J. Phys. Chem. B 2008, 112, 12795–12800
12795
Ultrafast Pump-Probe Study of the Primary Photoreaction Process in pharaonis Halorhodopsin: Halide Ion Dependence and Isomerization Dynamics Takumi Nakamura,† Satoshi Takeuchi,† Mikihiro Shibata,‡ Makoto Demura,§ Hideki Kandori,‡ and Tahei Tahara*,† Molecular Spectroscopy Laboratory, RIKEN (The Institute of Physical and Chemical Research), 2-1 Hirosawa, Wako 351-0198, Japan, Department of Materials Science and Engineering, Nagoya Institute of Technology, Showa-ku, Nagoya 466-8555, Japan, and DiVision of Biological Science, Graduate School of Science, Hokkaido UniVersity, Sapporo 060-0812, Japan ReceiVed: April 15, 2008; ReVised Manuscript ReceiVed: June 30, 2008
Halorhodopsin is a retinal protein that acts as a light-driven chloride pump in the Haloarchaeal cell membrane. A chloride ion is bound near the retinal chromophore, and light-induced all-trans f 13-cis isomerization triggers the unidirectional chloride ion pump. We investigated the primary ultrafast dynamics of Natronomonas pharaonis halorhodopsin that contains Cl-, Br-, or I- (pHR-Cl-, pHR-Br-, or pHR-I-) using ultrafast pump-probe spectroscopy with ∼30 fs time resolution. All of the temporal behaviors of the Sn r S1 absorption, ground-state bleaching, K intermediate (13-cis form) absorption, and stimulated emission were observed. In agreement with previous reports, the primary process exhibited three dynamics. The first dynamics corresponds to the population branching process from the Franck-Condon (FC) region to the reactive (S1r) and nonreactive (S1nr) S1 states. With the improved time resolution, it was revealed that the time constant of this branching process (τ1) is as short as 50 fs. The second dynamics was the isomerization process of the S1r state to generate the ground-state 13-cis form, and the time constant (τ2) exhibited significant halide ion dependence (1.4, 1.6, and 2.2 ps for pHR-Cl-, pHR-Br-, and pHR-I-, respectively). The relative quantum yield of the isomerization, which was evaluated from the pump-probe signal after 20 ps, also showed halide ion dependence (1.00, 1.14, and 1.35 for pHR-Cl-, pHR-Br-, and pHR-I-, respectively). It was revealed that the halide ion that accelerates isomerization dynamics provides the lower isomerization yield. This finding suggests that there is an activation barrier along the isomerization coordinate on the S1 potential energy surface, meaning that the three-state model, which is now accepted for bacteriorhodopsin, is more relevant than the two-state model for the isomerization process of halorhodopsin. We concluded that, with the three-state model, the isomerization rate is controlled by the height of the activation barrier on the S1 potential energy surface while the overall isomerization yield is determined by the branching ratios at the FC region and the conical intersection. The third dynamics attributable to the internal conversion of the S1nr state also showed notable halide ion dependence (τ3 ) 4.5, 4.6, and 6.3 ps for pHR-Cl-, pHR-Br-, and pHR-I-). This suggests that some geometrical change may be involved in the relaxation process of the S1nr state. 1. Introduction Halorhodopsin (HR) is a membrane protein that acts as a light-driven chloride pump in a haloarchaeal cell. In this protein, the all-trans retinyl chromophore is connected to a lysine residue through a Schiff base linkage,1-6 and it undergoes all-trans f 13-cis photoisomerization with the absorption of light. This photoisomerization of the chromophore triggers the unidirectional chloride ion pump from the extracellar to cytoplasm side, which proceeds over a millisecond time scale.1-6 An X-ray diffraction study for salinarum halorhodopsin (sHR) clarified the binding position of the chloride ion as well as the local structure around the Schiff base in the ground state.7 The structure is depicted in Figure 1a. The primary photoisomerization process in sHR has been studied with time-resolved spectroscopy by several groups.8-12 Especially, Kandori et al. conducted the first femtosecond time* To whom correspondence should be addressed. Phone: 81-43-467-4592. Fax: 81-43-467-4539. E-mail:
[email protected]. † RIKEN. ‡ Nagoya Institute of Technology. § Hokkaido University.
resolved absorption study of sHR8 and observed four transients in the picosecond time region, (1) Sn r S1 absorption (460-510 nm), (2) ground-state bleaching (around 578 nm), (3) K intermediate absorption (around 600 nm), and (4) stimulated emission (near-infrared region). They reported that the Sn r S1 absorption, ground-state bleaching, and stimulated emission decayed with a time constant of 2.3 ps, whereas the transient absorption due to the K intermediate grew up with a time constant of ∼1 ps. To explain the faster formation of the K intermediate than the decay of the S1 state, they introduced twodimensional coordinates to the reaction mechanism and proposed that the initial excited state generated in the Franck-Condon (FC) region is branched into two relaxation pathways along different coordinates, that is, the isomerization coordinate accessing the 13-cis form and another coordinate leading to the nonreactive S1 state. Arlt et al. studied the same system with an improved time resolution10 and found a 170 fs component in the stimulated emission. In addition, they revealed that the decay kinetics on the picosecond time scale was actually not single exponential but biexponential (1.5 and 8.5 ps). On the basis of these observation, they drew a model potential that also
10.1021/jp803282s CCC: $40.75 2008 American Chemical Society Published on Web 09/13/2008
12796 J. Phys. Chem. B, Vol. 112, No. 40, 2008
Figure 1. Retinyl chromophore in halorhodopsin; (a) the structure determined by X-ray crystallography17 and (b) a sketch of the hydrogenbonding alternation accompanied with all-trans f 13-cis isomerization of the chromophore.16
rationalizes the reaction mechanism proposed by Kandori et al.8 (Figure 2a) and explained the initial isomerization process as follows; the S1 state is initially generated at the FC region, and it relaxes along two different coordinates to generate two types of S1 states, that is, reactive (S1r) and nonreactive (S1nr) states. The S1r state is subsequently converted to the K intermediate, while the S1nr state goes back to the ordinary all-trans ground state by the S1 f S0 internal conversion. It was considered that the isomerization of the S1r state proceeds without any activation barrier in the S1 state. Very recently, Peters et al. investigated the isomerization dynamics of sHR with visible pump/IR probe spectroscopy and observed that the infrared bands due to the C-C and CdC stretching vibrations of the K intermediate (13cis form) grew up with a time constant of 1.5 ps.12 According to the above-mentioned mechanism, they concluded that the K intermediate is generated from the S1r state with the time constant of 1.5 ps after the population branching at the FC region. HR does not pump only Cl- but also other anions such as Br-, I-, and NO3-.13-18 Recently, Shibata et al. studied the halide ion dependence of the N-D stretching frequency of the (deuterated) Schiff base in pharaonis halorhodopsins (pHR) that contains Cl-, Br-, or I-.17 Low-temperature FTIR spectra of pHR exhibited different vibrational frequencies, depending on the anion. In the all-trans form, the N-D stretching frequency of the Schiff base decreased as the ion radius became larger (2488 cm-1 (Cl-), 2480 cm-1 (Br-), and 2458 cm-1 (I-)), whereas the corresponding frequency of the K intermediate showed the opposite trend (2338 cm-1 (Cl-), 2375 cm-1 (Br-), and 2427 cm-1 (I-)). On the basis of this finding, they concluded that a hydrogen-bonding alteration occurs between the all-trans form and K intermediate; a hydrogen bond is formed between the Schiff base and a water molecule in the all-trans form, while
Nakamura et al.
Figure 2. Model potential energy surfaces and mechanisms for the initial photoisomerization process of halorhodopsin; (a) the present potential model based on the two-state model10 and (b) the potential model based on the three-state model. S1n: reactive S1 state; S1nr: nonreactive S1 state.
the bonding partner of the Schiff base is switched to the halide ion in the K intermediate (Figure 1b). The halide ion dependence of the N-D stretching frequency indicates that interaction between the Schiff base and the halide ion significantly changes with the change of the radius of the halide ion and affects the force field of the chromophore. Naturally, it is intriguing to examine the halide ions dependence of the dynamics of the alltrans f 13-cis isomerization process because such a study is expected to provide new information about the mechanism of the primary photoisomerization process in HR. In this paper, we report the halide-ion dependence of the initial isomerization process of Natronomonas pharaonis halorhodopsin containing Cl-, Br-, or I- (pHR-Cl-, pHR-Br-, and pHR-I-, respectively). The dynamics of the Sn r S1 absorption, ground-state bleaching, K intermediate absorption, and stimulated emission were monitored by visible pump/visible probe spectroscopy with 30 fs time resolution. The improved time resolution enabled us to observe the ∼50 fs component in the Sn r S1 absorption as well as that in the stimulated emission. More importantly, the observed time-resolved data provided information about halide ion dependence of the rate and yield of the all-trans f 13-cis isomerization process. These results indicated that the present model potential cannot rationalize the observed data and that an activation barrier for isomerization should be introduced in the S1 potential to explain the primary process of the photoinduced halide ion pumping of HR. 2. Experimental Section Materials. The pHR samples were prepared as described previously.17 Briefly, the pHR protein was expressed in E. coli
Primary Photoreaction Process in pharaonis Halorhodopsin
Figure 3. Steady-state absorption spectra of pHR-Cl-, pHR-Br-, and pHR-I- in Tris (1 × 10-2 mol dm-3) (top) and the spectra of the pump and probe pulses (bottom).
BL-21 (DE3) cells, and a fraction of the protein using Ni-NTA agarose was collected by elution with buffer E′ {50 mM tris (pH 7.0), 300 mM NaCl, 300 mM imidazole, and 0.1% n-dodecyl β-D-maltopyranoside (DM)}. The replacement of halides was performed by three dialysis cycles. The buffer conditions for the measurements were 10 mM Tris (pH 7.0) and 0.1% DM in the presence of 150 mM NaCl, NaBr, or NaI. Methods. Two-Color Pump-Probe Measurements. Twocolor pump-probe measurements were carried out by using a setup based on two home-built noncollinear optical parametric amplifiers (NOPA).19 Briefly, a femtosecond mode-locked Ti: Sapphire oscillator (Coherent, Mira-900F; >0.8 W, 800 nm) was used to seed a regenerative amplifier (Coherent, Legend; 1 kHz, 1 mJ, 100 fs, 800 nm). The amplified pulse was split into two equal-energy pulses, and they were used to pump the two NOPAs. The output of the first NOPA (16 fs, 610 nm) was used as a pump pulse for photoexcitation of the sample. The output of the second NOPA was divided into two, and they were used as a probe and a reference pulse. At the sample position, the energies of the pump and probe pulses were about 30 and 0.2 nJ, respectively. The sample solution was circulated in a flow cell (0.2 mm path length, 2 mm thick quartz window) by a peristaltic pump. The probe pulse after the sample and the reference pulse were detected by photodiodes, and the electric signals were processed by an A/D converter. The pump and probe polarizations were set at the magic angle condition. The instrumental response of the experiments was determined to be ∼30 fs by cross-correlation measurements of the pump and probe pulses. Steady-State Absorption Spectra. Steady-state absorption spectra were measured by commercial spectrometer (U-3310, Hitachi). 3. Results and Discussion Population Dynamics in the Initial Isomerization Process. We first discuss pump-probe transients of pHR-Cl- and its isomerization process. Figure 3 shows the absorption spectra of pHR-Cl-, pHR-Br-, and pHR-I-, as well as the spectra of the pump and probe pulses. The pump wavelength used in this experiment matches the red part of the lowest-energy absorption band of pHR. The probe wavelength was tuned to 490, 560, 650, or 690 nm. Roughly, these four probe wavelengths
J. Phys. Chem. B, Vol. 112, No. 40, 2008 12797 correspond to the region of the Sn r S1 absorption, groundstate bleaching, K intermediate absorption, and stimulated emission, respectively, which were reported in the femtosecond absorption study of Kandori et al.8 The pump-probe transients of pHR-Cl- observed at the four wavelengths are displayed in Figure 4. At 490 nm, we observed a positive signal due to the Sn r S1 absorption that decays in the picosecond time region (Figure 4a).8-12 The observed signal in the picosecond time range was well fitted by a biexponential function with time constants of τ2 ) 1.4 ( 0.2 ps and τ3 ) 4.5 ( 0.2 ps. The same picosecond decay components were also observed in the pump-probe transients at 560, 650, and 690 nm (Figure 4b-d). This observation is fully consistent with previous reports;10,12 therefore, the τ2 and τ3 components are assignable to the decays of the S1r and S1nr states, respectively. In addition to these picosecond components, a long-lived component was clearly seen after 20 ps at the probe wavelengths of 560 and 650 nm. This component indicates the formation of the K intermediate. In fact, the long-lived component was observed as a positive signal at 650 nm, where the absorption of the K intermediate appears. In the femtosecond time region, it was found that the Sn r S1 absorption at 490 nm showed a finite rise time, as shown in Figure 5a. The time constant of this ultrafast component (τ1) was evaluated as ∼50 fs by a fitting analysis taking account of the instrumental response. Figure 5b shows that the same ultrafast component also appeared at 560 nm, where the groundstate bleaching mainly contributes to the signal. Since the ground-state bleaching itself should appear instantaneously with photoexcitation, the observation of the ultrafast component at 560 nm indicates that a positive signal due to the Sn r S1 absorption overlaps with the ground-state bleaching signal at this wavelength. In other words, the 560 nm probe pulse monitors not only the ground-state bleaching but also the red part of the Sn r S1 absorption band. As seen in Figure 5, the amplitude of the ultrafast component is positive at 560 nm, but it is negative at 490 nm. It means that this ultrafast component disappears at 560 nm and appears at 490 nm with a time constant of ∼50 fs. Therefore, this ultrafast component can be attributed to the blue shift of the Sn r S1 absorption band, which takes place with the time constant of ∼50 fs. We note that it was also necessary to include a small but significant contribution of this ultrafast component to obtain a good fit at 690 nm, that is, the blue part of the stimulated emission band. Consequently, the ultrafast component of τ1 ) ∼50 fs is attributable to the initial relaxation process from the FC state to the reactive/ nonreactive S1 state, which induces the spectral blue shift of the Sn r S1 absorption band as well as the red shift of the stimulated emission band. The improved time resolution of the present study enabled us to successfully determine the relaxation time of the initial FC state as short as 50 fs, which is significantly shorter than the values estimated in previous reports (17010 and 300 fs12). Halide Dependence of the Ultrafast Dynamics. To discuss the halide ion dependence of the primary reaction dynamics of HR, we measured pump-probe transients for pHR-Cl-, pHRBr-, and pHR-I- and compared the obtained signals in Figure 6. Obviously, the obtained signals showed significant halide ion dependence in the picosecond time scale, that is, the τ2 and τ3 time constants changed depending on the halide ion. By a fitting analysis, the two picosecond time constants were evaluated as τ2 ) 1.4, 1.6, and 2.2 ps and τ3 ) 4.5, 4.6, and 6.3 ps for pHRCl-, pHR-Br-, and pHR-I-, respectively. The halide ion dependence of the τ1 value was not noticeable within our
12798 J. Phys. Chem. B, Vol. 112, No. 40, 2008
Nakamura et al.
Figure 4. Pump-probe transients for pHR-Cl- at (a) 490, (b) 560, (c) 650, and (d) 690 nm in Tris (1 × 10-2 mol dm-3) observed with 610 nm excitation. Black curves indicate the fitting results.
Figure 5. Pump-probe transients of pHR-Cl- at (a) 490 and (b) 560 nm in the early time region. The solid black curves are the best fits. The dotted curves represent the simulated signals that appear instantaneously.
experimental accuracy. The amplitudes and time constants of the three components for pHR-Cl-, pHR-Br-, and pHR-I- are summarized in Table 1. Our experiments on the halide ion dependence showed that the τ2 time constant becomes longer as the radius of the halide ion increases. This means that the isomerization from the S1r state (all-trans) to the K intermediate (13-cis) needs a longer time with the increase of the ion radius. It was observed that the N-D stretching frequency in the all-trans form substantially changed with the change of halide ion in the previous lowtemperature FTIR study.17 This indicates that the interaction between the halide ion and the positively charged nitrogen atom of the Schiff base is strengthened as the electron cloud of the halide ion is spatially extended, thereby perturbing the local structure and/or force field around the Schiff base. Therefore, it is readily expected that the interaction also affects the dynamics of the isomerization. In this sense, the halide ion dependence observed in the present study is consistent with the assignment of the τ2 component to the isomerization process from the S1r state to the K intermediate. It should be noted that the isomerization dynamics may be influenced also by the
change in the mass. Nevertheless, the halide ion does not move before and after isomerization, and it is pumped after photoisomerization.1-6 Thus, it is reasonable to consider that the change of the radii (or the change of the interaction) predominantly causes the halide ion dependence of the dynamics observed. We also observed a halide ion dependence of the τ3 time constant that corresponds to the internal conversion from the S1nr state to the ground state of the all-trans form. In this paper, we do not make further discussion on the halide ion dependence of the τ3 time constant because τ3 is not directly related to the isomerization dynamics in question. Nevertheless, the observed halide ion dependence seems to suggest that some geometrical change is involved in the relaxation process of the S1nr state and the activation barrier along the nonreactive coordinate is affected by the change of the halide ion. Halide Ion Dependence of the Branching Ratio and Isomerization Yield. In addition to the halide ion dependence of the ultrafast dynamics (i.e., τ2 and τ3), the pump-probe data provided information about the halide ion dependence of the population branching ratio at the FC region as well as the total
Primary Photoreaction Process in pharaonis Halorhodopsin
J. Phys. Chem. B, Vol. 112, No. 40, 2008 12799
Figure 6. Pump-probe transients of pHR-Cl-, pHR-Br-, and pHR-I- at (a) 490, (b) 560, (c) 650, and (d) 690 nm (pump 610 nm, in Tris, 1 × 10-2 mol dm-3). Black lines are the fitting results.
TABLE 1: Time Constants and Amplitudes for the Pump-Probe Signals sample -
pHR-Cl
pHR-Br-
pHR-I-
probe wavelength
Α1
490 nm -0.55 ( 0.09 560 nm 0.18 ( 0.01 650 nm -a 690 nm -0.019 ( 0.01 averaged time constants 490 nm -0.53 ( 0.08 560 nm 0.19 ( 0.01 650 nm -a 690 nm -0.022 ( 0.01 averaged time constants 490 nm -0.53 ( 0.12 560 nm 0.10 ( 0.02 650 nm -a 690 nm -0.035 ( 0.02 averaged time constants
τ1/ps
Α2
τ2/ps
0.038 ( 0.02 0.052 ( 0.03 -a 0.045b 0.045 0.051 ( 0.013 0.047 ( 0.03 -a 0.049b 0.049 0.044 ( 0.02 0.060 ( 0.02 -a 0.052b 0.052
0.23 ( 0.02 -0.14 ( 0.03 -0.48 ( 0.07 -0.21 ( 0.06 0.18 ( 0.02 -0.13 ( 0.06 -0.42 ( 0.07 -0.23 ( 0.01 0.16 ( 0.04 -0.20 ( 0.07 -0.38 ( 0.05 -0.24 ( 0.02
1.4( 0.2 1.4 ( 0.3 1.2 ( 0.2 1.4 ( 0.3 1.4 1.7 ( 0.2 1.7 ( 0.4 1.5 ( 0.3 1.6 ( 0.3 1.6 2.2 ( 0.3 2.3 ( 0.5 2.1 ( 0.2 2.2 ( 0.5 2.2
Α3
τ3/ps
Α4
0.32 ( 0.02 -0.33 ( 0.03 -0.26 ( 0.06 -0.29 ( 0.06
4.5 ( 0.2 4.6 ( 0.3 4.3 ( 0.4 4.4 ( 0.6 4.5 4.6 ( 0.2 4.6 ( 0.5 4.5 ( 0.2 4.6 ( 0.4 4.6 6.7 ( 0.2 6.1 ( 1.0 6.5 ( 0.5 5.9 ( 1.5 6.3
-0.048 ( 0.002 0.076 ( 0.04 -0.006 ( 0.003
0.37 ( 0.04 -0.34 ( 0.06 -0.35 ( 0.05 -0.27 ( 0.01 0.37 ( 0.06 -0.25 ( 0.07 -0.35 ( 0.04 -0.27 ( 0.02
-0.055 ( 0.002 0.086 ( 0.03 -0.016 ( 0.002 -0.060 ( 0.004 0.11 ( 0.02 -0.026 ( 0.004
a The fitting was not carried out for the early time region due to the quartz cell signal. b Averaged value of τ1 at 490 and 560 nm was used for the fitting.
quantum yield of the all-trans f 13-cis isomerization of the chromophore. The population branching ratio at the FC region can be discussed by the amplitude ratio of the τ2 and τ3 components (A2/A3) because these amplitudes reflect the initial populations of the S1r and S1nr states that are formed after the population branching at the FC state. This amplitude ratio was evaluated as 0.72, 0.49, and 0.43 for pHRCl-, pHR-Br-, and pHR-I-, respectively, from the pump-probe signals measured at 490 nm (Figure 6a). The amplitude ratio clearly showed significant halide ion dependence, and the value for pHRCl- is much larger than those for pHR-Br- and pHR-I-. This result means that the initial population branching is most favorable for isomerization in pHR-Cl-. At delay times later than 20 ps, both of the S1r and S1nr states are completely quenched. In this time region, the residual longlived signal was observed, which is ascribed to the sum of the ground-state bleaching (all-trans form) and K intermediate absorption (13-cis). Although the sign of the long-lived com-
ponent is determined by a relative value of the extinction coefficient of the two forms at each wavelength, its absolute amplitude directly reflects the amount of the K intermediate formed with relaxation of the S1 state, that is, the total quantum yield of the isomerization. As seen in Figure 6, the absolute amplitude of the long-lived component increases in the order of pHR-Cl-, pHR-Br-, and pHR-I-, which indicates that the isomerization quantum yield becomes larger in this order. In other words, the K intermediate is most efficiently formed in pHR-I- among the three HR samples. It is noteworthy that the order of the total isomerization quantum yield is opposite to that of the initial branching ratio at the FC state (vide infra). Discussion on Halide Ion Dependence and Potential Models. The present study of halide ion dependence of all-trans f 13-cis isomerization of pHR showed that the total isomerization quantum yield is the largest in pHR-I-, although the time constant of the isomerization (τ2) is the shortest. This result does not fit the existing potential model (Figure 2a) that has been adopted for HR.
12800 J. Phys. Chem. B, Vol. 112, No. 40, 2008 In this model, the τ2 time constant corresponds to the time that the excited-state nuclear wavepacket slides down along the isomerization coordinate and relaxes to the ground state by the internal conversion. As the Landau-Zener tunneling model suggests,20,21 the tunneling probability in the potential curve crossing region is expected to be higher as the velocity of the nuclear motion becomes faster. Therefore, as the inverse of τ2 (1/τ2) is considered to represent the wavepacket velocity, it is natural to think that the isomerization quantum yield increases as the rate of the isomerization increases. However, the present results contradict this expectation. The relation between the rate and the quantum yield cannot be straightforwardly rationalized on the basis of the present model potential. For the primary process of bacteriorhodopsin,22,23 it was first considered that ultrafast isomerization takes place in a barrierless way on the S1 potential (the two-state model). Later, however, it was pointed out that this two-state model cannot explain the lack of the temporal spectral shift of the emission in bacteriorhodopsin after 100 fs,24-27 and the three-state model was introduced as a more relevant model.24,25 In the three-state model, the avoided crossing between the S1 and S2 states gives rise to a small activation barrier on the S1 potential energy surface along the isomerization coordinate. This three-state model has been supported by several experiments that indicated the presence of the activation barrier.28-30 Especially, Ruhman et al. carried out pump-dump-probe experiments and found that the excited-state depletion was insensitive to the timing of the dump pulse and that depletion of the fluorescent state led to an equal reduction of the photoproduct.29 Their experiments revealed that the excited state has a constant structure throughout the fluorescence lifetime, which implies that the population in the excited state is dammed up by an activation barrier. In other words, the rate-determining step in the three-state model is the process of surmounting the activation barrier in the S1 state. In both HR and BR, the same all-trans f 13-cis isomerization takes place to trigger the photocycle, and a mutant of BR also acts as a light-driven chloride-pumping retinal protein.31 Because of the high similarity between HR and BR, it is natural to consider that the S1 potential feature based on the three-state model is also more relevant to photoisomerization in HR. Actually, to rationalize our present observation, it is much more natural to consider the presence of an activation barrier along the isomerization coordinate on the S1 potential because the rate-determining step and the yielddetermining step are separated. In the three-state model, the isomerization rate is controlled by the height of the activation barrier in the S1 state. The overall isomerization quantum yield is determined by both the branching ratio at the FC state and that at the conical intersection, which exists on the isomerization coordinate.22-30 Therefore, we concluded that the model potential based on the three-state model is relevant also for the primary isomerization process of HR. The model potential concluded in the present study is depicted in Figure 2b. In the present study, significant halide ion dependence was found for the isomerization time constant of all-trans f 13-cis isomerization (τ2). Because the activation barrier determines the isomerization rate in the three-state model, this observation indicates that the activation barrier height is notably affected by the interaction between the Schiff base and the halide ion. The halide ion dependence was also observed for the branching ratio at the FC state as well as the overall isomerization quantum yield. The overall isomerization quantum yield was highest in pHR-I-, although the branching at the FC state was least favorable for isomerization. This means that, in pHR-I-, the branching ratio at the conical intersection to form the K
Nakamura et al. intermediate is so large that it compensates the small branching ratio at the FC state, which results in the highest isomerization quantum yield. Thus, it is clear that the branching ratio at the conical intersection also depends on the halide ions, although it was not directly measured in the present study. Because the branching ratio is expected to be sensitive to the structure of the potential energy surfaces, the present study indicates that a change of the halide ion induces a notable change in the local environment around the Schiff base of the retinyl chromophore, which significantly affects the S1 potential energy surface. Acknowledgment. This work was supported by Grant-inAid for Science Research on Priority Area (No. 19056009) from MEXT. S.T. acknowledges financial support by a Grant-in-Aid for Scientific Research (B) (No. 19350017) from JPSP. Supporting Information Available: Normalized absorption spectra. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Matsuno-Yagi, A.; Mukohata, Y. Arch. Biochem. Biophys. 1980, 199, 297. (2) Schobert, B.; Lanyi, J. K. J. Biol. Chem. 1982, 257, 10306. (3) Lanyi, J. K. FEBS Lett. 1984, 175, 475. (4) Oesterhelt, D. Isr. J. Chem 1995, 35, 475. (5) Varo, G.; Brown, L. S.; Sasaki, H.; Kandori, H.; Maeda, A.; Needleman, R.; Lanyi, J. K. Biochemistry 1995, 34, 14490. (6) Varo, G.; Brown, L. S.; Sasaki, H.; Kandori, H.; Maeda, A.; Needleman, R.; Lanyi, J. K. Biochemistry 1995, 34, 14500. (7) Kolbe, M.; Besir, H.; Essen, L.-O.; Oesterhelt, D. Science 2000, 288, 1390. (8) Kandori, H.; Yoshihara, K.; Tomioka, H.; Sasabe, H. J. Phys. Chem. 1992, 96, 6066. (9) Kandori, H.; Yoshihara, K.; Tomioka, H.; Sasabe, H.; Shichida, Y. Chem. Phys. Lett. 1993, 211, 385. (10) Arlt, T.; Schmidt, S.; Zinth, W.; Haupts, U.; Oesterhelt, D. Chem. Phys. Lett. 1995, 241, 559. (11) Kobayashi, T.; Kim, M.; Taiji, M.; Iwasa, T.; Nakagawa, M.; Tsuda, M. J. Phys. Chem. B 1998, 102, 272. (12) Peters, F.; Herbst, J.; Tittor, J.; Oesterhelt, D.; Diller, R. Chem. Phys. 2006, 323, 109. (13) Steiner, M.; Oesterhelt, D.; Ariki, M.; Lanyi, J. K. J. Biol. Chem. 1984, 4, 2179. (14) Walter, T. J.; Braiman, M. S. Biochemistry 1994, 33, 1724. (15) Rudiger, M.; Haupts, U.; Gerwert, K.; Oesterhelt, D. EMBO J. 1995, 14, 1599. (16) Hutson, M. S.; Shilov, S. V.; Krebs, R.; Braiman, M. S. Biophys. J. 2001, 80, 1452. (17) Shibata, M.; Muneda, N.; Sasaki, T.; Shimono, K.; Kamo, N.; Demura, M.; Kandori, H. Biochemistry 2005, 44, 12279. (18) Magyari, K.; Simon, V.; Varo, G. J. Photochem. Photobiol., B 2006, 82, 16. (19) Takeuchi, S.; Tahara, T. J. Chem. Phys. 2004, 120, 4768. (20) Landau, L. D.; Lifshitz, E. M. Quantum Mechanics (Non-relatiVistic Theory); 3rd ed.; Pergamon Press: Oxford, U.K., 1981. (21) Wittig, C. J. Phys. Chem. B 2005, 109, 8428. (22) Dobler, J.; Zinth, W.; Kaiser, W. Chem. Phys. Lett. 1988, 144, 215. (23) Mathies, R. A.; Brito Cruz, C. H.; Polland, W. T.; Shank, C. V. Science 1988, 240, 777. (24) Hasson, K. C.; Gai, F.; Anfinrud, P. A. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 15124. (25) Gai, F.; Hasson, K. C.; McDonald, J. C.; Anfinrud, P. A. Science 1998, 279, 1886. (26) Haran, G.; Wynne, K.; Xie, A.; He, Q.; Chance, M.; Hochstrasser, R. M. Chem. Phys. Lett. 1996, 261, 389. (27) Du, M.; Fleming, G. R. Biophys. Chem. 1993, 48, 101. (28) Kobayashi, T.; Saito, T.; Ohtani, H. Nature 2001, 414, 531. (29) Ruhman, S.; Hou, B.; Friedman, N.; Ottolenghi, M.; Sheves, M. J. Am. Chem. Soc. 2002, 124, 8854. (30) McCamant, D. W.; Kukura, P.; Mathies, R. A. J. Phys. Chem. B 2005, 109, 10449. (31) Sasaki, J.; Brown, L. S.; Chon, Y. S.; Kandori, H.; Maeda, A.; Needleman, R.; Lanyi, J. K. Science 1995, 269, 73.
JP803282S
