The caseinolytic protease complex component CLPC1 in Arabidopsis maintains proteome and RNA homeostasis in chloroplasts

iTRAQ based proteomics analysis identified new mis-regulated proteins in clpc1 chloroplasts

In Arabidopsis plastids (including chloroplasts), currently 2374 proteins have been identified according to the PPDB database [30]. Among them, the CLP protease complex is crucial to chloroplast development and embryogenesis [31]. CLP proteases are ATP-dependent caseinolytic proteases, consisting of a single proteolytic core complex with 11 distinct subunits including ClpP1, ClpP3–6, ClpR1–4, and ClpT1–2. Moreover, three potential chaperone partners ClpC1, ClpC2, and ClpD and an adaptor protein, CLPS [28] may facilitate the protease complex activity. A proteomics analysis indicated that some proteins in the clpc1 mutant were mis-regulated. For example, photosystem proteins were found with reduced abundance whereas Hsp70, Cpn60, and some RNA-binding proteins were up-regulated [11]. The clpc1 mutant in the WS background had similar morphological phenotypes to those of clpc1 in the Col-0 background such as pale green leaves and retarded growth33 (Fig. 1). Interestingly, the N-terminus-deleted CLPC1 (ΔN) could not complement these phenotypes but full-length CLPC1 could (Fig. 1)28.

Besides participating in the degradation of chloroplast proteins, CLPC1 was suggested to be involved in importing pre-proteins with inner membrane translocation complex components such as TIC110 and TIC40 [29, 34]. Our data showed that both TIC40 and TIC110 were over-accumulated in the clpc1 mutant and in the ΔN line, and their levels were restored to those of the wild type in the full-length CLPC1 complementation line (Table 4b). In accordance to the import function of TIC110 and TIC40, the clpc1 mutant also accumulated more stromal proteins Hsc70–1 and Hsc70–2, both of which are known to mediate pre-protein transport and folding following pre-protein TIC complex transport [34, 35] (Table 1a).

Accumulation of chloroplast RNA metabolism proteins in the clpc1 mutant

RNA homeostasis in chloroplasts is sustained by its biogenesis and degradation and mediated by chloroplast RNA polymerases, RNA-binding proteins, RNases and other proteins. We found that most of these RNA metabolism-related proteins were over-accumulated in the clpc1 mutant as well as in the ΔN plants (Table 2). These proteins include PPR proteins (MEE40, SVR7, and MRL1), RNA-binding proteins (CP29, CP33, RH3, etc.), chloroplast RNases (PRORP1, RNAse J, CSP41B), as well as RNA modification proteins (RNA 3′-end phosphate cyclase, RIF10, and 16S rRNA processing protein). In the full-length CLPC1 complementation line, most of these proteins were restored almost to the wild-type level (Table 2). These results suggest that CLPC1 may have functions in maintaining the homeostasis of these RNA metabolism factors, likely by degrading them when they are damaged or over-accumulated.

Besides the above nucleus-encoded, chloroplast-localized RNA metabolism proteins, all plastid-encoded RNA polymerase (PEP) subunits identified in our proteomic profiling are also over-accumulated in the clpc1 mutant. In addition, several plastid transcriptionally active chromosome proteins (pTACs), which facilitate PEP transcription [18], accumulated in the clpc1 mutant, and their levels could be restored to those of the wild type by reintroducing the full-length CLPC1 into the mutant (Table 3). However, for unknown reasons, rpoA, rpoB, and rpoC2 did not restore to the wild type level in the 2-week-old samples and remained at a relative high level in the full-length CLPC1 complementary line (CP line) (Table 2).

Accumulation of transcripts of chloroplast genes in the clpc1 mutant

Transcription of the plastid genome is accomplished by two different phage-type RNA polymerases (NEP) (RPOTp and RPOTmp) [36–38] along with one eubacterial-type RNA polymerase (PEP) consisting of rpoA, rpoB, rpoC1 and rpoC2 subunits [39, 40]. The activity of PEP is regulated by six sigma-type nucleus-encoded transcription initiation factors [16, 41–44]. Nonetheless, the level of chloroplast transcripts is determined both by transcription and by their metabolism regulated by many RNA processing factors [22]. In our proteomics profiling, we found that the PEP proteins were over-accumulated in the clpc1 mutant. Several PPR proteins, RNA-binding proteins, and RNA modification and degradation proteins were also over-accumulated in the mutant (Table 2). Similarly, there were several over-accumulated pTACs (Table 3). These data imply that CLPC1 may play a role in chloroplast RNA homeostasis. To test this hypothesis, we used gene-specific primers to perform qRT-PCR to specifically examine the level of sense transcripts in the wild type, clpc1 mutant, and the two complementation lines. Our results showed that all chloroplast sense transcripts examined were over-accumulated in the clpc1 mutant and the ΔN line, while they remained at the wild-type levels in the full-length CLPC1 complementation line (Fig. 2, Additional file 1: Figure S1).

Decoupling of the transcript levels and protein levels in the chloroplast photosystem genes

Down-regulation of photosystem proteins is associated with over-accumulation of CLPC2 in the clpc1 mutant

It has been suggested that photosynthesis genes (photogenes) in chloroplasts are transcribed by chloroplast-encoded eubacteria-like RNA polymerases (PEP) [39, 47, 48]. Although PEP subunit proteins (Table 2) as well as the sense transcripts of the photogenes were over-accumulated in the clpc1 mutant and ΔN (Fig. 2) line, proteins encoded by photogenes were accumulated less in these plants than in the wild type and the full-length CLPC1 complementation line (Table 5b). The observation of reduced accumulation of photosystem proteins also was confirmed in the clpp3 knock-out line [31]. These results suggest that there are probably mechanisms limiting the accumulation of those proteins even in the absence of components of the CLPRT complex. Interestingly, there is a concurrent accumulation of the CLPC2 protein in the clpc1 mutant (Table 4a), a phenomenon that was also noted earlier [49]. CLPC2 has been suggested to act antagonistically to FTSH2 (VAR2), a protease involved in photosystem II repair during photoinhibition [50], and thus accelerate photooxidative stress. Accordingly, both the clpc1 mutant and the ΔN line over-accumulated CLPC2 proteins and had pale green leaves with reduced levels of photosystem proteins. The under accumulation of these photosystem proteins could be due to over accumulation of CLPC2 although we cannot rule out that this could be an indirect effect caused by the clpc1 mutation. In contrast, the clpc2 mutant had dark green leaves, and plants over-expressing CLPC2 showed accelerated photooxidative stress and leaf chlorosis (Fig. 3) [50], especially when the seedlings were grown under normal or high light conditions. It was reported that only a subset of plants overexpressing CLPC2 had the leaf chlorosis phenotype [51]. That all the CLPC2-overexpression plants [51] in our hands exhibited chlorosis may be because the seeds we used were from a progenitor with the chlorosis phenotype.

The PPR protein SVR7 as a direct target of CLPC1

SVR7, a PPR protein, was found to accumulate in the clpc1 mutant (Table 2). This protein is required for FtsH-mediated chloroplast biogenesis [23] and the accumulation of ATP synthases and their functional transcripts [52]. Its RNA-binding ability and potential involvement in chloroplast RNA processing make us to ask whether SVR7 is a target of CLPC1. To this end, we examined whether SVR7 interacts with CLPC1. We conducted co-immunoprecipitation (Co-IP) assays using GFP-tagged SVR7. Six peptides belonging to CLPCs were identified. Two out of the four unique peptides identified are CLPC1-specific peptides and the other two could be from either CLPC1 and/or CLPC2 since these regions are identical between the two proteins (Fig. 4). These two CLPC1 unique peptides have high Mascot ion score (Additional file 1: Table S3). Since CLPC2 has a far lower expression level than CLPC1 in the wild type background, it is likely that the other two peptides that are common to both proteins are also from CLPC1. Whereas the negative control GFP-tagged AtYAK1 (a cytoplasm localized protein kinase, At5g35980) did not immunoprecipitate with any CLPC proteins, although other chloroplast proteins also were pulled down with the negative control. The results show that SVR7 may be targeted by CLPC1 and mutation in CLPC1 would lead to SVR7 protein accumulation in the clpc1 mutant. As a result, ATP synthase transcripts were also over-accumulated in the mutant (Additional file 1: Figure S2).

Plant materials

The wild-type Arabidopsis (WS ecotype), clpc1 mutant (WS background), and the ΔN (N-terminal deleted CLPC1 complementary line) and full-length CLPC1 complementation (CP) lines (with the CLPC1 genes driven by the cauliflower mosaic virus 35S promoter) were described previously [29]. The hsp93v (clpc1, sail_873_G11) was from the Arabidopsis Biological Resource Center, 1.4.3 (CLPC2 overexpressing in the clpc1 knockout background), 1.4.4 (CLPC2 overexpressing in the clpc1 knockout background) were from Dr. Paul Jarvis. Seeds were sterilized with 50% bleach with 0.01% Trion X-100, and then washed 5 times with sterilized double distilled H₂O. The sterilized seeds were placed onto a half-strength Murashige and Skoog (MS) salt medium, supplemented with 3% sucrose and 0.6% agar. After 4 days of cold stratification, plates were incubated at 22 °C under constant white light for seed germination and seedling growth. Around 14 days old seedlings were documented and transplanted to soil and further grew for 2 to 4 weeks under long-day (16 h light/ 8 h dark) conditions before chloroplast harvesting. Two independent proteomics experiments were performed. The first set used 4-week-old seedlings and the second set used 2-week-old seedlings (with two biological replicates). These growth periods correspond to the period when significant expression of CLPC1 has been documented.

Chloroplast isolation

Chloroplasts were isolated as described by Wilson et al. (2011) [60]. Briefly, plants were incubated in the dark for 12 h before chloroplast isolation. Large rosette leaves were cut and immediately immersed in a protoplast buffer (20 mM MES-KOH pH 5.2, 400 mM Sorbitol, 0.5 mM CaCl2 with 1.5% cellulase and 0.4% macroenzyme, 0.1% BSA) for 3 h. The protoplasts were then filtered with a 70-μm cell strainer and centrifuged. The materials were then resuspended/rinsed in 5 ml protoplast buffer by gentle swirling and centrifuged for 2 min at 100 g at 4 °C. The pellets were resuspended in 5 ml of buffer protoplast breaking buffer (20 mM Tricine-KOH pH 8.4, 300 mM Sorbitol, 5 mM EDTA, 5 mM EGTA, 10 mM NaHCO3, and 0.1% BSA). The suspension was passed through a 20-μm mesh and collected onto a chilled 40/85 percoll step column. The column was then centrifuged in a swinging rotor for 10 min at 2500 g at 4 °C with the brake off. The lower band was harvested using a pipette and transferred to a 50 ml tube and diluted with 40–45 ml of HEPES-sorbitol buffer (50 mM HEPES-KOH, pH 8.0, 330 mM Sorbitol). The sample was centrifuged for 5 min at 700 g at 4 °C and re-suspended in 200 μl of HEPES-sorbitol buffer (pH 8.0).

RT-PCR

One μg of total RNAs from each of WS, clpc1, ΔN, and the full-length CLPC1 complementation line was used for gene-specific reverse transcription using the Superscript III first strand synthesis kit (Invitrogen). We used the reverse primers for 49 chloroplast and nuclear (CLPC1 and CLPC2) genes and one reverse primer for the ACTIN2 gene in quantitative PCR (qPCR) for first strand cDNA synthesis (100 μM of each reverse primer were mixed, which gave a final concentration of 2 μM for each of the 50 reverse primers). The reverse-transcribed cDNA was first used for PCR to check whether the expected fragment was obtained and then used for quantitative RT-PCR to assess the transcript abundance. The primers used in the study are listed in Additional file 1: Table S2.

Co-immunoprecipitation (co-IP) experiments

Two-week-old Arabidopsis seedlings (ecotype Col-0) harboring the 35S::SVR7-GFP transgene were digested with protoplast buffer (20 mM MES-KOH pH 5.2, 400 mM Sorbitol, 0.5 mM CaCl2 with 1.5% cellulase and 0.4% macroenzyme, 0.1% BSA) for 3 h. Seedlings expressing 35S promoter driven YAK1 tagged with GFP at its C-terminus (35S::YAK1-GFP) were used as a control for the Co-IP. The digestion solution was filtered with a 70-μm cell strainer, and centrifuge at 100 x g for two minutes to pellet the protoplasts. After being washed three times with ice-cold PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, and 1.8 mM KH2PO4, pH 7.4), 200 μl of lysis buffer (10 mM Tris/Cl pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 0.5% NP40, 1 × protease inhibitor cocktail, and 1 mM PMSF) were added and the pellet was re-suspended by extensive pipetting. The sample was incubated on ice for 30 min with extensive pipetting every ten minutes and spun for 10 min at 4 °C at 16100 x g. The supernatant was transferred to a pre-cooled tube, and the volume was adjusted with dilution buffer (10 mM Tris/Cl pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 1 × protease inhibitor cocktail, and 1 mM PMSF) to 1 ml. This cell lysate was added to equilibrated GFP-Trap_A beads and incubated under constant mixing for 2 h at room temperature. The beads were washed three times with washing buffer (10 mM Tris/Cl pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 1× protease inhibitor cocktail, and 1 mM PMSF), and, after the first washing, the NaCl concentration was increased to 500 mM. The bound proteins were eluted by adding 50 μl 0.2 M glycine (pH 2.5) and incubated for 30 s under constant mixing followed by centrifugation. The supernatant was transferred to a new tube, and 5 μl 1 M Tris base (pH 10.4) were added for neutralization. The sample was subjected to electrophoresis in 2 x SDS sample buffer for 12 min and the gel was excised for in-gel digestion and LC-MS/MS analysis.

Peptide preparation, iTRAQ labeling and strong Cation exchange fractionation

Two hundred μl of chloroplasts in HEPES-sorbitol buffer (pH 8.0) were sonicated three times each for ten seconds at two minutes intervals using Qsonica LLC XL-2000 with the power output set at 8. Then the solution was acetone-precipitated (acetone:sample = 5:1 v/v) overnight at − 20 °C. The protein pellet was recovered by centrifugation at 12,000 g at 4 °C for 10 min, rinsed with cold acetone three times, and air-dried. The protein pellet was then resuspended in the buffer containing SDS-PAGE sample buffer without dye. The protein concentration was determined using a 2D Quant kit (GE Healthcare). Approximately 100 μg proteins of each sample were then loaded into a 10% SDS-PAGE gel and run for 25 min to separate proteins from other non-proteins/small molecules. After Coommassie blue staining, the total proteins were used for in-gel digestion with trypsin. The eluted peptides were dried using a Speedvac (Eppendorf, Hamburg, Germany) and labeled with iTRAQ reagents (Applied Biosystems, Framingham, MA, USA) according to the manufacturer’s protocol. Briefly, peptides were reconstituted in 30 μl of dissolution buffer (0.5 M TEAB) and mixed with 70 μl of ethanol-suspended iTRAQ reagents (one iTRAQ reporter tag per sample). Labeling reactions were carried out at room temperature for 60 min before all four samples were mixed in a single tube and dried using a SpeedVac. Strong cation exchange fractionation of the combined peptide mixture was carried out as previously described [61, 62]. Ten fractions were finally obtained, desalted, and dried.

Mass spectrometric analysis using LTQ-Orbitrap

Each dried fraction was reconstituted in 20 μl of 0.1% formic acid and acetonitrile just before mass spectrometric analysis. The labeled sample was analyzed three times on an LTQ-Orbitrap Velos (Thermo Scientific, Germany) coupled with an Easy-nLC (Thermo Scientific). Five microliters of the sample were injected for each analysis and concentrated in a preconditioned column (0.3 × 50 mm) packed with C18 AQ (5 μm particles, 200 Å pore size) (Bruker-Michrom, Auburn, CA, USA). The peptide separation was performed in a preconditioned capillary column (0.1 × 150 mm, with C18 AQ of 3 μm particles and 200 Å pore size (Bruker-Michrom)). The peptide was separated using a 60-min gradient comprised of 35 min of 0–35% mobile phase B (0.1% formic acid in acetonitrile (ACN)), 10 min of 35–80% B, and 15 min of 80% B. The total flow rate of the gradient was set at 400 nl/min. The sample was introduced into the LTQ-Orbitrap through a Nanospray Flex (Thermo Scientific) with an electrospray potential of 1.5 kV. The ion transfer tube temperature was set at 160 °C. The LTQ-Orbitrap was set to perform data acquisition in the positive ion mode. A full MS scan (350–1600 m/z range) was acquired in the Orbitrap at a resolution of 30,000 (at 400 m/z) in the profile mode with a maximum ion accumulation time of 1 s and a target value of 1 × e6. Charge state screening for the precursor ion was activated. The six most intense ions above a 1000-count threshold and carrying multiple charges were selected for a paralleled fragmentation (MS/MS) in the collision-induced dissociation (CID) in the linear ion trap and the higher energy collision dissociation (HCD) in the Orbitrap. Dynamic exclusion for both CID and HCD fragmentation was activated with a repeat count of 2, a repeat duration of 30 s, an exclusion duration of 45 s, and ± 5 ppm mass tolerance. The additional CID settings included a maximum ion accumulation time of 200 ms for MS/MS spectrum collection, a target value of 1 × e4, a normalized collision energy at 35%, an activation Q at 0.25, an isolation width of 3.0 and an activation time of 10 ms. The HCD settings included a full scan with the Orbitrap at a resolution of 7500 (at 400 m/z) in a centroid mode, a maximum ion accumulation time of 200 ms for MS/MS spectrum collection, a target value of 5 × e4, a normalized collision energy at 40%, an isolation width of 3.0 and an activation time of 0.1 ms.

Mass spectrometric data analysis

The MS raw data were processed using the Proteome Discoverer software (version 1.2, Thermo Scientific) to extract mascot generic files (mgf) from HCD and CID spectra separately. The four iTRAQ reporter ions had m/z of 114.112, 115.108, 116.116 and 117.115 respectively. These reporter ions and their intensities for each parent ion were extracted from the HCD mgf files. The mass tolerance for the extraction was set at 10 mDa. The extracted reporter ions were inserted back into both HCD and CID mgf files, whereas their original iTRAQ mass region (114.0–117.5) was cleared. The modified HCD and CID mgf files were analyzed using Mascot (Matrix Science, London, UK; version 2.4.0) [63], which searched the concatenated target-decoy Arabidopsis protein database TAIR10 [30] with common contaminants (71,248 entries). The enzyme limits were set at full tryptic cleavage at both ends, and a maximum of one missed cleavage was allowed. The mass tolerances were set to 10 ppm for the peptide precursors and 0.5 Da for the fragment ions. Variable modifications for the search included iTRAQ (4-plex, 144.10) at tyrosine and oxidation (+ 15.99) at methionine. The fixed modifications were carbamidomethylation (57.02) at cysteine and iTRAQ (4-plex) reagent labeling at N-terminal and lysine.

The mascot search results were exported in csv files and only peptides with expectation value less than 0.05 were included and used for quantitation. Peptide quantification was normalized based on the total intensity of the assigned mass spectrum according to Mascot searching result. The protein ratios were calculated accordingly from the weighted sums of the normalized peptide intensity.