Excited-state relation dynamics of a PYP chromophore model in solution: Influence of the thioester group

Article abstract of DOI:10.1016/S0009-2614(02)01480-X

Cis-trans photoisomerization of a photoactive yellow protein chromophore model, the deprotonated trans S-phenyl thio-p-hydroxycinnamate, is studied in aqueous solution by subpicosecond transient absorption and gain spectroscopy. The excited-state deactiva

Full text of DOI:10.1016/S0009-2614(02)01480-X

Chemical Physics Letters 365 (2002) 285–291
www.elsevier.com/locate/cplett
Excited-state relaxation dynamics of a PYP
chromophore model in solution: influence of the
thioester group
*
Pascale Changenet-Barret , Agathe Espagne, Nathalie Katsonis,
Sandrine Charier, Jean-Bernard Baudin, Ludovic Jullien, Pascal Plaza,
Monique M. Martin
Ecole Normale Supeꢀrieure, Deꢀpartement de Chimie, UMR-CNRS 8640 PASTEUR, 24 rue Lhomond, 75231 Paris Cedex 05, France
Received 5 August 2002; in final form 12 September 2002
Abstract
Cis–trans photoisomerization of a photoactive yellow protein chromophore model, the deprotonated trans S-phenyl
thio-p-hydroxycinnamate, is studied in aqueous solution by subpicosecond transient absorption and gain spectroscopy.
The excited-state deactivation is found to involve the formation, in 1.7 ps, of an intermediate state which decays in 2.8
ps. A persistent bleaching signal is observed at longer times indicating that the excited state not only relaxes to the
ground state but also partly forms a stable photoproduct, possibly the cis isomer. This behavior is analogous to that of
the native photoactive yellow protein.
ꢀ 2002 Elsevier Science B.V. All rights reserved.
1. Introduction
protein [2–5]. Upon photoexcitation, PYP un-
dergoes a photocycle characterized by a series of
intermediates formed on a timescale spanning
from several hundred femtoseconds to seconds
[6–18]. Like for rhodopsins, the trans to cis
isomerization of the chromophore was shown to
be the first overall step of the PYP photocycle
[19]. Along this step, an early intermediate with a
highly distorted geometry, involving the flipping
of the thioester group and a partial isomerization
of the vinyl bond, was reported to be trapped at
low temperature and identified as a precursor (I0)
of the cis isomer ðI1Þ [20]. In solution, ultrafast
events on the subpicosecond and picosecond
The photoactive yellow protein (PYP) is the
blue-light photoreceptor that is thought to me-
diate the negative photoaxis of the halophilic
purple bacteria Ectothiorhodospira halophila [1].
Its chromophore is a deprotonated trans-couma-
ric acid (pCA Fig. 1a) covalently linked, via a
thioester bond, to the unique cystein group of the
* Corresponding author. Fax: +33-1-44-32-33-25.
E-mail address: pascale.changenet@ens.fr (P. Changenet-
Barret).
0009-2614/02/$ - see front matter ꢀ 2002 Elsevier Science B.V. All rights reserved.
PII: S0009-2614(02)01480-X

286
P. Changenet-Barret et al. / Chemical Physics Letters 365 (2002) 285–291
2. Materials and methods
timescales were assigned to the formation of I0
[14,21].
2.1. Materials
In a first attempt to approach the intrinsic
properties of the PYP chromophore, subpicosec-
ond transient absorption spectroscopy was carried
out on the fully deprotonated trans-coumaric acid
(pCA2À) in solution (Fig. 1b). Cis-pCA2À was
found to be formed with a time constant of 10 ps
without any detectable intermediate [22]. This re-
action is quite different from that in the native
PYP but its kinetics is comparable to the excited-
state decay of the denatured PYP [21]. These ob-
servations suggest that pCA2À is a good model
compound of the PYP chromophore, despite the
absence of the thioester group. However, two
major differences remain. The absorption maxi-
mum of pCA2À is about 50-nm blue-shifted from
that of the denatured PYP [22]. Additionally, time-
resolved fluorescence experiments carried out on
the denatured PYP gave evidence that a charge
redistribution takes place in the excited state [21].
There was no evidence for such process in pCA2À
[23].
In this Letter, we report an experimental study
of the excited-state relaxation dynamics of a
thioester derivative of trans-pCA, the trans
S-phenyl thio-p-hydroxycinnamate (pCT, sket-
ched in Fig. 1c) in a basic aqueous solution, by
subpicosecond transient absorption spectroscopy
[24]. The compound has been synthesized to
provide a better representation of the PYP
chromophore than pCA, by including the thio-
ester function.
Trans-pCT was synthesized according to the
following procedure [25]. Oxalyl chloride (1.5 ml,
17 mmol) was added dropwise under N2 at 0 ꢁC to
a solution containing anhydrous dimethylforma-
mide (0.46 ml, 6 mmol) in CH2Cl2 (9 ml). After gas
evolution, the mixture was stirred for 1 h at 0 ꢁC.
The solvent was evaporated to give a white powder
that was dissolved in tetrahydrofuran (THF, 12
ml). A mixture of pCA (885 mg, 5 mmol) and
pyridine (0.49 ml, 6 mmol) in THF (9 ml) was
added dropwise to preceding solution. The reac-
tion mixture was stirred for 1 h at )30 ꢁC. Then a
solution of C6H5SLi 0.5 M (from C4H9Li 1.6 M in
hexane (2.5 ml, 4 mmol) and thiophenol (0.41 ml,
4 mmol) in THF (10 ml) under N2 at 0 ꢁC) was
added dropwise at )30 ꢁC. The mixture was stirred
for 2 h at )15 ꢁC, then hydrolyzed and extracted
with diethylether. The organic layer was washed
with NaHCO3 10% up to pH 6 and dried
(MgSO4). The solvent was removed under high
vacuum and the residue was chromatographed on
a silica gel column (eluent: acetone). The solid was
recrystallized from CH2Cl2 to give trans-pCT as
1
yellow crystals (531 mg, 2 mmol, 38%). H NMR
(400 MHz, CDCl3): d(ppm) ¼ 7.67, (d 3J ¼ 15:7
3
Hz, 1H), 7.45–7.55 (m, 7H), 6.87 (d, J ¼ 8:6 Hz,
3
2H), 6.70 (d, J ¼ 15:7 Hz, 1H), 5.49 (s, 1H); MS
(CI, NH3Þ: m=z ¼ 257 [M þ Hþ], calcd. for
C15H13O2S m=z ¼ 257:33; Anal. Calcd. for
Fig. 1. Structures of (a) the native PYP chromophore, (b) trans-pCA and (c) trans-pCT.

P. Changenet-Barret et al. / Chemical Physics Letters 365 (2002) 285–291
287
C15H12O2S: C, 70.29; H, 4.72; Found: C, 70.20; H,
4.82.
sum of exponentials convoluted with a Gaussian
representing the experimental response function.
The FWHM was found to be 1:5 ꢀ 0:1 ps.
Steady-state spectra were recorded with a dou-
ble-beam UV–Vis spectrophotometer (SAFAS
UVmc2) and a spectrofluorimeter (Jobin Yvon
Fluoromax 3).
Samples of pCTÀ were prepared in a borate
Britton–Robinson buffer [26] at pH 10.6 (ionic
strength 0.1 M, HPLC grade water) with an ab-
sorbance close to 1 per mm at the excitation
wavelength for pump–probe experiments. The
pKa of pCT was determined to be 8:85 ꢀ 0:05 by
means of a spectrometric titration. Hydrolysis of
the thioester bond was found to occur at basic pH,
leading to the formation of pCA2À. Since the hy-
drolysis rate increases strongly with the pH, we
fixed it at 10.6 to obtain both a complete depro-
tonation of pCT and a good chemical stability of
the solution over a few hours.
3. Results and discussion
3.1. Steady-state spectroscopy
Absorption and spontaneous emission spectra
of trans-pCTÀ in a buffer solution at pH 10.6 are
shown in Fig. 2. The absorption maximum at
395 nm (e ¼ 29; 000 l molÀ1 cmÀ1) is comparable
to that of the denatured protein (kmax ꢂ 400 nm)
[21] but 60-nm red-shifted from the pCA2À ab-
sorption peak (kmax ¼ 333 nm) [22]. Car Parrinello
ab initio molecular dynamics on pCTÀ shows that
the substitution of the oxygen atom by a sulfur
atom induces a red-shift of the electronic transi-
tion and that the p electron delocalization does not
extend to the phenyl ring attached to the sulfur
atom [28]. Thus, the thioester group most signifi-
cantly contributes to the red-shift of the pCTÀ
absorption. For comparison, the native PYP ab-
sorption spectrum undergoes an additional 50-nm
red-shift (PYP, kmax ¼ 446 nm), which should be
2.2. Pump–probe measurements
Transient absorption and gain experiments
were performed by the pump–probe technique
using a subpicosecond homemade dye laser system
described elsewhere [27]. Two intense subpicosec-
ond beams were generated at 570 and 428 nm.
Typically, 50 lJ of the 428 nm beam were focused
to a diameter of 2 mm in the sample cell and used
as the pump. The probe beam was a white light
continuum generated in a 1-cm water cell from the
300-lJ pulses at 570 nm. The continuum was then
split in two; one part was used as a reference sig-
nal, while the other part passed through the sam-
ple with a diameter close to 1 mm. The reference
and signal beams were sent to the entrance slit of a
polychromator (Jobin-Yvon Spex 270M, entrance
slit 64 lm) through 600-lm optical fibers and were
simultaneously analyzed by a 128 ꢁ 1024 CCD
camera (Princeton Instruments). Pump and probe
beams crossed in the sample, held in a 1-mm cell,
at an angle of 10ꢁ and were set at the magic po-
larization angle. The sample solution was recircu-
lated at room temperature. The absence of
significant photodegradation was checked by
measuring the absorption spectrum of the sample
before and after each pump–probe experiment.
The time-resolved differential absorbance spec-
tra (DA) were averaged over 500 shots and cor-
rected from the group velocity dispersion in the
probe beam. The kinetics at selected wavelengths
were fitted (Levenberg–Marquardt algorithm) to a
Fig. 2. Absorption and emission spectra of deprotonated trans-
pCT in a borate buffer at pH 10.6, at room temperature.

288
P. Changenet-Barret et al. / Chemical Physics Letters 365 (2002) 285–291
attributed to the specific interaction with the pro-
tein environment.
Finally, it is worth noting that quantum chemical
calculations showed that the cis configuration (I1)
of the chromophore in the native PYP is 25 kcal/
mol higher in energy than that of the trans con-
figuration [28]. In contrast, the cis configuration of
pCA2À is only 5 kcal/mol higher in energy than the
trans. This difference was attributed to the local
constraints of the protein environment that de-
stabilize the cis configuration. It comes out from
our steady-state photolysis of pCTÀ that the
presence of the thioester group itself might be, in
part, responsible of the destabilization of the cis
isomer.
The fluorescence spectrum of trans-pCTÀ in
water at pH 10.6 is centered around 500 nm. For
comparison the fluorescence spectra of the dena-
tured and native PYP are located in the same re-
gion, respectively, at 496 and 494 nm [21]. The
pCTÀ fluorescence spectrum exhibits positive sol-
vatochromism. In a low polar solvent such as
decanol ðe ¼ 7:2Þ, the fluorescence spectrum is
20-nm blue-shifted from that measured in water
ðe ¼ 78:3Þ indicating that an important charge re-
distribution occurs in the excited state. Such a
polarity effect was not observed for pCA2À [23].
On the other hand, it should be noted that the
fluorescence spectrum of pCTÀ in water is not the
mirror image of the absorption spectrum, the lat-
ter being about 2000 cmÀ1 broader. It is worth
mentioning that pCA2À and its amide analog,
p-hydroxycinnamide (pCMÀ), in water exhibit the
same feature, the absorption spectrum being 2800
cmÀ1 broader than the emission spectrum in both
cases. The absorption band of pCA2À and pCMÀ
in water have a noticeable shoulder on its blue side
and in dimethylformamide pCMÀ exhibits two
fully separated bands [23]. This is a strong indi-
cation that two electronic bands lie under the main
absorption band of the PYP chromophore analogs
in water. The presence of two electronic excited
states lying under the main absorption band has
also been discussed for the native PYP [14,16,18].
3.3. Time-resolved transient absorption spectra
Fig. 3 shows the time-resolved differential ab-
sorbance spectra ðDAÞ of trans-pCTÀ in water at
pH 10.6, after excitation at 428 nm (resolution: 1.5
3.2. Steady-state photolysis
Unlike for pCA2À [22], in a basic aqueous so-
lution steady-state photolysis experiments do not
show the formation of a stable cis isomer of pCTÀ.
The absorption spectrum recorded after a few
hours irradiation at 430 nm, remains almost un-
changed. One might think that an efficient back
reaction to the trans isomer occurs, possibly due to
the high energy of the cis isomer. We nevertheless
found that the steady-state photolysis of the amide
analog of pCTÀ, N-phenyl p-hydroxycinnamide
(pCMPhÀ) leads to a stable cis isomer [23]. This
rules out the steric hindrance of the phenyl ring
attached to the thioester group as the cause for a
possible destabilization of the cis isomer in pCTÀ.
Fig. 3. Time-resolved differential absorption spectra DAðk; tÞ of
trans-pCT in a borate buffer at pH 10.6, at room temperature,
after excitation at 428 nm, for different pump–probe delays
between (a) )10 and +1 ps, (b) 1.5 and 4 ps, (c) 4.5 and 40 ps.
The scattered pump light signal around 428 nm was masked.

P. Changenet-Barret et al. / Chemical Physics Letters 365 (2002) 285–291
289
ps). The spectral changes are characterized by
three different phases:
The kinetics recorded between )10 and +40 ps
in the three regions of the DA spectra are repre-
sented in Fig. 4. All kinetics have been fitted si-
multaneously to a biexponential function followed
by a small plateau, convoluted with the apparatus
response function. The plateau accounts for the
observed persistent bleaching signal. The UV ESA,
the bleaching and the SE bands decay with both
1.7 and 2.8 ps time constants. The 450-nm band
appears to develop in 1.7 ps and to decay in 2.8 ps.
• During the excitation pulse (Fig. 3a), three
bands are present. The positive band below
400 nm is assigned to excited-state absorption
(ESA). This UV ESA band is quite noisy be-
cause in that region the probe light is strongly
absorbed by the sample but its presence is con-
sistent in various experiments. Between 400
and 450 nm, a small negative band due to the
ground-state bleaching peaks at 420 nm. This
band is slightly red-shifted compared to the stea-
dy-state absorption due to the overlapping UV
ESA band that modifies its shape. Between 450
and 650 nm, in the steady-state fluorescence re-
gion, the DA spectra are dominated by stimu-
lated emission (SE). The continuous red-shift
up to about 5 nm of the SE band is attributed
to solvation dynamics in the excited-state. The
charge-redistributed character of the excited
state is indeed expected from the steady-state
fluorescence solvatochromism (see Section 3.1).
In water, the average solvation time lies in the
subpicosecond regime [29,30]. With the present
experimental resolution, only the slower compo-
nent (close to 1 ps and accounting for 35% of the
total solvation dynamics) can be detected. This
could explain the small amplitude of the red-
shift.
3.4. Nature of the 450 nm transient absorption band
Since the bleaching, SE and 450-nm ESA bands
overlap in the 400–500 nm region, we examined
the possibility that the delayed rise at 450 nm
could be due to the solvation-induced red-shift of
the SE band. In that picture, the SE and the 450-
nm ESA bands would come from the same excited
state and would decay with the same time con-
stant, roughly 2.8 ps. Simulating a red-shift of the
SE band of 5 nm at a rate constant of 1 psÀ1, we
calculated that the amplitude of the 450-nm band
should be dominant (i.e., DA > 0Þ at times below
1 ps. Since we did not observe such a behavior, we
consider another explanation which involves the
• For probe delays between 1.5 and 4 ps (Fig.
3b), the former three bands decay simulta-
neously while a transient absorption band rises
at 450 nm, between the bleaching and the SE
bands. This rising band leads to an apparent
red-shift of the SE band and a decrease of its
bandwidth. On the other hand, the decrease
of the UV ESA band leads to an apparent
blue-shift of the bleaching band and to the ob-
servation of an isosbestic point at 400 nm
ðDA < 0Þ.
• Finally, at longer probe delays (>4 ps, Fig. 3c),
the SE band and the 450-nm ESA band decrease
simultaneously in a few picoseconds, with an
isosbestic point (DA ¼ 0) around 480 nm. In
the ground-state absorption region, a small per-
sistent bleaching signal remains after 10 ps but
the isosbestic point at 400 nm is no longer ob-
served.
Fig. 4. DAðtÞ kinetics measured at 420, 450, 510 nm for trans-
pCT in a borate buffer at pH 10.6, after excitation at 428 nm.
The data were fitted simultaneously with a multiexponential
function convoluted by a 1.5-ps (FWHM) Gaussian function
representing the experimental response function. The fits led to
two short components, respectively, 1.7 and 2.8 ps followed by
a small plateau attributed to the persistent bleaching signal
observed after 10 ps. The fits are represented by the solid lines.

290
P. Changenet-Barret et al. / Chemical Physics Letters 365 (2002) 285–291
and proceeds through an intermediate state prior
to the formation of a long-lived photoproduct.
This observation stresses the important role of the
thioester group on the excited-state deactivation of
the PYP chromophore.
Interestingly, the excited-state relaxation kinet-
ics of pCTÀ differs markedly from that of the de-
natured PYP. The decay rate of the latter in
solution is in fact comparable to that of pCA2À
[22]. This was interpreted as a strong indication
that an isomerization around the vinyl bond also
takes place in the denatured PYP [22]. The differ-
ence between the decay rates of pCTÀ and the
denatured PYP remain then obscure since both
compounds contain the thioester group. However,
pCTÀ provides a better representation of the
charge distribution of the native PYP chromo-
phore than pCA2À. As matter of fact, a dynamical
Stokes shift of the fluorescence spectrum is ob-
served for pCTÀ, indicating that a photoinduced
charge redistribution occurs like in the denatured
PYP [21]. For pCA2À in water and alcohols, no
solvation dynamics have been detected indicating
little electronic redistribution in the excited state,
likely due to the presence of the charged carboxylic
group [23].
It is worth noting that the transient absorption
spectroscopy of pCTÀ is apparently similar to that
of the native PYP. In the protein, a transient ab-
sorption band was observed to appear in 3 ps be-
tween the bleaching and the stimulated emission
bands and associated to the formation of an in-
termediate state, possibly I0, prior to the cis isomer
I1. This comparison might nevertheless be fortu-
itous because the I0 intermediate of PYP is a
ground state and that of pCTÀ is, from the present
work, an excited state in equilibrium with the
fluorescent state.
Fig. 5. Four electronic state model proposed to explain the
evolution of DA spectra of pCTÀ. The transient state X is
formed in 1.7 ps from S1 and decays non-radiatively toward Y
in 2.8 ps after equilibration with S1.
formation of a transient state from the first emit-
ting state.
In a first attempt to interpret the time evolution
of the DA spectra, we propose a four-state kinetic
scheme (Fig. 5). The quenching of the stimulated
emission is attributed to the formation of a tran-
sient non-emissive state ðreX ꢂ 0Þ which absorbs at
450 nm. Since, after 4 ps, both transient absorp-
tion and SE bands decay identically, a fast equi-
librium between the fluorescent state and the non-
emissive state must be reached. Although, one
does not know the nature of this transient dark
state, the present results bring a definitive evidence
that it is a non-emissive excited state. In addition,
the small persistent bleaching observed after 10 ps
gives evidence that there is no complete ground-
state repopulation after the decay of the excited
states. This is an indication of the formation of a
long-lived photoproduct. The most likely photo-
product should be the cis isomer of pCTÀ.
3.5. Comparison with pCA2À and PYP
The intrinsic relaxation of pCTÀ in water is
different from that of pCA2À. The transient ab-
sorption experiments carried out on pCA2À gave
evidence for the formation of a cis isomer in 10 ps,
with the same rate as the excited-state decay [22].
No intermediate state was detected as in the case
of the well-documented trans-stilbene isomeriza-
tion process [31]. The excited-state relaxation of
pCTÀ in water is much faster than that of pCA2À
4. Conclusion
The deprotonated trans S-phenyl thio-p-hy-
droxycinnamate (pCT) has been studied in a basic
aqueous solution by subpicosecond transient ab-
sorption and gain spectroscopy. The excited-state
relaxation of pCTÀ is found to involve the for-
mation in 1.7 ps of a dark excited-state that decays

P. Changenet-Barret et al. / Chemical Physics Letters 365 (2002) 285–291
291
[11] A. Baltuska, I.H.M. van Stokkum, A. Kroon, R. Mons-
houwer, K.J. Hellingwerf, R. van Grondelle, Chem. Phys.
Lett. 270 (1997) 263.
in 2.8 ps, partly forming a long-lived (>40 ps)
photoproduct. Our experiment provides evidence
for a rapid equilibrium between the initial fluo-
rescent state and the dark excited state. The pho-
toproduct has not been characterized yet but is
thought to be the cis isomer. Surprisingly, the ex-
cited-state dynamics of pCTÀ differs markedly
from that of the denatured PYP but is close to that
of the native PYP. The present study suggests that
both the intrinsic properties of the chromophore,
especially the thioester group, and the effect of the
protein environment account for the primary
processes in the native PYP.
[12] H. Chosrowjan, N. Mataga, N. Nakashima, Y. Imamoto,
F. Tokunaga, Chem. Phys. Lett. 270 (1997) 267.
[13] P. Changenet, H. Zhang, M.J. van der Meer, K.J.
Hellingwerf, M. Glasbeek, Chem. Phys. Lett. 282 (1998)
276.
[14] S. Devanathan, A. Pacheco, L. Ujj, M. Cusanovich, G.
Tollin, S. Lin, N. Woodbury, Biophys. J. 77 (1999) 1017.
[15] H. Hanada, Y. Kanematsu, S. Kinoshita, M. Kumauchi, J.
Sasaki, F. Tokunaga, J. Lumin. 94 (2001) 593.
[16] Y. Imamoto, M. Kataoka, F. Tokunaga, T. Asahi, H.
Masuhara, Biochemistry 40 (2001) 6047.
[17] Z. Ren, B. Perman, V. Srajer, T. Teng, C. Pradervand, D.
Bourgeois, F. Schotte, T. Ursby, R. Kort, M. Wulff, K.
Moffat, Biochemistry 40 (2001) 13788.
[18] T. Gensch, C.C. Gradinaru, I.H.M. van Stokkum, J.
Hendricks, K.J. Hellingwerf, R. van Grondelle, Chem.
Phys. Lett. 356 (2002) 347.
References
[19] R. Kort, H. Vonk, X. Xu, W.D. Hoff, W. Crielaard, K.J.
Hellingwerf, FEBS Lett. 382 (1996) 73.
[1] W.W. Sprenger, W.D. Hoff, J.P. Armitage, K.J. Helling-
werf, J. Bacteriol. 175 (1993) 3096.
[20] U.K. Genick, S.M. Soltis, P. Kuhn, I.L. Canestrelli, E.D.
Getzoff, Nature 392 (1998) 206.
[2] M. Baca, G.E.O. Borgstahl, M. Boissinot, P.M. Burke,
D.R. Williams, K.A. Slater, E.D. Getzoff, Biochemistry 33
(1994) 14369.
[21] H. Chosrowjan, N. Mataga, Y. Shibata, Y. Imamoto, F.
Tokunaga, J. Phys. Chem. B 102 (1998) 7695.
[22] P. Changenet-Barret, P. Plaza, M. Martin, Chem. Phys.
Lett. 336 (2001) 439.
[3] G.E.O. Borgstahl, D.R. Williams, E.D. Getzoff, Biochem-
istry 34 (1995) 6278.
[4] M. Kim, R.A. Mathies, W.D. Hoff, K. Hellingwerf,
Biochemistry 34 (1995) 12669.
[23] P. Changenet-Barret, A. Espagne, S. Charier, J.-B. Baudin,
L. Jullien, P. Plaza, M.M. Martin, in preparation.
[24] P. Changenet-Barret, N. Katsonis, J.-B. Baudin, L. Jullien,
P. Plaza, M.M. Martin, Femtochemistry V, Toledo (2001)
Poster No. PII44.
[5] W.D. Hoff, B. Devreese, R. Fokkens, I.M. Nugteren-
Roodzant, J. Van Beeumen, N. Nibbering, K.J. Helling-
werf, Biochemistry 35 (1996) 1274.
[6] T.E. Meyer, E. Yakali, M.A. Cusanovich, G. Tollin,
Biochemistry 26 (1987) 418.
[25] E. Duran, H. Duran, L. Cazaux, L. Gorrichon, P. Tisnes,
F. Sarni, Bull. Soc. Chim. Fr. (1987) 143.
[7] W.D. Hoff, I.H.M. van Stokkum, H.J. van Ramesdonk,
M.E. van Brederode, A.M. Brouwer, J.C. Fitch, T.E.
Meyer, R. van Grondelle, K.J. Hellingwerf, Biophys. J. 67
(1994) 1691.
[26] C. Frugoni, Gazz. Chim. Ital. 87 (1957) 403.
[27] N. Dai Hung, P. Plaza, M.M. Martin, Y.H. Meyer, Appl.
Opt. 31 (1992) 7046.
€
[28] A. Sergi, M. Gruning, M. Ferrario, F. Buda, J. Phys.
Chem. B 105 (2001) 4386.
[8] A. Xie, W.D. Hoff, A.R. Kroon, K.J. Hellingwerf,
Biochemistry 35 (1996) 14671.
[29] W. Jarzeba, G.C. Walker, A.E. Johnson, P.F. Barbara,
Chem. Phys. 152 (1991) 57.
[9] U.K. Genick, G.E.O. Borgstahl, K. Ng, Z. Ren, C. Prad-
ervand, P. Burke, V. Srajer, T. Teng, W. Schildkamp, D.E.
McRee, K. Moffat, E. Getzoff, Science 275 (1997) 1471.
[10] B. Perman, V. Srajer, Z. Ren, T. Teng, C. Pradervand, T.
Ursby, D. Bourgeois, F. Schotte, M. Wulff, R Kort, K.
Hellingwerf, K. Moffat, Science 279 (1998) 1946.
[30] R. Jimenez, G.R. Fleming, P.V. Kumar, M. Maroncelli,
Nature 369 (1994) 471.
[31] R.J. Sension, S.T. Repinec, A.Z. Szarka, R.M. Hochst-
rasser, J. Chem. Phys. 98 (1993) 6291.