l" ‘\'

0 N

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

THE ELECTRON PARAMAGNETIC RESONANCE

CHARACTERIZATION OF FERROUS NITROSYL NON-

HEME IRON MODELS AND ENZYMES

presented by

Matthew D. Krzyaniak

has been accepted towards fulfillment
of the requirements for the

Ph.D degree in Chemistry

 

 

M‘fiW/AM

Major Professor’s Signature

W30,10|D
a ,

Date

MSU is an Afiirmative Action/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:IProj/Acc&Pres/ClRC/Dateoue.indd

 

 

 

THE ELECTRON PARAMAGNETIC RESONANCE CHARACTERIZATION OF
FERROUS NITROSYL NON-HEME IRON MODELS AND ENZYMES

By

Matthew D. Krzyaniak

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Chemistry

2010

ABSTRACT

THE ELECTRON PARAMAGNETIC RESONANCE CHARACTERIZATION OF
FERROUS NITROSYL NON-HEME IRON MODELS AND ENZYMES

By

Matthew D. Krzyaniak

Non-heme iron oxygenases represent a class of enzyme that activate dioxygen and
catalyze a diverse range of thermodynamically challenging reactions. Phenylalanine
hydroxylase(PheH) is a non-heme Fe enzyme that catalyzes the hydroxylation of the C4
position of the phenol side chain of phenylalanine(phe) to yield tyrosine and Tyrosine
hydroxylase(TyrH) catalyzes the hydroxylation of the C3 position of the phenol side chain
of tyrosine(tyr) to yield L-DOPA. These two enzymes contain a highly conserved iron
binding motif termed the facial triad with 2 histidine ligands and a carboxylate facially
coordinating the iron. This work focuses on the EPR characterization of these two
enzymes and a series of model complexes containing the same N, N, O facial binding
motif. The catalytically relevant oxidation state for non-heme iron oxygenases is Fe(II)
which is EPR silent, this problem was overcome by studying Fe(II)-NO complexes where
the NO acts as an oxygen surrogate and couples to the Fe(II) generating an EPR active
S=3/2 system. Using 2H electron spin echo envelope modulation(ESEEM) spectroscopy
we have shown that in TyrH and PheH there is a structural change that occurs in the
active site upon substrate addition which results in movement of the pterin co—susbtrate
closer to the F e(II) center. 2H ESEEM has also shown that in TyrH and PheH a change in

the orientation of the Fe-NO bond with respect primary substrate occurs on co-substrate

binding. Hyperfine sub-level correlation spectroscopy was used to characterize the
magnetic environment of the model complexes including the chelating ligand and solvent
molecules. These results were applied to aid in the understanding of the HYSCORE

spectra collected on TyrH and PheH.

ACKNOWLEDGEMENTS

First and foremost I'd like thank my collaborators without whom none of this
work would have been possible. My work on phenylalanine hydroxylase was in
collaboration with Giilbenk Anarat, a graduate student in the lab of John Caradonna at
Boston University. The non-heme iron model system was provided by Patrick Cappillino
another graduate student in the Caradonna lab. The work on tyrosine hydroxylase was in
collaboration with Bekir Eser, a graduate student in the lab of Paul Fitzpatrick at Texas
A&M. It was through these collaborations that I was able to pursue my real interests in
spectroscopy while maintaining relatively dry hands and avoiding smelly biochem labs.

I'd also like to thank my advisor John McCracken for patiently teaching me about
EPR theory and instrumentation. It has been a pleasure working for/with him and I am
thankful for all the knowledge and entertainment he has provided me. He has this special
talent to spin the most mundane of things into an entertaining story.

Also thanks to all the friends I've made along the way at MSU, with a special
shout out to the current and former members of the Weliky and McCusker groups. You
helped me to maintain my sanity through grad school.

I need to put a special thanks in here to my former undergraduate mentors, Dr.
Yiwei Deng and Dr. Sheila Smith of the University of Michigan — Dearborn. It was
through the work in your labs that motivated me to pursue a graduate education.

Lastly my family, your support and encouragement has kept me going through

school.

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. vii
LIST OF FIGURES .......................................................................................................... viii
LIST OF ABBREVIATIONS .............................................................................................. xi
CHAPTER I
1.1 Oxygen Activation .............................................................................................. 1
1.2 N,N,O Facial Triad ............................................................................................. 3
1.3 Structure of Tetrahydrobiopterin Dependent Hydroxylases ............................... 5
1.4 Mechanism ......................................................................................................... 7
1.5 Fe-NO as a Spectroscopic Probe ........................................................................ 8
1.6 Purpose ............................................................................................................... 9
1.7 Experimental ...................................................................................................... 9
1.7.1 TyrH Preparation ................................................................................. 9
1.7.2 PheH Preparation ............................................................................... 12
1.7.3 CW-EPR ............................................................................................ 16
1.8 Theory ............................................................................................................... 16
1.9 Results .............................................................................................................. 18
1.10 Simulation Notes ............................................................................................ 23
1.11 Discussion ...................................................................................................... 26
1.12 Conclusions .................................................................................................... 32
1.13 References ...................................................................................................... 33
CHAPTER II
2.1 Introduction ...................................................................................................... 38
2.2 Fe-NO Electronic Structure .............................................................................. 38
2.3 N,N,O Model System ....................................................................................... 42
2.4 Water Analysis .................................................................................................. 43
2.5 Hydrogen Hyperfine Coupling ......................................................................... 44
2.6 Experimental .................................................................................................... 47
2.7 Results .............................................................................................................. 49
2.7.1 Ferric CW—EPR .................................................................................. 49
2.7.2 {FeNO}7 CW—EPR ............................................................................ 51
2.7.3 {FeNO}7 HYSCORE ........................................................................ 52
2.7.4 Water Analysis ................................................................................... 56
2.7.5 Ligand Hyperfine Analysis ................................................................ 61
2.8 Discussion ........................................................................................................ 62
2.9 Conclusions ...................................................................................................... 68
2.10 References ...................................................................................................... 70
CHAPTER III
3.1 Introduction ...................................................................................................... 73
3.2 Experimental .................................................................................................... 76
3.3 Results .............................................................................................................. 80

3.4 2H ESEEM Simulation Notes .......................................................................... 85

3.5 Discussion ........................................................................................................ 88

3.6 Conclusions ...................................................................................................... 95

3.7 References ........................................................................................................ 97
CHAPTER IV

4.1 Introduction .................................................................................................... 100

4.2 HYSCORE Results ........................................................................................ 100

4.3 Histidine Hyperfine Coupling Analysis ......................................................... 103

4.4 Exchangeable Hydrogen Analysis .................................................................. 107

4.5 2H ESEEM on PheH ...................................................................................... 109

4.6 Conclusions .................................................................................................... 114

4.7 References ...................................................................................................... 115
APPENDIX I

MATLAB Simulation Scripts ............................................................................... 116
APPENDIX 11

Electron Spin Echo Envelope Modulation Spectroscopy ..................................... 124

vi

LIST OF TABLES

Table 1.1: Simulation Parameters for TyrH ....................................................................... 23
Table 1.2: Simulation Parameters for PheH ....................................................................... 23
Table 2.1: {F eNO}7 Forrnalisms ....................................................................................... 39
Table 3.1: Deuterium Hyperfine Couplings ....................................................................... 89

vii

Figure 1.1:
Figure 1.2:

Figure 1.3:

LIST OF FIGURES
Images in this dissertation are presented in color

General Oxygenase Mechanism ........................................................................ 2
Tetrahydrobiopterin Dependent Aromatic Amino Acid Hydroxylases .............. 4

Mechanism for Tetrahydrobiopterin Dependent Aromatic Amino Acid

Hydroxylases ........................................................................................................................ 8
Figure 1.4: CW—EPR Spectra of TyrH ................................................................................ 19
Figure 1.5: CW—EPR Spectra of PheH ............................................................................... 21
Figure 1.6: CW-EPR Spectra of APheH ............................................................................ 22
Figure 2.1: Structures of FeNzOl, FeN202 and FeN203 .................................................. 42

Figure 2.2: CW-EPR of ferric N201, N202 , and N203 .................................................... 50
Figure 2.3: CW—EPR of {FeNO}7 N20,, N202, and N203 .............................................. 51

Figure 2.4: HYSCORE spectrum of {F eNO}7-N20l recorded at 1980 G ........................ 53
Figure 2.5: HYSCORE spectrum of {FeNO}7-N201 recorded at 3100 G ........................ 53
Figure 2.6: HYSCORE spectrum of {FeNO}7-N202 recorded at 1980 G ........................ 54
Figure 2.7: HYSCORE spectrum of {F eNO}7-N202 recorded at 3100 G ........................ 54
Figure 2.8: HYSCORE spectrum of {FeNO}7-N203 recorded at 1980 G ........................ 55
Figure 2.9: HYSCORE spectrum of {FeNO}7-N203 recorded at 3100 G ........................ 56
Figure 2.10: HYSCORE spectrum of {FeNO}7-N20l in D20 recorded at 3100 G ......... 57

Figure 2.11: HYSCORE spectrum of {FeNO}7-N20I recorded at 2800 G ...................... 58
Figure 2.12: HYSCORE spectrum of {FeNO}7-N20l recorded at 2900 G ...................... 58
Figure 2.13: HYSCORE spectrum of {FeNO}7-N20l recorded at 3000 G ...................... 59
Figure 2.14: HYSCORE spectrum of {FeNO}7-N201 recorded at 3100 G ...................... 59

viii

Figure 2.15: HYSCORE spectrum of {F eNO}7-N201 recorded at 3200 G ...................... 59

Figure 2.16: ‘H HYSCORE simulation at 2800 G ............................................................ 60
Figure 2.17: 'H HYSCORE simulation at 2900 G ........................................................... 60
Figure 2.18: 1H HYSCORE simulation at 3000 G ............................................................ 60
Figure 2.19: ‘H HYSCORE simulation at 3100 G ............................................................ 61
Figure 2.20: lH HYSCORE simulation at 3200 G ............................................................ 61
Figure 2.21: 14N HYSCORE simulation at 3100 G .......................................................... 62
Figure 3.1: TyrH reaction ................................................................................................... 73
Figure 3.2: TyrH Catalytic Mechanism .............................................................................. 74
Figure 3.3: Representative 2H ESEEM of {FeNO}7 TyrH treated with d-6MPH4 and
treated with d-6MPH4 and tyr ............................................................................................ 81
Figure 3.4: Representative 2H ESEEM of {FeNO}7 TyrH treated with d-tyr and treated
with 6MPH4 and d-tyr ........................................................................................................ 83
Figure 3.5: Representative 2ii ESEEM of the {FeN0}7 TyrH mutant 13332.4 ................. 84
Figure 3.6: 2H simulations with an angle between the dipole vector and g2 of 0.3 rad ..... 86
Figure 3.7: 2H simulations with an angle between the dipole vector and gZ of 1.3 rad ..... 86
Figure 3.8: Summary of ESEEM Results .......................................................................... 94
Figure 4.1: HYSCORE spectrum of TyrH[6MPH4,tyr] at 1980 G .................................. 101
Figure 4.2: HYSCORE spectrum of TyrH[6MPH ,tyr] at 3100G ................................... 101
Figure 4.3: HYSCORE spectrum of PheH[Phe, 5d] at 1980 G ....................................... 102
Figure 4.4: HYSCORE spectrum of PheH[Phe, 5d] at 31000 ........................................ 103
Figure 4.5: 1H HYSCORE simulations for TyrH at 1980 G and 3100 G ........................ 104
Figure 4.6: 1H HYSCORE simulations for PheH at 1980 G and 3100 G ....................... 105
Figure 4.7: 14N HYSCORE simulations at 3100 G ......................................................... 106

ix

Figure 4.6: 1H HYSCORE simulations at 3100 G for TyrH and PheH ........................... 107

Figure 4.9: 2H ESEEM collected on PheH[d-Sd] and PheH[d-Sd, phe] .......................... 110
Figure 4.10: 2H ESEEM collected on APheH[d-5d] and APheH[d-5d, phe] .................. 111
Figure 4.11: 2H ESEEM collected on PheH[d-phe] ......................................................... 113
Figure A2.1: Energy level diagram and corresponding EPR spectrum ............................ 125
Figure A2.2: EPR pulse sequences .................................................................................. 126

LIST OF ABBREVIATIONS

5d — 5-deaza-6MPH4

6MPH4 — 6-Methyltetrahydropterin

CT — Charge Transferred

CW-EPR — Continuous Wave EPR

d-5d - 2H-5-dcaza-6MFH4

d—6MPH4 — 6-CH3-6,7-2H-tetrahydropterin

d-phe — L-para—zH-Phenylalanine

d-tyr — L-3,5-2H-Tyrosine

DOPA - dihydroxy-phenylalanine

EPR — Electron Paramagnetic Resonance

ESEEM — Electron spin echo envelope modulation
EXAFS — Extended X-ray absorption fine structure
FWHM — Full-width half-max

HYSCORE — Hyperfine sub-level correlation
MCD — Magnetic circular Dichroism

PheH — Phenylalanine Hydroxylase

phe — Phenylalanine

PKU — Phenylketonuria

TrpH — Tryptophan Hydroxylase

tyr — Tyrosine

TyrH — Tyrosine Hydroxylase

xi

XAS — X-ray absorption spectroscopy

xii

CHAPTER I
CONTINUOUS WAVE ELECTRON PARAMAGNETIC RESONANCE

INVESTIGATION OF TWO AROMATIC AMINO ACID HYDROXYLASES

1.1 OXYGEN ACTIVATION

Molecular oxygen serves as the terminal electron acceptor in a variety of
biological processes and serves as the source for the oxygen found in many essential
molecules in nature.1 Thermodynamically, the cleaving of the oxygen-oxygen bond of
molecular oxygen and its subsequent reaction with organic molecules is energetically
favorable.2 The thermodynamics are held in check by the slow reactivity of oxygen at
ambient temperatures. If this weren't the case, organic molecules would spontaneously
oxidize making our atmosphere deadly to life. Thankfully, there is a spin mismatch
between the ground states, dioxygen is a triplet while most organic molecules tend to
have a singlet ground state, this results in slow kinetic reactivity between the two.3 To
overcome this kinetic barrier, nature has developed a variety of catalysts in the form of
metalloenzymes.4 Transition metals such as iron or copper are able to coordinate the
triplet oxygen and through electron transfer processes are able to activate the oxygen and
form intermediates that can participate in reactions with organic substrates.5 This work is
focused on these metalloenzymes and specifically those containing a non-heme iron, so
only that aspect of oxygen activation will be further discussed.

The mono-nuclear metalloenzymes which incorporate oxygen atoms into organic

molecules can broadly be classified as oxygenases and can be sub divided into two

different classes, mono-oxygenases and di-oxygenases. These two classes of oxygenases
are distinguished by the number of oxygen atoms that they incorporate into the organic
substrate, mono-oxygenases add a single oxygen atom to the product, on the other-hand
both oxygen's are added to the substrates in di-oxygenases. The di-oxygenases can be
divided further into two groups intra-molecular di-oxygenases in which both oxygen
atoms are incorporated into a single product and inter-molecular di-oxygenases which

utilize a second substrate which is also hydroxylated. 2

o Fe(l|)
2\V ¢ 0

o”
I

\' Fe(ll)
Di-oxygenases i 6 Mono-oxygenases

OI

Fe(ll) fl Falwfi Fe(lll)flge(lV) Sflrem)

S-OH s S'-OH s' 2H+ H20 3014

Figure 1.1: General Oxygenases Mechanism

A general mechanism for the reaction of mono-nuclear oxygenases is shown in
figure 1.1.3 The mono-nuclear oxygenases function, first, through the direct binding of
dioxygen to the metal in the active site. Many of the proposed mechanisms suggest that
the oxygen is then reduced by two electrons, one provided by the metal and the other
from a co-factor or co-substrate.l This intermediate, a ferric peroxy adduct, contains an
activated oxygen that can now readily participate in chemical reactions. In the case of the
monooxygenases, this adduct is doubly protonated, releasing a water molecule and
forming what has been proposed as a high-valent ferryl-oxo intermediate. This iron-0x0

intermediate is highly reactive and will readily hydroxylate unactivated organic

substrates.3 The di-oxygenases are suggested to proceed through a very similar
mechanism, instead of the release of water, the ferric peroxy adduct reacts directly with
the substrate. 1’ 3 In both cases, the systems are generally designed to tightly control this

oxygen activation and minimize side reactions.6

1.2 N,N,O FACIAL TRIAD

Based on structural studies of non-heme iron(II) oxygenases, a common structural
motif has emerged as the foundation for catalyzing a highly versatile group of chemical
reactions. 7 This motif coordinates the iron with 2 histidines and a carboxylate in a facial
fashion, leaving three labile coordination sites onto which substrates or co-substrates can
bind. This coordination geometry represents a contrast to the heme oxygenases which
generally only have one open coordination site where oxygen binds.5 As a result the non-
heme Fe di-oxygenases allow for a very different set of hydroxylation reactions than the
heme oxygenases.8 However, these non-heme di-oxygenases have not been studied
nearly as extensively as their heme counterparts.9 The non-heme iron active sites have
ligands that do not exhibit the intense pi -> pi* absorption features seen from the
porphyrin ligand in the heme enzymes.10 In addition, the reduced F e(II) sites do not
exhibit charge transfer (CT) transitions and are generally not detectable by Electron
Paramagnetic Resonance(EPR), and while the oxidized Fe(III) sites do exhibit EPR
signals reflecting spin Hamiltonian parameters and CT features, their information content
has not been developed. “

One particularly interesting family of mono-nuclear non-heme iron hydroxylases

uses the co-factor tetrahydrobiopterin as a reductant to catalyze the hydroxylation of

aromatic amino acids. '2 '3 This family of tetrahydrobiopterin dependent aromatic amino
acid hydroxylases includes tyrosine hydroxylase(TyrH), phenylalanine
hydroxylase(PheH) and tryptophan hydroxylase(TrpH), each of which catalyze a very

important hydroxylation reaction that can be seen in figure 1 .2.14 '5

m: PheH Mmz
. +
H0 NH3
NH3

TrpH
QD/YNFr—‘a NH;
I
HN

2Co-substrate
O H O
H
TN 4 O
| —> \E \ 1
TN 8a N/ N

Figure 1.2: Tetrahydrobiopterin Dependent Aromatic
Amino Acid Hydroxylases

TyrH catalyzes the hydroxylation of tyrosine into dihydroxy-phenylalanine(DOPA), this
reaction is the rate limiting step to the biosynthesis of the catecholamine
neurotransmitters which includes dopamine, norepinephrine and epinephrine. ‘3 This
enzyme has been implicated with a number of neurological disorders such as Parkinson's
or Segawa's syndrome, both of which have been successfully treated with DOPA

replacement therapy. '2 Dysfunction of PheH causes Phenylketonuria(PKU), a disorder

which manifests itself in severe mental retardation caused by the neurotoxicity of phe and
its metabolites.12 TrpH produces 5-hydroxytryptophan, the precursor and rate limiting
step for the biosynthesis of serotonin and melatonin. '3 Understanding the structural
characteristics of these three enzymes will be important for addressing a range of
different neurological and psychological disorders.
1.3 STRUCTURE OF TETRAHYDROBIOPTERIN DEPENDENT
HYDROXYLASES

Crystal structures have been described for the three aromatic amino acid
hydroxylases. These crystal structures have largely been determined for different
truncated variants of the inactive oxidized enzyme containing Fe(III). PheH is one
exception, three structures exist of the truncated catalytic domain which contains Fe(II).
One structure is of the binary complex with co-substrate,
tetrahydrobiopterin(PDB:1J8U),l6 the other two are of the ternary complexes with co-
substrate, BH4, and primary substrate analogs thienylalanine(PDB:1KW0,1MMK)17 or
norleucine(PDB:1MMT)‘8. In addition to these ferrous structures, a number of lower
resolution ferric structures exist for PheH. The other two enzymes, TyrH and TrpH, have
not been crystallographically characterized to the same extent as PheH. The catalytic
domain of ferric TyrH, with(PDB: 1TOH)19 and without the co-substrate(PDB:2TOH)20
have been reported. TrpH has been crystallized in the ferric form with co-
substrate(PDB:1MLW)2' and with primary substrate trp(PDB:3E2T)22.

All three hydroxylases have significant sequence similarity and based on the

above structural studies have shown structural homology of their catalytic cores. ‘3 Based

on these two pieces of evidence and the similarity of the chemical reaction of these
enzymes much of what has been learned from the more characterized PheH has been used
interchangeably with the other two enzymes in the family. 6

A comparison of the catalytic domain of ferric PheH to that of ferrous PheH
shows no difference in the relative structure at the active site. '6‘ 23 This comparison
however is only of the binary enzyme with co-substrate, so while no large scale changes
are observed in the catalytic domain, this does not necessarily translate to the ternary or
quaternary complex.

Examination of the PheH crystal structures show a range of structural changes in
the catalytic core as a result of substrate and co-substrate binding.18 '7 For example
comparison of the binary PheH structure with that of the ternary enzyme reveals that a
flexible loop consisting of residues 131-145 and specifically Y138 packs against the
amino acid binding site. 17 A site directed mutagenesis study on the analogous loop in
TyrH indicated that the residue analogous to Y138 was responsible for optimizing the
coupling between co-substrate oxidation and the hydroxylation of primary substrate. 24
Further florescence anisotropy studies suggest that presence of the pterin modulates the
motion of this loop. 25

A second important observation from the PheH crystal strrictures is that the
presence of the primary substrate affects a change in the active site which brings the
position of the pterin closer to the iron. The binary structure of PheH has a distance of 6
A between the Fe and the position of the pterin that gets hydroxylated, the C 4a position.

Comparison of the binary enzyme to the ternary enzyme shows that binding of the amino

acid decreases this distance to 4.5 A. Thus, the binding of the amino acid substrate analog
initiates a conformational change in PheH that pushes the pterin co-substrate closer to the
iron. '7

Interestingly, in the ternary crystal structure, the pterin co-substrate and the
primary substrate never directly coordinate the iron. These substrates do however,
displace two of the iron coordinating solvent molecules, leaving the iron five coordinate:
2 his, 1 glu(bound bidentate) and a single water. The pterin is positioned in the active site
through a network of hydrogen bonds to the enzyme backbone consisting of residues
248-251. On addition of substrate, the pterin is brought closer to the iron by a movement
of the backbone and its position near the iron is defined by hydrogen bonding to glu286.
This movement of the pterin displaces two solvent molecules. The Fe ligands, Glu330
and Hi5285, together with the pterin approximately form a plane perpendicular to the
third Fe ligand, Hi5290. There is no structural evidence in regard to primary substrate
binding in the absence of co-substrate; however, in the ternary complex, the substrate
analog was seen to bind in a relatively hydrophobic pocket, stacking on Hi5285 and
positioning it near the remaining open coordination site. Both substrate and co-substrate
must be present for enzyme turnover.
1.4 MECHANISM

Based on the general mechanism given above and what is known about the
structure and reaction products, a specific chemical mechanism for the action of the

tetrahydrobiopterin aromatic amino acids has been proposed. Seen in figure 1.3 is the

proposed mechanism for the hydroxylation of phe in PheH.l3 The cycle begins when both

substrate and co-substrate bind in the enzyme, this binding converts the iron from 6 to 5
coordinate providing an open coordination site to accommodate oxygen. This portion of
the mechanism is supported by both the crystal structure evidence and other

spectroscopic techniques including MCD. 26

H H HN H H
H N YNHz H /N\|NrNH H ,NYNH H N /NINH
R N R R
H H HH HN 0O
02O
Fe(II)C “WC —’ 3:3,“;(3 —’ Fem)

M33 @333 ©J3,3.,0©_/k3,3.

Figure 1.3: Mechanism for Tetrahydrobiopterin Dependent Aromatic Amino Acid
Hydroxylation

Following oxygen binding to the iron, an iron-peroxy-pterin species forms,
oxidizing the pterin and reducing the di-oxygen.'5“’ 27 The iron-peroxy-pterin species
undergoes heterolytic bond cleavage producing 4a-hydroxypterin and an F e(IV)-oxo

species.28 An electrophilic substitution reaction occurs between the ferryl-oxo
intermediate and the substrate that then undergoes a 1,2-hydrogen shift and
tautomerization to yield the hydroxylated amino acid product.29 6‘ 30 After this
electrophilic substitution the ferrous enzyme is regenerated.

1.5 Fe-NO AS A SPECTROSCOPIC PROBE

As suggested by the mechanism, the ferrous oxidation state is the catalytically
relevant form of iron in these enzymes and as mentioned above, spectroscopically

studying the enzymes in this oxidation state is challenging. One method, first utilized in

the study of heme's and later in non-heme systems, involves binding the iron with nitric
oxide.31 The binding of NO to the iron serves two purposes: first, NO acts as an oxygen
surrogate binding in a similar fashion competitively blocking oxygen binding 32 ; second,
the NO antiferromagnetically couples with the iron converting the EPR silent S = 2 metal
into an EPR active S = 3/2 complex.31 This method has been used in a recent study by the
Abu-Omar group to examine a strain of bacterial PheH. 33 They presented evidence that
in the ternary complex NO mimics 02 binding and displaces the sole water molecule
observed in the crystal structure.33
1.6 PURPOSE
In this thesis, the F e(II)-NO complex will be used to study the tetrahydrobiopterin

dependent aromatic amino acid hydroxylases using a variety of EPR methods. Through
collaborations, our group has obtained two of the three hydroxylases, TyrH and PheH. As
an initial experiment to determine if we could produce the Fe(II)-NO adduct for both of
these enzymes, continuous wave EPR experiments were performed. These experiments
will tell us if we were successful in producing an EPR active species and if there are
quantifiable changes in the spectrum due to substrate interactions.
1.7 EXPERIMENTAL
1.7.1 TyrH PREPARATION

6-Methyltetrahydropterin (6MPH4) and 6-methylpterin were purchased from Schircks
Laboratories (Jona, Switzerland). 6-(2-Hydroxy-1-methyl-2-nitrosohydrazino)-N-methyl-
hexanamine (MAHMA NONOate), ethylenediamine tetraacetic acid (EDTA), 3-(N-

morpholino)propanesulfonic acid (Mops), L-tyrosine and glycerol were purchased from

Sigma-Aldrich (St. Louis, MO). L-3,5-2H-Tyrosine and deuterium gas were from
Cambridge Isotopes (Andover, MA). Potassium chloride and ferrous ammonium sulfate
were from Fisher (Pittsburg, PA). 6-CH3-6,7-2H-tetrahydropterin (d-6MPH4) was
synthesized by reduction of the commercially available 6-methylpterin to the level of
tetrahydropterin using deuterium gas, as previously described.34 All other chemicals were
of the highest purity commercially available.

\Vlld-type TyrH was expressed in E. coli and purified as previously described.35 36 In
order to remove the ferric iron from the protein, the ammonium sulfate pellet at the end
of the purification was resuspended in 5 mM EDTA, 200 mM Hepes, (pH 7.5), 10%
glycerol and 0.1 M KCl, and incubated on ice for one hour. The enzyme solution was
then dialyzed against the same buffer without EDTA and concentrated using Amicon
Ultra—15 and Ultra—4 centrifugal filters (Millipore Corp., MA). The enzyme samples for
ESEEM were brought to a final glycerol concentration of 30% (v/v) during the
concentration stage. The iron content of the apo-enzyme was measured using a Perkin-
Elmer Aanalyst 600 atomic absorption instrument.37 The typical iron content of an apo
enzyme preparation was 50.1 equivalent.

NO Samples for EPR were prepared using MAHMA NONOate as the nitric oxide
donor. Stock solutions of MAHMA NONOate were prepared in 0.01 M NaOH just
before the experiment and were always kept on ice. Exact concentrations of the MAHMA
NONOate solutions were determined from the UV absorbance at 250 nm in 0.01 M
KOH, using an extinction coefficient of 7.3 mM'lcm‘]. 38 Highly concentrated stock

solutions of tyrosine and 3,5-2H-tyrosine (~50 mM) were prepared by bringing the

10

solution to a final pH of ~1 0. The exact concentrations of the tyrosine stock solutions
were determined using an extinction coefficient of 1.34 mM'lcm'l at 275 nm in 0.1 M
HCl. Stock solutions of the protiated and deuterated 6MPH4 were prepared in 2 mM
HCl, and an extinction coefficient of 17.8 mM'lcm‘1 in 2 M perchloric acid was used to
determine the concentrations. Ferrous ammonium sulfate solutions were prepared fresh
by dissolving the appropriate amount of powder in 2 mM HCl. ESEEM samples were
prepared inside an anaerobic cuvette at 25 °C. Apo-TyrH (0.9-1.2 mM) and tyrosine (if
the complex contained tyrosine) were placed at the bottom of the cuvette. Ferrous
ammonium sulfate solution (0.9 equivalent in ~10 pl) was placed on the lower neck of
the cuvette. 6MPH4 and MAHMA NONOate solutions were either placed in the side
arms or on the upper neck of the cuvette, for large and small volumes, respectively.
Bufi‘er conditions were 100 mM Mops (pH 7.0), 0.3 M KCl and 30% glycerol. Each
MAHMA NONOate molecule releases two molecules of NO,38 and the half-life of
MAHMA NONOate is ~35 5 under these buffer and temperature conditions (data not
shown). The contents of the cuvette (a total volume of ~250 pl) were made anaerobic by
the application of argon-vacuum exchange for at least 20 minutes. The anaerobic enzyme
solution was then mixed with ferrous ammonium sulfate and incubated for 10 minutes.
This was followed by mixing with 6MPH4 (if the complex contained 6MPH4). After a
few minutes of incubation, MAHMA NONOate (~0.6 equivalent of the enzyme) at the
upper neck of the cuvette was introduced to the enzyme-substrate mixture. Afier ~3
minutes, ~200 u] of the reaction mixture was quickly transferred to the quartz EPR tubes

(4 mm OD, 707-SQ-250M, Wilmad, Buena, NJ) using a glass pipette and immediately

11

frozen in liquid nitrogen. (UV-Visible experiments performed under similar conditions
indicated that the maximum absorbance at 450 nm, which is indicative of NO binding,
was reached at about 3 minutes and stayed constant for at least a few minutes). UV-
Visible spectra collected at increasing concentrations of MAHMA NONOate showed that
NO was saturating under the concentrations used.
1.7.2 PheH Preparation

All commercial reagents were of the highest grade available and were used without
fiirther purification, with the exception of glycerol, which was treated with activated
carbon to remove contaminants”. Glycerol, L-Phe, 4-morpholinepropanesulfonic acid
(MOPS), KCl, sodium dithionite, ascorbate, sodium nitrite, ferrous ammonium sulfate, p-
I-L-Phe, were from Sigma (St. Louis, MO). Phenylsepharose and Superdex 200 were
purchased from Pharmacia (Uppsala, Sweden). D20 (99.9 atom % D) and per-ZH-L-Phe
were from Cambridge Isotopes Laboratories (Andover, MA). Diethylamine NONOate
(DEA/NO) was purchased from Cayman Chemicals (Ann Arbor, MI). 6-MPH4 4° and 5-
deaza-6-MPH4 4' was synthesized and characterized by standard methods42 and was
stored at —20°C until use. 5-deaza-6-CH3-5,6,7,8-2H-pterin (ZH-S-deaza-6MPH4) was
synthesized as previously described“, except that deuterium gas is used in the reduction
step. The para-ZH-L-phe was also synthesized as previously described43 using p-I-L-phe
as the starting material. E.coli cells were grown in a New Brunswick Bioflo 2000
fermentor (Edison, NJ). Total iron content of samples were routinely quantified using a
Varian AA280 atomic absorption spectrometry, with Zeeman GTAIZOZ graphite furnace

attachment, at 248.3 nm. The iron standard was purchased from Fisher (Pittsburgh, PA).

12

Automated protein purification was performed using a Pharmacia LKB F PLC (Uppsala,
Sweden). All electronic absorption spectroscopy (UV/V is) analyses were performed on a
HP-8453 diode array spectrophotometer (Palo Alto, CA). All anaerobic work was
performed in a Labconco inert box that was maintained at 4°C (Kansas City, Missouri).
thheH Fe? + -N0

Recombinant rat thheH was over-expressed in E. coli BL21(DE3) cells and was
purified using a variation of the hydrophobic affinity method”, as previously described.44
The specific activities of thheH ranged between 6-7 units/mg (units = micromoles of
tyrosine formed/min) with active iron content between 0.7-0.9 Fe/subunit. All protein
manipulations required for the preparation of the samples were performed in an inert
atmosphere box at 4°C. Approximately 35 mg of protein was used for one sample
preparation. The following is the general procedure for thheH sample preparation.

All enzyme manipulations and NO formation were prepared in buffer (50 mM
MOPS, 300 mM KCl, pH 7.2 at 4 °C) unless otherwise noted. In order to precisely
quantitate the concentration of NO produced in our buffer solutions, a Clark-type NO

1.45 A commercially

electrode was developed using a method adapted from Stetter et.a
available Clark-type electrode, ISO-NOP (World Precision Instruments, Sarasota FL),
was connected to a potentiostat (Bioanalytical Systems Inc., IN) located in an inert
atmosphere box. The poise voltage was set to 865 mV for NO detection.46 The resulting
current is proportional to the concentration of NO in solution. The electrode was

calibrated daily with fresh solutions of sodium nitrite and potassium iodide (resulting in

the formation of 0025-25 mM NO) according to the method suggested by the

13

manufacturers. The calibration factor uA/uM was determined with a linear fit program.

NO saturated buffer solutions were prepared by dissolving DEA/NONOate in 10
mL of buffer (50 mM MOPS, 300 mM KCl, pH 7.2 at 4 °C) to a final 2-2.5 mM
concentration. The solution was placed in a 10 mL Wheaton vial with a butyl-rubber
stopper, crimped with an aluminum seal and left to incubate for one hour with frequent
stirring. This sample volume was used to minimize the amount of void volume present
between the top of the solution and the bottom of the stopper of the vial.Afier one hour
incubation, the NO concentration reached approximately 2.5-2.7 mM, determined with
Clark-type NO-electrode. The maximum saturation of NO at 4°C is reported as 3.2 mM.46
Once the buffer was saturated with NO, it was then added to thheH sample
(approximately 0.1 mM), which was reduced with 0.5 equivalents of 6-MPH4 prior to the
addition of NO solution. After 5 minutes of incubation in a vial sealed under N2
atmosphere, the resulting intense yellow colored solution (Fe2+PheH-NO) was
transferred to Centricon (30K) microconcentrators (Millipore) and concentrated to a final
concentration of 1.5- 2 mM F e/subunit. Approximately 200 uL of the reaction mixture
was quickly transferred to a quartz EPR tube (4 mm OD, 707-SQ-250M, Mlmad, Buena,
NJ) using a glass pipette and immediately frozen in liquid N2.

If the enzyme needed to be manipulated with protiated or deuterated substrate
and/or cofactor, final concentration of 10 —- 15 mM of the corresponding chemical (i.e L-
Phe, para-zH-L-Phe, per-ZH-L-Phe, 5-deaza-6MPH43 2H-5-deaza-6MPH4) was added to
the buffer solutions (50 mM MOPS, 300 mM KCl, pH 7.2 at 4 °C) prior to sample

preparation. Additionally, in order to achieve the activated “R” state, samples were

14

incubated at 25 0C for 10 minutes when necessary.
A] -11 7PheH

Recombinant rat A1-117PheH was over-expressed in E. coli BL21(DE3) cells, similar
to thheH, except that no iron was present during cell growth resulting in expression of
apo-enzyme. Apo-1-117PheH was purified using methods reported earlier47 with the
following adaptations. The enzyme was precipitated with ammonimn sulfate (fractions
between 25% - 45% saturation) and purified by anion exchange chromatography
followed by size exclusion chromatography. Size exclusion chromatography performed at
0.25 mM subunit, confirm A1-117PheH elutes at a position consistent with the expected
tetrameric state of the enzyme that contains the C-terminal domain known to facilitate the
oligomeric state (apparent molecular mass, 154 kDa). The enzyme was concentrated back
to 1 mM (40-50 mg/ml), snap-frozen in liquid nitrogen, and transferred to an inert
atmosphere box for reconstitution with Fe2+. All anaerobic solutions were thoroughly
degassed by three freeze-pump-thaw cycles on a vacuum line utilizing N2 as the inert
atmosphere. Protein samples were diluted to 0.05 mM using deoxygenated buffer (50
mM MOPS, 0.3 M KCl, pH 7.2). Ferrous ammonium sulfate (10 mM stock dissolved in 5
mM HCl) was added to a final concentration of 0.9:] F e2+zenzyme and incubated for 15
minutes for reconstitution. Reconstituted Fe2+A1-117PheH was transferred to Centricon
(30K) microconcentrators (Millipore) in N2 atmosphere and concentrated to a final
concentration of 1.5 - 2 mM Fe/subunit. All protein manipulations required for the
sample preparation were identical to the thheH sample preparations, that are discussed

above, except that the addition of 6-MPH4 was not necessary in order to reduce the

15

enzyme, since it was reconstituted in the ferrous oxidation state. The specific activities of
the truncated enzyme used for kinetic analysis ranged between 10-14 units/mg with 0.7-
0.9 F e/subunit.
1.7.3 CW—EPR

CW—EPR spectra were recorded on a Bruker ESP300E X-band EPR spectrometer
equipped with a TE102 cavity and an Oxford Instruments ESR-900 liquid helium cryostat
with a model ITC-502 temperature controller. CW—EPR measurements were made with
the following conditions: sample temperature, 4.0 K; microwave frequency, 9.46 GHz;
microwave power, 1 mW; magnetic field modulation frequency, 100 kHz; field
modulation amplitude, 5 G; conversion time, 163 ms. CW—EPR data were analyzed using
the X—SOPHE program available from Bruker Biospin Corporation.
1.8 THEORY

The complex formed between Fe(II) and NO produces a characteristic EPR

spectrum with principle g values of g = 2 and g = 4. This spectrum can be interpreted as a
S = 3/2 spin state and can be represented by the Hamiltonian:

HzS-D-S+g08 H-S [1.1]

e

The first term in the Hamiltonian, D, is the term that represents interaction between the
electron dipoles of the unpaired electrons, referred to as the zero field interaction. The
interaction matrix D, is a symmetric and traceless tensor, and in its principle axis can be

described by two parameters, D the magnitude of the interaction and E/D the shifi away

16

from axial symmetry which can range from 0 to 1/3. At zero magnetic field the electron
spin states are degenerate pairs or Kramers doublets, for an S = 3/2 system there are two
doublets, i 1/2 and t 3/2, when D is non zero the degeneracy between these doublets is
lost and the spin states are separated by |D|. Fe-NO complexes have a [D] that is about 10

cm‘1

, our operating frequency at x-band is about 0.3 cm'l, this is much too small to
excite inter-kramers(between i 1/2 and at 3/2) doublet transitions and only intra-
kramers(in the :t 1/2 and i 3/2 manifolds) transitions should be observed.

The second term in the Hamiltonian is the electron Zeeman interaction, this term
only appears when the magnetic field is turned on and represents a field dependent
splitting within the Kramers doublets. An approximation for the electron Zeeman
interaction in the Fe-NO systems is to treat the g factor as a scalar and allow the principle
axis of the system to be defined by the zero field interaction and specifically the principle
axis of the zero field interaction is defined by the F e-NO bond.

One way to discuss highly anisotropic CW-EPR spectra is in terms of their
effective g values. In the case of the F e-NO S = 3/2 spin system, the observed spectrum is
highly anisotropic and is both a function of the electron Zeeman interaction and the zero
field interaction. Based on the approximations made above for the electronic Zeeman, it
is not exactly applicable to describe the spectrum in terms of its principle g values.
Instead a set of effective g values, which depend upon both the zero field interaction and
electron Zeeman interaction, need to be described. Afier diagonalization of equation 1,

the eigenvalues can be used to obtain the effective g values within the Krarners doublets.

In the limit where |D| >> gBeHO, these effective g value are described by the following

17

three equations“:

 

 

 

 

 

g =go-1— (1+3(E/D)) 32

x'efl \[1+3(E/D)23 [' a]

33 ___g _ (1— 3'<E/D>> 3333

“ff 0 J11+3 (E/D)2 ' 1
2 1

.e = 1- .

g" f g0 \/1+3(E/D)2 [12c]

Using these equations, the three principle features of the EPR spectrum can be discussed
in terms of the rhombicity of the zero field interaction. The magnitude of the zero field
interaction |D|, in the limit where lDl >> gl333H0 has no influence on the EPR features and
the rhombicity, E/D, only has an appreciable effect on gx and gy serving to split the two
features.
1.9 RESULTS

To study the effects of substrate and co-substrate binding on the electronic
structure of Fe center in TyrH and PheH various derivatives treated with NO were
examined. Four derivatives of TyrH were examined: TyrH treated with 6-
methyltetrahydropterin, TyrH[6MPH4], TyrH treated with L-tyrosine, TyrH[tyr], TyrH
treated with L-tyrosine and 6—methyltetrahydropterin, TyrH[tyr,6MPH4] and the TyrH
mutant E332A treated with L-tyrosine and 6-methyltetrahydropterin, E332A[tyr,6MPH4].
Four derivatives of PheH were examined: PheH, PheH[ ], PheH treated with 5-
deazapterin, PheH[Sd], PheH treated with L-phenylalanine, PheH[Phe], and PheH treated

with L- phenylalanine and 5-deazapterin, PheH[Phe, 5d]. In addition to the full length

18

PheH, three derivatives of a truncated variant were examined: A1-117PheH treated with

5-deazapterin, APheH[5d], A1-117PheH treated with L-phenylalanine, APheH[Phe], and

 

A1-117PheH treated with L- phenylalanine

j\fi and 5-deazapterin, APheH[Phe, 5d].

Shown in figure 1.4 are the CW—EPR
JP spectra of the 4 derivatives of TyrH; (a)
W TyrH[6MPH4], (b) TyrH[tyr], (c)

1600 1800 2000 2200(Figure 1.4a) is characteristic of an axial S =
Field (G)

 

 

 

dX"/dH (a.u.)

TyrH[tyr,6MPH4], (d) E332A[tyr,6MPH4],

 

 

experiment is shown in black and the

corresponding simulation is shown in red.

 

 

The CW-EPR spectrum of TyrH[6MPH4]

 

 

 

1200 1400

3/2 species yielding spectral features at

Figure 1.4: CW-EPR spectra of (a)
TyrH[6MPH4]. (b) TyrH[tyr]. (c)

TyrH[tyr,6MPH4], (d) E332 A[tyr,6MPH 4] red trace of Figure 1.4a is a simulation of the
shown in black with corresponding

effective g values of g i = 4 and gH = 2. The

, , , spectrum performed using spin Hamiltonian
Simulattons shown in red.

(1.1), with |D| = 10 em" and E/D = 0. The
line shape in the g = 4 region was described with two parameters, an intrinsic Gaussian
line shape with a FWHM of 1 x 10-3 cm'1 that was applied uniformly across the EPR
spectrum, and an E/D strain, oE/D = 0.0130, that was adjusted to account for the line

shape at g = 4.

Figure 1.4b shows the EPR spectrum of TyrH[tyr]. This was modeled as a mixture

19

of two components in our simulation (red trace). The major component, 91% of the
observed signal, was simulated with gO = 2.028 and E/D = 0.0175.The inflection-like
characteristic of the line shape at g = 4 required an E/D strain broadening of (SE/D =
0.0107. The minor component, 9% of the observed signal, is more rhombic with g0 =
2.023, E/D = 0.0425, and (SE/D = 0.0040. Simulations of the major and minor components
were added with the appropriate weightings to obtain the simulation.

The CW-EPR spectrum of TyrH treated with reduced pterin and L-tyrosine,
TyrH[6MPH4,tyr] (Figure 1.4c), is composed of two slightly rhombic species. The more
axial of the two accounts for 76% of the observed signal and was simulated with g0 =
2.024, E/D = 0.0204 and GE/D = 0.0035. The second, slightly more rhombic component,
was simulated with gO = 2.02, E/D = 0.051, and GE/D = 0.0040. As in Figure 1.4b,
simulations of the major and minor components were added with the appropriate
weightings to obtain the simulation shown as the red trace in Figure 1.4c.

The CW-EPR spectrum of the TyrH mutant E332A treated with reduced pterin
and L-tyrosine, E332A[6MPH4,tyr], is composed of two slightly rhombic species, as
shown in Figure 1.4d. The more axial of the two accounts for 76% of the observed signal
and was simulated with gO = 2.024, E/D = 0.0201 and (SE/D = 0.005. The second, slightly
more rhombic component, was simulated with gO = 2.019, E/D = 0.05, and (SE/D =
0.0040. As in Figure 1.4b, simulations of the major and minor components were added
with the appropriate weightings to obtain the simulation shown as the red trace in Fig
1.4d.

Shown in figure 1.5 are the CW-EPR spectra of the 4 derivatives of PheH; (a)

20

 

 

 

PheH[ ], (b) PheH[5d], (c) PheH[Phe], (d)
3 AK PheH[Phe, 5d], experiment is shown in
black and the corresponding simulation is
b shown in red. The CW-EPR spectrum of
PheH[] (Figure 1.5a) shows the

 

 

 

 

 

 

it
3 c characteristic nearly axial spectrum of a S =
3/2 species. The red trace of Figure 1.5a is a
simulation of the spectrum done using the
above spin Hamiltonian, with |D| = 10 cm‘l,
1200 1401600 1800 2000 2200g0— — 2. 0175 and E/D= 0.0135. The line

Field (G)
shape in the g = 4 region was described with

Figure 1.5: CW—EPR spectra of (a) PheH[ ], _ , , , ,
(b) PheH[S d], (C) PheH[Phe], (d) PheH[Phe, tWO parameters, an lIllIlI'lSlC Gaussran line
5d] shown in black with corresponding

simulations shown in red. shape with a FWHM of 1.5 x 10‘3 cm'1 that

was applied uniformly across the EPR

spectrum, and an E/D strain, (SE/D = 0.0030, that was adjusted to account for the line
shape at g = 4.

Figure 1.5b shows the EPR spectrum of PheH[5d]. The spectrum was simulated
with a single component with g0 = 2.023 and E/D = 0.0220. The line shape at g = 4
required an intrinsic Gaussian line shape with a FWHM of 2.0 x 10"3 cm'1 that was
applied uniformly across the EPR spectrum, and an E/D strain, (SE/D = 0.0060.

The CW—EPR spectrum of PheH[Phe] (Figure 1.50), is composed of three slightly

rhombic species. The most rhombic of the three accounts for 54% of the observed signal

21

and was simulated with gO = 2.019, E/D = 0.0330, Gaussian line shape 1.5 x 10’3 cm'I
and oE/D = 0.003. The second, slightly axial more component which accounts for 29%
of the signal, was simulated with g0 = 2.019, E/D = 0.0225, Gaussian line shape 1.5 x 10'
3 cm'1 and (SE/D = 0.0060. The last, axial component, which accounts for 16% of the
signal, was simulated with g0 = 2.019, E/D = 0.000, Gaussian line shape 1.5 x 10'3 cm'1
and (SE/D = 0.0060. As in Figure 1.4b, simulations of the major and minor components
were added with the appropriate weightings to obtain the simulation shown as the red
trace in Figure 1.50.

The CW—EPR spectrum of the PheH[Phe, 5d], is composed of two slightly
rhombic species, as shown in Figure 1.5d. The more axial of the two accounts for 66% of

the observed signal and was simulated with g0 = 2.0225, E/D = 0.0245, Gaussian line

shape 1.5 x 10'3 cm“1 and (IE/D = 0.008. The second, slightly more rhombic component,

 

was simulated with gO = 2.021, E/D =

0.0645, Gaussian line shape 2.0 x 10'3 cm"

 

 

1 , and (SE/D = 0.0075. As in Figure 1.4b,

simulations of the major and minor

 

dXVUH(au)

components were added with the

appropriate weightings to obtain the

 

simulation shown as the red trace in Fig

 

 

 

1200 1400 1600 1800 2000 2200

F19" (G) 1.5d.
Figure 1.6: CW—EPR spectra of (a) Lastly shown in figure 1'6 are the
APheH[5d], (b) APheH[Phe], (c) . .
APheH[Phe, 5d] shown in black with CW-EPR spectra of the 3 derivatives of A]-

corresponding simulations shown in red.

22

117 truncated PheH; (a) APheH[5d], (b) APheH[Phe], (c) APheH[Phe, 5d], experiment is

shown in black and the corresponding simulation is shown in red. The parameters used to

simulate the truncated PheH spectra were identical to those used for wt enzyme.

1.10 SIMULATION NOTES

The parameters used to simulate the CW-EPR spectra of TyrH are shown in table

1.1 and for PheH in table 1.2. These parameters were determined through visual

comparison of the experimental data and simulation.

Table 1.1: Simulation Parameters for TyrH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

% E/D (IE/D g0 Line-width
TyrH[6MPH4] 100 0.0000 0.0130 2 15
TyrH[tyr] 91 0.0175 0.0107 2.028 15

9 0.0425 0.0040 2.023 15
TyrH[6MPH4,tyr] 76 0.0204 0.0035 2.024 15

24 0.0510 0.0040 2.02 15
E332A[6MPH4, tyr] 76 0.0201 0.0050 2.024 15

24 0.0500 0.0040 2.019 15
Table 1.2: Simulation Parameters for PheH

% E/D OE/D g0 Line-width
PheH[ ] 100 0.0135 0.0030 2.0175 15
PheH[S-deazapterin] 100 0.0220 0.0060 2.0230 20
PheH[L-phe] 54 0.0330 0.0030 2.0190 15

29 0.0225 0.0060 2.0190 15

16 0.0000 0.0060 2.0190 15
PheH[L-phe,5- 66 0.0245 0.0080 2.0225 15
demptefinl 34 0.0645 0.0075 2.0210 20

 

 

 

 

 

 

 

23

 

 

The simulations were performed in the Bruker simulation program Xsophe using
the Hamiltonian in equation 1.1. Based on this Hamiltonian there are five parameters to
fit, the zero field parameters, D and E/D and the principle g values, gx, gy, g2. In addition -
to the Hamiltonian parameters, there were two line-width parameters, the intrinsic line-
width and the strain in D and E/D, that need to be considered.

In the limit, where |D| >> gBeH, it can be seen from equation 2a-c that the EPR
spectrum for a S = 3/2 spin system can be characterized in terms of two principle g
features, the feature at g = 2 which corresponds to gz and the feature at g = 4 which
corresponds to gx and gy or g_L. When considering the features of the F e-NO spectrum, gZ
was largely disregarded for several reasons. The samples were prepared with NO and in
some cases there was residual NO left in solution. The NO radical produces an EPR
signal which overlaps with the gZ feature obscuring any information we could gain in that
region. While this could be easily remedied through more careful sample preparation, the
second reason further supports neglecting gz. In equation 20, the position of the g2 feature
only has a minor dependance on EfD. This is further supported by observations from
sample simulations, changes in E/D have a negligible effect on the position of the g2
feature. Based on these observations the gz feature was considered unimportant to
accurately fit and instead the simulation efforts were focused on the g_L or g = 4 region.

In the simulation program, it has also been observed that when |D| < 6 cm'l, the
position of g_1_ is raised to a higher field. This is a result of the higher order terms which
manifest themselves upon diagonalization of equation 1.1 that don't entirely cancel. In

theory the value of |D| could be less than 6 cm'1 however due to our operating frequency,

24

~0.3 cm'l, we can't directly measure the transitions necessary assign a value to |D|, so we
have arbitrarily set the value at 10 cm'l. This value of 10 cm'1 was one utilized in

32” 49 and is large enough so that |D| has no effect on the simulation.

previous studies

It was found that often the g_L feature produced in the simulations did not properly
line up with experiment, when the exact operating frequency and field was used. In order
to adjust the position of g_1_ small adjustments to the value of g0 were necessary. This
overall is an approximation, as is the inclusion of g0 in equations 1.2a-c. A more accurate
interpretation would include three distinct g value, however, due to the resolution of the
experiment at x-band, much like with |D|, its impossible to accurately determine these
values.

The Hamiltonian parameter which has the most substantial influence on the
observed spectrum is the rhombicity of the zero field interaction, E/D. As can be seen in
equation 1.2a-c and using the above approximation for go, the only parameter which
separates gx and gy is E/D. This parameter can have a maximum value of 1/3 however for
those systems we've examined this value rarely exceeds 1/20. This result exemplifies the
axial nature of these F e-NO complexes.

After the Hamiltonian parameters are fit it is necessary to consider the line-shape
to overall produce a satisfying simulation. The line-shape in this system can be
influenced by two parameters intrinsic line-width and the zero field strain. The intrinsic
line-width represents a Gaussian broadening which is applied uniformly to the spectrum,

this broadening is difficult to fully interpret and directly relate back to a physical quantity,

such as T2 or T2,". When deciding the extent of the intrinsic line broadening, this

25

parameter was adjusted in a rough fashion with particular attention paid to the slope
between the maximum near g3, and the minimum near gy.

The line-shape characteristics are also influenced by the zero field strain, which
includes two different effects, strain in |D| and the strain in E/D. Based on the magnitude
of |D| that we've seen in the Fe-NO complexes, its reasonable to assume that only strain
in |D| that is of the magnitude of |D| will have any effect on the spectrum. The second
parameter, the E/D strain, represents an uncertainty in ED or in a physical sense, a
distribution of structures which contribute to a rhombicity in the zero field interaction.
1.11 DISCUSSION

From the presence of the characteristic g=4/g=2 CW-EPR spectrum seen in
figures 1.3-1.5, we've shown that we were able to produce the S = 3/2 Fe-NO complex
with both TyrH and PheH in multiple different states. It can be seen from both the spectra
and the parameters seen in table 1.1, when these enzymes are treated with substrate and
co-substrate there is a change observed in the rhombicity, E/D, and the strain associated
with E/D.

The CW-EPR spectra of TyrH were acquired on wt enzyme in various states, with
and without substrate and co-substrate This study of TyrH did not include the empty,
unloaded enzyme, so there is no “blank” to compare the effect that substrate might have
on the EPR spectrum. However, according to the crystal structures recorded for TyrHZO,
the pterin co-factor is not shown to directly coordinate the iron and based on this we
could make the assumption that the active site structure of TyrH[6MPH4] is close to that

of the empty state. While this assumption is not perfect it does provide us a basis with

26

which to compare the other TyrH spectra. The CW-EPR spectrum of TyrH[6MPH4]
shows a featureless axial line-shape, the result of having an E/D = 0, this line-shape
seems to suggest that the pterin binding has little influence on the electronic structure of
the Fe-NO center and does little to perturb it away from axial. In addition, it has an E/D
strain greater than any other sample examined, this could be a result of structural
dispersion at the active site.

A comparison of the spectrum of TyrH[6MPH4] to the spectrum of TyrH[tyr]
shows considerable differences between the two. TyrH[tyr] is composed of a mixture of
two species, the major component, which composes over 90% of the CW-EPR spectrum,
shows a significant increase of E/D compared to TyrH[6MPH4] this is accompanied by a
20% decrease in the E/D strain. The changes observed indicate that the substrate tyr has a
modest effect on the electronic structure of the TyrH Fe-NO center in comparison to the
co-substrate. This change in electronic structure is most likely a result of a structural
change due to the presence of tyr near the active site either displacing a Fe ligand or
having an influence on the first shell coordination. The decrease in strain suggests that
there is more local order with tyr present. The minor component of the spectra shows an
increase in rhombicity more than double the one seen in the major component and a
much larger decrease in strain.

When TryH is treated with both the substrate and the co-substrate, the CW-EPR
spectrum shows a mixture of two non-axial species at g = 4. The majority species shows
an increase in rhombicity over the majority species seen in the spectrum of TyrH[tyr] and

a decrease of about 60% in E/D strain. These changes can be interpreted in two ways: the

27

addition of the tyr affects a structural change which shifts the position the pterin further
increasing the rhombicity of the electronic structure of the Fe-NO; or the addition of the
pterin shifts the position of the tyr affecting this change. The decrease in strain does
suggest that the quaternary complex has a significantly more ordered structure than the
ternary complexes. The minor species also shows an increase in MD in comparison to the
minor species TyrH[tyr] and an increase in amplitude.

The TyrH mutant E3 32A was also examined, this mutant does not hydroxylate tyr
but does produce an oxidized pterin species, possibly 4a-hydroperoxypterin.35 M The
CW—EPR spectrum of E332A[6MPH4, tyr] shows a mixture of two species and splitting
much like that seen in wild type TyrH[6MPH4, tyr]. Both the major and minor component
have an E/D parameter within about 5% of the wild type enzyme, this result seems to
indicate that the electronic structure of the mutant and the effects the substrates have on
the Fe-NO center are similar to that seen in the wt enzyme.

Overall the CW-EPR spectra acquired on TyrH show differences in the electronic
structure of the Fe-NO active site when different substrates are added and differences
between the ternary and quaternary complexes. In addition to the natural substrates,
selectively deuterated substrates were examined, however no changes were observed. If
the labeled position had a significant interaction with the Fe-NO center a change in line-
width would have been observed, none was seen so the interaction with the labeled
position was too weak to resolve. The quaternary complex in D20 was also examined and
again no influence from isotopic labeling was observed within the resolution of the

experiment.

28

The CW—EPR spectra of PheH shows very similar changes with respect to
substrate binding as was observed in TyrH with a key difference; the major component of
the spectra in the PheH samples all show a relative increase in E/D in comparison to the
analogous sample of TyrH. Differences between the two enzymes are expected, while
their active site structures are similar, they do catalyze different reactions and studies
have suggested that they proceed through different kinetic mechanisms. ‘3

In the PheH system a “blank”, which consists of the empty state enzyme treated
with NO, was examined. The spectrum for this sample shows a splitting at g_L, the value
of E/D ¢ 0, this means that the electronic structure of the F e-NO in PheH[ ] is influenced
by the protein framework and is restricted from obtaining a purely axial environment like
the one seen in TyrH[6MPH4]. This difference is most likely an indication of the
structural differences between the two enzyme and the broader regulatory differences. ‘2

The addition of pterin to PheH[ ] increases the rhombicity of the Fe-NO zero field
interaction by 60% or an increase of E/D of 0.009 and broadens the line, we see an
increase in both the strain and the intrinsic line-width. These changes indicate that the
pterin affects a minor structural change within the enzyme. The increase in line-width
suggests that there may be a larger distribution of structures with pterin present and the
the increase in rhombicity confirms that pterin has a affect on the electronic structure of
the Fe-NO.

The spectrum of PheH[ ] treated with phe shows a mixture of three species all of

which have a rhombicity different than the one seen in PheH[ ] suggesting that phe has an

influence on the electronic structure of Fe-NO. The major species shows an increase in

29

rhombicity and a decrease in strain over both PheH[ ] and PheH[5d], these changes
indicate that in this species the presence of phe has a greater effect on the local structure
at the iron than the addition of pterin. There is a decrease in the E/D strain compared to
PheH[Sd] which means that with phe in the active site there is a slightly less disperse
interaction. The intermediate component of PheH[phe] has a set of parameters similar to
those obtained for the major component PheH[5d], from this similarity it could be
assumed that the phe is binding in a similar fashion and causing the same type of
interaction as the pterin. This is merely an assumption and without any higher resolution
experiments can't actually be determined. In addition to these two species there is a third
more minor component that was largely obscured by the other two. A simulation was
provided for this component however there is relative uncertain in the accuracy of that fit.

Overall the simulations provided for PheH[phe] are not entirely satisfying. While
its true that the simulations do accurately account for the spectral features, the presence
of three distinct species is difficult to accept. In part this idea comes from the overlap of
the different species seen in the spectrum, a feature that was simulated as two distinct
species could have in fact been a single extremely strain broadened(large °E/D) species.
Our ability to distinguish between these two different possibilities is limited by the
resolution of the experiment, so while three species may not be satisfying, the fits
provided are the best available.

When PheH[ ] is treated with both phe and 5d, it displays a mixture of two
slightly rhombic features at g = 4. The major component is 25% less rhombic and has

more than double the strain in E/D than the major, most rhombic component of

30

PheH[phe]. The minor component has a rhombicity greater than any seen in the CW-EPR
spectra collected in this work, with E/D =0.0645. Both of these changes are similar to
those observed in TyrH, and can be interpreted in a similar fashion. The presence of a
single substrate perturbs the electronic structure away from the initial state and the
addition of a second substrate further perturbs the Fe-NO active site. These changes in
electronic structure are most likely tied to a physical structural change near the Fe-NO
however the zero-field parameters are not very well understood and it is difficult to draw
any direct structural correlations.

PheH which has had residue 1-107 truncated was also examined to determine if
there were any significant differences in the electronic structure compared to firll length
enzyme. To within the resolution of the experiment the truncated variety of the enzyme
shows no significant differences compared to full length enzyme. This similarity suggests
that the active site structure of the full length and truncated enzyme are the same, this
however is just an assumption based on the zero field splitting.

Several of the CW-EPR spectra we've collected are composed of mixtures species
with different values of ED. A multi-component CW-EPR spectra has been observed in
both the ferric and ferrous nitrosyl forms of PheH and soybean lipoxygenase.50 4" 33 5 1 52
Based on the observation of a mixture in both ferrous and ferric PheH, it is reasonable to
assume that the mixtures we've observed are independent of the orientation of the NO
bond but instead reflects differences in the coordination environment of the Fe-NO. The

Fe center in these enzymes is known to be coordinated by 2 histidines and a glutamate;

the other sites include solvent and NO. Its possible that the two species observed in the

31

CW-EPR are the result of freezing in different conformers of bound water or of capturing
a mixture, either mono or bidentate, in the ligation of the glutamate side chain. An
alternate explanation was suggested from work performed on a series of models. From a
combined study of both the EPR spectra and the crystal structures, a relationship between
the angle of the Fe-N-O bond and the value of E/D was observed.53 With a decrease in the
F e-N-O bond angle from 178° —> 172° —e 160°, the value of E/D was observed to
increase from 0.062 —> 0.072 —> 0.090.53 This change in E/D is of a similar magnitude
(~0.03) to the changes that we've observed upon substrate additions to TyrH and PheH;
suggesting that these additions might simply be influencing the coordination sphere near
the Fe and in turn having an effect the F e-N-O bond angle. In addition, the mixtures that
we've observed could be a result of freezing in two or more different Fe-N-O bond
angles.
1.12 CONCLUSIONS

We've shown that we have been able to produce an EPR active species from N0
treated ferrous enzyme which shows changes in the electronic structure in response to
substrate additions. The CW-EPR spectra of TyrH and PheH show similarities but are
inherently different. These results while interesting do not provide structural data about
the enzymes and further work using more advanced, higher resolution EPR techniques

will have to be performed.

32

1.13 REFERENCES

l. Kovaleva, E. G.; Lipscomb, J. D., Versatility of Biological Non-Heme Fe(Ii)
Centers in Oxygen Activation Reactions. Nat Chem Biol 2008, 4 (3), 186-193.

2. Abu-Omar, M. M.; Loaiza, A.; Hontzeas, N., Reaction Mechanisms of
Mononuclear Non-Heme Iron Oxygenases. Chem Rev 2005, 105 (6), 2227-2252.

3. Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H., Heme-Containing
Oxygenases. Chem Rev 1996, 96 (7), 2841-2887.

4. Holm, R. H.; Kennepohl, P.; Solomon, E. 1., Structural and Functional Aspects of
Metal Sites in Biology. Chem Rev 1996, 96 (7), 2239-2314.

5. Decker, A.; Solomon, E. I., Dioxygen Activation by Copper, Heme and Non-
Heme Iron Enzymes: Comparison of Electronic Structures and Reactivities. Curr Opin
Chem Biol 2005, 9 (2), 152-163.

6. Fitzpatrick, P. F., Mechanism of Aromatic Amino Acid Hydroxylation.
Biochemistry-Us 2003, 42 (48), 14083-14091.

7. Hegg, E. L.; Que, L., The 2-His-1-Carboxylate Facial Triad - an Emerging
Structural Motif in Mononuclear Non-Heme Iron(Ii) Enzymes. European Journal of
Biochemistry 1997, 250(3), 625-629.

8. (a) Que, L., One Motif - Many Different Reactions. Nat Struct Biol 2000, 7(3),
182-184; (b) Bruijnincx, P. C. A.; van Koten, G.; Gebbink, R. J. M. K., Mononuclear
Non-Heme Iron Enzymes with the 2-His-l-Carboxylate Facial Triad: Recent
Developments in Enzymology and Modeling Studies. Chem Soc Rev 2008, 37 (12), 2716-
2744.

9. Solomon, E. I.; Decker, A.; Lehnert, N., Non-Heme Iron Enzymes: Contrasts to
Heme Catalysis. P Natl Acad Sci USA 2003, 100(7), 3589-3594.

10. Solomon, E. I.; Pavel, E. G.; Loeb, K. E.; Campochiaro, C., Magnetic Circular-
Dichroism Spectroscopy as a Probe of the Geometric and Electronic-Structure of
Nonheme Ferrous Enzymes. Coordin Chem Rev 1995, 144, 369-460.

11. Solomon, E. I.; Brunold, T. C.; Davis, M. 1.; Kemsley, J. N.; Lee, S. K.; Lehnert,
N.; Neese, F.; Skulan, A. J .; Yang, Y. S.; Zhou, J ., Geometric and Electronic
Structure/Function Correlations in Non-Heme Iron Enzymes. Chem Rev 2000, 100 (l ),
235-349.

12. Kappock, T. J .; Caradonna, J. P., Pterin-Dependent Amino Acid Hydroxylases.
Chem Rev 1996, 96 (7), 2659-2756.

13. Fitzpatrick, P. F., Tetrahydropterin-Dependent Amino Acid Hydroxylases. Annu

33

Rev Biochem 1999, 68, 355-381.

14. Lazarus, R. A.; Dietrich, R. F.; Wallick, D. E.; Benkovic, S. J ., On the Mechanism
of Action of Phenylalanine-Hydroxylase. Biochemistry— Us 1981, 20 (24), 6834-6841.

15. (a) Dix, T. A.; Kuhn, D. M.; Benkovic, S. J ., Mechanism of Oxygen Activation by
Tyrosine-Hydroxylase. Biochemistry-Us 1987, 26 (12), 3354-3361; (b) Moran, G. R.;
Daubner, S. C.; Fitzpatrick, P. F., Expression and Characterization of the Catalytic Core
of Tryptophan Hydroxylase. J Biol Chem 1998, 273 (20), 12259-12266.

16. Andersen, O. A.; Flatmark, T.; Hough, B, High Resolution Crystal Structures of
the Catalytic Domain of Human Phenylalanine Hydroxylase in Its Catalytically Active
F e(Ii) Form and Binary Complex with Tetrahydrobiopterin. J Mol Biol 2001, 314 (2),
279-91.

17. Andersen, O. A.; F latmark, T.; Hough, E., Crystal Structure of the Ternary
Complex of the Catalytic Domain of Human Phenylalanine Hydroxylase with
Tetrahydrobiopterin and 3-(2-Thienyl)-L-Alanine, and Its Implications for the
Mechanism of Catalysis and Substrate Activation. J Mol Biol 2002, 320(5), 1095-1108.

18. Andersen, O. A.; Stokka, A. J .; F latrnark, T.; Hough, E., 2.0 Angstrom Resolution
Crystal Structures of the Ternary Complexes of Human Phenylalanine Hydroxylase
Catalytic Domain with Tetrahydrobiopterin and 3-(2-Thienyl)-L-Alanine or L-
Norleucine: Substrate Specificity and Molecular Motions Related to Substrate Binding. J
Mol Biol 2003, 333 (4), 747-757.

19. Goodwill, K. E.; Sabatier, C.; Marks, C.; Raag, R.; Fitzpatrick, P. F.; Stevens, R.
C., Crystal Structure of Tyrosine Hydroxylase at 2.3 Angstrom and Its Implications for
Inherited Neurodegenerative Diseases. Nat Struct Biol 1997, 4 (7), 578-5 85.

20. Goodwill, K. E.; Sabatier, C.; Stevens, R. C., Crystal Structure of Tyrosine
Hydroxylase with Bound Cofactor Analogue and Iron at 2.3 Angstrom Resolution: Self-
Hydroxylation of Phe300 and the Pterin-Binding Site. Biochemistry- Us 1998, 3 7 (39),
13437-13445.

21. Wang, L.; Erlandsen, H.; Haavik, J.; Knappskog, P. M.; Stevens, R. C., Three-
Dimensional Structure of Human Tryptophan Hydroxylase and Its Implications for the
Biosynthesis of the Neurotransmitters Serotonin and Melatonin. Biochemistry- Us 2002,

41 (42), 12569-74.

22. W'mdahl, M. S.; Petersen, C. R.; Christensen, H. E. M.; Harris, P., Crystal
Structure of Tryptophan Hydroxylase with Bound Amino Acid Substrate. Biochemistry-
Us 2008, 47 (46), 12087-12094.

23. Erlandsen, H.; F usetti, F.; Martinez, A.; Hough, E.; F latrnark, T.; Stevens, R. C.,

Crystal Structure of the Catalytic Domain of Human Phenylalanine Hydroxylase Reveals
the Structural Basis for Phenylketonuria. Nat Struct Biol 1997, 4 (12), 995-1000.

34

24. Daubner, S. C.; McGinnis, J. T.; Gardner, M.; Kroboth, S. L.; Morris, A. R.;
Fitzpatrick, P. F., A Flexible Loop in Tyrosine Hydroxylase Controls Coupling of Amino
Acid Hydroxylation to Tetrahydropterin Oxidation. J M01 Biol 2006, 359(2), 299-307.

25. Sura, G. R.; Lasagna, M.; Gawandi, V.; Reinhart, G. D.; Fitzpatrick, P. F., Effects
of Ligands on the Mobility of an Active-Site Loop in Tyrosine Hydroxylase as Monitored
by Fluorescence Anisotropy. Biochemistry-Us 2006, 45 (31), 9632-9638.

26. (a) Loeb, K. E.; Westre, T. E.; Kappock, T. J .; Mitic, N.; Glasfeld, E.; Caradonna,
J. P.; Hedman, B.; Hodgson, K. 0.; Solomon, E. I., Spectroscopic Characterization of the
Catalytically Competent Ferrous Site of the Resting, Activated, and Substrate-Bound
Forms of Phenylalanine Hydroxylase. J Am Chem Soc 1997, 119 (8), 1901-1915; (b)
Kemsley, J. N.; Mitic, N.; Zaleski, K. L.; Caradonna, J. P.; Solomon, E. 1., Circular
Dichroism and Magnetic Circular Dichroism Spectroscopy of the Catalytically
Competent Ferrous Active Site of Phenylalanine Hydroxylase and Its Interaction with
Pterin Cofactor. J Am Chem Soc 1999, 121 (7), 1528-1536; (c) Wasinger, E. C.; Mitic, N.;
Hedman, B.; Caradonna, J .; Solomon, E. I.; Hodgson, K. O., X-Ray Absorption
Spectroscopic Investigation of the Resting Ferrous and Cosubstrate-Bound Active Sites
of Phenylalanine Hydroxylase. Biochemistry-Us 2002, 41 (20), 6211-6217; ((1) Chow, M.
S.; Eser, B. E.; Wilson, S. A.; Hodgson, K. 0.; Hedman, B.; Fitzpatrick, P. F.; Solomon,
E. I., Spectroscopy and Kinetics of Wild-Type and Mutant Tyrosine Hydroxylase:
Mechanistic Insight into O-2 Activation. J Am Chem Soc 2009, 131 (22), 7685-7698.

27. Eberlein, G.; Bruice, T. C.; Lazarus, R. A.; Henrie, R.; Benkovic, S. J ., The
Interconversion of the 5,6,7,8-Tetrahydropterin, 7,8-Dihydropterin, and Radical Forms of
6,6,7,7-Tetramethyldihydropterin - a Model for the Biopterin Center of Aromatic Amino-
Acid Mixed-Function Oxidases. J Am Chem Soc 1984, 106 (25), 7916-7924.

28. Ellis, H. R.; McCusker, K. P.; Fitzpatrick, P. F., Use of a Tyrosine Hydroxylase
Mutant Enzyme with Reduced Metal Affinity Allows Detection of Activity with Cobalt in
Place of Iron. Arch. Biochem. Biophys. 2002, 408 (2), 305-307.

29. Guroff, G.; Levitt, M.; Daly, J .; Udenfrie.S, Production of Meta-Tritiotyrosine
fi'om P-Tritiophenylalanine by Phenylalanine Hydroxylase. Biochem. Biophys. Res.
Commun. 1966, 25 (2), 253-&.

30. Renson, J .; Daly, J .; Weissbac.H; Witkop, B.; Udenfiie.S, Enzymatic Conversion
of 5-Tritiotryptophan to 4-Tritio-5-Hydroxytryptophan. Biochem. Biophys. Res.
Commun. 1966, 25 (5), 504-&.

31. Rich, P. R.; Salerno, J. C.; Leigh, J. S.; Bonner, W. D., Spin 3-2 Ferrous-Nitric
Oxide Derivative of an Iron-Containing Moiety Associated with Neurospora-Crassa and
Higher Plant-Mitochondria. Febs Lett 1978, 93 (2), 323-326.

32. (a) Hishinuma, Y.; Kaji, R.; Akirnoto, H.; Nakajima, F.; Mori, T.; Kamo, T.;

Arikawa, Y.; Nozawa, S., Reversible Binding of No to Fe(Ii)Edta. Bull. Chem. Soc. Jpn.
1979, 52 (10), 2863-2865; (b) Arciero, D. M.; Lipscomb, J. D.; Huynh, B. H.; Kent, T.

35

A.; Munck, E., Electron-Paramagnetic-Res and Mossbauer Studies of Protocatechuate
4,5-Dioxygenase - Characterization of a New Fe-2+ Environment. J Biol Chem 1983,
258 (24), 4981-4991.

33. Han, A. Y.; Lee, A. Q.; Abu-Omar, M. M., Epr and Uv-Vis Studies of the Nitric
Oxide Adducts of Bacterial Phenylalanine Hydroxylase: Effects of Cofactor and
Substrate on the Iron Environment. Inorg Chem 2006, 45 (10), 4277-4283.

34. Fitzpatrick, P. F., The Ph-Dependence of Binding of Inhibitors to Bovine Adrenal
Tyrosine-Hydroxylase. J Biol Chem 1988, 263 (31), 16058-16062.

35. Frantom, P. A.; Seravalli, J .; Ragsdale, S. W.; Fitzpatrick, P. F., Reduction and
Oxidation of the Active Site Iron in Tyrosine Hydroxylase: Kinetics and Specificity.
Biochemistry-Us 2006, 45 (7), 2372-2379.

36. Daubner, S. C.; Fitzpatrick, P. F., Site-Directed Mutants of Charged Residues in
the Active Site of Tyrosine Hydroxylase. Biochemistry- Us 1999, 38 (14), 4448-4454.

37. Ramsey, A. J .; Hillas, P. J .; Fitzpatrick, P. F., Characterization of the Active Site
Iron in Tyrosine Hydroxylase - Redox States of the Iron. J Biol Chem 1996, 271 (40),
24395-24400.

38. Keefer, L. K.; Nims, R. W.; Davies, K. M.; Wink, D. A., "Nonoates" (1-
Substituted Diazen-l-lum-l,2-Diolates) as Nitric Oxide Donors: Convenient Nitric Oxide
Dosage Forms. Method Enzymol 1996, 268, 281-293.

39. Shiman, R.; Gray, D. W.; Pater, A., Simple Purification of Phenylalanine-
Hydroxylase by Substrate-Induced Hydrophobic Chromatography. J Biol Chem 1979,
254 (22), 1300-1306.

40. Storm, C. B.; Shiman, R.; Kaufman, S., Preparation of 6-Substituted Pterins Via
Isay Reaction. J. Org. Chem. 1971, 36 (25), 3925-&.

41. Stark, E.; Breitmai.E, 5-Desazapteridines - Synthesis and Nmr-Spectroscopy.
Tetrahedron 1973, 29 (14), 2209-2217.

42. Shiman, R.; Akino, M.; Kaufman, S., Solubilization and Partial Purification of
Tyrosine Hydroxylase from Bovine Adrenal Medulla. J Biol Chem 1971, 246 (5), 1330-
&.

43. Guroff, G.; Reifsnyd.Ca; Daly, J ., Retention of Deuterium in P-Tyrosine Formed
Enzymatically from P-Deuterophenylalanine. Biochem. Biophys. Res. Commun. 1966, 24
(5), 720-&.

44. Kappock, T. J .; Harkins, P. C.; Friedenberg, S.; Caradonna, J. P., Spectroscopic

and Kinetic Properties of Unphosphorylated Rat Hepatic Phenylalanine Hydroxylase
Expressed in Escherichia Coli - Comparison of Resting and Activated States. J Biol

36

Chem 1995, 270 (51), 30532-30544.

45. Stetter, J. R.; Li, J ., Amperometric Gas Sensors - a Review. Chem Rev 2008, 108
(2), 352-366.

46. F eelisch, M.; Stamler, J. S., Methods in Nitric Oxide Research. J. Wiley:
Chichester ; New York, 1996; p xix, 712 p., 1 leaf of plates.

47. (a) Daubner, S. C.; Hillas, P. J .; Fitzpatrick, P. F., Characterization of Chimeric
Pterin-Dependent Hydroxylases: Contributions of the Regulatory Domains of Tyrosine
and Phenylalanine Hydroxylase to Substrate Specificity. Biochemistry- Us 1997, 36 (39),
11574-11582; (b) Hertzler. Comparative Studies of Select Mutants of Rat Phenylalanine
and Tyrosine Hydroxylases: Mechanistic Investigations. Ph. D. Thesis, Boston
University, 2006.

48. Pilbrow, J. R., Effective G-Values for S = 3/2 and S = 5/2. J Magn Reson 1978, 31
(3), 479-490.

49. Brown, C. A.; Pavlosky, M. A.; Westre, T. E.; Zhang, Y.; Hedman, B.; Hodgson,
K. 0.; Solomon, E. I., Spectroscopic and Theoretical Description of the Electronic-
Structure of S=3/2 Iron-Nitrosyl Complexes and Their Relation to O-2 Activation by
Nonheme Tron Enzyme Active-Sites. J Am Chem Soc 1995, 11 7(2), 715-732.

50. Bloom, L. M.; Benkovic, S. J .; Gaffney, B. J ., Characterization of Phenylalanine-
Hydroxylase. Biochemistry-Us 1986, 25 (15), 4204-4210.

51. Wallick, D. E.; Bloom, L. M.; Gaffney, B. J .; Benkovic, S. J ., Reductive
Activation of Phenylalanine-Hydroxylase and Its Effect on the Redox State of the Non-
Heme Iron. Biochemistry- Us 1984, 23 (6), 1295-1302.

52. Salerno, J. C.; Siedow, J. N., Nature of the Nitric-Oxide Complexes of
Lipoxygenase. Biochim Biophys Acta 1979, 5 79 (1), 246-251.

53. Ray, M.; Golombek, A. P.; Hendrich, M. P.; Yap, G. P. A.; Liable-Sands, L. M.;

Rheingold, A. L.; Borovik, A. S., Structure and Magnetic Properties of Trigonal
Bipyramidal Iron Nitrosyl Complexes. Inorg Chem 1999, 38 (13), 3110-3115.

37

CHAPTER 11
AN EPR INVESTIGATION OF A N,N,O FACIALLY COORDINATED NON-
HEME IRON MODEL SYSTEM

2.1 INTRODUCTION

In the previous chapter we presented the CW—EPR spectra of two aromatic amino
acid non-heme iron hydroxylases that were treated with NO to produce a Fe-NO complex
that was consistent with a S = 3/2 spin state that had a large axial zero field interaction. It
was also shown that these CW-EPR spectra change in rhombicity and line-width upon
substrate and co-substrate addition. In order to more thoroughly understand the electronic
structure of these non-heme Fe enzymes and relate the observed changes to structure, it
would be USCfill to examine a magnetic model system. A Fe(II)-NO model system would
allow us to more easily characterize the EPR spectrum without having to consider the
complication presented by the protein framework or the substrates.
2.2 Fe-NO ELECTRONIC STRUCTURE

In their review of metal nitrosyl complexes and their bonding, Enemark and
Feltham coined a notation which is useful for describing metal nitrosyls. This notation is
written as {MNO}“, where M is a transition metal and n stands for the number of d-type
electrons in the complex, which includes those from the metal and one from the NO.1 For
example, the F e(II)-NO system we've examined in the previous chapter would be
classified as {FeNO}7, this includes the 6 valence electrons of the Fe(II) plus one from
the NO. While this notation is useful as a basic description of these systems, it does not

accurately describe the specific electronic structure of these complexes. Knowledge of

38

the the specific electronic structure is necessary for the structural characterization of
metal nitrosyls using EPR methods.

Table 2.1 : {FeNO}7 F ormalisms

 

 

 

 

 

 

Reference Fe Fe spin NO spin Coupling
d-electrons

Brown 7- 5 S=5/2 S=1 anti-ferromagnetic

Bill3 7 S=3/2 S=0 none

Hoffman4

Pohl5 5 S=1/2 S= ferromagnetic

Salerno 6 6 S=2 S=l/2 anti-ferromagnetic

 

 

 

 

 

 

The electronic structure of non-heme {FeNO}7 has been examined in a variety of
studies using different spectroscopic methods, magnetic measurements and calculations.
These studies have proposed a number of formalisms to describe the {FeNO}7
complexes, the formalisms which are consistent the CW-EPR results, namely the S = 3/2
spectrum, are shown in table 2.1.

One of the more recent and accepted interpretations was proposed by the Solomon
group. It was suggested that {FeNO}7 is best described as a high spin(S = 5/2), ferric(d5)
iron which is anti-ferromagnetically coupled to a reduced nitric oxide(S = 1), this yields a
total spin, S = 3/2.7 This interpretation is strongly supported by a number of different
experimental approaches including X-ray absorption spectroscopy, Raman spectroscopy
and Magnetic Circular Dichroism.2

Fe K-edge X-ray absorption spectroscopy(XAS) performed on F e(EDTA)(N O)

and Fe(Me_,3TACN)(NO)(N3)2 produced a pre-edge feature with a line shape and energy

39

consistent with an iron in a high spin ferric oxidation state.2 In addition, the extended X-
ray absorption fine structure(EXAF S) analysis of the first coordination shell provided
distances which, when compared to F e(III)(EDTA)(HZO) and Fe(II)(EDTA)(H20), are
shorter than those of a ferrous complex, closer to that of a ferric species.8 This description
of the {FeNO}7 unit was further supported by resonance Raman spectroscopy. The NO
and the NO-F e vibrational frequencies were measured and gave values between the
expected values for NO0 and NO", consistent with a partially reduced NO ligand.2
Charge transfer transitions observed in absorption and MCD spectroscopy also agree with
the description of {FeNO}7 composed of a S = 5/2 ferric anti-ferromagnetically coupled
to a S = 1 NO‘ ligand.2

There are conflicting results for this interpretation from Méissbauer spectroscopy
and EPR. Mfissbauer spectra collected on the non-heme cofactor iron of putidarnonooxin
provided an isomer shift which, when compared to shifts of known electron
configuration, was too large for a d5 configuration and was assigned as d7.3 These results
combined with the observed S = 3/2 EPR spectrum prompted the interpretation proposed
by Bill et a1, where the iron is d7 and contains all the unpaired spin density.3 Similar
results were obtained from a Méssbauer study of Deoxy Hemerythrin, again the isomer
shift for the {Fe-NO}7 center was not consistent with the Solomon interpretation of a d5
Fe(S = 5/2).9 In this work the authors suggested that rather than using an idealized
electronic configuration of Fe or NO that it is more appropriate to consider the {FeNO}7
as a complete unit where some valence electrons are shared across the moiety.

Further evidence for the breakdown of the idealized electronic configuration

40

suggested by Solomon comes from EPR measurements made in the Hofiman lab.
Deuterium Electron Nuclear Double Resonance (ENDOR) measurements were
performed on the non-heme ferrous sites in naphthalene 1,2-dioxygenase4a and ACC
oxidase4b treated with NO. The 2H hyperfine couplings measured in these ENDOR
studies provided dipole-dipole distances for the coupling between specifically deuterated
substrates and the Fe center that compared well with X-ray crystallographic studies. In
addition, a recent 2H-electron spin echo envelope modulation (ESEEM) spectroscopy
study of taurine dioxygenase in our lab provided deuterium dipolar couplings for
specifically deuterated substrate taurine to the iron that also agreed well with predictions
from crystal structures.10

The two formalisms presented here would interpret the structural EPR results,
specifically the dipolar interaction, differently. An interaction between a nuclear dipole
and the first formalism, a S = 5/2 ferric anti-ferromagnetically coupled to a S = 1 NO—
ligand, cannot be explained by the point dipole approximation, since the individual spins
of the Fe and NO and their coupling in the {FeNO}7 unit need to be considered. This
leads to the necessity to use spin projection. factors to properly describe the dipole-dipole
coupling.” This entails calculating multiple point dipole approximations and scaling
these interactions by the degree of coupling between the two spin containing species.12
The second formalism, in which the NO donates an electron to the Fe producing a d7 S =
3/2 center, can properly utilize the point dipole approximation. This formalism works in
the context of the previously reported ENDOR and ESEEM results where the dipolar

coupling was reported in terms of the distance between Fe and the deuterated substrate

41

using the point dipole approximation, results which agreed well with distances observed
in crystallographic studies of the same systems.4a 4b '0
2.3 N,N,O MODEL SYSTEM

Our collaborators in the lab of John Caradonna at Boston University devised a
series of three models in order to mimic the magnetic structure of the His-His-Glu facial
triad observed in the non-heme iron hydroxylases. The series of ligands, shown in figure
2.1, contains a facially coordinating N,N,O group that is modified by one or two
additional carboxyl groups, this allows for a comparison of the properties of iron
complexes containing fac-N,N,O(N 201) with three fac exchangeable sites,
N,N,O,O(N202) with cis exchangeable sites, and N,N,O,O,O(N203) with one
exchangeable site, in which the number of coordinated carboxylate ligands is varied from

one to three. The CW-EPR spectra acquired on these three models should provide

insights into the Hamiltonian parameters we've collected for TyrH and PheH.

3 Ti
\ /
0 O
(N2F’e/L \N\O\ /L NK—V—é
N \ (N’FIG‘L < /Fe\

Ltof . ero LID
O 0

Figure 2.1: Structures of F eNZOl (left), FeNZO2
(middle) and FeN203 (right)

As suggested in the previous chapter, the CW—EPR spectra of { F eNO}7 systems
do not contain any easily measurable structural information. Much of the ligand
hyperfine structure is obscured by inhomogeneous broadening in the CW-EPR spectrum.

Pulsed EPR techniques such as electron spin echo envelope modulation(ESEEM)

42

spectroscopy or pulsed electron nuclear double resonance(ENDOR) have been developed
to address this problem. Application of these techniques to the model system will provide
more detailed information about the local magnetic environment of the {FeNO}7 and in
combination with the known structural data we can characterize the contributions from
the ligand and solvent. These assignments will be useful in helping to understand the
more complicated magnetic environment of the enzymes.
2.4 WATER ANALYSIS

One important aspect of this model system is the variable number of exchangeable
coordination sites which will allow for the development of a method to characterize the
water coordination in non-heme iron. Some previous work for the determination of water
coordination in non-heme iron enzymes includes 170 line broadening in a variety of
different systems13 14 and ENDOR studies of naphthalene 1,2-dioxygenase4a and ACC
oxidase4b. Although the 170 line broadening studies provide an indication whether or not
the {FeNO}7 has a coordinated water ligand, it does not provide any quantitative
information regarding the water. The basis behind this technique is to compare the CW-
EPR lineshape of {FeNO}7 in 170 enriched H20 to the line shape in un-enriched H20, if
a change in the linewidth is observed, the result of hyperfine broadening, it is concluded
that the {FeNO}7 has a coordinated water. ‘3 The problem with this technique is that it is
impossible to directly measure the hyperfine interaction and all information regarding the
number of waters and the location with respect to the magnetic axes is lost.

These problems were overcome in the Hoffman lab with the application of more

advanced EPR techniques such as ENDOR which can be used to directly probe the

43

hyperfine values. In ACC oxidase the presence of Fe bound water was established using
ENDOR to detect the hyperfine coupling to H2170 and D20.4b These two methods did
provide hyperfine coupling values for the interaction with a coordinated water however
they both face complications. The 17O resonance faced significant overlap with non-
exchangeable nitrogen features, which was compounded with overlap from non-
exchangeable hydrogen at fields lower than g~2.8 Additionally, due to the broad
linewidth of the 170 feature it was impossible to determine the orientation or the number
of waters coordinated.4b The 2H ENDOR also provided couplings that were consistent
with a coordinated water however they were unable to perform a field dependent analysis
due to the signal rapidly attenuating at g < 3.8.4b

The problems seen in the ENDOR experiments, namely the resonance overlap,
could be avoided with a 2-dimensional experiment. One such method, hyperfine sublevel
correlation(HYSCORE) spectroscopy largely avoids the problem of resonance overlap by
correlating the hyperfine frequencies of the nuclei in 2 dimensions hopefully separating
features that would otherwise overlap in l dimension.15 Application of HYSCORE
spectroscopy to the model system should allow for the systematic characterization of the
interaction with coordinated water without the use of extensive isotopic labeling.
2.5 HYDROGEN HYPERFINE COUPLING

The following discussion of the hydrogen hyperfine interaction is based on the

interpretation by Hutchinson and McKay. '6 17 The interaction of the {FeNO}7 center

with a distant hydrogen(I = 1/2) can be characterized with the following Hamiltonian:
H=He—g33,833Ho-I+S-24°I [2.1]

44

The first term, the electron Hamiltonian, is described in the previous chapter in equation
1.1. This term determines the energy differences between the sub-levels of the i1/2
Kramers doublet of the {F eNO}7 unit. As was shown in equations 1.2a-c, the zero field
interaction can be rewritten in terms of three effective g-values. Using these three
effective g-values and the projection of the magnetic field into the principle axis system

of the zero field interaction:

1
x

H0=Ho I, [2.2]
I.

in which ll=cos (pe sin 08, 12: sin (06 sin 033, l3= cos 0e are the direction cosines, the

electron Hamiltonian can be rewritten as:

A A A . gx3efl 0 0 I333
He=BeHo Sx 5y 5:1 0 8,3,7 0'1, [2.3]
0 0 g3 efl I:

an effective electron Zeeman interaction. Diagonalizing this portion of the Hamiltonian
allows for the electron spin angular momentum operator to be rewritten in the eigenbasis

of the effective g-tensor principle axis system as:

 

 

33 S", gx,eff x
se_-ge—fl- g333fl13 [2.4]
gz,efllz
where
_ 2 2 2_ hv
gefl—l(g,3efll,) +(g333fl13) +(g,3,firl,) _B H [2.5]
e 0

This effective electron Zeeman interaction is the dominant term in equation 2.1 being

several orders of magnitude greater than the remaining two terms which can be treated as

45

perturbations on the eigenstates. The first term, the nuclear Zeeman interaction, is
independent of the g-frame and represents a field dependent splitting of the nuclear spin
sub-levels. The second term is the hyperfine interaction, were A represents the sum of the
dipolar interaction and the isotropic Fermi contact interaction. In the case of a distant
hydrogen, the Fermi contact term is near zero and this term can be neglected. The
remaining term, the dipolar interaction, is the energy of the interaction between the field
induced by the {F eNO}7 at the hydrogen and the nuclear magnetic moment. By assuming
that the {FeNO}7 unit is a point dipole located at the Fe, this field is given as

Hd=—-(3rF—E) [2.6]

where r is the distance between the Fe and the hydrogen, r is the unit vector pointing in

the direction of the vector connecting the Fe and hydrogen, E is the unit dyadic and ue is

the electron magnetic moment given as

Iii-8.85“. [2.71

where g is a diagonal matrix composed of the effective g values of the dipole and S3, is

the electron spin operator given in equation 2.4. Substituting equation 2.7 into 2.6 and

f.

 

expressing in the g axis system the dipolar field is given as:
2
"'B 5’? gx,efl(3rl_1) 3gx,eflrl.r2 3gJt,eflrl.r3
" _ e e3 3 2_ 3
Hd— r3 3gy333fi.rl r2 gy333fi,(3r2 1) 3gy333flr3 r2 [2.8]

2
3gz.efi’1"’3 38:.efl’2"3 g:,efl(3r3—1)

where rl=cos (0N sin 0N, r2: sin (0N sin 0N, r3: cos 0N, are the three direction cosines
relating the vector that connects the Fe and the hydrogen to the principle axis of the g

tensor. This field interacts with the nuclear dipole producing the hyperfine term seen in

46

equation 2.1:

Hd-quSe-A-I [2.9]

I 2
gx.efl(3rl_1) 3gx,efi’rl°r2 3gx,efl'rl'r3

-B B 3
A—.e—N_H. 3r§_1)

— 3 3gy.eflrl.r2 gy.efl( 3gy,eflr3.r2
2
3gz,efl'rl°r3 3gz,eflr2.r3 g:,efl(3r3—1)

[2.10]

r
Often the dipole-dipole coupling is referenced to g e = 2.0023, however as can be seen in
equation 2.10 the dipolar coupling is scaled by the effective g-values, this means that
there is an enhancement of the interaction depending on its orientation with respect to the
g principle axis system.

Due to the anisotropy in the effective g tensor, when operating at a fixed field and
frequency, the resonance condition seen in equation 2.5 is only satisfied with a limited set
of 0 6'5 and (03's resulting in single crystal like spectra. This also results in a field
dependance in the effective spin operator, ge’ which scales the hyperfine interaction by
the effective g value. For example a hyperfine coupling recorded near g=4 would provide
a value double that of the same coupling at g = 2.

2.6 EXPERIMENTAL

All reagents used are commercially available and most were used as received.
Prior to its use methanol was distilled over magnesium metal and freeze-pump-thawed
(FPT) for four cycles. All manipulations involving ferrous complexes and nitrosyl
adducts were carried out in an inert atmosphere (N 2) glovebox, except where noted.

Nitric oxide (N0) gas (99+%) was purchased from Air Gas (Maumee, OH) and used

without further purification.

47

The ferrous complexes F e”(NZO3)(L), F e”(N202)(L2) and Fe"(N201)(L3) were
prepared by literature procedures. '8 NO adducts of Fe"(N203)(L), Fe"(N202)(L2) and
Fe"(N201)(L3) were prepared as follows. Stock solutions of Fe"(N203)(L), Fe"(N202)
(L2) and Fe"(N201)(L3) were prepared in degassed H20 and D20 pH 6.5 MOPS buffer
and brought out of the glovebox in a schlenk flask. The flask was evacuated and the
headspace above the stirring sample was subsequently charged with 3-5 PSI of N0 gas at
which point the colorless solution turned immediately orange with the formation of
{FeNO}7. The sample was allowed to stir for ~1 min, the flask was evacuated and the
sample was brought back into the inert atmosphere glovebox for further manipulation.
Samples for EPR analyses were diluted to 5mM with degassed 50:50 glycerol/H(D)ZO as
glassing agent for ~40% final glycerol content. The samples were loaded into EPR tubes,
frozen in N2 (1), and shipped in a dewared vessel at N2 (1) temperatures.

CW—EPR spectra were recorded on a Bruker ESP300E X-band EPR spectrometer
equipped with a TE102 cavity and an Oxford Instruments ESR—900 liquid helium cryostat
with a model [TC-502 temperature controller. CW—EPR measurements were made with
the following conditions: sample temperature, 4.0 K; microwave frequency, 9.46 GHz;
microwave power, 1 mW; magnetic field modulation frequency, 100 kHz; field
modulation amplitude, 5 G; conversion time, 163 ms. CW-EPR data were analyzed using
the X-SOPHE program available from Bruker Biospin Corporation.

Pulsed EPR measurements were made on a Bruker E-680X spectrometer
operating at X-band and equipped with a model ER 4118X-MD-X5-W1 probe that

employs a 5mm dielectric resonator. The temperature was maintained at 4.0 K using an

48

Oxford Instruments liquid helium flow system equipped with a CF -935 cryostat and an
ITC-503 temperature controller. HYSCORE data were collected using a four-pulse
sequence, 90°-r-90°-t1-180°-t2-90°, with 90° microwave pulse widths of 16 ns (FWHM)
and the 180° microwave pulse widths was adjusted between 24-36 ns for optimal turning
angle. The tau value was chosen to suppress the hydrogen matrix contribution; due to
short phase memory times, tau values were restricted to less than 200 ns. An integration
window of 28 ns was used to acquire spin echo amplitude, and data set lengths were 128
points with a 24 ns time increment. The data were processed using tools provided by
Bruker. The time domain data was corrected by subtracting a third order polynomial,
optimized at each slice, from both time dimensions. The resulting data was tapered with a
Hamming window function in both dimensions and the real part was 2-dimensional
Fourier transformed. The HYSCORE spectra were obtained by taking the absolute value
of the transforms.

Simulations of the hyperfine couplings were accomplished using scripts written in
MATLAB. The calculations were performed in the time domain using the density matrix
formalism of Mirns.19 The 1H HYSCORE scripts were based on the analytical formula
given by Gemperle,20 while the I = 1 script is based on a simplified summation formula
given by Dikanov2'. The orientation-dependent Hamiltonian used was modeled after
work by Hutchison and McKay and is described above.16 Processing of the simulations,
also performed in Matlab, was identical to that described above for the experimental
spectra.

2.7 RESULTS

49

2.7.1 FERRIC CW-EPR

As an initial study of the ligand system, we examined CW-EPR spectra of the
ferric complexes, this was to determine if in solution the ligand coordinates the iron in a
mono-nuclear fashion as observed in the crystal structures obtained on the complexes.
Seen in figure 2.2 are the CW—EPR spectra of Fe(III)-N20l (a), Fe(III)-N202 (b) and

Fe(III)-NZO3 (c). All three model

 

complexes show spectra typical of a single
high-spin Fe(III) in a ligand field of

3 rhombic symmetry which is indicated by a

 

low field feature at g ~ 9 and a broad

,3,
' b
g feature with an inflection at g = 4.3. Shown
33 .
v in red are the corresponding simulations for
a S = 5/2 system which utilized a
c

Hamiltonian similar to the one described in

Chapter 1. All three simulations utilized

 

 

 

50° 1000 150033332383 250° 300° 350° identical g-,D-, E/D-, D-strain and intrinsic

linewidth values, in order to account for the
Figure 2.2: CW—EPR of ferric N201 (a),
N202 (b), and N203 (0) shown in black with

corresponding srmulatrons shown in red. adjusted. In the calculations the g-value was

changes observed the E/D strain was

taken to be isotropic with gO = 2.00, the zero field interaction parameters were taken to be
rhombic with D = 0.30 cm'1 and E/D = 0.23. The spectra show broad features at g = 4.3,

because of this the intrinsic EPR linewidth selected for all three simulations was chosen

50

to be large, 0.030 cm-1 with a strain in D of 0.1 cm". The E/D strain was adjusted to
account for the lineshape changes, a strain in E/D of 0.20 was used in fig. 2.2a, a strain in
E/D of 0.16 was used in fig 2.2b, and a strain in E/D of 0.06 was used in fig 2.2c.
2.7.2 {FeNO}7 CW-EPR

To properly compare the results from the model system to those from the enzyme

it is necessary to examine the {F eNO}7 adduct. The model system was prepared using

 

ferrous iron in H20 and was treated with

NO. The CW-EPR spectra of the three

 

 

models {FeNO}7-N20l (a), {FeNO}7-
N202 (b) and {FeNO}7-N203 (c) are

D shown in figure 2.3. The three spectra show

 

 

dX'IdH (a.u.)

line-shapes indicative of the lower Kramers
doublet of a S = 3/2 spin system with |D| >>

geBeHo. Shown in red are the simulations

 

generated using the procedure outlined in

 

 

1 L

1200 1400 1600 1800 2000 mthe previous chapter. Similar to the previous V
Field (G)

 

results of the ferric model system all three

Figure 23: CW-EPR of {FeNO}7 N20I (a), simulations utilized identical g-, 0-, E/D-,
N202 (b), and N203 (c) shown in black with

corresponding simulations shown in red. D-strarn and 1ntr1nsrc lrnewrdth values. To

account for the changes in the observed
lineshape, the E/D strain was adjusted. In the calculations the g-value was taken to be

isotropic with gO = 2.00, the zero field interaction parameters were taken to be nearly

51

axial with D = 10.0 cm'1 and E/D = 0.0115. The intrinsic EPR linewidth selected for all
three simulations was taken to be 0.0015 cm'1 with no strain in D. The E/D strain was
adjusted to account for the lineshape changes, a strain in E/D of 0.0075 was used in fig.
2.3a, a strain in ED of 0.0045 was used in fig 2.3b, and a strain in E/D of 0.0035 was
used in fig 2.3c
2.7.3 {FeNO}7 HYSCORE

As suggested in the previous chapter, the CW-EPR results do not provide details
of metal ligation. More advanced techniques are necessary to determine specific 3
structural information. One specific technique which has proven useful for evaluating and
interpreting complicated spectra is HYSCORE spectroscopy.22 HYSCORE spectra of the
{FeNO}7 model complexes were acquired at multiple positions in the field range from g
= 4 to g = 2. This was done in order to properly take into account the single crystal like
selection of orientations imposed by the anisotropy in the EPR spectrum. The analysis of
these spectra utilized multiple field positions; however, to give a brief overview, only
representative spectra acquired at 1980 G(geflc = 3.489) and 3100 G(geff= 2.229) will be
shown in this section.

Seen in figures 2.4 and 2.5 are the HYSCORE spectra of {FeNO}7-NZOl at 1980
G and 3100 G, respectively. Both spectra show features in the (++) quadrant in the
frequency regime expected for hydrogen hyperfine couplings. The spectrum collected at
1980 G shows disperse clouds along the frequency diagonal with intensity at 9.5 MHz
and 11 MHz. Weak crosspeaks were observed in the (++) quadrant in range from (5,13)

MHz to (7,9) MHz, this feature at most had 30% of the maximum intensity of the feature

52

 

zobBt 1 1 1 1 1 r r T T I
18
16-

14-

F2 (MHz)
0 3
I l

 

 

 

 

6" O

41—

2F _

0 l L 1 J 1 1 n 1 1 1 1 1 1 1 n 1 1 l 1

-20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20
F1(MHz)

Figure 2.4: HYSCORE spectrum of {F eNO}7-NZOl recorded at 1980 G
20 I l

 

.1
_4
.1
q
q
—1
.1
.1
_1

18'

 

 

 

 

O
8 #- Q o 0 _
._ °° t1 -
O
4 '- w -l
o. e % n .
2 - _
9
o l l l l l l l l l l A l l l I l l J I
20 1a 16 14 12 -10 -e e 4 2 o 2 4 10 12 14 1e 1e 20
F1(MHZ)

Figure 2.5: HYSCORE spectrum of {FeNO}7-NZOl recorded at 3100 G

observed along the diagonal. At 3100 G a set of two distinct crosspeaks were observed in
the (++) quadrant, the more intense feature has a maximum at (115,155) MHz, the other

higher frequency feature is 30% less intense and has a maximum at (11.5,17) MHz.

53

18

16

14

F2 (MHz)
3

 

 

L 1

 

J

 

 

-18

-16

-14

-12 -

0
F1 (MHZ)

12

Figure 2.6: HYSCORE spectrum of {FeNO}7-N202 recorded at 1980 G

F2 (MHZ)

20

18

16

14

14

16

18 2O

 

l

 

I

l

I

I

I

l

l

J

I

l

l

r

 

 

 

F1 (MHz)

Figure 2.7: HYSCORE spectrum of {FeNO}7-N202 recorded at 3100 G

Seen in figures 2.6 and 2.7 are the HYSCORE spectra of {F eNO}7-N202 at 1980

G and 3100 G, respectively. In figure 2.7, non-diagonal features are observed in both the

(++) and (-+) quadrants. The crosspeaks in the (++) quadrant fall in a region which would

54

be consistent with hydrogen hyperfine couplings and they do show the expected field
dependent behavior. The crosspeaks themselves are composed of several different
components with varying maxima. In the (-+) quadrant a pair of crosspeaks at (11,13)
MHz is observed, these features do not show the field dependent behavior of hydrogen
and therefore must be from strongly coupled nitrogen, the only other magnetic nuclei in
the system. At 3100 G crosspeaks are also observed in both quadrants. Similar to the
features observed in figure 2.5, two sets of hydrogen crosspeaks are seen in the (++)
quadrant. The more intense feature has a maximum at (115,155) MHz and a line shape
similar to the analogous peak in figure 2.5. This intense feature largely obscures a second
component of higher fiequency at (12,17) MHz, compared to the same component in
figure 2.5 there is a decrease in amplitude relative to the lower frequency hydrogen
feature. In the (-+) quadrant, two sharp features which would be consistent with strongly

coupled nitrogen are observed, one at (-3,10) MHz and the other at (-8.5,12) MHz.

 

20 1 1 I 1 1 1 T
18 - Q -

16-

F2 (MHz)
3
I
O

 

 

8- 3 . -
O

6. Pest -

4- 8 -

21— 8 Q -

0 l l l #1 l l l

 

 

L l
-20 -10 -16 -14 -12 -10 -8 -6 .4 -2
F1(MHz)

Figure 2.8: HYSCORE spectrum of {FeNO}7-N203 recorded at 1980 G

55

Seen in figures 2.8 and 2.9 are the HYSCORE spectra of {FeN0}7-N203 at 1980
G and 3100 G, respectively. These two spectra show similarities to the spectra acquired
on N202 at the same field positions. At 1980 G there are the hydrogen crosspeaks in the
(++) quadrant and the strong nitrogen peaks in the (-+) quadrant at (-10.5,13) MHz. The

spectrum acquired at 3100 G contains only a single hydrogen crosspeak in (++) quadrant

 

20 I I I l I l l I I l l I 1 l I I l T T
18 — -
16 —

14-

F2 (MHz)
8
I

 

 

 

 

18 20

F1 (MHZ)

Figure 2.9: HYSCORE spectrum of {FeN0}7-N203 recorded at 3100 G

which has a local maxima at (115,155) MHz, N203 lacks the higher fi'equency hydrogen
feature seen in the other two models. The (-+) quadrant at 3100 G contains three sharp
features which would be consistent with strongly coupled nitrogen, one at (-3,9), (-5,7.5)
and at (-9,12) MHz.

These HYSCORE spectra provide hyperfine coupling information about the
magnetic nuclei near the {FeNO}7 center. The ligand systems provide two nitrogens, a

variable number of non-exchangeable hydrogens and in the open coordination sites water

56

molecules. In order to properly assign the features observed in the HYSCORE it is
necessary to perform simulations.
2.7.4 WATER ANALYSIS

In the HYSCORE spectra at both low field and high field features consistent with
hydrogen were observed. In order to differentiate the non-exchangeable ligand hydrogens
from those on the solvent, samples of the three models in D20 were prepared.
HYSCORE spectra were acquired on the three model systems at multiple field positions
to identify the exchangeable features. Upon examination of the spectra it was found that
exchangeable features are only present in the spectra of N201 and N202 acquired in the
range from g = 2.5 to g = 2.

Shown in figure 2.10 is the HYSCORE spectrum from N201 in a D20 buffer at

3100 G. In comparison to figure 2.5 we

 

observe a disappearance of the higher
4 frequency feature at (11.5,17) MHz. The D20
3 buffer also gives rise to two new crosspeak

' features at (9.5.13) and (135,175). Based on

F2 (MHz)

3 the observation that they lie on the same

‘ diagonal as the previously reported peak at

 

 

 

 

1‘7 19(11.5,15.5) its reasonable to assume that they

F1 (MHz)

are in fact combination lines. The presence of

Figure 2'710: HYSCORE spectrum 0f the D20 buffer results in an extremely intense
{ F eNO} -N20l in D20 recorded at 3100

G feature at the Larmour fiequency of the

57

deuteron arising from dipolar coupling to the bulk solvent and generally referred to as a
matrix line. Experimentally, it is possible to suppress the frequency associated with the
H20 matrix peak however the low frequency of the D20 matrix line makes application of
this trick difficult and at the expense of signal to noise ratio so it is not used. The matrix
line at 2.1 MHz from the D20 adds new frequency components at sums and differences
of existing, real, features. The two new sets of crosspeaks are approximately at (+2,+2)
MHz and (-2,-2) MHz from initial real crosspeaks. These results show that the higher
frequency feature at (11.5,17) MHz seen in of the spectrum of N20] in H20 are fi'om
solvent molecules.

To properly simulate the feature identified as water we needed to examined its
field dependent behavior. Shown in figures 2.11 — 2.15 is the (++) quadrant of the
HYSCORE of N20] at 2800 G, 2900 G, 3000 G, 3100 G, and 3200 G, respectively. The

water feature is barely resolved at the two extreme field positions shown and is not

 

 

 

 

 

 

 

 

 

19 19

17» 17»

151 15' i
’11? Tu‘
I I
E13 E13-
N N
u. LL

11* 11.

9. 4 9»

7\ 4 ’P J r 4 7 n I r r r

7 9 11 13 15 17 19 7 9 11 13 15 17 19

F1 (MHz) F1 (MHZ)

Figure 2.11: HYSCORE spectrum of Figure 2.12: HYSCORE spectrum of
{FeN0}7-N201 recorded at 2800 G {1~“eNo‘}7-Nzol recorded at 2900 G

58

 

 

 

 

 

 

 

 

 

19 . . . . . 19 . f
17' 17.
15* 15-
if E“
E13: 1 E13}
N N
LL LL
11» 11 J
9 1 9
7 t r r n r 7 n 4 n n 1
7 9 11 13 15 17 19 7 9 11 13 15 17 19

F1 (MHz) F1 (MHz)

Figure 2.13: HYSCORE spectrum of Figure 2.14: HYSCORE spectrum of

 

 

 

 

 

{FeNO}7-N201 recorded at 3000 G {FeN0}7-N20] recorded at 3100 G
19 . . . - '
17 _ 3 visible at field positions recorded higher or
15 3 3 lower than those. At 2800 G and 2900 G the
g 13 t 3 lineshape of the water feature appears to be
E 1 1 _ 3 composed of two slightly different
9 3 3 components, at the higher field positions these
7 3 3 two components overlap and are

7 9 11 13 15 17 19 3 3
F1 (MHz) indrstingurshable. It was observed that the

water feature moves with field along a

Figure 2.15: HYSCORE spectrum of
{F eN0}7-N201 recorded at 3200 G constant frequency, approximately along F2

above the diagonal or F1 below the diagonal.
This observation is key in identifying the proper orientation of the Fe

hydrogen dipole-dipole vector with respect to the Fe-NO bond. The overall higher

59

frequency characteristic of the water feature indicates that it has a stronger hyperfine
interaction than the other non-exchangeable hydrogen feature.

Shown in figures 2.16 — 2.20 are the simulations of the water feature at

2800 G, 2900 G, 3000 G, 3100 G, and 3200 G, respectively. The simulations utilized two
19

 

sets of hyperfine parameters to distinguish

‘ between the two features seen at 2800 G and
3 2900 G. One simulation used a dipolar
distance of 2.58 A and an angle between the
‘ dipole-dipole vector and Fe-N0 bond of 1.2

3 radians, the second feature used a dipolar

 

 

 

 

7 9 11 13 15 17 19distance of 2.56 A and an angle between the
F1 (MHZ)

dipole-dipole vector and Fe-NO bond of 1.25

3 3 1 . .
5 31332116- H HYSCORE srmulatron at radians. To generate the spectra these two

19,,,,3 19-.-

 

 

17-

 

 

 

 

 

 

 

 

9 11 13 15 17 19
F1 (MHz) F1 (MHZ)

7 9 11

Figure 2.17: lH HYSCORE simulation at Figure 2.18: 1H HYSCORE simulation at
2900 G 3000 G

60

 

 

   

 

 

 

 

 

 

19 . - 9 9 - 19
15, 15-
71? ’11?
I I
E13» g 13
N N
LI. Ll.
11 _ 11 '
9' 9.
7 1 1 A 1 L 7 A l n n A
7 9 11 13 15 17 19 7 9 11 13 15 17 19
F1 (MHz) F1 (MHz)

Figure 2.19: lH HYSCORE simulation at Figure 2.20: lH HYSCORE simulation at
3100 G 3200 G

simulations were assembled according to the product rule. 19 23

Seen in figure 2.7, we observe that N202 contains the higher frequency
feature that we've attributed to water. This was further confirmed when the spectrum of
{FeNO}7-N202 in D20 buffer was examined and the water feature was no longer
present. The simulation obtained for the second hydrogen feature accurately represented
the water feature observed in {FeNO}7-N202. The spectrum of N203 seen in figure 2.9,
as expected, does not contain the higher frequency feature that we've attributed to water
hydrogen a result further confirmed when examining the spectrum in D20.

2.7.5 LIGAND HYPERFINE ANALYSIS
In addition to the exchangeable features attributed to water ligands several other
non-exchangeable features are observed in the HYSCORE spectra. It can be assumed that

these non-exchangeable features are either from interactions with the chelating ligand or

61

the NO. Preliminary analysis of these features was performed on the N203 model system
in order ensure that the F e-NO axis was fixed and the uncertainty in the orientation is
minimized. Shown in figure 2.21 is a simulation performed at 3100 G using two different
nitrogens, the first nitrogen utilized the following parameters: A=[6.5,6.5,8.5] MHz, an
angle between the dipolar vector and the F e-NO bond of 0.23 radians, equ= 4.5 MHz
with an n=0.1; the second nitrogen used the following parameters: A=[7.5,7.5,9.5] MHz,

an angle between the dipolar vector and the Fe-NO bond of 1.3 radians, equ= 4.5 MHz

 

 

 

 

 

withann=0.1.
20 I I I I I T I I I I I I I I I T I I I
18- ~ ‘ -
16.. . . -1
14- ~ QR ° , —
A 12 _. . . ,o . q
I; 10 I ‘
5 N o :2 $3
8. . " . i d
Q“ q . o
6" . . . t o . . ..
' 0, %
4— o O V O 7 —
2 — _
o l l l l I l 41 41 L ¥ 1 L I L
-20 -18 -16 -14 -12 - o 2 4 6 8 1o 12 14 16 18 20
F1 (MHz)

Figure 2.21; 14N HYSCORE simulation at 3100 G

The ligand system contains 13 different hydrogens at a variety of orientations
found in the range between 3.2 — 4.0 A. The resolution of our experiment does not allow
us to distinguish between individual hydrogens making their assignment nearly
impossible. Preliminary simulations suggest the observed features arise only from the

closest hydrogens, those near 3.2 A.

62

2.8 DISCUSSION

EPR spectroscopic methods were used to examine a series of mono-nuclear non-
heme iron model complexes. These complexes were synthesized to mimic the facial
N,N,O coordination motif common in mono-nuclear non-heme iron oxygenases and to
provide a spectroscopic model to help in the analysis of more complicated enzyme
systems. To more thoroughly characterize the coordination sphere, the base N,N,O frame
was modified by one or two additional coordinating carboxylate arms.

The CW-EPR spectra recorded on the Fe(III) complexes were consistent with
what is expected for high-spin(ds) mono-nuclear complexes and the parameters used in
the simulations are commensurate with previous work on F e(III) model complexes24 and
non-heme Fe(III) sites in proteins.25 26 The increase in E/D strain observed in the EPR
measurements shows that the distribution of electronic structures represented in the
samples is increasing with the number of open coordination sites on the metal ion. This
result is not entirely unexpected as the increase in open coordination sites on the iron
allows for more solvent which, unlike the ligand, can adopt a range of different
conformers which adds to the distribution of structures. While these results are interesting
they do not provide any tangible structural information about the model complexes.
Pulsed EPR methods were also attempted on the ferric complexes with no success, a
result likely due to the large structural dispersion observed in the CW-EPR spectra.

When the model system is reduced and treated with NO, the CW-EPR spectra
show sharp resonances at g = 4 and g = 2, characteristic of a S = 3/2, {F eNO}7 unit. The

parameters used in the simulations were largely held constant in order to reflect the

63

structural similarities between the three models and the E/D strain was adjusted to
account for the lineshape differences. This leads to a trend for the {F eNO}7 complexes
similar to the one observed in Fe(III) namely that there is an increase in ED strain with
increasing number of open coordination sites or with an increase in ligand chelation the
E/D strain decreases. Again a similar justification can be used, the increase in the number
of solvent molecules, which can adopt a range of different conformers, adds to the
distribution of electronic structures observed. The observation of significant E/D strain in
both the ferric and ferrous-NO systems provides evidence that this effect is in part
independent of the NO coordination however the presence of the NO cannot be
neglected.

The three models, N203, N202 and N201 have 1, 2 and 3 open coordination sites
respectively, clearly in N203, the NO binds at the lone, open coordination site. However,
in the latter two models, the binding position of the NO is not defined by the ligand
structure. The observed increase in E/D strain with number of open coordination sites
could be a result of this ambiguity in the NO binding, instead of observing line
broadening we could be be observing a mixture of overlapping species. A recent study by
Borovik et al showed a correlation between the value of E/D and the angle of the Fe-N-O
bond, as the projection of the oxygen onto the plane perpendicular to the Fe-N bond
increases so does the value of E/D.27 Based on this result, the Fe-N-O angle would be
expected to decrease as the ligand becomes less chelating, allowing the NO ligand to bind
in a more bent fashion, the result of which would be an increase in E/D.27 This

contradicts the experimental results, the lineshape becomes more axial as the

64

carboxylates are replaced by solvent across the series. So while a mixture of discrete E/D
species is a possible reason for the change in lineshape, we would expect N202 and N201
to have mixture of rhombic species rather than axial, lending to the further support that
the E/D strain, the result of a broad distribution of electronic structures, is the cause for
line broadening observed.

The near axial nature of the zero field interaction in the {FeNO}7 model
complexes when the ligand backbone is clearly not axially symmetric is rather
interesting. The shift away from being purely axial, designated by E/D, is only 0.0115 of
a maximum of 1/3, so the model complexes only show 3.5% rhombicity. This near axial
interaction remains even with the relatively major changes to the iron coordination
moving through the ligand series. Lineshape changes aside, this suggests that the zero
field interaction of {FeNO}7 unit is relatively insensitive to changes in the Fe
coordination environment and, as shown by Borovik et al, is more sensitive to changes in
the NO bonding.27

The HYSCORE spectra collected on the {FeNO}7 model complexes show two
types of features, one from hydrogen and the other from strongly coupled nitrogen, both
of these results are expected based on the structure of the ligand complex. The
assignments of these features in the model complex will aid in the analysis of HYSCORE
results acquired on non-heme iron enzymes, in addition these assignments should provide
the details necessary to determine the orientation of NO bond and thus the gz axis with
respect to the Fe ligands.

A set of preliminary simulations for the nitrogen features in N203 at 3100 G were

65

obtained using similar parameters of two different orientations, one approximately
perpendicular to the gz axis and the other nearly co-linear. In order to shifi the phase of
the modulation function and produce features in the (-+) quadrant as seen in the
HYSCORE a large isotropic hyperfine coupling, Aiso > 2v“, is necessary. The isotropic
hyperfine component used in the simulations, Aiso = 7.5 MHz or Aiso = 8.5 MHz are
sufficiently large to produce features in the (-+) quadrant; however, this interaction alone
is not sufficient to produce the observed features. The two other components, the
anisotropy in the hyperfine interaction and the nuclear quadrupole interaction both show
an angular dependence that could be used to differentiate the nitrogen orientation with
respect to the gz axis. One approximation for the anisotropic component of the hyperfine
interaction is a dipole-dipole interaction, the F e-N distances given in the crystal structure
of N203 are 2.2 A for the nitrogen trans to the open coordination site and 2.4 A for the
cis nitrogen. Utilizing the point dipole approximation, these Fe-N distance would
translate to ~05 MHz; however, in the simulations this amount of anisotropy only
broadens the peaks and adds no significant angular dependence.

The nuclear quadrupole interaction is the interaction of the nitrogen nucleus with
electric field gradient produced by the lone pair of electrons on the nitrogen. The
principle axis of this interaction is defined by the lone pair on the nitrogen which, in the
model system, is coincident with the Fe-N bond. The magnitude of the nuclear
quadrupole interaction used in the simulations, eZQq = 4.5 MHz, is consistent with values
reported for tri-coordinate nitrogen and falls between the literature values for ammonia

and trimethylamine.28 Based on the similarity to tri-coordinate nitrogen, the F e-N bond

66

can be interpreted as not very covalent and closer to an ionic bond. Changing the
orientation of the nitrogen with respect to the gz axis shifted the features seen in the
HYSCORE, a nitrogen perpendicular to gz produces the feature at (-9,l2), while a co-
linear nitrogen produces a peak (-3,9) and a weak feature near (-9,12). The changes seen
in the nuclear quadrupole interaction provides a method to differentiate between the cis
and trans nitrogen.

Based on the crystal structure of N203 the assignment of the perpendicular
nitrogen is straightforward; however, the co-linear nitrogen could either be from the
chelating ligand or the NO. In the treatment used for the hydrogen hyperfine coupling
calculations the Fe was treated as a point dipole which means that it contains all of the
unpaired spin density of the {FeNO}7 unit; in this case the NO-nitrogen is just as likely
to contribute to the HYSCORE as the ligand nitrogen. If there is any delocalization of the
unpaired spin density onto the NO, as suggested by the various {FeNO}7 electronic
structure formalismsz, the isotropic component of the hyperfine interaction would be
expected to increase dramatically and the F e-N bond would become increasingly covalent
lowering the quadrupole energy. Based on the similarity of the hyperfine parameters and
the quadrupole interaction for the two different nitrogen’s it is reasonable to assume that
the co-linear nitrogen is ligand based and that the coupling to the NO-nitrogen is too large
to be detected in the experiment. Confirmation of these assumption could be made
through 15N labeling of the ligand and NO.

The assignment of the non-exchangeable hydrogen features seen in the

HYSCORE to the ligand hydrogen’s has shown to be difficult. As a rough approximation,

67

a dipole distance of 3.2 A to at most 3.5 A could account for the observed hydrogen
features. In the His-His—Glu active site of non-heme Fe oxygenases, this number of non-
exchangeable hydrogen’s is significantly reduced, down to 4 from the two histidine
ligands and assignment should be simplified. In non-heme enzymes a combination of the
orientation data for the histidine nitrogen and hydrogen should provide enough
constraints to determine the gz axis within the enzyme center and thus the NO bond axis.
The HYSCORE spectra of both N201 and N202, in the range of 2800 G to3200
G, had shown presence of exchangeable hydrogen. Simulations of these hydrogen
features provided a dipole-dipole distance of ~2.6 A and an angle relating the dipole-
dipole vector and the Fe-NO bond of ~70°, both of these parameters are consistent with
those expected for coordinated H20. The angle given more than likely represents a
distribution of orientations which culminates with a maximum that can be described by
the parameters used for the simulations. Support for a wide distribution of distances and
angles is given if the ratio of amplitudes are considered. The observed at 3100 G that the
ratio of the amplitude of non-exchangeable to exchangeable max was ~1.4:1, comparing
the amplitudes of simulation for a hydrogen 3.2 A from the iron to one of 2.6 A gives a
ratio of ~1 :6. This difference in ratio's could be accounted for in two different ways: The
coordinated water adopts a range of different of orientations and only a small percentage
of those contribute to the observable features; second, clearly more than one hydrogen is
contributing to the non-exchangeable features which would raise the amplitude in that
region. Based on some trial simulations the first reason, a range of water orientations, is

expected to have a greater influence on the HYSCORE amplitude. This result is not

68

entirely unexpected since there is nothing in the chelating ligand which restricts the water
orientation.
2.9 CONCLUSIONS

A set of three non-heme Fe models were characterized using EPR spectroscopy. In
the CW-EPR spectra a correlation can be drawn between the number of labile sites on the
iron and E/D strain, an increase in open coordination sites resulted in an increase in E/D
strain. The HYSCORE spectra of the three models were recorded and the non-
exchangeable features were assigned to chelating ligand hydrogen and nitrogen. The
exchangeable hydrogen observed in the HYSCORE spectra of N20] and N202 were

simulated and assigned to coordinated water.

69

2.10 REFERENCES

l. Enemark, J. H.; Feltham, R. D., Principles of Structure, Bonding, and Reactivity
for Metal Nitrosyl Complexes. Coordin Chem Rev 1974, I3 (4), 339-406.

2. Brown, C. A.; Pavlosky, M. A.; Westre, T. B; Zhang, Y.; Hedman, B.; Hodgson,
K. 0.; Solomon, E. I., Spectroscopic and Theoretical Description of the Electronic-
Structure of S=3/2 Iron-Nitrosyl Complexes and Their Relation to O-2 Activation by
Nonheme Tron Enzyme Active-Sites. J Am Chem Soc 1995, 11 7(2), 715-732.

3. Bill, E.; Bernhardt, F. H.; Trautwein, A. X.; Winkler, H., Mossbauer Investigation
of the Cofactor Iron of Putidamonooxin. European Journal of Biochemistry 1985, 14 7
(1), 177-182.

4. (a) Yang, T. C.; Wolfe, M. D.; Neibergall, M. B.; Mekmouche, Y.; Lipscomb, J.
D.; Hoffman, B. M., Substrate Binding to No-Ferro-Naphthalene 1,2-Dioxygenase
Studied by High-Resolution Q-Band Pulsed H-2-Endor Spectroscopy. J Am Chem Soc
2003, 125 (23), 7056-7066; (b) Tierney, D. L.; Rocklin, A. M.; Lipscomb, J. D.; Que, L.;
Hoffman, B. M., Endor Studies of the Ligation and Structure of the Non-Heme Iron Site
in Acc Oxidase. J Am Chem Soc 2005, 127 (19), 7005-7013.

5. Pohl, K.; Wieghardt, K.; Nuber, B.; Weiss, J ., Preparation and Magnetism of the
Binuclear Iron(Ii) Complexes (F e(C9h21n3)Ncs2)2 , (Fe(C9h21n3)Nc02)2 ,
(Fe(C9h21n3)N32)2 and Their Reaction with No - Crystal-Structures of (Fe(C9h21n3)
(Ncs)2)2 and Fe(C9h21n3)(No)(N3)2. J. Chem. Soc. -Dalton Trans. 1987, (1), 187-192.

6. Salerno, J. C.; Siedow, J. N., Nature of the Nitric-Oxide Complexes of
Lipoxygenase. Biochim Biophys Acta 1979, 579(1), 246-251.

7. Zhang, Y.; Pavlosky, M. A.; Brown, C. A.; Westre, T. E.; Hedman, B.; Hodgson,
K. 0.; Solomon, E. I., Spectroscopic and Theoretical Description of the Electronic-
Structure of the S = 3/2 Nitrosyl Complex of Nonheme Iron Enzymes. J Am Chem Soc
1992, 114 (23), 9189-9191.

8. Westre, T. E.; Dicicco, A.; F ilipponi, A.; Natoli, C. R.; Hedman, B.; Solomon, E.
I.; Hodgson, K. 0., Determination of the F e-N-O Angle in (Feno)(7) Complexes Using
Multiple-Scattering Exafs Analysis by anas. J Am Chem Soc 1994, 116 (15), 6757-
6768.

9. Rodriguez, J. H.; Xia, Y. M.; Debrunner, P. G., Mossbauer Spectroscopy of the
Spin Coupled F e2+-{F eno}(7) Centers of Nitrosyl Derivatives of Deoxy Hemerythrin
and Density Functional Theory of the {F eno}(7)(S=3/2) Motif. J Am Chem Soc 1999,
121 (34), 7846-7863.

10. Muthukumaran, R. B.; Grzyska, P. K.; Hausinger, R. P.; McCracken, J ., Probing

the Iron—Substrate Orientation for Taurine/Alpha-Ketoglutarate Dioxygenase Using
Deuterium Electron Spin Echo Envelope Modulation Spectroscopy. Biochemistry- Us

70

2007, 46 (20), 5951-5959.

11. Bertrand, B; More, C.; Guigliarelli, B.; Foumel, A.; Bennett, B.; Howes, B.,
Biological Polynuclear Clusters Coupled by Magnetic-Interactions - from the Point
Dipole Approximation to a Local Spin Model. J Am Chem Soc 1994, 116 (7), 3078-3086.

12. Elsasser, C.; Brecht, M.; Bittl, R., Pulsed Electron-Electron Double Resonance on
Multinuclear Metal Clusters: Assignment of Spin Projection Factors Based on the Dipolar

Interaction. J Am Chem Soc 2002, 124 (42), 12606-12611.

13. Arciero, D. M.; Orville, A. M.; Lipscomb, J. D., O-17 Water and Nitric-Oxide
Binding by Protocatechuate 4,5-Dioxygenase and Catechol 2,3-Dioxygenase - Evidence

for Binding of Exogenous Ligands to the Active-Site F e-2+ of Extradiol Dioxygenases. J
Biol Chem 1985, 260 (26), 403 5-4044.

l4. Han, A. Y.; Lee, A. Q.; Abu-Omar, M. M., Epr and Uv-Vis Studies of the Nitric
Oxide Adducts of Bacterial Phenylalanine Hydroxylase: Effects of Cofactor and
Substrate on the Iron Environment. Inorg Chem 2006, 45 (10), 4277-4283.

15. Hofer, P.; Grupp, A.; Nebenfuhr, H.; Mehring, M., Hyperfine Sublevel Correlation
(Hyscore) Spectroscopy - a 2d Electron-Spin-Resonance Investigation of the Squaric
Acid Radical. Chem. Phys. Lett. 1986, I32 (3), 279-282.

16. Hutchison, C. A.; Mckay, D. B., Determination of Hydrogen Coordinates in
Lanthanum Nicotinate Dihydrate Crystals by Nd+3 -Proton Double-Resonance. J Chem
Phys 1977, 66(8), 3311-3330.

17. Hurst, G. C.; Henderson, T. A.; Kreilick, R. W., Angle-Selected Endor
Spectroscopy .1. Theoretical Interpretation of Endor Shifts from Randomly Orientated
Transition-Metal Complexes. J Am Chem Soc 1985, 107 (25), 7294-7299.

18. Cappillino, P. J. M., J. R.; Tyler, L. A.; Lo, W.; Kasibhatla, B. S. T.; Krzyaniak, M.
D.; McCracken, J .; Armstrong, W. H.; Caradonna, J. P., Manuscript in submission 2010.

19. Mims, W. B., Envelope Modulation in Spin-Echo Experiments. Phys Rev B 1972,
5 (7), 2409-&.

20. Gemperle, C.; Aebli, G.; Schweiger, A.; Ernst, R. R., Phase Cycling in Pulse Epr.
J Magn Reson 1990, 88 (2), 241-256.

21. Dikanov, S. A.; Xun, L. Y.; Karpiel, A. B.; Tyryshkin, A. M.; Bowman, M. K.,
Orientationally-Selected Two-Dimensional Eseem Spectroscopy of the Rieske-Type Iron-

Sulfur Cluster in 2,4,5-Trichlorophenoxyacetate Monooxygenase from Burkholderia
Cepacia AC] 100. J Am Chem Soc 1996, 118 (35), 8408-8416.

22. Schweiger, A.; Jeschke, G., Principles of Pulse Electron Paramagnetic
Resonance. Oxford University Press: Oxford, UK ; New York, 2001; p xxvi, 578 p.

71

23. Mims, W. B.; Davis, J. L.; Peisach, J ., The Accessibility of Type-I Cu(Ii) Centers
in Laccase, Azurin, and Stellacyanin to Exchangeable Hydrogen and Ambient Water.
Biophys J 1984, 45 (4), 755-766.

24. Weisser, J. T.; Nilges, M. J.; Sever, M. J .; \Vrlker, J. J ., Epr Investigation and
Spectral Simulations of Iron-Catecholate Complexes and Iron-Peptide Models of Marine
Adhesive Cross-Links. Inorg Chem 2006, 45 (19), 7736-7747.

25. Doctor, K. S.; Gaffney, B. J.; Alvarez, G.; Silverstone, H. J ., Epr Spectroscopy of
Interdoublet Transitions in High-Spin Iron - Applications to Transferrin Oxalate. J Phys
Chem-Us 1993, 97 (12), 3028-3033.

26. Yang, A. S.; Gaffney, B. J ., Determination of Relative Spin Concentration in
Some High-Spin Ferric Proteins Using E/D-Distribution in Electron-Paramagnetic
Resonance Simulations. Biophys J 1987, 51 (1), 55-67.

It I 'I. IMJIS m...

27. Ray, M.; Golombek, A. P.; Hendrich, M. P.; Yap, G. P. A.; Liable-Sands, L. M.;
Rheingold, A. L.; Borovik, A. 8., Structure and Magnetic Properties of Trigonal ;
Bipyramidal Iron Nitrosyl Complexes. Inorg Chem 1999, 38 (13), 3110-3115. i;

 

28. Lucken, E. A. G, Nuclear Quadrupole Coupling Constants. Academic P.: London,
New York,, 1969; p x, 360 p.

72

CHAPTER III
STRUCTURAL CHARACTERIZATION OF NO ADDUCTS OF THE NON-HEME
F E(II) ACTIVE SITE OF TYROSINE HYDROXYLASE
3.1 INTRODUCTION
Tyrosine hydroxylase (TyrH) is a non-heme iron enzyme that requires

tetrahydrobiopterin (BH4) and molecular oxygen to catalyze the hydroxylation of

@005 002' aromatic ring of L-tyrosine (tyr) to
NH I I . .
HO 3 mm“ produce L-drhydroxyphenylalanrne

(DOPA) (Figure 3.1).1 The production of

 

TyrH 4' 02 I-5
A DOPA as described above is the rate
0 . . . . . .
R n R n 8 lrmrtrng step in the brosynthesrs of the
I 4| ‘ N” I It"
7 N Ba NANH N \NJKNH catecholamine neurotransmitters.2 TyrH is
2 2
H H

a member of the small family of
Figure 3-13 TyrH “33°90“ tetrahydrobiopterin-dependent aromatic
amino acid hydroxylases that also includes phenylalanine hydroxylase (PheH) and
tryptophan hydroxylase (TrpH). This family of enzymes are structurally similar in that
they all share similar catalytic sites which contain a non-heme iron coordinated facially
by two histidines and a glutamate.3 4 5 Based on their structural similarities and chemistry,
these enzymes have been proposed to follow the same catalytic mechanism (Figure 3.2). 2
This mechanism involves two sequential steps, first, the activation of oxygen using
electrons from a tetrahydropterin to yield a high valent, Fe(IV)-oxo hydroxylating

intermediate and hydroxylated pterin 6 7 and second the hydroxylation of the side chain of

73

Figure 3.2: TyrH Catalytic Mechanism

amino acid substrate via electrophilic aromatic substitutions

A steady-state kinetic mechanism has been described for rat TyrH that is
consistent with an ordered sequential mechanism for substrate binding; the
tetrahydropterin binds first followed by molecular oxygen and then finally, substrate
tyrosine.9 No reaction is observed until all three substrates are bound. Fluorescence
anisotropy studies have shown that pterin binding to TyrH initiates the movement of an
active site loop that packs against the amino acid binding site, providing a structural
explanation for the ordered kinetic mechanism. '0 Further, recent single turnover studies
have shown that binding of both a tetrahydropterin and tyrosine increase the reactivity of
the iron site with oxygen by several orders of magnitude.ll A combination of
spectroscopic approaches have identified changes in the iron ligands when both substrates
are bound, wherein the iron converts from hexacoordinate to pentacoordinate and opening
up a potential site for oxygen.11

While several structures have been described of the aromatic amino acid

hydroxlyases, there is none with only the primary amino acid substrate bound. Most of

74

the available structures are of the catalytic domains containing ferric iron in the presence
or absence of an oxidized pterin. For example, for TyrH, the only crystal structures
reported are for ferric forms of the enzyme in the resting state (lTOH)3 or with the
oxidized cofactor, 7,8 dihydrobiopterin, bound (2TOH)12. Exceptions are provided by the
structures of PheH. Specifically, X-ray crystallographic studies of ferrous PheH with BH4
alone13 and with BH4 and isoleucine or the phenylalanine analog thienylalanine 4 '4 have
been reported. Comparison of these crystal structures reveals that the binding of the
amino acid to PheH treated with BH4 reduces the distance between the iron and the C 4a
position of BH4 from 6 A to 4.5 A. Thus, the binding of the amino acid substrate analog
initiates a conformational change in PheH that pushes the reduced pterin cofactor closer
to the iron.4 However, it is not known whether similar changes occur in TyrH or even
whether the order of substrates binding to the two enzymes is the same. Additionally,
there is also nothing known about the structure of the reactive complex of oxygen, amino
acid and pterin in any of these enzymes.

One of the most successfiil approaches for the EPR characterization of F e(II)
active sites in non-heme enzymes calls for the metal ion to be studied in complex with
nitric oxide. 15 '6 This technique is ideal for F e(II) hydroxylases as NO acts as an oxygen
surrogate, binding to the iron and preventing catalysis by inhibiting oxygen binding. In
addition, NO antiferromagnetically couples with the iron generating an S = 3/2 spin
system that is readily detected by EPR spectroscopy. '5 '6 17 More recently, Hoffman and
co-workers have used the {FeNO}7, S = 3/2 spin system together with deuterium

Electron Nuclear Double Resonance (ENDOR) spectroscopy to study the non-heme

75

 

ferrous sites in naphthalene 1,2-dioxygenasel8 and ACC oxidase”. The 2H hyperfine
couplings measured in these ENDOR studies provided deuterium dipole-dipole distances
for specifically deuterated substrates that compared well with X-ray crystallographic
studies. In addition, a recent 2H-electron spin echo envelope modulation (ESEEM)
spectroscopy study of taurine dioxygenase provided deuterium dipolar couplings for
specifically deuterated substrate taurine that agreed well with predictions from crystal
structures.20

In this chapter, the results of 2H—ESEEM studies carried out on ferrous TyrH
treated with NO, and specifically deuterated cofactor, 6-methyl tetrahydropterin (6-
MPH 4) and substrate L-tyrosine are reported. The EPR studies provide new data on how
TyrH controls catalytic activity as the {FeNO )7 metal center mimics the Fe-02
interaction, structural information can be gained from tertiary, TyrH-{FeNO}7-6MPH4,
and quaternary, TyrH-{FeNO}7-6MPH4-tyr complexes. In addition to wild type enzyme
the TyrH mutant E332A, in which oxygen activation isonly poorly coupled to amino acid
hydroxylation, was examined.
3.2 EXPERIMENTAL

6-Methyltetrahydropterin (6MPH4) and 6-methylpterin were purchased from
Schircks Laboratories (Jona, Switzerland). 6-(2-Hydroxy-1-methyl-2-nitrosohydrazino)-
N-methyl-hexanamine (MAHMA NONOate), ethylenediamine tetraacetic acid (EDTA),
3-(N-morpholino)propanesulfonic acid (Mops), L-tyrosine and glycerol were purchased
from Sigma-Aldrich (St. Louis, MO). L-3,5-2H-Tyrosine and deuterium gas were from

Cambridge Isotopes (Andover, MA). Potassium chloride and ferrous ammonium sulfate

76

were from Fisher (Pittsburg, PA). 6-CH3-6,7-2H-tetrahydropterin (Figure 3.1) (d-6MPH4)
was synthesized by reduction of the commercially available 6-methylpterin to the level of
tetrahydropterin using deuterium gas, as previously described.” All other chemicals were

of the highest purity commercially available.

Wild-type TyrH was expressed in E. coli and purified as previously described.22 23
In order to remove the ferric iron from the protein, the ammonium sulfate pellet at the end
of the purification was resuspended in 5 mM EDTA, 200 mM Hepes, (pH 7.5), 10%
glycerol and 0.1 M KCl, and incubated on ice for one hour. The enzyme solution was then
dialyzed against the same buffer without EDTA and concentrated using Amicon Ultra-15
and Ultra-4 centrifugal filters (Millipore Corp., MA). The enzyme samples for ESEEM
were brought to a final glycerol concentration of 30% (v/v) during the concentration
stage. The iron content of the apo-enzyme was measured using a Perkin-Elmer Aanalyst
600 atomic absorption instrument.24 The typical iron content of an apo enzyme
preparation was $0.1 equivalent.

NO Samples for EPR were prepared using MAHMA NONOate as the nitric oxide
donor. Stock solutions of MAHMA NONOate were prepared in 0.01 M NaOH just
before the experiment and were always kept on ice. Exact concentrations of the MAHMA
NONOate solutions were determined from the UV absorbance at 250 nm in 0.01 M KOH,
using an extinction coefficient of 7.3 mM'lcm‘l. 25 Highly concentrated stock solutions
of tyrosine and 3,5-2H-tyrosine (~50 mM) were prepared by bringing the solution to a
final pH of ~10. The exact concentrations of the tyrosine stock solutions were determined

using an extinction coefficient of 1.34 mM'lcm’1 at 275 nm in 0.1 M HCl. Stock

77

 

solutions of the protiated and deuterated 6MPH4 were prepared in 2 mM HCl, and an
extinction coefficient of 17.8 mM'lcm'1 in 2 M perchloric acid was used to determine the
concentrations. Ferrous ammonium sulfate solutions were prepared fresh by dissolving
the appropriate amount of powder in 2 mM HCl. ESEEM samples were prepared inside
an anaerobic cuvette at 25 °C. Apo-TyrH (0.9-1.2 mM) and tyrosine (if the complex
contained tyrosine) were placed at the bottom of the cuvette. Ferrous ammonium sulfate
solution (0.9 equivalent in ~10 ul) was placed on the lower neck of the cuvette. 6MPH4
and MAHMA NONOate solutions were either placed in the side arms or on the upper
neck of the cuvette, for large and small volumes, respectively. Buffer conditions were 100
mM Mops (pH 7.0), 0.3 M KCl and 30% glycerol. Each MAHMA NONOate molecule
releases two molecules of NO,25 and the half-life of MAHMA NONOate is ~35 8 under
these buffer and temperature conditions (data not shown). The contents of the cuvette (a
total volume of ~250 1.11) were made anaerobic by the application of argon-vacuum
exchange for at least 20 minutes. The anaerobic enzyme solution was then mixed with
ferrous ammonium sulfate and incubated for 10 minutes. This was followed by mixing
with 6MPH4 (if the complex contained 6MPH4). After a few minutes of incubation,
MAHMA NONOate (~0.6 equivalent of the enzyme) at the upper neck of the cuvette was
introduced to the enzyme-substrate mixture. After ~3 minutes, ~200 u] of the reaction
mixture was quickly transferred to the quartz EPR tubes (4 mm OD, 707-SQ-250M,
Wilmad, Buena, NJ) using a glass pipette and immediately frozen in liquid nitrogen.
(UV-Visible experiments performed under similar conditions indicated that the maximum

absorbance at 450 nm, which is indicative of NO binding, was reached at about 3 minutes

78

and stayed constant for at least a few minutes). UV-Visible spectra collected at increasing
concentrations of MAHMA NONOate showed that NO was saturating under the
concentrations used.

Pulsed EPR measurements were made on a Bruker E-680X spectrometer
operating at X band and equipped with a model ER 4118X-MD-X5-W1 probe that
employs a 5mm dielectric resonator. The temperature was maintained at 4.0 K using an
Oxford Instruments liquid helium flow system equipped with a CF-935 cryostat and an
ITC-503 temperature controller. ESEEM data were collected using a three-pulse
(stimulated echo) sequence, 90°-r-90°-T—90°, with 90° microwave pulse widths of 16 ns
(F WHM). The tau value was chosen to suppress the hydrogen matrix contribution; due to
short phase memory times, tau values were restricted to less than 200ns. An integration
window of 24 ns was used to acquire spin echo amplitude, and data set lengths were 512
points. The deuterium contributions to ESEEM spectra were elucidated using the ratio
method introduced by Mims26 together with processing tools provided by Bruker. Three-
pulse ESEEM data were normalized by dividing each data set by its maximum amplitude.
2H-ESEEM data were then divided by the corresponding, normalized 1H ESEEM data,
resulting in ESEEM data dominated by 2H. The time domain data that resulted from these
ratios were tapered with a Hamming window function and Fourier transformed. ESEEM
spectra were obtained by taking the absolute value of the transforms.

Simulations of the 2H hyperfine couplings were accomplished using scripts
written in MATLAB. The calculations were performed in the time domain using the

density matrix formalism of Mims.27 The orientation-dependent Hamiltonian used was

79

 

modeled after work by Hutchison and McKay.28 Processing of the simulations, also
performed in Matlab, was identical to that described above for the experimental spectra.
3.3 RESULTS

In the previous chapter, to study the effects of substrate and cosubstrate binding on
the structure of the TyrH Fe site, x-band CW-EPR spectra were examined of four
derivatives of {F eNO}7 TyrH: TyrH treated with 6-methyltetrahydropterin,
TyrH[6MPH4], TyrH treated with L-tyrosine, TyrH[tyr], TyrH treated with L-tyrosine and
6-methyltetrahydropterin, TyrH[tyr,6MPH4] and the TyrH mutant E332A treated with L-
tyrosine and 6-methyltetrahydropterin, E3 32A[tyr,6MPH 4]. Changes in the spectra were
observed; however, no specific structural information, in regards to the substrates, could
be obtained from the CW—EPR spectra.

To measure the structural relationship between substrate L-tyrosine, cosubstrate
tetrahydrobiopterin and Fe(II) at the catalytic site ESEEM experiments were undertaken.
ESEEM data from {FeNO}7 TyrH include hyperfine interactions from all nearby
magnetic nuclei, including the nitrogens from the ligand histidines and hydrogens from
nearby solvent and ligands. To isolate interactions due to L-tyr or 6MPH , these co-
substrates were selectively deuterated, and the ESEEM data from the deuterated samples
were divided by that for the corresponding protonated samples. For this procedure the
time domain ESEEM data were normalized to the amplitude at the earliest time point and
then divided, leaving predominately the deuterium ESEEM contribution.

To define the geometry of these deuterium hyperfine interactions, ESEEM

experiments were performed at multiple field positions across the EPR spectrum. When

80

 

—I

 

Amplitude (a.u.)

0

2
1.6
1.2

0.8

Amplitude (a.u.)

0.4

{FeNO}7 TyrH: treated with d-6MPH4 (A)

og.

0
0

1

 

 

 

2345

Freqmncy (MHz)

 

 

 

ls...

 

 

J
1

2345

Frequency (MHz)
Figure 3.3: Representative 2H ESEEM of

Amplitude (a.u.)

Amplitude (a.u.)

o-.-
carom»

p
A

 

working at discrete field positions, this
effective g-anisotropy provides a single
.crystal-like selection of orientations that

‘satisfy the resonance condition.{Hutchison,

45s....

 

 

0 r
o 1 2 a 4 51977#25}181929 Forthemostpart,the2H-

2

1.6:

0.

0.4 r

.2'

8 .

Frequency (MHz)
ESEEM peaks observed in the samples were

 

. weak and consisted of only single peaks
' centered at the 2H-Larmor frequency.

Consequently, only spectra taken near g = 2,

 

 

3200 G, and near g = 4, 1980 G, will be

 

0...;
012345

me’ (MHz) presented.

Figures 3.3a and 3.3b show the 2H-

at 198OG and (B) at 3200G; and treated with

d-6MPH4 and tyr (C) at 19806 and (D) at

ESEEM spectra of {Fe-NO}7 TyrH

3200G. Experiment shown in black with the

corresponding simulation shown in red. The

prepared with deuterated 6-CH3-6,7-2H-

Hamiltonian parameters used in the

simulation of TyrH[d-6MPH4] were as

tetrahydropterin, TyrH[d-6MPH4], at 1980

follows: principle g values, 4.0, 4.0 and 2.0;

principle deuterium hyperfine values, -0.06,

G and 3200 G, respectively. The deuterium

-0.06, and 0.12 MHz; Euler angles for the

hyperfine tensor 0, 1, and 0 radians;

Larmor frequencies at these two magnetic

equ,

0.20 MHz; 1], 0; Euler angles relating nqi to

hfi, 0, 1, and 0 radians. The Hamiltonian

field strengths are 1.3 and 2.1 MHz. The

parameters used in the simulation of TyrH[d-

6MPH4,tyr] were as follows: principle g

quotient spectrum at 1980 G (Figure 3.3a)

values, 4.0, 4.0 and 2.0; principle deuterium

hyperfine values, -0.16, -0.16, and 0.32

is characterized by a signal-to-noise ratio of

Mhz; Euler angles for the hyperfine tensor 0,

1.15, and 0 radians; equ, 0.20 MHz; 11, 0;

about 2, but shows peaks at 1.0 and 1.8

Euler angles relating nqi to hfi, O, 0.75, and
0 radians.

MHz. These peaks are observed in the

81

corresponding ESEEM spectrum of the protiated sample prior to division and are
assigned to 14N contributions that have not been eliminated by the division procedure.
No ESEEM intensity is resolved at 1.3 MHz. At 3200 G (Figure 3.3b) the overall
modulation intensity is lower and the division process does a more complete job of
eliminating ESEEM components that are common to both the TyrH[d-6MPH4] and
TyrH[6MPH4] samples. Little or no ESEEM was detected at 2.1 MHz. The 2H ESEEM
spectra of {Fe-NO}7 TyrH treated with both d—6MPH4 and L-tyrosine, TyrH[d—6MPH4
,tyr], are shown in Figures 3.3c and 3.3d. Unlike the TyrH[d-6MPH4]/TyrH[6MPH4]
quotient spectra, the TyrH[d-6MPH4,tyr]/TyrH[6MPH4,tyr] quotients show resolved
peaks centered at the deuterium Larmor frequency of 1.3 (1980 G) and 2.1 MHz (3200
G), respectively.

The red traces of Figure 3.3 show the results of 2H-ESEEM simulations for our
studies of d-6MPH4 coupling to the {F eNO}7 paramagnetic center. When L-tyr is bound
to the enzyme, figures 3.3c and 3.3d, our simulations were done assuming a single 2H
was coupled to the {F eNO}7 center with a dipole-dipole distance of 4.2 :1: 0.2 A and an
angle between the principal axis of the hyperfine interaction and the gZ axis of 66 :1: 5°.
We also used simulations to gauge the change in dipolar distance that must occur to
reduce the 2H-ESEEM to the levels observed for Figures 3.3a and 3.3b when d-6MPH4
was present without L-tyr. The red traces in these figures were obtained by increasing the
dipolar distance to 5.9 A. The other parameters used in our calculations are given in the

figure caption.

Figures 3.4a and 3.4b show the 2H-ESEEM of {Fe-NO}7 TyrH treated with 3,5-

82

 

 

2 A - . . - 2 BI ‘ - T 2H-L-tyrosine, TyrH[d-tyr]. ESEEM peaks
3 1: :5: :: attributed to 2H are clearly resolved at 1.3
g (,8. . g 0,. . and 2.1 MHZ at 1980 and 3200 G,
E 0-1 ’ ‘ E 0'4 ‘respectively. Simulations of these spectra

 

 

 

 

 

 

o - - - - o . . - . .
0 1 2 3 4 5 0 1 2 3 4 5(red traces) assummg asrngle, coupled

 

 

 

 

Freqwncy (MHz) Frequency (MHz)
2 2 fi deuteron yielded a dipolar distance of 4.3 i
c o
A 1.6 . A 1.6 .02 A and an orientation of the hyperfine
=2 =2
31.2» .312. . . . . . o
o g prrncrpal axrs vvrth respect to gz of 72 :b 5 .
.3 ..
a 0.8 . g, 0.8 .
E 3 When {F e-NO}7 TyrH is treated with 6-
0.4 . 0.4 -
MPH4 and 3,5-2H tyrosine, TyrH[6MPH4,d-

 

 

 

 

 

 

0012345 0012345
WWW (MHZ) Fwy (MHZ) tyr], the 2H- ESEEM spectra, Figures 3.4c

Figure 3.4: Representative 2H ESEEM of

{FCNO}7 TyrH: treated with d-tyr (A) at and 3.4d, show a modest change in

1980G and (B) at 3200G; and treated with

6MPH4 and d-tyr (C) at 19300 and (D) at amplitude as compared to the

3200G. Experiment shown in black with the

corresponding simulation shown in red. The corresponding TyrH[d-tyr] spectra shown

hamiltonian parameters used in the

simulation of TyrH[d-tyr] were as follows: in Figures 3-43 and 3413- A more

principle g values, 4.0, 4.0 and 2.0; principle

deuterium hyperfine values, -0.15, -0. 15, and Significant change is observed in the line

0.30 MHz; Euler angles for the hyperfine

tensor 0, 1.25, and 0 radians; e2qQ, 0.20 shape 0f TyrH[6MPH4,d-tyr], which

MHz; 1], 0; Euler angles relating nqi to hfi,

0, 0,7, and o radiansjhe hamiltonian narrows and shows a single peak compared
parameters used in the simulation of
TyrH[6MPH4,d-tyr] were as follows: to the spectra 0f TyrH[d-tyr] at bOth 1930

principle g values, 4.0, 4.0 and 2.0; principle . .
deuterium hyperfine values, -0.17, -0.17, and and 3200 G- The red traces 1n Frgures 3.4c

0.34 Mhz; Euler angles for the hyperfine .
tensor 0, 0.45, and 0 radians; eZqQ, 0.20 and 349 show the corresponding

MHz; 1], 0; Euler angles relating nqi to hfi,
0, 1,35, and 0 radians. simulations for TyrH[6MPH4,d-tyr] that

83

 

 

 

N
N
l

 

 

 

 

 

 

 

 

 

   

 

 

 

 

3' . - - A . - Y result from changing the orientation between
1 6 r 1 6 .
3; 1} g 1.2, ‘ the PAS of the hyperfine tensor and g2 to
g ogl . é ogl . 25:1: 5° and decreasing the 2H-dipolar
E E
94’ 04 ‘ distance to 4.1 :t 0.2 A.
o - - - t o - r .
0 1 2 3 4 5 0 1 2 3 4 5 Figures 3.5a and 3.5b show the 2H-
Frequency (MHz) Frequency (MHz)
2 j 2 ESEEM of the {Fe-NO}7 TyrH mutant
c D
A 1 6l . A 1.6 . E3 32A treated with deuterated 6-MPH4 and
e 1.2» . 5"- 1,2. . .
o 1» tyrosrne, E332A[d-6MPH4 tyr]. ESEEM
'3 o 8 . é o 8 ,
O. - a, . .
E 5 peaks attributed to 2H are clearly resolved at
0.4 « 0.4 .
o .__.___ A 0 - - 1.3 and2.l MHZat1980 and 3200 G,
o 1 2 3 4 5 o 1 2 3 4 5
mey (MHZ) mey (MHZ) respectively. Simulations of these spectra

Figure 3.5: Representative 2H ESEEM of

the {FeNO}7 TyrH mutant E332A: treated (red traces) assuming a single, coupled
with d-6MPH4 and tyr (A) at 1980G and (B) . . .

at 3200c}; and treated with 6MPH4 and d-tyr deuteron yielded a dipolar distance of 4.2 i
(C) at 1980G and (D) at 3200G. Experiment
shown in black with the corresponding
simulation shown in red. The Hamiltonian
parameters used in the simulation of
E332A[d-6MPH4,tyr] were as follows: 7
principle g values, 4.0, 4.0 and 2.0; principle ESEEM of the E332A mutant of {Fe-NO}
deuterium hyperfine values, -0.16, -0.16, and

0.32 MHz; Euler angles for the hyperfine TyrH treated With 6"MPH4 and 395'2H'L'
tensor 0, 0.60, and 0 radians; eZqQ, 0.20

MHz; 1], 0; Euler angles relating nqi to hfi, tyrosine, E332A[6MPH4,d-tyr], is shown in
0, 0.35, and 0 radians.The Hamiltonian
parameters used in the simulation of
E332A[6MPH4,d-tyr] were as follows: ' .
principle g values, 4.0, 4.0 and 2.0; principle 2H peaks. The red traces in Frgures 3.5c
deuterium hyperfine values, -0.13, -0.13, and

0.26 Mhz; Euler angles for the hyperfine and 3'59 show the corresponding

tensor 0, 0.45, and 0 radians; eZqQ, 0.20

MHz; 11, 0; Euler angles relating nqi to hfi, simulations for TyrH[6MPH4,d-tyr] that
0, 0.65, and 0 radians.

0.2 A and an orientation of the hyperfine

PAS with respect to gz of 34 i 5°. The 2H

Figures 3.5c and 3.5d with clearly resolved

84

result from a single deuteron with an orientation between the PAS of the hyperfine tensor
and g2 of 25 i 5° and a 2H-dipolar distance of 4.5 a: 0.2 A.
3.4 2H ESEEM SIMULATION NOTES

2H-ESEEM simulations were undertaken to reveal the structural changes that
occur upon substrate addition that underlie the experimental observations in Figures 3.3-
3.5. The deuterium hyperfine couplings were modeled using the following spin
Hamiltonian:

HzSetA-i—ganHO-1+1-Q-1 [3.1]
which consists of the electron-nuclear hyperfine, nuclear Zeeman, and nuclear quadrupole
interactions (nqi). Using the methodology of Hutchison and McKay28 the spin
Hamiltonian was set up in the principal axes system of the effective g-matrix with the
effective g-values being used to account for the orientation selection as described
previously, in chapter 2. Following previous 2H-ESEEM work on taurine dioxygenase”,
the calculations assumed that the 2H-ligand hyperfine interaction could be modeled by a
through-space, dipole-dipole interaction and that the nqi could be modeled with values of
the nqi coupling constant, eZqQ/h = 0.2 MHz, and asymmetry parameter, r1 = 0.30. The
relative axial symmetry of the zerofield interaction, together with the axial symmetry
assumed for the 2H-hyperfine interaction and 2H-nqi, resulted in the simulations being
sensitive to just four parameters, the effective g-value of the measurement, the 2H dipolar
distance, the Euler angle describing the relative orientation of the principal axis of the

hyperfine interaction with the gz = 2.00 axis, and the Euler angle describing the relative

orientation of the 2H-hyperfine and nqi principal axes.

85

 

In this scheme, the electron nuclear hyperfine couplings are weighted according to
their orientation with respect to the g-tensor and the electron spin operator is weighted by
the effective g-value of the measurement as can be seen in equations 2.10 and 2.4,
respectively. In addition, when working at a fixed magnetic field position only a restricted
set of orientations will satisfy the resonance condition of equation 2.5. The resulting
single crystal-like spectra makes it necessary to examine multiple field positions to
establish the orientation of the hyperfine tensor with respect to gz axis.

In general, the 2H ESEEM examined in this work consisted of a single featureless
peak at the 2H Larmour frequency, a result consistent with a weak hyperfine interaction.
The relatively featureless lineshape of the ESEEM makes it necessary to consider the

amplitude of the 2H peaks which are proportional to the magnitude of the hyperfine

 

 

 

 

 

 

 

 

 

2 . fl . . . 2 fl . iii.
1.5 r 1.5+
3" 3’
e a
d) 0
g 1 ‘ g 1 ’
5 a_-'
r e ,
< < /
0.5 : 05 _ ,
\\ //l
o % g . o i m g 4
0 0.5 1 1.5 0 0.5 1 1.5 2.5 3
Frequency (MHz) Frequency (MHz)

Figure 3.6: 2H simulations at 1780 G(blue), Figure 3.7: 2H simulations at 1780 G(blue),

1880 G(green), 1980 G(red), 2080 G(teal),

1880 G(green), 1980 G(red), 2080 G(teal),

2180 G(purple) using a dipole distance of 2180 G(purple) using a dipole distance of
4.5A and an angle between the dipole vector 4.5A and an angle between the dipole vector

and g2 of 0.3 rad.

86

and g2 of 1.3 rad.

interaction. With an increase in the dipolar coupling the amplitude also increases, this is
related to a greater mixing of the nuclear spin sub-levels which increase the transition
probabilities. Due to the weighting of the hyperfine interaction by the g-matrix the
amplitudes are also a function of the orientation between the PAS of the hyperfine tensor
and gL. Based on the single crystal-like character of the spectra they also show a field
dependent change in amplitude, this effect can be used to “fingerprint” the experimental
data because the field dependent amplitude changes are dependent on the angle between
the PAS of the hyperfine tensor and gz,. These amplitude changes as a function of
orientation and field are only observed in the low field end of the spectrum geff< ~3. An
example of this field dependence is shown in figures 3.6 and 3.7 where two sets of
simulations, which differ by orientation(simulation parameters in the caption), are shown
at multiple field positions between g = 4 and g = 3. Figure 3.6 shows an angle 0.3 rad
between the principle axis of the hyperfine tensor and gZ axis, it can be seen that the
amplitude has a minimum at the low field end(1780G — g = 3.881) shows a maximum at
1980 G(g = 3.489) and then begins to dip again. Figure 3.7, with an angle of 1.3 rad
between the principle axis of the hyperfine tensor and gZ axis, shows different field
dependent behavior, in addition to an overall greater amplitude, the relative amplitude
difference between field positions is smaller than seen in figure 3.6.

The quadrupole interaction which represents the interaction of the 2H nucleus
with the electric field gradient produced by the electron in its o-bond to the C and is
characterized by a principle axis which lies along the 2H-C bond. This interaction is axial

and therefore the orientation between the principle axis of the quadrupole and the

87

Aug-'1};

[gut 1. v..-

 

 

 

principle axis of the hyperfine is described using a single Euler angle. In practice it was
found that this angle largely influences the lineshape and was generally adjusted until
satisfying spectra at multiple field positions were obtained. The relative certainty and
confidence in these values were low, an unavoidable problem which can be attributed to
the frequency resolution. A more accurate determination of the nuclear quadrupole
interaction might be realized by working at higher fields, where the nuclear Zeeman
interaction is of sufficient magnitude to separate the spin manifolds so that the individual
components could be measured.
3.5 DISCUSSION

While the present understanding of the mechanisms of the aromatic amino acid
hydroxylases is based mainly on studies of TyrH, direct evidence for structural changes
accompanying substrate binding to an aromatic amino acid hydroxylase comes from a
limited number of structures of truncated PheH. These show significant differences
between the ferrous PheH-tetrahydrobiopterin complex and the ferrous PheH-
tetrahydrobiopterin-amino acid complex, including movement of loops over the active
site and a change in the metal ligands.4 There is evidence that the effect of substrate
binding differs between TyrH and PheH. The structures of PheH show that the active site
loop containing Tyr138 closing down on the active site only in the ternary complex. In
contrast, solution studies of TyrH show that the comparable loop in that enzyme responds
to pterin binding and is unaffected by binding of the amino acid substrate.lo Moreover,
even the structures of the ferrous PheH ternary complexes do not establish the orientation

of the substrates in the reactive complex containing amino acid, pterin, and oxygen. The

88

 

data reported in the present work establish the relative orientations of the pterin and the
amino acid in the active site of TyrH and the changes accompanying binding of each
substrate. In addition, the use of NO as an O2 mimic allows us to model the active
quaternary complex.

The results of the ESEEM studies on TyrH are summarized in Tables 1, these
results are given in terms of their dipolar distance, an imaginary vector which connects
the Fe to the deuteron; an angle, GHFI’ which represents the angle between the dipolar
vector and the axis defined by the Fe-NO bond; and an angle, eNQI’ which is the angle
that relates the principle axis system of the nuclear quadrupole interaction to that of the
hyperfine interaction. In this work the nqi was held at a constant value of 0.2 MHz with

axial symmetry, where the

Table 3.1: Deuterium Hyperfine Coupling Parameters

 

 

 

 

 

 

 

Distance(A)a 911121 91,1le principle axis of the nqi is along

a 0.2 A i 5, i 5, the C-2H bond. This
TyrH[d-6MPH4] > 5.9 570 570 approximation for the nqi
TyrH[tyr,d-6MPH4] 4.2 650 420 allowed successful simulation of
TyrH[d-tyr] 4. 3 710 400 the data and falls in line with
TyrH[d-tyr,6MPH4] 4.1 25° 770 other studies in similar
E332A[tyr,d-6MPH4] 4.2 34° 20° systems.18 ‘9 20 In the ESEEM
E332A[d-tyr,6MPH4] 4,5 250 370 simulations, the principal g

 

 

 

 

 

 

values were held constant at
a Distance measured between the Fe and the 2H using

the point dipole approximation [4,4,2] despite the experimental
b Quadrupole parameters held constant, equ = 0.2, . . .
11:0 deterrnrnatron of the effectrve g

89

 

values (chapter 1). Simulations were run using both the experimentally determined g
values and the axial set [4,4,2]. This approximation has a negligible effect near g = 2 and
gives rise to less than a 10% increase in the amplitude (for E/D = 0.05) near the g = 4
extreme. Although the CW-EPR spectra of TyrH[tyr], TyrH[6MPH4,tyr] and
E332A[6MPH4,tyr] are mixtures of two species, the E/D values for all species are < 0.05
with the majority species having E/D < 0.02. Given the uncertainty in the ESEEM
amplitude that arises at g = 4 from the data processing, the assumption of an axial g-
matrix is justified.

Multi-component CW-EPR spectra haves been observed in both the ferric and
ferrous nitrosyl forms of PheH and soybean lipoxygenase.31 32 33 34 35 In chapter 1 we
observed that the CW-EPR spectra of TyrH[tyr], TyrH[6MPH4,tyr] and
E3 32A[6MPH 4,tyr] is composed of a mixture of two slightly rhombic species. Based on
the observation of a mixture in both ferrous and ferric PheH, it is reasonable to assume
that the mixture we observe is independent of the orientation of the NO bond but instead
reflects differences in the coordination environment of the {FeNO}7. The Fe center in
TyrH is known to be coordinated by 2 histidines and a glutamate; the other coordinate
sites are occupied by solvent and NO. Its possible that the two species observed in the
CW-EPR are the result of freezing in different conformers of bound water or of capturing
a mixture, either mono or bidentate, in the ligation of the glutamate side chain.

The results from the study with 6,7-2H-6MPH4 show that in TyrH there is a
decrease in the 2H-Fe distance upon the addition of tyr. In TyrH[d-6MPH4] the 2H

ESEEM signal is not clearly resolved in our quotient spectra. Simulations indicate that

90

 

 

the closest 2H-Fe distance that can account for this “failed observation” is 5.9 A. Our
simulations assumed that only a single 2H of the two on d-6MPH4 is interacting strongly
enough with the Fe to produce ESEEM. This assumption is backed by crystal structures
of TyrH and PheH that show the pterin oriented with C 4a closest to the iron. This
orientation of the pterin places the C6 2H at least 1 A closer to the iron than the C7 2H.
This difference in 2H distance would result in an ESEEM spectrum that is composed
almost entirely of the coupling to the closer of the two deuterons. The distance between

the Fe and the C6 2H determined by ESEEM agrees fairly well with the TyrH and PheH

 

crystal structures that give Fe-C6 distances of 6.5 A and 7.3 A, respectively.‘2 '3 This
range of distances precludes direct binding of the pterin to the Fe and is consistent with
the CW-EPR result (Figure 1.4a) that shows that pterin binding to TyrH by itself has little
or no effect on the symmetry of the zero field interaction of the {FeNO}7 center.

Upon tyr addition to TyrH[6MPH4], the 2H-F e distance decreases to 4.2 A, a
change of at least 1.7 A. This result agrees with changes observed in x-ray crystal
structures of PheH, where the distance between the Fe and the C 4 position of pterin is
reduced by approximately 1.5 A when a substrate analog is bound (lka), as compared
to the structure when just pterin is bound to the enzyme (1dmw).4' '3 The orientation of
the 2H-hyperfine principle axis suggests that the NO does not bind between the pterin and
the Fe, but rather in a more axial position with respect to the plane formed by the Fe and
pterin. These results fit with current ideas on how TyrH controls chemical reactivity to
ensure efficient preparation of DOPA and minimize side reactions that could occur if the

F e(IV)=O species were generated in the absence of tyr. Specifically, the binding of tyr to

91

the TyrH{FeNO}7 6MPH4 complex moves the pterin cosubstrate closer to the {FeNO}7
and presumably into a position where the initial 02-pterin chemistry can occur.

The interaction of tyr with TyrH is considerably different than that of the pterin.
The dipolar distance between 3,5-2H-tyr and the Fe center only changes a small amount
with the addition of the other substrate, 6MPH4. The ESEEM simulation for d-tyr in TyrH
assumed only one of the two deuterons contributes significantly to the ESEEM amplitude.
This assumption is not backed by existing structural evidence as it was with the pterin.
The assumption is made based on two experimental observations, the ESEEM amplitude
and the CW-EPR spectrum, that when taken in concert suggest that tyr is rotated to bring
one of the meta positions closer to the Fe with the other rotated away from the iron. The
observed ESEEM amplitude is inversely related to the dipolar distance and directly
related to the number of interacting nuclei. With two equivalent 2H's, such as would be
the case if the hydroxyl group of the tyr were symmetrically positioned toward the iron,
the minimal dipolar distance required to reproduce the ESEEM amplitude would be ~5 A.
If the tyr is rotated, with one of the positions brought closer while the other is pushed
away, the further of the two deuterons contributes an insignificant amount to the ESEEM
once it reaches ~5.5 A from the Fe. The ESEEM simulations were performed assuming a
single deuteron and therefore represent the closest approach of tyr to the Fe. The CW
EPR spectrum of TyrH[tyr](figure 1.4b) shows a shift away from the purely axial
spectrum observed for TyrH[6MPH4]. This shows that tyr binding to TyrH has an effect
on the symmetry of the {FeNO}7 zero field interaction and is most consistent with the

closer positioning of the C3-2H to the Fe. The range of 2H-Fe distances for tyr in TyrH

92

corroborate what has been suggested by x-ray crystallographic studies of PheH and other
data that indicate Fe does not directly coordinate the primary substrates.4’ ”

The addition of 6MPH4 to TyrH[d-tyr] only has a minor effect on the intensity of
the 2H modulations. Specifically the ESEEM amplitude increases by 18% at g = 2.2
which corresponds to a slight decrease in the 2H-Fe distance from 4.3 A in TyrH[d-tyr] to
4.1 A in TyrH[6MPH4,d-tyr]. The more significant change to the ESEEM spectrum
appears in the line shape. The splitting observed in TyrH[d-tyr] is no longer present when
6MPH4 is added, consistent with a change in the orientation of the 2H-F e vector with
respect to the F e-NO bond. Based on the small change observed in the 2H-Fe distance, it
is reasonable to suggest that the position of substrate tyr with respect to the Fe and the
relative orientation of the 2H HF I and NO] principle axes remain the same with and
without the pterin. The change in orientation observed would then correspond to a
reorientation of the F e-NO bond when the pterin is bound to TyrH. Our results are
consistent with the NO moving from a roughly equatorial position to one that is more
axial with respect to tyr. This change in relative orientation would facilitate the
hydroxylation of tyr upon the formation of the Fe(IV)=O species.

Glu332 in TyrH is part of the pterin binding site in TyrH,22 23 and the structure of
PheH shows that the corresponding residue Glu286 forms hydrogen bonds with the pterin
in that enzyme.4 The E332A mutant shows several changes compared to the wild type
enzyme with respect to the binding of 6MPH4 and tyr. For E332A TyrH [d-6MPH4,tyr]
the proximity of 6MPH4 toithe Fe compared to the wild-type protein is conserved at 4.2

A. However, the orientation of the dipole-dipole vector with respect to the NO bond is

93

eil'l‘fl .

‘l .11.?!) I‘

'I'I?

 

changed considerably, shifting from 65° to 34°, or from equatorial to more axial. On the
other hand, studies of E332A[6MPH4,d-tyr] show that tyr maintains the same orientation
seen in WT-TyrH[6MPH4,d-tyr], but with a longer Fe-ZH distance of 4.5 A. The CW EPR
spectrum of E332A[6MPH4,tyr] is nearly identical to TyrH[6MPH4,tyr] suggesting that

the first coordination sphere of the Fe changes very little as a result of the mutation.

d-tyr

4.1 25No

A @ 65° d-6MPH

4_3 W 2.1
A 71° T
@ ”0 +6MPH, d-6MPH
d-terO
E332 E332
A

.5 ' °N 34, d-6MPH
A
® 4.2
A

Figure 3.8: Summary of ESEEM Results

.h

The structural relationships between the dipole-dipole vectors and the F e-NO
bond for the four cases we examined with 2H ESEEM are summarized in figure 3.8.
These dipole-dipole vectors in theory represent cones about the Fe-NO bond on which the
substrate deuterons can be found. We cannot with certainty define the orientation of the
NO in the enzyme; however, using the crystal structure of PheH with pterin and substrate
analog as a model, it is very probable that the substrates lie opposite one another with
respect to the F e-NO bond. If NO is indeed mimicking O2 binding, we can then use its
structural relationship with the substrate deuterons to draw some conclusions about the

mechanism of hydroxylation in TyrH. The formation of the initial Fe-OO-pterin has been

94

 

 

 

suggested to involve a concerted reaction of oxygen, the tetrahydropterin, and the iron.ll
The positioning of NO in TyrH[6MPH ,tyr] is appropriate for such a reaction. The Fe-Oz-
pterin intermediate then heterolytically cleaves when the oxygen atom distal to the Fe(lI)
is protonated, forming the F e(IV)=O species.‘1 The structural model in Scheme 3 would
place the F e(IV)O oxygen in close proximity to the tyrosine carbon that is hydroxylated.

For productive hydroxylation of their substrates, enzymes such as TyrH that utilize
highly electrophilic Fe(IV)O intermediates must not only catalyze the formation of such
an intermediate, but also suppress all unproductive reactions both before and after its
formation. The TyrH mutant E3 32A, does not hydroxylate tyr but does produce an
oxidized pterin species, possibly 4a-hydroperoxypterin.22 ” The structural evidence
obtained from the ESEEM measurements suggest that this uncoupled reaction is a
consequence of the different pterin orientation in the mutant enzyme, resulting in a
geometrically altered F e-Oz-pterin activated complex. This altered orientation results in
cleavage of the oxygen-iron bond instead of the oxygen-oxygen bond, preventing
formation of the F e(IV)O intermediate. The previous suggestion that this is due to
protonation of the distal rather than the proximal oxygen11 is fully consistent with the
present results. In addition, the altered position of the pterin may weaken the oxygen-iron
bond, facilitating its cleavage.
3.5 CONCLUSIONS

The present results for the first time establish the relative positions of the three
substrates, tetrahydropterin, tyrosine, and oxygen, in the active site of an aromatic amino

acid hydroxylase. The resulting structural details extend the understanding of the

95

 

mechanisms of these enzymes.

 

 

96

 

3.6 REFERENCES

1. Haavik, J.; Flatmark, T., Isolation and Characterization of Tetrahydropterin
Oxidation Products Generated in the Tyrosine 3-Monooxygenase (Tyrosine Hydroxylase)
Reaction. European Journal of Biochemistry 1987, 168 (1), 21 -6.

2. Fitzpatrick, P. F., Mechanism of Aromatic Amino Acid Hydroxylation.
Biochemistry-Us 2003, 42 (48), 14083-14091.

3. Goodwill, K. E.; Sabatier, C.; Marks, C.; Raag, R.; Fitzpatrick, P. F.; Stevens, R.
C., Crystal Structure of Tyrosine Hydroxylase at 2.3 Angstrom and Its Implications for
Inherited Neurodegenerative Diseases. Nat Struct Biol 1997, 4 (7), 578-585.

4. Andersen, O. A.; F latmark, T.; Hough, E., Crystal Structure of the Ternary
Complex‘of the Catalytic Domain of Human Phenylalanine Hydroxylase with
Tetrahydrobiopterin and 3-(2-Thienyl)-L-Alanine, and Its Implications for the Mechanism
of Catalysis and Substrate Activation. J Mol Biol 2002, 320(5), 1095-1108.

5. Windahl, M. S.; Petersen, C. R.; Christensen, H. E. M.; Harris, P., Crystal
Structure of Tryptophan Hydroxylase with Bound Amino Acid Substrate. Biochemistry-
Us 2008, 47 (46), 12087-12094.

6. Eser, B. E.; Barr, E. W.; Frantom, P. A.; Saleh, L.; Bollinger, J. M.; Krebs, C.;
Fitzpatrick, P. F., Direct Spectroscopic Evidence for a High-Spin F e(Iv) Intermediate in
Tyrosine Hydroxylase. J Am Chem Soc 2007, 129 (37), 11334-+.

7. Eser, B. E.; Fitzpatrick, P. F., Measurement of Intrinsic Rate Constants in the
Tyrosine Hydroxylase Reaction. Biochemistry- Us 2010, 49(3), 645-652.

8. Hillas, P. J .; Fitzpatrick, P. F., A Mechanism for Hydroxylation by Tyrosine
Hydroxylase Based on Partitioning of Substituted Phenylalanines. Biochemistry- Us 1996,
35 (22), 6969-6975.

9. Fitzpatrick, P. F., Steady-State Kinetic Mechanism of Rat Tyrosine-Hydroxylase.
Biochemistry-Us 1991, 30 (15), 3658-3662.

10. Sura, G. R.; Lasagna, M.; Gawandi, V.; Reinhart, G. D.; Fitzpatrick, P. F., Effects
of Ligands on the Mobility of anActive-Site Loop in Tyrosine Hydroxylase as Monitored
by Fluorescence Anisotropy. Biochemistry-Us 2006, 45 (31), 9632-9638.

11. Chow, M. S.; Eser, B. E.; Wilson, S. A.; Hodgson, K. O.; Hedman, B.; Fitzpatrick,
P. F.; Solomon, E. I., Spectroscopy and Kinetics of Wild-Type and Mutant Tyrosine
Hydroxylase: Mechanistic Insight into O-2 Activation. J Am Chem Soc 2009, 131 (22),
7685-7698.

12. Goodwill, K. E.; Sabatier, C.; Stevens, R. C., Crystal Structure of Tyrosine
Hydroxylase with Bound Cofactor Analogue and Iron at 2.3 Angstrom Resolution: Self-

97

 

 

Hydroxylation of Phe3 00 and the Pterin-Binding Site. Biochemistry- Us 1998, 3 7 (39),
13437-13445.

13. Erlandsen, H.; Bjorgo, E.; Flatmark, T.; Stevens, R. C., Crystal Structure and Site-
Specific Mutagenesis of Pterin-Bound Human Phenylalanine Hydroxylase. Biochemistry-
Us 2000, 39 (9), 2208-2217.

14. Andersen, O. A.; Stokka, A. J.; F latrnark, T.; Hough, E., 2.0 Angstrom Resolution
Crystal Structures of the Ternary Complexes of Human Phenylalanine Hydroxylase
Catalytic Domain with Tetrahydrobiopterin and 3-(2-Thienyl)-L-Alanine or L-Norleucine:
Substrate Specificity and Molecular Motions Related to Substrate Binding. J Mol Biol
2003, 333 (4), 747-757.

15. Rich, P. R.; Salerno, J. C.; Leigh, J. S.; Bonner, W. D., Spin 3-2 Ferrous-Nitric
Oxide Derivative of an Iron-Containing Moiety Associated with Neurospora-Crassa and
Higher Plant-Mitochondria. Febs Lett 1978, 93 (2), 323-326.

16. Arciero, D. M.; Lipscomb, J. D.; Huynh, B. H.; Kent, T. A.; Munck, E., Electron-
Paramagnetic-Res and Mossbauer Studies of Protocatechuate 4,5-Dioxygenase -
Characterization of a New Fe-2+ Environment. J Biol Chem 1983, 258 (24), 4981-4991.

17. Brown, C. A.; Pavlosky, M. A.; Westre, T. B; Zhang, Y.; Hedman, B.; Hodgson,
K. 0.; Solomon, E. I., Spectroscopic and Theoretical Description of the Electronic-
Structure of S=3/2 Iron-Nitrosyl Complexes and Their Relation to O-2 Activation by
Nonheme Tron Enzyme Active-Sites. J Am Chem Soc 1995, 11 7(2), 715-732.

18. Yang, T. C.; Wolfe, M. D.; Neibergall, M. B.; Mekmouche, Y.; Lipscomb, J. D.;
Hoffman, B. M., Substrate Binding to No-Ferro-Naphthalene 1,2-Dioxygenase Studied by
High-Resolution Q-Band Pulsed H-2-Endor Spectroscopy. J Am Chem Soc 2003, 125
(23), 7056-7066.

19. Tierney, D. L.; Rocklin, A. M.; Lipscomb, J. D.; Que, L.; Hoffman, B. M., Endor
Studies of the Ligation and Structure of the Non-Heme Iron Site in Acc Oxidase. J Am
Chem Soc 2005, 127 (19), 7005-7013.

20. Muthukumaran, R. B.; Grzyska, P. K.; Hausinger, R. P.; McCracken, J., Probing
the Iron-Substrate Orientation for Taurine/Alpha-Ketoglutarate Dioxygenase Using
Deuterium Electron Spin Echo Envelope Modulation Spectroscopy. Biochemistry— Us
2007, 46 (20), 5951-5959.

21. Fitzpatrick, P. F., The Ph-Dependence of Binding of Inhibitors to Bovine Adrenal
Tyrosine-Hydroxylase.J Biol Chem 1988, 263 (31), 16058-16062.

22. Frantom, P. A.; Seravalli, J.; Ragsdale, S. W.; Fitzpatrick, P. F., Reduction and

Oxidation of the Active Site Iron in Tyrosine Hydroxylase: Kinetics and Specificity.
Biochemistry-Us 2006, 45 (7), 2372-2379.

98

23. Daubner, S. C.; Fitzpatrick, P. F., Site-Directed Mutants of Charged Residues in
the Active Site of Tyrosine Hydroxylase. Biochemistry- Us 1999, 38 (14), 4448-4454.

24. Ramsey, A. J.; Hillas, P. J .; Fitzpatrick, P. F., Characterization of the Active Site
Iron in Tyrosine Hydroxylase - Redox States of the Iron. J Biol Chem 1996, 271 (40),
24395-24400.

25. Keefer, L. K.; Nims, R. W.; Davies, K. M.; Wink, D. A., "Nonoates" (l-Substituted
Diazen-l-Ium-l ,2-Diolates) as Nitric Oxide Donors: Convenient Nitric Oxide Dosage
Forms. Method Enzymol 1996, 268, 281-293.

26. Mims, W. B.; Davis, J. L.; Peisach, J ., The Accessibility of Type-I Cu(Ii) Centers
in Laccase, Azurin, and Stellacyanin to Exchangeable Hydrogen and Ambient Water.
Biophys J 1984, 45 (4), 755-766.

27. Mims, W. B., Envelope Modulation in Spin-Echo Experiments. Phys Rev B 1972,
5 (7), 2409-&.

28. Hutchison, C. A.; Mckay, D. B., Determination of Hydrogen Coordinates in
Lanthanum Nicotinate Dihydrate Crystals by Nd+3-Proton Double-Resonance.J Chem
Phys 1977, 66(8), 3311-3330.

29. Hurst, G. C.; Henderson, T. A.; Kreilick, R. W., Angle-Selected Endor
Spectroscopy .1. Theoretical Interpretation of Endor Shifts from Randomly Orientated
Transition-Metal Complexes. J Am Chem Soc 1985, 107 (25), 7294-7299.

30. Edmonds, D. T., Nuclear-Quadrupole Double-Resonance. Phys Rep 1977, 29(4),
234-290.

31. Bloom, L. M.; Benkovic, S. J .; Gaffney, B. J., Characterization of Phenylalanine-
Hydroxylase. Biochemistry- Us 1986, 25 (15), 4204-4210.

32. Kappock, T. J.; Harkins, P. C.; F riedenberg, S.; Caradonna, J. P., Spectroscopic
and Kinetic Properties of Unphosphorylated Rat Hepatic Phenylalanine Hydroxylase
Expressed in Escherichia Coli - Comparison of Resting and Activated States. J Biol
Chem 1995, 270 (51), 30532-30544.

33. Han, A. Y.; Lee, A. Q.; Abu—Omar, M. M., Epr and Uv-Vis Studies of the Nitric
Oxide Adducts of Bacterial Phenylalanine Hydroxylase: Effects of Cofactor and Substrate
on the Iron Environment. Inorg Chem 2006, 45 (10), 4277-4283.

34. Wallick, D. E.; Bloom, L. M.; Gaffney, B. J.; Benkovic, S. J., Reductive
Activation of Phenylalanine-Hydroxylase and Its Effect on the Redox State of the Non-
Heme Iron. Biochemistry-Us 1984, 23 (6), 1295-1302.

35. Salerno, J. C.; Siedow, J. N., Nature of the Nitric-Oxide Complexes of
Lipoxygenase. Biochim Biophys Acta 1979, 579(1), 246-251.

99

 

CHAPTER IV
HYSCORE CHARACTERIZATION OF TYROSINE HYDROXYLASE AND
PHENYLALANINE HYDROXYLASE

4.1 INTRODUCTION

This chapter is a collection of results and discussions from TyrH and PheH. These
are interesting results that are worth noting but they don't stand on their own like the
results presented in the previous chapters. Attempts were made to collect HYSCORE
spectra on the binary, ternary and quaternary complexes of both enzymes; however, only
the quaternary complexes provided suitable crosspeaks that allowed for analysis. These
HYSCORE results of the quaternary complex of PheH and TyrH will be presented
accompanied by analysis of the exchangeable hydrogen and non-exchangeable hydrogen.
In addition, the preliminary analysis of 2H ESEEM collected on PheH treated with
selectively deuterated substrates will be presented.
4.2 HYSCORE RESULTS

In chapter 2, we characterized the HYSCORE spectrum of a non-heme {F eNO}7
model system. We provided the identification of the features we could expect to observe
in the HYSCORE spectra of non-heme iron enzymes which included directly coordinated
nitrogen, solvent molecules and non-exchangeable hydrogen. These results will provide a
guide to help understand HYSCORE spectra collected on TyrH and PheH. Similar to the
model system, HYSCORE spectra were acquired at multiple field positions from g = 4 to
g = 2 and utilized in the analysis; however, only representative spectra at 1980 G(g =

3.489) and 3100 G(g = 2.229) will be shown.

100

 

 

Shown in figures 4.1 and 4.2 are the HYSCORE spectra of the TyrH quaternary
complex acquired at 1980 G and 3100 G, respectively. In the (++) quadrant of both
spectra, we observe features in the region expected for hydrogen. At 1980 G, two sets of

hydrogen features are observed, one set with a maximum at (59,122) MHz and a second

 

 

 

 

14 r I l I 1 I
12"
10”
g g
E 6’ o
4" 80
2' . . Free

 

 

o
-14 -12 -1o -8 -6 4 -2
F1 (MHz)

Figure 4.1: HYSCORE spectrum of TyrH[6MPH4,tyr] at 1980 G

 

 

 

 

 

 

18 I I I I I Ifif f I I I I
16-
14-
Q
12-
.N. “a
£10-
5 Br I ,° 1
LL
6- .
4— _
ob .
2— O .o d
O 4 I I I l J J L l
-18-16-14-12-10-8-6-4-2 0 2 4 6 81012141618

F1 (MHz)

Figure 4.2: HYSCORE spectrum of TyrH[6MPH ,tyr] at 3100G

101

 

set with a maximum at (68,102) MHz. With an increase of the magnetic field to 3100
G, the two sets of hydrogen features are observed to now have maxima at (121,144)
MHz and (10.7,15.95) MHz, in addition, there is a new higher frequency peak at (12,16)
MHz. At 3100 G in the (-+) quadrant the directly coordinated nitrogen is observed with a
peak at (-3.8,8.1) MHz. In addition to the expected HYSCORE features, a new group of
features, not observed in the model, can be seen in the range from 0 — 4 MHz at both field
positions. The resolution of these features is largely obscured by broad intense features
along the diagonal.

Shown in figure 4.3 and 4.4 are the HYSCORE spectra acquired on the quaternary
complex of PheH at 1980 G and 3100 G, respectively. These two spectra show features
similar to those observed in the HYSCORE spectra of TyrH. At 1980 G, the (++)
quadrant has two sets of hydrogen features, one set with a maximum at (6.2,12) MHz and

the second set with a maximum at (6.8,10. 1) MHz. At 3100 G, the two sets of hydrogen

 

F2 (MHz)

 

 

. g ' r
l l 1 l I

18-16-14-12-10 -8 -8 -4 -2 o 2 4 6 81012141618
F1 (MHz)

 

 

Figure 4.3: HYSCORE spectrum of PheH[Phe, 5d] at 1980 G

102

 

 

 

 

 

 

 

 

18 I I I
16-
14-
12-
’1»?
I 10
a F
N 8”
LI.
6..
4..
2- r“-
0 I l L I I J l l l I
-18-16-14-12-10 -8 -6 -4 -2 0 2 4 6 81012141618
F1 (MHz)
Figure 4.4: HYSCORE spectrum of PheH[Phe, 5d] at 3100G
5:4

features are observed to now have maxima at (12,146) MHz and (105,161) MHz, in
addition, there is a new higher frequency peak at (112,168) MHz. A strongly coupled
nitrogen feature is also observed in the (-+) quadrant with a maximum at (-3.8,8.3) and a
second maximum at (-4.3,6.2). Again, features consistent with weakly coupled nitrogen
can be seen from 0-4 MHz and the resolution of these features is largely obscured by an
intense feature along the diagonal.
4.3 HISTIDINE HYPERFINE COUPLING ANALYSIS

As noted previously, the Fe center in TyrH and PheH is facially coordinated by
two histidine ligands and a glutamate. Further examination of the crystal structure1
suggests that the histidine ligands provide the only enzyme based magnetic nuclei within
~45 A. Each histidine provides a directly coordinated nitrogen, a remote nitrogen ~4.4 A

from the iron and two non-exchangeable hydrogens on the 2 and 5 positions which are

~3.2 A from the iron. Based on the work in chapter 2, it is possible to immediately assign

103

 

 

 

 
   

 

 

 

 

 

 

14 . . . - 18 -
b
12* 16
74‘ 74‘
I I
g 10- g 141
8 5
C C
8 8
5’ at Q g 12 r
u. L
6 10+
4 . r . . 8 r r . ¥
4 6 8 10 12 14 8 10 12 14 16 18
Frequency(MHz) Frequency(MHz)

Figure 4.5: 1H HYSCORE simulations for TyrH at 1980 G (a) and 3100 G (b).

the features attributed to the directly coordinated nitrogen and the hydrogen. Through a
process of elimination, it is reasonable to assume that the hyperfine frequencies observed
in the 0 — 4 MHz range represent the remote nitrogen; however, due to the poor resolution
it cannot be characterized.

Shown in figures 4.5a and 4.5b are the 1H HYSCORE simulations at 1980 G and
3100 G, respectively, for the hydrogen features seen in the quaternary complexes of TyrH.
The simulations for TyrH, seen in figure 4.5, are composed of two different hydrogen
interactions which represent different orientations for the PAS of the hydrogen hyperfine
tensor with respect to the g2 axis. At 1980 G, the hydrogen that can be characterized with
the maximum at (59,122) used a dipole-dipole coupled hydrogen with a distance of 3.2
d: 0.2 A and an orientation of the hyperfine principal axis with respect to gZ of 1.55 :t 0.1
rad. This hydrogen with an orientation nearly perpendicular to the gl axis generated the

cross peaks with maximum near (12.1,14.4) MHz at 3100 G. The second hydrogen,

104

 

 

   

 

 

 

 

 

 

14 13 . I . -
b
12* 16%
11? RT
I I
g 10 § 3 14 .
>4 >5
0 O
C C
“3’ 8 1 g 12
5 % 53’
U- u.
6 « 1o
4 . r r r 8 . . . .
4 6 8 1o 12 14 8 10 12 14 16 18
Frequency(MHz) . Frequency(MHz)

Figure 4.6: 1H HYSCORE simulations for PheH at 1980 G (a) and 3100 G (b).

characterized the maximum near (6.8,10.2) MHz at 1980G and (10.7,15.95) MHz at
3100G, used a dipole-dipole distance of 3.16 :t 0.2 A and an orientation of the hyperfine
principal axis with respect to gz of 0.34 i 0.1 rad.

The 1H HYSCORE simulation for PheH at 1980 G and 3100 G, shown in figures
4.6a and 4.6b, respectively, used parameters similar to those utilized in the TyrH
simulations. The hydrogen that can be characterized by the maximum at (6.2,12) at 1980
G and (12,146) MHZ at 3100 G, assumed a dipole-dipole coupled hydrogen with a
distance of 3.2 :I: 0.2 A and an orientation of the hyperfine principal axis with respect to
gZ of 1.52 :t 0.1 rad. The second hydrogen with maxima at (68,101) MHz and(10.5,16.1)
MHz used a dipole-dipole distance of 3.12 :1: 0.2 A and an orientation of the hyperfine
principal axis with respect to gz of 0.34 :1: 0.1 rad.

The Fe-H dipolar couplings obtained from these simulations correspond well with

the distances expected for the two hydrogens on the 2 and 5 position of the histidine. An

105

 

 

 

 

 

 

14 ' . E estimation, based on the ternary crystal
12 - .
structure of PheH“ for the distance between
10 r 1
I”? 8 . E ‘ the rron and the hydrogen 8 on the 2 and 5
2
E 6 , @O . position of the histidines is 3.1 A — 3.8 A. Our
4 ~ @ a . HYSCORE results only provide distances for
2 ' ' two distinct hydrogen populations, this can be

 

 

 

5)14 -12 -1‘0 .8 -6 .4 -2 ointerpreted in one of two ways: the hydrogen's
F1 (MHZ)
on the 2 and 5 positions share nearly the same
Figure 4.7: 14N HYSCORE simulations at Fe-H distances and we only have two types of
3100 G
hydrogen which differs by orientation or only
two of the four hydrogen's significantly contribute to the HYSCORE, obscuring the other
two more distant and weaker components. Depending on the method that these distances
are interpreted will determine how the orientation information is utilized to provide
restrictions on the position of the NO bond within the active site of these enzymes.
Proper simulation of the directly coordinated nitrogen should provide more clarity
to the above results and allow for a more restricted position for the NO bonding. A
sample simulation for the directly coordinated nitrogen in TyrH at 3100G is shown in
figure 4.7. The simulation utilized the following parameters: aiso = 7.5 MHz, T = 1 MHz,
an angle between the dipolar vector and the Fe-NO bond of 0.2 radians, and equ= 3.0
MHz with an n=0.1. These parameters are consistent with those observed for the directly

coordinated nitrogen in the model system.

The problem with a complete analysis of the Fe-NO bond orientation within the

106

 

active site is the lack of a second nitrogen coupling unless both nitrogen's share a similar
orientation with respect to gz. In PheH we did observe a second strongly coupled nitrogen
feature; however, it is likely only a small change in orientation, ~0.2 radians. Determining
the orientation of the Fe-NO within the active site will require more work on model
systems, such a TauD, in which we know the NO orientation with respect to the histidine
ligands
4.4 EXCHANGEABLE HYDROGEN

In the field range of 2800 G — 3200 G we were able to identify the frequency
components in the HYSCORE spectra of the model system which represent coordinated
water. In figures 4.2 and 4.4 we see a higher frequency component, one that we've
previously attributed to H20, represented by the peaks at (12,16) MHz and (112,168)
MHz, respectively. To verify that these higher frequency features were solvent, samples in
D20 were examined. It was observed that the quaternary complex of TyrH in D20 buffer

did not contain the higher frequency feature at (12,16) MHz; on the other hand, the

 

 

 

 

 

 

 

 

 

 

18 . . . . 18 a . 4
a b

16- 16 0‘
$714— 1714- 1
I I
a .2,
LC: 12' E 12F

10- 10

8 r L . . 8 r . n r
8 1O 12 14 16 18 8 10 12 14 16 18
F1 (MHz) F1 (MHz)

Figure 4.6: 1H HYSCORE simulations at 3100 G for TyrH(a) and PheH(b).

107

quaternary complex of PheH in D20 buffer still contained this high frequency feature.
This result was unexpected and it initially led us to conclude that TyrH had a coordinated
water in the quaternary complex while PheH did not. The question to the identity of this
high frequency component in PheH led us to reexamine the crystal structure of the ternary
complex of PheH. It was seen that the only hydrogen, aside from water, that could be
closer than 3 A would be from the pterin, specifically if the N5 position were protonated.l
Our observations make sense when the specific sample preparations were considered, the
PheH samples were prepared using 5-deazapterin in which the 5 position is occupied by a
carbon which would contain a non-exchangeable hydrogen while the TyrH samples
utilized 6-Methyltetrahydropterin with a nitrogen in the 5 position which could contain an
exchangeable hydrogen.

Shown in figures 4.8a and 4.8b are the 1H HYSCORE simulations, at 3100 G, for
the features assumed to be the hydrogen on the 5 position of the pterin in the quaternary
complex of TyrH and PheH, respectively. The TyrH simulation(fig. 4.8a) assumed a single
dipole-dipole coupled hydrogen with a distance of 2.8 i 0.2 A and an orientation of the
hyperfine principal axis with respect to gz of 1.15 i 0.1 radians. The PheH simulation(fig.
4.8b) assumed a single dipole-dipole coupled hydrogen with a distance of 2.7 :l: 0.2 A and
an orientation of the hyperfine principal axis with respect to gz of 1.10 d: 0.1 radians.

The distances provided by these results are consistent with the observation made
previously(chapter 3) and those distances suggested by the crystal structure of the ternary
complex of PheH. In the crystal structure of PheH, the distance measured between the Fe

and the N5 position on the pterin is 3.7 A. 1 With a N-H bond length of approximately 1

108

 

 

 

 

A, our HYSCORE result for PheH of 2.7 i 0.2 A represents the minimum F e-H distance
one could expect based on the known crystallographic data. Further confirmation of this
assignment was obtained with the disappearance of this feature when the HYSCORE of
quaternary PheH prepared with 5,6,7-2H-5-deazapterin was examined.
4.5 2H ESEEM ON PHEH

A study similar to the one described in chapter 3 on TyrH was undertaken to
examine the position of the substrate and co-substrate binding in PheH. In chapter 1 we
examined the CW-EPR spectra of three derivatives of {F eNO}7-PheH: PheH treated with
5-deazapterin, PheH[5d], PheH treated with L-phenylalanine, PheH[Phe], and PheH
treated with L- phenylalanine and 5-deazapterin, PheH[Phe, 5d]; and three derivatives of
truncated {FeNO}7-Al-117PheH were examined: A1-117PheH treated with 5-
deazapterin, APheH[5d], A1-117PheH treated with L-phenylalanine, APheH[Phe], and
A1-117PheH treated with L- phenylalanine and 5-deazapterin, APheH[Phe, 5d]. The CW-
EPR spectra showed changes with substrate additions; however, they these changes did
not provide structural information. In order extract the substrate hyperfine information
ESEEM experiments were undertaken. To isolate the interactions between the substrates
and the Fe center ESEEM experiments were run in parallel with protonated substrate and
selectively deuterated substrate. The time domain ESEEM results were normalized to the
amplitude at the earliest time point and then divided, leaving predominately the
deuterium ESEEM contribution.

Shown in figures 4.9a and 4% are the 2H-ESEEM spectra of {FeNO}7-PheH

treated with 5,6,7-2H-5-deazapterin, PheH[d-Sd], at 1980 G and 3100 G, respectively.

109

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

7 . 7 . . - -
6 L a 6 ’ b ‘ Simulations of these spectra (red traces)
E 5i i E 5i ‘ assuming a single, coupled deuteron yielded
V 41 t 5; 4. .
2; 3 ' g 3 ' . a dipolar distance of4.5 :1: 0.2 A, an
5 2 » i 2 . -
1 . 1 . . orientation of the hyperfine principal axis
0 - , 0 ~
0 1 2 3 4 5 0 1 2 3 4 5with respect to gZ of 1.1 i 0.1 rad and an
Frequency (M Hz) Frequency (M Hz)
7 - 7 f orientation between the PAS of the nuclear
c d
6 i 6 i ‘ quadrupole and the PAS of the hyperfine
3" 5’ . 3" 5 .
9’, 9L
0 4' ‘ g 4 ‘tensor of0.6 :t 0.1 rad. When {FeNO}7-
i 3‘ ‘ 'a 3 ‘
E 2 . E 2 PheH is treated with 5,6,7-2H-5-deazapterin
1 t 1 .
o , o and L-phenylalanine, PheH[d-Sd, phe], the

0 1 2 8 4 5 o 1 2 8 4 5
“my (MHZ) Fresmy (MHZ) ZH-ESEEM spectra acquired at 1980 G and
Figure 4.9: 2H ESEEM collected on

PheH[d-Sd] at (8)1980 G and (b) 3100 G); 3100 0. Figures 49c and 49d, show a
and PheH[d-Sd, phe] at (c)l980 G and (d)
3100 G. Experimental results shown in significant change in amplitude as
black with corresponding simulations in red

compared to the corresponding PheH[d-Sd]
spectra shown in Figures 4.9a and 4%. The red traces in figures 4.9c and 4.9d show the
corresponding simulations for PheH[d-Sd, phe] that result from a single deuteron with an
angle between the PAS of the hyperfine tensor and & of 1.10 i 0.1 radians, a 2H-dipolar
distance of 2.7 i 0.2 A, and an angle between the PAS of the nuclear quadrupole and the
PAS of the hyperfine tensor of 0.1 :1: 0.1 rad.

Shown in figures 4.10a and 4.10b are the zH-ESEEM spectra of {FeNO}7-Al -

ll7PheH treated with 5,6,7-2H-5-deazapterin, APheH[d-5d], at 1980 G and 3100 G,

 

respectively. Simulations of these spectra (red traces) assuming a single, coupled deuteron

110

 

 

 

 

 

 

 

 

 

 

7 - 7 - . . -
at a 6_ 1, yielded a dipolar distance of 4.5 :t 0.2 A, an
If 5* ‘ E 5’ orientation of the hyperfine principal axis
27 4r . I 4’ *
éa- 13:13) twithrespecttogzofl.2i0.1rad,andan
5. 2 - t 5 2 . -
1 . 1 1m 1 orientation between the PAS of the nuclear
o - - o - ’“
o 1 2 3 4 5 0 1 2 3 4 5quadrupole and the PAS of the hyperfine
Frequency (MHz) Frequency (MHz)
7 - f 7 - - - # tensor of 0.4 d: 0.1 rad. When {FeNO}7-
6 . ° 6 » d .
A 5 A 5 PheH is treated with 5,6,7-2H-5-deazapterin
=i ’ ‘ =1 ' ‘
s 3
g 4' ‘ 3 4' ‘and L-phenylalanine, APheH[d-5d, phe], the
-_ 3 - . -._ 3 . .
Q Q
5 2 . 5 2 . rzH-ESEEM spectra acquired at 1980 G and
1 » 1 . .
0 0 31006, Figures 4.10c and 4.10d, show a

 

 

 

 

 

 

012345 012345
Frequency (MHZ) mey (”'11) modest change in amplitude as compared to

Figure 4.10: 2H ESEEM collected on
APheH[d-5d] at (a)1980 G and (b) 3100 G); the corresponding APheHld-Sd] spectra
and APheH[d-5d, phe] at (c)l980 G and (d)
3100 G. Experimental results shown in shown in Figures 4.9a and 4.9b. The red
black with corresponding simulations in red

traces in figures 4.10c and 4.10d show the

corresponding simulations for APheH[d-5d, phe] that result from a single deuteron with
an orientation between the PAS of the hyperfine tensor and g2 of 1.10 :1: 0.1 radians, a 2H-
dipolar distance of 2.7 :I: 0.2 A, and an angle between the PAS of the nuclear quadrupole
and the PAS of the hyperfine tensor of 0.1 i 0.1 rad.

These simulations are an initial approximation for the pterin distances. Based on
the deuterium labeling scheme of the pterin, up to 3 deuterons could be interacting with
the Fe; however, our simulation only assumed a single deuteron. Previously(chapter 3),

we discussed disregarding the interaction with the C7 deuteron based on its distance and

111

 

likely ESEEM contribution. If this deuteron is disregarded, there still remains potentially
two interacting deuterons. The HYSCORE results described above provided a distance
and orientation for what was assumed the C5 hydrogen on the pterin, using these
parameters 2H ESEEM simulations were performed and compared to the experimental
results. The simulations performed with only a single deuteron provided a reasonable fit
at multiple field positions. It was also recognized that the pterin most likely binds in a
similar fashion to both TyrH and PheH, this is based on the similarity of the parameters
used to simulate the HYSCORE feature we assigned to the CS-pterin hydrogen. Inclusion
of C6 deuteron, using the parameters obtained in TyrH, resulted in at most a 5% change
in amplitude and a minimal line shape effect. It was therefore concluded that the ESEEM
contributions from the C6 deuteron is negligible and the spectrum is dominated by the
more intense C5 deuteron.

These results show that, much like our results in TyrH, the presence of the primary
substrate effects a change in the active site which allows for closer pterin binding to the
iron. It also shows that the truncated enzyme has behavior similar to that of the full length
enzyme, a result which is supported by the CW—EPR spectra.

An attempt was made to examine the 2H ESEEM spectra of PheH treated with
para-zH-phe(d-phe). With the exception of PheH[d-phe], the ESEEM results could be
characterized by a signal to noise ratio <1 and in many cases did not even have a weak
line at the deuterium Larmour frequency. This led us to examine samples prepared with
per-zH-phe(5d-phe), these samples did provide ESEEM, however the analysis is difficult

and they were only treated qualitatively.

112

 

Shown in figures 4.1 la and 4.11b are the 2H ESEEM of PheH[d-phe] at 1980 G
and 3100 G, respectively. The red overlay represents a simulation which utilized a single
deuteron with a dipole-dipole distance to the Fe of 4.2 i 0.2 A, an orientation of the
hyperfine principal axis with respect to g.Z of 1.2 i 0.1 rad, and an orientation between the
PAS of the nuclear quadrupole and the PAS of the hyperfine tensor of 1.4 a: 0.1 rad.

Assuming that PheH behaves in a similar fashion as TyrH with respect to primary
substrate binding, PheH[d-phe] binds in a location within the active site that remains
largely unchanged on co-substrate binding. If this is the case, then the above results in

combination with the results from the samples treated with 5d-phe can be used to show

2-.-- 2‘4

 

 

a h f I the relative changes that occur on substrate
I; 1.5 t ‘ g 1-5 ’ i . . .
3 3; binding. Usrng the field dependent
i 1 t g 1 . . . .
:51 '7; amplrtude changes described in chapter 3,
5 0.5 45 0.51 .
we were able to observe that 5d-phe

 

 

 

 

 

 

- - - r o - r - - . . . .
0 1 2 3 4 5 0 1 2 3 4 5underwent a change 1n onentatron wrth
Frequency (MHz) Frequency (MHz)

Figure 4.11: 2H ESEEM collected on respect to the gz axis upon co-substrate

PheH[d-phe] at (a)1980 G and (b) 3100 G). . . . .
Experimental results shown in black with addition. It was seen that the onentatron of

corresponding simulations in red

the net dipole-dipole vector went from one
more perpendicular to the g2 axis to one which was one more co-linear. This result agrees
with our failed observation of PheH[d-phe, 5d], if the dipole-dipole vector and the g2 axis
were colinear the number of crystal orientations which contribute to the ESEEM would

be greatly limited and any distribution of structures would result in a loss of ESEEM

amplitude.

113

4.6 CONCLUSIONS

Using a combination of 1H HYSCORE and 2H ESEEM a portion of the control
mechanism within PheH was revealed. It was shown that PheH controls pterin binding
much like we've seen in TyrH. The pterin is initially positioned at a non-reactive distance
from the iron with the addition of primary substrate a structural rearrangement occurs and
the pterin is brought closer to the iron. These results agree very well with observations

made on the crystal structures obtained on PheH.

 

114

4.7 REFERENCES

1. Andersen, O. A.; Flatmark, T.; Hough, E., Crystal Structure of the Ternary
Complex of the Catalytic Domain of Human Phenylalanine Hydroxylase with
Tetrahydrobiopterin and 3-(2-Thienyl)-L-Alanine, and Its Implications for the Mechanism
of Catalysis and Substrate Activation. J Mol Biol 2002, 320(5), 1095-1108.

115

 

APPENDIX I
MATLAB SIMULATION SCRIPTS
The simulation programs were written in matlab based off the original structure written
by John McCracken in F ortran. The programs include a number of improvements and

fixes over the original code.

The simulation scripts are written to accept the input parameters in terms of 3 sets, the
experimental parameters, the hyperfine parameters and the electron parameters. These

sets of input parameters are arranged for ease of storage and usage into three matrices.

mfile.m - command line syntax — description

parameters.m — [expp,hfip,gp] =parameters — This mfile contains the three matrices and
their structure used in the simulation routines. The default values in these matrices can be
edited or the matrices themselves could be edited once in the workspace. expp contains
the experimental parameters, hfip contains the coupling parameters, gp contains the

parameters used when there is anisotropy in EPR spectrum requiring angle selection.

sims.m — Sims — Program to interface the ESEEM and HYSCORE simulations into a user
friendly gui. Overall the input parameters should be self explanatory.
simplot.m — Utility script for sims.m, used to plot within the gui, helps to simplify the

main code.

116

simprint.m — Utility script for sims.m, used to "print" a simulation to a figure for

printing.

esehh.m — [rawA,rawB]=esehh(expp,hfip) — I=l/2 3P ESEEM simulation. With an axial
hyperfine interaction this will use an analytical formula for speed. When the hyperfine
interaction is rhombic this will use a full density matrix calculation.

esehhang.m — [rawA,rawB]=esehhang(expp,hfip,gp) —- I=1/2 Angular selection 3P
ESEEM using a density matrix calculation

eseho.m — [rawA,rawB]=eseho(expp,hfip) — I=1 3P ESEEM using a density matrix
calculation

esehoang.m — [rawA,rawB]=esehoang(expp,hfip,gp) — I=1 Angular selection 3P ESEEM

using a density matrix calculation

The output from the ESEEM simulations comes into two parts pieces, rawA the
alpha spin manifold and rawB the beta spin manifold. If S=3/2 is usedonly the S=1/2
manifold will be evaluated, specific for F e-NO. For an S=5/2, such as Mn”, the
simulations will evaluate the five spin transitions allowed, outputting both the upper and

lower manifolds into rawA and rawB respectively.

eseassemble.m — [spectra spectraft] =eseassemble(expp, rawA,rawB, weights) —— Used to
process the raw simulation data. This will apply the product rule for multiple nuclei. Each

column in rawA and rawB corresponds to a new nuclei. For example to assemble the

117

modulation function for two identical nuclei, rawA and rawB would each be two columns
wide where each column contains the same simulation output. Alternatively with two
different nuclei, column one would contain the output for nuclei 1 the second column for
nuclei 2.

After the modulation functions have been assembled, the DC component is
removed and the spectrum is hammed. This is followed by a fast Fourier transform which
will at minimum zero-fill the spectra to maintain the number of points, further zero-filling
can be added in expp. The output includes both the time domain spectrum and the

frequency domain spectrum.

hyschh.m — [raw] =hyschh(expp,hfip) — I=1/2 HYSCORE using an analytical solution
hyschhang.m — [raw] =hyschhang(expp,hfip,gp) -—I=1/2 angle selected HYSCORE using
an analytical solution
hyscho.m — [raw] =hyscho(expp, hfip) — I =1 HYSCORE using a summation form of a
density matrix calculation
hyschoang.m — [raw] =hyschoang(expp,hfip,gp) — I=1 angle selected HYSCORE using a
summation form of a density matrix calculation

Output from the simulation programs comes in a single matrix of two columns,
the first column is the alpha manifold, second is the beta manifold. This probably won't
properly handle S=5/2 since the specific weightings for the manifolds won't be taken into
account.

The I=1 simulation programs are horrendously slow and do not calculate the DC

118

 

terms, this is to help with the speed. This means that the I=1 simulations will not work

properly with the product rule.

hyscprocess.m -— [ timedom, fieqdom ]=hyscprocess(expp, raw) — Used to process the raw
HYSCORE spectrum. M11 apply the product rule in a fashion similar to that of
eseassemble.m, the only difference being the raw input comes in terms of two columns
per nuclei, so each two columns in raw correspond to a new nuclei.

The residual DC component is removed and the spectrum is harnmed in two
dimensions. This is followed by a 2D fast Fourier transform which will at minimum zero-
fill the spectra to maintain the number of points, further zero-filling can be added in expp.

The output includes both the time domain spectrum and the frequency domain spectrum.

hyscunpackm — [ Z, fiqu ] =hyscunpack(spectra) — This routine will take the HYSCORE
output generated in Xepr, which consists of the data in a column format and converts it
into a matrix for ease of plotting in Matlab.

simtobrukm — spec =simtobruk(spectrum,frequencyscale) — This routine will convert the
matrix representation of the HYSCORE used in Matlab to a format consistent with the

Bruker output.

endorangm — rawspec =endorang(expp,hfip,gp) — A generic angle selected ENDOR
routine that will calculate the frequency spectrum.

endorpowder.m — rawspec=endorang(expp,hfip) — A generic powder ENDOR routine

119

 

that will calculate the frequency spectrum.
The two ENDOR simulations are generic in the sense that they will calculate the
frequency spectrum irregardless of the nuclear spin. At present however it is only
equipped to calculate the AI=1 and AI=2 transitions if they are present. We have not been
faced with higher order transitions so those have been neglected in the code.

The output from these simulation routines consists of two columns, the first is the
frequency histogram for the alpha spin manifold, the second for the beta spin manifold. If
faced with S25/2 the output will consist of a column for each spin manifold much like the

output from the HYSCORE simulations. To properly utilize this output it needs to be sent

 

through engausconv.m.

engausconv.m — spec=engausconv(rawspec,expp) — This routine will assembles the raw
ENDOR frequency histograms from the simulations. The spectrum is further convoluted
with a Gaussian to remove any “fence-like” appearance due to coarse integrations and too
few points. The output consists of two columns, the frequency scale and the normalized

intensity.

The following scripts are used within the simulation programs themselves to perform
repeated functions.

eulerrotationm — [rot-hfi,rot-eeqq]=eulerrotation(hfip) — This routine will take the
nuclear hyperfine parameters stored in hfip and generate hyperfine and quadrupole

tensors in the principle axis system.

120

 

 

 

nucval.m -— [3, g, Ix, 1y, [2] =nucval(a) — Used to store and recal the nuclear spin, g values
and pauli spin matrices for a handful of commonly used nuclei.

pareng.m — [ theta, ncross ] =pareng(freq, dth, phi,mi,/ield, g1 , g2, g3,a1 ,a2, a3) — This search
routine is called by the simulation scripts when angle selection is needed. This is operated
by taking a phi value and cycling through theta, 0 ——> 1r/2 until the frequency is at
resonance then it will cycle from n/2 ——> 0 the determine if there is a second theta which
satisfies the resonance condition. Outputting the theta values in theta and the number of
crossing points in ncross.

eseft.m — FT =esefl(raw) — This routine will take partially processed spectra output from
Xepr and complete the processing. Ideally the raw spectrum with have the background
removed. The script will further remove any residual DC component, apply a hamming

window and Fourier transform the time time domain, maintaining the number of points.

loadfiles.m - loadfiles — Useful command line script to import into Matlab text files that
may contain a header.

loadcw.m — loadcw — Useful command line script to import into Matlab, remove the
header and normalize CW spectra exported from Xepr.

loadtd.m - loadtd — Useful command line script which imports into Matlab, removes the
header, and runs through eseft.m, the time domain spectra exported from Xepr

gval.m — [gx, gy, gz]=gval(g1, g2, g3, E/D) — This routine will calculate effective g values

in a S=3/2 Fe-NO system.

121

 

Notes on simulation structure

All of the simulations are structured in a similar fashion, within an orientation
integration loop the Hamiltonian matrix is diagonalized and the spectrum is calculated.
The orientation integration 100p is a numerical integration performed over theta and phi
using simpsons rule. It is within this loop the that Hamiltonian matrix for the specific
orientations is assembled and diagonalized. The Eigenvalues and Eigenvectors are then :-
used either in an analytical formula or a matrix expression derived through the use of the
density matrix formalism of Mims, these modulation fimctions are then used to calculate
the modulation depth over time.
Future Work

The simulation scripts work and with very low instance of error produce spectra
which properly reflect the input Hamiltonian parameters. This package still requires more
work to be considered complete. Following is a list of work that will be necessary to
reach that goal:

0 A single format for the raw data output for all the different simulations.

0 Re-implementation of S=5/2 calculations using the ideas presented by
Astashkin and Raitsimring 2002 for use with Mn2+.

° Optimize and increase the speed of the simulations. The HYSCORE
simulations for I=1 are a specific example of one program in desperate need of
optimization.

0 Expand the number of different simulation to include various other pulse

schemes such as 2-Pulse ESEEM, 5-Pulse ESEEM, 6-Pulse HYSCORE, Mims-

122

 

 

ENDOR, remims-ENDOR , etc.

 

 

 

123

 

APPENDIX 11
ELECTRON SPIN ECHO ENVELOPE MODULATION SPECTROSCOPY

This appendix is to provide a brief description of the pulsed EPR experimental
methods used in this thesis. These pulsed methods help to resolve the electron-nuclear
hyperfine couplings that are masked by inhomogeneous broadening in the CW-EPR
spectra. First, a semi-classical description of the 2-Pulse ESEEM experiment for a S=1/2
I=1/2 spin system, loosely based on the treatments by McCrackenl and/or Schweigerz,
will be presented. Following this will be a basic description of the 3-Pulse ESEEM and
the 4-Pulse HYSCORE experiments.

In a magnetic field the interaction between an S = 1/2 electron with an I = 1/2
nucleus can be represented by the Hamiltonian:

HzgefleHo‘siz-ganHofz't‘Siz'A'j [AZ-1]
For this treatment of the Hamiltonian, unlike the one in chapter 2, we assume that the g-
matrix is isotropic and the hyperfine interaction is axially symmetric. The diagonalization
of Hamiltonian A2.1 can be performed within the individual electron spin manifolds,
independent of one another, to yield the eigenvalues and eigenvectors for the NMR
transitions. This is possible because i is the only spin operator within the Hamiltonian
that contains off diagonal elements, these off diagonal terms mix the nuclear states but
have no effect on the electron states. The resulting NMR transitions within the a and B

electron spin manifolds are given by:

A 2 B 2]“2

war: 3) +( 2 [A2.2a]

 

(wN+

124

 

I:~

£1.41 1.!

 

rmmr A I'-
‘L—, fl

 

 

 

A 2 B 2 l/2
wB=[(wN——2-) +(—2) ] [A2.2b]

- _ 2
where (01" —gan Holh, the nuclear Larmour frequency; A-aiso+T(3°°S 9‘1)

and B :3 T31“ QCOSB’ with aiso and T being the isotropic and the dipolar hyperfine
couplings and 9 is the angle between the electron-nuclear vector and the magnetic field

H0. The two electron spin manifolds are separated by (”5:3 3 e HO/h’ the electron

Zeeman splitting. The resulting four level system is shown in figure A2.la:

 

 

 

 

 

 

 

 

 

 

 

 

 

a |1> b
w
0
V A K 12>
-- IUI M
935 wi-
w-
Lib- |3> ¢ —+
(”32 (”31 (1)8 (”42 (1)41
“’5 Frequency
I V L I4>

Figure A2.1: Energy level diagram(a) and the corresponding EPR spectrum(b) for an
S=1/2 I=1/2 spin system. '

The off diagonal elements introduced into the Hamiltonian by the nuclear spin operator,
A , mix the nuclear sub-levels within each electron spin manifold resulting in rnI no

longer being a good quantum number. The four energy levels are therefore labelled by

their Eigenvector, |n> where n=1-4. The normalized probability amplitude for the EPR

transitions is then given by lul and M. where:

125

 

lul=<2|§xl3>/(0.5g8H1)=sin{(<pa—<pB)/2} [A2.3a]

|v|=<1|s‘x|3>/(o.5gBHl)=cosI(cpa—cp

t )/2} [A2.3b]

B

In which sm (pa: BIZ wa and sm (pa: BIZ wB' Shown in figure A2.lb is the EPR
spectrum corresponding to the transitions seen in the energy level diagram, where (0+ =

wa+wfiand w+=wa—wB.

 

 

 

 

 

 

 

 

 

 

 

 

7!
'2' 7:
a
_. ,
‘ ‘/\
l I E
2 2 2
b
<————><—————>
. T ‘/\
.75 35 1
2 2 ” 2
C

 

 

 

 

 

 

 

Figure A2.2: EPR pulse sequences for (a) 2-pulse ESEEM, (b) 3-
pulse ESEEM, and (c) HYSCORE

The two pulse ESEEM experiment, shown in figure A2.2a, consists of a n/2 pulse
and 71: pulse separated by the time increment 1:. The intensity of the echo at time 1: after the
R pulse is recorded as a function of incrementally increasing the pulse spacing I. The
effect that the incremental spacing of the pulses has on the echo intensity can be
understood by combining the quantum mechanical description of a four level spin

system, given above, with a classical model for echo formation.

126

 

 

 

 

Prior to the application of the microwave pulses, the bulk magnetization of the
sample, M, is aligned with the laboratory magnetic field, H0, which is directed along the
z axis. In order to simplify the problem, the coordinate system is taken to rotate about the
z axis at an angular frequency of (05. The microwave pulses, H], are directed
perpendicular to the z axis, in the direction of +y in the rotating frame. Application of the
first 1r/2 pulse torques the magnetization into the plane perpendicular to the lab axis,
along +x; in addition, provided the pulse has the necessary bandwidth, it excites all four
transitions seen in fig. A2.la. The result of these transitions splits the magnetization into
four spin packets with the angular frequency of the transition. During the free precession
time 1, these spin packets accumulate a phase angle with respect to the x axis. This can be
seen in figure A2.lb, in a frame rotating with a frequency of (us, two will precess faster
moving ahead of the rotating frame and two will fall behind.

The second pulse torques the magnetization vectors 180° about the y-axis and
leads to the formation of an electron spin echo along -x, at time t after the 1: pulse. Again,
the pulse excites all four transitions seen in fig. A2.la and the spin packets precess with
the fiequency of the transition. The spins packets that proceed though the same transition
refocus along -x at time t producing an echo. There is however a finite probability that a
portion of each spin packet branches, following a different transition, and precesses at a
different angular frequency. At time t after the 7: pulse, the packets which underwent
branching will not have refocused along the -x axis with the formation of the echo. They
will instead have accumulated a phase at two different frequencies during the free

precession periods and will contribute to the echo based on their projection onto the x-

127

 

axis and weighted by the probability for the branching.
This description of the two pulse ESEEM experiment can be more completely
understood on examination of the analytical expression for the echo modulation of an S =

1/2, I = 1/2 spin system:

Emod(T):|u|4+|v|4+|u|2|v|2[2cos w“ T+ 2cos 003 T
[A2.4]

—cos(wa—w )T-cos(wa+wB)T]

B

This equation shows that the echo is composed of the sum of the transition probabilities
of the spin packets which proceed through the same transition with each pulse, |u|4 and
M4 and the transition probability that the spin packets branch, lulzlvlz. It can also be seen
that the contribution to the echo amplitude by the branching transitions is modulated in
time by the hyperfine frequencies and their sums and differences. Substituting A2.3a-b

into A2.4 yields:

E

m0d('r)=l—§[2—200swaT—2005w 'r

B [A2.5]

+cos(wa—w T+cos(wa+wB)-r]

3)
Where the depth of the modulation is given at:
k=[w1 B/(waw8)]2 [A2.6]
Two pulse ESEEM, while useful to describe the basic principles of these
techniques has proven to not be nearly as useful in the {F eNO}7 systems. This is due to
rather fast relaxation(T2*) of the spins, in practice the transverse magnetization is lost
within approximately 200 ns. The principle l-dimensional technique used in this thesis,

shown in figure A2.2b, is three-pulse ESEEM which can be described with the following

equation:

128

 

 

E (T,T)=l—%[l—COS(DBT][1—C08wa(T+T)]

mod

[A2.7]

—-§[1 ~coswaTHI—cosw (T+T)]

B

The three-pulse ESEEM experiment consists of a preparation period with two n/2 pulses
separated be a fixed time period T. The reason for this two pulse preparation is to allow
for a longitudinal evolution period, making the time dependence of the echo relatively
independent of the relaxation effects from T2*- This preparation period however adds a
secondary time dependence into the modulation function, due to the fixed nature of this
time it can have a negative effect on the observed spectrum by contributing blind spots.
Specific frequencies of (on will suppress the modulation of the related 03B, or vice versa.
In this work we exploited this blind spot behavior to suppress the strongly modulating
hydrogen matrix line which allowed for the higher resolution of weakly modulating
nuclei.

The other method used extensively in this thesis was HYSCORE spectroscopy.
The basic pulse scheme for this technique is shown in fig. A2.2c. Similar to three-pulse
ESEEM it begins with preparation period of two rt/2 pulses separated be a fixed time
period 1:; this style of preparation is performed for the same reasons, removal of T2* time
dependence. HYSCORE differs from three-pulse ESEEM by separating the single
evolution period into two by a mixing 1: pulse, this allows for the collection of two
separate time evolution periods. The resulting two-dimensional data set contains the
frequency correlation between the a and [3 electron spin manifolds. The portion of the

modulation function due exclusively to these frequency correlations is given as:

129

 

 

2
t2+watl+6+)+s cos(w t —watl—6_)

k
S(tl,t2):—A:C czcos(w B 2

B [A2.8]

 

 

 

 

+c2cos(wmt2+wBtl +6+)+szcos(watz—w3tl —6_)
Where the amplitude factor Cz_2 srn war/25m “)8 Tu and the phase factors
6+:(w0‘ HUB) Tu and 6-:(w“_wB)T/2 contains the t dependence. The actual
amplitudes of the modulations are equal to:
2 2
w —(w +w ) /4
s2: 1 0‘ B [A2.9a]
wa w,3
2 2
w —(w -w ) /4
cz: l 0‘ B [A2.9b]
wa wB
k=4czs2 [A2.9c]

This portion of the modulation function is useful for understanding the correlations;
however, when dealing with more than a single nuclei the full modulation function which

includes the DC components needs to be utilized.

130

 

 

REFERENCES

1. Poole, C. P.; Farach, H. A., Handbook of electron spin resonance .' data sources,

computer technology, relaxation, and ENDOR. American Institute of Physics: New York,
1 994.

2. Hanson, G., Metals in biology. Springer: New York, 2009.

131

MICHIGAN STATE UNIVERSITY

LIBRARIES

3 1293 03220 8724

 

q——v.——