The PUB4 E3 Ubiquitin Ligase Is Responsible for the Variegated Phenotype Observed upon Alteration of Chloroplast Protein Homeostasis in Arabidopsis Cotyledons

During a plant’s life cycle, plastids undergo several modifications, from undifferentiated pro-plastids to either photosynthetically-active chloroplasts, ezioplasts, chromoplasts or storage organelles, such as amyloplasts, elaioplasts and proteinoplasts. Plastid proteome rearrangements and protein homeostasis, together with intracellular communication pathways, are key factors for correct plastid differentiation and functioning. When plastid development is affected, aberrant organelles are degraded and recycled in a process that involves plastid protein ubiquitination. In this study, we have analysed the Arabidopsis gun1-102 ftsh5-3 double mutant, lacking both the plastid-located protein GUN1 (Genomes Uncoupled 1), involved in plastid-to-nucleus communication, and the chloroplast-located FTSH5 (Filamentous temperature-sensitive H5), a metalloprotease with a role in photosystem repair and chloroplast biogenesis. gun1-102 ftsh5-3 seedlings show variegated cotyledons and true leaves that we attempted to suppress by introgressing second-site mutations in genes involved in: (i) plastid translation, (ii) plastid folding/import and (iii) cytosolic protein ubiquitination. Different phenotypic effects, ranging from seedling-lethality to partial or complete suppression of the variegated phenotype, were observed in the corresponding triple mutants. Our findings indicate that Plant U-Box 4 (PUB4) E3 ubiquitin ligase plays a major role in the target degradation of damaged chloroplasts and is the main contributor to the variegated phenotype observed in gun1-102 ftsh5-3 seedlings.


Introduction
Albinotic and variegated mutants have been widely used to investigate chloroplast biogenesis and the communication pathways between the nuclear-cytosolic compartment and chloroplasts [1][2][3]. Among these, Arabidopsis ftsh5 (var1) and ftsh2 (var2) mutants, devoid of two subunits of the thylakoid transmembrane metalloprotease complex FTSH (Filamentation Temperature Sensitive protein H), responsible for photosystem II (PSII) biogenesis and repair, and characterized by a leaf variegated phenotype, contributed greatly in revealing the strict link between chloroplast protein homeostasis and proper chloroplast biogenesis [4][5][6][7][8]. Transmission electron microscopy analyses of ftsh2 and ftsh5 leaves revealed, indeed, correctly-shaped chloroplasts in the green sectors, while the white sectors showed miss-shaped plastids with highly vacuolated organelle structures [6,9]. Furthermore, second-site suppressor screens aimed to identify mutations able to suppress the ftsh leaf variegated phenotype, so-called Suppressors of Variegation (SVR), allowed the identification of several nuclear genes encoding plastid-located proteins mostly involved in rRNA maturation, translation, protein folding and protein degradation [3,[10][11][12][13][14]. Overall, these findings indicated that the differentiation of functional chloroplasts requires an

Plant Material and Growth Conditions
Arabidopsis (Arabidopsis thaliana, genetic background Col-0) wild-type and mutant seeds were grown on soil in a climate chamber under long-day conditions (16 h at 100 µmol photons m −2 s −1 light and 8 h dark, at 22 • C temperature). Genetic loci and T-DNA flanking regions of insertional mutant lines used in this work are described in Figure S1. Primer sequences for genotype determination are listed in Table S1. Multiple mutants were generated by manual crossing and identified by PCR-based segregation analyses of F2 populations, with the only exception being the gun1-102 ftsh5-3 pub4-7 mutant, where the pub4-7 mutant allele was generated using the CRISPR-Cas9 genome editing strategy. In particular, the gun1-102 ftsh5-3 pub4-7 triple mutant was generated by targeting the fourth exon of the PUB4 locus in the gun1-102 ftsh5-3 mutant background using the pHEE401E vector described by Xing et al. [28]. Mutant plants carrying the mutation of interest and devoid of the Cas9 endonuclease were selected in the T3 generation. Primer sequences used for guide RNA design are listed in Table S1. Cotyledon and leaf area were determined by the ImageJ software (http://imagej.nih.gov/ij/index.html, accessed on 15 August 2021). The Variegation Index (V.I.) was calculated as the ratio of green area over the total area of the organ.

Chlorophyll Fluorescence Measurements and Chlorophyll Quantification
The imaging Chl fluorometer (Walz Imaging PAM; https://walz.com/, accessed on 15 August 2021) was used to measure in vivo Chl a fluorescence. Six plants of each genotype were analyzed at 6 and 12 days after sowing (DAS) and average values plus-minus standard deviations were calculated. Dark-adapted plants were exposed to blue measuring beam (intensity 4) and a saturating light flash (intensity 4) to obtain the maximum quantum yield of PSII, Fv/Fm. A 5-min exposure to actinic light (36 µmol photons m −2 s −1 ) allowed for the calculation of the effective quantum yield of PSII, Y II . For Chl quantification, 100 mg (fresh weight) of 6 DAS seedlings were ground in liquid nitrogen and extracted in 90% acetone. Chlorophyll (Chl) a and b concentrations were measured according to Porra et al. [29]. Measurements were performed in triplicate.
Intact chloroplasts were isolated from 100 mg (fresh weight) 6 DAS seedlings according to Kunst [31], with a few changes. Samples were homogenized in 1 mL of 45 mM sorbitol, 20 mM Tricine-KOH pH 8.4, 10 mM EDTA, 10 mM NaHCO 3 and 0.1% (w/v) BSA fraction V, supplemented with proteinase inhibitor cocktail (cOmplete™, COEDTAF-RO, Roche; https://www.roche.com/, accessed on 15 August 2021), and centrifuged for 7 min at 700× g. The supernatant was discarded, while the pellet was washed twice with 1 mL of homogenization buffer. After a centrifugation step (7 min at 700× g), the pellet was collected as the intact chloroplast fraction. The level of signals was quantified by the ImageJ software (http://imagej.nih.gov/ij/index.html, accessed on 15 August 2021).

Accession Numbers
The Arabidopsis Genome Initiative accession numbers for the genes mentioned in this work can be found at TAIR (https://www.arabidopsis.org/, accessed on 15 August 2021): GUN1 (AT2G31400), FTSH5 (AT5G42270), FUG1 (AT1G17220), PRPS21 (AT3G27160), cpHSC70-1 (AT4G24280), SP1 (AT1G63900), PUB4 (AT2G23140), PP2AA3 (AT1G13320) and HSFA2 (AT2G26150). To dissect the molecular mechanisms responsible for the variegated phenotype observed in gun1-102 ftsh5-3 cotyledons and true leaves, as a consequence of chloroplast protein homeostasis perturbation ( [16]; Figures 1a and 2a), we introgressed mutations in genes involved in: (i) plastid translation, fug1-3 and prps21-1 [32,33]; (ii) plastid protein folding/import, cphsc70-1 [34,35]; (iii) cytosolic protein ubiquitination, sp1-3 and pub4-7 [22,24] (see Figure S1 and Table 1), with the aim of identifying Arabidopsis triple mutants able to suppress the gun1-102 ftsh5-3 double mutant phenotype. The large majority of the mutations investigated are caused by T-DNA insertions in the transcribed regions of the genes of interest, resulting in knock-out alleles. The only exception is represented by fug1-3 knock-down allele caused by a T-DNA insertion in the promoter region of At1g17220 locus (see Figure S1). Moreover, since the PUB4 and GUN1 loci are located on adjacent regions of chromosome 2, the PUB4 gene was silenced in the gun1-102 ftsh5-3 genetic background by using the CRISPR-Cas9 gene editing strategy. In particular, the pub4-7 allele is due to a frameshift mutation caused by the insertion of a Thymine in the fourth exon of the At2g23140 gene, in position +1804 from the transcription starting site (see Figure S1). Table 1. List of mutant lines analysed in this study. Gene names, AGI accession numbers, subcellular localization of the gene products, functions and references where mutant alleles have been described are included.

SP1
At1g63900 cytosolic side of the plastid outer envelope sp1-3 [22] E3 ubiquitin ligase involved in the regulation of plastid protein import and component of CHLORAD, i.e., ubiquitination and retro-translocation of outer membrane proteins for proteasome-dependent degradation [22].

PUB4 E3 Ubiquitin Ligase Plays a Major Role in Chloroplast Degradation upon Alteration of Plastid Protein Homeostasis
To further investigate the role of PUB4 as a component of the chloroplast quality control machinery (see also [16]), thin sections of cotyledons from the Col-0, gun1-102 ftsh5-3 double mutant, pub4-2 T-DNA insertional mutant ( Figure S1) and gun1-102 ftsh5-3 pub4-7 triple mutant were analysed by Transmission Electron Microscope (TEM) (Figure 3a). Chloroplasts of Col-0 and pub4-2 cotyledons appeared to be correctly shaped, with the proper thylakoid ultrastructure organization in grana stacks and stroma lamellae and abundant starch granules in the stroma, as in the case of gun1-102 and ftsh5-3 chloroplasts (Figure 3a; see also Tadini et al. [16]). In agreement with the V.I. values, several plastids devoid of thylakoid membranes and with large vesicles either inside the stroma (Figure 3a) or budding from the envelope, as also reported in Tadini et al. [16], were observed in gun1-102 ftsh5-3 cotyledon cells, indicating advanced chloroplast degradation.
On the contrary, gun1-102 ftsh5-3 pub4-7 cotyledons contained functional chloroplasts with severely altered shape, but that were still able to accumulate thylakoid membranes organized in grana and stroma lamellae and that were capable of performing photosynthesis, as proven by the accumulation of starch granules (Figure 3a). The suppression of gun1-102 ftsh5-3 variegated phenotype, achieved by the introgression of pub4-7 mutation, was also investigated by measuring the cotyledon chlorophyll content through a spectrophotometerbased assay (Figure 3b). Chlorophyll content was similar in Col-0, gun1-102, ftsh5-3 and pub4-2 cotyledons, while it was reduced to about 60% of control levels in gun1-102 ftsh5-3 seedlings. In agreement with previous observations (Figures 1 and 3a), gun1-102 ftsh5-3 pub4-7 cotyledons showed almost double the amount of chlorophyll compared to the variegated gun1-102 ftsh5-3 cotyledons.

The Absence of PUB4 E3 Ubiquitin Ligase Activity Increases Cytosolic Protein Folding Stress
Immunoblot analyses of total protein extracts from Col-0, gun1-102, ftsh5-3, pub4-2,  gun1-102 ftsh5-3 and gun1-102 ftsh5-3 pub4-7 cotyledons were performed to monitor the accumulation of components of the thylakoid electron transport chain, i.e., LHCB5, of thylakoid membrane biogenesis/maintenance, such as VIPP1 [37,38], of the plastid protein homeostasis machinery, such as CLPB3, cpHSC70-1 and CPN60 [34,39,40] and of the plastid pre-protein guidance complex and molecular markers of the cytosolic folding stress, as in the case of HSC70-4 and HSP90-1 cytosolic chaperones [17,26,41] (Figure 4a).  -0, gun1-102, ftsh5-3, gun1-102 ftsh5-3, pub4-2 and gun1-102 ftsh5-3 pub4-7  In agreement with previous observations, the higher level of LHCB5 in gun1-102 ftsh5-3 pub4-7 cotyledons with respect to gun1-102 ftsh5-3 further supported the ability of pub4-7 mutation to suppress the variegated phenotype of gun1-102 ftsh5-3 cotyledons. Furthermore, the accumulation of VIPP1 appeared comparable among all the genotypes except for gun1-102 ftsh5-3 pub4-7, which showed a slight increase in VIPP1 levels, while CLPB3, cpHSC70-1 and CPN60 resulted unchanged in all the genotypes. Interestingly, the accumulation of HSC70-4 and HSP90-1 cytosolic chaperones increased markedly in both the gun1-102 ftsh5-3 double mutant and gun1-102 ftsh5-3 pub4-7 triple mutant (Figure 4a) in line with what observed in gun1 single mutant upon Lincomycin treatment [16,17], pointing to a relatively high cytosolic protein folding stress in both mutant backgrounds. Finally, the total protein ubiquitination level was assessed by a UBQ11 specific antibody. Overall, no evident difference in total signal intensity was observed between wild-type and the different mutant samples. However, the protein ubiquitination pattern in gun1-102 ftsh5-3 cotyledons showed two relatively strong bands at around 70 and 55 kDa (see arrow heads in Figure 4a), not detectable in Col-0 and in the single mutant cotyledons. Strikingly, the band at 55 kDa was also visible in gun1-102 ftsh5-3 pub4-7 cotyledons, while the signal at 70 kDa disappeared in the triple mutant genetic background (Figure 4a). The lack of important discrepancies in the ubiquitination levels among all of the genotypes, even in the absence of PUB4 E3 ubiquitin ligase enzyme, prompted us to investigate whether major differences could be observed at the chloroplast level, given the ability of PUB4 to promote chloroplast degradation upon alteration of plastid protein homeostasis (see above). To this aim, intact purified chloroplasts were isolated from cotyledons of Col-0 and mutant seedlings and chloroplast proteins probed with the UBQ11 specific antibody (see Figure 4b). While chloroplast ubiquitination levels of Col-0, gun1-102 and ftsh5-3 chloroplasts were rather similar and most of the ubiquitination signal was concentrated in a single band migrating at around 30 kDa, the plastid ubiquitination levels increased largely in the gun1-102 ftsh5-3 genetic background, where signals of ubiquitinated proteins could be also detected at molecular weights different from the 30 kDa main band (Figure 4b). Intriguingly, a large part of the increased ubiquitination signals disappeared upon depletion of PUB4 activity in gun1-102 ftsh5-3 pub4-7 chloroplasts (Figure 4b). In addition, the relative expression of HSFA2 transcription factor, a key regulator of the chloroplast-related nuclear stress response [42,43], was assessed via qRT-PCR (Figure 4c). Consistently with the increase in cytosolic chaperone accumulation (see Figure 4a), HSFA2 transcripts were strongly up-regulated in gun1-102 ftsh5-3 and, to an even larger extent, in gun1-102 ftsh5-3 pub4-7 seedlings.

Discussion
Variegated mutants represent an important genetic tool to investigate chloroplast biogenesis and to dissect the functional interactions of different pathways involved in plastid differentiation, chloroplast protein homeostasis, chloroplast quality control and degradation. In this manuscript, we studied the GUN1-mediated plastid development and degradation, in the context of altered plastid protein homeostasis, as shown by the variegated phenotype of gun1-102 ftsh5-3 seedlings.
As the gun1-102 ftsh5-3 double mutant possesses chloroplasts prone to suffer terminal damages, the recovery of gun1-102 ftsh5-3 pub4-7 phenotype observed in cotyledons and true leaves (Figures 1 and 2) suggests that PUB4-mediated ubiquitination is the preferential pathway for degrading and recycling damaged chloroplasts upon alteration of plastid protein homeostasis ( Figure 5). These observations are in agreement with previous studies, in which ROS-damaged chloroplasts in Arabidopsis plants lacking the plastid ferrochelatase 2 (fc2-1) exhibited a PUB4-dependent increase in poly-ubiquitinated proteins [24]. The altered shape of chloroplasts observed in gun1-102 ftsh5-3 pub4-7 cotyledons (Figure 3a) is compatible with the early steps of the degradation process, not effectively achieved or delayed due to the absence of PUB4 protein activity, as previously reported [24]. In line with this, the up-regulation of the stress-responsive gene HSFA2, here used as molecular marker of damaged chloroplasts and the related nuclear stress response [42,43], observed in gun1-102 ftsh5-3, was comparable or even higher in gun1-102 ftsh5-3 pub4-7, proving that pub4 mutation does not suppress the damage itself ( Figure 4). It is tempting to explain the higher HSFA2 expression as a consequence of the increased number of chloroplasts in cells that, although defective, are still performing photosynthesis and, consequently, generating ROS and other stress-related signals. Worth noticing, HSFA2 expression is triggered by a GUN1-indepentent retrograde signalling pathway, as previously observed [43]. Furthermore, the introgression of pub4-7 mutation does not suppress the increased cytosolic folding stress observed in gun1-102 ftsh5-3 cotyledons (Figure 4), and previously described in gun1 seedlings upon lincomycin treatment [16,17] as proven by the over-accumulation of the cytosolic chaperones HSC70-4 and HSP90-1 and the increased total protein ubiquitination level. Noteworthy is that the comparison between gun1-102 ftsh5-3 and gun1-102 ftsh5-3 pub4-7 ubiquitination profiles, performed, for example, by mass spectrometry analyses of ubiquitin-enriched protein extracts, seems to be an optimal strategy to identify those chloroplast outer envelope proteins that are modified by the PUB4 ubiquitin ligase.
On the contrary, the sp1-3 mutation, which abolishes the activity of a transmembrane ubiquitin ligase responsible for controlling TOC protein complex composition and activity, does not suppress variegation in gun1-102 ftsh5-3 cotyledons and leaves (Figures 1 and 2). SP1 activity has been recognized as essential for chloroplast biogenesis in etiolated seedlings and to play a role in the tolerance of abiotic stresses, by modulating the import rate of photosynthetic proteins, and, therefore, to prevent ROS accumulation [22,25,26]. In light of this, the gun1-102 ftsh5-3 sp1-3 triple mutant is devoid of an additional level of chloroplast quality control mechanism that results in an overall decrease in plant growth with respect to gun1-102 ftsh5-3 (Figures 1 and 2). Taken together, these observations corroborate the model in which ubiquitination of specific chloroplast proteins triggers different mechanisms of stress tolerance, in cotyledons and leaves, not only in response to abiotic stresses but also as adaptation to genetic defects.
The introgression of cphsc70-1 mutation, impaired in plastid protein import and protein folding [34,35], into the gun1-102 ftsh5-3 genetic background led to a complete seedling lethal phenotype. This is in agreement with previous results, as GUN1 was shown to have functional interactions with FTSH5 and cpHSC70-1 loci [16,17]. The fact that ftsh5-3 cphsc70-1 double mutant displays no additive phenotype compared to cphsc70-1 ( Figure S3), while the gun1-102 ftsh5-3 cphsc70-1 shows seedling-lethality (Figure 1), suggests that plastid protein import, protein folding and plastid proteolysis act synergistically for achieving the correct chloroplast development and GUN1 is essential for such orchestration.
At last, in line with previous observations [10,32,33], the suppression of variegated phenotype was effectively achieved in gun1-102 ftsh5-3 true leaves by reducing the plastid translation rate, driven by the introgression of either fug1-3 or prps21-1 mutated alleles. Strikingly, this compensatory effect failed to occur in cotyledons, proving further that the pathways that underlie chloroplast biogenesis in cotyledons and leaves are characterized by substantial differences. The electron dense material inside the vacuole might derive from plastid degradation. The expression of the stressresponsive HSFA2 nuclear gene is strongly up-regulated, in response to signals from chloroplasts with altered protein homeostasis. Furthermore, the accumulation of cytosolic chaperones, namely HSP90 and HSC70, is increased, while the removal of damaged chloroplasts is favoured by PUB4 ubiquitin ligase. (c) The absence of PUB4 in the gun1-102 ftsh5-3 pub4-7 genetic background largely prevents the degradation of damaged chloroplasts, resulting in the suppression of variegation in both cotyledons and true leaves. Nevertheless, damaged chloroplasts, still functioning, trigger stress-related retrograde signals and lead to even higher HSFA2 expression levels.

Supplementary Materials:
The following are available online at www.mdpi.com/2073-4425/12/9/1387/s1, Figure S1: Graphical representations of the loci mentioned in this study, Figure  S2: Photosynthetic performance of Arabidopsis Col-0 and mutant seedlings, Figure S3: Visible On the contrary, chloroplasts in advanced degradation stages and vacuoles filled with electron-dense material are visible in gun1-102 ftsh5-3 variegated cotyledons. The electron dense material inside the vacuole might derive from plastid degradation. The expression of the stressresponsive HSFA2 nuclear gene is strongly up-regulated, in response to signals from chloroplasts with altered protein homeostasis. Furthermore, the accumulation of cytosolic chaperones, namely HSP90 and HSC70, is increased, while the removal of damaged chloroplasts is favoured by PUB4 ubiquitin ligase. (c) The absence of PUB4 in the gun1-102 ftsh5-3 pub4-7 genetic background largely prevents the degradation of damaged chloroplasts, resulting in the suppression of variegation in both cotyledons and true leaves. Nevertheless, damaged chloroplasts, still functioning, trigger stress-related retrograde signals and lead to even higher HSFA2 expression levels.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/genes12091387/s1, Figure S1: Graphical representations of the loci mentioned in this study, Figure S2: Photosynthetic performance of Arabidopsis Col-0 and mutant seedlings, Figure S3: Visible phenotypic characteristics and Variegation Index of Arabidopsis Col-0 and mutant seedlings at cotyledon stage, Table S1: Sequences of oligonucleotides employed for the molecular characterization of mutant lines.
Author Contributions: P.P. and L.T. designed the study. N.J., L.R., A.C. and L.T. took care of isolation of single and higher-order mutants. N.J., L.R. and G.F. took care of the phenotypical description of the mutants. N.J., L.R. and L.T. performed the molecular biology and biochemical characterization of mutants. N.J., L.R., P.P. and L.T. helped drafting the manuscript. L.T. and P.P. coordinated the study and took care of the final version of the manuscript. All authors have read and agreed to the published version of the manuscript.