Introduction

The unique light-harvesting antennas of oxygenic photosynthetic organisms contain pigments such as chlorophyll (Chl), carotenoid, and phycobilin. The light energy captured by these antennas is transferred to the reaction centers of the photosystems (PSs), where photochemical reactions proceed. A photochemical reaction in the reaction center induces electron flow from water to NADP+, generating a proton-motive force across the thylakoid membranes for ATP synthesis. Efficient oxygenic photosynthesis relies on the proper control of energy transfer from the light-harvesting antennas to both PSs (PSI and PSII).

Cyanobacteria and red algae possess a large light-harvesting antenna called phycobilisome (PBS) on their thylakoid membrane surfaces (Gantt 1981). PBS comprises an allophycocyanin (APC) core surrounded by phycocyanin (PC)-containing rods, which absorb visible light. The PC may be pure or combined with phycoerythrin (PE) (Gantt 1981; Adir 2005). Although the morphology of PBS differs between cyanobacteria and red algae (Arteni et al. 2008, 2009), the energy transfer efficiency of PBS in both taxa approaches 100% (Glazer 1984). PBS is generally known as the light-harvesting antenna of PSII. The energy transfer pathways from PBS to PSII have been investigated in fluorescence measurements of CP43-deficient and CP47-deficient mutants of Synechocystis (Bittersmann and Vermaas 1991; Shimada et al. 2008). However, as neither of these mutants transfer energy from PBS to PSII (PBS→PSII transfer) (Bittersmann and Vermaas 1991; Shimada et al. 2008), the details of the energy transfer remain unclear. PBS can supply light energy to PSI under intense light conditions, or under light that excites PSII (Mullineaux 2008). Energy transfer from PBS to PSI is thought to occur by two pathways: direct energy transfer from PBS to PSI (PBS→PSI transfer) (Allen et al. 1985) or indirect energy transfer from PBS to PSI via PSII (spillover after PBS→PSII transfer, i.e., PBS→PSII→PSI transfer) (Murata 1969). The occurrences of both transfers have been studied by various methods such as steady-state fluorescence spectra at 77 K (Bruce et al. 1989), photooxidation rates of P700 (Ley and Butler 1977), electron microscopy (Olive et al. 1997), and fluorescence recovery after photobleaching (Mullineaux et al. 1997). However, the energy transfer processes from PBS to PSI remained incompletely understood.

Recently, we demonstrated the existence of PSII→PSI transfer by measuring the delayed fluorescence (Yokono et al. 2011), which originates from charge recombination at the PSII reaction center (Mimuro et al. 2007). More recently, we measured the delayed fluorescence spectra under PBS-selective excitation, and confirmed the occurrence of PBS→PSII→PSI transfer in vivo in cyanobacteria; Synecocystis sp. PCC 6803 (hereafter referred to as Synechocystis), Synechococcus sp. PCC 7002 (Synechococcus), and the red alga Cyanidioschyzon merolae (Ueno et al. 2016). Such interactions were never observed in isolated PBS–PSII–PSI megacomplexes (Liu et al. 2013). However, 5–7 ns after 400-nm excitation of cyanobacterial cells, we observed shorter PSI peak wavelength in the delayed fluorescence than in the time-resolved fluorescence spectrum (TRFS) (Akimoto et al. 2014). Therefore, the PBS→PSI and PBS→PSII→PSI pathways may transfer light energy to PSI Chls with different energy levels. In the present study, we examined the energy transfer processes from PBS to PSs in various cyanobacterial and red algal cells. To this end, we supplemented our previous samples with the cyanobacterium Arthrospira platensis, and a wild-type (WT) and a green mutant (GM) of the red alga Pyropia yezoensis. A. platensis exhibits two peaks around 730 and 760 nm in the PSI fluorescence region; the latter is attributed to PSI trimers (Shubin et al. 1993). Unlike the PBS of C. merolae, the PBS of P. yezoensis contains APC, PC, and PE, of which the PE contents are reduced in the mutant (Niwa et al. 2009). We also selectively excited the PSs or PBS of the above organisms, and measured and compared their resulting time-resolved fluorescences at 77 K. Fluorescence peaks from the different components are better-resolved at 77 K than at room temperatures since there is less thermal energy to help excitons on long-wavelength pigments escape at 77 K.

Materials and methods

Cultivation conditions

Cyanobacteria Synechocystis and Synechococcus, and red alga C. merolae 10D were grown under previously described conditions (Ueno et al. 2016). The growth conditions of the cyanobacterium A. platensis NIES39 are also described elsewhere (Akimoto et al. 2012). In the present study, we employed the WT strain (HG-5) and the GM strain (IBY-G3) of Pyropia yezoensis. The WT was a pure line of cultivated P. yezoensis isolated by the method of Niwa et al. (2004). The GM was isolated from a green-cell cluster of the WT blade mutated by carbon-ion beam (50 Gy, 23 keV/μm) irradiation as described in Niwa et al. (2009). Gametophytic blades of WT and GM were laboratory cultured under the conditions described in Niwa et al. (2009).

Measurements and analyses

Steady-state fluorescence spectra were measured by a spectrofluorometer (JASCO FP-6600/PMU-183) operated at 77 K (Ueno et al. 2015). The excitation wavelength was set to 408 nm (exciting PSs) or 635 nm (exciting PBS). As the steady-state fluorescence spectra of P. yezoensis, we used the time-integrated fluorescence spectra constructed by integrating the TRFS. The TRFS were measured by a time-correlated single-photon counting system at 77 K (Akimoto et al. 2012), with the excitation wavelength set to 408 or 633 nm (exciting PSs or PBS, respectively). The repetition rate of the pulse train was 5 MHz. To accurately evaluate the delayed fluorescence, we set the time interval of data acquisition to 24.4 ps/channel (total time window = 100 ns). The fluorescence kinetics were measured at 1-nm intervals (675–750 or 675–775 nm in the case of A. platensis). To accurately evaluate the fluorescence kinetics, we also measured the fluorescence rise and decay curves up to 10 ns (2.44 ps/channel × 4096 channels), which covers the delayed fluorescence peaks of PSI and PSII. The fluorescence rise and decay curves at different detection wavelengths were obtained at different excitation wavelengths, and globally fitted by sums of exponentials with common time constants as follows:

$$F(t,\,\lambda )=\sum\limits_{n=1}^{7}{{{A}_{n}}(\lambda )\exp \,\left( -\frac{t}{{{\tau }_{n}}} \right)}.$$
(1)

Here, A n (λ) is the amplitude at the detection wavelength λ and τ n is the time constant.

For both steady-state and time-resolved fluorescence measurements, we obtained homogenous ice at 77 K by adding polyethylene glycol (average molecular weight 3350, final concentration 15% (w/v), Sigma-Aldrich, St Louis, MO, USA) to the sample solutions (Mimuro et al. 1989). Prior to measurements, all samples were dark adapted for ~20 min.

Results

Steady-state fluorescence spectra

Figure 1 shows the steady-state fluorescence spectra of various cyanobacteria and red algae after selective excitation of PSs and PBS. The spectra are divided into two fluorescence regions: PSII (675–700 nm) and PSI (longer than 700 nm). In the PSII fluorescence region, the peaks around 685 and 695 nm are assigned to CP43 and CP47, respectively (Murata and Satoh 1986). The PSI fluorescence region generally exhibits a single peak; the exception is A. platensis, which yields two fluorescence bands around 730 and 760 nm. The latter band derives from PSI trimers (Shubin et al. 1993). Although PBS-selective excitation induces detectable PSI fluorescence, it yields a lower fluorescence intensity ratio of PSI to PSII than PSs-selective excitation. The two peaks in the PSI fluorescence region of the A. platensis spectrum indicate that in this species, energy is transferred from PBS to both monomers and trimers of PSI.

Fig. 1
figure 1

Steady-state fluorescence spectra of Synechocystis sp. PCC 6803 (a), Synechococcus sp. PCC 7002 (b), Arthrospira platensis (c), Cyanidioschyzon merolae (d), and Pyropia yezoensis wild-type (e) and green mutant (f) at 77 K. Each spectrum is normalized by its maximum intensity. Blue and red lines indicate PSs- and PBS-selective excitation, respectively. The Pyropia yezoensis spectra are the time-integrated fluorescence spectra. Labels in each spectrum indicate the PSI fluorescence peak position

Full size image

Time-resolved fluorescence spectra

To examine the energy transfer processes from PBS to PSI and PSII, we measured the TRFS after excitation at 408 or 633 nm. Figure 2 shows TRFS of the various cyanobacteria and red algae after excitation at both wavelengths. The peak assignments are essentially those of the steady-state fluorescence spectra. The 408-nm excitation is mainly absorbed by PSs, indicating energy transfer between and within PSI and PSII. Immediately after the PSs-selective excitation (blue lines in Fig. 2), two bands are observed, but the PSI band is more intense than the PSII band. The CP47 fluorescence band in the PSII fluorescence region clearly persists over time, whereas the PSI fluorescence band shifts to longer wavelengths, reaching its longest wavelength band after 3.5–4.0 and 9.4–11 ns in cyanobacteria and red algae, respectively. In A. platensis, the fluorescence intensity around 760 nm increases after 610–730 ps, forming a clear band that persists over time. The TRFS at 47–76 ns are the delayed fluorescence spectra derived from charge recombination at the PSII reaction center (Mimuro et al. 2007). In previous studies of isolated PSI and light-harvesting antennas, no delayed fluorescence was observed in the time region of the present study (Mimuro et al. 2010; Yokono et al. 2015a). Therefore, in the absence of PSII→PSI transfer, the delayed fluorescence is contributed by PSII alone. Conversely, PSII→PSI transfer will manifest as delayed fluorescence in the PSI fluorescence region (Yokono et al. 2011). The observed PSI fluorescence band in the delayed fluorescence spectra indicates that PSII→PSI transfer occurs in our algae.

Fig. 2
figure 2

Time-resolved fluorescence spectra of Synechocystis sp. PCC 6803 (a), Synechococcus sp. PCC 7002 (b), Arthrospira platensis (c), Cyanidioschyzon merolae (d), and Pyropia yezoensis wild-type (e) and green mutant (f) at 77 K. The excitation wavelength was 408 nm (blue lines) or 633 nm (red lines). Zero time corresponds to the time of maximum intensity of the excitation laser pulse. Each spectrum is normalized by its maximum intensity. Numbers in each panel indicate the magnification factors relative to the most intense spectrum. Vertical lines indicate the delayed fluorescence peaks of PSI. The PSI fluorescence peak positions are displayed in TRFS at −24–0 ps, and 3.5–4.0 and 9.4–11 ns in cyanobacteria and red algae, respectively

Full size image

The 633-nm laser excitation is mainly absorbed by PBS. At this excitation wavelength, the TRFS reflect the energy transfer from PBS to PSs and subsequent energy transfer between and within PSI and PSII. Immediately after the PBS-selective excitation (red lines in Fig. 2), PSI and PSII bands are detected, and CP43 fluorescence appears in the PSII fluorescence region of all spectra except that of A. platensis. The fluorescence intensities of CP43 and CP47 are comparable in A. platensis. The CP47 fluorescence band is detectable after 200 ps in Synechocystis, Synechococcus, and P. yezoensis, whereas in C. merolae, the PSII fluorescence overtakes the PSI fluorescence after 200 ps and the CP47 fluorescence band is clearly observed after 610 ps. Meanwhile, the PSI fluorescence band is redshifted over time, and its longest wavelength band coincides with the wavelength of PSs-selective excitation. The TRFS at 47–76 ns after PBS-selective excitation are also delayed fluorescence spectra (Mimuro et al. 2007). Previously, we identified PBS→PSII→PSI transfer in Synechocystis, Synechococcus, and C. merolae by observing PSI fluorescence in their delayed fluorescence spectra after PBS-selective excitation (Ueno et al. 2016). Here, we also recognize PBS→PSII→PSI transfer in A. platensis and P. yezoensis. The delayed fluorescence spectra in the PSI fluorescence region after PBS-selective excitation peak at the wavelength of PSs-selective excitation, indicating that the spillover pathway is conserved regardless of connection or disconnection of PBS and PSII.

Immediately after the PBS-selective excitation, Synechocystis, Synechococcus, and C. merolae exhibit a broader band in the PSI fluorescence region than under PSs-selective excitation. The peak in the A. platensis and P. yezoensis WT spectra appears at shorter wavelengths than in the PSs-selective excitation, whereas the P. yezoensis GM spectrum exhibit no clear differences. In addition, the fluorescence intensity ratio of CP47 to CP43 in the delayed fluorescence is higher under PBS-selective excitation than under PSs-selective excitation, except in C. merolae.

Fluorescence rise and decay kinetics

To evaluate the fluorescence kinetics in PSI and PSII, we globally analyzed the measured fluorescence rise and decay curves at different detection wavelengths for different excitation wavelengths. The curves were fitted by seven exponentials with common time constants. The results are listed in Table 1. After PBS-selective excitation, the two shortest-lived components (τ 1 and τ 2) typically took negative amplitudes, indicating that CP43, CP47, and PSI receive energy from PBS in both fast phase (τ 1) and slow phase (τ 2). The fast-phase contributions are negligible only in A. platensis, and fast-phase energy transfer to CP43 is absent in C. merolae. However, the time constant of the slow phase was shorter in C. merolae than in P. yezoensis. Therefore, compared with P. yezoensis, the efficiency of PBS→CP43 transfer in C. merolae is moderately large. Synechococcus alone exhibits a fast phase in PBS→CP47 transfer. Fast energy transfer to PSI operates only in C. merolae and P. yezoensis GM. The τ 4 and τ 5 components represent the fluorescence lifetimes of red Chls with different energy levels. The τ 5 component probably originates from the red Chl with lowest energy, because its PSI fluorescence peak was redshifted over time. The τ 4 and τ 5 lifetimes differ between cyanobacteria (~500–750 ps and ~1.5 ns, respectively) and red algae (~1 and ~2 ns, respectively). These differences may reflect the environmental differences around red Chl in the two organisms.

Table 1 Time constants of the 1st to 5th components, and their amplitudes, after global fitting of the fluorescence rise and decay curves after 408- or 633-nm excitation (Ex 408 or Ex 633, respectively) of Synechocystis sp. PCC 6803, Synechococcus sp. PCC 7002, Arthrospira platensis, Cyanidioschyzon merolae, and Pyropia yezoensis wild-type (WT) and green mutant (GM)
Full size table

Discussion

Energy transfer from PBS to PSII

Immediately after PBS-selective excitation, differences between the CP43 and CP47 bands were observed. The spectra of all algae except A. platensis displayed a CP43 fluorescence band, whereas the fluorescence intensities of CP43 and CP47 were comparable in A. platensis. This result confirms PBS→CP43 transfer as the main pathway in Syenchocystis, Synechococcus, C. merolae, and P. yezoensis. In A. platensis, the energy transfer was almost equally split between PBS→CP43 and PBS→CP47. The intensity of the CP47 fluorescence band of A. platensis increased over time, as reported for the filamentous cyanobacterium Halomicronema hongdechloris (Tomo et al. 2014).

Delayed PSII fluorescence originates from PSII complexes that do not connect with PSI (i.e., PSII complexes and PBS–PSII supercomplexes) (Yokono et al. 2011). The information from delayed florescence spectra can be altered by changing the excitation wavelength. Under PSs-selective excitation, the delayed florescence spectra characterize the PSII complexes and PBS–PSII supercomplexes, whereas under PBS-selective excitation, they mainly characterize the PBS–PSII supercomplexes. Apart from C. merolae, the delayed fluorescence intensity at CP47 to CP43 is larger under PBS-selective excitation (c.f. PSs-selective excitation) (TRFS at 47–76 ns, Fig. 2). In their electron microscopy studies, Mörschel and Schatz (1987) reported that PBS mainly combines with the PSII dimer. Strong interaction between the PSII monomers induces energy transfer from CP43 to its neighboring CP47 (Yokono et al. 2012). Our study reveals that light energy flows from CP43 to CP47 by energy transfer between the neighboring PSII monomers in intact PBS–PSII supercomplexes. Energy transfer to CP47 might relate to spillover, as suggested previously (McConnell et al. 2002). In contrast to other algae, the CP43 and CP47 of C. merolae fluoresced with almost identical intensity under PBS-selective excitation. Although the CP43 fluorescence in C. merolae is large in the early time region (TRFS at 200–240 ps, Fig. 2), energy transfer between neighboring PSII monomers functions in this red alga.

At 77 K, excitation energy is transferred from higher-energy pigments to lower-energy pigments, and fluorescence is mainly emitted from terminal energy traps. On the other hand, up-hill energy transfer occurs at physiological temperatures. Therefore, energy is shared among CP43, CP47, and the reaction center via energy transfer between neighboring PSII monomers. Interactions between the PSII monomers in the PSII dimer change depend not only on species but also on pH (Yokono et al. 2015b). Therefore, by controlling the interactions between the PSII monomers in their PSII dimers, the algae might effectively perform charge separation at their reaction centers.

Energy transfer from PBS to PSI

Energy transfer from PBS to PSI occurs by two pathways: PBS→PSI transfer (Allen et al. 1985) and PBS→PSII→PSI transfer (Murata 1969). The delayed fluorescence spectra of all algae (TRFS at 47–76 ns, Fig. 2) exhibit a clear PSI fluorescence band, indicating that PBS→PSII→PSI transfer occurs in all of the investigated algae. Under PBS-selective excitation, the fluorescence intensity ratio of PSI to PSII in the delayed fluorescence reflects the contribution of the PBS→PSII→PSI transfer. The larger the intensity, the higher the contribution (Yokono et al. 2011). The fluorescence intensity ratios of PSI to PSII in the delayed fluorescence after PSs- and PBS-selective excitations are listed in Table 2. Under PBS-selective excitation, the contribution of PBS→PSII→PSI transfer was largest in P. yezoensis WT and smallest in Synechococcus. The intensity ratios of PSI to PSII in the delayed fluorescence were almost halved in the PBS-selective excitation, relative to the PSs-selective excitation. The sole exception was Synechococcus, for which the delayed fluorescence intensities were almost identical under both excitations (Table 2). This suggests that most of the PSII–PSI supercomplexes connect with PBS in Synechococcus.

Table 2 Fluorescence intensity ratio of PSI to PSII in the delayed fluorescence spectra (TRFS at 47–76 ns) of Synechocystis sp. PCC 6803, Synechococcus sp. PCC 7002, Arthrospira platensis, Cyanidioschyzon merolae, and Pyropia yezoensis wild-type and green mutant excited at 408 or 633 nm (Ex 408 or Ex 633, respectively)
Full size table

The relative amplitude of the τ 5 component was higher under PBS-selective excitation than under PSs-selective excitation (Table 1), indicating that energy is transferred from PBS to low-energy red Chls. In the cyanobacteria Synechocystis and Synechococcus, the PSI peak is clearly blueshifted as the TRFS advances from 3.5 to 4.0 ns to delayed fluorescence (Fig. 2). Similarly, the red algae C. merolae and P. yezoensis exhibit a slight blueshift in the PSI peak wavelength as the TRFS progresses from 9.4 to 11 ns to delayed fluorescence (Fig. 2). These results suggest that PBS→PSI transfer supplies light energy to lower-energy red Chls than PBS→PSII→PSI transfer in Synechocystis, Synechococcus, C. merolae, and P. yezoensis. PSI exists as a monomer in red algae (Gardian et al. 2007), but can be monomeric or trimeric in cyanobacteria (Kruip et al. 1994). Under PBS-selective excitation, the delayed fluorescence spectrum of A. platensis showed two peaks around 730 and 760 nm (TRFS at 47–76 ns, Fig. 2). The peak around 760 nm is unique to PSI trimers (Shubin et al. 1993), implying that PBS–PSII–PSI megacomplexes contain both monomeric and trimeric PSI in vivo.

Immediately after the PBS-selective excitation, the Synechocystis and Synechococcus spectra showed a broad, intense PSI fluorescence band at the longer wavelength side. This suggests that the PBS→PSI transfer delivers light to low-energy red Chls. Conversely, C. merolae exhibits a broad enhanced band at the shorter wavelength side, whereas A. platensis shows a blueshifted peak. These results suggest that light flows from PBS to higher-energy red Chls in these organisms. The clear PSI band in the spectrum of P. yezoensis indicates a large contribution by PSII→PSI transfer (Table 2). However, under PBS-selective excitation, the peak PSI shifts to shorter wavelengths in WT, but not in GM. Beyond 200 ps, the TRFS of WT and GM peak at similar wavelengths. PE level is considerably lower in the GM than in the WT (Niwa et al. 2009). Therefore, the PE contents affect the energy transfer pathways from PBS (APC) to PSI, but not the energy transfers within the PSI.

Summary

We examined the excitation energy transfer processes from PBS to PSs in cyanobacteria (Synechocystis, Synechococcus, and A. platensis) and red algae (C. merolae and P. yezoensis). Observing the delayed fluorescence spectra after PBS-selective excitation, we confirmed PBS→PSII→PSI energy transfer in all of the examined algae (TRFS at 47–76 ns, Fig. 2). We also demonstrated that the PBS–PSII–PSI megacomplexes of cyanobacteria contain both monomeric and trimeric PSI, both of which undergo PBS→PSII→PSI transfer (Fig. 2). The contribution of PBS→PSII→PSI transfer was largest in P. yezoensis WT and smallest in Synechococcus (Table 2). We suggest that in vivo, the PBS→PSI pathway transfers light energy to lower-energy red Chls than the PBS→PSII→PSI pathway (Fig. 2; Table 1). Comparing the time-resolved fluorescences after PSs- and PBS-selective excitation, we revealed that CP43→CP47 energy transfer occurs between neighboring PSII monomers in the PBS–PSII supercomplexes (TRFS at 47–76 ns, Fig. 2). We also suggest that most of the PSII–PSI supercomplexes connect with PBS in Synechococcus (Table 2).