Modeling the active site of cytochrome oxidase: Synthesis and characterization of a cross-linked histidine-phenol

Article abstract of DOI:10.1021/ja011852h

A cross-linked histidine-phenol compound was synthesized as a chemical analogue of the active site of cytochrome c oxidase. The structure of the cross-linked compound (compound 1) was verified by IR, ¹H and ¹³C NMR, mass spectrometry

Full text of DOI:10.1021/ja011852h

Published on Web 01/31/2002
Modeling the Active Site of Cytochrome Oxidase: Synthesis
and Characterization of a Cross-Linked Histidine-Phenol
Jenny A. Cappuccio,† Idelisa Ayala,‡ Gregory I. Elliott,† Istvan Szundi,†
James Lewis,† Joseph P. Konopelski,† Bridgette A. Barry,‡ and OÄ lo¨f Einarsdo´ttir*,†
Contribution from the Department of Chemistry and Biochemistry, UniVersity of California
Santa Cruz, California 95064, and Department of Biochemistry, Molecular Biology, and
Biophysics, UniVersity of Minnesota, 1479 Gortner AVenue, St. Paul, Minnesota 55108
Received July 30, 2001
Abstract: A cross-linked histidine-phenol compound was synthesized as a chemical analogue of the active
site of cytochrome c oxidase. The structure of the cross-linked compound (compound 1) was verified by
1
IR, H and 13C NMR, mass spectrometry, and single-crystal X-ray analysis. Spectrophotometric titrations
indicated that the pKa of the phenolic proton on compound 1 (8.34) was lower than the pKa of tyrosine
(10.1) or of p-cresol (10.2). This decrease in pKa is consistent with the hypothesis that a cross-linked
histidine-tyrosine may facilitate proton delivery to the binuclear site in cytochrome c oxidase. Time-resolved
optical absorption spectra of compound 1 at room temperature, generated by excitation at 266 nm in the
presence and absence of dioxygen, indicated a species with absorption maxima at ∼330 and ∼500 nm,
which we assign to the phenoxyl radical of compound 1. The electron paramagnetic resonance (EPR)
spectra of compound 1, obtained after UV photolysis, confirmed the generation of a paramagnetic species
at low temperature. Because the cross-linked compound lacks â-methylene protons, the EPR line shape
was dramatically altered when compared to that of the tyrosyl radical. However, simulation of the EPR line
shape and measurement of the isotropic g value was consistent with a small coupling to the imidazole
nitrogen and with little spin density perturbation in the phenoxyl ring. The ground-state Fourier transform
infrared (FT-IR) spectrum of compound 1 showed that addition of the imidazole ring perturbs the frequency
of the tyrosine ring stretching vibrations. The difference FT-IR spectrum, associated with the oxidation of
the cross-linked compound, detected significant perturbations of the phenoxyl radical vibrational bands.
We postulate that phenol oxidation produces a small delocalization of spin density onto the imidazole nitrogen
of compound 1, which may explain its unique optical spectral properties.
Introduction
residue (Tyr244) and the ꢀ-nitrogen of histidine 240, which is
a ligand to CuB.
Cytochrome c oxidase, the terminal protein in the electron
transport chain, catalyzes the sequential four-electron transfer
from cytochrome c to dioxygen.1 The reduction of dioxygen to
water is coupled to the translocation of four protons across the
inner mitochondrial membrane (or the bacterial cytoplasmic
membrane).2 The resulting electrochemical proton gradient is
used by ATP synthase to generate ATP. The O2 reduction takes
place at the heme a3/CuB binuclear site. Recent X-ray crystal-
lographic studies on the bovine heart3 and Paracoccus denitri-
ficans4 cytochrome oxidases have shown that the active site
contains an unprecedented cross-link between C6 of a tyrosine
The details of the mechanism of the reduction of dioxygen
to water are still unresolved. A “peroxy” intermediate (P) has
been proposed following the binding of O2 to heme a3. While
previous studies proposed that the O-O bond in P was intact,
more recent studies showed that the O-O bond is already
broken and that P is in fact an oxyferryl state, Fea34+dO2- 5-8
.
The discovery of the tyrosine-histidine cross-link at the active
site led to the proposal that the tyrosine may function as an
electron9 and a proton donor to the dioxygen bound to heme
a3,7,10,11 thus facilitating the cleavage of the dioxygen bond. One
of the protons taken up by cytochrome oxidase upon reduction
(5) Proshlyakov, D. A.; Ogura, T.; Shinzawa-Itoh, K.; Yoshikawa, S.;
Appelman, E. H.; Kitagawa, T. J. Biol. Chem. 1994, 269, 29385-29388.
(6) Proshlyakov, D. A.; Ogura, T.; Shinzawa-Itoh, K.; Yoshikawa, S.; Kitagawa,
T. Biochemistry 1996, 35, 76-82.
* To whom correspondence should be addressed. E-mail: olof@
chemistry.ucsc.edu. Fax: 831-459-2935.
† University of California Santa Cruz.
‡ University of Minnesota.
(7) Proshlyakov, D. A.; Pressler, M. A.; Babcock, G. T. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 8020-8025.
(1) Ferguson-Miller, S.; Babcock, G. T. Chem. ReV. 1996, 96, 2889-2907.
(2) Wikstro¨m, M. K. F. Nature 1977, 266, 271-273.
(3) Yoshikawa, S.; Shinzawa-Itoh, K.; Nakashima, R.; Yaone, R.; Yamashita,
E.; Inoue, N.; Yao, M.; Fei, M. J.; Libeu, C. P.; Mitzushima, T.; Yamaguchi,
H.; Tomizaki, T.; Tsukihara, T. Science 1998, 280, 1723-1731.
(4) Ostermeier, C.; Harrenga, A.; Ermler, U.; Michel, H. Proc. Natl. Acad.
Sci. U.S.A. 1997, 94, 10547-10553.
(8) Fabian, M.; Wong, W. W.; Gennis, R. B.; Palmer, G. Proc. Natl. Acad.
Sci. U.S.A. 1999, 96, 13114-13117.
(9) Blomberg, M. R. A.; Siegbahn, P. E. M.; Babcock, G. T.; Wikstro¨m, M.
J. Inorg. Biochem. 2000, 80, 261-269.
(10) Gennis, R. B. Biochim. Biophys. Acta 1998, 1365, 241-248.
(11) Sucheta, A.; Szundi, I.; Einarsdo´ttir, OÄ . Biochemistry 1998, 37, 17905-17914.
9
1750 VOL. 124, NO. 8, 2002 J. AM. CHEM. SOC.
10.1021/ja011852h CCC: $22.00 © 2002 American Chemical Society

Modeling the Active Site of Cytochrome Oxidase
A R T I C L E S
triacetate (3) (400 mg, 0.77 mmol). The solution was allowed to stir at
room temperature for 12 h, and the reaction was subsequently quenched
with an aqueous solution of sodium sulfide. The black precipitate was
filtered through Celite. The layers were separated and the aqueous layer
was extracted with CH2Cl2. The organic layer was dried with MgSO4
and purified by flash chromatography to provide (4) as a clear oil in
35% yield. [R]D25 +6.8 (c 0.95, MeOH); 1H NMR (500 MHz, CDCl3)
δ 7.706 (d, J ) 1 Hz, 1H), 7.328 (dt, J ) 6.5 Hz, 1H), 7.247 (d, J )
7.5 and 2 Hz, 2H), 7.041-6.993 (m, 2H), 6.001-5.924 (m, 1H), 5.338
(dq, J ) 17.5 and 1.5 Hz, 1H), 5.269 (dq, J ) 10 and 1 Hz, 1H), 4.868
(dt, J ) 7.5 and 5 Hz, 1H), 4.567-4.524 (m, 2H), 3.693 (s, 3H), 3.196
(dq, J ) 57.5, 14.5, and 5.5 Hz, 2H), 2.047 (s, 3H); 13C NMR (125
Hz, CDCl3) δ 172.1, 170.1, 151.5, 137.6, 137.1, 132.4, 128.9, 126.7,
125.5, 121.5, 118.0, 114.0, 69.7, 52.5, 52.3, 29.7, 23.3; IR (neat) 3246,
1721, 1669, 1245, 1003 cm-1. HRMS Calcd for [M]+ C18H21N3O4:
343.1532. Found: 343.1526.
of the binuclear center has been proposed to protonate the
tyrosine, thereby preparing it for its role in the cleavage of the
O-O bond.10
While recent experiments have provided support for the role
of a tyrosyl radical in the P form of the enzyme,7,12-14 direct
evidence for the participation of the tyrosyl radical in the
mechanism is lacking. The expected EPR signals from the S )
1/2 CuB and S ) /2 Tyr244• have not been detected in the P
form of the bovine heart enzyme. This has been attributed to a
possible spin coupling between the two species that is mediated
by the ligated cross-linked His240.7
2+
1
It is clear that further studies are needed to establish whether
the cross-linked tyrosine plays a key role in the mechanism of
the reduction of dioxygen to water by cytochrome oxidase. In
particular, it is important to understand whether the cross-link
between the histidine and tyrosine changes the properties of
the tyrosine to facilitate its role in dioxygen reduction. Recent
studies on 2-imidazol-1-yl-4-methylphenol have suggested that
the cross-link will facilitate proton delivery to the binuclear
center in the enzyme.15 It is also of interest to determine whether
the proposed radical resides on the tyrosine or the histidine. In
this work, we present the synthesis and spectroscopic studies
of a His-phenol cross-linked compound (compound 1), a
chemical analogue of the active site of cytochrome oxidase. Our
spectrophotometric titrations show that the pKa of the phenolic
proton on the cross-linked compound is lowered compared to
that of tyrosine or p-cresol. This is consistent with the cross-
linked tyrosine facilitating proton delivery to the binuclear site
in cytochrome c oxidase.7,10,15 In addition, the presence of a
UV-generated radical in compound 1 is investigated at room
temperature by time-resolved UV-vis spectroscopy and con-
firmed at low temperature by electron paramagnetic resonance
(EPR) and Fourier transform infrared (FT-IR) difference
spectroscopy. The data are consistent with the radical residing
primarily on the phenoxyl ring with a small delocalization of
spin density onto the imidazole.
(B) 1-o-Phenol(acetyl)histidine Methyl Ester (1). Tributyltin
hydride (60 µL, 0.226 mmol) was slowly added to a solution of 4 (66
mg, 0.19 mmol) and bis(triphenylphosphine) palladium chloride (3 mg,
4.7 µmol) in 10 mL of dry tetrahydrofuran (THF). The solution darkens
upon the final addition of tin hydride. The reaction mixture was allowed
to stir at room temperature for 2 h and then concentrated in a vacuum
to give an amber oil. Purification by flash chromatography provides
the title compound (1) in 99% yield as a white solid. Material suitable
for single-crystal X-ray analysis was obtained by recrystallization from
25
1
MeOH: mp 201-203 °C; [R]D +32.9 (c 1.35, MeOH); H NMR
(500 MHz, CDCl3) δ 7.739 (d, J ) 1.5 Hz, 1H), 7.239 (dt, J ) 8.5 and
1.5 Hz, 1H), 7.190 (dd, J ) 13 and 1.5 Hz., 1H), 7.099-7.052 (m,
3H), 6.931 (dt, J ) 8 and 1.5 Hz, 1H), 4.833 (dq, J ) 6 and 1.5 Hz,
1H), 3.674 (s, 3H), 3.098 (m, 2H), 1.965 (s, 3H); 1H NMR (500 MHz,
DMSO-d6) δ 10.17 (br s, 1H), 8.238 (d, J ) 7.5 Hz, 1H), 7.824 (s,
1H), 7.284 (dd, J ) 7.5 and 1.5 Hz, 1H), 7.185 (m, 2H), 7.033 (dd, J
) 8.5 and 1 Hz, 1H), 6.902 (dt, J ) 7 and 1 Hz, 1H), 4.519 (q, J )
8.5 and 1 Hz, 1H), 3.601 (s, 3H), 2.942 (dq, J ) 14 and 8 Hz, 2H),
1.822 (s, 3H); 13C NMR (125 Hz, CDCl3) δ 172.0, 170.9, 151.0, 137.4,
136.5, 129.5, 125.5, 124.9, 120.1, 118.4, 117.8, 52.8, 52.5, 29.9, 23.1;
IR (KBr) 3417, 1728, 1653, 1289, 1242 cm-1. HRMS Calcd for [M]+
C15H17N3O4: 303.1219. Found: 303.1215.
UV-Visible Spectra. Ground-state UV-visible spectra of tyrosine,
3-aminotyrosine, and compound 1 in 0.1 M sodium bicarbonate/0.06
M tert-butyl alcohol buffer, pH 10, were recorded at room temperature
on a Hewlett-Packard (8452) diode-array spectrophotometer in a 0.2-
cm path length quartz cuvette. Time-resolved absorbance difference
spectra (post- minus pre-photolysis) of tyrosine and compound 1 were
recorded at room temperature in the same deoxygenated buffer by an
optical multichannel analyzer.11 The spectra were collected in a 0.2 ×
0.05 cm quartz cuvette at 5-8 delay times between 30 ns and 10 ms
following excitation at 266 nm (Nd:YAG, 7 ns pulse, 300 µJ/mm2).
The spectral changes were probed along the 0.2-cm path length at 90°
to the laser photolyzing beam. Prior to recording of the time-resolved
spectra, the samples were saturated with nitrous oxide, which removes
the spectral contribution of a solvated electron according to: N2O +
e-(aq) f OH• + N2 + OH-.16 Because the hydroxyl radical may initiate
unwanted reactions and thus interfere with the transient spectra, tert-
butyl alcohol was added to quench the hydroxyl radical. The resulting
â-hydroxyl radical produced from the alcohol is unlikely to interfere
with the transient spectra.16 Each difference spectrum represents an
average of 16 and 20 individual spectra for the tyrosine and the cross-
linked His-phenol complex, respectively. Baseline fluorescence was
subtracted automatically. The time-resolved difference spectra were
analyzed by singular value decomposition (SVD) and global exponential
fitting with Matlab software (Mathworks), as previously described.11,17,18
Materials and Methods
Reagents. 3-Aminotyrosine, p-cresol, and L-tyrosine were purchased
from Sigma-Aldrich (St. Louis, MO). The H-His-Tyr-OH dipeptide
was purchased from Bachem (King of Prussia, PA). 13C(6)-Tyrosine
([ring-13C(6)]-L-tyrosine) was purchased from Cambridge Isotope
Laboratories (Andover, MA) and was 98-99% labeled.
Synthesis of the His-Phenol Cross-Linked Compound: (A) 1-o-
Allyoxyphenyl(acetyl)histidine Methyl Ester (4). N-R-Acetylhistidine
methyl ester (2) (136 mg, 0.55 mmol) was stirred in 5 mL of dry CH2-
Cl2. To this suspension was added Cs2CO3 (195 mg, 0.6 mmol) along
with Cu(OAc)2 (12 mg, 0.077 mmol) and o-(prop-2-enyloxy)phenyllead
(12) Chen, Y. R.; Sturgeon, B. E.; Gunther, M. R.; Mason, R. P. J. Biol. Chem.
1999, 274, 3308-3314.
(13) MacMillan, F.; Kannt, A.; Behr, J.; Prisner, T.; Michel, H. Biochemistry
1999, 38, 9179-9184.
(16) Bent, D. V.; Hayon, E. J. Am. Chem. Soc. 1975, 97, 2599-2606.
(14) Proshlyakov, D. A.; Pressler, M. A.; DeMaso, C.; Leykam, J. F.; DeWitt,
D. L.; Babcock, G. T. Science 2000, 290, 1588-1591.
(15) McCauley, K. M.; Vrtis, J.; Dupont, J.; Van der Donk, W. A. J. Am. Chem.
Soc. 2000, 122, 2403-2404.
(17) Sucheta, A.; Georgiadis, K. E.; Einarsdo´ttir, OÄ . Biochemistry 1997, 36,
554-565.
(18) Van Eps, N.; Szundi, I.; Einarsdo´ttir, OÄ . Biochemistry 2000, 39, 14576-
14582.
9
J. AM. CHEM. SOC. VOL. 124, NO. 8, 2002 1751

A R T I C L E S
Cappuccio et al.
The difference spectra of the intermediates were extracted on the basis
of a unidirectional sequential scheme. The time-resolved spectra of
compound 1 were also recorded in the presence of dioxygen in 10 mM
borate-NaOH, pH 11.
A0, is the sum of the concentrations of all the components with different
pKa values present in the system:
A0 ) A1 + A2 + A3 + ... + Ak-1 + Ak
Spectrophotometric Titrations. Spectrophotometric titrations of
p-cresol (pH ) 3.6-11.2), tyrosine (pH ) 8.08-11.82), and compound
1 (pH ) 2.5-10.63) were recorded on a diode-array spectrophotometer
(HP 8452). The samples were prepared in aqueous solutions of 0.1 M
potassium chloride, and the spectra were corrected for dilution by the
base titrant.
Since each concentration can be calculated from the previous one
A0 ) A1 + (K1/H)A1 + (K1K2/H2)A1 + ... +
(K1K2...Kk-2Kk-1/Hk-1)A1
or in a more compact form
The pKa values of tyrosine, p-cresol, and compound 1 were
determined from the experimental absorption spectra recorded over a
wide pH range by SVD and global pK fitting. The SVD significantly
reduces the number of parameters required for the global fitting.19 Below
is a brief description of the technique used to analyze the titration of
compounds with multiple dissociation steps.
k-1
A0 ) (1 +
Pi/Hi)A1
∑
i)1
where
The absorption data matrix of a mixture of k components can be
written as a product of two matrixes:
i
Pi )
Kj
∏
j)1
A(λ, pH) ) E(λ)C(pH)
The expressions for the concentrations of the individual components
are as follows:
where E(λ) is an m × k matrix whose columns are the extinction
coefficients of the components measured at m wavelength (λ) values,
and C(pH) is a k × n matrix whose rows are the concentrations of the
components at n pH values. For simplicity, the arguments will be
omitted from subsequent formulas. The term component used in the
analysis does not necessarily imply that a single chemical entity is
present. The spectrum of a given component may represent the
superposition of spectra from two or more related structures, such as
a neutral and a zwitterionic structure having the same protonation state.
In the SVD analysis, three data matrixes, U, S, and V, are generated
such that A ) USV′ where U is an m × n matrix and V′, the transpose
of V, is an n × n matrix. The columns of U are the orthogonal spectra
and the columns of V contain the pH dependence of these spectra. The
column vectors in the U and V matrixes are arranged in the order of
their significance, which is reflected by the corresponding singular
values in the S diagonal matrix. Only a limited number of the most
significant U and V vectors are needed to reproduce the data matrix
within the experimental noise. The scaled, significant orthogonal vectors
can be represented by the product of the U and S matrixes, US.
The spectra of the components, E, which are usually not orthogonal,
can be represented by a linear combination, represented by a matrix
L, of the significant, scaled orthogonal vectors: E ) USL. Accordingly,
the data matrix can be written as A ) USLC, where the pH dependence
is given by V′ ) LC. As shown below, the entries in C are calculated
by use of a function that describes the dissociation process, and the
pKa values are obtained from the best fit to this function. The
calculations become greatly simplified compared to fitting the data
matrix directly, because only a limited number of the significant V
vectors are involved in the fitting.
k-1
A1 ) A0/(1 +
Pi/Hi)
∑
i)1
k-1
A2 ) A0/(P1/H)/(1 +
Pi/Hi)
∑
i)1
and finally
k-1
Ak ) A0(Pk-1/Hk-1)/(1 +
Pi/Hi)
∑
i)1
In practice, we use pK and pH as variables, and Pi/Hi is replaced by
the power function of 10, where the power is (ipH - Σij)1pKj).
Each row of the C matrix contains the concentrations of the
corresponding component calculated at n pH values. If two or more
compounds are present with multiple dissociation steps, then the
concentration matrix is repeated for each compound. For compounds
with a single dissociation step, these formulas reduce to the Henderson-
Hasselbalch equation. The best least-squares fit of the LC matrix to
the V′ matrix provides a set of pKa values and an L matrix, which is
used to calculate the spectra of the components, E ) USL, as already
discussed.
EPR Spectroscopy. EPR spectra were acquired on a Bruker
(Billerica, MA) EMX spectrometer equipped with a TE cavity and a
Wilmad (Buena, NJ) temperature control dewar. The sample temper-
ature was maintained at 77 K. UV photolysis flashes at 266 nm were
supplied by a frequency-quadrupled pulsed YAG laser (Continuum,
Santa Clara, CA). The sample was illuminated with 5 (for tyrosine,
13C(6) tyrosine, and H-His-Tyr-OH dipeptide) or 10 (compound 1)
flashes in the cavity at a frequency of 10 Hz; the energy of the laser
pulses was 35-37 mJ. Samples were 100 mM and contained 10 mM
borate-NaOH, pH 11. To mimic the conditions for the FT-IR data
acquisition, samples were concentrated in the EPR tube with a stream
of cold nitrogen before data acquisition. Spectral conditions were as
follows: frequency, 9.21 GHz; microwave power, 0.2 mW; modulation,
100 kHz; modulation amplitude, 2 G; time constant, 1.3 s; data
acquisition time, 168 s; number of scans, 4. Spectra were obtained from
two different samples. EPR spectra were simulated with a program
previously described.20 Spin quantitation was performed by double
integration through the use of Igor Pro software (Lake Oswego, OR).
The function describing the concentration pH dependence of the
components reflects the chemical nature of the compounds under study.
For a mixture of independent acids and bases with a single dissociation
step, the simple Henderson-Hasselbalch formula is used.19 For
compounds with multiple dissociation steps, the formula is more
complicated. The derivation below applies to the so-called macroscopic
dissociation constants, which reflect the overall protonation changes
in the system. We assume that an acid has k forms with different pKa
values. The proton dissociation of the ith state is described by the
dissociation constant, Ki:
Ki ) (Ai+1H)/Ai
where H and A denote the concentrations of the protons and acid in its
different dissociation forms, respectively. The total acid concentration,
(19) Shrager, R. I.; Hendler, R. W. Anal. Chem. 1982, 54, 1147-1152.
(20) Hoganson, C. W.; Babcock, G. T. Biochemistry 1992, 31, 11874-11880.
9
1752 J. AM. CHEM. SOC. VOL. 124, NO. 8, 2002

Modeling the Active Site of Cytochrome Oxidase
A R T I C L E S
Figure 1. X-ray crystal structure of the cross-linked model histidine-
phenol compound, 1-o-phenol(acetyl)histidine methyl ester (compound 1).
FT-IR Spectroscopy. FT-IR ground-state spectra and difference FT-
IR spectra associated with the photolysis reaction were obtained through
the use of a Nicolet (Madison, WI) 60SXR spectrometer, equipped
with a liquid nitrogen-cooled MCT-B detector and a Hansen (Santa
Barbara, CA) cryostat. The experimental conditions were the same as
those described for the EPR measurements. The sample holder was
equipped with two CaF2 windows. Samples were concentrated on a
window with a stream of cold nitrogen before data acquisition.
Difference spectra were constructed from data obtained before and after
photolysis. Ground-state spectra were obtained before photolysis.
Spectral conditions were as follows: resolution, 4 cm-1; apodization
function, Happ-Genzel; zero-filling, 1 level; data acquisition time, 1
min. Before the beginning of FT-IR data acquisition, the FT-IR sample
was mounted in a Hitachi (Danbury, CT) 3000 UV-vis spectropho-
tometer and the absorption spectrum was measured. The absorption at
294 nm was used to correct the FT-IR spectrum for concentration and
path length. Alternatively, the FT-IR spectrum was corrected for the
amplitude of the C-C stretching vibration (1500 cm-1) or the C-O
stretching vibration (∼1260 cm-1) in the ground-state absorption
spectrum. Similar results were obtained with all three correction
methods. Difference spectra were obtained from 2-8 different samples
and averaged.
Figure 2. UV-visible spectra of tyrosine (3.24 mM) (-‚-), the histidine-
tyrosine dipeptide (6.65 mM) (s), 3-aminotyrosine (1.79 mM) (- - -),
and compound 1 (1.78 mM) (-‚‚-). The buffer was 0.1 M NaHCO3/0.06
M tert-butyl alcohol, pH 10.
Results
Structure of the His-Phenol Complex. Complete spectro-
scopic data (IR, 1H and 13C NMR, and mass spectrometry) were
obtained in order to verify the structure of the His-phenol
compound. The structure was confirmed by single-crystal X-ray
analysis, the results of which are shown in Figure 1. The dihedral
angle (-36.3°) between the phenol and imidazole functionalities
in this structure is similar in magnitude to that observed in the
native enzyme (44°).3
Ground-State Absorption Spectra. Figure 2 shows the
ground-state UV-visible spectra of tyrosine, the histidine-
tyrosine dipeptide, 3-aminotyrosine, and compound 1 in 0.1 M
NaHCO3/0.06 M tert-butyl alcohol buffer, pH 10. Because pH
10 is below the pKa of tyrosine (10.1), there are two maxima at
280 and 298 nm, corresponding to the protonated and depro-
tonated forms of tyrosine, respectively. The same maxima are
observed for the histidine-tyrosine dipeptide. On the other hand,
a single maximum is observed in the spectra of both compounds
with amine substitution ortho to the phenolic hydroxyl group,
3-aminotyrosine and compound 1, indicating that both are in
the fully deprotonated anionic form at pH 10.
Spectrophotometric Titrations. To verify that the pKa of
the phenolic proton on tyrosine is lowered as a result of the
cross-link, spectrophotometric titrations were carried out on
p-cresol (pH ) 3.6-11.2) (Figure 3A), tyrosine (pH ) 8.08-
11.82; not shown), and compound 1 (pH ) 2.5-10.63) (Figure
4A). SVD and global pK fitting provided the spectra of the
Figure 3. (Panel A) Spectrophotometric titration of an aqueous solution
of p-cresol (0.5 mM) from pH 3.6 to 11.2. The data were analyzed by SVD
and global pKa fitting, which provided the spectral forms of the different
protonation states of the compound (panel B: solid line, protonated; dashed
line, deprotonated). The residuals show the difference between the fit and
the data for a single pKa of 10.05 (panel C).
intermediate forms and their pKa values. For p-cresol, a
single pKa of 10.05 was obtained, while two pKa values,
8.65 and 10.10, were observed for tyrosine. The two pKa
values are attributed to the terminal amine and the phenolic
side chain, respectively, and are in good agreement with
previously reported values,15,16,21 thus validating our analysis
method. Analysis of the titration spectra of compound
(21) CRC Handbook of Chemistry and Physics, 58th ed.; CRC Press: Boca
Raton, FL, 1978.
9
J. AM. CHEM. SOC. VOL. 124, NO. 8, 2002 1753

A R T I C L E S
Cappuccio et al.
ns, 400 ns, 1 µs, 10 µs, 100 µs, 1 ms, and 10 ms. The difference
spectra show a unique absorption band at ∼500 nm. Another
peak was observed at ∼330 nm. The peak at 500 nm clearly
decays more rapidly than the 330 nm band, indicating that at
least two species contribute to the spectra. The same spectra
were observed in the presence of dioxygen (not shown).
The time-resolved difference spectra of the tyrosine (Figure
5A) were satisfactorily fitted with a single exponential (77 µs)
by SVD and global exponential fitting. The spectrum of the
intermediate is shown in Figure 5B and is consistent with the
formation of a tyrosyl radical.16,22,23 Time-resolved difference
spectra of the histidine-tyrosine dipeptide also revealed a single
peak at ∼400 nm (not shown).
To fit the transient data of compound 1, two exponentials
with apparent lifetimes of 1.4 and 9.8 µs were required,
supporting the presence of two species, the spectra of which
are shown in Figure 5D. The first species, with a lifetime of
∼1 µs, was found to have absorption maxima at ∼325 and ∼500
nm. The second species, which decayed with a lifetime of ∼10
µs, displayed a peak at ∼330 nm and a shoulder at ∼380 nm.
EPR Spectra. EPR spectroscopy and UV photolysis at 77
K have been shown to generate Tyr• radicals in basic aqueous
solutions.24 In Figure 6, the EPR spectra of the radicals produced
by photolysis in solutions of tyrosine (A), 13C(6) tyrosine (B),
the His-Tyr dipeptide (C), and compound 1 (D) are shown as
solid lines. The EPR line shape of the radical produced in frozen
tyrosinate solutions (Figure 6A) was similar to results previously
obtained.24-26 As expected, there was no detectable signal
produced from the borate buffer alone (Figure 6A, dotted line).
Spectral broadening upon 13C labeling of the ring (Figure 6B)
was consistent with the assigned distribution of unpaired spin
density.24,25 While the EPR line shape derived from the His-
Tyr dipeptide (Figure 6C) resembled spectra derived from
tyrosinate (Figure 6A, solid line), a significant line shape
perturbation was observed in compound 1 (Figure 6D, solid
line).
The g values of the radicals were 2.0050 ( 0.0001 for
tyrosine, 2.0054 ( 0.0001 for 13C(6) tyrosine, 2.0046 ( 0.0001
for the His-Tyr dipeptide, and 2.0052 ( 0.0001 for compound
1. Given the standard deviation, the g values of radicals
produced in tyrosine and in compound 1 are indistinguishable.
Quantitation of the amount of signal produced in tyrosine and
in compound 1 gave results that were also indistinguishable,
given the standard deviation (∼(10%). These results suggest
that each radical is deprotonated and that there is little change
in spin density on the phenolic oxygen when tyrosine and
compound 1 are compared.27
Figure 4. (Panel A) Spectrophotometric titration of an aqueous solution
of the His-phenol cross-linked compound (compound 1) (1.0 mM) from
pH 2.5 to 10.63. The data were analyzed by SVD and global pKa fitting,
which provided the spectral forms of the different protonation states of
compound 1 (panel B: s, fully protonated; - - -, imidazole deprotonated;
-‚‚-, fully deprotonated). The residuals show the difference between the
fit and the data for pKa values of 5.54 and 8.34 (panel C).
1 resulted in two pKa values of 5.54 and 8.34, which are assigned
to the imidazole group and phenol, respectively. The two pKa
values represent a unique fit and converged on the same values
regardless of the initial starting values. Figures 3B and 4B show
the spectra of the different protonation states of the p-cresol
and compound 1, respectively. The accuracy of the fit is
reflected in the residuals (Figures 3C and 4C), which represent
the difference between the reproduced data and experimental
data.
Time-Resolved UV-Vis Difference Spectra. The radical
forms of tyrosine and compound 1 were generated in vitro by
UV photolysis. UV photolysis should generate a tyrosyl radical
in compound 1, because the oxidation potential is expected to
be only 66 mV more positive, when compared to tyrosine.15
Figure 5, panels A and C, show the UV-visible (350-675 nm)
time-resolved difference spectra (post- minus pre-photolysis)
of tyrosine and compound 1, respectively, in a deoxygenated
buffer solution following excitation at 266 nm at room tem-
perature. The tyrosine time-resolved difference spectra were
recorded at five delay times: 1 µs, 10 µs, 30 µs, 100 µs, and
1 ms. The difference spectra, with a maximum at ∼400 nm
(Figure 5A), are consistent with the formation of a tyrosyl
radical.16,22,23 The time-resolved difference spectra of compound
1 (Figure 5C) were recorded at eight delay times: 30 ns, 100
To test this idea, the EPR signals arising from tyrosine (Figure
6A, solid line) and compound 1 (Figure 6D, solid line) were
simulated, and the simulations are shown superimposed in
Figure 6A,D as dashed lines. The parameters (Table 1) used
for the tyrosyl radical simulation (Figure 6A, dashed line) were
similar to those previously employed.25 The EPR line shape
derived from the cross-linked compound could be simulated
(Figure 6D, dashed line) by removal of the couplings to the
(24) Barry, B.; El-Deep, M. K.; Sandusky, P. O.; Babcock, G. T. J. Biol. Chem.
1990, 265, 20139-20143.
(25) Hulsebosch, R. J.; Van der Brink, J. S.; Niewenhuis, A. M.; Gast, P.; Raap,
J.; Lugtenburg, J.; Hoff, A. J. J. Am. Chem. Soc. 1997, 119, 8685-8694.
(26) Warncke, K.; Perry, M. S. Biochim. Biophys. Acta 2001, 1545, 1-5.
(27) Dixon, W. T.; Murphy, D. J. Chem. Soc., Faraday. Trans. 2 1976, 72,
1221-1229.
(22) Land, E. J.; Porter, G. Trans. Faraday Soc. 1963, 59, 2016-2026.
(23) Tripathi, G. N. R.; Schuler, R. H. J. Chem. Phys. 1984, 81, 113-121.
9
1754 J. AM. CHEM. SOC. VOL. 124, NO. 8, 2002

Modeling the Active Site of Cytochrome Oxidase
A R T I C L E S
Figure 5. (A, C) Time-resolved difference spectra (post- minus pre-photolysis) of aqueous solutions of (A) tyrosine (3.24 mM) and (C) the His-phenol
cross-linked compound (1.78 mM) following excitation at 266 nm (Nd:YAG, fourth harmonic 7 ns pulse) at room temperature. The laser power was 300
mJ/mm2. The probe beam was a xenon flash lamp, and the signals were detected by an optical multichannel analyzer with a 100 µm slit and 200 ns gate.
The buffer was 0.1 M NaHCO3/0.06 M tert-butyl alcohol (pH 10), and the samples were purged with N2O just prior to use (to remove absorption contribution
from the solvated electron). (B, D) Intermediate spectra of tyrosine (B) and the His-phenol complex (D) resulting from the SVD/global exponential fitting
and on the basis of a unidirectional sequential mechanism. A single apparent lifetime of 77 µs was obtained for tyrosine. The apparent lifetimes for compound
1 were 1.4 and 9.8 µs.
tyrosine â-protons and addition of a relatively isotropic, large
hyperfine coupling to the hydrogen at ring position 4 (Table
1). An isotropic hyperfine tensor is not expected for an R-proton
but was required to fit the EPR line shape with this parameter
set. A small hyperfine coupling for the imidazole nitrogen was
included, but the exact value between 0 and 1.5 G did not have
a dramatic effect on the simulation result. A small nitrogen
coupling is consistent with the unperturbed g value of this
radical, compared to tyrosine, and with a relatively unperturbed
phenoxyl spin density distribution. The results suggest that the
EPR spectrum of the tyrosyl radical is comparatively insensitive
to o-imidazole addition.
FT-IR Ground-State Spectra. Another optical property of
compound 1 was compared to tyrosine. Figure 7 shows the
ground-state FT-IR spectra of tyrosine (A), 13C(6) tyrosine (B),
the His-Tyr dipeptide (C), and compound 1 (D). Spectral
contributions from the borate buffer blank were subtracted.
Residual water contributions are presumably due to solvent
interaction with the tyrosine-related compounds and were
observed at ∼2000 (combination band, not shown) and ∼1650
cm-1 (Figure 7, O-H bend). When the compound contains a
terminal amino group, a contribution from the N-H bending
mode is also expected in the 1650 cm-1 region.28
at ∼1260 cm-1, and a C-C-H bending mode (υ9a), expected
at ∼1170 cm-1 29,30
.
In the tyrosinate spectrum (Figure 7A), a
band was observed at 1266 cm-1, which shifted upon 13C
labeling (Figure 7B). The frequency and the magnitude of the
13C shift support the assignment of this band to υ7a′ of the
tyrosinate species.29,30 A spectral feature with similar frequency
and intensity was observed in the His-Tyr dipeptide. In
compound 1, a band at 1265 cm-1 was much less intense, and
a band at 1241 cm-1 was also observed (Figure 7D). The fact
that two bands are observed may be attributed to a unique
intermolecular hydrogen-bonding interaction in the ground state
of compound 1.29,30
The 1173-1172 cm-1 band in tyrosine (Figure 7A) and the
His-Tyr dipeptide (Figure 7C) is assigned to υ9a, a C-H bending
mode.29,30 As expected, this band was shifted in compound 1
(Figure 7D), where one of the o-hydrogens has been removed
and replaced with an imidazole nitrogen.
(B) Tyrosinate Ring Stretching Vibrations. Bands at 1602-
1601 and 1501-1500 cm-1 were observed in both tyrosine
(Figure 7A) and the His-Tyr dipeptide (Figure 7C). These
spectral features downshifted to 1562 (∆ ) -40 cm-1) or 1549
(∆ ) -53 cm-1) and 1464 (∆ ) -36 cm-1) cm-1, respectively,
upon 13C(6) tyrosine labeling (Figure 7B). The observed
frequencies and isotope shifts support the assignment of these
bands to ring stretching modes, υ8a and υ19a, of the tyrosinate
species.23,29-32 In compound 1, bands at 1591 and 1509 cm-1
(A) Tyrosinate C-O Stretching Vibration. Several tyro-
sinate vibrational modes are anticipated in the 1300-1100 cm-1
region,29,30 including a C-O stretching mode (υ7a′), expected
(28) Bellamy, L. J. The infrared spectra of complex molecules; Chapman and
Hall: London, 1980; Vol. I.
(29) Takeuchi, H.; Watanabe, N.; Satoh, Y.; Harada, I. J. Raman Spectrosc.
1989, 20, 233-237.
(30) Mukherjee, A.; McGlashen, M. L.; Spiro, T. G. J. Phys. Chem. 1995, 99,
4912-4917.
(31) Chipman, D. B.; Lie, R.; Zhou, X.; Pulay, P. Chem. Phys. Lett. 1997, 272,
155-161.
9
J. AM. CHEM. SOC. VOL. 124, NO. 8, 2002 1755

A R T I C L E S
Cappuccio et al.
Two additional ring and C-C/C-H vibrational modes are
29,30
expected in phenolate-containing compounds, υ8b and υ19b
.
A 1560 cm-1 band was observed in the tyrosinate spectrum
(Figure 7A), which may exhibit a 13C(6) isotope shift of 11
cm-1 (Figure 7B). This spectral feature may correspond to υ8b.
Note that the asymmetric stretching vibration of the free
carboxylate group is also expected to contribute to the spectrum
at approximately 1560 cm-1 28
Spectra of tyrosine-containing
.
dipeptides (data not shown) consistently exhibited a band at
approximately 1580 cm-1 (Figure 7C) but did not show the 1560
cm-1 line. Therefore, the 1580 cm-1 band is a possible candidate
for the υ8b vibrational mode, which may have a shifted frequency
in dipeptides. Accordingly, the spectral component at 1584 cm-1
can be attributed to υ8b in compound 1 (Figure 7D). Assignment
of υ19b awaits a detailed normal coordinate analysis, but possible
candidates are between 1440 and 1350 cm-1; bands in this
region exhibited small 13C isotope shifts (Figure 7).
(C) Amide Bands. FT-IR spectra derived from both the His-
Tyr dipeptide and compound 1 should exhibit amide I and II
vibrational bands.33 The spectral component at 1631 cm-1 in
Figure 7C and at 1619 cm-1 in Figure 7D may be attributable
to amide I in the dipeptide and the cross-link, respectively. The
1537 cm-1 band in Figure 7C is a candidate for the amide II
vibrational mode in the His-Tyr dipeptide. The location of the
amide II band in compound 1 is not obvious, but this band may
overlap the putative tyrosinate ring mode at ∼1580 cm-1 (Figure
7D).
Figure 6. EPR spectra of free radicals produced in solutions of tyrosine
(A, solid line), 13C(6)-tyrosine (B), His-Tyr dipeptide (C), and His-phenol
cross-linked compound (Cpd 1) (D, solid line). Samples were 100 mM and
also contained 10 mM borate-NaOH, pH 11. The radicals were generated
by UV photolysis at 77 K; see Materials and Methods for conditions. EPR
simulations are superimposed as the dashed lines in panels A and D. See
Table 1 for parameters employed in the simulations. The dotted line in
panel A is a negative control showing the results obtained with borate buffer
alone.
(D) Other Vibrational Modes. Bands at 1437, 1419, 1360,
and 1331 cm-1 in tyrosine (Figure 7A) may originate from υ19b
,
C-C/C-H bending modes of the tyrosinate R and â carbon
atoms, or from the carboxylate symmetric stretching vibration.
This spectral region was altered in the His-Tyr dipeptide (Figure
7C) and in compound 1 (Figure 7D). Remaining unique bands
in the spectrum of compound 1 (Figure 7D) at 1478, 1451, and
1308 cm-1 are potentially attributable to vibrations of the
o-substituted imidazole moiety (see below), to C-CH3 bending
modes, and to coupled vibrations arising from both the imidazole
and the phenolate moieties. The ester CdO of compound 1
should also make a spectral contribution between 1750 and 1720
cm-1; this contribution may be obscured by the broad band at
Table 1. EPR Simulation Parametersa
hfi
Axx (G)
Ayy (G)
Azz (G)
Aiso (G)
φ
Tyrosineb
2.7
C2′-H
C3′-H
C5′-H
C6′-H
C1′-Hâ1
C1′-Hâ2
1.7
-9.6
-9.6
1.7
0.62
11.0
0.4
-7.0
-7.0
0.4
0.58
11.0
1.6
-6.5
-6.5
1.6
0.59
11.0
30
-23
23
-30
0
-2.8
-2.8
2.7
0.58
11.0
0
Compound 1c
C2′-N
C3′-H
C4′-H
C5′-H
C6′-H
2.0
1.7
1.1
2.7
1.5
0.4
1.5
1.6
23
-30
0
30
-23
approximately 1660 cm-1 28
.
FT-IR Spectra Associated with Photolysis. UV photolysis
was performed to examine the vibrational spectrum of the
tyrosyl radical. Vibrational spectroscopy has previously shown
that oxidation of phenolate and tyrosinate upshifts the C-O
stretching vibration and perturbs aromatic ring stretching
vibrations.23,29,30,34,35 Our EPR control experiments verify that
the neutral tyrosyl radical can be generated in tyrosinate
solutions under these conditions (see discussion above). Figure
8B shows the FT-IR difference spectrum associated with the
oxidation of tyrosinate at 77 K. Positive features in these data
arise from unique vibrational modes of the neutral Tyr• radical,
while negative features arise from unique modes of the tyrosinate
ground state. As expected, photolysis experiments performed
on the borate buffer alone gave a flat baseline (Figure 8A).
Photolysis-induced difference spectra were also acquired from
13C(6)-tyrosine (Figure 8C), the His-Tyr dipeptide (Figure 8D),
and compound 1 (Figure 8E). Each spectrum (Figure 8B-E)
-10
-10
-10
-10
1.7
2.7
0.4
1.6
-9.6
-2.8
-7.0
-6.5
a EPR simulations were performed through the use of a computer program
described in ref 20. Hyperfine interactions were obtained from ref 25 with
minor modifications and are reported in Gauss. Euler angle φ represents
the in-plane rotation of the hyperfine component about the gz axis and is
defined as positive clockwise rotation. Euler angles θ and ψ were set to 0
in each case. The isotropic g values were obtained experimentally. For
numbering purposes, the hydroxyl group is at carbon number 4 for tyrosine,
Tyr, and carbon number 1 for compound 1. b Tyrosine: gxx ) 2.0070, gyy
) 2.0046, gzz ) 2.0024; line width ) 2.7; frequency ) 9.21 GHz.
c Compound 1: gxx ) 2.0078, gyy ) 2.0054, gzz ) 2.0024; line width )
2.0; frequency ) 9.21 GHz.
may arise from υ8a and υ19a (Figure 7D). This preliminary
assignment suggests that the effect of o-N substitution on
ground-state phenolate vibrational modes is a downshift of the
highest energy ring mode and an upshift of the υ19a C-C/C-H
vibration.
(33) Krimm, S.; Bandekar, J. AdV. Protein Chem. 1986, 38, 181-364.
(34) Qin, Y.; Wheeler, R. A. J. Chem. Phys. 1995, 102, 1689-1698.
(35) Kim, S.; Barry, B. A. J. Phys. Chem. B 2001, 105, 4072-4083.
(32) Nwobi, O.; Higgins, J.; Xuefeng, Z.; Ruifeng, L. Chem. Phys. Lett. 1997,
272, 155-61.
9
1756 J. AM. CHEM. SOC. VOL. 124, NO. 8, 2002

Modeling the Active Site of Cytochrome Oxidase
A R T I C L E S
Figure 7. FT-IR spectra of tyrosine (A), 13C(6)-tyrosine (B), His-Tyr dipeptide (C), and His-phenol cross-linked compound (compound 1) (D). Samples
were 100 mM and also contained 10 mM borate-NaOH, pH 11. See Materials and Methods for conditions. The borate contributions have been subtracted
from these data. The tick marks on the y-axis correspond to 0.4 absorbance unit.
exhibited a broad positive band between 1665 and 1650 cm-1
which we assign to perturbations of OH (water) and NH (amine)
bending modes upon tyrosinate oxidation. The NH perturbation
will be described in detail in a future publication.
,
these may correspond to coupled vibrational modes, as discussed
above. Likely candidates had frequencies of 1479, 1451, and
1303 cm-1 (Figure 8D). These spectral features were unique to
the His-phenol cross-link and had similar frequencies in the
ground-state spectrum (Figure 7D). Observation of these bands
in the difference spectrum suggests that the photolysis reaction
perturbs force constants in the imidazole group of compound
1.
(B) Positive Spectral Contributions. The tyrosyl radical is
expected to contribute positive lines in Figure 8. Candidates
for tyrosyl ring stretching modes were observed at 1620 and
1577 cm-1 (Figure 8B). These bands shifted either to 1611 or
to 1527-1505 cm-1, respectively, upon 13C ring labeling (Figure
8C). On the basis of a previous Raman study of phenoxyl radical
and ab initio calculations for phenoxyl and tyrosyl radicals, a
frequency of ∼1620-1560 cm-1 and an isotope shift of ∼50-
60 cm-1 are expected for υ8a.29-32,37 We tentatively assign the
1577 cm-1 band, which may exhibit a large 13C isotope shift,
to the υ8a ring stretching vibration and the 1620 cm-1 band to
another ring stretching vibration, possibly υ8b. In the His-Tyr
dipeptide and compound 1 (Figure 8D,E), the 1620 and 1577
cm-1 bands were not evident, perhaps because of cancellation
in the difference spectrum.
(A) Negative Spectral Contributions. Perturbed tyrosinate
vibrations should be observed in the difference spectrum as
negative bands (see section above). In Figure 8B, negative lines
at 1606, 1500, 1263, and 1173 cm-1 were observed in the
tyrosine photolysis spectrum. Corresponding bands (frequencies
within 5 cm-1) were found in the ground-state tyrosinate
spectrum (Figure 7A) and were assigned to C-C ring stretching,
C-C/C-H stretching, C-O stretching, and C-H bending
modes, respectively. These negative bands in Figure 8B were
downshifted to 1560 or 1547, 1466, 1235, and 1167 cm-1 in
13C(6)-tyrosine (Figure 8C). They were observed at 1603, 1502,
1263, and 1173 cm-1 in the His-Tyr dipeptide (Figure 8D) and
at 1592 (shoulder), 1511, and 1265 cm-1 in compound 1 (Figure
8E). A negative band between 1570 and 1530 cm-1 in tyrosine
(Figure 8B) and 1577-1574 cm-1 in the dipeptide (Figure 8D)
and compound 1 (Figure 8E) may be assignable to υ8b. This
band may have a frequency of 1547 cm-1 in the 13C isotopomer
(Figure 8C). A band at 1415-1410 cm-1 in the dipeptide (Figure
8D) and in compound 1 (Figure 8E) may be assignable to υ19b
,
The C-O stretch of the neutral phenoxyl radical has been
previously assigned to a band at approximately 1505 cm-1, while
the C-O stretch of a tyrosyl radical has been assigned to a
although this assignment awaits further analysis (see discussion
above).
Unique negative features in spectra derived from compound
1 may be assigned to C-C/C-N stretching and C-H bending
vibrations of the o-substituted imidazole ring.36 Alternatively,
(36) Majoube, M.; Vergoten, G. J. Mol. Struct. 1992, 266, 345-352.
(37) Qin, Y.; Wheeler, R. A. J. Am. Chem. Soc. 1995, 117, 6083-6092.
9
J. AM. CHEM. SOC. VOL. 124, NO. 8, 2002 1757

A R T I C L E S
Cappuccio et al.
Figure 8. Difference FT-IR spectra associated with the production of radicals by UV photolysis at 77 K. The samples were borate alone (A), tyrosine (B),
13C(6)-tyrosine (C), His-Tyr dipeptide (D), and His-phenol cross-linked compound (1) (E). The conditions are described in the Materials and Methods
section and are identical to those employed for Figure 7. Individual difference spectra correspond to a total of 4 min of data acquisition, 2 min before UV
photolysis and 2 min after photolysis. The difference spectra were constructed by directly ratioing data obtained after 5 (A-D) or 10 (E) laser flashes to data
obtained before photolysis. To obtain the final signal-to-noise, spectra were averaged over two (A, C), eight (B), or three (D, E) different samples. Tick
marks on the y-axis represent 2 × 10-2 absorbance unit.
band at 1515 cm-1 23,29,30,34,38,39
.
The tyrosyl radical produced
intermediate of cytochrome bo from Escherichia coli was
recently attributed to the C-O stretching mode of a tyrosyl
radical.40
in tyrosinate solutions exhibited a positive band at 1516 cm-1
(Figure 8B); this band apparently shifted to 1480 cm-1 upon
13C labeling (Figure 8C). The frequency and magnitude of the
isotope shift support the assignment of this band or a component
of the band to a C-O stretching vibration (υ7a).29,30
Other vibrational bands were not observed in all samples.
For example, spectra derived from the His-Tyr dipeptide and
the His-phenol cross-link show a positive band at 1548 cm-1
,
which was not observed in tyrosine. This band was also observed
in dipeptides with other sequences (data not shown). The origin
of this band and other unique spectral features will be examined
in other work. Positive spectral bands between 1450 and 1300
cm-1 are tentatively assigned to C-C/C-H bending modes;
this region was compound-specific, as expected. Overall, the
photolysis spectrum obtained from compound 1 had a much
more complex pattern of positive lines than those from tyrosine
or the dipeptide. A subset of the positive bands may arise from
perturbation of the structure of the substituted imidazole ring36
upon tyrosinate oxidation. These vibrations either may be
localized on the imidazole or may arise from coupled vibrational
motions of the imidazole and phenoxyl ring. Candidates were
The His-Tyr dipeptide photolysis spectrum (Figure 8D)
exhibited a band at 1514 cm-1, which may have the same origin
as the 1516 cm-1 line in tyrosine. In the photolysis spectrum
of compound 1, there was no vibrational mode observed at
1516-1514 or 1505 cm-1 (Figure 8E). However, bands at 1522,
1498, and 1487 cm-1 were observed. Our EPR studies, described
above, showed that the g values of the radicals produced in
tyrosine and in compound 1 are indistinguishable. This result
is consistent with a largely unperturbed spin density distribution,
so the most likely candidates for the C-O vibration are either
the 1522 or the 1498 cm-1 band. Such a modest (∼1%) shift
of the C-O vibrational mode in compound 1 is qualitatively
consistent with the small nitrogen hyperfine coupling employed
in the EPR simulations. A shift of the C-O vibration to 1522
or 1498 cm-1 also accounts for the lack of a 2115 cm-1 C-O/
C-C-C combination band23 in the photolysis spectrum of the
cross-link (data not shown). A band at 1489 cm-1 in the P
observed at 1522, 1498, 1487, 1464, 1457, 1443, and 1368 cm-1
.
One explanation of the cross-link FT-IR results is that phenolate
oxidation results in a small delocalization of spin density onto
the imidazole nitrogen of the cross-linked compound (see EPR
results above) and that this spin delocalization alters force
constants in the imidazole moiety.
(38) Johnson, C. R.; Ludwig, M.; Asher, S. A. J. Am. Chem. Soc. 1986, 108,
905-912.
(39) Berthomieu, C.; Boullais, C.; Neumann, J. M.; Boussac, A. Biochim.
Biophys. Acta 1998, 1365, 112-116.
(40) Uchida, T.; Mogi, T.; Kitagawa, T. Biochemistry 2000, 39, 6669-6678.
9
1758 J. AM. CHEM. SOC. VOL. 124, NO. 8, 2002

Modeling the Active Site of Cytochrome Oxidase
A R T I C L E S
Discussion
attribute the first intermediate to a phenoxyl radical species.
The large spectral shift (∼100 nm) between intermediate 1 and
the unperturbed tyrosyl radical may also suggest significant
mixing of the imidazole and phenoxyl electronic states in
compound 1.22,23
The nature of the second species (∼10 µs lifetime) is
unknown, but we can exclude the formation of an oxidized
histidine radical because its spectrum is inconsistent47,48 with
the spectrum we observe. Moreover, electron transfer from
histidine to tyrosine would appear unlikely on the basis of their
reported redox potentials.49,50 The spectrum of the second
species is quite similar to that of oxidized N-methylimidazole
(pH 10) reported previously,51 suggesting that the second species
might be a secondary oxidation product.
EPR Results. The EPR spectrum, derived by photolysis of
compound 1, clearly establishes the generation of a paramagnetic
species in this sample. Our EPR analysis of the paramagnetic
species produced in compound 1 gives an isotropic g value of
2.0052. This g value is not significantly altered relative to the
tyrosyl radical. However, it is distinct from the g values of an
o-aminophenol radical, a histidine-OH radical, and a tryptophan
radical, which are in the 2.0026-2.0037 range.27,52,53 In addition,
the value of the isotropic hyperfine coupling to nitrogen (∼1.5
G) is much smaller than the nitrogen coupling obtained in
o-aminophenol (∼5-7 G).54,55 These results suggest that while
nitrogen substitution in o-aminophenol withdraws a significant
amount of spin density from the phenoxyl ring, in compound 1
there is a modest perturbation of the spin density distribution.
This is supported by recent calculations by Himo et al.,56 which
showed that the cross-link will have minor effects on the
hyperfine coupling constants and spin distribution of the tyrosyl
radical.
The EPR spectrum of the radical of compound 1 in Figure
6C differs from that published by McCauley et al.15 of the
radical form of 2-imidazol-1-yl-4-methylphenol. The difference
is attributed to the 4-methyl group in the latter, which gives
rise to three isotropic methyl protons (1:3:3:1 pattern) with Aiso
of 10-11 G.15 The EPR spectrum of compound 1 is also
significantly different from the radical-generated Paracoccus
denitrificans cytochrome oxidase EPR spectrum reported by
MacMillan et al.13 This signal was attributed to the cross-linked
tyrosine radical. The putative cross-linked tyrosyl radical in
cytochrome c oxidase exhibited large, nearly isotropic â-
methylene proton couplings.13 These protons are absent in
compound 1 employed here.
Synthesis of Compound 1. We report the synthesis of a
model for a post-translationally modified amino acid residue,
which is found at the active site of cytochrome c oxidase. The
preparation of compound 1 and related compounds needed for
future studies requires a method of synthesis that is mild enough
to maintain the protection scheme and the absolute stereochem-
istry of amino acid precursors. Recently, we described such a
method in the copper-catalyzed coupling of substituted imida-
zoles with aryllead(IV) reagents.41 The mild reaction conditions
allow for the incorporation of amino acid residues without
racemization or adventitious deprotection. Reaction of N-R-
acetylhistidine methyl ester (2)42,43 and aryllead reagent (3)44
at room temperature in CH2Cl2 produced coupled product 4 in
35% yield. Compound 4 was the only isomer isolated from the
reaction mixture. Removal of the allyl protection group provided
1 in near-quantitative yield. As opposed to the compounds
prepared in the previous study, compound 1 is devoid of
additional aromatic moieties in the protection groups that could
complicate spectroscopic analysis.45 Efforts to synthesize the
cross-linked histidine-tyrosine complex are underway.
Spectrophotometric Titrations. To evaluate the properties
of compound 1 we carried out spectrophotometric titrations. The
titrations resulted in two pKa values of 5.54 and 8.34, which
were assigned to protons on the ꢀ-nitrogen of the imidazole and
phenolic hydroxide, respectively. The pKa values are similar to
those reported by McCauley et al.15 for 2-imidazol-1-yl-4-methyl-
phenyl (5.54 and 8.60, respectively) but are slightly different
from the pKa values recently reported for 1-(o-hydroxyphenyl)-
imidazole (7.86 and 6.12 for the phenol and imidazole,
respectively).46 However, in all cases, the pKa of the phenolic
proton is significantly lower than for the unsubstituted tyrosine,
which is consistent with the cross-linked histidine-tyrosine
facilitating proton delivery to the binuclear site in cytochrome
c oxidase.7,10,15 Therefore, the change in the electronic structure
resulting from the substitution appears to lower the pKa of the
phenolic hydroxyl group compared to unsubstituted tyrosine.
Origin of the Transient Optical Intermediates. The analysis
of the time-resolved optical absorption difference spectra of
compound 1 indicated that two species were present. Our
primary focus will be on the first, which showed absorption
maxima at ∼500 and ∼325 nm and an apparent lifetime of 1.4
µs. We can exclude the possibility that this species represents
a triplet, since identical time-resolved difference spectra were
observed in the presence and absence of dioxygen. This is
consistent with previous reports that no transient spectrum due
to the triplet state can be observed at high pH.16 The 1.4 µs
species does not show an absorption band at ∼400 nm
characteristic of unperturbed phenoxyl radical.16,23 However,
o-substitution on phenol enables appreciable conjugative effects
that significantly red-shift the 400 nm peak of unperturbed
phenol. For example, the spectrum of o-phenylphenol has been
shown to have maxima at 500 and 360 nm.22 Therefore, we
FT-IR Results. Vibrational spectroscopy was also used to
probe the structure of compound 1. The ground-state FT-IR
spectrum of the cross-link was significantly altered relative to
(47) Adams, G. E.; Aldrich, J. E.; Bisby, R. H.; Cundall, R. B.; Redpath, J. L.;
Wilson, R. L. Radiat. Res. 1972, 49, 278-289.
(48) Parsons, B. J.; Al-Hakim, M.; Phillips, G. O. J. Chem. Soc., Faraday Trans.
1 1986, 1575-1588.
(49) Brabec, V.; Mornstein, V. Biophys. Chem. 1980, 12, 159-165.
(50) Harriman, A. J. Phys. Chem. 1987, 91, 6102-6104.
(51) Rao, P. S.; Simic, M.; Hayon, E. J. Phys. Chem. 1975, 79, 1260-1263.
(52) Lassmann, G.; Eriksson, L. A.; Lendzian, F.; Lubitz, W. J. Phys. Chem.
2000, 104, 9144-9152.
(41) Elliott, G. I.; Konopelski, J. P. Org. Lett. 2000, 2, 3055-3057.
(42) Mathews, H. R.; Rapoport, J. J. Am. Chem. Soc. 1973, 95, 2297-2303.
(43) Kasaf´ırek, E.; Bart´ık, M. Collect. Czech. Chem. Commun. 1980, 45, 442-
451.
(53) Lendzian, F.; Sahlin, M.; MacMillan, F.; Bittl, R.; Fiee, R.; Po¨tsch, S.;
Sjo¨berg, B.-M.; Gra¨slund, A.; Lubitz, W.; Lassman, G. J. Am. Chem. Soc.
1996, 118, 8111-8120.
(44) Morgan, J.; Pinhey, J. T. J. Chem. Soc., Perkin Trans. 1 1993, 1673-
(54) Dixon, W. T.; Moghimi, M.; Murphy, D. J. Chem. Soc., Faraday Trans. 2
1974, 70, 1713-1720.
1676.
(45) Elliott, G. I.; Konopelski, J. P. Tetrahedron 2001, 57, 5683-5705.
(46) Collman, J. P.; Wang, Z.; Zhong, M.; Zeng, L. J. Chem. Soc., Perkin Trans.
1 2000, 1217-1221.
(55) Neta, P.; Fessenden, R. W. J. Phys. Chem. 1974, 78, 523-529.
(56) Himo, F.; Eriksoon, l. A.; Blomberg, M. A. A.; Siegbahn, P. E. M. Int. J.
Quantum Chem. 2000, 76, 714-723.
9
J. AM. CHEM. SOC. VOL. 124, NO. 8, 2002 1759

A R T I C L E S
Cappuccio et al.
tyrosine or to a histidine-tyrosine dipeptide and showed
frequency shifts of bands assigned to ring C-C and C-O
stretching vibrations, as well as other spectral changes. The
frequency shifts of phenolate vibrational modes may be caused
by a mass effect or by changes in intermolecular hydrogen
bonding. New vibrational bands in the 1490-1290 cm-1 region,
when the spectrum of compound 1 is compared to the His-
Tyr dipeptide or tyrosinate spectrum, may arise from unique
structural elements of the cross-linked compound as well as
coupled vibrational motions of the phenolate and imidazole
rings.
The difference FT-IR spectrum associated with the oxidation
of compound 1 was also obtained and was found to be altered
compared to the tyrosyl radical or to the tyrosyl radical produced
in the dipeptide. New vibrational bands had frequencies
consistent with a direct contribution of imidazole to the
photolysis spectrum. Shifts of C-C and C-O bands assigned
to the phenoxyl ring were also observed. These shifts may be
associated with the small delocalization of spin density onto
the imidazole moiety, which is predicted by the EPR simula-
tions. This and other possible explanations will be evaluated in
future work.
substitution, the optical and vibrational properties are distinct
when compared to an unmodified tyrosyl radical. The origin of
these unique spectral properties will be explored with isotopic
labeling studies. This research and future studies on the cyclic
pentapeptide (including the His-phenol cross-link) alone and
incorporated into a Cu-containing ligand system will provide
the foundation for understanding the structural and functional
properties of the cross-linked His-Tyr cofactor in cytochrome
c oxidase.
Acknowledgment. J.A.C. and I.A. contributed equally to this
work. This work was supported by National Institutes of Health
Grants GM53788 (OÄ .E.), GM43273 (B.A.B), and GM19541
(I.A.), the American Chemical Society Petroleum Research Fund
(32051-AC1), the American Cancer Society (RGP-93-017-06-
CDD), and the California Department of Health Services, Cancer
Research Program (99-00632V-10129) (J.P.K). J.A.C. was
supported by the Graduate Assistance in the Areas of National
Need (GAANN) grant, and G.I.E. acknowledges a Roche
Bioscience (Palo Alto, CA) Fellowship in synthetic organic
chemistry. We thank Dr. John S. Winterle and Professors
Rebecca Braslau and Eugene Switkes, Department of Chemistry
and Biochemistry, University of California at Santa Cruz, for
helpful discussions.
Summary
While the EPR properties of the radical derived from
compound 1 are not dramatically perturbed by imidazole
JA011852H
9
1760 J. AM. CHEM. SOC. VOL. 124, NO. 8, 2002