Electronic Structure of a Transient Histidine Radical in Liquid Aqueous Solution: EPR Continuous-Flow Studies and Density Functional Calculations

Article abstract of DOI:10.1021/jp9843454

Transient histidine radicals formed in aqueous solutions by oxidation of histidine with a Ti³⁺/H₂O₂ Fenton system at pH 2.0 have been studied by EPR using a fast continuous-flow setup and a dielectric ring resonator equipped with a mixing chamber. A histidine peroxy radical with a single EPR line at g = 2.0151 and a histidine cation radical with a complex hyperfine structure and g = 2.0023 have been detected. The hyperfine structure of the latter radical was analyzed by investigating two selectively deuterated histidines and using an EPR simulation and fit program for analysis of the spectra. Isotropic hyperfine coupling constants of two β-protons, three ring protons, and two nitrogen nuclei have been determined in this way and assigned to a histidine-OH adduct cation radical. Density functional theory (DFT) calculations at the B3LYP and PWP86 levels have been performed on protonated cation radicals of 4-ethylimida zole (as histidine models), yielding isotropic hyperfine coupling constants for three different positions of OH addition. The C5 position for OH addition (a 5-oxohistidine cation radical) is clearly supported by the calculated hyperfine coupling constants. The agreement between DFT and EPR is further improved when hydrogen-bonding interactions to the N1 and C2 protons are introduced in the calculations.

Full text of DOI:10.1021/jp9843454

J. Phys. Chem. A 1999, 103, 1283-1290
1283
Electronic Structure of a Transient Histidine Radical in Liquid Aqueous Solution: EPR
Continuous-Flow Studies and Density Functional Calculations
,†
‡
§
†
G u1 nter Lassmann,* Leif A. Eriksson, Fahmi Himo, Friedhelm Lendzian, and
Wolfgang Lubitz†
Max-Volmer-Institut fur Biophysikalische Chemie und Biochemie, Technische UniVersit a¨ t Berlin,
D-10623 Berlin, Germany, Department of Quantum Chemistry, Uppsala UniVersity, Box 518,
S-75120 Uppsala, Sweden, and Department of Physics, Box 6730, Stockholm UniVersity,
S-11385 Stockholm, Sweden
ReceiVed: NoVember 3, 1998
3
+
Transient histidine radicals formed in aqueous solutions by oxidation of histidine with a Ti /H
O
2 2
Fenton
system at pH 2.0 have been studied by EPR using a fast continuous-flow setup and a dielectric ring resonator
equipped with a mixing chamber. A histidine peroxy radical with a single EPR line at g ) 2.0151 and a
histidine cation radical with a complex hyperfine structure and g ) 2.0023 have been detected. The hyperfine
structure of the latter radical was analyzed by investigating two selectively deuterated histidines and using an
EPR simulation and fit program for analysis of the spectra. Isotropic hyperfine coupling constants of two
â-protons, three ring protons, and two nitrogen nuclei have been determined in this way and assigned to a
histidine-OH adduct cation radical. Density functional theory (DFT) calculations at the B3LYP and PWP86
levels have been performed on protonated cation radicals of 4-ethylimidazole (as histidine models), yielding
isotropic hyperfine coupling constants for three different positions of OH addition. The C5 position for OH
addition (a 5-oxohistidine cation radical) is clearly supported by the calculated hyperfine coupling constants.
The agreement between DFT and EPR is further improved when hydrogen-bonding interactions to the N1
and C2 protons are introduced in the calculations.
1
. Introduction
are still missing. Meanwhile, a tyrosine-based radical is
discussed alternatively in this system.
1
6
In many metal enzymes histidine is a ligand of a metal ion
in the active site and influences the redox properties of these
complexes. Radical enzymes exhibiting radicals of certain amino
acid side chains, which are essential for the catalytic function,
also frequently contain redox-active metal complexes as a part
of the functional electron/radical transfer pathway. Catalytically
essential amino acid radicals in radical enzymes have been
Nevertheless, there is no direct EPR spectroscopic evidence
for a protein-linked histidine radical as reported for tyrosine
2
,11
radicals in R2 protein of RR
and tryptophan radicals in
1
7,18
mutants of R2 proteins.
This could have several reasons.
(i) One main reason may be the higher redox potential of
histidine (1.4 V) in comparison with tyrosine (0.85 V) and
tryptophan (0.90 V; measured in amino acids at pH 4.0 with
1
,2
observed from tyrosine in ribonucleotide reductase (RR),
3
4
19
prostaglandin H synthase, photosystem II, and galactose
electrooxidation using a graphite electrode ). (ii) The lifetime
5
6
oxidase, from tryptophan in DNA photolyase and cytochrome
c peroxidase, and from glycine in pyruvate formate lyase and
of a protein-linked radical depends strongly on the protein
environment; it can be long-lived if it is buried in a hydrophobic
pocket and shielded from the external aqueous medium, e.g.
7
8
9
anaerobic RR. A transient cysteine radical has been postulated
for the catalytic turnover in RR.1 It is reasonable to assume
0,11
Tyr in RR, or very short-lived if it is exposed to the medium,
•
2
•
6
that transient histidine radicals may also occur as metal ligands
e.g. Trp in DNA photolyase. In the latter case, detection by
EPR would be difficult. (iii) Another difficulty in EPR studies
of protein radicals arises from the oftenly observed magnetic
interaction with paramagnetic metal centers, which in many
radical enzymes are close to the radical site. In such a case,
spin exchange leads to a strong increase of relaxation rates of
the radical and EPR can be detected only at low temperatures.
An example is Trp in cytochrome c peroxidase.7 These
difficulties, in particular the close proximity of a possible
histidine radical to a ligating paramagnetic metal, may probably
be the reason that so far protein-linked histidine radicals have
not been observed and analyzed by EPR. An example for a
strong magnetic interaction between a metal site and a radical
from an amino acid ligand is found in galactose oxidase. Here
11
in the electron or radical transfer pathways of radical enzymes.
More generally, transient histidine ligand radicals may play a
role during redox reactions of metalloproteins.
In Ca-depleted samples of the oxygen-evolving photosystem
II complex, radicals from an oxidized histidine have been
postulated from an optical absorption band at 310 nm (similar
1
2
•
to that observed from a histidine OH-adduct radical ) and a
broad unstructured EPR line which has been simulated as a
species with a weak spin-exchange between a radical and a metal
1
3,14
site.
ligands for the Mn cluster in photosystem II
hyperfine coupling constants of this postulated histidine radical
Although ESEEM studies have suggested histidines as
14,15
data on
*
To whom correspondence should be addressed.
Technical University of Berlin.
University of Uppsala.
1
2+
a tyrosine radical (S ) /2) is ligated to a paramagnetic Cu (S
†
1
‡
§
) /2) (distance: 2 Å) which, in this case, leads to an anti-
ferromagnetic spin coupling and an EPR silent ground state.5
University of Stockholm.
1
0.1021/jp9843454 CCC: $18.00 © 1999 American Chemical Society
Published on Web 02/13/1999

1284 J. Phys. Chem. A, Vol. 103, No. 9, 1999
Lassmann et al.
as a result of H-atom addition and deamination only, and there
are little data on species created by oxidation.
The aim of the present paper is to identify the radical formed
by oxidation of histidine in aqueous medium and determine the
isotropic hyperfine coupling constants and spin density distribu-
tion from highly resolved EPR spectra. A comparison with
density functional theory (DFT) calculations of isotropic hy-
perfine coupling constants and spin density at the ring atoms
for radicals of the histidine model 4-ethylimidazole with OH
addition at three different position is presented.
This is the first report on a detailed electronic structure,
including hyperfine couplings of â-protons, ring protons, and
nitrogens, of a histidine radical in aqueous solution determined
by EPR spectroscopy and confirmed by density functional
calculations.
Figure 1. Histidine derivatives as used for generation of radicals by
Fenton oxidation in aqueous solution at pH 2, with numbering scheme
of ring positions; A, histidine in H
O, deuterated at dissociable proton positions; C, R,â,â-deuterated
histidine (in H O).
2
O, fully protonated; B, histidine in
D
2
2
From histidine being not involved in a polypeptide chain,
EPR and ENDOR data on different types of histidine radicals
have been reported earlier. In single crystals of histidine
hydrochloride X-irradiated at low temperature, primary radicals
were observed upon decarboxylation and H-atom addition to
the carbonyl oxygen or by one-electron oxidation yielding a
histidine cation radical on the imidazole ring.20 The hyperfine
coupling tensors for two ring R-protons of the cation radical
2. Materials and Methods
3
+
2
.0
.1. Generation of Radicals. A Ti Fenton system at pH
2
6-28
2
was used in order to generate histidine radicals by
oxidation of histidine by OH radicals according to the following
reaction sequence:
3+
•
-
4+
Ti + H O f OH + OH + Ti
(
isotropic values: 1.06 and 1.21 mT) have been determined,
2
2
•
•
but no hyperfine data were given for â-protons and imidazole
nitrogens. A stable secondary histidine radical in histidine
hydrochloride crystals irradiated at room temperature was
formed by hydrogen addition to the imidazole ring resulting in
OH + His f HisOH
(1)
The two solutions which were mixed in the continuous-flow
system at pH 2.0 contained the following components. Reactant
I: histidine (10, 40, and 80 mM, respectively); H2O2 (500 mM);
pH 2.0; Reactant II: histidine (10, 40, and 80 mM, respectively);
TiCl3 (14 mM).
21
two large proton couplings (isotropic value: 4.7 and 5.0 mT).
A deamination radical and carboxyl anion radicals have also
been observed in X-irradiated single crystals of histidine
dihydrochloride.22
L-Histidine was obtained from Bachem; TiCl3 and H2O2 from
Fluka. D2SO4, NaOD, and D2O were bought from Sigma; DL-
R,â,â-d3-histidine was from CDN Isotopes, Quebec, Canada.
Generation of histidine radicals in D2O was performed using
D2SO4 and NaOD for pH adjustment. The given concentration
of H2O2 diluted in D2O contributes only 2.5% H in the assay.
2.2. Continuous-Flow EPR Setup. A Bruker ESP 300E EPR
spectrometer was used in combination with the dielectric mixing
resonator ER 4117 D-MVT. This accessory combines a two-
component mixer with a dielectric ring X-band resonator and
allows continuous-flow EPR studies with low consumption of
reactants. The EPR cell is a thick-walled quartz capillary of
0.4 mm i.d. surrounded by a sapphire ring. This setup has a
high sensitivity for a small volume (2 µL) of aqueous samples.
The cross-section of the EPR tube is 11 times less than the 5-mm
quartz flat cell and 16 times less than the 10-mm flat cell.
Considering the same linear flow rate, e.g. the same time
resolution, the volume flow rate in this mixing resonator which
determines the consumption of reactants is less by a factor 11
and 16, respectively, than in the corresponding conventional
EPR X-band mixing devices mentioned above. This advantage
allows the use of expensive isotopically substituted substances
(see Figure 3C) with acceptable consumption of reactants. The
reactants I and II in the two 30 mL syringes (polypropylene) to
be mixed are driven simultaneously by a syringe pump (Harvard
22, Kleinfeld Laboratories, Hannover, Germany) via connecting
tubes (0.6 mm i.d. PEEK) through the mixer and the EPR
sample tube. The speed of the stepping motor of the pump is
programmable and microchip-controlled.
EPR and ENDOR data of X-irradiated single crystals of
imidazole were published for a radical which is formed in a
reducing step upon H-addition to the C2 position (for C-
numbering of the imidazole ring, see Figure 1). Hyperfine (hf)
tensors for all R-protons were determined, but no data for
nitrogens were submitted. Experimentally derived spin densities
at the ring atoms of imidazole were compared with those
2
3
obtained by H u¨ ckel, McLachlan, and INDO calculations.
Radicals from imidazole in the liquid state formed upon
reaction at pH 9-10 with hydroxyl radicals by steady-state in-
situ radiolysis24 or by a Fenton reaction at pH 2 in a steady-
2
5
state flow-system have been studied by EPR. In both cases,
OH addition at the C5 position leads to the imidazole radical,
and isotropic hyperfine coupling constants for R-protons and
nitrogens were determined. The kinetics of formation and decay
of cation radicals from imidazole and histidine formed after
reaction with OH radicals under radiolysis conditions were
studied by optical spectroscopy monitoring the 310 nm absorp-
12
tion peak. For histidine radicals formed by oxidation, however,
a comparable set of hf data in the liquid state including â-protons
does not exist. Therefore, EPR studies on histidine radicals in
model systems are necessary in order to elucidate typical spectral
features of such radicals in order to recognize them more easily
in different proteins.
In electron transfer reactions during enzyme catalysis, histi-
dine radicals with the unpaired electron localized on the
imidazole ring are expected upon one-electron oxidation. In the
case of protein damage, however, upon oxidative stress, e.g.
by an attack on the protein by active oxygen species (hydroxyl
radicals, superoxide radicals, or singlet oxygen), or upon
radiation damage (UV light or X-irradiation), OH radicals may
attack the imidazole ring leading to histidine-OH adduct
radicals. Previously, formation of histidine radicals was reported
A volume flow rate of 34 mL/min (corresponding to a linear
flow rate of 4.4 m/s) was used for generation of the histidine
radicals. The dead-time between the mixer cell and the EPR-
active volume at the fastest flow used is 1 ms, but a sufficiently
high steady-state concentration of these short-lived histidine

Electronic Structure of a Histidine Radical
J. Phys. Chem. A, Vol. 103, No. 9, 1999 1285
radicals is achieved probably because of the intentional incom-
plete mixing efficiency of the Y-mixer so that mixing takes place
mainly within the EPR cell rather than in the mixer. With a
flow rate of 34 mL/min two accumulated 20 s scans consumed
about 150 mg of ââ-deuterated histidine (40 mM in both of the
two syringes) in order to acquire the EPR spectrum shown in
Figure 3C.
Fremy’s salt (potassium peroxylamine disulfonate) with g )
2
.0055 ( 0.0001 was used as an appropriate g-factor standard
for the flow resonator.
2
.3. Analysis of Spectra. The simulation of isotropic EPR
spectra was performed on a HP work station using a simulation
program for EPR described in ref 17. This program allows a
nonlinear least-squares fit routine to find the optimum set of
parameters for the best simulation of a given experimental EPR
2
9
Figure 2. EPR continuous-flow spectra of histidine after reaction with
spectrum.
OH radicals in a Ti3 /H
+
O
2
Fenton system at pH 2.0 with various
2
2.4. Density Functional Theory Calculations. Complemen-
concentrations of histidine. Experimental conditions: microwave power,
mW; modulation, 100 kHz/0.05 mT); sweep time, 20 s (A, C), 40 s
(B); time constant, 20 ms (A, C), 40 ms (B); 1 scan (A, B), 2 scans
2
(C); flow rate, 34 mL/min. A: no histidine (control); reactant I, [H O ]
tary to the experimental study, detailed theoretical analyses were
conducted on a number of hydroxylated imidazole radicals, as
described below. By comparing theoretical and experimental
hyperfine coupling constants, it is possible to select the type of
radical fitting optimally to the experimentally observed radical.
Theoretical studies of radical hyperfine properties, although
conducted at various levels during some 30 years, have recently
witnessed a tremendous breakthrough with the development and
implementation of so-called gradient-corrected density functional
3
2
5
[
[
00 mM; reactant II, [Ti3 ] 14 mM. B: reactant I, [H
+
O
2 2
] 500 mM,
3+
His] 10 mM; reactant II, [Ti ] 14 mM, [His] 10 mM. C: reactant I,
2 2
H O
] 500 mM, [His] 40 mM; reactant II, [Ti3 ] 14 mM, [His] 40
+
mM.
above). Without any substrate (histidine), the OH radicals
4+
30
generated in the Fenton system form transient Ti -peroxy
theory (DFT) or hybrid Hartree-Fock (HF)-DFT approaches.
complexes showing an unresolved single EPR line (line width
Due to a more favorable scaling with increasing size of the
system compared with conventional accurate ab initio based
methods, it is possible with the DFT-based approaches to
investigate systems of far larger size than previously possible,
while still retaining accuracy.
∆
H ) 0.12 mT) with g ) 2.0135, a g-factor typical for peroxide
3
9,40
radicals as observed earlier
(Figure 2A). At histidine
concentrations between 5 and 20 mM another single EPR line
∆H ) 0.17 mT) appears at g ) 2.0151 (Figure 2B), which is
(
absent without the substrate histidine (Figure 2A). The amplitude
of this line increases with increasing concentration of histidine
from 5 to 20 mM. Above 20 mM histidine it decreases and a
complex EPR spectrum is formed with an isotropic g-factor of
Through detailed investigations and comparisons between
various DFT functionals, two have been singled out as being
by far the most accurate for hyperfine properties, B3LYP and
31
PWP86. The former is based on a combination of HF exchange
and gradient-corrected DFT exchange and correlation (Slater
local and Becke’s 1988 gradient corrected exchange; Vosko-
2
.0023 (Figure 2C). Up to 80 mM histidine, the intensity of
the complex EPR spectrum increases with a concomitant
decrease of the intensity of the EPR line at g ) 2.0151. Since
both signals exist only in presence of histidine and depend on
its concentration, they are regarded as belonging to histidine-
derived radicals. Both are short-lived on a submillisecond time
scale because their intensities increase with faster flow rates
Wilk-Nusair local and Lee-Yang-Parr gradient corrected
correlation).32 Their relative contributions were determined from
a least-squares fit to atomization energies, ionization potentials,
and electron affinities of a large number of systems. PWP86,
on the other hand, is a “pure” DFT functional, using the gradient
(
up to 34 mL/min).
correction to the exchange by Perdew and Wang and to cor-
4+
relation by Perdew.33 In the present paper, all systems inves-
The Ti -peroxy complex competes for the OH radicals with
the substrate histidine and disappears with higher concentrations
tigated were optimized at the B3LYP/6-31G(d,p) level, using
the Gaussian 94 program.34 These were followed by single point
4
7
of histidine (Figure 2C). A hyperfine structure from Ti and
4
9
4+
3
5
Ti isotopes as seen earlier on an intense line of the Ti -
PWP86/6-311G(2d,p) calculations using the deMon code, in
order to evaluate the hyperfine coupling constants for the
systems. This particular combination of approaches has recently
3
9,40
peroxy complex
was not observed on the histidine-derived
line at g ) 2.0151. From the close similarity of the g-factor of
the histidine single line to those of the peroxytitanium complex
been employed in several studies of radicals of biophysical
4
1
interest36-38
and to known organic peroxy radicals as well as from the
narrow line width and the absence of hf splitting, we assume
that the single line at g ) 2.0151 belongs to a histidine peroxy
and proven to yield highly accurate results.
The systems investigated are cation radicals from 4-ethyl-
imidazole at low pH (both nitrogens protonated) hydroxylated
at three different positions as models of the experimentally
investigated hydroxylated histidine cation radicals at pH 2. For
all systems we report atomic spin densities and isotropic hyper-
fine couplings; we have also investigated effects of hydrogen
bonding.
•
radical (his-O-O ) with spin density localized mainly at the
4
2
two oxygens. This is in accordance with the observation that
the histidine peroxy radical dominates at higher concentrations
of H2O2.
At higher concentrations of histidine, however, the preferred
reaction of OH is the formation of the histidine radical which
exhibits the complex EPR spectrum (Figure 2C). The large
number of lines suggests that the unpaired spin is localized
mainly in the imidazole ring of histidine and interacts with
several nuclei (Figure 1A). This histidine radical is the subject
of a detailed analysis of its hyperfine structure, as outlined
3. Results and Discussion
3
.1. Identification of Radicals upon Fenton Oxidation of
3
+
Histidine. The histidine radicals have been generated in a Ti
Fenton reaction at pH 2.0 using a flow system (as described

1286 J. Phys. Chem. A, Vol. 103, No. 9, 1999
Lassmann et al.
about 3 orders of magnitude smaller than for the narrow single
line of the proposed histidine peroxy radical at g ) 2.0151.
The steady-state concentration of the transient histidine radical
depends on its lifetime, the geometry of the mixer and the EPR
cell, the flow rate, and on the concentrations of the reactants.
Since the amplitude of the EPR signal is maximal at fastest
flow rate and collapses immediately after stopping the flow,
the lifetime of the radical is probably shorter than 1 ms.
In order to obtain EPR spectra of sufficient intensity and
resolution, it was necessary to perform a detailed optimization
of the concentration of the reactants. Optimal conditions for
sufficient concentration of histidine radicals are found for
3
+
concentrations [H2O2] . [S] > [Ti ] with the used Fenton
system which is in agreement with those reported for other
2
5-27
organic compounds
too (S: substrate to be oxidized in the
Fenton system, here histidine). The results of this procedure,
e.g. the optimal concentrations of all reactants, are given in the
legend of Figure 3.
3.2. Structure of a Histidine OH-Adduct (Cation) Radical
in Liquid Aqueous Solution. 3.2.1. Hyperfine Structure
Analysis of the Isotropic EPR Spectra. For unraveling the
complex EPR hyperfine spectrum of histidine radicals, two
selectively deuterated histidines have been used in addition to
the normal histidine: (i) All dissociable NH and OH protons
were replaced by dissolving histidine in D2O (Figure 1B). (ii)
Histidine with deuterated â-protons (dissolved in H2O) was used
(
Figure 1C). The EPR spectra at pH 2 from Fenton-oxidized
histidine in H2O, histidine in D2O, and ââ-deuterated histidine
in H2O) are shown in Figure 3A-C, respectively. The single
(
EPR line at g ) 2.0151 (from a histidine peroxy radical; see
also Figure 2B,C) has been subtracted in all three spectra shown
in Figure 3. Although the relative concentration of the species
from the single EPR signal is less than 2% of the concentration
of the multiline pattern, this approach was necessary for the
fitting process in the spectral simulations.
Considering the large splittings, the spectral analysis shows
five spectral groups with an integral intensity of about 1:2:2:
2
:1 for both histidine in H2O as well as in D2O (Figure 3A,B)
which corresponds to one large proton splitting of about 3.0
mT and two nearly equivalent proton splittings of about 1.5
mT. All these protons are nondissociable. Histidine radicals with
deuterated â-protons show only the large proton splitting and
lack the other two (Figure 3C). This indicates that the two, nearly
equivalent proton splittings in the unmodified histidine sample
belong to the two â-protons of the histidine radical.
The smallest hyperfine couplings are best visible in the wings
of the spectrum for histidine in D2O (Figure 3B) and correspond
to two nitrogen nuclei (I ) 1) with different splittings (0.14
and 0.27 mT). For histidine in H2O an additional small doublet
splitting (0.18-0.20 mT) from an exchangeable proton (I )
1
Figure 3. EPR spectra of transient histidine radicals formed by oxi-
dation in a Ti /H O Fenton system at pH 2.0, recorded by continuous
2 2
flow. The single line occurring at 341.5 mT (see Figure 2B) has been
subtracted prior to the simulation and fit procedure. (Dotted line: best
simulation using the hfc’s given in Table 1.) Experimental conditions:
microwave power, 3 mW; modulation, 100 kHz/0.05 mT; time constant,
/2) is additionally involved. Accordingly, this proton-splitting
3+
is missing in the total width of the D2O-histidine spectrum
(
(
7.16 mT) as compared to that of histidine in H2O (7.33 mT)
see Table 1). It turned out in the spectral simulations that a
further proton splitting of 0.58 mT was required in addition.
The most complicated spectrum, of â-deuterated histidine in
H2O, was fitted using the data set of histidine in H2O and by
replacing the two â-proton coupling by the correspondingly
smaller (resolved) deuteron splittings (see Figure 3C and Table
2
0 ms. A: histidine in H
2
O; reactant I, [H
reactant II, [Ti ] 14 mM, [His] 80 mM; flow rate, 38 mL/min; sweep
time, 40 s; 5 scans. B: histidine in D O; reactant I, [H ] 500 mM,
His] 80 mM; reactant II, [Ti ] 14 mM, [His] 80 mM; flow rate, 34
mL/min; sweep time, 20 s; 3 scans. C: R,â,â-d -histidine in H O;
reactant I, [H ] 500 mM, [His] 40 mM; reactant II, [Ti ] 14 mM,
His] 40 mM; flow rate, 34 mL/min; sweep time, 20 s; 2 scans.
2 2
O ] 500 mM, [His] 80 mM;
3
+
2
2 2
O
3
+
[
3
2
1
). The smaller total width of the spectrum (5.48 mT) for the
ââ-deuterated histidine radical corresponds to the larger value
7.33 mT) from the ââ-protonated radical and is nicely
3
+
2
O
2
[
(
below. Since this EPR spectrum extends over 7.4 mT, the
amplitude of the first derivative is, at the same concentration,
reproduced by the simulation which further supports the given
assignment.

Electronic Structure of a Histidine Radical
J. Phys. Chem. A, Vol. 103, No. 9, 1999 1287
TABLE 1: Isotropic Hyperfine Coupling Constants (mT) of
the Histidine-OH Adduct Cation Radical Derived from
EPR Experiments and Used for Spectra Simulation (Error:
(0.02 mT)
molecular
posn
His in H
2
Oa
His in D
2
O
R,â,â-deuterated
a
(fully protonated) (NH-deuterated) His (in H
2
O)
a(H5)
3.02
1.41
1.23
0.58
0.21
0.14
0.27
3.02
1.41
1.23
0.59
<0.03
0.14
3.02b
0.22
0.19b
0.59
0.21
0.14
0.28
a(Hâ1)
a(Hâ2)
a(H2)
a(HNH or HOH
a(N1)
)
a(N3)
0.27
c
line width
tot. width
0.08
7.33
0.09
7.16
0.06
5.48
d
a
Corresponding formulas in Figure 1. b Deuteron coupling constants.
c
d
Width of individual lines used for simulation. Between turning points
of outermost EPR lines in spectra of Figure 3.
Figure 4. Electronic structure (hyperfine coupling constants and π-spin
densities) determined experimentally by EPR spectroscopy for the
histidine-OH adduct cation radical at pH 2.0. Isotropic hyperfine
coupling constants (in brackets and italic letters) are given next to the
corresponding nuclei; the respective π-spin densities are given in bold
letters.
The complete set of hyperfine coupling constants obtained
from the best least-squares fit of the spectra (dotted lines in
Figure 3) are listed in Table 1 for the three different histidine
radicals. The hfc’s of two nitrogens and five protons in the case
of histidine in H2O (Figure 1A) are in excellent agreement with
those of the corresponding radical of histidine in D2O (two N’s,
four H’s; Figure 1B) and those of â-deuterated histidine in H2O
(
two N’s, two D’s, three H’s; Figure 1C).
The position of the exchangeable proton with a splitting of
0.21 mT of the histidine cation radical is still open. It may
belong to one of the two ring NH protons or to the OH proton
at the OH adduct. DFT data for the 4-ethylimidazole cation
radical with C5-OH addition favor the OH proton (0.22 mT)
or the N1-H proton (0.29 mT) but not the N3-H proton (0.005
mT) (Figure 5D). It is not yet clear why the second exchangeable
proton, which according to the calculations should have a
resolvable splitting, is not visible in the EPR spectra.
Both nitrogen hyperfine coupling constants (0.14 and 0.27
mT) have been resolved in the EPR spectrum of the OH-adduct
histidine cation radical (see Table 1). The assignment of the
smaller value to N1 and the larger to N3 in the histidine cation
radical is based on DFT calculations (see section 3.2.3). Figure
4 summarizes the assignment of the experimentally observed
hfc’s to the individual nuclei of the histidine radical (in italic
letters). These hfc’s are directly correlated with the distribution
of the spin density in the radical. This experimentally established
electronic structure of a histidine cation radical forms the basis
for a comparisons with calculated hyperfine data, in order to
test advanced quantum chemical methods.
3
.2.2. Assignment of Hyperfine Splittings to Protons and
Nitrogens of the Histidine Radical. The two isotropic splittings
of 1.41 and 1.23 mT in the EPR spectrum of normal histidine
in H2O, which reduce to 0.22 and 0.19 mT in case of
â-deuterated histidine, belong without any doubt to the two
â-protons. A large isotropic proton splitting of 2.64 mT was
previously observed by EPR in the liquid state after OH attack
on imidazole upon in-situ pulse radiolysis at pH 9-10. In the
case of oxidation of imidazole using the same Ti /H2O2 Fenton
system at pH 2.0 as employed here, and a similar EPR
continuous-flow device in Q-band, a large proton splitting of
•
2
4
3+
3
.03 mTsvery close to our value (3.02 mT)shas been
2
5
observed. In both cases, this large splitting has been assigned
to a ring proton (bent out-of-ring plane) linked to the tetragonal
C5 ring carbon atom in the imidazole radical formed after
addition of an OH radical to that carbon which originally was
trigonal. Thus, the R-proton on C5 is in a â-position with respect
to C4 where the maximum spin density resides (see below).
Large isotropic hfc’s up to 4.0 mT have also been reported from
ring protons after H-atom addition, in X-irradiated histidine
single crystals.2 Thus, we interpret the large proton splitting
in the present sample as originating from a CH proton of the
imidazole ring after addition of an OH radical to the C5 (or
C2) carbon (for the position of the OH addition, see also section
1,23
Spin Densities DeriVed from Isotropic EPR Hyperfine Cou-
pling Constants. For R-protons of an aromatic ring the π-spin
density FC at the connecting C-atom can be estimated from the
R
isotropic proton coupling constant a H using the McConnell
R
43,44
3
.2.3). Taking this into account, the type of observed histidine
relation a H ) FCQ, with Q ) 2.48 mT.
Following this
radical upon Fenton oxidation is an OH-adduct radical. Ac-
cording to the pK ) 6.5 of histidine and pK ) 5.5 of the
corresponding OH-adduct radical,12 the protonation of both
imidazole nitrogens of histidine at pH 2.0 suggests that after
addition of a neutral OH radical from the Fenton system to the
positively charged histidine we observe a histidine cation radi-
cal.
empirical relation a π-spin density at C2 of F(C2) ) 0.24 is
derived from the C2 R-proton. The isotropic â-proton coupling
â1,2
constant a H is related via a hyperconjugation mechanism to
the π-spin density FC at the neighboring C-atom according to
â1,2
2
43
a
H ) FCQ′ cos θ1,2, with Q′ ) 5.0 mT and the dihedral
â
angles θ1,2 between the projection of the C-H bond and the
direction of the pz-orbital perpendicular to the ring plane.
Assuming rapid rotational fluctuations (waggling) of the me-
thylene group of the histidine radical in liquid solution at room
temperature about the C4-Câ bond, the θ angles of the two
The proton splitting of 0.58 mT which appears in all three
histidine EPR spectra (Table 1) belongs to the R-proton on the
•
C2 (or C5) ring carbon which is not attacked by OH . In the
2
similar imidazole cation radical at pH 2.0 a corresponding
R-proton splitting of 0.66 mT was observed,25 which is close
to our value. In the neutral imidazole radical at pH 9-10, a
â-protons are averaged out to a certain extent and cos θ1,2
1
â1
â2
1
attains the mean value /2. Thus, a H ≈ a H ) FC4Q′ /2, with
the mean value of the two â-hfc’s (1.32 mT) yielding F(C4) )
0.53. The small but significant inequivalence of a H and a
H
2
4
â1
â2
larger value (0.99 mT) was found for the C2 R-proton.

1288 J. Phys. Chem. A, Vol. 103, No. 9, 1999
Lassmann et al.
can be explained by the asymmetry near the methylene group.
Particularly, the chiral carbon center adjacent to the methylene
group occurring in all amino acids has been shown to be
responsible for the inequivalence of the two methylene â-protons
4
5
in case of a tyrosyl radical in liquid aqueous systems. This
situation seems to be quite similar also in case of the histidine
radical as observed here.
The π-spin density at the nitrogens can be estimated using
N
π
N
N
43
the corresponding formula a ) F NQ with Q ) 2.5 mT
π
π
yielding to F N1 ) 0.06 and F N2 ) 0.11.
3
.2.3. Density Functional Calculations of Histidine Model
Radicals: Comparison of Hyperfine Data DeriVed from Density
Functional Theory and from EPR Experiments. In previous DFT
4
6
studies of biological radicals, we have investigated tyrosine,
42
47,48
37
tryptophan, glycine,
quinones, and the DNA base cyto-
sine. In these studies we have found that an ethyl group is a
sufficient and very suitable model for the amino acid or quinone
3
8
“
tail” to the active side group. This is primarily related to the
fact that with the ethyl group we are able to obtain a correct
description of the two â-protons rather than with a methyl group.
As a model for the histidine radical at low pH, studied
experimentally in this work, we chose, for the DFT calculations,
the 4-ethylimidazole cation radical in which both nitrogens are
protonated. Of the various systems, histidine/imidazole is unique
in the respect that there are significant data for the nitrogen
couplings.
From EPR data, particularly from the occurrence of a large
proton coupling of 3.02 mT, an addition of OH radicals to C2,
C4, or C5 of the imidazole ring is suggested. The isotropic hfc’s,
spin densities, and relative energies of these three radicals of
the histidine model 4-ethylimidazole have been computed at
the PWP86/6-311G(2d,p)//B3LYP/6-31G(d,p) and B3LYP/6-
311+G(2df,p)//B3LYP/6-31G(d,p) levels, respectively. The
three possible OH adducts were optimized and the hyperfine
properties evaluated. The resulting isotropic hfc’s and spin
densities are shown in Figure 5A-C; they are given in the figure
with italic and bold letters, respectively. The energetically lowest
lying structure is that where OH attack has occurred on C2,
whereas the C5 and C4 adducts lie 10.3 and 14.4 kcal/mol above
the one at C2, respectively.
Since advanced DFT approaches, as used here, are able to
yield hyperfine coupling constants directly, we compare the
isotropic hyperfine coupling constants (hfc’s) measured by EPR
for the histidine cation radical (see Figure 4) with the hfc’s of
the protonated 4-ethylimidazole cation radicals calculated by
DFT (Figure 5). This procedure is preferred rather than a
comparison of theoretical and experimentally determined spin
densities, because the Q and Q′ factors in the semiempirical
formulas for the experimentally estimated π-spin densities,
derived from measured isotropic hfc’s of R- and â-protons,
respectively, are only approximate values and not uniformly
valid for different radicals. This is supported also from a
comparison of DFT calculations of spin densities with those
Figure 5. Electronic structure (hyperfine coupling constants and spin
densities) calculated by DFT of different protonated OH-adduct cation
radicals of 4-ethylimidazole, OH additions at the positions C2 (A), C4
(B), and C5 (C) and OH addition at position C5 including a hydrogen-
bonded formic acid molecule (D) are shown. Isotropic hyperfine
coupling constants (in brackets and italic letters) are given next to the
corresponding nuclei; total spin densities are in bold letters. Key: ( ),
static values for energy-minimized side chain geometry; Key: [ ],
averaged values for side chain motion.
data. The calculated nitrogen hfc’s in this radical exhibit the
same ratio of 1:2 (and also similar hfc’s values) as found by
EPR; this is not the case for C2 addition. Furthermore, in the
C5-adduct radical the calculated hfc’s for the dissociable protons
fit better to the experimental one, whereas in case of C2 addition
they are larger than the experimental value. Taken together,
theoretical data support clearly that the formation of the histidine
cation radical at pH 2.0 takes place by addition of OH radicals
to the aromatic imidazole ring at the C5 carbon. An oxidation
of a ligand histidine has been observed in Cu,Zn superoxide
dismutase by HPLC technique upon treatment with H2O2.49
From comparison with 2-oxohistidine the position of attacking
the imidazole ring was reported at C2; a comparison with
5-oxohistidine, however, was not included in this study.
4
6
from experimentally available hfs data for tyrosine radicals.
Imidazole Ring. From the calculated isotropic hfc’s for three
•
different species with addition of OH to C2, C4, or C5, the
C4-OH adduct can be ruled out, because of the absence of
any large proton coupling constant of about 3.0 mT in the
calculation and also because of the near zero spin density at
C4. The latter leads to very small â-proton couplings which is
in contradiction to the experimental result. The calculated hfc’s
for the radicals with OH at C2 or C5 both show a large proton
hfc (2.67 and 2.71 mT, respectively), but the remaining hfc’s
of the C5 addition radical fit better to the experimental EPR

Electronic Structure of a Histidine Radical
J. Phys. Chem. A, Vol. 103, No. 9, 1999 1289
There is a good, although not perfect, match with the
experimental hfc’s of the ring protons and nitrogens for the C5
adduct (Figure 5C), except the C2 R-proton for which the theory
predicts a larger value (-0.92 mT) than found in the experiment
A clear indication to rapid rotational motion of the side group
is the fact that none of the calculated â-proton hfc’s for the
energy-minimized static conformation are in agreement with
the experimentally observed values, irrespective of rotational
angle. The calculated motional aVerage values for the two
â-protons (1.74 and 1.83 mT), however, are not equivalent and
are in much closer agreement with the experimental data set.
Thus, the computed average data of Hâ in the hydrogen-bonded
(-0.59 mT).
Since in aqueous systems the nitrogens of the histidine ring
are able to form hydrogen bonds which may influence the spin
densities and the hfc’s, DFT calculations have been extended
to a hydrogen-bonded molecule. In a first attempt to improve
the theoretical model, a hydrogen-bonding moiety (formic acid,
HCOOH) was added to the C5 adduct, such that the two formic
acid oxygens would form hydrogen bonds to the N1 and C2
protons. The system was again fully optimized, and the
hyperfine couplings were evaluated. The results shown in Figure
4
-ethyl-5-hydroxyimidazole radical cation, 1.74 and 1.83 mT
(see values in brackets in Figure 5D), are regarded as a suitable
fit to the experimentally observed couplings. The remaining
discrepancy may be due to the smaller model employed; we
did not taken into account the full effects of the bulkier amino
and carboxylic acid groups of histidine, and all calculations have
been performed in vacuum; i.e., we did not include the solvent
effects and hydrogen bonds between amino and carboxylic
groups of neighboring molecules.
5D demonstrate that the calculated hfc’s, particularly those for
the two nitrogens, the C2 R-proton, and the C5 proton are
significantly closer to the experimental EPR values than those
without the formic acid molecule. The nitrogen couplings at
N1 and N3 are slightly reduced and are now in better agreement
with the experiment. The H(O) and H1 couplings are probably
too close to each other to be separable, and each of them may
be responsible for the observed 0.21 mT coupling. The value
for H2 is decreased to -0.79 mT (experiment: -0.59 mT),
and the H5 coupling increases to 2.82 mT (experiment: 3.02
mT); in both cases these are closer to the experimental values.
4
. Conclusions
Upon oxidation of histidine in a Ti /H2O2 Fenton system, a
3
+
histidine-OH adduct cation radical is formed by addition of
OH radicals to the C5 carbon of the imidazole ring. The isotropic
hyperfine coupling constants of two â-protons, three ring
protons, and two nitrogen nuclei of this 5-oxohistidine cation
radical have been determined by EPR. This is the first report
on a detailed electronic structuresincluding hyperfine structure
of â-protons, protons, and nitrogens of the imidazole ringsof
a transient histidine radical in liquid aqueous solution determined
by EPR spectroscopy, which is complemented by density
functional theory (DFT) calculations. DFT calculations for all
hyperfine coupled nuclei and spin densities at ring atoms for
three cation radicals of the histidine model 4-ethylimidazole with
different positions of OH addition were performed and support
the assignment of experimental hfs splittings. Excellent agree-
ment of experimental and theoretical hfs data was achieved for
the nuclei of the imidazole ring, particularly when a H-bond to
the ring-N1 is included in the calculation of the hfc’s (ap-
proximating the aqueous environment of the radical).
Considering that an oxidation of the imidazole ring of
49
histidine was reported at the C2 position, a formic acid residue
H-bonded to the NH proton and the C2-OH proton for the
C2-addition radical (see Figure 5A) was also included in the
calculation. The nitrogen hfcc’s aN1 ) 0.290 mT and aN3 )
0
)
.374 mT (without formic acid) and aN1 ) 0.358 mT and aN3
0.451 mT (with formic acid) are even further from the
experimental values (0.14 and 0.27 mT) in the case of a
H-bonding molecule to the radical with 2-OH addition. The hfc’s
for NH protons (0.69 and 0.54 mT) in the C2 addition radical
remain almost unchanged with H-bonding formic acid and are
much larger than the experimental value (0.21 mT). Thus, only
the C5 position for histidine oxidation is in accordance with
the experimental data. Taking into account the fact that the DFT
calculations were performed for a molecule “in vacuum”,
whereas the experiments were done in aqueous phase at pH 2,
the agreement is quite good. The better agreement with the
experimental values in the case of additional modelling of the
water environment by a H-bond shows that the highly sophis-
ticated DFT approaches are able to describe the actual experi-
mental situation quite well.
The question may arise about which are the typical spectral
features of histidine radicals. This is important for recognizing
them easier in the protein-bound state. As shown in detail for
2
,11
protein-linked tyrosine radicals
tryptophan radicals,
and also for protein-linked
17,18
the conformation of â-methylene protons
of protein radicals depends primarily on structural constraints
in the immediate protein environment of the radical. This will
probably also be valid for protein-linked histidine radicals. A
special feature of histidine radicals is the occurrence of two
different nitrogen hfc’s. In powderlike EPR spectra of protein-
linked histidine radicals (in the frozen state and in the im-
mobilized liquid state), however, small and strongly anisotropic
As indicated above, the presence of a hydrogen-bonding
moiety to the N1/C2 fragment of the molecule, as employed in
this work, may well be the reason for the favored C5 radical
attack. On the basis of the improved agreement between
experimental and computed data, we conclude that at least one
hydrogen-bonding group is indeed present in the system.
nitrogen hfc’s as observed in different protein-linked tryptophan
17,18
radical
are expected which will usually be resolved only
by ENDOR or pulsed EPR (ESEEM) techniques.
â-Protons. In order to investigate the dependence of the
â-proton hfc’s on the orientation relative to the ring plane, the
ethyl tail of the C5-OH adduct was rotated about the C4-Câ
bond in steps of 10°, at each step reoptimizing the remaining
geometric parameters of the ethyl tail. From the resulting energy
curve of the variation in the isotropic hfc’s of â-protons as a
function of the CR-Câ-C4-N3 rotational angle (not shown),
we note that the potential energy for rotation is relatively low,
only about 2.5 kcal/mol. In solution at room temperature one
may expect that the system undergoes rapid waggling fluctua-
tions about the C4-Câ bond such that the observed â-proton
couplings of 1.23 and 1.41 mT are motionally averaged values.
The OH-adduct radical of oxidized histidine, however, which
exhibits an unusual large proton hf splitting of about 3.0 mT
may be recognized easily as an EPR spectroscopic fingerprint.
Since a ring proton on a tetragonal-hybridized carbon exhibits
properties of â-protons including their small hfs anisotropy,23
this splitting would certainly be resolved in powderlike EPR
spectra of immobilized histidine radicals in proteins.
OH-adduct radicals of histidine as studied here may occur
upon oxidative stress when reactive oxygen species are formed
in case of a cellular malfunction by reaction of H2O2 with free
ferrous iron or under conditions of UV photolysis or radiolysis
of proteins, in general.

1290 J. Phys. Chem. A, Vol. 103, No. 9, 1999
Lassmann et al.
Histidine radicals which may appear transiently during
(21) Box, H. C.; Freund, H. G.; Lilga, K. T. J. Chem. Phys. 1967, 46,
2
130.
22) Westhof, E.; Flossmann, W.; M u¨ ller, A. Int. J. Radiat. Biol. 1975,
7, 51.
(23) Lamotte, B.; Gloux, P. J. Chem. Phys. 1973, 59, 3365.
electron transport processes in radical enzymes and metallo-
proteins will probably keep the imidazole ring intact. To model
such histidine radicals in solution, the application of milder
oxidants than OH radicals is necessary. Another option to model
histidine radicals may be UV irradiation of frozen solution of
histidine models, but in this case the EPR spectra are much
less resolved50 and require more sophisticated techniques like
ENDOR or pulsed EPR (ESEEM) for hfs analysis.
(
2
(
(
(
24) Samuni, A.; Neta, P. J. Phys. Chem. 1973, 77, 1629.
25) Schmidt, J.; Borg, D. Radiat. Res. 1976, 65, 220.
26) Paul, H.; Fischer, H. Ber. Bunsen-Ges. Phys. Chem. 1969, 73, 972.
(27) Czapski, G. J. Phys. Chem. 1971, 75, 2957.
(28) Czapski, G.; A. Samuni, A.; Meisel, D. J. Phys. Chem. 1971, 75,
3
271.
29) Fiege, R. Dissertation, Technical University of Berlin, 1997.
30) Malkin, V. G.; Malkina, O. L.; Eriksson, L. A.; Salahub, D. R. In
(
(
Theoretical and Computational Chemistry, Vol 2, Modern Density Func-
tional Theory-A Tool for Chemistry; Politzcr, P., Seminario, J. M., Eds.;
Elsevier: New York, 1995; pp 273-347.
Acknowledgment. The work was supported by the Deutsche
Forschungsgemcinschaft (DFG) Project La 751/2-1, the Fonds
der Chemischen Industrie to G.L. and W.L., and the Swedish
Natural Science Research Council (NFR). Grants of computer
time at the Center for Parallel Computing (PDC) in Stockholm
are gratefully acknowledged. We thank Prof. Astrid Gr a¨ slund,
Department of Biophysics, Stockholm University, for valuable
discussion.
(
31) Barone, V. In Recent AdVances in Density Functional Theory, Part
1
; Chong, D. P., Ed.; World Scientific: Singapore, 1995; pp 287-334.
Engels, B.; Eriksson, L. A.; Lunell, S. AdV. Quantum Chem. 1996, 27, 297.
32) Slater, J. C. Quantum Theory of Molecules and Solids; McGraw-
(
Hill: New York, 1974. Becke, A. D. Phys. ReV. 1988, A38, 3098. Vosko,
S. H.; Wilk, L.; Nusiar, M. Can. J. Phys. 1980, 58, 1200. Lee, C.; Yang,
W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. Becke, A. D. J. Chem. Phys.
1993, 98, 1372.
(33) Perdew, J. P.; Wang, Y. Phys. ReV. B 1986, 33, 8800. Perdew, J.
P. Phys. ReV. B 1986, 33, 8822. Perdew, J. P. Phys. ReV. 1986, 34, 7406.
(34) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T. A.; Peterson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Pang, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; DeFrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94; Gaussian Inc.:
Pittsburgh, PA, 1995.
(35) St-Amant, A.; Salahub, D. R. Chem. Phys. Lett. 1990, 169, 387.
St-Amant, A. Ph.D. Thesis, Univ. de Montreal, 1991. Salahub, D. R.;
Fournier, R.; Mlynarski, P.; Papai, I.; St-Amant, A.; Ushio, J. In Density
Functional Methods in Chemistry; Labanowski, J., Andzelm, J., Eds.;
Springer-Verlag: New York, 1993; pp 77-100.
(36) Eriksson, L. A. Mol. Phys. 1997, 91, 827.
(37) Eriksson, L. A.; Himo, F.; Siegbahn, P. E. M.; Babcock, G. T. J.
Phys. Chem. A 1997, 101, 9496.
References and Notes
(1) Sj o¨ berg, B. M.; Reichard, P.; Gr a¨ slund, A.; Ehrenberg, A. J. Biol.
Chem. 1978, 253, 6863.
(
2) Gr a¨ slund, A.; Sahlin, M. Annu. ReV. Biophys. Biomol. Struct. 1996,
5, 259.
3) Dietz, R.; Nastaincyk, W.; Ruf, N. H. Eur. J. Biochem. 1988, 171,
2
(
3
21.
(
(
(
4) Berry, B. A. Photochem. Photobiol. 1993, 57, 179.
5) Whittaker, M. M.; Whittaker, J. W. J. Biol. Chem. 1990, 265, 9610.
6) Kim, S.; Sancar, A.; Essenmacher, C.; Babcock, G. T. Proc. Natl.
Acad. Sci. U.S.A. 1993, 90, 8023.
7) Huyett, J. E.; Doan, P. E.; Gurbiel, R.; Houseman, A. L. P.; Sivaraja,
M.; Goodin, D. B.; Hoffmann, B. M. J. Am. Chem. Soc. 1995, 117, 9033.
8) Wagner, A. F. V.; Frey, M.; Neugebauer, F. A.; Sch a¨ fer, W.;
Knappe, J. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 996.
9) Mulliez, E.; Fontecave, M.; Gaillard, J.; Reichard, P. J. Biol. Chem.
993, 268, 2296.
10) Mao, S. S.; Yu, G. X.; Chalfoun, D.; Stubbe, J. Biochemistry 1992,
1, 9752.
11) Sj o¨ berg, B. M. Nucleic Acids and Molecular Biology; Eckstein,
F., Lilley, D. M. J., Eds.; Springer: Berlin, Heidelberg, 1995; Vol. 9, pp
(
(
(
1
(38) Wetmore, S. D.; Himo, F.; Boyd, R. J.; Eriksson, L. A. J. Phys.
Chem. B 1998, 102, 7484.
(39) Fischer, H. Ber. Bunsen-Ges. Phys. Chem. 1967, 71, 685.
(40) Norman, R. O. C.; West, P. R. J. Chem. Soc. B (Phys. Org.) 1969,
389.
(
3
(
1
92-221.
(41) Hellwege, K. H. Landolt-B o¨ rnstein: Numerical Data and Func-
tional Relationships in Science and Technology, Group II, Vol. 9, Magnetic
Properties of Free Radicals; Springer-Verlag: New York, 1979; Part c2,
pp 8-25.
(
12) Rao, P. S.; Simic, M.; Hayon, E. J. Phys. Chem. 1975, 79, 1260.
13) Boussac, A.; Zimmermann, J. L.; Rutherford, A. W.; Lavergne, J.
(
Nature 1990, 347, 303.
14) Zimmermann, J. L.; Boussac, A.; Rutherford, A. W. Biochemistry
993, 32, 4831.
15) Tang, X. S.; Diner, B. A.; Larsen, B. S.; Gilchrist, M.L., Jr.; Lorigan,
G. A.; Britt, R. D. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 704.
16) Tang, X. S.; Randall, D. W.; Force, D. E.; Diner, B. A.; Britt, R.
D. J. Am. Chem. Soc. 1996, 118, 7638.
17) Lendzian, F.; Sahlin, M.; MacMillan, F.; Bittl, R.; Fiege, R; P o¨ tsch,
(
(42) Himo, F.; Eriksson, L. A. J. Phys. Chem. B 1997, 101, 9811.
(43) Wertz, J. G.; Bolton, J. R. Electron Spin Resonance; Chapman and
Hall: New York, 1986.
1
(
(44) Bender, C. J.; Sahlin, M.; Babcock, G. T.; Barry, B. A.; Chan-
drashekar, T. K.; Salowe, S. P.; Stubbe, J.; Lindstr o¨ m, B.; Peterson, L.;
Ehrenberg, A.; Sj o¨ berg, B. M. J. Am. Chem. Soc. 1989, 111, 8076.
(45) Sealy, R. C.; Harman, L.; West, P. R.; Mason, R. P. J. Am. Chem.
Soc. 1985, 107, 3401.
(46) Himo, F.; Gr a¨ slund, A.; Eriksson, L. A. Biophys. J. 1997, 72, 1556.
(47) Himo, F.; Eriksson, L. A. J. Chem. Soc., Perkin Trans. 2, 1997,
305.
(
(
S.; Sj o¨ berg, B.-M.; Gr a¨ slund, A.; Lubitz, W.; Lassmann, G. J. Am. Chem.
Soc. 1996, 118, 8111.
(18) P o¨ tsch, S.; Lendzian, F.; Ingemarson, R; H o¨ rnberg, A.; Thelander,
L.; Lubitz, W.; Lassmann, G.; Gr a¨ slund, A. J. Biol. Chem., submitted for
publication.
(48) Choi, C. H.; Kertzesz, M.; Karpfen, A. Chem. Phys. Lett. 1997,
276, 266.
(19) Brabek, V.; Mornstein, V. Biophys. Chem. 1980, 12, 159.
(
20) Ngo, F. Q.; Budzinski, E. E.; Box, H. C. J. Chem. Phys. 1974, 60,
(49) Uchida, K.; Kawakishi, S. J. Biol. Chem. 1994, 269, 2405.
(50) Berthomieu, C.; Boussac, A. Biospectroscopy 1995, 1, 187.
3
373.