- © 2019 American Society of Plant Biologists. All rights reserved.
Abstract
Autophagy is a major catabolic pathway whereby cytoplasmic constituents including lipid droplets (LDs), storage compartments for neutral lipids, are delivered to the lysosome or vacuole for degradation. The autophagic degradation of cytosolic LDs, a process termed lipophagy, has been extensively studied in yeast and mammals, but little is known about the role for autophagy in lipid metabolism in plants. Organisms maintain a basal level of autophagy under favorable conditions and upregulate the autophagic activity under stress including starvation. Here, we demonstrate that Arabidopsis (Arabidopsis thaliana) basal autophagy contributes to triacylglycerol (TAG) synthesis, whereas inducible autophagy contributes to LD degradation. We found that disruption of basal autophagy impedes organellar membrane lipid turnover and hence fatty acid mobilization from membrane lipids to TAG. We show that lipophagy is induced under starvation as indicated by colocalization of LDs with the autophagic marker and the presence of LDs in vacuoles. We additionally show that lipophagy occurs in a process morphologically resembling microlipophagy and requires the core components of the macroautophagic machinery. Together, this study provides mechanistic insight into lipophagy and reveals a dual role for autophagy in regulating lipid synthesis and turnover in plants.
INTRODUCTION
Eukaryotic cells use a highly evolutionarily conserved mechanism named autophagy to deliver cytoplasmic constituents into lysosomes or vacuoles for degradation. This catabolic process enables the removal of obsolete or damaged macromolecules, defective organelles, and invading microorganisms and at the same time the recycling of cellular components into needed nutrients and is therefore essential for homeostasis, development, and survival (Jaishy and Abel, 2016; Anding and Baehrecke, 2017; Dikic, 2017). Autophagy is functional at basal levels in virtually all cell types under favorable growth conditions but can also be massively induced by a wide array of developmental and environmental stimuli including nutrient starvation, senescence, pathogens, metabolic stress, and many other abiotic and biotic cues (Levine and Kroemer, 2008; Jaishy and Abel, 2016; Anding and Baehrecke, 2017; Wang et al., 2017; Nakamura and Yoshimori, 2018). The functional importance of both inducible and basal autophagy is well illustrated in plants using autophagy-defective mutants of Arabidopsis (Arabidopsis thaliana) and other plants (Liu et al., 2005; Guiboileau et al., 2012; Yoshimoto, 2012; Minina et al., 2014; Yang and Bassham, 2015; Barros et al., 2017; Üstün et al., 2017; Wang et al., 2017, 2018; Elander et al., 2018; Enrique Gomez et al., 2018). These mutants display early senescence, shortened lifespan, reduced seed yield, defective reproductive growth, and altered phytohormone signaling under normal growth conditions and are also hypersensitive to nutrient deprivation, oxidative stress, pathogen infection, drought, and high salinity, although many mechanistic details underlying these phenotypes and the stress sensitivity remain largely unknown.
Two major types of autophagy-related pathways, namely, macroautophagy and microautophagy, have been described in yeast, plants, and mammals (Yoshimoto, 2012; Noda and Inagaki, 2015; Yang and Bassham, 2015; Antonioli et al., 2017; Galluzzi et al., 2017; Üstün et al., 2017; Elander et al., 2018; Wang et al., 2018). Macroautophagy is the most extensively studied form of autophagy in diverse organisms and is characterized by the formation of a double-membrane structure named an autophagosome. Initial steps in macroautophagy involve the assembly and expansion of an isolated membrane structure called the phagophore. This membrane structure eventually closes to enwrap a portion of cytoplasmic content as cargo. Subsequently, the outer membrane of the autophagosome is fused with the vacuolar membrane to release the inner membrane along with cargo as an autophagic body into the vacuole for degradation and recycling. Macroautophagy is mediated by a group of proteins encoded by AUTOPHAGY-RELATED (ATG) genes that are highly conserved from yeast to mammals and plants. Among them, ATG3, ATG4, ATG5, ATG7, ATG8, ATG10, ATG12, and ATG16 are components of two ubiquitin-like conjugation systems essential for the formation of autophagosomes, their fusion with the tonoplast, and vacuolar degradation (Yoshimoto, 2012; Üstün et al., 2017; Elander et al., 2018; Soto-Burgos et al., 2018). These two systems facilitate the formation of ATG8-phosphatidylethanolamine (PE) conjugates, whose abundance is often used as an indicator of autophagic activity (Suzuki and Ohsumi, 2007; Yoshimoto, 2012). In particular, because ATG8 proteins are membrane associated, residing in phagophores, autophagosomes, and autophagic bodies, green fluorescent protein (GFP)–tagged ATG or autophagic cargo proteins are often used to monitor the delivery of autophagosomes to or the breakdown of autophagic cargos in the vacuole or lysosome (Yoshimoto, 2012; Klionsky et al., 2016; Galluzzi et al., 2017).
In contrast to macroautophagy, our understanding of microautophagy in terms of the underlying mechanism, the molecular machinery involved, and its functional role is still fragmentary, even in simple model systems such as yeast. At the ultrastructural level, microautophagy involves direct engulfment of cytoplasmic components via tonoplast invagination and subsequent release of the cargo into the vacuolar lumen for degradation (Noda and Inagaki, 2015; Dikic, 2017; Galluzzi et al., 2017; Oku and Sakai, 2018). In yeast, various cellular organelles including peroxisomes, mitochondria, the endoplasmic reticulum (ER), and the nucleus were identified as targets of microautophagy (Suzuki and Ohsumi, 2007; Reggiori and Klionsky, 2013; Noda and Inagaki, 2015; Galluzzi et al., 2017). Several forms of yeast microautophagy have been shown to require at least some components of the core machinery of macroautophagy (Suzuki and Ohsumi, 2007; Reggiori and Klionsky, 2013; van Zutphen et al., 2014). In plants, microautophagy has been shown to participate in the degradation of cytoplasmic anthocyanin aggregates (Chanoca et al., 2015) and starch granules (Toyooka et al., 2001) and chloroplasts under oxidative stress (Nakamura et al., 2018), but its underlying mechanism remains largely unknown.
Recent studies indicate that autophagy is functionally connected to lipid metabolism and storage in diverse model systems (Jaishy and Abel, 2016; Shatz et al., 2016; Wang, 2016; Zechner et al., 2017; Elander et al., 2018). Lipids in membranous organelles are used as alternative substrates for energy production via β-oxidation of fatty acids in mitochondria in mammals (Shatz et al., 2016) and peroxisomes in yeast (Kohlwein, 2010) and plants (Graham, 2008) during times of nutrient scarcity. Emerging evidence suggests that rather than directly being used for β-oxidation, fatty acids released from cellular membranes are first stored in the form of triacylglycerol (TAG) in lipid droplets (LDs), and TAG in LDs is then hydrolyzed by a process named lipolysis to supply the cell with fatty acids for the generation of energy (Cabodevilla et al., 2013; Fan et al., 2014, 2017; Rambold et al., 2015). In mammalian cells, autophagic digestion of membranous organelles is the major source of fatty acids for TAG synthesis and LD biogenesis under starvation conditions (Rambold et al., 2015; Nguyen et al., 2017). In addition, autophagy plays an important role in the cellular mobilization and degradation of neutral lipids in LDs, in a process termed lipophagy (Wang, 2016; Zechner et al., 2017). Recent studies have demonstrated a functional link between lipolysis and lipophagy (Martinez-Lopez et al., 2016; Peng et al., 2016), but the exact contribution of each of these two pathways to LD breakdown remains unknown (Zechner et al., 2017). Mammalian lipophagy depends on the core macroautophagy machinery (Jaishy and Abel, 2016) and is morphologically similar to macroautophagy; thus, it is referred to as macrolipophagy (Singh et al., 2009). Consequently, disruption of the core ATG genes increases LD accumulation in various organs (Singh et al., 2009). Unlike the situation in mammals, lipophagy in yeast resembles microautophagy and therefore is referred to as microlipophagy (van Zutphen et al., 2014). It has been suggested that microlipophagy plays an important role in maintaining cell viability (van Zutphen et al., 2014) and membrane integrity (Wang et al., 2014) under carbon starvation in yeast, but controversy exists as to whether this process depends on (van Zutphen et al., 2014; Wang et al., 2014) or can occur independently of (Oku et al., 2017) the core ATG proteins. Autophagy-like processes have also been shown to participate in TAG synthesis and/or LD breakdown in microalgae (Zhao et al., 2014; Schwarz et al., 2017) and in rice (Oryza sativa) during anther development (Kurusu et al., 2014). Disruption of autophagy has been shown to affect lipid turnover in maize (Zea mays; McLoughlin et al., 2018) and Arabidopsis seedlings under carbon starvation (Avin-Wittenberg et al., 2015), but whether lipophagy occurs in Arabidopsis and the molecular mechanism involved remain unknown.
In plants, as in yeast and mammals, TAG is assembled in the ER and stored in LDs in the cytosol (Chapman and Ohlrogge, 2012). In Arabidopsis, phospholipid:diacylglycerol acyltransferase1 (PDAT1) is a key enzyme catalyzing the last step of TAG assembly (Zhang et al., 2009). TAG breakdown is catalyzed by cytosolic lipases including SUGAR-DEPENDENT1 (SDP1), a patatin domain lipase responsible for the initiation of TAG catabolism (Eastmond, 2006). Disruption of SDP1 blocks TAG hydrolysis in germinating seeds (Eastmond, 2006) and in vegetative tissues such as mature leaves and roots (Kelly et al., 2013; Fan et al., 2014), suggesting that cytosolic lipolysis plays a dominant role in TAG breakdown in both seed and nonseed tissues under normal growth conditions. Under extended darkness, TAG levels in leaves of sugar dependent1 (sdp1) mutants increased rapidly and then declined, suggesting an activation of unknown, alternative pathways for TAG hydrolysis under starvation conditions (Fan et al., 2017).
Lipid metabolism in photosynthetic tissues such as leaves is geared toward the supply of building blocks for organellar membrane biogenesis and maintenance. As a result, leaf tissues do not accumulate TAG to significant amounts, although they do possess a high capacity for its synthesis and metabolism (Xu and Shanklin, 2016). In Arabidopsis, two parallel pathways, compartmentalized in either the ER or the chloroplast, contribute to membrane lipid biosynthesis (Browse and Somerville, 1991; Ohlrogge and Browse, 1995). Disruption of either pathway causes drastic changes in lipid metabolism including an increase in fatty acid synthesis and turnover and an accumulation of TAG (Fan et al., 2013a, 2015). In the trigalactosyldiacylglycerol1 (tgd1) mutant, a defect in the ER pathway also results in a compensatory increase in the chloroplast pathway activity (Xu et al., 2003, 2005). Similarly, overexpressing PDAT1 draws lipids from the ER pathway to TAG synthesis, causing an increase in the biosynthesis of thylakoid lipids via the chloroplast pathway (Fan et al., 2013b). On the other hand, the plastidic glycerol-3-phosphate acyltransferase1 (act1) mutant is defective in the initial step in the chloroplast pathway of membrane lipid synthesis (Kunst et al., 1988; Xu et al., 2006), and the vast majority of membrane lipids in act1 are assembled via the ER pathway (Kunst et al., 1988).
To understand the role of autophagy in lipid metabolism at the mechanistic level, we generated a series of double mutants defective in autophagy in the tgd1-, sdp1-, or PDAT1-overexpressing-line background. Using these mutants along with transgenic plants coexpressing an LD-targeted, GFP-tagged OLEOSIN1 (OLE1) fusion protein (Fan et al., 2013b) and autophagic or tonoplast markers, we demonstrate here an important role of autophagy in regulating TAG synthesis, membrane lipid turnover, and fatty acid synthesis under normal growth conditions and in mediating LD degradation under starvation. We show that lipophagy occurs in a process morphologically resembling microautophagy in yeast and requires key core players in macroautophagy. This study demonstrates the functional importance of autophagy in TAG metabolism and storage and the mechanistic basis for lipophagy in plants.
RESULTS
Basal Autophagy Contributes to TAG Synthesis in Leaves of Adult Plants
To test the role of autophagy in lipid metabolism in plants, we first compared TAG levels in mature seeds, young seedlings, and leaves of adult plants between the wild type and two atg mutants defective in ATG2 or ATG5, two core protein components of the macroautophagic machinery. Disruption of autophagy caused small but significant decreases in TAG content in seeds (Figure 1A) and 4-d-old seedlings (Figure 1B). Seed weight was slightly decreased in atg2-1 (169.3 ± 5.3 µg/10 seeds) and atg5-1 (181.3 ± 11.7 µg/10 seeds) compared with the wild type (186.7 ± 3.3 µg/10 seeds). TAG levels were low in developing leaves but increased as leaves matured and aged. Compared with the wild type, TAG content was reduced by an average of 29%, 52%, and 42% in developing, fully mature, and senescing leaves of atg mutants, respectively (Figure 1C). In all tissues examined, there were no significant differences in TAG content between atg2-1 and atg5-1, suggesting the decreased TAG levels in atg mutants are associated with defects in basal autophagy.
Disruption of Autophagy Reduces TAG Accumulation.
(A) to (C) TAG levels in dry seeds (A), 4-d-old seedlings (B), and leaves of 5-week-old plants (C). Data are means of three replicates with sd. Asterisks indicate statistically significant differences from the wild type based on Student’s t test (*P < 0.05, **P < 0.01; Supplemental Data Set). FW, fresh weight; WT, wild type.
Mutants defective in the core components of autophagy often display pleiotropic phenotypes including early senescence and defects in nutrient remobilization. Therefore, it is possible that the observed decrease in TAG content in seeds in atg mutants is due to a decrease in resource allocation to seeds rather than to a change in seed TAG metabolism. Similarly, a decreased TAG storage in seeds may also affect TAG content in young seedlings. To test these possibilities, we performed radiotracer labeling experiments using two different labeled substrates, 14C-acetate and 3H2O, substrates that label nascent fatty acids with 14C or 3H during the initial or reduction steps of fatty acid synthesis, respectively (Browse et al., 1981). Under our growth conditions, the incorporation of the radiolabel from14C-acetate or tritiated water (3H2O) into fatty acids of developing embryos was linear for at least 1 h (Supplemental Figure 1). The rate of incorporation of 14C or 3H into TAG calculated following 1 h of incubation was similar between the wild-type and atg embryos (Supplemental Figure 2). Likewise, there was no significant difference in the rate of radiolabeled TAG accumulation between the wild-type and atg seedlings. On the other hand, the rate of radiolabel incorporation into TAG was significantly reduced in mature and senescing leaves, with the largest effect being observed in mature leaves and the least in developing leaves (Figure 2), mirroring the differences in TAG content in leaves at different ages (Figure 1). Again, leaf TAG levels and rates of radiolabel incorporation into TAG were similar between two atg mutants.
Disruption of Autophagy Reduces TAG Synthesis in Mature and Senescing, But Not in Growing Leaves.
(A) and (B) Detached leaves of 5-week-old plants were incubated with 14C-acetate (A) or 3H2O (B) for 1 h, and total radioactivity in TAG was measured by scintillation counting following separation by thin layer chromatography. Data are means of three replicates with sd. Asterisks indicate statistically significant differences from the wild type based on Student’s t test (P < 0.01). FW, fresh weight; WT, wild type.
Disruption of Basal Autophagy Compromises Membrane Lipid Turnover
The decreased rate of radiolabel incorporation into TAG in atg leaves may be due to a decrease in fatty acid synthesis or a decline in the mobilization of fatty acids from organellar membranes to TAG via autophagy. The rate of fatty acid synthesis can be assessed by measuring the rate of 14C-acetate or 3H2O incorporation into total fatty acids (Browse et al., 1981). As shown in Supplemental Figure 3, growing leaves incorporated 14C from 14C-acetate or 3H from 3H2O into total lipids at a higher rate than did mature and senescing leaves, likely reflecting a higher demand for fatty acids to support membrane expansion and organellar biogenesis during rapid growth. Rates of radiolabel incorporation following 1 h of incubation were similar in the wild-type and atg leaves. These results suggest that the decreased TAG synthesis in atg mutants is not due to a decline in the rate of fatty acid synthesis.
We next tested whether disruption of autophagy affects membrane lipid turnover. To this end, we first incubated leaves with 14C-acetate for 1 h (pulse). After thoroughly washing with water to remove 14C-acetate, the leaves were incubated in unlabeled solution for an additional 3 d (chase). The radiolabel in leaf total membrane lipids following 1 h of pulse was similar between the wild type and atg mutants (Supplemental Figure 4). Quantification of radioactivity in total membrane lipids showed significant decreases in rates of radiolabeled fatty acid loss, particularly in mature and senescing leaves of atg mutants compared with the wild-type leaves of the same age during 3 d of chase (Figure 3). Together, results from pulse-chase labeling experiments suggest that disruption of autophagy results in a decrease in membrane lipid turnover and hence the accumulation of leaf TAG.
Disruption of Autophagy Slows Down Membrane Lipid Turnover in Mature and Senescing, but Not in Growing Leaves.
Radiolabel loss was calculated as percentage of loss of radioactivity in total membrane lipids during 3 d of chase following 1 h of 14C-acetate pulse of detached leaves of 5-week-old plants. Data are means of three replicates with sd. Asterisks indicate statistically significant differences based on Student’s t test (P < 0.01). WT, wild type.
The Contribution of Autophagy to TAG Synthesis Differs in Mutants Defective in the Chloroplast or the ER Lipid Biosynthesis Pathway
To provide additional evidence for the involvement of autophagy in TAG synthesis and also to test the relative contribution of the chloroplast versus the ER lipid assembly pathway to autophagy-mediated TAG synthesis, we constructed double mutants between tgd1 and atg2-1 or atg5-1. In addition, we crossed the PDAT1 overexpression (OE) line 4 in the act1 mutant background (act1/PDAT1-OE4; Fan et al., 2013b) with atg2-1 or atg5-1 and obtained the PDAT1-OE4 line in backgrounds of the wild type or double mutants of atg2-1 act1 or atg5-1 act1. Assays for PDAT activity in microsomal membranes revealed that disruption of autophagy had no significant effect on TAG formation from 14C-labeled phosphatidylcholine (PC), whereas the activity was more than fourfold higher in transgenic plants overexpressing PDAT1 compared with the wild type (Supplemental Figure 5). Analysis of lipid extracts from mature leaves of 5-week-old plants showed that TAG content was higher in tgd1, as expected (Figure 4). Interestingly, there was also a significant increase in TAG in act1 compared with the wild type. Disruption of autophagy caused significant decreases in TAG content in atg2-1 tgd1 and atg5-1 tgd1. TAG levels were 1.9- and 3.5-fold higher in PDAT1-OE4 and act1/PDAT1-OE4, respectively, compared with act1. Importantly, disruption of ATG2 or ATG5 resulted in more than 50 and 70% decreases in TAG content in PDAT1-OE4 and act1/PDAT-OE4, respectively. To confirm these results, we crossed atg2-1 or atg5-1 with another independent line act1/PDAT1-OE5 (Fan et al., 2013b) and recovered the PDAT1-OE5 in backgrounds of the wild type or double mutants of atg2-1 act1 or atg5-1 act1. Lipid analysis revealed that TAG levels were an average of 46% and 67% lower in atg/PDAT1-OE5 and atg/act1/PDAT1-OE5 compared with PDAT1-OE5 and act1/PDAT1-OE5, respectively (Supplemental Figure 6).
Disruption of Autophagy Reduces Leaf TAG Accumulation.
TAG content in mature leaves of 4-week-old PDAT1-overexpressing line 4 (PDAT1-OE4) in the wild-type, act1, atg2-1, atg5-1, atg2-1 act1, or atg5-1 act1 background. Data are means of three replicates with sd. Asterisks indicate statistically significant differences based on Student’s t test (*P < 0.05, **P < 0.01). FW, fresh weight.
Labeling experiments using 14C-acetate revealed that disruption of ATG2 or ATG5 caused significant decreases in rates of radiolabel incorporation into total fatty acids in tgd1, PDAT1-OE4, and act1/PDAT1-OE4 (Figure 5A). In addition, rates of the decay in labeled fatty acids of membrane lipids during 3 d of chase following 1 h of pulse with14C-acetate (Figure 5B) were significantly lower in atg tgd1, atg PDAT1-OE4, and atg/act1/PDAT1-OE4 than in the wild type, PDAT1-OE4, or act1/PDAT1-OE4. Together, these results suggest that basal autophagy plays an important role in regulating both fatty acid synthesis and membrane lipid turnover and that the ER lipid biosynthesis pathway contributes more to autophagy-mediated leaf TAG synthesis than the chloroplast pathway.
Disruption of Autophagy Reduces Fatty Acid Synthesis and Membrane Lipid Turnover in Growing Leaves of the tgd1 Mutant and PDAT1-Overexpressing Lines.
(A) Rate of 14C-acetate incorporation into total fatty acids in growing leaves of the 4-week-old PDAT1-overexpressing line 4 (PDAT1-OE4) in the wild-type, act1, atg2-1, atg5-1, atg2-1 act1, or atg5-1 act1 background.
(B) Radiolabel loss during the 3-d chase following 1 h incubation with 14C-acetate.
Data are means of three replicates with sd. Asterisks indicate statistically significant differences based on Student’s t test (*P < 0.05, **P < 0.01). FW, fresh weight; WT, wild type.
Lipophagy Is Induced under Starvation
Our data so far indicate that autophagy contributes to TAG synthesis and membrane lipid turnover, but it is not clear whether this mechanism is also involved in the breakdown of TAG stored in LDs. As a first step toward answering this important question, we took advantage of OLE1-GFP-overexpressing lines (Fan et al., 2013b). OLE1 is one of the most abundant LD proteins in seeds (Huang, 1996). When ectopically expressed in leaves, OLE1-GFP is specifically targeted to the surface of LDs (Wahlroos et al., 2003), and ectopic expression of OLE1-GFP in the tgd1 (tgd1/OLE1-GFP) or wild-type background (WT/OLE1-GFP) boosted TAG accumulation and induced the formation of clusters of small LDs (Fan et al., 2013b). To facilitate epifluorescence imaging of LD and autophagic dynamics, we generated stable transgenic lines constitutively overexpressing the autophagosome marker ATG8e, one of the nine isoforms of ATG8, fused at the N terminus with a Discosoma sp red fluorescent protein (DsRed) in the tgd1/OLE1-GFP background (DsRed-ATG8e/tgd1/OLEI-GFP). Under normal growth conditions, very few, if any, DsRed-ATG8e–labeled punctuate structures, likely representing immature and mature autophagosomes and their precursors (Yoshimoto, 2012), were detected in DsRed-ATG8e/tgd1/OLEI-GFP lines (Figure 6). When exposed to extended darkness, a starvation condition known to induce autophagy (Breeze et al., 2011), the number of DsRed-ATG8e–labeled puncta significantly increased (Figure 6; Supplemental Figure 7), suggesting an increase in autophagic activity. In addition, while the OLE1-GFP signals rarely overlapped with DsRed-ATG8e–labeled structures under normal growth conditions, some of the OLE1-GFP signals colocalized with DsRed-ATG8e after 3 d of dark treatment (Figure 6). The extent of colocalization was quantified using the Costes image randomization test (Costes et al., 2004). The Pearson’s correlation coefficient (PCC) was determined on 15 image pairs from at least three independent experiments. The average PCC for OLE1-GFP colocalization with DsRed-ATG8e was 0.51 ± 0.15 (n = 15) with an average Costes P-value of 1.00 ± 0.01 (n = 15; Figure 6), confirming that a subpopulation of LDs colocalize with DsRed-ATG8e–labeled autophagic structures. The relatively low PCC most likely reflects the large difference in size between the DsRed-labeled structures (less than 100 nm in diameter, Figure 7A) and the OLE1-GFP–labeled LD clusters (5 to 10 µm in diameter, Figure 7).
Colocalization of LDs With Autophagic Structures in Leaves Under Dark-Induced Starvation.
Confocal images of mature leaves of 4-week-old transgenic plants coexpressing OLE1-GFP (green) and DsRed-ATG8e (red) in tgd1 before and after 3 d of dark treatment. Boxed areas show colocalization of green and red signals under higher magnification. Quantification of colocalization is provided by the PCC and the Costes P-value below the images. Bars = 20 µm.
Autophagy of Leaf LDs Under Dark-Induced Starvation.
(A) Confocal images of mature leaves of 4-week-old transgenic plants coexpressing OLE1-GFP and DsRed-ATG8e in tgd1 after 3 d of dark treatment. Bars = 1 µm.
(B) to (D) Electron micrographs of LD clusters in leaf cells of tgd1 overexpressing OLE1-GFP before (B) and after (see [C] and [D]) 3 d of dark treatment. (D) Enlargement of the boxed area in (C). Arrows indicate LDs. Bars = 1 µm in (B) and (D) and 2 µm in (C).
Under higher magnification, DsRed-ATG8e–labeled autophagic structures were clearly found to be associated with LDs (Figure 7A). Transmission electron microscopy analysis of leaves of dark-treated tgd1/OLEI-GFP plants showed that, before dark treatment, LDs were often found to be tightly packed in large clusters with diameters of up to 10 µm (Figure 7B). After dark treatment for 3 d, autophagic vacuoles (AVs) appeared in LD clusters, some of which contained LDs (Figure 7C), which appeared to be partially degraded (Figure 7D).
Free GFP is relatively resistant to degradation within the vacuole or lysosome (Yoshimoto, 2012; Klionsky et al., 2016; Galluzzi et al., 2017). Therefore, if OLE1-GFP–coated LDs are degraded in the vacuole, we would expect to observe an increased accumulation of free GFP under dark treatment. To test this possibility, we exposed the WT/OLEI-GFP and tgd1/OLE1-GFP lines to extended darkness. Immunoblot analysis using antibody against GFP showed that free GFP levels markedly increased in both WT/OLEI-GFP and tgd1/OLE1-GFP lines following 3 d of darkness (Figure 8A). Autophagic activity can also be assessed by monitoring the protein level of ATG8-Phosphatidylethanolamine (PE; Suzuki and Ohsumi, 2007; Yoshimoto, 2012), which migrates faster on SDS-PAGE in the presence of urea than does the unmodified form (Chung et al., 2009, 2010). Immunoblot analysis using antibody against ATG8 showed that ATG8-PE conjugates were absent in leaves prior to dark treatment but accumulated after 3 d of darkness (Figure 8A), indicating an overall increase in autophagic activity during dark-induced starvation conditions, as expected. Time-course analysis showed that free GFP levels were low under normal growth conditions but increased steadily during 5 d of darkness, similar to dark-induced accumulation of ATG8-PE (Figure 8B). Importantly, blocking autophagy by disruption of ATG5 largely prevented the accumulation of free GFP in dark-treated WT/OLE1-GFP and tgd1/OLE1-GFP plants (Figure 8A), suggesting that free GFP accumulation is a result of lipophagic activity. ATG5 has been shown to be essential for ATG8 lipidation (Chung et al., 2010). Consistent with this, no ATG8-PE conjugates were detected in atg5-1 leaves following 3 d of dark treatment (Figure 8A). Together, these results provide evidence that lipophagy is induced during dark-induced starvation.
OLE1-GFP–Coated Leaf LDs Are Degraded in Vacuoles Under Dark-Induced Starvation.
(A) Accumulation of free GFP and ATG8-PE in mature leaves of the 4-week-old wild-type and tgd1 plants, but not in mature leaves of atg5-1 tgd1 double mutant overexpressing OLE1-GFP following 3 d of darkness.
(B) Time course of free GFP and ATG8-PE accumulation in mature leaves of tgd1 overexpressing OLE1-GFP under dark treatment.
Equal amounts of proteins were subjected to SDS-PAGE followed by immunoblot analysis with antibodies against GFP, ATG8, or the loading control actin. The dashed lines and asterisks locate free ATG8 proteins and ATG8-PE conjugates, respectively. WT, wild type.
Vacuolar Degradation of LDs Occurs in a Process That Resembles Microlipophagy
To dissect the mechanistic basis for lipophagy, we generated transgenic plants coexpressing the tonoplast marker δ tonoplast intrinsic protein fused to a Discosoma sp red fluorescent protein (δTIP-DsRed) and OLE1-GFP in the wild-type background (δTIP-DsRed/OLE1-GFP). When these plants were exposed to dark treatment for 2 d, individual LDs or LD clusters were observed inside (Figures 9A to 9D) or within invagination of (Figures 9E and 9F) δTIP-DsRed–labeled tonoplasts. Analysis of max-intensity projection images of z-stacks acquired by confocal microscopy revealed that the LDs were clearly enclosed by the tonoplast (Figures 9G and 9H).
LD and Vacuole Interactions During Dark-Induced Starvation.
(A) to (F) Confocal images of cotyledon cells of the 7-d-old wild-type transgenic plants coexpressing the tonoplast marker δTIP-DsRed (red) and OLE1-GFP (green) after 2 d of darkness in the presence of 0.5 µM concA. Overlay of red and green fluorescence showing the presence of LDs in vacuoles (see [A] to [D]) or within tonoplast invagination (see [E] and [F]). Bars = 10 µm.
(G) and (H) Three-dimensional images reconstructed from a series of confocal z-stack images. Bars = 10 µm
Since typical autophagosomes are small vesicles with diameters of only ∼1 µm (Merkulova et al., 2014), the large size of LD clusters in OLEI-GFP transgenic plants (Figure 7; Fan et al., 2013b) may exceed the size capacity of autophagosomes, thereby impeding the recruitment of LD clusters into these structures. Therefore, to further examine the process leading to lipophagy in plants, we took advantage of sdp1 mutants, which accumulated small LDs under dark-induced starvation conditions (Fan et al., 2017) and performed ultrastructural analysis of leaf cells of dark-treated sdp1 plants by transmission electron microscopy. Because lipophagy and autophagy appeared to be induced after, but not within, the first 1 d of darkness (Figure 8B), we focused on the subcellular morphological changes in leaf samples between 1 and 2 d of dark treatment. Consistent with changes in ATG8-PE abundance (Figure 8B), very few autophagic structures were seen after 1 d of darkness (Supplemental Figure 8A). Following 2 d of dark treatment, however, we observed the occurrence of autophagosomes (Supplemental Figures 8B and 8C) and many small vacuoles with diameters of 0.5 to 2 µm (Supplemental Figure 8D). Many of these structures contained autophagic bodies or remnants of cytoplasmic materials, suggesting that they are AVs.
Following 1 d of darkness, small LDs with an ∼100-nm diameter were frequently seen as spherical structures present in the cytosol (Figure 10A). LDs increased in size after 2 d of dark treatment (Figures 10B to 10F) and were frequently found to be in close contact with AVs (Figure 10C) or within invagination of AV membranes (Figures 10B and 10C) or inside AVs (Figure 10D) or the central vacuole (Figure 10E). Immunoelectron microscopy of dark-treated sdp1 plants with ATG8 antibody revealed the presence of immunogold particles on LDs (Figure 10F). Interestingly, LDs appeared to undergo degradation prior to being fully internalized into AVs, along with other sequestered materials (Figure 10C). Dark treatment in the presence of concanamycin A (concA) also led to the appearance of LDs in the central vacuole in leaves of wild-type seedlings (Supplemental Figure 9). On the other hand, we did not detect association of macroautophagic membrane structures with LDs as observed during macrolipophagy in mammals (Singh et al., 2009), suggesting that autophagic degradation of LDs during dark-induced starvation in plants is a microlipophagy-like process. Importantly, disruption of ATG2 (Figures 11A and 11B) or ATG5 (Figures 11C and 11D) in sdp1 largely blocked the formation of AVs and hence the interaction between LDs and AVs. In atg sdp1 double mutants, most of LDs were still present in the cytosol after 2 d of dark treatment.
Vacuolar Uptake of LDs Under Dark-Induced Starvation.
(A) to (F) Electron micrographs of leaf cells of 4-week-old sdp1-4 plants dark treated for 1 d (A) and 2 d (see [B] to [F]). Arrows indicate LDs. Bars = 0.5 µm in (A), 1 µm in (B) to (D), 2 µm in (E), 0.2 µm in (F), and 0.1 µm in the inset in (F).
(B) to (D) Various stages of LD internalization into AVs (see [B] and [D]). Note that LD is partially degraded within invagination of the AV membrane in (C).
(E) Presence of LDs in the central vacuole. CV, central vacuole.
(F) Immunogold labeling of sdp1-4 seedlings dark treated for 2 d in the presence of 0.5 µM concA using the ATG8 antibody. Arrowheads indicate gold particles. The inset shows higher magnification of the boxed region.
Disruption of Autophagy Blocks Autophagic Degradation of LDs.
(A) to (D) Electron micrographs of leaf cells of 4-week-old atg2-1 sdp1-4 (see [A] and [B]) and atg5-1 sdp1-4 (see [C] and [D]) plants dark treated for 2 d. Bars = 1 µm in (A), (B), and (D) and 2 µm in (C).
Autophagic Flux Remains Unchanged in sdp1 under Starvation
We next tested whether deficiency in cytosolic lipolysis affects autophagy under dark-induced starvation in plants as in mammals (Sathyanarayan et al., 2017). To do so, we crossed the wild type or sdp1 with tgd1 overexpressing DsRed-ATG8e and recovered the DsRed-ATG8e line in the wild-type or sdp1 background. As expected, the number of DsRed-ATG8e–labeled puncta increased under dark-induced starvation (Supplemental Figure 10A). Quantitative analysis showed that there was no significant difference in the number of puncta between the wild type and sdp1 after 4 d of darkness. In addition, there was an increase in levels of faster migrating forms of ATG8-PE during dark-induced starvation conditions (Supplemental Figure 10B); and again, there were no discernible differences in levels of starvation-induced ATG8-PE between the wild type and sdp1 mutants. Together, these data suggest that disruption of SDP1 does not affect autophagic flux under dark-induced starvation conditions in Arabidopsis.
Inhibition of Autophagy Increases Leaf TAG Accumulation in sdp1 under Starvation
Under dark-induced starvation conditions, TAG accumulated rapidly within the initial 1 d and then started to decline in leaves of sdp1 plants, likely reflecting the induction of lipophagy after dark treatment for 1 d (Fan et al., 2017). To test this possibility, we treated detached leaves of sdp1 mutants with 3-methyladenine (3-MA), a widely used inhibitor of autophagy in mammals (Blommaart et al., 1997) and plants (Yoshimoto, 2012). In untreated control leaves, TAG content increased by more than sixfold during the initial 2 d of dark treatment (Figure 12A). Treatment with 3-MA did not affect TAG levels during the initial 2 d of dark incubation, suggesting an involvement of an autophagy-independent mechanism in TAG synthesis. However, TAG content declined after day 2 of dark treatment in the untreated control but continued to increase toward the end of the experiment in 3-MA–treated leaves, such that TAG content was significantly higher at days 3 and 4 in 3-MA–treated leaves compared with the untreated control. These results suggest that autophagy contributes to TAG hydrolysis under severe starvation.
Inhibition of Autophagy Enhances TAG Accumulation in sdp1-4 under Extended Darkness.
(A) Changes in TAG levels in detached sdp1-4 mature leaves during dark treatment in the presence or absence of 3-MA.
(B) Changes in TAG levels in mature leaves of 4-week-old sdp1-4, atg2-1 sdp1-4, and atg5-1 sdp1-4 plants during dark treatment.
Data are means of three replicates with sd. Asterisks indicate statistically significant differences from controls (A) or sdp1 (B) based on Student’s t test (*P < 0.05, **P < 0.01). FW, fresh weight.
Lipophagy Requires the Core Components of Autophagy
To provide genetic evidence for the induction of lipophagy during dark-induced starvation and also to test whether lipophagy depends on the core autophagic machinery, we generated double mutants between sdp1 and atg2-1 or atg5-1. Under normal growth conditions, TAG levels were lower in leaves of atg sdp1 double mutants compared with sdp1 (Figure 12B). During dark treatment, TAG levels in sdp1 peaked at day 1 following dark exposure and started to decline thereafter (Figure 12B), consistent with our previous report (Fan et al., 2017). In contrast to sdp1, TAG content in atg sdp1 double mutants increased steadily during the first 2 d of dark treatment and remained largely unchanged at day 3. Statistical analysis confirmed that atg sdp1 double mutants accumulated significantly more TAG at days 2 and 3 following dark treatment compared with the sdp1 single mutant (Figure 12B). TAG levels remained largely unchanged in the wild type and atg single mutants following dark incubation for 3 d (Supplemental Figure 11).
The increased TAG accumulation in atg sdp1 double mutants could result from a decrease in TAG hydrolysis or an increase in the conversion of membrane lipids to TAG. To test these possibilities, we analyzed the changes in levels of total membrane lipids during dark treatment. We detected no significant differences in leaf membrane lipid content among wild type, single, and double mutants prior to or during 3 d of darkness (Supplemental Figure 12). Since fatty acid synthesis is completely inactive in the dark (Bao et al., 2000), the absence of change in membrane lipid level supports the hypothesis that disruption of autophagy does not affect membrane lipids to TAG conversion in sdp1 under dark treatment. Total membrane lipid levels were decreased to a similar extent following 3 d of dark treatment in all genotypes analyzed (Supplemental Figure 12), apparently because of an increase in fatty acid β-oxidation (Fan et al., 2017). Together, these results suggest that the increased TAG accumulation in atg sdp1 double mutants compared with sdp1 is due to decreased lipophagic activity and that lipophagy relies on the core machinery such as ATG2 and ATG5.
DISCUSSION
We have shown that autophagy plays an important role in organellar membrane turnover, TAG synthesis, and LD accumulation under normal growth conditions. Lipophagy, the autophagic degradation of LDs, was induced following extended dark treatment as evident from increased colocalization of LDs and autophagic structures, an increase in accumulation of free GFP derived from OLE1-GFP–coated LDs, the presence of LDs in vacuoles, the association of autophagic marker protein ATG8 with LDs, and an increase in TAG levels in atg sdp1 double mutants compared with sdp1. We show that lipophagy occurs in a process resembling microlipophagy as described in yeast and requires the core components of macroautophagy. These results provide mechanistic insight into the role of autophagy in lipid metabolism in plants and lend further support for a critical role of autophagy in quality control of cellular organelles (Yang and Bassham, 2015; Wang et al., 2018) and an immediate role of TAG metabolism in membrane lipid turnover (Fan et al., 2014, 2017).
Role of Basal Autophagy in TAG Synthesis
Our results show that disruption of autophagy impedes membrane lipid turnover and hence TAG synthesis under normal growth conditions. Importantly, the role of basal autophagy in lipid metabolism is tissue and/or development specific: the contribution of autophagy to TAG synthesis is insignificant in young seedlings, rapidly expanding leaves, and developing seeds but significant in mature and senescing leaves of adult plants. These results are perhaps not surprising because, in contrast to the situation in mature and senescing leaves, organellar membranes in growing cells are newly formed and therefore may not be targeted for autophagy-mediated degradation under normal growth conditions. In developing embryos, fatty acids in membrane lipids are known to be directed to TAG synthesis via acyl editing and headgroup exchange (Bates et al., 2012).
In plants, autophagy has been implicated in the degradation of peroxisomes (Kim et al., 2013; Shibata et al., 2013), mitochondria (Li et al., 2014), ER (Liu et al., 2012), and chloroplasts (Ishida et al., 2014), and fatty acids released from membranes of these and other organelles may be used for TAG synthesis (Figure 13). The contribution of autophagy to TAG synthesis is higher in act1 defective in the chloroplast pathway of glycerolipid biosynthesis but lower in tgd1 disrupted in the parallel ER pathway. These results suggest a key role of ER in the provision of fatty acids and/or lipid precursors for autophagy-mediated TAG synthesis. The importance of ER in autophagy-mediated TAG synthesis may reflect not only the role of autophagy in the degradation of this organelle (Liu et al., 2012) but also roles of ER as a membrane source for autophagosome biogenesis (Zhuang et al., 2018) and as the entry point for fatty acids exported from the plastid into the ER pathway of lipid biosynthesis (Bates et al., 2012).
A Proposed Model for the Role of Autophagy in Lipid Metabolism in Plants.
De novo fatty acid (FA) synthesis in chloroplasts is mediated by a series of enzymatic reactions collectively referred to as fatty acid synthase. The resultant FAs feed into membrane lipid synthesis via two parallel pathways localized in the chloroplast or the ER. Autophagy-mediated degradation of cellular organelles other than chloroplasts provides a source of FAs for TAG synthesis under normal and starvation conditions. Thylakoid lipids are broken down by hydrolytic enzymes inside the chloroplast, and the released FAs are used for TAG synthesis. TAG is packaged in LDs in the cytosol. Under normal growth conditions, TAG stored in LDs is hydrolyzed by SDP1. Nutrient starvation triggers microlipophagy, which functions together with cytosolic lipolysis catalyzed by SDP1 to mediate LD breakdown into FAs for energy production through β-oxidation. Black arrows represent processes occurring in both normal and starvation conditions. The red arrow is specific to starvation. FAS, fatty acid synthase; HEs, hydrolytic enzymes.
Previous studies have shown that during autophagy-mediated chloroplast breakdown, stromal proteins (Ishida et al., 2014), but not thylakoid components (Spitzer et al., 2015; Wang et al., 2018), are delivered into vacuoles for degradation. In line with these observations, our results showed that disruption of autophagy had no significant impact on the dark-induced synthesis of TAG (Figure 12B), which is mainly derived from thylakoid lipids (Kunz et al., 2009; Fan et al., 2017). Similarly, treatment with 3-MA did not affect TAG content during the initial 2 d of dark treatment (Figure 12A), suggesting that autophagy-independent breakdown of chloroplasts serves as a main source of fatty acids for TAG synthesis. In addition, our microscopy analysis showed that the number of chloroplasts per cell remained unaltered during dark treatment (Supplemental Figure 13), consistent with previous reports (Keech et al., 2007; Evans et al., 2010). These results exclude the possibility of whole chloroplast autophagy as observed in plants under photooxidative stress (Izumi et al., 2017; Nakamura et al., 2018) or in mutants defective in plastid protein import (Niwa et al., 2004) as a means for thylakoid degradation under dark-induced starvation. The autophagy-independent degradation of thylakoids is also consistent with previous reports showing an internal dismantling of thylakoid systems during senescence-induced chloroplast breakdown (Evans et al., 2010; van Doorn and Papini, 2013).
In addition to reduced organellar membrane turnover and TAG synthesis, disruption of basal autophagy results in significant decreases in fatty acid synthesis in tgd1 or PDAT1-OE lines (Figure 5A). Although the exact mechanistic basis as to how autophagy impacts fatty acid synthesis remains unclear, it is possible that blocking autophagy results in a buildup of fatty acids in the cytosol due to reduced cellular fatty acid needs for organellar membrane lipid turnover, which act as feedback signals to negatively regulate fatty acid synthesis in the chloroplast. On the other hand, overexpression of PDAT1 or blocking the chloroplast lipid biosynthesis pathway in act1 accelerates autophagy-mediated membrane lipid turnover and hence increases the cellular demand for fatty acids. This increased fatty acid demand may cause a decrease in fatty acids in the cytosol, thereby partially relieving feedback inhibition on plastid fatty acid synthesis. In this context, it is worth noting that inefficient utilization of fatty acids for glycerolipid biosynthesis in the ER has been shown to cause a feedback inhibition on fatty acid synthesis by an unknown mechanism (Bates et al., 2014), and exogenous fatty acid applications to cell cultures or isolated chloroplasts can also elicit feedback inhibition on fatty acid synthesis (Ohlrogge and Jaworski, 1997), likely involving 18:1-acyl carrier protein as a signal molecule (Andre et al., 2012).
TAG and fatty acid synthesis are increased in tgd1 mutants (Fan et al., 2013a), and disruption of SDP1 causes a significant increase (Fan et al., 2014), whereas blocking autophagy caused a significant decrease (Figure 4) in TAG accumulation in tgd1. These results suggest that under normal growth conditions, autophagy functions in TAG synthesis, whereas the cytosolic pathway mediated by neutral lipases including SDP1 is the major mechanism for TAG catabolism (Figure 13).
Inducible Autophagy Is Involved in LD Breakdown via Microlipophagy
Under extended darkness, TAG content decreases when autophagy is induced but increases when autophagy is disabled in sdp1. In addition, disruption of SDP1 does not impact autophagic flux under either normal growth or starvation conditions (Supplemental Figure 10). These results suggest an important and general role of lipophagy in mediating TAG hydrolysis under starvation conditions (Figure 13). TAG did not accumulate in atg mutants under extended darkness (Supplemental Figure 11). This result suggests that the SDP1-mediated cytosolic lipolytic pathway can functionally compensate for the lack of lipophagy in TAG hydrolysis under starvation.
Previous studies showed that plant autophagic organelles contain hydrolytic enzymes, including proteases and lipases, for cargo degradation at the onset of their formation (Marty, 1978, 1999; Buvat and Robert, 1979) and are functionally sufficient to break down the sequestered materials on their own (Rose et al., 2006; van Doorn and Papini, 2013). In accordance with the autophagosome-autonomous hydrolysis, our ultrastructural analysis showed that LDs and other cellular constituents were degraded in AVs (Figures 7D and 10C), in addition to the central vacuole (Figure 10E). These results point to the unique aspects of plant autophagy in comparison with this catabolic process in yeast and mammals, where the autophagosome itself lacks degradative enzymes and its cargo is broken down following fusion with lytic compartments such as vacuoles and lysosomes, respectively (Eskelinen, 2005; Suzuki and Ohsumi, 2007; Reggiori and Klionsky, 2013; Dikic, 2017; Galluzzi et al., 2017).
Our ultrastructural analysis showed that the autophagic degradation of LDs in Arabidopsis occurs in a process resembling microlipophagy in yeast. Disruption of autophagy genes increased TAG content in sdp1 under starvation conditions (Figure 12). These results suggest that microlipophagy in Arabidopsis depends on the core machinery of macroautophagy, similar to the situation in yeast (van Zutphen et al., 2014). At present, the exact mechanism underlying microautophagy and the role of ATG gene products in microlipophagy remain largely unknown (Noda and Inagaki, 2015; Galluzzi et al., 2017; Oku and Sakai, 2018). Our results showed that microautophagy-like LD degradation occurs in AVs, key autophagic structures in macroautophagy (Eskelinen, 2005). Therefore, it is possible that the observed dependence of starvation-induced TAG and LD accumulation on the macroautophagic machinery in Arabidopsis may simply reflect the essential role of core ATG proteins in the formation of autophagosomes and hence AVs. In support of this possibility, disruption of the core ATG genes blocks the formation of both AVs and microlipophagy (Figure 11). Recently, vacuolar membrane lipid rafts enriched in sterols have been shown to be necessary for microlipophagy in yeast (Oku and Sakai, 2018). Further studies are needed to test whether the sterol-enriched membrane rafts are involved in microlipophagy in plants, to determine how TAG is hydrolyzed in vacuoles, and to establish the regulation and physiological functions of lipophagy.
Methods
Plant Materials and Growth Conditions
The Arabidopsis (Arabidopsis thaliana) plants used in this study were of the Columbia ecotype. The tgd1 mutant was previously described by Xu et al. (2003), act1 mutant by Kunst et al. (1988), and sdp1 mutants by Fan et al. (2017). The PDAT1-overexpressing lines 3 and 4 were described in Fan et al. (2013b). The primers used for genotyping sdp1 were as described previously (Fan et al., 2014). The atg2-1 (SALK_076727) and atg5-1 (SAIL_129_B07) mutant lines were ordered from the Arabidopsis Biological Research Center at Ohio State University. The primers used for genotyping atg mutants are 5′-GTGGGGCTCATAGCTTAGACC-3′ and 5′-CACTTTCCATCAGCTACTCGC-3′ for atg2-1 and 5′-ATTTGCTATTTGTTTGGCACG-3′ and 5′-ATAATGGCAAACCAATTGCAG-3′ for atg5-1. Genotyping of tgd1 and act1 mutants was as described previously (Xu et al., 2005; Fan et al., 2015).
For plant growth in soil, surface-sterilized seeds of Arabidopsis were germinated on 0.6% (w/v) agar-solidified half-strength Murashige and Skoog (MS) medium (Murashige and Skoog, 1962) supplemented with 1% (w/v) Suc in an incubator with a photon flux density of 50 to 80 μmol m–2 s–1 (cool white lamps), a light period of 16 h (22°C), and a dark period of 8 h (18°C). After 10 d of growth, the seedlings were transferred to soil and grown under a photosynthetic photon flux density of 80 to 150 μmol m−2 s−1 (a combination of cool white fluorescent lamps and incandescent lamps) at 22/18°C (day/night) with a 16-h-light/8-h-dark period, unless stated otherwise. For starvation treatment, whole plants, unless stated otherwise in Figure 12A, were transferred to continuous darkness at 24°C for the time indicated.
Plasmid Construction
To construct the DsRed-ATG8e expression vector, the coding region of ATG8e was amplified with primers 5′-agaggtaccAATAAAGGAAGCATCttt-3′ and 5′-agaggatccTTAGATTGAAGAAGCAC-3′. To construct the tonoplast marker δTIP-DsRed, the coding sequence of At5g47450 was amplified using the primers δTIP-fw-Sac1, 5′-agagagctcATGGTGAAGATCGAAGTTGG-3′ and δTIP-Rv-kpn1, 5′-ctaggtaccCACTCGGATCTCACGGGTTT-3′. The PCR products were cloned into a binary vector pPZP212 (Fan et al., 2015) with DsRed fused to the N terminus of ATG8e or to the C terminus of δTIP. After confirming the integrity of the construct by sequencing, plant stable transformation was performed according to Clough and Bent (1998).
Lipid and Fatty Acid Analyses
Lipids were extracted from leaves of 4-week-old plants grown in soil as described by Fan et al. (2013a))). Polar and neutral lipids were separated on silica plates (Silica Gel 60, EMD Millipore) by thin layer chromatography (TLC) using acetone:toluene:water (91:30:7, by volume) and/or hexane:diethyl ether:acetic acid (70:30:1, by volume), respectively. To quantitate low TAG levels in leaves of wild type and atg mutants, total lipid extracts were first fractionated through silica columns (Discovery DSC-Si SPE tube, volume 6 mL, Supelco) as described by James et al. (2010), and the collected neutral lipid-containing fraction was separated by TLC predeveloped using methanol:chloroform (1:1, by volume). Fatty acid methyl esters were prepared as described by Li-Beisson et al. (2013). Separation and identification of the fatty acid methyl esters were performed on an HP5975 gas chromatograph-mass spectrometer (Hewlett-Packard) fitted with a 30 m × 250-μm DB-23 capillary column (Agilent) with helium as the carrier gas as described by Fan et al. (2013a). Fatty acid methyl esters were quantified using heptadecanoic acid as an internal standard as described by Fan et al. (2013a).
Immunoblot Analysis
Equal fresh weight of mature leaves of 4-week-old plants grown in soil was ground in liquid nitrogen, homogenized with 2× Laemmli sample buffer. The extracts were incubated for 5 min in boiling water and clarified by centrifugation at 12,000g for 5 min at 22°C. For ATG8 lipidation analysis, proteins were subjected to 15% SDS-PAGE with 6 M urea in the separating gel as described by Chung et al. (2010). For OLE1-GFP, proteins were subjected to 10% SDS-PAGE and blotted to a polyvinylidene difluoride membrane. Immunoblot analyses were performed according to the ECL Western Blotting procedure (32,106, Thermo Fisher Scientific) with antibodies against GFP (catalog no. 902603, lot no. E11LF02512, BioLegend), ATG8a (catalog no. AS142811, lot no. 1604, Agrisera), and actin (catalog no. MBS8500610, lot no. M14L06, MyBioSource). Targeted proteins were visualized using an ImageQuant LAS 4000 biomolecular imager (GE Healthcare Life Sciences).
Assays for Fatty Acid Synthesis and Degradation
In vivo labeling experiments with 14C-acetate or 3H2O were done as described previously by Fan et al. (2013a); Yu et al. (2018). Developing seeds of 50 siliques were directly harvested into labeling medium containing 20 mM MES, pH 5.5, one-tenth strength of MS salts, and 0.01% (v/v) Tween 20 on ice. For labeling, developing seeds or 4-d-old seedlings or detached leaves were incubated in the light of 80 µmol m−2 s−1 (cool white fluorescent lamps) at 22°C in 10 mL of the labeling medium. The assay was started by the addition of 0.1 mCi of 14C-acetate or 0.2 mCi 3H2O (both radiochemicals were from American Radiolabeled Chemicals). After incubation for 1 h, tissues were washed two times with water and immediately used for lipid extraction. For pulse-chase labeling experiments, leaves were labeled for 1 h with 14C-acetate. After washing three times with water, the leaves were incubated further with unlabeled solution under a 16-h-light/8-h-dark cycle for 3 d. Total lipids were extracted and separated as described previously by Fan et al. (2013a), and radioactivity associated with total lipids or different lipid classes was determined by liquid scintillation counting or phosphor imaging. Radiolabel loss was calculated by correcting for the dilution of radioactivity caused by tissue growth during the chase period.
PDAT Activity Assays
Microsomal membranes were isolated from 3-week-old seedlings as described previously (Xu et al., 2005). Radioactive PC for PDAT activity assays was prepared after incubating 2-week-old seedlings overnight in 20 mM MES-KOH, pH 6.0, with 0.2 mCi of 14C-acetate (American Radiolabeled Chemicals). Lipids were extracted and separated by TLC as described by Fan et al. (2013a). Radiolabeled PC was eluted from silica gel using chloroform:methanol:formic acid (1:2:0.1, by volume) and redissolved in chloroform. The reaction mixture contained 0.1 mg of microsomal proteins in 50 mM potassium phosphate buffer, pH 7.8, 250 µM 14C-labeled PC (56 MBq/mmol), and 250 µM 18:1-diacylglycerol (Avanti Polar Lipids) in a final volume of 200 µL. The reaction solution was thoroughly mixed and incubated at room temperature for 30 min. Lipid extraction and TLC separation were done as described previously by Fan et al. (2013a). Radioactivity in TAG was determined by scintillation counting.
Treatment with 3-MA
Detached leaves of 4-week-old plants grown in soil were floated on water with or without the addition of 5 mM 3-MA (dissolved in water, Sigma-Aldrich) and 0.01% (v/v) Tween 20 in the dark at 24°C. Samples were taken every 24 h over 4 d for lipid analysis as described previously (Fan et al., 2013a).
Microscopy
For LD imaging, leaf tissues were stained with a neutral lipid-specific fluorescent dye, Nile red (Sigma-Aldrich), or BODIPY493/503 at a final concentration of 10 or 5 µg/mL in PBS, pH 7.0, respectively, and observed under an epifluorescence microscope (Axiovert 200M, Carl Zeiss) with a GFP filter (Zeiss filter set 38). GFP was excited with a 470/40 nm band-pass filter and the emission was captured with a 525/50 nm band-pass filter.
For the colocalization study, leaf samples were mounted in water on slides and were directly examined using a Leica TCS SP5 laser scanning confocal microscope with sequential scanning. GFP was excited with a wavelength of 488 nm and detected at 500 to 530 nm. DsRed was excited at 543 nm and detected at 560 to 630 nm.
For tonoplast imaging, transgenic plants coexpressing OLE1-GFP and δTIP-DsRed were germinated on 0.6% (w/v) agar-solidified half-strength MS medium lacking Suc. Six-day-old seedlings were dark treated for 1 d and then transferred to half-strength MS medium with or without 0.5 µM concA and incubated in the dark for additional 20 h. The hypocotyls or cotyledons were observed under confocal microscopy.
For transmission electron microscopy, leaf tissues were fixed with 2.5% (v/v) glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.4, for 2 h and then postfixed with 1% (v/v) osmium tetroxide in the same buffer for 2 h at room temperature. After dehydration in a graded series of ethanol, the tissues were embedded in EPON 812 resin (Electron Microscopy Sciences), sectioned, and stained with 2% (w/v) uranyl acetate and lead citrate before viewing under a JEM-1400 LaB6 120-KeV transmission electron microscope (JEOL).
For chloroplast counting, leaf tissues were fixed and embedded. Thin sections produced from the embedded leaf tissues were stained with 1% (w/v) toluidine blue and imaged using an epifluorescence microscope (Axiovert 200M, Carl Zeiss). The number of chloroplasts was counted from at least 60 mesophyll cell cross sections for each time point of dark treatment.
Colocalization Analysis
Colocalization analysis of OLE1-GFP and ATG8e-DsRed signals was done with the Coloc 2 plugin for ImageJ. Background subtraction from image pairs was performed using rolling ball subtraction with a 50-pixel ball size. Statistical significance of the PCC of the image pairs was analyzed using the Costes image randomization test as described previously (Costes et al., 2004). Regions of interest were selected for colocalization analysis with 100 Costes randomizations using a point spread function of 3.
Immunogold Labeling
Five-day-old seedlings grown on 0.6% (w/v) agar-solidified half-strength MS medium with 1% (w/v) Suc were dark treated for 2 d. The seedlings were then transferred to half-strength MS medium containing 0.5 µM concA and incubated in the dark for 20 h. Hypocotyls of concA-treated seedlings were harvested and fixed overnight at 4°C in 4% (v/v) paraformaldehyde, 0.25% (v/v) glutaraldehyde, and 0.1 M cacodylate buffer, pH 7.4. The fixed hypocotyls were washed twice with 0.1 M cacodylate buffer and once in distilled water, each wash step for 30 min. The tissues were then dehydrated using a graded ethanol series (50, 70, 80, 95% (v/v), twice in 100% (v/v)) for 30 min each. After dehydration, the tissues were embedded in LR White resin (14381-CA, Electron Microscopy Sciences, London Resin Company) in gelatin capsules. Resin polymerization was performed at 50 to 55°C.
Ultrathin sections (70 to 90 nm) of LR White–embedded hypocotyls were collected with formvar-coated 300 mesh nickel grids. The grids were first washed with 1× PBS containing 0.2% (w/v) Gly two times, 3 min each, and then blocked with 1% (w/v) BSA and 0.2% (w/v) Gly for 30 min. After blocking, the grids were incubated with the primary antibody:rabbit polyclonal anti-ATG8a (catalog no. AS142811, lot no. 1604, Agrisera) of Arabidopsis (dilute 1:60, in blocking solution) at 4°C overnight in a wet chamber. After rinsing with blocking solution five times, 1 min each, the grids were then incubated in the secondary antibody of goat anti-rabbit immunoglobulin G conjugated with 10-nm gold particles (catalog no. G7402, lot no. SLBW9500, Sigma-Aldrich; 1:20 dilution in blocking solution) for 1 h at room temperature. Following washing with 1× PBS and 0.2% (w/v) Gly five times, 1 min each, the grids were immersed in a drop of 2.5% (v/v) glutaraldehyde solution for 30 min. After rinsing with distilled water two times, 2 min each, the grids were stained with 2% (w/v) uranyl acetate for 15 min and lead citrate for 3 min, and then observed using electron microscopy .
Accession Numbers
Sequence data from this article can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: ACT1, At1g32200; ATG2, At3g19190; ATG5, AT5g17290; ATG7, At5g45900; PDAT1, At5g13640; SDP1, At5g04040; TGD1, At1g19800; δTIP, At5g47450.
Supplemental Data
Supplemental Figure 1. Time course of the incorporation of radiolabel from 14C-acetate or 3H2O into total fatty acids in wild-type developing embryos.
Supplemental Figure 2. Rate of the incorporation of radiolabel from 14C-acetate or 3H2O into TAG in developing embryos and seedlings.
Supplemental Figure 3. Rate of the incorporation of radiolabel from 14C-acetate or 3H2O into total fatty acids in leaves.
Supplemental Figure 4. Rate of the incorporation of radiolabel from 14C-acetate into total membrane lipids in leaves.
Supplemental Figure 5. PDAT activity in microsomal membranes isolated from seedlings.
Supplemental Figure 6. Disruption of autophagy reduces TAG content in mature leaves of 4-week-old PDAT1-overexpressing transgenic plants.
Supplemental Figure 7. Increased accumulation of DsRed-ATG8e–labeled structures in leaves of tgd1 plants under dark treatment.
Supplemental Figure 8. Accumulation of autophagosomes and autophagic vacuoles in mature leaves of 4-week-old sdp1-4 plants under dark treatment.
Supplemental Figure 9. The appearance of LDs in the central vacuole in wild-type seedlings after dark treatment in the presence of concA.
Supplemental Figure 10. Autophagic activity in 4-week-old sdp1-4 plants under dark-induced starvation.
Supplemental Figure 11. TAG levels in mature leaves of 4-week-old wild type, atg2-1 and atg5-1 plants under dark-induced starvation.
Supplemental Figure 12. Membrane lipid levels in mature leaves of 4-week-old sdp1-4, atg2-1 sdp1-4, and atg5-1 sdp1-4 plants under dark-induced starvation.
Supplemental Figure 13. Chloroplast number in mature leaves of sdp1-4 plants under dark-induced starvation.
Supplemental Data Set. Results of statistical analyses.
Dive Curated Terms
The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:
Acknowledgments
This work was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences (DE-SC0012704), specifically through the Physical Biosciences program of the Chemical Sciences, Geosciences and Biosciences Division. Use of the transmission electron microscope and the confocal microscope at the Center of Functional Nanomaterials was supported by the Office of Basic Energy Sciences, U.S. Department of Energy (DE-SC0012704).
AUTHOR CONTRIBUTIONS
C.X. and J.F. designed the experiments. J.F., L.Y., and C.X. performed the research. J.F. and C.X. participate in data analysis. C.X. wrote the article with contributions from J.F. and L.Y.
Footnotes
- Received March 13, 2019.
- Revised March 26, 2019.
- Accepted April 19, 2019.
- Published April 29, 2019.