasmid pKK/pfl (1 11L) was transformed into 50 ML of BL21(DE3)plysS

competent cells. Luria Broth (LB) media (100 11L) was added to the cells thereafter. The

26

bacteria were incubated at 37 °C for 1 h. The growth was spread on 2 LB media plates
containing 50 ug/mL AMP. The cells were incubated at 37 °C overnight. One single
colony from the plates was used to inoculate 50 mL of LB media containing 50 ug/mL
AMP. The cells were grown at 37 °C overnight. This saturated culture was added into 10
L of LB media with 50 ug/mL AMP. The 10 L growth was conducted in a ferrnentor
(New Brunswick Scientific) at 37 °C, with 250 rpm stirring and 10 L/min air flow. The
optical density (0D) of the culture, read as the absorption at 600 nm in the UV-Vis
spectrometer (HP 8453), was used to monitor cell growth. When the OD600 reached 0.8,
isopropyl B-D-thiogalacto-pyranoside (IPTG) was introduced into the growth to 1 mM
final concentration. The culture was allowed to grow for another 2 h at 37 °C. The PFL
overexpression was checked by SDS-PAGE. The cells were then harvested by
centrifugation at 4 °C, 8000 rpm for 10 minutes in a Sorvall GS3 rotor. The cell pellet

was weighed and stored at -80 °C.

11.2.3 Purification of PFL

The cell pellet was lysed in a lysis buffer (1 g cell paste / 2 mL buffer) at 4 °C.
The buffer contained 20 mM HEPES, pH 7.2, 1 mM DTT, 1% (w/v) Triton X-100, 5%
(w/v) glycerol, 10 mM MgCl2, 1 mM PMSF, 0.2 mg/mL lysozyme, 0.002 mg/mL DNAse
1 and RNAse A. The suspension was stirred for 2 h until it was homogenized.

The mixture was centrifuged at 4 °C, 15,000 rpm for 30 minutes. The crude
extract (50 mL) was loaded directly onto an AP-5 column (5 x 30 cm, Waters) packed
with Accell Plus QMA anion exchange resin (Waters), and equilibrated with buffer A (20

mM HEPES, pH 7.2 and 1 mM DTT). The column was run isocratically with buffer A

27

for 60 min, followed by a linear gradient from 0% to 100% buffer B (20 mM HEPES, pH
7.2, 500 mM NaCl, and 1 mM DTT) over 180 min and isocratic 100% B for another 60
min. Fraction collection (10 mL/fraction) started at 180 min. The purification was
conducted by a BioLogic HR workstation chromatography system at 4 °C, and a flow rate
5 mL/min. The fractions were analyzed by SDS-PAGE (12% Tris-HCl gel) to identify
PFL and roughly quantify the relative amount of the protein in the fraction. The fractions
containing >~ 75% of the soluble proteins as PFL were combined and concentrated to
less than 10 mL on a Millipore YM-30 membrane using an AmiconTM concentrator.

The concentrated PFL was then purified by hydrophobic interaction
chromatography at 4°C and 1.0 mL/min. Approximately 5 mL of concentrated PFL
fractions were dialyzed against 3 L (2 x 1.5 L exchange) of buffer C (40 mM HEPES, pH
7.2, 1 M (NH4)2SO4, and 1 mM DTT) for total 4 h. After the dialysis, PFL was loaded
onto a Hi-load phenylsepharose hydrophobic interaction column (1.6 x 10 cm, Pharmacia)
equilibrated with buffer C. The column was run isocratically with 100% C for 50 min,
followed by a linear gradient to 100% buffer D (40 mM HEPES, pH 7.2 and 1 mM DTT)
over the next 50 min and an isocratic wash at 100% buffer D for 50 min. The fraction
collection (2 mL/fraction) started at 75 minutes. The fractions were analyzed by SDS-
PAGE (12% Tris-HCl gel). Those with PF L >~ 95% of total soluble proteins were pooled
together and concentrated. The concentrated protein was flash frozen in liquid N2 and

stored at -80 °C.

28

11.2.4 Overexpression of PF L-AE

Expression plasmid pCal-n-AE was transformed into 50 uL of BL21(DE3)plysS
competent cells. LB media (100 11L) was added into the cells thereafter. The bacteria
were incubated at 37 °C for 1 hour. The cells were spread on 2 LB media plates,
containing 50 ug/mL AMP, and incubated at 37 °C overnight. One single colony from the
plates was used to inoculate 50 mL of LB media containing 50 ug/mL AMP. The cells
were grown at 37 °C overnight. This saturated culture was added into 10 L of minimal
media with 50 ug/mL AMP and 34 ug/mL of chloramphenicol (CAM). 1" '2 The minimal
media contained (per 10 L) 110 g Casarnino acids, 84.2 g MOPS, 29.3 g NaCl, 8.0 g
TRICINE, 16.0 g KOH, 5.1 g N1-14Cl, 40 g glucose, 2 mM KH2P04, 0.276 mM K2804,
0.5 mM CaCl2, 9 vitamins (biotin, thiamine, riboflavin, niacinamide, pantothenic acid,
folic acid, vitamin Bl2, thioctic acid, pyridoxine, 10 mg of each), and trace amount of
minerals (Fe2+, 3033', Mn“, Co”, Cu”, Zn”, and MoO42'). “~ '2 The 10 L growth was
conducted in a fermentor (New Brunswick Scientific) at 37 °C, with 250 rpm stirring and
10 L/min air flow. When the OD600 reached 0.5, IPTG was introduced into the growth to
0.5 mM final concentration and 0.75 g Fe(NHa)2(SOa)2 was also added into the growth.
The culture was allowed to grow for another 2 h aerobically. After 2 h the culture was
cooled down to 30 °C, air flow was turned off, and N2 passed through at 10 L/min while
continuing to cool the culture. After 30 minutes, 0.75 g of Fe(NH4)2(SO.i)2 was added.
Once the temperature reached 20 °C, the fermentor was moved into a 4 °C refrigerator
and N2 was bubbled through the culture for at least 14 h at 4 °C. The cells were then
harvested by centrifugation at 4 °C, 8000 rpm for 10 minutes in a Sorvall GS3 rotor. The

cell pellet was weighed and stored at -80 °C.

29

11.2.5 Purification of PFL-AE

The entire purification of PFL-AE was carried out under strictly anaerobic
conditions in a COY anaerobic chamber (COY Laboratories, Grass Lake, MI). Cell pellet
(10 — 15 g) was suspended in 45 mL of lysis buffer, which contained 50 mM Tris-HCl,
pH 7.5, 200 mM NaCl, 5% (w/v) glycerol, 1% (w/v) Triton X-100, and 10 mM MgCl2.
The buffer was degassed thoroughly on a Schlenk line before taking into the anaerobic
chamber. To the anaerobic lysis buffer was added 39 mg DTT, 9 mg PMSF, 8 mg
lysozyme, and 0.1 mg DNAseI and RNAseA. The pellet was then added to the buffer,
and the mixture was stirred for 1 hour in an ice-water bath until it was totally
homogenized. This suspension was centrifuged at 18,000 rpm at 4 °C for 30 minutes. The
resulting crude extract (50 mL) was loaded onto an AP-5 column (5 x 60 cm, Waters)
packed with Superdex-75 (Pharmacia) gel filtration resin. If the PF L-AE was well
prepared, the crude extract would have a distinct dark brown color from the [4Fe-48F
clusters in the enzyme. The purification was conducted by an AKTA Basic 10
chromatography system (GE Healthcare).

The Superdex-75 column was pre-equilibrated with elution buffer (50 mM Tris-
HCl, pH 7.5, 200 mM NaCl, and 1 mM DTT). Elution was carried out at room
temperature with flow rate 5 mL/min. Fraction collection (10 mL/fraction) started at 2 h.
The dark brown fiactions were combined together and concentrated to about 10 mL on a
Millipore YM-10 membrane in an AmiconTM concentrator. The concentrated fractions
were flash frozen in liquid N2 and stored at —80 °C. Usually three batches of the
concentrated fractions were pooled together and loaded onto the same column again, and

purified using the same method as the first time. The UV-Vis absorbance of the dark

30

brown fractions was measured at 426 nm, indicative of the [4Fe-4S]2+ clusters, and at 280
nm, indicative of the protein concentration. Only the fractions with A426/A230 ratio over
0.19 were combined together and concentrated to less than 10 mL. The purified enzyme

was flash frozen in liquid N2 and stored at -80 °C.

11.2.6. Determination of protein concentration

The concentration of the proteins was determined by the method of Bradford. 13
The absorbance at 595 nm of Bradford dye in the presence of protein was measured as
the indicator of the protein concentration. Bovine serum albumin (BSA) was used as the
standard. According to Beer’s law, a standard curve of A595 corresponding to a series of
BSA concentrations was established. The concentration of PFL or PFL-AE was

calculated from the absorbance using the standard curve.

11.2.7. Determination of iron content in PFL-AB

Iron assays were conducted using the method of Fish or Beinert. 14’ '5 The iron in
PFL-AE were completely oxidized to Fe3+ using KMn04 (4.5% w/v) and 1.2 M HCl. The
absorbance at 562 nm of ferrozine-Fe3+ complex was measured as the indicator of the
iron concentration. Fe(N03)3 or F eCl3 solution was used as the standard. A standard
curve of A562 corresponding to the ferrozine-Fe3+ complex concentrations was established.

The iron concentration in PFL-AE was calculated from A562 using the standard curve.

31

11.3 Results and discussion
11.3.1 Overexpression and purification of PFL

The PKK/pfl plasmids were transformed into E. coli BL21(DE3)plysS strains.
Usually 10 L of growth could produce 30 — 40 g of cell pellet. PFL was well expressed in
the bacteria, determined by SDS-PAGE (Fig 11.3). PFL showed a band at the monomeric

molecular weight of 85 kDa.

PFL

 

Figure 11.3: SDS-PAGE of PFL overexpression. 1 mL of culture was taken out for the test.
Lane 1 is the protein standard. Lane 2 is the grth before IPTG induction. Lane 3 is the
growth 2 h after induction. PFL monomer band is at 85 kDa.

32

The ovexpressing cells were lysed using an enzymatic lysis procedure. The crude
extract was first purified by an anion exchange column. The collected fractions
containing PF L >~ 75% of total soluble proteins as identified by SDS-PAGE (12% Tris-
HCl gel), were combined together and concentrated (Fig 11.4). The concentrated fractions
were purified by hydrophobic interaction chromatography. The PFL content in the
collected fractions was determined by SDS-PAGE (12% Tris-HCl gel) again, and the
fractions that had >~ 95% of soluble proteins as PFL were pooled together and

concentrated (Fig 11.5). A typical 10 L growth would yield 500 — 600 mg of purified PFL.

 

Figure 11.4: SDS-PAGE of PFL obtained after anion exchange chromatography. Lane 1 is
the protein standard. Lane 2 is PFL after the anion exchange column.

33

97.000

PFL
66.000

 

Figure 11.5: SDS-PAGE of PFL obtained after hydrophobic interaction chromatography.
Lane 1 is the protein standard. Lane 2 is 0.9 ug PFL loaded onto the gel. Lane 3 is 1.4 pig
PFL loaded onto the gel
11.3.2 Overexpression and purification of PFL-AE

The PFL-AB plasmid pCal-n-AE was transformed into E. coli BL21(DE3)plysS
to overproduce the enzyme. The culture was first grown aerobically to express the protein
after IPTG induction, and then purged with N2 with cooling. The process is thought to
allow assembly of the [4Fe-4S]2+ clusters in the presence of ferrous iron. A 10 L growth
can produce 40 — 60 g of cell pellet. Afler the pellet was lysed in an anaerobic chamber
by an enzymatic lysis buffer, the crude extract was passed through a gel filtration column.
The fractions with a dark brown color, indicating the presence of iron-sulfur clusters,
were pooled together and concentrated. The fractions were then purified by the same gel
filtration column again. The fractions with A426/A230 ratio over 0.19 were combined
together and concentrated (Fig 11.6). A typical 10 L growth can produce 250 - 350 mg of

purified PFL-AB.

34

 

Figure 11.6: SDS-PAGE showing the overexpression and purification of PFL-AE. Lane 1
is the protein standard. Lane 2 is PFL-AE growth before induction. Lane 3 is PFL-AB
grth after IPTG induction. Lane 4 is the lysis of overpressing cells. Lane 5 is PFL-AB
obtained after the first gel filtration column. Lane 6 is PFL-AB obtained after the second
filtration column.

A successful growth and purification of PFL-AE has been found to produce
protein containing close to 4 mol of Fe per mol of protein (3.8 — 4.0 Fe/PFL-AE for
different high quality preparations). The state of the iron-sulfur cluster in the enzyme
depends on how strictly anaerobic conditions were maintained during the growth and
purification. The unique Fe in the [4Fe-4S]2+ is highly sensitive to O2, and exposure of
the enzyme to air can deplete the unique Fe and form a [3Fe-4S]+ cluster. 10‘ '2 Afler
purification, PFL-AB is tested by EPR spectroscopy to monitor the state of the cluster.
The [4Fe-4S]2+ cluster is EPR silent, whereas [3Fe-4S]+ cluster shows a strong and nearly
isotropic signal at g = 2.02 (Fig 11.7). ‘6 For the protein shown in Fig 11.7, the spin

quantification of the [3Fe-4S]+ cluster, conducted according to the published method, '7 is

35

362 11M in a total protein concentration of 0.89 mM. High quality PFL-AE should have

no such signal or a very small signal.

 

EPR Intensity

 

l l 1 1 J

3000 3200 3400 3600 3800 4000
Field (G)

 

 

 

Figure 11.7: X-band EPR spectrum of [3Fe-4S]+ cluster in PFL-AE. Temperature 12 K,
microwave power 1.9 mW, microwave frequency 9.36 GHz, modulation frequency 100
kHz, modulation amplitude 5 G, resolution 1024.

11. 4 Conclusions

PFL was successfully overexpressed in E. coli BL21(DE3)plysS strains. The
protein was purified initially by anion exchange chromatography and subsequently by
hydrophobic interaction chromatography. Purified PF L showed a monomeric molecular
weight band at 85 kDa in SDS-PAGE. Although PFL is only active under anaerobic
conditions, it can be prepared aerobically and is stable in the presence of air as an

inactive form.

36

PFL-AE was successfully overexpressed in E. coli BL21(DE3)plysS strains. The
protein was purified by gel filtration chromatography twice under strictly anaerobic
conditions. The enzyme has a distinct dark brown color, which is from the [4Fe-4S]2+
cluster. The absorption ratio A426/A230 was used to choose the fraction that had most
intact [4Fe-4S]2+ cluster in PFL-AB. The purified protien showed a 28 kDa band on SDS-
PAGE as the primary structure predicted '8. The best preparation of PFL-AE had close to

4 mol of Fe per mole of protein, and had no [3Fe-4S]+ cluster signal in the EPR spectrum.

37

References

1. Voet, D. V., J. G.,, Biochemistry. 2nd ed.; John Wiley & Sons: New York: 1995.

2. Winstanley, C. R., R., Protein Derived Cloning Vectors. In Molecular
Biomethods Handbook. Eds. Humana Press: Totowa, New Jersey, 1998; p pp 165-179.

3. Jacob, F.; Monod, J., JMol Biol 1961, 3, 318-56.

4. Porath, J.; F lodin, P., Nature 1959, 183, (4676), 1657-9.

5. Scopes, R., Protein Purification. 3rd ed.; Springer: 1993.

6. Muller, W., Journal of Chromatography 1990, 510, 133-144.

7. Chilla, R. ,Doering, K. M. ,,Domagk G. F. ,Rippa, M., Arch Biochem Biophys
1973, 159, (1), 235- 9.

8. Ochoa, J. L., Biochimie 1978, 60, 1-15.

9. Parast, C. V.; Wong, K. K.; Lewisch, S. A.; Kozarich, J. W.; Peisach, J.;
Magliozzo, R. 8., Biochemistry 1995, 34, (8), 2393-9.

10. Broderick, J. B.; Henshaw, T. F.; Cheek, J.; Wojtuszewski, K.; Smith, S. R.;
Trojan, M. R.; McGhan, R. M.; Kopf, A.; Kibbey, M.; Broderick, W. E., Biochemical
and Biophysical Research Communications 2000, 269, (2), 451-456.

11. Garboczi, D. N.; Hullihen, J. H.; Pedersen, P. L., J Biol Chem 1988, 263, (30),
15694-8.

12. Krebs, C.; Henshaw, T. F .; Cheek, J.; Huynh, B. H.; Broderick, J. B., Journal of
the American Chemical Society 2000, 122, (50), 12497-12506.

13. Bradford, M. M., Anal Biochem 1976, 72, 248-54.
14. Fish, W. W., Methods Enzymol 1988, 158, 357-64.
15. Beinert, H., Methods Enzymol 1978, 54, 435-45.

16. Kennedy, M. C. Kent, T. A.; Emptage, M M,erkle, H. ,Beinert, H.; Munck, E., J
Biol Chem 1984, 259, (23), 14463- 71.

17. Aasa, R.; Vanngard, T.; Dunford, H. .B., Biochim Biophys Acta 1975, 391, (2),
259-64.

38

l8. Roedel, W.; Plaga, W.; Frank, R.; Knappe, 1., European Journal of Biochemistry
1988, 177, (1), 153-8.

39

CHAPTER III

SULFUR LIGAND K-EDGE X-RAY ABSORPTION SPECTROSCOPIC STUDY OF
PYRUVATE FOMATE-LYASE ACTIVATING ENZYME

111.1 Introduction

Under anaerobic conditions PF L-AE activates PFL via reductively cleaving
AdoMet, 1'3 in which the electron required for the reduction is provided by the [4Fe-48]”
cluster in PFL-AE and a putative deoxyadenosyl radical intermediate is generated. 4 This
extremely reactive radical abstracts a pro-S hydrogen from Gly734 in PFL to yield a
glycyl radical, 4’ 5 thus activating PFL to its radical form. 6‘ 7 The cleavage of AdoMet
occurs between the sulfonium sulfur and the 5’ carbon, when one electron is transferred
from the [4Fe-4S]1+ cluster to AdoMet (Fig 111.1). One interesting questions are how the
interaction between AdoMet and the reduced cluster causes the electron transfer and
which atom in the cluster, the unique iron or the bridging sulfide, ejects the electron to
AdoMet. In another word, what is the mechanism behind the process?

Although previous evidence showed that AdoMet binds to the iron-sulfur cluster
of PFL-AE through a unique Fe site by the carboxylate O and amino N, 8'” the evidence
of the interaction between the sulfonium S and the cluster is ambiguous. In the
investigation of LAM, both the Se X-ray absorption spectroscopy (XAS) data of the
interaction between Se-AdoMet and the iron-sulfur cluster and the crystal structure of
LAM indicate that the Se of Se-Met is in the distance (2.7 A away from an Fe site in Se

XAS study and 3.2 A away fi'om the unique Fe in the crystal structure) to interact with

40

the unique Fe. ‘2‘ '3 However, Se XAS studies of Se-AdoMet in PFL-AB and BioB
showed no evidence for a nearby iron to the Se, '4 and ENDOR studies of PFL-AB
implied the sulfonium S in AdoMet had orbital overlap with the cluster, which could

occur either via a bridging sulfide or through one of the iron atoms. ”

1+
Cys—S

 

 

Methionine

Figure 111.1: The cleava e of AdoMet by radical SAM enzymes. One electron is
transferred from [4Fe-48] + to sulfonium S. A deoxyadenosyl radical and a methionine
are generated.

To illustrate the interaction between the sulfonium S and the cluster, a relatively
new technique sulfur ligand K-edge XAS was used. This method is able to elucidate the

covalency of iron-sulfur bonds, whose quantification demonstrates the exchange coupling

between Fe 3d orbitals and S 3p orbitals, in Fe-S clusters. 15'” Upon AdoMet binding, the

41

electronic property of the Fe-S bond in the cluster could be disturbed accordingly, which
is partly due to the interaction between the sulfonium sulfur and the cluster. The
calculation based on the change of the covalency in the absence and presence of AdoMet
might illustrate the distance between the S+ to the atoms in the cluster, and the orbital
coupling between AdoMet and the cluster, which implies the mechanism of electron
transfer from the cluster to the substrate.

A high-energy X-ray photon can be absorbed by an atom in a sample if the
photon’s energy matches the binding energy of an inner shell electron. The electron can
be ejected into a vacant molecular orbital, which is related to the molecule’s electronic
structure and geometry. Hence, distinct absorptions provide information about the bonds
surrounding the atom. A typical XAS spectrum shows a sharp rise in absorbance, a so-
called edge, when the photon is absorbed. The absorption is sensitive to the electronic
structure of the atom and demonstrates special features in the edge region. A scan of the
X-ray photon energy range can provide several absorption edges for one atom as each
photon energy is equal to the ionization potential of the inner shell electron. The
different edges are called K, L, M, etc, whose names are derived from the shells of the
Bohr atom (K edge for n = 1, L edge for n = 2, M edge for n = 3). The K-edge absorption
of sulfur is at 2.5 keV.

The ligand—metal bond covalency can be estimated from a pre K-edge feature in
the ligand K-edge X-ray absorption spectroscopy. '8 The ligand K-edge absorption is due
to the ligand ls 9 4p transition. The ligand 4p orbitals can, however, covalently overlap
with the unoccupied metal 3d antibonding orbitals, and the ligand filled ls electrons are

able to transit into the metal 3d orbitals, due to the electric dipole allowed ligand Is 9 4p

42

transition (Fig 111.2). The intensity of the Is -) 3d forbidden transition, which occurs on
the pre-edge region, depends on how much the ligand covalent character overlaps with
the metal unfilled antibonding 3d orbital. Therefore, the intensity of the pre-edge feature
can quantify the ligand—metal bond covalency. The energy position of the pre-edge
transition is determined by both the shift of ligand ls orbital energy and the energy of the
lowest unoccupied molecular orbital (LUMO) of the complex. The shift of ligand ls
orbital energy is related to the effective charge on the nucleus (Zen) of the ligand. In
addition, the Is orbital shifts to lower energy due to the ligand electron density
delocalization onto the metal in the complex bonds. Two primary factors contribute to the
complex LUMO energy: the geometry of the ligand field affects the d orbital energy
splitting, and the overall (1 energy can shift too, which is related to the coordination
number and the Zen of the metal.

Ligand 4p orbita1 T—

; “4‘ Ligand field splitting

 

 

 

‘I
- ~
‘s- -5

---
-~
Q

 

Ligand K-edge
transiti on Ligand
3p orbital

/
/// /

Ligand pre
K-edge transition

 

 

Ligand ls orbital 4

 

Figure 111.2: Schematic representation of the ligand pre K-edge transition. The energy of
the transition is determined by both the ligand ls orbital and the metal unoccupied 3d
antibonding orbital.

43

Sulfur K-edge XAS has been applied to investigate the characteristics of the Fe-S
bonds and electron transfer in the active sites of various iron-sulfur proteins and their
model complexes, such as rubredoxin, ‘9 F e2S2 ferredoxin, 2° F e384 ferredoxin, 2' Fe484
ferredoxin and high potential iron-sulfur protein (HIPIP). '6 In those XAS spectra, two
major pre-edge features between 2469 eV and 2471 eV are deconvoluted from the
absorption of each protein, except for rubredoxin which only has one pre-edge feature.
The lower energy transition is from the bridging sulfide ls -) Fe 3d antibonding orbital,
and the higher energy absorption is the transition between the terminal thiolate ls -> Fe
3d antibonding orbital. The primary edge rises afier 2472 eV, which is assigned to the
thiolate Is to C—S 0. transition. The covalency of the Fe-S bond is calculated from the
pre-edge transition intensity.

Although the ligand pre K-edge XAS monitors the electron transfer from the
ligand to the metal, its physical principles are different from ligand to metal charge
transfer (LMCT). LMCT involves the transition of an electron fi'om a complex bonding
molecular orbital (MO) with primary ligand character to an antibonding MO with
predominantly metal character. The electric dipole-allowed transfer shows a broad and
intense absorption band in the UV-Vis region. In the ligand pre-edge XAS, the electron
donor has only ligand character since the ligand ls orbital has no overlap with the metal.
Its transition dipole moment is oriented on the ligand alone, while the transition dipole
moment in LMCT is mixed between the ligand and metal. Therefore, the LMCT is bond
length dependent, but the ligand pre-edge XAS is not.

Another traditional method to investigate the ligand—metal bond and the

coordination environment is EPR. EPR involves the spectroscopic detection of species

44

that have unpaired electrons. These species are either radicals or transition metal ions
coordinated in complexes. A single electron has a magnetic moment because of its
electronic spin. If the electron is surrounded by an external magnetic field, this magnetic
moment can be aligned parallel or antiparallel to the external field, with parallel state
having lower energy than the antiparallel state (Zeeman effect). The electron can
resonantly absorb a photon, if the photon energy matches the separation between the
parallel and antiparallel states (Fig 111.3). This energy separation is AB = hv = geuBB,
where g, is called g factor, which is the ratio between the electron magnetic dipole
momentum and the angular momentum, up is the Bohr magneton, and the B is the
external magnetic field. The g factor for a free electron in vacuum is 2.002319304386;
the g factor can, however, deviate from this number if the environment changes. The
actual g factor is, therefore, one indicator of the electronic structure surrounding the
unpaired electron. The EPR signal is generally shown as a first derivative of the
resonance absorbance, and the shape of the signal is frequently reflective of the geometry
of the coordination environment. To enhance the resonance, EPR is usually measured at

low temperature, due to the more favorable Boltzmann distribution and slower relaxation

E,//////////’//’ +l/2

AB = hV = E172 ‘ E-I’Z

times.

-1/2

Figure 111.3: Schematic representation of the energy separation of an electron magnetic
moment placed in an external magnetic field.

45

The Fe484 cluster has been considered as a pair of two Fe2S2 subunits. 22 The
electrons are delocalized between two irons in each subunit. 23 The F e3+Fe2+ dimer has
one electron shared, so the overall status of the subunit is two Fez'SI. The delocalized
electron keeps the same spin state when it is in the presence of each iron site, and aligns
antiparallel to the d5 core electrons on the metal. Thus the spins of the d5 electrons on
each Fe3+ are aligned parallel, which is called ferromagnetic coupling (Fig 111.4). The
total spin of the F e2S2 is S = +9/2. In the Fe484 cluster, the two Fe2S2 subunits are coupled
antiferromagneticlly, meaning the delocalized electrons are aligned antiparallel. The
[4F e-4S]2+ has a total spin S = 0 and is EPR silent, since each F e2S2 cluster has a spin S =
9/2. In the reduced [4Fe-4S]1+ cluster, one more electron is delocalized in one of the
Fe2S2 subunits, and makes the overall spin of the Fe484 cluster S = + 1/2. This cluster is

observable in EPR.

Site a

m
.3“

(D

6

3+
Fe

'11

(D
"N
.1,

+++++
+++++

l

Ferromagnetic coupling in a F e282 subunit

Figure 111.4: Schematic representation of an electron delocalization in a F ez'5 F ez‘5 dimer.
The spin state of the unit is S = +9/2.

46

111.2 Materials and methods
111.2.1 Chemicals and materials

All chemicals were commercially purchased and of the highest quality except as
noted otherwise. E. coli BL21 (DE3)plysS competent cells were purchased from
NovagenTM. The pCAL-n-EK expression vector was purchased from Stratagene. pMG-
AE plasmid was inserted into the expression vector to produce the pCAL-n-AE plasmid.

SDS-PAGE gels were purchased from Bio-Rad Scientific.

111.2.2 Overexpression and purification of PFL-AE

Growth and purification of PFL-AE were carried out following the procedure
previously described in 11.2.4 and 11.2.5, except the modification that the reducing agent
in the purification buffer was 1 mM ascorbic acid instead of 1 mM DTT. Because the
sulfur XAS measures the absorbance from all types of S, DTT in the buffer would likely

have interfere with the XAS results and thus was avoided.

111.2.3 Synthesis of S-adenosylmethionine

AdoMet was synthesized in a reaction catalyzed by SAM synthetase (scheme
111.1). A crude extract of cells expressing SAM synthetase was used in the reaction. The
reaction was carried out in 10 mL of 100 mM Tris-HCl pH 8.0 buffer, containing 50 mM
KCl, 26 mM MgCl2, 13 mM adenosine triphosphate (ATP), 8% (v/v) B-mercaptoethanol,
1 mM EDTA, 10 mM methionine, 2.5 11L of inorganic pyrophosphatase (0.25 U) and 1
mL of SAM synthetase crude lysate. It was stirred at room temperature for 16 hours and

terminated by adding 1 mL of 1 M HCl. The solution was centrifuged at 18,000 rpm, 4 °C

47

for 30 minutes to remove the proteins. The clear solution (6 mL) was loaded onto a
Source 158 cation exchange column (Pharmacia, 1 cm x 10 cm), precharged with 1 M
HCl and equilibrated with H2O. The purification of AdoMet was conducted at room
temperature with flow rate 2 mL/min. The column was run isocratically at 100% A (MQ
H20) for 14 min, followed by a linear gradient to 10% B (1 mM HCl)over 5 min and then
isocratically at 10% B for 8 min. A linear gradient from 10% to 100% B over 60 min was
used to elute the AdoMet. The column was then washed with 100% B for 10 min
followed by re-equilibration with buffer A. The eluate between 40% and 60% B was
collected. The fractions were lyophilized on a Shlenk line and weighed after being
completely dried. The solid was then dissolved in 50 mM Tris-HCl pH 7.5 in the

anaerobic chamber (Mbraun). NaOH (1 M) was used to adjust the pH to near 7.0.

SAM synthetase
Methionine + ATP > S-adenosylmethionine
stir, 16 hours

 

Scheme 111.1: Schematic representation of AdoMet synthesis catalyzed by SAM
synthetase.

111.2.4 Sample preparation for sulfur XAS
All procedures were carried out in the anaerobic glove box (Mbraun) with 02
levels under 1 ppm. PFL-AE (300 11L of 3.4 mM) was mixed with 50 11L of 70 mM

AdoMet to make the final concentration of PFL-AE 2.9 mM and AdoMet 10 mM. PFL-

48

AB (300 11L of 3.4 mM) alone as well as the prepared PFL-AE and AdoMet mix were

flash frozen in liquid N2 in the box and stored at —80 °C.

111.2.5 PFL-AE photo-reduction

All procedures were carried out in EPR tubes in the anaerobic glove box
(Mbraun). Photoreduction was carried out in two identical samples, 300 11L each,
containing 100 mM Tris-HCl pH 7.5, 100 mM KCl, 50 11M PFL-AE, 100 11M 5-
deazariboflavin (the photo reducing agent), and 100 11M dithionite. A 300 W halogen
lamp was put 5 cm away from the sample to conduct the photo reduction. The EPR tubes
were placed in an ice-water bath (4 °C) to prevent over heating the samples. After
illumination, one sample was flash frozen in liquid N2. AdoMet (1.6 11L, 55 mM) was
added into the other sample to make the final AdoMet concentration 150 11M. The sample
was then flash frozen in liquid N2.

Another set of samples was photoreduced following the same procedure above.
The samples contained either 10 mM DTT or 10 mM cysteine in addition during the
illumination, after which AdoMet (final concentration 150 11M) was added into one of the
two identical samples.

The buffer (300 11L of 100 mM Tris-HCl, 100 mM KCl, pH 7.5) was illuminated

at the same condition as the blank sample for the background measurement in EPR.

111.2.6 UV-Vis spectroscopy

UV-Vis spectra were recorded on an HP-8453 diode-array spectrometer. PFL-AE

was diluted in 100 mM Tris-HCI, 100 mM KCl, pH 7.5 to final concentration 100 11M in

49

the anaerobic glove box (Mbraun). The diluted samples (300 11L) were loaded into a
quartz cuvette with 1 mm optical path length. The cuvette was sealed with a rubber
septum in the box. The absorbance between 200 and 1100 nm was measured in the
spectrometer at room temperature. During the measurement, a nitrogen flow was purging

from a funnel on top of the cuvette.

111.2.7 EPR spectroscopy

EPR spectra were recorded in a Bruker ER-200D-SRC spectrometer at 12 K. X-
band microwave frequency 9.37 GHz was used. The microwave power was 1.9 mW. The
external magnetic field was scanned from 2000 to 4000 G. The static field was 2000 G.
The modulation frequency was 100 kHz, and modulation amplitude was 5 G. Two scans

were accumulated.

111.2.8 Sulfur K-edge XAS

The XAS spectra were recorded at the Stanford Synchrotron Radiation Laboratory
under ring conditions of 3 GeV and 60-100 mA. The experiments were performed on the
54-pole wiggler beam line 6-2 operating in high field (10 kG) mode with a fully tuned
Si(l 11) double crystal monochromator followed by a Ni-coated harmonic rejection
mirror. The X-ray beam was optimized for low-energy studies, with all optical elements
under a differentially-pumping maintained vacuum, and with a single, in-hutch 127-um
beryllium window separating the experiment from ring vacuum. This Be window is
protected from oxidation by a ~4” He gas path terminated with a 6.35-um polypropylene

window. The beam thereafier passes in a He gas atmosphere through aperturing slits, an

50

ionization chamber for 10 measurements, and a 6.35-1tm polypropylene window into the

sample compartment (Fig 111.5).

Inside the experimental hutch

 

 

 

   

 

 

 

 

 

   

 

I
I
6 independently adjustable O |
sample chamber positioner l
:- (é.- IIIIIIIIIII .12).: “xi”. I
phosphorous - sampIe - flexible tubing I
,e,..." : box with ; tubing I
forboam . rotating lid

augment -_- §H° 9”: Ho gas I

\x _ ; e L a - Beam

ppw . line

second ionization | OptICS
ionizaton chamber vertical and I
chamber for incident horizontal I
Z 2 . t Magoo) hutch slits I
in ensr I
g ................ o I
I

 

 

motorized optical rail (vert./horz./tiltlskew)

Figure 111.5: Schematic representation of the experimental setup of X-ray absorption
spectrometer (ppw: polypropylene window, Alw: aluminized Mylar window, Bew:
Beryllium window, EY: electron yield , FF: fluorescence). 24

For fluorescence measurements, a N2-filled Lytle detector is used without filter or
Soller slits. Due to the high absorption at low energies (2-4 keV), external energy
calibration using Na2S203-5H2O as the standard was performed before and afier each
sample. The sharp maximum of the first K-edge feature is assigned at 2472.02 eV. The
monochromator step size over the edge region was 0.08 eV, the energy resolution was
~0.5 eV, and the reproducibility in energy determination was ~0.1 eV, as determined
from repeated calibration and sample measurements during separate experimental runs.

Protein samples (~300 11L) were transferred into a Pt-coated Al block sample cell,

sealed with a polypropylene window on the front and with septum-sealed loading and

51

gas-release holes on the top. The samples were maintained at ~ 4 °C by passing pre-
cooled N2 through the sample cell block.

Pre- and post-data calibration scans were compared and the energy scale of the
data is adjusted accordingly. The acceptable data scans were then averaged, and a smooth
second-order polynomial background was fit to the pre-edge region and subtracted from
the entire spectrum. Normalization of the data was accomplished by fitting a flattened
second-order polynomial or straight line to the post-edge region and normalizing to an
edge jump of 1.0 at 2490 eV. The protein spectra were re-normalized by the ratio of the
total sulfurs (including the cysteines and methionines) and the contributing absorbers for
the covalency determinations. Energy positions of features in the normalized spectrum

were obtained from the minima in the second derivative of the data.

111.3 Results and discussion
111.3.1 Overexpression and purification of PF L-AE

PFL-AB has been successfully overexpressed in E. coli strains BL21(DE3)plysS.
After purification in the presence of 1 mM ascorbic acid, the enzyme showed a distinct
dark brown color. The protein was over 95% pure as determined by densitometry of an
SDS-PAGE gel (Fig 111.6). The concentration of the protein was 3.4 mM and one mole of

PF L-AE contained 3.7 mole of Fe.

52

31.000

21,500

 

Figure 111.6: SDS-PAGE gel of PFL-AE purified in the presence of 1 mM ascorbic acid.
Lane 1: E. coli cells overexpressing PFL-AE after IPTG induction; Lane 2: Soluble lysate
of cells expressing PFL-AE. Lane 3: PFL-AE after lSt run over gel filtration column;
Lane 4: PFL-AE after 2"Cl run over gel filtration column; Lane 5: Protein standard.

111.3.2 Synthesis of S-adenosylmethionine

AdoMet synthesis and purification followed the procedure described in the
experimental methods. AdoMet was eluted out from the cation exchange column as a
broad and intense peak between 40% and 60% B monitored by UV absorbance at 260 nm
(Fig 111.7). After lyophilization, AdoMet was a colorless solid. A total of 36 mg of

product was obtained, corresponding to a 72% yield. Since AdoMet eluted between 0.4

53

and 0.6 M HCl, the solution made from the solid dissolved in the buffer was still acidic,

and thus neutralization using NaOH (1M) was necessary.

 

     

 

 

 

 

 

 

 

 

 

 

 

, 500.0

2 0° :— 100% buffer a ;
1_76 :1. % B $450.0

1 E4000
150-" E

3 AdoMet r3600
1251, Z-300.0
11101 325010

:_ j-zooio
0:757 conductlvlty E

:_ ,1500
0505 E

1 7100.0
0251 950.0

: t
0.00-‘ ; on

i E--50.0
0.25: r

‘ r r r v I v . v , r

0 1 2

AU Hours mSIcm

Figure 111.7: Chromatography of AdoMet elution from the Source 158 cation exchange
column. The AdoMet peak eluted between 0.4 and 0.6 M HCl.

111.3.3 UV-Vis spectra of PFL-AE purified in ascorbic acid and DTT

UV-Vis spectra of the purified PFL-AB showed a broad absorption band, with a
shoulder from 410 nm (Fig 111.8). This type of feature is characteristic of iron-sulfur
cluster proteins. 25‘ 26 The spectrum of PFL-AB purified in the presence of 1 mM ascorbic
acid was almost the identical to that of PFL-AB purified in the presence of 1 mM DTT,

suggesting that the change in reductant did not significantly effect cluster composition.

54

 

 

20 -

A 1 Ascorbic Acid
“1‘5 :
”3% 15'n
T." :
10 —

5 : \\
: \E

 

vrjfir *1

300' . 350 I 1460' I I 145.0. 1"5110‘ H 550
Wavelength (nm)

Figure 111.8: UV-Vis spectra of PFL-AB purified in ascorbic acid and DT T. The protein
was 100 11M in a quartz cuvette with 1 mm optical path length. The spectra were recorded

under anaerobic condition at room temperature.

To test the possibility that excess thiols in the buffer might bind to the unique Fe
of the Fe4S4 cluster in PFL-AE, the UV-Vis spectra of the protein in the presence of 10
mM DTT or 10 mM cysteine were also recorded. The reason to conduct this experiment
is to compare PFL-AE with Fe4S4 cluster model compounds and other Fe484 cluster
proteins (each Fe was coordinated by 4 sulfurs), whose S pre-edge spectra were published
previously. 16’ 17 The difference spectra (PFL-AE/ 10 mM RSH — PFL-AB) showed only
slight changes upon adding excess thiols (Fig 111.9), while the EPR spectra of reduced

PFL-AE and reduced PFL-AE with excess thiols are different (shown in 111.3.4),

55

indicating that UV-Vis absorption spectroscopy may not be sensitive enough to detect the

Fe484 cluster structural change under the conditions.

 
   
   
   
    

 

 

 

_A 05 A
E
-9 4
25 s ‘1.
, 5M;
PFL—AE + 10 mM DTT 3 1
20; 3011350400430 sinsédm
4 Wavelength(nm)
t-i“ ‘ A
Is , d2 0 B
FIG 15 d _.9
l j s 0
E I go:
“* dfihfihfifi"
_ PFL-AE + 10 mM C s
10 . y Wavelength(nm)
5 j

 

 

mvashuhtéoTsthsdo
Wavelength(nm)

Figure 111.9: UV-Vis spectra of PFL-AE with exogenous thiols (10 mM DTT or 10 mM
Cys). The protein was purified in the presence of 1 mM ascorbic acid. Protein
concentrations were 100 11M and RSH concentrations were 10 mM. The inset spectra are
difference spectra of PF L-AE/RSH - PF L-AE (purified in the presence of ascorbic acid).
A. RSH = DTT. B. RSH = Cys. All the spectra were recorded under anaerobic conditions
at room temperature.

111.3.4 EPR spectra of photo-reduced PF L-AE
The reduced [4Fe-4S]1+ cluster (S = +1/2) of PFL-AE purified in the presence of 1
mM ascorbic acid had a rhombic EPR signal (g = 2.01, 1.92 and 1.87, Fig III.10.A). The

shape of the signal and the g values were very similar to the published EPR spectrum (g =

56

2.02, 1.94, 1.88) of PFL-AE purified in the presence of 1 mM DTT. '0 For both the DTT
and ascorbic acid purified proteins, the presence of AdoMet resulted in a dramatic change
to a more rhombic signal (g = 2.00, 1.92 and 1.89, Fig III.10.B). This effect demonstrates
a significant perturbation of the cluster electronic environment by AdoMet, consistent
with the direct interaction between the substrate and the active site previously
demonstrated. 8'”

When the protein was photoreduced in the presence of exogenous thiol (10 mM
DTT or 10 mM cysteine), the [4Fe-48]” cluster exhibited a more axial EPR signal
compared to the protein in the absence of thiol (Fig 111.11). The axial signal indicates the
cluster has a more symmetric electronic environment, implying that the exogenous thiol
might coordinate to the unique Fe. In addition, there were subtle differences in the g
values and line shape depending on whether PF L-AE was reduced in the presence of
DTT (g = 2.03, 1.93) or cysteine (g = 2.00, 1.93). These slight difference are consistent
with the possibility of DTT and cysteine coordinating to the cluster, as they are distinct
molecules with different electron density on the sulfurs.

To compare PFL-AE purified in the presence of DTT with PFL-AE purified in the
presence of ascorbic acid, the former was photoreduced in the presence of 10 mM
cysteine or DTT. The EPR spectra and g values (g = 2.03, 1.93 in the presence of 10 mM
DTT and g = 2.00, 1.93 in the presence of 10 mM cysteine) were very similar to the
protein purified in the presence of ascorbic acid (Fig 111.11). The results further
demonstrate that the electronic structure of PFL-AE/[4Fe-4S]1+ purified in ascorbic acid

is similar to that purified in DTT.

57

 

 

 

 

 

 

 

 

A
.... B
m
E
9
E
m
9-1
H __
3000 3200 3400 3600 3800 4000

Field (gauss)

Figure 111.10: X-band EPR spectra of reduced PFL-AE and reduced PFL-AE with 3
equivalents of AdoMet added. The protein was purified in 1 mM ascorbic acid. The
buffer in both A and B contained 100 mM Tris-HCl pH 7.5, 100 mM KCl, 50 11M 5-
deazariboflavin and 100 11M dithionite. A. PFL-AE (50 11M) was photoreduced for 1 h. B.
PFL-AE (50 11M) with 150 11M AdoMet added afier photoreduction for 1h. EPR
parameters: T = 12 K, frequency 9.37 GHz, microwave power 1.9 mW, magnetic field
scanned from 2000 to 4000 G, static field 2000 G, modulation frequency 100 kHz,
modulation amplitude 5 G, scan twice.

58

 

 

 

 

 

 

A
’5
=
0
fl
.5
2‘.
a B

3000 32200 3400 3600 3800 4000

Field (gauss)

Figure 111.11: X-band EPR spectra of PFL-AE (purified in the presence of 1 mM
ascorbic acid) photoreduced in the presence of exogenous thiols. The buffer in both A
and B contained 100 mM Tris-HCl pH 7.5, 100 mM KCl, 50 11M 5-deazariboflavin and
100 11M dithionite. A. PFL-AE (50 11M ) photoreduced for l h with 10 mM DTT. B.
PFL-AE (50 11M) photoreduced for 1h with 10 mM cysteine. EPR parameters: T = 12 K,
frequency 9.37 GHz, microwave power 1.9 mW, magnetic field scanned from 2000 to
4000 G, static field 2000 G, modulation frequency 100 kHz, modulation amplitude 5 G,
scan twice.

59

 

 

EPR Intensi

 

 

 

 

l L 1 l

3000 3200 3400 3600 3800 4000
Field (gauss)

 

Figure 111.12: X-band EPR spectra of PFL-AE (purified in 1 mM DTT) and photoreduced
with exogenous thiol. The buffer in both A and B contained 100 mM Tris-HCl pH 7.5,
100 mM KCl, 50 11M 5-deazariboflavin and 100 uM dithionite. A. PFL-AE (50 11M )
photoreduced for 1 h with 10 mM DTT. B. PF L-AE (50 11M) photoreduced for 1h with
10 mM cysteine. EPR parameters: T = 12 K, fi'equency 9.37 GHz, microwave power 1.9
mW, magnetic field scanned from 2000 to 4000 G, static field 2000 G, modulation
frequency 100 kHz, modulation amplitude 5 G, scan twice.

For more comparisons between the PFL-AE purified in the presence of ascorbic

acid and the PFL-AE purified in the presence of DTT, AdoMet (3 equivalents) was added

60

after the PF L-AE was photoreduced in the presence of exogenous thiol. The spectra of
the protein purified in the presence of ascorbic acid (Fig 111.13.A and Fig III.13.B) are
almost identical to the spectra of the protein purified in the presence of DTT (Fig III.13.C
and Fig 111.13.D). The results demonstrate that AdoMet binds to PF L-AE the same way
in all cases, regardless of the reductant (DTT or ascorbic acid) used during purification,
and regardless of the presence of excess thiols in the sample. Together with the UV-vis
results and the EPR spectra of the proteins without AdoMet, these results demonstrate
that the F e4S4 cluster of PFL-AE purified in the presence of ascorbic acid has a similar
electronic structure as that of PFL-AE purified in the presence of DTT. Thus, the sulfur
K-edge XAS investigation of PFL-AE purified in the presence of ascorbic acid (described
in the next section) is directly relevant to previous studies of PFL-AE purified in the
presence of DTT.

The spectra in Fig 111.13 are distinct from those in Fig 111.11 and Fig 111.12,
indicating that the [4Fe-4S]1+ cluster in PFL-AE has a different electronic structure upon
AdoMet binding. The results are consistent with AdoMet binding to the cluster,
suggesting that AdoMet might replace an exogenous thiol ligand, even though the thiol
(10 mM) was in a much higher concentration than AdoMet (150 11M). The spectra in Fig
111.13 are similar to the spectrum in Fig III.10.B, which was photoreduced PFL-AE
(without 10 mM thiols) with AdoMet added after illumination. The results demonstrate
that the [4Fe-48]” cluster in PFL-AE has the same electronic structure in the presence of
AdoMet, regardless of the presence of large excesses of thiols. These results suggest that
AdoMet may be able to coordinate to the unique iron of the cluster, even though some

other ligands in large excess possibly bind to that iron in the absence of AdoMet.

61

 

 

A C
W
D

3000 3200 3400 3600 3300 4000 3000 3200 3400 3600 3800 4000
Field (831135) Field (gauss)

Cd
EPR Intensity

 

 

 

 

 

 

Figure 111.13: X-band EPR spectra of photoreduced PFL-AB in the presence of 10 mM
exogenous thiol with 3 equivalents of AdoMet added after illumination. The PFL-AE in
panel A and B was purified in the presence of 1 mM ascorbic acid. The PFL-AE in panel
C and D was purified in the presence of 1 mM DTT. The buffer contained 100 mM Tris-
HCl pH 7.5, 100 mM KCl, 50 M 5-deazariboflavin and 100 11M dithionite. AdoMet
(150 uM) was added into each sample after the PFL-AB was photoreduced for 1 h. A.
PFL-AB (50 11M, 1 mM ascorbic acid) photoreduced with 10 mM DTT + AdoMet (150
uM). B. PFL-AE (50 11M, 1 mM ascorbic acid) with 10 mM cysteine + AdoMet (150
11M). C. PFL-AE (50 11M, 1 mM DTT) with 10 mM DTT + AdoMet (150 11M). D. PFL-
AE (50 1.1M, 1 mM DTT) with 10 mM cysteine + AdoMet (150 11M). EPR parameters: T
= 12 K, frequency 9.37 GHz, microwave power 1.9 mW, magnetic field scanned from
2000 to 4000 G, static field 2000 G, modulation frequency 100 kHz, modulation
amplitude 5 G, and scan twice.

62

111.3.5 Sulfur K-edge XAS spectra of PF L-AE and PF L-AE with AdoMet

The S K-edge XAS data on PFL-AB (Fig. III.14.A) show an intense pre-edge
transition at ~2470 eV characteristic of a [4Fe-4S]2+ cluster. ‘7‘ 24 This broad feature
envelops transitions to all the un-occupied Fe3d based orbitals from the Is orbitals of four
u3Ssuifid, and three Simona, ligands. The rising-edge (at 2473.5 eV) is comprised of S Is to
OS 0* transitions of all bound and unbound cysteines and methionines present in the
protein. The data for the protein upon AdoMet binding (Fig. 111.14.A, dotted line) show
an increase in the intensity of the pre-edge feature as well as the rising-edge feature.
While the increase in the rising-edge features is expected as there are additional 3
equivalents of AdoMet (AdoMet has three C-S bonds), the increase in pre-edge feature
must result from AdoMet binding to the cluster as AdoMet has no feature at 2470 eV (Fig.
111.14.A, grey line). This increase is demonstrated in Fig. III.14.B, where the
renorrnalized difference spectrum shows a clear peak at 2470.3 eV. This increase in the
pre-edge intensity could originate from a change of cluster ligation, AdoMet interaction
with u3Ssumde, or other effects. To evaluate these effects, geometry optimized DFT
calculations were used.

The calculated covalency shows that substituting a OH' ligand (a proposed
exogenous ligand to the unique Fe in the absence of AdoMet) on the cluster with AdoMet
results in a 37% increase in S3,, character in the Few manifold (Table 111.1, row 1 and 2).
This increase does not occur when substituting OH' by a glycine or a methionine bound
form (Table 111.1, rows 3 and 4). This computational result is consistent with the
increased experimental observation of pre-edge intensity upon AdoMet binding. One

interesting result is that 25% of the 37% increase is derived from the Fe-thiolate

63

covalency (Table 111.1, rows 1 and 2, column 2) although AdoMet is proposed to have

more interaction with the Fe-Sb bonds than with the Fe-St bonds.

3. A
:27, ——PFL—AE _' ‘,
,3, 2.4 - """" PFL—AE + AdoMet
gm - ——AdoMet (310.4)
§ 1.8 -
151.5 -
51-2 ‘ pie-«U9
20.9 -
o 0.6 '1
5 0.3 7
o

2467 2469 2471 2473 2475 2477
Energy (eV)

C-S'

   
 
  
 
  

 

 

    
   
 

0.3 -
0.1 -

1.7- —PFL-AE
§ 1.5- -----PFL-AE+AdoMet
E 1.3 . —mrrerenee -'
g 1...
u 0.9.
5 0.7-
E 0.5- 2470.3 eV
0
C
O
I!

 

f T — I I I I I I

 

'0'54‘67 24 68 2469 2470 2471 24 72 2473 2474
Energy(eV)

Figure 111.14. Sulfur K-edge XAS spectra of PFL-AE, PF L-AE + AdoMet and AdoMet
(A) and difference spectrum PFL-AE/AdoMet — PFL-AE (B). A. Black solid line is the
absorbance of PF L-AE (3.4 mM). Black dashed line is the absorbance of PFL-AE (2.9
mM)+ AdoMet (10 mM). The grey line is the absorbance of AdoMet (20mM). B. The
black solid line is the absorbance of PFL-AE (3.4 mM). The black dashed line is the
absorbance of PFL-AE (2.9 mM) + AdoMet (10 mM). The grey line is the difference
spectrum PFL-AE/AdoMet — PF L-AE. XAS parameters: The XAS spectra were recorded
in Stanford Synchrotron Radiation Laboratory under ring conditions of 3 GeV and 60-
100 mA at 4 °C. The X-ray beam was optimized for low-energy studies. The
monochromator step size over the edge region was 0.08 eV, and the energy resolution
was ~0.5 eV. Solid Na2S2O3-5H2O was used as the standard to calibrate the spectra.

64

 

 

[Fe4sfi“(SMe)3L u3Sulfide Thiolate Total A
L = OH 369 90 459

L = AdoMet 381 115 496 37
L = Glycine 375 85 460 1
L = Methionine 370 80 450 -9

 

 

 

 

 

Table [11.1 DFT calculated covalencies (%) of [4Fe-4S]2+ cluster in PFL-AB with
different exogenous ligands.

The optimized structure of AdoMet bound [4Fe—4S]2+ cluster (Fig. III.15, left)
shows close contact between the AdoMet sulfonium S and the recent cluster bridging
sulfide (3.78 A) and the unique Fe (3.45 A). This is consistent with our crystal structure
study (manuscript in preparation) and ENDOR studies of PFL-AB. 9'” Energy levels of
the AdoMet bound [4Fe-4S]2+ cluster indicate that in addition to Fe3d based acceptor
orbitals present in a resting cluster, there are three AdoMet OS 6* orbitals in the same
energy region as these Fe3d based orbitals. These orbitals are ~3 eV higher in energy
relative to the C-S 0* of a thiolate (feature at 2473.5 eV) due to the positive charge on the
S atom in AdoMet. The AdoMet C-S 0* orbital between the sulfonium S and 5’-C has
back-bonding interactions with filled Fe-sulfide orbitals of the cluster. This shifis the
charge density to the OS 0* orbital from the cluster bridging sulfide, making the later a
poorer donor to the cluster. The reduced donation by the sulfide is compensated by
enhanced donation from the terminal thiolates which become more covalent. Although
the sulfide becomes a poorer donor to the cluster, its overall donation increases due to its
donation to the OS 0* orbital. Thus the pre-edge feature of the AdoMet bound [4Fe-

4S]2+ cluster includes transitions from the sulfide and the thiolate to the Fe3d based

65

orbitals and the sulfide to the OS 6* orbitals of AdoMet (which now has Sgp character of
sulfide mixed in due to back-bonding), increasing the overall pre-edge intensity of the

AdoMet bound cluster relative to a native cluster.

 

Figure III.15. DFT optimized structures of AdoMet bound cluster in the [4Fe—4S]2+ state
(left) and its one electron reduced form [4Fe-4S]1+ (right).

III.4 Conclusions

PFL-AE purified in the presence of 1 mM ascorbic acid had very similar UV-Vis
and EPR spectroscopic properties as the PFL-AB purified in the presence of 1 mM DTT,
the latter of which has been intensely investigated in our lab. Upon addition of exogenous
thiols (10 mM DTT or 10 mM cysteine), the EPR signals from both the proteins became
axial, indicating that the electronic structure of the cluster was more symmetric. This

would be consistent with coordination of the exogenous thiols to the unique site of the

66

cluster to form a F e484 cluster with 4 sulfur ligands (3 bridging sulfide and 1 terminal
thiolate) to each iron. Upon addition of AdoMet, the EPR spectra of the reduced PFL-AE
(purified in the presence of 1 mM ascorbic acid or 1 mM DTT) became rhombic,
suggesting that AdoMet significantly perturbed the cluster electronic structure,
presumably due to the carboxylate O and amine N coordinating to the unique iron. 9'”
Therefore AdoMet appears to bind the unique Fe of the Fe4S4 cluster even in the presence
of excess thiols as possible exogenous ligands. The results that PFL-AB purified in the
presence of ascorbic acid had very similar UV-Vis and EPR spectra to those of PFL-AE
purified in the presence of DTT indicate that the electronic structures in the two proteins
are the same.

The sulfur K-edge XAS data of PF L-AE showed a K-edge feature at 2473.5 eV
and a broad pre-edge feature at ~2470 eV, characteristic of [4Fe-4S]2+ clusters. The K-
edge is due to all sulfur (including all cysteines and methionines) ls electrons to C-S
0‘ transitions, and the pre-edge feature is due to the transitions from ls electrons in
bridging sulfide and terminal thiolates to unoccupied F63d orbitals. Both the K-edge
absorbance and the pre-edge feature had intensity increase upon addition of 3 equivalents
of AdoMet. The increase of the K-edge intensity is due to the absorbance from AdoMet,
however the increase of the pre-edge intensity is from the covelancy change in the cluster
in the presence of AdoMet, as AdoMet itself has no absorbance in this region. The DFT
calculation shows a 37% covalency increase of the Fe-S bonds, in which 25% is
contributed from the terminal thiolate, although the interaction between AdoMet and Fe-

Sb bonds is proposed more than the interaction between AdoMet and Fe—St bonds.

67

To explain this result, a model structure was constructed based on the DFT
calculation, and the energy level diagram indicates that the AdoMet OS 0* orbital
between the sulfonium S and 5’-C has back-bonding interaction with filled Fe-sulfide
orbitals of the cluster. This effect shifts charge density from the sulfide to the OS 0*
orbital, causing bridging sulfide to be a weaker donor to the cluster. The enhanced
donation from the terminal thiolate compensates for this change, inducing more
covalency into the F e-S bonds. The pre-edge feature has the characters of sulfur to Fe3d
antibonding transition and bridging sulfide to AdoMet S’C-S+ 0* transition, and therefore,
the overall donation from the u3su1f1de increases because of the back-bonding from
AdoMet. The XAS results demonstrate the interaction between the F e-S bonds and the
5’C-S+ bond, the position to be reductively cleaved to generate the putative
deoxyadenosyl radical. The charge shift from the cluster to AdoMet might be the initial

step to eject one electron from the cluster to the substrate, triggering the catalysis.

68

References

1. Knappe, J .; Schmitt, T., Biochemical and Biophysical Research Communications
1976, 71, (4), 1110-17.

2. Knappe, J .; Schacht, J .; Moeckel, W.; Hoepner, T.; Vetter, H., Jr.; Edenharder, R.,
European Journal of Biochemistry 1969, l 1, (2), 316-27.

3. Conradt, H.; Hohmann-Berger, M.; Hohmann, H. P.; Blaschkowski, H. P.;
Knappe, J ., Archives of Biochemistry and Biophysics 1984, 228, (1), 133-42.

4. Henshaw, T. F .; Cheek, J .; Broderick, J. B., Journal of the American Chemical
Society 2000, 122, (34), 8331-8332.

5. Frey, M.; Rothe, M.; Wagner, A. F. V.; Knappe, J ., Journal of Biological
Chemistry 1994, 269, (17), 12432-7.

6. Knappe, J .; Neugebauer, F. A.; Blaschkowski, H. P.; Gaenzler, M., Proceedings
of the National Academy of Sciences of the United States of America 1984, 81, (5), 1332-
5.

7. Knappe, J .; Elbert, S.; Frey, M.; Wagner, A. F. V., Biochemical Society
Transactions 1993, 21, (3), 731-4.

8. Krebs, C.; Broderick, W. E.; Henshaw, T. F.; Broderick, J. B.; Huynh, B. H.,
Journal of the American Chemical Society 2002, 124, (6), 912-913.

9. Walsby, C. J .; Ortillo, D.; Broderick, W. E.; Broderick, J. B.; Hoffman, B. M.,
Journal of the American Chemical Society 2002, 124, (3 8), 11270-11271.

10. Walsby, C. J.; Ortillo, D.; Yang, J.; Nnyepi, M. R.; Broderick, W. E.; Hoffman, B.
M.; Broderick, J. B., Inorganic Chemistry 2005, 44, (4), 727-741.

11. Walsby, C. J.; Hong, W.; Broderick, W. E.; Cheek, J .; Ortillo, D.; Broderick, J. B.;
Hoffman, B. M., Journal of the American Chemical Society 2002, 124, (12), 3143-3151.

12. Cosper, N. J .; Booker, S. J .; Ruzicka, F.; Frey, P. A.; Scott, R. A., Biochemistry
2000, 39, (51), 15668-73.

13. Lepore, B. W.; Ruzicka, F. J .; Frey, P. A.; Ringe, D., Proceedings of the National
Academy of Sciences of the United States of America 2005, 102, (3 9), 13819-13824.

14. Cosper, M. M.; Cosper, N. J .; Hong, W.; Shokes, J. E.; Broderick, W. E.;
Broderick, J. B.; Johnson, M. K.; Scott, R. A., Protein Science 2003, 12, (7), 1573-1577.

69

15. Glaser, T.; Rose, K.; Shadle, S. E.; Hedman, B.; Hodgson, K. 0.; Solomon, E. I.,
J Am Chem Soc 2001, 123, (3), 442-54.

16. Dey, A.; Glaser, T.; Couture, M. M.; Eltis, L. D.; Holm, R. H.; Hedman, B.;
Hodgson, K. 0.; Solomon, E. I., J Am Chem Soc 2004, 126, (26), 8320-8.

17. Dey, A.; Roche, C. L.; Walters, M. A.; Hodgson, K. 0.; Hedman, B.; Solomon, E.
I., Inorg Chem 2005, 44, (23), 8349-54.

18. Glaser, T.; Hedman, B.; Hodgson, K. 0.; Solomon, E. 1., Acc Chem Res 2000, 33,
(12), 859-68.

19. Rose, K. S., S. E.; Eidsness, M. K.; Kurtz, D. M., Jr.; Scott, R. A.; Hedman, B.;
Hodgson, K. 0.; Solomon, E. 1., Journal of American Chemical Society 1998, 120, (41),
10743-47.

20. Rose, K. S., S. E.; Glaser, T.; de Vries, S.; Cherepanov, A.; Canters, G. W.;
Hedman, B.; Hodgson, K. 0.; Solomon, E. 1., Journal of American Chemical Society

1999, 121, (11), 2353-63.

21. Dey, A.; Glaser, T.; Moura, J. J .; Holm, R. H.; Hedman, B.; Hodgson, K. 0.;
Solomon, E. I., J Am Chem Soc 2004, 126, (51), 16868-78.

22. Louis Noodleman, J. G. N., , Jr. Joseph H. Osborne, Arie Aizman, David A. Case,
Journal of American Chemical Society 1985, 107, (12), 3418-26.

23. Beinert, H.; Holm, R. H.; Munck, E., Science (Washington, D. C.) 1997, 277,
(5326), 653-659.

24. Solomon, E. 1., Hedman, B., Hodgson, K. 0., Dey, A., Szilagyi, R. K.,
Coordination Chemistry Reviews 2005, 249, 97-129.

25. Broderick, J. B.; Duderstadt, R. E.; Fernandez, D. C.; Wojtuszewski, K.; Henshaw,
T. F .; Johnson, M. K., Journal of the American Chemical Society 1997, 119, (31), 7396-
7397.

26. Ollagnier-de Choudens, S.; F ontecave, M., F EBS Lett 1999, 453, (1-2), 25-8.

70

CHAPTER IV

NUCLEAR RESONANCE VIBRATIONAL SPECTROSCOPY STUDY OF
PYRUVATE FORMATE-LYASE ACTIVATING ENZYME

IV.1 Introduction

PF L-AE utilizes a Fe4S4 cluster to catalyze the reductive cleavage of AdoMet to
form a putative deoxyadenosyl radical. 1‘ 2 The cluster is coordinated to the enzyme by
three cysteines, Cys-29, Cys-33, and Cys-36. 3 This CX3CX2C binding motif of the
cluster is highly conserved in all radical SAM superfamily enzymes. 4’ 5 Three of the four
irons in the cluster are coordinated by these three conserved cysteines, while the fourth
iron is coordinated by a non-cysteined ligand, making it chemically and spectroscopically
unique. The 4th iron binds AdoMet to anchors the substrate, however, the identity of the
exogenous ligand in the absence of AdoMet is unknown (scheme IV. 1 ). (*9

Another F e484 cluster enzyme, aconitase, also uses a site-differentiated Fe4S4
cluster to catalyze its reaction, although it is not a radical SAM superfamily protein. ‘0
ENDOR and resonance Raman spectroscopic evidence for aconitase show that a
hydroxide ion from the solvent binds to the unique Fe site in the absence of the substrate.
11' '2 The crystal structure of aconitase with the substrate isocitrate bound demonstrated
that the unique Fe was coordinated by one hydroxyl group, and one carboxylate group
13, 14 It

from isocitrate as well as one H20 to form a six ligand coordination environment.

has been proposed that the hydroxide ligand was protonated by a histidine and changed

71

position to let the substrate enter the active site. 8 The coordination of the unique Fe in

aconitase therefore changes from 4 to 6 ligands upon substrate binding.

 

 

 

 

 

 

 

Cys-S
Fe 3
/ I AIS-CY?»
S I Fe l
S...—
/ I/e\S-Cys
Fe S
X/
H
- + H
ooc>/\//s A
*H3N H3C °
AdoMet
OH OH
I
Cys-S
0 Fe 5
/ I /,l./S-Cys
S | Fe |
O S‘—
\l/ I/e\S-Cys + X
:\Fe S
/S\Ado
CH3

Scheme IV.1: Schematic representation of AdoMet binding to the unique iron of PFL-AE.
The fourth ligand to the unique Fe in the absence of AdoMet is unknown. AdoMet
replaces the unknown ligand when it binds to the Fe site.

72

Evidence suggests that PFL-AE has a similar change in coordination of the unique
Fe upon substrate binding. 6 To investigate whether a solvent hydroxide might bind to the
unique Fe and be replaced by AdoMet, a relatively new vibrational spectroscopic method,
nuclear resonance vibrational spectroscopy (N RVS), was used.

Vibrational spectroscopy provides some important information on the structure
and dynamics of the active sites in metalloproteins. ‘5 Resonance Raman (RR) in
particular has provided powerful insights into Fe4S4 cluster model compounds and
proteins. '6 However, RR spectroscopy is restricted by the selection rules, and as a site-
select technique it is sensitive only to those vibrations that are coupled to electronic
excitations. NRVS, sometimes called nuclear resonant inelastic X-ray scattering, detects
the complete vibrational spectrum of a Mossbauer-active nucleus, such as 57Fe. ”"9 This
technique couples nuclear excitation and molecular vibrations. A recoil-free absorption of
a photon can happen if the energy of the photon equals the energy difference between the
ground and excited nuclear states, as discovered by Mossbauer in 1958; 20 this energy is
14.413 keV for 57Fe. 2' A typical NRVS scan usually starts from —20 meV below the
recoil-free resonance energy and ends at +100 meV above that energy (-160 cm“1 — 800
cm!) Photons that have energy equal to the nuclear excitation state plus or minus
vibrational states (Stokes or anti-Stokes shift) can be absorbed by the nuclei (Fig IV.1).
Although this is an absorption spectroscopy, the detection is via measuring the resulting
6.4 keV atomic fluorescence emitted from the excited 57Fe atoms. The intensity of the

signal is directly related to the magnitude and direction of the vibration.

73

 

 

 

Nuclear Excited State

llllllllll llllllllll IIIIIIIIIIIIIII-

 

 

 

 

llllllllll llllllllll Illlllllll llllllllll IIIIIIIIH Illlllllll‘IIIIIIIIIIHIII"

 

Nuclear Ground State

 

Intensity

 

 

I
I
I
l
I
4 I

 

E

14.413 keV

Figure IV.1: Schematic representation of NRVS photon absorption in 57Fe. The solid line
is recoil-free energy absorption. The dashed line is the absorption involving one phonon
creation or annihilation. The broken line is the absorption involving two phonons creation
or annihilation.

Although this type of resonance absorption was predicted in 1960, 22 the initial
experiment was not reported until the mid 1990’s. 23 Technical difficulties arise from the

fact that the shifts of the vibrational bands are on the order of 1 meV, therefore requiring

source velocity of hundreds of meters per second to produce incident radiation of

74

appropriate energy (see late discussion of conventional Mossbauer spectroscopy). This is
not plausible with current Mossbauer instrumentation. Another technical difficulty arises
from that the intensity of the NRVS absorption is only about 1% of the recoil-free
resonance, and therefore an ultra bright source is required. These difficulties have been
overcome by a new generation synchrotron X-ray source. The bandwidth of the
excitation X-ray beam (with energy around 14.4 keV) can be narrowed down to 1.0 meV
via two monochromators, a diamond crystal as the premonochromator and a Si crystal as
the high-resolution monochromator. The wavelength (energy) of the X-ray is tuned
according to the Bragg reflection of the single crystal, n)» = 2dsin0, where A is the
wavelength, d is the separation between the lattice planes, 0 is the angle of incidence and
n = 1 is the fundamental reflection. The corresponding resolution of the spectrum is 7 — 8

cm], (Fig 1V.2).

hi gh-resolution

 

 

 

 

 

 

 

 

 

 

 

 

bandwidth 1 eV bandwidth 1 meV

X-ray ‘ ‘

source 9 PM I'IRM sample
‘ 4|

\ : ,’
x : - avalanche
X : APD photodiode
‘, : I’ , . detector
5 '

data acruisition and timing
motion control v lecmt

 

 

 

 

 

 

Figure 1V.2: Schematic diagram of NRVS experimental set-up. The bandwidth of the
irradiation beam is cut down to 1 meV by two monochromators. The APD is turned off
during the X-ray pulse in order to remove the elastically scattered photons. It is
controlled by the timing circuit to collect the delayed atomic fluorescence corresponding
to the nuclear absorption.

75

The relevant transition for 57F e is from the nuclear spin state I = 1/2 to the spin
state I = 3/2, corresponding to an energy of 14.413 keV. Conventional Mossbauer
spectroscopy monitors the electronic structure of 57Fe produced by the interaction
between the nucleus and electrons. 24 The excitation source is 57Co, with decay to 57Fe to
generating the 14.413 keV incident radiation with a bandwidth of 4.7x10'9 eV. To probe
the different energy transitions of the sample, the source is moved to produce a Doppler
shift of the incident radiation. A movement at a velocity of 1 mm/sec toward the sample
can increase the energy of the beam by 4.8x10'8 eV, 10 times the original bandwidth.
This velocity, called the isomer shift, is used as the unit in Mossbauer spectroscopy. The
actual spectrum is a plot of transmission counts versus the isomer shift.

There are several significant interactions between the nucleus and the electrons,
including the electrostatic attraction. The excited 57Fe nucleus is 0.1% smaller than the
ground state nucleus, and therefore the interaction between the electrons and nucleus
changes upon excitation. The difference in electron density on the nucleus therefore can
shift the Mossbauer transition energy, and is recorded as the isomer shift. The Is and 25
electrons are not affected significantly by the electronic environment of the 57Fe, but the
3d electrons shield the 3s electrons and decrease the electron density on the nucleus. The
isomer shift therefore decreases when there are more delocalized 3d electrons. The 43
character in the bonding orbital also reduce the isomer shift, which makes interpretation
of results more complicated. 25

The excited state of 57F e has quadrupole degeneracies, and the ground state has
double degeneracies in the absence of external electric field. The quadrupole

degeneracies can be partially lifted if the nucleus is surrounded by a low symmetric

76

electric field, due to the interaction of the nuclear quadrupole moment and the electric
field gradient. In the absence of magnetic interactions, the nuclear excited state is split
into two degenerate doublets. The ground state can have transitions to both of the
doublets that are separated by the quadrupole splitting energy (AE).

The nuclear spin also possesses a magnetic moment, which can be influenced by
the external electron magnetic field, a sum of the magnetic moments from the unpaired
electrons. For 5 7F e, the electron magnetic perturbation at the nucleus is approximately
proportional to the sum of the 3d electron interference. However, the 4s character in the
bonding orbitals mixes with the filled molecular orbital made from 3d orbitals, which
makes the calculation more complicated. Therefore, the values of the magnetic
interaction are determined empirically in many cases. 26 Mossbauer spectroscopy,
combining these three factors, can provide information on the oxidation state, spin state,
and coordination environment of the atom.

The NRVS spectra of 57Fe complexes are often compared with the resonance
Raman spectra of the same compounds. In Raman spectroscopy, an incident beam excites
a molecule from a ground vibrational state to a virtual excited vibrational state. The re-
emitted photon keeps the same energy if the molecule returns to the ground vibrational
state (v = 0); this is called Raleigh scattering. If the molecule returns to the lowest excited
vibrational state (v = l), the re-emitted photon will have less energy than the incident
photon; this is referred as Stokes scattering. If the molecule transits from the v = 1 state
to v = 0 state because of the irradiation, the scattered photon will have more energy than
the incident photon, a phenomenon called anti-Stokes scattering. Stokes and anti-Stokes

scatterings are two types of Raman scattering (Fig IV.3 A). Raman scattering can be

77

greatly enhanced (102 to 106 times) if the energy of the incident photon matches the
difference between the electronic ground state and excited state of the molecule (Fig IV.3
B). This is resonance Raman spectroscopy. Resonance Raman is a very powerful method
to study active sites in metalloproteins, as proteins in aqueous solution in the
concentration range of uM to mM can be investigated due to resonance enhancement.
Secondly, only the vibrational modes coupled to the electronic transitions are enhanced in
resonance Raman, which provides high selection for the metal center even in very large
proteins. However, resonance Raman also has some significant restrictions. Only the
vibrations that change the polarizability of the molecule are Raman active, which means
some vibrational modes are not observable in Raman spectroscopy. A further limitation
of resonance Raman is that not all the Raman active vibrations are enhanced because the
intensity increase of the bands is due to the electronic state transition induced
polarizability change. Those vibrations coupled with molecular geometry change of the
excited state are promoted, but others are not.

Resonance Raman of iron-sulfur cluster proteins utilizes excitation at the
absorption wavelength range of the sulfur to iron charge transfer. '6 Therefore, the Raman
bands involving Fe-S vibrations are enhanced, however the intensity of the vibration
between the unique iron and a non-sulfur ligand such as OH' or H20 may not be
increased. In fact, no Fe to exogenous ligand vibration has been detected yet in radical
SAM superfamily proteins. On the other hand, NRVS is not subject to the selection rule
or the electronic absorption enhancement, since it directly monitors the all movements of
the 5 7F e nucleus. Thus, it provides a greater opportunity to observe the unique S7Fe

vibration involving the exogenous ligand.

78

Stokes scattering Anti-Stokes scattering

 

 

hva

 

 

 

 

 

 

 

 

 

Vs = Vex ‘ Vv Vex vs : Vex + Vv

Frequency

A
............. virtual
11 ll
state
thX hvs hvex
V V = 1
v = 0 V
Raleigh
5:?” l1ne
2
B
E Stokes
5 shift .
'6' Ann-Stokes
52 shift

 

 

 

 

 

 

 

l
Excited
electronic *
state
hvex hvs
Ground

 

ClCCU'OIlIC _
state

Figure IV.3: Schematic diagrams of Raman scattering. A: Raleigh (vex), Stokes (vex - vv)
and anti-Stokes (vex - vv) scattering from the incident beam. B: Resonance Raman
scattering. The Raman shift (V.,x - v.) is enhanced by the electronic absorption.

1V.2 Materials and methods

IV.2.1 Chemicals and materials

All chemicals were purchased from commercial sources and were of the highest

quality except as noted otherwise. The isotope 57Fe metal powder was purchased from

Cambridge Isotope Laboratories. The solution of 57FeSO4 was prepared by dissolving the

metal in 3:1 H20:H2$O4. The pH of the solution was adjusted to 6.5 by 6 M NaOH. The

79

Fe concentration was determined by the method described in chapter 11. Natural
abundance 56Fe PFL-AE was overexpressed and purified following the procedure

described in chapter 11.

1V.2.2 Preparation of 57Fe labeled PFL-AE

Preparation of unique-iron site ”Fe labeled PFL-AE. All procedures were
carried out under strictly anaerobic conditions. Purified natural abundance PFL-AE (1
mL) was loaded on an Amersham HiTrap Desalting column and eluted with buffer (50
mM Tris, 200 mM NaCl, pH 8.5) to remove DTT. The colored fraction was collected.
The protein concentration and Fe content were determined via the methods described in
chapter 11. Three equivalents of K3Fe(CN)6 (in H20, 1.9 mM final conc, relative to [4F e-
4S]2+ cluster) were added into the protein and one equivalent of EDTA was added 10
minutes later. The oxidized protein was gel filtered through the same column again after
10 minutes to remove EDTA, extra oxidizing agent and released iron. The Fe content was
determined by the same method again. 57FeSOa, 1.1 equivalent relative to the [3Fe-4S]1+
cluster (converted from Fe content, determined by Fe assay as described in Chapter II,
presumably in [3Fe-4S]1+ state), was then introduced into the protein to reconstitute the
[4Fe-4S]2+ cluster. To provide reducing environment, DTT (10 mM final cone) was
added after 10 minutes. The protein was concentrated on a YM-lO ultra filtration
membrane (Millipore). The final concentration of the reconstituted protein was 3.6 mM
and the Fe content was 4.3 F e/PFL-AE.

Preparation of all sites 5 7F e labeled PFL-AE. PFL-AE was overexpressed and

purified as described in Chapter 11 except that 57FeS04, as the only iron source, was

80

introduced into the cell growth to a final concentration of 20 uM. The final concentration
of purified [57Fe]PFL-AE was 1.8 mM and the Fe content was 3.1 Fe/PFL-AE. The
purified PFL-AB was concentrated on YM-10 centrifugal filters (Millipore) at 10,000

rpm for 1 hour at 4 °C under anaerobic conditions to make final concentration 10 mM.

1V.2.3 NRVS sample preparation

All procedures were carried out under strictly anaerobic conditions. For S7Fe
unique site labeled PFL-AE, protein alone (3.6 mM) or protein (3.5 mM) plus AdoMet
(6.9 mM) were loaded into NRVS cells. The same samples were used for Mossbauer and
resonance Raman spectroscopy. For 57Fe all sites labeled PFL-AB, 10 mM PFL-AE or
6.5 mM PFL-AB plus 19.3 mM AdoMet were loaded into NRVS cells. The sample cells

were sealed with Teflon washers and sapphire windows, and stored at —80 °C.

IV.2.4 Mossbauer spectroscopy

Mossbauer spectra were recorded at Emory University in a strong field
spectrometer at 4.2 K in a magnetic field of 50 mT applied parallel to the incident beam.
The spectrometer operated in a constant acceleration mode in a transmission geometry.
The zero velocity of the spectra refers to the centroid of a room temperature spectrum of

a metallic iron foil.
1V.2.5 Resonance Raman spectroscopy

The resonance Raman spectra were recorded at Northeastern Univeristy. The

samples were excited by a laser beam at 413.3 nm from a Kr+ laser source (Coherent

81

Innova 302C), and measured at 20 K in a helium flow cryostat (Oxford Instruments). The
Raman scattering was collected in a LabRam HR800 spectrometer (J Y Horiba). The laser
power to excite the PFL—AE sample was 3.3 mW. The collection time was 4 minutes and
spectra were accumulated 12 times. The laser power to excite the sample of PFL-AE with
AdoMet was 5.3 mW. The collection time was 4 minutes and spectra were accumulated

1 1 times.

IV.2.6 NRVS spectroscopy

The NRVS measurement was conducted at sector 3-ID of the Advanced Photon
Source at Argonne National Laboratory. A high-resolution X-ray beam with energy
around 14.4 keV and resolution of 0.85 meV (7 cm'l) was used. The beam energy was
scanned above and below the nuclear excitation energy of 57Fe. The bandwidth was
narrowed by a diamond premonochromator followed by a dispersive pair of asymmetric

2 area was used to collect the atomic

Si crystals. An APD detector with a 1 cm
fluorescence from the sample. Pulses from the APD were recorded by a counter that

discriminated the fluorescence photons 140 ns later than the strong coincident X-ray

scattering. The sample was cooled down to 20 K in a He flow cryostat.

IV.3 Results and discussion
IV.3.1 Mossbauer spectra of PF L-AE
The Mossbauer spectrum of the unique site 5 7F e labeled PF L-AE showed 84% of

total 5 7F e was in the [4Fe-4S]2+ state (isomer shift 6 = 0.42 mm/sec, quadrupole splitting

82

AEQ = 1.12 mrn/sec), and 16% of the total 57F e was in the solution as free ferrous (data
not shown). It was not possible to distinguish the unique Fe from the other irons in the
cluster in the absence of AdoMet, as the electronic environment was almost the same for
each iron in the cluster. In the presence of AdoMet, the spectrum indicated 16% of the
total 57Fe was in the unique iron site with AdoMet bound (isomer shift 5 = 0.72 mm/sec,
quadruple splitting AEQ = 1.19 mm/sec) (Fig 1V.4). The increase of the isomer shift and
quadrupole splitting of the unique Fe implied an increase of coordination number to the
Fe due to AdoMet binding, as we have previously published. 6 However, the unique Fe
signal was not separated clearly from the irons in the other positions of the cluster, which
indicated that the AdoMet binding was not complete. Compared to previously published
Mossbauer spectra, 6 it was estimated that 15 — 55% of the [4Fe-4S]2+ cluster was not
bound to AdoMet. In the current experiments 68% of the total 57Fe was in other positions
of the cluster, or in the unique site without AdoMet bound.

The Mossbauer data demonstrated that 57Fe was reconstituted not only into the
unique site, but also into other sites in the cluster. Earlier Mossbauer experiments with
PFL-AE showed no exchange between the unique site and the other iron sites. 6 In these
earlier experiments, the 57Fe labeled sample was flash fiozen in liquid N2 immediately
after it was reconstituted. However, in the NRVS sample preparation, the protein was
concentrated for more than 30 minutes after the reconstitution, providing more time for
exchange of iron between sites. This implies that the cluster is not a static structure, but

rather dynamic under the reducing conditions (10 mM DTT).

83

 

Absorption (%)
o)
?

 

 

 

 

-4 -2 0 2 4

\b Io c ity (m m ls)

Figure IV .4: Mossbauer spectrum of ”Fe labeled unique site PFL-AB in the presence of
AdoMet. Data were recorded at 4.2 K in a magnetic field of 50 mT parallel to the
incident beam. The black dashed line is the experimental signal. The black solid line is
the deconvoluted 57Fe in the [4Fe-4S]2+ cluster other than the unique site. The dark grey
§7olid line is the deconvoluted 57Fe in the unique site. The light grey line is the free ferrous
Fe.

IV.3.2 Resonance Raman spectra of PFL-AE
Resonance Raman spectra of PFL-AE and PFL-AE with AdoMet showed typical
[4Fe-4S]2+ signals (Fig IV.5), compared to published [4Fe-4S]2+ model compounds and
proteins 1‘ 27'29. The strongest band at 338 cm'1 is the totally symmetric breathing mode of
the Fe484 cubane. 1' 16’ 29 The band at 370 cm'1 is assigned to the asymmetric stretching
mode of the bond between iron and the terminal sulfur. 28’” The band at 396 cm‘1 is the
symmetric stretching mode of the bond between iron and the terminal sulfur. 28' 29 The
bands at 290 cm'1 and 320 cm" are the stretching modes of the bond between iron and the

bridging sulfide. ”'29

84

   

36:38 6"}: 3:333JJ'

M
c. .-..-'.'—'.'-'-..'L‘-‘—'-‘.‘.‘-.'-

 

 

 

 

 

 

 

 

 

.5 u
a 5 5 ' E
a e =
8 a a :. 3
-H C) H .
“ l: ., .:
i! i; It
no I' ::
”I :: II
RN. 2: A
55 EV'H—I
II .I
5:.— - 1‘ t
38 3 55 T".
:3t :: ;:
:ll 1 2: :1
:m: ::
.VIE B
“1+
nnnnn Ill-IIAAIAIAA-lllll-I-l.------‘----‘----
250 300 350 400 450

Wavenumber ( c m '1)

Figure IV.5: Resonance Raman spectra of PFL-AE and PFL-AE with AdoMet. The
spectra were recorded at 20 K. The excitation wavelength was 413.3 nm. The solid bands
were the vibrational modes from the protein, and the dashed bands were from the
sapphire window of the sample cell. A: Spectrum of PFL-AE. Excitation power 3.3 mW.
48 min accumulation. B: Spectrum of PFL-AE with AdoMet. Excitation power 5.3 mW.
44 min accumulation.

In the presence of AdoMet, the resonance Raman spectrum of PFL-AE did not

1 split to two bands at 391 and 401 cm".

change signifcantly. The band at 396 cm'
AdoMet binding to the unique Fe changes the coordination number of this site, which

decreases the symmetry of the cluster and might lead to splitting of this band. The

relative intensity of the bands at 338 cm’1 and 370 cm'1 also changes upon addition of

AdoMet, with intensity of the band at 370 cm'1 increasing in the presence of AdoMet.
These two changes point to some differences in the Fe-St bonds when AdoMet is
coordinated to the unique Fe. Upon AdoMet binding, the intensities of the bands at 287,
320 and ~345 cm'l, mainly F e-Sb stretching modes, are also increased. Coordination of

AdoMet might induce more tense in the [4Fe-4S] cube, changing the Fe-Sb bonds.

IV.3.3 NRVS spectra of PF L-AE

The NRVS spectra of 5 7F e reconstituted PF L-AE and PFL-AB with AdoMet
reveal the complete vibrational modes of the [4Fe-43]2+ cluster from o -— 660 cm" (Fig
1V.6). The experimental recording shows the inelastic nuclear absorption of synchrotron
radiation as a function of energy of the incident radiation. Various features contributes to
the energy spectrum, in which the inelastic absorption is accompanied by excitation or
annihilation of different vibrational states. To obtain the spectrum corresponding to single
vibrational state change, decomposition of multiple probabilities of vibrational states
density is used to deconvolute the spectrum (both the experimental spectra and
deconvoluted spectra are shown in Fig 1V.6). Four major features are observed: a triplet
between 340 and 390 cm'l, a doublet between 250 and 300 cm'], a strong band at 160 cm’
1, and a broad band below 100 cm". The vibrations above 200 cm"1 are assigned to Fe-S
bond stretching modes, the vibrations between 100 and 200 cm'1 are assigned to S-Fe-S’
bending modes, and those below 100 cm"1 are assigned to cubane twist modes. 3° The
detailed assignments are listed in table IV.1, compared with calculated normal modes of

model compounds. 30‘ '4

86

Upon AdoMet binding, there were some changes in the vibrational modes. The
intensity of the band at 432 cm", which is the Fe-ligand plus Fe-S' stretching mode,
increased dramatically, indicating that the coordination of the amino N and carboxylate O
in AdoMet to the unique Fe enhanced the Fe-ligand vibration. The intensity of the bands
at 388 crn'l ( Fe-St stretching) and 158 cm'1 (S-Fe-S’ bending) also increased, indicating
that the magnitude of these two vibrations was increased. There was a split of the peak at
278 cm], also a Fe-St stretch. As with the resonance Raman spectra, AdoMet binding
causes some change to the Fe-St bonds. Since the terminal sulfur ligands to the cluster in
PFL-AE are three cysteines, this indicates that the details of cluster coordination to the

protein may change in the presence of AdoMet.

 

 

 

 

 

energy (meV)

0 40 60 80
’g . I ' I ' l ' l ' I V r v I I I - y
“g 0.015
“6‘ 0.01
"E

0.005

8 :
.6 I
'S 0

0 200 400 _ 600 800

cm’1

Figure 1V.6: NRVS spectra of 5 7F e reconstituted PFL-AB and PFL-AE with AdoMet. The
shadow line is the experimental data of PFL-AE with AdoMet. The black line was the
deconvoluted spectrum of PFL-AE with AdoMet. The grey line was the deconvoluted
spectrum of PF L-AE.

The stretching of the Fe-O bond between the unique Fe and the exogenous OH" in

aconitase was predicted to occur at 408 cm"1 in aconitase. '2 However, resonance Raman

87

spectroscopy has never detected this type of vibration in either aconitase or radical SAM
superfamily proteins, presumably because the Fe-O stretching might not be enhanced by
the excitation of sulfur to iron charge transfer absorption. The resonance Raman spectra
of PFL-AB were no exception. On the other hand, NRVS is not limited by the
requirement for resonance enhancement. The increase in intensity of the peak at 432 cm'1
and the up-shift of the band at 458 cm‘1 in the presence of AdoMet could be indicative of
a change in the exogenous ligand. The ligands from AdoMet could cause stronger and
more rigid vibrations than the vibrations between the unique iron and the (non-AdoMet)

exogenous ligand.

 

 

 

Wavenumber (cm'l)
Mode Symmetry PF L-AE PF L-AE + AdoMet Calculated
Fe-Ligand + Fe-St stretch E 432 432 429.3
Fe-St stretch A; 390 388 390.4
Mainly Fe-St stretch B2 368 364 361.0
Fe-Sb stretch A. 344 348 347.1
A1 300 299.9
A2 278 278 278.4
E 254 266.1
S-Fe-S’ bend A1 180 184 179.0
E 160 160 157.2
Cubane twist B1 58 60 54.3

 

 

 

 

 

Table IV.1 NRVS normal mode assignments of PFL-AE and PFL-AE with AdoMet,
compared with published calculations.

88

The NRVS spectrum of all sites labeled [57Fe]PFL-AE is different from that of the
unique site labeled sample (Fig 1V.7). The most intense band shifts from 160 cm'1 to 142
cm'l. The peak at 278 cm'1 is broader and has a small split. The peaks at 368 and 390 cm'
‘ combine to form a broad band. The intensity of the bands above 400 cm'1 are decreased.
The difference observed between the all-sites labeled and unique labeled samples might
be due to different forms of iron-sulfur clusters existing in the protein. The Fe content of
the purified all-sites labeled enzyme was only 3.1 F e/PFL-AE. Previous Mbssbauer
studies of PFL-AE with similar iron content showed that there were [2Fe-ZS]2+, [3Fe-
4S]1+ and [4Fe-4S]2+ clusters present. 25 Unfortunately, we did not obtain Méssbauer data
on the current sample, so we cannot be certain as to the cluster content. However it is
reasonable to suspect that the sample also had several cluster forms. The assignment of
the vibrational modes would thus be very difficult due to the lack of quantification of

different clusters present.

 

 

 

 

energy (meV)

V 0.015
g 3 0.01
“a l ‘
.3‘ ,
g l 1 0.005
'8 i. 3
ae' t ....................... _ _ . 0
> ‘ l I l l _ l -

0 2.00 400 600 800

cm'1

Figure 1V.7: NRVS spectrum of all sites labeled [57Fe]PFL-AE. The shadow line was the
experimental signal. The grey solid line was the deconvoluted spectrum of the protein.

89

1V.4 Conclusions

The isotope 57Fe has been reconstituted into the [4Fe-4S]2+ cluster in PFL-AE
after the purified protein was oxidized to the [3Fe-4S]l+ state under controlled conditions.
The Méssbauer spectrum showed that 84% of the total 5 7F e was incorporated into the
cluster, and 16% of the total iron was in the solution as free ferrous. The Méssbauer
spectrum of PF L-AE with AdoMet bound demonstrated that only 16% of the 57Fe
occupied the unique site with AdoMet bound, while the other 68% of 57Fe was scrambled
into other sites of the cluster and/or on the unique site without AdoMet bound. We have
previously show that 57F e could be taken into the [3Fe-4S]1+ cluster to form the [4Fe-
4S]2+ cluster, however the prolonged incubation in the presence of 10 mM DTT utilized
in the current experiment may have resulted in scrambling of the 57Fe into all sites of the
cluster. The mechanism of label scrambling is unknown.

The resonance Raman spectra of PFL-AE and PF L-AE with AdoMet bound are
typical [4Fe-4S]2+ signals. There are strong symmetric Fe-S stretching modes and weak
asymmetric Fe-S stretching modes. In the presence of AdoMet, there are small changes in
the vibrational modes of Fe-St stretching, but nothing that definitively identifies the
unique site ligand in the absence of AdoMet.

The NRVS spectra of the protein and the protein with substrate provide a
complete set of the cluster vibrational movements, including the Fe-ligand stretching, Fe-
S stretching, S-Fe-S’ bending, and the cubane twist. Compared with resonance Raman
spectra, NRVS not only demonstrates the vibrations that are missing in Raman
spectroscopy, but also provides more detailed information of the modes. Upon AdoMet

binding, significant changes of Fe-St vibrations are observed. The resonance Raman and

90

NRVS results imply the coordination of the cluster to the protein is altered when AdoMet
binds, based on the increases of the intensity and energy shift of the bands.

NRVS spectra show the vibrational information between Fe and exogenous
ligands above 400 cm'l, which are missing in the resonance Raman results. The changes
of the bands at 432 and 458 cm'1 in the presence and absence of AdoMet binding might
provide clues as to identify the exogenous ligand to the unique Fe site.

The NRVS spectrum of all sites labeled [57Fe]PFL-AE is different fiom the 57Fe
reconstituted protein. The explanation might be that the former spectrum is a mixture of
the vibrations of [2Fe-2S]2+, [3Fe-4S]1+ and [4Fe-4S]2+ clusters. Without the Mdssbauer

quantification of each cluster form, it is difficult to interpret these results.

91

References

1. Broderick, J. B.; Duderstadt, R. E.; Fernandez, D. C.; Wojtuszewski, K.; Henshaw,
T. F.; Johnson, M. K., Journal of the American Chemical Society 1997, 119, (31), 7396-
7397.

2. Henshaw, T. F .; Cheek, J.; Broderick, J. B., Journal of the American Chemical
Society 2000, 122, (34), 8331-8332.

3. Kulzer, R.; Pils, T.; Kappl, R.; Huttermann, J .; Knappe, J., Journal of Biological
Chemistry 1998, 273, (9), 4897-4903.

4. Sofia, H. J .; Chen, G.; Hetzler, B. G.; Reyes-Spindola, J. F.; Miller, N. E., Nucleic
Acids Research 2001, 29, (5), 1097-1106.

5. Check, J.; Broderick, J. B., Journal of biological inorganic chemistry .' JBIC : a
publication of the Society of Biological Inorganic Chemistry 2001, 6, (3), 209-26.

6. Krebs, C.; Broderick, W. E.; Henshaw, T. F.; Broderick, J. B.; Huynh, B. H.,
Journal of the American Chemical Society 2002, 124, (6), 912-913.

7. Walsby, C. J.; Ortillo, D.; Broderick, W. E.; Broderick, J. B.; Hoffman, B. M.,
Journal of the American Chemical Society 2002, 124, (3 8), 11270-1 1271.

8. Walsby, C. J .; Hong, W.; Broderick, W. E.; Cheek, J .; Ortillo, D.; Broderick, J. B.;
Hoffman, B. M., Journal of the American Chemical Society 2002, 124, (12), 3143-3151.

9. Walsby, C. J.; Ortillo, D.; Yang, J .; Nnyepi, M. R.; Broderick, W. E.; Hoffman, B.
M.; Broderick, J. B., Inorganic Chemistry 2005, 44, (4), 727-741.

10. Beinert, H.; Kennedy, M. C.; Stout, C. D., Chemical Reviews (Washington, D. C.)
1996, 96, (7), 2335-2373.

11. Werst, M. M.; Kennedy, M. C.; Beinert, H.; Hoffman, B. M., Biochemistry 1990,
29, (46), 10526-32.

12. LaTonya K. Kilpatrick, M. C. K., Helmut Beinert, Roman S. Czernuszewicz,
Thomas G. Spiro, Di Qiu, Journal of American Chemical Society 1994, 116, (9), 4053-61.

13. Lauble, H.; Kennedy, M. C.; Beinert, H.; Stout, C. D., Biochemistry 1992, 31,
(10), 2735-48.

14. Lauble, H.; Kennedy, M. C.; Beinert, H.; Stout, C. D., J Mol Biol 1994, 237, (4),
437-51.

92

15. Barth, A.; Zscherp, C., Q Rev Biophys 2002, 35, (4), 369-430.

16. Spiro, T. G. C., Roman S., Resonance Raman spectroscopy. Physical Methods in
Bioinorganic Chemistry. University Science Books: Sausalito, CA, 2000; p pp 97 - 112.

17. Chumakov, A. I., Sturhahn, W., Hyperfine Interactions 1999, 123/124, 781-808.

18. Leu, B. M.; Zgierski, M. Z.; Wyllie, G. R.; Scheidt, W. R.; Sturhahn, W.; Alp, E.
E.; Durbin, S. M.; Sage, J. T., J Am Chem Soc 2004, 126, (13), 4211-27.

19. Scheidt, W. R., Durbin, S. M., Sage, J. T., Journal of Inorganic Biochemistry
2005, 99, 60-71.

20. Méssbauer, R., Z. Physik 1958, 151, 124-143.

21. Edwards, P. R., Johnson, C. E. and Williams, R. J. P., J. Chem. Phys. 1967, 47,
2074-82.

22. A. Rahman, K. S. S., and A. delander, Phys. Rev. 1960, 120, (4), 1093-1102.

23. Sturhahn, W.; Toellner, T. S.; Alp, E. B; Zhang, X.; Ando, M.; Yoda, Y.; Kikuta,
S.; Seto, M.; Kimball, C. W.; Dabrowski, B., Physical Review Letters 1995, 74, (19),
3832-3835.

24. Cranshaw, T. E., Dale, B. W., Longworth, C. O. and Johnson, C. E., Mossbauer
Spectroscopy and its Applications. Cambridge Univ. Press.: 1985.

25. Greenwood, N. N. a. G., T. C., Mossbauer Spectroscopy. Chapman and Hall:
London, 1971.

26. Lang, G. a. M., W., Proc. Phys. Soc. 1966, 87, 3-34.

27. Johnson, M. K.; Czernuszewicz, R. S.; Spiro, T. G.; Ramsay, R. R.; Singer, T. P.,
J Biol Chem 1983, 258, (21), 12771-4.

28. Vittal K. Yachandra, J. H., Andrew Gewirth, Roman S. Czernuszewicz, T.
Kimura, Richard H. Holm, Thomas G. Spiro, Journal of American Chemical Society
1983, 105, 6462-6468.

29. Roman S. Czernuszewicz, K. A. M., Michael K. Johnson, Andrew Gewirth,
Thomas G. Spiro, Journal of American Chemical Society 1987, 109, (23), 7178-87.

30. Xiao, Y.; Koutrnos, M.; Case, D. A.; Coucouvanis, D.; Wang, H.; Cramer, S. P.,
Dalton Trans 2006, ( 18), 2192-201.

93

CHAPTER V

STRUCTURAL CHANGE IN PYRUVATE FORMATE-LYASE DURING
ACTIVATION

V. 1 Introduction

Pyruvate formate-lyase (PF L) utilizes a glycyl radical on G1y-734 to initiate
catalysis. 1'3 This radical is generated by a putative 5’-deoxyadenosyl (5’-dAdo) radical
intermediate, which is produced from the reductive cleavage of AdoMet catalyzed by
PF L-AE. 4’ 5’ 6 There has been no direct detection of the putative 5’-dAdo radical as yet,
although an allylic analogue of the deoxyadenosyl radical was identified by EPR during
LAM catalysis. 7 PFL is a homodimer, with each monomer comprising a 10 strand B/or

barrel. 8' 9

A short loop containing Gly-734 is buried in the non-radical form of PFL, with
G1y-734 8 A from the surface. 8'10 Considering the extremely high reactivity expected for
the deoxyadenosyl radical intermediate, it is expected that it must be generated in the
immediate vicinity of Gly-734 of PFL. A major unresolved question is how the PFL-AE
active site accesses Gly-734 of PFL to catalyze activation.

The structure of PFL is similar to subunit R1 of the class I ribonucleotide
reductase (RNR), 7‘ H which uses a thiyl radical in catalysis. The activation of class I
RNR is initiated by a tyrosyl radical in subunit R2, 35 A apart from the thiyl radical. '2 It
has been proposed that there is a long-range proton-coupled electron transfer to transport

the radical from the tyrosine to the cysteine. '3 Ten residues total are proposed to be

involved in the pathway. However, PFL probably has a different mechanism for

94

activation. Direct hydrogen transfer from Gly-734 to the product of AdoMet cleavage
was observed, 14 by which a direct interaction between the G1y-734 residue and the
putative deoxyadenosyl radical intermediate was proposed. This implies that PFL must
undergoe a conformational change prior to activation, in which the glycine loop is moved
out from the protein core. '0

EPR Spin quantification of the glycyl radicals in activated PFL, activity assays of
the enzyme, and circular dichroism (CD) spectroscopy were used to investigate the
proposed mechanism of PFL activation. The experiments provide evidence to support a

PFL conformational change. An examination of the published PFL crystal structure

suggests possible residues which may be involved in the glycine loop position change.

V.2 Materials and methods
V.2.1 Chemicals and materials

All chemicals were commercially purchased with the highest quality except as
noted otherwise. AdoMet was synthesized according to the procedure described in
chapter III. Overexpression and purification of PFL and PF L-AE, as well as protein and

iron assays were carried out following the procedures in chapter II.

V.2.2 PFL activation
All procedures were carried out under strictly anaerobic conditions (Mbraun glove

box, 02 level < 2 ppm). All components were degassed completely on a Shlenk line

before the reaction. An activation mix contained 100 mM Tris/HCl pH 7.6, 100 mM KCl,

95

10 mM oxamate, 10 mM DTT, 50 or 100 uM PFL, 0.5 — 100 11M PFL-AB, 10
equivalents of AdoMet relative to PFL, and 100 11M 5-deazariboflavin. The mix was
transferred into an EPR tube and illuminated by a 300 W halogen light approximately 5
cm away from the tube for different time periods. The EPR tube was put in an ice-water
bath (4 °C) to keep the protein from over-heating by the lamp. The mix was flash frozen

in liquid N2 immediately after illumination and later examined by using EPR spectrocopy.

V.2.3 EPR spectroscopy

EPR spectra were recorded at X-band (9.37 GHz) in a Bruker ER-200D-SRC
spectrometer at 60 K. The microwave power was 19 uW. The external magnetic field
was scanned from 3000 to 4000 G. The static field was 1000 G. The modulation
frequency was 100 kHz, and modulation amplitude was 5 G. The resolution was 1024.
K2(SO3)2NO was used as a standard to determine the quantity of the PFL glycyl radical,

according to the published method. 6‘ ‘5

V.2.4 Hydrogen-deuterium exchange of activated PFL in D20

All the procedures were conducted under strictly anaerobic conditions. After the
EPR spectra were measured, the samples were thawed. Buffer (150 11L of 100 mM
Tris/DCl, 100 mM KCl in D20, pD 8.2) was added into 150 11L of the EPR samples with
PFL to PFL-AE ratios of 100:1, 10:1 and 1:1. Immediately after mixing, the sample was
pipetted into an EPR tube and flash frozen in liquid N2. Another sample was prepared by
adding 225 uL of the D20 buffer into 75 11L of the same EPR sample, following the

procedure above.

96

V.2.5 PF L activity assay

All the procedures were conducted under strictly anaerobic conditions. The
protein activity was measured by using a modification of published procedures. ‘6 The
rate of acetyl-CoA production was monitored by a coupled enzymatic reaction (Fig V.1).
Completely activated PFL (1 uL of 100 uM), as determined by EPR, was added to 99 11L
of buffer (100 mM Tris/HCl, pH 7.6, 100 mM KCl). This diluted sample (5 11L) was
mixed with 795 11L of the coupling enzyme mix, containing 100 mM Tris-HCl pH 8.1, 10
mM pyruvate, 55 pM CoA, 10 mM malate, 3 mM NAD+, 0.1 mg BSA, 300 U malic
dehydrogenase and 20 U citrate synthase. The NAD+ reduction rate, recorded as the
increase in absorption at 340 nm due to production of NADH, was used as the indicator
of PFL activity. As can be seen in Fig V.1, the reduction of NAD+ to NADH by malic
dehydrogenase is at equilibrium if the oxaloacetate concentration remains stable. In the
presence of acetyl-CoA (produced by PFL), citrate synthase converts oxaloacetate to
citrate, which shifts the equilibrium to the right. The only source of acetyl-CoA is from
PFL catalysis, so the rate of NADH production could be converted to PF L activity.

The rate of production of NADH was recorded at 340 nm in an HP-8453 diode-
array spectrometer. The reaction is considered as a zero order reaction. The experiment

was conducted under anaerobic conditions at room temperature.

97

C oASH

     

 

 

 

 

Coupled Enzyme Mix:
Pyruvate 63 (“$0902-
C oA it

2
Malate g
NAD+ 2
Malic Dehydrogenase V HC‘O;
Citrate Synthase O

2 //
MK you (GAS—Q CoASH
Malate a ‘ Oxaloacetate . A Citrate
Malic Dehydrogenase Citrate Synthase

Figure V.1: Schematic representation of PFL activity assay. Activated PFL was mixed
with pyruvate, CoA, and two coupling enzymes. The rate of NADH production was
converted to PFL activity.

V.2.6 CD sample preparation

All procedures were carried out under strictly anaerobic conditions. The EPR
samples described previously were thawed after experiments. PFL samples (35.3 11L of
50 11M) were mixed with 64.7 uL of 10 mM phosphate buffer, pH 7.5 to make the final
concentration of PFL 3 mg/mL. Samples were then frozen in liquid N2. Each sample was
loaded into a 0.1 mm optical-path CD measuring cell immediately after it was thawed in
an anaerobic chamber (COY). The cells were sealed by rubber septa in the chamber, and
taken out for measurement

To subtract the CD absorption of PFL-AB in the PFL activation mix, background
solutions were prepared. The background solutions contained everything in the PFL

activation mix except PFL itself. They had the same concentration of PFL-AB as those

98

used to activate PFL, and were illuminated for the same time as for PFL activation. After
illumination, 35.3 uL of the solution was mixed with 64.7 uL of 10 mM phosphate buffer,

pH 7.5. The mixture was stored and measured the same way as PFL samples.

V.2.7 CD spectroscopy

CD spectra were measured by using a JASCO J-710 spectropolarimeter at room
temperature. The spectra were recorded in the wavelength range 170 — 300 nm. The scan
speed was 50 nm/min. The data pitch was 0.1 nm. The band width was 0.5 nm and 4
scans were accumulated. To keep anaerobic conditions during the measurement, the
spectrometer was purged with N2 gas at 60 L/min, and the measuring chamber was

purged with N2 from a separate line.

V.3 Results and discussion
V.3.1 Spin quantification of glycyl radical in activated PFL

PFL (50 11M) was activated for 2 h in the presence of 0.5, 5, 10, 20, 30, 40, 50,
and 100 11M of PFL-AB, as described in Materials and Methods. The EPR spectra of
these samples (having PFL to PFL-AE ratio of 100:1, 10:1, 5:1, 5:2, 5:3, 5:4, 1:1, and 1:2)
are shown in Fig V.2.A, and the spin quantifications are shown in Fig V.2.B and Table
V. 1. At higher PFL to PF L-AE ratios (100:1 and 10:1), the PFL is fully activated after 2 h
illumination, showing 50 (i5) and 49 (21:5) uM spin for 50 11M PFL respectively. The
smaller the PF L to PFL-AB ratio, the less glycyl radical was present after a 2 h activation.

These results were initially unexpected, because one would expect the most rapid and

99

 

complete activation to occur with higher ratios of catalyst (PFL-AB) to substrate.
Repeating this experiment, however, revealed that the results were very reproducible. We
thus began to suspect that these unexpected results might be telling us some important
information about the interaction of PF L-AE with its substrate PFL. We hypothesized
that when these two proteins interact, the glycine loop of PFL might be pulled out of the
core of the PFL structure, so that it is accessible to the PFL-AE active site. This would
leave the glycyl radical more exposed to solvent and more prone to quenching, resulting
in less glycyl radical being present after a 2 h illumination (activation) reaction. In this
model, the lower glycyl radical quantification at lower PFLzPFL-AE ratios reflects not
incomplete activation, but rather complete activation followed by faster quenching

process occurring in the presence of more PFL-AE.

100

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A W l : 2
W 1 : l
W 5 : 4
E W— 5 3 3
t:
8
E. 5 : 2
8‘.
Lu W 5 : I
10 : 1
. 1 r 1 a l l l l l
3200 3250 3300 3350 3400 3450 3500
Field (G)
60
B
50 ~
,7: .0 _ l
“O
E.’
'34
Q
.2.“
C3
30 _
20 l n l . J l l l
0 20 4O 60 80 100

Figure V.2: EPR spectra (top) and spin quantification (bottom) of 50 uM PFL activated
with different amounts of PFL-AE for 2 h. A. EPR spectra of 50 uM PFL activated by
0.5, 5, 10, 20, 30, 40, 50, and 100 11M of PFL-AE, 500 11M AdoMet, and 100 11M 5-
deazariboflavin for 2 h, in a buffer containing 100 mM Tris/HCl pH 7.6, 100 mM KCl,
10 mM oxamate, and 10 mM DTT. B. Spin quantification of the glycyl radicals in panel
A. EPR parameters: T = 60 K, microwave frequency 9.37 GHz, power 19 11W,
modulation amplitude 5 G, scan twice.

101

 

 

PFL : PFL-AE ratio [Gly-l (uM)
100:1 50
10:1 49
5:1 43
5:2 33
5:3 31
5:4 29
l :1 27
1:2 26

 

 

Table V.1: Spin quantification of glycyl radical in 50 11M PFL activated with 0.5, 5, 10,
20, 30, 40, 50, and 100 1.1M PFL-AB, 500 M AdoMet, and 100 11M 5-deazariboflavin in
a buffer containing 100 mM Tris/HCl pH 7.6, 100 mM KCl, 10 mM oxamate, and 10
mM DTT, after 2 h illumination.

In order to further examine this hypothesis, we illuminated 1:1 and 100:1
PFL:PFL-AE for times ranging from 0 to 120 min. Specifically, 50 11M PFL with either
0.5 11M or 50 11M PFL-AB was illuminated for 0, 20, 40, 60, 80, 100 and 120 minutes.
The spin quantifications of glycyl radicals in PFL activated by PFL-AE in a 100:1 ratio
increase during illumination (Fig V.3.A). The spin quantifications of glycyl radicals in
PFL activated by 1 equivalent of PFL-AB are maximum at the shortest time points but
decrease with longer illumination time (Fig V.3.B). The spin quantifications of the two
samples are listed in Table V.2. The results show that PFL is completely activated by

0.01 equivalent of PFL-AB in 100 minutes, and is stable up to 120 minutes. In the

102

presence of a 1:1 ratio of PFL-AB, however, the fully activated PFL present at 20
minutes is slowly inactivated during subsequent illumination.

The spin quantification demonstrated that a 1:1 ratio of PFL to PFL-AE resulted
in rapid and complete activation of PFL, as expected based on previous results. 6
However, the activated PFL was deactivated with continuing incubation. On the other
hand, the glycyl radical in PFL activated with 0.01 equivalents of PFL-AB was stable
after PFL was fully activated. The results rule out the possibility that the glycyl radicals
are destroyed by intense photo illumination, but are consistent with the hypothesis that
the presence of PFL-AE changes the environment of the glycyl radical loop, such that it
is more solvent exposed and thus more prone to quenching.

The PFL activated with 1 equivalent of PFL-AB for 2 h had 21 11M glycyl radical,
less than that shown in Fig V.2. The difference could be due to a slight difference in
experimental procedure. For the experiments in Fig V.3, all the samples were illuminated
in one preparation. After each time point, the sample was frozen in liquid N2, which took
approximately 2 — 3 minutes, while rest of the samples were kept in dark. Therefore, each

PFL sample was exposed to the solvent longer than the illumination time.

103

 

 

 

 

60
A
1-
P %
A 40 '—
2
E:
'5 30 —
”-6'
a I
5* 2o —
.2:-
O l-
10 _ i
0 — I
l l l n l L l n 1 n J a L
0 20 4O 60 80 100 120
Illumination time (min)
60

 

_B
50- f

A
O
1
l—I—l
H—l

 

 

 

2 E
3
E 30 r- i
.2 L
'3
‘—
">"s 20 P E
o
.23 ,
C3
10 -
0 - I
1 1 L 1 . 1 A 1 m 1 . I
O 20 4O 60 80 100 120

Illumination time (min)

Figure V.3: Spin quantification of glycyl radicals in PFL activated by different ratios of
PFL-AB as a function of time. PFL (50 1.1M) was illuminated with 0.5 11M (A) or 50 1.1M
(B) PFL-AE, 500 11M AdoMet, and 100 M 5-deazariboflavin in a buffer containing 100
mM Tris/HCl pH 7.6, 100 mM KCl, 10 mM oxamate, and 10 mM DTT. Samples were
activated for 0, 20, 40, 60, 80, 100, 120 minutes. EPR pararnaters: T = 60 K, microwave

frequency 9.37 GHz, power 19 11W, modulation amplitude 5 G, scan twice.

104

 

 

 

Illumination time (min) [Gly-] (11M)
PFL:PFL-AE (100:1) PFL:PFL-AE (1:1)

0 0 0

20 ll 52

40 25 42

60 43 37

80 47 33
100 51 28
120 51 21

 

 

 

Table V.2: Spin quantification of glycyl radical in 50 11M PFL activated by 0.5 and 50
11M PFL-AE, 500 11M AdoMet, and 100 11M 5-deazariboflavin in a buffer containing 100
mM Tris/HCl pH 7.6, 100 mM KCl, 10 mM oxamate, and 10 mM DTT.

In order to investigate the effect of AdoMet, PFL (50 uM) was activated for 2 h

by 50 11M PFL-AE with 500, 250, 200, 100, and 50 M AdoMet (Fig V.4, Table V3).
The results are similar to those in Fig V.2 and demonstrated that AdoMet does not
completely protect the glycyl radical from quenching. However, more radical is present
when there was more AdoMet present. This could be explained by AdoMet partially
protecting the glycyl radical from quenching or by reactivating of PFL after quenching by

reduced PFL-AE and excess AdoMet.

105

30

25-
L }

 

 

 

 

3‘20—
3;
— t
.8
E15-
3.. _ E
o
.3:
(D10...
51—
n 1 l .L L J J 1
0 2 4 6 8 1O
SAMzPFL

Figure V.4: Spin quantification of glycyl radicals in 50 11M PFL activated for 2 hours by
50 M of PFL-AB with 50, 100, 200, 250 and 500 11M of AdoMet, and 100 M 5-
deazariboflavin in a buffer containing 100 mM Tris/HCl pH 7.6, 100 mM KCl, 10 mM
oxamate, and 10 mM DTT. EPR parameters: T = 60 K, microwave frequency 9.37 GHz,
power 19 uW, modulation amplitude 5 G, scan twice.

 

 

AdoMet concentration (11M) [Gly-l (uM)
50 7
100 12
200 19
250 23
500 25

 

 

Table V.3: Spin quantification of glycyl radicals in 50 M PFL illuminated for 2 hours
by 50 M of PFL-AE with 50, 100, 200, 250 and 500 11M of AdoMet, and 100 M 5-
deazariboflavin in a buffer containing 100 mM Tris/HCl pH 7.6, 100 mM KCl, 10 mM
oxamate, and 10 mM DTT.

106

V.3.2 Hydrogen-deuterium exchange of activated PFL in D20

As previously reported, 2’ '7’ '8 the or-H on the Gly-734 radical can exchange with
solvent, as evidenced by the doublet EPR signal becoming a singlet in D20 (Fig V5).
The doublet EPR signal of the glycyl radical is due to the interaction between the radical
electronic spin (S = 1/2) and the glycine hygrogen nuclear spin (1 = 1/2). When the glycyl
hydrogen is replaced by a deuteron (I = l), the doublet signal becomes a singlet. The
active site residue Cys-419 was shown to be essential for the H/D exchange at G1y-734,
as the mutant C419A-PFL showed no exchange. ‘6 A similar Gly—734/Cys-4l9 radical
transfer is believed to occur in the PFL catalytic mechanism. 18’ ‘9’ 2° The H/D exchange
has been explained by the glycyl radical abstracting a hydrogen from Cys-4l9 to generate
a thiyl radical. Because the thiol hydrogen is exchangeable with solvent, the glycyl

radical will abstract either H (in H20) or D (in D20).

 

 

 

Activated PFL in H20 Activated PFL in D20
ii
.5 ,,__.,..__.—_._ 3
ad
0..
LL]

 

3250 3300 3350 3100 31503250 3300 ' 3350 3100 3150
Field (G) Field (G)

 

 

 

Figure V.5: EPR spectra of activated PFL in H20 and D20. PFL (50 11M) was activated
by 50 11M PFL-AB and 500 M AdoMet in either H2O or D20 buffer, containing 0.1 M
Tris-HCl pH 7.6 (in H2O) or pD 8.2 (in D20), 0.1 M KCl, 10 mM oxamate, 10 mM DTT,
100 11M 5-deazariboflavin. The spin quantification of glycyl radical of PFL in H2O is 39
11M, and the spin quantification of glycyl radical of PFL in D20 is 33 11M. EPR
parameters: T = 60 K, microwave frequency 9.37 GHz, power 19 11W, modulation
amplitude 5 G, scan twice.

107

Based on our previously stated hypothesis that interaction with PFL-AE perturbs
the position of the Gly-734 loop, we proposed that this position change might be reflected
in the H/D exchange results. If PFL-AB perturbs the position of the glycine loop, we
would expect less H/D exchange at higher PFL-AB to PFL ratios. We therefore fully
activated 100 11M PFL with either 1, 0.1 or 0.01 equivalents of PFL-AE and then mixed
the sample with D20 to examine HfD exchange by EPR. Fig V.6.A shows the change in
EPR signal of activated PFL mixed with one volume of D20 (final [D20] = 50%). The
PFL activated with 1 equivalent of PFL-AE has a partial doublet signal, which is
decreased in PFL activated with 0.1 and 0.01 equivalent of PFL-AB. Fig V.6.B shows the
change in EPR signal of activated PFL mixed with 3 volumes D20 (final [D20] = 75%).
The signal of PFL activated with 1 equivalent of PFL-AB is broader than those of PFL
with 0.1 and 0.01 equivalents of PFL-AB. The results demonstrate that glycyl on-
hydrogen exchange to deuteron is slower in PFL with 1:1 PFL-AB than that in PFL with
only 0.1 and 0.01 equivalents of PFL-AB, supporting our hypothesis that interaction with
PFL-AB pulled the glycine loop away from its position in the active site of PFL.

The EPR signal change of PFL diluted in 3 volumes of D20 was more significant
than the PF L diluted in 1 equivalent volume of D20, due to the greater extent of H/D

exchange in the former case.

108

 

 

 

 

 

 

 

 

 

 

 

 

 

A

.E‘

:9.

3 _.
E
at
G...
LL]

. 1 . 1 1 1 1
3300 3320 3340 3360 3380
Field (G)

.é‘

U)

c

.93

E.

0:

0.

DJ

1 1 1
3300 3320 3340 3360 3380

Field (G)

Figure V.6: EPR spectra of PFL activated with PFL-AE and then mixed with D20. In
both A and B, the solid line is 100 11M PFL activated by 100 11M PFL-AB for 30 min;
dashed line is 100 11M PFL activated by 10 11M PFL-AE for 2 h; dotted line is 100 11M
PFL activated by 1 11M PFL-AE for 2 h. The PFL activation mix contained 100 mM Tris-
HCl pH 7.6, 100 mM KCl, 500 11M AdoMet, 100 11M 5-deazariboflavin, 10 mM oxamate,
and 10 mM DTT. A. D20 buffer (150 11L, 100 mM Tris/HCl pD 8.2) was added into 150
11L of 100 11M activated PFL B. The D20 buffer (225 pL, 100 mM Tris/HCl pD 8.2) was
added into 75 11L of 100 11M activated PFL. EPR condition: T = 60 K, microwave
frequency 9.37 GHz, power 19 11W, modulation amplitude 5 G, scan twice.

109

V.3.3 PF L activity assay

In order to further examine our hypothesis that the position of the glycine loop
changes when PFL interacts with PFL-AB, the activity of PFL activated with 1, 0.1 or
0.01 equivalents of PFL-AE was investigated. It has been shown that Cys-418 and Cys-
419, both on another loop separated from the glycine loop, are required for PFL catalysis.
8’ 18’ 2' It has been proposed that the bond cleavage of pyruvate is initiated by a thiyl
radical on one of the cysteines, transferred from the glycyl radical. If the glycine loop is
pulled out of the core of PFL structure during the interaction with PFL-AB, the contact
between Gly-734 and Cys-419 (remaining in its position in the PFL active site) will be
lost, and the thiyl radical will not be generated. Therefore, in PFL activated with higher
ratio of PFL-AB, there might be less thiyl radical transfer to the Cys-419, and PFL would
have less activity accordingly. The experiments were designed such as to obtain maximal
PFL activation for each PF L to PFL-AE ratio. Thus 100 11M PFL was activated with 100
11M PFL-AE for 30 min, and in the two other samples 100 11M PFL was activated with
either 10 11M or 1 11M PFL-AB for 2 h. The spin quantifications of the glycyl radicals in
the activated PFL were 98 (PFL:PFL-AE 1:1), 105 (PFL:PFL-AE 10:1) and 102 11M
(PFL:PFL-AE 100:1). The specific activities of the three PFL samples were 6.6
(PFL:PFL-AE 1:1), 21.7 (PFL:PFL-AE 10:1), and 22.6 (PFL:PFL-AE 100:1) U/nmol
PFL (1U = 1 umol pyruvate tumover/min) respectively (Table V.4). Although the
samples have similar quantities of glycyl radical, the activity of PFL activated with 1
equivalent of PFL-AB, is only approximately 30% of the that observed for other two PFL
samples activated with 0.1 and 0.01 equivalents of PFL-AB. The data are consistent with

our hypothesis that more interaction between PFL and PFL-AB results in a greater

110

population of the PF L molecules having the glycine loop in the exposed position, which
reduces the contact between Gly-734 and Cys-419, generates less thiyl radical, and

results in less PFL activity.

 

PFL:PFL-AE Illumination (min) [Gly-] (11M) Activity (U/nmol PFL)

 

1:1 30 98 6.6
10:1 120 105 21.7
100:1 120 102 22.6

 

 

 

 

Tabel V.4: Activities of 100 11M PFL that was completely activated by 100, 10, and 1 uM
of PFL-AE, 500 11M AdoMet, and 100 11M 5-deazariboflavin in the buffer containing 100
mM Tris/HCl pH 7.6, 100 mM KCl, 10 mM oxamate, and 10 mM DTT.

V.3.4 CD spectra of PFL activated with different amounts of PFL-AB
To provide more evidence for the position change of PFL glycine loop during
interaction with PFL-AE, circular dichroism (CD) spectroscopy, a technique providing

information on g the secondary structures of proteins, 22

was used. The movement of Gly-
734, from the center of PFL into the active site of PFL-AB, requires a conformational
change of PFL, which might be observed in CD. Fig V.7 shows the CD spectra of PFL in
the activation mix with no PFL-AB in present, and PFL activated with 0.01 to 2
equivalents of PFL-AE (from bottom to top). Each spectrum is the result of taking the
spectrum of activated PF L with PFL-AE subtracted by the activation mix background (no
PFL in present) spectrum. The intensity of the negative peak at 209 nm, one of the

indicators of a-helical structure, 22 consecutively decreases as the amount of PFL-AB

increases. We were unable to obtain CD signal below 200 nm, the spectral region which

111

contains another a-helix indicator. The reason could be that the absorption from either
DTT, AdoMet, or 5-deazariboflavin saturated the detector.

The CD spectra demonstrated that PFL lost some of its a-helical structure during
interaction with PFL-AB. More PFL-AB might promote conversion of PFL to an
open .confonnation with the Gly loop removed from the PFL core, and thus results in

change of PFL secondary structure as observed in the CD spectra.

 

 

 

 

 

 

 

 

 

 

1 :2
*— : l
| Ar 5 : 4
‘ 5 2 3
g, —“r—A~— : 2
'c w A ‘ 5 : 1
E, 10 :1
3 ~— .__100 . l
g A no PF L-AE
g-
E
l
1 I 1 l l l 1 l 1 I 1 1
200 220 240 260 280 300 320

Wavelength (nm)

Figure V.7: CD spectra of 50 uM PFL activated by 0.5, 5, 10, 20, 30, 40, 50,100 uM of
PFL-AB, 500 1.1M AdoMet, and 100 11M 5-deazariboflavin in the buffer containing 100
mM Tris-HCl pH 7.6, 100 mM KCl, 10 mM oxamate, and 10 mM DTT for 2 h. Each line
was made from the spectrum of PFL (3mg/mL) in activation mix subtracted by the
background of the same activation mix in the absence of PFL. The ratios are between
PFL and PFL-AE. CD parameters: wavelength range 170 — 300 nm, scan speed 50
nm/min, data pitch 0.1 nm, band width 0.5 nm, 4 scans accumulated.

112

V.3.5 PF L structure analysis

The PFL crystal structure in the protein data bank (1H16) was used to analyze
which parts of PF L might change in conformation to move Gly-734 out from and back to
the core of PFL. Residues 710-759 comprise a domain of PFL containing the Gly loop as
well as 2 helixes before and after the loop; the sequence of this domain is highly
homologous to YfiD, a 14 kDa enzyme in E. coli that is able to resecue PFL activity after
PFL is damaged by 02. 23 Pymol and ViewerLite software were used to generate the
structural images for analysis (Fig V.8). There are several interesting H-bonds in the
domain. The H-bonds I751-R712 (6.9 A between a-carbons) and N740-P722 (7.4 A
between a-carbons) hold the helices before and after the glycine loop nearly parallel. Ile-
751 also communicates with Pro-722 through H-bonds I75l-Q747-F739-N740-P722,
which connects the two helixes and one end of the glycine loop (F739). Another route for
helix communication with the loop is via H-bonds R753-V732-Y735, in which the
distance between Arg-753 and Val-732 a-carbons is 9.9 A. The structure of the glycine
loop is maintained by four H-bonds, V732-Y735, I730=V737, and F739-L728, which put
Gly-734 at the tip of the loop pointing to Cys-418 and Cys-419. All these residues are
conserved in PFL from different organisms as well as in E. coli YfiD (Fig V.9). 23’ 24 We
propose that this C-terminal region might be the “switch” to turn on the activation
process. The movement might transfer the signal to the two helixes, which are then
straighter, and flip over to expose the glycine loop. Meanwhile, the change of the helixes
could be passed to the glycine loop by the H-bond chains starting from Arg-753 and Ile-

751. The loop might be pulled out by the two H-bond chains, and lose its structure if the

113

4 H-bonds, holding the loop structure, are broken in the interaction between PFL and

PFL-AE.

R712

 

Y735

Figure V.8: Structure of PFL active site taken from PFL crystal structure (PDB 1H16).
Dashed lines are hydrogen bonds. Solid lines are distances between or—carbons.

 

 

 

 

 

 

 

R712 P722 L728 V732 F739 0747 1751
EcPFL NREMLLDAMENPEKYPQLTIRVSGYAVRFNSLTKQQQDVITRTFTQSM
CPFL R P L----RV----Y F --Q -------- I--R
AfPFLZ R P L---RV---5 s Q -I--R
CCPFL R P 6---R6 Y F Q —I--R
EcYfiD R P L---RV----Y F Q I--R

 

 

Figure V.9: Sequence conservations of the C-terminal region in PFL from different
organisms and E. coli YfiD. The residues are numbered according to the E. coli PFL
sequence.

114

The position of the C-terminus appears to be fixed tightly in the absence of PFL-
AE. The C-terminus is inserted between two helixes from Gly-630 to Asp-690 (Fig V.10),
which hold the C-terminal region by five H-bonds, M759-V645, S75 8-A644, T756-V670,
T756-S668, and T754-D640. Binding of PFL-AE might break these H-bonds to release
the C-terminus and trigger the conformational change.

The two helixes before and after the glycine loop are on t0p of Gly-734 in the
PF L structure. To move the loop out of the protein, these two helices have to undergo a
significant structure change, as proposed previously. The sequence 700 — 709 (before the
helixes) and the sequence 750 — 759 (after the helixes) do not have real structure, but
their positions are tightly held by other two helixes (4 H-bonds E700-A657, E700-F656,
N706-Y665, and N706-F667 to fix the residues 700 - 710). The lack of secondary
structure in these two sequences could make the conformational change of the glycine
loop region easier. The hydrogen bonds between the sequences and other parts of the
protein might help to lock the correct position of the glycine loop after activation.

Our proposal is that the PFL activation process might be triggered by PFL-AB
binding, which could cause structural change in the C-terminal region, thus signaling the
conformational change of the helixes. The structural change could pull the glycine loop
out from PF L and render it accessible to the PFL-AE active site to generate the glycyl
radical. After activation, Gly-734 is re-buried in the PFL core, and the hydrogen bonds
between the two sequences (700 — 710 and 754 — 759) and other parts of the protein

might help PFL return to its activated structure.

115

  
  
 

C terminal

 

helix 2

A
>.
‘

Figure V.10: Crystal structure of PFL (PDB 1H16). Gly-734 is black and at the center of
PFL. Helix 1 is the structure after the glycine loop, and helix 2 is the structure before the
loop. The C-terminal is inserted into two other helixes.

V.4 Conclusions

The investigation reported in this chapter grew from the unexpected observation
that 100 11M PFL had 100 M glycyl radical (from EPR spin quanitificaiton) when it was
activated with 1 equivalent of PFL-AB for 30 min, and activated with 0.1 or 0.01

equivalents of PFL-AE for 2 h, whereas 100 11M PFL only had 30 11M glycyl radical if it

116

was activated with 1 equivalent of PFL-AE for 2 h. This unexpected result led us to
examine in determine the quantification of glycyl radical produced by different ratios of
PFL-AB to PFL when activated for varying lengths of time. The results showed that PFL
could be fully activated low ratios of PFL-AE during a 2 h activation. Higher ratios of
PFL-AE to PF L resulted in more rapid activation, however the protein lost its glycyl
radical during extended incubation. These results implied that the short loop containing
Gly-734 might be pulled out from the core of PFL as a result of interaction with PFL-AB,
and more PFL-AE would mean that more PFL is in this open conformation. This open
conformation would lead glycyl radical being quenched by the solution.

The (it-hydrogen on glycyl radical in activated PFL undergoes H/D exchange with
solvent D20 via Cys-4l9, 16 which is on an opposing loop to the glycine loop. This H/D
exchange was used to examine the hypothesis that interaction with PFL-AB produces a
more open conformation of PFL in which the Gly-734 is not in the core of PFL, and thus
not in contact with Cys-4l9. The results demonstrated that the H/D exchange was slower
in PFL with 1 equivalent of PFL-AB than that in PFL with 0.1 and 0.01 equivalents. PFL
catalysis also requires interaction between Gly-734 and two cysteines Cys-418 and Cys-
419; 6’ ‘6’ ‘9 we thus used PFL activity assays to further probe the position of the glycine
loop during the interaction between PFL and PFL-AB. The activity of 100 M PFL
completely activated with 100 11M PFL-AB was only 1/3 of 100 11M PFL fully activated
with 10 11M and 1 11M PFL-AE. These results support the hypothesis that Gly-734 in PFL
is in an alternate more open position during interaction with PFL-AE, but less interaction

with the cysteines, and thus less activity.

117

In order for the glycine loop to move out of the core of PFL, PFL has to undergo a
conformational change; CD spectroscopy has provided evidence for such a change. The
CD spectra demonstrated that more helical structure is lost in PFL upon interaction with
more PFL-AE. Inspection of the PF L crystal structure provides support for the idea that
the two (it—helixes before and after the glycine loop might change in structure to pull Gly-
734 out for activation. These two helixes could also be the site for binding to PFL-AE,
and the movement of the PFL C-terminus might trigger a PFL conformational change

through several H-bonds.

118

References:

1. Knappe, J .; Neugebauer, F. A.; Blaschkowski, H. P.; Gaenzler, M., Proceedings
of the National Academy of Sciences of the United States of America 1984, 81, (5), 1332-
5.

2. Unkrig, V.; Neugebauer, F. A.; Knappe, J., European Journal of Biochemistry
1989, 184, (3), 723-8.

3. Wagner, A. F. V.; Frey, M.; Neugebauer, F. A.; Schaefer, W.; Knappe, J.,
Proceedings of the National Academy of Sciences of the United States of America 1992,
89, (3), 996-1000.

4. Roedel, W.; Plaga, W.; Frank, R.; Knappe, J ., European Journal of Biochemistry
1988, 177, (1), 153-8.

5. Wagner, A. F. V.; Demand, J .; Schilling, G.; Pils, T.; Knappe, J ., Biochemical and
Biophysical Research Communications 1999, 254, (2), 306-310.

6. Henshaw, T. F.; Cheek, J.; Broderick, J. B., Journal of the American Chemical
Society 2000, 122, (34), 8331-8332.

7. Magnusson, O. T., Reed, G. H., Frey, P. A., JAm Chem Soc 1999, 121, 9764-
9765.

8. Becker, A.; Fritz-Wolf, K.; Kabsch, W.; Knappe, J.; Schultz, S.; Wagner, A. F. V.,
Nature Structural Biologt 1999, 6, (10), 969-975.

9. Becker, A.; Kabsch, W., Journal of Biological Chemistry 2002, 277, (42), 40036-
40042.

10. Lehtioe, L.; Leppaenen, V. M.; Kozarich, J. W.; Goldman, A., Acta
Crystallographica, Section D: Biological Crystallography 2002, D58, (12), 2209-2212.

11. Uhlin, U.; Eklund, H., Nature 1994, 370, (6490), 533-9.

12. Eklund, H.; Uhlin, U.; Famegardh, M.; Logan, D. T.; Nordlund, P., Prog Biophys
Mol Biol 2001, 77, (3), 177-268.

13. Stubbe, J.; Nocera, D. G.; Yee, C. S.; Chang, M. 0, Chem Rev 2003, 103, (6),
2167-201.

14. Frey, M.; Rothe, M.; Wagner, A. F. V.; Knappe, J., Journal of Biological
Chemistry 1994, 269, (17), 12432-7.

119

15. Aasa, R.; Vanngard, T., J Magn. Reson 1969, 19, (3), 308-15.

16. Knappe, J .; Blaschkowski, H. P.; Groebner, P.; Schmitt, T., European Journal of
Biochemistry 1974, 50, (1), 253-63.

17. Knappe, J.; Elbert, S.; Frey, M.; Wagner, A. F. V., Biochemical Society
Transactions 1993, 21, (3), 731-4.

18. Parast, C. V.; Wong, K. K.; Lewisch, S. A.; Kozarich, J. W.; Peisach, J.;
Magliozzo, R. 3., Biochemistry 1995, 34, (8), 2393-9.

19. Parast, C. V. Mechanistic studies of pyruvate formate-lyase: evidence for a
catalytic radical interchange between Gly 734, Cys 419 and Cys 418. 1996.

20. Plaga, W.; Frank, R.; Knappe, J ., European Journal of Biochemistry 1988, 178,
(2), 445-50.

21. Plaga, W.; Vielhaber, G.; Wallach, J .; Knappe, J ., FEBS Letters 2000, 466, (1),
45-48.

22. Greenfield, N. J ., Methods Mol Biol 2004, 261, 55-78.

23. Wagner, A. F. V.; Schultz, S.; Bomke, J.; Pils, T.; Lehmann, W. D.; Knappe, J.,
Biochemical and Biophysical Research Communications 2001, 285, (2), 456-462.

24. Lehtioe, L.; Goldman, A., Protein Engineering, Design & Selection 2004, 17, (6),
545-552.

120

CHAPTER VI

INVESTIGATIONS OF THE PYRUVATE FORMATE-LYASE CATALYTIC
MECHANISM

V1.1 Introduction

PFL catalyzes the reversible conversion of pyruvate and CoA into formate and
acetyl-CoA. 1 The mechanism proceeds via a radical route, in which a glycyl radical on
Gly-734 initiates the reaction. 2’ 3 Several lines of evidence combine to implicate two
cysteines, Cys-418 and Cys-419, as catalytic residues. Inactivation of PFL by
hypophosphite, an analogue of pyruvate, results in a l-hydroxyethyl phosphate covalently
bound to Cys-418. 4 Two PFL mutants, C418S-PFL and C419S-PFL, exhibit normal
glycyl radical signals, but no activity. 5 Cys-419 is also essential for the glycyl radical
H/D exchange in D20. 6 Therefore, Cys-418 and Cys-419 were suggested to be involved
in PFL catalysis.

Based on the crystal structure and other studies of PFL, two mechanisms have
been proposed by different research groups (scheme V1.1). 7'10 Both mechanisms involve
an initial glycyl radical that abstracts a hydrogen from Cys-419 to generate a thiyl radical.
The Kozarich mechanism proposes that this thiyl radical attacks the carbonyl carbon in
pyruvate, causing the C-C bond to be broken. The acetyl group thereafter attaches to Cys-
419, and the formate radical abstracts a hydrogen atom from Gly-734. The acetyl group is
then transferred to Cys-418. After formate leaves the active site, the thiol end of CoA

comes in and takes the acetyl group from Cys-418. Acetyl-CoA, the second product,

121

leaves the protein and PFL remains in the active form. In the Knappe mechanism, the
thiyl radical on Cys-419 is transferred to Cys-418 to yield another thiyl radical, which
then attacks pyruvate. The substrate is cleaved and the protein is acetylated directly on
Cys-418. After that, the reaction follows the same route as the Kozarich mechanism. The
difference between the two proposals is the role Cys-419 that plays in the mechanism,
whether it is a residue to initiate pyruvate cleavage (Kozarich mechanism) or just a

radical shuttle (Knappe mechanism).

 

 

 

 

 

 

 

 

CoA-S-COCl-l3 Co -511
0
CH SH SH
| l | ‘
G734 C419 C418
0C02' CH3
O\C/
. |
CH2 8 SH CH3C0C02 CH2 S SH CH
__-_—> 3
I i l I I I HC02' o§ /
0734 C419 C418 G734 C419 C418 T
' s
Kozarich Proposal TH TH
CH3 ,
ocoz' 0% / G734 C419 C418
0 HC02
([3142 TH S CH2COC02 ([142 TH T
—+>
G734 C419 C418 G734 C419 C418

Knappe Proposal
Scheme V1.1: Schematic representation of the PFL catalytic mechanism. Kozarich

proposed that Cys 419 attacks the substrate, while Knappe proposed that Cys 418
initiated the first step.

122

The structure of PFL is similar to the subunit 1 of class I ribonucleotide reductase
(RNR), which also uses a thiyl radical to initiate the reaction. H However, the first step in
class I RNR is that the thiyl radical abstracts a 3’-hydrogen from the nucleotide. ‘2 This
led us to question whether PFL catalysis might also involve a hydrogen abstraction from

the methyl group in pyruvate, rather than a nucleophilic attack by a thiyl radical (scheme

 

 

 

 

 

v1.2).
CH,
CH3COCO: 0 . K/
CH2C0C02 HCO' C
'l-l H H H H H 2' '
ill—“'31 1 1:4“ l”
0734 C419 C418 0734 C419 C418 0734 c419 c413

Scheme V1.2: Schematic representation of proposed PFL mechanism initiating with one
hydrogen abstraction.

We have used a deuterium labeled substrate, 3,3,3-d3-pyruvate to investigate the
possibility of H atom abstraction PFL catalysis. The reaction rate of PFL with this
substrate was compared with the rate of natural abundance pyruvate. The molecular
weight of the acetyl-CoA produced from the labeled pyruvate during the PFL enzymatic
reaction was determined by using mass spectrometry. To inspect which cysteine triggers
the reaction, the radical signals of two activated PFL mutants C418A-PFL and C419A- -

PFL in H20 and D20 were measured by EPR.

123

V1.2 Materials and methods
V1.2.1 Chemicals and materials

All chemicals were of the highest quality commercially available except as noted
otherwise. Overexpression and purification of PFL and PFL-AE were carried out
following the procedures in chapter 11. Protein and iron assays were conducted as

described in chapter 11.

VI.2.2 Synthesis of sodium 3,3,3-d3-pyruvate

Ethyl pyruvate-3,3,3-d3 was purchased from C/D/N isotopes INC. Sodium 3,3,3-
d3pyruvate was made from ethyl pyruvate-3,3,3-d3 according to a Japanese patent. '3
Ethyl pyruvate-3,3,3-d3 (0.3006 g) was dissolved in 100 mL H2O to make a 25 mM
solution. The solution was vigorously stirred in a 500 mL flask. NaOH (100 mL, 25 mM)
was added drop wise into the solution to maintain the pH at 6 to 8, at room temperature.
After the NaOH addition was complete, the reaction was further stirred for another 1 h.
The solvent was removed under reduced pressure. The sodium 3,3,3-d3-pyruvate was
crystallized by addition of excess isopropanol. The solid was flitered and washed with
isopropanol (3 x 50 mL). The yield was 0.2482 g (87.1%). When dissolved in D20 (10
mg per 800 uL D20), no hydrogen peak was detected in lH-NMR. The solid was stored

at 4 °C.
VI.2.3 PF L activity assay

All procedures were carried out under strictly anaerobic condition in the glove

box (Mbraun). PFL activation mix (500 1.1L) contained 100 mM Tris-HCl pH 7.6, 100

124

mM KCl, 8 mM DTT, 10 mM oxamate, 0.2 mM AdoMet, 50 uM 5-deazariboflavin, 1.5
11M PFL, and 0.015 11M PFL-AB. 5-Deazariboflavin was added last in the dark, and the
reaction was initiated via ambient light illumination, for 10 to 70 minutes. The mixture
was then wrapped with aluminum foil after each illumination time and IOuL of solution
was taken out for the PFL activity assay, and added to 790 uL of coupling enzyme mix
described in Chapter V. The PFL activity was tested following the procedure described
previously in Chapter V. The activity assay of activated PFL with 3,3,3-d3-pyruvate was
carried out following the same procedure except that the coupling enzyme mix contained

10 mM 3,3,3-d3-pyruvate instead of natural abundance pyruvate.

VI.2.4 PFL reaction with 3,3,3-dg-pyruvate

All procedures were carried out under strictly anaerobic conditions in the glove
box (Mbraun). The enzyme reaction mix (500 11L) was prepared, containing 100 mM
Tris-HCl pH 7.6, 100 mM KCl, 8 mM DTT, 10 mM sodium 3,3,3-d3-pyruvate, 0.5 mM
CoA tri-lithium salt, 2 mM AdoMet, 50 11M 5-Deazariboflavin, 6 11M PFL, and 0.06 11M
PF L-AE. The solution was illuminated by ambient light for 2 hours and then moved out
of the anaerobic chamber. It was exposed to air for 5 minutes on ice bath, shocked at
87°C for 30 seconds, and centrifuged at 12,000 rpm for l min, in order to cleave activated
PFL, denature the fragments and release the product. The clear solution was stored at —80

°C for mass spectrometry

125

VI.2.5 Mass spectrometry measurement of acetyl-CoA from PFL enzyme reaction

The mass spectra were recorded on a linear quadrupole ion trap mass spectrometer
(model LTQ, Thermo, San Jose, CA), equipped with a nanospray ionization (nanoESI)
source. Samples were introduced to the mass spectrometer at an infusion flow rate of 0.2
uL/min. The nanoESI conditions were optimized to maximize the intensity of the ion of
interest and to minimize the extent of in-source fragmentation. Spectra were acquired in a
negative ionization mode. MS conditions were: spray voltage 1.5 kV, heated capillary
temperature 200°C, capillary voltage —20 V, and tube lens voltage —75 V. MS/MS
spectra of singly and doubly deprotonated precursor ions were acquired on

monoisotopically isolated ions using standard isolation and excitation procedures.

V1.2.6 H/D exchange in mutants after reaction with pyruvate

All procedures were carried out under strictly anaerobic conditions in a glove box
(Mbraun). C418A-PFL and C419A-PFL were provided by J ian Yang. The activation mix
(200 uL), containing 100 mM Tris-HCl pH 7.6, 100 mM KCl, 200 11M PFL, 200 1.1M
PFL-AE, 200 11M 5-deazariboflavin, 500 1.1M AdoMet, 10 mM DTT and 10 mM oxamate,
was illuminated by a 500 W halogen light for 30 min in EPR tubes inserted in an
ice/water bath (4 °C). After illumination, 200 11M pyruvate (final concentration) was
added into each sample, which was then incubated for 10 min at room temperature. The
samples were split in half, and 300 1.1L of H20 or D20 (7 5% D20 final concentration) was
added into the samples, followed by incubation for 2 min at room temperature. The

samples were flash frozen in liquid N2 in the glove box for EPR.

126

VI.2.7 EPR spectroscopy

EPR spectra were recorded in a Bruker ER-200D-SRC spectrometer at 60 K. X-
band microwave frequency 9.37 GHz was used. The microwave power was 19 11W. The
external magnetic field was scanned from 3000 to 4000 G. The static field was 1000 G.
The modulation frequency was 100 kHz, and modulation amplitude was 5 G. The

resolution was 1024.

V1.3 Results and discussion
VI.3.1 PFL activity assay of sodium pyruvate and sodium 3,3,3-d3-pyruvate

In order to investigate the possibility of hydrogen atom abstraction as a key step
of PFL catalysis, the H/D isotope effect of 3,3,3-d3-pyruvate in PFL activity was
determined. The turnover rates of sodium pyruvate and sodium 3,3,3-d3-pyruvate in PFL
catalysis are provided in Table V1.1. The average H/D isotope effect on the tumover rate
was 1.1. The primary isotope effect (kn/kn) of a hydrogen abstraction reaction, which
involves carbon-hydrogen bond breaking, usually has a value of 6 or even higher. ‘4 For
example, the H/D isotope effect was found to be 7.4 in ethanolarnine deaminase, an
enzyme that utilizes a 5’-deoxyadenosyl radical generated fiom cobalamine coenzyme to
abstract a hydrogen atom from ethanolarnine. 15‘ ‘6 The small isotope effect observed for
3,3,3-d3-pyruvate tells us that C-H bond cleavage does not contribute to the rate-
deterrnine step of the reaction. It is therefore most likely that hydrogen abstraction is not

a step in pyruvate cleavage by PF L. The small isotope effect observed is in fact consistent

127

with a secondary H/D isotope effect, suggesting that the C-H bond cleavage is not

involved in the mechanism.

 

 

Illumination Time (min) pmol PYR‘/min pmol d3-PYRb/min PYR/dg-PYR
10 0.0019 0.0016 1. 2
15 0.052 0. 046 l. 1
20 0. 083 0. 080 1. 1
25 0.11 0.097 1.1
35 0.13 0.10 1. 2
40 0.14 0.13 1.1
50 0.18 0.15 1. 2
60 0.19 0.18 1.1
70 0.21 0.18 1. 2

Average 1. l

 

 

 

 

Table V1.1: Turnover rates of sodium pyruvate and 3,3,3-d3-pyruvate in PFL enzymatic
reaction for various illumination times. ‘ Sodium pyruvate. 1’ Sodium 3,3,3-d3-pyruvate.
V1.3.2 Mass spectra of reaction products

To further investigate whether hydrogen abstraction from pyruvate is a step in
PF L catalysis, mass spectra of the products from the PFL reaction with 3,3,3-d3-pyruvate
and CoA as substrates were acquired. If the radical form of PFL abstracts a deuteron from
3,3,3-d3-pyruvate, the pyruvate radical might subsequently abstract a hydrogen from Gly-
734, Cys-418 or Cys-419, or even from the solvent. Therefore, the product would have 1
dalton less than the d3-acetyl-CoA. The MS study indicated that one of the products after
a 2 h reaction PFL enzymatic reaction was 2’,2’,2’-d3-acetyl-C0A, with a molecular mass
of 811 ([M-H]', Fig VI.1.A). Fragrnentations of the deprotonated molecule are shown in
Fig VI.2.A. The mass of the fragments match the predicted degradation of 2’,2’,2’-d3-
acetyl-CoA (Fig VI.2.B). The results show that 2’,2’,2’-d3-acetyl-C0A was produced
fiom 3,3,3-d3-pyruvate and CoA, with all three deuterium labels remaining in the product.

No peak with one mass unit smaller than the molecule ion was detected (Fig VI.1.B),

128

providing no evidence for replacement of D by H of the final product. The results
corroborate the isotope effect studies and further support the proposal that hydrogen atom

abstraction is not a step in the PFL-catalyzed reaction.

A ”if?
100- MS of d3Ac-C‘oA ill H20 (after reaction)

‘ lM-Hl'
- 811

 

 

Relative Abundance
‘8
l

 

 

 

 

 

100 200 300 ' 400 '00 600 700 800 900 1600

miz
B
100_ 810.92
90- ll
80-

 

 

811.96

 

   

Relative Abundance
LII
5’

81292 .
81408 816.28

81524 816.92 81820

  
 

 

 

 

WWTWWWW
805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820
mlz

Figure V1.1: Electrospray mass spectra of 2’,2’,2’-d3-acetyl-C0A after 2 hour reaction. A:
Negative mode mass spectrum of 2’,2’,2’-d3-acetyl-C0A itself, [M-H]' at m/z 811 and
[M-2H]2' at m/z 405. B: Expanded m/z scale around [M-H]' peak. The intensity of the
peak at m/z 809.96, 1 mass unit less than the [M-H]', was no more than intensity
expected from the natural isotope distribution.

129

 

 

 

 

A 100 - 408 464
- MSt-MS of [M-Hj-
3 ..
g ' 426
a _
< _ _
p — 675 793
g - 482 713 731
g _ 272173 328 384 446 I 544 618
- 506 l
I ' i340 J 'n I 1 ‘11 I .
300 400 500 600 ’00 800 900
B 482
506
0 11 H 0H CH3 0

l
c—ii—s—(Crbl —1|1—c—(CH2]2—N—c —1:H—c —c O—P—
D3 2 I ”2

0 CH3 0

 

 

HEI—P=O

0H
Figure V1.2: Fragrnentations of 2’,2’,2’-d3-acetyl-C0A. A: MS/MS fragrnentations of the
deprotonated molecule [M-H]'. B: Structure of 2’,2’,2’-d3-acetyl-C0A. How bond
cleavage leading to observed fragments during CAD-MS/MS.
VI.3.3 Activity assay of C418A-PFL and C419A-PFL
To examine which cysteine residue, Cys—419 or Cys-418, initiates the reaction,
two PFL mutants, C418A-PFL and C419A-PFL, were prepared, and their activities were

tested to confirm the requirement of these two residues in PFL catalysis. Wild-type PFL,

C418A-PFL and C419A-PFL were activated for the same amount of time as described in

130

Materials and Methods. The resulting activities of the enzymes are provided in Table
V1.2. The activities of the two mutants are only approximately 1 — 2% of the wild type
PFL. The radical on Gly-734 has been proposed to transfer to Cys-419 to generate a thiyl
radical. 6 A mutation from cysteine to alanine would stop this radical transport, so the
mutant would not be expected to have activity. Cys-418 has been proposed as the
position for acetylation in the acetylated PFL intermediate, with CoA picking up the
acetyl group from this residue. 6’ 9’ 10 An alanine instead of a cysteine at this position
would prevent the acetylation of the enzyme and thereby prevent turnover. The small
amount of activity remaining in the two mutants likely comes from wild type PFL
contamination. The PFL gene in the host cells was not knocked out, so a small amount of
wild-type PFL from E. coli cells would still be produced during the growth of the two

mutants. The purification is not able to separate the wild type PFL from C418A-PFL and

 

 

C419A-PFL.
Illumination Time Wild PFL C418A-PFL C419A-PFL
(min) (umol PYR/min) (pmol PYR/min) (pmol PYR/min)
30 0. 16 0. 0010 0. 0024
40 0. 15 0. 0012 0. 0027
50 O. 17 0. 0013 0. 0033

 

 

 

 

Table V1.2: Activity assay of wild type PFL, C418A-PFL and C419A-PFL. The enzyme
was activated by photoreducing PFL-AB for 30, 40, 50 minutes. The activity of the
enzyme was represented by the turnover rate of pyruvate (umol/min).
VI.3.4 H/D exchange of PFL and PF L mutants afier reaction with pyruvate

A free Cys-419 has been shown to be required for H/D exchange. 6 A mechanism
has been proposed in which the a-hydrogen on the glycyl radical is exchanged with the

solvent deuterium via the labile thiol hydrogen on Cys-419. 6 We have used H/D

exchange in C419A-PFL and C418A-PFL to investigate which cysteine residue initiates

131

the reaction. If Cys-419 performs the initial attack on pyruvate, as proposed in the
Kozarich mechanism (Scheme V1.1), then Cys-419 would be acetylated upon reaction
with pyruvate. In the case of the C418A-PFL mutant, this initial reaction should still
occur, and the acetyl moiety should occupy Cys-419 since it could not be transferred to
Cys-418. In this mutant, therefore, Cys-419 would not undergo H/D exchange once it had
reacted with pyruvate. On the other hand, if the thiyl radical on Cys-418 performs the
initial attack on pyruvate, as proposed in the Knappe mechanism, then neither Cys-418
nor Cys-419 should be acetylated and the C418A-PFL should exhibit a singlet EPR
signal in D20. Our results support the mechanism put forward by Knappe. Activated
wild-type PFL exhibits a doublet EPR signal in H20 and a singlet in D20 (Fig V1.3.A) as
reported previously. 3’ 6 C419A-PFL exhibits a doublet glycyl radical EPR signal in both
H20 and D20 (Fig V1.3.B) as expected, since the cysteine was mutated to an alanine, and
thus no labile hydrogen was available for H/D exchange. Fig V1.3.C shows that after
reaction with, C418A-PFL has a doublet EPR signal in H20 and has a singlet EPR signal
in D20, indicating that Cys-419 is not acetylated and thus is able to transfer deuterium to
the glycyl radical. The result suggests that the fimction of Cys-419 is not to attack
pyruvate, but to transfer the radical to Cys-418, implying Cys-418 is likely the residue

that initiates the cleavage of pyruvate.

132

 

A H20 020

 

 

 

 

PFL

 

 

 

 

C419A-PFL

l

KW

4

ii

1

l

l

1

(')

 

 

 

C418A-PFL
\l
1

L__ J 1 l l I

3200 3250 3300 3350 3400 3450 3500 3200 3250 3300 3350 3400 3450 3500
Field (zauss) Field (sauss)

 

 

 

 

Figure V1.3: EPR spectra of PFL and PFL after reaction with pyruvate and addition of
D20. EPR condition: T = 60 K, microwave frequency 9.37 GHz, power 19 11W,
modulation amplitude 5 G, resolution 1024, scan twice. Left panel: protein in H20; right
panel: protein in D20. A: activated PFL. B: activated C419A-PFL. C: activated C418A-
PFL.

VI.4 Conclusions
PFL is the first enzyme that was shown to require a backbone glycyl radical to

initiate catalysis. 3‘ ‘7 The radical mechanism led us to propose that the reaction might be

133

triggered by hydrogen abstraction from pyruvate. To investigate this possibility, a
deuterium labeled substrate, 3,3,3-d3-pyruvate was ued. The reaction rates of activated
PFL with natural abundance pyruvate and deuterium labeled pyruvate as the substrates
were measured, allowing us to determine an isotope effect of 1.1. This value is much
smaller than a normal primary H/D isotope effect, which tells us that hydrogen atom
abstraction does not contribute to the rate limiting step of the reaction. The small isotope
effect is, however, consistent with a secondary isotope effect, suggesting the C-H bond
broken is not involved in PF L reaction.

The product of PFL enzymatic reaction with 3,3,3-d3-pyruvate and CoA as the
substrates was examined by using mass spectrometry. The production of 2’,2’,2’-d3-
acetyl-CoA was confirmed by its molecular weight ([M-H]’) at m/z 811, and the CAD-
MS/MS fragmentations of the deprotonated molecule. If the reaction was initiated by
hydrogen abstraction, there might be expected to be a D/H exchange on the final product,
which would have one mass unit less than the d3-1abeled product. This was not observed
in the mass spectrum of the product. This study again provided support for the idea that
the reaction was not initiated via hydrogen abstraction.

Two published mechanistic proposals suggest a thiyl radical at either Cys-418 or
Cys-419 attacks the carbonyl carbon in pyruvate to cleave the C—C bond. 6’ 8'10 Two PFL
mutants, C418A-PFL and C419A-PFL, were used to investigate these possibilities. As
previously reported, 5’ 7’ '8 due to the essential roles of these two cysteines in catalysis,
these mutants had very little enzymatic activity. 5’ 7’ '8 The small amount of activity is
likely due to wild-type PFL contamination, which is expressed in the host E. coli cells

and could not be separated from the mutants during purification. The small amount of

134

wild-type PF L would be activated in the activity assay, leading to observed activity in the
mutants.

A glycyl radical could be generated on both of the mutants in the presence of
reduced PF L-AE and AdoMet, and the activated mutants showed doublet EPR signals in
H20 after reacting with pyruvate as the substrate. The doublet EPR signal in wild-type
PFL changes to a singlet signal in D20, to the solvent H/D exchange on glycyl radical
mediated by Cys-419. 6 This result was re-produced in our lab, and H/D exchange in
pyruvate-reacted C418A-PFL and C419A—PFL was used to investigate which cysteine
residue initiates the attack on pyruvate. The activated and pyruvate-reacted C419A-PFL
remained a doublet signal in D20, which was expected because the Cys-419 mediates the
IUD exchange at Gly-734. The doublet EPR signal for activated C418A-PFL reacted with
pyruvate changed to a singlet signal in D20, indicating that Cys-419 in this mutant was
free to exchange the glycyl a-hydorgen with solvent deuterium. The results show that
Cys-419 is not occupied by the acetyl group as the product from the pyruvate cleavage.
These results suggest that Cys-419 in C418A-PFL does not attack pyruvate, although the
mutant could be activated to generate glycyl radical. Since a thiyl radical has been
proposed to initiate the pyruvate cleavage, the results imply that Cys-418 is likely the '

residue that initiates the first step of the reaction.

135

References

1. Knappe, J.; Blaschkowski, H. P.; Groebner, P.; Schmitt, T., European Journal of
Biochemistry 1974, 50, (1), 253-63.

2. Unkrig, V.; Neugebauer, F. A.; Knappe, J ., European Journal of Biochemistry
1989, 184, (3), 723-8.

3. Wagner, A. F. V.; Frey, M.; Neugebauer, F. A.; Schaefer, W.; Knappe, J.,
Proceedings of the National Academy of Sciences of the United States of America 1992,
89, (3), 996-1000.

4. Plaga, W.; Frank, R.; Knappe, J ., European Journal of Biochemistry 1988, 178,
(2), 445-50.

5. Knappe, J.; Elbert, S.; Frey, M.; Wagner, A. F. V., Biochemical Society
Transactions 1993, 21, (3), 731-4.

6. Parast, C. V.; Wong, K. K.; Lewisch, S. A.; Kozarich, J. W.; Peisach, J.;
Magliozzo, R. S., Biochemistry 1995, 34, (8), 2393-9.

7. Brush, E. J .; Kozarich, J. W., Enzymes (3rd Ed.) 1992, 20, 317-403.

8. Lehtioe, L.; Leppaenen, V. M.; Kozarich, J. W.; Goldman, A., Acta
Crystallographica, Section D: Biological Crystallography 2002, D58, (12), 2209-2212.

9. Becker, A.; Fritz-Wolf, K.; Kabsch, W.; Knappe, J.; Schultz, S.; Wagner, A. F. V.,
Nature Structural Biology 1999, 6, (10), 969-975.

10. Becker, A.; Kabsch, W., Journal of Biological Chemistry 2002, 277, (42), 40036-
40042.

11. Uhlin, U.; Eklund, H., Nature 1994, 370, (6490), 533-9.

12. Stubbe, J .; van der Donk, W. A., Chemical Reviews (Washington, D. C.) 1998, 98,
(2), 705-762.

13. Synthesis of sodium pyruvate. JP 63-132856, 6/4/1988, 1988.
14. Westheimer, F. H., Chem Rev 1961, 61, (3), 265-273.
15. Babior, B. M., J Biol Chem 1969, 244, (2), 449-56.

16. Weisblat, D. A.; Babior, B. M., J Biol Chem 1971, 246, (19), 6064-71.

136

17. Knappe, J .; Neugebauer, F. A.; Blaschkowski, H. P.; Gaenzler, M., Proceedings

0 the National Academy of Sciences of the United States of America 1984, 81, (5), 1332-

5
18. Brush, E. J .; Lipsett, K. A.; Kozarich, J. W., Biochemistry 1988, 27, (6), 2217-22.

137

CHAPTER VII

GENERAL CONCLUSIONS AND FUTURE DIRECTIONS

VII.1 The interaction between AdoMet and the iron-sulfur cluster in PFL-AE
Iron-sulfur cluster proteins have been identified in almost all types of organisms. I
The chemical versatility of iron and sulfur makes iron-sulfur proteins great candidates for
electron transfer. 2 The enzymes also play catalytic roles, both redox (e.g. nitrogenase,
hydrogenase), 3’ 4 and non-redox (e. g. aconitase). 5 E. coli Endonuclease III is an iron-
sulfirr cluster protein, in which the cluster appears to have neither an electron transfer nor
a catalytic role. 6 Other functions of iron-sulfur cluster proteins include serving as sensors
of iron, dioxygen, superoxide and possibly nitric oxide. 7'10 One of the intriguing roles for
iron-sulfur cluster proteins is the reductive cleavage of AdoMet to generate the putative
deoxyadenosyl radical. ”"3 AdoMet is cleaved at the C5’-S bond (~ 60 kcal mol") H;
how the reduced enzyme injects one electron into AdoMet to yield the putative radical is
an interesting question.

Sulfur pre-K-edge XAS can provide information on the Fe-S bond electronic
structure. We expect that the presence of AdoMet might change the covalency of the F e-
S bonds of the cluster, due largely to the proximity of the sulfonium-sulfur. We use sulfur
K-edge XAS of purified PFL-AB to examine this possibility. The PFL-AE used in these

studies was purified in the presence of ascorbic acid (to provide a reducing environment

without exogenous thiols), so we first used UV-Vis and EPR spectroscopy to determine

138

whether this protein had similar properties to PFL-AB purified in the presence of DTT
(our standard purification conditions). The UV-Vis spectrum of the protein purified in the
presence of ascorbic acid shows a broad band from 360 nm, with a shoulder from 410 nm,
a distinctive feature for [4F e-4S]2+ cluster. The spectrum is very similar to that of PFL-
AE purified in 1 mM DTT, which has been extensively characterized by chemical and
spectroscopic methods. The reduced [4Fe-48]” cluster shows (g = 2.01, 1.92 and 1.87),
which is also similar to the results (g = 2.02, 1.94, 1.88) for PFL-AE purified in 1 mM
DTT. '4 In the presence of 3 equivalents of AdoMet, the EPR signal undergoes a
significant change, with the g factors shifted to 2.00, 1.92 and 1.89. The change is similar
to that observed for the PFL-AE purified in DTT and suggests that AdoMet directly
coordinates the iron-sulfur cluster in a manner consistent with our previous reports

The effect of external thiolate on the cluster was examined by photoreducing the
enzyme in the presence of 10 mM DTT or 10 mM cysteine. The reduction results in an
axial signal with g-factors of g = 2.03 and 1.93 for DTT, and g = 2.00 and 1.93 for
cysteine. The results suggest a more symmetric structure of the cluster in the presence of
these thiols, implying that the exogenous thiolate might coordinate to the cluster to
generate a cluster with four thiolate ligands. Addition of AdoMet to these samples results
in a change in g-factors to g = 2.00, 1.92 and 1.89, suggesting that AdoMet binds to the
unique iron in the cluster, replacing the putative thiolate ligand. Together, the EPR and
UV-Vis data indicate that PFL-AB purified with ascorbic acid as reductant should be the
same as the PFL-AB on which we have previously published.

The sulfur K-edge XAS spectra of PFL-AE and PFL-AE with AdoMet show a

sharp rising feature at 2473.5 eV which is the sulfur K-edge absorption due to S Is to C-S

139

0* transition, and a broad pre-edge feature at 2470.3 eV due to Is (u3-Ssulfide and
terminal-Sthiolm) to Fegd unoccupied antibonding orbital transition. The intensity of the
pre-edge increases upon AdoMet binding, indicating an increase in covalency of the F e-S
bonds. The corresponding DFT calculations demonstrate that the total covalency of the
bonds increases 37%, in which 12% is from the bridging sulfide and 25% is fi'om the
terminal thiolate. A DFT optimized model shows that the AdoMet 5’C-Ssu1romum+ 0* bond
has a backbonding interaction with the filled Fe-S bond in the cluster, which shifts the
charge density from the cluster to the 5’C-S+ bond. This effect makes the sulfides poorer
donors to the Fe-S bond, and the terminal thiolates increase donation to compensate the
change. Therefore, the pre-edge feature has the characteristics of a S 1s electron to R3,:
antibonding orbital transition and the bridging sulfide to AdoMet 5’C-S+ 6* bond
transition. The charge density shifis from the cluster to AdoMet might be thought of as a

precursor to electron transfer from the cluster to the substrate.

VII.2 The vibrations of the iron-sulfur cluster in PFL-AB

Vibrational spectroscopy is an important tool to investigate the structures of iron-
sulfur clusters. There are many resonance Raman studies of Fe4S4 cluster model
compounds and proteins. 15 The resonance Raman spectra of PFL-AB in this thesis
showed a typical [4F e-4S]2+ signal, with Fe-Sb stretching modes at 290 cm'1 and 320 cm",
a Fe4S4 cluster cubane symmetric breathing mode at 338 cm", a Fe-St stretching mode at
370 cm], and a F e-St symmetric stretching mode at 396 cm". ‘6‘” Upon AdoMet binding,

the band at 396 cm'1 split into two bands at 391 and 401 cm", and the relative intensity of

140

the bands at 338 cm”l and 370 cm'1 was also changed. These changes indicate that the
coordination of AdoMet to the unique Fe causes some changes in the bonding between
the irons and terminal sulfurs.

We have proposed an oxygen from either H20 or 0H coordinates to the unique
Fe in the absence of AdoMet, however no clear evidence for this was obtained from
resonance Raman spectra. To investigate the proposal, nuclear resonance vibrational
spectroscopy (NRVS), was used. 57F e was reconstituted into the unique site of PFL-AB
after the cluster was oxidized to [3Fe-4S]1+ state, however Mossbauer spectroscopy
showed that 68% of the ”Fe was scrambled into other sites of the cluster and/or at the
unique site with no AdoMet bound. Due to this scrambling, we were not able to obtain
good NRVS data on unique-site-labeled PFL-AB.

The vibrational modes of the 57Fe labeled [4Fe-4S]2+ cluster were obtained by
NRVS from 0 — 660 cm". A broad band comprising the cubane twist modes was
observed below 100 cm", and a strong band of the S-Fe-S’ bending mode was observed
at around 160 cm]. 20 The F e-S stretching modes were detected as a doublet between 250
and 300 cm", and a triplet between 340 and 390 cm". 20 The Fe-O stretching mode night
be one of the bands above 400 cm". In the presence of AdoMet, the intensity of the Fe-
ligand + F e-St stretching mode at 432 cm'1 increased. The coordination of the amino N
and carboxylate O from AdoMet to the unique Fe enhanced the F e-ligand vibration. The
intensity of the bands at 388 cm'1 (Fe-St stretching), 158 cm'1 (S-Fe-S’ bending)
increased, and the Fe-St stretching mode at 278 cm'l was split in the presence of AdoMet.
Both the NRVS and resonance Raman results provide evidence for AdoMet binding

changing the Fe-St bonds, which coordinates the cluster to the protein. Neither techniques,

141

however, has allowed the identification of the exogenous ligand in the absence of

AdoMet.

VII.3 PF L structure change during activation

Our results in Chapter V demonstrate that the presence of PFL-AB results in a
lowered stability of the PFL glycyl radical. Several lines of evidence point to a change in
PFL conformation to a more open structure in the presence of PFL-AB, in which the
gylcyl radical is more solvent exposed, as the reason for the decreased stability. First,
with stoichiometric quantities of PFL-AB present, the H/D exchange at Gly-734 is
diminished, suggesting that there is less direct interaction between Gly-734 and Cys-419
at the active site. Second, fully activated PFL (as measured by glycyl radical content) has
lower activity in the presence of stoichiometric PFL-AB than in the presence of catalytic
amounts of PFL-AB, again suggesting that PFL-AE draws the Gly-734 loop out of the
active site. Finally, CD spectra changes of PFL in the presence of PFL-AB are consistent

with movement of the glycyl loop upon interaction with PFL-AB.

V11.4 The first step in PFL catalysis

It was proposed in our lab that PFL might initiate catalysis by hydrogen atom
abstraction from the methyl group in pyruvate. The H/D isotope effect on pyruvate
turnover (using 3,3,3-d3-pyruvate) was much smaller (1:1) than a primary H/D isotope

effect in hydrogen abstraction reaction, which should be 6 or even larger. 21 The small

142

isotope effect might be due to a secondary H/D isotope effect on the C-C bond broken in
pyruvate.

The product of the PF L enzymatic reaction with 3,3,3-d3-pyruvate and CoA as the
substrates was confirmed as 2’,2’,2’-d3-acetyl-C0A by mass spectrometry. It had a
molecular weight at 811 ([M-H]'), with fragrnentations in CAD-MS/MS that could be
interpreted as the degradation of the deprotonated molecule. If PFL catalysis involved
hydrogen abstraction, a hydrogen could be taken from the solvent or the protein into the
pyruvate radical. There was no deuterium to hydrogen exchange detected by mass
spectrometry, which should have the molecular weight at 810. Combined with the KIE
study of the deuterium labeled substrate, hydrogen abstraction could be ruled out as the
first step in the PFL mechanism.

PFL catalysis has been proposed to involve a thiyl radical, either on Cys-418 or
Cys-419, attacking the carbonyl carbon in pyruvate. 22’ 23 The two mutants C418A-PFL
and C419A-PF L, exihibit no catalytic activity, although a glycyl radical could be
generated in both of the proteins. In the presence of pyruvate, the activated C419A-PFL
had no H/D exchange at the glycyl radical, whereas the activated C418A-PFL did
exchange. The results show that the Cys-419 is not acetylated by pyruvate in activated
C418A-PFL, as a free thiyl group on Cys-419 is required for the exchange. The results
imply that Cys-419 does not attack the substrate. Our results support the proposal that a

key step in the PFL mechanism is a Cys-418 thiyl radical attacking pyruvate.

143

VII.5 Overview of PFL-AE and PFL catalytic mechanisms

A Fe4S4 cluster with three irons coordinated by three cysteines is the catalytic
center of PFL-AE. A hydroxide ion might bind to the unique Fe of the cluster in the
absence of AdoMet. Upon AdoMet binding, the cluster shifts some charge density to
AdoMet 5’C-S+ 0* bond, the position where AdoMet is reductively cleaved. This might
be the initial step for the cluster to transfer one electron to AdoMet. The produced
putative deoxyadenosyl radical abstracts one hydrogen from Gly-734 which we believe is
pulled out from PFL to bind the PF L-AE active site for activation. The radical would be
passed to Cys-418, using Cys-419 as the radical shuttle. The thiyl radical then attacks

pyruvate to initiate PFL catalysis.

VII.6 Future work on sulfur XAS of reduced PF L-AE

The current sulfur XAS spectra are from [4Fe-4S]2+, which is not the active state
of PFL-AB. It will be interesting to see the XAS data from the reduced [4Fe-4S]1+ cluster.
The spectral change of the protein with and without AdoMet can provide more

information about how the electron is transferred from the cluster to the substrate. One

problem might be the slow cleavage of AdoMet by PFL-AE.

VII.7 Future work on NRVS studies of PFL-AB in HzlsO
The Fe-ISO stretching mode should have a down-shift in wavenumbers compared

to the Fe-mO stretching if the exogenous ligand to the unique iron is from the solvent.

144

The comparison between the experimental data and calculated vibrational frequency

might provide the clue of which one, the OH' or the H20, is the actual ligand.

VII.8 Future work on the PFL and PFL-AB interaction

The current evidence of the PFL conformational change during activation is
indirect. The structures of PFL for PFL-AB binding and the hydrogen bond chains to
transfer the activation signal are both proposals. The best way to examine the idea might
be co-crystallize PFL, PF L-AE and AdoMet together. A successfirl experiment can tell
exactly how PF L and PFL-AB contact with each other. Site-directed mutagenesis of the
key residues, supposed to be involved in the conformational change transfer, can be used

to test the proposal.

VII.9 Future work on the investigation of the first step in PFL catalysis

The evidence of the thiyl radical on Cys-418 attacking pyruvate is also indirect.
More convincing evidence will be a direct observation of a thiyl radical on the residue,
which is not transferred from Cys-419, and a confirmation that this radical is able to

attack pyruvate and get acetylated.

145

 

References:

l. Beinert, H.; Holm, R. H.; Munck, B., Science (Washington, D. C.) 1997, 277,
(5326), 653-659.

2. Lovenberg, W., Iron-Sulfur Proteins. Academic Press: New York, 1973; Vol. I
and H.

3. Burgess, B. K.; Lowe, D. J ., Chem Rev 1996, 96, (7), 2983-3012.
4. Vignais, P. M.; Colbeau, A., Curr Issues Mol Biol 2004, 6, (2), 159-88.

5. Beinert, H.; Kennedy, M. C.; Stout, C. B., Chemical Reviews (Washington, D. C.)
1996, 96, (7), 2335-2373.

6. Thayer, M. M.; Ahem, H.; Xing, D.; Cunningham, R. P.; Tainer, J. A., Embo J
1995, 14, (16), 4108-20.

7. Hentze, M. W.; Kuhn, L. C., Proc Natl Acad Sci U S A 1996, 93, (16), 8175-82.
8. Beinert, H.; Kiley, P., FEBS Letters 1996, 390, (2), 239.
9. Gaudu, P.; Weiss, B., Proc Natl Acad Sci U S A 1996, 93, (19), 10094-8.

10. Hidalgo, E.; Bollinger, J. M., Jr.; Bradley, T. M.; Walsh, C. T.; Demple, B., J Biol
Chem 1995, 270, (36), 20908-14.

ll. Frey, P. A.; Magnusson, O. T., Chemical Reviews (Washington, DC, United
States) 2003, 103, (6), 2129-2148.

12. Sofia, H. J .; Chen, G.; Hetzler, B. G.; Reyes-Spindola, J. P.; Miller, N. E., Nucleic
Acids Research 2001, 29, (5), 1097-1106.

13. Cheek, J.; Broderick, J. B., Journal of biological inorganic chemistry .° JBIC .° a
publication of the Society of Biological Inorganic Chemistry 2001, 6, (3), 209-26.

14. Walsby, C. J .; Ortillo, D.; Yang, J .; Nnyepi, M. R.; Broderick, W. E.; Hoffman, B.
M.; Broderick, J. B., Inorganic Chemistry 2005, 44, (4), 727-741.

15 . Spiro, T. G. C., Roman S., Resonance Raman spectroscopy. Physical Methods in
Bioinorganic Chemistry. University Science Books: Sausalito, CA, 2000; p pp 97 - 112.

146

16. Broderick, J. B.; Duderstadt, R. E.; Fernandez, D. C.; Wojtuszewski, K.; Henshaw,
T. F.; Johnson, M. K., Journal of the American Chemical Society 1997, 119, (31), 7396-
7397.

17. Duin, E. C.; Lafferty, M. E.; Crouse, B. R.; Allen, R. M.; Sanyal, 1.; Flint, D. H.;
Johnson, M. K., Biochemistry 1997, 36, (39), 11811-11820.

18. Roman S. Czernuszewicz, K. A. M., Michael K. Johnson, Andrew Gewirth,
Thomas G. Spiro, Journal of American Chemical Society 1987, 109, (23), 7178-87.

19. Johnson, M. K.; Czernuszewicz, R. S.; Spiro, T. G.; Ramsay, R. R.; Singer, T. P.,
J Biol Chem 1983, 258, (21), 12771-4.

20. Xiao, Y.; Koutrnos, M.; Case, D. A.; Coucouvanis, D.; Wang, H.; Cramer, S. P.,
Dalton Trans 2006, (18), 2192-201.

21. Westheimer, F. H., Chem Rev 1961, 61, (3), 265-273.

22. Becker, A.; Fritz-Wolf, K.; Kabsch, W.; Knappe, J .; Schultz, S.; Wagner, A. F. V.,
Nature Structural Biology 1999, 6, (10), 969-975.

23. Parast, C. V.; Wong, K. K.; Lewisch, S. A.; Kozarich, J. W.; Peisach, J.;
Magliozzo, R. 8., Biochemistry 1995, 34, (8), 2393-9.

147

      
 

    

    
  

M

will

 

”Ill/ll!”

987

111111

56

. 1111117111

 

IE ......

   

. - .....