CLPB3 is required for the removal of chloroplast protein aggregates and thermotolerance in Chlamydomonas

Abstract In the cytosol of plant cells, heat-induced protein aggregates are resolved by the CASEIN LYTIC PROTEINASE/HEAT SHOCK PROTEIN 100 (CLP/HSP100) chaperone family member HSP101, which is essential for thermotolerance. For the chloroplast family member CLPB3 this is less clear, with controversial reports on its role in conferring thermotolerance. To shed light on this issue, we have characterized two clpb3 mutants in Chlamydomonas reinhardtii. We show that chloroplast CLPB3 is required for resolving heat-induced protein aggregates containing stromal TRIGGER FACTOR (TIG1) and the small heat shock proteins 22E/F (HSP22E/F) in vivo, and for conferring thermotolerance under heat stress. Although CLPB3 accumulation is similar to that of stromal HSP70B under ambient conditions, we observed no prominent constitutive phenotypes. However, we found decreased accumulation of the PLASTID RIBOSOMAL PROTEIN L1 (PRPL1) and increased accumulation of the stromal protease DEG1C in the clpb3 mutants, suggesting that a reduction in chloroplast protein synthesis capacity and an increase in proteolytic capacity may compensate for loss of CLPB3 function. Under ambient conditions, CLPB3 was distributed throughout the chloroplast, but reorganized into stromal foci upon heat stress, which mostly disappeared during recovery. CLPB3 foci were localized next to HSP22E/F, which accumulated largely near the thylakoid membranes. This suggests a possible role for CLPB3 in disentangling protein aggregates from the thylakoid membrane system.


Introduction
The casein lytic proteinase/heat shock protein 100 (Clp/ Hsp100) chaperones belong to the large family of AAA+ proteins (ATPases associated with various cellular activities; Neuwald et al., 1999). Clp/Hsp100 proteins are divided into two classes, with class II members containing one (ClpX, ClpY), and class I members two AAA+ modules in tandem (ClpA to E; Schirmer et al., 1996). The AAA+ module features the Walker A and B motifs. Clp/Hsp100 assemble into homohexameric rings with a central pore through which protein substrates are threaded (Guo et al., 2002;Lee et al., 2003;Mogk et al., 2015Mogk et al., , 2018. ClpA, ClpC, ClpE, and ClpX contain a conserved tripeptide [LIV]-G- [FL] crucial for binding to an associated protease like ClpP, that is lacking in ClpB/Hsp101 and ClpY family members (Kim et al., 2001). ClpB/Hsp101 family members contain two additional domains: the N-terminal domain and the middle domain, which forms a coiled-coil structure that is inserted into the first AAA+ module (Mogk et al., 2015). The threading activity of E. coli ClpB can initiate at the N-or C-termini or at internal sites of substrate proteins in protein aggregates, such that entire peptide loops are translocated through the pore (Avellaneda et al., 2020). Translocation is mediated by mobile loops in the central pore that contact the substrate via conserved aromatic residues (Deville et al., 2017;Rizo et al., 2019). Driven by ATP hydrolysis, the loops move downwards along the translocation channel. Axial staggering of the loops facilitates substrate handover and prevents substrate backsliding. Once displaced to the opposite side of the ClpB hexameric ring, the substrate can fold to the native state by itself or aided by Hsp70 and/or chaperonins (Mogk et al., 2018).
The ability of cytosolic ClpB/Hsp100 members to disentangle individual proteins from aggregates has been shown to be crucial for the resolving of protein aggregates formed upon heat stress in yeast and in Arabidopsis (Parsell et al., 1994;Agarwal et al., 2003;McLoughlin et al., 2016). Eliminating cytosolic HSP101 activity in land plants had no effect on plant growth under ambient temperatures, but resulted in reduced basal and acquired thermotolerance and, vice versa, increasing HSP101 levels resulted in enhanced basal and acquired thermotolerance Vierling, 2000, 2001;Queitsch et al., 2000;Nieto-Sotelo et al., 2002;Katiyar-Agarwal et al., 2003). Deletion of the single-copy gene encoding ClpB in Synechococcus sp. had no phenotype under optimal growth conditions, and did not affect basal thermotolerance, but strongly impaired the capacity of the mutant to develop thermotolerance Clarke, 1996, 2000). Hence, to survive heat stress it is crucial that cells resolve heat-induced cytosolic protein aggregates during recovery from heat. The engineering of yeast cytosolic Hsp104 to deliver substrates directly to an associated peptidase abolished thermotolerance, suggesting that for thermotolerance, a reactivation of aggregated cytosolic proteins is required, not just their removal (Weibezahn et al., 2004).
In plants, chloroplast ClpB (termed CLPB3, ClpB-p, or APG6) and mitochondrial ClpB (termed CLPB4 or ClpB-m) are both derived from cyanobacterial ClpB (Lee et al., 2007;Mishra and Grover, 2016). Suppressing the production of chloroplast CLPB3 in tomato did not result in visible phenotypes under optimal growth conditions, but strongly impaired acquired thermotolerance. This suggests that in the chloroplast too, protein aggregates formed under heat stress must be resolved during the recovery phase to promote survival (Yang et al., 2006). Interestingly, Arabidopsis clpb3 knock-out mutants under ambient conditions were pale green, with smaller and rounder chloroplasts lacking starch grains, and showing an abnormal development of thylakoid membranes, when compared with the wild type (Myouga et al., 2006;Lee et al., 2007;Zybailov et al., 2009). Accordingly, clpb3 mutant plants exhibited lower PSII activity and were seedling-lethal if not sup-plied with sucrose, pointing to a role of CLPB3 as a general housekeeping chaperone in chloroplasts, at least in Arabidopsis. Surprisingly, unlike in tomato, thermotolerance was not impaired in Arabidopsis clpb3 mutants, and was not enhanced when CLPB3 was overexpressed (Myouga et al., 2006;Lee et al., 2007). Nevertheless, recombinant Arabidopsis CLPB3 was found to interact with heat-denatured, aggregated GLU-COSE-6-PHOSPHATE DEHYDROGENASE (G6PDH), and to support the disentangling and refolding of a large part of the protein to the native state in an ATP-dependent reaction in vitro (Parcerisa et al., 2020). Moreover, Arabidopsis CLPB3 promoted the refolding of aggregation-prone 1-DEOXY-D-XYLULOSE 5-PHOSPHATE SYNTHASE (DXS; the rate-determining enzyme for the production of plastidial isoprenoids) under ambient conditions in vivo (Pulido et al., 2016;Llamas et al., 2017).
Chlamydomonas reinhardtii has five CLPB genes (Schroda and Vallon, 2009). CLPB1, CLPB3, and CLPB4 proteins are produced at low levels under ambient temperature (CLPB1 lower than CLPB3/4) and accumulate rapidly and with similar kinetics during heat stress at 42 °C, with a plateau reached after 2 h at 42 °C (Mühlhaus et al., 2011). CLPB2 and CLPB5 have not been detected in proteomics studies, lack expressed sequence tag (EST) support, and it is therefore not clear whether they are produced only under certain conditions, or not at all (Schroda and Vallon, 2009). CLPB1 is predicted to be localized to the cytosol, and CLPB3 to the chloroplast, while the localization of CLPB4 is not clear. The abundance of CLPB3, together with HSP22E, HSP22F, HSP22C, VESICLE INDUCING PROTEIN IN PLASTIDS 1 and 2 (VIPP1 and 2), and DEG1C, increased when chloroplasts experienced stresses likely to disturb chloroplast protein homeostasis. These stresses include high light intensities or elevated cellular H 2 O 2 concentrations (Nordhues et al., 2012;Blaby et al., 2015;Perlaza et al., 2019;Theis et al., 2019Theis et al., , 2020, depletion of chloroplast-encoded ClpP (Ramundo et al., 2014), depletion of thylakoid membrane transporters/integrases (Theis et al., 2019(Theis et al., , 2020, addition of nickel ions (Blaby-Haas et al., 2016) or the alkylating agent methyl methanesulfonate (Fauser et al., 2022), and the inhibition of chloroplast fatty acid synthesis (Heredia-Martínez et al., 2018). These seven chloroplast proteins appear to represent a core set of proteins involved in coping with disturbed chloroplast protein homeostasis (Ramundo et al., 2014;Perlaza et al., 2019). Their up-regulation appears to be triggered by mis-folded/mis-assembled proteins inducing lipid packing stress in chloroplast membranes that is sensed and dealt with by the VIPP1/2 proteins (Theis et al., 2020;Kleine et al., 2021). HSP22E/F were found to interact with thermolabile stromal proteins and chaperones in heat stressed cells, and with VIPP1/2 and stromal HSP70B, especially at chloroplast membranes, in cells exposed to H 2 O 2 (Rütgers et al., 2017a;Theis et al., 2020). DEG1C localizes to the stroma and the periphery of thylakoid membranes. Purified DEG1C exhibited high proteolytic activity against unfolded model substrates, which increased with temperature and pH (Theis et al., 2019). So far, there are no functional studies on Chlamydomonas CLPB3. Therefore, the aim of this work was to shed light on CLPB3 function and in particular, its possible role in maintaining chloroplast protein homeostasis. We show that CLPB3 is crucial for removing protein aggregates in the chloroplast, which contributes to enhanced thermotolerance under conditions of severe heat stress.

Strains and cultivation conditions
Chlamydomonas reinhardtii wild-type strain CC-4533 (cw15, mt-) and mutant strains clpb3-1 (LMJ.RY0402.250132_1) and clpb3-2 (LMJ. RY0402.104257_1) from the Chlamydomonas Library Project (Li et al., 2016) were obtained via the Chlamydomonas Resource Center (https:// www.chlamycollection.org/). Cultures were grown mixotrophically in TRIS-ACETATE-PHOSPHATE (TAP) medium (Kropat et al., 2011) on a rotatory shaker at 25 °C and ~40 μmol photons m -2 s -1 . For complementation, clpb3 mutant cells were transformed via the glass bead method (Kindle, 1990) as described previously (Hammel et al., 2020), with the constructs linearized via EcoRV digestion. Transformants were selected on TAP medium containing 100 µg ml -1 spectinomycin. Cell densities were determined using the Z2 Coulter Counter (Beckman Coulter, Germany) or photometrically by optical density measurements at 750 nm (OD 750 ). For heat stress experiments, cultures of exponentially growing cells were placed into a water bath heated to 41 °C if not stated otherwise, and incubated under agitation and constant illumination at ~40 μmol photons m -2 s -1 for 1 h, with subsequent recovery at 25 °C for 6 h. To determine survival rates via colony forming units, cultures of wild type and clpb3-1 were grown at 25 °C to a density of 1 × 10 6 cells ml -1 and split into equal parts. For each strain, one culture was left at 25 °C for 2 h while the other was incubated at 41 °C for 2 h in a water bath. After the treatment, cells were diluted to 1 cell µl -1 and 200 µl were plated out in triplicates on TAP agar plates in four independent experiments. For spot tests, cells were grown to a density of 3-5 × 10 6 cells ml -1 and diluted in TAP medium or high salt medium (HSM) such that 10 µl contained 10 4 , 10 3 or 10 2 cells. Following this, 10 µl of each dilution were spotted onto agar plates containing TAP medium or HSM, and incubated in low light (30 µmol photons m -2 s -1 ) for 7 d, in high light (600 µmol photons m -2 s -1 ) for 4 d, in the dark for 7 d, or exposed to three ~24 h heat shock treatments at 40 °C with ≤24 h recovery in between. HSM was prepared according to Sueoka (1960) but using the trace solutions from Kropat et al. (2011).

Extraction of Chlamydomonas genomic DNA and verification of the insertion sites
For the extraction of Chlamydomonas genomic DNA, 5 ml of exponentially growing cells were pelleted and resuspended in 250 μl water. Following this, 250 μl of 2× lysis buffer [20 mM Tris-HCl, 40 mM Na 2 EDTA, 1% (w/v) SDS] and 3 μl proteinase K (NEB, USA: P8102S, 20 mg ml -1 ) were added and incubated under agitation at 55 °C for 2 h. The lysate was supplemented with 80.9 μl of 5 M NaCl and mixed by vortexing. After the addition of 70 μl of pre-warmed CTAB/NaCl [2% (w/v) CTAB; 1.4 M NaCl], lysates were vortexed and incubated under agitation for 10 min at 65 °C. Nucleic acids were extracted by addition of 1 volume phenol: chloroform: isoamylalcohol (25:24:1; Roth, Germany), mixing the two phases and separating for 5 min at 18 000×g and 4 °C. Phenol/chloroform extraction of the aqueous phase was repeated once. An equal volume of chloroform: isoamylal-cohol (24:1; Roth) was added to the upper phase, and the mixture was centrifuged as above. Recovery of nucleic acids was achieved by precipitating with an equal volume of isopropanol. Finally, the pellet was resuspended in TE-buffer (10 mM Tris-HCl, pH 8.0; 1 mM EDTA), from which 1 ng of DNA was used for PCR. Validation of the aphVIII cassette insertion site within the genes of the Chlamydomonas Library Project (CLiP) mutant lines was performed using the specific primers listed in Supplementary Table S1 according to the manual provided by the CLiP (Li et al., 2016). Amplified products were analysed by agarose gel electrophoresis. Electrophoresed DNA was stained with Gelred (Biotium, USA) or HDGreen Plus DNA Stain (INTAS Science Imaging, Germany) and visualized under UV light using a gel documentation system [FUSION-FX7 Advance™ imaging system (PeqLab, Germany)/ECL ChemoStar V90D+ (INTAS Science Imaging, Germany)].

Cloning, production, and purification of recombinant CLPB3 to obtain anti-CLPB3 serum
The CLPB3 coding region lacking the chloroplast transit peptide was amplified by PCR from EST clone AV631848 (Asamizu et al., 2000) with primers CLPB3-Eco and CLPB3-Hind. The 2827 bp PCR product was digested with EcoRI and HindIII, and cloned into EcoRI-HindIIIdigested pETDuet-1 vector (Novagen, Germany) lacking two nucleotides upstream from the BamHI site, producing pMS976. CLPB3 was expressed with an N-terminal hexahistidine tag in E. coli Rosetta cells after induction with 1 mM IPTG for 16 h at 20 °C, and purified by cobaltnitrilotriacetic acid affinity chromatography, according to the manufacturer's instructions (G-Biosciences, USA), including a washing step with 5 mM Mg-ATP. Eluted CLPB3 was gel filtrated using an Enrich SEC650 column. Fractions containing CLPB3 were pooled and concentrated in Amicon® Ultra-4 Centrifugal Filter Units (Ultracel®-3K, Merck Millipore Ltd, Germany), with a subsequent buffer exchange using 6 M Urea, 50 mM NaCl, 20 mM Tris-HCl, pH 7.5. Protein concentrations were determined via a NanoDrop 2000 (ThermoFisher Scientific, Germany). Proteins were frozen in liquid nitrogen and stored at −80 °C. From this, 2.6 mg of the protein were used for the immunization of a rabbit via the 3 month standard immunization protocol of Bioscience bj-diagnostik (Göttingen, Germany).

Isolation of protein aggregates
Protein aggregates were isolated as described previously (Koplin et al., 2010) with minor modifications. Briefly, Chlamydomonas cells were grown to a density of approximately 5 × 10 6 cells ml -1 and a total of 2 × 10 8 cells were used. Cells were supplemented with sodium azide at a final concentration of 0.002% and harvested by centrifugation at 3500 × g for 2 min at 4 °C, and cell pellets were frozen in liquid nitrogen and stored at −80 °C. Cell pellets were thawed on ice and resuspended in lysis buffer [20 mM sodium phosphate pH 6.8, 10 mM DTT, 1 mM EDTA, and 0.25× protease inhibitor cocktail (Roche, Germany)]. Cells were sonicated on ice once for 10 s (cycle 6×, minimal output, MS73 tip; Sonopuls, Bandelin, Germany) and centrifuged for 10 min at 500 × g and 4 °C to remove intact cells and cell debris. Protein concentrations in the supernatant were measured by the Bradford assay (Bradford, 1976), and samples were diluted to match the sample with the lowest protein concentration. Samples for total input were taken and supplemented with 2× Laemmli sample buffer (125 mM Tris-HCl pH 6.8, 20% glycerol, 4% SDS, 0.1 M DTT, and 0.005% bromophenol blue). Samples were then centrifuged for 30 min at 19 000×g and 4 °C. Pellets were washed four times by the addition of washing buffer containing 20 mM sodium phosphate pH 6.8 and 2% Nonidet-P40, sonication for 6 s with the same settings as above, and centrifugation for 30 min at 19 000×g and 4 °C. At last, pellets were dissolved in 1× Laemmli sample buffer containing 3 M urea. Samples were separated on 12% SDS-polyacrylamide gels followed by Coomassie staining or immunoblotting.

Blue native PAGE analysis
Blue native (BN) PAGE with whole cell proteins was carried out according to published protocols (Schagger and von Jagow, 1991;Schagger et al., 1994) with minor modifications. Briefly, cells were exposed for 1 h to 41 °C heat shock with a subsequent 6 h recovery at 25 °C, as described above. Approximately 10 8 cells were harvested by centrifugation, washed with TMK buffer (10 mM Tris-HCl, pH 6.8, 10 mM MgCl 2 , 20 mM KCl), and resuspended in 500 μl ACA buffer (750 mM ε-aminocaproic acid, 50 mM Bis-Tris pH 7.0 and 0.5 mM EDTA supplemented with 0.25× protease inhibitor). Cells were disrupted by sonication. Intact cells and cell debris were removed by centrifugation for 5 min at 300×g and 4 °C. Whole cell lysates (equivalent to 0.25 μg μl -1 of chlorophyll) were solubilized for 20 min with 1% (w/v) β-dodecyl maltoside (Roth, Germany) on ice in the dark and insolubilized material was precipitated by centrifugation at 18 500×g for 10 min at 4 °C. Following this, supernatants were supplemented with native sample buffer [750 mM ε-aminocaproic acid and 5% (w/v) Coomassie Brilliant Blue G250] and separated on a 4-15% (w/v) blue-native (BN) polyacrylamide gel, followed by immunoblotting.

Microscopy
For immunofluorescence microscopy, cells were fixed and stained as described previously (Uniacke et al., 2011) with minor modifications: microscopy slides were washed three times with 100% ethanol and coated with 0.1% poly-L-lysine. Cells were fixed with 4% formaldehyde for at least 1 h at 4 °C on an overhead rotator. Aliquots of 40 μl cell suspension were allowed to adhere to the microscope slides for 15 min at 25 °C, followed by incubation in 100% methanol for 6 min at -20 °C. Subsequently, slides were washed five times with phosphate-buffered saline (PBS). Cells were permeabilized by incubating the slides with 2% Triton X-100 in PBS for 10 min at 25 °C. Slides were washed three times with PBS containing 5 mM MgCl 2 and with PBS-BSA (PBS, 1% BSA) for at least 30 min at 25 °C. Slides were incubated over night at 4 °C with antisera against HSP22EF and the HA-tag using 1:1000 dilutions, in PBS-BSA. Slides were then washed five times with PBS-BSA at 25 °C followed by incubation in a 1:200 dilution of the tetramethylrhodamine-isothiocyanate-labelled goat anti-rabbit antibody (TRITC, Sigma-Aldrich) and fluorescein isothiocyanate-labelled goat anti-mouse antibody (FITC, Sigma-Aldrich) in PBS-BSA for 1.5 h at 25 °C in the dark. Finally, the slides were washed five times with PBS and mounting solution containing 4',6-diamidino-2-phenylindole (DAPI; Vectashield; Vector Laboratories, USA) was dispersed over the cells. HSP22E/F and the HA-tag images were captured with a Zeiss LSM880 Axio Observer confocal laser scanning microscope equipped with a Zeiss C-Apochromat 40×/1.2 W AutoCorr M27 water-immersion objective (Zeiss, Germany). Fluorescent signals of FITC (excitation/emission 488 nm/491-589 nm) and TRITC (excitation/emission 633 nm/647-721 nm) were processed using the Zeiss software ZEN 2.3 or ImageJ. Light microscopy images were taken with an Olympus BX53 microscope.

Results
The Chlamydomonas CLPB3 gene encodes a pre-protein with 1043 amino acids, of which the N-terminal 115 aa are predicted to serve as a chloroplast transit peptide ( Supplementary  Fig. S1A). The mature CLPB3 protein has a mass of 101 kDa and shares 54% identical and 72% similar residues with E. coli ClpB, and 68% identical and 82% similar residues with mature Arabidopsis CLPB3. We produced mature recombinant Chlamydomonas CLPB3 in E. coli with an N-terminal hexahistidine tag (Supplementary Figs S1A, S2) and raised a polyclonal antibody. The antibody revealed that CLPB3 is produced constitutively in Chlamydomonas as a protein with an apparent molecular mass of ~102 kDa that migrated little below the full-length recombinant protein, indicating processing of the transit peptide at the predicted site ( Fig. 1; Supplementary Fig.  S1A). We observed some degradation of the recombinant protein. Quantification of the immunoblot signals in three independent replicates (including degradation products) revealed that CLPB3 accounts for approximately 0.2 ± 0.024% (SD, n=3) of the total cell proteins.

Two clpb3 mutants accumulate a truncated form of CLPB3 and less CLPB3
To obtain insights into the function of CLPB3 in Chlamydomonas, we ordered two clpb3 CLiP mutants (Li et al., 2016) with insertions of the mutagenesis cassette in exon 12 (clpb3-1) and in intron 4 (clpb3-2; Fig. 2A). We could amplify both flanking regions of the cassette in the clpb3-2 mutant and the flanking region 5' of the cassette in the clpb3-1 mutant ( Supplementary  Fig. S3). However, we could not amplify the flanking region 3' of the cassette in the clpb3-1 mutant, even with staggered flanking primers, but we could show that the cassette is intact. Most likely, additional DNA sequences were inserted between the 3' end of the cassette and the insertion site in the CLPB3 gene, which is not uncommon for Chlamydomonas insertional mutants (Zhang et al., 2014;Spaniol et al., 2022).
We first analysed CLPB3 production in the mutants and in wild type under ambient conditions, after a 60 min exposure to 41 °C, and after 6 h of recovery from heat stress. As shown in Fig. 2B, C, CLPB3 production in the wild type increased 4-fold during the heat treatment, compared with ambient conditions in the same strain and declined by ~14% during recovery, corroborating findings from a large-scale proteomics study (Mühlhaus et al., 2011). We found two putative heat shock elements (HSEs) about 60 nt and 90 nt upstream of a putative TATA box in the CLPB3 gene that show a degree of degeneration typical for HSEs in Chlamydomonas ( Supplementary Fig. S4) (Lodha et al., 2008). These HSEs are most likely driving heat-induced expression of the CLPB3 gene via the heat shock transcription factor HSF1 (Schulz-Raffelt et al., 2007).
In both mutants, CLPB3 accumulated to only ~20% of wild-type levels under ambient conditions. While in the clpb3-1 mutant, CLPB3 after heat treatment and recovery accumulated like in wild type, CLPB3 production barely increased in the clpb3-2 mutant (Fig. 2B, C). Apparently, intron splicing in this mutant is impaired and results in overall lower protein production. CLPB3 in the clpb3-1 mutant had an apparent molecular mass of ~96 kDa compared with ~102 kDa in the wild type and the clpb3-2 mutant, in line with the predicted truncation of its C-terminus ( Supplementary Fig. S1A). Unlike in Arabidopsis clpb3 mutants, we observed no obvious phenotypes in chloroplast development or photosystem II activity in the two Chlamydomonas clpb3 mutants (Supplementary Fig. S5).

CLPB3 abundance in the mutants can be partially restored in complemented lines
To complement the mutants, we synthesized the genomic sequence encoding the entire CLPB3 protein as a level 0 part for the Modular Cloning system (Crozet et al., 2018). All introns were kept, except introns 7, 8, and 11, because they contain highly repetitive sequences (CA in introns 7 and 8, and GT in intron 11). In addition, intron 11 is particularly long with 1269 nt. The CLPB3 genomic sequence was then assembled into a level 1 module with the HSP70A-RBCS2 promoter, RPL23 terminator, and a sequence encoding a C-terminal 3×HA tag ( Fig. 2A). We used the constitutive HSP70A-RBCS2 promoter because it was part of the MoClo toolkit and strongly enhances chances for transgene expression in Chlamydomonas (Strenkert et al., 2013;Crozet et al., 2018). Moreover, in our hands, promoters of genes encoding chloroplast-targeted HSPs were inefficient in driving expression of their gene to native expression levels in a transgene setting, presumably because they require a specific chromatin environment for full activity (Schroda et al., 2000;Strenkert et al., 2011;Rütgers et al., 2017b). Since cytosolic HSP101 fused C-terminally to GFP or to a Strep tag was fully functional (McLoughlin et al., 2016(McLoughlin et al., , 2019, we did not expect the C-terminal 3×HA sequence to interfere with CLPB3 function. After adding a spectinomycin resistance cassette in a level 2 device, the latter was transformed into both mutants, and spectinomycin-resistant transformants were screened using antibodies against CLPB3 and the HA epitope ( Supplementary Fig. S6). Despite using the HSP70A-RBCS2 promoter, less than 10% of the transformants expressed HA-tagged CLPB3 that was clearly detectable. Under ambient conditions, the best-expressing transformants accumulated CLPB3 like wild type (clpb3-1c) or to 85% of wild type (clpb3-2c; Fig. 2B, C). After heat shock, CLPB3 production in clpb3-1c exceeded that in the wild type by ~1.5-fold, while CLPB3 production in clpb3-2c amounted to only ~30% of wild type. Fig. 1. Quantification of CLPB3. The indicated amounts of whole-cell (WC) protein from Chlamydomonas wild type grown at 25 °C and of recombinant CLPB3 produced in E. coli were separated on a 8% SDSpolyacrylamide gel and analysed by immunoblotting using an antibody raised against Chlamydomonas CLPB3. The arrowhead points to intact CLPB3, and the asterisk to a degradation product.

Loss of function of CLPB3 results in strongly elevated DEG1C production
We next analysed the accumulation of selected proteins involved in chloroplast protein homeostasis (protein biosynthesis, folding, and degradation) to understand whether or not their accumulation was affected by the reduced CLPB3 production in the mutants. Chloroplast chaperones CPN60A and HSP22E/F (Bai et al., 2015;Rütgers et al., 2017a) strongly accumulated after heat stress and declined after recovery, and thus behaved similar to CLPB3, with little differences between mutants and wild type (Fig. 2B, C). Production of trigger factor TIG1, a thermolabile chaperone involved in protein biogenesis (Ries et al., 2017;Rütgers et al., 2017a;Rohr et al., 2019), declined by ~30% after heat stress in the wild type (Fig 2B,  C). There was a trend of a more pronounced decrease in both mutants that appeared to be relieved in the complemented lines. Similarly, production of chloroplast ribosome subunit PRPL1 (Ries et al., 2017) appeared to be overall lower in the mutants, compared with the wild type, with some restoration of PRPL1 production especially in the complemented line clpb3-2c (Fig. 2B, C). The most striking difference between clpb3 mutants and wild type was observed for stromal protease DEG1C (Theis et al., 2019). DEG1C accumulation was much higher in the mutants compared with the wild type under all conditions (more than 3-fold under ambient conditions), with a trend towards restoration of lower DEG1C production, especially in the clpb3-1c line (Fig. 2B, C). To substantiate these findings, we focused on investigating the accumulation of DEG1C in wild type, clpb3-2 mutant and clpb3-2c only under ambient conditions (Fig. 3). Here the clpb3-2 mutant showed 2.3-fold higher DEG1C production than the wild type, which was reduced to 1.5-fold higher production in the complemented line clpb3-2c. The lack of full complementation can be explained by the accumulation of CLPB3 to only ~80% of that seen in wild type.

CLPB3 partitions into aggregates of high molecular weight after heat stress
To assess the oligomeric state of CLPB3, we subjected wild type, clpb3 mutants and complemented lines to the 1 h 41 °C/6 h 25 °C heat shock/recovery regime as before, and analysed whole-cell proteins by blue-native PAGE and immunoblotting. We detected specific signals for CLPB3 that we assigned to monomers and aggregates of high molecular weight (Fig. 4). Although a signal was observed at the height of photosystem (PS) I that could correspond to CLPB3 hexamers (PSI has a molecular mass of ~600 kDa; Amunts et al., 2007), the equal intensity of this signal in all lines rather argues for a cross-reaction of the CLPB3 antibody with a PSI subunit. In wild type, the signals for monomers and aggregates increased strongly after heat stress, and remained strong after the recovery phase. In the clpb3-1 mutant the monomer was virtually absent indicate transcriptional start sites. WT, wild type; clpb3-c, complemented mutants; P A , P R , HSP70A and RBCS2 promoters, respectively; T RPL23 , RPL23 terminator. (B) Immunoblot analysis of the accumulation of CLPB3 and selected chloroplast proteins. Cells were grown in continuous light at 25 °C (CL), exposed to 41 °C for 1 h (HS), and allowed to recover at 25 °C for 6 h after the heat treatment (R). For the analysis, 10 µg of whole-cell proteins (100%) were used. (C) Quantification of immunoblot analyses. Values are means from three independent experiments (including two technical repeats for CLPB3 and HSP22E/F), normalized first by the median of all signals obtained with a particular antibody in the same experiment, and then by the mean signal of the heat-stressed wild type. Error bars represent standard deviation. Asterisks indicate significant differences with respect to the WT (two-tailed, unpaired t-test with Bonferroni-Holm correction, P<0.05). The absence of an asterisk indicates no significant differences. (3-2c). Cells were grown in continuous light at 25 °C. For the analysis, 10 µg (100%), 5 µg (50%) and 2.5 µg (25%) of whole-cell proteins were used. (B) Quantification of immunoblot analyses as described for Fig. 2C with normalization of protein levels relative to WT. Error bars represent standard deviation (n=3). Asterisks indicate significant differences with respect to the WT (two-tailed, unpaired t-test with Bonferroni-Holm correction, P<0.05). The absence of an asterisk indicates that there were no significant differences. under all conditions, while a very strong signal was detected in aggregates after heat shock and recovery. The same pattern was observed also in the clpb3-1c line, but there the transgenic CLPB3 monomer was detected. Low production of the monomer was detected in both the clpb3-2 mutant and complemented line clpb3-2c. Both these lines exhibited much weaker signals for CLPB3 in aggregates than those observed in wild type, clpb3-1, and clpb3-1c (Fig. 4).

CLPB3 and HSP22E/F localize in stromal foci and to the area occupied by the thylakoid membrane system, respectively
We next employed immunofluorescence to localize CLPB3 in cells of complemented lines exposed to the 1 h 41 °C/6 h 25 °C heat shock/recovery regime as before (for clpb3-2c) or to 1 h 41 °C only (for clpb3-1c). Since we expected a co-localization of HSP22E/F and CLPB3 in aggregates, we employed mouse antibodies against the HA tag to detect transgenic CLPB3, and rabbit antibodies against HSP22E/F, on the same cells. In all cells, HSP22E/F was weakly detectable under ambient conditions, and gave rise to strong signals in the chloroplast after heat shock and recovery (Fig. 5), corroborating earlier findings (Rütgers et al., 2017a). As expected, the HA antibody produced no signals in wild-type cells, but recognized HA-tagged CLPB3 in the complemented lines. In clpb3-2c, CLPB3 was evenly dispersed throughout the chloroplast under ambient conditions, but partitioned into stromal foci after heat stress. These foci largely vanished after the recovery phase (Fig.  5, top panel). Since transgenic CLPB3-HA production did not increase during heat stress (Fig. 2), the stromal foci must be formed by the condensation of existing CLPB3-HA pro-tein under heat and redistribution during recovery. In clpb3-1c, CLPB3 localized to stromal foci already under ambient conditions that became stronger and more condensed after heat stress (Fig. 5, bottom panel). To our surprise, in both complemented lines, HSP22E/F and CLPB3 hardly co-localized after heat shock. Rather, the CLPB3 stromal foci were located in close proximity to HSP22E/F, which appeared to be largely in an area occupied by the thylakoid membrane system.

The removal of aggregated proteins during recovery from heat stress is impaired in clpb3 mutants
Chloroplast CLPB3 from Arabidopsis was shown to exhibit disaggregase activity in vitro (Parcerisa et al., 2020). To elucidate a disaggregase function of Chlamydomonas CLPB3 in vivo, we exposed wild type and clpb3 mutants to our 1 h 41 °C/6 h 25 °C heat shock/recovery regime, and purified protein aggregates were analysed by SDS-PAGE and Coomassie staining (Fig. 6A). In wild type, the abundance of insoluble proteins increased after the heat treatment, but during recovery this was reduced to the same levels as before the heat treatment. This was not the case in the clpb3 mutants, in which insoluble proteins persisted after recovery. To quantify this finding, we performed the same experiment with wild type, clpb3 mutants and complemented lines, and analysed the abundance of CLPB3, HSP22E/F, and thermolabile TIG1 in purified aggregates (Fig. 6B, C). In all lines, the three proteins were barely detectable in non-soluble proteins prepared from cells kept under ambient conditions, but accumulated strongly in aggregates collected after the heat treatment. In the wild type, after 6 h recovery from heat, the abundance of CLPB3, HSP22E/F, and TIG1 in aggregates was reduced to Fig. 4. Analysis of the oligomeric state of CLPB3. Whole-cell proteins from wild-type (WT), clpb3 mutants and complemented lines exposed to the heat shock/recovery regime used in Fig. 2B were solubilized with 1% β-DDM and subjected to BN-PAGE. A lane of the gel after electrophoresis is shown at the left with PSII supercomplexes (I+II), PSI-LHCI (II), PSII dimers (III), ATP synthase (IV), PSII monomers/Cyt b 6 f complex (V), LHCII trimers (VI), and LHCII monomers (VII) visible as prominent bands. On the right is an immunoblot of the gel probed with antibodies against CLPB3. A, aggregates; M, CLPB3 monomers. The asterisk indicates a protein, presumably of PSI, that cross-reacts with the CLPB3 antibody. CL, 25 °C; HS, 1 h at 41 °C; R, 6 h recovery at 25 °C.

CLPB3 improves thermotolerance in Chlamydomonas
We wondered whether the impaired ability of the clpb3 mutants to remove aggregates was associated with a growth phenotype. To test this, we spotted serial dilutions of cultures of wild type, clpb3 mutants, and complemented lines onto agar plates and monitored growth under mixotrophic and photoautotrophic conditions in low and high light, heterotrophic conditions, and mixotrophic conditions with three ~24 h heat shock expo-sures at 40 °C and ≤24 h recovery periods in between (Fig.  7A). Under all growth conditions at ambient temperatures, we found no growth phenotype for the clpb3-2 mutant. The clpb3-1 mutant exhibited a mild growth phenotype under photoautotrophic conditions in low light and high light, which was ameliorated in the complemented line clpb3-1c. Clearly reduced growth was observed for both clpb3 mutants after the repeated prolonged heat stress treatments. This phenotype was ameliorated in clpb3-2c, but not in clpb3-1c. It was observed that growth phenotypes were only revealed after longer heat exposures. To substantiate this finding, we exposed cultures of wild type, clpb3 mutants, and complemented lines to 40 °C for 72 h and allowed them to recover for 120 h at 25 °C. As shown in Fig. 7B, the heat treatment strongly impaired growth of the wild type, but cells resumed growth during the recovery phase. Growth during heat treatment and recovery was abolished in the clpb3-1 mutant and impaired in the clpb3-2 mutant when compared with the wild type. This phenotype was ameliorated in both complemented lines, but the wild-type phenotype was Fig. 5. Sub-cellular localization of CLPB3 and HSP22EF. Cells were exposed to the heat shock/recovery regime used in Fig. 2B for wild type (WT) and clpb3-2c, or only to a 1 h heat shock treatment (for WT and clpb3-1c). HSP22E/F (22EF) and HA-tagged CLPB3 (HA) were detected by immunofluorescence using antibodies against HSP22E/F (magenta) and the HA epitope (green). Merge: overlay of both signals. Scale bars=2 µm.
not restored, most likely because the expression levels achieved with the transgenic CLPB3 gene controlled by the constitutive HSP70A-RBCS2 promoter were insufficient under heat stress conditions. We also determined survival rates for wild type and the clpb3-1 mutant after exposure at 41 °C for 2 h, and found significantly lower survival in the mutant (60%) versus the wild type (89.5%; P=0.025) (Fig. 7C). Hence, the growth phenotype after heat stress in the different lines correlated with their ability to remove aggregated proteins during recovery. Fig. 6. Analysis of aggregate formation and removal in wild type (WT), clpb3 mutants, and complemented lines. (A) Cells were exposed to the heat shock/recovery regime used in Fig. 2B. Total cell proteins and purified aggregates for each condition were separated by SDS-PAGE and stained with Coomassie blue. (B) Immunoblot analysis using antibodies against CLPB3, HSP22E/F, and TIG1 on total cell proteins and aggregates. (C) Quantification of the immunoblot analyses. Values represent the percentage of protein left in aggregates after 6 h of recovery from three independent experiments. Error bars represent standard deviation. Asterisks indicate significant differences between mutant and its respective complemented line (two-tailed, unpaired t-test, P<0.05). The absence of an asterisk means that there were no significant differences. CL, 25 °C; HS, 1 h at 41 °C; R, 6 h recovery at 25 °C.

The resolving of heat-induced protein aggregates by chloroplast CLPB3 is required for thermotolerance in Chlamydomonas
Cytosolic HSP101 is required for the resolving of heat-induced protein aggregates, and this activity is essential for basal and acquired thermotolerance in land plants Vierling, 2000, 2001;Queitsch et al., 2000;Nieto-Sotelo et al., 2002;Agarwal et al., 2003;Katiyar-Agarwal et al., 2003;McLoughlin et al., 2016). The situation in chloroplasts is not as clear: while chloroplast CLPB3 is required for thermotolerance in tomato (Yang et al., 2006), this is not the case for chloroplast CLPB3 in Arabidopsis (Myouga et al., 2006;Lee et al., 2007). Nevertheless, Arabidopsis CLPB3 has been shown to be capable of resolving aggregates of model substrate G6PDH in vitro, and aggregates formed by aggregation-prone DXS in vivo Pulido et al., 2017;Parcerisa et al., 2020). Cyanobacterial ClpB, from which chloroplast CLPB3 is derived (Lee et al., 2007;Mishra and Grover, 2016), has also been shown to be required for thermotolerance Clarke, 1996, 2000). In this study, we show that chloroplast CLPB3 is required for the resolving of heat-induced protein aggregates in Chlamydomonas (Fig. 6), and that this activity is required for conferring thermotolerance under severe heat stress conditions (Fig. 7). It is possible that a role for chloroplast CLPB3 in conferring thermotolerance in Arabidopsis is obscured by the strong chloroplast development phenotype in Arabidopsis clpb3 mutants and CLPB3 overproduction lines (Myouga et al., 2006;Lee et al., 2007;Zybailov et al., 2009). It is also possible that in the Arabidopsis clpb3 mutant other chaperones are overproduced that prevent a thermosensitive phenotype. , clpb3 mutants, and complemented lines were grown to log phase, diluted, and spotted onto agar plates with the cell numbers indicated. TAP plates were used for monitoring mixotrophic growth (light) or heterotrophic growth (dark), HSM plates for monitoring photoautotrophic growth. LL: low light at 30 µmol photons m -2 s -1 ; HL: high light at 600 µmol photons m -2 s -1 ; HS: three ~24 h heat shock exposures at 40 °C with ≤24 h recovery in between. The cell genotypes and number of cells spotted are shown. (B) Liquid cultures of WT, clpb3 mutants and complemented lines were grown to log phase, exposed to 40 °C for 72 h (HS) and allowed to recover at 25 °C for 120 h (R). Before the treatment, part of the culture was diluted and grown at 25 °C for 72 h (CL). Photographs of the cultures taken right after the corresponding treatment are shown. (C) WT and clpb3-1 mutant were grown to log phase at 25 °C and exposed to 41 °C for 2 h. Aliquots taken for each condition were diluted, plated on agar plates, and colony-forming units counted after 4 d at 25 °C to determine survival rates. Values are from four independent experiments done in triplicate. Error bars represent standard deviation. Differences were significant at P<0.05 (two-tailed, unpaired t-test).
Chlamydomonas cells appear to compensate for the loss of CLPB3 function by up-regulating the stromal DEG1C protease, and perhaps also by reducing chloroplast protein synthesis capacity, as suggested by a lower abundance of the PRPL1 plastid ribosomal subunit (Figs 2, 3). A reduced abundance of cytosolic and chloroplast ribosome subunits was observed in the Chlamydomonas deg1c mutant which, however, did not display elevated levels of CLPB3 (Theis et al., 2019). The loss of chloroplast CLPB3 function had no effect on the accumulation of other chloroplast chaperones in Chlamydomonas and tomato, including CPN60, HSP70, trigger factor, and sHSPs ( Fig. 2) (Yang et al., 2006). These observations suggest that a loss of chloroplast disaggregase activity appears to be compensated to some part by lowering the protein synthesis capacity and increasing protease activity, rather than by increasing other chaperone systems. However, further research is needed to draw such a conclusion.

CLPB3 dynamically localizes to stromal foci
We found that heat stress causes CLPB3 to organize in stromal foci by the condensation of the existing CLPB3 protein (Fig.  5). Although HSP22E/F were found in protein aggregates with stromal TIG1 (Fig. 6), and to interact with numerous stromal proteins after heat stress (Rütgers et al., 2017a), HSP22E/F localized largely to the area occupied by the thylakoid membrane system, with little overlap between CLPB3 and HSP22E/F signals (Fig. 5). While the stromal foci formed by CLPB3 in the clpb3-2c line largely vanished after the recovery phase, the HSP22E/F signals in the thylakoid membrane area persisted. These results are unexpected, since cytosolic HSP101 and sHSPs in Arabidopsis were found to largely co-localize in cytoplasmic foci (McLoughlin et al., 2016(McLoughlin et al., , 2019. It is possible that HSP22E/F play a dual role during heat stress, with their largest fraction partitioning to, and stabilizing thylakoid membranes, and a smaller fraction intercalating with stromal proteins in small aggregates to be resolved by CLPB3 in stromal foci. With the bulk HSP22E/F signal coming from the thylakoid system occupied area, this would explain why there is little overlap between the HSP22E/F and CLPB3 signals. Indeed, up to two thirds of Arabidopsis chloroplast Hsp21 have been shown to interact with thylakoid membranes during heat stress (Bernfur et al., 2017) and Hsp21 has been shown to stabilize thylakoid membranes and intrinsic protein complexes during heat stress (Chen et al., 2017).
In this scenario, the potential functions of HSP22E/F during heat stress would be divided into stabilizing thylakoid membranes and supporting CLPB3-mediated resolution of stromal aggregates. Since the CLPB3 stromal foci look like blobs sitting on HSP22E/F at stroma-exposed regions of the thylakoid system, could CLPB3 play a role there as well? We have previously shown that considerable amounts of HSP22E/F and DEG1C partition to chloroplast membranes upon oxidative stress, where HSP22E/F interact with VIPP1/2 and HSP70B (Theis et al., 2020). We proposed that misassembled, unfolded and aggregated proteins might induce lipid packing stress at chloroplast membranes that is sensed by the N-terminal amphipathic α-helix of VIPP2. VIPP2 might then serve as a nucleation point for VIPP1 and HSP22E/F to populate areas suffering from lipid packing stress, and prevent membrane leakage. In addition, these proteins might organize membrane domains that serve as interfaces between membrane and soluble chaperones and proteases for the handling of unfolded/aggregated membrane proteins and of aggregates of stromal proteins sticking to the membranes. In this case, CLPB3 might act by resolving such aggregates for refolding or degradation, e.g. via DEG1C. In fact, CLPB3 was found in the proxiome of VIPP1 in cells exposed to oxidative stress (Kreis et al., 2022). Moreover, cytosolic HSP101 has been shown to cooperate with the proteasome system, albeit only on a small sub-set of aggregated proteins, while refolding was the preferred path (McLoughlin et al., 2019). It is possible that thylakoid membrane proteins threaded through the CLPB3 pore might even be handed over to the membrane protein integrase AL-BINO3 (ALB3) for reinsertion into thylakoid membranes, to favour refolding over degradation. Definitely, more work is required to provide evidence for such a bold hypothesis. It is nevertheless attractive, as it provides a coherent function for the main players of the 'chloroplast unfolded protein response' regulon, VIPP1/2, HSP22C/E/F, DEG1C, and CLPB3.

Chlamydomonas CLPB3 appears not to be required for chloroplast development
We estimated chloroplast CLPB3 to account for ~0.2% of total cell proteins (Fig. 1). In comparison, the Hsp70 chaperone in the Chlamydomonas chloroplast, HSP70B, makes up ~0.19% of total cell proteins (Liu et al., 2007). When considering the molar masses, this results in a ratio of 1.4 HSP70B monomers per CLPB3 monomer, or about 10 HSP70B monomers per CLPB3 hexamer. Upon heat stress, the abundance of CLPB3 increases ~4-fold, while that of HSP70B increases ~2.5-fold (Fig. 2), and as shown previously (Mühlhaus et al., 2011). Hence, Chlamydomonas CLPB3 is a rather abundant chloroplast protein under ambient conditions, suggesting that it might perform housekeeping functions, as is the case in Arabidopsis (Myouga et al., 2006;Lee et al., 2007;Zybailov et al., 2009). However, in our Chlamydomonas clpb3 mutants we found no obvious chloroplast development phenotype ( Supplementary Fig. S5A), and no PSII phenotype ( Supplementary Fig. S5B) under ambient conditions. A mild growth phenotype was observed, especially under photoautotrophic conditions in the mutant clpb3-1 (Fig. 7). Obvious phenotypes under ambient conditions were neither observed in tomato clpb3 antisense lines (Yang et al., 2006) nor in Synechococcus sp. clpb3 knockout lines (Eriksson and Clarke, 1996).
Both Chlamydomonas clpb3 mutants accumulate CLPB3 to ~20% of wild-type levels (Figs 2, 3). While the 20% residual CLPB3 in mutant clpb3-2 represents wild-type protein, residually accumulating CLPB3 in mutant clpb3-1 has a truncated C-terminal domain (Figs 2, 3). If the mutagenesis cassette is indeed flanked by random DNA at its 3' end, as indicated by genotyping ( Supplementary Fig. S3), the truncation removes a stretch of ~20 amino acids that is highly conserved among ClpB family members, as well as a non-conserved stretch of 52 amino acids ( Supplementary Fig. S1A). These sequences are most likely replaced by some junk sequence until a random stop codon is encountered. The C-terminal domain of E. coli ClpB (Supplementary Fig. S1A) was shown to primarily support protein self-association and thus hexamer formation, which is required for ATP binding and chaperone activity. E. coli ClpB lacking the C-terminal domain could not form hexamers and had no chaperone activity (Barnett et al., 2000;Barnett and Zolkiewski, 2002;Mogk et al., 2003). Although the truncation in the clpb3-1 mutant removes only a small portion of the conserved part in the C-terminal domain ( Supplementary Fig. S1A), the missing sequences appear to be quite important for stability and functionality of CLPB3, as judged by the following observations: (i) the truncation obviously leads to a reduced accumulation of the protein (Fig.  2); (ii) truncated CLPB3 appears to localize in aggregates already under ambient conditions to which complementing wild-type CLPB3 is attracted (Fig. 5); (iii) truncated CLPB3 massively accumulates in aggregates during heat stress (Fig. 6); (iv) the clpb3-1 mutant is much more impaired in its ability to resolve heat-induced protein aggregates than the clpb3-2 mutant, albeit both accumulate similar levels of residual CLPB3 (Figs 2, 6); and (v) the clpb3-1 mutant is more thermosensitive than the clpb3-2 mutant (Fig. 7). Since wild-type CLPB3 can complement the clpb3-1 mutant to a significant extent (Figs 6, 7), the truncated protein variant apparently exerts no dominant negative effect, suggesting that it is severely impaired in oligomer formation, just like E. coli ClpB deprived of its C-terminal domain. According to the cryo-EM structure of Mycobacterium ClpB (Yin et al., 2021), the missing sequences in the clpb3-1 mutant could directly mediate contacts between neighbouring protomers or indirectly by stabilizing the C-terminal domain ( Supplementary Fig. S1B). Our data are consistent with a scenario where truncated CLPB3 engages with substrate proteins in a monomeric form, but cannot form oligomers, and thus has no chaperone activity. Nevertheless, to rule out any residual chaperone activity, and thus a role of Chlamydomonas CLPB3 in chloroplast development, clean knockout lines generated, for example, by CRISPR/ Cas9, will be required in future work.

Supplementary data
The following supplementary data are available at JXB online. Fig. S1. Alignment of amino acid sequences of CLPB proteins from E. coli and chloroplasts. Fig. S2. Production of recombinant CLPB3 in E. coli. Fig. S3. Analysis of the CIB1 integration sites in the CLPB3 gene by PCR. Fig. S4. Putative heat shock elements (HSEs) in the CLPB3 promoter. Fig. S5. Chlamydomonas clpb3 mutants display no obvious phenotype regarding chloroplast development and PSII activity. Fig. S6. Screening for complemented clpb3 mutant lines. Table S1. Primers used for cloning and genotyping.

Author contributions
EK performed all experiments, assisted by JN and MM; DS took the immunofluorescence images; MS conceived and supervised the project, and wrote the manuscript with contributions from all other authors.