The ATP-dependent Clp protease in plant chloroplasts consists of a heterogeneous proteolytic core containing multiple ClpP and ClpR paralogues. In this study, we have examined in detail the only viable knockout mutant to date of one of these subunits in Arabidopsis thaliana, ClpR1. Loss of ClpR1 caused a slow-growth phenotype, with chlorotic leaves during early development that later partially recovered upon maturity. Analysis of the Clp proteolytic core in the clpR1 mutant (clpR1-1) revealed approx. 10% of the wild-type levels remaining, probably due to a relative increase in the closely related ClpR3 protein and its partial substitution of ClpR1 in the core complex. A proteomic approach using an in organello proteolytic assay revealed 19 new potential substrates for the chloroplast Clp protease. Many of these substrates were constitutive enzymes involved in different metabolic pathways, including photosynthetic carbon fixation, nitrogen metabolism and chlorophyll/haem biosynthesis, whereas others function in housekeeping roles such as RNA maturation, protein synthesis and maturation, and recycling processes. In contrast, degradation of the stress-related chloroplast proteins Hsp21 (heat-shock protein 21) and lipoxygenase 2 was unaffected in the clpR1-1 line and thus not facilitated by the Clp protease. Overall, we show that the chloroplast Clp protease is principally a constitutive enzyme that degrades numerous stromal proteins, a feature that almost certainly underlies its vital importance for chloroplast function and plant viability.

INTRODUCTION

Molecular chaperones and proteases are integral components of the quality control processes in the actively changing cellular environment of all living organisms. Chaperones comprise a large group of proteins with diverse functions and are involved in such processes as protein folding/unfolding and in complex assembly/disassembly [1]. Additionally, some chaperones interact with proteases and are responsible for recognizing polypeptide substrates and unfolding them prior to degradation [2]. Proteases have an equally important role and are responsible for degrading damaged proteins that might otherwise accumulate to potentially harmful levels or begin to impair associated processes. Besides these housekeeping activities, proteases are also involved in recycling amino acids and regulating the stability of key enzymes and regulatory proteins (reviewed in [3]).

Plant chloroplasts are dynamic organelles which, consistent with their endosymbiotic origin, contain various proteases of bacterial ancestry. Of these chloroplast proteases, the best characterized to date are FtsH proteases anchored to the stromal surface of the thylakoid membranes, Deg proteases bound extrinsically to both sides of the thylakoid membrane, and the stromal-localized Clp protease (reviewed in [4]). The Clp protease is an ATP-dependent serine-type protease present in eubacteria, plants and mammals [5]. This two-component enzyme, composed of one Hsp100 (heat-shock protein 100) chaperone partner complexed to a proteolytic core, is best described in Escherichia coli. The central proteolytic core of the E. coli Clp protease is comprised of two apposing heptameric rings of ClpP with the proteolytic active sites sequestered within the internal cavity [6]. This barrel-shaped core is flanked on one or both ends by a single homogeneous hexameric ring of an Hsp100 partner, either ClpA or ClpX [7]. The Hsp100 partner recognizes, unfolds and translocates the protein substrate through the narrow entrance of the Clp proteolytic core into the central chamber in an energy-dependent manner [810]. Substrates are then rapidly degraded into small peptides that eventually diffuse out of the core complex [11].

Higher plants have by far the greatest diversity of Clp proteins, with at least 23 different types in the model species Arabidopsis thaliana [5,12]. Of these, nine are Hsp100 chaperones (ClpB1–B3, ClpC1–C2, ClpD, and ClpX1–X3), six are paralogues of the proteolytic subunit ClpP (ClpP1–P6), and four are paralogues of a ClpP-like variant (ClpR1–R4), which lacks the Ser-type catalytic triad and has recently been shown to be proteolytically inactive in cyanobacteria (F. I. Andersson and A. K. Clarke, unpublished work). In addition, Arabidopsis has an orthologue to the bacterial adaptor protein ClpS as well as two unique Clp proteins with sequence similarity to the N-terminal domain of Hsp100 proteins, which here are renamed ClpT1–T2 (from ClpS1–S2) to avoid confusion with the functionally distinct ClpS adaptor. The bulk of Clp proteins in Arabidopsis are located in the chloroplast stroma (ClpB3, ClpC1–C2, ClpD, ClpP1, ClpP3–P6, ClpR1–R4, ClpS and ClpT1–T2), with all except the plastomic ClpP1 encoded in the nucleus [1315]. Despite the numerous Clp proteins residing in the stroma, only one Clp proteolytic core has been identified [14,16]. This 325–350 kDa complex consists of two sub-complexes, presumably representing two heptameric rings by analogy to the ClpP core complexes in other organisms. One ring consists of ClpP1 and ClpR1–R4, while the other contains ClpP3–P6 [17]. In addition, a single subunit each of ClpT1 and ClpT2 associates peripherally to the proteolytic core [16]. ClpT1 has been further shown to bind to the ClpP3–P6 sub-complex [17].

Previous studies using various transgenic plants have demonstrated the crucial role of the chloroplast Clp protease. Reduced amounts of the plastomic ClpP1 or other components of the Clp proteolytic core results in plants impaired in chloroplast differentiation, shoot development and overall plant viability [1722]. In clpP6 antisense plants, for example, the amount of the chloroplast Clp proteolytic core decreased by 80–90%, consistent with the level of antisense repression of the ClpP6 subunit. These antisense lines were also used in the identification of the first few putative protein substrates for the chloroplast Clp protease [17]. In contrast, no viable Arabidopsis mutants for any of the nuclear-encoded ClpP or ClpR paralogues have been reported to date, despite extensive screening of all available T-DNA insertion lines [5]. The one exception to this was an EMS (ethylmethanesulfonate) mutant, clpR1-1, the characterization of which has so far been concentrated on 5-day-old seedlings [22]. In this study, we have now characterized in detail the clpR1-1 mutant at a later stage of development. The absence of ClpR1 leads to a chlorotic slow-growing phenotype as previously observed [22], and plants with reduced photosynthetic rates. Despite the loss of ClpR1, a small amount of intact Clp proteolytic core is formed in the clpR1-1 line, probably due to partial compensation by the ClpR3 subunit. Moreover, we have now used the clpR1-1 line to identify many new potential protein substrates for the stromal Clp protease, some of which are involved in general homoeostatic roles, whereas others are involved in more specific metabolic pathways. We also reveal that certain stress-inducible stromal proteins are not degraded by the Clp protease, emphasizing its role principally as an essential constitutive housekeeping enzyme.

EXPERIMENTAL

Plant growth conditions

Arabidopsis thaliana wild-type (ecotype Columbia-0) and clpR1-1 mutant [22] were grown in soil containing 20% (v/v) perlite in controlled environment chambers. Seeds were vernilized at 4 °C for a minimum of 48 h to break dormancy. All plants were grown individually in pots or as lawns under the following standard conditions: 8 h photoperiod with white light (approx. 150 μmol·m−2·s−1), 23/18 °C day/night temperatures, and 65% RH (relative humidity). These conditions were found to improve the growth of the clpR1-1 line from those used previously, in which seeds were first germinated on ½ Murashige and Skoog medium containing 1% (w/v) sucrose, grown under 12 h photoperiod for the first 10 days, then transferred to soil and placed in 16 h photoperiod [22]. Unless stated otherwise, all experiments were performed with wild-type Arabidopsis and clpR1-1 plants at the same developmental age (5 and 7 weeks old respectively).

Chlorophyll and photosynthetic measurements

Chlorophyll was extracted from leaf discs or small leaves and measured as described previously [23]. FV/FM and ETR (electron transport rate) parameters were measured on leaves using a pulse-amplitude modulated fluorimeter (PAM-2000; Heinz-Waltz, Effeltrich, Germany) also as described earlier [23]. Statistical significance was determined using the t test in which P<0.01 was significantly different.

Protein complex separation by native PAGE

Intact chloroplasts were isolated from wild-type and clpR1-1 leaves as described previously [17,24]. Protein concentration in each sample was determined using the BCA (bicinchoninic acid) protein assay according to the manufacturer's protocol (Pierce). The Tris/borate-based PAGE system [25] was then used to separate native soluble protein complexes (60 μg of protein) from wild-type and clpR1-1. Samples were electrophoresed to ensure that proteins reached their pore size limitation within the gel matrix as previously described [26]. Molecular-mass standards were ferritin (440 kDa monomer, 880 kDa dimer), urease (272 kDa trimer), and BSA (66 kDa monomer, 132 kDa dimer). After native PAGE, proteins were transferred to nitrocellulose and the Clp protein complexes were detected using specific polyclonal antibodies as previously described [15,17,23]. Primary antibodies were detected with the horseradish peroxidase-linked anti-rabbit IgG secondary antibody from donkey (Amersham Pharmacia) and visualized by enhanced chemiluminescence (ECL Advance; GE Healthcare) using the ChemiGenius2 imaging system (Syngene).

2D-PAGE (two-dimensional PAGE)

Stromal protein fractions were obtained from isolated chloroplasts [17] and purified further using the ReadyPrep 2D Cleanup Kit (Bio-Rad). Purified protein (450 μg) was resuspended in IEF (isoelectric focusing) buffer [8 M urea, 2 M thiourea, 50 mM DTT (dithiothreitol), 4% CHAPS, 0.2% (v/v) Bio-Lyte pH 3–10 ampholytes and 0.0002% (w/v) Bromophenol Blue] and loaded on 24 cm IEF ReadyStrips [IPG (immobilized pH gradient) nonlinear pH 4 to 7; Bio-Rad]. Strips were rehydrated for 12 h at 20 °C in 24 cm strip holders, followed by IEF on an Ettan IPGphore using the standard separation program. After focusing, strips were immediately equilibrated for SDS/PAGE in DTT buffer [6 M urea, 2.5% (v/v) SDS, 20% (v/v) glycerol, 2% (w/v) DTT, 0.002% (w/v) Bromophenol Blue and 50 mM Tris/HCl, pH 8.8] for 15 min followed by a 10 min incubation in iodoacetamide buffer [6 M urea, 2.5% (v/v) SDS, 20% (v/v) glycerol, 2.5% (w/v) iodoacetamide, 0.002% (w/v) Bromophenol Blue and 50 mM Tris/HCl, pH 8.8]. IEF strips were then placed on 12% (v/v) acrylamide SDS/PAGE gels (Tris/Gly) and run overnight at 20 °C in a Ettan DALT II system (GE Healthcare). Gels were electrophoresed in simultaneous sets of four to reduce variation, with triplicate gels run for each treatment. Gels were stained with Flamingo™ fluorescent stain (Bio-Rad), scanned, and spots were then analysed using specialized software ImageMaster™ 2D Platinum 5.0 (GE Healthcare). Spot volumes were normalized to the total spot volume on each gel. Spots were selected for mass-spectrometric identification based on significant differences in the three replicates between sample pairs; identification was performed with MALDI–TOF MS (matrix-assisted laser-desorption ionization–time-of-flight MS) or HPLC-MS/MS (Finnigan LTQ-FT; Thermo Electron) using the software Mascot (at Matrix Science, U.K.) at the SweGene Proteomics Center (Göteborg University).

Protein degradation assay

Intact chloroplasts from the lawns of wild-type and clpR1-1 plants at the same developmental age (3 and 5 weeks respectively) were isolated and counted as previously described [17]. An equal number of intact chloroplasts were incubated for 0 and 3 h in 60 μmol·m−2·s−1 light at 25 °C in the presence of an ATP regeneration system [17]. Samples were then ruptured in 5 volumes of rupture buffer (10 mM MgCl2, 20 mM Hepes/NaOH, pH 7.6) and frozen in liquid nitrogen. Thawed samples were centrifuged at 20000 g for 10 min, after which the stromal supernatants were transferred to new tubes. Protein concentration was determined using the BCA protein assay described above. For proteins larger than 60 kDa, 1D-PAGE (one-dimensional PAGE; precast 3–8% polyacrylamide Tris/acetate gels; Invitrogen) was used, followed by staining with Coomassie Brilliant Blue G 250. For proteins less than 60 kDa, 2D-PAGE (described in detail above) was used. All proteins identified as potential substrates for the Clp protease were quantified. Proteins exhibiting >25% degradation in wild-type chloroplasts during the time course of the assay were identified by MALDI–TOF MS and HPLC-MS/MS (SweGene Proteomics Center, Göteborg University).

For the Hsp21 degradation assay, fully developed wild-type and clpR1-1 plants grown under standard conditions were heat stressed under conditions similar to that described previously [27]. Plants were first preheated at 40 °C for 30 min at 4 h into the light period. After 20.5 h recovery, the plants were again heat shocked at the start of the next light period by gradually raising the growth temperature to 40 °C during the first hour, maintaining at 40 °C for a further 4 h, then gradually lowering the temperature to 25 °C during the following hour. During all high-temperature treatments, the light was kept at an irradiance of 350 μmol·m−2·s−1. Approx. 3 days after the heat shock treatment (i.e. 65 h), intact chloroplasts were isolated in three independent batches each containing chloroplasts from three individual plants as described [17]. For the degradation assays, intact chloroplasts (1.5×105 chloroplasts/μl) were incubated at 25 °C under 80 μmol of photons·m−2·s−1 for 0, 1.5 and 3 h. At each time point, an equal number of chloroplasts (4.5×105) were solubilized in NuPAGE sample buffer (Invitrogen) and separated by denaturing PAGE as previously described [17]. Immunoblotting was performed as detailed above using antibodies specific for Hsp21.

RESULTS

Loss of ClpR1 causes leaf chlorosis and impaired photosynthesis

In an earlier study, we identified an Arabidopsis EMS mutant lacking ClpR1 protein, termed clpR1-1 [22]. However, since only 5-day-old seedlings with chlorotic cotyledons were used for most of this previous work, we undertook a more comprehensive analysis of clpR1-1 in more mature plants. The growth conditions were also changed from that used previously [22] to optimize the growth of the clpR1-1 line relative to the wild-type. As shown in Figure 1(A), the clpR1-1 line had a slow-growing chlorotic phenotype. Chlorosis was most severe in younger inner leaves, and as the leaves expanded the severity of chlorosis gradually lessened. The growth in clpR1-1 was delayed by approx. 2 weeks relative to the wild-type. Owing to this, all comparisons between wild-type Arabidopsis and clpR1-1 were performed on plants of the same developmental size rather than age.

Loss of ClpR1 produces a chlorotic phenotype, and reduces chlorophyll content and photosynthetic performance

Figure 1
Loss of ClpR1 produces a chlorotic phenotype, and reduces chlorophyll content and photosynthetic performance

(A) Phenotypic comparison of plants at the same developmental stage: 5- and 7-week-old wild-type (Wt) Arabidopsis and clpR1-1, respectively. Plants were grown under the same standard conditions of 23/18 °C day/night temperatures, 8 h photoperiod with 150 μmol·m−2·s−1 light, and 65% RH. Chlorophyll content (B) and photochemical efficiency of PSII (FV/FM) (C) in outer and inner leaves of wild-type (Wt) and clpR1-1 plants. Values shown are averages±S.E.M. (n=3). (D) Photosynthetic ETR rates in outer and inner leaves of Wt and clpR1-1 plants. ETR rates were measured at different PAR (photosynthetically active radiation) levels between 0 and 400 μmol·m−2·s−1. Values shown are from three independent wild-type (Wt) and clpR1-1 plants. In (B)–(D), plants were compared at the same developmental stage as shown in (A).

Figure 1
Loss of ClpR1 produces a chlorotic phenotype, and reduces chlorophyll content and photosynthetic performance

(A) Phenotypic comparison of plants at the same developmental stage: 5- and 7-week-old wild-type (Wt) Arabidopsis and clpR1-1, respectively. Plants were grown under the same standard conditions of 23/18 °C day/night temperatures, 8 h photoperiod with 150 μmol·m−2·s−1 light, and 65% RH. Chlorophyll content (B) and photochemical efficiency of PSII (FV/FM) (C) in outer and inner leaves of wild-type (Wt) and clpR1-1 plants. Values shown are averages±S.E.M. (n=3). (D) Photosynthetic ETR rates in outer and inner leaves of Wt and clpR1-1 plants. ETR rates were measured at different PAR (photosynthetically active radiation) levels between 0 and 400 μmol·m−2·s−1. Values shown are from three independent wild-type (Wt) and clpR1-1 plants. In (B)–(D), plants were compared at the same developmental stage as shown in (A).

The amount of chlorophyll in clpR1-1 compared with wild-type was analysed in both outer and inner leaf whorls, given the difference in severity of chlorosis. In wild-type plants, the chlorophyll content of older outer leaves was significantly lower [1.59±0.03 nmol·mg of FW (fresh weight)−1; n=3] than in younger inner leaves (2.72±0.03 nmol·mg of FW−1; n=3) (Figure 1B). The opposite was true in clpR1-1, with the younger inner leaves having much less chlorophyll (0.59±0.04 nmol·mg of FW−1; n=3) compared with the outer leaves (1.1±0.04 nmol·mg of FW−1; n=3), consistent with their more severe chlorotic appearance. Despite the partial recovery in the older outer leaves of clpR1-1, the chlorophyll content remained significantly lower than in the outer wild-type leaves.

We next compared different photosynthetic parameters in the clpR1-1 leaves relative to those of the wild-type. Measuring first the photochemical efficiency of PSII (photosystem II) (FV/FM) revealed no significant differences between inner (0.839±0.001; n=3) and outer (0.837±0.002; n=3) leaves in the wild-type (Figure 1C). In contrast, the chlorotic inner leaves of clpR1-1 showed a marked inhibition of PSII photochemical efficiency (0.723±0.006; n=3). Although this recovered somewhat in the older outer leaves (0.758±0.011; n=3), the photochemical efficiency remained less than that in wild-type outer leaves. Photosynthetic electron transport rates also showed a decrease in both photosynthetic efficiency (i.e., quantum yield as determined by the initial slope of the light-response curve) and capacity (as measured at the highest irradiance) in both the inner and outer leaves of clpR1-1 (Figure 1D). Again, the younger inner leaves were more drastically affected, correlating with the partial recovery of the phenotype in more developed clpR1-1 plants.

Reduced levels of photosynthetic proteins in clpR1 mutant

Owing to the reduced photosynthetic performance and chlorophyll content in the clpR1 mutant, we next analysed the relative amounts of photosynthetic protein complexes using antibodies specific to the following marker proteins: SSU (small subunit) of Rubisco (ribulose-1,5-bisphosphate carboxylase/oxygenase); Lhcb2 and D1 for the PSII outer antennae and reaction centre respectively; PsaL for PSI and β-subunit of ATPase (Figure 2). The levels of all marker proteins decreased significantly (40–55%) in the inner leaves of clpR1-1 relative to the wild-type, consistent with the chlorotic phenotype and reduced chlorophyll levels. In comparison, no significant change in Rubisco, PSI, PSII reaction centre or ATPase contents were observed in the outer leaves of clpR1-1 relative to the wild-type, whereas a small decrease (approx. 15%) was observed in the levels of PSII outer antennae (results not shown). Interestingly, a larger-molecular- mass form of Lhcb2 accumulated in the clpR1 mutant in both inner (Figure 2A) and outer leaves (results not shown). This had been observed previously [22] as well as in a clpR2 knock-down mutant in which it was demonstrated to be an unprocessed precursor of Lhcb2 containing the chloroplast transit peptide [20]. When summed together, the abundance of Lhcb2 was only reduced by 20% in the inner leaves of clpR1-1 and unchanged in the outer leaves.

Changes in levels of photosynthetic proteins in the clpR1 mutant

Figure 2
Changes in levels of photosynthetic proteins in the clpR1 mutant

(A) Amounts of marker proteins for different photosynthetic protein complexes in inner leaves of 5-week-old wild-type Arabidopsis and 7-week-old clpR1-1 were determined by immunoblotting. Total cell extracts were isolated and separated based on equal protein content by 1D-PAGE. Antibodies were used to detect specific marker proteins for each of the photosynthetic protein complexes: SSU for Rubisco; Lhcb2 and D1 for PSII; PsaL for PSI; and the β-subunit of ATPase. (B) Quantification of the amount of each photosynthetic marker protein in the inner leaves of clpR1-1 relative to wild-type plants. Values shown are averages±S.E.M. (n=3) with the wild-type values set at 100%.

Figure 2
Changes in levels of photosynthetic proteins in the clpR1 mutant

(A) Amounts of marker proteins for different photosynthetic protein complexes in inner leaves of 5-week-old wild-type Arabidopsis and 7-week-old clpR1-1 were determined by immunoblotting. Total cell extracts were isolated and separated based on equal protein content by 1D-PAGE. Antibodies were used to detect specific marker proteins for each of the photosynthetic protein complexes: SSU for Rubisco; Lhcb2 and D1 for PSII; PsaL for PSI; and the β-subunit of ATPase. (B) Quantification of the amount of each photosynthetic marker protein in the inner leaves of clpR1-1 relative to wild-type plants. Values shown are averages±S.E.M. (n=3) with the wild-type values set at 100%.

Reduced levels of Clp proteolytic core in clpR1 mutant

The plastidic Clp proteolytic core contains multiple ClpP and ClpR paralogues in addition to ClpT1 and ClpT2 in an approx. 335 kDa complex [16]. The core is composed of two distinct subcomplexes, one containing ClpP1 and ClpR1–R4, and the other containing ClpP3–P6 [17]. Given that all subunits of the proteolytic core other than ClpR1 appear to be essential for plant viability, we next investigated how the loss of ClpR1 affected the formation of the core complex. Stromal proteins isolated from both wild-type and clpR1-1 plants were separated by native PAGE and then used for immunoblotting to detect the intact Clp proteolytic core (335 kDa) and the ClpP1, ClpR1–R4 sub-complex (230 kDa) (Figure 3A and Supplementary Figure S1 at http://www.BiochemJ.org/bj/417/bj4170257add.htm). Immunoblotting confirmed that, even though ClpR1 was totally absent in the mutant line, a small amount of the Clp proteolytic core remained [9.9%±4.7% (n=23) as quantified from the ClpP1, ClpR2, ClpR4, ClpP3–P6 and ClpT1 antibodies]. Analysis of the ClpP1/ClpR1–R4 sub-complex revealed a two-fold decrease in the ClpP1, ClpR2 and ClpR4 proteins in the clpR1 mutant, but little change to ClpR3 content (Figure 3A). This infers that ClpR3 is probably able to compensate partially for the lack of ClpR1, forming some stable ClpP1/ClpR2–R4 sub-complex and thereby the residual amount of Clp proteolytic core.

Clp proteolytic core complexes in wild-type Arabidopsis and clpR1 mutant

Figure 3
Clp proteolytic core complexes in wild-type Arabidopsis and clpR1 mutant

Clp proteolytic core complexes from 3-week-old wild-type (Wt) Arabidopsis and 5 week clpR1-1 were separated by native PAGE on the basis of equal protein content. (A) The intact Clp proteolytic core complex (335 kDa) and the sub-complex containing ClpP1 and ClpR1–R4 (230 kDa) were visualized by immunoblotting using specific antibodies as indicated below each panel. The graph shows the quantification of the amount of the 230 kDa Clp proteolytic core sub-complex in clpR1-1 relative to wild-type. Values shown are averages±S.E.M. (n=3), with the wild-type values set at 100%. (B) The intact Clp proteolytic core complex (335 kDa) and the sub-complexes containing ClpP3–P6 (180 kDa) or ClpP3–P6, ClpT1 (200 kDa) were visualized by immunoblotting using specific antibodies as indicated below each panel. The graph shows the quantification of the amount of the 180 and 200 kDa Clp sub-complexes in clpR1-1 relative to Wt. Values shown are averages±S.E.M. (n=3), with the wild-type values set at 100%.

Figure 3
Clp proteolytic core complexes in wild-type Arabidopsis and clpR1 mutant

Clp proteolytic core complexes from 3-week-old wild-type (Wt) Arabidopsis and 5 week clpR1-1 were separated by native PAGE on the basis of equal protein content. (A) The intact Clp proteolytic core complex (335 kDa) and the sub-complex containing ClpP1 and ClpR1–R4 (230 kDa) were visualized by immunoblotting using specific antibodies as indicated below each panel. The graph shows the quantification of the amount of the 230 kDa Clp proteolytic core sub-complex in clpR1-1 relative to wild-type. Values shown are averages±S.E.M. (n=3), with the wild-type values set at 100%. (B) The intact Clp proteolytic core complex (335 kDa) and the sub-complexes containing ClpP3–P6 (180 kDa) or ClpP3–P6, ClpT1 (200 kDa) were visualized by immunoblotting using specific antibodies as indicated below each panel. The graph shows the quantification of the amount of the 180 and 200 kDa Clp sub-complexes in clpR1-1 relative to Wt. Values shown are averages±S.E.M. (n=3), with the wild-type values set at 100%.

In addition to the ClpP1, ClpR1–R4 sub-complex, the ClpP3–P6 sub-complexes with (200 kDa) or without ClpT1 (180 kDa) were also examined for possible changes resulting from the loss of ClpR1. As shown in Figure 3(B), there was a significant shift in the proportion of both ClpP3–P6 sub-complexes in clpR1-1 relative to the wild-type. There was a 2–3-fold increase in the larger 200 kDa complex containing ClpT1 in clpR1-1, while the smaller 180 kDa sub-complex decreased by 35–50%. Overall, there was a marginal increase in the combined amount of ClpP3–P6 sub-complexes (on average 25%) in clpR1-1 compared with the wild-type. Despite this, the amount of intact Clp proteolytic core complex (335 kDa) detected in the clpR1-1 mutant using the various ClpP3–P6 and ClpT1 antibodies was again only approx. 10% of the wild-type level (Figure 3B), indicating that the slightly higher amount of the ClpP3–P6 sub-complex could not compensate for the loss of ClpR1. Moreover, no ClpP3–P6 protein was detected in the 230 kDa ClpP1, ClpR1–R4 sub-complex, again indicating that neither of the ClpP3–P6 proteins could substitute for ClpR1 within the Clp proteolytic core complex.

Identification of new protein substrates for the chloroplast Clp protease

Recent research on clpP6 antisense lines revealed the first six putative substrates for the chloroplast Clp protease in higher plants [17]. Given that these substrates were all involved in various constitutive metabolic and homoeostatic functions, we first tested if the Clp protease also targeted chloroplast polypeptides associated with certain stress responses. The first protein examined was Lox2 (lipoxygenase 2), which is required for the wound-induced accumulation of jasmonic acid in Arabidopsis [28]. Since Lox2 is constitutively expressed, in addition to being stress inducible, we simply examined the degradation rate of Lox2 in chloroplasts isolated from non-stressed leaves. Comparing wild-type and clpR1-1 chloroplasts, however, revealed no significant difference in the degradation of Lox2 during the 3 h time course (Figure 4). We next examined an exclusively stress-inducible protein that is not synthesized under normal growth conditions, the major small Hsp localized in chloroplasts of higher plants (Hsp21) during heat stress. It is well known that, following its induction and accumulation in heat-stressed leaves, stromal Hsp21 is steadily degraded during the recovery period at lower temperatures [27,29]. To determine if the Clp protease was responsible for this degradation of Hsp21, we isolated intact chloroplasts from both wild-type Arabidopsis and clpR1-1 plants at the early stages of this recovery phase. As shown in Figure 4, however, no significant difference was observed in the degradation rate of Hsp21 in wild-type and clpR1-1 chloroplasts.

Chloroplast Clp protease does not degrade certain stress-related polypeptides

Figure 4
Chloroplast Clp protease does not degrade certain stress-related polypeptides

(A) Equal amounts of intact chloroplasts from wild-type (Wt) and clpR1-1 plants were incubated for 0–3 h in the presence of light and ATP. For Lox2, plants were grown under standard growth conditions. For Hsp21, plants were first heat shocked at 40 °C as described in the Experimental section and then allowed to recover for 65 h under the standard growth conditions. Isolated chloroplast proteins from different time points were separated by denaturing PAGE, and the Hsp21 and Lox2 proteins were visualized by immunoblotting using specific antibodies. (B) Degradation of Hsp21 and Lox2 in wild-type (Wt) and clpR1-1 chloroplasts. Values shown are averages±S.E.M. (n=3), where protein content from the 0 h sample was set to 100%.

Figure 4
Chloroplast Clp protease does not degrade certain stress-related polypeptides

(A) Equal amounts of intact chloroplasts from wild-type (Wt) and clpR1-1 plants were incubated for 0–3 h in the presence of light and ATP. For Lox2, plants were grown under standard growth conditions. For Hsp21, plants were first heat shocked at 40 °C as described in the Experimental section and then allowed to recover for 65 h under the standard growth conditions. Isolated chloroplast proteins from different time points were separated by denaturing PAGE, and the Hsp21 and Lox2 proteins were visualized by immunoblotting using specific antibodies. (B) Degradation of Hsp21 and Lox2 in wild-type (Wt) and clpR1-1 chloroplasts. Values shown are averages±S.E.M. (n=3), where protein content from the 0 h sample was set to 100%.

Given the lack of degradation of the stress-related proteins Lox2 and Hsp21 by the chloroplast Clp protease, we next pursued a more comprehensive search for additional constitutive protein substrates. For this, a similar approach to that described previously [17] was used, except that the clpR1-1 line was used instead of the clpP6 antisense lines, since it has less of the Clp proteolytic core remaining and a more stable phenotype among individual plants. The methodology employed was first to identify significant changes in stromal protein content in clpR1-1 relative to the wild-type. This was based on the rationale that substrates should accumulate when the levels of Clp protease are considerably reduced. 1D denaturing PAGE gels were used for proteins more than 70 kDa (results not shown) and 2D-PAGE was used to resolve smaller proteins, less than 70 kDa. Ten proteins accumulated more than 3-fold in the clpR1 mutant (Figure 5, Table 1). Of these, all were detected in the wild-type except for a putative ribulose 5-phosphate isomerase (protein number 6). The identity of the remaining nine proteins were three other Calvin cycle enzymes, FBP ALD (fructose bisphosphate aldolase) and two paralogues of the Rubisco SSU, SSU 1A and SSU 2B; as well as two RNA-binding proteins (both cp29 types), a UPRT (uracil phosphoribosyltransferase), a cyclophilin PPIase (peptidyl-prolyl cis-trans isomerase), a NDP (nucleoside diphosphate) kinase, and a putative fibrillin. Five of these proteins, ribulose 5-phosphate isomerase, RNA-binding protein cp29, UPRT, PPIase and NDP kinase, were previously identified as proteins that also accumulated in the clpP6 antisense lines [17].

Changes in stromal protein composition in wild-type Arabidopsis and clpR1-1 line

Figure 5
Changes in stromal protein composition in wild-type Arabidopsis and clpR1-1 line

Stromal proteins isolated from 3-week-old wild-type (Wt) Arabidopsis and 5-week-old clpR1-1 plants were separated by 2D-PAGE and visualized by Flamingo™ staining. Proteins consistently more abundant in the clpR1-1 plants relative to the wild-type are circled and numbered. Shown are representative results from three replicates. The identity of each numbered protein is detailed in Table 1.

Figure 5
Changes in stromal protein composition in wild-type Arabidopsis and clpR1-1 line

Stromal proteins isolated from 3-week-old wild-type (Wt) Arabidopsis and 5-week-old clpR1-1 plants were separated by 2D-PAGE and visualized by Flamingo™ staining. Proteins consistently more abundant in the clpR1-1 plants relative to the wild-type are circled and numbered. Shown are representative results from three replicates. The identity of each numbered protein is detailed in Table 1.

Table 1
Identification of stromal proteins accumulating in clpR1-1 line

Identification by MALDI–TOF MS and HPLC-MS/MS of those proteins most abundant in the clpR1 mutant relative to the wild type. Each numbered protein corresponds to the same numbered protein circled in Figure 4. 1Molecular mass calculated from gel size markers. 2Protein identification as Arabidopsis Genome Initiative gene code. 3Protein scores >66 are significant (P<0.05). 4Peptide match at mass tolerance of 6100 p.p.m., allowing a maximum of one missed cleavage. 5Fold upregulation based on quantification of stained gel spots.

Spot number Identity Mass (kDa)1 Protein identified2 Protein score3 Peptides matched4 Fold5 
FBP ALD 42 At2g21330 817 14 
Putative RNA-binding protein (cp29) 33 At3g53460 571 11 
Putative RNA-binding protein (cp29) 31 At2g37220 460 
Putative fibrillin 30 At4g04020 316 3.5 
Putative ribose 5-phosphate isomerase 28 At3g04790 534 11 
PPIase 28 At3g62030 581 12 ∞ 
UPRT 22 At3g53900 432 18 10 
NDP kinase 19 At5g63310 182 
Rubisco SSU 2B 14 At5g38420 134 13 
10 Rubisco SSU 1A 14 At1g67090 119 15 
Spot number Identity Mass (kDa)1 Protein identified2 Protein score3 Peptides matched4 Fold5 
FBP ALD 42 At2g21330 817 14 
Putative RNA-binding protein (cp29) 33 At3g53460 571 11 
Putative RNA-binding protein (cp29) 31 At2g37220 460 
Putative fibrillin 30 At4g04020 316 3.5 
Putative ribose 5-phosphate isomerase 28 At3g04790 534 11 
PPIase 28 At3g62030 581 12 ∞ 
UPRT 22 At3g53900 432 18 10 
NDP kinase 19 At5g63310 182 
Rubisco SSU 2B 14 At5g38420 134 13 
10 Rubisco SSU 1A 14 At1g67090 119 15 

An in organello chloroplast degradation assay was then used to compare the stability of different stromal proteins in wild-type Arabidopsis and the clpR1 mutant. Both 1D- and 2D-PAGE were again used to resolve the degradation of stromal proteins over a 3 h time course. The amounts of protein at zero time and 3 h were quantified for both wild-type and clpR1-1, based on staining with either Coomassie Blue (1D-PAGE) or Flamingo™ (2D-PAGE). A slower degradation rate in clpR1-1 was observed for three large proteins: EF (elongation factor)-Ts, the molecular chaperone Hsp90, and a RNA helicase (Figure 6A and Table 2). After 3 h, the amount of each of these proteins decreased in wild-type by 25–45%, whereas in clpR1-1, no significant change was observed from the zero time levels (Figure 7A). Each of these proteins were also more abundant in clpR1-1 relative to wild-type, albeit by less than 2-fold. These three proteins were all identified earlier as substrates in the clpP6 antisense plants [17]. In addition, 22 small-molecular-mass substrates were also identified (Figure 6B and Table 2), including eight of the proteins that accumulated in clpR1-1 (Table 1); the other two proteins, SSU 1A and SSU 2B (spots 9 and 10), did not decrease in the wild-type during the 3 h degradation assay and were therefore not considered potential substrates. As shown in Figure 7(A), all the putative substrates significantly decreased (25–80% loss) in the wild-type throughout the 3 h degradation assay. In the case of one substrate (RNA-binding protein, spot 2), the protein could not be detected in wild-type chloroplasts after 3 h, indicating that most if not all of it was degraded during the assay time course. Most of these identified proteins (16 of the 22) were very stable in clpR1-1, and their levels did not significantly change during the 3 h time course. No degradation rate could be determined for PPIase (protein number 6) in the wild-type, as it was not detected at either time point with the fluorescent stain used. However, since PPIase was earlier identified as a putative substrate in the wild-type using the more sensitive silver staining [17], its lack of degradation in the clpR1 mutant confirmed this original finding, and therefore it was included as a substrate in this study. For the six remaining putative substrates, a slight decrease in the amount of each protein was observed in clpR1-1, ranging from 10–20%. Of the 25 substrate proteins identified, all but six were also more abundant in clpR1-1 (Table 2).

Identification of protein substrates for the chloroplast Clp protease

Figure 6
Identification of protein substrates for the chloroplast Clp protease

Intact chloroplasts from 3-week-old wild-type (Wt) Arabidopsis and 5-week-old clpR1-1 plants were incubated for 3 h in the presence of ATP and light. Chloroplasts were then ruptured and fractionated. Stromal proteins were separated by SDS/PAGE and visualized by Coomassie Blue staining (A) or by 2D-PAGE and visualized by Flamingo™ (B). Those proteins whose abundance significantly decreased over the 3 h time course in wild-type, but not in clpR1-1, were identified by MALDI–TOF MS. Shown are representative gels from three independent degradation assay replicates. (A) Arrows at the right indicate the three putative protein substrates. (B) Circled and numbered spots indicate 23 putative protein substrates. Where appropriate, the number corresponds to that in Figure 5. The identity of each numbered protein is detailed in Table 2.

Figure 6
Identification of protein substrates for the chloroplast Clp protease

Intact chloroplasts from 3-week-old wild-type (Wt) Arabidopsis and 5-week-old clpR1-1 plants were incubated for 3 h in the presence of ATP and light. Chloroplasts were then ruptured and fractionated. Stromal proteins were separated by SDS/PAGE and visualized by Coomassie Blue staining (A) or by 2D-PAGE and visualized by Flamingo™ (B). Those proteins whose abundance significantly decreased over the 3 h time course in wild-type, but not in clpR1-1, were identified by MALDI–TOF MS. Shown are representative gels from three independent degradation assay replicates. (A) Arrows at the right indicate the three putative protein substrates. (B) Circled and numbered spots indicate 23 putative protein substrates. Where appropriate, the number corresponds to that in Figure 5. The identity of each numbered protein is detailed in Table 2.

Putative protein substrates of the chloroplast Clp protease and their abundance during development

Figure 7
Putative protein substrates of the chloroplast Clp protease and their abundance during development

(A) Extent of degradation for 25 putative protein substrates of the Clp protease in intact chloroplasts from wild-type (Wt) Arabidopsis and clpR1-1 plants over the 3 h time course. Stromal proteins were separated by 1D- or 2D-PAGE and quantified after visualization by Coomassie Blue or Flamingo™ respectively. The relative amount of protein remaining after 3 h is shown as an average±S.E.M. (n=3) plotted as a percentage of the zero time value, which was set to 100%. Each number corresponds to a numbered protein in Figure 5, the identity of which is detailed in Table 2. *Protein that accumulated in clpR1-1 but was not detected in wild-type at t=0 or at t=3. It has previously been identified as a protein substrate using clpP6 antisense lines [17]. (B) Stromal proteins from inner and outer leaves of 8-week-old wild-type Arabidopsis were separated by 2D-PAGE and quantified by Flamingo™. The relative amounts of each newly identified substrate are shown as an average±S.E.M. (n=3) plotted as a percentage of the inner leaf value, set to 100%. Each number corresponds to a numbered protein in Figure 6, which is further detailed in Table 2.

Figure 7
Putative protein substrates of the chloroplast Clp protease and their abundance during development

(A) Extent of degradation for 25 putative protein substrates of the Clp protease in intact chloroplasts from wild-type (Wt) Arabidopsis and clpR1-1 plants over the 3 h time course. Stromal proteins were separated by 1D- or 2D-PAGE and quantified after visualization by Coomassie Blue or Flamingo™ respectively. The relative amount of protein remaining after 3 h is shown as an average±S.E.M. (n=3) plotted as a percentage of the zero time value, which was set to 100%. Each number corresponds to a numbered protein in Figure 5, the identity of which is detailed in Table 2. *Protein that accumulated in clpR1-1 but was not detected in wild-type at t=0 or at t=3. It has previously been identified as a protein substrate using clpP6 antisense lines [17]. (B) Stromal proteins from inner and outer leaves of 8-week-old wild-type Arabidopsis were separated by 2D-PAGE and quantified by Flamingo™. The relative amounts of each newly identified substrate are shown as an average±S.E.M. (n=3) plotted as a percentage of the inner leaf value, set to 100%. Each number corresponds to a numbered protein in Figure 6, which is further detailed in Table 2.

Table 2
Identity of putative protein substrates for the chloroplast Clp protease as determined by MALDI–TOF MS and HPLC-MS/MS

Each numbered protein corresponds to the same numbered protein circled in Figure 5. 1Molecular mass calculated from gel size markers. 2Protein identification as Arabidopsis Genome Initiative gene code. 3Protein scores >66 are significant (P<0.05). 4Peptide match at mass tolerance of 6100 p.p.m., allowing a maximum of one missed cleavage. 5Fold upregulation based on quantification of stained gel spots. *Not detected in wild-type, previously identified as a substrate by Sjögren et al. [17].

Spot No. Identity Mass (kDa)1 Protein identified2 Protein score3 Peptides matched4 Fold5 
FBP ALD 42 At2g21330 817 14 
RNA-binding protein 34 At3g53460 571 11 
Putative RNA-binding protein (cp29) 31 At2g37220 460 
Putative fibrillin 30 At4g04020 316 3.5 
Putative ribose 5-phosphate isomerase 28 At3g04790 534 11 
6* PPIase 28 At3g62030 581 12 ∞ 
UPRT 22 At3g53900 432 18 10 
NDP kinase 19 At5g63310 182 
11 EF-Ts 150 At4g29060 222 14 1.5 
12 Hsp90-5 87 At2g04030 184 19 1.6 
13 RNA helicase 3 85 At5g26742 178 16 1.8 
14 EF-G 79 At1g62750 176 17 1.6 
15 Hsp70-7 70 At5g49910 100 11 1.7 
16 Hsp70-6 70 At4g24280 103 11 1.6 
17 Chaperonin 60α 59 At2g28000 198 17 
18 Chaperonin 60β 59 At1g55490 98 11 1.7 
19 EF-Tu 50 At4g20360 108 12 1.7 
20 Phosphoglycerate kinase 48 At3g12780 129 11 
21 GS2 46 At5g35630 808 16 
22 Phosphoribulose kinase 42 At1g32060 132 11 
23 Coproporphyrinogen III oxidase 41 At1g03475 660 15 
24 Putative UTase 36 At1g16880 470 10 1.2 
25 Hypothetical protein 35 At2g37660 663 14 
26 Triose phosphate isomerase 32 At2g21170 716 12 
27 2-cys peroxiredoxin A 27 At3g11630 476 1.4 
Spot No. Identity Mass (kDa)1 Protein identified2 Protein score3 Peptides matched4 Fold5 
FBP ALD 42 At2g21330 817 14 
RNA-binding protein 34 At3g53460 571 11 
Putative RNA-binding protein (cp29) 31 At2g37220 460 
Putative fibrillin 30 At4g04020 316 3.5 
Putative ribose 5-phosphate isomerase 28 At3g04790 534 11 
6* PPIase 28 At3g62030 581 12 ∞ 
UPRT 22 At3g53900 432 18 10 
NDP kinase 19 At5g63310 182 
11 EF-Ts 150 At4g29060 222 14 1.5 
12 Hsp90-5 87 At2g04030 184 19 1.6 
13 RNA helicase 3 85 At5g26742 178 16 1.8 
14 EF-G 79 At1g62750 176 17 1.6 
15 Hsp70-7 70 At5g49910 100 11 1.7 
16 Hsp70-6 70 At4g24280 103 11 1.6 
17 Chaperonin 60α 59 At2g28000 198 17 
18 Chaperonin 60β 59 At1g55490 98 11 1.7 
19 EF-Tu 50 At4g20360 108 12 1.7 
20 Phosphoglycerate kinase 48 At3g12780 129 11 
21 GS2 46 At5g35630 808 16 
22 Phosphoribulose kinase 42 At1g32060 132 11 
23 Coproporphyrinogen III oxidase 41 At1g03475 660 15 
24 Putative UTase 36 At1g16880 470 10 1.2 
25 Hypothetical protein 35 At2g37660 663 14 
26 Triose phosphate isomerase 32 At2g21170 716 12 
27 2-cys peroxiredoxin A 27 At3g11630 476 1.4 

Relative abundance of identified substrates in developing leaves

Since the chlorotic phenotype of the clpR1 mutant was more severe in younger leaves relative to mature ones, we next analysed the relative amount of each newly identified protein substrate in young (inner) and mature (outer) leaves of the wild-type. It should be noted that the corresponding analyses for substrates 6–8 and 11–13 have already been shown [17]. As shown in Figure 7(B), around half of the new protein substrates (spots 3, 14, 16, 20, 21, 23–25) showed no significant variation in amount between the younger inner and mature outer leaves. Most of these unchanged polypeptides were enzymes within the different metabolic processes, whereas some, such as EF-G and Hsp70-6, were involved in protein synthesis and maturation. Of the remaining protein substrates, four (19, 22, 26 and 27) were comparatively more abundant in wild-type outer leaves, with again a mixture of metabolic enzymes and general housekeeping proteins. In contrast, five protein substrates (2, 4, 15, 17 and 18) were significantly less abundant in the outer leaves, all of which are involved in homoeostasis or structural functions within the chloroplast. For the two remaining putative substrates (1 and 5), no data could be obtained, since both had relatively low abundance and could not be reliably detected in the replicate experiments. Altogether, when combining the new substrates with those identified previously [17], more than a third of the total number of substrates (9 out of 25) were more abundant in the younger inner leaves, correlating with the more severe chlorotic phenotype in these tissues.

DISCUSSION

In this study, we have detailed the phenotypic effects resulting from the loss of ClpR1 in Arabidopsis, as well as identified many more possible protein substrates for the chloroplast Clp protease. The clpR1-1 line is apparently the only viable knockout mutant for any of the chloroplast-localized ClpP and ClpR proteins. The phenotype of clpR1-1 was very similar to that previously observed using clpP4 and clpP6 antisense lines [17,21] and a clpR2 knock-down T-DNA insertion line [20]. Plants were chlorotic, with the younger inner leaves more affected than the older outer leaves. This chlorosis in young leaves affected their rate of development, being approx. 2 weeks behind those of the wild-type. The overall phenotype of clpR1-1 was, unsurprisingly, more severe than that of the clpP6 antisense lines [17], since all the ClpR1 protein was lost in the mutant line. Although leaf chlorosis lessened with age in clpR1-1, the outer leaves never completely greened, whereas the outer leaves of clpP6 antisense lines fully recovered to become indistinguishable from the wild-type. The resulting delay in growth was also approx. 1 week longer in clpR1-1 relative to the clpP6 antisense lines. The clpR1-1 line had significantly lower chlorophyll content and impaired photosynthetic activity in both younger inner and older outer leaves, again consistent with the somewhat more severe phenotype of this mutant compared to the clpP6 antisense lines, in which significant changes only occurred in inner leaves relative to wild-type [17]. The amounts of the major photosynthetic protein complexes were also reduced by 40–50% in the chlorotic inner leaves of clpR1-1, again more severe than that in the corresponding leaves of the clpP6 antisense lines (30–40% reduction). These features suggest that although the Clp protease obviously plays a more critical role in chloroplasts of younger developing leaves, it still performs an important function in more mature leaves.

The model Clp proteolytic core in E. coli is composed of 14 identical ClpP subunits arranged in two face-to-face heptameric rings [6]. In the chloroplast stroma of higher plants, a single heterogeneous Clp proteolytic core exists, consisting of five ClpP and four ClpR paralogues [14]. Additionally, this oligomer contains two novel proteins, ClpT1 and ClpT2, that associate peripherally to the main ClpP/ClpR complex, forming a proteolytic core of 325–350 kDa [16]. The Clp proteolytic core in Arabidopsis chloroplasts consists of two sub-complexes, presumably heptameric rings, one containing ClpP1 and ClpR1–R4 and the other with ClpP3–P6, which can also associate to ClpT1 [17]. Interestingly, loss of ClpR1 greatly reduced the amount of the core complex (by 90%), but did not completely abolish it. Indeed, the 230 kDa sub-complex normally containing ClpR1 had less than half the wild-type levels of ClpP1, ClpR2 and ClpR4 in clpR1-1, but unchanged amounts of ClpR3. This relative increase of ClpR3 inferred that it was partially compensating for the absent ClpR1 within the 230 kDa sub-complex. This modified sub-complex must then associate, albeit with much lower affinity, to the ClpP3–P6 sub-complex, forming sufficient amounts of functional, or partially functional, Clp proteolytic core (i.e. 10%) to enable the clpR1-1 line to remain viable. The possibility that ClpR3 can partially substitute for ClpR1 function is supported by the fact that ClpR3 has the greatest sequence similarity to ClpR1 relative to the other subunits of the Clp proteolytic core [14]. The partial replacement of ClpR1 with ClpR3 in the clpR1-1 line is the only evidence to date for any degree of redundancy regarding the stoichiometry and oligomerization of the Clp proteolytic core. Mutant lines for any of the other subunits of the proteolytic core complex have so far proven lethal, suggesting they are each essential for the formation of an active chloroplast Clp protease.

Interestingly, loss of ClpR1 caused a slight increase in the amount of the ClpP3–P6 sub-complex, with a significantly greater proportion of the larger sub-complex containing ClpT1. This is in contrast with when the level of the ClpP3–P6 sub-complex is greatly reduced, as in Arabidopsis clpP6 antisense lines, with the ClpP1, ClpR1–R4 sub-complex also decreasing but to a lesser extent (i.e. by 25%; [17]). The accumulation of the larger ClpP3–P6, ClpT1 sub-complex in clpR1-1 suggests that this is an intermediate assembly step prior to the association of the two different sub-complexes (i.e. ClpP1, ClpR1–R4 and ClpP3–P6) into the intact Clp proteolytic core complex, although the mechanistic details of how this process occurs remains unclear. Overall, it appears that the fully active form of the chloroplast Clp proteolytic core requires the correct stoichiometry of each ClpP and ClpR subunit, which recently has been proposed for each sub-complex as 1:(3):1:2 for ClpP1, (-R1+-R3), -R2, -R4, and 1:3:2:1 for ClpP3, -P4, P5 and -P6 [17].

One of the more exciting advances in the field of the ATP-dependent Clp protease in chloroplasts has been the development of an in organello proteolytic assay that can be used for identifying potential native protein substrates [17]. Until recently the only known substrate for chloroplastic Clp proteases was the cytochrome b6/f complex identified in the green alga Chlamydomonas reinhardtii during nitrogen starvation [30]. Using this assay to compare wild-type Arabidopsis with clpP6 antisense lines, six substrates were identified for the stromal Clp protease. The function of all six proteins was generally related to chloroplast homoeostasis and not to any specific metabolic pathways, supporting the hypothesis that the Clp protease has a major ‘housekeeping’ role. In this study, we have expanded upon this early work to now identify many more putative protein substrates for the chloroplast Clp protease. Up to 25 soluble proteins were found as potential substrates, including the six that were earlier identified (EF-Ts, Hsp90, RNA helicase, UPRT, PPIase and NDP kinase [17]). Most of the identified proteins accumulated in the clpR1 mutant relative to the wild-type, as would be expected for a substrate whose degradation was impaired. Of the 25 substrate proteins, most (approx. 60%) were related to general housekeeping activities, again consistent with the Clp protease being primarily involved in chloroplast homoeostasis. Most of the housekeeping representatives were molecular chaperones. Hsp90 was one of the original substrates identified, and is involved in regulating the activity of key regulatory proteins in addition to co-operating with other chaperone systems, such as Hsp70 and PPIase in the correct folding of nascent polypeptides [31]. Interestingly, as with Hsp90, PPIase was also identified as a substrate in the previous search and we have now found two Hsp70 proteins, Hsp70-7 and Hsp70-6. In addition, the two subunits of the chloroplast chaperonin 60 (Cpn60α/β) were identified as substrates, which together are essential for the correct folding, and assembly when necessary, of many soluble proteins in the chloroplast [32].

Several components of chloroplast protein synthesis were also identified as putative substrates of the Clp protease. EF-Ts, another of the initial substrates found [17], acts as a nucleotide exchange factor for the EF-Tu. EF-Tu itself, along with the homologous EF-G, also now appear to be substrates. EF-Tu mediates the entry of the aminoacyl-tRNA into a free site of the ribosome, then EF-Ts catalyses the release of GDP from EF-Tu, enabling EF-Tu to bind to a new GTP molecule and go on to catalyse another aminoacyl-tRNA addition. EF-G then catalyses the translocation of the tRNA and mRNA down the ribosome at the end of each round of polypeptide elongation [33]. Participation of these elongation factors significantly accelerates the elongation of nascent polypeptides by ribosomes [34], and as such their degradation by the Clp protease might play a key regulatory role in the process of chloroplast protein synthesis. Proteins involved in RNA maturation were also identified as possible Clp proteolytic substrates. RNA-binding proteins play a major role in regulating mRNA metabolism in chloroplasts, and three types are found in Arabidopsis, cp29, cp31 and cp33 [35]. Here we identified two RNA-binding proteins (both cp29 types) that accumulated in clpR1-1, both of which turned out also to be substrates. Moreover, RNA helicase 3 (another of the original substrates [17]), was also found as a Clp proteolytic substrate in Arabidopsis. RNA helicases are essential for many homoeostatic processes such as transcription, mRNA processing, initiation of translation and ribosome biogenesis [36]. Given their functions, it is possible that impaired degradation of these RNA-related substrates might be related to the delay in chloroplast ribsosomal RNA maturation previously observed in the clpR1-1 line [22]. The final two substrates involved in ‘housekeeping’ activities were UPRT and NDP kinase, proteins also previously identified as substrates [17]. UPRT is involved in salvaging uracil for recycling to allow for continued synthesis of pyrimidine nucleotides [37], whereas NDP kinase helps to regulate the available nucleotide pool by generating nucleoside triphosphates [38].

Although many of the proteins identified in this study as putative substrates for the chloroplast Clp protease perform in general housekeeping functions, almost half (approx. 40%) had more specific cellular functions. Five substrates were enzymes of the photosynthetic Calvin cycle (FBP ALD, ribose 5-phosphate isomerase, phosphoribulokinase, phosphoglycerate kinase and triose-phosphate isomerase), whereas two were enzymes involved in nitrogen metabolism [GS2 (glutamine synthetase 2) and a regulatory protein of GS2 termed UTase (uridylyltransferase)]. GS2 plays an essential role in nitrogen metabolism by catalysing the condensation of glutamate and ammonia to form glutamine. Because of its critical role, GS2 is an enzyme that is tightly regulated by a complex signal transduction pathway. The internal cellular ammonium levels are first detected by UTase, which in turn alters the uridylylation state of PII (GlnB) proteins. These proteins occupy a pivotal position in the network regulating the activity of GS [39,40]. Another substrate of the Clp protease was an enzyme of the chlorophyll and haem biosynthetic pathway, coprogen oxidase (coproporphyrinogen III oxidase). Coprogen oxidase catalyses the oxidative decarboxylation of coprogen (coproporphyrinogen III) to proto-porphyrinogen IX. Coprogen is a very photosensitive tetrapyrrole and regulation of this enzyme guarantees a constant flux of metabolic intermediates and avoids photodynamic damage by accumulating porphyrins [41]. Indeed, it is possible that accumulation of coprogen oxidase in clpR1-1 might result in accumulation of the downstream product Mg-protoIX (Mg-protoporphyrin IX). Accumulation of Mg-protoIX inhibits the accumulation of both nuclear- and plastid-encoded proteins in chloroplasts [42,43], and this might be one explanation for the previously reported low amounts of mRNA encoding for plastid-destined proteins in clpR1-1 [22]. The three remaining protein substrates were the antioxidant 2-Cys periredoxin A, a putative fibrillin and a hypothetical protein predicted to be a nucleoside diphosphate sugar epimerase. Peroxiredoxins are a large group of proteins that participate in cell proliferation, differentiation, apoptosis, and photosynthesis. 2-Cys periredoxin A regulates the concentration of hydrogen peroxide to protect against oxidative damage, as well as regulating peroxide-mediated signal transduction pathways [44]. In plastids of plants, fibrillins are thought to be involved in maintaining the structural stability of plant lipid bodies [45], and are believed to play a role in plant responses to environmental stress [46]. Epimerases or ‘dehydratases’ are enzymes that regulate the stereochemistry of carbohydrates. Carbohydrates play important roles in photo-assimilation as well as serving as structural constituents of galactolipids and glycoproteins and transport metabolites. Therefore epimerases are involved in many metabolic pathways [47].

Of the various protein substrates now identified for the chloroplast Clp protease, less than half were more abundant in younger leaves of wild-type plants compared with mature leaves. Of those that were, however, almost all are proteins intricately involved in general housekeeping roles such as protein synthesis, folding and maturation, as well as RNA maturation. Greater amounts of these regulatory proteins are almost certainly related to more metabolically active chloroplasts in younger leaves, with relatively high rates of transcription, translation and protein turnover. As such, therefore, it is consistent that the near absence of the Clp protease due to repression or inactivation of one of the constituent subunits causes more severe phenotypic changes in younger leaves, as observed in this study and others [2022]. The function of these protein substrates in such housekeeping duties is also consistent with the Clp protease being present in plastids other than chloroplasts, including those in the non-photosynthetic parasitic plant Epifagus [48] and Apicomplexan protozoan [49]. Interestingly, although the Clp protease is also present in Arabidopsis roots [16], its relative content is considerably lower than that in leaves [15]. The greater abundance of Clp protease in leaves might not only be related to higher plastid numbers per cell compared with roots, but it might also be due to many of the protein substrates identified in this study functioning in chloroplast-specific processes such as photosynthesis.

We thank Cecilia Emanuelsson (Department of Biochemistry, Lund University, Sweden) for advice on the heat-shock experiments and the Hsp21 antibody, Ivo Feussner (Albrecht-von-Haller Institute for Plant Sciences, Georg-August University, Göttingen, Germany) for the Lox2 antibody, and Stefan Jansson and Gunnar Öquist (Umeå Plant Science Centre, Umeå University, Sweden) for antibodies to the different photosynthetic proteins.

Abbreviations

     
  • BCA

    bicinchoninic acid

  •  
  • 1D

    one-dimensional

  •  
  • 2D

    two-dimensional

  •  
  • DTT

    dithiothreitol

  •  
  • EF

    elongation factor

  •  
  • EMS

    ethylmethanesulfonate

  •  
  • ETR

    electron transport rate

  •  
  • FW

    fresh weight

  •  
  • FBP ALD

    fructose bisphosphate aldolase

  •  
  • GS2

    glutamine synthetase 2

  •  
  • Hsp

    heat-shock protein

  •  
  • IEF

    isoelectric focusing

  •  
  • Lox2

    lipoxygenase 2

  •  
  • MALDI–TOF MS

    matrix-assisted laser-desorption ionization–time-of-flight MS

  •  
  • NDP

    nucleoside diphosphate

  •  
  • PPIase

    peptidyl-prolyl cis-trans isomerase

  •  
  • PS

    photosystem

  •  
  • RH

    relative humidity

  •  
  • Rubisco

    ribulose-1,5-bisphosphate carboxylase/oxygenase

  •  
  • SSU

    small subunit

  •  
  • UPRT

    uracil phosphoribosyltransferase

  •  
  • UTase

    uridylyltransferase

FUNDING

This work was supported by grants to A. K. C. from the Swedish Research Council for Environment, Agricultural Science, and Spatial Planning [grant number 229-2006-948].

References

References
1
Ellis
R. J.
Molecular chaperones: assisting assembly in addition to folding
Trends Biochem. Sci.
2006
, vol. 
31
 (pg. 
395
-
401
)
2
Wickner
S.
Maurizi
M. R.
Gottesman
S.
Posttranslational quality control: folding, refolding, and degrading proteins
Science
1999
, vol. 
286
 (pg. 
1888
-
1893
)
3
Vierstra
R. D.
Proteolysis in plants: mechanisms and functions
Plant Mol. Biol.
1993
, vol. 
32
 (pg. 
275
-
302
)
4
Sakamoto
W.
Protein degradation machineries in plastids
Annu. Rev. Plant Biol.
2006
, vol. 
57
 (pg. 
599
-
621
)
5
Clarke
A. K.
MacDonald
T. M.
Sjögren
L. L. E.
The ATP-dependent Clp protease in chloroplasts of higher plants
Physiol. Plant.
2005
, vol. 
123
 (pg. 
406
-
412
)
6
Wang
J.
Hartling
J. A.
Flanagan
J. M.
The structure of ClpP at 2.3 Å resolution suggests a model for ATP-dependent proteolysis
Cell
1997
, vol. 
91
 (pg. 
447
-
456
)
7
Grimaud
R.
Kessel
M.
Beuron
F.
Stevens
A. C.
Enzymatic and structural similarities between the Escherichia coli ATP-dependent proteases, ClpXP and ClpAP
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
12476
-
12481
)
8
Ishikawa
T.
Beuron
F.
Kessel
M.
Wickner
S.
Maurizi
M. R.
Steven
A. C.
Translocation pathway of protein substrates in ClpAP protease
Proc. Natl. Acad. Sci. U.S.A.
2001
, vol. 
98
 (pg. 
4328
-
4333
)
9
Kim
Y. I.
Levchenko
I.
Fraczkowska
K.
Woodruff
R. V.
Sauer
R. T.
Baker
T. A.
Molecular determinants of complex formation between Clp/Hsp100 ATPases and the ClpP peptidase
Nat. Struct. Biol.
2001
, vol. 
8
 (pg. 
230
-
233
)
10
Ortega
J.
Lee
H. S.
Maurizi
M. R.
Steven
A. C.
Alternating translocation of protein substrates from both ends of ClpXP protease
EMBO J.
2002
, vol. 
21
 (pg. 
4938
-
4949
)
11
Gottesman
S.
Wickner
S.
Maurizi
M. R.
Protein quality control: triage by chaperones and proteases
Genes Dev.
1997
, vol. 
11
 (pg. 
815
-
823
)
12
Adam
Z.
Adamska
I.
Nakabayashi
K.
Ostersetzer
O.
Haussuhl
K.
Manuell
A.
Zheng
B.
Vallon
O.
Rodermel
S. R.
Shinozaki
K.
Clarke
A. K.
Chloroplast and mitochondrial proteases in Arabidopsis: a proposed nomenclature
Plant Physiol.
2001
, vol. 
125
 (pg. 
1912
-
1918
)
13
Nakabayashi
K.
Ito
M.
Kiosue
T.
Shinozaki
K.
Watanabe
A.
Identification of clp genes expressed in senescing Arabidopsis leaves
Plant Cell Physiol.
1999
, vol. 
40
 (pg. 
504
-
514
)
14
Peltier
J. B.
Ytterberg
J.
Liberles
D. A.
Roepstorff
P.
van Wijk
K. J.
Identification of a 350 kDa ClpP and protease complex with 10 different Clp isoforms in chloroplasts of Arabidopsis thaliana
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
16318
-
16327
)
15
Zheng
B.
Halperin
T.
Hruskova-Heidingsfeldova
O.
Adam
Z.
Clarke
A. K.
Characterization of chloroplast Clp proteins in Arabidopsis: localization, tissue specificity and stress responses
Physiol. Plant.
2002
, vol. 
114
 (pg. 
92
-
101
)
16
Peltier
J. B.
Ripoll
D. R.
Friso
G.
Rudella
A.
Cai
Y.
Ytterberg
J.
Giacomelli
L.
Pillardy
P.
van Wijk
K. J.
Clp protease complexes from photosynthetic and non-photosynthetic plastids and mitochondria of plants, their predicted three-dimensional structures, and functional implications
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
4768
-
4781
)
17
Sjögren
L. L. E.
Stanne
T. M.
Zheng
B.
Sutinen
S.
Clarke
A. K.
Structural and functional insights into the chloroplast ATP-dependent Clp protease in Arabidopsis
Plant Cell
2006
, vol. 
18
 (pg. 
2635
-
2649
)
18
Shikanai
T.
Shimizu
K.
Ueda
K.
Nishimura
Y.
Kuroiwa
T.
Hashimoto
T.
The chloroplast clpP gene, encoding a proteolytic subunit of ATP-dependent protease, is indispensable for chloroplast development in tobacco
Plant Cell Physiol.
2001
, vol. 
42
 (pg. 
264
-
273
)
19
Kuroda
H.
Maliga
P.
The plastid clpP1 protease gene is essential for plant development
Nature
2003
, vol. 
425
 (pg. 
86
-
89
)
20
Rudella
A.
Friso
G.
Alonso
J. M.
Ecker
J. R.
van Wijk
K. J.
Downregulation of ClpR2 leads to reduced accumulation of the ClpPRS protease complex and defects in chloroplast biogenesis in Arabidopsis
Plant Cell
2006
, vol. 
18
 (pg. 
1704
-
1721
)
21
Zheng
B.
MacDonald
T. M.
Sutinen
S.
Hurry
V.
Clarke
A. K.
A nuclear-encoded ClpP subunit of the chloroplast ATP-dependent Clp protease is essential for early development in Arabidopsis thaliana
Planta
2006
, vol. 
224
 (pg. 
1103
-
1115
)
22
Koussevitzky
S.
Stanne
T. M.
Peto
C. A.
Giap
T.
Sjögren
L. L. E.
Zhao
Y.
Clarke
A. K.
Chory
J.
An Arabidopsis thaliana virescent mutant reveals a role for ClpR1 in plastid development
Plant Mol. Biol.
2007
, vol. 
63
 (pg. 
85
-
96
)
23
Sjögren
L. L. E.
MacDonald
T. M.
Sutinen
S.
Clarke
A. K.
Inactivation of the clpC1 gene encoding a chloroplast Hsp100 molecular chaperone causes growth retardation, leaf chlorosis, lower photosynthetic activity, and a specific reduction in photosystem content
Plant Physiol.
2004
, vol. 
136
 (pg. 
4114
-
4126
)
24
Aronsson
H.
Jarvis
P.
A simple method for isolating import competent Arabidopsis chloroplasts
FEBS Lett.
2002
, vol. 
529
 (pg. 
215
-
220
)
25
Sokolenko
A.
Lerbs-Mache
S.
Altschmied
L.
Herrmann
R. G.
Clp protease complexes and their diversity in chloroplasts
Planta
1998
, vol. 
207
 (pg. 
286
-
295
)
26
Clarke
A. K.
Critchley
C.
The identification of a heat shock protein complex in chloroplasts of barley leaves
Plant Physiol.
1992
, vol. 
100
 (pg. 
2081
-
2089
)
27
Härndahl
U.
Hall
R. B.
Osteryoung
K. W.
Vierling
E.
Bornman
J. F.
Sundby
C.
The chloroplast small heat shock protein undergoes oxidation-dependent conformational changes and may protect plants from oxidative stress
Cell Stress Chaperones
1999
, vol. 
4
 (pg. 
129
-
138
)
28
Bell
E.
Creelman
R. A.
Mullet
J. E.
A chloroplast lipoxygenase is required for wound-induced jasmonic acid accumulation in Arabidopsis
Proc. Natl. Acad. Sci. U.S.A.
1995
, vol. 
92
 (pg. 
8675
-
8679
)
29
Osteryoung
K. W.
Vierling
E.
Dynamics of small heat shock protein distribution within the chloroplasts of higher plants
J. Biol. Chem.
1994
, vol. 
269
 (pg. 
28676
-
28682
)
30
Majeran
W.
Wollman
F. A.
Vallon
O.
Evidence for a role of ClpP in the degradation of the chloroplast cytochrome b6f complex
Plant Cell
2000
, vol. 
12
 (pg. 
137
-
149
)
31
Sangster
T. A.
Queitsch
C.
The HSP90 chaperone complex, an emerging force in plant development and phenotypic plasticity
Curr. Opin. Plant Biol.
2005
, vol. 
8
 (pg. 
86
-
92
)
32
Levy-Rimler
G.
Bell
R. E.
Ben-Tal
N.
Azem
A.
Type I chaperonins: not all are created equal
FEBS Lett.
2002
, vol. 
529
 (pg. 
1
-
5
)
33
Nilsson
J.
Nissen
P.
Elongation factors on the ribosome
Curr. Opin. Struct. Biol.
2005
, vol. 
15
 (pg. 
349
-
354
)
34
Chetverin
A. B.
Spirin
A. S.
Bioenergetics and protein synthesis
Biochem. Biophys. Acta
1982
, vol. 
683
 (pg. 
153
-
179
)
35
Ohta
M.
Sugita
M.
Sugiura
M.
Three types of nuclear genes encoding chloroplast RNA-binding proteins (cp29, cp31 and cp33) are present in Arabidopsis thaliana: presence of cp31 in chloroplasts and its homologue in nuclei/cytoplasms
Plant Mol. Biol.
1995
, vol. 
27
 (pg. 
529
-
539
)
36
Rocak
S.
Linder
L.
DEAD-box proteins: the driving forces behind RNA metabolism
Nat. Rev. Mol. Cell Biol.
2004
, vol. 
5
 (pg. 
232
-
241
)
37
Zrenner
R.
Stitt
M.
Sonnewald
U.
Boldt
R.
Pyrimidine and purine biosynthesis and degradation in plants
Annu. Rev. Plant Biol.
2006
, vol. 
57
 (pg. 
805
-
836
)
38
Hasunuma
K.
Yabe
N.
Yoshida
Y.
Ogura
Y.
Hamada
T.
Putative functions of nucleoside diphosphate kinase in plants and fungi
J. Bioenerg. Biomembr.
2003
, vol. 
35
 (pg. 
57
-
65
)
39
van Heeswijk
W. C.
Rabenberg
M.
Westerhoff
H. V.
Kahn
D.
The genes of the glutamine synthetase adenylylation cascade are not regulated by nitrogen in Escherichia coli
Mol. Microbiol.
1993
, vol. 
9
 (pg. 
443
-
457
)
40
van Heeswijk
W. C.
Wen
D.
Clancy
P.
Jaggi
R.
Ollis
D. L.
Westerhoff
H. V.
Vasudevan
S. G.
The Escherichia coli signal transducers PII (GlnB) and GlnK form heterotrimers in vivo: fine tuning the nitrogen signal cascade
Proc. Natl. Acad. Sci. U.S.A.
2000
, vol. 
97
 (pg. 
3942
-
3947
)
41
Kruse
E.
Mock
H. P.
Grimm
B.
Coproporphyrinogen III oxidase from barley and tobacco–sequence analysis and initial expression studies
Planta
1995
, vol. 
196
 (pg. 
796
-
803
)
42
Strand
Å.
Asami
T.
Alonso
J.
Ecker
J. R.
Chory
J.
Chloroplast to nucleus communication triggered by accumulation of Mg-protoporphyrinIX
Nature
2003
, vol. 
421
 (pg. 
79
-
83
)
43
Ankele
E.
Kindgren
P.
Pesquet
E.
Strand
A.
In vivo visualization of Mg-protoporphyrinIX, a coordinator of photosynthetic gene expression in the nucleus and the chloroplast
Plant Cell
2007
, vol. 
19
 (pg. 
1964
-
1979
)
44
Wood
Z. A.
Poole
L. B.
Karplus
P. A.
Peroxiredoxin evolution and the regulation of hydrogen peroxide signaling
Science
2003
, vol. 
300
 (pg. 
650
-
653
)
45
Ting
J. T. L.
Wu
S. S. H.
Ratnayake
C.
Huang
A. H. C.
Constituents of the tapetosomes and elaioplasts in Brassica campestris tapetum and their degradation and retention during microsporogenesis
Plant J.
1998
, vol. 
16
 (pg. 
541
-
551
)
46
Rey
P.
Gillet
B.
Romer
S.
Eymery
F.
Massimino
J.
Peltier
G.
Kuntz
M.
Over-expression of a pepper plastid lipid-associated protein in tobacco leads to changes in plastid ultrastructure and plant development upon stress
Plant J.
2000
, vol. 
21
 (pg. 
483
-
494
)
47
Dörmann
P.
Benning
C.
The role of UDPglucose epimerase in carbohydrate metabolism of Arabidopsis
Plant J.
1998
, vol. 
13
 (pg. 
641
-
652
)
48
Wolfe
K. H.
Morden
C. W.
Palmer
J. D.
Function and evolution of a minimal plastid genome from a nonphotosynthetic parasitic plant
Proc. Natl. Acad. Sci. U.S.A.
1992
, vol. 
89
 (pg. 
10648
-
10652
)
49
Roos
D. S.
Crawford
M. J.
Donald
R. G. K.
Fraunholz
M.
Harb
O. S.
He
C. Y.
Kissinger
J. C.
Shaw
M. K.
Striepen
B.
Mining the Plasmodium genome database to define organellar function: What does the apicoplast do?
Philos. Trans. R. Soc. Lond. B Biol. Sci.
2002
, vol. 
357
 (pg. 
35
-
46
)

Author notes

1

These authors contributed equally to this work.

Supplementary data