ANALYSIS OF WATER LIGANDS WITHIN THE ALLOSTERIC FORMS OF
PHENYLALANINE HYDROXYLASE
By
Michael Howart

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Chemistry – Doctor of Philosophy
2013

ABSTRACT
ANALYSIS OF WATER LIGANDS WITHIN THE ALLOSTERIC FORMS OF
PHENYLALANINE HYDROXYLASE
By
Michael Howart
Phenylalanine Hydroxylase (PheH) is a liver enzyme that catalyzes the conversion of Lphenylalnine to L-tyrosine using tetrahydrobiopterin and molecular oxygen. Deficiencies in this
enzyme cause phenylketonuria (PKU), an autosomal recessive metabolic disorder that occurs in
~1/10,000 live births. The general mechanism of PheH consists of two major steps: (1) formation
of the ferryl-oxo species and (2) hydroxylation of the aromatic amino acid L-phenylalanine.
Before the coupled hydoxylation can occur, PheH is activated by the presence of L-phe from the
T-(unactivated) to R-(activated)-state. Here activation is marked by a distinct increase in the
specific activity, which corresponds to a 2.2 and 8-fold increase in the specific activity for
hPheH and rPheH, respectively. Accompanying the activation of PheH is a global
conformational change, which results in a ligand rearrangement of the active site. Mixed signals
about the presence of water in the literature have lead to multiple accounts concerning the role
water plays in the formation of the ferryl-oxo species.
This work provided the basis for studying water ligands in pterin dependent non-heme
7

1

iron enzymes with an [FeNO] spin system using H ESEEM and HYSCORE. Based on
1

7

7

qualitative argument with the H ESEEM spectra from the [FeNO] -(N2O)(H2O)2 and [FeNO] (N2O2)(H2O)1 model complexes, it was determined that the activation of PheH, without BH4

complexed, lead to a one water dehydration of PheH’s active site. The water ligands were then
quantitatively analyzed to show characteristic dipolar distances (from the iron to the water
protons) of 2.52 – 2.62 Å and β angles between 58° - 83° and 97° - 122°. The observed dipolar
distances and angles are consitent with water ligands that are located cis to the binding site of
1

NO. Overall, this analysis demonstrated the ability of using H ESEEM and HYSCORE to study
water ligands bound to non-heme iron enzymes.

DEDICATION

This work is dedicated to my better half
!
!

!

iv!

ACKNOWLEDGEMENTS

I would like to first acknowledge my advisor Dr. John McCracken for all of his guidance
and patience. Whether it was explaining the ESEEM experiment or the history of Freon, our
conversations were always informative to me and I thank you for them.
I thank my lab mates, Tom Casey and Matt Krzyaniak, for their helpful insight and
knowledge of the EPR experiments and analysis. Without their help I would have had a thousand
more questions to interrupt Dr. McCracken’s day.
I’d also like to acknowledge my undergraduate advisors at the University of Michigan-Flint, Dr. Jessica Tischler and Dr. Jie Song, who helped prepare me for this adventure.
Of course a big debt of gratitude goes to my entire family for their continual support and
unrelenting ability to ask ‘have you graduated yet?’ Soon I’ll be responding to this question with
a resounding ‘YES!’
Last but not least, I would like to say thanks to my better half, Kristie Howart, for all of
her support and encouragement. Although I know I was absolutely delightful to be around during
the production of this work, I would still like to tell my new wife thank you and I love you.

!

v!

TABLE OF CONTENTS
LIST OF TABLES ....................................................................................................................... viii
LIST OF FIGURES .........................................................................................................................x
KEY TO SYMBOLS AND ABBREVIATIONS ...................................................................... xviii
Chapter 1: Introduction ...................................................................................................................1
Section 1.1: Biochemical Reactions of Several Pterin-Dependent Hydroxylases ...............1
Section 1.2: General Mechanism of Pterin-Dependent Hydroxylases.................................2
Section 1.3: Structural Features of PheH .............................................................................9
Section 1.4: Regulation of PheH........................................................................................21
Section 1.5: The Focus.......................................................................................................27
BIBLIOGRAPHY ..............................................................................................................29
Chapter 2: Electron Paramagnetic Resonance Experiments .........................................................37
Section 2.1: A Historical Perspective ................................................................................37
Section 2.2: The Spin System ............................................................................................37
Section 2.3: Continuous Wave – Electron Paramagnetic Resonance ................................42
Section 2.4: Electron Spin Echo Envelope Modulation.....................................................56
Section 2.5: Hyperfine Sublevel Correlation Experiment .................................................74
BIBLIOGRAPHY ..............................................................................................................81
Chapter 3: Qualitative Analysis of the Allosteric Forms of Phenylalanine Hydroxylase .............86
Section 3.1: Introduction....................................................................................................86
Section 3.2: Materials and Methods...................................................................................87
Section 3.3: Results and Analysis ......................................................................................88
Section 3.3.1: Model Complexes ...........................................................................88
T
R
Section 3.3.2: PheH [L-phe] vs PheH [L-phe] ....................................................97
R
Section 3.3.3: PheH [L-phe,5d-6MPH4] vs
R
2
PheH [L-phe,5- H-5d-6MPH1D3] ........................................................104
Section 3.4 Discussion .....................................................................................................106
APPENDICES ................................................................................................................110
7
Appendix 1: Preparation of [FeNO] Samples ...................................................111
Appendix 2: MATLAB Scripts ...........................................................................115
BIBLIOGRAPHY ............................................................................................................127

vi

Chapter 4: Quantitative HYSCORE Analysis of the Allosteric forms of Phenylalanine
Hydroxylase ................................................................................................................................131
Section 4.1: Introduction..................................................................................................131
Section 4.2: Materials and Methods.................................................................................132
Section 4.3: Results and Analysis ....................................................................................133
7
T
Section 4.3.1: CW-EPR Analysis of [FeNO] -PheH [L-phe] and
7
R
[FeNO] -PheH [L-phe]..............................................................................133
7
T
Section 4.3.2: HYSCORE Data of [FeNO] -PheH [L-phe] and
7
R
[FeNO] -PheH [L-phe]..............................................................................135
7
T
Section 4.3.3: Assignment of P1 and P2 in [FeNO] -PheH [L-phe] and
7
R
[FeNO] -PheH [L-phe]..............................................................................141
7
T
Section 4.3.4: HYSCORE Analysis of [FeNO] -PheH [L-phe] .........................144
7
R
Section 4.3.5: HYSCORE Analysis of [FeNO] -PheH [L-phe] ........................150
Section 4.3.6: CW-EPR Analysis of the Model Complexes ...............................157
7
Section 4.3.7: HYSCORE Analysis of [FeNO] -(N2O2)(H2O)1 ........................159
7
Section 4.3.8: HYSCORE Analysis of [FeNO] -(N2O1)(H2O)2 ........................164
Section 4.3.9: Visibility of the High-Frequency Correlation Peaks ....................168
Section 4.4: Discussion ....................................................................................................170
BIBLIOGRAPHY ............................................................................................................180

vii

LIST OF TABLES

Table 2.1: The 4-step phase cycling program used to eliminate the unwanted echoes in 3pulse ESEEM ................................................................................................................... 73
Table 2.2: To remove the unwanted echoes from the experiment this phase cycling
program was utilized. ....................................................................................................... 74
Table 4.1: The representative E/D values for the unactivated, T-State, and activated, Rstate, enzyme samples .................................................................................................... 134
Table 4.2: The frequency positions at each magnetic field strength for the three points
defined in figure 4.3 ....................................................................................................... 137
7

T

Table 4.3: Just as with the unactivated form of PheH, [FeNO] -PheH [L-phe], the
7
R
1
activated form, [FeNO] -PheH [L-phe], has three distinct sets of H correlations within
the HYSCORE spectra that are defined here as P1, P2, and P3 figure 4.5 and listed in
table 4.3 .......................................................................................................................... 139
Table 4.4: Shows the optimum values that were obtained from the fit ......................... 150
Table 4.5: The optimum parameters for the two frequency positions analyzed are listed
........................................................................................................................................ 151
Table 4.6: A summary of the CW-EPR results from the model complexes can be found in
table 4.6 .......................................................................................................................... 157
Table 4.7: The frequency position of these points can be seen in figure 4.23 and are listed
in table 4.7...................................................................................................................... 161
Table 4.8: Fitting the frequency positions P1, P2, and P3 in the HYSCORE spectra of
7
[FeNO] -(N2O2)(H2O)1 collected at 240 mT, 260 mT, and 280 mT resulted in the
optimized parameters with their standard deviations listed in table 4.8 ........................ 161
7

Table 4.9: The frequency positions of for the three points of [FeNO] -(N2O)(H2O)2 at
each magnetic field ........................................................................................................ 165
Table 4.10: The optimized parameters and their corresponding standard deviations are
listed ............................................................................................................................... 165
!

viii!

Table 4.11: The dipolar distances and β angles listed in table 4.11 were determined for
7
7
7
T
[FeNO] -(N2O2)(H2O)1, [FeNO] -(N2O)(H2O)2, [FeNO] -PheH [L-phe], and
7
R
[FeNO] -PheH [L-phe] with a frequency based HYSCORE analysis ......................... 172

!

ix!

LIST OF FIGURES

Figure 1.1: The specific reactions catalyzed by three different pterin-dependent aromatic
amino acid hydroxylase enzymes ...................................................................................... 1
Figure 1.2: The homology of their catalytic domains suggests that all three enzymes use a
common chemical mechanism ........................................................................................... 3
Figure 1.3: The interconversion between the three pterin forms previously described ..... 5
Figure 1.4: X-ray crystallography studies done using mammalian PheH containing the
catalytically relevant ferrous iron, PDB 1J8T, without BH4 ........................................... 10
Figure 1.5: X-ray crystallography studies done using mammalian PheH containing the
catalytically relevant ferrous iron, PDB 1J8U, with BH4 ................................................ 11
T

R

Figure 1.6: Fe K edge XAS spectra for rat PheH [] and PheH [L-phe] along with
comparative model complexes ......................................................................................... 12
Figure 1.7: The geometry change resembles a distorted octahedron transforming into a
distorted square pyramidal structure ................................................................................ 15
Figure 1.8: The strength of the transitions increase as the overlap between the ligand and
metal orbitals increase...................................................................................................... 16
Figure 1.9: A five-coordinate structure was also observed in a 2.5Å crystal structure of
truncated PheH that contained both BH4 and a substrate analogue, 3-(2-Thienyl)-Lalanine (PDB 1KW0) ....................................................................................................... 17
2

Figure 1.10: Refinement of the crystal structure revealed an even higher B-factor, 53Å ,
which was on the same order as the noise level and is the reason it was omitted from the
crystal structure ................................................................................................................ 18
Figure 1.11: In solution, mammalian PheH forms a tetrameric complex of monomer
peptides that have a molar mass of ~50 kDa .................................................................. 21
Figure 1.12: The general mechanism for the regulation of PheH reported in the literature
.......................................................................................................................................... 22

!

x!

Figure 1.13: This global movement can be seen by comparing the ‘unactivated’ and
‘activated’ crystal structures, which corresponds to 1J8U and 1MMK ........................... 25
Figure 1.14: This global movement can be seen by comparing the ‘unactivated’ and
‘activated’ crystal structures, which corresponds to 1J8U and 1MMK ........................... 26
2-

-

Figure 2.1: The Fe K edge XAS spectra of [Fe(H2O)EDTA] , [Fe(H2O)EDTA] , and
FeEDTA-NO .................................................................................................................... 39
Figure 2.2: The molecular orbital diagram of the spin system as predicted by the DFT
calculations ...................................................................................................................... 41
Figure 2.3: The relative strength of these individual interactions described by the spin
Hamiltonian in equation 2.1 ............................................................................................. 43
Figure 2.4: Exposure to an external magnetic field results in an energy separation
between the spin states that is described by the electronic Zeeman interaction .............. 44
Figure 2.5: CW-EPR spectrum with a temperature of 4 K and microwave frequency of
7
9.68 GHz for the [FeNO] spin system ........................................................................... 49
Figure 2.6: With the ZFS interaction defining the coordinate system, the external
magnetic field is now thought of as having multiple orientations ................................... 50
Figure 2.7: The sign of D determines which set of eigenvectors values make up the
ground state ...................................................................................................................... 52
Figure 2.8: As the rhombicity of the ZFS becomes non-zero, E≠0, the spectrum splits
near the g = 4 feature ....................................................................................................... 55
Figure 2.9: If we assume that β = 70° and T = 4.5 MHz, the powder pattern created by the
orientation dependence of the external magnetic field would result in the powder pattern
observed in the spectrum ................................................................................................. 62
Figure 2.10: The inhomogeneous broadening of the individual hyperfine frequencies
equation 2.27 and 2.28 can be seen in figure ................................................................... 63
Figure 2.11: The 2-pulse ESEEM experiment has a pulse train that consists of a π/2-pulse
followed by a π-pulse with a time separation of τ ........................................................... 64

!

xi!

Figure 2.12: The magnetic field of the resonant microwave pulses, B1, rotates the net


magnetization vector, M , through an angle θ................................................................. 64
Figure 2.13: Lets now analyze the formation of the spin echo with this vector approach ..
.......................................................................................................................................... 66
Figure 2.14: The mixing of the nuclear states in each manifold via the hyperfine
interaction produces a probability (equation 2.5) for the transition with the amplitude, |v|,
that are no longer equal to zero ........................................................................................ 67
Figure 2.15: To illustrate the echo modulation we can track a population of spin packets
during the course the experiment ..................................................................................... 68
Figure 2.16: Tracing the modulation pattern and taking the Fourier transform of this time
domain signal results in a frequency domain spectrum that displays the hyperfine
frequencies (ωα and ωβ) ................................................................................................... 69
Figure 2.17: By replacing the π-pulse with two π/2-pulses separated by an incremental
time value, T, the T1-dependent ESEEM experiment is formed ..................................... 71
Figure 2.18: Shows the unwanted echoes that are formed during the 3-pulse ESEEM
experiment........................................................................................................................ 72
Figure 2.19: To remove this congestion the experiment is converted into a twodimensional correlation experiment by splitting the incremental time period, T, into two
by inserting a π-pulse ....................................................................................................... 74
Figure 2.20: The two-dimensional 4-pulse HYSCORE experiment leads to a frequency
spectrum that correlates the hyperfine frequencies, ωα to ωβ, to one another ................. 75
Figure 2.21: The powder pattern results in a spectrum that has correlated peaks spanning
a range of frequencies up to 3T/2 .................................................................................... 76
Figure 2.22: The asymmetry raises the correlations off of the anti-diagonal in the
HYSCORE spectrum ....................................................................................................... 78
Figure 2.23: Shows the small population of spins that meet the resonance condition
(orange) and results only in a small section of the powder pattern signal being visible in
the HYSCORE spectrum ................................................................................................. 79

!

xii!

Figure 3.1: The ferrous model complexes studied in this work are illustrated ................ 88
Figure 3.2: Probing the N2O2 ferrous iron-nitrosyl complex with CW-EPR results in the
green spectrum presented ................................................................................................. 89
7

Figure 3.3: A 3-pulse ESEEM spectrum collected for the [FeNO] -(N2O2)(H2O) model
complex at 240mT ........................................................................................................... 90
Figure 3.4: The corresponding 4-pulse HYSCORE spectrum taken at the same magnetic
field strength and τ-value ................................................................................................. 92
1

7

1

7

Figure 3.5: Parallel H HYSCORE studies were done on [FeNO] -(N2O)(H2O)2 and
7
[FeNO] -N2O3 model complexes and HYSCORE spectra collected at 240 mT with τ =
96 ns are shown................................................................................................................ 93
Figure 3.6: Parallel H HYSCORE studies were done on [FeNO] -(N2O)(H2O)2 and
7
[FeNO] -N2O3 model complexes and HYSCORE spectra collected at 240 mT with τ =
96 ns are shown................................................................................................................ 94
Figure 3.7: Corresponding 3-pulse ESEEM spectra for the three model
complexes collected using the same magnetic field strength and τ-value as used for the
HYSCORE studies above ................................................................................................ 95
7

Figure 3.8: The g = 4 features of the CW-EPR spectra from the [FeNO] forms of
T
R
PheH [L-phe] and PheH [L-phe] ................................................................................... 97
Figure 3.9: Shows the 4-pulse HYSCORE spectra collected at 240 mT using a τ-value of
T
96 ns for PheH [L-phe] ................................................................................................... 98
Figure 3.10: Shows the 4-pulse HYSCORE spectra collected at 240 mT using a τ-value
R
of 96 ns for PheH [L-phe]............................................................................................... 99
Figure 3.11: The breadth of the cross-peak contours resolved for bound water ligands of
the model complexes encompasses both sets of cross-peaks observed in the HYSCORE
spectrum of the T-state PheH complex .......................................................................... 101
7

7

Figure 3.12: Shows the 3-pulse ESEEM spectra of [FeNO] -N2O and [FeNO] -N2O2
T
R
overlaid along with PheH [L-phe] and PheH [L-phe] ................................................. 103

!

xiii!

R

Figure 3.13: The 4-pulse HYSCORE spectra collected at 240mT using quaternary PheH
complexed to 5d-6MPH4 ............................................................................................... 104

R

Figure 3.14: The 4-pulse HYSCORE spectra collected at 240mT using quaternary PheH
2
complexed to 5- H-5d-6MPHD3 ................................................................................... 105
7

T

Figure 4.1: Shows the CW-EPR g = 4 feature of the [FeNO] -PheH [L-phe] and
7
R
[FeNO] -PheH [L-phe] samples in blue, while the simulations are represented by the
green spectra .................................................................................................................. 133
Figure 4.2: The HYSCORE spectra collected at magnetic field strengths of 240, 260, and
7
T
280 mT for [FeNO] -PheH [L-phe].............................................................................. 135
1

Figure 4.3: From the figure (4.2) three sets of correlations peaks associated to H can be
seen, which in this work will be define as P1, P2, and P3 as described by figure 4.3... 136
Figure 4.4: The HYSCORE spectra from the activated form of phenylalanine hydroxylase
collected at magnetic field strengths of 240, 260, and 280 mT are shown .................... 138
7

T

Figure 4.5: Just as with the unactivated form of PheH, [FeNO] -PheH [L-phe], the
7
R
1
activated form, [FeNO] -PheH [L-phe], has three distinct sets of H correlations within
the HYSCORE spectra that are defined here as P1, P2, and P3 figure 4.5 and listed in
table 4.3 .......................................................................................................................... 139
Figure 4.6: Comparing the relative intensity of P1 and P2 to that of P3 between the
allosteric forms reveals a decrease in the intensity of P1 and P2 during the activation of
PheH............................................................................................................................... 140
7

Figure 4.7: To accomplish this task three [FeNO] -(N2Ox)(H2O)3-x model complexes
were used ....................................................................................................................... 141
7

7

Figure 4.8: Starting with an inspection of [FeNO] -(N2O2)(H2O)1 and [FeNO] -(N2O3)
helps to demonstrate the result of completely dehydrating the first coordination sphere
........................................................................................................................................ 142
Figure 4.9: Although the frequency position of the high-frequency correlation (outlined in
7
7
red) between the [FeNO] -(N2O1)(H2O)2 and [FeNO] -(N2O2)(H2O)1 samples is
maintained, figure 4.9, the relative intensity of the high-frequency correlation pair to that
of the correlation pair center at (8 MHz, 13 MHz) and (13 MHz, 8 MHz) decreases by
about a factor of 2 .......................................................................................................... 143

!

xiv!

Figure 4.10: It can be seen that both high-frequency correlations (i.e. P1 and P2)
observed in the enzyme sample are positioned in such a fashion that they can be
1
understood to be the result of H couplings assigned to water ligands ......................... 144
Figure 4.11: A graphical representation of equation 4.1 can be seen in figure 4.11 ..... 146
7

T

7

T

7

T

Figure 4.12: After running the fitting routine for [FeNO] -PheH [L-phe] using the three
magnetic field strengths, 240, 260, and 280 mT, the optimum parameters were used to
simulate P1 and P2, figures 4.12, 4.13, and 4.14 ........................................................... 148
Figure 4.13: After running the fitting routine for [FeNO] -PheH [L-phe] using the three
magnetic field strengths, 240, 260, and 280 mT, the optimum parameters were used to
simulate P1 and P2, figures 4.12, 4.13, and 4.14 ........................................................... 148
Figure 4.14: After running the fitting routine for [FeNO] -PheH [L-phe] using the three
magnetic field strengths, 240, 260, and 280 mT, the optimum parameters were used to
simulate P1 and P2, figures 4.12, 4.13, and 4.14 ........................................................... 149
Figure 4.15: Using these parameters, the HYSCORE simulations at each magnetic field
strength are shown in figure 4.15, 4.16, and 4.17 .......................................................... 152
Figure 4.16: Using these parameters, the HYSCORE simulations at each magnetic field
strength are shown in figure 4.15, 4.16, and 4.17 .......................................................... 152
Figure 4.17: Using these parameters, the HYSCORE simulations at each magnetic field
strength are shown in figure 4.15, 4.16, and 4.17 .......................................................... 153
Figure 4.18: By extending the simulations to include two protons with the parameters
observed in the one proton fittings of the HYSCORE spectra, figure 4.18, 4.19, and 4.20,
reveals that this technique is fairly accurate .................................................................. 154
Figure 4.19: By extending the simulations to include two protons with the parameters
observed in the one proton fittings of the HYSCORE spectra, figure 4.18, 4.19, and 4.20,
reveals that this technique is fairly accurate .................................................................. 154
Figure 4.20: By extending the simulations to include two protons with the parameters
observed in the one proton fittings of the HYSCORE spectra, figure 4.18, 4.19, and 4.20,
reveals that this technique is fairly accurate .................................................................. 155

!

xv!

Figure 4.21: Looking at a stack plot of the 3-pulse ESEEM spectra of the T- and R-state,
it is easy to see that there is an amplitude difference between the allosteric states ....... 156
Figure 4.22: Shows the data (green) versus simulated spectra (blue) at the g = 4......... 158
7

Figure 4.23: The HYSCORE spectrum for [FeNO] -(N2O2)(H2O)1 collected with a
magnetic field strength of 240 mT and a temperature of 4 K ........................................ 160
Figure 4.24: Using the optimized parameters listed in table 4.8 produces the HYSCORE
simulations seen in figure 4.24, 4.25, and 4.26, which correspond to magnetic field
strengths of 240, 260, and 280 mT ................................................................................ 162
Figure 4.25: Using the optimized parameters listed in table 4.8 produces the HYSCORE
simulations seen in figure 4.24, 4.25, and 4.26, which correspond to magnetic field
strengths of 240, 260, and 280 mT ................................................................................ 163
Figure 4.26: Using the optimized parameters listed in table 4.8 produces the HYSCORE
simulations seen in figure 4.24, 4.25, and 4.26, which correspond to magnetic field
strengths of 240, 260, and 280 mT ................................................................................ 163
7

Figure 4.27: As was done previously, the upper left high-frequency peak of [FeNO] (N2O)(H2O)2 can be emulated using three points, figure 4.27, that outline the contour
shape of the high-frequency correlation ........................................................................ 164
Figure 4.28: The optimized parameters and their corresponding standard deviations are
listed in table 4.10, while the simulated spectra are shown along side of the data in figure
4.28, 4.29, and 4.30 ........................................................................................................ 166
Figure 4.29: The optimized parameters and their corresponding standard deviations are
listed in table 4.10, while the simulated spectra are shown along side of the data in figure
4.28, 4.29, and 4.30 ........................................................................................................ 167
Figure 4.30: The optimized parameters and their corresponding standard deviations are
listed in table 4.10, while the simulated spectra are shown along side of the data in figure
4.28, 4.29, and 4.30 ........................................................................................................ 167
Figure 4.31: As can be seen at 180 and 340 mT (figures 4.31 and 4.32, respectively), the
high-frequency peaks seemingly disappear at magnetic field strengths near the ends of the
CW-EPR spectrum ......................................................................................................... 168

!

xvi!

Figure 4.32: As can be seen at 180 and 340 mT (figures 4.31 and 4.32, respectively), the
high-frequency peaks seemingly disappear at magnetic field strengths near the ends of the
CW-EPR spectrum ......................................................................................................... 169
7

T

Figure 4.33: Using the information determined for the [FeNO] -PheH [L-Phe] system,
presented in table 4.11, and the knowledge that the Fe—NO bond forms nearly colinear
16
to the Dz axis, then the model shown in figure 4.33 can be built .............................. 176
Figure 4.34: Manipulating the protons to resemble the picture of water described by
theory results in the water ligand shown........................................................................ 177
Figure 4.35: By!placing!two!of!these!ligands!directly!into!the!coordination!spots!of!
water!1!and!water!3!from!1J8U,!the!active!site!takes!on!the!arrangement!as!observed!
in!the!figure!4.35 .......................................................................................................... 179

!

xvii!

KEY TO SYMBOLS AND ABBREVIATIONS

PheH

Phenylalanine Hydroxylase

TyrH

Tyrosine Hydroxylase

TrpH

Tryptophan Hydroxylase

BH4

(6R)-L-erythro-5,6,7,8-tetrahydrobiopterin

PKU

Phenylketonuria

q-6-MPH2
6-MPH4
4a-OH-6-MPH4

quinonoid 6-methyldihydropterin
6-methyltetrahydropterin
4a-hydroxy-6-methyltetrahydropterin

NMR

Nuclear Magnetic Resonance

ppm

part per million

CD

circular dichroism

DFT

density functional theory

EAS

electrophilic aromatic substitution

hPheH

Human Phenylalanine Hydroxylase

rPheH

Rat Phenylalanine Hydroxylase

Glu
MCD

!

Glutamate
Magnetic Circular Dichroism

xviii!

XAS

X-ray Absorption Spectroscopy

bPheH

Bacterial Phenylalanine Hydroxylase

DMPH4

6,7-dimethyl-5,6,7,8-tetrahydropterin

L-phe
kDa
cePheH

L-phenylalanine
1 kiloDalton = 1 g/mol
nematode Phenylalanine Hydroxylase

Tyr

Tyrosine

His

Histidine

Tih

3-(2-Thienyl)-L-alanine

NO

Nitric Oxide

CW-EPR

Continuous Wave – Electron Paramagnetic Resonance

XAS

X-ray Absorption Spectroscopy

ZFS

Zero Field Splitting

|S,ms>

Electron Spin Basis

|I,mI>

Nuclear Spin Basis

!
!

Nuclear Spin Operator

[Dx , Dy ,Dz]

!

Electron Spin Operator

ZFS Coordinate System

xix!

D

Secular ZFS Term

E

Rhombic ZFS Term

Bo

External Magnetic Field

B1

Applied Magnetic Field

ωo

Larmor Frequency

ω1

Rabi Frequency

h

Planck Constant

ge

Free Electron Spin g-factor (2.0023)

geff

Effective g-value

ms

Electron Magnetic Spin Quantum Number

mI

Nuclear Magnetic Spin Quantum Number

X-band

(8.2-12.4 GHz)

ESEEM

Electron Spin Echo Envelope Modulation

T

Dipolar Hyperfine Coupling

aiso

Isotropic Hyperfine Coupling

µo

Vacuum Permeability (4πE-7 N/A )

2

r
βe

!

Dipolar Distance
Electronic Bohr Magneton (9.274E-24 J/T)

xx!

βn

Nuclear Bohr Magneton (5.051E-27 J/T)

gn

Nuclear g-value (5.586 for H)

νI

Nuclear Frequency (14.9 MHz at 350 mT)

ωI

Angular Nuclear Frequency

k

Modulation Depth

HYSCORE

1

Hyperfine Sublevel Correlation Experiment

N2 O1
N2 O2

2,2′-(ethane-1,2-diylbis (methyl-azanediyl))diacetic acid

N2 O3

!

2-((2-(dimethyl-amino)ethyl)(methyl)amino)acetic acid

2,2′-(2-((carboxymethyl)-(methyl)amino) ethyl-azanediyl)diacetic acid

xxi!

Chapter 1: Introduction
Section 1.1: Biochemical Reactions of Several Pterin-Dependent Hydroxylases
The hydroxylation of aromatic amino acids, using molecular oxygen and the cofactor
(6R)-L-erythro-5,6,7,8-tetrahydrobiopterin (BH4), is catalyzed by a class of non-heme iron

Figure 1.1: Above is three reactions carried out by pterin-dependent aromatic amino acid
hydroxylases (AAH) using molecular oxygen and (6R)-L-erythro-5,6,7,8-tetrahydrobiopterin
(BH4). (a) Phenylalanine Hydroxylase (PAH) converts L-phenylalanine to L-tyrosine which is
then hydroxylated by (b) Tyrosine Hydoxylase (TyrH) to form L-3,4-dihydroxyphenylalanine
(L-DOPA). (c) Tryptophan Hydroxylase (TrpH) catalyzes the formation of 5-hydroxytryptophan from tryptophan.
!

1!

enzyme that includes Phenylalanine Hydroxylase (PheH), Tyrosine Hydroxylase (TyrH), and
Tryptophan Hydroxylase (TrpH). The specific reactions catalyzed by these tetrameteric enzymes
are shown in figure 1.1. PheH, figure 1.1a, is a liver enzyme that catalyzes the hydroxylation of
L-phenylalanine to produce L-tyrosine and has been linked to the autosomal recessive metabolic
disorder Phenylketonuria (PKU).

1-4

Tyrosine Hydroxylase (TyrH), figure 1.1b, catalyzes the

conversion of L-tyrosine to L-3,4-dihydroxyphenylalanine (L-DOPA). The formation of LDOPA is the rate-limiting step in the biosynthesis of the neurotransmitters dopamine, adrenaline,
5

and noradrenaline. TyrH is found in the central nervous system and has been associated with
various neurological disorders. Also in this class of enzymes is Tryptophan Hydroxylase (TrpH),
figure 1.1c, which catalyzes the hydroxylation of tryptophan to 5-hydroxy-tryptophan and is the
rate-limiting step in the production of the neurotransmitter serotonin.

1-4

Section 1.2: General Mechanism of Pterin-Dependent Hydroxylases
Comparison of the truncated forms of PheH, TyrH, and TrpH, expressed in rat, shows
1

that the sequence of the functional domains are essentially identical. In the tertiary forms of
these truncated enzymes, the pterin-dependent hydroxylases all have a ferrous iron in their
catalytic domain(s) that is facially ligated by the side chains of two histidine residues and a
carboxylate from the side chain of glutamic acid. The homology of their catalytic domains
6

suggests that all three enzymes use a common chemical mechanism. This mechanism, depicted
in figure 1.2, is initiated by the binding of the substrate amino acid and cofactor, BH4, to the

!

2!

Figure 1.2: Above is the general mechanism of phenylalanine hydroxylase. After the
formation of the (a) quaternary complex it is thought that the (b) molecular oxygen
forms a bridge between the aliphatic carbon of tetrahydrobiopterin and ferrous iron. (c)
Eventually, molecular oxygen is heterolytically cleaved resulting in the formation of a
4a-hydoxypterin and a high-valent Fe(IV)-oxo species. (d) The superoxide then
electrophillically attacks the aromatic amino acid eventually carrying out an
electrophillic aromatic substitution, EAS, resulting in the hydroxylated amino acid
product.

!

3!

Fe(II) site. When the substrate binds it causes a change in the conformation of the enzyme

7, 8

that facilitates O2 binding to Fe(II) and initiates a two-step coupled hydroxylation of BH4 and
2, 3, 6, 9

substrate

12

of the O-O bond
intermediate.
12

substrate

10,11

. The hydroxylation of BH4 occurs first

10, 13

and involves heterolytic cleavage

leading to the formation of 4a-hydroxypterin and an Fe(IV)=O
The ferryl-oxo species in turn promotes hydroxylation of the amino acid

via electrophilic aromatic substitution

1, 5,14

.

The current understanding of the coupled hydroxylation mechanism is supported by
several key spectroscopic observations. The first of which comes from mass spectrometry
measurements preformed by Kaufman et al. Utilizing isotopically labeled molecular oxygen,
18

O2, and water, H2O

18

, as variables in the conversion of L-phenylalanine to L-tyrosine with rat

PheH, a series of experiments were run to determine the source of the hydroxyl oxygen in the
amino acid product. A reaction carried out in an
resulted in the inclusion of

18

18

O2 enriched atmosphere, ~50% enrichment,

O in 59% of the amino acid product, tyrosine. Decarboxylation of

the aromatic amino acid revealed that the

18

O was located on the phenyl group and not on the
18

carboxylic group. When the reaction was carried out with enriched water, H2O , the presence of
isotopically labeled oxygen,

18

O, in tyrosine was negligible. These studies showed that water

does not play a role in the hydroxylation of the amino acid product and that molecular oxygen
was the source of the hydroxyl-oxygen on the L-tyrosine of rat PheH.

12

The second oxygen was initially thought to be converted into water and a reduced form
15

of the pterin, quinonoid dihydrobiopterin (q-BH2).

!

4!

Lazurus et al. used

13

C NMR to trace the

11

response of the intermediates over the course of 580 min at -30°C.

Labeling the C4a position

of quinonoid 6-methyldihydropterin (q-6-MPH2), 6-methyltetrahydropterin (6-MPH4) and 4aOH adduct (4a-OH-6-MPH3) with

13

C produced characteristic chemical shifts at 148.0, 99.0 and

Figure 1.3: Here the interconversion of 6-methyltetrahydropterin (6-MPH4). (a) 6MPH4 is hydroxylated at the 4a position to form (b) 4a-OH-6-MPH4 that is
eventually dehydrated in solution to form (c) q-6-MPH2. With the addition of heat
the quinonoid species is rehydrated to form the 6-MPH4.
72.2 ppm, respectively. Before the addition of PheH the

13

C NMR spectrum revealed the

existence of only the pterin cofactor, 6-MPH4. After the PheH was mixed, approximately 20
min, the spectrum showed signs for the production of the 4a-OH-6-MPH4 and to a lesser extent
the q-6-MPH2. Eventually, the q-6-MPH2 became the only species present in the solution, t =
360 min. Once the solution was warmed to 30°C the q-6-MPH2 began the transformation back

!

5!

to 6-MPH4. The interconversion between the three pterin forms previously described can be seen
in figure 1.3. This data is the first observation of the 4a-hydroxypterin intermediate and provided
an explanation for the appearance/disappearance of the reduced pterin species, q-6-MPH2. Dix et
al. confirmed that PheH produces this 4a-hydroxypterin species and used isotopic shits from
18

O, near 73.3 ppm, in the

13

C NMR spectrum to prove that molecular oxygen is the source of

the 4a-hydroxyl oxygen on the pterin species.
enriched atmosphere of

18

10

The experiment was carried out in a 50%

O2 with a resulting 0.023 ppm up-field shift. Additionally, they were

able to use circular dichroism (CD) spectroscopy to determine the chirality at the C4a position of
the hydroxylated pterin generated with PheH, 4a-OH-6-MPH4. Comparison of the CD spectrum
to that of 4a(S)-hydroxy-5-deaza-6-methyltetrahydropterin showed similarities that confirmed
the 4a-hydroxy group binds stereospecifically with an S-type configuration.
Although the source of the hydroxyl-oxygen on the aromatic amino acid product was
identified earlier, the current model suggests the hydroxylation of the pterin species occurs
6

before the hydroxylation of the aromatic substrate. The result of the pterin hydroxylation is the
formation of a high-valent Fe(IV)-oxo species. A similar ferryl-oxo species has been shown to
exist in Taurine Dioxygenase (TauD) and is characterized by an isomer shift (δ) of 0.25 mm/s in
the Mössbauer absorption spectrum.

16

Computational studies approximated the isomer shift for

the high-valent Fe(IV)-oxo species at ~0.3 mm/s.

17

Using the method to observe the TauD ferryl

intermediate as a model, the isomer shifts were measured for bacterial TyrH and PheH using
Mössbauer spectroscopy with samples that were prepared by rapid freeze-quenching.

!

6!

13, 18

In

both enzyme samples an intermediate appeared at 20 ms, similar to TauD, after the O2 was
integrated into the sample. Analysis of the intermediate in the TyrH and PheH spectra resulted in
isomers shifts of 0.25 mm/s and 0.28 mm/s, respectively; thereby, providing evidence for the
formation of a high-valent Fe(IV)-oxo species in the mechanism of these pterin-dependent
hydroxylases.
The total sum of these spectroscopic observations demonstrates that molecular oxygen is
being heterolytically cleaved to hydroxylate both the substrate
oxo intermediate

13, 18

12

and cofactor

11

through a ferryl-

.

Figure 1.2b shows an iron-peroxypterin bridge intermediate before the formation of the
ferryl-oxo species. The concept of the peroxypterin species initially comes from the comparison
of UV spectra from the 4a-adduct formed with 5-deaza-6-methylpterin in the presence of
electrophiles.

19

The 4a-adduct produced a UV spectrum that is almost identical to that of the 4a-

peroxypterin species with the exception that the absorbance maxima were shifted by
approximately 2 nm. During an attempt to electrochemically characterize the 4a-peroxypterin
species, it was shown to reduce into the 4a-adduct. This observation suggests that the pterindependent reactions could possibly work through a peroxy intermediate. The 4a-peroxypterin
intermediate gained support through the production of H2O2 in the autoxidation reaction of
pterin in the presence of O2.

20

In the same study, a thermodynamic/kinetic inquiry of the

autoxidation reaction using cyclic voltammetry concluded that the transition state, 4a-

peroxypterin, is formed through a rate-limiting 1 e oxidation of the cofactor. The predicted

!

7!

mechanism is assigned through a comparative analysis involving the rate constant. The rate
.-

constant of the annihilation of the pterin and O2 radicals following the rate-limiting 1 e
oxidation was determined to be diffusion controlled, k = 3.6x10

10

-

-1 -1

M s . A similar rate

constant was observed during formation of the 4a-peroxyflavin species in an oxygenated
21

solution.

Adding the observation of the ferryl-oxo species

13, 18

, known production of a 4a-

10

hydroxypterin species , and the formation of the 4a-peroxypterin intermediate

20, 21

together,

there is a compelling argument for the formation of an iron-peroxy-pterin bridge, Fe(II) µperoxypterin. Currently in the literature there is no direct evidence for the involvement of the
Fe(II) in the formation of the 4a-peroxypterin species. In the absence of Fe(II) in the bacterial
form of PheH, the catalysis of dimethyltetrahydrobiopterin was shown to produce quinonoid
22

dihydropterin and hydrogen peroxide.

However, the ‘turnover’ with the missing ferrous iron

was characterized at about 5% of the wild type enzyme. The diminishing return and the
production of H2O2 suggest that the chemistry occurring at the catalytic Fe(II) site is important.
From density functional theory (DFT) it was shown that a favorable pathway starts with the
-

production of Fe(III)O2 from the catalytically relevant Fe(II).

23

-

The Fe(III)O2 was then

predicted to attack the C4a position of the pterin to form the Fe(II) µ-peroxypterin, which would
6

avoid the spin-forbidden reaction of triplet O2 with the pterin species . Although this is just
theoretical, it raises some interesting points about what chemistry is occurring at the Fe(II) site
before the formation of the high-valent Fe(IV)=O species.

!

8!

In the class of pterin-dependent Hydroxylases, the ferryl-oxo species is thought to
hydroxylate the amino acid substrate via an electrophilic aromatic substitution (EAS), figure
1.2d. It is difficult to measure the kinetics for the hydroxylation of the amino acid substrate
6

because steps leading to the the formation of the ferryl-species are rate-limiting. Instead of
using steady-state kinetics to investigate the hydroxylation of the amino acid with TyrH, a
5

technique employing product partitioning for a series of substituted phenylalanines was used. In
this empirical investigation a coupling factor, C (the number of moles of the hydroxylated amino
acid per mole of hydroxylated pterin), was used to create Hammett plots.

24

The Hammett plots

correlate the coupling factor to the substituent constant, σ, and a reaction constant, ρ. From the
product partitioning technique, a reaction constant, ρ, equal to approximately -5 was determined
5

and is indicative of an electron deficient transition state. The electron deficient transient state
was interpreted as a carbocation intermediate. The end result is a coupled hydroxylation of the
amino acid substrate and the pterin cofactor.

Section 1.3: Structural Features of PheH
The general mechanism as depicted in figure 1.2 avoids the ligand structure of the bound
Fe(II). Since the majority of the x-ray crystallographic studies concern PheH, the rest of this
work will assume a common mechanism between the pterin-dependent AAHs and focus on the
structural evidence presented for PheH.

!

9!

Two of the first x-ray crystallography studies done using mammalian PheH containing
the catalytically relevant ferrous iron are PDB 1J8T and 1J8U, figure 1.4 and 1.5—respectively.
These structures are from the double truncated human PheH (truncated hPheH), which means
both the regulatory and tetramerization domains were removed. From this work two crystal
structures with the truncated hPheH were determined: either complexed (1) with (PDB 1J8U) or

Figure 1.4: X-ray crystal structure of truncated human PheH with the catalytically
relevant ferrous iron (orange) in the active site resolved to 1.7Å. (PDB: 1J8T). For
interpretation of the references to color in this and all other figures, the reader is referred
to the electronic version of this dissertation.
25

(2) without (1J8T) BH4.

Both crystal structures shown have three common residues in the

active site, one glutamate and two histidines that are facially ligated to the ferrous iron. These
three residues are also known as the ‘facial triad’ and are characteristic to this class of non-heme
iron enzymes. The other ligands present are the oxygen(s) presumably from water ligands. From

!

10!

the 1J8T structure only one of the three water ligands present in 1J8U has a well defined electron
density. The two poorly defined water ligands have been omitted from the 1J8T PDB file.
Anderson et al. explains that presence of the poorly defined electron density may be the result of
a small population that has a similar 6-coordinate structure to the 1J8U structure. The water
ligand that is present in both structures, termed Wat3, is hydrogen bound to Glu 286, which

Figure 1.5: X-ray crystal structure of truncated human PheH with the catalytically
relevant ferrous iron (orange) containing the cofactor, BH4 (PDB: 1J8U).
through site directed mutagenesis was found to be critically important for the binding of pterin.

26

When the Glu 286 was replaced with alanine the Km of the pterin increased 70 fold, suggesting
the affinity for the binding of pterin decreased substantially. The other major difference between
the 1J8T and 1J8U crystal structures is the placement of the Glu 330 residue.

!

11!

25

In the 1J8T

Figure 1.6: On the left is the Fe K edge XAS results from the comparison of (top left) rPheH to
(bot left) three model complexes forming six-, five- and four-coordinated structures. (Right
top) The MCD spectra for distorted octahedral (6C), square pyramidal (5C), and tetrahedral (4)
geometries. (Right bot) MCD spectra for a variety of rPheH samples: (black) rPheH[], (blue)
rPheH[L-phe], (green) rPheH[6-MPH4], and (red) rPheH[L-phe,6-MPH4].
structure, the Glu 330 has a higher degree of disorder and cannot be fit into the observed electron
density; the reason for this is still unclear.
The four-coordinate complex described by 1J8T
from x-ray absorption spectroscopy (XAS).
T

R

spectra for rat PheH [] and PheH [L-phe],

!

27

27

25

is in contradiction with the conclusions

The top left spectra of figure 1.6 shows the XAS
the T- and R-state designations

12!

28

will be described

in more detail later. Comparing these spectra to the bottom left spectra of figure 1.6 describes the
overall geometry. The major features used here to assign the geometry of the enzyme structures
are the pre-edge features at ~7112 eV resulting from 1s!3d transition, highlighted in the figures
insets, and the edge feature at ~7125 eV resulting from 1s!4p transitions. Empirically, it can be
seen that the enzyme samples match the distorted octahedral geometry of [Fe(imidazole)6]Cl2 in
the pre-edge and edge feature. These data

27

suggest that the enzyme with an empty active site
25

takes on a six-coordinate geometry unlike the predicted geometry from the 1J8T

crystal

structure.
Analysis of Magnetic Circular Dichroism (MCD) spectra for a variety of rat PheH (rPheH)
samples against four-, five- and six-coordinate geometries gives more insight into the structure of
the PheH active site, figure 1.6 (right).

29

The geometries are investigated using electronic
2+

transitions associated to PheH. Non-heme Fe

takes on a high spin configuration in the presence
5

of nitrogen and oxygen ligands coordinated to the active site and results in a splitting of the D
5

5

state into the T2g ground state and the Eg excited state.

30

Following the selection rules for the

electronic spectroscopy, Δms = 0 and Δml = ±1, there exists a one-electron transfer from
5

5

4

2

3

3

T2g! Eg that corresponds to the (t2g) (eg) !(t2g) (eg) transition. The low symmetry around
2+

the Fe

removes the degeneracy of the dz2 and dx2-y2 orbitals in the eg state and reflects the
5

splitting of the Eg state, which is dependent on the active site geometry.
29

Figure 1.6 (top right) shows the MCD spectra

of three model complexes: a six-

coordinate distorted octahedral, five-coordinate square pyramidal, and four-coordinate

!

13!

tetrahedral geometry. From the spectra it can be ascertained that the distorted octahedral
5

-1

structure has a Δ Eg splitting of approximately 1600 cm , the distorted square pyramidal
-1

structure has a splitting of ~5800 cm and the tetrahedral geometry resulted in a splitting of
-1

~500 cm (it is unclear from the picture but the number shown is the position of a spectral
5

-1

feature and not the energy difference; Δ Eg = 6500 - 6000 cm ). Comparing these model
geometries to the range of samples created with rat PheH, figure 1.6 (right bot), shows that the
5

enzyme by itself (black line) conforms to a distorted octahedral geometry with a Δ Eg splitting
-1

<2000 cm . These data are in contradiction to the result of the 1J8T
consistent with the XAS conclusions

27

25

crystal structure but is

previously discussed. A similar geometry, six-coordinate

distorted octahedral, is predicted with MCD when only the pterin analogue, 5-deaza-6methyltetrahydrobiopterin, is present with the enzyme (green spectra)
the 1J8U

25

29

and is consistent with

crystal structure. When just the amino acid substrate, L-phenylalanine, is in the

enzymatic sample (blue spectrum) the ligand field is perturbed but the overall geometry cannot
29

be assigned using MCD;

27

however, previously, with XAS , it was assigned as a six-coordinate

distorted octahedral structure. Interestingly, with the addition of the substrate, L-phenylalanine,
and the cofactor, 5-deaza-6-methyltetrahydropterin, to rPheH the MCD spectra depicted an
alteration to the Fe(II) ligation, causing it to change from a six coordinate to a five coordinate
structure.

29

The geometry change resembles a distorted octahedron transforming into a distorted

square pyramidal structure, figure 1.7.

!

14!

Figure 1.7: Above shows the suspected loss of a water ligand when rPheH is in the presence of
5-deaza-6-methyltetrahydrobiopterin and L-phenylalanine.
The open coordination site created after the ligand rearrangement, seen in figure 1.7, has
been suggested to be the position for the binding of O2

29

but has not been confirmed for the non-

31

heme pterin-dependent hydroxylases . Previous studies from other non-heme iron enymzes,
32

such as isopenicillin N-synthase

33

and extradiol dioxygenases , show nitric oxide, an O2

mimic, coordinating to the open position on the ferrous iron after the substrate binds. Han et al.
used

17

O line broadening with continuous wave-electron paramagnetic resonance (CW-EPR)

and UV-visible measurements to propose that the NO replaces one of the water ligands left in the
five-coordinate structure.
Olsson et al.

34

31

presented a model for water dissociation that ultimately leads to a ferryl-

oxo species. The dissociation of water ligands was suggested to come from the direct ligation of
the pterin cofactor. However, BH4 was found to be complexed to the enzyme in the second
coordination sphere of ferrous iron.

25

The line broadening experiments were done using bacterial PheH (bPheH), from C.
violaceum, in the presence of nitric oxide to create an EPR active complex, S = 3/2.
7

31

With

enriched water the CW-spectrum was collected for [FeNO] -bPheH[6,7-dimethyl-5,6,7,8-

!

15!

17

O-

7

tetrahydropterin (DMPH4)] and [FeNO] -bPheH[phe, DMPH4]. The sample containing only the
pterin species showed a 0.4 mT – 0.5 mT broadening of the g = 4 region, ~160 mT to 175 mT,
when in the presence of

17

O versus

16

7

O. Conversely, the [FeNO] -bPheH[phe, DMPH4]

revealed no line broadening form the isotopically labeled solution; thereby, suggesting that the
fully loaded enzyme has a dehydrated first coordination sphere.
Starting from the 1J8U

25

structure, the loss of water ligand(s) through the conversion
from six- to five-coordinate

29

leaves the

possibility of two waters in the active site. The
nitric oxide can either replace ligated water or
causes a structural change to form the
31

dehydrated active site observed previously.

-

Raman spectroscopy was used to assign NO 3+

Fe

charge-transfer transitions at 575, 509
-

*

and 443 nm with NO 2π to F
and dz2, respectively.
Figure 1.8: The suspected rotation of the
nitrosyl group in the presense of both the
substrate and the cofactor.

35

3+

dx2-y2, dxz,

The strength of the

transitions increase as the overlap between the
ligand and metal orbitals increase, as seen in
7

figure 1.8. The UV-visible spectrum of the [FeNO] -PheH[phe, DMPH4] shown an increased
7

7

intensity ranging across these spectral features relative to the [FeNO] -PheH[] and [FeNO] PheH[phe] samples. The increased intensity between 500 and 600 nm was attributed to an

!

16!

increase in the overlap of the orbitals as a result of the nitric oxide binding equatorially rather
than axial as depicted in figure 1.7.

31

From the MCD spectra a five-coordinate structure was observed.

29

A five-coordinate

structure was also observed in a 2.5Å crystal structure of truncated hPheH, figure 1.9, that
36

contained both BH4 and a substrate analogue, 3-(2-Thienyl)-L-alanine (PDB 1KW0).

The

ferrous iron was found to be ligated by the 2-Histidine-1-Carboxylate motif, with the glutamate
(Glu 330) coordinated to Fe(II) in a bidentate fashion, and one water ligand. When the substrate

Figure 1.9: X-ray crystal structure of truncated human PheH with the catalytically relevant
ferrous iron (orange) containing the cofactor, BH4, and the substrate analogue, 3-(2-Thienyl)-Lalanine (PDB: 1KW0).
!

17!

and cofactor are present the glutamate ligates bidendate resulting in the loss of two of the waters
relative to the 1J8U structure. Accompanying the loss of these water ligands is the movement of
pterin resulting in a shortening of the distance between the iron and the position of
hydroxylation, C4a. The 1J8U crystal structure has the Fe-C4a distance at 5.9Å,

25

while the

36

1KW0 crystal structure places this distance at 4.5 Å . Anderson et al. improved the resolution
of this structure to 2 Å using the truncated hPheH, figure 1.4, and the same ligands—BH4 and 3(2-Thienyl)-L-alanine (PDB 1MMK).

37

The oxygen from the water ligand in the 1KW0

Figure 1.10: X-ray crystallographic structure of truncated human PheH with
thienylalanine and tetrahydrobiopterin complexed to the catalytically relevant
ferrous iron (PDB: 1MMK).

!

18!

36

2

crystal structure has a slightly high B-factor, 44 Å . The B-factor is a measurement of dispersion
associated to the electron density for a particular atom. A high B-factor, like what was observed
for the water in 1KW0

36

makes it unclear if the water ligand is actually present within the
37

structure. Refinement of the crystal structure, 1MMK (figure 1.10),

revealed an even higher B-

2

factor, 53 Å , which was on the same order as the noise level and is the reason it was omitted
from the crystal structure.
As was previously discussed, the oxygen is thought to form an iron-peroxypterin bridge.
However, there is no direct proof of the oxygen bridge. Upon close inspection of the 1MMK
crystal structure, the distance from the Fe(II) to the C4a position on the pterin can be measured
as a distance of 4.4 Å. The gap between these atoms is large enough to account for the formation
of an O2 bridge assuming molecular oxygen is bound offset from the distance label in figure 1.4.
Based on the crystal structure it looks as though the O2 could be bound axial or equatorially. It is
typically thought of as binding axially to ensure the proper S-type stereocenter is formed at the
C4a position.
Both 1KW0

36

and 1MMK

37

were made using a post-crystallization diffusion soaking

technique. The method entailed treating the binary crystals, hPheH[BH4], with the solid form of
the substrate and giving the diffusion soaking process 24 hours. In these structures a substrate
analogue, 3-(2-Thienyl)-L-alanine, was used in lieu of L-phenylalanine. Treatment with the
thienylalanine seemed to leave the binary crystals unaffected. However, the diffraction pattern
from the enzymes treated with thienylalanine showed a mosacity (disorder calculated from

!

19!

peakwidth) of that was two- to three-fold higher than the binary crystals. When the truncated
hPheH was treated with solid L-phe the crystals were found to deform/disintegrate. The loss of
crystals suggests that a major conformational change is occurring upon the addition of the
substrate to the enzyme. A conformational change accompanying the addition of thienylalanine
could explain the increased disorder that was observed in these structures.

!

20!

Section 1.4: Regulation of PheH
In solution, mammalian PheH forms a tetrameric complex, figure 1.11a, of monomer
peptides that have a molar mass of ~50 kDa. This corresponds to a molar mass of ~200 kDa for

Figure 1.11: (a) X-ray crystal structure from the tetrameric form of human PheH with ferric
iron in the catalytic centers of the monomers (PDB: 2PAH). (b) Designation of the domains,
using rat PheH, in the monomer: residue 1-118 represents the regulatory domain (white),
118-410 is the catalytic domain, and 410-453 is the tetramerization domain.

!

21!

2, 38

the tetrameric form.

Each monomer consists of three domains, figure 1.11b, containing a

tetramerization domain (white), a functional domain (gray) and a regulation domain (black).
40

39,

Figure 1.11a represents the tetrameric structure of human PheH (PDB: 2PAH) with truncated

monomers that included residue 118 to 452. The monomeric units associated to this tetramer had
their regulatory domain removed and the active site was loaded with the non-catalytic ferric
iron.

40

This crystal structure illustrates that hPheH forms a tetrameric unit with the four

monomers attached through the tetramerization domain.

37

PheH is regulated through cooperative binding of L-phenylalanine, cofactor inhibition,
and phosphorylation.

2, 3, 6, 39, 41

Figure 1.12 shows the general mechanism for the regulation of

Figure 1.12: Mechanism of the regulation of phenylalanine hydroxylases.

PheH reported in the literature.

38, 39, 42, 43

When elevated levels of the cofactor are introduced

T

to the ‘inactive’ form of PheH, E , the enzyme is inhibited. Inhibition of the enzyme is
illustrated with a 10-fold decrease of the specific activity. These inhibition studies were done
T

using rat PheH and the inhibited form is represented in figure 1.12 as E /BH4.

36, 39

Specific

activities were also determined for the ‘inactive’ form with and without a 3-5 minute pre-

!

22!

R

incubation step, which allows for activation (phe/E in figure 1.12). The range of time used
during the pre-incubation is the result of two different analyses within the literature, which
pertained to either rat PheH (3 min) or mammalian PheH (5 min) forms of the enzyme.
Exclusion of the pre-incubation step resulted in a specific activity, represented as µmol of
tyrosine produced/min/mg of enzyme, of 0.74 and 0.42 µmol/min/mg for hPheH and rPheH—
respectively. By including the pre-incubation step the specific active of the enzyme was found to
be 1.66 µmol/min/mg for hPheH and 3.36 µmol/min/mg for rPheH, which is a 2.2 and 8-fold
increase in the specific activity. With phenylalanine pre-incubation experiments it was shown
T

that there are two different states of the enzyme, an ‘inactive’ form (E ) and an ‘activated’ form
R

(E ). The conversion between the two states is an equilibrium that requires ~50 and 60 µM of
phenylalanine to drive that activation for hPheH and rPheh, respectively.

44, 45

From the

activated state of the enzyme the substrate and cofactor are both hydroxylated with a coupled
mechanism as previously discussed.
Stoichiometry dictates that at least one molecule of phenylalnine must be present to
produce one tyrosine in the hydroxylation reaction that is facilitated using PheH. When
considering that activation of PheH is dependent upon the concentration of phenylalanine, it
becomes entirely possible that there might be a secondary (or regulatory) binding site for
phenylalanine. Another binding site within the tetramer would increase the stoichiometry
required to a value of more than one phenylalanine per monomeric unit. Using filter binding
46

assasys the stoichiometry of phenylalanine per monomer was determined as ~1.5 , while

!

23!

47

isothermal titration calorimetry predicted only 1 phenylalanine per monomer . The filter
46

binding assays were done with rat PheH , while the calorimetry experiments were
47

accomplished with nematode PheH (cePheH) . It is worth noting that the cePheH form is
known to lack an activated complex and a cooperative response to L-phe, which suggests there
48

may not be a regulatory binding site for phenylalanine.

Conversely, rat PheH was studied

using hydrophobic chromatography to show that the absorption properties of the enzyme change
upon activation, which implies there is a conformational change that occurs with the addition of
substrate.

49

The sum of these observations suggests that the behavior of rPheH is closer to

mammalian PheH than is the cePheH form. Eventually, NMR studies were able to confirm the
1

14

presence of a phenylalanine binding site in the isolated regulator domain using H- N
heteronuclear single quantum correlation spectroscopy.
The allosteric interaction of L-phe with PheH
results in a partial closure of the active site.

37

51

50

causes a global rearrangement

44

that

Upon activation from the T- to the R-state using a

substrate analogue, thienylalanine, the residues from 131-155 were seen to fold in closer towards
36

the active site.

This global movement can be seen by comparing the ‘unactivated’ and

‘activated’ crystal structures, which corresponds to 1J8U (figure 1.13)

25

and 1MMK

37

(figure

1.14) –respectively. Within the residue loop the Tyr 138, highlighted red in figure 1.13 and 1.14,
is observed to go through the largest displacement during activation of PheH. It is displaced 18.5
Å in the direction of the iron, which brings the oxygen from the hydroxyl group on the tyrosine
(Tyr138) within 6.5 Å of the iron. This oxygen on the tyrosine residue is close enough to form an

!

24!

interaction with an internal water molecule through hydrogen bonds, which is also in close
enough proximity to form a hydrogen bond to the carbonyl group on the pterin.

36

With the

activation, a global conformational change is seen that coincides with the loss of the up to three
water ligands. The structural rearrangement of the active site during the activation of PheH may
be explained by the steric interactions created by the Tyr 138 when the residue is moved closer

Figure 1.13: X-ray crystal structure of truncated human PheH with the catalytically relevant
ferrous iron (orange) containing the cofactor, BH4 (PDB: 1J8U).
!

25!

Figure 1.14: X-ray crystal structure of truncated human PheH with the catalytically relevant
ferrous iron (orange) containing the cofactor, BH4, and 3-(2-Thienyl)-L-alanine (PDB:
1MMK).
to the iron center.

36, 37

The effect of this activation on the ligand geometry has been previously investigated with
XAS spectroscopies using rPheH. Loeb et al. concluded that transformation from the T- and R2+

T

2+

R

states, [Fe ]-rPAH [BH4] and [Fe ]-rPAH [L-Phe, BH4], brings about a ligand

!

26!

rearrangement that transforms the active site geometry from a distorted octahedral to distorted
square pyramidal type structure.

27

From the crystal structures the activation creates a ligand

rearrangement that results in only a four coordinate iron, but minus one equatorial ligand the
general shape is loosely consistent with the distorted square pyramidal geometry.

25, 36, 37

Currently, there are no crystal structures with just phenylalanine present to outline the activation
of the enzyme; However, XAS spectroscopy suggests that activation in the absence of BH4 does
not change the ligand geometry

27

but it is unclear based on the literature.

Section 1.5: The Focus
Currently the literature is unclear about the existence of a water in the first coordination
sphere of the Fe(II) during the coupled hydroxylation carried out by pterin-dependent AAH. The
36

catalytically relevant crystal structures for PheH, 1KW0

37

and 1MMK , depict one and zero

water ligands respectively. Both structures were obtained from crystals that resulted from the
post-crystallization soaking of “slow substrate,” 3-(2-Thienyl)-L-alanine, into crystals grown
anaerobically in the presence of cofactor BH4. These authors noted that attempts to soak the
physiologically relevant substrate, L-phe, into these crystals were unsuccessful.

36

The fact that it

is not the physiological substrate raises the question about the validity of the catalytic structures
presented. Other spectroscopies, XAS and EPR, suggest that at least one water ligand is present
when both the substrate and cofactor are integrated into the sample.

27, 31

In the literature there is

very little information about what occurs at the active site when just phenylalanine is present.

!

27!

The only spectroscopic information that aims to answer this question is XAS spectroscopy,
which shows that the activation causes no major geometric change at the active site. However,
the results are unclear if there is a ligand rearrangement occurring. Using pulse EPR techniques
these questions will be investigated.
In the present work, we have used the pulsed EPR methods of ESEEM and HYSCORE,
described in chapter 2, to study recombinant rat phenylalanine hydroxylase in the presence of the
actual substrate, L-phenylalanine. To study this enzyme, an O2 surrogate, NO, is used to produce
an EPR active S = 3/2 spin system, which will also be described in chapter 2. This approach
52-58

allows for the study of the enzymes with the biologically relevant substrate.
7

use this spin system in PheH along with a series of model [FeNO] complexes

59

Chapter 3 will
of known

structure to empirically determine the number of coordinated water molecules. A more analytical
description of the water ligands present in the model and PheH samples that describe the
activation are discussed in chapter 4.

!

28!

BIBLIOGRAPHY

!

29!

BIBLIOGRAPHY

1.!Fitzpatrick,!P.!F.!(1999)!TetrahydropterinAdependent!amino!acid!hydroxylases,!Annu$Rev$
Biochem!68,!355A381.!
2.!Flatmark,!T.,!and!Stevens,!R.!C.!(1999)!Structural!Insight!into!the!Aromatic!Amino!Acid!
Hydroxylases!and!Their!DiseaseARelated!Mutant!Forms,!Chem$Rev!99,!2137A2160.!
3.!Kappock,!T.!J.,!and!Caradonna,!J.!P.!(1996)!PterinADependent!Amino!Acid!Hydroxylases,!
Chem$Rev!96,!2659A2756.!
4.!Solomon,!E.!I.,!Brunold,!T.!C.,!Davis,!M.!I.,!Kemsley,!J.!N.,!Lee,!S.!K.,!Lehnert,!N.,!Neese,!F.,!
Skulan,!A.!J.,!Yang,!Y.!S.,!and!Zhou,!J.!(2000)!Geometric!and!electronic!
structure/function!correlations!in!nonAheme!iron!enzymes,!Chem$Rev!100,!235A350.!
5.!Hillas,!P.!J.,!and!Fitzpatrick,!P.!F.!(1996)!A!mechanism!for!hydroxylation!by!tyrosine!
hydroxylase!based!on!partitioning!of!substituted!phenylalanines,!Biochemistry!35,!
6969A6975.!
6.!Fitzpatrick,!P.!F.!(2003)!Mechanism!of!aromatic!amino!acid!hydroxylation,!Biochemistry!
42,!14083A14091.!
7.!Flatmark,!T.,!Stokka,!A.!J.,!and!Berge,!S.!V.!(2001)!Use!of!surface!plasmon!resonance!for!
realAtime!measurements!of!the!global!conformational!transition!in!human!
phenylalanine!hydroxylase!in!response!to!substrate!binding!and!catalytic!activation,!
Analytical$biochemistry!294,!95A101.!
8.!Stokka,!A.!J.,!and!Flatmark,!T.!(2003)!SubstrateAinduced!conformational!transition!in!
human!phenylalanine!hydroxylase!as!studied!by!surface!plasmon!resonance!
analyses:!the!effect!of!terminal!deletions,!substrate!analogues!and!phosphorylation,!
Biochem$J!369,!509A518.!
9.!Solomon,!E.!I.,!Brunold,!T.!C.,!Davis,!M.!I.,!Kemsley,!J.!N.,!Lee,!S.AK.,!Lehnert,!N.,!Neese,!F.,!
Skulan,!A.!J.,!Yang,!Y.AS.,!and!Zhou,!J.!(1999)!Geometric!and!Electronic!
Structure/Function!Correlations!in!NonAHeme!Iron!Enzymes,!Chemical$Reviews!100,!
235A350.!

!

30!

10.!Dix,!T.!A.,!Bollag,!G.!E.,!Domanico,!P.!L.,!and!Benkovic,!S.!J.!(1985)!PhenylalanineA
Hydroxylase!A!AbsoluteAConfiguration!and!Source!of!Oxygen!of!the!4aA
Hydroxytetrahydropterin!Species,!Biochemistry!24,!2955A2958.!
11.!Lazarus,!R.!A.,!Debrosse,!C.!W.,!and!Benkovic,!S.!J.!(1982)!PhenylalanineAHydroxylase!A!
Structural!Determination!of!the!Tetrahydropterin!Intermediates!by!CA13!NmrA
Spectroscopy,!Journal$of$the$American$Chemical$Society!104,!6869A6871.!
12.!Kaufman,!S.,!Bridgers,!W.!F.,!Friedman,!S.,!and!Eisenberg,!F.!(1962)!Source!of!Oxygen!in!
Phenylalanine!Hydroxylase!and!DopamineABetaAHydroxylase!Catalyzed!Reactions,!
Biochem$Bioph$Res$Co!9,!497A&.!
13.!Eser,!B.!E.,!Barr,!E.!W.,!Frantorn,!P.!A.,!Saleh,!L.,!Bollinger,!J.!M.,!Krebs,!C.,!and!Fitzpatrick,!
P.!F.!(2007)!Direct!spectroscopic!evidence!for!a!highAspin!Fe(IV)!intermediate!in!
tyrosine!hydroxylase,!Journal$of$the$American$Chemical$Society!129,!11334A+.!
14.!Moran,!G.!R.,!Phillips,!R.!S.,!and!Fitzpatrick,!P.!F.!(1999)!Influence!of!steric!bulk!and!
electrostatics!on!the!hydroxylation!regiospecificity!of!tryptophan!hydroxylase:!
characterization!of!methyltryptophans!and!azatryptophans!as!substrates,!
Biochemistry!38,!16283A16289.!
15.!Lazarus,!R.!A.,!Debrosse,!C.!W.,!and!Benkovic,!S.!J.!(1982)!Structural!Determination!of!
Quinonoid!Dihydropterins,!Journal$of$the$American$Chemical$Society!104,!6871A6872.!
16.!Hoffart,!L.!M.,!Barr,!E.!W.,!Guyer,!R.!B.,!Bollinger,!J.!M.,!and!Krebs,!C.!(2006)!Direct!
spectroscopic!detection!of!a!CAHAcleaving!highAspin!Fe(IV)!complex!in!a!prolylA4A
hydroxylase,!P$Natl$Acad$Sci$USA!103,!14738A14743.!
17.!Neese,!F.!(2006)!Theoretical!spectroscopy!of!modelAnonheme![Fe(IV)OL5](2+)!
complexes!in!their!lowest!triplet!and!quintet!states!using!multireference!ab!initio!
and!density!functional!theory!methods,!Journal$of$Inorganic$Biochemistry!100,!716A
726.!
18.!Panay,!A.!J.,!Lee,!M.,!Krebs,!C.,!Bollinger,!J.!M.,!and!Fitzpatrick,!P.!F.!(2011)!Evidence!for!a!
HighASpin!Fe(IV)!Species!in!the!Catalytic!Cycle!of!a!Bacterial!Phenylalanine!
Hydroxylase,!Biochemistry!50,!1928A1933.!

!

31!

19.!Moad,!G.,!Luthy,!C.!L.,!Benkovic,!P.!A.,!and!Benkovic,!S.!J.!(1979)!Studies!on!6AMethylA5A
Deazatetrahydropterin!and!Its!4a!Adducts,!Journal$of$the$American$Chemical$Society!
101,!6068A6076.!
20.!Eberlein,!G.,!Bruice,!T.!C.,!Lazarus,!R.!A.,!Henrie,!R.,!and!Benkovic,!S.!J.!(1984)!The!
interconversion!of!the!5,6,7,8AtetrahydroA,!6,7,8AdihydroA,!and!radical!forms!of!
6,6,7,7Atetramethyldihydropterin.!A!model!for!the!biopterin!center!of!aromatic!
amino!acid!mixed!function!oxidases,!Journal$of$the$American$Chemical$Society!106,!
7916A7924.!
21.!Massey,!V.!(1994)!Activation!of!molecular!oxygen!by!flavins!and!flavoproteins,!J$Biol$
Chem!269,!22459A22462.!
22.!Chen,!D.!W.,!and!Frey,!P.!A.!(1998)!Phenylalanine!hydroxylase!from!Chromobacterium!
violaceum!A!Uncoupled!oxidation!of!tetrahydropterin!and!the!role!of!iron!in!
hyroxylation,!Journal$of$Biological$Chemistry!273,!25594A25601.!
23.!Bassan,!A.,!Blomberg,!M.!R.,!and!Siegbahn,!P.!E.!(2003)!Mechanism!of!dioxygen!cleavage!
in!tetrahydrobiopterinAdependent!amino!acid!hydroxylases,!Chemistry!9,!106A115.!
24.!Hammett,!L.!P.!(1937)!The!Effect!of!Structure!upon!the!Reactions!of!Organic!
Compounds.!Benzene!Derivatives,!Journal$of$the$American$Chemical$Society!59,!96A
103.!
25.!Andersen,!O.!A.,!Flatmark,!T.,!and!Hough,!E.!(2001)!High!resolution!crystal!structures!of!
the!catalytic!domain!of!human!phenylalanine!hydroxylase!in!its!catalytically!active!
Fe(II)!form!and!binary!complex!with!tetrahydrobiopterin,!J$Mol$Biol!314,!279A291.!
26.!Dickson,!P.!W.,!Jennings,!I.!G.,!and!Cotton,!R.!G.!(1994)!Delineation!of!the!catalytic!core!of!
phenylalanine!hydroxylase!and!identification!of!glutamate!286!as!a!critical!residue!
for!pterin!function,!J$Biol$Chem!269,!20369A20375.!
27.!Loeb,!K.!E.,!Westre,!T.!E.,!Kappock,!T.!J.,!Mitic,!N.,!Glasfeld,!E.,!Caradonna,!J.!P.,!Hedman,!
B.,!Hodgson,!K.!O.,!and!Solomon,!E.!I.!(1997)!Spectroscopic!characterization!of!the!
catalytically!competent!ferrous!site!of!the!resting,!activated,!and!substrateAbound!
forms!of!phenylalanine!hydroxylase,!Journal$of$the$American$Chemical$Society!119,!
1901A1915.!

!

32!

28.!Monod,!J.,!Wyman,!J.,!and!Changeux,!J.!P.!(1965)!On!Nature!of!Allosteric!Transitions!A!a!
Plausible!Model,!Journal$of$Molecular$Biology!12,!88A&.!
29.!Solomon,!E.!I.,!Decker,!A.,!and!Lehnert,!N.!(2003)!NonAheme!iron!enzymes:!Contrasts!to!
heme!catalysis,!P$Natl$Acad$Sci$USA!100,!3589A3594.!
30.!Solomon,!E.!I.,!Pavel,!E.!G.,!Loeb,!K.!E.,!and!Campochiaro,!C.!(1995)!Magnetic!circular!
dichroism!spectroscopy!as!a!probe!of!the!geometric!and!electronic!structure!of!nonA
heme!ferrous!enzymes,!Coordin$Chem$Rev!144,!369A460.!
31.!Han,!A.!Y.,!Lee,!A.!Q.,!and!AbuAOmar,!M.!M.!(2006)!EPR!and!UVAvis!studies!of!the!nitric!
oxide!adducts!of!bacterial!phenylalanine!hydroxylase:!Effects!of!cofactor!and!
substrate!on!the!iron!environment,!Inorganic$Chemistry!45,!4277A4283.!
32.!Roach,!P.!L.,!Clifton,!I.!J.,!Hensgens,!C.!M.,!Shibata,!N.,!Schofield,!C.!J.,!Hajdu,!J.,!and!
Baldwin,!J.!E.!(1997)!Structure!of!isopenicillin!N!synthase!complexed!with!substrate!
and!the!mechanism!of!penicillin!formation,!Nature!387,!827A830.!
33.!Sato,!N.,!Uragami,!Y.,!Nishizaki,!T.,!Takahashi,!Y.,!Sazaki,!G.,!Sugimoto,!K.,!Nonaka,!T.,!
Masai,!E.,!Fukuda,!M.,!and!Senda,!T.!(2002)!Crystal!structures!of!the!reaction!
intermediate!and!its!homologue!of!an!extradiolAcleaving!catecholic!dioxygenase,!J$
Mol$Biol!321,!621A636.!
34.!Olsson,!E.,!Martinez,!A.,!Teigen,!K.,!and!Jensen,!V.!R.!(2011)!Formation!of!the!IronAOxo!
Hydroxylating!Species!in!the!Catalytic!Cycle!of!Aromatic!Amino!Acid!Hydroxylases,!
ChemLEur$J!17,!3746A3758.!
35.!Brown,!C.!A.,!Pavlosky,!M.!A.,!Westre,!T.!E.,!Zhang,!Y.,!Hedman,!B.,!Hodgson,!K.!O.,!and!
Solomon,!E.!I.!(1995)!Spectroscopic!and!Theoretical!Description!of!the!ElectronicA
Structure!of!S=3/2!IronANitrosyl!Complexes!and!Their!Relation!to!OA2!Activation!by!
Nonheme!Tron!Enzyme!ActiveASites,!Journal$of$the$American$Chemical$Society!117,!
715A732.!
36.!Andersen,!O.!A.,!Flatmark,!T.,!and!Hough,!E.!(2002)!Crystal!structure!of!the!ternary!
complex!of!the!catalytic!domain!of!human!phenylalanine!hydroxylase!with!
tetrahydrobiopterin!and!3A(2Athienyl)ALAalanine,!and!its!implications!for!the!
mechanism!of!catalysis!and!substrate!activation,!J$Mol$Biol!320,!1095A1108.!

!

33!

37.!Andersen,!O.!A.,!Stokka,!A.!J.,!Flatmark,!T.,!and!Hough,!E.!(2003)!2.0A!resolution!crystal!
structures!of!the!ternary!complexes!of!human!phenylalanine!hydroxylase!catalytic!
domain!with!tetrahydrobiopterin!and!3A(2Athienyl)ALAalanine!or!LAnorleucine:!
substrate!specificity!and!molecular!motions!related!to!substrate!binding,!J$Mol$Biol!
333,!747A757.!
38.!Xia,!T.,!Gray,!D.!W.,!and!Shiman,!R.!(1994)!Regulation!of!rat!liver!phenylalanine!
hydroxylase.!III.!Control!of!catalysis!by!(6R)Atetrahydrobiopterin!and!phenylalanine,!
J$Biol$Chem!269,!24657A24665.!
39.!Fitzpatrick,!P.!F.!(2012)!Allosteric!regulation!of!phenylalanine!hydroxylase,!Archives$of$
biochemistry$and$biophysics!519,!194A201.!
40.!Fusetti,!F.,!Erlandsen,!H.,!Flatmark,!T.,!and!Stevens,!R.!C.!(1998)!Structure!of!tetrameric!
human!phenylalanine!hydroxylase!and!its!implications!for!phenylketonuria,!J$Biol$
Chem!273,!16962A16967.!
41.!AbuAOmar,!M.!M.,!Loaiza,!A.,!and!Hontzeas,!N.!(2005)!Reaction!mechanisms!of!
mononuclear!nonAheme!iron!oxygenases,!Chemical$Reviews!105,!2227A2252.!
42.!Shiman,!R.,!Xia,!T.,!Hill,!M.!A.,!and!Gray,!D.!W.!(1994)!Regulation!of!rat!liver!
phenylalanine!hydroxylase.!II.!Substrate!binding!and!the!role!of!activation!in!the!
control!of!enzymatic!activity,!J$Biol$Chem!269,!24647A24656.!
43.!Shiman,!R.,!Gray,!D.!W.,!and!Hill,!M.!A.!(1994)!Regulation!of!rat!liver!phenylalanine!
hydroxylase.!I.!Kinetic!properties!of!the!enzyme's!iron!and!enzyme!reduction!site,!J$
Biol$Chem!269,!24637A24646.!
44.!Shiman,!R.,!and!Gray,!D.!W.!(1980)!Substrate!Activation!of!PhenylalanineAHydroxylase!A!
a!Kinetic!Characterization,!Journal$of$Biological$Chemistry!255,!4793A4800.!
45.!Kowlessur,!D.,!Citron,!B.!A.,!and!Kaufman,!S.!(1996)!Recombinant!Human!Phenylalanine!
Hydroxylase:!Novel!Regulatory!and!Structural!Properties,!Archives$of$biochemistry$
and$biophysics!333,!85A95.!
46.!Kaufman,!S.,!and!Mason,!K.!(1982)!Specificity!of!amino!acids!as!activators!and!
substrates!for!phenylalanine!hydroxylase,!J$Biol$Chem!257,!14667A14678.!

!

34!

47.!Flydal,!M.!I.,!Mohn,!T.!C.,!Pey,!A.!L.,!SiltbergALiberles,!J.,!Teigen,!K.,!and!Martinez,!A.!
(2010)!Superstoichiometric!binding!of!LAPhe!to!phenylalanine!hydroxylase!from!
Caenorhabditis!elegans:!evolutionary!implications,!Amino$acids!39,!1463A1475.!
48.!Calvo,!A.!C.,!Pey,!A.!L.,!Ying,!M.,!Loer,!C.!M.,!and!Martinez,!A.!(2008)!Anabolic!function!of!
phenylalanine!hydroxylase!in!Caenorhabditis!elegans,!FASEB$journal$:$official$
publication$of$the$Federation$of$American$Societies$for$Experimental$Biology!22,!
3046A3058.!
49.!Shiman,!R.,!Gray,!D.!W.,!and!Pater,!A.!(1979)!A!simple!purification!of!phenylalanine!
hydroxylase!by!substrateAinduced!hydrophobic!chromatography,!J$Biol$Chem!254,!
11300A11306.!
50.!Li,!J.,!Ilangovan,!U.,!Daubner,!S.!C.,!Hinck,!A.!P.,!and!Fitzpatrick,!P.!F.!(2011)!Direct!
evidence!for!a!phenylalanine!site!in!the!regulatory!domain!of!phenylalanine!
hydroxylase,!Archives$of$biochemistry$and$biophysics!505,!250A255.!
51.!Shiman,!R.!(1980)!Relationship!between!the!Substrate!Activation!Site!and!Catalytic!Site!
of!PhenylalanineAHydroxylase,!Journal$of$Biological$Chemistry!255,!29A32.!
52.!Arciero,!D.!M.,!Orville,!A.!M.,!and!Lipscomb,!J.!D.!(1985)![17O]Water!and!nitric!oxide!
binding!by!protocatechuate!4,5Adioxygenase!and!catechol!2,3Adioxygenase.!Evidence!
for!binding!of!exogenous!ligands!to!the!active!site!Fe2+!of!extradiol!dioxygenases,!J$
Biol$Chem!260,!14035A14044.!
53.!Twilfer,!H.,!Bernhardt,!F.!H.,!and!Gersonde,!K.!(1985)!DioxygenAactivating!iron!center!in!
putidamonooxin.!Electron!spin!resonance!investigation!of!the!nitrosylated!
putidamonooxin,!European$journal$of$biochemistry$/$FEBS!147,!171A176.!
54.!Arciero,!D.!M.,!and!Lipscomb,!J.!D.!(1986)!Binding!of!17OAlabeled!substrate!and!
inhibitors!to!protocatechuate!4,5AdioxygenaseAnitrosyl!complex.!Evidence!for!direct!
substrate!binding!to!the!active!site!Fe2+!of!extradiol!dioxygenases,!J$Biol$Chem!261,!
2170A2178.!
55.!Nelson,!M.!J.!(1987)!The!nitric!oxide!complex!of!ferrous!soybean!lipoxygenaseA1.!
Substrate,!pH,!and!ethanol!effects!on!the!activeAsite!iron,!J$Biol$Chem!262,!12137A
12142.!

!

35!

56.!Chen,!V.!J.,!Orville,!A.!M.,!Harpel,!M.!R.,!Frolik,!C.!A.,!Surerus,!K.!K.,!Munck,!E.,!and!
Lipscomb,!J.!D.!(1989)!Spectroscopic!studies!of!isopenicillin!N!synthase.!A!
mononuclear!nonheme!Fe2+!oxidase!with!metal!coordination!sites!for!small!
molecules!and!substrate,!J$Biol$Chem!264,!21677A21681.!
57.!Rocklin,!A.!M.,!Tierney,!D.!L.,!Kofman,!V.,!Brunhuber,!N.!M.,!Hoffman,!B.!M.,!
Christoffersen,!R.!E.,!Reich,!N.!O.,!Lipscomb,!J.!D.,!and!Que,!L.,!Jr.!(1999)!Role!of!the!
nonheme!Fe(II)!center!in!the!biosynthesis!of!the!plant!hormone!ethylene,!Proc$Natl$
Acad$Sci$U$S$A!96,!7905A7909.!
58.!Tierney,!D.!L.,!Rocklin,!A.!M.,!Lipscomb,!J.!D.,!Que,!L.,!Jr.,!and!Hoffman,!B.!M.!(2005)!
ENDOR!studies!of!the!ligation!and!structure!of!the!nonAheme!iron!site!in!ACC!
oxidase,!J$Am$Chem$Soc!127,!7005A7013.!
59.!Cappillino,!P.!J.,!Miecznikowski,!J.!R.,!Tyler,!L.!A.,!Tarves,!P.!C.,!McNally,!J.!S.,!Lo,!W.!N.,!
Kasibhatla,!B.!S.!T.,!Krzyaniak,!M.!D.,!McCracken,!J.,!Wang,!F.,!Armstrong,!W.!H.,!and!
Caradonna,!J.!P.!(2012)!Studies!of!iron(II)!and!iron(III)!complexes!with!facAN2O,!cisA
N2O2!and!N2O3!donor!ligands:!models!for!the!2AHis!1Acarboxylate!motif!of!nonA
heme!iron!monooxygenases,!Dalton$Transactions!41,!5662A5677.!
!

!

36!

Chapter 2: Electron Paramagnetic Resonance Experiments
Section 2.1: Historical Perspective
In 1944, the first electron paramagnetic resonance was observed by E. K. Zavoisky at the
University of Kazan in the Soviet Union but the group was unable to develop the idea.

1

Independently in 1948, Brebis Bleany along with D. M. S. Bagguley, R. P. Penrose, J. H. E.
Griffiths, and B. I. Plumpton showed a clear electron spin resonance using chromium and copper
2

salts. Over the years the technique has been developed to investigate biological samples using a
range of methods. This chapter’s focus will be on the specific spin system and experiments used
in our paramagnetic resonance study of PheH.

Section 2.2: The Spin System
Pheynlalanine Hydroxylase, a non-heme pterin-dependent enzyme, forms a tetrameric
2+

complex that contains a single Fe

3

in each of the four functional domains. In nature, the

typical ligands surrounding the iron metal of non-heme enzyme complexes are usually ligated
through either nitrogen or oxygen atoms. From the crystal structures presented in chapter 1, the
ligation can easily be observed as N, N, and O from the ‘facial triad’ with up to three additional
O atoms from the water ligands.
2+

with the Fe

4-6

The electric field produced by the interaction of these ligands
7

is comparable to the field produced by the free ion. The result is an active site that
6

has a localized weak-field as defined by the d Tanabe-Sugano diagram, which means that the
6

metal ion takes on a high electron spin configuration. The high-spin iron is a d ion with four

!

37!

2+

unpaired electrons that result in a quintet spin multiplicity, S = 2. Even though the Fe

ion is

paramagnetic it is an electron paramagnetic resonance (EPR) ‘silent’ species at X-band (8.2-12.4
8

GHz).

2+

Treatment of the catalytic Fe

with nitric oxide, an O2 analogue, inhibits the oxygen

reactive non-heme iron enzymes. The inhibition provides a method to study enzymes in the
2+

presence of biologically relevant substrates. Binding of the NO to the Fe
characteristic CW-EPR response.

8, 9

metal produces a

The absorption signal from the CW-EPR experiments have

been attributed to an S = 3/2 species by the observation of spectral features ranging from about g
= 4 to g = 2. Utilizing the S = 3/2 spin system created by the formation of the iron-nitrosyl
complex provided a methodology that enabled the use of EPR spectroscopy to investigate a
range of non-heme iron enzymes.

8-16
7

The electronic description of the iron-nitrosyl complex, [FeNO] , comes from a
compilation of theoretical calculations and a range of experimental techniques.
7

17

Currently, the

3+

[FeNO] complex is best represented by an Fe , S = 5/2, antiferromagnetically coupled to an
-

NO , S = 1, forming an S = 3/2 spin system.

17, 18

Experimentally, the Fe K pre-edge feature from XAS spectra allow for the determination
of the oxidation state. The Fe K pre-edge features are a result of 1s!3d transitions. With the
1

final state, 1s 3d

n+1

2+

3+

, of the ferrous, Fe , and ferric, Fe , forms of iron being different, their Fe

K pre-edge should also be different. Differentiation between the two pre-edge features allows for
the detection of the oxidation state of the iron in the sample. The pre-edge features associated

!

38!

with the two different oxidation states will arise from the maximum spin multiplicities that can
be reached. Excitation of a 1s electron to the 3d orbitals of a high-spin ferrous and ferric iron
1

7

1

6

produce the 1s 3d and 1s 3d electronic states, respectively. The high-spin forms of these
7

electronic states respectively correspond to the quartet and quintet multiplicities. For a d ,
resulting from the ferrous iron, the spectroscopic states available with a quartet spin multiplicity
4

4

are F and P; these two states are split by an energy difference of 2 eV. The ferric form with a
5

quintet multiplicity only has the D spectroscopic state available. The Fe K edge XAS spectra of
2-

-

[Fe(H2O)EDTA] , [Fe(H2O)EDTA] , and FeEDTA-NO can be seen in figure 2.1 with the inset
2-

showing the pre-edge features. [Fe(H2O)EDTA] (solid line in figure 2.1) shows two peaks in

2-

Figure 2.1: Fe K XAS edge spectra of [Fe(H2O)EDTA] (solid), [Fe(H2O)EDTA](dashed), and FeEDTA-NO (dotted) with the inset showing the pre-edge feature.

!

39!

its pre-edge feature with maxima near 7111.5 eV and 7113.4 eV. The presence of two transition
-

peaks suggests that the sample contains a ferrous iron. Conversely, the [Fe(H2O)EDTA] and
FeEDTA-NO samples contain only one pre-edge feature at 7112.9 eV, which suggests both
samples contain ferric iron.

17

Density functional theory was utilized to study the electronic structure of the quartet ironnitrosyl complex formed at the catalytic site of taurine dioxygenase (TuaD), a non-heme iron
enzyme. Calculations utilizing the B3LYP hybrid functional predicted a bond length of 1.88 Å
7

18

between the Fe—N in TauD-[FeNO] and has an Fe—N—O bond angle of 144°.

A 1.54 Å

resolution x-ray crystal structure of the α-ketoglutarate depending non-heme iron enzyme,
clavamate synathase, in the presence of α-ketoglutarate (cofactor),
deoxyguanidinoproclavaminate (substrate), and nitric oxide depicted the Fe—N as 1.79 Å with
the bond angle between Fe—N—O as 146°.

19

The similarities between the experiment and

calculation are thought to be empirical proof that the DFT functional and basis sets provide an
accurate description of the spin system.

!

18

40!

Figure 2.2: A molecular orbital diagram of the spin system created within the TauD nonheme Iron catalytic center when complexed to NO.
Figure 2.2 shows the molecular orbital diagram of the spin system as predicted by the
DFT calculations. The calculation describes the spin system as having five spin-up singly
*

occupied molecular orbits surrounding the iron, with two spin-down electrons in the NO-π . The
*

large spatial overlap observed between the irons dxz and dyz with the NO-π (S = 0.6 - 0.78)
molecular orbitals indicates there is a partial covalent bond between the ferric iron and the nitric

!

41!

oxide.

18

The resulting S = 3/2 spin system is in full agreement with the XAS studies previously

discussed.

17
-

The high-spin ferric iron (S = 5/2) antiferromagnetically coupled to a triplet NO (S=1)
*

results in the S = 3/2 spin system as described previously. The coupling between NO-π

molecular orbitals and the ferric iron’s dxz and dyz leaves three singly occupied orbitals left in
the spin system. The calculation describes the singly occupied orbitals as one non-bonding (dxy)
and two σ-antibonding (dx2-y2 and dz2) orbitals that result in the majority of unpaired spin
density being located on the iron.

18

Section 2.3: Continuous Wave – Electron Paramagnetic Resonance (CW-EPR)
Electron paramagnetic resonance (EPR) spectroscopy provides an effective way to study
the S = 3/2 spin system created with PheH in the presence of nitric oxide. The interactions that
involve the iron-nitrosyl paramagnetic spin system are described by the spin Hamiltonian in
equation 2.1. The relative strength of these individual interactions described by the spin
Hamiltonian can be seen in figure 2.3. The first two terms of the spin Hamiltonian are the axial

β
g β
S(S + 1)
ˆ
ˆ
ˆ
ˆ
ˆ
ˆ ˆ
ˆ
H = D(S 2 −
) + E(S 2 − S 2 ) + e B ⋅ g ⋅ S − n n B ⋅ I + S ⋅ A ⋅ I
z
x
y
3
h
h
and rhombic zero field splitting (ZFS) terms, respectively.

20

(eq 2.1)
7

In the [FeNO] spin system the ZFS

term has the largest interaction, relative to the rest of the Hamiltonian, with D ≈ 300000 MHz

!

42!

-1 17, 18

(10 cm ).

The next largest interaction is the electronic Zeeman, which is the third term in

equation 2.1. Application of a 0.3 T external magnetic field to the iron-nitrosyl paramagnetic

7

Figure 2.3: Relative strength of the interactions that occur in the [FeNO] spin
system.
-1

spin system creates a splitting of ~10 GHz (0.33 cm ) between the ms = ±1/2 spin states as a
result of the electronic Zeeman interaction. Both the ZFS and electronic Zeeman interactions are
dependent upon the magnetic moment of the electron but reveal nothing about the surrounding
atoms. Investigating magnetic nuclei close to the spin system can help to reveal structural
characteristics about catalytic site in an enzyme. Information regarding magnetic nuclei close to
the paramagnetic spin center is encoded into the

!

43!

nuclear Zeeman and hyperfine interactions. These interactions are the fourth and fifth terms in
equation 2.1, respectively. The magnitudes of these interactions are approximately on the same
-3

-1

order of ~10 MHz (10 cm ), in a magnetic field of 0.3 T.!

21

Collecting information on the paramagnetic spin system typically starts with the
acquisition of a continuous wave electron paramagnetic resonance (CW-EPR) spectrum. CWEPR is a type absorption spectroscopy that utilizes microwave radiation to investigate a spin

Figure 2.4: The splitting of the S=1/2 spin system when an external magnetic field is
present. The arrow represents the absorption process and shows a Δms = +1 spin
transition.
system in the presence of an external magnetic field. Assuming a spin of S = 1/2, exposure to an
external magnetic field results in an energy separation between the spin states, figure 2.4, that is
described by the electronic Zeeman interaction. The electronic Zeeman interaction is the third
term in the spin Hamiltonian, equation 2.1. Evaluation of the electronic Zeeman interaction in
the spin basis, |S,ms>, for an S = 1/2 spin system leads to the matrix in equation 2.2. The energy

!

44!

⎛
⎜
⎜
⎜
⎜
⎝

1 1 ˆ 1 1
, H ,
2 2
2 2

1 1 ˆ 1 1
, H ,−
2 2
2 2

1 1 ˆ 1 1
,− H ,
2 2
2 2

1 1 ˆ 1 1
,− H ,−
2 2
2 2

⎞ ⎛ g β B
⎟ ⎜ e e o
2
⎟ =⎜
⎟ ⎜
⎟ ⎜
0
⎠ ⎜
⎝

⎞
⎟
⎟
g β B ⎟
− e e o ⎟
⎟
2
⎠
0

(eq 2.2)

difference, in frequency units, between the two spin states when the field is aligned along the zaxis is equal to geβeBo. The energy difference has a linear relationship to the external magnetic
field, Bo, meaning as the field increases so does the energy separation. When the energy
difference matches the energy of the radiation then a transition can occur. This absorption of the
radiation results in a CW-EPR spectrum.

20

The absorption process entails the spin system absorbing a photon causing a transition

β + hν → α

(eq 2.3)

α + hν → β + 2hν

(eq 2.4)

from the lower to the upper state, figure 2.4. However, the same quanta of energy can incite the
stimulated emission of a photon driving the system back to the ground state. These two processes
can be seen in equation 2.3 and 2.4, respectively. The rates associated with these processes in an
EPR experiment are tied to the transition
probability, equation 2.5. For the transition to occur, the resonant condition, represented by the δ-

2

Pαβ = σ α H ′ β δ (ΔE − hυ )

(eq 2.5)

function, must first be met. This means the radiation applied must have an energy that equals the

!

45!

separation of the energy levels associated to the transition. The probability of a transition is also
related to the modulus of the transition moment integral squared and the population (σ) of the
level. When the populations are equal the probabilities of the absorption and stimulated emission
are equivalent, which results in no signal and the sample is said to be in a state of ‘saturation.’
This implies that in order to acquire a signal from CW-EPR there must exist some population
difference between the α and β spin states. Based on the explanation it would seem that the spin
states will eventually reach a state of saturation, but as long as the relaxation rate is faster than
the transition rate the population difference will last. Increasing the power of the radiation
applied to the sample will increase the rate of transitions, which could eventually lead to a
saturated state.

20

The population difference between energy levels is described by the Boltzmann
distribution in equation 2.6. Assuming we have a two-level spin system, S = 1/2, in a 0.3 T
ΔE
N
k T
u =e B
N
l
−

(eq 2.6)

magnetic field, the upper (Nu) and lower (Nl) states refer to the ms = +1/2 and ms= -1/2 energy
levels, respectively. The energy separation, ΔE, between these ms levels is described by the
difference in electron Zeeman field interaction (figure 2.4). Using the energy level difference
(~10 GHz), Boltzmann constant (k=1.38·10

-23

J·s), and a temperature (T = 298 K) the ratio of

the populations comes out to 0.99984; which translates to the lower level, ms = -1/2, having an

!

46!

~0.008% excess population. Lowering the temperature to 4 K creates a larger population
difference, ~0.6% excess in the lower level, and directly corresponds to an increased signal
response. Addition of these excess magnetic moments divided by the sample volume results in
the net magnetization vector.
In the CW-EPR experiment, a spin system is exposed an external magnetic field that
defines the lab axis (z-axis). Using electromagnetic radiation drives either the absorption or
stimulated emission process, assuming the probability of the transition is not zero. The result is a
transition of the spin state, which has a probability that can be calculated using equation 2.5. The
probability amplitude, or transition moment integral, within equation 2.5 is written in Dirac
notation with the Hamiltonian representing the applied magnetic field attributed to the
electromagnetic radiation used. It is easily seen that if the Hamiltonian describing the transition
utilized the Sz operator the probability of the transition would be zero. Therefore, to bring about
the transition either the Sx and/or Sy operator(s) must be implemented in the Hamiltonian, which
means the radiation is applied perpendicular (i.e. in the xy-plane) to the external magnetic field.
The time-dependent Hamiltonian used in the transition moment integral is described by equation
ˆ
ˆ
ˆ
H ' = g β B [ S cos(ω t) + S sin(ω t)]
e e 1 x
o
y
o

(eq 2.7)

2.7. In the equation, ge represents the g-factor of a free electron (= 2.0023) and βe is the Bohr
magneton of the electron. The Hamiltonian describes the excitation magnetic field,

!

47!

B1, as rotating in the xy-plane at the Larmor frequency, ωo. The Larmor frequency is the angular
frequency attributed to the energy separation created by the electronic Zeeman interaction,
equation 2.8. Equation 2.7 describes a rotating magnetic field, B1, but a more accurate portrayal

g β B
ω = e e o
o


(eq 2.8)

of the instrumentation used in this work contains two counter-rotating fields to produce a linearly
oscillating magnetic field, say along the x-axis.

22

To effectively use the Hamiltonian described by equation 2.7, the Sx and Sy operators are
rewritten in terms of the raising and lowering operators, S+ and S- operators, equation 2.9. The
raising operator describes the transition from |S,ms> to the |S,ms+1> state, while the lowering
operator describes the transformation from the |S,ms> to |S,ms-1> state (equation 2.10).
ˆ
ˆ
ˆ
ˆ
S +S
S −S
ˆ
ˆ
S = + − and S = + −
x
y
2
2i

(eq 2.9)

ˆ
S S, m = S(S + 1) − m (m + 1) S, m + 1
+
s
s s
s
ˆ
S S, m = S(S + 1) − m (m − 1) S, m − 1
−
s
s s
s

(eq 2.10)

Evaluation of the transition moment integrals using the raising and lowering operators leads to
the formulation of the selection rules for EPR spectroscopy, Δms = ±1 and ΔmI = 0.

20, 23

Collecting a CW-EPR spectrum with a temperature of 4 K and microwave frequency of
7

9.68 GHz for the [FeNO] spin system results in figure 2.5. The modulated absorption signal

!

48!

ranges from approximately 170 mT to 350 mT. Using equation 2.11 to calculate the effective gvalue at the extremes of the spectrum it can be seen that the absorption peak ranges from an
effective g value of 2ge to ge, where ge represent the free electron spin g-factor (= 2.0023). The
absorption signal is not a δ-function, but it has some intrinsic linewidth associated with it, which

hν
g =
eff β B
e o

(eq 2.11)
7

is classified as homogenous broadening. However, the CW-EPR spectrum of the [FeNO] spin
system shows inhomogenous broadening that is a combined effect due to ZFS and the electronic
Zeeman interaction.

7

Figure 2.5: The CW-EPR spectrum for a [FeNO] spin system collected at a temperature
of 4 K and microwave frequency of 9.68 GHz.
!

49!

The range of the absorption signal observed in the CW-EPR spectrum of the S = 3/2 ironnitrosyl spin system can best be explained by evaluating the Hamiltonian with just the ZFS and
electronic Zeeman interaction, equation 2.12, in the spin basis. D and E are ZFS terms that are
defined by the principle values of the ZFS tensor [Dx,Dy,Dz]. Equation 2.12 has the electronic

S(S + 1) E ˆ 2 ˆ 2
ˆ
ˆ
ˆ
ˆ
ˆ
H = D(S 2 −
) + (S + S ) + g β B S + g β B S + g β B S
z
−
e e x x
e e y y e e z z
3
2 +
3
1
D = D and E = (D − D )
y
2 z
2 x

(eq 2.12)

Zeeman interaction expanded out into the canonical orientations because we are working within
the ZFS axis system. The spectrum above was
collected at 4 K, which creates a powder with
the spin systems distributed randomly. With the
ZFS interaction defining the coordinate system,
the external magnetic field is now thought of as
having multiple orientations, figure 2.6. The four
spin states create a 4x4 matrix in the spin basis
set, equation 2.13, and can be evaluated with the
magnetic field along the canonical orientations
Figure 2.6: The orientation of the external
magnetic field, Bo, within the zero-field
splitting axis system.

!

(x, y, and z).

50!

!
(eq 2.13)
!

!

51!

⎛
3g β B
⎜ D+ e e z
2h
⎜
⎜
⎜
0
⎜
⎜
⎜
3E
⎜
⎜
⎜
0
⎜
⎝

0

3E

g β B
−D + e e z
2h

0

0

g β B
−D − e e z
2h

3E

⎞
⎟
⎟
⎟
⎟
3E
⎟
⎟
⎟
0
⎟
⎟
3g β B ⎟
e e z
D−
⎟
2h
⎠
0

0

(eq 2.14)

For a compact view, the spin basis set, |S,ms>, is represented with just the ms value, |ms>. The
matrix (equation 2.13) evaluated with the magnetic field aligned along the z-axis is shown in
equation 2.14. If the number of unpaired electrons in the spin system is odd, like it is for the
7

[FeNO] system, then the spin states are split into doublets according to Kramer’s theorem by
the zero-field splitting interaction. The doublet degeneracy can be seen in equation 2.14 by just
looking at the components associated to the zero-field splitting, D and E. The energy difference
between the ms = ±1/2 and ms = ±3/2, assuming E ~ 0, is |2D|. The sign of D determines which

Figure 2.7: The axial zero field splitting interaction from a S = 3/2 spin
system.
!

52!

20, 24, 25

set of ms values make up the ground state, figure 2.7.

From Mössbauer spectroscopy

and temperature dependent magnetic susceptibility experiments the zero field splitting
-1

parameter, D, was determined to be approximately 12 cm , which is consistent with a ground
7

state of ms = ±1/2 for the [FeNO] spin system.

17, 18

As can be seen from equation 2.14, when

the external magnetic field is turned on, directed along the z-axis in the ZFS axis system, the
degeneracies are split by the electronic Zeeman interaction. However, the 2D energy splitting
between the Kramer’s doublets cannot be matched with the X-band frequencies used in our EPR
experiment, which without state mixing (E~0) allows us to focus on only the ms = ±1/2 spin
states. By condensing equation 2.14 to the ms = ±1/2 spin states you are left with the 2x2

⎛
g β B
⎜ −D + e e z
2h
⎜
⎜
⎜
0
⎜
⎝

⎞
⎟
0
⎟
g β B ⎟
−D − e e z ⎟
⎟
2h
⎠

(eq 2.15)

matrix in equation 2.15. The eigenvalues can be seen directly from the matrix because it is
already diagonalized and this leads to an energy splitting of geβeBz between the ms = +1/2 and
ms = -1/2 spin states. Although the external magnetic field is stationary, it can be thought of as
rotating within the ZFS axis system (figure 2.6). If the external magnetic field focused along the
x or y-axis relative to the ZFS axis system results in equation 2.16 and 2.17, respectively.
⎛
⎜
−D
⎜
⎜ g β B
⎜ e e x
⎜
h
⎝

!

g β B ⎞
e e x ⎟
h
⎟
⎟
⎟
−D
⎟
⎠

53!

(eq 2.16)

Determining the eigenvalues for thesematrices results in equation 2.18. Their respective energy

⎛
⎜
−D
⎜
⎜ g β B
⎜ e e y
⎜
h
⎝

g β B ⎞
e e y ⎟
⎟
h
⎟
⎟
−D
⎟
⎠

(eq 2.17)

g β B
2g β B
ν = −D ± e e x ⇒ Δν = e e x
±
h
h
g β B
2g β B
e e y
e e y
ν = −D ±
⇒ Δν =
±
h
h

(eq 2.18)

difference, in frequency units, is represented as Δν in the equations. From the frequency
difference formulations, it can be seenthat the effective g-value for the perpendicular axes, x and
y, is 2ge (~4). This means that the feature in the CW-EPR (figure 2.5) spectrum at g = 4
represents magnetic fields that are perpendicular to the Dz-axis. Conversely, the g = 2 feature
results from a situation where the external magnetic field is parallel to the Dz-axis. Therefore, the
ZFS and electronic Zeeman field interaction create a range of absorptions from g⟂ = gx = gy =
20, 24

2ge ~ 4 to g|| = gz = ge ~ 2.

As the rhombicity of the ZFS becomes non-zero, E≠0, the spectrum splits near the g = 4
feature, figure 2.8. Using X-band radiation (~9.68 GHz in our case) the value of D cannot be
sampled because the energy splitting is much larger than the energy provided by this radiation.
Although D is too large to measure directly, with X-band CW-EPR the ratio of E/D is available.
The top spectrum in figure 2.8 shows the g = 4 feature when E = 0, which means the ratio of E/D

!

54!

Figure 2.8: The effect of an E/D value that is not equal to zero on the g = 4 feature:
(top) E/D = 0; (bot) E/D = 0.1

!

55!

is also equal to 0. The bottom spectrum represents an E/D value of 0.1. Sometimes it is easier to
referrer to the splitting at the g = 4 feature as a splitting in the effective g-values. The canonical
effective g-values are listed in equation 2.19 and were derived from the energy difference

⎡
⎤
1+ 3 E D ⎥
⎢
g = g ⎢1 ±
⎥
x
e⎢
E 2⎥
1+ 3 D ⎥
⎢
⎣
⎦

( )
( )

⎡
⎤
1− 3 E D ⎥
⎢
g = g ⎢1 ±
⎥
y
e⎢
E 2⎥
1+ 3 D ⎥
⎢
⎣
⎦

( )
( )

(eq 2.19)

⎡
⎤
⎢
⎥
−2
g = g ⎢1 ±
⎥
z
e⎢
E 2⎥
1+ 3 D ⎥
⎢
⎣
⎦

( )

of the exact eigenvalues in the | D | >>

g β B
e e o

h limit. As you can see from the figure, an

E/D splitting can cause an appreciable change in the gx and gy values; however, gz is only
slightly changed from free electron g-value (2.0023), which is the reason figure 2.8 focuses on
26

the g⟂ region.

Section 2.4: Electron Spin Echo Envelope Modulation (ESEEM)
For paramagnetic transition metal ions with S > 1/2, the CW-EPR experiment is typically
7

dominated by the ZFS and electron Zeeman interactions. In the case of the [FeNO] spin system,
these interactions create an absorption peak that spans from g = 4 to g = 2.

10

These terms are

dependent on the electron spin and orbital angular momentum and provide little to no
!

56!

information regarding the ligand structure. To utilize EPR as a structural biology technique,
information pertaining to the surrounding ligands needs to be accessible, which is buried under
the features created by the ZFS and electronic Zeeman term in the CW-EPR spectrum. Pulsed
EPR techniques allow for the extraction of information related to the nuclear Zeeman and
Hyperfine interaction, which help tell the story about magnetic nuclei surrounding the spin
system. With the ability to obtain information about ligands, pulsed EPR provides a unique
structural perspective to study the catalytic domains of metalloenzymes.

23

The last two terms described by the spin Hamiltonian, equation 2.1, are the nuclear

g β
ˆ
ˆ ˆ
ˆ
H = − n n B⋅I + S⋅A⋅I
h

(eq 2.20)

Zeeman and hyperfine interactions, respectively. These terms have been rewritten in equation
2.20. The nuclear Zeeman interaction describes the coupling of the magnetic moments of the
nuclei to the external magnetic field, which results in an energy
separation between the nuclear states (αn and βn). In the future, the energy separation between
nuclear states will be translated to frequency units to keep the notation compact, equation 2.21.
The hyperfine coupling arises because of the interaction between the unpaired electron

ν =
n

Eβ n − Eα n
h

(eq 2.21)

spin and nuclear magnetic moment. It is useful to separate the hyperfine interaction into isotropic
and anisotropic terms, equation 2.22. The isotropic hyperfine is represented as aiso in the matrix

!

57!

⎛ a − T (1+ ρ )
0
0
⎜ iso
0
a − T (1− ρ )
0
A=⎜
iso
⎜
0
0
a + 2T
⎜
iso
⎝

⎞
⎟
⎟
⎟
⎟
⎠

(eq 2.22)

and is an interaction that occurs when the unpaired spin density is inside the nucleus, which
occurs in an S-type orbital.

23

This work will be focused on water ligands, which typically show

µ g β g β
T= o e e n n
4π
r3

(eq 2.23)

small aiso values because only a small fraction of unpaired spin character centered on the metal
1

ion reaches the H nucleus. The isotropic hyperfine coupling from protons bound to water
ligands is generally less than 1 MHz.

27

The anisotropic or ‘dipolar’ term, T, is shown in equation

2.23 and represents the magnitude of the interaction between the electron and nuclear magnetic
⎛ cos α cos β cos γ − sin α sin γ
⎜
R(α , β , γ ) = ⎜ − cos α cos β sin γ − sin α cos γ
⎜
cos α sin β
⎝

sin α cos β cos γ + cos α sin γ
− sin α cos β sin γ + cos α cos γ
sin α sin β

− sin β cos γ ⎞
⎟
sin β sin γ ⎟
⎟
cos β
⎠

(eq 2.23)
moment. ρ, in equation 2.22, is an asymmetry parameter that describes the rhombicity of the
dipolar hyperfine interaction.

21

If the hyperfine tensor is not collinear with the principle axis system, the hyperfine tensor
must be transformed into that axis system. In this work the principle axis system is defined by
the ZFS interaction, figure 2.6, because its z-axis has been shown to be nearly collinear to the

!

58!

28

molecular frame; Dz was found to be ~5° off the Fe-NO bond.

To rotate the hyperfine tensor

into the principle axis system a unitary transformation is preformed with the rotation operator in
equation 2.23. This rotation matrix is the combination of three orthogonal rotations and is

A Z = R(α , β , γ )A d R(α , β , γ )−1

(eq 2.24)

described using three angles (α, β, and γ), termed ‘Euler angles.’ To build this super-rotation
matrix the ‘zyz’ convention was used. Transforming the hyperfine tensor into the ZFS coordinate
Z

system results in A , and is obtained by carrying out the operation described by equation 2.24.
With the Fe-NO bond nearly collinear to the Dz axis, the angular dependence of the principle
hyperfine axis relative to the ZFS coordinate system has a physical significance.

⎛ a −T
⎜ iso
0
A=⎜
⎜
0
⎜
⎝

0
a −T
iso
0

⎞ ⎛ A
⎟ ⎜ ⊥
0
⎟ =⎜ 0
⎟ ⎜
a + 2T ⎟ ⎜ 0
iso
⎠ ⎝
0

0 ⎞
⎟
0 ⎟
⎟
A ⎟
 ⎠

0
A
⊥
0

23

(eq 2.25)

Lets illustrate this by rotating an axial hyperfine tensor, ρ = 0, represented in equation
2.25, into the ZFS axis system. By using an axial hyperfine tensor, one angle, β, is enough to
describe the rotation of the hyperfine tensor into the ZFS axis system (assuming E = 0). The

⎛
2
2
⎜ A⊥ cos β + A sin β
⎜
0
R(0, β ,0)−1 AR(0, β ,0) = ⎜
⎜
⎜ (A − A )cos β sin β

⊥
⎝

0
A
⊥
0

⎞
(A − A )cos β sin β ⎟

⊥
⎟
0
⎟
⎟
A sin 2 β + A cos2 β ⎟
⊥

⎠
(eq 2.26)

!

59!

other Euler angles are zero, which leads to the transformation of the axial hyperfine about the yaxis (equation 2.26). The choice of rotating the tensor about the y-axis is arbitrary, it can be done
⎛
2
0
⎜ aiso + T (3cos β − 1)
⎜
0
a −T
R(0, β ,0)−1 AR(0, β ,0) = ⎜
iso
⎜
3T cos β sin β
0
⎜
⎝

⎞
⎟
⎟
0
⎟
⎟
a + T (3cos2 β − 1) ⎟
iso
⎠
3T cos β sin β

(eq 2.27)
about any axis in the transverse (perpendicular) plane. By plugging in the definition of an axial
tensor leads to the equation 2.27, where the secular, Azz, and pseudo-secular, B, term can be
defined (equation 2.28).

23

A = a + T (3cos2 β − 1)
zz
iso
B = 3T cos β sin β

(eq 2.28)

The strength of the electronic Zeeman interaction is so much stronger than the hyperfine
interaction that only the secular operator, Sz, is used in the hyperfine Hamiltonian. However, the
hyperfine interaction mixes the nuclear spin states and forces the use of the entire nuclear spin
operator in the Hamiltonian (equation 2.29). Using the hyperfine tensor in equation 2.27 and the

ˆ
ˆ
S = [0,0, S ]
z
ˆ ˆ ˆ ˆ
I = [I , I , I ]
x y z

(eq 2.29)

definitions in equation 2.28, the Hamiltonian in equation 2.20 can be evaluated in the spin basis
to yield equation 2.30. In the matrix, νn represents the nuclear Lamor frequency, while the lx

!

60!

|α, αn>

<α, αn|
<α, βn|
<β, αn|
<β, αn|

⎛
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎜
⎝

|α, β n>
ν (l − il ) B
ν l A
n x y
− n z + zz
−
+
2
4
2
4
ν (l + il ) B
ν l A
n x y
n z − zz
−
+
2
4
2
4
0

0

0

0

|β, αn>

|β, βn>

⎞
⎟
0
0
⎟
⎟
⎟
0
0
⎟
⎟
ν (l − il ) B ⎟
ν l A
n x y
− n z − zz
−
− ⎟
2
4
2
4 ⎟
⎟
ν (l + il ) B
ν l A
n x y
n z + zz
⎟
−
−
⎟
2
4
2
4
⎠

(eq 2.30)

(cos θ cos ϕ), ly (cos θ sin ϕ), lz (cos θ) are the direction cosines describing the position of the
magnetic field within the ZFS coordinate system. The difference between the upper and lower
eigenvalues for the α-block (top left eq 2.30) and β-block (bottom right) are presented in
equations 2.31 and 2.32, respectively.!

23

2
A ⎞2 ⎛
⎛
B⎞2
ν = ⎜ ν l − zz ⎟ + ⎜ ν l − ⎟ + ⎛ ν l ⎞
⎝ n y⎠
α
⎝ n x 2⎠
2 ⎠
⎝ nz

(eq 2.31)

2
A ⎞2 ⎛
⎛
B⎞2
ν = ⎜ ν l + zz ⎟ + ⎜ ν l + ⎟ + ⎛ ν l ⎞
⎝ n y⎠
β
⎝ n x 2⎠
2 ⎠
⎝ nz

(eq 2.32)

Although the external magnetic field is stationary, by defining our principle axis system
as being collinear to the ZFS tensor the magnetic field takes on an orientation dependence that is
described by the direction cosines. If we say the field is parallel to the Dz-axis then lz = 1 and lx

!

61!

= ly = 0, which means equations 2.31 and 2.32 transform into equation 2.33 and 2.34,
respectively.

A ⎞ 2 B2
⎛
ν = ⎜ ν − zz ⎟ +
α
2 ⎠
4
⎝ n

(eq 2.33)

A ⎞ 2 B2
⎛
ν = ⎜ ν + zz ⎟ +
β
2 ⎠
4
⎝ n

(eq 2.34)

Similarly, the external magnetic field can be thought of as being parallel to the Dx and Dy-axis as
well; this corresponds to [lx, ly, lz] equaling [1, 0, 0] and [0, 1, 0], respectively. The orientation
dependence of the external

Figure 2.9: The ENDOR powder pattern observed for an S = 3/2 spin system simulated
with a microwave frequency of 9.68 GHz and a temperature of 4 K.

!

62!

magnetic field within the principle axis of the ZFS interaction leads to inhomogeneous
broadening due to the hyperfine interaction according to equation 2.31 and 2.32. If we assume
that β = 70° and T = 4.5 MHz, the powder pattern created by the orientation dependence of the
external magnetic field would result in the powder pattern observed in figure 2.9. The
inhomogeneous broadening of the individual hyperfine frequencies equation 2.31 and 2.32 can
be seen in figure 2.10. These spectra assume that all orientations of the external magnetic field
are accessible. However, the pulsed experiments used to measure the hyperfine parameters
are ran at a constant magnetic field strength, which results in the selection of only a small portion
of orientations; thereby, only portion of the spectra presented in figure 2.9 and 2.10 will be

Figure 2.10: The ENDOR powder pattern observed for an S = 3/2 spin system collected
with a microwave frequency of 9.68 GHz and a temperature of 4 K the transitions
separated.
!

63!

Figure 2.11: The 2-pulse ESEEM experiment.

present due to orientation selection.

23

Hyperfine frequencies can be measured using a pulsed EPR method known as Electron
Spin Echo Envelope Modulation (ESEEM). The one dimensional ESEEM experiment encodes
information about the hyperfine frequencies into
the amplitude of the spin echo produced by pulses
of microwave radiation. This pulsed experiment is
run at a constant external magnetic field. To get an
intuitive picture of the experiment it is typically
easiest to start with the 2-pulse experiment. The 2pulse ESEEM experiment has a pulse train that
consists of a π/2-pulse followed by a π-pulse with
a time separation of τ, figure 2.11. The result of
Figure 2.12: Microwave radiation causes a
rotation of the magnetization vector.

this pulse train is a spin echo that has a time

!

64!

separation of τ from the π-pulse. A simple vector analysis can help to illustrate the formation of a
spin echo.

23, 29

Before we complete a vector analysis we need to have a vector to discuss. In these
pulsed experiments the net magnetization vector, derived from the sum of the electron magnetic
dipole moments, is manipulated using nanosecond pulses of microwave radiation. The magnetic



field of the resonant microwave pulses, B1, rotates the net magnetization vector, M , through an
angle θ, figure 2.12. B1, is directed along the x-axis, which torques the magnetization vector,
initially aligned to the z-axis, to rotate towards the –y-axis. The angle of rotation, θ (equation

θ =ω t
1p

(eq 2.35)

2.35), is a product of the time the radiation is left on, tp, and the angular frequency, ω1, which is
proportional to the magnetic field strength of the radiation, B1. ω1 is termed the Rabi frequency.
As the rotation about the B1 axis occurs, the magnetization vector as a whole is torqued about the
external magnetic field, Bo. With the microwave radiation still on, both precessions, about Bo
and B1, occur simultaneously and result in the magnetization vector spiraling down towards the
–y-axis. To remove the added complication of the spiraling magnetization vector about the
external magnetic field, the axis system is rotated (about the z-axis) at the Larmor frequency
(ωo); thereby, creating what is known as the rotating frame. Within the rotating frame, the
applied magnetic field (B1) appears to be a simple vector directed along one axis, while the
23

magnetization vector appears to precess about B1 at the Rabi frequency (ω1).

!

65!

Figure 2.13: A vector description of the spin vectors as the experiment is carried out,
which leads up to the formation of a spin echo as the spin vectors collapse on the
detection axis (y-axis).
Lets now analyze the formation of the spin echo, figure 2.13, with this vector approach.
Start by assuming the sample contains an S = 1/2 coupled to an I = 1/2 spin. The first pulse in
the sequence is a π/2 (90°) pulse, which means that the nutation field from the pulse rotates the
magnetization through an angle equal to π/2. If the nutation field, B1, is along the x-axis the
magnetization vector will end up on the –y-axis, as shown above. After the pulse period, there is
a free precession time equal to τ where the magnetization vector starts to break up into spin
packets or groups of spins that ‘feel’ different effective magnetic fields, and precess at different
angular frequencies, ωi. These spin packets accumulate phase shifts, (ωi- ωo)τ, prior to
application of the second pulse. After this free precession time period, τ, the πx-pulse then flips
the spin packets by 180° about the x-axis. Assuming that the spin packets maintain the same
precession frequencies, ωi, they refocus along the +y-axis at time 2τ.

!

66!

23

The above description of electron spin echo formation leads to an echo signal that decays
*

with an increasing τ-value according to equation 2.36. In the equation, T2 is the spin-spin

e

2τ
T*
2

(eq 2.36)

relaxation for an inhomogenously broadened EPR spectrum. However, the anisotropy in the
hyperfine coupling that leads to inhomogeneous broadening also serves to mix the nuclear spin
states within each electron spin manifold. With an S = 1/2 coupled to an I = 1/2 spin system and
an external magnetic field strength of ~0.3 T, the electronic Zeeman interaction is so much larger

Figure 2.14: The four level energy diagram of an S = 1/2 coupled to an I = 1/2.
!

67!

than the hyperfine interaction that ms = 1/2 and ms = -1/2 are ‘good’ quantum numbers. The
jargon here just means there is essentially no state mixing in the electron spin states. The mixing
of the nuclear states in each manifold via the hyperfine interaction produces a probability
(equation 2.5) for the transition with the amplitude, |v|, that are no longer equal to zero—figure
2.14. With four accessible transitions, spin packets may have an accumulated phase offset, while
other spin packets refocus to form an echo along the detection axis. The projection of the offset
onto the detection axis manipulates the echo amplitude, which will modulate according to the
hyperfine frequencies as the pulse spacing (τ) changes.
To illustrate the echo modulation we can track a population of spin packets during the

Figure 2.15: A vector description of the spin packets going through the |1>!|3>
and |3>!|2> transitions during the 2-pulse ESEEM experiment. The end result is
an accumulated phase offset that is related to the hyperfine frequency (ϕ= ωβτ) from
the detection axis (y-axis).

!

68!

course the experiment, figure 2.15. Say we have a population of spins that make the transition
from |1>!|3> during the initial π/2-pulse and call their angular frequency ω13. During the free
precession period, the spin packets accumulate a phase offset, ϕ= (ω13- ωo)τ. If the π-pulse now
drives the |3>!|2> transition, the total phase offset after another free precession period equal to
τ is ϕ= (ω13- ω23)τ, which is equal to ϕ= ωβτ. Similarly, there will be some spin packets that
accumulate a phase relative to ωα. By changing the pulse spacing (τ), the phase offset changes
relative to the hyperfine frequencies. This causes the spin packet’s projection onto the detection
axis to modulate the spin echo amplitude. Tracing the modulation pattern and taking the Fourier
transform of this time domain signal results in a frequency domain spectrum that displays the

Figure 2.16: (Top) The 2-pulse ESEEM experiment creates an echo with a τ from the πpulse. As the τ value is incremented the echo height modulates according to the phase
offset of the hyperfine frequencies (ϕ=ωτ). (Bottom) Taking a Fourier transform of the
time trace acquired from the ESEEM experiment results in a frequency spectrum that
depicts the hyperfine frequencies centered about the nuclear Larmor.
!

69!

hyperfine frequencies (ωα and ωβ), figure 2.16.

23

The analytical form of the modulated spin

echo in the 2-pulse ESEEM experiment for a coupled S = 1/2, I = 1/2 can be seen in equation
2.37. It is easy to see from the expression that the signal will modulate according to the hyperfine
frequencies. In the analytical form, there is a k-parameter that represents the modulation depth

⎧2 − 2 cos(ωα τ ) − 2 cos(ω β τ )
⎫
⎪
k⎪
V = 1− ⎨
⎡⎛
⎞ τ ⎤ + cos ⎡⎛ ω − ω ⎞ τ ⎤ ⎬
2P
4 ⎪+ cos ω + ω
⎢⎝ α
⎢⎝ α
β⎠ ⎥
β ⎠ ⎥⎪
⎣
⎦
⎣
⎦⎭
⎩

(eq 2.37)

parameter and is tied to the B-term from the hyperfine interaction (equation 2.38). The inclusion

⎛ Bω ⎞
I ⎟
k=⎜
⎜ ωα ω β ⎟
⎝
⎠

2
(eq 2.38)

of the pseudo-secular B-term makes it clear that the nuclear states must be mixed in order to get
an ESEEM pattern.

23, 29

S =V
2P
2p

⎛
⎞
⎜ − 2τ ⎟
⎜ T*⎟
⎝ 2⎠
e

(eq 2.39)

The 2-pulse ESEEM spin echo decays exponentially (equation 2.39) because of spin*

spin relaxation and is described by the decay factor, T2 . Another type of relaxation is termed
spin-lattice relaxation and refers to the coupled interaction between the spin system and lattice.
In spin-lattice relaxation, fluxuating magnetic fields created by thermal motion of the lattice are

!

70!

Figure 2.17: The 3-pulse ESEEM experiment
coupled to a spin system and provide a mechanism for relaxation. In a powder sample the spinlattice relaxation time, T1, is typically at least an order of magnitude longer than the spin-spin
relaxation time, T2. To obtain the longest lasting ESEEM modulations a slow decay of the echo
is best, which means the optimum ESEEM experiment would decay according to the T1 value
rather than T2. Using the 2-pulse ESEEM experiment as a starting point, an ESEEM experiment
where the echo decays relative to the T1 value can be created. By replacing the π-pulse with two
π/2-pulses separated by an incremental time value, T, the T1-dependent ESEEM experiment is
formed, figure 2.17.
The 3-pulse ESEEM experiment produces an echo referred to as a stimulated echo that
decays exponentially relative to T1.The first two pulses in this experiment, separated by a
constant time τ, store information about the spin packets as projections along the lab field. The
information encoded into the spin packets is sampled with another π/2 pulse after a time of T. By
storing spin packet information about lab field, the decay of information is governed by T1. In
the following chapters, ESEEM will refer to the 3-pulse ESEEM experiment unless otherwise
noted. The analytical form of the 3-pulse ESEEM signal for an S = 1/2, I = 1/2 coupled spin

!

71!

⎧⎡
⎤ ⎫
⎤⎡
⎪ ⎣1− cos(ωα τ ) ⎦ ⎢1− cos(ω β (τ + T )) ⎥ ⎪
k⎪
⎣
⎦ ⎪
V = 1− ⎨
⎬
3P
4⎪ ⎡
⎤ ⎡1− cos(ω (τ + T )) ⎤ ⎪
+ 1− cos(ω τ ) ⎥
β ⎦⎣
α
⎦⎭
⎪ ⎣
⎪
⎩ ⎢

(eq 2.40)

system is presented in equation 2.40. The T1-dependent decay of the stimulated echo described
in equation 2.40 can be written down in an analogous fashion to equation 2.39.

23, 29, 30

When collecting 3-pulse ESEEM data, unwanted 2-pulse spin echoes can distort the
stimulated echo. These interferences produce glitches in the data that distort the modulated
signal. Figure 2.18 shows the unwanted echoes that are formed during the 3-pulse ESEEM
experiment. To remove these unwanted echoes the pulses are phase cycled. In our case the phase
program is implemented using a 4-channel bridge that is setup so the pulses have a relative phase
of 0°, 90°, 180°, or 270°, which are termed the x, y, -x, and -y axes—respectively. By varying
the phase of the first two pulses, all of the unwanted echoes phases are also manipulated.
Through addition or subtraction of the signals created from these phased experiments, the
unwanted echoes can be eliminated leaving only the stimulated echo. The 4-step phase cycling

Figure 2.18: The 3-pulse ESEEM experiment with the unwanted echoes.
!

72!

program used to eliminate the unwanted echoes in 3-pulse ESEEM is listed in table 2.1 and the
31

final signal, containing only the stimulated echo, is obtained by A-B-C+D.

Table 2.1: The relative phases of each echo during the phased experiments
Cycle

P1

P2

P3

SE

A

x

x

x

+1

+1

-1

+1

+1

B

x

-x

x

-1

+1

-1

-1

+1

C

-x

x

x

-1

-1

+1

+1

-1

D

-x

-x

x

+1

-1

+1

-1

-1

4

0

0

0

0

A-B-C+D

!

73!

UE1 UE2 UE3 UE4

Section 2.5: Hyperfine Sublevel COrrelation Experiment (HYSCORE)
The 3-pulse ESEEM experiment is a method that was shown to modulate according to the
hyperfine frequencies from coupled nuclei. However, when multiple magnetic nuclei are present
the spectrum quickly becomes congested, making it difficult to resolve individual and assign

Figure 2.19: The pulse train of the 4-pulse Hyperfine Sublevel Correlation Experiment
(HYSCORE).

hyperfine frequencies. To remove this congestion the experiment is converted into a twodimensional correlation experiment by splitting the incremental time period, T, into two by
!
Table!2.2:!43pulse!HYSCORE!Phase!Cycling!Program!
Cycle

P1

P2

P3

P4

x

x

x

x

+

B

x

x

x

-x

_

C

x

x

-x

x

+

D

x

x

-x

-x

_

is referred to as Hyperfine

!
!

The two-dimensional experiment

Signal

A

inserting a π-pulse—figure 2.19.

74!

Sublevel Correlation Experiment
(HYSCORE). By using the
mixing π-pulse the 4-pulse
echo’s phase is inverted.

23, 32

Analogous to the 3-pulse ESEEM experiment, this inverted echo is susceptible to distortion
through the combination of unwanted echoes with this 4-pulse echo. To remove the unwanted
echoes from the experiment the phase cycling program in table 2.2 was utilized. The ‘Signal’
column describes how the phased experiment are combined to obtain the final signal, which
33, 34

specifically for this experiment is A-B+C-D.

Figure 2.20: Correlation peaks (orange) resulting from the coupled hyperfine
frequencies from both electron spin manifolds on the anti-diagonal, which is centered
about the nuclear Larmor frequency on the diagonal.
!

75!

The two-dimensional 4-pulse HYSCORE experiment leads to a frequency spectrum that
correlates the hyperfine frequencies, ωα to ωβ, to one another—figure 2.20. These frequencies
are correlated on the line perpendicular to the diagonal with their bisect being the nuclear Larmor
frequency. This line is referred to as the anti-diagonal. To help visualize the correlation peaks of
the couplings, simplistic 3-pulse ESEEM spectra are shown as ‘skylines’ in the two dimensions.
Drawing lines directly from the hyperfine frequencies to find their bisect illustrates the
placement of the correlations within the HYSCORE spectrum. The separation of the hyperfine

Figure 2.21: The powder pattern spectrum shows broad correlation peaks (red) that
reflect the disorder of the orientations within the sample.
!

76!

frequencies based on the nuclei and strength of their coupling to the spin system is a powerful
aspect of HYSCORE.

23

The spectrum present in the figure 2.20 is the result of an ideal system with only one
orientation. In practice, working at 4 K creates a powder sample with a range of orientations
present. The powder pattern results in a spectrum that has correlated peaks spanning a range of
frequencies up to 3T/2, figure 2.21. This figure represents a situation
were the isotropic hyperfine coupling is larger than that of the dipolar hyperfine (|aiso|>|T|),
which creates a separation between the two peaks attributed to the different spin manifolds. In
this work, the dipolar hyperfine interaction is larger than the isotropic term. When the dipolar
term is larger than the isotopic hyperfine the correlated peaks cross through the nuclear Larmor
frequency as a result of powder pattern that resemble figure 2.9. For this reason, this discussion
will assume |aiso|>|T| to keep it clean.

23

If the dipolar coupling is large, then the pseudo-secular hyperfine interaction, B, is
‘turned on.’ Introduction of this term creates an asymmetric coupling about the nuclear
Larmor frequency (as seen in figure 2.10) in the ESEEM spectra described by equations 2.31 and
2.32. The asymmetry raises the correlations off of the anti-diagonal in the HYSCORE spectrum,
figure 2.22, which results in an arc-like structure. The maximum frequency shift between the arc
and the anti-diagonal is described by equation 2.41.

Δω =

!

9T 2
32 ω

23

(eq 2.41)

I

77!

With the powder pattern intact and isolated, it is easy to calculate the dipolar hyperfine
interaction and subsequently extract the structural information. However, remember these pulsed
experiments are being run at a constant external magnetic field. With a constant magnetic field
only certain orientations meet the resonance condition. Figure 2.23 shows the small population

Figure 2.22: The asymmetrical nature of the spectra create arc-like structures out of the
correlation peaks.

!

78!

Figure 2.23: The limited number of orientations that meet the resonance condition
(orange) result in correlated peaks with only a small portion of the arc being visible.
of spins that meet the resonance condition (orange) and results only in a small section of the
powder pattern signal being visible in the HYSCORE spectrum. It is for this reason that when
determining the dipolar coupling the spectra must be collected at multiple field strengths.

23

In the rest of this work, these EPR experiments will be used to collect information
7

regarding the coupling of the spin system, [FeNO] , to magnetic nuclei. Specifically, protons

!

79!

bound to ligated water molecules, which will help to determine the role water plays in the
coupled hydroxylation carried out by PheH.

!

80!

BIBLIOGRAPHY

!

81!

BIBLIOGRAPHY

1.!Elliott,!R.!(2009)!Brebis!Bleaney.!6!June!1915!—!4!November!2006,!Biographical+Memoirs+
of+Fellows+of+the+Royal+Society!55,!3311.!
2.!Bagguley,!D.!M.!S.,!Bleaney,!B.,!Griffiths,!J.!H.!E.,!Penrose,!R.!P.,!and!Plumpton,!B.!I.!(1948)!
Paramagnetic!Resonance!in!Salts!of!the!Iron!Group!3!A!Preliminary!Survey:!I.!
Theoretical!Discussion,!Proceedings+of+the+Physical+Society!61,!542.!
3.!Fitzpatrick,!P.!F.!(2012)!Allosteric!regulation!of!phenylalanine!hydroxylase,!Archives+of+
biochemistry+and+biophysics!519,!1943201.!
4.!Andersen,!O.!A.,!Stokka,!A.!J.,!Flatmark,!T.,!and!Hough,!E.!(2003)!2.0A!resolution!crystal!
structures!of!the!ternary!complexes!of!human!phenylalanine!hydroxylase!catalytic!
domain!with!tetrahydrobiopterin!and!33(23thienyl)3L3alanine!or!L3norleucine:!
substrate!specificity!and!molecular!motions!related!to!substrate!binding,!J+Mol+Biol!
333,!7473757.!
5.!Andersen,!O.!A.,!Flatmark,!T.,!and!Hough,!E.!(2002)!Crystal!structure!of!the!ternary!
complex!of!the!catalytic!domain!of!human!phenylalanine!hydroxylase!with!
tetrahydrobiopterin!and!33(23thienyl)3L3alanine,!and!its!implications!for!the!
mechanism!of!catalysis!and!substrate!activation,!J+Mol+Biol!320,!109531108.!
6.!Andersen,!O.!A.,!Flatmark,!T.,!and!Hough,!E.!(2001)!High!resolution!crystal!structures!of!
the!catalytic!domain!of!human!phenylalanine!hydroxylase!in!its!catalytically!active!
Fe(II)!form!and!binary!complex!with!tetrahydrobiopterin,!J+Mol+Biol!314,!2793291.!
7.!Solomon,!E.!I.,!Pavel,!E.!G.,!Loeb,!K.!E.,!and!Campochiaro,!C.!(1995)!Magnetic!circular!
dichroism!spectroscopy!as!a!probe!of!the!geometric!and!electronic!structure!of!non3
heme!ferrous!enzymes,!Coordin+Chem+Rev!144,!3693460.!
8.!Rich,!P.!R.,!Salerno,!J.!C.,!Leigh,!J.!S.,!and!Bonner,!W.!D.!(1978)!Spin!332!Ferrous3Nitric!
Oxide!Derivative!of!an!Iron3Containing!Moiety!Associated!with!Neurospora3Crassa!
and!Higher!Plant3Mitochondria,!Febs+Letters!93,!3233326.!
9.!Arciero,!D.!M.,!Lipscomb,!J.!D.,!Huynh,!B.!H.,!Kent,!T.!A.,!and!Munck,!E.!(1983)!EPR!and!
Mossbauer!studies!of!protocatechuate!4,53dioxygenase.!Characterization!of!a!new!
Fe2+!environment,!J+Biol+Chem!258,!14981314991.!

!

82!

10.!Arciero,!D.!M.,!Orville,!A.!M.,!and!Lipscomb,!J.!D.!(1985)![17O]Water!and!nitric!oxide!
binding!by!protocatechuate!4,53dioxygenase!and!catechol!2,33dioxygenase.!Evidence!
for!binding!of!exogenous!ligands!to!the!active!site!Fe2+!of!extradiol!dioxygenases,!J+
Biol+Chem!260,!14035314044.!
11.!Twilfer,!H.,!Bernhardt,!F.!H.,!and!Gersonde,!K.!(1985)!Dioxygen3activating!iron!center!in!
putidamonooxin.!Electron!spin!resonance!investigation!of!the!nitrosylated!
putidamonooxin,!European+journal+of+biochemistry+/+FEBS!147,!1713176.!
12.!Arciero,!D.!M.,!and!Lipscomb,!J.!D.!(1986)!Binding!of!17O3labeled!substrate!and!
inhibitors!to!protocatechuate!4,53dioxygenase3nitrosyl!complex.!Evidence!for!direct!
substrate!binding!to!the!active!site!Fe2+!of!extradiol!dioxygenases,!J+Biol+Chem!261,!
217032178.!
13.!Nelson,!M.!J.!(1987)!The!nitric!oxide!complex!of!ferrous!soybean!lipoxygenase31.!
Substrate,!pH,!and!ethanol!effects!on!the!active3site!iron,!J+Biol+Chem!262,!121373
12142.!
14.!Chen,!V.!J.,!Orville,!A.!M.,!Harpel,!M.!R.,!Frolik,!C.!A.,!Surerus,!K.!K.,!Munck,!E.,!and!
Lipscomb,!J.!D.!(1989)!Spectroscopic!studies!of!isopenicillin!N!synthase.!A!
mononuclear!nonheme!Fe2+!oxidase!with!metal!coordination!sites!for!small!
molecules!and!substrate,!J+Biol+Chem!264,!21677321681.!
15.!Rocklin,!A.!M.,!Tierney,!D.!L.,!Kofman,!V.,!Brunhuber,!N.!M.,!Hoffman,!B.!M.,!
Christoffersen,!R.!E.,!Reich,!N.!O.,!Lipscomb,!J.!D.,!and!Que,!L.,!Jr.!(1999)!Role!of!the!
nonheme!Fe(II)!center!in!the!biosynthesis!of!the!plant!hormone!ethylene,!Proc+Natl+
Acad+Sci+U+S+A!96,!790537909.!
16.!Tierney,!D.!L.,!Rocklin,!A.!M.,!Lipscomb,!J.!D.,!Que,!L.,!Jr.,!and!Hoffman,!B.!M.!(2005)!
ENDOR!studies!of!the!ligation!and!structure!of!the!non3heme!iron!site!in!ACC!
oxidase,!J+Am+Chem+Soc!127,!700537013.!
17.!Brown,!C.!A.,!Pavlosky,!M.!A.,!Westre,!T.!E.,!Zhang,!Y.,!Hedman,!B.,!Hodgson,!K.!O.,!and!
Solomon,!E.!I.!(1995)!Spectroscopic!and!Theoretical!Description!of!the!Electronic3
Structure!of!S=3/2!Iron3Nitrosyl!Complexes!and!Their!Relation!to!O32!Activation!by!
Nonheme!Tron!Enzyme!Active3Sites,!Journal+of+the+American+Chemical+Society!117,!
7153732.!

!

83!

18.!Ye,!S.,!Price,!J.!C.,!Barr,!E.!W.,!Green,!M.!T.,!Bollinger,!J.!M.,!Jr.,!Krebs,!C.,!and!Neese,!F.!
(2010)!Cryoreduction!of!the!NO3adduct!of!taurine:alpha3ketoglutarate!dioxygenase!
(TauD)!yields!an!elusive!{FeNO}(8)!species,!J+Am+Chem+Soc!132,!473934751.!
19.!Zhang,!Z.!H.,!Ren,!J.!S.,!Harlos,!K.,!McKinnon,!C.!H.,!Clifton,!I.!J.,!and!Schofield,!C.!J.!(2002)!
Crystal!structure!of!a!clavaminate!synthase3Fe(II)323oxoglutarate3substrate3NO!
complex:!evidence!for!metal!centred!rearrangements,!Febs+Letters!517,!7312.!
20.!Atherton,!N.!M.!(1993)!Principles+of+electron+spin+resonance,!Ellis!Horwood!:!PTR!
Prentice!Hall,!New!York.!
21.!Atherton,!N.!M.,!and!Horsewill,!A.!J.!(1979)!Proton!ENDOR!of!Cu(H2O)6!2+!in!
Mg(NH4)2(SO2)4.6H2O,!Molecular+Physics!37,!134931361.!
22.!Wertz,!J.!E.,!and!Bolton,!J.!R.!(1972)!Electron+spin+resonance;+elementary+theory+and+
practical+applications,!McGraw3Hill,!New!York,.!
23.!Schweiger,!A.,!and!Jeschke,!G.!(2001)!Principles+of+pulse+electron+paramagnetic+
resonance,!Oxford!University!Press,!Oxford.!
24.!Boca,!R.!(2004)!Zero3field!splitting!in!metal!complexes,!Coordin+Chem+Rev!248,!7573815.!
25.!Drago,!R.!S.!(1992)!Physical+methods+for+chemists,!2nd!ed.,!Saunders!College!Pub.,!Ft.!
Worth.!
26.!Pilbrow,!J.!R.!(1978)!Effective!G3Values!for!S!=!3/2!and!S!=!5/2,!J+Magn+Reson!31,!4793
490.!
27.!Maurelli,!S.,!Ruszak,!M.,!Witkowski,!S.,!Pietrzyk,!P.,!Chiesa,!M.,!and!Sojka,!Z.!(2010)!
Spectroscopic!CW3EPR!and!HYSCORE!investigations!of!Cu(2+)!and!O(2)(3)!species!in!
copper!doped!nanoporous!calcium!aluminate!(12CaO.7Al(2)O(3)),!Phys+Chem+Chem+
Phys!12,!10933310941.!
28.!Aquino,!F.,!and!Rodriguez,!J.!H.!(2009)!Accurate!Calculation!of!Zero3Field!Splittings!of!
(Bio)inorganic!Complexes:!Application!to!an!{FeNO}(7)!(S=3/2)!Compound,!Journal+
of+Physical+Chemistry+A!113,!915039156.!

!

84!

29.!Mims,!W.!B.!(1972)!Amplitudes!of!Superhyperfine!Frequencies!Displayed!in!the!
Electron3Spin3Echo!Envelope,!Physical+Review+B!6,!354333545.!
30.!Mims,!W.!B.!(1972)!Envelope!Modulation!in!Spin3Echo!Experiments,!Physical+Review+B!
5,!240932419.!
31.!Fauth,!J.!M.,!Schweiger,!A.,!Braunschweiler,!L.,!Forrer,!J.,!and!Ernst,!R.!R.!(1986)!
Elimination!of!unwanted!echoes!and!reduction!of!dead!time!in!three3pulse!electron!
spin3echo!spectroscopy,!Journal+of+Magnetic+Resonance+(1969)!66,!74385.!
32.!Shane,!J.!J.,!Höfer,!P.,!Reijerse,!E.!J.,!and!de!Boer,!E.!(1992)!Hyperfine!sublevel!correlation!
spectroscopy!(HYSCORE!)!of!disordered!solids,!Journal+of+Magnetic+Resonance+
(1969)!99,!5963604.!
33.!Gemperle,!C.,!Aebli,!G.,!Schweiger,!A.,!and!Ernst,!R.!R.!(1990)!Phase!cycling!in!pulse!EPR,!
Journal+of+Magnetic+Resonance+(1969)!88,!2413256.!
34.!Stoll,!S.,!and!Schweiger,!A.!(2006)!EasySpin,!a!comprehensive!software!package!for!
spectral!simulation!and!analysis!in!EPR,!J+Magn+Reson!178,!42355.!
!

!

85!

Chapter 3: Qualitative Analysis of the Allosteric Forms of Phenylalanine Hydroxylase
Section 3.1: Introduction
Phenylalanine Hydoxylase uses tetrahydrobiopterin and molecular oxygen to catalyze the
1

conversion of L-phenylalanine to L-tyrosine. Phenylalanine allosterically regulates the enzyme
leading to a global conformational change that results in ligand rearrangement at the catalytic
2, 3

site.

As stated in section 4 and 5 of chapter 1, it is unclear if the rearrangement dehydrates the

active site, which has lead to some confusion about the role water plays in the formation of the
high-valent Fe(IV)-oxo species.
(ESEEM)

6, 7

4, 5

In this chapter, Electron Spin Echo Envelope Modulation
8

7

and Hyperfiine Sublevel Correlation (HYSCORE) studies of a series of [FeNO] 9

N2O1-3 model complexes were used to develop a means for detecting and quantifying water
ligands bound to Fe(II). This methodology was then extended to studies of recombinant rat
Phenylalanine Hydroxylase (rPheH) to reveal the presence of two water ligands bound to the
7

[FeNO] paramagnetic center in the active site of the unactivated, or allosteric T-(Tense) state of
the enzyme. Further addition of substrate, L-phe, to poise the enzyme in its fully active, or
allosteric R-(Relaxed) state, resulted in the loss of one bound water ligand. Finally, addition of
7

the cofactor 5-deaza-6-methyltetrahydropterin to activated, R-state, [FeNO] -PheH, resulted in
the loss of the remaining water ligand leaving the metal site void of bound water.

!

86!

Section 3.2: Materials and Methods
Our collaborators at Boston University under the direction of Dr. John P. Caradonna
made the samples. Their preparation is cataloged in appendix 1.
Continuous Wave (CW) and pulsed EPR experiments were performed with a Bruker E680X spectrometer operating at X-band. Measurements were done using an ER4118X-MD-5W1 probe equipped with a 5mm dielectric resonator. An Oxford Instruments, liquid helium
temperature control system equipped with a CF-935 cryostat and an ITC-503 temperature
controller was used to maintain the sample temperature at 4 K. For CW-EPR experiments, data
were collected using the following conditions: microwave frequency, 9.68 GHz; microwave
power, 6.3 µW; field modulation frequency, 10 kHz; and field modulation amplitude, 0.8 mT.
CW-EPR simulations utilized the ‘Pepper’ module within the EasySpin package

10

for MATLAB

(The Mathworks, Natick, MA).
6

3-pulse ESEEM experiments (90°-t-90°-T-90°) utilized a 4-step phase cycle

11

(Table

2.1) to eliminate unwanted 2-pulse echoes and DC offset voltages from the stimulated echo. The
acquisition of ESEEM data included: microwave frequency, 9.68 GHz; 90° pulse length, 16ns
full-width at half maximum; pulse power, 160W; time increment, 12 ns; and dataset length, 512
points. ESEEM frequency spectra were obtained by a procedure that involved: (1) subtraction of
the background decay; (2) the application of a Hamming Window function; (3) zero-filling to
1024 points and (4) Fourier transformation. Absolute value spectra are displayed. 4-Pulse
HYSCORE experiments (90°-t-90°-T1-180°-T2-90°) were performed using similar conditions to

!

87!

those given above for 3-pulse ESEEM. Data acquisition in the T1 and T2 domain was limited to
128 points collected at 16 ns intervals. A 4-step phase cycle was used (Table 2.2) and data were
processed in the same manner as provided above for 1-dimensional ESEEM experiments prior to
12

a 2-D FFT.

t -values for both 3-pulse ESEEM and 4-pulse HYSCORE measurements were

restricted to the first or second harmonic of the proton Larmor frequency for each magnetic field
strength studied owing to rapid T2* values.

Section 3.3: Results and Analysis
Section 3.3.1: Model Complexes

Figure 3.1: The three model complexes (a) Fe(II)(N2O3)(L), (b) Fe(II)(N2O2)(L)2,
and (c) Fe(II)(N2O)(L)3
9

The ferrous model complexes studied in this work are illustrated in figure 3.1.
Abbreviated forms of the major ligands are used to reference the individual ferrous complexes:
(figure 3.1a) 2,2′-(2-((carboxymethyl)-(methyl)amino) ethyl-azanediyl)diacetic acid (N2O3);
(3.1b) 2,2′-(ethane-1,2-diylbis (methyl-azanediyl))diacetic acid (N2O2); and (3.1c) 2-((2-

!

88!

(dimethyl-amino)ethyl)(methyl)amino)acetic acid (N2O1). The labile coordination sites are
labeled L in the figure with the N, N, O facially ligated (red). Nitric oxide occupies
one of the labile sites, which leaves the others open to exchange with the solvent, H2O.

7

Figure 3.2: The CW-EPR spectrum of [FeNO] -(N2O2)(H2O) (blue trace) collected at
4.0 K using a microwave frequency of 9.68 GHz. The green trace is a computer
simulation of the experimental spectrum using an E/D value of 0.0121. B1 marks the
field value, 240 mT, where the pulsed EPR data shown in this manuscript were
collected. (Inset) The feature near g = 4 is magnified to highlight the differences
between the simulation and the data
!

89!

Probing the N2O2 ferrous iron-nitrosyl complex with CW-EPR results in the green
spectrum presented in figure 3.2. The observed spectrum is typical for high spin
S = 3/2 centers dominated by zero field splitting and described by the Hamiltonian in equation
2.7.

13

The presence of a large D-value leads to CW-EPR spectra that arise from the lowest

energy Kramer’s doublet, previously described in chapter 2. The CW-EPR spectrum at the
microwave frequency used in our measurements is only sensitive to the ratio of E/D. For spectral
-1

simulations, the D value was set to 10 cm commensurate with values derived from
7

experimental and theoretical studies of other [FeNO] model complexes.

7

14

Parameters obtained

Figure 3.3: The 3-Pulse ESEEM spectrum of [FeNO] -(N2O2)(H2O) collected at 240mT with
a τ = 96 ns and a sample temperature of 4K.

!

90!

from the simulation, blue line in figure 3.2, included an axial g-tensor ([g⊥ g||]), [2.015 2.0023],
7

and an E/D value of 0.0121 for the [FeNO] -N2O2 complex. Small differences between the
N2O2 data and the simulation were found in the g = 4 region of the EPR spectrum and are shown
in the upper right corner of figure 3.2. A better fit can be achieved by considering the spectrum
to be a composite of two different paramagnetic complexes characterized by E/D values of
0.0186 and 0.0118 present in a 3:1 ratio. Because the effect of this modest E/D difference on the
analysis of ESEEM and HYSCORE spectra is negligible, the speciation is assumed to be
7

7

singular for the purpose of this work. Similar analyses of [FeNO] -N2O1 and [FeNO] -N2O3
complexes resulted in E/D values of 0.0083 and 0.0123, respectively.
7

A 3-pulse ESEEM spectrum collected for the [FeNO] -(N2O2)(H2O) model complex at
240mT, B1 from figure 3.2, is shown in figure 3.3. The spectrum shows a complicated mixture
of peaks from the coupled

14

1

N and H nuclei of the N2O2 and water ligands. The corresponding

4-pulse HYSCORE spectrum taken at the same magnetic field strength and τ-value is shown in
figure 3.4. Cross-peaks (outlined red) symmetric about the frequency diagonal are resolved at
(7.9 MHz, 13.1 MHz) and (12.7 MHz, 8.2 MHz); and (5.7 MHz, 17.6 MHz) and (17.6 MHz, 5.9
MHz). Because these cross-peaks are centered about the proton Larmor frequency of 10.2 MHz
1

they are assigned to H couplings. Off-diagonal correlations representative of

14

N coupling(s)

were not resolved at this field position and τ-value in either the (+,+) or (-,+) quadrants. The
strength of anisotropic hyperfine coupling, or “dipolar” coupling, can be gauged for these two
correlations by their position relative to a line that can be constructed perpendicular to the

!

91!

7

Figure 3.4: The HYSCORE spectra of [FeNO] -(N2O2)(H2O) collected at 240mT with τ =
1
96 ns and a sample temperature of 4K. The H correlations related to bound water are
outlined.
1

frequency diagonal that runs through the H Larmor frequency (an “anti-diagonal”), and by the
1

magnetic field dependence of their cross-peaks. Of the two different H correlations observed,
1

1

the set resolved at is substantially raised off of the H-anti-diagonal. This H correlation is also
characterized by a magnetic field dependence showing broader arcs that are barely resolved
above the noise floor at 220 mT, which travel along an arc-like trajectory that leads them into the

!

92!

frequency diagonal as the magnetic field strength is increased to 320 mT. This large anisotropy
1

in the H hyperfine interaction is typical of bound water and likely results from the proximity of
1

the ligand protons to the metal ion. The H correlation resolved at (8.0 MHz, 12.9 MHz) is
visible across the entire EPR spectrum and shows a modest dependence on magnetic field
strength commensurate with the change in 1H Larmor frequency.
1

7

7

Parallel H HYSCORE studies were done on [FeNO] -(N2O)(H2O)2 and [FeNO] -N2O3

7

Figure 3.5: The HYSCORE spectra of [FeNO] -(N2O)(H2O)2 collected at 240mT with τ =
1
96 ns and a sample temperature of 4K. The H correlations related to bound water are
outlined.
!

93!

7

Figure 3.6: The HYSCORE spectra of [FeNO] -(N2O3) collected at 240mT with τ = 96
1
ns and a sample temperature of 4K. The H correlations related to bound water are
outlined.
model complexes and HYSCORE spectra collected at 240 mT with τ = 96 ns are shown in
figures 3.5 and 3.6, respectively. The proton correlation centered at (5.9, 17.6 MHz) with a
7

contour shape indicative of strong dipolar hyperfine couplingis clearly resolved for [FeNO] (N2O)(H2O)2 (figure 3.6). The cross-peaks arising from this coupling cover a larger frequency
7

range and show an increased intensity relative to the intensity resolved for the [FeNO] -

!

94!

(N2O2)(H2O) model (fig 3.3). These differences, boxed in figure 3.3 and 3.5, are commensurate
7

with the presence of an additional bound water molecule in the [FeNO] -(N2O)(H2O)2 complex.
These cross-peaks were found to exhibit the same magnetic field dependence as described above
7

7

for the [FeNO] -(N2O2)(H2O) complex. The 4-pulse HYSCORE spectrum of [FeNO] -N2O3
1

(figure 3.6) collected at 240 mT is characterized by the absence of cross-peaks due to the H
correlation at (5.9, 17.6 MHz) and verifies our assignment of this high-frequency correlation to
1

H couplings arising from bound water.

7

Figure 3.7: The ESEEM spectra of the three model complexes: (fine dash) [FeNO] 7
7
(N2O3), (dash) [FeNO] -(N2O2)(H2O), and (solid) [FeNO] -(N2O)( H2O)2 collected at
240mT, τ = 96 ns, and sample temperature of 4K.

!

95!

Figure 3.7 shows corresponding 3-pulse ESEEM spectra for the three model complexes
collected using the same magnetic field strength and τ-value as used for the HYSCORE studies
above. The cross-peaks resolved in the HYSCORE spectra that were assigned to bound water
should give rise to peaks in the corresponding 3-pulse ESEEM spectra at approximately 4-6
MHz and 17-18.5 MHz. Figure 3.7 shows that the 4-6 MHz region is too congested to enable the
extraction of useful amplitude information. However, the intensity of the higher frequency peak
near 18 MHz should fall in a clean spectral window and allow 3-pulse ESEEM amplitudes to be
used to quantify bound water. The overlaid spectra of figure 3.7 shows the expected trend with
7

the solid trace due to [FeNO] -N2O showing the largest intensity, while spectrum arising from
7

the [FeNO] -N2O2 sample has an intermediate ESEEM signal strength at 18 MHz. The spectrum
7

from the [FeNO] -N2O3 complex in figure 3.6 shows no signal at 18 MHz. This observation
confirms the peak assignment to bound water and demonstrates a clean spectral window in the
16-18 MHz range that should aid in determining the number of waters bound to the [FeNO]
center of PheH.

!

96!

7

T

R

Section 3.3.2: PheH [L-phe] vs PheH [L-phe]

T

Figure 3.8: Above is the g = 4 spectral features from the CW-EPR spectra of (a) PheH [LR

phe] and (b) PheH [L-phe] collected at 4K.
!

97!

2

Activation of PheH induces a global conformational change that is thought to result in a
15

ligand rearrangement at the Fe(II) site . To study the active site rearrangement, the wild-type
rPheH was studied in the presence of the activator, L-phe. The g = 4 features of the CW-EPR
7

T

R

spectra from the [FeNO] forms of PheH [L-phe] and PheH [L-phe] are shown in figure 3.8a
and 3.8b, respectively. The similarities between the spectra suggest that the parameters from the
7

individual [FeNO] centers are consistent. At low concentrations of L-phe the enzyme is in the
T-state, figure 3.8a, and is best described by a 2:1 mixture of species with E/D values of 0.0330

T

1

Figure 3.9: HYSCORE spectrum of PheH [L-phe] with two H cross-peaks outlined.

!

98!

R

and 0.0250. The activation produces PheH [L-phe], figure 3.8b, with a speciation containing a
1:1 ratio with E/D values of 0.0330 and 0.0263. Speciation between the samples appears to be
approximately identical; however, the mixture ratios differ and that maybe due in part to the
3+

presence of Fe

in the R-state sample, which is indicated by the peak at g = 4.3 in figure 3.8b.

Figure 3.9 and 3.10 show 4-pulse HYSCORE spectra collected at 240 mT using a τ-value
T

R

1

of 96 ns for PheH [L-phe] and PheH [L-phe], respectively. Both samples show two pairs of H
T

cross-peaks arising from strong dipolar couplings. The first set of cross-peaks in the PheH [L-

R

1

Figure 3.10: HYSCORE spectrum of PheH [L-phe] with two H cross-peaks outlined.

!

99!

phe] HYSCORE spectrum are centered near (4.6 MHz, 17.3 MHz) and (17.3 MHz, 4.5 MHz)
1

and outlined with a red box in figure 3.9. The second set of H cross-peaks are outlined with a
black box in figure 3.9 and are centered at (7.6 MHz, 17.5 MHz) and (17.6 MHz, 7.6 MHz). The
cross-peaks attributed to the water ligand in the model complexes, figures 3.5 and 3.4, show
broad contours characterized by complex shapes with the largest cross-peak intensities centered
near (6 MHz, 18 MHz) and (18 MHz, 6 MHz) at 240 mT. These contours span frequency ranges
from 3-8 MHz and 16-18.5 MHz. The breadth of the cross-peak contours resolved for bound
water ligands of the model complexes encompasses both sets of cross-peaks observed in the
HYSCORE spectrum of the T-state PheH complex, figure 3.11. This suggests that each crosspeak resolved for the enzyme sample may belong to different protons attached to the same
ligated water molecule(s). Resolution of two distinct protons in the enzymes suggests that the
orientations of the ligand water(s) are limited to a small subset of the orientations that are
available to the water ligand protons of the model complexes, possibly through hydrogen
bonding interactions.
Higher concentrations of L-phe activate the enzyme by facilitating the transformation
2

from the T-state to the R-state. The HYSCORE spectrum of the R-state also shows two sets of
1

H cross-peaks arising from strong dipolar coupling. The first set of correlated peaks from figure

3.10, outlined with red boxes, are centered near (4.7 MHz, 17.4 MHz) and (17.3 MHz, 4.6 MHz)
T

and closely resemble the first set of correlated peaks observed for the PheH [L-phe] complex of
figure 3.9. Highlighted with a pair of black boxes, the second set of correlated peaks centered

!

100!

7

Figure 3.11: HYSCORE spectral overlay of PheHR[L-phe] ontop of [FeNO] -(N2O2)(H2O)
close to (4.7 MHz, 17.5 MHz) and (17.3 MHz, 4.6 MHz) mimic the second set of correlated
peaks in the T-state complex from figure 3.9. The placements of these proton correlations are
R

7

evidence for the existence of H2O bound to the iron in PheH [L-phe]. The [FeNO] -(N2O),
7

figure 3.5, and [FeNO] -(N2O2)(H2O), figure 3.4, highlight the difference between 2 and1 water
ligand(s), respectively. Comparing the relative intensity of the peaks resulting from the ligated
water protons to that of the other protons, centered near (8.2 MHz, 13.4 MHz) and (13.2 MHz,

!

101!

8.3 MHz), between the T- and R-state, figures 3.9 to 3.10, matches the trend observed for the
loss of a water ligand in the model complexes.
7

7

Figure 3.12a shows the 3-pulse ESEEM spectra of [FeNO] -N2O and [FeNO] -N2O2
7

7

overlaid. The intensity ratio of the high frequency peak of [FeNO] -N2O to [FeNO] -N2O2 at 18
MHz is 2:1. Figure 3.12b shows a parallel set of 3-pulse ESEEM spectra for T-state PheH (solid
trace) and R-state PheH (dashed trace). Calculating the intensity ratio observed for the T- to Rstate transition of the enzyme, for both high frequency peaks attributed to bound water protons
yields a ratio of 2.2 for the peaks at 17.7 MHz and 18.8 MHz. The loss of intensity during the
activation of the enzyme indicates that some population of bound water protons coupled to the
7

7

iron is being lost as well. Direct comparison of the intensity ratio of [FeNO] -N2O to [FeNO] N2O2, to the ratios calculated from the activation of the enzyme show they are approximately
equivalent, and that wild-type rPheH loses one bound water ligand upon activation.

!

102!

Figure 3.12: The (top) overlaid 3-pulse ESEEM spectra comparing the models (green)
7
7
[FeNO] -(N2O)(H2O)2 and (blue) [FeNO] -(N2O2)(H2O) highlights the change in the
relative peak intensities through the loss of a ligand water. Analogously, the (bot) overlay
T
R
of (green) PheH [L-phe] and (blue) PheH [L-phe] 3-pulse spectra show the loss of a
water ligand through with the intensity change of the two high-frequency peaks.

!

103!

R

2

Section 3.3.3: PheH [L-phe,5d-6-MPH4] vs PheHR[L-phe,5- H-5d-6-MPH1D3]
R

CW-EPR analysis of the quaternary forms of PheH reveals spectra that can only be
7

R

explained with a mixture of two species. The CW-EPR spectrum of [FeNO] -PheH [L-phe, 5d6-MPH4] in the g = 4 region is shown as an inset to figure 3.13. These data can be interpreted as
a composite of two paramagnetic species characterized by E/D values of 0.0255 and 0.0636 with
an approximate mixture ratio of 2:1, respectively. The CW-EPR spectrum of the quaternary

R

2

Figure 3.13: HYSCORE spectrum of PheH [L-phe, 5- H-5d-6-MPH4] at 240 mT and 4 K
with non-water high-frequency proton ligands. Inset is the CW-EPR spectrum for the same
sample.
!

104!

2

enzyme complex with deuterated pterin, 5- H-5d-6MPHD3, shown as an inset to figure 3.14, is
R

essentially identical to the PheH [L-phe,5d-6MPH4] sample. This spectrum can be interpreted as
a 2:1 mixture of species with E/D values of 0.0233 and 0.0650. The variability of the E/D values
attributed to the speciation in both the T- and R-states are too small to affect the orientation
dependence by an appreciable amount when analyzed using Easyspin

16

for the pulsed

experiments.

R

2

Figure 3.14: HYSCORE spectrum of PheH [L-phe, 5- H-5d-6-MPHD3] at 240 mT and 4
K. Inset is the CW-EPR spectrum for the same sample.

!

105!

R

The 4-pulse HYSCORE spectra collected at 240mT using quaternary PheH complexed
to 5d-6-MPH4 and are shown in figures 3.13 and 3.14, respectively. The spectrum attributed to
R

the PheH complexed with the protonated pterin analogue shows cross-peaks at (6 MHz, 16
MHz) and (16 MHz, 6 MHz) that are indicative of a large dipolar coupling, similar to those
R

2

assigned to bound H2O in the N2Ox models. A parallel analysis of PheH [L-phe,5- H-5d-62

MPHD3] shows that the H substitution, most likely at the 5-position, resulted in the loss of the
1

H correlation at (6,16 MHz). The crystal structure of truncated PheH complexed with BH4 and

thienylalanine shows that the proton bound to the nitrogen at the 5 position of the BH4 is 2.8 Å
from the Fe(II).

21

This is the proton closest to the iron in the 1MMK crystal structure. Therefore,

the cross-peaks at (6 MHz, 16 MHz) and (16 MHz, 6 MHz) in figure 3.13 can be assigned to the
proton bound to carbon at the 5-position of the reduced 5-deaza-6-methylpterin. Further 3-pulse
ESEEM and 4-pulse HYSCORE experiments performed at other magnetic field positions for the
2

1

quaternary complex treated with 5- H-5d-6MPHD3 provide no evidence for strong H dipolar
7

couplings indicative of a water molecule bound to the [FeNO] center. This absence of bound
7

R

water for [FeNO] -PheH [L-phe,5d-6-MPH4] is consistent with the 1MMK crystal structure of
the truncated enzyme.

Section 3.4: Discussion
A combination of ESEEM and HYSCORE spectra were used in this study to demonstrate
that the activation of phenylalanine hydroxylase leads to a structural change at the Fe2+ site that

!

106!

results in the loss of a bound H2O molecule. This conclusion is supported by model compound
2+

2+

2+

studies where nitrosyl derivatives of [Fe ]-(N2O)(H2O)2, [Fe ]-(N2O2)(H2O) and [Fe ]1

(N2O3) were used to develop an empirical approach to quantifying bound water. The H crosspeaks assigned to bound water in these model complexes are characterized by an unusual contour
shape. Specifically, they cover a frequency range of nearly 5 MHz, from 3-8 MHz, in their low
frequency hyperfine component, while the high-frequency components cover a range of about 2
MHz from about 16.5-18.5 MHz. As a result, the HYSCORE cross-peaks appear to be nearly
parallel to either the Frequency 1 or Frequency 2 axes. This is a consequence of orientation
1

selection, strong anisotropy in the H hyperfine coupling and, most likely, dispersion of water
molecule orientations with respect to the Fe-O bond. These details and an analysis of HYSCORE
cross-peak shapes measured across the EPR spectrum is described in chapter 4. While cross-peak
intensities in HYSCORE spectra cannot be used to quantify coupled nuclei, they do show that
the high-frequency component of the coupling is in a clean spectral window. As a result, 3-pulse
ESEEM spectra could be used for this task. A comparison of the intensity of the peak centered at
18 MHz for the series of model compounds studied showed a nearly linear relationship with the
number of bound waters.
1

The HYSCORE spectra collected for the T- and R-states of PheH show a H cross-peak
pattern with two resolved correlations that overlap the broader region of correlations covered by
bound water protons of the N2Ox model complexes. These data support our interpretation of the
contour shapes for the model complexes and our assignment of the correlations resolved for the

!

107!

T

R

PheH [L-phe] and PheH [L-phe] samples. Because the ESEEM and HYSCORE data collected
for the model complexes arise from a distribution of conformers, quantification of the number of
water ligands using spectral simulations from a spin Hamiltonian model would be difficult.
Therefore, we favor the empirical analysis used here that shows the loss of a water ligand as
PheH transitions from the T- to the R-state.
Putting the loss of the water ligand into context of other spectroscopic information on
PheH in the literature provides a detailed description of the catalytic site as it prepares to carry
out hydroxylation. MCD and XAS spectroscopic studies describe rat PheH as maintaining a
distorted octahedral geometry in the presence of just the substrate, L-phenyalanine.

14

The

literature lacks crystal structures with only phenylalanine present; however, the 1J8T and 1J8U

17

crystal structures of truncated hPheH with an empty active site and truncated hPheH with bound
cofactor, respectively, are commensurate with the distorted octahedral geometry predicted for rat
2

PheH. Adding phenylalanine causes a global conformational change, which also facilitates a
ligand rearrangement with the glutamate coordination to the Fe(II) changing from mono- to
bidentate.

15, 18

During the activation of PheH the chelating glutamate displaces one of the water

ligands maintaining the distorted octahedral geometry observed previously. When the activation
of rPheH is carried out in the presence of the cofactor, MCD studies observed the transition into
a five-coordinate square pyramidal geometric structure.

19

The conversion from six-coordinate to

a five-coordinate structure is thought to remove another water ligand from the first coordination
sphere. It has been suggested that ligand rearrangement coupled to the activation of the enzyme

!

108!

creates an open coordination site where molecular oxygen binds leaving one water molecule
7

coordinated to Fe(II). Previous CW-EPR studies of [FeNO] derivatives of bacterial-PheH using
17

O-labelled water and line broadening to detect water coordination showed that in the presence

of both substrate and cofactor, no water was coordinated to Fe. These authors concluded that the
NO had displaced the final water molecule when it bound to the metal center.

20

In this study, the

HYSCORE spectrum in figure 10b provides evidence that no bound water molecules are present
in the R-state quaternary complex of PheH with nitric oxide. The resulting dehydrated first
coordination sphere is in agreement with the previous CW-EPR study of bacterial PheH and the
X-ray crystallographic results obtained using truncated PheH and the slow-substrate,
thienylalanine. Provided that NO and O2 coordinate to Fe(II) in the same fashion, these data
show that the hydroxylation of tetrahydrobiopterin and phenylalanine occur without the presence
of an Fe(II)-bound water ligand.

!

109!

APPENDICES

!

110!

7

Appendix 1: Preparation of [FeNO] Samples
This work comes directly from our collaborators at Boston University (Gulbenk Aranat,
Joshua McNally, Patrick Cappolino, and John P. Caradonna).
Model Complex Preparation
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 (N2) glovebox, except were noted. Nitric oxide (NO) gas (99+%) was
purchesed from Air Gas (Maumee, OH) and used without further purification.
2+

2+

2+

The ferrous complexes [Fe ]-(N2O3)(L), [Fe ]-(N2O2)(L2), and [Fe ]-(N2O1)(L3)
9

were prepared by literature procedures. NO adducts of these ferrous complexes were prepared
2+

2+

2+

as follows. Stock solutions of [Fe ]-(N2O3)(L), [Fe ]-(N2O2)(L2), and [Fe ]-(N2O1)(L3)
were prepared in degassed H2O, pH 6.5 was obtained with 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 NO gas at which point the colorless solution
7

turned immediately orange with the formation of [FeNO] . 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 5 mM with
degassed 50:50 glycerol as glassing agent for ~40% final glycerol content. The samples were
loaded into EPR tubes, frozen in N2 (l) and stored at N2 (l) temperatures.

!

111!

Enzyme Sample Preparation
All commercial reagents were of the highest grade available and were used without
further purification, with the exception of glycerol, which was treated with activated carbon to
21

removed 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
2

(Uppsala, Sweden). D2O (99.9 atom % D) and per- H-L-phe were from Cambridge Isotopes
Laboratories (Andover, MA). Diethylamine NONOate (DEA/NO) was purchased from Cayman
Chemicals (Ann Arbor, MI). 6-MPH4
characterized by standard methods.
2

22

24

and 5-deaza-6-MPH4

23

were synthesized and

Both pterin species were stored at -20°C until required. 5-

2

deaza-6-CH3-5,6,7,8- H-pterin ( H-5-deaza-6-MPH4) was synthesized as previously
23

2

described , except that deuterium gas is used in the reduction step. The para- H-L-phe was also
synthesized as previously described

25

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
was routinely quantified using a Varian AA280 atomic absorption spectrometry, with Zeeman
GTA12OZ graphite furnace attachment, at 248.3 nm. The iron standard was purchased form
Fisher (Pittsburgh, PA). Automated protein purification was performed using a Pharmacia LKB
FPLC (Uppsala, Sweden). All electronic absorption spectroscopy (UV/Vis) 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).

!

112!

Recombinant rat wtPheH was over-expressed in E. coli BL21 (DE3) cells and purified
21

using a variation of the hydrophobic affinity method , as previously described.

26

The specific

activities of wtPheH range between 6 – 7 units/mg (units = micromoles of tyrosin 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
wtPheH 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 electrode was developed
using a method adapted from Stetter et. al.

27

A commercially 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.

28

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 0.025 - 2.5 mM NO) according to the method suggest by the
manufacturers. The calibration factor µA/µM 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

!

113!

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. After 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.

28

Once the buffer was saturated with NO, it was

then added to wtPheH 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 ( [FeNO]7-PheH) was
transferred to Centricon (30 K) microconcentrators (Millipore) and concentrated to a final
concentration of 1.2-2 mM Fe/subunit. Approximately 200 µL of the reaction mixture was
quickly transferred to a quartz EPR tube (4 MM OD, 707-SQ-250M, Wilmad, Buena, NJ) using
a glass pipette and immediately frozen in liquid N2.
If the enzyme needed to be manipulated with protiated for deuterated substrate and/or
2

cofactor, final concentration of 10 – 15 mM of the corresponding chemical (i.e. L-phe, para- H2

L--phe, 5-deaza-6-MPH4, H-5-deaza-6-MPH4) was added to the buffer solution (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 incubated at 25°C for 10 minutes when necessary.

!

114!

Appendix 2: MATLAB Scripts
Wrapper Script
The wrapper script was used as a workstation to run simulation and fitting routines.
Within this script all parameters were defined and feed to a function that carried out the task
specified here. The specific wrapper script used to run the 2 proton simulation/fitting routine for
the [FeNO]7-PheHR[L-phe] sample was:
%%%%%%%% Spin System
par.g=[2.02 2.0024];
par.S=3/2;
par.D=[300000 10000];
par.DStrain=[0 0];
par.lw=1.8;
%%%%%%%%
%%%%%%%% Experimental parameters
par.Sequence='HYSCORE';
par.mwFreq=9.68;
par.nPoints=128;
par.ExciteWidth=67.5;
par.dt=0.016;
par.nKnots=91;
par.Field=[240 260 280 180 200 320];
par.tau=[0.096 0.092 0.084 .128 .116 0.142];
%%%%%%%%
%%%%%%% fitting/simulation information
par.N='1H,1H'; % coupled nuclei
%[aiso T rho alpha beta gamma]…
a=[0.3,4.33,1,80,64,60;... % proton 1 - parameters
0.8,4.84,1,0,78,4];
% proton 2 - parameters
par.change=[1 1 1 1 1 1]; % Defines the parameters that will be varied
% in a fit. 1 = varied, 0 = not varied
par.aConst=a; % Defines the values in second place for reference
par.lb=[0 3 0 0 0 0;... % proton 1 - Lower Bounds
0 3 0 0 0 0];%...
% proton 2
par.ub=[1 13 1 180 180 180;... % proton 1 - Upper Bounds
1 13 1 180 180 180];%...
% proton 2
par.pick=320; % Field to Simulate
par.basic=1; % 1 for simulation, 0 for fit
par.oneFit=1; % 1 to fit on one field, 0 for entire range (par.Field)
par.proc=2;
% 1 = decay 2 = cta % anything else is just simple fft
par.split=1; % 1 = no split **Use only for different spin centers**
par.numArray=[2 1]; % # of nuclei in each calc from split simulations
par.factorArray=[1 1.5]; % weighting factors

!

115!

par.HyscFit=1; % 1=P1, 3=P2 to be fit
par.HyscExperimentalMax.Left=[4.72
6.51 8.46; ...
17.42 17.58 17.58];
par.HyscExperimentalMax.Right=[7.7
8.47 9.77; ...
17.45 17.90 17.90];

%x
%y
%x
%y

(MHz)
(MHz)
(MHz)
(MHz)

P1
P1
P2
P2

par.beg=[102 104 93]; % The index to fit 3P ESEEM
par.end=[131 127 113]; % calcs residuals from par.beg:1:par.end
par.sample='wtPAHrphe42'; % Sample variable
par.k=['DS.' par.sample]; % Personally stored all data in DS struct
par.dir='.'; % useful on hpcc
%angle distribution information
par.ang=0; % 1 = multiple angles; 0 = one orientation
par.FWHM=3; % FWHM of distribution
par.numAngles=5; %The number of angles to use for Gaussian distribution
%%%%%%%
%%%%%%% viewing window
par.constraints=[3 8.5 16 19;...%[xmin xmax ymin ymax] for B=240
6 10 16 19;... %xmin xmax ymin ymax] for B=260
8 11 16 19]; %xmin xmax ymin ymax] for B=280
par.plotCon=0; % 1 = display data and simulation; 0 = false
par.con=1; % 1 = specify bounds on contours; 0 = use whole Z range
par.conTop=7E6; % upper bound contour
par.conBot=7E5; % lower bound contour
par.norm=0; % 0 = no 1 = yes % normalizes the proton peak
par.proMax=1929064; % value of the normalization
par.resMe=0; % turns reporting of residue on = 1 off = 0
% Declares the position of textbox – display for correlations max/min
if par.plotCon==1
par.h=subplot(1,2,1);
par.expMax=annotation('textbox',[0 .2 .1 .1]);
par.simMax=annotation('textbox',[0 .1 .1 .1]);
par.font=12;
end
%%%%%%%
%%%%%%% Method Run
if par.basic==0
if strcmp(par.Sequence,'HYSCORE')==1
options = psoptimset('Display','iter');
dpar=patternsearch(@(a)HYSCORE_Simple(a,par,DS),a,[],[],[],[],par.lb,par.ub,o
ptions);
elseif strcmp(par.Sequence,'3pESEEM')==1
options = psoptimset('Display','iter','TolFun',1e-12);

dpar=patternsearch(@(a)N2O2_ESEEM_Simple(a,par,DS),a,[],[],[],[],par.lb,par.u
b,options);

!

116!

%dpar=simulannealbnd(@(a)N2O2_ESEEM_Simple(a,par,DS),a,par.lb,par.ub,options)
;
end
else
if strcmp(par.Sequence,'HYSCORE')==1
[~,p]=HYSCORE_Simple(a,par,DS,par.plotCon);
title(strcat('------'))
xlabel('Frequency 1 (MHz)')
ylabel('Frequency 2 (MHz)')
grid on;
elseif strcmp(par.Sequence,'3pESEEM')==1
N2O2_ESEEM_Simple(a,par,DS)
end
end
%%%%%%%

The wrapper script called on the functions ‘HYSCORE_Simple’ and ‘ESEEM_Simple’ to carry
out the duties described within the above script.
‘HYSCORE_Simple’:
function [res,p]=HYSCORE_Simple(a,par,DS,plotMe)
%======================================================================
% Outputs
% res = normalized sum of the square of the residuals
% p = HYSCORE data set of the simulation p.f1 (x) p.f2 (y) and p.fd (z)
%
% Inputs
% a = parameters
% par = information defined by wrapper script
% DS = data set – contains experimental data
% plotMe = binary switch to turn off graphics (useful for hpcc)
%
%======================================================================
if nargin < 4
plotMe=1;
end
res='NA';
a=BuildA(a,par);
%Spin Build
Sys.S=par.S;
Sys.g=par.g;
Sys.D=par.D;
Sys.DStrain=par.DStrain;
Sys.lw=par.lw;
%Experimental Build
Exp.Sequence = par.Sequence;
Exp.mwFreq=par.mwFreq;

!

117!

Exp.nPoints = par.nPoints;
Exp.dt = par.dt;
Exp.ExciteWidth = par.ExciteWidth;
%Options
Opt.nKnots = par.nKnots;
%Sets the parameters and determine how many run-throughs are needed
if par.basic==1 && par.resMe==1
[~,rt]=size(par.Field);
s=zeros(1,rt);
if par.HyscFit==1
section='Left';
elseif par.HyscFit==2
section='Mid';
elseif par.HyscFit==3
section='Right';
end
xyData=eval(strcat('par.HyscExperimentalMax.',section));
elseif par.oneFit==1 && par.resMe==1 || par.basic==1 && par.resMe==1
[~,rt]=size(par.Field);
for j=1:rt
if par.Field(j)==par.pick
temp.Field=par.Field(j);
par.tau=par.tau(j);
par.beg=par.beg(j);
par.end=par.end(j);
if par.HyscFit==1
section='Left';
elseif par.HyscFit==2
section='Mid';
elseif par.HyscFit==3
section='Right';
end
xyData=eval(strcat('par.HyscExperimentalMax.',section,'(:,j)'));
end
end
par.Field=temp.Field;
rt=1;
s=zeros(1,rt);
clear('temp.Field');
elseif par.oneFit==1 && par.resMe==0 || par.basic==1 && par.resMe==0
[~,rt]=size(par.Field);
for j=1:rt
if par.Field(j)==par.pick
temp.Field=par.Field(j);
par.tau=par.tau(j);
end
end
par.Field=temp.Field;
rt=1;
s=zeros(1,rt);
clear('temp.Field');
else
[~,rt]=size(par.Field);
s=zeros(1,rt);
if par.HyscFit==1
section='Left';

!

118!

elseif par.HyscFit==2
section='Mid';
elseif par.HyscFit==3
section='Right';
end
xyData=eval(strcat('par.HyscExperimentalMax.',section));
end
%creation of data house
store=zeros(2*Exp.nPoints);
%main run
for i=1:rt
Exp.Field = par.Field(i);
Exp.tau = par.tau(i);
%accounts for weightings of multiple species
for runTime=1:par.split
par.num=par.numArray(runTime);
%sets par.nucs
if par.split>1
%nulear split
par.nums=par.num;
par.Nu=regexp(par.N,',','split');
par.Nucs='';
for parRun=1:par.num
if parRun==1
par.Nucs=par.Nu(1,parRun);
else
par.Nucs=strcat(par.Nucs,',',par.Nu(1,parRun));
end
end
par.Nucs=char(par.Nucs);
else
par.Nucs=par.N;
end
Sys.Nucs=par.Nucs;
%selection of the variables and factors
if runTime>1
par.size=0;
for sum=1:length(par.numArray)
par.size=par.size+par.numArray(sum);
end
[row,col]=size(a);
for rp=par.size:row
zp=1;
for cp=1:col
temp.a(rp,cp)=a(zp,cp);
end
zp=zp+1;
end
else
par.size=0;

!

119!

for sum=1:length(par.numArray)
par.size=par.size+par.numArray(sum);
end
[row,col]=size(a);
for rp=1:row
for cp=1:col
temp.a(rp,cp)=a(rp,cp);
end
end
end
%Finish setting up spin system
clear('Sys.A');
temp.a(:,3)=abs(temp.a(:,3)); %makes rho a positive value
[Sys]=spinBuildnew(temp.a,par,Sys);
%Angle distribution
if par.ang==1
for angRun=1:par.num
par.center=Sys.Apa(angRun,2)*180/pi;
strBeta=strcat('par.betaAngles.group',num2str(angRun));
eval(strcat(strBeta,'=zeros(2,',num2str(par.numAngles),');'));
eval(strcat(strBeta,'=gaussianDistribution(par);'));
%eval(strcat(strBeta,'(2,:)'));
end
for angDist=1:par.numAngles
avg=0;
for angRun=1:par.num
strBeta=strcat('par.betaAngles.group',num2str(angRun));
eval(strcat('Sys.Apa(',num2str(angRun),',2)=',...
strBeta,'(2,',num2str(angDist),');'));
avg=avg+eval(strcat(strBeta,'(1,',num2str(angDist),');'));
end
avg=avg/par.num;
[~,~,~,p]=saffron(Sys,Exp,Opt);
normData=avg*par.factorArray(runTime)*p.fd/max(max(p.fd));
store=store+normData;
end
else
[~,~,~,p]=saffron(Sys,Exp,Opt);
normData=par.factorArray(runTime)*p.fd/max(max(p.fd));
store=store+normData;
end
end
%Data with correct field
String=strcat(par.k,'.B', num2str(par.Field(i)),'mT.HYSCORE.');
%Graphing info
if par.Field(i)==par.pick && plotMe==1
%Simulation
if par.norm==1
normTop=(max(max(p.fd))*(par.conTop/par.proMax));

!

120!

normBot=(max(max(p.fd))*(par.conBot/par.proMax));
incF=abs(normTop-normBot)/par.numCon;
p.v=normBot:incF:normTop;
subplot(1,2,1);
contour(p.f1,p.f2,p.fd,p.v)
xlim([0 22])
ylim([0 22])
drawnow;
p.z=p.fd;
p.x=p.f1;
p.y=p.f2;
p.max=normTop;
p.min=normBot;
else
subplot(1,2,1);
contour(p.f1,p.f2,p.fd)
xlim([0 22])
ylim([0 22])
drawnow;
end
%Data
if par.con==1 && par.plotCon==1
incCon=abs((par.conTop-par.conBot)/par.numCon);
v=par.conBot:incCon:par.conTop;
subplot(1,2,2);
contour(eval(strcat(String,'x')),eval(strcat(String,'y')),eval(strcat(String,
'z')),v)
xlim([0 22])
ylim([0 22])
drawnow;
par.plotCon=0;
elseif par.plotCon==1
subplot(1,2,2);
contour(eval(strcat(String,'x')),eval(strcat(String,'y')),eval(strcat(String,
'z')))
xlim([0 22])
ylim([0 22])
drawnow;
par.plotCon=0;
end
elseif par.Field(i)==par.pick && plotMe==0
%Simulation
if par.norm==1
normTop=(max(max(p.fd))*(par.conTop/par.proMax));
normBot=(max(max(p.fd))*(par.conBot/par.proMax));
incF=abs(normTop-normBot)/par.numCon;
p.v=normBot:incF:normTop;
contour(p.f1,p.f2,p.fd,p.v)
xlim([0 22])
ylim([0 22])
drawnow;
p.z=p.fd;
p.x=p.f1;
p.y=p.f2;

!

121!

p.max=normTop;
p.min=normBot;
xlabel('Frequency 1 (MHz)')
ylabel('Frequency 2 (MHz)')
grid on;
else
contour(p.f1,p.f2,p.fd)
xlim([0 22])
ylim([0 22])
drawnow;
end
end
%residue calculations
temp.x=p.f1;
temp.y=p.f2;
temp.z=p.fd;
[~,xy]=muteSpecCorner(temp,par.Field(i));
xfreq=xy(2,1);
yfreq=xy(2,2);
if par.Field(i)==par.pick && plotMe==1
content = sprintf(strcat('Data X: ',num2str(xyData(1,i)),'\n','Data
Y: ',num2str(xyData(2,i))));
set(par.expMax,'String',content,'Fontsize',par.font,'FitHeightToText','on');
set(par.simMax,'String',content,'Fontsize',par.font,'FitHeightToText','on');
end

if par.resMe==1
s(1,i)=(((round(100*xfreq)/100)xyData(1,i))^2)+(((round(100*yfreq)/100)-xyData(2,i))^2);
end
end
if par.resMe==1
res=0;
for i=1:rt
res=s(1,i)+res;
end
res=res/(3*0.04); % normalization and reweighting
% 3 = # Fields, 0.04 = square of variance
end
end

!

122!

‘ESEEM_Simple’:
function [res,x,y]=N2O2_ESEEM_Simple(a,par,DS)
%======================================================================
% Outputs
% res = normalized sum of the square of the residuals
% x = ESEEM time-domain data
% y = ESEEM time-domain data
%
% Inputs
% a = parameters
% par = information defined by wrapper script
% DS = data set – contains experimental data
%
%======================================================================
a=BuildA(a,par);
%Spin Build
Sys.S=par.S;
Sys.g=par.g;
Sys.D=par.D;
Sys.DStrain=par.DStrain;
Sys.lw=par.lw;
%Experimental Build
Exp.Sequence = par.Sequence;
Exp.mwFreq=par.mwFreq;
Exp.nPoints = par.nPoints;
Exp.dt = par.dt;
Exp.ExciteWidth = par.ExciteWidth;
Opt.nKnots = par.nKnots;
%Sets the parameters and determine how many run-throughs are needed
if par.oneFit==1 || par.basic==1
[~,rt]=size(par.Field);
for j=1:rt
if par.Field(j)==par.pick
temp.Field=par.Field(j);
par.tau=par.tau(j);
par.beg=par.beg(j);
par.end=par.end(j);
end
end
par.Field=temp.Field;
rt=1;
clear('temp.Field');
else
[~,rt]=size(par.Field);
s=zeros(1,rt);
end
%main run
for i=1:rt

!

123!

%creation of data house
store=zeros(1,Exp.nPoints);
z=strcat(par.k,'.B',num2str(par.Field(i)),'mT'); %Defines Data being fit
Exp.Field = par.Field(i);
Exp.tau = par.tau(i);
%accounts for weightings of multiple species
for runTime=1:par.split
par.num=par.numArray(runTime);
%sets par.nucs
if par.split>1
%nulear split
par.nums=par.num;
par.Nu=regexp(par.N,',','split');
par.Nucs='';
for parRun=1:par.num
if parRun==1
par.Nucs=par.Nu(1,parRun);
else
par.Nucs=strcat(par.Nucs,',',par.Nu(1,parRun));
end
end
par.Nucs=char(par.Nucs);
else
par.Nucs=par.N;
end
Sys.Nucs=par.Nucs;
%selection of the right variables and factors
if runTime>1
par.size=0;
for sum=1:length(par.numArray)
par.size=par.size+par.numArray(sum);
end
[row,col]=size(a);
for rp=par.size:row
zp=1;
for cp=1:col
temp.a(rp,cp)=a(zp,cp);
end
zp=zp+1;
end
else
par.size=0;
for sum=1:length(par.numArray)
par.size=par.size+par.numArray(sum);
end
[row,col]=size(a);
for rp=1:row
for cp=1:col
temp.a(rp,cp)=a(rp,cp);
end

!

124!

end
end
%Finish setting up spin system
clear('Sys.A');
temp.a(:,3)=abs(temp.a(:,3)); %makes rho a positive value
[Sys]=spinBuildnew(temp.a,par,Sys);
%Angle distribution/not
if par.ang==1
for angRun=1:par.num
par.center=Sys.Apa(angRun,2)*180/pi;
strBeta=strcat('par.betaAngles.group',num2str(angRun));
eval(strcat(strBeta,'=zeros(2,',num2str(par.numAngles),');'));
eval(strcat(strBeta,'=gaussianDistribution(par);'));
end
for angDist=1:par.numAngles
avg=0;
for angRun=1:par.num
strBeta=strcat('par.betaAngles.group',num2str(angRun));
eval(strcat('Sys.Apa(',num2str(angRun),',2)=',...
strBeta,'(2,',num2str(angDist),');'));
avg=avg+eval(strcat(strBeta,'(1,',num2str(angDist),');'));
end
avg=avg/par.num;
[x,tempY]=saffron(Sys,Exp,Opt);
normData=tempY/max(abs(tempY));
store(1,:)=store(1,:)+(normData(1,:)*par.factorArray(runTime)*avg);
end
else
[x,tempY]=saffron(Sys,Exp,Opt);
normData=tempY/max(abs(tempY));
store(1,:)=store(1,:)+(normData(1,:)*par.factorArray(runTime));
end
end
y=store;
%Sets the processing mode
if par.proc==1
timeX=eval(strcat(z,'.x'));
AmpY=eval(strcat(z,'.y'));
warning off all; % command window becomes cluttered without
[~,~,~,yb]=simUprocDecay(timeX,AmpY,Exp.dt,6);
[x,y,~,~]=simUprocDecay(x,y,Exp.dt,6,yb(1,:));
warning on all;
xStr=strcat(par.k,'.B', num2str(par.Field(i)),'mT.xf');
yStr=strcat(par.k,'.B', num2str(par.Field(i)),'mT.yf');
yValue='.yf(1,';
elseif par.proc==2

!

125!

warning off all;
[x,y]=simUprocCTA(x,y,Exp.dt,6,20);
warning on all;
if exist('DS.wtPAHrphe42.B240mT.CTAxf','var')==0
timeX=eval(strcat(z,'.x'));
AmpY=eval(strcat(z,'.y'));
[temp.x,temp.y]=simUprocCTA(timeX,AmpY,Exp.dt,6,20);
xStr='temp.x';
yStr='temp.y';
else
xStr=strcat(par.k,'.B', num2str(par.Field(i)),'mT.CTAxf');
yStr=strcat(par.k,'.B', num2str(par.Field(i)),'mT.CTAyf');
end
yValue='.CTAyf(1,';
else
[x,y]=simUproc(x,y,Exp.dt,0);
xStr=strcat(par.k,'.B', num2str(par.Field(i)),'mT.xf');
yStr=strcat(par.k,'.B', num2str(par.Field(i)),'mT.yf');
yValue='.yf(1,';
end
% subtracts residual baseline left
if par.ManipulateY==1
y(:)=y(:)-y(par.nPoints/2);
yExp=eval(yStr);
yExp(:)=yExp(:)-yExp(par.nPoints/2);
else
yExp=eval(yStr);
end
if par.Field(i)==par.pick
plot(eval(xStr),yExp,x,y)
title(strcat(par.k,'.B', num2str(par.Field(i)),'mT'))
xlabel('Frequency (MHz)')
ylabel('Modulated Amplitude')
grid on;
drawnow
clear ('xStr','yStr','yExp');
end
res=0;
for j=1:(par.end(i)-par.beg(i))
zNum=eval(strcat(z,yValue,num2str(j+par.beg(i)),')'));
res=res+((y(1,j+par.beg(i))-zNum)^2);
end
s(i)=sqrt(res/(par.end(i)-par.beg(i)-1));
end
res=0;
for j=1:3
res=(res+s(i));
end
res=res/3;
end

!

126!

BIBLIOGRAPHY

!

127!

BIBLIOGRAPHY

1.!Kaufman,!S.,!Bridgers,!W.!F.,!Friedman,!S.,!and!Eisenberg,!F.!(1962)!Source!of!Oxygen!in!
Phenylalanine!Hydroxylase!and!DopamineMBetaMHydroxylase!Catalyzed!Reactions,!
Biochem(Bioph(Res(Co!9,!497M&.!
2.!Shiman,!R.,!and!Gray,!D.!W.!(1980)!Substrate!Activation!of!PhenylalanineMHydroxylase!M!a!
Kinetic!Characterization,!Journal(of(Biological(Chemistry!255,!4793M4800.!
3.!Stokka,!A.!J.,!and!Flatmark,!T.!(2003)!SubstrateMinduced!conformational!transition!in!
human!phenylalanine!hydroxylase!as!studied!by!surface!plasmon!resonance!
analyses:!the!effect!of!terminal!deletions,!substrate!analogues!and!phosphorylation,!
Biochem(J!369,!509M518.!
4.!Bassan,!A.,!Blomberg,!M.!R.,!and!Siegbahn,!P.!E.!(2003)!Mechanism!of!aromatic!
hydroxylation!by!an!activated!FeIV=O!core!in!tetrahydrobiopterinMdependent!
hydroxylases,!Chemistry!9,!4055M4067.!
5.!Olsson,!E.,!Martinez,!A.,!Teigen,!K.,!and!Jensen,!V.!R.!(2010)!Water!Dissociation!and!
Dioxygen!Binding!in!Phenylalanine!Hydroxylase,!Eur(J(Inorg(Chem,!351M356.!
6.!Mims,!W.!B.!(1972)!Envelope!Modulation!in!SpinMEcho!Experiments,!Physical(Review(B!5,!
2409M2419.!
7.!Mims,!W.!B.!(1972)!Amplitudes!of!Superhyperfine!Frequencies!Displayed!in!the!ElectronM
SpinMEcho!Envelope,!Physical(Review(B!6,!3543M3545.!
8.!Schweiger,!A.,!and!Jeschke,!G.!(2001)!Principles(of(pulse(electron(paramagnetic(resonance,!
Oxford!University!Press,!Oxford.!
9.!Cappillino,!P.!J.,!Miecznikowski,!J.!R.,!Tyler,!L.!A.,!Tarves,!P.!C.,!McNally,!J.!S.,!Lo,!W.!N.,!
Kasibhatla,!B.!S.!T.,!Krzyaniak,!M.!D.,!McCracken,!J.,!Wang,!F.,!Armstrong,!W.!H.,!and!
Caradonna,!J.!P.!(2012)!Studies!of!iron(II)!and!iron(III)!complexes!with!facMN2O,!cisM
N2O2!and!N2O3!donor!ligands:!models!for!the!2MHis!1Mcarboxylate!motif!of!nonM
heme!iron!monooxygenases,!Dalton(Transactions!41,!5662M5677.!

!

128!

10.!Stoll,!S.,!and!Schweiger,!A.!(2006)!EasySpin,!a!comprehensive!software!package!for!
spectral!simulation!and!analysis!in!EPR,!J(Magn(Reson!178,!42M55.!
11.!Fauth,!J.!M.,!Schweiger,!A.,!Braunschweiler,!L.,!Forrer,!J.,!and!Ernst,!R.!R.!(1986)!
Elimination!of!unwanted!echoes!and!reduction!of!dead!time!in!threeMpulse!electron!
spinMecho!spectroscopy,!Journal(of(Magnetic(Resonance((1969)!66,!74M85.!
12.!Stoll,!S.,!and!Kasumaj,!B.!(2008)!Phase!Cycling!in!Electron!Spin!Echo!Envelope!
Modulation,!Applied(Magnetic(Resonance!35,!15M32.!
13.!Brown,!C.!A.,!Pavlosky,!M.!A.,!Westre,!T.!E.,!Zhang,!Y.,!Hedman,!B.,!Hodgson,!K.!O.,!and!
Solomon,!E.!I.!(1995)!Spectroscopic!and!Theoretical!Description!of!the!ElectronicM
Structure!of!S=3/2!IronMNitrosyl!Complexes!and!Their!Relation!to!OM2!Activation!by!
Nonheme!Tron!Enzyme!ActiveMSites,!Journal(of(the(American(Chemical(Society!117,!
715M732.!
14.!Loeb,!K.!E.,!Westre,!T.!E.,!Kappock,!T.!J.,!Mitic,!N.,!Glasfeld,!E.,!Caradonna,!J.!P.,!Hedman,!
B.,!Hodgson,!K.!O.,!and!Solomon,!E.!I.!(1997)!Spectroscopic!characterization!of!the!
catalytically!competent!ferrous!site!of!the!resting,!activated,!and!substrateMbound!
forms!of!phenylalanine!hydroxylase,!Journal(of(the(American(Chemical(Society!119,!
1901M1915.!
15.!Andersen,!O.!A.,!Stokka,!A.!J.,!Flatmark,!T.,!and!Hough,!E.!(2003)!2.0A!resolution!crystal!
structures!of!the!ternary!complexes!of!human!phenylalanine!hydroxylase!catalytic!
domain!with!tetrahydrobiopterin!and!3M(2Mthienyl)MLMalanine!or!LMnorleucine:!
substrate!specificity!and!molecular!motions!related!to!substrate!binding,!J(Mol(Biol!
333,!747M757.!
16.!Stoll,!S.,!and!Britt,!R.!D.!(2009)!General!and!efficient!simulation!of!pulse!EPR!spectra,!
Physical(Chemistry(Chemical(Physics!11,!6614M6625.!
17.!Andersen,!O.!A.,!Flatmark,!T.,!and!Hough,!E.!(2001)!High!resolution!crystal!structures!of!
the!catalytic!domain!of!human!phenylalanine!hydroxylase!in!its!catalytically!active!
Fe(II)!form!and!binary!complex!with!tetrahydrobiopterin,!J(Mol(Biol!314,!279M291.!
18.!Andersen,!O.!A.,!Flatmark,!T.,!and!Hough,!E.!(2002)!Crystal!structure!of!the!ternary!
complex!of!the!catalytic!domain!of!human!phenylalanine!hydroxylase!with!
tetrahydrobiopterin!and!3M(2Mthienyl)MLMalanine,!and!its!implications!for!the!
mechanism!of!catalysis!and!substrate!activation,!J(Mol(Biol!320,!1095M1108.!

!

129!

19.!Solomon,!E.!I.,!Decker,!A.,!and!Lehnert,!N.!(2003)!NonMheme!iron!enzymes:!Contrasts!to!
heme!catalysis,!P(Natl(Acad(Sci(USA!100,!3589M3594.!
20.!Han,!A.!Y.,!Lee,!A.!Q.,!and!AbuMOmar,!M.!M.!(2006)!EPR!and!UVMvis!studies!of!the!nitric!
oxide!adducts!of!bacterial!phenylalanine!hydroxylase:!Effects!of!cofactor!and!
substrate!on!the!iron!environment,!Inorganic(Chemistry!45,!4277M4283.!
21.!Shiman,!R.,!Gray,!D.!W.,!and!Pater,!A.!(1979)!A!simple!purification!of!phenylalanine!
hydroxylase!by!substrateMinduced!hydrophobic!chromatography,!J(Biol(Chem!254,!
11300M11306.!
22.!Kaufman,!S.,!Storm,!C.!B.,!and!Shiman,!R.!(1971)!Preparation!of!6Msubstituted!pterins!via!
the!Isay!reaction,!The(Journal(of(Organic(Chemistry!36,!3925M3927.!
23.!Stark,!E.,!and!Breitmaier,!E.!(1973)!5MDesazapteridine,!synthese!und!NMRM
spektroskopie,!Tetrahedron!29,!2209M2217.!
24.!Shiman,!R.,!Akino,!M.,!and!Kaufman,!S.!(1971)!Solubilization!and!Partial!Purification!of!
Tyrosine!Hydroxylase!from!Bovine!Adrenal!Medulla,!Journal(of(Biological(Chemistry!
246,!1330M1340.!
25.!Guroff,!G.,!Reifsnyder,!C.!A.,!and!Daly,!J.!(1966)!Retention!of!deuterium!in!pMtyrosine!
formed!enzymatically!from!pMdeuterophenylalanine,!Biochem(Bioph(Res(Co!24,!720M
724.!
26.!Kappock,!T.!J.,!Harkins,!P.!C.,!Friedenberg,!S.,!and!Caradonna,!J.!P.!(1995)!Spectroscopic!
and!Kinetic!Properties!of!Unphosphorylated!Rat!Hepatic!Phenylalanine!Hydroxylase!
Expressed!in!Escherichia!coli:!COMPARISON!OF!RESTING!AND!ACTIVATED!STATES,!
Journal(of(Biological(Chemistry!270,!30532M30544.!
27.!Stetter,!J.!R.,!and!Li,!J.!(2008)!Amperometric!Gas!SensorsA!Review,!Chemical(Reviews!
108,!352M366.!
28.!Howland,!J.!L.!(1997)!Methods!in!nitric!oxide!research:!Edited!by!M!Feelisch!and!J!S!
Stamler.!pp!712.!John!Wiley!and!Sons,!New!York.!1996.!$69.95!ISBN!0M471M95524M8,!
Biochemical(Education!25,!184M185.!
!

!

130!

Chapter 4: Quantitative HYSCORE Analysis of the Allosteric forms of Phenylalanine
Hydroxylase
Section 4.1: Introduction
Phenylalanine Hydroxylase facilitates the coupled hydroxylation of L-phenylalanine and
tetrahydrobiopterin using molecular oxygen.

1, 2

The general mechanism of the coupled
5

hydroxylation starts with the hydroxylation of the 4a-C on the tetrahydrobiopterin, which
6

results in the formation of a high-valent iron(IV)-oxo species. This ferryl-oxo complex attacks
L-phenylalanine and results in the hydroxylation of the aromatic amino acid through an
7

electrophilic aromatic substitution mechanism. Currently, the literature is unclear on the role
water plays during the formation of the ferryl-oxo species.

8, 9

In the previous chapter it was

shown, using HYSCORE, that water ligand protons produce high-frequency correlation peaks
that are in a clean spectral window. By monitoring these water ligand protons using HYSCORE,
10, 11

it was determined that through the activation of PheH, transition from the T- to the R-state

one water ligand is lost in the presence of L-phenylalanine. In this chapter, these peaks will be
analytically investigated to obtain structural information including distances and ligand
7

T

orientations. The analysis will start with the enzyme samples [FeNO] -PheH [L-phe] and
7

R

[FeNO] -PheH [L-phe], while initially using the model complexes as support. From here the
7

[FeNO] -(N2Ox)(H2O)x model complexes

12

will be analyzed using the prescribed method

developed with the enzyme samples.

!

131!

,

Section 4.2 Materials and Methods
Our collaborators at Boston University under the direction of Dr. John P. Caradonna
made the samples. Their preparation is cataloged in appendix 1.
The methods mentioned in chapter 2 were implemented here to obtain ESEEM and
HYSCORE spectra. CW-EPR spectra in this chapter were fit and presented in chapter 3 but in
some cases will be presented again for completeness. They were fit with the ‘Pepper’ module in
the EasySpin package

13

for MATLAB (The Mathworks, Natick, MA), while the pulsed

experiments (i.e. ESEEM and HYSCORE) utilized the ‘Saffron’ module.

14

With this function,

scripts were created to allow for a frequency-based fitting using the ‘patternsearch’ function
2

within MATLAB. Best-fits were obtained by minimizing χ , which were calculated based on the
difference between the frequency position of the upper left correlation peak of the data and
simulation.

!

132!

Section 4.3: Results and Analysis
7

T

7

R

Section 4.3.1: CW-EPR Analysis of [FeNO] -PheH [L-phe] and [FeNO] -PheH [L-phe]
Before the completion of a qualitative analysis on the enzyme samples that represents the
activation of phenylalanine hydroxylase, a CW-EPR analysis must be completed. Here is a quick
summary of the CW-EPR analysis carried
out in section 3.2 of chapter 3. Figure 4.1
shows the CW-EPR g = 4 feature of the
7

T

7

[FeNO] -PheH [L-phe] and [FeNO] R

PheH [L-phe] samples in blue, while the
simulations are represented by the green
spectra. The representative E/D values
for the unactivated, T-State, and
activated, R-state, enzyme samples are
listed in table 4.1. The electron spin
system in the enzyme samples is best
modeled as a mixture of two different S
Figure 4.1: Above is the g = 4 spectral features from
T
= 3/2 spin systems, with each species
the CW-EPR spectra of (a) PheH [L-phe] and (b)
R

PheH [L-phe] collected at 4K.

being defined by a specific E/D value.

The speciation between the enzyme samples seems to be consistent and exhibits only a small
difference (0.0013) in one of the two species. However, there is an observed difference in the

!

133!

3+

ratio of the species. This is most likely due to the presence of Fe

7

R

in [FeNO] -PheH [L-phe],

which is indicated by the spectral feature at g = 4.3. Although there were two species present in
the CW-EPR spectra, the HYSCORE analysis assumed that 0.0330 was the only species present
Table!4.1:!Summary!of!the!CW8EPR!results!of!the!T8!and!R8!State!Enzyme!Samples!
!
Sample
♯

7
7

2:1

0.0330:0.0250

1:1

0.0330:0.0263

T

♯

E/D Values

R

(a) [FeNO] -PheH [L-phe]
(b) [FeNO] -PheH [L-phe]
*
♯

*

Mixture Ratio

!
3, 4
A D value of 10 cm (300,000 MHz) was assumed for the simulation
-1

corresponds to figure 3.8
!

for both enzyme samples. The validity of this assumption was demonstrated by the negligible
frequency difference of the correlations observed between the two species in the HYSCORE
spectra. With the same hyperfine parameters, the different E/D values seem to negligibly alter
the orientations selected at these intermediate magnetic fields. With the description of the
electron spin system completed we can begin the pulsed EPR analysis of the allosteric forms of
PheH.

!

134!

7

T

7

R

Section 4.3.2: HYSCORE Data of [FeNO] -PheH [L-phe] and [FeNO] -PheH [L-phe]

!
7
T
Figure 4.2: The HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at (a) 240, (b)
260, and (c) 280 mT using a microwave frequency of 9.68 GHz and temperature of 4 K.

!

135!

!

HYSCORE is a pulsed EPR experiment that correlates the spin manifold frequencies,
which is a combined result of the nuclear Larmor and Hyperfine interactions, perpendicular to
the diagonal (typically at the nuclear Larmor frequency). This 2 dimensional experiment
deconvolves some of information by separating the correlated peaks into regions based on the
types of magnetic nuclei in the sample. Investigating the role of water ligands means that here
1

the interest is on H couplings (I=1/2). Their correlations will be focused about the diagonal
1

position between 10 and 12 MHz, which encompasses the nuclear Larmor frequencies of the H
that are associated to the magnetic field strengths 240, 260, and 280 mT. The HYSCORE
7

T

spectra collected at these magnetic field strengths for [FeNO] -PheH [L-phe] can be seen in
1

figure 4.2. From the figure three sets of correlations peaks associated to H can be seen, which in
this work will be define as P1, P2, and P3 as described by figure 4.3. The frequency positions at

!
!
7
T
Figure 4.3: The (left) HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at 240
1

mT, with the (right) position labels used to describe the H correlations observed in the
spectra.
!

136!

each magnetic field strength for the three points defined in figure 4.3 are listed in table 4.2.
Because the correlation pairs are approximately symmetric about the diagonal, the peaks above
!

7
T
Table!4.2:!The!Frequency!Positions!for!1H!Correlated!Peaks!of![FeNO] -PheH [L-phe]

in!the!HYSCORE!Spectra!at!Each!Magnetic!Field!

!

P1

P2

P3

B (mT)

d1F1

d1F2

d2F1

d2F2

d3F1

d3F2

240

4.72

17.42

7.70

17.45

8.20

13.00

260

6.51

17.58

8.47

17.90

8.50

14.25

280

8.46

17.58

9.77

17.90

10.20

15.30

the diagonal will be used to designate the correlated pair.
Upon inspection of these features, it can easily be observed that P1 and P2 peaks are
raised far off the ‘anti-diagonal’ (i.e. the line perpendicular to the diagonal at the nuclear Larmor
frequency). This indicates that these couplings are the result of a strong dipolar coupling. Bound
water ligand protons would produce such dipolar couplings, so these couplings could be the
1

result of water ligands. Other H couplings observed in the HYSCORE spectra can be seen near
P3. These protons are in a position that is commensurate with histidine ligand protons; however,
it is unclear what other proton couplings may be convolved within this position. The rest of the
features in these spectra seem to be focused along the diagonal, which is contaminated with 3pulse ESEEM and will not be analyze in this treatment.

!

137!

1

7

R

Similarly the H correlations from the activated form of PheH, [FeNO] -PheH [L-phe],
can be investigated through HYSCORE. The HYSCORE spectra from the activated form of

!
!
7
R
Figure 4.4: The HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at (a) 240,
(b) 260, and (c) 280 mT using a microwave frequency of 9.68 GHz and temperature of 4
K.

!

138!

7

!

R

Figure 4.5: The (left) HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at 240
mT, with the (right) positions used to investigate the two bound water proton peaks
labeled.
phenylalanine hydroxylase collected at magnetic field strengths of 240, 260, and 280 mT are
7

T

shown in figure 4.4. Just as with the unactivated form of PheH, [FeNO] -PheH [L-phe], the
7

R

1

activated form, [FeNO] -PheH [L-phe], has three distinct sets of H correlations within the
HYSCORE spectra that are defined here as P1, P2, and P3 in figure 4.5 and listed in table 4.3.

!
Table!4.3:!The!Frequency!Positions!for!the!Upper!left!Correlated!Peaks!of!
7
R
[FeNO] -PheH [L-phe] in!the!HYSCORE!Spectra!at!Each!Magnetic!Field!
!
P1

P2

P3

B (mT)

d1F2

d2F1

d2F2

d3F1

d3F2

240

4.72

17.42

7.70

17.45

8.30

13.43

260

6.59

17.58

9.15

17.20

8.54

14.40

280

!

d1F1

8.46

17.50

10.25

17.33

9.52

14.89

139!

Comparing only the frequency position of these three correlations between the T- and R-states
1

suggest little happens to H positioning during rearrangement. However, comparing the relative
intensity of P1 and P2 to that of P3 between the allosteric forms reveals a decrease in the
intensity of P1 and P2 during the activation of PheH, figure 4.6. To understand the significance
1

of this change, assignment of H correlation pairs needs to be made.

!
!
7
T
Figure 4.6: The (left) HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at 240
7
R
mT, with the (right) HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at 240
mT. The color bar in the middle defines the intensity scale for both contour plots.

!

140!

7

T

Section 4.3.3: Assignment of P1 and P2 in [FeNO] -PheH [L-phe] and
7

R

[FeNO] -PheH [L-phe]
Insight into the role water ligands plays in the allosteric activation of PheH using
1

HYSCORE requires assignment of H correlations that relate to water ligands. To accomplish
7

this task three [FeNO] -(N2Ox)(H2O)3-x model complexes were used, figure 4.7. These
7

7

7

complex, [FeNO] -(N2O1)(H2O)2, [FeNO] -(N2O2)(H2O)1, and [FeNO] -(N2O3), have two,

!
7
7
Figure 4.7: The three model complexes (a) [FeNO] -(N2O1)(H2O)2, (b) [FeNO] 7
(N2O2)(H2O)1, and (c) [FeNO] -(N2O3)

!

7

one, and zero water ligands, respectively. Starting with an inspection of [FeNO] -(N2O2)(H2O)1
7

and [FeNO] -(N2O3) helps to demonstrate the result of completely dehydrating the first
coordination sphere, figure 4.8. By visually comparing these spectra it is easily seen that the
high-frequency correlation peaks, centered near (6 MHz, 18 MHz) and (18 MHz, 6 MHz), are
lost when the water ligand is removed; this loss of the high-frequency crosspeaks is

!

141!

!
!
7
Figure 4.8: The (left) HYSCORE spectrum of [FeNO] -(N2O2)(H2O)1 collected at 240
7
mT, with the (right) HYSCORE spectrum of [FeNO] -(N2O3) collected at 240 mT.
actually the loss of protons bound to the water ligand. Although the frequency position of the
7

7

high-frequency correlation (outlined in red) between the [FeNO] -(N2O1)(H2O)2 and [FeNO] (N2O2)(H2O)1 samples is maintained, figure 4.9, the relative intensity of the high-frequency
correlation pair to that of the correlation pair center at (8 MHz, 13 MHz) and (13 MHz, 8 MHz)
1

decreases by about a factor of 2. The loss of intensity illustrates that the high-frequency H
correlation observed in the models (outlined in red) direct correlates to the loss of a water ligand
7

7

between these model complexes, [FeNO] -(N2O)(H2O)2 and [FeNO] -(N2O2)(H2O)1.
In the HYSCORE spectra of the enzyme samples there are two high-frequency
correlation that are indicative of strong dipolar couplings; however, in the model complexes
there seems to be only one strong dipolar coupling. If you overlap one of the enzyme samples

!

142!

!
!
7
Figure 4.9: The (left) HYSCORE spectrum of [FeNO] -(N2O)(H2O)2 collected at 240 mT,
7
with the (right) HYSCORE spectrum of [FeNO] -(N2O2)(H2O)1 collected at 240 mT. The
color bar in the middle defines the intensity scale for both contour plots.
7

T

7

and a model complex, say [FeNO] -PheH [L-phe] and [FeNO] -(N2O)(H2O)2, it can be seen
that both high-frequency correlations (i.e. P1 and P2) observed in the enzyme sample are
1

positioned in such a fashion that they can be understood to be the result of H couplings assigned
to water ligands, figure 4.10. The distinct correlations observed in the enzyme sample and the
disperse contour in the model complexes make sense by remembering the samples are frozen in
solution. Without a rigid second coordination sphere, like in the models, the water ligands can
take on multiple orientations relative to the ZFS tensor. Conversely, the side chains in the
enzyme may help orient the water molecules through hydrogen bond contacts, which will limit
1

the positioning of the two H associated to water ligands and result in two resolved correlations.
1

In the next section theses H correlations attributed to bound waters, designated as P1 and P2 in
the enzyme samples, will be analytically investigated.

!

143!

7

T

Section 4.3.4: HYSCORE Analysis of [FeNO] -PheH [L-phe]
The individual peaks of the enzyme samples are positioned within the correlations
7

attributed to water ligand(s) observed in the [FeNO] -(N2O)(H2O)2 model complex (figure

!
7
R
7
Figure 4.10: The HYSCORE spectrum of [FeNO] -PheH [L-Phe] over [FeNO] (N2O)(H2O)2 collected at (a) 240, (b) 260, and (c) 280 mT using a microwave
frequency of 9.68 GHz and temperature of 4 K.
!

144!

!

4.10). The distinct peaks could result from aspects of the environment that restrict water ligand
conformation. For example, hydrogen bonding in the second coordination sphere could restrict
water ligand(s), which would limit the dipolar distances and orientations that are accessible. As a
consequence, the P1 and P2 correlations, defined in figure 4.3, are resolved so that they may be
fit individually. The frequency positions associated with the maximum intensity of P1 and P2 are
listed in table 4.2.
7

T

The CW-EPR spectrum of [FeNO] -PheH [L-phe] is characteristic of an S = 3/2 spin
system with an inhomogenously broadened absorption peak that spans from g = 4 to g = 2;

15

however, the high-frequency correlations are only resolved at 240, 260, and 280 mT. The origins
of this will be investigated later using simulations after the P1 and P2 correlations in the
HYSCORE spectra are analyzed. Analysis of these high-frequency correlations, P1 and P2, at the
three magnetic field strengths will result in the extraction of the hyperfine parameters (aiso, T, ρ,
α, β, and γ) associated to protons attached to ligand water(s). It is important to note that the Euler
angles will relate the protons back to the ZFS tensor. With the Dz axis defined by the Fe-NO
bond, as they were previously found to be nearly collinear, the Euler angles also relate to the
16

molecular structure.

g ⋅ 39.47
T= e
MHz ⋅ Å 3
3
r
eff

!

(eq 4.1)

145!

As described in chapter 2, the dipolar coupling is inversely proportional to the distance
between the coupled magnetic moment cubed, equation 4.1. A graphical representation of this
equation can be seen in figure 4.11 using ge = 2.0023, which is the free electron g-value. Adding
protons to the water ligands within the 1J8U crystal structure, using pymol, produced iron-proton
distances from 2.5 to 3 Å, which correspond to dipolar hyperfine couplings, T, of ~ 5 and 3 MHz
(dashed lines in figure 4.11), respectively. When determining the optimum reff by simulating the

Figure 4.11: The relation between the dipolar distance and the dipolar hyperfine
coupling as described by equation 4.1. This graph was created using a ge = 2.

!

HYSCORE spectra across three magnetic fields simultaneously, a range straddling the predicted
couplings from 1 to 10 MHz was used. With nothing to really judge the rhombicity, ρ, it was
varied from 0 to 1 (it’s full range). A non-zero rhombicity, means that in the analysis of the

!

146!

HYSCORE spectra all of the Euler angles (α, β, and γ) relating to the ZFS axis could potentially
be non-zero. Initially, the β-value was varied from 0° to 180° but the simulated spectra were
found to be symmetrical about 90°, which is why only one angle is reported. The other Euler
angles (i.e. α and γ) were varied from 0° to 180°. The last term that was analyzed by the fitting
routine was aiso and it was constrained to a value between 0 and 1 MHz. This range comes from
single crystal studies of bound water protons.

17

The simulations assumed only one proton to limit the time and complexity of the
2

analysis. From the fits, optimum parameters were found by minimizing the χ value with the
2

‘patternsearch’ function, equation 4.2. The χ values were calculated based on the frequency

N

∑ ⎡(sF1i − dF1i )2 + (sF2i − dF2i )2 ⎤
⎢
⎥
⎣
⎦

χ2 = i

(eq 4.2)

N

position from the most intense portion of the upper left high-frequency correlation peaks. In
equation 4.2, F1 represents ‘Frequency 1’ (the x-value) and F2 corresponds to ‘Frequency 2’ (the
y-value). A prefix of ‘s’ or ‘d’ is added to the frequency labels, which define the values as
correspond to either the simulation or experimental data, respectively. N represents the total
number of magnetic fields analyzed, which is three in this work. The data values used in
1

equation 4.2, dF1i and dF2i, correspond to the frequency positions of the H peaks that are
labeled P1 and P2 at each magnetic field strength for the enzyme samples. The experimental data
7

T

7

points used are listed in table 4.2 and table 4.3 for [FeNO] -PheH [L-phe] and [FeNO] R

PheH [L-phe], respectively.

!

147!

!!
!
7
T
Figure 4.12: The (left) HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at 240
mT, with the (right) simulated points layered onto the experimental spectrum.
!

!
!
7
T
Figure 4.13: The (left) HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at 260
mT, with the (right) three simulated points layered onto the experimental spectrum.!

!

148!

!

7

!

T

Figure 4.14: The (left) HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at 280
mT, with the (right) simulated points layered onto the experimental spectrum.
7

T

After running the fitting routine for [FeNO] -PheH [L-phe] using the three magnetic
field strengths, 240, 260, and 280 mT, the optimum parameters were used to simulate P1 and P2,
figures 4.12, 4.13, and 4.14. The simulations for P1 and P2 were added together at each field
position and shown overlaid on the experimental data to allow for a visual inspection of the fit.
Table 4.4 shows the optimum values that were obtained from the fit. To judge the parameters the
standard deviations (σ) were calculated using equation 4.3.
2

18

The x and y values in the equation

2

refer to χ verse parameter plots. From theses χ verse parameter plots, standard deviations for
y −y
σ =± 1 2
2

⎛
χ2 ⎞
y = y = ⎜ 1+
⎟υ
1 2 ⎜
υ ⎟
⎝
⎠
x ≠x
1
2

!

149!

(eq 4.3)

!

!
7
T
Table!4.4:!Best8fit!Hyperfine!Parameters!for!the![FeNO] -PheH [L-phe] sample at
the Two Positioned Polled!
P1

P2

Best-fit

σ

aiso (MHz)

0.35

T (MHz)

*

*

Best-fit

σ

±0.25

0.85

±0.5

4.40

±0.15

4.90

±0.05

ρ

1

-

1

-

α (°)

35

-

51

-

β (°)

66

±0.3

76

±0.6

γ (°)

38

±25

50

±15

2

χ
*

0.021

0.192

: A dash (-) represents that the standard deviation could not be determined !

aiso, T, β and γ were determined. However, the standard deviations for ρ and α included the
entire range of these values, which suggests that HYSCORE is not sensitive to them. Based on
2

percent error of aiso, T, β and γ, the parameters T and β had the largest affect on the χ value.
The optimum values for T determined from this analysis lead to the two different dipolar
coupling values (T), which were measured as 4.40 and 4.90 MHz for P1 and P2, respectively.
These values correspond to dipolar distances of 2.62 and 2.52 Å. The β angles determined for
these points were 66 and 76°, respectively.

7

R

Section 4.3.5: HYSCORE Analysis of [FeNO] -PheH [L-phe]
7

R

Just as with the T-state enzyme sample from section 4.3.4, [FeNO] -PheH [L-phe]
exhibits two distinct contour features associated to the bound water protons. The maximum
intensity of these peaks are marked as P1 and P2 in figure 4.5 and have a frequency position at

!

150!

each magnetic field that are provided in table 4.3. A quick comparison of the P1 and P2
correlations provided in table 4.2 and 4.3, shows little difference in frequency position of the
crosspeaks between the two enzyme samples, which suggest the couplings are similar.
!
7
R
Table!4.5:!Best8fit!Hyperfine!Parameters!for!the![FeNO] -PheH [L-phe] sample at
the Two Positioned Polled!
!
P1

P2

Best-fit

σ

aiso (MHz)

0.30

T (MHz)

*

*

Best-fit

σ

±0.2

0.80

±0.2

4.33

±0.15

4.84

±0.35

ρ

1

-

0

-

α (°)

80

-

0

-

β (°)

65

±1.5

78

±0.4

γ (°)

60

±25

4

±15

2

χ
*

0.022

0.210

: A dash (-) represents that the standard deviation could not be determined !

By running the HYSCORE analysis, the optimum parameters were determined to have a
2

χ value of 0.022 and 0.210 for points P1 and P2, respectively. The optimum parameters for the
two frequency positions analyzed are listed in table 4.5. From the table, the dipolar coupling
value describing the first frequency point is 4.33 MHz, while the second point is best fit by 4.84
MHz. These couplings correspond to distances of 2.63 and 2.54 Å, while the β values are 64°
7

R

and 78°, respectively, which is very close to the parameters determined for [FeNO] -PheH [Lphe]. Using these parameters, the HYSCORE simulations at each magnetic field strength are
shown in figure 4.15, 4.16, and 4.17.

!

151!

!

!!
7
R
Figure 4.15: The (left) HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at 240
mT, with the (right) simulated points layered onto the experimental spectrum.
!

!

7

R

Figure 4.16: The (left) HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at 260
mT, with the (right) simulated points layered onto the experimental spectrum.
!

!

152!

!

!

!
7
R
Figure 4.17: The (left) HYSCORE spectrum of [FeNO] -PheH [L-phe] collected at 280
mT, with the (right) three simulated points layered onto the experimental spectrum.
In this analysis, the correlation peaks were treated independently to extract hyperfine
parameters. This approximation was implemented to help simplify the fitting process. However,
by extending the simulations to include two protons with the parameters observed in the one
proton fittings of the HYSCORE spectra, figure 4.18, 4.19, and 4.20, reveals that this technique
is fairly accurate. Interestingly, from the two proton HYSCORE simulations, it seems as though
combination of the strong dipolar couplings produce more peaks within the spectra than just the
high-frequency correlations analyzed. One such set of peaks fall near the correlation defined as
P3, which seems to fall with in the general shape the P3 contour observed in the experimental
data. Aside from these correlations, there is an autopeak seen along the diagonal at
approximately (18 MHz, 18 MHz), which could be the result of a hyperfine sum combination

!

153!

(ωα+ωβ). The large insensity of this peak observed in the experimental HYSCORE data could be
the result of contamination with 3-pulse ESEEM. Even with
!

Figure 4.18: Two proton HYSCORE (right) simulation compared to the experimental
7
R
(left) data for [FeNO] -PheH [L-phe] collected at 240 mT

!!

!

!

Figure 4.19: Two proton HYSCORE (right) simulation compared to the experimental
7
R
(left) data for [FeNO] -PheH [L-phe] collected at 260.
!

!

154!

!

!

Figure 4.20: Two proton HYSCORE (right) simulation compared to the experimental
7
R
(left) data for [FeNO] -PheH [L-phe] collected at 280 mT.!
phase cycling in HYSCORE, the echo resulting from 3-pulse ESEEM is not removed from the
echo, which creates an ESEEM pattern directed along the diagonal in these spectra. This could
be the origins of the large intensity of the combination peak. Overall, the similarities between the
two proton simulation and the experimental data suggest that this frequency based analysis can
be used to provide information about ligand water protons within non-heme iron enzymes using
the physiological relevant substrates.
With the reliablity of this technique established, the hyperfine parameters determined in
7

T

7

R

the anslysis for the [FeNO] -PheH [L-phe] and the [FeNO] -PheH [L-phe] can be used to
demonstrate the number of water ligands lost during the allosteric activation of PheH.
1

Previously, it was shown in figure 4.9 that the intensity of the H correlations associated to water

!

155!

ligands decreases upon activation of PheH with L-phe. Looking at a stack plot of the 3-pulse
ESEEM spectra of the T- and R-state, it is easy to see that there is an amplitude difference
between the allosteric states, figure 4.21a and 4.21b. Using the hyperfine information that was
extracted from the T- and R-state to generate a 4 and 2 proton simulation (figure 4.21c), to
mimic the respectively states, it can be seen that the ratio of the amplitude change is matched

!
7
T
7
Figure 4.21: The (a) 3-pulse ESEEM spectra of [FeNO] -PheH [L-phe] and [FeNO] R

PheH [L-phe] collected at 240 mT overlaid. (b) Enhanced view of the water proton
peaks and (c) a comparable view of the 4 and 2 proton simulations.
!
!

!

156!

!

reasonably well. This suggests that the allosteric activation of the enzyme, using L-phe,
results in the loss of a water ligand from the active site.

Section 4.3.6: CW-Analysis of the Model Complexes
The [FeNO]7-(N2Ox)(H2O)3-x model complexes

12

shown in figure 4.7 have a

ethylenediamine structure that has been substituted with acetyl group to restrict coordination to
the ferrous iron. These complexes allowed for the assignment of high-frequency correlation to
ligand water protons, which were used to extract hyperfine information from the HYSCORE
spectra of allosteric forms of PheH. To be complete, the model complexes are quantitatively
analyzed to show that the complex contour shape is the result of a distribution in both the T and
β parameters rather than simply orientation selection. Simulation of these high-frequency
correlation peaks in the HYSCORE spectra utilized the ‘Saffron’ module in the EasySpin
14

package

for MATLAB (The Mathworks, Natick, MA). Using EasySpin allowed for a
Table!4.6:!Summary!of!the!CW8EPR!results!of!the!Model!Complexes!
!
Sample
♯

7

♯

7

♯

E/D Value

7

(a) [FeNO] -(N2O1)(H2O)2

0.0083

(b) [FeNO] -(N2O2)(H2O)1

0.0121

(c) [FeNO] -(N2O3)(H2O)0
*
♯

0.0123

-1

A D value of 10 cm (300,000 MHz) was assumed for the simulation
corresponds to figure 4.22

!

!

*

157!

3, 4

parametric analysis of both the CW and pulsed spectra that ultimately resulted in the
determination of E/D, aiso, T, ρ, α, β, and γ for these model complexes.

13

7

Before any structural information related to the [FeNO] spin system can be determined
from pulsed spectra, the CW-EPR spectra first need to be analyzed. The CW-EPR spectra of the
three model complexes were previously analyzed in chapter 3. A summary of the results can be
found in table 4.6. Figure 4.22 shows the data (green) versus simulated spectra (blue) at the g =

!
7
7
Figure: 4.22: The CW-EPR spectra of (a) [FeNO] -(N2O1)(H2O) 2, (b) [FeNO] 7
(N2O2)(H2O)1, and (c) [FeNO] -(N2O3) (blue) with the simulated spectra (green)
corresponding to the E/D values in table 4.1.
!

158!

!

4. Beyond the E/D values, the experimental data was simulated using a slightly anisotropic gtensor that equates to [2.015 2.0023].
For the purpose of this work, the samples were thought to consist of just one type of
paramagnetic spin system; however, better fits of the CW-EPR spectra were obtained by utilizing
two different spin centers. The modest difference in the E/D values obtained when assuming the
presence of multiple spin systems did not change the positioning of the correlation peaks in the
simulated HYSCORE spectra. Therefore, using two different spin systems when simulating the
pulsed EPR spectra creates no advantage in the analysis; but adds to the total time of the
simulation.

7

Section 4.3.7: HYSCORE analysis of [FeNO] -(N2O2)(H2O)1
Out of the three model complexes, only two of them exhibited the high-frequency
7

7

correlation peaks in their HYSCORE spectra, [FeNO] -(N2O)(H2O)2 and [FeNO] 7

(N2O2)(H2O)1, that was previously attributed to water ligand proton(s). [FeNO] -(N2O2)(H2O)1
is analyzed first because with only one ligand water there are less spin Hamiltonian parameters
to take into consideration. The rational is that this reduces the complexity of the analysis.
However, unlike the enzyme samples the waters coordinated to the model complexes are not
held into specific orientations, which results in one large disperse contour peak rather than two
distinct couplings as previously seen in the enzyme samples. The convolution of these water

!

159!

!
7
Figure 4.23: The (left) HYSCORE spectrum of [FeNO] -(N2O2)(H2O)1 collected at 240
mT, with the (right) positions used to investigate the dispersion of the water ligands
labeled.

!

couplings creates a scenario where only a range of hyperfine values are accessible through the
analysis of the HYSCORE spectra.
7

The HYSCORE spectrum for [FeNO] -(N2O2)(H2O)1 collected with a magnetic field
strength of 240 mT and a temperature of 4 K is shown in figure 4.23. Outlined in red is a highfrequency crosspeak that was assigned to bound water protons earlier in section 4.3.3. The other
correlation pair, seen at ~ (8 MHz, 13 MHz), is from other ligand protons in the sample and
exhibit a weaker dipolar coupling relative to the bound water protons. This means they are
further away from the paramagnetic center. In this work, the analysis will focus on the highfrequency correlated peaks associated with the bound water ligands only. To accurately describe
the experimental HYSCORE spectra for these model complexes, three points on the upper left
high-frequency correlation peak were fit in a similar manor as was done before with P1 and P2

!

160!

!
Table!4.7:!The!Frequency!Positions!for!the!Upper!left!Correlated!Peak!in!the!
HYSCORE!Spectra!at!Each!Magnetic!Field!
!
P1

P2

P3

B (mT)

d1F1

d1F2

d2F1

d2F2

d3F1

d3F2

240

4.52

18.19

5.86

17.58

7.20

16.72

260

6.83

17.90

7.48

17.57

8.46

16.93

280

8.79

17.70

9.27

17.45

9.64

17.01

from the enzyme samples. The frequency position of these points can be seen in figure 4.23 and
are listed in table 4.7.
7

Fitting the frequency positions P1, P2, and P3 in the HYSCORE spectra of [FeNO] (N2O2)(H2O)1 collected at 240 mT, 260 mT, and 280 mT resulted in the optimized parameters
with their standard deviations listed in table 4.8. With a low percent error (<5%), the T and β
!
7
Table!4.8:!Best8fit!Hyperfine!Parameters!for!the![FeNO] -(N2O2)(H2O)1 Model
Complex at the Three Positioned Polled!
!
P1

P2

Best-fit

σ

aiso (MHz)

0.3

T (MHz)

*

P3

Best-fit

σ

±0.08

0.3

4.7

±0.05

ρ

0

α (°)

*

*

Best-fit

σ

±0.18

0.3

±0.08

4.7

±0.10

4.45

±0.15

-

1

-

1

-

60

-

60

-

52

±60

β (°)

66

±0.19

71

±0.88

76

±0.35

γ (°)

0

±20

0

±50

0

±50

2

χ
*

!

0.064

0.0076

0.155

: A dash (-) represents that the standard deviation could not be determined !
161!

parameters are the most reliable hyperfine parameters that can be extracted from the HYSCORE
spectra. Interestingly, at P3 a standard deviation for α was able to be determined but the percent
error was >100%, which suggests the experiment is not sensitive to it. Simulations revealed a
dipolar coupling range from 4.45 - 4.7 MHz in the sample, which corresponds to effective
distances of 2.60 and 2.56 Å, respectively. Across the three points analyzed, the β angles,
relative to the ZFS axis, have a range that spans from 66 - 76°, while the aiso seems to stay near
0.3 MHz. The rest of the hyperfine parameters (ρ, α, and γ) have too large of a deviation to be
confidently assigned a value. Using the optimized parameters listed in table 4.8 produces the
HYSCORE simulations seen in figure 4.24, 4.25, and 4.26, which correspond to magnetic field
strengths of 240, 260, and 280 mT.

!
!
7
Figure 4.24: The (left) HYSCORE spectrum of [FeNO] -(N2O2)(H2O)1 collected at 240
mT, with the (right) three simulated points layered onto the experimental spectrum.

!

162!

!
!
7
Figure 4.25: The (left) HYSCORE spectrum of [FeNO] -(N2O2)(H2O)1 collected at 260
mT, with the (right) three simulated points layered onto the experimental spectrum.
!

!
7
Figure 4.26: The (left) HYSCORE spectrum of [FeNO] -(N2O2)(H2O)1 collected at 280
mT, with the (right) three simulated points layered onto the experimental spectrum.
!

!

163!

!

7

Section 4.3.8: HYSCORE Analysis of [FeNO] -(N2O)(H2O)2
7

[FeNO] -(N2O)(H2O)2 is another ferrous iron model complex, figure 4.7a, which has
7

waters ligands present. Using the previous HYSCORE analysis of [FeNO] -(N2O2)(H2O)1 as a
guide, the hyperfine information can be extracted from the HYSCORE spectra using a frequency
7

position analysis. Similarly in this model complex, [FeNO] -(N2O)(H2O)2, the high-frequency
correlation peaks are only resolved in the HYSCORE spectra collected with magnetic field
strengths of 240, 260, and 280 mT; Therefore, the frequency based analysis will occur at these
three magnetic field strengths.
7

As was done previously, the upper left high-frequency peak of [FeNO] -(N2O)(H2O)2
can be emulated using three points, figure 4.27, that outline the contour shape of the high-

!
7
Figure 4.27: The (left) HYSCORE spectrum of [FeNO] -(N2O)(H2O)2 collected at 240
mT, with the (right) positions used to investigate the dispersion of the water ligands
labeled.
!

164!

!

!
7
Table!4.9:!The!Frequency!Positions!for!the!Upper!left!Correlated!Peak!of![FeNO] (N2O)(H2O)2 !in!the!HYSCORE!Spectra!at!Each!Magnetic!Field!

!

P1

P2

P3

B (mT)

d1F1

d1F2

d2F1

d2F2

d3F1

d3F2

240

3.91

18.01

5.86

17.58

7.32

17.09

260

6.35

18.07

7.32

17.58

8.30

17.58

280

7.81

18.07

8.79

17.58

10.25

17.09

frequency correlation. The frequency positions of these three points at each magnetic field
strength are listed in table 4.9. From the table it can be seen that each point seems to travel
parallel to the x-axis as the magnetic field strength is changed. The observed trajectory is another
indication that the sample contains a range of T and β values.
Using the MATLAB function ‘patternsearch,’ to fit the three points (P1, P2, P3)
7

Table!4.10:!Best8fit!Hyperfine!Parameters!for!the![FeNO] -(N2O)(H2O)2 Model
Complex at the Three Positioned Polled!
!
P1

P2

Best-fit

σ

aiso (MHz)

0.3

T (MHz)

*

P3

Best-fit

σ

±0.3

0.13

4.51

±0.2

ρ

1

α (°)

*

*

Best-fit

σ

±0.35

0.55

±0.1

4.73

±0.18

4.70

±0.1

-

1

-

1

-

63

±80

60

±30

62

±55

β (°)

63

±1.8

70

±0.5

77

±0.4

γ (°)

0

-

0

-

3

±20

2

χ
*

!

0.153

0.058

0.098

: A dash (-) represents that the standard deviation could not be determined !

165!

listed in table 4.9 at each magnetic field strength, produced three sets of optimized parameters.
The optimized parameters and their corresponding standard deviations are listed in table 4.10,
while the simulated spectra are shown along side of the data in figure 4.28, 4.29, and 4.30. From
the figures, it can be seen that the three-point analysis mimics the contour shape well. In these
7

simulations aiso, T, α, and β have a standard deviation for [FeNO] -(N2O)(H2O)2 that can be
estimated. However, only the dipolar coupling and β angles have a percent error less than 100%
for all the points fit. The HYSCORE analysis resulted in dipolar couplings that range from 4.51 –
7

4.70 MHz in the [FeNO] -(N2O)(H2O)2 model complex. These couplings correspond to
distances of 2.60 and 2.56 Å, respectively. The β angles associated to this dipolar coupling range
are distributed between 63 and 77°, respectively.

!

7

Figure 4.28: The (left) HYSCORE spectrum of [FeNO] -(N2O)(H2O)2 collected at 240
mT, with the (right) three simulated points layered onto the experimental spectrum.
!

166!

!

!

7

!

Figure 4.29: The (left) HYSCORE spectrum of [FeNO] -(N2O)(H2O)2 collected at 260
mT, with the (right) three simulated points layered onto the experimental spectrum.
!

!

!
7
Figure 4.30: The (left) HYSCORE spectrum of [FeNO] -(N2O)(H2O)2 collected at 280
mT, with the (right) three simulated points layered onto the experimental spectrum.
!

!

167!

Section 4.3.9: Visibility of the High-Frequency Correlation Peaks
As can be seen at 180 and 340 mT (figures 4.31 and 4.32, respectively), the highfrequency peaks seemingly disappear at magnetic field strengths near the ends of the CW-EPR
7

spectrum. Using the optimized hyperfine parameters from table 4.8 for P1 of [FeNO] (N2O2)(H2O)1, the lack of the high-frequency peak can be investigated.
7

Using the parameters previously determined for [FeNO] -(N2O2)(H2O)1, a spectrum can
be simulated for a HYSCORE experiment collected with a magnetic field strength of 180 mT.
7

Figure 4.31 shows the HYSCORE spectrum for [FeNO] -(N2O2)(H2O)1 collected at 180 mT on
the left with the simulated data overlaid on the right. The simulated spectrum depicts a
correlation peak centered on (4 MHz, 16 MHz), while it is absent in the data. This observation

!
!
7
Figure!4.31:!The!(left)!HYSCORE!spectrum!of![FeNO] -(N2O2)(H2O)1collected!at!
180!mT,!with!the!(right)!corresponding!simulation!using!parameters!listed!in!table!
4.8.!

!

168!

can be explained based on the intensities of these proton crosspeaks from the simulated spectra
7

of [FeNO] -(N2O2)(H2O)1. In the simulated spectra, the intensities of the crosspeaks from the
previously investigated magnetic field strengths, 240, 260, and 280 mT, are seen to be
approximately the same size. Compared to these intensities, the magnitude of the intensity of the
crosspeak centered at (4 MHz, 16 MHz) in the simulated spectrum (figure 4.31) is approximately
half the size. The simulation has no noise, so such a drastic drop in the intensity could result in
this crosspeak being buried under the noise floor in the data. Although this peak may be under
the noise floor, there is another peak that is present in the simulated HYSCORE spectra at 180
mT, which is centered at (4 MHz, 12 MHz). These crosspeaks have an intensity that is double

!
!
7
Figure!4.32:!The!(left)!HYSCORE!spectrum!of![FeNO] -(N2O2)(H2O)1collected!at!340!
mT,!with!the!(right)!corresponding!simulation!using!parameters!listed!in!table!4.8.!
the magnitude of the high-frequency correlation peaks observed in the simulations at 240, 260,
7

and 280 mT. The HYSCORE spectrum of the [FeNO] -(N2O2)(H2O)1 (figure 4.31) have high

!

169!

intensity correlation peaks centered at (6 MHz, 11 MHz) that is commensurate with couplings
19

previously observed for histidine protons . Although the model complex does not have
histidines, it does have crosspeak resulting from the proton coupling at (4 MHz, 12 MHz) for
spectra collected at 180 mT, most likely from the N2O2 ligand protons, which could be
convolved into the high intensity correlation.
The high-frequency correlation peak assigned to the bound water protons is no longer
resolved in the HYSCORE spectra collected at magnetic field strengths near the g = 2 region of
the CW-EPR spectrum. Using the parameters in table 4.8 listed for P2, the HYSCORE spectrum
at a magnetic field strength of 340 mT was simulated and compared with data collected at the
same field, figure 4.32. The simulation only predicts one set of crosspeaks, which are center at
(13 MHz, 16 MHz). These peaks seem to be convolved with the histidine protons just like
before. The combination of multiple couplings in a similar region and the drop in intensity
accounts for the limited viewing range of the correlations assigned to bound water protons.

Section 4.4: Discussion
In recent years, theoretic investigations into the formation of the high-valent iron(IV)-oxo
20, 21

species have predicted multiple mechanisms.

The divergent point between the predicted

mechanisms occurs as L-phenylalanine activates PheH. These computational studies were seeded
with information derived from spectroscopic data. Within the literature the role water ligands
play or even their existence is unclear because it is hard to directly observed them. From XAS

!

170!

and MCD spectroscopy, the activation results in no change from the distorted octahedral
geometry. In the presence of BH4, the activation changes the geometric structure of the ligands
3

about the iron from a distorted octahedral to a distorted square pyramidal structure. If it is
assumed that this geometric change is the observation of one water ligand being removed from
the first coordination sphere, then there are two waters still present. On the other hand, x-ray
crystallography, with thienylalanine and norluecine, not the physiological substrate, resulted in
no ligand waters present.

8, 9, 22

The crystal structures show Glu 330 chelating the ferrous iron,

which is a possible reason the XAS and MCD predict no change in the active site’s geometry
upon the activation of PheH. Figure 4.21 shows that a water ligand is lost during the activation of
PheH with L-phe from the T- to the R-state. This makes the chealation of Glu 330 an attractive
rearrangment during the transition from the T- to the R-state. With the pterin species present, the
rearrangment of Glu 330 and the change in the geometry accounts for the loss of two waters
from the active site. The last water was shown to be displaced by NO, an O2 surrogate, using
17

O line broadening effects;

23

this observation was confirmed in chapter 3. Chapter 3 helped to

tie the spectrocopic literature together by counting the number of water ligands present in a
qualitative fashion. In this work, a frequency position analysis of HYSCORE spectra was able to
demonstrate characteristics of bound water protons and provides a insight into the role of water
ligands in non-heme iron enzymes.
7

The dipolar distances and β angles listed in table 4.11 were determined for [FeNO] 7

7

T

7

R

(N2O2)(H2O)1, [FeNO] -(N2O)(H2O)2, [FeNO] -PheH [L-phe], and [FeNO] -PheH [L-phe]

!

171!

with a frequency based HYSCORE analysis. The dipolar distances and β angles associated with
the model complexes are reported as
!
Table!4.11:!Summarized!Dipolar!Distances!and!β!angles!
reff!(Å)!

β!!(°)!

7

2.56!–!2.60!

66!8!76!or!104!8!114!

7

2.56!–!2.60!

63!8!77!!or!103!8!117!

Sample!
[FeNO] -(N2O2)(H2O)1!
[FeNO] -(N2O)(H2O)2!
7

T

2.52!/!2.62!

66!/!76!or!104!/!114!

7

R

2.54!/!2.63!

64!/!78!or!102!/!116!

[FeNO] -PheH [L-phe]!
[FeNO] -PheH [L-phe]!
!

a range because the high-frequency correlation contour shape had to be described using a range
of T and β values. Conversely, in the enzyme samples two distinct sets of high-frequency
crosspeaks were resolved, so for these samples two distinct dipolar distances and β angles were
reported.
With orientation selected HYSCORE, only a small portion of orientaions are observed at
any given magnetic field strength. In order to determine the hyperfine parameters from these
spectra, a range of spectra collected at different magnetic fields need to be analyzed
simultaneously. The high-frequency correlation peaks assigned to bound water protons in both
the model complexes and enzyme samples are only seen with a limited range of magnetic field
strengths. An accurate potrayal of the hyperfine values becomes difficult with just a few points to
analyze. From this frequency based analysis, the only values that consistently had less than a
percent error of ~5% was T and β. The aiso value consistently had a large standard deviation, but

!

172!

the values were determined to be within the range of values previously observed for bound water
protons.

17

The rest of the parameters seemed to have a large range of values that satisfied the fit,

which is evident by the large standard deviation or the lack of a standard deviation.
As has been shown in the analysis, the standard deviation for the hyperfine parameters ρ,
α, and γ either could not be determined or were typically large, which suggests the experiment is
not very sensitive to these parameters. Previously in this chapter, values for ρ, α, and γ were
listed but these discrete values had little affect on the high-frequency correlations being
analyzed. In some cases, the ‘patternsearch’ fitting routine assigned non-zero values to these
parameters during optimization even though they have little effect on the frequency position. So
with the simulation program requiring an input for all parameters, the optimized ρ, α, and γ
values, zero or non-zero, determined by the fitting routine were used in the simulations to
generate figures. The majority of the bounds used for these parameters were typically found to be
2

within one standard deviation unit of the minimized χ value, which is why these parameters are
not reported in the discussion.
2

The standard deviations caluclated were based on plots of the χ verses the parameters.

24

2

χ values were calculated using the frequency position of the high-frequency correlation peaks.
The MATLAB fitting routine, ‘patternsearch,’ resulted in optimal parameters by minimizing the
2

χ values through a polling method. Upon close inspection, the frequency pair defined as P2 in
2

the model complexes had the smallest (best) χ values. This is because the central frequency
position of the model complexes high-frequency correlation was determined by figuring out

!

173!

where the maximum intensity occurred; while the side positions (P1 and P3) were picked
manually to follow the contour. Similarly, the frequency positions of P1 in the allosteric forms
7

T

7

R

2

of PheH, [FeNO] -PheH [L-phe] and [FeNO] -PheH [L-phe], have a smaller χ value relative
to P2. In both enzyme samples, the crosspeak centered at P1 seems to be more focused, while the
peak near P2 has some associated frequency dispersion that could be indicative another
population of couplings. It may be helpful to analyze the peak centered at P2 in a similar fashion
as the model complexes to determine a range of couplings that describe the contour shape. If the
2

2

χ values determined in this analysis are divided by the estimated variance squared (0.04 MHz ),
only the P2 positions from the model complexes and P1 positions from the enzyme samples are
approximately less than or equal to 1. Suggesting only their simulations predict frequency
positions within the estimated variance for all three magnetic field strengths.
The frequency based HYSCORE analysis was done using three different magnetic field
strengths, 240, 260, and 280 mT. At other magnetic field strengths, the resolution of bound water
protons is lost. Using the parameters determined in the analysis, the absence of the strong dipolar
couplings was reasoned. When working at magnetic fields near the g = 4 end of the CW-EPR
spectrum, two correlation peaks were predicted by the simulation with one centered at (4 MHz,
16 MHz) and the other at (4 MHz, 12 MHz). In the simulation, the intensity of the first crosspeak
dropped by a factor of 2 relative to the peaks predicted for magnetic field strengths near the
middle of the CW-EPR spectrum (240, 260, and 280 mT). With the simulation having no noise, a
drop in the intensity could put the peak below the noise floor of the data. The loss of intensity

!

174!

could be a result of an increased dipolar coupling (T). This dipolar coupling is increased by
reweighting it with the effective g value, which makes the coupling increase the closer as the
magnetic field strength approches the g = 4 region. As the dipolar coupling becomes larger so
does the hyperfine coupling. If the hyperfine frequency becomes two times larger than the
nuclear Larmour frequency, the correlations can move from the (+,+) to (+,-) quadrant of the
HYSCORE spectrum.

25

At the moment, with the spectra collected, no crosspeaks attributed to

bound water protons have been idendified in the (+,-) quadrant. However, a division of the
couplings between quadrants can reduce the intensity of the resulting correlation peaks. Another
explination for the loss of intensity could be the result of the dipolar copuling being so large near
the g = 4 region that the mixing of nuclear states is removed, which would remove the ESEEM
modulation. The other correlation, seen near the g = 4 region of the CW-EPR spectrum, is in a
position to overlap with proton crosspeaks from typical histidine protons.

19

In the case of the

model complexes, with protons associated to the substituted diethylamine (N2O2). On the other
end of the CW-EPR spectrum, near g = 2, the only predicted crosspeaks are positioned such that
the coupling associated to the substituted diethylamine (N2O2) are overlapping them. Just as in
the g = 4 region, this convolution makes it difficult to use these peaks to extract reliable
information.
Using the quantitative information obtained in the frequency-based analysis of the
HYSCORE spectra, the role of water ligand’s can be investigated. Initially starting with the 1J8U
crystal structure, the dipolar distances and relative orientations of the water bound protons

!

175!

extracted from this analysis can be used to build a water ligand surrounding the iron. The 1J8U
crystal structure depicts three waters coordinated to the ferrous iron in the presence of
tetrahydrobiopterin. With the position of the NO unknown, it is assumed that NO will displace
one of the water ligands to form the Fe—NO bond along one of the Fe—O bonds. It was
previously shown that the hydroxylation of the C4a position on the pterin species will form an Stype stereocenter,

26

which implies that O2 binds in the coordiation position of water 2 (as

22

defined by 1J8U) . As NO is an O2 surrogate, it will be assumed that the NO bond replaces
7

T

water 2. Using the information determined for the [FeNO] -PheH [L-Phe] system, presented in
table 4.11, and the knowledge that the Fe—NO bond forms nearly colinear to the Dz axis,

16

then

the model shown in figure 4.33 can be built. This figure shows cones that represent the possible
7

T

position of water bound protons based on the parameters extracted from [FeNO] -PheH [LPhe]. With four possible proton positions there are 10 configurations that exist when excluding

!
Figure!4.33:!Possible!position!of!water!bound!protons!based!on!the!parameters!
7
T
extracted!from!the!HYSCORE!spectra!of![FeNO] -PheH [L-Phe]!
!

176!

!

duplicates. However, it was shown previously that activation of the enzyme resulted in the loss
of a water ligand. If one water ligands is removed from the active site with the presence of NO
and another during the activation of the enzyme,then one water ligand is present in the R-state of
the enzyme. With the presence of two distinct couplings resulting from one water ligand, it is

!
!
Figure!4.34:!A!water!ligand!developed!from!the!hyperfine!parameters!observed!in!the!
HYSCORE!spectra!of![FeNO]78PheHT[L8phe].!

clear that configurations including both of the prontons on the same contour or symmetrical
contours can be elminated as possibilities. So if the cones were labeled 1, 2, 3, and 4 from the top
down in figure 4.33 the remaining possibilities would have protons on 1-2, 1-3, 2-4, and 3-4. To
maitain a resemblence of the accepted picture for water, 104.5° for the HOH angle and 0.958 Å

!

177!

for the OH bond lengths,

27

the possibilities that include both protons on the same hemisphere of

the iron can be eliminated (1-2 and 3-4). If they are both the same side either the angle or the
bond lengths will be extremely far off. With two pseudo-symmetric possibilities left, the
placement of the protons now becomes the question.
Utilizing only the 1-3 possibility, a ligand water molecule can be built from the
information extracted from the HYSCORE analysis of the enzyme complexes. A water molecule
will bind the ferrous iron through the lone pairs on the oxygen, assuming that this bond is 2 Å
long and located 90° from the Fe—NO bond, Dz axis, there is a scaffolding present to attach
protons. Manipulating the protons to resemble the picture of water described by theory results

27

in the water ligand shown in figure 4.34. The water ligand formed had O—H bond lengths of
1.08 and 1.10 Å, while the HOH bond angle is 115.6°. Although the water ligand formed does
not match the exact structure of water discussed, this gas-phase picture of water could change in
solution or when water ligates a transition metal ion. The water ligand was created using the
hyperfine parameters extracted from [FeNO]78PheHT[L8phe].!It!was!previously!determined!
that!two!the!T8state!of!the!enzyme!contains!two!water!ligands,!which!have!similar!
configurations.!By!placing!two!of!these!ligands!directly!into!the!coordination!spots!of!water!
1!and!water!3!from!1J8U,!the!active!site!takes!on!the!arrangement!as!observed!in!the!figure!
4.35.!The!figure!shows!that!these!water!ligands!form!possible!hydrogen!bonding!contacts!
with!the!ptering!carbonyl,!which!could!help!to!orient!it!for!hydroxylation.!

!

178!

!
!
Figure!4.35:!Water!1!(W1)!and!Water!3!(W3)!replaced!with!the!water!ligand!
created!using!the!hyperfine!parameters!extracted!from!the!HYSCORE!spectra.!

!

179!

BIBLIOGRAPHY

!

180!

BIBLIOGRAPHY

1.!Kaufman,!S.,!and!Mason,!K.!(1982)!Specificity!of!amino!acids!as!activators!and!substrates!
for!phenylalanine!hydroxylase,!J"Biol"Chem!257,!14667814678.!
2.!Kappock,!T.!J.,!and!Caradonna,!J.!P.!(1996)!Pterin8Dependent!Amino!Acid!Hydroxylases,!
Chem"Rev!96,!265982756.!
3.!Brown,!C.!A.,!Pavlosky,!M.!A.,!Westre,!T.!E.,!Zhang,!Y.,!Hedman,!B.,!Hodgson,!K.!O.,!and!
Solomon,!E.!I.!(1995)!Spectroscopic!and!Theoretical!Description!of!the!Electronic8
Structure!of!S=3/2!Iron8Nitrosyl!Complexes!and!Their!Relation!to!O82!Activation!by!
Nonheme!Tron!Enzyme!Active8Sites,!Journal"of"the"American"Chemical"Society!117,!
7158732.!
4.!Ye,!S.,!Price,!J.!C.,!Barr,!E.!W.,!Green,!M.!T.,!Bollinger,!J.!M.,!Jr.,!Krebs,!C.,!and!Neese,!F.!
(2010)!Cryoreduction!of!the!NO8adduct!of!taurine:alpha8ketoglutarate!dioxygenase!
(TauD)!yields!an!elusive!{FeNO}(8)!species,!J"Am"Chem"Soc!132,!473984751.!
5.!Lazarus,!R.!A.,!Dietrich,!R.!F.,!Wallick,!D.!E.,!and!Benkovic,!S.!J.!(1981)!On!the!mechanism!
of!action!of!phenylalanine!hydroxylase,!Biochemistry!20,!683486841.!
6.!Panay,!A.!J.,!Lee,!M.,!Krebs,!C.,!Bollinger,!J.!M.,!and!Fitzpatrick,!P.!F.!(2011)!Evidence!for!a!
High8Spin!Fe(IV)!Species!in!the!Catalytic!Cycle!of!a!Bacterial!Phenylalanine!
Hydroxylase,!Biochemistry!50,!192881933.!
7.!Hillas,!P.!J.,!and!Fitzpatrick,!P.!F.!(1996)!A!mechanism!for!hydroxylation!by!tyrosine!
hydroxylase!based!on!partitioning!of!substituted!phenylalanines,!Biochemistry!35,!
696986975.!
8.!Andersen,!O.!A.,!Flatmark,!T.,!and!Hough,!E.!(2002)!Crystal!structure!of!the!ternary!
complex!of!the!catalytic!domain!of!human!phenylalanine!hydroxylase!with!
tetrahydrobiopterin!and!38(28thienyl)8L8alanine,!and!its!implications!for!the!
mechanism!of!catalysis!and!substrate!activation,!J"Mol"Biol!320,!109581108.!
9.!Andersen,!O.!A.,!Stokka,!A.!J.,!Flatmark,!T.,!and!Hough,!E.!(2003)!2.0A!resolution!crystal!
structures!of!the!ternary!complexes!of!human!phenylalanine!hydroxylase!catalytic!
domain!with!tetrahydrobiopterin!and!38(28thienyl)8L8alanine!or!L8norleucine:!
!

181!

substrate!specificity!and!molecular!motions!related!to!substrate!binding,!J"Mol"Biol!
333,!7478757.!
10.!Shiman,!R.,!and!Gray,!D.!W.!(1980)!Substrate!Activation!of!Phenylalanine8Hydroxylase!8!
a!Kinetic!Characterization,!Journal"of"Biological"Chemistry!255,!479384800.!
11.!Fitzpatrick,!P.!F.!(2012)!Allosteric!regulation!of!phenylalanine!hydroxylase,!Archives"of"
biochemistry"and"biophysics!519,!1948201.!
12.!Cappillino,!P.!J.,!Miecznikowski,!J.!R.,!Tyler,!L.!A.,!Tarves,!P.!C.,!McNally,!J.!S.,!Lo,!W.!N.,!
Kasibhatla,!B.!S.!T.,!Krzyaniak,!M.!D.,!McCracken,!J.,!Wang,!F.,!Armstrong,!W.!H.,!and!
Caradonna,!J.!P.!(2012)!Studies!of!iron(II)!and!iron(III)!complexes!with!fac8N2O,!cis8
N2O2!and!N2O3!donor!ligands:!models!for!the!28His!18carboxylate!motif!of!non8
heme!iron!monooxygenases,!Dalton"Transactions!41,!566285677.!
13.!Stoll,!S.,!and!Schweiger,!A.!(2006)!EasySpin,!a!comprehensive!software!package!for!
spectral!simulation!and!analysis!in!EPR,!J"Magn"Reson!178,!42855.!
14.!Stoll,!S.,!and!Britt,!R.!D.!(2009)!General!and!efficient!simulation!of!pulse!EPR!spectra,!
Physical"Chemistry"Chemical"Physics!11,!661486625.!
15.!Rich,!P.!R.,!Salerno,!J.!C.,!Leigh,!J.!S.,!and!Bonner,!W.!D.!(1978)!Spin!382!Ferrous8Nitric!
Oxide!Derivative!of!an!Iron8Containing!Moiety!Associated!with!Neurospora8Crassa!
and!Higher!Plant8Mitochondria,!Febs"Letters!93,!3238326.!
16.!Aquino,!F.,!and!Rodriguez,!J.!H.!(2009)!Accurate!Calculation!of!Zero8Field!Splittings!of!
(Bio)inorganic!Complexes:!Application!to!an!{FeNO}(7)!(S=3/2)!Compound,!Journal"
of"Physical"Chemistry"A!113,!915089156.!
17.!Atherton,!N.!M.,!and!Horsewill,!A.!J.!(1979)!Proton!ENDOR!of!Cu(H2O)6!2+!in!
Mg(NH4)2(SO2)4.6H2O,!Molecular"Physics!37,!134981361.!
18.!Roe,!B.!P.!(2001)!Probability"and"statistics"in"experimental"physics,!2nd!ed.,!Springer,!
New!York.!
19.!Garcia8Rubio,!I.,!Fittipaldi,!M.,!Trandafir,!F.,!and!Van!Doorslaer,!S.!(2008)!A!
Multifrequency!HYSCORE!Study!of!Weakly!Coupled!Nuclei!in!Frozen!Solutions!of!
High8Spin!Aquometmyoglobin,!Inorganic"Chemistry!47,!11294811304.!

!

182!

20.!Bassan,!A.,!Blomberg,!M.!R.,!and!Siegbahn,!P.!E.!(2003)!Mechanism!of!dioxygen!cleavage!
in!tetrahydrobiopterin8dependent!amino!acid!hydroxylases,!Chemistry!9,!1068115.!
21.!Olsson,!E.,!Martinez,!A.,!Teigen,!K.,!and!Jensen,!V.!R.!(2011)!Formation!of!the!Iron8Oxo!
Hydroxylating!Species!in!the!Catalytic!Cycle!of!Aromatic!Amino!Acid!Hydroxylases,!
ChemNEur"J!17,!374683758.!
22.!Andersen,!O.!A.,!Flatmark,!T.,!and!Hough,!E.!(2001)!High!resolution!crystal!structures!of!
the!catalytic!domain!of!human!phenylalanine!hydroxylase!in!its!catalytically!active!
Fe(II)!form!and!binary!complex!with!tetrahydrobiopterin,!J"Mol"Biol!314,!2798291.!
23.!Han,!A.!Y.,!Lee,!A.!Q.,!and!Abu8Omar,!M.!M.!(2006)!EPR!and!UV8vis!studies!of!the!nitric!
oxide!adducts!of!bacterial!phenylalanine!hydroxylase:!Effects!of!cofactor!and!
substrate!on!the!iron!environment,!Inorganic"Chemistry!45,!427784283.!
24.!Roe,!B.!P.!(1992)!Probability"and"statistics"in"experimental"physics,!Springer,!New!York!
etc.!
25.!Schweiger,!A.,!and!Jeschke,!G.!(2001)!Principles"of"pulse"electron"paramagnetic"
resonance,!Oxford!University!Press,!Oxford.!
26.!Dix,!T.!A.,!Bollag,!G.!E.,!Domanico,!P.!L.,!and!Benkovic,!S.!J.!(1985)!Phenylalanine8
Hydroxylase!8!Absolute8Configuration!and!Source!of!Oxygen!of!the!4a8
Hydroxytetrahydropterin!Species,!Biochemistry!24,!295582958.!
27.!Csaszar,!A.!G.,!Czako,!G.,!Furtenbacher,!T.,!Tennyson,!J.,!Szalay,!V.,!Shirin,!S.!V.,!Zobov,!N.!
F.,!and!Polyansky,!O.!L.!(2005)!On!equilibrium!structures!of!the!water!molecule,!The"
Journal"of"Chemical"Physics!122,!2143058214310.!
!

!

183!