Structure of the maize photosystem I supercomplex with light-harvesting complexes I and II

See allHide authors and affiliations

Science  08 Jun 2018:
Vol. 360, Issue 6393, pp. 1109-1113
DOI: 10.1126/science.aat1156

Antenna switches partners in the shade

A cloudy day or an overshadowing tree causes fluctuations in light that can throw off the balance of energy flow in plant photosystems I and II (PSI and PSII). Pan et al. solved structures of PSI bound to two light-harvesting complexes (LHCs). One LHC is permanently associated with PSI. The other LHC delivers light energy to PSII under optimal conditions but can switch to a PSI-associated state after phosphorylation by a kinase that senses the redox environment of the chloroplast. The movement of LHCs between the photosystems helps maintain even energy flux. Two chlorophyll-containing subunits are visible in the structure that connect the PSI core to each LHC.

Science, this issue p. 1109


Plants regulate photosynthetic light harvesting to maintain balanced energy flux into photosystems I and II (PSI and PSII). Under light conditions favoring PSII excitation, the PSII antenna, light-harvesting complex II (LHCII), is phosphorylated and forms a supercomplex with PSI core and the PSI antenna, light-harvesting complex I (LHCI). Both LHCI and LHCII then transfer excitation energy to the PSI core. We report the structure of maize PSI-LHCI-LHCII solved by cryo–electron microscopy, revealing the recognition site between LHCII and PSI. The PSI subunits PsaN and PsaO are observed at the PSI-LHCI interface and the PSI-LHCII interface, respectively. Each subunit relays excitation to PSI core through a pair of chlorophyll molecules, thus revealing previously unseen paths for energy transfer between the antennas and the PSI core.

During oxygenic photosynthesis, photosystems II (PSII) and I (PSI) operate in series and are tightly coupled to allow efficient light-driven electron transport. Both photosystems are multisubunit supramolecular complexes containing a core complex and the peripheral antenna system (1, 2). In plants, the peripheral antennas are composed of light-harvesting complexes (LHCs). LHCIs (containing the Lhca apoproteins) are associated with the PSI core, forming the PSI-LHCI complex, whereas LHCIIs (containing the Lhcb apoproteins) are mostly attached to the PSII core, constituting the PSII-LHCII complex. The antenna systems of PSI or PSII have different pigment compositions and hence different light absorption properties. Red and far-red light preferentially stimulate PSII and PSI, respectively, and fluctuating illumination can cause unequal excitation of the two photosystems. Balanced light harvesting is crucial for efficient photosynthesis; thus, plants have developed both short-term and long-term adaptation to the constantly changing light conditions in the natural environments. State transitions are a short-term response occurring on a time scale of minutes (3) and allow a balanced energy distribution between the two photosystems when light quality changes (46). During state transitions, the trimeric LHCII (composed of different combinations of Lhcb1-3) is reversibly phosphorylated and dephosphorylated, a process controlled by the redox state of plastoquinone (PQ) and regulated by a chloroplast kinase (STN7) and phosphatase (PPH1, also known as TAP38) in plants (5, 6). In state 1, LHCIIs are mostly associated with PSII and transfer excitation energy to the PSII core. Under light conditions favoring PSII excitation, overexcitation of PSII leads to a reduction of the PQ pool, activation of STN7 kinase, and subsequent phosphorylation of the N-terminal region of LHCII. A portion of the phosphorylated LHCIIs (mobile LHCII) move laterally within the thylakoid membrane from PSII to PSI, forming the PSI-LHCI-LHCII supercomplex and resulting in a switch from state 1 to state 2. The mobile LHCII serves as a peripheral antenna of PSI in addition to LHCI, increasing energy transfer toward the PSI core. Under natural light conditions, state transitions are necessary for optimal plant growth and fitness (79).

A two-dimensional projection map of the PSI-LHCI-LHCII supercomplex at 16-Å resolution (10) revealed a single LHCII trimer associated with the PSI core on the opposite side of LHCI (10, 11), yet the protein-protein and pigment-pigment interactions between LHCII and PSI remain unclear. Although crystal structures of plant LHCII had been solved previously (12, 13), the N-terminal tail of LHCII containing the phosphorylation site was not observed in these structures. It remains an unresolved question as to how the phosphorylated LHCII enhances its interaction with PSI. Plant PSI-LHCI includes a core complex made up of 14 subunits (PsaA to L, PsaN, and PsaO) and a peripheral antenna system containing four LHCI proteins organized as two heterodimers (Lhca1-Lhca4 and Lhca2-Lhca3) (2). Recent crystal structures of pea PSI-LHCI (14, 15) revealed the architecture and location of 16 subunits, but two plant-specific subunits, PsaN and PsaO, were not observed in these structures. An intact PSI-LHCI structure should allow for a better understanding of the energy transfer within the PSI-LHCI complex.

We purified PSI-LHCI-LHCII from maize leaves treated with orange light, which induces the shift into state 2 (fig. S1). Using single-particle cryo–electron microscopy (cryo-EM), we solved the PSI-LHCI-LHCII structure at an overall resolution of 3.3 Å (fig. S2 and table S1). One LHCII trimer binds PSI at the PsaA side (Fig. 1). In addition to the 16 subunits present in the previously reported structures (14, 15), we identified two areas of density corresponding to the subunits PsaN and PsaO (Figs. 1B and 2 and fig. S3). The PSI-LHCI moiety in the PSI-LHCI-LHCII supercomplex shows the most complete protein composition, with no known subunit missing (Fig. 1, fig. S4A, and table S2). LHCII exhibits lower resolution than the PSI-LHCI portion; however, its location and orientation are assigned unambiguously as the transmembrane helices are well resolved (fig. S3D). In our PSI-LHCI-LHCII structure, the membrane-spanning regions of PSI-LHCI and LHCII are not coplanar but instead form an angle of about 10° (fig. S3A). The curved structure of the PSI-LHCI-LHCII supercomplex may indicate a role for the membrane-embedded complexes in influencing the native membrane architecture, as previously suggested for the mitochondrial adenosine 5′-triphosphate (ATP) synthase (16). Superposition of the PSI-LHCI moiety in our maize supercomplex structure with the two previously reported pea PSI-LHCI crystal structures (14, 15) reveals that, except for PsaN and PsaO, all other 16 PSI-LHCI subunits are similar to their counterparts from pea and exhibit small variations in their hydrophilic loop regions (fig. S4). On the basis of the cryo-EM density, we were able to build several subunits with more complete structures (fig. S4, B to D, and table S2) and assigned a few more chlorophyll molecules (fig. S4, E and F, and table S3).

Fig. 1 The maize PSI-LHCI-LHCII supercomplex.

(A) View from the luminal side. One LHCII monomer with its N-terminal region contacting PSI is labeled as M1. (B) Side view along membrane plane. PsaN and PsaO are highlighted in cartoon mode.

Fig. 2 Structures and interactions of individual subunits in the supercomplex.

(A) PsaN binds two chlorophylls; Chl a1001 is coordinated by the main-chain carbonyl oxygen of Phe106. (B) PsaN is located at the luminal side and interacts with PsaA, PsaF, PsaJ, Lhca2, and Lhca3. (C) PsaO is associated with two chlorophyll molecules; Chl a2001 is ligated by His103. A fragment in the luminal loop region of PsaO is partly disordered in the structure and is shown as a dashed line. (D) PsaO bridges PsaA and LHCII through contacts within the membrane and on the stromal and luminal surface. (E) The interactions of the N-terminal region of pLhcb2 with PsaL, PsaH, and PsaO. Hydrogen bonds with distances between 2.6 and 3.4 Å are shown as black dashed lines.

PsaN exists exclusively in the eukaryotic PSI (17) and is the only membrane-extrinsic PSI subunit found at the luminal side (18). PsaN is easily detached from PSI by salt washing (19). We purified the supercomplex at low ion concentration through one-step isolation after mild solubilization of the thylakoid membranes to keep PsaN attached. We further confirmed the presence of PsaN in our PSI-LHCI-LHCII sample by immunoblotting (fig. S1E). PsaN is composed of three α helices and binds two chlorophylls (named a1001 and a1002) (Fig. 2A and fig. S5). PsaN is located near the luminal regions of PsaA, PsaF, and PsaJ, where it fills the empty space between Lhca2 and PsaA (Fig. 2B and fig. S6, A and B). This location is consistent with previous biochemical and genetic studies (2022) and suggests a role in connecting these subcomplexes. In crystal structures of pea PSI-LHCI (14, 15), a large gap is present between Lhca2 and the core complex at the luminal side. Because of the longer distances between interfacial chlorophylls, the excitation energy transfer (EET) from Lhca2 to the core was suggested to mainly flow via Lhca3, an inefficient pathway involving multiple steps. In our structure, two “linker chlorophylls” from PsaN create direct EET pathways between Lhca2 and the PSI core (Fig. 3, A and B, and table S4). Chl a1001 from PsaN is strongly connected with the red form Chl a603 from Lhca2 and is in close contact with Chl a808 from PsaA, thus establishing the major EET pathway from Lhca2 to the PSI core (Fig. 3C). In addition, both chlorophylls from PsaN can receive excitation energy from luminal chlorophylls of Lhca2, further facilitating the energy transfer between Lhca2 and the PSI core. The arrangement of chlorophylls in PSI-LHCI moiety in our structure is consistent with previous picosecond fluorescence spectroscopy showing that Lhca2 can efficiently transfer the absorbed energy to the core in a direct manner (23, 24). PsaN is also necessary for the efficient interaction between PSI and plastocyanin (Pc) (25), a soluble protein that transfers electrons in the thylakoid lumen from cytochrome b6f to PSI complex. The efficient binding of Pc to eukaryotic PSI requires the N-terminal extension domain of PsaF (26). In agreement, our structure shows that PsaN forms close contacts with the N-terminal extension of PsaF (fig. S6C). Therefore, PsaN may be directly involved in the formation of the Pc binding site or assist the association of Pc by stabilizing the N-terminal domain of PsaF (fig. S6D).

Fig. 3 Chlorophyll arrangement and proposed energy transfer pathways in the PSI-LHCI-LHCII supercomplex.

(A) Distribution of chlorophylls within the PSI-LHCI-LHCII supercomplex at the stromal layer. Three chlorophylls (a1001 from PsaN, a2001 from PsaO, and a616 from Lhca1) located at the luminal side can receive energy from and transfer energy to chlorophylls at the stromal layer, and are shown in ball-and-stick model. The black and red lines indicate the pathways composed of chlorophyll molecules within the same layer and at different layers, respectively. (B) Chlorophylls at the luminal layer. In (A) and (B), the possible EET pathways from Lhca2 to PSI core (mediated by PsaN) and from LHCII to PSI core are highlighted by green and purple circles, respectively. The interfacial chlorophyll pairs with Mg-to-Mg distances shorter or longer than 17 Å are linked with solid or dashed lines, respectively. (C) The plausible EET pathways from Lhca2 to PSI core (shown as dashed lines). (D) The plausible EET pathways from LHCII to PSI core (shown as dashed lines). For clarity, the phytol chains of chlorophylls are omitted. The distances of interfacial chlorophyll pairs are summarized in table S4.

We also identified density corresponding to PSI subunit PsaO, with two transmembrane helices and N- and C-termini loops along stromal surface. Two chlorophylls, named a2001 and a2002, are located near the luminal and stromal surfaces, respectively (Fig. 2C and fig. S7). An earlier report suggested that PsaO is involved in the formation of the docking site for LHCII (27), and state transitions were reduced by 50% when PsaO expression was abolished in Arabidopsis (28). Our structure shows a direct connection of PsaO with LHCII and several PSI core subunits (Fig. 2D), in agreement with cross-linking studies (27, 28). PsaO interacts with PsaL, PsaA, and PsaK at the stromal surface, PsaA at the luminal side, and LHCII at their membrane-spanning interface. The N-terminal region of one LHCII monomer (M1) (Fig. 1A) forms close contacts with core subunits PsaL, PsaH, and PsaO at the stromal side (fig. S8). Recent reports showed that the mobile LHCII is mainly composed of (Lhcb1)2Lhcb2 trimer, and the phosphorylated Lhcb2 (pLhcb2) is responsible for the association of LHCII with PSI under state 2 conditions (2931). The specific features of the cryo-EM density around the N-terminal tail of monomer M1 match with the Lhcb2 sequence (figs. S3D and S8C) but not with that of Lhcb1, which is five residues longer than Lhcb2 at the N-terminal region (fig. S9). We therefore assigned the monomer M1 as pLhcb2 and the other two monomers as Lhcb1 (Fig. 1A). The high-quality cryo-EM density (fig. S8C) allowed the construction of the complete N-terminal region of pLhcb2, including its phosphorylated Thr3 (pThr3) residue. The phosphate group of pThr3 strongly interacts with several residues from PsaL (Fig. 2E). These residues are conserved in plant PsaL and located at an extended stromal loop, which is absent in cyanobacteria (fig. S10). The two basic residues (Arg1 and Arg2) preceding pThr3 in Lhcb2 are also pivotal for PSI subunits recognizing LHCII. Arg2 binds residues from PsaL and PsaO. Arg1 interacts with the N-terminal regions of PsaH (Fig. 2E). The first three residues are completely conserved among Lhcb2 from different plants, and Lhcb1 always has a Lys residue at its second position (fig. S9), which may be less favorable for interacting with PsaL. The relatively shorter N-terminal region and the specific residue at the second position (Arg2) of Lhcb2 are likely to account for the specificity of Lhcb2 interacting with the PSI core. In addition, our structure shows that the plant-specific subunits (PsaO and PsaH) and the extended stromal loop of PsaL are essential for PSI to interact with Lhcb2, explaining the crucial role of PsaH and PsaL in state transitions (32). The strong interactions at the stromal side between PSI and LHCII are probably responsible for the curved architecture of the supercomplex (fig. S3A).

In addition to its structural role, PsaO also mediates EET from LHCII to the PSI core through its two chlorophylls (Fig. 3, A and B, and table S4). PsaO may receive the excitation energy from Lhcb2 of LHCII at both stromal and luminal layers (Fig. 3D). Chl a2001 from PsaO is located close to Chl a611 from Lhcb2. The Chl a611/a612/a610 cluster in LHCII was proposed to be at the lowest-energy state (33), so the excitation energy equilibrated within the LHCII trimer may be focused on this Chl cluster in Lhcb2 and further transferred to PsaO. In addition, PsaK may be involved in alternative EET pathways by receiving the excitation energy from Lhcb1 at both stromal and luminal sides. The excitation energy collected by LHCII will be further transferred to the stromal cluster a823/a824/a845 and the luminal pair a836/a837 in PsaA. The network of chlorophylls within LHCII and PSI core observed in our structure agrees with previous spectroscopic results demonstrating that the LHCII trimer associated with PSI is a highly efficient light harvester (11, 34, 35). These well-connected interfacial pigments should provide multiple EET pathways between LHCII and the PSI core, similar to the connections mediated by PsaN between LHCI and the PSI core.

Supplementary Materials

Materials and Methods

Figs. S1 to S10

Tables S1 to S4

References (3649)

References and Notes

Acknowledgments: We thank J. Zhang and X. Zhao for assistance in preparing thylakoid samples. Cryo-EM data collection was carried out at the Center for Biological Imaging, Core Facilities for Protein Science, at the Institute of Biophysics (IBP), Chinese Academy of Sciences (CAS). We thank X. Huang, G. Ji, Z. Guo, B. Zhu, D. Fan, T. Niu, F. Sun, and other staff members at the Center for Biological Imaging (IBP, CAS) for their support in data collection; L. Niu, X. Ding, F. Yang for mass spectrometry; J. Li for assistance in spectral measurement; and L. Kong for cryo-EM data storage and backup. Funding: The project was funded by the National Key R&D Program of China (2017YFA0503702, 2017YFA0504700, 2016YFA0502900), the Strategic Priority Research Program of CAS (XDB08020302, XDB08030204), the Key Research Program of Frontier Sciences of CAS (QYZDB-SSW-SMC005), and National Natural Science Foundation of China (31770778, 31700649, 31600609, 31570724). Z.L. and X.Z. received scholarships from the “National Thousand (Young) Talents Program” from the Office of Global Experts Recruitment in China. Author contributions: W.C., Z.L., M.L., and X.Z. conceived the project; X.S. performed the sample preparation, characterization, and data collection; J.M. and X.Z. processed the cryo-EM data and reconstructed the cryo-EM map; X.P. built and refined the structure model; P.C. assisted in the structural refinement; X.P., Z.L., X.Z., and M.L. analyzed the structure; M.L., Z.L., and X.Z. wrote the manuscript; all authors discussed and commented on the results and the manuscript. Competing interests: The authors declare no competing interests. Data and materials availability: The atomic coordinate of the PSI-LHCI-LHCII supercomplex has been deposited in the Protein Data Bank with accession code 5ZJI. The cryo-EM map of the supercomplex has been deposited in the Electron Microscopy Data Bank with accession code EMD-6932.
View Abstract

Stay Connected to Science

Navigate This Article