Content uploaded by Peter Faller
Author content
All content in this area was uploaded by Peter Faller
Content may be subject to copyright.
Rapid formation of the stable tyrosyl radical in
photosystem II
Peter Faller*, Richard J. Debus
†‡
, Klaus Brettel*, Miwa Sugiura
§
, A. William Rutherford*
‡
, and Alain Boussac*
*Section de Bioe´ nerge´ tique, De´ partement de Biologie Cellulaire et Mole´ culaire (DBCM), Commissariat a` l’Energie Atomique (CEA) Saclay, Centre National
de la Recherche Scientifique (CNRS) Unite´ de Recherche Associe´ e (URA) 2096, 91191 Gif-sur-Yvette Cedex, France;
†
Department of Biochemistry,
University of California, Riverside, CA 92521; and
§
Department of Applied Biological Chemistry, Osaka Prefecture University,
1-1 Gakuen-cho, Sakai, Osaka 599-8531, Japan
Edited by Christopher T. Walsh, Harvard Medical School, Boston, MA, and approved October 12, 2001 (received for review July 24, 2001)
Two symmetrically positioned redox active tyrosine residues are
present in the photosystem II (PSII) reaction center. One of them,
TyrZ, is oxidized in the ns–
s time scale by P680
ⴙ
and reduced
rapidly (
s to ms) by electrons from the Mn complex. The other
one, TyrD, is stable in its oxidized form and seems to play no direct
role in enzyme function. Here, we have studied electron donation
from these tyrosines to the chlorophyll cation (P680
ⴙ
) in Mn-
depleted PSII from plants and cyanobacteria. In particular, a mutant
lacking TyrZ was used to investigate electron donation from TyrD.
By using EPR and time-resolved absorption spectroscopy, we show
that reduced TyrD is capable of donating an electron to P680
ⴙ
with
t
1/2
⬇ 190 ns at pH 8.5 in approximately half of the centers. This rate
is ⬇10
5
times faster than was previously thought and similar to the
TyrZ donation rate in Mn-depleted wild-type PSII (pH 8.5). Some
earlier arguments put forward to rationalize the supposedly slow
electron donation from TyrD (compared with that from TyrZ) can
be reassessed. At pH 6.5, TyrZ (t
1/2
ⴝ 2–10
s) donates much faster
to P680
ⴙ
than does TyrD (t
1/2
> 150
s). These different rates may
reflect the different fates of the proton released from the respec-
tive tyrosines upon oxidation. The rapid rate of electron donation
from TyrD requires at least partial localization of P680
ⴙ
on the
chlorophyll (P
D2
) that is located on the D2 side of the reaction
center.
T
yrosyl radicals play key roles in the mechanisms of a wide
range of enzymes. Understanding these roles and how pro-
teins are able to control these reactive species to carry out quite
specific chemical reactions has been an important aim of re-
searchers in this area (for review, see ref. 1). One of the earliest
demonstrations of redox active tyrosines was in photosystem II
(PSII), the water oxidizing enzyme, in which a tyrosyl radical,
designated tyrosine Z
䡠
(TyrZ
䡠
), is thought to play a key role in
the active site, abstracting electrons and possibly protons from
the substrate water that is bound to the highly oxidized Mn
cluster (2–5).
PSII contains a second tyrosyl radical, TyrD
䡠
, which is stable
during enzyme function and which is located in a 2-fold rota-
tionally symmetrical position to TyrZ on a subunit (D2) adjacent
to that (D1) in which water oxidation takes place. TyrZ and TyrD
are equally situated relative to the central chlorophyll (Chl) pair
(P
D1
and P
D2
) with the center-to-center distances for TyrZ–P
D1
and TyrD–P
D2
being ⬇12.4 Å (6). It is this pair of Chls that bears
the photogenerated cation (P680
⫹
) that is considered to be the
oxidant for the tyrosines (refs. 6–12; for recent reviews see refs.
2 and 3). The existing enzyme may have evolved from an ancestor
in which the core was a homodimer with the two tyrosines having
identical redox functions (13).
Despite homologous positions in subunits D1 and D2, TyrZ
and TyrD exhibit extremely different kinetics, different redox
potentials, and play completely different functional roles in the
enzyme. Thus, they constitute an ideal system for understanding
how the protein environment controls the reactivity of these
species.
Several studies have focused on the differences in the kinetics
of formation and decay of the two radicals. Most of these studies
were conducted on Mn-depleted PSII because TyrZ
䡠
is much
longer lived in this material, decaying with t
1/2
values of 20–600
ms, in contrast with oxygen-evolving PSII in which the t
1/2
values
are 0.03–1 ms (14–15). Compared with TyrZ
䡠
,TyrD
䡠
is very
stable; it decays with t
1/2
in the minutes–hours time range (refs.
16–18; for reviews, see refs. 2 and 3).
TyrD
䡠
is thought to be immobilized in a site with a well defined
and ordered H-bond interaction between its phenol oxygen and
a proton from D2 His-189 (in cyanobacteria; D2 His-190 in
higher plants) (19). Moreover, the site is well shielded from the
lumen (20–23) and is thought to be relatively hydrophobic. The
TyrZ
䡠
site is quite different. In the absence of Mn, TyrZ
䡠
is
readily accessible from the lumen (22–24) and is thought to be
in a more hydrophilic environment. In addition, TyrZ
䡠
exhibits
a much more disordered H-bond, and the ring seems to be more
mobile (25, 26). The hydrogen bond to TyrZ
䡠
seems to involve
D1 His-190 directly or indirectly (i.e., by means of one or more
water molecules) (for discussion, see refs. 2 and 3).
In the reduced state, TyrD and TyrZ are likely to be proton-
ated and to form a hydrogen bond to D2 His-189 and D1 His-190,
respectively (27, 28). For TyrZ, indications were obtained for a
minor population in which the hydrogen bonds to either water,
a hydroxylated amino acid side chain, or to the protonated
imidazole side chain of His (i.e., HisH
2
⫹
instead of HisH) (27).
Because TyrZ and TyrD are protonated in the reduced state
and deprotonated in the oxidized state (Tyr
䡠
), the oxidation
involves both electron and proton transfer. In PSII lacking the
Mn-cluster, the t
1/2
(190 ns–30
s) of the electron transfer from
TyrZ to P680
⫹
was reported to depend on the biological origin
of PSII, the presence of Ca
2⫹
, and the pH (23, 24, 29–31). The
rate increased with the pH from tens of
s at acidic pH to
sor
sub-
s at basic pH, exhibiting pK
a
values of 8.3 (23), 7.0 (24), and
7.5 (31). This pK
a
was ascribed to the ionization of either TyrZ
or D1-His-190 (23, 24, 31). In the latter case, the suggestion is
that at low pH, the His is protonated and has to be deprotonated
before it is able to accept a proton from TyrZ. In contrast, at high
pH (⬎8–9), the His is thought to be deprotonated already and,
therefore, the electron transfer can occur more rapidly (sub-
s
or
s) (31, 32).
It is widely accepted that electron donation from TyrD to
P680
⫹
is relatively slow compared with TyrZ, although its redox
potential is much lower (TyrD, ⬇0.75 V; TyrZ, 0.95 V–1.1 V).
This notion is largely based on the work of Buser et al. (33), who
measured the yield of oxidation of all of the electron donors to
P680
⫹
at pH 5.0–7.5 in Mn-depleted PSII. A global fit gave a t
1/2
This paper was submitted directly (Track II) to the PNAS office.
Abbreviations: Mes, 2-(N-morpholino)ethanesulfonic acid; Chl, chlorophyll; P680, pho-
tooxidizable chlorophyll; P
D1
,P
D2
, two central chlorophylls bound to polypeptide D1 and
D2, respectively; PSII, photosystem II.
‡
To whom reprint requests should be addressed. E-mail: Richard.Debus@ucr.edu or
rutherford@dsvidf.cea.fr.
The publication costs of this article were defrayed in part by page charge payment. This
article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.
§1734 solely to indicate this fact.
14368–14373
兩
PNAS
兩
December 4, 2001
兩
vol. 98
兩
no. 25 www.pnas.org兾cgi兾doi兾10.1073兾pnas.251382598
⫽ 10–20 ms for the electron transfer from TyrD to P680
⫹
at pH
values between 5.0 and 7.5. Thus, the oxidation yield appeared
to be practically independent of pH over this range (33).
Intriguingly, indications for TyrD electron donation to P680
⫹
with a rate similar to TyrZ exist in the older literature, but these
have been largely overlooked by the field (34, 35). The present
work was performed to readdress the electron donation rate
from TyrD to P680
⫹
by using kinetic absorption and EPR
spectroscopies and a mutant lacking TyrZ.
Materials and Methods
Construction of D1-Y161F Mutants and Isolation of PSII Particles from
Synechocystis sp.
The D1-Y161F mutation was constructed in the
psbA-2 gene of the cyanobacterium Synechocystis sp. strain PCC
6803 (36). The plasmid bearing this mutation was transformed
into a host strain of Synechocystis that lacks all three psbA genes
(37) and contains a hexahistidine-tag (His-tag) fused to the C
terminus of CP47 (38). Single colonies were selected for their
ability to grow on solid media containing 5
g兾ml kanamycin
monosulfate (37). The control wild-type⬘ strain was constructed
in identical fashion as the D1-Y161F mutant except that the
transforming plasmid carried no site-directed mutation. The
designation wild-type⬘ differentiates this strain from the native
wild-type strain that contains all three psbA genes and is
sensitive to antibiotics. Cells were propagated and PSII particles
were isolated as described (38). The PSII particles were isolated
in a buffer containing 50 mM 2-(N-morpholino)ethanesulfonic
acid (Mes)–NaOH (pH 6.0), 20 mM CaCl
2
, 5 mM MgCl
2
, 0.03%
(wt兾vol) N-dodecyl

-D-maltoside, 25% (vol兾vol) glycerol, and
concentrated to ⬇1 mg of Chl per ml (38). The concentration of
Mes was decreased to 5 mM by 10-fold dilution and was followed
by reconcentration. The final solution contained 5 mM Mes–
NaOH (pH 6.0), 20 mM CaCl
2
, 5 mM MgCl
2
, 0.03% (wt兾vol)
N-dodecyl

-D-maltoside, and 25% (vol兾vol) glycerol. Mn-
depleted wild-type⬘ PSII particles were treated with NH
2
OH and
EDTA, as described (32). After concentration to ⬇1mgofChl
per ml, the concentration of Mes in the Mn-depleted PSII
particles was decreased to 5 mM by dilution and reconcentration,
as described above. The mutant and wild-type⬘ PSII cores were
stored at 77 K at a concentration of about 1.5 mg Chl per ml.
After thawing, 20 mM Tris䡠 HCl (pH 8.5) was added to the
PSII. The PSII particles then were dark-adapted for2hatroom
temperature with 8 mM K
3
Fe
3⫹
(CN)
6
兾8mMK
4
Fe
2⫹
(CN)
6
.
These conditions result in an almost complete reduction of
TyrD. For time-resolved absorption measurements, the samples
were diluted to 0.15 mg Chl per ml in 15 mM NaCl兾0.3 M sucrose
and 50 mM of one of the following buffers: CAPSO兾NaOH (pH
9.5), Tris䡠HCl (pH 8.5 and pH 8.0), Hepes兾NaOH (pH 7.5), or
Mes兾NaOH (pH 6.5). For EPR, 40
l of concentrated PSII was
adjusted to pH 8.5 (pH 6.5) by adding 12.5
l of 0.4 M Tris䡠HCl
at pH 8.5 (or Mes兾NaOH pH 6.5) and then put into a small
volume (⬇50
l) quartz flat cell.
EPR Spectroscopy. X-band EPR spectra and kinetics were re-
corded with an ESP 300 spectrometer (Bruker, Karlsruhe,
Germany) in a flat cell [small size for Synechocystis sp., normal
size (⬇400
l) for spinach]. Excitation flashes were provided
with a Spectra-Physics GCR-230–10 Nd-YAG laser (532 nm, 550
mJ, 8 ns).
Time-Resolved Absorption at Room Temperature. Excitation flashes
(300 ps, 532 nm) were provided with an Nd-YAG laser (Quantel,
Les Ulis, France) with an energy of about 1.5 mJ兾cm
2
. The
measuring light was a continuous laser diode emitting at 820 nm
(SDL-5411-G1, Spectra Diode Labs, San Jose, CA). A cut-off
filter RG780 was placed before the cuvette and an 820-nm
interference filter (10-nm bandwidth) after the cuvette (path
length ⫽ 10 mm, 2 mm broad). The measuring-light intensity and
the flash-induced changes were monitored behind the sample
with a photodiode (FND 100, EG & G, Salem, MA) connected
to an amplifier (HCA, 28dB, DC-325 MHz; FEMTO, Berlin,)
and a digital storage oscilloscope DSA 602A with plug-in 11A52
(DC-20 MHz, Tektronix, Les Ulis, France). The kinetic data
were fitted into a multiexponential decay with a Marquart
least-square algorithm. The program was kindly provided by P.
Setif (Saclay). The very-fast phase with a t
1/2
⫽⬇3nswas
assigned to exited states of loosely bound Chls and were not
considered in the fittings.
GEPASI 3 was downloaded from http:兾兾gepasi.dbs.aber.ac.uk兾
softw兾gepasi.html (for review, see ref. 39). The simulations were
based on Scheme 1. The rate constant k
Rec
⫽ 800 s
⫺1
was taken
from the literature (33). No direct measurement of k
TyrZ
back
is
available in the literature. k
TyrZ
back
was calculated based on the
equilibrium between TyrZ
⫹
and P680
⫹
, which was estimated to
be about 10
3
at pH 8.5 (23, 40, 41). This value yields a k
TyrZ
back
of
about 3.4 ⫻ 10
3
s
⫺1
based on a k
TyrZ
⫽ 3.4 ⫻ 10
6
s
⫺1
at pH 8.5,
as measured in the present study.
Results
Fig. 1 A, C, and E shows the EPR experiments at pH 8.5
performed with the D1-Y161F PS II mutant (i.e., the TyrZ-less
mutant), in which the typical tyrosine radical spectrum arises
solely from TyrD. Fig. 1 B, D, and F show the results obtained
with wild-type⬘ PSII.
Dark incubation of PSII from wild-type and D1-Y161F mu-
tant from Synechocystis sp. for2hatpH8.5inabuffercontaining
K
3
Fe
3⫹
(CN)
6
and K
4
Fe
2⫹
(CN)
6
(see Materials and Methods)
resulted in the reduction of TyrD. Fig. 1 A and B (traces labeled
‘‘dark’’) shows that in both samples, the amplitude of the TyrD
䡠
signal is less than 5% of that measured after 50 flashes (Fig. 1
A and B, traces labeled ‘‘after 50 fl’’). A further series of 50
flashes did not result in an additional increase (data not shown),
indicating that TyrD was fully oxidized after 50 flashes.
The kinetics of TyrD
䡠
and兾or TyrZ
䡠
were measured by mon-
itoring the EPR signal at 3,465 G (1 G ⫽ 0.1 mT; see Fig. 1 A
and B arrow) in these samples. This magnetic field value is
outside the magnetic-field range in which signals from Chl and
carotenoid cations could contribute significantly (42, 43).
Fig. 1C shows the time course of TyrD
䡠
formation and decay
during a series of five flashes separated by 8 s each (flashes are
indicated by arrows) in D1-Y161F PSII. The first flash induced
a signal with an amplitude similar to that measured after 50
flashes (Fig. 1A). This observation indicates that TyrD is
transiently fully oxidized after one flash. The slow time-
resolution of the EPR spectrometer (i.e., 82 ms) used did not
allow the kinetics of TyrD
䡠
to be resolved. About 30% of the
signal decayed within 1 s, whereas the other 70% was relatively
stable. The second flash oxidized the re-reduced fraction of
TyrD, and again, about one third of this fraction decayed within
1 s. With each flash, the stable signal approached the level of
fully oxidized TyrD
䡠
(Fig. 1A). After the fifth flash, the signal
was monitored for 42 s (Fig. 1C shows only 22 s), showing the
beginning of a very slow decay (t
1/2
estimated to be 30 min).
Immediately after the recording of this trace, the TyrD
䡠
spec-
trum was measured (Fig. 1A, trace ‘‘after 5 fl’’), confirming that
almost all TyrD was oxidized after five flashes.
Scheme 1.
Faller et al. PNAS
兩
December 4, 2001
兩
vol. 98
兩
no. 25
兩
14369
BIOPHYSICS
Fig. 1E corresponds to the light-induced signal (averaged over
32 flashes) measured immediately after the recording of the
trace in C. The absence of a reversible Tyr
䡠
signal is as expected,
because TyrZ is absent in this mutant.
Fig. 1D shows the time course of Tyr
䡠
formation (TyrD
䡠
and兾or TyrZ
䡠
) and decay after five flashes separated by 8 s
each (flashes indicated by arrows) in wild-type⬘ PSII. The first
flash oxidized one equivalent of Tyr. About 15% of Tyr
䡠
decayed within a few seconds; the other 85% were relatively
stable. On the following flashes, one equivalent of Tyr
䡠
is
formed,
¶
but the fraction of the signal that was stable became
smaller from flash to flash. The stable signal is attributed to
TyrD
䡠
because of the similar spectrum and decay rate as seen
for TyrD
䡠
in the Z-less mutant (Fig. 1 A and B; trace ‘‘after 5
fl’’) (see also ref. 14). Thus, after the first flash, about 85% of
the TyrD was oxidized
储
, while the remainder was oxidized by
the subsequent flashes.
Fig. 1F shows the flash-induced TyrZ
䡠
signal of wild-type⬘
PSII, in which TyrD is fully oxidized after 50 preflashes (see also
ref. 44). Because it could be averaged (16 flashes, spaced by 8s),
this trace was measured with a higher time resolution and shows
that all of the TyrZ was oxidized upon each flash (i.e., it has the
same intensity as the fully oxidized TyrD in Fig. 1A).
Fig. 2 shows the flash-induced absorption changes at 820 nm
of wild-type⬘ and D1-Y161F PSII mutant from Synechocystis sp.
at pH 8.5. Chl cations absorb at this wavelength; thus, all of the
kinetics show an immediate increase caused by the fast forma-
tion of P680
⫹
followed by a decay that reflects the re-reduction
of P680
⫹
by electron donor(s).
Before the first flash, TyrD was reduced (see above) and thus
available as an electron donor to P680
⫹
. For the time window
shown in Fig. 2, the kinetic traces could be satisfactorily fitted
with three components. The initial spike present during the first
10 ns was omitted for the fits (see Material and Methods). For
the first flash on the wild-type⬘, the fit yielded a fast phase with
t
1/2
⫽ 169 ns (70%), a slower phase with t
1/2
⫽ 895 ns (21%),
and a constant (9%). The constant reflects the slower decays,
which were also measured here by using a longer time range
(data not shown) and were fitted with t
1/2
⫽ 112
s (4%), 820
s
(5%), and 4.3 ms (1%) (see also refs. 30 and 31). The kinetics
after the tenth flash of the wild-type⬘ were not significantly
different from the first flash [t
1/2
⫽ 186 ns (67%), 950 ns (22%),
constant (11%)] (Fig. 2).
Surprisingly, the kinetic traces of the D1-Y161F mutant upon
the first flash revealed about the same t
1/2
values as the wild-
type⬘, although the amplitudes were somewhat different [t
1/2
⫽
183 ns (45%), 675 ns (22%), constant (33%)]. The trace after the
tenth flash was very different; i.e., the constant (
s兾ms phases)
increased dramatically at the expense of the sub-
s phases [fit:
t
1/2
⫽ 222 ns (7%), 1160 ns (13%), constant (80%)]. All of the
fits of the wild-type⬘ and mutant samples at different pHs
revealed similar rates, which were in the following range: fast
phase, t
1/2
⫽ 190 ⫾ 35 ns; slow phase, t
1/2
⫽ 900 ⫾ 300 ns. EPR
experiments showed that almost all TyrD is oxidized after the
first flash (see above). Therefore, the fast phase that occurs upon
¶
In Fig. 1F, TyrZ
䡠
is detected in almost all of the PSII centers. The apparent smaller amplitude
of the reversible signal in Fig. 1D is only due to a lower time resolution.
储
We obtained similar values for the yield of TyrD oxidation after one flash at pH 8.5 for PSII
from Synechococcus elongatus (i.e., 85%) and spinach (i.e., 55%) (data not shown). For
chloroplasts from spinach, values from 65–85% were reported (34, 35).
Fig. 1. Room temperature EPR studies of TyrZ
䡠
兾TyrD
䡠
kinetics on Mn-
depleted PSII particles from Synechocystis sp. strain PCC 6803 at pH 8.5. The
measurements were performed on the mutant D1-Y161F (A, C, E), where
the electron donor Tyr
Z
was replaced by a Phe, and on the wild-type⬘ PSII
(B, D, F). (A and B) Traces ‘‘dark’’ show the spectrum of the dark-adapted
samples. Trace ‘‘after 5 fl’’ is the spectrum measured after 5 saturating
flashes (immediately after the experiment shown in C and D), whereas the
trace ‘‘after 50 fl’’ shows the spectrum taken after 50 flashes (four scans per
spectrum). (C and D) The time course of the EPR signal from (oxidized) Tyr
䡠
measured at 3,464 G (arrows A and B). The five arrows indicate the
saturating laser flashes given to the samples. (E and F) The time course of
the EPR signal from (oxidized) Tyr
䡠
measured at 3,464 G (arrows in A and B)
after illumination with 50 flashes. The signal was averaged over 16 or 32
flashes (repetition rate ⫽ 8 s). EPR conditions: power, 20 mW; microwave
frequency, 9.6 GHz; modulation amplitude, 4.5 G; modulation frequency,
100 kHz. Conversion time: 82 ms for C兾D, 5.12 ms for E兾F. Time constant: 41
ms for C兾D, 2.56 ms for E兾F.
Fig. 2. Flash-induced absorption transients at 820 nm in PSII particles from
Synechocystis sp. strain PCC 6803 at pH 8.5. The traces from the wild-type⬘ PSII
are denoted wt⬘; the traces from the TyrZ-less mutant are denoted D1-Y161F.
The traces show the first and the tenth flash after dark adaptation of the
samples with a mixture of 8 mM K
4
[Fe
2⫹
(CN
⫺
)
6
]and8mMK
3
[Fe
3⫹
(CN
⫺
)
6
]to
reduce TyrD. The initial amplitudes of the traces from the wild-type⬘ were
normalized to that of the D1-Y161F.
14370
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.251382598 Faller et al.
the first flash, which is almost absent after the tenth flash, can
be attributed to electron donation from TyrD to P680
⫹
, with a
t
1/2
⫽⬇190 ns. The phase with a t
1/2
⫽⬇900 ns is also diminished
after the tenth flash, suggesting its assignment to P680
⫹
reduc-
tion by TyrD in a fraction of centers. The higher amplitude of the
s兾ms phases (i.e., the constant in our fit) in the D1-Y161F
mutant than in the wild-type⬘ that occurs upon the first flash is
partially due to the presence of some residual oxidized TyrD
(⬇5% measured by EPR) before the flash experiment (see Fig.
1 A and B and Discussion).
Treatment of the D1-Y161F PSII mutant and wild-type⬘ PSII
with NH
2
OH [instead of K
3
Fe
3⫹
(CN)
6
兾K
4
Fe
2⫹
(CN)
6
] to reduce
TyrD
䡠
did not significantly change the rates and amplitudes of
the traces, when taking into account the longer lifetime of Q
A
⫺
in
the presence of NH
2
OH.
The P680
⫹
reduction in the TyrZ-less mutant (D1-Y161F) was
also measured at different pH values (6.5–9.5; Fig. 3B). TyrD was
prereduced by dark-incubation in the presence of 1 mM NH
2
OH
(as verified by the lack of a TyrD
䡠
signal in EPR). The amplitude
of the long-lived P680
⫹
[i.e., the constant at this time scale (0–4
s); assigned to P680
⫹
兾Q
A
⫺
recombination] increased at the
expense of the sub-
s phases (assigned to electron donation
from TyrD) upon lowering the pH. At pH 6.5, fitting suggested
re-reduction of P680
⫹
by TyrD in 10% and by recombination
with Q
A
⫺
in 90% of the centers. This estimate is in line with the
corresponding EPR experiments, which revealed that only about
16% of all TyrD is oxidized after one flash at pH 6.5 (Fig. 3A).
Fig. 3C shows the sum of the amplitudes of the two sub-
s
phases (corresponding to electron donation from TyrD to
P680
⫹
) plotted against the pH. The experimental data show a
strong pH dependence and were fitted to a single proton
titration, which led to a pK
a
of 7.7 (Fig. 3C).
Discussion
The use of a D1-Y161F (i.e., TyrZ-less) mutant allowed a direct
measurement of the re-reduction rate of P680
⫹
by TyrD. Two
rates with t
1/2
⫽⬇190 ns and 900 ns could be attributed to this
electron donation. These rates are four to five orders of
magnitude faster than the generally accepted rate (reviewed
in ref. 45).
Buser et al. (33) reported a TyrD donation rate to P680
⫹
slower than that reported here (t
1/2
⫽ 10–20 ms at pH 5.0–7.5).
The two measurements agree at low pH (ⱕ6.5) but disagree at
higher pH. Buser et al. (33) measured low yields of TyrD
䡠
at pH
7.5 in the presence of 0.4 mM ascorbate and 20 s after the flash,
and this led them to calculate a much slower rate. Experiments
(data not shown) performed under these conditions with similar
materials (PSII membranes from spinach) gave similar yields.
We suggest that the low yield is due to the ascorbate that may
reduce TyrD
䡠
after the flash and Q
A
before the flash, leading to
an underestimation of TyrD
䡠
yield.
The electron transfer rate from TyrD is very similar to that
from TyrZ to P680
⫹
at higher pH [measured in the presence
of oxidized TyrD (Fig. 2 and ref. 30)]. The sub-
s electron
donation from TyrD to P680
⫹
was measured in a significant
percentage of centers at pH 8.5, i.e., 67%. In the remaining
33% of the centers, recombination between Q
A
⫺
and P680
⫹
is
proposed to occur. In wild-type⬘ PSII with preoxidized TyrD,
the rate assigned to the recombination between Q
A
⫺
and P680
⫹
occurred in only 12% of the centers. This difference in the
amplitude of the recombination phase between wild-type⬘ and
mutant can in part be ascribed to the presence of some TyrD
䡠
(⬇5%) in the TyrZ-less mutant before the first actinic flash
(Fig. 1 A and B, ‘‘dark’’). The reason for the remaining 16%
difference is not known.**
The amplitude of the sub-
s electron donation from TyrD to
P680
⫹
is pH dependent, with an apparent pK
a
close to pH 7.7
(Fig. 3C). At low pH, the TyrD electron donation rate is slower
than the Q
A
⫺
兾P680
⫹
recombination (t
1/2
⬎ 0.15 ms). This pK
a
could represent the pK
a
value of a single amino acid residue. The
D2-His-189 residue is a candidate for the protonable group.
Protonation of the D2-His-189 nitrogen proximal to TyrD would
be expected to have drastic effect on TyrD electron donation. In
this case, at low pH, TyrD would have to donate its electron to
P680
⫹
either, without releasing its proton, which is energetically
very unfavorable (e.g., ref. 46) or with kinetic limitations due to
deprotonation (and reorientation) of the D2-His-189. However,
this explanation is in contradiction with Fourier transform
infrared (FTIR) measurements, suggesting that D2-His-189 is
deprotonated at pH 6.0 (in the presence of formate and phos-
phate) (27).
Another possibility, which is in line with the FTIR measure-
ments, is that the protonable group is TyrD itself and that its
protonation state is mediated by the protonation state of D2-
His-189 at its distal nitrogen. The faster electron-donation rate
of TyrD to P680
⫹
at higher pH (pK
a
of 7.7) is due to the
deprotonation of TyrD before oxidation, i.e., the proton is
already on the proximal nitrogen of D2-His-189 when the distal
nitrogen is deprotonated. At low pH the electron transfer is slow,
the TyrD is neutral and H-bonded to the proximal nitrogen of
D2-His-189, while the distal nitrogen of the D2-His-189 is
protonated. Thus, oxidation of TyrD is limited by H
⫹
transfer to
the D2-His-189, which in turn is limited by the release of H
⫹
from its distal nitrogen (see below).
It is known that TyrZ electron donation also shows a pH
dependence with a pK
a
between 7.0 and 8.3 (23, 24, 31), similar
to that for TyrD reported here. As for TyrD, titration of the TyrZ
itself or the neighboring D1-His-190 were suggested reasons for
the pH dependence (23, 24, 31). At high pH, the donation rates
of TyrZ and TyrD to P680
⫹
seem to be similar (⬇190 ns, see
above). In contrast, at low pH, the rates of the dominating kinetic
**Two possible explanations were tested: (i) Working in total darkness and blocking the
measuring light beam until 2.5 ms before the measurement did not change the results.
Thus it is unlikely that TyrD was oxidized to a significant extent by the measuring and兾or
stay light before the measurement occurred (ii) The incubation time (20 h instead of 2 h)
at pH 8.5 before the measurement did not affect the result, indicating the TyrD site was
pH equilibrated.
Fig. 3. (A) Room temperature EPR studies of TyrD
䡠
kinetics on Mn-depleted
PSII TyrZ-less mutant from Synechocystis sp. at pH 6.5 and pH 8.5. The kinetics
were measured at 3,464 G, and a flash was given at time 0. Other parameters
were as in Fig. 1D.(B) Flash-induced absorption transients at 820 nm in the
Mn-depleted TyrZ-less mutant (D1-Y161F) from Synechocystis sp. at different
values of pH as indicated in the figure. TyrD was prereduced by incubating for
90 min in 1 mM NH
2
OH. (C) The relative amplitude of the sum of the fast
components (⬍1
s, corresponding to TyrD electron donation) at the different
pH values from Fig. 3B. The solid line shows a single proton titration fitofthe
data points.
Faller et al. PNAS
兩
December 4, 2001
兩
vol. 98
兩
no. 25
兩
14371
BIOPHYSICS
phases are very different, with a t
1/2
⫽ 10–30
s for TyrZ (23,
24, 31) and a t
1/2
⬎ 0.2–0.9 ms for TyrD. As suggested above, the
deprotonation of the neighboring His at its distal nitrogen could
be linked to the protonation of its proximal nitrogen by the
proton from the Tyr. This process could be easier with TyrZ than
with TyrD, because D1-Glu-189 appears to be (indirectly)
associated with a TyrZ deprotonation pathway (47), possibly
positioning the proton acceptor for D1-His-190. No equivalent
carboxylic acid is conserved in the D2 sequence associated with
TyrD, where it is replaced by the acid兾base inactive Phe residue
(4, 46, 48–51). Another explanation for TyrZ oxidation being
faster than TyrD oxidation at low pH could be the possibility that
TyrZ
䡠
has multiple H-bonding partners (for review, see ref. 3;
refs. 25, 26, 52), whereas TyrD
䡠
is considered to have D2-His-189
as its unique H-bonding partner (19).
Whereas in the TyrZ-less PSII mutant (D1-Y161F), TyrD is
oxidized directly by P680
⫹
(see above), in the wild-type PSII,
TyrD could be oxidized in two ways: (i) directly by P680
⫹
, and
(ii) by rapid formation of a TyrZ兾P680
⫹
º TyrZ
䡠
(H
⫹
)兾P680
equilibrium followed by a slower oxidation of TyrD by the
equilibrium population of P680
⫹
. Because of the lower redox-
potential of TyrD (18, 34, 53), it would act as a thermodynamic
sink for the electron hole (see Scheme 1). The second of these
mechanisms has been the generally accepted view. Our obser-
vation of a rapid oxidation of TyrD by P680
⫹
(t
1/2
⫽⬇190 ns)
in the D1-Y161F mutant strongly suggests that a direct rapid
route also exists in wild type, but it does not rule out the other
mechanism.
In this context, an observation of Boussac and Etienne (34) is
relevant. They showed in Tris-washed thylakoids with prere-
duced TyrD that the yield of TyrD
䡠
after one flash was almost
the same (⬇65%) in the absence and in the presence of
phenylenediamine, a rapid electron donor to TyrZ
䡠
. We have
performed similar experiments by using Mn-depleted PSII mem-
branes from spinach and confirmed these results (data not
shown). Because phenylenediamine donates to TyrZ
䡠
withat
1/2
⬍ 1 ms (34, 54), both electron donation rates from TyrD to
P680
⫹
and from P680
⫹
to TyrZ
䡠
(i.e., k
TyrD
and k
TyrZ
back
) must
be substantially faster than this; otherwise, the yield of TyrD
䡠
would decrease dramatically in the presence of phenlyenedia-
mine. Simulation of Scheme 1 using the program
GEPASI (see
Materials and Methods) in the presence of phenylenediamine
(i.e., k
Don
⫽ 700 s
⫺1
instead of 4 s
⫺1
) resulted in a formation of
65% TyrD
䡠
only when k
TyrD
was larger than 1 ⫻ 10
6
s
⫺1
(i.e., t
1/2
⬍ 700 ns). Formation of 85% TyrD
䡠
needs a k
TyrD
⬇3.7 ⫻ 10
6
(i.e., t
1/2
⫽⬇190 ns). These simulations provide an argument
that TyrD in wild-type PSII is oxidized in the sub-
s range at pH
8.5, and that in a fraction of centers, TyrD is oxidized directly by
P680
⫹
(mechanism 1) with a similar rate as in the D1-Y161F
mutant.
Before the present work, the reports that electron donation
from TyrD to P680
⫹
were slow had been rationalized by
assuming a high reorganization energy caused by the trapping of
the departing proton from TyrD by D2-His-189 (4). Thus, the
more rapid donation from TyrZ to P680
⫹
was taken as indicating
that no such charge was accumulated close to TyrZ
䡠
. This line of
thought in turn was used as an argument in favor of the
translocation of the proton away from TyrZ
䡠
, a feature that could
be taken as being in favor of certain aspects of the H-atom
(electron ⫹ proton) abstraction model for oxygen evolution (46).
The present work questions the need to invoke such a large
reorganization energy associated with the putative charge close
to TyrD
䡠
and thus weakens this argument for proton transloca-
tion away from the TyrZ
䡠
site.
The reassessment of the relative donation rates of TyrZ and
TyrD to P680
⫹
has repercussions on the question of the local-
ization of P680
⫹
(for instance, see ref. 45). There are four Chls
in the inner reaction center symmetrically positioned around a
C
2
axis (ref. 6; Protein Databank accession no. 1FE1). It was
suggested that the much faster donation rate of TyrZ compared
with TyrD could be explained if P680
⫹
were located on the D1
side (i.e., closer to TyrZ than TyrD) of the reaction center (for
review, see ref. 45). The similar donation rates found here
question this argument. According to Page et al. (55), who
describe a correlation between the distance and the rate of an
electron transfer in proteins, an electron donation rate of t
1/2
⫽
⬇190 ns gives an upper limit of 14 Å for the edge-to-edge
distance between the donor and acceptor. In the structural
model of PSII, only the Chls on the D2 side (P
D2
and Chl
D2
) are
located within this distance from TyrD (the distance P
D1
to TyrD
is ⬎18 Å edge to edge), and thus a direct electron transfer from
TyrD to P
D1
⫹
is unlikely to occur with a t
1/2
⫽ 190 ns. This
argument suggests that P680
⫹
is at least partially localized on the
Chl P
D2
(located on the D2 side). A recent study, performed in
the presence of TyrD
䡠
, suggested that the cation is predominantly
localized on P
D1
(D1 side; ref. 12). These results are not
contradictory, because the present studies suggest only partial
localization of P680
⫹
on P
D2
and only in the presence of reduced
TyrD (TyrD was oxidized in ref. 12). An effect of the oxidation
state of TyrD on the localization of P680
⫹
is conceivable (see
below).
In the present work, the measured electron-transfer rates to
P680
⫹
from TyrD and TyrZ were the same in Mn-depleted
PSII at pH 8.5 (t
1/2
⬇ 190 ns for the dominating phase). Thus,
when both TyrD and TyrZ are able to donate electrons to
P680
⫹
(e.g., first flash in wild type), the reduction rate P680
⫹
is expected to be twice as fast, i.e., t
1/2
⬇95 ns. The measured
rate, however, exhibited again a t
1/2
⬇190 ns (Fig. 2, ‘‘wt’:1.
flash’’). The most straightforward explanation for this finding
is that the redox-state of TyrD affects the TyrZ donation. The
following assumptions would allow for such a situation: (i) the
TyrZ donation rate is slower than t
1/2
⫽ 190 ns when both
tyrosines are reduced. This assumption seems reasonable given
the environment and potential of the two tyrosines (see above)
(ii) The presence of TyrD
䡠
accelerates the TyrZ to P680
⫹
electron donation. It seems reasonable that the positive charge
formed close to the TyrD site upon its oxidation (for example,
the protonated D2-His-189) has an electrostatic effect on
P680
⫹
(55), increasing its potential and perhaps shifting the
location of the positive charge (P680
⫹
) over to the D1 side
(P
D1
) (see ref. 56) where it is now supposed to be (12). A
similar role was suggested for D2-Arg-181 (57). Such a role for
TyrD
䡠
on the redox-potential of P680
⫹
, and thus on the kinetics
of TyrZ oxidation, could provide a mechanistic raison d’eˆtre
for TyrD in addition to its postulated redox role in photoac-
tivation and oxidation of the lower S states (17). Experimental
indications of an interaction of TyrD and P680 have been
reported: (i) TyrD oxidation induces an electrochromic shift in
the absorption spectrum of the P680 Soret band (58); (ii)
charge recombination between Q
A
⫺
and TyrZ
䡠
was slower in
wild-type⬘ PSII (in the presence of TyrD
䡠
) than in TyrD-less
mutants (59). These findings can be explained by an electro-
static effect of TyrD
䡠
H
⫹
-His on P680
⫹
that would raise the
redox potential of P680
⫹
and thus depopulate P680
⫹
(shift the
equilibrium of TyrZ
䡠
º P680
⫹
to the left). This depopulation
of P680
⫹
, which is an intermediate in the Q
A
⫺
兾TyrZ
䡠
recom-
bination (45), would slow down the recombination, as ob-
served. However, Hays et al. (28) compared the TyrZ donation
in wild-type (TyrD
䡠
present) and D2-Y161F mutant at differ-
ent pHs (7.0–9.0) and found no significant differences, indi-
cating that the presence of TyrD
䡠
does not affect the TyrZ
donation. Further comparative studies on the TyrD-less and
TyrZ-less mutant, as well as on the wild-type⬘, should clarify
whether the redox chemistry of TyrD has an electrostatic
effect on P680
⫹
.
14372
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.251382598 Faller et al.
We thank Dr. P. Se´tif for providing the software for the kinetic analyses;
A. P. Nguyen for isolating the thylakoid membranes from and maintaining
the wild-type⬘ and mutant cultures of Synechocystis 6803; Drs. F. Rappaport,
S. Un, T. Mattioli, A. Ivancich, C. Berthomieu, and R. Hienerwadel for
helpful discussions; Drs. B. Diner, E. Schlodder, P. Nixon, W. Coleman, F.
Rappaport, J. Lavergne, W. Vermaas, and D. Chisholm for providing us
with a preprint of ref. 12. The research was supported by a grant from the
Swiss National Science Foundation (to P.F.), a European Union Training
and Mobility of Researchers (TMR) Research Network grant (FMRX-
CT98–2014), a cooperation grant ‘‘metalloproteins Japan’’ from the Min-
istre des Affaires E
´
trangers (PO N°2000兾1577), and National Institutes of
Health Grant GM43496 (to R.J.D.).
1. Stubbe J. & van der Donk, W. A. (1998) Chem. Rev. (Washington, D.C.) 98,
705–762.
2. Diner, B. A. (2001) Biochim. Biophys. Acta 1503, 147–163.
3. Debus, R. J. (2001) Biochim. Biophys. Acta 1503, 164–186.
4. Babcock, G. T., Espe, M., Hoganson, C., Lydakis-Simantriris, N., McCracken,
J., Shi, W., Styring, S., Tommos, C. & Warncke, K. (1997) Acta Chem. Scand.
51, 533–540.
5. Rappaport, F. & Lavergne, J. (2001) Biochim. Biophys. Acta 1503, 246–259.
6. Zouni, A., Witt, H.-T., Kern, J., Fromme, P., Krauss, N., Saenger, W. & Orth,
P. (2001) Nature (London) 409, 739–743.
7. Debus, R. J., Barry, B. A., Babcock, J. T. & McIntosh, L. (1988) Proc. Natl.
Acad. Sci. USA 85, 427–430.
8. Vermass, W. F. J., Rutherford, A. W. & Hansson, O. (1988) Proc. Natl. Acad.
Sci. USA 85, 8477–8481.
9. Hirsh, D. J. & Brudvig, G. W. (1993) J. Phys. Chem. 97, 13216–13222.
10. Koulougliotis, D., Tang, X.-S., Diner, B. A.&&Brudvig,G.W.(1995)
Biochemistry 34, 2850–2856.
11. Astashkin, A. V., Kodera, Y. & Kawamori, A. (1994) Biochim. Biophys. Acta
1187, 89–93.
12. Diner, B. A., Schlodder, E., Nixon, P. J., Coleman, W. J., Rappaport, F.,
Lavergne, J., Vermaas, W. F. J. & Chisholm, D. A. (2001) Biochemistry 40,
9265–9281.
13. Rutherford, A. W. & Nitschke W. (1996) in Origin and Evolution of Biological
Energy Conversion, ed. Baltscheffsky, H. (VCH, New York), pp. 143–75.
14. Babcock, G. T. & Sauer, K. (1975) Biochim. Biophys. Acta 376, 315–328.
15. Brettel, K., Schlodder, E. & Witt, H. T. (1984) Biochim. Biophys. Acta 766,
403–415.
16. Babcock, G. T. & Sauer, K. (1973) Biochim. Biophys. Acta 325, 483–503.
17. Styring, S. & Rutherford, A. W. (1987) Biochemistry 26, 2401–2405.
18. Vass, I. & Styring, S. (1991) Biochemistry 30, 830–839.
19. Campell, K. A., Peloquin, J. M., Diner, B. A., Tang, X. S., Chisholm, D. A. &
Britt, R. D. (1997) J. Am. Chem. Soc. 119, 4787–4788.
20. Rodriguez, I. D., Chandrashekar, T. K. & Babcock, G. T. (1987) in Progress in
Photosynthesis Research, ed. Biggins, J. (Martinus Nijhoff, Dordrecht, The
Netherlands), Vol. I, pp. 471–474.
21. Tang, X.-S., Chisholm, D. A., Dismukes, G. C., Brudvig, G. W. & Diner, B. A.
(1993) Biochemistry 32, 13742–13748.
22. Tommos, C., McCracken, J., Styring, S. & Babcock, G. T. (1998) J. Am. Chem.
Soc. 120, 10441–10452.
23. Diner, B. A., Force, D. A., Randall, D. W. & Britt, R. D. (1998) Biochemistry
37, 17931–17943.
24. Ahlbrink, R., Haumann, M., Cherapanov, D., Bogenshausen, O., Mulkidjanian,
A. & Junge W. (1998) Biochemistry 37, 1131–1142.
25. Force, D. A., Randall, D. W., Britt, R. D., Tang, X. S. & Diner, B. A. (1995)
J. Am. Chem. Soc. 117, 10547–10554.
26. Tommos, C., Tang, X. S., Warncke, K., Hoganson, C. W., Styring, S.,
McCracken, J. & Diner, B. A. (1995) J. Am. Chem. Soc. 117, 10325–10335.
27. Hienerwadel, R., Boussac, A., Breton, J., Diner, B. & Berthomieu, C., (1997)
Biochemistry 36, 14712–14723.
28. Berthomieu, C., Hienerwadel, R., Boussac, A., Breton, J. & Diner, B. (1998)
Biochemistry 37, 10547–10554.
29. Conjeaud, H. & Mathis, P. (1980) Biochim. Biophys. Acta 590, 353–359.
30. Schlodder, E. & Meyer, B. (1987) Biochim. Biophys. Acta 890, 23–31.
31. Hays, A. A., Vassiliev, I. R., Golbeck, J. H. & Debus, R. J. (1999) Biochemistry
38, 11851–11865.
32. Hays, A. A., Vassiliev, I. R., Golbeck, J. H. & Debus, R. J. (1998) Biochemistry
37, 11352–11365.
33. Buser, C. A., Thompson, L. K., Diner, B. A. & Brudvig, G. W. (1990)
Biochemistry 29, 8977–8985.
34. Boussac, A. & Etienne, A. L. (1982) Biochem. Biophys. Res. Commun. 109,
1200–1205.
35. Boussac, A. & Etienne, A. L. (1982) FEBS Lett. 148, 113–116.
36. Debus, R. J., Barry, B. A., Sithole, I., Babcock, G. T. & McIntosh, L. (1988)
Biochemistry 27, 9071–9074.
37. Debus, R. J., Nguyen, A. P. & Conway, A. B. (1990) in Current Research in
Photosynthesis, ed. Baltscheffsky, M. (Kluwer, Dordrecht, The Netherlands),
Vol. I, pp. 829–832.
38. Debus, R. J., Campbell, K. A., Gregor, W., Li, Z.-L., Burnap, R. L. & Britt,
R. D. (2001) Biochemistry 40, 3690–3699.
39. Mendes, P. (1997) Trends Biochem. Sci. 22, 361–363.
40. Yerkes, C. T., Babcock, G. T. & Crofts, A. R. (1983) Biochim. Biophys. Acta
158, 359–363.
41. Rappaport, F. & Lavergne, J. (1997) Biochemistry 36, 15294–15303.
42. Visser, J. W. M., Rijgersberg, C. P. & Gast, P. (1977) Biochim. Biophys. Acta
460, 36–46.
43. Hanley, J., Deligiannakis, Y., Pascal, A., Faller, P. & Rutherford, A. W. (1999)
Biochemistry 38, 8189–8185.
44. Blankenship, R. E., Babcock, G. T., Warden, J. T. & Sauer, K. (1975) FEBS
Lett. 51, 287–293.
45. Diner, B. A. & Babcock, G. T. (1996) in Oxygenic Photosynthesis: The Light
Reactions, eds. Ort, D. R. & Yocum, C. F. (Kluwer, Dordrecht), pp. 213–247.
46. Tommos, C. & Babcock, G. T. (2000) Biochim. Biophys. Acta 1458, 199–219.
47. Debus, R. J., Campbell, K.A., Pham, D. P., Hays, A.-M. A. & Britt, R. D. (2000)
Biochemistry 39, 627–6287.
48. Svensson, B., Vass, I., Cedergren, E. & Styring, S. (1990) EMBO J. 9,
2051–2059.
49. Ruffle, S., Donnelly, D., Blundell, T. L. & Nugent, J. H. A. (1992) Photosynth.
Res. 34, 287–300.
50. Chu, H. A., Nguyen, A. P. & Debus, R. J. (1995) Biochemistry 34, 5839–5858.
51. Svensson, B., Etchebest, C., Tuffery, P., van Kan, P., Smith, J. & Styring, S.
(1996) Biochemistry 35, 14486–14502.
52. Un, S., Tang, X. S. & Diner, B. A. (1996) Biochemistry 35, 679–684.
53. Boussac, A. & Etienne, A. L. (1984) Biochim. Biophys. Acta 766, 576–584.
54. Diner, B. A. (1977) in Proceedings of the 4
th
International Congress on
Photosynthesis, ed. Hall, D. O., Coombs, J. & Goodwin, T. W. (Biochemical
Society, London), pp. 359–372.
55. Page, C. C., Moser, C. C., Chen, X. & Dutton, P. L. (1999) Nature (London)
402, 47–52.
56. Nugent, J. H. A., Bratt, P. J., Evans, M. C. W., MacLachlan, D. J., Rigby, S. E. J.,
Ruffle, S. V. & Turconi, S. (1994) Biochem. Soc. Trans. 22, 327–331.
57. Mulkidjanian, A. (1999) Biochim. Biophys. Acta 1410, 1–6.
58. Diner, B., Tang, X-S., Zheng, M., Dismukes, G. C., AnnForce, D., Randall,
D. W. & Britt, R. D. (1995) in Photosynthesis: From Light to Biosphere, ed.
Mathis, P. (Kluwer, Dordrecht, The Netherlands) pp. 229–234.
59. Boerner, R. J., Bixby, K. A., Nguyen, A. P., Noren, G. H., Debus, R. J. & Barry,
B. A. (1993) J. Biol. Chem. 268, 1817–1823.
Faller et al. PNAS
兩
December 4, 2001
兩
vol. 98
兩
no. 25
兩
14373
BIOPHYSICS