Correlation of Autophagosome Formation with Degradation and Endocytosis Arabidopsis Regulator of G-Protein Signaling (RGS1) through ATG8a

Damaged or unwanted cellular proteins are degraded by either autophagy or the ubiquitin/proteasome pathway. In Arabidopsis thaliana, sensing of D-glucose is achieved by the heterotrimeric G protein complex and regulator of G-protein signaling 1 (AtRGS1). Here, we showed that starvation increases proteasome-independent AtRGS1 degradation, and it is correlated with increased autophagic flux. RGS1 promoted the production of autophagosomes and autophagic flux; RGS1-yellow fluorescent protein (YFP) was surrounded by vacuolar dye FM4-64 (red fluorescence). RGS1 and autophagosomes co-localized in the root cells of Arabidopsis and BY-2 cells. We demonstrated that the autophagosome marker ATG8a interacts with AtRGS1 and its shorter form with truncation of the seven transmembrane and RGS1 domains in planta. Altogether, our data indicated the correlation of autophagosome formation with degradation and endocytosis of AtRGS1 through ATG8a.


Introduction
Degradation of proteins in eukaryotic cells is mediated either through autophagy or the ubiquitin proteasome system (UPS). The UPS clears short-lived proteins, whereas autophagy degrades whole organelles and individual proteins that cannot be processed by the UPS [1]. Autophagy is a ubiquitous and highly conserved process, in which degrading misfolded and long-lived proteins, and damaged or old organelles are engulfed with a double membrane and then infused to the vesicle for lysosomal degradation, (or vacuole in plant cell). Stresses including starvation and oxidation, or stimulation by glucose and hormones accelerate autophagy processes [2][3][4][5]. Autophagic activity is low and balanced when cells contain sufficient nutrients, but increases rapidly when cells are starved. Cells survive starvation by recycling fatty acids and amino acids to meet the energy demand, and remove misfolded proteins and abnormal organelles [6]. The basic autophagy process is conserved among eukaryotes from yeast to animals and plants [7][8][9]. Several types of autophagy, including microautophagy [10], macroautophagy [9], chaperone-mediated autophagy [11], and organelle-specific autophagy, have been described in species [12]. Cell autophagy is normally maintained at a relatively low level, but increases abruptly when facing a disturbance. Autophagy preserves cell homeostasis and is one of the crucial safeguards of eukaryotic organisms.
The best indicator for autophagy is the formation of autophagosomes or double membrane vesicles that sequester other organelles or proteins. As the UPS and autophagy are both vital for Seven-day-old seedlings expressing GFP-ATG8a grown in ½ MS medium containing 1% sucrose for 7 days. (B) Seven-day-old seedlings grown as described for panel A except for use of sucrose-free medium for 2 days. (C) Seven-day-old seedlings grown as described for panel A remove sucrose being transferred to media with added 2% glucose for 1 h. Fluorescence from GFP-ATG8a visualized with a 488 nm excitation and 505-550 nm emission settings. A, B, C autophagosomes labeled by GFP-ATG8a in root cells incubation within 1 µM CA for 12 h. Scale bars = 50 µm. (D) Quantification of autophagosomes in root cells of four seedlings: "no S" means no starvation (A); "S 2d", starvation for 2 days (B); "G 1h", 2% glucose treatment for 1 h (C). (E) Seven-day-old seedlings grown in liquid ½ MS, and 1% sucrose under dim, continuous light. "Starved" indicates the seedlings treated with ½ MS media without sucrose for 2 h. Shown above is a typical Western blot probed with antiserum against GFP. Quantification of bands from replicate Western blots is shown below. MG-115, starvation plus 100 mM MG115 for 2 h CHX, starvation plus 70 µM CHX for 2 h; "S 2 h", starvation with no sucrose for 2 h as control. (F) Glucose indicated by "G" was added at the indicated times. β-actin was used as loading controls. Each value was the mean ± S.D. of four independent replicates. Asterisks indicate significant differences (* p < 0.05 or *** p < 0.001).

RGS1 Promoted the Production of Autophagic Flux and Autophagosomes
To determine the relationship between RGS1 and autophagosome and autophagic flux, we performed screening by crossing rgs1 and GFP-ATG8a and obtained homozygous seeds containing GFP-ATG8a in rgs1. Autophagic flux was measured by Western blotting using GFP antibodies in rgs1 (GFP-ATG8a staining) transgenic Arabidopsis. As shown in (Figure 2A,B) in GFP-ATG8a, compared with the control, the content of free GFP was lower in 1% glucose treatment for 6 and 12 h. In rgs1 (GFP-ATG8a staining), 1% glucose treatment was performed separately for 0, 6 and 12 h, and no significant change was observed in free GFP expression, suggesting that RGS1 promotes the production of autophagic flux. We further investigated the root cells of GFP-ATG8a transgenic (A) Seven-day-old seedlings expressing GFP-ATG8a grown in 1 2 MS medium containing 1% sucrose for 7 days. (B) Seven-day-old seedlings grown as described for panel A except for use of sucrose-free medium for 2 days. (C) Seven-day-old seedlings grown as described for panel A remove sucrose being transferred to media with added 2% glucose for 1 h. Fluorescence from GFP-ATG8a visualized with a 488 nm excitation and 505-550 nm emission settings. A, B, C autophagosomes labeled by GFP-ATG8a in root cells incubation within 1 µM CA for 12 h. Scale bars = 50 µm. (D) Quantification of autophagosomes in root cells of four seedlings: "no S" means no starvation (A); "S 2d", starvation for 2 days (B); "G 1h", 2% glucose treatment for 1 h (C). (E) Seven-day-old seedlings grown in liquid 1 2 MS, and 1% sucrose under dim, continuous light. "Starved" indicates the seedlings treated with 1 2 MS media without sucrose for 2 h. Shown above is a typical Western blot probed with antiserum against GFP. Quantification of bands from replicate Western blots is shown below. MG-115, starvation plus 100 mM MG115 for 2 h CHX, starvation plus 70 µM CHX for 2 h; "S 2 h", starvation with no sucrose for 2 h as control. (F) Glucose indicated by "G" was added at the indicated times. β-actin was used as loading controls. Each value was the mean ± S.D. of four independent replicates. Asterisks indicate significant differences (* p < 0.05 or *** p < 0.001).

RGS1 Promoted the Production of Autophagic Flux and Autophagosomes
To determine the relationship between RGS1 and autophagosome and autophagic flux, we performed screening by crossing rgs1 and GFP-ATG8a and obtained homozygous seeds containing GFP-ATG8a in rgs1. Autophagic flux was measured by Western blotting using GFP antibodies in rgs1 (GFP-ATG8a staining) transgenic Arabidopsis. As shown in (Figure 2A,B) in GFP-ATG8a, compared with the control, the content of free GFP was lower in 1% glucose treatment for 6 and 12 h. In rgs1 (GFP-ATG8a staining), 1% glucose treatment was performed separately for 0, 6 and 12 h, and no significant change was observed in free GFP expression, suggesting that RGS1 promotes the production of autophagic flux. We further investigated the root cells of GFP-ATG8a transgenic seedlings to observe whether autophagosomes were induced by 1% glucose treatment for 0.5 h in Arabidopsis thaliana. Therefore, the green punctate structures were examined using laser confocal scanning microscopy, before and after glucose treatment. Glucose led to dramatic changes, showing a significant increase in autophagosomes ( Figure 2C,E). However, in 1% glucose treatment for 0.5 h a small amount of autophagosomes were induced in rgs1 (GFP-ATG8a staining) ( Figure 2D,E). These results indicate that RGS1 promotes the formation of autophagosomes. seedlings to observe whether autophagosomes were induced by 1% glucose treatment for 0.5 h in Arabidopsis thaliana. Therefore, the green punctate structures were examined using laser confocal scanning microscopy, before and after glucose treatment. Glucose led to dramatic changes, showing a significant increase in autophagosomes ( Figure 2C,E). However, in 1% glucose treatment for 0.5 h a small amount of autophagosomes were induced in rgs1 (GFP-ATG8a staining) ( Figure 2D,E).
These results indicate that RGS1 promotes the formation of autophagosomes. Asterisks indicate significant differences of the starved seedlings treated with glucose from normal ones. Scale bars = 10 µm (** p < 0.01 or *** p < 0.001).

RGS1 Degradation is Correlated with Increased Autophagic Flux Independent of the Proteasome
Interestingly, we observed that the change in autophagic flux was due to degradation and re-feeding of RGS1 protein resulted in the opposite event. After starvation of sucrose for 2 h, the steady-state level of AtRGS1 decreased, and autophagic flux increased ( Figure 3A-C). Upon addition of glucose, the level of AtRGS1 recovered and autophagic flux returned to the baseline. Thus, alteration of RGS1 steady state levels was induced by nutrient starvation and addition of 1% glucose. The starvation-induced decrease of total AtRGS1 occurred in the presence of proteasome inhibitor MG-115 ( Figure 3D,E), suggesting that the decrease in AtRGS1 results from autophagy and not by proteasome. As MG-115 itself induces autophagy, the addition of this inhibitor further Figure 2. RGS1 promoted the production of autophagosomes and autophagic flux. Quantification of autophagic flux and GFP-ATG8a and rgs1 (GFP-ATG8a staining) seedlings. Seven-day-old seedlings of GFP-ATG8a treated by liquid MS medium without sugar following stimulation by 1% glucose for 0, 6, and 12 h. (A) Equal amounts of protein extracted from the seedlings were used in SDS-PAGE, followed by Western blotting with anti-GFP and anti-β-actin antibodies. (B) Quantification of changes in free GFP normalized with the expression of β-actin. Asterisks indicate significant differences compared to starved seedlings treated with 1% glucose for 0 h ** p < 0.01. Error bar represent S.D. obtained from three independent replicates. (C) Observation of autophagosomes. Autophagosomes labeled by GFP-ATG8a and rgs1 (GFP-ATG8a staining), in roots of GFP-ATG8a plants incubated with 1 µM CA for 12 h. Normal seedlings and GFP-ATG8a in seedlings treated by 1% glucose for 0.5 h in GFP-ATG8a. (D) Normal seedlings and 1% glucose seedlings for 0.5 h in rgs1 (GFP-ATG8a staining) seedlings treated by 1% glucose for 0.5 h. (E) Quantification of GFP-ATG8a and rgs1 (GFP-ATG8a staining), labeled autophagosomes per root cell at the indicated times were used to calculate autophagic activity. Mean and S.D. values were calculated from roots of six seedlings per time point. Results from three parallel experiments were used for quantification. Asterisks indicate significant differences of the starved seedlings treated with glucose from normal ones. Scale bars = 10 µm (** p < 0.01 or *** p < 0.001).

RGS1 Degradation is Correlated with Increased Autophagic Flux Independent of the Proteasome
Interestingly, we observed that the change in autophagic flux was due to degradation and re-feeding of RGS1 protein resulted in the opposite event. After starvation of sucrose for 2 h, the steady-state level of AtRGS1 decreased, and autophagic flux increased ( Figure 3A-C). Upon addition of glucose, the level of AtRGS1 recovered and autophagic flux returned to the baseline. Thus, alteration of RGS1 steady state levels was induced by nutrient starvation and addition of 1% glucose. The starvation-induced decrease of total AtRGS1 occurred in the presence of proteasome inhibitor MG-115 ( Figure 3D,E), suggesting that the decrease in AtRGS1 results from autophagy and not by proteasome. As MG-115 itself induces autophagy, the addition of this inhibitor further decreased the steady-state level of AtRGS1 in comparison with nutrient starved conditions. To verify whether AtRGS1 vesicles can fuse with autophagosomes containing ATG8a, we designed the following experiments to show the co-localization of RGS1 and autophagosomes.

Co-Localization of RGS1 and Autophagosomes
Seedlings of GFP-ATG8a and RGS1-red fluorescent protein (RFP) were observed for colocalization of autophagosomes with RGS1 in Arabidopsis root cells. In comparison with the autophagosome formation and AtRGS1 under normal growth conditions ( Figure 4A) ( Figure S1), after half an hour post-treatment with 6% glucose of cells, autophagosomes and AtRGS1 were found to overlap ( Figure 4B) ( Figure S1), indicating their co-localization was internalized by glucose induced endocytosis.
BY-2 cells are relatively homogenous and are suitable for studies of autophagy [34,35]. LysoTracker Red (LTR) dye is an effective autophagosome stain used to quantitate autophagy activity [36][37][38], and it can slightly stain Nicotiana tobaccos RGS1 (NtRGS1) in cytoplasm especially on the plasma membrane. The results showed that the punctate fluorescent signal of GFP-tagged tobacco RGS1 (NtRGS1-GFP) overlapped with the LTR punctate signals, with an estimate of ~35% co-localization of the two markers ( Figure 4C) ( Figure S2), whereas starvation increased co-localization to ~70% ( Figure 4D) ( Figure S2). This finding is consistent with results obtained by Hanamata et al. using the autophagic flux marker yellow fluorescent protein (YFP)-tagged NtATG8a for autophagosome formation [35]. We found that autophagosomes were increased by 0.5 h sucrose treatment after 2 days starvation and that a portion of RGS1-GFP was located in LTR punctate signals (autophagosome), implying that parts of RGS1-GFP were recycled in autophagosomes.
Starvation of BY2 cells increased the NtRGS1-GFP punctate signals and specific punctate structures stained with LTR ( Figure 4D). The increased co-localization suggests that starvation

Co-Localization of RGS1 and Autophagosomes
Seedlings of GFP-ATG8a and RGS1-red fluorescent protein (RFP) were observed for colocalization of autophagosomes with RGS1 in Arabidopsis root cells. In comparison with the autophagosome formation and AtRGS1 under normal growth conditions ( Figure 4A) ( Figure S1), after half an hour post-treatment with 6% glucose of cells, autophagosomes and AtRGS1 were found to overlap ( Figure 4B) ( Figure S1), indicating their co-localization was internalized by glucose induced endocytosis.
BY-2 cells are relatively homogenous and are suitable for studies of autophagy [34,35]. LysoTracker Red (LTR) dye is an effective autophagosome stain used to quantitate autophagy activity [36][37][38], and it can slightly stain Nicotiana tobaccos RGS1 (NtRGS1) in cytoplasm especially on the plasma membrane. The results showed that the punctate fluorescent signal of GFP-tagged tobacco RGS1 (NtRGS1-GFP) overlapped with the LTR punctate signals, with an estimate of~35% co-localization of the two markers ( Figure 4C) ( Figure S2), whereas starvation increased co-localization to~70% ( Figure 4D) ( Figure S2). This finding is consistent with results obtained by Hanamata et al. using the autophagic flux marker yellow fluorescent protein (YFP)-tagged NtATG8a for autophagosome formation [35]. We found that autophagosomes were increased by 0.5 h sucrose treatment after 2 days starvation and that a portion of RGS1-GFP was located in LTR punctate signals (autophagosome), implying that parts of RGS1-GFP were recycled in autophagosomes.
Starvation of BY2 cells increased the NtRGS1-GFP punctate signals and specific punctate structures stained with LTR ( Figure 4D). The increased co-localization suggests that starvation drives NtRGS1 into the autophagosome. Our BY-2 cellular experiment showed the significant increase in autophagosome after 3% sucrose treatment ( Figure 4E), consistent with induction of autophagosomes in Arabidopsis by glucose ( Figure 1C). We observed that NtRGS1 1-248 -GFP, NtRGS1 249-413 -GFP, and NtRGS1 249-459 -GFP in BY-2 cells overlapped with LTR punctate signals ( Figure 5), suggesting that co-localization of NtRGS1 truncations and autophagosomes occurred by 0.5 h sucrose induction.
FM4-64 dye is widely used in the study of plasma membrane and vesicles; it combines with plasma membrane and endometrial organelles to produce high-intensity fluorescence. This dye is used for the observation of autophagosomes, such as in tobacco cells, under sucrose starvation treatment, the flow of plasma membrane to the autophagosome membrane was observed by FM4-64 dye, and the central vacuole component was autophagosome, indicating that endocytosis and supply from central vacuoles may aid in the formation of autophagosomes [39]. Related studies have shown that accumulation of autophagosomes in Arabidopsis root cells and yeast cells can be detected using FM4-64 [40,41]. Here, we observed the confocal of Arabidopsis thaliana root cells treated with 3% glucose and stained with FM4-64 dye. The results of the FM4-64 dye encapsulating the yellow fluorescent RGS1-YFP illustrates the close relationship between autophagy and RGS1 (Supplementary Video). drives NtRGS1 into the autophagosome. Our BY-2 cellular experiment showed the significant increase in autophagosome after 3% sucrose treatment ( Figure 4E), consistent with induction of autophagosomes in Arabidopsis by glucose ( Figure 1C). We observed that NtRGS1 1-248 -GFP, NtRGS1 249-413 -GFP, and NtRGS1 249-459 -GFP in BY-2 cells overlapped with LTR punctate signals ( Figure 5), suggesting that co-localization of NtRGS1 truncations and autophagosomes occurred by 0.5 h sucrose induction. FM4-64 dye is widely used in the study of plasma membrane and vesicles; it combines with plasma membrane and endometrial organelles to produce high-intensity fluorescence. This dye is used for the observation of autophagosomes, such as in tobacco cells, under sucrose starvation treatment, the flow of plasma membrane to the autophagosome membrane was observed by FM4-64 dye, and the central vacuole component was autophagosome, indicating that endocytosis and supply from central vacuoles may aid in the formation of autophagosomes [39]. Related studies have shown that accumulation of autophagosomes in Arabidopsis root cells and yeast cells can be detected using FM4-64 [40,41]. Here, we observed the confocal of Arabidopsis thaliana root cells treated with 3% glucose and stained with FM4-64 dye. The results of the FM4-64 dye encapsulating the yellow fluorescent RGS1-YFP illustrates the close relationship between autophagy and RGS1 (Supplementary Video). showing numerous autophagosomes. Differential contrast microscopy (DIC), merged image of green (or yellow) and red. NtRGS1-GFP BY-2 cells were standardly cultured for 3 days, and then supplemented with 3% sucrose (control) or without sucrose (starvation). Con A (1 µM) and LTR stain were added for confocal imaging. LTR fluorescence appeared red, and RGS1-GFP appeared green. In the merged images, the overlap of LTR and RGS1 fluorescence appeared yellow. (C) (D) (E) Scale bar = 20 µm. Autophagosomes stained with LTR.

ATG8a Interacts with Full-Length and Truncated RGS1
Our previous paper showed that ATG2 and ATG5 inhibited AtRGS1 recovery after D-glucose treatment, whereas 1% glucose treatment induced ATG gene expression [3]. Both ATG2 and ATG5 Differential contrast microscopy (DIC), merged image of green (or yellow) and red. NtRGS1-GFP BY-2 cells were standardly cultured for 3 days, and then supplemented with 3% sucrose (control) or without sucrose (starvation). Con A (1 µM) and LTR stain were added for confocal imaging. LTR fluorescence appeared red, and RGS1-GFP appeared green. In the merged images, the overlap of LTR and RGS1 fluorescence appeared yellow. (C) (D) (E) Scale bar = 20 µm. Autophagosomes stained with LTR.

ATG8a Interacts with Full-Length and Truncated RGS1
Our previous paper showed that ATG2 and ATG5 inhibited AtRGS1 recovery after D-glucose treatment, whereas 1% glucose treatment induced ATG gene expression [3]. Both ATG2 and ATG5 play important roles in autophagosome formation [3,42]. AtRGS1-YFP is located on the plasma membrane in WT and atg2 and atg5 mutant plants ( Figure 6A,D). After addition of 6% glucose for 0.5 h, autophagy inhibitor 3-methyladenine (3-MA) affected the subcellular localization of AtRGS1-YFP ( Figure 6B,E), as shown by an increase in AtRGS1-YFP vesicles. This result suggests the internal location of AtRGS1-YFP in the plasma membrane. In the absence of 3-MA, the number of AtRGS1-YFP vesicle was reduced in atg2 and atg5 mutants compared with the WT (Figure 6C,F), implying that sequestration of autophagy-related proteins (at least ATG2 and ATG5) participates in AtRGS1 metabolism. As AtRGS1 endocytosis is a type of stimulation, we would assume that ATG2 and ATG5 might provoke the stimulation of RGS1 endocytosis. To further confirm that the autophagosome marker GFP-ATG8a interacts with RGS1 and its truncations, we carried out bimolecular fluorescence complementation (BiFC) and pull-down experiments.    In the BiFC experiment ( Figure 6G), fluorescence was complemented in the cells expressing AtRGS1-cYFP, which was tagged with the autophagy protein nYFP-ATG8a, but not in cells expressing the negative control P31-nYFP. ATG8 is a ubiquitin-like protein involved in cargo recruitment and biogenesis of autophagosomes. Autophagosome size is determined by the amount of ATG8. As ATG8 is selectively enclosed by autophagosomes, its breakdown would allow measurement of the autophagic rate [30].

Discussion
Throughout the eukaryotic kingdom, RGS proteins act as negative regulators in G protein signaling. In animal cells, RGS proteins participate in cellular processes, including cell growth, mitosis, neuron signaling, membrane diffusion, embryo development, and inflammatory and neurodegenerative diseases [20,[43][44][45]. In humans and nematodes, RGS proteins are involved in almost all signaling transmission and adjustment processes. AtRGS1 serves as an important sensor for glucose in Arabidopsis [19,20]. The study of the relationship between plant cell autophagy and RGS1 is instrumental to our understanding of the role of RGS1 in sugar signaling.
Our study showed the metabolism of RGS1 is negatively correlated with autophagic flux (Figure 3A,C). Autophagic flux and autophagosomes of rgs1 mutant express all rarely processing at different times. (Figure 2). In addition, we noted that in co-localization of AtRGS1 and ATG8a after glucose-induced ( Figure 4B) co-localization of NtRGS1 and autophagosomes (with LTR stain) under

Discussion
Throughout the eukaryotic kingdom, RGS proteins act as negative regulators in G protein signaling. In animal cells, RGS proteins participate in cellular processes, including cell growth, mitosis, neuron signaling, membrane diffusion, embryo development, and inflammatory and neurodegenerative diseases [20,[43][44][45]. In humans and nematodes, RGS proteins are involved in almost all signaling transmission and adjustment processes. AtRGS1 serves as an important sensor for glucose in Arabidopsis [19,20]. The study of the relationship between plant cell autophagy and RGS1 is instrumental to our understanding of the role of RGS1 in sugar signaling.
Our study showed the metabolism of RGS1 is negatively correlated with autophagic flux (Figure 3A,C). Autophagic flux and autophagosomes of rgs1 mutant express all rarely processing at different times. (Figure 2). In addition, we noted that in co-localization of AtRGS1 and ATG8a after glucose-induced ( Figure 4B) co-localization of NtRGS1 and autophagosomes (with LTR stain) under normal conditions, starvation treatment and sucrose recovery ( Figure 4C-E), the rate of co-localization under starvation treatment was higher than that under normal conditions, that is, sucrose induced more autophagosomes (Figure 4C,E). We further confirmed that ATG8a can interact with RGS1 and its truncations (Figure 7). FM4-64 as a marker of endocytosis is able to trace the formation of autophagosomes, with RGS1-YFP being wrapped by this dye (Supplementary Video). These results suggest that autophagy may participate in RGS1 degradation during starvation and RGS1 endocytosis by inducement with glucose. Autophagosomes are induced by treatment with glucose ( Figure 1C,D) and sucrose ( Figure 4E) at the early phase stage. We speculate that exuberant metabolism may be associated with increased autophagosomes. Xiong et al. observed that 30 mM (about 0.54%) glucose treatment inhibited autophagy through inducing activity of protein kinase target of rapamycin (TOR) [46]. RGS1 promotes ATG expression after glucose induction [3]. Our results show that 1-6% glucose treatment in the early phase promoted autophagosome formation ( Figure 1C) and negatively regulated autophagic flux in Arabidopsis ( Figure 2B). The 3% sucrose treatment induced the production of autophagosomes in BY2 cells after starvation ( Figure 4E). We used mannitol as a control treatment and observed that autophagosomes treated by 3% glucose were induced by glucose rather than osmotic stress ( Figure S3).

Plant Materials and Growth Conditions
All Arabidopsis lines were of the Columbia ecotype. Seed surface was sterilized with 70% ethanol for 10 min, 95% ethanol for 10 min, and finally washed with water. Seeds were vernalized at 4 • C for 2 days in 1 2 X Murashige and Skoog (MS) liquid medium supplemented with 1% sucrose. A total of 100 Arabidopsis seed were grown in 100 mL liquid medium in a 250 mL flask with rotary shaking at 140 RPM under dim continuous light (40 µE m −2 ·s −1 ) at 23 • C for 7 days.
After sterile rinsing thrice with water and starvation using 1 2 MS lacking sucrose for 2 h, the seedlings were transferred to 1 2 MS with 100 mM MG115 or 70 µM CHX in continuous dim light for 6 h, whereas the same medium with 1% glucose was used as control. All experiments were performed at least thrice. For immunoblot analyses, seedlings were harvested and flash frozen in liquid nitrogen.
WT Nicotiana benthamiana plants were used for BiFC and pull-down experiments. Plants were grown at 23 • C and 70% relative humidity under a 16 h light/8 h dark cycle for 1-1.5 months before infiltration. After infiltration, plants were kept under the same growth conditions.

Arabidopsis Thaliana Mutants and Transgenic Lines
atg2 (SALK_076727), and atg5 (SALK_020601) were obtained from the Arabidopsis Biological Resource Center. AtRGS1 (encoding amino acids 1-459) was subcloned to pEarleyGate205 (C-terminal TAP). The 35S::AtRGS1-YFP construct was transformed into Agrobacterium EHA105, which was then used to transform WT, atg2, and atg5 by the floral-dip method [47], homozygous lines of transgenic plants were used in this study.
The coding regions of RGS1 were amplified with TaKaRa Ex Taq DNA Polymerase (Fisher Scientific, R001A) using specific primers containing Gateway attB sites and then cloned into pDONR207 ( Figure S4) entry vector using the BP Clonase II (Life Technologies) to create RGS1 entry clones. After verification by sequencing, each clone was mobilized using the LR Clonase II (Life Technologies) into the Gateway destination vector pK7RWG2 ( Figure S5). The primers used for pENTR clones in autophagy-related gene. The 35S::AtRGS1-RFP construct were transformed into Agrobacterium EHA105, which was then used to transform the transgenic Arabidopsis expressing GFP-ATG8a (Dr. Faqiang Li providing) by the floral-dip method [47]; homozygous line RGS1-RFP/GFP-ATG8a of transgenic plants was used in this study ( Figure S1A).
Homozygous rgs1 (GFP-ATG8a staining) was successfully crossed through crossing rgs1 and GFP-ATG8a. In general, first cross rgs1 and GFP-ATG8a and the next generation were screened in Basta-resistant medium to confirm the cross success (rgs1 is a hybrid, confocal observation was performed to select the strong seedlings and transfer them to the soil, and obtain seeds). Then, the materials collected from each plant were grown on 10 µg/mL Basta-resistant medium and the seedlings on all the grown plates were selected. Confocal observation confirmed a relatively strong GFP fluorescence, the seedlings were transferred to the soil, and each strain was confirmed to be a rgs1 mutant ( Figure S6). In theory, the seedlings were homozygous rgs1 (stain GFP-ATG8a), and the next generation confirmed rgs1 (stain GFP-ATG8a) as homozygous.

AtRGS1-YFP Internalization Analysis
Fluorescence quantification for AtRGS1-YFP internalization was performed as described by Urano et al. [48] and Fu et al. [49]. WT, atg2, and atg5 seedlings (7 days old) were treated with 0% or 6% D-glucose (w/v) for 30 min. Root epidermal cells located 2-4 mm below the cotyledon were imaged (Z stacks obtained) using a Zeiss LSM710 confocal laser scanning microscope equipped with a 20 × Plan-NeoFluor numerical aperture (N.A. = 0.5) objective and a 40 × C-Apochromat (N.A. = 1.2) water immersion objective. YFP fluorescence was excited by a 514 nm argon laser and detected at 526-569 nm by a photomultiplier detector. At least 10 sets of images from seven seedlings were obtained for internalization quantification analysis by ImageJ software.

BiFC
BiFC was pe work of Klopffleisch et al. [50]. The coding regions of ATG8a were amplified with TaKaRa Ex Taq DNA Polymerase (Fisher Scientific, RR001A, Invitrogen, Waltham, MA, USA) using specific primers containing Gateway attB sites and then cloned into the pDONR207 entry vector ( Figure S4) using the BP Clonase II (rformed as described in the Life Technologies, Invitrogen) to create ATG8a entry clones. After verification by sequencing, each clone was mobilized using the LR Clonase II (Life Technologies) into the Gateway destination vector pCL112_JO ( Figure S7) for BiFC. The primers used for pENTR clones in autophagy-related gene ATG8a are listed as follows: GGGGACAAGTTTGTACAAAAAAGCAGGCTTCATGATCTTTGCTTGCTTGAAATT and GGGGACCACTTTGTACAAGAAACTGGGTCTCAAGCAACGGTAAGAGATCCAAAAGT. The open reading frame of AtRGS1 in BiFC vectors was as previously described by Grigston et al. [51].
Agrobacterium tumefaciens strain EHA105 was cultured on Luria Bertani (LB) medium for 2 days, and a single colony was inoculated into 5 mL LB medium supplemented with the appropriate antibiotics (spectinomycin) and grown at 28 • C in a shaker for 48 h. The culture was transferred to the infiltration buffer with 10 mM 2-(N-morpholine)-ethanesulfonic acid (MES; pH 5.6) and 40 µM acetosyringone (1:100 ratio, v/v) for growth at 28 • C for 16 h. When growth reached A 600 = 3.0, the bacteria were spun down gently (3200g, 10 min), and the pellets were resuspended in 10 mM MgCl 2 at a final A 600 = 1.5 (A 600 = 1 for p19). A final 150 µM acetosyringone was added and the bacteria were kept at room temperature for at least 4 h without shaking.
Leaf disks were obtained from infiltration sites and imaged by confocal microscopy as previously described [48]. Tobacco leaf epidermal cells were imaged using a Zeiss LSM710 confocal laser scanning microscope equipped with an Apochromat 40× water immersion objective (N.A. = 1.2). YFP fluorescence was excited by a 514 nm argon laser and detected at 526-569 nm by a photomultiplier detector, and Mt-rk and RFP fluorescence was excited by a 543nm HeNe laser and detected at 565-621 nm.

Transformation of BY-2 Cells
BY-2 cells (Nicotiana tabacum "Bright Yellow") were cultured in modified MS medium supplemented with 3% (w/v) sucrose, 1 µg/mL thiamine-HCl, 0.2 µg/mL 2,4-D, 100 µg/mL myo-inositol, and 200 µg/mL KH 2 PO 4 , with a final pH 5.8 adjusted by KOH. Cell lines were cultured in either liquid MS medium with continuous shaking, or in the form of calli on MS media solidified with 0.8% (w/v) agar in the dark at 26 • C. Suspensions were sub-cultured every 7 days by transferring 1.5 mL culture into 30 mL fresh MS medium; calli were sub-cultured every 3-4 weeks. NtRGS1 was expressed under 35S promoter in our constructs. The tobacco homolog of AtRGS1, the NtRGS1 gene was amplified from BY-2 cDNA and cloned into pDrive cloning vector with 12 bp of Kozak sequence to achieve seamLess expression. The fragment was subsequently cloned into the pGreen binary vector using BamHI and HindIII to create C-terminal fusion with enhanced GFP. Partial digestion was required for this procedure.
Exponential cell suspension (3-4 days after sub-culturing) was filtered and resuspended in 30 mL of fresh MS medium. Acetosyringone (15 mL, 40 mM) was added to the suspension and thoroughly mixed by pipetting. Total of 3 mL Agrobacterium suspension was then added to the cell suspension and cultivated for 3 days in the dark at 26 • C. The cells were washed with 300 mL 3% (w/v) sucrose and 100 mL MS medium supplemented with 100 mg/L cefotaxime. Finally, the cells were resuspended in 2-3 mL liquid MS medium containing 100 mg/L cefotaxime and 50 mg/L kanamycin and cultured in a Petri dish for 3-4 weeks in the dark at 26 • C. The calli cultures were transferred onto fresh MS medium with the same antibiotics. Grown cells (3 d sub-culturing) were centrifuged at 5000g for 10 min, rinsed thrice with sucrose-free BY-2 medium and subsequently starved in MS medium for 2 days. BY-2 cells grown in the MS medium supplemented with 3% sucrose served as control.
CA (1 µM) and LTR (1 µM) Red (Invitrogen) were added into BY-2 media for 12 and 3 h, respectively, before confocal imaging. GFP fluorescence was excited by a 488 nm argon laser and detected at 505-550 nm by a photomultiplier detector, and LTR fluorescence was excited by a 543 nm HeNe laser and detected at 565-621 nm. At least 10 sets of images were obtained thrice BY-2 cells for quantification analysis. GFP or LTR punctae were counted per cell according to 10 sets of images field of vision.

Pull Down Assays
pENTR clones in ATG8a were mobilized using the LR Clonase II (Life Technologies) into the Gateway destination vector pDEST17 ( Figure S9) for His-ATG8a. His-ATG8a was expressed in E. coli at