Article Cite This: J. Phys. Chem. B XXXX, XXX, XXX−XXX
pubs.acs.org/JPCB
Origin of the Reactive and Nonreactive Excited States in the Primary Reaction of Rhodopsins: pH Dependence of Femtosecond Absorption of Light-Driven Sodium Ion Pump Rhodopsin KR2 Shinya Tahara,† Satoshi Takeuchi,†,‡ Rei Abe-Yoshizumi,§ Keiichi Inoue,§,∥,⊥ Hiroyuki Ohtani,# Hideki Kandori,§,∥ and Tahei Tahara*,†,‡ †
Molecular Spectroscopy Laboratory and ‡Ultrafast Spectroscopy Research Team, RIKEN, 2-1 Hirosawa, Wako 351-0198, Japan Department of Frontier Materials and ∥OptoBioTechnology Research Center, Nagoya Institute of Technology, Showa-ku, Nagoya 466-8555, Japan ⊥ PRESTO, Japan Science and Technology Agency, 4-1-8 Honcho Kawaguchi, Saitama 332-0012, Japan # Graduate School of Bioscience and Biotechnology, Tokyo Institute of Technology, Yokohama 226-8501, Japan §
S Supporting Information *
ABSTRACT: KR2 is the first light-driven Na+-pumping rhodopsin discovered. It was reported that the photoexcitation of KR2 generates multiple S1 states, i.e., “reactive” and “nonreactive” S1 states at physiological pH, but their origin remained unclear. In this study, we examined the S1 state dynamics of KR2 using femtosecond time-resolved absorption spectroscopy at different pH′s in the range from 4 to 11. It was found that the reactive S1 state is predominantly formed at pH >9, but its population drastically decreases with decreasing pH while the population of the nonreactive S1 state(s) increases. The pH dependence of the relative population of the reactive S1 state correlates very well with the pH titration curve of Asp116, which is the counterion of the protonated retinal Schiff base (PRSB) in KR2. This strongly indicates that the deprotonation/protonation of Asp116 is directly related to the generation of the multiple S1 states in KR2. The quantitative analysis of the timeresolved absorption data led us to conclude that the reactive and nonreactive S1 states of KR2 originate from KR2 proteins having a hydrogen bond between Asp116 and PRSB or not, respectively. In other words, it is the ground-state inhomogeneity that is the origin of the coexistence of the reactive and nonreactive S1 states in KR2. So far, the generation of multiple S1 states having a different photoreactivity of rhodopsins has been mainly explained with the branching of the relaxation pathway in the Franck− Condon region in the S1 state. The present study shows that the structural inhomogeneity in the ground state, in particular that of the hydrogen-bond network, is the more plausible origin of the reactive and nonreactive S1 states which have been widely observed for various rhodopsins.
■
yield in solution (0.2−0.3),25 which facilitates the high efficiency of photoinduced biological functions. Interestingly, the photoexcitation of PRSB in the protein often generates multiple S1 states which have been attributed to the “reactive” and “nonreactive” S1 states, and it has been widely believed that they originate from the branching of the relaxation pathway on the S1 potential energy surface.9,10,12,15,29−33 The reactive S1 state has been defined as the S1 state that undergoes isomerization and forms the K intermediate, whereas the nonreactive S1 state has been defined as the S1 state that shows no isomerization and produces no photoproduct. Because it is critically important to understand the effect and role of the protein matrix in controlling and optimizing the primary reaction of the chromophore, it is highly desirable to elucidate
INTRODUCTION
Rhodopsins constitute a group of photoreceptor proteins which are ubiquitous in various animals and microorganisms.1,2 They possess the protonated retinal Schiff base (PRSB) as the chromophore, and the photoexcitation of PRSB triggers photocycle reactions to realize various biological functions such as ion pumps and light sensing. In such photocycles, the initial photoexcitation generates the S1 state of PRSB, which subsequently undergoes ultrafast S1 → S0 relaxation inducing trans−cis isomerization of the chromophore. This photoisomerization process is the key primary event in the photocycle of rhodopsins, and hence it has been attracting much interest for many years.3−7 An intriguing feature of this process is that the isomerization of PRSB proceeds very quickly on a time scale of hundreds of femtoseconds in the proteins,8−20 although the isomerization in solution takes as long as 5 ps.21−25 Moreover, the isomerization quantum yield in proteins is significantly high (0.3−0.7)26−28 compared to the © XXXX American Chemical Society
Received: February 25, 2018 Revised: April 10, 2018
A
DOI: 10.1021/acs.jpcb.8b01934 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
isomerization, whereas the slower picosecond S1 decays are those of the nonreactive S1 states that relax directly to the initial S0 state without generating the photoproduct. As mentioned already, such multiple S1 states having different reactivities have been widely observed for various rhodopsins,9,10,12,15,29−33 but their origin remains unclear. Therefore, it is important to clarify the molecular origin of the reactive and nonreactive S1 states, not only for understanding the primary process of KR2 but also for elucidating the complex primary process of rhodopsins. In this study, we recorded femtosecond time-resolved absorption spectra of KR2 with changing pH, aiming to clarify the molecular origin of the multiple S1 states of KR2. It was reported that Asp116 of KR2, which acts as the counterion of PRSB, can have protonated and deprotonated states, and they coexist under the neutral condition.35 Thus, we particularly examined the effect of protonation/deprotonation of Asp116 on the ultrafast dynamics. The obtained data clearly show that the change in the protonation state of Asp116 drastically changes the photoisomerization reactivity of PRSB in KR2, indicating that the inhomogeneity of the ground state is the origin of the multiple S1 states exhibiting very different reaction yields.
the mechanism that makes the photoisomerization of PRSB distinct in proteins. KR2 is a newly found rhodopsin that realizes Na+ pumping. Although a large number of proton-pumping and anionpumping rhodopsins have been discovered so far, no cationpumping rhodopsin had been found before KR2.34,35 In fact, the discovery of KR2 was a big surprise because it had been thought that no cation-pumping rhodopsins exist in nature. Since PRSB is positively charged and hence is electrostatically repulsive to the cation, this electrostatic repulsion was thought to prevent the active cation transport through the Schiff base region. To elucidate the Na+-pumping mechanism, the photoreaction of KR2 has been studied by various methods, including IR,36 time-resolved visible,16,35,37−39 and Raman spectroscopies40 as well as solid-state NMR,41 X-ray crystallography,42,43 and pump activity measurements.35,44 On the basis of these studies, the following photocycle has been proposed for the neutral pH condition:16,35 Upon photoexcitation, PRSB having an all-trans configuration in the unphotolysed state undergoes trans−cis photoisomerization, and the J intermediate is generated as the first intermediate in the S0 state.16 Subsequently, the J intermediate is converted to the K, L, and M intermediates.35 In the M intermediate, the proton of PRSB is transferred to Asp116, which is the counterion of PRSB (Figure 1). With this proton transfer, Asp116 is directed
■
METHODS Sample Preparation. KR2 was overexpressed in Escherichia coli and purified with Ni2+-NTA affinity chromatography as described previously.35 The buffer compositions were carefully selected so that the sodium ion concentration is kept at 100 mM irrespective of the pH value as shown below: Table 1. Buffer Composition at Each pHa pH 4−7 8 9 10 11
Figure 1. Crystallographic structure of KR2 (3X3C). The retinal is shown with a yellow stick. The side chains of Asp116, Ser70, Asn112, Asp251, and Lys255 are also shown. Oxygen atoms and nitrogen atoms are shown in red and blue, respectively. The direction of the Na+ transport is also shown.
composition 50 50 50 50 50
mM mM mM mM mM
citrate-Tris, 100 mM NaCl, 0.05% DDM Tris-HCl, 100 mM NaCl, 0.05% DDM Tris-citrate, 100 mM NaCl, 0.05% DDM CAPS-NaOH, 89 mM NaCl, 0.05% DDM CAPS-NaOH, 62 mM NaCl, 0.05% DDM
a
Tris: tris(hydroxymethyl) aminomethane. CAPS: N-cyclohexyl-3aminopropanesulfonic acid. DDM: n-dodecyl-β-D-maltoside.
Femtosecond Time-Resolved Absorption Measurements. The apparatus of the femtosecond time-resolved absorption measurements has been described elsewhere.45 Briefly, a Ti:sapphire regenerative amplifier (800 nm, 1.2 mJ, 80 fs, 1 kHz, Legend Elite, Coherent) was used as the light source. The 0.2 mJ portion of its output was attenuated and focused into a 3-mm-thick calcium fluoride (CaF2) plate to generate a visible supercontinuum. The supercontinuum pulse was divided into two, and they were used as the probe and reference pulses. The remaining 1 mJ portion was used for driving an optical parametric amplifier (TOPAS-C, Light Conversion), and its signal output was frequency-doubled to generate the pump pulse centered at 575 nm. The polarization of the pump pulse was set at the magic angle (54.7°) with respect to the horizontally polarized probe pulse. Both the pump and probe pulses were focused into the sample solution in a flow cell (1 mm path length). The pump pulse energies were 240, 160, and 120 nJ/pulse for measurements at pH 11−6, 5, and 4, respectively, and the diameter of the pump beam was 250 μm at the sample position. The sample solution was continuously circulated during the measurements, and the flow speed was adjusted so that the sample solution in the excited volume was
toward Ser70 and Asn112,43 which sequesters the proton from the Na+ transport pathway, enabling Na+ to pass through the Schiff base region. After passing through the Schiff base region, Na+ binds to Asn112 and Asp251 on the extracellular side in the O intermediate.37 At the same time, the proton transfers back to the Schiff base, preventing the back flow of Na+. Finally, Na+ is released to the extracellular side, and the photocycle reaction of KR2 is completed. Very recently, we examined the ultrafast dynamics of KR2 under the physiological condition of pH 8 by femtosecond time-resolved absorption spectroscopy.16 We observed that the S1 → S0 stimulated emission, which reflects the population dynamics of the S1 state, exhibited a multiexponential decay: it consists of a rapid decay within 1 ps and slower decays occurring on a picosecond time scale. However, the transient absorption due to the photoproduct, the J intermediate, appeared within 1 ps and did not show any further rise on the picosecond time scale. This implies that the fast S1 decay within 1 ps corresponds to the deactivation of the reactive S1 state to form the J intermediate with trans−cis photoB
DOI: 10.1021/acs.jpcb.8b01934 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
PRSB and the resultant formation of the J intermediate. The peak of the photoproduct absorption is located at around 620 nm at 0.2 ps, but it subsequently shifts to 590 nm. This blue shift is attributable to the conversion from the J intermediate to the K intermediate.16 In bacteriorhodopsin, the vibrational cooling process of the chromophore, accompanied by the distortion of the chromophore, has been considered to take place in the J → K process.47,48 The high similarity between the spectral changes for the J → K conversion in BR and KR2 strongly suggests that similar relaxation processes occur in KR2. At 200 ps, the transient band due to the K intermediate peaks at around 590 nm, and no further spectral change was observed within the delay time range of our measurements (400 ps). Figure 2c depicts time-resolved absorption spectra measured at pH 4. Similar to the data obtained at pH 11, we observed the excited-state absorption at around 400−530 nm, the groundstate bleaching at around 540−620 nm, and the stimulated emission in the >650 nm region at 0.1 ps. However, the stimulated emission decays on a time scale of a few tens of picoseconds, which is significantly slower than the stimulated emission decays at pH 11. This implies that the S1 state deactivates much more slowly at pH 4. Another significant difference from the data at pH 11 is the weakness of the photoproduct absorption. In fact, because of the longer lifetime of stimulated emission and the smaller amplitude of photoproduct absorption, the time-resolved absorption signal in the >600 nm region is negative up to 1 ps. Only in the spectra after 10 ps does photoproduct absorption become noticeable as a positive band peaked at around 620 nm. The amplitude of the photoproduct absorption band is ∼1/3 of the corresponding band at pH 11, indicating a much lower yield of the photoproduct at pH 4. Femtosecond time-resolved absorption spectra measured at pH 7 exhibit the spectral change between those observed at pH 11 and 4, as shown in Figure 2b. The femtosecond time-resolved absorption measurements at different pH′s reveal that the photochemical dynamics of PRSB in KR2 significantly change with the change in pH: as the pH value becomes lower, the apparent lifetime of the S1 state becomes longer and the photoproduct yield (and hence the isomerization quantum yield) becomes lower. pH Dependence of the Temporal Traces. To obtain more insight into the pH dependence of the S1 dynamics, we analyzed the temporal traces of the time-resolved absorption at four typical wavelengths. Figure 3a−d depicts the transient signals at 460, 550, 650, and 720 nm observed in the pH range from 4 to 11. These wavelengths correspond to the wavelength region of the excited-state absorption, ground-state bleaching, photoproduct absorption, and stimulated emission. The pH dependence of time-resolved absorption can be seen more clearly in these temporal traces. For instance, the stimulated emission signal at 720 nm observed at pH 11 exhibits a bimodal decay consisting of a major ultrafast component vanishing within 1 ps and a minor slow component decaying on a picosecond/subnanosecond time scale. Upon decreasing the pH value, the amplitude of the ultrafast decay component becomes smaller while the slow decay components become dominant. To obtain quantitative information, we performed a fitting analysis of these four temporal traces. The fitting was done using a sum of exponential functions with common time constants for each pH value (i.e., global fitting analysis). The multiexponential function was convoluted with the Gaussian function with a 0.1 ps fwhm to take into account the time
replaced between the laser shots. The probe pulse, after passing through the sample as well as the reference pulse, was spectrally analyzed by a polychromator (500is/sm, Chromex) equipped with a CCD camera (PIXIS-400F, Princeton Instruments). The group delay dispersion of the probe pulse was determined by the optical Kerr effect (OKE) measurement on the buffer solution,46 and it was used to determine the time origin at each probing wavelength. The time resolution was evaluated to be ca. 100 fs on the basis of the fwhm of the OKE signal.
■
RESULTS AND DISCUSSION pH Dependence of the Femtosecond Time-Resolved Absorption Spectra. In a previous study, Inoue et al. examined the pH dependence of the steady-state absorption spectrum of KR2.35 They found that when the pH value is decreased to 650 nm region. The stimulated emission decays within 1 ps, indicating that most of the S1 population deactivates in the femtosecond time region. Concomitantly with the decay of the stimulated emission, a photoproduct absorption band shows up at around 620 nm, which indicates all-trans-to-13-cis photoisomerization of the C
DOI: 10.1021/acs.jpcb.8b01934 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
components were assigned to the deactivation of the nonreactive S1 states. The τ0 component is attributable to the relaxation from the Franck−Condon state to the quasistationary reactive and nonreactive S1 states. The τ2 component corresponds to the blue shift of the photoproduct absorption, which appears as a decrease in the absorption signal at 650 nm (Figure 3c) and is assignable to the J → K conversion. The τ5 component represents the decay of the long-lived K intermediate, but it is treated as infinity because the corresponding signal change cannot be recognized due to the microsecond time constant of the K → L conversion process within the delay time range of the present measurements ( 7, showing that only the τ1 component contributes to photoproduct formation.16 We note that, at pH
