Abstract
While ATG8ylation, the lipidation of ATG8-family proteins, is canonically linked to double-membrane autophagosome formation, emerging studies demonstrate its non-canonical association with single-membrane organelles. The functional significance of ATG8ylation in these compartments, however, remains unclear. Here, we demonstrate that ionophores rapidly trigger ATG8 conjugation to the vacuolar membrane (tonoplast), a process reliant on the ATG conjugation system rather than the upstream autophagic regulators. Inhibiting reactive oxygen species (ROS) generation or V-ATPase function greatly impedes the targeting of ATG8 to the tonoplast. Intriguingly, the attachment of ATG8 to the tonoplast enhances its invagination and fosters the formation of intraluminal vesicles within vacuoles, which is achieved independently of the ESCRT machinery or cytoskeletal components. The emergence of ATG8-positive vesicles may facilitate the restoration of vacuolar acidification by redirecting proton flow from the vacuole-to-cytoplasm to an intravacuolar direction, which aids in the rapid recovery of plant growth after removal of monensin. Furthermore, under alkaline stress, ATG8 targets the tonoplast and induces vacuolar membrane invagination via a regulatory mechanism similar to that of monensin, indicating that ATG8ylation-mediated vacuolar remodeling represents an adaptive mechanism against environmental alkalinization in plants.
Similar content being viewed by others
Introduction
Autophagy is an evolutionarily conserved mechanism that removes damaged organelles and protein aggregates to maintain cellular homeostasis. Over the past few decades, ground-breaking studies have uncovered autophagy-related (ATG) genes and molecular details of autophagy1,2. Among them, the ATG8-family proteins serve as critical components of the autophagic machinery, exerting influence in numerous facets, including cargo selection, phagophore expansion, autophagosome closure, and fusion with vacuole/lysosome3,4. Interestingly, emerging lines of evidence demonstrate that ATG8 can also be conjugated with various single-membrane structures, such as the Golgi apparatus5,6, phagosome7,8,9, endosome10, and lysosome11, leading to alternative pathways collectively known as non-canonical autophagy. ATG8ylation has recently garnered attention as a putative membrane stress signal12, displaying broad functional significance in the immune response7,8,13, antigen presentation14, cancer, and neurodegeneration15. In plants, despite recent studies have revealed the translocation of ATG8 to swollen Golgi cisternae to aid its reassembly after heat stress5,16, our understanding as to non-canonical autophagy and the roles of ATG8ylation in the pathway remains limited.
Recent investigations from non-plant systems have made significant advances in elucidating the regulatory mechanisms of non-canonical autophagy6,7,8. Generally, the vacuolar-type ATPase (V-ATPase) has been found to directly recruit ATG16L1 (ATG16 in plants) to mediate lipidation of ATG8-family proteins to single membrane structures6,7,8,17. Deletion of the C-terminal WD40 domain or a single point mutation (K490A) in ATG16L1 impairs its functionality specifically in non-canonical autophagy, while exerting no influence on macroautophagy17,18. In agreement with this, the inhibition of V-ATPase using Bafilomycin A1 disrupts the conjugation of ATG8 to single membrane structures. Thus, the V-ATPase-ATG16L1 axis is characterized as a universal mechanism governing non-canonical autophagy8,18,19. V-ATPase is a highly conserved proton pump among eukaryotes responsible for acidification in various endomembrane compartments such as endosomes and vacuoles/lysosomes20. Isoforms of the subunit a in the membrane-integral V0 subcomplex dictate the diverse subcellular localization of the V-ATPase. For example, VHA-a1 targets the V-ATPase to the trans-Golgi network (TGN)/early endosomes, while VHA-a2 and VHA-a3 are localized to the tonoplast in Arabidopsis21. Currently, the role of V-ATPase in regulating non-canonical autophagy in plants remains largely unexplored.
One of the critical functions of ATG8 is the recognition and binding of cargoes/receptors, primarily facilitated through the interaction of ATG8-interacting motifs (AIMs) with the AIMs docking site (ADS) on ATG8 proteins22. From the topological orientation, the incorporation of ATG8 proteins on the single-membrane vesicles seems to take place solely on their side facing the cytosol. In this context, it is speculated that ATG8 on these single-membrane structures is unlikely to sequester cargoes from the cytosol as it does in canonical autophagosomes23. At present, the specific functions associated with the non-canonical conjugation of ATG8 have not been clearly established. In this study, we demonstrated that ionophores rapidly induce the translocation of ATG8 to the tonoplast and markedly enhance its invagination. By redirecting proton transport from the vacuole-to-cytoplasm to an intravacuolar direction, the formation of ATG8-positive vesicles in the vacuole may assist in the resumption of vacuolar acidification. On the other hand, ATG8 works in concert with ATG2 to recruit the endoplasmic reticulum (ER) to the vicinity of the tonoplast, potentially facilitating vacuolar membrane repair by mediating the transfer of lipids. Our research unveiled previously uncharacterized functionalities of ATG8 in alleviating vacuolar damage by influencing membrane curvature, providing a model for further investigation into the diverse roles of ATG8 in single membrane structures in plants.
Results
Ionophores induce conjugation of ATG8 to tonoplast
Previously, we found that ATG8 proteins concentrate to swollen Golgi stacks following acute heat stress5. This finding prompted us to investigate whether the disruption of the Golgi could induce a similar response. We conducted a screening of several chemicals, including brefeldin A, concanamycin A (ConcA), and monensin, which are known to disrupt the Golgi structure24,25. Treatment with each of the three drugs resulted in the aggregation of the cis-Golgi marker GFP-SYP32, yet none induced the translocation of mCherry-ATG8f to the Golgi apparatus (Supplementary Fig. 1). Interestingly, monensin treatment caused mCherry-ATG8f to display a membranous pattern (Supplementary Fig. 1).
We then examined the pattern changes of ATG8 in response to other carboxylic ionophores (i.e., nigericin and salinomycin). These ionophores share similar molecular structures (Fig. 1a), and their carboxyl and hydroxyl groups facilitate the formation of electrically neutral zwitterionic complexes with cations26. The resulting chelate complexes exhibit lipid solubility, enabling their diffusion across biological membranes for mediating ion exchange26. Indeed, localization of GFP-ATG8a on membranous structures matching those after monensin treatment was also observed by incubation with either nigericin or salinomycin (Fig. 1b). Afterward, we focused on monensin to explore how its varying concentrations impact the subcellular distribution of ATG8. It was found that YFP-ATG8e exhibited such membranous distribution at monensin concentrations exceeding 10 μM (Fig. 1c). The time-lapse confocal imaging showed that treatment with monensin rapidly induced subcellular alterations in GFP-ATG8a, with obvious membrane-like signal observed within 20 min (Fig. 1d). Moreover, we assessed the subcellular localization variations among different ATG8 isoforms (ATG8a to ATG8i) (Supplementary Fig. 2a, b), and all ATG8 isoforms displayed a comparable response to monensin, forming similar membrane-like structures (Supplementary Fig. 2c).
a An overview of the chemical structures of monensin sodium, nigericin sodium, and salinomycin sodium. b The formation of membrane-like structures of GFP-ATG8a in response to three ionophores. Scale bar, 20 μm. c The impact of different concentrations of monensin on the subcellular localization of YFP-ATG8e. 5-day-old YFP-ATG8e transgenic seedlings were treated with different concentrations of monensin (0, 5, 10, 20, 40 μM) in 1/2MS liquid medium for 1 h, followed by confocal microscopy observation. Scale bar, 20 μm. d Time series analysis of the dynamics of GFP-ATG8a in response to monensin (20 μM). Time was presented in minutes. Scale bar, 10 μm. e–i Colocalization analyses between ATG8 and the tonoplast marker VAMP711-mCherry (e), the late endosome marker Rab7-GFP (f), the early endosome marker YFP-ARA7 (g), the TGN marker VHA-a1-RFP (h), as well as the cis-Golgi marker GFP-SYP32 (i). Scale bars in e–i, 20 μm. j Quantification of the colocalization ratios shown in (e–i). The Pearson correlation (PSC) coefficient was analyzed by ImageJ with the PSC colocalization plugin. The data represent means ± SD. n = 6 confocal images (112.5 µm × 112.5 µm) of individual roots. Significance analysis using unpaired two-sided Student’s ttest. Similar confocal imaging results were obtained from at least six individual roots, with three replicates. Source data are provided as a Source Data file.
To figure out the subcellular location of ATG8 following monensin treatment, we generated double transgenic lines of ATG8 with different organelle markers. Upon monensin treatment, the fluorescent signals of ATG8 exhibited good colocalization with Vamp711-mCherry (tonoplast) as well as Rab7-GFP (late endosome/tonoplast), while no significant colocalization was observed with other endomembrane markers, including YFP-ARA7 (early endosome/multivesicular body, EE/MVB), VHA-a1-RFP (TGN) and GFP-SYP32 (cis-Golgi) (Fig. 1e–j). It is noteworthy that monensin treatment caused intracellular aggregations of both VHA-a1-RFP and GFP-SYP32 (Fig. 1h, i). Such aggregates were further characterized as clusters of dilated vesicles derived from the Golgi/TGN complexes by transmission electron microscopy (TEM) (Supplementary Fig. 3). These vesicle clusters were specifically labeled with 10 nm nanogold particles conjugated to anti-GFP antibodies, which target the Golgi-localized Man1-GFP marker (Supplementary Fig. 3c, d). While MVB morphology was unaffected by monensin treatment (Supplementary Fig. 3e), their proximity to swollen Golgi vesicles correlated with the organelle clustering observed in confocal microscopy (Fig. 1g). Noteworthily, unlike plant cells challenged by acute heat stress, no ATG8 was associated with the swollen Golgi membrane5. These results indicated that ATG8 selectively incorporated into the vacuolar membrane following treatment with ionophores.
ATG8 targeted to tonoplast relies on the ATG conjugation system
In canonical autophagy, several core complexes such as the ATG1 complex, ATG9 vesicles, ATG2 complex, phosphatidylinositol 3-kinase (PI3K) complex, and ATG conjugation systems synergistically regulate the formation of autophagosomes (Fig. 2a). To test whether these upstream regulators were involved in the conjugation of ATG8 to tonoplast, we evaluated the subcellular location of GFP-ATG8a in atg1abct, atg11-1, atg9-4, and atg2-1 mutants. Unexpectedly, after treatment with monensin, an obvious vacuolar membrane signal can still be observed in these mutant lines (Fig. 2b). Similarly, inhibition of PI3K activity with wortmannin did not prevent monensin-induced translocation of ATG8 to vacuolar membranes (Fig. 2b). Therefore, the core upstream autophagy regulators, including ATG1, ATG11, ATG9, ATG2, and PI3K, appear to be dispensable for the targeting of ATG8 to the tonoplast.
a Schematic depiction of the core protein complexes that regulate the formation of autophagosomes, encompassing the ATG1 complex, ATG9 vesicles, PI3K complex, ATG2-ATG18 complex, and ATG conjugation systems. b The impact of monensin on the subcellular localization of GFP-ATG8a in atg1abct, atg11-1, atg9-4 and atg2-1 mutants, as well as inhibition of PI3K activity with wortmannin. c Analysis of the impact of ATG4 on the subcellular localization of YFP-ATG8a. d Effect of mutation in ATG5, ATG7, and ATG16 genes on the subcellular localization of ATG8 under monensin treatment. Five-day-old GFP-ATG8a/atg1abct, GFP-ATG8a/atg11-1, GFP-ATG8a/atg9-4, GFP-ATG8a/atg5-1, YFP-ATG8e/atg7-2 and YFP-ATG8a(G132A) x mCherry-ATG8f double transgenic seedlings were treated with 20 μM monensin for 1 h, followed by confocal microscopy observation. Wortmannin (16.5 μM) was pre-incubated for 10 min and then added with 20 μM monensin for 1 h, followed by confocal imaging. Scale bars in (b–d), 20 μm. Similar confocal imaging results were obtained in at least six individual roots with three replicates. e–h The impact of monensin on ATG8 lipidation in wild-type Col-0 and autophagy mutants including atg11-1, atg7-2, atg5-1 and atg16-c1. The ratios (n = 3 biological replicates) between the lipidation form of ATG8 and actin was quantified with ImageJ (f and h). An asterisk indicates an unknown band. Mon, Monensin. Data are mean ± SD. Significance analysis using an unpaired two-sided Student’s ttest. The immunoblotting assays were independently replicated three times with consistent results. Source data are provided as a Source Data file.
The C-terminal cleavage of ATG8 at a conserved glycine residue by the cysteine protease ATG4 is a prerequisite for lipidation (Fig. 2a and Supplementary Fig. 2b)27. Using YFP-ATG8a(G132A) x mCherry-ATG8f double transgenic plants, we analyzed the impact of the point mutation of the ATG4 recognition site on ATG8 localization. Notably, in contrast to mCherry-ATG8f, the tonoplast localization of YFP-ATG8a(G132A) was completely abolished with monensin treatment (Fig. 2c), indicating that ATG4 plays a crucial role in the conjugation of ATG8 to the tonoplast.
Next, we evaluated alterations in the localization of ATG8 in mutants defective in ATG conjugation systems, specifically atg5-1, atg7-2, and atg16-c1. Monensin treatment did not affect the cytosolic distribution of ATG8 in atg5-1, atg7-2, or atg16-c1 mutants (Fig. 2d). Furthermore, the application of monensin significantly enhanced the lipidation level of ATG8 in the Col-0 and atg11-1 mutants, while no such increase was observed in atg5-1, atg7-2, and atg16-c1 mutants (Fig. 2e–h). Collectively, these results indicate that the ATG conjugation system is critical for the targeting of ATG8 to vacuolar membranes.
ROS and V-ATPase are required for ATG8 translocation to tonoplast
Previous studies have shown that cells accumulate a significant amount of ROS under monensin treatment28. To investigate the potential involvement of ROS in monensin-induced targeting of ATG8 to the tonoplast, we assessed ROS generation using a cell-permeant indicator H2DCFDA. Accompanied by the translocation of mCherry-ATG8f to the vacuolar membrane, a robust burst of ROS was observed in the root cells (Fig. 3a). Notably, the exogenous application of ascorbic acid (AsA), which eliminates ROS29, resulted in a great reduction in ATG8 binding to the tonoplast followed by monensin treatment (Fig. 3b). Interestingly, pretreatment with diphenyleneiodonium (DPI), an inhibitor of NADPH oxidase30, markedly attenuated the recruitment of GFP-ATG8a to the tonoplast (Fig. 3c), indicating the requirement of NADPH oxidase-derived ROS for ATG8 conjugation to the tonoplast upon monensin treatment.
a Detection of ROS generation after monensin treatment. 5-day-old mCherry-ATG8f seedlings were subjected to monensin treatment at a concentration of 20 μM. The ROS-sensitive dye H2DCFDA (1 μM) was incubated 10 min prior to confocal imaging. The images were shown with LUT pseudocolor scale (Rainbow RGB). Scale bar, 50 μm. b The ROS scavenger ascorbic acid (AsA) reduced GFP-ATG8a response to monensin. The left schematic diagram illustrates the timing of the addition of AsA and the subsequent monensin treatment. Scale bar, 20 μm. c Treatment with diphenyleneiodonium chloride (DPI) inhibited GFP-ATG8a translocation to the tonoplast. The schematic diagram on the left illustrates the timing of DPI addition and the subsequent monensin treatment. Scale bar, 20 μm. d Pretreatment of ConcA inhibited the translocation of GFP-ATG8a to the tonoplast. ConcA was added 10 min prior to monensin treatment. Scale bar, 20 μm. e Western blotting analysis of the effect of ConcA on the ATG8 lipidation. The immunoblotting assays were independently replicated three times with consistent results. f Statistical analysis of the ratio of lipidation form ATG8 to ATG8 in (e). Ratio values were calculated from n = 3 biologically independent experiments. Data are mean ± SD. Significance analysis using an unpaired two-sided Student’s ttest. g A schematic diagram showing the structure of V-ATPase. The subunit a isoforms VHA-a2 and VHA-a3 are specifically located at vacuolar membranes. h Monensin-induced the translocation of GFP-ATG8a to the tonoplast was abolished in the vha-a2vha-a3 double mutant. Scale bar, 20 μm. Similar confocal imaging results were obtained in at least six individual roots with three replicates. Source data are provided as a Source Data file.
In mammalian cells, V-ATPase is considered a universal regulator for the conjugation of ATG8-family proteins to single membranes8,19. To test whether the V-ATPase was also involved in ATG8 targeting to the tonoplast, we firstly investigated the subcellular localization of GFP-ATG8a in the presence of ConcA, a specific inhibitor of V-ATPase31. As shown in Fig. 3d, the localization of GFP-ATG8a to vacuolar membranes was completely blocked by ConcA. Correspondingly, the treatment with ConcA resulted in a significant reduction in the lipidation form of ATG8 following monensin treatment (Fig. 3e, f). Next, ATG8 targeting to tonoplast was studied in the vha-a2 vha-a3 double mutant, which lacks the tonoplast-localized isoforms of the V-ATPase subunit VHA-a (Fig. 3g)32,33. Under normal physiological conditions, a large amount of mCherry-ATG8f punctate structures were observed within the vacuoles of the mature root cells of vha-a2 vha-a3 mutants (Supplementary Fig. 4a). This phenomenon is likely attributed to vacuolar alkalization resulting from the deficiency of V-ATPase32, which subsequently leads to reduced vacuolar hydrolase activities. Noteworthily, in comparison to wild-type Col-0 plants, the vha-a2 vha-a3 mutants failed to exhibit the conjugation of mCherry-ATG8f to vacuolar membranes after treatment with monensin (Fig. 3h). Taken together, these results demonstrated the essential requirement of both NADPH oxidase-derived ROS and V-ATPase for the incorporation of ATG8 into vacuolar membranes, resembling the mechanisms observed in non-canonical autophagy processes in mammalian cells8.
ATG8ylation enhances the invagination of vacuolar membranes
To gain insights into the potential roles of ATG8 in tonoplast dynamics, we examined vacuolar morphology using GFP-ATG8a x Vamp711-mCherry double transgenic plants. After monensin treatment, massive vesicles labeled with both GFP-ATG8a and Vamp711-mCherry appeared in the vacuoles, some of which were attached to the vacuolar membrane, while others were randomly distributed within the vacuolar lumen (Fig. 4a). To further obtain the ultrastructure of the vacuoles, we subjected seedlings treated with monensin to high-pressure freeze fixation for TEM analysis. Interestingly, after monensin treatment, numerous micrometer-sized vesicles containing cytoplasmic components accumulated inside the vacuole (Fig. 4b). Further analysis using 3D electron tomography revealed that these vesicles, formed by invaginations of the vacuolar membrane, not only exhibited spherical shapes but also included some large tubular structures (Fig. 4c and Supplementary Movie 1). It is worth mentioning that some of these vesicles also contained organelles such as mitochondria and ER (Fig. 4c). In addition, the time-lapse confocal imaging further showed that the vacuolar membrane was highly dynamic, with masses of invagination occurring (Fig. 4d and Supplementary Movie 2). An unexpected observation was that some ATG8-conjugated vacuolar membranes, which appeared to be ruptured, were capable of curving back to create intraluminal vesicles (Fig. 4d and Supplementary Movie 2). By contrast, inhibition of GFP-ATG8a targeting to vacuolar membranes by application with ConcA obviously decreased the vacuolar invaginations (Fig. 4e). Consistently, a reduction in vacuolar membrane invaginations was also observed in atg5-1 mutant (Fig. 4f and Supplementary Movie 3), indicating that the incorporation of ATG8 facilitates the invagination of the tonoplast.
a 3D projection of GFP-ATG8a and tonoplast marker Vamp711-mCherry after monensin treatment. Arrowheads indicated representative intralumenal vesicles. A plot profile analysis of the overlapping signals between GFP-ATG8a and Vamp711-mCherry was shown on the right. Scale bar, 20 μm. b Representative electron microscopic images illustrated the morphological changes in the vacuole following a 1 h treatment with monensin. Invaginating vesicles were denoted by magenta triangles. Similar TEM results were obtained in six individual blocks with three replicates. Scale bars, 1 μm. c Electron tomography analysis of the vacuolar morphology after monensin treatment for 1 h. The 3D model reconstructed from the tomography was presented on the right. Magenta triangles in individual tomogram slices indicate invaginating vesicles. In the 3D model, red color indicates mitochondria, yellow color indicates ER, and blue color indicates small vesicles. The 3D tomography was repeated three times independently, with similar results. Scale bars, 1 μm. d Time-series analysis of the ATG8-positive vesicles invaginated from the vacuolar membrane. A representative invaginating vesicle was indicated by the arrowhead. The magenta arrow pointed to a ruptured membrane structure that curved back to form a vesicle. Scale bar, 10 μm. e Time-series analysis of the effect of ConcA on ATG8-positive vesicles invaginated from the vacuolar membrane. Scale bar, 10 μm. f Time-series analysis of the vacuolar membrane invagination in atg5-1 mutant upon monensin treatment. Scale bar, 10 μm. Similar confocal imaging results in (a, d–f) were obtained in at least six individual roots with three replicates. Source data are provided as a Source Data file.
Formation of ATG8-positive invaginated vesicles is independent of ESCRT and cytoskeleton
It is well-characterized that the endosomal sorting complexes required for transport (ESCRT) play crucial roles in membrane curvature and the final fission process34. To determine whether ESCRT was involved in formatting ATG8-positive intraluminal vesicles, we investigated the subcellular location of FYVE domain protein required for endosomal sorting 1 (FREE1), a plant-specific ESCRT component that regulates the closure of autophagosomes via direct interaction with ATG84. Intriguingly, GFP-FREE1 did not localize to the vacuolar membrane as clearly as mCherry-ATG8f after treatment with monensin (Fig. 5a). To analysis the impact of monensin on the vacuolar morphology in free1 mutants, we introduced YFP-ATG8e into heterozygous T-DNA insertional mutants free1(+/−) and screened the albino seedlings (i.e., the homozygous free1(−/−)34) to conduct subsequent confocal analyzes. The time-lapse imaging showed that some small vesicles were still generated and closely associated with the vacuolar membrane in the free1(-/-) mutant (Fig. 5b and Supplementary Movie 4). To further ascertain the role of ESCRT in tonoplast invagination, two ESCRT-III dominant-negative plants expressing SNF7.1(L22W) and VPS4(E232Q) were used to assess intraluminal vesicle formation. Under the induction of dexamethasone (DEX), both transgenic plants showed chlorosis similar to that of free1(-/-) mutant (Fig. 5c), and time-lapse imaging confirmed the formation of ATG8-positive vesicles in the vacuoles after monensin treatment (Fig. 5d–f and Supplementary Movie 5). Importantly, under electron microscopy, we noted the development of invaginated vesicles inside the vacuoles of free1(-/-) mutants and DEX:VPS4(E232Q) transgenic plants following monensin treatment (Fig. 5g). Collectively, these results indicated that the ATG8-mediated invagination of the vacuolar membrane may operate independently of ESCRT functionality.
a Analysis of the subcellular location of GFP-FREE1 before and after monensin treatment. Scale bar, 20 μm. b Time-series analysis of the invagination of ATG8-positive vesicles in free1(-/-) mutant. The magenta arrows indicated a vesicle that adhered inside the vacuolar membrane; white arrows indicated invaginated vesicles. Scale bar, 10 μm. c Confirmation of the validity of the dexamethasone (DEX)-inducible dominant-negative mutants, DEX:SNF7.1(L32W) and DEX:VPS4(E232Q). The transgenic seeds were germinated on agar plates containing 10 μM DEX. Representative images were captured after 5 days of growth under light. Scale bar, 1 cm. d A schematic diagram showing the treatment flow of DEX-inducible dominant-negative mutants. e, f Time-series analysis of the invaginated vesicles in YFP-ATG8e/DEX:SNF7.1(L32W) (e) and YFP-ATG8e/DEX:VPS4(E232Q) (f) plants after DEX induction. The magenta arrows indicated vesicles that adhered inside the tonoplast. Scale bars, 20 μm. g TEM analysis of vacuolar morphology in free1(-/-) mutant and DEX:VPS4(E232Q) transgenic plants. The magenta arrows indicate invaginating vesicles. Scale bars, 1 μm. h, i Analysis of the effects of cytoskeletal disruption on the formation of ATG8-positive vesicles. Microtubules were visualized using a mCherry-TUA5 fusion protein (h), while microfilaments were labeled with ABD2-GFP (i). Oryzalin (10 μM) and latrunculin B (10 μM) were employed to disrupt microtubule and microfilament structures, respectively. The invaginated vesicles were denoted by yellow arrowheads. Scale bars, 20 μm. j Confocal time-lapse analysis the effects of microfilament disruption on the dynamics of ATG8-positive vesicle invagination. White arrowheads indicated the invaginated vesicles. Scale bar, 20 μm. Similar confocal imaging results were obtained in at least six individual roots with three replicates.
The organization of the cytoskeleton, particularly actin filaments, within cells can significantly influence vacuole dynamics and morphology35. To investigate the role of the cytoskeleton in ATG8-mediated invagination processes, we generated double-transgenic plants expressing ATG8 with microtubule marker mCherry-TUA5 and actin marker ABD2-GFP, respectively. In contrast to microtubules, actin filaments were found to have a closer association with the vacuolar membrane (Fig. 5h, i and Supplementary Movie 6). Disrupting microtubules with oryzalin or actin filaments with latrunculin B did not affect the formation of monensin-induced ATG8-positive invaginated vesicles (Fig. 5h, i). Confocal time-lapse analysis further revealed that the disassembly of actin filaments inhibited the overall movement of the vacuole, but did not affect the invagination process (Fig. 5j and Supplementary Movie 7). These findings indicate that ATG8-mediated invagination is independent of microtubules or actin filaments.
ATG8ylation induces redistribution of the ER to the tonoplast
From the electron tomography results, we noted that some organelles, such as the ER, were encapsulated within ATG8-positive invaginated vesicles (Fig. 4c). This observation prompted us to investigate whether the ER serves as a cargo delivered to the vacuoles. We developed double transgenic plants expressing mCherry-ATG8f and CNX-GFP, and monitored the localization changes of the ER in response to monensin treatment. Strikingly, after monensin treatment, the fluorescence signal of CNX-GFP colocalized with mCherry-ATG8f at the tonoplast (Fig. 6a). Notably, no ER fluorescence signal was detected around the ATG8-positive invaginated vesicles (Fig. 6a), suggesting that invagination and recruitment of the ER might be two separate events. To obtain a more precise understanding of the spatial relationship between the ER and vacuoles at the ultrastructural level, we conducted serial section electron tomography. The 3D reconstruction revealed a very clear reticular structure of the ER in close proximity to the tonoplast (Fig. 6b, and Supplementary Movie 8). Subsequently, to explore whether inhibiting the translocation of ATG8 to the vacuolar membrane would affect the redistribution of the ER, we constructed double transgenic plants expressing CNX-RFP and YFP-ATG8e in the atg7-2 mutant background. During monensin treatment, the absence of CNX-RFP signal around the vacuoles, where mCherry-ATG8f was localized (Fig. 6c), revealed that ATG8 plays a crucial role in the recruitment of the ER to the vicinity of the vacuolar membrane.
a Monensin treatment induced the redistribution of the ER to the vicinity of the vacuolar membrane. A plot profile analysis of the co-localization between CNX-GFP and mCherry-ATG8f was presented, with the white dashed line in the inset 1 indicating the region of interest. The arrow in inset 2 indicates an ATG8-positive invaginated vesicle. Scale bar, 20 μm. b 3D electron tomography analysis of the structural organization between the ER and vacuole. Magenta arrowheads in the individual slices (N = 60 and 110) indicated the ER. 3D models reconstructed from the tomography were presented on the right. The blue arrows indicated the invaginating vesicles. Scale bars, 1 μm. c Impact of ATG7 mutation on ER localization following monensin treatment. Co-localization analysis of CNX-RFP and YFP-ATG8e in the atg7-2 mutant background is presented on the right. d Monensin treatment induces the co-localization of YFP-ATG2 with mCherry-ATG8f at the vacuolar membrane. e The impact of ATG2 mutation on ER localization following monensin treatment. f Time-lapse analysis of ATG8-positive invaginated vesicle formation in atg2-1 mutant background following monensin treatment. White arrowheads indicated ATG8-positive invaginated vesicles. Scale bars in (c–f), 20 μm. Similar confocal imaging results were obtained in at least six individual roots with three replicates. Source data are provided as a Source Data file.
In canonical autophagy, lipids transfer from the ER by ATG2 is involved in the expansion of autophagosomes36,37. To determine the role of ATG2 in the relocalization of the ER to the tonoplast, we first examined the localization changes of ATG2 following monensin treatment. We observed that the fluorescence signal of YFP-ATG2 colocalized with mCherry-ATG8f at the tonoplast after monensin treatment (Fig. 6d). Importantly, monensin failed to induce the colocalization of CNX-RFP with GFP-ATG8a in the atg2-1 mutant (Fig. 6e), indicating the essential function of ATG2 in the recruitment of the ER by ATG8 to the vacuolar membrane. Notably, confocal time-lapse imaging showed that the mutation of ATG2 did not affect the formation of ATG8-positive invaginated vesicles (Fig. 6f ), further suggesting that ATG8ylation-mediated invagination and the recruitment of the ER to the vacuolar membrane are two separate processes. The latter may involve lipid transfer, potentially playing a role in the repair of the vacuolar membrane.
ATG8ylation facilitates the restoration of vacuolar acidification
Given that ATG8ylation possessed the aforementioned dual functions, we then wondered their potential to reshape the physiological functions of plant vacuoles. Initially, we detected the distribution of ATG8 on the vacuolar membrane after washing away monensin. The GFP-ATG8a fluorescence signal started to detach from the vacuolar membrane, presenting a diffuse distribution in the cytoplasm, with the signal becoming nearly undetectable on the tonoplast after recovery for 3 h (Fig. 7a). This result indicated that monensin-triggered ATG8 relocation to the vacuolar membrane is reversible. As an ionophore, monensin promotes the exchange of metal cations (such as Na+ and K+) with H+ across the membrane, resulting in the disruption of the proton gradient on the vacuolar membrane26. Consistent with the help of the Na+-specific fluorophore CoroNa Green, a noticeable enhancement in Na+ fluorescence within the vacuole can be detected (Fig. 7b). We then analyzed the vacuolar acidification recovery in wild-type Col-0 and atg5-1 mutant plants following washing out monensin. The atg5-1 mutants exhibited a markedly slower recovery of vacuolar acidity compared to Col-0 (Fig. 7c, d). Furthermore, short-term treatment with monensin resulted in the suppression of primary root growth to differing extents in the Col-0, atg11-1, and atg5-1 plants (Fig. 7e–g). After 2 days of recovery from the removal of monensin, atg11-1 mutants exhibited comparable root elongation to Col-0 (Fig. 7e, f), while atg5-1 mutants showed significant growth retardation (Fig. 7g, h). Since the mutation of ATG5 inhibited the binding of ATG8 to the vacuolar membrane (Fig. 2d), the above results imply that ATG8ylation plays a positive role in the rapid recovery of vacuolar acidity and the resumption of normal plant growth after the removal of monensin.
a Monensin-induced ATG8 binding to the vacuolar membrane is reversible. Scale bar, 20 μm. b Monensin induces ion exchange across the vacuolar membrane. The fluorescent dye CoroNa Green AM (1 μM) was used for staining Na+. Scale bar, 20 μm. c Analysis of vacuolar acidification dynamics in Col-0 and atg5-1 roots during the recovery phase after monensin treatment. Vacuolar acidification was measured by detecting the ratio of BCECF fluorescence. Scale bar, 20 μm. d Quantitative statistical analysis of the ratio (488/458) of BCECF fluorescence intensities. The ratio of BCECF fluorescence intensities from ten individual roots (n = 10) was quantified with ImageJ. Data are mean ± SD. Significance analysis using an unpaired two-sided Student’s t-test. e–h Analysis of the effect of monensin treatment on the root elongation of Col-0, atg11-1, and atg5-1 mutants. Five-day-old seedlings were treated with 20 μM monensin for 2 h and then transferred to fresh 1/2MS agar plates. Images (e and g) were captured at 0 and 2 days after transfer to the plate. Statistical analysis of root elongation (f and h) was performed by quantifying six roots per genotype (n = 6) using ImageJ. Scale bar, 10 mm. i Analysis the subcellular localization of ATG8 in response to alkaline stress induced by (NH₄)₂CO₃. Five-day-old seedlings were immersed in 10 mM (NH₄)₂CO₃ solution (pH 8.5) for 1 h, followed by confocal imaging. Magenta arrows indicate the rupture of an invaginated vesicle, white arrows mark the complete vesicle formation from a detached membrane. Scale bar, 10 μm. j Comparative analysis of alkaline stress sensitivity in Col-0, atg11-1, and atg5-1 mutants. Five-day-old seedlings were treated with 10 mM (NH₄)₂CO₃ solution (pH 8.5) for 12 h, then transferred to 1/2 MS liquid medium for 5-day recovery. Scale bar, 10 mm. k Statistical analysis of chlorophyll content in (j). Three biological replicates (n = 3) were used for quantification. Data are mean ± SD. Significance analysis using an unpaired two-sided Student’s ttest. Source data are provided as a Source Data file.
The ability of monensin to alkalinize vacuoles motivated us to test whether environmentally relevant alkaline stress would trigger ATG8 recruitment to the vacuolar membrane. Intriguingly, ammonium carbonate ((NH4)2CO3)-induced alkalinization (pH 8.5) robustly promoted GFP-ATG8a localization to the tonoplast and stimulated vacuolar membrane invagination vesicle formation (Fig. 7i). Mechanistically paralleling monensin, this process required the ATG conjugation system but was independent of canonical upstream autophagy regulators, including ATG9 and ATG11 (Supplementary Fig. 5a). Furthermore, both ROS and V-ATPase activity were also essential for alkaline stress-triggered ATG8ylation on the tonoplast (Supplementary Fig. 5b, c). Notably, atg5-1 mutants exhibited heightened alkaline stress sensitivity compared to Col-0 and atg11-1, displaying accelerated leaf chlorosis and reduced chlorophyll content (Fig. 7j, k). Collectively, these results suggest that ATG8ylation-mediated vacuolar membrane remodeling represents a physiologically adaptive mechanism under alkaline environmental conditions.
Discussion
In the process of canonical autophagy, ATG8 performs diverse and essential functions, encompassing cargo recognition and binding, membrane tethering and extension, as well as autophagosome closure and maturation2,4,8. However, recent investigations have unveiled an intriguing aspect wherein ATG8 demonstrates the ability to associate with single-membrane structures independently of the classical upstream regulatory factors involved in autophagy5,7,8,9,11,18. Recruitment of ATG8 onto single-membrane vesicles is generally believed to result in eventual degradation23. Currently, the specific roles played by ATG8 in these single-membrane structures remain largely unknown. In this study, we have elucidated that the conjugation of ATG8 to vacuolar membranes, induced by ionophores, actively promoted invagination and the formation of intraluminal vesicles. These vesicles may potentially redirect the proton transfer that typically occurs from the vacuole to the cytoplasm, turning it into an internal vacuolar event, which could expedite the restoration of vacuolar acidity (Fig. 8).
Alkaline stress or ionophores (e.g., monensin) trigger NADPH oxidase-dependent ROS production, which may act as signaling molecules mediating ATG8ylation on vacuolar membranes. ATG8 conjugates to the vacuolar membrane via the V-ATPase-ATG16 axis, serving dual functions: (1) ATG8ylation enhances vacuolar membrane invagination. The formation of invaginated vesicles may help to sequester monensin-induced proton exchange within the vacuole, thereby preventing cytoplasmic proton influx and facilitating rapid restoration of vacuolar acidity; (2) ATG8 mediates ER recruitment to the vacuolar membrane, where ATG8 may engage ATG2 to facilitate lipid transfer from the ER, ensuing in vacuolar membrane repair.
In mammalian cells, the V-ATPase is generally considered to be a universal regulator of non-canonical autophagy8,19. It recruits ATG16L1 to assemble into a functional E3 complex (known as the ATG12-ATG5-ATG16L1 complex), which catalyzes the lipidation of ATG8-family proteins on single membranes. Consistent, our pharmacological and genetic studies demonstrated that inhibiting V-ATPase significantly impeded the targeting of ATG8 to the tonoplast (Fig. 3d, h). This finding suggested that plants share a common mechanism with animals for inducing non-canonical autophagy. In Arabidopsis, the V-ATPase exhibits dual subcellular localization, which is regulated by distinct isoforms of the subunit a. VHA-a1 directs it to the TGN/EE compartments, while VHA-a2 and VHA-a3 are predominantly localized to the tonoplast20,21. Notably, monensin treatment resulted in the disruption of TGN/EE morphology, leading to the formation of vesicular structures with a bubble-like appearance (Fig. 1h and Supplementary Fig. 3b). However, this cellular perturbation did not elicit the translocation of ATG8 to the swollen TGN/EE membrane (Fig. 1h). Conversely, our previous investigations unveiled that extremely high temperatures can evoke the translocation of ATG8 to the dilated Golgi apparatus rather than the vacuolar membrane5. Thus, although both the TGN and tonoplast harbor V-ATPase, they are possibly governed by discrete regulatory mechanisms. Such regulatory diversification may facilitate the selective targeting of ATG8 towards distinct single-membrane organelles under disparate environmental stimulus conditions.
Membrane curvature exerts a substantial influence on the structural configuration of the cell membrane, contributing significantly to its overall shape and organization. In vitro reconstitution investigations employing purified proteins and synthetic giant unilamellar vesicles have provided valuable insights into the multifaceted membrane-associated functionalities facilitated by ATG8 conjugation, such as tethering, hemi-fusion, tubulation, perturbation, and in/out-bud38,39,40,41. The direction of membrane curvature induced by the covalent anchorage of ATG8 is predominantly determined by the difference in the membrane area between the outer and inner layers of the lipid bilayer38,39. It is noteworthy that ATG8 conjugation is inadequate to initiate membrane invagination. Instead, it relies on the recruitment of other proteins within the ATG conjugation system, including ATG3, ATG7, and ATG12-ATG5-ATG16, to provide the necessary driving force for the induction of in-bud formation38. However, ATG16, ATG5, and ATG7 lacked tonoplast enrichment in our transgenic lines (Supplementary Fig. 6a), in striking contrast to ATG8’s pronounced tonoplast localization under identical conditions. Since the ATG12-ATG5-ATG16 E3 ligase complex is essential for ATG8 lipidation (Fig. 2), we proposed that this complex may execute its catalytic activity in the membrane-proximal cytosol and dissociate from the vacuolar membrane upon ATG8ylation completion. This hypothesis is supported by faint ATG5-GFP signals detected near the vacuolar membrane during early monensin treatment stages (Supplementary Fig. 6b).
Recent findings demonstrate that ATG8 can directly interact with the ESCRT component FREE1 to participate in the closure of autophagosomes4. However, in the presence of monensin, FREE1 was not recruited to the vacuolar membrane by ATG8, and the formation of invaginated vesicles was still observed in the vacuoles of free1(-/-) mutant and two ESCRT-III dominant negative lines DEX:SNF7.1(L32W) and DEX:VPS4(E232Q) (Fig. 5b–g). Therefore, we speculate that ESCRT may not directly participate in ATG8-mediated tonoplast invagination. Nevertheless, it is noteworthy that the intraluminal vesicles in the free1(-/-) mutant exhibited a prolonged adherence to the vacuolar membrane (Fig. 5b). This suggested a potential reliance on ESCRT-mediated scission processes for the detachment of invaginated vesicles from the vacuolar membrane. Undoubtedly, further experimental evidences are required to substantiate this hypothesis.
ATG8 shares structural and modification similarities with ubiquitin, as both molecules undergo sequential enzymatic reactions involving E1, E2, and E3 enzymes. However, their respective targets differ, with ubiquitin primarily marking proteins and ATG8 specifically targeting membrane lipids. Like the role of ubiquitination as a general signal for protein degradation, mounting evidence suggests that ATG8ylation also serves as a signaling mechanism in response to membrane stress events12. Ionophores disrupt membrane ion permeability, resulting in the inability of the vacuole to maintain a normal proton gradient26. In this context, ATG8-mediated vacuolar membrane invagination may redirect proton transport from the vacuole-to-cytoplasm to an intravacuolar direction (Fig. 8). The formation of invaginated vesicles were proposed to function as transient proton reservoirs, and subsequent vesicle rupture, as showed in Fig. 7i, allows these sequestered protons to be released back into the vacuolar lumen, which likely explains why Col-0 exhibits faster pH recovery after monensin removal compared to atg5-1.
Building upon the mechanistic link wherein monensin pharmacologically induces vacuolar alkalinization, we demonstrated that environmentally relevant alkaline stress analogously drives ATG8 recruitment to the tonoplast through conserved regulatory machinery (Fig. 7i). The heightened alkaline stress sensitivity of atg5-1 mutants compared to atg11-1 mutants (Fig. 7j), genetically dissects the canonical autophagy from ATG8lylation’s contribution to restoring vacuolar function. Notably, during the revision of this manuscript, Julian et al. (2023) reported that cell wall damage triggers turgor pressure-dependent ATG8ylation on tonoplast, which safeguards vacuolar integrity during such stress42. Thus, ATG8ylation can be considered a self-preservation mechanism by which the vacuole responds to the disruption of proton gradients caused by environmental stress.
In collaboration with ATG2, ATG8 facilitates ER recruitment to the vacuolar membrane (Fig. 6), a mechanism that parallels observations in mammalian systems11. However, a key divergence exists in which ATG8ylation predominantly facilitates tonoplast invagination (Fig. 4), contrasting with the lysosomal tubulation observed in animals. The opposing membrane curvature directionality (inward vs. outward remodeling) may suggest distinct regulatory paradigms. We proposed that ATG8ylation possesses an intrinsic ability to reverse membrane curvature, as evidenced by the bending of ruptured vacuolar membranes into spherical structures within vacuoles (Figs. 4d, 7i). In addition, as a lipid transfer protein, ATG2 probably aids in the transportation of lipids from the ER to the vacuolar membrane, counteracting the membrane impairment caused by monensin (Fig. 8). In future, it will be of interest to assess and quantify the damages inflicted on vacuolar membrane lipids, which would contribute to better understanding of the functional consequences of ATG8ylation.
Collectively, our results presented a non-canonical autophagic function of ATG8 in the field of plants. Under ionophores or alkaline stress treatment, ATG8 conjugation to tonoplast is rely on the ATG conjugation system rather than the upstream autophagic regulators. The association of ATG8 exerts a pronounced influence on membrane curvature, actively promoting invagination processes and facilitating the subsequent development of intraluminal vesicles within the vacuoles. This study broadened the scope of understanding regarding the diverse functions of the core autophagy protein ATG8 in plant cells, extending beyond its canonical role in autophagy.
Methods
Plant materials and growth conditions
Arabidopsis thaliana wild-type (Col-0) and mutant lines derived from this ecotype were used throughout the study. The autophagy mutants, including atg5-143, atg7-2 43, atg16-c15, and atg11-144, as well as the vha-a2 vha-a333 double mutants, were described previously. The single transgenic plants GFP-ATG8a45, mCherry-ATG8f46, YFP-ATG8b/c/d/e/g/h/i5, YFP-ATG165, ATG5-GFP5, GFP-ATG8a/atg1abct45, GFP-ATG8a/atg5-15, YFP-ATG8e/atg7-25, GFP-ATG8a/atg9-445, GFP-ATG8a/atg11-144, GFP-ATG8a/atg2-145, GFP-ATG8a/atg16-c15 and GFP-FREE134 were reported previously. The double transgenic plants YFP-ATG8a(G132A) x mCherry-ATG8f[ 5, GFP-ATG8a x Vamp711-mCherry5, Rab7-GFP x mCherry-ATG8f[ 5, YFP-ARA7 x mCherry-ATG8f[ 5, GFP-ATG8a x VHA-a1-RFP5, GFP-SYP32 x mCherry-ATG8f[ 5, YFP-ATG8e/DEX: VPS4(E232Q)4, YFP-ATG8e/DEX: SNF7.1(L32W)4 have been previously described. The GFP-FREE1 x mCherry-ATG8f, mCherry-TUA5 x GFP-ATG8a, ABD2-GFP x mCherry-ATG8f, YFP-ATG2 x mCherry-ATG8f, mCherry-ATG8f x CNX-GFP, YFP-ATG8e/atg7-2 x CNX-RFP/atg7-2, GFP-ATG8a/atg2-1 x CNX-RFP/atg2-1, mCherry-ATG8f/vha-a2 vha-a3 and Vamp711-mCherry/atg5-1 were obtained by cross-pollination, and the homozygous mutant backgrounds were verified by PCR (Supplementary Fig. 4b). YFP-ATG2, ATG7-GFP were cloned into the pCAMBIA1300 vector, and the transgenic plants were generated through floral dip. The seeds were surface-sterilized with 70% (v/v) ethanol containing 0.05% Triton X-100 and then sown on 1/2 Murashige and Skoog (MS) plates. After being kept at 4 °C for 48 h, the plates were transferred to a culture chamber maintained at 22 °C with a photoperiod of 16 h light and 8 h dark, and cultivate for 5 d for subsequent experiments.
Chemical treatment
The ionophores, including monensin sodium salt (MCE, #HY-N0150), salinomycin sodium salt (MCE, #HY-17439), and nigericin sodium salt (MCE, #HY-100381), were prepared as 10 mM stock solutions in ethanol and stored at − 20 °C. The 5-day-old seedlings were immersed in liquid 1/2 MS medium containing 0.2% ethanol (control) or 20 μM monensin for 1 h, and then the plant materials were subjected to confocal imaging. For alkaline stress treatment, 5-day-old seedlings were placed in 10 mM (NH4)2CO3 (Sangon, #A600058) solution (pH 8.5) for 1 h, followed by confocal imaging acquisition. The treatment of salinomycin and nigericin is the same as monensin. ConcA (MCE, #HY-N1724) and wortmannin (MCE, #HY-10197), oryzalin (MCE, #HY-147092), and latrunculin B (MCE, #HY-101848) stock solutions were prepared at concentrations of 1, 16.5, 10, and 10 mM, respectively, in DMSO and stored at − 20 °C. ConcA (1 μM), wortmannin (16.5 μM), oryzalin (10 μM), or latrunculin B (10 μM) was pre-incubated for 10 min and then added with 20 μM monensin for 1 h, followed by confocal imaging. For the ROS detection, a 1 μM concentration of the ROS-sensitive dye H2DCFDA (MCE, #HY-D0940) was incubated in the darkness for 10 min prior to confocal imaging. L-Ascorbic acid sodium salt (MCE, #HY-B0166A) was freshly prepared in 1/2 MS medium and added 1 h before monensin treatment. DPI (MCE, #HY-100965) was prepared as 10 mM stock solutions in DMSO and added 0.5 h before monensin treatment. CoroNa Green AM (Invitrogen, #C36676) and BCECF AM (Invitrogen, #B1170) were prepared as a 1 mM stock solution in DMSO and diluted to a final working concentration of 1 µM.
ATG8 lipidation assay
Briefly, 0.2 g of 5-day-old Col-0, atg5-1, atg7-2, atg11-1, and atg16-c1 seedlings, with or without 20 μM monensin treatment for 1 h, were grounded thoroughly in pre-cooled mortar with 1.5 ml of membrane buffer (40 mM HEPES, 1 mM EDTA, 10 mM KCl, 0.4 M Sucrose, pH 7.4) on ice. The crude lysates were collected in 2 ml tubes, then centrifuged at 1000 x g for 10 min to remove large cell fragments. The supernatants were transferred and mixed with loading buffer (250 mM Tris-HCI, pH 6.8, 10% (w/v) SDS, 0.5%(w/v) Bromophenol blue, 50% (v/v) Glycerol, 5% (v/v) β-Mercaptoethanol), and boiled at 95 oC for 10 min. The protein solution was subjected to 12% SDS-PAGE and immunoblotted with anti-ATG8 antibody (Agrisera, #AS14 2769). Uncropped western blots are available in supplementary files (Supplementary Fig. 7).
TEM analysis
The TEM assay was performed following our previously established protocols47,48. Briefly, 5-d-old seedlings were germinated on a 1/2 MS plate and then treated with or without 20 μM monensin in liquid 1/2 MS before dissecting. For high-pressure freezing, the root tips were collected and immediately frozen with a high-pressure freezer (EM ICE, Leica). For freeze substitution, the root tips were substituted with 2% osmium tetroxide in anhydrous acetone and maintained at − 80 °C for 24 h using an AFS2 temperature-controlling system (Leica). Subsequently, the samples were subjected to three washes with precooled acetone and gradually warmed to room temperature over a period of 60 h. Infiltration with increasing concentrations of EPON resin mix (50% Epon resin monomer, 15% dodecenyl succinic anhydride, and 35% nadic methyl anhydride) was carried out at room temperature. The root tips were then transferred into tin foil molds and polymerized by curing at 60 °C for 2 days. The embedded samples were sectioned into 90 nm-thick slices using an ultramicrotome (Leica UC7). Micrographs were acquired using a transmission electron microscope (Hitachi H-7650) operating at 80 kV, coupled with a charge-coupled device (CCD) camera.
3D electron tomography
Electron tomography was conducted using a 200 kV Tecnai F20 electron microscope (FEI Company) following previously established procedures49. Briefly, the tilt images were obtained from 250 nm-thick sections across a range of − 60° to 60°, with 1.5° increments, while the grid was rotated by 90° for the collection of the other axis of the tilt image stack. Dual-axis tomograms were generated by utilizing pairs of image stacks with the etomo program of the IMOD software (v.4.11.25). The contours of vacuolar membranes and intraluminal vesicles were manually delineated and subsequently meshed using the 3Dmod program within the IMOD software suite.
Confocal imaging and image processing
The confocal images were acquired using the Zeiss LSM880 laser scanning confocal system with 63X/1.4 NA or 40X/1.4 NA oil objective. The excitation and emission wavelengths for YFP, GFP, CoroNa Green, and H2DCFDA were 488 nm and 500–550 nm, respectively. For mCherry and RFP, the excitation and emission wavelengths were 561 nm and 570–650 nm, respectively. For dual-channel scanning, the “line” scanning mode was used. The images from different channels were exported separately using ZEN2.5 (blue edition) for further analysis. The co-localization analysis was performed using the PSC plugin in Image J software (NIH).
Vacuolar acidification analysis
BCECF fluorescence detection of vacuolar acidification was performed as described previously32. Roots of monensin-treated/untreated seedlings (20 μM, 1 h) were incubated with BCECF-AM (1 μM, Invitrogen #B1170) under light-protected conditions (aluminum foil wrapping) for 30 min, followed by a 5-min wash in dye-free 1/2 MS medium. Recovery periods (0/1/2/3 h) were synchronized across biological replicates. BCECF was excited at two wavelengths, 458 nm and 488 nm, and the emitted fluorescence was captured within the range of 500–550 nm. The vacuolar acidification was measured by calculating the ratio of the fluorescence intensities at these two excitation wavelengths (488/458).
Phenotypic Assays
For root elongation assay, five-day-old wild-type Col-0, atg11-1, and atg5-1 seedlings were placed in a liquid 1/2MS solution containing 20 µM monensin for 2 h. Next, the seedlings were gently rinsed three times with a newly prepared 1/2MS solution, with each rinse lasting for 5 min. Subsequently, the seedlings were transferred onto 1/2 MS solid media and vertically cultured in a growth chamber. The root tip position was marked with a marker pen, and images were taken to record the roots at 0- and 2-d after the transfer. Using the Image J software (NIH), the freehand line tool was employed to trace the outline of the root elongation zone, and the length of each root was measured.
For analysis of plant growth under alkaline stress, Col-0, atg11-1, and atg5-1 seedlings cultured in 1/2MS liquid medium for 5 d were placed in 10 mM (NH4)2CO3 solution (pH 8.5) for 12 h, and then transferred to freshly prepared 1/2MS liquid medium for recovery culture for 5 days. The chlorophyll content was measured spectrophotometrically following extraction in 90% ethanol at 65 °C for 2 h43.
Quantification and statistical analysis
All experiments were repeated at least three times with consistent results. The unprocessed western blots and DNA gel were given in Supplementary Fig. 7. The co-localization ratio and western band intensities were quantified using Image J software (NIH). Charting and statistical analysis were performed using GraphPad Prism 8 software. The P-values were determined with two-tailed unpaired Student’s t tests, and the asterisks represent significance levels (ns, not significant; *P < 0.05; ***P < 0.001).
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The data that support the results in this study are available in the article and its Supplementary Information files. The data supporting this research are also available from the corresponding author without any restrictions. Source data are provided in this paper.
References
Noda, N. N. & Inagaki, F. Mechanisms of autophagy. Annu. Rev. Biophys. 44, 101–122 (2015).
Li, H. et al. Shedding light on the role of phosphorylation in plant autophagy. FEBS Lett. 596, 2172–2185 (2022).
Johansen, T. & Lamark, T. Selective autophagy: ATG8 family proteins, LIR motifs and cargo receptors. J. Mol. Biol. 432, 80–103 (2020).
Zeng, Y. et al. The plant unique ESCRT component FREE1 regulates autophagosome closure. Nat. Commun. 14, 1768 (2023).
Zhou, J. et al. A non-canonical role of ATG8 in Golgi recovery from heat stress in plants. Nat. Plants 9, 749–765 (2023).
Gao, Y. et al. Golgi-associated LC3 lipidation requires V-ATPase in noncanonical autophagy. Cell Death Dis. 7, e2330 (2016).
Xu, Y. et al. A bacterial effector reveals the V-ATPase-ATG16L1 axis that initiates xenophagy. Cell 178, 552–566 (2019).
Hooper, K. M. et al. V-ATPase is a universal regulator of LC3-associated phagocytosis and non-canonical autophagy. J. Cell Biol. 221, e202105112 (2022).
Durgan, J. et al. Non-canonical autophagy drives alternative ATG8 conjugation to phosphatidylserine. Mol. Cell 81, 2031–2040 (2021).
Jia, M. et al. Noncanonical ATG8-ABS3 interaction controls senescence in plants. Nat. Plants 5, 212–224 (2019).
Cross, J. et al. Lysosome damage triggers direct ATG8 conjugation and ATG2 engagement via non-canonical autophagy. J. Cell Biol. 222, e202303078 (2023).
Kumar, S., Jia, J. & Deretic, V. Atg8ylation as a general membrane stress and remodeling response. Cell Stress 5, 128–142 (2021).
Wang, Y. et al. Non-canonical autophagy functions of ATG16L1 in epithelial cells limit lethal infection by influenza A virus. EMBO J. 40, e105543 (2021).
Romao, S. et al. Autophagy proteins stabilize pathogen-containing phagosomes for prolonged MHC II antigen processing. J. Cell Biol. 203, 757–766 (2013).
Heckmann, B. L. et al. LC3-associated endocytosis facilitates beta-amyloid clearance and mitigates neurodegeneration in murine Alzheimer’s disease. Cell 178, 536–551 (2019).
Zheng, X., Chen, S., Gao, C. & Zhou, J. An emerging role of non-canonical conjugation of ATG8 proteins in plant response to heat stress. Autophagy 20, 946–948 (2024).
Fletcher, K. et al. The WD40 domain of ATG16L1 is required for its non-canonical role in lipidation of LC3 at single membranes. EMBO J. 37, e97840 (2018).
Fischer, T. D., Wang, C., Padman, B. S., Lazarou, M. & Youle, R. J. STING induces LC3B lipidation onto single-membrane vesicles via the V-ATPase and ATG16L1-WD40 domain. J. Cell Biol. 219, e202009128 (2020).
Durgan, J. & Florey, O. Many roads lead to CASM: Diverse stimuli of noncanonical autophagy share a unifying molecular mechanism. Sci. Adv. 8, eabo1274 (2022).
Schumacher, K. & Krebs, M. The V-ATPase: small cargo, large effects. Curr. Opin. Plant Biol. 13, 724–730 (2010).
Lupanga, U. et al. The Arabidopsis V-ATPase is localized to the TGN/EE via a seed plant-specific motif. Elife 9, e60568 (2020).
Stephani, M. & Dagdas, Y. Plant selective autophagy-still an uncharted territory with a lot of hidden gems. J. Mol. Biol. 432, 63–79 (2020).
Nieto-Torres, J. L., Leidal, A. M., Debnath, J. & Hansen, M. Beyond autophagy: the expanding roles of ATG8 proteins. Trends Biochem. Sci. 46, 673–686 (2021).
Ritzenthaler, C. et al. Reevaluation of the effects of brefeldin A on plant cells using tobacco Bright Yellow 2 cells expressing Golgi-targeted green fluorescent protein and COPI antisera. Plant Cell 14, 237–261 (2002).
Zhang, G. F., Driouich, A. & Staehelin, L. A. Effect of monensin on plant Golgi: re-examination of the monensin-induced changes in cisternal architecture and functional activities of the Golgi apparatus of sycamore suspension-cultured cells. J. Cell Sci. 104, 819–831 (1993).
Painter, G. R., & Pressman, B. C. Host Guest Complex Chemistry II, 83–110 (2005).
Yoshimoto, K. et al. Processing of ATG8s, ubiquitin-like proteins, and their deconjugation by ATG4s are essential for plant autophagy. Plant Cell 16, 2967–2983 (2004).
Charvat, R. A. & Arrizabalaga, G. Oxidative stress generated during monensin treatment contributes to altered Toxoplasma gondii mitochondrial function. Sci. Rep. 6, 22997 (2016).
Yu, Y. et al. Ascorbic acid integrates the antagonistic modulation of ethylene and abscisic acid in the accumulation of reactive oxygen species. Plant Physiol. 179, 1861–1875 (2019).
Zhou, J., Sun, A. & Xing, D. Modulation of cellular redox status by thiamine-activated NADPH oxidase confers Arabidopsis resistance to Sclerotinia sclerotiorum. J. Exp. Bot. 64, 3261–3272 (2013).
Yu, J. & Zhou, J. Vacuolar accumulation and colocalization is not a proper criterion for cytoplasmic soluble proteins undergoing selective autophagy. Plant Signal Behav. 16, 1932319 (2021).
Krebs, M. et al. Arabidopsis V-ATPase activity at the tonoplast is required for efficient nutrient storage but not for sodium accumulation. Proc. Natl. Acad. Sci. USA 107, 3251–3256 (2010).
Liang, G., Song, H., Xiao, Y. & Zhang, Z. Ammonium accumulation caused by reduced tonoplast V-ATPase activity in Arabidopsis thaliana. Int. J. Mol. Sci. 22, 2 (2020).
Gao, C. et al. A unique plant ESCRT component, FREE1, regulates multivesicular body protein sorting and plant growth. Curr. Biol. 24, 2556–2563 (2014).
Higaki, T. et al. Actin microfilaments regulate vacuolar structures and dynamics: dual observation of actin microfilaments and vacuolar membrane in living tobacco BY-2 cells. Plant Cell Physiol. 47, 839–852 (2006).
Valverde, D. P. et al. ATG2 transports lipids to promote autophagosome biogenesis. J. Cell Biol. 218, 1787–1798 (2019).
Osawa, T. et al. Atg2 mediates direct lipid transfer between membranes for autophagosome formation. Nat. Struct. Mol. Biol. 26, 281–288 (2019).
Alam, J. M. et al. Complete set of the Atg8-E1-E2-E3 conjugation machinery forms an interaction web that mediates membrane shaping. Nat. Struct. Mol. Biol. 31, 170–178 (2023).
Maruyama, T. et al. Membrane perturbation by lipidated Atg8 underlies autophagosome biogenesis. Nat. Struct. Mol. Biol. 28, 583–593 (2021).
Nakatogawa, H., Ichimura, Y. & Ohsumi, Y. Atg8, a ubiquitin-like protein required for autophagosome formation, mediates membrane tethering and hemifusion. Cell 130, 165–178 (2007).
Wang, X. et al. Membrane morphology is actively transformed by covalent binding of the protein Atg8 to PE-Lipids. PLoS ONE 9, e115357 (2014).
Julian, J. et al. ATG8ylation of vacuolar membrane protects plants against cell wall damage. Nat. Plants 11, 321–339 (2025).
Yin, R. et al. Up-regulation of autophagy by low concentration of salicylic acid delays methyl jasmonate-induced leaf senescence. Sci. Rep. 10, 11472 (2020).
Li, F., Chung, T. & Vierstra, R. D. AUTOPHAGY-RELATED11 plays a critical role in general autophagy- and senescence-induced mitophagy in Arabidopsis. Plant Cell 26, 788–807 (2014).
Huang, X. et al. Genetic analyses of the Arabidopsis ATG1 kinase complex reveal both kinase-dependent and independent autophagic routes during fixed-carbon starvation. Plant Cell 31, 2973–2995 (2019).
Zhuang, X. et al. A BAR-domain protein SH3P2, which binds to phosphatidylinositol 3-phosphate and ATG8, regulates autophagosome formation in Arabidopsis. Plant Cell 25, 4596–4615 (2013).
Kang, B. H. Electron microscopy and high-pressure freezing of Arabidopsis. Methods Cell Biol. 96, 259–283 (2010).
Ma, J. et al. Friendly mediates membrane depolarization-induced mitophagy in Arabidopsis. Curr. Biol. 31, 1931–1944 (2021).
Toyooka, K. & Kang, B. H. Reconstructing plant cells in 3D by serial section electron tomography. Methods Mol. Biol. 1080, 159–170 (2014).
Acknowledgements
We appreciate Prof. Zhenhua Zhang (Hunan Agricultural University) for providing us with the vha-a2 vha-a3 double mutants and Prof. Liwen Jiang (The Chinese University of Hong Kong) for providing us with the YFP-ATG8e/DEX:SNF7.1(L32W) and YFP-ATG8e/DEX:VPS4(E232Q) transgenic plants. This work was supported by grants from the Guangdong Basic and Applied Basic Research Foundation (2025A1515010706) and the National Science Foundation of China (31600288) to J.Z., the National Natural Science Foundation of China (32270291, 32470797), the Agricultural and Rural Department of Guangdong Province (Grant No. 2024-NPY-00-004) and the Open Competition Program of Top 10 Critical Priorities of Agricultural Science and Technology Innovation for the 14th Five-Year Plan of Guangdong Province (2022SDZG05) to C.G., and Hong Kong Research Grant Council (GRF14113921, GRF14109222, GRF14110823, GRF14113424, N_CUHK462/22, and C4014-23GF) to B.-H.K. We would like to acknowledge the Open Fund of the MOE Key Laboratory of Laser Life Science and the Institute of Laser Life Science.
Author information
Authors and Affiliations
Contributions
J.Z., J.M., and C.G. designed the experiments. X.Z., J.M., J.Li, J.Luo, J.W., S.C., K.Z., and J.Z. performed the experiments. X.Z., J.M., Y.Z., B.-H.K., C.G., and J.Z. analyzed the data. J.M., J.Li., K.Z., and B.-H.K. contributed to the TEM analysis. J.Z., X.Z., J.M., F.L., C.P., Y.Z., B.-H.K., and C.G. wrote and edited the manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Hugo Zheng, who co-reviewed with Jiaqi Sun and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Zheng, X., Ma, J., Li, J. et al. ATG8ylation-mediated tonoplast invagination mitigates vacuole damage. Nat Commun 16, 6621 (2025). https://doi.org/10.1038/s41467-025-62084-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-025-62084-3
