PSB27: A thylakoid protein enabling Arabidopsis to adapt to changing light intensity

Identification of a Mutant with a Defect in the Assembly of PS II Supercomplexes.

Our earlier work suggests that FKBP20-2 and CYP38 chaperones function in the assembly of PSII SCs (14, 15). Specifically, these proteins appear to be localized to the chloroplast thylakoid lumen where a major portion of the PSII complex resides. These findings raise the possibility that additional luminal proteins may be required for PSII assembly. To gain information on this point, we have taken a systematic reverse genetic approach and isolated 58 Arabidopsis T-DNA insertion mutants in genes encoding 30 luminal proteins (16, 17). The mutants were screened for defects in PSII complex organization by blue native gel electrophoresis. The mutant line with the most marked defect in PSII organization (SALK_004769) contained a T-DNA insertion in the 5′ UTR of the PSB27 gene (At1g03600, Fig. S1A). As reported earlier (18), this mutant was devoid of PSB27 mRNA determined by RT-PCR [Fig. S1B, compare wild type (WT) and mutant line (ML)] as well as the corresponding protein probed by SDS/PAGE and Western blot (Fig. S1C). Protein fractionation of subchloroplast compartments confirmed that PSB27 is located in the thylakoid lumen (Fig. S1D).

In an effort to localize the defect in PSII, we next conducted more quantitative blue native gel analysis with WT and mutant thylakoid membrane proteins and found pronounced differences in the abundance of individual PSII SC bands: notably, the mutant plants were almost completely devoid of PSII SCs that were readily detected in the WT (Fig. 1A). By contrast, the band representing the free forms of the corresponding LHC trimers (LHCII-T) showed a relative increase in abundance. When subjecting the blue native gel slices to the second electrophoresis dimension by SDS/PAGE, we observed a dramatic decrease in abundance of protein subunits associated with the PSII SCs (Fig. 1B, see red boxes).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Analysis of thylakoid membrane protein complexes. (A) Blue native gel analysis of thylakoid proteins of WT and ML plants (15 μg chlorophyll as loading standard). (B) Thylakoid proteins separated by BN gel in A were further subjected to SDS PAGE and silver-stained. The red frames indicate the individual protein components of the PSII SCs. Protein identification was according to Fu et al. (15). (C) 2D-SDS PAGE immunoblots of thylakoid proteins. Antibodies were applied as indicated. (D) 1D-SDS PAGE Western blots of thylakoid proteins (2.5 μg chlorophyll) from WT and ML plants grown for 4 wk in a growth chamber under constant light intensity. (E) 1D-SDS PAGE immunoblots of thylakoid proteins (2.5 μg chlorophyll) from WT and ML grown in a greenhouse for 4 wk. Antibodies applied and plants grown as indicated.

To further visualize changes in the PSII complexes, we used immunoblots to examine the distribution and abundance of subunits of the four major thylakoid membrane complexes. To this end, the individual proteins resolved by 2D-SDS PAGE were transferred to a nitrocellulose membrane and probed with antibody against subunits of each of the four thylakoid membrane complexes: PSII: D1, D2, CP43, CP47, PsbO, and LHCII; PSI: PsaA/B, PsaD, and PsaF; cytochrome b₆ f complex: CytF, b6, and PetC; and CF₀-CF₁ complex: ATPα. The results confirmed that the mutant (ML) lacked the protein subunits in the PSII SCs that were observed in the WT (Fig. 1C). The subunit distribution pattern of other complexes including PSI, Cyt b₆ f, and CF₀-CF₁ remained unchanged (Fig.1C).

It was difficult to compare the abundance of each individual protein between WT and mutant using 2D Western blots because the gels were processed separately. To circumvent this difficulty, we quantified these proteins by loading WT and mutant samples on the same SDS/PAGE gel, followed by Western blot. The results showed no significant difference in the abundance of PSII subunits between the WT and mutant (ML) cultivated under growth chamber (Fig. 1D) or greenhouse conditions (Fig. 1E). Taken together, these results suggested although the total amount of individual PSII subunits was not affected, the mutant plants were incapable of assembling the individual subunits to form PSII SCs.

We noted that the mutant showed two new weak bands in the PSII SCs region of the blue native gel (Fig. 1A): one was located between 1,250 and 1,100 kDa and the other was just below the 880 kDa band of WT PSII SC. To determine whether these bands represented altered forms of the PSII SC or were unrelated to this photosystem, we performed blue native gel analysis of both WT and mutant (ML) with a larger amount of total protein (Fig. S2A). The major bands above the PSII dimer in both WT and mutant samples in Fig. S2A were isolated and subjected to SDS PAGE analysis. The results showed that the top band, which was identified as NDH (labeled WT-1 and ML-1 in Fig. S2A), and the bottom band, which was identified as PSI SC (19, 20) (labeled WT-6 and ML-4), displayed identical patterns of subunit composition in the WT and ML (Fig. S2B). The four major PSII SC bands of WT (WT-2–WT-5) all contained typical PSII subunits (red lettered a–d bands in Fig. S2B), whereas the two unique bands (ML-2 and ML-3) in mutant lacked these subunits. We further showed by mass spectrometry that these proteins (a–d bands in the WT sample in Fig. S2B) are true PSII subunits (Fig. S2C). Interestingly, both the 880-kDa band of the mutant (ML-3) and the upper band (ML-2), neither of which is obvious in the WT gel, displayed a subunit pattern similar to PSI SC (WT-6) instead of PSII SC (Fig. S2 A and B). This result again supports our conclusion that all detectable forms of PSII SCs are eliminated in the mutant.

Mutant Line Is Deficient in Adaptation to Changing Light Intensity.

When cultivated in a growth chamber under normal growth conditions (Fig. 2A), the ML failed to show a noticeable phenotypic difference from WT in either soil (Fig. 2E) or MS medium (Fig. 2I) despite lacking PSII SCs. Similar results were reported previously (18). By contrast, when grown in the greenhouse (Fig. 2B), leaves of ML plants were much smaller and less green compared with WT (Fig. 2 F and J). Clearly, environmental differences (growth chamber vs. greenhouse) caused a dramatic change in mutant phenotype. To identify the factor responsible for such a difference, we cultivated the plants under several growth conditions—i.e., temperature, day length (8 h light/16 h dark for short day, 16 h light/8 h dark for long day), and light intensity (20 μmol m⁻² s⁻¹ to 250 μmol m⁻² s⁻¹). The results indicated that none of these factors was responsible for the observed difference in phenotype. We then realized that the growth chamber supplies light at constant intensity during the light cycle, whereas the greenhouse experiences light at fluctuating intensity during the day akin to a natural setting.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Mutant line is compromised in light adaptation. (A–D) Light intensity regimes for plant growth. The light/dark cycle was 16 h light/8 h dark for all regimes. (A) Growth chamber conditions with constant light intensity (100 μmol m⁻² s⁻¹) during the light periods. (B) Greenhouse conditions with light intensity changes during the day time. (C) Growth chamber conditions with light intensity changes from 75 μmol m⁻² s⁻¹ to 250 μmol m⁻² s⁻¹ during the light periods. (D) Growth chamber conditions with light intensity switched between 50 μmol m⁻² s⁻¹ and 150 μmol m⁻² s⁻¹ every hour during the light periods. (E–H) WT and ML were grown in soil for 4 wk under the various light regimes in A–D. (I–L) WT and ML were grown on 1/2 MS medium for 10 d under light regimes in A–D. Growth conditions depicted in A–D were used, respectively, for growth in E–H and I–L.

To pursue the possibility that change in light intensity led to altered growth of the mutant, we designed growth chamber conditions such that plants experienced a gradient of light intensity during the light cycle (Fig. 2C), mimicking greenhouse conditions (Fig. 2B). Indeed, the ML plants subjected to this light intensity regime clearly differed from WT, similar to those cultivated in the greenhouse (Fig. 2 G and K). When the light intensity was changed every hour from 50 to 150 μmol m⁻² s⁻¹ during the light cycle (Fig. 2D), growth of ML plants was also slow relative to WT (Fig. 2H). A similar phenotypic difference was observed with plants grown on agar plates containing 1/2 MS medium (Fig. 2L). These results suggested the ML plants have defects in adaptation to changing light intensity.

Mutant Line Is Defective in PSII Function and Starch Biosynthesis.

The stunted growth of ML plants observed under changing light intensity suggests that the mutant had lower photosynthetic capacity under these conditions. To test this idea, photosynthetic parameters were determined by measuring chlorophyll fluorescence in leaves of whole plants. The ratio of variable fluorescence to maximum fluorescence (F_v/F_m), which represents the maximum photochemical efficiency of PSII (21), was lower in the ML than in the WT even with plants cultivated in the growth chamber which showed no visible phenotypic difference (Fig. 3A, Upper). The decrease in the F_v/F_m ratio became more pronounced with plants grown in the greenhouse (Fig. 3B, Upper). ФPSII, which measures the proportion of PSII-associated light harvesting, was also lower in the ML (Fig. 3A, Lower). Like the F_v/F_m ratio, ФPSII was further decreased when plants were grown in the greenhouse (Fig. 3B, Lower). Taken together, the chlorophyll fluorescence assays suggested that ML plants are compromised in PSII function.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Mutant line is deficient in utilization of light. (A) Chlorophyll fluorescence analyses of WT and ML plants grown under constant light intensity conditions in the growth chamber (80 μmol m⁻² s⁻¹ for 16 h and dark for 8 h) for 4 wk. (B) Chlorophyll fluorescence analyses of WT and ML plants grown under greenhouse conditions for 4 wk. F_v/F_m: the maximum efficiency of PSII photochemistry; ФPSII: efficiency of PSII electron transport. Data represent mean ± SEM (n = 4). (C) Electron micrographs of chloroplasts from 4 wk-old WT and ML grown under constant light intensity conditions. (D) Electron micrographs of chloroplasts from WT and ML grown in greenhouse for 4 wk. Transmission Electron Microscope (FEI) was used to examine leaf cross sections as described (42).

The fluorescence results prompted us to test the possibility that PSII SC formation is needed for thylakoid membrane organization. If so, lack of PSII SCs could lead to defects in chloroplast morphology in mutant plants. To pursue this idea, we examined leaf sections of both WT and ML under the electron microscope. Under growth chamber conditions, the chloroplasts from ML and WT showed similar size, grana, and starch accumulation (Fig. 3C). By contrast, electron microscopy revealed a striking difference between WT and ML specimens grown in the greenhouse: The chloroplasts in wild type accumulated copious amounts of starch, whereas, consistent with its stunted growth, the ML was almost completely devoid of starch grains (Fig. 3D), indicating a much lower efficiency of starch biosynthesis under changing light conditions. This inefficiency could result, in part, from an impairment in the regulation of enzymes of starch biosynthesis due to defective PSII activity and the attendant reduction of thioredoxin (22).

Mutant Line Accumulates Less Anthocyanin and Produces More ROS Under High Light Stress.

Because PSII deficiency and photoinhibition can lead to damage under high light, we investigated whether the mutant line is defective in adaptation to this condition. To this end, WT and ML were grown on 1/2 MS medium for 10 d under low light (20 μmol m⁻² s⁻¹). These plants showed no obvious difference in phenotype. Further, when the plants were transferred to continuous high light (350 μmol m⁻² s⁻¹), both the WT and ML grew well. However, surprisingly, we observed a pronounced difference in leaf color within 2 d after the transfer to high light. WT plants turned dark purple, whereas the ML remained green (Fig. 4A, Lower), indicating that mutant is defective in the accumulation of anthocyanin, a pigment that serves as a “filter” for protection against high light (23). A similar pattern was observed when plants were cultivated in soil (Fig. 4B). The 4-wk-old, soil-grown plants showed no difference when grown under normal light intensity (60 μmol m⁻² s⁻¹) (Fig. 4B, Top). However, after transfer to high light (350 μmol m⁻² s⁻¹), a difference in the color of leaves of WT and mutant plants was clearly visible (Fig. 4B, Lower). In quantitative terms, WT rosettes accumulated 165 μg total anthocyanin per g fresh weight, whereas the ML accumulated only 18 μg, a drop of almost 90% (Fig. 4C).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Mutant line is defective in light-induced anthocyanin accumulation and displays enhanced ROS accumulation. (A) WT and ML plants were grown on 1/2 MS medium under constant low light conditions (20–30 μmol m⁻² s⁻¹) for 10 d, and then transferred to higher light intensity (60 μmol m⁻² s⁻¹, upper plate or 350 μmol m⁻² s⁻¹, lower plate) for 2 d. (B) WT and ML were grown in soil under constant light intensity (16 h light/8 h dark, 60 μmol m⁻² s⁻¹) for 4 wk, and then were transferred to high light conditions (350 μmol m⁻² s⁻¹) for 2 d. (C) Anthocyanin content of plants in B. Data represent mean ± SEM (n = 4). (D) Composite curves of gene expression patterns of WT and ML after high light treatment. Expression levels of individual genes are shown in Fig. S3. (E) DAB (3, 3-diaminobenzidine) staining for hydrogen peroxide in high light treated (350 μmol m⁻² s⁻¹ for 2 d) plants of WT and mutant. (F) The relative expression levels of ROS marker genes measured by real-time RT-PCR before and after high light (350 μmol m⁻² s⁻¹) treatment. The results are shown as mean ± SEM from three experiments.

To obtain insight into the anthocyanin deficiency, we applied real time q-PCR to examine the transcript levels of genes responsible for accumulation of the pigment, including two regulatory genes (PAP1 and PAP2) and 12 biosynthesis genes (Table S1) (24). Each of these genes was induced by high light in both WT and ML (Fig. S3). However, the kinetics of mRNA accumulation differed between the two groups. In the mutant, transcript levels increased at the beginning and rapidly dropped to the basal level, whereas in WT, transcripts increased to a higher level, and were sustained for a longer time (Fig. S3 and Fig. 4D). The results suggest that the mutant can respond to a high light signal, but the response seems inadequate to sustain the production of anthocyanin.

The data above show that the mutant has a reduction in PSII function (Fig. 3 A and B). The mutant should, therefore, sustain more severe photoinhibition due to the formation of higher levels of reactive oxygen species (ROS). Moreover, the observed inhibition of anthocyanin biosynthesis in the ML may also have resulted from elevated levels of ROS as reported earlier (24, 25). To follow up on this possibility, we measured hydrogen peroxide in WT and ML using 3, 3-diaminobenzidine (DAB) stain. Elevated levels of H₂O₂ were observed in the mutant compared with WT plants following high light treatment (Fig. 4E). These results were confirmed in experiments with ROS-responsive gene markers (25) (Fig. 4F). The mRNA levels of five ROS marker genes were similar in WT and ML before the treatment. However, after exposing plants to high light for two days, the expression level of each of the five marker genes was significantly higher in ML compared with WT. The results support the conclusion that the excessive H₂O₂ production taking place in high light inhibits anthocyanin biosynthesis as previously reported (24, 25).

Lack of CP26 Causes Deficiency of PS II Supercomplexes.

To confirm the link between phenotypic changes in the T-DNA mutant and the function of PSB27, we transformed the mutant with the wild type PSB27 gene driven either by the 35S promoter or native promoter. Several phenotypic changes in the mutant were fully complemented by PSB27, including defects in light adaptation, photosynthetic capacity and anthocyanin biosynthesis. The transgenic plants also grew similarly to WT under greenhouse conditions (Fig. S4 A and B) and accumulated comparable amounts of anthocyanin after high light treatment (Fig. S4B, Right). In like manner, the chlorophyll florescence parameters representing PSII capacity (F_v/F_m) and efficiency (ФPSII) were recovered in the complemented plants grown under greenhouse conditions (Fig. S4C).

However, to our surprise, the defect in PSII SCs remained in the complemented plants (Fig. S4D). For a better resolution of PSII antenna components, the thylakoid proteins were, therefore, separated by Tris-Tricine SDS PAGE. This procedure revealed that a protein band of 26 KDa was missing in ML (Fig. S4E). Mass spectrometry identified the missing protein as CP26, a monomeric antenna protein shown earlier to be required for the assembly of PSII SCs (1, 4). Lack of CP26 would explain why the mutant failed to form PSII SCs and also why PSB27 was unable to complement this defect. In short, the CP26 defect was unrelated to PSB27 function. As a consequence, we followed up by examining the genomic sequence of the CP26 gene in the mutant and found a point mutation in the first exon of CP26, which caused the absence of the corresponding mRNA and protein (Fig. S4 E–G).

PSB27, but Not PSII Supercomplex, Is Critical for Growth Adaptation to Changing Light Intensity.

Although finding the CP26 defect in the mutant line was unexpected, it provided an ideal genetic system to compare the contribution of PSII SCs and PSB27 to light intensity adaptation. To this end, we segregated the cp26 and psb27 mutations by backcrossing ML with WT. The F2 population of plants was genotyped to identify the cp26 and psb27 single mutants. As seen in Fig. 5, the psb27 mutant, but not the cp26 mutant, showed retarded growth under greenhouse conditions (Fig. 5 A and B middle) and produced less anthocyanin under high light (Fig. 5B, Right). The cp26 mutant, again not psb27 mutant, was deficient in PSII SCs, which could be rescued by expressing a wild type copy of CP26 gene in the mutant (Fig. 5 C and D).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Genetic linkage of cp26 and psb27 mutant alleles with plant growth defects. (A) The original T-DNA ML containing both cp26 and psb27 mutant alleles was compared with the WT and with segregated single mutants of cp26 and psb27 under constant light intensity in a growth chamber (Upper) or under varying light intensity in a greenhouse (Lower). (B) The plants were grown on 1/2 MS medium in growth chamber (Left) or greenhouse (Center) for 10 d. The same plants grown under constant low light for 10 d were then transferred to high light (350 μmol m⁻² s⁻¹) for 2 d (Right). (C) Blue native gel analysis showing PSII SCs of thylakoid proteins of WT, ML, psb27, cp26, ML complemented with CP26 or/and PSB27. (D) Tris-Tricine SDS PAGE analyses showing CP26 proteins of thylakoids of the same plants in C.

We then compared psb27,cp26 with state transition mutants including stn7 (10), stn8 (26), and stn7 stn8 double mutants (27). After growing for 4 wk in the greenhouse, cp26 and WT plants were comparable, the state transition mutants were slightly smaller than WT (particularly in the double mutant), whereas psb27 displayed marked growth retardation (Fig. 6A). A similar phenotypic difference was observed with plants grown on 1/2 MS medium plates for 10 d (Fig. 6B). The phosphorylation of LHCII and CP43 which is regulated by STN7 and STN8 suggests that state transitions (10, 26), were not affected in psb27 mutant (Fig. 6C). The results suggest that PSB27 plays a fundamental role in enabling plants to adapt to change in light intensity. Furthermore, PSII SCs or state transitions do not appear to significantly contribute to light adaptation at least under the conditions used in this study.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Comparison of cp26, psb27, stn7, stn8, and stn7 stn8 double mutant in their adaptation to light conditions. (A) WT, cp26, psb27, stn7, stn8, and stn7 stn8 double mutant were grown in soil in growth chamber (Upper) and greenhouse (Lower) for 4 wk (the growth conditions were as in Fig. 2 A and B). (B) The same plants were grown on 1/2 MS medium for 10 d under light regimes in Fig. 2A (Growth Chamber) and Fig. 2B (Greenhouse). (C) phosphorylation of thylakoid proteins. Total proteins of 2-wk-old seedlings were separated by 12% SDS PAGE and immunoblotted with anti-phosphothreonine antibodies.

Concluding Remarks.

The present experiments have uncovered a protein of the chloroplast thylakoid lumen, PSB27, that is required for adaptation to changing light intensity. Using an Arabidopsis mutant lacking PSB27, we found that, although not essential under conditions of constant light intensity, this protein was required for normal growth under changing light regimes in either a natural setting or a growth chamber. The stunted growth observed under changing light was associated with lower PSII efficiency and defective starch accumulation. In response to high light exposure, mutant plants also displayed enhanced ROS production and reduced synthesis of the protectant anthocyanin although such changes did not affect plant growth.

We observed that the monomeric antenna protein CP26, a component of PSII SCs, is required for the assembly of the photosynthetic apparatus. An Arabidopsis mutant lacking CP26 was found to be essentially devoid of PSII SCs. Although it was reported that the LHCII S trimer could associate with the PSII core in the absence of CP26 (28), our results suggest that this association is not stable, consistent with the sucrose gradient results (1).

The impact of state transitions on energy balance between PSI and PSII has been established for the green alga Chlamydomonas (29). However, state transitions play a less prominent role in land plants than in green algae. Growth is only marginally affected in Arabidopsis mutants impaired in state transitions (27, 30), even in fluctuating light (10). Our results have shown that state transition mutants also display minor, if any, phenotypic change under conditions that severely inhibit growth of psb27 mutant plants. Our results suggest that land plants adapt to changing light intensity through a PSB27-dependent mechanism that is independent of both state transitions and PSII SCs. Thus, based on our results it seems possible that the recently reported effects of PSB27 on PSII SC formation and state transitions (31) may have been the result of the cp26 point mutation that was present together with the psb27 T-DNA insert in the mutant line that was used (SALK_004769).

A final point, PSB27 is present in cyanobacteria as well as land plants (16, 32). In cyanobacteria PSB27 is tightly associated with the thylakoid membrane via an N-terminal lipid modification that is absent in PSB27 from land plants (33). Therefore, PSB27 of Arabidopsis is present exclusively in the soluble phase of the thylakoid lumen (Fig. S1D). This difference in location may, in part, be responsible for functional differences of PSB27 in the two groups of organisms. In cyanobacteria, the membrane-bound PSB27 facilitates the assembly of the manganese cluster of PSII (34⇓⇓⇓–38) by interacting tightly with the CP43 core protein in the assembly of intermediates. In land plants, on the other hand, PSB27 functions in acclimation to changing light. Elucidation of further details of PSB27 function may well add insight into the evolution of oxygenic photosynthesis.