Insights into the functional role of the tyrosine-histidine linkage in cytochrome c oxidase [17]

Full text of DOI:10.1021/ja993774s

J. Am. Chem. Soc. 2000, 122, 2403-2404
Insights into the Functional Role of the
2403
Tyrosine-Histidine Linkage in Cytochrome c
Oxidase
Kevin M. McCauley, Jennifer M. Vrtis, Joseph Dupont,† and
Wilfred A. van der Donk*
Department of Chemistry
UniVersity of Illinois at Urbana-Champaign
600 South Mathews AVenue, Urbana, Illinois 61801
ReceiVed October 21, 1999
Recent crystallographic studies of cytochrome c oxidase (CcO)
have revealed a unique and unexpected posttranslational modi-
fication in the enzyme active site. This modification results in a
tyrosine cross-linked at C6 to the ꢀ-nitrogen of a histidine residue
(Figure 1).1 This finding has raised the question of the role of
this novel cross-link in enzyme catalysis. We report here the first
experimental studies to address this issue. Cytochrome c oxidase
catalyzes the four-electron reduction of molecular oxygen to water,
an essential step in the respiratory chain.2 The active site of the
enzyme contains a binuclear center composed of heme a3 and
CuB (Figure 1). Since its discovery, a number of critical functional
roles have been offered for the unique cross-linked residue. In
most models, the tyrosine, which is located at the end of a possible
proton channel to the surface (K-channel), provides a hydrogen
to the distal oxygen of a dioxygen species bound to the heme.
The detailed mechanism of this process (proton or hydrogen atom
transfer) differs in the various proposals. In their seminal work
on the structure of the enzyme from bovine heart mitochondria,
Yoshikawa and co-workers proposed1a proton transfer from
Tyr244 to a ferric peroxide to generate a hydroperoxo adduct. In
Figure 1. Binuclear center of cytochrome c oxidase from bovine heart in
the reduced state.1a Coordinates 10CR from RCSB PDB. The axial his-
tidine ligand to the heme and the heme side chains are omitted for clarity.
1+
a subsequent step, CuB was suggested to provide an electron
via the His240-Tyr244 cross-link to cleave the O-O bond of the
ferric hydroperoxide. On the other hand, several groups have
speculated that the cross-linked tyrosine might serve as a hydrogen
atom donor during the reduction of O2.1b,2b,c,3 In this model, the
metals in the active site provide three of the four electrons required
for O-O bond cleavage by oxidation of the Fe3a2+/CuB+ binuclear
center to oxoferryl (Fe4+dO) and HO-CuB2+. On the basis of
the proximity of the cross-linked tyrosine, it was suggested that
this residue might supply both the proton and the fourth electron
required for this transformation, via hydrogen atom transfer from
the neutral tyrosine. Thus, a tyrosyl radical would be generated,
placing CcO in the company of a growing number of proteins
that utilize redox active amino acids during catalysis.4 The
feasibility of these functional roles, as well as others,5 critically
depends on the perturbation of the physicochemical properties
of the tyrosine as a result of the covalent bond to histidine. We
provide the first experimental insight into these issues.
Figure 2. Spectrophotometric titrations of aqueous solutions of 1 (0.4
mM, closed circles) and p-cresol (0.56 mM, open circles). The titration
was monitored at 314 nm for 1 at which wavelength only the deprotonated
form absorbs. Inset shows pH dependence of the absorbance at 230 nm.
For p-cresol the titration was monitored at 297 nm.
moiety (Figure 2).7 Careful examination of the spectra at 230 nm
revealed a second pKa at 5.54 ( 0.12, assigned to the imidazole
group of 1 (inset Figure 2). Titration of p-cresol under identical
conditions produced a pKa of 10.23 ( 0.09. Thus, the covalently
linked imidazole group perturbs the acid dissociation constant of
the phenol by more than 1.5 orders of magnitude. This modulation
is most likely due to stabilization of the phenolate through an
inductive effect that may be even more pronounced in the protein
by coordination of the imidazole to CuB.8
Electrochemical experiments were performed to assess the
influence of the cross-link on the redox properties of 1. Cyclic
voltammetry at pH 11.5 produced irreversible waves for both 1
and p-cresol, with the anodic peak potential of the former 66 (
3 mV more positive.9 This increase of the peak potential is
To evaluate the modulation induced by the cross-link, we have
prepared 2-imidazol-1-yl-4-methylphenol (1).6 The UV-vis
spectrum of 1 showed a λmax at 286 nm at pH 3-7, and a λmax at
304 nm in basic solution (pH >10). Spectrophotometric titration
of 1 at 314 nm revealed a pKa of 8.60 ( 0.04 for the phenol
(5) In addition to involvement in O-O bond cleavage, the Tyr-His cross-
link may serve other roles. Yoshikawa1a suggested that the cross-linked tyrosine
is deprotonated in the fully oxidized state of the enzyme. Thus, protonation
of Tyr244 is required to prime the enzyme for catalysis.1a Indeed, in the
reductive half of the catalytic cycle, the fully oxidized enzyme has been
reported to take up two protons as well as two electrons to reduce the binuclear
site to its Fe2+/Cu+ oxidation state (R-state) (reviewed in ref 2b). It has been
proposed that these protons are delivered to the binuclear center via Tyr244,
and that one of the protons is used to protonate Tyr244 and prepare it for its
role in O-O cleavage.2b An alternative view of proton movement is presented
in refs 2c,d.
(6) Full experimental details can be found in the Supporting Information.
See also: Strehlke, P. Eur. J. Med. Chem. 1979, 14, 227-230.
(7) The pKa values were determined by two independent titrations of 0.4
mM analyte using 17 different buffer solutions adjusted to 0.3 M ionic strength
with KCl.
* To whom correspondence should be addressed.
† Undergraduate Snyder Scholar at the University of Illinois.
(1) (a) Yoshikawa, S.; Shinzawaitoh, K.; Nakashima, R.; Yaono, R.;
Yamashita, E.; Inoue, N.; Yao, M.; Fei, M. J.; Libeu, C. P.; Mizushima, T.;
Yamaguchi, H.; Tomizaki, T.; Tsukihara, T. Science 1998, 280, 1723-1729.
(b) Ostermeier, C.; Harrenga, A.; Ermler, U.; Michel, H. Proc. Natl. Acad.
Sci. U.S.A. 1997, 94, 10547-10553. (c) Buse, G.; Soulimane, T.; Dewor,
M.; Meyer, H. E.; Bluggel, M. Protein Sci. 1999, 8, 985-990.
(2) (a) Ferguson-Miller, S.; Babcock, G. T. Chem. ReV. 1996, 96, 2889-
2907. (b) Gennis, R. B. Biochim. Biophys. Acta 1998, 1365, 241-248. (c)
Michel, H. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 12819-12824. (d) Michel,
H. Biochemistry 1999, 38, 15129-15140.
(3) Proshlyakov, D. A.; Pressler, M. A.; Babcock, G. T. Proc. Natl. Acad.
Sci. U.S.A. 1998, 95, 8020-8025.
(4) Stubbe, J.; van der Donk, W. A. Chem. ReV. 1998, 98, 705-762.
10.1021/ja993774s CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/26/2000

2404 J. Am. Chem. Soc., Vol. 122, No. 10, 2000
Communications to the Editor
Figure 4. X-band EPR spectrum of the radical form of 1. Instrument
settings: frequency 9.27 GHz, power 100 µW, modulation amplitude 1
G, time constant 0.032 s, and temperature 77 K.
Figure 3. pH dependence of the midpoint potential of 1 in aqueous buffer
determined by DPV. Potential values are versus NHE. Working electrode,
glassy carbon; counter electrode, Pt wire; reference electrode, Ag+/AgCl.
All buffers were adjusted to identical ionic strengths using KNO3 electro-
lyte. Solid line produced by fitting the data to the following equation:18
Em ) Cst + 0.059 log{[H+]2 + Ka1[H+] + Ka1Ka2} - 0.059 log{[H+]}
+ Ka3}.
reaction by lowering the tyrosine pKa. In the oxidative half
reaction, the hydrogen atom transfer model requires that hydrogen
atom abstraction is facilitated by the cross-link without stabilizing
the resulting His-Tyr radical so much that its subsequent one-
electron reduction becomes less favorable. Compound 1, with its
lowered pKa and slightly increased oxidation potential, provides
such a delicately balanced system.14
Scheme 1
Very recently, MacMillan and co-workers reported the EPR
signal of a radical generated in CcO.15 The signal displayed a
giso value of ∼2.0055 with a spectral width of about 80 G, and
was proposed to originate from the cross-linked tyrosine.16 While
the functional importance of this signal remains to be established,
its assignment by double resonance techniques will be assisted
by the availability of spectroscopic data on the radical form of 1.
In a preliminary experiment, we have generated the latter by UV-
photolysis of a frozen solution at pH 10 resulting in the EPR
signal shown in Figure 4. The overall width and giso of the signal
(2.0058) are similar to those reported for the enzyme.17 High-
frequency EPR and double resonance techniques on the enzyme
signal and 1• should enable the assignment of the hyperfine inter-
actions and establish whether the former is indeed associated with
the cross-linked tyrosine residue. Such experiments as well as incor-
porating 1 into a copper-containing model are currently underway.
consistent with the imidazole withdrawing electron density from
the phenolate. The pH dependence of the midpoint potential for
one-electron oxidation of 1 determined by differential pulse
voltammetry (DPV) is shown in Figure 3. The data are best fit
as a one-electron oxidation process that is affected by three
different dissociation constants.10 Two of the three pKa values
thus obtained (pKa2 ) 5.7 ( 0.20, and pKa1 ) 8.90 ( 0.08) agree
reasonably well with those determined by spectrophotometric
titration of 1 (Scheme 1).11 On the basis of the direction of the
change in slope at the third pKa (pKa3 ) 4.80 ( 0.24), it is
associated with protonation of the imidazole of the oxidized
species. The alternative, protonation on oxygen of the phenol
radical, would require significantly lower pH values.12
Combining the experimentally determined values of the equi-
librium acidities (pKa) and oxidation potentials of their conjugate
anions (E1/2 at pH 11.5) for 1 and p-cresol allows an estimation
of the influence of the cross-link on the bond dissociation energy
of the phenol oxygen-hydrogen bond using a thermodynamic
cycle.13 The increase in the oxidation potential, coupled with the
depression of the pKa, leads to a 0.7 kcal/mol decrease in BDE
energy of the O-H bond in 1 compared to p-cresol.13 Thus, this
study supports previous proposals5 that the cross-link will facilitate
proton delivery to the binuclear center in the reductive half
Acknowledgment. This work was supported by the Department of
Chemistry and the Snyder Scholar Fund at the University of Illinois. We
are grateful to Profs. Gennis and Martinez (UIUC) and Prof. Babcock
(Michigan State) for sharing unpublished data and for stimulating
discussions. We thank Prof. Suslick (UIUC) for use of electrochemical
equipment and Prof. Gerlt for help with preparing Figure 1.
Supporting Information Available: Synthetic procedures for the
preparation of 1 and fits of the DPV data to alternative models (PDF).
This material is available free of charge via the Internet at http://pubs.acs.org.
JA993774S
(13) This procedure has been used widely by Bordwell and co-workers
and provides the following equation: ∆BDE(ArO-H) ) F∆E01/2(ArO-) +
2.302RT∆pK1(ArOH). Bordwell, F. G.; Bausch, M. J. J. Am. Chem. Soc. 1986,
108, 1979-1985. For a rigorous application reversible redox potentials are
required. Despite this Bordwell13b and Parker13c have suggested that the error
introduced by using irreversible potentials for a family of compounds is not
significant or is fortuitously canceled. (b) Bordwell, F. G.; Cheng, J.-P.; Ji,
G.-Z.; Satish, A. V.; Zhang, X. J. Am. Chem. Soc. 1991, 113, 9790-9795.
(c) Parker, V. D. J. Am. Chem. Soc. 1992, 114, 7458-7462 and references
therein; correction: Parker, V. D. J. Am. Chem. Soc. 1993, 115, 1201.
(14) An alternative interpretation is that the relatively minor perturbation
in the BDE indicates that the cross-link merely serves a structural role. The
importance of the cross-link for proper assembly of the binuclear center has
been shown by site-directed mutagenesis: Das, T. K.; Pecoraro, C.; Tomson,
F. L.; Gennis, R. B.; Rousseau, D. L. Biochemistry 1998, 37, 7, 14471-14476.
(15) MacMillan, F.; Kannt, A.; Behr, J.; Prisner, T.; Michel, H. Biochemistry
1999, 38, 9179-9184.
(8) An N-linked imidazole is not expected to lower the pKa of the phenol
by resonance stabilization of the conjugate anion. The increased acidity of
the phenol is therefore best explained as an inductive electron-withdrawing
effect. The influence of copper coordination will depend on its oxidation state
as well as coordination geometry. Generally, coordination will increase the
inductive electron-withdrawing effect, but dπ-pπ back-bonding may offset
this. Note also that the cross-link lowers the basicity of the imidazole nitrogen,
which will weaken its coordinating ability.
(9) This value was derived from five independent scans at 100 mV s-1
and probably represents an upper value. At 10 mV s-1, 1 showed a peak
potential 73.0 ( 4.9 mV more positive than p-cresol.
(10) Two independent determinations of the pH dependence of the half-
wave potential both showed a significantly better fit to three than to two
dissociation constants. Figures showing alternative fits are provided in the
Supporting Information.
(16) If the signal is indeed associated with the cross-linked tyrosine one
might have expected significant broadening of the signal or even elimination
of any signal due to magnetic coupling between the radical and CuB2+ via the
direct covalent connection (Figure 1).
(11) While the pKa values determined from two independent electrochemical
experiments proved reproducible within 0.2, the spectrophotometric values
are more accurate since the former are extracted from an irreversible redox
system, which precludes accurate determination of E°.
(12) Dixon, W. T.; Murphy, D. J. Chem. Soc., Faraday Trans. 2 1976, 72,
1221-1230.
(17) The two signals differ in line shape since the signal in CcO is largely
governed by the orientation of the two â-protons of the Tyr residue, while
the line shape for 1• is dominated by three isotropic methyl protons (1:3:3:1
pattern with Aiso of 10-11 G).
(18) Clark, W. M. Oxidation-Reduction Potentials of Organic Systems;
Krieger Publishing Co.: Huntington, NY, 1972.