Dark Excited States of Carotenoid Regulated by Bacteriochlorophyll in

Mar 9, 2011 - Dark Excited States of Carotenoid Regulated by Bacteriochlorophyll in ... ratus of plants and bacteria, they act mainly as light harvest...
0 downloads 0 Views 2MB Size
Download PDF
ARTICLE pubs.acs.org/JPCB

Dark Excited States of Carotenoid Regulated by Bacteriochlorophyll in Photosynthetic Light Harvesting Ryosuke Nakamura,*,†,‡,^ Katsunori Nakagawa,§,‡ Mamoru Nango,§,‡,|| Hideki Hashimoto,||,‡ and Masayuki Yoshizawa†,‡ †

Department of Physics, Graduate School of Science, Tohoku University, 6-3 Aramaki-Aza-Aoba, Aoba-ku, Sendai 980-8578, Japan JST, CREST, 4-1-8 Hon-chou, Kawaguchi, Saitama 332-0012, Japan § Department of Life and Materials Engineering, Graduate School of Engineering, Nagoya Institute of Technology, Gokiso-cho, Showaku, Nagoya 466-8555, Japan Department of Physics, Graduate School of Science, Osaka City University, 3-3-138 Sugimoto, Sumiyoshi-ku, Osaka 558-8585, Japan

)

‡

bS Supporting Information ABSTRACT: In photosynthesis, carotenoids play important roles in light harvesting (LH) and photoprotective functions, which have been described mainly in terms of two singlet excited states of carotenoids: S1 and S2. In addition to the “dark” S1 state, another dark state, S*, was recently identified and its involvement in photosynthetic functions was determined. However, there is no consistent picture concerning its nature or the mechanism of its formation. One particularly anomalous behavior obtained from femtosecond transient absorption (TA) spectroscopy is that the S*/S1 population ratio depends on the excitation intensity. Here, we focus on the effect of nearby bacteriochlorophyll (BChl) on the relaxation dynamics of carotenoid in the LH complex. We performed femtosecond TA spectroscopy combined with pre-excitation of BChl in the reconstituted LH1 complex from Rhodospirillum rubrum S1. We observed that the energy flow from S1, including its vibrationally excited hot states, to S* occurs only when nearby BChl is excited into Qy, resulting in an increase in S*/S1. We also examined the excitation-intensity dependence of S*/S1 by conventional TA spectroscopy. A comparison between the pre-excitation effect and excitation-intensity dependence shows a strong correlation of S*/ S1 with the number of BChls excited into Qy. In addition, we observed an increase in triplet formation as the S* population increased, indicating that S* is an electronic excited state that is the precursor to triplet formation. Our findings provide an explanation for observed spectroscopic features, including the excitation-intensity dependences debated so far, and offer new insights into energy deactivation mechanisms inherent in the LH antenna.

’ INTRODUCTION Carotenoids are a class of natural pigments that play significant roles in various biological systems. In the photosynthetic apparatus of plants and bacteria, they act mainly as light harvesting (LH) and photoprotecting pigments. Carotenoids absorb light in the blue-green region where (bacterio)chlorophyll (BChl) molecules do not. That energy is transferred to nearby BChls with high efficiency, resulting in enhancement of the absorption cross section for photosynthesis.1-3 In the photoprotective function, carotenoids efficiently quench the BChl triplet state to prevent sensitized generation of reactive singlet oxygen.2 These functions of carotenoids can be qualitatively understood in terms of their electronic structure. The generally accepted energy-level scheme of carotenoid consists of three singlet electronic states: the ground state S0 with Ag- symmetry and two excited states of S1 and S2 with Ag- and Buþ symmetry, respectively, in the idealized C2h point group.1 The specific strong absorption in the visible region is due to the S0 f S2 r 2011 American Chemical Society

transition, while the one-photon transition of S0 f S1 is dipoleforbidden. Therefore, the S1 state is referred to as the “dark” state. The energy transfer between the S1 state of carotenoid and the BChl states has been investigated intensively and the involvement of the S1 state in LH has been revealed.4 Besides the S1 and S2 excited electronic states, an additional dark state, termed S*, has been identified, and has attracted considerable interest.3,5 The S* state was first identified by Gradinaru et al. in femtosecond transient absorption (TA) data of spirilloxanthin in solution and in the LH1 complex from Rhodospirillum (Rs.) rubrum S1.6 The S* state is characterized by a spectral band that appears on the high-energy side of the S1 f Sn transition and usually has a longer decay time than the S1 state. The S* state is the precursor on the reaction pathway toward Received: December 9, 2010 Revised: February 13, 2011 Published: March 09, 2011 3233

dx.doi.org/10.1021/jp111718k | J. Phys. Chem. B 2011, 115, 3233–3239

The Journal of Physical Chemistry B triplet formation and plays a critical role in efficient excitation energy deactivation in the LH1 complex. It has been reported that the S* state is actively involved in the energy transfer process to BChl in LH2 complexes and contributes to efficient LH.7,8 These observations support the hypothesis that the S* state is an electronic excited state with energy close to that of the S1 state and is most likely formed in the relaxation from the S2 state, in parallel with the S1 state formation originally proposed by Gradinaru et al.6 It has also been proposed that the S* state is associated with the S1 state of a twisted carotenoid conformation based on a series of studies involving femtosecond TA experiments and quantum chemical calculations.9-11 However, consistent explanation of its origin has not been available until now. One of the most striking observations that generates considerable debate about the S* state formation in the LH complexes is that the S* and S1 populations have different excitationintensity dependences.12-14 This behavior has been observed in various kinds of carotenoids in LH complexes, but there is no corresponding observation for carotenoids free in solution.12,13 This experimental result seems to require an alternative model for the S* state formation. To account for these excitationintensity dependence studies, various models for the S* formation have been proposed.12-14 (i) the inhomogeneous model, (ii) the sequential two-photon model, and (iii) the impulsive stimulated Raman scattering (ISRS) model. The ISRS model assumes that the S* state is the vibrationally excited hot S0 state, which is formed via the ISRS process, rather than an electronic excited state. The model was originally proposed to account for the results obtained by multipulse pump-deplete-probe spectroscopy.15,16 However, a recent TA experiment using narrowband excitation seems to exclude the ISRS process as a mechanism for S* formation.13 Although the inhomogeneous model, where the S* state is the result of a distorted conformation of the carotenoid already present in the ground state, has been favored,9-13,17,18 this model alone is not enough to quantitatively explain the observed spectroscopic features. Moreover, in solution, triplet states are apparently not formed from the S* state.6 Accordingly, the possibility that the S* state in solution differs from that observed in the LH complexes should be considered. The nature of the S* state is still a matter of debate, and there is no consistent model for all the experimental observations. In this work, we studied the formation of the S* and S1 states of carotenoids in a LH complex. We focus on the effect of nearby BChl on the relaxation dynamics of carotenoids on the basis of the hypothesis that energy flow related to the carotenoid dark states depends on the nearby BChl state. We performed femtosecond TA spectroscopy combined with pre-excitation of BChl in the LH1 complex from Rs. rubrum S1, containing spirilloxanthin as main carotenoid. Here, we used reconstituted LH1 with spirilloxanthin to remove other kinds of carotenoids that may be involved. BChl molecules were excited into the Qy state by a preexcitation pulse, and remained there with a lifetime of less than a nanosecond, while the induced changes in energy flow from the S2 state of spirilloxanthin in the LH1 complex were monitored by the excitation and probe pulses. This method based on the prepump-pump probe spectroscopy and analysis of induced changes in energy flow is expected to give a detailed picture of the S* state formation. Here, we emphasize that the role of the preexcitation pulse in TA spectroscopy is solely to prepare the BChl Qy population prior to excitation of spirilloxanthin, and that the TA signal changes induced by pre-excitation can be simply attributed to interaction between spirilloxanthin and the BChl

ARTICLE

Qy state in LH1 from Rs. rubrum. For further details, see the Supporting Information. We also measured the excitation-intensity dependence of the S* state formation by conventional TA spectroscopy for comparison purposes.

’ EXPERIMENTAL METHODS Sample Preparation. The reconstituted LH1 with spirilloxanthin was prepared as described elsewhere,19 to remove other kinds of carotenoids that may be involved. Cells of the purple photosynthetic bacterium, Rs. rubrum wild-type S1, were cultivated anaerobically in the light for 4 days. The LH1 subunit-type complex from Rs. rubrum was solubilized in 40 mM Tris-HCl buffer (pH 8.0) containing 0.03% LDAO. Carotenoids were isolated from the cells of Rs. rubrum, and all-trans-spirilloxanthin was purified using alumina column chromatography. The LH1 complex was reconstituted using the LH1 subunit-type complex and all-trans-spirilloxanthin, and then was purified by using an ion-exchange column (DE52) and sucrose-density gradient ultracentrifugation. 50 mg of poly(vinyl alcohol) (PVA, Kuraray Co., Ltd., PVA217) was dissolved into 1 mL of sample solution of the reconstituted LH1 complex (20 mM Tris-HCl buffer, pH 8.0, containing 0.01% LDAO). After a small portion of the sample solution was dropped onto a micro cover glass (MATSUNAMI GLASS Ind., Ltd., 18  18 mm), any residual solvent was removed under reduced pressure at 4 C. During the laser spectroscopic measurements, the sample was rapidly translated to avoid sample degradation and the accumulation of any potential photoproducts. Laser Spectroscopy. The femtosecond dispersed TA spectroscopy setup is based on an amplified mode-locked Ti:sapphire laser system operating at 1 kHz.20 A fraction of the amplified pulses was used to drive two independent optical parametric amplifiers (OPA): a collinear OPA pumped by the fundamental pulses (800 nm) and a noncollinear OPA (NOPA) pumped by the second harmonics (400 nm). In the excitationintensity dependence experiments, the excitation pulses resonant to the S0 f S2 transition of spirilloxanthin (545 nm, 100 fs) were obtained by sum-frequency mixing a fundamental pulse with an idler pulse at 1710 nm from the collinear OPA. In the pre-excitation experiments, the pre-excitation pulse (880 nm, 100 fs) for excitation to the Qy state of BChl was obtained by frequency doubling an idler pulse at 1760 nm from the collinear OPA, while the excitation pulses (520 nm, 20 fs) for spirilloxanthin were generated by the NOPA. In both experiments, the femtosecond supercontinuum generated by a sapphire plate was used as a broadband probe pulse. The full widths at half-maximum (fwhms) of the cross correlation traces between the excitation and probe pulses were 150 fs in the excitation intensity experiment and 80 fs in the preexcitation experiment. The polarizations of all the beams were set to be parallel to one another. The spot sizes at the sample position were about 150 μm for excitation and pre-excitation pulses and about 50 μm for the probe pulse. In the pre-excitation experiment, a delay time of the preexcitation pulse was fixed at 0.6 ps prior to the excitation pulse. A probe pulse after the sample was dispersed onto a linear image sensor with a spectrometer. The output signals were digitized and collected at the repetition rate of the laser system (1 kHz). The pre-excitation and excitation beams were modulated at 250 and 500 Hz, respectively, by a mechanical chopper, which was frequency locked to the laser pulse train. The noise level of the obtained absorbance change (ΔA) was smaller than 10-4 in the probe region of 460-1400 nm. Data Analysis. The data set consisting of 108 time-gated spectra (-1 to 450 ps) at 221 wavelengths (460-680 nm) obtained from TA experiments was fitted by using the global and target analysis technique.21 The temporal and spectral data set, Ψ(t,λ), is given by a superposition of the contributions of n different components: Ψðt, λÞ ¼ Σnl cl ðtÞεl ðλÞ 3234

dx.doi.org/10.1021/jp111718k |J. Phys. Chem. B 2011, 115, 3233–3239

The Journal of Physical Chemistry B

ARTICLE

Figure 1. Stationary absorption spectrum of LH1 from Rs. rubrum S1 used in TA spectroscpoies.

Figure 2. TA spectra measured without and with pre-excitation for the BChl Qy state (solid and dashed lines, respectively) at delay times of 1.0 and 5.0 ps after photoexcitation of spirilloxanthin in LH1. Intensities of both the excitation and the pre-excitation were set to 40 nJ/pulse (5.9  1014 photons cm-2 pulse-1). where cl(t) and εl(λ) are the concentration and species-associated difference spectrum (SADS) of the lth component, respectively. The shapes of cl(t) and εl(λ) can be independently determined by introducing a certain kinetic model expressed by the differential equation: ðd=dtÞCðtÞ ¼ KcðtÞ þ jðtÞ where K is the transfer matrix containing the rate constants, and j(t) is the input vector to the system which includes the instrumental response function (IRF). We used a Gaussian profile with a fwhm of 80 fs as IRF.

’ RESULTS AND DISCUSSION Pre-Excitation Effect on TA Spectra and Kinetics. Figure 1 shows the stationary absorption spectrum of the reconstituted LH1 complex from Rs. rubrum S1 used in TA spectroscopies. The peaks located at 450-570 nm are ascribed to the vibrational structure of the S0 f S2 transition of spirilloxanthin. The two peaks at 589 and 881 nm are the Qx and Qy bands of BChl, respectively. The solid lines in Figure 2 show the TA spectra in the visible region at 1.0- and 5.0 ps delay times after photoexcitation of spirilloxanthin in LH1. Negative signals below 560 nm are assigned to the ground state bleach of the S0 f S2 transition. Positive signals above 560 nm show a characteristic spectral feature with a two peak structure: the positive band peaking at 620 nm is assigned to the S1 f Sn transition, while the positive band located at the high-energy side of the S1 f Sn transition corresponds to the S* state. The amplitude ratio of the S* band to the S1 band increases at the longer delay time, indicating that the

Figure 3. TA kinetics measured without and with pre-excitation for the BChl Qy state (circles and triangles, respectively) at typical wavelengths of 660 (A), 620 (B), and 580 nm (C). The solid lines denote the result of the global fitting procedure. The inset in part B represents the data normalized to the peak on a logarithmic scale. The inset in part C shows the data normalized at a longer delay time than 6 ps. The open squares represent the difference between them. The solid line is an exponential function with a decay time of 1.4 ps convoluted with the instrumental response function.

S* state has a longer lifetime than the S1 state. These assignments will be discussed later in more detail. The dashed lines in Figure 2 show the TA spectra measured with the pre-excitation pulse. Remarkably, pre-excitation to the BChl Qy state affects the S*/S1 amplitude ratio: the S* band increases while the S1 band decreases when pre-excitation is applied. Figure 3 shows the TA kinetics changes in response to preexcitation observed at three typical wavelengths of 660, 620, and 580 nm, which mainly correspond to the hot S1 and S1, S1, and S* states, respectively, in this time window. The hot S1 state is a vibrationally excited state belonging to the S1 state, and appears as a shoulder at the low-energy side of the S1 f Sn transition in the TA spectrum. In addition to changes in the amplitudes of the S1 and S* states, it is recognized that the TA kinetics drastically change in response to pre-excitation. It is clearly seen that the decay time of the S1 state becomes shorter as shown in the inset in Figure 3B. Global fitting analysis discussed later gives time constants of 1.6 and 1.4 ps for the data without and with pre-excitation, respectively. On the other hand, it is seen that the S* amplitude increases with a slow rise component, whose time constant is determined to be approximately 1.4 ps, as shown in the inset in Figure 3C. These results imply that the energy flow of S1 f S* occurs when BChl is in the Qy state. The hot S1 state corresponds to a rapid decay component in the decay profile at 660 nm in Figure 3A. Although it can be seen that the amplitude decreases over a period longer than 0.5 ps in response to pre-excitation, it is 3235

dx.doi.org/10.1021/jp111718k |J. Phys. Chem. B 2011, 115, 3233–3239

The Journal of Physical Chemistry B

ARTICLE

Figure 4. Results of global and target analysis of TA data. (A) Species-associated difference spectra (SADS) obtained for the TA data taken without and with pre-excitation (solid and dashed lines, respectively). Time constants of 170 fs and 1.4 ps indicated in the second and third panels, respectively, correspond to the decay times of the data taken with pre-excitation. Other SADS components have the same decay times for both data sets. (B) Timedependent concentrations of the corresponding excited states. (C, D) Summaries of the energy level scheme and energy flow of spirilloxanthin and BChl obtained for the data without (C) and with (D) pre-excitation. Energy flows indicated by solid arrows are explicitly included in the global and target analysis. The associated decay time of each state is provided in parentheses. The additional energy flows that occur when BChl is pre-excited are indicated by thick red arrows with their rate constants.

difficult to recognize whether the hot S1 component changes or not. Global and Target Analysis. To obtain more insight into the kinetic behaviors without and with pre-excitation, the spectral (460-680 nm) and temporal (-1 to 450 ps) data sets were fitted using a global and target analysis procedure.21 We tested various target models for relaxation kinetics of spirilloxanthin in LH1 with different numbers of components and branching positions. We adopted the kinetic model depicted by solid black arrows in Figure 4C, which describes well our data taken without pre-excitation. The kinetic model is similar to that initially proposed by Gradinaru et al.,6 except for an additional component corresponding to the hot S1 state, and is one of the most accepted models for relaxation kinetics of spirilloxanthin in LH1. To simplify the analysis, the S* f S0 transition was not taken into account. In addition, we assumed that the energy transfer from the S2 state to BChl Qx was 30%,22 and that the energy transfer from the S1 and S* states to BChl Qy was 0%.6,23 The solid lines in Figure 4A show species-associated difference spectra (SADS) of five components with decay times of 80 fs, 180 fs, 1.6 ps, 5.7 ps, and 1 ns. These components are assigned to the S2, hot S1, S1, S*, and T1 states, respectively, because of their characteristic spectral profiles and decay times.6 The decay time of the fifth component was set to 1 ns in the global fitting analysis for convenience, although it did not decay on the time scale of our experiment. To elucidate changes in the energy flow induced by preexcitation, we also fitted the spectral and temporal data sets taken with pre-excitation and obtained five SADS components with decay times of 80 fs, 170 fs, 1.4 ps, 5.7 ps, and 1 ns. The decay times of the second and third SADS corresponding to the hot S1 and S1 states, respectively, became shorter. The data is reasonably explained by the model depicted in Figure 4D. An essential feature of this model is the existence of new energy flow pathways of (hot) S1 f S*. When the (hot) S1 f S* transition is not taken into account, the SADS obtained have spectral interference from

other SADS components. As can be seen in Figure 4A, the spectral profiles of each SADS component without and with preexcitation are almost identical. We determined the energy flow rates of hot S1 f S* and S1 f S* to be (3.1 ps)-1 and (11.2 ps)-1, respectively. Therefore, the branching yields of hot S1 and S1 states are 5.5% for hot S1 f S* and 12.5% for S1 f S*. These changes in the energy flow induced by pre-excitation are also seen in the time-dependent concentration of each state shown in Figure 4B. It can be recognized that triplet formation increases as the S* population increases. This clearly indicates that the S* state is an electronic excited state that is the precursor to triplet formation. On the other hand, the studies on S* state formation of carotenoids in solution propose that the S* state is the vibrationally hot S0 state.15,16,24 It seems that the possibility that the S* state in solution partially differs from that observed in the LH complexes is required to be taken into account. Excitation Intensity Dependences of TA Spectra and Kinetics. Figure 5A shows the excitation intensity dependence of the TA spectra in the visible region at 1.0 and 5.0 ps delay times. These spectra are normalized to negative signals below 560 nm assigned to the ground state bleach of the S0 f S2 transition. It can be seen that, under strong excitation conditions, the relative amplitude of the S* band increases while that of the S1 band decreases, consistent with the results reported so far.12-14 These spectral behaviors are quite similar to the results obtained by TA spectroscopy with pre-excitation of BChl as shown in Figure 2. The TA kinetics measured at different excitation intensities are shown in Figures 5B-D. The plots are scaled in the same way as in Figure 5A. The kinetics changes induced by strong excitation are similar to those induced by pre-excitation as shown in Figures 3A-C. In response to pre-excitation, the amplitude decreases over a time period longer than 0.5 ps at 660 nm, and also decreases at 620 nm, whereas at 580 nm it increases with a slow rise component. Correlation between S*/S1 and BChl Qy Population. A comparison between the two different TA spectroscopies indicates 3236

dx.doi.org/10.1021/jp111718k |J. Phys. Chem. B 2011, 115, 3233–3239

The Journal of Physical Chemistry B

ARTICLE

Figure 5. Excitation intensity dependences in conventional TA spectroscopy for spirilloxanthin in LH1: Weak excitation at 10 nJ/pulse (1.6  1014 photons cm-2 pulse-1) and strong excitation at 400 nJ/pulse (6.2  1015 photons cm-2 pulse-1). (A) TA spectra measured at delay times of 1.0 and 5.0 ps. TA kinetics measured at typical wavelengths of 660 (B), 620 (C), and 580 nm (D).

that (i) excitation of carotenoid to the S2 state at high intensity and (ii) pre-excitation to the BChl Qy state have the same effect on the population ratio of the S* to S1 states. Accordingly, this result raises the hypothesis that the excited state Qy population of BChls within the LH1 ring is a key parameter regulating the S*/ S1 population ratio of the carotenoid. The BChl Qy population is formed via excitation energy transfer from the carotenoid in the excitation-intensity dependence experiment, whereas the Qy population is mainly formed by direct photoexcitation in the pre-excitation experiment. We studied the S*/S1 population ratio at a delay time of 1.0 ps in relation to the relative population of the BChl Qy state, [BChl Qy]. Here, we define the S*/S1 population ratio as the ratio of the 580 to 620 nm amplitudes at a 1.0 ps delay time because these peaks are attributed to the S* and S1 components, respectively. On the other hand, [BChl Qy] is defined as the ratio of the Qy state population to total BChl population. Therefore, it is necessary to know the amount of the ground state bleach of BChl and the sample absorbance. The ground state bleach appears in the spectral regions of the S0 f Qy transition at around 890 nm and also in the S0 f Qx transition at around 590 nm. By assuming that the spectrum of the ground state bleach is the same as that of the ground state absorption for BChl, the amount of the ground state bleach can be estimated by separation from the stimulated emission and excited state absorption signals. Figure 6A shows [BChl Qy] at 0.1 ps as a function of excitation intensity for the spirilloxanthin S0 f S2 transition determined by conventional TA spectroscopy. Figure 6B shows [BChl Qy] at 0.1 ps as a function of pre-excitation intensity for the BChl S0 f Qy transition determined in the pre-excitation experiment. The dashed line indicates [BChl Qy] without pre-excitation, which is formed via energy transfer from spirilloxanthin. In Figure 6C, we plotted S*/S1 at 1.0 ps as a function of [BChl Qy] at 0.1 ps measured by the two different TA spectroscopies. This plot strongly indicates that the S*/S1 population ratio is correlated with the BChl Qy population. This is clear evidence that the population ratio between the dark states of S* and S1 is regulated by the BChl Qy state in LH1. It should be also noted that the correlation shown in Figure 6C supports that the role of the preexcitation pulse in TA spectroscopy is solely to prepare the BChl Qy population as discussed in Supporting Information. On the basis of our observations, anomalous behaviors of the excitation-intensity dependence of the S*/S1 ratio reported so far can be qualitatively explained as follows. (i) The excitation to the carotenoid S2 state leads to energy transfer to nearby BChl. Under strong excitation conditions, there is a probability that

Figure 6. Correlations between the S*/S1 amplitude ratio and the BChl Qy state population, [BChl Qy]. (A) [BChl Qy] at a delay time of 0.1 ps as a function of the excitation intensity in conventional TA spectroscopy. (B) [BChl Qy] at a delay time of 0.1 ps as a function of the pre-excitation intensity. The excitation intensity was set to 40 nJ/pulse (5.9  1014 photons cm-2 pulse-1). The dashed line indicates the value of [BChl Qy] measured without pre-excitation. (C) S*/S1 amplitude ratio, which is defined as the ratio of the TA amplitude at 580 nm to that at 620 nm, measured at a delay time of 1.0 ps is plotted as a function of [BChl Qy]. These are obtained by TA spectroscopies combined with excitation intensity dependence (circles) or with pre-excitation of BChl (triangles).

both carotenoid and BChl are in the excited states within the same LH1 ring. (ii) The excitation of BChl moves within a LH1 ring with wavelike motion or incoherent hopping, and then interacts with carotenoid in the excited state. (iii) As a result, the S* state formation is enhanced in the excited state of carotenoid due to interaction with the BChl Qy state. The proposed model explains why the anomalous excitation-intensity dependences have been observed only in complexes.12,13 The existence of BChl and formation of a super complex are necessary for the intensity dependent S*/S1 ratio. Possible Mechanism of S*/S1 Regulation. It has been shown experimentally for various kinds of carotenoids in solution that 3237

dx.doi.org/10.1021/jp111718k |J. Phys. Chem. B 2011, 115, 3233–3239

The Journal of Physical Chemistry B the spectroscopic properties of the S* state are greatly affected by the carotenoid structure, e.g., distortion from the planar π-conjugated backbone.9-11,18 Solvent dependence experiments and quantum chemical calculations for spirilloxanthin have shown that the S* state is more pronounced in a polar solvent than in a nonpolar solvent, which is explained by its optimized corkscrew conformation in a polar solvent.10 In addition, it has been indicated that the S* population is more pronounced in a protein matrix than in solution, which is also explained by the carotenoid backbone distortion caused by binding to the protein and by the electrostatic environmental effect.6 These studies strongly suggest that the spectroscopic properties of the S* state are largely determined by the carotenoid conformation in the ground state. Therefore, the inhomogeneous model where the conformation inhomogeneity in the ground state determines the S*/S1 population ratio is likely favorable. Although our experimental results do not completely exclude the possibility of ground state inhomogeneity, they clearly show that the regulation of the S*/S1 population ratio occurs in the electronic excited state. Quantum chemical calculations indicate that the S2 (1Buþ) state has a similar conformation to the ground state except that the bond order alternation is reduced. On the other hand, the S1 (2Ag-) state has a large amount of doubly excited character, and forms a more planar conformation with bond order reversal.9,10,25 It is thus expected that the distorted conformation in the ground state unravels to form a more planar conformation during relaxation to the S1 state after photoexcitation. During this relaxation process, the other stable conformation, the S* state, on the S1 potential is possibly formed, as proposed by Frank et al.9,10 As a result, the spectroscopic properties of the S* state are considered to be dependent on the conformation in the ground state. To reveal the nature of the S* state, it is necessary to gain insight into the mechanism of S*/S1 regulation by the BChl Qy state. A possible effect of pre-excitation of BChl to the Qy state is a change in the local electric field near the carotenoid induced by the transition dipole moment of BChl. Herek et al. first observed a Stark shift of carotenoid induced by the photoexcitation of BChl in LH complexes, and estimated that the induced local field near the carotenoid is more than a few MV/cm.26 It is expected that such a large electric field affects the relaxation dynamics to the S1 state, accompanied by conformational change if the carotenoid has a distorted geometry in the ground state and a net dipole moment. Other kinds of interactions between the carotenoid dark excited states and the BChl Qy state, such as excitonic coupling27 or a radical ion pair,28,29 do not account for our experimental results. If there is excitonic mixing between the carotenoid dark states and the BChl Qy state, the excitation energy will be transferred between them. However, there is no report on energy transfer in either direction (S1,S* f BChl Qy and S1,S* r BChl Qy) for LH1 from Rs. rubrum S1. It has been reported that carotenoid radical cations absorb in the near-IR region.30 The TA spectrum measured in the near-IR region exhibited no signal corresponding to the spirilloxanthin radical. This result is consistent with the result previously reported by Papagiannakis et al.23 Moreover, we did not observe any additional spectroscopic components in the detected region of 460-1400 nm. Therefore, the formation of a radical ion pair is unlikely to explain the interaction between the carotenoid dark excited states and the BChl Qy state.

ARTICLE

The results presented in this work show that excitation energy deactivation via the S* state of spirilloxanthin is more enhanced as the number of BChl molecules excited into the Qy state in the LH1 ring increases. Our observations seem to correspond to the regulation of the energy deactivation inherent in the core antenna LH1 under high light illumination conditions, where the excitation energy is transferred from the surrounding LH2. There may also be a possibility that a few BChl molecules are excited within the same LH1 ring. Although it is not clear if the regulation system of the energy deactivation is actually active and biologically significant in photosynthetic apparatuses, a detailed study of the regulation mechanism will provide new insights into the complicated network of energy flow pathways and the influence of carotenoid structure and its surrounding environment in the LH complex.

’ CONCLUSIONS In this work, we have shown that the energy flow from the S1 state, including its vibrationally excited states, to the S* state occurs only when nearby BChl is excited into the Qy state in the reconstituted LH1 complex from Rs. rubrum S1. In addition, we observed that the triplet population increases with the S* population, supporting the idea that the S* state is an electronic excited state and a precursor to triplet formation. An energy flow pathway regulated by the BChl Qy state in the LH1 ring accounts for the anomalous behavior that the S* and S1 states have different excitation-intensity dependences, which has generated considerable debate on the nature and formation mechanism of the S* state. ’ ASSOCIATED CONTENT

bS

Supporting Information. Details of effects of the preexcitation pulse on the main observations. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Telephone: þ81-22-795-6421. Fax: þ81-22-795-6425. E-mail: [email protected]. Present Addresses ^

Center for advanced science and innovation, Osaka University, 2-1 Yamada-oka, Suita, Osaka 565-0871, Japan.

’ ACKNOWLEDGMENT This work was partly supported by Grants-in-Aid from JSPS (No. 22740271 to R.N., No. 19340076 to M.Y., and Nos. 17204026 and 17654083 to H.H.). ’ REFERENCES (1) Koyama, Y.; Kuki, Y.; Andersson, P. O.; Gillbro, T. Photochem. Photobiol. 1996, 63, 243. (2) Frank, H. A.; Cogdell, R. J. Photochem. Photobiol. 1996, 63, 257. (3) Polívka, T.; Sundstr€om, V. Chem. Rev. 2004, 104, 2021. (4) Walla, P. J.; Linden, P. A.; Hsu, C. P.; Scholes, G. D.; Fleming, G. R. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 10808. (5) Polívka, T.; Sundstr€om, V. Chem. Phys. Lett. 2009, 477, 1. (6) Gradinaru, C. C.; Kennis, J. T. M.; Papagiannakis, E.; van Stokkum, I. H. M.; Cogdell, R. J.; Fleming, G. R.; Niederman, R. A.; van Grondelle, R. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 2364. 3238

dx.doi.org/10.1021/jp111718k |J. Phys. Chem. B 2011, 115, 3233–3239

The Journal of Physical Chemistry B

ARTICLE

(7) Papagiannakis, E.; Kennis, J. T. M.; van Stokkum, I. H. M.; Cogdell, R. J.; van Grondelle, R. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 6017. (8) Cong, H.; Niedzwiedzki, D. M.; Gibson, G. N.; LaFountain, A. M.; Kelsh, R. M.; Gardiner, A. T.; Cogdell, R. J.; Frank., H. A. J. Phys. Chem. B 2008, 112, 10689. (9) Niedzwiedzki, D. M.; Sullivan., J. O.; Polívka, T.; Birge, R. R.; Frank, H. A. J. Phys. Chem. B 2006, 110, 22872. (10) Niedzwiedzki, D.; Koscielecki, J. F.; Cong, H.; Sullivan, J. O.; Gibson, G. N.; Birge, R. R.; Frank, H. A. J. Phys. Chem. B 2007, 111, 5984. (11) Cong, H.; Niedzwiedzki, D. M.; Gibson, G. N.; Frank, H. A. J. Phys. Chem. B 2008, 112, 3558. (12) Papagiannakis, E.; van Stokkum, I. H. M.; Vengris, M.; Cogdell, R. J.; van Grondelle, R.; Larsen, D. S. J. Phys. Chem. B 2006, 110, 5727. (13) Jailaubekov, A. E.; Song, S.-H.; Vengris, M.; Cogdell, R. J.; Larsen, D. S. Chem. Phys. Lett. 2010, 487, 101. (14) Savolainen, J.; Buckup, T.; Hauer, J.; Jafarpour, A.; Serrat, C.; Motzkus, M.; Herek, J. L. Chem. Phys. 2009, 357, 181. (15) Wohlleben, W.; Buckup, T.; Hashimoto, H.; Cogdell, R. J.; Herek, J. L.; Motzkus, M. J. Phys. Chem. B 2004, 108, 3320. (16) Buckup, T.; Savolainen, J.; Wohlleben, W.; Herek, J. L.; Hashimoto, H.; Correia, R. R. B.; Motzkus, M. J. Chem. Phys. 2006, 125, 194505. (17) Christensson, N.; Milota, F.; Nemeth, A.; Sperling, J.; Kauffmann, H. F.; Pullerits, T.; Hauer, J. J. Phys. Chem. B 2009, 113, 16409. (18) Chabera, P.; Fuciman, M.; Hríbek, P.; Polívka, T. Phys. Chem. Chem. Phys. 2009, 11, 8795. (19) Nakagawa, K.; Suzuki, S.; Fujii, R.; Gardiner, A. T.; Cogdell, R. J.; Nango, M.; Hashimoto, H. J. Phys. Chem. B 2008, 112, 9467. (20) Kosumi, D.; Abe, K.; Karasawa, H.; Fujiwara, M.; Cogdell, R. J.; Hashimoto, H.; Yoshizawa, M. Chem. Phys. 2010, 373, 33. (21) van Stokkum, I. H. M.; Larsen, D. S.; van Grondelle, R. Biochim. Biophys. Acta 2004, 1657, 82. (22) Noguchi, T.; Hayashi, H.; Tasumi, M. Biochim. Biophys. Acta 1990, 1017, 280. (23) Papagiannakis, E.; van Stokkum, I. H. M.; van Grondelle, R.; Niederman, R. A.; Zigmantas, D.; Sundstr€om, V.; Polívka, T. J. Phys. Chem. B 2003, 107, 11216. (24) Lenzer, T.; Ehlers, F.; Scholz, M.; Oswald, R.; Oum, K. Phys. Chem. Chem. Phys. 2010, 12, 8832. (25) Hudson, B. S.; Kohler, B. E.; Schulten, K. Linear polyene electronic structure and potential surfaces In Excited States; Lim, E. C., Ed.; Academic Press: New York, 1982; Vol. 6, p 1. (26) Herek, J. L.; Wendling, M.; He, Z.; Polívka, T.; Garcia-Asua, G.; Cogdell, R. J.; Hunter, C. N.; van Grondelle, R.; Sundstr€om, V.; Pullerits, T. J. Phys. Chem. B 2004, 108, 10398. (27) van Amerongen, H.; van Grondelle, R. J. Phys. Chem. B 2001, 105, 604. (28) Pan, J.; Xu, Y.; Sun, L.; Sundstr€om, V.; Polívka, T. J. Am. Chem. Soc. 2004, 126, 3066. (29) Holt, N. E.; Zigmantas, D.; Valkunas, L.; Li, X.-P.; Niyogi, K. K.; Fleming, G. R. Science 2005, 307, 433. (30) Jeevarjan, J. A.; Wei, C. C.; Jeevarajan, A. S.; Kispert, L. D. J. Phys. Chem. 1996, 100, 5637.

3239

dx.doi.org/10.1021/jp111718k |J. Phys. Chem. B 2011, 115, 3233–3239