Oxyphototroph organisms sustain most of the life forms on earth by absorbing solar energy and converting it into the chemical energy and releasing oxygen. Photosystems I and II (PSI and PSII) are two thylakoid membrane-embedded complexes essential for the photosynthetic light reactions. Both photosystems contain a structurally conserved core complex and a highly variable antenna system. While most eukaryotic phototrophs utilize the membrane-embedded light-harvesting complexes (LHCs) as peripheral antennas for both photosystems, prokaryotic cyanobacteria contain huge membrane-extrinsic phycobilisomes (PBSs) serving as the antenna system of PSII 1. Red algae possess both PBSs and LHCs as antennas for PSII and PSI, respectively 2.
Cryptophytes acquired their chloroplasts from red algae by secondary endosymbiosis millions years ago 3. These microalgae utilize both membrane-embedded LHCs and membrane-extrinsic phycobiliproteins (PBPs) as antennas 4. However, PBPs in cryptophytes form rhombic α1α2ββ-tetramers5 and are located at the thylakoid lumen 6, differing from the stroma-localized ring-shaped PBPs in cyanobacteria and red algae. Cryptophyte LHCs are similar to those in red algae, but contain unique carotenoid alloxanthin, as well as chlorophyll (Chl) a and c2, thus were named ACP (alloxanthin and chlorophyll a/c binding protein) 7 or generally termed CAC (chlorophyll a/c binding protein) 8 .
Cryptophyte algae are single-cellular organisms that go through logarithmic and stationary growth phase (L-phase and S-phase) 9,10. While L-phase cells contain higher amounts of PBPs and exhibit high photosynthetic efficiency 9,10, cells in S-phase are characterized by greatly reduced amount of PBPs 9. The cryptophyte cells may regulate their light-harvesting capability via utilizing a specialized antenna system according to their particular growth phase. The recently determined structures of PSI-ACPI from Chroomonas placoidea (CpPSI-ACPI) provide a basis for understanding the assembly and pigment arrangement of CACs in cryptophyte PSI at L-phase 7. However, the precise details of how the antennas of PSI are organized at S-phase remain unknown.
Here, we solved four structures of cryptophyte PSI-CAC complex purified from Rhodomonas salina (Rs) cells grown at L- and S-phase, containing either 14 or 11 CACs (Fig. 1, Fig. S1-S6, Table S1). We termed these four structures as PSI-14CACL-phase and PSI-11CACL-phase for L-phase models, and PSI-14CACS-phase and PSI-11CACS-phase for S-phase models. Two PSI-14CAC structures closely resemble each other, containing 14 CAC subunits which encircle the core complex, and were named CAC-a to CAC-n in a clockwise way when looked at the lumenal side (Fig. 1a). Except CAC-a and CAC–h, all CACs are grouped as four heterotrimers. Three trimers, combined with CAC-a and CAC–h, constitute the inner antenna layer surrounding the core, whereas the fourth heterotrimer (CAC-l/m/n) forms the outer layer. Within CAC trimers, CACs in the corresponding positions exhibit high similarity in both protein folding and pigment arrangement (Fig. S7). It is noteworthy that CAC-h constitutes a RedCap (red lineage chlorophyll a/b-binding-like protein) 11, which exhibits slightly different conformation and binds fewer pigment molecules compared to canonical CACs (Fig. S8). One subunit (termed chain-s) contains one transmembrane helix and binds two of each Chl a, Chl c2 and α-carotene (Figs. S6a, S9a). Chain-s is sandwiched by the inner CAC-a/b/c and the outer CAC-l/m/n (Fig. 1a), and forms multiple hydrogen bond interactions with CAC-a/b/c/l at the lumenal side (Fig. S9b). The only difference between the two PSI-14CAC structures is that in the PSI-14CACL-phase complex, one four-helix-bundle protein is located at the lumenal side, while in the PSI-14CACS-phase structure, it appears to be missing (Fig. 1b). The overall folding of this protein is similar to the PSII extrinsic subunit PsbQ; we therefore termed it PsaQ (Fig. 2a, Fig. S6b).
Consistent with the trimeric organization of CACs, both PSI-11CAC structures lack one CAC trimer, but at distinct positions (Fig. 1b). PSI-11CACL-phase loses the inner trimer-e/f/g which attaches to PsaO, whereas PSI-11CACS-phase lacks the outer trimer, which links to the inner belt through chain-s. Accordingly, PsaO is untraceable in the PSI-11CACL-phase structure, and chain-s is absent in our PSI-11CACS-phase structure. While PsaQ is also absent in our PSI-11CACS-phase structure, it binds to PSI-11CACL-phase at the same lumenal position as that in PSI-14CACL-phase (Fig. 1b). Moreover, the two CpPSI-ACPI complexes purified from L-phase cells exhibited structures almost identical to our two L-phase PSI-CAC structures, and appeared to contain the lumen-localized PsaQ (Unk1 in CpPSI-ACPI) 7 (Fig. S10). Together, these results suggested that PsaQ is strongly associated with the PSI core in L-phase cells.
Based on our high-quality cryo-EM map and cDNA sequencing result, we revealed the identity of PsaQ, which is an uncharacterized yet highly conserved protein in cryptophytes (Fig. S11). PsaQ harbors three chlorophyll molecules, which are located at the interfacial region between CAC-i and PsaB (Fig. 2a,b), thus stabilizing the core-CAC association. A long loop between helix 2 and 3 of PsaQ inserts into the membrane plane and forms multiple hydrogen bond interactions with CAC-i/PsaB (Fig. 2b-d), tethering PsaQ to the thylakoid membrane.
Several chloroplast proteins with PsbQ-like folding (Fig. S12) were found binding to the lumen surface of photosynthetic supercomplexes, and were suggested to facilitate the assembly, folding, stabilization of these supercomplexes 12. However, to the best of our knowledge, a pigment-bound PsbQ-like protein has never been identified in phototrophs. PsaQ is unique to cryptophytes, and the chlorophyll-coordinating residue N234 is extremely conserved in PsaQ homologs (Fig. S11), indicating that PsaQ and the bound chlorophylls are crucial for the proper functioning of cryptophyte cells during L-phase. The three chlorophylls of PsaQ are located in close proximity to chlorophylls of CAC-i and PsaB (Fig. 2e). These observations suggest that in L-phase cells, PsaQ facilitates efficient excitation energy transfer (EET) at the lumenal side of PSI-CAC. The fact that PsaQ is absent in PSI-CAC from S-phase cells is in line with the idea that photosynthesis is less efficient in S-phase cells 13.
We also found that PsaQ and PE545 (PBP in R. salina) share several similar features, including their exclusive presence in cryptophytes 14, their lumenal localization, and the higher stability/abundance in L-phase cells, suggesting that PsaQ might be involved in PE545 function. While EET from PE545 to both photosystems was previously suggested 15, precisely where PE545 is positioned relative to PSI and PSII remains unknown. Direct binding of PE545 to the PSII core appears to be sterically hindered by the presence of extrinsic subunits PsbO/PsbU/PsbV and the lumen-extruding domains of core subunits CP43/CP47. Furthermore, the central region of PSI core constitutes the docking site for the lumenal electron donor, thus preventing direct association of PE545 with the PSI core.
Interestingly, we found that both RsPSI-11CACL-phase and CpPSI-11ACPI structures superposed well with the PSI-LHCI moiety of the recently reported in-situ structure of PBS-PSII-PSI-LHC megacomplex from red alga Porphyridium purpureum (PpPSI-LHCI) 16 (Fig. S13). When we assessed the PpPBS-PSII-PSI-LHC megacomplex structure for features of PSI-PSII association, we found that the PSI interacts with the PSII core through the PsaL-PsaO-PsaK side (Fig. 2f). RsPSI-11CACL-phase and CpPSI-11ACPI structures also feature an exposed PsaL-PsaO-PsaK side without bound CACs, implying that a similar PSII-PSI-CAC megacomplex is present in the L-phase cryptophyte cells. While direct evidence remains absent, we assume that the PsaQ-bound PSII-PSI-CAC megacomplex in L-phase R. salina cells provides a platform for binding PE545. We thus superposed our PSI-11CACL-phase structure onto the PSI part in PpPBS-PSII-PSI-LHC structure, generating a hypothetic RsPSII-PSI-CAC model (Fig.2f).
When we analyzed our hypothetic RsPSII-PSI-CAC model, we found that the arrangement pattern of PsaQ and the PSII lumenal subunit PsbQ’ is almost symmetrical (Fig. 2f,g). Together with the membrane-embedded RedCap, these subunits shape a lumenal shallow groove with the width of approx. 78 Å. The PE545 is a tetramer characterized by a pseudo-two-fold symmetry, with a dimension of approx. 75 Å ´ 60 Å ´ 45 Å5. These features enable PE545 to fit into the lumenal groove (Fig. 2g). Based on our hypothetic model of RsPE545-PSII-PSI-CAC, PE545 may transfer energy to both photosystems through CACs and PsaQ, and may balance the energy distribution between the two photosystems. This model is in agreement with an earlier suggestion that PE545 transfers the energy to photosystems through CACs 17,18. Moreover, our model explains previous data measured by the steady-state and time-resolved fluorescence anisotropy, showing that PBPs funnel the excitation energy to both photosystems with similar efficiency 15,19.
We assume that the proposed RsPE545-PSII-PSI-CAC megacomplex is more abundant in L-phase cells, as it may meet the energy requirement for the quick growth of L-phase cells. When cryptophytes enter the S-phase, changes of physiological conditions inside the chloroplast, such as lumen pH values, may weaken interactions between PsaQ and PsaB/CAC-i, resulting in the detachment of PsaQ and the reorganization of the antennas of both photosystems. It is likely that PSI-11CACL-phase constitutes the preferential form for the potential RsPE545-PSII-PSI-CAC megacomplex formation, whereas other populations of PSI-CAC complexes simultaneously exist in the chloroplast, where they facilitate the dynamic regulation of light harvesting of PSI under different physiological conditions.
In conclusion, the findings of our study should allow to propose a regulatory mechanism by which R. salina cells adjust the light-harvesting capability and photosynthetic efficiency in line with their growth phase, namely through changing the types of antennas used as well as through rearranging the PSI-associated CACs. Future structural and biochemical assessment of the proposed RsPE545-PSII-PSI megacomplex will provide the detailed information required for verifying our hypothesis.