SUMOylation contributes to proteostasis of the chloroplast protein import receptor TOC159 during early development

Chloroplast biogenesis describes the transition of non-photosynthetic proplastids to photosynthetically active chloroplasts in the cells of germinating seeds. Chloroplast biogenesis requires the import of thousands of nuclear-encoded preproteins by essential import receptor TOC159. We demonstrate that the small ubiquitin-related modifier (SUMO) pathway crosstalks with the ubiquitin–proteasome pathway to affect TOC159 stability during early plant development. We identified a SUMO3-interacting motif (SIM) in the TOC159 GTPase domain and a SUMO3 covalent SUMOylation site in the membrane domain. A single K to R substitution (K1370R) in the M-domain disables SUMOylation. Compared to wild-type TOC159, TOC159K1370R was destabilized under UPS-inducing stress conditions. However, TOC159K1370R recovered to same protein level as wild-type TOC159 in the presence of a proteasome inhibitor. Thus, SUMOylation partially stabilizes TOC159 against UPS-dependent degradation under stress conditions. Our data contribute to the evolving model of tightly controlled proteostasis of the TOC159 import receptor during proplastid to chloroplast transition.


Introduction
Chloroplasts are unique organelles that carry out photosynthesis. Although chloroplasts contain their own genome, the majority of chloroplast proteins are encoded by the nuclear genome and synthesized as preproteins in cytosol, and these preproteins are imported into the chloroplast through TOC-TIC complexes (Translocon at the Outer or Inner membrane of the Chloroplast) (Jarvis and Ló pez-Juez, 2013). The core of the TOC complex contains two related GTP-dependent preprotein receptor GTPases, TOC159 and TOC33, which interact with a b-barrel membrane protein, TOC75, that forms a protein-conducting channel, and is regulated by specific interactions with nuclearencoded preproteins Kessler et al., 1994;Jarvis et al., 1998). TOC159 is a major point of entry for highly abundant photosynthesis-associated preproteins arriving at the translocon complex. It is therefore regarded as the major chloroplast protein import receptor. TOC159 has a three-domain structure: a highly acidic N-terminal domain (A-domain), a central GTP-binding domain (G-domain), and a C-terminal membrane anchor domain (M-domain). Chloroplast biogenesis, the transition of a non-photosynthetic proplastid to a photosynthetically active chloroplast, depends on the essential import receptor TOC159 and its mutation results in non-photosynthetic albino plants (Bauer et al., 2000).
Recent studies have shown that during the etioplast to chloroplast transition, the TOC components are ubiquitinated by a novel chloroplast RING-type E3 ubiquitin ligase SP1 (suppressor of ppi1 locus 1). Ubiquitinated TOC components are extracted from the chloroplast outer envelope membrane with the help of SP2 (an Omp85-like b-barrel protein) and Cdc48 (a cytosolic AAA+ chaperone) providing the extraction force. The ubiquitinated TOC component is degraded by the 26S proteasome in the cytosol (Ling et al., 2012;Ling et al., 2019). This proteolytic pathway has been named CHLORAD. In a different context, chloroplast biogenesis takes place during the proplastid to chloroplast transition. It is dependent on the plant hormone gibberellic acid (GA). Under unfavorable seed germination conditions when gibberellic concentrations are low, a DELLA (RGL2) (a negative regulator of GA signaling) promotes the ubiquitylation and degradation of TOC159 by the 26S proteasome to insertion into the outer membrane of the chloroplast. This mechanism delays the onset of chloroplast biogenesis at an early developmental stage and has been shown to be independent of SP1 and presumably CHLORAD (Shanmugabalaji et al., 2018).
A recent study employed a yeast-two-hybrid screen to identify putative Arabidopsis small ubiquitin-like modifier (SUMO) substrates. The E2 SUMO conjugating enzyme (SCE1) was used as bait, and TOC159 was identified as a high-probability interaction candidate. TOC159 contains a predicted SUMO attachment site (yKXE/D) (Elrouby and Coupland, 2010). The SUMO, a 11 kDa protein, covalently modifies a large number of proteins. SUMO-dependent regulation is involved in many cellular processes, including gene expression, signal transduction, genome maintenance, protein localization, and activity (Gill, 2004). SUMOylation plays a crucial role in plant development and stress responses (Elrouby, 2017;Augustine and Vierstra, 2018;Morrell and Sadanandom, 2019). This reversible and dynamic SUMOylation starts with the attachment of SUMO to the target protein by a conjugation pathway, mechanistically analogous to the ubiquitylation system (Augustine and Vierstra, 2018;Novatchkova et al., 2012). The target protein covalently modified by SUMO performs specific functions, which may subsequently be reversed by SUMO proteases that hydrolyze the isopeptide bond between SUMO and the target protein (Yates et al., 2016). In addition, SUMO can also interact with target proteins through a SUMO-interacting motif (SIM). The non-covalent SUMO-SIM interaction may work as a molecular signal for protein-protein interaction and affect stability of proteins (Geiss-Friedlander and Melchior, 2007).
In this study, we address the role of SUMO modification of TOC159 in the context of the proplastid to chloroplast transition and investigate both covalent SUMOylation and non-covalent SUMO interaction. The biochemical and genetic evidence showed that the TOC159 G-domain contains a SIM that interacts with the SUMO3 and that SUMOylation of TOC159 at the M-domain is the key regulatory mechanism to protect it from further depletion under low GA conditions. The data provides new insight for TOC159 SUMO-binding and SUMOylation, demonstrating that SUMOylation positively influences protein stability with regard to the UPS. Thereby SUMOylation contributes to the proteostatic fine-tuning of TOC159 levels during TOC159-dependent chloroplast biogenesis in early plant development.

SUMO3 interacts with the TOC159 G-domain and is SUMOylated at the TOC159 M-domain
An earlier study using an in vitro SUMOylation assay showed that of the three SUMO isoforms (SUMO1, SUMO2, and SUMO3), only SUMO3 SUMOylated TOC159 (Elrouby and Coupland, 2010). TOC159 has a N-terminal A-(acidic-), a central G-(GTP-binding domain), and a C-terminal M-(membrane) domain (Bauer et al., 2000). The A-domain is known to be non-essential for TOC159 function and exquisitely sensitive to protease activity Agne et al., 2010). It was excluded from our DNA constructs and only the TOC159 G-and M-domains (TOC159GM) were used. In addition to a covalent SUMOylation site, TOC159 may also have SIM. To investigate the possibility of SUMO interaction with TOC159GM, we used the GPS-SUMO prediction algorithm (http://sumosp. biocuckoo.org/online.php) to search for SIM in TOC159GM (Zhao et al., 2014). Based on the search results, there is a predicted SIM (VKVLP) in the G-domain ( Figure 1A). To analyze the physical interaction between TOC159GM and SUMO isoforms, we performed yeast two-hybrid assays using TOC159G and TOC159M separately as baits. TOC159G interacted exclusively with SUMO3, but not SUMO1 and SUMO2. None of the SUMO isoforms interacted with TOC159M in the yeast two-hybrid assay ( Figure 1B, Figure 1-figure supplement 1). We further confirmed the TOC159GM-SUMO3 interaction by co-immunoprecipitation of GFP-TOC159GM and SUMO3-MYC by in planta transient co-expression using the Nicotiana benthamiana system ( Figure 1C).
We used the GPS-SUMO algorithm (http://sumosp.biocuckoo.org/online.php) to search for covalent SUMOylation sites in TOC159GM. A high scoring consensus SUMOylation site with a strongly conserved motif (TGVKLED) and containing a potentially SUMOylatable lysine (K1370) was identified  Figure 1E). To investigate the SUMOylation of TOC159GM, we selected the SUMO3 isoform based on the earlier in vitro study (Elrouby and Coupland, 2010). We infiltrated Nicotiana benthamiana with 35S-GFP-TOC159GM or GFP-TOC159GM-K/R (replacing lysine with a non-sumoylatable arginine residue at position 1370) each together with or without 35S-SUMO3-MYC. To analyze the infiltration experiments identical amounts of total extracts were subjected to immunoprecipitation using anti-GFP-beads followed by western blotting. An anti-GFP antibody was used to indicate total expression of GFP-TOC159GM and GFP-TOC159GM-K/R and resulted in bands of similar intensities in all four experiments. Anti-MYC and anti-SUMO3 were used to analyze conjugation of SUMO3-MYC to GFP-TOC159GM and GFP-TOC159GM-K/R. The western blotting using anti-MYC and anti-SUMO3 antibodies resulted in strong signals for GFP-TOC159GM ( Figure 1F, lane 3) but only very weak ones for GFP-TOC159GM-K/R ( Figure 1F, lane 4) when co-expressed with SUMO3-MYC. No signals were detected when the GFP-TOC159GM constructs were expressed in the absence of SUMO3-MYC ( Figure 1F, lanes 1 and 2). Note that GFP-TOC159GM co-infiltration with SUMO3 resulted in a much higher molecular mass band when analyzed with anti-MYC or -SUMO3 than the main, strong GFP-TOC159GM band detected with anti-GFP. It therefore appears that upon co-expression with SUMO3-MYC only a small fraction of GFP-TOC159GM was present in the SUMOylated form and that it had a higher molecular mass than GFP-TOC159GM alone ( Figure 1F). We further checked the SUMOylation assay by co-infiltration of GFP alone (control), GFP-TOC159GM, and GFP-TOC159GM-K/R together with SUMO3 in the N. benthamiana system. The co-immunoprecipitation gave strong signals for GFP-TOC159GM (

The non-SUMOylatable TOC159GM-K/R mutant complements the ppi2 mutation
It has been demonstrated that the A-domain of TOC159 is dispensable but that the M-domain of TOC159 is essential for protein import into the chloroplast (Lee et al., 2003). Furthermore, it was demonstrated that TOC159GM alone without the A-domain could complement the albino ppi2 phenotype . To characterize the effect of the K1370R mutation on the M-domain in vivo, we engineered transgenic lines expressing GFP-TOC159GM as well as GFP-TOC159GM-K/R under the TOC159 promoter in the ppi2 background. Two independent transgenic lines of pTOC159-GFP-TOC159GM:ppi2 and pTOC159-GFP-TOC159GM-K/R:ppi2 (called GFP-TOC159GM: ppi2 and GFP-TOC159GM-K/R:ppi2 plants hereafter) were isolated and the genotypes were confirmed by PCR using specific primer pairs (  replaced with R) with and without SUMO3-MYC in N. benthamiana leaves. Total protein extracts were subjected to immunoprecipitation with anti-GFP beads. The immunoprecipitated proteins from the expression of GFP-TOC159GM (lane 1) and GFP-TOC159GM-K/R (lane 2) alone as well as the coexpression with SUMO3 (lanes 3 and 4) were analyzed by western blotting using anti-GFP, anti-MYC and anti-SUMO3 antibodies. The online version of this article includes the following source data and figure supplement(s) for figure 1: Source data 1. Source data for Figure 1E.  Homozygous GFP-TOC159GM:ppi2 and GFP-TOC159GM-K/R:ppi2 transgenic lines gave green plants with almost identical chlorophyll concentrations that were very close to those of the wild type ( Figure 2C). This indicated complementation of the albino ppi2 phenotype. Western blotting analysis using GFP antibodies specifically detected GFP-TOC159GM and GFP-TOC159GM-K/R proteins. The blots revealed that TOC159GM and GFP-TOC159GM-K/R proteins accumulated to very similar levels in the respective transgenic lines. The blots were also probed with antibodies against TOC75 and TOC33, showing no significant differences between the two transgenic lines ( Figure 2B Figure 2D).

TOC159GM-K/R accumulation is diminished when compared to TOC159GM under low GA conditions
Addition of paclobutrazol (PAC) to the Murashige-Skoog (MS) growth medium can be used to inhibit GA biosynthesis and results in low GA conditions (Piskurewicz et al., 2008). Under low GA during early germination, the DELLA (RGL2) destabilizes TOC159 via the ubiquitin-proteasome system (Shanmugabalaji et al., 2018). To explore a possible role of SUMOylation of K1370 on protein stability, GFP-TOC159GM:ppi2 and GFP-TOC159GM-K/R:ppi2 were allowed to germinate in the presence or absence of PAC. As reported earlier, the GFP-TOC159GM protein level was severely reduced in GFP-TOC159GM:ppi2 seeds under low GA. In comparison, the level of GFP-TOC159-K/R in GFP-TOC159GM-K/R:ppi2 seeds was even more diminished. Moreover, TOC75 and TOC33 levels were also lower in mutant TOC159GM-K/R:ppi2 than in wild-type GFP-TOC159GM:ppi2 under low GA, whereas their levels were the same in the untreated lines ( Figure 3A,B, Figure 3-figure supplements 1 and 2). PAC-treated seeds, low in GA, accumulated very high levels of the RGL2 protein (Piskurewicz et al., 2008). We also compared RGL2 protein levels between TOC159GM:ppi2 and TOC159GM-K/R:ppi2 in the presence of PAC. The results revealed that there was no difference in RGL2 accumulation between the two lines ( Figure 3C, Figure 3-figure supplement 3A,B).

SUMOylation partially stabilizes the TOC159 under low GA conditions
To determine whether the diminished stability of GFP-TOC159GM-K/R under low GA in the presence of PAC is also due to UPS-mediated degradation, GFP-TOC159GM-K/R:ppi2 and GFP-TOC159GM:ppi2 seeds were germinated in the presence of PAC and subjected to treatment with or without the proteasome inhibitor MG132. Western blot analysis demonstrated that both GFP-TOC159GM-K/R and GFP-TOC159GM were rescued by MG132 and accumulated to the same level ( Figure 3E,F, Figure 3-figure supplement 4). The results suggest that SUMOylation partially protects GFP-TOC159GM against UPS-mediated degradation under low GA when compared to GFP-TOC159GM-K/R.
Accumulation of photosynthesis-associated proteins trended lower in TOC159GM-K/R:ppi2 than TOC159GM:ppi2 The results so far demonstrated that non-SUMOylatable TOC159GM-K/R is significantly more susceptible to UPS-mediated degradation than wild-type TOC159GM under low GA. To address whether this has an effect on the accumulation of photosynthesis-associated proteins (the presumed Source data 1. Source data for Figure 2C.

Discussion
It has been demonstrated previously that TOC159 physically interacts with the SUMO E2 enzyme in a yeast two-hybrid screen and that the SUMO3 isoform covalently SUMOylated TOC159 in an in vitro assay (Elrouby and Coupland, 2010). Apart from a covalent SUMOylation motif, the GPS-SUMO algorithm identified a conserved non-covalent SIM ('VKVLP') in the G-domain of TOC159. We confirmed this prediction using a yeast two-hybrid assay and co-immunoprecipitation experiment. It also revealed that the G-domain specifically interacted with the SUMO3 isoform and not with SUMO1 and 2. ( Figure 1B). It has been shown in Arabidopsis that the GA receptor GID1 interacted with SUMO1 through a SIM that prevented its interaction with a DELLA protein (Conti et al., 2014).
It is tempting to hypothesize that non-covalent binding of TOC159(SIM) to SUMO3 may modify interactions with SUMOylated complex partners such as TOC33/TOC75 or the DELLA (RGL2) under low GA conditions as these proteins also have predicted SUMOylation sites (http://sumosp.biocuckoo.org/online.php). TOC159 has highly conserved SUMOylation motif ('TGVKLED') with a lysine at position 1370 (K1370) within the M-domain of TOC159. We established SUMOylation at K1370 using an in planta SUMOylation assay in N. benthamiana ( Figure 1F, Figure 1-figure supplement 2B). Expression of non-SUMOylatable GFP-TOC159GM-K/R (containing the K1370R substitution) restored a wild-type green phenotype in the GFP-TOC159GM-K/R:ppi2 Arabidopsis plants. Note that ppi2 mutant plants normally have an albino phenotype. Based on the presence of wild-type levels of TOC75 and TOC33, it appeared that the TOC complex assembled normally in the GFP-TOC159GM-K/R:ppi2 plants (Figure 2A,B). Confocal laser microscopy localized GFP-TOC159GM-K/R at the envelope of Error bars indicate ± SEM (n = 3). two-tailed t test; *p<0.05; ***p<0.005. (C) Immunoblotting of total protein extracts from 3 days old GFP-TOC159GM:ppi2 and GFP-TOC159GM-K/R:ppi2 seedlings grown in the presence or absence of PAC (5 mM). The blot was probed with anti-GFP and -RGL2 antibodies. Anti-actin was used for a loading control. (D) Total protein extracts of 3 days old GFP-TOC159GM:ppi2 or GFP-TOC159GM-K/R:ppi2 grown seedling grown on PAC and subsequently treated with or without MG132 were analyzed by immunoblotting using anti-GFP antibodies and anti-UGPase for a loading control. (E) The specific bands corresponding to GFP and UGPase were quantified from three independent experiments (C). The quantified bands were normalized to GFP-TOC159GM in GFP-TOC159GM:ppi2 without MG132. Error bars indicate ± SEM (n = 3). two-tailed t test; **p<0.01. The online version of this article includes the following source data and figure supplement(s) for figure 3: Source data 1. Source data for Figure 3B, E, Figure 3-figure supplements 3 and 5. Figure supplement 1. Immunoblotting of total protein extracts from two independent GFP-TOC159GM:ppi2 lines (3.1 and 3.12) and two independent GFP-TOC159GM-K/R:ppi2 lines (4.2 and 4.8). Figure supplement 2. Immunoblotting of total protein extracts from GFP-TOC159GM:ppi2 and GFP-TOC159GM-K/R:ppi2 grown in the presence or absence of PAC (5 mM).   the chloroplast ( Figure 2D). Taken together, these results indicate that outer membrane insertion, TOC complex assembly, as well as chloroplast biogenesis take place normally under standard growth conditions despite the K1370R mutation and disabled SUMOylation.
In Arabidopsis, SUMOylation is triggered by environmental stimuli including biotic and abiotic stress. Hormone signaling (GA, auxin, brassinosteroids, and ABA) and development (root, seed, embryo, and meristem) are the main biological processes associated with SUMOylation in plants (Conti et al., 2014;Miura et al., 2009;Ishida et al., 2009;Augustine et al., 2016;Orosa-Puente et al., 2018;Kwak et al., 2019;Srivastava et al., 2020). While SUMOylation is independent of the ubiquitination pathway, there is the complex crosstalk between these two pathways. SUMOylation can promote the UPS-dependent degradation through SUMO-targeted Ub ligases (Perry et al., 2008). Contrary to this, SUMOylation may also protect from UPS-dependent degradation by blocking the ubiquitination of lysine residues (Creton and Jentsch, 2010;Jentsch and Psakhye, 2013;Ramachandran et al., 2015;Rott et al., 2017). Here, we explored the connection between TOC159 SUMOylation and the UPS-mediated TOC159 degradation under low GA during early developmental stages in seed germination.
We previously demonstrated that low GA concentrations brought about by PAC promote the DELLA (RGL2)-dependent TOC159 degradation via the UPS (Shanmugabalaji et al., 2018). Under low GA, GFP-TOC159GM-K/R accumulated to significantly (<50%) lower levels than wild-type GFP-TOC159GM in the respective overexpression lines. In addition, the TOC159-interacting TOC-complex core proteins TOC75 and À33 also accumulated to considerably lower levels in the GFP-TOC159GM-K/R:ppi2 line under low GA presumably to maintain complex stoichiometry ( Figure 3A, B). Typically, a small fraction of the SUMO target protein pool undergoes SUMOylation under a specific cellular condition (Johnson, 2004). The reduced levels under low GA of GFP-TOC159GM-K/R when compared to the wild-type protein were attributed to increased UPS-mediated protein degradation as both proteins recovered to the same protein levels in the presence of the proteasome inhibitor MG132 ( Figure 3D,E).
We conclude that SUMOylation partially stabilizes TOC159 against UPS-dependent degradation under specific conditions such as low GA during early development. The photosynthesis-associated proteins are considered the preferred transport cargoes of TOC159 because they fail to accumulate in the ppi2 mutant lacking TOC159. They tended to accumulate to a lesser degree in GFP-TOC159GM-K/R:ppi2 plants under low GA in the presence of PAC, but not in statistically significant fashion. The explanation may be that the levels of TOC159 are already low in the GFP-TOC159GM-K/R:ppi2 plants in the absence of PAC and allow only for relatively small changes in the accumulation of photosynthesis-associated proteins in the presence of PAC. Nevertheless, the results suggest that SUMOylation at the M-domain of TOC159 serves to fine tune preprotein import under low GA.
Based on the results, we propose a hypothetical model for the role of SUMOylation during early developmental stages, when environmental conditions are unfavorable and GA concentrations are low. As previously demonstrated, the chloroplast import receptor TOC159 is ubiquitylated prior to outer membrane insertion by an unknown E3-ligase other than SP1 and degraded via the UPS. In addition, its G-domain interacts with SUMO3, the physiological consequences of which remain unknown but may regulate interaction with DELLA system. The M-domain may be SUMOylated by SUMO3, protecting to some extent against UPS-dependent degradation. When the environmental conditions become more favorable, GA levels increase, and the GA receptor-DELLA complex is degraded by the UPS and TOC159 liberated for outer membrane insertion. The non-ubiquitinated TOC159 is assembled into the TOC complex thus allowing proplastids to differentiate into chloroplasts ( Figure 4). In the CHLORAD pathway, cytosolic Cdc48 extracts ubiquitinated TOC proteins from the outer membrane of the chloroplast (Ling et al., 2019;Shanmugabalaji and Kessler, 2019). The intriguing connection between the Cdc48 and SUMO pathways in chromatin dynamics (Franz et al., 2016) suggests that the SUMO pathway might also act on the CHLORAD pathway at specific developmental stage, but we currently do not offer evidence for such a mechanism. Our data specifically implicate SUMOylation and probably SUMO interaction in the control of proplastid to chloroplast transition by regulation of TOC159 levels in early plant development.  Figure 4. Hypothetical model for the SUMOylation-dependent fine-tuning of chloroplast biogenesis at the level of the TOC159 import receptor in early plant development. Environmental conditions influence the concentrations of gibberellic acid (GA) that accumulate upon seed imbibition. When active GA levels are reduced by stress (-GA; left-hand panel), DELLA (RGL2) accumulates and sequesters TOC159, which is then degraded via the UPS. In addition, TOC159 interacts and is also covalently SUMOylated by SUMO3. Covalent SUMOylation protects TOC159 against UPS-mediated degradation and supports the accumulation of photosynthesis-associated proteins in the chloroplast. Any unimported preproteins are degraded in the cytosol via the UPS. When GA concentrations increase (+GA, right-hand panel), the GA receptor complex binds to DELLA, which is degraded by the UPS. TOC159 is then free to assemble into the TOC complex. Protein import is thus enabled, allowing proplastids to differentiate into chloroplasts. Continued on next page

Plant transformation and transgenic lines
The K1370R point mutation was introduced into the binary construct pTOC159-GFP-TOC159GM using a site-directed mutagenesis kit (Agilent-QuikChangeII) with the primers TOC159S3F and TOC159S3R and resulted in pTOC159-GFP-TOC159GM-K/R. The pTOC159-GFP-TOC159GM-K/R construct was introduced into Agrobacterium tumefaciens (C58 strain) and stably transformed into heterozygous ppi2 plants using the floral dip method (Clough and Bent, 1998). Transformed plants were selected on phosphinothricin-containing medium, and lines homozygous for the transgene as well as the ppi2 mutation were isolated and named pTOC159:GFP-TOC159GM-K/R:ppi2 (referred to as GFP-TOC159GM-K/R:ppi2 plants).