3.2. Energy and electron transfer dynamics of PSII-dimer excited at 670 nm
Figure 2(A) presents the photo-induced absorption spectra of PSII-dimer at 140 K given by excitation to the Qy state of Chl a at 670 nm. To avoid the appearance of the annihilation processes between excited Chl a molecules in excited state dynamics, the excitation intensity was set to 3 µW (corresponding to 6×1013 photon/cm2), which is sufficiently low intensity to suppress annihilation processes in PSII-dimer (Müller, et al. 1996; Yoneda, et al. 2016). The negative absorbance changes around 670 nm appeared immediately after photo-excitation. These signals include both the ground state bleaching and the stimulated emission of Chl a bound to PSII-dimer. The negative signal peaks at 678 nm at 0.1 ps and slightly shifts to longer wavelength (683 nm) with delay times until 10 ps. On the other hand, the negative peak again shifted to 678 nm at a delay time of 3000 ps. The broad positive signal appeared in the visible region of the spectrum and can be assigned to the transient absorption from the Qy state of Chl a to the higher-lying excited state (Kosumi, et al. 2018; Niedzwiedzki and Blankenship 2010). Further, the negative signal at 542 nm was observed at delay times of 100 and 3000 ps. This signal has been assigned to the Qx bleaching of Pheo reduced in a primary charge separation (Durrant, et al. 1993; Greenfield, et al. 1997; Hastings, et al. 1992; Holzwarth, et al. 2006). The Pheo bleaching appears at a delay time of 10 ps and slowly rises until 100 ps.
Figure 2(B) shows the kinetic traces of the photo-induced absorption signals of the PSII-dimer at 140 K excited at 670 nm. The solid lines in Fig. 2(B) are the best-fit curves for the rise and decay phases convoluted with the instrument response function assuming a Gaussian temporal profile. The obtained kinetics traces of the photo-induced absorbance changes were analyzed globally, and then five exponential components were required to fit the kinetics traces fully. The kinetic trace at 668 nm rises instantaneously, and initially decays with the time constant of 0.71 ps. The Qy bleaching signal at 680 nm rises with the time constant of 0.71 ps and multi-exponentially decays. In the 542 nm kinetic trace, the transient absorption due to the Qy state of Chl a appear immediately after excitation. The signal translated to a negative phase corresponding to the Pheo Qx bleaching with a time constant of several ten picoseconds.
To visualize the detailed electron and energy transfer dynamics in PSII-dimer, we show the decay-associated difference spectra (DADS) at 670 nm in Fig. 3. The observed signals in the Chl a Qy bleaching region are mainly negative part corresponding to the ground state bleaching and stimulated emission. Thus, negative part of DADS represents decay of the signal (recovery of absorption bleaching here) while positive part corresponds to rise of signal elsewhere (onset of bleaching there) with the same time constant. At 140 K, the first DADS (0.71 ps) denotes energy transfer from a Chl a site with 670 nm absorption (negative signal) to a site with 682 nm absorption (positive signal). The second DADS (3.28 ps) exhibits the negative peak at 680 nm. The second DADS also has the positive peak at 688 nm but its amplitude is small. This positive signal is assignable to an energy transfer component from a Chl a site with 680 nm absorption to a site with 688 nm absorption or a characteristic transient absorption of a Chl a site with 680 nm absorption. The third (43.0 ps) and fourth (358 ps) DADS have the negative peak at 682 nm and 684 nm, respectively. The last DADS (non-decay) has the negative peak at 678 nm and the transient absorption in both blue and red sides. Then, the last DADS exhibits the bleaching signal of the Pheo Qx band at 542 nm. The observed transient absorption of 650 ~ 800 nm is characteristic for pigment-protein complexes and has not been detected for isolated Chl a in solution (Kosumi, et al. 2018; Niedzwiedzki and Blankenship 2010). This absorption band has been assigned to PheoD1− and PD1+ (Durrant, et al. 1993), thus the last DADS corresponds to the charge separated P680+Pheo− state. The third DADS involves the rise component of the Pheo Qx bleaching at 542 nm, suggesting that the charge separation takes place with a time constant of 43.0 ps. The fourth DADS has neither rise nor decay component of the Pheo Qx bleaching at 542 nm. It has been reported that the secondary electron transfer from Pheo− QA to Pheo QA− takes place in a timescale of several hundred ps (Groot, et al. 2005; Holzwarth, et al. 2006; Schatz, et al. 1988). However, the absence of decay component of the Pheo Qx bleaching in the fourth DADS demonstrate that the 358 ps component does not originate from the secondary electron transfer to QA. The fourth DADS exhibits the longest-wavelength (684 nm) absorption band in the Qy region of Chl a. It has been well-known that the CP47 core complex from higher plants contributes to the longer-wavelength fluorescence (F695) (Acharya, et al. 2010; Andrizhiyevskaya, et al. 2004; Komura, et al. 2006; Mohamed, et al. 2016). Shibata et al., have suggested that the lowest-energy site in the CP47 core complex from thermophilic cyanobacteria act as an energy sink (designated as C695, see Fig.S3 in the Supplemental Information) and that this sink site can donate an energy to a reaction center at temperature only above 77 K (Shibata, et al. 2013). Indeed, the fourth DADS exhibits the positive component at 666 nm, implying an uphill energy transfer from a 684 nm site to a 666 nm site. Therefore, the forth DADS is assignable to the sink site corresponding likely to C695 in the CP47 core complex.
The energy and electron transfer dynamics of PSII-dimer measured at 296 K are essentially identical with that at 140 K as shown in Fig. 3(B) (see also Fig. S2 in the Supplementary Information). The observed kinetics were analyzed by six exponential decay/rise components and the obtained time-constants at 296 K are shorter than those at 140 K. The third DADS (14.8 ps) at 296 K shown in Fig. 3(B) exhibits the rise component of the Pheo Qx bleaching at 544 nm. This observation demonstrates that the primary charge separation takes place with 14.8 ps at 296 K. It should be noted that ChlD1 has been recently suggested to be a primary electron donor (Groot, et al. 2005). However, the bleaching peak of Chl a Qy in the fifth DADS corresponding to the charge separated state was observed at 680 nm (296 K) or 678 nm (140 K), which agrees well with the PD1 bleaching reported previously (Lozier and Butler 1974). When the negative signal of the third DADS at 682 nm (140 K or 296 K) is assignable to ChlD1*, the electron transfers of ChlD1*PheoD1 → intermediate charge separated state → PD1+ PheoD1− take place within 43.0 ps at 140 K (14.8 ps at 296 K). Especially, it is likely that the electron transfer time of ChlD1*PheoD1 → intermediate charge separated state is shorter than the ChlD1* formation time (0.71 ps or 3.28 ps) because the intermediate state was temporally and spectrally hidden (see Fig. S4 in the Supplementary Information). It has been reported that the absorption band of was ChlD1 located around 685 nm that is slightly longer wavelength than PD1 (Judd, et al. 2020; Raszewski, et al. 2008). Based on these results, it is most likely that the bleaching at 682 nm in the third DADS originates from ChlD1 and that ChlD1 is the primary electron donor in PSII.
A kinetic compartment model of PSII-dimer excited at 670 nm is summarized in Fig. 4. The Pheo Qx at 140 K seems not to decay in the temporal range of 3 ns in this study, while the fourth DADS at 296 K (650 ps) exhibits the bleaching-like signal at 544 nm. This signal is likely to be assigned to the decay of the Pheo Qx bleaching signal implying secondary electron transfer to QA at 296 K with the time constant of 650 ps (see the detail in the Supplementary Information). Electron paramagnetic resonance (EPR) spectroscopic investigations have revealed that the presence of the photoinduced QB− state even at 77 K (De Causmaecker, et al. 2019; Fufezan, et al. 2005). Therefore, the secondary electron transfer to QA should occur at 140 K even though its electron transfer dynamics was not observed in this study due to its longer time-constant.
3.3. β-car→Chl a energy transfer dynamics of PSII-dimer excited at 500 nm
Next, we examined roles of β-car in photosynthetic functions in PSII-dimer. Figures 5(A) and (B) show the photo-induced absorption spectra and their kinetic traces of the PSII-dimer at 140 K after excitation into the S2 state of β-car at 500 nm (see also the photo-induced absorption spectra and their kinetic traces of the PSII-dimer at 296 K in the “Supplemental Information”). In the photo-induced absorption spectra, the negative change appeared below 530 nm at 0.1 ps. This change originates both from the bleaching due to the depletion of the S0 ground state and the stimulated emission from the S2 excited state of β-car. Then, the positive absorption band was observed at 580 nm, which increased and shifted to shorter wavelength until 1.0 ps. This signal has been assigned to the transient absorption from S1 to the higher-lying Sn state (Kosumi, et al. 2009; Polívka and Sundström 2009). The S1-Sn transient absorption band was broad at early delay times, then its band became narrower and its peak shifted to shorter wavelength owing to the vibrational relaxation (Billsten, et al. 2005; Kosumi, et al. 2005; Polívka and Sundström 2009). Further, the bleaching of Chl a Qy was observed at 680 nm, suggesting the β-car→Chl a energy transfer.
The solid lines in Fig. 5(B) are the best-fit curves for the rise and decay phases convoluted with the instrument response function assuming a Gaussian temporal profile. The obtained kinetics traces were analyzed globally, and six exponential components were required to fit the kinetics traces fully. The kinetic trace of the S1-Sn transient absorption of β-car at 570 nm rises with the time-constants of 0.10 and 0.55 ps. The former corresponds to the S2 lifetime of β-car bound to PSII and the latter is assignable to the vibrational cooling time in S1. The kinetic trace of the S1-Sn transient absorption decays with two time-constants of 9.91 and 24.3 ps. The amplitude of the 24.3 ps component is rather small in the trace probed at 570 nm, thus the time-constants of 9.91 ps can be assigned to the S1 lifetime of β-car bound to PSII. It has been reported that the S1 lifetime of β-Car in solution was about 9 ps at room temperature (Jailaubekov, et al. 2011; Kosumi, et al. 2005) and 15 ps at 77 K (Jailaubekov, et al. 2011; Ostroumov, et al. 2011). Then, the kinetic trace of the Chl a Qy bleaching as shown in Fig. 5(B) has no rise component corresponding to the S1 decay, suggesting that the S1→Qy energy transfer channel is closed in PSII-dimer.
The DADS of PSII-dimer excited at 500 nm at 140 K are represented in Fig. 5(C). The first DADS (0.10 ps) exhibits a pronounced negative signal ranging from 450 to 650 nm. This negative signal contains the S0 ground state bleaching, the stimulated emission of S2, and the rise component of the S1 transient absorption of β-car. Then, the positive change slightly appears at 660 ~ 680 nm corresponding to the opposite change of the Chl a Qy bleaching. Therefore, the result suggests the energy transfer from β-Car S2 to Chl a. The second (0.55 ps) and third (9.91 ps) DADS correspond to the hot S1 and S1 of β-Car, respectively. It should be noted that the positive signal at 682 nm in the second DADS does not exhibit the hot S1→Chl a energy transfer. The positive spectral component at 682 nm with the time-constant of 0.55 ps corresponds to the decay at 660 nm in the second DADS. Indeed, the energy transfer between Chl a in the core complex (CP43) was observed with the time-constant of 0.71 ps after excitation at 670 nm as shown in Fig. 3(A). The fourth DADS (24.3 ps) contains the S0 ground state bleaching of β-Car and the broad transient absorption ranging from 510 nm to 590 nm. This spectral feature has been reported as the postulated S* state of carotenoids (Polívka and Sundström 2009; Polivka and Sundstrom 2004). The nature of the S* state has been a matter of controversial yet, while the S* state was assumed to be an excited state or the hot ground state.
The fifth (416 ps) and sixth (non-decay component) DADS exhibit the broad transient absorption band in the 450 ~ 600 nm spectral region (see also Fig. S7). This transient absorption was observed after excitation at 670 nm as shown in Fig. 6, thus this signal is likely to be assigned to PD1+PheoD1− and long-lived Chl a bound to the core-complexes. On the other hand, the sixth DADS excited at 500 nm exhibits the S0 ground state bleaching of β-Car, suggesting the involvement of the β-Car excited state. Therefore, we examined the difference between sixth DADS (500 nm exc.) and fifth DADS (670 nm) at 140 K as shown the inset in Fig. 6 to extract the contribution of β-Car from the six DADS excited at 500 nm. The difference spectrum shows the S0 ground state bleaching of β-Car and the transient absorption band at 538 nm that is distinct from S1. This transient absorption has been assigned to the transition from the triplet excited T1 state to a higher-lying triplet Tn state of β-Car (Groot, et al. 1995). The T1-Tn transient absorption band of β-Car has been reported to be peaked at 530 nm in the CP47 core complex at 40 K (Groot, et al. 1995), while to be 500 ~ 530 nm in solution at room temperature (Fujii, et al. 2000; Jailaubekov, et al. 2011). It has been reported that carotenoid T1 is generated by the triplet-triplet energy transfer from (B)Chl T1 following the (B)Chl Qy→(B)Chl T1 intersystem crossing in photosynthetic pigment-protein complexes. In this process, carotenoid T1 appears after several ten nanoseconds following photo-excitation to (B)Chl Qy (Kosumi, et al. 2016; Kvicalova, et al. 2016; Mandal, et al. 2017; Niedzwiedzki, et al. 2020). However, the spectral characteristics of β-Car T1 in PSII-dimer appeared at least 100 ps (see Fig. S6 in the “Supplemental Information”). The ultrafast triplet generation of carotenoids via S* has been observed in purple bacterial antenna complexes (Gradinaru, et al. 2001; Papagiannakis, et al. 2002; Simova, et al. 2022), while the recent investigations have suggested the singlet fission of Car S2 or Bchl Qy to Car T1 (Niedzwiedzki, et al. 2017; Uragami, et al. 2020). Moreover, Jailaubekov et al., has reported the ultrafast triplet state generation directly from S2 of β-Car in solution (Jailaubekov, et al. 2011), while the observation cannot be explained by a singlet fission due to a low molecular density. In PSII-dimer, it is most likely that the triplet state of β-Car was generated directly from S2, because such triplet generation of β-Car was not observed after excitation at 670 nm.
A kinetic model for excited states and energy transfer dynamics of β-Car is illustrated in Fig. 7. The S2 lifetime of β-Car has been reported to be 133 fs at 77 K (Jailaubekov, et al. 2011) and 120 ~ 175 fs at 80 K (Akimoto, et al. 2002). Thus, we assumed the intrinsic S2 lifetime of β-Car bound to PSII to be 150 fs both at 140 and 296 K. On the basis of this assumption, the excitation energy transfer rate and efficiency from S2 of β-Car to Chl a in PSII-dimer are determined to be (300 fs)−1 an 30 % both at 140 and 296 K. As mentioned before, the S1 energy transfer channel is inactive in PSII-dimer, thus the overall energy transfer efficiency from β-Car to Chl a is determined to e 30 %, which agrees well with that determined by fluorescence excitation measurements (de Weerd, et al. 2003; Kwa, et al. 1992). The kinetic model shown in Fig. 7 considers the S2→T1 intersystem crossing (or a singlet fission), while this process does not affect on the S2→Chl a energy transfer efficiency due to the low T1 yield (~ 10− 3) (Jailaubekov, et al. 2011). The nature of the S* state has been a matter of controversial and this study did not clarify whether S* originated from an excited state or the hot S0 ground state. Therefore, we tentatively assigned S* with the decay time-constant of 24.3 ps at 140 K (23.6 ps at 296 K) to the hot S0 ground state, as the previous works postulated (Kosumi, et al. 2012; Lenzer, et al. 2010). Comparison of the fifth (416 ps) and sixth (inf.) DADS around the 500 nm region of the spectra (Figs. 5(C) and S5(A)) implies that these DADS associate to each other. Thus, the fifth DADS excited at 500 nm is likely to be representable to the vibrationally hot T1 state.