Widespread adaptive evolution in angiosperm photosystems provides insight into the evolution of photosystem II repair

Abstract

Oxygenic photosynthesis generates the initial energy source that fuels nearly all life on Earth. At the heart of the process are the photosystems, which are pigment binding multiprotein complexes that catalyze the first step of photochemical conversion of light energy into chemical energy. Here, we investigate the molecular evolution of the plastid-encoded photosystem subunits at single-residue resolution across 773 angiosperm species. We show that despite an extremely high level of conservation, 7% of residues in the photosystems, spanning all photosystem subunits, exhibit hallmarks of adaptive evolution. Through in silico modeling of these adaptive substitutions, we uncover the impact of these changes on the predicted properties of the photosystems, focusing on their effects on cofactor binding and intersubunit interface formation. By analyzing these cohorts of changes, we reveal that evolution has repeatedly altered the interaction between Photosystem II and its D1 subunit in a manner that is predicted to reduce the energetic barrier for D1 turnover and photosystem repair. Together, these results provide insight into the trajectory of photosystem adaptation during angiosperm evolution.

Figure 1.

Highly conserved regions of the plastid-encoded photosystem proteins in angiosperms. A) Schematics of the PSI (left) and PSII (right) reaction centers with the proteins encoded by the 20 genes analyzed in this study highlighted in green. Proteins in gray were not ubiquitously conserved in the plastid genome among the species analyzed. B) Bar chart showing the percentage of each protein that was variant (gray) versus invariant (black) in the 773 angiosperm species dataset. Genes are ordered by decreasing percentage of protein that is variable from top to bottom. C) Crystal structure of the proteins analyzed in PSI with invariant residues highlighted in black. Red, PsaA; blue, PsaB; dark green, PsaC; pink, PsaI; and teal, PsaJ. For reference, the electron transfer pathway pigments are shown as green sticks and iron–sulfur centers as orange/yellow spheres. D) Crystal structure of the genes analyzed in PSII with invariant residues highlighted in black. Red, PsbA; light blue, PsbB; salmon, PsbC; blue, PsbD; pale green, PsbE; dark green, PsbF; purple, PsbH; teal, PsbI; dark blue, PsbK; green, PsbL; orange, PsbM; pale yellow, PsbT; and pink, PsbZ. Pigments are shown as green sticks and the Fe²⁺ ion as an orange sphere.

Figure 2.

Amino acid substitutions in plastid-encoded photosystem genes during radiation of the angiosperms. A) Horizontal bar chart of the raw count of amino acid substitutions inferred to have occurred across the phylogeny for each protein. Proteins are ordered by decreasing number of amino acid substitutions from top to bottom, respectively. B) Horizontal bar chart of the substitution count per residue for each protein. Proteins are ordered as in (A). C) Venn diagram showing the overlap between invariant residues (yellow) with cofactor binding residues (blue). D) Venn diagram showing the overlap between residues that are invariant (yellow) with those that are at the protein–protein interfaces between subunits (purple). P-values shown below Venn diagrams are the results of hypergeometric tests.

Figure 3.

Adaptive evolution in the plastid-encoded photosystem proteins during the radiation of the angiosperms. A) Left two bar charts show amino acid substitution counts (left) and percentage of substitutions (right) that are single occurrence (orange) or recurrent (blue), i.e. a specific amino acid replacement that has occurred once or more than once at a given residue, respectively. Proteins are ordered by decreasing number of substitutions inferred to have occurred. Right two bar charts show the proportion of recurrent substitutions that occurred more frequently than expected by chance (P < 0.001, Monte Carlo test, dark blue) as compared to those that did not (light blue). Raw counts are shown on the left and percentage of recurrent substitutions shown on the right. B) Bar charts showing the number (left) and percentage (right) of residues under positive selection (red), purifying selection (blue), or subject to neither (gray). Proteins are ordered by decreasing number of sites under positive selection from top to bottom. All proteins below the dashed line have no sites under positive selection. C and D) Structures of PSI and PSII with modeled sites of recurrent evolution highlighted in blue and sites subject to positive selection shown in red, respectively.

Figure 4.

Analyzing regional and structural enrichment of adaptive evolution in the photosystems. A) Bar chart showing the number of sites under adaptive evolution in the 20 photosystem proteins analyzed corrected for protein length in decreasing order from left to right. Proteins in PSI and PSII are in light green and dark green, respectively. B) Boxplot showing data in (A) grouped by photosystem complex. Two-sided two-sample t-test, P > 0.05 (PSI n = 5 and PSII n = 15). C) Boxplot showing data in (A) grouped by photosystem region (Core n = 9 and Accessory n = 11). Boxplots show median (dark line), shaded areas encompass the second and third quartiles, the whiskers indicate the boundaries of the first and fourth quartiles, and dots indicate outliers. Core proteins included PsaA, PsaB, PsaC, PsbA, PsbD, PsbC, PsbD, PsbE, and PsbF. Peripheral proteins included PsaI, PsaJ, PsbH, PsbI, PsbJ, PsbK, PsbL, PsbM, PsbN, PsbT, and PsbZ. P-value shown derived from two-sided two-sample t-test. D–F) Venn diagrams showing the overlap between adaptively evolving sites (orange) and transmembrane residues (green), cofactor binding residues (blue) and intersubunit residues (purple), respectively. P-values shown are the results of hypergeometric tests.

Figure 5.

The position and impact of adaptive substitutions on intersubunit interactions with the D1 (psbA) protein. A) Schematic mapping the location of substitutions that destabilize intersubunit interactions with D1. The D1 protein is shown in orange for clarity and sites of substitutions are marked with a red star. A close-up view of the cluster of destabilizing substitutions in/around the DE-loop of D1 in the 7OUI PSII structure is also provided. Residues at sites of destabilizing substitutions are shown as red sticks. For reference, the two plastoquinones (Q_A and Q_B), Fe²⁺, and HCO₃⁻ involved in electron transfer are also shown. B) Loss of hydrogen bonds (dashed cyan lines) upon E → A substitution between D1 235 and D2 (psbD) N264 and W267. C) Reduced hydrophobic interactions upon L → V substitution at 36 in D1 with F15 and F19 in PsbI. Yellow dashed lines show distance between residues given in Å.