Ubiquitin-based pathway acts inside chloroplasts to regulate photosynthesis

Abstract

Photosynthesis is the energetic basis for most life on Earth, and in plants it operates inside double membrane-bound organelles called chloroplasts. The photosynthetic apparatus comprises numerous proteins encoded by the nuclear and organellar genomes. Maintenance of this apparatus requires the action of internal chloroplast proteases, but a role for the nucleocytosolic ubiquitin-proteasome system (UPS) was not expected, owing to the barrier presented by the double-membrane envelope. Here, we show that photosynthesis proteins (including those encoded internally by chloroplast genes) are ubiquitinated and processed via the CHLORAD pathway: They are degraded by the 26S proteasome following CDC48-dependent retrotranslocation to the cytosol. This demonstrates that the reach of the UPS extends to the interior of endosymbiotically derived chloroplasts, where it acts to regulate photosynthesis, arguably the most fundamental process of life.

Fig. 1.. Proteins of the chloroplast interior are polyubiquitinated.

(A and B) Chloroplasts isolated from transgenic plants expressing Myc-tagged ubiquitin (6Myc-Ub), and from wild type, were analyzed by immunoblotting (A). Similar chloroplasts were treated with thermolysin protease (Ther), or buffer lacking protease (Mock), before immunoblotting analysis (B). The asterisk in (B) indicates a nonspecific band. (C) Chloroplasts isolated from 6Myc-Ub plants were separated into membrane pellet and soluble supernatant (predominantly stroma) fractions by centrifugation at 18,000g and then analyzed by immunoblotting. Analysis of control proteins confirmed the efficacy of the protease treatment and fractionation steps. Positions of molecular weight markers (sizes in kDa) are shown to the right of the images.

Fig. 2.. Photosynthesis and other proteins of the chloroplast interior are prominent in the chloroplast ubiquitinome.

(A) Dot plot showing significantly overrepresented gene ontology (GO) terms in the chloroplast ubiquitinome, as determined using chloroplasts purified from CDC48-DN plants after estradiol induction. Dot size indicates overrepresentation (fold enrichment) compared to the whole genome. Dot color indicates false discovery rate [FDR; −log₁₀ (P value)], where higher FDR values indicate more statistically significant enrichment. Dots are not shown for terms lacking statistically significant (P < 0.05) enrichment. (B) Suborganellar and functional distribution of the chloroplast ubiquitinome, showing also ubiquitination sites as determined by di-Gly analysis. Localizations were assigned manually, and only proteins localized in a defined chloroplast compartment (OEM, IEM, stroma, and thylakoid), or in an internal chloroplast membrane fraction (comprising IEM and thylakoids), are shown. Boxes indicate individual proteins (white, nucleus-encoded; orange, chloroplast-encoded), and red circles with numbers indicate which amino acids showed ubiquitination. (C) Histogram showing the number of ubiquitination sites detected per protein in the chloroplast ubiquitinome. (D) Pie chart showing the relative abundance (based on peptide intensity) of different polyubiquitin linkage types in the chloroplast ubiquitinome. Values are means from three experiments. (E) Logo plot showing motif analysis of ubiquitinated peptides in the chloroplast ubiquitinome. Six residues either side of the modification site (position 0) are shown.

Fig. 3.. Identification of numerous CHLORAD substrates using quantitative proteomics.

(A) Pie chart showing the differential accumulation of chloroplast (cp) proteins upon CHLORAD inhibition, as determined by quantitative proteomic analysis of CDC48-DN and CDC48-WT transgenic plants after estradiol induction. (B) Dot plot showing significantly overrepresented GO terms for chloroplast proteins that are overaccumulated in CDC48-DN plants. Dot size indicates overrepresentation (fold enrichment) compared to the whole genome. Dot color indicates false discovery rate [FDR; −log₁₀ (P value)], where higher FDR values indicate more statistically significant enrichment. Dots are not shown for terms lacking statistically significant (P < 0.05) enrichment. (C) Pie chart showing the differential accumulation of ubiquitinated chloroplast proteins in the quantitative proteomic analysis depicted in (A). Only proteins in the chloroplast ubiquitinome (Fig. 2) are shown. (D) Pie chart showing the differential expression of mRNAs corresponding to the chloroplast proteins identified by proteomics in (A), as determined by RNA-seq transcriptomics. (E) Dot plot showing a lack of correlation between chloroplast protein abundancies [from (A)] and corresponding mRNA abundancies [from (D)]. The coefficient of determination (R²) value is shown.

Fig. 4.. Proteins of the chloroplast interior are processed by CHLORAD.

(A and B) Proteasome dependency of PrfB3 degradation was analyzed. Wild-type protoplasts expressing PrfB3-HA were incubated with/without cycloheximide (CHX) and proteasome inhibitor bortezomib (Btz) and analyzed by immunoblotting (A). Bands for PrfB3-HA were quantified and normalized to histone H3 control data (B). Time 0 was taken as 100%. Data are means ± SEM from at least two experiments. (C to E) SP2 dependency of chloroplast-encoded protein degradation was analyzed. Wild-type and sp2 plants were incubated with lincomycin (Lin) and analyzed by immunoblotting (C) and Coomassie staining (D). Bands were quantified as in (B). Data are means ± SEM from at least four experiments. Asterisks in (E) indicate significance according to unpaired two-tailed Student’s t tests comparing the two genotypes at each time point (*P < 0.05, **P < 0.01, and ****P < 0.0001; ns, not significant). (F to H) CDC48 dependency of chloroplast-encoded protein degradation was analyzed. Data are equivalent to those in (C) to (E), except that CDC48-WT/DN transgenic plants were used. (I to K) Interaction of substrates with the CHLORAD apparatus was assessed by coimmunoprecipitation (co-IP). In (I), wild-type protoplasts were cotransfected with SP2-6Myc and either PrfB3-HA, YFP-HA, or SP1-HA. In (J), wild-type or SP2-6Myc seedlings were analyzed without protoplastation. In (K), wild-type protoplasts were transfected with YFP-HA or CDC48-Trap-HA [AtCDC48A^E581Q that stabilizes substrate binding (8)]. Immunoprecipitations used anti-Myc (I and J) or anti-HA (K) resin. YFP-HA acted as a negative control. Positions of molecular weight markers (sizes in kDa) are shown. Asterisk in (I) indicates a nonspecific band. TL, total lysate; poly-Ub, poly-ubiquitinated.

Fig. 5.. CDC48 is required for the extraction of polyubiquitinated photosynthesis proteins.

(A) The role of CDC48 in the extraction of chloroplast-encoded PsaA and PsbC substrates was analyzed using an in vivo retrotranslocation assay. Protoplasts from CDC48-WT and CDC48-DN transgenic plants that were transiently expressing FLAG-Ub were treated, after estradiol induction, with bortezomib proteasome inhibitor and then separated into cytosol and chloroplast fractions. In this assay, retrotranslocation occurred in intact cells, and the extracted substrates were protected by bortezomib inhibition, which initiated the experiment. After fractionation, ubiquitinated proteins were immunoprecipitated from both fractions and detected by immunoblotting. Tic40, which is not a substrate of CHLORAD, served as a negative control. Typical immunoblotting results are shown. Positions of molecular weight markers (sizes in kDa) are shown to the left of the images. The asterisk indicates a nonspecific band. (B and C) Retrotranslocation efficiency was assessed by quantifying the relative amounts of ubiquitinated PsaA (B) and PsbC (C) in the cytosol and chloroplast fractions described in (A). Data are means ± SEM from at least three experiments.

Fig. 6.. CHLORAD is required for normal photosynthetic function and lipid homeostasis.

(A to E) Chloroplast development in 8-day-old CDC48-WT/DN seedlings, after 2-day estradiol induction, was studied. Cotyledons of typical plants (A) were analyzed by transmission electron microscopy (B). Upper images in (B) are at the same magnification (scale bar, 2 μm); higher-magnification (×4) images corresponding to the boxed regions are below. These micrographs were used to determine plastoglobule diameter (C), chloroplast cross-sectional area (D), and thylakoid lamellae per granum (E). Values are means ± SEM from 50 chloroplasts. (F to I) Photosynthetic electron flow through PSI [ETR(I)] and PSII [ETR(II)] in 4-week-old CDC48-WT and CDC48-DN plants (F and G), and in 5-week-old wild-type and sp2 mutant plants (H and I), was measured. Here, the CDC48 plants received a 1-day estradiol induction. Plants were dark-adapted before exposure to actinic light [photosynthetically active radiation (PAR)] of the indicated intensities, followed by saturating pulses for the calculations. Values are means ± SEM from five to seven plants. (J) Free FA and polar lipid species in CDC48-WT/DN seedlings, after 2-day estradiol induction, were quantified. Data were normalized to internal standards and per unit fresh weight, and the normalized values were expressed relative to CDC48-WT (taken as 1). Numbers alongside the FA/lipid names describe the constituent FA chains, in the format (number of FA carbons):(number of FA double bonds). Selected FA/lipid species showing significant differences (P < 0.05) in CDC48-DN relative to CDC48-WT are shown; the complete dataset is in fig. S10. Values are means ± SEM from six biological replicates.